Mechanical Circulatory Support

for Advanced Heart Failure

A Texas Heart Institute/

Baylor College of
Medicine Approach
Jeffrey A. Morgan
Andrew B. Civitello
O.H. Frazier
Editors

123
Mechanical Circulatory Support
for Advanced Heart Failure
Jeffrey A. Morgan  •  Andrew B. Civitello
O.H. Frazier
Editors

Mechanical Circulatory
Support for Advanced
Heart Failure
A Texas Heart Institute/Baylor
College of Medicine Approach
Editors
Jeffrey A. Morgan, MD Andrew B. Civitello, MD
Division of Cardiothoracic Medical Director, Heart Transplant
Transplantation Program
Surgical Director, Mechanical Co-Director, Advanced Heart
Circulatory Support and Cardiac Failure Center
Transplantation Texas Heart Institute at Baylor
Texas Heart Institute at Baylor St. St. Luke’s Medical Center, Baylor
Luke’s Medical Center, Baylor College College of Medicine
of Medicine Houston, TX, USA
Houston, TX, USA

O.H. Frazier, MD
Center for Cardiac Support
Texas Heart Institute at Baylor St.
Luke’s Medical Center, Baylor College
of Medicine
Houston, TX, USA

ISBN 978-3-319-65363-1    ISBN 978-3-319-65364-8 (eBook)

[Link]

Library of Congress Control Number: 2017957648

© Springer International Publishing AG 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or
part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way,
and transmission or information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in
this book are believed to be true and accurate at the date of publication. Neither the publisher nor
the authors or the editors give a warranty, express or implied, with respect to the material
contained herein or for any errors or omissions that may have been made. The publisher remains
neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
We would like to dedicate our book Mechanical Circulatory Support for
Advanced Heart Failure: A Texas Heart Institute/Baylor College of
Medicine Approach to Dr. Denton A. Cooley.
Dr. Cooley is considered to be the world’s greatest heart surgeon. His
accomplishments include expanding therapeutic potential for patients with
congenital heart conditions, pioneering the artificial heart and heart
transplantation, developing prosthetic heart valves, and establishing new
methods for aortic aneurysm repair. In the words of Dr. Walt Lillehei,
“Dr. Cooley was the first to demonstrate the safety of heart surgery with the
heart-lung machine. He performed more heart surgery than any heart
surgeon in the world every year from 1956 to 1994.” Dr. Cooley also
developed the first bundled services plan for cardiac surgery, called the
CardioVascular Care Providers, which was influential in the structuring
of cardiac services for Medicare.
Dr. Cooley founded the Texas Heart Institute (THI) in 1962 and was
instrumental in THI rising to become one of the premier institutions for
cardiac surgery in the world. Over 120,000 cardiac surgeries using the
cardiopulmonary bypass circuit were performed at THI during Dr. Cooley’s
lifetime. Dr. Cooley published over 1400 scientific articles and was a
member in more than 30 professional medical societies. He founded the
Cullen Cardiovascular Surgical Research Laboratory, which under
Dr. O.H. Frazier’s leadership, a trainee and devotee of Dr. Cooley, was
instrumental in developing nearly all of the left ventricular assist devices
used in clinical practice today. Among his numerous honors and awards,
Dr. Cooley received the Presidential Medal of Freedom from President
Reagan in 1984 and the National Medal of Technology and Innovation from
President Clinton in 1998, as well as the Lifetime Achievement Award in
2016 from the American Association for Thoracic Surgery.
It is our belief that every cardiologist, cardiac surgeon, and cardiac
patient owes a great degree of gratitude to Dr. Cooley for his enormous
contribution to the field. We are greatly honored to have been given the
opportunity to dedicate our book to the memory of the late Dr. Cooley.
Respectfully,
Jeffrey A. Morgan, M.D.; Andrew B. Civitello, M.D.; and O.H. Frazier, M.D.
Foreword

I am proud and honored to have been asked to write this foreword. It seems
only fitting that a book about mechanical circulatory support (MCS) should
be published by experts from Baylor College of Medicine and the Texas
Heart Institute (THI). Since the 1960s, these two institutions, at first sepa-
rately and now jointly, have been involved in almost every major advance in
this field.
Not so long ago, a book written in collaboration between THI and Baylor
physicians would have been unimaginable. In 1969, professional rivalry
between myself and Dr. Michael E. DeBakey, chairman of Baylor’s
Department of Surgery, caused me to resign my long-standing professorship
at Baylor and devote my full attention to THI, which I had founded in 1962.
Baylor and THI each continued to make outstanding contributions to cardio-
vascular medicine, but they lacked the advantage of a mutually beneficial
collaboration. Not until 2007 was a cordial relationship reestablished.
Instrumental in that reconciliation were Dr. George P. Noon of Baylor, Dr.
O.H. Frazier of THI, and several other physicians at both institutions. In late
2007, Dr. DeBakey joined me in the THI research laboratory to watch Dr.
Frazier implant a total artificial heart into a calf. The heart comprised dual
MicroMed DeBakey left ventricular assist devices. This occasion marked a
breakthrough in both MCS research and Baylor-THI relations. By the time of
Dr. DeBakey’s death, at age 99 in July 2008, the new rapport was firmly
established.
In a modest way, this rapprochement might be compared to the ending of
the twentieth-century “space race” between the US astronauts and the Soviet
cosmonauts. Elsewhere, I have related how the space race influenced my
response to the unique scientific challenge posed by the first TAH implanta-
tion [1]. With the end of the Cold War, former rivalries were laid aside, and
old boundary lines were dissolved. Since then, unprecedented spaceflight
cooperation between the USA and Russia has led to progress in education,
research, and technology. Today, unprecedented cooperation between Baylor
and THI is leading to advances in education, research, and patient care. The
current book is a result—and a symbol—of that cooperation.
I congratulate Drs. Morgan, Civitello, and Frazier and all the other con-
tributors to this superb volume, which covers every aspect of clinical cardiac
support. The experience related here is based on the largest single-center
MCS series in the USA. As a clear, comprehensive, and authoritative guide to
device therapy, this book will be an indispensable resource for physicians,

vii
viii Foreword

other medical personnel, and anyone else interested in support of the failing
heart.

Houston, TX, USA Denton A. Cooley, M.D.

Reference
1. Cooley DA. Some thoughts about the historic events that led to the first clinical implan-
tation of a total artificial heart. Tex Heart Inst J. 2013;40(2):117–119.
Preface

The first successful LVAD was implanted by Dr. DeBakey at Baylor College
of Medicine/Methodist Hospital in 1966. In 1968, Dr. Denton Cooley per-
formed the first successful human heart transplant in the USA at Texas Heart
Institute, St. Luke’s Hospital. Dr. Cooley subsequently performed the first
successful artificial heart implantation in 1969 at the Texas Heart Institute.
The first LVAD as a bridge to transplant and the first combined heart/kidney
transplant were also performed by Dr. Cooley in 1978 at the Texas Heart
Institute. In 1988, Dr. Frazier implanted the first successful continuous-flow
LVAD and has subsequently been instrumental in the development of nearly
all continuous-flow devices used clinically, including the Jarvik, HeartMate
2, HeartMate 3, and HeartWare HVAD.
With the popularization of continuous-flow LVADs, mechanical circula-
tory support has evolved into the standard of care for patients with refractory,
end-stage heart failure. Advancements in patient selection, device design,
surgical techniques, and postoperative management have led to significant
improvements in survival and a reduction in device-related complications,
such as bleeding, infection, stroke, device malfunction, and device
thrombosis.
Each chapter in our text Mechanical Circulatory Support for Advanced
Heart Failure: A Texas Heart Institute/Baylor College of Medicine Approach
was authored by staff members from the Texas Heart Institute, Baylor College
of Medicine. Our LVAD program has grown significantly over the years with
greater than 1300 LVADs implanted to date, including over 850 continuous-
flow LVADs. Our goal in writing this text was to provide a framework for
physicians evaluating patients for LVADs, caring for patients perioperatively,
and/or managing patients with LVADs long-term by sharing the cumulative
experience of the Texas Heart Institute, Baylor College of Medicine LVAD
program.

Houston, TX, USA Jeffrey A. Morgan, M.D.

 Andrew B. Civitello, M.D.
 O.H. Frazier, M.D.

ix
Contents

1 History of Mechanical Circulatory Support��������������������������������    1

O.H. Frazier
2 Who Is an Appropriate Candidate for Long-Term
MCS?: The Art of Patient Selection ��������������������������������������������   15
Carol S.C. Lai and Andrew B. Civitello
3 Optimization of Right Ventricular Function Preoperatively
for LVAD Implantation������������������������������������������������������������������   35
Salman Gohar, Samar Sheth, and Reynolds Delgado III
4 Bridge-to-Bridge Strategies with IABP, Impella,
and TandemHeart��������������������������������������������������������������������������   57
Samar Sheth, Salman Bandeali, and Joggy George
5 LVAD Surgical Implant Technique: Extraperitoneal
Approach����������������������������������������������������������������������������������������   69
Rachel A. Beaupré, Gabriel Loor, and Jeffrey A. Morgan
6 LVAD Surgical Implant Technique: Infradiaphragmatic
Approach����������������������������������������������������������������������������������������   79
Jeffrey A. Morgan and O.H. Frazier
7 Intraoperative Anesthesia Management��������������������������������������   83
Marissa Wagner Mery, Siavosh Saatee, and Charles D. Collard
8 Perioperative Management of LVAD Patients����������������������������   95
Krishna Ayyagari, William Patrick Mulvoy, Arthur W. Bracey,
Cesar A. Castillo, and James P. Herlihy
9 Management of Fluid Balance and Perioperative Renal
Complications��������������������������������������������������������������������������������  129
Whitson B. Etheridge and Sarah A. Shearer
10 Chronic Management of Patients with Left Ventricular
Assist Devices����������������������������������������������������������������������������������  145
Luke C. Cunningham and Ajith P. Nair
11 Surveillance Echocardiography for LVAD Patients��������������������  161
Raymond F. Stainback

xi
xii Contents

12 Diagnosis of  Device Thrombosis ��������������������������������������������������  191

Cyril Varughese, Ajith P. Nair, and Jordan Chaisson
13 Device Exchange: THI Technique Involving
a Left Subcostal Approach������������������������������������������������������������  199
Tadahisa Sugiura and Masashi Kawabori
14 Mechanical Circulatory Support to Bridge
to a Long-Term Continuous-Flow Left Ventricular
Assist Device as a Bridge to Heart Transplantation ������������������  203
Chitaru Kurihara
15 Outcomes Using LVADs for Destination Therapy����������������������  209
Priyanka Sen and Selby Oberton
16 Cardiac Regenerative Strategies for Advanced Heart Failure��������  221
Vivekkumar B. Patel, Megumi Mathison, Vivek Singh,
Jianchang Yang, and Todd K. Rosengart
17 Long-Term Complications of Ventricular Assist Devices ����������  239
George V. Letsou
18 Right Ventricular Failure in Patients Undergoing
LVAD Placement����������������������������������������������������������������������������  251
David Kuten and Joggy K. George
Index��������������������������������������������������������������������������������������������������������   263
Contributors

Krishna Ayyagari, M.D.  Texas Heart Institute at Baylor St. Luke’s Medical

Center, Baylor College of Medicine, Houston, TX, USA
Salman Bandeali, M.D.  Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Rachel A. Beaupré, M.D.  University of Cincinnati, Cincinnati, OH, USA
Arthur W. Bracey, M.D.  Clinical Pathology, Texas Heart Institute at Baylor
St. Luke’s Medical Center, Baylor College of Medicine, Houston, TX, USA
Cesar A. Castillo, M.D.  Anesthesia Critical Care, Texas Heart Institute at
Baylor St. Luke’s Medical Center, Baylor College of Medicine, Houston,
TX, USA
Jordan Chaisson, M.D.  Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Andrew B. Civitello, M.D.  Heart Transplant Program, Texas Heart Institute
at Baylor St. Luke’s Medical Center, Baylor College of Medicine, Houston,
TX, USA
Advanced Heart Failure Center, Texas Heart Institute at Baylor St. Luke’s
Medical Center, Baylor College of Medicine, Houston, TX, USA
Charles D. Collard, M.D. Division of Cardiovascular Anesthesiology,
Texas Heart Institute at Baylor St. Luke’s Medical Center, Baylor College of
Medicine, Houston, TX, USA
Luke C. Cunningham, M.D. Texas Heart Institute at Baylor St. Luke’s
Medical Center, Baylor College of Medicine, Houston, TX, USA
Reynolds Delgado III, M.D. Mechanical Assist Devices, Texas Heart
Institute at Baylor St. Luke’s Medical Center, Baylor College of Medicine,
Houston, TX, USA
Whitson B. Etheridge, M.D.  Renal Transplantation, Texas Heart Institute at
Baylor St. Luke’s Medical Center, Baylor College of Medicine, Houston, TX,
USA
O.H. Frazier, M.D. Center for Cardiac Support, Texas Heart Institute at
Baylor St. Luke’s Medical Center, Baylor College of Medicine, Houston, TX,
USA

xiii
xiv Contributors

Mechanical Circulatory Support and Cardiac Transplantation, Texas Heart

Institute at Baylor St. Luke’s Medical Center, Baylor College of Medicine,
Houston, TX, USA
Joggy K. George, M.D.  Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Salman Gohar, M.D. Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
James P. Herlihy, M.D. Critical Care Services, Texas Heart Institute at
Baylor St. Luke’s Medical Center, Baylor College of Medicine, Houston,
TX, USA
Masashi Kawabori, M.D. Cardiothoracic Transplant and Mechanical
Circulatory Support, Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Chitaru Kurihara, M.D. Cardiothoracic Transplant and Mechanical
Circulatory Support, Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
David Kuten, M.D. Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Carol S.C. Lai, M.D.  Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
George V. Letsou, M.D.  Texas Heart Institute, Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Gabriel Loor, M.D.  Lung Transplantation, Texas Heart Institute at Baylor
St. Luke’s Medical Center, Baylor College of Medicine, Houston, TX, USA
Megumi Mathison, M.D., Ph.D. Division of Surgical Research, Baylor
College of Medicine, Houston, TX, USA
Marissa Wagner Mery, M.D. Texas Heart Institute at Baylor St. Luke’s
Medical Center, Baylor College of Medicine, Houston, TX, USA
Jeffrey A. Morgan, M.D.  Division of Cardiothoracic Transplantation, Texas
Heart Institute at Baylor St. Luke’s Medical Center, Baylor College of
Medicine, Houston, TX, USA
Mechanical Circulatory Support and Cardiac Transplantation, Texas Heart
Institute at Baylor St. Luke’s Medical Center, Baylor College of Medicine,
Houston, TX, USA
Center for Cardiac Support, Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
William Patrick Mulvoy III, M.D., M.B.A.  Texas Heart Institute at Baylor
St. Luke’s Medical Center, Baylor College of Medicine, Houston,
TX, USA
Contributors xv

Ajith P. Nair, M.D. Texas Heart Institute at Baylor St. Luke’s Medical

Center, Baylor College of Medicine, Houston, TX, USA
Selby Oberton, M.D. Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Vivekkumar B. Patel, M.D. Texas Heart Institute at Baylor St. Luke’s
Medical Center, Baylor College of Medicine, Houston, TX, USA
Todd K. Rosengart, M.D.  Department of Surgery, Texas Heart Institute at
Baylor St. Luke’s Medical Center, Baylor College of Medicine, Houston,
TX, USA
Siavosh Saatee, M.D.  Mechanical Circulatory Support Critical Care, Texas
Heart Institute at Baylor St. Luke’s Medical Center, Baylor College of
Medicine, Houston, TX, USA
Priyanka Sen, M.D. Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Sarah A. Shearer, M.D.  Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Samar Sheth, M.D. Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Vivek Singh, Ph.D. Division of Surgical Research, Baylor College of
Medicine, Houston, TX, USA
Raymond F. Stainback, M.D. Non-invasive Cardiology, Texas Heart
Institute at Baylor St. Luke’s Medical Center, Baylor College of Medicine,
Houston, TX, USA
Tadahisa Sugiura, M.D. Cardiothoracic Transplant and Mechanical
Circulatory Support, Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Cyril Varughese, D.O.  Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine, Houston, TX, USA
Jianchang Yang, M.D., Ph.D. Division of Surgical Research, Baylor
College of Medicine, Houston, TX, USA
 History of Mechanical Circulatory
Support 1
O.H. Frazier

This introduction focuses on the role of the The meaningful pursuit of heart replacement
Baylor College of Medicine (BCM) and the began in 1964 when Dr. Michael DeBakey secured
Texas Heart Institute (THI) in the evolution and funding, mainly through the auspices of then
development of heart replacement and circula- President Lyndon B. Johnson, to pursue the devel-
tory assist technology. This is appropriate because opment of an artificial heart. It was unusual for the
both the first successful LVAD and the first suc- National Institutes of Health (NIH) to support
cessful artificial heart were implanted at these such a project; in general, they confined their
institutions in Houston, Texas. In addition, the grants to pure research without any immediate
initial experimental work on the continuous-­flow clinical objective. So, this funding was unique in
pumps now in use (Jarvik, HeartMate II, that regard and probably would not have been
HeartWare, Impella) began at our institute. My granted without Dr. DeBakey’s leadership. Also, I
experience has been unique in this regard, as I remember well those heady times, when we were
have been personally involved in this journey going to the moon, among other grandiose objec-
from 1963 to the present. My only absence was tives. Creating an artificial heart, comparatively
during 1968–1970, when I served with an assault speaking, seemed like a simple side project.
helicopter company engaged in active combat in The funding for the artificial heart went
the central highlands of Vietnam. In this same primarily to BCM, where I was then a medical
period (April 1969), Dr. Denton Cooley “relo- student. During that time, BCM required us to
cated” Dr. DeBakey’s artificial heart from BCM’s participate yearly in research projects as part of
labs to THI-St. Luke’s Hospital and successfully our medical school education. Although I had no
implanted it as the first bridge to transplant with particular interest in surgery, my research proj-
an artificial heart (or any device) (Fig. 1.1). ects, by sheer chance, began in 1963 with Dr.
Thereafter, Dr. DeBakey and Dr. Cooley did not Domingo Liotta, who was developing heart
speak to each other for more than 30 years. My replacement pumps. Dr. Liotta was mainly inter-
friends who were in Houston at the time assured ested in the total artificial heart [1] but was pri-
me that Vietnam was probably a safer place for marily occupied with developing temporary
me to be. left ventricular assist devices (LVADs). This
work was initiated by Dr. DeBakey in 1964; it
continued after 1972 in the THI research labs.
This research initially was dedicated exclusively
O.H. Frazier, M.D. (*) to pulsatile pumps. By 1989, the NIH had spent
Center for Cardiac Support, Texas Heart Institute at
more than $266 million developing pulsatile
Baylor St. Luke’s Medical Center, Baylor College of
Medicine, Houston, TX, USA pumps, and the companies contracted to develop
e-mail: ofrazier@[Link]; lschwenke@[Link] this technology had spent at least as much. In all,

© Springer International Publishing AG 2018 1

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
2 O.H. Frazier

By the early 1980s, it seemed to me that the

limiting factor in developing pulsatile pumps
might be as simple as the durability of the mem-
branes. The normal heart of an inactive adult beats
approximately 100,000 times every 24 h. This
poses quite a challenge to the membrane technol-
ogy in pulsatile pumps, as well as to the additional
technological complexity that a completely
implantable total artificial heart would require.
The technological challenge of making this
device fully implantable was further compounded
by the fact that the left and right ventricles do not
pump the same amount of blood. The left heart
receives blood directly from the bronchial artery
circulation; thus, in the normal adult, the amount
of blood ejected from the left heart with each
heartbeat is 1–2 cc more than the right heart [4].
This is not much, but in the course of a 24 h period,
the difference amounts to more than 100,000 cc.
This necessitated finding a way for a totally
implantable artificial heart to adjust automatically
for the imbalance between the left and right flow.
Fig. 1.1  Domingo Liotta and the Liotta Artificial Heart The AbioCor total artificial heart addressed this
(1969) problem primarily by shifting the blood to the
right side when the left-sided pressures became
elevated [5]. Although this solution seemed effec-
probably more than $450 million was spent by tive in the short term, its long-term application was
the NIH and an equal amount by the private com- never tested beyond one 17-month survivor.
panies on developing pulsatile pumps [2]. At the The durability of the membranes seemed to be
time, the pulsatile pumps seemed to be logical limited to about 24 months in the pump made by
candidates for both temporary and total artificial Thermo Cardiosystems, Inc. (TCI) and a bit longer
heart development. in the Novacor pump. The Jarvik 7 total artificial
When I completed my surgery training at heart had a similar durability problem.
BCM in 1974, I renewed my direct involvement By the mid-1980s, it became apparent to me
with the development of cardiac replacement that the best approach to the durability and flow
pumps at THI. The work at THI at that time imbalance problems would be a continuous-flow
(1972–1980) was directed by Dr. Jack Norman, heart pump. Continuous-flow pumps are inher-
a capable Harvard-trained physician. This was ently inflow sensitive in that the higher the inflow
the only research on pumps being conducted in becomes, the more they will pump (if the outflow
the Texas Medical Center at the time; Dr. resistance is constant) without increasing the
DeBakey had suspended work on the artificial pump speed (Fig. 1.2). This would allow more or
heart in 1969 after his dispute with Dr. Cooley. less a physiologic Starling-type response, as well
Under Dr. Norman’s direction, we implanted 22 as physiologic adjustment, to control flow imbal-
intra-­
abdominal LVADs between 1976 and ance between the right and left heart.
1979, and one of the patients became the first to However, probably the most pressing reason to
be bridged to transplant with an LVAD [3]. pursue implantable long-term continuous-flow
Unfortunately, none of the 22 patients were pumps was the durability problem. I realized that if
long-term survivors, but the pump itself worked a pump had only a 2-year life span, the pump could
well in all cases. serve only as a prolonged bridge to transplant;
1  History of Mechanical Circulatory Support 3

10,000 rpm
14
HeartMate II
Jarvik 2000
12

10
Flow rate (L/min)

0
0 5 10 15 20 25 30 35 40
Preload (mmHg)

Fig. 1.2  Inflow sensitivity results in a Starling-like response without changing the pump speed

therefore, although the device could be lifesaving of such implantable continuous-flow pump
in individual cases, it would have no epidemiologic technology cited numerous potential problems,
impact on the heart failure population. The prob- mechanical as well as physiologic. The physio-
lem of changing the pump every 2 years, or else logic aberration of the baroreceptor response and
simply adding another patient to the transplant potential disruption of the juxtaglomerular
list, was and remains a barrier to the further devel- response were only a few of the many physiologic
opment of pulsatile pump technology. changes that would be produced by implantable
I had become interested in continuous-flow long-term continuous-flow pumps.
pumps in the late 1970s and early 1980s, when I In addition to these physiologic challenges,
used the Biomedicus pump (a constrained vortex there were two important engineering barriers
continuous-flow pump) in my extracorporeal that were thought to be insurmountable. In the
membrane oxygenation patients, as well as for mid-1980s, the only type of implantable
temporary LVAD support. I had used this pump continuous-­flow pump available used axial-flow
in 1987 in a 9-year-old patient who became the technology. Axial-flow pumps require bearings,
first pediatric patient to be bridged to transplant. and you could not have a nonlubricated bearing
Using this device not only enabled our patient to in the bloodstream (or anywhere else)—at least,
survive, but as I stated in the discussion of the that was the conventional thought. This was an
report of the case, it also “prompted us to specu- engineering axiom. (In fact, the only nonlubri-
late about broader application of nonpulsatile cated bearings I know to be in use today are those
flow, to the development of fully implantable in axial-flow blood pumps.) In addition, the pump
devices for long-term cardiovascular support of speed required to produce significant flow seemed
the terminal heart disease patient….The potential to be, by definition, a barrier to using axial-flow
for long-term benefit lies in meeting the require- technology: Speeds of more than 2500 rpm in a
ments of the circulatory system with a nonpulsa- small device were believed to be too damaging to
tile pump [italics added]” [6]. the blood (the “Waring blender effect”), causing
Making a continuous-flow pump implantable too much hemolysis to have any practical value
seemed to be a significant challenge. During that in producing meaningful blood flow.
era, I became involved in numerous debates and At a National Heart, Lung, and Blood Institute
discussions at meetings on this subject. Skeptics (NHLBI) contractor’s meeting in Louisville,
4 O.H. Frazier

Kentucky, in 1985, I was approached (separately) transplant survivor. We used this pump in several
by Drs. Richard Wampler and Robert Jarvik. more patients, with excellent results [8].
Although they were not acquainted, they were The Hemopump became the first implantable
both looking independently at engineering continuous-flow pump to be presented to the US
solutions to this problem. Dr. Wampler showed Food and Drug Administration (FDA) for
me his concept for a temporary implantable approval. It was developed without any NIH
continuous-­ flow device that would spin at funding. I funded the laboratory work (done in
25,000 rpm (although at the time I thought he had my lab), and Nimbus, a small research company,
said 2500 rpm). Shortly afterward, Dr. Jarvik funded the manufacture of the pump. The com-
showed me an implantable, long-term axial-flow pany received the bulk of its money from inves-
pump that would use blood-washed bearings. I tors who, naturally, wanted to apply this pump to
recommended to Dr. Jarvik that this smaller the largest patient population possible. Therefore,
pump be placed in the ventricle to avoid the inlet for the initial clinical trial of this device (which
problems that had plagued the pulsatile pumps. I had excellent results), the entry criterion was
agreed with both of these investigators, indepen- heart failure of any cause. The FDA, however,
dently, to proceed with this research in our labs at wanted more precisely defined entry criteria, and
THI. (Although I am not sure I would have pro- they recommended performing a new trial with
ceeded with Wampler’s design had I really under- such criteria. However, rather than fund further
stood that it spun at 25,000 rpm!) studies, the venture capitalists withdrew their
Initial work with what Dr. Wampler called the funding and invested in a more profitable stent
Hemopump was very promising. This small technology.
pump—the size of the eraser on a #2 lead pencil Fortunately, I found new support for the devel-
(Fig.  1.3)—could produce 4–5 L of outflow. opment of continuous-flow pumps in Helmut
Furthermore, the device caused only minimal Reul, a German friend of mine who earned, at the
hemolysis in the experimental animal. Because University of Houston, what was probably the
of these promising experimental findings, we first Ph.D. in bioengineering. I had met him dur-
introduced this pump clinically in April 1988 in a ing his time in Houston, after which he had
patient dying of heart allograft transplant rejec- returned to Aachen, Germany, and initiated a
tion [7]. We were able to support this patient with research program. At a medical meeting in
the Hemopump for 5 days, during which time we Germany in 1994, I advised him of the potential
reversed his organ rejection. The patient survived this of the Hemopump technology and that it would
potentially mortal event and became a long-­term not be further pursued in the United States.
Subsequently, at his research base in Aachen, he
began developing similar technology based on
the Hemopump principle. The resulting device
subsequently was acquired by the Abiomed com-
pany in Boston and is now in widespread use as
the Impella pump, a temporary assist device.
Dr. Jarvik began working on long-term
implantable axial-flow pump technology in my
lab in 1985. The development was much more
challenging than it had been for the Hemopump.
The first few pumps made by Dr. Jarvik lasted
only a short time before the nonlubricated
bearing would accumulate debris and occlude the
Fig. 1.3 The Hemopump, a tiny axial-flow pump
pump. However, after many revisions and experi-
designed to provide temporary circulatory support mental animal implantations, Dr. Jarvik produced
1  History of Mechanical Circulatory Support 5

Fig. 1.4  Drawing of an

implanted Jarvik 2000.
This device was first
placed in a patient in
April 2000

a workable nonlubricated, blood-washed bearing Company, which was a very small research com-
in an axial-flow pump by the early 1990s [9]. pany based in Sacramento, California (Fig. 1.4).
This research showed the feasibility of At that time, the engineering leader at Nimbus
continuous-­ flow implantable pumps for both was Dr. John Moise, a recognized expert and one of
long-term and temporary use. All of this research the best engineers in his field. He was struggling to
on continuous-flow pumps was funded internally develop a magnetically levitated axial-­flow pump.
with personal research funds of mine, by the At that time (the early 1990s), we were a small
Nimbus Company, and by Dr. Jarvik’s company, group—never more than 20 people—and we
Jarvik Heart, Inc. No NIH funding supported the worked collegially with one another. I had shared
feasibility studies performed in the 1980s and Dr. Wampler’s research success with the
early 1990s. This work formed the foundation for Hemopump with Dr. Jarvik, and I thought nothing
all future clinical applications of continuous-flow of doing the same with Dr. Jarvik’s success with
blood pumps. blood-washed, nonlubricated bearings. Our pri-
Soon after the initial clinical success of the mary goal was to make a pump that would ulti-
Hemopump, the Nimbus Company also became mately benefit patients. I had never thought of or
involved with the development of an implantable had any business interest in any of these projects.
long-term continuous-flow pump. Because I was I suggested to Dr. Moise that they put bear-
the only clinician involved in developing this ings on the rotor and not continue with the then
technology at that time, I was the medical advisor futile attempts at creating a maglev axial-flow
for both Dr. Jarvik’s company1 and the Nimbus pump. He replied, politely, that I did not know
anything about engineering and that you could
not have a nonlubricated bearing in the blood-
It may be of interest to note that Dr. Jarvik’s company
1 

initially consisted of only Dr. Jarvik and his wife, Marilyn

vos Savant, famed for having the highest recorded IQ it without doubt the company with the highest average IQ
according to the Guinness Book of World Records, making in the world.
6 O.H. Frazier

stream. But we had already shown in the Jarvik him that I had been working experimentally and
pump that blood-­washed bearings were possible, clinically with a short-term centrifugal force
so I stated that I did not know that it could not be continuous-flow pump and if we could develop a
done, that Dr. Jarvik did not know that it couldn’t long-term, magnetically levitated, bearingless,
be done, and that, most importantly, there was a implantable pump, it would potentially be an
calf in Houston that had had the pump for more important advancement in the field. I believed
than 8 months and that seemed not to know that this because such a pump would not require the
it couldn’t be done. At that point, the Nimbus controversial blood-washed bearings at all. Even
Company began working on what is now known though Dr. Jarvik had shown the feasibility of
as the HeartMate II. blood-washed bearings, bearings still had the
In closing this section, I would be remiss in potential for wear. And although I anticipated
not emphasizing that this whole field (implant- that these pumps would last far longer than the
able continuous-flow pump technology) was pulsatile pumps, I believed they would have a
initiated primarily by the engineering work of finite life span of 5–10 years. (This has proved to
two individuals. Dr. Wampler showed that you be erroneous, because these pumps have now
could, in fact, use a pump speed of not only more been in patients for longer than 10 years, and
than 2500 rpm but up to 25,000 rpm in the blood- none of those that were properly fabricated
stream without causing hemolysis. Dr. Jarvik’s and implanted have been pumped to failure.)
seminal contribution of creating a nonlubricated However, I did see (and continue to see) the
bearing was essential for the development of all advantages of a magnetically suspended, bear-
axial-flow implantable continuous-flow pumps. I ingless centrifugal force pump. A particularly
was privileged to work on both of these projects important advantage of this type of pump was
and have been fortunate to introduce both into the that it could be easily implanted intrapericardi-
clinical arena. More than 40,000 continuous-flow ally and therefore could be used for long-term
blood pumps have now (as of mid-2017) been right-sided, as well as left-sided, support.
implanted in otherwise mortally ill patients. Before this time, we had no right-sided long-­
Other than the three of us, there was no one, to term implantable pumps. The axial-flow pumps
my knowledge, actively pursuing implantable did not seem easily applicable to right-sided
continuous-flow pumps in the experimental ani- support, although I used a Jarvik pump success-
mal at that time (the mid 1980s). fully (in 2003) in the first patient to receive biven-
tricular implantable pump support [10]. I knew
the centrifugal force pump could be made flat so
 evelopment of Magnetically
D that it could easily fit inside the pericardium. Dr.
Levitated Centrifugal Force Fine asked me to recommend an engineer who
Continuous-Flow Pumps could work with him on this project, and I told
him that, in fact, there were only two engineers in
The investor who initiated funding for the world qualified to do so: Rich Wampler and
continuous-­ flow, centrifugal force, bearingless Rob Jarvik. Although Dr. Jarvik was busy further
pumps was Dr. Robert Fine, who, after earning developing his long-term pump, Richard
his medical degree, had also obtained a master’s Wampler had more freedom because the Nimbus
degree in business and became a Wall Street bro- Company, for which he worked, was no longer
ker specializing in medical investments. I had involved with the Hemopump.
met him as a result of this involvement. He had Dr. Wampler subsequently began working on
successfully invested in the first pump to be what ultimately became the first implantable cen-
approved (in 1994) by the FDA, the TCI pneu- trifugal force pump, known today as the
matic LVAD (which was developed in our facil- HeartWare. The company, originally called
ity). Dr. Fine then came to me and asked what I Kriton, reformed in the early 2000s and was
thought would further advance the field. I told renamed HeartWare, Inc. HeartWare began intro-
1  History of Mechanical Circulatory Support 7

ducing its device clinically in Australia and After this change was made, I implanted this
Europe in 2005. Implantation of these pumps in iteration of the HeartMate II in experimental ani-
the United States began in 2008, and this device mals. After success in this, I implanted the first
became the first FDA-approved magnetically HeartMate II clinical pump in November 2003.
levitated rotary pump. It proved to be easily This experience, I feel, is important to detail, as it
applicable to both right and left ventricular sup- shows the difficulty in developing these pumps.
port. This pump has subsequently received wide- The slightest even seemingly inconsequential
spread clinical acceptance and is recognized as mistake may, despite good experimental results,
an important contribution to the field. turn into a clinical failure.
Shortly after my encounter with Dr. Fine, I This was a well-run company, and once the
was at a meeting with Victor Poirier and Kurt HeartMate II was clinically reintroduced in 2003,
Dasse, who were, at that time, the leading engi- it subsequently became the most widely used of
neers with TCI. I had worked with them for more all continuous-flow pumps. To date, more than
than a decade on developing pulsatile pumps. I 25,000 patients worldwide have been implanted
suggested to them to also start looking at a mag- with this device. Of these, 196 were supported
netically levitated centrifugal-force pump. These for more than 8 years, including 135 patients who
two capable engineers began working on this had the same device (i.e., never required pump
project in the late 1990s. Their work eventually exchange) for that entire period.
resulted in a short-term pump, the CentriMag, Another reason for the success of this pump is
and an implantable maglev pump, the HeartMate that its inflow cannula acts as a relative restrictor to
III, being clinically introduced. pump inflow. This factor is important in ensuring a
As noted earlier, the pump now known as the satisfactory reservoir, which is important for limit-
HeartMate II began with John Moise and the ing inflow turbulence. Also, the position of the
Nimbus Company, which eventually was absorbed inlet cannula, designed by Vic Poirier, ensures that
into Thoratec. This pump underwent further devel- the cannula moves with the motion of the heart,
opment at Pittsburgh Medical School. thereby giving it further protection from pump
Implantations began in Europe and Israel, with inlet turbulence and consequent pump failure.
poor results. Vic Poirier brought me the pump. I
pointed out that they had placed sintered titanium
on the inside of the pump, causing it to become  linical Application of Rotary Blood
C
coated with a cellular layer and resulting in plate- Pumps
let activation. Both of these factors increased the
potential for pump thrombosis. After their feasibility was demonstrated in our
The layering of the cellular elements, par- lab, these pumps went directly to the manufactur-
ticularly mast cells, on the sintered titanium ers: Jarvik Heart, Thoratec for the HeartMate II,
was well demonstrated in the early experience and HeartWare for the centrifugal force
with the initial pulsatile pumps. However, the HeartWare pump. (The implantable Impella
cellular layering was important in avoiding pump, a descendent of the Hemopump, was sub-
anticoagulation in these large pulsatile pumps. sequently bought by Abiomed and is widely used
However, the much smaller continuous-flow for short-term support.) More than 150 hospitals
pumps like the HeartMate II had little clear- in the United States alone are implanting these
ance. Therefore, the cellular layer formed was pumps. Another major reason for the widespread
obstructive, and the increased turbulence and use of these small pumps was their ease of
shear stress thus engendered promoted implantation, which was far greater than that of
increased platelet activation. I agreed to the much larger and more complicated pulsatile
implant the pump experimentally and then in pumps. This allowed surgeons who had relatively
patients if the sintered titanium was removed little experience with continuous-flow pump
from the interior of the pump. technology to implant the pumps without difficulty.
8 O.H. Frazier

These pumps’ longer durability and reliability The only pressure that can be measured—with
proved another important factor in their wide- the Doppler apparatus—is the systolic pressure.
spread acceptance. The actual pressure difference between systole
The NIH spent $400–450 million in develop- and diastole remains unknown when the pulse is
ing the pulsatile pumps, and the companies not present (unless there is an arterial pressure
involved spent at least an equivalent amount. line). What contribution this abnormal physiology
This involved more than 10 years of intense study makes to continued pump thrombosis and the
of the physiologic parameters of the pulsatile ever-present, although reduced, incidence of
pumps, which were, after all, intended to mimic stroke has not been determined.
the native circulation. The continuous-flow pumps, The phenomenon of gastrointestinal bleeding
however, introduced an entirely new physiology. (GI) from arteriovenous malformations in the
This significant alteration in the normal circula- small and large bowel was first described by
tion contributes, in my opinion, to complications Heyde in 1958 in preterminal aortic stenosis [11].
that our medical community has still not fully We first reported GI bleeding in a minority of
addressed. patients supported with the Jarvik pump. We
Indeed, in the 1980s, there was much criticism thought that the decreased pulsatility induced by
from my medical colleagues as to the altered the continuous-flow pumps and the decreased
physiology these pumps induced. They voiced pulsatility noted in patients with severe aortic ste-
questions such as how will the pressure-sensitive nosis could be related. This problem with GI
baroreceptor response be affected, and what will bleeding remains. In our experience, it can gener-
its impact be on the normal blood pressure? This ally be addressed by decreasing the pump flow,
response would obviously be modified by a thereby increasing pulsatility. As the aortic valve
continuous-­flow pump. In addition, the juxtaglo- opening time is increased, minimal anticoagula-
merular apparatus of the kidneys should also be tion is required; thus, this complication is usually
affected, since they are believed to be pressure managed successfully [12].
sensitive, as well. What would be the effect on Numerous cases of complications have been
the right heart function? Could it be impaired by associated with pump thrombosis. Nonetheless,
the continuous unloading of the ventricle more than 250 patients have survived with a
throughout the cardiac cycle? These and many single continuous-flow pump for more than
other concerns were legitimately raised before 8 years, and 36 patients have been supported by
this technology was clinically introduced. the HeartMate II for more than 10 years. We
The effects of continuous-flow pumps, particu- know of no pump failures due to inherent
larly on the blood pressure, remain to be properly mechanical flaws. Rather, all of the complica-
investigated. Earlier experience with the pumps tions we see seem to be related to either the ana-
from 2003 to 2005, particularly with the tomic placement of the pump or other clinical
HeartMate II, saw hemorrhagic strokes in as many factors, such as hypotension due to sepsis or
as 20% of patients. We determined that although hemorrhagic shock. Improper pump placement
the systolic blood pressure was diminished, the can result in turbulence at the inflow or obstruc-
introduction of positive flow throughout diastole, tion at the outflow; either of these problems can
when pressure is normally passive, could contrib- contribute to stasis within the pump and
ute to an altered but hypertensive state that would increased platelet activation, both of which can
increase stroke risk. We addressed this complica- promote pump thrombosis. This problem high-
tion by aggressively lowering the blood pressure, lights the importance of proper implantation
which dramatically reduced the incidence of this technique. So, clearly, these pumps have over-
often fatal complication. come the durability problem that was a barrier
An additional problem results if the aortic to the clinical application of the pulsatile pumps.
valve is not opening. In this case, the pneumatic However, the abnormal physiology induced by
cuff will not yield an accurate blood pressure. continuous-flow pumps remains to be addressed.
1  History of Mechanical Circulatory Support 9

Fig. 1.5  The totally implantable version of the Jarvik 2000. Two power leads exit off the blood pump and are connected
to the internal power and control unit. Primary and secondary transcutaneous energy transmission system (TETS) coils
are placed in different locations in the abdominal wall. The external power and control are provided by the primary
TETS, and the secondary TETS is for backup operation. (Reproduced with permission from Myers TJ, Gregoric I,
Tamez D, et al. Development of the Jarvik 2000 intraventricular axial-flow left ventricular assist system. J Congest
Heart Fail Circ Support. 2000;1(3):133–140)

I am hopeful that medical academic leaders, already been shown with the Jarvik pump.
with NHLBI support, will be better able to This would insure a degree of pulsatility and
understand the physiologic problems associated perhaps lessen the problems of aortic insuffi-
with this technology. ciency and GI bleeding.
Many obvious problems could be addressed. If the aortic valve is closed, the pressure dif-
The most persistent problem is that of driveline ference between systole and diastole cannot be
infection. The percutaneous driveline was the directly measured without an arterial line. This
most expeditious and inexpensive approach in difference should be maximized to minimize
the feasibility studies. However, transcutane- diastolic pressure. In fact, the pump speed (in
ous power, which is as old as Tesla, has proved rpms) should be minimized because these pumps
effective in both the AbioCor and LionHeart are most effective as a true assist device and
pulsatile pumps and experimentally with the operate optimally at the lowest speed that can
Jarvik (Fig. 1.5) [9, 13]. Intermittent speed normalize circulation and maintain aortic valve
control can be done rather simply and has opening (Fig. 1.6).
10 O.H. Frazier

Fig. 1.6  Pulse pressure readings at various continuous-­ tion of the heart has less influence on the pulse pressure as
flow pump speeds. As pump speed is increased, the aortic the pump takes on more of the workload
valve ceases to open and close, and the rhythmic contrac-

offer the best answer to total heart replacement.

The Future Many patients still would benefit from total artifi-
cial heart technology. In the 1970s, we developed
I began my original involvement in this field as a a plutonium-powered internal battery that could
student under Drs. Michael E. DeBakey and power a 50-W pump for more than 82 years.
Domingo Liotta. The goal was to develop an artifi- Obviously, this was not pursued because we did
cial heart. In 1965, Dr. DeBakey told me that by not have a pump that would last more than 2 years.
1980, there would be “a hundred thousand These continuous-flow pumps, in contrast, have
Americans with a functional artificial heart.” not yet been pumped to mechanical failure, and
Likewise, NIH studies from the late 1960s pre- their long durability evidences their potential as
dicted that a clinically practical artificial heart meaningful long-term pumps.
would be in widespread use by the mid-1980s In 2005, Dr. William Cohn and I replaced the
(Fig. 1.7). But the problems associated with devel- total heart in an experimental animal with two
oping such a device proved to be far more formi- continuous-flow pumps [14]. We repeated these
dable than was commonly assumed, on the basis experiments numerous times and found that
of the perception at the time that an artificial heart animals with continuous-flow pumps could per-
could be a simple pump. The continuous-­ flow form well, grew normally, and had a normal
pumps now in widespread use as LVADs also may activity response on the treadmill; many of them
 NHLI–IMPLANTABLE CIRCULATORY SUPPORT SYSTEM –DEVELOPMENT STRATEGY

DECISION ON DECISION TO BUILD DECISION TO

CLINICAL SYSTEM APPLY CLINICALLY PROGRAM MISSION ACCOMPLISHED:
SYSTEM FEASIBILITY
• CLINICAL ACCEPTANCE – 10 YR. LIFE SYSTEM
• APPLICATION AND MAINTENANCE ESTABLISHED
COMPLETELY IMPLANTABLE TESTING INITIATED IN MEDICAL COMMUNITY
THERMAL ART. HEART COMPLETE SYSTEM • ESTABLISHED INDUSTRIAL SOURCES
AVAILABLE FOR CLINICAL WITH 5 YEAR LIFE
TESTING 2½ YEAR LIFE HIGH RELIABILITY
FIRST COMPLETELY
IMPLANTED THERMAL
ART. HEART SYSTEM CLINICAL
CONTROL
SUBSYSTEM
TEST PROGRAM OBJECTIVE ACCOMPLISHED –
TESTING INITIATED ON TESTING INITIATED WITH 10 YR. LIFE
1  History of Mechanical Circulatory Support

IMPLANTED COMPLETE SYSTEM WITH HIGH RELIABILITY SYSTEM

BLOOD PUMP ONE YEAR HIGH RELIABILITY TESTING INITIATED
(4 MO TESTING) COMPLETE SYSTEM
FIRST IMPLANTED 2½ YEAR LIFE
RADIOISOTOPE HIGH RELIABILITY WITH 5 YEAR LIFE
ACCEPTABLE ENDOGENOUS HIGH RELIABILITY
HEAT LOAD ESTIMATED ARTIFICIAL HEART
LESS THAN 60 WATTS (TH) IN ANIMAL

ANIMAL
FIRST COMPLETE THERMAL
BLOOD PUMP SYSTEM TESTING INITIATED ON
(ELECTRICALLY HEATED) COMPLETE SYSTEM WITH TESTING INITIATED
ONE YEAR RELIABILITY COMPLETE SYSTEM
FIRST
FIRST CONTROL WITH 5 YEAR LIFE
THERMAL SYSTEM FIRST IMPLANTED 2½ YEAR LIFE HIGH RELIABILITY
ENGINE RADIOISOTOPE HIGH RELIABILITY
OPERATIONAL ARTIFICIAL
HEART SYSTEM
BLOOD
PUMP

ENGINEERING

1967 '68 '69 '70 '71 '72 '73 '74 '75 '76 '77 '78 '79 '80 '81 '82 '83 '84 '85
CLINICAL SYSTEM DEVELOPMENT AND APPLICATION

Fig. 1.7  Diagram of the strategy and timeline for developing an implantable circulatory support system. (Modified from the National Heart and Lung Institute (later the
National Heart, Lung, and Blood Institute))
11
12 O.H. Frazier

Fig. 1.8  The AbioCor

totally implantable
pump (left) compared
with the BiVACOR
centrifugal pump (adult,
center, and child, right)

survived long term. We began working in 2012 lifesaving effort. However, it must be reiterated
with an investigator in Australia, Daniel Timms, that this represents a unique physiology never
who had devised a continuous-flow total artificial before encountered in mammalian species. We
heart (Fig. 1.8). This pump is small but can pro- have patients doing well who have not had a
duce up to 20 L of flow. It has only one moving pulse in more than 9 years and yet are totally
part, which is magnetically levitated. It perfuses asymptomatic. We must, however, study and
the pulmonary and systemic circulations simulta- address the complications seen with the use of
neously. We have demonstrated the feasibility of this technology, in both its short-term and long-
this pump in experimental animals and have even term application, to optimally benefit the heart
showed a Starling response, much like that of the failure patient.
normal heart, without changing the pump speed, In conclusion, I greatly appreciate the contri-
when calves implanted with this pump are on the butions of the THI faculty to the creation of this
treadmill. This technology offers great promise book—particularly Dr. Jeffrey Morgan—who, as
for the future and for the meaningful prevention a new arrival to our center, perhaps appreciates
of premature death from the loss of natural heart more than ourselves the more than 30 years of
function without the need for a heart transplant. I work on implantable continuous-flow pumps that
am confident that this technology will soon be originated here. I am glad to have had the oppor-
available for clinical use. tunity to document the history of these pumps, as
This book primarily addresses the current well as to highlight some of the early contributors
widespread use of the continuous-flow pump. It to the field.
is based on more than 50 years of experimental
and clinical work and a single-center experience
(one of the largest in the world) of more than References
1300 pump implantations and 1500 heart trans-
plants. In 2016, the number of continuous-flow 1. Cooley DA, Liotta D, Hallman GL, Bloodwell RD,
Leachman RD, Milam JD. Orthotopic cardiac pros-
pump implantations was twice that of heart thesis for two-staged cardiac replacement. Am
transplants, and I am personally gratified to J Cardiol. 1969;24:723–30.
know that more than 40,000 of these pumps 2. Institute of Medicine Committee to Evaluate the
have been implanted in patients worldwide as a Artificial Heart Program of the National Heart
1  History of Mechanical Circulatory Support 13

Lung and Blood Institute. The National Academies Hemopump, a catheter-mounted ventricular assist
Collection: reports funded by National Institutes device. Ann Thorac Surg. 1990;49:299–304.
of Health. In: Hogness JR, VanAntwerp M, edi- 9. Myers TJ, Gregoric I, Tamez D, Jarvik R, Frazier
tors. The artificial heart: prototypes, policies, and O, Inman RW, et al. Development of the Jarvik
patients. Washington, DC: National Academies 2000 intraventricular axial-flow left ventricular
Press; 1991. assist system. J Congest Heart Fail Circ Support.
3. Norman JC, Duncan JM, Frazier OH, Hallman GL, 2000;1:133–40.
Ott DA, Reul GJ, et al. Intracorporeal (abdominal) 10. Frazier OH, Myers TJ, Gregoric I. Biventricular

left ventricular assist devices or partial artificial assistance with the Jarvik FlowMaker: a case report.
hearts: a five-year clinical experience. Arch Surg. J Thorac Cardiovasc Surg. 2004;128:625–6.
1981;116:1441–5. 11. Heyde EC. Gastrointestinal bleeding in aortic stenosis
4. Hall JE. Guyton and hall textbook of medical physiol- (letter). N Engl J Med. 1958;259:196.
ogy. 13th ed. Philadelphia, PA: Elsevier; 2016. 12. Letsou GV, Shah N, Gregoric ID, Myers TJ, Delgado
5. Kung RT, Yu LS, Ochs B, Parnis S, Frazier OH. An R, Frazier OH. Gastrointestinal bleeding from arte-
atrial hydraulic shunt in a total artificial heart. A bal- riovenous malformations in patients supported by the
ance mechanism for the bronchial shunt. ASAIO Jarvik 2000 axial-flow left ventricular assist device.
J. 1993;39:M213–7. J Heart Lung Transplant. 2005;24:105–9.
6. Frazier OH, Bricker JT, Macris MP, Cooley DA. Use 13. Parnis SM, Conger JL, Fuqua JM Jr, Jarvik RK,
of a left ventricular assist device as a bridge to Inman RW, Tamez D, et al. Progress in the devel-
transplantation in a pediatric patient. Tex Heart Inst opment of a transcutaneously powered axial flow
J. 1989;16:46–50. blood pump ventricular assist system. ASAIO
7. Frazier OH, Nakatani T, Duncan JM, Parnis SM, J. 1997;43:M576–80.
Fuqua JM. Clinical experience with the Hemopump. 14.
Frazier OH, Tuzun E, Cohn W, Tamez D,
ASAIO Trans. 1989;35:604–5. Kadipasaoglu KA. Total heart replacement with
8. Frazier OH, Wampler RK, Duncan JM, Dear WE, dual centrifugal ventricular assist devices. ASAIO
Macris MP, Parnis SM, et al. First human use of the J. 2005;51:224–9.
 Who Is an Appropriate Candidate
for Long-Term MCS?: The Art 2
of Patient Selection

Carol S.C. Lai and Andrew B. Civitello

HF hospitalization in the past 6 months, and the

Indications for MCS presence of all the previous features despite
attempts to optimize therapy (Table 2.1) [4].
Heart failure (HF) is a chronic and complex dis- Patients with advanced heart failure refractory
ease that has reached epidemic proportions to medical management may be eligible for
worldwide. An estimated 6.5 million Americans advanced therapy, including heart transplantation
have HF, and it is a leading cause of morbidity and mechanical circulatory support (MCS). Heart
and mortality, with 50% mortality within 5 years transplantation is considered the definitive ther-
of diagnosis [1]. Approximately less than 10% of apy for advanced HF. However, shortage of donor
this population will progress to advanced organs and prolonged wait times remain a signifi-
HF. These patients experience poor quality of cant limitation. The development of MCS, such
life, frequent hospitalizations, and a 1-year mor- as the left ventricular assist devices (LVAD), has
tality of 25–50% [2, 3]. Advanced HF is charac- emerged as an effective and viable form of ther-
terized by severe symptoms of heart failure with apy. Though this field is quickly evolving, ther-
dyspnea and/or fatigue at rest or with minimal apy with an LVAD is not free of complications,
exertion, episodes of fluid retention, objective making appropriate patient selection imperative
evidence of severe cardiac dysfunction, severe for successful therapy.
impairment of functional capacity, history of ≥1 Generally, LVAD implantation is considered
reasonable in “highly selected patients with
advanced end-stage HF and an estimated 1-year
mortality >50% with medical therapy” [5].
C.S.C. Lai, M.D. Four major indications for LVAD implanta-
Texas Heart Institute at Baylor St. Luke’s Medical
Center, Baylor College of Medicine tion exist: (1) bridge to transplantation (BTT), (2)
Houston, TX, USA destination therapy (DT), (3) bridge to recovery,
e-mail: [Link]@[Link] and (4) bridge to decision. Bridge to transplanta-
A.B. Civitello, M.D. (*) tion is considered in patients with advanced HF
Heart Transplant Program, Texas Heart Institute at who are candidates for heart transplantation but
Baylor St. Luke’s Medical Center, Baylor College of are hemodynamically unstable despite maximum
Medicine, Houston, TX, USA
medical therapy, including inotropes and intra-­
Advanced Heart Failure Center, Texas Heart Institute aortic balloon pumps. Due to hemodynamic
at Baylor St. Luke’s Medical Center, Baylor College
of Medicine, Houston, TX, USA instability, prolonged wait time, and increased
e-mail: andrewc@[Link] risk mortality, they are too ill to wait for a donor

© Springer International Publishing AG 2018 15

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
16 C.S.C. Lai and A.B. Civitello

Table 2.1  Definition of advanced heart failure [4] transplantation. Like BTT, LVAD implantation as
1. Severe symptoms of HF with dyspnea and/or DT has also been shown to improve survival. The
fatigue at rest or with minimal exertion (NYHA landmark Randomized Evaluation of Mechanical
functional class III or IV)
Assistance for the Treatment of Congestive Heart
2. Episodes of fluid retention (pulmonary and/or
systemic congestion, peripheral edema) and/or
Failure (REMATCH) trial randomized 129
reduced cardiac output at rest (peripheral patients with end-stage HF who were not eligible
hypoperfusion) for cardiac transplantation to LVAD therapy
3. Objective evidence of severe cardiac dysfunction, (HeartMate VE device) versus optimal medical
shown at least by one of the following: management. LVAD therapy demonstrated a sub-
 (a) Low LVEF (<30%) stantial survival benefit with 1-year survival of
 (b) Severe abnormality of cardiac function on
52% versus 25% and 2-year survival of 23% ver-
Doppler echocardiography with a pseudonormal
or restrictive mitral inflow pattern sus 8%. Patients in the LVAD group also had
 (c) High LV filling pressures (mean PCWP improved quality of life, though they were two
>16 mmHg and/or mean RAP >12 mmHg by times more likely to develop serious adverse
pulmonary artery catheterization) events including infection, bleeding, and mechan-
 (d) High BNP or NT-proBNP plasma levels, in the ical failure of the device [10]. The development
absence of non-cardiac causes
of newer left ventricular assist devices has further
4. Severe impairment of functional capacity shown by
one of the following: improved survival and quality of life. The
 (a) Inability to exercise HeartMate II trial randomized 200 patients with
 (b) 6-MWT distance <300 m or less in females and/ refractory HF who were ineligible for transplan-
or patients aged ≥75 years tation to the newer continuous-flow device versus
 (c) Peak VO2 < 12–14 mL/kg/min pulsatile-flow device. Patients with the
5. History of ≥1 HF hospitalization in the past continuous-­ flow device demonstrated superior
6 months survival rates with 1-year survival of 68% versus
6. P
 resence of all the previous features despite “attempts 55% and 2-year survival of 58% versus 24%.
to optimize” therapy including diuretics, inhibitors of
the renin-angiotensin-aldosterone system, and While both groups had improvements in NYHA
beta-blockers, unless these are poorly tolerated or class and quality of life scores, the continuous-­
contraindicated, and CRT, when indicated flow group had less major adverse effects [11].
Survival for continuous-flow pumps has contin-
heart to be identified and require mechanical sup- ued to improve to an overall 1-year survival of
port in the interim. MCS in this population 90% with 2-year survival of 70% [9, 12, 13].
improves quality of life and reduces mortality Improved survival is not only secondary to
[6]. In a multicentered trial, Frazier et al. demon- advancements in LVAD technology but also to an
strated LVAD implantation improves NYHA improved understanding of heart failure, patient
functional class, organ dysfunction, and survival selection, surgical technique, and postoperative
to transplantation compared to medical manage- care [14]. Bridge to recovery is used in patients
ment in a cohort of transplant candidates. with potentially reversible etiologies of heart fail-
Furthermore, this survival benefit is sustained ure. LVAD implantation allows for recovery of
1-year posttransplantation [7, 8]. Bridge to trans- myocardial function and is removed once the
plantation remains the most common indication myocardium has recovered. Finally, bridge to
for implantation [9]. However, due to the grow- decision is reserved for patients with hemody-
ing population of patients with advanced HF, namic instability requiring urgent mechanical
scarcity of donor organs, and improved durability support in which candidacy for transplantation or
of the newer designed devices, there is an increas- destination therapy cannot be made at the time of
ing trend toward LVAD implantation in patients implantation. MCS at this time may allow for
for destination therapy. Destination therapy is change in transplant eligibility with reversal of
offered as a permanent device in patients with pulmonary hypertension, improvement in renal
advanced HF who are not candidates for heart function and/or hepatic function, and weight loss.
2  Who Is an Appropriate Candidate for Long-Term MCS?: The Art of Patient Selection 17

 eneral Criteria for MCS Patient

G The challenge lies with selecting patients with suf-
Selection ficient severity of illness to achieve a benefit while
avoiding patients who are too ill or too early in the
Patients must undergo a comprehensive evalua- disease course to derive any benefit. Additionally,
tion prior to LVAD implantations. The severity of the risk of device-related complications including
heart failure must be assessed by the patient’s thromboembolic events, gastrointestinal bleeding,
clinical presentation, cardiopulmonary stress and infection must also be considered.
testing, and hemodynamic studies. Cardiac and The National Institutes of Health-sponsored
anatomic factors are also evaluated such as right Interagency Registry for Mechanically Assisted
heart function, presence of arrhythmias, and ana- Circulatory Support (INTERMACS) is a US reg-
tomic and body size. Non-cardiac factors includ- istry which has acquired data on patients with
ing coexisting life-limiting illness, psychosocial, FDA-approved MCS devices. It has developed a
and age-related considerations are also assessed. validated classification scheme based on patients’
Patients must also be evaluated for cardiac trans- hemodynamic status and has been established to
plant candidacy and LVAD operative risk. predict outcomes in patients undergoing
General indications for LVAD implantation are MCS. The classification scheme is felt to be more
based on inclusion and exclusion data from clini- specific than NYHA classification and allows for
cal trials. LVADs are indicated in patients with optimal patient selection and timing. Moreover, it
severe HF with NYHA functional IV symptoms is easily assessed at the bedside (Table 2.2) [17].
and left ventricular ejection fraction <25%, who Currently LVADs are approved by the FDA for
have failed response to optimal medical manage- INTERMACS profile 1–5, with the greatest amount
ment. This includes beta-blockers and ACE inhibi- of implants in patients with profile 1–2 [9].
tors if tolerated, inotropes, and intra-­aortic balloon
pumps. In those who are not inotrope- or balloon
pump-dependent and physically able, functional Risk Scores
limitation is demonstrated with peak oxygen con-
sumption ≤14 mL/kg/min [15]. Absolute contra- No single variable exists to select candidates for
indications include sepsis or current active LVAD therapy. While the INTERMACS classifi-
infections, severe right heart failure, untreated and cation scheme has been shown to predict progno-
severe carotid artery disease, severe obstructive sis after MCS, it lacks specificity and objectivity.
and/or restrictive pulmonary disease, irreversible Furthermore, it does not incorporate severity of
severe cerebral injury, dialysis-­ dependent renal multi-organ dysfunction [18]. Several risk scores
failure, elevated INR from liver failure or dissemi- have been developed, which incorporate individ-
nated intravascular coagulation, any severe end- ual data. They have been shown to predict both
organ failure, heart failure expected to recover mortality in advanced heart failure and survival
without mechanical circulatory support, and non- after LVAD therapy.
cardiac illness that would limit survival to less The Acute Physiology and Chronic Health
than 2 years. Relative contraindications include Evaluation II (APACHE II) score was developed
morbid obesity, small body size (BSA <1.5 m2), from a multi-institutional cohort of 5815 criti-
chronic renal dysfunction but not dialysis-depen- cally ill patients. The score consists of 13 vari-
dent, malnutrition, and severe or untreated mitral ables including temperature, mean arterial
stenosis and aortic regurgitation. Of note, small pressure, heart rate, respiratory rate, partial pres-
body size may no longer be a contraindication as sure of arterial O2 or A-a gradient, arterial pH,
the newer devices are able to accommodate smaller sodium, potassium, creatinine, hematocrit, white
body sizes [16]. blood cell count, Glasgow coma score, and age.
Patient selection for LVAD remains a challenge Patients with HF and an APACHE II score greater
as successful therapy is dependent on time of than 20 have been shown to have a significant
implantation and appropriate patient selection. increase in hospital deaths [19]. In a retrospective
18 C.S.C. Lai and A.B. Civitello

Table 2.2  INTERMACS classification [17]

Time frame for
Level Hemodynamic status intervention
1. “Crash and burn” Persistent hypotension despite rapidly escalating support Within hours
and eventually IABP and critical organ hypoperfusion
2. “Sliding on inotropes” Intravenous inotropic support with acceptable values of Within days
blood pressure and continuing deterioration in nutrition,
renal function, or fluid retention
3. “Dependent stability” Stability reached with mild to moderate doses of inotropes Elective over weeks to
but demonstrated failure to wean from them due to months
hypotension, worsening symptoms, or progressive renal
dysfunction
4. “Frequent flyer” Possible weaning of inotropes but experiencing recurrent Elective over weeks to
relapses, usually fluid retention months
5. “Housebound” Severe limited tolerance for activity: comfortable at rest with Variable urgency,
some volume overload and often with some renal dependent on nutrition
dysfunction and organ function
6. “Walking wounded” Less severe limited tolerance for activity and lack of volume Variable urgency,
overload. Fatigue easily dependent on nutrition
and organ function
7. “Placeholder” Patient without current or recent unstable fluid balance. Not currently
NYHA class II or III indicated

cohort of patients with advanced HF, each unit patients referred for transplant by Kalogeropoulos
increase in the APACHE II score independently et al. Overall, the SHFM accurately discriminates
predicted death. Furthermore, patients with a between low- and high-risk patients. However, in
medium APACHE II score of 11–20 had the terms of absolute risk (observed versus predicted
greatest benefit from LVAD placement [20]. event rate), the model overestimates survival and
The Seattle Heart Failure Model (SHFM) was underestimates risk, particularly in black patients
derived from a cohort of 1125 NYHA IIIb–IV and those with implantable devices. The adverse
patients. Twenty variables are weighted by haz- events in this case were primarily driven by
ard ratio: age, gender, NYHA class, weight, ejec- death. The model showed improved calibration
tion fraction, systolic blood pressure, presence of when evaluating for mortality alone [22]. One
ischemic cardiomyopathy, daily furosemide limitation of the SHFM is that the model is not
equivalent dose, inotrope use, statin use, allopuri- developed from a cohort of patients being consid-
nol use, angiotensin-converting enzyme/angio- ered for MCS. However, when applied to the
tensin receptor blocker use, beta-blocker use, REMATCH database, in which patients with
potassium-sparing diuretic use, implantable car- advanced HF were ineligible for cardiac trans-
dioverter defibrillator, hemoglobin, lymphocyte plant, the 1-year SHFM-predicted survival was
percentage, serum uric acid, serum cholesterol, similar to actual survival for both the medical and
and serum sodium. The SHFM has been updated LVAD therapy groups [23]. Furthermore, the
specifically for LVAD patients with the addition SHFM is able to predict important features of the
of intra-aortic balloon pump and inotrope ther- patient’s hospital course after LVAD implanta-
apy. Web-based calculators are available to con- tion. Patients in the lower-risk group had shorter
vert the SHFM score to a mortality estimation length of stay, higher rate of discharge, and dis-
where a score > 3.53 is considered high risk with charge within 60 days of LVAD placement [24].
a 50% predicted survival at 6 months. SHFM has The REMATCH trial demonstrated that
been shown to predict 1-, 2-, and 3-year survival patients who were ineligible for cardiac trans-
in low- to high-risk HF patients [21]. The perfor- plant had improved 1-year survival with LVAD
mance of the SHFM was applied to advanced HF therapy compared to patients receiving optimal
2  Who Is an Appropriate Candidate for Long-Term MCS?: The Art of Patient Selection 19

medical therapy. The Destination Therapy Risk HMRS provided significantly higher risk dis-
Score (DTRS) was derived from the DT registry crimination in both DT and BTT populations and
in the post-REMATCH era. In an analysis of 222 across all risk groups [27]. However, this could
patients who received pulsatile LVAD, nine pre- not be confirmed in a single-center study by
operative risk factors were identified to predict Thomas et al. In this retrospective analysis of 205
90-day in-hospital mortality by multivariable patients who received HMII for either BTT or
analysis. This includes platelet count ≤148, DT, the HMRS demonstrated poor discrimina-
serum albumin ≤3.3, INR >1.1, vasodilator ther- tion in 90-day and 1-year survival across all risk
apy, mean pulmonary artery pressure ≤25, AST groups [28].
>45, hematocrit ≤34%, BUN >51, and no intra- The above risk scores were applied to a cohort
venous inotrope use. Each variable is assigned a of 86 patients with continuous-flow LVADs to
weighted risk score, creating a cumulative score determine their ability to predict mortality after
ranging from 0 to 27. Patients can be stratified to LVAD implantation. INTERMACS appeared to
four risk categories based on probability of differentiate high-risk populations if patients with
90-day mortality. The score offers good discrimi- profile 1 and 2 were combined. The APACHE II
nation between low- and high-risk groups with a and SHFM scores successfully differentiated
1-year survival of 69% versus 13%. Limitations between high-risk and low-risk groups. The DTRS
of the DTRS include generalizability as the regis- failed to show significant survival between low-
try is composed from an older population and risk and high-risk groups. Overall, the SHFM
predominantly Caucasian males. Mechanical score was the best predictor of mortality [29].
ventilation, intra-aortic balloon pump, and patient The development of risk scores based on pre-
size were also not represented in the model due to operative risk is a useful guide to patient selec-
small sample size. Lastly DTRS was not applied tion. However, no single risk score has been
to newer generations of continuous-flow LVADs shown to be conclusively predictive and must be
[25]. This was examined by Teuteberg et al. who used in context with the patient’s clinical status.
applied the DTRS to 1124 patients with
continuous-­flow LVADs for BTT and DT. The
DTRS in this lower-risk cohort demonstrated Pulmonary Hypertension
modest discriminatory capacity for 90-day in-­
house mortality. The score was able to discrimi- Pulmonary hypertension is common in patients
nate between those at low versus high risk, with advanced HF and is most commonly second-
however failed to discriminate between low- and ary to left heart disease. Pulmonary hypertension
intermediate-risk groups. The ability of DTRS to in this population is defined by mean pulmonary
predict 2-year survival was also examined. DTRS artery pressure ≥25 at rest and pulmonary arterial
moderately predicted survival for DT populations wedge pressure (PAWP) >15 mmHg assessed by
stratified by risk group, though it was unable to right heart catheterization [30]. Pulmonary artery
predict survival in the BTT population [26]. hypertension with elevated pulmonary vascular
The HeartMate II risk score (HMRS) was resistance (PVR) is considered a contraindication
developed from patients enrolled either in the to heart transplantation when PVR >5 Wood units
HeartMate II (HMII) BTT or DT clinical trials. or the indexed pulmonary vascular resistance
The score used preoperative patient-specific fac- (PVRI) is >6 or the transpulmonary gradient
tors for predicting 90-day mortality in LVAD exceeds 16–20 mmHg, due to the high risk of right
candidates. In a multivariate analysis, age (per ventricular failure posttransplantation [31]. This
10 years), albumin, creatinine, INR, and center occurs because the grafted right heart is unable to
volume demonstrated good discrimination and tolerate an abrupt increase in pulmonary vascular
calibration in predicting 90-day mortality. resistance in the immediate postoperative period.
Patients were able to be stratified to three risk cat- The Cardiac Transplant Research Database has
egories. When comparing HMRS to DTRS, shown that preoperative vascular resistance is an
20 C.S.C. Lai and A.B. Civitello

independent risk factor for early and late mortality lowing LVAD implantation. Increased cardiac
after cardiac transplantation [32]. output from left ventricular unloading leads to
While pulmonary hypertension is a contrain- increased venous return to the right ventricle,
dication for cardiac transplantation, it is not con- thereby unmasking preexisting right ventricular
sidered a contraindication for LVAD therapy. In dysfunction. Left ventricular decompression may
fact, LVAD implantation may improve the sever- also result in leftward shift of the interventricular
ity of pulmonary hypertension, thereby reversing septum, decreasing septal contribution to right
a patient’s contraindication to transplant candi- ventricle contraction [39].
dacy. Numerous studies have demonstrated that The incidence of RVF ranges from 9.4 to 44%,
LVAD implantation is safe and efficacious when depending on the definition of RVF [40]. The
used as bridge to transplant [33–35]. Tsuashita INTERMACS defines right heart failure as symp-
et al. conducted one of the largest studies evaluat- toms or findings of persistent right ventricular
ing the effect of continuous-flow LVADs on pul- failure characterized by both of the following: (1)
monary hypertension as well as posttransplantation documentation of elevated central venous pres-
outcomes in patients bridged with LVADs. The sure (CVP) and (2) manifestation of elevated
study found that PVR decreased significantly central venous pressure. Elevated CVP may be
post-LVAD implantation. This was also observed measured either directly (e.g., right heart cathe-
in patients considered to have severe refractory terization) with CVP or right atrial pressure
pulmonary hypertension. As a result, 66% of (RAP) >16 mmHg, findings of the significantly
patients in this cohort, who would have otherwise dilated inferior vena cava with the absence of
not been candidates for transplantation, were able inspiratory variation by echocardiography, or
to have their eligibility reversed and undergo car- clinical findings of elevated jugular venous dis-
diac transplantation. Despite improvements in tension. Elevated CVP is manifested by clinical
PVR post-implantation, patients with high PVR findings of peripheral edema, the presence of
pre-LVAD implantation still had increased in-­ ascites or palpable hepatomegaly on physical
hospital mortality posttransplantation. A poten- examination or diagnostic imaging, or laboratory
tial explanation for this finding is that the evidence of worsening hepatic or renal function.
pulmonary vasculature undergoes heterogenous The severity of RVF may be further categorized
or incomplete remodeling and therefore is sus- as mild, moderate, severe, and severe-acute RVF
ceptible to early postoperative insult including (Table 2.3) [41].
myocardial ischemia, metabolic acidosis, hypox- Predicting patients at risk of developing RVF
emia, inflammatory response, or blood transfu- post-LVAD implantation would improve patient
sion. Long-term survival posttransplantation, selection. It would also allow clinicians to imple-
however, was similar in patients with high PVR ment strategies to avoid RVF. Several studies
pre-LVAD compared to those with low PVR [36]. have identified predictors of post-implantation
RVF (Table 2.4) [42–47]. However, predicting
RV response to LVAD implantation remains chal-
Right Ventricular Failure lenging as there is no consensus among these
studies, and the data is limited by different defini-
Right ventricular failure (RVF) after LVAD tions used by each trial to define RVF.
implantation is a cause of significant morbidity Matthews et al. developed a RVF risk score
and mortality. It results in longer hospitalization, (RVFRS) based on independent predictors of
higher transfusion requirements, need for reop- RVF. Each variable was weighted by odds ratio:
eration, and end-organ damage [37]. Concurrent vasopressor requirement (4 points), AST ≥80 (2
placement of a right ventricular assist device points), bilirubin ≥2 (2 points), and creati-
(RVAD) for RVF has been identified as the most nine ≥3.0 (3 points). The RVFRS is calculated as
significant risk factor for death after implantation the sum of the pointed awarded for the presence
[38]. Several mechanisms contribute to RVF fol- of each of the preoperative variables. Patients
2  Who Is an Appropriate Candidate for Long-Term MCS?: The Art of Patient Selection 21

Table 2.3  Right heart failure severity score [41] retrospectively to continuous-flow LVAD recipi-
Mild – Post-implant inotropes, inhaled nitric ents, a higher PAPi (>2.0) was associated with
oxide, or intravenous vasodilators not lower rates of RVAD implantation. Interestingly,
continued beyond post-op day 7 after
PAPi more strongly predicted early RVAD
VAD implant
– No inotropes continued beyond
requirement when hemodynamics were mea-
post-op day 7 after VAD implant sured on inotropes versus off inotropes.
Moderate – Post-implant inotropes, inhaled nitric Furthermore, the predictive ability of PAPi
oxide, or intravenous vasodilators remained valid regardless of the timing between
continued beyond post-op day 7 and right heart catheterization and LVAD implanta-
up to post-op day 14 following VAD
implant
tion, where the maximum time was 6 months
Severe – Central venous pressure or right atrial [49]. In another cohort of patients with
pressure >16 mmHg continuous-­flow LVADs, PAPi initially decreased
– Prolonged post-implant inotropes, in the immediate postoperative setting. However,
inhaled nitric oxide, or intravenous in patients without RVF, PAPi significantly
vasodilators continued beyond
increased after 24 h. Therefore, PAPi may also be
post-op day 14 following VAD
implant used as an indicator for right ventricular recovery
Severe-­ – Central venous pressure or right atrial following LVAD implantation [50].
acute pressure >16 mmHg Attempts have also been made to identify pre-
– Need for right ventricular assist dictors of RVF by echocardiography. This is
device at any time following VAD challenging due to the retrosternal position of the
implant
right ventricle and its complex geometry, as well
– Death during VAD implant
hospitalization with RHF as the as the lack of standardization of echocardio-
primary cause of death graphic protocols [40]. Tricuspid annular plane
systolic excursion (TAPSE) <7.5 mm [51],
reduced right ventricular fractional change area
with a score ≥5.5 have a 15-fold greater risk of (<35%) [52], short to long axis ratio ≥ 0.6, and
developing RVF compared to patients with a more severe tricuspid regurgitation [53] have
score ≤3.0. When compared to commonly used been found to be predictive of RVF in small,
predictors of RVF-RVWSI, transpulmonary gra- single-­center studies. However, these parameters
dient, PVR, RAP, and PASP, RVFRS was shown could not be validated in larger subsequent stud-
to be superior [46]. ies, potentially because measurements of these
Invasive hemodynamic monitoring also has a load-dependent values are made in patients under
role in predicting RVF. Elevated CVP, low pul- severe hemodynamic stress.
monary artery systolic pressure (PASP), low RV Strain, strain rate, and speckle tracking are
stroke work index, low cardiac index, and ele- promising predictors of RVF. Unlike standard
vated PVR have all been shown to be predictors echocardiographic predictors, strain imaging is
of RVF. Interestingly, unlike cardiac transplanta- less sensitive to loading conditions and can rea-
tion, low PASP rather than high PASP is a sonably assess right ventricular systolic function.
­predictor of RVF. Caution must be used however Strain imaging was applied to patients undergoing
as hemodynamic parameters may fluctuate in LVAD implantation. Global longitudinal right
unstable patients who may be on inotropes. ventricular strain was found to be an independent
A novel hemodynamic marker, the pulmonary predictor of RVF where a right ventricular free
artery pulsatility index (PAPi), has been found to wall strain <−9.6% predicted RVF with 68% sen-
identify patients at high risk of developing severe sitivity and 76% specificity. When combined with
RVF after inferior wall myocardial infarction the RVF score described above, it provided incre-
[48]. PAPi is defined as [(systolic pulmonary mental value on its predictive ability [54]. The
artery pressure − diastolic pulmonary artery right ventricular function after left ventricular
pressure)/central venous pressure]. When applied assist device (RFV-LVAD) study prospectively
22 C.S.C. Lai and A.B. Civitello

Table 2.4  Preoperative risk factors for RVF

Study (first RVF
author) n VAD type Definition of RVF incidence Preoperative risk factors
Ochai [42] 245 Pulsatile –  RVAD implantation 9% –  Female gender (OR 4.5)
– Preoperative circulatory
support (OR 5.3)
– Nonischemic etiology
(OR 3.3)
Dang [43] 108 Pulsatile –  RVAD implantation 38.9% –  Female gender
– Inotropes/pulmonary – Elevated intraoperative CVP
vasodilator therapy (OR 1.2)
≥14 days
Drakos [44] 175 Pulsatile –  RVAD implantation 44% –  Destination therapy (OR 3.3)
(86%), – Inotrope therapy >14 –  IABP (or 3.9)
Continuous consecutive days
(14%) –  Inhaled NO ≥48 h – PVR ≥4.3 WR (or 4.1)
–  PVR 2.8–4.2 Wu (or 3.0)
–  Higher RAP
– Increased LV end-diastolic
diameter
–  Lower platelets
–  Higher bilirubin
Fitzpatrick 266 Pulsatile –  RVAD implantation 37% – Cardiac index <2.2 L/min/m2
[45] (98%), (OR 5.7)
Continuous – RVSWI <0.25 mmHg/L/m2
(2%) (OR 5.1)
– Severe pre-VAD RV
dysfunction (OR 5.0)
– Cr ≥1.9 (OR 4.8)
– Previous cardiac surgery
(OR 4.5)
– SBP ≤96 mmHg (OR 2.9)
Matthews 194 Pulsatile –  RVAD implantation 35% – Vasopressor requirement
[46] (86%), (OR 3.9)
Continuous – Inotrope therapy – AST ≥80 (or 2.1)
(14%) >14 days
– Inhaled NO ≥48 h – Bilirubin ≤2 (OR 2.4)
– Hospital discharge with – Cr ≥2.3 (OR 2.9)
inotrope therapy
Kormos [47] 484 Continuous –  RVAD implantation 20% –  CVP/PCWP >0.63 (or 2.3)
– Inotrope therapy – Preoperative ventricular
>14 days after support (OR 5.5)
implantation
– Inotrope support starting –  Bun >39 (or 2.1)
14 days after
implantation
AST aspartate aminotransferase, BUN blood urea nitrogen, CVP central venous pressure, Cr creatinine, IABP intra-­
aortic balloon pump, LV left ventricular, NO nitric oxide, PCWP pulmonary capillary wedge pressure, PVR pulmonary
vascular resistance, RAP right atrial pressure, RV right ventricular, RVSWI RV stroke work index, SBP systolic blood
pressure
2  Who Is an Appropriate Candidate for Long-Term MCS?: The Art of Patient Selection 23

evaluated LVAD candidates with standard echo- right ventricle fails despite these measures and is
cardiographic parameters and strain imaging. unable to maintain adequate LVAD flow, implanta-
Standard echocardiographic parameters included tion of a right VAD is necessary.
TAPSE, pulse tissue Doppler peak systolic veloc-
ity, right ventricular myocardial performance
index, and right ventricular fractional area change. Renal and Hepatic Dysfunction
Preliminary findings revealed right ventricular
global longitudinal strain was the most important Renal dysfunction is common in patients hospi-
predictor of RVF [55]. talized for heart failure. Acute kidney injury
With available predictive risk factors for RVF, (AKI), defined as an increase in creatinine
the goal for the clinician should be to optimize the >0.3 mg/dL, is observed in 50% of patients
patient prior to LVAD implantation as well as cer- admitted for acute decompensated heart failure
tain measures during implantation and postopera- and 70% of patients with cardiogenic shock [60].
tively [56, 57]. In the preoperative stage, patients AKI is an independent risk factor for all-cause
with evidence of right ventricular dysfunction mortality, cardiac-specific mortality, pump fail-
should undergo aggressive treatment to reduce right ure death, and increased hospitalization in
ventricular wall stress with a goal of decreasing patients with heart failure [61, 62]. The etiology
right atrial pressure to <12 mmHg. The benefit of of renal dysfunction in patients with heart failure
reducing pulmonary artery pressure and pulmonary is multifactorial due to intrinsic disease from
vascular resistance remains uncertain as they have chronic comorbidities and cardiorenal syndrome.
been shown to be inconsistent predictors of In cardiorenal syndrome, poor cardiac output
RVF. Though they may be decreased with PDE5 results in poor renal perfusion. Venous conges-
inhibitors, this has yet to demonstrate a clear clini- tion is also thought to play a role via activation of
cal benefit. Those who remain at high risk despite the renin-angiotensin system and renal arterial
medical optimization should be considered for vasoconstriction to maintain glomerular filtration
planned biventricular support or total artificial heart rate (GFR). The autoregulatory capability cannot
as elective right VAD implantation has been shown be sustained and GFR ultimately declines.
to have better long-term survival versus emergency Hemodynamic changes also result in decreased
implantation [58]. Intraoperatively, correction of tri- GFR secondary to inflammation, endothelial
cuspid regurgitation has been theorized to improve ­dysfunction, and anemia [63].
right ventricular function, though this remains con- Not surprisingly, preoperative renal dysfunc-
troversial. In a review of 2000 patients who under- tion is associated with complications and mortality
went LVAD implantation, concomitant correction post-LVAD implantation. While renal dysfunction
of moderate-to-severe tricuspid regurgitation did is a relative contraindication for LVAD and dialy-
not reduce early death or right VAD requirement. In sis dependence remains an absolute contraindica-
fact, this was associated with worse early postoper- tion, renal function has been shown to improve
ative outcomes including renal failure, dialysis, after LVAD implantation. This is due to improved
reoperation, transfusion requirement, and greater renal perfusion and correction of the neurohor-
length of stay [59]. Surgical hemostasis during monal dysregulation. Studies have shown patients
LVAD implantation is critical to minimize blood who had recovery in GFR had a slightly increased
product transfusion, thereby preventing volume survival compared to those who had no improve-
overload of the right ventricle [57]. Inotropic agents ments in GFR. In fact, patients who had a recovery
in the immediate postoperative period are essential of GFR to that of >60 mL/min/1.73 m2 had com-
due to increased preload to the right ventricle. parable survival to patients with normal renal
Pulmonary vasodilators may also be employed to function post-­implantation. Positive predictors of
decrease right ventricular afterload. Optimization of improved renal function include the absence of
pump speed also avoids excessive leftward shift of diabetes, lower cardiac index preimplantation,
the septum, thereby decreasing venous return. If the lower body mass index, and use of intra-aortic
24 C.S.C. Lai and A.B. Civitello

b­ alloon pump. Negative predictors include older Like renal dysfunction, hepatic dysfunction has
age and the use of angiotensin-converting enzyme been shown to improve after LVAD implantation.
inhibitors/angiotensin receptor blockers. These Russell et al. showed a decrease in AST, ALT, and
variables however have not been externally vali- total bilirubin in patients after continuous-­flow
dated [64–66]. LVAD implantation. A more significant decrease
While the majority of patients will have was seen in patients with above-normal liver func-
improvements in renal function after LVD implan- tion prior to implantation [71]. This reduction in
tation, a subset of patients may develop AKI, liver function was seen by 1-month and continued
which confers a threefold increased risk of 1-year up to 1-year post-­implantation [72].
mortality [67]. Factors such as acute blood loss,
volume shifts, arrhythmias, and multiple vasoac-
tive medications may negatively affect renal Patient Size Considerations
hemodynamics. Management in this scenario
should focus on optimizing the patient’s volume Data shows that patients at both extremes of body
status, maintaining goal mean arterial pressure, mass index (BMI) are associated with worse out-
balancing inotropic and vasopressor medications, comes after LVAD implantation. Cardiac cachexia,
and optimizing right ventricular systolic function defined as BMI <20 kg/m2 or <80% ideal body
[68]. Should renal function continue to decline, the weight, is a common complication in patients with
indications for renal replacement therapy are the heart failure and portends a poor prognosis. This
same as other patients without LVADs. However, state is marked by poor nutritional status and
the type of renal replacement therapy, hemodialy- hypoalbuminemia, which is independently associ-
sis versus peritoneal dialysis, vascular access, and ated with an increased risk of death from heart fail-
hemodynamic monitoring during dialysis, remains ure [73]. Cachexia prior to cardiac transplantation
challenging [69]. has also been associated with increased risk of
The etiology of hepatic dysfunction in patients morbidity and mortality posttransplantation [74].
with heart failure ranges from cardiac cirrhosis or Malnutrition increases the risk of postoperative
congestive hepatopathy to ischemic hepatitis. complications such as infection and poor func-
Cardiac cirrhosis results from chronic and pro- tional capacity. Interestingly, studies have shown
gressive right ventricular dysfunction. As dis- that underweight patients who underwent LVAD
cussed elsewhere, hepatic function is used in implantation have no difference in survival com-
various risk scores to predict right ventricular fail- pared to patients with normal BMI. However,
ure. Ischemic hepatitis on the other hand is caused underweight patients are at significantly higher
by cardiogenic shock or other hemodynamic col- risk of bleeding and procedure failure [75, 76].
lapse, which results in hepatocellular necrosis. Nevertheless, all candidates for LVAD therapy
Hepatic dysfunction is associated with increased should undergo a nutritional assessment to develop
morbidity and mortality. It is associated with coag- a strategy customized to each patient [77].
ulopathies resulting in increased risk of bleeding, Small body size defined as body surface area
vasodilation, and poor nutrition. (BSA) <1.5 m2 is considered a relative contrain-
The Model for End-Stage Liver Disease dication to LVAD implantation. This is the case
(MELD) score is a scoring system, which mea- as the older, pulsatile-flow generation LVADs
sures the progression of liver dysfunction using are larger and can only be placed in patients with
creatinine, total bilirubin, and international nor- BSA >1.5 m2. This limitation results in an under-
malized ratio (INR). The MELD score has been served population of smaller adults, women, and
shown to successfully risk stratify patients and children. The newer-generation, continuous-
predict 1-year mortality in patients with heart flow LVADs are much smaller. In fact, the
failure [70]. As previously noted, elevated AST HeartMate II is one seventh the size and one
and bilirubin have also been associated with poor quarter the weight of the older-generation
outcomes following LVAD implantation. HeartMate XVE [78]. Therefore, the HeartMate
2  Who Is an Appropriate Candidate for Long-Term MCS?: The Art of Patient Selection 25

II has been approved to be implanted in patients Age

with a lower limit of BSA at 1.2 m2 and greater.
There remains limited data available on implant- The age cutoff for cardiac transplantation varies
ing LVADs in smaller patients as few studies by transplant center. In general, it is agreed that
have a cohort of patients with a BSA <1.5 m2. patients less than 70 years old should be consid-
Ono et al. demonstrated no difference in survival ered for cardiac transplants. Patients older than
at 1 year in a cohort of 104 Japanese patients 70 years old may also be considered for trans-
with pre-LVAD BSA of <1.5 m2 compared to plantation, though they must be carefully selected
patients with BSA >1.5 m2. Both groups experi- [82]. Data on survival in the older post-LVAD
enced a significant decrease in NYHA class. population remains limited. Though there is no
There was no difference in rate of thromboem- age cutoff for LVAD implantation, post-LVAD
bolism; however, patients with BSA <1.5 m2 outcomes play an important factor in patient
were at increased risk of driveline infection [79]. selection, particularly as there is an increase in
Morbid obesity, defined as body mass index older patients who are referred for LVAD
(BMI) >35 kg/m2, is a relative contraindication therapy.
for LVAD implantation, as it is associated with In a retrospective analysis of 55 patients who
worse outcomes post-implantation. In a retro- received HeartMate II as either bridge to trans-
spective analysis of 3856 patients that underwent plantation or destination therapy, survival at
LVAD implantation as bridge to transplant, there 36 months was comparable between patients <70
was no difference in mortality across all BMI versus ≥70 years of age. Both group of patients
groups. However, there was a trend toward had improved quality of life and similar length of
increased risk of infection and thromboembolism stay with no significant differences in adverse
in patients with a higher BMI [80]. Another study effects [84]. Another retrospective analysis evalu-
also showed no differences in 1- and 2-year sur- ated outcomes in 128 patients implanted with con-
vival after LVAD implantation in all BMI groups, tinuous-flow LVAD for either bridge to
though obese patients had a higher incidence of transplantation or destination therapy. Patients
sepsis and device-related infection as well as were divided into patients <65 years of age and
higher rate of rehospitalization [75]. The higher ≥65 years of age. In the bridge to transplantation
incidences of driveline infection may be second- group, patients ≥65 years old had a longer length
ary to excess adipose tissue in the abdominal of stay in the ICU. However, there was no differ-
area, resulting in less blood supply to the drive- ence in the two groups in terms of 1-year survival
line area and therefore poor wound healing [81]. as well as 30-day adverse events including infec-
Morbid obesity is also a relative contraindica- tion, re-exploration for bleeding, ischemic or hem-
tion for heart transplantation as it is associated orrhagic stroke, and renal failure. In the destination
with worse outcomes posttransplantation. therapy group, patients ≥65 years old had a higher
Therefore, weight loss is recommended to achieve incidence of stroke. However, there were no sig-
a BMI ≤35 kg/m2 before listing for cardiac trans- nificant differences in terms of survival at 2 years,
plantation [82]. A small study evaluated LVAD length of stay, and other adverse events. In other
implantation in 19 obese patients with advanced studies, there was a trend toward decreased sur-
heart failure as a “bridge to weight loss.” Majority vival at 6 and 12 months in the older population,
of obese patients experienced significant weight though this did not reach statistical significance.
loss post-implantation and therefore reversal of Nonetheless, adverse events remained comparable
their eligibility for cardiac transplantation [83]. in both age groups [85, 86]. These studies suggest
However, this was contradicted in a larger study in that age alone should not be used as criteria for
which only 15% of obese patients were able to exclusion of LVAD therapy.
lose enough weight to be recategorized in a lower It is felt that frailty rather than age may be a
BMI group [80]. better measure to aid in patient selection for
26 C.S.C. Lai and A.B. Civitello

LVAD therapy, as age itself has not demonstrated Rockwood index, gait speed alone was the stron-
to be associated with worse outcomes. Though gest predictor of 6-month mortality in patients
frailty is associated with advanced age, it is not with coronary artery disease [91].
confined to old age. Likewise, advanced age does Frailty was applied to 99 patients undergoing
not equate frailty. Frailty is defined as a “decrease LVAD implantation as destination therapy. Frailty
reserve and reactivity to internal and external was fined by the “deficit index,” which consisted
stressors—physically, psychologically, and of 31 impairments, disabilities, and comorbidi-
socially” [87]. Markers of frailty include decline ties. Patients were divided into tertiles based on
in lean body mass, strength, endurance, balance, deficit index ranging from frail, intermediate
walking performance, and low activity. Frailty frail, and not frail. The study observed a stepwise
has been shown to be predictive of both short- increase in 1-year mortality with increasing defi-
and long-term mortality, disability, and hospital- cit index. In fact, those in the highest frailty ter-
ization when applied to patients with heart tile had a threefold increase in death compared to
failure, coronary artery disease, and percutane- those who were not frail. Frailty prior to LVAD
ous coronary intervention [88]. implantation was also associated with increased
Several screening tools are available to identify risk of rehospitalization [92].
frailty, though this remains difficult as no gold stan-
dard definition exists. Perhaps the most widely used
criteria to identify frailty is the Fried frailty pheno- Psychosocial Evaluation
type. This was developed based on observations of
progressive weakness and decline in activity in Psychosocial morbidity is common in patients
older adults. Frailty is defined with three or more of with heart failure. Depression is the most com-
the following criteria: (1) weight loss, (2) weakness, mon issue with a prevalence ranging from 15 to
(3) poor endurance, (4) slow gait speed, and (5) low 36% [93]. Risk factors for depression in this pop-
physical activity (Table 2.5). The Fried frailty phe- ulation include female gender, living alone, and
notype has been shown to be predictive of falls, hos- poor social support [94]. Depression is associated
pitalization, disability, and mortality [89]. with increased mortality. Vaccarino et al. demon-
The Rockwood index is another screening tool strated patients with higher level of depressive
to diagnose frailty. This uses multiple domains symptoms had a higher rate of either functional
including disability, comorbidities, nutritional sta- decline or death at 6 months. Though noncompli-
tus, cognitive function, and physical performance. ance with medical regime likely contributes to
Though effective, this tool may be too difficult to increased mortality, it is not felt to play a signifi-
apply in the clinical setting [90]. A simpler tool cant role. Rather, depression may worsen heart
would be to use gait speed as a single measure of failure prognosis via direct physiologic mecha-
frailty. When compared with the Fried criteria and nisms [95]. Other psychiatric morbidities includ-
ing anxiety, prior suicide attempts, drug or
Table 2.5  Fried frailty phenotype [89] alcohol dependence or history of rehabilitation,
Weight loss Unintentional of >10 lb in prior
and prior psychiatric hospitalization are also
year or ≥5% of body weight in associated with poor survival [96].
prior year Mood and anxiety disorders are also common
Weakness Grip strength in the lower 20% at after cardiac transplantation, particularly in the
baseline, adjusted for gender and first year posttransplantation. Predictors of poor
BMI
functioning posttransplantation include poor
Poor endurance Indicated by self-report of
and energy exhaustion physical functioning at the time of cardiac trans-
Slow gait speed >6–7 s to walk 15 ft plant, pretransplant history of psychiatric disor-
Low physical Weighted scores of kilocalories der, poor social support, poor coping strategies,
activity expended per week lower sense of personal control, and lower opti-
Frailty defined with three or more of the above criteria mism. Collectively, this is associated with
2  Who Is an Appropriate Candidate for Long-Term MCS?: The Art of Patient Selection 27

increased mortality and risk of acute and chronic Interestingly, when this cohort of patients was
graft rejection and graft loss [97]. compared to patients with restrictive or hypertro-
Psychological evaluation prior to LVAD phic cardiomyopathy without LVAD therapy, the
implantation is essential for patient selection. LVAD group had slightly improved survival [99].
Due to the complexity of care, candidates for
LVAD must have both the cognitive and psycho-
social abilities to care for themselves after receiv- Valvular Heart Disease
ing LVAD therapy. It is recommended that
psychosocial selection criteria for LVAD implan- Mechanical aortic valve is considered a relative
tation should follow heart failure transplant contraindication to LVAD placement. Aortic valve
guidelines [98]. immobility leads to stagnant blood flow near the
valve, therefore increasing risk of thrombus forma-
tion and thromboembolism [16]. It is recommended
Structural Heart Disease that the mechanical valve be replaced with a bio-
prosthetic valve or the aortic valve prosthesis be
Left Ventricular Structure patched with pericardium during LVAD insertion,
and Function though there is no consensus on which method is
preferred. One small case series showed replacing
It is felt that dilated ventricles are better suited for a mechanical valve with a bioprosthetic did not
LVAD implantation. Dilated ventricles allow the offer any additional protection against thrombosis,
inflow cannula to be seated in the long axis of the and therefore patching the aortic valve prosthesis
left ventricle and avoid contact with the ventricu- may be more beneficial [100, 101]. Another small
lar septum or free wall [16]. Patients with restric- case series examined outcomes between mechani-
tive or hypertrophic cardiomyopathy have not cal and bioprosthetic aortic valves at the time of
been represented in LVAD trials, and therefore LVAD implantation. Incidence of thromboembo-
data remains limited. Topilsky et al. demon- lism was low and comparable in both groups [102].
strated LVAD implantation in this population Similar results were seen in other small studies,
may be feasible. In this study, eight patients with where survival rates were similar to patients with-
either restrictive or hypertrophic cardiomyopa- out mechanical valves [103–105].
thy, who received LVAD placement, were evalu- Aortic stenosis does not require surgical inter-
ated. Myomectomy was performed at the time of vention as LVAD flow is not dependent on the
LVAD insertion and enabled the placement of the flow through the aortic valve [56]. More than
inflow cannula. When this group was compared mild aortic regurgitation on the other hand should
to patients with dilated cardiomyopathy, there be addressed at the time of LVAD implantation.
was no difference in 1-year survival rates. Lower left ventricular pressures following LVAD
However, the patients with restrictive or hyper- implantation create a greater pressure gradient
trophic cardiomyopathy had increased rates of across the aortic valve, thereby worsening aortic
right ventricular dysfunction as suggested by regurgitation. The regurgitate flow produces a
right atrial pressures, decreased pump flow, and closed-loop system, which leaves the device inef-
increased duration of inotrope use. One explana- fective for hemodynamic support. Furthermore,
tion of this finding is the possibility of myopathic aortic insufficiency has been observed to p­ rogress
involvement of the right ventricle and therefore even after LVAD implantation [106]. Diminished
preexisting pulmonary hypertension. Another flow through the aortic valve results in dimin-
explanation is increased “suck-down” events sec- ished valve motion and fusion of the aortic valve
ondary to contact between the inflow cannula and commissures, resulting in worsening aortic valve
intraventricular septum. Finally, there is concern insufficiency [107]. Current strategies to rectify
that the myocardium may be too stiff; therefore, aortic insufficiency include aortic valve closure,
“suck-down” events may not be detected. repair, and replacement with bioprosthesis. In a
28 C.S.C. Lai and A.B. Civitello

retrospective analysis of patients with aortic between both groups [111]. However, these
insufficiency who underwent aortic valve proce- results could not be replicated by other studies.
dure, aortic valve closure was associated with the Saeed et al. demonstrated no difference in the
highest 1-year mortality, followed by replace- incidence of right ventricular failure as well as
ment, and then repair. Patients with aortic valve long-term survival between patients who received
closure were more susceptible due to pump dys- concomitant TVP and LVAD implantation alone
function as patients in this group become com- [112]. In addition, TVP has been also shown to
pletely dependent on the pump [108]. be associated with increased risk of postoperative
Moderate to severe mitral stenosis should be renal failure, greater transfusion requirement,
replaced with a prosthetic valve at the time of reoperation, prolonged ventilation, prolonged
LVAD implantation. Mitral stenosis after LVAD ICU stay, and prolonged hospitalization. These
implantation results in reduced left ventricular inconsistences may reflect different management
filling and decreased LVAD flow. Pulmonary strategies for preoperative optimization in
artery pressure also remains elevated, increasing patients at risk of right ventricular failure [59].
the risk of right ventricular failure. Mitral regur- Alternatively, there is also evidence that LVAD
gitation on the other hand does not generally implantation may improve tricuspid regurgitation
require surgical intervention. LVAD implantation in a subset of patients due to left ventricular
results in decreased left ventricular pressure, unloading and improvement in pulmonary artery
thereby reducing the severity of the regurgitant pressures [113, 114]. Altogether, there has been
volume. A prosthetic or mechanical mitral valve no demonstration of a significant survival benefit
is not considered a contraindication for LVAD to concomitant TVP. Though TVP improves the
implantation. There has been no increased risk of severity of tricuspid regurgitation, the data has
thrombus formation demonstrated [109]. failed to show a clinical benefit, and the risks and
Significant tricuspid has been shown to be a benefits of TVP should be carefully weighted
predictor of right ventricular failure after LVAD before LVAD implantation.
implantation. Correction of moderate or greater
tricuspid regurgitation with either tricuspid valve
repair or replacement at the time of LVAD Congenital Heart Disease
implantation has been proposed to reduce the risk
of right ventricular failure postoperatively. Intracardiac shunts including patent foramen
Maltais et al. demonstrated that tricuspid valve ovale and atrial septal defects should be closed at
procedure (TVP) for moderate-severe tricuspid time of LVAD implantation. Reduced left ven-
regurgitation produced right ventricular geome- tricular filling following implantation may result
try changes consistent with reverse remodeling in right-to-left shunting and therefore thrombo-
and improved right ventricular function [110]. embolism [115, 116].
The clinical benefits of concomitant tricuspid LVAD use in the adult congenital heart disease
valve procedure were examined in a small cohort population is limited without established out-
of patients with significant tricuspid regurgitation comes. Furthermore, implantation in this popula-
who received continuous-flow LVAD. tion may not be possible in this population due to
Echocardiography after 1 month of implantation their complex anatomy, presence of pulmonary
demonstrated reduced tricuspid regurgitation and hypertension and biventricular heart failure, and
right ventricular volume in the group who prior cardiac procedures. In patients following an
received concomitant TVP compared to the atrial switch procedure with a failing systemic
group who received LVAD alone. Postoperative right ventricle, placement of a VAD into the right
right ventricular failure was also reduced in the ventricle was shown to be successful. This is
TVP group. On the other hand, there was no dif- technically challenging though, as the right ven-
ference in duration of hospitalization, need for tricle apex is not as well developed as the left
rehospitalization, and 30-day or 1-year mortality ventricle apex. Furthermore, the trabeculae and
2  Who Is an Appropriate Candidate for Long-Term MCS?: The Art of Patient Selection 29

moderator band may obstruct the inflow cannula This has led to an increasing trend to offer
and must be carefully resected [117, 118]. mechanical circulatory support to patients with
less severe heart failure. Several studies have
demonstrated patients who are inotrope depen-
Infections dent have superior outcomes with LVAD implan-
tation when compared to patients receiving
Active infection is a contraindication for LVAD optimal medical therapy [10, 122]. However, out-
placement. Assessment for risk of infection must comes in LVAD implantation in INTERMACS
be assessed prior to LVAD implantation. This profile 4–7 are less established. The ROADMAP
includes evaluating for leukocytosis or leukope- study is a prospective, nonrandomized, observa-
nia, recent infections, BMI >40 kg/m2, dental tional study comparing LVAD implantation ver-
exam, indwelling catheters, and poor nutrition sus optimal medical therapy in patients who are
which is defined by prealbumin <15 mg/dL and high risk but not inotrope dependent
BMI <20 kg/m2 [119]. (INTERMACS 4–7). At 1 year, survival in the
Infection is one of the most frequent adverse LVAD group was significantly higher than the
events following LVAD implantation. Driveline optimal medical therapy group. However, this did
infections are the most common type of LVAD-­ not reach statistical significance on the intention
associated infection. However, the severity of to treat analysis. The LVAD group also experi-
infection can range from driveline or pump enced greater improvement and quality of life.
pocket infection to sepsis. Expectedly, adverse events were more common
in the LVAD group. Bleeding was the primary
driver in the LVAD group, while worsening heart
Moving to Patients with Less Severe failure was the primary driver in the optimal
Heart Failure medical therapy group [123]. Earlier implanta-
tion may produce survival benefit and improve
LVAD implantation was initially reserved for quality of life at the cost of increased adverse
patients with hemodynamic compromise and events, primarily bleeding.
refractory heart failure, with the majority of
recipients categorized as INTERMACS profile
1 and 2. However, this group is also associated References
with the highest 12-month mortality [9]. In a
HeartMate II post-approval study, those with an 1. Benjamin EJ, Blaha MJ, Chiuve SE, et al. Heart
disease and stroke statistics—2017 update: a report
INTERMACS profile 1–3 had a significantly from the American Heart Association. Circulation.
lower survival at 24 months compared to 2017;135(10):e146–603.
patients in the INTERMACS 4–7 profile [120]. 2. Metra M, Carubelli V, Ravera A, et al. Heart fail-
When patients with LVAD placement were ure 2016: still more questions than answers. Int
J Cardiol. 2017;227:766–77.
stratified by INTERMACS profile, more 3. Gustafsson F, Rogers JG. Left ventricular
patients with a higher INTERMACS profile assist device therapy in advanced heart failure:
survived to discharge compared with critically patient selection and outcomes. Eur J Heart Fail.
ill patients with a lower profile. Patients with 2017;19(5):595–602.
4. Metra M, Ponikowski P, Dickstein K, et al. Advanced
higher INTERMACS profile also had a shorter chronic heart failure: a position statement from
length of stay [18]. Earlier implantation also the Study Group on Advanced Heart Failure of the
confers other advantages such as reduced noso- Heart Failure Association of the European Society
comial infection and improved functional of Cardiology. Eur J Heart Fail. 2007;9:684–94.
5. Pleura JL, Colvin-Adams M, Francis GS, et al.
capacity prior to implantation [14]. Furthermore, Recommendations for the use of mechanical circu-
health-related quality of life is significantly latory support: device strategies and patient selec-
improved with LVAD therapy in all tion. A scientific statement from the American Heart
INTERMACS profiles [121]. Association. Circulation. 2012;126:2648–67.
30 C.S.C. Lai and A.B. Civitello

