NBS SPECIAL PUBLICATION 450

U.S. DEPARTMENT OF COMMERCE / National Bureau of Standards

 NATIONAL BUREAU OF STANDARDS

The National Bureau of Standards^ was established by an act of Congress March 3, 1901. The Bureau's overall goal is to
strengthen and advance the Nation's science and technology and facilitate their effective application for public benefit. To this
end, the Bureau conducts research and provides: (1) a basis for the Nation's physical measurement system, (2) scientific and
technological services for industry and government, (3) a technical basis for equity in trade, and (4) technical services to pro-
mote public safety. The Bureau consists of the Institute for Basic Standards, the Institute for Materials Research, the Institute
for Applied Technology, the Institute for Computer Sciences and Technology, the Office for Information Programs, and the
Office of Experimental Technology Incentives Program.

THE INSTITUTE FOR BASIC STANDARDS provides the central basis within the United States of a complete and consist-
ent system of physical measurement; coordinates that system with measurement systems of other nations; and furnishes essen-
tial services leading to accurate and uniform physical measurements throughout the Nation's scientific community, industry,
and commerce. The Institute consists of the Office ofMeasurement Services, and the following center and divisions:
Applied Mathematics — — Mechanics — Heat — Optical Physics — Center
Electricity for Radiation Research — Lab-
oratory Astrophysics^ — Cryogenics" — Electromagnetics" — Time and Frequency".
THE INSTITUTE FOR MATERIALS RESEARCH conducts materials research leading to improved methods of measure-
ment, standards, and data on the properties of well-characterized materials needed by industry, commerce, educational insti-
tutions, and Government; provides advisory and research services to other Government agencies; and develops, produces, and
distributes standard reference materials. The Institute consists of the Office of Standard Reference Materials, the Office of Air
and Water Measurement, and the following divisions:

Analytical Chemistry — Polymers — Metallurgy — Inorganic Materials — Reactor Radiation — Physical Chemistry.
THE INSTITUTE FOR APPLIED TECHNOLOGY provides technical services developing and promoting the use of avail-
able technology; cooperates with public and private organizations in developing technological standards, codes, and test meth-
ods; and provides technical advice services, and information to Government agencies and the public. The Institute consists of
the following divisions and centers:

Standards Application and Analysis —

Electronic Technology —
Center for Consumer Product Technology: Product
Systems Analysis; Product Engineering —
Cpnter for Building Technology: Structures, Materials, and Safety; Building
Environment; Technical Evaluation and Application —
Center for Fire Research: Fire Science; Fire Safety Engineering.

THE INSTITUTE FOR COMPUTER SCIENCES AND TECHNOLOGY conducts research and provides technical services
designed to aid Government agencies in improving cost effectiveness in the conduct of their programs through the selection,
acquisition, and effective utilization of automatic data processing equipment; and serves as the principal focus wthin the exec-
utive branch for the development of Federal standards for automatic data processing equipment, techniques, and computer
languages. The Institute consist of the following divisions:

Computer Services — Systems and Software — Computer Systems Engineering — Information Technology.
THE OFFICE OF EXPERIMENTAL TECHNOLOGY INCENTIVES PROGRAM seeks to affect public policy and process
to facilitate technological change in the private sector by examining and experimenting with Government policies and prac-
tices in order to identify and remove Government-related barriers and to correct inherent market imperfections that impede
the innovation process.

THE OFFICE FOR INFORMATION PROGRAMS promotes optimum dissemination and accessibility of scientific informa-
tion generated within NBS; promotes the development of the National Standard Reference Data System and a system of in-
formation analysis centers dealing with the broader aspects of the National Measurement System; provides appropriate services
to ensure that the NBS staff has optimum accessibility to the scientific information of the world. The Office consists of the
following organizational units:

Office of Standard Reference Data — Office of Information Activities — Office of Technical Publications — Library —
Office of International Standards — Office of International Relations.

1 Headquarters and Laboratories at Gaitbersburg, Maryland, unless otherwise noted; mailing address Washington, D.C. 20234.
' Located at Boulder, Colorado 80302.
Blood pH, Gases, and Electrolytes

Proceedings of the workshop on

pH and Blood Gases, held at the
National Bureau of Standards,
Gaithersburg, Maryland,
July 7-8, 1975

Richard A. Durst, Editor

Analytical Chemistry Division
Institute for Materials Research
National Bureau of Standards
Washington, D.C. 20234

Sponsored by:

American Association for Clinical Chemistry

American Society of Clinical Pathologists
International Federation of Clinical Chemistry
National Committee for Clinical Laboratory Standards
U.S. Department of Commerce, National Bureau of Standards

U.S. DEPARTMENT OF COMMERCE, Juanita M. Kreps, Secretary

Dr. Sidney Harman, Under Secretary

Jordan J. Baruch, Assistant Secretary for Science and Technology

issued June 1977

 —.

