Endocrine Facets of Ageing.

Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

ENDOCRINE FACETS
OF AGEING
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

The Novartis Foundation is an international scienti¢c and educational

charity (UK Registered Charity No. 313574). Known until September 1997
as the Ciba Foundation, it was established in 1947 by the CIBA company
of Basle, which merged with Sandoz in 1996, to form Novartis. The
Foundation operates independently in London under English trust
law. It was formally opened on 22 June 1949.
The Foundation promotes the study and general knowledge of
science and in particular encourages international co-operation in
scienti¢c research. To this end, it organizes internationally
acclaimed meetings (typically eight symposia and allied open
meetings and 15^20 discussion meetings each year) and publishes
eight books per year featuring the presented papers and discussions
from the symposia. Although primarily an operational rather than
a grant-making foundation, it awards bursaries to young scientists
to attend the symposia and afterwards work with one of the other
participants.
The Foundation’s headquarters at 41 Portland Place, London W1N 4BN,
provide library facilities, open to graduates in science and allied disciplines.
Media relations are fostered by regular press conferences and by articles
prepared by the Foundation’s Science Writer in Residence. The Foundation
o¡ers accommodation and meeting facilities to visiting scientists and their
societies.

Information on all Foundation activities can be found at

http://www.novartisfound.org.uk
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Novartis Foundation Symposium 242

ENDOCRINE FACETS
OF AGEING

2002

JOHN WILEY & SONS, LTD

 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2
Copyright & Novartis Foundation 2002
Published in 2002 byJohn Wiley & Sons Ltd,
Ba⁄ns Lane, Chichester,
West Sussex PO19 1UD, England

National 01243 779777

International (+44) 1243 779777
e-mail (for orders and customer service enquiries): [email protected]
Visit our Home Page on http://www.wiley.co.uk
or http://www.wiley.com
All Rights Reserved. No part of this book may be reproduced, stored in a retrieval
system, or transmitted, in any form or by any means, electronic, mechanical, photocopying,
recording, scanning or otherwise, except under the terms of the Copyright, Designs and
Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency,
90 Tottenham Court Road, London,W1P 9HE, UK, without the permission in writing
of the publisher.

Other Wiley Editorial O⁄ces

John Wiley & Sons, Inc., 605 Third Avenue,
NewYork, NY 10158-0012, USA

WILEY-VCH Verlag GmbH, Pappelallee 3,

D-69469 Weinheim, Germany

Jacaranda Wiley Ltd, 33 Park Road, Milton,

Queensland 4064, Australia

John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01,
Jin Xing Distripark, Singapore 129809

John Wiley & Sons (Canada) Ltd, 22 Worcester Road,

Rexdale, Ontario M9W 1L1, Canada

Novartis Foundation Symposium 242

viii+290 pages, 46 ¢gures, 12 tables

Library of Congress Cataloging-in-Publication Data

Chadwick, Derek.
Endocrine facets of ageing / DerekJ. Chadwick, Jamie A. Goode.
p.cm. ^ (Novartis Foundation sympoisum ; 242) (Ciba Foundation symposium)
Includes bibliographical references and index.
ISBN 0-471-48636-1 (alk. paper)
1. Aging^Endocrine aspects ^Congresses. I. Goode, Jamie. II.Title. III. Series. IV.
Series: Ciba Foundation symposium
QP187.3.A34 C475 2002
612.6’7^dc21 2001046746

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

ISBN 0 471 48636 1

Typeset in 101
2 on 121
2 pt Garamond by DobbieTypesetting Limited, Tavistock, Devon.

Printed and bound in Great Britain by Biddles Ltd, Guildford and King’s Lynn.
This book is printed on acid-free paper responsibly manufactured from sustainable forestry,
in which at least two trees are planted for each one used for paper production.
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Contents

Novartis Foundation symposium on Endocrine facets ofageing, held atthe Novartis Foundation,
London, 30 January^1February 2001
Editors: Derek J. Chadwick (Organizer) and Jamie A. Goode
This symposium is based on a proposal made by Johannes D.Veldhuis and Zvi Laron

Johannes D.Veldhuis Chair’s introduction 1

Annewieke W. van den Beld and Steven W. J. Lamberts Endocrine aspects of

healthy ageing in men 3
Discussion 16

Bernard J. Carroll Ageing, stress and the brain 26

Discussion 36

Per Bj˛rntorp Alterations in the ageing corticotropic stress-response axis 46

Discussion 58

David J. Handelsman Male reproductive ageing: human fertility, androgens

and hormone dependent disease 66
Discussion 77

Christina Wang, Amiya Sinha Hikim, Monica Ferrini, Juan J. Bonavera,

DoloresVernet, Andrew Leung,Yan-He Lue, Nestor F. Gonzalez-Cadavid
and Ronald S. Swerdlo¡ Male reproductive ageing: using the Brown Norway
rat as a model for man 82
Discussion 95

J. D.Veldhuis, A. Iranmanesh,T. Mulligan and C.Y. Bowers Mechanisms of

conjoint failure of the somatotropic and gonadal axes in ageing men 98
Discussion 118

Zvi Laron E¡ects of growth hormone and insulin-like growth factor 1 de¢ciency
on ageing and longevity 125
Discussion 137
v
vi CONTENTS

Antonio Ruiz-Torres The role of insulin-like growth factor 1 and insulin in ageing
and atherosclerosis 143
Discussion 153

Henry G. Burger, Emma Dudley, Pam Mamers, David Robertson,

Nigel Groome and Lorraine Dennerstein The ageing female reproductive
axis I 161
Discussion 167

Jerilynn C. Prior The ageing female reproductive axis II: ovulatory changes with
perimenopause 172
Discussion 186

Holger Leitolf, Jens Behrends and Georg Brabant The thyroid axis in
ageing 193
Discussion 201

Greet Van den Berghe and Stephen M. Shalet Critical illness as a model of
hypothalamic ageing 205
Discussion 216

Dariush Elahi, Denis C. Muller, Josephine M. Egan, Reubin Andres,

JohannesVeldhuis and Graydon S. Meneilly Glucose tolerance, glucose
utilization and insulin secretion in ageing 222
Discussion 242

B. Lawrence Riggs Endocrine causes of age-related bone loss and

osteoporosis 247
Discussion 260

David Robertson, Jens Jordan, Giris Jacob,Terry Ketch, John R. Shannon and
Italo Biaggioni Ageing and water homeostasis 265
Discussion 275

Johannes D.Veldhuis Summing-up 279

Index of contributors 280

Subject index 282

 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Participants

Per Bj˛rntorp Department of Heart and Lung Diseases, University of

G˛teborg, Sahlgrenska University Hospital, S-413 45 G˛teborg, Sweden

Cyril Y. Bowers Department of Endocrinology,Tulane Medical School,

1430 Tulane Avenue, SL53, New Orleans, Louisiana 70112, USA

Georg Brabant Abteilung Klinische Endokrinologie, Medizinische

Hochschule Hannover, Carl-Neubergstr 1, Hanover, D-30623, Germany

Henry G. Burger Prince Henry’s Institute of Medical Research, Monash

Medical Centre, Post O⁄ce Box 5152, Clayton,VIC 3168, Australia

Bernard Carroll Paci¢c Behavioral Research Foundation, 26386 Carmel

Rancho Lane, Suite 202, PO Box 223040, Carmel, CA 93922-3040, USA

Dariush Elahi Geriatric Research Laboratory, GRBSB-0015, Massachusetts

General Hospital, 55 Fruit Street, Boston, MA 02114, USA

Andrea Giustina Department of Internal Medicine, 2A Medicina - Spedali

Civili, P Le Spedali Civili I, Brescia, Italy

David J. Handelsman ANZAC Research Institute, Department of Andrology,

Concord Hospital, University of Sydney, Sydney, NSW 2139, Australia

Erhard Haus Department of Laboratory Medicine and Pathology, University of

Minnesota, Department of Anatomic and Clinical Pathology, Regious Hospital,
640 Jackson Street, St Paul, MN 55101-2595, USA

Zvi Laron Endocrinology and Diabetes Research Unit, Schneider Children’s

Medical Center of Israel (SCMCI),Tel Aviv University, 14 Kaplan Street, 49202,
PetachTiqva, Israel
vii
viii PARTICIPANTS

John Monson Department of Medicine and Endocrinology, St Bartholomew’s

and the Royal London School of Medicine,West Smith¢eld, London
EC1A 7BE, UK

John E. Morley Division of Geriatric Medicine, St Louis University School of

Medicine, 1402 S Grand Boulevard, Room M238, St Louis, MO 63104-1083,
USA

Eugenio E. Mˇller University of Milan, Department of Pharmacology,

ViaVanvitelli 32, 10129 Milan, Italy

Jerilynn C. Prior The University of British Columbia, Department of

Endocrinology, 380-575 W 8th Avenue,Vancouver, Canada BC V5Z 1C6

B. Lawrence Riggs Division of Endocrinology/Metabolism, Mayo Clinic and

Foundation, 200 1st Street SW, Rochester, MI 80220-3706, USA

David Robertson Departments of Medicine, Pharmacology, and Neurology,

Vanderbilt University School of Medicine, AA 3228 MCN, Nashville,
TN 37232-2195, USA

Antonio Ruiz-Torres Instituto Universitario de Investigacio¤n, Gerontologica y

Metabolica, Hopital de la Princesa, Diego de Leon 62, 28006 Madrid, Spain

Stephen M. Shalet Department of Endocrinology, Christie Hospital,Wilmslow

Road, Manchester M20 4BX, UK

Annewieke van den Beld Department of Internal Medicine, University

Hospital Dijkzigt, 40 Dr Molewaterplein, 3015 GD, Rotterdam,
The Netherlands

Johannes D.Veldhuis (Chair) Division of Endocrinology , Department of

Internal Medicine, University of Virginia School of Medicine , PO Box 800202,
Charlottesville,VA 22908-0202, USA

Christina Wang General Clinical Research Center (Box 16), Harbor-UCLA

Medical Center, 1000 W Carson Street,Torrance, CA 90509, USA
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Chair’s introduction
Johannes Veldhuis

University of Virginia Health System, NIH General Clinical Research Center, Department of
Internal Medicine, Division of Endocrinology and Metabolism, PO Box 800202,
Charlottesville, Virginia 22908-0202, USA

The idea from this symposium arose from some conversations that Professor Zvi
Laron and I had about a year and a half ago, when we considered meeting in Europe
to discuss some of the issues in the ageing ¢eld that were puzzling us. From these
notions grew a more formalized structure, which I am delighted to see hosted today
by the Novartis Foundation. The glory of the format for this type of small meeting
is that it is very open to query and enquiry.
I think that it is always productive ¢rst to try to focus on some of the unresolved
issues in a ¢eld. One aspect of research philosophy that I like is that one can identify
areas of ignorance honestly: it is considered a point of brilliance to be aware of
ignorance, because one can then address the corresponding issue. Thus, this
conference will examine some unresolved issues, some of which perhaps were
thought to be fait accompli, but in fact are not clearly understood. This
foundation of fact building is very important in generating new avenues for
research. Smaller discussion groups spawn an interchange of techniques and
occasionally stimulate emergence of a new technique: this is something that has
happened for me in the past, and it has been a privilege to go away, ¢nd a
mathematician or physiologist as a collaborator and then perhaps come up with a
new method to answer a question that previously couldn’t be addressed directly.
Thus a corollary aim of this conference is to de¢ne possible areas wherein
techniques are de¢cient in ageing research to explore important queries that
remain unresolved.
A second intent of this symposium is to represent themes from many branches of
endocrinology and selected non-endocrine facets of ageing research. Ecumenism
in science promotes interdisciplinary collaboration. One of the joys of my career so
far has been viewing research areas that initially appear disparate and trying to
catalyse some interaction. This often results in synergistic outcomes and novel,
exciting insights. The present symposial format accomplishes this objective.
A third challenge in contemporary ageing research is to coalesce ‘among-axes
uniformities’ in ageing, as distinguished from the between-axes di¡erences. In a
fundamental way, clarifying how di¡erent neuroendocrine axes behave according

1
2 VELDHUIS

to some similar template as they age is going to be important. Then, identifying

their distinctions may teach us something vital.
Lastly, modern work is concerned with how neuroendocrine control systems
interact. A number of recent papers show that researchers are beginning to
examine expressly how the gonadotrophic axis in ageing is interconnected at
several levels with the somatotrophic axis. This is a somewhat obvious
connection, but I think that we all know from our neuroscience experience that
other axes interface in subtle and basic ways. In the broadest sense, axes
collaborate via common and parallel neurotransmitter pathways that drive
feedforward and feedback signalling. Interactions emerge at the target tissue,
too, and this is where one of the big challenges remains: how does one de¢ne
ageing e¡ects across axes at the distal e¡ector-site level? Most of us study just a
single axis, and this is complicated enough. Thus, to begin to integrate results
across di¡erent axes is one aim of this symposium.
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Endocrine aspects of healthy ageing

in men
Annewieke W. van den Beld and Steven W. J. Lamberts

Department of Internal Medicine, Erasmus University Hospital, Rotterdam,

The Netherlands

Abstract. Frailty is characterized by generalized weakness, impaired mobility and balance

and poor endurance. Loss of muscle strength is an important factor in the process of
frailty, and is the limiting factor for an individual’s chances of living an independent life
until death. In men, several hormonal systems show a decline in activity during ageing.
Serum bioavailable testosterone and oestradiol, dehydroepiandrosterone and its sulfate,
and growth hormone and insulin-like growth factor 1 concentrations all decrease during
ageing in men. Physical changes during ageing have been considered physiologic, but
there is evidence that some of these changes are related to this decline in hormonal
activity. Studies on hormone administration in the elderly appear to be promising.
However, until now, hormone replacement has not yet been proven to be bene¢cial and
safe.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 3^25

The concept of frailty and successful ageing

Age-related disability is characterized by generalized weakness, impaired mobility,
impaired balance and poor endurance. This state is also called ‘frailty’ and is de¢ned
as a syndrome of multi-system reduction in physiological capacity as a result of
which an older person’s function may be severely compromised by minor
environmental challenges, giving rise to the condition ‘unstable disability’
(Campbell & Buchner 1997). The increase in heterogeneity with age makes
research ¢ndings more di⁄cult to generalize, therefore it might be better for
certain research to focus on the least frail and ‘non-diseased’, which implies the
successfully aged. Older persons with minimal physiological loss, or none at all,
when compared to the average of their younger counterparts, can be regarded in
physiological terms as having aged more broadly successfully (Rowe & Kahn
1987). The concept of frailty focuses mainly on the physical or physiological
aspects of ageing, while the concept of successful ageing comprises a broader
range of aspects. Both concepts are not easy to de¢ne in a single measure.
3
4 VAN DEN BELD & LAMBERTS

Although de¢nitions of successful ageing in gerontology are numerous, there is

still no generally accepted de¢nition. Rowe and colleagues de¢ned it as including
three main components: low probability of disease and disease-related disability,
high cognitive and physical functional capacity, and active engagement with life
(Rowe & Kahn 1997). Fries, amongst others, de¢ned successful ageing as
optimizing life expectancy while simultaneously minimizing physical,
psychological and social morbidity (Fries 1988). Vaillant argued that in addition
to physical health, there are three further dimensions, or outcomes, of successful
ageing: mental health, psychosocial e⁄ciency and life satisfaction (Vaillant &
Vaillant 1990).
Since many of the predictors of the physical functional status appear to be
potentially modi¢able, research must be done to re¢ne diagnostic criteria and
elucidate practical methods of measurement of key physiological capacities in
order to identify proper targets for interventions with ‘normal’ elderly and thus
to enhance the proportion of the older population that ages successfully.
In a research we performed among 403 independently living elderly men (aged
73^94 years) from a blue-collar suburb of Rotterdam, The Netherlands, we have
measured several physical characteristics. We assessed subjective and objective
functional ability using the modi¢ed health assessment questionnaire and a
physical performance test, respectively. Furthermore, we measured bone mineral
density, body composition, muscle strength and cognition. Muscle strength was
independently, positively related to lean body mass, bone mineral density and
physical performance, and inversely related to the number of problems in
activities of daily living. Muscle strength and functional ability can be considered
the key characteristics of physical functional status in independently living elderly
men. These ¢ndings con¢rm previous studies in which it was demonstrated that
loss of muscle strength is a strong predictor of physical functional problems
(Fiatarone et al 1994, Guralnik et al 1994). The clinical correlates of sarcopenia,
the age-related loss in skeletal muscle, include falls, fractures, loss of mobility,
and development of independence in basic activities of daily living.
Successful ageing also encompasses terms like psychological well-being, quality
of life (QoL) and life satisfaction, which are all used interchangeably. QoL is
measured by subjective indicators only, while successful ageing can be measured
by both objective and subjective indicators. There are numerous general and non-
speci¢c QoL questionnaires. In our study among elderly men, we used a QoL
questionnaire recently developed by Henrich and Herschbach, which includes a
weighting for the relative importance of each dimension for the individual
concerned (Herschbach 1995, Huber et al 1988).
Part of the ageing process a¡ecting body composition (loss of muscle size and
strength, loss of bone, and increase in fat mass) might well be related to changes in
the endocrine system (Korenman et al 1990, Rudman et al 1990).
HEALTHY AGEING IN MEN 5

Endocrinology of ageing

In men, several hormonal systems show a gradual decline in activity during ageing,
represented by a decrease in their bioactive hormone concentrations. The
‘andropause’ is characterized by a gradual decline in serum total and bioavailable
testosterone, due to a decrease in testicular Leydig cell numbers and in their
secretory capacity, as well as by an age-related decrease in episodic and stimulated
gonadotropin secretion (Vermeulen 1991). Both cross-sectional (Vermeulen 1991)
and longitudinal (Morley et al 1997) studies have shown that in healthy males mean
serum total testosterone (T) levels decrease by about 30% between age 25 and 75,
whereas mean serum free T levels decrease by as much as 50% over the same period.
The steeper decline of free T levels is explained by an age-associated increase in sex-
hormone binding globulin (SHBG) binding capacity (Vermeulen & Verdonck
1972). Con£icting results have been reported concerning the question of whether
luteinizing hormone (LH) increases with age or remains relatively stable (Morley
et al 1997, Ongphiphadhanakul et al 1995, van den Beld et al 1999). One reason may
be that the ageing-induced decrease in T is primarily testicular in some men, mainly
due to hypothalamo-pituitary insu⁄ciency in others, and of mixed origin in a third
group. In our study we found a signi¢cant increase of LH with age and an inverse
relationship between serum LH and testosterone concentrations.
It has recently become clear that not only T decreases with age, but that serum
oestradiol (E2) also signi¢cantly decreases in ageing males (Ferrini & Barrett-
Connor 1998, Khosla et al 1998). In our population of elderly men, a signi¢cant
decrease in serum oestradiol levels was also observed, while serum oestrone (E1)
decreased to an even greater extent. In normal men small amounts of oestradiol are
derived from direct secretion by the testes and indirectly from adrenal androgens.
Most E2, however, is formed from testicular androgens by peripheral
aromatization of T to E2 (MacDonald et al 1979).
The second hormonal system demonstrating age-related changes is the circu-
lating levels of dehydroepiandrosterone (DHEA) and its sulfate (DHEAS), which
gradually decline resulting in ‘adrenopause’ (Ravaglia et al 1996, Herbert 1995). At
age 30, DHEAS levels are approximately ¢ve times higher than at age 85. The
decline in DHEA(S) levels contrasts with the maintenance of plasma cortisol
concentrations at the same level, and seems to be caused by a selective decrease in
the number of functional zona reticularis cells in the adrenal cortex rather than
regulated by a central (hypothalamic) pacemaker of ageing (Herbert 1995).
The third endocrine system that gradually declines in activity with ageing is the
growth hormone (GH)/insulin-like growth factor 1 (IGF1) axis (Corpas et al
1993). Mean pulse amplitude, duration, and free fraction of GH secreted, but not
pulse frequency, gradually decrease during ageing. In parallel, there is a
progressive fall in circulating IGF1 levels (Corpas et al 1993). The IGF1
6 VAN DEN BELD & LAMBERTS

reduction probably results from reduced stimulation of the liver to produce IGF1
rather than an age-related insensitivity or inability of the liver to respond to
circulating GH. The predominant IGF-binding protein (IGFBP) concentration
in the blood, IGFBP3, reaches maximum levels at puberty and decreases between
18 and 79 years (Corpas et al 1993). The second most abundant IGFBP, IGFBP2,
decreases after birth until puberty, after which it gradually increases again,
especially after the age of 60 (Clemmons 1997). At age 80 concentrations are
nearly twice as high as in young adults. In agreement with this, our study in
subjects aged between 73 and 94 years showed a decrease in serum IGF1 and
IGFBP3 levels and an increase in both serum IGFBP1 and IGFBP2 levels.

Clinical consequences of the decline in activity of the hormonal systems

Andropause
Testosterone has long been known for its anabolic e¡ects (Brodsky et al 1996).
Muscle weakness, anaemia, lowered bone mass, and mood disturbances rapidly
normalize in mid-adult hypogonadal men during T replacement therapy. Since
the decrease in serum T concentrations occurs in parallel with the decrease in
muscle mass, strength and bone mass, it has been suggested that these are
causatively related. Several cross-sectional and longitudinal studies have
demonstrated relationships in older men between serum T levels and muscle
strength, changes in body composition and bone mineral density (Rudman &
Shetty 1994, Murphy et al 1993, Ongphiphadhanakul et al 1995). In agreement,
in a cross-sectional study of 403 elderly men, we found positive relationships
between non-SHBG-bound T and muscle strength and bone mineral density,
and an inverse relationship with fat mass. A cross-sectional study among 856
elderly men demonstrated that bioavailable T levels were signi¢cantly lower for
the 25 men with categorically de¢ned depression than levels observed in all other
men (Barrett-Connor et al 1999). This suggests that T treatment might improve
depressed mood in older men who have low levels of bioavailable testosterone.
On the other hand, in the population of elderly men we studied, no signi¢cant
relation was observed between general life satisfaction and T levels. However,
non-SHBG-bound T was positively related to some separate questions of the
questionnaire, like the satisfaction related to mobility, physical performance and
independency.
It is not known whether T therapy in older men has bene¢cial e¡ects on muscle
function, sexual function, sense of well-being and quality of life, and whether this
therapy can be done safely. Studies involving physiological replacement therapy of
testosterone in older men are limited. However, the results of these studies are
consistent; reported are treatment-related declines in body fat mass ranging from
HEALTHY AGEING IN MEN 7

6.4^14%, and increases in lean mass ranging from 3.2^5.0% are both reported
(Tenover 1998, Snyder et al 1999a). Several studies evaluating muscle strength in
older men during testosterone replacement have demonstrated a statistically
signi¢cant increase in strength with therapy (Tenover 1998, Sih et al 1997).
However, Snyder et al (1999a) evaluated lower extremity strength in a double-
blind, placebo-controlled study involving 108 men, and did not demonstrate a
signi¢cant change compared to the placebo group, nor did they observe an
increase in physical performance in response to testosterone (Snyder et al 1999a).
In the same study, testosterone replacement did not increase lumbar spine bone
density in the overall group, but it did in men with low pretreatment serum
testosterone concentrations (Snyder et al 1999b). It has to be mentioned,
however, that the men in this study did not have very low T concentrations.
Although it has been thought for years that T administration adversely a¡ects
serum lipid concentrations, recent research shows that T probably has a favourable
e¡ect on total and LDL cholesterol concentration, and probably also on
triglycerides concentration (Tenover 1992). So far it is unclear whether T
replacement in older men can be considered safe with regard to cardiovascular
risk. In young men, T administration generally decreases HDL cholesterol levels,
while LDL and triglycerides levels remain unchanged. Physiological doses of
androgens in elderly men reduced total and LDL cholesterol and had no e¡ect on
HDL cholesterol (Zgliczynski et al 1996).
Additionally, the e¡ect of T supplementation on the prostate remains unclear. It
has been claimed that about 50% of men aged over 50 years have a subclinical
prostate carcinoma, which can be activated by T administration (Jackson et al
1989). However, Snyder et al (1999b) reported that T treatment was associated
with a small increase in mean serum prostate-speci¢c antigen concentrations, but
not with increases in other parameters that re£ect prostate disease (Snyder et al
1999b). Furthermore, it is known that T increases haematocrit within the normal
range (Sih et al 1997), while a negative role in the sleep apnoea syndrome has been
reported in a few individuals (Matsumoto et al 1985). Overall, the bene¢cial e¡ects
of T replacement in older men seem to be promising, and T administration can
probably be given safely to elderly men, provided they are monitored frequently.
However, since only few double-blind placebo-controlled studies have been done
on T replacement in the elderly, more research is necessary to de¢ne dose, form and
subjects for T therapy.
In our study of 403 elderly men, serum LH levels were inversely related to
muscle strength and lean mass, and positively to self-reported disability (van den
Beld et al 1999). All relationships were independent of T, suggesting that LH
re£ects the serum androgen activity in a di¡erent manner than T, possibly more
closely re£ecting the combined feedback e¡ect of T, dihydrotestosterone and
oestrogen.
8 VAN DEN BELD & LAMBERTS

In most women, the period of decline in oestrogens during menopause is

accompanied by vasomotor reactions, depressed mood, and changes in skin and
body composition (increase in body fat and decrease in muscle mass). In the
subsequent years, the loss of oestrogen is followed by a high incidence of
cardiovascular disease, loss of bone mass and cognitive impairment (Lindsay et al
1996). Only recently has it become evident that oestrogens may not only play an
important role in regulating bone turnover in women, but also in men. Smith et al
(1994) described a male with a homozygous mutation in the oestradiol receptor
gene who, even in the presence of normal T levels, had unfused epiphysis and
marked osteopenia, along with elevated indexes of bone turnover. A few studies
now have demonstrated signi¢cant relations between serum (bioavailable)
oestradiol levels and bone mineral density in elderly men (Khosla et al 1998).
In normal men small amounts of oestradiol (15% of the total daily production)
are derived from direct secretion from the testis; about equal amounts of oestradiol
are synthesized in adipose tissue, muscle, skin, breast and liver, as well as indirectly
from adrenal androgens via the conversion of androstenedione to oestrone to
oestradiol or by peripheral aromatization of testosterone to oestradiol
(MacDonald et al 1979). Bioavailable oestradiol decreases dramatically with age
in community-dwelling men, independent of body size, health and chronic
disease (Ferrini & Barrett-Connor 1998). In our population of elderly men, we
also found strong positive relations between serum oestradiol and bone mineral
density. In addition, however, we observed a strong positive relation between
serum oestradiol concentrations and life satisfaction. It remains to be examined
whether oestradiol has a central action on the brain, or whether this relationship
represents an indirect e¡ect of oestradiol on physical characteristics, which in turn
lead to a better quality of life. Only a few studies have examined the relation
between oestrogens and atherosclerosis in men (Price et al 1997). Oestrogens may
o¡er some degree of protection against cardiovascular disease by in£uencing the
lipid pro¢le (Bagatell et al 1994, Giri et al 1998). The relationship between this
decline in endogenous oestradiol levels and fragility, impaired functioning and
chronic diseases (such as osteoporosis, diabetes, cancer, prostate and heart
disease) should be the focus of future research.

Adrenopause
DHEA in its free, sulfated and lipoidal forms is the most abundant steroid secreted
by the zona reticularis of the adult human adrenal. Circulating DHEAS levels in
healthy adults are more than 10 times higher than those of cortisol (Ravaglia et al
1996). It is well known that ageing is associated with a marked decrease of the
adrenal androgens DHEA and its sulfate (Orentreich et al 1984, Labrie et al
1997). DHEA is a universal precursor for androgenic and oestrogenic steroid
HEALTHY AGEING IN MEN 9

formation in peripheral tissues, which contain a number of DHEA-metabolizing

enzymes (Herbert 1995). A variety of factors in£uence cortisol and DHEA levels:
obesity, meals, insulin, stress, alcohol and smoking all alter adrenal steroid levels.
DHEA has been called ‘the Fountain of Youth’. However, despite the
abundance of DHEA and DHEAS in human serum, it remains unclear whether
they play a functional role in the ageing process, either directly or through
conversion into other steroids. Our knowledge of the functions of adrenal
hormones is mainly derived from animal studies. Studies in rodents, which have
very low circulating DHEAS levels, suggest that DHEA administration prevents
obesity, diabetes mellitus, cancer and heart disease, while it enhances immune
function. Supportive data in humans are limited, highly controversial and as yet
unresolved.
Functional parameters of daily living in males over 90 years old were lowest in
those with the lowest DHEAS levels (Ravaglia et al 1996). In a group of
independent community-dwelling older men, men in the highest quartile of
serum DHEAS concentrations were leaner, more ¢t and had a favourable lipid
pro¢le compared with those in the lowest quartile (Abbasi et al 1998). In our
population of independently living men, a signi¢cant relationship between serum
DHEAS and bone mineral density became non-signi¢cant after adjustment for
oestradiol or testosterone, suggesting that any potential e¡ect on bone mineral
density is indirect through the conversion of DHEAS into androgens and
oestrogens. Serum DHEA(S) levels in the same population were not related to
self-reported disability, physical performance, muscle strength or body
composition, nor to quality of life.
Administration of 50 mg DHEA to elderly men leads to serum DHEAS
concentrations similar to those in healthy adults, while circulating serum
oestrogens signi¢cantly increase (Arlt et al 1999). This DHEA-induced increase
in oestrogenic activity may contribute to the bene¢cial e¡ects of DHEA in men.
Two randomized placebo-controlled studies support the concept that oral
administration of DHEA has bene¢cial e¡ects (Baulieu 1995). Three months of
daily treatment with 50 mg of DHEA in 20 adults, most of whom were non-
elderly individuals (40^70 years old), increased DHEA(S) levels to young adult
levels, increased plasma androgen and IGF1 concentrations, and induced a
remarkable increase in perceived physical and psychological well-being in both
sexes without an e¡ect on libido. In a subsequent study, 100 mg of DHEA given
daily for 6 months to 9 men and 10 women increased lean body mass in both sexes
but increased muscle strength in men only (Morales et al 1998). In a randomized,
double-blind, placebo-controlled trial among 140 elderly men and 140 elderly
women (Baulieu et al 2000) it appeared that 50 mg DHEA a day for one year led
to changes in bone mineral density in women but not in men. In addition, no e¡ect
of DHEA administration on vascular properties in elderly men was observed. A
10 VAN DEN BELD & LAMBERTS

nine month cross-over study in 39 elderly men, did not con¢rm the e¡ects of the
drugs publicized by others, such as on the sense of well-being (Flynn et al 1999).
DHEA might in£uence CNS activity (Wolf et al 1998), although short-term
DHEA replacement does not appear to improve cognition, memory, mood or
well-being (Wolf et al 1997, 1998).
Higher DHEAS levels are accompanied by a modestly reduced risk of death
from cardiovascular disease in males (Barrett-Connor & Goodman-Gruen 1995,
Feldman et al 1998). Data with regard to a protective e¡ect of DHEA on
cardiovascular disease in humans are con£icting and unresolved; studies using
animal models, however, are quite promising. Animal studies also show that
DHEA reverses the immuno-incompetence of aged animals. One study in
humans indeed shows an activation of the immune system in elderly men.
DHEA administration increased the number of monocytes and natural killer
cells, and the functional activation of T cells (Khorram et al 1997).
Although neither Flynn et al (1999) or Baulieu et al (2000) demonstrated a
change in prostate speci¢c antigen concentration after DHEA replacement, it is
so far unknown whether the increase in sex steroid levels induced by DHEA is
safe with regard to the development of prostate, or other types of cancer.
DHEA has received great attention from the general population as potentially
being able to increase the sense of well-being and sexual function, and to partially
reverse the ageing process. So far, however, results are neither clear nor consistent.
It seems prudent to await more trials, in order to be certain that exogenous DHEA
is bene¢cial and can be used safely.

Somatopause
Growth hormone (GH) is an important anabolic hormone with stimulatory e¡ects
on protein synthesis and on lipolysis. In man both ageing and GH de¢ciency are
associated with reduced protein synthesis, decreases in lean body mass and bone
mass, and increases in body fat (Corpas et al 1993). In several studies of healthy
individuals of a broad age range, an association was observed between the
maximum aerobic capacity and circulating IGF1 levels (Papadakis et al 1995).
Within the elderly population, however, cross-sectional studies have
demonstrated weak or no correlations between measures of body composition
and bone mass on the one hand, and serum IGF1 or GH levels on the other
(Goodman-Gruen & Barrett-Connor 1997, Rudman & Shetty 1994). Similarly,
largely negative ¢ndings have been reported concerning the association between
IGF1 and muscle strength, and measures of functional ability (Papadakis et al 1995,
Welle et al 1996). In our study in elderly men, we also failed to demonstrate a
signi¢cant relationship between IGF1 and IGFBP3, and measures of physical
functional status. We did, however, demonstrate strong inverse relationships
HEALTHY AGEING IN MEN 11

between serum IGFBP2 concentrations and physical performance, muscle

strength, lean body mass, fat mass and bone mineral density. A positive
relationship was observed between IGFBP2 and self-reported disability. Serum
IGFBP2 levels in an individual represent an integrative response to the
nutritional state, 24 h GH-secretion, serum IGF1 and IGF2 concentration and, to
a lesser extent, serum insulin concentrations (Clemmons 1997). Therefore low
IGFBP2 levels might serve as an indicator of better overall functional physical
status. As mentioned above, we did not ¢nd an association between the physical
characteristics of ageing and serum IGFBP3 levels, which are mainly regulated by
GH and therefore serve as an indicator of serum GH levels. We did, however, ¢nd a
positive relationship between serum IGFBP3 levels and quality of life, measured
with the questionnaire developed by Herschbach and Huber (Herschbach 1995,
Huber et al 1988). In a study of 52 elderly men and 80 elderly women, positive
correlations were observed between HDL cholesterol and IGFBP3
concentrations, which served as an index of GH secretion (Ceda et al 1998).
The expectation that somatopause contributes to the decline of functional
capacity in the elderly is mainly derived from studies in which GH replacement
therapy of GH-de¢cient adults was shown to increase muscle mass, muscle
strength, bone mass and quality of life. A bene¢cial e¡ect on the lipid pro¢le and
an important decrease in fat mass were also observed in such patients (Salomon et al
1989). The GH/IGF1 axis may also play a role in the age-related decline of certain
cognitive functions (Aleman et al 1999).
GH administration for 3 to 6 months to healthy elderly individuals increased
IGF1 levels to those observed in control individuals half their age, while muscle
strength, muscle mass, skin thickness and bone mineral content signi¢cantly
increased and fat mass decreased (Rudman et al 1990, Welle et al 1996).
Unfortunately, GH replacement in 52 healthy older men did not improve muscle
strength and functional ability (Papadakis et al 1996). If GH was administered in
combination with resistance exercise training, however, a signi¢cant positive e¡ect
on muscle mass, muscle strength and bone mineral density was recorded but did
not di¡er from that seen with placebo treatment, which suggests that GH does not
add to the bene¢cial e¡ects of exercise (Yarasheski et al 1995, 1997). Preliminary
results of a randomized placebo-controlled trial of 6 weeks GH administration in
elderly individuals with an acute hip fracture indicate that in individuals over 75
years old, GH administration causes a statistically signi¢cant earlier return to
independent living after the fracture. Finally, it has to be mentioned that side
e¡ects of GH therapy are common, namely carpal tunnel syndrome,
gynaecomastia and hyperglycaemia.
Other components in the regulation of the GH/IGF1 axis are e¡ective in
activating GH and IGF1 secretion. Long-acting derivates of the hypothalamic
peptide growth hormone-releasing hormone (GHRH), given twice daily
12 VAN DEN BELD & LAMBERTS

subcutaneously for 14 days to 70 year old healthy men, increased GH and IGF1
levels to those encountered in 35 year olds (Corpas et al 1992). These studies
support the concept that somatopause is primarily hypothalamically driven and
that pituitary somatotrops retain their capacity to synthesize and secrete high
levels of GH. GH-releasing peptides (GHRPs) are oligopeptides with even more
powerful GH-releasing e¡ects. Originally developed by design, it has recently been
suggested that they mediate their GH-secretory e¡ects through endogenous
speci¢c receptors (Howard et al 1996). Non-peptide analogues such as MK-677
and L-692,429 have powerful GH-releasing e¡ects, restoring IGF1 secretion in
the elderly to levels encountered in young adults (Chapman et al 1996). Long-
term oral administration of MK-677 to healthy elderly individuals increased lean
body mass but not muscle strength. If proven to be GH-speci¢c, these orally active
GHRP derivates might be important alternatives to subcutaneously administered
GH in the reversal of somatopause, the prevention of frailty, and the reversal of
acute catabolism. The long-term safety of the activation of GH/IGF1 levels
remains uncertain with regard to tumour growth, as most human solid cancers
express IGF1 receptors (Chapman et al 1996).

Andro-, adreno- and somatopause

Some of the changes in hormones concentrations of the di¡erent hormonal axes
which take place during ageing occur in parallel. However, it is not known
whether interrelations exist between the hormones of the di¡erent axes, or
whether they independently in£uence the physiological changes during ageing.
Interactions between these hormonal systems may play a physiologically relevant
role in the elderly population in the management of age-associated alterations in
physical performance, muscle strength and body composition. Compared with
healthy young men, 80% of healthy young men are hyposomatomedinaemic
(lowered IGF1 levels) and hypogonadal (lowered testosterone levels); compared
with healthy old men, 30% of chronically institutionalized old men are both
hyposomatomedinaemic and hypogonadal. In institutionalized old men,
hyposomatomedinia and hypogonadism are usually of central origin, but their
occurrences are not signi¢cantly associated (Abbasi et al 1993). Although few
relationships were present between hormones of the di¡erent axes in the study of
403 independently living elderly men, the associations between hormone
concentrations and physical characteristics were all independent. This implies
that subjects with low IGF1/GH levels do not necessarily have low androgens or
adrenal hormones. This suggests that changes in hormones concentrations of the
di¡erent axes in£uence the ageing process independently, which is extremely
important for the selection of subjects for the respective replacement therapies.
HEALTHY AGEING IN MEN 13

Conclusions
In elderly men, ageing is accompanied by a decrease in objective and subjective
physical function, as well as changes in body composition and bone mass. In
parallel, important changes in the endocrine system occur. Testosterone
administration seems to in£uence bone mass and perhaps muscle strength,
although insu⁄cient evidence exists to con¢rm that testosterone substitution in
the elderly is indicated. Also, oestradiol seems to in£uence bone mineral density
in elderly men. Although DHEA and its sulfate have been regarded as the
hormone of youth, studies performed so far do not yet indicate a role for these
adrenal hormones in maintaining physical functional status. DHEA,
nevertheless, might have bene¢cial e¡ects on cardiovascular and immunological
processes. It remains unclear how hormones of the somatotropic axis contribute
to the ageing process. Traditionally, the ageing process has been considered
physiological and unavoidable. In recent years, however, it has become evident
that it might not be necessary to accept the grim stereotype of ageing as an
unalterable process of decline and loss (Fiatarone et al 1994, Rowe & Kahn
1987). When su⁄cient research has been done on hormonal replacement therapy
in elderly men, hormonal substitution might be used to delay the ageing process
and to allow us to live for a longer period in a (relatively) independent state of
successful ageing.

References
Abbasi AA, Drinka PJ, Mattson DE, Rudman D 1993 Low circulating levels of insulin-like
growth factors and testosterone in chronically institutionalized elderly men. J Am Geriatr
Soc 41:975^982
Abbasi A, Duthie EH Jr, Sheldahl L et al 1998 Association of dehydroepiandrosterone sulfate,
body composition, and physical ¢tness in independent community-dwelling older men and
women. J Am Geriatr Soc 46:263^273
Aleman A, Verhaar HJ, De Haan EH et al 1999 Insulin-like growth factor-I and cognitive
function in healthy older men. J Clin Endocrinol Metab 84:471^475
Arlt W, Haas J, Callies F et al 1999 Biotransformation of oral dehydroepiandrosterone in elderly
men: signi¢cant increase in circulating estrogens. J Clin Endocrinol Metab 84:2170^2176
Bagatell CJ, Knopp RH, Rivier JE, Bremner WJ 1994 Physiological levels of estradiol stimulate
plasma high density lipoprotein2 cholesterol levels in normal men. J Clin Endocrinol Metab
78:855^861
Barrett-Connor E, Goodman-Gruen D 1995 The epidemiology of DHEAS and cardiovascular
disease. Ann N Y Acad Sci 774:259^270
Barrett-Connor E, von Mˇhlen DG, Kritz-Silverstein D 1999 Bioavailable testosterone and
depressed mood in older men: the Rancho Bernardo Study. J Clin Endocrinol Metab
84:573^577
Baulieu EE 1995 Studies on dehydroepiandrosterone (DHEA) and its sulphate during aging.
C R Acad Sci III 318:7^11
14 VAN DEN BELD & LAMBERTS

Baulieu EE, Thomas G, Legrain S et al 2000 Dehydroepiandrosterone (DHEA), DHEA sulfate,

and aging: contribution of the DHEAge study to a sociobiomedical issue. Proc Natl Acad Sci
USA 97:4279^4284
Brodsky IG, Balagopal P, Nair KS 1996 E¡ects of testosterone replacement on muscle mass and
muscle protein synthesis in hypogonadal men
a clinical research center study. J Clin
Endocrinol Metab 81:3469^3475
Campbell AJ, Buchner DM 1997 Unstable disability and the £uctuations of frailty. Age Ageing
26:315^318
Ceda GP, Dall’Aglio E, Magnacavallo A et al 1998 The insulin-like growth factor axis and
plasma lipid levels in the elderly. J Clin Endocrinol Metab 83:499^502
Chapman IM, Bach MA, Van Cauter E et al 1996 Stimulation of the growth hormone (GH)-
insulin-like growth factor I axis by daily oral administration of a GH secretogogue (MK-677)
in healthy elderly subjects. J Clin Endocrinol Metab 81:4249^4257
Clemmons DR 1997 Insulin-like growth factor binding proteins and their role in controlling
IGF actions. Cytokine Growth Factor Rev 8:45^62
Corpas E, Harman SM, Pineyro MA, Roberson R, Blackman MR 1992 Growth factor (GH)-
releasing hormone-(1-29) twice daily reserves the decreased GH and insulin-like growth
factor-I levels in old men. J Clin Endocrinol Metab 75:530^535
Corpas E, Harman SM, Blackman MR 1993 Human growth hormone and human aging. Endocr
Rev 14:20^39
Feldman HA, Johannes CB, McKinlay JB, Longcope C 1998 Low dehydroepiandrosterone
sulfate and heart disease in middle-aged men: cross-sectional results from the Massachusetts
Male Aging Study. Ann Epidemiol 8:217^228
Ferrini RL, Barrett-Connor E 1998 Sex hormones and age: a cross-sectional study of testosterone
and estradiol and their bioavailable fractions in community-dwelling men. Am J Epidemiol
147:750^754
Fiatarone MA, O’Neill, Ryan ND et al 1994 Exercise training and nutritional supplementation
for physical frailty in very elderly people. N Engl J Med 330:1769^1775
Flynn MA, Weaver-Osterholtz D, Sharpe-Timms KL, Allen S, Krause G 1999 Dehydro-
epiandrosterone replacement in aging humans. J Clin Endocrinol Metab 84:1527^1533
Fries JF 1988 Aging, illness, and health policy: implications of the compression of morbidity.
Perspect Biol Med 31:407^428
Giri S, Thompson PD, Taxel P et al 1998 Oral estrogen improves serum lipids, homocysteine
and ¢brinolysis in elderly men. Atherosclerosis 137:359^366
Goodman-Gruen D, Barrett-Connor E 1997 Epidemiology of insulin-like growth factor-I in
elderly men and women. The Rancho Bernardo Study. Am J Epidemiol 145:970^976
(erratum: 1997 Am J Epidemiol 146:357)
Guralnik JM, Simonsick EM, Ferrucci L et al 1994 A short physical performance battery
assessing lower extremity function: association with self-reported disability and prediction
of mortality and nursing home admission. J Gerontol 49:M85^M94
Herbert J 1995 The age of dehydroepiandrosterone. Lancet 345:1193^1194
Herschbach P 1995 The di¡erence between ill persons and patients (translated from German).
Psychother Psychosom Med Psychol 45:83^89
Howard AD, Feighner SD, Cully DF et al 1996 A receptor in pituitary and hypothalamus that
functions in growth hormone release. Science 273:974^977
Huber D, Henrich G, Herschbach P 1988 Measuring the quality of life: a comparison between
physically and mentally chronically ill patients and healthy persons. Pharmacopsychiatry
21:453^455
Jackson JA, Waxman J, Spiekerman AM 1989 Prostatic complications of testosterone
replacement therapy. Arch Intern Med 149:2365^2366
HEALTHY AGEING IN MEN 15

Khorram O, Vu L, Yen SS 1997 Activation of immune function by dehydroepiandrosterone

(DHEA) in age-advanced men. J Gerontol A Biol Sci Med Sci 52:M1^M7
Khosla S, Melton LJ III, Atkinson EJ, O’Fallan WM, Klee GG, Riggs BL 1998 Relationship of
sex steroid levels and bone turnover markers with bone mineral density in men and women: a
key role for bioavailable estrogen. J Clin Endocrinol Metab 83:2266^2274
Korenman SG, Morley JE, Mooradian AD et al 1990 Secondary hypogonadism in older men: its
relation to impotence. J Clin Endocrinol Metab 71:963^969
Labrie F, Be¤ langer A, Cusan L, Candas B 1997 Physiological changes in dehydroepiandrosterone
are not re£ected by serum levels of active androgens and estrogens but of their metabolites:
intracrinology. J Clin Endocrinol Metab 82:2403^2409
Lindsay R, Bush TL, Grady D, Spero¡ L, Lobo RA 1996 Therapeutic controversy: estrogen
replacement in menopause. J Clin Endocrinol Metab 81:3829^3838
MacDonald PC, Madden JD, Brenner PF, Wilson JD, Siiteri PK 1979 Origin of estrogen in
normal men and in women with testicular feminization. J Clin Endocrinol Metab 49:905^916
Matsumoto AM, Sandblom RE, Schoene RB et al 1985 Testosterone replacement in
hypogonadal men: e¡ects on obstructive sleep apnoea, respiratory drives, and sleep. Clin
Endocrinol 22:713^721
Morales AJ, Haubrich RH, Hwang JY, Asakura H, Yen SS 1998 The e¡ect of six months
treatment with 100 mg daily dose of dehydroepiandrosterone (DHEA) on circulating sex
steroids, body composition and muscle strength in age-advanced men and women. Clin
Endocrinol 49:421^432
Morley JE, Kaiser F, Raum WJ et al 1997 Potentially predicative and manipulative blood serum
correlates of aging in the healthy human male: progressive decreases in bioavailable
testosterone dehydroepiandrosterone sulfate, and the ratio of insulin-like growth factor 1 to
growth hormone. Proc Natl Acad Sci USA 94:7537^7542
Murphy S, Khaw KT, Cassidy A, Compston JE 1993 Sex hormones and bone mineral density in
elderly men. Bone Miner 20:133^140
Ongphiphadhanakul B, Rajatanavin R, Chailurkit L et al 1995 Serum testosterone and its
relation to bone mineral density and body composition in normal males. Clin Endocrinol
43:727^733
Orentreich N, Brind JL, Rizer RL, Vogelman JH 1984 Are changes and sex di¡erences in serum
dehydroepiandrosterone sulfate concentrations throughout adulthood. J Clin Endocrinol
Metab 59:551^555
Papadakis MA, Grady D, Tierney MJ, Black D, Wells L, Grunfeld C 1995 Insulin-like growth
factor 1 and functional status in healthy older men. J Am Geriatr Soc 43:1350^1355
Papadakis MA, Grady D, Black D et al 1996 Growth hormone replacement in healthy older men
improves body composition but not functional ability. Ann Intern Med 124:708^716
Price JF, Lee AJ, Fowkes FG 1997 Steroid sex hormones and peripheral arterial disease in the
Edinburgh Artery Study. Steroids 62:789^794
Ravaglia G, Forti P, Maioli F et al 1996 The relationship of dehydroepiandrosterone slufate
(DHEAS) to endocrine-metabolic parameters and functional status in the oldest-old.
Results from an Italian study on healthy free-living over-ninety-year-olds. J Clin
Endocrinol Metab 81:1173^1178
Rowe JW, Kahn RL 1987 Human aging: usual and successful. Science 237:143^149
Rowe JW, Kahn RL 1997 Successful aging. Gerontologist 37:433^440
Rudman D, Shetty KR 1994 Unanswered questions concerning the treatment of
hyposomatotrophism and hypogonadism in elderly men. J Am Geriatr Soc 42:522^527
Rudman D, Feller AG, Nagraj HS et al 1990 E¡ects of growth hormone in men over 60 years old.
N Engl J Med 323:1^6
16 DISCUSSION

Salomon F, Cuneo RC, Hesp R, Sonksen PH 1989 The e¡ects of treatment with recombinant
human growth hormone on body composition and metabolism in adults with growth
hormone de¢ciency. N Engl J Med 321:1797^1803
Sih R, Morley JE, Kaiser FE, Perry HM III, Patrick P, Ross C 1997 Testosterone replacement in
older hypogonadal men: a 12-month randomized controlled trial. J Clin Endocrinol Metab
82:1661^1667
Smith EP, Boyd J, Frank GR et al 1994 Estrogen resistance caused by a mutation in the estrogen-
receptor gene in man. N Engl J Med 331:1056^1061
Snyder PJ, Peachey H, Hannoush P et al 1999a E¡ect of testosterone treatment on body
composition and muscle strength in men over 65 years of age. J Clin Endocrinol Metab
84:2647^2653
Snyder PJ, Peachey H, Hannoush P et al 1999b E¡ect of testosterone treatment on bone mineral
density in men over 65 years of age. J Clin Endocrinol Metab 84:1966^1972
Tenover JS 1992 E¡ects of testosterone supplementation in the aging male. J Clin Endocrinol
Metab 75:1092^1098
Tenover JS 1998 Androgen replacement therapy to reverse and/or prevent age-associated
sarcopenia in men. Baillie' re’s Clin Endocrinol Metab 12:419^425
Vaillant GE, Vaillant CO 1990 Natural history of male psychological health, XII: a 45-year study
of predictors of successful aging at age 65. Am J Psychiatry 147:31^37
van den Beld A, Huhtaniemi I T, Pettersson KS et al 1999 Luteinizing hormone and di¡erent
genetic variants, as indicators of frailty in healthy elderly men. J Clin Endocrinol Metab
84:1334^1339
Vermeulen A 1991 Clinical review 24: androgens in the aging male. J Clin Endocrinol Metab
73:221^224
Vermeulen A, Verdonck L 1972 Some studies on the biological signi¢cance of free testosterone.
J Steroid Biochem 3:421^426
Welle S, Thornton C, Statt M, McHenry B 1996 Growth hormone increases muscle mass and
strength but does not rejuvenate myo¢brillar protein synthesis in healthy subjects over 60
years old. J Clin Endocrinol Metab 81:3239^3243
Wolf OT, Neumann O, Hellhammer DH et al 1997 E¡ects of a two-week physiological
dehydroepiandrosterone subsitiution on cognitive performance and well-being in healthy
elderly women and men. J Clin Endocrinol Metab 82:2363^2367
Wolf OT, Naumann E, Hellhammer DH, Kirschbaum C 1998 E¡ects of dehydro-
epiandrosterone replacement in elderly men on event-related potentials, memory, and well-
being. J Gerontol A Biol Sci Med Sci 53:M385^M390
Yarasheski KE, Zachwieja JJ, Campbell JA, Bier DM 1995 E¡ect of growth hormone and
resistance exercise on muscle growth and strength in older men. Am J Physiol 268:E268^E276
Yarasheski KE, Campbell JA, Kohrt WM 1997 E¡ect of resistance exercise and growth
hormone on bone density in older men. Clin Endocrinol 47:223^239
Zgliczynski S, Ossowski M, Slowinska-Srzednicka J et al 1996 E¡ect of testosterone
replacement therapy on lipids and lipoproteins in hypogonadal and elderly men.
Atherosclerosis 121:35^43

DISCUSSION
Veldhuis: When you use the term ‘unselected’, surely these are patients who are
self-selected; they agreed to come in. In this sense it is not fully random sampling.
One of the long-standing problems in this area is getting full population
HEALTHY AGEING IN MEN 17

representation, but your numbers are extraordinary and the relationships are
excellent.
This is sort of an elementary statistician’s nightmare, because one has multiple
possibilities for correlation. What I suspect you’ve probably done is to use some of
the modern principal-component or cluster-variable-type statistical approaches,
which lead to the question: can I select two or three variables which (in unique
combination) explain one of several critical outcomes? There are statisticians who
make their livelihood just out of this sort of principal component analysis. Can you
tell us anything about that?
van den Beld: In comparing all the di¡erent axes, we tried to see which components
are dependent on each other. It might not be just one hormone that is an indicator
for an overall good physical functional status. These relationships are all
independent of each other. The relationship of IGFBP2 with all the physical
characteristics was very strong. For example, testosterone remains signi¢cantly
related to muscle strength as well as bone mineral density after adjustment for
IGFBP2. With regard to just the physical characteristics, I think the physical
performance test that we measured by a method described by Guralnik et al
(1994) was a very good one, as was the muscle strength test. These are probably
the best indicators: if you measure them you know whether somebody is ageing
healthily or not.
Veldhuis: There are three technical terms that I would suggest you take to your
statistician. These are ‘cluster’ (not the cluster that we invented, this is a pre-
existing cluster), ‘principal component’, and ‘pathway’ or ‘critical component’.
There are only a few statisticians who work with this, but they’re looking at n-
dimensional correlations. These are mind boggling to even visualize. You are
asking yourself, ‘Is there anyway in this constellation of features to ¢nd little
nebulae that control the behaviour of the constellation?’ You presented
extraordinary data that a statistician of that ilk would be extremely enthusiastic
about.
Laron: What is the e¡ect of testosterone on visceral adipose tissue? It seems that
the increase of adipose tissue with advancing age is as visceral fat. Are there any data
on this? It is said that GH a¡ects mainly visceral fat. As there are data showing that
testosterone reduces adipose tissue, I wonder whether it has a di¡erential e¡ect on
general versus visceral adipose tissue.
van den Beld: I could not determine that in my study because we measured only
total fat mass and could not discriminate the visceral fat. I understand from the
literature that testosterone does indeed also have an e¡ect on the visceral fat.
Bj˛rntorp: Some years ago we carried out a 9 month study looking at testosterone
e¡ects in which we measured adipose tissue volumes by CT scan. We found a
signi¢cant decrease only in visceral fat, which seemed to be a speci¢c e¡ect on
visceral tissue. This might be because the density of androgen receptors is higher
18 DISCUSSION

in visceral fat than other adipose tissue. On this theme, I missed the Snyder report
that you referred to in your paper (Snyder et al 1999a,b), but there are problems
associated with performing studies like this. One of them is how to administer the
testosterone: if you inject it you get peaks and non-optimal concentration curves.
We tried various preparations in our studies and found that the transdermal
administration was quite good because this generates a reasonably stable level
within normal levels.
How were these men screened? When we screened our population we were
careful to screen potential subjects with prostate-speci¢c antigen (PSA) levels and
ultrasound of prostate size: if there was any suspicion of something abnormal then
we didn’t include them. We found, for example, two men with elevated PSA,
which would not have been good had they been included.
van den Beld: The subjects were strictly selected for this study.
Bj˛rntorp: As Professor Laron mentioned, there are similar studies of GH. I
think it would be extremely interesting to try a combination of testosterone and
GH treatment because there is evidence that these hormones amplify each other’s
e¡ect in the periphery. We are planning to do a study like this, but we can’t get the
funding.
Veldhuis: We’re very interested in the concept of low-dose GH treatment
bringing an older individual to a mid-adult level. Targeted IGF1 combined with
targeted testosterone would have low toxicity and may exert anabolic synergy. We
performed a one month study in 10 men, each of whom was studied at baseline and
then for one month on blinded transdermal 5 mg T-patch which brought the
testosterone to about 20 nM (600 ng/decilitre). IGF1 was stimulated through
6.25 mg/kg GH per day. This was only a one-month study, but the muscle IGF1
peptide gene expression was increased in nine out of 10 men. We picked such
biochemical endpoints, because it was a single-month study and we didn’t expect
functional, bone mineral or body fat changes. I think this needs to be explored
further.
Bj˛rntorp: To add more to this from our studies, we looked at the risk factor
pattern. Blood lipids went down (but the HDL cholesterol fraction did not
change as far as I remember), and blood pressure decreased.
van den Beld: In our study we also measured the intima/medial thickness of the
carotid artery. We also found a bene¢cial relationship between testosterone and the
intima/medial thickness: the higher the testosterone levels the lower the intima/
medial thickness.
Prior: The concern I have is that we are talking here about relationships at one
point in time. What I ¢nd much more physiologically important and valuable
are changes within people over time. I hope that you are carrying on these studies.
van den Beld: Last summer I performed a four year follow-up of all the elderly
men, but I’m still waiting for the hormone results.
HEALTHY AGEING IN MEN 19

Carroll: I have a couple of questions about your sample. First of all, were these
men medically screened at all?
van den Beld: Yes.
Carroll: They came from a working class suburb of Rotterdam. Did you screen
for alcoholism, for example? Or diabetes?
van den Beld: Yes.
Carroll: Were these all eliminated?
van den Beld: There were ¢ve people with non-insulin dependent diabetes, and
about 60 people taking medication for hypertension. We adjusted for medica-
tion.
Carroll: The correlations that you showed us are statistically signi¢cant by virtue
of the large sample size, but in truth they don’t account for more than about 10% of
the variance in any one case. A lot of other things are going on here.
van den Beld: Yes, that’s a problem.
Carroll: If you do a path analysis can you get to the directional links among these
variables because the correlations themselves are not striking.
van den Beld: Exactly. This is still quite a selected group: the age range is narrow
and there is an element of having to deal with survival of the ¢ttest as well. They
were all independently living men and they could all walk to the centre themselves.
Secondly, I think a lot of the associations are determined genetically, so we are
working in a restricted range.
Carroll: With regard to the QoL instrument, it was disappointing that this did
not to seem to relate to very much except IGFBP2. What is in that instrument? Is
there a speci¢c mood score?
van den Beld: I haven’t shown all the data. For example, questions about the
patients’ health, ¢nancial status and living conditions. The health part contains
questions about how well they can see, or whether they have problems with their
mobility or hearing. Within the ¢rst part they are asked how important they think
it is, and in the second part they are asked how satis¢ed they are about this same
question. Then the hormone-related questions are about sleep, anxiety, and this
sort of thing.
Veldhuis: Can you summarize the main correlates of quality of living? I thought
oestradiol was a strong endocrine correlate of quality of life. Was IGFBP3 a
positive or negative correlate?
van den Beld: Positive. IGFBP1 was negative, as was BP2, but I didn’t mention
them because when we adjusted for muscle strength and physical performance,
they were also strongly related to these QoL scores and there was no relationship.
The relationship between IGFBP2 and QoL was probably explained by the good
physical functional status of the subjects.
Veldhuis: So oestradiol and BP3 were both positive.
van den Beld: Yes.
20 DISCUSSION

Veldhuis: Here you have a possible linkage that is potentially very interesting.
You could ¢nd these clusters and then look for some theoretical explanation that
you could test further to try to explain successful living.
van den Beld: Testosterone was also positively related to quality of life, but this
relationship also disappeared after adjustment for muscle strength and physical
performance. Thus the e¡ect of testosterone on quality of life is probably
through the physical status of these men.
Carroll: Do you do any better if you look subscores on that instrument?
van den Beld: Yes, and then there are many more relationships. For example,
testosterone was strongly positive with the question about mobility. There was
also a question about sexual function and testosterone was positive with this.
Veldhuis: You might expect testosterone to have no real direct relationship with
QoL within the normal range.
Laron: You show that mobility is vitally important for the well-being of the
elderly males. When we take into consideration that increased exercise builds up
the muscles, increases GH secretion and reduces adipose tissue, how much can we
achieve by postponing the negative e¡ects of ageing by increasing exercise, thus
preventing the undesirable e¡ects of the changes in hormone secretion?
van den Beld: Quite a lot. If you can try to get them to exercise it is much better
than giving them hormone replacement therapy.
Laron: Have you tried to do this?
van den Beld: No.
Bj˛rntorp: I heard Steve Blair recently describing his phenomenal studies in
which they followed 10 000 people performing exercise tests for 10^20 years. He
found it was worse to be lean and untrained than fat and trained.
Morley: A meta analysis in the Journal of Gerontology clearly shows that exercise has
no outcome or long-term e¡ect on disability (Keysor & Jette 2001). It has
outcomes on strength, but when you look at all the studies that have been done
that look at function, beyond a couple of months you cannot see this. I was
surprised, but this is a very good meta analysis. People tend to look at the short-
term e¡ects and say they’re wonderful, but in studies in which exercise is
maintained long-term in controlled situations the results are disappointing.
Following up on Zvi’s point about mobility, there is absolutely no evidence that
mobility is important for function. All the function tests have been built around
mobility, but if you look at basic activities of daily living, while we are focused on
lower limb strength it is in fact upper limb strength and cognition that decide
whether you can function or not. We focus on legs, but legs get exercised all the
time, so hormones are much less likely to make a major di¡erence there than they
are in the upper limbs. If you look at the literature for some of the hormones,
upper limb strength tends to improve more than you will ¢nd with lower limb
strength.
HEALTHY AGEING IN MEN 21

Handelsman: We have recently completed two placebo-controlled trials using

alternative androgen delivery. There was no e¡ect at all on upper limb strength
over three months in either study.
Morley: I think we have overlooked a couple of important general points. Firstly,
there are a number of longitudinal studies out there. We have got to pay attention
to these. For instance, a study in Wisconsin showed short-term, in healthy elderly
people, DHEA tends to go up a little (E. Duthie, E. Burns, unpublished observa-
tions). Over a four year period in healthy older people the changes are di¡-
erent than if you take big groups of people where clearly it goes down. The
longitudinal studies in addition have to take into account seasonal e¡ects. It is
clear that most hormones have not only daily e¡ects but they also have circ-
annual e¡ects. Franz Halberg has recently pointed out that if you look at 10 year
e¡ects, you can plot myocardial infarction very nicely against sunspots which are
about a quasi 10^11 year period, so when we look at a single period of time we are
probably looking at nothing in an ageing process that takes place over 70 or 80
years (Halberg et al 2001). Until we start to pay a little more attention to this we
are going to get into trouble. At least in our cross-sectional study, the New Mexico
Process Study, we showed that not only testosterone but also IGF1 in males did
correlate with exercise, physical activity and food intake (Baumgartner et al 1999).
There are actually very few data on females, and these are mostly on oestrogen. In
fact, oestrogen has nothing to do with muscle strength. At the time of the meno-
pause, although there is a rapid loss of muscle strength, it is not related to oestrogen.
Handelsman: I wanted to pick up the issue of the nature of the sample. You
described them as unselected, but while they may not have been selected by you,
they were selected by themselves. What was the population from which you drew
the 400 and why did these 400 participate?
van den Beld: I did not refer to them as unselected. They are only unselected with
regard to hormone replacement studies. We sent invitations to all men aged above
73 in the town. 25% participated but of the 75% who did not participate some of
them already lived in nursing homes, so it’s di⁄cult to de¢ne the potential group
that could have participated.
Handelsman: Were they paid?
van den Beld: No.
Riggs: It seems to me that the term ‘healthy ageing’ is a bit of an oxymoron. In
ageing research, if you just study healthy people you are going to have a rather
severe bias. I also have an issue with epidemiological studies in general. I think
they’re very good for generating hypotheses, but at the end of the day you are
going to have to intervene: this is the only way you can be absolutely sure. The
pathway analysis that you showed is so complex that it is going to take quite a
while for all of us working together to be able to back this up with interventional
studies. I think this is the only way we will learn.
22 DISCUSSION

Veldhuis: How would you design an ideal intervention study? What points
would you tend to focus on given the analysis you have so far of these Rotterdam
data?
van den Beld: We have already performed an intervention study, with DHEA
replacement therapy. We took hundreds of really healthy men, not being treated
for hypertension, and we selected them on muscle strength. We tried to get subjects
with low muscle strength because four years ago we thought that this would make a
good endpoint to look at, since it was so highly correlated with all the other
physical characteristics and it is easy to measure. We also measured all the other
physical characteristics to see what would make the best endpoint. However we
could not ¢nd any e¡ect of DHEA replacement therapy for 9 months.
Wang: Was this with 50 mg DHEA?
van den Beld: Yes.
Morley: Baulieu et al (2000) showed the same thing in a huge study. It is very
di⁄cult to support the DHEA concept, at least at 50 mg. The studies that
showed any positive e¡ects were performed at 100 mg, except for mood. They
had a problem with their placebo which will have a¡ected mood.
Veldhuis: I have a comment on the converse strategy. Picking a group at high
risk for catabolism, such as the perioperative hip study that you did with GH which
appeared to show a subgroup response that was dramatic as an interventional
strategy, is interesting. However, one of the tough issues for all of us is which
group to pick for an intervention study. Does the choice risk a type II error, risk
confounds and risk a short lifetime on our part to ever publish the data?
van den Beld: We gave subjects GH therapy or placebo within 24 h of hip fracture,
continuing for 6 weeks. I’m not sure about the dose we gave but the subjects over
the age of 80 who received GH returned home signi¢cantly earlier compared to the
placebo-treated subjects.
Wang: Are these all men?
van den Beld: Men and women.
Wang: Bill Bremner’s group has data from Seattle, where they treated men after
knee surgery with high doses of testosterone (600 mg i.m. weekly for around 6
weeks) (abstract presented at the Endocrine Society Annual Meeting, June 2000,
Toronto, Canada). Their stay in hospital was remarkably decreased.
Morley: The best testosterone study so far published is the one by Bakhshi et al
(2000), where they looked at a rehabilitation group and clearly showed
improvement of the rate of rehabilitation with an adequate but not massive
replacement dose. To take frail people and give massive doses of anything is
going to get you into trouble if you continue for long enough, we shouldn’t be
doing it. One of the major problems with this ¢eld is that people have taken non-
hypogonadal, non-GH de¢cient people and given them huge pharmacological
doses. We should remember that the Snyder et al (1999a,b) study was actually in
HEALTHY AGEING IN MEN 23

non-hypogonadal men on the whole. I would never show the Snyder study as a
study for testosterone replacement because it didn’t look at real people in whom
you would be replacing testosterone.
van den Beld: Snyder also showed that there might be a threshold. Indeed, this is a
problem: if you have healthy men you don’t have the threshold.
Haus: We have carried out a study, which in some aspects was similar to Dr Van
den Beld’s, in cooperation with the C. I. Parhon Institute of Endocrinology of the
Romanian Academy of Medical Sciences in Bucharest, Romania. We studied the
circadian time organization in about 300 elderly subjects of both sexes, of ages
ranging from the 6th to the 10th decade (mean age 77 11 SEM). The subjects
were regarded as clinically healthy by their physicians and were not su¡ering any
acute, active and debilitating disease, although partially age-dependent conditions
like atherosclerosis, osteoarthritis and mild hypertension were present as to be
expected in any average population of this age. We studied some 15 hormonal
variables and about 30 other biochemical variables at six time points over one 24
h span, and in 32 subjects of this group four times over a 24 h span, each time
during a di¡erent season (Nicolau et al 1984, Haus et al 1988).
We found that circadian periodicity is well maintained in clinically healthy
elderly and old subjects (Haus et al 1988, 1989). However, there are some
characteristic di¡erences. If the temporal order of 39 endocrine and biochemical
variables is analysed by cluster (pattern) analysis, the timing of the rhythms varies
predominantly with gender, while the amplitude (expressed as percentage of the
circadian mean value) varies with age (Ticher et al 1994). It is of interest that in
women of di¡erent age groups, the epidemiologically determined risk to develop
breast cancer was characterized by clustering of the rhythm’s timing (the peak
times or acrophases) and not according to the amplitude, although age is
regarded as one of the prominent risk factors for the development of breast
cancer (Ticher et al 1994, 1996). The majority of endocrine variables showed a
decrease in circadian amplitude with ageing, which was pronounced in
melatonin, growth hormone, aldosterone, adrenal androgens, catecholamines,
TSH, serum iron, and in others in contrast to LH and insulin, which showed an
increase of level and amplitude with advancing age (Haus et al 1989, Haus &
Touitou 1997).
We explored the relation of rhythm parameters to the functional state of the
elderly subjects, as measured by the activities of daily living (ADL) index and a
mental status index developed by psychiatrists of the C. I. Parhon Institute. We
found a better functional state in subjects with higher (although not
hypertensive) blood pressure, higher (although not obese) body weight, lower
aldosterone (with higher salt excretion) and lower GH. Elderly men with higher
circadian mean testosterone concentrations did better than the ones with lower
testosterone, while the women with higher testosterone did worse.
24 DISCUSSION

Unfortunately, we did not have sex hormone binding protein determinations

available at that time.
In the elderly, we found absence of circannual rhythms or seasonal variations in
several variables as a group phenomenon, for example, in catecholamines and some
thyroid functions. If we assume that these seasonal variations are due to external
factors (e.g. cold), their absence may indicate a defect of adaptation in the elderly,
or if we assume that these variations are the expression of endogenous circannual
rhythms, their absence may indicate a desynchronization within the group, e.g.
with free running circannual rhythms in some of the subjects as we could observe
in the elderly, e.g. in blood pressure.
A word of warning has to be said about the comparison of populations of
di¡erent ethnic/geographic background. We had the opportunity to study under
comparable conditions groups of subjects in Japan, the USA and Romania. Using
cluster analysis, there was a similar acrophase clustering in the Americans and
Romanians, while the Japanese appeared to be di¡erent in their temporal
organization (Ticher et al 1996). Also the ageing patterns in the pulsatile
secretion of prolactin and of cortisol were di¡erent in American women as
compared with Japanese. Any generalizations concerning ageing changes in
ethnically/geographically di¡erent populations have to be treated with caution.
Laron: There may be a resistance to aldosterone/GH hormone developing with
age as there is with insulin.
Haus: Unfortunately we did not have IGF1 levels. This could tell us whether this
may be a resistance to GH function at the receptor level.
Morley: The Paris study (Maison et al 1998) showed that high GH levels were
associated with death in Parisian men. I think that’s the problem with high GH:
none of the studies, including the animal studies, have shown that it does anything
but kill you as you get older.
Veldhuis: It does raise the question of stress confounds, one of which is nutrition,
lowering IGF1 and stimulating some GH elevation. Do you have cortisol or
interleukin/cytokine data?
van den Beld: That’s something we are working on at the moment. We did
measure cortisol but the only correlation we found was that it was inversely
correlated with quality of life.

References
Bakhshi V, Elliott M, Gentili A, Godschalk M, Mulligan T 2000 Testosterone improves
rehabilitation outcomes in ill older men. J Am Geriatr Soc 48:550^553
Baulieu EE, Thomas G, Legrain S et al 2000 Dehydroepiandrosterone (DHEA), DHEA sulfate,
and aging: contribution of the DHEAge Study to a sociobiomedical issue. Proc Natl Acad Sci
USA 97:4279^4284
HEALTHY AGEING IN MEN 25

Baumgartner RN, Waters DL, Gallagher D, Morley JE, Garry PJ 1999 Predictors of skeletal
muscle mass in elderly men and women. Mech Ageing Dev 107:123^136
Guralnik JM, Simonsick EM, Ferrucci L et al 1994 A short physical performance battery
assessing lower extremity function: association with self-reported disability and prediction
of mortality and nursing home admission. J Gerontol 49:M85^M94
Halberg F, Cornelissen G, Watanabe Y et al 2001 Near 10-year and longer periods modulate
circadians: intersecting anti-aging and chronoastrobiological research. J Gerontol 56:M304^
M324
Haus E, Touitou Y 1997 Chronobiology of development and aging. In: Redfern PH, Lemmer B
(eds) Handbook of experimental pharmacology: physiology and pharmacology of biologic
rhythms. Springer-Verlag, Heidelberg, p 95^134
Haus E, Nicolau GY, Lakatua D, Sackett-Lundeen L 1988 Reference values for
chronopharmacology. Annu Rev Chronopharmacol 4:333^424
Haus E, Nicolau G, Lakatua DJ, Sackett-Lundeen L, Petrescu E 1989 Circadian rhythm
parameters of endocrine functions in elderly subjects during the seventh to the ninth decade
of life. Chronobiologia 16:331^352
Keysor JJ, Jette AM 2001 Have we oversold the bene¢t of late-life exercise? J Gerontol
56:M412^M423
Maison P, Balkau B, Simon D, Chanson P, Rosselin G, Eschwege E 1998 Growth hormone as a
risk for premature mortality in healthy subjects: data from the Paris prospective study. Br Med
J 316:1132^1133
Nicolau GY, Lakatua DJ, Sackett-Lundeen L, Haus E 1984 Circadian and circannual rhythms of
hormonal variables in elderly men and women. Chronobiol Int 1:301^319
Snyder PJ, Peachey H, Hannoush P et al 1999a E¡ect of testosterone treatment on body
composition and muscle strength in men over 65 years of age. J Clin Endocrinol Metab
84:2647^2653
Snyder PJ, Peachey H, Hannoush P et al 1999b E¡ect of testosterone treatment on bone mineral
density in men over 65 years of age. J Clin Endocrinol Metab 84:1966^1972
Ticher A, Sackett-Lundeen L, Ashkenazi IE, Haus E 1994 Human circadian time structure in
subjects of di¡erent gender and age. Chronobiol Int 11:349^355
Ticher A, Haus E, Ron IG, Sackett-Lundeen L, Ashkenazi IE 1996 The pattern of hormonal
circadian time structure (acrophase) as an assessor of breast-cancer risk. Int J Cancer 65:
591^593
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Ageing, stress and the brain

Bernard J. Carroll

Paci¢c Behavioral Research Foundation, 26386 Carmel Rancho Lane, Suite 202, PO Box
223040, Carmel, CA 93922-3040, USA

Abstract. Ageing of the brain is an important factor in overall ageing and mortality, and
new insights have clari¢ed the relationship between neuroregulation and ageing. First,
neuronal loss in normal ageing is now known to be a minor change. Loss of synapses
through dystrophic neuronal change is the hallmark of normal ageing. Second, similar
dystrophic changes occur in the brain with chronic stress. In both instances, forebrain
sites experience loss of synaptic input from brainstem regulatory nuclei. Third,
functional ageing is attributed in part to lifetime stress, under the concept of ‘allostatic
load’. Being inseparable from the functions of appraising and responding to stress, the
brain is an ultimate mediator of stress-related mortality, through hormonal changes that
lead to proximate pathologies like hypertension, glucose intolerance, cardiovascular
disease and immunological impairment. In chronic stress the brain shows clear allostatic
compensations that lead to pathology. Two subtle and chronic mechanisms that may
mediate brain pathology and accelerated ageing in chronic stress are proposed. These
are abnormal glucocorticoid receptor (GR) occupancy over the 24 h cycle, and elevated
body temperature. These factors lead to GR-mediated tissue changes and to acceleration
of general cellular ageing mechanisms. Human depression is discussed as an exemplary
demonstration of these principles.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 26^45

In 1988 Joseph Meites described a neuroregulatory theory of ageing, emphasizing

the integrative role of the nervous system for neuroendocrine axes and circadian
rhythms (Meites 1988). As the brain aged, Meites proposed, so would the
hypothalamus age, leading to menopause, andropause, somatopause and
dysregulated circadian rhythms. Meites tested pharmacological strategies to
augment hypothalamic neurotransmitter function, which he found could reverse
these ‘biomarkers of ageing’, a result that comports with the age-associated decline
of hypothalamic monoamine neurotransmitter systems (Rodr|¤ guez-Go¤mez et al
1995). Though it can be modi¢ed by recent insights, the general perspective of
Meites that we endocrinologists are also neuroscientists remains valid. Ageing of
the brain is an important factor in overall ageing and mortality.
Since that time, four new insights have changed our understanding of
neuroregulation and ageing. First, neuronal loss in normal ageing is minor; loss
26
AGEING, STRESS AND THE BRAIN 27

of synapses through dystrophic neuronal change is the hallmark of normal ageing.

Second, similar dystrophic changes occur in the brain with chronic stress. Third,
functional ageing is attributed in part to lifetime stress, under the concept of
‘allostatic load’. Being inseparable from the functions of appraising and
responding to stress, the brain is an ultimate mediator of stress-related mortality,
through hormonal changes that lead to proximate pathologies like hypertension,
glucose intolerance, cardiovascular disease and immunological impairment.
Fourth, neurogenesis does occur in the adult mammalian brain, which opens new
possibilities for treatments. The topic of neurogenesis is beyond the scope of this
chapter, though we may say with Pasko Rakic (1998) that, as a result of this
discovery, ‘the word impossible is not in the vocabulary of contemporary
neuroscience’.

Ageing of the normal brain

Until recently, ageing and senility were considered ¢rm partners. We now know,
however, that normal ageing of the brain is a process distinct from the degenerative
dementias. Early neuropathology studies led to the dogma that up to 40% of
neurons inexorably died over the lifespan. That belief was overturned by
improved methodology using unbiased stereological techniques (Morrison &
Hof 1997). In ageing primates and ageing rodents, rather than a loss of cortical
pyramidal cells there is decreased neuronal size and loss of synaptic arborization
(Masliah et al 1993, Peters et al 1994). Go¤mez-Isla et al (1996) found no age-
related loss of cells in the entorhinal cortex of the temporal lobe in normal
subjects aged from 60 to 90, in contrast to losses up to 65% in early Alzheimer’s
disease (AD).
A modest loss of brain volume occurs in normal ageing. In men, total brain
volume at 60 is 10% less than at 25 years (Murphy et al 1992). This change a¡ects
cortical and subcortical white matter rather than grey matter (Guttmann et al 1998,
Peters et al 1994) and is associated with myelin pathology in vertical ¢bres
traversing deeper layers of the cortex. This myelin pathology will slow nerve
conduction in association pathways, and may lead to functional de¢cits in
reaction time and working memory (Peters 1996, Peters et al 1994, 2000). In the
rhesus monkey, neurons of prefrontal cortex and substantia nigra show severe
dendritic pathology, with loss of organelles, vacuolation of cytoplasm,
membranous whorls and dense inclusion bodies (Siddiqi & Peters 1999, Peters et
al 1998). There is a reduction of 30^60% in the density of apical synapses on
pyramidal cells in prefrontal cortex. These changes correlate with age-related
cognitive impairment (Peters et al 1996, 1998).
It was also accepted wisdom that age-related neuronal loss occurs in subcortical
nuclei of the human brainstem, such as locus ceruleus (LC), which project
28 CARROLL

widely to the cerebral cortex and to subcortical limbic system sites, including the
hypothalamus (Chan-Palay & Asan 1989). Recent results with unbiased
stereological methods are con£icting, with predominantly negative studies (Ohm
et al 1997, Kubis et al 2000), though some continue to report substantial cell loss in
LC with age (Manaye et al 1995). Whether LC cells do or do not die with age, their
forebrain and hypothalamic projections certainly are dystrophic, with reduced
terminal arborization, loss of synaptic contacts and abnormal axonal branching
(Ishida et al 2000). A similar pattern of preserved neuronal number but aberrant
axonal formations with ageing occurs in the serotonin-containing rostral
projections of the dorsal raphe nucleus (Nishimura et al 1998, van Luijtelaar et al
1992).
A particular focus of recent work has been the trisynaptic perforant pathway
(PP) circuit that projects from layer II of the entorhinal cortex (EC) to granule
cells in the dentate gyrus of the hippocampus. These neurons in turn send mossy
¢bre projections to the CA1 and CA3 regions of Ammon’s horn, with further
relays to the hippocampal out£ow paths via the subiculum and the ¢mbria-fornix
(see Morrison & Hof 1997). This circuit is critical for associative memory, as the
EC receives highly processed information from heteromodal cortical association
areas, which it funnels into the hippocampus. Layer II of the EC is a¡ected by
neuro¢brillary tangles (NFTs) very early in AD, but it is minimally a¡ected in
normal ageing. Likewise, there is up to a 50% neuronal loss in the EC in early
AD but not in normal ageing.
Nevertheless, there is evidence of functional impairment of the perforant
pathway in normal ageing. Smith et al (2000) reported that spatial learning
de¢cits in aged rats correlated with reduced synaptophysin staining in the PP
entry zone, as well as in the CA3 region, which suggests loss of synaptic integrity
between EC and hippocampus via the PP. There is also a signi¢cant decrease of
N-methyl-D-aspartate (NMDA) receptor subunit 1 (NMDAR1) in the distal
dendrites of the dentate gyrus granule cells that receive the PP input (Gazzaley et
al 1996)
the inference is that distal dendritic pruning has occurred in the granule
cells. It is now clear that glucocorticoid excess and chronic stress produce similar
dendritic changes (McEwen 2000). The role of the PP and NMDA receptors in
long-term potentiation in the hippocampus is well known, and the changes just
described have been suggested as a basis of benign, age-related memory decline
(Morrison & Hof 1997).
In the hypothalamus itself, ageing and AD also are quite distinct with respect to
NFTs (Swaab et al 1992) and cell survival. Age-related changes of neuronal
number in human hypothalamic nuclei are quite variable (Swaab 1995, Hofman
1997). The sexually dimorphic interstitial nucleus of the anterior hypothalamus
displays a greater than 80% decrease of cell number in both sexes. In contrast, the
vasopressin- and oxytocin-producing cells of the supraoptic nucleus (SON) and
AGEING, STRESS AND THE BRAIN 29

the paraventricular nucleus (PVN) remain intact in old age. In the infundibular
nucleus there is increased cell number and activity, with hypertrophic neurokinin
B neurons in postmenopausal women.
Age-related neuronal dystrophic changes occur in the hypothalamus, similar to
those described in EC, hippocampus and cerebral cortex. For example, in the
arcuate nucleus the number of dendritic segments, total dendritic length,
branching and spine densities are reduced (Leal et al 1998). Likewise, in the SON
of the rat, marked dendritic regression is seen even though there is no neuronal
loss. The dendritic regression is thought to result from dea¡erentation due to the
preceding age-related loss of the noradrenergic input to the SON from the
brainstem (Flood & Coleman 1993). A related contribution is the decrease of a
key growth factor, brain-derived neurotrophic factor (BDNF) in ageing (Croll
et al 1998). These dystrophic changes in the ageing hypothalamus resemble those
seen in the hippocampus and the frontal cortex with stress (see below). The
morphologic similarities are accompanied by activation of hypothalamic nitric
oxide synthase (NOS) through NMDA receptor activation in both stress and
ageing (Kishimoto et al 1996, Vernet et al 1998). This process has been identi¢ed
as responsible for corticosterone-produced dystrophic dendritic changes in CA3
hippocampal pyramidal cells (Reagan et al 1999), and may be a general
mechanism by which stress and glucocorticoids cause neuronal dystrophy.
Likewise, both stress and ageing are associated with a decrease of BDNF in
hippocampus, hypothalamus and cortex. Furthermore, in ageing rats the
dynamic responses of BDNF and its receptor, TrkB, to stress are impaired
(Smith et al 1995, Smith & Cizza 1996), a dysregulation that appears to be
glucocorticoid-mediated (Cosi et al 1993).
In summary, age-associated changes in the brain are not like the pathology of
AD. They more closely resemble the changes caused by stress. There is no marked
loss of cortical neurons and only minimal appearance of NFTs in the PP^
hippocampal circuit or the hypothalamus. There is signi¢cant pathology of
myelin and of glial cells, which may slow nerve conduction velocities, reduce the
formation of synapses, and impair normal associative functioning. Most striking is
the dystrophy of forebrain projections from key brainstem nuclei. Axons display
reduced terminal arborization and abnormal branching, which is followed by
dendritic atrophy in the terminal ¢elds. This dea¡erentation and the loss of
synapses result in functional disconnection and consequent dysregulation of
hypothalamic functions, including the gonadal, growth hormone and
hypothalamus^pituitary^adrenal (HPA) axes. Candidate mechanisms for the
functional disconnection of the forebrain and hypothalamus in ageing are
NMDA receptor^NOS activation, compounded by impaired responsiveness of
mRNA for BDNF and TrkB, which would normally counter-regulate the
dendritic and axonal dystrophy in brainstem and forebrain sites. Both these
30 CARROLL

neurochemical changes are glucocorticoid-mediated and are seen in chronic

stress.

Allostasis, allostatic load and ageing

Another shift in the neuroregulatory perspective is the proposal that ageing is
related to lifetime stress, under the concepts of allostasis and allostatic load.
Sterling & Eyer (1988) introduced the term allostasis to explain morbidity and
mortality associated with chronic stress, especially social stress in human
populations. They de¢ned allostasis as ‘maintaining stability through change’.
The boundaries between allostasis and related constructs such as homeostasis,
adaptation and pathology are sometimes unclear. It is a heuristic construct that is
still being re¢ned (for example by Koob & Le Moal 2001). Speaking generally,
allostasis is distinct from homeostasis in maintaining a compensated equilibrium
rather than a physiological equilibrium: stability is maintained at a price. The
allostatic set point is abnormal relative to the homeostatic set point, the system is
inherently less stable, and it has a relatively narrow dynamic range. Finally, a
system in allostasis leads to pathology whereas a system in homeostasis does not.
The term allostatic load was introduced by McEwen & Stellar (1993) to denote
the cost of responding repeatedly or chronically to stress. McEwen (1998)
emphasized four types of responses that produce this ‘wear and tear’: (i) repeated
acute stress challenges; (ii) failure to adapt (i.e. to extinguish) the acute (or chronic)
stress responses normally with repeated (or chronic) exposure; (iii) excessively
prolonged acute stress responses; and (iv) inadequate acute stress responses
leading to elevated activity of other, normally counter-regulated allostatic
systems after stress (for which the Lewis rat is a proposed model). Both
developmental and genetic dimensions have been proposed as modi¢ers of
allostatic load that may determine vulnerability and risk of pathology (McEwen
2000).
When we view the brain’s response to stress, all the features of allostasis are
discernible (McEwen 2000). Chronic stress produces changes in brain function
and structure consistent with a compensated equilibrium and altered set points.
These include increased activity of the HPA axis; impaired fast and delayed
glucocorticoid feedbacks; and elevated circulating glucocorticoids. There is a
restricted dynamic range, with diminished circadian amplitude and elevated nadir
values. The resultant pathology includes decreased neurogenesis in the dentate
gyrus; loss of terminal arborization in the forebrain projections of the LC and the
raphe nuclei (Kitayama et al 1994, Duman et al 1997); dendritic regression in
hippocampus; and loss of synapses in hypothalamus and cortex. Antidepressant
treatments that induce mRNA for BDNF and TrkB can reverse these changes
(Kitayama et al 1997, Nibuya et al 1995). The functional result of these changes is
AGEING, STRESS AND THE BRAIN 31

a feed-forward cascade of disinhibited HPA activity, and other dysregulations such

as loss of oestrous cycling, impaired hedonic function (decreased spontaneous
motor activity, caloric intake and sweet food preference, and intracranial self-
stimulation), and decreased adaptive behaviours (Hatotani et al 1977, 1979, Katz
1982). These e¡ects of chronic stress resemble those seen in ageing. Consistent with
this similarity is the decreased longevity of inbred rat strains that are hyperreactive
to stress. Mean lifespan is 15 months in the spontaneously hypertensive rat (SHR),
and 21.5 months in the Wistar^Kyoto (WKY) rat, compared with 31 months for
the Brown Norway rat (Gilad & Gilad 1987, Brandle et al 1997).
In summary, the neuropathological and neuroregulatory changes of normal
ageing resemble those associated with chronic stress. These ¢ndings are
consistent with the proposal that longevity is a¡ected by allostatic load, and give
new meaning to Selye’s famous phrase, ‘the stress of life’. The glucocorticoid-
mediated allostatic e¡ect on lifespan must be mediated through relatively subtle
and chronic processes. Comparisons often invoked, for example by Sapolsky
(2000), between chronic stress and grossly pathological hypercortisolaemic states
like Cushing’s disease are not appropriate or informative. The exposure of tissues
to glucocorticoids in human and animal chronic stress paradigms is well below the
cushingoid range. Two more relevant candidate mediators of allostatic pathology
in chronic stress are proposed below.
First, the critical allostatic link between chronic stress, brain dystrophic change
and longevity is not likely to be the exposure of tissues to pathologically raised
concentrations of glucocorticoids. A chronobiological explanation is more likely,
namely, an altered HPA circadian rhythm that leads to the continuous occupancy
of glucocorticoid receptors (GRs) in target tissues. An important consideration
here is that normally GRs are essentially unoccupied during 25^30% of the day
(Dallman & Akana 1991). With the allostatic resetting of the circadian HPA
programme by stress, elevated nadir values of cortisol or corticosterone are
observed, which are su⁄cient to produce continuous occupancy of GRs over
24 h (e.g. Lo¤pez et al 1998). Continuous GR occupancy is abnormal, as evidenced
by tissue change such as thymic atrophy, even though the mean 24 h cortisol value
is held normal (Akana et al 1991). GR-containing target sites in the brain a¡ected
by this process will include the LC, raphe nuclei and hippocampus, where stress-
related dystrophic change occurs. Other GR-responsive tissues such as skin, bone
and liver will also be a¡ected, with consequent changes such as hair loss, skin
ulceration, osteoporosis, decreased muscle mass, increased susceptibility to
infections and glucose intolerance that are characteristic of chronically stressed
rats.
A second candidate allostatic mechanism leading to pathology is the stress-
associated phenomenon of elevated body temperature, which is most marked
during the quiet phase of the activity and HPA circadian rhythms (Meerlo et al
32 CARROLL

1997). The persistent, subtle hyperthermia results in a hypermetabolic state that

will accelerate general cellular ageing mechanisms. These include mitochondrial
failure and mitochondrial DNA damage, accumulation of oxygen free radicals
and protein conformational changes (Drachman 1997). The combination of
chronic hyperthermia with abnormal GR occupancy may be additive for
pathology. The signi¢cance of the allostatic load of nocturnal hyperthermia as a
link between stress and ageing is apparent from studies that ¢nd body
temperature di¡erences associated with di¡erential longevity (Hunter et al 1999).
The SHR rat strain provides an especially good example of stress-associated
hyperthermia, end organ pathology and shortened lifespan (Berkey et al 1990,
Morley et al 1990). Parenthetically, calorie restriction, which dramatically
increases longevity, leads to a persistent reduction of body temperature (Lane et
al 1996). Thus, these two factors, abnormal GR occupancy and hyperthermia,
which occur together during the quiet phase of the HPA circadian cycle, are
candidate mechanisms for the subtle and chronic e¡ects of stress on the brain and
on longevity. Both factors operate also in major depression, a clear instance of an
allostatic disorder.