6. Frazier OH, Rose EA, McCarthy P, et al. Improved 21. Levy WC, Mozaffarian D, Linker DT, et al. The
mortality and rehabilitation of transplant candidates Seattle Heart Failure Model: prediction of survival
treated with long-term implantable left ventricular in heart failure. Circulation. 2006;113:1421–33.
assist system. Ann Surg. 1995;222:327–38. 22. Kalogeropoulos AP, Gerogiopoulou VV, Giamouzis
7. Frazier OH, Rose EA, Oz MC, et al. Multicenter G, et al. Utility of the Seattle heart failure model
clinical evaluation of the HeartMate vented elec- in patients with advanced heart failure. J Am Coll
tric left ventricular assist system in patients await- Cardiol. 2009;120:835–42.
ing heart transplantation. J Thorac Cardiovasc Surg. 23. Levy WC, Mozaffarian D, Linker DT, et al. Can the
2001;122:1186–95. Seattle Heart Failure Model be used to risk-stratify heart
8. Pagani FD, Miller LW, Russell SD, et al. Extended failure patients for potential left ventricular assist device
mechanical circulatory support with a continuous-­ therapy? J Heart Lung Transplant. 2009;28:231–6.
flow rotary left ventricular assist device. J Am Coll 24. Ketchum ES, Moorman AJ, Fishbein DP, et al.
Cardiol. 2009;54:312–21. Predictive value of the Seattle Heart Failure Model
9. Kirklin JK, Naftel DC, Pagani FD, et al. in patients undergoing left ventricular assist device
Seventh INTERMACS annual report: 15,000 placement. J Heart Lung Transplant. 2010;29:1021–5.
patients and counting. J Heart Lung Transplant. 25. Lietz K, Long JQ, Kfoury AG, et al. Outcomes of
2015;34:1495–504. left ventricular assist device implantation as destina-
10. Rose EA, Gelijns AC, Moskowitz AJ, et al. Long-­ tion therapy in post-REMATCH era: implications for
term use of a left ventricular assist device for end-­ patient selection. Circulation. 2007;116:497–505.
stage heart failure. N Engl J Med. 2001;345:1435–43. 26. Teuteberg J, Ewald G, Adamson R, et al. Risk
11. Slaugther MS, Rogers JG, Milano CA, et al. assessment for continuous flow left ventricular assist
Advanced heart failure treated with continuous-­ devices: does the destination therapy risk score
flow left ventricular assist device. N Engl J Med. work? J Am Coll Cardiol. 2012;60:44–51.
2009;361:2241–51. 27. Cowger J, Sundareswaran K, Rogers JG, et al.
12. Starling RC, Naka Y, Boyle AJ, et al. Results of the Predicting survival in patients receiving continu-
post-U.S. Food and Drug Administration-approval ous flow left ventricular assist devices. J Am Coll
study with a continuous flow left ventricular assist Cardiol. 2013;61:313–21.
device as a bridge to heart transplantation: a pro- 28. Thomas SS, Nahumi N, Han J, et al. Pre-operative
spective study using the INTERMACS (Interagency mortality risk assessment in patients with
Registry for Mechanically Assisted Circulatory continuous-­flow left ventricular assist devices: appli-
Support). J Am Coll Cardiol. 2011;57:1890–8. cations of the HeartMate II risk score. J Heart Lung
13. Williams ML, Trivedi JR, McCants KC, et al. Heart Transplant. 2014;33:675–81.
transplantation vs left ventricular assist device in 29. Schaffer JM, Allen JG, Weiss ES, et al. Evaluation of
heart-transplant-eligible patients. Ann Thorac Surg. risk indices in continuous-flow left ventricular assist
2011;91:1330–3. device patients. Ann Thorac Surg. 2009;88:18889–96.
14. Miller LW, Guglin M. Patient selection for ventricular 30. Galie N, Humbert M, Vachiery JL, et al. 2015 ESC/
assist devices. J Am Coll Cardiol. 2013;61:1209–21. ERS Guidelines for the diagnosis and treatment
15. Center for Medicare & Medicaid Services. Medicare of pulmonary hypertension: The Joint Task Force
national coverage determinations manual. 2010. for the Diagnosis and Treatment of Pulmonary
[Link] Hypertension of the European Society of Cardiology
[Link]. Accessed September 27, 2016. (ESC) and the European Respiratory Society (ERS):
16. Khazanie P, Rogers JG. Patient selection for left ventric- Endorsed by: Association for European Paediatric
ular assist device. Congest Heart Fail. 2011;17:227–34. and Congenital Cardiology (AEPC), International
17. Alba AC, Rao V, Ross HJ, et al. Usefulness of the Society for Heart and Lung Transplantation (ISHLT).
INTERMACS scale to predict outcomes after Eur Heart J. 2016;37:67–119.
mechanical assist device implantation. J Heart Lung 31. Mehra MR, Kobashigawa J, Starling R, et al. Listing
Transplant. 2009;28:827–33. criteria for heart transplantation: International Society
18. Boyle AJ, Ascheim DD, Russo MJ, et al. Clinical for Heart and Lung Transplantation guidelines for
outcomes for continuous-flow left ventricular the care of cardiac transplant candidates—2006.
assist device patients stratified by pre-operative J Heart Lung Transplant. 2006;25:1024–42.
INTERMACS classification. J Am Coll Cardiol. 32. Bourge RC, Naftel DC, Costanzo-Nordin MR,
2013;61:313–21. et al. Pretransplantation risk factors for death after
19. Knaus WA, Draper EA, Wagner DP, et al. APACHE heart transplantation: a multiinstitutional study. The
II: a severity of disease classification system. Crit Transplant Cardiologists Research Database Group.
Care Med. 1985;13:818–29. J Heart Lung Transplant. 1993;12:549–62.
20. Gracin N, Johnson MR, Spokas D, et al. The use of 33. Martin J, Siegenthaler MP, Friesewinkel O, et al.
APACHE II scores to select candidates for left ven- Implantable left ventricular assist device for treat-
tricular assist device placement. Acute Physiology ment of pulmonary hypertension in candidates for
and Chronic Health Evaluation. J Heart Lung orthotopic heart transplant—a preliminary study.
Transplant. 1998;17:1017–23. Eur J Cardiothorac Surg. 2004;25:971–7.
2  Who Is an Appropriate Candidate for Long-Term MCS?: The Art of Patient Selection 31

34. Salzberg S, Lachat ML, Harbou KV, et al. 48. Korabathina R, Heffernan KS, Paruchuri V, et al.
Normalization of high pulmonary vascular resistance The pulmonary artery pulsatility index identifies
with LVAD support in heart transplantation candi- severe right ventricular dysfunction in acute inferior
dates. Eur J Cardiothorac Surg. 2005;27:222–5. myocardial infarction. Catheter Cardiovasc Interv.
35. Knutty RS, Parameshwar J, Lewis C, et al. Use of 2012;80:593–600.
centrifugal left ventricular assist device as bridge 49. Kang G, Ha R, Banerjee D. Pulmonary artery pulsa-
to candidacy in severe heart failure with secondary tility index predicts right ventricular failure after left
pulmonary hypertension. Eur J Cardiothorac Surg. ventricular assist device implantation. J Heart Lung
2013;43:1237–42. Transplant. 2016;35:67–73.
36. Tsuashita M, Takayama H, Takeda K, et al. Effect of 50. Morine KJ, Kiernan MS, Pham DT, et al. Pulmonary
pulmonary vascular resistance before left ventricular Artery Pulsatility Index is associated with right ven-
assist device implantation on short-and long-term tricular failure after left ventricular assist device sur-
post-transplant survival. J Thorac Cardiovasc Surg. gery. J Card Fail. 2016;22:110–6.
2015;150:1352–60. 51. Puwanant S, Hamilton KK, Klodell CT, et al. Tricuspid
37. Kavarana MN, Pessin-Minsley MS, Urtecho J, et al. annular motion as a predictor of severe right ventricu-
Right ventricular dysfunction and organ failure in lar failure after left ventricular assist device implanta-
left ventricular assist device recipients: a continuing tion. J Heart Lung Transplant. 2008;27:1102–7.
problem. Ann Thorac Surg. 2002;73:745–50. 52. Raina A, Seetha Rammohan HR, Gertz ZM, et al.
38. Deng MC, Edwards LB, Hertz MI, et al. Mechanical Postoperative right ventricular failure after left
circulatory support device database of the International ventricular assist device placement is predicted by
Society for Heart and Lung Transplantation: third preoperative echocardiographic structural, hemo-
annual report—2005. J Heart Lung Transplant. dynamic, and functional parameters. J Card Fail.
2005;24:1182–7. 2013;19:16–24.
39. Lampert BC, Teuteberg JJ. Right ventricular fail- 53. Potapov EV, Stepanenkko A, Dandel M, et al.
ure after left ventricular assist device. J Heart Lung Tricuspid incompetence and geometry of the right
Transplant. 2015;34:1123–30. ventricle as predictors of right ventricular function
40. Hayek S, Sims DB, Markham DW, et al. Assessment of after implantation of a left ventricular assist device.
right ventricular function in left ventricular assist device J Heart Lung Transplant. 2008;27:1275–81.
candidates. Circ Cardiovasc Imaging. 2014;7:379–89. 54. Grant AD, Smedira NG, Starling RC, et al.
41. Interagency Registry for Mechanically Assisted Independent and incremental role of quantitative
Circulatory Support (INTERMACS). Appendix A: right ventricular evaluation for the prediction of right
Adverse event definitions: adult and pediatric patients. ventricular failure after left ventricular assist device
2013. [Link] implantation. J Am Coll Cardiol. 2012;60:521–8.
images/protocol_4.0/protocol_4.0_MoP/Appendix_A_ 55. Kalogeropoulos AP, Al-Anbari R, Pekarek A, et al.
INTERMACS_AE_Definitions__05152013.docx. The right ventricular function after left ventricular
Accessed September 28, 2016. assist device (RVF-LVAD) study: rationale and pre-
42. Ochiai Y, McCarthy PM, Smedira NG, et al. Predictors liminary results. Eur Heart J Cardiovasc Imaging.
of severe right ventricular failure after implantable 2016;17:429–37.
left ventricular assist device insertion: analysis of 245 56. Feldman D, Pamboukian SV, Teuteberg JJ, et al.
patients. Circulation. 2002;106:I192–202. The 2013 International Society for Heart and Lung
43. Dang NC, Topkara VK, Mercando M, et al. Right Transplantation Guidelines for mechanical circu-
heart failure after left ventricular assist device latory support: executive summary. J Heart Lung
implantation in patients with chronic congestive Transplant. 2013;32:157–87.
heart failure. J Heart Lung Transplant. 2006;25:1–6. 57. Romano MA, Cowger J, Aaronson KD, et al.
44. Drakos SG, Janicki L, Horne BD, et al. Risk fac- Diagnosis and management of right-sided heart
tors predictive of right ventricular failure after left failure in subjects supported with left ventricular
ventricular assist device implantation. Am J Cardiol. assist devices. Curr Treat Options Cardiovasc Med.
2010;105:1030–5. 2010;12:420–30.
45. Fitzpatrick JR 3rd, Frederick JR, Hsu VM, et al. 58. Fitzpatrick JR, Frederick JR, Hiesinger W, et al.
Risk score derived from pre-operative data analy- Early planned institution of biventricular mechani-
sis predicts the need for biventricular mechani- cal circulatory support results in improved outcomes
cal circulatory support. J Heart Lung Transplant. compared with delayed conversion of a left ven-
2008;27:1286–92. tricular assist device to a biventricular assist device.
46. Matthews JG, Koelling TM, Pagani FD, et al. J Thorac Cardiovasc Surg. 2009;137:971–7.
The right ventricular failure risk score. J Am Coll 59. Robertson JO, Grau-Sepulveda MV, Okada S, et al.
Cardiol. 2008;51:2163–72. Concomitant tricuspid valve surgery during implan-
47. Kormos RL, Teuteberg JJ, Pagani FD, et al. Right tation of continuous-flow left ventricular assist
ventricular failure in patients with HeartMate II devices: a Society of Thoracic Surgeons database
continuous-flow left ventricular assist device: inci- analysis. J Heart Lung Transplant. 2014;33:609–17.
dence, risk factors, and effect on outcomes. J Thorac 60. Zannad F, Mabazaa A, Juilliere Y, et al. Clinical
Cardiovasc Surg. 2010;139:1316–24. profile, contemporary management and one-year
32 C.S.C. Lai and A.B. Civitello

mortality in patients with severe acute heart fail- 76. Musci M, Loforte A, Potapov EV, et al. Body mass
ure syndromes: the EFICA study. Eur J Heart Fail. index and outcome after ventricular assist device
2006;8:697–705. placement. Ann Thorac Surg. 2008;86:1236–42.
61. Smith GL, Lichtman JH, Bracken MB, et al. Renal 77. Holdy K, Dembitsku W, Eaton LL, et al. Nutrition
impairment and outcomes in heart failure: sys- assessment and management of left ventricular
temic review and meta-analysis. J Am Card Coll. assist device patients. J Heart Lung Transplant.
2006;47:1987–96. 2005;24:1690–6.
62. Damman K, Navis G, Voors AV, et al. Worsening 78. John R. Current axial-flow devices—the HeartMateII
renal function and prognosis in heart failure: sys- and Jarvik 2000 left ventricular assist devices. Semin
tematic review and meta-analysis. J Card Fail. Thorac Cardiovasc Surg. 2008;20:264–72.
2007;13:599–608. 79. Ono M, Sawa Y, Nakatani T, et al. Japanese multi-
63. Damman DK, Testani JM. The kidney in heart fail- center outcomes with the HeartMate II left ventricu-
ure: an update. Eur Heart J. 2015;36:1437–44. lar assist device in patients with small body surface
64. Butler J, Geisberg C, Howser R, et al. Relationship area. Circ J. 2016;80:1931–6.
between renal function and left ventricular assist 80. Clerkin KJ, Naka Y, Mancini DM, et al. The impact
device use. Ann Thorac Surg. 2006;81:1745–51. of obesity on patients bridged to transplantation
65. Sandner SE, Zimpfer D, Zrunek P, et al. Renal func- with continuous-flow left ventricular assist devices.
tion and outcome after continuous flow left ven- JAAC Heart Fail. 2016;4:761–8.
tricular assist device implantation. Ann Thorac Surg. 81. Raymond AL, Kfoury AG, Bishop CJ, et al. Obesity
2009;87:1072–8. and left ventricular assist device driveline exit site
66. Hasin T, Topilsky Y, Schirger JA, et al. Changes in infection. ASAIO. 2010;56:57–60.
renal function after implantation of continuous-flow 82. Mehra MR, Canter CE, Hannan MM, et al. The 2016
left ventricular assist devices. J Am Coll Cardiol. International Society for Heart Lung Transplantation
2012;59:26–36. listing criteria for heart transplantation: a 10-year
67. Genovese EA, Dew MA, Teuteberg JJ, et al. Early update. J Heart Lung Transplant. 2016;35:1–23.
adverse events as predictors of 1-year mortality dur- 83. Dhesi P, Simsir SA, Daneshvar D, et al. Left ventric-
ing mechanical circulatory support. J Heart Lung ular assist devices as “bridge to weight loss” prior to
Transplant. 2010;29:981–8. transplantation in obese patients with advanced heart
68. Coffin ST, Waquespack DR, Haglund NA, et al. Kidney failure. Ann Transplant. 2011;16:5–13.
dysfunction and left ventricular assist device support: a 84. Adamson RM, Stahovich M, Chillcott S, et al.
comprehensive perioperative review. Cardiorenal Med. Clinical strategies and outcomes in advanced heart
2015;5:48–60. failure patients older than 70 years of age receiv-
69. Patel AM, Adeseun GA, Ahmed I, et al. Renal failure ing the HeartMate II left ventricular assist device: a
in patients with left ventricular assist devices. Clin community hospital experience. J Am Coll Cardiol.
J Am Soc Nephrol. 2013;8:484–96. 2011;57:2487–95.
70. Kim MS, Kato TS, Wu C, et al. Hepatic dysfunction 85. Sandner SE, Zimpfer D, Zrunek P, et al. Age and
in ambulatory patients with heart failure: a applica- outcome after continuous-flow left ventricular assist
tion of the MELD scoring system for outcome pre- device implantation as bridge to transplantation.
diction. J Am Coll Cardiol. 2013;61:2253–61. J Heart Lung Transplant. 2009;28:367–72.
71. Russel SD, Rogers JG, Carmelo AM, et al. Renal 86. Huang R, Deng M, Rogers JG, et al. Effect of age on
and Hepatic function improve in advanced heart outcomes after left ventricular assist device place-
failure patients during continuous-flow support ment. Transplant Proc. 2006;38:1496–8.
with the HeartMate II left ventricular assist device. 87. Sieber CC. Frailty—from concept to clinical prac-
Circulation. 2009;120:2352–7. tice. Exp Gerontol. 2017;87(Pt B):160–7.
72. Deo SV, Sharma V, Altarabsheh SE, et al. Hepatic 88. Flint KM, Matlock DD, Lindenfeld J, et al. Frailty
and renal function with successful long-term support and the selection of patients for destination ther-
on a continuous flow left ventricular assist device. apy left ventricular assist device. Circ Heart Fail.
Heart Lung Circ. 2014;23:229–33. 2012;5:286–93.
73. Horwich TB, Kalantar-Zadeh K, MacLellan 89. Fried LP, Tangen CM, Walston J, et al. Frailty in
RW, et al. Albumin levels predict survival in older adults: evidence for a phenotype. J Gerontol A
patients with systolic heart failure. Am Heart Biol Sci Med Sci. 2001;56:146–57.
J. 2008;155:883–9. 90. Rockwood K, Stolee P, McDowell I, et al. J Am
74. Lietz K, John R, Burke EA, et al. Pretransplant Geriatr Soc. 1996;44:578–82.
cachexia and morbid obesity are predictors of increased 91. Purser JL, Kuchibhatla MN, Fillenbaum GG, et al.
mortality after heart transplantation. Transplantation. Identifying frailty in hospitalized older adults with
2001;72:277–83. significant coronary artery disease. J Am Geriatr
75. Brewer RJ, Lanfear DE, Sai-Sudhaker CB, et al. Soc. 2006;54:1674–81.
Extremes of body mass index do not impact mid-­ 92. Dunlay SM, Park SJ, Joyce LD, et al. Frailty and
term survival after continuous-flow left ventricular outcomes after implantation of left ventricular
assist device implantation. J Heart Lung Transplant. assist device as destination therapy. J Heart Lung
2012;31:167–72. Transplant. 2014;33:359–65.
2  Who Is an Appropriate Candidate for Long-Term MCS?: The Art of Patient Selection 33

93. Konstam V, Moser DK, De Jong MJ. Depression and 107. Letsou GV, Connely JH, Delgado RM, et al. Is
anxiety in heart failure. J Card Fail. 2005;11:455–63. native aortic valve commissural fusion in patients
94. Tsuchihashi-Makaya M, Kato N, Chishaki A, et al. with long-term left ventricular assist devices associ-
Anxiety and poor social support are independently ated with clinically important aortic insufficiency?
associated with adverse outcomes in patients with J Heart Lung Transplant. 2006;25:395–9.
mild heart failure. Circ J. 2009;73:280–7. 108. Robertson JO, Naftel DC, Myers SL, et al.
95. Vaccarino V, Kasl SV, Abramson J, et al. Depressive Concomitant aortic valve procedures in patients
symptoms and risk of functional decline and death undergoing implantation of continuous-flow left
in patients with heart failure. J Am Coll Cardiol. ventricular assist devices: an INTERMACS database
2001;38:199–205. analysis. J Heart Lung Transplant. 2015;34:797–805.
96. Owen JE, Bonds CL, Wellisch DK. Psychiatric 109. Wang TS, Hernandez AF, Felker M, et al. Valvular
evaluations of heart transplant candidates: pre- heart disease in patients supported with left ventricu-
dicting post-transplant hospitalizations, rejec- lar assist devices. Circ Heart Fail. 2014;7:215–22.
tion episodes, and survival. Psychosomatics. 110. Maltais S, Topilsky Y, Tchantchaleishivili V, et al.
2006;47:213–22. Surgical treatment of tricuspid valve insufficiency
97. Cupples S, Des MA, Grady KL, et al. Report of the promotes early reverse remodeling in patients
Psychosocial Outcomes Workgroup of the Nursing with axial-flow left ventricular devices. J Thorac
and Social Sciences Council of the International Cardiovasc Surg. 2012;143:1370–6.
Society for Heart and Lung Transplantation: pres- 111. Piacentino V 3rd, Ganapathi AM, Stafford-Smith
ent status of research on psychosocial outcomes in M, et al. Utility of concomitant tricuspid valve pro-
cardiothoracic transplantation: review and recom- cedures for patients undergoing implantation of a
mendations for the field. J Heart Lung Transplant. continuous-flow left ventricular device. J Thorac
2006;25:716–25. Cardiovasc Surg. 2012;144:1217–21.
98. Eshelman AK, Mason S, Nemeh H, et al. LVAD 112. Saeed D, Kidambi T, Shalli S, et al. Tricuspid valve
destination therapy: applying what we know about repair with left ventricular assist device implantation: is
psychiatric evaluation and management from it warranted? J Heart Lung Transplant. 2011;30:530–5.
cardiac failure and transplant. Heart Fail Rev. 113. Atluri P, Fairman AS, MacArthur JW, et al.
2009;14:21–8. Continuous flow left ventricular assist device
99. Topilsky Y, Pereira NL, Shah DK, et al. Left ventric- implant significantly improves pulmonary hyper-
ular assist device therapy in patients with restrictive tension, right ventricular contractility, and tricuspid
and hypertrophic cardiomyopathy. Circ Heart Fail. valve competence. J Card Surg. 2013;28:770–5.
2011;4:266–75. 114. Lee S, Kamdar R, Madlon-Kay R, et al. Effects of
100. Rose AG, Connelly JH, Park SJ, et al. Total left the HeartMate II continuous-flow left ventricular
ventricular outflow tract obstruction due to left ven- assist device on right ventricular function. J Heart
tricular assist device-induced sub-aortic thrombosis Lung Transplant. 2010;29:209–15.
in 2 patients with aortic valve bioprosthesis. J Heart 115. Loforte A, Violini R, Musumeci F. Transcatheter clo-
Lung Transplant. 2003;22:594–9. sure of patent foramen ovale for hypoxemia during
101. Pelletier MP, Chang CP, Vagelos R, et al. Alternative left ventricular assist device support. J Card Surg.
approach for use of a left ventricular assist device 2012;27:528–9.
with a thrombosed prosthetic valve. J Heart Lung 116. Bartoli CR, McCants KC, Birks EJ, et al.
Transplant. 2002;21:402–4. Percutaneous closure of a patent foramen ovale to
102. Mokashi SA, Schmitto LS, James LD, et al. prevent paradoxical thromboembolism in a patient
Ventricular assist device in patients with prosthetic with a continuous-flow LVAD. J Invasive Cardiol.
heart valves. Artif Organs. 2010;34:1030–4. 2013;25:154–6.
103. Swartz MT, Lowdermilk GA, Moroney DA, 117. Wiklund L, Svensson S, Berggren H. Implantation
et al. Ventricular assist device support in patients of a left ventricular assist device, back-to-front, in
with mechanical heart valves. Ann Thorac Surg. an adolescent with a failing mustard procedure.
1999;68:2248–51. J Thorac Cardiovasc Surg. 1999;118:755–6.
104. Tisol WB, Mueller DK, Hoy FB, et al. Ventricular 118. Menachem JN, Swaminathan AC, Bashore TM, et al.
assist device use with mechanical heart valves: an Initial experience of left ventricular assist device
outcome series and literature review. Ann Thorac support for adult patients with transposition of the
Surg. 2001;72:2051–4. great vessels. Congenit Heart Dis. 2015;10:382–6.
105. Rao V, Slater JP, Edwards NM, et al. Surgical man- 119. Halbreiner MS, Soltesz E, Starling R, et al. Current prac-
agement of valvular disease in patients requiring tice in patient selection for long-term mechanical circu-
left ventricular assist device support. Ann Thorac. latory support. Curr Heart Fail Rep. 2015;12:120–9.
2001;71:1448–53. 120. Jorde UP, Kushwaha SS, Tatooles AJ, et al. Results
106. Holley CT, Fitzpatrick M, Roy SS, et al. Aortic of the destination therapy post-food and drug admin-
insufficiency in continuous-flow left ventricu- istration approval study with continuous flow left
lar assist device support patients is common but ventricular assist device: a prospective study using
does not impact long-term mortality. J Heart Lung the INTERMACS registry. J Am Coll Cardiol.
Transplant. 2017;36(1):91–6. 2014;63:1751–7.
34 C.S.C. Lai and A.B. Civitello

121. Grady KL, Naftel D, Stevenson L, et al. Overall ­transplant candidates: results of the INTrEPID Trial.
Quality of life improves to similar levels after J Am Coll Cardiol. 2007;50:741–7.
mechanical circulatory support regardless of sever- 123. Estep JD, Starling RC, Horstmanshof DA, et al.
ity of heart failure before implantation. J Heart Lung Risk assessment and comparative effectiveness of
Transplant. 2014;33:412–21. left ventricular assist device and medical manage-
122. Rogers JG, Butler J, Lansman SL, et al. Chronic ment in ambulatory heart failure patients: results
mechanical circulatory support for inotrope-­ from the ROADMAP study. J Am Coll Cardiol.
dependent heart failure patients who are not 2015;66:1747–61.
 Optimization of Right Ventricular
Function Preoperatively for LVAD 3
Implantation

Salman Gohar, Samar Sheth,
and Reynolds Delgado III

medical management, there has been an exponen-

Introduction tial increase in the number of mechanical circula-
tory assist devices for NYHA class IIIB/IV, stage
Heart failure (HF) continues to place a significant D heart failure [2]. The HeartMate II study
burden on the current healthcare system. In 2012, reported a significant survival advantage of
the total direct medical cost for HF was $20.9 bil- continuous-­ flow ventricular assist devices
lion, and this is expected to increase to $53.1 billion (CF-VAD) over the older pulsatile HeartMate
in 2030 (representing a 2.5-fold increase). The XVE device in a BTT strategy and opened the
majority of these costs are related to hospitalization door for continuous-­flow VADs (CF-VAD) that
[1]. Left ventricular assist devices (LVAD) were have since become standard of care [3].
developed to provide end-stage heart failure patients Since 2001, more than 15,000 LVADs have
opportunities to wait for a heart on a bridge-to-heart been implanted in humans of whom more than
transplant (BTT) strategy or improve survival for 12,000 were CF-VADs. The current 1-year and
those who were not eligible for heart transplanta- 2-year survival with CF-VADs is 80% and 70%,
tion on a destination therapy (DT) pathway. Since respectively. Survival among DT patients con-
the Randomized Evaluation of Mechanical tinues to remain poor. In the recent era, survival
Assistance for the Treatment of Congestive Heart with DT therapy at 1 and 3 years is 76% and
Failure (REMATCH) trial reported a 48% reduc- 57%, respectively [4]. One way to minimize the
tion in risk of death in patients with advanced heart risk is to affect immediate perioperative compli-
failure with the HeartMate XVE LVAD versus cations. Hence, once a patient has been selected
as a candidate for LVAD, the patients must be
optimized to prevent perioperative complica-
tions. RV failure after LVAD implantation is a
S. Gohar, M.D. • S. Sheth, M.D. serious complication associated with increased
Texas Heart Institute at Baylor St. Luke’s Medical length of ICU and hospital stay [5] and 50%
Center, Baylor College of Medicine,
Houston, TX, USA mortality at 1 year for those requiring a
e-mail: Salmangoharmd@[Link]; RVAD. From the seventh INTERMACS registry,
salmangohar196@[Link]; RVF is associated with an immediate hazard for
samarshethMD@[Link] death; however, by 3 months, death rate
R.M. Delgado III, M.D. (*) decreases dramatically, highlighting the need to
Mechanical Assist Devices, Texas Heart Institute at prevent RVF in the immediate postoperative
Baylor St. Luke’s Medical Center, Baylor College of
Medicine, Houston, TX, USA period [4]. Although LVAD design has transi-
e-mail: reynoldsdelgadomd@[Link] tioned from pulsatile to continuous-flow devices,

© Springer International Publishing AG 2018 35

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
36 S. Gohar et al.

there has been no clinically significant reduction in a non-dilated RV, even after replacing the
in RV failure [6]. Current reported incidence of free wall with a noncontractile material (Dacron
RV failure after LVAD implantation remains patch), the septum was able to maintain suffi-
between 20 and 50% depending on the patient cient RV cardiac output for hemodynamic sta-
series and is associated with overall poor clinical bility [10].
outcome.
This chapter aims to discuss the current defini-
tion of RVF after LVAD implantation as well as Normal RV Physiology
the current state of knowledge on its predictors,
outcomes, and management. We will then discuss The RV contracts by three separate mecha-
how to optimize RV function preoperatively from nisms, the first of which is an inward movement
lessons learned during the history of mechanical of the free wall followed by contraction of
support development and clinical research and, deeper longitudinal muscle fibers that draw the
finally, via our own experience at the Texas Heart tricuspid annulus toward the apex and finally
Institute. traction of the free wall from LV contraction
generating a forward stroke volume. Normal
baseline RV ejection fraction (RVEF) ranges
Normal RV Anatomy from 40 to 76% and varies based on loading
conditions [7].
The right ventricle is designed to pump blood in a RV function, like the LV, is affected by pre-
high-compliance, high-volume, and low-­pressure load, afterload, and contractility. This complex
system. It is approximately one-sixth the mass of relationship is best illustrated by differences in
LV with a paper-thin free wall and a septal wall pressure-volume curves. The RV pressure-­
shared with the LV. Unlike the conical left ventri- volume loop is more triangular in shape with the
cle, the RV is a triangular-shaped structure in sag- pulmonary valve opening early in systole once
ittal sections and crescent shaped in cross section RV pressure reaches the low pulmonary pressure.
[7]. It can be divided into a muscular trabeculated As there is very little time spent in isovolumetric
body (sinus) and a smooth infundibular outflow contraction, the loop assumes a triangular con-
tract (Fig. 3.1). The intraventricular septum (IVS) figuration in contrast to a square LV loop. This
also influences the RV shape. Under normal load- also hints to the fact that the RV performs primar-
ing and electrical conditions, the IVS is concave ily volume work [11] (Fig. 3.2).
to the LV during both systole and diastole The slope of the end-systolic pressure-volume
(Table 3.1). relationship is called ventricular elastance and is a
RV myocardium is composed of two layers: relatively load-independent measure of ventricular
a superficial layer with circumferential fibers contractility. Dell’Italia demonstrated that the nor-
running parallel to the atrioventricular groove mal maximal RV elastance was 1.3 ± 0.84 mmHg/
and a deep layer with fibers aligned longitudi- mL, four times lower than the LV. Consequently,
nally from base to apex. This differs from the the RV is more afterload sensitive than the LV
LV, where oblique fibers are superficial, longi- [12]. This is demonstrated in the acute setting (i.e.,
tudinal fibers are in the subendocardium, and massive pulmonary embolism), where RV stroke
circumferential fibers are in between. The sep- volume decreases ­ significantly with sudden
tum is shared and structurally similar to the LV, increase in pulmonary artery pressure.
providing an intimate anatomic and functional Similar to the LV, in a normal RV, based on the
relationship between the two, a basis for ven- Frank-Starling principle, an increase in preload
tricular interdependence. Multiples studies improves contraction. However, in RV failure, the
have indicated that septal contribution to RV curves flatten and move down and to the right
cardiac output is significant, ranging between depict a drop in RV output with increase in pre-
20 and 60% [8, 9]. Hoffman demonstrated that load (Fig. 3.3).
3  Optimization of Right Ventricular Function Preoperatively for LVAD Implantation 37

Fig. 3.1 (a) The inlet, trabeculated apical myocardium Quintana, University of Extremadura, Spain). SMT indi-
and infundibulum of the RV. The tricuspid and pulmonary cates septomarginal trabeculation with its anterior (a) and
valves are separated by the ventriculoinfundibular fold posterior (p) arm; A-S anterosuperior leaflet of the tricus-
(VIF). (b) Short-axis plane of the RV demonstrating its pid valve, PT pulmonary trunk, Ao aorta, RA right atrium,
crescentic shape. (c) The four-chamber anatomic plane of LA left atrium (Reproduced with permission from Ho SY,
the heart showing the moderator band (MB) and the more Nihoyannopoulos P. Anatomy, echocardiography, and
apical insertion of the tricuspid valve. (d) Superficial normal right ventricular dimensions. Heart. Copyright ©
muscle layer of the RV (dissection by Damian Sanchez-­ 2006, BMJ Publishing Group Ltd)

The pericardium encases both ventricles and pathogen invasion. In addition, it also imparts a
under normal circumstances provides an important diastolic interdependence effect on both ventricles
protective barrier from overdistention and direct [13, 14]. Bernheim first hypothesized the impor-
38 S. Gohar et al.

Table 3.1  Structural findings of abnormal RV

Structural characteristics of RV
Feature Criteria Interpretation
Dilation Volume > 101 mL/m2 Volume overload
RV max SAX >43 mm Pressure overload
RVEDA/LVEDA >2/3 Intrinsic myocardial disease
D-shaped LV a
Eccentricity index >1 RV pressure or volume overload
Diastolic D-shaped LV suggests
RV volume overload
Systolic D-shaped LV is RV
pressure overload
Hypertrophy Mass > 35 g/m Pressure-overloaded RV
RV infarction wall >5 mm Hypertrophic cardiomyopathy
infiltrative disease, exclude
double-chambered RV
Aneurysm Localized RV dilation AVRD, RVMI; localized absence
of pericardium
TV septal insertion Septal insertion > 1 cm or 8 mm/m Consider Ebstein’s anomaly
Delayed hyperenhancement Area of delayed contrast uptake and Suggests myocardial fibrosis
washout in MRI
Fatty infiltration High-intensity signal on MRI Consider ARVD
Haddad, F., et al., Circulation, 2008. 117 (11): p. 1436–48 [7]
RV max SAX indicates RV maximal short-axis diameter, RVEDA/LVEDA ratio of RV to LV end-diastolic area, ARVD
arrhythmogenic RV dysplasia, RVMI RV myocardial infarction, TV tricuspid valve
a
The eccentricity index measures the degree of septal displacement and is defined as the ratio of the minor axis diameter
of the LV parallel to the septum to that perpendicular to it

CO RV
LV

RVF
6
Pressure

2 8 14 AP

Volume Fig. 3.3  Comparison of RV and LV Starling curves. LV

requires higher atrial filling pressures (AP) to produce
Fig. 3.2 The solid line depicts the square-shaped equivalent cardiac output (CO). In RV failure (RVF), the
pressure-­volume loop for the LV (solid line) and the curve moves downward and to the right.
triangular-­shaped pressure-volume loop for the RV (bro-
ken line)
the output of the other [16]. This was clinically
tance of the interventricular ­ relationship [15]. demonstrated in 1956 by Dexter as deterioration of
Henderson and Prince later demonstrated that vol- LV function in patients with atrial septal defects
ume and pressure loading of one ventricle decreased who developed RV pressure and volume overload,
3  Optimization of Right Ventricular Function Preoperatively for LVAD Implantation 39

Fig. 3.4  RV size and function. Two perpendicular sec- ventricle [LV] and RV during systole relies on interven-
tions of a 3D TEE reconstruction of the right ventricle tricular septum position as shown in cross section for dif-
from tricuspid valve [TV] to pulmonary [PV] valve are ferent clinical scenarios (Reproduced with permission
shown. The cross section [a] demonstrates the crescent from Meineri M, Van Rensburg AE, Vegas A. Right ven-
shape and the sagittal section and [b] the triangular shape tricular failure after LVAD implantation: prevention and
of the RV. Ventricular interdependence between the left treatment. Best Pract Res Clin Anaesthesiol. 2012)

a phenomenon called “reverse Bernheim effect” in RVEF but the most clinically used. Two-­
which he postulated that the leftward septal shift dimensional echo assessment by Simpson’s rule
resulted in impaired LV filling [17] (Fig. 3.4). can be used and correlates well with MRI; how-
ever, this is dependent on the quality of the images.
RV fractional change area can be measured in four-
Measuring Normal RV Function chamber views and easily incorporated into most
echo reports. Tricuspid annular plane systolic
Multiple methods have been used to measure RV excursion (TAPSE) is another useful quantitative
function clinically. Cardiac MRI is the most accu- measurement of RV systolic performance [18, 19].
rate tool to assess RV diastolic and systolic vol- RV myocardial performance index, a ratio of iso-
umes as well as the RVEF [7]. By MRI, RVEF volumic time interval to ventricular ejection time,
ranges 47–76%. A technique not used frequently is doesn’t involve ventricular geometry and is a load-
radionuclide angiography by which RVEF is usu- independent measure of RV function. Finally,
ally 40–45%. Echo is least accurate in assessing tissue Doppler imaging also allows for quantitative
40 S. Gohar et al.

Table 3.2  RV function echocardiography indices, Haddad, F., et al., Circulation, 2008. 117(11): p. 1436–48 [7]
RV contractility indices
Functional parameters Normal value Load dependencea Clinical use
RVEF, % 61 ± 7% (47–76%) +++ Clinical validation, wide
acceptance
>40–45% Prognostic value in
cardiopulmonary disorders
RVFAC, % >32% +++ Good correlates with RVEF
Prognostic value in MI and
bypass surgery
TAPSE, mm >15 +++ Simple measure, not limited by
endocardial border recognition:
Good correlation well with
RVEF
Sm annular, cm/s >12 +++ Good sensitivity and specificity
for RVEF <50%
Strain Basal: 19 ± 6 +++ Correlates with stroke volume
Mid: 27 ± 6
Apical: 32 ± 6
Strain rate, s−1 Basal: 1.50 ± 0.41 ++ Correlates with contractility
Mid: 1.72 ± 0.27
Apical: 2.04 ± 0.41
RVMPI 0.28 ± 0.04 ++ Global nongeometric index,
index of systolic and diastolic
function, prognostic value in PH
and CHD
dP/dt max, mmHg/s 100–250 ++ Not a reliable index of
contractility
More useful in assessing
directional change when preload
accounted for
IVA, m/s2 1.4 ± 0.5 + Promising new noninvasive
index of contractility, studies in
CHD
Maximal RV elastance 1.30 ± 0.84 + Most reliable index of
mmHg/mL contractility
RVFAC indicates RV fractional area change, MI myocardial infarction, TAPSE tricuspid annular plane systolic excur-
sion, Sm tissue Doppler maximal systolic velocity at the tricuspid annulus, RVMPI RV myocardial perfusion index, PH
pulmonary hypertension, CHD congenital heart disease
a
Should be viewed as a general indication of load dependence

RV assessment. Finally, myocardial strain and monary artery pulsatility index (PAPi), and a
speckle-tracking analysis can also be used to define RA to pulmonary capillary wedge pressure ratio
RV function [20, 21] (Table 3.2). have all been used to measure RV function [22]
Invasive hemodynamic assessment is the and predict RV failure post-LVAD implantation.
gold standard to assess RV function and can At the Texas Heart Institute, we routinely use
often tease out acute RVF from chronic RV dys- noninvasive parameters including TAPSE, tis-
function. Direct measurement of RA, RV, and sue Doppler imaging (TDI), PAPi, and RA/
PA pressures can clarify inconclusive noninva- PCWP ratio to aid in clinical decision-making
sive data. RV stroke work index (RVSWI), pul- (Table 3.3).
3  Optimization of Right Ventricular Function Preoperatively for LVAD Implantation 41

Table 3.3  Assessing normal RV performance using hemodynamic parameters

Parameter Formula Desirable value
RV size N/A RVEDV <200 mL, RVESV <177 mL
Central venous pressure (CVP) N/A <15 mmHg, 5 mmHg < PCWP
Transpulmonary gradient (TPG) MPAP − CVP <15 mmHg
Pulmonary vascular resistance MPAP − CVP/co <4WU
(PVR)
RV stroke work index (RVSWI) MPAP – CVP × (CI/HR) >300–600 mmHg mL/m2
Pulmonary artery pressure index PASP − PADP/CVP >2
(PAPi)
Right atrial pressure to pulmonary CVP/PCWP <0.7
capillary wedge pressure ratio
MPA mean pulmonary artery pressure, CO cardiac output, RVEDV RV end-diastolic volume, RVESV RV end-systolic
volume, PASP pulmonary artery systolic pressure, PADP pulmonary artery diastolic pressure

atinine >2.0 mg/dL) are present [29] as illus-

Right Heart Failure Definition trated in Table 3.4.
RV failure results from a number of reasons
The definition of RVF remains nebulous due to and similar to the LV can be both systolic and
the use of inconsistent criteria in different pub- diastolic in nature. While the RV can accommo-
lications. Some authors have described RVF as date increased preload, it is sensitive to increased
a need for intravenous inotrope or pulmonary afterload and elevations in PA pressures [29, 30].
vasodilator therapy for 14 days postoperatively As PA pressures increase due to worsening left-­
and/or need for RVAD, while others defined it sided function, there is delayed pulmonary valve
as simply a requirement for RVAD. Others still opening, leading to increased RV work and O2
have used two or more of the following hemo- consumption. Moreover, this leads to progressive
dynamic parameters to define RVF: central RV dilation, wall stress, and impaired coronary
venous pressure greater than 16 mmHg, mean perfusion pressure. As the dilation progresses,
arterial pressure lower than 55 mmHg, cardiac geometry changes lead to tricuspid annular dila-
index less than 2.0 L/min/m2, inotrope support tion and functional tricuspid regurgitation due to
>20 units, and mixed venous saturation lower non-coaptation of leaflets. Abnormal septal acti-
than 55%, all in the absence of cardiac tampon- vation also disrupts normal IVS function [29,
ade [23–28]. 30]. Over time, if the heart failure remains
According to the INTERMACS registry, untreated, cardiomyocyte stress and hypertrophy
RVF is present if symptoms or findings charac- lead to irreversible apoptosis [29].
terized by both elevation of central venous RV infarction due to obstructive coronary dis-
pressure (right atrial pressure > 16 mmHg on ease is another mechanism, which can lead to
right heart catheterization, significantly dilated RVF although the incidence of RV infarct post-
inferior vena cava with no inspiratory variation ­MI is low and is usually due to an isolated infe-
on echocardiography, and elevated jugular rior wall MI. The likely reason for the lower
venous pressure) and manifestations of elevated incidence of ischemic RV failure compared to
CVP (peripheral edema, ascites, or hepatomeg- ischemic LV failure is probably lower RV wall
aly on exam or diagnostic imaging and labora- stress and stroke work in addition to smaller mass
tory evidence of worsening hepatic total requiring lower resting coronary flow and O2
bilirubin >2.0 mg/dL and renal dysfunction cre- extraction [31].
42 S. Gohar et al.

Table 3.4 Interagency Registry for Mechanically Table 3.4 (continued)

Assisted Circulatory Support definition of right ventricu-
lar failure Interagency Registry for mechanically assisted
circulatory support definition of right ventricular failure
Interagency Registry for mechanically assisted Severe Death during VAD implant
circulatory support definition of right ventricular failure acute hospitalization with RVF as primary
RVF Symptoms or findings of persistent cause
definition RVF characterized by both of the CVP central venous pressure, RVF right ventricular fail-
following: ure, VAD ventricular assist device (Reproduced with per-
  – Right atrial pressure > 16 mmHg mission [29, 47])
on right heart catheterization
 – Significantly dilated inferior vena
cava with no inspiratory variation
on echocardiography
 – Elevated jugular venous pressure  athophysiology of Right Heart
P
Manifestations of elevated CVP Failure After LVAD Implantation
characterized by:
 – Peripheral edema (>2+) RV failure after LVAD implantation results from
 – Ascites or hepatomegaly on exam a complex sequence of events in the setting of
or diagnostic imaging underlying risk factors. Multiple mechanisms
 – Laboratory evidence of worsening have been suggested. LV decompression and
hepatic (total bilirubin >2.0 mg/
dL) or renal dysfunction increased cardiac output increase venous return
(creatinine >2.0 mg/dL) to the RV. Intraoperative volume resuscitation
Severity scale including blood transfusions also contributes to
Mild Patient meets both criteria for RVF this increase in RV preload and can aggravate a
plus: decompensated RV [32]. Abnormal interventric-
 – Post-implant inotropes, inhaled ular septum (IVS) geometry due to excessive
nitric oxide, or intravenous
vasodilators not continued beyond
leftward shift at high LVAD speeds can also
post-op day 7 after VAD implant worsen RV function due to loss of IVS contribu-
 – No inotropes continued beyond tion to RV output [33].
post-op day 7 after VAD implant Ischemic injury is another mechanism seen
Moderate Patient meets both criteria for RVF after prolonged bypass times, coronary ischemia,
plus: and/or loss of coronary bypass grafts or coronary
 – Post-implant inotropes, inhaled embolism. Between 30 and 64% of patients with
nitric oxide, or intravenous
vasodilators continued beyond advanced HF will have associated tricuspid
post-op day 7 and up to post-op regurgitation, which improves after LV decom-
day 14 after VAD implant patient pression with a LVAD [34, 35]. However, a
meets both criteria for RVF plus: dilated tricuspid annulus or an incompetent valve
 – CVP or right atrial
generally worsens TR after LV decompression
pressure > 16 mmHg
 – Prolonged post-implant inotropes,
and increased preload. Severe TR further contrib-
inhaled nitric oxide, or intravenous utes to RV failure through the development of
vasodilators continued beyond right-sided volume overload and reduced RV
post-op day 14 after VAD implant ejection. Although pulmonary artery pressures
Severe Patient meets both criteria for RVF improve after LV decompression, perioperative
plus:
ischemia-related pulmonary endothelial injury
 – CVP or right atrial
pressure > 16 mmHg and transfusion-related lung injury often con-
 – N eed for right ventricular assist versely increase pulmonary vascular resistance
device at any time after VAD implant and can result in RV failure.
(continued) Supraventricular arrhythmias are seen in more
than 20% of patients after LVAD implantation
3  Optimization of Right Ventricular Function Preoperatively for LVAD Implantation 43

and have been associated with twice the risk of 83% in the no-RVAD group. However, the inci-
RV failure [36]. More sinister rhythms like ven- dence and underlying mechanisms of RV failure
tricular fibrillation have been associated with a changed after the introduction of continuous-­
32% drop in cardiac output. Incessant postopera- flow LVADs (CF-LVAD) [39].
tive VT can therefore adversely affect RV func- The right ventricular failure risk score (RVFRS)
tion in LVAD patients and should be avoided as evaluated 197 patients undergoing HM II
far as possible [37]. CF-LVAD implantations. Sixty-eight cases (35%)
were complicated by postoperative RV failure.
Points were given for need for vasopressors, eleva-
 redicting Right Heart Failure
P tion in aspartate aminotransferase (>80 IU/L), bili-
After LVAD Implantation rubin (>2.0 mg/dL), and creatinine (>2.3 mg/dL).
All were found to be independent predictors of RV
Over the last three decades, many centers have failure. The odds ratios for RV failure for patients
worked to develop algorithms and risk scores to with an RVFRS of 3.0, 4.0–5.0, and 5.5 were 0.49
predict RVF after LVAD implantation. Early iden- (95% confidence interval [CI], 0.37–0.64), 2.8
tification of high-risk patients remains important (95% CI, 1.4–5.9), and 7.6 (95% CI, 3.4–17.1),
as it allows for the formulation of strategies to respectively, and 180-day survival of 90 ± 3%,
avoid RV failure. Unfortunately, most risk scores 80 ± 8%, and 66 ± 9%, respectively (P < 0.0045)
devised from retrospective, small single-­ center [40]. The different studies and RV failure risk
experiences provide a variable spectrum of pre- models are listed in Table 3.5.
dictors including hemodynamic, echocardio- The HeartMate II risk model is the only large
graphic, biochemical, and intra- and postoperative multicenter study with 484 patients, all of whom
parameters with no single model dependably received CF-LVADs with results likely applicable
forecasting RVF. Many early studies incorporated in the current era. RVF in the trial was defined
pulsatile devices in BTT cohorts and therefore did based on three groups; group 1 included those who
not accurately reflect outcomes in the current needed RVAD support postoperatively, group 2
CF-VAD era. In addition, validation of many of included those who required inotropes support for
these scores has demonstrated the modest real- ≤14 days, and group 3 included patients who
world application [38]. In our center, we have needed inotrope support for ≥14 days after implan-
noted that a preoperative, systemic inflammatory tation. Groups 1 and 2 together comprised the
syndrome associated with a leukocytosis and “early RVF” cohort, and group 3 was defined as
thrombocytopenia may prime the RV for failure. “late RVF.” The cumulative incidence of RVF was
20% with any early RVF noted in 13% of patients.
The incidence of late RVF was 7%. The model
Hemodynamic Models found the CVP/PCWP ratio > 0.63, preoperative
ventilator support, and BUN >39 mg/dL were
Fukamachi et al. reported RVAD support require- independent predictors of RVF. Actuarial survival
ment for 11 out of 100 patients after HeartMate at 1 year was also significantly better for patients
XVE pulsatile LVAD implantation. RVAD use without RVF (79%) compared with that in patients
was significantly higher in young, female patients requiring RVADs (group 1, 59%; P 1⁄4 0.004) or
with small BSA and those with myocarditis. extended inotropes (group 2, 56%; P 1⁄4 0.007),
There was no significant difference in the cardiac whereas there was no difference for patients with
index, RV ejection fraction, or right atrial pres- late inotrope use (group 3, 75%; P 1⁄4 0.81).
sure between groups preoperatively. Low preop- Decreased survival for patients with early RVF is
erative mean pulmonary arterial pressure (PAP) evident in the grouped Kaplan-Meier survival
and RV stroke work index (RVSWI) were associ- curve (Fig. 3.5). Hospital length of stay for dis-
ated with the need for post-op RVAD. Survival to charged patients was longer for those requiring an
transplant was poor in the RVAD group, 27% vs. RVAD than for those without RVF (32 vs. 22 days,
44 S. Gohar et al.

Table 3.5  Clinical trials evaluating hemodynamic parameters for RV failure post-LVAD implantation (Reproduced
with permission [47])
Study N VAD type RVF definition RVFi Risk factors/scores Outcomes
Ochai [27] 245 100% Need for RVAD 9% Pre-op circulatory
pulsatile support (OR 5.3)
1991–2001 VAD Female gender (OR 4.4)
BTT 98% Nonischemic etiology
(OR 3.3)
Drakos [40] 175 86% Need for RVAD 44% (1 point for each) 365-day post-­
pulsatile LVAD survival:
VAD
1993–2008 14% ≥14 days Destination therapy (OR ≤5.0 = 83%
CF-VADs inotropes 3.31)
Single-center iNO ≥ 48 h Inotrope dependency 5.5–8.0 = 77%
(OR 2.47)
Retrospective Obesity (BMI ≥ 30 kg/ 8.5–12.0 = 71%
analysis m2) (OR 1.99)
BTT 58% IABP (or 3.88) ≥12.5 = 61%
PVR RVF % for risk
score categories:
 1.8–2.7 Wu (or 1.95) ≤5.0 = 11%
 2.8–4.2 Wu (or 3.01) 5.5–8.0 = 37%
 ≥4.3 Wu (or 4.14) 8.5–12.0 = 56%
ACE or ARB (OR 0.49) ≥12.5 = 83%
Beta-blocker (OR 1.60)
Fitzpatrick [41] 266 98% Need for RVAD 37% Cardiac index (≤2.2 L/ Total score ≥ 50
pulsatile min/m2) (OR 5.7) score predicts need for
VAD 18 BiVAD
1995–2007 2% RSWI ≤0.25 mmHg/L/
CF-VADs m2 (OR 5.1) score 18
Single-center Severe pre-VAD RV
dysfunction (OR 5.0)
score 16
Retrospective Pre-op creatinine
≥1.9 mg/dL (OR 4.8)
score 17
Previous cardiac surgery
(OR 4.5) score 16
SBP ≤ 96 mmHg (OR
2.9) score 13

(continued)
3  Optimization of Right Ventricular Function Preoperatively for LVAD Implantation 45

Table 3.5 (continued)

Study N VAD type RVF definition RVFi Risk factors/scores Outcomes

Dang [23] 108 100% Need for RVAD, 39% Elevated intraoperative
1996–2004 pulsatile ≥14 days (OR 1.2)
BTT 73% VAD inotropes and/or
pulmonary
vasodilators
Matthews [25] 197 86% Need for RVAD/ 35% Vasopressor use (OR 180 day post-LVAD
pulsatile ECMO 3.9) score 4 survival for RVFRS
VAD
RVFRS 14% ≥ 14 days AST ≥ 80 IU/L (OR Total:
CF-VAD inotropes, inhaled 2.1) score 2
iNO ≥ 48 h, or
hospital discharge
on an inotrope
1996–2006 Bilirubin ≥2 mg/dL ≥5.5 = 66 ± 9%
(OR 2.4) score 2.5
Single-center Creatinine ≥2.3 mg/dL 4.0–5.0 = 80 ± 8%
(OR 2.9) score 3
Retrospective 3.0 = 90 ± 3%
BTT 94%
Atluri [28] 218 59% Need for RVAD 23% (1 point for each) Score 0–1: LVAD
pulsatile only
VAD
2003–2011 41% CVP > 15 (or 2.0) Score 2–3: May
CF-LVAD tolerate isolated
LVAD with temp
medical or RVAD
support
CRITT score Severe RV dysfunction Score 4–5: BiVAD
(OR 3.7)
Pre-op mechanical
ventilation (OR 4.3)
Severe tricuspid
regurgitation (OR 4.1)
Heart rate > 100 beats/
min (OR 2.0)
Kormos [26] 484 100% Need for RVAD, 20% CVP/PCWP >0.63 (or 365-day survival
CF-LVAD ≥14 days 2.3)
Multicenter inotropes, late Early Pre-op ventilator No RVF = 78%
inotropes, later RVF support (OR 5.5)
inotrope support 13%
2005–2008 starting >14 days BUN >39 mg/dL (OR early RVF = 59%
after implant 2.1)
Retrospective
HM II BTT trial
BTT 100%
CRITT score [C]VP [R]V dysfunction [I]ntubation preoperatively [T]ricuspid regurgitation [T]achycardia; RVFRS RV
failure risk score; ACEI/ARB angiotensin-converting enzyme inhibitor/angiotensin receptor blocker, iNO inhaled
nitrous oxide; RSWI
46 S. Gohar et al.

Survival Curve
1.0
0.94 ± 0.01
No RVF
0.87 ± 0.02
0.78 ± 0.03
0.8
0.89 ± 0.04
RVF 0.66 ± 0.06
Survival Probability

0.59 ± 0.07
0.6

0.4

0.2
Remaining at Risk
386 348 206 105
65 52 30 18
0.0
30 180 365
Days

Fig. 3.5  HeartMate II LVAD indicating worse 1-year survival for patients with RVF (Modified from Kormos et al. [26])

P < 0.001). Those who required inotropic support quantitative RV functional assessment on

for more than 14 days after LVAD implantation echocardiography is difficult to reproduce with
and those with late inotropic support had an aver- potential for significant interobserver variability.
age length of stay of 35 and 32 days, respectively. Tricuspid annular plane systolic excursion
Thus, any RVF resulted in a significantly longer (TAPSE) is a validated parameter of global right
hospitalization time before discharge than seen in ventricular function. Puwanant et al. reviewed
those without any RVF (P < 0.001) [26]. This trial preoperative echo parameters in 33 patients and
indicates that patients with a significant preopera- noted that TAPSE <7.5 mm was associated with
tive RV dysfunction as characterized by a high- increased risk of RVF with a 91% specificity and
resting RAP to LAP ratio may be at higher risk for 46% sensitivity. Given the small cohort size with
RVF after LVAD implantation and need closer half devices pulsatile, extrapolation of the results
monitoring. In addition, patients with established remains limited in the CF-VAD era [42]. The RV/
acute organ dysfunction especially if associated LV diameter ratio is a surrogate of disproportion-
with a systemic inflammatory syndrome are very ate RV remodeling, similar to CVP/PCWP ratio.
high-risk LVAD candidates as well. In a study involving 115 patients with CF-VADs,
the RV-to-LV diastolic diameter (R/L) ratio was
measured on transesophageal echocardiogram.
Echocardiographic Models The odds ratio of developing RVF was 11.4 in
patients with an R/L ratio 0.72 (P 0.0001) [24]
Similar to hemodynamic parameters, various (Fig. 3.6). Another trial using transthoracic echo-
echocardiographic models have been used to pre- cardiogram demonstrated a R/L ratio ≥ 0.75 and
dict post-LVAD RHF. Fitzpatrick et al. graded was associated with a fivefold increase in
RV function on preoperative echo as none, mild, RVF. The R/L diameter ratio ≥ 0.75 (AUC = 0.68)
moderate, or severe. He noted that in addition to was as optimal as the Matthews (AUC = 0.69)
RV stroke work index (RSVWI), severe pre-op [25] and Kormos (AUC = 0.63) [26] risk scoring
RV dysfunction was one of the strongest predic- systems in predicting RVF alone and the compos-
tors of BiVAD placement [41]. However, semi- ite of RVF and death [43].
3  Optimization of Right Ventricular Function Preoperatively for LVAD Implantation 47

Prevention of Right Heart Failure

Based on an understanding of RV physiology,

many institutions have developed strategies to
prevent post-LVAD RHF. These strategies are
multifactorial as variables that relate to patient
selection determine the need and type of medical
plan prior to surgery. This can be thought of as
optimizing the patient for LVAD implantation.
What follows is our practice at the Texas Heart
Institute (THI), which is informed by much of the
previous discussion in this chapter. As such, it
should not be considered a guideline or standard
of care but rather a proposed pathway for practi-
tioners to manage these complex patients.

Preload Optimization

In a study by Cordtz et al., the pathophysiologic

steps leading to RVF after LVAD implantation
were shown to initiate early on in the postopera-
Fig. 3.6  Upper panel: mid-esophageal (ME) four-­ tive period as RVF incidence was related to the
chamber view for measurements of right ventricular max- immediate cardiac index post-implant [46]. Based
imal end-diastolic diameter (RVEDD max). Lower panel:
transgastric (TG) two-chamber view for measurements of on the data presented earlier and our clinical expe-
left ventricular end-diastolic diameter (LVEDD). Both rience, we have developed a uniform process to
parameters are used for the R/L ratio calculation. A R/L optimize RV function prior to LVAD implant.
ratio > 0.72 was associated with RVF (Reproduced with Preoperative hemodynamic assessment and
permission from Kukucka M et al. [24])
volume optimization with a PA catheter-guided
strategy are recommended for patients at risk for
Strain imaging is a load-independent tech- RVF. The RV is a volume-dependent ventricle,
nique for measuring RV function. Although data and optimizing central venous pressure (CVP) is
is limited, many small studies have suggested key to mitigate the risk of RVF. Volume optimiza-
decreased RV peak, and free wall longitudinal tion with aggressive diuresis and if needed ultra-
strain may predict RVF [21]. In a small study of filtration are the mainstay of therapy. A
19 patients, speckle-tracking echocardiography CVP > 16 mmHg has been associated with post-
was used to assess RV performance pre- and post- ­op RVF [25, 40, 47, 48]. At our institution, we
LVAD implant and noted that those with the low- use an aggressive diuretic regimen employing
est strain value pre-implant had worse function loop diuretic infusions in combination with thia-
post-implant [44]. Quantitative 3D echocardiog- zide diuretics to block the renal tubules in a
raphy (3DE) is a promising method for pre-LVAD multi-targeted approach.
RV assessment. Echocardiographic indices asso- In addition to increasing urine output, attempts
ciated with RVF included 3DE indexed RV end- should be made to limit blood product transfu-
diastolic and end-systolic volumes (RVEDVI and sions during surgery by correcting coagulopa-
RVESVI) and RV ejection fraction (RVEF). In a thies prior to surgery to reduce SIRS-associated
small study, preoperative RV volumes were asso- PVR elevation. In a single-center study, with pul-
ciated with RVF in continuous-­flow LVAD recipi- satile LVADs, vitamin K given preoperatively
ents, independently from hemodynamic correlates was found to reduce bleeding and the need for
of RV function (RVSWI) [45]. blood product transfusions [49, 50]. For known
48 S. Gohar et al.

coagulopathies, we should proceed in ­conjunction Inotropic Support

with clinical pathology consultation to correct
the bleeding diathesis prior to surgery. Perioperative inotropic support for a tired, dis-
tended RV is important to prevent postoperative
RV failure. Milrinone and dobutamine are the
Afterload Reduction two approved inotropes used for cardiac support.
Milrinone is an intravenous PDE3 inhibitor that
Optimization of RV afterload can further help improves cardiac output by elevating intra-­
improve hemodynamics and assist with weaning cardiomyocyte cAMP levels. It has a long half-­
from cardiopulmonary bypass and chest closure. life and has minimal effect on the heart rate. Its
As noted earlier, the RV is very afterload sensitive, vasodilation of both the pulmonary and systemic
yet some reports suggest that this sensitivity vascular beds can result in hypotension at high
increases further after LVAD placement [51, 52]. doses, limiting its use in many situations. Dose
Attempts to reduce PVR prior to LVAD implant adjustment is required in patients with renal dys-
can lower the incidence of RV with likelihood of function due to risk of toxicity. Dobutamine is a
benefit [48, 49]. Pulmonary vasodilators including β-1 agonist with a short half-life. It increases
inhaled nitrous oxide (iNO) and phosphodiester- both cardiac contractility and heart rate with
ase-­5 inhibitors have been used for RV afterload some hypotension. In stable patients not on
optimization during intraoperative and postopera- ­inotropes, given their negative inotropic effects,
tive period. A prospective, randomized, double- we recommend avoiding beta-blockers in the
blind, placebo-controlled, multicenter trial perioperative period at our institute.
demonstrated that iNO initiated before weaning Vasopressors are used cautiously in specific
from cardiopulmonary bypass (CPB) and contin- scenarios involving hypotension and post-op
ued for 48 h post-LVAD implantation decreased vasoplegia after LVAD implant. Epinephrine
mean pulmonary artery pressure (mPAP) and and dopamine are the two commonly employed
increased LVAD flow. However, it did not reduce agents for vasodilatory shock after LVAD
the incidence of RVF, with most benefit seen in implantation. They also possess some intrinsic
patients with higher mean pulmonary artery pres- inotropic properties and increase blood pressure
sures and low pump flow during weaning from by arteriolar and splanchnic vasoconstriction.
CPB [53]. Sildenafil is a phosphodiesterase-­ 5A We try to limit the overall duration of inotrope
(PDE5) inhibitor with pulmonary artery vasodila- use given the association with increased mortal-
tory properties. In patients with persistent pulmo- ity in clinical trials. Our inotrope weaning strat-
nary hypertension after recent LVAD placement, egy at Texas Heart Institute (THI) involves
sildenafil use resulted in a significant decrease in vasopressor weaning first. The patient is aggres-
PVR when compared with control patients [54, sively diuresed during this time while PDE5is
55]. A recent systematic review noted insufficient are uptitrated. Once clinical euvolumia is
evidence supporting PDE5 inhibitor use to attenu- achieved (increase in BUN/Cr ratio and
ate RV failure in patients requiring an LVAD [56]. MvO2 > 60%), inotrope weaning is initiated and
Lastly, strategies to avoid sudden elevation in pul- completed within a few days.
monary vascular resistance including hypoxia, Sinus rhythm and AV synchrony are important
hypercarbia, or severe acidosis should be imple- in right ventricular function. A lack of sinus rhythm
mented during peri- and early postoperative period. leads to suboptimal RV mechanical function [7].
At THI, we initiate intraoperative iNO in all There is strong evidence that cardiac resynchroni-
patients and oral phosphodiesterase inhibition in zation therapy in the appropriate patient population
most patients to optimize PVR prior to surgery. (LBBB, QRS >150, NYHA class III–IV) improves
These are continued into the postoperative period. LV and RVEF and NYHA symptoms.
3  Optimization of Right Ventricular Function Preoperatively for LVAD Implantation 49

 emporary Mechanical Circulatory

T implant demonstrated clinical stabilization with
Support an IABP, and outcomes post-VAD implant were
improved as compared to those that did not and
RVF may still occur despite all necessary precau- required escalation of therapy [63].
tions after LVAD implantation. Prompt identifi- Femoral access for IABP is useful in emergent
cation and surgical support in such scenarios can scenarios due to ease of access and caliber of ves-
be crucial to improve long-term outcomes. sel but has important limitations. Due to the need
Elective RVAD implantation has been correlated for limb immobilization, restriction on patient
with better long-term survival than an emergent ambulation often leads to significant decondition-
implant [57]. Concomitant RVAD implant at the ing. An alternative approach is axillary or subcla-
time of LVAD implantation has also been shown vian (SCA) IABP insertion. In a single-center
to improve survival to transplantation in small study from the University of Chicago, Tanaka
studies [58]. et al. demonstrated the benefit of a subclavian
Percutaneous ventricular assist devices (SCA) IABP. Ninety percent of the 70 patients
(pVADs) are an important adjunct to medical who received a SCA IABP were bridged to either
therapy at THI. At our institution, we routinely transplant, LVAD or recovery [60]. A statistically
utilize mechanical circulatory support for optimi- significant improvement in CVP, pulmonary
zation of hemodynamics and end-organ function artery pressure, wedge pressure, renal function,
before durable LVAD implantation. Intra-aortic as well as cardiac index was noted with a median
balloon pump (IABP) is associated with improved use of 21 days. Moreover, with the axillary bal-
RV dysfunction in postcardiotomy shock [59]. It loon pump, the patients were able to ambulate
has also shown to improve LV and RV hemody- and maintain muscle tone. We utilize IABP in
namics in patients with RV pressure overload patients with INTERMACS 2 extensively at our
[60, 61]. We utilize the balloon pump as a bridge institution. Imamura et al. have also described a
to decision for both elective durable LVAD similar practice of insertion of IABP in all
implantation and in critically sick ICU patients INTERMACS 2 patients <1 week prior to LVAD
with slowly progressive cardiogenic shock. Our implantation. They demonstrated that compared
experience has demonstrated that IABP improves to matched controls, patients who received an
MPAP, PCWP, and RV geometry similar to previ- IABP had shorter ICU stays, improved markers
ously published work [49]. of perfusion, and lower cost of care [64].
Severe biventricular heart failure remains a In those patients in which IABP is not an
difficult clinical scenario requiring complex option or who are in acute profound shock, we
management in the VAD era. In a small, single-­ have utilized the Impella pVAD (Abiomed)
center study, Ntalianis et al. demonstrated the device series as a bridge to decision or bridge to
prolong use of a sheathless femoral IABP (mean LVAD. The Impella also may afford a period of
duration of 73 days), in 15 high-risk patients time to precondition the RV to increase flows
(INTERMACS 1 or 2). The study demonstrated while also decreasing PA pressures. The Impella
an improvement in RAP, PAP, and cardiac index RP (Abiomed) is a small micro-axial percutane-
as well as improvement in echocardiographic ously inserted pump that has been designed for
indices of RVF (RVSWI, TAPSE). None of the short-term RV support and can assist with bridge
six patients (40%) who were bridged to LVAD to decision for long-term durable support without
implantation developed post-implant RVF, while the need for significant surgery for implantation
three patients who were weaned from the IABP [65] (Fig. 3.7). Impella CP is a left-sided pVAD
preserved satisfactory RV function 6 months after inserted across the aortic valve in the LV to opti-
IABP removal. An interesting finding from the mize LV end-diastolic pressures. Its utility in
analysis noted that patients with early shock seemed RVF is questionable given the limited ability of
to derive the most benefit from the IABP [62]. the device to provide full cardiac output.
At the University of Washington in St. Louis, 54 The Impella 5.0 (Impella) is a larger pVAD,
patients with cardiogenic shock prior to LVAD which can deliver up to 5 L/min of flow and pro-
50 S. Gohar et al.

AAO
1 5
PVRn
LPA • Endothelial injury
•Transfusion related
lung injury

6 SVC
Arrhythmia
SVT/VT LA
AV

RA
MV LV
4
Tricuspid valve PV
• Dilated annulus LVAD
• Incompetent valve
nRegurgitation

RCA
2
Septal position
3 • (contribution)
Ischemic injury ventricular
• Loss of bypass graft interdependence
• Plaque rupture • Abnormal position
• Thrombo-embolic IVC with leftward pull of
events IVS due to high
RV
LVAD speeds. Loss
1 of RV septal support
Increased preload
• Increased venous
return
• Blood transfusion
• Intra-op volume

Fig. 3.7  Pathophysiology of RV failure after LVAD implantation

vide full cardiac support in cardiogenic shock. Its Similar to the axillary IABP [68], the Impella
major limitation remains the need for surgery/ 5.0 pVAD can be inserted via the axillary artery
vascular graft implantation for access. In a study via surgical graft if the vessel is of adequate cali-
involving 90 patients with abnormal renal or bor- ber to facilitate mobilization [69] (Fig. 3.8). At
derline RV function, 40 patients received the THI, we utilize the Impella 5.0 as bridge to deci-
Impella 5.0 pVAD out of which 75% survived to sion or bridge to transplantation where needed.
transplant or LVAD. Of those who survived to We also frequently upgrade support from IABP
LVAD, 87% survived to hospital discharge [66]. to the axillary Impella 5.0 in patients with pro-
Hall et al. in Baylor University Medical Center, gressive cardiogenic shock and multi-organ
Dallas, and the Ochsner Clinic utilize the Impella failure.
5.0 as the device of choice in patients with car- For refractory shock and the need for escalation
diogenic shock and high MELD scores. Their of therapy, the CentriMag (Thoratec, Pleasanton,
data demonstrates improved hemodynamics and CA) [70] and TandemHeart (CardiacAssist Inc.,
end-organ function as demonstrated by improved Pittsburgh, PA) [65] have successfully been used
MELD score [66, 67]. for temporary RV support. Both however require
3  Optimization of Right Ventricular Function Preoperatively for LVAD Implantation 51

Fig. 3.8  Thoracic X-ray

film of a patient shows
an Impella 5.0 pump in
the left ventricle,
crossing the aortic valve,
and the driving cable
exiting from the right
axillary artery
(Reproduced with
permission from Ann
Thorac Surg. 2008
Apr;85(4):1468–70)

surgical implantation. Long-term durable biven- RVF. CVP/PCWP ratio > 0.63, preoperative ven-

tricular support options include Thoratec PVAD tilator support, and BUN >39 mg/dL are proba-
and Syncardia Total Artificial Heart as bailout bly the strongest predictors of RVF along with
options on a BTT strategy. The use of TAH as a TAPSE <7.5 mm on echocardiography.
bridge to transplant has demonstrated a 79% sur- Perioperative management of RV failure should
vival to transplantation vs. 46% in patients not include preload optimization, afterload reduc-
receiving a TAH in a small observational prospec- tion, and inotropic support in all patients. RVF
tive study [71]. It is not currently approved for des- can still occur despite all necessary precautions,
tination therapy. Detailed discussion on durable and prompt identification with early RVAD
RVADs will be addressed in a later chapter. implantation may help improve long-term out-
comes (Fig. 3.9).
It is still to be seen if in the future, minimally
Conclusion invasive surgery and off-pump pVAD implanta-
tion will mitigate the problem of right heart fail-
Optimization of right ventricular function prior ure. The field of mechanical support for heart
to LVAD implant is important and should be indi- failure continues to grow however through clini-
vidualized for each patient. Early identification cal research and experience to improve outcomes
of high-risk patients remains critical for imple- and drive innovation. Interinstitutional collabora-
menting strategies to avoid RV failure. tion experience in this field is expected to drive
Unfortunately, most RVF risk scores have been this growth further. This, combined with fast-­
derived from small single-center retrospective moving innovation in the field, will likely
trials and provide a variable spectrum of predic- improve RVF-related outcomes significantly in
tors with no single model dependably forecasting the future.
52 S. Gohar et al.

Patient Selected for LVAD

Assess RV

ECHO RHC

RVEF RA mmHg
TAPSI RV mmHg
Volume PA mmHg
E/A PCWP mmHG
Strain
Diameter PAPi
Mass RA/PCWP
Septal motion RVSWI

Optimize RV

Preload Contractility Afterload

Diuresis Sinus Rhythm Volume Status

CVVHD Hold BB PDE-5 inhibitors
Correct Coagulopathy Inotropy Inhaled Flolan
Limit bypass pump runs

Percutaneous Mechanical Circulatory Support

Fig. 3.9  Proposed algorithm for the prevention and management for RV failure after LVAD implantation

Baldwin JT, Young JB. Seventh INTERMACS annual

References report: 15,000 patients and counting. J Heart Lung
Transplant. 2015;34(12):1495–504.
1. Heidenreich PA, Albert NM, Allen LA, et al. 5. Takeda K, et al. Outcome of unplanned right
Forecasting the impact of heart failure in the United ventricular assist device support for severe right
States: a policy statement from the American Heart heart failure after implantable left ventricular
Association. Circ Heart Fail. 2013;6(3):606–19. assist device insertion. J Heart Lung Transplant.
2. Rose E, Gelijns A, Moskowitz A, et al. Long-term 2014;33(2):141–8.
mechanical left ventricular assistance for end-stage 6. Kirklin JK, et al. Second INTERMACS annual report:
heart failure. N Engl J Med. 2001;345:1435–43. more than 1,000 primary left ventricular assist device
3. Miller LW, Pagani FD, Russell SD, et al. Use of a implants. J Heart Lung Transplant. 2010;29(1):1–10.
continuous-flow device in patients awaiting heart 7. Haddad F, et al. Right ventricular function in cardio-
transplantation. N Engl J Med. 2007;357:885–96. vascular disease, part I: anatomy, physiology, aging,
4. Kirklin JK, Naftel DC, Pagani FD, Kormos RL, and functional assessment of the right ventricle.
Stevenson LW, Blume ED, Myers SL, Miller MA, Circulation. 2008;117(11):1436–48.
3  Optimization of Right Ventricular Function Preoperatively for LVAD Implantation 53

8. Santamore WP, Dell’Italia LJ. Ventricular interde- tation in patients with chronic congestive heart fail-
pendence: significant left ventricular contributions to ure. J Heart Lung Transplant. 2006;25:1–6.
right ventricular systolic function. Prog Cardiovasc 24. Kukucka M, Stepanenko A, Potapov E, Krabatsch
Dis. 1998;40(4):289–308. T, Redlin M, Mladenow A, et al. Right-to-left ven-
9. Santamore WP, Gray L Jr. Significant left ventricu- tricular end-diastolic diameter ratio and prediction
lar contributions to right ventricular systolic func- of right ventricular failure with continuous-flow left
tion. Mechanism and clinical implications. Chest. ventricular assist devices. J Heart Lung Transplant.
1995;107(4):1134–45. 2011;30:64–9.
10. Hoffman D, Sisto D, Frater RW, Nikolic SD. Left-to- right 25. Matthews JC, Koelling TM, Pagani FD, et al. The
ventricular interaction with a noncontracting right ven- right ventricular failure risk score a pre-operative
tricle. J Thorac Cardiovasc Surg. 1994;107:1496–502. tool for assessing the risk of right ventricular failure
11. Leo G. Kevin, Matthew Barnard; right ventricu- in left ventricular assist device candidates. J Am Coll
lar failure. Contin Educ Anaesth Crit Care Pain. Cardiol. 2008;51:2163–72.
2007;7(3):89–94. 26. Kormos RL, Teuteberg JJ, Pagani FD, et al. Right
12. Dell’Italia LJ, Walsh RA. Application of a time vary- ventricular failure in patients with the HeartMate II
ing elastance model to right ventricular performance continuous-flow left ventricular assist device: inci-
in man. Cardiovasc Res. 1988;22:864–74. dence, risk factors, and effect on outcomes. J Thorac
13. Feneley MP, Gavaghan TP, Baron DW, Branson JA, Roy Cardiovasc Surg. 2010;139:1316–24.
PR, Morgan JJ. Contribution of left ventricular contrac- 27. Ochiai Y, McCarthy PM, Smedira NG, et al.

tion to the generation of right ventricular systolic pres- Predictors of severe right ventricular failure after
sure in the human heart. Circulation. 1985;71:473–80. implantable left ventricular assist system insertion:
14. Klima UP, Lee MY, Guerrero JL, et al. Determinants analysis of 245 patients. Circulation. 2002;106(12
of maximal right ventricular function: role of septal Suppl 1):I198–202.
shifts. J Thorac Cardiovasc Surg. 2002;123:72–80. 28. Atluri P, Goldstone AB, Fairman AS, et al. Predicting
15. Bernheim D. De l’asystolie veineuse dans rhyper tro-­ right ventricular failure in the modern, continuous
phie due coer gauche par stenose concomitante du flow left ventricular assist device era. Ann Thorac
ventricule droit. Rev Med. 1910;39:785. Surg. 2013;96:857–64.
16. Henderson Y, Prince AL. The relative systolic dis- 29. Interagency Registry for Mechanically Assisted

charges of the right and left ventricles and their bear- Circulatory Support (INTERMACS). Appendix A:
ing on pulmonary congestion and depletion. Heart. adverse event definitions: adult and pediatric patients;
1914;5:217–22. 2013.
17. Dexter L. Atrial septal defect. Br Heart J. 1956;18:209. 30. Dandel M, Krabatsch T, Falk V. Left ventricular vs.
18. Lang RM, et al. Recommendations for cardiac
biventricular mechanical support: decision making
chamber quantification by echocardiography in and strategies for avoidance of right heart failure
adults: an update from the American Society of after left ventricular assist device implantation. Int
Echocardiography and the European Association of J Cardiol. 2015;198:241–50.
Cardiovascular Imaging. J Am Soc Echocardiogr. 31. Ondrus T, et al. Right ventricular myocardial infarc-
2015;28(1):1–39.e14. tion: from pathophysiology to prognosis. Exp Clin
19. Lang RM, et al. Recommendations for chamber
Cardiol. 2013;18(1):27–30.
quantification: a report from the American Society 32. Farrar DJ, Compton PG, Hershon JJ, et al. Right heart
of Echocardiography’s Guidelines and Standards interaction with the mechanically assisted left heart.
Committee and the Chamber Quantification Writing World J Surg. 1985;9:89–102.
Group, developed in conjunction with the European 33. Moon MR, Bolger AF, DeAnda A, et al. Septal func-
Association of Echocardiography, a branch of tion during left ventricular unloading. Circulation.
the European Society of Cardiology. J Am Soc 1997;95:1320–7.
Echocardiogr. 2005;18(12):1440–63. 34. Morgan JA, Paone G, Nemeh HW, et al. Impact of
20. Cameli M, et al. Evaluation of right ventricular func- continuous-flow left ventricular assist device support
tion in the management of patients referred for left on right ventricular function. J Heart Lung Transplant.
ventricular assist device therapy. Transplant Proc. 2013;32:398–403.
2015;47(7):2166–8. 35. Piacentino V 3rd, Williams ML, Depp T, Garcia-­

21. Grant AD, et al. Independent and incremental role Huerta K, Blue L, Lodge AJ, Mackensen GB,
of quantitative right ventricular evaluation for the Swaminathan M, Rogers JG, Milano CA. Impact
prediction of right ventricular failure after left ven- of tricuspid valve regurgitation in patients treated
tricular assist device implantation. J Am Coll Cardiol. with implantable left ventricular assist devices. Ann
2012;60(6):521–8. Thorac. 2011;91:1342–7.
22. Morine KJ, et al. Pulmonary artery pulsatility index 36. Brisco M, Sundareswaran K, Milano, et al. The inci-
is associated with right ventricular failure after dence, risk, and consequences of atrial arrhythmias in
left ventricular assist device surgery. J Card Fail. patients with continuous-flow left ventricular assist
2016;22(2):110–6. devices. J Card Surg. 2014;29:572–80.
23. Dang NC, Topkara VK, Mercando M, et al. Right 37. Cantillon DJ, Saliba WI, Wazni OM, Kanj M,

heart failure after left ventricular assist device implan- Starling RC, Tang WH, Wilkoff BL. Low cardiac
54 S. Gohar et al.

output associated with ventricular tachyarrhythmias hemodynamic analysis. J Heart Lung Transplant.
in continuous-flow LVAD recipients with a concomi- 2016;35(7):868–76.
tant ICD (LoCo VT study). J Heart Lung Transplant. 52. Marzec LN, Ambardekar AV. Preoperative evaluation
2014;33(3):318–20. and perioperative management of right ventricular
38. Kalogeropoulos AP, et al. Validation of clinical scores failure after left ventricular assist device implantation.
for right ventricular failure prediction after implanta- Semin Cardiothorac Vasc Anesth. 2013;17(4):249–61.
tion of continuous-flow left ventricular assist devices. 53. Potapov E, Meyer D, Swaminathan M, et al. Inhaled
J Heart Lung Transplant. 2015;34(12):1595–603. nitric oxide after left ventricular assist device implan-
39. Fukamachi K, McCarthy PM, Smedira NG, Vargo tation: a prospective, randomized, double-blind,
RL, Starling RC, Young JB. Preoperative risk fac- multicenter, placebo-controlled trial. J Heart Lung
tors for right ventricular failure after implantable left Transplant. 2011;30(8):870–8.
ventricular assist device insertion. Ann Thorac Surg. 54. Tedford RJ, Hemnes AR, Russell SD, et al. PDE5A
1999;68(6):2181–4. inhibitor treatment of persistent pulmonary hyperten-
40. Drakos SG, Janicki L, Horne BD, et al. Risk fac- sion after mechanical circulatory support. Circ Heart
tors predictive of right ventricular failure after left Fail. 2008;1(4):213–9.
ventricular assist device implantation. Am J Cardiol. 55. Klodell CT Jr, et al. Effect of sildenafil on pulmo-
2010;105:1030–5. nary artery pressure, systemic pressure, and nitric
41. Fitzpatrick JF 3rd, Frederick JR, Hsu VM, et al. Risk oxide utilization in patients with left ventricular assist
score derived from preoperative data analysis predicts devices. Ann Thorac Surg. 2007;83(1):68–71; discus-
the need for biventricular mechanical circulatory sup- sion 71.
port. J Heart Lung Transplant. 2008;27:1286–92. 56. Baker WL, Radojevic J, Gluck JA. Systematic review
42. Puwanant S, Hamilton KK, Klodell CT, et al. Tricuspid of phosphodiesterase-5 inhibitor use in right ven-
annular motion as a predictor of severe right ventricu- tricular failure following left ventricular assist device
lar failure after left ventricular assist device implanta- implantation. Artif Organs. 2016;40(2):123–8.
tion. J Heart Lung Transplant. 2008;27:1102–7. 57. Fitzpatrick JR, Frederick JR, Hiesinger W, et al.
43. Vivo RP, Cordero-Reyes AM, Qamar U, et al.
Early planned institution of biventricular mechan-
Increased right-to-left ventricle diameter ratio is a ical circulatory support results in improved
strong predictor of right ventricular failure after left outcomes ­ compared with delayed conversion
ventricular assist device. J Heart Lung Transplant. of a left ventricular assist device to a biven-
2013;32:792–9. tricular assist device. J Thorac Cardiovasc Surg.
44. Cameli M, Lisi M, Righini FM, et al. Speckle track- 2009;137:971–7.
ing echocardiography as a new technique to evaluate 58. Morgan JA, John R, Lee BJ, et al. Is severe right
right ventricular function in patients with left ventric- ventricular failure in left ventricular assist device
ular assist device therapy. J Heart Lung Transplant. recipients a risk factor for unsuccessful bridging to
2013;32:424–30. transplant and post-transplant mortality. Ann Thorac
45. Kiernan MS, French AL, DeNofrio D, Parmar
Surg. 2004;77:859–63.
YJ, Pham DT, Kapur NK, Pandian NG, Patel 59. Boeken U, et al. Intraaortic balloon pumping

AR. Preoperative three-dimensional echocardiog- in patients with right ventricular insufficiency
raphy to assess risk of right ventricular failure after after cardiac surgery: parameters to predict fail-
left ventricular assist device surgery. J Card Fail. ure of IABP support. Thorac Cardiovasc Surg.
2015;21(3):189–97. 2009;57(6):324–8.
46. Cordtz J, et al. Right ventricular failure after implanta- 60. Tanaka A, et al. The subclavian intraaortic balloon
tion of a continuous-flow left ventricular assist device: pump: a compelling bridge device for advanced heart
early haemodynamic predictors. Eur J Cardiothorac failure. Ann Thorac Surg. 2015;100(6):2151–2157;
Surg. 2014;45(5):847–53. discussion 2157–8.
47. Lampert BC, Teuteberg JJ. Right ventricular failure 61. Darrah WC, et al. Intraaortic balloon counterpulsation
after left ventricular assist devices. J Heart Lung improves right ventricular failure resulting from pres-
Transplant. 2015;34(9):1123–30. sure overload. Ann Thorac Surg. 1997;64(6):1718–
48. Van Meter CH Jr. Right heart failure: best treated by 23; discussion 1723–4.
avoidance. Ann Thorac Surg. 2001;71(3):S220–2. 62. Ntalianis A, et al. Prolonged intra-aortic balloon

49. Argiriou M, et al. Right heart failure post left ven- pump support in biventricular heart failure induces
tricular assist device implantation. J Thorac Dis. right ventricular reverse remodeling. Int J Cardiol.
2014;6(Suppl 1):S52–9. 2015;192:3–8.
50. Kaplon RJ, et al. Vitamin K reduces bleeding in left 63. Sintek MA, et al. Intra-aortic balloon Counterpulsation
ventricular assist device recipients. J Heart Lung in patients with chronic heart failure and cardiogenic
Transplant. 1999;18(4):346–50. shock: clinical response and predictors of stabiliza-
51. Houston BA, et al. Right ventricular afterload
tion. J Card Fail. 2015;21(11):868–76.
sensitivity dramatically increases after left ven- 64. Imamura T, et al. Prophylactic intra-aortic balloon
tricular assist device implantation: a multi-center pump before ventricular assist device implantation
3  Optimization of Right Ventricular Function Preoperatively for LVAD Implantation 55

reduces perioperative medical expenses and improves 68. Wang Y, et al. Decision tree for adjuvant right ventricu-
postoperative clinical course in INTERMACS profile lar support in patients receiving a left ventricular assist
2 patients. Circ J. 2015;79(9):1963–9. device. J Heart Lung Transplant. 2012;31(2):140–9.
65. Kapur NK, Paruchuri V, Korabathina R, et al. Effects 69.
Schibilsky D, et al. Impella 5.0 support in
of percutaneous mechanical circulatory support INTERMACS II cardiogenic shock patients using
device for medically refractory right ventricular fail- right and left axillary artery access. Artif Organs.
ure. J Heart Lung Transplant. 2011;30:1360–7. 2015;39(8):660–3.
66. Lima B, et al. Effectiveness and safety of the
70. Bhama JK, Kormos RL, Toyoda Y, et al. Clinical
Impella 5.0 as a bridge to cardiac transplantation or experience using the Levitronix CentriMag system
durable left ventricular assist device. Am J Cardiol. for temporary right ventricular mechanical support.
2016;117(10):1622–8. J Heart Lung Transplant. 2009;28:971–6.
67. Bansal A, et al. Using the minimally invasive Impella 71. Copeland JG, Smith RG, Arabia FA, et al. Cardiac
5.0 via the right subclavian artery cutdown for acute replacement with a total artificial heart as a bridge to
on chronic decompensated heart failure as a bridge to transplantation. N Engl J Med. 2004;351:859.
decision. Ochsner J. 2016;16(3):210–6.
 Bridge-to-Bridge Strategies
with IABP, Impella, 4
and TandemHeart

Samar Sheth, Salman Bandeali, and Joggy George

Earlier implantation (early meaning in patients

Introduction with ambulatory heart failure) is currently being
evaluated and the data from the ROADMAP
As our awareness of long-term consequences of study (Risk Assessment and Comparative
advanced heart failure improves, there has been Effectiveness of Left Ventricular Assist Device
an institutional push to pursue durable solutions and Medical Management in Ambulatory Heart
for these failing hearts. These solutions include Failure Patients) is pivotal in demonstrating the
left ventricular assist device (LVAD) implanta- benefit of early Heart Mate II LVAD implantation
tion as a destination therapy or as bridge to trans- relative to optimal medical therapy. However in
plantation, as well as listing for orthotropic heart comparison to stable outpatient heart failure
transplantation. patients, there are still a number of patients who
Because pre-LVAD patients often have com- are admitted to the hospital with worsening heart
plex hemodynamics and multiple comorbid con- failure and cardiogenic shock. This subset of
ditions, there have been multiple risk scores patients includes long-term HF patients who
designed to risk stratify and predict outcomes of have worsened over a period of time despite opti-
patients after LVAD implantation. Few of the mal medical therapy as well as patients present-
most commonly used scores include the ing with cardiogenic shock secondary to acute
Interagency Registry for Mechanically Assisted myocardial infarction (AMI) or acute cardiomy-
Circulatory Support (INTERMACS) profile and opathies. Such patients with refractory cardio-
the Heart Mate risk score [1, 2]. As the field con- genic shock, colloquially known as “crash and
tinues to grow, heart failure programs across the burn” [3, 4] could benefit from durable LVAD
country are currently working on better clarify- placement once they are stabilized.
ing these risks and working on patient optimiza- Data from the seventh INTERMACS report
tion prior to implant. suggests that the total number of patients receiv-
ing LVADs considered to be profile 1 is 15%, and
this has remained stable since 2003 [5]. Those
with profile 2 have gone from accounting for
S. Sheth, M.D • S. Bandeali, M.D. • J. George, M.D. (*) 43% of the implants to 37%, whereas those with
Texas Heart Institute at Baylor St. Luke’s profile 3 have increased from 22% to 28%. Data
Medical Center, Baylor College of Medicine,
Houston, TX, USA
has shown that those patients with lower
e-mail: samarshethMD@[Link]; INTERMACS profiles have higher risk of death
jkgeorgemd@[Link] post implant and remain in the hospital longer [5,

© Springer International Publishing AG 2018 57

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
58 S. Sheth et al.

6]. The overall number of LVADs placed as res- pressure of >15 mmHg, (c) systolic BP <90 mmHg
cue therapy has remained constant according to for at least 30 min or SBP >90 mmHg on support-
registry data but general trends suggest decreas- ive measures [10]. In addition to hemodynamic
ing use in this setting. The mortality of such parameters, clinical signs of pulmonary edema
patients is significantly higher ranging anywhere and impaired organ function are required with at
between 60 and 84% [7, 8]. The poorer outcomes least one of the following: altered mental status,
may explain why fewer patients with refractory cold clammy skin and extremities, oliguria, or
shock (profile 1) are being considered for LVADs. serum lactate >2.0 mmol L−1 [11]. The decreased
The Heart Mate risk score is a means of cardiac output leads to compensatory increase in
risk-­stratifying patients across all INTERMACS circulatory volume. As the filling pressures in the
profiles. Patients with higher INTERMACs pro- heart increase, the stretch of ventricular wall fur-
file can have a mortality risk similar or even ther worsens the pumping function starting a
higher than patients with a lower INTERMACS cycle of decompensation. With further decreases
profile depending on the Heart Mate risk score in cardiac output, signs of multi-organ system
[1]. Therefore, despite the higher risk of poor failure and lactic acidosis ensue. Dr. Kapur writes
outcomes, a lower INTERMACS profile cannot “at this point the cardiogenic shock has transi-
and should not be considered a contraindication tioned from a potentially reversible problem to a
for durable LVAD implantation. more complex ‘hemo-­ metabolic’ problem that
A significant number of patients we select at may not respond to traditional treatments” [12]. It
Texas Heart Institute® for LVAD implantation are is this multisystem organ dysfunction that is asso-
INTERMACS profile 1 or 2. ciated with poor outcomes after LVAD implant. In
As discussed in Chap. 3, we are strong propo- fact, scoring systems for shock related to systemic
nents of temporary percutaneous mechanical cir- inflammatory responses and biomarkers
culatory support (pMCS) at the Texas Heart® (APACHE, SAPSII, SOFA, IL-6) do a better job
Institute. While there is scarcity of data on risk of predicting outcomes than basic hemodynamic
stratification and postoperative outcomes of parameters for CHF [13–15].
patients who get pMCS as a bridge to a durable The most common cause of cardiogenic shock
LVAD or heart transplant [9], we employ this is AMI (acute myocardial infarction), with an
strategy to stabilize these patients with hopes of incidence of 10% [16, 17]. Data from the seventh
improving end organ function prior to a durable INTERMACS registry suggests that as many as
LVAD implant/heart transplant. 50% of patients referred for LVAD are in cardio-
In this chapter, we plan to discuss our approach genic shock [5]. Management of such patients
to patients in refractory shock and the use of lack large clinical trials and is mostly based on
pMCS as a bridge-to-bridge strategy. institutional experience and practices. Despite
new revascularization techniques and treatment
algorithms, morbidity and mortality among these
Background patients remain significantly high [18]. While
inotropes and vasopressors may help to improve
Cardiogenic shock is defined as inadequate end hemodynamics, they increase myocardial
organ perfusion due to poor cardiac output in a demand, and this has been shown in multiple tri-
setting of adequate circulatory volume. The clini- als to be associated with increased mortality and
cal and hemodynamic parameters of cardiogenic morbidity. In such circumstances, institution of
shock defined in the SHOCK (Should We MCS can help lower myocardial demand and
Emergently Revascularize Occluded Coronaries improve end organ perfusion as well as poten-
for Cardiogenic Shock) trial are very well tially reverse the damage.
accepted and used. Hemodynamic parameters The use of durable LVADs as rescue therapy
include: (a) cardiac index <1.8 L/min/m2 without for refractory cardiogenic shock has decreased;
supportive measures or <2.2 L/min/m2 with sup- however, this has translated to the increased the
portive measures, (b) pulmonary capillary wedge use of pMCS devices as a means of stabilizing
4  Bridge-to-Bridge Strategies with IABP, Impella, and TandemHeart 59

the patients and preventing irreversible multi-­ a wire/pressure lumen. Helium is used due to its
organ failure. rapid transit into and out of the balloon in addi-
Guidelines have recently started to include tion to the ability to absorb rapidly into blood in
percutaneous MCS as options in high-risk PCI case of rupture. The balloon is triggered to inflate
as well as cardiogenic shock. Given the limited during diastole either by the ECG or pressure
direction however presented in the guidelines, in triggers. With the onset of systole, the balloon is
2015 the Society for Cardiovascular Angiography rapidly deflated [12, 19, 21]. Tachycardia and or
and Interventions (SCAI) presented a consensus arrhythmia can effect timing of the pumps func-
document on the use of MCS [19]. While the tion and cause it to be ineffective, by either inef-
data on use in patients pre-LVAD is limited, it is ficient triggering or rapid deflation due to
reasonable to use nondurable MCS in patients tachycardia [22, 23]. Inflation during diastole
falling into the INTERMACS category 1 and 2 provides a displacement of blood in the aorta and
[19]. However, the timing of initiation of hemo- hence increased pressure in the aortic root leading
dynamic support devices in these patients is not to increased coronary perfusion [21]. It should be
well studied. It should be noted that the majority noted however that in coronary arteries with fixed
of data on MCS and the hemodynamic effects stenotic lesions there is no improvement in coro-
evaluated are on that of the left ventricle and the nary flow [21, 24]. In systole, the rapid deflation
patients we are discussing often have biventricu- of the device leads to a pressure sink thereby
lar failure and MOSF which make the overall reducing LV afterload and increasing LV output.
hemodynamic effects difficult to predict. Contraindications to placement include mod-
Interestingly, data has recently shown that early erate or more severe aortic insufficiency as well
use of MCS in patients with cardiogenic shock as severe peripheral vascular disease. The major-
secondary to MI have an improved survival [20]. ity of complications are vascular in nature and
In fact, the earlier the device was implanted, include limb ischemia, stroke, or access site
even before the use of inotropes, the better the issues. Thrombocytopenia has been seen due to
outcomes. platelet deposition on the membrane of the bal-
At Texas Heart® for our patients in cardio- loon or issues related to anticoagulation with
genic shock we are rapid to escalate to nondura- heparin. If femoral IABP is selected, prolonged
ble MCS to improve end organ function and use can lead to complications associated with
possibly post-LVAD outcomes. Moreover, for immobility.
those patients with refractory shock or consid- Small trials have defined the safety of using
ered for salvage LVAD placement, we utilize IABP prior to LVAD placement as a bridge device
MCS to stabilize these patients and provide a [25, 26]. Recent literature has suggested however
bridge-to-bridge decision. that the more dysfunctional the LV, the less ben-
Currently, our percutaneous MCS options efit from the IABP [27]. The lack of overall ben-
include intra-aortic balloon pump (IABP), efit seen in the various IABP trials is likely
Impella, TandemHeart, and ECMO. because the subpopulation that would benefit the
most is yet to be defined. In one particular study,
ten patients referred for LVAD placement had an
Intra-aortic Balloon Pump (IABP) IABP placed. In those patients that were consid-
ered responders there was a 20% increase in car-
The IABP is the oldest and most commonly used diac index, lower SVR and lower right heart
hemodynamic support device. It is a pulsatile pressures [28]. In another small study 27 patients
device that can either be inserted via the axillary with non-ischemic cardiomyopathy had an IABP
artery or the common femoral artery and is used placed for cardiogenic shock. In patients that
to augment pulsatile flow. It has two major com- were considered responders (67%), within the
ponents, the balloon catheter and the pump con- first 24 h their urine output increased. In both
sole. The catheter is a 7.5–8-French device with responders and non-responders, lactate levels
two lumens: a closed lumen with helium gas and improved [29]. Interestingly, patients with higher
60 S. Sheth et al.

bilirubins or C-reactive protein (markers of sys- and PROTECT2 studies. In the ISAR-SHOCK
temic inflammation or end organ dysfunction) trial (Impella LP 2.5 versus IABP in Cardiogenic
were more likely to be non-responders suggest- Shock), the Impella LP 2.5 showed a greater
ing there is a point in the shock cascade in which increase in cardiac index and mean arterial blood
the IABP will no longer be effective. In another pressure with a significant reduction in lactate;
study with similar design, IABP was placed prior however, there was no difference in mortality and
to LVAD. In those patients who showed further major adverse events between the Impella 2.5
decompensation (defined by the need for and IABP arms [32]. The Impella CP is a
increased vasopressors and inotropes) had worse 14-French device with limited data in the litera-
outcomes post LVAD [27, 30]. When the authors ture but has the ability to provide 3–4 L of blood
evaluated the difference in two populations, they flow (compared to 2.5 L by the Impella LP 2.5).
noted that patients with contractile reserve, as The Impella 5.0 is usually placed via the axillary
defined by LV and RV power indices, were those artery after placement of a graft and surgical cut
most likely to respond to IABP. These trials sug- down. It can however also be placed through the
gest that early use of IABP is potentially of ben- femoral artery.
efit in selected patients prior to LVAD While the Impella 2.5 pump is the device that
placement. was initially studied in trials and received an
As discussed in Chap. 3, IABP has been FDA indication, the use of device is decreasing at
showed to improve both RV and LV function and our institution. This is largely because both
can improve outcomes when placed prophylacti- Impella CP and Impella 5.0 provide more cardiac
cally in patients with INTERMACS profile 2 output. In addition, the recent FDA approved
[30]. But as mentioned above, the overall benefit indication for Impella CP in high-risk PCI and
of those in florid cardiogenic shock is debatable. cardiogenic shock has resulted in most operators
Our approach at Texas Heart® Institute is to use using the device as opposed to Impella 2.5.
IABP as a first-line bridge in patients with early Trials evaluating the Impella CP are limited.
cardiogenic shock to allow improvement in The IMPRESS (Impella CP versus IABP in
hemodynamics and reduce the use of vasopres- Acute Myocardial Infarction Complicated by
sors/inotropes. If patient’s clinical status does not Cardiogenic Shock) trial recently published in
improve or worsens, the next step is to escalate October 2016 and compared Impella CP to the
support to stronger pMCS devices. IABP in patients with cardiogenic shock second-
ary to AMI. There were interesting similarities
to studies comparing Impella 2.5 with IABP.
Impella (ABIOMED Inc., Danvers, MA) No difference in mortality was seen at 30 days
and 6 months. However, notably 92% of the
The impella is a non-pulsatile continuous axial patients in the study had cardiac arrest that
flow device based on the Archimedes screw required resuscitation prior to pMCS implanta-
pump that propels blood from the implanted tion, and this may have affected the overall results
chamber [31] .Unlike the IABP there is no trig- [33]. In one small, single center trial the Impella
gering of this device by ECG or pressure. In fact, CP pump was placed in 28 patients with refractory
the flow generated is independent of ventricular cardiogenic shock, mostly due to acute coronary
function. There are currently three versions of the syndrome. The estimated mortality rate based on
impella pump for left ventricular support: Impella SOFA scores was 87.1%; however, 36% of the
LP 2.5 (low power), Impella CP (cardiac power), patients survived, hence suggesting a benefit
and Impella 5.0. The Impella 2.5 is a 12-French secondary to Impella CP placement [34].
device placed via the femoral artery that has been The Impella 5.0 requires a surgical cut down
studied extensively. It is FDA approved for car- given the device design. It is usually implanted
diogenic shock and high risk percutaneous coro- using the axillary approach which can allow for
nary intervention as demonstrated the PROTECT patient ambulation. The Impella 5.0 has been
4  Bridge-to-Bridge Strategies with IABP, Impella, and TandemHeart 61

used as a bridge in small series of patients, the pump implantation in patients with INTERMACS
largest of which is from Dr. Hall’s group in profile 1. In cases of relative urgency and emer-
Dallas, Texas. They examined 40 patients with gency, Impella 2.5 and CP are the devices choice
biventricular failure out of which 66% were given ease of placement and no requirement for a
INTERMACS profile 2 and placed them on surgical cut down or general anesthesia. The
Impella 5.0 support as a bridge therapy. Of these approved duration of left ventricular support with
75% survived to the next therapy (13 received Impella 2.5 and CP is 4 days. If we anticipate a
cardiac transplant, 15 durable LVADs, 2 had sys- prolonged course of hemodynamic support or if
tolic function recovery) [35]. This study demon- there is a requirement for greater LV unloading
strated the efficacy of using Impella 5.0 device as despite the presence of Impella 2.5/CP, the Impella
a bridge to decision in patients with questionable 5.0 pump is chosen as the bridge to bridge.
benefit of LVAD placement, including those in
the INTERMACS 1 profile. A single center
German study also demonstrated the feasibility TandemHeart (Tandem-Life,
and benefit of the Impella 5.0 as a bridge in Pittsburgh, PA)
patients with INTERMACS 2 profile [36]. At a
small trial at the Ochsner Clinic, the Impella 5.0 Our Center has one of the large TandemHeart
was used as a bridge to decision in patients with experiences in the country. The left-sided
elevated MELD scores (marker of outcome for TandemHeart is an extracorporeal centrifugal-
those undergoing VAD support). Prior to implant flow pump that draws blood from the left atrium
the MELD score was 21 and improved to 14 post and returns it in the descending or ascending
implant thus suggesting improved end organ per- aorta. Removing blood from the left atrium
fusion. Moreover, there was 70% success rate in reduces left ventricular preload and direct arterial
bridge to decision and a 63% survival to dis- return pressurizes the aorta improving mean arte-
charge noted, which compared to historical out- rial pressure. Of the various pMCS options, it is
comes of cardiogenic shock suggests a significant the most technically challenging as it requires
improvement [37]. All these trials and studies transseptal puncture and placement of a large
have resulted in impella pumps being the only bore transseptal inflow cannula. There are four
pMCS pumps approved by FDA for cardiogenic components to the s­ ystem; the inflow cannula is a
shock and high-risk PCI. 21-French catheter placed via transseptal tech-
Contraindications to placing an impella include nique in the left atrium. The inflow cannula goes
severe peripheral vascular disease, mechanical to a centrifugal pump, which can have an oxy-
aortic valve, or the presence of left ventricle genator added to the circuit if needed, and then
mural thrombus. Common complications include this empties into the aorta. Depending on the
access site bleeding and limb ischemia. placement of the aortic outflow cannula, there
Hemolysis occurs in the first 24 h in up to 10% of can be two opposite effects on the left ventricular
patients and can respond to device repositioning afterload. If placed in the ascending aorta, after-
[38]. Persistent hemolysis can lead to AKI and is load will increase along with an increase in LV
an indication for removal. While IABP has been stroke work. If placed in the descending aorta,
shown to improve right ventricular function, data there is an increase in retrograde perfusion of the
on RV improvement is limited with impella and great vessel, mesenteric and renal arteries thereby
hence one must be wary of RV function while the decreasing LV stroke work [39]. Similar to
LV is being supported by impella. impella, left-sided tandem works independently
Based on our institutional experience and of heart rate and native cardiac function. The
practices at the Texas Heart institute, our first-­ total flow delivered can be affected by the size of
line pMCS for the majority of INTERMACS pro- the outflow cannula (15–19 French) but the
file 2 or higher is the IABP. However, we TandemHeart pump has the potential to deliver
generally go straight to the Impella 2.5, CP, or 5.0 3–5 L of blood flow per minute. Given the pres-
62 S. Sheth et al.

ence of large bore outflow catheter through the ists available so we are readily able to implant a
femoral artery, an antegrade sheath is placed in TandemHeart in appropriate clinical situations.
the superficial femoral artery to perfuse the ipsi- However, given ease of use we are likely to first
lateral limb and limb ischemia. implant an Impella CP.
A poorly functioning right ventricle is a rela-
tive contraindication to use of left side
TandemHeart pump, as the RV is required to Veno-Arterial Extracorporeal
deliver blood to the left atrium. Severe aortic Membrane Oxygenation (VA-ECMO)
insufficiency and ventricular septal defect may
also limit the utility of the pump. Left atrial In patients with both cardiovascular and respira-
thrombus is a contraindication to placement. In tory failure or who suffer from severe biventricular
addition, the patient must be able to tolerate anti- failure, VA-ECMO is the circulatory support of
coagulation required for pump operation. choice. Similar to the IABP it can be placed, if
Movement of the transseptal catheter can result in need be, at the bedside without fluoroscopic guid-
left atrial wall trauma leading to cardiac tampon- ance. The circuit consists of an 18–22-French
ade. It is sometimes difficult to secure the inflow venous inflow cannula originating in the RA to a
cannula that traverses from the left atrium through non-pulsatile centrifugal pump for blood propul-
the femoral vein. With unintended patient move- sion to a 15–22-French arterial outflow cannula
ment or patient repositioning, the cannula can that ends in the aorta with a membrane oxygenator
retract into the right atrium leading to a large right for gas exchange in series [12, 19, 21]. The device
to left shunt and systemic desaturation. requires a full-time perfusionist for management.
Our center has published extensively on the use In order to ensure adequate retrograde flow, a right
of left side TandemHeart. In one series, 117 patients radial arterial line is required. Of the various per-
with refractory cardiogenic shock despite IABP cutaneous LVADs, VA-ECMO is the only to
placement were studied. As with other pMCS data increase afterload. Often a second device for LV
sets, the majority of the patients were in cardio- venting is required (IABP or impella). Large bore
genic shock due to ischemia. Similar to the access sites for both venous and a­ rterial limbs are
IMPRESS trial, a large percentage suffered cardiac required. The sizes of the cannulas used effect the
arrest, in this case 50%, prior to implant [40]. In amount of flow. As with TandemHeart, an ante-
this study, TandemHeart institution was associated grade sheath on the ipsilateral arterial side is
with rapid improvement in hemodynamic parame- required to ensure lower limb perfusion.
ters including improvement in mixed venous oxy- Access site complications are the major draw-
genation, pulmonary capillary wedge pressure, and back to ECMO placement as bleeding is com-
a decrease in creatinine. The TandemHeart investi- mon. Blood products are often required which
gators directly compared TandemHeart to IABP in can interfere with future transplant options.
patients with cardiogenic shock. Seventy percent of Peripheral vascular disease and the inability to
the patients enrolled were shock secondary to tolerate anticoagulation are relative contraindica-
ACS. Compared to IABP in this trial, wedge pres- tions to ECMO placement. Various scoring sys-
sure, creatinine, and cardiac power all improved. tems have been evaluated to determine the benefit
However, similar to impella trials, 30-day survival of ECMO. Despite the utility and awareness in
was not statistically different than the IABP arm placing VA-ECMO, mortality of patients that
[41]. Another study done at our center showed that receive the device still is as high as 50% [44].
left-sided TandemHeart is an effect bridge to more Data for ischemic cardiomyopathy and ECMO
definitive therapy in those with end-stage cardio- to facilitate reperfusion has shown promising
myopathy [42, 43]. results but data on use in refractory cardiogenic
At Texas Heart®, many of the interventionalist shock is limited [45, 46]. In a single center trial of
are skilled at transseptal punctures as well as 87 patients who underwent extracorporeal life
large bore access. In addition, we have perfusion- support (ECLS) implantation as a last treatment
4  Bridge-to-Bridge Strategies with IABP, Impella, and TandemHeart 63

option for refractory cardiogenic shock, 60% of ECLS arm required temporary right VAD place-
these were in shock secondary to acute coronary ment and blood product transfusion as well as a
syndrome and 30% were due to a primary cardio- prolonged intensive care unit course. However,
myopathy. It should be noted that in 31% of these survival between the two groups was similar [49].
patients, ECLS was implanted during CPR [47]. At Texas Heart®, we use ECMO for our
Patients with a higher lactate prior to implant or at crashing and burning patients and will often
6 h post implant had a higher mortality rate. have to add it to patients already on some other
Moreover, patients with a pH closer to the normal form of MCS such as IABP or impella. We try
range were more likely to survive. The overall not to use ECMO as a bridge to LVAD unless the
survival rate was 47%. Thirty-eight percent of the patients is awake and or has been fully consented
patients died on ECLS and 12% died after explant for LVAD as we are proponents of a philosophy
[47]. In another single center study 15 pts with that no patient should “wake up with a VAD.”
ECLS were bridged to LVAD. Eighty percent of
these patients were INTERMACS 1. Fourteen
patients improved with temporary cardiac support Conclusion
to INTERMACS profile 3. At the end of the study,
87% were still alive and none had right ventricu- Patients in refractory cardiogenic shock,
lar failure [48]. The study has an important impli- INTERMACS 1 and 2 have been shown to have
cation that if unstable patients receive timely poorer outcomes post-LVAD placement. We try
VA-ECMO implant they can be considered for and improve these outcomes by bridging these
durable LVAD if stabilized. In another study with patients with percutaneous nondurable ventricular
similar findings, 58 consecutive patients undergo- assist devices that improve cardiac output, improve
ing LVAD implant were assessed and divided ret- end organ perfusion, and improve LV loading con-
rospectively into two groups: one group that ditions. While the data on this strategy is limited,
required ECLS prior to placement of durable we have had good success with this approach and
LVAD and the other that did not. APACHE III will continue to provide percutaneous mechanical
scores were higher in the ECLS group. During circulatory support for our sickest patients as a
LVAD implant, larger number of patients in the bridge to further therapies.