Library of Congress Cataloging in Publication Data

Workshop on pH and Blood Gases, National Bureau of Standards,

1975.
Blood pH, gases and electrolytes.
(NBS special publication; 450)
Supt. of Docs, no.: €13.10:450
1. Blood — —
Analysis and chemistry Congresses. 2. Blood gases
—
Analysis Congresses. 3. Hydrogen-ion concentration Measure- —
— — —
ment Congresses. 4. Electrolytes Analysis Congresses. I. Durst,
Richard A., 1937-11. American Association of Clinical Chemists.
III. Title. IV. Series: United States. National Bureau of Standards.
Special publication; 450.
QC100.U57 no. 450 [RB45] 602'.ls [616.07'561] 76-608179

National Bureau of Standards Special Publication 450

Nat. Bur. Stand. (U.S.), Spec. Publ. 450, 338 pages (June 1977)

CODEN: XNBSAV

U.S. GOVERNMENT PRINTING OFFICE

WASHINGTON: 1976

For sale by the Superintendent of Documents, U.S. Government Printing OflSce, Washington, D.C. 20402
(Order by SD Catalog No. C13.10 :450) . Stock No. 003-003-01792^2 Price $4.75
(Add 25 percent additional for other than U.S. mailing)
 FOREWORD

The Analytical Chemistry Division of the National Bureau of Standards (NBS), Institute
for Materials Research has as its goal the development of new techniques of analysis,
standard materials and instruments to improve the accuracy and precision with which analyses
are performed. To obtain as broad input as possible from the user community on the needs
and trends in analytical chemistry and to disseminate information outside NBS, groups of
experts are invited to attend workshops at the Bureau. These workshops often are cosponsored
by other scientific and professional organizations. These proceedings include the pre-
sentations and deliberations of one of a series of Analytical Chemistry Division sponsored
workshops and are concerned with the accurate measurement of blood pH and gases. Other
workshops in this series have included Secondary Ion Mass Spectrometry, Aerosol Measure-
ments, Marine Pollution Monitoring (Petroleum), Monte Carlo Calculations in Ion Probe
Microanalysis and Scanning Electron Microscopy, Standards for Spectrophotometry and Lu-
minescence, Standards Required in Marine Science, Standards Required for Offshore Oil
Drilling, and Standards Required for Environmental Monitoring of Oil Shale Processing.

The international character of the attendance at this workshop and of the interest
exhibited by the cosponsoring organizations: the American Association of Clinical Chemists,
the American Society of Clinical Pathologists, the International Federation of Clinical
Chemistry, and the National Committee for Clinical Laboratory Standards, are indicative of
the interest and importance of this subject throughout the clinical community. This pub-
lication should serve to characterize the current state of accuracy for blood pH and gases
measurements and to point out the standards required in this field.

P. D. LaFleur
Chief
Analytical Chemistry Division

iii
 PREFACE

This publication is the formal report of the proceedings of the Workshop on pH and
Blood Gases sponsored by the American Association of Clinical Chemists, American Society of
Clinical Pathologists, International Federation of Clinical Chemistry, National Committee
for Clinical Laboratory Standards, and the National Bureau of Standards. The purpose of
this meeting was to discuss the status and needs in this very important area of clinical
measurement and to provide a starting point for future cooperative efforts on an inter-
national level toward the standardization of pH and blood gas measurements and the various
quantities and terms used in this field.

The importance of the acid-base status for clinical diagnosis is reflected in the fact
that the determination of pH and blood gases is one of the most widely and frequently
performed clinical tests. However, even though it has been more than a half century since
Henderson and Van Slyke described the chemical and physiological relationships and mechanisms
involved in acid-base balance, there still remain many uncertainties as to which indices are
best measured, what parameters should be derived from the measured data and how to calculate
them, and finally how these data should be interpreted by the clinicians.

The first requirement is an internationally agreed-upon set of definitions, terms, and

symbols. The conflict of ideas that still exists is serious, not only because of the need
for uniformity in nomenclature, but also for the more pragmatic reason that, as highly
sophisticated blood pH/gas analyzers are developed, the derived parameters will be critically
dependent upon the assumptions and "standard values" used in the calculations. It is hoped
that this workshop and the resulting proceedings will provide a preliminary step toward
achieving international accord in this area. Of course, a two-day workshop allowed little
time for problem solving or the resolution of details. However, it did permit the exposure
of diverse viewpoints on the critical needs in this field so that subsequent efforts, such
as those of the newly established IFCC Expert Panel on pH and Blood Gases, can be directed
toward the most important topics.