Depression: an allostatic disorder with premature mortality

Human depression is a clear example of the allostatic link between chronic stress
and reduced longevity. The importance of psychosocial stress for provocation of
depressive episodes, for vulnerability to, and for risk of depression has been
reviewed by Checkley (1996). Depression is an established outcome of stress, and
the ongoing depressive episode itself constitutes a chronic stress. In large,
prospective studies of adverse medical outcomes in depression, the customary
psychiatric distinction between depressive symptoms and a depressive syndrome
appears not to be important. Stress is a risk factor for both major depression
(syndrome) and minor depression (symptoms). Minor depression is both a
prodrome of major depression and an outcome of major depression; and major
depression is a major outcome of minor depression.
The neuroendocrinology of human depression closely resembles that of chronic
stress models in the laboratory (Checkley 1996, Lo¤pez et al 1998). One sees
increased central HPA activity, reduced feedback, and adrenocortical
hypertrophy. Mean 24 h plasma cortisol concentrations are normal to slightly
elevated, as are urinary free cortisol excretion, plasma and cerebrospinal £uid free
cortisol concentrations, and salivary cortisol. Nocturnal plasma cortisol
concentrations are signi¢cantly raised, whereas daytime plasma cortisol
concentrations usually are not. Moreover, body temperatures, both nocturnal
and diurnal, are signi¢cantly elevated in depression by 0.4^0.6 8C (Szuba et al
1997, Rausch et al 2000). These are classical allostatic changes.
AGEING, STRESS AND THE BRAIN 33

Recent studies suggest a linear relationship between the number of depressive

symptoms and premature mortality. Koenig et al (1999) found in a nine-year
prospective study of 1001 male general medical inpatients that the adjusted
mortality hazard increased 10% for each one-point increase in baseline depression
score. In a prospective study of 7518 elderly community-resident women followed
for six years, Whooley & Browner (1998) found 7% mortality in those with no
depressive symptoms, 17% in those with three to ¢ve symptoms, and 24% in
those with six or more depressive symptoms. The excess mortality was associated
with both cardiovascular and non-cardiovascular deaths, but not with cancer.
Depression is associated with many cardiovascular risk factors. These include
hypertension, elevated plasma noradrenaline concentrations, decreased heart rate
variability (an index of cardiac vagal tone), increased platelet reactivity, increased
plasma ¢brinogen, myocardial ischaemia and myocardial infarction. Non-cardiac
morbidity associated with depression includes Type II diabetes mellitus,
osteoporosis, increased risk of falls (a correlate of decreased muscle mass),
increased intra-abdominal fat (a corticosteroid-related risk factor for cardio-
vascular disease), impaired wound healing and immune system suppression with
increased incidence of in£uenza in the elderly (see Whooley & Browner 1998).
Several of these depression-associated pathologies can be related to abnormal
GR occupancy (for example, osteoporosis, increased intra-abdominal fat,
immune system impairment and impaired glucose tolerance). Others may re£ect
accelerated ageing through the interaction of allostatic nocturnal hyperthermia
with general cellular degenerative mechanisms, as was discussed above in relation
to chronic stress. Thus, this human disorder well illustrates the connections among
ageing, stress and the brain. Human depression provides an excellent model for
re¢ning the concept of allostasis and for advancing our understanding of the
long-term e¡ects of stress on lifespan.

References
Akana SF, Bradbury MJ, Cascio S et al 1991 Low levels of exogenous corticosterone (B) impair
regulation of ACTH and B in intact rats. Endocrinology 128A:41
Berkey DL, Meeuwsen KW, Barney CC 1990 Measurements of core temperature in
spontaneously hypertensive rats by radiotelemetry. Am J Physiol 258:R753^R759
Brandle M, al Makdessi S, Weber RK, Dietz K, Jacob R 1997 Prolongation of life span in
hypertensive rats by dietary interventions. E¡ects of garlic and linseed oil. Basic Res Cardiol
92:223^232
Chan-Palay V, Asan E 1989 Quantitation of catecholamine neurons in the locus ceruleus in
human brains of normal young and older adults and in depression. J Comp Neurol
287:357^372
Checkley S 1996 The neuroendocrinology of depression and chronic stress. Br Med Bull 52:597^
617
34 CARROLL

Cosi C, Spoerri PE, Comelli MC, Guidolin D, Skaper SD 1993 Glucocorticoids depress
activity-dependent expression of BDNF mRNA in hippocampal neurons. Neuroreport
4:527^530
Croll SD, Ip NY, Lindsay RM, Wiegand SJ 1998 Expression of BDNF and trkB as a function of
age and cognitive performance. Brain Res 812:200^208
Dallman MF, Akana SF 1991 Negative feedback of glucocorticoids on the brain and pituitary:
role of feedback and its interactions with circadian and stress signals. In: Costa E, Paul SM
(eds) Neurosteroids and brain function. Thieme Verlag, Stuttgart (Fidia Research Found
Symp 8) p 25^29
Drachman D 1997 Aging and the brain: a new frontier. Annals of Neurology 42:819^828
Duman RS, Heninger GR, Nestler EJ 1997 A molecular and cellular theory of depression. Arch
Gen Psychiatry 54:597^606
Flood DG, Coleman PD 1993 Dendritic regression dissociated from neuronal death but
associated with partial dea¡erentation in aging rat supraoptic nucleus. Neurobiol Aging
14:575^587
Gazzaley AH, Siegel SJ, Kordower JH, Mufson EJ, Morrison JH 1996 Circuit-speci¢c
alterations of N-methyl-D-aspartate receptor subunit 1 in the dentate gyrus of aged
monkeys. Proc Natl Acad Sci USA 93:3121^3125
Gilad GM, Gilad VH 1987 Age-related reductions in brain cholinergic and dopaminergic indices
in two rat strains di¡ering in longevity. Brain Res 408:247^250
Go¤mez-Isla T, Price JL, McKeel DW Jr, Morris JC, Growdon JH, Hyman BT 1996 Profound
loss of layer II entorhinal cortex neurons occurs in very mild Alzheimer’s disease. J Neurosci
16:4491^4500
Guttmann CR, Jolesz FA, Kikinis R et al 1998 White matter changes with normal aging.
Neurology 50:972^978
Hatotani N, Nomura J, Yamaguchi T, Kitayama I 1977 Clinical and experimental studies on the
pathogenesis of depression. Psychoneuroendocrinology 2:115^130
Hatotani N, Nomura J, Inoue K, Kitayama I 1979 Psychoendocrine model of depression.
Psychoneuroendocrinology 4:155^172
Hofman MA 1997 Lifespan changes in the human hypothalamus. Exp Gerontol 32:559^575
Hunter WS, Croson WB, Bartke A, Gentry MV, Meliska CJ 1999 Low body temperature in
long-lived Ames dwarf mice at rest and during stress. Physiol Behav 67:433^437
Ishida Y, Shirokawa T, Miyaishi O, Komatsu Y, Isobe K 2000 Age-dependent changes in
projections from locus coeruleus to hippocampus dentate gyrus and frontal cortex. Eur J
Neurosci 12:1263^1270
Katz RJ 1982 Animal model of depression: pharmacological sensitivity of a hedonic defect.
Pharmacol Biochem Behav 16:965^968
Kitayama I, Nakamura S, Yaga T et al 1994 Degeneration of locus ceruleus neurons in stress-
induced depression model. Brain Res Bull 35:573^580
Kitayama I, Yaga T, Kayahara T et al 1997 Long-term stress degenerates, but imipramine
regenerates, noradrenergic axons in the rat cerebral cortex. Biol Psychiatry 42:687^696
Kishimoto J, Tsuchiya T, Emson PC, Nakayama Y 1996 Immobilization-induced stress
activates neuronal nitric oxide synthase (nNOS) mRNA and protein in hypothalamic-
pituitary-adrenal axis in rats. Brain Res 720:159^171
Koenig HG, George LK, Larson DB, McCullough ME, Branch PS, Kuchibhatla M 1999
Depressive symptoms and nine-year survival of 1,001 male veterans hospitalized with
medical illness. Am J Geriatr Psychiatry 7:124^131
Koob GF, Le Moal M 2001 Drug addiction, dysregulation of reward, and allostasis.
Neuropsychopharmacology 24:97^129
Kubis N, Faucheux BA, Ransmayr G et al 2000 Preservation of midbrain catecholaminergic
neurons in very old human subjects. Brain 123:366^373
AGEING, STRESS AND THE BRAIN 35

Lane MA, Baer DJ, Rumpler WV et al 1996 Calorie restriction lowers body temperature in
rhesus monkeys, consistent with a postulated anti-aging mechanism in rodents. Proc Natl
Acad Sci USA 93:4159^4164
Leal S, Andrade JP, Paula-Barbosa MM, Madeira MD 1998 Arcuate nucleus of the
hypothalamus: e¡ects of age and sex. J Comp Neurol 401:65^88
Lo¤pez JF, Chalmers DT, Little KY, Watson SJ 1998 Regulation of serotonin 1A, gluco-
corticoid, and mineralocorticoid receptors in rat and human hippocampus: implications for
the neurobiology of depression. Biol Psychiatry 43:547^573
Manaye KF, McIntire DD, Mann DM, German DC 1995 Locus ceruleus cell loss in the aging
human brain: a non-random process. J Comp Neurol 358:79^87
Masliah E, Mallory M, Hansen L, DeTeresa R, Terry RD 1993 Quantitative synaptic alterations
in the human neocortex during normal aging. Neurology 43:192^197
McEwen BS 1998 Protective and damaging e¡ects of stress mediators. N Eng J Med 338:
171^179
McEwen BS 2000 Allostasis and allostatic load: implications for neuropsychopharmacology.
Neuropsychopharmacology 22:108^124
McEwen BS, Stellar E 1993 Stress and the individual. Mechanisms leading to disease. Arch
Intern Med 153:2093^2101
Meerlo P, Overkamp GJ, Koolhaas JM 1997 Behavioural and physiological consequences of a
single social defeat in Roman high- and low-avoidance rats. Psychoneuroendocrinology
22:155^168 (erratum: 1997 Psychoneuroendocrinology 22:385)
Meites J 1988 Neuroendocrine biomarkers of aging in the rat. Exp Gerontol 23:349^
358
Morley RM, Conn CA, Kluger MJ, Vander AJ 1990 Temperature regulation in biotelemetered
spontaneously hypertensive rats. Am J Physiol 258:R1064^R1069
Morrison JH, Hof PR 1997 Life and death of neurons in the aging brain. Science 278:
412^419
Murphy DG, DeCarli C, Schapiro MB, Rapoport SI, Horwitz B 1992 Age-related di¡erences in
volumes of subcortical nuclei, brain matter, and cerebrospinal £uid in healthy men as
measured with magnetic resonance imaging. Arch Neurol 49:839^845 (erratum: 1994 Arch
Neurol 51:60)
Nibuya M, Morinobu S, Duman RS 1995 Regulation of BDNF and trkB mRNA in rat brain by
chronic electroconvulsive seizure and antidepressant drug treatments. J Neurosci 15:7539^
7547
Nishimura A, Ueda S, Takeuchi Y, Matsushita H, Sawada T, Kawata M 1998 Vulnerability to
aging in the rat serotonergic system. Acta Neuropathol 96:581^595
Ohm TG, Busch C, Bohl J 1997 Unbiased estimation of neuronal numbers in the human nucleus
coeruleus during aging. Neurobiol Aging 18:393^399
Peters A 1996 Age-related changes in oligodendrocytes in monkey cerebral cortex. J Comp
Neurol 371:153^163
Peters A, Leahu D, Moss MB, McNally KJ 1994 The e¡ects of aging on area 46 of the frontal
cortex of the rhesus monkey. Cereb Cortex 4:621^635
Peters A, Rosene DL, Moss MB et al 1996 Neurobiological bases of age-related cognitive decline
in the rhesus monkey. J Neuropathol Exp Neurol 55:861^874
Peters A, Sethares C, Moss MB 1998 The e¡ects of aging on layer 1 in area 46 of prefrontal cortex
in the rhesus monkey. Cereb Cortex 8:671^684
Peters A, Moss MB, Sethares C 2000 E¡ects of aging on myelinated nerve ¢bers in monkey
primary visual cortex. J Comp Neurol 419:364^376
Rakic P 1998 Young neurons for old brains? Nat Neurosci 1:645^647
Rausch JL, Johnson M, Fei Y et al 2000 Depressed outpatients have elevated temperatures. Biol
Psychiatry 47:93S
36 DISCUSSION

Reagan LP, McKittrick CR, McEwen BS 1999 Corticosterone and phenytoin reduce neuronal
nitric oxide synthase messenger RNA expression in rat hippocampus. Neuroscience 91:
211^219
Rodr|¤ guez-Go¤mez JA, de la Roza C, Machado A, Cano J 1995 The e¡ect of age on the
monoamines of the hypothalamus. Mech Ageing Dev 77:185^195
Sapolsky RM 2000 Glucocorticoids and hippocampal atrophy in neuropsychiatric disorders.
Arch Gen Psychiatry 57:925^935
Siddiqi ZA, Peters A 1999 The e¡ect of aging on pars compacta of the substantia nigra in rhesus
monkey. J Neuropathol Exp Neurol 58:903^920
Smith MA, Cizza G 1996 Stress-induced changes in brain-derived neurotrophic factor
expression are attenuated in aged Fischer 344/N rats. Neurobiology of Aging 17:
859^864
Smith MA, Makino S, Kim SY, Kvetnansky R 1995 Stress increases brain-derived neurotrophic
factor messenger ribonucleic acid in the hypothalamus and pituitary. Endocrinology
136:3743^3750
Smith TD, Adams MM, Gallagher M, Morrison JH, Rapp PR 2000 Circuit-speci¢c alterations in
hippocampal synaptophysin immunoreactivity predict spatial learning impairment in aged
rats. J Neurosci 20:6587^6593
Sterling P, Eyer J 1988 Allostasis: a new paradigm to explain arousal pathology. In: Fisher S,
Reason J (eds) Handbook of life stress, cognition and health. Wiley, Chichester, p 629^649
Swaab DF 1995 Ageing of the human hypothalamus. Horm Res 43:8^11
Swaab DF, Grundke-Iqbal I, Kremer HP, Ravid R, van de Nes JA 1992 Tau and ubiquitin in the
human hypothalamus in aging and Alzheimer’s disease. Brain Res 590:239^249
Szuba MP, Guze BH, Baxter LR Jr 1997 Electroconvulsive therapy increases circadian
amplitude and lowers core body temperature in depressed subjects. Biol Psychiatry 42:1130^
1137
van Luijtelaar MG, Tonnaer JA, Steinbusch HW 1992 Aging of the serotonergic system of the
rat forebrain: an immunocytochemical and neurochemical study. Neurobiol Aging 13:201^
215
Vernet D, Bonavera JJ, Swerdlo¡ RS, Gonzalez-Cadavid NF, Wang C 1998 Spontaneous
expression of inducible nitric oxide synthase in the hypothalamus and other brain regions of
aging rats. Endocrinology 139:3254^3261
Whooley MA, Browner WS 1998 Association between depressive symptoms and mortality in
older women. Arch Intern Med 158:2129^2135

DISCUSSION
Veldhuis: One of the things I am struggling with is this relationship between
stress and depression: which comes ¢rst? These both aggravate each other. Can
you separate this a little bit further for us? In a sense, just being alive and
producing cortisol must be a chronic stressor.
Carroll: That’s what Selye meant when he talked about the ‘stress of life’. As far
as the relationship with depression is concerned it doesn’t matter whether it’s the
chicken or the egg
there are probably several paths to it. One of these paths is
genetic. A small number of individuals have the genes for bipolar disorder or
familial unipolar depression. They often become depressed in the ¢rst instance
AGEING, STRESS AND THE BRAIN 37

through some kind of precipitating life stress, but after that the illness seems to run
a course of its own and the recurrences often come independently of stress. By the
time we get to study them they’ve been depressed for a month or two and, as I said,
that in itself constitutes a new chronic stress for the individual. The state of being
depressed is about the most painful existential condition that humans can endure.
Veldhuis: Are there data on glucocorticoid receptor-de¢cient mice that would
help in establishing this hypothesis? Right now the challenge I see, although it is
extremely attractive from an associative viewpoint, is causal connectivity. Maybe
you could demonstrate that a mouse with a constitutively activated glucocorticoid
receptor system, independently of peripheral steroid levels, tended to develop all
these diseases and die early and that the resistant receptor genotype had the
converse experience. Are people gathering such information?
Carroll: It’s a very good idea; I don’t know of any data on this.
Veldhuis: In humans we don’t think of the type I receptor as connected to the
metabolic features of cortisol activity except at the level of sodium retention and
potassium excretion. However, the ratio of these receptors changes in ageing.
What is your assessment of the role of the type I mineralocorticoid receptor (MR)
in the rodent during the ageing process? This is sort of a black box to me as an
endocrinologist.
Carroll: The best answer I can give to that is to say that MRs are concerned
with the amplitude of the rhythms, there is down regulation of MRs as I pointed
out in the stressed animal and in the depressed patient and that may well be the
mechanism by which the nadir cortisol values are elevated. It’s the MR that raises
the nadir of the values, but the elevated occupancy of glucocorticoid receptor
round the clock is what causes the tissue damage.
Burger: I wondered if you could say a little bit more about the direction of the
changes you’ve described in the male reproductive axis. You mentioned in the rat
the loss of cycling, which is probably also true of the human female. What about the
e¡ects of chronic stress and depression on the male reproductive axis and on
testosterone levels?
Carroll: Acute stress clearly leads to a decrease in testosterone production. In
chronic stress this may come back a little, but it still remains suppressed. We have
to be careful about which chronic stresses we are discussing, because if it’s a social
stress where a male rat is defeated by a stronger male, in that context the
testosterone stays very low. That’s the way the animal hierarchies of dominance
work.
Burger: Do they also have a shorter lifespan?
Carroll: Yes, if you put these animals in the visible burrow system where you
can observe their interactions, you can see that the ones that are pushed to the
periphery get less of everything: less food, less shelter and less access to the
opposite sex. They develop stress-related pathologies.
38 DISCUSSION

Burger: Has any attempt been made to actually replace or increase the
testosterone levels of animals in that situation?
Carroll: There have been a few experiments along these lines. The studies that
I am aware of look at male animals that have been through a social defeat.
Their testosterone is greatly reduced. If they are put back into a group of rats
they already knew, their stress parameters come back pretty rapidly, but if they
are kept in isolation then their stress parameters stay low. There are social
interactions with the rate of recovery from a severe stress.
Burger: Is that true for exogenous replacement as well?
Carroll: I am not aware that anyone has looked at exogenous replacement.
Veldhuis: That would be very interesting.
Handelsman: I’ve been quite struck by the number of depressed people on
selective serotonin reuptake inhibitors (SSRIs) who have relatively low
testosterone levels. Against that, in placebo-controlled studies there seems little
change in healthy young men in testosterone. The depression in testosterone
seems unrelated to whether they get better or not.
Laron: Do SSRIs reduce growth hormone (GH) as well?
Veldhuis: No, they stimulate GH slightly.
Handelsman: This may be a na|« ve question, but what is the e¡ect on life
expectancy of adrenalectomy, with or without maintenance of glucocorti-
coids?
Carroll: I don’t know.
Veldhuis: This could be interesting. Giving glucocorticoids back according
to the rhythmic hypothesis would be necessary, or giving a modern anti-
glucocorticoid that’s fairly selective to an intact animal could work as well.
Mˇller: There are data showing that adrenalectomy reduces hippocampal
damage (Stein & Sapolsky 1988). In referring to the senile brain, you
mentioned impairment in the monoaminergic systems (Gottfries 1992). What
about impairment of acetylcholine neurotransmission (Sherman & Friedman
1990, Pepeu et al 1993)? As far as depression is concerned we know that there
is a cholinergic theory of depression: typical antidepressant drugs are endowed
with anticholinergic activity (Baldessarini 1996).
Carroll: That idea has been around for a long time. In the cholinergic system
there are similar changes as in the monoaminergic system in normal ageing, but
these changes are very di¡erent from the loss of cholinergic cells in the nucleus
basalis that occurs in AD. Ageing is not like AD.
Bj˛rntorp: You are the pioneer in dexamethasone suppression testing. This is
useful for Cushing’s disease and depression. When we are looking at a functional
increase of cortisol in patients with chronic stress, which is the best way to assay the
feedback loop and the receptor loop? We have tried a lower dose than the
conventional one (0.5 mg rather than 1 mg) but I’m not sure that we really know
AGEING, STRESS AND THE BRAIN 39

what we are measuring. It might be useful to look at the escape of the inhibition;
you might see something there.
Carroll: Even though I did introduce the dexamethasone suppression test for
depression, I would never use dexamethasone again, because a major confound
in all of the dexamethasone work in psychiatry has been accelerated metabolism
of the steroid. I think we know why that happens now, because we give
dexamethasone at 11 p.m. and we sample blood the next day. With the
hyperthermia there is increased pharmacokinetic clearance, and we know that
these escapers very often have lower plasma dexamethasone concentrations when
the cortisol samples are drawn. I would use some other paradigm, such as a
hydrocortisone infusion in the human, and look for indices of
adrenocorticotropic hormone (ACTH) suppression.
Bj˛rntorp: You say that the metabolism of dexamethasone explains all this, but
are you sure that the circulating dexamethasone is mirroring what’s happening? Is
this a measurement of bound dexamethasone?
Carroll: It depends who you ask. If you ask Ron de Kloet he would say that with
dexamethasone the brain is like an adrenalectomized brain because the synthetic
dexamethasone doesn’t get to the brain, and that the site of action of
dexamethasone suppression of the HPA axis is the anterior pituitary. I don’t
think everyone agrees with that. If you put dexamethasone into the cerebral
ventricles it will certainly act on glucocorticoid receptors. Don’t forget about the
pituitary: it is sitting there, between the brain and the adrenal gland and makes
human experiments on direct central nervous system (CNS) feedback di⁄cult
because whatever you think you are doing to the brain, at the same time you are
also doing it to the anterior pituitary.
Prior: The experience I have as a clinician is that many women who present with
ovulatory disturbances of the menstrual cycle often give a story of childhood
sexual, psychological or physical abuse. It seems that this sets their hyper-
responsiveness to other life stresses for a long time. One of the common threads
besides disturbances of ovulation is sleep disturbance. Again, when we talk about
sleep we are talking about higher cortisol during the night when it should be at its
nadir.
Carroll: That is an emerging story; it certainly happens in a signi¢cant number of
depressed women. Be careful not to over-generalize from that sample because there
are many women with depression where this clearly does not apply. Then there are
all the depressed men and it doesn’t apply to them either.
Prior: To re£ect on the idea that those who have fewer children have less
allostatic load, I’ll agree with that in terms of the outcomes after pregnancy, but
also there is an ampli¢cation e¡ect of oestradiol on corticotropin-releasing
hormone (CRH). It is further ampli¢ed by social stress, as shown by
Kirschbaum’s group in young men (Kirschbaum et al 1996). The question is one
40 DISCUSSION

of homeostatic balance. Until you understand why women have fewer children I
wouldn’t necessarily jump to the conclusion it was related to long life and
homeostasis!
Carroll: Go back and read Westendorp & Kirkwood (1998) and I think you will
get the point. Look at those telling data from the English aristocracy. You mostly
won’t see it in modern times because very few families have 8^12 children, but they
used to.
Veldhuis: Perinatal imprinting is another subtlety that could confound
retrospective and even prospective studies. The Plotsky data published on
maternal^infant separation producing delayed adult di¡erences in stress
responses, tell me there are sparingly plastic changes in CNS feedback that occur
early in life. These are a bit frightening to us in clinical research, because we don’t
know what they all are. They are not always acknowledged even for the elements
we know, such as sexual abuse in childhood. They can confound and add
heterogeneity to these cross-sectional studies and perhaps add some strange
autocorrelation to the prospective studies. The oestrogen issue is complicated

I’d love to have a separate symposium on sex steroid interactions with the stress
axis.
Handelsman: I did some work with David Phillips looking at two separate birth
cohorts. In these, low birth weight (which is a well known predictor of
cardiovascular disease), was also a strong predictor of lifelong non-married state.
This suggests there are a lot of complexities that are not necessarily related to
traumatic experiences.
Morley: Clearly when we talk about cortisol we have to look at plasma clearance.
In ageing there is a decrease in plasma cortisol clearance time which makes
measurements of cortisol very di⁄cult to interpret. The Kirkwood data
(Westendorp & Kirkwood 1998) were strongly contested at a meeting in
Holland about 14 months ago by data that looked exactly the opposite. You can
look at the two sets of data and take whichever set you like: women who have lots
of kids learn to adapt and function well if they live in Holland, but if they live in
England they can’t adapt. The thing that really interested me is that you alluded to
dietary restriction as everybody does at an ageing meeting. The real problem with
dietary restriction is what you are really doing is taking animals who are living the
worlds most gluttonous, although reasonably stressful, life in unusually small
cages. We’ve recently completed a study looking at baboons in Ethiopia and
compared them to the average American baboon kept in relatively poor
conditions in a farm. Using leptin as a surrogate for fat, it turned out that leptin
is almost unmeasurable in the average baboon living in a normal environment.
With regard to dietary restriction, nobody will ever get fat levels down to that
level: it would be considered too cruel and unacceptable by any animal review
board so we have to recognize that dietary restriction isn’t going to be valid for
AGEING, STRESS AND THE BRAIN 41

humans. Also, in mice and rats restriction really produces premature ageing if you
want to look at the hormones. Almost all the hormone changes in rodents look just
like the hormone changes you see in old animals and the possible interpretation of
this, which makes a lot of sense, is that if you slow down life so that almost nothing
happens, you can live a long time. These data also hint that we shouldn’t replace
hormones because replacing hormones may hurry life up: we may have two great
years, but on the other hand we might die 10 years earlier.
Carroll: It depends what you call ‘ageing’. This takes us back to Meites, who
talked about menopause and somatopause as biomarkers of ageing. It all depends
on the biologic context in which it occurs. Your point about food restriction
illustrates the same thing: animals in the wild who are existing under high
foraging demand for food are very stressed and calorie restricted at the same
time. Animals in the lab have it easy, when we cut back their calories by 30%
they’ve still got much more than the high-foraging-demand animals in the wild,
so they are not stressed even though they have equivalent calorie intake.
Veldhuis: Are you saying that over-feeding is stressful and conducive to shorter
life?
Handelsman: Remember to keep in perspective that people have never lived
longer or been better fed than now.
Morley: If you look at the epidemiological studies that show that low
cholesterol is good for the heart, the only way they could ever do those is to
take out the amount of food eaten because the studies clearly showed that the
more food you ate the less likely you were to have a heart attack. In fact the
conundrum is easily explainable, because the very heavy eaters are the people
who exercise a lot, so the only way they could get rid of the high food intake
was to take out the exercise. These things are never easy and the pitfalls we all
have to deal with demonstrate that we shouldn’t accept that just because we
overeat that it is bad for us. Gross overeating is clearly bad, somewhere in
between is most probably OK.
Bj˛rntorp: Having gone through some of the literature on cortisol in ageing,
I found a study looking at cortisol turnover with ageing. If you are fat then
the adipose tissue transforms cortisol to cortisone by 11b-hydroxysteroid
dehydrogenase.
Veldhuis: I agreed with that review. Our deconvolution methods depend upon
an assumption of a stable distribution. I haven’t found consistent data reporting
altered distributions in ageing.
Riggs: I was particularly interested in your comments that normal ageing is
associated with a disproportionate decrease in synaptic connections, with a
smaller decrease in the number of neurons. I was reminded of the data in rodents
showing that synaptic connections can be related to the degree of stimulus the
rodents are exposed to. Is there any evidence in humans that those elderly people
42 DISCUSSION

that stay busy, that read, and remain active and stimulated have less a¡ected
synaptic connections?
Carroll: The only data I know that touch tangentially on this are those on AD.
The risk of AD is clearly related in Snowdon et al’s (1996) study of nuns to the
complexity of their mental operations at age 18, when they wrote essays about
why they wanted to enter the convent. They were later followed through to
death and autopsy, and those who developed AD late in life had less complex
mental operations in their teenage years. You can interpret that several ways, you
can say AD begins at birth, or you can say that the ones who were better educated
and stimulated intellectually somehow managed to mitigate their independent risk
of AD.
Morley: The problem with this is that the more active and higher your education
level, the less likely you are to develop AD or dementia quickly because the
diagnosis takes time. There was a recent paper that actually looked at plaques in
people who did and didn’t have AD and there was no di¡erence. You can come to
the conclusion that there might be very little di¡erence between normal ageing and
AD. Our animal model in the SAMP8 mice clearly suggest that it is just an
exaggeration: if everybody and every animal overproduces b-amyloid it’s how
much you over produce and for how long
if you excessively overproduce you
will get AD (Morley et al 2000, Kumar et al 2000). But in bright people the
diagnosis will not be made as early. All the educational epidemiological studies
show the higher your education the less likely you are to get AD. I don’t think
this is because your synapses are necessarily better, its just that its harder to
diagnose someone who is very bright.
Carroll: I think you have overstated the case here; the data from autopsy
con¢rm the in vivo clinical diagnoses with about 80% agreement
Mˇller: Although there are data showing that education level is inversely related
to the development of AD (Geerlings et al 1999), this may be due to the improved
lifestyle of better educated people. Several studies have shown that oestrogen
therapy can improve cognitive function or prevent AD in elderly women
(Paganini-Hill & Henderson 1994, Jacobs et al 1998, Tang et al 1996), whereas
other studies have not found an association (Brenner et al 1994, Shaywitz &
Shaywitz 2000). Anyway, women with high serum concentrations of non-protein
bound and bioavailable oestradiol are less likely to develop cognitive impairment
than women with low concentrations (Ya¡e et al 2000). Reportedly, oestrogens
induce synapsis formation in the hippocampal pyramidal neurons (McEwen &
Alves 1999). I think that we have to consider all these aspects together.
Robertson: In human epidemiology, I wonder whether there are data from agents
that a¡ect body temperature, for example non-steroidal anti-in£ammatory agents
and barbiturates. Is there any evidence that these might be having a favourable
impact on ageing?
AGEING, STRESS AND THE BRAIN 43

Carroll: There are data showing that NSAIDs reduce the risk of AD. I do not
know of any similar data for ageing in general.
Robertson: You gave the impression that you wished the glucocorticoid receptor
in human subjects was a little less active. Do you think a slightly less e¡ective
receptor or a slightly less e¡ective adrenal cortex might be bene¢cial for ageing in
human beings?
Carroll: I would rather say let’s eliminate that nocturnal secretion of cortisol.
There are ways this can be done. For example, one of the current experimental
approaches of treating depression is to give metyrapone, or ketoconazole, but
this is mostly done with daytime administration. I would prefer to go with a
midnight or 10 p.m. administration; there you might have a chance of success
because this would truly uncouple the GR from circulating steroids for a certain
length of time at the right period of the circadian cycle.
Bj˛rntorp: I heard that there is now a CRH inhibitor available for treating
depression.
Carroll: Every pharmaceutical corporation that I know is trying to develop
CRH antagonists, and to get them into clinical trials as rapidly as possible. The
potential indications include depression, but also anxiety states, for example. Any
data currently available are very preliminary and so we just have to wait.
Handelsman: Can you comment on the point that Dr Mˇller made about
oestradiol and/or other oestrogens preventing AD? Are there are any studies of
testosterone in men having similar e¡ects?
Carroll: The data show that oestrogen is worthless as a treatment of already
diagnosed AD patients, but the epidemiologic studies suggest that women who
are maintained on oestrogen after menopause have a reduced incidence of AD.
Prior: However, it is wise to remember that these studies are highly biased by the
di¡erences between oestrogen-taking women and non-oestrogen taking women
(Barrett-Connor 1991).
Handelsman: This may be like the cardiovascular story all over again.
Prior: Are the temperature di¡erences that you are talking about measurable in a
practical way? Can you detect the di¡erence between depressed and non-depressed
patients?
Carroll: Yes. There have even been studies measuring daytime temperatures in
depressed patients when they come in for an outpatient visit. There are clear 0.3^
0.4 8C di¡erences between depressed and control subjects even in the daytime.
Prior: What would be the best time to measure it?
Carroll: I would go for the night time, when the di¡erences will be greatest.
Prior: Could it be measured at the patient’s usual bedtime? Or do you have to
wake them in the middle of the night?
Carroll: We could give them a capsule to swallow and get telemetric data
from it.
44 DISCUSSION

Prior: We’re not talking epidemiology here! I guess we could get the equipment
for small studies.
Veldhuis: Those capsules are about US$80 a piece and the monitoring equipment
is a couple of hundred dollars.
Carroll: But the data they produce are wonderful.
Prior: They have been used in the past to show that core temperature rises prior
to vasomotor episodes (Freedman & Woodward 1996).

References
Baldessarini RJ 1996 Drugs and the treatment of psychiatric disorders: depression and mania. In:
Hardman G, Linmbird LE, Molino¡ PB, Ruddon RY, Goodman Gilman A (eds) Goodman
& Gilman’s the pharmacological basis of therapeutics, 9th edition, MacGraw-Hill, New
York, p 431^459
Barrett-Connor E 1991 Postmenopausal estrogen and prevention bias. Ann Intern Med
115:455^456
Brenner DE, Kukull WA, Stergachis A et al 1994 Postmenopausal estrogen replacement therapy
and the risk of Alzheimer’s disease: a population-based case-control study. Am J Epidemiol
140:262^267
Freedman RR, Woodward S 1996 Core body temperature during menopausal hot £ushes. Fertil
Steril 65:1141^1144
Geerlings MI, Schmand B, Jonker C, Lindeboom J, Bouter LM 1999 Education and incident
Alzheimer’s disease: a biased association due to selective attrition induce of a two-step
diagnostic procedure? Int J Epidemiol 28:492^497
Gottfries CG 1992 Clinical and Neurochemical aspects of diseases with cognitive impairment.
Rev Neurosci 3:191^205
Jacobs DM, Tang MX, Stern Y et al 1998 Cognitive function in nondemented older women who
took estrogen after menopause. Neurology 50:368^373
Kirschbaum C, Schommer N, Federenko I et al 1996 Short-term estradiol treatment enhances
pituitary^adrenal axis and sympathetic responses to psychosocial stress in healthy young men.
J Clin Endocrinol Metab 81:3639^3643
Kumar VB, Farr SA, Flood JF et al 2000 Site-directed antisense oligonucleotide decreases the
expression of amyloid precursor protein and reverses de¢cits in learning and memory in aged
SAMP8 mice. Peptides 21:1769^1775
McEwen BS, Alves SE 1999 Estrogen actions in the central nervous system. Endocr Rev
20:279^307
Morley JE, Kumar VB, Bernardo AE et al 2000 b-amyloid precursor polypeptide in SAMP8
mice a¡ects learning and memory. Peptides 21:1761^1767
Paganini-Hill A, Henderson VW 1994 Estrogen de¢ciency and risk of Alzheimer’s disease in
women. Am J Epidemiol 140:256^261
Pepeu G, Casamenti F, Pepeu IM, Scali C 1993 The brain cholinergic system in ageing mammals.
J Reprod Fertil Suppl 46:155^162
Shaywitz BA, Shaywitz SE 2000 Estrogen and Alzheimer’s disease: plausible theory, negative
clinical trial. JAMA 283:1055^1056
Sherman KA, Friedman E 1990 Pre-and post-synaptic cholinergic dysfunction in aged rodent
brain regions: new ¢ndings and an interpretative review. Int J Dev Neurosci 8:689^708
Snowdon DA, Kemper SJ, Mortimer JA, Greiner LH, Wekstein DR, Markesbery WR 1996
Linguistic ability in early life and cognitive function and Alzheimer’s disease in late life.
Findings from the Nun Study. JAMA 275:528^532
AGEING, STRESS AND THE BRAIN 45

Stein BA, Sapolsky RM 1988 Chemical adrenalectomy reduce hippocampal damage induced by
kainic acid. Brain Res 473:175^180
Tang MX, Jacobs D, Stern Y et al 1996 E¡ect of oestrogen during menopause on risk and age at
onset of Alzheimer’s disease. Lancet 348:429^432
Westendorp RG, Kirkwood TB 1998 Human longevity at the cost of reproductive success.
Nature 396:743^746
Ya¡e K, Lui L-Y, Grady D, Cauley J, Kramer J, Cummings SR 2000 Cognitive decline in
women in relation to non-protein-bound oestradiol concentrations. Lancet 356:708^712
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Alterations in the ageing corticotropic

stress-response axis
Per Bj˛rntorp

Department of Heart and Lung Diseases, Sahlgren’s Hospital, University of G˛teborg, SE

413 45, Sweden

Abstract: The problem with determining whether ageing is followed by perturbations of

the regulation of the hypothalamic^pituitary^adrenal (HPA) axis is that ageing per se is
very di⁄cult to separate from the e¡ects of environmental insults. Data in ageing rodents
indicate that with age the winding-down of cortisol after challenges is prolonged. This is
probably due to an insu⁄cient feedback regulation by glucocorticoid receptors in
speci¢c ¢elds in the hippocampus. The functional and topographic characteristics of
these changes are identical to those seen after prolonged stress, suggesting that such
factors might be of signi¢cance. Data in humans also suggest that with age basal
cortisol secretion, diurnal rhythm, and the early response to challenges are not
a¡ected but similar to animal data the return to baseline values after stimulation
seems to be delayed, probably due to a diminished feedback control. Several studies
suggest that common diseases of age, for example cardiovascular disease and type 2
diabetes mellitus, are associated with similar HPA axis perturbations as those seen in
old subjects. Recent population studies indicate that adrenal hyperactivity, associated
with a stressful environment, is generating risk factors for these diseases. This is likely
to be dependent on genetic susceptibility, and associated polymorphisms have been
found in several candidate genes of importance for neuroendocrine and autonomic
regulations.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 46^65

Stress and the function of the hypothalamic^pituitary^adrenal axis

The hypothalamic^pituitary^adrenal (HPA) axis regulates the secretion of cortisol.
The net result of the output of cortisol from the adrenals is dependent on central
stimulatory events and feedback inhibitory mechanisms. The central regulation
takes the shape of a diurnal rhythm, which in humans consists of high activity in
the early morning hours and low activity in the afternoon/evening. This basic
pattern is changed by the experience of a variety of factors which are often
grouped together under the common description of ‘stress’.
46
CORTICOTROPIC STRESS-RESPONSE AXIS 47

‘Stress’ may be de¢ned as factors which disturb the homeostasis of several

neuroendocrine and autonomic systems. A more recent de¢nition is the
disturbance of allostasis. An example of allostasis is the maintenance of a normal
adaptation of blood pressure to posture, or insulin secretion after a meal (McEwen
1998). Stress reactions have been de¢ned by Henry & Stephens (1977) and may be
divided into two distinct classes. One is the ¢ght^£ight reaction where a challenge
is met by a typical neuroendocrine^autonomic reaction, primarily involving the
sympathetic nervous system with elevation of heart rate and blood pressure.
When the threat is dealt with successfully additional hormonal changes are
occurring, which in males include an increase in testosterone production.
Another type of reaction is a depressive, defeat reaction, also described as a
situation of helplessness. This reaction is seen when the threat is perceived as
overwhelming and cannot be coped with. Mental depression also follows this
reaction as well as typical neuroendocrine^endocrine reactions, including
activation of the HPA axis with elevated cortisol secretion. When this is
frequently repeated or chronic, other consequences follow. A primary reaction
seems to be a delayed ‘down-winding’ of the response, with cortisol elevation
remaining for longer than normal after a challenge. A second step is diminished
responsiveness of the HPA axis. An early consequence of this seems to be
decreased morning cortisol levels in humans. Finally a low, poorly responsive
HPA axis results. Such neuroendocrine ‘burn-out’ has been described in war
veterans and holocaust victims (McEwen 1998).
Along this pathway other neuroendocrine and autonomic adaptations are
occurring. The growth hormone and hypothalamo^pituitary^gonadal axes are
inhibited by the elevated HPA axis activity at several levels (Chrousos & Gold
1992). This results in decreased secretion and low concentrations of sex steroid
and growth hormones. In addition, when the activity of the HPA axis is
su⁄ciently down-regulated there seems to be a parallel activation of the
sympathetic nervous system, which might be seen as a compensatory adjustment
in an attempt to maintain homeostatic conditions (Plotsky et al 1989).
The activity of the HPA axis is controlled by a feedback inhibition mechanism
where central glucocorticoid receptors are involved. When cortisol is bound to
these receptors the activity of the HPA axis is diminished. When challenged with
repeated activation of the HPA axis this control mechanism becomes less e⁄cient,
¢rst apparently by a decrease of glucocorticoid receptor (GR) density and later by
structural changes, including atrophy of the dendrites coupled to the inhibitory
mechanism of the GR. Eventually brain substance loss occurs, which is visible as
atrophy and the formation of lacunae. This is seen in Cushing’s syndrome (with
severe, constant overproduction of cortisol), in melancholic depression (another
condition with chronically elevated cortisol), and in war veterans (presumably as a
result of long-term, severe stress reactions in the battle¢eld). At which stage such
48 BJO«RNTORP

FIG. 1. The associations between stress reactions and pathways to disease generation.
Activation of the HPA axis is followed by cortisol secretion, particularly when insu⁄ciently
controlled by the feedback loop involving central glucocorticod receptors (GRs). Androgens,
probably partly of adrenal origin, are important in women. When prolonged, this activation
causes insulin resistance and visceral accumulation of body fat, both cornerstones of the
metabolic syndrome. A parallel activation of the central sympathetic nervous system is
followed by hypertension and increased mobilisation of free fatty acids (FFA), which amplify
insulin resistance. For details, see Bj˛rntorp (1996, 2002). CRH, corticotropin-releasing
hormone; ACTH, adrenocorticotropic hormone; n.s., nervous system.

changes are reversible is not certain, but it seems unlikely that severe topographic
losses can be normalized (McEwen 1998).
As described above, the reaction to stress involves a large number of reactions
where the central sympathetic nervous system and the HPA axis are primary
players. The inhibitions of other central neuroendocrine axes are probably
secondary to the activation of the HPA axis but form an integrated part of the
stress reaction participating in the damaging somatic consequences.
Both of these reaction types are often mixed, particularly in humans, and it may
be di⁄cult to separate one from the other. The peripheral consequences of stress
reactions are therefore often a mixture of neuroendocrine/endocrine and
autonomic reactions. As will be described later, such prolonged reactions are
damaging to bodily systems and may end up in serious diseases (see schematic
summary in Fig. 1).

Stress activators
What are the inducers of stress reactions? They consist of a large number of
complex, variable external and internal events. External, psychological stressors
may include socioeconomic handicaps such as demanding work environments or
poor economic conditions. Psychosocial handicaps include divorce, living alone
and bereavement. Other external stressors are a poor physical working situation
CORTICOTROPIC STRESS-RESPONSE AXIS 49

with noise, polluted air, toxicological hazards, etc. Internal stressors might be
infections or other diseases, trauma and sleep deprivation. In addition to the
large number of stress factors, di¡erent personalities perceive pressure di¡erently:
what one person counts as stressful does not necessarily apply to another person.
The complexities of this ¢eld have led researchers to de¢ne ‘stress’ as any factor
that activates the stress axes: the HPA axis and/or the central sympathetic nervous
system.

Ageing and the HPA axis

The background provided above is necessary for understanding the potential
in£uence of ageing on the HPA axis and its associated tight network of other
neuroendocrine reactions. One may ask whether this system is ageing by what
might be called a ‘normal’ age-related process or if the wear and tear of today’s
complex, hectic living conditions are involved. As will be seen, it is di⁄cult to
separate these two factors.
In order to separate out the e¡ects of a ‘normal’ ageing process from the e¡ects of
external factors, one would have to examine the HPA axis and its associated cascade
of reactions in other neuroendocrine and autonomic systems in older subjects who
have not been exposed to the wear and tear of life. This is of course not possible
even if one were to ¢nd a population totally separated from modern, urbanized life,
because there are always surrounding factors which activate our stress systems.

Animal studies
An approach to this problem might be to study the regulation of the HPA axis in
the ageing laboratory rat. Such animals are probably less exposed to varying
external stressors than humans are. Furthermore, the primitive mechanisms
involved are probably common to most mammals because they are essential for
survival, and the results from rats should therefore be reasonably applicable to
humans. The results of such studies indicate that the ageing rat has an increased
corticosterone secretion after a stress challenge (Sapolsky et al 1983). The down-
winding of the corticosterone response seems to be delayed after a stress response.
This varies among di¡erent rat strains (De Kloet 1992) and between genders (Brett
et al 1986). The reason for the slow return to basal values seems to be a down-
regulation of central GR density, particularly in certain ¢elds of the hippocampus
region (Sapolsky et al 1984a). This is apparently an e¡ect of extended exposure to
glucocorticoids (Sapolsky et al 1984b) and is ampli¢ed by toxins (Sapolsky 1985),
including alcohol (Walker et al 1980). The mechanisms involved have been
suggested to be e¡ects of glucocorticoids on glucose uptake and utilization in the
cells involved, in analogy with the well known sensitivity of brain cells to such
50 BJO«RNTORP

insults. This is seen after prolonged administration of glucocorticoids at

concentrations corresponding to those after stress, resulting in about 50% lower
GR density of hippocampal receptors (Sapolsky et al 1985a). Interestingly, the
neonatal rat, which has a low number of GRs in critical areas of the brain that
develop slowly to adult density, is sensitive to impacts during this developmental
period. Interference with this development results in persistent hypersecretion of
cortisol (Sapolsky et al 1985b).
The end result of such down-regulation of GRs will then be an ine⁄cient
feedback control of corticosterone production which will amplify further
damage to the regulatory system. When exposure to noxious agents occurs for a
shorter time, the changes are reversible, while longer periods of exposure are
followed by slower or no restitution (Sapolsky 1985). This corresponds to a loss
of glucocorticoid binding cells in the critical areas of the hippocampus which are
replaced by glial cells. When the insults are su⁄ciently large and extended, loss of
brain substance occurs.
Interestingly, these are the same changes that are seen in the ageing rat, and are
the basis for the ‘glucocorticoid cascade hypothesis’ (Sapolsky et al 1986). This
hypothesis suggests that repeated stress insults, particularly when accompanied
with exposure to toxins, down-regulate GR density in speci¢c areas of the
hippocampus region in the brain. This is reversible up to a certain point, beyond
which permanent, irreversible loss of hippocampal neurons occurs with persistent
oversecretion of glucocorticoids as a result.
This hypothesis suggests that with ageing and the experience of repeated stress
periods during life, such changes may occur. The best evidence for this possibility
seems to be the identical functional and anatomical changes seen with ageing with
less or no such exposure.

Studies in humans
The brain functions involved are those responsible for primitive survival
reactions. Studies in non-human primates have shown essentially the same
reaction pattern as in rodents (Uno et al 1989). Therefore a similar chain
of reaction would also be expected in humans. The situation is, however, more
di⁄cult to evaluate in humans, who typically experience a wide variety of
external challenges during a long life, with presumably varying vulnerability.
Several di⁄culties are apparent with the available human studies. These include
variables such as the selection of study groups, whether the study group is healthy
or not, sample size, the extent of the age di¡erences in the groups, and perhaps the
most important problem, the techniques utilized for assaying HPA axis function.
This latter problem is particularly important in human studies because of the
exquisite sensitivity of the reactions to be followed, as will be illustrated.
CORTICOTROPIC STRESS-RESPONSE AXIS 51

Furthermore, fully controlled experiments are di⁄cult to perform in this area of

research. These problems make many studies of limited value, which has been
taken into account in the following evaluation.
Although the results are not totally consistent, they indicate that the basal, non-
stimulated HPA axis and circadian rhythm are not disturbed in old subjects
(Seeman & Robbins 1994). Stimulation by adrenocorticotropic hormone
(ACTH) is apparently also followed by a normal initial response of cortisol while
there might be a delayed return to normal values (Blichert-Toft 1975, West et al
1961). In general, studies with stimulation by corticotrophin-releasing hormone
are di⁄cult to evaluate, largely because of small sample sizes. However, one such
study also suggests that the return to baseline cortisol values after such stimulations
is prolonged in older individuals (Pavlov et al 1986).
Tests of the function of the feedback mechanism in humans usually involve the
administration of dexamethasone and followed by the inhibition of cortisol
secretion. These results are of particular importance in view of the evidence for a
de¢cient feedback regulation in old animals. With the conventional dose of
dexamethasone (1 mg) results have been largely normal, when calculated as
inhibition below a certain given ‘normal’ level (Green & Friedman 1968), while
old subjects seem to have higher mean absolute values after suppression
(Tourigny-Rivard et al 1981). When inhibition with lower doses (0.5 mg) is
tested to improve sensitivity, there is better evidence for a blunted inhibition in
older subjects (Oxenkrug et al 1983, Branconnier et al 1984).
Most of the studies referred to have only measured the concentrations of circu-
lating hormones, which are a result of secretion and clearance. One study indicates
that there is no di¡erence in clearance of cortisol (Barton et al 1993). The cortisol
levels after various challenges are most likely a result of secretion without much
in£uence of removal rates, although the latter cannot be stated with certainty.
Several studies have been performed in subjects with various diseases that might
be suspected to be consequences of HPA axis activation over prolonged periods.
Conditions with increased cortisol secretion due to tumours (such as Cushing’s
syndrome) or with elevated exposure to glucocorticoids due to therapeutic inter-
ventions are followed by insulin resistance, abdominal obesity, hypertension,
dyslipidaemia, osteoporosis, cognitive impairments and immune de¢ciencies
(Seeman & Robbins 1994). These are consequences of elevated glucocorticoid
exposure where the mechanisms are essentially known.
Conditions with similar phylogenetic expression to these overexposures to
glucocorticoids are hypertension, cardiovascular disease and type 2 diabetes
mellitus. These conditions are suggested to have impairments of HPA axis
regulation (Seeman & Robbins 1994). This might be a consequence of the
diseased state, but it is possible that these diseases are in fact at least partly
consequences of primary dysregulation of the HPA axis.
52 BJO«RNTORP

In an attempt to summarize the human data, there is no conclusive evidence that

ageing per se is associated with a faulty regulation of the HPA axis, although this
must be considered impossible to study in a meaningful way because it requires all
external in£uences to be absent. There is, however, suggestive evidence that in the
ageing person there is often a diminished rate of return of cortisol secretion to basal
values after challenge, which may be due to an ine⁄cient feedback regulation. This
would then agree with the better-controlled animal data on the input of ageing. It is
impossible, however, to judge whether this is a consequence of the ageing process
or due to repeated exposure to stressful events during a long life. Normal ageing is
unavoidably associated with exposure to the wear and tear of daily life. There are
clearly individual di¡erences in the sensitivity and perception of such challenges,
and these di¡erences most likely have a genetic basis. Increased frequency of
periods of stress during a lifetime would be expected to increase the vulnerability
to diseases via e¡ects on metabolic and other processes, as will be discussed in a
following section.
Another approach to this problem is the following. If it is supposed that the
HPA axis and its associated cascade of events are a¡ected by ageing, then it is
interesting to know what characterizes subjects who have been able to maintain a
normal HPA axis function into an advanced age. The function of the HPA axis has
been followed longitudinally for a long period in a group of healthy subjects.
Those without signs of deterioration of this function were characterized by good
memory and selective attention (Lupien et al 1994) as well as preserved
hippocampal structure (Lupien et al 1998). Whether these di¡erences are
genetically determined and localized to the HPA axis regulation itself, or to a
resistance to environmental ‘stressors’, or both, is not known.

The impact of the HPA axis on human health

Since it is almost impossible to separate the e¡ects of ageing and cumulative stress
on the regulation of the HPA axis, it might be of interest to examine how the HPA
axis activity is associated with diseases that are characteristic of ageing. This
assumes that the insults on the HPA axis of wear and tear are an unavoidable
consequence of ageing. Longevity runs in families and ‘successful’ ageing is
likely to have genetic components. Perturbations of the regulation of the HPA
axis and associated neuroendocrine reactions and disease risk factors are linked to
molecular genetic characteristics. Although this molecular genetic information is
so far preliminary and super¢cial, it might serve to generate ideas of new
approaches to the problem of neurendocrine^endocrine and autonomic ageing.
This information has largely been obtained from two population studies, one in
men and one in women. Since the primary aim of these studies was not to study the
e¡ects of ageing, they included subjects born in the same year, the men born in 1944
CORTICOTROPIC STRESS-RESPONSE AXIS 53

and the women in 1956. This approach was chosen to avoid the confounding
in£uence of age. Nevertheless, the information obtained might be useful for the
interpretation of the in£uence of daily life stressors on neuroendocrine health, or
abnormalities and their associations to disease or disease precursors, characteristic
of older subjects.
In these studies it was thought to be important to examine the status of the HPA
axis during ordinary daily life. Measuring the easily disturbed HPA axis during
abnormal, potentially stressful conditions, such as in a laboratory or hospital,
does not provide the information needed in an attempt to examine the impact of
daily life on, for example, cortisol secretion. If, in addition, venipuncture is
involved, then the risk for disruption of a steady state increases due to the
exquisite sensitivity of the systems being examined. For these reasons saliva
cortisol was measured. The advantages of this procedure are that saliva can easily
be collected under any conditions (except sleep) and saliva cortisol contains the free
active fraction of circulating cortisol. Saliva was collected during several pre-
determined time points including morning, afternoon and evening, and perceived
stress was registered by the proband. The reaction to a standardized lunch was
determined. A dexamethasone suppression test was performed, also under ‘free-
living’ conditions, with a lower (0.5 mg) than conventional (1.0 mg) dose in order
to improve sensitivity. In this way, an impression of the basal and stimulated (food
intake and reported perceived stress) activity of the HPA axis was obtained as well
as the function of the feedback inhibitory loop. These measurements were then set
in relation to a number of psychological, socioeconomic and lifestyle factors,
other hormones, as well as conventional metabolic and haemodynamic risk
factors for disease (Rosmond et al 1998, Baghaei et al 2001).
The sensitivity of the HPA axis as mirrored by saliva cortisol measurements is
clearly demonstrated by another recent study where saliva cortisol was measured at
random times during a working day, with perceived stress and mood being
reported before each sampling. The results showed that the measurements were
in£uenced not only by the stress perceived immediately before sampling, but also
by the memory of previously experienced stress, anticipated stress and mood.
Cortisol levels varied 20-fold over the day and were proportional to the reported
perceived psychological impacts (Smyth et al 1998). This study demonstrates the
necessity of measuring the activity of the HPA axis under ordinary, daily
conditions, which are those presumably a¡ecting disease-generating mechanisms
over a prolonged period of time.

Men
The results in the men and women showed interesting di¡erences. In the men it was
possible to single out groups of HPA axis reactions. A normal group was
54 BJO«RNTORP

characterized by having high morning cortisol, rapidly decreasing before lunch, a

response to lunch and then low cortisols in the afternoon/evening. This group
consisted of about two-thirds of these essentially randomly selected men. About
25% of the men reported stress during the cortisol sampling and these men had
lower morning values which did not wind down before lunch, and an
accentuated lunch response. Afternoon and evening values did not di¡er from
the normal group. This resulted in an elevation of total secreted cortisol over the
day. These di¡erences from the norm are characteristic of an HPA axis reaction
which has been subjected to prolonged stress (Dallman et al 1992) and therefore
suggest that these men had been repeatedly stressed by various factors before the
day of measurements. The feedback inhibition also showed signs of a less than
optimal function, presumably another sign of previous stress experiences.
There were statistical associations to socioeconomic and psychosocial handicaps
as well as depressive and anxiety symptoms in these men, suggesting background
factors to a perturbed function, or a hypersensitivity of the HPA axis.
In addition to these neuroendocrine perturbations, there were associations with
conventional risk factors for disease, including abdominal accumulation of body
fat, insulin resistance, dyslipidaemia and hypertension, a cluster of factors now
called the ‘metabolic syndrome’. It is predicted that these men have an increased
risk of developing type 2 diabetes mellitus, cardiovascular disease and/or stroke.
The risk factor pattern can largely be explained by the elevated cortisol secretion,
because cortisol directs an increased fraction of body fat to central depots through
known mechanisms, where an increased density of GRs in visceral depots is a
crucial characteristic (Bj˛rntorp 1996). This is seen dramatically in Cushing’s
syndrome. Furthermore, cortisol is well known to induce insulin resistance, and
the associated dyslipidaemia may originate from the elevated insulin levels (Reaven
1988). The increased blood pressure is likely to be caused by a parallel activation of
the sympathetic nervous system through mechanisms mentioned in the
introduction. Taken together it is thus possible to visualize a chain of events
starting out from environmental factors which activate the HPA axis, resulting
in peripheral risk factors for diseases which increase dramatically in prevalence
with age.
Among the men there were also a small group (less than 10%) where the HPA
axis showed a low, constant, rigid activity, such as has been described for a
neuroendocrine ‘burn-out’ (McEwen 1998). There were also signs of a poor
suppression by dexamethasone. These men had low values of testosterone and
growth hormone as well as robust associations to risk factors including elevated
blood pressure. It is unlikely that elevated cortisol is responsible for these
associations. Instead other hormonal and autonomic factors, taking part in
the cascade of events following regulatory perturbations of the HPA axis, might
be involved. We do not know the background to these changes, although there are
CORTICOTROPIC STRESS-RESPONSE AXIS 55

also associations to the presumed stress-generating factors mentioned above.

Other factors, yet unknown, are probably also involved.
This description of detailed phylogenetic characteristics made it possible to
approach genetic background factors by the candidate gene approach. Since the
results pointed at a central origin of the ¢ndings, we focused on genes that
control central neuroendocrine and autonomic regulations. This might be of
interest from the aspect of ageing, because molecular genetic mechanisms might
increase the susceptibility to develop various age-associated diseases.
A ¢rst target was the GR gene, because if the central GR is malfunctioning, the
sensitivity of the HPA axis to various challenges would be increased due to a
de¢cient feedback control, which would explain many of the phenomena
described. The ¢ndings, in brief, have so far been the following. With restriction
fragment length polymorphism technique it is possible to demonstrate
polymorphisms that are strongly associated with basal and stimulated cortisol
secretion as well as several of the crucial risk factors associated with a sensitive
HPA axis. Furthermore, polymorphisms which are both activating or protecting
from elevated activity of the sympathetic nervous system have been revealed (for
review and detailed references, see Bj˛rntorp 2002). It would be interesting to
examine whether these polymorphisms are associated with enhanced longevity. If
they are of importance for the creation of disease risk factors, it seems likely that
they would be associated with a shorter life due to development of morbidity and
mortality.

Women
The study of women shows interesting di¡erences in comparisons with the study in
men. It turns out that androgens are the major associates to disease risk factors in
women, showing stronger associations than various cortisol measurements
(Baghaei et al 2001). This is also combined with low oestrogens. Such elevated
androgen levels may have an adrenal origin because they are statistically
connected with other hormones of adrenal origin. Furthermore, in women the
endocrine perturbations also correlate with essentially the same psychosocial and
socioeconomic handicaps as in men, including tendencies to anxiety and
depression, expected to activate the HPA axis. It is likely that stressed men
also secrete elevated amounts of adrenal androgens, but this is of no
signi¢cance in the presence of a large input of gonadal testosterone. It seems
likely from the results now available that certain common factors in men and
women activate the HPA axis, with cortisol as the main trigger for adverse
peripheral e¡ects in men, while adrenal androgens are of more importance in
women. These androgens in women may also have ovarian origin, but this has
not yet been examined.
56 BJO«RNTORP

Androgens have several adverse e¡ects on females. Muscular insulin resistance is

created via known mechanisms and distribution of androgens to physically normal
women is followed by accumulation of fat in the abdomen, both of which are
corner stones of the metabolic syndrome (For review and references, see
Bj˛rntorp 1996).
In women the candidate gene approach has followed di¡erent lines to the
molecular genetic studies in men, guided by the results of the phylogenetic
expression of symptoms. Special emphasis has been placed on the status of micro-
satellites in genes involved in steroid hormone regulations, because this has
recently been found to be of particular interest in problems concerning steroid
hormone metabolism and e¡ects (Comings 1998). Studies have focused on the
pathway of androgen metabolism and signalling to cells. Aromatase is the
enzyme converting androgens to oestrogens, and might therefore be considered
to protect women from a too pronounced an exposure to androgens. Elevated
androgens in combination with low oestrogens, found in the women who were
a¥icted by disease risk factors in the population, may well be an indirect sign of
less than optimal aromatase activity. The aromatase gene has a microsatellite in the
¢fth intron, which when short, previously has been found to be associated with
increased risk of developing breast cancer and osteoporosis in women (Comings
1998). With hyperandrogenicity this microsatellite is also short (Baghaei et al
2001). This polymorphism seems particularly interesting because when women
are examined who have signs of perfect somatic and psychological health, includ-
ing low androgens and high oestrogens, their microsatellite in the aromatase gene
is longer than average. In fact, both groups of women, one with a high
androgen:oestrogen ratio and a number of disease risk factors as a consequence,
and the other with general health and a low androgen:oestrogen ratio, have
not only opposite phylogenetic characteristics, they form totally separate
subgroups in their length of the microsatellite in the aromatase gene (L. Baghaei,
R. Rosmond, L. Westberg, M. Hellstrund, M. Landen, E. Eriksson, G. Holm &
P. Bj˛rntorp, unpublished results). This is such a striking, non-overlapping
¢nding that it should motivate more detailed studies of the mechanisms involved.
The androgen receptor transfers the androgen signals to cells. Its gene has a
microsatellite in the ¢rst exon, coding for the signalling part of the androgen
receptor. Here a short repeat has also been found to be associated with serious
disease both in men and women. Such a short stretch is associated with strong
signals of androgens (Comings 1998). With hyperandrogenicity this
microsatellite is also abnormally short, suggesting that the receptor transfers
ampli¢ed signals to the cell, and would therefore worsen the e¡ects of elevated
androgens (Westberg et al 2001).
The stress reactions along the HPA axis and the central sympathetic nervous
system are thus likely to be followed by predictors and subsequent precipitation
CORTICOTROPIC STRESS-RESPONSE AXIS 57

of diseases such as cardiovascular disease and type 2 diabetes mellitus. The

presumed pathways for such a development are summarized schematically in Fig.
1. The neuroendocrine^endocrine pro¢les of such perturbations seem to be
di¡erent in men and women. These perturbations have similarities with the
characteristics of ageing. Detailed references and further discussion are found in
reviews dealing with this problem (Bj˛rntorp 1996, 2002).

References
Baghaei F, Rosmond R, Westberg L et al 2001 Microsatellite stretches in the CYP450 aromatase
gene are associated with androgens and risk factors in women, submitted
Barton RN, Horan MA, Weijers WM, Sakkee AN, Roberts NS, van Bezooijen CFA 1993
Cortisol production rate and the urinary excretion of 17-hydroxycorticosteroids, free
cortisol and 6b-hydroxycortisol in healthy elderly men and women. J Gerontol 48:M213^
M218 (erratum: 1994 J Gerontol 49:B35)
Bj˛rntorp P 1996 The regulation of adipose tissue distribution in humans. Int J Obes Relat
Metab Disord 20:291^302
Bj˛rntorp P 2002 Hypothalamic origin of prevalent human disease. In: Pfa¡ D, Arnold A, Etgen
A, Fahrbach S, Rubin R (eds) Hormones, brain and behavior. Academic Press, San Diego, in
press
Blichert-Toft M 1975 Secretion of corticotrophin and somatotrophin by the senescent
adenohypophysis in man. Acta Endocrinol Suppl 78:15^154
Branconnier RJ, Oxenkrug GF, McIntyre I, Pomara N, Harto NE, Gershon S 1984 Prediction
of serum cortisol response to dexamethasone in normal volunteers: a multivariate approach.
Psychopharmacology 84:274^275
Brett LP, Levine R, Levine S 1986 Bidirectional responsiveness of the pituitary-adrenal system in
old and young male and female rats. Neurobiol Aging 7:153^159
Chrousos GP, Gold PW 1992 The concept of stress and stress system disorders. JAMA
267:1244^1252
Comings DE 1998 Polygenic inheritance and micro/minisatellites. Mol Psychiatry 3:21^31
Dallman MF, Akana SF, Scribner KA et al 1992 Stress, feedback and facilitation in the
hypothalamo-pituitary-adrenal axis. J Neuroendocr 4:516^526
De Kloet ER 1992 Corticosteroids, stress, and aging. Ann NY Acad Sci 663:337^371
Green MF, Friedman M 1968 Hypothalamic-pituitary-adrenal function in the elderly. Gerontol
Clin 10:334^339
Henry JP, Stephens PM 1977 Stress, health and the psychosocial environment: a sociobiological
approach to medicine. Springer-Verlag, Berlin
Lupien S, Lecours AR, Lussier I, Schwartz G, Nair NPV, Meaney MJ 1994 Basal cortisol levels
and cognitive de¢cits in human aging. J Neusrosci 14:2893^2903
Lupien SJ, DeLeon M, DeSanti S et al 1998 Cortisol levels during human aging predict
hippocampal atrophy and memory de¢cits. Nat Neurosci 1:69^73 (erratum: 1998 Nat
Neurosci 1:329)
McEwen BS 1998 Protective and damaging e¡ects of stress mediators. New Engl J Med
338:171^179
Oxenkrug GF, Pomara N, McIntyre IM, Branconnier RJ, Stanley M, Gershon S 1983 Aging and
cortisol resistance to suppression by dexamethasone: a positive correlation. Psychiatry Res
10:125^130
58 DISCUSSION

Pavlov EP, Harman SM, Chrousos GP, Loriaux DL, Blackman MR 1986 responses of plasma
adrenocorticotropin, cortisol, and dehydroepiandrosterone to ovine corticotropin-releasing
hormone in healthy aging men. J Clin Endocrinol Metab 62:767^772
Plotsky PM, Cunningham ET Jr, Widmaier EP 1989 Catecholaminergic modulation of
cortocotropin-releasing factor and adrenocorticotropin secretion. Endocr Rev 10:437^458
Reaven GH 1988 Role of insulin resistance in human disease. Banting lecture 1988. Diabetes
37:1595^1607
Rosmond R, Dallman MF, Bj˛rntorp P 1998 Stress-related cortisol secretion in men:
relationships with abdominal obesity and endocrine, metabolic and hemodynamic
abnormalities. J Clin Endocrinol Metab 83:1853^1859
Sapolsky RM 1985 A mechanism for glucocorticoid toxicity in the hippocampus: increased
vulnerability to metabolic insults. J Neurosci 5:1228^1232
Sapolsky RM, Krey LC, McEwen BS 1983 The adrenocortical stress-response in the aged male
rat: impairment of recovery from stress. Exp Geront 18:55^62
Sapolsky RM, Krey LC, McEwen BS 1984a Glucocorticoid-sensitive hippocampal neurons are
involved in terminating the adrenocortical stress response. Proc Natl Acad Sci USA 81:6174^
6180
Sapolsky RM, Krey LC, McEwen BS 1984b Stress down-regulates corticosterone receptors in a
site-speci¢c manner in the brain. Endocrinology 114:287^292
Sapolsky RM, Krey LC, McEwen BS 1985a Prolonged glucocorticoid exposure reduces
hippocampal neuron number: implications for aging. J Neurosci 5:1222^1227
Sapolsky RM, Meaney MJ, McEwen BS 1985b The development of the glucocorticoid receptor
system in the rat limbic rain III. Negative feedback regulation. Brain Res 350:169^173
Sapolsky RM, Krey LC, McEwen BS 1986 The neuroendocrinology of stress and aging: the
glucocorticoid cascade hypothesis. Endocr Rev 7:284^301
Seeman TE, Robbins RJ 1994 Aging and the hypothalamic-pituitary-adrenal response to
challenge in humans. Endocr Rev 15:233^260
Smyth J, Ockenfels MC, Porter L, Kirschbaum C, Hellhammer DH, Stone AA 1998 Stressors
and mood measured on a momentary basis are associated with salivary cortisol secretion.
Psychoneuroendocrinology 23:353^370
Tourigny-Rivard MF, Raskind M, Rivard D 1981 The dexamethasone suppression test in an
elderly population. Biol Psychiatry 16:1177^1184
Uno H, Tarara R, Else JG, Suleman MA, Sapolsky RM 1989 Hippocampal damage associated
with prolonged and fatal stress in primates. J Neurosci 9:1705^1711
Walker DW, Barnes DE, Zornetzer SF, Hunter BE, Kubanis P 1980 Neuronal loss in
hippocampus induced by prolonged ethanol consumption in rats. Science 209:711^721
West CD, Brown H, Simmins EL et al 1961 Adrenocortical function and cortisol metabolism in
old age. J Clin Endocrinol Metab 21:1197^1207
Westberg L, Baghaei F, Rosmond R et al 2001 Polymorphisms in the androgen receptor gene
and the estrogen receptor b gene are associated with androgen levels in women. J Clin
Endocrinol Metab 86:2562^2568

DISCUSSION
Veldhuis: This area presents a lot of challenges to me. One of the di⁄culties I’m
having as an endocrinologist with the testosterone theory, is that I don’t see this
syndrome X (metabolic syndrome) appearing across puberty in boys. At this stage
their testosterone goes up 10^30-fold, but they get just mild insulin resistance. This
makes me wonder whether insulin resistance is a marker for some other activity in
CORTICOTROPIC STRESS-RESPONSE AXIS 59

the system. Polycystic ovary syndrome (PCOS) patients (bearing in mind that this
is a diverse category) appear to have primary insulin resistance. If PCOS patients
who are not particularly obese but who are hyperandrogenaemic are given drugs or
other interventions to lower their insulin levels, this also lowers androgen levels. If
androgens were the only thing one measured and one knew nothing about insulin
action through its tyrosine kinase pathways, one would say that it is the androgen
that is the prime cause here.
Bj˛rntorp: I didn’t have time to show that in the stressed men. Particularly in
the small subset of ‘burned out’ men I referred to, testosterone levels are
lowered. The hyperandrogenicity problem is probably a separate female issue.
There is a sort of chicken and egg discussion about which comes ¢rst, the changes
in the androgens or the changes in the oestrogens. But when we give androgens in
moderate amounts to female rats, they become highly insulin resistant. When
investigators have given androgens to women
there is one US study (Lovejoy
et al 1996) and one Dutch study on transsexuals (Elbers et al 1997)
they also
become insulin resistant. I believe that when you give testosterone, these things
happen.
Veldhuis: There are a few exceptions. One data set from John Nestler showed
that, when he reduced insulin with diazoxide, androgen levels also fell (Nestler et al
1989, 1990). Most experiments show the converse, but it is hard to give long-term
androgens. When we monitor androgens, we may be looking at a marker of
another underlying event that is driving syndrome X. We certainly see syndrome
X all the time without any hint of hyperandrogenism. Could you clarify your
thinking on the relationship between this group of women who are stressed and
hyperandrogenaemic and the PCOS patients? Is PCOS simply an extended
subgroup of this type of woman, or is PCOS a separate disease? It is such a tough
area.
Bj˛rntorp: Yes. PCOS is a mess. First of all, I think we have to separate out PCOS
and polycystic ovaries. There are reports now of polycystic ovaries in up to 20% of
women. We have found around 25% of women to be hyperandrogenic. We are
going to look at their ovaries. On the other hand, I have recently seen hints in
the literature that stress might be involved in polycystic ovaries.
Veldhuis: There are all kinds of mild hints of HPA axis dysregulation in PCOS.
Almost any mild degree of dysregulation feature that you wish to propose will have
been reported, but not consistently. This is one of the problems that I have.
Prior: I’d like to expand the concepts a little more. Some time ago Kaplan and
Adams showed that stressed, non-dominant premenopausal cyclic female monkeys
continued to have regular cycles but had disturbed ovulation (Kaplan et al 1986,
Adams et al 1985). Disturbed ovulation was associated with obliteration of
the normal female protection against atherosclerosis. In those early studies,
androgens weren’t reported.
60 DISCUSSION

Bj˛rntorp: I know Carol Shively very well; we have a close contact. I asked her if
she had looked at androgens. She tells me that they are elevated in females.
Prior: There are two types of non-ovulation that di¡er remarkably in many of
their outcomes, especially in terms of what happens to bone. A simplistic way of
looking at non-ovulation is as ‘turned on’ (meaning high luteinizing hormone
[LH], ovaries enlarged and increased androgen manifestations) or ‘turned o¡’
(low gonadotropins, normal or low-ish oestrogen levels). But these two types of
non-ovulation have a common morphology with polycystic ovaries. I think we
should hone our terminology and stop de¢ning a state that means androgen
excess by the same non-speci¢c morphological feature that also characterizes the
very di¡erent hypothalamic non-ovulation (Prior 1997).
Burger: PCOS is a very di⁄cult model or analogy to use in the present discussion.
I have been quite taken by the sort of analysis that has come from Bart Fauser’s
group in Rotterdam, who has adopted a much more pragmatic approach to
the whole concept of what we call PCOS (Imani et al 2000). He looks at the
characteristics of women presenting because of infertility or coming to the
gynaecological endocrine clinic. He takes a number of the classic features that we
associate and then looks at their prevalence in individual women. He sees
overlapping Venn diagrams of a group who have insulin resistance, a group who
have hyperandrogenaemia, a group who have clinical hirsutism and a group with
obesity. They overlap, but it is only a small proportion who have the full
complement of all these features that were originally associated with the Stein^
Leventhal description of PCOS. It is this sort of analysis that will help us to
clarify what sort of terminology we should use. He uses it as a much more
pragmatic approach to say, ‘What is the prognosis of this individual and how
should they be managed? Let’s forget about trying to make a diagnosis of a
category to which only very few will actually belong.’ This is an illuminating
perspective.
Veldhuis: To me, this is a more honest treatment of the heterogeneity that exists.
This may help reduce some of the confusion caused by the clumping of patients in
reports. I’m personally intrigued that insulin is driving androgen secretion. When
we do in vitro studies with porcine theca cells, and transfect these with cDNA
encoding the 17a-hydroxylase promoter, if we add minute amounts of insulin
gene transcription goes up within 30^60 minutes. It is gorgeous! If you add a
tiny bit of LH with the insulin, there is synergy. Most clinicians ¢nd that there
are patients who are overly responsive to either insulin, or LH, or both. I want to
be careful not to attribute something downstream, such as androgens, to events
that are proximal or upstream. The theory of the serine phosphorylation defect
has not actually been extended, but this is the kind of thinking that may
eventually help us aggregate where appropriate and separate where appropriate.
A particular patient might have a polymorphism for the androgen receptor,
CORTICOTROPIC STRESS-RESPONSE AXIS 61

because a slight decrease in androgen feedback due to a slightly defective receptor

based on a polymorphism for the receptor causes her to oversecrete androgen. This
may not be the same as another patient who is also anovulatory and
hyperandrogenaemic with say a glutamine/alanine substitution in the promoter
of the 17a-hydroxylase gene. She in turn is di¡erent from an obese syndrome X
patient who has a primary defect in the insulin signalling pathway or in MAP
kinase. I am struggling with the looseness of the biological endpoints that all of
us depend on, and I am trying to drive some further precision in our focus of where
the defects are, which would help us ¢nd out why the Venn diagrams that Henry
Burger referred to are potentially separable.
Burger: Per Bjo«rntorp, how solidly established and how reproducible is the
evidence of dehydroepiandrosterone sulfate (DHEAS) hypersecretion in chronic
stress? My impression from having looked at the literature a while back is that it is
not very consistent. Is this how we can explain some patients with hirsutism who
otherwise have regular cycles and don’t seem to ¢t in the PCOS grouping?
Bj˛rntorp: I can’t answer that question.
Veldhuis: What might the notion of ‘burn-out’ better be de¢ned to mean? Are
you describing the chronically stressed animal without a novel stressor being
introduced? Even a chronically stressed animal with a novel stressor will show a
new stress response.
Bj˛rntorp: This is a confusing ¢nding. This kind of burned-out HPA axis is seen
in Vietnam veterans and Holocaust victims, for example, along with depression.
People who are experts in the ¢eld believe that this is because the HPA axis is just
worn out. In this situation there is apparently an increased activity of the
sympathetic nervous system as a compensatory event. In the burned out men I
showed here, which is about 9% of the population, we ¢nd background factors
which we interpret as causing stress, but not dramatically more than in the less
damaged men. My guess is that there is something else involved here.
Veldhuis: There are three clinical situations I can think of where the axis looks
fairly £at. In the post-op Cushing’s disease patients, the axis has been suppressed
and is fairly £at, unless you infuse CRH and then the whole axis wakes up. Post-
partum, there is a £at, monotonic, minimally pulsatile output. And there is a £at
HPA axis in some cases of chronic fatigue syndrome (which is a bit like PCOS in
that it is a poorly de¢ned syndrome). Those three conditions involve inferential
CRH de¢ciency states. If you model it on this basis, then you have to introduce
something that I think is important in discussing these axes, and that is to view
them as joint feedback and feed forward ensembles. For example, it would be
inappropriate to talk only about feedforward, because this is irrelevant unless
you can tell me the feedback state concurrently. If you give me any feedback
state, I can make it irrelevant by driving feedforward appropriately. Neither
feedback nor feedforward can be viewed in isolation. One of the things we may
62 DISCUSSION

need is a way to describe burn out pro¢les more formally. On the basis of all the
known data on feedback, are there some conditions that would never produce that
output in any sensible, easy manner? And are there other conditions that might? If
you give me this information and I have the laboratory funds, I might then do the
research to clarify some of the plausible options, and not immediately spend
precious sponsorship funds on the least likely theories. One of the challenges we
have clinically is that we have never assembled a clear, well de¢ned, mapped
network based on reasonable, sensible, minimal feedback connections.
Carroll: It is a very di⁄cult area to nail down in humans for the reason that I
mentioned before: the pituitary is sitting right in there interposed between your
sampling zone and your zone of physiological interest
the brain. As long as
that is the case, it is di⁄cult to interpret. Most of the literature says that if you use
ACTH and cortisol as your dependent variables for response to a stress, then in
the context of chronic stress, acute stress challenges lead to impairment of the
expected response. The same is true when CRH is administered. It is clear from
the depression literature that if you provoke insulin hypoglycaemia, the ACTH
and cortisol responses are blunted, and if you give CRH their response is blunted
both in cortisol and ACTH. But you can’t interpret that with reference to the
brain because the pituitary is sitting right there. If what I was saying earlier
today is correct, the pituitary is a prime site of GR action, and this may be the
answer. We have to come up with alternative experimental strategies. Johannes
Veldhuis and I have been looking at a low feedback strategy, but we don’t have
data analysed yet for that. This would be the approach: knockout the feedback
signal with metyrapone, use ACTH as the dependent variable and then do things
to the brain and see what happens. Plasma CRH is not a worthwhile measure in
this context, for short-term studies, because it gets diluted in the general
circulation.
Handelsman: Isn’t it better for that experiment to use adrenalectomy rather than
using drugs?
Carroll: That’s not so easy to do in humans!
Velduis: This opens the question of stressing the axis, trying to pin-point
responses and residual elements without watching the whole system change
simultaneously. Even this is di⁄cult, because of the triple nodes.
Carroll: Paul Plotsky and Paul Sovchenko refer to this as pharmacological
adrenalectomy. You have to be very careful about timing and dosage to make it
work.
Laron: The aetiology of so-called PCOS is more complicated. We have data that
shows insulin-like growth factor (IGF)1 over-dosage causes a PCOS-like
syndrome in females (Klinger et al 1998), and in males it increases LH (Laron &
Klinger 1998). I don’t think there are enough data showing whether insulin
competes on the IGF1 receptor in the patients with so-called PCOS. There may
CORTICOTROPIC STRESS-RESPONSE AXIS 63

be interplay, but it seems that the synergism of IGF1 with androgens is stronger
than that between insulin and androgens.
Veldhuis: When you say it produces a PCOS-like syndrome, given what we have
heard about the Venn diagram multiplicity, what kinds of features are you
referring to?
Laron: It causes hyperandrogenaemia, acne and interrupted menstruation,
which are all reversible when the dose is reduced.
Veldhuis: Have the ovaries in these patients been examined by ultrasound?
Laron: Some have, but they did not show the typical changes of PCOS, namely
the ¢brosis.
Veldhuis: This highlights the complexity of watching an endproduct when you
don’t know which receptor pathways are triggered.
Prior: I was going to make the clinical observation that when women become
overweight, they tend towards increased androgenicity; when men become
overweight they tend towards decreased androgenicity. There are some
fascinating epidemiological data taking the reported weight and height at age 18,
and showing anovulatory infertility in those who have a body mass index higher
than 24.5 (Rich-Edwards et al 1994). The relationship between body weight, LH
and androgenicity is a key here, somehow.
Veldhuis: I would add insulin to this.
Bj˛rntorp: I know what you are saying. The problem with these studies as I see
them is that people lump obesity into one pot. You have to separate central obesity
and peripheral obesity.
Prior: That is the way I thought of it. Also, we must consider muscularity
causing higher weight but not obesity.
Morley: Besides the paper on resistance, there is also a paper by Brˇning et al
(2000), in which the insulin receptor was speci¢cally knocked out in the brain.
These mice are very clearly hypogonadal. Females basically become fat, whereas
males don’t. This was used as an argument for insulin a¡ecting feeding. In fact, if
you knock out oestrogen you actually get fat females.
Brabant: It is unclear whether this is an e¡ect mediated by insulin on body fat or
simply via insulin acting on the gonadotroph.
Morley: The other thing in the recent literature that is going to change all of this
is a paper showing that fat cells produce a PPARg-related compound called resistin
(Steppan et al 2001). If that is hormonal-dependent or changes hormone it will
totally change the way we think of this. Most probably this is what is going to
happen.
Haus: We studied over 750 24 h pro¢les (6^8 samples/24 h) of hormonal and
biochemical variables in clinically healthy diurnally active subjects between 9 and
490 years of age. In comparison to young adult subjects (21 1.5 years of age), the
elderly (77 7 years of age) showed a drop in their circadian mean of serum
64 DISCUSSION

ACTH levels (44.6 6.4 pg/ml vs. 25.0 1.6 pg/ml) while serum b-endorphin
showed in the same subjects a statistically signi¢cant increase (3.9 0.2 pmol/l vs.
5.3 0.3 pmol/l) (Haus et al 1989, Nicolau et al 1991). This raises the question of a
possible di¡erence with ageing in the splitting of the parent molecule of
proopiomelanocortin from which both ACTH and b-endorphin are derived.
Plasma cortisol 24 h mean values varied very little between the same age groups
(10.5 0.3 mg/dl vs. 9.5 0.2 mg/dl) (Haus et al 1989, Lakatua et al 1992). However,
in the plasma ACTH:plasma cortisol ratio there was a very marked age di¡erence,
which was strictly time dependent. In the morning, the ratio was comparable
between the young and old subjects. However, during the afternoon and evening
there was a substantial di¡erence between the age groups. The younger subjects
showed a high amplitude circadian rhythm with rising values of the
ACTH:cortisol ratio in the afternoon reaching a peak at 00:00. With increasing
age the amplitude of this rhythm became less and the rhythm almost disappeared
in subjects above 80 (Lakatua et al 1992). This observation suggests in the elderly a
higher sensitivity of the adrenal to ACTH during the evening and early night hours
when the adrenal in the young subjects becomes less responsive to ACTH.
Veldhuis: This is the feedforward concept that I want to visualize.

References
Adams MR, Kaplan JR, Clarkson TB, Koritnik DR 1985 Ovariectomy, social status and
atherosclerosis in cynomolgus monkeys. Arteriosclerosis 5:192^200
Brˇning JC, Gautam D, Burks DJ et al 2000 Role of brain insulin receptor in control of body
weight and reproduction. Science 289:2122^2125
Elbers JM, Asscheman H, Seidell JC, Megens JA, Gooren LJ 1997 Long-term testosterone
administration increases visceral fat in female to male transsexuals. J Clin Endocrinol Metab
82:2044^2047
Haus E, Nicolau G, Lakatua DJ, Sackett-Lundeen L, Petrescu E 1989 Circadian rhythm
parameters of endocrine functions in elderly subjects during the seventh to the ninth decade
of life. Chronobiologia 16:331^352
Imani B, Eijkemans MJ, de Jong FH et al 2000 Free androgen index and leptin are the most
prominent endocrine predictors of ovarian response during clomiphene citrate induction of
ovulation in normogonadotropic oligomernorrheic infertility. J Clin Endocrinol Metab
85:676^682
Kaplan JR, Adams MR, Koritnik DR, Rose JC, Manuck SB 1986 Adrenal responsiveness and
social status in intact and ovariectomized Macaca fascicularis. Am J Primatology 11:181^193
Klinger B, Anin S, Silbergeld A, Eshet R, Laron Z 1998 Development of hyperandrogenism
during treatment with insulin-like growth factor I (IGF-I) in female patients with Laron
syndrome. Clin Endocrinol 48:81^87
Lakatua DJ, Nicolau GY, Sackett-Lundeen L, Petrescu E, Ortmeier T, Haus E 1992 Circadian
rhythm in adrenal response to endogenous ACTH in clinically healthy subjects of di¡erent age
groups. 5th International Meeting on Chronopharmacology and Chronotherapeutics, Amelia
Island, Florida, July 12^16, 1992, p 9^11
CORTICOTROPIC STRESS-RESPONSE AXIS 65

Laron Z, Klinger B 1998 E¡ect of insulin-like growth factor I treatment on serum androgens and
testicular and penile size in males with Laron syndrome (primary growth hormone resistance).
Eur J Endocrinol 138:176^180
Lovejoy JC, Bray GA, Bourgeois MO et al 1996 Exogenous androgens in£uence body
composition and regional body fat distribution in obese postmenopausal women
a
clinical research center study. J Clin Endocrinol Metab 81:2198^2203
Nestler JE, Barlascini CO, Matt DW et al 1989 Suppression of serum insulin by diazoxide
reduces serum testosterone levels in obese women with polycystic ovary syndrome. J Clin
Endocrinol Metab 68:1027^1032
Nestler JE, Singh R, Matt DW, Clore JN, Blackard WG 1990 Suppression of serum insulin level
by diazoxide does not alter serum testosterone or sex hormone binding globulin levels in
healthy, nonobese women. Am J Obstet Gynecol 163:1243^1246
Nicolau GY, Haus E, Lakatua DJ et al 1991 Circadian periodicity of pituitary-adrenal
parameters in di¡erent age groups. 20th International Conference on Chronobiology, Tel
Aviv, Israel, June 16^21, 1991
Prior JC 1997 Ovulatory disturbances: do they matter? Can J Diagnosis (February) p 64^80
Rich-Edwards JW, Goldman MB, Willett WC et al 1994 Adolescent body mass index and
infertility caused by ovulatory disorder. Am J Obstet Gynecol 171:171^177
Steppan CM, Bailey ST, Bhat S et al 2001 The hormone resistin links obesity to diabetes. Nature
409:307^312
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Male reproductive ageing: human

fertility, androgens and hormone
dependent disease
David J. Handelsman

ANZAC Research Institute, Department of Andrology, Concord Hospital, University of

Sydney, Sydney, NSW 2139, Australia

Abstract. The waning of male virility with age, in all its ambiguities, has always intrigued
humans, prompting innumerable approaches to staving it o¡. In modern terms,
advancing age impacts on all aspects of male reproductive health
sexuality, fertility
and androgenization
with di¡ering extents and tempo. With ageing, male sexual
function declines predominantly due to vasculogenic defects in cavernosal
haemodynamics, whereas libido and ejaculation are less a¡ected. This raises the
potential for prevention and treatment of erectile dysfunction as an early clinical
manifestation of atherosclerosis. After maturity, male fertility persists throughout life
but decreases modestly with age presumably due to concomitant decline in sexual
activity rather than in sperm output or function, although systematic population studies
of the latter are di⁄cult. Male ageing is associated with a progressive, partial and variable
degree of androgen de¢ciency, but the clinical and public health signi¢cance remain to be
established. Available evidence suggests that androgen supplementation is unlikely to
prolong life expectancy but might improve quality of life through prevention of
apparently age-related declines in androgen-sensitive tissues. The appropriate target
population, treatment modality and objectives remain to be established by further
controlled clinical studies.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 66^81

Over 50 years ago WHO de¢ned health as ‘A state of complete physical, mental and
social well-being and not merely the absence of disease or in¢rmity’. Analogously,
male reproductive health can be considered as having three dimensions

sexuality, fertility and androgenization
each a¡ected by ageing.
This review highlights two historical themes. The ¢rst is the importance of the
great contributions of 20th century statistics to medicine
randomization and
placebo controls
as the fundamentals of prospective or experimental studies.
These features control not only known, but crucially unknown covariables, that
in£uence outcomes. By contrast, retrospective or observational studies even if
based on the counsel of perfection
valid population-based sampling frames
66
MALE REPRODUCTIVE AGEING 67

and objective validation of study measures
only control for known variables.
This unique power gives salience to experimental over observational designs
thereby creating a distinction between strong and weak inference in biomedical
science. These ideas lie at the root of evidence-based medicine, a welcome
modern trend in medical science even if somewhat misnamed. Its focus is not so
much on evidence per se (as subjective or anecdotal experience constitute
evidence), as on the quality of evidence. Ironically, the rigour of evidence-based
medicine rests upon subjective judgement of the qualities of evidence.
Nonetheless, the robustness of the criteria of randomization and placebo controls
are themselves ultimately susceptible to objective judgement.
The second historical theme is the rami¢cations of Santayana’s aphorism that
those not familiar with their history are condemned to repeat it. This has
particular resonance for male reproductive ageing research as medical history
over the last two centuries has experienced repeated colourful episodes of
rejuvenation quackery, including doctors, masquerading as science. Preying on
the profound and endlessly resurgent but unattainable human desire for
rejuvenation will always be fertile territory for the plausible quack to harvest rich
pickings
and the crop is £owering once again.