DEVICE CONTRAINDICATIONS COMPLICATIONS

All devices Severe peripheral vascular disease Bleeding
Irreversible neurologic disease Vascular injury
Sepsis* Infection
Neurologic injury
IABP Moderate to severe aortic insufficiency Thrombocytopenia
Aortic dissection Thrombosis
Abdominal aortic aneurysm Obstruction of arterial flow due to malposition
Contraindication to anticoagulation* Aortic rupture or dissection
Air or plaque embolism
ECMO Mechanically ventilated >7 days Thrombosis of circuit
Contraindication to anticoagulation Upper body hypoxia due to incomplete retrograde oxygenation
LV dilatation
Systemic gas embolism
CentriMag Contraindication to anticoagulation Thromboembolic events
Air embolism
TandemHeart Ventricular septal defect Cannula migration
Moderate to severe aortic insufficiency Tamponade due to perforation
Contraindication to anticoagulation Thromboembolism
Air embolism during cannula insertion
Inter-atrial shunt development
Impella LV thrombus Hemolysis
Moderate to severe aortic stenosis Pump migration
Moderate to severe aortic insufficiency Aortic value injury
Mechanical aortic valve Aortic insufficiency
Recent TIA or stroke Tamponade due to LV perforation
Aortic abnormalities Ventricular arrhythmia
Contraindication to anticoagulation
64 S. Sheth et al.

Pulsatile

IABP

Intracorporeal Axial-flow

Impella CP PHP *

Centrifugal flow

Continuous
flow pumps

Extracorporeal

Tandem heart VA-ECMO

4  Bridge-to-Bridge Strategies with IABP, Impella, and TandemHeart 65

Hemodynamic Profiles in Shock

Left Heart Filling Pressures

14
50
Hemodynamic Profiles in Heart Failure
40

(PCWP)
30
LV Congestion BiV Congestion
Warm/Dry Warm/Wet
Perfusion

20 18
Hypovolemic
10 RV Congestion
Cold/Dry Cold/Wet 0
0 10 20 30
Right Heart Filling Pressures
Congestion (CVP or RA)

Table 2. Clinical applications for non-durable, percutaneously-delivered mechanical circulatory support

Advanced heart failure / cardiogenic shock ( refractory to IABP or ≥1 inotrope) Acute MI
LV Failure RV Failure Biventricular Failure Cardio-Pulm Failure Killip IV MI
Recovery Decision DT-LVAD Recovery Decision Recovery Decision DT-LVAD Recovery Decision LV RV
Impella CP X X X X
Impella 5.0 (axillary) X X X X X X
Impella RP (investigational) X X X
TandemHeart LVAD X X X X X X
X X X
TandemHeart RVAD X X X X X
VA-ECMO X X X + Vent X + Vent X + Vent X + Vent X + Vent

ventricular assist devices. Circulation. 1995;92(9

References Suppl):II169–73.
8. Oz MC, Rose EA, Levin HR. Selection criteria for
1. Adamo L, et al. The HeartMate Risk Score identi- placement of left ventricular assist devices. Am Heart
fies patients with similar mortality risk across all J. 1995;129(1):173–7.
INTERMACS profiles in a large multicenter analysis. 9. Shah P, et al. Clinical outcomes of advanced heart
JACC Heart Fail. 2016;4(12):950–8. failure patients with cardiogenic shock treated with
2. Adamo L, et al. The Heartmate Risk Score predicts temporary circulatory support before durable LVAD
morbidity and mortality in unselected left ven- implant. ASAIO J. 2016;62(1):20–7.
tricular assist device recipients and risk stratifies 10. Hochman JS, et al. Early revascularization in acute
INTERMACS class 1 patients. JACC Heart Fail. myocardial infarction complicated by cardiogenic
2015;3(4):283–90. shock. Should we emergently revascularize occluded
3. Trachtenberg BH, Estep JD. Roads, maps, and des- coronaries for cardiogenic shock. N Engl J Med.
tinations: the journey of left ventricular assist device 1999;341(9):625–34.
implantation in ambulatory patients with advanced 11. Thiele H, et al. Intraaortic balloon counterpulsation
heart failure. Curr Cardiol Rep. 2016;18(12):132. in acute myocardial infarction complicated by cardio-
4. Starling RC, et al. Risk assessment and comparative genic shock: design and rationale of the Intraaortic
effectiveness of left ventricular assist device and med- Balloon Pump in Cardiogenic Shock II (IABP-­
ical management in ambulatory heart failure patients: SHOCK II) trial. Am Heart J. 2012;163(6):938–45.
the ROADMAP study 2-year results. JACC Heart 12. Esposito ML, Kapur NK. Acute mechanical circula-
Fail. 2017;5(7):518–27. tory support for cardiogenic shock: the “door to sup-
5. Kirklin JK, et al. Seventh INTERMACS annual port” time. F1000Res. 2017;6:737.
report: 15,000 patients and counting. J Heart Lung 13. Mandawat A, Rao SV. Percutaneous mechanical cir-
Transplant. 2015;34(12):1495–504. culatory support devices in cardiogenic shock. Circ
6. Boyle AJ, et al. Clinical outcomes for continuous-­ Cardiovasc Interv. 2017;10(5).
flow left ventricular assist device patients stratified 14. Landis ZC, et al. Severity of end-organ damage as a
by pre-operative INTERMACS classification. J Heart predictor of outcomes after implantation of left ven-
Lung Transplant. 2011;30(4):402–7. tricular assist device. ASAIO J. 2015;61(2):127–32.
7. Oz MC, et al. Screening scale predicts patients 15. Werdan K, et al. Mechanical circulatory support in
successfully receiving long-term implantable left cardiogenic shock. Eur Heart J. 2014;35(3):156–67.
66 S. Sheth et al.

16. Menees DS, et al. Door-to-balloon time and mortal- 31. Basra SS, Loyalka P, Kar B. Current status of per-
ity among patients undergoing primary PCI. N Engl cutaneous ventricular assist devices for cardiogenic
J Med. 2013;369(10):901–9. shock. Curr Opin Cardiol. 2011;26(6):548–54.
17. McNamara RL, et al. Predicting in-hospital mortal- 32. Seyfarth M, et al. A randomized clinical trial to

ity in patients with acute myocardial infarction. J Am evaluate the safety and efficacy of a percutaneous
Coll Cardiol. 2016;68(6):626–35. left ventricular assist device versus intra-aortic bal-
18. Wayangankar SA, et al. Temporal trends and out- loon pumping for treatment of cardiogenic shock
comes of patients undergoing percutaneous coronary caused by myocardial infarction. J Am Coll Cardiol.
interventions for cardiogenic shock in the setting of 2008;52(19):1584–8.
acute myocardial infarction: a report from the CathPCI 33. Ouweneel DM, et al. Experience from a random-
Registry. JACC Cardiovasc Interv. 2016;9(4):341–51. ized controlled trial with Impella 2.5 versus IABP in
19. Rihal CS, et al. 2015 SCAI/ACC/HFSA/STS STEMI patients with cardiogenic pre-shock. Lessons
Clinical Expert Consensus Statement on the Use of learned from the IMPRESS in STEMI trial. Int
Percutaneous Mechanical Circulatory Support Devices J Cardiol. 2016;202:894–6.
in Cardiovascular Care: Endorsed by the American 34. Lackermair K, et al. Retrospective analysis of

Heart Assocation, the Cardiological Society of India, circulatory support with the Impella CP(R) device in
and Sociedad Latino Americana de Cardiologia patients with therapy refractory cardiogenic shock.
Intervencion; Affirmation of Value by the Canadian Int J Cardiol. 2016;219:200–3.
Association of Interventional Cardiology-Association 35. Lima B, et al. Effectiveness and safety of the Impella
Canadienne de Cardiologie d’intervention. J Am Coll 5.0 as a bridge to cardiac transplantation or durable
Cardiol. 2015;65(19):e7–26. left ventricular assist device. Am J Cardiol. 2016;
20. Basir MB, et al. Effect of early initiation of mechani- 117(10):1622–8.
cal circulatory support on survival in cardiogenic 36.
Schibilsky D, et al. Impella 5.0 support in
shock. Am J Cardiol. 2017;119(6):845–51. INTERMACS II cardiogenic shock patients using
21. Morine KJ, Kapur NK. Percutaneous mechanical
right and left axillary artery access. Artif Organs. 2015;
circulatory support for cardiogenic shock. Curr Treat 39(8):660–3.
Options Cardiovasc Med. 2016;18(1):6. 37. Bansal A, et al. Using the minimally invasive Impella
22. de Waha S, et al. Intra-aortic balloon counterpulsa- 5.0 via the right subclavian artery cutdown for acute
tion—basic principles and clinical evidence. Vascul on chronic decompensated heart failure as a bridge to
Pharmacol. 2014;60(2):52–6. decision. Ochsner J. 2016;16(3):210–6.
23. Papaioannou TG, Stefanadis C. Basic principles
38. Lauten A, et al. Percutaneous left-ventricular support
of the intraaortic balloon pump and mechanisms with the Impella-2.5-assist device in acute cardio-
affecting its performance. ASAIO J. 2005;51(3): genic shock: results of the Impella-EUROSHOCK-­
296–300. registry. Circ Heart Fail. 2013;6(1):23–30.
24. Kern MJ, et al. Augmentation of coronary blood flow 39. Kapur NK, et al. Hemodynamic effects of left atrial
by intra-aortic balloon pumping in patients after coro- or left ventricular cannulation for acute circulatory
nary angioplasty. Circulation. 1993;87(2):500–11. support in a bovine model of left heart injury. ASAIO
25. Koudoumas D, et al. Long-term intra-aortic bal-
J. 2015;61(3):301–6.
loon pump support as bridge to left ventricular assist 40. Kar B, et al. The percutaneous ventricular assist

device implantation. J Card Surg. 2016;31(7):467–71. device in severe refractory cardiogenic shock. J Am
26. Norkiene I, et al. Intra-aortic balloon counterpulsation Coll Cardiol. 2011;57(6):688–96.
in decompensated cardiomyopathy patients: bridge to 41. Burkhoff D, et al. A randomized multicenter clini-
transplantation or assist device. Interact Cardiovasc cal study to evaluate the safety and efficacy of the
Thorac Surg. 2007;6(1):66–70. TandemHeart percutaneous ventricular assist device
27. Sintek MA, et al. Intra-aortic balloon counterpulsa- versus conventional therapy with intraaortic balloon
tion in patients with chronic heart failure and car- pumping for treatment of cardiogenic shock. Am
diogenic shock: clinical response and predictors of Heart J. 2006;152(3):469 e1–8.
stabilization. J Card Fail. 2015;21(11):868–76. 42. Idelchik GM, et al. Use of the percutaneous left

28. Annamalai SK, et al. Acute hemodynamic effects ventricular assist device in patients with severe refrac-
of intra-aortic balloon counterpulsation pumps in tory cardiogenic shock as a bridge to long-term left
advanced heart failure. J Card Fail. 2017;23(8):606–14. ventricular assist device implantation. J Heart Lung
29. den Uil CA, et al. First-line support by intra-aortic bal- Transplant. 2008;27(1):106–11.
loon pump in non-ischaemic cardiogenic shock in the 43. Bruckner BA, et al. Clinical experience with the

era of modern ventricular assist devices. Cardiology. TandemHeart percutaneous ventricular assist device
2017;138(1):1–8. as a bridge to cardiac transplantation. Tex Heart Inst
30. Imamura T, et al. Prophylactic intra-aortic balloon J. 2008;35(4):447–50.
pump before ventricular assist device implantation 44. Schmidt M, et al. Predicting survival after ECMO for
reduces perioperative medical expenses and improves refractory cardiogenic shock: the survival after veno-
postoperative clinical course in INTERMACS profile arterial-ECMO (SAVE)-score. Eur Heart J. 2015;
2 patients. Circ J. 2015;79(9):1963–9. 36(33):2246–56.
4  Bridge-to-Bridge Strategies with IABP, Impella, and TandemHeart 67

45. Abrams D, Combes A, Brodie D. Extracorporeal

extracorporeal life support in therapy-­ refractory
membrane oxygenation in cardiopulmonary dis- cardiogenic shock. Eur J Cardiothorac Surg.
ease in adults. J Am Coll Cardiol. 2014;63(25 Pt 2016;49:802–9.
A):2769–78. 48. Schibilsky D, et al. Extracorporeal life support prior

46. Guenther S, et al. Percutaneous extracorporeal life to left ventricular assist device implantation leads to
support for patients in therapy refractory cardiogenic improvement of the patients INTERMACS levels and
shock: initial results of an interdisciplinary team. outcome. PLoS One. 2017;12(3):e0174262.
Interact Cardiovasc Thorac Surg. 2014;18(3):283–91. 49. Marasco SF, et al. Extracorporeal life support bridge
47. Guenther SPW, Brunner S, Born F, Fischer M,
to ventricular assist device: the double bridge strategy.
Schramm R, Pichlmaier M, et al. When all else fails: Artif Organs. 2016;40(1):100–6.
 LVAD Surgical Implant Technique:
Extraperitoneal Approach 5
Rachel A. Beaupré, Gabriel Loor,
and Jeffrey A. Morgan

Introduction Sections

As the number of LVAD implantations continues 1 . Traditional Approach to LVAD Insertion

to rise and the technology advances, alternate 2. Minimally Invasive Implantation Techniques
approaches and new techniques have continued 3. Selection of Outflow Graft Anastomosis
to develop. This chapter aims to not only sum- 4. Off-pump LVAD Insertion
marize the steps common to the traditional LVAD
implantation approach but also to introduce,
explore, and summarize the increasing minimally  raditional Approach to LVAD
T
invasive approaches, the variants in outflow graft Insertion
(OG) anastomosis sites and their perceived ben-
efits, and the varying approaches to “off-pump” The process for inserting a LVAD through tradi-
LVAD implantation that circumvent the operative tional approach using an extended sternotomy
risks associated with the use of cardiopulmonary incision and placing the patient on cardiopulmo-
bypass (CPB). nary bypass has 12 common steps which are
introduced and summarized below.

R.A. Beaupré, M.D. Common Steps in LVAD Insertions

University of Cincinnati, Cincinnati, OH, USA
e-mail: rbeaupre@[Link]
1. Skin incision
G. Loor, M.D. 2. Creation of pre-peritoneal pocket
Lung Transplantation, Texas Heart Institute at Baylor
St. Luke’s Medical Center, Baylor College of 3. Device tunneling
Medicine, Houston, TX, USA 4. Mediastinal exposure
J.A. Morgan, M.D. (*) 5. Cannulation of the aorta and venous system
Division of Cardiothoracic Transplantation, Texas 6. Cardiopulmonary bypass
Heart Institute at Baylor St. Luke’s Medical Center, 7. Coring of the left ventricle, placement of
Baylor College of Medicine, Houston, TX, USA core sutures, inserting inflow into apex
Mechanical Circulatory Support and Cardiac 8. Outflow graft anastomosis to the ascending
Transplantation, Texas Heart Institute at Baylor aorta
St. Luke’s Medical Center, Baylor College of
Medicine, Houston, TX, USA 9. De-airing the device
e-mail: [Link]@[Link] 10. Wean from CPB and actuation of LVAD

© Springer International Publishing AG 2018 69

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
70 R.A. Beaupré et al.

1 1. Establishment of hemostasis
12. Closure of sternotomy and pre-peritoneal

pocket

Primary Incision

In the traditional approach, the incision begins in

the usual fashion used for sternotomy; starting at
the sternal notch, the incision traverses down-
ward at the midline; however, in lieu of ending at
the xiphoid process, caudal extension progresses
to create space for a pre-peritoneal pocket to
accommodate the size of the particular device to
be implanted. Sternotomy is then conducted with
Bovie electrocautery with careful attention to
avoid the pleural space and not to enter the peri-
toneal cavity.

Creation of the Pre-peritoneal Pocket

Creation of the pre-peritoneal pocket is generally

conducted using one of two methods: by placing Fig. 5.1  Insertion of LVAD in extraperitoneal and extra-­
pericardial pocket
the LVAD posterior to the posterior rectus sheath
or by placing the LVAD between the posterior
rectus sheath and the muscle (Fig. 5.1). Taking and clips. The pericardium is opened along the
down a portion of the hemidiaphragm is usually right side of the heart to the diaphragm and over
required to accommodate the device’s size. Once the left apex, while the superior pericardium is
the pocket has been developed, the LVAD is brought over the aorta to the pericardial reflec-
placed. tion. Pericardial stays secure the pericardium and
provide exposure of the heart.

Device Tunneling
 annulation of Aorta and Venous
C
The LVAD is then screwed to the tunneler which System
is brought into the incision and through the fascia
just left of the midline of the pocket. The tunneler Cannulation proceeds in the usual fashion using
is exited through a previously created incision in the right atrial appendage and ascending aorta
the right upper quadrant between the umbilicus for implantation. If concomitant valve work or
and anterior iliac spine. The driveline exits with ASD closure is to be performed, bicaval venous
the “felt” portion remaining inside, and the LVAD cannulation of the IVC and SVC is performed.
is situated within the pocket. When heparinization has progressed to an ACT
of 400 s or higher, cannulation of the ascending
aorta at the pericardial reflection is performed,
Mediastinal Exposure and de-­airing and securing of the cannula fol-
lowed by line testing is performed. Venous can-
Insertion proceeds with division of the retroster- nulation is performed and connected to the
nal fat and peri-thymic tissue utilizing the Bovie circuit.
5  LVAD Surgical Implant Technique: Extraperitoneal Approach 71

Cardiac Bypass

The patient is placed on CPB in the usual fashion

and kept warm. Volume is removed from the
heart and the field is flooded with CO2.

 oring of Left Ventricle, Placement

C
of Core Sutures, Inserting Inflow
into Apex

In order to begin the coring process, the left ven-

tricle apex is exposed and brought closer to the
midline of the sternotomy. An incision in the
apex is made left of the left anterior descending
artery where the heart dimples. Coring proceeds
with the coring knife directed toward the LV cav- Fig. 5.2  Placement of LV apical LVAD sutures
ity (not the septum). Any large trabeculations and
thrombus are removed from the ventricular cav-
ity; next, 2–0 pledgets are placed in horizontal
mattress circumferentially around the ventricu-
lotomy (Fig. 5.2). The inflow cuff is then seated
and tied into place with circumferential sutures
(Fig.  5.3). BioGlue (CryoLife Inc., Kennesaw,
GA) is applied to pledgets and along the inflow
cuff. The cannula is inserted in the inflow hous-
ing and secured. The heart is returned to its nor-
mal position and the LVAD unit is rested into the
pre-peritoneal pocket.

 utflow Graft Anastomosis

O
to Ascending Aorta

The correct length for the outflow graft is mea-

sured and cut with a bevel edge. The aorta is par-
tially occluded with side-biting clamp aortotomy Fig. 5.3  Securing of LVAD inflow cannula
proceeds. The OG is then anastomosed to the
ascending aorta with mattress sutures which are replaces volume. Ventilation of the patient pro-
tied, and BioGlue (CryoLife Inc., Kennesaw, GA) ceeds, and the head of the table is moved upward
is applied to the single-layer running anastomosis as the table’s left side is tilted downward. The
(Fig. 5.4a–c). The graft is de-aired and clamped, outflow graft is connected to the outflow housing,
and the anastomosis is checked for bleeding. and a “de-airing” hole is placed in the outflow
graft. The cross-clamp is kept distal to the de-­
airing hole on the graft (Fig. 5.5). Transesophageal
De-airing the Device Circuit echocardiography is used to monitor the de-­
airing process until its completion.
De-airing begins when the outflow housing is Inotropes are started for optimization of right
opened, and the heart is filled as the perfusionist heart function, and pressors (levophed and
72 R.A. Beaupré et al.

Fig. 5.4 (a) Side-biting clamp placed on ascending aorta for outflow graft anastomosis. (b). Evaluation of outflow graft
anastomosis. (c) Removal of side-biting clamp from ascending aorta

v­ asopressin) are titrated as necessary to maintain i­nterventricular septum, confirming that no bow-
hemodynamic stability and mean arterial pres- ing is present. These findings are used to adjust
sures of 60–80 mmHg. the RPM settings of the LVAD and optimize vol-
ume status and/or increase the dose of inotropes.

 eaning from CPB and LVAD

W
Actuation Establishing of Hemostasis

After placement of the LVAD, weaning off CPB It is essential that all surgical, cannulation, and
is performed. The HeartMate II device speed anastomoses sites are evaluated, and bleeding is
begins at 6000 RPM once the CPB flow decreases addressed with pledgeted sutures. Nonsurgical
to 2 L. The cross-clamp is released from the out- bleeding is surveyed until full reversal of prot-
flow graft to allow forward flow. As CPB weans, amine and baseline ACT are reached, as this
RPM is increased to reach between 8800 and occurs light packing and the use of hemostatic
9600 RPM. Further, de-airing occurs through the agents may aid the process. Electrocautery may
hole in the outflow graft. be used on soft tissue, LVAD pocket, and the ster-
TEE assesses decompression of the LV and num. Examination and proper control of bleeding
degree of mitral regurgitation, evaluates flow in the LVAD pocket should be thorough and is
across the inflow and outflow cannula, monitors necessary to proceed.
for the incidence of aortic insufficiency, assesses If diffuse coagulopathy and excess bleeding
right ventricular function, and evaluates the cannot be controlled with these methods, the
5  LVAD Surgical Implant Technique: Extraperitoneal Approach 73

Modifications of Technique for HVAD

and HeartMate 3

For the HeartWare HVAD device and HeartMate

3, we do not create a preperitoneal pocket. The
device is place intrapericardially. Thus, the ster-
notomy incision does need to extend beyond the
xyphoid, like for HM2. Additionally, the left hemi-
diaphragm does need to be taken down to accomo-
date for the device. Unlike the HM2, where a “cut
and sew” technique is used, for the HVAD and
HM3, a “sew and cut” technique is employed. This
involves identfying the LV true apex/dimple,
marking it, placing the felt ring over it, with the
center of the ring over the true apex, marking the
outer portion of the ring, and then placing plegeted
sutures circumferentially. After all sutures are
placed, they are passed through the felt ring, the
ring is seated, sutures are tied, and cut. A cruciate
incision is then made in the LV apex, followed by
spreading with a Tonsil clamp, followed by coring
of the LV, followed by resection of LV trabeculae.
The device is then inserted with the outflow graft
Fig. 5.5  LVAD in pocket, outflow connected to aorta,
clamped. The device is then secured in place while
protective graft secured over outflow graft
ensuring there is good apposition between the
pump and felt ring, without a space. The device is
then deaired. As side biting clamp is then placed
mediastinum can be packed, and patch is placed on the ascending aorta and the proximal anasto-
over the sternotomy site with the plan to return to mosis is performed. For additional deairing, a hole
the operating room for true closure in 24 h after is made in the outflow graft both proximal and dis-
adequate resuscitation and resolution of the coag- tal to the clamp on the aotflow graft.
ulopathy. A GoreTex pericardial membrane
(Gore Medical Products, Flagstaff, AZ) is sutured
to pericardial edges to minimize injury at reentry Other Techniques
during reoperation (Morgan).
As LVADs have become more widely used, tech-
niques have been introduced and developed to
 losure of Sternotomy and
C decrease invasiveness of the procedure and elimi-
Pre-­peritoneal Pocket nate the need for sternotomy and/or cardiopul-
monary bypass (CPB). The following sections of
Chest tubes are placed in the mediastinum and this chapter will describe and compare the myr-
pleural space. Closure of the sternotomy pro- iad of minimally invasive approaches to LVAD
ceeds in the usual fashion. Abdominal fascia is implantation and selection of outflow graft anas-
closed with interrupted figures of eight. The tomosis sites in the minimally invasive approach
remaining layers and the skin are closed in the and describe techniques used in “off-pump”
usual fashion. LVAD implantation to avoid the use of CPB.
74 R.A. Beaupré et al.

Minimally Invasive Approaches common cannulation strategy employed in mini-

mally invasive procedures is femoro-femoro
The advent and successful use of minimally inva- (Frazier [3], Ghosizad, [4]); however, central can-
sive methods of VAD implantation have the poten- nulation is still utilized [2] dependent on inci-
tial to improve the postoperative recovery in sional approach and can even be done using the
comparison to traditional sternotomy. Heart fail- Seldinger technique (Anyanwu). Thus, there are a
ure patients requiring VAD often have a host of variety of both incisional approaches and cannu-
other comorbidities such as poor pulmonary func- lation site choices that can be utilized for mini-
tion, diabetes, malnutrition, obesity, and decon- mally invasive VAD implantation.
ditioning that add to the risks of major surgery and Approach to minimally invasive implantation
can delay wound healing. The use of minimally can be chosen strategically for specific operative
invasive implantation therefore has the potential to goals. Popov et al. [4] have standardized their
decrease recovery time and avoid the stress of approach using a single left thoracotomy
sternotomy. approach and describe its use as able to protect
In addition, minimally invasive techniques are the outflow graft at the time of reentry and, by
ideal to avoid reentering the chest cavity after one maintaining a short path of OG to the aorta on the
or more prior sternotomies or to further preserve side of the heart, to avoid creating difficulties dis-
the sternotomy site in patients such as those in secting the OG from the right AV groove as may
which the VAD is used in a bridge to transplant be created by sternotomy approach. Cohn and
strategy. Owing to the large number of ischemic Frazier [5] use a transdiaphragmatic approach for
heart failure, which compromises the majority inflow cannulation and return of OG graft to the
receiving LVADs in the Unites States, many VAD supraceliac aorta to avoid not only the peritoneal
patients are likely to have one or more previous cavity but the mediastinum and the left hemitho-
sternotomies for coronary artery bypass graft or rax entirely.
valve repair/replacement surgeries. Minimally One benefit of the minimally invasive tech-
invasive techniques may be utilized in patient niques over traditional sternotomy is that this
with prohibitive operative risk for sternotomy approach often allows for less extensive pericar-
(Frazier). Thus, minimally invasive techniques diotomy. Often through the subcostal approach,
can prevent the complications and increased the pericardiotomy extends along the base of the
operative time and stress caused by sternotomy. heart and apex. The pericardial sac over the right
The most common minimally invasive opera- heart can then be left more intact. By avoiding
tive approach consists of implanting the VAD into opening of the pericardium over the right heart, it
the pre-peritoneal cavity through a left subcostal is thought that right heart function can be better
incision with the patient positions in left antero- preserved.
lateral thoracotomy. This approach is often paired Cheung showed that the duration of CPB was
with right mini-thoracotomy to gain access to the on average 30 min less in minimally invasive
aorta. In Makdisi et al. [1] review of the literature, cases as compared to the sternotomy group
the most common surgical incision variations for though there was no difference in ICU stay or
minimally invasive VAD implantation incisions. hospital stay length or inotropic use. Frazier
The two incisions used for inflow are subdia- described shorter ICU stays and less 12-h blood
phragmatic or left thoracotomy, whereas outflow loss for patients who underwent LVAD implanta-
graft insertion site incision approaches are upper tion through the subcostal as opposed to the redo
hemi-sternotomy, right mini-thoracotomy, and sternotomy group. Sorensen et al. [6] showed that
right upper hemi-­ sternotomy combined with a redo sternotomy predicted the use of RBC trans-
right mini-­thoracotomy in a J-shaped incision, or fusion and increased ICU; in contrast patients
axillary. When planning implantation combined that underwent minimally invasive implantation
with tricuspid or aortic valve work, an upper after prior sternotomy saw significant reductions
hemi-­sternotomy should be used [2]. The most in the need of the both.
5  LVAD Surgical Implant Technique: Extraperitoneal Approach 75

Using minimally invasive techniques may specific circumstances in which the native aorta
lead to the possibility of avoiding full general is either hostile to graft anastomosis or is not nec-
anesthesia for implantation as it has been done essarily readily accessible through the chosen
for other non-cardiac surgeries. Bottio et al. [7, 8] incisional approach to implantation. Benefits of
have used paravertebral blocks combined with this technique include the ease of exposure and
mild anesthesia in such cases and report success- the low incidence of atherosclerosis in the ves-
ful extubation in the OR in 75% of cases with sels. Popov et al. [4] further postulate that OG to
adequate pain control. The possibility of per- axillary artery anastomosis has the potential to
forming LVAD implantation in such a fashion reduce postoperative aortic regurgitation as the
would further support faster recovery times with entry of blood flow from the OG is quite distal to
less pulmonary complications in the LVAD the aortic root.
population. However, extra-aortic OG anastomosis does
carry some considerable concerns relating to pos-
sible postoperative complications or undesired
 pproach to Outflow Graft
A sequelae. One concern is the risk of the patient
Anastomosis kinking the OG or the effects that elevate the arm
may have on flow through the system. To prevent
The site of outflow graft anastomosis varies kinking, reinforced grafts with ring may be used
depending on the surgical approach to exposure, [10, 11]. Cohn et al. [12] have found success in
patient anatomy and state of the native aorta, and preventing kinking by using 10–15 cm of polytet-
surgeon preference. Makdisi and Wang [1] rafluoroethylene (PTFE) to cover the OG before
describe the most frequent outflow graft insertion anastomosis to the descending or supraceliac
sites to be to the ascending aorta, the axillary aorta. Patients can be informed to avoid excessive
artery, the descending aorta, and the supraceliac elevation of the arm, and blood pressure should
aorta. Most surgical approaches make use of the not be taken on the left arm to avoid potential dis-
ascending and descending thoracic aorta. turbances to flow [11]. Magdy describes tunnel-
A simulation of the hemodynamics comparing ing the outflow graft from the innominate under
outflow graft anastomosis to the ascending or the IVC and innominate vein to protect it from
descending aorta was undertaken by Kar et al. direct contact with the sternum. If anastomosis to
[9]. At high VAD outputs with output graft anas- the left axillary artery tunneling occurred through
tomosis to the descending aorta, areas of stagna- the left 2nd intercostal space, then graft to graft
tion, which may be potentially thrombogenic in anastomosis can be performed in the mediastinum
the native aorta, particularly in the region of the (Magdy). When using the subclavian, the graft
aortic root, were present which were not present can be tunneled through the 1st intercostal space
when outflow graft anastomosis to the ascending [11]. In addition, the use of the axillary proximal
aorta was simulated due to re-circulatory flow. to the innominate has the potential for cerebrovas-
This model correlated clinically with findings in cular complications; side-biting clamps are
Jarvik (2000) patients with descending aorta out- applied to the vessel to avoid this complication
flow anastomosis that showed stagnation on TEE (Magdy).
while the aortic valve was not opening, whereas Strategies can be employed to reduce swelling
these areas were not present in patients with an and excessive blood flow into the arm. For the
ascending anastomosis even at high VAD speeds purpose of avoiding excessive blood flow to the
[9]. These findings and their clinical correlation arm when anastomosing to the axillary artery,
suggest that the site of outflow graft anastomosis Bortolussi et al. [10] use double radial pressure
may affect both hemodynamics and the risk for monitoring and use restrictive distal anastomosis
thrombosis in VAD patients. to ensure even pressure to both arms, have also
When the outflow graft is anastomosed to ves- described reducing distal axillary caliber if nec-
sel sites other than the aorta, this often occurs in essary with division of the artery and end-to-end
76 R.A. Beaupré et al.

anastomosis to the outflow graft, and have docu- is employed to limit the potential for air embolus.
mented the success of this technique with postop- In order to further minimize this risk and opti-
erative follow-up. Riebandt et al. [11], when mize function of the heart, augmented fluid infu-
performing anastomosis to the subclavian artery, sion rates and a bolus of inotropic drugs may be
avoid excessive blood flow to the arm by banding necessary [16]. Thus, the surgeons, assistants,
the subclavian when there is a pressure difference and anesthesiology team must be focused and
of 20 mmHg or more. united in this monumental step of the procedure.
Given that proper off-pump technique is fol-
lowed, the postoperative benefits of beating heart
Off-Pump Insertion Techniques LVAD implantation can be delineated. Sileshi
et al. [17] compared on-pump LVAD insertions
The use of cardiopulmonary bypass is known to to minimally invasive off-pump implantation
activate systemic inflammatory mediators. When with left hemi-sternotomy and left anterolateral
conducting off-pump implantation, one should thoracotomy patients and found statistically sig-
always be prepared to initiate CPB. The surgical nificant decrease in postoperative days on ino-
team may choose to prepare sites for cannulation tropes and substantially decreased need for
in advance should the necessity arise later in the perioperative and postoperative blood products.
procedure to minimize the time to emergent CPB. This study however compared the minimally
When conducting off-pump LVAD implanta- invasive off-­pump technique to traditional ster-
tion, it is crucial to create the best visualization of notomy implantation on CPB, therefore con-
the LV apex possible to prepare for safe, quick, founding the results. Yet, Gregoric et al. [3]
and accurate coring of the LV. Proper positioning, found that with CPB use being the only variable,
visualization, and uninhibited access to the site of the average blood product requirement in the
LV coring must be ensured. Heparinization to tar- off-pump group was 7 units as compared to
geted activated clotting time interval must be 26 units in the CPB group.
completed before this portion of the surgery. As
this process is occurring without the use of CPB,
it is absolutely essential that clean coring of the
LV myocardium and rapid placement of the
References
pump into the newly fashioned ventriculotomy 1. Makdisi G, Wang IW. Minimally invasive is the future
proceed quickly, minimizing any loss of blood of left ventricular assist device implantation. J Thorac
through the core. In lieu of placing packing Dis. 2015;7(9):E283–8.
underneath the heart to produce stabilization, 2. Haberl T, Riebandt J, Mahr S, Laufer G, Rajek A,
Schima H, Zimpfer D. Viennese approach to minimize
other techniques may be used. The use of a suc- the invasiveness of ventricular assist device implanta-
tion device (like the Guidant XPOSE [Boston tion. Eur J Cardiothorac Surg. 2014;46(6):991–6; dis-
Scientific, Natick, MA]) in off-pump median cussion 996.
sternotomy may be used to retract the heart into 3. Gregoric ID, La Francesca S, Myers T, Cohn W,
Loyalka P, Kar B, Gemmato C, Frazier OH. A less
the surgical field [13, 14]. Cohn [15] have invasive approach to axial flow pump insertion.
described an experimental technique that uses a J Heart Lung Transplant. 2008;27(4):423–6.
vacuum stabilizer for the sewing ring with an 4. Popov AF, Mohite PN, Sabashnikov A, Weymann A,
over-the-wire vacuum-assisted myocardial cor- Patil NP, Sáez DG, Simon AR. Minimally invasive
HeartWare LVAD implantation through single left
ing tool, and an endoventricular occlusion bal- thoracotomy. J Artif Organs. 2015;18(2):170–2.
loon, that ensures exact, clean coring of the 5. Cohn WE, Frazier OH. Off-pump insertion of an
myocardium, while the balloon can tamponade extracorporeal LVAD through a left upper-quadrant
the core site as the pump is navigated into place. incision. Tex Heart Inst J. 2006;33(1):48–50.
6. Sorensen EN, Pierson RN 3rd, Feller ED, Griffith
The whole operative team must be well coor- BP. University of Maryland surgical experience with
dinated at this point in the procedure, as the heart the Jarvik 2000 axial flow ventricular assist device.
is still beating; steep Trendelenburg positioning Ann Thorac Surg. 2012;93(1):133–40.
5  LVAD Surgical Implant Technique: Extraperitoneal Approach 77

7. Bottio T, Bejko J, Guariento A, Tarzia V, Pittarello 12. Cohn WE, Gregoric ID, Frazier OH. Reinforcement
D, Gerosa G. Bilateral mini-thoracotomy off-pump of left ventricular assist device outflow grafts to pre-
Jarvik 2000 implantation in regional asymmetric para- vent kinking. Ann Thorac Surg. 2007;84(1):301–2.
vertebral analgesia. J Cardiovasc Med (Hagerstown). 13. Anyanwu AC, Fischer GW, Plotkina I, Pinney S,

2016;17(2):160–4. Adams DH. Off-pump implant of the Jarvik 2000
8. Bottio T, Bejko J, Gallo M, Bortolussi G, Gerosa ventricular assist device through median sternotomy.
G. Less invasive implantation of HeartWare left ven- Ann Thorac Surg. 2007;84(4):1405–7.
tricular assist device. Multimed Man Cardiothorac 14. Piacentino V 3rd, Jones J, Fisher CA, et al. Off-pump
Surg. 2014. technique for insertion of a HeartMate vented elec-
9. Kar B, Delgado RM 3rd, Frazier OH, Gregoric ID, tric left ventricular assist device. J Thorac Cardiovasc
Harting MT, Wadia Y, Myers TJ, Moser RD, Freund Surg. 2004;127:262–4.
J. The effect of LVAD aortic outflow-graft place- 15. Cohn WE. New tools and techniques to facilitate off-­
ment on hemodynamics and flow: implantation tech- pump left ventricular assist device implantation. Tex
nique and computer flow modeling. Tex Heart Inst Heart Inst J. 2010;37(5):559–61.
J. 2005;32(3):294–8. 16. Bejko J, Guariento A, Bortolussi G, Tarzia V, Gerosa G,
10. Bortolussi G, Lika A, Bejko J, Gallo M, Tarzia V, Bottio T. Minimally invasive surgical Jarvik 2000 off-
Gerosa G, Bottio T. Left ventricular assist device pump implantation. Multimed Man Cardiothorac Surg.
end-to-end connection to the left subclavian 2015;2015. doi: [Link]
artery: an alternative technique. Ann Thorac Surg. 17. Sileshi B, Haglund NA, Davis ME, Tricarico NM,
2015;100(4):e93–5. Stulak JM, Khalpey Z, Danter MR, Deegan R,
11. Riebandt J, Sandner S, Mahr S, Haberl T, Rajek A, Kennedy J, Keebler ME, Maltais S. In-hospital out-
Laufer G, Schima H, Zimpfer D. Minimally invasive comes of a minimally invasive off-pump left thora-
thoratec Heartmate II implantation in the setting of cotomy approach using a centrifugal continuous-flow
severe thoracic aortic calcification. Ann Thorac Surg. left ventricular assist device. J Heart Lung Transplant.
2013;96(3):1094–6. 2015;34(1):107–12.
 LVAD Surgical Implant Technique:
Infradiaphragmatic Approach 6
Jeffrey A. Morgan and O.H. Frazier

within the center of the left ventricular (LV)

Introduction ­cavity, thus being free of potentially obstructive
surfaces and oriented toward the mitral valve ori-
The evolution of mechanical circulatory support fice. Traditionally, the pump inlet was inserted
(MCS) systems from early-generation volume-­ through the LV apex to take advantage of the lon-
displacement pumps to smaller and more durable gest measured span within the ventricle.
continuous-flow (CF) devices has dramatically Surgeons at our institution have developed a
reduced mortality rates and device-related compli- method for implanting the HeartMate II along the
cations. The advantages afforded by CF left ven- diaphragmatic surface of the heart [1–3]. Although
tricular assist devices (LVADs) have led to their this method is a significant departure from tradi-
widespread application in the treatment of severe tional implantation techniques, it has proved valu-
heart failure. However, the perpetuation of implan- able in both eliminating the need for a preperitoneal
tation strategies developed in an era of pulsatile pump pocket and establishing a geometrically
MCS makes some patients vulnerable to compli- advantageous pump alignment.
cations unique to CF physiology and design.
Proper orientation of the inlet cannula is an
essential component of overall device function. Methods
Ideally, the mouth of the cannula should rest
A vertical midline incision is made, incorporat-
ing a 6-cm subxiphoid extension. Via a standard
J.A. Morgan, M.D. (*) median sternotomy, the pericardium is opened
Division of Cardiothoracic Transplantation, Texas both in the midline and along the length of the
Heart Institute at Baylor St. Luke’s Medical Center,
Baylor College of Medicine, Houston, TX, USA diaphragm. The anterior border of the diaphragm
is then incised from the midline to the apex of the
Center for Cardiac Support, Texas Heart Institute at
Baylor St. Luke’s Medical Center, Baylor College of heart, providing access to the peritoneal cavity.
Medicine, Houston, TX, USA After systemic heparinization and initiation of
e-mail: [Link]@[Link] cardiopulmonary bypass (CPB), the LV apex is
O.H. Frazier, M.D. brought out of the chest and controlled with a suc-
Mechanical Circulatory Support and Cardiac tion stabilizer device. The ventricular coring site is
Transplantation, Texas Heart Institute at Baylor St. then identified approximately one third of the dis-
Luke’s Medical Center, Baylor College of Medicine,
Houston, TX, USA tance from the apex to the base of the heart (thus,
e-mail: ofrazier@[Link]; lschwenke@[Link] anterior to the origin of the papillary ­muscles). The

© Springer International Publishing AG 2018 79

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
80 J.A. Morgan and O.H. Frazier

LV

RV

PD

[Link]. PV

IVC

Fig. 6.1  The left ventricular (LV) apex is reflected in a cephalad direction to expose the diaphragmatic surface of the
heart. The LV cavity is outlined by the dotted line, and the enclosed crosshair depicts “Frazier’s point,” the optimal
location for transdiaphragmatic insertion of the inlet cannula (Modified from Gregoric ID et al. [6])

medial edge of the sewing ring is placed 0.5–1.0 cm After a diaphragmatic myotomy has been cre-
lateral to the posterior descending artery to ensure ated to correspond to the selected coring site, the
a parallel orientation with respect to the short axis inlet cannula is then guided through the dia-
of the left ventricle (Fig. 6.1). The ventriculotomy phragm, inserted within the Silastic ring, and
is performed with a circular coring knife, taking secured in place with two ratcheting cable ties.
care to stay parallel to the septum and to follow Proper orientation of the device is achieved by
posteriorly in the direction of the mitral valve. The first pulling the pump housing into the abdomen
ventricular cavity must then be digitally explored until the heart lies flush with the diaphragm and
and inspected for evidence of thrombus or obstruc- then positioning the outflow graft to course above
tive trabeculae. the left lobe of the liver. Our preference is to wrap
Next, the Silastic inflow cuff is secured in the the body of the pump in available omentum to
standard fashion, using 12 pledgeted, full-­thickness protect the bowel from erosive injury.
horizontal mattress sutures placed circumferen- The outflow graft then is measured and bevel
tially around the coring site. Hemostasis around the cut with enough length to allow a gentle curve
inflow cannula is bolstered by using a large-caliber toward the right side of the chest without exces-
monofilament suture to place a full-­ thickness sive redundancy. The outflow anastomosis is then
purse-string stitch through the pledgeted ring. performed in a standard end-to-side fashion with
6  LVAD Surgical Implant Technique: Infradiaphragmatic Approach 81

the aid of a partial occluding clamp along the nology necessitates consideration of specific ana-
ascending aorta. After externalization of the tomic and mechanical challenges associated with
driveline, the system is thoroughly de-aired by the use of these pumps.
using a 19-gauge needle placed at the highest Inlet placement along the diaphragmatic sur-
point of the outflow graft. Cardiopulmonary face of the heart was first reported nearly 10 years
bypass flows are gradually decreased, the heart is ago, when surgeons from our institution described
allowed to fill, and the pump is started at its low- a subcostal approach for implanting the Jarvik
est setting (6000 rpm). The patient is weaned 2000 pump (Jarvik Heart Inc., New York, NY)
from CPB, which is eventually terminated with [5]. Although used sparingly, the procedure
the aid of transesophageal echocardiographic revealed both the feasibility and potential advan-
guidance to allow optimization of LVAD speeds, tage of an alternative cannulation site for LVAD
chamber size, interventricular septal position, implantation. The HeartWare HVAD (HeartWare
and right ventricular function. Inc., Framingham, MA) was developed in the
Once proper function and orientation are veri- Texas Heart Institute laboratories during the mid-­
fied, protamine is administered, the CPB cannu- 1990s and was specifically designed to fit within
las are removed, and drains are placed in the the ­pericardial space. Whereas a familiarity with
mediastinum and pleural spaces. The bare por- apical positioning led surgeons to prefer this
tion of the outflow graft is covered with a 20-mm-­ method in early clinical trials, diaphragmatic
diameter ringed Gore-Tex graft (Gore Medical, implantation of the HVAD and HM3, as previ-
Newark, DE) to prevent kinking and damage dur- ously described elsewhere, is also performed at
ing future sternal reentry. The defect in the dia- our institution [6].
phragm is partially reapproximated, and the
sternum and soft tissues are closed in the stan- Disclosures  None of the authors has any com-
dard fashion. mercial conflicts of interest.

Discussion References

Early in the development of MCS, inflow cannula- 1. El-Sayed Ahmed MM, Aftab M, Singh SK,
Mallidi HR, Frazier OH. Left ventricular assist device
tion was performed through the LV apex to accom- outflow graft: alternative sites. Ann Cardiothorac
modate lengthy inlet-cannula designs. Although Surg. 2014;3:541–5.
inlet conduits were eventually shortened in 2. Riebandt J, Sandner S, Mahr S, et al. Minimally inva-
response to excessive inflow-graft occlusion seen sive thoratec Heartmate II implantation in the setting
of severe thoracic aortic calcification. Ann Thorac
in experimental testing, device implantation Surg. 2013;96:1094–6.
­techniques changed little over time [4]. As a result, 3. Umakanthan R, Haglund NA, Stulak JM, et al. Left
apical cannulation became standard practice after thoracotomy HeartWare implantation with outflow
widespread adoption of the HeartMate XVE graft anastomosis to the descending aorta: a simpli-
fied bridge for patients with multiple previous ster-
LVAD (Thoratec Corp.). Although care had to be notomies. ASAIO J. 2013;59:664–7.
taken at the time of implantation to avoid mechani- 4. Pool GE, Parnis SM, Creager GJ, et al. Evaluation
cal inlet obstruction, few complications arose from of occlusive inlet pannus formation: comparison of
this orientation because of the obligatory preserva- conduit designs. Trans Am Soc Artif Intern Organs.
1985;31:408–10.
tion of a ventricular reservoir with pulsatile 5. Frazier OH, Gregoric ID, Cohn WE. Initial experi-
devices. Familiarity with this implantation tech- ence with non-thoracic, extraperitoneal, off-pump
nique resulted in its subsequent application to CF insertion of the Jarvik 2000 heart in patients with pre-
LVADs―a practice bolstered by the inclusion vious median sternotomy. J Heart Lung Transplant.
2006;25:499–503.
of an inlet cannula identical to that of the HeartMate 6. Gregoric ID, Cohn WE, Frazier OH. Diaphragmatic
XVE in the design of the HeartMate II. However, implantation of the HeartWare ventricular assist
the unique physiology associated with CF tech- device. J Heart Lung Transplant. 2011;30:467–70.
 Intraoperative Anesthesia
Management 7
Marissa Wagner Mery, Siavosh Saatee,
and Charles D. Collard

ized even among cardiac procedures [1]. Indeed,

Introduction LVAD candidates often present with not only
cardiac failure but also secondary pulmonary,
­
Indications for left ventricular assist device renal, and/or hepatic pathophysiology that
(LVAD) placement have expanded in recent impacts fluid regulation, coagulopathy, and drug
years. Bridge-to-transplantation (BTT) and desti- pharmacokinetics. Moreover, the anesthesiolo-
nation therapy (DT) comprise the majority of gist will typically assume intraoperative respon-
indications, but an increasing number of patients sibility for detecting intracardiac shunts, valvular
are receiving LVAD support as a bridge-to-­ dysfunction, right heart failure, and correct can-
recovery (BTR) or even bridge-to-decision-mak- nula placement via transesophageal echocardiog-
ing (BDM). While the development of short-term raphy (TEE). This chapter will provide a
and percutaneously placed assist devices (e.g., framework for thinking about the anesthesiolo-
Impella, TandemHeart) has changed the manage- gist’s preparation and management for intraop-
ment of heart failure and post-­cardiopulmonary erative LVAD placement.
bypass (CPB) cardiogenic shock, the anesthetic
management for LVAD implantation is special-
Preoperative Evaluation
and Considerations
M.W. Mery, M.D. (*)
Texas Heart Institute at Baylor St. Luke’s Medical The LVAD recipient is typically characterized by
Center, Baylor College of Medicine,
decompensated cardiac failure refractory to med-
Houston, TX, USA
e-mail: [Link]@[Link] ical management. In addition to a severely
reduced ejection fraction (EJ), these patients are
S. Saatee, M.D.
Mechanical Circulatory Support Critical Care, Texas often characterized by an elevated pulmonary
Heart Institute at Baylor St. Luke’s Medical Center, vascular resistance, right ventricular (RV) dys-
Baylor College of Medicine, Houston, TX, USA function, coagulopathy, hepatorenal dysfunction,
e-mail:saatee@[Link]
and decreased responsiveness to catecholamines.
C.D. Collard, M.D. Thus, the preoperative exam informs an anes-
Division of Cardiovascular Anesthesiology, Texas
thetic plan tailored for a patient’s comorbidities
Heart Institute at Baylor St. Luke’s Medical Center,
Baylor College of Medicine, Houston, TX, USA and provides the opportunity for a personalized
e-mail: cdcollard@[Link] informed consent.

© Springer International Publishing AG 2018 83

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
84 M.W. Mery et al.

Cardiac Evaluation Table 7.1  Preimplantation TTE/TEE red-flag findings

and Considerations Left ventricle and interventricular septum
Small LV size, particularly with increased LV
The cardiac evaluation encompasses not only the trabeculation
etiology of the heart failure but also perioperative LV thrombus
assessment of LV and right ventricular (RV) LV apical aneurysm
function, valvular status, lesions with the poten- Ventricular septal defect
Right ventricle
tial for intracardiac shunting, arrhythmias, ven-
RV dilatation
tricular thrombus, and the presence of pulmonary
RV systolic dysfunction
hypertension. A history of previous surgeries is
Atria, interatrial septum, and inferior vena cava
warranted as many LVAD candidates have under-
Left atrial appendage thrombus
gone prior sternotomy which may influence the
PFO or atrial septal defect
surgical approach to CPB cannulation and/or the
Valvular abnormalities
potential for a subxiphoid or thoracotomy surgi- Any prosthetic valve (especially mechanical AV or
cal approach [2]. If redo sternotomy is planned, MV)
there is a higher risk of both an increased periop- >Mild AR
erative transfusion requirement and iatrogenic ≥Moderate MS
cardiovascular injury. ≥Moderate TR or >mild TS
Preoperative echocardiographic determination >Mild PS; ≥moderate PR
of LV internal dimension at end diastole (LVIDd) Other
is essential, as the comparison to post-LVAD Any congenital heart disease
placement signifies the extent of LVAD-assisted Aortic pathology: Aneurysm, dissection, atheroma,
coarctation
LV unloading (Table 7.1) [3]. A small LVIDd
Mobile mass lesion
(<63 mm) is associated with increased 30-day
Other shunts: Patent ductus arteriosus,
morbidity and mortality, and often encountered intrapulmonary
in smaller women and patients with infiltrative
AR aortic regurgitation, AV aortic valve, LV left ventricu-
cardiomyopathies [3]. lar, MS mitral stenosis, MV mitral valve, PFO patent fora-
RV failure occurs in up to 30% of patients fol- men ovale, PR pulmonary regurgitation, PS pulmonary
lowing LVAD placement and is a harbinger of a stenosis, RV right ventricle, TR tricuspid regurgitation, TS
tricuspid regurgitation
poor postoperative course complications [4–6].
There is no preoperative predictive model for RV
failure with both good sensitivity and specificity;
the right ventricular failure risk score (RVFRS) pulmonary pressures used to be a contraindica-
and TEE are often used together to identify tion for LVAD placement, but emerging evi-
potential high-risk patients for postoperative RV dence suggests fixed pulmonary hypertension
failure [4–7]. Postoperative RV failure following secondary to left heart failure can improve sig-
LVAD placement can be mitigated by concomi- nificantly in the first 6 months post-LVAD
tant placement of an RVAD, but the outcomes for placement [9]. Regardless, inhaled nitric oxide
these patients are substantially worse [8]. (iNO) and prostacyclin vasodilators should be
Therefore the presence of preoperative RV isch- available.
emic disease may warrant stenting or bypass Given the advanced heart failure in this popu-
grafting prior to LVAD placement. Preoperative lation, atrial and ventricular arrhythmias are
inotropic support should be also considered if the common and not contraindications to LVAD
RV is marginal [8]. placement. The majority of patients will present
The presence of pulmonary hypertension with automatic implantable defibrillators
should be assessed preoperatively. Severe fixed (AICDs) which will need to be deactivated in the
7  Intraoperative Anesthesia Management 85

operating room, and external defibrillation pads Other

placed. All pacemakers should be interrogated
prior to surgery for battery life and underlying Patients should have a complete preoperative
heart rhythm. Pacer-dependent patients should be neurologic exam to rule out deficits, an anes-
placed in asynchronous mode. thetic history evaluation (assessing for personal or
Potential intracardiac shunts and valvular family anesthetic complications), and an airway
pathology should be ruled out by echocardiogra- exam. Difficulty in ventilation or intubation can
phy. A patent foramen ovale (PFO) or other sep- result in a host of cardiopulmonary complications
tal defect requires pre-LVAD implantation repair, as hypoxia- or hypercarbia-related pulmonary
as the transluminal pressure gradient post-LVAD hypertension and can precipitate cardiovascular
predisposes a right-to-left shunt and intractable collapse.
hypoxia. Similarly, moderate to severe aortic
insufficiency (AI), moderate to severe mitral ste-
nosis (MS), and severe tricuspid regurgitation Preoperative Laboratory Testing
(TR) are valvular lesions requiring repair prior to and Imaging
LVAD implantation.
Finally, the patient’s cardiac medications Preoperative labs consisting of an ECG, CXR,
should be reviewed for potential anesthetic inter- pulmonary function tests, complete metabolic
actions. ACE inhibitors’ impact on afterload panel (including LFTs), complete blood count,
reduction, cardiac remodeling, and mortality in and coagulation profile, including fibrinogen and
heart failure renders them a recommended medi- a functional coagulation assessment such as a
cation for this patient population. However, their TEG or ROTEM, should be ordered.
potential to blunt catecholamine response has Echocardiogram and cardiac catheterization
been reported in the literature and may contribute should assess the transpulmonary gradient, pul-
to refractory vasoplegia. Withholding an ACE-I monary vascular resistance, pulmonary vascular
on the day of LVAD placement should be response to vasodilators, right ventricular func-
discussed. tion, cardiac output, valvular function, and LV
filling pressures. Imaging including head, chest,
and abdominal CT scans should be negative for
Renal and Hepatic Function malignancy, terminal process, or hemorrhage.
Patients should be typed and crossed, and blood
Patients classified as NYHA III or IV experience products should be immediately available in the
a decrease in volume of distribution (VD) and operating room.
reduced clearance of many intravenous medica-
tions by 50% or more [10]. Additionally, second-
ary renal and hepatic dysfunction are common in Consent/Family Discussions
patients presenting for LVAD placement. The net
result is that many common anesthetic drugs Despite their success in extending the quality and
require dose adjustments. Both renal and hepatic duration of life for a vast majority of patients,
end-organ functions can improve post-LVAD LVAD placement comes with the assumption of
implantation, but their presence preoperatively significant risk for not only mortality but also pro-
are independently associated with worsened out- longed ICU stay, renal failure, progressive cardiac
comes [11–13]. Elevated bilirubin is the labora- failure, and fatal hemorrhage. Patients and their
tory value most strongly associated with families need to be aware of this during the
mortality, and primary liver disease should be informed consent process, and the discussion
ruled out prior to LVAD placement. Pre-existing should address the patient’s wishes in the event of a
coagulopathy or electrolyte/acid-base imbal- catastrophic complication or decompensation.
ances should be corrected. There is a high degree of family member ­confusion
86 M.W. Mery et al.

at the time of end of life for LVAD patients given agent, but it comes with a risk of adrenal insuffi-
the complexity and life-sustaining nature of these ciency. If this is a concern, ketamine can be used.
devices [14]. Esmolol should be available to manage any tachy-
cardia resulting from laryngoscopy. Increased time
for circulation is required for onset of all intrave-
Monitors nous medications, and intraoperative awareness is
more frequent in patients undergoing cardiac sur-
Large-bore peripheral IV access is recommended gery [15]. Maintaining cardiac output while restrict-
in patients with prior sternotomy. If present, an ing fluids to avoid unnecessary increases in RV
AICD should be inactivated in the OR, and exter- end-diastolic pressure and maintaining adequate
nal defibrillation pads placed. An arterial line anesthetic depth are the pre-CBP goals.
should be confirmed prior to induction. A PA Baseline labs including basic chemistry panel,
catheter recommended and is useful to measure arterial blood gas, and ACT should be obtained
ventricular pressures, mixed venous saturation, following induction. Hypokalemia and hypergly-
and CVP/PCMP ratio. Unlike the LVAD flow cemia should be addressed immediately.
reading which is estimated and does not include Antibiotics should be initiated at the time of skin
the native heart’s contribution, a PA catheter is prep (or earlier with vancomycin) and redosed
also useful for cardiac output. Intraoperative TEE accordingly during the procedure. Leukocyte-­
is useful for detecting factors that may affect reduced blood should be available to reduce anti-­
VAD performance and patient outcome, includ- HLA antibody production [16]. Replacement
ing septal defects, aortic valve regurgitation, products for patients with an iatrogenic anti-
mitral stenosis, RV dysfunction, intracavitary thrombin III deficiency should be available.
thrombus, and aortic atheroma. Additionally, Patients should be anticoagulated with hepa-
TEE is useful for confirming VAD cannulae posi- rin (300–400 units/kg) and an appropriate ACT
tioning and cardiac de-airing. (>350 s) confirmed with the surgeon and perfu-
sionist prior to initiating CPB [17]. For patients
with a suspected heparin resistance manifested
I nduction and Management Prior by inappropriate ACT elevation, an additional
to Bypass dose of heparin may be administered. If this is
unsuccessful, a presumed antithrombin III defi-
I nduction and Preparation ciency can be treated with antithrombin III con-
for Cardiopulmonary Bypass centrate or FFP if the former is unavailable. If the
patient remains refractory to heparin or there is a
End-stage cardiac failure patients presenting for contraindication to heparin, bivalirudin is the
LVAD implantation are often dependent on high preferred substitute for use on CPB.
circulating concentrations of catecholamines to
maintain vasoconstriction. Acute decreases in LV
preload or increases in LV afterload are poorly Intraoperative Transesophageal
tolerated and should be avoided on anesthetic Echocardiography
induction. Decreases in heart rate (HR) are espe-
cially deleterious, as these patients cannot com- The American Society of Echocardiography rec-
pensate by increasing stroke volume [1]. Thus, ommends a TEE checklist for both pre- and post-
these patients may benefit from low-dose norepi- implantation (Table 7.2). TEE evaluation of the
nephrine or epinephrine infusion at the time of RV should begin with a qualitative assessment of
induction in order to maintain HR and CO. size, tricuspid regurgitation (TR) severity, TR eti-
Lidocaine and fentanyl are often given to blunt ology (iatrogenic or secondary to dilated annu-
the sympathetic response to laryngoscopy. lus), and ventricular motion from the tricuspid
Etomidate is the most commonly used induction annulus to the apex, inclusive of the free wall [3, 18].
Table 7.2  Perioperative TEE protocol/checklist (Adapted with permission from [3])
Two-part exam
1. Preimplantation perioperative TEE exam
Goals: Confirm any preoperative echocardiography (TTE or TEE) findings; detect unexpected abnormal findings
prior to LVAD implantation
Blood pressure: If hypotension is present, consider vasopressor agent to assess AR severity
LV: Size, systolic function, assess for thrombus
LA: Size, assess for LA appendage/LA thrombus
RV: Size, systolic function, catheters/leads
RA: Size, assess for thrombus, catheters/leads
Interatrial septum: Detailed 2D, color Doppler, IV saline contrast. Red flag: PFO/ASD
Systemic veins: Assess SVC, IVC
Pulmonary veins: Insepect
Aortic valve: red flags: >mild AR, prosthetic valve
Mitral valve: red flags: ≥moderate mitral stenosis, prosthetic mitral valve
Pulmonary valve: red flags: >mild PS, ≥moderate PR, if RVAD planned; prosthetic valve
Pulmonary trunk: red flags: Congenital anomaly (PDA, pulmonary atresia or aneurysm)
Tricuspid valve: TR, systolic PA pressure by TR velocity. Red flags: ≥moderate TR, >mild TS, prosthetic valve
Pericardium: Screen for effusion; consider constrictive physiology
Aorta: Root, ascending, transverse, and descending thoracic aorta; screen for aneurysm, congenital anomaly,
dissection, or complex atheroma at each level
2. Postimplantation perioperative TEE exam
Goals: Monitor for intracardiac air; rule out shunt; confirm device and native heart function
Pump type and speed: Confirm
Blood pressure: Via arterial line; for hypotension (MAP of <60 mmHg), consider vasopressor agent before
assessing AR severity and other hemodynamic variables
Intracardiac air: Left-sided chambers and aortic root during removal from CPB
LV: Size, inflow-cannula position and flow velocities, septal position. Red flags: Small LV (over-pumping or RV
failure), right-to-left septal shift; large LV (obstructed or inadequate pump flows)
Inflow-cannula position: 2D/3D, assess for possible malposition
Inflow-cannula flow: Spectral and color Doppler (red flag: Abnormal flow pattern/high/low velocities, especially
after sternal closure)
LA: Assess LA appendage
RV: Size, systolic function. Red flags: Signs of RV dysfunction
RA: Size, assess for thrombus, catheters/leads
Interatrial septum: Repeat IV saline test and color Doppler evaluation of IAS (red flags: PFO/ASD)
Systemic veins: (SVC, IVC)
Pulmonary veins: Inspect
Aortic valve: Degree of AV opening and degree of AR (red flags: >mild AR)
Mitral valve: Exclude inflow-cannula interference with submitral apparatus; assess MR
Pulmonary valve: Assess PR, measure RVOT SV if able
Pulmonary trunk: (if applicable, demonstrate RVAD outflow by color Doppler); assess PR
Tricuspid valve: Assess TR (red flags: >moderate TR); systolic PA pressure by TR velocity (if not severe TR)
Pericardium: Screen for effusion/hematoma
Aorta: Exclude iatrogenic dissection
Outflow graft: Identify conduit path adjacent to RV/RA with color and spectral Doppler (when able)
Outflow graft-to-aorta anastomosis: Assess patency/flow by color and spectral Doppler (when able). Red flags:
Kinked appearance/turbulent flow/velocity > 2 m/s, particularly after sternal closure
2D two-dimensional, 3D three-dimensional, AR Aortic regurgitation, ASD Atrial septal defect, AV Aortic valve, CPB
Cardiopulmonary bypass, IAS Interatrial septum, IV Intravenous, IVC Inferior vena cava, LA Left atrium, LV Left ven-
tricle, LVAD Left ventricular assist device, LVOT Left ventricular outflow tract, MAP Mean arterial pressure, MR Mitral
regurgitation, PA Pulmonary artery, PFO Patent foramen ovale, PDA Patent ductus arteriosus, PR Pulmonary regurgita-
tion, PS Pulmonary stenosis, RA Right atrium, RV Ventricle, RVAD Right ventricular assist device, RVOT Right ven-
tricular outflow tract, SV Stroke volume, SVC Superior vena cava, TEE Transesophageal echocardiography, TR Tricuspid
regurgitation, TS Tricuspid stenosis, TTE Transthoracic echocardiography
88 M.W. Mery et al.

Systolic function and dilatation should be noted, placement and to identify the risk for RV over-
as RV end-diastolic diameter was one of the two load in the event of pulmonic insufficiency.
echocardiographic variables recently identified Unaddressed moderate or severe mitral stenosis
as independently predictive of RV failure [3, 19]. will impair LVAD filling and can hamper RV
3D volume assessment, tricuspid annular plane function. Mitral regurgitation does not impact
systolic excursion (TAPSE), global and regional LVAD function. Finally baseline tricuspid regur-
RV fractional change area, and the maximum gitation (TR) and annular dilatation should be
derivative of the RV pressure (dP/dt max) have noted because the LVAD’s presence and its resul-
also been mentioned as quantitative options for tant LV decompression can worsen TR. Tricuspid
evaluating systolic function [3, 18]. However, repair or replacement should be considered if TR
these tools are not always available, and the tech- is moderate or severe [3].
niques are challenging in this population [3]. At Severe aortic atherosclerotic disease (ather-
present there is no consensus on reliable predic- oma >5 mm or protruding) and calcifications
tors of RV failure following LVAD placement, should be identified as they are associated with
but given the high morbidity and mortality asso- an increased risk of embolic events. Likewise
ciated with that event, thorough pre-­bypass and atrial and ventricular thrombi should be ruled out
postimplant examinations of both ventricles are in this population, as they are often located near
imperative. the apical implantation site and may increase the
Any potential intracardiac shunt should be risk of perioperative stroke [3, 18].
identified [3]. LVAD placement leads to a pre-
cipitous decline in LV and left atrial (LA) pres-
sures, and a patent foramen ovale (PFO) or other  anagement on Cardiopulmonary
M
septal defect could result in right-to-left shunting, Bypass
systemic hypoxia, and paradoxical thromboem-
boli [3, 20]. Identification prior to bypass is General Anesthetic Management
essential because the repair may alter the sur-
geon’s cannulation technique [18]. Our preferred While on bypass, blood pressure control is the pri-
technique for visualizing PFOs is via agitated mary goal with titration of vasopressors as neces-
saline in the mid-esophageal bicaval view. While sary, favoring vasopressin and norepinephrine. In
pre-bypass identification is ideal, Valsalva the event of refractory vasoplegia, there are case
maneuver is not always successful in identifying reports of success with methylene blue adminis-
PFOs in patients with severe heart failure and tration, which is hypothesized to inhibit guanylyl
elevated atrial pressures. Therefore confirming cyclase, although its use in small case series has
the absence of a PFO following LVAD placement not been shown to improve overall mortality [22].
is advised [18]. Moreover methylene blue’s pulmonary vasocon-
Valvular deficiencies need to be identified and strictive properties are concerning. Electrolytes,
often repaired prior to implantation. Aortic insuf- glucose, and hemoglobin should be monitored
ficiency (AI) is especially problematic. The frequently. Some institutions recommend magne-
LVAD decreases LV end-diastolic pressure, sium (4 g) and lidocaine (100 mg) loading prior to
thereby increasing the aortic transvalvular gradi- inflow cannula insertion to reduce ventricular
ent. Flow through the LVAD is increased com- arrhythmogenicity [23].
pared to normal, but the systemic and coronary
flow is inadequate because the leaky AV results in
preferential LV filling [21]. Aortic stenosis is less Cannula Positioning and De-airing
of an acute problem, but an aortic valve that does
not intermittently open increases the risk of pump TEE can guide optimal positioning of the LVAD
thrombosis [18]. Similar examination should be inflow cannula toward the mitral valve and away
done for the pulmonic valve in the event of RVAD from the ventricular septum (Fig. 7.1). The tip
7  Intraoperative Anesthesia Management 89

Fig. 7.1 TEE
verification of LVAD
inflow cannula (shown
by arrow) placement
(Panel a, b)

should rest in the center of the chamber, away caused for concern [3]. De-airing needs to occur
from the ventricular free wall, to minimize suc- before weaning from bypass [3]. Air migration
tion events. Aortic dissection can occur, rarely, into the right coronary artery can precipitate RV
during suturing of the outflow graft to the aorta ischemia and failure, and air in the LVAD outflow
(Fig.  7.2). Intimal tears, aortic valve pathology, track can result in a cerebrovascular event. Air
and true and false lumens can be assessed. When should be excluded from the pulmonary veins,
the LVAD is functional, outflow should be unidi- left atrium, LV, LVAD outflow graft, proximal
rectional and laminar; any flow >2 m/s should be ascending aorta, and right coronary sinus [3, 18].
90 M.W. Mery et al.

Fig. 7.2 TEE
verification of LVAD
outflow cannula
placement (arrow marks
outflow cannula entering
into aorta (AO))

Fig. 7.3 TEE
verification of right
ventricular septal
bowing (arrow) due to
right ventricular failure
and/or underfilling of
the left ventricle

 VAD Speed Determination

L promise RV contractility, whereas inadequate LV
and Ventricular Function decompression and a rightward septal shift indi-
cate inadequate LVAD flows (Fig. 7.3) [3]. The
Once the LVAD is implanted and functioning, the aortic valve (AV) should be evaluated for opening
interventricular septum should be in a neutral to during systole. Fixed AV closure can be accom-
leftward position with the LV moderately decom- panied by adequate systemic flow but increases
pressed [3]. Exaggerated decompression and sig- the risk of AV thrombotic events. A permanently
nificant leftward shift of the interventricular closed AV may be inevitable in severe heart fail-
septum can predispose suction events and com- ure (opening is associated with degree of LVAD
7  Intraoperative Anesthesia Management 91

support), repaired aortic insufficiency, or aortic [2, 23, 26, 27]. Sternal-sparing surgeries decrease
stenosis, or it may be able to be ameliorated with the future sternotomy risk for patients awaiting a
speed changes on the LVAD device. Finally, the transplant and protect bypass grafts or previous
heart should be examined again for intracardiac repairs for congenital heart disease. Options cited
shunts, aortic regurgitation, and right ventricular in the literature include implantation via thora-
function [3]. cotomy, hemi-sternotomy, and diaphragmatic
approaches [23, 28, 29]. Anesthetic consider-
ations for these approaches need to include
Post-cardiopulmonary Bypass potential for massive blood loss, ventilation strat-
Management egies, and optimization of the left ventricular
cavity for inflow graft placement in a systemi-
Post-CPB implantation care now focuses on cally pumping heart.
LVAD speed optimization, assessment and aug- An initial case series suggests blood product
mentation of RV function, and determining utilization may be less for patients undergoing
whether a BiVAD will be necessary. In conjunc- implantation via thoracotomy than traditional ster-
tion with this, CPB-associated vasoplegia, hyper- notomy, but the potential for massive bleeding
tension, and coagulopathy will need to be treated. with reduced visualization is possible [23, 27].
While hypertension is rare, it can impede sys- With a thoracotomy approach, both single- and
temic LVAD flow. double-lumen tubes have been used successfully
RV failure occurs in up to 30% of patients fol- for patients, and the discussion regarding prefer-
lowing LVAD placement and is a significant ence should happen with the surgeon prior to intu-
cause of morbidity and mortality [4–6, 24]. While bation. Patients for whom interrupted ventilation
an LVAD may provide RV afterload reduction, it may result in decompensation, such as those with
can mechanically impair RV contractility and COPD, may be better candidates for double-­lumen
elevate RV filling pressures. Optimizing inotro- endotracheal tubes. Separately, a team from
pes, reducing pulmonary afterload, and rectifying Vanderbilt cites an off-pump approach as a means
volume status offer the best hope for avoiding of RV protection, as it spares pericardial opening
RVAD placement. However, if an RVAD is and uncontrolled RV dilatation [23]. Moreover, the
needed, it is best placed during the initial surgery adenosine used in their technique (to assist with
as outcomes worsen for patients who need to inflow cannula placement off-­pump) is postulated
return to the OR for RVAD placement as a subse- to protect the RV further via adenosine-induced
quent operation [25]. pulmonary vasodilatation [23].
Following bypass cessation and confirmed
hemodynamic stability, protamine will be given
to reverse heparin anticoagulation. An ACT  evices Placed Unexpectedly
D
should be <150 s. A full coagulation panel includ- for Inability to Come Off Bypass
ing PT, PTT, INR, fibrinogen, and TEG or
ROTEM should be run. Derangements should be Short-term options for cardiac or cardiopulmonary
managed. support include extracorporeal VADs (i.e.,
Levitronix, Bio-Medicus, Abiomed AB5000) and
extracorporeal membrane oxygenation (ECMO).
Special Situations Both benefit from TEE support to rule out intracar-
diac shunts and AI and confirm cannula place-
 inimally Invasive and Off-Pump
M ment. ECMO in the setting of significant
Approaches post-cardiotomy bleeding is a challenge, and
coagulation management will be adjusted based
Minimally invasive and off-pump approaches to on the risk of ongoing bleeding vs. circuit
LVAD implantation are increasing in frequency thrombosis.
92 M.W. Mery et al.

 ercutaneous Devices Placed

P 9. Mikus E, Stepanenko A, Krabatsch T, Loforte A,
Dandel M, Lehmkuhl HB, et al. Reversibility of fixed
in the Catheterization lab pulmonary hypertension in left ventricular assist
device support recipients. Eur J Cardiothorac Surg.
Percutaneous devices such as Impellas and 2011;40:971–7.
TandemHearts are being used increasingly as a 10. Ogawa R, Stachnik JM, Echizen H. Clinical pharma-
cokinetics of drugs in patients with heart failure. Clin
BTR or BTD in the catheterization lab. A BTD
Pharmacokinet. 2013;52:169–85.
placement might involve the options of recovery, 11. Sandner SE, Zimpfer D, Zrunek P, Rajek A, Schima
placement of a long-term LVAD, or transplant H, Dunkler D, et al. Renal function and outcome after
[30, 31]. The Impella in particular is also being continuous flow left ventricular assist device implan-
tation. Ann Thorac Surg. 2009;87:1072–8.
invoked for support and cardioprotection during
12. Xuereb L, Go PH, Kaur B, Akrawe S, Borgi J, Paone
high-risk percutaneous interventions and ventric- G, et al. Should patients with hepatic fibrosis undergo
ular tachycardia ablations. Despite the popularity LVAD implantation: a comparative analysis. ASAIO
of these devices, in general, they are short-term J. 2016;62:498–500.
13. Maltais S, Stulak JM. Right and left ventricular assist
devices and do not provide the same level or
devices support and liver dysfunction: prognostic
duration of ventricular support as an LVAD. and therapeutic implications. Curr Opin Cardiol.
2016;31:287–91.
14. McIlvennan CK, Jones J, Allen LA, Swetz KM,

Nowels C, Matlock DD. Bereaved caregiver perspec-
References tives on the end-of-life experience of patients with
a left ventricular assist device. JAMA Intern Med.
1. Mets B. Anesthesia for left ventricular assist device place- 2016;176:534–9.
ment. J Cardiothorac Vasc Anesth. 2000;14:316–26. 15. Barry AE, Chaney MA, London MJ. Anesthetic man-
2. Karaca N, Sahutoglu C, Kocabaş S, Orhaner BT, agement during cardiopulmonary bypass: a system-
Askar FZ, Ertugay S, et al. Anesthetic management atic review. Anesth Analg. 2015;120:749–69.
for left ventricular assist device implantation without 16. Jackman RP, Deng X, Bolgiano D, Utter GH,
using cardiopulmonary bypass: case series. Transplant Schechterly C, Lebedeva M, et al. Leukoreduction
Proc. 2015;47:1503–6. and ultraviolet treatment reduce both the magnitude
3. Stainback RF, Estep JD, Agler DA, Birks EJ, and the duration of the HLA antibody response.
Bremer M, Hung J, et al. Echocardiography in the Transfusion. 2014;54:672–80.
management of patients with left ventricular assist 17. Gravlee GP, Haddon WS, Rothberger HK, Mills
devices: recommendations from the american soci- SA, Rogers AT, Bean VE, et al. Heparin dosing and
ety of echocardiography. J Am Soc Echocardiogr. monitoring for cardiopulmonary bypass. A com-
2015;28:853–909. parison of techniques with measurement of subclini-
4. Haddad F, Hunt SA, Rosenthal DN, Murphy DJ. Right cal plasma coagulation. J Thorac Cardiovasc Surg.
ventricular function in cardiovascular disease, part I: 1990;99:518–27.
anatomy, physiology, aging, and functional assessment 18. Patangi SO, George A, Pauli H, O’Leary D, Roysam
of the right ventricle. Circulation. 2008;117:1436–48. C, Butt T, et al. Management issues during HeartWare
5. Haddad F, Doyle R, Murphy DJ, Hunt SA. Right left ventricular assist device implantation and the
ventricular function in cardiovascular disease, part role of transesophageal echocardiography. Ann Card
II: pathophysiology, clinical importance, and man- Anaesth. 2013;16:259–67.
agement of right ventricular failure. Circulation. 19. Aissaoui N, Salem J-E, Paluszkiewicz L, Morshuis
2008;117:1717–31. M, Guerot E, Gorria GM, et al. Assessment of right
6. Matthews JC, Koelling TM, Pagani FD, ventricular dysfunction predictors before the implan-
Aaronson KD. The right ventricular failure risk score. tation of a left ventricular assist device in end-stage
A pre-­operative tool for assessing the risk of right heart failure patients using echocardiographic mea-
ventricular failure in left ventricular assist device can- sures (ARVADE): combination of left and right ven-
didates. J Am Coll Cardiol. 2008;51:2163–72. tricular echocardiographic v. Arch Cardiovasc Dis.
7. Kalogeropoulos AP, Kelkar A, Weinberger JF, 2015;108:300–9.
Morris AA, Georgiopoulou VV, Markham DW, et al. 20. McCarthy PM, Savage RM, Fraser CD, Vargo R,
Validation of clinical scores for right ventricular James KB, Goormastic M, et al. Hemodynamic and
failure prediction after implantation of continuous-­ physiologic changes during support with an implant-
flow left ventricular assist devices. J Heart Lung able left ventricular assist device. J Thorac Cardiovasc
Transplant. 2015;34:1595–603. Surg. 1995;109:409–17; discussion 417–8.
8. Taghavi S, Jayarajan SN, Komaroff E, Mangi AA. Right 21. Kellman SE, Feider AJ, Jeevanandam V, Chaney MA.
ventricular assist device results in worse post-transplant Can intraoperative transesophageal echocardiography
survival. J Heart Lung Transplant. 2016;35:236–41. predict postoperative aortic insufficiency in patients
7  Intraoperative Anesthesia Management 93

receiving implantable left ventricular assist devices? demonstrating an integrated multidisciplinary strat-
J Cardiothorac Vasc Anesth. 2015;29:901–5. egy. J Cardiothorac Vasc Anesth. 2015;29:271–4.
22. Hosseinian L, Weiner M, Levin MA, Fischer
27. Khalpey Z, Bin Riaz I, Marsh KM, Ansari MZA,
GW. Methylene blue: magic bullet for vasoplegia? Bilal J, Cooper A, et al. Robotic left ventricular assist
Anesth Analg. 2016;122:194–201. device implantation using left thoracotomy approach
23. Maltais S, Davis ME, Haglund N. Minimally inva- in patients with previous sternotomies. ASAIO
sive and alternative approaches for long-term LVAD J. 2015;61:e44–6.
placement: the Vanderbilt strategy. Ann Cardiothorac 28. Riebandt J, Sandner S, Mahr S, Haberl T, Rajek
Surg. 2014;3:563–9. A, Laufer G, et al. Minimally invasive thoratec
24. Harjola V-P, Mebazaa A, Čelutkienė J, Bettex D, Heartmate II implantation in the setting of severe
Bueno H, Chioncel O, et al. Contemporary manage- thoracic aortic calcification. Ann Thorac Surg. 2013;
ment of acute right ventricular failure: a statement 96:1094–6.
from the heart failure association and the working 29. Anyanwu AC. Technique for less invasive implanta-
group on pulmonary circulation and right ventricular tion of Heartmate II left ventricular assist device with-
function of the european society of cardiology. Eur out median sternotomy. Semin Thorac Cardiovasc
J Heart Fail. 2016;18:226–41. Surg. 2011;23:241–4.
25. Fitzpatrick JR, Frederick JR, Hiesinger W, Hsu VM, 30. Lima B, Kale P, Gonzalez-Stawinski GV, Kuiper JJ,
McCormick RC, Kozin ED, et al. Early planned insti- Carey S, Hall SA. Effectiveness and safety of the
tution of biventricular mechanical circulatory support impella 5.0 as a bridge to cardiac transplantation or
results in improved outcomes compared with delayed durable left ventricular assist device. Am J Cardiol.
conversion of a left ventricular assist device to a 2016;117:1622–8.
biventricular assist device. J Thorac Cardiovasc Surg. 31. Pieri M, Contri R, Winterton D, Montorfano M,

2009;137:971–7. Colombo A, Zangrillo A, et al. The contemporary role
26. Wagner CE, Bick JS, Kennedy J, Haglund N, Danter of Impella in a comprehensive mechanical circulatory
M, Davis ME, et al. Minimally invasive thoracic left support program: a single institutional experience.
ventricular assist device implantation; case series BMC Cardiovasc Disord. 2015;15:126.
 Perioperative Management
of LVAD Patients 8
Krishna Ayyagari, William Patrick Mulvoy III,
Arthur W. Bracey, Cesar A. Castillo,
and James P. Herlihy

A recent INTERMACS (Interagency Registry for mately 5% were due to respiratory failure [3].
Mechanically Assisted Circulatory Support) The remaining deaths were mainly due to device
annual report on the outcomes of patients who malfunction, arrhythmias, or other less common
underwent continuous-flow left ventricular assist complications [3, 4]. Table 8.1 presents the nota-
device (CF-LVAD) or biventricular assist device ble postoperative complications and their fre-
implantation showed that these patients had a quency of occurrence. These data suggest that
1-year survival rate of 80% [1]. Most of the mor- poor LVAD implantation outcomes can often be
talities occurred within the first 30 days after prevented or effectively managed. Therefore,
device implantation or during the postoperative diligent perioperative care is paramount to ensure
period of the index hospitalization [1, 2]. The positive outcomes for these patients.
major causes of death during this period were as This chapter will focus on the perioperative
follows: 60–65% were due to multisystem organ management of patients who have received one of
failure (MSOF), which was driven primarily by the two prevalent and FDA-approved CF-LVADs:
poor oxygen delivery (DO2) and, often, specifi- the HeartMate II (HM2), which was approved for
cally by right heart failure (RHF) [1, 2]; 15–20% bridge to transplantation in 2008 and for destina-
were due to embolic and hemorrhagic stroke [1, 2]; tion therapy in 2010, and the HeartWare HVAD
10–15% were due to bleeding events [1, 2]; (HW HVAD), which was approved for bridge to
5–10% were sepsis related [3, 4]; and approxi- transplantation in 2012.

K. Ayyagari, M.D. • W.P. Mulvoy III, M.D., M.B.A

Texas Heart Institute at Baylor St. Luke’s Medical C.A. Castillo, M.D.
Center, Baylor College of Medicine, Anesthesia Critical Care, Texas Heart Institute at
Houston, TX, USA Baylor St. Luke’s Medical Center, Baylor College of
e-mail: [Link]@[Link]; Medicine, Houston, TX, USA
willmulvoy@[Link] e-mail: ccastillo5@[Link]
A.W. Bracey, M.D. J.P. Herlihy, M.D. (*)
Clinical Pathology, Texas Heart Institute at Baylor Critical Care Services, Texas Heart Institute at Baylor
St. Luke’s Medical Center, Baylor College of St. Luke’s Medical Center, Baylor College of
Medicine, Houston, TX, USA Medicine, Houston, TX, USA
e-mail: a2bracey@[Link] e-mail: jpherlihy@[Link]

© Springer International Publishing AG 2018 95

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
96 K. Ayyagari et al.

Table 8.1  Early postoperative complications of left ven- A substantial percentage of patients consid-
tricular assist device (LVAD) implantation
ered for LVAD implantation are malnourished or
Complication % of cases at risk for malnourishment through the patho-
Right heart failure 10–39%; up to 50% physiology of cardiac cachexia [6]. Indeed, only
when LVAD emergently 10% of patients being evaluated for heart trans-
placed
plantation or LVAD implantation are considered
Respiratory failure 6–40%
to be “well nourished” [7], and insufficient nutri-
Cerebral vascular accident 10–15%
Transient ischemic attack 4–12%
tional status is associated with poor LVAD
Delirium 10%
implantation outcomes [8]. Therefore, preopera-
Renal failure 3–33% tive nutritional assessment of LVAD candidates is
Hepatic failure 2–8% recommended, including at least a prealbumin
Arrhythmia 30–60% screening and possibly more advanced measures
 Atrial 25% of nutritional status, such as caloric expenditure,
 Ventricular 22–52% along with specific, nutritionist-directed inter-
Thromboembolism 6% pulmonary or ventions [6–9].
systemic vasculature All unnecessary lines and catheters should be
Hemolysis 3–5% removed from patients who are to undergo LVAD
Infection Up to 42% of index implantation. A dental evaluation and treatment
hospitalizations of any active or potential infections is also rec-
Bleeding requiring 31–81% ommended before LVAD placement. In addition,
transfusion
any other active infections should be fully treated
Bleeding requiring 31%
reoperation preoperatively [5, 10]. The administration of pro-
Tamponade 15–28% phylactic antibiotics within 60 min of the inci-
sion and continued use for 48 h postoperatively
has become standard practice [5, 11]. The most
recent guidelines suggest that antibiotic regimens
 reoperative Management of LVAD
P include coverage for both gram-positive and
Patients gram-negative organisms [5]. The most aggres-
sive regimens, which we favor, include preop-
The perioperative management of an LVAD erative administration of vancomycin, a
recipient begins with preoperative optimization broad-spectrum cephalosporin, and fluconazole,
of the patient’s clinical status. Clear risk factors followed by a 2-day course of vancomycin in
for poor outcomes after LVAD implantation have combination with cephalosporin. Patients with a
emerged over the past decade or so [1, 2, 5]. high preoperative risk for or who had a preop-
These are discussed in detail in Chap. 2. However, erative nasal swab positive for methicillin-resis-
several of these adverse conditions are often, at tant Staphylococcus aureus should be given
least to some degree, remedial. Effective inter- rifampin and nasal topical mupirocin preopera-
ventions for such conditions can sometimes allow tively and for 7 days postoperatively [4]. Standard
for positive outcomes after LVAD placement. topical skin preps should include the use of
The recently published guidelines for mechanical chlorhexidine solution [3, 12].
circulatory support (MCS) from the International The function of major nonheart organs, includ-
Society for Heart and Lung Transplantation pro- ing the kidneys, liver, and lungs, should be
vide a helpful summary of these preoperative optimized preoperatively. Renal failure, depend-
treatment strategies [5]. Optimization of cardiac ing on degree, is a major risk factor for poor
status, particularly addressing RHF, is a major LVAD outcomes [13–16]. In fact, end-stage renal
focus of preoperative care. Right heart failure is disease requiring hemodialysis is currently con-
discussed in separate chapters, but we discuss sidered a contraindication for LVAD implanta-
remediation strategies for other conditions below. tion [5]. Therefore, it is generally recommended
8  Perioperative Management of LVAD Patients 97

that renal function be optimized before device and possibly biopsy of the liver [5, 20]. Cirrhosis
implantation [5]. In the case of advanced conges- and having a high MELD score are considered
tive heart failure (CHF), renal perfusion is often contraindications for LVAD implantation [5].
compromised. In this circumstance, it is recom- Patients with these contraindications are typi-
mended that renal function be supported by opti- cally suggested to undergo combined heart and
mizing the patient’s hemodynamics with liver transplantation [5, 20]. However, for patients
pharmacologic therapies and, perhaps, temporary with hepatic dysfunction that is not severe enough
MCS [5]. The evolving practice of using tempo- to exclude them from consideration for LVAD
rary MCS as a bridge to LVAD implantation in implantation, the current recommendations sug-
patients with MSOF is covered in another chap- gest pharmacologic improvement of hepatic per-
ter, but early studies have suggested that such fusion and treatment of RHF [20]. As suggested
strategies are very promising [5, 17]. However, for patients with renal failure, one strategy that
studies are needed to assess the specific strategy may be successful for improving outcomes of
of targeting preoperative renal function to patients with liver dysfunction could be to use
improve LVAD implantation outcomes. Volume temporary mechanical right and left heart support
overload, usually demonstrated by a central to improve or normalize liver function before
venous pressure (CVP) of ≥16, has been shown LVAD implantation [20, 26]. A recent publica-
to increase the risk of poor LVAD implantation tion from a German hospital reported excellent
outcomes [18], and it has generally been agreed outcomes (75% survival at 1 year) in a large
that CVP should be controlled preoperatively by cohort of relatively young patients (mean age,
using diuretics or hemodialysis techniques [5]. 35 ± 12 years) with acute hepatic failure who
However, it is sometimes difficult to separate out- underwent HM2 or HW HVAD implantation
right RHF and volume overload from renal dys- [27]. Interestingly, 85% were on pharmacologi-
function when conducting a study. Interestingly, cal support, and 41% were on some form of MCS
volume overload has also been shown to increase going into surgery. For any patient with a demon-
the risk of acute kidney injury after LVAD strated liver abnormality, we recommend admin-
implantation [13, 19]. istering vitamin K (10 mg, intramuscular or
Heart failure adversely affects liver function intravenous, though it is worth noting the risk of
and can lead to liver injury by inducing both rare but potentially dramatic anaphylaxis with
hypoperfusion (ischemic hepatitis) and venous such) preoperatively, as well as maintaining vigi-
congestion (leading to cardiac cirrhosis) [20–22]. lance for coagulopathy both intraoperatively and
Of these two possibilities, it appears that RHF postoperatively, because the liver makes many of
and passive venous congestion have a more pro- the key coagulation factors.
found adverse effect. Advanced liver dysfunction Pulmonary function is often impaired in
can cause coagulopathy and vasodilation [20, patients with advanced heart failure due to inter-
22]. It has been well established that patients stitial and alveolar edema, cardiomegaly, pleural
with liver cirrhosis or a high model for end-stage effusions, and secondary pulmonary hyperten-
liver disease (MELD) score are at high risk for sion, which can cause restrictive and obstructive
adverse outcomes after LVAD implantation [20, lung defects, impaired gas exchange, decreased
23–25]. In fact, recent data from a single-center lung compliance, increased work to breath [28–
study suggest that significantly elevated levels of 32], and in the case of cardiac cachexia, respira-
aspartate aminotransferase and alanine transami- tory muscle dysfunction. Studies correlating
nase and the need for a preoperative liver biopsy preoperative pulmonary function test results to
are powerful independent predictors of survival LVAD outcomes are lacking. A limited number of
for HM2 and HW HVAD recipients [26]. studies have suggested that pulmonary function
Therefore, patients with abnormal liver function may be reduced after LVAD implantation, in
tests are recommended to undergo ultrasound spite of otherwise controlled CHF, particularly in
evaluation of the liver, hepatology consultation, HM2 recipients [31, 32]. Limited data suggest
98 K. Ayyagari et al.

that the HM2 device may affect diaphragm func- actually declines immediately after CF-LVAD
tion [33], presumably because implantation implantation and continues to decline over time [37].
requires transection of the left anterior hemidia- Low cardiac output state, a well-recognized
phragm and because the pump for the HM2 lies complication of CPB, is essentially characterized
directly below the left hemidiaphragm, impairing by poor ventricular performance [38]. The cause
its movement. However, there is substantial evi- is thought to be multifactorial, including myocar-
dence correlating poor preoperative pulmonary dial ischemia during cross-clamping, reperfusion
function to adverse outcomes after cardiac sur- injury, cardioplegia-induced myocardial dys-
gery not involving ventricular assist device function, and activation of inflammatory cas-
(VAD) implantation, and standard recommenda- cades, resulting in suppression of myocardial
tions suggest optimizing pulmonary function function. In our experience, patients who undergo
before cardiac surgery [34, 35]. We recommend LVAD implantation have the same complica-
controlling pulmonary edema and any airway tions, except that the LVAD, presumably, com-
obstructions due to CHF or underlying obstruc- pensates for impaired left ventricular (LV)
tive lung disease as well as possible before the performance. The effects of LVAD implantation
operation. However, again, no specific interven- on the right heart are complex; these are discussed
tion studies have confirmed that this approach is in Chap. 18.
beneficial. Additionally, preoperative respiratory Cardiac vasoplegia syndrome (CVS) is a form of
muscle training may be beneficial for patients vasodilatory shock that occurs in up to 44% of CPB
with advanced heart failure and respiratory mus- patients [37]. This syndrome is caused by blood
cle dysfunction, as has been shown for a high-­ exposure to the CPB circuit and the consequent neu-
risk cohort of patients undergoing general cardiac rohumoral factor and inflammatory mediator activa-
surgery [36]. tion. The clinical picture of CVS is typical of a
Finally, coagulopathy and platelet dysfunction systemic inflammatory response, with a low sys-
should be corrected before a patient undergoes temic vascular resistance and hypotension, but it can
LVAD implantation [5, 18]. Many LVAD candi- potentially be characterized by its resistance to typi-
dates are on anticoagulation and antiplatelet regi- cal vasopressor dosing [39]. Patients who undergo
mens preoperatively (see “Bleeding and LVAD implantation are at greater risk of developing
Hemostasis Considerations” section below). CVS, and it appears that CVS may more signifi-
cantly affect the outcomes of LVAD recipients than
those of typical cardiac surgery patients. A recent
 ffects of Cardiopulmonary Bypass
E study showed that norepinephrine-­resistant CVS is
(CPB), Anesthesia, and Surgery associated with a 25% mortality rate in patients who
on the Postoperative Course have undergone LVAD implantation [39]. Cardiac
After LVAD Implantation vasoplegia syndrome should be treated with vaso-
pressors, particularly vasopressin and norepineph-
CPB rine. For refractory cases, methylene blue,
cyanocobalamin, and steroids may be used [39].
Immediately after CPB is initiated and the aortic Another factor to consider during the postop-
cross clamp is placed, pulmonary capillary wedge erative care of LVAD recipients is the fluid shifts
pressure (PCWP) and pulmonary artery pressure associated with CPB and surgery. When patients
(PAP) increase. This can cause pulmonary vascular are initially placed on bypass, the circuit priming
endothelial dysfunction and result in pulmonary volume adds to the intravascular volume. On the
vasoconstriction and, ultimately, right heart strain other hand, intraoperative bleeding may decrease
when the patient is being weaned from intravascular volume during surgery. For these
CPB. However, a number of studies have demon- reasons, it is very difficult to predict intravascular
strated that pulmonary vascular resistance (PVR) volume and, hence, right ventricular (RV) load-
ing at the end of CPB. This is among the reasons
8  Perioperative Management of LVAD Patients 99

why it is standard to perform transesophageal recipients leave the operating room with arterial
echocardiography (TEE) at the termination of and pulmonary artery catheters for monitoring
CPB and specifically important to view the right pressure and measuring mixed venous oxygen
ventricle and intraventricular septum. The posi- saturation (MVO2) and cardiac output by thermo-
tion of the septum and other aspects of the right dilution. The patients are also on inotropic sup-
ventricle are more fully discussed below, but port, often consisting of low doses of milrinone
essentially, a midline intraventricular septum is and a catecholamine (usually dobutamine or epi-
consistent with optimal intravascular volume and nephrine). As discussed in detail below, the ino-
RV loading [40]. tropes are primarily for supporting the right
ventricle, which is at risk of stress and decom-
pensating after LVAD placement [44]. In addi-
Anesthesia tion, the patients are usually on vasopressors to
treat CVS; a combination of vasopressin and nor-
Anesthetic agents and rewarming can compound epinephrine is standard at our institution.
the vasodilatory effects of CPB. Short-acting,
less potent anesthetics are less likely to cause
vasodilation. Some practitioners choose to avoid Postoperative Course,
volatile agents altogether and use propofol Complications, and Management
instead. However, propofol can also significantly
decrease systemic vascular resistance. The postoperative period is often characterized
by labile hemodynamics (typically caused by
volume shifts), vasoplegia, and RV dysfunction.
Surgery Partly as a consequence of this, major organ
function can change quickly in the postoperative
Placement of the inflow cannula in the apical seg- period. Not surprisingly, postoperative complica-
ments of the left ventricle causes increased LV tions occur frequently after LVAD implantation.
dysfunction and loss of apical contractility. The major postimplantation complications are
Analysis of LV tissue obtained up to a year after listed in Table 8.1. Because of the possible com-
LVAD implantation showed persistent damage to plications, our care model is for the anesthesiolo-
myocytes and contractility derangements [41]. gist and surgeon to directly hand off the care of
Because the majority of the LV contraction is the patient to a postanesthesia care unit (PACU)/
based on twisting and untwisting of the apical por- ICU intensivist and nursing team at the bedside.
tion to eject the blood, coring out the LV apex in Typically, the required nurse-to-patient ratio is
order to place the inflow cannula invariably wors- either 1:1 or 2:1 for these patients. It is impera-
ens LV contractility, compliance, and output. The tive that all team members are well trained and
loss of apical contractility adds to the decrease in experienced. The complexity and dynamic nature
inotropy seen once the LVAD is implanted [42], of the immediate postoperative course requires
although LV function can certainly improve over the proximate presence of the ICU team and clear
time after LVAD implantation [43]. lines of communication to the cardiologist and
surgical team for best outcomes.