Editorial changes have been made to achieve some measure of internal consistency,
especially with regard to the International System of Units (SI). However no changes are
made in the nomenclature used by the authors since one of the purposes of this workshop was
to reveal the diversity of symbols, terms, definitions, etc., in order to arrive at some
future uniformity in this subject. Also, certain fundamental discussions appear in more
than one paper but, again, in view of the diversity of approaches to this subject, this
duplication was intentionally preserved. Although all the papers have been reviewed by the
editor, the views expressed are entirely the responsibility of the individual authors.

In order to specify the procedures adequately, it has been necessary to identify

commercial materials and equipment in this report. In no case does such identification
imply recommendation or endorsement by the National Bureau of Standards, nor does it imply
that the material or equipment is necessarily the best available for the purpose.

The organization of the workshop and publication of these proceedings would not have
been possible without the cooperation and assistance of a large group of people within the
National Bureau of Standards. Particular thanks are given to Philip D. LaFleur, Chief of
the Analytical Chemistry Division, for his encouragement and support in this effort, and to
Ronald B. Johnson and Robert F. Martin of the Institute for Materials Research for their
assistance in the business matters associated with this project. I also want to express my
appreciation to Ellen N. Ring and her staff in the IMR Text-Editing Facility. The NBS
Office of Information Activities, with special help from Sara R. Torrence, and also Rebecca
J. Morehouse and Miriam K. Gland of the Office of Technical Publications gave invaluable
assistance in many phases of the effort, varying from the initial meeting arrangements to
the final publication of these proceedings. Special thanks are given to Gloria Burdick for
her help in preparing the original workshop correspondence and program, Rosemary Hormuth for
transcription of the recorded discussions, and Carolyn A. Shipley for her continuous effort
in typing the coded manuscripts.

Richard A. Durst

iv
 ABSTRACT

On July 7-8, 1975, a workshop was held at the National Bureau of Standards to discuss
the status and needs of this very important area of clinical measurement. A major goal of
this workshop was the initiation of cooperative efforts on an international level toward the
standardization of pH and blood gas measurements and the various quantities and terms used
in this field.

To this end, the first technical session was concerned with the acid-base status of
blood and included the topics: Definitions of Quantities and Concepts; Recommendations of
Nomenclature, Physiological Terminology and Symbols; Reference Values; and the Evaluation of
Nomograms and Algorithms. The second session addressed itself to the more practical aspects
of this subject and included the topics: Blood Sampling, Handling, and Storage; Instrument
Specifications; Quality Control and Standards; and the Development of Reference Methods.
Finally, a brief session was held on the newer topic of the electrometric measurement of
blood electrolytes.

This volume contains all of the papers invited for presentation at the workshop by some
of the leading clinical and medical authorities on this subject and also includes a trans-
cription of the extensive discussion sessions.

Key words: acid-base status; blood electrolytes; blood gases; blood pH; calcium; carbon
dioxide; hydrogen ion concentration; nomograms; oxygen; P-^ ; pH; P^ ; potassium; sodium.

V
 CONTENTS

PAGE

FOREWORD iii

PREFACE IV

Part I. Acid-Base Status

DEFINITIONS OF ACID-BASE QUANTITIES: TERMINOLOGY, SYMBOLS, AND SI UNITS

01 e Siggaard-Andersen 1

THE BUFFER VALUE OF PLASMA, ERYTHROCYTE FLUID AND WHOLE BLOOD

Ole Siggaard-Andersen, M. R0rth, and D. A. P. Strickland 11

ACID-BASE ALGORITHMS
Ole Siggaard-Andersen 21

DETERMINATION OF TOTAL CO2 CONCENTRATION IN BLOOD OR PLASMA

P. Rispens, E. J. van Kampen, and W. G. Zijlstra 27

THE APPARENT OVERALL FIRST DISSOCIATION CONSTANT OF CO2 IN PLASMA

P. Rispens and W. G. Zijlstra 33

QUANTITATIVE RELATIONSHIPS BETWEEN TOTAL CO2 CONCENTRATION IN BLOOD AND PLASMA,

PLASMA BICARBONATE CONCENTRATION, PLASMA pH AND CARBON DIOXIDE TENSION
BETWEEN 16-42 °C
P. Rispens, J. P. Zock, and W. G. Zijlstra 39

P^Q INDEPENDENT QUANTITIES, WHY OR WHY" NOT

P. Rispens, W. G. Zijlstra, and E. J. van Kampen 47

BASE EXCESS. WHY REOPEN THE ACID-BASE DEBATE?