Sexuality
Since antiquity, history and literature are replete with references to the waning of
male virility, in all its ambiguities, with age. The seemingly inexhaustible
repertoire of remedies to stave it o¡ testi¢es to the ubiquity of this human
¢xation. In particular, the decline in male sexuality in old age, being both
emblematic and problematic, it is not surprising that aphrodisiacs are among the
most ancient and ubiquitous of folk remedies. The modern educated layman,
however, having lost faith in rhino horn and tiger penis, appears now bedazzled
by the misty allure of hormones, the contemporary folkloric embodiment of the
fabled aphrodisiac. And for hormones, our social organization dictates that
doctors remain its monopolistic dispensers.
Systematic sociological studies of human sexuality were pioneered by Kinsey in
mid 20th century followed by more recent comprehensive national data from the
UK (Wellings et al 1994), USA (Laumann et al 1994) and France (Messiah &
Pelletier 1996). Their focus has been on men under 60 years leaving an
incomplete picture. As an observational methodology, such research aspires to
high standards in representative sampling; its limitations are in its lacking
independent, objective validation of ¢ndings. It is hardly surprising that
uncorroborated self-report (to strangers) of the most emotive, and hence least
likely to be reliably reported, aspect of human behaviour would produce glaring
discrepancies. For example, *50% of men but only *5% of women over the age
68 HANDELSMAN

of 80 report ongoing sexual activity with a partner (Laumann et al 1994). If these

facts are dubious, more subtle motivations and manifestations must be well beyond
reach. Bedevilled by its subjectivity and imperilled by proximity to social mores,
the prospects for more objective research into male sexual activity in the general
population seem remote. Yet large lacunae in understanding of male sexuality
remain, among them the role of prostitution which, given its ubiquity and
pervasiveness, is remarkably little studied. If prostitution is the oldest profession,
then use of its services must be the oldest male pastime. While the sociology of
prostitutes is widely studied, their male clients remain obscure and ignored. It is
hard to imagine that studies of such a ubiquitous and pervasive facet of male
sexuality would not enlighten the understanding of male sexuality.
By contrast, medical complaints of male sexual dysfunction undoubtedly
increase with age. While all functional aspects of male sexuality
libido,
potency and ejaculation
are reduced in older men, the major impact of ageing
is the exponential increase of erectile dysfunction with age (Ayta et al 1999). The
last decade has seen a major paradigm shift in understanding of male sexual
dysfunction. This includes a public relations makeover in nomenclature
(impotence becoming erectile dysfunction), a shift in beliefs about causation
(from prevailing Freudian psychoanalytic beliefs on psychosomatic causation, to
erectile dysfunction being due to organic neurovascular disorders of cavernosal
hydraulics) and development of the ¢rst e¡ective pharmacotherapy.
The development of vasoactive cavernosal pharmacotherapy was inaugurated
by the accidental discovery of papaverine e¡ects during vascular surgery and
soon surpassed by the equally fortuitous discovery of oral sildena¢l during
clinical drug evaluation for other purposes. Both are striking instances of
Pasteur’s dictum of chance favouring the prepared mind rather than outcomes of
goal-oriented research. Stemming from recognizing the vasculogenic basis of most
age-related erectile dysfunction, important issues arising include the role of erectile
dysfunction as a sentinel presentation of cardiovascular disease, the potential for
primary and secondary prevention strategies based on cardiovascular disease
pathogenesis (Ayta et al 1999) and elucidating the pathogenic mechanisms of the
most frequent iatrogenic cause of erectile dysfunction, anti-hypertensive drug
therapy (Feldman et al 2000). While the lucrative new pharmaceutical market for
vasoactive drugs will ensure expanding options for treatment of individuals with
erectile dysfunction, the high and increasing prevalence means the prospective
impact on public health policy and ¢nancing is daunting (Ayta et al 1999).

Fertility
Paternity is well established at the oldest age so the natural history of male fertility
is concluded only by death. By contrast, female fertility, terminating naturally at
MALE REPRODUCTIVE AGEING 69

menopause, occupies only half of average adult life expectancy. Despite the
enduring fertility potential, fathers older than 50 years are responsible for only
*1% of births in developed countries. Communal procreative patterns are
determined by the similarity of couple’s age and the overwhelming age-restriction
of female fertility. Nevertheless, fertility concerns of men cannot be disregarded at
any set age particularly with increasing remarriage to younger women.
Unfortunately the biomedical literature usually regards male fertility evaluation
as synonymous with semen analysis. Such reductive logic risks overlooking
important details of the whole picture and, coupled with the fact that men only
provide semen samples when concerned about their fertility (Handelsman 1997),
virtually all research on human sperm involves convenience samples inherently
unrepresentative of the general male population. Tunnel vision in appraising the
limitation of such research leads predictably to grievous misunderstanding
(Handelsman 2000).
In evaluating age e¡ects on male fertility, it is necessary to note that male fertility
is measured by counting conceptions, not sperm. Consequently, estimation of male
fertility has been largely the domain of demographers. They, in turn, focus almost
exclusively on female fertility, for reasons only partly explained by the simplistic
axiom that only maternity is ascertainable. The little evidence available suggests
that male fertility declines modestly with age (Anderson 1975, Ford et al 2000).
In these studies, it remains unclear if attempts to adjust for the highly correlated
e¡ects of wife’s age are su⁄cient to exclude that the modest decline in male fertility
is only illusory. If real, such modest declines in male fertility may be due to age-
related reduction in coital rate, sperm production or function, none being easy to
study in the general population.
Apart from indirect evidence from paternity, the few studies of sperm output
available are restricted to small convenience samples of infertile, older men.
These suggest sperm output is undiminished up to the age of 50 years but
systematic studies of any quality are limited beyond that age to small convenience
samples. Similarly, as both testicular histology and sperm function require biopsy
and semen samples, respectively, there are equally no valid population-based data
to evaluate age e¡ects on spermatogenic histology or sperm function.
Given the di⁄culties of evaluating spermatogenesis by studies requiring semen
analysis, valid surrogate variables would be useful. Indeed, semen analysis itself is
an imperfect surrogate marker for male fertility for which questionnaire
instruments have been developed (Levine 1988, Jo¡e 1997). Testicular volume
provides an excellent surrogate marker of spermatogenesis because seminiferous
tubules comprise the bulk of testis volume as noted in the clinical observation that
testicular atrophy reliably connotes impaired spermatogenesis in infertile men.
Lacking any longitudinal studies of testis size, the largest cross-sectional study
indicates that testis volume is little a¡ected by age until the eighth decade, if the
70 HANDELSMAN

e¡ects of terminal illness or vascular disease (Regadera et al 1985) are distinguished

(Handelsman & Staraj 1985). Smaller post-mortem (Johnson 1986) and
ultrasound (Lenz et al 1993) studies are consistent with these observations.
Cross-sectional studies showing consistently that blood follicle-stimulating
hormone (FSH) increases and inhibin B decreases progressively with advancing
age while their inverse relationship is maintained suggest that these hormone
markers, together with testis ultrasound, could constitute useful surrogate
markers of spermatogenesis for future population-based studies of human
ageing.

Androgen e¡ects
The 20th century saw dramatic prolongation of life expectancy in developed
countries. More people living longer in retirement creates a premium on
strategies to promote healthy ageing (meaning, in this context, maintenance of
enjoyable and independent living for the longest time with compression of
morbidity into the shortest timeframe at the end of life). A focus on gainful
coexistence with chronic ailments supplants the eradication of disease.
Ameliorating disabilities that accrue during age could enhance quality of life
regardless of its prolongation. Among strategies to improve health and well-
being of the elderly, the judicious evidence-based use of hormonal therapy clearly
warrants exploration. An important caveat is that since the best available evidence
suggests that androgen replacement therapy does not in£uence life-expectancy
(Handelsman 1998), bene¢cial e¡ects of physiological androgen dosages are
likely to be restricted to quality of life, for which instrumental measures remain
inadequate.

‘You can’t make a pig grow by weighing it’: observational studies

As testicular endocrine function can be evaluated from blood samples, large
community-based epidemiological studies were feasible since the advent of
radioimmunoassay in the 1960s. The now accepted consensus that blood
testosterone concentrations decline gradually after mid-life (Gray et al 1991a) was
secured by decisive evidence from the population-based Massachusetts Male
Ageing Study (Gray et al 1991b). This decline of *1% per annum, accompanied
by rising blood sex-hormone-binding globulin (SHBG), luteinizing hormone
(LH) and FSH concentrations, is accelerated by concomitant medical disorders
which accumulate with age. It is sobering that three generations of observational
studies using various convenience samples had not achieved de¢nitive
conclusions, generating futile controversy for two decades, before the pivotal
importance of representative sampling was fully appreciated. Key aspects of the
MALE REPRODUCTIVE AGEING 71

descriptive epidemiology of androgenic status of older men still requiring

clari¢cation include (a) whether androgenic thresholds, sensitivity and e¡ects
di¡er between ethnic groups, individuals and tissues, and (b) which older
men and what biological e¡ects are the best targets for androgen therapy.

‘The Emperor’s New Clothes’: measuring testosterone

Like most steroids and drugs, testosterone binds with varying a⁄nity to
circulating proteins notably SHBG and albumin. It is often stated or assumed
that various measures of ‘free’ testosterone are superior markers of androgen
supply to, and/or net e¡ects upon, androgen sensitive tissues than measurement
of total testosterone, the gold standard for biochemical con¢rmation of androgen
de¢ciency. The so-called ‘free hormone’ hypothesis has a long and tortuous history
(Edwards & Ekins 1988, Pardridge 1988, Mendel 1989). Regarding testosterone,
claims of superiority for derived testosterone assays are based exclusively on
theoretical arguments without clinical validation; yet even in theory, freer
movement of unbound testosterone into tissues would be just as likely a priori to
result in faster metabolic inactivation as enhanced biological e¡ects on target
tissues.
Derived testosterone assays include direct measurement of free (non-protein-
bound) testosterone by equilibrium dialysis, centrifugal ultra¢ltration or
analogue immunoassay, or indirect measures of ‘bioavailable’ (non-SHBG
bound) testosterone. ‘Free’ testosterone can also be calculated from total
testosterone and SHBG concentrations either as a simple ratio (Kapoor et al
1993) or a more complex calculation (S˛dergrd et al 1982). Claims of superior
clinical utility have not been supported by direct empirical proof; some derived
testosterone measures are invalid on theoretical (Kapoor et al 1993) and/or
empirical (Winters et al 1998) grounds. Given their weak rationale, absent
empirical validation of alleged superiority and less accurate measurements, there
seems little to recommend such measures in routine clinical use. The curious gulf
between their validation and popularity may re£ect their common distinctive
feature, that of demonstrating a faster fall with age than total testosterone.
Clearer separation between younger and older men serves to in£ate the
proportion of older men de¢ned in this manner as having biochemical ‘androgen
de¢ciency’. This circular logic however cannot evade the need to prove e⁄cacy and
safety of androgen therapy in older men. Such confusion of ends and means may be
alleviated by distinguishing androgen replacement therapy, which assumes a
de¢ciency state to be recti¢ed by physiological testosterone doses, and
pharmacological androgen therapy which aims for proof of e⁄cacy and safety
regardless of drug, dose or de¢ciency state.
72 HANDELSMAN

‘If I had a hammer, I’d hammer in the morning’: interventional therapeutic studies

Following the adage that when one’s only tool is a hammer, remarkably soon all
problems turn into nails, it is an understandable that androgen administration is
proposed for older men. Such proposals long pre-date the ¢rst availability of
testosterone (Hamilton 1937) with many bouts of rejuvenation quackery
associated with the names of Brown-Se¤ quard, Steinach and Vorono¡ into the
early 20th century. Standard clinical endocrine practice includes replacement
therapy for unequivocal hormonal de¢ciencies of the pancreas (insulin), thyroid
(thyroxine), adrenal (glucocorticoid, mineralocorticoid) and gonads (oestrogen,
androgen). Conversely, hormone replacement is not provided for other classical
hormones such as prolactin, glucagon, somatostatin, calcitonin, calcitonin-gene
related peptide and adrenalin, and other hormones (thyroid-stimulating
hormone, LH, FSH, parathyroid hormone) are not replaced but substituted by
simpler non-peptide end products. Generally, the criteria for a treatable
hormonal de¢ciency include (a) a well de¢ned clinical de¢ciency state, (b)
availability of su⁄cient clinical grade hormones, and (c) convincing evidence for
therapeutic e⁄cacy and safety. The £exibility of this categorization is illustrated by
the changing status of adult growth hormone replacement therapy based on
emerging evidence (Carroll et al 1998).
Certain features of ageing men resemble those observed in androgen-de¢cient
younger men, notably decreased lean body mass (muscle) and bone; reduced body
hair growth, skin thickness and dermal sebum secretion; impaired cognitive
function and mood; increased adiposity; and reduced strength, endurance,
initiative, virility and sense of well-being. Since androgen replacement in
younger men can reverse muscle, bone and mental changes of androgen
de¢ciency, it is a reasonable postulate that the partial androgen de¢ciency may
contribute to the physical frailty and mental torpor of older men. Nevertheless, it
remains unclear whether (a) blood testosterone concentrations fall far enough to
warrant replenishment, (b) tissue androgenic thresholds change with age, and (c)
older tissues remain su⁄ciently androgen responsive. In short, the biological
signi¢cance of the gradual, partial and variable decline in blood testosterone
concentrations in older men cannot be established from observational studies
alone but requires critical evaluation by interventional therapeutic studies.
Ad hoc trials of androgen therapy in ageing men dating back to the ¢rst
availability of testosterone (Heller & Myers 1944) were unconvincing as they
lacked the decisive features of placebo controls and randomization. Several
recent small controlled studies of at least three months’ duration demonstrate
inconsistent bene¢ts of testosterone supplementation (Marin et al 1992, Tenover
1992, Morley et al 1993, Sih et al 1997) but most had signi¢cant limitations in
design (inadequate masking, poorly de¢ned clinical end-points, low power).
MALE REPRODUCTIVE AGEING 73

A major study of generally healthy men over 65 years with a low plasma
testosterone (516.5 nmol/l) treated with daily transdermal testosterone for three
years showed an improved sense of well-being, increased lean mass and decreased
fat mass compared with placebo (Snyder et al 1999a,b). Crucially, however,
testosterone did not consistently improve muscular function or bone density
compared with placebo. Secondary analyses showed any bene¢ts of testosterone
in older men were limited to those with overt androgen de¢ciency.
Consequently, at present androgen supplementation in ageing men is appropriate
for overt androgen de¢ciency, but more liberal use is still best considered
promising but unproven.
Future studies should aim for more powerful design, better focus on appropriate
subgroups of men and end-points likely to bene¢t, and/or alternative hormonal
regimens. Pharmacological androgen therapy, using supraphysiological doses or
novel synthetic androgens, might improve muscle, bone or other androgen-
dependent functions in older men regardless of androgen de¢ciency status,
nature or dose of androgen. Viewed like any other anti-ageing treatment, this
would require evidence of e⁄cacy, safety and cost-e¡ectiveness from controlled
trials rather then relying on supposed replacement status to lighten the burden of
proof for e⁄cacy and safety. This approach would diversify androgen therapies to
allow enhanced targeting of androgen therapy via exploiting variations in tissue
selectivity and metabolic activation (5a reduction, aromatization) pro¢les
(Sundaram et al 1994) that could be developed in novel potent designer
androgens (Dalton et al 1998, Edwards et al 1998). Regardless of the speci¢c
androgen, the safety issues remain whether androgen therapy in£uences
progression of prostate or cardiovascular disease or precipitates idiosyncratic
e¡ects (polycythaemia, sleep apnoea, £uid retention, behavioural disturbance).
The key long-term e¡ects on cardiovascular and prostate disorders would require
large, long-term vigilance studies as were undertaken for oestrogen therapy in
menopause (Handelsman 1998).

Androgen-dependent disorders (prostate and cardiovascular disease)

A salient fact of ageing men’s health is that cardiovascular disease is the major cause
of death, and prostate disease the leading cause of both major surgery and new
internal cancers. Both prostate and cardiovascular disease demonstrate distinctive
hormone-dependent features, with long incubation periods during early life
culminating in exponentially increasing age-related incidence during late life.
These features create opportunities for novel preventive strategies (Handelsman
1998). Among promising options for future developments is the potential to
develop designer androgens (Dalton et al 1998, Edwards et al 1998) that exploit
tissue-selective metabolic activation of testosterone (Sundaram et al 1994) via
74 HANDELSMAN

the 5a reductase pathway, a local androgen ampli¢cation mechanism, or

diversi¢cation of androgen action via aromatase which leads to e¡ects mediated
via the oestrogen receptor.
In the near future the most important developments with predictable potential
for improved management of prostate disease in ageing men include (a)
development of better tests to make wide-scale prostate cancer screening feasible
and (b) Finasteride Prostate Cancer Chemoprevention study (Brawley & Parnes
2000) which will determine whether selective inhibition of 5a reductase can
reduce prostate cancer diagnosis and mortality with great implications for
development of prostate-sparing androgens for supplementation.
Arguably the most fundamental fact in cardiovascular epidemiology is that men
have earlier and more severe morbidity and mortality than women of similar age.
Despite prevailing orthodoxy based on retrospective case^control studies, it has
long been evident that natural menopause does not accelerate cardiovascular
disease (Heller & Jacobs 1978, Tunstall-Pedoe 1998), unlike breast cancer or
bone density, where oestrogen dependency is clearer. This uncertainty is
accentuated by the ¢rst prospective studies of oestrogen replacement therapy in
menopausal women demonstrating no bene¢t, and even detrimental e¡ects, on
secondary prevention of cardiovascular disease (Hulley et al 1998). As life-long
castration appears not to alter life-expectancy (Handelsman 1998), endogenous
testosterone has at best only modest e¡ects in promoting cardiovascular disease,
perhaps only in subpopulations. While recent studies have rediscovered the
vasodilatory properties of androgens (Ja¡e 1977) and that low endogenous
testosterone is a risk factor for cardiovascular disease (Alexandersen et al 1996),
these ¢ndings need to be reconciled with the likelihood that androgens may also
foster early stages of atherogenesis (McCrohon et al 1997, 1999, 2000, McCredie
et al 1998).

Diagnosis and management of androgen de¢ciency in older men

The diagnosis of androgen de¢ciency, primarily a clinical diagnosis con¢rmed by
appropriate hormone assays, is generally unambiguous in younger men with
hypothalamic^pituitary or testicular disorders. Among older men the non-
speci¢c symptoms of androgen de¢ciency may have other causes and biochemical
diagnosis is uncertain unless blood testosterone concentrations are unambiguously
low. Appropriate reference ranges for blood testosterone concentrations are
controversial
adopting the reference range of healthy young men assumes age-
dependent changes in blood testosterone levels are inherently pathological whereas
an age-adjusted reference range might overlook recti¢able androgen de¢ciency. A
reliable, independent marker of net tissue androgen e¡ect would help resolve this
dilemma but none exists. The androgenic threshold for sexual function is low
MALE REPRODUCTIVE AGEING 75

whereas those for muscle, bone and cognitive function are unknown but likely to
be higher, and may vary within and between individuals. Consequently, driven by
the need for practical guidance, free of marketing hype in an environment of
commercially driven disease-mongering, the Endocrine Society of Australia has
recently published consensus best practice guidelines for androgen prescribing
(Conway et al 2000) which were adopted by the national Pharmaceutical Bene¢ts
Scheme for subsidy of e¡ective medicines. Revision based on further evidence and
other national guidelines will undoubtedly emerge.

References
Alexandersen P, Haarbo J, Christiansen C 1996 The relationship of natural androgens to
coronary heart disease in male: a review. Atherosclerosis 125:1^13
Anderson BA 1975 Male age and fertility: results from Ireland prior to 1911. Population Index
41:561^567
Ayta IA, McKinlay JB, Krane RJ 1999 The likely worldwide increase in erectile dysfunction
between 1995 and 2025 and some possible policy consequences. BJU Int 84:50^56
Brawley OW, Parnes H 2000 Prostate cancer prevention trials in the USA. Eur J Cancer 36:1312^
1315
Carroll PV, Christ ER, Bengtsson BA et al 1998 Growth hormone de¢ciency in adulthood and
the e¡ects of growth hormone replacement: a review. Growth Hormone Research Society
Scienti¢c Committee. J Clin Endocrinol Metab 83:382^395
Conway AJ, Handelsman DJ, Lording DW, Stuckey B, Zajac JD 2000 Use, misuse and abuse of
androgens: The Endocrine Society of Australia consensus guidelines for androgen
prescribing. Med J Aust 172:220^224
Dalton JT, Mukherjee A, Zhu Z, Kirkovsky L, Miller DD 1998 Discovery of nonsteroidal
androgens. Biochem Biophys Res Commun 244:1^4
Edwards JP, West SJ, Pooley CL, Marschke KB, Farmer LJ, Jones TK 1998 New nonsteroidal
androgen receptor modulators based on 4-(tri£uoromethyl)-2(1H)-pyrrolidino[3,2-g]
quinolinone. Bioorg Med Chem Lett 8:745^750
Edwards P, Ekins R 1988 The ‘‘Pardridge’’ hypotheses relating to the role of hormone binding
proteins in hormone delivery: a critique. Steroids 52:367^368
Feldman HA, Johannes CB, Derby CA et al 2000 Erectile dysfunction and coronary risk factors:
prospective results from the Massachusetts male aging study. Prev Med 30:328^338
Ford WC, North K, Taylor H, Farrow A, Hull MG, Golding J 2000 Increasing paternal age is
associated with delayed conception in a large population of fertile couples: evidence for
declining fecundity in older men. The ALSPAC Study Team (Avon Longitudinal Study of
Pregnancy and Childhood). Hum Reprod 15:1703^1708
Gray A, Berlin JA, McKinlay JB, Longcope C 1991a An examination of research design e¡ects
on the association of testosterone and male aging: results of a meta-analysis. J Clin Epidemiol
44:671^684
Gray A, Feldman HA, McKinlay JB, Longcope C 1991b Age, disease, and changing sex
hormone levels in middle-aged men: results of the Massachussetts Male Aging Study. J Clin
Endocrinol Metab 73:1016^1025
Hamilton JB 1937 Treatment of sexual underdevelopment with synthetic male hormone
substance. Endocrinology 21:649^654
Handelsman DJ 1997 Sperm output of healthy men in Australia: magnitude of bias due to self-
selected volunteers. Hum Reprod 12:2701^2705
76 HANDELSMAN

Handelsman DJ 1998 The safety of androgens: prostate and cardiovascular disease. In: Wang C
(ed) Male reproductive function. Kluwer Academic Publishers, Norwell, MA, p 173^190
Handelsman DJ 2000 Myth and methodology in the evaluation of human sperm output. Int J
Androl 23:50^53
Handelsman DJ, Staraj S 1985 Testicular size: the e¡ects of aging, malnutrition and illness. J
Androl 6:144^151
Heller CG, Myers GB 1944 The male climacteric, its symptomatology, diagnosis and treatment.
J Am Med Assoc 126:472^427
Heller RF, Jacobs HS 1978 Coronary heart disease in relation to age, sex, and the menopause. Br
Med J 1:472^474
Hulley S, Grady D, Bush T et al 1998 Randomized trial of estrogen plus progestin for secondary
prevention of coronary heart disease in postmenopausal women. Heart and Estrogen/
progestin Replacement Study (HERS) Research Group. JAMA 280:605^613
Ja¡e MD 1977 E¡ect of testosterone cypionate on postexercise ST segment depression. Br Heart
J 39:1217^1222
Jo¡e M 1997 Time to pregnancy: a measure of reproductive function in either sex. Asclepios
Project. Occup Environ Med 54:289^295
Johnson L 1986 Spermatogenesis and aging in the human. J Androl 7:331^354
Kapoor P, Luttrell BM, Williams D 1993 The free androgen index is not valid for adult males. J
Steroid Biochem Mol Biol 45:325^326
Laumann EO, Gagnon JH, Michael RT, Michaels S 1994 The social organization of sexuality.
University of Chicago Press, Chicago, IL
Lenz S, Giwercman A, Elsborg A et al 1993 Ultrasonic testicular texture and size in 444 men from
the general population: correlation to semen quality. Eur Urol 24:231^238
Levine RJ 1988 Monitoring fertility to detect toxicity to the male reproductive system. Reprod
Toxicol 2:223^227
Marin P, Krotkiewski M, Bj˛rntorp P 1992 Androgen treatment of middle-aged, obese men:
e¡ects on metabolism, muscle and adipose tissues. Eur J Med 1:329^336
McCredie RJ, McCrohon JA, Turner L, Gri⁄ths KA, Handelsman DJ, Celermajer DS 1998
Vascular reactivity is impaired in genetic females taking high-dose androgens. J Am Coll
Cardiol 32:1331^1335
McCrohon JA, Death AK, Nakhla S et al 2000 Androgen receptor expression is greater in
macrophages from male than from female donors. A sex di¡erence with implications for
atherogenesis. Circulation 101:224^226
McCrohon JA, Jessup W, Handelsman DJ, Celermajer DS 1999 Androgen exposure increases
human monocyte adhesion to vascular endothelium and endothelial cell expression of
vascular cell adhesion molecule-1. Circulation 99:2317^2322
McCrohon JA, Walters WA, Robinson JT et al 1997 Arterial reactivity is enhanced in genetic
males taking high dose estrogens. J Am Coll Cardiol 29:1432^1436
Mendel CM 1989 The free hormone hypothesis: a physiologically based mathematical model.
Endocr Rev 10:232^274
Messiah A, Pelletier A 1996 Partner-speci¢c sexual practices among heterosexual men and
women with multiple partners: results from the French national survey, ACSF. Analyse des
Comportements Sexuel en France. Arch Sex Behav 25:233^247
Morley JE, Perry HM 3rd, Kaiser FE et al 1993 E¡ects of testosterone replacement therapy in old
hypogonadal males: a preliminary study. J Am Geriatr Soc 41:149^152
Pardridge WM 1988 Selective delivery of sex steroid hormones to tissues by albumin and by sex
hormone-binding globulin. Oxf Rev Reprod Biol 10:237^292
Regadera J, Nistal M, Paniagua R 1985 Testis, epididymis, and spermatic cord in elderly men.
Correlation of angiographic and histologic studies with systemic arteriosclerosis. Arch Pathol
Lab Med 109:663^667
MALE REPRODUCTIVE AGEING 77

Sih R, Morley JE, Kaiser FE, Perry HM 3rd, Patrick P, Ross C 1997 Testosterone replacement in
older hypogonadal men: a 12-month randomized controlled trial. J Clin Endocrinol Metab
82:1661^1667
Snyder PJ, Peachey H, Hannoush P et al 1999a E¡ect of testosterone treatment on bone mineral
density in men over 65 years of age. J Clin Endocrinol Metab 84:1966^1772
Snyder PJ, Peachey H, Hannoush P et al 1999b E¡ect of testosterone treatment on body
composition and muscle strength in men over 65 years of age. J Clin Endocrinol Metab 84:
2647^2653
S˛dergrd R, Bckstr˛m T, Shanbhag V, Carstensen H 1982 Calculation of free and bound
fractions of testosterone and estradiol-17 beta to human plasma proteins at body
temperature. J Steroid Biochem 16:801^810
Sundaram K, Kumar N, Bardin CW 1994 7-a-methyl-19-nortestosterone: an ideal androgen for
replacement therapy. Recent Prog Horm Res 49:373^376
Tenover JS 1992 E¡ects of testosterone supplementation in the aging male. J Clin Endocrinol
Metab 75:1092^1098
Tunstall-Pedoe H 1998 Myth and paradox of coronary risk and the menopause. Lancet 351:
1425^1427
Wellings K, Field J, Johnson AM, Wadsworth J 1994 Sexual behaviour in Britain: the national
survey of sexual attitudes and lifestyles. Penguin Books, London
Winters SJ, Kelley DE, Goodpaster B 1998 The analog free testosterone assay: are the results in
men clinically useful? Clin Chem 44:2178^2182 (erratum: 1999 Clin Chem 45:444)

DISCUSSION
Veldhuis: This brings us well into the androgen discussion in ageing at the level
of the target tissue. John Morley, I thought it would be good to give you the
chance to comment.
Morley: What always amazes me is that David Handelsman and I agree on the
conclusions, but for totally di¡erent reasons. Of the available data, I don’t believe
Snyder’s study (Snyder et al 1999) was very rigorous for a variety of reasons. One
was the addition of Ca2+ and vitamin D willy-nilly in a group who were chosen to
be osteoporotic. This can a¡ect muscle strength among other things. That alone
questions that study. The other problem is that many of the men were not
hypogonadal: as David Handelsman says (and I agree), the data suggest that you
should treat hypogonadal males, not non-hypogonadal males who have drifted
down into the normal range. I think that the strong data that we are about to see
come from the unpublished study by Tenover. This study shows an improvement
in strength as well as improvement in bone mineral density. Similar unpublished
data are coming from the Baltimore group for upper limb strength. I think that the
improvement of upper limb strength with testosterone may well be real in
hypogonadal males. Bebb’s data look very similar, but he is also going to take a
while to publish. The Kenny paper is in press, and shows clear bone e¡ects without
Ca2+ and vitamin D addition in patients who were more hypogonadal than
Snyder’s. In a very small group, she found lower and upper limb strength
increased, but not quite to a signi¢cant level. If you put these unpublished
78 DISCUSSION

studies together with the other published data, you can conclude that testosterone
most probably has bene¢cial e¡ects, but it is in people who are hypogonadal. This
does not mean that every single man should get testosterone supplementation. The
other unanswered question
which I think is the big one
is whether high,
almost pharmacological doses are needed to produce e¡ects, or whether lower
doses, such as those delivered by patch techniques, are su⁄cient. We have a long
way to go; we clearly need long, well-controlled trials. We also need to know who
we are entering into these trials. I know some of these are in development, but so
far I’m not aware of any funding agency that has been prepared to study the 3000^
4000 men needed, equivalent to the women’s health studies, to really answer these
questions. Until we have these, we should treat men who are clearly hypogonadal,
no matter what their age, and there will be a good outcome. There are three
controlled trials that David didn’t mention. Two of these are from Portland
(Janowsky et al 1994, 2000) and there is another one from Matsumoto and
Bremner’s group (Cherrier et al 2001), all showing very positive e¡ects on
cognition. In fact, if you were to choose one thing that testosterone really seems
to help for the majority of middle-aged to elderly men, it is cognitive function.
Handelsman: Both of our studies have a variety of cognitive measures. We also
had gait, mobility and cognitive tests and quality of life measures: none were
signi¢cantly improved. In particular, in the human chorionic gonadotropin
(hCG) study we also had actigraphy measures for spontaneous physical activity
as well as insulin clamps in 40 men before and after treatment. There was no
signi¢cant change.
Veldhuis: And do you agree that the data are sparse on the route and dose-
dependency of androgen repletion?
Handelsman: The data are sparse altogether. There is just one other comment I’d
like to add putting the Snyder study in perspective, he got less than 100 patients out
of screening 1000. You have to keep this high degree of selectivity in mind. This is
with a testosterone level of 516 nM. If you insist on subjects being really
hypogonadal (i.e. 510 nM), you will have 1 in 1000. This would be a very small
subpopulation.
Morley: That is not necessarily true. He also acquired people who were
osteopaenic to some extent. They had to have a lower bone mineral density. He
had a number of other exclusion criteria. If you take the Massachusetts male
ageing study, this does not ¢t at all with what the population numbers are. It is
not what we found in California, and it is not what has been found by other
people in other areas.
Handelsman: In our study we also recruited about 10%. The main reason for
rejection was baseline testosterone concentrations.
Veldhuis: This permeates the area until we get a broad prospective open study. It
makes the ¢nal results very subgroup dependent.
MALE REPRODUCTIVE AGEING 79

Burger: How does one de¢ne hypogonadism in the male who is said to be in the
andropause? What are the criteria, and what range of testosterone concentrations
are we to use? Is the young normal range the appropriate one or should we be
working from a di¡erent one? These are crucial issues in this debate.
Veldhuis: Let me try to simplify this, does anyone have the nerve to give a
number?
Handelsman: The Endocrine Society of Australia was bold enough to make a stab
at this, and came up with a ¢gure of 8 nM, independent of age (Conway et al 2000).
Above this testosterone concentration, the diagnosis in absence of underlying
pituitary or testis pathology was regarded as unequivocal.
Morley: The bioavailable testosterone or free testosterone index may be better.
We use the bioavailable level of 70 ng/dl, which is about 3 nM. This is one of the
choices.
Veldhuis: Is that independent of age? We would consider that level profoundly
low in a young man.
Morley: In our lab, that is two standard deviations below the mean. We see no
young men below 3 nM.
Wang: We have lots of debate in the USA about the cut-o¡ level of testosterone
to de¢ne the andropause. If serum testosterone levels are below 250 ng/dl, the men
are hypogonadal and should be treated. The problem is when the levels are between
250 and 300. Unless we have large studies that can show that the bene¢t is
outweighing the risk, we are hesitant to proceed.
Shalet: Is it possible to build a gonadotropin component into these de¢nitions?
What about raised LH?
Veldhuis: In the allostatic state that we learned about this morning, if the LH is
elevated, you could argue that this is a physiological marker as an elevated thyroid
stimulating hormone (TSH) is in early thyroid failure.
Laron: I think that is a wrong terminology. We measure male hypogonadism by
millilitre volume of the testes. This shouldn’t be confused with the quantitative
secretion of androgens. They involve di¡erent cells.
Veldhuis: We heard that these volumes are pretty stable.
Morley: You could use andropause to get around this.
Veldhuis: I would argue that this is an age-dependent population-sensitive
measure. It is assay dependent; you can’t come up with a single number. I am
with Henry Burger: we have to develop norms that are pertinent for the patient
in question who is being compared. If the patient is a 78 year old individual, that
clearly represents a potential population base that is readily distinguishable under
usual health de¢nitions for all body systems. Their glomerular ¢ltration rate, lung
di¡usion capacity and bone mineral density are all reduced. If you are going to say
whether a person has a disease, beyond the fact that they have survived to age 78,
you have to show me other 78 year olds who are not complaining. He is entitled to
80 DISCUSSION

complain, but I may not attribute the complaint to that level solely if I ¢nd it
matches those of all other men of similar age and disposition who are not
complaining.
Riggs: Just to take the position of devil’s advocate, we have struggled in the bone
¢eld about whether t scores or z scores are more important. In other words, should
the normal range be that of young adults, or should it be age adjusted? We have
come down strongly on the side of the former, because there is a critical value for
bone density below which there is a marked increase in fracture risk, regardless of
age.
Veldhuis: We don’t have the targeted end point to tell us. I haven’t heard this
acknowledged. We need the dose-dependent target for di¡erent end points. They
are not necessarily the same for osteopaenia as for muscle weakness and for
decreased cognition.
Handelsman: If you are thinking about growth hormone studies in ageing, the
advantage is that dose can be titrated with insulin-like growth factor (IGF)1 as an
independent end-point. We don’t have any equivalent form of titration with
testosterone, which is a serious limitation. It is quite likely that the thresholds
and/or sensitivity are di¡erent for di¡erent tissues. The sensitivity may not be the
same for di¡erent age groups and populations.
Veldhuis: Absolutely. We have heard innuendos that the cognitive end-points
might be more responsive, but this remains to be proven. It’s an exciting issue. We
picked muscle gene expression because Randy Urban in Texas showed increased
IGF1 mRNA by two- to threefold in three weeks by titrated androgen replacement
in six men.
Ruiz-Torres: We are interested in the question of why prostrate volume is so
closely related to age, despite the fact that testosterone secretion declines. We
thought that this could be due to an increased number of androgen receptors of
the epithelium cells. However, we have found that the cytosolic receptors are
increased, but the nuclear receptors do not show any change. One possible
explanation is that the transfer from the nucleus to the cytoplasm is altered by
ageing. In any case, it seems that androgens are not the cause of hypertrophic
prostrate. Nevertheless, the benign prostatic hyperplasia (BPH) is mainly due to
an increase in smooth muscle cells where testosterone could play a role as an
anabolic e¡ector. But normally in ageing the testosterone levels are low, so that
this possibility concerns the treatment with androgens only.
Handelsman: It takes quite a bit of looking into what BPH and prostate growth
with age means. BPH is actually nodular growth: I have always thought of this
disease process as very much like ¢broids. It doesn’t necessarily occur smoothly,
even though there is that famous study which suggests that there is a small rise
(Berry et al 1984). In fact, it probably occurs in punctated bursts of growth, just
like ¢broids. The changes in the central zone are much more marked than those that
MALE REPRODUCTIVE AGEING 81

occur in the total prostate volume: we can see changes earlier and more
prominently in the central zones. The fact is that with dihydrotestosterone
(DHT) treatment, we didn’t see any prostrate growth. My explanation for this is
that using DHT you don’t get any intraprostrate ampli¢cation of androgen action.
Thus androgens which are not 5a reducible or activated would be expected to have
less e¡ect on the prostate. In balance, however, I have to add this may not actually
be an advantage. There are reasons why having a prostate that doesn’t have 5a
reduction or full growth is a disadvantage. The disadvantage is that ejaculate
volume is dependent on prostate £uids. If you don’t have this, it could be that
ejaculation may be experienced like in retrograde ejaculation, which may be
symptomatically unacceptable.
Veldhuis: Are there any data on androgen receptor localization or activity in
ageing? I recall diverse reports, one showing a change in the brain that was
opposite to that in the prostate, and one showing no change. Per Bj˛rntorp, you
alluded to an androgen receptor polymorphism. Do you know of any data on the
e¡ect of ageing on androgen receptor expression?
Bj˛rntorp: Yes, there are some microsatellites in the ¢rst exon which codes for the
transduction domain. When this is too short, there is an increased risk for prostate
cancer.
Veldhuis: That’s a polymorphism without a known change in receptor function.
Bj˛rntorp: I think these recent microsatellite developments are very interesting.
In the androgen receptor gene, for example, these microsatellites have been altered
in transgenic mice. The shorter the microsatellite, the stronger the e¡ect of the
produced protein.
Ruiz-Torres: The number of androgen receptors in the prostatic epithelium
depends on the location. We have found that the content of nuclear receptors of
cells from the posterior zone
where cancer usually appears
is higher than in
the case of the central zone, which is mostly a¡ected by BPH.

References
Berry SJ, Co¡ey DS, Walsh PC, Ewing LL 1984 The development of human benign prostatic
hyperplasia with age. J Urol 132:474^479
Cherrier MM, Asthana S, Plymate S et al 2001 Testosterone supplementation improves spatial
and verbal memory in healthy older men. Neurology 57:80^88
Conway AJ, Handelsman DJ, Lording DW, Stuckey B, Zajac JD 2000 Use, misuse and abuse of
androgens. The Endocrine Society of Australia consensus guidelines for androgen
prescribing. Med J Aust 172:220^224
Janowsky JS, Oviatt SK, Orwoll ES 1994 Testosterone in£uences spatial cognition in older
men. Behav Neurosci 108:325^332
Janowsky JS, Chavez B, Orwoll E 2000 Sex steroids modify working memory. J Cogn Neurosci
12:407^414
Snyder PJ, Peachey H, Hannoush P et al 1999 E¡ect of testosterone treatment on bone mineral
density in men over 65 years of age. J Clin Endocrinol Metab 84:1966^1972
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Male reproductive ageing: using the

Brown Norway rat as a model for man
Christina Wang*, Amiya Sinha Hikim*, Monica Ferrini{, Juan J. Bonavera*, Dolores
Vernet{, Andrew Leung*, Yan-He Lue*, Nestor F. Gonzalez-Cadavid{ and Ronald S.
Swerdlo¡*

Divisions of *Endocrinology and {Urology, Departments of Medicine and Surgery, Harbor-

UCLA Medical Center and Research and Education Institute, 1000 West Carson Street,
Carson, CA 90509, USA

Abstract: The Brown Norway (BN) rat is an excellent model for male reproductive
ageing. We and others have shown that with ageing, the BN rat exhibits low serum
testosterone, low Leydig cell steroidogenic capacity, decreased Sertoli cell function and
number, marked reduction in seminiferous tubule volume and sperm content, and
accelerated germ cell apoptosis. These testicular changes are the result of a combination
of a primary testicular defect and a secondary hypothalamic dysfunction. Leydig cell
dysfunction results from decreased activities of the steroidogenic enzymes and Leydig
cell secretory capacity and is not corrected by daily administration of replacement
luteinizing hormone (LH), suggesting a primary testicular defect. However ageing in
male BN rats is associated with decreased LH pulse amplitude, reduced gonadotropin
releasing hormone (GnRH) and gonadotropin responsiveness to excitatory amino
acids, and decreased GnRH mRNA and peptide in the hypothalamus. We have further
shown in the hypothalamus of ageing BN rats that while the excitatory amino acid
receptor content is reduced, nitric oxide synthase (NOS) activity is increased which is
due to increased inducible (iNOS) but not neuronal NOS (nNOS). The increased iNOS
protein in the hypothalamus is associated with increased peroxynitrite formation and
neuronal cell apoptosis. We conclude that increased hypothalamic levels of iNOS may
result in neurotoxicity in the hypothalamus leading to loss of hypothalamic GnRH
secretory cells and impaired GnRH pulsatile secretion that contributes to the abnormal
Leydig cell function characteristic of male reproductive ageing.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 82^97

Cross-sectional and longitudinal studies in men showed that serum testosterone

declined with ageing (Vermeulen 1991). Even though serum total testosterone
may be in the normal range, the free or bioavailable testosterone levels are
frequently lower in elderly men (Gray et al 1991, Korenman et al 1990, Baker et
al 1976, Harman & Tsitouras 1980). Low testosterone levels in older men are
associated with sexual dysfunction, loss of bone, decreased muscle mass and
82
REPRODUCTIVE AGEING IN THE MALE BROWN NORWAY RAT 83

strength, increased body fat, frailty, and poorer quality of life (Swerdlo¡ & Wang
1993). The low circulating serum testosterone levels are usually associated with
elevated serum luteinizing hormone (LH) and follicle-stimulating hormone
(FSH) levels in aged men. However, such increases in LH may be
inappropriately low compared to those of young men with similar low serum
testosterone levels. In search of an animal model to examine the process and
mechanisms of reproductive ageing, we and others have de¢ned the BN rat as the
most suitable model for male reproductive ageing (Zirkin et al 1993, Wang et al
1993, Gruenewald & Matsumoto 1991). The BN rat is a better model to study
male reproductive ageing than other strains because these rats have a longer
lifespan, do not develop pituitary or testicular tumours, are not excessively
obese, and manifest both testicular and hypothalamic-pituitary dysfunction with
ageing. In this chapter, we will describe the work in our laboratory using the BN
rat as a model for studying human male reproductive ageing.

Testicular dysfunction
Reproductive ageing in the BN rat is characterized by low serum testosterone
levels (Zirkin et al 1993, Wang et al 1993, 1999). Low serum testosterone levels
are also a hallmark of the ageing male demonstrated in both cross-sectional and
longitudinal studies (Vermeulen 1991, Gray et al 1991). We studied serum
testosterone levels and sperm concentration in 6, 9, 12, 15, 18 and 31 month-old
BN rats. Beginning at 15 months, plasma testosterone and inhibin levels both
showed a progressive decline with age (Fig. 1). Intratesticular testosterone and
inhibin concentrations were not lower, and in fact appeared to be higher in testes
that showed marked regression, because of the loss of seminiferous tubule content
in the testis (Wang et al 1993, 1999). The decrease in serum testosterone was due to
decreased Leydig cell secretory capacity as demonstrated by Zirkin et al (1993),
Zirkin & Chen (2000) and Chen et al (1994, 1996). The Leydig cell dysfunction
resulted from reductions in the levels and activities of steroidogenic enzymes (Luo
et al 1996). Leydig cells isolated from old rats produced signi¢cantly less basal and
LH-stimulated testosterone in vitro. If the Leydig cells were depleted by treatment
with ethane dimethanesulfonate and then allowed to regenerate, the capacity of the
Leydig cells in the old rats to produce testosterone was similar to young animals
for up to 10 weeks. Whether the capacity of these repopulated Leydig cells in old
rats would continue to produce testosterone in the older BN rats was not studied
(Chen et al 1996). The same investigators administered contraceptive doses of
testosterone (implants) to young and old rats for 8 months. Two months after
removal of testosterone implants, the Leydig cells of both the young and old rats
secreted high levels of testosterone. The investigators suggest that by placing the
Leydig cells in ‘hibernation’ the decreased Leydig cell steroidogenesis associated
84 WANG ET AL

FIG. 1. Plasma LH (A), FSH (B), testosterone (C), and inhibin (D) concentrations in ageing
rats. Data represent mean SE of 5^10 animals per group. *Indicates P50.05 when compared
with young (6 month old) rats. Reproduced with permission from Wang et al (1999).
REPRODUCTIVE AGEING IN THE MALE BROWN NORWAY RAT 85

with ageing did not occur (Chen & Zirkin 1999). These gradual changes in serum
testosterone levels in the BN rat mimicked closely the ‘andropause’ associated with
ageing in men. Low serum testosterone levels are the cause of the low muscle mass,
decreased bone density, frailty and might be partially responsible for the erectile
dysfunction commonly encountered in older men as described in the chapter by
Handelsman (2002, this volume).
We, as well as others (Wang et al 1993, 1999, Wright et al 1993), have shown that
testicular weights were reduced in 21^22 month old BN rats to about 60^70% of
those of young rats. Frequently a di¡erential decrease in testis size was observed in
the same animal. One testis was often much smaller (‘regressed’) and the other was
relatively normal in weight. By 30 months the mean testicular weights were very
small: only about 25^30% of those at 6 months. Mean testicular sperm
concentration and total sperm content decreased progressively from 15 months
(Fig. 2). Testicular histology in the relatively normal looking testis of 21^22
month old BN rats showed active spermatogenesis with large lumina and
marked variation in the appearance of the germinal epithelium from relatively
normal to a £attened epithelium lined by a single layer of Sertoli cells and a few
spermatogonia. Many tubules had areas consisting of normal-looking tubules
intermingled with groups of damaged tubules. In the regressed testis, the
seminiferous tubule showed complete cessation of spermatogenesis. By 30
months, the testes were very small and showed features similar to those of the
regressed testes of the 21^22 month old rats. Testicular stereologic analysis
showed a marked reduction in the volumes of seminiferous tubules and tubular
diameter. In old animals, Leydig cell volume was markedly lower but their
number remained unchanged. There was a marked decrease in the number of
Sertoli cell per testis in the regressed testes of 21^22 and 30 month old rats
(Wang et al 1993, Wright et al 1993). We also showed that the decrease in germ
cells was due to accelerated germ cell apoptosis (Fig. 3) involving spermatogonia,
spermatocytes and spermatids (Wang et al 1999). In addition, there was marked
variability in the apoptosis rate in di¡erent tubules. Germ cell apoptosis was most
evidenced in stages XII^XIV when compared with young animals. It should be
noted that in the regressed testis, the seminiferous tubules were lined with Sertoli
cells and few apoptotic germ cells as well as Leydig and Sertoli cells (Wang et al
1999). The variable and irregular depletion of germ cells by apoptosis in the
seminiferous tubules suggests that in addition to low testosterone levels, other
factors such as decreased blood supply and increased cytotoxic agents (e.g.
reactive oxygen species) might also play a role in germ cell degeneration. The
changes in the seminiferous tubules in the BN rats are more pronounced than
those reported in men. Semen analyses of healthy active grandfathers have been
reported to be relatively normal (Nieschlag et al 1982), but others reported that the
daily sperm production was decreased in ageing men (Neaves et al 1984).
86 WANG ET AL

FIG. 2. Body weight (A), testicular parenchymal weight (B), sperm content (C), and sperm
concentration (D) in ageing BN rats. Data represent mean SE of 5^10 animals per group.
*Indicates signi¢cant di¡erence, P50.05, when compared with young (6 month old) rats.
Reproduced with permission from Wang et al (1999).

When rat LH was administered as a twice daily injection to 15 month old BN rats
for 6 months, plasma hormone levels, testis weight, sperm concentration and
content, and germ cell apoptosis rate remained the same as those of the untreated,
control group. In the control group, three out of 10 testes were regressed whereas
in the LH-treated group only one out of 12 testes was regressed. These results
REPRODUCTIVE AGEING IN THE MALE BROWN NORWAY RAT 87

FIG. 3. Light micrographs of testicular sections from a 21 month old rat showing apoptotic
germ cells (arrow) in the relatively normal looking (A) and regressed (B) testes. Visualization of
apoptotic germ cells was by terminal deoxynucleotidyl transferase-mediated dUTP nick end
labelling (TUNEL). Methyl green was used as a counterstain. Magni¢cation

suggest that impaired hypothalamic^pituitary function may not be the only cause
of testicular germ cell loss associated with ageing (Wang et al 1999). However the
experimental conditions cannot exclude the possibility that the LH treatment
might be inadequate because it was given as twice daily doses and not in a
pulsatile fashion, and the testosterone response to the dose of LH administered
might be insu⁄cient to support spermatogenesis.
88 WANG ET AL

Hypothalamic^pituitary dysfunction

In addition to a primary testicular failure, the aged BN rat also showed features
suggestive of a hypothalamic^pituitary dysfunction. Serum FSH levels were
elevated in association with low serum inhibin and decreased spermatogenesis in
older rats (Wang et al 1993, Gruenewald et al 1994). In contrast, despite the low
serum testosterone levels, serum LH showed no change (Chen et al 1994,
Gruenewald et al 1994) or a decrease in aged rats (Wang et al 1993, 1999) (Fig. 1).
Moreover, the blunted rise in serum LH and FSH after castration in the old, when
compared with the young animals provided further evidence for a hypofunctional
hypothalamic^pituitary axis (Gruenewald et al 1994). We also showed that the
pulsatile secretion of LH was characterized by a shortened pulse interval and
reduced areas of the pulses in the old rats. LH pulse amplitude and total area of
the LH pulses were also signi¢cantly lower in old than in young rats. In contrast,
mean serum FSH levels in old rats were signi¢cantly higher than those observed in
young rats. Mean areas but not amplitude of FSH pulses decreased signi¢cantly in
the old rats. FSH pulse frequency increased and pulse interval decreased in the old
rats (Fig. 4). Our results were corroborated by similar studies by Gruenewald et al
(2000) who also demonstrated blunted circadian rhymicity of LH and testosterone
secretion due to decreased gonadotropin releasing hormone (GnRH) production
rather than decreased pituitary responsiveness to GnRH. These changes in the
pulsatile secretion of the gonadotropins in the BN rat are similar to the human
and are consistent with a hypothalamic GnRH pulse generator dysfunction. In
man, high amplitude LH peaks tend to fall and frequency of low amplitude LH
peaks tend to rise in older men (Veldhuis et al 1992, Mulligan et al 1995). The
decreased LH secretory burst amplitude correlated with serum free testosterone
in elderly men. In contrast, serum FSH showed an increase in basal secretion rate
and increased FSH secretory burst mass and amplitude in old compared to young
men (Veldhuis et al 1999). The neuroendocrine mechanisms underlying the
discrepant LH and FSH pulsatile secretions occurring with ageing are not known.
It has been previously shown that excitatory amino acids acting through their
receptors have a primary role in the regulation of the pulsatile secretion of GnRH
and LH. GnRH peptide content, synthetic capacity and gene expression are
reduced in aged male rats (Gruenewald & Matsumoto 1991, Gruenewald et al
2000). We hypothesize that decreased GnRH responsiveness to excitatory amino
acids may occur with ageing. To test this hypothesis, we ¢rst demonstrated that in
vivo serum FSH or LH responsiveness to GnRH was not altered in the old rats. We
then showed that administration of an excitatory amino acid (glutamate) receptor
agonist N-methyl-D-aspartate (NMDA) induced higher LH and prolactin releases
in young versus old animals. Moreover, using hypothalamus fragments, we
showed that the in vitro GnRH e¥ux in response to NMDA was lower in old rats
FIG. 4. Plasma luteinizing hormone (LH), follicle-stimulating hormone (FSH), and testosterone (T) levels over the 8 h sampling period of one
representative male rat of each age group. *Denotes pulses identi¢ed by cluster analysis.
90 WANG ET AL

compared to young rats. In the hypothalamus from old rats there was signi¢cant
reduction in the content of glutamine and g-aminobutyric acid (Bonavera et al
1998). Taken together, these results showed that the NMDA^GnRH^LH axis
was altered in old rats, and the decreased hypothalamic content of some of the
excitatory amino acids and reduced responsiveness of GnRH neurons to NMDA,
both in vivo and in vitro, may play a role in the altered LH pulsatile secretion
observed in older rats.

Nitric oxide synthase and reproductive ageing

The excitatory amino acid receptors (NMDA receptors) are considered to be the
main neurotransmitter receptors mediating fast synaptic excitation in the central
nervous system (CNS) through the opening of Ca2+ channels triggering a series
of neuronal cascades. One of the mechanistic pathways is the stimulation of
neuronal nitric oxide synthase (nNOS) resulting in the synthesis of nitric oxide
which has been presumed to be the mediator of some of the neuroendocrine
e¡ects of the excitatory amino acid glutamine speci¢cally on GnRH and LH
pulsatile secretions (Braun et al 1997). In the hypothalamus, nNOS is the
predominant NOS isoform and has been demonstrated to colocalize with or
adjacent to the NMDA receptors and GnRH neurons in the medial preoptic area
(Grossman et al 1994). Thus in ageing, it is conceivable that the hypothalamic
NMDA receptor/nNOS pathway may play a role in hypothalamic dysfunction.
NMDA receptors might be decreased accompanied by a decrease in nNOS
activity in the brain and other organs outside the CNS.
An alternative hypothesis is that ageing is associated with an excessive synthesis
of NO resulting in the accumulation of its cytotoxic metabolites such as
peroxynitrite, leading to neuronal apoptosis. The process might a¡ect the
hypothalamic neurons including those that secrete GnRH. Such cytotoxic e¡ects
of excessive levels of NO could result in accelerated apoptosis of other components
of the reproductive axis such as the testes. The high NO levels in tissues of ageing
animals may occur as a result of excessive stimulation of nNOS by activation of
NMDA receptors or the spontaneous expression of inducible NOS (iNOS), the
NOS isoform that is induced during autoimmunity, in£ammation and
degeneration. In normal physiological conditions iNOS is undetectable in the
organs of adult laboratory animals and is expressed in high levels only after
exogenous cytokine stimulation and in in£ammatory or infectious processes. To
study the role of NOS in reproductive ageing, our laboratories investigated the
NMDA receptor content and binding, nNOS and iNOS levels, and activity in
the brain and in the testes of the ageing BN rat.
The NMDA receptor binding activity in the hypothalamus of old rats was 66%
lower than that of adult animals. The results of the binding activity were con¢rmed
REPRODUCTIVE AGEING IN THE MALE BROWN NORWAY RAT 91

by the demonstration that the NMDA receptor content was also decreased by 34%
compared to young animals. To our surprise, the decrease in NMDA receptors was
associated with a 67% increase in NOS activity in the hypothalamus of old rats
when compared to the adult animals, while nNOS content was not di¡erent
between the two groups. In contrast, iNOS content in the hypothalamus of old
rats were increased by 3.8-fold compared with adult animals (Fig. 5). The
increase in iNOS content was demonstrated not only in the hypothalamus but
also in the frontal and parietal cortex, and in the cerebellum (Vernet et al 1998).
Our results showed that ageing in the BN rat was associated with high NO
synthesis in the hypothalamus and other regions of the brain. This occurred
independently of the NMDA receptors (which were decreased) and nNOS
activity (which was unchanged). We thus concluded that increased iNOS might
result in neurotoxicity which could be involved in the impaired GnRH pulsatile
secretion and also a possible inducer of age-associated cell loss in the brain and
other organs such as the testes.
We then proceeded to investigate the role of iNOS in male reproductive ageing
in the hypothalamus and the end organ, the testis. Using immunohistochemistry
we found signi¢cant increases in iNOS immunostaining in the supraoptic and
paraventricular nucleus and the preoptic area of the hypothalamus in the old
rats. Nitrotyrosine, a marker for peroxynitrite formation (a cytotoxic product of
excess NO and reactive oxygen species interaction) was also elevated in the same
areas of the hypothalamus of old rats. The accumulation of peroxynitrite was
accompanied by an increase in the apoptotic index of neurons in the supraoptic,
paraventricular and arcuate nucleii as well as in the preoptic area of the hypo-
thalamus of old rats. Apoptosis of neurons is extremely rare in the hypothalamus
and other areas of the brain of young animals. We thus hypothesized that neuronal
apoptosis could be the cause of the reduction of GnRH neurons detected by in situ
hybridization by Gruenewald et al (2000). In contrast, ageing did not a¡ect nNOS
expression. We further examined the anatomical relationship of iNOS and GnRH-
positive neurons in the hypothalamus using double immuno£uorescence
technique in combination with confocal laser scanning microscopy. The iNOS
staining co-localized with GnRH staining in the same regions of the
hypothalamus of the rat brain. These preliminary studies showed iNOS
expression in the hypothalamus of the a¡ected regions of the brain known to
control the synthesis and release of GnRH, con¢rming our hypothesis that
iNOS may indeed play a role in the reduction of GnRH pulsatile secretion
resulting in reproductive dysfunctions such as lowered testosterone in the ageing
males. Ongoing studies aim to demonstrate iNOS co-localization with apoptotic
cells in the hypothalamus of old rats.
We also demonstrated that similar changes in NOS activity occurred in the testes
of old BN rats. In the regressed testes of old animals NOS activity was increased
92 WANG ET AL

FIG. 5. E¡ect of ageing on iNOS levels in the rat hypothalamus. Top panel, autoradiography
of the 130 kDa bands on a typical Western blot of the postmitochondrial supernatants (80 mg
protein/lane) with an antibody against mouse iNOS and visualization with a luminol reaction.
Middle panel, Mean intensity of the respective bands determined by densitometry

compared with younger adults. The expression of iNOS assessed by Western blot
assay was present in 3 month old animals but increased by 2.5-fold in the relatively
normal testes and fourfold in the regressed testes of the old rats. No signi¢cant
changes were noted in nNOS levels between the testes obtained from young or
REPRODUCTIVE AGEING IN THE MALE BROWN NORWAY RAT 93

old animals. Immunohistochemistry showed that iNOS expression was most

striking in Leydig cells with little or no immunoreactivity in the Sertoli and germ
cells of young rats. In contrast, there was a marked increase in iNOS
immunoreactivity in the Sertoli cells and some staining in the germ cells of old
rats. Neuronal NOS was detectable in Leydig and Sertoli cells in both young and
old animals but unlike iNOS, no age-related di¡erences were apparent. Our results
showed that ageing in the BN rats, in a similar fashion to changes in the
hypothalamus and other brain regions, was associated with iNOS induction and
high iNOS synthesis in the testis. We also speculate that the increased iNOS
through the formation of cytotoxic products may be an important mediator of
age-related activation of germ cell apoptosis.

Summary
In this chapter we have described the changes in the male reproductive axis with
ageing in the BN rat as a model for human reproductive dysfunction. We have
shown that the reproductive axis in the rat sustained dual hits at the testis and the
hypothalamus. We showed that these hits caused an accelerated neuronal and germ
cell apoptosis presumably as a result of oxidative damage by excessive
accumulation of the inducible NOS. The dual dysfunction at both the testicular
and hypothalamic regions possibly resulted in impaired Leydig cell function and
decreased spermatogenesis characteristic of male reproductive ageing in men.

References
Baker HWG, Burger HG, de Kretser DM et al 1976 Changes in the pituitary-testicular system
with age. Clin Endocrinol 5:349^372
Bonavera JJ, Swerdlo¡ RS, Sinha-Hikim AP, Lue YH, Wang C 1998 Aging results in attenuated
gonadotropin releasing hormone^luteinizing hormone axis responsiveness to glutamate
receptor agonist N-Methyl-D-aspartate. J Neuroendocrinol 10:93^99
Braun DW, Bhat GK, Lamar CA, Mahesh VB 1997 Gaserous transmitters and neuroendocrine
regulation. Neuroendocrinology 65:385^395
Chen H, Zirkin BR 1999 Long-term suppression of Leydig cell steroidogenesis prevents Leydig
cell aging. Proc Natl Acad Sci USA 96:14877^14881
Chen H, Hardy MP, Huhtaniemi I, Zirkin BR 1994 Age-related decreased Leydig cell
testosterone production in the Brown-Norway rats. J Androl 15:551^557
Chen H, Huhtaniemi I, Zirkin BR 1996 Depletion and repopulation of Leydig cells in the testes
of aging Brown Norway rats. Endocrinology 137:3447^3452
Gray A, Berlin JA, McKinlay JB, Longcope C 1991 An examination of research design e¡ects on
the association of testosterone and male aging: results of a meta-analysis. J Clin Epidemiol
44:671^684
Grossman AB, Rosmanith WG, Kabigtin EB, Cadd G, Clifton D, Steiner RA 1994 The
distribution of hypothalamic nitric oxide synthase mRNA in relation to gonadotropin-
releasing hormone neurons. J Endocrinol 140:R5^R8
94 WANG ET AL

Gruenewald DA, Matsumoto AM 1991 Age-related decreases in serum gonadotropin levels and
gonadotropin-releasing hormone gene expression in the medial preoptic area of the male rat
are dependent upon testicular feedback. Endocrinology 129:2442^2450
Gruenewald DA, Naai MA, Hess DL, Matsumoto AM 1994 The Brown Norway rat as a model
of male reproductive aging: evidence for both primary and secondary testicular failure. J
Gerontol 49: B42^B50
Gruenewald DA, Naai A, Marck BT, Matsumoto AM 2000 Age-related decrease in
hypothalamic gonadotropin-releasing hormone (GnRH) gene expression but not pituitary
responsiveness to GnRH, in the male Brown Norway rat. J Androl 21:72^84
Handelsman DJ 2002 Male reproductive ageing: human fertility, androgens and hormone
dependent disease. In: Endocrine facets of ageing. Wiley, Chichester (Novartis Found Symp
242) p 66^81
Harman SM, Tsitouras PD 1980 Reproductive hormones in aging men. I. Measurement of sex
steroids, basal luteinizing hormone and Leydig cell response to human chorionic
gonadotropin. J Clin Endocrinol Metab 51:35^40
Korenman SG, Morley JE, Mooradian AD et al 1990 Secondary hypogonadism in old men: its
relation to impotence. J Clin Endocrinol Metab 71:963^969
Luo L, Chen H, Zirkin BR 1996 Are Leydig cell steroidogenic enzymes di¡erentially regulated
with aging? J Androl:509^515
Mulligan T, Iranmanesh A, Gheorghiu S, Godschalk M, Veldhuis JD 1995 Ampli¢ed nocturnal
lutienizing hormone (LH) secretory burst frequency with selective attenuation of pulsatile
(but not basal) testosterone secretion in healthy aged men: possible Leydig cell
desensitization to endogenous LH signalling
a clinical research center study. J Clin
Endocrinol Metab 80:3025^3031
Neaves WB, Johnson L, Porter JC, Parker CR, Petty CS 1984 Leydig cell numbers, daily sperm
production and serum gonadotropin levels in aging men. J Clin Endocrinol Metab 59:756^
763
Nieschlag E, Lammers U, Freischem CW, Langer K, Wickings EJ 1982 Reproductive functions
in young fathers and grandfathers. J Clin Endocrinol Metab 55:675^681
Swerdlo¡ RS, Wang C 1993 Androgens and aging in men. Exp Gerontol 28:435^446
Veldhuis JD, Urban RJ, Lizarralde G, Johnson ML, Iranmanesh A 1992 Attenuation of
luteinizing hormone secretory burst amplitude is a proximate basis for the
hypoandrogenism of healthy aging in men. J Clin Endocrinol Metab 75:707^713
Veldhuis JD, Iranmanesh A, Demers LM, Mulligan T 1999 Joint basal and pulsatile
hypersecretory mechanisms drive the monotropic follicle-stimulating hormone (FSH)
elevation in healthy older men: concurrent preservation of the orderliness of the FSH
release process: a general clinical research center study. J Clin Endocrinol Metab 84:3506^3514
Vernet D, Bonavera JJ, Swerdlo¡ RS, Gonzalez-Cadavid NF, Wang C 1998 Spontaneous
expression of inducible nitric oxide synthase in the hypothalamus and other brain regions of
aging rats. Endocrinology 139:3254^3261
Vermeulen A 1991 Androgens in the aging male. J Clin Endocrinol Metab 73:221^224
Wang C, Leung A, Sinha-Hikim AP 1993 Reproductive aging in the male Brown-Norway rat: a
model for the human. Endocrinology 133:2773^2781
Wang C, Sinha-Hikim AP, Lue YH, Leung A, Baravarian S, Swerdlo¡ RS 1999 Reproductive
aging in the Brown Norway rat is characterized by accelerated germ cell apoptosis and is not
altered by luteinizing hormone replacement. J Androl 20:509^518
Wright WW, Fiore C, Zirkin BR 1993 The e¡ect of aging on the seminiferous epithelium of the
Brown-Norway rat. J Androl 14:110^117
Zirkin BR, Santulli R, Strandberg JD, Wright WW, Ewing LI 1993 Testicular steroidogenesis
in the aging Brown-Norway rat. J Androl 14:118^123
REPRODUCTIVE AGEING IN THE MALE BROWN NORWAY RAT 95

Zirkin BR, Chen H 2000 Regulation of Leydig cell steroidogenic function during aging. Biol
Reprod 63:977^981

DISCUSSION
Handelsman: Can I draw you out on one aspect you didn’t mention, which is the
opioid control of GnRH neurons. As far as I understand it, what you are working
on is within the GnRH neuron itself. Is there any relationship between opioid
e¡ects and iNOS?
Wang: We haven’t studied this. Opiates exert negative e¡ects on GnRH
hormones. We don’t know whether the iNOS is important as the mechanism of
action, but this is what we are pursuing. We are looking at both the brain and the
testis.
Mˇller: You have shown an increase of iNOS in both the testis and the brain. In
the brain iNOS increases in many areas beside the hypothalamus. Can you speculate
on this ¢nding a little bit? Could it be that NOS increase is not only a marker of
decreased fertility but also a general phenomenon? We also have some indirect data
in this respect as far as growth hormone (Rigamonti et al 1999), and sexual function
(Melis et al 2001) are concerned.
Wang: There are data even in the human brain showing that iNOS activity may
be increased in older subjects, but these data are not as solid as the rat data. The
NMDA receptor has also been implicated in a stress-related increase in cortisol,
causing damage to the brain, especially in rat models.
Bj˛rntorp: I am going to ask a terribly ignorant question. How do you imagine
that the NOS is acting here?
Wang: We think that ageing causes decreasing blood £ow and increased ROS
generation, including iNOS. I believe there is also an increase in cytokines in the
brain. The peroxynitrite products of iNOS are cytotoxic and kill the neurons that
secrete GnRH and oxytocin, which causes some of the changes we observed. We
also have evidence that the iNOS is colocalized in cells undergoing apoptosis.
Laron: You said that this is the best model for ageing studies in humans. We
learned earlier about the characteristics of ageing in humans. What are the
characteristics of ageing in the rat?
Wang: I think osteoporosis is present. I have no idea about muscle. The only
reason we think this is a good model is that the testosterone is low, caused by
both testicular and hypothalamic^pituitary dysfunction.
Morley: You are looking at 50 year olds. These could very well be human data,
but you would have to go to 27^28 month old rats to be looking at the equivalent
of ‘old’ subjects. The use of ‘old’ is a misnomer in this situation.
Wang: Initially we did experiments on 30 month old rats, but in these rats the
testis is so small we cannot do many experiments. The supply of very old rats can be
96 DISCUSSION

limited and 50% of the rats die by the age of 30 months. With the iNOS knockout
mice we wish to study them when they are very young and then sacri¢ce them at
di¡erent ages until old age.
Morley: You also didn’t look at 12 month old rats. In our paper we found that
some of those increases were present at 12 months (Morley et al 1996). It really is a
maturational change with the iNOS. We have found a reduction of NOS mRNA in
older animals.
Wang: We have looked at 3, 6, 9, 12 and 18 month old animals, because we want
to characterize the changes so that we can do interventions. This is why we need to
know when they start to change with ageing.
Handelsman: A point of clari¢cation. You mentioned transgenics in your paper.
Are you planning to do knockouts in the mouse? If so, is the mouse the same as the
Brown Norway rat?
Wang: We plan to use iNOS knockout mice. We have studied both the testis and
the brain of the mouse, we can see an increase in iNOS with ageing. We have not
done the pulsatile LH secretion; this is very di⁄cult in the mouse.
Brabant: With the maturational change of iNOS decreasing in the old
animals, it would be interesting to know what happens if you castrate young
animals.
Wang: We haven’t done castration experiments, although we really should. Al
Matsumotos’s group did castration experiments in the young and old rats, and
compared the GnRH content in hypothalamic neurons (Gruenewald &
Matsumoto 1991, Gruenewald et al 1994, 2000). They showed that GnRH
mRNA and content are decreased in castrated ageing BN rats. They did not
measure iNOS. We also haven’t done testosterone replacement experiments in
these rats.
Veldhuis: There may be a slightly di¡erent question here: in the female rat,
dozens of reports have shown that high dose oestradiol is neurotoxic to certain
of these centres, while at the same time being protective against brain ischaemia
and hypoxia. Georg Brabant is raising an interesting point: could you prevent
ageing of the GnRH neurons by castration?
Wang: Castration has not been shown to be protective in the hypothalamus.
There was a recent paper in which researchers gave contraceptive doses of
testosterone into the old BN rat, which caused the suppression of testosterone
and spermatogenesis (Chen & Zirkin 1999). The treatment caused the Leydig
cells to go into what they called ‘hibernation’. If this was done for 6 months and
then the treatment was stopped, the testosterone production in the young and old
rats became similar.
Veldhuis: What makes you say that the upstream activators of this iNOS are
probably cytokines? What evidence do you have that cytokines are increased in
the ageing male rat in these areas?
REPRODUCTIVE AGEING IN THE MALE BROWN NORWAY RAT 97

Wang: There is scattered information about increased cytokine levels, both in the
rat brain and the human brain. iNOS is traditionally induced by in£ammatory
responses or cytokines.
Veldhuis: So the neuronal death could be a glial event?
Wang: Yes.
Laron: What do you know about the in£uence of insulin-like growth factor
(IGF)1 on the ageing brain? A few months ago I went through the data on IGF1
in the brain, and it seems to be anti-apoptotic. You showed increased apoptosis.
Wang: I don’t know.
Veldhuis: Is this the same model in which a single group has reported that fetal
neuronal transplantation has restored potency in the old male rat? This raises the
possibility that there is structural loss of GnRH neurons, and it is not just
functional. I didn’t believe that this was the case. I had always preferred the
hypothesis that there was functional loss of input signals to the GnRH ensemble,
or loss of coordinate secretion. Your data are suggesting that there is increased cell
death.
Prior: Tomorrow we will be talking about the perimenopause in women. To use
similar language and call it the ‘andropause’ is to make light of the huge number of
changes that occur over 5^10 years in women, versus the very gradual, slow
changes that occur in men. I wish we could take away that word and not use it.
Perhaps the ‘andropause’ applies to the Norway rat, but I don’t think it applies to
human males.

References
Chen H, Zirkin BR 1999 Long-term suppression of Leydig cell steroidogenesis prevents Leydig
cell aging. Proc Natl Acad Sci USA 96:14877^14881
Gruenewald DA, Matsumoto AM 1991 Age-related decreases in serum gonadotropin levels and
gonadotropin-releasing hormone gene expression in the medial preoptic area of the male rat
are dependent upon testicular feedback. Endocrinology 129:2442^2450
Gruenewald DA, Naai MA, Hess DL, Matsumoto AM 1994 The Brown Norway rat as a model
of male reproductive aging: evidence for both primary and secondary testicular failure. J
Gerontol 49: B42^B50
Gruenewald DA, Naai A, Marck BT, Matsumoto AM 2000 Age-related decrease in
hypothalamic gonadotropin-releasing hormone (GnRH) gene expression but not pituitary
responsiveness to GnRH, in the male Brown Norway rat. J Androl 21:72^84
Melis MR, Succu S, Spano MS et al 2001 Penile erection induced by EP 80661 and other
hexarelin peptide analogues: involvement of paraventricular nitric oxide. Eur J Pharmacol
411:305^310
Morley JE, Kumar VB, Mattammal MB, Farr S, Morley PM, Flood JF 1996 Inhibition of
feeding by a nitric oxide synthase inhibitor: e¡ects of aging. Eur J Pharmacol 311:15^19
Rigamonti AE, Cella SG, Marazzi N, Mˇller EE 1999 Nitric oxide modulation of the growth
hormone-releasing activity of hexarelin in young and old dogs. Metabolism 48:176^182
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Mechanisms of conjoint failure of the

somatotropic and gonadal axes in
ageing men
J. D. Veldhuis*, A. Iranmanesh{, T. Mulligan{ and C. Y. Bowers}

*Division of Endocrinology, University of Virginia School of Medicine, Charlottesville,

Virginia 22908-0202, {Endocrine Section, Salem Veterans A¡airs Medical Center, Salem,
Virginia 24153, {Geriatrics and Extended Care Service, McGuire Veterans A¡airs Medical
Center, Richmond, Virginia 23249, and }Department of Medicine SL53, Tulane University
Medical Center, New Orleans, Louisiana 70112-2699, USA

Abstract. Endogenous growth hormone (GH) production falls by 50% every 7 years and
bioavailable testosterone concentrations decline concomitantly by 12^15% every decade
in ageing men. Despite this temporal parallelism, the neuroendocrine bases of the
somatopause and gonadopause are not known. This knowledge de¢cit contrasts with
the recent unfolding of new insights into the nature of oestrogen-dependent control of
the GH^insulin-like growth factor (IGF)1 axis in pre- and postmenopausal women. The
present overview examines the postulate that the pathophysiology of somatopause
and gonadopause in ageing men is bidirectionally linked. According to this broader
thesis, hyposomatotropism accentuates Leydig cell steroidogenic failure and,
conversely, progressive androgen de¢ciency exacerbates the decline in GH^IGF1
output in ageing.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 98^124

Clinical features of ageing include variably impaired psychological well being,

cognitive function and quality of life; decreased libido and/or sexual function;
reduced physical productivity; diminished muscle and bone mass; and increased
visceral obesity, dyslipidemia and risk of cardiovascular events. These attributes
collectively heighten the potential for frailty, disability, su¡ering and loss of
independent living status. From an endocrine perspective, ageing in the male is
accompanied by a progressive and dual decline in the bioavailability of growth
hormone/insulin-like growth factor (IGF)1 and testosterone, which otherwise
jointly support tissue anabolism (Iranmanesh et al 1994, Veldhuis et al 1995;
Fig. 1). The terms somatopause and gonadopause highlight the foregoing age-
related impoverishment of GH/IGF1 and androgen output, respectively.
98
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 99

FIG. 1. (Left) Relationship between overnight serum bioavailable testosterone and GH

concentrations in a group of 10 young and older healthy men (Mulligan et al 1999a, Veldhuis
et al 1999b, 2000b). (Right) Correlations between serum total testosterone and 24 h GH and IGF1
output in a cohort of 46 healthy pubertal boys (Martha et al 1992).

However, virtually nothing is known about the basic clinical pathophysiology

underlying the interlinked attrition of these pivotal trophic^hormone axes in
ageing.
The present overview develops the bipartite thesis that ageing-associated
hyposomatotropism and hypogonadism arise from conjoint pathophysiologies.
We hypothesize that failing GH/IGF1 output worsens the age-related decline in
luteinizing hormone (LH)-stimulated testicular steroidogenesis, and, conversely,
that waning androgen availability blunts hypothalamo^pituitary drive of the GH/
IGF1 axis. This notion re£ects extensive basic science and clinical data (examined
below), which collectively indicate that GH and androgen show anabolic synergy
and mechanistic coupling in normal physiology.

Physiological synergy between the actions of GH and testosterone

GH and testosterone stimulate skeletal and muscular growth synergistically in late
puberty (Giustina & Veldhuis 1998, Keenan et al 1993, Klindt et al 1990,
Zachmann 1992). Recent studies also document target-tissue synergy between
GH and androgen in healthy older men. Thus, restoring both GH and
testosterone output in the older male by nearly physiological means would be
important. However, the precise mechanistic basis for joint failure of GH and
testosterone secretion in ageing is not known. To this end, a better
comprehension of their interactive neuroendocrine control will be required.
100 VELDHUIS ET AL

Basic linkages between the somatotropic and gonadal axes

Extensive clinical data a⁄rm that testosterone bioavailability and GH secretion
rise in parallel in puberty and fall concurrently in ageing (Fig. 1; Copinschi &
Van Cauter 1994, Dudl et al 1973, Giustina & Veldhuis 1998, Corpas et al 1992,
Fryburg et al 1997, Hartman et al 1991, Iranmanesh et al 1991, 1994, Martha et al
1992, Mulligan et al 1999a, Veldhuis et al 1991, 1995). Causality is inferable since, at
least in androgen-de¢cient individuals, administration of testosterone e¡ectually
drives the secretion of both GH and IGF1, and augments local tissue-speci¢c
IGF1 activity (Fryburg et al 1997, Gentili et al 2000, Snyder et al 1999, Urban et
al 1995, Wennink et al 1990). Conversely, GH and IGF1 amplify the biosynthesis
and tissue actions of androgens (Balducci et al 1993, Carani et al 1999, Giustina &
Veldhuis 1998, Kulin et al 1981, Lin et al 1986). Thus, testosterone positively
modulates the secretion and activity of the GH/IGF1 system (Fryburg et al 1997,
Gentili et al 2000, Veldhuis et al 1997), whereas GH and IGF1 impact the
production and e¡ects of testosterone. De¢cits in one or more of the foregoing
interactions may contribute to the pathophysiology of combined GH and
androgen de¢ciency in ageing, systemic illness and debilitated states.
Our recent clinical studies document that testosterone retains full e⁄cacy in
driving threefold ampli¢cation of pulsatile, 24 h rhythmic and entropic GH
secretion in older men, as otherwise observed in boys and young men (Fryburg
et al 1997, Gentili et al 2000, Shah et al 1999a, Veldhuis et al 1997, 2000a).
However, the exact neuroendocrine mechanisms that mediate these stimulatory
actions of testosterone are not established at any age. The following observations
underscore possible clues to and the implications of elucidating such basic
mechanistic pathways.

GH and IGF1 versus co-gonadotropins

GH and IGF1 can be considered as ‘co-gonadotropins’, in view of their ability to
facilitate LH and/or follicle-stimulating hormone (FSH)-stimulated gonadal
steroidogenesis in the rat and human (Lin et al 1986, Veldhuis 1996). Clinical
studies corroborate a co-gonadotropic action of GH in ovulation induction in
hypopituitary, but not eusomatotropic women. Analogously, hypopituitary, but
not GH-su⁄cient, boys and men respond to GH repletion with enhanced human
chorionic gonadotropin (hCG)-stimulated testosterone secretion (Balducci et al
1993, Carani et al 1999, Kulin et al 1981). Thus, maximal hCG/LH-driven
gonadal steroidogenesis appears to require a GH-su⁄cient state. This emergent
notion has never been tested in hyposomatotropic ageing men, who are in fact
hypopituitary by young-adult standards (Giustina & Veldhuis 1998, Martha et al
1992, Veldhuis et al 1995, 1997, 1999a). To address this issue, we have developed a
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 101

FIG. 2. Illustrative serum LH and testosterone concentration pro¢les obtained in one young
and one older man pretreated with leuprolide 3^4 weeks earlier to down-regulate endogenous
gonadotropin secretion. Data re£ect 10 min blood sampling and eight consecutive intravenous
pulses of recombinant human LH (50 IU, Serono) (Mulligan et al 2000).

novel experimental regimen of physiologically i.v. pulsatile GH ‘addback’ for 2

weeks to test the GH dependence of hCG-stimulated Leydig-cell steroidogenesis
in older men. This mechanistic consideration will be important to address
de¢nitively (Blackman 1987, degli Uberti et al 1997, Giustina & Veldhuis 1998,
Haji et al 1980, Mulligan et al 1999a, Shah et al 1999b, Veldhuis 1996).
Our pilot studies show that a pulsatile i.v. infusion of recombinant human (rh)
GH every 90 min for 14 days (0.33 mg/kg/pulse) will restore 24 h serum GH
concentration pro¢les, elevate plasma IGF1 concentrations and heighten 4 h
mean Leydig-cell testosterone secretion stimulated by a single i.m. injection of
2000 IU hCG. In preliminary experiments designed to eliminate the use of the
less physiological hCG stimulus, we have tested Leydig-cell responsiveness to
pulsatile i.v. infusions of rh LH in 15 leuprolide-down-regulated young and
older men (Mulligan et al 2000; Fig. 2). To improve study design further, we
102 VELDHUIS ET AL

have shown that a single s.c. dose (2 mg) of a gonadotropin-releasing hormone

(GnRH) antagonist, ganirelix, lowers the serum testosterone concentration
within 6 h by 485% for 16^24 h in young men (T. Mulligan, A. Iranmanesh &
J. D. Veldhuis, unpublished data). We earlier reported a comparably rapid time
course of D-Nal-Glu GnRH’s inhibitory actions in young men. Studies by
Tenover et al (1990), also showed that young and older men manifest equivalent
(age-independent) gonadal-axis suppression by a GnRH antagonist.
Documenting GH-dependent enhancement of gonadotropin-supported
Leydig-cell steroidogenesis in older men during controlled pulsatile LH drive
would directly link hyposomatotropism with impaired testicular steroidogenesis
in older men. To our knowledge, such a causal association would represent a novel
¢nding in any ageing mammalian species. Conversely, we posit that androgen
feedback on the GH axis enhances somatotropic activity. This thesis is discussed
next from a general and then mechanistic perspective.

Overview of the ensemble GH/IGF1 axis

The GH/IGF1 axis comprises a feedback ensemble controlled jointly by: (i) GH-
releasing hormone (GHRH) feedforward; (ii) somatostatin (SS) inhibition; (iii) a
GH-releasing oligopeptide (GHRP) signalling pathway; and (iv) GH and IGF1
autonegative feedback (Mueller et al 1999, Argente et al 1991, Baumann &
Maheshwari 1997, Baumbach et al 1998, Bowers et al 1990, Bowers 1993,
Carlsson et al 1990, Frohman & Jansson 1986, Giustina & Veldhuis 1998, Ja¡e
et al 1993, Kojima et al 1999, Smith et al 1997, Ho£and & Lamberts 1996). This
simpli¢ed core construct is illustrated in Fig. 3.
GHRH is an established primary agonist driving the biosynthesis and pulsatile
secretion of GH in all mammalian species (Giustina & Veldhuis 1998). Conversely,
somatostatin (SS) is a dominant inhibitory signal, which antagonizes the
exocytotic secretion of GH, but not its biosynthesis or storage (Ho£and &
Lamberts 1996). GHRPs are potent and selective oligopeptidyl GH
secretagogues (Bowers et al 1984), mimicked by certain non-peptidyl agonists
and exempli¢ed endogenously by a 3Ser-octanoylated 28-amino acid GHRP-like
ligand (Smith et al 1997, Howard et al 1996, Kojima et al 1999, Mueller et al 1999).
The foregoing trilogy of neuropeptidyl regulators controls GH secretion via
topographically and biochemically distinct receptors and secondary signalling
molecules (Barinaga et al 1985, Giustina & Veldhuis 1998, Mayo 1992, Mueller
et al 1999). These diverse, but uniquely interactive, features create: (a)
cooperative mechanisms of biological control, and (b) multiple vulnerable loci
for disrupted neuroregulation in ageing and/or hypogonadism. We believe that
such interactive properties further mandate the particularly careful design of
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 103

b 2-

FIG. 3. Working experimental schema of primary tri-peptidyl control of the human GH axis
(Giustina & Veldhuis 1998).

experimental interventions, e.g. wherein one ‘clamps’ two (of the three) input
signals in order to assess (sex-steroid induced) changes in the third.

Altered GH neuroregulation in ageing and/or hypogonadism

Precisely how ageing and/or androgen de¢ciency impacts the GH/IGF1
neuroregulatory unit is not known (Giustina & Veldhuis 1998). However, both
animal and human studies make an ensemble perspective essential. For example,
clinically, only L-arginine (an agent believed to withdraw SS) combined with (a)
GHRH, (b) GHRP or (c) both GHRH and GHRP can normalize GH secretion
acutely, when judged against similarly stimulated young adults (Ghigo et al
1990, Khorram et al 1997, Veldhuis & Giustina 2000). Speci¢cally, no single
GHRH or GHRP agonist, or SS antagonist, can restore GH/IGF1 output
completely in older volunteers (Veldhuis & Giustina 2000, degli Uberti et al
1997, Giustina & Veldhuis 1998). Notably, sustained exogenous GHRH
stimulation or GHRP2 drive only partially reconstitutes the GH/IGF1 axis in
elderly humans, thus pointing to a relative de¢ciency of each agonist (Corpas et
al 1992, Evans et al 2000, Iranmanesh et al 1998, Khorram et al 1997). However,
the GH-stimulating e¡ect of GHRH infused i.v. over 3 days is blunted by
increasing age and relative hypoandrogenaemia (Iranmanesh et al 1998; Fig. 4).
FIG. 4. (Left) Testosterone-dependent decline in 24 h GH secretion driven by pulsatile i.v. infusion of GHRH (0.33 mg/kg/pulse) every 90 min
for 3 days in 19 men (Iranmanesh et al 1998). (Right) Partial reconstitution of daily GH and IGF1 production by continuous s.c. GHRP2 infusion
for 30 days in 12 healthy older men.
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 105

Thus, we propose that multifold neuroregulatory failure underlies impoverished

GH secretion in ageing, namely, combined GHRH and/or GHRP de¢ciency and
SS excess. Several plausible mechanistic considerations render this thesis more
compelling.

GHRP and GHRH

GHRPs act at both hypothalamic and pituitary loci (Smith et al 1997). At CNS
sites, these agonists stimulate electrical ¢ring and c-fos gene expression in rodent
NPY and GHRH neurons, elicit GHRH secretion acutely into sheep portal blood,
and induce somnolence or alter appetite (Arvat et al 1998, Dickson et al 1995,
Giustina & Veldhuis 1998, Guillaume et al 1994, Iranmanesh et al 2000, Locke et
al 1995). At the pituitary level, GHRPs stimulate GH secretion directly (albeit less
powerfully) in vitro, and putatively act at joint hypothalamo^pituitary loci to
enhance somatotrope GH gene expression in the infant rat in vivo (Bowers et al
1984). Central (hypothalamic) actions of GHRPs are critical clinically, since
hypothalamo^pituitary interruption virtually abolishes GHRP-stimulated GH
secretion, even when somatotrope responsiveness to GHRH is preserved
(Giustina & Veldhuis 1998, Mueller et al 1999). Maximal e¡ects of GHRPs also
require a functional GHRH receptor, as inferred from genetic studies in mice (lit/
lit mutation) and humans (dwarfs of Sindh) (Baumann & Maheshwari 1997,
Frohman 1996), and based on the ability of a selective GHRH-receptor
antagonist to suppress GHRP-stimulated GH secretion in young men.
Most studies document prominent in vivo synergy between the secretagogue
e¡ects of GHRP and GHRH in the human, pig and rodent (Bowers 1998,
Giustina & Veldhuis 1998, Mueller et al 1999, Bowers et al 1990; Fig. 5). How
testosterone modulates the foregoing GHRH^GHRP synergy at any age remains
unknown.

Somatostatin and GHRPs

GHRPs act as so-called ‘functional SS antagonists’ by increasing the ID50 of SSs
inhibition of spontaneous and GHRH-stimulated GH secretion by two^threefold
in vitro and by 8^10-fold in vivo (Bowers 1998, Smith et al 1997). GHRPs partially
oppose central SSergic activity, but do not directly block SS binding to pituitary
cells or SS secretion into portal blood (Bowers 1998, Guillaume et al 1994).