Coming Out of the Operating Room

Pulmonary Considerations
Once the patient is stable and the LVAD speed
has been set based on appropriate TEE imaging As with any patient who has undergone cardiac
[40], it is time to transport the patient to the inten- surgery, the initial PACU assessment for LVAD
sive care unit (ICU) and immediately begin car- recipients begins with an evaluation of gas
diac intensive care management. All of our LVAD exchange and pulmonary status; however, for
100 K. Ayyagari et al.

these patients, evaluation and management of effects on LV wall tension. Our general approach
mechanical ventilation, which virtually all LVAD is to assure a low peak airway pressure
recipients are on when they leave the operating (<20 cmH2O), as concerns about PVR and RHF
room, is particularly important. Continuous-flow are paramount. Typically, high RV preload and
MCS, though not extensively studied, does not high LV afterload are dealt with by using inter-
appear to affect the ventilation/perfusion relation- ventions other than mechanical ventilation
ships in the lungs or gas exchange [45]. Gas manipulation. The best postoperative ventilator
exchange is evaluated with hemoglobin oxygen management involves using settings that promote
saturation and arterial blood gas testing. However, gas exchange, prevent and/or treat atelectasis,
the standard pulse oximetry test for assessing and avoid a high Paw. As of yet, no studies have
oxygen saturation of the blood may be unreliable specifically sought to determine the optimal ven-
in CF-LVAD recipients [46]. Obviously, the goal tilator practice in the LVAD postimplantation set-
for the arterial hemoglobin oxygen saturation ting [44], so we follow standard post-cardiac
level is greater than 90% for optimal DO2. The surgery guidelines for ventilation management.
goal for the carbon dioxide arterial tension is to be Our usual mechanical ventilation mode is vol-
near but not over 40 mmHg in order to optimize ume cycled with a Vt of 6–8 mL/kg and a PEEP
the acid-base balance. But for LVAD recipients, setting of 5 cmH2O. The respiratory rate is typi-
these goals are particularly important because cally set at 10–12 breaths per minute, with a frac-
failure to meet them will result in increased PVR tion of inspired oxygen (FiO2) of 50% or titrated
[47], which can add afterload burden to a right to an O2 saturation level of >92%. Adaptive sup-
ventricle already burdened by an increased pre- port ventilation, a newer mode of ventilatory sup-
load from the increased cardiac output of the port, has been used successfully to shorten the
new LVAD. As discussed below and elsewhere, postoperative ventilator times of cardiac surgery
RHF is an extremely important complication to patients [48]. Our preliminary experience with
avoid in hearts with a newly implanted LVAD. using adaptive support ventilation for LVAD
Physical exams and chest x-rays are used to patients has been good; however, close attention
assess for atelectasis and pulmonary edema, both must be paid to gas exchange and the potential
of which can affect gas exchange. Both should be development of atelectasis and alveolar overdis-
aggressively treated, but we are particularly dili- tension. As with other types of cardiac surgery
gent about correcting any significant atelectasis patients, if LVAD patients demonstrate hemody-
because it is a frequent consequence of cardiac namic stability, rewarm properly, come out of
surgery and can significantly increase PVR [34, anesthesia appropriately, and meet the basic
36]. Usually, we treat atelectasis by increasing mechanical ventilation weaning criteria, we aim
either the ventilator-delivered tidal volume (Vt) to extubate the patient and remove the mechani-
or the positive end-expiratory pressure (PEEP), cal ventilation within 6–8 h of surgery. One
although sometimes therapeutic bronchoscopy is caveat to note for the extubation of LVAD patients
required in cases of airway obstruction. is that extubation may cause an increase in RV
Mechanical ventilation can have three distinct preload. That is, removal of the positive thoracic
effects on cardiac physiology through the effects pressure of mechanical ventilation may increase
of applied mean airway pressure (Paw) [34, 47]. venous return to the right ventricle, placing it at
First, increased Paw, via transmission of such risk of decompensation. Therefore, it is impor-
pressure to the pleural space, can reduce the RV tant to monitor the CVP and RV function by
preload by discouraging venous inflow into the echocardiography, along with other hemody-
higher pressure chest cavity. Second, increased namic parameters, immediately after extubation
Paw can potentially cause the alveoli to overdis- of these patients. Prolonged postoperative respi-
tend, thereby causing compression of the pulmo- ratory failure (PPRF) (the definition for which
nary capillaries and increasing the PVR. Finally, varies but usually includes a mechanical ventila-
increased Paw can decrease LV afterload via tion requirement of at least 6 days) has been
8  Perioperative Management of LVAD Patients 101

reported to occur in 9–40% of LVAD recipients tion cardiac output, and laboratory measurements
postimplantation [3, 49–53]. However, the exact (adequate and optimal values are provided).
causes of prolonged ventilation have not been Additionally, the key LVAD operational parame-
well delineated in the literature. In our experi- ters (displayed on the bedside inpatient LVAD
ence, conditions/procedures often associated monitors) and the important heart-LVAD interac-
with PPRF include poor DO2 with MSOF, bleed- tion parameters (obtained by point-of-care echo-
ing requiring reoperation, open chest, acute lung cardiography) are reviewed immediately after the
injury associated with MSOF, transfusions, and operation (expected values are provided in
sepsis. Thus far, there have been no sophisticated Table 8.3).
studies evaluating the effects of CF-LVAD The bedside inpatient LVAD monitors for the
implantation on respiratory load and respiratory HM2 and HW HVAD are connected to the device
muscle power output in humans. Patients with controllers, which operate the pumps and serve
PPRF should be cared for by following the same as the user interfaces [55, 56]. The inpatient mon-
guidelines as above, but additionally, vigilance itors are specifically designed to provide bedside
for ventilator-associated pneumonia should be clinicians with a real-time, optimized data dis-
increased, as these patients are at significant risk play from the controllers. The LVAD speed,
for this complication [3]. which refers to the revolutions per minute
(RPMs) of the device’s impellers, is the only set
operational parameter for the devices. The HM2
 emodynamic and LVAD Function
H is an axial-flow device that “pushes” blood
Considerations through the pump casing, moving it from the
inflow cannula through to the outflow cannula
As LVAD and total artificial heart technology with a turning propeller. The propeller is turned
progresses, there are likely to be significant by magnets and supported by mechanical bear-
advancements made in device auto adjustments ings. In contrast, the HW HVAD is a centrifugal-­
to respond to the fluctuating hemodynamic states flow device that “throws” blood from the pump.
in the postoperative and other periods [54]. Specifically, this device takes in blood from the
However, for now, the bedside hemodynamic inflow cannula, pushes it between the blades of a
assessments and interventions made by the care rotating disk that is housed in the pump casing
team during the immediate postoperative period and levitated with magnets and hydrodynamic
are often critical to the outcomes of these patients. forces, and then throws it tangentially out the outflow
The initial postoperative assessment of hemody- cannula. The HW HVAD acts much like a discus
namics consists of a review of the standard thrower, who releases the disk after generating
data presented in Table 8.2, including clinical energy through a spinning motion [44, 55–57].
parameters, catheter filling pressure, thermodilu- The HM2 operates at a set speed between 6000

Table 8.2  Postoperative hemodynamic parameters

Adequate Optimal
Cardiac index 2.2 L/min/m2 ≥2.5 L/min/m2
Mean arterial pressure 60–90 mmHg 70–80 mmHg
Mixed venous oxygen >50% >70%
Central venous pressure ≤15 mmHg 5–10 mmHga
Pulmonary artery occlusion pressure ≤15 mmHg 8–12 mmHg
Cardiac rhythm – Normal sinus rhythm
Lactate <4 mmol/L –
Hemoglobin – ≥10 g/dL
a
In right heart failure, central venous pressure may need to be higher for adequate preload
102 K. Ayyagari et al.

Table 8.3  Expected left ventricular assist device (LVAD) device and is not directly measured [55, 56]. The
and echocardiography parameters after device
only way to directly measure LVAD output is via
implantation
Doppler TEE determination of flow at the out-
Clinical screen of
flow cannula. This data, together with the diam-
LVAD monitor HM2 HW HVAD
eter of the outflow cannula, allows the output
LVAD speed 8000–10,000 rpm 2400–
3200 rpm from the device to be calculated [58, 59]. The
LVAD flow 2.5–7 L/min 2.5–7 L/min HW HVAD also uses hematocrit, a measure of
LVAD power 4–9 W 2.5–8.5 W blood viscosity (which is manually entered), to
Pulsatility index 3.5–5.5 – estimate flow. The default hematocrit is 30.
Pulsatility – ∆ flow Generally, the calculated flow is quite accurate
waveform >2–4 L/min for the HW HVAD over the usual range of speeds
Bedside point-of- [57, 60]. However, flow estimation is much less
care TTE HM2 and HW HVAD accurate for the HM2, which shows substantial
Interventricular Midline variability between patients [61, 62]. In the usual
septum range of flows (i.e., 2.5–6 L/min), the calculated
Left ventricular size Reduction of LVID by 20–30% flow is typically 15–20% below the actual flow
Right ventricular size Variable effects on right
rate. However, for both devices, the accuracy of
ventricular size
Right ventricular Normal RV EFXN >45%
the calculated flows cannot be assured at high or
function low LVAD speeds or when there is an obstruction
Inflow cannula Directed at mitral valve at the inflow or outflow cannula or within the
Doppler: Turbulence minimal pump [55, 56]. This is key to understanding and
or less troubleshooting the devices. Both the HM2 and
Aortic valve opening Every 2–3 beats HW HVAD have a measure of pulsatility or vari-
Aortic valve Doppler: Regurgitation ability in flow. It is important to understand that
minimal or less
flow through these devices, like that through the
Mitral valve Doppler: Regurgitation
minimal or less heart, is determined, in part, by preload and after-
Tricuspid valve Doppler: Regurgitation load, as well as by the pump speed [3, 44, 55–58,
minimal or less 63]. Or to put it another way, the flow rate partly
Pericardial effusion Minimal and no evidence of depends upon the differential pressure (or “head
right ventricular or right atrial pressure”) between the inflow and outflow can-
collapse (a sign of tamponade)
nulae. The pressure at the inflow cannula is equal
HM2 HeartMate 2, HW HVAD HeartWare HVAD, LVID
to the pressure in the left ventricle, and the pres-
left ventricular internal diameter, RV EFXN right ventricu-
lar ejection fraction, TTE transthoracic echocardiography sure at the outflow cannula is equal to the pres-
sure in the proximal aorta. The equation for
differential pressure (Diff P) is as follows:
and 15,000 RPMs, but the speed is usually set
Diff P = ( Pa − Pv ) + delta P pump,
between 8000 and 10,000 RPMs. The HW HVAD
can operate at a speed between 1800 and 4000
RPMs, but the usual setting is between 2200 and where Pa = pressure in the aorta, Pv = pressure in
3200 RPM. These speeds result in the typical the left ventricle, and delta P pump is the change
optimal blood flow of between 2.5 and 6 L/min, in pressure as blood flows through the pump (typ-
but both devices can provide up to 10 L/min of ically, this is negligible).
flow. The power input to the pump is measured Flow rate is inversely related to the differen-
and displayed in wattage and varies according to tial pressure. If the LV volume (and hence the
the pump speed and volume or flow through the pressure in the left ventricle and at the inflow
pump. For both devices, the flow displayed is an cannula) increases, then the flow though the
estimate that is calculated using an algorithm pump will increase. Similarly, if the LV volume
based on the speed and measured power of the decreases, then the flow will drop. If the LV
8  Perioperative Management of LVAD Patients 103

contractile force (and hence the pressure at the line position is consistent with an LVAD flow
inflow cannula) increases, then the flow through adequate enough to appropriately decompress the
the pump increases. If the pressure in the aorta failing, distended left ventricle but not so high as
(and hence at the outflow cannula) decreases to empty the left ventricle to the point where the
due to low vasotone, then the flow will increase. inflow cannula is at risk of coming up against the
If the aortic vasotone increases, then the flow LV wall and causing dynamic inflow obstruction.
will decrease. So, normally, although the speed Furthermore, when the septum is at midline, the
is set, flow through the device is somewhat pul- speed and flow are typically appropriate for opti-
satile, and the pulsatility is driven primarily by mal RV function. If the LVAD flow is too high,
cyclical differences in LV pressure. It is impor- the septum is pulled leftward toward the over-­
tant to note here that the HW HVAD and all the emptied and small left ventricle, causing the sep-
centrifugal-­flow pumps are, by design, more tal contribution of the RV contraction to be
sensitive to head pressure and usually demon- impaired. A leftward-shifted septum can also
strate greater pulsatility. The design of these compromise tricuspid valve geometry (annular
devices also makes them particularly sensitive dilatation and chordae tendineae tension) and
to afterload. The HM2 provides a pulsatility function (papillary muscle). If the LVAD flow is
index, whereas the HW HVAD displays a pulsa- too low, the left ventricle is inadequately emp-
tility waveform as a marker of pulsatility. The tied, thus causing the left ventricle to enlarge and
pulsatility index (PI) for the HM2 is calculated the septum to bulge rightward, similarly impair-
as follows (note that the value has no units): ing the septal contribution to RV contraction. A
leftward shift of the septum may also signal vol-
PI = ( flow max − flowmin ) / flowaverage × 10,
ume overload of the right ventricle. An overdis-
tended right ventricle essentially “pushes” the
where the flow max and flow min are the aver- septum leftward. An LVAD-driven increase in
aged peak and valley flows over a 10- to 15-s circulatory flow may cause the blood to be deliv-
interval and the flow average is the total averaged ered at a higher pressure and volume than the
flow over this interval (refer to the device manual right ventricle can pump through and into the
for the expected ranges). pulmonary circulation. This may be due, in part,
The HW HVAD displays flow pulsatility as a to a high PVR or poor RV function, as is common
continual waveform on the bedside monitor in patients with advanced heart failure before
(refer to the device manual for ranges) [55, 56]. they undergo LVAD implantation, or to incom-
Pulsatility measures are used, along with other pletely understood negative effects of LVAD
measures, to identify disturbances in preload implantation on RV function. Overdistention of
and afterload conditions that affect flow of the right ventricle can lead to a cycle of progres-
CF-LVADs. sive RHF, which is among the most common and
Standard echocardiography windows (para- feared complications of LVAD implantation.
sternal and four-chamber views) are usually ade- Overdistension can force the right ventricle to
quate for the initial bedside point-of-care operate on the descending limb of the Starling
assessment by the intensivist [58, 59, 64]. The curve, producing RV wall stress and injury and a
first and key echocardiography parameter to decrease in RV wall perfusion, creating ­conditions
assess is the position of the interventricular sep- for ischemic injury. A midline septum also typi-
tum. Fig. 8.1 demonstrates the optimal change in cally insures that the inflow cannula is appropri-
septal position after LVAD implantation (i.e., ately aligned with the mitral valve to allow for
midline between both ventricles) [58]. When the unobstructed flow into the pump.
septum is midline, the LVAD speed and flow are When determining the appropriateness of
generally within the desired ranges, and the LV LVAD flow, other factors to assess, besides septal
preload and afterload conditions are appropriate position, are the gross sizes of the left and right
for the settings [44, 57–59, 64–66]. Septal mid- ventricles, especially in relationship to the preload
104 K. Ayyagari et al.

Fig. 8.1  Right ventricular size and function. Two perpen- (RV). Ventricular interdependence between the left ven-
dicular sections of a 3D transesophageal echocardiogra- tricle (LV) and RV during systole relies on interventricular
phy reconstruction of the right ventricle from tricuspid septum position as shown in cross section for different
valve (TV) to pulmonary (PV) valve are shown. The cross clinical scenarios. Used with permission from Meineri
section (a) demonstrates the crescent shape and the sagit- et al. [58]
tal section (b) the triangular shape of the right ventricle

and afterload of the heart-LVAD system. An tion of the inflow cannula. The outflow cannula is
enlarged left ventricle may signal that the LVAD usually difficult to visualize on a standard trans-
flow is too low to decompress the left ventricle or, thoracic echocardiogram (TTE).
alternatively, it may indicate a volume overload Initial echocardiographic inspection by the
state with adequate DO2 from the LVAD flow. In intensivist should include visualization of the
both of these circumstances, the right ventricle aortic and mitral valves, both by two-dimensional
may be distended, as well, but it certainly would and Doppler imaging. Though somewhat contro-
not be expected to decrease in size. An underfilled versial, it is generally recommended that the
left ventricle may signify an LVAD speed too high LVAD speed be adjusted to allow the aortic valve
for the volume state or RHF with an inability to to open every two to three beats, if the DO2 is
deliver adequate preload to the left ventricle. In the otherwise adequate after LVAD implantation [59,
latter case, the right ventricle would be expected to 67]. When the LVAD speed is sufficiently low,
be distended. enough blood is allowed to build up in the ven-
On gross inspection, the LVAD inflow cannula tricle to be ejected through the valve, creating a
should be directed at the mitral valve to allow for parallel flow to the devices. This can help to min-
the most linear flow through the ventricle into the imize the risk of thrombus formation at or imme-
pump. Additionally, Doppler signals should show diately above the valve and to decrease the
minimal turbulence at the inflow cannula, ruling chance of valve fusion due to disuse [67]. A fused
out significant anatomic or thrombotic obstruc- aortic valve may eventually degenerate, and
8  Perioperative Management of LVAD Patients 105

under pressure from LVAD flow into the proxi- the possibility of developing cardiac tamponade
mal aorta, become incompetent, eventually lead- should become a concern. Tamponade is much
ing to aortic valve regurgitation. Left ventricular more common in patients who have undergone
assist devices may also exacerbate pre-existing LVAD implantation (occurring in 15–28% of
aortic valve regurgitation. After LVAD implanta- patients) than in those who have undergone most
tion, proximal aortic flow and pressure increase, other types of cardiac surgery. It is usually caused
typically along with a decrease in LV diastolic by postoperative bleeding into the pericardium or
pressure due to LV unloading. The change in by a mediastinal hematoma extrinsic to the peri-
pressure gradient across the aortic valve can pro- cardial space. With either pathophysiology, the
mote an increase in aortic valve regurgitation. echocardiogram almost always shows collapse of
Severe aortic valve regurgitation can lead to con- the right atrium and right ventricle. Occasionally,
tinuous recirculation of blood from the proximal the echocardiography signs of tamponade are
aorta to the left ventricle and back to the proxi- atypical in this patient population [69], and other
mal aorta again, thereby decreasing systemic signs are used to diagnose tamponade (see dis-
DO2 [59, 67]. As such, severe aortic valve regur- cussion below).
gitation after LVAD implantation may prompt the The principal concern regarding hemodynam-
need for surgical correction of the valve. ics during the postoperative period is poor DO2,
Moderate-to-severe mitral valve regurgitation which is generally defined as having a thermodi-
is common in patients with dilated cardiomyopa- lution cardiac index of <2.5 (certainly when
thy, occurring in 76% of the patients in one study <2.2) [5] or otherwise is suggested by having a
[67, 68]. Mitral valve regurgitation is a function MVO2 level of <70% (certainly when <50%) or a
of annular dilation and LV end-diastolic pressure lactic acid level of 4 mmol or greater. Our algo-
[67]. Left ventricular assist device implantation rithm for assessing poor DO2 is presented in
typically mitigates mitral valve regurgitation via Table 8.4. This algorithm follows several recently
LV unloading. Failure to do so may suggest that published guides for assessing hypotension and
the LVAD inflow cannula is interfering with the low DO2 in patients immediately after LVAD
mitral valve apparatus and that it may be neces- implantation [3, 5, 44, 58, 63, 64]. Essentially,
sary to decrease the LVAD speed or even surgi- our algorithm follows standard hemodynamic
cally intervene. evaluation protocols by looking at cardiac pre-
Tricuspid valve regurgitation is also very com- load and afterload (arterial vasotone), as well as
mon in patients with dilated cardiomyopathy “central” (cardiac equivalent in normal physiol-
(30–60% of cases, depending upon the series). ogy) LVAD output or flow. The LVAD flow is key
This is due to the elevated PAP resulting from LV to our algorithmic assessment.
failure and a dilated right ventricle [67]. Typically, A decrease in LVAD flow during the postop-
LVAD implantation would be expected to erative period and a subsequent decrease in DO2
decrease PAP via LV unloading. However, per- caused by a decrease in LV preload is usually the
sistent pulmonary vascular bed remodeling and result of bleeding. Bleeding requiring transfusion
pulmonary vasoconstriction from CPB may in this period is common, occurring in 31–81%
cause PVR to increase after the operation. This of cases. Hemodynamically significant bleeding
may contribute to RHF during the postoperative and other significant causes of hypovolemia are
period. Significant tricuspid valve regurgitation characterized by hypotension and low PCWP and
should prompt at least pharmacologic treatment CVP. Echocardiogram findings of this complica-
of pulmonary hypertension and possibly surgical tion include diameter decrease or collapsibility of
correction of the valve. the inferior vena cava and possibly a leftward
Finally, after LVAD implantation, echocar- shift in the cardiac intraventricular septum, as the
diography should be performed to ensure that LVAD continues to unload the LV. However, the
there is only a minimal amount of pericardial septum may maintain an appropriate midline
fluid. If pericardial effusion is significant, then position in this circumstance. In addition, the
Table 8.4  Decreased oxygen delivery: assessment and treatment
106

↓CO, ↓MVO2, ↑LA

Low LVAD flow High
Volume depletion Severe aortic regurgitation
Hemodynamics Differential diagnosis: Hemodynamics Differential diagnosis:
 MAP ↓ − Hypovolemia  MAP ↓ − Severe AR
 PCWP ↓ − Bleeding  PCWP ↑ Treatment:
Treatment: Device
 CVP ↓ Volume  CVP ↑ or no change − Trial ↓ speed
Echo findings −  RBCs for Hgb ≥10 Echo findings Surgical
 IVC ↓ − Bolus 250 mL 5% albumin and  IVC ↑ or no change − Aortic valve repair
 LV ↓ reassess  LV ↑ or no change Medications
Surgical − Nicardipine
 RV ↓  RV ↑ or no change
− Possible surgery for bleeding − Nitroprusside
 Septum Midline or left shifted  Septum Right shifted
 LVAD LVAD
 PI ↓  PI ↑
 PP ↓  PP ↑
RHF or PE Vasodilation
Hemodynamics Differential diagnosis: Hemodynamics Differential diagnosis:
 MAP ↓ − PHTN (SPAP  ≥ 40)  MAP ↓ − Cardiac vasoplegia
− RHF − Sepsis
 PCWP ↓  PCWP ↓ or no change
− TR − Systemic inflammatory
 CVP ↑ − PE (massive or submassive)   CVP ↓ or no change response syndrome
Echo findings Treatment (for RHF/PHTN): Echo findings Treatment:
 IVC ↑ Device  IVC ↓ or no change Volume
 LV ↓ − Trial ↓ speed  LV ↓ or no change − 250 mL 5% albumin or
Pulmonary vasodilators crystalloid and reassess
 RV ↑  RV ↓ or no change
− NO/inhaled epoprostenol Medications
 Septum Left shifted − Sildenafil  Septum No change or left shifted − Vasopressin
 LVAD − Milrinone  LVAD − Levophed
 PI ↓ Inotropes for RHF  PI ↓ Suspected sepsis
 PP ↓ − Epinephrine  PP ↑ − Culture and use empiric
− Dopamine broad-­spectrum antibiotics
− Dobutamine
Pulmonary embolism
− AC/possible catheter-directed therapy
Surgical
K. Ayyagari et al.

− Possible surgery for TR

 ↓CO, ↓MVO2, ↑LA
Low LVAD flow High
Cardiac Tamponade Pump dysfunction (flow erroneously high)
Hemodynamics Differential diagnosis: Hemodynamics Differential diagnosis:
 MAP ↓ −  Pericardial effusion  MAP ↓ − Rotor thrombosis
 PCWP ↑* − Mediastinal hematoma  PCWP ↑ − Mechanical failure of
− Tension pneumothorax  rotor
 CVP ↑  CVP ↑
Treatment: Treatment:
Echo findings Surgical Echo findings Medication
 IVC ↑ −  Pericardial window  IVC ↑ − AC or thrombolytics
 LV ↓ − Hematoma evacuation  LV ↑ Surgical
 RV Collapsed Tension PTX treatment  RV ↑ − Pump replacement
−  Tube thoracostamy
 RA Collapsed  RA ↑
*Note: PCWP may be ↓ early due to LV
 Septum N/A unloading, ↑ PCWP is a late finding  Septum Right shifted
8  Perioperative Management of LVAD Patients

LVAD LVAD
 PI ↓  PI ↓
 PP ↓  PP ↑
Increased LV afterload
Hemodynamics Differential diagnosis:
 MAP ↑ − Hypertension
 PCWP ↑ − Vasopressors 
Treatment:
 CVP ↑ or no change
− Stop or reduce vasopressors and/or
Echo findings start nicardipine or nitroprusside
 IVC ↑ or no change
 LV ↑ or no change
 RV ↑ or no change
 Septum No change or right
shifted
LVAD
 PI ↓ or no change
 PP ↓ or no change
(continued)
107
Table 8.4 (continued)
108

↓CO, ↓MVO2, ↑LA

Low LVAD flow High
Decreased pump function
Hemodynamics Differential diagnosis:
 MAP ↓ − Speed settings too low
 PCWP ↑ −  Suction event
− Inflow/outflow obstruction
 CVP ↑ or no change
−  Pump obstruction
Echo findings     Pump thrombosis
 IVC ↑ or no change     Pump malfunction
 LV ↑ or no change − Arrhythmia 
 RV ↑ or no change Treatment:
Device
 Septum Variable
− Trial ↑ speed
LVAD Suction event
 PI Variable −  ↓ pump speed and give volume if
 PP Variable indicated
Flow obstruction
−  Advanced echo
Medication
−  ↑ AC or thrombolytics for thrombosis
Surgical
− Possible surgical correction of
obstruction or pump change
Arrhythmia control
− Medications or ablation techniques
AC anticoagulation, AR aortic valve regurgitation, CO cardiac output, CVP central venous pressure, echo echocardio­graphy, Hgb hemoglobin, HTN hypertension, IVC inferior
vena cava, LA lactic acid, LV left ventricle, LVAD left ventricular assist device, MAP mean arterial pressure, MVO mixed venous oxygen saturation, NO nitric oxide, PCWP
pulmonary capillary wedge pressure, PE pulmonary embolism, PHTN pulmonary hypertension, PI pulsatility index, PP pump power, PTX pneumothorax, RA right atrium, RHF
right heart failure, RBCs red blood cells, RV right ventricle, septum intraventricular septum, TR tricupid valve regurgitation
K. Ayyagari et al.
8  Perioperative Management of LVAD Patients 109

chamber sizes of both the left and right ventricle emergently placed, and is a major driver of poor
typically decrease. As with any hypovolemic outcomes for LVAD patients [70]. Chapter 18
state, the LVAD flow pulsatility would decrease, specifically deals with preoperative assessment
reflecting an underfilled LV and, hence, low LV for postoperative RHF risk and risk remediation
intraventricular pressures. But in contrast to other strategies. However, it should be noted that RHF
causes of hypovolemia, bleeding would be prediction models are still not highly sensitive or
expected to cause a decrease in the hemoglobin specific [58, 65, 66]. Additionally, therapy for
level, although this might not be the case with RHF has evolved to the point where there are
brisk bleeding. Chest tube output (from draining well established and even more effective treat-
the mediastinum and, usually, the pleural cavi- ment protocols involving volume management,
ties) is often the earliest and clearest indication of device adjustment, medication, and MCS for the
significant bleeding. Chest tube output rates right ventricle. In fact, device support for the fail-
>100 mL/h could indicate a significant bleed. ing right ventricle is one of the most active areas
Interventions for significant bleeding include of MCS research and development. As such, high
reversal of platelet dysfunction and coagulopathy vigilance to the development of RHF is war-
(see “Additional Hemostasis Considerations” ranted in the immediate postoperative period.
below), volume resuscitation with red blood cells Diagnostic indicators of RHF include a low DO2,
(leukoreduced to limit allosensitization in hypotension, low LVAD flow, a low PCWP, and a
­consideration of the future potential for cardiac high CVP. Echocardiography findings consistent
transplantation) to achieve a hemoglobin level of with RHF are a dilated inferior vena cava, a
>10 g/dL, and other volume-support therapies leftward-­shifted intraventricular septum, a small
(5% albumin is favored) if red blood cells are not left ventricle, and an enlarged right ventricle.
immediately available. It is critically important Echocardiography measures of RV function may
to continually assess the effects of the resuscita- show impaired contractility. Additionally, sub-
tion by monitoring clinical parameters, labora- stantial tricuspid valve regurgitation may be seen
tory measurements, filling pressures, by Doppler echocardiography. Because the left
echocardiography findings, and LVAD flow ventricle is typically underfilled in this circum-
parameters, not only to determine the effect on stance, device flow pulsatility can also be
DO2 but also to avoid over-resuscitation, specifi- decreased.
cally RV overload. Such RV overload can lead to Initial therapeutic maneuvers for RV decom-
RHF, a key complication of perioperative bleed- pensation include the following [3, 5, 44, 58, 63,
ing [3, 44, 63]. Right ventricular failure leads to a 65, 66]:
decrease in DO2 and is one of the primary causes
of poor outcomes in these patients. The surgical –– Promote RV volume unloading (typically if
team should obviously be apprised of significant the CVP is >15 mmHg) by using diuresis or
bleeding and ultimately may need to take the dialysis techniques (usually continuous renal
patient back to the operating room if the bleeding replacement therapy) and/or by decreasing the
is refractory to coagulopathy correction or is of overall circulatory flow and volume delivery
such volume or duration to suggest the need for to the right ventricle by decreasing the device
surgical intervention. speed.
Other causes of hypovolemia during the post- –– Promote RV afterload unloading by admin-
operative period include over-diuresis and exces- istering pulmonary vasodilators, typically
sive volume removal by dialysis. These conditions inhaled nitric oxide and epoprostenol, as
are also treated with a volume expander, usually well as sildenafil. There are multiple new
crystalloid, with the caveats delineated above. drugs available for treating pulmonary
Right ventricular failure during the postopera- hypertension, including drugs in the prosta-
tive period is common, occurring in 10–39% of cyclin, endothelin antagonist, phosphodies-
patients and in even up to 50% when the VAD is terase inhibitor, and guanylate cyclase
110 K. Ayyagari et al.

stimulator classes. These have not yet been tended inferior vena cava, and the CVP is elevated.
systematically studied in this population. Often, the PCWP is initially low, and the LV size
–– Provide RV inotropic support with milrinone is decreased due to ongoing LVAD unloading of a
(which is also capable of vasodilating the pul- poorly filled left ventricle. However, elevations in
monary vasculature) and catecholamines, PCWP and pulmonary artery diastolic pressure are
including epinephrine, dobutamine, and usually late findings that are ominous for circula-
dopamine. tory collapse [5, 69]. Pulsatility of the LVAD is
–– If volume and medical treatments for RHF are also decreased in patients with tamponade due to
unsuccessful, MCS of the right ventricle the decreased LV filling. Treatment of cardiac
maybe be used for specific indications. The tamponade almost always involves a return to the
evolving selection of MCS devices for this operating room to drain the pericardium or the
purpose and the specific indications for use mediastinal hematoma causing it, as well as to
are discussed in Chap. 18 “Management of look for bleeding sites in need of surgical treat-
RV Failure After LVAD Implantation.” ment. Occasionally, the tamponade may be
relieved by percutaneous drainage.
Massive or submassive pulmonary embolism Tension pneumothorax (PTX) physiology is
can mimic RHF in LVAD recipients. Fortunately, very similar to that of cardiac tamponade and has
it is an uncommon occurrence in these patients. similar effects on the hemodynamic profile of
Signs indicative of a pulmonary embolism LVAD recipients. A key diagnostic indicator for
include an echocardiogram showing a right- PTX is an abrupt elevation of mechanical venti-
sided thrombus and failure of typical maneuvers lator pressures. Clinical signs that may also sug-
to improve RHF. Pulmonary embolisms are gest the presence of PTX include diminished
diagnosed by finding a thrombus in the pulmo- hemithorax breath sounds and tracheal deviation.
nary arteries using imaging techniques, usually Although radiology remains the usual means for
computerized tomography (CT) angiography but diagnosing PTX, thoracic ultrasound has a simi-
occasionally TEE or even TTE. The base therapy lar sensitivity and specificity, and point-of-care
for these patients is anticoagulation, of course, ultrasound testing is typically faster. Signs to
but for patients who do not respond to standard look for include “lung pointing” and the “strato-
anticoagulation protocols or who need emergent spheric sign” and the absence of the “sliding
clearance of a hemodynamically threatening lung” sign and “B-lines.” The treatment for PTX
thrombus, we use catheter-based interventions to is immediate tube thoracostomy [71].
administer local lytic drug therapy, as well as Vasodilation during the immediate postopera-
thrombus fragmentation and suction removal tive period can result in poor DO2 due to
techniques. We do not use systemic lytic therapy decreased filling volumes. Vasodilation can also
for LVAD recipients because of the high risk of promote an increase in DO2 in LVAD recipients
causing major bleeding in the immediate postim- by decreasing afterload. Independent of the DO2
plantation period. Rarely, patients may need to effects, vasodilation can cause poor tissue perfu-
undergo surgical thrombectomy. sion by inducing maldistribution of blood flow
As previously mentioned, cardiac tamponade is (distributive shock) to the tissues. In LVAD
not an uncommon postoperative complication that recipients (as in all other patients), the net result
results from bleeding after LVAD implantation. of vasodilation on DO2 to the tissues depends
Tamponade physiology causes a decrease in DO2 upon the net balance of these effects. The diagno-
and blood pressure and, if not addressed, can lead sis of vasodilation can be suggested by a physical
to cardiovascular collapse. It is primarily diag- exam showing hypotension, a widened pulse
nosed by using echocardiography to show collapse pressure, or warm or overperfused skin, but the
of the right atrium and ventricle, often cyclically diagnosis is usually confirmed by a decreased
with the positive pressure ventilator cycle. systemic vascular resistance. The cause of vaso-
Echocardiography also typically shows a dis- dilation during the immediate postoperative
8  Perioperative Management of LVAD Patients 111

period is almost always cardiac vasoplegia (dis- There are many reasons why pump function
cussed in detail above). This is typically treated may decrease, including controller dysfunction
with vasopressors, including vasopressin and the (either intrinsic or due to poor electrical connec-
catecholamines norepinephrine and phenyleph- tions to the power source or device), obstruction
rine. Rarely, methylene blue and steroids are used of the inflow or outflow of the pump, thrombosis
for refractory cases. Other causes of vasodilation in the pump rotors, mechanical failure of the
are systemic inflammatory response syndrome pump impeller or bearings, significant arrhyth-
from the operation itself and sepsis. For LVAD mias, or cardiac arrest. A decrease in pump func-
recipients who develop sepsis during the imme- tion will, in turn, decrease pump flow and can
diate postoperative period, the sources of infec- markedly decrease DO2. In patients who experi-
tion are potentially the same as those of other ence such events, the hemodynamic profile is
patients who undergo similar procedures, such as similar: the systemic blood pressure is usually,
those on ventilator support (e.g., pneumonia), though not always, decreased, but the filling
those who have undergone device implantation pressures, thermodilution cardiac output, and
(e.g., line infection and catheter-related urinary echocardiographic parameters look like those of
tract infection), or those who have undergone patients with CHF. That is to say that the filling
surgery in general (e.g., wound infection). De pressures increase, cardiac output decreases, and
novo ­ device-­related infections can occur, but the left ventricle (and usually the right ventricle)
these usually take at least a couple of days to dilates. Additionally, the pulsatility usually
develop to the point of sepsis. However, in our decreases with the overall drop in flow.
experience, patients who undergo LVAD For both the HM2 and HW HVAD, the con-
exchange because of device-related infection troller display shows alarms for battery failure, a
often manifest sepsis syndrome during the imme- poor connection to the electrical outlet, or
diate postoperative period. “controller failure.” These alarms respectively
An elevated LV afterload (hypertension) can call for battery replacement, correction of the
decrease LVAD flow and DO2, particularly in faulty electrical connection, or controller replace-
patients who have received a HW HVAD, which ment. If the controller ceases to display data, it
is more sensitive to the differential pressure could be completely powerless (due to a battery
across the pump [45, 57]. In LVAD recipients, and/or electrical connection failure) or it may not
elevation of afterload is usually due to either the be working properly, in which case it should be
inotropes or the vasopressors typically adminis- replaced with a backup controller, which should
tered postimplantation to provide RV support and always be immediately available. For both
prevent postoperative vasoplegia, although it can devices, the LVAD controller can be connected to
also result from intrinsic hypertension. The fill- a monitor that displays not only the alarms men-
ing pressures and echocardiography profiles of tioned above but also pump flow, speed, power,
these patients are similar to those of patients with and either pulsatility index (HM2) or a flow-time
CHF. Left ventricular assist device pulsatility curve (HW HVAD), which serves as a “pulsatil-
may decrease if the afterload is high enough to ity wave.” These parameters also have alarm
substantially decrease the pressure differential functions that can be set for high or low values.
between the left ventricle and the proximal aorta. Table 8.5 displays changes in these measures that
It is important to note that having a high pump can be used to help sort out pump-related mal-
speed and flow can cause hypertension. So when functions or physiologic derangement.
the blood pressure and LVAD flow are elevated, Obstruction of the LVAD pump inflow can be
one should consider decreasing the speed and due to malpositioning of the inflow cannula,
flow. Having a mean arterial pressure above 90 allowing anatomic structures of the left ventricle,
has been associated with cerebral vascular acci- such as the septum, to block flow. Additionally,
dents (CVAs) in this population [72]. intraventricular pathologies, such as ruptured
112 K. Ayyagari et al.

Table 8.5  LVAD troubleshooting

Monitored parameters
Speed Power Flow
A constant speed should ↓ Power ↓ Flow
always be maintained by the  Inflow/outflow obstruction  Decrease in preload
controller unless there is a  Suction event  Increase in afterload
drop in power, most  Controller malfunction  Decrease in speed
commonly from a suction  Power connection problem  Inflow/outflow obstruction
event. Speed will decrease  Decrease in   preload ↑ Flow
during a suction event. A  Increase in afterload  Increase in preload
suction event can be caused ↑ Power  Decrease in afterload
by (1) the pump speed being  Thrombus/mechanical impairment of rotor  Increase in speed
too high, (2) a decrease in  Increase in preload  Mechanical impediment due to
preload, or (3) poor inflow  Increase in inotropy thrombus on rotor (erroneous high)
cannula positioning  Increase in pump speed
 Decrease in afterload

chordae tendineae, can block flow to the inflow anticoagulation, glycoprotein IIb/IIIa inhibitors,
cannula. Obstruction of the inflow cannula can be thrombolysis (significantly risky), catheter-based
diagnosed with echocardiography (Table 8.6), CT interventions, and surgical interventions (up to
angiography, or left heart catheterization with left and including LVAD replacement) [74, 75]. Of
ventriculogram. Obstruction of the graft outflow note, the International Society of Heart and Lung
may be caused by graft kinking or twisting or by Transplantation (ISHLT) has developed a clinical
thrombosis. Echocardiography can be useful for algorithm for treating LVAD thrombosis (as well
diagnosing graft outflow obstruction (Table 8.6), as for making the diagnosis) [75].
but we have found CT angiography to be more
helpful. It is more difficult to diagnose an obstruc-
tion within the pump itself or mechanical failure Pump Stop
of the rotor because our imaging modalities
(echocardiography and CT angiography) do not A complete pump stop is an immediately life-­
allow us to see inside the metal encased pumps. threatening event. An alarm will probably go
However, certain echocardiography findings can off on the LVAD controller or monitor, but one
still be suggestive of these problems, especially may not if the controller is completely dysfunc-
when the imaging is done in conjunction with tional or disconnected from a power source.
LVAD speed changes (Table 8.6). Specifically, if Clinically, a pump stop can be determined by
increasing the LVAD speed does not result in a auscultating over the device. If there is no
decrease in LV volume, then there may be an mechanical hum, the device has stopped. As
obstruction in the LVAD pump [59]. Occasionally, mentioned above, when there is a pump stop,
internal pump obstruction can be diagnosed with the first assessment/intervention should be to
the help of a left ventriculogram showing inflow examine the electrical connections of the drive-
to the pump and no outflow from the pump [73]. line and power line to the controller. An emer-
Specifics regarding the interventions for pump gent trial of controller change should be
inflow, outflow, and internal obstructions are considered. If the connections are intact, a new
beyond the scope of this chapter. These interven- controller trial has failed, and central cardiac
tions require input from cardiologists and sur- output has ceased to be adequate (i.e., DO2 and
geons, as well as an individualized strategy. blood pressure are severely decreased or
However, depending upon the specific cause of absent), then cardiopulmonary resuscitation
the obstruction, the basic treatment options are (CPR) should be begun.
8  Perioperative Management of LVAD Patients 113

Table 8.6  Continuous-flow LVAD postimplant compli- Table 8.6 (continued)

cations and device dysfunction detected by
a. 2D/3D: small or crowded inflow zone with or
echocardiography
without evidence of localized obstructive muscle
Pericardial effusion trabeculation, adjacent MV apparatus or thrombus;
With or without cardiac tamponade including RV malpositioned inflow cannula
compression. Tamponade: respirophasic flow changes; b. High-velocity color or spectral Doppler at inflow
poor RVOT SV orifice. Results from malposition, suction event/other
LV failure secondary to partial LV unloading inflow obstruction: aliased color-flow Doppler, CW
Doppler velocity > 1.5 m/s
(by serial exam comparison) c. Low-velocity inflow (markedly reduced peak systolic
a. 2D/3D: increasing LV size by linear or volume and nadir diastolic velocities) may indicate internal
measurements; increased AV opening duration, inflow cannula thrombosis or more distal obstruction
increased left atrial volume within the system. Doppler flow velocity profile may
b. Doppler: increased mitral inflow peak E-wave appear relatively “continuous” (decreased phasic/
diastolic velocity, increased E/A and E/e′ ratio, pulsatile pattern)
decreased deceleration time of mitral E velocity,
worsening functional MR, and elevated pulmonary Outflow graft abnormality
artery systolic pressure Typically due to obstruction/pump cessation
RV failure a. 2D/3D imaging: visible kink or thrombus
(infrequently seen)
a. 2D: increased RV size, decreased RV systolic b. Doppler: peak outflow graft velocity ≥ 2 m/sa if near
function, high RAP (dilated IVC/leftward atrial obstruction site; however, diminished or absent spectral
septal shift), leftward deviation of ventricular septum Doppler signal if sample volume is remote from
b. D oppler: increased TR severity, reduced RVOT SV, obstruction location, combined with lack of RVOT SV
reduced LVAD inflow cannula and/or outflow graft change and/or expected LV-dimension change with
velocities (i.e., <0.5 m/s with severe failure); inflow pump speed changes
cannula high velocities if associated with a suction
event. Note: a “too-high” LVAD pump speed may Hypertensive emergency
contribute to RV failure by increasing TR (septal New reduced/minimal AV opening relative to baseline
shift) and/or by increasing RV preload exam at normal BP, especially if associated with new/
Inadequate LV filling or excessive LV unloading worsened LV dilatation and worsening MR. Note:
hypertension may follow an increase in pump speed
Small LV dimensions (typically <3 cm and/or marked
deviation of interventricular septum toward LV). Note: Pump malfunction/pump arrest:
May be due to RV failure and/or pump speed too high a. Reduced inflow cannula or outflow graft flow
for loading conditions velocities on color and spectral Doppler or, with pump
LVAD suction with induced ventricular ectopy arrest, shows diastolic flow reversal.
b. Signs of worsening HF: including dilated LV,
Underfilled LV and mechanical impact of inflow worsening MR, worsened TR, and/or increased TR
cannula with LV endocardium, typically septum, velocity; attenuated speed change responses: decrease
resolves with speed turndown or absence of expected changes in LV linear dimension,
LVAD-related continuous aortic insufficiency AV opening duration, and RVOT SV with increased or
Clinically significant—at least moderate and possibly decreased pump speeds; for HVAD, loss of inflow
severe—characterized by an AR proximal jet-to-LVOT cannula Doppler artifact
height ratio > 46% or AR vena contracta ≥3 mm; 2D, two-dimensional; 3D, three-dimensional; A, mitral
increased LV size and relatively decreased RVOT SV valve late peak diastolic velocity; AR, aortic regurgitation;
despite normal/increased inflow cannula and/or outflow AV, aortic valve; BP, blood pressure; CW, continuous wave;
graft flows E, mitral valve early peak diastolic velocity; e′, mitral annu-
LVAD-related mitral regurgitation lar velocity; HVAD, HeartWare ventricular assist device;
a. Primary: inflow cannula interference with mitral IVC, inferior vena cava; LV, left ventricular; LVAD, left ven-
apparatus tricular assist device; LVOT, left ventricular outflow tract;
b. S econdary: MR functional, related to partial LV MR, mitral regurgitation; MV, mitral valve; RAP, right atrial
unloading/persistent heart failure pressure; RV, right ventricular; RVOT, right ventricular out-
Note: Elements of both a and b may be present flow tract; SV, stroke volume; TR, tricuspid regurgitation
Intracardiac thrombus
a
Note: based on observational data. The “normal” outflow
graft peak velocities are not well defined. Because the
Including right and left atrial, LV apical, and aortic root
HVAD outflow graft diameter is smaller than that of the
thrombus
HM II device (see discussion in text). Therefore, the normal
Inflow cannula abnormality Doppler-derived HVAD outflow velocities may be some-
(continued) what higher on average than those observed for the HM II
LVAD. Used with permission from Stainback et al. [59]
114 K. Ayyagari et al.

Cardiac arrest and drugs (particularly catecholamines). We

manage atrial fibrillation by addressing the
There is debate as to whether CPR, administered potential underlying causes and by using proto-
by compressing the chest, is safe and effective in cols provided in the guidelines recently published
patients with an LVAD. If the device is not by multiple cardiac societies [83]. However, the
obstructed, patients may benefit from CPR aggressiveness of therapy is largely determined
because it may promote flow through the device. by the effects on device output and the strength of
Even if the device is obstructed, CPR may pro- evidence suggesting that it is promoting RHF. The
mote flow through the aortic valve, unless, of first pharmacologic agents used to control heart
course, the valve has been sewn closed. Although rate (target rate, 80–100 beats per minute) are
there would seem to be a risk of cannulae dis- beta blockers (typically esmolol and metoprolol).
lodgment and damage to the LVAD outflow con- Calcium channel blockers are avoided because of
duit during CPR, a recent retrospective analysis their potential negative inotropic effects. In fact,
of LVAD patients who received chest compres- careful monitoring for negative inotropic effects
sions for cardiac arrest did not support the theory is warranted even when using beta blockers. In
that LVADs would be harmed by conventional our practice, the typical second-line agent used to
resuscitation algorithms [76]. There have been control heart rate is amiodarone. Rapid atrial
cases published about alternative means of exter- arrhythmias that result in shock or acute RHF are
nal CPR, like abdominal-only compressions. treated with cardioversion. Amiodarone is also
However, no studies have compared these alter- usually our first-line drug for pharmacologic car-
native means of external CPR to conventional dioversion of atrial fibrillation or for keeping a
CPR. Further research is required to address both patient in sinus rhythm once electrically con-
the safety and efficacy of chest compressions in verted. Atrioventricular node ablation and other
this population. ablation techniques may be needed for refractory
arrhythmias [82].
Preoperative predictors of postoperative ven-
Arrhythmias tricular arrhythmias for patients undergoing
LVAD implantation are a history of such arrhyth-
Arrhythmias are common in the postoperative mias, nonischemic cardiomyopathy, and
period, especially on day 1, occurring in 30–60% increased age [78]. Interestingly, the preopera-
of patients [49, 51, 77–81]. The common arrhyth- tive INTERMACS profile has not been found to
mias are atrial fibrillation, atrial flutter, ventricu- be predictive of postoperative ventricular
lar tachycardia, and ventricular fibrillation. Atrial arrhythmias. A recent multi-institutional study
fibrillation and flutter can affect device output, identified the following specific causes of post-
predominantly by affecting RV filling and, con- operative ventricular arrhythmias in LVAD
sequently, LV preload, especially in patients with patients and their relative frequencies: inotropic
RHF. Atrial arrhythmias also increase the risk of agents (43%), suction events (whereby the sep-
thromboembolic events in this population [82]. tum is pulled up against the inflow cannula)
The specific effect of atrial fibrillation or loss of (10%), electrolyte abnormalities (4%), ischemia
sinus rhythm on RV function has not been well (1%), and no specific cause (42%) [78]. Our
studied in this population. Many patients who experience has shown that an acute decrease in
undergo LVAD implantation have a history of blood volume (usually due to bleeding, inade-
atrial tachycardias. The causes of postoperative quate unloading of the left ventricle, or RHF)
atrial fibrillation/flutter in LVAD recipients are and device malpositioning can also precipitate
the same as those for patients without a VAD: ventricular arrhythmias. However, for a large
elevated atrial pressure from inadequate LV percentage of patients, a specific cause is not
unloading or RHF, electrolyte disturbance (par- readily identifiable. Notably, and in regards to
ticularly potassium, magnesium, and acid-base this, there is evidence that LVAD implants can be
disturbance), hypoxemia, myocardial ischemia, arrhythmogenic themselves by directly creating
8  Perioperative Management of LVAD Patients 115

new areas of myocardial scarring around the ative period. Such frequent firing can lead to
inflow cannula insertion site and by altering RHF [78, 79]. The ICD may eventually be turned
myocardial ion channels. Conversely, LVADs back on, but reconsideration of the shock thresh-
can also decrease arrhythmogenicity (via LV old is recommended. Cardiac resynchronization
unloading) in some patients [79]. therapy (CRT) has been shown to improve out-
Ventricular fibrillation is an immediate threat comes in select advanced heart failure patients.
to life in patients without an LVAD. However, At this time, few studies have assessed the effects
patients with an LVAD will often tolerate this of CRT on LVAD implantation outcomes.
rhythm acutely since device output and systemic However, one recent study suggests that CRT, in
DO2, for the most part, do not depend upon LV situ, at the time of LVAD implantation improves
contraction. If the LVAD receives adequate pre- ventricular arrhythmias and decreases ICD
load from a heart in fibrillation, essentially acting shocks in these patients [81].
as a conduit for blood flow, it can deliver ade-
quate output. However, the fibrillation may
markedly decrease RV output (especially in
­  roubleshooting LVAD Malfunction
T
patients with RHF) and, consequently, LV pre- by Using Echocardiography
load, which can lead to shock. Regardless of the
immediate effects, ventricular fibrillation and Table 8.6, from the recently published guidelines
ventricular tachycardia can ultimately result in on the echocardiographic assessment of LVADs,
RHF, although the exact mechanism for this is shows the problems associated with LVAD func-
not clear and is likely multifactorial [78, 79]. Our tion and heart-LVAD interactions that can be
approach to treating ventricular fibrillation and found with echocardiography [59].
ventricular tachycardia includes an assessment
for remedial causes. But, as with atrial fibrillation
treatment, the aggressiveness of treatment for  leeding and Hemostasis
B
ventricular fibrillation and tachycardia depends Considerations
on the effect of the rhythm on LVAD output and
RV function. For acute, severe effects, we per- Management of perioperative hemotherapy in
form emergent external cardioversion. When patients requiring LVAD support is challenging
ventricular arrhythmias are not emergently for several reasons, including preoperative organ
threatening, we use beta blockers and amioda- (hepatic and/or renal) dysfunction, the complex
rone as our primary pharmacologic interventions. nature of the operative procedure, preoperative
Second-line agents include mexiletine, sotalol, antithrombotic exposure, and postoperative anti-
and lidocaine. Ventricular arrhythmias refractory coagulation requirement in a setting of high
to treatment should prompt concern of possible bleeding risk [84, 85]. Registry data show that
device malpositioning and should be assessed the majority of LVAD patients require transfu-
with echocardiography and, perhaps, CT imag- sion support, often characterized as large-volume
ing. Electrophysiologic ablation may be required transfusions [86, 87].
for ventricular arrhythmias not corrected with
pharmacology or device repositioning. One
important point of consensus to emerge from the Preoperative Management
literature recently is that, although implantable
cardioverter defibrillators (ICDs) implanted pre- As a component of the preoperative evaluation,
operatively may ultimately be beneficial for patients scheduled for LVAD implantation should
select LVAD recipients, it is best to turn them off be screened for factors associated with increased
during the perioperative period. Those implanted bleeding risk. A careful assessment of medical
preoperatively may unnecessarily fire quite fre- history should include previous spontaneous
quently during the dynamic immediate postoper- bleeding episodes, response to previous surgical
116 K. Ayyagari et al.

challenges, family history, and exposure to anti- reversing its effect. A second dose will be needed
thrombotic drugs. It is important to sort routine in most instances. For emergent correction of the
patients from those requiring in-depth input from effects of warfarin, treatment with prothrombin
a hematology consultant. Antiplatelet therapy complex concentrates is preferred, the dosing for
with aspirin is routine for patients with ischemic which is weight based and adjusted for INR cat-
heart disease and is typically maintained due to egory [93]. These agents promptly reduce the
the associated benefits and the relatively low INR while minimizing the volume challenge
additional risk of perioperative bleeding [88]. In associated with plasma-based correction. Plasma-­
all but urgent cases, when the patient is on more based correction is slower and has been associ-
potent antiplatelet agents (e.g., P2Y12 ADP ated with an increased incidence of pulmonary
receptor inhibitors), those agents should be with- reactions [94].
held for the interval of the inhibitor effects (at If a patient needs to undergo preoperative anti-
least 5 days in the case of P2Y12 inhibitors), or coagulation with a heparin analogue, unfraction-
the patient should undergo an evaluation of plate- ated heparin is the easiest agent to manage. Given
let function to determine if it is acceptable [89]. the short half-life of unfractionated heparin
In many cases, routine laboratory testing is not (1–2 h), it can be withheld for a short interval
recommended as a part of the preoperative assess- before surgery to achieve adequate operative
ment when the patient’s medical history is not hemostasis. Managing low-molecular-weight
suggestive of a bleeding diathesis [86]. The inter- heparins (LMWH) is more challenging in the
national normalized ratio (INR) is a test designed perioperative interval because of the longer half-­
primarily to monitor warfarin-based anticoagula- life of these agents and the potential for impaired
tion. Warfarin is typically dosed to achieve an renal clearance, which would further prolong the
INR value of 2–3. The INR has a nonlinear anticoagulant effects of the LMWH. Patients
response to clotting factor activity, but it can be managed with a LMWH should have the agent
used to determine when clotting factor activity withheld for at least 12 h before an operation if a
reaches the level required for hemostasis, which prophylactic dose is being used. If a therapeutic
is approximately 30%. It is important to know the dose (1 mg/kg) of LMWH is being used, the dose
response of the INR to clotting factor activity in administered within 12 h of an operation should
each facility due to the variation in response of be 50% of the standard dose to avoid the risk of
different reagents: instrument pairs. Although the bleeding [95]. Given the renal clearance mecha-
INR has not been shown to be a general predictor nisms and possibility of decreased glomerular
of perioperative bleeding in the preoperative eval- filtration rate in candidates for VAD implanta-
uation of broad patient populations, it is used to tion, we prefer to use unfractionated heparin in
determine reversal of warfarin effect and to assess the immediate preoperative period.
the degree of hepatic derangement in patients Preoperative transfusion to bolster hemostatic
with hepatic dysfunction [90, 91]. potential is not typically used in mild coagulo-
Patients who receive the anticoagulant warfa- pathic states. Clinical staff should understand the
rin before an operation should be treated with relationships between coagulation parameters and
vitamin K to reverse the effects. The most effec- hemostatic factor activity. For example, an INR
tive way to administer vitamin K is intravenously value of 1.6 may be consistent with adequate pro-
[92]. Slow infusion should be used to avoid the coagulant clotting factor activity (40–50%),
risk of adverse reactions. Correction of the INR depending on the reagent being used. A bleeding
commences 6–8 h after intravenous therapy is patient or a patient with severe congenital or
begun. If there is ample time for correction (i.e., acquired coagulopathy will need to be treated
>24 h), oral vitamin K may be used. The half-life before the operation to prepare for the hemostatic
of warfarin (40 h) should be considered when challenge.
8  Perioperative Management of LVAD Patients 117

Operative Management available test of fibrinolysis at this time. We treat

all patients undergoing CPB with antifibrinolytic
The operating room suite should have access to agents. Therefore, we truncate our VE panel
point-of-care or near-site coagulation laboratory because the therapy will not be affected.
testing to facilitate data-driven transfusion ther- Traditional coagulation laboratory measure-
apy. The use of these systems has been associ- ments and VE results may be used jointly or
ated with reduced blood transfusions and separately. We use both traditional point-of-care
improved patient outcomes [96, 97]. Important coagulation tests and VE assay measures in our
measures include prothrombin ratio/INR, fibrin- center (algorithm in Fig. 8.2). Although a PTT
ogen, platelet count, and whole blood viscoelas- test is useful for detecting an individual clotting
tic (VE) test parameters (e.g., Thromboelastograph factor deficiency (e.g., factor VIII, factor IX, or
[TEG®] R, TEG alpha, TEG MA, and TEG factor XI), it is of limited value in the periopera-
EPL). The TEG R value represents the time until tive setting for patients without a congenital
initiation of clotting and is analogous to INR and factor deficiency. We use the difference in
partial thromboplastin time (PTT) data, although VE-based TEG R data (kaolin versus heparinase
the correlation is not high [98]. The TEG alpha cups) to check for heparin effects after the ini-
value reflects the reaction rate and is attributed to tial anesthesia-­managed point-of-care activated
coagulation protein activity and fibrinogen con- clotting time screen used for protamine titration.
tent. The TEG MA value represents the maxi- The thrombin time test may also be used to
mum clot strength and is affected by platelet detect the effects of heparin, but this assay is
count and fibrinogen. The VE assays can be typically only available in the main coagulation
modified by inhibition of platelet activity. laboratory and, thus, not ideal for rapid deci-
Viscoelastic fibrinogen results highly correlate sion-making. Point-­of-­care assays, such as acti-
with routine laboratory (Claus method) fibrino- vated clotting time or modified forms of this
gen results. Estimated percentage lysis reflects test, are routinely used in the operating room to
fibrinolytic activity and is the only clinically assess for heparin effects [99].

Massive Transfusion
BSLMC Operating
Room Transfusion Microvascular Bleeding by
Algorithm Observation of Surgical Field

Surgical
Coagulation and
Protamine ACT > Baseline All normal Re-exploration of
Platelet Tests
chest

PLT Fibrinogen
TEG MA < 48 INR > 1.6 PTT > 57
< 102K < 144
and/or and/or

Cryoprecipitate
Platelet transfusion Plasma transfusion
transfusion

Fig. 8.2  Intraoperative transfusion algorithm. ACT activated clotting time, INR international normalized ratio, PLT
platelet, PTT partial thromboplastin time, TEG Thromboelastograph
118 K. Ayyagari et al.

Postoperative Management ate whether to initiate anticoagulation on the

second postoperative day. We use reduced chest
Patients should be transferred to the recovery tube output (<0.5 mL × kg body weight) and lack
room/ICU only when the bleeding rate has been of ongoing demand for red blood cells as gauges
controlled, which should be achieved by follow- to help make this decision. Anticoagulation is
ing a defined treatment algorithm involving surgi- typically initiated by administering unfraction-
cal techniques and hemostatic therapy (Fig. 8.3). ated heparin because of its rapid onset.
When the patient is admitted to the ICU, the coag- Unfractionated heparin should not be used at
ulation system should be assessed using tradi- high doses to avoid bleeding; we start with a
tional coagulation assays and/or VE testing. dose around 10–15 U/kg and then gradually
Recent studies of cases involving frequent, large- increase it until a PTT of 50–60 s is attained. It is
volume bleeding show that careful attention to important for the team to know the dose response
VE measures and maintaining optimal fibrinogen of their particular reagent system to unfraction-
content is very helpful for optimizing hemostasis ated heparin. Although it has not been validated
[100]. Data on optimal parameters for other mea- for this purpose, the difference in TEG R time
sures are needed. Each team should know what between kaolin and heparinase cups can be
the target INR should be for the critical therapeu- effective for assessing onset of heparin activity.
tic set points of 30–50% clotting factor activity. Anti-Xa activity may also be used to determine
For example, at our center, an INR of 1.6 is con- the effect of heparin therapy. On postoperative
sidered an acceptable therapeutic target. There are day 3, we typically begin administering aspirin
no validated data regarding therapeutic thresholds (81 or 325 mg) and warfarin, with an INR goal
for platelet count or PTT. of 2.5–3.5. In this setting, there is no need to
If bleeding is controlled on the first postop- rush to attain the therapeutic INR goal of 2–3, so
erative day, the management team should evalu- a large initial loading dose is not recommended.

BSLMC CVRR
Transfusion
Algorithm

Chest Tube Output < 150 ml/hr

(2 ml/kg/hr) Chest Tube Output > 150 ml/hr
Add PEEP
(> 2ml/kg/hr)

Surgical
Do Nothing - Observe Coagulation and
All normal Re-exploration of
Platelet Tests
chest

PLT Fibrinogen
TEG MA < 55 INR > 1.6 PTT > 45
< 102K < 200
and/or and/or

Cryoprecipitate
Platelet transfusion Plasma transfusion
transfusion

Fig. 8.3  Postoperative transfusion algorithm. INR international normalized ratio, PEEP positive end-expiratory pres-
sure, PLT platelet, PTT partial thromboplastin time, TEG thromboelastograph
8  Perioperative Management of LVAD Patients 119

Additional Hemostasis Postoperative Review of Systems

Considerations
After the immediate post-LVAD implantation
 lterations in von Willebrand Factor
A assessment to establish adequate gas exchange
(VWF) and hemodynamics, we perform a complete
The high flow states that occur with VAD use review of the systems at risk for complications.
have been associated with abnormal VWF mul- As part of this review, we perform a brief neurol-
timer patterns and clinical bleeding [101, 102]. ogy exam to screen for a CVA when the patient
If unexpected bleeding occurs in the early wakes from anesthesia. The exam consists of a
postoperative interval, this possibility should level of arousal assessment and a gross motor and
be considered, and a laboratory assessment of sensory exam of the extremities. Cerebral vascu-
VWF multimers should be performed. If bleed- lar accident is one of the leading complications of
ing occurs but measured coagulation parame- LVAD implantation [1, 22, 105–107] and, as pre-
ters are normal, treatment with cryoprecipitated viously stated, is the second leading cause of
antihemophilic factor concentrate may be con- death in the early postoperative period. Transient
sidered. It is helpful to have expert hematology ischemic attacks have also been reported after
consultants to aid in managing these issues. LVAD implantation at an incidence rate of
approximately 12% [3, 49, 51]. Specific risk fac-
Heparin-Induced Thrombocytopenia tors for CVAs in LVAD recipients have been well
Because unfractionated heparin is used for described and include both hypotension and
patients undergoing CPB and often for periopera- hypertension (MAP >90), infection, pump dys-
tive anticoagulation, staff caring for LVAD function or thrombosis, both excessive anticoag-
patients should be aware of the risk of heparin-­ ulation and insufficient anticoagulation [2, 72,
induced thrombocytopenia, as well as the limita- 105, 107, 108], and heparin-induced thrombocy-
tions of the heparin antibody test and the pitfalls topenia. Outcomes of CVAs are worse for intra-
related to inappropriate use this test [103]. In the cerebral hemorrhage than for ischemic CVAs and
context of cardiac surgery, especially in the set- correlate, not surprisingly, with the extent of the
ting of heart failure and the application of devices neurological deficit [108]. An excellent protocol
associated with increased platelet consumption has been published regarding the management of
(i.e., during the use of intra-aortic balloon pumps CVAs in LVAD patients and the outcomes for
or extracorporeal membrane oxygenation or such [107, 108]. These patients are generally
early after VAD implantation), the use of the hep- managed by following standard CVA treatment
arin antibody test to assess a decrease in platelet guidelines, with some notable caveats. For intra-
count requires considerable caution. The 4T pre- cerebral hemorrhages, the anticoagulation is
test probability score should be taken into reversed, and standard guidelines for surgical
account, and testing should be reserved when intervention are followed. For ischemic strokes,
indicated by this index [104]. The heparin-PF4 thrombolytics are not given in the perioperative
ELISA test is a sensitive test for diagnosing period because of the risk of bleeding. However,
heparin-­induced thrombocytopenia, but it is non-­ patients with large-territory ischemic CVAs may
specific, frequently showing false positives. benefit from catheter thrombectomy within 8 h of
Reports have demonstrated a relationship the event. Antiplatelet and anticoagulation thera-
between assay optical density (strength of test pies are instituted or continued for any CVA (or
result) and true-positive status. Given the serious transient ischemic attack) that involves an infarct
consequences of heparin-induced thrombocyto- area of less than a third of the hemispheric vol-
penia, specifically the need to alter the CPB ume. Infarcts involving a larger area are at risk
approach and perhaps even limit access to trans- for hemorrhagic conversion, so antiplatelet drugs
plantation, it would be wise to confirm the immu- and, particularly, anticoagulants are stopped.
nologic results with the serotonin-release assay There are rare reports of spinal cord infarcts and
(as is the practice in our hospital). peripheral nerve injuries after LVAD implantation,
120 K. Ayyagari et al.

as well [3]. Delirium is relatively common in apy is the preferred form of renal replacement
these patients, occurring in 10% of them [109]. therapy during the postoperative period and has
We manage delirium in the usual manner but proven to be very effective [16, 110–113].
have found dexmedetomidine to be a particularly Postoperative liver dysfunction occurs in
helpful pharmacological therapy in the i­ mmediate 2–8% of patients [3, 20, 23–27, 51]. Therefore,
postoperative period, and we use exercise physi- we monitor liver function tests during the postop-
ologists and physical therapists to aggressively erative period; in particular, we use the transami-
promote the mobilization of patients soon after nases as early markers of liver dysfunction.
extubation as a primary prevention strategy. Although liver dysfunction can be induced in
Renal failure has been shown to occur in these patients by sepsis or the use of toxic drugs,
3–33% of patients after LVAD implantation, the primary causes of liver function that we look
depending upon the series [3, 15, 16, 26, 51]. We for are poor DO2 and RV dysfunction, and any
monitor renal function by using standard indica- issues found are remediated accordingly and
tors: urine output and serum creatinine level. The expeditiously. Progressive liver dysfunction can
Foley catheter output is monitored particularly lead to encephalopathy, coagulopathy, and bleed-
closely during the immediate postoperative ing, as well as vasodilation [20, 24, 25].
period as the earliest marker of renal function. Gastrointestinal bleeding is a very common
The standard of ≥0.5 mL/kg/h of urine output is complication of CF-LVAD implantation, occur-
used as preliminary evidence of adequate renal ring in 15–30% of patients [45, 114]. Although
function. Creatinine measurement is used in the gastrointestinal bleeding is rare in the early post-
standard fashion to determine renal failure (50% operative period, the risk increases the longer an
increase above baseline or 50% decline in glo- LVAD is implanted [45, 114]. The reasons for
merular filtration rate). Evidence suggesting gastrointestinal bleeding are multifactorial and
renal dysfunction prompts a standard evaluation include the development of CF-LVAD-associated
to categorize the dysfunction as occurring from bowel angiodysplasia and coagulopathy, which
prerenal, renal (i.e., acute kidney injury), or post-­ can be caused by anticoagulant use or the loss of
renal problems; this is determined by performing VWF due to the sheer stress from continuous
a urinalysis, urine electrolyte tests, and some- flow. As part of our standard postoperative proce-
times an ultrasound assessment of the urinary dures, we provide either proton pump inhibitors
tract. However, at the first sign of renal dysfunc- or H2 blockers as a prophylaxis against gastritis.
tion, we also reassess DO2 because inadequate Postoperative ileus is common, occurring in
DO2 is the principal cause of renal failure in this nearly 20% of patients. This risk can be mitigated
patient population. If DO2 is found to be inade- largely by instituting early feeds (starting on
quate (Table 8.4), we promptly intervene to cor- postoperative day 2 or 3) and a bowel regimen
rect this. In addition to being at increased risk for [115]. It is helpful to remember that the patients
acute kidney injury due to poor DO2, LVAD undergoing LVAD implantation are often mal-
patients are specifically at risk for emboli to the nourished at the time of surgery, so it is important
kidney if adequate anticoagulation is not achieved to ensure that these patients are well nourished
and for LVAD-associated hemolysis [16]. We are during the postoperative period.
particularly careful to avoid using potentially The most concerning hematologic complica-
nephrotoxic drugs in the at-risk LVAD popula- tion of LVAD placement, aside from gastrointes-
tion. For this patient population, we use the stan- tinal bleeding (discussed above), is hemolysis.
dard indications to determine the need for However, hemolysis is rare during the immediate
hemodialysis, except that we often resort to using postimplantation period, occurring in 3–5% of
continuous renal replacement therapy quicker patients. Typically, hemolysis is associated with
than usual for patients with RHF to control vol- excessively high device speeds and, particularly,
ume that is refractory to diuretics (Chap. 20). with device thrombosis [45, 116]. It is diagnosed
Continuous venovenous renal replacement ther- by showing elevated levels of plasma-free hemo-
8  Perioperative Management of LVAD Patients 121

globin and lactate dehydrogenase [45, 116]. associated bacteremia, pneumonia, urinary tract
Consequences of hemolysis include anemia, infection, Clostridium difficile bowel infection,
decreased DO2, elevated pulmonary and systemic and sternal wound infection. Sepsis syndrome
vascular resistance, disordered coagulation, renal occurs in up to about 20% of patients and is asso-
failure, and systemic inflammatory response ciated with a very high mortality rate of approxi-
­syndrome. It is important to note that CPB itself mately 50%. Recommendations to decrease the
can cause hemolysis, although it is mild and risk of these infections include early extubation
occurs in the immediate postoperative period (up to 18% of patients in this population on pro-
[117, 118]. Heparin-induced thrombocytopenia longed mechanical ventilation acquire pneumo-
syndrome (introduced above) is uncommon in nia) [51] and removal of early invasive lines and
LVAD patients but can be catastrophic by pro- Foley catheters [5]. Emerging data suggesting
moting pump thrombosis [119]. When heparin- that the use of proton pump inhibitors is a risk
induced thrombocytopenia is diagnosed or even factor for Clostridium difficile infection [125] has
strongly suspected, we immediately discontinue prompted some to consider avoiding these agents
heparin and begin administering a direct throm- in favor of H2 blockers for gastritis prophylaxis.
bin inhibitor; bivalirudin has served us well for Other interventions used to decrease the risk of
this purpose. infection include early enteral nutrition and glu-
At this time, there are no known clinically sig- cose control. We also practice thorough culturing
nificant effects of LVAD placement on the endo- and investigation of potential sources when early
crine system [45]. However, hyperglycemia signs of infection appear, such as fever or ele-
complicates the postoperative course of LVAD vated white blood cell count. It is hoped that
recipients, just as it does for patients who have evolving molecular markers for infection will
undergone other cardiac surgeries and CPB. We allow us to identify infections in these patients
know from many studies conducted over the past even earlier. At the first signs of sepsis syndrome
15 years that patients whose glucose is controlled in these patients, we immediately institute
by continuous insulin infusions suffer less infec- empiric broad-spectrum antibiotic therapy, as has
tions and have better outcomes [120]. Therefore, been a standard recommendation of infectious
we employ standard post-cardiac surgery insulin disease experts who work with these patients.
protocols for these patients. Typically, vancomycin and either a third- or
Prophylactic administration of antibiotics is fourth-generation cephalosporin, advanced peni-
recommended for 48 h after LVAD implantation, cillin combination drug, or carbapenem is used;
as described above. Surgical wounds and the occasionally, depending upon the risk factors,
driveline should be meticulously cared for by antifungal therapy will be added [3, 4,
cleaning them with chlorhexidine daily and using 122–124].
sterile dressings. Recently, one group demon-
strated good outcomes when using silver-­
impregnated gauze dressings [69]. Additionally, Delayed Sternal Closure
it has been shown that fixing the LVAD driveline
in position by using various anchoring devices The incidence rate of delayed sternal closure
decreases the incidence of site infections [121]. (DSC) has been reported to be 3–50% for LVAD
However, for up to 42% of LVAD patients, the patients [126]. Although recent series have
index hospitalization is complicated by infection; shown rates in the lower range of these num-
these infections are typically non-LVAD related bers, delayed sternal closure is much more
[3, 4, 122–124]. In contrast, infections occurring common in LVAD patients than in the general
after the initial hospitalization typically are population of adult cardiac surgery patients and
LVAD related (driveline and pump pocket infec- is primarily associated with intraoperative
tions) [4]. The common infections seen during bleeding and RHF severe enough to require
the initial hospitalization period include line-­ RVAD implantation for hemodynamic support.
122 K. Ayyagari et al.

Risk factors reported for DSC include increased 4. Topkara VK, Kondareddy S, Malik F, Wang IW,
Mann DL, Ewald GA, et al. Infectious complications
number and severity of preoperative morbidi-
in patients with left ventricular assist device: etiol-
ties in general and, more specifically, factors ogy and outcomes in the continuous-flow era. Ann
that increase the chance of bleeding and Thorac Surg. 2010;90:1270–7.
RHF. The risk of bleeding and DSC is increased 5. Feldman D, Pamboukian SV, Teuteberg JJ, Birks E,
Lietz K, Moore SA, et al. The 2013 International
in patients on IIb/IIIa inhibitors, with a high
Society for Heart and Lung Transplantation
MELD score or renal failure, or supported by Guidelines for mechanical circulatory support:
extracorporeal membrane oxygenation immedi- executive summary. J Heart Lung Transplant.
ately before the operation. Our experience sug- 2013;32:157–87.
6. Holdy K, Dembitsky W, Eaton LL, Chillcott S,
gests that having a history of sternotomy
Stahovich M, Rasmusson B, et al. Nutrition assess-
increases the risk, as well. Patients who have ment and management of left ventricular assist device
DSC when undergoing LVAD implantation patients. J Heart Lung Transplant. 2005;24:1690–6.
have been reported, in most studies, to have a 7. Yost G, Gregory M, Bhat G. Nutrition assessment
with indirect calorimetry in patients evaluated for
significantly increased risk of mortality [126].
left ventricular assist device implantation. Nutr Clin
This is presumably due to the patients’ higher Pract. 2015;30:690–7.
acuity of illness going into the device implanta- 8. Liszkowski M, Teuteberg JJ, Myers SL, Rogers JG,
tion procedure and the complications necessi- Starling R, Ascheim DD, et al. 63: INTERMACS
profiles of nutrition and organ function in relation to
tating DSC. However, there is also data
outcomes. J Heart Lung Transplant. 2010;29:S27.
suggesting that DSC is an independent risk fac- 9. Aggarwal A, Kumar A, Gregory MP, Blair C,
tor for mortality in the LVAD population [126]. Pauwaa S, Tatooles AJ, et al. Nutrition assessment in
Why this would be is unclear, but it has been advanced heart failure patients evaluated for ventric-
ular assist devices or cardiac transplantation. Nutr
suggested that the higher incidence of sepsis in
Clin Pract. 2013;28:112–9.
this group may be the prime driver [126]. 10. Holman WL, Skinner JL, Waites KB, Benza RL,
Our specific approach to DSC) is to correct McGiffin DC, Kirklin JK. Infection during circula-
the coagulopathy and RHF that often precipi- tory support with ventricular assist devices. Ann
Thorac Surg. 1999;68:711–6.
tates the need for DSC, to keep the patient well
11. Acharya MN, Som R, Tsui S. What is the opti-
sedated (and occasionally paralyzed) to mum antibiotic prophylaxis in patients undergo-
decrease the risk of a movement-related sternal ing implantation of a left ventricular assist device?
complication, to stringently enforce barrier Interact Cardiovasc Thorac Surg. 2012;14:209–14.
12. George S, Leasure AR, Horstmanshof
protection, and to continue aggressive use of
D. Effectiveness of decolonization with chlorhexi-
prophylactic antibiotics for the duration of time dine and mupirocin in reducing surgical site infec-
that the chest is open. tions: a systematic review. Dimens Crit Care Nurs.
2016;35:204–22.
13. Brisco MA, Testani JM, Cook JL. Renal dysfunc-
References tion and chronic mechanical circulatory support:
from patient selection to long-term management and
prognosis. Curr Opin Cardiol. 2016;31:277–86.
1. Kirklin JK, Naftel DC, Pagani FD, Kormos 14. Gilotra NA, Russell SD. Patient selection for
RL, Stevenson LW, Blume ED, et al. Seventh mechanical circulatory support. Heart Fail Rev.
INTERMACS annual report: 15,000 patients 2013;18:27–34.
and counting. J Heart Lung Transplant. 15. Mao H, Katz N, Kim JC, Day S, Ronco
2015;34:1495–504. C. Implantable left ventricular assist devices and the
2. Stulak JM, Mehta V, Schirger JA, Aaronson KD, kidney. Blood Purif. 2014;37:57–66.
Joyce LD, Daly RC, et al. Temporal differences in 16. Tromp TR, de Jonge N, Joles JA. Left ventricular
causes of mortality after left ventricular assist device assist devices: a kidney’s perspective. Heart Fail
implantation. Ann Thorac Surg. 2015;99:1969–72. Rev. 2015;20:519–32.
discussion 1972–4 17. Lima B, Kale P, Gonzalez-Stawinski GV, Kuiper JJ,
3. Allen SJ, Sidebotham D. Postoperative care and Carey S, Hall SA. Effectiveness and safety of the
complications after ventricular assist device Impella 5.0 as a bridge to cardiac transplantation or
implantation. Best Pract Res Clin Anaesthesiol. durable left ventricular assist device. Am J Cardiol.
2012;26:231–46. 2016;117:1622–8.
8  Perioperative Management of LVAD Patients 123

18. Rao V, Oz MC, Flannery MA, Catanese KA, during left ventricular assist device therapy. J Artif
Argenziano M, Naka Y. Revised screening scale Organs. 2016;19:330–5.
to predict survival after insertion of a left ven- 32. Mohamedali B, Bhat G, Yost G, Tatooles A. Changes
tricular assist device. J Thorac Cardiovasc Surg. in spirometry after left ventricular assist device
2003;125:855–62. implantation. Artif Organs. 2015;39:1046–50.
19. Borgi J, Tsiouris A, Hodari A, Cogan CM, Paone 33. Arena R, Humphrey R, McCall R. Altered exercise
G, Morgan JA. Significance of postoperative acute pulmonary function after left ventricular assist device
renal failure after continuous-flow left ­ventricular implantation. J Cardpulm Rehabil. 1999;19:344–6.
assist device implantation. Ann Thorac Surg. 34. Herlihy J, Cooper J, Reul R. Cardiac surgery and
2013;95:163–9. weaning: Clinical and therapeutic implications. In:
20. Maltais S, Stulak JM. Right and left ventricular Rodriguez E, Matias A, Volsko TA, editors. Yearbook
assist devices support and liver dysfunction: prog- respiratory care clinics and applied technologies. San
nostic and therapeutic implications. Curr Opin Franciso, S.A.: Tipografia; 2008. p. 907–20.
Cardiol. 2016;31:287–91. 35. Wynne R, Botti M. Postoperative pulmonary
21. Wadia Y, Etheridge W, Smart F, Wood RP, Frazier dysfunction in adults after cardiac surgery with
OH. Pathophysiology of hepatic dysfunction and cardiopulmonary bypass: clinical significance
intrahepatic cholestasis in heart failure and after and implications for practice. Am J Crit Care.
left ventricular assist device support. J Heart Lung 2004;13:384–93.
Transplant. 2005;24:361–70. 36. Weiner P, Zeidan F, Zamir D, Pelled B, Waizman
22. Yuan N, Arnaoutakis GJ, George TJ, Allen JG, Ju J, Beckerman M, et al. Prophylactic inspira-
DG, Schaffer JM, et al. The spectrum of complica- tory muscle training in patients undergoing
tions following left ventricular assist device place- coronary artery bypass graft. World J Surg.
ment. J Card Surg. 2012;27:630–8. 1998;22:427–31.
23. Concha PM, Mertz KV. Perioperative risk 37. Houston BA, Kalathiya RJ, Hsu S, Loungani R,
among patients with cirrhosis. Rev Med Chil. Davis ME, Coffin ST, et al. Right ventricular after-
2010;138:1165–71. load sensitivity dramatically increases after left ven-
24. Matthews JC, Pagani FD, Haft JW, Koelling TM, tricular assist device implantation: a multi-center
Naftel DC, Aaronson KD. Model for end-stage liver hemodynamic analysis. J Heart Lung Transplant.
disease score predicts left ventricular assist device 2016;35:868–76.
operative transfusion requirements, morbidity, and 38. Epting CL, McBride ME, Wald EL, Costello JM.
mortality. Circulation. 2010;121:214–20. Pathophysiology of post-operative low cardiac out-
25. Modi A, Vohra HA, Barlow CW. Do patients with put syndrome. Curr Vasc Pharmacol. 2016;14:14–23.
liver cirrhosis undergoing cardiac surgery have 39. Omar S, Zedan A, Nugent K. Cardiac vasoplegia
acceptable outcomes? Interact Cardiovasc Thorac syndrome: pathophysiology, risk factors and treat-
Surg. 2010;11:630–4. ment. Am J Med Sci. 2015;349:80–8.
26. Morgan JA, Go PH, Xuereb L, Kaur B, Akrawe S, 40. Flores AS, Essandoh M, Yerington GC, Bhatt
Nemeh HW, et al. Outcomes on continuous flow left AM, Iyer MH, Perez W, et al. Echocardiographic
ventricular assist devices: a single institutional 9-year assessment for ventricular assist device placement.
experience. Ann Thorac Surg. 2016;102:1266–73. J Thorac Dis. 2015;7:2139–50.
27. Weymann A, Patil NP, Sabashnikov A, Mohite PN, 41. de Jonge N, van Wichen DF, Schipper ME, Lahpor
Garcia Saez D, Bireta C, et al. Continuous-flow left JR, Gmelig-Meyling FH, Robles de Medina EO,
ventricular assist device therapy in patients with pre- et al. left ventricular assist device in end-stage heart
operative hepatic failure: are we pushing the limits failure: persistence of structural myocyte damage
too far? Artif Organs. 2015;39:336–42. after unloading. An immunohistochemical analysis
28. Agostoni P, Cattadori G, Guazzi M, Palermo P, of the contractile myofilaments. J Am Coll Cardiol.
Bussotti M, Marenzi G. Cardiomegaly as a possible 2002;39:963–9.
cause of lung dysfunction in patients with heart fail- 42. Buckberg G, Hoffman JI, Mahajan A, Saleh S,
ure. Am Heart J. 2000;140:e24. Coghlan C. Cardiac mechanics revisited: the rela-
29. Dimopoulou I, Daganou M, Tsintzas OK, Tzelepis tionship of cardiac architecture to ventricular func-
GE. Effects of severity of long-standing congestive tion. Circulation. 2008;118:2571–87.
heart failure on pulmonary function. Respir Med. 43. Ambardekar AV, Buttrick PM. Reverse remodeling
1998;92:1321–5. with left ventricular assist devices: a review of clini-
30. Hosenpud JD, Stibolt TA, Atwal K, Shelley cal, cellular, and molecular effects. Circ Heart Fail.
D. Abnormal pulmonary function specifically 2011;4:224–33.
related to congestive heart failure: comparison of 44. Pratt AK, Shah NS, Boyce SW. Left ventricular
patients before and after cardiac transplantation. Am assist device management in the ICU. Crit Care
J Med. 1990;88:493–6. Med. 2014;42:158–68.
31. Imamura T, Kinugawa K, Kinoshita O, Nawata K, 45. Healy AH, McKellar SH, Drakos SG, Koliopoulou
Ono M. Reversible decline in pulmonary function A, Stehlik J, Selzman CH. Physiologic effects of
124 K. Ayyagari et al.

continuous-flow left ventricular assist devices. 59. Stainback RF, Estep JD, Agler DA, Birks EJ, Bremer
J Surg Res. 2016;202:363–71. M, Hung J, et al. Echocardiography in the manage-
46. Sheu R, Joshi B, High K, Thinh Pham D, Ferreira ment of patients with left ventricular assist devices:
R, Cobey F. Perioperative management of patients recommendations from the American Society
with left ventricular assist devices undergoing non- of Echocardiography. J Am Soc Echocardiogr.
cardiac procedures: a survey of current practices. 2015;28:853–909.
J Cardiothorac Vasc Anesth. 2015;29:17–26. 60. Pagani FD. Continuous-flow rotary left ventricular
47. Cheifetz IM. Cardiorespiratory interactions: the assist devices with “3rd generation” design. Semin
relationship between mechanical ventilation and Thorac Cardiovasc Surg. 2008;20:255–63.
hemodynamics. Respir Care. 2014;59:1937–45. 61. Slaughter MS, Bartoli CR, Sobieski MA,
48. Tam MK, Wong WT, Gomersall CD, Tian Q, Ng Pantalos GM, Giridharan GA, Dowling RD, et al.
SK, Leung CC, et al. A randomized controlled trial Intraoperative evaluation of the HeartMate II flow
of 2 protocols for weaning cardiac surgical patients estimator. J Heart Lung Transplant. 2009;28:39–43.
receiving adaptive support ventilation. J Crit Care. 62. Slaughter MS, Pagani FD, Rogers JG, Miller LW,
2016;33:163–8. Sun B, Russell SD, et al. Clinical management of
49. Slaughter MS, Rogers JG, Milano CA, Russell SD, continuous-flow left ventricular assist devices in
Conte JV, Feldman D, et al. Advanced heart failure advanced heart failure. J Heart Lung Transplant.
treated with continuous-flow left ventricular assist 2010;29:S1–39.
device. N Engl J Med. 2009;361:2241–51. 63. Blum FE, Weiss GM, Cleveland JC Jr, Weitzel
50. John R, Kamdar F, Liao K, Colvin-Adams M, Boyle A, NS. Postoperative management for patients
Joyce L. Improved survival and decreasing incidence with durable mechanical circulatory support
of adverse events with the HeartMate II left ventricu- devices. Semin Cardiothorac Vasc Anesth.
lar assist device as bridge-to-transplant therapy. Ann 2015;19:318–30.
Thorac Surg. 2008;86:1227–34. discussion 1234–5 64. Scolletta S, Biagioli B, Franchi F, Muzzi
51. Genovese EA, Dew MA, Teuteberg JJ, Simon MA, L. Echocardiography and hemodynamic monitoring
Kay J, Siegenthaler MP, et al. Incidence and patterns tools for clinical assessment of patients on mechani-
of adverse event onset during the first 60 days after cal circulatory support. In: Komamura K, editor.
ventricular assist device implantation. Ann Thorac Recent advances in the field of ventricular assist
Surg. 2009;88:1162–70. devices. Croatia: InTech; 2013.
52. Tsiouris A, Paone G, Nemeh HW, Borgi J, Williams 65. Lampert BC, Teuteberg JJ. Right ventricular failure
CT, Lanfear DE, et al. Short and long term out- after left ventricular assist devices. J Heart Lung
comes of 200 patients supported by continuous-­ Transplant. 2015;34:1123–30.
flow left ventricular assist devices. World J Cardiol. 66. Rich JD. Right ventricular failure in patients
2015;7:792–800. with left ventricular assist devices. Cardiol Clin.
53. Miller LW, Pagani FD, Russell SD, John R, Boyle 2012;30:291–302.
AJ, Aaronson KD, et al. Use of a continuous-flow 67. Wang TS, Hernandez AF, Felker GM, Milano
device in patients awaiting heart transplantation. N CA, Rogers JG, Patel CB. Valvular heart disease
Engl J Med. 2007;357:885–96. in patients supported with left ventricular assist
54. Horvath D, Byram N, Karimov JH, Kuban B, devices. Circ Heart Fail. 2014;7:215–22.
Sunagawa G, Golding LA, et al. Mechanism of self-­ 68. Morgan JA, Brewer RJ, Nemeh HW, Murthy R,
regulation and in vivo performance of the Cleveland Williams CT, Lanfear DE, et al. Left ventricu-
Clinic continuous-flow total artificial heart. Artif lar reverse remodeling with a continuous flow left
Organs. 2016;41(5):411–7. ventricular assist device measured by left ventricu-
55. HeartWare. HeartWare Ventricular Assist System: lar end-diastolic dimensions and severity of mitral
Instructions for Use. [Link] regurgitation. ASAIO J. 2012;58:574–7.
sites/default/files/uploads/docs/ifu00001_rev_15. 69. Topilsky Y, Price TN, Atchison FW, Joyce LD.
pdf. Accessed 1 Feb 2017. Atypical tamponade hemodynamic in a patient with
56. Thoratec Corporation. Resource Library (USA): temporary left ventricular assist device. Interact
HeartMate II LVAD. [Link] Cardiovasc Thorac Surg. 2011;12:832–4.
ical-professionals/resource-library/ifus-manuals/ 70. Kimmaliardjuk DM, Ruel M. Cardiac passive-­
[Link]. Accessed 1 Feb 2017. aggressive behavior? The right ventricle in patients
57. Moazami N, Fukamachi K, Kobayashi M, Smedira with a left ventricular assist device. Expert Rev
NG, Hoercher KJ, Massiello A, et al. Axial and cen- Cardiovasc Ther. 2017;15:267–76.
trifugal continuous-flow rotary pumps: a translation 71. Irwin Z, Cook JO. Advances in point-of-care tho-
from pump mechanics to clinical practice. J Heart racic ultrasound. Emerg Med Clin North Am.
Lung Transplant. 2013;32:1–11. 2016;34:151–7.
58. Meineri M, Van Rensburg AE, Vegas A. Right ven- 72. Harvey L, Holley C, Roy SS, Eckman P, Cogswell
tricular failure after LVAD implantation: prevention R, Liao K, et al. Stroke after left ventricular assist
and treatment. Best Pract Res Clin Anaesthesiol. device implantation: outcomes in the continuous-­
2012;26:217–29. flow era. Ann Thorac Surg. 2015;100:535–41.
8  Perioperative Management of LVAD Patients 125

73. Cheng A, Swartz MF, Massey HT. VADoscopy: a medical issues. Journal of Congestive Heart Failure
novel intraoperative technique to evaluate HeartMate & Circulatory Support. 2001;2:13–7.
II left ventricular assist device inflow obstruction 86. Bunte MC, Blackstone EH, Thuita L, Fowler J,
and thrombosis. ASAIO J. 2013;59:671–4. Joseph L, Ozaki A, et al. Major bleeding dur-
74. Doligalski CT, Jennings DL. Device-related throm- ing HeartMate II support. J Am Coll Cardiol.
bosis in continuous-flow left ventricular assist device 2013;62:2188–96.
support. J Pharm Pract. 2016;29:58–66. 87. Haglund NA, Davis ME, Tricarico NM, Ahmad
75. Toeg H, Ruel M, Haddad H. Anticoagulation strat- RM, DiSalvo TG, Keebler ME, et al. Perioperative
egies for left ventricular assist devices. Curr Opin blood product use: a comparison between HeartWare
Cardiol. 2015;30:192–196. and HeartMate II devices. Ann Thorac Surg.
76. Shinar Z, Bellezzo J, Stahovich M, Cheskes S, 2014;98:842–9.
Chillcott S, Dembitsky W. Chest compressions 88. Myles PS, Smith JA, Forbes A, Silbert B, Jayarajah
may be safe in arresting patients with left ven- M, Painter T, et al. Stopping vs. continuing aspi-
tricular assist devices (LVADs). Resuscitation. rin before coronary artery surgery. N Engl J Med.
2014;85:702–4. 2016;374:728–37.
77. Bartoli CR, Ghotra AS, Pachika AR, Birks EJ, 89. Society of Thoracic Surgeons Blood Conservation
McCants KC. Hematologic markers better pre- Guideline Task F, Ferraris VA, Brown JR, Despotis
dict left ventricular assist device thrombosis than GJ, Hammon JW, Reece TB, et al. 2011 update to
echocardiographic or pump parameters. Thorac the Society of Thoracic Surgeons and the Society of
Cardiovasc Surg. 2014;62:414–8. Cardiovascular Anesthesiologists blood conserva-
78. Garan AR, Levin AP, Topkara V, Thomas SS, tion clinical practice guidelines. Ann Thorac Surg.
Yuzefpolskaya M, Colombo PC, et al. Early post-­ 2011;91:944–82.
operative ventricular arrhythmias in patients with 90. Chokshi A, Cheema FH, Schaefle KJ, Jiang J,
continuous-flow left ventricular assist devices. Collado E, Shahzad K, et al. Hepatic dysfunction and
J Heart Lung Transplant. 2015;34:1611–6. survival after orthotopic heart transplantation: appli-
79. Healy C, Viles-Gonzalez JF, Sacher F, Coffey JO, cation of the MELD scoring system for outcome pre-
d'Avila A. Management of ventricular arrhythmias diction. J Heart Lung Transplant. 2012;31:591–600.
in patients with mechanical ventricular support 91. Tamim H, Habbal M, Saliba A, Musallam K,
devices. Curr Cardiol Rep. 2015;17:59. Al-Taki M, Hoballah J, et al. Preoperative INR and
80. McIlvennan CK, Magid KH, Ambardekar AV, postoperative major bleeding and mortality: a ret-
Thompson JS, Matlock DD, Allen LA. Clinical rospective cohort study. J Thromb Thrombolysis.
outcomes after continuous-flow left ventricular 2016;41:301–11.
assist device: a systematic review. Circ Heart Fail. 92. Meehan R, Tavares M, Sweeney J. Clinical experi-
2014;7:1003–13. ence with oral versus intravenous vitamin K for war-
81. Schleifer JW, Mookadam F, Kransdorf EP, Nanda U, farin reversal. Transfusion. 2013;53:491–8. quiz 490
Adams JC, Cha S, et al. Effect of continued cardiac 93. Sarode R, Milling TJ Jr, Refaai MA, Mangione A,
resynchronization therapy on ventricular arrhyth- Schneider A, Durn BL, et al. Efficacy and safety
mias after left ventricular assist device implantation. of a 4-factor prothrombin complex concentrate in
Am J Cardiol. 2016;118:556–9. patients on vitamin K antagonists presenting with
82. Enriquez AD, Calenda B, Gandhi PU, Nair AP, major bleeding: a randomized, plasma-controlled,
Anyanwu AC, Pinney SP. Clinical impact of phase IIIb study. Circulation. 2013;128:1234–43.
atrial fibrillation in patients with the HeartMate II 94. Marshall AL, Levine M, Howell ML, Chang Y,
left ventricular assist device. J Am Coll Cardiol. Riklin E, Parry BA, et al. Dose-associated pulmo-
2014;64:1883–90. nary complication rates after fresh frozen plasma
83. Wann LS, Curtis AB, January CT, Ellenbogen KA, administration for warfarin reversal. J Thromb
Lowe JE, Estes NA 3rd, et al. 2011 ACCF/AHA/ Haemost. 2016;14:324–30.
HRS focused update on the management of patients 95. Douketis JD, Woods K, Foster GA, Crowther
with atrial fibrillation (updating the 2006 guideline): MA. Bridging anticoagulation with low-molecular-­
a report of the American College of Cardiology weight heparin after interruption of warfarin therapy
Foundation/American Heart Association Task is associated with a residual anticoagulant effect
Force on Practice Guidelines. J Am Coll Cardiol. prior to surgery. Thromb Haemost. 2005;94:528–31.
2011;57:223–42. 96. Nuttall GA, Oliver WC, Santrach PJ, Bryant S,
84. Radiovancevic R, Radovancevic B, Bracey A, Riggs Dearani JA, Schaff HV, et al. Efficacy of a simple
S, Frazier OH. Coagulations parameters in patients intraoperative transfusion algorithm for noneryth-
undergoing left ventricular assist device implanta- rocyte component utilization after cardiopulmonary
tion. ASAIO J. 2001;47:134. bypass. Anesthesiology. 2001;94:773–81. discus-
85. Radovancevic B, Bracey AW, Riggs SA, sion 5A–6A
Radovancevic R, Frazier OH. Left ventricular assist 97. Shore-Lesserson L, Manspeizer HE, DePerio
devices and bleeding diatheses: hematologic and M, Francis S, Vela-Cantos F, Ergin MA.
126 K. Ayyagari et al.

Thromboelastography-guided transfusion algorithm assist device implantation on renal function in

reduces transfusions in complex cardiac surgery. patients requiring post-implant renal replacement
Anesth Analg. 1999;88:312–9. therapy. J Heart Lung Transplant. 2011;30:182–7.
98. Haas T, Spielmann N, Mauch J, Madjdpour C, Speer 111. Ford RM, Book W, Spivey JR. Liver disease related
O, Schmugge M, et al. Comparison of thromboelas- to the heart. Transplant Rev (Orlando). 2015;29:33–7.
tometry (ROTEM(R)) with standard plasmatic coag- 112. Kaltenmaier B, Pommer W, Kaufmann F, Hennig
ulation testing in paediatric surgery. Br J Anaesth. E, Molzahn M, Hetzer R. Outcome of patients
2012;108:36–41. with ventricular assist devices and acute renal fail-
99. Despotis GJ, Gravlee G, Filos K, Levy ure requiring renal replacement therapy. ASAIO
J. Anticoagulation monitoring during cardiac sur- J. 2000;46:330–3.
gery: a review of current and emerging techniques. 113. Topkara VK, Dang NC, Barili F, Cheema FH,
Anesthesiology. 1999;91:1122–51. Martens TP, George I, et al. Predictors and out-
100. Rahe-Meyer N, Pichlmaier M, Haverich A, Solomon comes of continuous veno-venous hemodialysis use
C, Winterhalter M, Piepenbrock S, et al. Bleeding after implantation of a left ventricular assist device.
management with fibrinogen concentrate targeting a J Heart Lung Transplant. 2006;25:404–8.
high-normal plasma fibrinogen level: a pilot study. 114. Stulak JM, Davis ME, Haglund N, Dunlay S,
Br J Anaesth. 2009;102:785–92. Cowger J, Shah P, et al. Adverse events in con-
101. Crow S, Chen D, Milano C, Thomas W, Joyce L, temporary continuous-flow left ventricular assist
Piacentino V 3rd, et al. Acquired von Willebrand devices: a multi-institutional comparison shows
syndrome in continuous-flow ventricular assist significant differences. J Thorac Cardiovasc Surg.
device recipients. Ann Thorac Surg. 2010;90:1263– 2016;151:177–89.
9. discussion 1269 115. Badami A, Fehrenbach Prell EA, Murray MA,
102. Jilma-Stohlawetz P, Quehenberger P, Schima H, Johnson MR, Akhter SA, Lozonschi L, et al. A
Stoiber M, Knobl P, Steinlechner B, et al. Acquired novel approach to prevent post-operative ileus
von Willebrand factor deficiency caused by LVAD is after continuous-flow left ventricular assist device
ADAMTS-13 and platelet dependent. Thromb Res. implantation: a retrospective cohort study. Int J Surg.
2016;137:196–201. 2015;20:135–9.
103. Warkentin TE. Heparin-induced thrombocytopenia. 116. Vercaemst L. Hemolysis in cardiac surgery patients
Curr Opin Crit Care. 2015;21:576–85. undergoing cardiopulmonary bypass: a review in
104. Crowther MA, Cook DJ, Albert M, Williamson D, search of a treatment algorithm. J Extra Corpor
Meade M, Granton J, et al. The 4Ts scoring system Technol. 2008;40:257–67.
for heparin-induced thrombocytopenia in medical-­ 117. Rother RP, Bell L, Hillmen P, Gladwin MT. The clin-
surgical intensive care unit patients. J Crit Care. ical sequelae of intravascular hemolysis and extra-
2010;25:287–93. cellular plasma hemoglobin: a novel mechanism of
105. Morgan JA, Brewer RJ, Nemeh HW, Gerlach B, human disease. JAMA. 2005;293:1653–62.
Lanfear DE, Williams CT, et al. Stroke while on long-­ 118. Wetz AJ, Richardt EM, Schotola H, Bauer M, Brauer
term left ventricular assist device support: incidence, A. Haptoglobin and free haemoglobin during car-
outcome, and predictors. ASAIO J. 2014;60:284–9. diac surgery-is there a link to acute kidney injury?
106. Pagani FD, Milano CA, Tatooles AJ, Bhat G, Anaesth Intensive Care. 2017;45:58–66.
Slaughter MS, Birks EJ, et al. HeartWare HVAD for 119. Awad H, Bryant R, Malik O, Dimitrova G, Sai-­
the treatment of patients with advanced heart fail- Sudhakar CB. Thrombosis during off pump LVAD
ure ineligible for cardiac transplantation: results of placement in a patient with heparin induced throm-
the ENDURANCE destination therapy trial. J Heart bocytopenia using bivalirudin. J Cardiothorac Surg.
Lung Transplant. 2015;34:S9. 2013;8:115.
107. Willey JZ, Demmer RT, Takayama H, Colombo PC, 120. Omar AS, Salama A, Allam M, Elgohary Y,
Lazar RM. Cerebrovascular disease in the era of Mohammed S, Tuli AK, et al. Association of time in
left ventricular assist devices with continuous flow: blood glucose range with outcomes following car-
risk factors, diagnosis, and treatment. J Heart Lung diac surgery. BMC Anesthesiol. 2015;15:14.
Transplant. 2014;33:878–87. 121. Cagliostro B, Levin AP, Fried J, Stewart S, Parkis
108. Willey JZ, Gavalas MV, Trinh PN, Yuzefpolskaya G, Mody KP, et al. Continuous-flow left ventricu-
M, Reshad Garan A, Levin AP, et al. Outcomes after lar assist devices and usefulness of a standardized
stroke complicating left ventricular assist device. strategy to reduce drive-line infections. J Heart Lung
J Heart Lung Transplant. 2016;35:1003–9. Transplant. 2016;35:108–14.
109. Baba A, Hirata G, Yokoyama F, Kenmoku K, 122. Gordon RJ, Quagliarello B, Lowy FD. Ventricular
Tsuchiya M, Kyo S, et al. Psychiatric problems of assist device-related infections. Lancet Infect Dis.
heart transplant candidates with left ventricular 2006;6:426–37.
assist devices. J Artif Organs. 2006;9:203–8. 123. Maniar S, Kondareddy S, Topkara VK. Left ventricu-
110. Demirozu ZT, Etheridge WB, Radovancevic R, lar assist device-related infections: past, present and
Frazier OH. Results of HeartMate II left ­ventricular future. Expert Rev Med Devices. 2011;8:627–34.
8  Perioperative Management of LVAD Patients 127

124. Schaffer JM, Allen JG, Weiss ES, Arnaoutakis GJ, the associated risk of recurrent Clostridium difficile
Patel ND, Russell SD, et al. Infectious complica- infection. JAMA Intern Med. 2015;175:784–91.
tions after pulsatile-flow and continuous-flow left 126. Quader M, LaPar DJ, Wolfe L, Ailawadi G, Rich
ventricular assist device implantation. J Heart Lung J, Speir A, et al. Delayed sternal closure after con-
Transplant. 2011;30:164–74. tinuous flow left ventricle assist device implantation:
125. McDonald EG, Milligan J, Frenette C, Lee analysis of risk factors and impact on outcomes and
TC. Continuous proton pump inhibitor therapy and costs. ASAIO J. 2016;62:432–7.
 Management of Fluid Balance
and Perioperative Renal 9
Complications

Whitson B. Etheridge and Sarah A. Shearer

c­ omplex disease process that is broadly referred

Introduction to as cardiorenal syndrome (Fig. 9.2). Renal and
systemic ischemia and congestive nephropathy
Renal disease and heart disease commonly coex- and hepatopathy from high central venous pres-
ist but also interact in a powerful way. Patients sure (CVP) and low cardiac output cause neuro-
with end-stage heart disease (ESHD) have a high humoral stimulation with upregulation of the
incidence of renal dysfunction [1]. In 84% of the renin-­angiotensin-­aldosterone system and
patients in whom end-stage renal disease (ESRD) increased catecholamines, antidiuretic hormone,
develops, left ventricular (LV) hypertrophy and and inflammatory cytokines (Figs. 9.3, 9.4, and
diastolic heart failure are also present (Fig. 9.1) 9.5) [5–8]. Renal function may also be negatively
[2]. The most common cause of morbidity and affected by the use of diuretics, angiotensin II
mortality in ESRD patients is cardiovascular dis- receptor blockers or angiotensin II-converting
ease. Renal disease coexisting with cardiac dis- enzyme inhibitors, and intravenous contrast
ease is frequently due to underlying comorbidities, agents [6]. Over time, this leads to a decline in
most commonly diabetes and hypertension. renal function with renal fibrosis [6]. ESHD fre-
Systemic diseases (e.g., amyloidosis) may cause quently occurs in older patients and therefore in
heart and renal disease or, less commonly, pri- the presence of low renal reserve (nephronope-
mary renal diseases that, although not the direct nia) and chronic metabolic acidosis, both of
cause of heart disease, may worsen heart failure which may cause renal fibrosis [9].
through hypertension and salt and fluid retention Preoperative abnormal renal function (i.e.,
as the glomerular filtration rate (GFR) falls. GFR <60 mL/min/1.73 m2) is associated with an
Chronic and acute heart failure patients fre- increased incidence of acute kidney injury/acute
quently have renal dysfunction caused by a tubular necrosis (AKI/ATN) and a reduced rate of
survival 1 year after left ventricular assist device
W.B. Etheridge, M.D. (*) (LVAD) placement [10]. Other studies have con-
Renal Transplantation, Texas Heart Institute at Baylor firmed that adverse outcomes are associated with
St. Luke’s Medical Center, Baylor College of
Medicine, Houston, TX, USA
preoperative renal dysfunction [10, 11]. An
e-mail: wetheridge@[Link] INTERMACS (Interagency Registry for
S.A. Shearer, M.D.
Mechanically Assisted Circulatory Support) risk
Texas Heart Institute at Baylor St. Luke’s Medical score of 1 vs. 2 or 3 predicts poor renal outcomes
Center, Baylor College of Medicine, [12]. Studies have shown that postoperative AKI/
Houston, TX, USA ATN is associated with the following preoperative
e-mail: sshearer@[Link]

© Springer International Publishing AG 2018 129

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
130 W.B. Etheridge and S.A. Shearer

GFR, ≥75.0 GFR, 60.0-74.9 GFR, 45.0-59.9 GFR, <45.0

ml/min/1.73 m2 ml/min/1.73 m2 ml/min/1.73 m2 ml/min/1.73 m2
60
P<0.001
50
Estimated Event Rate (%)

40

30

20

10

0
Death from Reinfarction CHF Stroke Resus- Composite
CV Causes citation End Point

Fig. 9.1  Kaplan-Meier estimates of the rates of death at cardiac arrest, and the composite end point, according to
3 years from cardiovascular (CV) causes, re-infarction, the estimated glomerular filtration rate (GFR) at baseline.
congestive heart failure (CHF), stroke, resuscitation after Modified from the original version [2, 3]

Gene Polymorphism

Inflammatory
BNP/ANP Renal ADH
Cytokines Prostaglandins
venous
RAAS Congestion
( CVP)

↓ Renal
Cardiorenal syndrome Blood Flow

↑↓N.O.