P. J. N. Howorth 53

USE OF IN VIVO CO2 TITRATION CURVES IN THE PHYSIOLOGICAL ASSESSMENT

OF ACID-BASE BALANCE
P. J. N. Howorth 57

RELATIVE ACTIVITY AND SI UNITS

B. F. Visser and A. H. J. Maas 69

SEMI-EMPIRICAL ACID-BASE PROGRAM

B. F. Visser, A. J. Hoelen, J. A. Kreuger, and A. H. J. Maas 73

"ACID-BASE SEMANTICS"--A CENTURY OF THE TOWER OF BABEL

Harry F, Weisberg 75

vii
 PAGE

TRI-SLIDE™ CALCULATOR FOR HENDERSON-HASSELBALCH EQUATION AND C02RREC°t-02-SLIDE™

FOR TEMPERATURE CORRECTIONS OF pH, Pco^, AND P02
Harry F. Weisberg 91

AIDS FOR EVALUATION OF ACID-BASE IMBALANCE— DIAGRAMS, NOMOGRAMS,

AND SLIDE-RULES
Harry F. Weisberg 103

THE OVERALL FIRST IONIZATION EQUATION OF CARBONIC ACID AS RELATED

TO CO2 IN GAS PHASE: A NEW pK
A. H. J. Maas and B. F. Visser 119

TOWARDS A PHYSIOLOGIC NOMENCLATURE FOR IN VIVO DISTURBANCES

OF ACID-BASE BALANCE
Jordan J. Cohen 127

MINIMAL ACCEPTANCE CRITERIA FOR ACID-BASE NOMOGRAMS

Jordan J. Cohen 131

A PHYSIOLOGICAL APPROACH TO ACID-BASE DIAGNOSTICS

Poul Kildeberg and Knud Engel 133

CO2 SOLUBILITY, pK' AND RELATED FACTORS IN ACID-BASE BALANCE

William H. Austin 143

DEFINITION OF OXYGEN SATURATION AND CHARACTERIZATION '

OF OXYGEN-HEMOGLOBIN AFFINITY
Arthur L. Mai enfant 153

PROBLEMS ASSOCIATED WITH THE DEFINITION OF MEASURED AND CALCULATED

QUANTITIES IN BLOOD pH AND GAS ANALYSIS
t

Robert W. Burnett 163

Part II. Instrumentation, Methodology, and Standards

ARTERIAL AND VENOUS BLOOD SAMPLES IN ACID-BASE BALANCE

William H. Austin . 167

BLOOD SAMPLING AND HANDLING IN THE DETERMINATION OF BLOOD pH

AND BLOOD GASES
Arthur H. Richards 171

NON-ANALYTICAL SOURCES OF LABORATORY ERROR IN pH AND BLOOD GAS ANALYSIS

Jack H. Ladenson 175

vi i i
 PAGE

INSTRUMENT SPECIFICATIONS
S. Raymond Gambino 191

EFFECTS OF THE LIQUID JUNCTION ON pH MEASUREMENT IN BLOOD;

THE 0.160 MOL/L SODIUM CHLORIDE BRIDGE

A. H. J. Maas 195

THE CRITICAL CARE LABORATORY: A 10- YEAR PERSPECTIVE

Myron B. Laver and Domenic R. Misiano 201

A THEORETICAL AND PRACTICAL ANALYSIS OF P^ MICROELECTRODE BEHAVIOR:

THE THREE-SHELL MODEL
R. G. Buckles, H. Heitmann, and M. B. Laver 207

DYNAMIC RESPONSE OF A pCOz ELECTRODE

S. J. Pace and M. J. D. Brand 227

MONITORING OF OXYGEN PRESSURE IN HUMAN AND ANIMAL BLOOD

H. P. Kimmich and F. Kreuzer 235

ALTERNATIVE METHODS OF CO2 MEASUREMENT, WITH PARTICULAR REFERENCE

TO CONTINUOUS RECORDING
L. H. J. van Kempen and F. Kreuzer ,
239

NBS STANDARDS FOR pH AND ION ACTIVITY MEASUREMENTS IN BIOLOGICAL FLUIDS

Richard A. Durst and Roger G. Bates 247

USE OF CARBON DIOXIDE- AND OXYGEN TONOMETERED PHOSPHATE-BICARBONATE-

CHLORIDE-GLYCEROL-WATER MIXTURES FOR CALIBRATION AND CONTROL OF pH,
PCO2, AND PO2 ELECTRODE SYSTEMS
A. H. J. Maas, A, H. Veefkipd, R. A. M. Van den Camp, A. J. Teunissen,
A. B. T. J. Boink, and T. J. C. Ruigrok 257