Somatostatin, GHRH and GHRP triology

Tri-peptidyl interactions operate in the rat and human, since experimentally
reducing SSergic input will markedly amplify in vivo dual-agonist (GHRH/GHRP)
106 VELDHUIS ET AL

FIG. 5. Synergistic stimulation of GH secretion by bolus injection of GHRH (1 mg/kg) and

GHRP2 (1 mg/kg) in middle-aged men, which can be enhanced further by post-SS (octreotide)
rebound. See text for details.

synergy (Arvat et al 1998, Bowers 1998, Smith et al 1997). Indeed, co-infusion of L-

arginine (to withdraw SS) and GHRH or GHRP, or all three stimuli combined, will
evoke supra-additive GH secretion in older humans, which is equivalent to that
achieved in young adults (Arvat et al 1998, Bowers 1998). These data are
consistent with excessive SSergic restraint and combined GHRH/GHRP
de¢ciency in ageing (Bowers 1998, Giustina & Veldhuis 1998). In contrast, the
magnitude of bihormonal GHRH/GHRP synergy (without L-arginine) wanes
substantially in older humans. Accordingly, we propose to examine the
endogenous control of all three (SS, GHRH, and GRHP) signalling pathways to
better explicate the mechanisms of hyposomatotropism in the older male.

GH autonegative feedback
An abrupt increase in the GH concentration feeds back physiologically to limit
further secretion (Berman et al 1994, Chapman et al 1997, Clark et al 1988,
Giustina & Veldhuis 1998, Harel & Tannenbaum 1992, Rosenthal et al 1986).
This time-lagged and reversible autoregulatory action probably sustains normal
GH pulsatility (Frohman 1996). Autonegative feedback is mediated via GH’s
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 107

stimulation of hypothalamic SS release and reciprocal inhibition of GHRH

secretion, without any direct action on somatotropes (Frohman 1996, Smith et al
1997). The ability of testosterone to stimulate GH and IGF1 production
simultaneously would suggest that testosterone may act to override such
autoinhibition by GH (Gentili et al 2000, Giustina et al 1997, Giustina &
Veldhuis 1998, Veldhuis et al 1997).

Integrative issues
The foregoing interactive features of GH neuroregulation highlight the crucial
need to explore interactive mechanisms subserving impoverishment of GH/
IGF1 output in the ageing and relatively hypogonadal male. Below we evaluate
several such presumptive neuroregulatory mechanisms, which could plausibly
mediate testosterone’s failing drive of the GH/IGF1 axis in older men.

Hypothesis 1: testosterone deprivation accentuates

GH autonegative feedback in the ageing male
Autonegative feedback denotes the recognized ability of a GH pulse to inhibit
subsequent GH secretion, whether driven spontaneously or by an exogenous
GHRH or GHRP stimulus (Carlsson et al 1990, Chihara et al 1981, Clark et al
1988, Rosenthal et al 1986). GH autoinhibition occurs by way of central nervous
system (CNS) (rather than pituitary) pathways, which rapidly stimulate SS and
repress GHRH secretion (Chihara et al 1981, Pellegrini et al 1996). Autofeedback
is non-trivial, since molecular defects of the GH receptor (or IGF1) gene evoke
marked secondary GH hypersecretion (Frohman 1996). Conversely, GH
secretagogues that putatively oppose SS release, such as L-arginine, overcome
GH autofeedback. Thus, GH autonegative feedback in£uences the activity of
each primary neuroregulatory peptide (SS, GHRH and GHRP; Fig. 3).
We have shown that a single i.v. pulse of rhGH reproducibly inhibits
subsequent spontaneous (endogenous) GH secretion by 75^80%, as well as that
stimulated acutely by bolus GHRH (by 60^75%) and GHRP2 (by 35^50%). GH
feeds back less e¡ectively to suppress the GHRP than GHRH stimulus,
presumptively because GHRPs partially antagonize SS’s actions (Guillaume et al
1994, Bowers 1998). These data suggest the experimental question: does
testosterone depletion in older men limit GH and IGF1 output in part by
accentuating (hypothalamic SS-mediated) GH autonegative feedback? This issue
is additionally noteworthy in view of the recent ¢nding that brain GH receptor
density falls in older humans. Since CNS GH receptors mediate GH
autofeedback (Giustina & Veldhuis 1998), their attrition in ageing could forecast
a countervailing interpretation of waning GH autofeedback. However, oestrogen
108 VELDHUIS ET AL

elevates hypothalamic (and represses liver) GH receptor expression, at least in

the rat. Thus, relative oestrogenization in ageing men might actually accentuate
GH receptor-dependent autonegative feedback. Given these provocative and
divergent regulatory issues, we believe that GH autofeedback studies will be
important to pursue in androgen-de¢cient and androgen-replete older men.

Hypothesis 2: testosterone depletion reduces the

potency and/or e⁄cacy of GHRH’s actions
GHRH stimulates three major responses in somatotropes: (i) exocytotic release of
stored GH (immediate e¡ect); (ii) de novo GH gene transcription and GH synthesis
(acute and short-term actions); and (iii) somatotrope cellular hypertrophy and
proliferation (longer-term response) (Giustina & Veldhuis 1998, Mueller et al
1999). Whereas few if any studies show consistently positive regulation of the
GHRHergic pathway by oestrogen (Argente et al 1991), in the male rat (non-
aromatizable) androgens up-regulate hypothalamic GHRH gene and peptide
expression and enhance pituitary responsiveness to GHRH (Jansson et al 1993,
Mueller et al 1999, Argente et al 1991). Likewise, in leuprolide-down-regulated
young men, testosterone addback restores the acute stimulatory e¡ect of GHRH
(Devesa et al 1991). Notably, several independent clinical studies suggest that
endogenous GHRHergic activity is reduced in the ageing male: (1) intermittent
i.v. infusion of GHRH (0.33 mg/kg every 90 min for 72 h) ampli¢es pulsatile GH
secretion by several fold in older men, albeit not to young-adult levels (Iranmanesh
et al 1998); (2) a GHRH-receptor antagonist inhibits GH secretion more e¡ectively
in older than young individuals, consistent with reduced hypothalamic GHRH
secretion and/or SS excess (Russell-Aulet et al 1999); and (3) post-SS rebound
GH secretion is impaired in the elderly, suggesting limited endogenous GHRH
drive (degli Uberti et al 1997). However, how testosterone modulates the
GHRHergic pathway in ageing humans remains unknown.
Based on the foregoing physiology, and testosterone’s ability to rescue hypo-
somatotropism in older men (Gentili et al 2000), we postulate that testosterone
can (a) enhance endogenous GHRH secretion and/or (b) facilitate somatotrope
responsiveness to GHRH. To this end, we are carrying out clinical experiments
to drive the GHRHergic signalling pathway in older men while simultaneously
¢xing inputs by SS and GHRP before and after testosterone supplementation.

Hypothesis 3: Testosterone de¢ciency increases

the hypothalamic release and/or actions of SS in ageing men
Beyond its primary repressive role (above), SS paradoxically maintains somatotrope
responsiveness to recurrent stimulation by secretagogues (Baumbach et al 1998,
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 109

Kraicer et al 1986, Sugihara et al 1989). Speci¢cally, intermittent SS exposure

obviates the biochemical down-regulation of GH secretion induced by repeated
GHRH and GHRP stimuli (Smith et al 1997). Thus, a critical mechanistic question
is whether testosterone depletion limits GH secretion in part by augmenting
sustained rather than intermittent SSergic activity (Fryburg et al 1997, Giustina et
al 1997, Giustina & Veldhuis 1998, Veldhuis et al 1997).
Clinical studies are particularly pertinent here, in view of possible species
di¡erences in SS and androgen physiology (Argente et al 1990, Painson et al
2000, Pincus et al 1997). For example, in the rat, non-aromatizable androgens
(but not oestrogen) stimulate GH secretion and up-regulate hypothalamic
SSergic activity. In contrast, in the human, non-aromatizable androgens
(stanozolol, £uoxymesterone, oxandrolone and 5a-DHT) do not stimulate GH
secretion consistently (Devesa et al 1991, Fryburg et al 1997, Giustina &
Veldhuis 1998, Keenan et al 1993, Veldhuis et al 1997). Thus, aromatizable and
nonaromatizable androgen actions are readily distinguishable in the two species,
whereas testosterone’s impact on SSergic signalling may not be distinctive.
Clinical studies of this issue are further relevant in older humans, in whom
accentuated SS inhibition is widely inferred, but its susceptibility to relief by sex-
steroid hormone repletion is unknown (Chihara et al 1981). Our pilot data
document that short-term testosterone administration in older men exerts
qualitatively identical actions to those reported in hypogonadal boys and young
men (Fryburg et al 1997, Gentili et al 2000, Giustina et al 1997, Veldhuis et al 1997);
speci¢cally: (i) augmentation of GH secretory burst mass and basal GH release;
(ii) ampli¢cation of the 24 h rhythmicity of GH secretion; (iii) heightening of the
irregularity (approximate entropy) of GH release; and (iv) elevation of plasma
IGF1 concentrations (Fig. 6). All four responses are strongly controlled by SS
under various pathophysiological conditions (Giustina & Veldhuis 1998,
Mueller et al 1999, Mulligan et al 1999b). For example, we have shown that SS or
octreotide infusions selectively suppress GH secretory burst mass and basal GH
release in young and older men (Mulligan et al 1999b). Likewise, SSergic inputs
likely in£uence the 24 h rhythmicity of GH secretion, since the latter persists
during an unvarying exogenous GHRH or GHRP2 infusion (Evans et al 2000,
Iranmanesh et al 1998, Shah et al 1999a, 2000). In corollary, SS signalling
governs the quanti¢able irregularity (entropy) of GH secretory patterns (Gevers
et al 1998, Pincus et al 1996, Straume et al 1995). Lastly, longer-term inhibition of
GH output by SS clearly lowers IGF1 production. Testosterone strongly impacts
each of these four categories of GH/IGF1 responsiveness in older men (Gentili et al
2000), thus pointing to (but not proving) its ability to control SSergic signalling
the elderly male.
Available clinical data do not exclude an opposing hypothesis that an
aromatizable androgen actually elevates SSergic input, as reported for both
110 VELDHUIS ET AL

FIG. 6. Fourfold stimulation by short-term (3 week) parenteral testosterone supplementation

of: (i) GH secretory burst mass; (ii) 24 h rhythmic GH release; (iii) GH ApEn (irregularity
measure); and (iv) serum IGF1 concentrations in older men

aromatizable and non-aromatizable androgens in the rodent (Giustina & Veldhuis

1998, Mueller et al 1999, Chihara et al 1981, Frohman 1996). Indeed,
diethylstilbestrol administration to young men and the normal preovulatory
milieu in young women enhance GH release stimulated by secretagogues that
putatively withdraw SSergic restraint (Frantz & Rabkin 1965). Given such
divergent clinical data, it will be crucial to clarify how testosterone de¢ciency
modulates SSergic signalling in the human.

Hypothesis 4: testosterone deprivations

attenuate activity of the GHRP pathway in older men
Several pivotal factors motivate the important, but unproven, consideration that
de¢cient endogenous GHRPergic activity contributes to the hyposomatotropism
of ageing. Foremost among these are partial reconstitution of GH/IGF1 axis
output by continuous s.c. GHRP2 infusion or oral MK0677 (a non-peptidyl
agonist) administration in older individuals, and diminished hypothalamic
GHRP receptor density in ageing humans (Chapman et al 1996, Mueller et al
1999). The timeliness of this query is highlighted by the recent cloning of a
natural 3Ser-octanoylated 28 amino-acid ligand of the GHRP receptor, ghrelin,
in the rat and human. Ghrelin gene transcripts are expressed by RT-PCR in the
brain, and ghrelin peptide is detected by immunocytochemistry in the arcuate
nucleus of the hypothalamus and by radioimmunoassay in peripheral human
blood (Kojima et al 1999). Since neither the native peptide nor derivative
antagonists are available for clinical use, we and others have probed
responsiveness of the human GHRP-receptor/e¡ector pathway using a potent
and highly speci¢c synthetic (hexapeptide) agonist, such as GHRP2 (our FDA
IND #38 149) (Bowers 1993, 1998, Bowers et al 1990, Pihoker et al 1998, Shah
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 111

et al 1998a,b, 1999a,c, 2000). GHRP2 is the most e¡ective GHRP-receptor agonist

available for investigational use in the human. GHRP2 synergizes with GHRH in
stimulating GH secretion in healthy men (Fig. 4), and enhances oestradiol’s drive
of GH secretion in postmenopausal women (Evans et al 2000, Shah et al 1998a,b,
1999c, 2000). Other clinical studies show that GHRPs exert maximal acute
stimulatory e¡ects in mid-to-late puberty when sex-steroid hormone
concentrations peak. Moreover, a single i.m. injection of testosterone in boys
and brief oestrogen exposure in girls double GH stimulation by a near-maximally
e¡ective dose of the GHRP, hexarelin (Loche et al 1997). Thus, sex steroids may
modulate GHRP-receptor/e¡ector activity in the human, as inferred recently in the
postnatal rat and GH-transgenic mouse. Indeed, the promoter of the human
GHRP-receptor gene contains a hemioestrogen-responsive element. Thus, we
postulate that testosterone could act either pre- or post-aromatization to up-
regulate responsiveness of this key secretagogue pathway.
Not all studies point to sex-steroid dependent control of the GHRP pathway.
For example, GH responses to a single (near-maximal) i.v. bolus dose of a GHRP
did not: (i) vary within the menstrual cycle; (ii) di¡er in women and men; or (iii)
change following low-dose (0.05 mg daily) transdermal oestradiol administration
in postmenopausal individuals . However, the foregoing analyses were carried out
in the face of variable (and, hence, potentially confounding) endogenous SSergic
and GHRH inputs. In fact, based on a simpli¢ed tri-peptidyl model of GH
neuroregulation (Fig. 3), we reason that testosterone’s reported facilitation of
maximal GHRP stimulation in prepubertal boys (Loche et al 1997) could re£ect
testosterone’s (non-exclusive) ability to: (a) facilitate GHRP-receptor/e¡ector
signalling (Mueller et al 1999); (b) reduce concomitant SSergic activity; and/or
(c) augment the release and/or actions of endogenous GHRH. Only the ¢rst
interpretation would be consistent with GHRP-receptor/e¡ector up-regulation.
The second consideration arises because SS partially opposes GHRP’s actions
(Bowers 1998, Iranmanesh et al 1999, Wideman et al 2000a,b). The third
postulate is signi¢cant, since GHRPs can release and act synergistically with
GHRH (above). Accordingly, further experimental studies will be needed to
resolve these key mechanistic distinctions.
In addition to receptor-level regulation, the secretion of endogenous GHRP-
like ligands could be controlled by androgen. Since ghrelin is also produced by
gastric (oxyntic) cells (Kojima et al 1999), simply assaying its concentration in
peripheral blood would not be fully informative. Accordingly, to explore
testosterone’s regulation of endogenous GHRP release and actions, one would
need to block SS secretion and ¢x GHRH inputs simultaneously (Bray et al 1999,
degli Uberti et al 1997, Dickerman et al 1993, Giustina & Veldhuis 1998). This
stratagem should appraise how testosterone modulates non-SS- and non-GHRH-
dependent (i.e. putative GHRP-like) endogenous drive of the human GH axis
112 VELDHUIS ET AL

(Bray et al 1999, Giustina & Veldhuis 1998, Iranmanesh et al 2000, Shah et al

1998a,b, 1999c, Wideman et al 2000a,b).

Distinctions in the actions of androgen and oestrogen on the GH/IGF1 axis

The neuroendocrine mechanisms by which testosterone governs pulsatile GH
secretion in the human have remained elusive, in part because of species
di¡erences in the nature of androgenic control of the GH/IGF1 axis and in part
due to the tripeptidyl control of this specialized axis (Giustina & Veldhuis 1998).
Moreover, known actions of oestrogen are not necessarily equivalent to those
postulated for testosterone. Notably, in clinical studies testosterone, but not
oestradiol, consistently stimulates both GH and IGF1 production and elevates
(non-pulsatile) GH secretion (Bellantoni et al 1991, Devesa et al 1991, Giustina
& Veldhuis 1998, Karlsson et al 1990, Shah et al 1999b, van Kesteren et al 1996,
Blumenfeld et al 1992, De Leo et al 1993). Thus, for example, the putative
neuroendocrine mechanism of oestrogen’s unleashing of GH secretion in
postmenopausal women by way of reduced systemic IGF1 feedback cannot be
facilely invoked to explicate testosterone’s combined stimulation of GH and
IGF1 secretion in ageing and/or hypogonadal men.

Summary
Advances in GHRH, GHRP and SS peptide chemistry, receptorology and
neuroregulatory physiology now create a unique platform for more informative
and insightful clinical studies of the mechanisms of testosterone’s control of the
ageing human GH/IGF1 axis. We suggest that the relatively hyposomatotropic
and hypoandrogaemic older male should a¡ord an excellent clinical context in
which to explore such issues. In addition, based on the cardinal role of sex steroid
hormones in sustaining GH secretion throughout the adult human lifetime
(Giustina & Veldhuis 1998), such studies should also provide signi¢cant
corollary insights into the regulatory pathophysiology of the GH/IGF1 axis
during early puberty and other hypogonadal states.

References
Argente J, Chowen BJ, Steiner RA, Clifton DK 1990 Somatostatin messenger RNA in
hypothalamic neurons is increased by testosterone through activation of androgen receptors
and not by aromatization to estradiol. Neuroendocrinology 52:342^349
Argente J, Chowen JA, Zeitler P, Clifton DK, Steiner RA 1991 Sexual dimorphism of growth
hormone-releasing hormone and somatostatin gene expression in the hypothalamus of the rat
during development. Endocrinology 128:2369^2375
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 113

Arvat E, Ceda GP, Di Vito L et al 1998 Age-related variations in the neuroendocrine control,
more than impaired receptor sensitivity, cause the reduction in the GH-releasing activity of
GHRP’s in human aging. Pituitary 1:51^58
Balducci R, Toscano V, Mangiantini A, Bianchi P, Guglielmi R, Boscherini B 1993 The e¡ect of
growth hormone administration on testicular response during gonadotropin therapy in
subjects with combined gonadotropin and growth hormone de¢ciencies. Acta Endocrinol
128:19^23
Barinaga M, Bilezikjian LM, Vale WW, Rosenfeld MG, Evans RM 1985 Independent e¡ects of
growth hormone releasing factor on growth hormone release and gene transcription. Nature
314:279^281
Baumann G, Maheshwari H 1997 The Dwarfs of Sindh: severe growth hormone (GH)
de¢ciency caused by a mutation in the GH-releasing hormone receptor gene. Acta Pediatr
Suppl 432:33^38
Baumbach WR, Carrick TA, Pausch MH et al 1998 A linear hexapeptide somatostatin antagonist
blocks somatostatin activity in vitro and in£uences growth hormone release in rats. Mol
Pharmacol 54:864^873
Bellantoni MF, Harman SM, Cho DE, Blackman MR 1991 E¡ects of progestin-opposed
transdermal estrogen administration on growth hormone and insulin-like growth factor-I in
postmenopausal women of di¡erent ages. J Clin Endocrinol Metab 72:172^178
Berman M, Ja¡e CA, Tsai W, DeMott-Friberg R, Barkan AL 1994 Negative feedback regulation
of pulsatile growth hormone secretion by insulin-like growth factor I: involvement of
hypothalamic somatostatin. J Clin Invest 94:138^145
Blackman MR 1987 Pituitary hormones and aging. Endocrinol Metab Clin North Am 16:981^
994
Blumenfeld Z, Barkey RJ, Youdim MB, Brandes JM, Amit T 1992 Growth hormone (GH)-
binding protein regulation by estrogen, progesterone, and gonadotropins in human: the
e¡ect of ovulation induction with menopausal gonadotropins, GH, and gestation. J Clin
Endocrinol Metab 75:1242^1249
Bowers CY 1993 GH releasing peptides
structure and kinetics. J Pediatr Endocrinol 6:21^31
Bowers CY 1998 Synergistic release of growth hormone by GHRP and GHRH: scope and
implication. In: Bercu BB, Walker RF (eds) Growth hormone secretagogues. Springer-
Verlag, New York, p 1^25
Bowers CY, Momany FA, Reynolds A, Hong A 1984 On the in vitro and in vivo activity of a
new synthetic hexapeptide that acts on the pituitary to speci¢cally release growth hormone.
Endocrinology 114:1537^1545
Bowers CY, Reynolds GA, Durham D, Barrera CM, Pezzoli SS, Thorner MO 1990 Growth
hormone (GH)-releasing peptide stimulates GH release in normal men and acts
synergistically with GH-releasing hormone. J Clin Endocrinol Metab 70:975^982
Bray M, Shah N, Veldhuis JD 1999 Impact of estrogen on the dose-dependent inhibition by
somatostatin of GH secretion in postmenopausal women. Presented at the 81st Annual
Endocrine Society Meeting, San Diego, California, June 12^15
Carani C, Granata AR, De Rosa M et al 1999 The e¡ect of chronic treatment with GH on gonadal
function in men with isolated GH de¢ciency. Eur J Endocrinol 140:224^230
Carlsson LMS, Clark RG, Robinson IC 1990 Sex di¡erence in growth hormone feedback in the
rat. J Endocrinol 126:27^35
Chapman IM, Bach MA, Cauter EV et al 1996 Stimulation of the growth hormone (GH)-insulin-
like growth factor I axis by daily oral administration of a GH secretagogue (MK-677) in
healthy elderly subjects. J Clin Endocrinol Metab 81:4249^4257
Chapman IM, Hartman ML, Pezzoli SS et al 1997 E¡ect of aging on the sensitivity of growth
hormone secretion to insulin-like growth factor-I negative feedback. J Clin Endocrinol
Metab 82:2996^3004
114 VELDHUIS ET AL

Chihara K, Minamitani N, Kaji H, Arimura A, Fujita T 1981 Intraventrically injected growth

hormone stimulates somatostatin release into rat hypophyseal portal blood. Endocrinology
109:2279^2281
Clark RG, Carlsson LMS, Robinson ICAF 1988 Growth hormone (GH) secretion in the
conscious rat: negative feedback of GH on its own release. J Endocrinol 119:201^209
Copinschi G, Van Cauter E 1994 Pituitary hormone secretion in aging: role of circadian
rhythmicity and sleep. Eur J Endocrinol 131:441^442
Corpas E, Harman SM, Pineyro MA, Roberson R, Blackman MR 1992 Growth hormone (GH)-
releasing hormone-(1-29) twice daily reverses the decreased GH and insulin-like growth
factor-I levels in old men. J Clin Endocrinol Metab 75:530^535
De Leo V, Lanzetta D, D’Antona D, Danero S 1993 Growth hormone secretion in
premenopausal women before and after ovariectomy: e¡ect of hormone replacement
therapy. Fertil Steril 60:268^271
degli Uberti EC, Ambrosio MR, Cella SG et al 1997 Defective hypothalamic growth hormone
(GH)-releasing hormone activity may contribute to declining GH secretion with age in man. J
Clin Endocrinol Metab 82:2885^2888
Devesa J, Lois N, Arce V, Diaz MJ, Lima L, Tresguerres JA 1991 The role of sexual steroids in
the modulation of growth hormone (GH) secretion in humans. J Steroid Biochem Mol Biol
40:165^173
Dickerman Z, Guyda H, Tannenbaum GS 1993 Pretreatment with somatostatin analog SMS
201-995 potentiates growth hormone (GH) responsiveness to GH-releasing factor in short
children. J Clin Endocrinol Metab 77:652^657
Dickson SL, Leng G, Dyball RE, Smith RG 1995 Central actions of peptide and non-peptide
growth hormone secretagogues in the rat. Neuroendocrinology 61:36^43
Dudl RJ, Ensinck JW, Palmer HE, Williams RH 1973 E¡ect of age on growth hormone
secretion in man. J Clin Endocrinol Metab 37:11^16
Evans WS, Bowers CY, Hull LT, Anderson SM, Veldhuis JD 2000 E¡ects of estradiol and of
continuous 24-hour GHRH- and GHRP-2 stimulation in women on pulsatile and entropic
GH secretion. Presented at the Endocrine Society 82nd Annual Meeting, Toronto, Canada,
June 21^24 Abstract # 2016
Frantz AG, Rabkin MT 1965 E¡ects of estrogen and sex di¡erence on secretion of human
growth hormone. J Clin Endocrinol Metab 25:1470^1480
Frohman LA 1996 New insights into the regulation of somatotrope function using genetic and
transgenic models. Metabolism 45:1^3
Frohman LA, Jansson JO 1986 Growth hormone-releasing hormone. Endocr Rev 7:223^253
Fryburg DA, Weltman A, Jahn LA et al 1997 Short-term modulation of the androgen milieu
alters pulsatile, but not exercise or growth hormone releasing hormone-stimulated GH
secretion in healthy men. J Clin Endocrinol Metab 82:3710^3719
Gentili A, Mulligan T, Godschalk M et al 2000 Testosterone supplementation increases growth
hormone secretion in older but not young men. Presented at the Endocrine Society 82nd
Annual Meeting, Toronto, Canada, June 21^24 Abstract # 1944
Gevers E, Pincus SM, Robinson IC, Veldhuis JD 1998 Di¡erential orderliness of the GH release
process in castrate male and female rats. Am J Physiol 274:R437^R444
Ghigo E, Go⁄ S, Nicolosi M et al 1990 Growth hormone (GH) responsiveness to combined
administration of arginine and GH-releasing hormone does not vary with age in man. J Clin
Endocrinol Metab 71:1481^1485
Giustina A, Veldhuis JD 1998 Pathophysiology of the neuroregulation of growth hormone
secretion in experimental animals and the human. Endocr Rev 19:717^797
Giustina A, Scalvini T, Tassi C et al 1997 Maturation of the regulation of growth hormone
secretion in young males with hypogonadotropic hypogonadism pharmacologically
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 115

exposed to progressive increments in serum testosterone. J Clin Endocrinol Metab 82:1210^

1219
Guillaume V, Magnan E, Cataldi M et al 1994 Growth hormone (GH)-releasing hormone
secretion is stimulated by a new GH-releasing hexapeptide in sheep. Endocrinology
135:1073^1076
Haji M, Kato KI, Nawata H, Ibayashi H 1980 Age-related changes in the concentrations of
cytosol receptors for sex steroid hormones in the hypothalamus and pituitary gland of the
rat. Brain Res 204:373^386
Harel Z, Tannenbaum GS 1992 Synergistic interaction between insulin-like growth factors-I
and -II in central regulation of pulsatile growth hormone secretion. Endocrinology
131:758^764
Hartman ML, Faria AC, Vance ML, Johnson ML, Thorner MO, Veldhuis JD 1991 Temporal
structure of in vivo growth hormone secretory events in man. Am J Physiol 260:E101^E110
Ho£and LJ, Lamberts SWJ 1996 Somatostatin receptors and disease: role of receptor subtypes.
Baillieres Clin Endocrinol Metab 10:163^176
Howard AD, Feighner SD, Cully DF et al 1996 A receptor in pituitary and hypothalamus that
functions in growth hormone release. Science 273:974^977
Iranmanesh A, Lizarralde G, Veldhuis JD 1991 Age and relative adiposity are speci¢c negative
determinants of the frequency and amplitude of growth hormone (GH) secretory bursts and
the half-life of endogenous GH in healthy men. J Clin Endocrinol Metab 73:1081^1088
Iranmanesh A, Grisso B, Veldhuis JD 1994 Low basal and persistent pulsatile growth hormone
secretion are revealed in normal and hyposomatotropic men studied with a new ultrasensitive
chemiluminescence assay. J Clin Endocrinol Metab 78:526^535
Iranmanesh A, South S, Liem AY et al 1998 Unequal impact of age, percentage body fat, and
serum testosterone concentrations on the somatotrophic, IGF-I, and IGF-binding protein
responses to a three-day intravenous growth-hormone-releasing-hormone pulsatile infusion
in men. Eur J Endocrinol 139:59^71
Iranmanesh A, Gupta PM, Bowers CY, Veldhuis JD 1999 GHRH and GHRP-2 interact to
stimulate GH secretion in men in the presence of low-dose synthetic somatostatin
(octreotide): evidence for a novel collaboration between GHRH and GHRP. Presented at
the 81st Annual Endocrine Society Meeting, San Diego, California, June 12^15
Iranmanesh A, Bowes CY, Hussein RH, Veldhuis JD 2000 Singular and interactive impact of
acute infusions of GHRH, GHRP-2, and octreotide pretreatment on GH neuroregulation in
middle-aged men. Presented at the Endocrine Society 82nd Annual Meeting, Toronto,
Canada, June 21^24 Abstract # 2019
Ja¡e CA, Friberg RD, Barkan AL 1993 Suppression of growth hormone (GH) secretion by a
selective GH-releasing hormone (GHRH) antagonist. Direct evidence for involvement of
endogenous GHRH in the generation of GH pulses. J Clin Invest 92:695^701
Jansson JO, Sta¡an E, Isaksson OGP 1993 Sexual dimorphism in the control of growth
hormone secretion. Endocr Rev 128:150^168
Karlsson R, Eden S, von Schoultz B 1990 Altered growth hormone secretion during oral
contraception. Gynecol Obstet Invest 30:234^238
Keenan BS, Richards GE, Ponder SW, Dallas JS, Nagamani M, Smith ER 1993 Androgen-
stimulated pubertal growth: the e¡ects of testosterone and dihydrotestosterone on growth
hormone and insulin-like growth factor-I in the treatment of short stature and delayed
puberty. J Clin Endocrinol Metab 76:996^1001
Khorram O, Laughlin GA, Yen SS 1997 Endocrine and metabolic e¡ects of long-term
administration of [Nle27]growth hormone-releasing hormone-(1-29)-NH2 in age-advanced
men and women. J Clin Endocrinol Metab 82:1472^1479
116 VELDHUIS ET AL

Klindt J, Ford JJ, Macdonald GJ 1990 Synergism of testosterone propionate with growth
hormone in promoting growth of hypophysectomized rats: e¡ect of sexual di¡erentiation. J
Endocrinol 127:249^256
Kojima M, Hosoda H, Date Y, Nakazato M, Matsuo H, Kangawa K 1999 Ghrelin is a growth-
hormone-releasing acylated peptide from stomach. Nature 402:656^660
Kraicer J, Cowan JS, Sheppard MS, Lussier B, Moor BC 1986 E¡ect of somatostatin withdrawal
and growth hormone (GH)-releasing factor on GH release in vitro: amount available for
release after disinhibition. Endocrinology 119:2047^2051
Kulin HE, Samojlik E, Santen RJ, Santner S 1981 The e¡ect of growth hormone on the Leydig
cell response to chorionic gonadotropin in boys with hypopituitarism. Clin Endocrinol
15:463^472
Lin T, Haskell J, Vinson N, Terracio L 1986 Direct stimulatory e¡ects of insulin-like growth
factor 1 on Leydig cell steroidogenesis in primary culture. Biochem Biophys Res Commun
137:950^956
Loche S, Colao A, Cappa M et al 1997 The growth hormone response to hexarelin in children:
reproducibility and e¡ect of sex steroids. J Clin Endocrinol Metab 82:861^864
Locke W, Kirgis HD, Bowers CY, Abdoh AA 1995 Intracerebroventicular growth-hormone-
releasing peptide-6 stimulates eating without a¡ecting plasma growth hormone responses in
rats. Life Sci 56:1347^1352
Martha PM Jr, Gorman KM, Blizzard RM, Rogol AD, Veldhuis JD 1992 Endogenous growth
hormone secretion and clearance rates in normal boys as determined by deconvolution
analysis: relationship to age, pubertal status and body mass. J Clin Endocrinol Metab
74:336^344
Mayo KE 1992 Molecular cloning and expression of a pituitary-speci¢c receptor for growth
hormone-releasing hormone. Mol Endocrinol 6:1734^1744
Mˇller EE, Locatelli V, Cocchi D 1999 Neuroendocrine control of growth hormone secretion.
Physiol Rev 79:511^607
Mulligan T, Iranmanesh A, Kerzner R, Demers LW, Veldhuis JD 1999a Two-week pulsatile
gonadotropin releasing hormone infusion unmasks dual (hypothalamic and Leydig cell)
defects in the healthy aging male gonadotropic axis. Eur J Endocrinol 141:257^266
Mulligan T, Jaen-Vinuales A, Godschalk M, Iranmanesh A, Veldhuis JD 1999b Synthetic
somatostatin analog (octreotide) suppresses daytime growth hormone secretion
equivalently in young and older men: preserved pituitary responsiveness to somatostatin’s
inhibition in aging. J Am Geriatr Soc 47:1422^1424
Mulligan T, Kuno H, Clore J, Iranmanesh A, Veldhuis JD 2000 Pulsatile infusions of
recombinant human LH in leuprolide-downregulated older vs young men unmask an
impoverished Leydig-cell secretory response in aging to mid-physiological LH stimuli.
Presented at the Endocrine Society 82nd Annual Meeting, Toronto, Canada, June 21^24
Abstract #249
Painson JC, Veldhuis JD, Tannenbaum GS 2000 Single exposure to testosterone in adulthood
rapidly induces regularity in the growth hormone release process. Am J Physiol 278:E933^
E940
Pellegrini E, Bluet-Pajot MT, Mounier F, Bennett P, Kordon C, Epelbaum J 1996 Central
administration of a growth hormone (GH) receptor mRNA antisense increases GH
pulsatility and decreases hypothalamic somatostatin expression in rats. J Neurosci 16:8140^
8148
Pihoker C, Kearns GL, French D, Bowers CY 1998 Pharmacokinetics and pharmacodynamics of
growth hormone-releasing peptide-2: a phase I study in children. J Clin Endocrinol Metab
83:1168^1172
Pincus SM, Gevers E, Robinson IC et al 1996 Females secrete growth hormone with more
process irregularity than males in both humans and rats. Am J Physiol 270:E107^E115
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 117

Pincus SM, Veldhuis JD, Mulligan T, Iranmanesh A, Evans WS 1997 E¡ects of age on the
irregularity of LH and FSH serum concentrations in women and men. Am J Physiol
273:E989^E995
Rosenthal SM, Hulse JA, Kaplan SL, Grumbach MM 1986 Exogenous growth hormone
inhibits growth hormone-releasing factor-induced growth hormone secretion in normal
men. J Clin Invest 77:176^180
Russell-Aulet M, Ja¡e CA, Friberg RD, Barkan AL 1999 In vivo semiquanti¢cation of
hypothalamic growth hormone-releasing hormone (GHRH) output in humans: evidence
for relative GHRH de¢ciency in aging. J Clin Endocrinol Metab 84:3490^3497
Shah N, Anderson SM, Azimi PP, Evans WS, Bowers CY, Veldhuis JD 1998a Estradiol
augments basal GH pulse mass and ampli¢es GH-secretory responsiveness to low doses of
GHRP-2. Presented at the Fifth International Pituitary Congress, Naples, Florida, June 28^30
Shah N, Anderson SM, Azimi PP, Evans WS, Bowers CY, Veldhuis JD 1998b Growth hormone
releasing peptide-2 (GHRP-2) induced growth hormone (GH) release is in£uenced by oral
estrogen replacement in postmenopausal women. Presented at the Growth Hormone
Research Society Conference, San Francisco, California, September 3^6
Shah N, Evans WS, Bowers CY, Veldhuis JD 1999a Tripartite neuroendocrine activation of the
human growth-hormone (GH) axis in women by continuous 24-hour GH-releasing peptide
infusion: pulsatile, entropic, and nyctohemeral mechanisms. J Clin Endocrinol Metab
84:2140^2150
Shah N, Evans WS, Veldhuis JD 1999b Actions of estrogen on the pulsatile, nyctohemeral, and
entropic modes of growth hormone secretion. Am J Physiol 276:R1351^R1358
Shah N, Bowers CY, Veldhuis JD 1999c Pulsatile nyctohemeral and entropic modes of GH
secretion after estradiol (E2), GHRP-2 or the combinatin in post-menopausal women.
Presented at the Southern Regional Meeting of the American Federation of Medical
Resarch, New Orleans, Louisiana, February 18^20
Shah N, Evans WS, Bowers CY, Veldhuis JD 2000 Oral estradiol administration modulates
continuous intravenous growth hormone (GH)-releasing peptide-2 driven GH secretion in
postmenopausal women. J Clin Endocrinol Metab 85:2649^2659
Smith RG, van der Ploeg LH, Howard AD et al 1997 Peptidomimetic regulation of growth
hormone secretion. Endocr Rev 18:621^645
Snyder PJ, Peachey H, Hannoush P et al 1999 E¡ect of testosterone treatment on body
composition and muscle strength in men over 65 years of age. J Clin Endocrinol Metab
84:2647^2653
Straume M, Chen L, Johnson ML, Veldhuis JD 1995 Systems-level analysis of physiological
regulation interactions controlling complex secretory dynamics of growth hormone axis: a
connectionist network model. Methods Neurosci 28:270^310
Sugihara H, Minami S, Wakabayashi I 1989 Post-somatostatin rebound secretion of growth
hormone is dependent on growth hormone-releasing factor in unrestrained female rats. J
Endocrinol 122:583^591
Tenover JS, Dahl KD, Vale WW, Rivler JE, Bremner WJ 1990 Hormonal responses to a potent
gonadotrophin hormone-releasing hormone antagonist in normal elderly men. J Clin
Endocrinol Metab 71:881^888
Urban RJ, Bodenburg YH, Gilkison C et al 1995 Testosterone administration to elderly men
increases skeletal muscle strength and protein synthesis. Am J Physiol 269:E280^E286
van Kesteren P, Lips P, Deville W et al 1996 The e¡ect of one-year cross-sex hormonal treatment
on bone metabolism and serum insulin-like growth factor-1 in transsexuals. J Clin Endocrinol
Metab 81:2227^2232
Veldhuis JD 1996 Male hypothalamic-pituitary-gonadal axis. In: Lipshultz LI, Howards SS
(eds) Infertility in the male. Mosby, Philadelphia, PA, p 23^58
118 DISCUSSION

Veldhuis JD, Giustina A 2000 Neuroendocrinology strategies for appraising suspected growth
hormone de¢ciency in the adult. In: Loriaux LD (ed) The endocrinologist. Lippincott
Williams & Wilkins, Baltimore, MD, vol 10:34^50
Veldhuis JD, Iranmanesh A, Ho KK, Waters MJ, Johnson ML, Lizarralde G 1991 Dual defects
in pulsatile growth hormone secretion and clearance subserve the hyposomatotropism of
obesity in man. J Clin Endocrinol Metab 72:51^59
Veldhuis JD, Liem AY, South S et al 1995 Di¡erential impact of age, sex-steroid hormones, and
obesity on basal versus pulsatile growth hormone secretion in men as assessed in an
ultrasensitive chemiluminescence assay. J Clin Endocrinol Metab 80:3209^3222
Veldhuis JD, Metzger DL, Martha PM Jr et al 1997 Estrogen and testosterone, but not a non-
aromatizable androgen, direct network integration of the hypothalamo-somatotrope (growth
hormone)-insulin-like growth factor I axis in the human: evidence from pubertal
pathophysiology and sex-steroid hormone replacement. J Clin Endocrinol Metab 82:3414^
3420
Veldhuis JD, Evans WS, Shah N, Story S, Bray MJ, Anderson SM 1999a Proposed mechanisms
of sex-steroid hormone neuromodulation of the human GH^IGF-I axis. In: Veldhuis JD,
Giustina A (eds) Sex-Steroid interactions with growth hormone. Springer-Verlag, New
York, p 93^121
Veldhuis JD, Iranmanesh A, Mulligan T, Pincus SM 1999b Disruption of the young-adult
synchrony between luteinizing hormone release and oscillations in follicle-stimulating
hormone, prolactin, and nocturnal penile tumescence (NPT) in healthy older men. J Clin
Endocrinol Metab 84:3498^3505
Veldhuis JD, Roemmich JN, Rogol AD 2000a Gender and sexual maturation-dependent
contrasts in the neuroregulation of growth hormone secretion in prepubertal and late
adolescent males and females
a general clinical research center-based study. J Clin
Endocrinol Metab 85:2385^2394
Veldhuis JD, Iranmanesh A, Godschalk M, Mulligan T 2000b Older men manifest multifold
synchrony disruption of reproductive neurohormone out£ow. J Clin Endocrinol Metab
85:1477^1486
Weltman A, Weltman JY, Hartman ML et al 1994 Relationship between age, percentage body
fat, ¢tness and 24-hour growth hormone release in healthy young adults: e¡ects of gender. J
Clin Endocrinol Metab 78:543^548
Wennink JMB, Delemarre-van de Waal HA, Schoemaker R, Blaauw G, van den Braken C,
Schoemaker J 1990 Growth hormone secretion patterns in relation to LH and testosterone
secretion throughout normal male puberty. Acta Endocrinol 123:263^270
Wideman L, Weltman JY, Patrie JT et al 2000a Synergy of L-arginine and GHRP-2 stimulation
of growth hormone in men and women: modulation by exercise. Am J Physiol 279:R1467^
R1477
Wideman L, Weltman JY, Patrie JT et al 2000b Synergy of L-arginine and growth hormone
(GH)-releasing peptide-2 on GH release: in£uence of gender. Am J Physiol 279:R1455^
R1466
Zachmann M 1992 Interrelations between growth hormones and sex hormones: physiology and
therapeutic consequences. Horm Res 38:1^8

DISCUSSION
Bj˛rntorp: You measured these things every second minute, and then you have a
pulse of say LH. This then has to reach the testosterone-producing site, causing
testosterone secretion. This then has to feedback. Is this actually happening? I can
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 119

understand the nervous circuit, because this is quick, but the circulatory circuit is
slower.
Veldhuis: Yes, there is feedback. One of the things that is strangely lacking is the
exact kinetics of the feedback time delays. I have spent a year reading 600 papers
and cannot ¢nd exact time delays. We are going to use the drug ketoconazole,
which blocks cytochrome P450 activity at high doses and thus steroidogenesis.
One dose of this drug lowers testosterone concentrations overnight from 25 nM
to about 3 nM. Now you have a testosterone-withdrawn state in which you can
clamp feedback. At midnight, we start a constant testosterone infusion that is
about one-third of the expected amount produced endogenously over the same
time interval. At 8 a.m. the next morning we pulse in a 6 min waveform of
testosterone and monitor feedback timing in young and older men. Currently,
we are guessing feedback timing on the basis of cross-correlation data, which
means that we take 15 older men and 15 younger men, and sample for LH and T
simultaneously for 24 h. We then have paired series. We ask the question,
‘whenever T goes down, how long does it take for LH to go up?’ In the human,
this negative feedback has about a 60^90 min delay in the young and 0^60 min
delay in the older male. Feedback is occurring, but this is the only way that we
have estimated how long it takes for the system to react.
Bj˛rntorp: You are not saying that the decrease of the LH peak is dependent on an
immediate feedback.
Veldhuis: No, that is a good point
it is unlike the cortisol^adrenocorticotropic
hormone (ACTH) axis where there is some evidence for that. This is not so rapid.
Bj˛rntorp: So the LH peaks are sort of automatic?
Veldhuis: That is what we believe, but we think the pulse generator frequency is
under T control in the human.
Handelsman: It is a wonderfully subtle illustration of entropy. The power of the
approach is enormous. I just wondered whether you have thought about this
another way.
Veldhuis: This is the advantage of the current formulation by Steve Pincus,
which we used in Pincus et al (1996). It is a lag-independent cross-approximate
entropy (ApEn) metric. One of the disadvantages of simple linear cross-
correlation is that it is assumed that each subject has roughly the same
relationship between the two hormones in time, and also that within any one
person there is a similar relationship across the day and night. This may not be
true. When we run windows of cross-correlation, we ¢nd that the strength of
feed-forward coupling varies across 24 h. It sounds intuitively obvious, but it is
very clear for ACTH^cortisol and LH^T. The beauty of cross-ApEn is that it is
lag-independent. It basically asks the question, given standardized z-score
transforms of the original time series to make them scale independent, if there are
some wiggles in the ¢rst series of hormones, do they ever happen in the second? If
120 DISCUSSION

they never do, they are not very synchronous. If they do happen a lot together, they
are fairly synchronous. If they are happening almost all the time, they are highly
synchronous. It matches templates up- and downstream independently of location.
Handelsman: One of the beauties of this technique is that it is model-independent.
However, the di⁄culty is that it makes entropy seem such an abstract notion. To be
a bit more concrete, in the studies where you deliberately gave LH pulses, how does
that look under your model? Can you override the apparent age-related increase in
entropy by administering clear and coherent LH pulses?
Veldhuis: We intend to look at this. When we use the GnRH pump in older men,
we see a result that at ¢rst is counterintuitive. There is a more random output of LH
in younger and older men under perfectly regular 90 min experimentally enforced
pulsatile GnRH drive (Mulligan et al 1999). Why is this? We are actually
monitoring minute-to-minute feedback activity between T and/or LH under the
GnRH stimulus and not the 90 min pulse. This is a microanalysis that is checking
feedback adjustment (Veldhuis 1999a). When the system is forced with ¢xed input,
this abolishes feedback. Thus, ApEn of LH actually goes up on the GnRH pump in
young men. The axis essentially becomes a clamped system with a non-dynamic
quality of feedback. The feedback elegance is abolished by the clamp.
Handelsman: The physiological pharmacodynamic models of Jusko and
colleagues are very similar; they are constructed of components like that. Are
your thoughts going in this direction?
Veldhuis: This is the idea. We have a couple of papers out using testosterone,
which took us ¢ve years to write, because there are surprisingly large subtleties
in how to build a dose^response curve and prove that the set of equations is
realizable mathematically (Keenan & Veldhuis 1998, 2001, Keenan et al 2000).
There are some things that are produced only in the square root of minus one,
which is an answer that has no utility to us as clinicians, being an unde¢ned term.
We have now done this for ACTH and we are just getting there for GH (Farhy et al
2001). The idea is only to take the core components. The ¢rst time we did this for
GH, we tried to do it comprehensively (Straume et al 1995). We ended up with 87
parameters, and it may take a few dozen years for computers to be developed with
the power to optimize this collection of parameters. Now we are down to 12
parameters for GH and about 10^12 for LH. Without infusing LH, I’d like to
know that the older Leydig cell of the mouse, human or rat is unresponsive to
LH. How do I do that? I have to watch LH and T move together in young and
older men and calculate the endogenous dose^response curve, without ever seeing
it. I can’t do that without a correct statement of how the dose^response curve
primarily operates. If I can do that, I can tell you the dose^response curve
without injecting anything.
Robertson: I’m not as familiar with chaos theory applied to the endocrine system
as I am for example with its application to heart rate. But looking at your data I
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 121

would never have guessed you would be able to account for nearly everything
that you see by the interactions of these two variables. What percentage of
in£uences do you think are outside this paradigm?
Veldhuis: That is a gorgeous question: I just wish you had been a reviewer. A
reviewer once said that we didn’t need any stochastic, random element that is
unexplained, but that we should just draw the correct feedback loop. My
response was that nothing is absolutely constant. If I stand up, my LH
distribution volume is slightly di¡erent. If I walk, my testis blood£ow is
changed. Everything is changing slightly outside the idealized dose^response.
The aggregate uncertainty can be instilled in a stochastic di¡erential equation.
Our stochastic term is only 2^3% of the data, but it is critical. Where is the
stochastic element? Firstly, the pulse generator doesn’t ¢re exactly as a clock
ticks. We allow it some variability (Veldhuis 1999a, Urban et al 1988). Second,
we say that the feedback equations we are talking about are idealized equations.
They are never observed. The feedback equations are dancing slightly, but at all
times. We put in a little stochastic term to allow the parameters of the feedback
equations to jiggle by 2^3%. This gives our realistic data pro¢les. Otherwise one
observes gorgeous curves that you and I only see one in 20 pro¢les.
Handelsman: This is such a highly deterministic system. One feature is that you
can explain virtually all T secretion by LH pulsatility as shown by the fact that you
can switch it completely o¡ with antagonists or steroids. This makes it reasonable
to say that you can predict nearly all the components, because we can easily identify
proximate determinants that can switch it completely o¡. This isn’t true for most
other physiological systems, where there is a very lower proportion of variance
explained.
Veldhuis: It is strongly deterministic with just 2^3% stochastic. This ¢ts with the
fact that we are not a couple of molecules reacting in free solution.
Giustina: One point we know from the ageing GH axis is that ageing is
interacting with obesity in creating a loss of GH secretion. What about your
model of LH and testosterone in obesity?
Veldhuis: All we know is that in general the literature agrees that LH pulse
amplitude is damped in some manner by visceral obesity in particular. Most
studies show this e¡ect to be quite strong, and at least as strong as the age e¡ect
in middle age. What isn’t clear to me is what is mediating this e¡ect. It could be the
insulin levels.
Shalet: Where does oestradiol ¢t in to this?
Veldhuis: That is a frightening question because the situation is getting more and
more complicated. The oestrogen receptor knockout or the aromatase gene-
defective animal have about a doubling of LH output. We ¢nd the same with the
drug anastrozole, an aromatase inhibitor. In situ aromatase activity appears to be
important in negative feedback. It isn’t clear whether this is controlling only
122 DISCUSSION

amplitude or frequency. We have had to study 31 men to try to get a clear answer.
We have used three di¡erent pulse methods to see whether we can get a consistent
opinion on it. The amplitude clearly changes, so there is amplitude drive. I had
been puzzled why men show an increased LH pulse frequency on clomifene or
tamoxifen, both antioestrogens, but when you infuse peripheral oestrogen in the
human, you can almost never demonstrate a suppression of LH pulse frequency
(Veldhuis et al 1984, Veldhuis & Dufau 1987, Urban et al 1988). The exception is
a study in which we put an oestrogen-containing silastic ring intravaginally in
postmenopausal women, delivering oestradiol (Veldhuis et al 1987). There, on
day 5, LH pulse frequency fell with amplitude and then recovered. In the
monkey, one can infuse peripheral oestrogen, which decreases hypothalamic
multiunit ¢ring within minutes. But, excluding those exceptions, people cannot
readily demonstrate oestradiol negative feedback on frequency in the male. This
is puzzling. Is it the in situ hypothalamic oestradial that suppresses the pulse
generator frequency? As far as I can tell, this would explain the tamoxifen/
clomifene and anastrozole data. These drugs would block oestradiol produced in
the hypothalamus. This would also explain the fact that peripheral oestradiol
infusion, in almost everybody’s hands, mainly blocks GnRH-driven LH
amplitude at the pituitary level. The aromatase inhibitor study will be key to try
to dissect whether that is true.
Laron: What down-regulates prolactin?
Veldhuis: I would love to know. Older men are hypoprolactinaemic (Iranmanesh
et al 1999). So are type I diabetic patients (Iranmanesh et al 1990). There are a
collection of curious situations where there is reversible hypoprolactinaemia. But I
don’t know other data that give us clear evidence for dopamine excess.
Mˇller: There could be an up-regulation of the dopamine receptor.
Veldhuis: That would be beautiful.
Mˇller: We have data from experiments in older rats showing increased pituitary
sensitivity to the prolactin-lowering e¡ect of bromocriptine (Cocchi et al 1984).
Moreover in aged female rats presenting with marked alterations in the
tuberoinfundibular dopaminergic neuronal function, pituitary binding sites for
[3H] spiroperidol, a neuroleptic, are increased (Govoni et al 1980).
Laron: This relates to my ¢rst question: how much of the dose^response is
actually linked to the number of receptors?
Veldhuis: The data in the rat do not consistently show loss of GnRH receptor or
GnRH activity with ageing.
Wang: The GnRH responsiveness is normal in old rats.
Veldhuis: In our hands it is also normal to enhanced in the human (Mulligan et al
1999, Zwart et al 1996).
Morley: Johannes, you have done a great job of trying to take the complex and
make it intelligible. Unfortunately, I prefer complexity. As I look at the literature,
FAILURE OF THE SOMATOTROPIC AND GONADAL AXES IN AGEING 123

when we start to look closely, young and old are not young and old: there are at
least four separate phases. There is the 20^32 year old who we could call normal
young. Somewhere after 32 we start to lose some testosterone in almost all studies.
When you get to early middle age (40^55), there are some hints that there is excess
opioid secretion at this stage and these people are very responsive to opiate
antagonism in their LH levels. Then we get the group you have been talking
about, the 60^75 year olds. In longitudinal studies by ourselves and others, when
you get to 80+ the LH suddenly takes o¡ and gets to 25^30 IU/L. These are clearly a
di¡erent group. I guess the question is, how do you put this together with the
entropy and what causes that sudden release in very old age to going from a quasi
secondary hypogonadism to a primary hypogonadism?
Veldhuis: That is why I was so excited by Christine Wang’s data suggesting that
there may be some ¢xed GnRH defect. I had assumed that the late LH rise was due
to end-stage Leydig cell failure (Veldhuis et al 1999b). Of course, no one has done
the kind of near-physiological drive of the Leydig cells that modern tools allow us
to do. The idea of formalizing several levels in the axis that are points of lesioning
in ageing is that you can then test them and their implications. Certain nodes clearly
don’t lead to exactly what you predict. The reason is that if there is a non-linear
interactive, time-delayed system, intuition is stymied (Keenan & Veldhuis 1998,
2001, Keenan et al 2000). However, what we thought we would do is create a
model in which we let the computer run overnight, producing a seven year
lifespan. We let it gradually trickle down one of the feedback constants, and then
introduce a couple of lesions on top of that. This gives an idea of whether it is
possible to unmask the phases. The other real challenge is the between-individual
variability. There you have individuals who appear to be absolutely normal in their
entropy scores, and yet their pulse generator looks awfully good.
Morley: Could you go back and look at the people you studied a long time ago?
This is probably the key to understanding all this.
Veldhuis: Believe it or not, since you like complexity, I will soon have data of
10 min LH ApEn analyses collected for four consecutive days, to watch the pulse
generator unfold and £uctuate over the day and night. We intend to compare this
in older and young men. We postulate that as one gets into these metastable
conditions (by which I mean that they are not absolutely normal, but they are not
pathological), the stability of the pulse generator over four days will be degraded
(Pincus et al 1996). We may be wrong, and it may prove to be even more stable in
the elderly, which would be a more exciting paper.

References
Cocchi D, Novelli A, Ganzetti I, Mˇller EE 1984 In vivo supersensitivity of the anterior
pituitary of old female rats to dopaminergic inhibition of prolactin secretion. Gerontology
30:345^349
124 DISCUSSION

Farhy LS, Straume M, Johnson ML, Kovatchev BP, Veldhuis JD 2001 A construct of
interactive feedback control of the GH axis in the male. Am J Physiol 281:R38^R51
Govoni S, Memo M, Saiani L, Spano PF, Trabucchi M 1980 Impairment of brain
neurotransmitter receptor in aged rats. In: Mechanisms of aging and development, chapter
1^2. Elsevier Sequoia, Lausanne, p 39^46
Iranmanesh A, Veldhuis JD, Carlson EC et al 1990 Attenuated pulsatile release of prolactin in
men with insulin-dependent diabetes mellitus. J Clin Endocrinol Metab 71:73^78
Iranmanesh A, Mulligan T, Veldhuis JD 1999 Mechanisms subserving the physiological
nocturnal relative hyperprolactinaemia of healthy older men: dual decline in prolactin
secretory burst mass and basal release with preservation of pulse duration, pulse frequency,
and interpulse interval: a general clinical research center study. J Clin Endocrinol Metab
84:1083^1090
Keenan DM, Veldhuis JD 1998 A biomathematical model of time-delayed feedback in the
human male hypothalamic^pituitary^Leydig cell axis. Am J Physiol 275:E157^E176
Keenan DM, Veldhuis JD 2001 Hypothesis testing of the aging male gonadal axis via a
biomathematical construct. Am J Physiol Regul Integr Comp Physiol 280:R1755^R1771
Keenan DM, Sun W, Veldhuis JD 2000 A stochastic biomathematical model of the male
reproductive hormone system. SIAM J Appl Math 61:934^965
Mulligan T, Iranmanesh A, Kerzner R, Demers LW, Veldhuis JD 1999 Two-week pulsatile
gonadotropin releasing hormone infusion unmasks dual (hypothalamic and Leydig cell)
defects in the healthy aging male gonadotropic axis. Eur J Endocrinol 141:257^266
Pincus SM, Mulligan T, Iranmanesh A, Georghiu S, Godschalk M, Veldhuis JD 1996 Older
males secrete luteinizing hormone and testosterone more irregularly, and jointly more
asynchronously, than younger males. Proc Natl Acad Sci USA 93:14100^14105
Straume M, Chen L, Johnson ML, Veldhuis JD 1995 Systems-level analysis of physiological
regulation interactions controlling complex secretory dynamics of growth hormone axis: a
connectionist network model. Methods Neurosci 28:270^310
Urban RJ, Evans WS, Rogol AD, Kaiser DL, Johnson ML, Veldhuis JD 1988 Contemporary
aspects of discrete peak detection algorithms. I. The paradigm of the luteinizing hormone
pulse signal in men. Endocr Rev 9:3^37
Veldhuis JD 1999a Male hypothalamic-pituitary-gonadal axis. In: Yen SSC, Ja¡e RB, Barbieri
RL (eds) Reproductive Endocrinology. W.B. Saunders Co., Philadelphia p 622^631
Veldhuis JD 1999b Recent insights into neuroendocrine mechanisms of aging of the human
male hypothalamo-pituitary-gonadal axis. J Androl 20:1^17
Veldhuis JD, Dufau ML 1987 Estradiol modulates the pulsatile secretion of biologically active
luteinizing hormone in man. J Clin Invest 80:631^638
Veldhuis JD, Rogol AD, Samojlik E, Ertel NH 1984 Role of endogenous opiates in the
expression of negative feedback actions of androgen and estrogen on pulsatile properties of
luteinizing hormone secretion in man. J Clin Invest 74:47^55
Zwart AD, Urban RJ, Odell WD, Veldhuis JD 1996 Contrasts in the gonadotropin-releasing
dose^response relationships for luteinizing hormone, follicle-stimulating hormone, and a
subunit release in young versus older men: appraisal with high speci¢city immunoradiatric
assay and deconvolution analysis. Eur J Endocrinol 135:399^406
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

E¡ects of growth hormone and

insulin-like growth factor 1
de¢ciency on ageing and longevity
Zvi Laron

Endocrinology & Diabetes Research Unit, Schneider Children’s Medical Center of Israel and
Sackler Faculty of Medicine, Tel Aviv University, Israel

Abstract: Present knowledge on the e¡ects of growth hormone (GH)/insulin-like growth

hormone (IGF)1 de¢ciency on ageing and lifespan are reviewed. Evidence is presented
that isolated GH de¢ciency (IGHD), multiple pituitary hormone de¢ciencies (MPHD)
including GH, as well as primary IGF1 de¢ciency (GH resistance, Laron syndrome)
present signs of early ageing such as thin and wrinkled skin, obesity, hyperglycemia and
osteoporosis. These changes do not seem to a¡ect the lifespan, as patients reach old age.
Animal models of genetic MPHD (Ames and Snell mice) and GH receptor knockout mice
(primary IGF1 de¢ciency) also have a statistically signi¢cant higher longevity compared
to normal controls. On the contrary, mice transgenic for GH and acromegalic patients
secreting large amounts of GH have premature death. In conclusion longstanding
GH/IGF1 de¢ciency a¡ects several parameters of the ageing process without impairing
lifespan, and as shown in animal models prolongs longevity. In contrast high GH/IGF1
levels accelerate death.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 125^142

In contrast to growth and development, ageing is a progressive process

orchestrated by decreasing synthesis and secretion of numerous factors and
hormones; among them growth hormone (GH) and its anabolic e¡ector
hormone, insulin-like growth factor 1 (IGF1). Therefore, ageing is often
compared with growth hormone de¢ciency (GHD) (Toogood & Shalet 1998).
This assumption is based on the evidence that pituitary GH secretion and serum
IGF1 concentrations decline with increasing age (Gil-Ad et al 1984, Arvat et al
2000), reaching low levels in late adulthood, and have similarities to changes of
body appearance, composition and function (Carroll et al 1998, Toogood &
Shalet 1998) (Table 1). These ¢ndings led to trials of GH treatment in elderly
people (Rudman et al 1990). The ¢nding that GH increased lean body mass,
decreased adiposity and improved apparent skin changes gave birth to an
125
126 LARON

TABLE 1 Similarities between GH de¢ciency and ageing

Thinning of skin (wrinkling)

Excess adipose tissue (obesity)
Rise in insulin resistance (tendency for diabetes)
Decline in b cell function
Reduced lean body (muscle) mass
Reduced physical performance
Reduced mineral density (osteoporosis)
Lowered venous access













Rise in serum cholesterol

approved (Butter¢eld et al 1997) and non-approved administration of GH to

ageing people, at ‘so-called’ rejuvenation clinics. These medical acts were
reinforced by reports that GH de¢ciency increases the risk for cardiovascular
disease (Rose¤ n et al 1993) and leads to premature mortality (Rose¤ n & Bengtsson
1990).
In order to analyse the present knowledge on the possible role of GH and IGF1
in ageing and lifespan, this paper reviews states of congenital (i.e. primary) GH
and/or IGF1 de¢ciency in humans and animals. Attention is paid as to whether
patients or animals with GH/IGF1 de¢ciency present early signs of ageing and
e¡ects on the duration of their lifespan.

Congenital isolated GH de¢ciency (IGHD)

Previously called idiopathic GHD, modern laboratory technology has shown that
this state can be caused by molecular defects of the GH releasing hormone
(GHRH) gene or receptor, or the hGH gene (Laron 2001). These patients are
presently diagnosed at an early age and treated with the unlimited amounts of
available biosynthetic hGH. Therefore, there is little information on studies of
adult patients with isolated GH de¢ciency (IGHD). In 1969 Merimee and
colleagues reported 31 patients with hereditary IGHD, whose age ranged from
13^78 years (Merimee et al 1969). Among the clinical descriptions Merimee &
Laron (1996) wrote that wrinkling of the skin often began early in life and these
patients consequently looked prematurely old (Table 2). Ten out of 13 males and 9
of 18 females had wrinkled skin. Rimoin et al (1966) performed skin biopsies and
found decreased soluble collagen in the dermis of two thirds of these patients. The
histological changes found were probably the underlying cause of thinning and
EFFECTS OF GH/IGF1 DEFICIENCY ON AGEING AND LIFESPAN 127

TABLE 2 Clinical characteristics of adult patients with untreated IGHD

n = 31 13 males 18 females

Age (years) 13^78

Height (cm) 110^140
Wrinkled skin 10 9
High pitched voice 12 5

Based on data from Merimee & Laron (1996).

wrinkling of the skin, characteristic of GH/IGF1 de¢ciency, and these in turn are
caused by the lack of anabolic e¡ect on collagen and hydroxyproline of these
hormones. The only other histological data on the skin have been obtained by
Abramovici et al (1983) who studied the skin biopsies of 35 children and
adolescents including 18 with IGHD. These latter investigators found that
patients with IGHD lack elastic ¢bres in the skin papillary layer and an uneven
distribution of elastic ¢bres in the reticular layer.

Congenital multiple pituitary hormone

de¢ciencies (MPHD) including growth hormone
In 1988, we had the opportunity to examine six out of 10 living dwarfed patients,
part of 24 related patients recorded on the island of Krk in the Adriatic Sea. They
belong to two villages near to each other, and were known to have existed since the
end of the 19th century (Hanhart 1925). DNA from these patients revealed a
mutation in the PROP1 gene (a transcription factor) causing MPHD (thyroid-
stimulating hormone, prolactin, luteinizing hormone, follicle-stimulating
hormone and GH de¢ciencies) (Krzisnik et al 1999). They were treated by
thyroxine, which some took irregularly. Only one 14 year old girl received GH.
The ¢ve adult patients ranged from 47 to 68 years (3 males, 2 females). In addition
to short stature (120^139 cm), they were obese, sexually immature and had a very
wrinkled skin (Fig. 1). Notably the patients did not have any grey hair despite their
advanced adult age. This was also seen on a picture of a 70 year old patient found on
his tombstone (Fig 2). The information obtained for four deceased patients
revealed that they had died at ages 68, 77, 83 and 91.

Laron syndrome (primary GH resistance)

The following describes adult patients with primary IGF1 de¢ciency due to
primary GH resistance or insensitivity (i.e. Laron syndrome). In 1966 and in
1968 our group described a new hereditary syndrome resembling IGHD, but
128 LARON

FIG. 1. Appearance of a 58 year old patient with GH de¢ciency due to a PROP1 gene
mutation. Note wrinkled and loose skin.

FIG. 2. Appearance of a 70 year old patient with GH de¢ciency due to a PROP1 mutation.
Note absence of grey hair and loose skin. For details see text. Reproduced with permission from
Krzisnik et al (1999).
EFFECTS OF GH/IGF1 DEFICIENCY ON AGEING AND LIFESPAN 129

FIG. 3. Early ageing appearance of a 39 year old female with Laron syndrome.

with very high serum hGH levels (Laron et al 1968). Since then we have been
following in Israel a cohort of 51 patients from infancy to adulthood (Laron
1999a, Laron & Parks 1993). Since the ¢rst description several hundred
patients, or their descendants, have been described with Laron syndrome, mainly
in Mediterranean and Mid-Eastern populations (Rosenfeld et al 1994, Laron
1999a). This syndrome is caused by deletions or mutations in the GH receptor or
postreceptor pathways (Godowski et al 1989, Amselem et al 1996, Laron 1999a,
1999b), leading to an inability by the liver and possible other tissues to generate
IGF1 (Laron et al 1971), the anabolic e¡ector hormone of GH (Laron 1999c).
Studying adult patients with Laron syndrome (Laron & Klinger 1993, 1994,
Laron 1999b) we observed that these patients remain very short (females, 108^
136 cm; males, 119^142 cm; adult height), have an early ageing appearance (such
as a wrinkled face at an early adult age; Fig. 3), and relatively thin skin on their
hands. Abramovici et al (1983) performed skin biopsies in six children and late
adolescents and found that patients with Laron syndrome had bundles of
thickened elastic ¢bres in the upper dermis.
Even young adult patients presenting with Laron syndrome develop marked
general and visceral obesity (Fig. 4), high cholesterol levels (Laron & Klinger
1993), reduced muscular strength (Brat et al 1997), insulin resistance (Laron et al
130 LARON

FIG. 4. Marked obesity in a 41 year old female with Laron syndrome.

1997), osteoporosis (Laron & Klinger 1994), and/or su¡er from psychological
de¢ciencies (Galatzer et al 1993), all features characteristic for normal ageing and
usually apparent at a later age. The oldest patient followed by us is a 73 year old
male; one lady examined by us only once and suspected (but not proven) to have
Laron syndrome died at age 53. She had su¡ered from asthma and coronary heart
disease (Laron 1999b). Also, adult patients in the large Ecuadorian cohort of Laron
syndrome patients have been reported to have reached ages of 70 years or more
(Rosenbloom et al 1999). It is of note that with one exception none of our adult
patients has grey hair. However, they have a tendency for baldness (in males) and
thin hair (in females) (Laron et al 2001).
In conclusion, the relatively small number of adult patients with IGHD or
MPHD never previously treated with GH, as well as patients with primary IGF1
de¢ciency (Laron syndrome) not treated by IGF1, show a series of early
developing characteristics compatible with ageing such as thinning and
wrinkling of skin, obesity, muscle weakness, osteoporosis and hyperlipidermia.
In contradistinction to the postulation of Rose¤n & Bengtsson (1990) that
hypopituitary patients have premature mortality due to cardiovascular disease
(Rose¤ n et al 1993), the patients with GH and IGF1 de¢ciency live a long life,
despite the signs of early ageing. One big di¡erence between our patients and
EFFECTS OF GH/IGF1 DEFICIENCY ON AGEING AND LIFESPAN 131

those reported by Rose¤ n et al (Rose¤ n & Bengtsson 1990, Rose¤ n et al 1993) is that
almost all of those reported by Rose¤ n and colleagues had tumours, mostly pituitary
adenomas, and were treated either by surgery or irradiation; all were MPHD and
received a combination of hormone replacement treatments with the exception of
GH. Therefore, those patients cannot be compared to the patients with IGHD and/
or IGF1 de¢ciency, and the statement that GH or IGF1 de¢ciency shortens the
lifespan seems incorrect.
A review of animal studies using models of GH or IGF1 de¢ciency also revealed
that the lifespan in these animals is prolonged compared to intact animals.

Ageing and lifespan in GH/IGF1-de¢cient mice

Several mouse models with GH/IGF1 de¢ciency are available to study the
in£uence of these hormones on ageing and longevity (lifespan). Due to the IGF1
de¢ciency all homozygous a¡ected mice are dwarfed and they divide as the human
models into MPHD including GH (e.g. the Ames dwarf mice and the Snell dwarf
mice) and primary IGF1 de¢ciency (the Laron mouse).

The Ames dwarf mice (df/df)

These mice ¢rst described by Schaible & Gowen (1961) have a mutation in a
transcription factor for all anterior pituitary hormones (GH, prolactin, thyroid-
stimulating hormone and sex hormones) called Prophet of Pit.1 (Prop-1)
(Sornson et al 1996), which is located on chromosome 11. Homozygous mice for
the df/df mutation are dwarfed, and have a longer life span than control animals,
which is not related to caloric intake or the reduced body temperature (Bartke
1998).

Snell dwarf mice (dw/dw)

Described in 1929 this type of dwarfed mouse has been subsequently shown to be
caused by a mutation of the transcription factor Pit-1 (Li et al 1990) which is
involved in the di¡erentiation of somatotrophs, lactotrophs and thyrotrophs
(Sornson et al 1996). Phenotypically, the Ames and Snell dwarfed mice are very
similar with the exception that in Snell mice the gonadal development is more
advanced.
Snell mice have extremely low serum levels of IGF1 (van Buul O¡ers et al 1986).
They have also been described to have delayed ageing and a longer lifespan than
normal animals from the same strain (Bartke 2000).
The ageing symptoms of these two types of MPHD mice are their retarded
sexual development, reduced activity (not in all), progressive obesity and hair
132 LARON

FIG. 5. Increased longevity in Ames dwarfed mice compared to normal controls of the same
breed. Reproduced with permission from Bartke (2000).

loss (in part of the animals). Nevertheless, these animals appear to remain in
excellent general condition for longer periods than their normal siblings (Bartke
2000). A group of Ames dwarfed mice outlived a control group of normal mice by
more than one year (Bartke 2000) (Fig. 5). This extension of lifespan was longer in
female mice.

The GH receptor/BP gene-disrupted mice (the Laron mouse)

A model of isolated IGF1 de¢ciency was created by disrupting the GH receptor
(GHR) gene in mice (Zhou et al 1997). This model bears many similarities to the
human primary GH resistance (GH insensitivity, i.e. Laron syndrome) (Kopchick
& Laron 1999), such as high GH and low IGF1 and IGF binding protein
(IGFBP)3 levels, dwar¢sm and organomicria, typical characteristics of Laron
syndrome (Laron 1999b). Calculating average lifespans for each genotype (+/+,
+/7, 7/7) and gender, we observed that the homozygous mice for the GHR
mutation had a signi¢cantly longer lifespan than the una¡ected and heterozygote
mice (Coschigano et al 2000) (Table 3).
EFFECTS OF GH/IGF1 DEFICIENCY ON AGEING AND LIFESPAN 133

TABLE 3 Lifespan of GHR/BP gene-disrupted mice

Gender Genotype n Lifespan (days)*

Males +/+ 7 629+72

+/7 8 668+51
7/7 7 975+106a
Females +/+ 13 749+41
+/7 19 701+36
7/7 11 1031+41b

*Mean SE.
a
P 5 0.02 compared to +/+.
b
P5 0.005 compared to +/+.
Reproduced with permission from Coschigano et al (2000).

GH transgenic mice
In contrast to the previously described observations, prolonged elevation of serum
GH, as occurs in GH transgenic mice, is associated with a reduced lifespan (Bartke
1998), which may reach half of that in normal mice of the same species. This is
similar to ¢ndings in patients with acromegaly. Thus, the question arises whether
high levels of GH increase mortality. In e¡ect, treatment of rats with high doses of
GH accelerates the death of the animals (Groesbeck et al 1987). Although the
conditions may be di¡erent, one should remember that GH treatment of
chronically ill patients in intensive care units was also found to increase mortality
(Takala et al 1999).
At present it is not clear how GH/IGF1 de¢ciency prolongs the lifespan in mice.
It is possible that certain genes are involved. Mutations of a recently described
insulin receptor like-gene, Daf2, result in increased longevity (Kimura et al
1997). This receptor, possibly homologous to the mammalian IGF1 receptor,
mimics primary IGF1 de¢ciency. Nor do we fully understand how GH excess
shortens the lifespan. This may be partly due to the water and electrolyte
retention induced by GH/IGF1 and/or by the well-documented cardiotrophic
e¡ects of these hormones.
Although it may sound anecdotal it should be mentioned that there is evidence
that within species, lifespan is negatively correlated with body size. Thus dogs
from small breeds live longer than dogs from large breeds and small mice live
longer than large mice (Bartke 2000). Last but not least, food-restricted animals
(which are smaller), live longer than those fed ad libitum (Masoro 1992), with the
exception of Ames and Snell mice.
134 LARON

The premature ageing syndromes

It was of interest to ¢nd out whether the rare genetic disorders known as premature
ageing syndromes (Pesce & Rothe 1996) are related to a disorder in the secretion of
GH or IGF1. All are characterized by marked growth retardation associated with
early and fast ageing, various dermal changes (wrinkling, loose skin), hypo-
trichosis and early greying of the hair, and early death mostly by heart attacks due
to atherosclerosis or congestive heart failure.

Progeria (Hutchinson^Gilford disease)

First described in 1886 by Hutchinson, the incidence is estimated at 1 per million
live births. The syndrome is characterized by increased hyaluronic acid excretion.
We found only one report by Villee et al (1969) who studied two boys with
classical progeria and found lack of GH response to insulin-induced
hypoglycemia.

Werner’s syndrome (adult progeria)

First described in 1904 this disease is clinically characterized by short stature, skin
changes (soft tissue wasting ulcerated hyperkeratosis, pigmentation) ¢ne hair and
alopecia; early greying of hair and a tendency to malignancies and insulin-resistant
diabetes. There is only one report on GH de¢ciency in Werner syndrome (Rubin &
Reed 1996).

Cockayne’s syndrome
Reported in 1936 and 1946 by Cockayne, it is characterized by dwar¢sm associated
with retinal atrophy and deafness as well as skin atrophy. Endocrine function in a
group of patients revealed normal GH response to stimulation tests in eight
patients, decreased response in four, and an exaggerated response in three
children (Nance & Berry 1992).

Bloom syndrome
Bloom syndrome is a rare autosomal recessive disorder characterized by growth
de¢ciency, skin changes, photosensitivity, variable degrees of immunode¢ciency,
predisposition to malignancies, type 2 diabetes and early death. It is caused by
mutations in the gene BLM. Growth retardation is a major characteristic of
Bloom syndrome but GH de¢ciency has so far not been documented.
EFFECTS OF GH/IGF1 DEFICIENCY ON AGEING AND LIFESPAN 135

Rothmund^Thomson syndrome
This disease ¢rst described in 1986 is a rare autosomal recessive disorder charac-
terized by short stature, skin changes (consisting of atrophy and telangiectasis) and
hair loss (Kaufmann et al 1986). The skin changes resemble Bloom syndrome.
Kaufmann et al (1986) reported GH de¢ciency in an 11 year old girl when tested
by arginine, L-Dopa and GH-releasing hormone (GRF 1^44). No similar reports
have come to our attention.
In summary, the ¢nding of only very few patients with GH de¢ciency among the
patients with ‘premature ageing syndromes’ of genetic origin, the majority of
whom have normal pituitary functions, indicates that their accelerated ageing
and various complications are not related to GH or IGF1. The rare instances of
GH de¢ciency must be considered coincidental.

Conclusion
From the clinical and experimental studies reviewed it ensues that longstanding
GH/IGF1 de¢ciency of genetic origin does not shorten lifespan. On the
contrary, it may prolong it, as is clearly evident from animal models. This occurs
simultaneously with the development of some characteristic changes of early
ageing (thinning of skin, wrinkling, obesity, reduction in lean body mass) but
arrest of other signs, such as greying of hair. It has also been shown that high
levels of GH accelerate death. How exactly GH and IGF1 a¡ect the ageing
process and duration of life remains to be established.

References
Abramovici A, Josefsberg Z, Mimouni M, Liban E, Laron Z 1983 Histopathological features of
the skin in hypopituitarism and Laron-type dwar¢sm. Isr J Med Sci 19:515^519
Amselem S, Sobrier ML, Dastot F, Duquesnoy P, Duriez B, Goossens M 1996 Molecular basis
of inherited growth hormone resistance in childhood. Baillie' re’s Clin Endocrinol Metab
10:353^369
Arvat E, Broglio F, Ghigo E 2000 Insulin-like growth factor I: implications in aging. Drugs
Aging 16:29^40
Bartke A 1998 Growth hormone and aging. Endocrine 8:103^108
Bartke A 2000 Delayed aging in Ames dwarf mice. Relationships to endocrine function and body
size. In: Hekimi (ed) The molecular genetics of aging. Results and problems in cell
di¡erentiation. Springer-Verlag, Berlin, vol 26:181^202
Brat O, Ziv I, Klinger B, Avraham M, Laron Z 1997 Muscle force and endurance in untreated
and human growth hormone or insulin-like growth factor-I-treated patients with growth
hormone de¢ciency or Laron syndrome. Horm Res 47:45^48
Butter¢eld GE, Thompson J, Rennie MJ, Marcus R, Hintz RL, Ho¡man AR 1997 E¡ect of
rhGH and rhIGF-I treatment on protein utilization in elderly women. Am J Physiol
272:E94^E99
136 LARON

Carroll PV, Christ ER, Bengtsson BA et al 1998 Growth hormone de¢ciency in adulthood and
the e¡ects of growth hormone replacement: a review. Growth Hormone Research Society
Scienti¢c Committee. J Clin Endocrinol Metab 83:382^395
Coschigano KT, Clemmons D, Bellush LL, Kopchick JJ 2000 Assessment of growth parameters
and life span of GHR/BP gene-disrupted mice. Endocrinology 141:2608^2613
Galatzer A, Aran O, Nagelberg N, Rubitzek J, Laron Z 1993 Cognitive and psychosocial
functioning of young adults with Laron syndrome. In: Laron Z, Parks JS (eds) Pediatric
and adolescent endocrinology, vol 24: Lessons from Laron syndrome (LS) 1966^1992.
Karger, Basel, p 53^60
Gil-Ad I, Gurewitz R, Marcovici O, Rosenfeld J, Laron Z 1984 E¡ect of aging on human plasma
growth hormone response to clonidine. Mech Ageing Dev 27:97^100
Godowski PJ, Leung DW, Meacham LR et al 1989 Characterization of the human growth
hormone receptor gene and demonstration of a partial gene deletion in two patients with
Laron-type dwar¢sm. Proc Natl Acad Sci USA 86:8083^8087
Groesbeck MD, Parlow AF, Daughaday WH 1987 Stimulation of supranormal growth in
prepubertal adult plateaued and hypophysectomized female rats by large doses of rat growth
hormone: physiological e¡ects and adverse consequences. Endocrinology 120:1963^1975
Hanhart E 1925 Uber heredodegenerativen Zwergwuchs mit Dystrophia adiposo-genitalis.
Arch Julius Klaus Stift Vererbungsforsch Sozialanthropol Rassenhyg 1:235^253
Kaufmann S, Jones M, Culler FL, Jones KL 1986 Growth hormone de¢ciency in the
Rothmund-Thomson syndrome. Am J Med Genet 23:861^868
Kimura KD, Tissenbaum HA, Liu Y, Ruvkun G 1997 daf-2, an insulin receptor-like gene that
regulates longevity and diapause in Caenorhabditis elegans. Science 277:942^946
Kopchick JJ, Laron Z 1999 Is the Laron mouse an accurate model of Laron syndrome? Mol
Genet Metab 68:232^236
Krzisnik C, Kolacio Z, Battelino T, Brown M, Parks JS, Laron Z 1999 The ‘Little People’ of the
Island of Krk
revisited. Etiology of hypopituitarism revealed. J Endocrine Genet 1:9^19
Laron Z 1999a Laron syndrome: primary growth hormone resistance. In: Jameson JL (ed)
Hormone resistance syndromes. Contemporary endocrinology. Humana Press, Clifton, NJ,
vol 2:17^37
Laron Z 1999b Natural history of the classical form of primary growth hormone (GH) resistance
(Laron syndrome). J Pediatr Endocrinol Metab 12:231^249
Laron Z 1999c Somatomedin-1 (recombinant insulin-like growth factor-1): clinical
pharmacology and potential treatment of endocrine and metabolic disorders. BioDrugs
11:55^70 (erratum: 1999 BioDrugs 11:114)
Laron Z 2001 Molecular mutations in the human growth hormone axis. In: Eugster EA,
Pescovitz OH (eds) Developmental endocrinology from research to clinical practice.
Humana Press, Clifton, NJ
Laron Z, Klinger B 1993 Body fat in Laron syndrome patients: e¡ect of insulin-like growth
factor I treatment. Horm Res 40:16^22
Laron Z, Klinger B 1994 IGF-I treatment of adult patients with Laron syndrome: preliminary
results. Clin Endocrinol 41:631^638
Laron Z, Parks JS 1993 Pediatric and adolescent endocrinology, vol 24: Lessons from Laron
syndrome (LS) 1966^1992. Karger, Basel
Laron Z, Pertzelan A, Karp M 1968 Pituitary dwar¢sm with high serum levels of growth
hormone. Isr J Med Sci 4:883^894
Laron Z, Pertzelan A, Karp M, Kowadlo-Silbergeld A, Daughaday WH 1971 Administration of
growth hormone to patients with familial dwar¢sm with high plasma immunoreactive growth
hormone: measurement of sulfation factor, metabolic and linear growth responses. J Clin
Endocrinol Metab 33:332^342
EFFECTS OF GH/IGF1 DEFICIENCY ON AGEING AND LIFESPAN 137

Laron Z, Avitzur Y, Klinger B 1997 Insulin resistance in Laron syndrome (primary insulin-like
growth factor-I (IGF-I) de¢ciency) and e¡ect of IGF-I replacement therapy. J Pediatr
Endocrinol Metab 10:105^115
Laron Z, Ben-Amitai D, Lurie R 2001 Impaired hair growth and structural defects in patients
with Laron syndrome (Primary IGF-I de¢ciency). Pediat Res 49:76A
Li S, Crenshaw BE 3rd, Rawson EJ, Simmons DM, Swanson LW, Rosenfeld MG 1990 Dwarf
locus mutants lacking three pituitary cell types result from mutations in the POU-domain
gene pit-I. Nature 347:528^533
Masoro EJ 1992 Retardation of aging processes by nutritional means. Ann NY Acad Sci USA
673:29^35
Merimee T, Laron Z 1996 Modern endocrinology and diabetes, vol 4: Growth hormone, IGF-I
and growth: new views of old concepts. Freund Publishing House Ltd, Tel Aviv
Merimee T, Rimoin DL, Hall JD, McKusick VA 1969 A metabolic and hormonal basis for
classifying ateliotic dwarfs. Lancet 1:963^965
Nance MA, Berry SA 1992 Cockayne syndrome: review of 140 cases. Am J Med Genet 42:68^84
Pesce K, Rothe MJ 1996 The premature aging syndromes. Clin Dermatol 14:161^170
Rimoin DL, Merimee TJ, McKusick VA 1966 Sexual ateliotic dwar¢sm: a recessively inherited
isolated de¢ciency of human growth hormone. Trans Assoc Am Physicians 79:297^304
Rose¤ n T, Bengtsson BA 1990 Premature mortality due to cardiovascular disease in
hypopituitarism. Lancet 336:285^288
Rose¤ n T, Ede¤ n S, Larson G, Wilhelmsen L, Bengtsson BA 1993 Cardiovascular risk factors in
adult patients with growth hormone de¢ciency. Acta Endocrinol (Copenh) 129:195^200
Rosenbloom AL, Guevara-Aguirre J, Rosenfeld RG, Francke U 1999 Growth hormone
receptor de¢ciency in Ecuador. J Clin Endocrinol Metab 84:4436^4443
Rosenfeld RG, Rosenbloom AL, Guevara-Aguirre J 1994 Growth hormone (GH) insensitivity
due to primary GH receptor de¢ciency. Endocr Rev 15:369^390
Rubin CD, Reed B 1996 Twenty-four hour growth hormone secretion in a patient with Werner’s
syndrome. Exp Gerontol 31:557^561
Rudman D, Feller AG, Nagraj HS et al 1990 E¡ects of human growth hormone in men over 60
years old. N Engl J Med 323:1^6
Schaible R, Gowen JW 1961 A new dwarf mouse. Genetics 46:896
Sornson MW, Wu W, Dasen JS et al 1996 Pituitary lineage determination by the prophet of Pit-1
homeodomain factor defective in Ames dwar¢sm. Nature 384:327^333
Takala J, Ruokonen E, Webster NR et al 1999 Increased mortality associated with growth
hormone treatment in critically ill adults. N Engl J Med 341:785^792
Toogood AA, Shalet SM 1998 Ageing and growth hormone status. Baillieres Clin Endocrinol
Metab 12:281^296
van Buul-O¡ers S, Veda I, van den Brande JL 1986 Biosynthetic somatomedin C (SM-C/IGF-I)
increases the length and weight of Snell dwarf mice. Pediatr Res 20:825^827
Villee DB, Nichols G Jr, Talbot NB 1969 Metabolic studies in two boys with classical progeria.
Pediatrics 43:207^216
Zhou Y, Xu BC, Maheshwari HG et al 1997 A mammalian model for Laron syndrome produced
by targeted disruption of the mouse growth hormone receptor/binding protein gene (the
Laron mouse). Proc Natl Acad Sci USA 94:13215^13220

DISCUSSION
Shalet: Just a point of fact. Elderly patients with pituitary disease are very
di¡erent in their GH secretion when compared with age-matched controls. Just
138 DISCUSSION

to equate a so-called somatopause with pituitary tumour patients that have organic
GH de¢ciency is incorrect. The di¡erence in the 24 h pro¢le is some 90%. The
patients with pituitary tumours and organic GH de¢ciency have a GH reduction
of 90%, mainly consisting of a decrease in the amplitude of the GH pulse. This
needs to be stated up-front.
I was surprised at the way you were pushing us at the end into thinking it is better
to be IGF1 and GH de¢cient. I ¢nd that puzzling for a man who has spent so many
years of his life ¢ghting to get IGF1 replacement for patients with GH
insensitivity. It is a curious contradiction in terms of your policy. Does this mean
that you will no longer replace GH and IGF1 in children who are GH-de¢cient or
insensitive, respectively, for fear of reducing their potential life expectancy?
Laron: You don’t have to exaggerate. If you have muscle weakness, short stature
and osteoporosis, this should be treated. What I wanted to point out is the
following. (a) The general statement that hyperpituitarism reduces the lifespan is
not true, unless you analyse the precise nature of the hyperpituitarism. (b) Too
much GH is dangerous. We should learn how to administer GH replacement in
order to prevent its negative e¡ects. I am not saying we shouldn’t treat true GH/
IGF1 de¢ciency. However, the dose one should use in ageing adults is still
controversial.
Giustina: I have a couple of points. The parallelism between greying hair and
longevity is not proven. I am not sure there are data showing that people with no
grey hair live longer. Moreover, acromegalic patients do not have very early
greying of hair. This fact raises some doubts in the relationship with GH. When
you quote the data from Rose¤n and Bengtsson on lifespan, if you look at the real
data, there is no big di¡erence between control and population studies. This means
that when you study a very small population such as yours, and say that lifespan is
not reduced, statistically this is not correct. This is because GH-de¢cient subjects
do not die at a very young age. However, they have been proven to have a reduced
lifespan with respect to a comparable population in the same country in the same
registry. If you want to demonstrate what you are saying you need to look at the
comparable population in your territory and see whether on a statistical basis there
is a reduction or not in lifespan. Otherwise, what you are saying is that these people
may sometimes live long and sometimes not. This is a descriptive concept that has
to be proven on a statistical basis.
Laron: One thing is clear. The animal models with isolated IGF1 de¢ciency or
multiple hormone de¢ciency, including GH, live longer. In humans, having
‘congenital GH or IGF1 de¢ciency’ does not shorten the lifespan, as stated by
Rose¤ n & Bengtsson (1990). The population they studied were patients after
operation or irradiation for pituitary tumours.
Giustina: But you didn’t prove it.
Laron: Their statements have been cited in many papers and even textbooks.
EFFECTS OF GH/IGF1 DEFICIENCY ON AGEING AND LIFESPAN 139

Giustina: You need to prove that in your control population in that territory, the
lifespan is the same. Otherwise you are not proving the concept. You are just
describing the fact that some of the patients may live a long time. There is no
statistical evidence supporting your concept.
Laron: With regard to cancer and IGF1, there was a meeting in Halle in
September 2000 which had a clear message. People genetically susceptible to
cancer are very susceptible to IGF1, as they are to sex hormones. In those who
are not genetically susceptible, IGF1 does not induce cancer.
Monson: Concerning the comparison of di¡erent populations, the Lund series
(Bˇlow et al 1997) and the Gothenburg series (Rose¤ n & Bengtsson 1990)
examined patients from the mid-1950s, when the concept of lipid lowering and
healthy lifestyle was less de¢ned. So these were observational, epidemiological
studies in patients who by current standards may have had suboptimal care. We
are superimposing on that background a group of patients who by virtue of
small size may not have su⁄cient power to show a di¡erence in mortality, and
who also by virtue of one’s interest in their clinical problem, are likely to have
had more interventions. It is therefore very di⁄cult to be certain that lifespan is
reduced.
We talk about the Lund study and the Gothenberg study as demonstrating
increased mortality related to GH de¢ciency, but these patients were
predominantly pan hypopituitary. This is a demonstration of increased mortality
in hypopituitary populations, who may have had variable quality of cortisol
replacement. We know that GH de¢ciency itself alters the relationship between
cortisol and cortisone conversion. This is likely to be accentuated in hepatocytes
and adipocytes. Arterial intima^media thickness is increased in hypopituitary
patients and this is partially reversed by GH replacement. Nonetheless, we
should be wary about concluding that GH of itself has any true impact on
atherogenesis and that there is reversibility in terms of GH replacement. Having
said this, I agree with you that the animal data are extremely compelling in terms of
the e¡ect of GH and IGF1 on longevity.
Handelsman: I would add that in several of those studies they did not have
adequate reproductive hormone replacement, either.
Laron: In view of the alleged neuroprotective and neurotrophic actions of IGF,
in people with Laron syndrome there is a decline in the central nervous system
(CNS) function and when they are treated with IGF there is amelioration, as
observed in the treated children. We have not treated adults long enough, but we
have recently found, by MRI, changes in the brains of IGF1-de¢cient adult
patients, and no defect whatsoever in patients treated from childhood onwards.
The longest treatment in paediatric patients is 10 years. It is di⁄cult to compare
adults with children, but judging by the size of the head (which indicates brain
growth; Laron et al 1992) and by psychological tests (Galatzer et al 1993), the
140 DISCUSSION

intrauterine and perinatal IGF1 is of great importance for its neurotrophic e¡ect.
This seems to be preventable by treatment.
Riggs: Are individuals with congenital IGF1 or GH de¢ciency truly susceptible
to osteoporosis? Outside of the bone ¢eld it is not always appreciated that what is
measured by dual-energy X-ray absorptiometry (DEXA) is areal bone density and
not volumetric density. There will be a built in error if you have small bones. It is
possible to correct this through formulae. It would be interesting to see whether
you could do this. I guess the compelling question is, is there enough follow-up on
these patients with regard to whether or not they develop vertebral fractures?
Laron: There aren’t vertebral fractures, but there is cervical stenosis which
develops in adulthood. In children who have been treated for a long time we
don’t see these changes. We know that IGF1 is an anabolic hormone that
in£uences the connective tissue. I wish to mention one more important issue that
relates to GH or IGF replacement therapy, namely quality of life. This is a di⁄cult
subject to study.
Shalet: Adults with childhood-onset GH de¢ciency are as a group signi¢cantly
osteopaenic. Middle-aged subjects with adult-onset GH-de¢ciency are also
osteopaenic, but less so. The elderly-onset GH-de¢cient patients are not
signi¢cantly di¡erent in terms of bone mineral density measurements from age-
matched controls.
Monson: Should we really be using the term osteoporosis in relation to
childhood-onset GH de¢ciency? Isn’t this a peak bone mass issue?
Riggs: Osteoporosis can arise from lack of developing optimal peak bone mass.
But the point is that even though DEXA is universally used, it is not widely
appreciated that if the size of the bone is di¡erent from normal, the normative
values are not useful.
Monson: I was thinking about structure, also. A similar bone mineral density
de¢cit in a child that has failed to reach peak bone mass, compared with a 60 year
old with hypopituitarism, will be associated with quite di¡erent bone morphology
in terms of what will have happened to the struts.
Riggs: That is correct.
Ruiz-Torres: From a gerontological point of view, there are two contradictory
facts that are apparent. On the one hand, GH de¢ciency is not related to short life.
Furthermore, there are experiments, such as those of A. V. Everitt in Australia,
which show that hypophysectomy has lifespan-prolonging e¡ects similar to
those of dietary restriction. On the other hand, the opposite point is that well
being is related to GH concentration, as we have seen in healthy people in
agreement with your GH de¢ciency results.
Laron: Patients with Laron syndrome are short and have reduced muscle mass,
which contributes to the development of osteoporosis. They are hindered in
normal life to varying degrees. These patients also have varying de¢cits in mental
EFFECTS OF GH/IGF1 DEFICIENCY ON AGEING AND LIFESPAN 141

abilities. They should be diagnosed prenatally, and ideally start treatment

prenatally
and certainly no later than at birth.
Veldhuis: Do they tend to have a slightly low core body temperature? We heard
yesterday how fasting tends to reduce mean core temperature. Does anyone know
whether the mice de¢cient in mitochondrial uncoupling protein are also
thermogenic? This would test the theory of thermogenesis, which is painfully
broad.
Handelsman: To avoid therapeutic nihilism about the use of GH, what about
using IGF1 as a way of dose titration? How important is this?
Laron: It is used as a marker for acromegaly. We use IGF1 as a marker for testing
whether the dose of IGF1 is su⁄cient.
Shalet: Remember that 50% of middle-aged hypopit patients start with a normal
IGF1 level. In adult patients we keep the IGF1 between 72 and +2 SDs. We don’t
give them supraphysiological doses of GH because we would then exceed the IGF1
SD score of +2. Clearly, at the other end, for a patent starting with a low IGF1
level, monitoring the IGF1 level helps assess compliance. Otherwise, in terms of
optimizing GH replacement therapy, IGF1 SD score doesn’t really help.
Laron: There is debate concerning whether one has to look at both IGF1 and
also IGF binding protein 3. The consensus now is that IGF1 su⁄ces, and only in
early infancy when IGF1 is low should IGFBP3 be measured.
Elahi: I have reviewed the literature and been unable to ¢nd a consistent number
for what is considered a low IGF1 at any age. Professor Shalet, what number do
you use? We use 4135 pg/mol as our low value.
Shalet: There is no number. We have a service provided by a pharmaceutical
company that makes the GH. Our patients are in an international surveillance
program. We take blood, the sample is sent o¡ to that particular lab and they
have normative data that are decade based and gender based, and they then
issue a standard deviation score that takes into account the normal age-matched
value.
Haus: The circadian peak in plasma GH decreases in the elderly and some
suggestions have been made that the functional state of elderly subjects could be
improved by GH substitution. This clearly is not supported by Professor Laron’s
observation on the life expectancy of subjects with genetic and/or acquired
conditions of habitually high or low GH concentrations. In this context, an
observation on the correlation of the circadian means and amplitudes in plasma
GH concentrations with the functional state of the subjects may be of interest.
Comparing the circadian mean and amplitude of plasma GH (obtained in 279
pro¢les in 149 subjects, 77 8 years of age with 6 measurements over a 24 h span)
with the functional state of the subjects as evaluated by the Index of Independence
in Activities of Daily Living (ADL, Katz) and the Mental Status Index (MSI) we
found a statistically signi¢cant positive correlation of the growth hormone levels
142 DISCUSSION

and amplitudes with functional impairment rather than with functional capacity
(Haus et al 1989).