↑ Endothelin

↓ GFR
Catecholamines ↓ K+

Anemia

Fig. 9.2  Congestive heart failure (CHF) and cardiorenal syndrome

9  Management of Fluid Balance and Perioperative Renal Complications 131

Distribution of central venous pressure (CVP) and the relationship between CVP and
estimated GFR in 2557 patients.
80 500

70
eGFR (mL/min/1.73m2) (95% CI)

400
60

Number of patients
50
300

40

200
30

20
100
10

0 0
0 5 10 15 20 25 30
CVP (mmHg)

Fig. 9.3  Effect of increasing central venous pressure tionship between CVP and estimated GFR (eGFR): first
(CVP) on glomerular filtration rate (GFR) in rats with order, Y = −25.8·(CVP + 1)/10 (Wald 28.2, p < 0.0001),
constant blood pressure. The curvilinear model had the and second order, Y = 35.7·([CVP + 1]/10)0.5 (Wald 17.4,
following individual polynomial components for the rela- p < 0.0001). Modified from the original version [4]

CVP and Renal Blood Flow on GFR

120

*†
100 *†
GFR (ml/min/1.73m2 )

80 *

60

40

20
LOW RBF LOW RBF HIGH RBF HIGH RBF
HIGH RAP LOW RAP HIGH RAP LOW RAP

Fig. 9.4  CVP and renal blood flow on GFR. Modified from Damman K et al., Eur J Heart Fail 2007; 9:872–878
132 W.B. Etheridge and S.A. Shearer

Pathophysiology of the relation between venous congestion and reduced glomerular filtration
rate (GFR).
1

High Central Venous Pressure

[11]
[15]
Renal Venous Pressure RAAS ANP
[13] 3
Intrarenal
2 Angiotensin II [24] 7
[18] [19]
4
[23]
Renal Interstitial [22]
Pressure SNS Activation 8

5
[14] [13] [16,17]
[21] [20,21] Hypo-
Hydrostatic
Filtration responsiveness to
Hypoxic pressure in Angiotensin II
Coefficient ANP in CHF
Trigger Bowman’s
capsule

Reduced GFR

Fig. 9.5 Pathophysiology of the relationship between receptor antagonists. (2) Ultrafiltration, diuretics, and
venous congestion and reduced glomerular filtration rate sodium and water restriction. (3) Angiotensin
(GFR). ANP atrial natriuretic peptide, SNS sympathetic II-converting enzyme (ACE) inhibitors and angiotensin
nervous system, RAAS renin-angiotensin-aldosterone II receptor blockers. (4) Statin therapy. (5) Beta-blocker
system. Numbers in circles represent the targets for spe- therapy. (6) Angiotensin II receptor blockers. (7) Neutral
cific therapies as follows: (1) ultrafiltration, diuretics, endopeptidase inhibitors. (8) Urodilatin. Modified from
sodium, and water restriction and arginine vasopressin the original version [5]

characteristics: angiotensin II-converting enzyme Table 9.1  Preoperative factors associated with postop-
inhibitor or angiotensin II receptor blocker use, erative acute kidney injury
renal size <10 cm, older age, small left ventricle, 1. Glomerular filtration rate <60 mL/min
and diastolic dysfunction with high CVP (both 2. INTERMACS score of 1 vs. 2 or 3
likely signs of right heart failure and known asso- 3. Preoperative use of angiotensin II-converting
ciation of diastolic dysfunction with chronic kid- enzyme inhibitors or angiotensin II receptor
ney disease) (Table 9.1) [13–15]. However, in a blockers
review of 100 consecutive continuous-­flow LVAD 4. Renal size <10 cm
(CF-LVAD) implantations, Borgi and colleagues 5. Diastolic dysfunction and high central venous
[14] did not find a statistically significant associa- pressure, likely reflective of right heart dysfunction
tion between postoperative AKI/ATN and preop-
erative diabetes mellitus, hypertension, or renal c­ ardiopulmonary bypass (CPB) time (122 ± 55
dysfunction (7). The difference in these study vs. 78 ± 17 min), higher intraoperative blood loss
findings may be attributed to the later era of the and replacement (>1 L), and need for reoperation
Borgi study, which has been proposed to be a all increase the risk of postoperative AKI/ATN
period when patients received implants earlier in (Table 9.2) [13, 16–19]. Postoperative AKI/ATN
their disease process at a higher INTERMACS is associated with a high risk of mortality, but
score. Operative and perioperative risk factors for death is most likely to occur within the first year
AKI/ATN have been studied, and longer
­ after AKI/ATN, as the survival rate after the first
9  Management of Fluid Balance and Perioperative Renal Complications 133

Table 9.2  Perioperative factors associated with postop- will discuss the successful care of these post-­
erative acute kidney injury
LVAD patients with ESRD by a specialized dial-
1. Cardiopulmonary bypass time >90 min ysis clinic and management team.
2. Blood loss >500–1000 mL Other renal-related syndromes in LVAD
3. Right heart dysfunction patients are worthy of attention in this new and
4. Need for return to the operating room relatively unstudied physiology, which is charac-
terized by minimal pulsatile flow with high dia-
stolic and low systolic pressure and low-grade,
year is not affected by the occurrence of postop- continuous hemolysis. Postoperative mediastinal
erative AKI/ATN [15]. and pericardial tamponade is a cause of sudden
Studies have shown that CF-LVAD placement oliguria. Partial-flow constriction from a clot,
can often lead to postoperative liver and renal pannus over the inflow, or “kinking” of the inflow
recovery, especially for (but not limited to) cannula or the aortic graft that causes acute or
patients with mild preoperative renal dysfunction subacute massive hemoglobinuria in the presence
(e.g., creatinine levels between 1.4 and 1.9 mg/ of reduced cardiac output can cause AKI [43].
dL) [10, 11, 20–30]. Despite an early recovery in Renal or splenic infarcts may develop in patients,
GFR (measured by serum creatinine) in patients accompanied by acute pain syndrome.
after CF-LVAD implantation, a late (i.e., greater Hyponatremia often continues after LVAD
than 1 year post implantation) decline in GFR implantation, and, although not yet studied, cor-
(measured by an increase in serum creatinine) has rection may lead to improved functional status.
been observed. The reason for this decline is
unclear [10, 23, 31–35]. Possible causes include
low muscle mass at the time of implantation with The First 48 h
a subsequent increase in muscle mass, continued
(albeit less intense) neurohumoral and inflamma- Performing LVAD surgery generally requires the
tory cytokine stimulation, and hypertensive dam- patient to undergo CPB. As in all cardiovascular
age from the “new physiology” of chronic high surgeries, limiting CPB time to <90 min can help
diastolic perfusion. Animal and human studies prevent postoperative AKI [13]. Perioperative
have shown an abnormal inflammatory response bleeding is a confounding complication with a
in the arterial wall after exposure to a CF-LVAD progressive increase in the risk of AKI with
[36–43]. Notably, at higher pump speeds, low-­ >500–1000 mL of blood loss and replacement
grade continuous hemolysis occurs, which has [13]. Right heart dysfunction after LVAD place-
been proposed as a cause of chronic hemoglobin- ment is a common risk factor for postoperative
uria, reduced availability of nitric oxide, and oxi- AKI because of the reduced cardiac output and
dative stress with peritubular inflammation [16, high venous pressure [6, 7, 14]. Pulmonary hyper-
18]. In an ovine model, micro-emboli have been tension frequently develops in patients with
seen in the renal microvasculature [17]. chronic heart failure; in fact, an LVAD is usually
Most contemporary pumps are continuous-­ placed as DT or as a “bridge to candidacy” in
flow assist devices. Therefore, this chapter will patients with increased pulmonary vascular resis-
focus on the care of patients with renal dysfunc- tance (>4 WU) to allow pulmonary pressure to
tion after placement of a CF-LVAD. Currently, normalize so that transplantation can be consid-
ESRD is an exclusion criterion for destination ered. Right ventricular failure causes high right-­
therapy (DT) LVAD placement, so patients with sided pressure and increased CVP. Right heart
ESRD have been approved for LVAD placement failure may be affected by the LVAD itself, as
only if they have been approved for dual-organ discussed in Chap. 18. The septum may be dis-
transplant or bridge to transplant (BTT). The placed to the left, adversely affecting right ven-
number of patients in whom ESRD develops tricular function. This is dependent on the LVAD
after LVAD implantation for DT has grown. We pump speed and its effect on the anatomy of the
134 W.B. Etheridge and S.A. Shearer

left ventricle. Ideally, to help protect the septum, [48] have shown improved outcomes in patients
the pump is adjusted so that the aortic valve opens undergoing abdominal vascular surgery with the
with each beat and the left ventricle remains use of a fluid-restricted protocol [45]. In addition,
mildly dilated. High venous return with increased at CHI/St. Luke’s Texas Heart Institute, we are
pump flow may further overload the right ventri- initiating a trial of a fluid-sparing regimen in
cle and cause dilation and strain, especially if CF-LVAD patients. The goal PA pressure is
there is not a concomitant reduction in the pulmo- <45 mmHg, and the goal CVP is initially
nary pressure and cardiac output. The end result <12 mmHg and is reduced to <10 mmHg once
may be right ventricular failure and increased the patient is hemodynamically stable. Milrinone
central and renal venous pressure. This outcome is often required to support right ventricular
may be further complicated by tricuspid regurgi- function and to lower PA pressure. In patients
tation, which is common in chronic heart failure, with a low GFR, which affects milrinone clear-
especially in patients with high preoperative ance, drug accumulation may cause low systemic
pulmonary artery (PA) pressure. Congestive and renal perfusion pressure. For patients with a
hepatopathy and nephropathy are also common low GFR (<30–40 mL/min/1.73 m2) or low urine
and may worsen renal function and cause a output (<0.5 mL/kg/h), we prefer to reduce the
poor diuretic response (Figs. 9.3, 9.4, and 9.5) milrinone dose to <0.25 μg/kg/min, but we con-
[5–8, 14]. fer regularly with the cardiology team (Table 9.3)
Low systemic blood pressure in postoperative (Figs. 9.6 and 9.7) [50].
LVAD patients compounds right heart failure and Although diuretics generally are not recom-
high central and renal venous pressures. Causes mended early because of the risk of venous dila-
of low systemic blood pressure include sedation, tation, low blood pressure, and reduced LV
pain medications, and frequent use of vasodila- filling, loop diuretics may be necessary to prevent
tory inotropes and pulmonary vasodilators. A severe volume overload, especially in patients
continuous-flow pump cannot pump against high with right heart dysfunction [51]. Judicious fluid
pressure, so it is important to keep the systemic management, however, should be the mainstay of
pressure high enough to provide adequate renal fluid therapy.
perfusion, taking the adverse effects of high In patients with fluid overload, pulmonary
venous pressure into account. Care should be congestion, and high right-sided filling pressures,
taken, however, not to increase the systemic pres- we initiate loop diuretics while minimizing fluid
sure high enough to reduce output from the intake. The latter requires coordination with the
continuous-­flow pump [44]. pharmacy and members of the intensive care unit
The nephrologist should confer closely with (ICU) team. A low dose of a loop diuretic (20–
the intensivist, cardiologist, and LVAD surgeon. 40 mg of furosemide or 0.5–1 mg of bumetanide)
The best plan for accomplishing this is to per- is given, while the patient’s response and blood
form daily team rounds. pressure are monitored. The dose can then be
In the immediate postoperative period, we
minimize the use of casual fluid (i.e., fluid given
as a carrier for drips, medications, and electro- Table 9.3  Fluid therapy choice in the AKI ICU
lytes) and administer therapeutic fluid according Alternatives
to a weight-based protocol by using a balanced Drug Hazard/disadvantages
electrolyte solution with some bicarbonate equiv- 0.9% saline Acidosis, ?AKI
alent rather than normal saline. Postoperative Lactated Ringer’s Hypotonic, Ca++
volume-related weight gain has been associated Plasmalyte Gluconate, acetate
with poor outcomes [45], and fluid-restricted Albumin Cost, ?AKI
protocols in the postoperative period have been HES AKI, bleeding, pruritus
associated with either worse [46] or improved Gelatin Anaphylaxis
[47] outcomes. González-Fajardo and colleagues AKI acute kidney infection
9  Management of Fluid Balance and Perioperative Renal Complications 135

Outcomes

Mortality
Gastrointestinal
Renal
Cardiac
Respiratory
Major hemorrhage
Major infection
Composite
0.1 1 10
Favors balanced crystalloid Odds ratio Favors 0.9% saline

Fig. 9.6  Favor-balanced fluid not NS. Modified from Shaw, et al. Ann Surg. 2012 Mar 30

Fig. 9.7 Relationship Relationship between chloride and survival

between change in
chloride and survival 800
time. Modified from
700
original [49] R2 = 0.3697
Survival Time (min)

600 p < 0.0001

500
400
300
200
100
0
0 5 10 15 20 25 30
Delta Chloride (mEq/L)

titrated either with higher intermittent doses or need to make pump speed adjustments. These
with a continuous drip (5–20 mg/h of furosemide adjustments may be made under echocardio-
or 0.25–1 mg/h of bumetanide). For patients who graphic or PA catheter guidance. The next step
do not respond, we confer with the team to ensure that we have found helpful is the addition of a
that no other cause for low renal perfusion pres- distal tubule blocker (usually chlorothiazide,
sure can be identified (e.g., excessive sedation, 250–500 mg administered intravenously). Loop
pain medication, vasodilators such as milrinone, diuretics and distal tubular blockers cause alkalo-
or tamponade). A lack of response, again, may sis and hypokalemia. We use potassium chloride,
indicate a poor right heart function or a need to carbonic anhydrase inhibitors, or amiloride in
increase the LV output, and the cardiologist may this situation. Amiloride has a shorter half-life
136 W.B. Etheridge and S.A. Shearer

Table 9.4  Ideal hemodynamics for renal function Table 9.6  Acute oliguria after continuous-flow LVAD
implantation: action plan
1. Central venous pressure <10 mmHg
2. Pulmonary pressure <45 mmHg 1. Ensure that the bladder is decompressed
3. “Mean arterial pressure” 70–80 mmHg 2. Look for a sudden decrease in hemoglobin levels
4. Adequate cardiac output to ensure stable and 3. Confer with the ICU team to adjust pump or drips
normal end-organ function; patient should be 4. Perform chest radiography: computed tomography
awake, alert, and neurologically stable (confer with or echocardiography
cardiology) 5. Check urine indices, if appropriate
6. Look for nephrotoxic drugs
Table 9.5  Acute oliguria after continuous-flow LVAD
implantation
1. Lower urinary tract obstruction
2. Severe right heart dysfunction with decreased Acute Oliguria Post Continuous-Flow LVAD
cardiac output 1. Lower urinary tract obstruction
3. Tamponade due to mediastinal bleeding 2. Severe RH dysfunction with decrease in car-
4. Bleeding (look for hemothorax) diac output
5. Pump malfunction, inflow or outflow obstruction 3. Tamponade/due to mediastinal bleeding
(rare in the postoperative period)
4. Bleeding/look for hemothorax
6. Sepsis- or drug-induced hypotension (consider 5. Pump malfunction/inflow or outflow obstruc-
milrinone)
tion—rare in post-op period
6. Sepsis or drug-induced hypotension—con-

and a more immediate onset than mineralocorti- sider milrinone
coid inhibitors and will ameliorate alkalosis and
hypokalemia as well. If the above-described measures do not improve
From a renal perspective, ideal postoperative the patient’s condition, intervention with renal
values include a CVP of <10 mmHg, PA pressure replacement is necessary. For patients who remain
of <45 mmHg with the LVAD adjusted so that the on the ventilator and need fluid removal, we prefer
valve is opening, and a mean systemic pressure to use continuous renal replacement therapy
of 70–80 mmHg, as well as an even fluid balance (CRRT), which allows more continuous ultrafiltra-
(Table 9.4). tion (UF) adjustment with excellent clearance. At
A sudden reduction in urine output should trig- our institution, we often keep the patient on CRRT
ger suspicion of bleeding, especially bleeding into in the operating room to control fluid volume and
the mediastinum with functional tamponade or electrolytes. Our ICU nurse, who is familiar with
into the pleural space with a sudden reduction in the patient, manages CRRT in the operating room.
cardiac output or mean pressure. Tamponade will Adjustments are made by the anesthesia team, but
usually be associated with increased central pres- the nephrologist is always available.
sures, but this is occasionally subtle and may cause We have found that the blood flow rate does
low urine output, even in the setting of a minimal not affect blood pressure or hemodynamic stabil-
change in PA pressure and CVP. Right heart dys- ity, so we recommend adjusting this rate as close
function is always a consideration; thus, we fur- to 300 mL/min as possible to prevent system
ther emphasize the importance of conferring with thrombosis. The clearance, which is controlled
the LVAD team (Table 9.5). Often, an adjustment by the dialysate flow and the dialysate content, is
of inotropic drugs or pump speed is successful. based on the concentrations of blood urea nitro-
However, the patient may require reoperation for gen, creatinine, and potassium, as well as the
bleeding and pericardial decompression or place- acid-base status, just as in the standard ICU
ment of a right ventricular support device, but we patient. However, we prefer a dialysate and UF
have a standard action plan (Table 9.6). fluid flow of at least 35 mL/kg/h (Fig. 9.8).
9  Management of Fluid Balance and Perioperative Renal Complications 137

Fig. 9.8  Dose of CVVH Dose of CVVH in AKI

in AKI. Modified from
[52]
100%
* p < 0.001 v. 20 mL/kg/hr
80%
57%* 58%*

Survival
60%
41%
40%

20%

0%
20 mL/kg/hr 35 mL/kg/hr 45 mL/kg/hr
Ultrafiltration Rate

Table 9.7  Benefits of nocturnal prolonged intermittent Continuous-Flow Physiology

renal replacement therapy for the Nephrologist and Dialysis
Early mobilization during the day when PT/OT staff Nurse
are available
Radiological or surgical procedures can be performed In a CF-LVAD patient, blood pressure does not have
during the day
the same meaning as it does in a patient without an
Less anxiety for patients
LVAD. Although we typically monitor and record a
“mean pressure” or, in some patients, an actual sys-
Potassium replacement should be given by tolic pressure, we do so with the understanding that
an enteral route if the patient has a feeding tube there is not an actual mean blood pressure. Because
and the gastrointestinal tract is functional. the flow is continuous, there is no true diastole. The
Using this route reduces the amount of fluid physiologic measures we are interested in for
that needs to be removed. Venous access is CF-LVAD patients are the degree of filling of the
obtained through a standard percutaneous line, left ventricle and the systemic flow. Similar to the
unless the patient is on extracorporeal mem- non-­LVAD patient, less blood is available for sys-
brane oxygenation (ECMO), in which case the temic perfusion as the volume in the left ventricle
arterial and venous lines can be placed in line decreases. The Doppler and the audible or palpable
with the ECMO circuit. The nephrologist pulse are monitored to detect a change in the flow
should monitor the patient closely because UF with a ventricular contraction. In the case of the pal-
and dialysate needs can change frequently. The pable or audible pulse, there is enough blood in the
cardiologist and the intensivist should be ventricle and a strong enough contraction to open
allowed to alter the UF rate on a short-term the valve (note that the palpable pulse and the
basis, but we recommend that the nephrologist Korotkoff sounds are from the closure of the aortic
see the patient at least twice per day and closely valve). When the valve opens, blood flows through
coordinate care with the other members of the the pump and also from the ventricle through the
ICU team. Once the patient is stable (and espe- aortic valve (parallel flow). When the contraction is
cially if the patient is off mechanical ventila- very weak or the ventricle is less full—which can
tion), we recommend initiating periodic occur during UF—the valve may remain closed, but
intermittent renal replacement, shift dialysis, a Doppler pulse can be detected because the ventri-
sustained low-efficiency dialysis (SLEDD), or cle continues to contract and increase input (albeit
standard hemodialysis (SHD) (Table 9.7). less) to the pump with each beat. Only two values
138 W.B. Etheridge and S.A. Shearer

on the LVAD monitor are actually measured and “­ suction event” that can cause ventricular tachy-
displayed on the pump: the speed in revolutions per cardia. The current hypothesis is that this occurs
minute (RPMs) and the power in watts. Power is because the ventricle is empty and the inlet can-
increased by the amount of flow across the pump or nula touches the ventricular wall.
by an increase in pump speed and is decreased by a The nurse can sometimes palpate or auscultate
decline in the flow across the pump or a decline in a pulse (as noted above) if the valve is opening,
speed. The flow is derived from a formula that is or there may be an arterial line to record a sys-
based on the power (because the power is depen- tolic pressure. More often, the arterial line has
dent on flow) in both the HeartMate II (HMII) and been removed, and the nurse will need another
the HeartWare ventricular assist device (HVAD). way to detect a change in the left ventricle filling
The HMII also displays the pulsatility index (PI), a (i.e., blood pressure). This can be done by using
value that can be helpful to the dialysis team. a standard pressure cuff, generally placed on the
A variation in power can be caused by an arm in the usual fashion, to detect the first appear-
increase or decrease in the speed set by the cardi- ance of an audible or palpable pulse or by detect-
ology team, a change in the contractile force of ing a Doppler signal, if the valve is not opening.
the ventricle, a change in the systemic vascular To monitor the patient during dialysis, the power
resistance (which can increase or decrease the and—indirectly—the flow and PI (in the HMII)
afterload), or a change in the fluid volume deliv- are watched closely, along with the blood pres-
ered to the pump through the left ventricle. An sure. Small reductions in power, flow, and PI (in
increase in power or flow will occur rhythmically the HMII) are expected during UF. We generally
with a ventricular contraction (whether the aortic accept a 5–10% reduction in power or PI as nor-
valve opens or not), when a fluid bolus is given or mal, but this variation is dependent on the patient
when the patient is overloaded with fluid. In the and the clinical situation. The nephrologist and
HMII, the variation in power with each cardiac the cardiologist will need to confer on these lim-
cycle is measured, and a larger power change is its and discuss parameters with the dialysis nurse.
reflected by a higher PI. During the UF process After the patient has achieved a physiologic dry
and as the ventricle is emptied, we can anticipate weight, patients generally remain stable on renal
that less blood will be available with each beat. replacement therapy. We usually set a “mean
Less augmentation of the power during ventricu- pressure” limit of 60–70 mmHg for patients on
lar systole vs. diastole may result in lower power SHD, shift dialysis, or SLEDD, but this limit is
overall, a lower power variation from beat to variable. A low-flow alarm or a suction event
beat, and a lower PI for the HMII, and just lower with ventricular tachycardia requires urgent
power for the HVAD. attention. The patient should be given 3–5 mL/kg
of intravenous fluids, and the nephrologist and
the cardiologist should be notified. Ventricular
 LEDD, Shift Dialysis, and SHD
S tachycardia is usually tolerated for a short time
in the LVAD Patient and generally resolves with volume replacement.
Chest compressions should not be performed.
During the dialysis process, the primary concerns Postoperative patients who are still taking pain
of the nephrologist and the dialysis nurse are the medication and may also be on vasodilatory ino-
changes in LV filling and systemic flow reflected tropes have low systemic vascular resistance, mak-
in the “pressure.” The pump speed is set by the car- ing UF difficult without a pressor. Using a dialysate
diology or surgical team. Dialysis nurses add or with a sodium concentration of 145 meq/L and/or
increase inotropes or adjust vasopressors only with a calcium concentration of 3 meq/L will allow
a direct order from a physician on the care team. A more UF by increasing cardiac output, vasocon-
decrease in power generally signifies reduced flow striction and better right ventricular filling, and
across the pump; if the power falls below a set blood pressure. We use this “modeled” dialysate
point, then a “low-flow” alarm activates. In extreme on a short-term basis until the patient is stable, and
cases, the patient can have what is termed a then we switch to a more standard dialysate.
9  Management of Fluid Balance and Perioperative Renal Complications 139

Patients may be on pressor agents, but, as noted, replacement should be administered immediately,
these agents are generally adjusted by a physician and UF should be stopped. In some patients, a
only unless adjustment is part of the order set. “suction event” may occur with a brief episode of
Two situations warrant special consideration. ventricular tachycardia that is usually terminated
First, patients with severe aortic insufficiency with the fluid treatment. The nurse should notify
may have the aortic valve closed at the time of the cardiologist and nephrologist. Because many
operation and, therefore, will have pulsatility patients are still sedated or on pain medications
through the pump but no dicrotic notch on the and may have low systemic vascular resistance,
arterial tracing. As a result, these patients will we have sometimes found it necessary to use low-
generally not have a palpable pulse or Korotkoff dose pressor agents to prevent peripheral blood
sounds, and the nurse will need to rely on the pooling and to maintain LV return. This is an
Doppler pressure reading. Second, some patients important option to consider in the ICU, but we
will need biventricular support. As with one always confer with the cardiologist first. In addi-
pump, the speed of both pumps will be set to opti- tion, during dialysis and UF, some patients may
mize the hemodynamics of the patient. Over-­ benefit from a small increase in pump speed,
pumping of the right side can cause pulmonary which is set by the cardiologist only. Alternatively,
congestion or a perfusion injury of the lungs. some patients because of a small-sized left ventri-
Once the patient is stable, the guidelines for mon- cle or unusual inflow cannula may require a reduc-
itoring the patient are essentially the same as tion in pump speed during treatment to prevent
those previously described for patients with an ventricular collapse and recurrent low flow or suc-
LVAD only. tion events and accompanying symptoms of syn-
Most patients are already anticoagulated; cope. Again this is coordinated through the LVAD
therefore, we do not add an anticoagulant. This team. After the patient is out of the ICU, using
aspect of care is managed by the cardiology team. dialysate with increased sodium concentration,
In summary, the dialysis nurse should monitor sodium modeling, or even increased calcium con-
the degree of LV filling by using the tools avail- centration may be necessary on a short-term basis.
able. This includes monitoring the pressure with The potassium is adjusted on the basis of the
the standard cuff, palpation, or auscultation, the patient’s pretreatment potassium level (as usual).
Doppler pulse, and, if discussed with the nephrol- Using the above plan, extending the standard
ogist or team, the power, flow, or PI (in the HMII). treatment to 4.5–5 h is sometimes required to
Reduced power, flow, or PI indicates reduced LV achieve the necessary UF, especially in patients
filling. When writing the dialysis orders, the with right heart dysfunction or a small ventricle.
nephrologist should instruct the nurse on the Although many patients have recovered after
expected UF volume and hemodynamic parame- 30 days, others continue to require renal replace-
ters. Although an expected “pressure” value is ment therapy. We declare these patients as having
determined on the basis of experience with the ESRD and prepare them for outpatient dialysis.
individual patient, the “mean pressure” is typi- Late recovery, however, is possible, and we con-
cally kept at >55–65 mmHg. Occasionally, espe- tinue to monitor their renal function and urine
cially early in the initiation of intermittent output.
treatment, we may ask the nurse to also monitor
the power or PI for a sudden decrease and notify
the nephrologist or the cardiologist of any con-  ransitioning the Patient
T
siderable (i.e., 20–30%) change. to Outpatient Dialysis and ESRD
If the low-flow alarm is activated, this indicates Care
an urgent situation in which the LV volume has
become dangerously low (see above), and the The literature on ESRD in LVAD patients is
power and therefore the measured flow have sparse, but we expect this unique patient popula-
decreased to critically low levels (the “low-flow” tion to increase in number. The expected increase
alarm threshold is set by the LVAD team). Volume is attributed to the growing number of LVADs
140 W.B. Etheridge and S.A. Shearer

Table 9.8  End-stage renal disease (ESRD) in patients insufficient disease education—contributes to
after acute kidney injury (mean age <60 years)
poor compliance, missed treatments, and large
Author Journal Patients ESRD (%) fluid weight gains between treatments. Eventually
Schmidt R AJKD 2008 Meta-­ 31 the patients realize that their kidneys will not
analysis
recover. In these patients, there is a high inci-
Thakar CV AJKD 110 70
dence of psychosocial complications such as
Palevsky PM NEJM 533 66
survived depression and disrupted family structure, which
Kurella NEJM 3702 >24 is not surprising given that education is priori-
Tamura M tized in LVAD care at the expense of poor ESRD
education. We have adopted a multidisciplinary
team approach that includes working with the
implanted as BTT in patients with ESRD await- LVAD team, nephrologists trained in LVAD care,
ing dual-organ transplant and LVADs implanted and ESRD social workers and prioritizing referral
as DT in patients who develop ESRD after ATN of patients to a designated LVAD/dialysis clinic.
post-op CF-LVAD implant (an estimated 50% Educating patients before they are discharged has
risk). This is in line with studies performed in the played a crucial role in promoting both dialysis
population of surgically treated patients as a awareness and acceptance.
whole (Table 9.8). In addition, the number of Most dialysis patients are referred for SHD.
patients living with an LVAD is expected to Although peritoneal dialysis has been reported to
increase. Thus, more patients will be experienc- have satisfactory outcomes, experience with this
ing a slow decline in renal function over time— technique has been limited. Theoretically, perito-
even after initial improvement after LVAD neal dialysis may have several advantages, but
implantation—and will need long-term renal this form of treatment has not been well studied
replacement therapy. As the population of patients and will most likely be limited to patients with an
with ESRD grows and ages, more patients will intrathoracic pump or a pre-peritoneal-positioned
present with or develop cardiac disease and will LVAD [51, 53].
be referred for dual-organ transplant. These
patients may require an LVAD as BTT and will
continue to need outpatient dialysis. Access
An increase in the number of LVAD patients
with ESRD will pose unique challenges in dis- Dialysis is generally initiated with a temporary
ease management. ESRD is complex, and patients catheter that is transitioned to a permanent cath-
can face psychological barriers. Many patients eter, tunneled into the internal jugular vein. A
arrive at the outpatient clinic with the belief that permanent arteriovenous access, however, may
their renal dysfunction was caused by heart dis- have been placed by the time of outpatient plan-
ease and that they will recover, when in fact (as ning and discharge. The patient, patient’s family,
noted above) there may be only a 50% chance of and health-care providers should be educated
recovery. If patients reach the point of discharge about the risks of intravenous catheters. This con-
without recovery, then the chance of recovery cern may be heightened by the theoretical risk of
may be even lower. The overly optimistic expec- pump infection, but this has not been well stud-
tation of patients can be exacerbated by optimism ied, and actual pump infection is rare.
of the LVAD team and the nephrologist for even- For outpatient ESRD treatment, including the
tual renal recovery. The most important result of treatment of LVAD patients with ESRD, the ideal
this is contribution to delayed or poor education access is an arteriovenous (AV) fistula (AVF). We
of the patient and family in ESRD care and man- avoid placing an AV access on the side of an
agement. In addition, patients often interpret automatic implantable cardioverter-defibrillator
urine output as evidence of impending renal because there is a risk of angioedema. In a review
recovery, which—along with the above mentioned by Patel et al. [1], the authors discuss the issue of
9  Management of Fluid Balance and Perioperative Renal Complications 141

access in LVAD patients with ESRD and con- in a review by Patel et al. [1] When combined
clude that an AV graft is the best option for these with a standard patient assessment, monitoring
patients, not an AVF. The authors base their rec- with an automated cuff is sufficient for most
ommendation on the theoretical concern that the patients. Patient monitoring during dialysis can
thin walls of the native vein would not receive the be adjusted on an individual basis and usually
typical pulsatile flow from an artery and, as a remains very stable. The disappearance of pulsa-
result, would not mature in a CF-LVAD patient. tility may indicate that the UF rate needs to be
Our experience, however, does not agree with reduced, but in our experience, patients usually
this conclusion. We have seen satisfactory do not become unstable.
­maturation of fistulas, even in patients with low Patient behavior and increased fluid and salt
pulsatility, so fistula remains our preferred access. intake have presented more of a challenge than
Nonetheless, we have not had a fistula in a patient the actual physiologic nuances of monitoring the
with a surgically closed aortic valve. Even in patient’s blood pressure. Just as in non-LVAD
patients who do not have a palpable pulse, our patients, the standard assessment of pretreatment
nurses have been able to palpate the fistula by vir- volume in LVAD patients is critical. An increased
tue of the bruit, and cannulation has not been a rate of fluid removal reduces pump power and
serious obstacle. Coagulation plays an unknown flow. Just as in non-LVAD patients, the rate of UF
role in the patency of the fistula, but the failure of needs to be controlled. UF should be limited to
warfarin to maintain patency in patients with 10 mL/kg/h. Patients are instructed that if they
multiple access thromboses suggests that the role gain more than their allowable weight goal, dial-
of anticoagulation may be minimal. Some ysis time will need to be increased accordingly.
patients are off anticoagulation because of gas- There are three special circumstances that
trointestinal bleeding, and they are treated as any occur infrequently but may need additional atten-
other patient with an AVF who is not on chronic tion. Patients with aortic insufficiency may
anticoagulation. The requirement for warfarin in require surgical closure of the aortic valve, in
most patients with a CF-LVAD complicates the which case there will be no palpable or audible
placement of AV access because this surgery pulse. For these patients, we recommend that a
requires heparin “bridging” so that the patient is Doppler instrument be used to detect and monitor
anticoagulant-free on the day of surgery. a “mean pressure,” understanding again that that
Therefore, as part of our education and prepara- there is not a true mean pressure in CF-LVAD
tion process, we place an AV access before the patients. The same may be true in the group of
patient’s original discharge. Our surgeons are patients who have poor LV function and an aortic
trained in access placement in these patients. valve that does not open regularly. Occasionally,
a patient will require a right and left continuous-­
flow pump. In general, these patients are stable,
Dialysis Monitoring and it is recommended that they be treated the
same as the abovementioned patients, whether
In the literature, it has been suggested that blood there is valve opening or not. Finally, LVAD
pressure monitoring in CF-LVAD recipients is patients will occasionally have continuous ven-
either inaccurate or not feasible. This would pres- tricular fibrillation or tachycardia. These patients
ent an obstacle for patients undergoing outpatient will generally have a closed aortic valve and
dialysis because patient monitoring would be should be treated accordingly. In our experience,
much more time consuming and would require these patients are stable on dialysis and can be
the added expense of using a Doppler machine. assessed with a standard exam, taking into
Most patients have enough pulsatile flow to allow account the skin color, level of consciousness,
for automated blood pressure readings. This, of and presence of diaphoresis. We have not had a
course, requires some degree of residual LV patient with continuous ventricular tachycardia or
function and valve opening and closing, as noted fibrillation on outpatient dialysis. Communication
142 W.B. Etheridge and S.A. Shearer

with the LVAD team is critical. We have a desig- 3. Anavekar NS, McMurray JJ, Velazquez EJ, et al.
Relation between renal dysfunction and cardiovas-
nated case manager for our LVAD patients on out-
cular outcomes after myocardial infarction. N Engl J
patient dialysis. Med. 2004;351(13):1285–95.
The dialysis team should be familiar with the 4. Firth JD, Raine AE, Ledingham JG. Raised venous
increased risk of gastrointestinal bleeding, primar- pressure: a direct cause of renal sodium retention in
oedema? Lancet. 1988;1(8593):1033–5.
ily from AV malformations of the small bowel.
5. Damman K, van Deursen VM, Navis G, Voors AA,
Any sudden decrease in hemoglobin should alert van Veldhuisen DJ, Hillege HL. Increased central
the renal team and prompt a referral to the LVAD venous pressure is associated with impaired renal
team managers. We initiate our erythropoiesis- function and mortality in a broad spectrum of patients
with cardiovascular disease. J Am Coll Cardiol.
stimulating agent protocol in LVAD patients if
2009;53(7):582–8.
they are on anticoagulation. For those not on anti- 6. Cruz DN, Schmidt-Ott KM, Vescovo G, et al.
coagulation, we discuss the case with the cardiol- Pathophysiology of cardiorenal syndrome type 2 in
ogy/LVAD team because there may be some stable chronic heart failure: workgroup statements
from the eleventh consensus conference of the Acute
increased risk of thrombosis.
Dialysis Quality Initiative (ADQI). Contrib Nephrol.
In conclusion, hemodialysis and—though 2013;182:117–36.
experience with it has been limited—peritoneal 7. McCullough PA, Kellum JA, Haase M, et al.
dialysis are satisfactory treatment modalities for Pathophysiology of the cardiorenal syndromes: exec-
utive summary from the eleventh consensus confer-
CF-LVAD patients in whom ESRD develops
ence of the Acute Dialysis Quality Initiative (ADQI).
after implantation or for CF-LVAD patients who Contrib Nephrol. 2013;182:82–98.
are already on dialysis (currently only patients 8. Mullens W, Abrahams Z, Francis GS, et al. Importance
approved for dual-organ transplantation). The of venous congestion for worsening of renal function
in advanced decompensated heart failure. J Am Coll
dialysis team should have a basic understanding
Cardiol. 2009;53(7):589–96.
of continuous-flow physiology, and it is prefera- 9. Vallet M, Metzger M, Haymann JP, et al. Urinary
ble to have clinics and case managers dedicated ammonia and long-term outcomes in chronic kidney
to the care of these patients. Monitoring the disease. Kidney Int. 2015;88(1):137–45.
10. Sandner SE, Zimpfer D, Zrunek P, et al. Renal func-
hemodynamics of patients on dialysis has not
tion and outcome after continuous flow left ventric-
been difficult, with the standard exam and assess- ular assist device implantation. Ann Thorac Surg.
ment before treatment and the use of the standard 2009;87(4):1072–8.
blood pressure cuff to monitor the “mean pres- 11. Haglund NA, Feurer ID, Dwyer JP, et al. Does renal
dysfunction and method of bridging support influ-
sure.” Our preferred access is AVF. Patient edu-
ence heart transplant graft survival? Ann Thorac Surg.
cation and training before the initial discharge 2014;98(3):835–41.
should include information on dialysis and 12. Yoshioka D, Sakaguchi T, Saito S, et al. Predictor of
ESRD, in addition to standard LVAD training. early mortality for severe heart failure patients with
left ventricular assist device implantation: signifi-
Patient satisfaction is good, with a low hospital-
cance of INTERMACS level and renal function. Circ
ization rate after CF-LVAD implantation. Some J. 2012;76(7):1631–8.
patients have chosen to remain on dialysis instead 13. Alba AC, Rao V, Ivanov J, Ross HJ, Delgado

of pursuing dual-organ transplantation. DH. Predictors of acute renal dysfunction after
ventricular assist device placement. J Card Fail.
2009;15(10):874–81.
14. Borgi J, Tsiouris A, Hodari A, Cogan CM, Paone
References G, Morgan JA. Significance of postoperative acute
renal failure after continuous-flow left ventricu-
1. Patel UD, Hernandez AF, Liang L, et al. Quality of lar assist device implantation. Ann Thorac Surg.
care and outcomes among patients with heart failure 2013;95(1):163–9.
and chronic kidney disease: a get with the guide- 15. Topkara VK, Dang NC, Barili F, et al. Predictors and
lines—Heart Failure Program study. Am Heart J. outcomes of continuous veno-venous hemodialy-
2008;156(4):674–81. sis use after implantation of a left ventricular assist
2. Berl T, Henrich W. Kidney-heart interactions: epide- device. J Heart Lung Transplant. 2006;25(4):404–8.
miology, pathogenesis, and treatment. Clin J Am Soc 16. Burke MA, Givertz MM. Assessment and manage-
Nephrol. 2006;1(1):8–18. ment of heart failure after left ventricular assist device
implantation. Circulation. 2014;129(10):1161–6.
9  Management of Fluid Balance and Perioperative Renal Complications 143

17. Cooper TK, Zhong Q, Nabity M, Rosenberg G, Weiss 31. Deo SV, Sharma V, Altarabsheh SE, et al. Hepatic and
WJ. Use of urinary biomarkers of renal ischemia in a renal function with successful long-term support on
lamb preclinical left ventricular assist device model. a continuous flow left ventricular assist device. Heart
Artif Organs. 2012;36(9):820–4. Lung Circ. 2014;23(3):229–33.
18. Moreno JA, Martin-Cleary C, Gutierrez E, et al. AKI 32. Feitell S, Patel J, Pirlamarla P, et al. Type 2 cardio-
associated with macroscopic glomerular hematuria: renal syndrome in the left ventricular assist device
clinical and pathophysiologic consequences. Clin J (LVAD) population. J Heart Lung Transplant.
Am Soc Nephrol. 2012;7(1):175–84. 2013;32(4):S227.
19. Morgan JA, Brewer RJ, Nemeh HW, et al. 547:
33. Hasin T, Topilsky Y, Schirger JA, et al. Changes in
impact of acute renal failure on survival after HM renal function after implantation of continuous-flow
II LVAD implantation. J Heart Lung Transplant. left ventricular assist devices. J Am Coll Cardiol.
2010;29(2):S178. 2012;59(1):26–36.
20. Aaronson KD, Patel H, Pagani FD. Patient selection 34. Jacobs S, Droogne W, Waelbers V, et al. Evolution
for left ventricular assist device therapy. Ann Thorac of renal function after partial and full mechanical
Surg. 2003;75(6):S29–35. support for chronic heart failure. Int J Artif Organs.
21. Bank AJ, Mir SH, Nguyen DQ, et al. Effects of left 2014;37(5):364–70.
ventricular assist devices on outcomes in patients 35. Lok SI, Martina JR, Hesselink T, et al. Single-centre
undergoing heart transplantation. Ann Thorac Surg. experience of 85 patients with a continuous-flow
2000;69(5):1369–74; discussion 1375. left ventricular assist device: clinical practice and
22. Burnett CM, Duncan JM, Frazier OH, Sweeney MS, outcome after extended support. Eur J Cardiothorac
Vega JD, Radovancevic B. Improved multiorgan Surg. 2013;44(3):e233–8.
function after prolonged univentricular support. Ann 36. Amir O, Radovancevic B, Delgado RM 3rd, et al.
Thorac Surg. 1993;55(1):65–71; discussion 71. Peripheral vascular reactivity in patients with pulsa-
23. Butler J, Geisberg C, Howser R, et al. Relationship tile vs axial flow left ventricular assist device support.
between renal function and left ventricular assist J Heart Lung Transplant. 2006;25(4):391–4.
device use. Ann Thorac Surg. 2006;81(5):1745–51. 37. Bartoli CR, Giridharan GA, Litwak KN, et al.

24. Demirozu ZT, Etheridge WB, Radovancevic R,
Hemodynamic responses to continuous versus pulsa-
Frazier OH. Results of HeartMate II left ventricular tile mechanical unloading of the failing left ventricle.
assist device implantation on renal function in patients ASAIO J. 2010;56(5):410–6.
requiring post-implant renal replacement therapy. J 38. Grosman-Rimon L, Tumiati LC, MA MD, et al.

Heart Lung Transplant. 2011;30(2):182–7. (194)—chronic inflammation in heart failure patients
25. Farrar DJ, Hill JD. Recovery of major organ func- with mechanical circulatory support. J Heart Lung
tion in patients awaiting heart transplantation Transplant. 2014;33(4):S76–7.
with Thoratec ventricular assist devices. Thoratec 39. Kihara S, Litwak KN, Nichols L, et al. Smooth muscle
Ventricular Assist Device Principal Investigators. J cell hypertrophy of renal cortex arteries with chronic
Heart Lung Transplant. 1994;13(6):1125–32. continuous flow left ventricular assist. Ann Thorac
26. Frazier OH, Rose EA, Oz MC, et al. Multicenter Surg. 2003;75(1):178–83; discussion 183.
clinical evaluation of the HeartMate vented elec- 40. Nishinaka T, Tatsumi E, Nishimura T, et al. Change in
tric left ventricular assist system in patients await- vasoconstrictive function during prolonged nonpulsa-
ing heart transplantation. J Thorac Cardiovasc Surg. tile left heart bypass. Artif Organs. 2001;25(5):371–5.
2001;122(6):1186–95. 41. Ohnishi H, Itoh T, Nishinaka T, et al. Morphological
27. Kamdar F, Boyle A, Liao K, Colvin-adams M, Joyce changes of the arterial systems in the kidney under
L, John R. Effects of centrifugal, axial, and pulsa- prolonged continuous flow left heart bypass. Artif
tile left ventricular assist device support on end- Organs. 2002;26(11):974–9.
organ function in heart failure patients. J Heart Lung 42. Ootaki C, Yamashita M, Ootaki Y, et al. Reduced pul-
Transplant. 2009;28(4):352–9. satility induces periarteritis in kidney: role of the local
28. Khot UN, Mishra M, Yamani MH, et al. Severe
renin-angiotensin system. J Thorac Cardiovasc Surg.
renal dysfunction complicating cardiogenic shock 2008;136(1):150–8.
is not a contraindication to mechanical support as a 43. Segura AM, Gregoric I, Radovancevic R, Demirozu
bridge to cardiac transplantation. J Am Coll Cardiol. ZT, Buja LM, Frazier OH. Morphologic changes in
2003;41(3):381–5. the aortic wall media after support with a continu-
29. Russell SD, Rogers JG, Milano CA, et al. Renal and ous-flow left ventricular assist device. J Heart Lung
hepatic function improve in advanced heart failure Transplant. 2013;32(11):1096–100.
patients during continuous-flow support with the 44. Feldman D, Pamboukian SV, Teuteberg JJ, et al.

HeartMate II left ventricular assist device. Circulation. The 2013 International Society for Heart and Lung
2009;120(23):2352–7. Transplantation guidelines for mechanical circulatory
30. Singh M, Shullo M, Kormos RL, et al. Impact of support: executive summary. J Heart Lung Transplant.
renal function before mechanical circulatory support 2013;32(2):157–87.
on posttransplant renal outcomes. Ann Thorac Surg. 45. Fulop T, Pathak MB, Schmidt DW, et al. Volume-
2011;91(5):1348–54. related weight gain and subsequent mortality in acute
144 W.B. Etheridge and S.A. Shearer

renal failure patients treated with continuous renal survival and acid-base balance with Hextend compared
replacement therapy. ASAIO J. 2010;56(4):333–7. with saline. Crit Care Med. 2002;30(2):300–5.
46. Vermeulen H, Hofland J, Legemate DA, Ubbink
50. Cox ZL, Calcutt MW, Morrison TB, Akers WS, Davis
DT. Intravenous fluid restriction after major abdomi- MB, Lenihan DJ. Elevation of plasma milrinone con-
nal surgery: a randomized blinded clinical trial. Trials. centrations in stage D heart failure associated with
2009;10:50. renal dysfunction. J Cardiovasc Pharmacol Ther.
47. Brandstrup B, Tonnesen H, Beier-Holgersen R, et al. 2013;18(5):433–8.
Effects of intravenous fluid restriction on postopera- 51. Guglielmi AA, Guglielmi KE, Bhat G, Siemeck R,
tive complications: comparison of two perioperative Tatooles AJ. Peritoneal dialysis after left ventricu-
fluid regimens: a randomized assessor-blinded multi- lar assist device placement. ASAIO J. 2014;60(1):
center trial. Ann Surg. 2003;238(5):641–8. 127–8.
48. González-Fajardo JA, Mengibar L, Brizuela JA,
52. Ronco C, Bellomo R, Homel P, et al. Effects of differ-
Castrodeza J, Vaquero-Puerta C. Effect of postopera- ent doses in continuous veno-venous haemofiltration
tive restrictive fluid therapy in the recovery of patients on outcomes of acute renal failure: a prospective ran-
with abdominal vascular surgery. Eur J Vasc Endovasc domised trial. Lancet. 2000;356(9223):26–30.
Surg. 2009;37(5):538–43. [Link] 53. Thomas BA, Logar CM, Anderson AE. Renal replace-
ejvs.2009.01.010. Epub 2009 Feb 20. ment therapy in congestive heart failure requiring left

49. Kellum JA. Fluid resuscitation and hyperchloremic ventricular assist device augmentation. Perit Dial Int.
acidosis in experimental sepsis: improved short-term 2012;32(4):386–92.
 Chronic Management of Patients
with Left Ventricular Assist Devices 10
Luke C. Cunningham and Ajith P. Nair

from the larger XVE device, weighing 1250 g

Introduction requiring intraperitoneal implantation, to much
smaller devices such as the HeartMate II (390 g)
Implantable left ventricular assist device (LVAD) and HeartWare HVAD devices (160 g) (Figs. 10.1
use has continued to increase since the first long-­ and 10.2) [1]. Currently, FDA-approved
term placement of such a device in 1988. Following continuous-­flow devices include the axial-flow
this event, the FDA approved implantable LVADs HeartMate II (Thoratec, Pleasanton, CA) and the
in 1994 as a bridge to transplantation. Due to a centrifugal-flow HeartWare HVAD (HeartWare
perpetual shortage of available organ donors for International, Framingham, MA). In addition,
end-stage congestive heart failure (CHF) patients, HeartMate III is a continuous-flow centrifugal
durable LVAD implantation is now also used for device that is currently under investigation in
destination therapy (DT) since 2002 after it clinical trials (Fig. 10.3). Given the increased use
received approval from the FDA. These popula- of continuous-flow devices, we will review care
tions of patients require close monitoring and con- and use of all the above devices.
tinued management of congestive heart failure. The HeartMate II device (Thoratec, Pleasanton,
Initial devices attempted to preserve pulsatile CA) was initially approved for use as a bridge-to-
flow such as the HeartMate XVE device, which transplant (BTT) therapy in 2008 and then as a
used a central blood chamber and inflow-outflow destination therapy (DT) for patients not eligible
conduits separated by 25 mm porcine valves. The for transplant in 2010. HeartWare became approved
Texas Heart Institute (THI) was key in the devel- as a BTT device in 2012 and is currently complet-
opment and testing of the device with successful ing trials to obtain DT approval. Increasing use of
implantation in 1991. However, continuous-flow LVADs for DT has been seen in the most recent
LVADs have shown to be more durable and now Interagency Registry for Mechanically Assisted
have replaced older models. As technology Circulatory Support (INTERMACS) data, which
improves, the size of these devices has decreased recorded an increase in patients receiving DT
devices from 28.3% in the 2008–2011 era to 45.7%
in 2014. However, approximately 60% of patients
still receive an LVAD as a BTT therapy with imme-
L.C. Cunningham, M.D. • A.P. Nair, M.D. (*) diate listing for heart transplantation or plans to
Texas Heart Institute at Baylor St. Luke’s Medical place on the list in the near future [2].
Center, Baylor College of Medicine, Those receiving a device as DT have been dem-
Houston, TX, USA onstrated to have improved survivals compared to
e-mail: lukec@[Link]; [Link]@[Link]

© Springer International Publishing AG 2018 145

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
146 L.C. Cunningham and A.P. Nair

Fig. 10.3  HeartMate III left ventricular assist device. CT

scan images shown. The device is implanted in the peri-
cardial space. Compared to the HVAD, it is larger with a
shorter inflow cannula

LVADs have improved long-­ term survival in

Fig. 10.1  HeartMate II left ventricular assist device. CT patients with end-stage heart failure who receive a
scan film demonstrating appearance of the device on radi- device for a BTT or DT indication. According to
ography, which is important to recognize in patients with
long-term devices in place. The inflow cannula is seen the most recent estimates, combined survival at
entering the left ventricle, and the pump is positioned out- 2 years with either an axial or centrifugal LVAD is
side the pericardial space 83%. Despite improvements in devices, chal-
lenges continue to emerge in the management of
these patients.

Device Management

General Management

Current generation devices utilize a single part

rotor in either an axial-flow rotor or centrifugal
design to counter earlier generation complica-
tions of mechanical failure, but the continuous
flow of these devices changes the nature of moni-
toring this subset of patients. While speed is the
main parameter adjusted in an LVAD patient
based on their hemodynamic status, there should
also be continued optimization of the CHF regi-
Fig. 10.2  HeartWare HVAD left ventricular assist device.
CT scan demonstrates the appearance of the device. The
men, mean arterial pressure, and possibly newly
pump is implanted in the pericardial space initiated anticoagulation regimen. This should
include maintenance of optimally tolerated neu-
rohormonal blockade, volume optimization, and
those end-stage heart failure patients who have no afterload reduction.
intervention. For DT patients, after implantation After implantation of an LVAD, monitoring of
of an LVAD, there is estimated survival of 76% the device parameters and alarms can provide
and 57% at 1 year and 3 years, respectively. insight into the function and hemodynamic status
Technical improvements in continuous-­ flow of the patient. HeartMate II device parameters
10  Chronic Management of Patients with Left Ventricular Assist Devices 147

include speed (RPMs), power (Watts), pulsatility as a right heart catheterization (RHC). The speed
index (PI), and estimated flow (L/min). These should be aimed to maintain peripheral pressure
parameters should be measured and recorded at and perfusion while minimizing right ventricular
each visit to assess for trends or acute changes in (RV) overload, aortic regurgitation (AI), or left
LVAD function. Additionally, review of any ventricular (LV) collapse [3, 4]. Since RV func-
recent alarms and the timing of those alarms is an tion, AI, and LV size are dynamic factors, fre-
important evaluation of an LVAD patient. The quent assessment should occur at least every
specific interpretation of alarms will be reviewed 6 months following implantation of an LVAD or
later in this chapter. sooner if symptoms arise.
Generally flow and power carry a linear rela-
tionship compared at a given speed. However,
this may not hold true in certain clinical situa- Role of Hemodynamic Assessment
tions. Given the linear relationship between flow
and power, flow is a calculated number based on A reduction in exertional capacity is a character-
a direct measurement of the power. Therefore, if istic feature of advanced heart failure. The use of
the power increases due to mechanical failure, it cardiopulmonary exercise testing (CPET) in
may not truly reflect increased flow from the patients with severe left ventricular impairment
LVAD. For example, when a thrombus is present prior to cardiac transplantation or LVAD place-
in the inflow cannula, it can produce an increase ment will demonstrate a markedly reduced exer-
in power without increased flow due to the cise time and peak oxygen consumption. Patients
obstruction by the thrombus. This may cause with marked reduction of VO2 below 14 mL/kg/
some confusion related to reported device read- min have been demonstrated to have a reduced
ings as both power and flow will be incorrectly survival when compared to those with left ven-
reported as increased. Similarly, if there was out- tricular impairment and a VO2 greater than
flow obstruction, this would cause a decrease in 14 mL/kg/min [5]. While completion of CPET is
flow as well as erroneously low power. typically completed prior to cardiac transplanta-
The PI is a measure of the assistance the tion, its use is also recommended in patients after
LVAD is providing to the LV and is provided in a implantation of an LVAD. It may be useful in
range from 1 to 10. The PI is a specific parameter providing clinicians an objective assessment of
for only the HeartMate devices. This PI is calcu- exercise capacity as well as helping guide recom-
lated based on flow pulses during systole, which mendations for exercise regimens. Studies have
are sensed by the LVAD and averaged over dura- consistently demonstrated improvement in exer-
tion of 15 s. Lower values indicate the pump is cise time in the first 6–8 weeks following LVAD
contributing more to systemic flow and will implantation; however, there has been variable
translate to less pulsatility, whereas higher values response to VO2. Age may be a large predictor of
indicate less pump contribution to this flow and VO2 improvement with patients reaching only
higher pulsatility. PI values should remain rela- 50–60% of their age- and sex-predicted VO2
tively constant, and a decrease should lead clini- after LVAD implantation [6, 7].
cians to consider a decrease in circulating blood Current guidelines recommend regular inter-
volume. Significant increases in PI should prompt val testing of hemodynamics by right heart cath-
further evaluation for possible fluid retention, eterization, particularly in those patients awaiting
hyperdynamic states (such as sepsis), significant transplantation. Serial evaluation can identify
aortic valvular insufficiency, and, in rare those patients with pulmonary hypertension
instances, cardiac recovery. (pHTN), which when irreversible has been
Additionally, after implantation, the necessary ­associated with a higher risk of allograft dysfunc-
speed of the LVAD should be evaluated at each tion [8–10]. No specific recommendations cur-
visit based on mean arterial pressure, echocar- rently exist on the interval of testing; however,
diography, and, if necessary, further testing such data has suggested that those with pHTN on right
148 L.C. Cunningham and A.P. Nair

heart catheterization 2–3 months after LVAD American Society of Echocardiography recom-

implantation are at the greatest risk for persistent mend surveillance TTE at postoperative week 2
or progressive pHTN in the following 6 months. and at 1, 3, 6, and 12 months if the patient remains
In addition, there appears to be little predictive clinically stable. Evaluation should then take
value of a right heart catheterization prior to place every 6–12 months thereafter [12].
LVAD implantation and development of pHTN Measurement of the left ventricular internal
following surgical implantation. diastolic dimension (LVIDd) is an important
Right heart catheterization has also proven parameter in the assessment of LV unloading fol-
useful for hemodynamic-guided LVAD optimiza- lowing LVAD implantation. While end-diastolic
tion and diagnosis of inadequate pump speed, volumes as obtained by Simpson’s method of
right heart failure, or volume overload. Studies disks have in some cases been shown to be a
have demonstrated the ability to decrease pulmo- more accurate measurement of LV unloading,
nary capillary wedge pressure and central venous reproducibility can be an issue due to shadowing
pressures as well as increase cardiac output/index from the outflow cannula of the LVAD. Therefore,
by evaluating increasing pump speeds while in the use of the parasternal long-axis images to
the catheterization lab [11]. Interval evaluation obtain the LVIDd has been found to be the most
may also provide a way to differentiate patients reproducible measurement. Following LVAD
with persistent heart failure symptoms into cate- implantation, approximately a 15% reduction in
gories of right heart failure (relatively normal LV size has been shown to be expected at
PCWP with persistently elevated central venous 3 months post-LVAD [13, 14]. This data was
pressures) versus those with left-sided volume obtained solely in HM II patients; therefore vari-
overload (persistently elevated PCWP despite ation may exist between devices.
increasing pump speeds) (Fig. 10.4). Also some In regard to LV ejection fraction, serial mea-
patients may respond to increasing pump speeds surement can provide clinicians data to evaluate
with decreasing PCWP and central venous pres- for myocardial recovery or worsening over time.
sures suggesting inadequate LVAD speeds. While As mentioned above, measurement of LVEF by
hemodynamic improvements have been achieved Simpson’s method of disks is the recommended
with right heart catheterization-directed studies, method; however, it can be difficult due to shad-
no direct correlations have been drawn to symp- owing in the apex from the LV outflow cannula,
tomatic improvement or decreases in morbidity paradoxical septal motion, or significant regional
and mortality. wall motion abnormalities. Therefore, in cases
when endocardial visualization is difficult, the
use of alternative methods including LV frac-
Echocardiography tional shortening, LV fractional area change, or
the Quinones method has been suggested (for the
This section will identify the basics of echocar- latter bearing in mind the LV apex should be con-
diography in the chronic management of patients sidered akinetic due to the presence of the LV
following implantation of an LVAD, while prior inflow cannula) (Table 10.1). It is important to
chapters will address preoperative assessment and remember that these methods are not validated in
issues immediately postoperatively. Transthoracic patients with LVAD in situ.
echocardiography (TTE) is essential in the opti- Assessment of the aortic valve mobility is an
mization of LV decompression, reduction of aor- alternative parameter in serial evaluation of LV
tic insufficiency, and evaluation of possible device unloading. If LVAD speed is set above which
malfunction. Using standard echocardiographic allow aortic valve opening, it may lead to aortic
views, 2D measurements, color Doppler, and regurgitation (AR) which impairs LVAD function.
spectral Doppler can collectively be used to pro- This occurs as a continuous loop of flow from the
vide LV size, valvular function, and interrogation LV outflow cannula to the aorta followed by regur-
of inflow/outflow cannulae. Guidelines by the gitation back into the LV. Therefore, significant
10  Chronic Management of Patients with Left Ventricular Assist Devices 149

Fig. 10.4  Interpretation of right heart catheterization in LVAD patients. CVP central venous pressure; PCWP pulmo-
nary capillary wedge pressure; RHF right heart failure

AR can affect the unloading of the LV and thus Parameters for defining the severity of AR
LVAD effectiveness. Development of AR has have not been specifically validated; however, the
important implications in morbidity and mortality, use of prior guidelines is generally followed.
which will be discussed in more detail later in this When there is a vena contracta of ≥0.3 cm or an
chapter. Additionally, if the aortic valve remains AR jet width of >46% of the LVOT diameter,
closed, it can predispose patients to aortic root there is likely at least moderate if not severe
thrombus and/or fusion of the aortic valve cusps). AR. The AR may also be present during only
Assessment of the aortic valve during LVAD diastole, nearly continuous when it extends into
surveillance TTE should begin with evaluation of systole, or continuous when it is holodiastolic
opening of the valve. This is most accurately and holosystolic. Due to the ability of AR to be
achieved with the use of M-mode echocardiogra- present into systole and the extracardiac circuit of
phy by recording the aortic valve in up to 5–6 the LVAD, neither the pressure halftime nor the
cardiac cycles, as the valve can open with every presence/absence of aortic diastolic flow reversal
cardiac cycle, open intermittently, or remain is a reliable method to quantitate AR.
closed with every cycle. While aortic valve open- Mitral regurgitation (MR) is also regularly
ing can occur with each cardiac cycle, it may only evaluated on surveillance TTE after LVAD
occur for a short duration. Therefore, the duration implantation as it can have implications for
of aortic valve opening should also be addressed device management. Quantification of the sever-
by averaging the duration of opening in multiple ity of MR can be made based on the general
cardiac cycles, usually in milliseconds (ms). echocardiography guidelines. Importantly, the
150 L.C. Cunningham and A.P. Nair

Table 10.1  Alternative methods of evaluation of LVEF in LVAD patients

Method Equation Benefits Limitations
Fractional area FAC = [(end-diastolic area − end-systolic Can be used in the Less reliable in the
change (%) area)/(end-diastolic area)] absence of adequate setting of significant LV
apical visualization wall motion or distortion
(i.e., aneurysm)
Fractional FS = [(LVIDd − LVIDs)/(LVIDd)] Can be used in the Less reliable in the
shortening absence of adequate setting of significant
apical visualization wall motion or LV
distortion (i.e.,
aneurysm)
Quinones method LVEF = [(LVIDd2 − LVIDs2)/(LVIDd2)] Decreases error by Less reliable in the
using multiple areas setting of significant
of measurement wall motion or LV
distortion (i.e.,
aneurysm)
FAC fractional area change; FS fractional shortening; LVIDd left ventricular internal diastolic dimension; LVIDs left
ventricular internal systolic dimensions; LVIDd2 indicates measurement of eight different LV dimensions at different
levels in the parasternal long axis, four-chamber, and long-axis views of end diastole; LVIDs2 indicates measurement of
eight different LV dimensions at different levels in the parasternal long axis, four-chamber, and long-axis views of end
systole

presence and severity of MR can be an indicator inflow may be visualized in the parasternal long
of adequate unloading provided by an axis or LV four-chamber views and should be eval-
LVAD. Appropriate LVAD speeds will ideally uated for its position in reference to the septum or
lead to reduction in LV size and in turn the mitral submitral apparatus. Flow through the cannula can
annulus. This reduction in the mitral annular size also be interrogated using pulsed and continuous
will improve coaptation and thus reduce mitral wave Doppler and should be obtained over 3–4
regurgitation. However, if mitral regurgitation is cardiac cycles. Normal Doppler waveforms will
persistent despite increasing LV unloading, eval- be pulsatile due to the contribution of the LV to
uation for LVAD malfunction should be sought as flow even if the aortic valve is closed (Fig. 10.5).
the outflow cannula may interfere with the sub- Doppler velocities should also be ≤1.5 m/s, and
mitral apparatus in some cases. when higher flows are present, it may indicate
The tricuspid and pulmonic valves are reliably obstruction or the presence of thrombus.
interrogated using standard methods in patients Evaluation of the aortic outflow graft anasto-
following LVAD implantation. Tricuspid regurgi- mosis is more difficult but can be seen in the
tation (TR), present in moderate to severe ranges, modified parasternal views, which focus on the
can provide indirect data on the function of the ascending aorta. In cases where this is not suffi-
LVAD. Assuming the absence of a concurrent cient, positioning patients in the right lateral
right ventricular assist device, significant TR in decubitus position and obtaining right parasternal
appropriate clinical scenarios can suggest inade- views may be helpful. Spectral velocities through
quate LV unloading, RV dysfunction, or exces- the graft can be used in calculation of flow using
sive LV unloading leading to intraventricular velocity time integral (VTI) and outflow graft
shift and distortion of the tricuspid valve mor- area method, keeping in mind that velocities can
phology. The presence of significant TR should vary between graft sizes. For example, HeartMate
therefore prompt review of serial changes in LV II tends to have a larger outflow graft (16 mm)
size, ejection fraction, measures of RV function, than HeartWare devices (10 mm). In general,
and intraventricular motion. flows >2 m/s are considered abnormal for the
Lastly, TTE can be useful in the evaluation of outflow graft, and further evaluation for obstruc-
the inflow and outflow cannula of the LVAD. The tion should be undertaken.
10  Chronic Management of Patients with Left Ventricular Assist Devices 151

Fig. 10.5  Pulsed wave Doppler is seen of a HeartMate II inflow cannula. Although it is a continuous-flow device,
contribution of left ventricular contraction leads to a systolic peak (lined arrow) and a diastolic nadir (arrow head).
Typically these can be obtained in a standard four-­chamber echo view or from the parasternal long view (upper left
panel). Velocities should be ≤1.5 m/s

Optimization of a patient’s LVAD speed can  eartMate II and HeartMate III Device

H
be completed in asymptomatic ambulatory Parameters
patients using TTE to obtain the above param-
eters. These optimization studies, sometimes After implantation of a HeartMate II or HeartMate
referred to as a “speed change” echo, can be III device, patients will receive a system control-
obtained by completing a TTE study at the ler with two sets of batteries, a primary operating
baseline speed and either lower or incremen- set, and a backup. Additionally, the patients will
tally increasing speeds. At each speed, the have a system controller and a power base unit
patient’s mean pressure as well as echo param- (PBU). A control monitor is required at the initial
eters should be obtained, including LVIDd, implantation and subsequent encounters to
septal position, frequency/duration of aortic review settings and alarms but is not a required
valve opening, and quantitative or qualitative component for discharge. Patients will be edu-
AR, MR, and TR. It should be noted that if a cated during their admission on the operation of
thrombus is visualized in the aortic root, the the system controller, which has several display
LVAD speed should not be changed as it can icons and buttons seen on the face (Fig. 10.6).
aid in mobilization of the thrombus, especially There are two buttons, which include the “Test
at lower speeds. Select” and “Silence Alarm” buttons, which
152 L.C. Cunningham and A.P. Nair

 eartMate II and HeartMate III Device

H
Alarms

Several alarms can occur within the HeartMate II

and HeartMate III systems and are divided into
hazard alarms and advisory alarms. The former
type of alarm is the more critical of the two and
can indicate imminent or current loss of support
from the device. Hazard alarms are indicated by a
continuous alarm tone and illumination of either
the red battery or heart symbols. When the red
heart symbol is illuminated with a continuous
Fig. 10.6  HeartMate II control system
alarm tone, this indicates low flow (≤2.5 L/min)
or pump off. It should prompt evaluation of the
pump connections between the controller and
allow patients to interact with the system. Lighted power source. Illumination of the red battery
icons on the controller include a power symbol, symbol indicates a low-voltage situation and
battery symbol with fuel gauge, battery module should prompt the user to replace current batter-
symbol, and red heart symbol. ies. In this situation, the system will revert to a
The battery symbol and fuel gauge are the power saver mode in which the RPMs will drop
most vital for day-to-day patient usage. The to 8000 despite the current settings. Lastly, if a
gauge has four green markers, which can provide continuous audio tone is present without an illu-
an approximate amount of battery life remaining. minated indicator, it suggests the system is not
When all four are lighted, it signifies that receiving power and should lead the operator to
75–100% of the battery energy remains with a check the connections between the controller and
reduction of 25% of battery energy with the loss power source or change the current battery.
of each marker. When a single green lighted Advisory alarms indicate a change in the con-
marker remains, it signifies <25% of battery dition or minor malfunction of the system, which
energy remains, and once the battery symbol will have little to no effect on hemodynamic sup-
appears as a yellow or red indicator, it signifies port. Instead of a continuous alarm tone, advisory
<15 min or <5 min of battery energy remains. alarms will emit alarms lasting seconds with or
Not all patients can be provided with similar without illumination of indicators on the control-
durations of battery life as this can vary depend- ler. A yellow battery illumination with one beep
ing on the set speed or the age of the battery. every 4 s indicates a low-voltage advisory with a
Higher set speeds will deplete a battery more battery life of <15 min. This should prompt the
quickly, and as the battery ages, it will hold its user to change the battery or switch to an alter-
charge for less time. nate power source. Similarly, when the battery
As previously described, the LVAD speed, powering the controller is depleted, a yellow bat-
flow, power, and pulsatility index are displayed tery alarm symbol will appear and alert one beep
when the device is hooked to the control monitor every 4 s; this requires replacement of the battery.
and can be adjusted by the clinician through this If the power lead is disconnected, an alarm will
interface (Fig. 10.7). The minimum and maxi- beep every second with green illumination of the
mum operating speeds for the HeartMate devices power button and flashing of the green fuel gauge
are 6000 and 15,000 RPMs, but typical operating lighting. In this situation, the power cord should
speeds usually range between 8800 and 10,000 be promptly reconnected. Lastly, if the module
RPMs. When making speed adjustments or speed provides one alarm tone every 4 s with no warn-
optimization on TTE, incremental changes of ing light, this is an indication that the pump is
200–400 RPMs are used. operating below the low speed limit. If appropri-
10  Chronic Management of Patients with Left Ventricular Assist Devices 153

Fig. 10.7  HeartMate II display panel. The pump speed, pump power, pulsatility index, and pump power are displayed

ate, the patient should be connected to the system  VAD Device Parameters
H
monitor and the speed adjusted. Additionally, if and Waveform Analysis
the system reverts to the backup operating sys-
tem, two alarm tones will be provided in 1 s fol- The HVAD device when implanted comes
lowed by 2 s of silence with no illuminated alarm equipped with two sets of batteries, a battery
indicators. In this situation, the system controller charger and a controller. Unlike the HeartMate
should be replaced. devices, patients receive a touch screen tablet
A daily self-test should be completed on the monitor at discharge, which allows for monitor-
system controller to ensure proper operation of ing of device function and adjustment of param-
the device. A self-test is accomplished by press- eters if needed. Recommended device speeds
ing and holding the “test select” button for at range from 2400 to 3200 RPM, which correlate
least 3 s at which time all indicators should light to approximately 3–8 L of flow. Speeds can be set
up and a continuous alarm tone should be heard. as low as 1800 RPM; however, these are not gen-
It is during this time the patient or clinician com- erally used except for during initial implantation
pleting the test should ensure all indicators are when patients are being weaned from cardiopul-
functioning properly. If malfunction of the alarm monary bypass. Speeds higher than 3200 RPM
system occurs, it will emit a 1-s tone every other are associated with higher risk of LV suction
second, or if malfunction of the controller is events.
present requiring replacement, it will emit an The HVAD controller face has two push but-
audio tone of two beeps every second, and no tons and four indicators as well as a display.
indicators will light up. If a normal self-test is Buttons include the “alarm mute” and “scroll”
performed, the alarm tone and indicators will buttons, while indicators include AC/DC indi-
turn off 5 s after releasing the “test select” but- cator, battery 1 and 2, and alarm indicator. A
ton. The LVAD will remain in its set parameters patient must be connected to two power sources
throughout the self-test. through the controller at all times, and this may
154 L.C. Cunningham and A.P. Nair

include two batteries or a single battery and an monitor power icon. Each of these allows for
AC/DC adapter. The battery icons on the con- review of current settings, review of alarms,
troller include a battery symbol with four review of temporal trends in flow/speed/power,
boxes, which will illuminate green when a fully and adjustment of LVAD speed or patient data,
charged battery is connected. As battery energy respectively.
is depleted, the boxes sequentially no longer Lastly, the home screen includes two real-
illuminate and indicate the percent of battery time waveforms, which display the pulsatility
energy remaining. For example, when three of both the flow and power. Normal pulsatility
boxes remain, there is 50–74% of battery of the flow should include a variation of at least
energy remaining, two boxes is 25–49% of 2 L/min from peak to trough in the waveform.
energy remaining, and one box is <25% energy In addition, the minimum value of flow should
remaining. Depending on whether one or both be greater than 2 L/min. Deviations of normal
batteries are depleted, different alarms may waveforms may provide suggestion of pump or
sound, and this will be discussed in the subse- hemodynamic dysfunction such as with regur-
quent section. gitant lesions, pump thrombosis, or suction
The “home screen” seen on the HVAD touch events. When regurgitant lesions arise, such as
screen monitor will display important parame- aortic insufficiency, the pulsatility waveform
ters that can help assess and evaluate the current may take on a large variability with trough
condition of the patient and pump. From the top flows reaching negative flow rates. In contrast,
left hand of the screen down will be displayed in patients with suction events or a small,
the pump flow (L/min), speed (RPM), and decompressed LV, there may be decreased vari-
power (Watts) (Fig. 10.8). Immediately to the ability due to the outflow cannula abutting the
right of this display bar appear five touch LV wall and reducing blood intake. Since the
responsive icons which allow the clinician to waveform display on the home screen may pro-
navigate between display screens. From top to vide only short periods of assessment, it can be
bottom, these include the “home screen,” “alarm useful to review the waveform from longer
screen,” “trend screen,” “system screen,” and a time periods through the trend screen which

Fig. 10.8  HeartWare HVAD waveform display

10  Chronic Management of Patients with Left Ventricular Assist Devices 155

can retrieve data from the prior 60 min, 4 h, Heart Failure Management
24 h, 14 days, or 30 days.
Implantation of an LVAD provides hemodynamic
support and reduces myocardial work improving
HVAD Device Alarms symptoms and the functional class of CHF
patients. This support also provides unloading of
Device alarms on the HVAD device can be sepa- LV and can allow for recovery of LV function in
rated into low-, medium-, and high-priority some cases [15–17]. Pharmacologic therapies that
alarms. Each alarm has a separate set of indica- augment the renin-angiotensin-aldosterone sys-
tors and alarm tones, and the controller will pro- tem and sympathetic nervous system have also
duce a text message to indicate the exact source been shown to improve LV function, decrease
of malfunction. fibrosis, and improve heart failure symptoms.
High-priority alarms indicate immediate Importantly, mortality has been shown to be
action is needed as loss of support has occurred reduced in CHF patients in large randomized tri-
or is imminent. For all alarms, the alarm indicator als using angiotensin-converting enzyme (ACE)
will be illuminated in a flashing red color, and a inhibitors [18–21], angiotensin receptor blockers
continuous alarm tone will sound which cannot (ARB) [22, 23], β-blockers [24–27], mineralocor-
be muted. Situations in which high-priority ticoid antagonists [28, 29], and the most recently
alarms occur include when the driveline is dis- validated angiotensin-­neprilysin inhibitor (ARNi)
connected or the connector is broken, when there and sacubitril/valsartan combination [30].
is controller failure, or when both battery sources At this time, no large randomized controlled tri-
have limited time remaining. In each of these, the als have been completed to evaluate pharmacologic
controller will produce a two-line message, for therapies in patients with LVADs. Generally, the use
example, if there is controller failure, the first line of diuretics, ACE inhibitors/ARBs, and β-blockers
may read “VAD Stopped” and the second line is encouraged for the management of volume over-
“Change Controller.” load, hypertension, and tachyarrhythmia, respec-
Medium-priority alarms will produce a flash- tively. Mineralocorticoid a­ntagonists may also be
ing yellow illumination of the alarm indicator used to limit the need for potassium replacement in
and a low-volume tone which will increase in LVAD patients with adequate renal function.
volume over the next 1 min if not silenced. This However, given the lack of direct evaluation of neu-
type of alarm can indicate a low-flow situation, rohormonal blockade in LVAD patients, the 2013
suction events, electrical fault, or high voltage of ISHLT guidelines provide a Class I indication for
the device. Typically these alarms prompt the the use of these agents for the above indications and
patient to call the clinician for further instruc- with only a level of evidence C (expert opinion) [8].
tions. Text messages may include “Low Flow” in Small studies have evaluated the use of combi-
line 1 and “Call” in line 2. Low-priority alarms nation therapies for LV recovery in LVAD patients.
will produce a constant yellow illumination of One single-center study initiated maximally toler-
the alarm indicator and a low-volume tone which ated lisinopril (maximum dose 40 mg daily),
will increase after 5 min if not silenced. These carvedilol (maximum dose 50 mg twice daily),
alarms occur when one of two power sources is spironolactone (maximum dose 25 mg daily), and
low, for example, one of the two batteries con- losartan (maximum dose 100 mg daily) in 15 non-
nected. It will also occur when one power source ischemic cardiomyopathy patients after device
is disconnected from the controller. Text mes- implantation. The patients were followed by echo-
sages may include “Low Battery 1” in line 1 and cardiography every 2 weeks for 6 months to evalu-
“Replace Battery 1” in line 2. For complete ate end-diastolic LV size with the device on and
review of text provided with each potential alarm, off. Once it was determined, maximal LV recovery
clinicians should review the HeartWare had occurred, and the patients were switched to
Instruction Manual. clenbuterol at a maximally tolerated dose of
156 L.C. Cunningham and A.P. Nair

700 μg three times daily in addition to bisoprolol. The etiology of late RVF has not been fully
At the end of the study, 11 patients were able to elucidated, but it may be due to hemodynamic
undergo device explantation with an improvement changes following LVAD implantation. Early
in LVEF from 12 ± 6% to 64 ± 8% (p = 0.001) and RHF has been associated with increases in car-
end-diastolic diameter from 75.1 ± 16.3 mm to diac output, which directly increases preload and
55.9 ± 8.3 mm (p = 0.002). The remaining four may increase the work of the right ventricle. In
patients did not meet the criteria for explantation addition, the geometry of the heart can be
but underwent transplantation. The survival rate changed when the left ventricle is decompressed
among those after explantation at 1 and 4 years and/or the interventricular septum is shifted left-
was 90.1% and 81.8%, respectively [31]. ward. However, given that these changes are
more acute after implantation of an LVAD, they
are unlikely to contribute to the late presentations
Right Heart Failure of RHF. Studies have shown that baseline mea-
surements of CVP and CVP/PCWP ratios are
Incidence of Late RV Failure similar among those who develop RHF implicat-
ing other mechanical stressors [33, 34]. Therefore,
Right ventricular failure (RVF) can be seen in one common link between those who develop
patients after LVAD implantation and carries a late RHF has been readmission for tachyarrhyth-
poor prognosis for these patients. Early right mia, bleeding, and infection or progression of
heart failure is a common complication following tricuspid regurgitation and pulmonary hyperten-
implantation and has been studied more fre- sion [33, 35].
quently in the literature. However, patients may
present months to years after their incident
admission. Generally, these patients develop Pulmonary Vasodilators
signs of increased peripheral edema, elevated in Pulmonary Hypertension
liver enzymes due to hepatic congestion, increas- and Right Heart Failure
ing diuretic requirement, and decreased LVAD
flow and/or pulsatility. The incidence of late RVF Prior evidence has shown a direct correlation
has been studied in a few small cohorts and dem- with pulmonary hypertension and transplant
onstrated a variable incidence between 11 and mortality [36] which has led to the International
45%; however, definitions of RVF varied signifi- Society of Heart and Lung Transplantation to
cantly among studies [32, 33]. provide a relative contraindication to cardiac
Varying severity of RVF has been defined pre- transplantation in patients with a pulmonary vas-
viously by the level of therapy needed for the cular resistance (PVR) >3 Woods units (WU).
patient. Those with mild RVF have two of four Studies have documented the success of pulmo-
signs of elevated central venous pressure (CVP) nary vasodilators in lowering PVR in patients
by Swan-Ganz measurement (>18 mmHg), car- after implantation of an LVAD with both silde-
diac index <2.3 L/min/m2, ascites or moderate to nafil and bosentan; however, their effects on out-
severe peripheral edema, or evidence of elevated comes remain unclear. Additionally, safety
CVP by physical examination or echocardiogra- concerns have been raised for the use of endothe-
phy. Mild RVF can usually be treated with lin receptor antagonists (ERAs) in patients with
increasing doses of diuretics. Moderate RVF has CHF due to results from the Research on
been defined as those meeting the mild criteria Endothelin Antagonism in Chronic Heart Failure
and requiring inotropic support or pulmonary (REACH-1) trial in which a large proportion of
vasodilators for a duration of >1 week at any time patients had elevation of their transaminases
following LVAD implantation. Severe RVF is without significant clinical benefit.
defined in those patients that necessitate a right In a retrospective study by LaRue et al., 50
ventricular assist device. patients were treated with bosentan after LVAD
10  Chronic Management of Patients with Left Ventricular Assist Devices 157

implantation. Predominantly these patients were less than 130 mmHg and a diastolic blood pres-
recipients of HeartMate II devices, and 70% were sure of <85 mmHg. In patients with nonpulsatile
implanted with a BTT strategy. Therapy was support devices, they recommend a mean BP goal
started at a median of 37 days after LVAD implan- of <80 mmHg. These recommendations both
tation and continued for a median of about come with a level of evidence of “C” as they are
16 months. A target dose of 125 mg twice daily provided as a recommendation of expert opinion.
was achieved in 78% of patients with a statisti- Recent studies in patients with CF-LVAD have
cally significant decrease in 6-month follow-up shown that achieving mean BPs <80 mmHg is
BNP, total bilirubin, and alkaline phosphatase. A correlated with decreased incidence of hemor-
significant decrease in PVR was also noted by rhagic stroke, thromboembolic events, and aortic
echocardiography measurement with a mean regurgitation. One retrospective study included
reduction of 1.4 Woods units (WU) (3.93 ± 1.53 123 patients who were at 30 days from CF-LVAD
to 2.58 ± 1.05, p < 0.0001). Of note, three patients implantation and in the outpatient setting. End
discontinued the medication due to elevated liver points including intracranial hemorrhage, throm-
transaminases, but all had normalization after boembolic events, or moderate to severe aortic
discontinuation of the mediation [37]. insufficiency were evaluated. Based on at least ten
outpatient Doppler-derived blood pressure read-
ings, subjects were stratified to either a controlled
Hypertension Management (<80 mmHg), intermediate (80–90 mmHg), or
high (>90 mmHg) group. Based on the results
As in those patients who are at risk or have devel- 18 months after LVAD implantation, there were
oped CHF, hypertension management is of clear 97%, 86%, and 70% ­survival-­free from combined
benefit. In addition, emerging evidence suggests adverse events in each of the controlled, interme-
blood pressure control is of importance to pre- diate, and high groups, respectively [40].
vent device complications or malfunction. Due to
the increasing use of continuous-flow LVADs,
most blood pressure goals are reported as  ortic Insufficiency and Valvular
A
achievement of a goal mean pressure given the Heart Disease
loss of pulsatility. Validation of blood pressure
measurements by dopplerable flow has been Greater than mild AI after LVAD implantation
shown to be correlated with both systolic blood has been demonstrated to occur in 25–31% of
pressure and mean arterial pressure when vali- patients at 1 year [41, 42]. This finding has been
dated to simultaneous intra-arterial measurement demonstrated to lead to worsening of heart fail-
[38, 39]. Measurement can be taken at the bra- ure symptoms and increased LVAD flows due to
chial artery with a pulse wave Doppler as the first the continuous circuit that occurs between the
indication of flow when a sphygmomanometer is LVAD and the LV [43]. A proposed etiology for
slowly deflated after obtaining occlusive pres- this clinical entity stems from the continuous
sure. This type of pressure is more accurate than positive pressure the outflow graft places on the
obtaining measurement from an automated blood aortic side of the valve and the negative pressures
pressure cuff or palpable pulsation, especially in the inflow cannula has on the LV side of the valve
those patients who do not achieve aortic valve which lead to remodeling, fusion of the aortic
opening at optimal speeds. valve leaflets, and eventually incompetence [44,
The International Society for Heart and Lung 45]. Current therapies for significant AI include
Transplantation (ISHLT) provides recommenda- medical management, aortic valve repair, aortic
tions for blood pressure goals in patients with valve closure, or aortic valve replacement.
durable mechanical circulatory support devices. Predictors for development of significant AI
In those with pulsatile devices, they recommend a have been evaluated in both preoperative and post-
blood pressure goal of a systolic blood pressure operative LVAD patients. Significant preoperative
158 L.C. Cunningham and A.P. Nair

predictors appear to include the use of continuous- 6. Dunlay SM, Allison TG, Pereira NL. Changes in
cardiopulmonary exercise testing parameters fol-
flow devices, lower body mass, lower preopera-
lowing continuous flow left ventricular assist device
tive ejection fraction, and female gender [41, 46, implantation and heart transplantation. J Card Fail.
47]. Interestingly, the presence of AI at baseline 2014;20:548–54.
has not been identified as a risk factor for progres- 7. Haft J, Armstrong W, Dyke DB, et al. Hemodynamic
and exercise performance with pulsatile and
sion after LVAD implantation. Postoperatively,
continuous-­ flow left ventricular assist devices.
greater risk is conferred for AI in patients with Circulation. 2007;166:I-8–I-15.
higher LVAD speeds, smaller LV size, lack of aor- 8. Feldman D, Pamboukian SV, Teuteberg JJ, et al.
tic valve opening, and presence of AV commis- The 2013 International Society for Heart and Lung
Transplantation guidelines for mechanical circulatory
sural fusion [41, 47–49].
support: executive summary. J Heart Lung Transplant.
Medical management for significant AI is 2013;32:157–87.
generally aimed at afterload reduction. AI 9. Goland S, Czer LS, Kass RM, et al. Pre-existing pul-
leads to increased LV diastolic pressures and monary hypertension in patients with end-stage heart
failure: impact on clinical outcome and hemodynamic
increased LVAD flow; however, when LVAD
follow-up after orthotopic heart transplantation.
speeds are increased to compensate, this may J Heart Lung Transplant. 2007;2007:312–8.
contribute to widening of transvalvular gradi- 10. Uriel N, Sims DB, Jorde UP. Fixed pulmonary hyper-
ent (by increasing the systemic diastolic pres- tension and mechanical support: an unclear opportu-
nity. J Heart Lung Transplant. 2011;30:600.
sure and decreasing the LV diastolic pressure).
11. Uriel N, Sayer G, Addetia K, et al. Hemodynamic
These hemodynamic changes are favorable for ramp tests in patients with left ventricular assist
worsening of the AI and promotion of aortic devices. J Am Coll Cardiol Heart Fail. 2016;4:208–17.
valve remodeling. Reduction of the LVAD 12.
Stainback RF, Estep JD, Agler DA, et al.
Echocardiography in the management of patients with
speed may actually improve AI by allowing at
left ventricular assist devices: recommendations from
least intermittent aortic valve opening, but if the American Society of Echocardiography. J Am Soc
systemic pressures are too high, end-organ per- Echocardiogr. 2015;28:853–909.
fusion may suffer. Therefore, afterload reduc- 13. Lam KM, Ennis S, O’Driscoll G, et al. Observations
from non-invasive measures of right heart hemody-
tion is the preferred treatment, aiming for the
namics in left ventricular assist device patients. J Am
ISHLT recommended goal of a mean arterial Soc Echocardiogr. 2009;22:1055–62.
pressure of ≤80 mmHg. Antihypertensive ther- 14.
Topilsky Y, Oh JK, Atchison FW, et al.
apies are reviewed in prior sections of this Echocardiographic findings in stable outpatients
with properly functioning HeartMate II left ven-
chapter.
tricular assist devices. J Am Soc Echocardiogr.
2011;24:157–69.
15. Dipla K, Mattiello JA, Jeevanandam V, et al. Myocyte
References recovery after mechanical circulatory support in
humans with end-stage heart failure. Circulation.
1998;97:2316–22.
1. Slaughter MS, Rogers JG, Milano CA, et al. Advanced
16. Terracciano CMN, Harding SE, Adamson D, et al.
heart failure treated with continuous-flow left ventric-
Changes in sarcolemmal Ca entry and sarcoplas-
ular assist device. NEJM. 2009;361:2241–51.
mic reticulum Ca content in ventricular myocytes
2. Kirklin JK, Naftel DC, Pagani FD, et al. Seventh
from patients with end-stage heart failure following
INTERMACS annual report: 15,000 patients and
myocardial recovery after combined pharmacologi-
counting. J Heart Lung Transplant. 2015;34:1495–504.
cal and ventricular assist device therapy. Eur Heart
3. Smith EM, Franzwa J. Chronic outpatient manage-
J. 2003;24:1329–39.
ment of patients with a left ventricular assist device.
17. Zafeiridis A, Jeevanandam V, Houser SR, et al.

J Thorac Dis. 2015;7:2112–24.
Regression of cellular hypertrophy after left ventricu-
4. Cowger J, Romano MA, Sulak J, et al. Left ventricular
lar assist device support. Circulation. 1998;98:656–62.
assist device management in patients chronically sup-
18. Lewis GR. Comparison of lisinopril versus pla-

ported for advanced heart failure. Curr Opin Cardiol.
cebo for congestive heart failure. Am J Cardiol.
2011;26:149–54.
1989;63:D12–6.
5. Mancini DM, Eisen H, Kussmaul W, et al. Value of
19. Lechat P, Garnham SP, Desche P, et al. Efficacy

peak exercise oxygen consumption for optimal timing
and acceptability of perindopril in mild to mod-
of cardiac transplantation in ambulatory patients with
erate chronic congestive heart failure. Am Heart
heart failure. Circulation. 1991;83:778–86.
J. 1993;136:798–806.
10  Chronic Management of Patients with Left Ventricular Assist Devices 159

20. Magnani B, Magelli C. Captopril in mild heart fail- 36. Hosenpud JD, Bennett LE, Keck BM, et al. The reg-
ure: preliminary observations of a long-term, double-­ istry of the international society for heart and lung
blind, placebo-controlled multicentre trial. Postgrad transplantation: seventeenth official report. J Heart
Med J. 1986;61:153–8. Lung Transplant. 2000;19:909–31.
21. Swedberg K, Kjekshus J. Effects of enalapril on mor- 37. LaRue SJ, Garcia-Cortes R, Nassif ME, et al.

tality in severe congestive heart failure: results of the Treatment of secondary pulmonary hypertension with
cooperative North Scandinavian Enalapril Survival bosentan after left ventricular assist device implanta-
study. Am J Cardiol. 1988;62:60A–6A. tion. Cardiovasc Ther. 2015;33:50–5.
22. Maggioni A, Anand I, Gottlieb SO, et al. Effects of val- 38. Bennett MK, Roberts CA, Russell SD. Ideal meth-
sartan on morbidity and mortality in patients with heart odology to assess systemic blood pressure in patients
failure not receiving angiotensin-­converting enzyme with continuous-flow left ventricular assist devices.
inhibitors. J Am Coll Cardiol. 2002;40:1414–21. J Heart Lung Transplant. 2010;29:593–4.
23. McMurray JJV, Ostergren J, Swedberg K, et al.
39. Lanier GM, Orlanes K, Hayashi Y, et al. Validity and
Effects of candesartan in patients with chronic heart reliability of a novel slow cuff-deflation system for
failure and reduced left-ventricular systolic function noninvasive blood pressure monitoring in patients
taking angiotensin-converting-enzyme inhibitors: the with continuous-flow left ventricular assist device.
CHARM-added trial. Lancet. 2003;362:767–71. Circ Heart Fail. 2013;6:1005–12.
24.
The cardiac insufficiency bisoprolol study II 40. Saeed O, Jermyn R, Kargoli F, et al. Blood pres-
(CIBIS-II): a randomised trial. Lancet. 1999;353:9–13. sure and adverse events during continuous flow left
25. MERIT-HF-investigators. Effect of metoprolol CR/ ventricular assist device support. Circ Heart Fail.
XL in chronic heart failure (MERIT-HF). Lancet. 2015;8:551–6.
1999;353:2001–7. 41. Cowger J, Pagani FD, Haft JW, et al. The devel-
26. Packer M, Bristow MR, Cohn JN, et al. The effect of opment of aortic insufficiency in left ventricular
carvedilol on morbidity and mortality in patients with assist device-­ supported patients. Circ Heart Fail.
chronic heart failure. NEJM. 1996;334:1349–55. 2010;3:668–74.
27. Packer M, Coats AJ, Fowler M, et al. Effect of
42. Soleimani B, Haouzi A, Manoskey A, et al.

carvedilol on survival in severe chronic heart failure. Development of aortic insufficiency in patients
NEJM. 2001;344:1651–8. supported with continuous flow left ventricu-
28. Pitt B, Remme W, Zannad F, et al. Eplerenone, a lar assist devices. Am Soc Artif Internal Organs
selective aldosterone blocker, in patients with left J. 2012;58:323–9.
ventricular dysfunction after myocardial infarction. 43.
Morgan JA, Brewer RJ, Nemeh HW, et al.
NEJM. 2003;348:1309–21. Management of aortic valve insufficiency in
29. Zannad F, McMurray JJ, Krum H, et al. Eplerenone patients supported by long-term continuous flow
in patients with systolic heart failure and mild symp- left ventricular assist devices. Ann Thorac Surg.
toms. NEJM. 2011;364:11–21. 2012;94:1710–2.
30. McMurray JJV, Packer M, Desai AS, et al. Angiotensin- 44. Jorde UP, Uriel N, Nahumi N, et al. Prevalence, sig-
Neprilysin inhibition versus enalapril in heart failure. nificance, and management of aortic insufficiency in
NEJM. 2014;371:993–1004. continuous flow left ventricular assist device recipi-
31. Birks EJ, Tansley PD, Hardy J, et al. Left ventricular ents. Circ Heart Fail. 2014;7:310–9.
assist device and drug therapy for the reversal of heart 45. Tuzun E, Rutten M, Dat M, et al. Continuous-flow
failure. NEJM. 2006;355:1873–84. cardiac assistance: effects on aortic valve function in
32. Kapelios CJ, Charitos C, Kaldara E, et al. Late-onset a mock loop. J Surg Res. 2011;171:443–7.
right ventricular dysfunction after mechanical support 46. Mano A, Gorcsan J, Teuteberg J, et al. Incidence

by a continuous-flow left ventricular assist device. and impact of de novo aortic insufficiency following
J Heart Lung Transplant. 2015;34:1604–10. continuous flow LVADs implantation. J Heart Lung
33. Takeda K, Takayama H, Colombo PC, et al. Incidence Transplant. 2012;31:S22.
and clinical significance of late right heart failure dur- 47. Rajagopal K, Daneshmand MA, Patel CB, et al.

ing continuous-flow left ventricular assist device sup- Natural history and clinical effect of aortic valve
port. J Heart Lung Transplant. 2015;34:1024–32. regurgitation after left ventricular assist device
34. Kormos RL, Teuteberg JJ, Pagani FD, et al. Right implantation. J Thorac Cardiovasc Surg. 2013.
ventricular failure in patients with the HEartMate II 48. Toda K, Fujita T, Domae K, et al. Late aortic insuf-
continuous-flow left ventricular assist device: inci- ficiency related to poor prognosis during left ven-
dence, risk factors, and effect on outcomes. J Thorac tricular assist device support. Ann Thorac Surg.
Cardiovasc Surg. 2010;2010:1316–24. 2011;92:929–34.
35. Garan AR, Iyer V, Whang W, et al. Catheter ablation 49. Westaby S, Bertoni GB, Clelland C, Nishinaka

for ventricular tachyarrhythmias in patients supported T, Frazier O. Circulatory support with attenuated
by continuous-flow left ventricular assist devices. Am pulse pressure alters human aortic wall morphology.
Soc Artif Internal Organs J. 2014;60:311–6. J Thorac Cardiovasc Surg. 2007;133:575–6.
 Surveillance Echocardiography
for LVAD Patients 11
Raymond F. Stainback

protocols and reporting worksheets that can be

Introduction modified to fit an LVAD center’s existing internal
standards.
Guidelines have been previously published for
the use of echocardiography during each of the
various phases of care for left ventricular assist  ole of Echocardiography
R
device (LVAD) patients [1]. This chapter focuses After LVAD Implantation
on the use of surveillance echocardiography
either before hospital discharge or during routine The significant variability among individual
scheduled follow-up appointments. Surveillance patients’ clinical courses after LVAD implanta-
examinations can confirm individual patients’ tion precludes taking a “one-size-fits-all”
baseline echo parameters, which can later be approach to postimplantation echocardiography.
used for problem solving if the patient presents Nevertheless, an overall framework can be rec-
with new abnormal signs or symptoms. Also, ommended to improve patient care while address-
surveillance examinations can detect abnormali- ing the efficiency of both the LVAD clinic and the
ties of the native heart and pump malfunctions echocardiography laboratory. Surveillance echo-
(Table 11.1) that might otherwise remain occult, cardiography generally refers to standard trans-
particularly in less active patients. This chapter thoracic echocardiography (TTE) imaging that is
provides educational material for mechanical performed in the echocardiography lab, in the
circulatory support (MCS) medical and technical LVAD clinic, or at the bedside.
staff team members, along with example imaging In LVAD-supported patients, surveillance
echocardiography is performed at the pump’s
baseline speed setting and includes LVAD-­
Substantial portions of this material were adapted from
specific views and Doppler flow assessments, in
Stainback RF, Estep JD, Agler DA, et al. Echocardiography
in the management of patients with left ventricular assist addition to all the elements of a standard TTE
devices: Recommendations from the American Society exam for heart failure (HF) patients. If an LVAD
of Echocardiography. J Am Soc Echocardiogr. 2015; optimization protocol is added, further limited
28(8):853–909. Used with permission.
imaging can be performed at pump speeds above
R.F. Stainback, M.D. (*) or below the baseline speed to optimize LVAD
Non-invasive Cardiology, Texas Heart Institute at
and native heart function, although standards for
Baylor St. Luke’s Medical Center, Baylor College of
Medicine, Houston, TX, USA optimal pump speeds may vary among centers
e-mail: rstainback@[Link] and according to patient-specific variables.

© Springer International Publishing AG 2018 161

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
162 R.F. Stainback

Table 11.1  Continuous-flow LVAD postimplant complications and device dysfunction detected by echocardiography
Pericardial effusion
With or without cardiac tamponade including RV compression. Tamponade: respirophasic flow changes; poor
RVOT SV
LV failure secondary to partial LV unloading
(by serial exam comparison)
 (a)  2D/3D: increasing LV size by linear or volume measurements; increased AV opening duration, increased
left atrial volume
 (b)  Doppler: increased mitral inflow peak E-wave diastolic velocity, increased E/A and E/e′ ratio, decreased
deceleration time of mitral E velocity, worsening functional MR, and elevated pulmonary artery systolic
pressure
RV failure
 (a)  2D: increased RV size, decreased RV systolic function, high RAP (dilated IVC/leftward atrial septal shift),
leftward deviation of ventricular septum
 (b)  Doppler: increased TR severity, reduced RVOT SV, reduced LVAD inflow cannula and/or outflow-graft
velocities (i.e., <0.5 m/s with severe failure), inflow-cannula high velocities if associated with a suction
event. Note: a “too-high” LVAD pump speed may contribute to RV failure by increasing TR (septal shift)
and/or by increasing RV preload
Inadequate LV filling or excessive LV unloading
Small LV dimensions (typically <3 cm and/or marked deviation of interventricular septum toward LV). Note: may
be due to RV failure and/or pump speed too high for loading conditions
LVAD suction with induced ventricular ectopy
Underfilled LV and mechanical impact of inflow cannula with LV endocardium, typically septum, resolves with
speed turndown
LVAD-related continuous aortic insufficiency
Clinically significant—at least moderate and possibly severe—characterized by an AR proximal jet-to-LVOT height
ratio >46% or AR vena contracta ≥3 mm; increased LV size and relatively decreased RVOT SV despite normal/
increased inflow-cannula and/or outflow-graft flows
LVAD-related mitral regurgitation
 (a)  Primary: inflow-cannula interference with mitral apparatus
 (b)  Secondary: MR functional, related to partial LV unloading/persistent heart failure
Note: Elements of both a and b may be present
Intracardiac thrombus
Including right and left atrial, LV apical, and aortic root thrombus
Inflow-cannula abnormality
 (a)  2D/3D: small or crowded inflow zone with or without evidence of localized obstructive muscle
trabeculation, adjacent MV apparatus, or thrombus; malpositioned inflow cannula
 (b)  High-velocity color or spectral Doppler at inflow orifice. Results from malposition, suction event/other
inflow obstruction: aliased color-flow Doppler, CW Doppler velocity >1.5 m/s
 (c)  Low-velocity inflow (markedly reduced peak systolic and nadir diastolic velocities) may indicate internal
inflow-cannula thrombosis or more distal obstruction within the system. Doppler flow-­velocity profile may
appear relatively “continuous” (decreased phasic/pulsatile pattern)
Outflow-graft abnormality
Typically due to obstruction/pump cessation
 (a)  2D/3D imaging: visible kink or thrombus (infrequently seen)
 (b)  Doppler: peak outflow-graft velocity ≥2 m/sa if near obstruction site; however, diminished or absent
spectral Doppler signal if sample volume is remote from obstruction location, combined with lack of RVOT
SV change and/or expected LV-dimension change with pump speed changes
Hypertensive emergency
New reduced/minimal AV opening relative to baseline exam at normal BP, especially if associated with new/
worsened LV dilatation and worsening MR. Note: hypertension may follow an increase in pump speed
(continued)
11  Surveillance Echocardiography for LVAD Patients 163

Table 11.1 (continued)
Pump malfunction/pump arrest
 (a)  Reduced inflow-cannula or outflow-graft flow velocities on color and spectral Doppler or with pump arrest
shows diastolic flow reversal
 (b)  Signs of worsening HF, including dilated LV, worsening MR, worsened TR, and/or increased TR velocity;
attenuated speed-change responses: decrease or absence of expected changes in LV linear dimension, AV
opening duration, and RVOT SV with increased or decreased pump speeds; for HVAD, loss of inflow-
cannula Doppler artifact
2D two dimensional, 3D three dimensional, A mitral valve late peak diastolic velocity, AR aortic regurgitation, AV aortic
valve, BP blood pressure, CW continuous wave, E mitral valve early peak diastolic velocity, e′ mitral annular velocity,
HVAD HeartWare ventricular assist device, IVC inferior vena cava, LV left ventricular, LVAD left ventricular assist
device, LVOT left ventricular outflow tract, MR mitral regurgitation, MV mitral valve, RAP right atrial pressure, RV right
ventricular, RVOT right ventricular outflow tract, SV stroke volume, TR tricuspid regurgitation. Adopted and modified
from Estep et al. [15]
a
Note: based on observational data. The “normal” outflow-­graft peak velocities are not well defined, because the HVAD
outflow-graft diameter is smaller than that of the HM-II device (see discussion in text). Therefore, the normal Doppler-
derived HVAD outflow velocities may be somewhat higher on average than those observed for the HM-II LVAD

Fig. 11.1  Sample schedule for initial and follow-up surveillance echocardiography of patients with no evidence of
device malfunction

Patients with an uncomplicated postoperative schedule for timing postimplantation surveil-

course (i.e., absence of HF symptoms, success- lance TTE.
ful weaning from intravenous pharmacologic Comparing serial surveillance exam results
inotropic and vasopressor agents within 14 days, with each other (for an individual patient) or to
absence of LVAD controller alarms, and no sero- population-based benchmarks can also help the
logic evidence of hemolysis or infection) should examiner understand a patient’s response to
undergo follow-up surveillance TTE at prespeci- LVAD therapy over time. Moreover, surveil-
fied intervals. Periodic LVAD surveillance echo lance data may allow early diagnosis of occult
exams are recommended, to establish patient-­ native heart abnormalities (e.g., development
specific “baseline” parameters for both LVAD of LVAD-­related aortic regurgitation [AR]) or
and native heart function. An LVAD surveillance device-­related problems, including a drift from
echo exam should be considered at approxi- previously optimal device speed settings. When
mately 2 weeks after device implantation or surveillance TTE is coordinated with the
before discharge from the index hospitalization patient’s routine LVAD clinic visits, HF special-
(whichever occurs first). Additional routine sur- ists can integrate the information obtained into
veillance TTE should be considered at 1, 3, 6, their clinical assessments and care plans. A
and 12 months after implantation and every putative benefit of routine LVAD surveillance
6–12 months thereafter; however, at this time, no echocardiograms (with optimization protocols
outcome data support a specific timetable. when indicated) is better patient outcomes,
Figure  11.1 summarizes a reasonable sample including early detection and ­ treatment of
164 R.F. Stainback

complications and fewer hospitalizations for which no pulse is p­ resent, a mean arterial BP
recurrent HF. can be obtained with a standard BP cuff, along
with a handheld audible Doppler device for
Key Points detecting brachial or radial artery flow [2].
• Patients with an uncomplicated postopera- Note that the arterial Doppler-derived BP read-
tive course should undergo LVAD surveil- ing lies between the systolic pressure and the
lance echocardiography at predetermined mean arterial BP [3]. For practical purposes, if
intervals after LVAD implantation to assess the patient has a pulse (i.e., the aortic valve
the patients’ response to mechanical circula- [AV] is opening), the Doppler-derived BP is
tory support (MCS) therapy and to screen the same as the systolic BP. If the patient does
for the development of subclinical not have a pulse (i.e., the AV is not opening),
complications. the Doppler BP is considered to be the mean
• When possible, LVAD surveillance echocar- arterial BP.
diography should be coordinated with routine A current BP measurement is necessary for
LVAD clinic visits. accurate echo interpretation and for safety rea-
sons during “speed change” protocols, particu-
larly when changing to higher speed settings.
 linical Data Acquisition Standards
C Susceptible patients may develop clinically sig-
and Sonographer Reproducibility nificant hypertension in response to increased
LVAD flow, and a mean arterial pressure of
Before initiating any LVAD echo exam, sonogra- <85 mmHg is recommended [4]. Hypotension is
phers should always annotate the LVAD type and generally defined as a mean arterial pressure of
baseline LVAD speed in rotations per minute <60 mmHg and may be associated with tradi-
(rpm) on the imaging screen (Fig. 11.2), in addi- tional symptoms or signs of hypoperfusion. With
tion to the standard patient demographic data. If CF-LVADs, one challenge is that a sonographer
the device speed is changed, this should be noted (or some other trained and available individual)
during the exam. The device type and speed needs to be facile at obtaining an arterial Doppler-­
information should also be routinely incorpo- derived BP reading. To facilitate the care of
rated into reporting templates. CF-LVAD recipients, better BP monitoring tech-
niques may be needed [5].