CALIBRATION OF BLOOD GAS ANALYZERS

Alan H. Runck 267

QUALITY CONTROL
S. Raymond Gambino 273

QUALITY CONTROL AND STANDARDS

Daniel C. Noonan 275

DEVELOPMENT OF REFERENCE METHODS: BLOOD GAS ANALYSIS

Arthur L. Mai enfant and Kevin D. Fall an 279

QUALITY CONTROL AND STANDARDS

SfJren K. Sfirensen 285

ix
 PAGE

DEVELOPMENT OF REFERENCE METHODS

Sfiren K. S^rensen 289

Part III. Electrolytes

STANDARDIZATION OF ION-SELECTIVE ELECTRODES FOR SERUM ANALYSIS

M. S. Mohan and Roger G. Bates , 293

ELECTROLYTE ACTIVITIES IN HUMAN BLOOD PLASMA

M. J. D. Brand and W. J. Scott 301

THE KING'S COLLEGE HOSPITAL ION-SELECTIVE ELECTRODE

SERUM ELECTROLYTE ANALYZER
A. D. Hirst, P. Gay, P. Richardson, and P. J. N. Howorth 311

Part IV. Discussion

WORKSHOP DISCUSSION 315

X
 ,

National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH

and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

DEFINITIONS OF ACID-BASE QUANTITIES: TERMINOLOGY, SYMBOLS,

AND SI UNITS

Ole Si ggaard-Andersen
Department of Clinical Chemistry
University of Copenhagen
Copenhagen County Hospital
Herlev, Denmark

In clinical medicine, patients are often encountered with disturbances characterized by

accumulation or loss of hydrogen ions and/or carbon dioxide. For routine clinical purposes,
such disturbances are evaluated by measuring various quantities, for example, the "blood pH"
the "hydrogen ion concentration", the "eucapnic pH", the "PCO2", the "total -C02"-, the "plasma
bicarbonate", the "CO2 combining power", the "CO2 capacity", the "standard bicarbonate", the
"buffer base", and the "base excess." Most of these designations are "clinical slang" and
they ought to be replaced by the more systematical chemical terminology which has been re-
commended by various international organizations [1-5]^.

1. Names and Symbols for Chemical Quantities

The name of a physical or chemical quantity must include a specification of the kind of
quantity and a sufficiently detailed characterization of the physical system to which the
quantity refers. This often involves a specification of one or more components of the
system or processes or reactions in the system.

Table 1 shows the names of some of the acid-base quantities according to the inter-
national recommendations [2], The names follow the scheme: System-component, kind of
quantity. For example: P-hydrogen ion, substance concentration equals 53 nmol/1. In sym-
bols this may be written [6]:

eH'^(P) = {fiH^(P)} • [cH^(P)] = 53 nmol/1 . (1)

The name as well as the symbol should (explicitly or implicitly) include the following
specifications:

(1) Kind of quantity The different kinds of quantities are defined by the Interna-
.

tional Organization for Standardization in a publication series: ISO 31 [3]. The total
number of different kinds of quantities amounts to several hundred. In clinical chemistry,
the kinds of quantities listed in table 2 are of special interest. The symbol for the
kind of quantity should always be a single letter and should always be printed in italics
(sloping type), while all other symbols are printed in Roman type (upright) [1], An excep-
tion to this rule is pH which unfortunately has been defined as a special kind of quanti-
ty (see later), but nevertheless is written in upright type.

According to the recommendations of the International Federation of Clinical Chemistry

(IFCC) and the International Union of Pure and Applied Chemistry (lUPAC) [3,4], substance
oonoentration (or simply concentration) should be employed rather than mass concentration in
all cases where a formula unit can be defined for the component. For example, the hemo-
globin concentration should be the substance concentration based on a molar mass, of
16114 g/mol referring to one quarter of a hemoglobin molecule with two a and two b chains.

Figures in brackets indicate literature references at the end of this paper.

1
Table 1. Names of various acid-base quantities according to the recommendations
of the International Federation of Clinical Chemistry (IFCC) and the Interna-
tional Union of Pure and Applied Chemistry (lUPAC) [2]. (The symbols are
author's suggestions.)

Name Symbol

U - Acid (H), substance concentration (method) ActH^(U)

aB - Base (H"'^-binding groups), substance concentration eBB' (aB)

(Singer and Hastings, 1948)

aB - Base (H^-binding groups), substance concentration AeB' (aB)

difference (method; Pt - Norm)

Gas (aB equil.) - Carbon dioxide, partial pressure pC02(aB)

(method; 37.0 °C)

(aB) P - Carbon dioxide, substance concentration cC02(Pc:aB)

P - Carbonate + carbon dioxide, substance con- ctC02(P)

centration

P - Hydrogen carbonate ion, substance concentration cHC03(PcstB)

(blood; aiOz) = 0.21 mmol/1; e(C02) = l-IQ mmol/1;
0 = 37 °C)

(aB)P - Hydrogen ion, substance concentration cH"^(PcaB)

aB - Plasma, pH (37 °C) paH"^(PcB)

Pt - Urine, pH paH'^(U)

Table 2. Various kinds of quantities generally employed in clinical chemistry.