References
Bˇlow B, Hagmar L, Mikoczy Z, Nordstrom CH, Erfurth EM 1997 Increased cerebrovascular
mortality in patients with hypopituitarism. Clin Endocrinol 46:75^81
Galatzer A, Aran O, Nagelberg N, Rbitzek J, Laron Z 1993 Cognitive and psychosocial
functioning of young adults with Laron syndrome. In: Laron Z, Parks JS (eds) Lessons
from Laron syndrome (LS) 1966^1992. Pediatr Adolesc Endocrinol 24:53^60
Haus E, Nicolau GY, Robu E, Petrescu E, Sackett-Lundeen L 1989 The functional state of
elderly subjects related to the circadian and circannual mean and amplitude in plasma
growth hormone. Chronobiologia 16:142^143
Laron Z, Anin S, Klipper-Aubach Y, Klinger B 1992 E¡ects of insulin-like growth factor on
linear growth, head circumference and body fat in patients with Laron-type dwar¢sm. Lancet
339:1258^1261
Rose¤ n T, Bengtsson BA 1990 Premature mortality due to cardiovascular disease in
hypopituitarism. Lancet 336:285^288
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

The role of insulin-like growth

factor 1 and insulin in ageing and
atherosclerosis
Antonio Ruiz-Torres

Instituto Universitario de Investigacio¤ n, Gerontolo¤ gica y Metabolica (University Research

Institute of Ageing), Hopital de la Princesa, Diego de Leo¤ n 62, 28006 Madrid

Abstract. With advancing age insulin-like growth factor (IGF)1 blood levels decrease
continuously, but with great interindividual di¡erences. There is a relationship between
the IGF1 serum concentration and biomarker behaviour, indicating that growth
hormone (GH) secretion is a determinant of organismic well being and surviving in
advanced age. In contrast, the secretion of insulin rises with age, which is related to
both increasing body fat mass and ageing itself. In vitro insulin stimulates the
proliferation, migration and collagen secretion of human vascular smooth muscle cells
(SMCs). The mechanism underlying these processes mainly involves occupancy of
IGF1 receptors by insulin, with the exception of migration. With advancing age of the
donor, the in vitro proliferation rate and migration capacity of SMCs decreases. When
insulin or IGF1 is added, there is no reversibility, so that there is no recovery to the
values of SMCs from young donors. The blockade of Ca2+ channels by diltiazem
inhibits the in vitro stimulation by IGF1 and insulin on SMC proliferation and
migration. We conclude that the acceleration of ageing is related to the decline of IGF1
in such a manner that ageing rates progress as GH secretion diminishes. Biomarkers are
a¡ected correspondingly. The role of insulin in atherogenesis is related to
hyperinsulinaemia, but the increase in insulin secretion belongs to the process of ageing
regulation. Nevertheless, the e¡ect of insulin in changing the phenotype of SMCs is
atherogenic. Diltiazem may therefore act as an antiatherogenic agent. In advanced age
the risk of atherogenesis decreases because of lowered propensity of SMCs to proliferate
and migrate, which is probably due to a greater proportion of senescent cells.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 143^160

After reaching a maximum level at the end of the puberty, insulin-like growth
factor 1 (IGF1) blood levels decline, indicating that the genetic programme of
growth and di¡erentiation has ¢nished and adulthood has begun. The involution
of the organism appears simultaneously with a decrease of IGF1 levels and growth
hormone (GH) secretion (Iranmanesh & Veldhuis 1992, Vahl et al 1996). From a
gerontological point of view, it is not possible to distinguish the signs of the
143
144 RUIZ-TORRES

involutive process and those of ageing. For this reason, the continuous decline of
IGF1 secretion may be considered a marker of biological ageing.

IGF1 and insulin secretion during ageing

The secretion of GH decreases with advancing age, mainly through a decrease in
the amplitude of GH pulses (Veldhuis et al 1995). Mean 24 h GH concentrations
decline from late puberty into old age (Rudman et al 1981) due to decreased GH
production and an increased clearance rate (Iranmanesh et al 1991). The changes in
IGF1 levels throughout life appear to mimic those of GH (Ho et al 1987).
Accordingly, in a cross-sectional study of healthy adults with similar lean body
mass (LBM), the IGF1 blood levels decreased with population age (Ruiz-Torres &
Corpas 1993), independently of adiposity (Copeland et al 1990). We observed that
the slope of the corresponding curve shows di¡erences dependent on the age range
considered. In the range between 20 and 50 years the slope is clearly steeper than
when all age groups up to a very advanced age are included. This indicates that
people with low levels of IGF1 die earlier so that the average value in the
surviving group is increased. Consequently, the curve obtained up to around 50
years age (and then extrapolated) represents a more realistic behaviour of IGF1 and
should be considered as a standard reference curve.
These results could help explain the role of IGF1 levels on age-related
di¡erences in, for example, testosterone, thyroid hormone and procollagen III
peptide blood concentrations that disappear when comparing young and old
persons with similar IGF1 values (Table 1). Furthermore, individuals of very
advanced age show IGF1 blood levels clearly above the standard regression
curve, the behaviour of which is not a¡ected by the selection of individuals by
mortality (Soares de Melo 1997).
In a similar de¢ned population of healthy individuals (as mentioned earlier)
insulin levels are usually increased with advancing age (Ruiz-Torres et al 1996).

TABLE 1 Di¡erences between younga and oldb depending on IGF1 blood level in
males

IGF1 Testosterone PIIIP LBM

Young versus old Decreased Decreased Increased Decreased

P50.0001 P50.0001 P50.0001 P50.0001
Lowest range in young versus NS NS Increased NS
highest in old P50.01
of IGF1
a
20^39 years age, n = 22.
b
70^92 years age, n = 33.
LBM, lean body mass; NS, di¡erence not signi¢cant; PIIIP, procollagen III peptide.
AGEING AND ATHEROSCLEROSIS 145

FIG. 1. IGF1 serum levels and insulin secretion of young (20^39 years old, n = 22) and old
(70^92, n = 33) healthy men with corresponding anthropometrical manifestations. On the right
are blood concentrations of the N-terminal peptide of procollagen type III (PIIIP). The ¢gure
shows the opposite age-dependent behaviour of the hormones mentioned. IGF1
concentrations were determined after alcohol extraction by radioimmunoassay and PIIIP.
Daily insulin secretion was by means of 24 h C-peptide excretion, corrections and
normalization of results as described (Ruiz-Torres et al 1996); LBM calculated according to
Forbes & Bruining (1976); and adipose mass worked out on the basis of skin fold thickness
and body density according to Durnin & Womersley (1974).

This elevation of insulin is related to the increased body fat mass and reduced
muscle mass, the latter primarily due to the progressive decrease of GH/IGF1
secretion (Fig. 1). Moreover, the reverse correlation between insulin secretion
and IGF1 blood levels could be understood as a wear and tear e¡ect.
Nevertheless, it needs to be stressed that all regulatory processes have some side
e¡ects, in this case those concerning excessive amounts of insulin with or without
hyperinsulinaemia.

Atherogenity of insulin
Clinical studies have demonstrated that those processes linked to
hyperinsulinaemia, such as type 2 diabetes or obesity, show a higher mortality
due to coronary or cerebral atherosclerosis (Py˛r
l
et al 1985). Furthermore,
146 RUIZ-TORRES

experimental results show that insulin acts on the vascular wall, either producing
hypertension and endothelial changes, or in£uencing the smooth muscle cells
(SMCs) to proliferate. It is well known that endothelial lesions and SMC
proliferation are basic steps of atherogenesis (Ross 1993).
For a better understanding of the role of SMCs in atherosclerosis, it is worth
mentioning that these cells and collagen represent the main content of the
atheroma plaque. SMCs migrate from the media crossing the intima to
accumulate and release collagen. Two distinctive phenotypes of SMC are known:
contractile and synthetic. Contractile SMCs respond to agents inducing vasomotor
changes, whereas the synthetic SMCs are capable of expressing genes for growth
regulators and collagen synthesis. Normally, in adult life the vascular wall has only
postmitotic SMCs which are contractile. Therefore, muscle cells of atheromas
should have changed their di¡erentiated phenotype to a synthetic one, with the
capability to proliferate, migrate and ¢nally secrete collagen. At present, it is
believed that oxidized low density lipoprotein (LDL) alters the endothelium,
producing a cascade of events including SMC migration (for review see, Massy
& Keane 1996). The question is whether insulin is able to change the SMC
phenotype and, if so, by what mechanism.

E¡ects of insulin and IGF1 on SMCs

More than a decade ago di¡erent publications showed that insulin stimulates SMC
proliferation in vitro (Stout 1990). We have observed that in non-cultured cells,
SMCs directly taken from the human artery, insulin stimulates collagen secretion.
This e¡ect was probably produced by activation of the IGF1 receptors, because
addition of insulin receptor-blocking antibodies did not show any inhibition. On
the contrary, antibodies blocking IGF1 receptors inhibited the insulin-induced
collagen secretion. We concluded that insulin is able to change the SMC
phenotype by acting as a growth factor (Ruiz-Torres et al 1998).
Moreover, insulin stimulates the chemotaxis of SMCs directly dispersed from
the human artery (Mu•oz et al 1998). In these experiments this migration could
be inhibited by insulin receptor-blocking antibodies, so we assumed that insulin
was acting here through its speci¢c receptors. Nevertheless, these results point out
a very close relationship between the stimulating e¡ect of insulin on both collagen
secretion and migration, in spite of apparently acting through di¡erent types of
receptor. Further experiments were required to clarify how insulin stimulates
SMC migration.

The e¡ect of insulin and IGF1 on the

cytoskeleton in relation to SMC migration
The cytoskeleton is a complex network of protein ¢laments extending throughout
the cytoplasm that is involved in a range of cell processes. Of the three main
AGEING AND ATHEROSCLEROSIS 147

FIG. 2. Insulin-induced F-actin reorganization near the membrane of human vascular smooth
muscle cells showing a ru¥ing which does not appear when Ca2+ channels of these cells are
blocked by diltiazem. The results are similar in the case of IGF1 (unpublished).

cytoskeletal types, the actin ¢laments are primarily responsible for many cell
movements, for example for SMC migration (Alberts et al 1994, p 787^803).
According to Bretscher’s model of ¢broblast locomotion, actin ¢laments
depolymerize ahead of the nucleus, generating actin subunits which di¡use to the
cell’s front where actin ¢laments polymerize at the leading edge (Bretscher 1996).
Chemoattractant receptors contribute to the promotion of actin assembly at the
leading cell edge (Stossel 1993). Ca2+ ions play an important role in the process of
assembly and deassembly of actin ¢laments. Insulin induces these speci¢c
rearrangements in mesangial cells in vitro. Fluorescent staining for F-actin with
phalloidin normally shows actin ¢laments spanning the entire cell in all
directions. After insulin treatment, the cells show peripheral ru¥ing and lifting
o¡ from the substrate as attachments are released (Ber¢eld et al 1996). We have
shown that insulin, like IGF1, induces membrane ru¥ing of F-actin in human
vascular SMCs in vitro. This e¡ect appears very quickly, just 5 minutes after
hormone application. As expected, the blockade of Ca2+ channels by diltiazem
does not allow insulin to induce ru¥ing (Fig. 2). The result is similar for IGF1.
148 RUIZ-TORRES

These data suggest that insulin in£uences SMC movements with a mechanism
probably similar to that of IGF1.

IGF1-like mechanism of insulin to

induce proliferation but not migration of SMCs
A large number of well known factors modulate SMC migration, such as platelet-
derived growth factor (PDGF) and IGF1. Most have similar proliferation e¡ects.
It is therefore possible that both the migratory and proliferative responses have a
common intracellular mechanism, but the results obtained postulate some
di¡erences (Schachter 1997). The concentration of IGF1 required to induce
migration is substantially lower than that needed for proliferation (Bornfeldt et al
1994). Regarding the e¡ect of Ca2+ channels blockade on migration, diltiazem
inhibits IGF1 and insulin stimulation in a similar manner (Fig. 3).
Insulin stimulates SMCs to proliferate similarly to IGF1. Our recent
unpublished results obtained from in vitro experiments with cultured human
arterial SMCs show that the mitotic induction of insulin occurs by occupation of
the IGF1 receptors. This has been demonstrated using receptor-blocking
antibodies. The blockade of insulin receptors does not inhibit the insulin-induced
induction of SMC proliferation, but the addition of IGF1 receptor blocking
antibodies does. The opposite results have been obtained in chemotaxis
experiments (Fig. 4). The addition of IGF1 receptor-blocking antibody has no
e¡ect on migration, whereas insulin receptor-blocking antibody does.
As expected, both the proliferation and migration capability of SMCs may be
inhibited by blockade of Ca2+ channels as with diltiazem. These results support
the possibility of preventing atherosclerosis by treating with Ca2+ antagonists
(Marche et al 1997).
The results above are in agreement with those  also mentioned in this
chapter  on the insulin mechanism stimulating collagen secretion. On this basis
and considering the induction of F-actin ru¥ing, we therefore postulate that there
are no di¡erences between IGF1 and insulin with respect to their e¡ect in changing
the phenotype of SMCs. Extrapolating these results to in vivo, insulin may be
behaving as a growth factor rather than a metabolic e¡ector because of its
presence in higher than normal concentrations. From this point of view, whether
insulin is atherogenic or not would depend primarily on its concentration.

Atherogenesis and ageing

The culture of SMCs is a useful model system because under these conditions the
cells show many of the functional changes characteristic of atherogenesis, in
particular the ability to proliferate and to secrete collagen (Campbell & Campbell
AGEING AND ATHEROSCLEROSIS 149

FIG. 3. Relationship between hormone-induced migration and corresponding inhibition due

to Ca2+ channel blockade by diltiazem in vascular smooth muscle cells from rat aorta and human
femoral artery. The inhibition decreases directly proportionally to stimulation level either by
IGF1 or insulin (bottom). Unpublished experiments performed in the Boyden chamber.

1987, Massy & Keane 1996). The in vitro proliferating SMCs develop cytoskeletal
features similar to those observed in fetal and pathological states (Skalli et al 1986).
Essentially, SMCs exhibit two phenotypes in culture: one able to proliferate and
the other to di¡erentiate, recalling the in vivo fetal and contractile phenotypes,
respectively. The terminal di¡erentiation in culture resembles the behaviour
observed in other cells such as melanocytes (Medrano et al 1994). Proliferating
SMCs in vivo assume fetal phenotypic features which are also observed in cultured
cells (Desmoulie' re & Gabbiani 1992).
As the proliferation of human vascular SMCs characterizes the atherogenesis, it
seems of interest to investigate the mitotic activity in cultured SMCs dependent on
150 RUIZ-TORRES

FIG. 4. Inhibition of insulin stimulation of proliferation, migration and collagen secretion in

human vascular smooth muscle cells (hvSMCs) due to receptor blocking antibodies, indicating
the mechanism of action. The addition of IGF1 receptor blocking antibody inhibits the
proliferative e¡ect of insulin, whereas the opposite occurs in the case of migration. Similarly to
proliferation, the insulin receptor blocking antibody does not show any e¡ect on collagen
secretion in spite of its inhibition of glucose uptake. White columns: incubations of hvSMCs
with insulin and speci¢c receptor blocking antibody (IRAb). Hatched columns: those with
IGF1-receptor blocking antibody (IGF1RAb). *Dispersed cells from artery; the others from
the culture (2^4 passage). ***P50.0001; NS, not statistically signi¢cant (results partially
published in Ruiz-Torres et al 1998, Mu•oz et al 1998).

the age of the donor. The proliferation rate of SMCs decreases in the culture as
donor age advances (Ruiz-Torres et al 1999). The decline reaches the zero point
that is the total loss of proliferative activity, at age over 100 years, near the limit
of maximum life potential for human beings. Therefore age is a factor which
decreases the ability of vascular SMCs to proliferate. The meaning is the same for
in vivo as in vitro. Nevertheless, cultured cells age according the passage numbers as
described by Hay£ick (Smith & Hay£ick 1974).
The cellular expression of ageing in vivo or in vitro is the progressive appearance
of senescent cells, so that cells have a ¢nite life potential (Hay£ick 1987). Senescent
cells lose their capacity to proliferate despite potentially remaining in the tissue or
culture for a relatively long time (Matsumura et al 1979). A similar phenomenon
occurs regarding migration capacity. We recently found that the basal migration of
human vascular SMCs decreases as donor’s age advances. The slope of the curve is
similar to that of proliferation, showing the total loss near the age of maximum life
potential (Ruiz-Torres et al 1999). Consequently, the interpretation is the same in
AGEING AND ATHEROSCLEROSIS 151

TABLE 2 Decrease of migration by age: chemotaxis (basal and

induced) of cells from young and old male donors

Age (years) 43 74

Number of experiments 3 3
Migrated cells:
Basal 930 36 540 43*
1 nM insulin 1110 165 740 106*/**
1 nM IGF1 1153 117 710 10*/***

Results using the Boyden chamber. *young:old P50.05; **with basal young
P50.07; ***with basal young P50.05.

both cases: during adulthood, increasing age decreases the sensitivity of SMCs to
phenotypic change (Table 2). The high content of senescent cells in advanced age
would remarkably diminish the risk for atherogenesis. The opposite would happen
when relatively young SMCs coexist with risk factors such as hyperchol-
esterolaemia and hyperinsulinaemia. This state corresponds to the age range within
the 3rd and 5th decades according to our studies and in agreement with
epidemiological results on cardiovascular degenerative diseases.

Acknowledgements
This work was made possible by a grant of the Fondo de Investigacio¤n Sanitaria (FIS), Spanish
Ministry of Health, Nr 98/0177. The unpublished results described here were obtained with the
collaboration of Ra¡aele Carraro, Rosario Lozano, Manuel Mac|¤ a and Marcia Soares de Melo.

References
Alberts B, Bray D, Lewis J, Ra¡ M, Roberts R, Watson JD 1994 Molecular biology of the cell.
Garland, London
Ber¢eld AK, Raugi GJ, Abrass CK 1996 Insulin induces rapid and speci¢c rearrangement of the
cytoskeleton of rat mesangial cells in vitro. J Histochem Cytochem 44:91^101
Bornfeldt KE, Raines EW, Nakano T, Graves LM, Krebs EG, Ross R 1994 Insulin-like growth
factor I and platelet-derived growth factor-BB induce directed migration of human arterial
smooth muscle cells via signaling pathways that are distinct from those of proliferation. J
Clin Invest 93:1266^1274
Bretscher MS 1996 Getting membrane £ow and the cytoskeleton to cooperate in moving cells.
Cell 87:601^606
Campbell GR, Campbell JH 1987 Smooth muscle cells. In: Olson AG (ed) Atherosclerosis and
clinical science. Churchill Livingstone, New York, p 105^115
Copeland KC, Colletti RB, Devlin JT, McAulife TL 1990 The relationship between insulin-like
growth factor-1, adiposity and aging. Metabolism 39:584^587
Desmoulie' re A, Gabbiani G 1992 The cytoskeleton of arterial smooth muscle cells during
human and experimental atheromatosis. Kidney Int Suppl 37:S87^S89
152 RUIZ-TORRES

Durnin JV, Womersley J 1974 Body fat assessed from total body density and its estimation from
skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr
32:77^97
Forbes GB, Bruining GJ 1976 Urinary creatinine excretion and lean body mass. Am J Clin Nutr
29:1359^1366
Hay£ick L 1987 Origins of longevity. In: Warner HR, Butler RN, Sprott RL, Schneider EL (eds)
Modern biological theories of aging. Raven Press, New York, p 21^34
Ho KY, Evans WS, Blizzard RM et al 1987 E¡ects of sex and age on the 24-hour pro¢le of
growth hormone secretion in man: importance of endogenous estradiol concentrations. J
Clin Endocrinol Metab 64:51^58
Iranmanesh A, Lizarralde G, Veldhuis JD 1991 Age and relative adiposity are speci¢c
negative determinants of the frequency and amplitude of growth hormone (GH) secretory
bursts and the half-life of endogenous GH in healthy men. J Clin Endocrinol Metab
73:1081^1088
Iranmanesh A, Veldhuis JD 1992 Clinical pathophysiology of the somatotropic (GH) axis in
adults. Endocrinol Metab Clin North Am 21:783^816
Marche P, He¤rembert TH, Zhu DL 1997 Pharmacologic treatment of atherosclerosis: beyond
lipid-lowering therapy. Int J Cardiol 62:S17^S22
Massy ZA, Keane WF 1996 Pathogenesis of atherosclerosis. Semin Nephrol 16:12^20
Matsumura T, Zerrudo L, Hay£ick L 1979 Senescent human diploid cells in culture: survival,
DNA synthesis and morphology. J Gerontol 34:328^334
Medrano EE, Yang F, Boissy R et al 1994 Terminal di¡erentiation and senescence in human
melanocyte: repression of tyrosine-phosphorylation of the extracellular signal-regulated
kinase 2 selectively de¢nes in two phenotypes. Mol Biol Cell 5:497^509
Mu•oz FJ, Jime¤ nez M, Melo¤n J, Ruiz-Torres A 1998 Phenotypic changes in vascular smooth
muscle cells during aging: insulin e¡ect on migration. Gerontology 44:149^152
Py˛r
l
KE, Savolainen E, Kaukolas S, Haapakoski J 1985 Plasma insulin as coronary heart
disease risk factor: relationship to other factors and predictive value during 9 and a half-
years follow-up of the Helsinki Policeman Study population. Acta Med Scand Suppl
701:38^52
Ross R 1993 The pathogenesis of atherosclerosis: a perspective for the 1990s. Nature 362:
801^809
Rudman D, Kutner MH, Rogers CM, Lubin MF, Flemming GA, Bain RF 1981 Impaired
growth hormone secretion in the adult population: relation to age and adiposity. J Clin
Invest 67:1361^1369
Ruiz-Torres A, Corpas E 1993 Anthropometrical changes in aging: Their relationship with GH
secretion and its possible metabolic consequences. In: Laron Z, Butenandt O (eds) Growth
hormone replacement therapy in adults: pros and cons. Freund Publishing House, London,
p 123^136
Ruiz-Torres A, Melo¤n J, Mun˛z FJ 1998 Insulin stimulates collagen synthesis in vascular
smooth muscle cells from elderly patients. Gerontology 44:144^148
Ruiz-Torres A, Vicent D, Sanchez de Paco G, Munoz FJ, Gimeno A, Carraro R 1996 Increase in
insulin secretion with age: its clinical importance in evaluating abnormal secretions focused on
diabetes type II and obesity. Arch Gerontol Geriatr 22:39^47
Ruiz-Torres A, Gimeno A, Melo¤n J, Mendez L, Mun˛z FJ, Mac|¤ a M 1999 Age-related loss of
proliferative activity of human vascular smooth muscle cells in culture. Mech Ageing Dev
110:49^55
Schachter M 1997 Vascular smooth muscle cell migration, atherosclerosis, and calcium channel
blockers. Int J Cardiol 62:S85^S90
AGEING AND ATHEROSCLEROSIS 153

Skalli O, Bloom WS, Ropraz p, Azzarone B, Gabbiani G 1986 Cytoskeletal remodeling of rat
aortic smooth muscle cells in vitro: relationships to culture conditions and analogies to in vivo
situations. J Submicrosc Cytol 18:481^493
Smith JR, Hay£ick L 1974 Variation in the life-span of clones derived from human diploid cell
strains. J Cell Biol 62:48^53
Soares de Melo M 1997 Concentracio¤n serica de IGF-1 durante la vida adulta. Su relacio¤n con el
estado de conservacio¤n biologica y capacidad de supervivencia. PhD thesis. Universidad
Autonoma de Madrid, Spain
Stossel TP 1993 On the crawling of animal cells. Science 260:1086^1094
Stout RW 1990 Insulin and atheroma, 20-yr perspective. Diabetes Care 13:631^654
Vahl N, Jorgensen JOL, Jurik AG, Christiansen JS 1996 Abdominal adiposity and physical
¢tness are major determinants of the age associated decline in stimulated GH secretion in
healthy adults. J Clin Endocr Metab 81:2209^2215
Veldhuis JD, Liem AY, South S et al 1995 Di¡erential impact of age, sex steroid hormones, and
obesity on basal versus pulsatile growth hormones secretion in men as assessed in an
ultrasensitive chemiluminiscence assay. J Clin Endocrinol Metab 80:3209^3222

DISCUSSION
Veldhuis: One of the issues I am perceiving is that the distinction is sometimes
blurred between what we implicitly view as the reversible facets of ageing and
those that aren’t. Your extrapolation of the IGF1 levels down to about age 110
was exciting. I guess for testosterone we’d live to be about 250, so we will run
out of other things before testosterone.
Handlesman: Tangential to this, there is a great deal of re-thinking going on in
atherogenesis, particularly as people are starting to recognize that a lot of the
accepted explanations about oestrogens preventing cardiovascular disease are
based almost solely on retrospective case control data. In collaboration with
some of our cardiology colleagues, we have done a series of studies looking at
elements of the process of atherogenesis. In particular, we have studied adhesion
and migration of macrophages to human endothelial cells. It really looks as if
testosterone itself has signi¢cant e¡ects, increasing adhesion by a VCAM
mechanism, including NF-kB and Ca2+. This is very similar to what you are
describing. It is somewhat surprising that the insulin e¡ects seem to be inhibiting
this when the epidemiology suggests that insulin levels rise when atherogenesis
increases. Perhaps part of the picture is that sex steroids, and particularly
androgens, have a role that is so far not fully recognized.
Veldhuis: In part, this illustrates the complexity of these system interactions. One
has conditions in the prostate where testosterone will induce binding proteins for
IGF1 and conversely GH alters testosterone’s conversion to dihydrotestosterone
in the liver. The gemisch of these changes across axes is going to be tough to sort
out, with the incremental changes within each axis being so small. I am struck by
the enormous range of similarity of mild glucocorticoid excess to ageing and
syndrome X. Working in the GH ¢eld I’ve always attributed this to GH
154 DISCUSSION

de¢ciency. Here we have two systems where arguing for a slight increase in corticol
net activity (associated with the stress in daily living), and a decline in GH. Each
directional change produces a similar outcome. Altogether, I’m surprised we’re
still alive here!
Prior: I would like to understand the interrelationships between insulin/IGF1
and body weight, and in particular changes with endurance or aerobic exercise.
There is some key related to exercise that will enlighten us about the rise in
insulin and decrease in IGF with ageing.
Elahi: I don’t know much about IGF1 with respect to body composition, but it
is well established that hyperinsulinaemia not only causes overall increased
adiposity, but also deposition of subcutaneous fat in the abdomen which may
consist of two metabolically active types.
Veldhuis: We found that ¢t young adults tend to undersecrete insulin,
presumably because of good insulin sensitivity in the periphery (Engdahl et al
1995). They have well-organized low amplitude insulin pulses that are su⁄cient
to maintain euglycaemia, whereas older people show some clear disruption of
several features of insulin release (Meneilly et al 1997, 1999).
Monson: Cortisol metabolism and GH are not necessarily distinct from each
other. Adrenocorticotropic hormone (ACTH) feedback controls normal
circulating free cortisol. However, at a tissue level, increased 11b-HSD1 activity
in the adipocyte consequent upon GH de¢ciency will shift the set point in favour of
increased local exposure to cortisol in both the liver and the adipocyte. It isn’t
surprising, therefore, that there may be phenotypic similarities. I am not saying
that all the e¡ects of GH on fat accumulation are mediated through cortisol, but
there is a potential link there. The other important hormonal modulator of 11b-
HSD1 is insulin.
Carroll: To add to that, there is yet another way that the two interact. Under
stress conditions, and certainly in human depression, GH secretion at night is
practically eliminated.
Veldhuis: There is a cortisol^GH connection at a couple of interesting
neuroendocrine levels, as well (Giustina & Veldhuis 1998, Giustina et al
1994).
Giustina: A slight cortisol excess may decrease GH secretion via an increase in
hypothalamic somatostatin. We have good evidence both in normal volunteers
and in patients with autoimmune diseases that you can get a decrease in GH
secretion after short- and long-term glucocorticoid administration. In addition,
GH inhibition is observed in patients with Cushing’s disease.
Veldhuis: This is where I betray my own intuition that with ageing being such a
minimally incremental, subtle process, these interactions between systems,
particularly when they have common adverse e¡ects on certain target cells, are
probably especially important to understand. We may be wrong in thinking of
AGEING AND ATHEROSCLEROSIS 155

ageing as a monohormonal problem. We do this only because we work more easily

in one hormone system at a time.
Handelsman: Is there evidence that mild hypercortisolism actually accelerates
atherogenesis?
Veldhuis: Can the Rotterdam group help us with this? Their data are more
correlated with decreased bone mass and decreased quality of life with high
cortisol, than death from atherosclerosis.
van den Beld: I investigated the relationship between cortisol and intima medial
thickness in the distensibility of the carotid artery, and couldn’t ¢nd any correlation
between this and low insulin/cortisol.
Carroll: In relation to that, there are many more ways to die than just by
atherosclerosis. Atherosclerosis is a major cause of premature mortality. We are
constantly riding the boundary between gerontology and geriatrics here.
Handelsman: In a population, the gender disparity in life expectancy is about 5^7
years. This is precisely the di¡erence in the peak and the rates of cardiovascular
deaths in the community. It isn’t a¡ecting just premature deaths, it is the overall
pattern of death. This gender disparity remains less well understood than it was
once thought.
Laron: One factor in atherosclerosis that hasn’t been mentioned is lipo-
protein (a), which is regulated by IGF1 and insulin. IGF1 suppresses it; GH
elevates it.
How do you explain the e¡ect on procollagen III? We know that IGF raises it. Is
it a competitive e¡ect with the insulin receptors?
Ruiz-Torres: You are right. The blood level curve of procollagen III peptide
shows a biphasic behaviour. During young adulthood, there is a decrease from a
maximum reached in the ¢rst phase of life. Then after age 30 there is another rise, in
spite of continuous diminution of IGF1. This dissociation is probably due to
insulin, whose secretion increases with advancing adult age. As shown, insulin
acts in a similar way to IGF1. Furthermore, the more insulin released, the higher
the level of procollagen III, as we have seen in hyperinsulinaemic type II diabetes
patients.
Laron: Procollagen III is also related to the collagen in the elastic ¢bres. We
know that there is a reduction in elastic ¢bres in IGF1 de¢ciency, but this does
not ¢t the levels that you showed.
Ruiz-Torres: The procollagen III levels we see in the second phase of adulthood
are mainly related to the accumulation of collagen in the arterial wall. The best
example is myocardial infarction patients with high procollagen III peptide
blood levels.
Giustina: I would like to make a general point about IGF1 and ageing. One
thing we haven’t discussed so far, but which is important, is the concept of
the frail elderly. These people are di¡erent from the normal elderly subjects.
156 DISCUSSION

Endocrinologists could help understand this. It has been suggested that IGF1
could be a marker of the frail elderly.
Ruiz-Torres: My intention was to explain the role of insulin in atherogenesis. I
have presented results which would lead to the conclusion that the way by which
insulin induces smooth muscle cells to proliferate is through IGF1 receptors. On
the other hand, I have also presented results that show that IGF1 serum levels are
related to well being in advanced age. These data indicate that not only muscle
mass but also other ageing parameters are less in£uenced when IGF1 levels are
relatively high.
Giustina: Is IGF1 acting on the endothelium or on the smooth muscle
cells?
Ruiz-Torres: We haven’t studied this point speci¢cally, but I know that both
endothelium and smooth muscle cells are sensitive to IGF1 as well as insulin.
From a gerontological point of view, the latter seems to be the more important
atherogenic e¡ector because it is increased in ageing, as opposed to IGF1 which
decreases. Regarding the question of whether atherogenesis is a normal
manifestation of ageing or a disease, I would like to underline how di⁄cult it is
to ¢nd a correct answer. My personal opinion is that atherosclerosis is a ‘side-
e¡ect’ of hormonal changes linked to the regulatory process of ageing.
Giustina: The frail elderly are the men or women who are likely to be prone to
immobility and diseases.
Ruiz-Torres: Compared with young individuals you would say that they are ill,
but considering a normal follow-up of ageing the state of these individuals would
only express the ¢nal phase of life.
Giustina: Clinically speaking there are scales that can be used to quantify the
frailty of the population.
Morley: The problem with frailty is that everybody knows what it is, but
there is no good de¢nition. The Baltimore group have a de¢nition that looks
pre-dominantly at muscle strength. We have to be careful about these de¢nitions,
because each group has gone out and de¢ned something. If I see a patient I know
if they are frail, but when stick to the de¢nitions they don’t necessarily categorize
the people we would say are frail into those de¢nitions. The problem with
IGF1 as a marker of frailty is that it is almost certainly a marker because it is a
marker of malnutrition. Frail people are almost always somewhat malnourished,
so this is where we run into trouble with it as a marker. There is a nice study
coming out on centenarians showing that the subjects with really low IGF1s are
those with low albumins and tend to be malnourished (Arai et al 2001). Again, you
are looking at the problem of interactions, and nutrition becomes a major
hormonal interactor for IGF1. This is where we are running into trouble. None
of this is easy: we desperately need a good de¢nition for frailty and also for
sarcopenia.
AGEING AND ATHEROSCLEROSIS 157

Riggs: We have de¢ned sarcopenia on the basis of established values in young

people, rather than on a functional basis.
Handelsman: As an outsider reading the gerontology literature I see the issue of
malnutrition raised all the time. I wonder how much of it is pushed by the for-pro¢t
US nursing homes. Is this the real issue? Is this true for non-institutionalized people
out in the community?
Morley: It is a huge problem. In my practice I see a lot of malnourished old
people. Old people run into trouble when they get sick, and many of them
become malnourished. 10^20% of the really frail old people are malnourished. As
they get more frail they get more malnourished. Depression is a big factor in
inducing malnutrition.
Laron: I would like to clarify the relationship between malnutrition and IGF1.
Malnutrition results in secondary GH resistance and low IGF1. I showed that
IGF1 de¢ciency does not cause malnutrition; on the contrary, it causes obesity,
but with reduced lean body mass. The frailty and lack of appetite in old people
must have another aetiology than IGF1 de¢ciency, as must their lack of thirst.
Morley: There is a large literature on the factors involved in causing anorexia in
ageing, which has nothing to do with IGF1. It is just one of the markers of
malnutrition, and it is one that may be less a¡ected by cytokines, but I still
haven’t seen enough data to convince me that it is not cytokine dependent.
Veldhuis: Malnutrition in ageing is a fascinating issue. It is a complex cascade
spanning factors from depression to socioeconomic status. I agree that it looks
like a mild GH resistance state. In a fasting animal, GH receptor signalling is
down-regulated (Giustina & Veldhuis 1998).
Bj˛rntorp: With regard to cortisol as a risk factor for atherosclerosis, as far as I can
see there are no prospective studies on this. This is probably due to the fact that this
whole ¢eld has been so focused on details of lipid fractions, hypertension, smoking
and insulin. On the other hand, an increased mass of visceral fat, which might be
considered as a surrogate measurement of chronic elevation of cortisol secretion, is
a powerful risk factor for atherosclerotic manifestations.
Veldhuis: This is what I mean when I talk about these devilish interrelationships.
Carroll: If you are going to do studies on the cortisol relationship, you want to be
looking at the night levels of cortisol, not the daytime measures. This is the
approach the MacArthur Foundation is taking with their epidemiological
measures of allostatic load. They are measuring overnight urinary free cortisol
excretion among other things.
Bj˛rntorp: Are they doing prospective studies?
Carroll: Yes.
Brabant: I wanted to come back to the question of the endothelium. What is NO
doing in your concept? NO de¢ciency is a marker of atherosclerosis in animal
models, where there is a decrease of NO along with atherosclerosis. On the other
158 DISCUSSION

hand, NO is closely controlled by the GH^IGF1 system. How does this ¢t into
your concept?
Ruiz-Torres: Our studies have been focused on ¢nding a relationship between
the hormonal state in ageing and the phenotypic change of vascular smooth muscle
cells representing the ¢rst step in atherogenesis. This change means that a
contractile, non-mitotic cell becomes able to proliferate, migrate and produce
collagen. It is known that NO is anti-atherogenic, but especially because of its
interaction with angiotensin II. Additional to this, NO is likely to help stabilize
the postmitotic state of smooth muscle cells, but its role here is not clear. IGF1
induces NO production in many cells, but it also inhibits NO production in
smooth muscle cells after interleukin 1b. In any case, we have to bear in mind
that in ageing both GH and IGF1 secretion are remarkably reduced.
Prior: I don’t know about the IGF1^insulin relationship to NO, but we recently
showed in an intra-arterial cross-over randomized study that oestrogen and
progesterone have similar NO-mediated e¡ects to increase forearm blood £ow
(Mather et al 2000).
Veldhuis: I think that there is prompt vasoarterial dilation in the human with
IGF infusion. It is thought to operate through NO. GH itself increases NO
products in the urine (Giustina & Veldhuis 1998). There are some exciting linkages.
Handelsman: Using the tracheal ultrasound technique to measure £ow-mediated
dilatation, it is possible to show that castration in men increases vascular activity as
do oestrogens. In transexuals and normal physiological young men, androgens
decrease vascular reactivity but also increase vessel diameter. There are quite a
number of direct sex steroid relationships with NO. There were placebo-
controlled studies in the 1950s and 1960s of cardiac ischaemia clearly showing
that testosterone is a vasodilator under some circumstances.
Morley: There is a study from Newcastle showing that angina can be treated with
testosterone (English et al 2000). The problem with NO is that it may actually be
bad to vasodilate. If there is an unstable plaque, vasodilation will increase shear
force which could result in popping the clot. It is one of those questions that
hasn’t been answered: most deaths from atherosclerosis involve the rupture of a
clot, not blocked arteries. This may be one of the reasons that the HERS study
showed early death (Hulley et al 1998).
Giustina: As to whether the dilatation e¡ect is good or bad, this probably
depends on the cardiovascular situation at the outset. If a patient has the mild
degree of heart failure that is often observed in older people, the vasodilating
e¡ect will probably be bene¢cial. Studies show that if GH is infused acutely or if
it is given on a subchronic basis, cardiovascular function is improved.
Morley: We are dependent on waiting for the MRI techniques to get good
enough to pick up the unstable plaques. When we can do this, we will have much
more insight as to who to give vasodilators to and who not to.
AGEING AND ATHEROSCLEROSIS 159

Mˇller: The involvement of the NO system in neuroendocrine control is a well

known phenomenon. For instance, the ability of growth hormone (GH)-releasing
peptides (GHRPs) to increase GH secretion in both old and young dogs is
strikingly increased by NO donors, and counteracted by NOS inhibitors
(Rigamonti et al 1999). GH secretagogues, in addition to their endocrine actions,
exert other important extra-endocrine actions, and they too appear to require NO
involvement. Some GHRPs stimulate sexual function in rats following
intracerebral administration, and NOS inhibitors prevent this action (Melis et al
2001); in addition, the NO system would be involved in the prevention of injury
and dysfunction of the vascular endothelium e¡ected by GHRPs in a rat heart
preparation undergoing low-£ow ischaemia and reperfusion (De Gennaro
Colonna et al 1997). Interestingly, in hearts from aged rats GHRP fail to interfere
positively with the endothelium-dependent relaxant mechanism, suggesting the
existence of an impaired NO function (Rossoni et al 1998).
Morley: We were discussing anorexia in ageing. To complete the cycle, our data
suggest fairly strongly that anorexia in ageing is related to NO malfunction both
within the hypothalamus and also in the fundus of the stomach. Many older people
cannot open up their fundus (the top part of their stomach), so they get early
satiation because the food just goes straight into the antrum where dilation
produces rapid feedback, resulting in anorexia. Our animal studies suggest NO is
involved, and the human studies agree.
Veldhuis: Is any more known about NO induction by sex steroids?
Morley: The simple answer is that testosterone is permissive for the development
of NO synthase. If you have no testosterone at all, you don’t get NO synthase. If
you have some, you tend to get NO synthase. We have a paper in press looking at
Viagra (sildena¢l) failures in people who are hypogonadal. If you also treat them
with testosterone, Viagra will work. I think oestrogen is similar.
Wang: Nestu Cadavid’s group has shown that NOS in the penis is responsive to
androgens (Garban et al 1995).

References
Arai Y, Hirose N, Yamamura K et al 2001 Serum insulin-like growth factor 1 in centenarians:
implications of IGF1 as a rapid turnover protein. J Gerontol 56:M79^M82
De Gennaro Colonna V, Rossoni G, Bernareggi M, Mˇller EE, Berti F 1997 Cardiac ischemia
and impairment of vascular endothelium function in hearts from growth hormone-de¢cient
rats: protection by hexarelin. Eur J Pharmacol 334:201^207
Engdahl J, Veldhuis JD, Farrell PA 1995 Altered pulsatile insulin secretion associated with
endurance training. J Appl Physiol 79:1977^1985
English KM, Steeds RP, Jones TH, Diver MJ, Channer KS 2000 Low-dose transdermal
testosterone therapy improves angina threshold in men with chronic stable angina: a
randomized, double-blind, placebo-controlled study. Circulation 102:1906^1911
160 DISCUSSION

Garban H, Marquez D, Cai L, Rajfer J, Gonzalez-Cadavid NF 1995 Restoration of the normal

adult penile erectile response in aged rats by long-term treatment with androgens. Biol Reprod
53:1365^1372
Giustina A, Veldhuis JD 1998 Pathophysiology of the neuroregulation of GH secretion in
experimental animals and the human. Endocr Rev 19:717^797
Giustina A, Bresciani E, Bossoni S et al 1994 Reciprocal relationship between the level of
circulating cortisol and growth hormone secretion in response to growth hormone-
releasing hormone in man: studies in patients with adrenal insu⁄ciency. J Clin Endocrinol
Metab 79:1266^1272
Hulley S, Grady D, Bush T et al 1998 Randomized trial of estrogen plus progestin for secondary
prevention of coronary heart disease in postmenopausal women. Heart and estrogen/
progestin replacement study (HERS) research group. JAMA 280:605^613
Mather KJ, Norman EG, Prior JC, Elliott TG 2000 Preserved forearm endothelial responses
with acute exposure to progesterone: a randomized cross-over trial of 17-b estradiol,
progesterone and 17-b estradiol with progesterone in healthy menopausal women. J Clin
Endocrinol Metab 85:4644^4649
Melis MR, Succu S, Spano MS et al 2001 Penile erection induced by EP 80661 and other
hexarelin peptide analogues: involvement of paraventricular nitric oxide. Eur J Pharmacol
411:305^310
Meneilly GS, Ryan AS, Veldhuis JD, Elahi D 1997 Increased disorderliness of basal insulin
release, attenuated insulin secretory burst mass, and reduced ultradian rhythmicity of insulin
secretion in older individuals. J Clin Endocrinol Metab 82:4088^4093
Meneilly GS, Veldhuis JD, Elahi D 1999 Disruption of the pulsatile and entropic modes of
insulin release during an unvarying glucose stimulus in elderly individuals. J Clin
Endocrinol Metab 84:1938^1943
Rigamonti AE, Cella SG, Marazzi N, Mˇller EE 1999 Nitric oxide modulation of the growth
hormone releasing activity of hexarelin in young and old dogs. Metabolism 48:176^182
Rossoni G, De Gennaro Colonna V, Bernareggi M, Polvani GL, Mˇller EE, Berti F 1998
Protectant activity of hexarelin and growth hormone against post ischemic ventricular
dysfunction in hearts from aged rats. J Cardiovasc Pharmacol 32:260^265
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

The ageing female reproductive

axis I
Henry G. Burger*, Emma Dudley{, Pam Mamers{, David Robertson*,
Nigel Groome} and Lorraine Dennerstein{

*Prince Henry’s Institute of Medical Research, Monash Medical Centre, 246 Clayton Road,
Clayton, Australia, {O⁄ce for Gender & Health, The University of Melbourne, 6th Floor,
Charles Connibere Building, Royal Melbourne Hospital, Vic 3050, Australia, {Department
of Obstetrics & Gynaecology, Monash Medical Centre, Clayton, Australia and }School of
Biological & Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus,
Headington, Oxford, UK

Abstract. The female reproductive axis includes the hypothalamo^pituitary unit, the
ovaries and the uterus. While changes in the brain may contribute to reproductive
ageing, the major focus of current research is on the ovary, where the progressive loss
of follicles ultimately leads to absent follicular function and consequent permanent
cessation of menstruation, the menopause. The pituitary gonadotropins, follicle-
stimulating hormone (FSH) and luteinizing hormone, stimulate ovarian secretion of
oestradiol and the inhibins from follicular granulosa cells, and androgens from
interstitial cells, including the theca. A primary event in the ageing of the reproductive
axis appears to be a decline in the secretion of inhibin B as follicle numbers fall. This leads
to a slow rise in FSH in women who continue to cycle regularly, particularly in the last
decade of reproductive life. As the menopause approaches, decreasing concentrations of
both oestradiol and inhibin B lead to more marked increases in the gonadotropins, which
reach their postmenopausal peak 2^3 years after ¢nal menses. In contrast, total
testosterone concentrations are maintained across the menopausal transition, with a
fall in sex hormone binding globulin (SHBG) and hence a rise in free testosterone.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 161^171

In women as opposed to men, there is a striking and readily observable marker of

reproductive ageing, the spontaneous cessation of menstrual bleeding or
menopause. In this paper, the characteristics of the hypothalamo^pituitary^
gonadal axis are summarized with respect to contemporary research on the
inhibins in particular. The overall perspective put forward is that the endocrine
changes which characterize female reproductive ageing are the consequence of
the progressive decline in ovarian follicle number, with the ultimate loss of
granulosa cell function. The precise mechanisms underlying follicular loss remain
to be characterized fully. While much research has concentrated on changes in
161
162 BURGER ET AL

oestradiol levels and their consequences, this review also summarizes data on the
inhibins and on androgens, particularly testosterone.

The hypothalamo^pituitary^ovarian axis

The hypothalamo^pituitary^ovarian axis is a classical example of a closed loop
endocrine negative feedback system. Pulsatile secretion of gonadotropin-
releasing hormone (GnRH), from hypothalamic peptidergic neurons drives the
pulsatile secretion of luteinizing hormone (LH) and follicle-stimulating hormone
(FSH) from the anterior pituitary. Whilst LH secretion depends on GnRH input,
there is a GnRH-independent component of FSH secretion, probably driven by
intrapituitary secretion of activin (Corrigan et al 1991), and illustrated by the
ability of dispersed pituitary cell cultures to continue to secrete FSH, but to cease
secreting LH (Farnworth et al 1988). The gonadotropins in turn drive ovarian
production of steroid and peptide hormones. Oestradiol is the major steroid
secretory product of the granulosa cell, where it is formed as a result of the action
of aromatase on testosterone derived from the surrounding theca under LH
control. Progesterone is the product of granulosa^lutein cell secretion following
formation of the corpus luteum. The inhibins are gonadal glycoprotein hormones,
again predominantly the product of the ovarian granulosa cell, though there is also
evidence of thecal inhibin secretion. Two major types of inhibin have been
isolated, inhibin A and inhibin B. Evidence from the localization of inhibin
subunits within the ovary (Roberts et al 1993) and from studies of circulating
inhibin levels (Groome et al 1996), indicates that inhibin A is a product primarily
of the dominant follicle. Inhibin B on the other hand is the product of the cohort of
growing follicles from which the dominant follicle is selected. The secretory
patterns of the two inhibins di¡er. Inhibin A levels are quantitatively low
throughout much of the follicular phase of the cycle and show a late follicular
phase rise in parallel with the preovulatory rise in serum oestradiol. Inhibin A
levels peak at mid cycle, fall brie£y and then rise to reach their highest levels
during the luteal phase, when inhibin A is a product of the corpus luteum, its
secretion being parallel to that of progesterone. In the late luteal phase,
circulating concentrations of oestradiol, progesterone and inhibin fall. Inhibin B,
on the other hand, shows an early follicular phase rise and fall, parallel to the rise
and fall in circulating FSH. A mid cycle peak occurs in parallel with that of inhibin
A, but following this peak, inhibin B falls to low concentrations and remains low
throughout the luteal phase until the initiation of the luteal^follicular transition,
when its levels rise, closely associated with the intercycle rise in FSH. The inhibins
speci¢cally inhibit the synthesis and secretion of pituitary FSH by mechanisms that
have not yet been clari¢ed. Recently, two types of molecules have been recognized
which may function as components of an inhibin receptor system, betaglycan
FEMALE REPRODUCTIVE AXIS I 163

(Lewis et al 2000) and p120 (Chong et al 2000). The precise signalling mechanisms
are still being elucidated. Oestradiol also exerts negative feedback e¡ects on
pituitary FSH and LH secretion, and acts predominantly at the hypothalamic
level. Under certain circumstances, oestradiol exerts a paradoxical positive
feedback e¡ect involved in the generation of the mid-cycle LH surge. When
operating as a negative feedback system, the pituitary^gonadal axis can be
conceptualized as a system in which FSH and LH drive the ovarian production
of oestradiol and the inhibins, which in turn feed back to negatively regulate
gonadotropin secretion. In this model, a primary defect in ovarian inhibin
secretion, for example, would be expected to lead to a monotropic increase in
circulating FSH levels.

Ovarian morphology as a function of age

The numbers of ovarian primordial follicles reach their peak during fetal life.
About one million are present at birth and the number then decreases
progressively with age. Semi-logarithmic plots of ovarian follicle numbers as a
function of age suggests that there is an increase in the rate of follicle loss at
approximately age 38, so that the curve appears to be biexponential (Richardson
et al 1987). This interpretation of the data has been challenged (Leidy et al 1998).
Few if any follicles remain at the time of the cessation of menses and follicular
numbers in older women whose cycles have become irregular are about 1/10 of
those in women of similar age in whom ovarian cyclicity continues. Whether or
not the rate of loss of follicles changes at age 38, there is clearly an alteration in
oocyte function as fecundity decreases markedly at around that time (Schwartz &
Mayaux 1982).

Endocrine markers of follicle number

Several studies of the pituitary^ovarian axis as a function of increasing age in
regularly cycling women have indicated that the most striking observable change
is a progressive rise in early follicular and mid cycle FSH levels without signi¢cant
change in the levels of LH, oestradiol or progesterone (Lee et al 1988). This
monotropic rise in FSH is explicable on the basis of changes in the circulating
concentrations of inhibin B (Klein et al 1996, Burger et al 2000a). Therefore the
authors (Burger et al 2000b) measured the circulating concentrations of oestradiol,
FSH and the inhibins in the early follicular phase of 66 regularly cycling women
aged 20^50. Serum FSH, inhibin A and oestradiol were all positively correlated
with age between years 20 and 50, the increase in FSH being particularly striking
in women over the age of 40. Inhibin B levels were not signi¢cantly correlated with
age, though they showed a tendency to fall slowly, but in women over the age of 40
164 BURGER ET AL

there was a highly signi¢cant inverse correlation between inhibin B and FSH
(r = 70.61, P50.001). When log FSH was modelled as a function of log inhibin
B and log oestradiol, with age ¢tted as a co-variate, only inhibin B was a signi¢cant
independent predictor of FSH. Other studies have also shown a decline in
circulating serum inhibin B levels as a function of age with no change or even an
increase in oestradiol and inhibin A (Danforth et al 1998, Welt et al 1999). Thus it
can be postulated that inhibin B is the main form of inhibin regulating FSH during
the follicular phase of the menstrual cycle. Inhibin B levels may therefore be a
marker of ovarian follicular numbers and/or function.
The importance of inhibin B as a regulator of the pituitary^gonadal axis becomes
more evident at the time of onset of menstrual irregularity, marking the beginning
of the menopausal transition or perimenopause. Studies from the authors’
laboratory have shown that the most clear-cut change in pituitary ovarian
function in women who had developed irregular menstrual cycles was a
profound fall in inhibin B without signi¢cant change in inhibin A and oestradiol,
and with a small but statistically non-signi¢cant increase in FSH (Burger et al
1998). As progression through the menopause transition occurs, inhibin A and
oestradiol levels also fall with further rises in serum FSH. From this prospective
study of a community based sample of women experiencing the menopause, it was
concluded that at the time of ¢nal menses, circulating FSH levels (48.4 Iu/l) were
approximately 50% of those which would ultimately be found postmenopausally,
whilst circulating oestradiol was also approximately 50% of its early follicular
phase levels at about 113 pmol/l (Burger et al 1999). Nadir oestradiol levels are
reached two to three years postmenopausally as are peak concentrations of FSH.
At this time inhibin B is undetectable and inhibin A is also very low or
undetectable.

Consequences of the fall in circulating oestradiol

The clinical markers signalling the end of reproductive function in the female are
symptoms resulting from loss of ovarian oestradiol secretion. The most striking
and characteristic symptom is the hot £ush which shows a variable prevalence in
di¡erent communities, but a¡ects 65^75% of women in developed countries.
Broad correlations between circulating oestradiol concentrations and hot £ush
frequency have been documented in various studies (Guthrie et al 1996). In the
Melbourne Women’s Midlife Health project, longitudinal observations have
con¢rmed that the only symptoms clearly related to the transition from early to
late perimenopause when the profound fall in oestradiol levels occurs are hot
£ushes, night sweats and vaginal dryness (Dennerstein et al 2000). The other
symptom which changes at this time is that of breast tenderness which falls
signi¢cantly as oestradiol levels fall. The profound drop in oestradiol
FEMALE REPRODUCTIVE AXIS I 165

concentrations, more than 90% across the menopausal transition, may have later
important health consequences. There is a clear-cut acceleration in the rate of loss
of bone mineral at the time of menopause (Guthrie et al 1998), continuing for ¢ve
to eight years and predisposing women, in particular, to later postmenopausal
osteoporotic fracture. Whether the change in circulating oestradiol also
predisposes to the occurrence of cardiovascular disease, independently of the
ageing process, remains a controversial issue. Long-term consequences of
oestrogen deprivation may result in an increased risk of the development of
Alzheimer’s dementia and possibly other disorders.

Changes in circulating androgens

The signi¢cance of changes in androgen secretion is a neglected area of female
reproductive ageing. One important study has documented a 50% fall in
circulating total and free testosterone concentrations in normal regularly cycling
women between the ages of 20 and 40 (Zumo¡ et al 1995). This has been postulated
to re£ect declining levels of adrenal androgen precursor secretion. Across the
menopausal transition itself, studies from the authors’ laboratory indicate that
there is no signi¢cant change in the circulating concentrations of total
testosterone, whilst there is a fall in sex hormone binding globulin and an
increase in free androgen index (Burger et al 2000b). Studies from other
investigators suggest that there may be a further increase in circulating androgen
levels in the late 50s and 60s (Laughlin 2000). The precise consequences of these
changes in androgen for womens’ health in general are unknown. On the other
hand, loss of androgen as may occur following ovariectomy may lead to
signi¢cant consequences of loss of drive and energy and loss of sexual interest
which can be corrected by androgen substitution (Shifren et al 2000).

Conclusions
The most striking endocrine marker of female reproductive ageing is a progressive
increase in the concentrations of circulating FSH, possibly beginning in the 20s and
becoming more obvious from approximately age 40. This increase in FSH may be
the result primarily of a decrease in the concentrations of circulating inhibin B,
acting as a marker of primordial follicle number. When follicle numbers fall to
critical levels, granulosa cell function becomes impaired and subsequently ceases
with falls in circulating oestradiol and inhibin A and further increases in FSH.
Changes in circulating androgens appear to occur particularly during
reproductive life when levels fall between ages 20 and 50. The implications of
ageing of the female reproductive axis for health and health policy remain to be
fully determined.
166 BURGER ET AL

References
Burger HG, Cahir N, Robertson DM et al 1998 Serum inhibins A and B fall di¡erentially as FSH
rises in perimenopausal women. Clin Endocrinol 48:809^813 (erratum: 1998 Clin Endocrinol
49:550)
Burger HG, Dudley EC, Hopper JL et al 1999 Prospectively measured levels of serum follicle-
stimulating hormone, estradiol and the dimeric inhibins during the menopausal transition in a
population-based cohort of women. J Clin Endocrinol Metab 84:4025^4030
Burger HG, Dudley E, Mamers P, Groome P, Robertson DM 2000a Early follicular phase serum
FSH as a function of age: the roles of inhibin B, inhibin A and estradiol. Climacteric 3:
17^24
Burger HG Dudley EC, Cui J, Dennerstein L, Hopper JL 2000b A prospective longitudinal
study of serum testosterone dehydroepiandrosterone sulfate and sex hormone-binding
globulin levels through the menopause transition. J Clin Endocrinol Metab 85:2832^2938
Chong H, Pangas SA, Bernard DJ et al 2000 Structure and expression of a membrane component
of the inhibin receptor system. Endocrinology 141:2600^2607
Corrigan AZ, Bilezikjian LM, Carroll RS et al 1991 Evidence for an autocrine role of activin B
within rat anterior pituitary cultures. Endocrinology 128:1682^1684
Danforth DR, Arbogast LK, Mroueh J et al 1998 Dimeric inhibin: a direct marker of ovarian
aging. Fertil Steril 70:119^123
Dennerstein L Dudley EC, Hopper JL, Guthrie JR, Burger HG 2000 A prospective populated-
based study of menopausal symptoms. Obstet Gynecol 96:351^358
Farnworth PG, Robertson DM, de Kretser DM, Burger HG 1988 E¡ects of 31 kilodalton
bovine inhibin on follicle-stimulating hormone and luteinizing hormone in rat pituitary
cells in vitro: actions under basal conditions. Endocrinology 122:207^213
Groome NP, Illingworth PJ, O’Brien M et al 1996 Measurement of dimeric inhibin B
throughout the human menstrual cycle. J Clin Endocrinol Metab 81:1401^1405
Guthrie JR, Dennerstein L, Hopper JL, Burger HG 1996 Hot £ushes, menstrual status and
hormone levels in a population-based sample of midlife women. Obstet Gynecol 88:437^442
Guthrie JR, Ebeling PR, Hopper JL et al 1998 A prospective study of bone loss in menopausal
Australian-born women. Osteoporos Int 8:282^290
Klein NA, Illingworth PJ, Groome NP, McNeilly AS, Battaglia DE, Soules MR 1996 Decreased
inhibin B secretion is associated with the monotropic FSH rise in older, ovulatory women: a
study of serum and follicular £uid levels of dimeric inhibin A and B in spontaneous menstrual
cycles. J Clin Endocrinol Metab 81:2741^2745
Laughlin GA, Barrett-Connor E, Kritz-Silverstein D, von Mˇhlen D 2000 Hysterectomy,
oophorectomy, and endogenous sex hormone levels in older women: the Rancho Bernado
study. J Clin Endocrinol Metab 85:645^651
Lee SJ, Lenton EA, Sexton L, Cooke ID 1988 The e¡ect of age on the cyclical patterns of plasma
LH, FSH, oestradiol and progesterone in women with regular menstrual cycles. Hum Reprod
3:851^855
Leidy LE, Godfrey LR, Sutherland MR 1998 Is follicular atresia biphasic? Fertil Steril 70:
851^859
Lewis KA, Gray PC, Blount AL et al 2000 Betaglycan binds inhibin and can mediate functional
antagonism of activin signalling. Nature 404:411^414
Richardson SJ, Senikas V, Nelson JF 1987 Follicular depletion during the menopausal
transition: evidence for accelerated loss and ultimate exhaustion. J Clin Endocrinol Metab
65:1231^1237
Roberts VJ, Barth S, el-Roeiy A, Yen SS 1993 Expression of inhibin/activin subunits and
follistatin messenger ribonucleic acids and proteins in ovarian follicles and the corpus
luteum during the human menstrual cycle. J Clin Endocrinol Metab 77:1402^1410
FEMALE REPRODUCTIVE AXIS I 167

Schwartz D, Mayaux MJ 1982 Female fecundity as a function of age: results of arti¢cial

insemination in 2193 nulliparous women with azoospermic husbands. Federation CECOS.
Federation CECOS. N Engl J Med 306:404^406
Shifren JL, Braunstein GD, Simon JA et al 2000 Transdermal testosterone treatment in women
with impaired sexual function after oophorectomy. N Engl J Med 343:682^688
Welt CK, McNicholl DJ, Taylor AE, Hall JE 1999 Female reproductive aging is marked by
decreased secretion of dimeric inhibin. J Clin Endocrinol Metab 84:105^111
Zumo¡ B, Strain GW, Miller LK, Rosner W 1995 24 h mean plasma testosterone concentration
declines with age in normal premenopausal women. J Clin Endocrinol Metab 80:1429^1430

DISCUSSION
Veldhuis: Is the quality of current ultrasonography such that one can argue
whether the amount of inhibin produced per antral follicle is declining, or there
are just fewer follicles entering the antral stage and then eventually being selected?
Burger: The evidence would be very strongly that the number of antral follicles
developing decreases as the total primordial follicle pool falls. This has been
modelled by Faddy & Gosden (1995) among others. Their model strongly
indicates that the number of follicles recruited gets progressively smaller. To my
knowledge there is only one study done ultrasonographically that speci¢cally
looked at this, and it does show a marked decline in the numbers of visible
follicles with increasing age. This is a potentially fertile area for more study.
Veldhuis: What is the evidence that follicle function is decreasing? You are
saying that it may not be. If one gave these women recombinant human FSH
under GnRH antagonist blockade, would one expect inhibin B secretion to
respond normally?
Burger: This is a di⁄cult study, which we have been talking about doing. My
prediction would be that the response would decrease, because there will be a
smaller number of follicles. This is precisely the problem I have had in trying to
think about how to design the study. How do you nail whether it is a per follicle or
total numbers issue?
Veldhuis: There was a paradoxical oestradiol elevation in the face of a granulosa
cell that under-secretes inhibin. What do you postulate is happening here?
Burger: I don’t have a good explanation, except the di¡erential regulation
phenomenon that I postulated. If the follicle starts to drop o¡ its inhibin
secretion, this allows FSH to rise. Our postulate is that oestradiol and inhibin
normally contribute about 50% each. This is based on physiological oestradiol
replacement therapy in post-menopausal women, for example, and looking at the
suppression of FSH that results from elevating oestradiol to follicular phase levels.
The feedback is about 50:50, and if you drop out one of the feedback factors, if the
oestradiol was thinking about switching o¡ as well, the higher FSH will stimulate
it into keeping going for a while.
168 DISCUSSION

Veldhuis: So you don’t ¢nd elevated oestradiol without elevated FSH.

Burger: Our data are on a once-a-year sample basis. They are not su⁄cient to
answer that. I am not far enough through the analysis to tell you too much about
another study which I have done in collaboration with Britt-Marie Landgren in
Stockholm. She has done a longitudinal study of about 15 women who have
been studied annually from age 45^47 until they reach their menopause. They
have had blood sampling three times a week through one cycle each year for up
to eight years. I am hoping to be able to dissect out some of those questions in that
group. Interestingly, it seems that the women who continue to have ovulatory
cycles with normal luteal phases don’t show much of a rise in FSH in the
follicular phase. It is only the women who start to show luteal insu⁄ciency, or
who have anovulatory cycles of various kinds, who do. My preliminary take on
that number is that about 50% of the women continue to have ovulatory cycles
right to the end of the transition, even though they start to become irregular,
whereas the others have all sorts of cycle patterns even though their degree of
irregularity doesn’t seem too di¡erent from those who don’t.
Prior: From some of the data, it looks like there are more follicular type cysts in
the ovaries of perimenopausal women. Are you ¢rm on the idea that each follicle
makes the same amount of inhibin?
Burger: I am not ¢rm on the idea; I have no reason to believe they wouldn’t. The
cysts may re£ect the higher FSH level.
Bj˛rntorp: Could you expand on the regulation of the androgen production?
Also, we are planning to treat our slightly hypoandrogenic women, who we
think are sick. They are all premenopausal. If we are going to treat them with
androgen, what is the best way of doing this?
Burger: The regulation of post-menopausal ovarian androgen secretion has not
been very fully worked on. There do appear to be a small number of subjects
(presumably those with the highest androgen levels) who have stromal
hyperplasia, even though by other criteria they haven’t had polycystic ovary
syndrome. There may be some degree of autonomy about the androgen secretion
in those with stromal hyperplasia. My hypothesis is that the stromal and interstitial
cells that remain in the ovary are the source of androgen. There must be some
aromatase there because some post-menopausal ovaries do continue to secrete a
small amount of oestrogen. If you measure the ovarian venous arterial di¡erence,
there is evidence that oestradiol is being made by the ovary. I suspect that the
androgen secretion is primarily LH driven, as it is premenopausally, and that
when oestrogen is given, it does two things: it suppresses gonadotropin secretion
and therefore also the androgen secretion to a degree, and if the oestrogen is given
orally it will also increase SHBG levels such that the free androgen levels will drop
disproportionately. Our suspicion is that a number of women who are given
standard hormone replacement therapy either continue to have or develop
FEMALE REPRODUCTIVE AXIS I 169

symptoms of androgen de¢ciency, including loss of libido, because their free

androgen levels have been dropped heavily. I don’t know of many systematic
studies that have really looked at the regulation of androgen secretion. It is not
an easy area to study.
Handelsman: Can I draw you out on the fall in testosterone in younger women?
Are there longitudinal data to support this, as claimed by some, and what do you
think the mechanisms are?
Burger: I don’t know of any longitudinal data in the younger age group. As you
saw, our studies began with women who were already 45. There are few
longitudinal studies in this area, and yet they are crucial. I have no idea about the
mechanism.
Riggs: Have you attempted to do regression analyses to see whether there is a
threshold level of oestradiol at which point bone loss begins? And does it
continue linearly or does it accelerate as oestradiol levels decrease?
Burger: The people primarily involved in the bone studies with us are John Wark
and Peter Ebeling. We didn’t see a high degree of correlation between change in
oestradiol and change in bone density between measurements. I don’t think we
have identi¢ed a threshold point, and from our limited data the rate of loss was
maximal in the group right at the time they became post-menopausal. Those who
remained post-menopausal already had a lesser rate of change than that very early
group.
Carroll: Women in this age group have a fairly high incidence of mood disorder.
Depending on whom you read, 15^30% may have depression at some time in the
period that you are studying. I see you have Lorraine Dennerstein (a psychiatrist) as
a collaborator. Can you make any comments on the timing of menopausal events
vis-a' -vis presence of mood disorder?
Burger: Lorraine Dennerstein is the principal investigator for our overall study,
and I represent the endocrine arm of it. When we did a longitudinal analysis of
symptoms, which included some of the symptoms you would associate with
mildly depressed mood, we couldn’t establish a correlation between these and the
occurrence of menopause. There are three things that come out of that longitudinal
analysis: hot £ushes/night sweats and vaginal dryness are the two symptoms
positively related to the occurrence of menopause, and the third one is
disappearance of breast tenderness.
Carroll: My question was whether a history of depression predicts an earlier
onset of the menopausal events.
Burger: I am not aware of it, and I think the relationship between mood disorder
and menopause has become clouded. To my knowledge it is much more related to
premenopausal experiences and psychosocial background than it is to the
menopause itself. In contrast to that, there was a paper published recently from
the group at NIH (Schmidt et al 2000). They did a prospective randomized
170 DISCUSSION

controlled trial in which they administered transdermal oestradiol to a group of

women who were moderately depressed at the time of menopause. This showed a
positive short-term e¡ect.
Shalet: I think you said you thought it was pretty useless to measure FSH and
oestradiol in terms of predicting the development of the menopause. Is this also
true for an inhibin B measurement? As I recollect, Faddy & Gosden (1996) describe
a change in the rate of atresia around about 37^38 years of age. Is that correct, and if
so what causes it?
Burger: My grip on the data as far as inhibin B is concerned is that it probably is
not a lot better. We need to examine this more critically. It may turn out to be
better. It certainly is an area which people working in assisted reproductive
technology (ART) programmes have addressed quite carefully. Is inhibin B
predictive of reproductive outcome in response to ART? Although this is
controversial, the best studies have shown that it is not predictive. In terms of
the rate of atresia, the semi-logarithmically plotted data do suggest that there is
a break in the line around age 38 with an increased rate of loss beyond this. This
has been challenged in a paper in Fertility and Sterility by Leidy et al (1998). They
argue that this was a mathematical artefact. In contrast, it is around age 38 that
the decline in fertility really kicks in. Clearly, the phenomenon that explains
follicle loss in atresia appears to be apoptosis. To me this only takes the
question back one stage: why does apoptosis occur?
Laron: The IGF1 decline may play a role in apoptosis; it is known that IGF1
delays or prevents apoptosis.
Morley: I was interested in your comment that testosterone goes up post-
menopausally. As you know, muscle also falls o¡ fairly dramatically post-
menopausally. It is not likely to be related to oestrogen. How convinced are you
that testosterone goes up over this period, and in particular bioavailable
testosterone? What is its role in replacement post-menopausally?
Burger: In our data, there was a fairly slow increase in free androgen index across
the menopause. Trying to continue a longitudinal study like this is extremely
di⁄cult to get funding for, because the agencies say that we have already had
enough money and want to start funding elsewhere. I can’t tell you any more
about our longitudinal study, which I think is the way to solve this. The data I
am talking about are cross-sectional and from at least three groups. The most
recent was a paper from Laughlin et al (2000) who looked at over 600 women.
They showed that from the late 50s right up to 80, there was an increase in total
testosterone and also an increase in SHBG. Thus free androgen index didn’t
change. I don’t know what the function of testosterone is at that stage. Perhaps
the most striking place where a function of testosterone is seen is in the young
woman who has had her ovaries removed. If you give her adequate oestrogen
and progesterone replacement therapy, often she will continue not to be in
FEMALE REPRODUCTIVE AXIS I 171

optimal health: she will lack energy and drive, and have low libido. If you give her
physiological level testosterone replacement it is strikingly bene¢cial. There was a
recent paper from Jan Shifren in the New England Journal of Medicine, who gave
testosterone transdermally strictly in the physiological dosage range to
oophorectomized women (Shifren et al 2000). This showed bene¢t in the older
age group. I believe from clinical practice that in a subset of women who present
peri- or post-menopausally with poor energy and loss of libido, in whom
testosterone levels are at the low end of the range and in whom you cannot
discern a relationship problem, testosterone is highly clinically bene¢cial.
Handelsman: The free androgen index in men is particularly inappropriate
because it cannot measure free testosterone (Kapoor et al 1993). The assumption
is that the testosterone is negligible related to SHBG concentrations (100-fold
di¡erence), which is not the case in men. However, this is true in women, and
free androgen index does not have the same theoretical problems in women.

References
Faddy MJ, Gosden RG 1995 A mathematical model of follicle dynamics in the human ovary.
Hum Reprod 10:770^775
Faddy MJ, Gosden RG 1996 A model conforming the decline in follicle numbers to the age of
menopause in women. Hum Reprod 11:1484^1486
Kapoor P, Luttrell BM, Williams D 1993 The free androgen index is not valid for adult males.
J Steroid Biochem Mol Biol 45:325^326
Laughlin GA, Barrett-Connor E, Kritz-Silverstein D, vov Muhlen D 2000 Hysterectomy,
oophorectomy, and endogenous sex hormone levels in older women: the Rancho Bernardo
Study. J Clin Endocrinol Metab 85:645^651
Leidy LE, Godfrey LR, Sutherland MR 1998 Is follicular atresia biphasic? Fertil Steril 70:
851^859
Shifren JL, Braunstein GD, Simon JA et al 2000 Transdermal testosterone treatment in women
with impaired sexual function after oophorectomy. N Engl J Med 343:682^688
Schmidt PJ, Nieman L, Danaceau MA et al 2000 Estrogen replacement in perimenopause-
related depression: a preliminary report. Am J Obstet Gynecol 183:414^420
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

The ageing female reproductive axis II:

ovulatory changes with perimenopause
Jerilynn C. Prior

Division of Endocrinology and Metabolism, University of British Columbia, Vancouver Hospital

and Health Sciences Centre, Vancouver, BC, Canada V5Z 1C6

Abstract. Perimenopause, a complex physiological transition for midlife women, begins

with changes in experiences many years before cycles become irregular, oestradiol levels
decrease or follicle-stimulating hormone levels increase. Erratic and average higher
oestradiol levels as well as shorter luteal phase lengths and lower progesterone levels
occur during perimenopause. These ovarian changes may be causally related to lower
inhibin production but the dynamic prospective inter-relationships within women are
not well documented. This review will ¢rst de¢ne perimenopause and then explore the
limited published data on ovulatory characteristics in perimenopause. In addition, it will
report preliminary prospective observational data on menstrual cycles and ovulation in
initially ovulatory women followed through the perimenopause. Prospective data
suggest that ovulation disturbances begin early in perimenopause and increase with
irregular cycles. Combined with higher oestradiol levels they may cause menorrhagia. It
is not yet known whether disturbances of ovulation relate to bone loss in perimenopausal,
as in premenopausal, women. It is also not known whether progesterone therapy can
e¡ectively counteract the end organ (breast, endometrial, brain) e¡ects of higher/erratic
oestradiol levels and e¡ectively treat perimenopausal vasomotor and other symptoms.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 172^192

Each woman is born with an average of over a million follicles in her two ovaries,
and each follicle contains an egg that could potentially be released and fertilized.
The life cycle of each woman’s cohort of follicles is not well known but includes
continuous maturation (that may manifest as ovarian cysts; Merrill 1963) and
atresia of immature follicles that begin long before puberty. Therefore,
independent of pituitary stimulation or ovulation, follicle numbers steadily
decrease. Prior to puberty, the ovaries enlarge and cystic activity increases but the
¢rst 10^12 years following menarche is required before the majority of women
consistently and normally ovulate (Vollman 1977). This review of ovulation will
focus on ovulation during perimenopause, the ¢nal portion of the life cycle of
ovarian follicles.
172
OVULATORY CHANGES WITH PERIMENOPAUSE 173

Phase A B C D E
FIG. 1. This time line diagram shows the approximate time intervals for the characteristic
changes of the perimenopause. Phases A and B occur before irregular periods begin. Phases C
and D occur before the year of ¢nal menstrual £ow that is phase E of the perimenopausal
transition (Prior 1998).

Perimenopause is understood to begin when the number of remaining ovarian

follicles reaches some (unknown) low point. Cross-sectional ovarian histological
studies suggest that the rate of follicle ‘loss’ accelerates abruptly in the late 30s or
early 40s (Richardson et al 1987). On the face of it, that information suggests that
ovulation becomes more prevalent during perimenopause. Perimenopause, for
this review, begins when a woman reports a change in experience (such as
increased breast tenderness and premenstrual symptoms with or without night
sweats and sleep disturbances, Fig. 1). That the ovaries become depleted of
follicles ¢ts with the prevalent notion that perimenopause is a time of ovarian
senescence during which ‘oestrogen deprivation’ develops. In fact, cross-
sectional data have been interpreted to support such an oestradiol decline
(Burger et al 1995) although a meta-analysis of all controlled studies in which
oestradiol levels have been systematically measured in perimenopausal and
premenopausal control women shows that oestradiol levels are signi¢cantly
higher in perimenopausal women (Prior 1998). These higher and erratic
oestradiol levels are postulated to be due to decreasing ovarian inhibin
production (Burger et al 2002, this volume) and others (Klein et al 1996, Welt et
al 1999). These lower inhibin levels allow follicle-stimulating hormone (FSH)
levels to rise slightly but not signi¢cantly, which stimulates multiple follicles
each of which secretes some oestradiol. The resulting oestradiol levels are not
only higher but because of disturbed pituitary^ovarian feedback, are also non-
suppressible (Fig. 2) (Prior 1998).
Rising and uninhibited FSH levels stimulate maturation and oestradiol
production by more follicles but it is not clear whether or not ovulation and
progesterone production are also hyperstimulated. More ovarian cysts, that
suggest anovulation, appear to be present (Djerassi et al 1995). These women,
were they to become pregnant, would more likely produce non-identical twins
(Gil¢llan et al 1996). However, all of the few studies in which ovulation has been
174 PRIOR

FIG. 2. Diagram of the ovaries in the follicular phase and the inter-relationships among
ovarian oestradiol, inhibin and FSH productions during the premenopausal and
perimenopausal years. Lacking adequate inhibin production from ovarian follicles (dark
ovoids), perimenopausal FSH levels rise slightly and the increasing oestradiol levels are not
su⁄cient, with lower inhibin levels, to suppress them. Dotted line, inhibition; solid line,
stimulation; shaded ovoids, growing follicles, striped bodies, atretic corpus luteum. Reprinted
from Prior (1998) with permission of the Endocrine Society.

documented prospectively and some cross-sectional studies indicate that

disturbances of ovulation are an integral part of the ageing of the reproductive
system for women.
The purpose of this review is to examine reproductive ageing in women as it
relates to ovulation and progesterone production throughout the several
postulated phases of the perimenopause (Table 1) (Prior 1998). This paper will
focus not only on available data concerning ovulation in perimenopause, but also
highlight the potential physiological and clinical consequences of such ovulation
disturbances. Finally, it will review the potential data from studies in progress and
propose therapy with progesterone for symptomatic perimenopausal women.
It is important ¢rst, to de¢ne ovulation disturbances and indicate how they are
diagnosed. The term ‘ovulation disturbance’ includes a spectrum of changes in
progesterone production by the corpus luteum including short luteal phase
(SLP) and anovulatory cycles (Anov) (Prior et al 1990a). An SLP lasts less than
12 days from the luteinizing hormone (LH) peak or has fewer than 10 days of
signi¢cant basal temperature elevation (Vollman 1977). Ovulation disturbances
and luteal phase length can be documented by quantitative basal temperature
measurement methods using least squares or mean temperature analysis to
determine whether and where in the cycle a signi¢cant temperature increase
occurs (Prior et al 1990b). Ovulation disturbances can also be diagnosed using
daily or intermittent measurements of blood (Welt et al 1999), saliva or urinary
progesterone breakdown products such as pregnanediol corrected for creatinine
OVULATORY CHANGES WITH PERIMENOPAUSE 175

(PdG) (Santoro et al 1996). However, no single gold standard for ovulation exists
and criteria for diagnosing ovulation and luteal phase length with each method are
not well described and validated. Importantly, no ovulation documentation
methods are suitable for continuous monitoring over long periods of time and in
random participants from populations.