Blood Pressure Key Points

• Although BP readings can be challenging
The patient’s blood pressure (BP), which to obtain in LVAD patients, this variable is
reflects peripheral vascular resistance, is an important because it significantly influ-
important parameter that greatly influences ences echo findings and their
ventricular unloading and the observed echo- interpretation.
cardiographic findings. Therefore, BP should • In the absence of a palpable pulse, BP mea-
be recorded just before the exam and immedi- surement may require audible Doppler inter-
ately afterward if pump speed changes were rogation by an appropriately trained individual
made. Patients with a continuous-flow LVAD before the echo exam.
(CF-LVAD) have a reduced and narrowed pulse • Susceptible patients can have marked hyper-
pressure and may not have a palpable pulse. tension after the LVAD pump speed is
Therefore, cuff-based BP assessment may be increased. Therefore, BP measurement should
difficult or impossible. In the intensive care be repeated after a significant pump speed
unit, BP may be obtained from invasive arterial increase, particularly if the patient’s BP is
monitoring devices. In other circumstances in elevated at the baseline pump speed.
Fig. 11.2  Side-by-side comparison of multiple imaging increased size by LVIDd after LVAD thrombosis; (c) AV
metrics in the same patient before and after HM-II LVAD M-mode, minimal opening (107 ms) during normal LVAD
impeller thrombosis. (a) LVIDd, normal LVAD; (b) function; (d) markedly increased AoV opening duration
166 R.F. Stainback

Fig. 11.2  (continued) (230 ms) after internal LVAD images, respectively, with nearly absent diastolic flow
thrombosis; (e, g) inflow-­cannula color-flow (arrow) and (view h) after development of impeller thrombosis; (i)
pulsed Doppler images, respectively, during normal RVOT pulsed Doppler VTI = 15 cm during normal LVAD
LVAD function; (f, h) very-­low-­velocity inflow-cannula function; (j) RVOT pulsed Doppler VTI = 7.9 cm after
systolic flow on color-flow (arrow) and pulsed Doppler LVAD thrombosis. Inflow inflow cannula; vel. velocity

• A mean arterial BP of <85 mmHg is this scenario, correlating the images to the elec-
recommended. trocardiographic signal can be helpful.
• Hypotension is generally defined as a mean Additionally, using a microbubble contrast agent
arterial pressure <60 mmHg. It may be associ- should be seriously considered when endocardial
ated with traditional symptoms or signs of definition is insufficient for accurate LVIDd mea-
hypoperfusion. surement [7]. Previous data from HeartMate II
(HM-II)-supported outpatients in stable condition
suggest that the LVIDd is probably at least 15%
LV Size and Systolic Function lower than the preimplantation value 3 months
after implantation [8, 9]. Care must be taken to
Linear and volumetric approaches for determin- correlate LV end-systolic versus end-­ diastolic
ing LV size and systolic function in non-LVAD diameters with the electrocardiographic signal.
patients have been described by Lang and col- Paradoxically, the LVIDd may be smaller than the
leagues [6]. These methods may or may not be left ventricular internal dimension at end-systole
appropriate for LVAD patients, as outlined below. (LVIDs); this is an important finding, as it may
indicate excessive LVAD unloading, severe right
ventricular (RV) dysfunction, or both.
LV Size Left ventricular volumes, as determined by the
Simpson’s biplane or single-plane method of
As mentioned above, the left ventricular internal disks (Fig. 11.3), reflect the LV size more accu-
dimension at end-diastole (LVIDd) from the 2D rately than do linear measurements. However,
parasternal long-axis image is considered the measuring the LV size by volume may be techni-
most reproducible measure of LV size in LVAD cally challenging after LVAD implantation
recipients (Fig. 11.2a, b). In patients with a nor- because of apical shadowing/dropout associated
mally functioning CF-LVAD, severely depressed with the inflow cannula. This is one reason why
native LV function, and altered mitral valve open- postimplantation LV volumes assessed by echo-
ing, determining end-diastole may be difficult. In cardiography are smaller than those assessed by
11  Surveillance Echocardiography for LVAD Patients 167

Fig. 11.3 Left ventricular end-diastolic volume linear measurements (e.g., Fig. 11.2). The inflow can-
(LVEDV), as measured by Simpson’s biplane method nula (arrow) and anterolateral papillary muscle
of disks. This method is preferred for LV size assess- (Asterisk) are excluded from the endocardial tracing.
ment when possible. Simpson’s single-plane LVEDV Note: In view b, aneurysmal remodeling of the LV apex
method (using the best-/least-foreshortened (a) four- (relative to the LV base), which would cause underesti-
chamber [4Ch] or (b) two-chamber [2Ch] view) may mation of LV size by parasternal long-axis-view linear
suffice for LV size assessment and may be superior to measurements (e.g., Fig. 11.2a, b)

cardiac computed tomography [10]. A reasonable rately and meaningfully assessing overall LV
LV diastolic volume assessment is possible in systolic function from the patient’s LV ejection
many ambulatory LVAD patients, and this metric fraction (LVEF). The limitations of LVEF mea-
can be incorporated into the surveillance exam, surement are both technical (with regard to imag-
particularly at the baseline pump speed setting. ing quality) and physiologic. The LV endocardium
However, LVIDd measurement, being more may be difficult to visualize because of apical
expediently acquired and reproducible, is practi- foreshortening, apical shadowing from the
cal for tracking the relative LV size over time at a device, or acoustic dropout (signal attenuation).
baseline pump speed (e.g., Fig. 11.2a, b vs. Physiologic challenges related to the LVAD
Fig.  11.3) and in the context of a speed-change include enhanced interventricular dependence
exam (see below) for quick problem solving. and discordant septal and inferolateral wall
That the serial LVIDd measurement (combined motion, which can vary considerably in the same
with the degree of AV opening) can be used as a patient at different pump speeds. If the LV endo-
surrogate marker for the degree of LV unloading cardium, including the apex, can be adequately
in CF-LVAD patients seems intuitive and is sup- visualized, with or without a microbubble con-
ported by the limited available literature, which is trast agent, the preferred method for calculating
derived primarily from HM-II studies. However, the LVEF is the biplane method of disks (modi-
robust outcome data are limited, and applicability fied Simpson rule; Fig. 11.3) [6]. Although other
to patients with a HeartWare LVAD (HVAD), parameters for LV systolic function can be con-
from whom there is less evidence, has not been sidered, LVEF is an important surrogate because
demonstrated at this time [11]. it can reveal possible LV worsening or recovery.
Therefore, surveillance and recovery LVAD
exam reports should include an LVEF assess-
LV Systolic Function ment, even if only a qualitative assessment is
possible. However, LVAD support markedly
Accurately determining LV volumes is challeng- reduces LV afterload, an important determinate
ing after device implantation. So, too, is accu- of LVEF. Therefore, the value of LVEF for
168 R.F. Stainback

d­etermining systolic function during LVAD u­ nloading that can be tracked over time and at
support must be taken into consideration during different pump speeds.
clinical decision-making. • The LV end-diastolic volume is a more accu-
Other methods: In patients with suboptimal rate representation of LV size than is the
apical but adequate parasternal views, the follow- LVIDd.
ing methods for measuring LV systolic function • After LVAD implantation, measuring LV
may be considered, although their accuracy has volumes and the LVEF can be technically
not been validated in LVAD patients: challenging. When the LVEF needs to be
obtained (particularly to assess for LV recovery),
1. The LV fractional area change (FAC) method at the Simpson biplane method of disks is rec-
the mid-papillary muscle level on 2D short-­axis ommended for use when possible.
views: FAC (%) = [(end-diastolic area − end-
systolic area)/(end-diastolic area)] [12].
2. The Quinones method for determining the
LV Diastolic Function
LVEF [13], with the assumption of an akinetic
apex given the presence of the apical inflow It can be assumed that LVAD patients have mark-
cannula. edly abnormal baseline diastolic function. Although
3 . The LV fractional shortening (%) method: the standard LV diastolic function parameters [16]
FS = [(LVIDd − LVIDs)/(LVIDd)], where can (and, in the context of clinical research, should)
FS = fractional shortening and LVIDs = the be measured and included in the report, there is a
LV internal dimension at end-systole [6], paucity of data validating their clinical usefulness
which has been used in LVAD patients in patients receiving LVAD support. However, the
[10, 14]. LVAD echo report should not include an assess-
ment of LV diastolic function, which has not been
The linear and volume measurements of sys- validated in LVAD-­supported patients. Using cer-
tolic function noted above represent possible tain acquired diastolic parameters could be helpful,
methods of tracking the course of individual particularly when they are correlated with symp-
patients, serving as their own controls, over time. toms in individual patients. This should be done at
However, routinely using the three methods the discretion of the interpreter, because these
described above may not be feasible or advisable parameters may reflect changes in the degree of LV
for many LVAD patients because of segmental unloading when compared with data recorded dur-
wall motion abnormalities, exaggerated paradox- ing previous examinations or at different pump
ical septal motion, ventricular dyssynergy, or speeds during the same exam.
ventricular septal shift, the extent of which could Previous data suggest that the mitral E velocity
change at varying pump speeds in the same (cm/s), left atrial volume (mL), pulmonary vascu-
patient. Note that calculating the LVEF from the lar resistance (Wood units), and pulmonary artery
LV stroke volume is not recommended, because systolic pressure (mmHg) are significantly reduced
many LVAD patients have beat-to-beat variations and that the mitral deceleration time (ms) is sig-
in this parameter [15]. Previous data suggest that nificantly prolonged in outpatients whose condi-
most outpatient HM-II recipients in stable condi- tion is stable 3–6 months after HM-II implantation
tion have persistent moderately to severely [8, 9]. How these parameters should be integrated
depressed LV systolic function during the first into postimplantation clinical management is cur-
6 months after device implantation [8, 9]. rently unclear, as is their prognostic value for
patient outcomes. For clinical LVAD echo report-
Key Points ing purposes, a practical approach at this time may
• After CF-LVAD activation, the LVIDd may be be to use the following (or a similar) statement:
the most reproducible measure of LV “Interpretation of the degree of LV diastolic
11  Surveillance Echocardiography for LVAD Patients 169

d­ ysfunction (presumed abnormal) is not provided at any LVAD speed in patients with extremely
because of continuous-flow LVAD support.” poor native LV function. The frequency of AV
opening is most accurately assessed by recording
Key Points multiple (5 or 6) cardiac cycles at a slow M-mode
• It can be assumed that LVAD patients have sweep speed (e.g., 25–50 mm/s) (Fig. 11.4); the
markedly abnormal baseline diastolic valve should be characterized as opening with
function. every cardiac cycle, opening intermittently, or
• How LV diastolic parameters should be inte- remaining closed [15, 22].
grated into the interpretation and report of an Many HF teams also request that the duration
LVAD echocardiography examination and any of AV opening be measured (in ms) from the same
resulting treatment decisions is currently M-mode acquisitions. This parameter may vary
undefined, as is these parameters’ prognostic from beat to beat, so it is best to measure several
value for patient outcomes. beats and report an average value. When the AV
opening duration is relatively constant, a faster
sweep speed (e.g., 75–100 mm/s) may be appro-
RV Size and Systolic Function priate (Fig. 11.4). An important potential pitfall of
using M-mode to assess the presence and duration
Many of the standard measures of RV size and sys- of aortic cusp separation is illustrated in Fig. 11.5.
tolic function [17], including linear dimensions, In this example, AV semilunar cusp conformation,
RV FAC, tricuspid annular plane systolic excur- combined with cardiac translational motion or
sion (TAPSE), and right-sided cardiac output, can slightly off-axis imaging, can create the false
feasibly be measured in LVAD patients [8, 18, 19]. appearance of aortic cusp separation when the
However, recent data suggest that the correlation cusps are not actually separating. Careful attention
of TAPSE with overall RV systolic function may and the additional use of color M-mode may be
be weaker after cardiothoracic surgery, so this useful in difficult cases to avoid M-mode “pseudo
variable may have less clinical utility than the AV opening” or an exaggerated AV opening dura-
other measures [20]. Current data regarding the tion. However, in some cases of “minimal” AV
expected response of RV systolic function after opening, the duration of AV cusp separation and
LVAD implantation are conflicting: one study the duration of forward systolic flow are not
showed a significant improvement in RV FAC at always the same, and using color M-mode can
3 months [9], but another study did not show a sig- help to document this finding (Fig. 11.6).
nificant difference in this parameter at either 1 or 6 In patients whose AV remains closed, it is
months [8]. important to evaluate for aortic root thrombus,
which may be transient or associated with com-
missural fusion. Continuously closed aortic cusps
Valvular Assessment have been associated with the development of
aortic root thrombosis and LVAD-associated AR
Aortic Valve [23], as discussed below. Fusion of the aortic
cusps, either surgical or secondary to chronic
Evaluating and reporting the degree of AV open- aortic cusp closure, can be recognized on speed-­
ing (if any) are important because it is affected by change echocardiograms (discussed below).
several other parameters, including LVAD speed, New-onset (“de novo”) AR is found in
native LV function, volume status, and peripheral approximately 25–33% of patients 12 months
vascular resistance. In addition, whether the AV after LVAD implantation [24, 25] and is a key
opens may have clinical implications. Whereas finding, given its adverse effects on LVAD per-
recent guidelines recommend that the LVAD formance and its association with morbidity and
speed be set low enough to allow at least inter- mortality [26–28]. Several studies suggest that
mittent AV opening [21], opening may not occur persistent AV closure is a risk factor for de novo
170 R.F. Stainback

Fig. 11.4  The duration of AV opening during LVAD sup- mal finding at a high LVAD pump speed (9800 rpm). (c–e)
port can be easily measured by using M-mode during either The expected progressively reduced duration of AV open-
TEE (a) or TTE (b). In view a, the AV “barely opens” inter- ing in the same patient during a ramp (speed-change) echo
mittently (arrows); this maybe related, in part, to an exam at different HM-II pump speeds: In view c (8000 rpm),
arrhythmia and suggests normal LVAD function at a pump the AV “barely opens,” in view d (8600 rpm), the AV “opens
speed of 9600 rpm. In view b, there is near-­normal AV intermittently” (arrows); in view e (9000 rpm), the AV
opening, with durations of >200 ms; this may be an abnor- “remains closed”

Fig. 11.5  Images suggestive of an exaggerated or “false” the cusp closure line (view b: red line), and translational
AV opening duration, as assessed by M-mode. This artifact motion of the aortic root. This pitfall could have negative
should be suspected when the aortic cusp opening shape is implications when the examiner relies solely on M-mode
fusiform (a). Although the apparent M-mode AV opening for selecting the AV closing speed during an LVAD optimi-
duration in this case appeared to be >200 ms (arrows), there zation protocol. M-mode should not be used in isolation.
was, in fact, little or no AV opening. (b) This error was due False M-mode AV opening can be identified by correlating
to several factors, including the semilunar shape of the AV M-mode findings with the 2D image and color M-mode (in
cusps, placement of the interrogating cursor to the left of the presence of AR) to validate the extent of AV opening
11  Surveillance Echocardiography for LVAD Patients 171

AR after LVAD implantation, even when no aor- nificant MR after LVAD support begins may
tic root thrombus is present (Fig. 11.7) [24, 29, indicate inadequate LV unloading or inflow-can-
30]. Standard methods for quantifying AR [31] nula malposition and interference with the sub-
may be challenging to use after LVAD implanta- mitral apparatus. If MR is present, it can be
tion. In the absence of definitive cutoff criteria to quantified by standard methods [32]. Incidental
define mild, moderate, and severe AR after post-LVAD MR may also represent LVAD mal-
LVAD implantation, one should perform an function and should be discussed with the clini-
aggregate assessment based on duration (pre- cal team.
dominantly diastolic vs. continuous AR flow by
spectral Doppler), AR jet vena contracta width
(significant AR vena contracta ≥3 mm, see
­ Tricuspid and Pulmonary Valves
Table 11.1), the presence or absence of holodia-
stolic flow reversal in the descending thoracic Like MR, moderate or greater tricuspid regurgi-
aorta on suprasternal notch views, jet height rel- tation (TR) is an important finding on LVAD
ative to the LV outflow tract (LVOT), compara- surveillance echocardiography, because this
tive LVAD and native-circuit flow measures condition can be associated with insufficient LV
(Figs.  11.8 and 11.9), and LV chamber size. unloading (functional TR), excessive LV
Additionally, significant AR noted on LVAD sur- unloading with a leftward shift of the interven-
veillance echocardiography can be further evalu- tricular septum (e.g., a suction event), elevated
ated with device controller data and the cardiac systolic pulmonary pressures, or intrinsic RV
response during LVAD problem-focused echo- systolic dysfunction. Distinguishing among
cardiography with speed changes, as described these causes by using echocardiographic param-
below. eters is discussed in further detail below.
Regardless of the cause, TR after LVAD implan-
Key Points tation can generally be assessed with standard
• Recording multiple cardiac cycles with color methods [32]. Furthermore, the native pulmo-
M-mode at a sweep speed of 25–50 mm/s is nary valve typically remains functionally nor-
recommended to accurately assess the fre- mal after LVAD implantation and can be
quency and duration of AV opening. interrogated by using standard methods when
• Persistent AV closure can be associated with significant stenosis or regurgitation is suspected
aortic root thrombus and de novo AR. [31, 32].
• If aortic root thrombus is suspected, decreas-
ing the LVAD pump speed (e.g., during a
planned speed-change exam) should be Interventricular Septal Position
avoided, because it could result in sudden AV
opening. The end-diastolic interventricular septal position,
• After LVAD implantation, AR is not uncom- which is dependent on the interventricular pres-
mon. Assessment of AR severity is partly sure gradient, should be routinely reported as neu-
based on careful color Doppler analysis in the tral, leftward-shifted, or rightward-shifted. A
parasternal long-axis view. leftward shift can be due to elevated RV end-­
diastolic pressures, reduced LV preload, or LV
over-decompression resulting from excessive
Mitral Valve LVAD speed; differentiation among these causes
is further discussed below. A rightward shift is
As noted above, LV unloading generally leads to generally due to elevated LV end-diastolic pres-
reduced mitral valve annular dilatation, improved sures resulting from an inadequate LVAD speed
leaflet coaptation, and, ultimately, reduced mitral setting, pump dysfunction, severe AR, or an
regurgitation (MR) severity. Persistence of sig- increased LV afterload.
172 R.F. Stainback

Fig. 11.6  Assessment of AR. (a) TEE shows at least mod- and the jet width/LVOT width is clearly >46%. Color-flow
erate—and possibly severe—continuous AR during LVAD Doppler reveals inflow-cannula systolic entrainment of the
support. The AR vena contracta (VC) is clearly >3 mm, AR jet (arrow). A closed MV and trace MR (Asterisk) are
11  Surveillance Echocardiography for LVAD Patients 173

I nflow-Cannula and Outflow-Graft The inflow cannula should be routinely inter-

Interrogation rogated with CW spectral Doppler at the baseline
pump speed and particularly during speed-change
Inflow Cannula exams (discussed below) to screen for inflow
obstruction. Note that in many cases, a normal
Usually, the apically inserted inflow cannula can inflow-cannula spectral Doppler flow-velocity
be adequately imaged in standard or modified 2D profile can be contaminated by low-velocity dia-
parasternal and apical TTE views. The sonogra- stolic AR or mitral inflow (Fig. 11.12). Moreover,
pher’s objective is to reveal the inflow cannula’s during TEE evaluation of the inflow cannula, the
location and orientation in relation to the inter- CW Doppler signal can be contaminated by MR
ventricular septum and other LV structures. The (Fig.  11.5c). The HVAD inflow-cannula flow
inflow cannula can often be visualized with 3D velocities typically cannot be evaluated by using
echo techniques, and this approach can be used as either color or spectral Doppler because of a
a complementary imaging method by examiners characteristic Doppler artifact (Fig. 11.13) related
experienced in 3D imaging. As noted above in the to the inflow cannula’s direct connection to the
section on perioperative TEE, color Doppler adjacent impeller housing.
interrogation of a properly aligned inflow cannula
should reveal laminar, unidirectional flow from
the ventricle to the inflow cannula, with no evi- Outflow Graft
dence of turbulence or regurgitation [33]. Pulsed
and continuous-wave (CW) spectral Doppler In contrast to inflow-cannula imaging, visualiz-
interrogation may require “off-axis” modification ing the outflow graft requires the use of atypical
of a standard parasternal, apical, or short-axis echocardiographic windows. The terminal por-
TTE view to achieve true coaxial alignment tion of the outflow conduit and its anastomosis to
between the sampling beam and inflow-cannula the aorta can generally be visualized from a high
flow; such interrogation should additionally left parasternal long-axis view (Figs. 11.8 and
reveal the flow to have a low peak velocity 11.14). The midportion of the outflow graft is
(<1.5 m/s). Because of native LV contractility, best visualized from a right parasternal view
cannula flow generally remains pulsatile to some while the patient is in a right lateral decubitus
degree even when the AV does not open [9, 15]. position. Color Doppler and spectral Doppler
Recording both the peak systolic and the nadir interrogations are usually possible from these
diastolic velocities over at least 3 or 4 cardiac views; as with the inflow cannula, recording both
cycles is recommended (Figs. 11.10 and 11.11). the peak systolic and nadir diastolic velocities
over at least 3–5 cardiac cycles is recommended

Fig. 11.6  (continued) indicative of marked systolic AR. RVOT, right ventricular outflow tract. (b, c) During LVAD
support, at least moderate continuous AR (arrow) is observed in the transthoracic parasternal long-axis view with color
Doppler (b) and color M-mode imaging (c); the inflow cannula is denoted by an asterisk. In view c, note the variance in
the early systolic (arrowhead) versus late systolic (arrow) AR VC width, as shown by M-mode. This finding is not
consistent among different patients; it is likely influenced by several variables and by the fact that the AV cusps can
exhibit augmented systolic opening, despite AR, at speeds close to (but less than) the AV “opening speed.” (d–e) The
AR VC width may increase at higher pump speeds in the same patient, as seen here. This may partially be due to an
increased systemic arterial pressure at higher pump speeds, which presumably increases the AR volume. At both speeds,
the VC is >3 mm, indicating at least moderate—and possibly severe—AR. The VC width is 4.2 mm at 8600 rpm in view
d and is 5.7 cm at 9600 rpm in view E (HM-II LVAD). (f) “Continuous” holosystolic and holodiastolic AR, as detected
by continuous-wave Doppler (TTE apical 5-chamber view). (g) Continuous-wave Doppler (TTE apical 5-chamber
view) reveals nearly continuous AR, which significantly extends into the electrical and mechanical systolic period with
a brief period of AV systolic forward flow (arrows). (h) Color M-mode shows minimal AV opening, with a brief dura-
tion of low-velocity systolic forward flow (arrows). (i) TTE parasternal long-­axis view of an AR jet on color-flow
Doppler imaging (arrow). (j) The AV opens widely, with forward flow that interrupts AR. However, the AR period
extends into the electrical and mechanical systolic period (arrows) during HVAD pumping at 2600 rpm
174 R.F. Stainback

Fig. 11.7  De novo AR after LVAD implantation. This at 14 months (c). All images are transthoracic parasternal
condition progressed from no AR on the baseline surveil- long-axis views with color Doppler. In this patient, the AV
lance study exam at 1 week (a) to trivial AR (arrow) at never opened at any pump speed during the LVAD support
1 month (b) to at least moderate AR (arrows, VC >3 mm) period; aortic root thrombus was not present

Fig. 11.8  Direct Doppler measurement of LVAD flow flow graft (LVAD output) can be derived by measuring the
from the distal outflow graft, as evaluated by TTE (a). graft’s diameter (arrow) and the pulsed Doppler VTI at
Flow (stroke volume and cardiac output) within the out- the same location, proximal to the anastomosis site (b)

(Fig. 11.14), depending on the uniformity of the dad vs. cephalad) within the graft. There is no
spectral Doppler signal. Note that the outflow-­ standard recommendation for a positive-versus-­
graft flow-velocity profile will appear either negative outflow-graft display other than to pro-
above or below the baseline in the spectral vide the most coaxial alignment and to ensure
Doppler display, depending on the sonographer’s that the flow direction (caudad vs. cephalad) is
positioning of the sample volume direction (cau- apparent. In some patients, the outflow graft may
11  Surveillance Echocardiography for LVAD Patients 175

Fig. 11.9 The total cardiac output (combined LVAD pulmonary regurgitation and to measure the RVOT VTI.
flow output and native LVOT flow output [if any]) is the Note: In the case shown above, the RVOT VTI is low
same as the RVOT cardiac output. The RVOT cardiac (7–9 cm) at a relatively high HM-II pump speed of
output is measured by using standard imaging tech- 9600 rpm; this was consistent with a low cardiac output,
niques including (a) measurement of the RVOT (pulmo- which was due to an obstructed (kinked) outflow graft. It
nary annulus) diameter (d). Color-flow (b) and spectral may be useful to average 3–5 VTIs, depending on their
Doppler (c) studies are performed to rule out significant variability

Fig. 11.10  After LVAD implantation, (a) TEE shows that Pulsed Doppler interrogation of the inflow cannula shows
the inflow cannula is somewhat directed toward the ven- a typical continuous, systolic dominant inflow pattern.
tricular septum (arrow). This can be acceptable but may Dashed arrow = peak systolic velocity; X = nadir diastolic
predispose to inflow-cannula obstruction after sternal clo- velocity. (c) Continuous-wave spectral Doppler interroga-
sure or later reduction in LV size. However, cannula posi- tion of the inflow cannula (to screen for inflow obstruction)
tion and flow velocities are shown to be acceptable shows normal inflow-cannula systolic flow (black arrow);
(normal) in this case. Simultaneous orthogonal plane “+” indicates a hybrid signal that results from overlapping
imaging reveals unobstructed, laminar inflow-cannula flow of continuous diastolic inflow-cannula flow and diastolic
on 2D and color-flow Doppler (blue) examination. (b) MV inflow; “*” indicates MR velocity

be visualized in subcostal or sternal notch views, pulsed Doppler time velocity integral (TVI)
depending on the body habitus. At similar flow combined with the expected or measured
rates, normal flow velocities within the HM-II outflow-­
graft area may be used to measure
outflow graft (16-mm diameter) are somewhat LVAD flow directly (see Fig. 11.8 and the dis-
lower than those within the smaller-caliber cussion below).
HVAD outflow graft (10-mm diameter).
Otherwise, phasic holosystolic and holodiastolic Key Points
laminar flow-velocity patterns should be similar • When 2D imaging is inconclusive, 3D echo-
between the two devices. The outflow-graft cardiography can help delineate the
176 R.F. Stainback

Fig. 11.11  Normal HM-II inflow-cannula flow, assessing flow and pulsed spectral Doppler profiles (b) are consis-
which, in this case, required using a modified parasternal tent with laminar flow. (c) Peak systolic velocity = 1.0 m/s
long-axis view for coaxial alignment of the sampling vol- (dotted arrow) and nadir diastolic velocity = 0.3 m/s (solid
ume. (a) Color-flow Doppler with pulsed Doppler sample arrow) as assessed by continuous-wave spectral Doppler
volume at the inflow cannula’s inflow zone. The color-­ (Normal peak inflow velocities are typically <2 m/s)

Fig. 11.12 (a) HM-II inflow cannula, systolic frame, show- (d) Continuous-wave Doppler shows a similar pattern and
ing normal color Doppler inflow (blue, downward arrow). rules out obstruction. Note: Hybrid/contaminated inflow-
(b) HM-II inflow cannula, diastolic frame, with color cannula Doppler signals may also be observed with AR jets.
Doppler showing prominent diastolic mitral inflow (orange, These types of low-velocity, normal-variant, contaminated
upward arrow) in a patient with a previous mitral annulo- inflow-­cannula spectral Doppler patterns can be explained
plasty repair. (c) Pulsed Doppler examination of the inflow with color Doppler and should not be confused with higher-­
cannula shows normal systolic inflow (dotted line). However, velocity signals (typically >2 m/s), which could signify
prominent bidirectional diastolic velocities are present inflow obstruction. Nonetheless, the pure continuous dia-
because of mitral inflow (arrow) and interaction between the stolic inflow, as shown in Fig. 11.11, is not seen, and the dia-
cannula and the adjacent interventricular septum (Asterisk). stolic nadir velocity cannot be reported

r­ elationship of the inflow cannula to the inter- from 3 to 5 cardiac cycles, depending on the
ventricular septum and other LV structures. regularity of the spectral Doppler contour.
• In patients with an HM-II LVAD, peak sys- • Outflow-graft velocities of >2 m/s at any level
tolic and nadir diastolic inflow-cannula and may be abnormal and suggest possible obstruc-
outflow-graft velocities can be derived from tion, although benchmark data are lacking.
coaxially aligned spectral Doppler.
• HM-II inflow-cannula peak systolic flow veloc-
ities are typically <1.5 m/s. Higher velocities  ative Heart Versus LVAD Flow
N
suggest possible inflow-cannula obstruction. Assessment
• HVAD inflow-cannula velocities cannot be
accurately measured because of a characteris- In the absence of significant pulmonary valve
tic Doppler artifact. regurgitation, the net cardiac output (combined
• Peak systolic and nadir diastolic inflow-­cannula native LV outflow and LVAD conduit flow) is the
and outflow-graft velocities should be derived same as the right-sided cardiac output. The
11  Surveillance Echocardiography for LVAD Patients 177

Fig. 11.13  HVAD inflow-cannula artifact. (a) Artifact hibits spectral Doppler interrogation of the inflow cannula.
(arrow) is visible in a 2D parasternal long-axis view on Inflow-cannula flow must be surmised by other means
TTE. (b) Typical color-flow Doppler artifact (arrow) asso- (e.g., outflow-graft and RVOT flow, AV opening, LV size
ciated with the HVAD inflow cannula. This artifact pro- changes during pump speed changes)

Fig. 11.14  Mild stenosis of the LVAD outflow graft-to-­ Doppler examination of the anastomotic region shows
ascending aorta anastomosis site, as assessed by TTE turbulent flow and an abnormally high peak systolic
with color and spectral Doppler. (a) 2D image: outflow velocity. (c) Continuous-wave Doppler reveals an abnor-
graft (Asterisk). The aliased color Doppler signal reveals mally high anastomotic velocity of 2.5 m/s
the site of anastomotic stenosis (arrow). (b) Pulsed

r­ ight-­sided output is calculated by using the fol- cant AR and no AV opening, the LVAD flow can
lowing commonly applied equation: RVOT car- be assumed to be significantly greater than the
diac output = RVOT pulsed Doppler RVOT-derived cardiac output, owing to a blind
VTI × [3.14 × (RVOT diameter/2)2 × HR] loop of LVAD-to-aorta flow as described above.
(Fig. 11.9), where RVOT is the right ventricular In cases of greater than mild AR, it may be useful
outflow tract, VTI is the velocity time interval, to calculate the LVAD cardiac output directly by
and HR is the heart rate. When the AV does not measuring flow within the outflow graft with
open and there is no significant AR, the RVOT- pulsed Doppler and the following equation:
derived cardiac output is the same as the LVAD LVAD output = outflow-­ conduit
cardiac output. When the AV opens significantly VTI × [3.14 × (outflow-graft diameter/2)2 × HR]
and an adequate LVOT VTI can be measured (Fig. 11.8) [9, 34, 35], although this approach has
with pulsed Doppler (and in the absence of sig- not been well validated for the HVAD. When
nificant AR), the LVAD cardiac output should using this formula, accuracy can be increased by
equal the RVOT-derived cardiac output minus measuring the outflow-graft diameter (area)
the LVOT cardiac output. In patients with signifi- directly at the site of Doppler interrogation,
178 R.F. Stainback

rather than using the manufacturer’s reported ing shows a possible intracardiac or aortic root
graft diameter (which could cause overestimation thrombus.
of flow) [9]. The aortic regurgitant volume would An experienced and knowledgeable member
then equal the LVAD stroke volume, measured of the MCS team should be immediately avail-
directly, minus the RVOT-derived stroke volume, able to solve potential problems and recognize
as described above and in Fig. 11.9. These key safety endpoints (discussed below) before
Doppler methods can be useful for validating an optimization or problem-focused echo exam
normal or abnormal LVAD flows reported by the is initiated. In the case of an optimization exam,
device’s controller (see discussion of alarms, unless the supervising MCS medical staff mem-
below) or detecting problems early, in advance of ber or an experienced echocardiography medi-
an alarm report. cal staff member is actively supervising the
exam, the ordering HF team must indicate pro-
Key Points spectively what speeds should be tested, which
• In the absence of AV opening, significant AR, or echo parameters should be measured at each
significant PR, the RVOT Doppler-derived car- speed, what defines the “optimal” LVAD speed
diac output equals the LVAD cardiac output. for that particular patient, and what the LVAD
• If the AV opens and the LVOT cardiac output speed should be at the conclusion of the study.
can be measured, the LVAD cardiac output A structured ordering template can facilitate
can be calculated as the difference between this process; a representative template is shown
the RVOT and LVOT outputs. in Table 11.2, which also outlines reasons to
• In patients with significant AR and no AV stop a speed-­change (ramp) test. These reasons
opening, it may be best to directly compute the include (1) completion of the test; (2) a suction
LVAD cardiac output by using pulsed-­wave event (at higher speeds); (3) new symptoms,
Doppler in the outflow graft. An estimate of including palpitations, dizziness, chest pain,
regurgitant volume can then be computed by shortness of breath, and headache, which may
subtracting the RVOT cardiac output. be related to hypoperfusion or hypotension; (4)
hypertension; (5) and cannula flow reversal.
Because increasing the pump speed can mark-
 chocardiography with Speed
E edly increase the mean arterial BP, the BP
Changes and Safety Concerns should be rechecked at higher pump speed set-
tings [36]. At lower pump speeds, particularly
“Speed-change testing” is part of either an opti- in patients with an elevated mean arterial pres-
mization protocol or a problem-focused (ramp) sure (hypertension), outflow-graft flow reversal
exam, both of which are outlined below. Before a can occur. The inflow-cannula color and spec-
speed-change exam is initiated, the patient’s anti- tral Doppler exam should be repeated at each
coagulant status should be considered. new pump speed to establish the following
Speed-change testing is typically performed parameters: (1) the expected progressive
only if a patient has been receiving therapeutic decrease in the peak systolic and nadir diastolic
doses of warfarin or parenteral anticoagulation flow-­velocity ratio with increasing pump speed
therapy. Risks of performing speed changes (Fig. 11.15), (2) possible flow reversal (at lower
include embolic events associated with sudden speeds as mentioned or with pump arrest [Fig.
AV opening (return to pulsatile flow) in the event 11.16]), (3) inflow-cannula flow obstruction
of undiagnosed aortic root thrombus or the poten- (Figs. 11.17 and 11.18: suction event), and (4)
tial liberation of peripheral or internal pump diminished or absent change in the flow-veloc-
thrombi, particularly at lower pump speeds. In ity profile at varying speeds in the case of inter-
general, strong consideration should be given to nal pump thrombosis or other mechanical
deferring speed-change exams if baseline imag- obstruction (Fig. 11.2) or significant AR.
11  Surveillance Echocardiography for LVAD Patients 179

Table 11.2  Sonographer checklist/ordering worksheet: During screening for inflow obstruction, both
LVAD-specific demographic data, image acquisition, and
pulsed and CW spectral Doppler interrogations
safety considerations particular to “speed-change” echo
exams (optimization, problem solving/ramp studies) of the inflow cannula are useful at baseline speed
and at each higher pump speed. When possible,
√ Study type being ordered
Doppler evaluation of the outflow graft is useful
 •  Surveillance, initial (± optimization,
preoperative/discharge) at baseline speed, but it may not be needed when
 •  Surveillance, post-discharge (± optimization, pump speed is changed (e.g., during optimization
number months post: 1, 3, 6, 12, every 6–12 or problem-focused exams, discussed below)
thereafter) unless the baseline values are abnormal or the
 •  Problem solving at baseline speed only information might be otherwise relevant for clini-
 •  Problem solving at baseline + other speed cal problem solving. The outflow-graft Doppler
settings
exam is of greater importance for HVAD patients,
 •  Recovery
because HVAD inflow-cannula velocities cannot
Ordering/responsible physician identified
be measured with Doppler.
Implantation date documented
Symptoms noted (if applicable)
Device alarms: if present, type of alarm identified
Other key clinical history/information related to LVAD Optimization
indication noted Echocardiography
Anticoagulation therapy adequate if low pump
speeds tested The LVAD optimization echo exam (with speed
LVAD name noted on worksheet and annotated on changes) is generally performed in asymptomatic
screen
or minimally symptomatic patients with no
LVAD speeds (baseline and changes) noted on
worksheet and annotated on screen
device alarms or other clinical indicators of
Blood pressure (cuff or Doppler) noted on abnormal LVAD or cardiac function. Optimization
worksheet and annotated on screen (obtained by echocardiography for the LVAD consists of rou-
designated trained individual at time of exam) tine comprehensive TTE, first at the baseline
Designated person to change pump speed available speed setting (Table 11.3) and then with stepwise
Supervision: appropriate staff to perform speed incremental adjustments to the LVAD speed (in
changes; safety endpoint recognition (e.g., low
rpm). At each new speed, prespecified echocar-
flow, suction event, hypo-/hypertension)
Endpoints for speed-change exams
diographic parameters (Tables 11.4 and 11.5) are
  • Protocol completion collected that reflect LVAD function, native LV
  • Hypotension function, or both (e.g., LVIDd, interventricular
  • Hypertension septal position, AV opening frequency/duration,
  • New symptoms
  • Device alarm
TR or MR severity) [15, 26, 36].
 •  Signs of a suction event
   –  Decrease in LV size (typically <3 cm)
   –  Interventricular septum shifting leftward HM-II Speeds
   –  Impeded flow into inlet cannula
   –  Worsening TR due to septal shifting, RV
enlargement, or both The minimum and maximum speed settings for
  – Cessation of AV opening the HM-II LVAD are 6000 and 15,000 rpm,
   –  Increased severity of AR (when present respectively. The speed can be changed in 200-­
at baseline)
 •  Signs of low cardiac output
rpm increments. Although patient dependent, the
 •  Aortic root thrombus (lowering speed could recommended range of speeds for normal pump
open AV) operation is 8800–10,000 rpm [37]. With the
 •  Cannula flow reversal (at low pump speeds) HM-II pump, speed changes for optimizing
AR aortic regurgitation, AV aortic valve, LV left ventricu- device function are usually made in small incre-
lar, LVAD left ventricular assist device, TR tricuspid
ments of 200–400 rpm.
regurgitation
180 R.F. Stainback

Fig. 11.15  LVAD outflow graft, mid right parasternal speed is systematically increased (narrowing of the pulse
window, on TTE. The flow velocity within the outflow pressure). Note: A similar diminution of the peak systolic
graft should appear laminar, with a characteristic diminu- velocity and increase in the nadir diastolic velocity occurs
tion of the peak systolic velocity (dotted line) and increase as the pump speed is systematically increased in the
in the nadir diastolic velocity (solid line) as the pump absence of any provoked inflow obstruction

Fig. 11.16  TTE, in the modified parasternal long-axis the inflow cannula shows reduced systolic forward flow
view, shows HM-II inflow-cannula diastolic flow reversal but pandiastolic regurgitant flow (aorta to left ventricle)
in a normally functioning LVAD at a relatively low pump due to improved diastolic recoil. Note: A similar Doppler
speed (8200 rpm) in a patient with significant LV recov- pattern is seen with LVAD pump arrest, although such
ery. (a) Color-flow Doppler image. The inflow cannula is arrest is associated with symptoms and echocardiographic
denoted by an arrow. (b) Pulsed spectral Doppler image of signs of HF

Fig. 11.17  (continued) the ventricular septum is bowed toward the left. (b) Aliased color-flow Doppler image of the
inflow cannula. (c) Aliased high-velocity pulsed spectral Doppler image of the inflow cannula. (d) Continuous-wave
spectral Doppler examination of the inflow cannula shows irregular flow, with systolic velocity “spikes” of up to 3.5 m/s
during tachycardia (HR = 154 bpm). (e) Right parasternal TTE view of the outflow graft shows low-­velocity laminar
flow, as evaluated by color Doppler. (f) Pulsed Doppler of the outflow graft shows an irregular pattern low-velocity flow
consistent with variable degrees of severe inflow-cannula obstruction. (g) Apical four-­chamber view shows severely
dilated right-sided chambers, a tiny LV cavity, and right-to-left bowing of the interventricular septum (arrows), with
associated severe TR on color-flow Doppler (h). The asterisk denotes a pacing lead
11  Surveillance Echocardiography for LVAD Patients 181

Fig. 11.17  Suction event at a relatively low pump speed ment, the patient received an RVAD after this exam.) (a)
(HM-II, 8200 rpm), consistent with severe RV failure. Modified parasternal RV inflow tract view. The tiny LV cavity
(Because this condition was refractory to medical manage- is “sucked down” around the inflow cannula (arrow), and
182 R.F. Stainback

Fig. 11.18  Mechanical ventricular tachycardia due to a suc- red arrows) and eliminated the ventricular tachycardia (by
tion event at a high pump speed (10,000 rpm), due to new reducing mechanical contact between the ventricular septum
hypovolemia resulting from a gastrointestinal illness. (a) and the inflow cannula). Normal low-­velocity inflow-can-
Small LV chamber size (LVEDd = 2.3 cm, red arrows), with nula flow is observed on color-flow (bb) and continuous-
frequent nonsustained ventricular tachycardia (white wave Doppler (cc) at the reduced pump speed. The LV size
arrows). (b) Turbulent, aliased inflow-cannula inflow, as remained small (3.1 cm) because of the hypovolemia, which
assessed by color Doppler. (c) Complex, “spiky,” high later resolved. Note: Mechanical ventricular tachycardia can
inflow-cannula inflow velocities up to 4 m/s on continuous-­ also be associated with excessive inflow-cannula angulation
wave Doppler examination. (aa) Reducing the pump speed toward the septum or other endocardial surfaces after sternal
(to 8600 rpm) immediately increased the LVEDd (to 3.1 cm, closure, particularly at higher pump speeds

cally performed at index hospitalization dis-

HVAD Speeds charge or 2 weeks after LVAD implantation) and
then only as needed when a routine surveillance
The minimum and maximum speed settings for the echo (without speed changes) reveals a less-than-­
HVAD are 1800 and 4000 rpm, respectively. The optimal LVAD speed according to predefined cri-
speed can be changed in 20-rpm increments. The teria [19, 26]. It is important to note that using
recommended range of speeds for normal pump echocardiography to optimize the LVAD speed is
operation is 2400–3200 rpm. With this device, speed relatively new, and the effect of echocardiography-­
changes for optimizing device function are usually guided LVAD speed optimization protocols on
made in small increments of 20 or 40 rpm [4]. short- and long-term clinical outcomes is cur-
Some LVAD implantation centers have chosen rently unknown. Table 11.6 shows summary
to include an optimization (speed-change) proto- benchmark echocardiography parameters from
col in all LVAD surveillance echo exams. Others three cohorts of patients from three different
have chosen to include the optimization protocol institutions, beginning before LVAD implanta-
in the initial surveillance echo examination (typi- tion and extending up to 12 months afterward.
11  Surveillance Echocardiography for LVAD Patients 183

Table 11.3  LVAD surveillance echo protocol: standard comprehensive TTE (or TEE) with additional LVAD-­specific
parameters
Blood pressure (if no pulse, Doppler-derived mean arterial pressure), pump type, and baseline speed
Degree of aortic valve opening/closure
Ventricular and interatrial septal position
LV inflow cannula
 • Location
 •  Note optimal view for visualization
 • Flow type
 • Flow direction
 •  Peak systolic and diastolic flow velocities (pulsed Doppler)
 •  Velocity flow pattern
LV outflow graft
 •  Location/note optimal view for visualization
 • Flow type
 • Flow direction
 •  Peak systolic and diastolic flow velocities (pulsed Doppler)
 •  Velocity flow pattern
LVAD output
 •  Outflow-graft pulsed Doppler VTI
 •  Cross-sectional area as calculated from the measured cannula diameter or from the known cannula diameter
Total cardiac output
 •  RVOT pulsed Doppler VTI
 •  Calculated cross-sectional area from the RVOT diameter
Pericardium: effusion/hematoma
Post-VAD placement red flag echo findings
 •  Ventricular and/or atrial septal shift from midline
 • Intracardiac shunt
 •  Excessive increase in cannula velocities
 •  Mechanical cannula obstruction
 •  Cannula suction event
 •  Worsening aortic or mitral regurgitation
 • Cardiac thrombus
 •  Pericardial hematoma/effusion, with or without tamponade
 •  RV dysfunction (multiple parameters in aggregate)
   –  Enlarged RV cavity size
   –  RV systolic dysfunction (qualitative, quantitative as far as possible)
  – Moderate or severe TR
  – Elevated RA pressure
LV left ventricular, LVAD left ventricular assist device, RA right atrial, RV right ventricular, RVOT right ventricular
outflow tract, TEE transesophageal echocardiography, TR tricuspid regurgitation, TTE transthoracic echocardiography,
VTI velocity time integral
Refer to Table 11.1 for guidance regarding the possible implications of abnormal/“red flag” findings

• The minimum speed is defined by echocar-

 etermining the “Optimal” LVAD
D diography parameters as the speed below which
Speed the LVIDd (measured in cm) exceeds its base-
line value. The interventricular septum may be
The definition of optimal LVAD speed varies shifted rightward, MR may become more
among implantation centers. However, there is a prominent, AV opening may occur or become
general consensus among centers that the opti- more frequent or sustained, and estimated right
mal speed lies between “minimum” and “maxi- atrial and systolic pulmonary artery pressures
mum” speeds, defined as follows: may increase. Clinically, the minimum speed is
184 R.F. Stainback

Table 11.4  LVAD optimization/ramp echo protocol

Perform baseline LVAD surveillance study (annotate BP, pump type, baseline pump speed)
At baseline pump speed, acquire the following:
 •  LVIDd in the parasternal long-axis view
 •  RV VTI (to calculate cardiac output) in the parasternal short-axis view
 •  AV opening by 2D and M-mode in the parasternal long-axis view (color Doppler M-mode if needed)
 •  2D imaging in the parasternal long- and short-axis views
 •  Color Doppler examination of AR and MR in the parasternal long-axis and apical views
 •  Color Doppler examination of TR in the RV inflow and apical four-chamber view
 •  Standard mitral valve PW Doppler inflow parameters
 •  Positioning of the interventricular and interatrial septa
Decrease pump speed to as low as 8000 rpm (for HM-II)
Or
Decrease pump speed to as low as 2400 rpm (for HVAD)
 •  Wait 2 min
 •  Repeat data acquisition
Increase pump speed by 400 rpm (for HM-II)
Or
Increase pump speed by 20–40 rpm (for HVAD)
 •  Wait 2 min
 •  Repeat data acquisition
HM-II: continue to increase pump speed in 400-rpm increments to a pump speed of up to 12,000 rpm or until
endpoint (below), acquiring data at each stage
HVAD: continue to increase pump speed in 20- to 40-rpm increments to a pump speed of up to 3200 rpm, or until
endpoint (below), acquiring data at each stage
Endpoints
 •  Completion of test
 •  Suction event: decrease in LV size (typically <3 cm), ± ventricular ectopy, ± inflow-cannula intermittent
obstruction, leftward ventricular septal shift, worsening TR
 •  Symptoms including, but not limited to, palpitations, dizziness, chest pain, shortness of breath, or headache
 •  Hypertension (e.g., MAP > 100 mmHg or symptoms)
 •  Hypotension (e.g., MAP < 60 mmHg or symptoms)
Note: inflow-cannula color and spectral Doppler (including CW Doppler) should be evaluated at each pump speed to
test for obstruction. Outflow-graft Doppler evaluation is needed at baseline but is optional at speed changes if LVAD
function is normal. When abnormal conditions are being evaluated, additional parameters may be assessed when pos-
sible, such as outflow-graft velocity profile/stroke volume (e.g., for obstruction, to assess AR volume) and outflow-
graft-to-aortic anastomosis to assess obstruction or flow reversal
2D two dimensional, AR aortic regurgitation, AV aortic valve, BP blood pressure, HM-II HeartMate II, HVAD HeartWare
ventricular assist system, LV left ventricular, LVAD left ventricular assist device, LVIDd left ventricular internal diam-
eter at end-diastole, MAP mean arterial pressure, MR mitral regurgitation, PW pulsed Doppler, RV right ventricular, TR
tricuspid regurgitation, TV tricuspid valve, VTI velocity time integral

that speed below which the patient’s functional enlargement, the AV may cease opening, and
capacity is reduced, congestion develops, or AR (when present) increases. Some or all of
end-organ function worsens. these changes above the maximal speed may
• The maximum speed is defined echocardio- constitute a “suction event,” with low-flow
graphically as the speed above which the alarms (see below).
interventricular septum shifts leftward or
impedes flow into the inflow cannula. To provide a margin of safety, implantation
Tricuspid regurgitation may worsen owing to centers that view maximal LV unloading as para-
the leftward interventricular septal shift with mount in HF management define the optimal
tricuspid valve annular distortion or RV LVAD speed as being just below the maximum
Table 11.5  Speed changes: LVAD optimization or problem-focused (ramp) protocol worksheet

CF-LVAD type: Implantation date: [PT INR = _____ PTT = _____]

Previous echo exam date and significant findings:

• Optimization protocol. Optimal speed based on MCS center’s own standard; sample order sets include the following: (a) attain at least
intermittent AV opening, (b) attain neutral IVS position and/or mild or less MR, (c) attain complete AV closure to maximize LV
unloading, or (d) adjust speed to below the maximum speed associated with complete AV closure and the minimum speed associated with
more prominent MR and rightward IVS
• Problem-focused protocol. Indication for exam: sample order sets include the following indications:
(a) Smoldering left- and/or right-sided heart failure
(b) Screen for pump function in patients with hemolysis and suspected pump thrombosis
(c) Other LVAD alarm troubleshooting

AV MV peak E (a) Symptoms (y/n)

Pump opening RVOT TR velocity IVS (b) Evidence of
speed (y/n/intermit LVIDd VTI Significant Significant Significant velocity (m/s), DT direction inflow- cannula
(rpm) BP tent) (cm) (cm) AR (y/n) MR (y/n) TR (y/n) (m/s) (ms) L/R/neutral obstruction (y/n)

Reason for termination: (E.g., signs of inflow-cannula obstruction, hypotension, hypertension, worsening RV or LV function)

Final speed setting = ____ rpm

Final BP = ______ mmHg

Note: parameters measured at each speed setting may vary according to an implantation center’s internal standards.
After examination at the baseline pump speed, most of the needed parameters at subsequent pump speeds can be
obtained primarily from parasternal views in most cases, as a limited exam.
AR aortic regurgitation, AV aortic valve, BP blood pressure, CF continuous flow, DT deceleration time, E early diastole,
INR international normalized ratio, IVS interventricular septum, LV left ventricular, LVAD left ventricular assist device,
LVIDd left ventricular internal diameter at end-diastole, MCS mechanical circulatory support, MR mitral regurgitation,
MV mitral valve, PT prothrombin time, PTT partial thromboplastin time, RVOT right ventricular outflow tract, TR tri-
cuspid regurgitation, VTI velocity time integral

Table 11.6  Magnitude and time course of echo LV parameter changes induced by CF-LVAD unloading
Post-LVAD Post-LVAD Post-LVAD Post-LVAD
Pre-LVAD 1 month 3 months 6 months 12 months
Study 1 (N = 21) Study 1 (N = 21) – Study 1 (N = 10) –
Study 2 (N = 63) – Study 2 (N = 63) Study 2 (N = 63) –
Variable Study 3 (N = 80) Study 3 (N = 68) Study 3 (N = 47) Study 3 (N = 32) Study 3 (N = 20)
LV parameters
 LV diastolic diameter
   Study 1 (mm) 66 ± 11 55 ± 11** – 52 ± 11* –
   Study 2 (mm) 68 ± 9 – 56 ± 11* 57 ± 12 –
   Study 3 (cm/m2) 3.2 (2.9, 3.6) 2.8 (2.3, 3.2) 2.9 (2.4, 3.4) 2.8 (2.2, 3.4) 2.6 (2.2, 3.0)*
 LV systolic diameter
   Study 1 (mm) 58 ± 10 47 ± 12 – 43 ± 13 –
   Study 2 (mm) 61 ± 9 – 47 ± 13* 49 ± 13 –
   Study 3 (cm/m2) 3.0 (2.6, 3.3) 2.6 (2.0, 3.1) 2.6 (2.1, 3.1) 2.5 (1.8, 2.9) 2.3 (1.9, 2.8)*
 LV end-diastolic volume
   Study 1 (mL) 242 ± 108 127 ± 68* – 113 ± 45* –
  Study 2 – – – – –
(continued)
186 R.F. Stainback

Table 11.6 (continued)
Post-LVAD Post-LVAD Post-LVAD Post-LVAD
Pre-LVAD 1 month 3 months 6 months 12 months
Study 1 (N = 21) Study 1 (N = 21) – Study 1 (N = 10) –
Study 2 (N = 63) – Study 2 (N = 63) Study 2 (N = 63) –
Variable Study 3 (N = 80) Study 3 (N = 68) Study 3 (N = 47) Study 3 (N = 32) Study 3 (N = 20)
   Study 3 (mL/m2) 113 (94, 141) 77 (54, 109)* 86 (62, 106)* 86 (52, 108)* 69 (45, 93)*
 LV end-systolic volume
   Study 1 (mL) 191 ± 93 100 ± 66* – 82 ± 42* –
  Study 2 – – – – –
   Study 3 (mL/m2) 3.0 (2.6, 3.3) 2.6 (2.0, 3.1)* 2.6 (2.1, 3.1) 2.5 (1.8, 2.9)* 2.3 (1.9, 2.8)*
 LV ejection fraction (%)
  Study 1 22 ± 5 25 ± 13 – 29 ± 10 –
  Study 2 19 ± 7 – 26 ± 12* 27 ± 14 –
  Study 3 17 (14, 23) 20 (15, 30) 20 (14, 26) 25 (18, 33)* 22 (15, 31)
 LV mass
  Study 1 – – – – –
   Study 2 (g) 383 ± 113 – 295.9 ± 188* 314 ± 134 –
   Study 3 (g/m2) 114 (93, 146) 95 (71, 114)** 92 (63, 118)** 111 (74, 134) 77 (50, 104)*
LV diastolic parameters
 LA size
   Study 1 (mm) 47 ± 7 37 ± 9** – 42 ± 13 –
   Study 2 (mL/m2) 69 ± 30 – 42 ± 15* – –
   Study 3 (mL/m2) 46 (35, 54) 28 (22, 36)* 32 (23, 38)* 25 (19, 39)* 28 (18, 38)*
 E-wave
   Study 1 (cm/s) 96 ± 23 73 ± 27** – 66 ± 12** –
   Study 2 (cm/s) 98 ± 35 – 100 ± 160 80 ± 20 –
   Study 3 (cm/s) 100 (80, 110) 80 (60, 100)* 80 (70, 100) 80 (70, 110) 100 (60, 120)
 E/A ratio
  Study 3 2.8 (2.1, 4.1) 2.2 (1.2, 3.6) 1.5 (1.0, 2.9)* 1.6 (1.3, 2.2)** 1.7 (1.0, 3.3)
 Mitral DT
  Study 1 124 ± 39 180 ± 53** – 164 ± 24 –
  Study 2 132 ± 27 – 188 + 70* 166 ± 48 –
  Study 3 133 (112, 165) 175 (137, 220)* 178 (141, 212)* 172 (121,220)* 170 (157, 225)
 Tissue Doppler e′ (cm/s)
  Study 1 – – – – –
   Study 2 (septal e′) 4 ± 1 – 4 ± 1 – –
   Study 3 (septal e′) 4 (3, 6) 6 (5, 9)* 7 (5, 9)* 7 (4, 9)* 7 (6, 10)**
   (lateral e′) 8 (5, 11) 9 (7, 10) 9 (6, 11) 10 (7, 13) 12 (8, 12)
 E/e′ (ratio)
  Study 1 – – – – –
   Study 2 (septal e′) 26 ± 11 – 20 ± 9** 13 ± 7 –
   Study 3 (septal e′) 23 (16, 30) 13 (9, 19)* 12 (9, 16)* 12 (9, 19)* 15 (7, 17)**
   (lateral e′) 14 (9, 19) 9 (16, 13**) 10 (6, 12) 9 (7, 13) 10 (6, 11)
Study 1 Lam et al. JASE 2009 [8], Study 2 Topilsky et al. JASE 2011 [9], Study 3 Drakos et al. JACC 2013;61:1985–94.
Values are mean ± SD for studies 1 and 2 and median (25th, 75th percentiles) for Study 3.
Study 2 P values only provided comparing pre-LVAD and post-LVAD 3-months measurements. A mitral valve late peak
diastolic velocity, CF continuous flow, DT deceleration time, E mitral valve early peak diastolic velocity, e′ mitral
annular velocity, LA left atrial, LV left ventricle, LVAD left ventricular assist device
*
P < 0.01 versus pre-LVAD; **P < 0.05 versus pre-LVAD
11  Surveillance Echocardiography for LVAD Patients 187

speed, even when the AV remains closed (typi- can be streamlined considerably when not all
cally at least 400 rpm below the maximum speed data columns are required according to a center’s
for the HM-II [37] and at least 40 rpm below the own protocol. On the other hand, data collection
maximum speed for the HVAD). Implantation may need to be more comprehensive at each
centers that desire AV opening will, when possi- pump speed if certain problems are suspected on
ble, choose a lower “optimal” LVAD speed, at the basis of baseline speed findings.
which AV opening occurs either intermittently or In summary, the components and timing of an
during every cardiac cycle, combined with other LVAD surveillance examination have only been
echocardiographic data to suggest clinically ade- recently defined by consensus in the recent rec-
quate (if not maximal) LV unloading. A subset of ommendations from the American Society of
these centers may elect to maximize the AV open- Echocardiography [1]. There remains little out-
ing duration. As noted above, Table 11.5 provides comes data with regard to optimal speed settings.
a typical set of parameters that can be measured Ideally, a program of routine surveillance echo-
at each speed during an LVAD optimization cardiography examinations could confirm normal
exam, including LVIDd, interventricular septal device function or detect occult device or native
position, AV opening frequency/duration, MR heart abnormalities that could be addressed early
severity, TR severity and velocity, and cannula on to prevent hospitalization for recurrent heart
flow velocities. failure or to identify patients who may require
Implementing safe and consistent LVAD sur- more frequent monitoring. Implementing consis-
veillance echocardiography in a busy clinical tent and safe surveillance exam protocols can be
environment with changing medical and techni- a challenge, and it is hoped that the material and
cal staff can be challenging. Accordingly, the fol- implementation protocols herein might be a help-
lowing implementation tools have been included. ful starting point for creating a more efficient
Table 11.3 is an “LVAD surveillance echo proto- LVAD patient care plan.
col” that provides a checklist for exam setup,
LVAD-specific parameters, and “red flag find-
ings.” Laboratory leaders may edit this protocol References
into their existing standard heart failure examina-
tion protocol for a “baseline speed surveillance 1. Stainback RF, Estep JD, Agler DA, Birks EJ, Bremer
M, Hung J, et al. Echocardiography in the manage-
echo protocol.” Table 11.4 can be used when an ment of patients with left ventricular assist devices:
LVAD optimization/ramp protocol is needed. recommendations from the American Society
Table  11.5 is a “speed-change worksheet” that of Echocardiography. J Am Soc Echocardiogr.
can be used to help organize data for interpreta- 2015;28:853–909.
2. Bennett MK, Roberts CA, Dordunoo D, Shah A,
tion and reporting purposes. The LVAD data are Russell SD. Ideal methodology to assess systemic
typically first obtained at the baseline pump blood pressure in patients with continuous-flow left
speed and then at subsequent speeds determined ventricular assist devices. J Heart Lung Transplant.
by the supervising physician. The data columns 2010;29:593–4.
3. Lanier GM, Orlanes K, Hayashi Y, Murphy J,
can be edited to conform to a lab’s internal stan- Flannery M, Te-Frey R, et al. Validity and reliabil-
dards. Typically, an LVAD examination at the ity of a novel slow cuff-deflation system for nonin-
baseline pump speed requires a similar time vasive blood pressure monitoring in patients with
allotment as a standard non-LVAD examination continuous-­flow left ventricular assist device. Circ
Heart Fail. 2013;6:1005–12.
once sonographers become familiar with these 4. HeartWare ventricular assist system: instructions for
devices and given some of the technical limita- use. Miami Lakes, FL: HeartWare; 2012.
tions on obtaining standard images as outlined 5. Markham DW, Drazner MH. Measuring nonpulsatile
above. Data acquired at subsequent pump speeds blood pressure: say goodbye to the Doppler? Circ
Heart Fail. 2013;6:879–80.
188 R.F. Stainback

6. Lang RM, Badano LP, Mor-Avi V, Afilalo J, tion by echocardiography. J Am Soc Echocardiogr.
Armstrong A, Ernande L, et al. Recommendations 2009;22:107–33.
for cardiac chamber quantification by echocardiogra- 17.
Rudski LG, Lai WW, Afilalo J, Hua L,
phy in adults: an update from the American Society Handschumacher MD, Chandrasekaran K, et al.
of Echocardiography and the European Association Guidelines for the echocardiographic assessment of
of Cardiovascular Imaging. J Am Soc Echocardiogr. the right heart in adults: a report from the American
2015;28:1–39.e14, e14. Society of Echocardiography endorsed by the
7. Porter TR, Abdelmoneim S, Belcik JT, McCulloch European Association of Echocardiography, a regis-
ML, Mulvagh SL, Olson JJ, et al. Guidelines for tered branch of the European Society of Cardiology,
the cardiac sonographer in the performance of con- and the Canadian Society of Echocardiography.
trast echocardiography: a focused update from the J Am Soc Echocardiogr. 2010;23:685–713; quiz
American Society of Echocardiography. J Am Soc 786–788.
Echocardiogr. 2014;27:797–810. 18. Lee S, Kamdar F, Madlon-Kay R, Boyle A, Colvin-­
8. Lam KM, Ennis S, O'Driscoll G, Solis JM, Adams M, Pritzker M, et al. Effects of the HeartMate
Macgillivray T, Picard MH. Observations from non-­ II continuous-flow left ventricular assist device on
invasive measures of right heart hemodynamics right ventricular function. J Heart Lung Transplant.
in left ventricular assist device patients. J Am Soc 2010;29:209–15.
Echocardiogr. 2009;22:1055–62. 19. Topilsky Y, Hasin T, Oh JK, Borgeson DD, Boilson
9. Topilsky Y, Oh JK, Atchison FW, Shah DK, Bichara BA, Schirger JA, et al. Echocardiographic variables
VM, Schirger JA, et al. Echocardiographic find- after left ventricular assist device implantation associ-
ings in stable outpatients with properly functioning ated with adverse outcome. Circ Cardiovasc Imaging.
HeartMate II left ventricular assist devices. J Am Soc 2011;4:648–61.
Echocardiogr. 2011;24:157–69. 20. Raina A, Vaidya A, Gertz ZM, Susan C, Forfia

10. Garcia-Alvarez A, Fernandez-Friera L, Lau JF, Sawit PR. Marked changes in right ventricular contractile
ST, Mirelis JG, Castillo JG, et al. Evaluation of right pattern after cardiothoracic surgery: implications for
ventricular function and post-operative findings using post-surgical assessment of right ventricular function.
cardiac computed tomography in patients with left J Heart Lung Transplant. 2013;32:777–83.
ventricular assist devices. J Heart Lung Transplant. 21. Feldman D, Pamboukian SV, Teuteberg JJ, Birks E,
2011;30:896–903. Lietz K, Moore SA, et al. The 2013 International
11. Sauer AJ, Meehan K, Gordon R, Abicht T, Rich Society for Heart and Lung Transplantation guide-
JD, Anderson AS, et al. Echocardiographic mark- lines for mechanical circulatory support: executive
ers of left ventricular unloading using a centrif- summary. J Heart Lung Transplant. 2013;32:157–87.
ugal-flow rotary pump. J Heart Lung Transplant. 22. Slaughter MS, Pagani FD, Rogers JG, Miller LW,
2014;33:449–50. Sun B, Russell SD, et al. Clinical management of
12. Schiller NB, Shah PM, Crawford M, DeMaria A,
continuous-flow left ventricular assist devices in
Devereux R, Feigenbaum H, et al. Recommendations advanced heart failure. J Heart Lung Transplant.
for quantitation of the left ventricle by two-­ 2010;29:S1–S39.
dimensional echocardiography. American Society 23. Mudd JO, Cuda JD, Halushka M, Soderlund KA,
of Echocardiography Committee on Standards, Conte JV, Russell SD. Fusion of aortic valve commis-
Subcommittee on Quantitation of Two-Dimensional sures in patients supported by a continuous axial flow
Echocardiograms. J Am Soc Echocardiogr. 1989;2: left ventricular assist device. J Heart Lung Transplant.
358–67. 2008;27:1269–74.
13. Quinones MA, Waggoner AD, Reduto LA, Nelson 24. Cowger J, Pagani FD, Haft JW, Romano MA,

JG, Young JB, Winters WL Jr, et al. A new, simplified Aaronson KD, Kolias TJ. The development of
and accurate method for determining ejection fraction aortic insufficiency in left ventricular assist
with two-dimensional echocardiography. Circulation. device-­supported patients. Circ Heart Fail. 2010;3:
1981;64:744–53. 668–74.
14. Mancini DM, Beniaminovitz A, Levin H, Catanese 25. Pak SW, Uriel N, Takayama H, Cappleman S, Song
K, Flannery M, DiTullio M, et al. Low incidence of R, Colombo PC, et al. Prevalence of de novo aor-
myocardial recovery after left ventricular assist device tic insufficiency during long-term support with left
implantation in patients with chronic heart failure. ventricular assist devices. J Heart Lung Transplant.
Circulation. 1998;98:2383–9. 2010;29:1172–6.
15. Estep JD, Stainback RF, Little SH, Torre G, Zoghbi 26. Estep JD, Chang SM, Bhimaraj A, Torre-Amione
WA. The role of echocardiography and other imag- G, Zoghbi WA, Nagueh SF. Imaging for ven-
ing modalities in patients with left ventricular assist tricular function and myocardial recovery on non-
devices. JACC Cardiovasc Imaging. 2010;3:1049–64. pulsatile ventricular assist devices. Circulation.
16. Nagueh SF, Appleton CP, Gillebert TC, Marino
2012;125:2265–77.
PN, Oh JK, Smiseth OA, et al. Recommendations 27. Hatano M, Kinugawa K, Shiga T, Kato N, Endo M,
for the evaluation of left ventricular diastolic func- Hisagi M, et al. Less frequent opening of the a­ ortic
11  Surveillance Echocardiography for LVAD Patients 189

valve and a continuous flow pump are risk factors tion with two-dimensional and Doppler echocardiog-
for postoperative onset of aortic insufficiency in raphy. J Am Soc Echocardiogr. 2003;16:777–802.
patients with a left ventricular assist device. Circ 33. Ammar KA, Umland MM, Kramer C, Sulemanjee
J. 2011;75:1147–55. N, Jan MF, Khandheria BK, et al. The ABCs of left
28. Toda K, Fujita T, Domae K, Shimahara Y, Kobayashi ventricular assist device echocardiography: a sys-
J, Nakatani T. Late aortic insufficiency related to poor tematic approach. Eur Heart J Cardiovasc Imaging.
prognosis during left ventricular assist device support. 2012;13:885–99.
Ann Thorac Surg. 2011;92:929–34. 34. Horton SC, Khodaverdian R, Chatelain P, McIntosh
29. Aggarwal A, Raghuvir R, Eryazici P, Macaluso G, ML, Horne BD, Muhlestein JB, et al. Left ven-
Sharma P, Blair C, et al. The development of aor- tricular assist device malfunction: an approach to
tic insufficiency in continuous-flow left ventricular diagnosis by echocardiography. J Am Coll Cardiol.
assist device-supported patients. Ann Thorac Surg. 2005;45:1435–40.
2013;95:493–8. 35. Schwarz KQ, Parikh SS, Chen X, Farrar DJ, Steinmetz
30. Jorde UP, Uriel N, Nahumi N, Bejar D, Gonzalez-­ S, Ramamurthi S, et al. Non-invasive flow measure-
Costello J, Thomas SS, et al. Prevalence, significance, ment of a rotary pump ventricular assist device using
and management of aortic insufficiency in ­continuous quantitative contrast echocardiography. J Am Soc
flow left ventricular assist device recipients. Circ Echocardiogr. 2010;23:324–9.
Heart Fail. 2014;7:310–9. 36. Uriel N, Morrison KA, Garan AR, Kato TS,

31. Baumgartner H, Hung J, Bermejo J, Chambers JB, Yuzefpolskaya M, Latif F, et al. Development of a
Evangelista A, Griffin BP, et al. Echocardiographic novel echocardiography ramp test for speed optimiza-
assessment of valve stenosis: EAE/ASE recommen- tion and diagnosis of device thrombosis in continuous-­
dations for clinical practice. J Am Soc Echocardiogr. flow left ventricular assist devices: the Columbia
2009;22:1–23; quiz 101–2. ramp study. J Am Coll Cardiol. 2012;60:1764–75.
32. Zoghbi WA, Enriquez-Sarano M, Foster E, Grayburn 37. HeartMate II LVAS: clinical operation & patient

PA, Kraft CD, Levine RA, et al. Recommendations for management. Pleasanton, CA: Thoratec Corporation;
evaluation of the severity of native valvular regurgita- 2011.
 Diagnosis of Device Thrombosis
12
Cyril Varughese, Ajith P. Nair,
and Jordan Chaisson

ters utilized still remains the international nor-

Introduction malized ratio (INR) with typical goals ranging
from 2 to 3 or even 2.5–3.5 with patients who
As continuous-flow left ventricular devices might be at higher risk from pump thrombosis.
become more streamlined and smaller, pump Current guidelines from ISHLT recommend an
thrombosis and hemolysis are emerging as seri- INR of 2.0–3.0 with HeartMate II and HeartWare
ous complications following device implanta- devices. However, despite aggressive anticoagu-
tion. Initial studies involving continuous-flow lant therapy, a patient might still be prone to
left ventricular assist devices (LVAD) revealed an developing device thrombosis, which with early
incidence of pump thrombosis following initial detection can play a key role with overall
implantation at 0.014–0.03 events per patient-­ survival.
year [1]. Additional studies reviewing data from
around 2005 to 2007 reported 4% of patients
with pump thrombosis [2]. More recent analysis Etiology
of INTERMACS data has shed further light in
revealing possible risk factors, such as age, gen- Although no single entity has been known to
der, and BMI, which might make a patient more produce left ventricular assist device thrombo-
prone to developing pump thrombosis [3]. One sis, several factors together may play a role for
challenge that clinicians face following LVAD thrombus formation in select patients. First
implantation remains balancing anticoagulation assessment should focus on identifying those
therapy with risk of bleeding. Anticoagulation patients that might be predisposed to thrombus
and antiplatelet therapy varies from institution to formation. Usually risk assessment for pump
institution, but the primary goal remains balanc- thrombosis is performed prior to device
ing a patient risk of bleeding with their risk for implantation based on various institutions’
pump thrombosis. The primary marker most cen- guidelines. Those individuals already at high
risk ­(hypercoagulable state, autoimmune dis-
ease, malignancy) are not a direct contraindi-
C. Varughese, D.O. • A.P. Nair, M.D. (*) cation for mechanical circulatory support but
J. Chaisson, M.D. do warrant a thorough investigation prior to
Texas Heart Institute at Baylor St. Luke’s Medical
device implantation. A patient’s body weight
Center, Baylor College of Medicine, Houston, TX, USA
e-mail: cyrilvarughese@[Link]; [Link]@[Link]; can also play a role for potential pump throm-
jchaiss@[Link] bosis. Just as prior studies have identified poor

© Springer International Publishing AG 2018 191

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
192 C. Varughese et al.

outcomes in low BMI patient population dence of pump thrombosis due to development of
(<18.5 kg/m2), those patients with obesity or a thin biofilm layer decreasing thrombin deposi-
even morbid obesity have been shown to be at tion. Unfortunately similar textured surfaces on
higher risk for pump thrombosis [3, 4]. Prior the inflow cannula of new left ventricular assist
retrospective studies reviewed patients with a devices have not been able to replicate the prior
BMI greater than 30 with increased risk for effect to a degree.
pump thrombosis [3]. Analysis of INTERMACS
data identified female gender as having early
phase incidence of pump thrombosis; however, Locating Pump Thrombosis
this may have been secondary to smaller LV
cavity size [3]. Even after device placement, Identifying the region of pump thrombosis can also
vigilance must be maintained to evaluate addi- be challenging. Usually mechanical continuous-­
tional factors that may predispose a patient to flow pumps can be divided into three areas of inter-
thrombus formation. Following LVAD implan- est for thrombosis: the inflow, pump, and outflow
tation, driveline infection and/or pump pocket graft. Laboratory testing can be useful to aid with
infections tend to promote prothrombotic diagnosing pump thrombosis, but 2D echocardiog-
states that can lead to an increased risk for raphy and CT angiography are also helpful to iden-
pump thrombosis [5]. tify the location of pump thrombosis.
In addition to assessing a patient’s inherent Positioning of the inflow cannula usually can
risk for pump thrombosis, external factors can be assessed with 2D echocardiography. Typically
play a role in predisposing to thrombus forma- angulation of the inflow cannula medially toward
tion. One key component to minimize pump the interventricular septum or next to papillary
thrombosis remains tight anticoagulation man- muscles can predispose to obstruction near the
agement. In one review of various LVAD implan- inlet. Depending on specific hardware type,
tation centers, an increase incidence of pump direct visualization of inflow cannula can be lim-
thrombosis in one part was thought to be second- ited by acoustic dropout from echocardiography.
ary to lower anticoagulation goals [6]. Continuous wave Doppler directed toward the
Anticoagulation therapy can become difficult if inflow cannula with velocities >1.5 m/s has been
the patient’s risk of bleeding increases as in the suggestive of inflow obstruction or even suction
setting of an acute gastrointestinal bleed. Usually event [9]. Identifying thrombosis in the pump
in an effort to reduce the incidence of GI bleed- itself can be difficult due to limitations with
ing, pump speeds are reduce to maintain some imaging. Severe streaking artifact on CTA of the
semblance of pulsatility. An indirect result of this chest from the metallic motor housing makes
can be less blood moving through the pump, direct visualization of pump thrombosis difficult.
which prevents adequate heat dissipation from Usually evaluation of the motor housing itself is
the motor itself leading to an increased risk of accomplished by direct visualization in the oper-
thrombus formation. Simulated low axial flow in ating room or disassembly of the device at the
HeartMate II pumps demonstrated prothrombotic manufacturing site.
states once flows were decreased below 3.8 lpm In a similar fashion of evaluating the inflow
which supported prior studies suggesting lower graft, the outflow graft of a continuous-flow
flows promoting increase risk for pump thrombo- LVAD can be visualized through CTA with 3D
sis [7, 8]. In addition to maintaining adequate reconstruction (Fig. 12.1). This can aid the clini-
anticoagulation goals and speed adjustments, cian to determine if any kinks are present
evaluating the device itself becomes crucial to obstructing flow through the outflow graft. Two-­
determine if specific positioning of the inflow or dimensional echocardiography can also assist in
outflow graft may predispose to pump thrombo- evaluating LVAD outflow graft velocities.
sis. Initially textured surfaces of the HeartMate Typically Doppler flow tends to accelerate if
XVE ventricular assist device minimized inci- obstruction is present to levels >2 m/s [9].
12  Diagnosis of Device Thrombosis 193

Fig. 12.1  CT angiogram with 3D reconstruction of left LVAD, outflow graft, and ascending aorta. (c) Orthogonal
ventricular assist device. (a) Following contrast adminis- visualization of outflow graft. (d) Further visualization of
tration, filling defect can be seen along the distal portion filling defect in distal outflow graft consisted with
of outflow graft. (b) Complete 3D reconstruction of thrombosis

study, hematological markers were a better

Diagnosing Pump Thrombosis marker to predict pump thrombosis before imag-
ing or pump parameters were identified [10].
Evaluating a patient for acute thrombosis usually Most advanced heart failure centers routinely
starts with a clinical evaluation but ultimately screen patients with left ventricular assist devices
requires a culmination of laboratory testing in by checking lactate dehydrogenase (LDH) levels
addition to various imaging modalities [11]. Long-term monitoring of a patient’s LDH
(Fig. 12.2). One of the primary goals for labora- level is key to evaluate trends that may suggest a
tory testing remains validating the presence of gradual pump thrombosis versus an acute event.
hemolysis. Clinical signs may be present that LDH levels tend to remain elevated but stable in
coincide with lab results suggestive of hemolysis. the former whereas levels can rise abruptly in the
Patients may describe darkened urine from latter. Typically once levels are greater than 2–2.5
increased hemoglobin destruction due to active times, partial or complete pump thrombus forma-
hemolysis from pump thrombosis. When pump tion should be excluded. Also plasma-free hemo-
thrombosis forms, reduced forward flow devel- globin levels are another laboratory test to
ops, preventing adequate unloading of the left identify active hemolysis. Typically two separate
ventricle. Soon the patient may experience simi- levels that are greater than 40 mg/dL are sugges-
lar symptoms of decompensated heart failure tive of ongoing hemolysis. Combining an ele-
they felt prior to device implantation. Once these vated LDH and plasma-free hemoglobin reading
clinical factors are present, active evaluation of with other common markers for hemolysis such
pump thrombosis must be performed starting as reduced levels of hemoglobin, hematocrit, and
with laboratory analysis. In one retrospective haptoglobin and increased indirect bilirubin lev-
194 C. Varughese et al.

Suspect LVAD
Thrombosis

Hemodynamically
Yes Stable

No
LVAD CHF
alarms Symptoms

Admit to ICU
Evidence of hemolysis:
± Inotropic Support
↑ LDH
± Temporary mechanical
↑ pfHB circulatory support
Urgent surgical evaluation

Above Yes No
Ramp Study: LV Imaging
baseline Unloading (CTA)
readings
No Adjust CHF meds
Yes Adjust anticoagulation
No
Adjust appropriate meds Inflow/Outflow
↑INR Goal obstruction
Complete Yes Resolved? Yes
diagnostic
with log
Reassess for Surgical No
file eval Dual antiplatelet
hemolysis/CHF Evaluation
Higher INR goals

Fig. 12.2  Flowchart for diagnosing and managing acute patient’s baseline. Evidence for hemolysis would be ele-
pump thrombosis. Abnormal LVAD alarms include power vated LDH levels 2–2.5 times the upper limit of normal
readings >10 W or readings greater than 2 W above and plasma-free hemoglobin (pfHb) > 40 mg/dL

els is strongly suggestive of ongoing hemolysis thrombus formation or kink present. Most
from pump thrombosis. patients with acute pump thrombosis tend to have
In addition to laboratory testing for hemolysis, elevated creatinine levels due to ongoing hemoly-
various imaging studies can be utilized to evalu- sis, which can make ordering a CTA difficult due
ate LVAD positioning and ability to unload the to the need for intravenous contrast. However, the
left ventricle. The first imaging study to perform risk of worsening renal perfusion from a low car-
on a patient with a suspected pump thrombosis diac output state is also an important factor to
should be a chest x-ray. The chest x-ray can be consider when working up a patient with an acute
useful to not only evaluate pulmonary congestion pump thrombosis.
from decompensated heart failure but also to Two-dimensional echocardiography can be a
evaluate any kinking or malposition of the inflow useful tool when evaluating a patient with
cannula. In addition to the standard chest x-ray, suspected pump thrombosis. When pump
­
CTA of the chest becomes useful in evaluating thrombosis occurs, adequate unloading of the
the inflow and outflow graft. As mentioned above, left ventricle becomes difficult. As a result, left
positioning of the inflow cannula can be visual- ventricular end-diastolic dimensions can
ized with CT angiography as well as LV cavity increase which is one sign of pump malfunction
size and possible location of pump thrombosis. if the left ventricular size is larger than previous
The outflow graft can also be adequately visual- baseline studies at similar speeds. Mitral regur-
ized to determine if any specific segment has gitation is also a useful parameter seen on 2D
12  Diagnosis of Device Thrombosis 195

echocardiogram which can help in determining a normal functioning pump, at higher speeds,
if adequate LV unloading is occurring at certain cardiac output increases, PCWP decreases, and
pump speeds. While there may be some mild right-sided filling pressure may increase due to
mitral regurgitation at baseline speeds in a nor- higher preload.
mal functioning LVAD, theoretically increasing Left ventricular device power surges or
the LVAD speed should unload the LV suffi- changes in pulsatility index can also serve to
ciently to reduce or minimize the presence of assist in diagnosing pump thrombosis. Typically
mitral regurgitation. Once pump thrombosis pump thrombosis usually causes high power
occurs, adequate LV unloading is not possible at spikes with low a low pulsatility index. Caution
higher speeds, so severity of mitral regurgitation should be used in avoiding making clinical deci-
may persist despite higher VAD settings. The sions based on single pump parameter changes
aortic valve can be another structure seen on 2D but evaluate changes in LVAD settings compared
echocardiography that can assist in determining to a patient’s baseline. Typically pump thrombo-
if adequate LV unloading is occurring at higher sis causes elevated power readings >10 W or sus-
speeds. At higher speeds, normal functioning tained readings greater than 2 W above a patient’s
left ventricular assist devices should unload the baseline [13]. Recent studies regarding pump
left ventricle to the point that a majority of blood thrombosis recommend identifying multiple fac-
flow is directed toward the LVAD instead of the tors such as active hemolysis, symptoms of
left ventricular outflow tract. As a result, mini- decompensated heart failure, abnormal imaging
mal aortic valve opening is seen at higher LVAD or ramp studies, and/or abnormal pump readings
speeds suggestive of adequate pump function. to help in making the diagnosis.
Caution should be used when performing a
ramp study on a patient with suspected pump
thrombosis. Depending on the location of Management and  Treatment
thrombus, distal embolization can occur once
speed adjustments are made. Usually HeartMate Sudden thrombus formation leading to pump
II LVADs are increased by 200 rpm, while stoppage can be life-threatening depending on
HeartWare HVAD speeds are increased by how much native cardiac function still exists. For
20 rpm increments. Initial studies reviewing the those patients completely dependent on their left
use of ramp echocardiography evaluation were ventricular assist device to maintain cardiac out-
helpful with HeartMate II LVEDD slope rela- put (i.e., oversewn aortic valve), acute thrombus
tionships but did not correlate well when extrap- formation leads to profound hemodynamic insta-
olated to HVAD function [12]. bility and possible death. Early detection and
In some instances when adequate information management become critical to stabilize the
is not obtained from CTA or 2D echocardiogra- patient and identify further therapeutic options.
phy, cardiac output and hemodynamics assessed The first step in managing a patient with pump
from right heart catheterization can assist in thrombosis is assessing the patient for hemody-
determining if adequate LV unloading is occur- namic stability (Fig. 12.2). As mentioned above,
ring with various pump settings. One advantage those patients fully dependent on cardiac output
to right heart catheterization versus 2D echocar- from their assist devices can clinically deteriorate
diography or CTA is the ability to measure both rapidly. These patients usually require inotropic
left- and right-sided pressures. Specific protocols support and in some cases temporary mechanical
for ramp right heart catheterization studies may circulatory support. Those patients who are in
vary from institution to institution; however, sim- shock and are not responsive to medical therapy
ilar to ramp echocardiogram study, various mea- require emergent surgical evaluation for pump
surements such as cardiac outputs, pulmonary exchange and should not be delayed to minimize
capillary wedge pressures, and right-sided filling end organ damage. In addition to inotropic sup-
pressure can be obtained at various speeds. With port, attention must be focused on anticoagulation
196 C. Varughese et al.

therapy. The first line of anticoagulation therapy subsequent sternotomies. However, a recent study
remains IV heparin infusion; if contraindications comparing postoperative complications in
are present or thrombus does not resolve, then con- patients with resternotomy versus primary ster-
sideration must be given for direct thrombin inhib- notomy found similar results between the two
itors. Laboratory testing can be used to diagnosis groups [14]. Another surgical approach for
hemolysis, which can be trended to determine if replacement of left ventricular assist device, par-
pump thrombosis persists or is resolving. Once a ticularly HeartMate II, is the subcostal approach.
patient has become hemodynamically stable, fur- If pump thrombosis is thought to be located to
ther assessment can be performed on the pump inflow cannula and sternotomy can be avoided,
itself in addition to ramp studies to determine if then the subcostal approach is generally consid-
appropriate LV unloading is occurring. ered for surgical removal [15].
If symptoms of hemolysis resolve with adjust-
ment to anticoagulation and antiplatelet regimen,
then careful evaluation must be undertaken to Conclusion
determine if pump thrombosis has truly resolved.
Serial ramp echocardiographic studies may be Left ventricular assist device thrombosis is a
necessary to verify adequate LV unloading, and complication that clinicians practicing in
symptoms of hemolysis should resolve. In these advanced heart failure centers should be comfort-
select patients, higher INR goals with additional able managing. While each institution may have
antiplatelet therapy may be required to minimize its own protocol for management of pump throm-
future risk of device thrombosis. In certain cir- bosis, a clear understanding of precipitating fac-
cumstances when devices are implanted as bridge tors and further clinical workup can become
to transplant, patients who present with active helpful in making a prompt diagnosis. Early
hemolysis, pump device dysfunction, or visually diagnosis and stabilization of the patient can lead
detected pump thrombosis may qualify for to better postoperative outcomes following surgi-
UNOS 1A transplant status. cal replacement.
A majority of patients with pump thrombosis
will ultimately require surgical replacement.
Multiple factors can play a role in the approach References
for surgical removal of a pump such as bleeding
risk, underlying pulmonary disease, BMI, and 1. Slaughter MS, Rogers JG, Milano CA, Russell SD,
Conte JV, Feldman D, et al. Advanced heart failure
any ongoing infections. One common discussion treated with continuous-flow left ventricular assist
that occurs when surgical replacement of the device. N Engl J Med. 2009;361:2241–51.
pump is deemed necessary is the risk associated 2. Park SJ, Milano CA, Tatooles AJ, Rogers JG, Adamson
with resternotomy. A resternotomy approach to RM, Steidley E, et al. Outcomes in advanced heart
failure patients with left ventricular assist devices for
surgical removal and replacement of an LVAD is destination therapy. Circ Heart Fail. 2012;5:241–8.
one approach to consider if complete visualiza- 3. Kirklan JK, Naftel DC, Pagani FD, Kormos RL,
tion of the device is required. Performing a ster- Myers S, Acker MA, et al. Pump thrombosis in the
notomy is particularly useful if inflow cannula Thoratec HeartMate II device: an update analysis of
the INTERMACS Registry. J Heart Lung Transplant.
position needs to be replaced or the entire pump 2015;34:1515–26.
requires exchange. This approach is particularly 4. Kilic A, Acker MA, Pavan A. Dealing with surgical
helpful if concern for outflow graft thrombosis is left ventricular assist device complications. J Thorac
present. However, this approach does have sev- Dis. 2015;7(12):2158–64.
5. Trachtenberg BH, Cordero-Reyes AM, Aldeiri M,
eral risk factors associated with it in a patient who Alvarez P, Bhimaraj A, Ashrith G, et al. Persistent
may have already undergone prior open-­ heart blood stream infection in patients supported with a
procedures. Due to buildup of scar and inflamma- continuous-flow left ventricular assist device is asso-
tion dissection, planes can become difficult to ciated with an increase risk of cerebrovascular acci-
dents. J Card Fail. 2015;21(2):119–25.
visualize and the risk of bleeding increases with
12  Diagnosis of Device Thrombosis 197

6. Starling RC, Moazami N, Silvestry SC, Ewald G, for mechanical circulatory support: task force 5: out-
Rogers JG, Milano CA, et al. Unexpected abrupt patient management of the mechanical circulatory
increase in left ventricular assist device thrombosis. N support device recipient. J Heart Lung Transplant.
Engl J Med. 2014;370:33–40. 2013;32(2):157–87.
7. Yang F, Kormos RL, Antaki JF. High-speed visual- 12. Uriel N, Levin AP, Sayer GT, Mody KP, Thomas
ization of disturbed pathlines in axial flow ventricu- SS, Adatya S, et al. Left ventricular decompression
lar assist device under pulsatile conditions. J Thorac during speed optimizations ramps in patients sup-
Cardiovasc Surg. 2015;150(4):938–44. ported by continuous-flow left ventricular assist
8. Hurst TE, Moazami N, Starling RC. Left ventricular devices: device specific performance characteristics
assist device thrombosis in the setting of left ventricu- and impact on diagnostic algorithms. J Card Fail.
lar recovery. J Heart Lung Transplant. 2014;34:622–3. 2015;21:785–91.
9. Stainback RF, Estep JD, Agler DA, Birks EJ, Bremer 13. Maltais S, Kilic A, Nathan S, Keebler M, Emani

M, Hung J, et al. Echocardiography in the manage- S, Ransom J, et al. PREVENtion of HeartMate II
ment of patient with left ventricular assist devices: pump thrombosis through clinical management:
recommendations from the American Society the PREVENT multi-center study. J Heart Lung
of Echocardiography. J Am Soc Echocardiogr. Transplant. 2017;36:1–12.
2015;28:853–909. 14. Tsiouris A, Brewer RJ, Borgi J, Hodari A, Nemeh
10. Bartoli CR, Ghotra AS, Pachika AR, Birks EJ,
HW, Cogan CM, et al. Is Resternotomy a risk for
McCants KC. Hematologic markers better predict continuous-­flow left ventricular assist device out-
left ventricular assist device thrombosis than echo- comes. J Card Surg. 2012;28:82–7.
cardiography or pump parameters. Thorac Cardiovasc 15. Halbreiner MS, Tong MZ, Moazami N. Subcostal

Surg. 2014;62(05):414–8. approach to replacement of a HeartMate II device:
11. Pamboukian SV, Elliot T, Mohacsi P, Potapov EV, indication and technique. Oper Tech Thorac
Russell SD, Teuteberg JJ. The 2013 International Cardiovasc Surg. 2014;19:443–53.
Society of Heart and Lung Transplantation guidelines
 Device Exchange: THI Technique
Involving a Left Subcostal 13
Approach

Tadahisa Sugiura and Masashi Kawabori

cal strategy. From this information, the surgeon

Introduction should consider whether total or partial replace-
ment of the device is required. If only isolated
Continuous-flow left ventricular assist devices pump exchange is planned, it can be performed
(CF-LVADs) are being used to support patients by a left subcostal approach.
with end-stage heart failure for longer periods Optimization of patient’s condition is man-
because of both longer transplant waiting times datory before the surgery. If the patient’s hemo-
and the use of these devices as destination ther- dynamic status is compromised, inotropic
apy. Longer durations of support make CF-LVAD support or temporary mechanical circulatory
exchange more common. In this chapter, we support should be started immediately.
present our technique of HeartMate II (St. Jude Mechanical respiratory support and pulmonary
Inc., St. Paul, MN) pump exchange through a left vasodilators such as nitric oxide and epopros-
subcostal approach. tenol are useful for managing right heart failure.
A Swan-Ganz catheter can be used to assess car-
diac function and optimize volume status. Major
Preoperative Considerations neurological issues should be ruled out before
the operation.
A CF-LVAD exchange is planned in cases of Patients with an active pump infection require
pump thrombosis, pump infection, or driveline aggressive preoperative antibiotic therapy. Any
issues. A preoperative complete system workup coagulation abnormalities should be corrected
is mandatory before all exchange cases [1]. and anticoagulant effects should be reversed
Physical examination, medical history, and before the procedure.
medication compliance should be reviewed.
Echocardiography should be performed in all
patients to assess hemodynamics. A computed  echnique of Pump Exchange
T
tomography scan is necessary for planning surgi- Through a Left Subcostal Approach

We perform isolated pump exchange through a

T. Sugiura, M.D. (*) • M. Kawabori, M.D. left subcostal approach by a technique that does
Cardiothoracic Transplant and Mechanical not require sternotomy [2]. The exchange can be
Circulatory Support, Texas Heart Institute at Baylor completed without CPB if sufficient hemody-
St. Luke’s Medical Center, Baylor College of namic parameters can be maintained at a pump
Medicine, Houston, TX, USA
e-mail: sugiura_sskr@[Link]; kawabori. speed of approximately 6000 rpm; nonetheless,
masashi@[Link] a femoral artery is exposed in case CPB becomes

© Springer International Publishing AG 2018 199

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
200 T. Sugiura and M. Kawabori

necessary. A left subcostal incision is created the exposed portion of the outflow graft and on
two fingerbreadths below the left costal margin, the silicone elastic bellows, and the old driveline
from the midline to the mid-axillary line, is transected. The outflow collar is rotated coun-
extending through the abdominal musculature terclockwise to detach the outflow graft
to the pump pseudo-capsule. A heavy-duty self- (Fig. 13.1d). Surgeons often need a tubing clamp
retaining retractor (Thompson Surgical to loosen the collar. The inflow collar is held with
Instruments, Inc., Traverse City, MI) is neces- a tubing clamp (Fig. 13.1e), while the old pump
sary for adequate exposure (Fig. 13.1a). One housing, now disconnected from the pump outlet
arm is used to pull the left costal margin anteri- and the driveline, is rotated counterclockwise to
orly and cephalad, and a second arm is used to unscrew it from the old inlet components
retract the left side of the rib cage laterally. The (Fig. 13.1f).
incision is often extended several centimeters to The new pump is then attached to the old
the right of the midline, and the rib junction is inflow and outflow components. If the patency
resected to expose circumferentially the connec- of the inflow segments is uncertain, the vascu-
tion between the outflow-­graft collar and the lar clamp on the silicone elastic bellows can be
detachable outflow bend-relief. The bend-relief released for a heartbeat to ensure brisk back-
is detached, and dissection is performed around flow of the blood. This maneuver can cause air
the outflow graft to make enough space for a entrainment, but this is unlikely to happen
vascular clamp to be applied (Fig. 13.1b). because these patients have elevated LV dia-
Excessive traction of the outflow graft should be stolic pressure. If the inflow becomes partly
avoided because this can damage the graft, mak- obstructed, the surgeon should consider start-
ing it vulnerable to stenosis. ing CPB and repairing the inflow, either
Next, additional dissection is done around the through the existing subcostal incision or by
inflow cannula to allow a vascular clamp to be performing a redo sternotomy. We attach the
applied across the white silicone elastic bellows new pump to the old inflow components by
between the sintered titanium inflow cannula and holding the inflow collar firmly with a tubing
the pump. When the vascular clamp is placed, the clamp, then screwing the new pump into the
bellows and graft are compressed to temporarily collar by rotating the entire pump clockwise.
occlude inflow (Fig. 13.1c). The old driveline is We keep the driveline from interfering with the
dissected circumferentially for several centime- attachment process by rewinding the new
ters so that the driveline can be easily transected pump housing several turns counterclockwise
before the pump is removed. before engaging the screw threads on the
Heparin is administered systemically, and inflow collar.
pump speed is gradually decreased to slowly The outflow graft is then attached to the new
wean the patient from the HeartMate II. pump, and we confirm the patency of the graft.
Transesophageal echocardiography is used to Typically, we do this by releasing the outflow-­
determine whether hemodynamic stability can be graft clamp to ensure brisk arterial backflow.
maintained for 10–15 min without CF-LVAD However, in some cases, we instead measure the
support. Inotropes and vasoactive agents are used pressure inside the graft by inserting a needle
as needed. Patients with extremely poor LV func- through the graft proximal to the clamp and
tion may require CPB. attaching the needle to a transducer. Any
The new driveline is tunneled through the ­difference between the pressure within the graft
abdominal wall and brought out in a suitable and the systemic arterial pressure is a sign of
position. The inflow components included with outflow-­graft stenosis, potentially indicating
the new pump are not used; if they are already outflow-­graft repair.
attached to the new pump’s housing, they should A 19-gauge needle is inserted into the out-
be removed. Two vascular clamps are placed on flow graft for de-airing. The LVAD pump is
13  Device Exchange: THI Technique Involving a Left Subcostal Approach 201

Fig. 13.1 (a) The pump is exposed through an extended unscrewed from the curved titanium tube at the pump out-
left subcostal incision. A self-retaining retractor is invalu- let. Brisk backflow is confirmed by transiently releasing
able in visualizing the white silicone rubber bellows that the clamp to ensure adequacy of the outflow graft. (e) A
surrounds the pump inlet graft. (b) The bend-relief is dis- pump clamp is used to firmly grasp the collar at the pump
connected from the pump outlet so that the outflow graft inlet. (f) While the collar is held, the pump is rotated coun-
can be exposed. After hemodynamic stability is ensured, terclockwise until it can be removed from the inlet. Brisk
the pump power is turned off, and the driveline is cut. (c) bleeding from the inlet is confirmed by transiently releas-
Clamps are immediately placed on the outlet graft and the ing the clamp on the inflow bellows to ensure adequacy of
silicone rubber inflow bellows to prevent retrograde flow the pump inflow. The new pump is implanted by perform-
through the nonfunctioning pump. (d) The outflow graft is ing these steps in the reverse order
202 T. Sugiura and M. Kawabori

initiated, and the pump speed is gradually References

increased, while transesophageal echocardiog-
raphy is used to monitor hemodynamics and LV 1. Gregoric ID. Exchange techniques for implantable
ventricular assist devices. ASAIO J. 2008;54(1):14–9.
size. Protamine is administered to maintain 2. Cohn WE, Mallidi HR, Frazier OH. Safe, effective
hemostasis. The wound is closed in layers in the off-pump sternal sparing approach for HeartMate II
standard fashion. exchange. Ann Thorac Surg. 2013;96(6):2259–61.
 Mechanical Circulatory Support
to Bridge to a Long-Term 14
Continuous-Flow Left Ventricular
Assist Device as a Bridge to Heart
Transplantation

Chitaru Kurihara

Baylor College of Medicine clinical LVAD

Introduction database indicated that low preoperative levels
of albumin or prealbumin and a high Model of
Over the past decade, the use of left ventricular assist End-Stage Liver Disease-eXcluding INR
devices (LVADs) for long-term therapy has increased (MELD-XI) score are predictors of poor patient
exponentially [1–3]. In addition, the indications for outcomes after LVAD implantation (Fig. 14.1a–
implantation of mechanical devices have expanded c). Preoperative malnutrition has been shown to
over time and now include cardiogenic shock, bridge increase postoperative morbidity and mortality
to transplant (BTT), bridge to decision, and destina- in patients who undergo cardiac surgery [4].
tion therapy (DT). On the basis of their features, The preoperative diagnosis of malnutrition
devices can be separated into various subcategories could help identify patients who may benefit
of use, including short-term versus long-term from improving nutritional status before implan-
devices and assist devices versus complete heart tation surgery. It is also important to implant
replacement (i.e., total artificial heart). This chapter LVAD in patients before development of end-
will focus on the use of LVADs for long-term sup- organ failure.
port as BTT and will cover critical factors relating to Contraindications for LVAD implantation
types of devices, including the use of short-term include irreversible end-organ failure, particu-
mechanical circulatory support (MCS) to bridge to larly renal, hepatic, and respiratory, which are
BTT, and the importance of patient selection. uniformly independent predictors of poor out-
come [5–7]. Severe, unrecoverable neurological
dysfunction is also a contraindication for LVAD
Patient Selection implantation. Systemic sepsis poses a signifi-
cant risk to patients who undergo LVAD implan-
Prudent patient selection is a critical element in tation because it can cause a profound,
achieving good clinical results with LVAD sup- refractory, vasodilatory state and an increased
port. Analysis of the Texas Heart Institute/ incidence of infections such as device-related
endocarditis [8, 9]. Patients who have had sepsis
C. Kurihara, M.D. (*) should have two negative blood cultures over a
Cardiothoracic Transplant and Mechanical 1-week period before LVAD implantation to
Circulatory Support, Texas Heart Institute at Baylor
indicate that the infection has been cleared from
St. Luke’s Medical Center, Baylor College of
Medicine, Houston, TX, USA the bloodstream. Another contraindication for
e-mail: chitaru1207@[Link] LVAD implantation is the presence of a malig-

© Springer International Publishing AG 2018 203

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
204 C. Kurihara

Fig. 14.1 (a) Survival a

curves for patients who 1.0 + Censored
had normal albumin
levels (≥3.5 g/dL),
moderate 0.8
hypoalbuminemia