Name Symbol Unit Definition (C = component, S = system)

Number N 1 (one) number of particles or molecules

Amount of substance n mol number of formula units:

6.023 1023
• nC/mol = BZ •

Mass m kg

Volume V 1 (or m^)

Number fraction X 1 Cone) ^Cj(S) = /vc. ;S)/EffC.(S)

i

Substance fraction X 1 Cone) a;C.(S) = nC [S)/EnC.(S)

•

J
i

= mC(S )/mS
Mass fraction w 1 Cone) wC(S)
= 7C(S )/7S
Volume fraction 1 Cone) <|)C(S)

Number concentration C 1-1 CC(S) = A/C(S' m

Substance concentration c mol/1 cC(S) = nC(S^ /ys

Mass concentration P kg/1 pC(S) - mC(S' /ys

Molality m mol /kg mC(S) = nc(s; /mSolvent(S)

pC(S) = xc(s;
(Partial) pressure P Pa

2
 ,

(2) Component The names of the components are specified in considerable detail accord-
.

ing to a certain set of rules [2]. Several of the names may be expressed in alternative
forms, e.g., carbonate + carbon dioxide = total CO2. The symbols for the inorganic sub-
stances should be the familiar chemical symbols. For certain organic substances, symbols
are also almost international, e.g., hemoglobin = Hb, protein = Pr.

(3) System The above mentioned recommendations include a list of symbols for the
.

systems, e.g. blood = B, arterial blood = aB, plasma = P, urine = U, etc.

(4) Number Braces {} around the symbol for the quantity indicate the pure number
. .

value of the quantity [1].

(5) Unit Brackets [] around the symbol for the quantity indicate the unit of the
.

quantity [1]. Whenever possible, it is recommended to use the SI units (m, kg, s, -system).
This means that the traditional unit for pressure, mmHg, should be replaced by Pa (= N/m^),
conversion factor: 1 mmHg = 0.133 kPa.

The names recommended by IFCC and lUPAC are especially designed for tables and labora-
tory records [2]. In a running text or in the spoken language, a different order is often
chosen, e.g., the concentration of hydrogen ions in the plasma, the concentration of plasma
hydrogen ion, the plasma hydrogen ion concentration, the plasma concentration of hydrogen
ions, the hydrogen ion concentration in the plasma. All combinations are seen in the
1 iterature.

2. Stock Components, Amount of Substance, and Chemical Potential

The fundamental basis for a description of acid-base physiology was the development of
physical chemistry and thermodynamics which led to an exact description of a chemical system
at the beginning of this century.

The organism is an open chemical system in a dynamic equilibrium with its surroundings.
A complete description of the system therefore requires not only a description of the momen-
tary composition of the different body phases, including the aoid-base status of the blood,
but also requires a description of the rate of intake, production, conversion, and excretion
of acids and bases, i.e. , a description of the aoid-base balance of the organism.

A chemical system, e.g., blood plasma, can theoretically be prepared from a limited num-
ber of stock components, for example, H2O, NasPOi^, HCl CO2, O2, NaCl etc.
, It is not the
,

purpose here to describe all the stock components of plasma. We are in the present connec-
tion especially interested in changes in the amount of hydrogen ions in the system.

Adding hydrogen ions to the system presents a special problem because they cannot be
added or removed alone, according to the law of electrical neutrality. With the present
choice of stock components, a hydrogen ion is always accompanied by a chloride ion. Alter-
natively, hydrogen ions couia be added or removed in exchange for sodium ions which virtually
means removing or adding NaOH. The first method is described as addition or removal of
strong acid, the second method as removal or addition of strong base, and the two methods
are considered equivalent, although the effects on the ionic strength will be slightly
different.

The stock components can react mutually giving rise to several derived components,
e.g.

CO2 + H2O ^ H2CO3 ^ H"^ + HCO3 • (2)

with the above-mentioned choice of stock components, HCO3 is considered a derived component.
However, the choice of stock components is arbitrary; one could also choose H2O, NaHCOs, and
HCl as stock components, in which case CO2 would be a derived component.

3
 The first mentioned choice of stock components is based on physiological considerations,
according to which H2O, CO2, and strong acid or base are ingested or formed in the organism
and excreted in the lungs and kidneys.