Ovulation disturbances and perimenopause

The following will brie£y describe the perimenopausal prevalence and incidence of
ovulation disturbances while expanding on a previous review (Prior 1998). Cross-
sectional data will be reviewed before the prospective and epidemiological data are
examined.
The earliest cross-sectional studies of ovulation in 40^50 year old women used
non-quantitative basal temperature methods. Collett et al (1954) documented 302
cycles from 146 women aged 17^50 using the basal temperature nadir (assessed
visually from graphed data) as the day of ovulation. By these imperfect
temperature analysis criteria, 15.1% of cycles in women ages 40^50 were
anovulatory. This contrasted with an average of about 2% anovulatory cycles in
women aged 25^34. Doring (1969), also using non-quantitative basal temperature
methods, showed SLP and anovulatory cycles were more prevalent in cyclic
women of increasing age.
Vollman, a Swiss physician, collected basal temperatures for 14 848 cycles in
over 500 women and analysed them using an innovative quantitative mean basal
temperature (QBT) analysis method (that has subsequently been validated against
the midcycle LH peak; Prior et al 1990b). Women were strati¢ed by gynaecological
age (years since menarche); the prevalence of ovulation disturbances was
documented to be almost 60% in the ¢rst year and about 21% in the cycles at
highest gynaecological ages (Vollman 1977) (Fig. 3). Results from the three large
basal temperature-based data sets di¡er from those of the majority of smaller cross-
sectional studies using hormonal measurements. Four careful studies did not show
lower progesterone or PdG levels in older women (Lee et al 1988, Reame et al 1996,
Reyes et al 1977, Welt et al 1999). However, two hormone-based cross-sectional
studies did show signi¢cantly lower progesterone levels or PdG excretions in older
women (Ballinger et al 1987, Santoro et al 1996). One study tested PdG from daily
overnight urine samples across one cycle in women with regular cycles, contrasting
11 women over age 47 with 11 under 36. The older women had signi¢cantly lower
PdG levels as well as higher urinary total oestrogen excretions (Santoro et al
1996). Similar results were obtained using weekly serum sampling (Ballinger et al
1987).
Thus cross-sectional studies suggest but do not support the hypothesis that
ovulation disturbances are more prevalent in perimenopause. However, in none
TABLE 1 Proposed phases of the perimenopause: clinical and hormonal characteristics

Phase A Phase B Phase C Phase D Phase E

Duration *2^24 months *2^24 months *1^2 years *1^2 years 1 year
Menstrual cycles Regular
cycle Regular
cycle Irregular
alternating Oligomenorrhoea. Amenorrhoea
length shorter; length shortest; short and long cycles, Rare normal
short FP+. short FP+. normal ovulation less ovulatory cycles
Usually ovulatory Onset of ovulatory than 50%
disturbances

short luteal
phase and
anovulation
Flow Increased or the same Often increased in 22 with £ooding or 3; Spotting alternating None
duration and often alternating with £ooding is
amount common
Menstrual cycle- 2 breast tenderness, 22 breast tenderness, Less cyclic breast, Breast tenderness Breast tenderness,
related symptoms mood swings and mood swings and mood and £uid persistent in some mood and £uid
swelling before swelling before symptoms*, or women+other symptoms less*.
£ow*. £ow*. present and premenstrual Sometimes occur
Dysmenorrhoea +2 Often unpredictable. symptoms*. without subsequent
Headaches 2 and/or 22dysmenorrhoea. Dysmenorrhoea less Dysmenorrhoea may be £ow.
migraines start Headaches in migraine related to £ow. May 3 or persistent in a Dysmenorrhoea
su¡erers often severe occur at midcycle or few. Sometimes is usually gone
for many days in the relieved with £ow
cycle
Vasomotor Often begin. First Cyclic related to £ow Night sweats still tend Night and day VMS Night and day VMS
symptoms onset, cyclic before and usually still to be cyclic, but less no longer predict may become daily
and during £ow night sweats during predictable. For some £ow. For some and incessant. In
usually in the night or at the end of sleep. women VMS women they may be others they are rare,
or very early Fewer VMS in symptoms 3 but others very severe both day have decreased or
morning hours. midcycle begin to have marked and night, especially generally disappear
May occur at Rare daytime VMS 2. in long cycles
midcycle. First daytime VMS
Rare daytime VMS
Hormonal 2 E2 in early FP and 22 E2 especially E2 normal alternating E2 normal except 22 E2 normal or slightly
characteristics premenstrually. with ovulatory with 22, especially levels intermittently low but 22
FSH normal disturbances. around £ow in long cycles. intermittently.
FSH intermittently 2FSH 22 FSH
LH normal LH normal. 2 FSH more common. 2LH now common. 22 LH much of the
? Inhibin B 3 Inhibin B 3. 2 LH occasionally. Inhibin B 33. time.
Inhibin A normal Inhibin A 3 with Inhibin B 33. Inhibin A 33. Inhibins B and A both
ovulatory Inhibin A 3. below assay
disturbances sensitivity

2, moderately increased; 22, very high; ?, uncertain; E2, oestradiol; +FP, follicular phase; *, premenstrual symptoms; VMS, vasomotor symptoms.
Modi¢ed from Prior (1998).
178 PRIOR

FIG. 3. This bar diagram of ovulation disturbances cross-sectionally shows the percentages of
cycles in women of di¡erent gynaecological ages that are anovulatory (open bars), have SLPs
(short LP, grey bars) and are normally ovulatory (normal, black bars). Note that anovulatory
cycles increase in women of gynaecological ages 35^40 but that SLPs increase only slightly.
Adapted to represent data by Vollman (1977).

of these cross-sectional studies are the necessary clinical or prospective menstrual

cycle data reported that would allow a diagnosis of perimenopause or assessment of
the phase of participants.
There are even fewer prospective studies of ovulation and its disturbances
during perimenopause. The majority of those that do exist poorly describe
recruitment criteria and the participants’ physiological characteristics and clinical
experiences. Most well designed longitudinal studies report menstrual cycle data
without documentation of ovulation (Brambilla et al 1994, Kaufert et al 1987,
Treloar 1981). A number of other important longitudinal studies collecting
intermittent hormonal data have also not documented ovulation (Guthrie et al
1998, Longcope et al 1986). Even if ovulatory status is determined, luteal phase
lengths are not assessed systematically by any longitudinal data sets except those
of Vollman (1977). Others have only documented anovulatory but not SLP cycles,
because weekly urine sampling does not allow luteal length quanti¢cation (Metcalf
1979, Brown 1985).
OVULATORY CHANGES WITH PERIMENOPAUSE 179

The available prospective data consistently show that ovulation disturbances

increase in perimenopause. Vollman (1977) diagrams the ¢nal 100 cycles in one
woman who experienced 16% of cycles as anovulatory. Vollman also noted that
within one woman the basal temperatures decreased (in both the follicular and
luteal phases) in cycles during the 3rd and 4th decades from menopause
compared with the 2nd and 1st (Vollman 1977). Temperatures were lower by
70.1 8C in the follicular and by ^0.15 8C in the luteal phase, suggesting an added
e¡ect of lower progesterone production in addition to the e¡ect of an apparently
lower metabolic rate (documented earlier, Collett 1949).
Although detailed cycle data are graphed for three or four cycles each from only
four women, Shideler shows the variability of both £ow and ovulation (Shideler et
al 1989). Figure 3 shows that each woman in phase C of perimenopause experienced
variations in both luteal phase and in cycle lengths and that each experienced
ovulation disturbances. These women, in contrast to many reported in the above
cross-sectional studies, were perimenopausal by cycle variability and/or vasomotor
symptoms (VMS) (Shideler et al 1989).
A much longer set of weekly urinary hormonal data are available from each of
two women over the ¢nal 6^7 years of the perimenopausal transition (Brown
1985), and from one woman over 108 consecutive weeks (Metcalf & Donald
1979). The two women studied by Brown recorded 15 and 23 episodes of £ow
and both had several months of no £ow at the end of monitoring. Using 5 as the
PdG cut-o¡ for ovulatory cycles, 8 of 15 and 10 of 23 cycles
only 48.4%
were
ovulatory (Brown 1985). The longitudinal data in one woman, however, appeared
to be earlier in the perimenopausal transition and showed that 86% of cycles were
ovulatory (18 of 21 £ow episodes were ovulatory by PdG) (Metcalf & Donald
1979).
In a larger and more systematic study, and the only one that had as its primary
focus as the prospective study of ovulation in perimenopause, Metcalf
systematically documented weekly urinary PdG over three cycles in 139
women over 40 years of age (Metcalf 1979). 93% of the 86 participants in
whom all of the documented cycles were of normal cycle length (21^35 days)
were ovulatory (although some probably were SLP cycles) as were 95% of the
cycles in the 81 women who reported no recent cycle changes. However, if cycles
were irregular either by history (in 58 women) or by prospective documentation
(in 53 women), the percentage of the three cycles that were consistently
ovulatory decreased to 39.7% and 34%, respectively (Metcalf 1979). These
latter women who were perimenopausal by epidemiological criteria based on
irregular cycles had an average of 36.9% of consistently ovulatory cycles
(Metcalf 1979). Despite this, PdG levels did not di¡er with age and 10.3% of
women who had gone three months without £ow subsequently ovulated
(Metcalf 1979).
180 PRIOR

The ¢nal prospective study documented three women aged 37^47 for whom one
cycle of daily hormonal data were available approximately 10 years previously and
graphically displayed two cycles of hormonal data 11 years apart from one woman
(Welt et al 1999). Mean progesterone levels were lower in women studied 10 years
later (4.9 versus 11.7 ng/ml) in addition to lower luteal phase inhibin A levels and
borderline lower follicular phase inhibin B levels. No signi¢cant prospective
di¡erences in LH, FSH or oestradiol were documented (Welt et al 1999). From
graphed data, luteal phase lengths shortened in one woman followed over 11
years (10 and 11 versus 12 and 13 days from the LH peak) (Welt et al 1999).
A large prospective epidemiological study from Malmo, Sweden, providing the
only population-based data of ovulation available, studied over 150 women for 12
years with serum hormone levels measured at 3^6 monthly intervals (Rannevik et al
1995). This study followed women through menopause and for several years
following it. In women who were 72^61 months before their ¢nal £ow, 62.2% of
cycles were ovulatory with a mean progesterone level of 27.3 nmol/l. However,
women in the 0^6 months before the ¢nal £ow experienced ovulatory cycles only
4.8% of the time although the average progesterone level was unchanged at
22.4 nmol/l (Rannevik et al 1995). These data were presented as a summary in the
text but no primary progesterone data were provided.
Thus prospective data consistently document ovulation disturbances as a part of
perimenopause and reproductive ageing in women. In women who were
symptomatic with new cycle irregularity or vasomotor symptoms the incidence
of ovulation disturbances was increased. However, the literature contains
information about prospective variations in ovulation in relatively few women
and includes no primary data from epidemiological samples. These prospective
results (Fig. 4) suggest that ovulation disturbances do not generally increase until
phase C of perimenopause when cycle intervals start becoming variable.

Clinical e¡ects of ovulation disturbances

The focus on ovulation disturbances in this paper has so far only discussed their
prevalence and incidence. It is likely however, that any SLP or persistent
ovulation disturbances present in phases A and B when cycles are still regular
causes dysfunctional bleeding and menorrhagia in women experiencing high
oestradiol levels. It is also likely that premenstrual symptoms are increased in
cycles with ovulation disturbances in addition to high oestradiol levels (Wang
et al 1996). Severe mood swings, increased stretchy mucus and breast
tenderness prior to and during £ow commonly coexist with disturbed
perimenopausal ovulation. Ballinger described heavy £ow in such cycles
(Ballinger et al 1987). In addition, the cycles with ovulation disturbances may
be at increased risk for VMS based on data suggesting increased VMS prior to
OVULATORY CHANGES WITH PERIMENOPAUSE 181

FIG. 4. The ovulatory characteristics of three or four consecutive cycles from three
perimenopausal women. The cycle lengths are shown as open bars and luteal phase lengths are
shown in black. Using 10 days as the lower limit of normal, each woman had at least one
SLP and two of three women had anovulatory cycles (*). Redrawn from data by Shideler et al
(1989).

£ow and that progesterone/progestin therapy is an e¡ective treatment. Ovulation

disturbances are related to rapid cancellous bone loss in regularly cycling,
initially ovulatory premenopausal women (Prior et al 1990a) and cyclic
medroxyprogesterone therapy increases bone density in women with menstrual
cycle and ovulation disturbances (Prior et al 1994). It is probable, therefore, that
ovulation disturbances in any phase of perimenopause, but especially phases C
and D contribute to the increased bone loss that occurs at that stage of the
transition (Okano et al 1998, Prior 1998).
In summary, ovulation disturbances are probably of clinical as well as
physiological consequence in the perimenopause and may relate to menorrhagia,
VMS, mood disturbances and accelerated bone loss. Studies are needed to
document this because therapy with cyclic progesterone or medroxyprogesterone
may well be e¡ective in symptomatic perimenopausal women.
182 PRIOR

Research in progress
Prospective data from the Vancouver Ovulation and Bone Change Cohort (Prior
et al 1996, 1990a) are still being collected as these women become perimenopausal
or menopausal. Some of those women continue to keep quantitative basal
temperature and Daily Perimenopause Diary (Prior 1999) records, and will
potentially provide important comparisons of ovulation and experiences in
premenopausal and perimenopausal cycles in relation to subsequent menopause
and bone density. As an illustration, one year of ovulatory characteristics
(analysed by the QBT least squares method; Prior et al 1990b) are shown from
the initial study and 10 years later in one woman (Fig. 5). This woman’s initial
cycles averaged 29.3 1.6 days in length and became two days shorter (27.0 1.5
days, P = 0.0004). The changes in luteal phase length were even more dramatic
with reduction from a mean of 11.2 2.0 to 5.4 2.7 days over the 10 years
(P50.0001). Note that she had no normally ovulatory cycles in 1994^1995 even
though she continued to have regular cycles.
A large selected population to represent the US ethnic mix is being studied with
US National Institutes of Health funding (Sowers et al 2000). Although over 3000
women will be followed prospectively, only limited, intermittent documentation
of hormone levels is being performed.
The Canadian Multicentre Osteoporosis Study (CaMOS), a national
population-based study that includes approximately 35 premenopausal and 60
perimenopausal women in each of nine centres (Kreiger et al 1999), is just
completing a three-year follow-up in women who were aged 40^60 at baseline.
The 3 year questionnaire includes information about cycle regularity and the
experience of premenstrual symptoms and night sweats as well as molimina
(Prior 1997). We are currently in the process of evaluating the molimina question
in relation with the characteristics in premenopausal women as well as validating its
assessment of ovulation. Eventually, prospective population-based data about
experiences, weight, diet and exercise as well as bone changes will be available
from CaMOS data. Unfortunately, no current methods for prospective
continuous ovulation documentation are su⁄ciently convenient and reliable that
they can be used in epidemiological samples.

Summary
This review of ovulation in midlife shows that ovulation disturbances are an
important characteristic of endocrine ageing in women and are likely to be
prevalent in all phases of perimenopause. At present only limited prospective
data are available and these data generally lack clinical correlation and recording
of women’s experiences. The ovarian hormonal physiology, sociocultural context,
OVULATORY CHANGES WITH PERIMENOPAUSE 183

perimenopausal women
(3 cycle window)
Percentage of

FIG. 5. This three-part ¢gure summarizes prospective data on ovulation disturbances during
the perimenopause. The top section shows the proportion of three women experiencing three
consecutive cycles that are normal (open bar) or showed ovulation disturbances (short luteal
phase [SLP] and/or anovulatory in black). Metcalf’s data in women with irregular cycles are
shown on the left (n = 58) and on the right prospective data for 3^4 consecutive cycles in three
women (see Fig. 4). The middle portion of the diagram shows prospective data drawn as
percentage of ovulatory (open bar, [includes SLP for Shideler data]) and anovulatory cycles
(black bar). The bottom panel shows the percentage of sera with luteal levels of progesterone
(Luteal Levels Prog, open bar) or low progesterone levels (Low Prog, black bar) during the 72^
61 versus 6^0 months before the ¢nal menstrual £ow from Rannevik data. mo, months. All data
redrawn from published work (Metcalf 1979, Shideler et al 1989, Brown 1985, Rannevik et al
1995).
184 PRIOR

FIG. 6. Diagram contrasting 14 consecutive cycles in one ovulatory woman during 1985^1986
and 1994^1995 showing cycle length (open bars) and luteal phase lengths (black) using
quantitative basal temperature analysis with least mean squares statistics (Prior et al 1990b).
The normal luteal phase length of 10 days is virtually consistent in the earlier years but is never
present despite 86% continued ovulatory cycles when she experienced vasomotor symptoms and
sleep disturbances although remained in phase B of perimenopause.
OVULATORY CHANGES WITH PERIMENOPAUSE 185

experiential and clinical consequences as well as genetic background of

perimenopausal ovulation disturbances need to be systematically documented in
prospective studies.

References
Ballinger CB, Browning NC, Smith AH 1987 Hormone pro¢les and psychological symptoms in
peri-menopausal women. Maturitas 9:235^251
Brambilla DJ, McKinlay SM, Johannes CB 1994 De¢ning the perimenopause for application in
epidemiologic investigations. Am J Epidemiol 140:1091^1095
Brown JB 1985 Return of fertility after childbirth and during lactation, and changes in the
climacteric. J Biosoc Sci Suppl 9:5^27
Burger HG, Dudley EC, Hopper JL et al 1995 The endocrinology of the menopausal transition:
a cross-sectional study of a population-based sample. J Clin Endocrinol Metab 80:3537^3545
Burger HG, Dudley E, Mamers P, Robertson D, Groome N, Dennerstein L 2002 The ageing
female reproductive axis I. In: Endocrine facets of ageing. Wiley, Chichester (Novartis Found
Symp 242) p 161^171
Collett M 1949 Basal metabolism at the menopause. J Appl Physiol 1:629^636
Collett ME, Wertenberger GE, Fiske VM 1954 The e¡ect of age upon the pattern of the
menstrual cycle. Fertil Steril 5:437^448
Djerassi A, Coutifaris C, West VA et al 1995 Gonadotroph adenoma in a premenopausal woman
secreting follicle-stimulating hormone and causing ovarian hyperstimulation. J Clin
Endocrinol Metab 80: 591^594
Doring GK 1969 The incidence of anovular cycles in women. J Reprod Fertil Suppl 6:77^81
Gil¢llan CP, Robertson DM, Burger HG, Leoni MA, Hurley VA, Martin NG 1996 The control
of ovulation in mothers of dizygotic twins. J Clin Endocrinol Metab 81:1557^1562
Guthrie JR, Ebeling PR, Hopper JL et al 1998 A prospective study of bone loss in menopausal
Australian-born women. Osteoporos Int 8:282^290
Kaufert PA, Gilbert P, Tate R 1987 De¢ning menopausal status: the impact of longitudinal data.
Maturitas 9:217^226
Klein NA, Illingworth PJ, Groome NP, McNeilly AS, Battaglia AS, Soules MR 1996 Decreased
inhibin B secretion is associated with the monotropic FSH rise in older, ovulatory women: a
study of serum and follicular £uid levels of dimeric inhibin A and B in spontaneous menstrual
cycles. J Clin Endocrinol Metab 81:2742^2745
Kreiger N, Tenenhouse A, Joseph L et al 1999 The Canadian Multicentre Osteoporosis Study
(CaMos): Background, rationale, methods. Can J Aging 18:376^387
Lee SJ, Lenton EA, Sexton L, Cooke ID 1988 The e¡ect of age on the cyclical patterns of plasma
LH, FSH, estradiol and progesterone in women with regular menstrual cycles. Hum Reprod
3:851^855
Longcope C, Franz C, Morello C, Baker RS, Johnston CC Jr 1986 Steroid and gonadotropin
levels in women during the peri-menopausal years. Maturitas 8:189^196
Merrill JA 1963 The morphology of prepubertal ovary: relationship to the polycystic ovary
syndrome. Southern Med Journal 56:225^231
Metcalf MG 1979 Incidence of ovulatory cycles in women approaching the menopause. J Biosoc
Sci 11:39^48
Metcalf MG, Donald RA 1979 Fluctuating ovarian function in a perimenopausal women. N Z
Med J 89:45^47
Okano H, Mizunuma H, Soda M et al 1998 The long-term e¡ect of menopause on
postmenopausal bone loss in Japanese women: results from a prospective study. J Bone
Miner Res 13:303^309
186 DISCUSSION

Prior JC 1997 Ovulatory disturbances: they do matter. Can J Diagnosis (Feb issue) 64^80
Prior JC 1998 Perimenopause: the complex endocrinology of the menopausal transition. Endocr
Rev 19:397^428
Prior JC 1999 The puzzle of perimenopause
using the daily perimenopause diary. Video
produced by University of British Columbia, distributed by BC Endocrine Research
Foundation, British Columbia
Prior JC, Vigna YM, Schechter MT, Burgess AE 1990a Spinal bone loss and ovulatory
disturbances. N Eng J Med 323:1221^1227 (erratum: 1990 N Eng J Med 328:1724)
Prior JC, Vigna YM, Schulzer M, Hall JE, Bonen A 1990b Determination of luteal phase length
by quantitative basal temperature methods: validation against the midcycle LH peak. Clin
Invest Med 13:123^131
Prior JC, Vigna YM, Barr SI, Rexworthy C, Lentle BC 1994 Cyclic medroxyprogesterone
treatment increases bone density: a controlled trial in active women with menstrual cycle
disturbances. Am J Med 96:521^530
Prior JC, Vigna YM, Barr SI, Kennedy S, Schulzer M, Li DK 1996 Ovulatory premenopausal
women lose cancellous spinal bone: a ¢ve year prospective study. Bone 18:261^267
Rannevik G, Jeppsson S, Johnell O, Bjerre B, Laurell-Borulf Y, Svanberg L 1995 A longitudinal
study of the perimenopausal transition: altered pro¢les of steroid and pituitary hormones,
SHBG and bone mineral density. Maturitas 21:103^113
Reame NE, Kelche RP, Beitins IZ, Yu MY, Zawacki CM, Padmanabhan V 1996 Age e¡ects of
follicle-stimulating hormone and pulsatile luteinizing hormone secretion across the menstrual
cycle of premenopausal women. J Clin Endocrinol Metab 81:1512^1518
Reyes FI, Winter JS, Faiman C 1977 Pituitary-ovarian relationships preceding the menopause. I.
A cross-sectional study of serum follicle-stimulating hormone, luteinizing hormone,
prolactin, estradiol and progesterone levels. Am J Obstet Gynecol 129:557^564
Richardson SJ, Senikas V, Nelson JF 1987 Follicular depletion during the menopausal
transition: evidence for accelerated loss and ultimate exhaustion. J Clin Endocrinol Metab
65:1231^1236
Santoro N, Brown JR, Adel T, Skurnick JH 1996 Characterization of reproductive hormonal
dynamics in the perimenopause. J Clin Endocrinol Metab 81:1495^1501
Shideler SE, DeVane GW, Kalra PS, Bernirschke K, Lasley BL 1989 Ovarian-pituitary hormone
interactions during the perimenopause. Maturitas 11:331^339
Sowers M, Crawford SL, Sternfeld B et al 2000 SWAN: a multicenter, multiethnic, community-
based cohort study of women and the menopausal transition. In: Lobo RA, Kelsey J, Marcus
R (eds) Menopause: biology and pathobiology. Academic Press, San Diego, p 175^188
Treloar AE 1981 Menstrual cyclicity and the pre-menopause. Maturitas 3:249^264
Vollman RF 1977 The menstrual cycle. Major Probl Obstet Gynecol 7:1^193
Wang M, Seippel L, Purdy RH, Backstrom T 1996 Relationship between symptom severity and
steroid variation in women with premenstrual syndrome: study on serum pregnenolone,
pregnenolone sulfate, 5alpha-pregnane-3,20-dione, and 3alpha-hydroxy-5a-pregnan-20-one.
J Clin Endocrinol Metab 81:1076^1082
Welt CK, McNicholl DJ, Taylor AE, Hall JE 1999 Female reproductive aging is marked by
decreased secretion of dimeric inhibin. J Clin Endocrinol Metab 84:105^111

DISCUSSION
Veldhuis: One of the distinctions I see between your work and Henry Burger’s is
that you propose that oestradial elevations occur before FSH. Therefore, you have
the task of accounting for an isolated oestradiol increase. Is that right?
OVULATORY CHANGES WITH PERIMENOPAUSE 187

Prior: Yes. My hypothesis would be that FSH isn’t measurably higher (Burger
1994), but this doesn’t mean that it isn’t occasionally higher and at a level su⁄cient
to recruit a larger cohort of follicles.
Veldhuis: We do come to a distinction that will probably need more careful
study, as to whether the antral follicle cohort is increased or not. High quality
ultrasonogrophy prospectively could be marvellous.
Is it your view that the luteal phase defects are a marker of the primary ovarian
disease associated with follicle attrition, or is this a neuroendrocrine disorder
re£ected in due course in an abnormal cytodi¡erentiation?
Prior: I would guess that it is the organization of the stimuli to the follicle, but I
don’t know the answer.
Burger: I don’t know the answer either, but I have normally interpreted luteal
phase defects as related to a defect in follicular function. However, it is very hard to
diagnose them in young reproductive-age women. It is a woolly area.
Veldhuis: I don’t know whether iNOS is in the ovary, but the NOergic pathway
is clearly there
it is partly driven by FSH. Can you think of a mechanism that
would drive inappropriate oestradiol episodically leading to some defect in follicle
growth, which is a mixture of oestradiol, FSH, insulin-like growth factor (IGF)1
and also negative regulators producing orderly follicle growth, yet enough
oestradiol to drive a surge, and then complex organizaton of that follicle into a
corpus luteum? Could the ovarian ageing seen in the human be the analogue of
the CNS ageing seen in the rat? In the female rat there is strong evidence for
CNS-dependent ageing, rather than primary ovarian ageing, based on the
ovarian transplantation studies (Evans et al 1992). Is there a possible model for
iNOS being involved in the early changes in the ovary? The species di¡erence is
not the basic issue to me, but rather the site where the defect ¢rst emerges.
Prior: I don’t know, but this is a useful thing to start thinking about.
Handelsman: The point that Jerilynn Prior made was that all the £ushing episodes
seem to start through sleep. It has always struck me that there are clearly oestrogen-
dependent symptoms and then there are the rest. I have always assumed that some
of these are due to sleep disturbance from £ushing during the night. If this was so,
it should be oestrogen sensitive. It appears that this is not the case. Is sleep
disturbance important?
Prior: I would say that it is very important. Something happens to sleep
physiology, because there is sleep disturbance in women who have never
reported hot £ushes. There is some neuroendrocrine change that occurs at that
stage of life.
Veldhuis: There aren’t many data. We wrote one paper showing this LH^FSH
synchrony loss in the monotrophic FSH elevation stage of the premenopause
(Matt et al 1998a) and in ageing men (Pincus et al 1997). It is clearly abnormal,
but I don’t know that it can’t be explained by a downstream problem in variable
188 DISCUSSION

oestrogen overproduction and then withdrawal, leading sometimes to high FSH.

In the men it seems clear that there is some central nervous system (CNS)
dysregulation of the out£ow control of several factors, some neurogenic, some
hormonal, and some across axes (Pincus et al 1996, Veldhuis et al 1999a,b, 2000).
But whether these derangements are in any sense primary is unclear. The sleep
disorder raises the question of whether CNS alterations are precursory.
Mˇller: Referring to this argument, in addition to their genomic actions, steroids
can have non-genomic e¡ects that occur with a shorter time lapse. They can act
directly at the level of hormone-sensitive neurotransmitter receptors present on
cellular membranes of neuronal cells, and can also be synthesized in the CNS,
independently from their peripheral sources (Baulieu 1997). For instance,
precursors (pregnenolone sulfate, dehydroepiandrosterone) or metabolites
(allopregnanolone) of progesterone, act respectively as antagonists and agonists
of the GABAA receptor, thus in£uencing neuronal activity within large parts of
the CNS (Genazzani et al 1996). If they can modulate GABAergic or
glutaminergic function, we should not be surprised that they can induce
behavioural alterations.
For instance, insomnia may be an expression of latent depression. Many of the
symptoms such as irritability and depressed mood can be interpreted as the action
of these steroids directly on the brain.
It is interesting that in rats, also, before the complete arrest of the oestrus
cyclicity there is a phase called constant oestrus, involving continuous activation
of oestrogen production (Kalra et al 1993). This is reminiscent of the high
variability of circulating oestradiol levels of women in the late perimenopause
(Burger 1999).
Veldhuis: In the rat there is a blunting of LH surge in middle-aged females, well
before their oestrus cyclicity changes, pointing to a CNS component (Matt et al
1998b). This may be a feature of the rodent (Evans et al 1992). Others would say
it is also a feature of the human.
Burger: I have debated this issue at some length with Phyllis Wise. I challenged
her about the relevance of the rodent model to an understanding of the human
endocrinology. I must say that she has persuaded me that there are similarities
that warrant the continued exploration of that model.
Laron: Coming back to the issue of sleep disturbance, we know that in puberty
children love to sleep for a long time. Would this mean that they have much more
of what the menopausal women lack?
Veldhuis: GH secretagogues are thought to be somnogenic (Giustina &
Veldhuis 1998). There are data showing that the actions of GH secretagogues
favour sleep (Thorner et al 1990).
Carroll: With regard to Dr Mˇller’s point about the neurosteroids, they tend to
have GABAergic and somewhat sedative properties. As I was preparing my review
OVULATORY CHANGES WITH PERIMENOPAUSE 189

I came across another reference to yet another alleged explanation for the night
sweats and vasomotor instability of menopausal women. This is increased cell
number and enlarged cell size of neurokinin B neurons in the infundibular
nucleus of the hyothalamus. This is another theory as to why they develop
vasomotor instability.
I agree that the menopausal period is one of stress. In general, in the
epidemiological studies of depression that I talked about yesterday, an important
point to make is that depressive symptoms seem to count, rather than a formal
psychiatric diagnosis of major depression. The traditional clinical psychiatric
distinction between minor depression and major depression appears to be
unimportant in these longitudinal studies.
Mˇller: There are data showing that the latest antidepressants are e¡ective in
blocking some of these postmenopausal symptoms. This has been shown for
instance with inhibition of hot £ushes by venlafaxine, a speci¢c blocker of
noradrenergic re-uptake (Loprinzi et al 2000).
Veldhuis: Does it a¡ect the sleep disturbance?
Mˇller: I don’t remember.
Haus: We have been talking about sleep disturbances, but nobody has
mentioned melatonin. In the USA, melatonin can be purchased over the counter
and many people who develop sleep disturbances take it, some under medical
direction, many others without. What role might melatonin play in the treatment
of sleep disturbances in the elderly? Some publications on this topic have
previously appeared (e.g. Dagan et al 1998).
Prior: I haven’t seen any data on this.
Riggs: I want to ask about the important studies you did in 1990 on the relation-
ship between decreased luteal phase duration and bone loss. The limitation of these
studies is that it takes so long to get an accurate estimate of bone loss by bone
densitometry. What has changed since then is that we now have the biochemical
markers, so it is possible to see changes in bone turnover within a single cycle. Have
you had a chance to look at this, to follow up on the earlier studies?
Prior: Progesterone works on bone formation and not on bone resorption. The
markers that we have for bone formation are in£uenced by bone resorption. This
means that we don’t have a speci¢c marker that we could use to document the
progesterone e¡ect on bone.
Riggs: I disagree with that. Formation and resorption are linked together, so the
markers for each will be linked together. If you measured resorption and
formation, and looked at the gap between them, you ought to be able to see a
di¡erence, if there is a di¡erence, within the individual cycle. This might be
worth doing.
Prior: There have been several attempts to look at changes in bone resorption
across normal ovulatory menstrual cycles. The data are wild. There is obviously a
190 DISCUSSION

lag of a few days between a time of change in oestrogen and a change in those
markers. It is complicated to do this study; I haven’t yet got su⁄cient funding to
do it. I have applied to do a study in which I will control resorption in
perimenopause, give progesterone cyclically or daily, and see what happens to
bone.
Robertson: The vasomotor symptoms of menopause are so dramatic and so
discrete. I am amazed we don’t have a better handle on them. What is the current
understanding of the kindling features: the things that are occurring in the minute
or two before the onset of these? What are the mediators of the vasomotor
symptoms?
Prior: That’s a good question. It is clear that there is an adrenergic activation
prior to any temperature change. The women will describe that they woke up
suddenly and they felt anxious or angry, and then they felt hot. The better studies
of the temperature change suggest that there is a tiny increase in core temperature
which the hypothalamus then decides is not OK. It begins the process of
vasodilitation that leads to heat dissipation. The core temperature then drops.
Some women will get quite chilled afterwards and will describe ‘cold sweats’.
Hot £ushes are increased by social stress (Swartzman et al 1990). For example,
people who are having a number of hot £ushes will sometimes get a £ush if they
hear an argument in the next o⁄ce. Early on in some women £ushes are triggered
by even a small sip of alcohol. We have used a diary system so that each day a person
integrates their assessment of the intensity of the hot £ush on a 0^4 scale, the
number of discrete episodes during the waking day, and the number and
intensity during sleep. We have those data tracked longitudinally in quite a large
number of women. I think we are getting a better descriptive idea about it, but the
endocrinology and neurobiochemistry are complex.
Burger: There are so many things about hot £ushes that are not understood. One
of the curiosities is that in population surveys of women there seems to be a baseline
hot £ush frequency reporting rate of about 10% in the normal community under
the age of 40. If you ask women whether they have had hot £ushes in the last two
weeks, about 10% will say they have. I see a lot of patients referred for hot £ushes
for which I can ¢nd no cause whatsoever.
Prior: To put a di¡erent perspective on it, I would think that many of those
women are in the earliest phases of perimenopause (Prior 1998). If you were to
follow them over time, they will begin to show short follicular phases and
ovulation disturbances.
Ruiz-Torres: Have you any data on leptin interactions?
Prior: No, but that’s an interesting question.
Handelsman: One of the purposes of this meeting is to identify things that we
don’t understand. The most striking thing for me is watching those log linear
declines in primordial follicles, which are disappearing at a phenomenal rate.
OVULATORY CHANGES WITH PERIMENOPAUSE 191

What controls this atresia? With the time decay in a log linear fashion, menopause
could be delayed by years. Yet we have no idea what controls this atresia that
dictates age at menopause.
Veldhuis: First, the ovarian physiologists are going to have to teach us what local
factors are produced by the granulosa cell that talk to the theca cell and egg, and
vice versa.

References
Baulieu EE 1997 Neurosteroids: of the nervous system, by the nervous system, for the nervous
system. Recent Prog Horm Res 52:1^32
Burger HG 1994 Diagnostic role of follicle-stimulating hormone (FSH) measurements during
the menopausal transition
an analysis of FSH, oestradiol and inhibin. Eur J Endocrinol
130:38^42
Burger HG 1999 The endocrinology of the menopause. J Steroid Biochem Mol Biol 69:31^35
Dagan Y, Yovel I, Hallis D, Eisenstein M, Raicik I 1998 Evaluating the role of melatonin in the
long-term treatment of delayed sleep phase syndrome (DSPS). Chronobiol Int 15:181^190
Evans WS, Sollenberger MJ, Booth RA Jr et al 1992 Contemporary aspects of discrete peak
detection algorithms. II. The paradigm of the luteinizing hormone pulse signal in women.
Endocr Rev 13:81^104
Genazzani AR, Petraglia F, Purdy RH 1996 The brain: source and target for sex steroid
hormones. Parthenon, New York
Giustina A, Veldhuis JD 1998 Pathophysiology of the neuroregulation of growth hormone
secretion in experimental animals and the human. Endocr Rev 19:717^797
Kalra SP, Sahu A, Kalra PS 1993 Ageing of the neuropeptidergic signals in rats. J Reprod Fertil
Suppl 46:11^19
Loprinzi CL, Kugler JW, Sloan JA et al 2000 Venlafaxine in management of hot £ashes in
survivors of breast cancer: a randomized controlled trial. Lancet 356:2059^2063
Matt DW, Kauma SW, Pincus SM, Veldhuis JD, Evans WS 1998a Characteristics of luteinizing
hormone secretion in younger versus older premenopausal women. Am J Obstet Gynecol
178:504^510
Matt DW, Gilson MP, Sales TE et al 1998b Characterization of attenuated proestrous luteinizing
hormone surges in middle-aged rats by deconvolution analysis. Biol Reprod 59:1477^1482
Pincus SM, Mulligan T, Iranmanesh A, Gheorghiu S, Godschalk M, Veldhuis JD 1996 Older
males secrete luteinizing hormone and testosterone more irregularly, and jointly more
asynchronously, than younger males. Proc Natl Acad Sci USA 93:14100^14105
Pincus SM, Veldhuis JD, Mulligan T, Iranmanesh A, Evans WS 1997 E¡ects of age on the
irregularity of LH and FSH serum concentrations in women and men. Am J Physiol
273:E989^E995
Prior JC 1998 Perimenopause: the complex endocrinology of the menopausal transition. Endocr
Rev 19:397^428
Swartzman LC, Edelberg R, Kemmann E 1990 Impact of stress on objectively recorded
menopausal hot £ushes and on £ush report bias. Health Psychol 9:529^545
Thorner MO, Vance ML, Hartman ML et al 1990 Physiological role of somatostatin on growth
hormone regulation in humans. Metabolism 39:40^42
Veldhuis JD, Iranmanesh A, Mulligan T, Pincus SM 1999a Disruption of the young-adult
synchrony between luteinizing hormone release and oscillations in follicle-stimulating
hormone, prolactin, and nocturnal penile tumescence (NPT) in healthy older men. J Clin
Endocrinol Metab 84:3498^3505
192 DISCUSSION

Veldhuis JD, Iranmanesh A, Demers LM, Mulligan T 1999b Joint basal and pulsatile
hypersecretory mechanisms drive the monotropic follicle stimulating hormone (FSH)
elevation in healthy older men: concurrent preservation of the orderliness of the FSH
release process: a general chemical research center study. J Clin Endocrinol Metab 84:3506^
3514
Veldhuis JD, Iranmanesh A, Godschalk M, Mulligan T 2000 Older men manifest multifold
synchrony disruption of reproductive neurohormone out£ow. J Clin Endocrinol Metab
85:1477^1486
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

The thyroid axis in ageing

Holger Leitolf, Jens Behrends and Georg Brabant1

Department of Clinical Endocrinology, Center for Internal Medicine and Dermatology,

Medizinische Hochschule Hannover (MHH), 30625 Hannover, Germany

Abstract. The hypothalmo^pituitary^thyroid axis, among various endocrine systems,

undergoes physiological alterations associated with the ageing process. Directly age-
related changes have to be distinguished from indirect modi¢cations which are caused
by simultaneous thyroidal or non-thyroidal illness or other physiological or
pathophysiological states whose incidence increases with age. In summary, direct
changes of the hypothalmo^pituitary^thyroid axis seem to be subtle and suggestive of a
decreased hypothalamic stimulation of thyroid function. In parallel, disease-speci¢c
alterations such as the development of thyroid autonomy or changes in energy intake or
sleep lead to pronounced alterations of thyroid function with age which may dominate the
underlying ageing of the hypothalmo^pituitary^thyroid axis itself. The following article
attempts to delineate some aspects of the interplay of the regulation of thyroid function
and the ageing process.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 193^204

Ageing is associated with physiological alterations in various endocrine systems

including the hypothalamo^pituitary^thyroid axis. A number of studies have
attempted to delineate alterations in thyroid physiology due to the ageing
process (Felicetta 1988, Levy 1991, Mooradian & Wong 1994). When focusing
on age-dependent changes, alterations of the hypothalamo^pituitary^thyroid axis
have to be considered which are not directly age-related but rather indirectly caused
by simultaneous thyroidal or non-thyroidal illness or other physiological and
pathophysiological modi¢cations whose incidence increases with age.
Nevertheless, these events have important in£uences on thyroid physiology and
function.
Physiological regulation of the hypothalamo^pituitary^thyroid axis is
characterized by pulsatile stimulation of pituitary thyrotropin (TSH) secretion
resulting from an interplay of stimulatory hypothalamic in£uences mediated by

1
This chapter was presented at the symposium by Georg Brabant, to whom correspondence
should be addressed.

193
194 LEITOLF ET AL

thyrotropin-releasing-hormone (TRH) pulses as well as inhibitory hypothalamic

in£uences mediated by dopamine and somatostatin release. After being converted
from the prohormone tetraiodothyronine (or thyroxine, T4) by speci¢c
deiodinases, triiodothyronine (T3) inhibits synthesis and secretion of TRH
(Segerson et al 1987) and stimulates the release of inhibitory factors such as
dopamine and somatostatin from the respective hypothalamic centres. Within the
thyrotropic cells of the pituitary gland, T3 directly inhibits the synthesis of the
TSH b-subunit and thereby TSH secretion by interaction with speci¢c T3
receptors (Gurr & Kourides 1985). Thyrotropin is released from the pituitary
gland in a pulsatile pattern with approximately one pulse every two hours.
However, these pulses are not uniformly distributed but rather cluster within the
evening hours. Fusion of TSH pulses with increasing pulse amplitude, leads to the
nightly increase in mean TSH serum levels with a maximum between 0200 h and
0400 h (Brabant et al 1986, Greenspan et al 1986, Clark et al 1987, Table 1).
Many factors have been described which in£uence the secretory pattern of TSH.
Pulse amplitude is rapidly dampened by increasing circulating T4 levels in healthy
subjects. A similarly rapid e¡ect can be observed after bolus application of
glucocorticoids but, in contrast to thyroid hormones, this e¡ect appears to be
speci¢cally exerted on the hypothalamic level whereas the responsiveness of the
pituitary to TRH remains unaltered (Re et al 1976, Brabant et al 1987, 1989).
Physiological alterations such as sleep and energy supply also exert profound
e¡ects on the hypothalamo^pituitary^thyroid system. In healthy volunteers, 36
hours of fasting signi¢cantly decreases TSH levels and almost completely
suppresses the normal TSH increase during the night. Despite this profound
e¡ect, the pattern of TSH pulses and their frequency remains constant, indicating
a selective e¡ect on TSH pulse amplitude (Romijn et al 1990). As it is known that
leptin plasma levels decrease with fasting, these may also serve as potential
mediators. When starving rodents received intraperitoneal leptin injections twice
daily, the fasting-dependent decrease in hypothalamic^pituitary^thyroid axis was
reversed (Seoane et al 2000). Experimental sleep alterations have been shown to
modify TSH secretion and change hypothalamic and pituitary responsiveness.
Acute sleep withdrawal in healthy young subjects induces an activation of the
nightly increase in TSH secretion whereas a longer lasting sleep deprivation
period and sleep fragmentation induces a dramatic decrease in the mean 24 h
TSH secretion (Brabant et al 1990, Behrends et al 1998, Spiegel et al 1999).
The examples described above are important as they all may a¡ect the
hypothalamo^pituitary^thyroid axis in ageing and may thus obscure physio-
logical alterations. Moreover, in iodine-de¢cient regions such as Germany, the
prevalence of goitre is estimated to range between 30^50% with an increasing
frequency with age (Berghout et al 1990, Hintze et al 1991a, Hampel et al 1995).
Alterations in iodine supply are known to change thyroid function. Most likely due
THE THYROID AXIS IN AGEING 195

TABLE 1 Age-related changes of the thyroidal axis and the peripheral responsive-
ness to thyroid hormones

Parameter Change

TSH response to TRH = or decreased

Nocturnal TSH peak Decreased
Thyroid response to TSH = or decreased
Radioactive iodine uptake Decreased
T4 production = or decreased
5’ deiodinase activity Decreased
T3 production = or decreased
T4 / T3 degradation = or decreased
Serum thyroid hormone binding =
Total or free T4 serum concentration = or decreased
Total or free T3 serum concentration =
Reverse T3 serum concentration = or increased
Metabolic rate Decreased
Lipid peroxidation = or decreased
Malic enzyme, Na/K-ATPase, S14 Decreased

= , stays the same.

Adapted from Levy (1991).

to a higher sensitivity of the thyroid gland to TSH, at least in short-term

experiments, circulating TSH levels drop by approximately 50% without any
e¡ect on the frequency of TSH pulses when iodine supply is experimentally
decreased (Brabant et al 1992). Prolonged iodine de¢ciency results in a high
number of patients with nodular goitre estimated to account for more than 15%
of the population in Germany beyond 60 years of age. This percentage is even
further increased if only hospitalized subjects are considered (Felicetta 1988).
Overt or occult thyroid diseases have to be considered when assessing the
physiological alterations of the hypothalamo^pituitary^thyroid axis associated
with ageing, and their clinical presentation, at least in part, may di¡er
substantially from the symptoms and signs found in younger patients
(K˛bberling et al 1981, Hennemann & Krenning 1987, Nordyke et al 1988,
Trivalle et al 1996). In many studies, thyroid autonomy and autoimmune thyroid
disease have been identi¢ed as the leading causes for hyperthyroidism in ageing
(Trzepacz et al 1989, van Coevorden et al 1989, Brabant et al 1991, Greenspan et
al 1991, Levy 1991, Mariotti et al 1993, Mooradian & Wong 1994, Samuels 1998).
A large study of 583 healthy subjects including 34 centenarians revealed an
196 LEITOLF ET AL

age-dependent increase in thyroid-speci¢c antibodies with a signi¢cant peak

between 70 and 85 years, accompanied by a decrease in total and CD5+ B cells
(Mariotti et al 1992).
The predominant cause of hyperthyroidism in old age, however, is related to the
widespread therapeutic use of thyroid hormones. Data from the Framingham
study demonstrated in an unselected population of 2575 adults beyond the age of
60 (mean age 68.6 years) that 6.9% (2.3% of men and 10.0% of women) were
receiving medical treatment with thyroid hormones. It is interesting to note that
apart from the therapeutic use in goitre patients and as substitution therapy for
hypothyroidism, approximately 20% of the subjects studied were using thyroid
hormones without appropriate identi¢able cause. Insu⁄cient treatment could be
demonstrated in 37% of the hypothyroid patients using thyroid hormones by
elevated TSH serum levels (Sawin et al 1989). Approximately 6% of the group
with lowered TSH serum levels in the Framingham study were hyperthyroid
when all laboratory and clinical indicators were combined. This suggests that
lowered TSH serum levels in the elderly may be under a dominant in£uence of
other factors than the thyroid status alone (Sawin et al 1991). The ratio of ‘truly’
hyperthyroid subjects as compared to other causes of lowered thyrotropin serum
levels changes even more dramatically when the study is focused on hospitalized
patients. Our group recently investigated patients of a rehabilitation clinic after
treatment of acute illness (Brabant et al 1996). Restricting this prospective
analysis to those 619 patients with no previous suspicion of active thyroid disease
during the primary hospital stay, 5 subjects (0.8%) were identi¢ed to have elevated
TSH serum levels indicating hypothyroidism. As expected for a region with
de¢cient dietary iodine supply, the prevalence of suppressed TSH serum
concentrations was much higher (22.6%). In 19% of the subjects from this group
(i.e. 4% of the total group), overt hyperthyroidism was diagnosed supporting the
insidious clinical signs of thyroid dysfunction in elderly subjects (Stolte et al 1998,
Sawin et al 1994). This ¢ts to recent data in hospitalized patients where a frequency
of 0.8% overt clinical (4.2% subclinical) hypothyroidism was found in an iodine-
de¢cient region, whereas 1.5% of the patients (10.4%) in a region with su⁄cient
dietary iodine supply and 7.6% of the patients (23.9%) with high dietary iodine
supply could be shown to be clinically (subclinically) hypothyroid. In contrast,
clinical and subclinical hyperthyroidism was observed in 3.4% of the patients in a
region with iodine de¢ciency, 3.0% in an area with su⁄cient dietary iodine and 0%
in a region with high dietary iodine supply (Szabolcs et al 1997).
In large population studies, the incidence of hypothyroidism in elderly patients
varies from less than 1% to 17% depending on the iodine supply. Women are more
commonly a¡ected than men, and subclinical hypothyroidism is more frequent
than overt hypothyroidism. Virtually all the cases of hypothyroidism found are
related to autoimmune thyroid disease (Hintze et al 1991b).
THE THYROID AXIS IN AGEING 197

Apart from the iodine supply, thyroid function may be altered by the availability
of other micronutrients such as selenium in the thyroid gland itself as well as in
major target tissues for thyroid hormones such as the liver. Ravaglia et al (2000)
demonstrated lower serum selenium levels in a group of healthy subjects beyond 90
years of age as compared to younger subjects. This was accompanied by other
changes such as lower zinc serum concentrations. Many studies underline the
importance of selenium for normal bioactivity of the type II deiodinase (Beckett
et al 1993, Meinhold et al 1993, Mitchell et al 1997, Hotz et al 1997) at least in
serious iodine de¢ciency.
Total energy intake exerts a profound in£uence on thyroid function, mainly on
the hypothalamic level but also on the pituitary and thyroid level. The importance
of inadequate energy intake in ageing has been discussed in large population
studies. In the Euronut^Seneca study (de Groot et al 1992) on 2600 subjects born
between 1913 and 1918 in 18 European communities, body mass index was below
20 kg/m2 in up to 15% of elderly men and 17% of elderly women, in addition, the
daily energy intake in these subjects was found to be lower than their respective
requirements.
Similarly, in severe non-thyroidal illness (NTI) the decrease in mean TSH
secretion (Custro et al 1994) follows a comparable pattern to fasting with
signi¢cantly reduced TSH pulse amplitude but an unchanged pulse frequency
(Romijn et al 1990). Malnutrition may at least play a partial role in the
explanation of disturbed thyroid function in NTI. The prevalence of chronic
diseases in the Euronut^Seneca study varied between 59% and 92% (Danforth &
Burger 1989, de Groot et al 1992), thereby indicating that NTI may introduce a
signi¢cant bias when assessing age-dependent changes of the thyroid axis. Under
these pathophysiological conditions, the glucocorticoid axis may be activated.
This may suppress, to a variable degree, TSH secretion, an e¡ect potentially
important in physiological ageing where an activation of glucocorticoid
secretion has been previously reported (van Cauter et al 1996, Harper et al 1999,
Bergendahl et al 2000).
Finally, sleep may play an important role. It is well known that the sleep
pattern changes with age. With increasing age, total sleep duration decreases
and sleep fragmentation increases. In addition the relative time intervals in
slow wave sleep are decreased as less deep sleep stages and REM sleep
dominate (van Cauter et al 2000). When allowed to sleep only four hours
every 24 h over one week, thereby mimicking the sleep pattern of elderly
subjects, a profound e¡ect on TSH secretion with very low circulating TSH
levels at the end of the experiment could be demonstrated in young volunteers
(Spiegel et al 1999). This indicates that sleep-dependent processes not directly
related to ageing may change TSH secretion and the functional activity of the
thyroid axis.
198 LEITOLF ET AL

Most of the epidemiological studies on thyroid function in ageing are based on

population studies including subjects where the above mentioned mechanisms
may very well result in a signi¢cant study bias due to disease-related alterations
or independent age-related problems (Mariotti et al 1995). Reliable studies
focusing on healthy subjects are scarce. A careful evaluation of healthy
centenarians selected according to criteria of the Eurage Senieur protocol
(Mariotti et al 1993) revealed subtle changes of the thyroid axis. Free
triiodothyronine and TSH serum concentrations were slightly decreased in
parallel to a small but signi¢cant increase in reverse triiodothyronine serum
concentrations. Free thyroxine serum levels were unchanged when comparing a
group of 41 healthy centenarians with 98 healthy subjects aged 20^64 years or
with 33 subjects aged 65^80 years, respectively. The changes observed were so
subtle that they would not have been evident if only subjects up to 80 years of
age had been included. These results are compatible with a decline in
hypothalamic^pituitary activity during ageing resulting in a decreased TSH
secretion. Studies on the 24 h rhythm of TSH in young and elderly subjects
showed that the nocturnal TSH peak was blunted in the elderly (Barreca et al
1985, van Coevorden et al 1989, Greenspan et al 1991). Analysis of the pulsatile
release pattern of thyrotropin revealed that this change selectively rests on a
decreased TSH pulse amplitude whereas the temporal structure of the circadian
TSH secretion with a nadir in the late afternoon and a peak around midnight in
conjunction with an increased frequency of TSH pulses appears to be preserved
(Brabant et al 1991). Thyrotropin response to a de¢ned TRH challenge in the
study group, male subjects from 67^84 years of age, was attenuated, pointing to a
decreased pituitary TSH secretory reserve despite comparable triiodothyronine
and thyroxine serum (van Coevorden et al 1989), which would be compatible
with a hypothalamic or pituitary mechanism.
Animal studies in male Fisher rats suggest an age-related decrease in
hypothalamic TRH synthesis followed by a decreased formation of thyrotropin b
subunit within the pituitary gland (Cizza et al 1992). Studies in humans using the
dopamine antagonist metoclopramide indicated a decreased responsiveness of
pituitary TSH secretion to TRH (Targum et al 1989). This may be explained by a
change in the nocturnal increase of TSH secretion due to an increased nightly
dopamine tone (Greenspan et al 1991). As dopamine has direct inhibitory e¡ects
on pituitary TSH synthesis (Shupnik et al 1986), the attenuated response to TRH
may ¢t into this concept. As an alternative or additional mechanism, an altered set
point for the feedback action of thyroxine in elderly subjects was proposed, due to a
putatively increased pituitary T4 to T3 conversion (Lewis et al 1991) and a selective
decrease of free serum T3 levels (Mariotti et al 1993). Such age-related alterations in
the activity of T4 to T3 conversion have not only been described for the activity of
the 5’ deiodinase in the pituitary, but also in the liver (Donda & Lemarchand-
THE THYROID AXIS IN AGEING 199

Beraud 1989). On the contrary, TSH appears to stimulate deiodinase activity and
may thus counteract these e¡ects (K˛hrle 1990).
In summary, changes of the hypothalamic^pituitary^thyroid axis with age seem
to be subtle and suggest a decreased hypothalamic stimulation of thyrotropin
release. In parallel, disease-speci¢c alterations such as the development of thyroid
autonomy, or changes in energy intake or sleep may dominate these small
physiological changes and lead to pronounced alterations of thyroid function
with age, which are only indirectly related to ageing of the hypothalamo^
pituitary^thyroid axis itself.

References
Barreca T, Franceschini R, Messina V, Bottaro L, Rolandi E 1985 24 hour thyroid-stimulating
hormone secretory pattern in elderly men. Gerontology 31:119^123
Beckett GJ, Nicol F, Rae PW, Beech S Guo Y, Arthur JR 1993 E¡ects of combined iodine and
selenium de¢ciency on thyroid hormone metabolism in rats. Am J Clin Nutr 57:240S^243S
Behrends J, Prank K, Dogu E, Brabant G 1998 Central nervous system control of thyrotropin
secretion during sleep and wakefulness. Horm Res 49:173^177
Bergendahl M, Iranmanesh A, Mulligan T, Veldhuis JD 2000 Impact of age on cortisol secretory
dynamics basally and as driven by nutrient-withdrawal stress. J Clin Endocrinol Metab
85:2203^2214
Berghout A, Wiersinga WM, Smits NJ, Touber JL 1990 Interrelationships between age, thyroid
volume, thyroid nodularity, thyroid function in patients with sporadic nontoxic goitre. Am J
Med 89:602^608
Brabant A, Brabant G, Schˇrmeyer T et al 1989 The role of glucocorticoids in the regulation of
thyrotropin. Acta Endocrinol 121:95^100
Brabant G, Ranft U, Ocran K, Hesch RD, von zur Mˇhlen A 1986 Thyrotropin
an
episodically secreted hormone. Acta Endocrinol 112:315^322
Brabant G, Brabant A, Ranft U et al 1987 Circadian and pulsatile thyrotropin secretion in
euthyroid man under the in£uence of thyroid hormone and glucocorticoid administration. J
Clin Endocrinol Metab 65:83^88
Brabant G, Prank K, Ranft U et a 1990 Physiological regulation of circadian and pulsatile
thyrotropin secretion in normal man and woman. J Clin Endocrinol Metab 70:403^409
Brabant G, Prank K, Hoang-Vu C, Hesch RD, von zur Mˇhlen A 1991 Hypothalamic
regulation of pulsatile thyrotopin secretion. J Clin Endocrinol Metab 72:145^150
Brabant G, Bergmann P, Kirsch CM, K˛hrle J, Hesch RD, von zur Mˇhlen A 1992 Early
adaptation of thyrotropin and thyroglobulin secretion to experimentally decreased iodine
supply in man. Metabolism 41:1093^1096
Brabant G, Mayr B, Lucke C 1996 Klinische und laborchemische Aspekte der latenten und
manifesten Hyperthyreose im h˛heren Lebensalter. Schilddrˇse 1995, deGruyter, Berlin
S241^S251
Cizza G, Brady LS, Calogero AE et al 1992 Central hypothyroidism is associated with advanced
age in male Fischer 344/N rats: in vivo and in vitro studies. Endocrinology 131:2672^2680
Clark PMS, Clark JDA, Holder R, Raggatt PR, Wheatley T 1987 Pulsatile secretion of TSH in
healthy subjects. Ann Clin Biochem 24:470^476
Custro N, Sca¢di V, Gallo S, Notarbartolo A 1994 De¢cient pulsatile thyrotropin secretion in
the low-thyroid-hormone state of severe non-thyroidal illness. Eur J Endocrinol 130:132^136
200 LEITOLF ET AL

Danforth E Jr, Burger AG 1989 The impact of nutrition on thyroid hormone physiology and
action. Ann Rev Nutr 9:201^227
de Groot LC, Hautvast JG, van Staveren WA 1992 Nutrition and health of elderly people in
Europe: the Euronut-Seneca study. Nutr Rev 50:185^194
Donda A, Lemarchand-Beraud T 1989 Aging alters the activity of 5’-deiodinase in the
adenohypophysis, thyroid gland, and liver of the male rat. Endocrinology 124:1305^1309
Felicetta JV 1988 The Thyroid and Aging. In: Sowers JR, Felicetta JV (eds) The endocrinology
of aging. Raven Press, New York, p 15^39
Greenspan SL, Klibanski A, Schoenfeld D, Ridgway EC 1986 Pulsatile secretion of thyrotropin
in man. J Clin Endocrinol Metab 63:661^668
Greenspan SL, Klibanski A, Rowe JW, Elahi D 1991 Age-related alterations in pulsatile
secretion of TSH: role of dopaminergic regulation. Am J Physiol 260:E486^E491
Gurr JA, Kourides IA 1985 Thyroid hormone regulation of thyrotropin alpha- and beta-subunit
gene transcription. DNA 4:301^307
Hampel R, Kˇlberg T, Klein K et al 1995 [Goiter incidence in Germany is greater than
previously suspected.] Med Klin 90:324^329
Harper AJ, Buster JE, Casson PR 1999 Changes in adrenocortical function with aging and
therapeutic implications. Semin Reprod Endocrinol 17:327^338
Hennemann G, Krenning EP 1987 Pitfalls in the interpretation of thyroid function tests in old
age and non-thyroidal illness. Horm Res 26:100^104
Hintze G, Windeler J, Baumert J, Stein H, K˛bberling J 1991a Thyroid volume and goitre
prevalence in the elderly as determined by ultrasound and their relationships to laboratory
indices. Acta Endocrinol 124:12^18
Hintze G, Burghardt U, Baumert J, Windeler J, K˛bberling J 1991b Prevalence of thyroid
dysfunction in elderly subjects from the general population in an iodine de¢ciency area.
Aging 3:325^331
Hotz CS, Fitzpatrick DW, Trick KD, L’Abbe¤ MR 1997 Dietary iodine and selenium interact to
a¡ect thyroid hormone metabolism of rats. J Nutr 127:1214^1218
K˛bberling J, Hintze G, Blossey HC 1981 Diagnostische Probleme der Hyperthyreose im
h˛heren Lebensalter. Dtsch med Wschr 106:973^978
K˛hrle J 1990 Thyrotropin (TSH) action on thyroid hormone deiodination and secretion:
one aspect of thyrotropin regulation of thyroid cell biology. Horm Metab Res Suppl 23:
18^28
Levy EG 1991 Thyroid disease in the elderly. Med Clin North Am 75:151^167
Lewis GF, Alessi CA, Imperial JG, Refeto¡ S 1991 Low serum free thyroxine index in
ambulating elderly is due to a resetting of the threshold of thyrotropin feedback
suppression. J Clin Endocrinol Metab 73:843^849
Mariotti S, Sansoni P, Barbesino G et al 1992 Thyroid and other organ-speci¢c autoantibodies in
healthy centenarians. Lancet 339:1506^1508
Mariotti S, Barbesino G, Caturegli P et al 1993 Complex alteration of thyroid function in healthy
centenarians. J Clin Endocrinol Metab 77:1130^1134
Mariotti S, Franchesi C, Cossarizza A, Pinchera A 1995 The aging thyroid. Endocr Rev 16:
686^715
Meinhold H, Campos-Barros A, Walzog B, K˛hler R, Mˇller F, Behne D 1993 E¡ects of
selenium and iodine de¢ciency on type I, type II and type III iodothyronine deiodinases and
circulating thyroid hormones in the rat. Exp Clin Endocrinol 101:87^93
Mitchell JH, Nicol F, Beckett GJ, Arthur JR 1997 Selenium and iodine de¢ciencies: e¡ects on
brain and brown adipose tissue selenoenzyme activity and expression. J Endocrinol 155:
255^263
Mooradian AD, Wong NC 1994 Age-related changes in thyroid hormone action. Eur J
Endocrinol 131:451^461
THE THYROID AXIS IN AGEING 201

Nordyke RA, Gilbert FI Jr, Harada AS 1988 Graves’ disease. In£uence of age on clinical
¢ndings. Arch Intern Med 148:626^631
Ravaglia G, Forti P, Maioli F et al 2000 Blood micronutrient and thyroid hormone
concentrations in the oldest-old. J Clin Endocrinol Metab 85:2260^2265
Re RN, Kourides IA, Ridgway EC, Weintraub BD, Maloof F 1976 The e¡ect of glucocorticoid
administration on human pituitary secretion of thyrotropin and prolactin. J Clin Endocrinol
Metab 43:338^346
Romijn JA, Adriaanse R, Brabant G, Prank K, Endert E, Wiersinga WM 1990 Pulsatile
secretion of thyrotropin during fasting: a decrease of thyrotropin pulse amplitude. J Clin
Endocrinol Metab 70:1631^1636
Samuels MH 1998 Subclinical thyroid disease in the elderly. Thyroid 8:803^813
Sawin CT, Geller A, Hershman JM, Castelli W, Bacharach P 1989 The aging thyroid. The use of
thyroid hormone in older persons. JAMA 261:2653^2655
Sawin CT, Geller A, Kaplan MM, Bacharach P, Wilson PW, Hershman JM 1991 Low serum
thyrotropin (thyroid-stimulating hormone) in older persons without hyperthyroidism. Arch
Intern Med 151:165^168
Sawin CT, Geller A, Wolf PA et al 1994 Low serum thyrotropin concentrations as a risk factor
for atrial ¢brillation in older persons. N Engl J Med 331:1249^1252
Segerson TP, Kauer J, Wolfe HC et al 1987 Thyroid hormone regulates TRH biosynthesis in the
paraventricular nucleus of the rat hypothalamus. Science 238:78^80
Seoane LM, Carro E, Tovar S, Casanueva FF, Dieguez C 2000 Regulation of in vivo TSH
secretion by leptin. Regul Pept 92:25^29
Shupnik MA, Greenspan SL, Ridgway EC 1986 Transcriptional regulation of thyrotropin
subunit genes by thyrotropin-releasing hormone and dopamine in pituitary cell culture. J
Biol Chem 261:12675^12679
Spiegel K, Leproult R, van Cauter E 1999 Impact of sleep debt on metabolic and endocrine
function. Lancet 354:1435^1439
Stolte H, von zur Mˇhlen A, Lucke C, Brabant G 1998 Atrial ¢brillation and hyperthyroidism in
geriatric patients
a prospective study in acute disease. J Endocrinol Invest 21:45
Szabolcs I, Podoba J, Feldkamp J et al 1997 Comparative screening for thyroid disorders in old
age in areas of iodine de¢ciency, long-term iodine prophylaxis and abundant iodine intake.
Clin Endocrinol 47:87^92
Targum SD, Marshall LE, Magac-Harris K, Martin D 1989 TRH tests in a healthy elderly
population. Demonstration of gender di¡erences. J Am Geriatr Soc 37:533^536
Trivalle C, Doucet J, Chassagne P et al 1996 Di¡erences in the signs and symptoms of
hyperthyroidism in older and younger patients. J Am Geriatr Soc 44:50^53
Trzepacz PT, Klein I, Roberts M, Greenhouse J, Levey GS 1989 Graves’ disease: an analysis of
thyroid hormone levels and hyperthyroid signs and symptoms. Am J Med 87:558^561
van Cauter E, Leproult R, Kupfer DJ 1996 E¡ects of gender and age on the levels and circadian
rhythmicity of plasma cortisol. J Clin Endocrinol Metab 81:2468^2473
van Cauter E, Leproult R, Plat L 2000 Age-related changes in slow wave sleep and REM sleep
and relationship with growth hormone and cortisol levels in healthy men. JAMA 284:861^
868
van Coevorden A, Laurent E, Decoster C et al 1989 Decreased basal and stimulated thyrotropin
secretion in healthy elderly men. J Clin Endocrinol Metab 69:177^185

DISCUSSION
Carroll: Your wonderful data bring us back to some of the points made earlier,
about nocturnal cortisol in ageing and stress. Can you tell us what the end of the
line functional status of the thyroid is in the elderly in terms of basal metabolic rate
202 DISCUSSION

(BMR)? Is there decreased BMR or is there a change in set point of body

temperature?
Brabant: I don’t have good data on body temperature. For BMR, you have to
bear in mind that it changes with body weight. You have to stratify for similar
weight and thyroid status, and to my knowledge this is not normally done. But it
adds another level of control and regulation. For example, leptin changes BMR, as
shown in rodent models. As you can rescue the thyroid axis in fasting with leptin,
there is an evident integration of systems. But to date, the sequence of events has
not been determined.
Veldhuis: The other confounder is that both growth hormone (GH) and
androgens increase BMR. In the male, both of these are going to be falling with
age.
Handelsman: There are always issues about how one constitutes a sample of
healthy elderly. If you looked at the original data, and if the disease wasn’t so
rigorously excluded, how much of an e¡ect is there? Is it a very marked e¡ect of
disease on thyroid function? Related to this, have there been any studies with
thyroxine supplementation in just the elderly?
Brabant: We need to discuss separately whether we have thyroid pathology or
hypothyroidism (as is common in the USA in population studies with rates of up to
10%) or we have hyperthyroidism (as we have in Germany). There we have de¢nite
data to show that correction is important. There are some German data, recently
generated by our group, showing 8^10% hyperthyroidism. If you include the
subclinical forms, there is a 1^2% rate of hypothyroidism.
Handelsman: I was thinking more of non-thyroidal hormone exclusions.
Brabant: The exclusion of non-thyroidal illness is not easy. It is estimated in large
population studies to add up to 15^20%.
Laron: We know that hypothyroidism reduces secretion of GH. If we don’t see
obvious hypothyroidism in old people, can we conclude that the reduction in GH/
insulin-like growth factor (IGF)1 is not thyroid dependent?
Brabant: The observation that population studies show constant thyroid
hormone levels up to age 80 but a prominent decrease in GH/IGF1 argues
against such interrelations. However, treating this large population of
hypothyroidal subjects may have e¡ects but I am not aware of data where this has
been formally investigated, and I do not know which change in thyroid hormone
concentrations will elicit alterations in GH/IGF1.
Shalet: I don’t think anyone has ever de¢ned the threshold level of thyroid
hormone that has a signi¢cant e¡ect on GH secretion.
Brabant: There is a huge range of normal thyroid hormone concentrations. In
our lab 4.5^12 mg/dl, a threefold variation, is considered to be normal. There are
individual set points that di¡er between individuals. It is di⁄cult to de¢ne the
thyroid status we would like to aim for.
THE THYROID AXIS IN AGEING 203

Ruiz-Torres: Two thyroid parameters are more or less dependent on age. Both
TSH and T3 decrease with age. But what role does the thyroid have in ageing?
Brabant: The data are con£icting. In the studies I referred to in my paper we ¢nd
TSH ¢rst goes up and then decreases. If TSH is increasing you have a more
prominent e¡ect from the thyroid itself. You may get primary hypothyroidism
which is sensed by the hypothalamic^pituitary complex, and responds by
increasing in TSH. In contrast, in Marrioti’s study there is a decrease in TSH by
concentrations of peripheral thyroid hormones, which means there is a dominant
defect in the hypothalamic^pituitary axis. But this doesn’t exclude the possibility
that there might be some additional problems with the thyroid. I deliberately put in
the data on thyroid autonomy. It takes so little in a country that is as iodine de¢cient
as Italy and Germany to really get some subclinical thyroid autonomy, with a
perfectly normal situation if you do the readings for T3 and T4, but a slightly
lowered TSH serum concentration. Is this already clinical hyperthyroidism?
Probably not, but there is certainly some problem in the hypothalamo^pituitary
interaction, with a primary change in the thyroid.
Ruiz-Torres: In old age, there is remarkable tendency to get hypothyroidism,
which shows the same manifestations as ageing itself.
Brabant: From the data up to the 90s there isn’t profound hypothyroidism.
There is an intensive discussion about whether subclinical hypothyroidism
should be treated. For subclinical hyperthyroidism the situation is less
controversial, because atrial ¢brillation increases threefold and cardiovascular
consequences develop. Anyway, I don’t think mild hypothyroidism explains
ageing, especially as we only see the condition in centenarians.
Haus: We have studied 194 clinically healthy children living in the endemic
goitre area of Romania and standardized in their daily routine by their school
activity. 60% of the children had endemic goitre, but were clinically and
chemically euthyroid. We compared these children with a group of 284 elderly
subjects between 60 and 90+ years of age also living in Romania. The circadian
mean of plasma TSH was signi¢cantly higher in the elderly than in the children
due to higher circadian trough values, while the circadian peak values were
essentially the same (Nicolau & Haus 1992). The circulating thyroid hormones
also showed low-amplitude circadian rhythms with a di¡erence between the
children and the elderly in the levels as well as in timing. The TSH peak in the
children usually occurred during the night, and the peak of the circulating
thyroid hormones followed in the morning. In contrast, in the elderly, although
the peak in TSH occurred also during night hours there was a delay of several hours
of the peak times (acrophases) of the thyroid hormones. It appears that in the
elderly higher levels of TSH lead to a lower output of thyroid hormones than in
children and after the nightly peak in TSH the thyroid responds slower with a
phase delay in the thyroid hormone output.
204 DISCUSSION

Another ¢nding was a di¡erence in the seasonal variation of the pituitary^

thyroid axis between children and the elderly. In the elderly, the seasonal
amplitude was substantially smaller and a seasonal (circannual) variation was
present only up to the eighth decade (Haus et al 1988). Beyond this it was absent.
The elderly also showed a di¡erence in the mix of the circulating thyroid
hormones with a statistically signi¢cantly lower T3:T4 ratio in the elderly than in
young adults or children (Dumitriu et al 1996a,b).

References
Dumitriu L, Bogdan C, Nicolau GY et al 1996a Seasonal variations in TSH, TT3, TT4, log
TSH/TT4 and T3/T4 ratio in elderly subjects. Proceedings of the 10th International
Congress of Endocrinology, June 12^15, San Francisco, CA, Abstract P2-1059:669
Dumitriu L, Plinga L, Nicolau GY et al 1996b Seasonal variations in iodine excretion, plasma
TSH, TT3, TT4, long TSH/TT4 and T3/T4 ratio in children with and without endemic
goiter. Proceedings of the 10th International Congress of Endocrinology, June 12^15, San
Francisco, CA, Abstract P2-1058:669
Haus E, Nicolau GY, Lakatua D, Sackett-Lundeen L 1988 Reference values for
chronopharmacology. Annu Rev Chronopharmacol 4:333^424
Nicolau GY, Haus E 1992 Chronobiology of the hypothalamic-pituitary-thyroid (HRT) axis.
In: Touitou Y, Haus E (eds) Biological rhythms in clinical and laboratory medicine. Springer-
Verlag, Heidelberg p 330^347
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Critical illness as a model of

hypothalamic ageing
Greet Van den Berghe and Stephen M. Shalet*1

Department of Intensive Care Medicine, University Hospital, Gasthuisberg, Catholic University

of Leuven, Belgium and *Department of Endocrinology, Christie Hospital, Manchester, UK

Abstract. In this review we shall consider the endocrine changes seen in the various
hypothalamic pituitary-target gland axes at di¡erent stages of critical illness and
conclude by comparing these changes with those seen in the normal ageing process.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 205^221

By de¢nition, ‘critical illness’ is any condition requiring support for failing vital
organ functions, either with mechanical aids or with pharmacological agents,
without which death would ensue
it is the ultimate example of acute severe
physical stress. If onset of recovery does not follow within several days of
intensive medical care, critical illness often becomes prolonged and intensive care
must be continued for weeks or even months.
Patients previously died from these severe challenges, which include septic
shock, multiple trauma or extensive burns. However, the survival mediated by
intensive care is also associated with negative aspects. The highly technological
intervention in the natural course of the dying process has unmasked previously
unknown conditions, including a non-speci¢c wasting syndrome: despite feeding,
protein continues to be lost from vital organs and tissues due to increased
proteolysis and decreased protein synthesis, whereas re-esteri¢cation of free fatty
acids (FFAs) allows fat stores to build up (Streat et al 1987, Gamrin et al 1996).
Moreover the muscle wasting is accompanied by hyperglycaemia, insulin
resistance, hypoproteinaemia, hypercalcaemia, intracellular water and potassium
depletion, and hypertriglyceridaemia, which often prompt symptomatic
treatment.

1This paper was presented at the symposium by Stephen Shalet, to whom correspondence should

be addressed.

205
206 VAN DEN BERGHE & SHALET

Growth hormone axis

In the acute phase circulating levels of growth hormone (GH) are elevated. The
peak GH levels as well as interpulse concentrations are high (Ross et al 1991,
Voerman et al 1992) and the GH pulse frequency is increased. Serum insulin-like
growth factor (IGF)1 concentrations are low, (Ross et al 1991). The combination
of high GH levels and low IGF1 levels has been interpreted as resistance to GH,
which may be related to decreased GH receptor expression (Hermansson et al
1997). The other GH-dependent peptides IGF binding protein (IGFBP)3 and
acid labile subunit (ALS) are also decreased in the circulation (Baxter 1997,
Timmins et al 1996), preceded by a drop in the GH binding protein (GHBP).
Circulating levels of the small IGFBPs such as IGFBP1, IGFBP2 and IGFBP6
are elevated (Baxter et al 1998, Rodr|¤ gues-Arnao et al 1996). It has been
suggested that these changes are brought about by the e¡ects of cytokines such as
tumour necrosis factor (TNF)a, interleukin (IL)1 and IL6. The hypothesis is that
reduced GH receptor expression and thus low IGF1 levels are the primary events
(cytokine-induced) which, in turn, through reduced negative feedback inhibition,
induce the abundant release of GH during acute stress, exerting direct lipolytic,
insulin-antagonizing and immune-stimulating actions, while the indirect IGF1-
mediated e¡ects of GH are attenuated. This explanation is plausible in that such
changes would prioritize essential substrates such as glucose, FFAs and amino
acids (glutamine) towards survival rather than anabolism. Increased IGFBP3
protease activity in plasma has also been reported, however, and is thought to
result in increased dissociation of IGF1 from the ternary complex, thereby
shortening the IGF1 half-life in the circulation (Baxter 1997, Gibson & Hinds
1997).
In prolonged critical illness (PCI) the pattern of GH secretion is very chaotic and
the quantity of GH released in pulses is much reduced compared with the acute
phase (Van den Berghe et al 1997a, 1998, 1999). Furthermore, although the non-
pulsatile fraction is still elevated and the number of pulses is still frequent, mean
nocturnal GH concentrations are scarcely elevated compared with the healthy non-
stressed condition. The mean nocturnal GH level is about 1 mg/l, though levels are
easily detectable and peak GH levels hardly ever exceed 2 mg/l; these results are
surprisingly, independent of age, gender, body composition and type of
underlying disease (Fig. 1).
The pulsatile component of GH secretion, which is substantially reduced,
correlates positively with circulating levels of IGF1, IGFBP3 and ALS, all of
which are low (Van den Berghe et al 1997a, 1998, 1999). Thus the more pulsatile
GH secretion is suppressed, the lower circulating levels of the GH-dependent
IGF1 and ternary complex binding proteins become. This clearly no longer
represents a state of GH resistance! The elevated serum levels of GHBP, assumed
HYPOTHALAMIC AGEING 207

FIG. 1. Nocturnal serum concentration pro¢les of GH, TSH and PRL illustrating the
di¡erences between the initial phase (interrupted line) and the chronic phase (dotted line) of
critical illness within an intensive care setting. The continuous lines illustrate normal patterns.

to re£ect GH receptor expression in peripheral tissues, in PCI patients are in line

with recovery of GH-responsiveness with time during severe illness. The low
levels of IGF1, IGFBP3, ALS and IGFBP5 are tightly related to biochemical
markers of impaired anabolism such as low serum osteocalcin and leptin (Van
den Berghe et al 1999). These ¢ndings suggest that relative GH de¢ciency,
epitomised by reduced pulsatile GH secretion participates in the pathogenesis of
the ‘wasting syndrome’ especially in the chronic phase of critical illness.
Furthermore there is a gender dissociation in that men show a greater loss of
pulsatility and regularity within the GH secretory pattern than women (despite
indistinguishable total GH output) and concomitantly lower IGF1 and ALS
levels (Van den Berghe et al 2000). It remains unknown whether the sexual
dimorphism within the GH^IGF1 axis and the fact that males seem to be at
higher risk for an adverse outcome from PCI than females is a casual or causal
association.
The pathogenesis of the secretory pattern of GH in PCI is probably complex.
One of the possibilities is a de¢ciency of the endogenous GH releasing peptide
(GHRP)-like ligand together with reduced somatostatin tone and maintenance of
some GH releasing hormone (GHRH) e¡ect. In reality the GH responses to a bolus
injection of GHRP are exuberant in long-stay intensive care patients and several-
fold higher than the response to GHRH, the latter being normal or often
subnormal. GHRH and GHRP evoke a clear synergistic response under these
circumstances (Van den Berghe et al 1996), revealing the highest GH response
ever reported in a human study (Fig. 2). The pronounced GH responses to
secretagogues exclude the possibility that the blunted GH secretion during the
chronic phase of critical illness is due either to a lack of pituitary capacity to
synthesise GH or to exaggerated somatostatin-induced suppression of GH
208 VAN DEN BERGHE & SHALET

FIG. 2. (Upper part) Nocturnal serum GH pro¢les in the prolonged phase of illness
illustrating the e¡ects of continuous infusion of placebo, GHRH (1 mg/kg per h), GHRP2
(1 mg/kg per h) or GHRH plus GHRP2 (1+1 mg/kg per h). The age range of the patients was
62^85 yrs; the duration of illness was between 13^48 days: infusions were started 12 h before the
onset of the respective pro¢le. (Lower part) Exponential regression lines have been reported
between pulsatile GH secretion and the changes in circulating IGF1, ALS and IGFBP3
obtained with a 45 h infusion of either placebo, GHRP2 or GHRH plus GHRP2. They
indicate that the parameters of GH responsiveness increase in proportion to GH secretion up
to a point beyond which a further increase in GH secretion has apparently little or no additional
e¡ect. In the chronic non-thriving phase of critical illness, GH sensitivity is clearly present in
contrast to the acute phase of illness, which is thought to be primarily a condition of GH
resistance (Van den Berghe et al 1998).

release. The combination of reduced somatostatin tone and de¢ciency of an

endogenous GHRP-like ligand would explain the reduced GH burst amplitude,
the increased frequency of spontaneous GH secretory bursts, and the elevated
interpulse levels as well as the striking responsiveness to GHRP alone or in
combination with GHRH. Females with PCI have a markedly greater GH
response to a bolus of GHRP compared with males, a di¡erence which is
eliminated when GHRH is injected together with GHRP. Lower endogenous
GHRH action in men with PCI possibly due to the concomitant profound
testosterone de¢ciency (Van den Berghe et al 2000) accompanying loss of action
of an endogenous GHRP-like ligand with prolonged stress in both genders, may
explain this ¢nding.
HYPOTHALAMIC AGEING 209

TSH^thyroid axis

Acute illness or trauma induces alterations in thyroid hormone equilibrium within

hours. Although serum TSH usually remains normal, circulating T3 rapidly drops,
partly due to reduced conversion of T4 to T3 and/or increased turnover of thyroid
hormones. The extent of the T3 drop within 24 hours re£ects the severity of illness
(Schlienger et al 1991, Rothwell & Lawler 1995). Serum reverse T3 (rT3) levels
increase partly due to reduced rT3 degradation. In animal models hepatic nuclear
T3 receptors appear to decrease in number and occupancy. The absence of a TSH
elevation in the presence of low circulating T3 levels suggests that there is also an
altered feedback setting at the hypothalamic^pituitary level. Experimental data
indicate that reduced TRH gene expression as well as enhanced nuclear T3
receptor occupancy within the thyrotrophs may be involved (Kakucska et al
1994, St Germain & Galton 1985).
The mechanism responsible for the low T3 syndrome remains speculative.
Increased levels of cytokines TNF-a, IL1 and IL6, increased endogenous thyroid
hormone analogues resulting from alternative deamination and decarboxylation,
and low concentrations of binding proteins and inhibition of hormone binding,
transport and metabolism by elevated levels of FFA and bilirubin, have all been
proposed as factors contributing to the low T3 syndrome at tissue level, although
de¢nitive proof is awaited.
In the prolonged phase of critical illness, the situation changes; pulsatile TSH
secretion is diminished (Fig. 1) and positively related to the low serum T3 levels
(Van den Berghe et al 1997b, 1998), suggesting that the reduced production of
thyroid hormone at this time may re£ect neuroendocrine dysfunction. Consistent
with this suggestion, hypothalamic TRH gene expression is positively related to
serum T3 and, in addition, an increase in serum TSH is a marker of the onset of
recovery from severe illness. The exact mechanisms underlying the neuro-
endocrine dysfunction responsible for low thyroid hormone levels in PCI remain
unknown, but the low thyroid hormone levels, per se, cannot be interpreted as a
bene¢cial response; low thyroid hormone levels have been found to correlate
inversely with urea production and bone degradation and therefore do not re£ect
an adaptive protective mechanism against hypercatabolism. Thus the restoration
of physiological levels of thyroid hormones by continuously infusing TRH,
together with a GH secretagogue, reduces hypercatabolism, an e¡ect related only
to changes in the thyroid hormone levels. During TRH infusion, the negative
feedback exerted by thyroid hormones upon the thyrotrophs was maintained,
thus avoiding overstimulation of the thyroid axis, which would inadvertently
aggravate catabolism. The co-infusion of TRH and GH-releasing factors appears
to be a better strategy than the infusion of TRH alone, since the combination but
not TRH alone, increases the pulsatile fraction of TSH release and avoids a rise in
210 VAN DEN BERGHE & SHALET

circulating rT3 (Van den Berghe et al 1998). The latter observation may re£ect the
well-known e¡ect of GH on type 1 deiodinase activity. It remains unknown
however if the normalization of circulating and tissue T3 concentrations has any
bene¢cial clinical consequences. Thus far pioneering studies with T4
administration have failed to demonstrate clinical bene¢ts within an intensive
care setting but in view of the impaired conversation of T4 to T3 this is not really
surprising. In this context a recent report on thyroid hormone treatment during the
entire stay on the intensive care unit involving substitution doses of T3 in
paediatric patients after correction of congenital heart defects revealed
improvement in post-operative cardiac function. In contrast to treatment with
thyroid hormones, infusing TRH encourages appropriate changes in thyroid
hormone metabolism thereby normalizing thyroid hormone levels in the
circulation and tissues (Van den Berghe et al 1998). Outcome bene¢t of TRH
infusion alone or in combination with GH secretagogues in PCI has yet to be
studied.

Prolactin
In response to acute physical or psychological stress the circulating prolactin level
rises (Noel et al 1972), a response that may be mediated by vasoactive intestinal
peptide, oxytocin, dopaminergic pathways and/or other still uncharacterized
factors; cytokines may also play a signalling role. Although prolactin appears to
have immunostimulatory properties in animal models as well as in humans, it
remains unclear whether the relative hyperprolactinaemia during the initial phase
of critical illness or post-trauma contributes to the initial activation of the
in£ammatory cascade.
In the chronic phase of critical illness, serum prolactin levels are no longer as
high as in the acute phase and the secretory pattern is characterised by a reduced
pulsatile fraction (Fig. 1) (Van den Berghe et al 1998, Gardner et al 1979). A role for
endogenous dopamine has been suggested. It is unknown whether the blunted
prolactin secretion in the chronic phase plays a role in the anergic immune
dysfunction or in the increased susceptibility for infections characterising the
chronically ill (Meakins et al 1977). However, exogenous dopamine often infused
as an inotropic drug in intensive care-dependent patients has been shown to further
suppress prolactin secretion and was found to aggravate concomitantly both T
lymphocyte dysfunction and impaired neutrophil chemotaxis (Van den Berghe
1994a).

Luteinizing hormone^testosterone axis

The low serum testosterone concentrations despite elevated luteinizing hormone
(LH) levels documented during the acute stress of surgery or myocardial infarction
HYPOTHALAMIC AGEING 211

suggest an immediate stress-induced Leydig cell suppression (Wang et al 1978a,b,

Dong et al 1992), the exact cause of which remains obscure. A role for
in£ammatory cytokines (IL1 and IL2) is possible, as suggested by experimental
studies. It may be considered appropriate that the secretion of anabolic
androgens be switched o¡ in circumstances of acute stress, in order to reduce the
consumption of energy and substrates at such a critical life-threatening time.
As critical illness becomes prolonged, hypogonadotropic hypogonadism ensues
(Vogel et al 1985, Woolf et al 1985). Circulating levels of testosterone become
extremely low, often undetectable in men whereas oestradiol concentra-
tions are reduced to a lesser degree, thereby increasing the oestradiol:testosterone
molar ratio. The progressive decline in serum gonadotropin levels, however,
appears to lag behind the rapid decline in serum testosterone. In men with PCI, a
high LH pulse frequency with an abnormally low LH pulse amplitude is seen and
this has been interpreted as an impaired LH response to very low circulating
testosterone levels (Van den Berghe et al 1994b). Endogenous dopamine, opiates
IL1, and the ‘relatively spared’ oestradiol level may be involved in the pathogenesis
of the gonadotropin de¢ciency. Androgen treatment in men a¡ected by prolonged
critical illness failed to induce conclusive clinical bene¢t (Tweedle et al 1972). An
alternative approach with exogenous gonadotropin-releasing hormone (GnRH)
has proved a little more promising in that the gonadotropin de¢ciency could be
partially reversed with pulsatile GnRH administered intravenously (Van den
Berghe et al 2001). The inability to completely reverse the profound
hypogonadotropic hypogonadal state may be due to relative pituitary
desensitization to GnRH or enhanced-feedback inhibition from the markedly
elevated oestradiol:testosterone molar ratio. Peripheral tissues were sensitive to
transient changes in sex steroids, as re£ected by anabolic and in£ammatory
responses (Van den Berghe et al 2001).

Pituitary^adrenal axis
It has been known for a long time that the vital stress-induced hypercortisolism
induced by surgery, trauma or sepsis is associated with augmented
adrenocorticotropic hormone (ACTH) release, driven presumably by
corticotropin-releasing hormone (CRH), cytokines and the noradrenergic
system. Concomitantly, circulating aldosterone rises markedly probably under
the control of an activated renin-angiotensin system. Hypercortisolism acutely
shifts carbohydrate, fat and protein metabolism, so that energy is instantly and
selectively available to vital organs such as the brain and so that anabolism is
delayed.
In PCI serum ACTH levels are low, while cortisol levels remain elevated,
indicating that cortisol release may in this phase be driven through an alternative
212 VAN DEN BERGHE & SHALET

pathway possibly involving endothelin (Vermes et al 1995). The mechanism

responsible for low ACTH levels in PCI is unknown although a role for atrial
natriuretic peptide or substance P has been suggested. In contrast, circulating
levels of adrenal androgens such as dehydroepiandrosterone sulfate (DHEAS)
(Van den Berghe et al 1995) and the mineralocorticoid, aldosterone, are reduced,
despite increased renin activity (Zipser et al 1981) in PCI. This steroid pro¢le
suggests a shift of pregnenolone metabolism away from both mineralocorticoid
and adrenal androgen pathways towards the glucocorticoid pathway. The
mechanism responsible for this change in steroid secretory pattern is unknown
but ultimately it may fail thereby accounting for the 20-fold higher incidence of
adrenal insu⁄ciency in critically ill patients over the age of 50 years and being
treated on the intensive care unit for more than 14 days. The fact that this type of
relative adrenal failure coincides with adverse outcomes suggests that high levels of
glucocorticoids remain essential for haemodynamic stability. Whether
hypercortisolium in the chronic phase of critical illness is exclusively bene¢cial
remains uncertain.

Protracted critical illness versus ageing

Pattern of abnormality
In PCI anterior pituitary functional status is impaired with decreased secretion of
GH, ACTH, LH, TSH and prolactin. The low normal serum TSH and prolactin
levels are generated mainly through tonically released TSH and prolactin. Pulsatile
TSH and prolactin release is signi¢cantly reduced quantitatively with absence of
the nocturnal hormone surges independent of concomitant sleep, although pulse
frequency is normal. Peripheral thyroid hormone levels are low and related to the
reduced nocturnal pulsatile TSH secretion. In ageing there are signi¢cant
similarities in that the balance of opinion suggests a blunted nocturnal TSH rise,
decreased TRH synthesis and a reduction in TSH levels; the most pronounced
change in this axis being the gradual age-dependent decline in serum T3
concentration as a consequence of decreased peripheral degradation of T4.
There are signi¢cant di¡erences in the LH^testosterone^oestradiol pro¢le seen
in men with PCI compared with the normal elderly male. In the former the
testosterone level is grossly reduced in the presence of only a modest reduction in
the oestradiol level, whereas the LH pulse frequency is increased although the
pulses are greatly reduced in amplitude. In contrast with advancing age normal
males show a slight decline in the serum testosterone level with an increased
proportion showing a testosterone level in the hypogonadal range. Basal LH
levels increase with age and there is a reduction in the amount and pulsatile
frequency of LH secretion by the pituitary in response to GnRH stimulation,
HYPOTHALAMIC AGEING 213

whereas oestradiol levels remain constant with increasing age. Nonetheless, despite
the quantitative di¡erence in degree of testosterone reduction, there are qualitative
similarities between PCI and ageing, primarily based on the suppressive action of
oestradiol on LH secretion, a feedback phenomenon which is facilitated by the
increase in aromatase activity observed in both situations.
In a similar fashion to TSH, prolactin and LH, the amount of GH secreted over
24 hours in PCI is reduced, with substantial reduction in the pulsatile component
but a non-pulsatile fraction that remains elevated. In normal subjects, after the age
of 40 years, GH production decreases gradually
at a rate of about 14% per
decade, primarily because of a decrease in the amplitude of nocturnal GH pulses.
At 65 years of age daily spontaneous secretion of GH is reduced by about 50^70%.
Thus there are considerable similarities between the pituitary functional changes
seen in PCI and normal ageing; major di¡erences include the intensity of the
testosterone de¢ciency, the elevated non-pulsatile fraction of GH secretion and
the reduction in ACTH secretion seen in PCI compared with normal ageing.
Despite the increased incidence of adrenal insu⁄ciency and the recognized
decrease in ACTH secretion, cortisol levels remain elevated in the majority of
patients with PCI, whereas in both humans and experimental animals there are no
major alterations of the hypothalamic^pituitary^adrenal axis, in terms of ACTH
and cortisol levels, during ageing. Interestingly adrenal androgen production is
signi¢cantly reduced both with ageing and PCI.

Site of defect
In both PCI and normal ageing the relatively normal and sometimes exuberant
pituitary hormone responses to a number of secretagogues indicate that any
degree of hypopituitarism is not a consequence of pituitary failure per se (Fig. 2).
Equally, the consistent relationships between pulsatile GH and TSH levels with
IGF1 and T3 levels respectively in PCI suggests that failure of peripheral
endocrine gland responsiveness is not contributing signi¢cantly to the
hypofunctional state. Similar conclusions have been drawn in the normal elderly
in whom the IGF1 response to an acute bolus of GH remains undiminished
compared with that seen in young adults.
These observations imply that the relative hypopituitarism in both PCI and
normal ageing is primarily hypothalamic in origin with the caveat that in certain
systems (i.e. pituitary^gonadal) there is some degree of end organ failure (i.e.
testis).
Thus impaired secretion of endogenous TRH, GnRH and hypothalamic factors
controlling GH secretion appear to be common to both PCI and normal ageing.
Reduced hypothalamic GHRH and probably decreased hypothalamic GH
secretagogue (i.e. ghrelin) are likely to be found in both ‘pathologies’. As for
214 VAN DEN BERGHE & SHALET

hypothalamic somatostatin, this appears reduced in PCI but the exact nature of any
change with ageing remains contentious.
In conclusion, PCI is a reasonable paradigm for studying the relative
hypopituitarism associated with normal ageing
more so for certain axes, such
as TSH and GH, but less so for others, such as gonadotropin secretion.