Survival Probability
(2.5–3.5 g/dL), and
severe hypoalbuminemia 0.6
(<2.5 g/dL) before
undergoing implantation
of a continuous-flow left 0.4
ventricular assist device.
(b) Survival curves for
0.2 Albumin > 3.5 g/dL
patients who had
preoperative prealbumin Albumin 2.5-3.5 g/dL P<0.001
levels ≥17 or <17 after Albumin < 2.5 g/dL
implantation of a 0.0
continuous-flow left 0 500 1000 1500 2000
ventricular assist device. days
(c) Survival curves for b
patients who had 1.0 + Censored
preoperative MELD
score ≥17 or <17 after
implantation of a 0.8
continuous-flow left
Survival Probability

ventricular assist device

0.6

0.4

Prealb<17
P=0.04
0.2 Prealb≥17

0.0
0 500 1000 1500 2000
days
c
1.0 + Censored

0.8
Survival Probability

0.6

0.4

MELD-XI<17
0.2 P<0.01
MELD-XI≥17

0.0
0 500 1000 1500 2000
days
14  Mechanical Circulatory Support to Bridge to a Long-Term Continuous-Flow Left Ventricular Assist… 205

nancy leading to a life expectancy of less than (VA-ECMO), and CentriMag (Thoratec
2 years. Each of these cases requires individual Corporation, Pleasanton, CA) devices.
attention and e­valuation for appropriate deci- Appropriate patient selection is the central
sion-making. A patient with human immunode- tenet of the current paradigm of MCS. Indeed,
ficiency virus (HIV) infection who is compliant compared with more stable patients on medical
with medical therapy and who has a normal therapy, inotrope-dependent patients with rapid
CD4 count and undetectable viral levels should deterioration and end-organ dysfunction have
be considered for LVAD implantation. unacceptable outcomes, including a 1-year mor-
tality rate of nearly 50% among those who sur-
vive to discharge while receiving long-term
Types of Devices LVAD support [12, 13]. Patients who receive
short-term MCS before LVAD implantation are
Using Short-Term MCS sicker at baseline than LVAD-only patients, and
short MCS is used to optimize patients’ condition
Data from the Interagency Registry for before LVAD placement and to increase their
Mechanically Assisted Circulatory Support suitability as long-term MCS candidates. Despite
(INTERMACS) patient profiles indicate that their poorer condition at baseline, MCS-­
mortality and morbidity rates are worst in pro- supported patients had similar outcomes to
file 1 patients (critical cardiogenic shock). LVAD-only patients, suggesting that short-term
Because INTERMACS profile 1 patients have a MCS decreases preoperative risk and achieves
poor survival (ranging from 65 to 76% at outcomes similar to those using LVAD support
1 year), the patient population undergoing only. In a meta-analysis of data from patients
LVAD implantation is shifting from with cardiogenic shock who were randomly
INTERMACS profile 1 toward those with less assigned to receive a percutaneous ventricular
severe illnesses such as INTERMACS profiles assist device (p-VAD, including TandemHeart
2 or 3. In 2008, 30% of LVAD implantations in and Impella) or IABP support, outcomes were
the INTERMACS registry were performed in similar between the groups, although p-VAD
profile 1 patients, whereas that percentage support appeared to improve hemodynamics [13,
decreased to 15% in 2013. Therefore, short- 14]. Clinical criteria for using MCS are difficult
term MCS has become a necessary component to determine, but in our center, short-term MCS
of the therapeutic strategy for patients in car- is initiated when patients show signs of signifi-
diogenic shock, and implanting a short-term cant hemodynamic instability or end-organ dys-
MCS device as a bridge to decision has gained function, such as renal or respiratory failure,
popularity. This approach is used to stabilize despite maximum medical support. Previous
the patient’s hemodynamic status and improve studies suggest that VA-ECMO is as safe and
end-organ function, thereby reducing the surgi- effective as a p-VAD to bridge patients to more
cal risk associated with implanting an LVAD advanced therapies such as heart transplantation
[10, 11]. Ideal strategy with short-term MCS or long-term LVAD support [15–18]. However,
would reduce surgical risk before LVAD patients with VA-ECMO had only a 40–50% rate
implant with stabilizing hemodynamics and of survival to recovery or next therapy. One
improving end-organ function. The most com- potential advantage of p-VADs over ECMO is
monly used short-term MCS device is the intra-­ that their design and mechanism of action allow
aortic balloon pump (IABP); other forms of direct ventricular unloading, thereby reducing
MCS include the TandemHeart (CardiacAssist the myocardial oxygen demand and the workload
Inc., Pittsburgh, PA), Impella (Abiomed, of the failing heart [19]. In contrast, a sizable pro-
Danvers, Massachusetts, MA), venoarterial portion of patients supported with ECMO may
extracorporeal membranous oxygenation develop refractory pulmonary edema, necessitat-
206 C. Kurihara

ing ventricular decompression secondary to ventricular assist device as bridge-to-transplant ther-

apy. Ann Thorac Surg. 2008;86(4):1227–34; discus-
increased afterload [20].
sion 1234–25.
2. Slaughter MS, Pagani FD, Rogers JG, et al. Clinical man-
agement of continuous-flow left ventricular assist devices
Using Long-Term MCS in advanced heart failure. J Heart Lung Transplant.
2010;29(4 Suppl):S1–39.
3. Slaughter MS, Rogers JG, Milano CA, et al.
 xial Flow Pump for Long-Term
A Advanced heart failure treated with continuous-­
Mechanical Support: Thoratec flow left ventricular assist device. N Engl J Med.
HeartMate II 2009;361(23):2241–51.
4. Rich MW, Keller AJ, Schechtman KB, Marshall
The most commonly used pump at our institution
WG Jr, Kouchoukos NT. Increased complications
is the Thoratec HeartMate II LVAD (HM II; and prolonged hospital stay in elderly cardiac surgi-
Thoratec Corp.), which is an axial flow rotary cal patients with low serum albumin. Am J Cardiol.
pump constructed of titanium. Smaller than the 1989;63(11):714–8.
5. Gracin N, Johnson MR, Spokas D, et al. The use of
Thoratec HeartMate XVE (XVE; Thoratec
APACHE II scores to select candidates for left ven-
Corp.), the HM II pump housing is implanted in tricular assist device placement. Acute Physiology and
the peritoneal space and requires a less invasive Chronic Health Evaluation. J Heart Lung Transplant.
operative approach. It can provide flow up to 1998;17(10):1017–23.
6. Oz MC, Goldstein DJ, Pepino P, et al. Screening
10 L/min at pump speed of 6000–15,000 RPM,
scale predicts patients successfully receiving long-­
with inflow via the left ventricular apex or dia- term implantable left ventricular assist devices.
phragm and outflow via the ascending aorta. A Circulation. 1995;92(9 Suppl):II169–73.
small percutaneous driveline exits the skin in the 7. Rao V, Oz MC, Flannery MA, Catanese KA,
Argenziano M, Naka Y. Revised screening scale
right upper abdomen. Patients are placed on sys-
to predict survival after insertion of a left ven-
temic anticoagulation with warfarin and antiplate- tricular assist device. J Thorac Cardiovasc Surg.
let therapy with aspirin to prevent thromboembolic 2003;125(4):855–62.
events. The Food and Drug Administration (FDA) 8. Argenziano M, Catanese KA, Moazami N, et al.
The influence of infection on survival and successful
has approved the HM II for BTT and DT.
transplantation in patients with left ventricular assist
devices. J Heart Lung Transplant. 1997;16(8):822–31.
 entrifugal Pump for Long-Term
C 9. Schulman AR, Martens TP, Christos PJ, et al.
Mechanical Support: HeartWare Comparisons of infection complications between con-
tinuous flow and pulsatile flow left ventricular assist
The HeartWare HVAD (HVAD; HeartWare
devices. J Thorac Cardiovasc Surg. 2007;133(3):841–2.
International Inc., Framingham, MA) is a cen- 10. Lima B, Kale P, Gonzalez-Stawinski GV, Kuiper JJ,
trifugal pump that has no mechanical bearings Carey S, Hall SA. Effectiveness and safety of the
and that weighs 145 g, with a displaced stroke Impella 5.0 as a bridge to cardiac transplantation or
durable left ventricular assist device. Am J Cardiol.
volume of 45 cm3. It can provide flow up to 10 L/
2016;117(10):1622–8.
min at 2000–3000 RPM. The device is implanted 11. Shah D. Does targeting ibutilide-resistant CFAE

in the pericardial space without the need for an improve outcomes for catheter ablation of persistent
abdominal incision. The inflow cannula is inte- AF? Eur Heart J. 2016;37(20):1622–5.
12. Boyle AJ, Ascheim DD, Russo MJ, et al. Clinical
grated into the left ventricle. A single, flexible
outcomes for continuous-flow left ventricular
driveline (4.2 mm in diameter) exits the anterior assist device patients stratified by pre-operative
abdomen. The HVAD is approved by the FDA for INTERMACS classification. J Heart Lung Transplant.
BTT and DT. 2011;30(4):402–7.
13. Kirklin JK, Naftel DC, Kormos RL, et al. Interagency
Registry for Mechanically Assisted Circulatory
Support (INTERMACS) analysis of pump thrombo-
References sis in the HeartMate II left ventricular assist device.
J Heart Lung Transplant. 2014;33(1):12–22.
1. John R, Kamdar F, Liao K, Colvin-Adams M, Boyle 14. Miller LW, Guglin M. Patient selection for ventricu-
A, Joyce L. Improved survival and decreasing inci- lar assist devices: a moving target. J Am Coll Cardiol.
dence of adverse events with the HeartMate II left 2013;61(12):1209–21.
14  Mechanical Circulatory Support to Bridge to a Long-Term Continuous-Flow Left Ventricular Assist… 207

15. Gray BW, Haft JW, Hirsch JC, Annich GM,

18. Rastan AJ, Dege A, Mohr M, et al. Early and late out-
Hirschl RB, Bartlett RH. Extracorporeal life sup- comes of 517 consecutive adult patients treated with
port: experience with 2,000 patients. ASAIO extracorporeal membrane oxygenation for refrac-
J. 2015;61(1):2–7. tory postcardiotomy cardiogenic shock. J Thorac
16. Hoefer D, Ruttmann E, Poelzl G, et al. Outcome
Cardiovasc Surg. 2010;139(2):302–11, 311.e1.
evaluation of the bridge-to-bridge concept in patients 19. Kawashima D, Gojo S, Nishimura T, et al. Left

with cardiogenic shock. Ann Thorac Surg. 2006; ventricular mechanical support with Impella pro-
82(1):28–33. vides more ventricular unloading in heart failure
17. Pagani FD, Aaronson KD, Swaniker F, Bartlett
than extracorporeal membrane oxygenation. ASAIO
RH. The use of extracorporeal life support in adult J. 2011;57(3):169–76.
patients with primary cardiac failure as a bridge to 20. Beurtheret S, Mordant P, Pavie A, Leprince P. Impella
implantable left ventricular assist device. Ann Thorac and extracorporeal membrane oxygenation: a demand-
Surg. 2001;71(3 Suppl):S77–81; discussion S82–75. ing combination. ASAIO J. 2012;58(3):291–3.
 Outcomes Using LVADs
for Destination Therapy 15
Priyanka Sen and Selby Oberton

In this chapter, we aim to present a summary of

Introduction the work of landmark trials that have studied the
outcomes of LVADs used for DT.
Most recent estimates cited by the AHA suggest
that >250,000 patients in the United States are in
end-stage heart failure that is medically refrac- Survival
tory [1]. This is a staggering number when com-
pared to the >2000 transplants that are performed REMATCH was the first landmark trial to dem-
annually in the United States [2]. Implantation of onstrate that implantation of an LVAD (HeartMate
a left ventricular assist device (LVAD) as destina- vented electric device (HM VE LVAD)) as DT
tion therapy (DT) is a hopeful alternative for provided a significant survival advantage com-
those who qualify. Approved for use as destina- pared to optimal medical therapy (OMT) in
tion therapy, continuous-flow LVADs are quieter patients with end-stage heart failure who were
and more streamlined and involve fewer moving ineligible for cardiac transplantation [8].
parts that may fail, giving them more longevity Conducted between 1998 and 2001, it random-
than their pulsatile counterparts and more suit- ized 129 patients (68 to the LVAD and 61 to the
ability for long-term use [3]. Though far superior OMT arms) who were deemed ineligible for
than optimal medical therapy alone, LVADs are transplantation with NYHA class IV heart failure
not without their problems which include but are symptoms, LVEF < 25%, either peak oxygen
not limited to bleeding, infection, thromboem- consumption <12 mL/kg/min or dependence on
bolic events, and pump thrombosis [4]. The use IV inotropic infusion, OMT use for at least 60 of
of risk stratification to carefully select patients the last 90 days, and a projected life expectancy
and time of LVAD implantation may help of less than 2 years. Analysis of the patients ran-
­minimize the potential complications [5, 6] and domized to the HM VE LVAD and OMT arms
help predict outcomes after implantation [1, 7]. demonstrated a rate of survival at 1 year of 52%
in the device group and 25% in the OMT group
(p = 0.002) with a median survival advantage of
8.5 months [8]. This amounted to a RR of 48% in
P. Sen, M.D. • S. Oberton, M.D. (*) the primary endpoint of death from any cause and
Texas Heart Institute at Baylor St. Luke’s Medical
therefore an absolute risk reduction of 27% at
Center, Baylor College of Medicine,
Houston, TX, USA 1 year (Fig. 15.1) [8]. The HM VE LVAD as DT
e-mail: psen1216@[Link]; [Link]@[Link] was thus touted to have a remarkable treatment

© Springer International Publishing AG 2018 209

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
210 P. Sen and S. Oberton

Fig. 15.1  Taken from

the REMATCH trial, this
Kaplan-Meier analysis
depicts a reduction of
48% in the risk of death
from any cause (primary
endpoint) in the group
that received the LVAD
versus the group that
received optimal medical
therapy alone [7]

effect nearly four times that of beta-blockers or Enhancements in LVAD design based on les-
an ACE inhibitor, which individually had been sons learned from the REMATCH trial led to the
estimated to prevent 70 deaths for every 1000 development of the HeartMate XVE (HM XVE).
patients [8]. In a nonrandomized, prospective trial, Lietz et al.
At the time of its publication, the REMATCH studied the outcomes of 280 patients who were
trial was noteworthy for having enrolled heart implanted with the modified HM XVE between
failure patients with the most severe clinical and 2001 and 2005 for DT [1]. This study aimed to
hemodynamic compromise and mortality rates investigate the impact of this improved pulsatile-­
to date [1, 9]. An unanticipated 71% of patients flow LVAD on DT outcomes and to identify clini-
were on inotropic infusions at the time of ran- cal predictors portending worse prognosis that
domization [9]. Post hoc analysis of the out- could then be made into a risk score to stratify
comes of patients on and off intravenous DT candidates. Rates of survival were 86.1%,
inotropic therapy at the time of randomization 56.0%, and 30.9% at 30 days, 1 year, and 2 years
confirmed that the patients on inotropes derived after LVAD implantation, respectively [1]. The
a near doubling of survival benefit from LVAD high 1-year survival rates among recipients of
implantation with a 1-year survival of 49% and HM XVE as DT were corroborated by a smaller,
24% in the LVAD and OMT groups, respectively albeit nonrandomized trial. This study by Long
(p = 0.0014) [9]. Survival of patients on baseline et al. compared the survival of 48 recipients of
inotrope therapy was equal to or better than that the HM XVE at four of the highest volume cen-
in the OMT group at all times, even in spite of ters participating in the Thoratec DT registry
the predicted excess of perioperative mortality with that seen in the historical LVAD arm of the
due to LVAD implantation [9]. The difference in REMATCH trial [10]. It found a statistically non-­
1-year survival rates was, however, not statisti- significant but nonetheless remarkable 40%
cally significant between the LVAD (57%) and decline in the rate of death per patient year from
OMT (40%) groups who were not on intrave- any cause in the HM XVE arm as compared to
nous inotropic support at baseline (p = 0.55) [9]. the HM VE arm of the REMATCH trial [10]. The
It was thus concluded that LVAD implantation lesser mortality rates were attributed to improved
was most beneficial for the sickest patients with LVAD design and patient management protocols
advanced heart failure. from years of experience.
15  Outcomes Using LVADs for Destination Therapy 211

While pulsatile-flow LVADs had earned Conducted between 2005 and 2007 and using
acceptance as therapy for refractory heart failure, similar eligibility criteria as the REMATCH trial
their limited long-term durability and large pump which have since been adapted into the CMS
size created a need for the simpler and smaller (Center for Medicare and Medicaid Services) cri-
design of continuous-flow LVADs (CF-LVADs) teria [12], this multicenter trial randomized 200
[4, 11]. The HeartMate II trial was among the DT patients to either CF-LVAD (HeartMate II) or
earliest of studies that looked at the outcomes pulsatile-flow HeartMate XVE groups [13]. The
of CF-LVADs, which were being tested only as primary endpoint of a composite of survival at
a bridge-to-transplant at the time. By 180 days 2 years, free of disabling stroke (stroke with a
of the study, 100 of the 133 patients implanted Rankin score > 3) or the need for reoperation to
with a HM II had either undergone cardiac trans- replace the device, was achieved by 46% and 11%
plantation, had significant cardiac recovery, or in the HM II and HM XVE groups, respectively,
were still on ongoing mechanical support while with p < 0.001 [13]. Subgroup analysis of 1- and
remaining eligible for transplantation [11]. 2-year survival rates showed similarly significant
Although the results of a study on BTT patients results of 68% and 58%, respectively, in the
cannot be directly applied to DT patients, as the CF-LVAD group and 55% and 24%, respectively,
degree of severity of heart failure and associ- in the pulsatile-flow LVAD group (Fig. 15.2) [13].
ated comorbidities making a patient ineligible
for transplant and hence a DT patient make for a
much sicker population, this study did show the Adverse Events
promise of CF-LVADs.
The staggering results of REMATCH and While both continuous and pulsatile-flow LVADs
other trials individually comparing the survival demonstrated significant survival benefits in
benefit among patients implanted with improved patients with end-stage heart failure who were
pulsatile-flow LVADs and continuous-flow ineligible for transplantation, their use was not
LVADs begged the need for a head-to-head com- without complications. In the REMATCH trial,
parison of pulsatile-flow and CF-LVAD therapies. for instance, patients randomized to the device

Fig. 15.2  Taken from

the HeartMate II
destination therapy trial,
this Kaplan-Meier
analysis depicts the
striking superior survival
rates seen up to
24 months after
implantation of the
continuous-flow
HeartMate II compared
to the pulsatile-flow
HeartMate XVE [12]
212 P. Sen and S. Oberton

group were 2.35 times as likely as those in the of device failure in 24 months. Within the device
OMT group to suffer a serious adverse event [8]. group, ultimately, 10 of the 68 patients required
Infection, suspected malfunction of the LVAD, replacement [8]. Largely due to improvements in
non-neurologic bleeding, and neurologic dys- LVAD design, CF-LVADs unsurprisingly fared
function were the most common adverse events better in this regard. As the HeartMate II study
associated with the LVAD group [8]. showed, fewer patients in the HM II group
required reoperation to repair or replace the pump
than those in the HM XVE group (p < 0.001) [13].
Infection

Infection was by far the most frequent adverse Non-neurologic Bleeding

event in LVAD studies and, in the era of pulsatile
pumps, was particularly commonplace. The Non-neurologic bleeding has also been a vexing
REMATCH trial stated that 28% of patients problem with LVADs. Balancing the risk of
implanted with HM VE developed infection device-related thromboembolic events with an
within the first 3 months, with most cases related increased propensity for bleeding due to antico-
to a local driveline tract and pocket infection. agulation and the theorized increased development
While most could be treated with local antibiot- of acquired von Willebrand syndrome and arterio-
ics, sepsis still claimed 17 of 68 lives [8]. A com- venous malformations in the setting of chronically
parison of the results of HM XVE and HM II low pulse-pressure has been a struggle with LVAD
recipients showed that CF-LVADs significantly management [17, 18]. In the REMATCH study,
reduced the rates of LVAD-associated infections even despite a lack of routine anticoagulation, the
by 50% as well as that for non-LVAD-associated frequency of bleeding within the first 6 months
local infections and sepsis [13]. The larger area following implantation with the HM VE was as
of surgical dissection required for implantation high as 42% [8]. During the early trials using HM
of a pulsatile-flow LVAD and the percutaneous II for BTT, some centers had adopted a stringent
driveline were felt to be the most likely causes of antiplatelet and anticoagulation regimen of aspirin
increased infection risk associated with pulsatile-­ and dipyridamole with a postoperative heparin
flow pumps [11, 13, 14]. Additionally, CF-LVADs bridge to warfarin to achieve a target international
lack the compliance chambers, polyurethane normalized ratio or INR of 2.0–3.0 [11]. But this
membranes, and prosthetic valves that may regimen resulted in a significantly increased rate
become niduses for infection [15]. Indeed, the of bleeding events, especially in the early postop-
transition to the use of CF-LVADs and the higher erative period [11]. A comparison of HM II with
cumulative experience of LVAD management at HM XVE outcomes showed that both types of
high-volume centers correlated with reduced LVAD were associated with ten times the rate of
overall LVAD-associated infections. Institutional bleeding as thromboembolic events, even though
changes to use management guidelines, abdomi- only those assigned to the HM II group were anti-
nal binders to stabilize the percutaneous drive- coagulated [13]. Based on these results, many cen-
line, and antibiotic prophylaxis likely contributed ters have since reduced the targeted INR to 1.5–2.5
to reduced rates of infection with LVADs over the for continuous-flow LVADs and have eliminated
years [10, 14, 16]. the heparin bridge [14].

LVAD Dysfunction Neurologic Events

Suspected LVAD dysfunction was the second Surprisingly, despite a lack of routine anticoagu-
most common adverse event seen in the lation use in the LVAD group of the REMATCH
REMATCH trial, amounting to a 35% probability trial, 76% of patients were free of serious neuro-
15  Outcomes Using LVADs for Destination Therapy 213

logic events [8]. Only 10% of patients in the Causes of Death

device group suffered an ischemic stroke [8]. The
low percentage was attributed to the textured sur- The primary causes of death among patients with
faces of the HM VE [8]. Comparison of the HM the pulsatile HeartMate VE were sepsis (41%)
XVE and HM II recipients showed similar rates followed by LVAD failure (17%) instead of termi-
of ischemic stroke between the two groups of nal heart failure which accounted for the majority
14% and 17%, respectively [13]. In fact, while on of deaths in the optimal medical therapy group of
Coumadin targeted to an INR of 2.0–3.0, the HM the REMATCH trial [8]. A trial looking at HM
II patients suffered a similar rate of ischemic XVE recipients in the post-REMATCH era found
stroke as that among other patients with advanced that the majority of in-hospital deaths (79%)
heart failure and atrial fibrillation who are not on occurred within the first 3 months [1]. During a
device support [13]. minimum follow-up period of 2 years after LVAD
Overall, head-to-head assessment of pulsatile implantation, the leading cause of death was hem-
and continuous-flow LVADs showed significant orrhagic stroke which was seen in 9 and 10% of
reductions in the rate of major adverse events patients in the pulsatile and continuous-­ flow
among CF-LVAD patients including device- and recipients, respectively [13]. This was followed
non-device-related infections, RV failure, respi- by right ventricular failure which occurred in 8
ratory failure, renal failure, and cardiac arrhyth- and 5% of the pulsatile and continuous-flow
mia [13]. The study concluded that implantation LVAD recipients, respectively [13]. External
of CF-LVADs as compared to pulsatile-flow power interruption, respiratory failure, cardiac
devices among advanced heart failure patients arrest, and bleeding each accounted for 3–4% of
being considered for destination therapy signifi- the rest of the mortalities in the HM II group [13].
cantly improved the chance of survival free of The estimated 1- and 2-year survival rates in this
stroke and the need for reoperation for device group were 68% and 58%, respectively [13].
repair or replacement at 2 years [13].
An understated confounder of the results
which showed an overwhelming superiority of  isk Score for Survival to Hospital
R
CF-LVADs when compared to pulsatile-flow Discharge and 1-Year Post HM XVE
models was the increased clinical experience that
was gained during the early years of device devel- Risk scores and risk calculators have become an
opment in the management of these complicated important tool in patient selection for LVAD and
patients. A follow-up study using the same patient in the education of patients that are interested in
pool from the REMATCH trial but followed for LVAD. In the cases in which the LVAD is used as
an additional period of 375 patient-­months found destination therapy, the patient as well as the phy-
lesser rates of adverse events among the group of sician must be aware of risk for in-hospital mor-
patients that had been enrolled during the second tality and 1-year outcomes. The authors of the
half of the trial [14]. A study of 377 patients who post-REMATCH era [1] developed a risk score
were implanted with the HeartMate I generation that can assist the discussion. Although the
of LVADs (HM VE and XVE) for DT a rising HeartMate XVE LVAD was used during the trial,
trend in 1-year survival rates related to the center the data does support the notion that sicker patient
volume of DT surgeries performed [16]. However, pre-LVAD implant does worse. Using univariate
when preoperative DT risk score adjusted analy- and multivariable analyses, the results were able
sis of 1-year survival rates showed that DT center to predict the risk factors for 90-day in-hospital
volume was not an independent predictor of sur- mortality following LVAD. Table 15.1 shows the
vival, it was surmised that other factors related to nine variables used in the multivariable model.
center experience such as better patient selection Each variable was assigned a weighted risk
and improved perioperative treatment accounted score and the cumulative risk score was calcu-
for the difference [19]. lated for each patient. A maximum score that can
214 P. Sen and S. Oberton

Table 15.1  Multivariable analysis of risk factors for 90-day in-hospital mortality after LVAD as DT [6]
Patient characteristics Odds ratio (CI) p Weighted risk score
Platelet count ≤148 × 103/μL 7.7 (3.0–19.4) <0.001 7
Serum albumin ≤3.3 g/dL 5.7 (1.7–13.1) <0.001 5
International normalization ratio >1.1 5.4 (1.4–21.8) 0.01 4
Vasodilator therapy 5.2 (1.9–14.0) 0.008 4
Mean pulmonary artery pressures ≤25 mmHg 4.1 (1.5–11.2) 0.009 3
Aspartate aminotransferase >45 U/mL 2.6 (1.0–6.9) 0.002 2
Hematocrit ≤34% 3.0 (1.1–7.6) 0.02 2
Blood urea nitrogen >51 U/dL 2.9 (1.1–8.0) 0.03 2
No intravenous inotropes 2.9 (1.1–7.7) 0.03 2

Table 15.2  Operative risk categories and risk score for 90-day in-hospital mortality after LVAD implantation as DT,
survival to hospital discharge, and 1-year survival [6]
In-hospital mortality within 90 days Survival, %
Operative risk % probability To discharge,
categories Risk score No. Observed, n Predicted, n (CI) % 90 days 1 year
Low 0–8  65  2 1.6 2 (1.1–5.4) 87.5 93.7 81.2
Medium 9–16 111 12 13.7 12 (8.0–18.5) 70.5 86.5 62.4
High 17–19  28 10 7.9 44 (32.8–55.9) 26 38.9 27.8
Very high >19  18 22 22.8 81 (66.0–90.9) 13.7 17.9 10.7

be obtained was 31. The patients were then physical-­function and emotional-role subscales
divided into four operative risk categories based of the SF-36 and the Beck Depression Inventory
on probability of mortality during the hospital- in the device arm when compared to the medical
ization. Table 15.2 and Fig. 15.3 provide the in-­ therapy group [8]. A follow-up study by
hospital mortality and survival for DT-LVAD Stevenson et al. showed that even the most
patients receiving a HeartMate XVE LVAD. gravely ill subset of patients from the REMATCH
Therefore, the cumulative risk score for in-­ trial who had been on inotropic support at the
hospital mortality after LVAD surgery ranged time of device implantation reported a decline in
from 81% in the lowest-risk candidates to a high the Minnesota Living with Heart Failure Quality
probability of postoperative death and 11% of Life Score of 77 at the time enrollment to 41
1-year survival in the highest-risk candidates. by 1 year suggesting less impairment [9]. The
HeartMate II study for destination therapy
showed that 80% of patients with CF-LVAD had
 uality of Life and Functional
Q an NYHA functional class of I or II at 24 months
Status and doubled the mean distance on the 6-min
walk test [13]. Members of both the pulsatile and
A number of studies have shown that survivors CF-LVAD groups demonstrated a statistically
of LVAD implantation had a sustained improve- significant (p < 0.001) improvement by over 30
ment in their quality of life and functional status. points in their scores from baseline on the
At 1 year since the time of surgery, the survivors Minnesota Living with Heart failure and Kansas
of HM VE implantation had a median NYHA City Cardiomyopathy questionnaires after fol-
functional class of II versus IV in the medical low-up of up to 12 months for the pulsatile-flow
therapy group (p < 0.001) [8]. The same study group and up to 24 months for the CF-LVAD
showed significantly better scores on the group [13].
15  Outcomes Using LVADs for Destination Therapy 215

Fig. 15.3  Survival after LVAD implantation [6]

LVAD Durability because of device end of life or expired as a result

of pump failure [1].
The lack of long-term durability was a concern CF-LVADs addressed some of the problems
beginning with the initial LVAD models as was that were noted with the pulsatile models. With
mentioned briefly before. Pulsatile pumps were only a single moving part, the internal rotator, the
notably bulky with a large-diameter percutaneous design was much simpler [20]. The HM II desti-
lead and pump and thus required a large patient nation therapy trial showed that 9% of the patients
habitus to accommodate the device and extensive who were supported by HM II required pump
surgical dissection for implantation [11, 13, 20]. replacement compared to 34% of those who had
In the REMATCH trial which used the HeartMate received a HM XVE (p < 0.001) [13]. Of the 59
VE, though no device had failed by 12 months of patients who were initially implanted with a
follow-up, the probability of device failure was PF-LVAD, 20 patients required a total of 21 pump
quoted to be 35% by 24 months [8]. The HM VE replacements: 3 were replaced with an alternate
was fraught with inflow-valve failure, late ero- PF-LVAD and 18 with a CF-LVAD. The most
sions of the outflow graft due to kinking, rupture common reasons for replacement were bearing
of the lining, motor failure, and wear on the bear- wear, valve malfunction, and infection [13]. Of
ings [8]. The HeartMate XVE was an upgrade the 133 patients who received a CF-LVAD ini-
with a better designed inflow valve and outflow tially, only 12 patients required a total of 13 pump
graft. Used in the post-REMATCH trial by Lietz replacements [13]. A pump replacement rate of 6
et al., the HM XVE exhibited a median time of events per 100 patients in the CF-LVAD group
LVAD support on the first pump of 18.6 months was approximately one-eighth the incidence seen
[1]. During the follow-up period ranging from in the pulsatile-flow group of patients, and the
1 day to 3.6 years (mean 10.3 months), 24.6% of most common reason was damage to the percuta-
patients either required LVAD replacement neous lead [13].
216 P. Sen and S. Oberton

Hospitalization with end-stage heart failure needing advance

heart failure therapies. Heart transplant is usu-
Among HeartMate VE recipients, the mean num- ally not a viable option in this age group.
ber of days spent in the hospital was 88 as Unfortunately, they are left with only optimal
opposed to 24 in the medical therapy group of the medical therapy. LVAD has become an option
REMATCH trial [8]. In the HM II destination for this age group but often with significant risk.
therapy trial, the median length of stay after This patient group is often left out of random-
LVAD implantation was 28 days in the HM XVE ized control trials or not the main focus which
group and 27 days in the HM II group [13]. makes advising and educating these patients
Significantly, Slaughter et al. noted a 38% rela- about the risk of implant difficult. However, a
tive reduction in the rate of rehospitalization recent retrospective study analyzed patients
among continuous-flow LVAD recipients as com- ≥70 years old. Taken from the Interagency
pared to pulsatile-flow LVAD recipients [13]. Registry for Mechanically Assisted Circulatory
Support (INTERMACS), Atluri et al. looked at
590 patients >70 years old (565 pts 70–79 years
Psychosocial Characteristics old, 25pts >80 years old) who received a contin-
and Outcome Predictors uous-flow left ventricular assist device
(CF-LVAD) and compared them to <70 years
Psychosocial factors play a key role in the man- old patients. These 70+-year-old patients were
agement of LVAD patients and can affect the out- healthier and hemodynamically stable evident
comes. Depression is often given the most by their INTERMACS score and dependence on
attention since the incidence is consistently higher inotrope. They were more common to undergo
than the general population. Prevalence and sever- re-sternotomy for the LVAD implant and have
ity of depression in heart failure vary considerably worse renal function on presentation. Length of
from 11–25% in outpatients to 35–70% among stay and mean bypass time were the same as
inpatients [21–23]. Although some studies have compared to patients <70 years old. Adverse
shown an increased risk of death and hospitaliza- events after LVAD implant revealed a signifi-
tions in patients with HF and depressive symp- cant increase in stroke (2.3% vs. 0.9%, p = 0.01)
toms, other failed to show any associations [24]. in the elderly when compared with younger,
One study looked at specific psychosocial charac- increased risk of gastrointestinal bleeding
teristics in 136 DT-LVAD patients and the asso- (19.84% vs. 13.39%, p < 0.001).
ciation with all-­cause readmission and death. In Rehospitalization (62.17% vs. 63.9%, p = 0.5),
the retrospective analysis, there was no statisti- renal dysfunction (14.72% vs. 12.78%, p = 0.2),
cally significant difference in the risk of death in respiratory failure (21.06% vs. 20.07%, p = 0.9),
regard to a multitude of psychosocial characteris- and right heart failure (14.31% vs. 14.23%,
tics. They did note current tobacco users had p = 0.9) were similar in both groups.
lower risk of readmission (adjusted HR, 0.57; Interestingly, there was a much lower incidence
95% CI, 0.38–0.88). On the other hand, illegal of driveline infection among the elderly when
drug use (HR, 1.55; 95% CI, 1.01–2.35) and compared with younger patients (5.7% vs.
depression (HR, 1.77; 95% CI, 1.40–2.22) had a 12.6%, p < 0.001). Survival (Fig. 15.4) was sig-
higher readmission risk [25]. nificantly different but acceptable, 1 year (75%
vs. 81%, p < 0.001) [26].
As more scoring systems come about and the
Elderly (>70 years old) frailty index evolves, this patient cohort will soon
gain an understanding on the outcomes and inci-
Given the increasing aging population and dence of adverse events. More importantly with a
increasing incidence of heart failure, there will study such as ROADMAP, early LVAD implant
be a large number of elderly patients presenting may help this population greatly.
15  Outcomes Using LVADs for Destination Therapy 217

Fig. 15.4 Kaplan-Meier
analysis for cumulative
survival undergoing
LVAD in patients less
than 70 or 70 and greater
[25]

hypertension in 12, 5-year cancer survival in 5,

Recovery recovery of renal function in 4, weight loss in 3,
resolution of infection in 4, and correction of
Mechanical volume and pressure unloading other issues in 16 patients [1]. In the HeartMate II
induced by LVADs allow reversal of stress-­ destination therapy trial, transplantation was ulti-
related compensatory responses triggering struc- mately possible in 9 of 66 patients in the pulsatile-­
tural and functional reverse remodeling. In flow LVAD group and 17 of 134 patients in the
theory, this would allow for cardiac recovery and continuous-flow LVAD group mainly due to a
explant of the LVAD. However, studies are lim- significant drop in pulmonary pressures [13].
ited specifically in the destination therapy group Based on data from INTERMACS as reported by
[27]. INTERMACS data suggest less than 5% of Teuteberg et al., 14.6% of patients in the destina-
patients in the DT group (553 total in the study) tion therapy group were deemed eligible for
recovered cardiac function [28]. transplant. At 6 months, 1% of DT group received
a transplant, and at 2 years only 6% of DT popu-
lation received a transplant [28].
 ransplant from Destination
T
Therapy
Conclusions and New Directions
Transplant in the destination therapy group is
uncommon but can occur. There are patients who The large discrepancy between the rising inci-
received a destination therapy device and trans- dence of medically refractory heart failure and
plant ineligible for various reasons. As age was the limited supply of available heart transplants is
the most common reason precluding patients in a challenging issue that will have to be dealt with
the REMATCH trial from cardiac transplanta- in the coming years. LVADs placed for destina-
tion, no patients enrolled in that study underwent tion therapy provide a means to mitigate the situ-
transplantation [8]. However, a study of HM ation. However, not enough LVADs are placed as
XVE DT recipients during the post-REMATCH destination therapy to meet the projected demand
era showed that 47 patients (17%) underwent for them. This may be due to an unawareness in
heart transplantation after a mean mechanical the general medical community about this tech-
support time of 10.2 months [1]. Transplant eligi- nology or due to a perception that LVADs serve a
bility occurred due to the reversal of pulmonary palliative option.
218 P. Sen and S. Oberton

Trials comparing pulsatile-flow LVADs to improve quality of life, and obviate the need for
medical therapy alone in transplant-ineligible heart transplant altogether.
patients have demonstrated the superiority of
LVADs in terms of lengthening survival though at
the cost of increasing adverse events. The bulki- References
ness and fragility of earlier LVAD models were
addressed by the simpler design of the continuous-­ 1. Lietz K, Long JW, Kfoury AG, et al. Outcomes of left
ventricular assist device implantation as destination
flow LVADs which were less prone to adverse therapy in the post-REMATCH era: implications for
events such as device infections, pump thrombo- patient selection. Circulation. 2007;116(5):497–505.
sis, cardiac arrhythmia, and RV failure than pul- 2. Lietz K. Destination therapy: patient selection and
satile models. However, the nonphysiologic current outcomes. J Card Surg. 2010;25(4):462–71.
3. Slaughter MS, Pagani FD, Rogers JG, et al. Clinical
continuous flow was attributed to increased rates management of continuous-flow left ventricular
of non-neurologic bleeding. assist devices in advanced heart failure. J Heart Lung
The next generation of left ventricular assist sup- Transplant. 2010;29(4S):S1–39.
port systems promises more compact design, more 4. Puehler T, Ensminger S, Schoenbrodt M, et al.
Mechanical circulatory support devices as destination
physiologic flow, and fewer hemocompatibility-­ therapy—current evidence. Ann Cardiothorac Surg.
related adverse events. Unlike the standard axial- 2014;3(5):513–24.
flow HeartMate II, the new HeartMate 3, a 5. Oz MC, Godstein DJ, Pepino P, et al. Screening scale
magnetically levitated centrifugal continuous-flow predicts patients successfully receiving long-term
implantable left ventricular assist devices. Circulation.
pump with intrinsic pulsatility, and the HeartWare, 1995;92:II169–73.
a small intra-pericardial centrifugal-flow pump with 6. Galvan M, Saeed O, Immekus J, et al. An interna-
magnetic levitation and hydrodynamic design, both tional survey to assess referral thresholds for desti-
lack bearings [29, 30]. They were made with the nation therapy in non-inotrope dependent patients:
results of the CONSENSUS-DT study. J Card Fail.
idea that eliminating the bearings may reduce fric- 2014;20:492–7.
tion and risk of thrombosis associated with other 7. Cowger J, Sundareswaran K, Rogers JG, et al.
LVADs. The 6-month follow-up of patients enrolled Predicting survival in patients receiving continuous
in the MOMENTUM 3 trial comparing the out- flow left ventricular assist devices—the HeartMate II
risk score. J Am Coll Cardiol. 2013;61(3):313–21.
comes of patients who received HeartMate 3 versus 8. Rose EA, Geligins AC, Moskowitz AJ, et al.
HeartMate II for bridge-to-transplant and destina- Randomized evaluation of mechanical assistance for
tion therapy showed that HeartMate 3 patients the treatment of congestive heart failure study group:
exhibited greater survival free of hemocompatibility-­ long-term use of a left ventricular assist device for end-­
stage heart failure. N Engl J Med. 2001;345:1435–43.
related adverse events defined as any nonsurgical 9. Stevenson LW, Miller LW, Desvigne-Nickens P, et al.
bleeding, thromboembolic event, pump thrombosis, Left ventricular assist device as destination for patients
or neurological event [29, 31]. The ENDURANCE undergoing intravenous inotropic therapy: a subset
trial which compares the HeartWare HVAD against analysis from REMATCH (Randomized Evaluation
of Mechanical Assistance in Treatment of Chronic
the HeartMate II for destination therapy has proven Heart Failure). Circulation. 2004;110:975–81.
the non-inferiority of the HeartWare HVAD in the 10. Long JW, Kfoury AG, Slaughter MS, et al. Long-­
primary endpoint of survival at 2 years free from term destination therapy with the HeartMate XVE
disabling stroke or device removal for malfunction left ventricular assist device: improved outcomes
since the REMATCH study. Congest Heart Fail.
or failure [30, 32]. 2005;11(3):133–8.
Durable LVAD for destination therapy is not 11. Miller LW, Pagani FD, Russell SD, et al. Use of a
without its risk. Considering the high mortality continuous-flow device in patients awaiting heart
rate among patients who are on medical therapy transplantation. N Engl J Med. 2007;357:885–96.
12. Centers for Medicare and Medicaid Services (CMS).
alone awaiting heart transplantation, the increased Decision memo for ventricular assist devices as des-
adverse events balanced by increased survival tination therapy (CAG-00119R2). [Link]
rates associated with LVADs may be a necessary gov/medicare-coverage-database/details/nca-deci-
trade. We look forward to future directions of [Link]?NCAId=243&ver=9&NcaName=V
entricular+Assist+Devices+as+Destination+Therapy
LVAD design that may one day minimize risk,
15  Outcomes Using LVADs for Destination Therapy 219

+(2nd+Recon)&bc=BEAAAAAAEAAA&&fromdb 24. Moraska AR, Chamberlain AM, Shah ND, Vickers KS,
=true. Accessed 27 Sept 2016. Rummans TA, Dunlay SM, Spertus JA, Weston SA,
13. Slaughter MS, Rogers JG, Milano CA, et al. Advanced McNallan SM, Redfield MM, Roger VL. Depression,
heart failure treated with continuous-flow left ventric- healthcare utilization, and death in heart failure: a
ular assist device. N Engl J Med. 2009;361:2241–51. community study. Circ Heart Fail. 2013;6(3):387–94.
14. Park SJ, Tector A, Piccioni W, et al. Left ventricular 25. Snipelisky D, Stulak JM, Schettle SD, Sharma S,

assist devices as destination therapy: a new look at Kushwaha SS, Dunlay SM. Psychosocial character-
survival. J Thorac Cardiovasc Surg. 2005;129:9–17. istics and outcomes in patients with left ventricular
15. Maniar S, Kondareddy S, Topkara V. Left ventricu- assist device implanted as destination therapy. Am
lar assist device-related infections: past, present and Heart J. 2015;170(5):887–94.
future. Expert Rev Med Devices. 2011;8(5):627–34. 26. Atluri P, Goldstone AB, Kobrin DM, Cohen JE,

16. Lietz K, Long JW, Kfoury AG, et al. Impact of center MacArthur JW, Howard JL, Jessup ML, Rame JE,
volume on outcomes of left ventricular assist device Acker MA, Woo YJ. Ventricular assist device implant
implantation as destination therapy: analysis of the in the elderly is associated with increased, but respect-
Thoratec HeartMate registry, 1998 to 2005. Circ able risk: a multi-institutional study. Ann Thorac
Heart Fail. 2009;2(1):3–10. Surg. 2013;96(1):141–7.
17. Uriel N, Pak SW, Jorde UP, et al. Acquired von 27. Drakos SG, Mehra MR. Clinical myocardial recov-
Willebrand syndrome after continuous-flow mechan- ery during long-term mechanical support in advanced
ical device support contributes to a high prevalence heart failure: insights into moving the field forward.
of bleeding during long-term support and at a time J Heart Lung Transplant. 2016;35(4):413–20.
of transplantation. J Am Coll Cardiol. 2010;56: 28. Teuteberg JJ, Stewart GC, Jessup M, Kormos RL, Sun
1207–13. B, Frazier OH, Naftel DC, Stevenson LW. Implant
18. Harvey L, Holley CT, John R. Gastrointestinal
strategies change over time and impact outcomes:
bleed after left ventricular assist device implanta- insights from the INTERMACS (Interagency Registry
tion: incidence, management, and prevention. Ann for Mechanically Assisted Circulatory Support).
Cardiothorac Surg. 2014;3(5):475–9. JACC Heart Fail. 2013;1(5):369–78.
19. Park SJ, Milano CA, Tatooles AJ, et al. Outcomes in 29. Heatley G, Sood P, Goldstein D, Uriel N, Cleveland
advanced heart failure patients with left ventricular J, Middlebrook D, Mehra MR, MOMENTUM 3
assist devices for destination therapy. Circ Heart Fail. Investigators. Clinical trial design and rationale of the
2012;5:241–8. Multicenter Study of MagLev Technology in Patients
20. Baldwin JT, Robbins RC. Executive summary for the Undergoing Mechanical Circulatory Support Therapy
National Heart, Lung, and Blood Institute Working with HeartMate 3 (MOMENTUM 3) investigational
Group on next generation ventricular assist devices device exemption clinical study protocol. J Heart
for destination therapy. Semin Thorac Cardiovasc Lung Transplant. 2016;35:528–36.
Surg. 2005;17:369–71. 30. Pagani FD, Milano CA, Tatooles AJ, et al. HeartWare
21. Joynt KE, Whellan DJ, O’Connor CM. Why is
HVAD for the treatment of patients with advanced
depression bad for the failing heart? A review of the heart failure ineligible for cardiac transplantation:
mechanistic relationship between depression and results of the ENDURANCE destination therapy trial.
heart failure. J Card Fail. 2004;10:258–71. J Heart Lung Transplant. 2015;34:S9. U.S. National
22. Jiang W, Alexander J, Christopher E, Kuchibhatla M, Institutes of Health.
Gaulden LH, Cuffe MS, Blazing MA, Davenport C, 31.
Uriel N, Colombo PC, Cleveland JC, et al.
Califf RM, Krishnan RR, O’Connor CM. Relationship Hemocompatibility-related outcomes in the
of depression to increased risk of mortality and rehos- MOMENTUM 3 trial at 6 months—a random-
pitalization in patients with congestive heart failure. ized controlled study of a fully magnetically levi-
Arch Intern Med. 2001;161:1849–56. tated pump in advanced heart failure. Circulation.
23. Freedland KE, Rich MW, Skala JA, Carney RM,
2017;135(21):2003–12.
Davila-Roman VG, Jaffe AS. Prevalence of depres- 32.
Rogers JG, Pagani FD, Tatooles AJ, et al.
sion in hospitalized patients with congestive heart Intrapericardial left ventricular assist device for
failure. Psychosom Med. 2003;65:119–28. advanced heart failure. NEJM. 2017;376:451–60.
 Cardiac Regenerative Strategies
for Advanced Heart Failure 16
Vivekkumar B. Patel, Megumi Mathison,
Vivek Singh, Jianchang Yang,
and Todd K. Rosengart

iCM Induced cardiomyocyte

Abbreviations iPS Induced pluripotent stem cells
mRNA microRNA
CABG Coronary artery bypass grafting
PAD Peripheral arterial disease
CAD Coronary artery disease
SCIPIO Stem cell infusion in patients
CADUCEUS Intracoronary cardiosphere-­
with ischemic cardiomyopathy
derived cells for heart regenera-
(phase 1 trial)
tion after myocardial infarction
VEGF Vascular endothelial growth
(phase 1 trial)
factor
cTnT Cardiac troponin T
ES Embryonic stem cells
FGF Fibroblasts growth factor
GFP Green fluorescence protein Introduction
control
GMT Gata4, Mef2c, Tbx5 cardio-­ Congestive heart failure, caused by a loss of car-
differentiating factors diac function typically as a result of myocardial
ischemia or infarction, is the common end point
for advanced coronary artery disease and is the
V.B. Patel, M.D. leading cause of mortality from heart disease [1].
Texas Heart Institute at Baylor St. Luke’s Medical Current therapies for end-stage coronary artery
Center, Baylor College of Medicine, disease, including transplant or assist devices, are
Houston, TX, USA
associated with considerable morbidity, limiting
e-mail: vivekkup@[Link]
cost and/or availability, and aggregate 5-year
M. Mathison, M.D., Ph.D. • V. Singh, Ph.D.
mortality rates of 50% [2–4].
J. Yang, M.D., Ph.D.
Division of Surgical Research, Baylor College of Despite some evidence of limited endogenous
Medicine, Houston, TX, USA myocyte replication or regeneration from resi-
e-mail: [Link]@[Link]; dent stem cells, native adult cardiac muscle does
[Link]@[Link]; [Link]@[Link]
not effectively regenerate itself de novo after car-
T.K. Rosengart, M.D. (*) diomyocyte death. Strategies have therefore been
Department of Surgery, Texas Heart Institute at
devised to administer exogenous stem cells or
Baylor St. Luke’s Medical Center, Baylor College of
Medicine, Houston, TX, USA reserve cells (e.g., embryonic stem cells, mesen-
e-mail: [Link]@[Link] chymal stem cells, or skeletal myoblasts) into the

© Springer International Publishing AG 2018 221

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
222 V.B. Patel et al.

infarct zone to enhance cardiac function. Stem blast cells), (3) direct reprogramming strategies,
cell clinical trials have however been largely and (4) resident progenitor cell and cardiomyo-
­disappointing, likely due to inadequate implant cyte replicative stimulation strategies [9–13]. The
phenotypes and/or poor implant survival and remainder of this chapter will focus on the first three
engraftment into the host myocardium [5, 6]. of these.
Given that post-infarct ventricular remodeling is Recent analyses of these efforts suggest that
characterized by the replacement of cardiomyo- the poor survival of exogenous stem cell implants
cytes with fibroblasts, a new strategy utilizing in the hostile, ischemic milieu of the myocardial
genetic reprogramming to regenerate (induced) scar is in large part related to the absence of an
cardiomyocytes (iCMs) from endogenous scar adequate blood supply to nourish these relatively
fibroblasts has thus evolved as an intriguing new great metabolic needs of these cells compared to
therapeutic paradigm for treating patients with that of the resident fibroblasts and poor engraft-
congestive heart failure arising from end-stage ment of these exogenous cells in the host syncy-
coronary artery disease. tium [5, 6]. Following consequently marginal
More specifically, the seminal finding that cel- improvement, at best, in cardiac function in stem
lular reprogramming strategies could be used to cell clinical trials, current rationalization for the
produce induced pluripotent stem (iPS) cells mechanism of action of stem cell implantation
from adult somatic cells was soon followed by centers on proposed angiogenic and other para-
findings that iPS cells could be redifferentiated crine effects of these cells upon host cardiomyo-
into cells with a cardiomyocyte-like phenotype. cytes and other cells, rather than the primary
The more recent discovery that “transdifferentiat- contractile or physio-mechanical contributions of
ing” transcription factors could be administered implanted cells [14].
to reprogram adult somatic cells (fibroblasts) into In comparison, although the direct administra-
“induced cardiomyocyte” (iCM) cells suggests tion of angiogenic factors has been shown to
that in situ autologous cardiomyocyte regenera- induce myocardial vascularization in animal
tion that completely bypasses stem cell or (poten- models and some clinical trials, it is improbable
tially immunogenic and/or tumorigenic) iPS that angiogenesis therapy alone cannot improve
staging is possible [7, 8]. the function of myocardial scar and on contrac-
“Direct” reprogramming thus offers the excit- tile scar fibroblasts [9, 10, 15, 16]. Thus, as
ing opportunity to convert cardiac fibroblasts into depicted in Fig. 16.1, a rational myocardial
functional iCMs in situ and transform scar tissue regeneration strategy could incorporate prevascu-
into functional myocardium. Knowledge that has larization of myocardial scar with angiogenic
been garnered from the false starts and disproved therapy, followed by direct, in situ reprogram-
premises of many of the therapeutic angiogenesis ming of resident scar fibroblasts into induced car-
and stem cell trials conducted over the past two diomyocytes, obviating entirely the challenges of
decades will hopefully provide a solid starting exogenous cell delivery and engraftment into the
point for well-designed studies to test the poten- host myocardium.
tial clinical efficacy of this new strategy—one Cardiac cellular reprogramming describes the
that is based on sound selection of delivery vec- process by which cardio-differentiating (tran-
tors, routes of delivery, and regulation of “dos- scription) factors or the genes encoding them can
age” [9–11]. be administered to terminally differentiated cells
in order to reprogram them into cardiomyocyte-­
like cells. The scientific principles underlying
Strategies for Cardiac Regeneration this strategy were dramatically put forth in the
revolutionary work of Nobel laureate Yamanaka,
The cardiac regeneration strategies that are the most who used the transcription factors Oct4, Sox2,
extensively studied to date include: (1) angiogenic Klf4, and c-myc to reprogram terminally differ-
therapy, (2) exogenous (stem) cell implantation entiated cells into induced pluripotent stem (iPS)
(e.g., using hematopoietic, bone marrow, or myo- cells [17]. Capitalizing on this strategy, Srivastava
16  Cardiac Regenerative Strategies for Advanced Heart Failure 223

Fig. 16.1  Multimodal strategy for cardiac regeneration. Printed with permission from Baylor College of Medicine

et al. soon thereafter showed that cardiac fibro- ble for tumor angiogenesis [19]. These growth
blasts could be directly reprogrammed into factors are most commonly identified today as
induced cardiomyocytes (iCMs) using the cardio-­ the angiogenic peptides vascular endothelial
differentiating factors Gata4, Mef2c, and Tbx5 growth factor (VEGF) and fibroblast growth fac-
(GMT), which were able to bypass a pluripotent tor (FGF) [9, 10].
intermediate state [11]. Most importantly, the Not long following Folkman’s discovery,
administration of GMT into infarcted myocar- compelling preclinical data in animal models
dium has more recently been shown to yield sig- demonstrated that angiogenic protein or gene
nificant improvements in post-infarct ventricular delivery improved peripheral and myocardial
function and fibrosis in rodent models [18]. Since collateralization, as well as peripheral perfusion
this time, a variety of gene- and small molecule-­ and cardiac function [20, 21]. Since then, numer-
based strategies have been successfully tested ous trials involving over 2000 patients and utiliz-
in vitro and in vivo, as described below. ing administration of angiogenic proteins and
their genes (e.g., VEGF-A 121, 165, and 189,
FGF-1, and FGF-4) for the treatment of coronary
Angiogenesis-Based Therapies and peripheral vascular disease have yielded
mixed results [10, 22–24].
The goal of therapeutic angiogenesis is to “bio- Likely, the disappointing outcomes of angio-
logically bypass” an occluded vessel in coronary genic therapy trials were often due to inappropri-
artery disease (CAD) or peripheral arterial dis- ate or ineffective routes of administration (e.g.,
ease (PAD) by forming collateral blood vessels, intravascular vs. intramyocardial) that provided
thereby relieving ischemia. The notion of thera- inadequate tissue concentrations of angiogenic
peutic angiogenesis for coronary and peripheral agents and/or were due to inappropriate end
arterial diseases originates from the seminal work points (e.g., relatively low-resolution myocardial
by Dr. Judah Folkman in the early 1970s, who perfusion studies vs. PET or treadmill exercise
demonstrated that growth factors were responsi- testing) [9, 10, 13, 14]. These limitations
224 V.B. Patel et al.

f­requently confounded data interpretation, lead- latter also confers the additional potential risk of
ing many to dismiss the promise of angiogenic disseminating angiogenic factors throughout the
(gene) therapy. The unfortunate death of study body, potentially producing “off-target” effects.
patient Jesse Gelsinger, who received an abnor- Likewise, the application of angiogenic thera-
mally high dose of adenovirus therapy for treat- pies to peripheral vascular vs. coronary disease
ment of ornithine transcarbamylase deficiency, creates the challenge of effectively treating vas-
likewise led to a brief moratorium on all gene cular obstruction at multiple levels over a rela-
therapy in the United States, which nevertheless tively vast tissue territory (i.e., central inflow
led to a prolonged period of diminished enthusi- disease, diffuse intermediate-level vascular
asm for gene therapy trials [25, 26]. obstruction, and peripheral outflow obstruction),
Despite these challenges in this new area of compared to more “geographically” localized
drug development, retrospective analyses of the coronary disease requiring less total amounts and
angiogenic gene therapy clinical trials have distribution of drug delivery. The treatment of
yielded useful insights that offer new opportuni- coronary disease may thus represent an ideal tar-
ties for advancing this field. First, it appears clear get for angiogenic therapy. While these consider-
that the type of angiogenic vector utilized (i.e., ations may seem obvious in retrospect, the failure
protein, plasmid, or virus) is critical to achieving to consider these basic pharmacokinetic princi-
an appropriately therapeutic angiogenic factor ples often led to analytic generalizations about
“dose” sufficient to induce angiogenesis [27, 28]. negative trial outcomes to the entire field and
In this regard, angiogenic protein delivery typi- (mis)perceptions that angiogenic therapy was
cally produced disappointing results likely due to ineffective.
the relatively short half-life and the dose-limiting A number of the early angiogenesis trials were
side effect of hypotension associated with large also dismissed because they lacked placebo con-
systemic dosing. Plasmid delivery likewise seems trols and objective, clinically relevant end points,
to be ineffective due to its low transduction effi- such as electrocardiographic changes associated
ciency into cells and transient expression. In con- with exercise treadmill testing [16]. Paradoxically,
trast, adenoviral delivery of angiogenic factors while inclusion of such placebo controls was rel-
allows for a higher transduction efficiency of atively easy with (pharmacokinetically ineffec-
cells and prolonged expression of genes up to tive) intravascular trials, performing sham
weeks after administration without integration surgery to allow (more effective) direct intramyo-
into the genome. Accordingly, trials incorporat- cardial delivery in patients with end-stage heart
ing adenoviral-mediated delivery were and may disease poses an ethical dilemma [15].
prove to be more effective in inducing angiogen- Unfortunately, this led many observers to dis-
esis than are those testing other agents [10, credit the positive findings demonstrated by
29–33]. (non-placebo) intramyocardial-based trials,
The use of intracoronary vs. direct intramyo- despite positive objective data such as changes in
cardial delivery of angiogenic factors also likely ischemia levels based on ECG data. While a ran-
undermined therapeutic efficacy, as did the choice domized, placebo-controlled trial may ultimately
of appropriate tissue treatment targets [15, 28]. be necessary to conclusively verify the clinical
More specifically, direct intramyocardial delivery benefit of angiogenic therapy, the development of
generally yielded more favorable outcomes com- a uniform study protocols with objective, vali-
pared to intracoronary delivery likely because of dated end points, more efficient methods of gene
the more favorable pharmacokinetics of “drug delivery, and improved methods to determine
delivery” via this more localized approach com- total gene delivery into the target tissue may well
pared to the more diffuse, systemic delivery of offer a pathway to effectively test this new bio-
drug associated with intracoronary or intravascu- logic treatment for patients with vascular disease
lar delivery (Table 16.1) [15]. Parenthetically, the [9, 10, 15, 27, 28].
Table 16.1  Representative cardiac angiogenesis clinic trials
Study
Trial title Author, year design Sample size Follow-up Vector Delivery Agent Results
Direct myocardial injectiona
Symes et al.,b Phase I 20 90 days Plasmid Epicardial; VEGF-165 (+) Angina
1999 minithoracotomy (nitroglycerin use)
(+) Perfusion (SPECT)
(+/−) Coronary
angiography
Rosengart Phase I 16 6 months Adenovirus Epicardial; VEGF-121 (+) Angina class
et al.,c 1999 minithoracotomy (+) (+) Exercise
duration
(+) (+) Perfusion
(sestamibi)
Vale,d 2000 Phase I 13 60 days Recombinant Epicardial; VEGF-165 (+) Angina
protein minithoracotomy (+) Exercise duration
(+) Perfusion (NOGA)
Losordo et al.,e DB-RCT 19 12 weeks Plasmid Endocardial; NOGA VEGF2 (+) Angina class
2002 (+) Exercise duration
(+/−) Perfusion
(NOGA/SPECT)
16  Cardiac Regenerative Strategies for Advanced Heart Failure

Euroinject One Kastrup et al.,f DB-RCT 80 3 months Plasmid Endocardial; NOGA VEGF-165 (−) Angina class
2005 (−) Perfusion (SPECT/
NOGA)
REVASC Stewart et al.,g RCT 67 26 weeks Adenovirus Epicardial VEGF-121 (+) Angina
2006 (+) Exercise duration
NORTHERN Stewart et al.,h DB-RCT 93 6 months Plasmid Endocardial; NOGA VEGF-165 (−) Angina class
2009 (−) Exercise duration
(−) Perfusion
Intracoronary deliverya
Intracoronary Laham et al.,i Phase I 52 180 days Recombinant Intracoronary FGF-2 (+) Angina score
2000 protein (Seattle)
(+) Exercise treadmill
(+) Ischemia (MRI)

(continued)
225
Table 16.1 (continued)
226

Study
Trial title Author, year design Sample size Follow-up Vector Delivery Agent Results
Hendel et al.,j Phase I 14 60 days Recombinant Intracoronary VEGF-165 (+/−) SPECT
2000 protein
FIRST Simons et al.,k DB-RCT 337 180 days Recombinant Intracoronary FGF2 (−) Angina class
2002 protein (−) Exercise time
AGENT-2 Grines,l 2003 DB-RCT 52 8 weeks Adenovirus Intracoronary FGF4 (−) Angina class
(−) SPECT
VIVA Henry et al.,m DB-RCT 178 120 days Recombinant Intracoronary VEGF (−) Angina class
2003 protein [(+) at high dose]
(−) Exercise time
(−) SPECT
AGENT-3, Henry et al.,n DB-RCT 532 12 months Adenovirus Intracoronary FGF4 (−) Angina class
AGENT-4 2007 (−) Exercise time
+, Statistically significant change; +/−, nonsignificant improvement; −, no improvements detected
DB-RCT indicates double-blind, randomized controlled trial, CABG coronary artery bypass graft, SPECT surface photon emission computerized tomography, REVASC random-
ized evaluation of VEGF for angiogenesis, NORTHERN NOGA angiogenesis revascularization therapy: assessment by radionuclide imaging, FGF indicates fibroblast growth
factor, MRI magnetic resonance imaging, FIRST FGF initiating revascularization trial, AGENT angiogenic gene therapy trial, VIVA VEGF in ischemia for vascular
angiogenesis
Note. Adapted from Tables 1 and 2 in Rosengart TK, Fallon E, Crystal RG. Cardiac Biointerventions: Whatever Happened to Stem Cell and Gene Therapy?, Innovations: Tech
in Cardiothoracic Surgery, May/June 2012; 7(3): 173–179. Please see this reference for further information, including citations for these trials. Reprinted with permission from
Wolters Kluwer
a
Does not include combined interventions (e.g., adjunct to CABG); includes latest of multiple reports on each trial
b
Symes, J. F., Losordo, D. W., Vale, P. R., Lathi, K. G., Esakof, D. D., Mayskiy, M., & Isner, J. M. (1999). Gene therapy with vascular endothelial growth factor for inoperable
coronary artery disease. Ann Thorac Surg, 68(3), 830–836; discussion 836–837
c
Rosengart, T. K., Lee, L. Y., Patel, S. R., Kligfield, P. D., Okin, P. M., Hackett, N. R., Isom, O. W., & Crystal, R. G. (1999). Six-month assessment of a phase I trial of angiogenic
gene therapy for the treatment of coronary artery disease using direct intramyocardial administration of an adenovirus vector expressing the VEGF121 cDNA. Ann Surg, 230(4),
466–470; discussion 470–462
d
Vale, P. R., Losordo, D. W., Milliken, C. E., Maysky, M., Esakof, D. D., Symes, J. F., & Isner, J. M. (2000). Left ventricular electromechanical mapping to assess efficacy of
phVEGF(165) gene transfer for therapeutic angiogenesis in chronic myocardial ischemia. Circulation, 102(9), 965–974
e
Losordo, D. W., Vale, P. R., Hendel, R. C., Milliken, C. E., Fortuin, F. D., Cummings, N., Schatz, R. A., Asahara, T., Isner, J. M., & Kuntz, R. E. (2002). Phase 1/2 placebo-­
controlled, double-blind, dose-escalating trial of myocardial vascular endothelial growth factor 2 gene transfer by catheter delivery in patients with chronic myocardial ischemia.
Circulation, 105(17), 2012–2018
f
Kastrup, J., Jorgensen, E., Ruck, A., Tagil, K., Glogar, D., Ruzyllo, W., Botker, H. E., Dudek, D., Drvota, V., Hesse, B., Thuesen, L., Blomberg, P., Gyongyosi, M., & Sylven,
C. (2005). Direct intramyocardial plasmid vascular endothelial growth factor-A165 gene therapy in patients with stable severe angina pectoris A randomized double-blind
placebo-­controlled study: the Euroinject One trial. J Am Coll Cardiol, 45(7), 982–988
g
Stewart, D. J., Hilton, J. D., Arnold, J. M., Gregoire, J., Rivard, A., Archer, S. L., Charbonneau, F., Cohen, E., Curtis, M., Buller, C. E., Mendelsohn, F. O., Dib, N., Page, P.,
V.B. Patel et al.

Ducas, J., Plante, S., Sullivan, J., Macko, J., Rasmussen, C., Kessler, P. D., & Rasmussen, H. S. (2006). Angiogenic gene therapy in patients with nonrevascularizable ischemic
heart disease: a phase 2 randomized, controlled trial of AdVEGF(121) (AdVEGF121) versus maximum medical treatment. Gene Ther, 13(21), 1503–1511. doi:[Link]
org/10.1038/[Link].3302802
h
Stewart, D. J., Kutryk, M. J., Fitchett, D., Freeman, M., Camack, N., Su, Y., Siega, A. D., Bilodeau, L., Burton, J. R., Proulx, G., Radhakrishnan, S., & Investigators, N. T. (2009).
VEGF Gene Therapy Fails to Improve Perfusion of Ischemic Myocardium in Patients With Advanced Coronary Disease: Results of the NORTHERN Trial. Mol Ther, 17(6),
1109–1115
i
Laham, R. J., Chronos, N. A., Pike, M., Leimbach, M. E., Udelson, J. E., Pearlman, J. D., Pettigrew, R. I., Whitehouse, M. J., Yoshizawa, C., & Simons, M. (2000). Intracoronary
basic fibroblast growth factor (FGF-2) in patients with severe ischemic heart disease: results of a phase I open-label dose escalation study. J Am Coll Cardiol, 36(7),
2132–2139
j
Hendel, R. C., Henry, T. D., Rocha-Singh, K., Isner, J. M., Kereiakes, D. J., Giordano, F. J., Simons, M., & Bonow, R. O. (2000). Effect of intracoronary recombinant human
vascular endothelial growth factor on myocardial perfusion: evidence for a dose-dependent effect. Circulation, 101(2), 118–121
k
Simons, M., Annex, B. H., Laham, R. J., Kleiman, N., Henry, T., Dauerman, H., Udelson, J. E., Gervino, E. V., Pike, M., Whitehouse, M. J., Moon, T., & Chronos, N. A. (2002).
Pharmacological treatment of coronary artery disease with recombinant fibroblast growth factor-2: double-blind, randomized, controlled clinical trial. Circulation, 105(7), 788–793
l
Grines, C. L., Watkins, M. W., Mahmarian, J. J., Iskandrian, A. E., Rade, J. J., Marrott, P., Pratt, C., Kleiman, N., & Angiogene, G. T. S. G. (2003). A randomized, double-blind,
placebo-controlled trial of Ad5FGF-4 gene therapy and its effect on myocardial perfusion in patients with stable angina. J Am Coll Cardiol, 42(8), 1339–1347
m
Henry, T. D., Annex, B. H., McKendall, G. R., Azrin, M. A., Lopez, J. J., Giordano, F. J., Shah, P. K., Willerson, J. T., Benza, R. L., Berman, D. S., Gibson, C. M., Bajamonde,
A., Rundle, A. C., Fine, J., McCluskey, E. R., & Investigators, V. (2003). The VIVA trial: Vascular endothelial growth factor in Ischemia for Vascular Angiogenesis. Circulation,
107(10), 1359–1365
n
Henry, T. D., Grines, C. L., Watkins, M. W., Dib, N., Barbeau, G., Moreadith, R., Andrasfay, T., & Engler, R. L. (2007). Effects of Ad5FGF-4 in patients with angina: an analysis
of pooled data from the AGENT-3 and AGENT-4 trials. J Am Coll Cardiol, 50(11), 1038–1046
16  Cardiac Regenerative Strategies for Advanced Heart Failure
227
228 V.B. Patel et al.

Stem Cell Implantation capacity, reduced risk of immune-mediated rejec-

tion, and, of course, more ethically sound pro-
Given the low regenerative capacity of cardio- curement [38, 39]. The disadvantages of iPS cells
myocytes (<1% per year), replacement of cardio- include the cost needed to generate patient-­
myocytes with exogenous stem cells (embryonic, specific iPS cells and their immature development
bone marrow-derived, or induced pluripotent) of sarcomeric structures following implantation,
following myocardial infarction has received immature electrophysiologic properties, electrical
much attention over the past two decades. Bone ectopy, and the formation of a heterogeneous mix-
marrow-derived hematopoietic progenitor cells ture of cell types, as well as the same implant sur-
were the first to be implanted, resulting in over vival and integration challenges associated with
100 phase I-II trials including thousands of all cell implants into the hostile milieu of the
patients. Despite the promising safety profile of infarcted myocardium [14, 38]. While the implan-
these cell implants, several meta-analyses tation of iPS cells into animal infarct models has
revealed that bone marrow-derived progenitor been shown to lead to 10–15% increases in left
cell implantation therapy led only to a non-­ ventricular ejection fraction, these results, as with
clinically relevant increase (i.e., not greater than other cell implants, may thus likely not be trans-
2–4%) in left ventricular ejection fraction [7, 8, latable to clinical efficacy [40–45].
34, 35]. This negative outcome has been attrib- In the context of these mixed results, efforts
uted to the low survival of implanted cells, poor are currently underway to improve implant cell
engraftment into the native myocardium, and the survival through improved mechanical adherence
formation of immature cardiomyocyte structures strategies (e.g., using engineered tissue scaffolds
(Table  16.2) [36]. The modest improvement in and adhesive biogels) and improved vasculariza-
ejection fraction seen in these trials has more tion of the tissue bed to enhance implanted cell
recently been alternatively ascribed to the secre- survival (e.g., using angiogenic pretreatment)
tion by implanted cells of paracrine signals that [39, 46–50].
might support resident myocyte survival, neovas-
cularization, recruitment of endogenous cardiac
progenitor cells, or cardiac remodeling induced Direct Cardiac Cellular
by these implants rather than their differentiation Reprogramming Studies In Vitro
into nascent cardiomyocytes [12, 35, 37].
The use of embryonic stem (ES) cells, which Given the challenges associated with cell implan-
possess a potentially more favorable degree of tation therapies for cardiac regeneration, the
plasticity and proliferative capacity than bone or advent of direct cardiac cellular reprogramming
circulation-derived stem cells, is not ideal as a represents a significant potential advancement for
stem cell therapy for obvious ethical reasons, as this field as it capitalizes on the presence of endog-
well as the risk of immune-mediated rejection enous, quiescent cardiac fibroblasts to generate
necessitating immunosuppression and the forma- induced cardiomyocytes (iCMs), avoiding nearly
tion of heterogeneous cardiac cell subtypes (ven- all of the challenges posed by exogenous cell
tricular cardiomyocytes, atrial cardiomyocytes, implant strategies [6]. In general, this entire field
or pacemaker cells) or other aberrant tissues such seeks to capitalize upon the identification and
as teratomas [14, 38]. administration of the differentiating factors associ-
Given these concerns, induced pluripotent ated with early embryonic cardiac development.
stem (iPS) cells, which can be derived directly The first iterations of direct cardiac repro-
from the patient and reimplanted into infarcted gramming using administration of the cardio-­
myocardium, have emerged as potentially more differentiating transcription factors Gata4,
suitable candidates for cell-based regeneration. Mef2c, and Tbx5 (GMT) yielded reprogramming
The advantages of using iPS cells include their efficiencies of about 7%, as evidenced by expres-
similarity to ES cells, unlimited proliferative sion of the cardiomyocyte marker cardiac tropo-
16  Cardiac Regenerative Strategies for Advanced Heart Failure 229

Table 16.2  Summary of current cardiac regenerative strategies

Bone marrow-derived Induced pluripotent (iPS) Direct cardiac
stem cells cells reprogramming
Type of therapy Cell implantation Cell implantation Gene therapy
Survival of cells Low Low Unknown beyond
12 weeks
Characterization of cells Immature sarcomeric Immature sarcomeric Mature sarcomeric
structures structures structures
Oncogenic risk Low High None
Risk of immunorejection High if allogeneic Low (viral reprogramming Low (viral
source factors) reprogramming factors)
Improvement in ejection +2–5% +10–15% +10–25%
fractiona
Advantages Proven safe in phase I/ Ex vivo expansion of cells, Avoids cell implantation,
II trials lower immunorejection uses endogenous cardiac
risk (patient-specific cells) fibroblasts
Disadvantages Marginal improvement Cardiac subtype Human cells resistant to
in ejection fraction, heterogeneity, arrhythmia, reprogramming, no
complicated acquisition teratoma, poor long-term studies
and delivery of cells, engraftment, low cell (>12 weeks)
low cell survival, survival
arrhythmia, potential
for teratoma
Modifications/future research Tissue scaffolds or Tissue scaffolds or Non-viral vectors,
improved improved vascularization epigenetic modifications,
vascularization to (VEGF, omental flap) to simpler combinations of
improve cell survival, increase cell survival factors
new delivery techniques
Represents the difference between treated vs. untreated (control) groups for each regenerative strategy
a

nin T (cTnT), with an even smaller subset of “singlet” GMT vectors [51, 52]. Optimized repro-
contractile iCMs [11]. In order to enhance the gramming transgene dosing via differential over-
reprogramming efficiency and the maturation of expression of Mef2c in relation to Gata4 and Tbx5
iCMs, several modifications have been proposed, provided by stoichiometric rearrangement or
including the improvement of the cardio-­ fusion of the MyoD domain to Mef2c has likewise
differentiating gene cocktail and incorporation of been shown to result in a higher proportion of con-
downregulation of native fibroblast gene expres- tractile iCMs, reemphasizing the principle that the
sion (Table 16.3). correct “dosing” of genes is important [53, 54].
Consistent with the results of the angiogenesis The addition of new potent select cardio-­
trials, it was quickly established that efficient differentiating factors to the original GMT for-
delivery of reprogramming factors, not surpris- mula has also produced improvements in efficacy.
ingly, improved the efficacy of reprogramming. The addition of Hand2 [55], Nkx2.5 [56],
The use of a “triplet” GMT vector which resolved Myocardin, and/or Mesp1 [57] to GMT has, for
the challenge of target cells needed to be infected example, been shown to increase the reprogram-
by three different viruses each encoding for a sin- ming efficiency and the contractile characteris-
gle reprogramming transgene, for example, tics of iCMs [57–59]. Other reprogramming
resulted in a twofold increase in reprogramming factor cocktails have also been favorably tested
efficiency (based on % cTnT+ cells), more mature (Table 16.3).
sarcomeric structures (immunofluorescence), and “Erasing” the preexistent somatic cell signa-
a threefold improvement in post-infarct ventricular ture of target cells by downregulating native
function when compared to the use of the original fibroblast gene expression has also been shown to
230 V.B. Patel et al.

Table 16.3  In vivo and in vitro reports of direct cardiac reprogramming

Reprogramming
Cell type Reprogramming factors Vector Contractile iCM efficiency (% cTnT+) In vitro ∆EF%
Murine
CF, DF Gata4, Mef2c, Tbx5 RV, LV ● 8% ● +10
[11, 18]
CF Gata4, Mef2c, Tbx5 RV ○ 7% ● a

[51]
CF, TTF Gata4, Mef2c, Tbx5 RV ● 10% ● +21
[53, 71]
CF, TTF Gata4, Mef2c, Tbx5 LV ○ − ○
[85]
CF, TTF Gata4, Mef2c, Tbx5, RV ● 25% ● +21
Hand2 [55]
MEF, TTF Gata4, Mef2c-MyoD RV ● 21% ○
fusion, Tbx5, Hand2
[54]
CF, TTF, Gata4, Mef2c, Tbx5, RV ● 30% ○
MEF Hand2, Akt1 [86]
MEF, CF Gata4, Mef2c, Tbx5, LV ● 5%b ○
Hand2, Nkx2.5 [56]
MEF Gata4, Tbx5, LV ○ 26%b ○
Myocardin [58]
MEF Mef2c, Tbx5, LV ○ 10% ○
Myocardin [59]
MEF Gata4, Mef2c, Tbx5, LV ○ 30%b ○
Myocardin, SRF,
Mesp1, SMARCD3
[57]
CF miR-1, miR-133, LV ○ − ● c

miR-208, and
miR-499 [66, 67]
CF, MEF Gata4, Mef2c, Tbx5, LV ● 17%b ○
Hand2, Nkx2.5,
SB431542 (TGFβ
inhibitor) [62]
MEF, CF, Gata4, Mef2c, Tbx5, RV, AV ● 60% ○
TTF Hand2, miR-1,
miR-133, A-83-01
(TGFβ inhibitor) [63]
MEF, TTF Oct4, small molecules LV ● − ○
(SB431542,
CHIR99021, Parnate,
Forskolin) [68]
MEF, TTF Small molecules − ● − ○
(CHIR99021,
RepSox, Forskolin,
VPA, Parnate,
TTNPB) [69]
CF, TTF Gata4, Mef2c, Tbx5, RV ● 30% ○
Bmi1 shRNA [65]
MEF, TTF Gata4, Mef2c, Tbx5, RV ● 3% ○
FGF2, FGF10, VEGF
[74]

(continued)
16  Cardiac Regenerative Strategies for Advanced Heart Failure 231

Table 16.3 (continued)
Reprogramming
Cell type Reprogramming factors Vector Contractile iCM efficiency (% cTnT+) In vitro ∆EF%
Rat CF Gata4, Mef2c, Tbx5, LV, AV ○ 8% ● +15
VEGF [52, 73]
Other
MEF, Gata4, Mef2c, Tbx5, RV, LV ● 30% ○
Human CF miR-133, Mesp1,
Myocardin [60]
Human DF Ets2, Mesp1 [87] RV ○ − ○
Human CF, Gata4, Mef2c, Tbx5, RV ○ 12% ○
DF ESRRG, Mesp1,
Myocardin, ZFPM2
[75]
Human CF, Gata4, Tbx5, Hand2, RV ● 20% ○
DF Myocardin, miR-1,
miR-133 [61]
Human CF, Gata4, Mef2c, Tbx5, RV, LV ○ 6% ○
DF Myocardin, Mesp1
[76]
Human DF Small molecules − ● 6% ○
(CHIR99021,
A83-01, BIX01294,
AS8351, SC1,
Y27632, OAC2,
SU16F,
JNJ10198409) [70]
∆EF% represents the difference or improvement in ejection fraction between GMT-treated groups and control groups
RV retrovirus, LV lentivirus, AV adenovirus, CF cardiac fibroblasts (neonatal or adult), DF dermal fibroblasts, TTF tail
tip fibroblasts, MEF mouse embryonic fibroblasts
Reprinted with permission of Springer. Current Treatment Options in Cardiovascular Medicine, Direct Cardiac Cellular
Reprogramming for Cardiac Regeneration, (2016) 18: 58, Patel V, Mathison M, Singh VP, Yang J, Rosengart TK
a
Inagawa et al. demonstrated formation of sarcomeric structures in fibroblasts (IF for α-actinin and cTnT) following
administration of intramyocardial GMT
b
Addis et al. and Ifkovits, % cells that express Troponin T-GCaMP-GFP (calcium reporter) by immunofluorescence (IF).
Christoforou, % cells that express Tnnt2 (precursor gene for cTnT) by flow cytometry. Zhou et al. % Tnnt2 cells by IF
c
Jayawardena et al. showed reduction in fibrosis and improvement in cardiac function (echocardiography) by other
metrics such as increased fractional shortening and decreased left ventricular dimensions

favor the formation of iCMs. Muraoka et al. and structures [62, 63]. By combining multiple
Nam et al., for example, demonstrated that cardio-­differentiating factors and multiple inhibi-
microRNA-133 (miRNA-133)-mediated sup- tors of pro-fibrotic genes, Zhao et al. reported a
pression of Snai1, a key activator of pro-fibrotic reprogramming efficiency of nearly 60% (%
genes, improved reprogramming efficiency at cTnT+ cells by flow cytometry) in mouse embry-
least fivefold in murine and human fibroblasts onic fibroblasts, many of which were contractile.
[60, 61]. Similarly, Ifkovits et al. and Zhao et al. As such efforts will only increase the com-
demonstrated that small molecule inhibitors of plexity and number of constituents in reprogram-
key pro-fibrotic pathways, transforming growth ming factor cocktails, which would be difficult to
factor-β (TGF-β) or Rho-associated kinase, sub- deliver in vivo, efforts are also underway to
stantially enhanced reprogramming efficiency ­consolidate the number of factors by silencing
and quality, as evidenced by the percentage of the intrinsic epigenetic barriers that oppose
contractile cells and maturation of sarcomeric reprogramming [64]. Zhou et al. demonstrated
232 V.B. Patel et al.

that silencing Bmi1, an epigenetic regulator, the potential clinical benefit of cardiac cellular
results in derepression of cardiogenic genes reprogramming [18, 52, 55, 71]. Given these
Gata4, Nkx2.5, Isl1, Pitx2, Tbx20, and Hand2, findings and the observation, for example, that up
enhancing both the efficiency and quality of to 35% of infarct border zone cardiac fibroblasts
reprogramming when combined with GMT [65]. may be reprogrammed into iCMs, it has been
Similarly, we have shown that silencing the “anti- postulated that the in vivo microenvironment is
plasticity gene” p63 alone results in derepression more permissive to the formation of mature car-
of Gata4, Mef2c, and Tbx5 and upregulation of diomyocytes than is that provided in cell culture
cardiomyocyte-­ specific markers [Patel et al. [18]. The in vivo milieu may, for example,
2016, unpublished]. include unknown paracrine signaling pathways
While most reprogramming approaches have as well as electrical and mechanical stimuli likely
utilized integrative lentiviral or retroviral vectors, to promote the maturation of iCMs [18, 51, 52,
many have reported that non-viral, transient 55, 67, 71]. Collectively, these studies indicate
expression of reprogramming factors is sufficient that the reprogramming of cardiac fibroblasts or
for cardiac reprogramming. Jayawardena et al. other non-myocytes into iCMs can result in
described that transient transfection of fibroblasts improved post-infarct ventricular function and
with miR-1, miR-133, miR-208, and miR-499 decreased fibrosis in vivo [18, 55, 72].
reprograms cardiac fibroblasts into iCMs, as evi- In the context of these encouraging in vivo
denced by the formation of sarcomeric structures, data, we have further shown that optimization of
demonstration of spontaneous calcium fluxes, in vivo results and potentiation of the effect of
and, most importantly, improvement in post-­ reprogramming cocktails such as GMT may be
infarct ventricular function (increased fractional obtained by adjuvant prevascularization of the
shortening and reduction in ventricular dimen- ischemic, scarred myocardium (with VEGF),
sions) [66, 67]. resulting in enhanced iCM population density
Consistent with these findings, small mole- and greater improvement in post-infarct ventricu-
cules regulating Wnt signaling, TGF-β signaling, lar function (17% vs. 4% relative increase in
and various other pathways may also reprogram ejection fraction) [73]. A subsequent study by
murine and human fibroblasts into iCMs [68–70]. Yamakawa et al. confirmed that VEGF and FGF
Additionally, we have demonstrated that non-­ also activate multiple cardiac transcriptional
integrative adenovirus-mediated delivery of pathways (p38 mitogen-activated protein kinase
GMT reprograms murine cardiac fibroblasts into and phosphoinositol 3-kinase/AKT pathways) in
iCMs [Mathison et al. 2016, unpublished]. cardiac fibroblasts, which may further enable the
Therefore, cardiac cellular reprogramming is maturation of iCMs [74].
also possible using non-viral vectors or small
molecules, both of which are likely more ideal
for clinical use than the use of chronic expression  uture Directions in Cardiac
F
vectors. Reprogramming

Although direct cardiac reprogramming has

 irect Cardiac Reprogramming
D moved closer to clinical implementation, signifi-
Studies In Vivo cant barriers still need to be addressed. Most
notably, human cardiac fibroblasts are more
Despite the low in vitro efficiency of reprogram- resistant to reprogramming than are those of
ming (~10% cTnT+ cells) demonstrated in initial lower-order species and are, for example, almost
studies, in vivo studies have demonstrated about completely resistant to reprogramming with
a 10–20% improvement in ejection fraction and GMT alone [60, 61, 70, 75, 76]. This finding is
50% reduction in fibrosis in rodent infarct mod- likely due to a more complex cardiac gene regu-
els, providing the most convincing evidence for latory network and greater epigenetic constraints
16  Cardiac Regenerative Strategies for Advanced Heart Failure 233

imposed on the human species relative to murine  ardiac Regenerative Strategies

C
models [64]. In this regard, a systematic evalua- Using Endogenous Cardiomyocytes
tion of reprogramming barriers in human cells or Cardiac Stem Cells
revealed that approximately 956 genes are
opposed to reprogramming, regulating a multi- Efforts to force (normally quiescent) adult human
tude of functions including transcription, chro- cardiomyocytes back into a replicative state,
matin modifications, and various cellular mimicking the persistent replicative state of
functions [77]. lower vertebrates such as zebrafish and neonatal
While most attempts to overcome such barri- mice, represent a final potential myocardial
ers have simply included the addition of more regenerative strategy now being investigated.
cardio-differentiating factors or anti-fibrotic Cardiomyocyte cell cycle reentry has been
agents, some strategies to overcome the epigene- attempted using miRNAs, cyclins, growth factors
tic constraints have been reported, including sup- including FGF, and other small molecules [12].
pression of Bmi1 [65] or p63 [Patel et al. 2016, Adenoviral-mediated administration of cyclin A2
unpublished]. However, our understanding of has most prominently been demonstrated in
epigenetic regulation in cardiac reprogramming murine and porcine infarct models to induce car-
is very limited, and extensive investigation is still diomyocyte proliferation with improvements in
needed. Encouragingly, our recent observations cardiac function (~27% by echocardiography in
suggest that a porcine model may be an appropri- porcine models) [78–80]. While these initial
ate surrogate to investigate the resistance of reports are compelling, therapeutic use of these
human cells to reprogramming [Singh et al. 2016, factors may incur the risk of neoplastic transfor-
unpublished]. mation, which has yet to be investigated.
The long-term survival of iCMs and demon- Initiatives to isolate and induce the ex vivo
stration of sustained improvement in post-infarct expansion of cardiac stem cells, which are usu-
ventricular function beyond 12 weeks are other ally quiescent but activated following myocar-
significant considerations in the potential clinical dial injury, represent one more cardiac
applicability of this strategy. Whether GMT-­ regeneration strategy under investigation [81,
mediated reprogramming prevents or reverses 82]. In the stem cell infusion in patients with
ventricular remodeling is an intriguing additional ischemic cardiomyopathy (SCIPIO) phase 1
consideration derived from early studies. In this trial, right atrial biopsies were performed at the
regard, further investigation is needed of the time of coronary artery bypass grafting, with
intriguing observation that reprogramming fac- subsequent intracoronary infusion of these car-
tors induce an unexpected 50% reduction in post-­ diac stem cells 4 months later. This intervention
infarct fibrosis [18, 55, 67, 71, 73], far exceeding resulted in a 14% increase in left ventricular
the extent of iCM generation. The potential ejection fraction compared to baseline and a
mechanisms underlying this observation are reduction in infarct size 12 months following
largely unexplored. Finally, while most studies treatment [83]. Similarly, the intracoronary car-
have utilized immediate administration of GMT diosphere-derived cells for heart regeneration
following coronary artery ligation, reflective of after myocardial infarction (CADUCEUS)
an acute infarct model [18, 51, 55, 67, 71], an phase 1 trial utilized percutaneous endomyocar-
increased focus on delayed administration stud- dial biopsies to generate cardiac stem cells for
ies reflective of the clinically relevant scenario of infusion and established safety data for further
established scar is still needed [52, 73]. trials [84]. Obstacles to clinical implementation
234 V.B. Patel et al.

of these techniques include the costly expansion 3. Baran DA, Jaiswal A. Management of the ACC/AHA
Stage D patient: mechanical circulatory support.
of cardiac stem cells and similar difficulties to
Cardiol Clin. 2014;32(1):113–24.
that seen with stem cell implantation, as well as 4. Kittleson MM, Kobashigawa JA. Management of the
challenges in differentiation into other cardiac ACC/AHA Stage D patient: cardiac transplantation.
subtypes, potential for arrhythmia, and electro- Cardiol Clin. 2014;32(1):95–112.
5. Fisher SA, Zhang H, Doree C, Mathur A, Martin-­
mechanical integration.
Rendon E. Stem cell treatment for acute myo-
cardial infarction. Cochrane Database Syst Rev.
2015;9:CD006536.
Conclusions 6. Nowbar AN, Mielewczik M, Karavassilis M, Dehbi
HM, Shun-Shin MJ, Jones S, et al. Discrepancies in
autologous bone marrow stem cell trials and enhance-
Initial disappointments in the field of cardiac ment of ejection fraction (DAMASCENE): weighted
regeneration have resulted in the formation of a regression and meta-analysis. BMJ. 2014;348:g2688.
more rigorous, tempered research methodology, 7. Xin M, Olson EN, Bassel-Duby R. Mending bro-
ken hearts: cardiac development as a basis for adult
which has significantly influenced novel regen-
heart regeneration and repair. Nat Rev Mol Cell Biol.
erative strategies such as direct cardiac cellular 2013;8:529–41.
reprogramming. In comparison to cell-based 8. Sadahiro T, Yamanaka S, Ieda M. Direct car-
interventions for cardiac regeneration, direct car- diac reprogramming: progress and challenges in
basic biology and clinical applications. Circ Res.
diac cellular reprogramming boasts a significant
2015;116(8):1378–91.
advantage by utilizing endogenous cardiac fibro- 9. Gupta R, Tongers J, Losordo DW. Human studies of
blasts, thereby offering a higher chance of nascent angiogenic gene therapy. Circ Res. 2009;8:724–36.
cardiomyocyte survival, maturation, electrophys- 10. Zachary I, Morgan RD. Therapeutic angiogenesis
for cardiovascular disease: biological context, chal-
iological integration into the host syncytium, and
lenges, prospects. Heart (British Cardiac Society).
durable long-term results. Optimal cardiac cellu- 2010;97(3):181–9.
lar reprogramming strategies will likely utilize a 11. Ieda M, Fu JD, Delgado-Olguin P, Vedantham V,
multimodal approach involving upregulation of Hayashi Y, Bruneau BG, et al. Direct reprogram-
ming of fibroblasts into functional cardiomyocytes by
cardio-differentiating factors, inhibition of pro-­
defined factors. Cell. 2010;3:375–86.
fibrotic gene expression, and induction of angio- 12. Sharma A, Zhang Y, Wu SM. Harnessing the induc-
genic factors as exemplified in our iteration tion of cardiomyocyte proliferation for cardiac regen-
shown in Fig. 16.1. erative medicine. Curr Treat Options Cardiovasc Med.
2015;17(10):404.
13. Oh H, Ito H, Sano S. Challenges to success in heart
Acknowledgments This study was funded by the failure: cardiac cell therapies in patients with heart dis-
National Heart, Lung, and Blood Institute eases. J Cardiol. 2016;pii:S0914-5087(16):30059–4.
(1R01HL121294-01A1 [TR]). 14. Lalit PA, Hei DJ, Raval AN, Kamp TJ. Induced plu-
ripotent stem cells for post-myocardial infarction
Disclosure  The authors declare that they have repair: remarkable opportunities and challenges. Circ
Res. 2014;114(8):1328–45.
no conflict of interest.
15. Rosengart TK, Fallon E, Crystal RG. Cardiac bioin-
terventions: whatever happened to stem cell and gene
therapy? Innovations. 2012;7(3):173–9.
References 16. Simons M, Bonow RO, Chronos NA, Cohen DJ,

Giordano FJ, Hammond HK, et al. Clinical tri-
als in coronary angiogenesis: issues, problems,
1. Galli A, Lombardi F. Postinfarct left ventricular remod-
consensus: an expert panel summary. Circulation.
elling: a prevailing cause of heart failure. Cardiol Res
2000;102(11):E73–86.
Pract. 2016:2579832.
17. Takahashi K, Yamanaka S. Induction of pluripotent
2. Yancy CW, Jessup M, Bozkurt B, Butler J, Casey
stem cells from mouse embryonic and adult fibroblast
DE, Drazner MH, et al. 2013 ACCF/AHA guide-
cultures by defined factors. Cell. 2006;126(4):663–76.
line for the management of heart failure: execu-
18. Qian L, Huang Y, Spencer CI, Foley A, Vedantham V,
tive summary: a report of the American College of
Liu L, Conway SJ, Fu JD, Srivastava D. In vivo repro-
Cardiology Foundation/American Heart Association
gramming of murine cardiac fibroblasts into induced
Task Force on practice guidelines. Circulation.
cardiomyocytes. Nature. 2012;485(7400):593–8.
2013;128(16):1810–52.
16  Cardiac Regenerative Strategies for Advanced Heart Failure 235

19. Folkman J. Tumor angiogenesis: therapeutic implica- treatment of chronic myocardial ischemia: phase
tions. N Engl J Med. 1971;285(21):1182–6. II results of the Kuopio Angiogenesis Trial (KAT).
20. Hammond HK, McKirnan MD. Angiogenic gene
Circulation. 2003;107(21):2677–83.
therapy for heart disease: a review of animal studies 32. Makinen K, Manninen H, Hedman M, Matsi P,

and clinical trials. Cardiovasc Res. 2001;49(3):561–7. Mussalo H, Alhava E, et al. Increased vascularity
21. Mack CA, Patel SR, Schwarz EA, Zanzonico P, Hahn detected by digital subtraction angiography after
RT, Ilercil A, et al. Biologic bypass with the use of VEGF gene transfer to human lower limb artery:
adenovirus-mediated gene transfer of the comple- a randomized, placebo-controlled, double-blinded
mentary deoxyribonucleic acid for vascular endothe- phase II study. Mol Ther. 2002;6(1):127–33.
lial growth factor 121 improves myocardial perfusion 33. Kapur NK, Rade JJ. Fibroblast growth factor 4 gene
and function in the ischemic porcine heart. J Thorac therapy for chronic ischemic heart disease. Trends
Cardiovasc Surg. 1998;115(1):168–76. Cardiovasc Med. 2008;18(4):133–41.
22. Isner JM, Pieczek A, Schainfeld R, Blair R, Haley L, 34. de Jong R, Houtgraaf JH, Samiei S, Boersma E,

Asahara T, et al. Clinical evidence of angiogenesis Duckers HJ. Intracoronary stem cell infusion after
after arterial gene transfer of phVEGF165 in patient acute myocardial infarction: a meta-analysis and
with ischaemic limb. Lancet. 1996;348(9024):370–4. update on clinical trials. Circ Cardiovasc Interv.
23. Schumacher B, Pecher P, von Specht BU, Stegmann 2014;7(2):156–67.
T. Induction of neoangiogenesis in ischemic myocar- 35. Abdel-Latif A, Bolli R, Tleyjeh IM, Montori VM, Perin
dium by human growth factors: first clinical results of EC, Hornung CA, et al. Adult bone marrow-derived
a new treatment of coronary heart disease. Circulation. cells for cardiac repair: a systematic review and meta-­
1998;97(7):645–50. analysis. Arch Intern Med. 2007;167(10):989–97.
24. Losordo DW, Vale PR, Symes JF, Dunnington CH, 36. Batty JA, Lima JA, Kunadian V. Direct cellular repro-
Esakof DD, Maysky M, et al. Gene therapy for gramming for cardiac repair and regeneration. Eur
myocardial angiogenesis: initial clinical results J Heart Fail. 2015;18(2):145–56.
with direct myocardial injection of phVEGF165 as 37. Segers VF, Lee RT. Stem-cell therapy for cardiac dis-
sole therapy for myocardial ischemia. Circulation. ease. Nature. 2008;451(7181):937–42.
1998;98(25):2800–4. 38. Barad L, Schick R, Zeevi-Levin N, Itskovitz-Eldor
25. Wilson JM. Lessons learned from the gene therapy J, Binah O. Human embryonic stem cells vs human
trial for ornithine transcarbamylase deficiency. Mol induced pluripotent stem cells for cardiac repair. Can
Genet Metab. 2009;96(4):151–7. J Cardiol. 2014;11:1279–87.
26. Yarborough M, Sharp RR. Public trust and research 39. Gupta MK, Illich DJ, Gaarz A, Matzkies M, Nguemo
a decade later: what have we learned since Jesse F, Pfannkuche K, et al. Global transcriptional pro-
Gelsinger’s death? Mol Genet Metab. 2009;97(1):4–5. files of beating clusters derived from human induced
27. Yla-Herttuala S, Alitalo K. Gene transfer as a tool pluripotent stem cells and embryonic stem cells are
to induce therapeutic vascular growth. Nat Med. highly similar. BMC Dev Biol. 2010;10:98.
2003;9(6):694–701. 40. Nelson TJ, Martinez-Fernandez A, Yamada S, Perez-­
28. Rincon MY, VandenDriessche T, Chuah MK. Gene Terzic C, Ikeda Y, Terzic A. Repair of acute myocardial
therapy for cardiovascular disease: advances in vec- infarction by human stemness factors induced plurip-
tor development, targeting, and delivery for clinical otent stem cells. Circulation. 2009;120(5):408–16.
translation. Cardiovasc Res. 2015;108(1):4–20. 41. Singla DK, Long X, Glass C, Singla RD, Yan

29. Stewart DJ, Hilton JD, Arnold JM, Gregoire J, Rivard B. Induced pluripotent stem (iPS) cells repair and
A, Archer SL, et al. Angiogenic gene therapy in regenerate infarcted myocardium. Mol Pharm.
patients with nonrevascularizable ischemic heart 2011;8(5):1573–81.
disease: a phase 2 randomized, controlled trial of 42. Pasha Z, Haider HK h, Ashraf M. Efficient non-viral
AdVEGF(121) (AdVEGF121) versus maximum reprogramming of myoblasts to stemness with a
medical treatment. Gene Ther. 2006;13(21):1503–11. single small molecule to generate cardiac progenitor
30. Rajagopalan S, Mohler ER, Lederman RJ, Mendelsohn cells. PLoS One. 2011;6(8):e23667.
FO, Saucedo JF, Goldman CK, et al. Regional angio- 43. Ahmed RP, Haider HK, Buccini S, Li L, Jiang S,
genesis with vascular endothelial growth factor in Ashraf M. Reprogramming of skeletal myoblasts
peripheral arterial disease: a phase II randomized, for induction of pluripotency for tumor-free car-
double-blind, controlled study of adenoviral delivery diomyogenesis in the infarcted heart. Circ Res.
of vascular endothelial growth factor 121 in patients 2011;109(1):60–70.
with disabling intermittent claudication. Circulation. 44. Xiong Q, Ye L, Zhang P, Lepley M, Tian J, Li J, et al.
2003;108(16):1933–8. Functional consequences of human induced pluripo-
31. Hedman M, Hartikainen J, Syvanne M, Stjernvall
tent stem cell therapy: myocardial ATP turnover rate
J, Hedman A, Kivela A, et al. Safety and feasibility in the in vivo swine heart with postinfarction remodel-
of catheter-based local intracoronary vascular endo- ing. Circulation. 2013;127(9):997–991008.
thelial growth factor gene transfer in the prevention 45. Kawamura M, Miyagawa S, Miki K, Saito A,

of postangioplasty and in-stent restenosis and in the Fukushima S, Higuchi T, et al. Feasibility, safety,
236 V.B. Patel et al.

and therapeutic efficacy of human induced pluripo- cardio-inducing effect of GATA4, TBX5, and MEF2C
tent stem cell-derived cardiomyocyte sheets in a por- during direct cellular reprogramming. PLoS One.
cine ischemic cardiomyopathy model. Circulation. 2013;8(5):e63577.
2012;126(11 Suppl 1):S29–37. 58. Zhou L, Liu Y, Lu L, Lu X, Dixon RA. Cardiac gene
46. Kawamura M, Miyagawa S, Fukushima S, Saito A, activation analysis in mammalian non-myoblasic
Miki K, Ito E, et al. Enhanced survival of transplanted cells by Nkx2-5, Tbx5, Gata4 and Myocd. PLoS One.
human induced pluripotent stem cell-derived cardio- 2012;7(10):e48028.
myocytes by the combination of cell sheets with the 59. Protze S, Khattak S, Poulet C, Lindemann D, Tanaka
pedicled omental flap technique in a porcine heart. EM, Ravens U. A new approach to transcription
Circulation. 2013;128(11 Suppl 1):S87–94. factor screening for reprogramming of fibroblasts
47. Yau TM, Kim C, Ng D, Li G, Zhang Y, Weisel RD, to cardiomyocyte-like cells. J Mol Cell Cardiol.
et al. Increased transplanted cell survival with cell 2012;53(3):323–32.
based angiogenic gene therapy. Ann Thorac Surg. 60. Muraoka N, Yamakawa H, Miyamoto K, Sadahiro
2005;80(5):1779–86. T, Umei T, Isomi M, et al. MiR-133 promotes
48. Retuerto MA, Beckmann JT, Carbray J, Patejunas G, cardiac reprogramming by directly repressing
Sarateanu S, Kane BJ, et al. Angiogenic pre-treatment Snai1 and silencing fibroblast signatures. EMBO
enhances myocardial function following cellular J. 2014;33(14):1565–81.
cardiomyoplasty with skeletal myoblasts. J Thorac 61. Nam YJ, Song K, Luo X, Daniel E, Lambeth K, West
Cardiovasc Surg. 2007;133(2):478–84. K, et al. Reprogramming of human fibroblasts toward
49. Ahmed RP, Ashraf M, Buccini S, Shujia J, Haider a cardiac fate. PNAS. 2013;110(14):5588–93.
HK h. Cardiac tumorigenic potential of induced plu- 62.
Ifkovits JL, Addis RC, Epstein JA, Gearhart
ripotent stem cells in an immunocompetent host with JD. Inhibition of TGF-beta signaling increases direct
myocardial infarction. Regen Med. 2011;6(2):171–8. conversion of fibroblasts to induced cardiomyocytes.
50. Zhang Y, Wang D, Chen M, Yang B, Zhang F, Cao PLoS One. 2014;9(2):e89678.
K. Intramyocardial transplantation of undifferentiated 63. Zhao Y, Londono P, Cao Y, Sharpe EJ, Proenza C,
rat induced pluripotent stem cells causes tumorigen- O’Rourke R, et al. High-efficiency reprogramming
esis in the heart. PLoS One. 2011;6(4):e19012. of fibroblasts into cardiomyocytes requires sup-
51. Inagawa K, Miyamoto K, Yamakawa H, Muraoka N, pression of pro-fibrotic signalling. Nat Commun.
Sadahiro T, Umei T, et al. Induction of cardiomyocyte-­ 2015;10(6):8243.
like cells in infarct hearts by gene transfer of Gata4, 64. Ebrahimi B. Reprogramming barriers and enhanc-
Mef2c, and Tbx5. Circ Res. 2012;111(9):1147–56. ers: strategies to enhance the efficiency and kinet-
52. Mathison M, Singh VP, Gersch RP, Ramirez MO, ics of induced pluripotency. Cell Regen (Lond).
Cooney A, Kaminsky SM, et al. “Triplet” polycis- 2015;4:10.
tronic vectors encoding Gata4, Mef2c, and Tbx5 65. Zhou Y, Wang L, Vaseghi HR, Liu Z, Lu R,

enhances postinfarct ventricular functional improve- Alimohamadi S, et al. Bmi1 is a key epigenetic bar-
ment compared with singlet vectors. J Thorac rier to direct cardiac reprogramming. Cell Stem Cell.
Cardiovasc Surg. 2014;148(4):1656–1664.e2. 2016;3:382–95.
53. Wang L, Liu Z, Yin C, Asfour H, Chen O, Li Y, et al. 66.
Jayawardena TM, Egemnazarov B, Finch EA,
Stoichiometry of Gata4, Mef2c, and Tbx5 influences Zhang L, Payne JA, Pandya K, et al. MicroRNA-­
the efficiency and quality of induced cardiac myocyte mediated in vitro and in vivo direct reprogramming
reprogramming. Circ Res. 2014;116(2):237–44. of cardiac fibroblasts to cardiomyocytes. Circ Res.
54. Hirai H, Katoku-Kikyo N, Keirstead SA, Kikyo
2012;110(11):1465–73.
N. Accelerated direct reprogramming of fibroblasts 67. Jayawardena TM, Finch EA, Zhang L, Zhang H,

into cardiomyocyte-like cells with the MyoD transacti- Hodgkinson CP, Pratt RE, et al. MicroRNA induced
vation domain. Cardiovasc Res. 2013;100(1):105–13. cardiac reprogramming in vivo: evidence for mature
55. Song K, Nam YJ, Luo X, Qi X, Tan W, Huang
cardiac myocytes and improved cardiac function. Circ
GN, et al. Heart repair by reprogramming non-­ Res. 2015;116(3):418–24.
myocytes with cardiac transcription factors. Nature. 68. Wang H, Cao N, Spencer CI, Nie B, Ma T, Xu T, et al.
2012;485(7400):599–604. Small molecules enable cardiac reprogramming of
56. Addis RC, Ifkovits JL, Pinto F, Kellam LD, Esteso P, mouse fibroblasts with a single factor, Oct4. Cell Rep.
Rentschler S, et al. Optimization of direct fibroblast 2014;6(5):951–60.
reprogramming to cardiomyocytes using calcium 69. Fu Y, Huang C, Xu X, Gu H, Ye Y, Jiang C, et al.
activity as a functional measure of success. J Mol Cell Direct reprogramming of mouse fibroblasts into
Cardiol. 2013;60:97–106. cardiomyocytes with chemical cocktails. Cell Res.
57. Christoforou N, Chellappan M, Adler AF, Kirkton 2015;25(9):1013–24.
RD, Wu T, Addis RC, et al. Transcription factors 70. Cao N, Huang Y, Zheng J, Spencer CI, Zhang Y, Fu
MYOCD, SRF, Mesp1 and SMARCD3 enhance the JD, et al. Conversion of human fibroblasts into func-
16  Cardiac Regenerative Strategies for Advanced Heart Failure 237

tional cardiomyocytes by small molecules. Science. tion and enhances cardiac function in ischemic heart
2016;352(6290):1216–20. failure. Circulation. 2006;114(1 Suppl):I206–13.
71. Ma H, Wang L, Yin C, Liu J, Qian L. In vivo cardiac 80. Shapiro SD, Ranjan AK, Kawase Y, Cheng RK, Kara
reprogramming using an optimal single polycistronic RJ, Bhattacharya R, Guzman-Martinez G, Sanz J,
construct. Cardiovasc Res. 2015;108(2):217–9. Garcia MJ, Chaudhry HW. Cyclin A2 induces car-
72. Fu JD, Srivastava D. Direct reprogramming of fibro- diac regeneration after myocardial infarction through
blasts into cardiomyocytes for cardiac regenerative cytokinesis of adult cardiomyocytes. Sci Transl Med.
medicine. Circ J. 2015;79(2):245–54. 2014;6(224):224–7.
73. Mathison M, Gersch RP, Nasser A, Lilo S, Korman 81. Le T, Chong J. Cardiac progenitor cells for heart