The important equation which expresses the change in free energy of the system (dc) when
the amounts of stock components are changed (at constant temperature and pressure) is the
following:

dG = yH"*" •
dntH"^

+ yC02 •
dntC02

+ yCl" •
dntCl"

+ yNa^ •
dntNa"*"

+ y02 •
dnt02 J

t = prefix designating total amount added,

n = amount of substance (unit: mole),
y = electrochemical potential (unit: joule/mole),
= y + 2 • F •
(j).

where y = chemical potential, z = charge number, F = Faraday constant, and ^ = inner elec-
trical potential.

This important equation forms the basis for derivation of important relationships among
the various quantities. In the present connection, the equation only serves to define the
ahemieal potential as the partial molar free energy of the different components, and the
equation serves to emphasize that each stock component of the system can be characterized by
two quantities: (1) an intensive quantity, i.e., the chemical potential, and (2) an extensive
quantity, i.e., the amount of substance of added component.

The component of primary interest is (strong acid or base), but it is necessary to

deal with another component which greatly influences the chemical potential of H"*" in an
aqueous solution, namely 002- In the following, the intensive and extensive quantities are
described in more detail for these two components.

3. Chemical Potential of Hydrogen Ions, AyH^ or pH

From a physico-chemical point of view, the chemical potential of the hydrogen ions is
the most important quantity for description of the "acidity" of a solution. The significance
of the chemical potential is that hydrogen ions tend to diffuse from a phase with a higher
chemical potential to a phase with a lower potential, unless an electrical potential dif-
ference exists between the phases. In that case, the equilibrium condition requires that
the electrochemical potential [i.e. the sum of the chemical potential and the electrical
,

potential times s times the Faraday constant) should be identical in the two phases.

It must be emphasized that it is impossible (theoretically) to measure absolute chemi-

cal potentials. Using two ideal H'*' electrodes: E|^lS|^|KCl |Sp|Ej^, we can measure yH'''(Sp) -

yH"*'(S|^) + FEA£'(LJ). It is possible to minimize ea£'(LJ) by means of a saturated KCl bridge.

If S. is a reference solution (S®), we can therefore measure (approximately):

AyH^(sQ/SJ = Ao^H'^(Sr) = yH'^(SR) mH^S^) (4)

i.e.:, the excess chemical potential of hydrogen ions in solution Sp. The reference solution
is chosen so that aH'*'(S®) = 1 (see below).

Often the activity of the hydrogen ions is used as a measure of the acidity. The acti-
vity is related to the chemical potential as follows:

aV(S) = exp^i^ (5)

ra*H"'(S®/S) = r.a*H^(S) = ^^^^ e aH^(S) (6)

where a* is absolute activity and a is relative activity; r is a convenient mathematical

symbol for ratio (compare a: Inra The reference solution (S®) is defined by
= Alnx). tl

following equations which also define the activity coefficient y:

^H+(s) = ltL(Sl^_|L(Sl. ^ o(S) - 1:^yH^(S) - (7)

1 .

mH (S^)

S® may be said to be a hypothetical ideal solution with mW^{s'^) = 1 mol/kg and yH^(S®) = 1.
Actually, aH"*" is about 1 in a hydrochloric acid solution with mW^ about 0.85 mol/kg, i.e.,
yW^ = 1.18. yH"*" varies with the ionic strength. For human plasma, yH''"(P) is often assumed
to be about 0.8.

Calling the mathematical operator -log for p, gives the familiar pH quantity:

-logaH"^ = paH* = pH . (8)

It will be seen that changes in pH (with opposite sign) are directly proportional to changes
in the chemical potential of the hydrogen ions.

A number of international standard buffers have been defined in order to enable re-
producible pH measurements by means of electrodes, and the pH concept has thereby been given
an international "operational definition." It was unfortunate, however, that lUPAC defined
pH as a special derived kind of quantity [3].

Occasionally the hydrogen ion aonoentTation [oW^) is used as a measure of the acidity.
The relationship between cH+ and ihW^ {i.e., molality) is the following:

cH'^(S) = mH'^(S) • pH20(S) , (9)

where pH20(S) is the mass concentration of H2O in the system [e.g., for plasma: pH20(P) =
0.94 kg/1). In the medical literature, opinion is divided as to whether pH or eH+ should be
preferred as a measure of the acidity [6]. For "medical purposes," the cH+ of the plasma is
generally calculated without distinguishing between molality and concentration and without

5
taking the activity coefficient into account: cH'*' = anti1og(9 - pH) nmo1/l. The correct
equation would be:

.
pH20(P)/(kg/1)
cH (P) = T • antilog(9 - pH) nmol/1 (10)
yH^P)

Now that SI units are strongly recommended, it seems logical to use the SI unit joule
also for the "pH quantity", i.e., to use excess chemical potential of hydrogen ions. For
human arterial blood plasma, the reference range would be:

AyH^: -43.76 to -44.23 kJ/mol , and the extreme pathological range:

AyH"^: -40.37 to -46.31 kJ/mol , (corresponding to pH 7.37 to 7.45, and 6.80 to

7.80, respectively).