References
Baxter RC 1997 Acquired growth hormone insensitivity and insulin-like growth factor
bioavailability. Endocrinol Metab 4:65-69
Baxter RC, Hawker FH, To C, Stewart PM, Holman SR 1998 Thirty day monitoring of insulin-
like growth factors and their binding proteins in intensive care unit patients. Growth Horm
IGF Res 8:455^463
Dong Q, Hawker F, McWilliam D, Bangah M, Burger H, Handelsman DJ 1992 Circulating
immunoreactive inhibin and testosterone levels in men with critical illness. Clin Endocrinol
36:399^404
Gamrin L, Esse¤ n P, Forsberg AM, Hultman E, Wernerman J 1996 A descriptive study of
skeletal muscle metabolism in critically ill patients free amino acids, energy-rich phosphates,
protein, nucleic acids, fat, water, and electrolytes. Crit Care Med 24:575^583
Gardner DF, Kaplan MM, Stanley CA, Utiger RD 1979 E¡ect of triiodothyronine replacement
on the metabolic and pituitary responses to starvation. N Eng J Med 300:579^584
Gibson FA, Hinds CJ 1997 Growth hormone and insulin-like growth factors in critical illness.
Intensive Care Med 23:369^378
Hermansson M, Wickelgren RB, Hammerqvist F et al 1997 Measurement of human growth
hormone receptor messenger ribonucleic acid by a quantitative polymerise chain reaction-
based assay: demonstration of reduced expression after elective surgery. J Clin Endocrinol
Metab 82:421^428
Kakucska I, Romero LI, Clark BD et al 1994 Suppression of thyrotrophin-releasing hormone
gene expression by interleukin-l-beta in the rat: implications for nonthyroidal illness.
Neuroendocrinology 59:129^137
Meakins JI, Pietsch JB, Bubenick O et al 1977 Delayed hypersensitivity: indicator of acquired
failure of host defenses in sepsis and trauma. Ann Surg 188:241^250
Noel GL, Suh HK, Stone SJG, Frantz AE 1972 Human prolactin and growth hormone release
during surgery and other conditions of stress. J Clin Endocrinol Metab 35:840^851
Rodr|¤ gues-Arnao J, Yarwood Y, Ferguson C, Miell J, Hinds CJ, Ross RJ 1996 Reduction in
circulating IGF-I and hepatic IGF-I mRNA levels after caecal ligation and puncture are
associated with di¡erential regulation of hepatic IGF-binding protein-1, -2 and -3 mRNA
levels. J Endocrinol 151:287^292
Ross R, Miell J, Freeman E et al 1991 Critically ill patients have high basal growth hormone
levels with attenuated oscillatory activity associated with low levels of insulin-like growth
factor-1. J Clin Endocrinol 35:47^54
Rothwell PM, Lawler PG 1995 Prediction of outcome in intensive care patients using endocrine
parameters. Crit Care Med 23:78^83
Schlienger H, Sapin R, Capgras T et al 1991 Evaluation of thyroid function after myocardial
infarction. Ann Endocrinol (Paris) 52:283^288
St Germain DL, Galton VA 1985 Comparative study of pituitary^thyroid economy in fasting
and hypothyroid rats. J Clin Invest 75:679^688
Streat SJ, Beddoe AH, Hill GL 1987 Aggressive nutritional support does not prevent protein
loss despite fat gain in septic intensive care patients. J Trauma 27:262^266
HYPOTHALAMIC AGEING 215

Timmins AC, Cotterill AM, Cwyfan-Hughes SC et al 1996 Critical illness is associated with low
circulating concentrations of insulin-like growth factors-I and -II, alterations in insulin-like
growth factor binding protein and induction of an insulin-like growth factor binding protein
3 protease. Crit Care Med 24:1460^1466
Tweedle D, Walton C, Johnston DA 1972 The e¡ect of an anabolic steroid on postoperative
nitrogen balance. Br J Clin Pract 27:130^132.
Van den Berghe G, de Zegher F, Lauwers P 1994a Dopamine and the euthyroid sick syndrome in
critical illness. Clin Endocrinol 41:731^737
Van den Berghe G, de Zegher F, Lauwers P, Veldhuis JD 1994b Luteinizing hormone secretion
and hypoandrogenemia in critically ill men: e¡ect of dopamine. Clin Endocrinol 41:563^569
Van den Berghe G, de Zegher F, Schetz M, Verwaest C, Ferdinande P, Lauwers P 1995
Dehydroepiandrosterone sulphate in critical illness: e¡ect of dopamine. Clin Endocrinol
43:457^463
Van den Berghe G, de Zegher F, Bowers CY et al 1996 Pituitary responsiveness to growth
hormone releasing hormone, GH-releasing peptide-2 and thyrotropin releasing hormone in
critical illness. Clin Endocrinol 45:341^351
Van den Berghe G, de Zegher F, Velduis JD et al 1997a The somatotropic axis in critical illness:
e¡ect of continuous growth hormone-releasing hormone and GH-releasing peptide-2
infusion. J Clin Endocrinol Metab 82:590^599
Van den Berghe, de Zegher F, Veldhuis JD et al 1997b Thyrotrophin and prolactin release in
prolonged critical illness: dynamics of spontaneous secretion and e¡ects of growth hormone
secretagogues. Clin Endocrinol 47:599^612
Van den Berge G, de Zegher F, Baxter RC et al 1998 Neuroendocrinology of prolonged critical
illness: e¡ects of exogenous thyrotropin-releasing hormone and its combination with growth
hormone secretagogues. J Clin Endocrinol Metab 83:309^319
Van den Berghe G, Wouters P, Weekers F et al 1999 Reactivation of pituitary hormone release
and metabolic improvement by infusion of growth hormone-releasing peptide and
thyrotrophin-releasing hormone in patients with protracted critical illness. J Clin Endocrin
Metab 84:1311^1323
Van den Berghe G, Baxter RC, Weekers F et al 2000 A paradoxical gender dissociation within the
growth hormone/insulin-like growth factor I axis during protracted critical illness. J Clin
Endocrinol Metab 85:183^192
Van den Berghe G, Weekers F, Baxter RC 2001 Five days pulsatile GnRH administration unveils
combined hypothalamic-pituitary-gonadal defects underlying profound hypoandrogenism in
men with prolonged critical illness. J Clin Endocrinol Metab 86:3217^3226
Vermes I, Bieshuizen A, Hampsink RM, Haanen C 1995 Dissociation of plasma
adrenocorticotropin and cortisol levels in critically ill patients: possible role of endothelin
and atrial natriuretic hormone. J Clin Endocrinol Metab 80:1238^1242
Voerman HJ, Strack van Schijindel RJM, de Boer H, van der Veen EA, Thijs IG 1992 Growth
hormone: secretion and administration in catabolic adult patients, with emphasis on the
critically ill patient. Neth J Med 41:229^244
Vogel AV, Peake GT, Rada RT 1985 Pituitary^testicular axis dysfunction in burned men. J Clin
Endocrinol Metab 60:658^665
Wang C, Chan V, Tse TF, Yeung RT 1978a E¡ect of surgical stress on pituitary^testicular
function. Clin Endocrinol 9:255^2266
Wang C, Chan V, Tse TF, Yeung RT 1978b E¡ect of acute myocardial infarction on pituitary^
testicular function. Clin Endocrinol 9:249^253
Woolf PD, Hamill RW, McDonald JV, Lee LA, Kelly M 1985 Transient hypogonadotrophic
hypogonadism caused by critical illness. J Clin Endocrinol Metab 60:444^450
Zipser RD, Davenport MW, Martin KI et al 1981 Hyperreninemic hypoaldosteronism in the
critically ill: a new entity. J Clin Endocrinol Metab 53:867^873
216 DISCUSSION

DISCUSSION

Veldhuis: You have tackled an interesting problem that few groups have
studied, partly because of the complex confounding by concurrent drugs. Dr
Greet Van den Berghe has managed to ¢nd a fairly homogeneous population that
has convinced many of us that there is excellent pituitary responsiveness to a host of
releasing factors, particularly given in combination (Van den Berghe et al 1997a,b,
1998, 1999a,b, 2000, 2001a,b).
Bowers: One of the surprises of the GHRP2 is that as one gives it continuously, it
would be expected to down-regulate the e¡ect. But it is only partially down-
regulated because there is a persistent increase of pulsatile secretion. We have
done this study in older individuals up to 30 days, and in one individual for 90
days. The increased pulsatile secretion of GH remains sustained. When we give
acute bolus injections of GHRP2, the GH response is markedly desensitized.
However, during chronic infusion pulsatile GH secretion is augmented. In the
elderly studies there was a di¡erence between males and females. In some of the
men, GHRP2 did not work very well, so we gave it in combination with
GHRH. My interpretation of these data is that the women do not have as much
endogenous GHRH de¢ciency. GHRP2 is ine¡ective unless one has endogenous
GHRH. I assume that these women have more GHRH reserve. Men have more of a
de¢ciency of both the natural GHRP hormone, ghrelin, and GHRH.
Laron: I have had discussions with several of our colleagues about the fact that to
achieve the anabolic e¡ects of GH, pulsatile secretion is not necessary. What is the
bene¢t of using secretagogues rather than IGF1 or GH?
Bowers: That is not a simple question. When one gives GH itself, there is a
sustained level and there will be more side e¡ects. By stimulating pulsatile
secretion one can get the same e¡ects as with GH. This means that there is much
less GH secreted and indicates how e⁄cacious the e¡ects of pulsatile GH are on
peripheral targets. Comparatively, when one administers GH the levels are high.
Another issue is the di⁄culty of determining the dose of GH that should be given if
GH is to be given exogenously. By the GHRP approach, GH levels are totally
determined by the feedback system. On each patient the e¡ect is individualized
internally by the patient’s feedback response.
Laron: You are saying that by using this technique, one would be able to reduce
the very high sensitivity to GH in the aged population, while we still don’t know
the lowest (minimal) e¡ective dose.
Bowers: With regard to the total quantity of GH needed, one can obtain the same
e¡ect with a small quantity of pulsatile GH secretion as with administration of GH
at larger doses.
Veldhuis: The IGF feedback on the GHRP2 system is so e¡ective that if one
happens to overdrive GH output and IGF comes up, it will restrain the GH
HYPOTHALAMIC AGEING 217

(Giustina & Veldhuis 1998). I was under the impression that GH levels rise
promptly under continued infusions, and as the IGF1 levels rise the GH levels
fall about 50%. The system re-sets at a higher level of IGF.
Mˇller: I was impressed by the ¢ndings shown by Dr Shalet that the pulsatile
secretion of GH changes so dramatically from the acute phase to the chronic
phase, in spite of similar serum concentrations of peripheral IGF1.
Carroll: I want to raise a philosophical question. We are used to thinking of stress
responses as adaptive. There is a long tradition of this viewpoint for acute stress
responses. We never evolved to survive in intensive care units: there was no
opportunity for selection on this success! Are we to look at these hypothalamic
changes as adaptive stress responses, or as allostatic changes? Are we to look on
them, therefore, as pathological and in need of treatment?
Shalet: I think these are unnatural changes, because this is a very abnormal set of
circumstances.
Bowers: But there is an under-stress response in that second phase, it is not a full
stress response. Since these patients are still under stress during the second phase,
the full capacity of normal stress responses appears to be impaired and the stress
response is inappropriately low.
Carroll: It is some kind of adaptive response, but whether it will promote
survival is another matter.
Shalet: It is a failing system.
Carroll: You moved very quickly over the cortisol data. There are very good
data, for example, from burns units, showing sustained elevation of cortisol
going on for weeks, and a severe drop in production of cortisol binding
globulin. How much of the other axis changes are attributable to a primary
change in cortisol?
Shalet: That is a plausible suggestion. You could argue the case in terms of
changes in gonadotrophin or GH secretion. There are a number of possible
hypotheses to explain the endocrine ¢ndings in protracted critical illness, and this
would be one. We hardly touched at all on the use of other drugs in these patients.
In critical care, even if we exclude dopaminergic drugs, we haven’t mentioned
opiates.
Handelsman: We have done a study similar to this. We don’t see testosterone
levels of as low as 1 nM in men unless there is something more even than this
huge stress response. For example, hypercortisolism alone does not seem to
produce that. What does is opiates. Burns patients in particular, are on opiates for
long periods of time. I don’t know the data that you present very well, but I suspect
that there is an opiate e¡ect there that we should take into account.
Shalet: In fairness, in some of the studies Dr Van den Berghe has tried to look at
statistical di¡erences in terms of opiate use or not. This is presumably a fairly crude
overview of opiate use.
218 DISCUSSION

Handelsman: With regard to the aromatase, I don’t doubt the observations, but
the interpretation of whether a change in the oestradiol:testosterone ratio in
peripheral blood indicates a change in aromatase is suspect. Only 0.2% of the
body’s production of testosterone is converted to oestradiol on a whole body
basis. Just this very fact alone should necessitate caution. In addition, clearance
rate di¡erences in oestradiol will give all sorts of changes, especially in this
setting. I would be hesitant to take this interpretation further.
Shalet: Dr Van den Berghe raises the issue of the increased fat mass that occurs in
protracted critical illness. This may well contribute to the 37-fold increase in
aromatase activity. Changes in catecholamines and TNFa levels may also be
contributory factors.
Elahi: I have been discussing some work with respect to hypoglycaemia that
persists during this time. The therapy is glucose, potassium and insulin. There
are other modes of therapy currently under investigation, which are less risky.
Would the therapies you mention exaggerate the glucose excursion?
Shalet: I don’t think Dr Van den Berghe’s data touch on this. I would say that she
has moved on, and another key hormone she is focusing on at the moment is
insulin. This will be a key measure in her future studies.
Ruiz-Torres: I am interested in the relationship between the hormonal changes
in critical illness and those which appear during ageing. I have some doubts
regarding the value of this model for studies on ageing. For instance, each critical
illness has high dysproteinaemia as a common manifestation. The consequence
is a decrease of plasma protein transport and an increase of the distribution
volume. For this reason, the blood levels of hormones, substances, or drugs are
decreased. What happens afterwards is dependent not only on feedback
mechanisms but also on a wide range of other factors produced by toxins and
intermediary products.
Shalet: Let me answer you simply. I hope I portrayed in my presentation that I
didn’t see this as a marvellous model for ageing. I tried to be careful there. But there
were some similarities in the behaviour of certain axes. Even if what you said is true
(and I’m sure that much of it is), you still have to explain things like the gender
variation in GH secretion which is very striking.
Veldhuis: A 30% drop in blood volume puts an adult into shock. If you are
talking about any hormone changes of several-fold, you don’t have to worry that
distribution volume is the sole explanation: it won’t be for changes of that
magnitude.
Prior: In a study of women post-premenopausal ovariectomy, their initial
fasting HDL levels were 71.5 standard deviations compared to the normal age
range. These were not women with cardiovascular risk factors and their HDL
levels subsequently came into the normal range over the next three months.
When I looked at this I found that there are reports in the literature of low HDL
HYPOTHALAMIC AGEING 219

in association with acute trauma or illness. What do people think the mechanisms
might be?
Veldhuis: This is a protein that is under negative control by insulin, like others
such as IGFBP1. Beyond this, and the high insulin output that occurs during
hyperalimentation regimens, it is not clear to me what would lower HDL in
these women. The half-life is presumably several weeks.
Prior: These women were not on hyperalimentation: they were just fasting post-
operatively.
Laron: Stephen Shalet, considering the Scandinavian data you have presented
in both acute phase and in chronic illness, what is your advice concerning GH
therapy?
Shalet: The simple answer is that I don’t know. The dose of GH used in
the intensive care study was something like 20 units per day, whereas a standard
GH replacement dose is of the order of 1.2 units daily. This was a very big dose of
GH. Considering what we have heard, I am attracted to the approach that Dr Van
den Berghe and her colleagues are following, which is to begin to look at
combinations of pituitary hormone secretagogues. I can’t say more about the
insulin side. I know Dr Van den Berghe feels that insulin is going to be a key
factor in terms of outcome. Do I personally see a big place in this discussion for
GH? ‘Not particularly’ would be my immediate reaction. Other strategies look
more attractive to me. The one area where I think the data are most promising
for GH therapeutically would be the burns patients. The data look reasonable
but are limited.
Giustina: You mentioned that the combination of GHRP2 and TRH may be
e¡ective in modifying some target organ e¡ects. You explain this by implying
that GHRP has an e¡ect on the GH axis, and TRH has an e¡ect on the thyroid
axis. If I remember correctly, in critical illness, as occurs in other situations such
as acromegaly and type 1 diabetes, TRH may have some GH-stimulating action.
Might this be a combination acting mainly via the GH axis?
Shalet: Individual pituitary hormone secretagogues contribute signi¢cantly to
the e¡ect on other pituitary hormone axes when used in combination.
Bowers: Van den Berghe found that TRH by itself had very little e¡ect; it was
only when she put it together with GHRP2 that she got an additional e¡ect.
Carroll: In terms of intervention, in the near future we will have CRH
antagonists for human use. These have already been looked at in animal models
of delirium tremens, alcohol withdrawal and drug withdrawal. Drugs like CRH
antagonists will probably show great promise for managing the centrally driven
hyperstimulated hypothalamic^pituitary^adrenal states. Critical illness might
well be a clinical context where they will ¢nd application.
Bj˛rntorp: I have a vague memory of having heard about studies in which GH has
been used to treat heart failure. What is the current situation?
220 DISCUSSION

Shalet: The ¢rst paper, as I recollect, was by Fazio et al (1996). There were seven
patients with dilated cardiomyopathy, and the results were encouraging in terms of
improvement in cardiac function. To my knowledge, at least two subsequent
controlled studies from Sweden (Isgaard et al 1998) and Germany (Osterziel et al
1998) did not show bene¢t.
Giustina: The issue is very complex. The selection of patients is probably critical
in understanding the results of giving GH to patients with heart failure. There are
probably some patients with a certain degree of GH resistance. These are the
patients who are not likely to bene¢t from standard doses of GH.
Elahi: I presume you were talking about the paper by Fazio et al (1996). We were
very stimulated by this study. So far we have looked at four patients with
congestive heart failure. We gave them 21 d of GHRH therapy. Their IGF1
levels rose following therapy. We administered GHRH 1 mg qd and a month
later, 2 mg qd. We now have switched to *2 mg/day given in four equal pulses,
every 2 h from 23:00 to 05:00 h. The total dose of 30 mg/kg is split into four pulses.
Whereas with a 1 mg dose, nothing happened, with 2 mg we saw a substantial
increase in IGF1 levels. Having said this, it is only one-third of the response seen
in normal age-matched low-IGF volunteers. We are doing positron emission
tomography (PET) scans to see whether we get any changes in protein synthesis
in the heart.

References
Fazio S, Sabatini D, Capaldo B et al 1996 A preliminary study of growth hormone in the
treatment of dilated cardiomyopathy. N Engl J Med 334:809^814
Giustina A, Veldhuis JD 1998 Pathophysiology of the neuroregulation of growth hormone
secretion in experimental animals and the human. Endocr Rev 19:717^797
Isgaard J, Bergh CH, Caidahl K, Lomsky M, Hjalmarson A, Bengtsson BA 1998 A placebo-
controlled study of growth hormone in patients with congestive heart failure. Eur Heart J
19:1704^1711
Osterziel KJ, Strohm O, Schuler J et al 1998 Randomised, double-blind, placebo-controlled trial
of human recombinant growth hormone in patients with chronic heart failure due to dilated
cardiomyopathy. Lancet 351:1233^1237
Van den Berghe G, de Zegher F, Veldhuis JD et al 1997a The somatotropic axis in critical illness:
e¡ect of continuous growth hormone (GH)-releasing hormone and GH-releasing peptide-2
infusion. J Clin Endocrinol Metab 82:590^599
Van den Berghe, de Zegher F, Veldhuis JD et al 1997b Thyrotrophin and prolactin release in
prolonged critical illness: dynamics of spontaneous secretion and e¡ects of growth hormone
secretagogues. Clin Endocrinol 47:599^612
Van den Berghe G, de Zegher F, Baxter RC et al 1998 Neuroendocrinology of prolonged critical
illness: e¡ects of exogenous thyrotropin-releasing hormone and its combination with growth
hormone secretagogues. J Clin Endocrinol Metab 83:309^319
Van den Berghe G, Wouters P, Bowers CY, de Zegher F, Bouillon R, Veldhuis JD 1999a
Growth hormone-releasing peptide-2 infusion synchronizes growth hormone, thyrotropin
and prolactin release in prolonged critical illness. Eur J Endocrinol 140:17^22
HYPOTHALAMIC AGEING 221

Van den Berghe G, Wouters P, Weekers F et al 1999b Reactivation of pituitary hormone release
and metabolic improvement by infusion of growth hormone-releasing peptide and
thyrotrophin-releasing hormone in patients with protracted critical illness. J Clin
Endocrinol Metab 84:1311^1323
Van den Berghe G, Baxter RC, Weekers F, Wouters P, Bowers CY, Veldhuis JD 2000 A
paradoxical gender dissociation within the growth hormone/insulin-like growth factor I
axis during protracted critical illness. J Clin Endocrinol Metab 85:183^192
Van den Berghe G, Wouters P, Weekers F et al 2001a Combined administration of growth
hormone-, thryotropin- and gonadotropin secretagogues to prolonged critically ill men:
endocrine and metabolic e¡ects. J Clin Endocrinol Metab
Van den Berghe G, Weekers F, Baxter RC 2001b Five-day pulsatile gonadotrophin-releasing
hormone administration unveils combined hypothalamic-pituitary-gonadal defects
underlying profound hypoandrogenism in men with prolonged critical illness. J Clin
Endocrinol Metab 86:3217^3226
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Glucose tolerance, glucose utilization

and insulin secretion in ageing
Dariush Elahi, Denis C. Muller*, Josephine M. Egan*, Reubin Andres*,
Johannes Veldhuis{ and Graydon S. Meneilly{

Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston,

MA 02114, *Laboratory of Clinical Physiology, Gerontology Research Center, National
Institute on Aging, National Institutes of Health, Baltimore 21224, MD, {University of
Virginia, Division of Endocrinology and Metabolism, Charlottesville, VA 22908, USA, and
{Department of Medicine, University of British Columbia, Vancouver, BC V6T 1Z3, Canada

Abstract. Ageing is associated with an increased incidence of hypertension, macrovascular

disease and type 2 diabetes (non-insulin-dependent diabetes). It has been suggested that a
common mechanism may be responsible for all of these pathological states since all of
these conditions often cluster in the same individual. Epidemiological and clinical data
have consistently demonstrated an association between insulin resistance and/or
hyperinsulinaemia and glucose intolerance, dyslipidaemia and elevated systolic blood
pressures. Therefore, insulin resistance and hyperinsulinaemia have been proposed as
the causal link among the elements of the clusters. The elderly are more glucose
intolerant and insulin resistant, but it remains controversial whether this decrease in
function is due to an inevitable consequence of ‘biological ageing’ or due to
environmental or lifestyle variables, noticeably increased adiposity/altered fat
distribution and physical inactivity. An increase of these modi¢able factors has been
shown to result in increases in insulin resistance and hyperinsulinaemia and vice versa.
However, insulin secretion appears to decrease with age even after adjustments for
di¡erences in adiposity, fat distribution and physical activity. The glucose intolerance of
ageing may be due, in part, to decreased insulin sensitivity of pancreatic b cells to
insulinotropic gut hormones (GLP1/GIP) and in part to alterations of hepatic glucose
production.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 222^246

Ageing is associated with a decline in function in many, if not all, human

physiological systems. Cardiovascular, haemodynamic, metabolic and renal
functions generally decrease with advancing age. Insulin has been causally related
in the aetiology of many of these decrements (Reaven 1988, Kaplan 1989, Stout
1990, DeFronzo & Ferrannini 1991). The reduction in whole-body carbohydrate
metabolism in the elderly is one of the hallmarks of the ageing process and
substantial evidence shows that increasing age is associated with decreased
222
GLUCOSE HOMEOSTASIS IN AGEING 223

FIG. 1. E¡ect of age on fasting and 2 h plasma glucose levels in men from the Baltimore
Longitudinal Study of Aging.

glucose tolerance (Andres 1971, Broughton & Taylor 1991, Davidson 1979,
DeFronzo 1981, Reaven et al 1989). Oral glucose tolerance tests (OGTTs),
conducted in healthy non-diabetic individuals across the age range clearly
demonstrate a decline in glucose tolerance, as judged from the fasting,
intermediate and the 2 h levels, in each age decade from the third decade (20^29
year old) to the ninth decade (80^89 year old) (Elahi & Muller 2000). This
decline is easily appreciated from the relationship of age and either fasting or 2 h
level, as shown in Fig. 1, and is observed both in men and women. The two-hour
plasma glucose level during an OGTT rises on average, 5.3 mg/dl per decade and
the fasting plasma glucose rises on average, 1 mg/dl per decade (Davidson 1979).
The decline in glucose tolerance is also re£ected in NHANES III survey on the
prevalence of diabetes and impaired fasting glucose and impaired glucose
tolerance in US adults (Harris et al 1998). Comparison of the percentage of
physician-diagnosed diabetes in middle-aged adults (40^49 years) and elderly
adults (575 years) reveals an increase from 3.9% to 13.2%. The percentage of
adults with (a) undiagnosed diabetes (fasting plasma glucose [FPG] 5126 mg/dl)
increased from 2.5 to 5.7%; and (b) impaired FPG (110^125 mg/dl) increased from
7.1% to 14.1%. Thus, approximately a third of the elderly adults in the USA have
abnormal glucose metabolism as de¢ned by the revised 1997 criteria of the
American Diabetes Association (The Expert Committee on the Diagnosis and
Classi¢cation of Diabetes Mellitus 1997). A detailed review of the e¡ects of
ageing on glucose homeostasis, in both humans and animals, has recently been
224 ELAHI ET AL

FIG. 2. Risk factors of the metabolic syndrome. Each line represents a statistically signi¢cant
association between each risk factor that is found in most populations.

published in a number of relevant articles (Evans & Farrell 2001). The reduction of
glucose tolerance associated with ageing is also accompanied with dyslipidaemia
and hypertension. The clustering of these conditions is commonly referred to as
syndrome X or the metabolic syndrome. These pathological conditions are more
prevalent in the elderly and are complex, interrelated and multifunctional. The
complexity of these inter-relationships is depicted in Fig. 2. The lines connecting
the di¡erent factors represent the statistically signi¢cant associations that exist in
the cross-sectional analysis of most populations. For example, a signi¢cant decrease
in glucose tolerance with increasing age can be demonstrated from all
epidemiological studies. However, this relationship is confounded by an increase
in adiposity and a decrease in physical activity with age and each of these factors are
also associated with a decrease in glucose tolerance. It is di⁄cult to demonstrate
how much of the glucose intolerance can be attributed solely to ageing, decreased
activity or increased adiposity; the combination of these certainly leads to glucose
intolerance. Similarly, the relationships of age, hyperinsulinaemia, dyslipidaemia
and insulin resistance are interwoven. This review focuses on the clinical evidence
of a change with age in insulin resistance and insulin secretion that is independent
of changes in other known factors.
GLUCOSE HOMEOSTASIS IN AGEING 225

Shimokata et al (1991) examined the relationship between age, obesity, physical

activity and glucose tolerance in a community dwelling with men and women
ranging in age from 17 to 92, from the Baltimore Longitudinal Study of Aging
(BLSA). The independent e¡ect of age on glucose tolerance was examined after
statistical adjustment for the confounding e¡ects of obesity, fat distribution and
physical activity. They found that the decline in glucose tolerance from early
adult (17^39 year) to middle age (40^59 year) is entirely explained by the
secondary in£uences of fatness and ¢tness. However, the decline from mid-life to
old age (60^92 year) was still in£uenced by chronological age. Other population
studies also attribute the decline in glucose tolerance to age-related environmental
factors (obesity, physical activity, dietary habits, diabetogenic drug use) (Zavaroni
et al 1986, Maneatis et al 1982, Zamboni et al 1997). However, it should be noted
that anthropometric determination of body composition and questionnaire-
derived assessment of impaired activity have their limitations. Furthermore,
when statistical adjustment is made, in addition to the variability in the precision
of the measured confounders, linear relationships between the confounders are
assumed, interactions between them are ignored, and sample size may not be
su⁄ciently large. Thus, in a complex statistical model, true cause^e¡ect
relationships may not be detectable due to under/over adjustment or
imprecision.

Diagnostic criteria and the elderly

It is desirable to have a screening test for detection of glucose intolerance which is
both sensitive (has a high probability of being positive when there is glucose
intolerance) and speci¢c (has a high probability of being negative when there is
no glucose intolerance). Often, there is compromise between the two. Increasing
sensitivity reduces speci¢city and vice versa. The screening tests should be reliable
and reproducible. The expert committee on the diagnosis and classi¢cation of
diabetes mellitus of the American Diabetes Association (ADA) provided new
criteria that simpli¢ed the method of detection of glucose intolerance and
diabetes (The Expert Committee on the Diagnosis and Classi¢cation of Diabetes
Mellitus 1997). For clinical diagnosis in asymptomatic individuals the ADA
recommends that diabetes be de¢ned as an FPG value 57.0 mmol/l (126 mg/dl).
This recommendation revises the 1985 World Health Organization (WHO)
diagnostic criteria (WHO 1985) and the National Diabetes Data Group
(NDDG)’s 1979 criteria for diabetes (National Diabetes Data Group 1979),
which relied on both fasting glucose and the 2 h oral glucose tolerance test. An
FPG 57.8 mmol/l (140 mg/dl) and/or a 2 h plasma glucose value 511.1 mmol/l
(200 mg/dl) was diagnostic of diabetes by the previous (1985/1979) criteria.
The 1997 ADA criteria also provided recommendations for an additional
226 ELAHI ET AL

diagnostic class. Impaired fasting glucose (IFG) is de¢ned as an FPG level

56.1 mmol/l (110 mg/dl) and 57.0 mmol/l (126 mg/dl); normal fasting glucose
is changed to an FPG 56.1 mmol/l. The OGTT is not recommended for routine
diagnosis of glucose intolerance or diabetes. The 1999 WHO report on ‘De¢nition,
diagnosis and classi¢cation of diabetes mellitus and its complications’ essentially
endorses the ADA’s 1997 recommendation with the exception that they strongly
advocate the use of the OGTT (WHO 1999). The 1985 WHO criteria had de¢ned
impaired glucose tolerance (IGT) as a fasting glucose 57.8 mmol/l and a two-hour
FPG 57.8 mmol/l and 511.1 mmol/l. Numerous reports (at least 25) have
examined the new ADA 1997 criteria in order to assess their impact on
prevalence of diabetes by age, sex and ethnicity. An example, is the report by
Harris et al (1998), who used the FPG and the 2 h glucose value from an OGTT
from the data obtained from the NHANES III survey to compare the percentage of
the population meeting the 1997 ADA and the 1985 WHO diagnostic criteria by
age, sex and ethnic groups.
It is not entirely clear how the various cut-o¡ points were agreed upon by the
various expert committees. It is absolutely evident from all the population data
available, almost no one with a fasting value of 140 mg/dl (7.8 mmol/l), has a 2 h
plasma value 4200 mg/dl (11.1 mmol/l). However, many individuals with a
fasting value of 5140 mg/dl will have a 2 H value 5200 mg/dl. It is also not
clear how exactly the new cut-o¡ fasting value of 126 mg/dl was agreed upon.
This new value is reported to be more in agreement for the diagnosis of
diabetes with the 2 h plasma glucose value of 200 mg/dl following the OGTT,
than the previous fasting value of 140 mg/dl. However, comparisons of the
results of FPG and the 2 h plasma glucose level following an OGTT from the
NHANES III survey using the 1985 WHO and the 1997 ADA criteria are still
often more than 50% discrepant (Harris et al 1998). The ADA’s 1997 report has
stated that the justi¢cation for the cut-o¡ point for the 2 h plasma glucose of the
OGTT of 200 mg/dl is derived from the evidence that the prevalence of
microvascular complications increases dramatically at this point. Another
justi¢cation has to do with the fact that the 2 h plasma glucose value following
an OGTT from several populations has a bimodal distribution. The nadir
intersection of the two modes is known as the antimode and it shifts to the
right with advancing age (Bennett et al 1976). The 200 mg/dl level,
purportedly, represents the average level of the antimodes from several large
population studies (Pima Indians, Naruans, Samoans, Mexican-Americans, and
East Indians). However, the antimodes from these populations range widely
from about 150 to over 300 mg/dl (Bennett et al 1976, Zimmet et al 1978, Loo
et al 1993, Dowse et al 1994, Rosenthal et al 1985) and do not support the
average 200 mg/dl level. They do, however, increase with age. Nevertheless,
the 200 mg/dl level remains the standard.
GLUCOSE HOMEOSTASIS IN AGEING 227

FIG. 3. Distributions for diagnosis of diabetes (D) made from FPG (57.0 mmol/l; x-axis) and
from two-hour plasma glucose level following OGTT (511.1 mmol/l; y-axis) in the BLSA
population.

We summarized 25 reports of studies that have computed the prevalence of

diabetes in a population when both fasting glucose and the glucose tolerance test
(GTT) were used (the WHO recommendation) in contrast to the prevalence when
only the fasting level is considered (the ADA recommendation). A very high
percentage (ranging from 11 to 80%) of the diagnosable diabetics are missed if
the GTT is not considered.
When both tests are used, there are nine possible diagnostic categories since each
test results in a classi¢cation of normal, impaired or diabetic. Figure 3 illustrates the
distribution of test results from our BLSA population in men and women over 65
years. The percentage of subjects classi¢ed as diabetic when both tests are used is
remarkably reduced when the fasting glucose alone is used: 65% of the men and
80% of the women would have been missed.
In other analyses, since the BLSA subjects are followed over many years with
repeated testing, we could estimate that, on average, the diagnosis of diabetes can
be established some 7 to 9 years earlier by glucose tolerance testing than by fasting
glucose testing alone.
The result of our study is in agreement with many of the reported studies which
show that the 1997 ADA diagnosis standards do not result in equal sensitivity for
fasting and 2 h glucose levels, especially in older individuals. A disturbingly high
percentage of older men and women will have to be informed either that they have
diabetic or impaired test results. Although the use of fasting plasma glucose alone
for diabetes diagnosis simpli¢es testing, the WHO criteria identify a much greater
percentage of elderly subjects with diabetes or impaired glucose metabolism.
228 ELAHI ET AL

We will now provide a short summary on the e¡ect of age on hepatic glucose
production, glucose uptake and insulin secretion. This topic has been recently
reviewed in detail, as previously noted, by Evans & Farrell (2001) and we will
focus mainly on our own work using the clamp technique. The methodology of
the technique has been previously reported (DeFronzo et al 1979) and its various
uses reviewed (Elahi 1996). We will additionally review our data with respect to the
insulinotropic e¡ect of incretins and our preliminary data using these agents to
stimulate insulin secretion and to regulate glucose homeostasis in type 2 diabetic
individuals.

E¡ect of age on hepatic glucose production

During the post-prandial state stable plasma glucose levels are maintained by a
coordinated balance between hepatic glucose production (HGP) and glucose
uptake by peripheral tissues (primarily muscle). The role of the liver in the
maintenance of plasma glucose level has been best evaluated with the
euglycaemic clamp. Dose^response relationships between plasma insulin level
and HGP clearly demonstrate that in normal tolerant individuals, the liver is
exquisitely sensitive to insulin and HGP is completely suppressed at insulin
levels well below the commonly used insulin dose (240 pmol.m72 .min71).
Additionally, as demonstrated in Fig. 4, there is no di¡erence in either the basal
HGP or in the dose^response curve of its suppression by insulin between young

FIG. 4. Dose^response curves for suppression of hepatic glucose output by insulin in young
and old participants.
GLUCOSE HOMEOSTASIS IN AGEING 229

and old individuals. HGP is suppressed over 90% in aged individuals with insulin
levels of 240 pmol/l (40 mU/ml) (Meneilly et al 1987). The half-maximal e¡ect is
approximately 150 pmol/l (25 mU/ml). This ¢nding is in general agreement with
¢ndings of other investigators in both young and elderly volunteers (DeFronzo
1979, Fink et al 1983, Rizza et al 1981). In our studies (Meneilly et al 1987), we
have also observed that with low dose insulin infusions, HGP is more rapidly
suppressed in the elderly. We have attributed this to a delayed suppression of
endogenous insulin release in the elderly individuals (as assessed with C-peptide
levels). Thus, early in the study, the liver is exposed to a greater insulin level in
the old than in the young. The liver is exquisitely sensitive to portal insulin
concentration and an increase of 30^60 pmol/l (*5 mU/ml) reduces HGP by 50%
(DeFronzo et al 1983). Thus, although total suppression of HGP does not di¡er
with age, reduction of HGP by insulin occurs more rapidly in the elderly at
physiological levels. The European Group for the Study of Insulin Resistance
(EGIR, described in more detail below) has reported their analysis of HGP in
344 non-diabetic subjects (212 men and 132 women) during a euglycaemic clamp
(240 pmol.m72 .min71) (Natali et al 2000). They ranged in age from 18^85 years
and in BMI from 15^55 kg/m2. Basal HGP showed a large variability and was
strongly related to lean body mass (LBM) (r = 0.63). HGP was 23% higher in
men than in women, which is entirely due to higher LBM in men and no longer
di¡erent when HPG is adjusted for LBM. Similarly HGP was also positively
related to BMI and percentage fat and again these associations were no longer
di¡erent when adjusted for LBM. There was a tendency for HGP to decline with
increasing age (71.1 0.7 mmol.m71 per year) which was not signi¢cant
(P = 0.10) and any signi¢cance was completely lost when adjusted for LBM
(0.002 mmolm71 .kg LBM71 per year). Therefore the variability of basal HGP is
due to di¡erences of LBM which also explains the di¡erences of HGP due to age,
sex and BMI. Additionally, independent of LBM and fasting plasma insulin,
peripheral insulin resistance is associated with higher HGP. The latter suggests
that regulation of peripheral glucose uptake and HGP are coupled.

E¡ect of age on peripheral tissue glucose utilization

Muscle is the primary sink for glucose uptake as adipose tissue is relatively
inert and accounts for only 2^3% of glucose uptake (DeFronzo 1981). Brain and
the splanchnic bed account for another 20^55% (DeFronzo 1988). The latter
two sinks for glucose uptake are insulin independent. The majority of reports
examining the role of age on glucose uptake, utilizing a variety of techniques
(clamp, MinMod, local forearm perfusion, glucose^insulin^somatostatin
infusion) have demonstrated decreased insulin sensitivity during
hyperinsulinaemia as summarized in the report by EGIR (Ferrannini et al 1996).
230 ELAHI ET AL

However, this decreased insulin sensitivity is not demonstrated for fasting insulin
levels (Coon et al 1989). The EGIR report is an analysis of euglycaemic clamps, at a
single dose of 240 pmol.m72 .min71, from 20 European clinical research centres.
In this study a total of 1146 clamps were performed (776 men and 380 women) for
2 h in individuals with normal glucose tolerance and arterial blood pressure. In
addition, anthropometric data were available. Glucose utilization, M, was
calculated during the last 40 min of the study. In univariate analysis, age was
associated with a signi¢cant decrease in insulin action of (P = 0.0002)
0.9 mmol.kg71 per decade of life, (*0.2 mg.kg71 per decade of life) which was
no longer signi¢cant when adjusted for BMI (P = 0.08). However, BMI was
strongly associated with a decrease in insulin action (5 mmol.kg71 .10 kg of body
weight, 0.9 mg.kg71 .10 kg of body weight, P50.001), which remained
signi¢cant, even after adjustment for age. Furthermore, when the relationship of
glucose utilization, M, was examined as a function of gender and obesity (BMI
425 or 425), age-related decrease in insulin action was demonstrated only in
lean women (cf. Fig. 3 of Ferrannini et al 1996), without a slope change after age
50 years (i.e. no menopause e¡ect). Thus, this large study demonstrates that glucose
uptake is not altered as a function of age per se except in lean women at this
hyperinsulinaemic level.
Despite the strength of the EGIR study, the issue of age-related insulin
resistance is still controversial, because it has been argued that complete dose-
response curves are necessary to resolve the issue. Several groups have conducted
dose-response studies as a function of age. The studies are rather small (n*50),
mainly in men, without a signi¢cant di¡erence in BMI between young and old.
One study (Rowe et al 1983) has examined the e¡ect of age on glucose utilization
over the insulin range of 60^6000 pmol/l (10^1000 mU/ml). An age-associated
decrease in glucose utilization was demonstrated with preservation of maximal
glucose uptake (i.e. a shift to the right). The half-maximal glucose uptake
occurred when plasma insulin levels were 324 pmol/l (54 mU/ml) in the young
and 678 pmol/l (113 mU/ml) in the old. When the glucose utilization was plotted
per kilogram of lean body mass, the relationship remained. This study is in
agreement with other studies where several insulin doses were employed
(DeFronzo 1979, Fink et al 1983). In addition, studies using various techniques,
including the hyperglycaemic clamp (DeFronzo 1979, Elahi et al 1993), the
forearm glucose uptake technique (Jackson et al 1982) and Minimal Model (Min
Mod) (Chen et al 1985) have revealed resistance to insulin-induced glucose disposal
in aged volunteers. It is obvious, that in addition to age, various factors in£uence
insulin sensitivity including fat mass, fat distribution, physical ¢tness, dietary
composition and genetic factors. When these factors are taken into account, it
is not clear that age per se still has an independent e¡ect on peripheral glucose
uptake.
GLUCOSE HOMEOSTASIS IN AGEING 231

The e¡ect of age on pancreatic b cell sensitivity

A major component of glucose homeostasis is the balance between insulin

secretion and tissue sensitivity to insulin. Again there are many reports that
either support or refute a decline in pancreatic b cell response to glucose as a
function of age. However, very few investigators have examined b cell secretion
in a dose^response manner. We have performed 230 hyperglycaemic clamps for 2 h
at four doses (3.0, 5.4, 7.9 and 12.8 mmol/l above basal glucose levels) in healthy
young, middle-aged and old volunteers (Elahi et al 1993). Deconvolution analysis,
which provides an analysis of insulin secretion from the plasma insulin
concentration curves, showed that insulin secretion is characterized by (1) a spike
lasting about one minute (¢rst phase), (2) cessation of secretion from 0^15 min, (3)
a step increase at 15 min to a level above basal, and (4) a gradual increase in secretion
from 15^120 min. In these dose^response studies, as the hyperglycaemic level
increased, there was a gradual increase in insulin response within each age group.
Furthermore, with each group, comparisons of insulin responses between doses
were statistically signi¢cant both for the early phase insulin response and for each
of the succeeding 30 min periods (30^60, 60^90 and 90^120 min) of the late-phase
insulin response. However, in neither the early phase, nor the later-phase insulin
responses were there any statistically signi¢cant di¡erences among the age groups
(Fig. 5). These results are consistent with the observation of DeFronzo (1979)
where hyperglycaemic clamps were performed at 6.9 mmol/l above basal in
young, middle-aged and old. Using the Min Mod technique, two studies (Chen
et al 1985, Pacini et al 1988) have also shown that neither ¢rst phase nor second
phase insulin response di¡ers signi¢cantly between young and old. Thus, the
redundant, unnecessary consensus is that insulin secretion does not di¡er in ageing.
As previously discussed, comparisons between groups (e.g. young vs. old) have
inherent di⁄culties due to the di¡erences between them other than age. In our own
studies, we have tried to ‘match’ for di¡erences using several approaches. In the
clamp studies reported above (Elahi et al 1993), we allowed a higher BMI entry
criterion in the young than the old in order to match for loss of LBM and
increased adiposity in the older group. In a study of insulin secretion of women
master athletes across the age span, we carefully restricted entry criteria for both
VO2 max and percentage in order to examine the e¡ect of ageing in very active
individuals who maintained their LBM. We found, again, that insulin secretion
does not di¡er in ageing (Ryan et al 2001). However, despite our strict entry
criteria for similarity for percentage fat, older female athletes had higher amount
of visceral fat than younger female athletes (*30 cm2) (Ryan & Elahi 1996). An
elegant cross sectional study (Pimenta et al 1995) enrolled 100 volunteers of
European ancestry with normal OGTT according to the 1985 WHO criteria.
There were two groups (n = 50 in each), one of which had at least one ¢rst-degree
232 ELAHI ET AL

FIG. 5. Bimodal time course of the plasma insulin response to ¢xed hyperglycaemia. There is a
graded increase in insulin as hyperglycaemia increases, and a delay in the fall of plasma insulin in
response to the instantaneous fall in glucose at the end of the clamp.
GLUCOSE HOMEOSTASIS IN AGEING 233

type 2 relative. The two groups were matched for gender, age, weight, height,
waist hip ratio, HbA1C , and fasting glucose and insulin levels. None were being
treated for hypertension and none were engaged in habitual exercise. A
hyperglycaemic clamp at a plasma glucose level 180 mg/dl (10 mmol/l) was
performed for three hours. In 62 individuals (26 from the type 2 group and 36
from the normal group) an additional hyperinsulinaemic^euglycaemic clamp
(240 m.Um^2 .min^1) was also performed. Again, the two groups were carefully
matched for all the confounding variables listed above. The group with a ¢rst-
degree type 2 relative had reduced ¢rst and second phase insulin response (cf.
Fig. 2 of Pimenta et al 1995). However, their peripheral tissue sensitivity to
insulin (from both clamps) was not signi¢cantly di¡erent. This study
demonstrates that middle-aged (*40 2 years) individuals of European ancestry
with normal OGTT but with a ¢rst-degree type 2 diabetic relative have impaired b
cell function without a signi¢cant reduction in peripheral tissue sensitivity to
insulin as compared with normal individuals without a type 2 diabetic relative.
Furthermore, b cell defects were not uniform. These results support the
hypothesis that a defect in insulin secretion precedes a defect in insulin sensitivity
in the development of type 2 diabetes.
During a hyperglycaemic clamp in normal and type 2 diabetic individuals,
second phase insulin response is seen to progressively increase throughout the
180 or 240 min duration of the clamp. Does the second phase response ever
plateau? We have performed a 10 h hyperglycaemic clamp at *190 mg/dl
(10.6 mmol/l) in young (age = 23 1 years, BMI = 23 0.6, WHR = 0.8, VO2
max = 44 ml.kg71 .min71, 5 males and 5 females) and old (age = 80 2,
BMI = 24 0.5, WHR = 0.89, VO2 = max 21 ml.kg71 .min71, 5 males and 5
females) with normal OGTTs (Meneilly et al 1999). First phase insulin
responses were not di¡erent between the two groups (Fig. 6). Second phase
insulin responses increased progressively until *120 to 150 min in the old
after which it reached a plateau. In the young a plateau was not reached until
*300 min. The second phase insulin responses were signi¢cantly di¡erent after
120 min. Glucose infusion rates necessary to maintain the stable hyperglycaemia
reached a plateau at *180^240 min in both groups, but at signi¢cantly di¡erent
rates. We also obtained one-minute samples for determination of plasma insulin
levels for 150 min during both the basal period and after 1 h of achievement of a
plateau in the rate of glucose infusion. Insulin release was evaluated by cluster
analysis (Porksen et al 1995, Engdahl et al 1977). Disorderly insulin release, a
reduction in the amplitude and mass of rapid insulin pulses and decreased
frequency of ultradian pulses is characteristic of normal ageing in the basal
state (Meneilly et al 1997). In the stimulated state, a reduction in mass and
amplitude of rapid pulses, decrease in frequency and regularity of ultradian
pulses occurs with ageing (Meneilly et al 1999).
234 ELAHI ET AL

FIG. 6. Plasma insulin response during a 10 h hyperglycaemic clamp at *11 mmol/l. The
insert on the bottom shows the ¢rst phase insulin response with an adjusted scale.

The incretin e¡ect in ageing and in type 2 diabetes

Incretin is the umbrella term to cover the multiple gut factors (now known to be
hormones) which augment insulin response above that which can be attributed to
glucose alone. This insulinotropic e¡ect has been demonstrated by matching the
time course of the plasma glucose excursion following an OGTT with both an i.v.
infusion of glucose (Perley & Kipnis 1967) and during a hyperglycaemic clamp. In
the hyperglycaemic clamp, plasma glucose was increased to *11 mmol/l for 2 h on
GLUCOSE HOMEOSTASIS IN AGEING 235

two di¡erent occasions. In one, only glucose was infused, while in the second, an
OGTT was administered at 60 min and the exogenous glucose infusion rate was
adjusted during the second hour as the ingested glucose was being absorbed; in this
manner the plasma glucose level remained at the same plateau level in the second
hour as in the ¢rst hour (Andersen et al 1978). During the second hour in the study,
despite the constancy of the plasma glucose concentration, a marked potentiation
of insulin secretion was observed. The increase in insulin level is preceded by an
increase in plasma concentration of a measured gut hormone, glucose-dependent
insulinotropic polypeptide (GIP) by 5 min, with a subsequent time course very
similar to the potentiation of insulin response. It was logical to attribute the
increase in insulin to this hormone. To test this further, we subsequently infused
GIP during the second hour of a hyperglycaemic clamp and we showed that GIP
can indeed potentiate insulin response in normal individuals in a similar fashion to
that observed following ingestion of glucose (Elahi et al 1979). We then examined
b cell sensitivity to endogenously released GIP as a function of age (Elahi et al 1984)
and b cell sensitivity to exogenously administered GIP as a function of age and
hyperglycaemia (Meneilly et al 1998) during hyperglycaemic clamps. b cell
sensitivity to endogenously released GIP was analysed from the association of
the ratio of insulin response after OGTT to that which would occur had OGTT
not been administered, divided by the ratio of GIP response after administration of
OGTT to GIP levels before administration of OGTT:

[IRI (90 ^ 120 min+GIP) / IRI (90 ^ 120 min ^ GIP)] / [GIP (90^120 min) / (GIP
(0^60 min)]

There was a signi¢cant negative age relationship, indicating that b cell sensitivity
to GIP is reduced with advancing age (Elahi et al 1984). In studies with exogenous
infusion of GIP, two hyperglycaemic clamps were performed at a level of
*11 mmol/l and at *18 mmol/l in young (19^26 years) and old (67^79 years)
volunteers (Meneilly et al 1998). A total of 93 clamps were performed. During
each clamp, GIP was infused for 60 min at a dose of 2 pmol.kg71 .min71 and
4 pmol.kg71 .min71. A clamp was also performed, at each glycaemic level,
without GIP administration. The GIP levels during the basal state, before
the infusion of GIP at hyperglycaemia and after infusion, were similar
between groups and between hyperglycaemic plateaus during the 2 and
4 pmol.kg71 .min71 infusions (60^120 min levels = *350 and 580 pmol/l,
respectively). In response to GIP infusions, signi¢cant increases in insulin
occurred in young and old at both glucose levels. The potentiation of the insulin
response caused by GIP was greater in the young subjects than the old in the
11 mmol/l glucose hyperglycaemic study. However, the insulin response to GIP
was similar in both young and old during the 18 mmol/l glucose hyperglycaemic
236 ELAHI ET AL

clamps. The insulinotropic e¡ect of this incretin was greater in both the young and
the old in the 18 mmol/l clamps than in the 11 mmol/l clamps. We concluded that
normal ageing is characterized by a decrease in b cell sensitivity to GIP during
modest hyperglycaemia. The age-related impairment in response to GIP may be
an important cause of the glucose intolerance of ageing, a precursor for diabetes
in this age group. The insulinotropic e¡ect of GIP is increased with increasing
levels of glycaemia, and the defect in b cell response to GIP disappears when
plasma glucose is increased to higher levels.
While we were examining the e¡ect of GIP on insulin secretion, Habener and
colleagues (Mojsov et al 1986) reported the identi¢cation of another gut hormone,
glucagon-like peptide 1 (GLP1), and subsequently showed it to have a potent
insulinotropic e¡ect in rats (Weir et al 1989). Infusions of GLP1 were also
subsequently shown to lower fasting plasma glucose levels in type 2 diabetic
patients (Nathan et al 1992). We examined the insulinotropic e¡ect of GLP1 in
normal glucose-tolerant and in type 2 diabetic volunteers during hyperglycaemic
clamp (*11 mol/l) and compared its e¡ect to that of GIP (Elahi et al 1994). We
demonstrated that GLP1 is indeed a more potent insulinotropic hormone than
GIP. Furthermore, there was an additive e¡ect of GIP and GLP1 on b cell
stimulation. Most importantly, we showed that while GLP1 has a potent
insulinotropic e¡ect in type 2 individuals, albeit less than in normal glucose
tolerant individuals, the GIP insulinotropic e¡ect is totally absent. Thus GLP1 is
being investigated as a potential therapeutic agent for the normalization of glucose
homeostasis in type 2 diabetes.
Recently, it has been shown during a hyperglycaemic clamp (*11 mmol/l)
GLP1 signi¢cantly potentiates insulin release in elderly type 2 diabetic volunteers
(age 570 years). The potentiation is clinically relevant and at least threefold greater
than insulin release with glucose alone (Fig. 7) (Meneilly et al 2001a). The same
volunteers were also examined with a hyperinsulinaemic^euglycaemic clamp
during infusion of somatostatin in the presence and absence of GLP1 (Meneilly
et al 2001b). The plasma glucose was allowed to fall to a normal level in both
euglycaemic clamps (*5.3 mmol/l). During these two clamp studies both plasma
insulin and glucose levels were similar. We showed that peripheral tissue sensitivity
to insulin was signi¢cantly greater when GLP1 was infused. This study
demonstrates that in states of glucose intolerance, such as type 2 diabetes, GLP1
has insulinomimetic, or at least insulin-augmenting, properties in peripheral tissues
which can not be attributed to the well known delayed gastric emptying properties
of GLP1. It was previously reported that while there was a small tendency for
GLP1 to augment peripheral tissue sensitivity to insulin in young normal
glucose tolerant men, the increase was not statistically signi¢cant (Ryan et al 1998).
We are currently examining the role of continuous subcutaneously administered
GLP1 for 12 weeks in type 2 diabetic volunteers. These volunteers were previously
GLUCOSE HOMEOSTASIS IN AGEING 237

FIG. 7. Plasma glucose, insulin and C-peptide levels during the hyperglycaemic clamp studies.

being treated with oral hypoglycaemic agents and none had received insulin for
control of blood glucose. Hyperglycaemic clamps (5.4 mmol/l above basal
glucose level) were performed before and at 12 weeks of GLP1 treatment. Our
preliminary data have shown that GLP1 is at least as good as the usual
hypoglycaemic agents in the control of glucose homeostasis, as demonstrated by
the amount of insulin released during the clamp and by a lowering of HbA1C
(Meneilly et al 2001c). More importantly, we obtained 2 min samples before and
238 ELAHI ET AL

6 weeks after treatment with GLP1 during a hyperglycaemic clamp in two

volunteers. Insulin release was evaluated by cluster analysis (Porksen et al 1995,
Engdahl et al 1977). The results show that GLP1 restores insulin burst amplitude
in the diabetic patients from a low level (*8 pmol/l) to those normally observed in
elderly individuals with a normal glucose tolerance (*30 pmol/l) (Meneilly et al
2001c). Thus, this hormone appears to be an excellent candidate for the treatment
of type 2 diabetes and studies from multiple research centres are currently
evaluating the long-term e⁄cacy of this hormone.
It should be noted that GLP1 has a relatively short half-life (*2 min).
Therefore, continuous infusion will most likely be necessary for its use as a
monotherapeutic agent. This will probably not be well tolerated by the patient.
However, its use will probably be most e⁄cacious as a secondary agent and this
use has several advantages which will become obvious as clinical trials continue.
We also note that several investigators/pharmaceutical companies are making
substitutions in the amino acid sequence of GLP1, which increase its half-life
substantially. However, human clinical trials with these agents have not been
reported other than for very acute administrations. Finally, there is a naturally
occurring analogue of GLP1, exendin 4, which is found in the salivary gland of
the Gila monster. This peptide has at least 10 times the potency of GLP1 and at
least 150-fold longer half-life. Acute administration of this peptide augments
insulin release markedly (Egan et al 1999). Additionally, limited experience in
type 2 diabetic patients receiving twice daily administration of this peptide for a
month has shown excellent control of glucose levels (J. M. Egan, G. S. Meneilly &
D. Elahi, unpublished results 2001). There are no data from humans on the e⁄cacy
of this peptide as a function of age, to our knowledge.
We re-a⁄rm that ageing is associated with a deterioration of glucose tolerance.
The deterioration in large part can be attributed to insulin resistance and not insulin
secretion. The cause of this deterioration is mainly attributable to lifestyle changes
(increased adiposity, loss of LBM, reduced activity, changes in diet) or to genetic
factors. That insulin resistance is not an obligatory result of ageing is best
exempli¢ed by the demonstration of maintained insulin secretion and action
across the age-span in women athletes (Ryan et al 2001) and by preserved insulin
action in healthy centenarians (102 0.8 years) compared to octogenarians
(78 0.7 years) and middle-aged volunteers (44 1.8 years) (Barbieri et al 2001,
Paolisso et al 1996).

Acknowledgements
We express our appreciation to all the volunteers who participated in our studies. We also thank
the sta¡ of the clinical research centre at our various medical centres and the excellent technical
support of our sta¡. We thank Restoragen, Inc. and Amylin Pharmaceutical, Inc. for the
generous donation of drugs and their support. The studies were also supported in part by NIH
GLUCOSE HOMEOSTASIS IN AGEING 239

grant AG-00599, the intramural research program of the NIA, a grant from the Canadian
Diabetes Association, and the Medical Research Council of Canada.

References
Andersen DK, Elahi D, Brown JC, Tobin JD, Andres R 1978 Oral glucose augmentation of
insulin secretion. Interactions of gastric inhibitory polypeptide with ambient glucose and
insulin levels. J Clin Invest 62:152^161
Andres R 1971 Aging and diabetes. Med Clin North Am 55:835^846
Barbieri M, Rizzo MR, Manzella D, Paolisso G 2001 Age-related insulin resistance: is it and
obligatory ¢nding? The lesson from healthy centenarians. Diabetes Metab Res Rev 17:19^26
Bennett PH, Rushfroth NB, Miller M, Lecompte PM 1976 Epidemiologic studies of diabetes in
the Pima Indians. Recent Prog Horm Res 32:333^376
Broughton DL, Taylor R 1991 Review: deterioration of glucose tolerance with age: the role of
insulin resistance. Age Ageing 20:221^225
Chen M, Bergman R, Pacini G, Porte DJ 1985 Pathogenesis of age-related glucose intolerance in
man: insulin resistance and decreased beta-cell function. J Clin Endocrinol Metab 60:13^20
Coon PJ, Bleecker ER, Drinkwater DT, Meyers DA, Goldberg AD 1989 E¡ects of body
composition and exercise capacity on glucose tolerance, insulin, and lipoprotein lipids in
healthy older men: a cross-sectional and longitudinal intervention study. Metabolism
38:1201^1209
Davidson MB 1979 The e¡ect of aging on carbohydrate metabolism: a review of the English
literature and a practical approach to the diagnosis of diabetes mellitus in the elderly.
Metabolism 28:688^705
DeFronzo RA 1979 Glucose intolerance and aging: evidence for tissue insensitivity to insulin.
Diabetes 28:1095^1101
DeFronzo RA 1981 Glucose intolerance and aging. Diabetes Care 4:493^501
DeFronzo RA 1988 Lilly lecture 1987. The triumvirate: beta-cell, muscle, liver. A collusion
responsible for NIDDM. Diabetes 37:667^687
DeFronzo R, Ferrannini E 1991 Insulin resistance. A multifaceted syndrome responsible for
NIDDM, obesity, hypertension, dyslipidemia, and atherosclerotic cardiovascular disease.
Diabetes Care 14:173^194
DeFronzo R, Tobin JD, Andres R 1979 Glucose clamp technique: a method for quantifying
insulin secretion and resistance. Am J Physiol 237:E214^E223
DeFronzo RA, Ferrannini E, Hendler R, Felig P, Wahren J 1983 Regulation of splanchnic and
peripheral glucose uptake by insulin and hyperglycemia in man. Diabetes 32:35^45
Dowse G, Spark R, Mavo B et al 1994 Extraordinary prevalence of non-insulin-dependent
diabetes mellitus and bimodal plasma glucose distribution in the Wanigela people of Papua
New Guinea. Med J Aust 160:767^774
Egan J, Clocquet A, Elahi D 1999 Insulinotropic e¡ect of exendin-4 in humans. Diabetologia
42:148A
Elahi D 1996 In praise of the hyperglycemic clamp. A method of assessment of beta-cell
sensitivity and insulin resistance. Diabetes Care 19:278^286
Elahi D, Muller DC 2000 Carbohydrate metabolism in the elderly. Eur J Clin Nutr 54: S112^
S120
Elahi D, Andersen DC, Brown J et al 1979 Pancreatic alpha- and beta-cell responses to GIP
infusion in normal man. Am J Physiol 237:E185^E191
Elahi D, Andersen DC, Muller DC, Tobin JD, Brown JC, Andres R 1984 The enteric
enhancement of glucose-stimulated insulin release. The role of GIP in aging, obesity and
non-insulin-dependent diabetes mellitus. Diabetes 33:950^957
240 ELAHI ET AL

Elahi D, Muller DC, McAloon-Dyke M, Tobin JD, Andres R 1993 The e¡ect of age on insulin
response and glucose utilization during four hyperglycemic plateaus. Exp Gerontol 28:393^
409
Elahi D, McAloon-Dyke M, Fukagawa N et al 1994 The insulinotropic actions of glucose-
dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (7-37) in normal
and diabetic subjects. Regul Pept 51:63^74
Engdahl J, Veldhuis J, Farrell P 1977 Altered pulsatile insulin secretion associated with
endurance training. J Appl Physiol 79:1977^1985
Evans W, Farrell P 2001 The aging pancreas: e¡ects of aging on insulin secretion action. In:
Je¡erson LS, Cherrington AD (eds) Handbook of Physiology. Section 7: The endocrine
system. Vol 2: The endocrine pancreas and regulation of metabolism. Oxford University
Press, Oxford, p 969^998
Ferrannini E, Vichi S, Beck-Nielsen H, Laakso M, Paolisso G, Smith U 1996 Insulin action and
age. European group for the study of insulin resistance (EGIR). Diabetes 45:947^953
Fink RI, Kolterman OG, Gri⁄n J, Olefsky JM 1983 Mechanisms of insulin resistance in aging. J
Clin Invest 71:1523^1535
Harris M, Flegal K, Cowie C et al 1998 Prevalence of diabetes, impaired fasting glucose, and
impaired glucose tolerance in US adults. The third national health and nutrition
examination survey, 1988^1994. Diabetes Care 21:518^524
Jackson RA, Blix PM, Matthews JA et al 1982 In£uence of ageing on glucose homeostasis. J Clin
Endocrinol Metab 55:840^848
Kaplan NM 1989 The deadly quartet. Upper-body obesity, glucose intolerance,
hypertriglyceridemia, and hypertension. Arch Intern Med 149:1514^1520
Loo SG, Dowse GK, Finch C, Zimmet P 1993 Bimodality analysis of frequency distributions of
2-hour plasma glucose concentrations in the urban Micronesian population of Kiribati. J
Diabetes Complicat 7:73^80
Maneatis T, Condie R, Reaven G 1982 E¡ect of age on plasma glucose and insulin responses to a
test mixed meal. J Am Geriatr Soc 30:178^182
Meneilly G, Minaker K, Elahi D, Rowe J 1987 Insulin action in aging man: evidence for tissue-
speci¢c di¡erences at low physiologic insulin levels. J Gerontol 42:196^201
Meneilly GS, Ryan AS, Veldhuis JD, Elahi D 1997 Increased disorderliness of basal insulin
release, attenuated insulin secretory burst mass, and reduced ultradian rhythmicity of insulin
secretion in older individuals. J Clin Endocrinol Metab 82:4088^4093
Meneilly GS, Ryan AS, Minaker KL, Elahi D 1998 The e¡ect of age and glycemic level on the
response of the b-cell to glucose-dependent insulinotropic polypeptide and peripheral tissue
sensitivity to endogenously released insulin. J Clin Endocrinol Metab 83:2925^2932
Meneilly G, Veldhuis J, Elahi D 1999 Disruption of the pulsatile and entropic modes of insulin
release during an unvarying glucose stimulus in elderly individuals. J Clin Endocrinol Metab
84:1938^1943
Meneilly G, McIntosh CHS, Pederson R et al 2001a Glucagon-like peptide-1 (7-37) augments
insulin release in elderly patients with diabetes. Diabetes Care 24:964^965
Meneilly GS, McIntosh CHS, Pederson RA et al 2001b E¡ect of glucagon-like peptide-1 on non-
insulin-mediated glucose uptake in the elderly patient with diabetes. Diabetes Care, in press
Meneilly G, Egan J, Greig N, Veldhuis J, Habener J, Elahi D 2001c Twelve Weeks of
Continuous Administration of GLP-1 in Elderly Diabetic Patients. Diabetologia 44:910A
Mojsov S, Heinrich G, Wilson I, Ravazzola M, Orci L, Habener J 1986 Preproglucagon gene
expression in pancreas and intestine diversi¢es at the level of post-translational processing.
J Biol Chem 261:11880^11889
Natali A, Toschi E, Camastra S, Gastaldelli A, Groop L, Ferrannini E 2000 Determinants of
postabsorptive endogenous glucose output in non-diabetic subjects. European group for
the study of insulin resistance (EGIR). Diabetologia 43:1266^1272
GLUCOSE HOMEOSTASIS IN AGEING 241

Nathan D, Schreiber E, Fogel H, Mojsov S, Habener J 1992 Insulinotropic action of

glucagonlike peptide-I (7-37) in diabetic and nondiabetic subjects. Diabetes Care 15:
270^276
National Diabetes Data Group 1979 Classi¢cation and diagnosis of diabetes mellitus and other
categories of glucose intolerance. Diabetes 28:1039^1057
Pacini G, Valerio A, Beccaro F, Nosadini R, Cobelli C, Crepaldi G 1988 Insulin sensitivity and
beta-cell responsivity are not decreased in elderly subjects with normal OGTT. J Am Geriatr
Soc 36:317^323
Paolisso G, Gambardella A, Ammendola S et al 1996 Glucose tolerance and insulin action in
healthy centenarians. Am J Physiol 33:E890^E894
Perley M, Kipnis D 1967 Plasma insulin responses to oral and intravenous glucose: studies in
normal and diabetic subjects. J Clin Invest 46:1954^1962
Pimenta W, Korytkowski M, Mitrakou A et al 1995 Pancreatic beta-cell dysfunction as the
primary genetic lesion in NIDDM. Evidence from studies in normal glucose tolerant
individuals with a ¢rst-degree NIDDM relative. JAMA 273: 1855^1861
Porksen N, Munn S, Steer J, Vore S, Veldhuis J, Butler P 1995 Pulsatile insulin secretion
accounts for 70% of total insulin secretion during fasting. Am J Physiol 269:E478^E488
Reaven G 1988 Banting Lecture 1988. Role of insulin resistance in human disease. Diabetes
37:1595^1607
Reaven G, Chen N, Hollenbeck C, Chen YD 1989 E¡ect of age on glucose tolerance and glucose
uptake in healthy individuals. J Am Geriatr Soc 37:735^740
Rizza RA, Mandarino LJ, Gerich JE 1981 Dose-response characteristics for e¡ects of insulin on
production and utilization of glucose in man. Am J Physiol 240:E630^E639
Rosenthal M, McMahan CA, Stern MP et al 1985 Evidence of bimodality of two hour plasma
glucose concentrations in Mexican Americans: results from the San Antonio Heart Study. J
Chronic Dis 38:5^16
Rowe J, Minaker K, Pallota J, Flier J 1983 Characterization of the insulin resistance of aging. J
Clin Invest 71:1581^1587
Ryan A, Elahi D 1996 The e¡ects of acute hyperglycemia and hyperinsulinemia on plasma leptin
levels: its relationships with body fat, visceral adiposity, and age in women. J Clin Endocrinol
Metab 81:4433^4438
Ryan AS, Egan JM, Habener JF, Elahi D 1998 Insulinotropic hormone glucagon-like peptide-
1-(7-37) appears not to augment insulin-mediated glucose uptake in young men during
euglycemia. J Clin Endocrinol Metab 83:2399^2404
Ryan A, Muller D, Elahi D 2001 Sequential hyperglycemic-euglycemic clamp to assess b-cell and
peripheral tissue: studies in female athletes. J Appl Physiol 91:872^881
Shimokata H, Muller DC, Fleg JL, Sorkin J, Ziemba AW, Andres R 1991 Age as an independent
determinant of glucose tolerance. Diabetes 40:44^51
Stout RW 1990 Insulin and atheroma. Diabetes Care 13:631^654
The Expert Committee on the Diagnosis and Classi¢cation of Diabetes Mellitus 1997 Report of
the expert committee on the diagnosis and classi¢cation of diabetes mellitus. Diabetes Care
20:1183^1197
Weir GC, Mojsov S, Hendrick GK, Habener JF 1989 Glucagonlike peptide I (7-37) actions on
endocrine pancreas. Diabetes 38:338^342
WHO 1985 Diabetes mellitus. World Health Organization, Geneva (Technical Report Series
727)
WHO 1999 De¢nition, diagnosis and classi¢cation of diabetes mellitus and its complications.
Report of a WHO consultation, part 1: Diagnosis and classi¢cation of diabetes mellitus.
World Health Organization, Geneva
Zamboni M, Armellini F, Harris T et al 1997 E¡ects of age on body fat distribution and
cardiovascular risk factors in women. Am J Clin Nutr 66:111^115
242 DISCUSSION

Zavaroni I, Dall’Aglio E, Bruschi F et al 1986 E¡ect of age and environmental factors on glucose
tolerance and insulin secretion in a worker population. J Am Geriatr Soc 34:271^275
Zimmet P, Whitehouse S 1978 Bimodality of fasting and two-hour glucose tolerance
distributions in a Micronesian population. Diabetes 27:793^800

DISCUSSION
Veldhuis: The growth hormone (GH) axis is of course much more closely related
to modulators produced by the gut, based on recent studies. For example, ghrelin,
a major GH-releasing hormone, is produced in the fundus of the stomach, except it
is inversely controlled by nutrition. It goes up with fasting and drives GH in that
circumstance.
One of the things that struck me about your paper, and which I have missed as an
endocrinologist over the years, is that the diabetic loses the ¢rst phase, but the
ageing individual appears to dominantly lose the second phase. Did I understand
this correctly?
Elahi: In the non-diabetic individual, using the hyperglycaemic clamp technique
and large numbers of patients (restricting the obesity to BMI of 425 we enrol older
people with even lower BMI), we do not see a discrepancy between ¢rst phase and
second phase in patients lower than 75 years of age. However, if we go above age 75
and don’t control for obesity we begin to see real discrepancies in the second phase.
Veldhuis: I had always supposed that ageing was a representation of an increased
probability of an non-insulin-dependent diabetes mellitus (NIDDM) phenotype
emerging on some genetic base, probably driven by increased visceral fat.
Therefore, one would have an insulin-release pro¢le matching NIDDM. Your
results suggest a di¡erent pathophysiology. On the other hand, a terribly messy
confound is the visceral obesity/changes in body composition with ageing. I
thought that it was interesting that the only study population that unmasks an
age e¡ect speci¢cally on insulin action in the periphery is the lean female
(Ferrannini et al 1996).
Elahi: That same group subsequently published a paper that looked at hepatic
glucose output as a function of this (Natali et al 2000). Total endogenous glucose
output is largely explained by the amount of lean mass, which explains di¡erences
due to obesity, sex, and age.
Veldhuis: This is somewhat reminiscent of the GH axis in adiposity studies (Vahl
et al 1997). If you take an obese individual, his GH levels are so low even in young
adulthood that one cannot see any age e¡ect at all. GH is already reduced to 20% of
normal. If however one considers the leanest cohort, the lowest quintile of
percentage body fat, there is a strong negative age e¡ect even in men (Veldhuis
et al 1995, Iranmanesh et al 1994, 1998), it isn’t restricted to women in this case
(Weltman et al 1994). This shows a strong confound with visceral obesity, again.
GLUCOSE HOMEOSTASIS IN AGEING 243

Elahi: We have just completed a study in which we examined approximately 65

females who are considered master athletes. These are Olympic level athletes who
because of their age can no longer compete with younger people. They are still very
active. We advertised, and got 65 people across the country. I can tell you
emphatically, despite the fact that the BMIs and percentage fat is the same, the
insulin release in the older group, aged 50^69, is only giving hints of not being
equal to the young.
Veldhuis: There is some visceral obesity even in well conditioned individuals. It
is a nasty confound for all of us working in this ¢eld.
Handelsman: Euglycaemic clamps essentially work on the idea of the body as a
black box. Most of the black box, as far as insulin sensitivity is concerned, is in
muscle. Are there ways to look at other tissues and their sensitivity? When we
think of sex steroids, there is increasing focus on the idea that each tissue may
have di¡erent threshold e¡ects, or at least sensitivities. Can this be done with
clamps, or are we forever having to think of the body as one big black box?
When you measure sensitivities you are talking about are essentially muscle
sensitivity, and mainly striated muscle sensitivity. Can you work out ways of
looking at the sensitivity of fat or skin, for example?
Elahi: Usually, the doses of insulin used are so high that it suppresses the hepatic
glucose output right away. In order to look at the liver you have to use very low
dose insulin clamps. We have done this, and it is possible to tease out changes in
liver sensitivity. The other way is during the insulin clamp itself. We have taken
muscle and fat biopsies and looked at incorporation of protein into muscle, for
example, before and after treatment.
Bj˛rntorp: One can measure glucose uptake in adipose tissue by using small
amounts of radioactive glucose. It is very small: less than 5%.
Laron: Would you say that the normal progression of ageing, during which the
muscle mass is decreased and adipose tissue mass is increased, plays a role in the
development of the insulin resistance?
Elahi: We have consecutive two year OGTTs for diagnoses of diabetes for
*10^25 years. If you look at the fasting glucose and decide when there is a
diagnosis of diabetes, and compare this with the 2 h value, do you have any
idea what the discrepancy is with respect to time, i.e. when you can detect
diabetes with the 2 h glucose level and the other one with fasting level? Nine
years. Similarly, you can pick up impaired tolerance seven years earlier if you
use the 2 h glucose level from the OGTT rather than the fasting glucose level.
No one knows when someone becomes diabetic; if you have a test you are either
normal or diabetic. We can also look at how long it takes to go from impaired to
diabetic. Let us imagine that you are impaired, as diagnosed either from fasting
levels or the 2 h OGTT levels. How long does it take to go from impaired to
diabetic? Again, on the average seven years. If you put the two together, i.e. both
244 DISCUSSION

fasting and OGTT, as well as impaired and diabetic diagnosis, one can have an
early warning something like 7^16 years before an individual progresses from a
normal test to an impaired and ¢nally a diabetic diagnosis. Imagine what we
could do in terms of interventions if we have this much time.
Prior: I remember the ¢rst criteria for diabetes that were created with data in
young people. These individuals weren’t carbohydrate-loaded so they had a
starved liver. How reproducible are the OGTTs? Does the reproducibility
decrease as you get older?
Elahi: We have looked at that. There is movement over two years. It is di⁄cult
to control because the doctors write reports to the participants telling them that
they are impaired or diabetic, and there is no way of knowing whether these people
have made any life function adjustments in the light of this. In general, there is
approximately 80% concurrence of the diagnosis and 30% who move back and
forth.
Prior: Has anyone studied this at weekly intervals over an extended period?
Elahi: Not that I know of. In general the reproducibility is terrible. This test is
essentially for diagnosis of diabetes.
Prior: It is a tremendous social and ¢nancial burden if you are diagnosed with
diabetes. It changes your health insurance and your concept of yourself. What I see
in clinical practice is people being diagnosed who are elderly who probably don’t
have diabetes. We need to be sure.
Veldhuis: This brings up the problem that I am still struggling with, namely the
notion of what is de¢ned as ‘normal’. This comes up in the bone ¢eld. Because of
the high predictive risk of z scores against the young population, we are willing to
intervene based on a comparison between older and young skeletal mass values. In
many endocrine areas we haven’t reached this predictive capability yet. With IGF
we clearly say that the age-related norms are appropriate because IGF collapses
over decades. This is an interesting issue. By one of your classi¢cations, 80% of
the elderly would have impaired glucose tolerance or diabetes. In a sense, this is a
philosophical issue of when to use a young adult range. For the thyroid it is clear
because the range looks ¢xed. We are struggling with testosterone.
Handelsman: This is an important issue. In the end, a lot of what we do is geared
to ill health-treatment and prevention of disease. In a sense the hard clinical
endpoints are an important reference point we mustn’t lose sight of. In the bone
¢eld, we are talking about fractures, disability and death. But the concept of a
surrogate variable is very important. Bone density is a well-accepted surrogate
variable. We have other surrogate variables that work well, such as total
cholesterol for heart disease. A lot of other things that we think of as
intermediary factors are not accepted because they are not sensitive or speci¢c
enough as surrogate variables. What we are really talking about is de¢ning
convenient surrogate variables. The question here is what is the right surrogate
GLUCOSE HOMEOSTASIS IN AGEING 245

variable for disability and death from diabetes in a 70 year old? Does it make sense
to talk about long-range micro- and macrovascular complications? What is the
predictive value?
Carroll: Of all the competing de¢nitions of diabetes, the only practical thing to
do is to look at them with respect to outcome. This is how one validates criteria.
Veldhuis: I think David Handelsman’s remark on this is compelling for clinical
issues at least. If you can ¢nd me a determinant that predicts adverse consequence,
for which I have either a putative intervention or a known intervention, I become
interested in that marker. It is embarrassing to be hung up at such a basic practical
level, and not to go a little bit deeper into what is the normal. We were wise to have
the Rotterdam group begin by asking how does a healthy older body behave? Give
me a few hundred such observations to compare with a subset in the general
population of similar age, who manifest measurable di¡erences in a body
parameter and are complaining about it clinically, and I will begin to construe a
possible relationship. I am beginning to think we need more normative data in
understanding ageing and normal physiology. Thus, as Dariush Elahi’s talk
illustrates, we need to know the predictive value of biochemical anomalies.
Prior: One of the things that may be of value in the elderly is standardizing
glycosylated haemoglobin levels. It gives a broader picture. My understanding is
that it hasn’t become a tool because the laboratory diagnosis is not secure.
Elahi: You are correct; that is the problem. I imagine that in the next 10 years it
will become the standard marker for diagnosis of diabetes. Having said this, I
currently know of at least three publications that cross-sectionally examine
haemoglobin as a function of age and BMI. There is an increase in haemoglobin
A1C (HbA1C) as a function of age both in the lean and the obese, and it is obesity
independent. The values di¡er between the populations because of the non-
standardization, but the shape of the curve doesn’t change.
Prior: Are there any HbA1C data in the Rotterdam study?
Haus: In some populations the HbA1C values may be less reliable than the
measurement of glycated haemoglobin due to the higher percentages of
abnormal haemoglobins. Which determination method would be preferable for
general use?
Elahi: Whichever one can be normalized across the world the fastest.
Laron: Ageing is a normal biological process. It includes an adaptation of the
body to the progressive changes; if we do not establish norms along each phase
of ageing, but rather take one point as the normal, we must conclude that ageing
is a disease and not a normal state.
Elahi: There is a nice paper that has a normogram that gives you a
recommendation for diagnosis diabetes not only as a function of age but also for
each gender. This was before the 1979 ADA criteria. Before 1979 the criterion for
diagnosis was a 2 h glucose level 5140 mg/dl. This is why this normogram was
246 DISCUSSION

good because it really shifted as you got older. The 1979 ADA raised the 2 h level to
200 mg/dl, so there was no more age adjustment necessary.
Laron: The basic question is whether we should adapt subjects to ‘normality’
along the ageing process?
Elahi: I think we should, but it would be hard to implement this in the general
practitioners’ o⁄ces across the world.
Handelsman: This is an important and familiar discussion, but it goes round in a
circle because it lacks the reference point of prevention of disability and death. This
breaks that vicious circle. We can easily say that all parameters should be age
adjusted or not, but in the end the criteria for which of these is better is going to
be which prevents disease and disability best. Without this framework, no
meaningful answer is possible; it is an internal vicious circle.

References
Ferrannini E, Vichi S, Beck-Nielsen H, Laakso M, Paolisso G, Smith U 1996 Insulin action and
age. European Group for the Study of Insulin Resistance (EGIR). Diabetes 45:947^953
Iranmanesh A, Grisso B, Veldhuis JD 1994 Low basal and persistent pulsatile growth hormone
secretion are revealed in normal and hyposomatotropic men studied with a new ultrasensitive
chemiluminescence assay. J Clin Endocrinol Metab 78:526^535
Iranmanesh A, South S, Liem AY et al 1998 Unequal impact of age, percentage body fat, and
serum testosterone concentrations on the somatotropic, IGF-I, and IGF-binding protein
responses to a three-day intravenous growth-hormone-releasing hormone pulsatile infusion
in men. Eur J Endocrinol 139:59^71
Natali A, Toschi E, Camastra S, Gastaldelli A, Groop L, Ferrannini E 2000 Determinants of
postabsorptive endogenous glucose output in non-diabetic subjects. European Group for
the Study of Insulin Resistance (EGIR). Diabetologia 43:1266^1272
Vahl N, Jorgensen JO, Skjaerbaek C, Veldhuis JD, Orskov H, Christiansen J 1997 Abdominal
adiposity rather than age and sex predicts the mass and regularity of GH secretion in healthy
adults. Am J Physiol 272:E1108^E1116
Veldhuis JD, Liem AY, South S et al 1995 Di¡erential impact of age, sex-steroid hormones, and
obesity on basal versus pulsatile growth hormone secretion in men as assessed in an
ultrasensitive chemiluminescence assay. J Clin Endocrinol Metab 80:3209^3222
Weltman A, Weltman JY, Hartman ML et al 1994 Relationship between age, percentage body
fat, ¢tness and 24 hour growth hormone release in healthy young adults: e¡ects of gender.
J Clin Endocrinol Metab 78:543^548
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Endocrine causes of age-related bone

loss and osteoporosis
B. Lawrence Riggs

Endocrine Research Unit, Mayo Clinic and Mayo Foundation, North 6 Plummer, Rochester,
MN 55905, USA

Abstract. Women have an early postmenopausal phase of rapid bone loss that lasts for
5^10 years after menopause, whereas both ageing women and men have a slow continuous
phase of bone loss that lasts inde¢nitely. In women, the rapid phase is mediated mainly by
loss of the direct restraining e¡ect of oestrogen on bone cell function, whereas the slow
phase is mediated mainly by the loss of oestrogen action on extraskeletal calcium
homeostasis leading to net calcium wasting and secondary hyperparathyroidism.
Because elderly men have low serum bioavailable oestrogen and testosterone levels, and
because recent data suggest that oestrogen is the main sex steroid regulating bone
metabolism in men, oestrogen de¢ciency may also be the principal cause of bone loss in
elderly men. Decreased bone formation contributes to bone loss in both genders and may
be caused by a decreased production of growth hormone and IGF1 as well as oestrogen
and testosterone de¢ciency. Other changes in endocrine secretion, although present in the
elderly, seem less important in the pathophysiology of age-related bone loss and
osteoporosis.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 247^264

The problem of osteoporosis

Osteoporosis is one of the major problems facing ageing women and men. In the
USA, osteoporosis annually causes *1.5 million fractures and costs *US$15
billion to the healthcare system. The lifetime risk of fractures of the spine
(symptomatic), hip and distal radius is 40% for white women and 13% for white
men (Melton et al 1988 and Table 1).

Patterns of age-related bone loss

The pattern of age-related bone loss varies between genders. Women undergo two
phases of bone loss
an early, accelerated transient phase followed by a slow
continuous phase
whereas men undergo only a slow continuous phase (Riggs
et al 1998). The accelerated bone loss in the early postmenopausal phase decreases
247
248 RIGGS

TABLE 1 Lifetime risk of major fractures due to osteoporosis for white

men and women at age 50

Women (%) Men (%)

Hip fracture 17.5 6.0

Vertebral fracture 15.6 5.0
Forearm fracture 16.0 2.5
Any of the three 39.7 13.1

exponentially over 5^10 years to merge asymptotically with the subsequent slow
phase that continues inde¢nitely. This early phase involves predominantly
cancellous bone loss; it accounts for losses of 20^30% of cancellous bone but only
of 5^10% of cortical bone. It is associated with high bone turnover and the increase
in bone resorption is greater than the increase in bone formation. The reason for
the slowing and eventual cessation of the rapid phase of bone loss is unclear.
However, it is likely that with the rapid depletion of the cancellous bone,
biomechanical forces act to limit further loss.
The late slow phase in women involves equal losses (*20^25% each) of cortical
and cancellous bone over life. The slow continuous phase in men is similar to the
late slow phase in postmenopausal women, both with respect to its rate and time
course and with respect to the type and amount of bone loss. The slow phases of
bone loss in both genders are associated with high bone turnover. These patterns
are shown schematically in Fig. 1.

Endocrine abnormalities and bone loss in women

Early accelerated phase

This phase begins at menopause, can be prevented by oestrogen replacement, and
almost certainly results from the cessation of ovarian function. Oestrogen acts
through high a⁄nity oestrogen receptors in osteoblasts and osteoclasts to
restrain bone turnover, and when this restraint is lost at menopause, overall bone
turnover increases and resorption increases more than formation. In addition, the
increased activity of osteoclasts and their prolonged lifespan lead to trabecular
plate perforation and to loss of structural elements, thus weakening bone out of
proportion to the loss of bone density. The high rate of bone resorption increases
skeletal calcium out£ow, which leads to a partial suppression of parathyroid
hormone (PTH) secretion and compensatory increases in urinary calcium
excretion (Riggs et al 1998). The reason for the cessation of the rapid phase of
ENDOCRINE CAUSES OF AGE-RELATED BONE LOSS 249

FIG. 1. Schematic representation of changes in bone mass over life in cancellous (broken line)
and cortical (solid line) bone in women (left panel) and men (right panel) from age 50 onward.
Note that men have only one phase of continuous bone loss but women have two
an early
accelerated phase and a late slow phase. Note also that the accelerated phase, but not the slow
phase, involves disproportionate loss of cancellous bone. (With permission from Riggs et al
1998.)

bone loss is unclear but may relate to the activation of biomechanical forces that
limit the rate of further bone loss when the amount of cancellous bone falls below
some critical level.

Late slow phase

As the rapid bone loss phase subsides, serum levels of PTH increase progressively
throughout the remainder of life. Markers for bone turnover also increase and
these increases are directly correlated with the increase in serum PTH (Table 2).
Suppression of PTH secretion by intravenous calcium infusion abolishes the
di¡erences in bone resorption markers between young and elderly women
(Ledger et al 1994), strongly suggesting that the increase in bone resorption in
ageing women is PTH-dependent. Also, a chronic, high calcium intake will also
reduce elevated values for serum PTH in elderly women into the premenopausal
range (McKane et al 1995). Thus, a considerable body of data implicates secondary
hyperparathyroidism as the cause of the late slow phase of bone loss in elderly
women. The secondary hyperparathyroidism has traditionally been assumed to
be secondary to age-related factors that impair calcium absorption and renal Ca2+
homeostasis. However, the recent ¢ndings that the increases in both serum PTH
and bone resorption in late postmenopausal women can be normalized by
oestrogen replacement suggests that oestrogen de¢ciency may play a causal role
(McKane et al 1997, Khosla et al 1998; Table 3).
250 RIGGS

TABLE 2 Changes in serum PTH and in biochemical markers for bone turnover in
304 women residents of Rochester, MN from the third into the tenth decade of life

Spearman correlation coe⁄cients

Variable Increase with D (%) vs. age vs. PTH

PTH 54% 0.354** 1.000

BSAP 38% 0.329** 0.192*
OC 64% 0.392** 0.206*
fPYD 76% 0.505%** 0.203*
NTx 86% 0.344** 0.190*

Samples are age-strati¢ed and population-based. Postmenopausal women receiving oestrogen replacement
are not included. There are age-related increases in serum PTH and in markers for bone formation (serum
bone speci¢c alkaline phosphatase [BSAP] and osteocalcin [OC]) and bone resorption (urine free
pyridinoline [fPYD], and cross-linked N-terminal telopeptide of type I collagen [NTx]). Furthermore, the
increases in biochemical markers are directly correlated with the increases in serum PTH. (Data are from re-
analysis of study of Khosla et al 1997.)
*P50.0001, **P50.001.

Both the early accelerated and the late slow phases of bone loss in women are
associated with low levels of serum oestrogen, and oestrogen treatment is
e¡ective in preventing further bone loss in both phases (McKane et al 1997,
Ettinger et al 1985). However, in the early phase of bone loss there is a trend to
decreased serum PTH and these levels increase following oestrogen treatment. By

TABLE 3 Comparative e¡ects of age and oestrogen status in women as described by

McKane et al (1997)

Postmenopausal Postmenopausal
Variable Premenopausal untreated treated

N 30 30 30
Age (years) 32.0 0.5 74.2 0.6 73.8 0.6
Serum:
PTH (pmol/l) 2.7 0.2 3.6 0.3* 2.5 0.2
Urine:
NTx (nmol/mmol Cr) 28.8 2.3 42.9 3.5** 24.6 2.3
PYD (nmol/mmol Cr) 45.6 1.6 61.2 3.2** 40.7 1.6
DPD (nmol/mmol Cr) 11.9 0.5 16.2 1.0** 9.4 0.5

Serum intact PTH was fasting morning value. Bone resorption markers were measured by ELISA kit for N-
telopeptide of type I collagen (NTx) and by £uorometric detection after HPLC for pyridinoline (PYD) and
deoxypyridinoline (DPD). All results are mean SEM.
For di¡erence from premenopausal groups: *P50.05, **P50.005.
ENDOCRINE CAUSES OF AGE-RELATED BONE LOSS 251

contrast, serum PTH levels increase progressively in the late slow phase of bone
loss and are decreased by oestrogen treatment. We have attempted to resolve this
paradox by hypothesizing that oestrogen has two major e¡ects on bone
a direct
action on bone cells and an action on external calcium homeostasis that increases
PTH secretion and indirectly increases bone loss (Riggs et al 1998). The loss of the
direct e¡ect of oestrogen on bone cells is responsible for the early rapid phase of
bone loss. This direct e¡ect is initiated by the large and relatively rapid fall in serum
oestrogen levels at menopause, and it becomes less important after the rapid
accelerated phase of bone loss subsides. In the late slow phase, the indirect e¡ect
of oestrogen de¢ciency on extra-skeletal calcium metabolism leads to the secondary
hyperparathyroidism that is the major cause of the bone loss.
The indirect e¡ect is initiated by loss of oestrogen action on the intestine
and kidney (Riggs et al 1998). The intestine contains oestrogen receptors and
oestrogen stimulates calcium absorption (Gennari et al 1990). Oestrogen also
increases renal calcium conservation (McKane et al 1995) through a PTH-
independent enhancement of tubular reabsorption of calcium. During oestrogen
de¢ciency, the loss of the intestinal and renal actions of oestrogen leads to external
losses of calcium and negative calcium balance. Unless these losses of calcium are
o¡set by large increases in dietary calcium (McKane et al 1996), they will lead to
secondary hyperparathyroidism and continued bone loss.