M, Fourman M, et al. In vivo cardiac cellular repro- repair. Cell Death Discov. 2016;2:16052.
gramming efficacy is enhanced by angiogenic pre- 82. Faiella W, Atoui R. Therapeutic use of stem cells
conditioning of the infarcted myocardium with for cardiovascular disease. Clin Transl Med.
vascular endothelial growth factor. J Am Heart Assoc. 2016;5(1):34.
2012;1(6):e005652. 83. Chugh AR, Beache GM, Loughran JH, Mewton N,
74. Yamakawa H, Muraoka N, Miyamoto K, Sadahiro T, Elmore JB, Kajstura J, et al. Administration of cardiac
Isomi M, Haginiwa S, et al. Fibroblast growth factors stem cells in patients with ischemic cardiomyopathy:
and vascular endothelial growth factor promote car- the SCIPIO trial: surgical aspects and interim analysis
diac reprogramming under defined conditions. Stem of myocardial function and viability by magnetic res-
Cell Reports. 2015;5(6):1128–42. onance. Circulation. 2012;126(11 Suppl 1):S54–64.
75. Fu JD, Stone NR, Liu L, Spencer CI, Qian L, Hayashi 84. Makkar RR, Smith RR, Cheng K, Malliaras K,

Y, et al. Direct reprogramming of human fibroblasts Thomson LE, Berman D, et al. Intracoronary
toward a cardiomyocyte-like state. Stem Cell Reports. cardiosphere-derived cells for heart regenera-
2013;1(3):235–47. tion after myocardial infarction (CADUCEUS):
76. Wada R, Muraoka N, Inagawa K, Yamakawa H,
a prospective, randomised phase 1 trial. Lancet.
Miyamoto K, Sadahiro T, et al. Induction of human 2012;379(9819):895–904.
cardiomyocyte-like cells from fibroblasts by defined 85. Chen JX, Krane M, Deutsch MA, Wang L, Rav-Acha
factors. PNAS. 2013;110(31):12667–72. M, Gregoire S, et al. Inefficient reprogramming of
77. Qin H, Diaz A, Blouin L, Lebbink RJ, Patena W, fibroblasts into cardiomyocytes using Gata4, Mef2c,
Tanbun P, et al. Systematic identification of barriers to and Tbx5. Circ Res. 2012;111(1):50–5.
human iPSC generation. Cell. 2014;158(2):449–61. 86. Zhou H, Dickson ME, Kim MS, Bassel-Duby R,

78. Chaudhry HW, Dashoush NH, Tang H, Zhang L,
Olson EN. Akt1/protein kinase B enhances transcrip-
Wang X, Wu EX, Wolgemuth DJ. Cyclin A2 medi- tional reprogramming of fibroblasts to functional car-
ates cardiomyocyte mitosis in the postmitotic myocar- diomyocytes. PNAS. 2015;112(38):11864–9.
dium. J Biol Chem. 2004;279(34):35858–66. 87. Islas JF, Liu Y, Weng KC, Robertson MJ, Zhang S,
79. Woo YJ, Panlilio CM, Cheng RK, Liao GP, Atluri Prejusa A, et al. Transcription factors ETS2 and MESP1
P, Hsu VM, Cohen JE, Chaudhry HW. Therapeutic transdifferentiate human dermal fibroblasts into car-
delivery of cyclin A2 induces myocardial regenera- diac progenitors. PNAS. 2012;109(32):13016–21.
 Long-Term Complications
of Ventricular Assist Devices 17
George V. Letsou

Heart failure is a common affliction in developed tancy of less than 2 years on optimal medical
countries. Approximately 2% of all adults in therapy were randomly assigned to undergo either
industrialized countries have heart failure, and maximal medical management or surgical implan-
the prevalence rises to 6–10% in adults older than tation with a pulsatile-flow HeartMate Vented
65 [1]. Of the 23 million people who have heart Electric (XVE) left ventricular assist device
failure worldwide, approximately 5.8 million live (LVAD). The 1-year survival rate was 52% in the
in the United States [2]. The risk of death is surgical group and 25% in the medical group. The
approximately 35% in the first year after diagno- 2-year survival rates were 23% and 8%, respec-
sis and then falls to approximately 10% each year tively. However, patients with a surgically
after that [3]. Orthotopic heart transplantation is implanted LVAD had a nearly twofold increase in
the treatment of choice for end-stage heart fail- the rate of long-term adverse events, including
ure, but a limited donor pool restricts the number infection, hemorrhage, and device malfunction
of cardiac transplantations that can be performed [6]. The survival advantage associated with LVAD
worldwide to approximately 4000 per year [4]. support was impressive, but improving LVAD reli-
Therefore, mechanical circulatory support is an ability and durability was clearly necessary.
area of intense interest. In a subsequent trial in 2009, patients with an
The first successful implantation of a mechani- implantable pulsatile-flow device were compared
cal cardiac assist device was performed in 1963 in with patients who had a continuous-flow device.
Houston [5] several years before the first success- Two hundred patients with a left ventricular ejec-
ful heart transplant. Progress with different tion fraction <25%, a peak oxygen consumption
mechanical cardiac assist devices proceeded inter- <14 cc/kg/min, New York Heart Association
mittently until the influential Randomized (NYHA) class IIIb or IV symptoms, or the need
Evaluation of Mechanical Assistance for the for inotropic therapy or intra-aortic balloon pump
Treatment of Congestive Heart Failure counterpulsation were randomly assigned to
(REMATCH) trial in 2001. In that trial, 129 undergo implantation with either the first-­
patients with severe heart failure and a life expec- generation pulsatile HeartMate XVE or the newer
continuous-flow HeartMate II device. Actuarial
survival was significantly greater in the
G.V. Letsou, M.D. (*) continuous-­flow HeartMate II group than in the
Texas Heart Institute, Baylor St. Luke’s Medical
pulsatile-flow HeartMate XVE group (68% vs.
Center, Baylor College of Medicine,
Houston, TX, USA 55%, respectively, at 1 year and 58% vs. 24%,
e-mail: [Link]@[Link] respectively, at 2 years) (Fig. 17.1). The adverse

© Springer International Publishing AG 2018 239

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
240 G.V. Letsou

According to the Interagency Registry for

Mechanically Assisted Circulatory Support
(INTERMACS) database, between 2012 and
2014, the number of continuous-flow pumps
implanted as a bridge to transplantation increased
from 404 to 734, and the number implanted as des-
tination therapy increased from 983 to 1108 [8].
Rates of mechanical failure markedly decreased
in the continuous-flow era. In the pivotal study by
Slaughter et al. [7] in 2009 that compared pulsa-
tile-flow and continuous-flow devices, pump
replacement for malfunction (including events
such as a driveline fracture) was necessary for 9%
of implanted continuous-­flow pumps and 34% of
implanted pulsatile-flow pumps. There were no
primary pump or bearing failures in patients with a
continuous-flow LVAD. However, pump thrombo-
sis occurred in 4% of patients with continuous-
flow pumps but in none of the patients with
pulsatile-flow devices [7].
Fig. 17.1  Estimated actuarial survival (continuous-flow
The findings of other studies have shown the
LVAD therapy vs. pulsatile-flow LVAD therapy). Modified
with permission from Hindawi Publishing Corporation [33] reliability of continuous-flow devices and free-
dom from replacement. In one retrospective
study [9], replacement of the device for failure or
event rate was significantly lower in the thrombosis was required for only 3.8% of
continuous-­ flow group (Fig. 17.2) [7]. patients. However, studies of earlier continuous-­
Continuous-flow pumps offered the greater dura- flow devices, such as the Jarvik 2000 and the
bility that investigators had been searching for. DuraHeart, have documented device replacement
Mechanical cardiac assist devices have become rates as high as 14% [10]. In the latter series,
significantly more reliable, but still have prob- thrombosis was the most common indication for
lems. The most common complications of long- replacement (66%), followed by driveline infec-
term VAD implantation include pump thrombosis, tion (10%), and other problems (22%).
mechanical failure necessitating device replace- When VAD replacement is required, outcomes
ment, stroke, LVAD-related infection, sepsis, are not compromised. Of 469 patients who under-
bleeding requiring blood transfusion, cardiac went 546 continuous-flow LVAD implantations
arrhythmia, renal failure, and aortic valve insuffi- from December 1999 to December 2013, 14%
ciency. Each of these issues will be addressed in required the exchange of one continuous-flow
the following sections. Complications after the LVAD for another [10]. Survival was not signifi-
implantation of continuous-­flow devices will be cantly different between the exchange and non-­
emphasized because those devices accounted for exchange groups (Fig. 17.3).
more than 90% of pumps implanted in 2015. In 2013, two separate reports documented an
increase in the incidence of HeartMate II pump
thrombosis that seemed to have begun in 2011
Pump Replacement and Thrombosis [11, 12]. These reports estimated the risk of pump
thrombosis at 6 months to be 6–12%, a substan-
The greater durability of continuous-flow LVADs tially greater risk than was previously reported.
led to an increase in the use of device implantation Another report indicated that the HeartWare
to treat heart failure. In the current era, more device may also have a higher-than-expected
LVADs are implanted than hearts are transplanted. pump thrombosis rate of 8% [13]. Many of the
17  Long-Term Complications of Ventricular Assist Devices 241

Fig. 17.2 Comparison
of adverse event rates of
continuous-flow LVADs
and pulsatile-flow
LVADs. *Significant
difference (p < 0.05)
between continuous-­
flow adverse event rate
and pulsatile-flow
adverse event rate.
Modified with
permission from
Hindawi Publishing
Corporation [33]

Fig. 17.3 Kaplan-Meier
analysis of survival
(VAD exchange vs. no
VAD exchange).
Survival from time of
primary VAD
implantation to time of
most recent follow-up.
Modified with
permission from
Elsevier [10]
242 G.V. Letsou

events occurred within 6 months of implantation. possible thrombosis. In addition, nonoperative

Although these rising rates of thrombosis were management with anticoagulant and thrombo-
not initially associated with a decrease in sur- lytic therapy has been introduced, with success
vival, they were associated with substantial mor- rates approaching 80% in some series with cen-
bidity, the need for pump exchange, and increased trifugal continuous-flow pumps [16].
healthcare costs. Nonoperative therapeutic strategies have included
Suggested explanations for the increase in the use of heparin, direct thrombin inhibitors, and
pump thrombosis rates included mechanical platelet glycoprotein IIb/IIIa inhibitors, as well as
defects in the devices and suboptimal inflow local and systemic thrombolytic therapy with tis-
cannula geometry, but these theories were not sue plasminogen activator (tPa).
independently supported. Other possible con- Compared with pulsatile-flow pumps,
tributing factors included clinical management continuous-­flow pumps offer greater durability
issues, such as the use of lower levels of antico- and have been a clear step forward in improving
agulation to minimize gastrointestinal (GI) LVAD outcomes. Although mechanical failure
bleeding and the reduced use of heparin bridging is much less common in patients with
after implantation to minimize postoperative continuous-­ flow pumps, pump thrombosis
bleeding. The lowering of pump speeds to facili- occurs in 4–10% of these patients, posing sub-
tate aortic valve opening and minimize aortic stantial challenges. Appropriate attention to
valve commissural fusion may have also been a anticoagulation protocols is important. In many
possible contributing factor. cases, monitoring LDH levels and early inter-
After these reports, management strategies vention (beginning with augmented anticoagu-
were changed to emphasize heparin bridging lation and thrombolytic therapy) can successfully
after device implantation and to increase post- prevent pump thrombosis. If pump replacement
operative anticoagulation. Despite the increased is necessary, acceptable results and survival can
awareness of the problem, actuarial freedom be anticipated.
from pump thrombosis decreased progressively
from 2009 to 2013. Freedom from pump throm-
bosis at 6 months fell from 98% in 2010 to 92% Stroke and Thromboembolic Events
in 2013 [14]. In 2014, freedom from pump
thrombosis improved to a level approaching Cerebrovascular events are a substantial cause of
that of 2011. Subsequent analyses of data pro- morbidity and mortality after implantation of
vided by the INTERMACS database suggested either a continuous-flow or a pulsatile-flow
that pump thrombosis rates remain elevated VAD. In 2009, a randomized trial in which
when compared to those reported before 2011 continuous-­flow and pulsatile-flow LVADs were
[15]. In this INTERMACS analysis, the most compared showed that the incidence of stroke
important predictors of pump thrombosis and was 17% in the continuous-flow group and 14%
pump exchange were age at the time of implan- in the pulsatile-flow group. However, the number
tation and body mass index. Patients younger of strokes per patient year was only 0.13 in the
than age 72 and those with a body mass index continuous-flow group and 0.22 in the pulsatile-­
greater than 25 kg/m2 were at increased risk. In flow group. This incidence of stroke in the
this analysis, pump thrombosis was associated continuous-­ flow group was similar to that in
with an elevated mortality risk of 18% at patients with end-stage heart failure who do not
1 month after the first thrombosis and 37% at have mechanical support [7].
1 year after the first thrombosis. Other studies have documented similar stroke
To prevent pump thrombosis, methods of early rates after continuous-flow LVAD implantation.
detection and intervention have been adopted, A review of 150 patients who underwent
including the monitoring of lactate dehydroge- continuous-­flow LVAD implantation revealed a
nase (LDH) levels as a measure of hemolysis and stroke rate of 18% [17]. The anticoagulation
17  Long-Term Complications of Ventricular Assist Devices 243

Fig. 17.4 Freedom
from stroke (clamped or
arrested vs. unclamped).
Modified with
permission from Wolters
Kluwer [18]

protocol in this study included aspirin (81 mg/ were not risk factors for stroke in this study, but
day) and warfarin (target international normal- prior cardiac surgery and infection were associ-
ized ratio [INR] of 2.0–2.5). In the 32 patients ated with a higher risk of stroke. Stroke compro-
who died, stroke was the second most common mises survival [18–20]: 12 months after
cause of death (n = 8). Six patients had hemor- implantation, survival in the stroke group was
rhagic strokes, and two had embolic strokes. significantly lower (71%) than that in the non-­
Risks factors for stroke include high blood stroke group (82%) (Fig. 17.6) [20].
pressure, infection, pump thrombosis, GI bleed- In a Columbia University Medical Center
ing, aortic cross-clamping with cardioplegic review of data from 301 patients who underwent
arrest during implantation, and insufficient or implantation with continuous-flow LVADs from
excessive anticoagulation. 2008 to 2015, strokes occurred in 40 (13%). The
A review of 100 consecutive continuous-flow study emphasized appropriate characterization of
pump implantation cases revealed a stroke rate of the type of stroke. A clear distinction was made
12% [18]. Patients who had a stroke had a signifi- between ischemic stroke and primary intracerebral
cantly higher prevalence of diabetes (66% vs. hemorrhagic stroke, which probably have different
41%), previous stroke (17% vs. 5%), and use of etiologies. Unlike patients in other studies, these
aortic cross-clamping with cardioplegic arrest patients were classified as having ischemic stroke,
during LVAD implantation (50% vs. 20%) than ischemic stroke with hemorrhagic conversion, or
did patients without stroke (Figs. 17.4 and 17.5). intracerebral hemorrhagic stroke after the careful
Notably, the mean INR at the time of stroke was review of the clinical presentation and radiologic
subtherapeutic in all four patients who had findings. In this study, ischemic stroke—presum-
embolic strokes. Mortality within 30 days of ably caused by embolic disease originating from
stroke was 25%. A University of Minnesota the LVAD—was the most frequent cause of death,
review of 230 patients in whom a HeartMate II occurring in 32 of the 40 patients who had a stroke;
continuous-flow LVAD was implanted as a bridge the remaining 8 patients had an intracerebral hem-
to transplant revealed a stroke rate of 17%. Of orrhage. On the basis of the Columbia protocol,
those stroke cases, 49% were embolic and 52% continuous-flow LVAD patients were maintained
were hemorrhagic. Diabetes and hypertension on a regimen of aspirin (81 mg/day) and warfarin
244 G.V. Letsou

Fig. 17.5 Freedom
from stroke (patients
with diabetes vs.
patients without
diabetes). Modified with
permission from Wolters
Kluwer [18]

Fig. 17.6 Actuarial
survival rates after
HeartMate II
implantation. Modified
with permission from
Elsevier [20]

(target INR of 2.0–2.5). However, a substantial Ischemic stroke patients often recovered suffi-
proportion of stroke patients had an INR at the ciently to proceed to transplant or discharge [21].
time of stroke that was either subtherapeutic or Identifying optimal anticoagulation strategies
above the target range. For several patients, warfa- continues to be an important area of investiga-
rin had been discontinued for various reasons. tion. Studies by John et al. [22] and Katz et al.
In-hospital mortality was 50% after intracerebral [23] showed that reduced amounts of anticoagu-
hemorrhagic stroke and 28% after ischemic stroke. lation may be acceptable. In the study by John
Survival correlated with the severity of the stroke, et al. [22], 45 patients who received the HeartMate
which was assessed clinically by using the II were anticoagulated with aspirin and warfarin,
National Institutes of Health Stroke Scale. but 41 of those patients had a mean INR less than
17  Long-Term Complications of Ventricular Assist Devices 245

2.0. Among the 21 patients who had a mean INR more common after continuous-flow pump
less than 1.6, only one stroke occurred [22]. In implantation than after pulsatile-flow pump
the intriguing study by Katz et al. [23], 94% of implantation. In a review from the University of
patients receiving reduced anticoagulation (i.e., Minnesota, GI bleeding developed in 12 of 55
only aspirin, only warfarin, or no agents at all) patients (22%) who underwent implantation with
because of bleeding complications were free a continuous-flow pump, whereas only 3 of 46
from ischemic stroke at 1 year [23]. patients (7%) who underwent implantation with a
Early thrombus formation after LVAD implan- pulsatile-flow pump had GI bleeding [26]. A sim-
tation is likely a risk factor for stroke. To minimize ilar predisposition to bleeding after continuous-­
this risk, appropriate attention to anticoagulation flow pump implantation was seen in a study from
with heparin is important in the early postopera- Washington University. The incidence of GI
tive period. With prolonged circulatory support, bleeding in 61 patients who underwent implanta-
the risk of stroke may be increased by comorbidi- tion with a HeartMate II continuous-flow pump
ties such as atrial fibrillation, appropriate aortic was 21% [27]. The overall incidence of GI bleed-
valve closure, and the anatomic position of the ing after continuous-flow pump implantation is
inflow cannula, device, and outflow graft. Stroke is typically between 15 and 30%.
a multifactorial problem that develops in approxi- The most common sources of GI bleeding
mately 10–20% of patients with continuous-flow after continuous-flow pump implantation are
LVADs. Hemorrhagic stroke is typically more arteriovenous malformations and ulcers.
lethal than ischemic stroke. Arteriovenous malformations were somewhat
more common than ulcers in both of the above-
mentioned series. Bleeding episodes are often
Bleeding significant and necessitate transfusion but typi-
cally are not lethal. Management with aggressive
Bleeding is the most common specific adverse medical therapy, including endoscopic interven-
event for patients with a continuous-flow device. tion, is often effective, but repeated interventions
According to the 2009 study by Slaughter et al. are frequently necessary.
[7], 81% of patients had bleeding that necessi- The mechanism of increased bleeding is
tated a packed red blood cell transfusion. unclear and multifactorial. Many investigators
Bleeding in the immediate surgical period is have suggested that a possible cause is acquired
common. Approximately 75% of patients under- von Willebrand factor (vWF) deficiency, which is
going LVAD implantation require blood transfu- characterized by a reduction in high molecular
sion. Re-exploration for bleeding is necessary for weight multimer production after continuous-­flow
20–30% of patients [24]. Reoperation for bleed- pump implantation. Rapidly rotating continuous-­
ing is associated with illness severity in these flow pumps may cause vWF to deform and
compromised patients. Often, patients undergo- degrade into smaller proteins that are then cleared
ing LVAD implantation have significant comor- from the bloodstream. Previously asymptomatic
bidities and have been anticoagulated GI lesions, such as arteriovenous malformations or
preoperatively. Some degree of platelet dysfunc- ulcers, may then become prone to bleeding. In a
tion, as well as hepatic dysfunction, is common. study performed at Duke University and the
Optimizing the patient’s hemodynamics and University of Minnesota, acquired vWF deficiency
overall condition preoperatively is important for was documented in all 37 patients who underwent
minimizing perioperative bleeding. axial continuous-flow pump (HeartMate II)
Gastrointestinal bleeding is common during implantation between July 2008 and April 2009
the recovery period after LVAD implantation. In [28]. However, only 10 of the 37 patients devel-
an early experience with a continuous-flow oped clinically significant GI bleeding.
LVAD, 3 of 21 patients (14%) developed GI Despite the differences in pump conforma-
bleeding [25]. Gastrointestinal bleeding may be tion and rotation speed between centrifugal and
246 G.V. Letsou

axial continuous-flow pumps, the implantation Infections

of either type of pump generally results in a
similar decrease in levels of high molecular Infection after LVAD implantation is a serious
weight multimers of vWF. In a review of 102 cause of patient morbidity and mortality.
consecutive implantations in Leipzig and Compared with pulsatile-flow pumps,
Hannover, Germany, a 34% decrement in continuous-­flow pumps have a markedly lower
plasma levels of high molecular weight multim- overall rate of infection. In the clinical trial by
ers of vWF was observed after implantation of Slaughter et al. [7], the number of infectious
a centrifugal continuous-flow pump events per patient year was almost 50% lower
(HeartWare), compared with a 30% decrement for patients with continuous-­flow pumps (0.48)
after implantation of an axial continuous-­flow than for patients with pulsatile-flow pumps
pump (HeartMate II) [29]. (0.91). This lower rate of infectious events after
Other likely mechanisms that may contribute the implantation of a continuous-flow pump was
to postoperative bleeding after continuous-flow an important factor in reducing the overall read-
pump implantation include the mechanical mission rate for patients after LVAD implanta-
destruction of platelet proteins, increased shear tion. The lower rate of overall infectious
stress on blood elements by the pump rotors, and complications associated with the use of contin-
decreased pulse pressure. Therapeutic anticoagu- uous-flow pumps is likely related to the elimina-
lation after continuous-flow pump implantation tion of the large pump pocket.
usually involves antiplatelet agents and warfarin; Most infections in the continuous-flow pump era
however, an optimal anticoagulation regimen are driveline infections. All continuous-flow and
remains to be determined. Strategies to minimize pulsatile-flow pumps employ a percutaneous drive-
postoperative bleeding include lowering the post- line to carry electrical signals and energy from the
operative target INR and reducing the use of anti- controller and battery unit to the pump. This places
platelet agents. the patient at risk for infection at the driveline
An intriguing strategy for reducing postopera- entrance site and along the subcutaneous tunnel of
tive bleeding is altering the pump speed to induce the driveline leading to the pump. To minimize com-
pulsatility. In a study of 134 patients from the plications from infection, a Dacron cuff surrounds
Utah Transplantation Affiliated Hospitals the driveline. Antibiotic-­ impregnated drivelines
(UTAH) program, nonsurgical bleeding was have been used to minimize infectious complica-
more than fourfold higher in patients with a lower tions. Nevertheless, driveline infections occur in
pulsatility index than in patients with a higher approximately 15–30% of LVAD recipients.
pulsatility index [30]. Driveline infections are most commonly due
Bleeding complications after continuous-flow to Pseudomonas aeruginosa or Staphylococcus
LVAD implantation are common. Early bleeding aureus. Each of these bacterial species accounts
is usually secondary to perioperative surgical fac- for approximately one-third of infections.
tors, including severity of patient illness and Fungal infections are unusual. Many infections
comorbidities. After the patient recovers from remain localized to the driveline exit site. Most
surgical implantation, GI bleeding is frequent. infections begin superficially and either remain
Gastrointestinal bleeding may be related to an localized or progress to involve the entire drive-
acquired VWF deficiency or to the anticoagula- line [31]. Infections occur at a steady rate of
tion regimen. Pulsatility may also play an impor- approximately 2% per month after LVAD
tant role in bleeding complications. Heightened implantation. After 24 months, a driveline infec-
attention to these factors is important for mini- tion may have developed in as many as 30% of
mizing bleeding complications. LVAD recipients (Fig. 17.7).
17  Long-Term Complications of Ventricular Assist Devices 247

Fig. 17.7 Freedom from driveline infection in tion. Error bars show standard deviation. Modified with
continuous-­flow LVAD recipients. Time (months) from permission from Elsevier [32]
LVAD implantation to the first percutaneous site infec-

Although driveline infections are often per- Evidence-based interventions that have been
ceived as benign events, they are associated with a shown to decrease the rate of driveline infection
distinct increase in mortality and hospital readmis- include preparing the surgical site with an appro-
sion rates. In studies by Koval et al. [31] and priate antibiotic agent, properly sterilizing and
Goldstein et al. [32], approximately 10% of disinfecting equipment, minimizing operating
patients in whom driveline or percutaneous site room traffic, and using appropriate ventilation
infections developed eventually died, with sepsis systems. In addition, appropriate glucose control
as the most common cause of death. The most in diabetic patients is important during the peri-
common cause of death in patients with a driveline operative period. Specific measures for postop-
infection is disseminated sepsis. Antibiotic ther- erative driveline care have been proposed, but
apy can be effective in a minority of these cases, none have been shown objectively to decrease the
but device removal and replacement can be incidence of infection. Driveline infection is a
required. The definitive treatment for driveline morbid event that should be taken seriously and
infection or pump infection is heart transplanta- prevented if possible.
tion. Despite immunosuppression in the transplant
recipient, subsequent infection is not common.
Driveline infection is an unusual surgical Conclusion
infection in that it is not typically associated with
traditional surgical risk factors, such as diabetes, Mechanical cardiac assist devices continue to be
acuity of illness, and malnutrition. Younger age developed as an effective treatment for the large
seems to be an important predisposing factor. population of patients living with heart failure.
Many driveline infections occur after patients are Continuous-flow pumps have been a major
discharged home. All of these associations sug- advancement. Nevertheless, long-term chal-
gest that localized minor trauma that occurs as an lenges remain, including reducing rates of pump
outpatient (which may be more likely for younger thrombosis, stroke, bleeding at sites distant from
patients) is an important risk factor for driveline the heart, and infection. Further attention to these
infection. Better outpatient support and driveline issues will be necessary as newer mechanical car-
care should thus be emphasized. diac assist devices are introduced.
248 G.V. Letsou

References for mechanically assisted circulatory support data.

J Heart Lung Transplant. 2015;34(12):1527–34.
16. Saeed D, Maxhera B, Albert A, Westenfeld R,

1. McMurray JJ, Pfeffer MA. Heart failure. Lancet.
Hoffmann T, Lichtenberg A. Conservative approaches
2005;365(9474):1877–89.
for HeartWare ventricular assist device pump throm-
2. Bui AL, Horwich TB, Fonarow GC. Epidemiology
bosis may improve the outcome compared with
and risk profile of heart failure. Nat Rev Cardiol.
immediate surgical approaches. Interact Cardiovasc
2011;8(1):30–41.
Thorac Surg. 2016;23(1):90–5.
3. National Clinical Guideline Centre. Chronic heart
17. Tsiouris A, Paone G, Nemeh HW, et al. Lessons

failure: the management of chronic heart failure
learned from 150 continuous-flow left ventricular
in adults in primary and secondary care. London:
assist devices: a single institutional 7 year experience.
National Clinical Guideline Centre; 2010.
ASAIO J. 2015;61(3):266–73.
4. Lund LH, Edwards LB, Kucheryavaya AY, et al. The
18. Morgan JA, Brewer RJ, Nemeh HW, et al. Stroke
Registry of the International Society for Heart and
while on long-term left ventricular assist device sup-
Lung Transplantation: Thirtieth Official Adult Heart
port: incidence, outcome, and predictors. ASAIO
Transplant Report—2013; focus theme: age. J Heart
J. 2014;60(3):284–9.
Lung Transplant. 2013;32(10):951–64.
19. Willey JZ, Demmer RT, Takayama H, Colombo PC,
5. DeBakey ME. Left ventricular bypass pump for car-
Lazar RM. Cerebrovascular disease in the era of
diac assistance. Clinical experience. Am J Cardiol.
left ventricular assist devices with continuous flow:
1971;27(1):3–11.
risk factors, diagnosis, and treatment. J Heart Lung
6. Rose EA, Gelijns AC, Moskowitz AJ, et al. Long-term
Transplant. 2014;33(9):878–87.
use of a left ventricular assist device for end-stage
20. Harvey L, Holley C, Roy SS, et al. Stroke after

heart failure. N Engl J Med. 2001;345(20):1435–43.
left ventricular assist device implantation: out-
7. Slaughter MS, Rogers JG, Milano CA, et al.
comes in the continuous-flow era. Ann Thorac Surg.
Advanced heart failure treated with continuous-­
2015;100(2):535–41.
flow left ventricular assist device. N Engl J Med.
21. Willey JZ, Gavalas MV, Trinh PN, et al. Outcomes
2009;361(23):2241–51.
after stroke complicating left ventricular assist device.
8. Kirklin JK, Naftel DC, Pagani FD, et al. Seventh
J Heart Lung Transplant. 2016;35(8):1003–9.
INTERMACS annual report: 15,000 patients and count-
22. John R, Kamdar F, Liao K, et al. Low thromboem-
ing. J Heart Lung Transplant. 2015;34(12):1495–504.
bolic risk for patients with the HeartMate II left
9. Morgan JA, Go PH, Xuereb L, et al. Outcomes on
ventricular assist device. J Thorac Cardiovasc Surg.
continuous-flow left ventricular assist devices: a sin-
2008;136(5):1318–23.
gle institutional 9-year experience. Ann Thorac Surg.
23. Katz JN, Adamson RM, John R, et al. Safety of

2016;102(4):1266–73.
reduced anti-thrombotic strategies in HeartMate II
10. Anand J, Singh SK, Hernandez R, et al. Continuous-­
patients: a one-year analysis of the US-TRACE Study.
flow ventricular assist device exchange is safe and
J Heart Lung Transplant. 2015;34(12):1542–8.
effective in prolonging support time in patients with
24. Quader M, LaPar DJ, Wolfe L, et al. Blood product uti-
end-stage heart failure. J Thorac Cardiovasc Surg.
lization with left ventricular assist device implantation: a
2015;149(1):267–75, 278.e61.
decade of statewide data. ASAIO J. 2016;62(3):268–73.
11. Starling RC, Moazami N, Silvestry SC, et al.

25. Letsou GV, Shah N, Gregoric ID, Myers TJ, Delgado
Unexpected abrupt increase in left ventricular assist
R, Frazier OH. Gastrointestinal bleeding from
device thrombosis. N Engl J Med. 2014;370(1):33–40.
­arteriovenous malformations in patients supported by
12. Kirklin JK, Naftel DC, Kormos RL, et al. Interagency
the Jarvik 2000 axial-flow left ventricular assist device.
Registry for Mechanically Assisted Circulatory
J Heart Lung Transplant. 2005;24(1):105–9.
Support (INTERMACS) analysis of pump thrombo-
26. Crow S, John R, Boyle A, et al. Gastrointestinal
sis in the HeartMate II left ventricular assist device.
bleeding rates in recipients of nonpulsatile and pulsa-
J Heart Lung Transplant. 2014;33(1):12–22.
tile left ventricular assist devices. J Thorac Cardiovasc
13. Najjar SS, Slaughter MS, Pagani FD, et al. An

Surg. 2009;137(1):208–15.
analysis of pump thrombus events in patients in the
27. Wang I-W, Guthrie T, Ewald GA, Geltman EM,

HeartWare ADVANCE bridge to transplant and con-
Joseph S, Moazami N. Gastrointestinal bleeding com-
tinued access protocol trial. J Heart Lung Transplant.
plications in continuous-flow LVAD patients: is it
2014;33(1):23–34.
device specific? J Heart Lung Transplant. 2010;29(2
14. Kirklin JK, Naftel DC, Pagani FD, et al. Pump throm-
Suppl):S8.
bosis in the Thoratec HeartMate II device: an update
28. Crow S, Chen D, Milano C, et al. Acquired von

analysis of the INTERMACS Registry. J Heart Lung
Willebrand syndrome in continuous-flow ven-
Transplant. 2015;34(12):1515–26.
tricular assist device recipients. Ann Thorac Surg.
15. Smedira NG, Blackstone EH, Ehrlinger J, et al.

2010;90(4):1263–9; discussion 1269.
Current risks of HeartMate II pump thrombosis:
29. Meyer AL, Malehsa D, Budde U, Bara C, Haverich
non-parametric analysis of interagency registry
A, Strueber M. Acquired von Willebrand syndrome
17  Long-Term Complications of Ventricular Assist Devices 249

in patients with a centrifugal or axial continuous cohort of continuous-flow ventricular assist device recip-
flow left ventricular assist device. JACC Heart Fail. ients. J Heart Lung Transplant. 2014;33(11):1164–72.
2014;2(2):141–5. 32. Goldstein DJ, Naftel D, Holman W, et al. Continuous-­
30. Wever-Pinzon O, Selzman CH, Drakos SG, et al.
flow devices and percutaneous site infections:
Pulsatility and the risk of nonsurgical bleeding in clinical outcomes. J Heart Lung Transplant. 2012;
patients supported with the continuous-flow left 31(11):1151–7.
ventricular assist device HeartMate II. Circ Heart 33. Anand J, Singh SK, Antoun DG, Cohn WE, Frazier
Fail. 2013;6(3):517–26. OH, Mallidi HR. Durable mechanical circulatory sup-

31. Koval CE, Thuita L, Moazami N, Blackstone port versus organ transplantation: past, present, and
E. Evolution and impact of drive-line infection in a large future. Biomed Res Int. 2015;2015:849571.
 Right Ventricular Failure
in Patients Undergoing LVAD 18
Placement

David Kuten and Joggy K. George

which was associated with a higher incidence of

Introduction end-organ dysfunction, longer intensive care unit
stays, and greater mortality [2]. RVF remains a
The introduction of left ventricular assist devices morbid complication in the era of continuous-­
(LVADs) has proven to reduce morbidity and flow devices. Data from the initial HeartMate II
mortality in patients with advanced heart failure bridge-to-transplant trial show that patients were
[1]. Despite their benefits, LVADs are associated significantly less likely to survive to transplant,
with a host of adverse events. These include right recovery, or ongoing device support if they had
ventricular failure (RVF), which is a particularly RVF [3]. In this chapter, we discuss issues related
common and morbid complication when it arises to early RVF, including its evolving definition,
in the early postoperative period. The hemody- myriad of causes, incidence, predictors, and
namic consequences, both positive and negative, diagnosis. Particular attention is paid to the pre-
of LVAD therapy on the right ventricle are vention and management of RVF.
addressed in greater detail elsewhere in this text.
Briefly, although the right ventricle benefits from
LV decompression of the pulmonary circulation, Definition
it remains vulnerable to multiple insults related
to LVAD-induced hemodynamic changes, in The right heart serves both as a low-pressure res-
addition to the complications associated with ervoir for blood as it returns from the systemic
major cardiac surgery. circulation and as a pump filling the left heart via
The impact of RVF on patient outcomes was the pulmonary circulation. RVF is caused by an
recognized early in the clinical experience with inability to perform one or both of these impor-
LVADs. A 2002 report found that almost one-­ tant roles. In patients with an LVAD, RVF trans-
third of HeartMate XVE recipients had RVF, lates into an inability to preserve the low-pressure
state of the right heart, resulting in Fontan-like
physiology, and failure to maintain adequate
­preload to the left heart. The INTERMACS crite-
D. Kuten, M.D. • J.K. George, M.D. (*) ria for RVF require documentation of elevated
Texas Heart Institute at Baylor St. Luke’s Medical central venous pressure (CVP) and clinical mani-
Center, Baylor College of Medicine,
Houston, TX, USA
festations of venous congestion [4]. The
e-mail: dkuten@[Link]; INTERMACS classifies early RVF as mild if the
davidkuten85@[Link]; jkgeorgemd@[Link] patient requires inotropes, inhaled nitric oxide, or

© Springer International Publishing AG 2018 251

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
252 D. Kuten and J.K. George

intravenous pulmonary vasodilators for less than increased RV preload, such that RV power output
7 days after implantation, and RVF is classified and RV myocardial efficiency were unchanged
as moderate if such therapies are required for [8]. Thus, the overall significance of changes in
7–14 days. Early RVF is classified as severe if septal geometry is not completely understood
CVP is greater than 16 mmHg, and the patient and may be affected by an individual patient’s
requires these therapies for more than 14 days unique physiologic milieu, including the degree
after LVAD implantation. RVF is considered of RV afterload reduction observed after LVAD
severe and acute if the patient requires an RV implantation.
assist device (RVAD) or dies during the index In addition to these structural alterations, the
hospitalization with RVF as the primary cause of RV is exposed to an increase in preload result-
death. ing from augmented cardiac output after LVAD
placement. Although a healthy, highly compli-
ant right ventricle can tolerate increases in pre-
Etiology load, particularly with a concomitant reduction
in RV afterload, factors such as baseline RV
The right ventricle is uniquely vulnerable to early dysfunction and significant tricuspid regurgita-
LVAD hemodynamic alterations for reasons tion (TR) may impair the right ventricle’s ability
related to its underlying anatomy and physiology. to compensate. TR due to annular dilatation, in
The anteriorly positioned right ventricle is a thin-­ particular, may deteriorate with increased RV
walled, highly compliant structure with a com- preload. A 2011 study of 176 patients who
plex shape that is affected by neighboring underwent either pulsatile or continuous-flow
structures and is critical to its function [5]. RV LVAD implantation without tricuspid annulo-
contractility depends on the function of the RV plasty found that TR did not improve in the
free wall and the interventricular septum, which, immediate postoperative period and that patients
in turn, is highly dependent on LV function. As a with at least moderate TR required longer ino-
result, the right ventricle is exquisitely affected tropic support, had longer hospitalization, and
by changes in loading conditions in both ventri- had a trend toward poorer survival [9]. Almost
cles, as well as septal position. 10% of patients with significant preoperative
RV function benefits from the decrease in RV TR required temporary RVAD support, whereas
afterload caused by reduced LV end-diastolic none of the patients with insignificant TR
pressure and pulmonary artery pressure with required such support.
LVAD therapy. However, leftward shift of the Along with these LVAD-specific factors, the
interventricular septum with reduced LV filling usual complications of cardiac surgery can also
pressures may diminish RV contractility [6]. cause RV dysfunction. For instance, myocardial
Moon et al. [7] elucidated the effect of acute LV ischemia in the RV distribution can be caused by
unloading on interventricular septal position and intraoperative air emboli, which preferentially
function in a canine model by clamping left atrial affect the right coronary artery because of its
pressure between 0 and 2 mmHg, thereby reduc- anterior location. In patients with ischemic car-
ing LV pressure (mean reduction, 30%) [7]. As a diomyopathy, ischemia can also result from the
result of this intervention, the septum moved left- exacerbation of epicardial obstructive disease
ward and flattened. Functionally, both systolic due to perioperative increases in myocardial
septal thickening and systolic septal output (the demand. In addition, pericardial tamponade due
product of septal thickening and heart rate) were to postoperative bleeding can impair RV filling,
reduced. However, an earlier study by the same resulting in venous congestion and diminished
group calls into question the impact of these LV preload.
changes. Among eight dogs that underwent Another potential cause of RVF is sustained
LVAD implantation, the reduction in RV contrac- tachyarrhythmia, which can disproportionately
tile function caused by septal displacement was affect the unsupported right ventricle. A 2015
counterbalanced by reduced RV afterload and publication by Garan et al. [10] found a high
18  Right Ventricular Failure in Patients Undergoing LVAD Placement 253

incidence of ventricular arrhythmia in the period a 2008 study comparing outcomes after
immediately after continuous-flow LVAD implan- HeartMate XVE and HeartMate II implantation
tation. Among 162 patients enrolled between at Johns Hopkins Hospital [12]. Consistent with
2012 and 2014, 38 (23%) had ventricular arrhyth- the aforementioned studies, early postoperative
mia. The investigators found that patients with RVF was defined as the need for inotrope or
ventricular arrhythmia had RVF at a higher rate inhaled pulmonary vasodilator therapy for more
than those without arrhythmia (44.7 vs. 23.4%; than 14 days, or RVAD insertion. Among 77
P = 0.01). patients included in the study, the investigators
Interestingly, findings suggest that appropriate found similar incidences of RVF in those sup-
implantable cardioverter-defibrillator (ICD) shocks ported by pulsatile and continuous-flow devices
may be inherently detrimental in such patients. (41.2% vs. 34.9%; P = 0.63). The authors posit
Among Garan and colleagues’ 129 patients with that their increasing use of inhaled pulmonary
ICDs, 25 had ICD shocks and 13 had antitachycar- vasodilators and inotropes to facilitate diuresis
dia pacing (ATP) for malignant arrhythmia. ICD toward the end of the study period may have
shocks were associated with RV dysfunction, evi- biased the results. To this point, patients with a
denced by increased CVP and initiation of therapy continuous-flow LVAD were more likely to
for RVF in 8 of the 25 patients. In contrast, none of receive milrinone (rather than epinephrine) than
the 13 patients who received ATP had RV dysfunc- were patients with pulsatile devices. Furthermore,
tion. Also informatively, the authors report that HeartMate II recipients were less likely to require
43% of the arrhythmia events were associated with RVAD insertion than HeartMate XVE recipients.
inotropic medications, implicating perioperative A larger, multicenter experience with the
vasoactive medications in the onset of ventricular HeartMate II among 484 patients in the initial
arrhythmia with resultant RVF. bridge-to-transplant trial of that device found a
20% incidence of RVF [3]. Here, RVF was defined
as the need for RVAD insertion (7% of patients),
Incidence continuous treatment with inotropes for more
than 14 days (6%), and the need for inotropic sup-
The reported incidence of RVF after LVAD port beginning 14 days after device implantation
placement varies, in large part, according to how (6%). Thus, only 13% of patients had early post-
RVF is defined in each study, the comorbidities operative RVF. Consistent with previous reports,
of the study cohort, the type of LVAD used, and the authors found that outcomes were worse in
the medical therapies available to the treating cli- patients with RVF. A smaller, single-­center expe-
nicians. In a 2002 publication detailing the expe- rience published contemporaneously found an
rience of 69 recipients of a HeartMate XVE at even lower incidence of RVF (2 of 40 patients),
Columbia University, the authors report a 30% which was similarly defined [13]. Study investi-
incidence of RVF, which they define as inotrope gators posit that the lower incidence of RVF in
use for more than 14 days or need for RVAD sup- these modern cohorts with axial-flow devices may
port [2]. Patients with RVF had higher postopera- relate to a confluence of several factors, including
tive rates of mortality and certain morbidities, the more favorable hemodynamic impact of these
including bleeding requiring reoperation and newer devices on the right ventricle, better patient
renal failure. In 2006, the same institution selection and preoperative optimization, and
reported a 38.9% incidence of RVF, similarly refined postoperative management.
defined, among 108 patients [11]. Of the 42 In summary, although RVF in the early post-
patients with RVF, 14 required RVAD insertion. operative course is less common in the era of
Whether continuous-flow LVADs might reduce continuous-flow LVADs than it was in the era of
the high incidence and morbidity of RVF in pulsatile devices, the incidence of postoperative
patients with pulsatile LVADs was investigated in RVF remains high.
254 D. Kuten and J.K. George

Predictors bridge-to-transplant trial data, found that an

elevated (i.e., >0.63) CVP-to-­ PCWP ratio was
Given the morbidity associated with RVF in more predictive of RVF than was elevated CVP
LVAD-supported patients, investigators have alone, implicating RV dysfunction, as opposed
sought to identify clinical, hemodynamic, bio- to left-sided congestion, as the culprit of
chemical, and echocardiographic risk factors for RVF. Predictably, these authors also found that a
its development. Risk calculators incorporating a low RVSWI was associated with RVF.
multitude of parameters have also been devel- Biochemical predictors of postoperative RVF
oped. The studies conducted to date are limited include those suggestive of elevated CVP, such as
by small sample size, retrospective design, and elevated bilirubin and transaminase levels and
inclusion of patients with pulsatile LVADs. markers of renal insufficiency, including elevated
Clinical predictors of RVF were first described blood urea nitrogen and creatinine levels [3, 16].
in a large cohort by Ochiai et al. [14], who in The usefulness of preoperative transthoracic
2002 published the Cleveland Clinic experience echocardiography in identifying patients at risk
with 245 patients who received a pulsatile for RVF after LVAD implantation has been inves-
LVAD. The authors found that risk factors for tigated in a few small cohorts. One study of 33
postoperative RVAD placement, which was per- patients found that reduced tricuspid annular
formed in 23 patients, included female gender motion predicted postoperative RVF, whereas Tei
(odds ratio [OR] 4.5), nonischemic cardiomyopa- index, RV fractional area change, and right atrial
thy (OR 3.3), and pre-LVAD circulatory support dimension were not predictive [17]. A subse-
(OR 5.3). The authors speculate that patients quent study published in 2011 found that among
with nonischemic cardiomyopathy are more 40 patients with various LVADs, more severe
likely to have preoperative biventricular failure preoperative tricuspid insufficiency correlated
than those with ischemic cardiomyopathy. with RVF [18]. A 2013 study of 55 patients iden-
Subsequent studies that included patients with tified three predictors of RVF: a reduced RV frac-
both pulsatile and continuous-flow LVADs have tional area change, elevated estimated right atrial
identified additional clinical predictors such as pressure, and low left atrial volume index; here,
preoperative intra-aortic balloon pump, mechani- tricuspid annular motion was not predictive [19].
cal ventilation, and vasopressor administration, The disparate findings of these studies reflect
suggesting that preoperative cardiovascular and their small sample sizes, variable definitions of
pulmonary instability plays an important role in RVF, and diverse patient populations.
the development of RVF [12, 15, 16]. Given the varied contributors to RVF in
Hemodynamic predictors of RVF have been patients undergoing LVAD, multiple investiga-
investigated, as well. Ochiai et al. [14] found that tors have attempted to develop a weighted risk
low RV stroke work index (RVSWI), but not ele- calculator. In 2008, Matthews et al. [16] pub-
vated pulmonary artery pressure, was a risk factor lished a risk prediction model developed from
for RVAD placement. That poor preoperative RV data obtained from 179 LVAD-supported patients
contractility, as reflected by low RVSWI, can pre- at the University of Michigan. Multivariate risk
dict RVF in patients undergoing LVAD support factors for RVF (defined as inotrope use for
was supported in a 2006 study from Columbia, >14 days postoperatively, inotrope use at dis-
which reported a lower intraoperative systolic charge, inhaled nitric oxide use for >48 h, or the
pulmonary artery pressure in patients with right need for a mechanical RVAD) included vasopres-
heart failure than in unaffected patients (51 ± 11 sor use and serum levels of creatinine >2.3 mg/
vs. 58 ± 11 mmHg; P = 0.047) [11]. Of note, this dL, bilirubin >2 mg/dL, and AST >80 IU/L. The
study identified elevated intraoperative CVP as an factors were compiled into a weighted risk calcu-
even stronger hemodynamic predictor of RVF lator that classifies patients’ risk of postoperative
(23 ± 8 vs. 17 ± 6 mmHg; P = 0.017). Kormos RVF as low, intermediate, or high. A more
et al. [3], in their analysis of the HeartMate II recently published risk score designed to predict
18  Right Ventricular Failure in Patients Undergoing LVAD Placement 255

need for biventricular assist device (BIVAD) sup- Prevention: Preoperative

port is based on patient data collected at the
University of Pennsylvania between 2003 and Instituting measures to prevent RVF in patients
2011 [20]. The investigators identified five multi- who are identified as high risk is essential.
variate predictors of postoperative RVF: CVP However, the difficulty of identifying patients
>15 mmHg, severe preoperative RV dysfunction, who will ultimately develop RVF requires that, in
preoperative intubation, severe tricuspid insuffi- all patients, clinicians optimize hemodynamics
ciency, and tachycardia. A risk calculator was preoperatively and maintain a high index of sus-
created that assigns each factor a 0 or 1 and sums picion for RV dysfunction during the postopera-
the results to produce a single score; higher tive period. An algorithm summarizing our
scores correspond with greater risk of approach to the prevention and management of
RVF. Although we do not systematically use RVF is provided (Table 18.1).
these risk scores, it is our opinion that they attest Preoperative measures to prevent RVF include
the overall acuity of a patient’s condition, which invasive hemodynamic monitoring, volume opti-
is the primary driver of RVF. mization, the perioperative use of pulmonary

Table 18.1  Prevention and management of RVF in patients undergoing LVAD placement
Preoperative measures
Risk stratification
Low Moderate High
Limit fluids
Pulmonary vasodilators
Cardiac inotropes
Consider elective RVAD
Operative measures
Preventive High-risk patient/early RVF Fulminant RVF
Limit transfusion
Expeditious operative time
TEE-guided optimization of pump speed
Optimize electrolytes/pH balance
TEE-guided titration of vasodilators
and cardiac inotropes
Consider elective RVAD
If severe TR, consider elective
tricuspid annuloplasty
Insert central RVAD
Postoperative measures
Preventive Early RVF Fulminant RVF
PA-guided volume and inotrope
management
Limit arrhythmias
Limit RV preload
Rule out reversible causes:
• VT
•  RV hematoma
•  Epicardial ischemia
•  Increased preload
Insert RVAD
•  Elective: central or peripheral
•  Emergent: VA-ECMO
256 D. Kuten and J.K. George

vasodilators and inotropic medications, and weaned from RVAD support by postoperative
careful selection of patients who could benefit day 18 and were successfully discharged with an
from empiric BIVAD support. We commonly LVAD alone.
employ indwelling pulmonary artery catheters in The advent of percutaneous RVADs (Impella
patients awaiting LVAD implantation in order to RP and TandemHeart) is actively changing the
titrate therapies. As mentioned, pulmonary artery discussion as it relates to the prophylactic use of
pressure alone may not be as useful as the over- temporary RVADs. Schmack et al. [22] describe
all hemodynamic picture, including pulmonary a practice of prophylactic RVAD support with the
vascular resistance, RVSWI, and the ratio of TandemHeart, using the Protek Duo cannula, in
CVP to PCWP. patients at risk for RVF. Potential benefits of this
Optimizing volume status before LVAD place- device include its lower invasiveness, the mobil-
ment is important. We strive to attain the lowest ity it affords to patients, and that it can be
right and left heart filling pressures that the explanted without reoperation. The authors do
patient can tolerate from a hemodynamic and not specify in how many patients this strategy has
renal standpoint. The choice of strategy for been used or on what basis patients were deemed
achieving this goal is informed by the individual high risk for RVF. Although we have used percu-
patient’s hemodynamic stability, renal function, taneous devices in the postoperative period after
and degree of volume overload. Intravenous loop RVF has developed, we have not yet used them
diuretics, usually administered via a continuous prophylactically in high-risk patients. We hope
drip, with or without thiazide diuretics, are the that future improvements in preoperative risk
mainstay of diuresis. Often, patients are already prediction models for RVF will help us to select
receiving low-dose inotropes, which can poten- patients who are most likely to benefit from such
tially improve renal perfusion. In patients who a strategy.
are profoundly volume overloaded or who have Another controversial issue is whether patients
severe renal dysfunction, ultrafiltration with con- with severe tricuspid annular dilatation benefit
tinuous renal replacement therapy is necessary, from undergoing tricuspid valve annuloplasty con-
with the goal of reducing filling pressures. comitantly with LVAD placement. In a 2014 sin-
When used judiciously, preoperative percuta- gle-center study involving 101 patients who
neous LVAD support can optimize organ perfu- underwent LVAD implantation, 14 patients who
sion and decompress both ventricles. Although had concomitant tricuspid valve repair (all of
we have experience with the TandemHeart whom had moderate or greater TR) were found to
(CardiacAssist, Inc., Pittsburgh, PA) and the have greater survival, but not less severe RVF, than
femorally implanted Impella (Abiomed, patients with similarly severe valvulopathy who
Danvers, MA), using these devices necessitates did not undergo annuloplasty [23]. A 2015 meta-
bedrest. We have, therefore, begun to use axil- analysis of six observational studies with a total of
lary intra-­aortic balloon pumps and the Impella 3249 patients found no difference in survival or
5.0 more frequently. RVF rates in LVAD recipients who underwent
A particularly controversial topic in the peri- concomitant tricuspid valve repair versus LVAD
operative management of LVAD recipients is that implantation alone [24]. We do not routinely per-
of empiric placement of BIVADs in select form tricuspid annuloplasty concomitantly with
patients. Proponents argue that, although RVAD LVAD implantation at our institution.
implantation for RVF is associated with increased
morbidity, planned BIVAD placement improves
outcomes in high-risk patients. A 2010 report Prevention: Postoperative
from an Italian group describes a strategy of
planned BIVAD placement in which central By the nature of their underlying disease, patients
RVAD is used at the time of LVAD placement in who undergo LVAD placement are uniformly high
patients at high risk of RVF [21]. All of the six risk with regard to major cardiac surgery. Thus,
patients described in the report were successfully they are particularly susceptible to intraoperative
18  Right Ventricular Failure in Patients Undergoing LVAD Placement 257

complications such as air embolus into the right tive period because high speeds may induce
coronary artery, aggressive transfusion of blood unnecessary RV strain. Consequently, we prefer
products, and ischemia and vasoplegia associ- to evaluate only one or two incrementally higher
ated with prolonged cardiopulmonary bypass. speeds, paying close attention to echocardio-
Expeditious procedural times and paying scrupu- graphic and invasive indicators of RV function.
lous attention to intraoperative bleeding while The prophylactic use of pulmonary vasodila-
limiting transfusion are imperative. Intraoperative tors after LVAD implantation in high-risk patients
transesophageal echocardiography (TEE) is is appealing. The effects of inhaled nitric oxide in
invaluable for optimizing cannula placement and LVAD patients with elevated pulmonary vascular
initial pump speed. resistance (PVR) were explored in 2011 by
On returning to the recovery unit, patients Potapov et al. [25], who randomly assigned 150
must be hemodynamically optimized. Acid-base patients with preoperative PVR greater than
status and electrolytes should be monitored and 200 dyn*s/cm−5 to receive either inhaled nitric
corrected as needed. Volume status should be oxide or placebo. Patients who received inhaled
assessed with invasive hemodynamic monitoring. nitric oxide had less RV dysfunction, time on
We strive to maintain a CVP as low as can be mechanical ventilation, and need for an RVAD
hemodynamically tolerated through the afore- than the placebo-treated patients, but these differ-
mentioned use of diuretics and, if needed, ences did not reach statistical significance.
ultrafiltration. Whether prophylactic use of inhaled prostacy-
Postoperative arrhythmia, while tolerated by clins, such as epoprostenol and iloprost, may be
the left ventricle, is a potential source of RV dys- protective has not been investigated in a placebo-­
function. Patients who develop supraventricular controlled trial. One group that administered epo-
or ventricular tachyarrhythmia should be aggres- prostenol to 37 consecutive LVAD recipients
sively treated with intravenous antiarrhythmics found that it reduced pulmonary pressures
and, if necessary, synchronized cardioversion. A whether it was initiated before or during weaning
search should be undertaken for the underlying from cardiopulmonary bypass [26]. Whether this
cause of the postoperative arrhythmia, such as strategy improved clinical outcomes has yet to be
high pump speeds with septal interference, sub- determined. Despite the lack of high-quality evi-
optimal cannula positioning, electrolyte abnor- dence to support their use, we have a low thresh-
malities, postoperative pericardial bleeding, or old for initiating either inhaled nitric oxide or
RV ischemia. Vasopressor administration should epoprostenol administration in patients who are
be minimized. The best time to resume tachyther- perceived to be at high risk for RV dysfunction
apy in patients with defibrillators is not clear. We after LVAD placement.
typically resume tachytherapy immediately; In addition to targeted pulmonary vasodila-
however, if a patient requires multiple defibrilla- tors, cardiac inotropes are frequently used in
tions, adjustments may be required, including the perioperative period to enhance RV sup-
increasing the defibrillation threshold or turning port. Milrinone, a phosphodiesterase III inhibi-
off shocks altogether. tor that increases cardiac inotropy and causes
Echocardiographic guidance is necessary to pulmonary and systemic vasodilation by
optimize pump settings and, thus, hemodynamic increasing tissue levels of cAMP, has a particu-
conditions. We use intraoperative TEE to select larly appealing hemodynamic profile. However,
initial device settings, and we obtain serial trans- when not used selectively, intravenous milri-
thoracic echocardiograms in the early postopera- none can cause systemic hypotension. Inhaled
tive period to ensure that increases in pump speed milrinone has recently been explored as an
are well tolerated. Although we frequently use alternative formulation that may not have this
ramp or “speed-change” studies to evaluate car- adverse effect. Haglund et al. [27] have
diac response to various pump speeds in patients described their experience with using inhaled
with an LVAD, caution should be taken in per- milrinone in ten consecutive patients who
forming such a study during the early postopera- underwent HeartMate II LVAD implantation.
258 D. Kuten and J.K. George

The authors note a reduction in pulmonary preference. Evidence for the efficacy of these
pressures and no episodes of sustained hypo- agents in patients with RVF after LVAD is mainly
tension. In addition, institutional costs were extrapolated from the general cardiac surgery lit-
significantly lower with the use of inhaled mil- erature. In 2002, Kihara et al. [28] described the
rinone than with the use of inhaled nitric oxide. use of intravenous milrinone in two patients with
At our institution, LVAD patients frequently RVF after LVAD. Used at low doses to avoid sys-
begin receiving intravenous milrinone, usually temic hypotension and malignant arrhythmia,
in conjunction with low-dose epinephrine, milrinone was associated with clinical benefit and
dobutamine, or both, on weaning from cardio- avoidance of RVAD placement in both patients. A
pulmonary bypass. The adverse effects of these more recent retrospective study of 149 patients
medications, including their proarrhythmic who underwent continuous-flow LVAD implanta-
properties, are well-known; thus, the risks and tion at Henry Ford Hospital found that, among 18
benefits of their use in an individual patient patients who developed postoperative RVF, mor-
should be continuously reassessed by using all tality was no different when patients were treated
available hemodynamic information. with prolonged milrinone; on the other hand,
patients who required RVAD support had poorer
survival [29]. We consider milrinone, dobuta-
Management mine, and epinephrine first-line agents in the
management of postoperative RVF, and we typi-
Should RVF develop after LVAD implantation cally begin administering them either prophylac-
despite preventive measures, it must be quickly tically or at the first signs of RV dysfunction.
identified and aggressively managed. Any of the Although randomized studies are lacking that
aforementioned reversible causes should be could validate the use of pulmonary vasodilators
excluded, and electrolyte levels, acid-base status, in patients with RVF after LVAD, observational
and cardiac rhythm should be optimized. reports support their use. A 2012 report by a
Here, again, transthoracic echocardiography Greek group describes the use of combination
is crucial to optimizing LVAD speed and identi- treatment with inhaled nitric oxide and iloprost in
fying reversible causes of RVF, such as pericar- seven patients with RVF refractory to inotropic
dial effusion. Absence of arterial pulsatility and support with dobutamine and epinephrine [30].
aortic valve opening, along with septal bowing The authors describe significant reductions in
into the left ventricle, suggests that the LVAD’s PVR and mean pulmonary artery pressure, along
pump speed may be too high. Reducing pump with increased tricuspid annular velocity as mea-
speed in such cases relieves RVF by reducing RV sured by echocardiogram, without a clinically
preload and allowing the interventricular septum significant drop in mean arterial pressure. None
to assume a midline position, particularly if there of the patients required RVAD support. A 2014
is septum-cannula interference causing suction study supporting a strategy of liberal pulmonary
events and ventricular ectopy. Patients with evi- vasodilator use in unselected LVAD patients
dence of RVF may need to be able to tolerate (91% received inhaled nitric oxide, iloprost, oral
some degree of incomplete LV decompression in sildenafil, or some combination of these) found
the early postoperative period. that, among six patients with postoperative RVF,
As mentioned previously, we have a low none required an RVAD [31]. Five of the six
threshold for administering pulmonary afterload-­ patients received inhaled nitric oxide, and all six
reducing agents as patients wean from cardiopul- received iloprost and sildenafil. At our institution,
monary bypass. We typically continue to give we have a similarly low threshold for using pul-
patients inotropic agents, including milrinone, monary vasodilators in high-risk patients, espe-
dobutamine, and epinephrine, during the early cially those with postoperative indicators of RVF.
postoperative period; individual regimens are In patients who are not selected for a strategy
determined by hemodynamic profile and clinician of planned BIVAD support, an RVAD is inserted
18  Right Ventricular Failure in Patients Undergoing LVAD Placement 259

Fig. 18.1  TandemHeart Protek Duo placed for RVF following LVAD. (a) Post-LVAD TTE demonstrating LV compres-
sion by the RV. (b) Tandem Protek Duo placed in the cath lab

when the aforementioned medical therapies for flows greater than 4 L/min. The TandemHeart
RVF are ineffective. Frequently, this decision is pump offers the benefit of an oxygenator to
made during the LVAD insertion procedure when facilitate early extubation; when used with the
the patient cannot be weaned from cardiopulmo- Protek Duo dual-lumen catheter, it can be placed
nary bypass despite aggressive use of inotropes through a single jugular access site, enabling
and pulmonary vasodilators, and RV function is patient ambulation (Fig. 18.1).
severely reduced on visual inspection. In such Options for permanent RV support are limited.
cases, the surgeon may elect to insert a temporary Currently, patients who require durable biventricular
RVAD centrally. support are considered candidates for the SynCardia
Once the patient is in the recovery unit, Total Artificial Heart (SynCardia Systems, Inc.,
options for mechanical RV support include Tucson, AZ) or temporary RVAD support as a bridge
extracorporeal membrane oxygenation (ECMO), to cardiac transplantation (Fig. 18.2).
percutaneous RVAD support, and returning to
the operating room for central RVAD placement.
We reserve ECMO for patients experiencing Conclusion
complete cardiopulmonary collapse, in which
case femoral cannulation can be performed and RVF is a common and morbid complication seen
full support initiated at the bedside. As men- in the early postoperative period. Although
tioned earlier, percutaneous temporary RVAD numerous demographic, biochemical, hemody-
support is an increasingly attractive option. namic, and imaging predictors of early RVF have
Available devices include the Impella RP and been identified, no patient is immune from this
TandemHeart, both of which we have used in adverse event. Thus, with the expanding use of
cases of refractory RVF as a bridge to RV recov- LVADs in an increasingly sick patient popula-
ery or permanent RV support. Inserted through tion, it is critical that the clinician make efforts to
the femoral vein, the Impella RP is a 22 Fr pump prevent RVF and identify early clinical and
on an 11 Fr catheter and is capable of producing hemodynamic signs as they develop. Further, the
260 D. Kuten and J.K. George

Fig. 18.2  HeartWare RVAD placed for RVF following Following immediate implantation of HeartWare device
LVAD. (a) RV dilatation and severe TR by intraoperative in the RV position, biventricular support was tolerated for
TEE immediately following LVAD placement. (b) 9 months pending successful heart transplant

clinician must be familiar with the growing arse- 6. Santamore WP, Gray LA Jr. Left ventricular contri-
butions to right ventricular systolic function during
nal of medications and mechanical interventions LVAD support. Ann Thorac Surg. 1996;61(1):350–6.
available for the timely and effective manage- 7. Moon MR, Bolger AF, DeAnda A, et al. Septal func-
ment of RVF. tion during left ventricular unloading. Circulation.
1997;95(5):1320–7.
8. Moon MR, Castro LJ, DeAnda A, et al. Right ven-
Acknowledgments We would like to thank Stephen tricular dynamics during left ventricular assistance in
Palmer, PhD., for his editorial support with this chapter. closed-chest dogs. Ann Thorac Surg. 1993;56(1):54–
66; discussion 66–7.
9. Piacentino V, 3rd, Williams ML, Depp T, et al. Impact
References of tricuspid valve regurgitation in patients treated
with implantable left ventricular assist devices. Ann
Thorac Surg. 2011;91(5):1342–6; discussion 1346–7.
1. Rose EA, Gelijns AC, Moskowitz AJ, et al. 10. Garan AR, Levin AP, Topkara V, et al. Early post-­
Long-term use of a left ventricular assist device operative ventricular arrhythmias in patients with
for end-stage heart failure. N Engl J Med. continuous-flow left ventricular assist devices. J Heart
2001;345(20):1435–43. Lung Transplant. 2015;34(12):1611–6.
2. Kavarana MN, Pessin-Minsley MS, Urtecho J, et al. 11. Dang NC, Topkara VK, Mercando M, et al. Right
Right ventricular dysfunction and organ failure in left heart failure after left ventricular assist device implan-
ventricular assist device recipients: a continuing prob- tation in patients with chronic congestive heart fail-
lem. Ann Thorac Surg. 2002;73(3):745–50. ure. J Heart Lung Transplant. 2006;25(1):1–6.
3. Kormos RL, Teuteberg JJ, Pagani FD, et al. Right 12. Patel ND, Weiss ES, Schaffer J, et al. Right heart dys-
ventricular failure in patients with the HeartMate II function after left ventricular assist device implanta-
continuous-flow left ventricular assist device: inci- tion: a comparison of the pulsatile HeartMate I and
dence, risk factors, and effect on outcomes. J Thorac axial-flow HeartMate II devices. Ann Thorac Surg.
Cardiovasc Surg. 2010;139(5):1316–24. 2008;86(3):832–40; discussion 832–40.
4. INTERMACS Executive Committee. INTERMACS 13. Lee S, Kamdar F, Madlon-Kay R, et al. Effects of the
adverse event definitions: adult and pediatric patients. HeartMate II continuous-flow left ventricular assist
[Link] device on right ventricular function. J Heart Lung
ces-4-0/appendix-a-4-0. Accessed 17 Nov 2016. Transplant. 2010;29(2):209–15.
5. Haddad F, Hunt SA, Rosenthal DN, Murphy 14. Ochiai Y, McCarthy PM, Smedira NG, et al.

DJ. Right ventricular function in cardiovascular dis- Predictors of severe right ventricular failure after
ease, part I: anatomy, physiology, aging, and func- implantable left ventricular assist device insertion:
tional assessment of the right ventricle. Circulation. analysis of 245 patients. Circulation. 2002;106(12
2008;117(11):1436–48. Suppl 1):I198–202.
18  Right Ventricular Failure in Patients Undergoing LVAD Placement 261

15. Drakos SG, Janicki L, Horne BD, et al. Risk fac- empt right heart failure. Med Sci Monit Basic Res.
tors predictive of right ventricular failure after left 2016;22:53–7.
ventricular assist device implantation. Am J Cardiol. 23. Brewer RJ, Cabrera R, El-Atrache M, et al.

2010;105(7):1030–5. Relationship of tricuspid repair at the time of left ven-
16. Matthews JC, Koelling TM, Pagani FD, Aaronson tricular assist device implantation and survival. Int
KD. The right ventricular failure risk score a pre-­ J Artif Organs. 2014;37(11):834–8.
operative tool for assessing the risk of right ventricu- 24. Dunlay SM, Deo SV, Park SJ. Impact of tricuspid
lar failure in left ventricular assist device candidates. valve surgery at the time of left ventricular assist
J Am Coll Cardiol. 2008;51(22):2163–72. device insertion on postoperative outcomes. ASAIO
17. Puwanant S, Hamilton KK, Klodell CT, et al.
J. 2015;61(1):15–20.
Tricuspid annular motion as a predictor of severe 25. Potapov E, Meyer D, Swaminathan M, et al. Inhaled
right ventricular failure after left ventricular assist nitric oxide after left ventricular assist device implan-
device implantation. J Heart Lung Transplant. tation: a prospective, randomized, double-blind,
2008;27(10):1102–7. multicenter, placebo-controlled trial. J Heart Lung
18. Baumwol J, Macdonald PS, Keogh AM, et al. Right Transplant. 2011;30(8):870–8.
heart failure and “failure to thrive” after left ventricu- 26. Groves DS, Blum FE, Huffmyer JL, et al. Effects
lar assist device: clinical predictors and outcomes. of early inhaled epoprostenol therapy on pulmonary
J Heart Lung Transplant. 2011;30(8):888–95. artery pressure and blood loss during LVAD placement.
19. Raina A, Seetha Rammohan HR, Gertz ZM, Rame JE, J Cardiothorac Vasc Anesth. 2014;28(3):652–60.
Woo YJ, Kirkpatrick JN. Postoperative right ventricu- 27. Haglund NA, Burdorf A, Jones T, et al. Inhaled milri-
lar failure after left ventricular assist device place- none after left ventricular assist device implantation.
ment is predicted by preoperative echocardiographic J Card Fail. 2015;21(10):792–7.
structural, hemodynamic, and functional parameters. 28. Kihara S, Kawai A, Fukuda T, et al. Effects of mil-
J Card Fail. 2013;19(1):16–24. rinone for right ventricular failure after left ven-
20. Fitzpatrick JR 3rd, Frederick JR, Hsu VM, et al. Risk tricular assist device implantation. Heart Vessel.
score derived from pre-operative data analysis predicts 2002;16(2):69–71.
the need for biventricular mechanical circulatory sup- 29. Tsiouris A, Paone G, Brewer RJ, Nemeh HW, Borgi J,
port. J Heart Lung Transplant. 2008;27(12):1286–92. Morgan JA. Outcomes of patients with right ventric-
21. Loforte A, Montalto A, Lilla Della Monica P,
ular failure on milrinone after left ventricular assist
Musumeci F. Simultaneous temporary CentriMag device implantation. ASAIO J. 2015;61(2):133–8.
right ventricular assist device placement in HeartMate 30. Antoniou T, Prokakis C, Athanasopoulos G, et al.
II left ventricular assist system recipients at high risk Inhaled nitric oxide plus iloprost in the setting of post-­
of right ventricular failure. Interact Cardiovasc Thorac left assist device right heart dysfunction. Ann Thorac
Surg. 2010;10(6):847–50. Surg. 2012;94(3):792–8.
22. Schmack B, Weymann A, Popov AF, et al. Concurrent 31. Critoph C, Green G, Hayes H, et al. Clinical outcomes
left ventricular assist device (LVAD) implanta- of patients treated with pulmonary vasodilators early
tion and percutaneous temporary RVAD support and in high dose after left ventricular assist device
via CardiacAssist Protek-Duo TandemHeart to pre- implantation. Artif Organs. 2016;40(1):106–14.
 Index

A Angiogenesis-based therapies, 222

Acute kidney injury (AKI), 23 angiogenic vector, 224
and ATN, 129, 132 CAD/PAD, 223
perioperative factors, 133 electrocardiographic changes, 224
preoperative factors, 132 FGF, 223
Acute myocardial infarction (AMI), 57 gene delivery, 223
Acute oliguria post continuous-flow LVAD, 136 gene therapy, 224
Acute Physiology and Chronic Health Evaluation II intracoronary vs. direct intramyocardial delivery, 224
(APACHE II), 17 intramyocardial-based trials, 224
Anesthesia, 99 limitations, 223
Anesthetic management placebo controls, 224
airway exam, 85 plasmid and protein delivery, 224
BTR, 83 VEGF, 223
BTT, 83 Angiotensin-neprilysin inhibitor (ARNi), 155
cardiac evaluation Anticoagulation
AICDs, 84 bleeding, 191
heart failure, 84 chronic management, 146
iatrogenic cardiovascular injury, 84 thrombosis, 192
intracardiac shunts and valvular pathology, 85 Aortic cusps, 169
LVIDd, 84 Aortic insufficiency, 157, 158
medications, 85 Aortic stenosis, 27
perioperative assessment, 84 Aortic valve (AV), 169–171
perioperative transfusion requirement, 84 fused, 104
PFO, 85 LVAD speed, 104
pre-LVAD implantation repair, 85 regurgitation, 105
prior sternotomy, 84 AR, assessment, 169, 172, 173
pulmonary hypertension, 84 Arteriovenous malformations, 212
RV failure, 84
consent/family discussions, 85, 86
CPB (see Cardiopulmonary bypass (CPB)) B
devices placement, 91 Baylor College of Medicine (BCM), 1
DT, 83 Beck Depression Inventory, 214
EJ reduction, 83 Biventricular assist device (BIVAD), 255
laboratory testing and imaging, 85 Biventricular heart failure, 28
LVAD implantation, 83 Blood pressure (BP), 164
minimally invasive and off-pump approaches, 91 Body mass index (BMI), 24, 25
monitoring, 86 Body surface area (BSA), 24, 25
neurologic examination, 85 Bridge-to-transplantation (BTT), 15, 35, 83
percutaneous devices, 92
preimplantation TTE, 84
preoperative laboratory testing and imaging, 85 C
renal and hepatic function, 85 Cardiac arrest, 114
TEE, 86, 88–90 Cardiac cachexia, 96
ventilation/intubation, 85 Cardiac cirrhosis, 24

© Springer International Publishing AG 2018 263

J.A. Morgan et al. (eds.), Mechanical Circulatory Support for Advanced Heart Failure,
[Link]
264 Index

Cardiac regeneration LV diastolic pressure, 200

angiogenesis therapy, 222 outflow-graft repair, 200
cardiac cellular reprogramming, 222 protamine, 202
cardiomyocyte death, 221 speed, 199
cell-based interventions, 234 vascular clamps, 200
congestive heart failure, 221 Chronic kidney disease, 132
direct (see Direct cardiac cellular reprogramming) Chronic management of patients, 148, 155–157
endogenous cardiomyocytes/stem cells, 233, 234 aortic insufficiency and valvular heart disease, 157,
end-stage coronary artery disease, 221 158
exogenous stem cell implants, 222 BTT therapy, 145
GMT, 223 CHF, 145
host syncytium, 234 echocardiography (see Echocardiography)
iCM cell, 222 end-stage heart failure, 146
iPS cells, 222 evaluation, LVEF, 150
post-infarct ventricular remodeling, 222 generation, devices, 146
pro-fibrotic gene expression, 234 heart failure (see Heart failure management)
rational myocardial, 222 HeartMate II device, 145, 146
reprogramming, 232, 233 HeartMate II and III device
research methodology, 234 alarms, 152, 153
stem cell implantation, 222, 228 parameters, 151, 152
strategies, 222 HeartMate XVE device, 145
therapeutic angiogenesis (see Angiogenesis-based HeartWare HVAD devices, 145
therapies) hemodynamic assessment, 147, 148
Cardiac reprogramming, 232, 233 HVAD device (see HVAD device)
Cardiomyocyte death, 221 hypertension (see Hypertension management)
Cardiopulmonary bypass (CPB), 48, 69, 73, 79, 81 INTERMACS data, 145
antithrombin III deficiency, 86 linear relationship, 147
cannula positioning and de-airing, 88–90 long-term placement, 145
esmolol, 86 LVAD
general anesthetic management, 88 continuous-flow, 145, 146
heparin, 86 FDA approval, 145
HR and CO, 86 implantation, 146
hypokalemia and hyperglycemia, 86 monitoring, 146
iatrogenic antithrombin III deficiency, 86 peripheral pressure and perfusion, 147
leukocyte-reduced blood, 86 PI values, 147
lidocaine and fentanyl, 86 RHF (see Right heart failure (RHF))
LVAD right heart catheterization, 149
implantation, 86 XVE device, 145
postoperative course, 98, 99 Circulatory shock, 110, 114, 115
speed determination and ventricular function, Clinical data acquisition standards, 164–166
90–91 Closed-loop system, 27
maintaining anesthetic depth, 86 Complications, 99
post-CPB implantation care, 91 Congenital heart disease, 28, 29
surgery, 86 Congestive heart failure (CHF), 97
Cardiorenal syndrome, 129, 130 Congestive hepatopathy, 134
Center for Medicare and Medicaid Services, 211 Congestive nephropathy, 129
Central venous pressure (CVP), 20, 47, 251 Continuous-flow left ventricular assist device
CentriMag devices, 205 (CF-LVAD), 43, 95, 164, 185, 186, 199–202,
Cerebral injury, 17 211, 213–215
CF-LVAD pump exchange active pump infection, 199
arm, 200 echocardiography, 199
CPB, 199 end-stage heart failure, 199
de-airing, outflow graft, 200 exchange, 199
dissection and driveline, 200 HeartMate II pump exchange, 199
excessive traction, outflow graft, 200 mechanical respiratory support, 199
heavy-duty self-retaining retractor, 200 optimization, patient’s condition, 199
hemodynamic parameters, 199 pulmonary vasodilators, 199
heparin, 200 pump exchange (see CF-LVAD pump exchange)
inflow and outflow components, 200 Swan-Ganz catheter, 199
left subcostal approach, 199 Continuous-flow pumps, 2–4, 10
left subcostal incision, 201 Continuous-flow ventricular assist devices (CF-VAD), 35
Index 265

Continuous renal replacement therapy (CRRT), 136 challenges, 228

Controller and power source, 152 cocktails, 231
Coring process, LVAD, 79, 80 cTnT, 228–229
CW spectral Doppler, 173 derepression, cardiogenic genes, 232
GMT, 228, 229, 232
iCMs, 228
D integrative lentiviral/retroviral vectors, 232
3D echocardiography (3DE), 47 Mef2c, 229
Delayed sternal closure (DSC), 121, 122 microRNA-133, 231
De novo AR after LVAD implantation, 169–171, 174 murine and human fibroblasts, 231
Destination therapy (DT), 15, 35, 83, 212, 213 MyoD domain, 229
adverse events pro-fibrotic genes, 231
infection, 212 Rho-associated kinase, 231
LVAD dysfunction, 212 somatic cell signature, 229
neurologic events, 212–213 TGF-β signaling, 232
non-neurologic bleeding, 212 Wnt signaling, 232
causes of death, 213 in vivo studies
CF-LVADs, 209 clinical benefits, 232
elderly (> 70 yo), 216, 217 ejection fraction, 232
hospital discharge and 1-year post HM XVE, 213–215 GMT, 232
hospitalization, 216 iCMs, 232
LVAD durability, 215 VEGF and FGF, 232
multivariable analysis of risk factors, 213, 214 Direct Doppler measurement of LVAD flow, 171, 174
outcome predictors, 216 Doppler-derived BP, 164
psychosocial characteristics, 216 Doppler flow assessments, 161
quality of life and functional status, 214 Driveline infections, 246, 247
recovery, 217 Driveline issue, pump exchange, 200
risk scores and calculators, 213–215 DT-LVAD patients, 214
survival, 209–211
transplant, 217
Destination therapy risk score (DTRS), 19 E
Dialysis Echocardiographic models, 46, 47
access, 140, 141 Echocardiography, 161
and ESRD care chronic management
acute kidney injury, 140 aortic outflow graft anastomosis, 150
advantages, 140 AR, 148, 149
challenges, 140 aortic valve mobility, 148, 149
education, 140 asymptomatic ambulatory patients, 151
LVAD patients, 139 Doppler velocities, 150
outpatient clinic, 140 LVAD speed, 151
renal recovery, 140 LVEF, Simpson’s method, 148
SHD, 140 measurement, LVIDd, 148
monitoring MR, 149, 150
aortic valve, 141 pulmonic valves, 150
AVF, 142 Quinones method, 148
blood pressure, 141 severity, AR, 149
CF-LVAD patients, 142 TR, 150
continuous ventricular fibrillation/tachycardia, 141 TTE, 148, 150
degree of residual LV function, 141 inflow cannula, 192
Doppler machine, 141 LVAD patients, 99–103, 105, 109–113, 115
fluid and salt intake, 141 surveillance (see Surveillance echocardiography)
hemodialysis, 142 Electrocautery, 72
mean pressure, 141 Endogenous cardiomyocytes/cardiac stem cells, 233, 234
outpatient, 141 End-stage renal disease (ESHD), 139
risk, gastrointestinal bleeding, 142 cardiovascular disease, 129
UF rate, 141 dialysis (see Dialysis)
valve opening and closing, 141 diastolic heart failure, 129
Direct cardiac cellular reprogramming LV hypertrophy, 129
in vitro studies older patients, 129
angiogenesis trials, 229 Extracorporeal life support (ECLS), 62
anti-plasticity gene, 232 Extracorporeal membrane oxygenation (ECMO), 205, 259
266 Index

F left ventricular assist device, 146

Fluid balance and perioperative renal complications, LVAD, 179
see Perioperative renal complications LVEDD slope relationships, 195
Fungal infections, 246 pulsed-wave Doppler, 151
pump thrombosis, 192
recipients, 213
G thoratec, 206
Gastrointestinal bleeding (GI), 8 trial, 16, 211
Gene therapy, heart failure, 224 HeartMate II risk score (HMRS), 19
Glomerular filtration rate (GFR), 23 HeartMate III left ventricular assist device, 146
HeartMate vented electric device (HM VE LVAD), 209
HeartMate XVE (HM XVE), 210, 212, 213
H device, 145
Heart assist devices, see Long-term complications of XVE LVAD, 35, 214
ventricular assist devices Heart transplantation, 15
Heart disease, 129 HeartWare HVAD, 206
renal complications (see Perioperative renal left ventricular assist device, 146
complications) waveform, 154
Heart failure (HF), 15, 35, 57, 84, 203, 239 HeartWare LVAD (HVAD), 167
abnormal RV, 38 alarms, 155
definition, 41 inflow-cannula artifact, 173, 177
hemodynamic parameters, 41, 42 parameters and waveform analysis, 153–155
left ventricle, 84 speeds, surveillance echocardiography, 182–183
long-term complications, devices (see Long-term Hemodynamic and LVAD function
complications of ventricular assist devices) assessments
LVAD chronic management, 147, 148
implantation, 43, 50, 52 and interventions, 101
placement, 84 cardiac tamponade, 110
management, 83 chest tube output rates, 109
ACE inhibitors, 155 controller failure, 111
ARB, 155 CT angiography, 112
ARNi, 155 differential pressure, 102
echocardiography, 155 echocardiography, 110, 112
2013 ISHLT guidelines, 155 flow rate, 102, 103
LVAD implantation, 155 HM2, 101, 102
mineralocorticoid antagonists, 155 HW HVAD, 101, 102
randomized controlled trials, 155 hypertension, 111
survival rate, 156 hypovolemia, 109
MCS (see Mechanical circulatory support (MCS)) inflow cannula, 104
mortality, 85 initial echocardiographic inspection, 104
perioperative assessment, 84 left ventricle, 104
right, 83 massive/submassive pulmonary embolism, 110
RV function, 40, 84 models, 43–46
severe, 88, 90 moderate-to-severe mitral valve regurgitation, 105
HeartMate monitoring, 101
risk score, 57 obstruction, pump inflow, 111
XVE, 253 operational parameters, 101
HeartMate I generation of LVADs (HM VE and XVE), pericardial effusion, 105
213 postoperative hemodynamic parameters, 101
HeartMate II (HMII), 7, 19, 25, 212 postoperative period, 105
CF-LVAD pump exchange, 199, 200 PTX physiology, 110
control system, 152 pulsatility index, 103
display panel, 153 pump function, 111
DT trial, 211 right ventricular failure, 109
and HMIII RV function, 103
alarms, 152, 153 septal midline position, 103
parameters, 151–152 standard echocardiography, 103
implantation, 253 therapeutic maneuvers, RV decompensation, 109
impeller thrombosis, LVAD, 165, 166 thrombosis, 112
inflow cannula, 173, 176 total artificial heart technology progresses, 101
Index 267

tricuspid valve regurgitation, 105 clinical data acquisition standards, 164

vasodilation, 110, 111 and CPB, 72
Hemopump, 4, 5 de-airing, 71
Hemostasis, 72, 73 device tunneling, 70
Heparin-induced thrombocytopenia, 119 direct Doppler measurement, 171, 174
Hepatic dysfunction, 23, 24 DT
HM VE, 212, 213 causes of death, 213
Hypertension management, 157 durability, 215
elderly (> 70 yo), 216
hospitalization, 216
I infection, 212
Iatrogenic cardiovascular injury, 84 neurologic events, 212–213
Impella, 57–65, 205 non-neurologic bleeding, 212
Implantable cardioverter-defibrillator (ICD), 253 outcome predictors, 216
Infections, ventricular assist devices psychosocial characteristics, 216
continuous-flow pumps, 246 quality of life and functional status, 214
driveline, 246, 247 recovery, 217
evidence-based interventions, 247 survival, 209–211, 213, 214
fungal, 246 suspected LVAD dysfunction, 212
LVAD implantation, 246 transplant, 217
Pseudomonas aeruginosa, 246 echocardiographic parameters, 179
pulsatile-flow pumps, 246 extraperitoneal and extra-pericardial pocket, 70
Inflow cannula flow assessment vs. native heart, 176–178
flow obstruction, 178, 180, 181 hemostasis, 72, 73
surveillance echocardiography, 173, 175–177 HM-II speeds, 179
Inhaled nitrous oxide (iNO), 48 implantation, 161–164, 203
Interagency Registry for Mechanically Assisted inflow cannula, 71, 173
Circulatory Support (INTERMACS), 17, 19, inotropic support, 48
20, 35, 57–61, 63, 145, 216, 217, 251 interventricular septal position, 171
International normalized ratio (INR), 24 long-term complications (see Long-term
The International Society for Heart and Lung complications of ventricular assist devices)
Transplantation (ISHLT), 157 long-term therapy, 203
Interventricular septal position, 171 LV diastolic function, 168–169
Interventricular septum (IVS), 42 LV size, 166–167
Intra-aortic balloon pump (IABP), 49, 59, 60 LV systolic function, 167–168
Intraoperative transesophageal echocardiography (TEE), MCS (see Mechanical circulatory support (MCS))
86–88 mediastinal exposure, 70
Intraventricular septum (IVS), 36, 99, 105, 109 minimally invasive approaches, 74, 75
Ischemic hepatitis, 24 mitral valve, 171
off-pump insertion techniques, 76
optimal speed, 183
K optimization echocardiography, 179
Kaplan-Meier analysis, 217 optimization/problem-focused (ramp) protocol
worksheet, 185
optimization/ramp echo protocol, 184
L outflow graft, 71, 72, 75, 173
Labile hemodynamics, 99 patient selection, 203–205
Left heart failure, 84 preload optimization, 47–48
Left subcostal approach, 199–202 perioperative management (see Perioperative
Left ventricular assist devices (LVAD), 1, 15, 16, 23, 35, management, LVAD patients)
83, 95, 145, 191, 203, 209–217, 239 pre-peritoneal pocket, 70
afterload reduction, 48 primary incision, 70
anesthesia (see Anesthetic management) RV size and systolic function, 169
aorta and venous system, 70 sonographer reproducibility, 164
AV, 169 sternotomy and pre-peritoneal pocket, 73
BP, 164 surgical technique
BTT, 203 apical cannulation, 81
cardiac bypass, 71 CF devices, 79
chronic management (see Chronic management CPB, 79, 81
of patients) diaphragmatic myotomy, 80
268 Index

Left ventricular assist devices (LVAD) (cont.) LV outflow tract (LVOT), 171
HeartMate XVE, 81 LV size, 166–167
implantation, 79, 81 LV systolic function, 167–168
inlet-cannula, 79, 81
MCS, 79, 81
outflow anastomosis, 80 M
silastic inflow cuff, 80 Mechanical circulatory support (MCS), 15, 16, 161, 205,
ventriculotomy, 80 206
vertical midline incision, 79 age, 25, 26
surveillance echo protocol, 179, 183 artificial heart, 1
thrombosis (see Thrombosis) axial-flow pumps, 6
tricuspid and pulmonary valves, 171 axial-flow technology, 3
Left ventricular end-diastolic volume (LVEDV), 166, 167 cellular layer, 7
Left ventricular internal dimension at end-diastole centrifugal force pump, 6
(LVIDd), 165–167 continuous-flow pumps, 2–4
Left ventricular structure and function (LVAD), 27 Hemopump, 4
Liotta artificial heart, 2 long-term
Long-term complications of ventricular assist devices, axial flow pump, thoratec HeartMate II, 206
242–245 centrifugal pump, HeartWare, 206
actuarial survival rates, HM II implantation, 244 patient selection, 17, 18
adverse event rates, 241 patient size considerations, 24, 25
bleeding psychosocial evaluation, 26, 27
centrifugal and axial continuous-flow pumps, pulsatile pumps, 2
245–246 rotary blood pumps, 7–10
continuous-flow device, 245 short-term
continuous-flow LVAD implantation, 246 CentriMag, 205
gastrointestinal, 245 development, end-organ function, 205
LVAD implantation, 245 hemodynamic instability and status, 205
minimizing, 246 IABP, 205
postoperative, 246 INTERMACS patients profiles, 205
red blood cell transfusion, 245 LVAD implantation, 205
re-exploration, 245 medical therapy, 205
surgical period, 245 reduction, preoperative risk, 205
therapeutic anticoagulation, 246 refractory pulmonary edema, 205
vWF deficiency, 245 surgical risk, 205
continuous-flow vs. pulsatile-flow LVAD therapy, TandemHeart, 205
240, 241 therapeutic strategy, 205
driveline infection, 247 VA-ECMO, 205
freedom from stroke, 243, 244 surgery training, 2
implantation, 239 surveillance echocardiography (see Surveillance
infections (see Infections, ventricular assist devices) echocardiography)
Kaplan-Meier analysis of survival, 241 Mechanical ventricular tachycardia, 178, 182
LVAD, 239 Microbubble contrast agent, 166
mechanical cardiac assist devices, 240 Mid-esophageal (ME), 47
MSC, 239 Minimally invasive approaches, 74, 75, 91
orthotopic heart transplantation, 239 Mitral regurgitation (MR) severity, 171
prevalence, heart failure, 239 Mitral valve, 171
progress, 239 Mixed venous oxygen (MVO2), 99, 101
pulsatile-flow device, 239 Model for end-stage liver disease (MELD), 24
pulsatile HeartMate XVE group, 239 Mood and anxiety disorders, 26
pump replacement and thrombosis, 240–242 Morbid obesity, 25
REMATCH trial, 239 Myocardium, 36
risk of death, 239 Myomectomy, 27
stroke (see Stroke and thromboembolic events)
treatment, 247
LV diastolic function, 168–169 N
LV ejection fraction (LVEF), 167, 168 National Heart and Lung Institute, 11
LV fractional area change (FAC) method, 168 National Heart, Lung and Blood Institute (NHLBI), 3
LV fractional shortening (%) method, 168 National Institutes of Health (NIH), 1
Index 269

Native heart vs. LVAD flow assessment, 176–178 PACU, 99

Nephropathy, 134 pathophysiology, cardiac cachexia, 96
Neurologic events, DT, 212–213 physical exams and chest x-rays, 100
Non-neurologic bleeding, 212 platelet dysfunction, 98
Normal HM-II inflow-cannula flow, 173, 176 poor oxygen delivery (DO2), 95
postoperative management, 118
preoperative evaluation, 115, 116
O pulmonary function, 97
Off-pump approaches, 91 pulmonary status, 99
Off-pump insertion techniques, 76 renal failure, 96, 97
One-size-fits-all approach, 161 review of systems (see Postoperative review of
Optimal LVAD speed, 183 systems)
Optimal medical therapy (OMT), 209, 210 RHF, 96, 97
Orthotopic heart transplantation, 239 right ventricular size and function, 104
Outflow graft (OG) risk factors, 96
anastomosis, 75 Staphylococcus aureus, 96
surveillance echocardiography, 173–177 surgery, 99
Oxygen delivery, LVAD patients, 95, 106–108 TEG R value, 117
transfusion algorithm, 118
VE assays, 117
P Perioperative renal complications, 129, 133–139
Percutaneous mechanical circulatory support (pMCS), 58 AKI/ATN, 132
Percutaneous ventricular assist devices (pVADs), 49 CF-LVAD, 132, 133
Pericardiotomy, 74 chronic and acute heart failure, 129
Perioperative hemotherapy management, 115 continuous-flow physiology, 137, 138
Perioperative management, LVAD patients, 101–112, CVP, 129
119–121 dialysis and ESRD care, 139–140
anesthesia, 99 diuretics and angiotensin II receptor blockers, 129
arrhythmias, 114, 115 DT, 133
bleeding and hemostasis, 115 ESHD (see End-stage renal disease (ESHD))
blood transfusions, 117 ESRD, 133
cardiac arrest, 114 first 48 hours
CF-LVADs, 95 abdominal vascular surgery, 134
CHF, 97 acute oliguria after continuous-flow LVAD, 136
coagulation laboratory measurements, 117 acute oliguria post continuous-flow LVAD, 136
coagulopathy, 98 bleeding, 136
complications, 95 blood flow rate, 136
continuous-flow LVAD postimplant complications, casual fluid, 134
113 congestive hepatopathy and nephropathy, 134
CPB, 98, 99 CVP, 133
decreased oxygen delivery, 106–108 dialysate flow, 136
device malfunction, 95 ECMO circuit, 137
DSC, 122 fluid overload, 134
echocardiographic assessment, 115 fluid-sparing regimen, CF-LVAD patients, 134
echocardiography parameters after device fluid therapy, ICU, 134
implantation, 102 hemodynamics, 136
hemodynamic (see Hemodynamic and LVAD high venous return, 134
function) low GFR, 134
hemostasis, 119 LVAD, 133, 134
HM2/HW HVAD implantation, 97 milrinone, 134
INTERMACS report, 95 PA pressure, 134
intraoperative transfusion algorithm, 117 potassium replacement, 137
LVAD implantation and placement, 96 renal replacement therapy, 137
MCS, 96, 97 right heart dysfunction, 136
mechanical ventilation, 100, 101 right heart function, 135
MELD score, 97 septum, 133
monitored parameters, 112 tamponade, 136
MSOF, 95 weight gain, 134
operating room, 99, 117 INTERMACS score, 129, 132
270 Index

LVAD patients Pulmonary artery systolic pressure (PASP), 21

anticoagulant, 139 Pulmonary hypertension, 19, 20, 28, 84, 156, 157
degree of LV filling, 139 Pulmonary vascular resistance (PVR), 19, 257
dialysis process, 138 Pulsatile-flow LVADs, 211, 212
Doppler pressure, 139 Pulsed and continuous-wave (CW) spectral Doppler
ESRD, 139 interrogation, 173, 179
hemodynamics, 139 Pump infection, 199
HMII, 138 Pump replacement and thrombosis
ICU, 139 anticoagulation protocols, 242
inotropes/adjust vasopressors, 138 continuous-flow LVADs, 240
limits, 138 GI bleeding, 242
LV volume, 139 HeartMate II, 240
pain medication, 138 HeartWare device, 240
palpate/auscultate, 138 heparin bridging after device implementation, 242
physiologic dry weight, 138 INTERMACS database, 240, 242
severe aortic insufficiency, 139 mechanical failure, 240
SHD, 138 monitoring, LDH levels, 242
shift dialysis, 138 outcomes, 242
SLEDD, 138 platelet glycoprotein IIb/IIIa inhibitors, 242
sodium concentration, 138 pulsatile-flow and continuous-flow devices, 240
standard pressure cuff, 138 pulsatile-flow pumps, 242
standard treatment, 139 speeds, 242
ventricular tachycardia, 139 suboptimal inflow cannula geometry, 242
LVAD placement, 129, 133 thrombin inhibitors, 242
preoperative abnormal renal function, 129 Pump thrombosis
systemic ischemia, 129 acute, 194
Perioperative transesophageal echocardiography (TEE), anticoagulation, 192
84 BMI, 192
LVAD inflow cannula placement, 89 CTA, 3D reconstruction, 192
LVAD outflow cannula placement, 90 2D echocardiography, 192
perioperative TEE protocol, 87 diagnosis
preimplantation, 84 adequate LV unloading, 195
right ventricular septal, 90 clinical factors, 193
PF-LVAD, 215 culmination, laboratory testing, 193
Phosphodiesterase-5A (PDE5), 48 formation, 193
Polytetrafluoroethylene (PTFE), 75 hemoglobin destruction, 193
Postoperative management, LVAD patients, 118 hemolysis, 194
Postoperative review of systems, LVAD implantation LDH level, 193
antibiotics, 121 LVAD, 195
arousal assessment, 119 pulsatility index, 195
assessment, 119 two-dimensional echocardiography, 194, 195
Clostridium difficile bowel infection, 121 GI bleeding, 192
consequences, hemolysis, 121 HeartMate II, 192
CVA treatment, 119 HeartMate XVE ventricular assist device, 192
delirium, 120 identification, region, 192
driveline and pump pocket infections, 121 inflow cannula, 192
early postoperative complications, 96 INTERMACS data, 192
gastrointestinal bleeding, 120 management and treatment, 195, 196
hematologic complication, 120 risk assessment, 191
hyperglycemia complications, 121 Pump’s baseline speed setting, 161
ileus, 120
liver dysfunction, 120
renal failure, 120 Q
risk, 119, 121 Quinones method, LVEF, 168
surgical wounds, 121
transient ischemic attacks, 119
vancomycin, 121 R
Pseudo AV opening, 169 Randomized Evaluation of Mechanical Assistance for the
Pulmonary arterial pressure (PAP), 43 Treatment of Congestive Heart Failure
Pulmonary arterial wedge pressure (PAWP), 19 (REMATCH) trial, 16, 18, 35, 209–217, 239
Pulmonary artery pulsatility index (PAPi), 21, 40 Renal and hepatic function, 85
Index 271

Renal arterial vasoconstriction, 23 embryonic, 228

Renal disease, see Perioperative renal complications exogenous, 228
Renal dysfunction, 23, 24 immature cardiomyocyte structures, 228
Renal replacement therapy, 136–140 iPS cells, 228
Renin-angiotensin system, 23 safety profile, 228
Right atrial pressure (RAP), 20 Stem cells, see Endogenous cardiomyocytes/cardiac stem
Right heart catheterization (RHC), 147 cells
Right heart failure (RHF), 95 Stroke and thromboembolic events
late RV Failure, 156 anticoagulation protocol, 242–243
pulmonary hypertension, 156 circulatory support, 245
Right ventricle (RV) continuous-flow group, 242
myocardium, 36 continuous-flow LVAD implantation, 242
septum, 36 continuous-flow pump implantation, 243
ventricular elastance, 36 diabetes and hypertension, 243
Right ventricular assist device (RVAD), 20 end-stage heart failure, 242
Right ventricular failure (RVF), 20–23, 84, 88, 91, HeartMate II, 243
255–258 hemorrhagic, 245
common and morbid complication, 251 heparin, 245
definition, 251–252 ischemic, 243
diagnostic indicators, 109 morbidity and mortality, pulsatile-flow VAD, 242
etiology, 252–253 optimal anticoagulation strategies, 244
HeartWare RVAD, 260 primary intracerebral hemorrhagic, 243
incidence, 253 risks factors, 243
management, 258–259 Subclavian (SCA), 49
MCS, 109 Surveillance echocardiography
morbid complication, 251 before hospital discharge, 161
patient outcomes, 251 BP, 164
poor oxygen delivery, 109 clinical data acquisition standards, 164–166
postoperative period, 109 heart and pump malfunctions, 161
predictors, 254–255 HVAD speeds, 182–183
prediction models, 109 inflow cannula, 173, 175–177
prevention initial and follow-up, 163
and management, 255 interventricular septal position, 171
postoperative, 256–258 after LVAD implantation, 161–164
preoperative, 255–256 LV diastolic function, 168–169
TandemHeart Protek Duo, 259 LV size, 166–167
VAD, 109 LV systolic function, 167–168
Right ventricular failure risk score (RVFRS), 43 mitral valve, 171
Right ventricular function after left ventricular assist native heart vs. LVAD flow assessment, 176–178
device (RFV-LVAD), 21 optimal LVAD speed, 183
ROADMAP study, 29 outflow graft, 173–177
RV assist device (RVAD), 252–254, 256–259 RV size and systolic function, 169
RV ejection fraction (RVEF), 36, 47 safety concerns, 162, 163, 178, 180–182
RV FAC, 169 sonographer reproducibility, 164–166
RVF risk score (RVFRS), 20 speed-change testing, 162, 163, 178, 180–182
RV size and systolic function, 169 surveillance examinations, 161
RV stroke work index (RVSWI), 40, 43, 46, 254 tricuspid and pulmonary valves, 171
RVOT, 177, 178 valvular assessment, 169–171
Surveillance examinations, 161, 187
Systemic vasodilation, 257
S
Seattle heart failure model (SHFM), 18
Septum-cannula interference, 258 T
Severe aortic atherosclerotic disease, 88 TandemHeart, 57–65, 256, 259
Simpson’s biplane/single-plane method, 166, 167 Tension pneumothorax (PTX) physiology, 110
Sonographer reproducibility, 164–166 Texas Heart Institute (THI), 1, 48, 58, 61
Speed-change testing, 162, 163, 178, 180–182 Thermo Cardiosystems, Inc. (TCI), 2
Stem cell implantation, heart failure Thoratec HeartMate II, 206
cardiac remodeling, 228 Thrombosis, 191, 195, 196
development, vascularization, 228 acute pump thrombosis, 194
ejection fraction, 228 anticoagulation therapy, 119, 191
272 Index

Thrombosis (cont.) Total cardiac output, 171, 175

assessment, 191 Transcutaneous energy transmission system (TETS), 9
CF-LVAD, 191 Transesophageal echocardiography (TEE), 255, 257, 260
complications, 196 Transthoracic echocardiography (TTE), 148, 161, 163,
CT angiogram with 3D reconstruction, 193 180
device, 120 Tricuspid and pulmonary valves, 171
formation, 192 Tricuspid annular plane systolic excursion (TAPSE), 21,
graft outflow, 112 39, 46, 169
guidelines, 191 Tricuspid regurgitation (TR), 171, 252
HeartMate II, 192 Tricuspid valve procedure (TVP), 28
and hemolysis, 191
INTERMACS data, 191
low-velocity inflow, 113 U
LVAD, 112, 191 US Food and Drug Administration (FDA), 4
management and treatment
hemodynamic stability, 195
hemolysis, 196 V
LV unloading, 196 Valvular assessment, AV, 169–171
sudden thrombus formation, 195 Valvular deficiencies, 88
surgical replacement, 196 Valvular heart disease, 27, 28, 157, 158
UNOS 1A transplant status, 196 Vasodilation, LVAD patients, 110, 111
postoperative outcomes, 196 Vasodilatory shock, 98
promoting pump, 121 Venoarterial extracorporeal membrane oxygenation
pump (see Pump thrombosis) (VA-ECMO), 62, 63, 205
pump rotors, 111 Ventricular elastance, 36
vigilance, 192 Ventriculotomy, 80
Thrombus formation, 27 von Willebrand Factor (VWF), 119, 245
Tissue Doppler imaging (TDI), 40 von Willebrand syndrome, 212