Decreased bone formation

Both the early accelerated and the late slow phases of bone loss are associated with
an absolute increase in bone resorption and a relative decrease in bone formation.
Normally, there is a tight coupling of bone formation to bone resorption. During
oestrogen de¢ciency, however, there is a failure of a compensatory increase in bone
formation to o¡set the increase in bone resorption and this leads to continued bone
loss. This abnormality is demonstrable soon after menopause suggesting that it is
caused by oestrogen de¢ciency. Oestrogen has been shown to increase collagen
synthesis in skin ¢broblasts and to increase production of insulin-like growth
factor (IGF)1 (Ernst et al 1989) and transforming growth factor (TGF)b
(Ashcroft et al 1997), growth factors that are anabolic for osteoblasts. Also,
oestrogen prolongs the lifespan of mature osteoblasts by decreasing apoptosis
(Manolagas 2000). Oestrogen de¢ciency could also account for the impaired
osteoblastic function in older women, although it is possible that age-related
abnormalities in hormones or growth factors regulating osteoblast function also
contribute.
The probable mechanisms by which oestrogen de¢ciency produces bone loss in
ageing women are shown in Fig. 2A.
252 RIGGS

FIG. 2. Schematic representation of unitary model for bone loss in postmenopausal women
(A) and in ageing men (B). See text for details. (With permission from Riggs et al 1998.)

Endocrine abnormalities and bone loss in ageing men

Except after orchiectomy, men do not have an equivalent of the rapid phase of
bone loss that women experience following menopause. After accounting for the
absence of this phase, the patterns of late bone loss and of the increases in serum
PTH and bone resorption markers in ageing men are virtually superimposable
upon those occurring in women (Riggs et al 1998). In the past, it has been
di⁄cult to attribute male bone loss to sex steroid de¢ciency because men do not
have an equivalent of menopause, and because serum total testosterone levels
ENDOCRINE CAUSES OF AGE-RELATED BONE LOSS 253

TABLE 4 Gender di¡erences in changes over life in sex steroids

Men % change Women % change

Lateral spine BMD 727** 745**

Serum:
Bio E 747** 783**
Bio T 764** 728*
SHBG +124** 71
LH +285** +731**
FSH +505** +1805**

*P50.05, **P50.005.
Data are adapted from Khosla et al (1998).

decrease only marginally with age except for a small subset of elderly men who
develop clinical hypogonadism. However, in the last 5 years, thinking on this
issue has undergone a sea change.
First, in population studies, we (Khosla et al 1998) and others have shown that
although levels of serum total testosterone and oestrogen decrease only slightly in
men with ageing, there are major decreases in biologically available levels of both
sex steroids (Table 4, Figs 3 and 4). This disassociation is due to a progressive
increase in men in levels of the fraction of sex steroids bound to serum sex
hormone binding globulin (SHBG) (Fig. 5) which is not available to tissues. The

FIG. 3. Changes in serum testosterone (T) in ageing men. The left panel shows that serum total
testosterone decreases only slightly in ageing men. Total T (P50.001), bioavailable T
(P50.001). The right panel shows the changes in serum bioavailable testosterone, which
decreased progressively with ageing. Values for premenopausal (Pre) and postmenopausal
(Post) women are given for comparison. Error bars represent SEM. (With permission from
Riggs et al 2000.)
254 RIGGS

FIG. 4. Changes in serum oestrogen (E) in ageing men. The left panel shows that serum total
oestrogen decreases only slightly in ageing men. Total E (not signi¢cant), bioavailable E
(P50.001). The right panel shows the changes in serum bioavailable oestrogen, which
decreased progressively with ageing. Values for premenopausal (Pre) and postmenopausal
(Post) women are given for comparison. Error bars represent SEM. (With permission from
Riggs et al 2000.)

FIG. 5. Changes in sex hormone binding globulin (SHBG) in ageing men. Note the
progressive increase in SHBG which binds to approximately 50% of total serum oestrogen or
serum testosterone, rendering it largely unavailable to target tissues. The remaining 50%, which
is bound to albumin or is free, is bioavailable. Thus, the progressive increases in serum SHBG
with ageing are a major reason for the progressive de¢ciency in bioavailable testosterone and
oestrogen in ageing men. Error bars represent SEM. (With permission from Riggs et al 2000.)

mechanisms for the increase in serum SHBG are complex but are probably due, in
part, to decreases in production of growth hormone and IGF1 coupled with an
impaired gondal secretory reserve (Table 5). Thus, as assessed by their free or
bioavailable levels, about half of elderly males have a substantial de¢ciency of
oestrogen and testosterone, and, in general, these are the ones that are losing
bone (Khosla et al 2000).
ENDOCRINE CAUSES OF AGE-RELATED BONE LOSS 255

TABLE 5 Di¡erences between genders: mechanisms of sex steroid de¢ciency

Females Males

Onset Begins acutely at menopause Gradual and progressive

Oestrogen de¢ciency ++++ ++
Testosterone de¢ciency ++ ++++
D, SHBG 0 +++
Mechanism (a) Inactivation by SHBG
Ovarian failure (b) Impaired hypothalamic,
pituitary, gonad axis

Data from several recent ‘experiments of nature’ are consistent with the concept
that oestrogen plays a major role in maintaining bone mass in men. A young adult
who was unable to respond to oestrogen because of homozygous mutations of
oestrogen receptor a genes (Smith et al 1994) and two young adults who were
unable to synthesize oestrogen because of homozygous mutations of the
aromatase genes (Carani et al 1997, Morishima et al 1997) had osteopenia despite
normal or elevated levels of testosterone. Moreover, Vanderschueren et al (1997)
found no di¡erences in the e¡ects of orchiectomy or treatment with aromatase
inhibitor on decreasing bone density in aged male rats, suggesting that the
aromatization of androgens to oestrogens was playing a major role in skeletal
maintenance.
Four recent population-based, observational studies (Khosla et al 1998,
Slemenda et al 1997, Center et al 1997, Greendale et al 1997) involving an
aggregate total of 1410 men from young adulthood to old age found by
multivariate analysis that free serum oestrogen rather than free serum
testosterone was the main predictor of bone mass at all measured sites except
some cortical bone sites in the appendicular skeleton. Khosla et al (2000) have
also demonstrated that this also applies to the rate of bone loss in elderly men.
Finally, our group (Falahati-Nini et al 2000) has recently shown that when a
group of elderly men were pharmacologically rendered hypogonadal and their
aromatase activity was blocked, oestrogen, but not testosterone, prevented an
increase in bone resorption markers. Collectively, these studies provide
convincing evidence that a de¢ciency in oestrogen is a major cause of bone loss
in ageing men. Interestingly, Bernecker et al (1995) found that mean levels of
serum oestrogen but not testosterone were signi¢cantly reduced in 56 men with
established idiopathic osteoporosis.
Collectively, these data support the hypothesis that oestrogen de¢ciency plays a
major role in involutional bone loss in men as well as in women. However,
testosterone clearly accounts for the sexual dimorphism of the skeleton that
256 RIGGS

develops following puberty and probably also stimulates the subsequent periosteal
growth of cortical bone (Seeman 1997). In addition, we (Khosla et al 1998) have
found that testosterone de¢ciency is the main determinant of the predominantly
cortical bone mass of the appendicular skeleton. More studies must be made to
de¢ne the additional e¡ects of testosterone on the male skeleton and to determine
the relative contributions of de¢ciencies of oestrogen and testosterone in causation
of the slow phase of bone loss in ageing men.
The probable mechanisms by which sex steroid de¢ciency produces bone loss in
ageing men are shown in Fig. 2B.

Other age-related endocrine abnormalities

Although decreases in serum sex steroids and increases in serum PTH are by far the
most important endocrine abnormalities causing age-related bone loss, there are
other abnormalities that contribute variably. The two most important of these
are those of the vitamin D^endocrine system and the growth hormone^IGF1
system. Reduced serum concentrations of both the active vitamin D
metabolites
25-hydroxyvitamin D (25[OH]D and 1,25-dihydroxyvitamin D
(1,25[OH]2D)
have been demonstrated in both genders with ageing. Serum
25(OH)D is an indicator of vitamin D nutrition. Several population based
studies have shown that 25(OH)D decreases by 30^60% with ageing in both
genders (Tsai et al 1987). This may contribute to the secondary
hyperparathyroidism of ageing because these decreases correlate inversely with
serum PTH levels (Khosla et al 1998). Elderly, housebound persons with
inadequate exposure to ultraviolet radiation and poor nutrition are particularly
prone to vitamin D de¢ciency which, unless severe, is manifested by
osteoporosis, rather than osteomalacia. This is particularly likely in populations
living in higher latitudes, such as Great Britain and France, that do not fortify
milk products with vitamin D. Indeed, Chapuy et al (1992) have demonstrated
that supplementing the diet of elderly, house-bound women from Lyon, France
with 800 U/day of vitamin D and 1000 mg/day of calcium for 18 months
decreased the incidence of hip fracture by 43% over the following 18 months.
Also, serum levels of the physiologically active vitamin D metabolite,
1,25(OH)2D decrease with ageing (Manolagas et al 1983), at least relative to the
concomitant increases in serum PTH (Eastell et al 1991). Because infusions with
the trophic hormone, PTH, result in a blunting of the increase in serum
1,25(OH)2D levels relative to young adults (Tsai et al 1984), ageing may result in
a primary de¢ciency in the renal enzyme, 25(OH)D 1a-hydroxylase, that is
responsible for the conversion of 25(OH)D to 1,25(OH)2D, and this may also
contribute to the secondary hyperparathyroidism and increased bone resorption
with ageing.
ENDOCRINE CAUSES OF AGE-RELATED BONE LOSS 257

There is indisputable evidence that ageing decreases the amplitude and

frequency of growth hormone secretion (Thorner & Vance 1988). There is also a
60% decrease in serum IGF1 levels with ageing and a smaller decrease in serum
IGF2 levels (Bennett et al 1984). Whether the decreased serum IGF1 levels are
due to a decreased production of hepatic or skeletal IGF1 production or
increased plasma clearance is still unclear. Nonetheless, these abnormalities may
contribute to the decreases in osteoblastic function in elderly men and women.
They may also contribute to the progressive, age-related increases in serum
SHBG levels in men which is a major cause of their decreases in serum
bioavailable sex steroids and, thus, to their slow phase of bone loss.
Other changes in endocrine function with ageing appear to make smaller
contributions to bone loss. There are decreases of almost 70% in serum
dehydroepiandrosterone (DHEA) and DHEA sulfate in elderly women and men
(Meikle et al 1991). However, this is a relatively weak adrenal androgen and, its
importance, if any, in bone loss is problematic. In contrast to the decrease in
bioavailable anticatabolic sex steroids with ageing, the concentration of plasma-
free cortisol, a catabolic steroid, increases by 20^50% (Van Cauter et al 1996).
Thus, the increased catabolic/anti-catabolic ratio of circulating steroid hormones
with ageing could contribute to bone loss.

Non-endocrine age-related abnormalities

Although endocrine factors appear to be the major cause of age-related bone loss,
there are important non-endocrine factors that also contribute. The level of bone
mass present prior to the onset of age-related bone loss is clearly important: those
persons who have high levels are relatively protected against osteoporosis whereas
those with low levels are clearly at a greater risk. As has been long recognized, there
are a number of episodic factors that increase bone loss in some, but not other,
members of the ageing population. These include use of certain drugs such as
corticosteroids, diseases such as malabsorption, anorexia nervosa and renal
hypercalciuria, and behavioural factors such as smoking, alcohol abuse and
inactivity
to enumerate but a few. These may make major contributions to
fractures in about 40% of men and 20% of women (Riggs et al 1986).

Conclusion
In women, oestrogen de¢ciency due to the menopause is the major cause of both
the early rapid and late slow phases of age-related bone loss. In ageing men,
oestrogen de¢ciency also appears to be the dominant cause and is due to age-
related increases in serum SHBG and to impaired gonadal production of sex
steroids. The role played by decreases in bioavailable testosterone in bone loss in
258 RIGGS

men is presently unclear. Thus, although other factors contribute, age-related

osteoporosis appears to be primarily an endocrine de¢ciency disease in both
genders.

Acknowledgement
This work was supported by National Institutes of Health grants AR27065 and AG04875.

References
Ashcroft GS, Dodsworth J, van Boxtel E et al 1997 Estrogen accelerates cutaneous wound
healing associated with an increase in TGF-b1 levels. Nat Med 3:1209^1215
Bennett A, Wahner HW, Riggs BL, Hintz RL 1984 Insulin-like growth factors I and II: aging
and bone density in women. J Clin Endocrinol Metab 59:701^704
Bernecker PM, Willvonseder R, Resch H 1995 Decreased estrogen levels in male patients with
primary osteoporosis. J Bone Miner Res 10:S445
Carani C, Qin K, Simoni M et al 1997 E¡ect of testosterone and estradiol in a man with aromatase
de¢ciency. N Engl J Med 337:91^95
Center JR, Nguyen TV, White CP, Eisman JA 1997 Male osteoporosis predictors: sex hormones
and calcitropic hormones. J Bone Miner Res 12:S368
Chapuy MC, Arlot ME, Duboeuf F et al 1992 Vitamin D3 and calcium to prevent hip fractures in
elderly women. N Engl J Med 327:1637^1642
Eastell R, Yergey AL, Vieira N, Cedel SL, Kumar R, Riggs BL 1991 Interrelationship among
vitamin D metabolism, true calcium absorption, parathyroid function and age in women:
evidence of an age-related intestinal resistance to 1,25-dihydroxyvitamin D action. J Bone
Miner Res 6:125^132
Ernst M, Heath JK, Rodan GA 1989 Estradiol e¡ects on proliferation, messenger ribonucleic
acid for collagen and insulin-like growth factor-I, and parathyroid hormone-stimulated
adenylate cyclase activity in osteoblastic cells from calvariae and long bones. Endocrinology
125:825^833
Ettinger B, Genant HK, Cann CE 1985 Long-term estrogen replacement therapy prevents bone
loss and fractures. Ann Intern Med 102:319^324
Falahati-Nini A, Riggs BL, Atkinson EJ, O’Fallon WM, Eastell R, Khosla S 2000 Relative
contributions of testosterone versus estrogen in regulating bone resorption and formation
in normal elderly men. J Clin Invest 106:1553-1560
Gennari C, Agnusdei D, Nardi P, Civitelli R 1990 Estrogen preserves a normal intestinal
responsiveness to 1,25-dihydroxyvitamin D3 in oophorectomized women. J Clin
Endocrinol Metab 71:1288^1293
Greendale GA, Edelstein S, Barrett-Connor E 1997 Endogenous sex steroids and bone mineral
density in older women and men: the Rancho Bernardo Study. J Bone Miner Res 12:1833^
1843
Khosla S, Melton LJ III, Atkinson EJ, O’Fallon WM, Klee GG, Riggs BL 1998 Relationship of
serum sex steroid levels and bone turnover markers with bone mineral density in men and
women: a key role for bioavailable estrogen. J Clin Endocrinol Metab 83:2266^2274
Khosla S, Atkinson EJ, Melton LJ III, Riggs BL 1997 E¡ects of age and estrogen status on
serum parathyroid hormone levels and biochemical markers of bone turnover in women: a
population-based study. J Clin Endocrinol Metab 82:1522^1527
Khosla S, Melton LJ III, Atkinson, EJ, O’Fallon WM, Riggs BL 2000 Relationship of sex
steroids to longitudinal changes in bone mineral density and bone resorption in young
ENDOCRINE CAUSES OF AGE-RELATED BONE LOSS 259

versus elderly men: e¡ects of estrogen on peak bone mass and on age-related bone loss. J Bone
Miner Res 15:S159
Ledger GA, Burritt MF, Kao PC et al 1994 Abnormalities of parathyroid hormone secretion in
elderly women that are reversible by short term therapy with 1,25-dihydroxyvitamin D3. J
Clin Endocrinol Metab 79:211^216
Manolagas SC 2000 Birth and death of bone cells: basic regulatory mechanisms and implications
for the pathogenesis and treatment of osteoporosis. Endocr Rev 21:115^137
Manolagas SC, Culler FL, Howard JE, Brickman AS, Deftos LJ 1983 The cytoreceptor assay for
1,25-dihydroxyvitamin D and its application to clinical studies. J Clin Endocrinol Metab
56:751^760
McKane WR, Khosla S, Burritt MF et al 1995 Mechanism of renal calcium conservation with
estrogen replacement therapy in women in early postmenopause
a clinical research center
study. J Clin Endocrinol Metab 80:3458^3464
McKane WR, Khosla S, Egan KS, Robins SP, Burritt MF, Riggs BL 1996 Role of calcium intake
in modulating age-related increases in parathyroid function and bone resorption. J Clin
Endocrinol Metab 81:1699^1703
McKane WR, Khosla S, Risteli J, Robins SP, Muhs JM, Riggs BL 1997 Role of estrogen
de¢ciency in pathogenesis of secondary hyperparathyroidism and increased bone resorption
in elderly women. Proc Assoc Am Physicians 109:174^180
Meikle AW, Daynes RA, Araneo BA 1991 Adrenal androgen secretion and biologic e¡ects.
Endocrinol Metab Clin North Am 20:381^400
Melton LJ III, Kan SH, Wahner HW, Riggs BL 1988 Lifetime fracture risk: An approach to hip
fracture risk assessment based on bone mineral density and age. J Clin Epidemiol 41:985^994
Morishima A, Grumbach MM, Bilezikian JP 1997 Estrogen markedly increases bone mass in an
estrogen de¢cient young man with aromatase de¢ciency. J Bone Miner Res 12:S126
Riggs BL, Melton LJ 1986 Involutional osteoporosis. N Engl J Med 314:1676^1686
Riggs BL, Khosla S, Melton LJ III 1998 A unitary model for involutional osteoporosis: estrogen
de¢ciency causes both type I and type II osteoporosis in postmenopausal women and
contributes to bone loss in aging men. J Bone Miner Res 13:763^773
Riggs BL, Khosla S, Melton LJ III 2000 Primary osteoporosis in men: role of sex steroid
de¢ciency. Mayo Clin Proc 75:S46^S50
Seeman E 1997 From density to structure: growing up and growing old on the surfaces of bone.
J Bone Miner Res 12:509^521
Slemenda CW, Longcope C, Zhou L, Hui SL, Peacock M, Johnston C 1997 Sex steroids and
bone mass in older men. Positive associations with serum estrogens and negative associations
with androgens. J Clin Invest 100:1755^1759
Smith EP, Boyd J, Frank GR et al 1994 Estrogen resistance caused by a mutation in the estrogen-
receptor gene in a man. N Engl J Med 331:1056^1061 (erratum: 1995 N Engl J Med 332:131)
Thorner MO, Vance ML 1988 Growth hormone, 1988. J Clin Invest 82:745^747
Tsai KS, Heath H III, Kumar R, Riggs BL 1984 Impaired vitamin D metabolism with
aging in women. Possible role in pathogenesis of senile osteoporosis. J Clin Invest 73:
1668^1672
Tsai KS, Wahner HW, O¡ord KP, Melton LJ III, Kumar R, Riggs BL 1987 E¡ect of aging on
vitamin D stores and bone density in women. Calcif Tissue Int 40:241^243
Van Cauter E, Leproult R, Kupfer DJ 1996 E¡ects of gender and age on the levels and circadian
rhythmicity of plasma cortisol. J Clin Endocrinol Metab 81:2468^2473
Vanderschueren D, Van Herck E, Nijs J, Ederveen AGH, De Coster R, Bouillon R 1997
Aromatase inhibition impairs skeletal modelling and decreases bone mineral density in
growing male rats. Endocrinology 138:2301^2307
260 DISCUSSION

DISCUSSION

Veldhuis: Even allowing for this marvellously interesting and prominent e¡ect
of oestrogen, the regressions don’t account for 100% of the variability in bone
mass. Is there room for non-genetic ageing-related factors independently of the
endocrine axes on bone? Would bone cells age in a perfectly normal milieu?
Riggs: Obviously, there is. It is interesting that sex steroid de¢ciency appears to
be much more important than I thought some ¢ve to 10 years ago. It does seem that
sex steroid de¢ciency is mostly responsible. Clearly, there are other non-endocrine
non-ageing factors involved. There are secondary causes of osteoporosis and
behavioural changes such as smoking and alcohol abuse that lead to bone loss.
There are genetic factors, but most of these appear to act in the development of
peak bone mass. They are much less important in the rate of bone loss.
Prior: I would like to raise a somewhat philosophical but important issue. The
concept of ‘oestrogen de¢ciency’ related to menopause is wrong. Menopause is not
a choice for women. We are born with a programmed set of ovarian signals, and it is
normal for every woman to go through menopause. If it is normal, how does that
make it ‘de¢ciency’?
Riggs: I think it is usual, but not ‘normal’. Except for advanced primates, other
animals do not have the equivalent of the menopause. But whether it is normal or
not, if it causes disease then it has to be treated, and I think it causes disease.
Prior: The model breaks down because there is good evidence now that
continually oestrogen-treated menopausal women do not preserve their bone
density (Prior et al 1997). They also lose bone density at a rapid rate when
oestrogen replacement is stopped. In addition, there is also good evidence from
many centres that the bone loss begins several years before the ¢nal menstrual
period and certainly before menopause (de¢ned as one year after the ¢nal
menstrual period) (Recker et al 2000, Okano et al 1998). During this time
statistically, oestradiol levels are not low (Prior 1998). Finally, if you are going to
talk about ‘oestrogen de¢ciency’, you also have to talk about its partner female
steroid hormone, progesterone, which is equally de¢cient and for which low
levels begin even earlier. The concept of oestrogen as a solo de¢ciency is wrong.
As a woman who has no choice about going through this process of life, it feels
prejudicial to continue to use ‘oestrogen de¢ciency’ rhetoric.
Riggs: I think you are mixing in social issues with medical issues.
Prior: Can you separate them?
Riggs: Yes, I think they can be separated. The key issue here is medical. Does this
de¢ciency state cause osteoporosis, and can this be prevented? The evidence is
overwhelming that this is so. Oestrogen replacement at the menopause is
certainly not going to prevent 100% of osteoporotic fractures, but the evidence
is strong that it will prevent the majority of them. You can then make the case that
ENDOCRINE CAUSES OF AGE-RELATED BONE LOSS 261

premenopausal loss, except possibly for the proximal femur, is related to decreases
in sex steroid production in the early pre-menopause. The major question here
between the physician and his or her patient is whether the patient will bene¢t
from treatment. Of course, there is the issue with risk and bene¢t with oestrogen
replacement therapy. If there wasn’t, it would be as easy to give hormone-
replacement therapy (HRT) as it would to give thyroid hormone replacement.
No one would dispute that.
Prior: We have a problem here of a concept that has become a paradigm. There is
only one small, one-year randomized placebo-controlled trial that sort of shows
that hormone therapy after menopause prevents fractures (Lufkin et al 1992).
Handelsman: Larry Riggs, I thought your paper was an elegant presentation of
the case for the importance of sex steroids, but I have a couple of issues I’d like to
raise. Dirk Vanderschueren has done a study in the Tfm (testicular feminized) rat,
which has completely non-functional androgen receptors (Vanderschueren et al
1994). On the basis of what you said, you would predict that there would be no
di¡erence between the Tfm genetic males and females. This is not the case, and
suggests that androgen e¡ects are not all due to aromatization. Although
periosteal bone growth due to androgen is an interesting alternative, it doesn’t
resolve the issue of these androgen receptors.
Riggs: Please don’t come away from here with the impression that we think that
testosterone has no e¡ect on bone. Clearly, it does. It is just di⁄cult to show.
Regarding the animal models, mice are very di¡erent from primates. Even the
ER^/ER^ double knockout mice don’t approach the degree of osteopenia
observed in women after ovariectomy.
Handelsman: I have real doubts about what is called the free hormone hypothesis.
It is a super¢cially attractive concept, but I think it is also vague and misleading. If a
‘free hormone’ is not bound to a protein, it may well be more accessible to tissues,
but it is also more accessible to metabolism. There is no real reason why such a
hormone should be more active rather than more rapidly metabolized. It is
impossible to predict a priori. This issue has never really been fully resolved.
Widely used measures of ‘free’ hormones are just manipulations of the data. In
particular, you showed nicely the SHBG levels rise: those bio testosterone and
oestrogen levels you measured are really an inverted measure of SHBG. If you
o¡ered that to the regression equations I wouldn’t be surprised if you ¢nd that
the reciprocal of SHBG is as good or better predictor than the supposedly
manipulated steroid levels.
Riggs: When we measure the total hormone we get weak correlations, but when
we measure the so called bioavailable T and E we get strong correlations. But I
agree with you: the sex hormone binding component is the main predictor.
Ruiz-Torres: The di¡erence between pathology and ageing is di⁄cult to work
out. Many years ago you published an important paper in which you gave the
262 DISCUSSION

normal values of bone density in women from age 20 to age 80+ (Riggs et al 1981).
If you put your data on a semilogarithmic scale you get a half-life of around 50
years. This means that those young women whose bone density values are clearly
below the mean, according with the half-life mentioned, will become osteoporotic
at the age (but without the in£uence) of the menopause. Moreover, during the life
of a rat, from adulthood onwards there is normally a continuous decrease of both
bone DNA and collagen, which is not reversible with oestrogens or testosterone.
Furthermore, using the model based on the chronic application of
aminoacetonitrile to rats, it is possible to detect the e¡ects of drugs, because the
osteoporosis produced is reversible when stopping this type of intoxication.
Oestrogen does not in£uence the recovery: the collagen or cell content of the
bones remain similar to the controls, as opposed to androgen treatment which
clearly improves the bone collagen de¢ciency. Finally, another point which
underlines the doubts of the e⁄ciency of oestrogen treatments against
osteoporosis, is the fact that osteoblasts
the cells producing bone collagen

don’t have oestrogen receptors but those for IGF1. In summary, the reduction of
the main contents of the bone is a normal manifestation of ageing and is therefore
not reversible. I consequently think that the best procedure against osteoporosis
which becomes manifest around the age of the menopause is to ensure that females
end adolescence reaching optimal bone values.
Riggs: I agree that the e¡ect of oestrogen on stimulating bone formation is
the weakest part of the hypothesis. This is why I think the paper recently
published by Khastgir et al (2001) is so interesting. Nonetheless, I agree that
oestrogen is weaker than testosterone with respect to stimulating collagen
synthesis. I wouldn’t agree that oesteoblasts don’t have oestrogen receptors. We
reported this back in 1988 (Eriksen et al 1988). They clearly have ERs and respond
to oestrogen in a number of ways. With regard to the rats, to make a rat or mouse
responsive to oestrogen, you need to ovariectomize them. If you do this, there is a
response.
Mˇller: You have clearly shown that oestrogen de¢ciency also plays a role in
male osteoporosis. Thus, it would be the rationale for administering oestrogens
to males also, but this is not feasible. Are there any trials involving SERM
(speci¢c oestrogen receptor modulator) administration, e.g. raloxifene, to males?
Riggs: To my knowledge the only study that has been done so far is the one that I
referred to. We identi¢ed a threshold level below which the markers decreased, and
above which they increased, so the net e¡ect was of no di¡erence. This is probably
additional evidence that this is the threshold between oestrogen de¢ciency and
su⁄ciency. Raloxifene is a much weaker bone agonist than oestrogen, so you are
displacing a stronger agonist with a weaker agonist and you will make an
oestrogen-de¢cient male worse. We were hoping for a better result, and there are
third generation SERMs in the pipeline that appear to have the same potency as
ENDOCRINE CAUSES OF AGE-RELATED BONE LOSS 263

oestradiol. This may produce better results in elderly men who are partially
oestrogen de¢cient.
Carroll: You mentioned the cortisol e¡ect. Actually it is trough cortisol that
correlates with rate of bone loss (Dennison et al 1999). This gets back to the
point I was making on Tuesday. What are the steps in your model where you
would see cortisol interacting with those other variables?
Riggs: We haven’t really worked this out. We are in a process of actually
measuring free cortisol excretion in our on-going studies. My guess is that at
least there will be additional factors that will be additive to the oestrogen
de¢ciency. Whether or not they interact with oestrogen de¢ciency in some
synergistic way is unclear.
Handelsman: With regard to the issue of predictions from your concept, if you
use a non-aromatizable androgen in men, it should cause bone loss. We have done a
placebo-controlled three month study with DHT, which suppresses endogenous
testosterone quite profoundly. We saw no e¡ect on bone alkaline phosphatase or
osteocalcin.
Wang: It suppresses oestradiol too.
Handelsman: We can ¢nd evidence in a more compelling model. If you take a
post-menopausal women who is oestrogen de¢cient and then use a non-
aromatizable androgen, such as nandrolone, this has profound e¡ects on bone.
These must be androgen receptor mediated.
Riggs: The osteoblasts and osteoclasts both have androgen receptors. However,
we believe that oestrogen de¢ciency is the dominant cause of bone loss in ageing
men, but that testosterone de¢ciency contributes also.
Laron: You mentioned the sex hormones. We know more these days about the
gastrointestinal hormones which certainly regulate Ca2+ intake or response to
hormones. What is the state of the art now?
Riggs: 1,25-dihydroxyvitamin D plays a key role, and this is regulatable by
oestrogen. Thus part of the e¡ect of oestrogen is through this mechanism. As for
the other gut hormones, I am not clear whether these have major e¡ects.
Haus: As pathologist, I am interested in the hyperparathyroidism you
mentioned. Is this a transient phenomenon at the time of the accelerated phase?
The reason I am asking is that morphologically the functional parathyroid
parenchyma undergoes a very marked atrophy with advancing age. At the age
when osteoporosis is most prevalent the parathyroid has lost a large percentage
of its epithelial cells which in this age group have been replaced by adipose tissue.
Riggs: There is a suppression at the beginning and then an age-related increase
that continues inde¢nitely. One autopsy study from Sweden showed an increase in
the size in the parathyroid glands with ageing. The level of the hypertrophy of the
parathyroid glands is certainly not to the extent that you see with secondary
hyperparathyroidism in chronic renal failure.
264 DISCUSSION

Haus: The gland may be of similar size, but it is composed largely of fat in older
people. Quite often there are just small sheets and clusters of active parathyroid
tissue left.
Ruiz-Torres: I have two explanations for why PTH increases with age. The ¢rst
is that ageing decreases the catabolic rate of hormones, as is well known in the case
of thyroxin. The second is related to the regulatory mechanism that would appear
when the mineral content of the bone decreases with age.
Riggs: We have looked at the secretory dynamics of the parathyroid gland, and
there is clearly an increase in secretion. It is compensatory. If you give enough Ca2+,
you can bring the PTH back to normal.

References
Dennison E, Hindmarsh P, Fall C et al 1999 Pro¢les of endogenous circulating cortisol and bone
mineral density in healthy elderly men. J Clin Endocrinol Metab 84:3058^3063
Eriksen EF, Colvard DS, Berg NJ et al 1988 Evidence of estrogen receptors in normal human
osteoblast-like cells. Science 241:84^86
Khastgir G, Studd J, Holland N, Alaghband-Zadeh J, Fox S, Chow J 2001 Anabolic e¡ect of
estrogen replacement on bone in postmenopausal women with osteoporosis: histo-
morphometric evidence in a longitudinal study. J Clin Endocrin Metab 86:289^295
Lufkin EG, Wahner HW, O’Fallon WM et al 1992 Treatment of postmenopausal osteoporosis
with transdermal estrogen. Ann Int Med 117:1^9
Okano H, Mizunuma H, Soda M et al 1998 The long-term e¡ect of menopause on
postmenopausal bone loss in Japanese women: results from a prospective study. J Bone
Miner Res 13:303^309
Prior JC 1998 Perimenopause: the complex endocrinology of the menopausal transition. Endocr
Rev 19:397^428
Prior JC, Vigna YM, Wark JD et al 1997 Premenopausal ovariectomy-related bone loss: a
randomized, double-blind one year trial of conjugated estrogen or medroxyprogesterone
acetate. J Bone Miner Res 12:1851^1863
Recker R, Lappe J, Davies K, Heaney R 2000 Characterization of perimenopausal bone loss: a
prospective study. J Bone Miner Res 15:1965^1973
Riggs BL, Wahner HW, Dunn WL, Mazess RB, O¡ord KP, Melton LJ 1981 Di¡erential
changes in bone mineral density of the appendicular and axial skeleton with aging:
relationship to spinal osteoporosis. J Clin Invest 67:328^335
Vanderschueren D, Van Herck E, Geusens P et al 1994 Androgen resistance and de¢ciency have
di¡erent e¡ects on the growing skeleton of the rat. Calcif Tissue Int 55:198^203
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Ageing and water homeostasis

David Robertson, Jens Jordan*, Giris Jacob{, Terry Ketch, John R. Shannon and
Italo Biaggioni

Autonomic Dysfunction Center, Departments of Medicine, Pharmacology, and Neurology

Vanderbilt University, Nashville TN 37232^2195, USA, *Clinical Research Center,
Franz Volhard Institut, Berlin, Germany and {Recanati Autonomic Dysfunction Center,
Rambam Medical Center, Haifa, Israel

Abstract. This review outlines current knowledge concerning £uid intake and volume
homeostasis in ageing. The physiology of vasopressin is summarized. Studies have been
carried out to determine orthostatic changes in plasma volume and to assess the e¡ect of
water ingestion in normal subjects, elderly subjects, and patients with dysautonomias.
About 14% of plasma volume shifts out of the vasculature within 30 minutes of upright
posture. Oral ingestion of water raises blood pressure in individuals with impaired
autonomic re£exes and is an important source of noise in blood pressure trials in the
elderly. On the average, oral ingestion of 16 ounces (473 ml) of water raises blood
pressure 11 mmHg in elderly normal subjects. In patients with autonomic impairment,
such as multiple system atrophy, strikingly exaggerated pressor e¡ects of water have
been seen with blood pressure elevations greater than 75 mmHg not at all uncommon.
Ingestion of water is a major determinant of blood pressure in the elderly population.
Volume homeostasis is importantly a¡ected by posture and large changes in plasma
volume may occur within 30 minutes when upright posture is assumed.

2002 Endocrine facets of ageing. Wiley, Chichester (Novartis Foundation Symposium 242)
p 265^278

Water is the major constituent of living beings. It engenders and facilitates

homeostasis by virtue of its e¡ects as a solvent, its ionizing potential, its high
thermal conductivity, its high speci¢c heat content, and its high latent heat of
evaporation. Its plasma proteins (via oncotic pressure) and its electrolytes (via
osmotic pressure) maintain body £uid homeostasis. (Greenleaf & Morimoto
1996.)
Total body water accounts for 50^70% of body weight (Fig. 1). The cell
membranes demarcate an intracellular compartment comprising about 50% of
body weight and the extracellular compartment comprising about 20% of body
weight. Extracellular £uid is further subdivided by the capillary endothelium
into interstitial £uid (*15%) and plasma volume (*5%).
The crucial cation in the extracellular £uid is Na+ (*150 mEq/l) and its
associated anions are Cl7 (*110 mEq/l) and bicarbonate (*28 mEq/l). The
265
266 ROBERTSON ET AL

FIG. 1. The distribution of body water.

osmotic environment of the extracellular £uid is largely de¢ned by these ions, and
its osmolality is usually about double the Na+ concentration.
Thirst is a subjective quality de¢ned as ‘a desire to drink potable liquids’. Thirst
appears to have several di¡erent causes, and its fundamental mechanisms remain
uncertain (Gordon et al 1997). Cellular, extracellular and volume factors all seem to
be involved. Osmotic stimuli account for about 70% and volume factors for about
25% of dehydration-induced drinking, but under circumstances where volume loss
is large, for example with haemorrhage or phlebotomy, the contribution of volume
factors increases considerably.
Greenleaf et al (1966) developed a linear regression equation for predicting
actual water intake in 87 young military trainees in ¢eld conditions which re£ects
(r = 0.79; P 5 0.01) the complexity of this functional expression of thirst:

Water intake (ml/day) = 711502+45.8 (serum osmolality)+1.2 (mean daily

urine output)718.9 (mean daily urinary potassium)+4.4 (mean daily urinary
chloride)718.7 (lying heart rate) + 1.8 (daily sweat rate).

Ageing has a signi¢cant e¡ect on body water compartments. With increasing age
beyond 30 years, there is a gradual fall in the fraction of body weight that consists of
water. This is accompanied by a less dramatic increase in the fraction of water in the
fat-free body. These changes are primarily due to loss of skeletal muscle tissue, a
relatively water-rich bodily component, with ageing.
WATER HOMEOSTASIS IN AGEING 267

FIG. 2. On assumption of upright posture after 60 minutes of supine posture, there is a rapid
loss of plasma volume from the vasculature into the interstitial tissue. This is re£ected in the
increased concentration of total protein. In this ¢gure, the per cent change in plasma volume
in the lower part of the ¢gure is calculated on change in haematocrit.

Postural plasma volume shift

Although only about 5% of the body weight is intravascular £uid, this is obviously
a very important 5% in terms of haemodynamic variables. The £uid component of
the blood is in equilibrium with the interstitial £uid. With changes in posture, the
amount of £uid in the vasculature versus the interstitial space may be altered by
gravity-induced e¡ects. The most important implication of this relates to changes
from supine to upright posture. Under normal circumstances, there is pooling of
blood in the vasculature of the lower part of the body within a few seconds of
assumption of upright posture. Normally, about 700 ml of blood moves to the
lower abdomen, pelvis, and lower extremities on standing. Compensatory
neurohumoral adjustments rapidly occur leading to increases in sympathetic
nerve tra⁄c, vasopressin, and the activation of the renin^angiotensin^
aldosterone axis.
Much less appreciated but nevertheless very signi¢cant is the loss following the
assumption of upright posture of 12^18% of plasma volume into the interstitial
£uid compartment over the succeeding 20^30 minutes (Fig. 2) (Jacob et al 1996,
1998). This £uid shift translates into loss of about a unit of blood within 30 minutes
of standing. This occurs in addition to the well-known intravascular pooling e¡ect
in the venous capacitance bed. The concurrence of these two e¡ects render the
period from 10 minutes after standing particularly vulnerable for maintenance of
adequate blood £ow to the brain and other important organs. This is when most
individuals who have a tendency to experience syncope due to low blood pressure
will have this become manifest.
268 ROBERTSON ET AL

FIG. 3. In this ¢gure, the enormous interindividual variability in per cent change in plasma
volume (upper register) and absolute value of volume change (lower register) is seen. The
volume loss in this population ranged from 240 to 770 ml within a period of 30 minutes.

Perhaps more important than the average shift in plasma volume from the
intravascular to the interstitial space is the substantial interindividual variability,
which in a recent study of 25 individuals ranged from a change in plasma volume of
7% to a change of 28% (Fig. 3). The latter resulted in a shift in haematocrit from
42% to 34% within 30 minutes of lying down, even though no bleeding had
actually occurred.
An important future direction for research is the understanding of
interindividual variation in plasma volume shift with changes in posture and
how pathophysiological conditions and drugs may alter this function. The
determinants of and the e¡ect of ageing on this shift, remain unknown.
WATER HOMEOSTASIS IN AGEING 269

Vasopressin pathophysiology

Vasopressin (AVP, ADH) is the principal hormone responsible for the body’s
narrow control of plasma osmolality (Baylis 2001). Vasopressin is synthesized in
the supraoptic (SON) and paraventricular (PVN) nuclei in the hypothalamus. The
synthesized hormone is released from the neurohypophysis (posterior pituitary
gland). It is a nonapeptide derived from a 155 amino acid precursor encoded in
chromosome 20 about 11 kilobases from the gene for oxytocin. Vasopressin is
excreted in approximately equimolar amounts with its hypophysin and then
circulates in the bloodstream with a half-life of 5^15 minutes.
Receptors for vasopressin have been identi¢ed (Baylis 2001, Preisser et al 2000).
A vasopressin 1A receptor utilizes phospholipase C/G protein, inositol phosphates
and diacylglycerol as intracellular messengers for vasopressin functions in smooth
muscle, platelets, liver and some sites in the central nervous system (CNS). A
vasopressin 1B receptor utilizes similar intracellular mechanisms and is primarily
located in the pituitary corticotroph. A vasopressin 2 receptor utilizes adenylate
cyclase and Gs protein through cAMP and protein kinase A to activate aquaporin
2 channel insertion in renal tubular cells.
Vasopressin release is powerfully stimulated by high osmolality, reduced blood
pressure (stretch), by nausea/vomiting, and by a variety of visceral traction stimuli,
but the correlation between plasma vasopressin and osmolality of blood is tightly
and directly linked over the range 285^305 mOsmol/kg (Helderman et al 1978,
Moses et al 1976). When osmolality is maintained, the e¡ect of changes in blood
pressure on plasma arginine vasopressin is also quite tight over the range 0^40 mm
Hg fall in blood pressure.
A number of studies have been undertaken to explore abnormalities in
vasopressin in ageing and in Alzheimer’s disease (Robertson & Rose 1980,
Hoogendijk et al 1985, Faull et al 1993, Frolkis et al 1999, Liu et al 2000). Not all
these studies are in agreement, but the following changes seem to be reasonably
well established. In ageing, the basal arginine vasopressin level in blood is
normal or increased, the plasma arginine vasopressin level following stimulation
is increased, the cerebrospinal £uid (CSF) arginine vasopressin level is normal and
the renal response to arginine vasopressin is diminished.
In Alzheimer’s disease, basal arginine vasopressin is normal or decreased,
stimulated arginine vasopressin is decreased and CSF levels of arginine
vasopressin are decreased. There are also changes with ageing in other hormones
of relevance to volume homeostasis (Lesser et al 1963, Shoeller 1989). Plasma
noradrenaline tends to increase with ageing, particularly in men, whereas there is
a lesser change in plasma adrenaline changes. Plasma renin activity and plasma
aldosterone decrease with age and plasma atrial natriuretic hormone is increased
in ageing.
270 ROBERTSON ET AL

End-organ changes also occur. Anatomic changes in the ageing kidney include
reduced size of the kidney, fewer glomeruli, a reduced tubular mass, and sclerosis
of pre- and post-glomerular arterioles. At the functional level, there is decreased
glomerular ¢ltration rate, decreased renal blood £ow, and a reduction in the
maximum urine concentration capability and the maximum urine dilution
capability. There are also sluggish responses to Na+ deprivation and to an acid
load.
It is not uncommon for elderly subjects to have changes in £uid homeostasis
related to low £uid intake (Miller 1995, Roth et al 2001). There are many reasons
for this lower £uid intake, including limited access to £uids due to immobility,
visual problems, or in some cases, restraints; £uid restriction (therapeutic,
procedural, or preventive); altered mental status (CNS disease, infections, drugs);
gastrointestinal disorders (swallowing problems, obstruction, drugs); and altered
thirst mechanisms (impaired thirst, CNS disease, drugs). These situations may
account for much of the hypernatraemia in the elderly. In one study, febrile
illness was present in 70% of elderly subjects presenting with hypernatraemia,
in¢rmity in 40%, surgery in 21%, nutritional supplementation in 20%,
intravenous solutions in 18%, diabetes mellitus in 15%, diarrhoea in 11%,
gastrointestinal bleed in 9% and diuretic use in 9%. Only 7% of such patients
actually had diabetes insipidus as the cause of the hypernatraemia.

The age-dependent pressor e¡ect of water

While individual patients with orthostatic hypotension due to autonomic failure
had described improvement in upright blood pressure and functional capacity
following ingestion of water, there was limited support for this concept based on
current understanding of human physiology. We examined the e¡ect of oral
ingestion of water in patients with two forms of autonomic failure, pure
autonomic failure (PAF) and multiple system atrophy (MSA) (Jordan et al 1999,
2000a,b). PAF and MSA patients often have severe autonomic dysfunction, but the
site of pathophysiology is distinct. In PAF, there is loss of peripheral autonomic
neuronal ¢bres, so that the intact CNS control mechanisms lack the peripheral
‘wiring’ needed to e¡ect changes in autonomic out£ow. In contrast, in patients
with MSA, the autonomic failure is not so much characterized by destruction of
peripheral sympathetic and parasympathetic nerves, but rather by a failure of CNS
mechanisms to appropriately engage the largely intact peripheral autonomic
innervation.
The testimony of patients about water had been surprising. Traditional
physiology did not provide a rationale for a pressor e¡ect of water drinking in
individuals who were not dehydrated. Indeed most attention of blood pressure
investigators has been aimed at examining and understanding the long-term
WATER HOMEOSTASIS IN AGEING 271

FIG. 4. In a population of patients with autonomic impairment, the ingestion of water (16
ounces; 473 ml) elicited an increase in systolic blood pressure of about 40 mmHg within 20
minutes. The diastolic pressure increment was about half that. Note that most of the e¡ects of
the water ingestion are dissipated within 90 minutes.

e¡ects of dietary Na+ on arterial pressure, and even this remained controversial. It
was therefore unexpected when under controlled conditions we did indeed observe
a substantial e¡ect of water on blood pressure in autonomic failure patients.
Water was given to PAF and MSA patients in a dosage of 16 ounces (473 ml),
ingested fairly rapidly over a period of 2^3 minutes (Fig. 4). In response to this,
blood pressure became detectably higher within 10 minutes and climbed to a
maximal pressure increase of about 40^50 mmHg at approximately 25 minutes
following ingestion (Robertson et al 2001). After that pressor e¡ect, a gradual
decline occurred that brought blood pressure back to baseline over the next 45
minutes. While some patients had responses smaller than 30 mmHg, in others
blood pressure rose more than 90 mmHg. The magnitude of this response was
astonishing. Indeed, it was greater than that observed with commonly used doses
of any pressor drug in autonomic failure. The e¡ect was most dramatic in the
supine posture but was also present in the seated and upright postures and
translated into a signi¢cant increase also in the standing blood pressure and
functional capacity.
Autonomic failure often increased the amplitude of e¡ects that were also present
albeit less prominently in normal subjects. Thus from a modelling standpoint,
autonomic dysfunction frequently made it possible to detect pressor and
depressor stimuli in small numbers of subjects that would be very di⁄cult to
detect in larger numbers of normal subjects. After observing such a dramatic
response in MSA patients, it was important to determine if these e¡ects could be
272 ROBERTSON ET AL

FIG. 5. E¡ect of oral water (480 ml) on blood pressure and heart rate in healthy older controls.
A mean increment of 10 mmHg in systolic blood pressure is noted.

replicated in normal individuals. Using a similar protocol in healthy young

subjects the pressor e¡ect of water was not observed, although a signi¢cant
increase in plasma noradrenaline was seen. In contrast, water administration in
older normal subjects elicited a signi¢cant and reproducible mean increase in
blood pressure of 11 mmHg (Fig. 5). In the older patients this pressor e¡ect was
accompanied by a trend toward reduced heart rate, but this did not reach statistical
signi¢cance.
To address the mechanism of this unexpected e¡ect of water, trimetaphan was
administered to determine if the e¡ect was dependent on the integrity of autonomic
function. In the presence of trimetaphan, no e¡ect of water on blood pressure was
observed. This suggests that the pressor e¡ect is dependent on a degree of integrity
of the autonomic nervous system for expression (Jordan et al 2000b, Rossi et al
1998). In this regard it is noteworthy that even in severe disorders of autonomic
dysfunction, some level of autonomic function remains characteristic, and, in the
absence of the bu¡ering capacity of a fully functioning barore£ex, autonomic
pressor and depressor re£exes can be surprisingly strong.
To determine whether oral administration was crucial, we gave a comparable
volume of 5% dextrose solution intravenously. This agent was selected because
the hypotonicity of water itself would be harmful if administered parenterally,
and use of physiological saline would add the confounding variable of Na+ to the
study. The 5% dextrose did not replicate the pressor e¡ect of oral water.
WATER HOMEOSTASIS IN AGEING 273

The temperature of oral water might be related to its pressor e¡ect. To test this
possibility, administration of cold (4 8C) and warm (37 8C) were compared. They
were e¡ective to the same degree as water at room temperature. These data suggest
that sympathetic activation is somehow elicited by oral water but that this is most
dramatically manifest with impairment in central autonomic control as in the case
of MSA.
It is remarkable that this e¡ect of water had been missed by physicians for so
many years, but we were probably so in£uenced by our physiology training that
we could not see this striking anomaly. However, once recognized, the e¡ect of
water has proven to be a signi¢cant therapeutic advance. Its value vis-a' -vis other
drugs lies in the potency of its action, the lack of major side e¡ects, the rapidity of its
onset and, since supine hypertension is often a limiting factor in therapy of patients
with MSA, the relatively fast return of blood pressure to normal. The depressor
e¡ect of food, especially carbohydrate, has long been recognized in MSA. Now
with recognition of the pressor e¡ect of water, a simple and patient-controlled
approach to blood pressure management is possible. Our experience so far
suggests that many patients need no other therapy for control of blood pressure
than judicious dosing of food and water. Careful further studies will be required in
order to elucidate the precise mechanism underlying the e¡ect of water, but an
action through gastrointestinal stretch receptors or osmoreceptors must be
addressed in future investigations. There remains the possibility that subtle
blood volume increases in the hour after water ingestion, and before it is fully
compensated by e¡ects of the vasopressin/renal system, might contribute to this
e¡ect.
For the older normal subjects, the importance of the pressor e¡ect of water is the
noise it adds to the monitoring of blood pressure. Until now recent water intake
has not been a variable controlled in studies of blood pressure and antihypertensive
agents in the older age group. It would now appear that water ingestion may
represent a major component of the noise inherent in blood pressure monitoring
in this population. Recognition of this e¡ect and control for it may reduce the
number of subjects needed in future studies of antihypertensive drugs.

Conclusions
Fluid handling in ageing is altered in many ways. There is a reduced fraction of the
body consisting of water in ageing. Postural plasma volume shifts occur although
the precise role of ageing in interindividual variation in this variable remains
unstudied. Oral water raises blood pressure in individuals with impaired
autonomic re£exes and is an important source of noise in blood pressure trials in
the elderly. On the average, oral ingestion of 16 ounces (473 ml) of water will raise
blood pressure 11 mmHg in elderly normal subjects. In patients with autonomic
274 ROBERTSON ET AL

impairment such as multiple system atrophy, strikingly exaggerated pressor e¡ects

of water have been seen with blood pressure elevations greater than 75 mmHg not
at all uncommonly. These responses tend to be maximal within 30 minutes of water
ingestion and to be largely dissipated within 90 minutes.

References
Baylis PH 2001 Vasopressin, diabetes insipidus and syndrome of inappropriate antidiuresis. In:
DeGroot LJ, Jameson JL (eds) Endocrinology. Saunders, Philadelphia, PA p 363^376
Faull CM, Holmes C, Baylis PH 1993 Water balance in elderly people: is there a de¢ciency of
vasopressin? Age Ageing 22:114^120
Frolkis VV, Kritnitskaya-Ryzhora TY, Dubiley TA 1999 Vasopressin, hypothalame-
neurohypophyseal system and ageing. Arch Gerontal Geriatr 29:193^214
Greenleaf JE, Averkin EG, Sargent F 1966 Water consumption by man in a warm environment:
A Statistical Analysis. J Appl Physiol 54:414^419
Greenleaf JE, Morimoto T 1996 Mechanisms controlling £uid ingestion: Thirst and drinking.
In: Buskirk ER, Puhl SM (eds) Body £uid balance: exercise and sport. CRC Press, Boca Raton,
FL, p 3^16
Gordon MS, Majzoub JA, Williams GH, Gordon MB 1997 Sodium balance modulates thirst in
normal man. Endocr Res 23:377^392
Helderman JH, Vestal RE, Rowe JW, Tobin JD, Andres R, Robertson GL 1978 The response
of arginine vasopressin to intravenous ethanol and hypertonic saline in man: the impact of
aging. J Gerontol 33:39^47
Hoogendijk JE, Fliers E, Swaab DF, Verwer RW 1985 Activation of vasopressin neurons in the
human supraoptic and paraventiruclar nucleus in senescence and senile dementia. J Neurol Sci
69:291^299
Jacob G, Ertl AC, Shannon JR et al 1996 Idiopathic orthostatic tachycardia; the role of dynamic
orthostatic hypovolemia and norepinephrine. Circulation 94:1^627
Jacob G, Ertl AC, Shannon JR, Robertson RM, Robertson D 1998 E¡ect of standing on
neurohumoral responses and plasma volume in healthy subjects. J Appl Physiol 84:914^921
Jordan J, Shannon JR, Grogan E, Biaggioni I, Robertson D 1999 A potent pressor re£ex elicited
by drinking water. Lancet 353:723
Jordan J, Tank J, Diedrich A, Robertson D, Shannon JR 2000a Vasopressin and blood pressure
in humans. Hypertension 36: E3^E4
Jordan J, Shannon JR, Black BK et al 2000b The pressor response to water drinking in humans:
a sympathetic re£ex? Circulation 101:504^509
Lesser GT, Kumar I, Steel JM 1963 Changes in body composition with age. Ann NY Acad Sci
110:576^588
Liu RY, Zhou JN, Hoogendijk WJ et al 2000 Decreased vasopressin gene expression in the
biological clock of Alzheimer disease patients with and without depression. J Neuropathol
Exp Neurol 59:314^322
Miller M 1995 Hormonal aspects of £uid and sodium balance in the elderly. Endocrinol Metab
Clin North Am 24:233^253
Moses AM, Miller M, Streeten DH 1976 Pathophysiologic and pharmacologic alterations in the
release and action of ADH. Metabolism 25:697^721
Preisser L, Teillet L, Aliotti S et al 2000 Downregulation of aquaporin-2 and -3 in aging kidney is
independent of V(2) vasopressin receptor. Am J Physiol Renal Physiol 279:F144^152
Robertson D, Shannon JR, Jordan J et al 2001 Multiple system atrophy: New developments in
pathophysiology and therapy. Parkinsonism Relat Disord 7:257^260
WATER HOMEOSTASIS IN AGEING 275

Robertson GL, Rose JW 1980 The e¡ect of aging on neurophypophysial function. Peptides
1:159
Rossi P, Andriesse GI, Oey PL, Wienke GH, Roelofs JM, Akkermans JM 1998 Stomach
distension increases e¡erent muscle sympathetic nerve activity and blood pressure in healthy
humans. J Neurol Sci 161:148^155
Roth J, Koch CA, Rother KI 2001 Aging, endocrinology and the elderly patient. In: DeGroot
LJ, Jameson JL (eds) Endocrinology. Saunders, Philadelphia, PA p 529^555
Schoeller DA 1989 Changes in total body water with age. Am J Clin Nutr 50:1176^1181

DISCUSSION
Handelsman: This is very interesting. A comment you made in passing about
pharyngeal receptors prompted me to think of some experiments. It would be
interesting to see in humans given £uid via a nasogastric tube whether you
would see this phenomenon or not
in other words, whether you could localize
the e¡ect to the pharynx. The second issue is whether chemical sympathectomy via
6-hydroxydopamine would have e¡ects like this in animals. It would also be
interesting to deliver the water into the stomach, bypassing the pharynx, or
actually allow the subject to drink the water, and then suck it out of the stomach
as quickly as it is drunk.
Robertson: Studies on rats have been done. I believe they were not
sympathectomized. There were some observations of increases in blood pressure
with water ingestion. It was more prominent with water than with £uid with Na+
in it. The question about pharyngeal receptors is very important, and I don’t have
an answer for that. It would be a wonderful study to do. Our current studies are
focused on blowing up a balloon in the stomach to observe the e¡ect of stretch. As
you know, there are osmoreceptors in the liver also that might detect small changes
in the diluteness of blood£ow from the liver. I don’t know any other purpose for
those osmoreceptors.
Veldhuis: It would have to involve more speci¢c receptors, because just passage
of food may not be relevant. I was thinking of the simple sugar experiment. It has
been a clinical adage that bad hiccups can be sometimes suppressed by swallowing
some granulated sugar irritating the glossopharangyeal nerve.
Laron: Your data on the orthostatic shift are extremely interesting. Elderly
people drink less, so they are relatively dry. They are more prone to have
arteriovascular incidents. If you take out so much £uid they could have more
plugging of small arteries. What is the practical application of this? We know
that if people don’t move, they get more osteoporosis. How do you meet the
needs of the population we are discussing here?
Robertson: That is a good point. There are data indicating there are more heart
attacks in the early hours of the morning, soon after people have woken up. The
main reason cited for this is stress, although other factors may be involved. But the
276 DISCUSSION

constituents of the blood are concentrated by standing. Although 12^15% increase

in concentration might not seem much, it is certainly possible that a 12% increase in
the platelet count might make blood clots in the heart more likely.
Laron: Would you say that elderly people should drink more before standing up?
Robertson: We view water as potentially a very dangerous drug! Seriously,
though, we do ask people who have these rare autonomic disorders not to
drink water after about 9 p.m. unless they are really thirsty. In some of these
patients with dysautonomia, water can raise blood pressure by 90 mmHg, and
if this occurs before they go to bed, the resulting hypertension could be
deleterious.
Burger: Is the di¡erence between oral versus intravenous £uid load in any way
explicable by the speed at which the plasma compartment is expanded by the two
routes?
Robertson: I wish I had a better way to do that experiment. In addition to the
problem you have identi¢ed there is another: dextrose has its own tendency to
lower blood pressure by calling out some of the gastrointestinal hormones that
may themselves be vasodilatory. On the other hand, if we gave saline, this
wouldn’t be a fair test of water either. We tried to spread the saline out over an
hour. We wish we could have just looked at volume.
Carroll: Could you do an isotope dilution study and look at the rate of entry of
that water volume into the circulation?
Roberston: We need to do that. Although the haematocrit didn’t change, the £uid
could be distributed both intracellularly and extracellularly.
Carroll: It is probably not operating through blood volume changes, but rather
through some sort of re£ex.
Elahi: We have done some studies where we have infused hypertonic saline in
young and old volunteers. There are great changes in atrial natriuretic peptide
(ANP). It is delayed in the elderly. We did the opposite experiments in which we
used half normal saline. There is very little change in the elderly. Do you have any
information about brain ANP with respect to age?
Robertson: No, I don’t.
Haus: In the elderly, there is a shift in urine excretion from daytime to night. This
is not a function of the enlarging prostate because it occurs just as much in women
as in men (Haus et al 1988). Do you have any explanation for this?
Robertson: It is in part circadian, because it isn’t just supine posture that is
responsible.
Haus: Other solutes in the urine such as adrenaline and noradrenaline do not
shift. Their circadian rhythm remains unchanged in its timing and is dissociated
from that in urinary volume excretion (Haus et al 1988).
Robertson: I don’t have the answer, though there is a substantial literature about
this.
WATER HOMEOSTASIS IN AGEING 277

Haus: The blood viscosity is higher in the morning, which may be one of the
factors contributing to the morning incidence of myocardial infarction. The
number of circulating granulocytes peaks in the late afternoon. Granulocytes
and lymphocytes show high amplitude circadian rhythms. The number of
circulating platelets peaks in the afternoon, but the amplitude of their rhythm
is relatively small (Haus 1996). In contrast, platelet aggregation and adhesion
peak in the morning (Haus et al 1990). This may be responsible for some of the
cardiovascular and cerebrovascular accidents occurring at that time. Also, the
plasminogen activator inhibitor (PAI), which determines the overall activity of
the ¢brinolytic system peaks in the morning decreasing the activity of the
¢brinolytic component of the haemostatic system (Andreotti & Patti 1997).
Robertson: I don’t know whether we really have data that will address the
postural e¡ects on platelets. For example, remember these changes in £uid occur
over 10^15 minutes. It may be that acutely standing up in the morning could
signi¢cantly increase the number of platelets within 10 minutes. This may be
separate from the circadian variation.
Haus: The change in position may increase the function of platelets probably
more than their number. To£er et al (1987) studied platelet aggregation and
found a rise with the change from recumbent to upright position which
paralleled the rise in plasma catecholamines. We studied clinically healthy
subjects in the morning while still lying down and during the day after 30
minutes in recumbent position and found the rise in platelet adhesion and
aggregation (Haus et al 1990) a little earlier than To£er et al (1987).
Robertson: The postural changes could induce additional changes on top of that.
Bj˛rntorp: As you mentioned, this work has practical clinical and research
implications. Most blood sampling is done with the patient lying down, but
some is when the patient is sitting. How long does it take before you reach
equilibrium after lying down?
Robertson: I think it is 30 minutes to equilibrium in either change of posture.
When I was an intern, patients would walk into the emergency room, get
screening blood work and be admitted. Then they would become supine. In
those days we didn’t watch our testing too much, and the next day another blood
test would show a 4% fall in the haematocrit, and it would appear that the patient
had lost a unit of blood. As an intern, I was frequently pushed by my resident
physician to ¢nd out where the patient was bleeding. We were probably
observing this volume shift.
Veldhuis: We learned this in our General Clinical Research Center. If the nurses
call you for a slightly low haematocrit, have the patient walk about for 30 minutes
and repeat the blood count: it will typically be normal.
Bj˛rntorp: What about sitting?
Robertson: I don’t have data on sitting.
278 DISCUSSION

Prior: We need to ¢nd out, because this is the posture in which most samples are
obtained.
Mˇller: Would it be worthwhile testing this water role in Parkinsonian patients:
many of them have a natural orthostatic hypotension. There are data showing that
the noradrenergic innervation of the heart of these patients is decreased (Goldstein
et al 2000). In your data you show that there is a blockade of dopamine
b-hydroxylase which will involve an increase in the dopaminergic tone.
Robertson: We recorded our ¢ndings about water in the literature about 18
months ago. Then, investigators in France gave water to patients with
Parkinson’s disease and did not see any e¡ect. I haven’t yet gone back and looked
at Tennessee Parkinson’s disease patients.

References
Andreotti F, Patti G 1997 Chronobiology of the haemostatic system. In: Redfern PH, Lemmer B
(eds) Handbook of experimental pharmacology: physiology and pharmacology of biologic
rhythms. Springer-Verlag, Heidelberg, vol 125:533^555
Goldstein DS, Holmes CMT, Li S-T, Bruce S, Metman LV, Cannon RO III 2000 Cardiac
sympathetic denervation in Parkinson disease. Ann Intern Med 133:338^347
Haus E 1996 Biologic rhythms in hematology. Path Biol 44:618^630
Haus E, Nicolau GY, Lakatua D, Sackett-Lundeen L 1988 Reference values for
chronopharmacology. Annu Rev Chronopharmacol 4:333^424
Haus E, Cusulos M, Sackett-Lundeen L, Swoyer J 1990 Circadian variations in blood
coagulation parameters, alpha-antitrypsin antigen and platelet aggregation and retention in
clinically healthy subjects. Chronobiol Int 7:203^216
To£er GH, Brenzinski D, Schafer AL et al 1987 Concurrent morning increase in platelet
aggregability and the risk of myocardial infarction and sudden cardiac death. N Engl J Med
316:15141^518
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Summing-up
Johannes Veldhuis

University of Virginia Health System, NIH General Clinical Research Center, Department of
Internal Medicine, Division of Endocrinology and Metabolism, PO Box 800202,
Charlottesville, Virginia 22908-0202, USA

Abraham Lincoln once said that it is better to remain silent and appear ignorant
than to speak freely and remove all doubt! Thus, I will attempt brie£y to highlight
some of the discussions we have had over the last few days. Ageing can be viewed as
an array of sequelae, some of which we interpret as undesirable, ranging from
decreased bone mass to atrophic skin and hair greying; and from relative
sarcopenia and variable cognitive defects to increased visceral fat and heightened
risk of cardiovascular disease. This panoply is somehow directed by an ensemble of
cellular and systemic factors. The resultant complexity of the ageing process is thus
challenging. Causal endocrine and non-endocrine factors undoubtedly overlap,
even in ways beyond those we recognize. From an endocrine perspective, there is
a substantial decline of GH and sex-steroid input to target cells. A gradual
reciprocal increase in integrated cortisol receptor drive throughout age seems to
contribute to catabolic changes, particularly in a waning anabolic context of
attenuated GH/IGF1 and sex-steroid production. Concomitantly, alterations in
the insulin pathway condition cellular signalling in ageing, which strongly
in£uences epidemiological risk. Operating on this fourfold hormonal
background is the genetic endowment and any number of environmental factors.
Only the thyroidal axis seems to be spared substantially. This symposium has tried
to unravel some of the intersecting causes and consequences of these
neurohormonal changes in ageing.

279
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Index of contributors

Non-participating co-authors are indicated by asterisks. Entries in bold type indicatepapers; other
entries refer to discussion contributions.

A G
*Andres, R. 222 Giustina, A. 121, 138, 154, 156, 158, 219,
220
*Gonzalez-Cadavid, N. F. 82
B
*Groome, N. 161
*Behrends, J. 193
*Biaggioni, I. 265 H
Bj˛rntorp, P. 17, 18, 20, 38, 39, 41, 43, 46, Handelsman, D. J. 21, 38, 40, 41, 43, 62, 66,
59, 60, 61, 63, 81, 95, 118, 119, 157, 168, 78, 79, 80, 95, 96, 119, 120, 121, 139, 141,
220, 243, 277 153, 155, 157, 158, 169, 171, 187, 190,
*Bonavera, J. J. 82 202, 217, 218, 243, 244, 246, 261, 263,
275
Bowers, C. Y. 98, 216, 217, 219 Haus, E. 23, 24, 63, 141, 189, 203, 245, 263,
Brabant, G. 63, 96, 158, 193, 202, 203 264, 276, 277
Burger, H. G. 37, 38, 60, 61, 79, 161, 167, *Hikim, A. S. 82
168, 169, 170, 187, 188, 190, 276
I
C *Iranmanesh, A. 98
Carroll, B. J. 19, 20, 26, 36, 37, 38, 39, 40,
41, 42, 43, 44, 62, 154, 155, 157, 169, 188, J
201, 217, 219, 245, 263, 276 *Jacob, G. 265
*Jordan, J. 265
D
K
*Dennerstein, L. 161
*Dudley, E. 161 *Ketch, T. 265

L
E
*Lamberts, S. W. J. 3
*Egan, J. M. 222 Laron, Z. 17, 20, 24, 38, 62, 63, 79, 95, 97,
Elahi, D. 141, 154, 218, 220, 222, 242, 243, 122, 125, 138, 139, 140, 141, 155, 157,
244, 245, 246, 276 170, 188, 202, 216, 219, 243, 245, 246,
263, 275, 276
F *Leitolf, H. 193
*Leung, A. 82
*Ferrini, M. 82 *Lue, Y. -H. 82

280
INDEX OF CONTRIBUTORS 281

M S
*Mamers, P. 161 Shalet, S. M. 79, 121, 137, 140, 141, 170,
*Meneilly, G. S. 222 202, 205, 217, 218, 219, 220
Monson, J. 139, 140, 154 *Shannon, J. R. 265
Morley, J. E. 20, 21, 22, 24, 40, 41, 42, 63,
77, 78, 79, 95, 96, 122, 123, 156, 157, 158, *Swerdlo¡, R. S. 82
159, 170
*Muller, D. C. 222 V
Mˇller, E. E. 38, 42, 95, 122, 159, 188, 189,
217, 262, 278 van den Beld, A. W. 3, 17, 18, 19, 20, 21, 22,
*Mulligan, T. 98 23, 24, 155
*Van den Berghe, G. 205
P Veldhuis, J. D. 1, 16, 17, 18, 19, 20, 22, 24,
36, 37, 38, 40, 41, 44, 58, 59, 60, 61, 62,
Prior, J. C. 18, 39, 43, 44, 59, 60, 63, 97, 154, 63, 64, 77, 78, 79, 80, 81, 96, 97, 98, 119,
158, 168, 172, 187, 189, 190, 219, 244, 120, 121, 122, 123, 141, 153, 154, 155,
245, 260, 261, 278 157, 158, 159, 167, 168, 186, 187, 188,
189, 191, 202, 216, 217, 218, 219, 222,
R 242, 243, 244, 245, 260, 275, 277, 279
Riggs, B. L. 21, 41, 80, 140, 157, 169, 189, *Vernet, D. 82
247, 260, 261, 262, 263, 264
Robertson, D. 42, 43, 120, 161, 190, 265, W
275, 276, 277, 278
Ruiz-Torres, A. 80, 81, 140, 143, 155, 156, Wang, C. 22, 79, 82, 95, 96, 97, 122, 159,
158, 190, 203, 218, 261, 264 263
 Endocrine Facets of Ageing.
Novartis 242
Copyright & 2002 John Wiley & Sons Ltd
Print ISBN 0-471-48636-1 Online ISBN 0-470-84654-2

Subject index

A receptor expression 81
abdominal fat 51, 54, 56 therapy 70, 72^73, 75
and depression 33 andropause 5, 6^8, 12^13, 85, 97
acid labile subunit (ALS) 206 angina, testosterone treatment 158
acromegaly 133, 138 anorexia in ageing 157, 159, 257
actin ¢laments 147 anovulatory cycles 63, 174
activin 162 antidepressants 30, 43, 189
adenylate cyclase 269 antihypertensives 68
adipose tissue, glucose uptake 229 antimode 226
adrenal cortex, zona reticularis cells 5 aphrodisiacs 67
adrenalectomy 38 aquaporin 269
adrenaline 269 L-arginine 103, 106
adrenocorticotropic hormone (ACTH) 64, aromatase 56, 74, 121^122, 162, 168, 218
154 associative memory 28
atherogenesis 74, 139, 149^151, 153, 155,
critical illness 211^212
156
adrenopause 5, 8^10, 12^13
atherosclerosis 155, 156, 158
alcohol 49, 190
calcium ion antagonist therapy 148
abuse 257
aldosterone 211, 212, 269 insulin and 146
allopregnanolone 188 and oestrogen 8
allostasis 30^32, 47 ovulation disturbance 60
allostatic load 27, 30^32 athletes 231, 238, 243
Alzheimer’s disease (AD) atrial natriuretic peptide 212, 269, 276
arginine vasopressin 269 autoimmune thyroid disease 196
and education level 42 autonomic failure 270^271
neuro¢brillary tangles 28
neuronal loss 27, 28 B
oestrogen deprivation 165 Baltimore Longitudinal Study of Aging
oestrogen therapy 42, 43 (BLSA) 225
Ames dwarf mice 131, 132 basal metabolic rate (BMR) 202
g-aminobutyric acid 90, 188 basal temperature 174, 175, 179
Ammon’s horn 28 benign prostatic hyperplasia 80^81
b-amyloid 42 betaglycan 162
anastrozole 121 bicarbonate 265
androgens 70^75, 112 blood pressure, water e¡ect 270^273
de¢ciency, diagnosis and management blood sampling and posture 277^278
74^75 blood viscosity 277
designer 73^74 Bloom syndrome 134
disease risk 55^57, 73^74 body composition 4, 6
female reproductive ageing 165, body temperature 31^32
168^169 and depression 32^33, 43^44

282
SUBJECT INDEX 283

bone loss 247^264 child abuse and ovulatory disturbances

cancellous bone 248 39
cortical bone 248 chloride ions 265
cortisol 263 chronic fatigue syndrome 61
luteal phase length 189^190 circadian time organization 23
medroxyprogesterone therapy 181 clamp technique 228
non-endocrine factors 257 clomifene 122
oestrogen de¢ciency 248, 249^251, 255, Cockayne’s syndrome 134
260^261 cognitive function
testosterone therapy 7 brain ageing 27
bone mineral density GH/IGF1 axis 11
glucocorticoid exposure 51
dehydroepiandrosterone sulfate levels 9
co-gonadotropins 100^102
oestradiol levels 8, 165, 169
cold sweats 190
testosterone role 6, 7, 85, 255^256 collagen 251
brain congestive heart failure, growth hormone
ageing 27^30 therapy 220
glucose uptake 229 constant oestrus 188
stress response 27, 28, 29, 30^32 corticosteroids 257
brain-derived neurotrophic factor 29 corticosterone 49
brainstem 27 corticotropin-releasing hormone 39, 211
breast antagonists 219^220
cancer 23, 56 cortisol 5, 40, 41, 47, 52, 54, 64, 154, 155
tenderness 164, 180 bone loss 263
Brown Norway rat 31, 82^97 burns patients 217
‘burn-out’ 47, 54, 59, 61^62 central regulation 46
burns patients, cortisol levels 217 critical illness 211^212, 217
and depression 32
C saliva measures 53
tumours 51
CA3 hippocampal pyramidal cells 29
critical illness 205^221
calcium 147, 148, 248, 249, 251
cAMP 269 adrenal insu⁄ciency 212
Canadian Multicentre Osteoporosis Study gender di¡erences 207, 208
182 growth hormone axis 206^208, 219
candidate gene approach 55, 56 hypogonadotropic hypogonadism 211
carbohydrate metabolism 222 luteinizing hormone^testosterone axis
cardiovascular disease 51, 54, 57, 73^74 210^211
dehydroepiandrosterone sulfate link 10 pituitary^adrenal axis 211^212
and depression 33 prolactin 210
erectile dysfunction 68, 85 thyrotropin^thyroid axis 209^210
cross-approximate entropy 119^120
oestradial and 165
Cushing’s syndrome 47, 51, 54
oestrogen-related protection 8
postoperative HPA axis 61
oestrogen replacement therapy 74 cytokines 90, 95, 96^97, 206, 210, 211
and testosterone 74 cytoskeleton 147
carotid artery intima/medial thickness, cytotoxins 90, 92, 93, 95
testosterone e¡ects 18
castration 74, 96, 158
D
CD5+ B cells 196
central nervous system 188 Daf2 133
dehydroepiandrosterone in£uence 10 defeat reaction 47
284 SUBJECT INDEX

dehydroepiandrosterone (DHEA) 5, 8^10, energy

188, 257 intake and thyroid function 197
bone mineral density 9 loss following ovariectomy 165
central nervous system activity 10 entorhinal cortex 27, 28
immune system 10 entropy 119^120
prostate e¡ects 10 erectile dysfunction 68, 85
therapy 9^10, 22 ethnicity 24
dehydroepiandrosterone sulfate (DHEAS) euglycaemic clamps 230, 243
5, 8^10, 61, 212, 257 Euronut^Seneca study 197
cardiovascular risk 10 European Group for the Study of Insulin
deiodinases 194, 198^199 Resistance 229^230
dementias 27, 42 evidence-based medicine 67
dentate gyrus 28, 30 excitatory amino acids 88, 90
depression 32^33, 188 exendin 4 238
see also antidepressants 30, 43, 189 exercise 154
body temperature 32^33, 43^44 and disability 20
cardiovascular risk 33
cholinergic theory 38
cortisol levels 32 F
defeat reaction 47
malnutrition and 157 fat mass and testosterone 6^7
melancholic 47 fertility 68^70
menopausal 169^170, 189 ¢ght^£ight reaction 47
mortality and 33 Fisher rats 198
non-cardiac morbidity 33 £uid intake, elderly 270
stress and 32, 36^37 follicle-stimulating hormone 70, 83, 88, 162,
testosterone production and 6 163, 165, 173
dexamethasone suppression test 38^39, 51, 53 ¢nal menses levels 164
diabetes mellitus food restriction 40^41, 133^134
diagnosis 225^227, 243^244 forearm glucose uptake 230
type I 122 ‘Fountain of Youth’ 9
type II 51, 54, 57 frailty 3, 85, 156
and depression 33 Framingham study 196
glucagon-like peptide 1 (GLP1) therapy free hormone hypothesis 71, 261
236^238 functional ability 4, 141
incretin e¡ect 234^238
diacylglycerol 269
diet restriction 40^41, 133^134 G
diethylstilbestrol 110
Gs protein 269
dihydrotestosterone 81
GABA 90, 188
diltiazem 147
ganirelix 102
dopamine 194, 198, 211
genes in ageing 52^57
intensive care use 210
microsatellites 56^57, 81
dual-energy X-ray absorptiometry (DEXA)
140 geography 24
dyslipidaemia 51, 54 germ cells 93
apoptosis 85, 93
Germany, goitre in 194, 195
E
ghrelin 110, 111, 242
Endocrine Society of Australia 75 glucagon-like peptide 1 (GLP1) 236^238
b-endorphin 64 glucocorticoid cascade hypothesis 50
SUBJECT INDEX 285

glucocorticoid excess GHRP2 110^111, 216

dendritic changes 28 testosterone e¡ects 110^112
disease links 51
glucocorticoid receptor H
abnormal occupancy 31, 32, 33
density 47, 49^50 haematocrit, testosterone therapy 7
gene 55 haemoglobin 245
glucose hair
hepatic production 228^229 greying 127, 130, 138
utilization 229^230 loss 31
glucose-dependent insulinotropic HDL cholesterol 11
polypeptide (GIP) 235^236 acute illness 219
glucose intolerance 31, 222^225 testosterone therapy 7
diagnosis 225^227 healthy ageing 21
glucose tolerance tests, see oral glucose heart attacks 155
tolerance tests early morning incidence 275, 277
glutamine 90 heart failure, growth hormone therapy
goitre 194, 195 220
gonadopause 98 helplessness 47
gonadotropin-releasing hormone (GnRH) hepatic glucose 228^229
88, 90, 92, 95, 97, 162 hiccups 275
granulocytes 277 hip fracture, growth hormone therapy 11, 22
growth hormone (GH) 10, 24, 100^102, hippocampus 28, 29, 30, 31, 49, 52
103^107, 125, 144, 154, 257 holocaust victims 47, 61
autonegative feedback 102, 106^108 hormone replacement therapy 72, see also
congenital isolated de¢ciency 126^127 speci¢c hormones
critical illness 206^208, 219 hot £ushes
HPA axis activity 47 and alcohol 190
primary resistance (Laron syndrome) 127, oestradiol levels 164
129^131, 139 social stress 190
receptor density and age 107 venlafaxine treatment 189
and sleep 188 11b-HSD1 154
testosterone synergy 99^100 Hutchinson^Gilford disease 134
therapy 11, 18, 72, 125^126, 141, 216^217, hyperglycaemic clamp 230, 231, 233,
219 234^235, 236, 237
heart failure 220 hyperinsulinaemic^euglycaemic clamp 233,
hip fractures 11, 22 236
intensive care 133, 219 hypernatraemia 270
side-e¡ects 11 hyperparathyroidism 249, 251, 256, 263
transgenic mice 133^134 hypertension 51, 54
growth hormone binding protein (GHBP), drug therapy and erectile dysfunction 68
critical illness 206 hyperthermia 31^32
growth hormone/insulin-like growth and depression 32^33, 43^44
factor 1 axis 5^6, 11, 98, 102^103, hyperthyroidism 195^196
112 hypogonadism 79, 103^107
de¢cient mice 131^134 hypothalamic^pituitary^adrenal axis
growth hormone-releasing hormone ageing and 49^52
(GHRH) 12, 102, 105^106 burn-out 47, 54, 59, 61^62
testosterone e¡ects 108 disease risks 51, 52^57
growth hormone-releasing peptides stress and 31, 46^48
(GHRPs) 12, 102, 105^106, 159 hypothalamic^pituitary dysfunction 88^90
286 SUBJECT INDEX

hypothalamo^pituitary^ovarian axis intestinal oestrogen receptors 251

162^163 intravascular pooling 267
hypothalamo^pituitary^thyroid axis iodine de¢ciency 194^195
193^204 irritability 188
hypothalamus 28^29, 30, 90, 92, 269
hypothyroidism 196, 202, 203 K
ketoconazole 119
I
in depression 43
IGF-binding protein (IGFBP) kidney
and critical illness 206 ageing 270
IGFBP1 6 oestrogen e¡ects 251
IGFBP2 6, 11, 206 Krk island 127
IGFBP3 6, 10, 206
IGFBP6 206 L
immune system
dehydroepiandrosterone e¡ects 10 L-692,429 12
depression and 33 Laron mouse 132^133
glucocorticoid exposure 51 Laron syndrome 127, 129^131, 139
imprinting 40 LDL cholesterol 146
incretin 234^238 testosterone therapy 7
inducible nitric oxide synthase 90, 92^93, 95, leptin 194
96, 187 Leydig cells 5, 83, 85, 93, 96, 123
infection susceptibility 31 stress-induced suppression 211
in£uenza and depression 33 life expectancy and gender 155
infundibular nuclus 29, 189 lifespan and body size 133^134
inhibins 88, 162, 173 lipid concentrations, testosterone therapy 7
A 162 lipoprotein (a) 155
liver, glucose production 228^229
B 70, 162, 163^164
locus ceruleus 27^28, 30, 31
inositol phosphates 269
luteal phase length 174
insulin 60^61, 145^146
bone loss 189^190
atherogenicity of 146
luteinizing hormone (LH) 5, 7, 70, 83, 88, 90,
resistance 51, 54, 56, 59
162, 163
smooth muscle cell stimulation 146^148
and critical illness 210^211
insulin-like growth factor 1 (IGF1) 10, 62^63,
therapy 86^87
97, 100^102, 125, 143, 144^145, 156, 257
lymphocytes 277
autonegative feedback 102
and cancer 139
critical illness 206 M
malnutrition and 156^157 macrophages 153
neurotrophic e¡ect 139 malnutrition
oestrogen e¡ects 251 frailty 156^157
smooth muscle cell e¡ects 146^147 thyroid function in non-thyroidal illness
see also growth hormone/insulin-like 197
growth factor 1 axis medroxyprogesterone 181
insulin-like growth factor 2 (IGF2) 257 melancholic depression 47
intensive care melatonin 189
dopamine use 210 memory 27, 28, 52
growth hormone therapy 133, 219 metabolic syndrome 54, 56, 224
interleukins 206, 211 N-methyl-D-aspartate (NMDA) 28, 88,
interstitial £uid 265 90^91
SUBJECT INDEX 287

metoclopramide 198 oestrogens 8, 55, 112

metyrapone 43, 62 Alzheimer’s disease 165
microsatellites 56^57, 81 atherosclerosis 8
mineralocorticoid receptor, type 1 37 bone loss 248, 249^251, 255, 260^261
Minimal Model 230, 231 cardiovascular protection 8
MK-667 12 therapy 42, 43, 74, 248, 250, 260^261
mobility 20 oestrone (E1) 5
molecular genetics 52^57 oral glucose tolerance tests 223, 226, 243^244
molimina 182 osmolality 266
monocytes 10 osteoporosis 31, 140, 165, 247
mood swings 180 aromatase gene 56
multiple pituitary hormone de¢ciencies 127 and depression 33
multiple system atrophy 270^271 glucocorticoid exposure 51
muscle see also bone loss
glucose uptake 229 ovaries
mass 6, 31, 85 androgen secretion 168
strength 4, 6, 7 cysts 172, 173
myelin 27 follicle numbers 163^164, 167, 172, 173
myocardial infarction 155 granulosa cells 162, 167
early morning incidence 275, 277 ovariectomy 165
polycystic 59^60, 62^63
N ovulation disturbance 174
natural killer cells 10 atherosclerosis and 60
neuro¢brillary tangles 28 child abuse 39
neurogenesis 27 clinical e¡ects 180^181
neurohypophysis 269 non-ovulation, turned-on and turned-o¡
neurokinin B neurons 189 60
neuronal nitric oxide synthase 90, 91^92, 93 perimenopause 175^180
neurons oxytocin 210
apoptosis 90, 92
loss 26, 27^28 P
synapses 27, 41^42 p120 163
NHANES III survey 223 pancreatic b-cells 231^234
night sweats 164, 189 papaverine 68
nitric oxide 158^159 parathyroid glands 263^264
nitric oxide synthase (NOS) 29, 90^93, 95, parathyroid hormone 248, 249, 251, 264
96, 159, 187 paraventricular nucleus 29, 269
nitrotyrosine 92 perforant pathway circuit 28
NMDA 28, 88, 90^91 perimenopause 173, 175^180
non-steroidal anti-in£ammatories 42^43 peroxynitrite 90, 92, 95
noradrenaline 269 pharyngeal receptors 275
normality 244^245 phospholipase C/G protein 269
O pituitary 62
disease 131, 137^138
obesity 63, 121 plasma volume 265
oesteoblasts 262 postural shift 267^268
oestradiol (E2) 5, 121^122, 162, 163, plasminogen activator inhibitor 277
167^168, 173 platelets 277
bone mineral density 8, 165, 169 function and posture 277^278
falling circulation of 164^165 polycystic ovary syndrome 59^60, 62^63
288 SUBJECT INDEX

polymorphisms 55, 56, 81 selenium 197

postpartum HPA axis £attening 61 semen analysis 69, 85
posture senescent cells 150^151
blood sampling 277^278 serine phosphorylation 60^61
plasma volume shift 267^268 Sertoli cells 85, 93
platelet function 277^278 sex-hormone binding globulin (SHBG) 5,
prefrontal cortex 27 70, 165, 170, 254, 257
pregnanediol corrected for creatinine (PdG) sexual dysfunction 68
174^175, 179 sexual interest, loss of 165
pregnenolone 188, 212 short luteal phase 174
premature ageing syndromes 134^135 sildena¢l 68, 159
procollagen III peptide 144, 155 skin
progeria 134 ¢broblasts 251
progesterone 162, 175 ulceration 31
vasomotor symptom therapy 181 wrinkling 126, 127, 129
progestin therapy 181 sleep
prolactin 122 apnoea, testosterone therapy 7
and critical illness 210 deprivation 194
proopiomelanocortin 64 disturbances 39, 187^189
PROP1 gene 127 fragmentation 194
prostate disease 73^74 thyrotropin and 197
benign prostatic hyperplasia 80^81 withdrawal 194
dehydroepiandrosterone therapy 10 smoking 257
testosterone therapy link 7 smooth muscle cells 146^148, 149^151, 156,
volume and age 80 158
prostitution 68 contractile 146
protein kinase A 269 synthetic 146
puberty 59, 188 Snell dwarf mice 131^132
pure autonomic failure 270^271 social stress 190
sodium ions 265
Q somatopause 10^13, 98
somatostatin 102, 105^106, 194
quality of life 4, 19^20 testosterone e¡ects 108^110
speci¢c oestrogen receptor modulator
R 262^263
raloxifene 262 spermatogenesis 69^70, 85, 88
raphe nucleus 28, 30, 31 splanchnic bed, glucose uptake 229
reaction time 27 spontaneously hypertensive rat 31, 32
5a reductase 74 stomach fundus 159
renal hypercalciuria 257 stress
renin 269 brain response 27, 28, 29, 30^32
renin^angiotensin^aldosterone axis 267 and depression 32, 36^37
resistin 63 HPA axis 31, 46^48
restriction length polymorphism 55 Leydig cell suppression 211
Rothmund^Thomson syndrome 135 reaction classi¢cation 47
stressors 48^49
testosterone levels 37^38
S
stroke 54
sarcopenia 4, 157 substance P 212
seasonal a¡ects 204 substantia nigra 27
selective serotonin reuptake inhibitors 38 successful ageing 3^4
SUBJECT INDEX 289

supraoptic nucleus 28, 29, 269 therapy 196, 209^210

sympathetic nervous system 47, 54, 55, 61, see also speci¢c hormones
267 thyrotropin (TSH) 193^194, 198, 203
synapses 27, 41^42 b-subunit 194
syncope 267 and critical illness 209^210
syndrome X (metabolic syndrome) 54, 56, non-thyroidal illness 197
224 secretory pattern 194
sleep and 197
T thyrotropin-releasing hormone (TRH) 194
T cells 10 thyroxine 194, 198
T3 (triiodothyronine) 194, 198 time-of-day e¡ects 275, 276^277
critical illness 209 toxins 49
T4 (thyroxine) 194, 198 transforming growth factor b 251
tamoxifen 122 triiodothyronine (T3) 194, 198
testis critical illness 209
dysfunction 83^87 trimetaphan 272
volume 69^70, 85 TrkB 29
testosterone 6, 144 tumour necrosis factor-a 206
age-related decline 5, 70, 82^83, 85, 98
angina treatment 158 U
assays 71
bone density 6, 7, 85, 255^256 urine excretion, time-of-day 276
cardiovascular risk 74
carotid artery intima/medial thickness V
18
critical illness 210^211 vaginal dryness 164
depression-related decrease 6 Vancouver Ovulation and Bone Change
fat mass 6^7 Cohort 182
female reproductive ageing 165, 169, vasoactive intestinal peptide 210
170^171 vasodilation 74, 158^159
growth hormone autonegative feedback vasomotor symptoms 177, 180^181, 189,
107^108 190
growth hormone-releasing hormone progesterone/progestin therapy 181
actions 108 vasopressin 267, 269^270
growth hormone-releasing peptide activity ageing 269^270
110^112 receptors 1A and 1B 269
growth hormone synergy 99^100 venlafaxine 189
measurement 71 Venn diagrams 60
somatostatin activity 108^110 Viagra (sildena¢l) 68, 159
stress-related decrease 37^38 visceral adipose tissue 17
therapy 6^7, 72^73, 77^78, 170^171 vitamin D metabolites 256, 263
visceral adipose tissue e¡ects 17
tetraiodothyronine (T4) 194, 198 W
thirst 266, 270
thyroid axis 193^204 war veterans 47, 61
critical illness 209^210 wasting syndrome 205, 207
thyroid disease 194^196 water homeostasis 265^278
thyroid hormones age-dependent pressor e¡ect 270^273
circulating hormone mix and age 204 body weight 265, 266
secretions, normal range 202 extracellular 265^266
290 SUBJECT INDEX

water homeostasis (cont.) working memory 27

intake prediction 266 wound healing and depression 33
intracellular 265
Werner’s syndrome 134
Z
white matter 27
Wistar^Kyoto rat 31 zinc 197