For conversion of pH values to AyH^ and vice versa, the following equation applies:

AvH"^ = -i?? • In 10 • pH , (n)

or for T = 310.15 K:

AyH^ = -5.937 . pH kJ/mol . (12)

4. Titratable Acid or Base; the Base Excess Concept, AetH^ or AcB'

The amount of added hydrogen ions in the system (ntH^) i.e., added strong acid, is an ,

arbitrary quantity which requires a definition of the system before the addition. If the
initial state of the system is defined, then the amount of added strong acid or base can
be determined by back titration with strong base, respectively, strong acid. In other words,
one can define and determine the difference (AntH''"): the amount of total hydrogen ion in
the system before titration minus the amount of total hydrogen ion in the system after ti-
tration. By dividing with the volume of the system, this extensive quantity can be expressed
as an "excess concentration" of total hydrogen ion:

AetH^(S) = AMtH"^(S)/yS . (13)

In the usual terminology, this quantity with opposite sign is designated the excess concen-
tration of base, or simply the "base excess":

ActH'^(S) = -AeB'(S) . (14)

The base excess concentration can also be expressed as being the concentration of
titratable base minus the concentration of titratable acid. For plasma, the endpoint of ti-
tration is defined at pH = 7.40, pCOz = 40 mmHg (= 5.33 kPa) and t = 37 °C. For whole ,

blood, the endpoint is the same, the pH value referring to the plasma phase. For urine, the

6
endpoint may be defined at pH = 7.40, pC02 = 0, pNHs = 0, and t = 37 °C; by this choice HCO3
is included as titratable base, while NH^ is included as titratable acid.

For urine, the titratable acid or base is generally determined directly by titration.
For blood or plasma, however, the quantity is generally determined indirectly by calculation
from the directly measured pH and pC02 values [6].

It appears that the concentration of total hydrogen ion in the plasma is witho t prac-
tical significance, contrary to, for example, the concentration of total calcium ion 1. the
plasma. The reason is that the solvent itself (H2O) contains large amounts of hydrogen luns
in bound form. Therefore, it is only relevant to determine changes or differences in the
concentration of strong acid or base in the solution. Apart from this, however, there is
no principal difference between the description of H"*" and Ca''""'': both ions can be added to or
removed from the system (together with an indifferent anion) and both are bound specifically
by binding groups, which in the case the hydrogen ions, are called base groups.

It is only in recent years that the base excess concentration has been used routinely
for description of the acid-base status of the blood. Originally, the bicarbonate concen-
tration or the total CO2 concentration was used as indicators of the accumulation of non-
carbonic acid or base, i.e., as measures of a metabolic acid-base disturbance. When dealing
with a pure bicarbonate solution, the change in the bicarbonate concentration equals the
change in the concentration of titratable acid or base, and the bicarbonate concentration
is unaffected by changes in PCO2; but in the presence of non-bicarbonate buffers, the bicar-
bonate concentration also changes during isolated changes in the pC02 even when the concen-
tration of titratable acid or base is constant. In order to compensate for this effect, the
PCO2 was standardized to a normal value of 40 mmHg (= 5.33 kPa), and the concentration of
total CO2 or HCO3 of such standardized blood or plasma was designated "the plasma CO2 com-
bining power," "the CO2 capacity of the blood," or "the standard bicarbonate." However,
none of these quantities accurately reflect the accumulation of noncarbonic acid or base
in the blood or plasma.

5. The Chemical Potential or the Partial Pressure of CO2,

AyC02, or PCO2

Generally, the partial pressure of carbon dioxide in a gas phase in equilibrium with
the blood is used as a measure of the chemical potential of CO2. Conversion of pC02 to
yC02 or vice versa follows the equation

AVCO2 = ifT.lnCpCOj/lOl .3 kPa) , (15)

or at T = 310.15 K:

AyCOz =^5.937-log(pC02/101.3 kPa) kJ/mol . (16)

To be exact, pC02 should be multiplied by a fugacity coefficient which can be taken to be 1

for pressures below 100 kPa. The reference system is a system in equilibrium with a hypo-
thetical ideal gas phase with pC02 = 101.3 kPa.

For chemical substances, lUPAC has recommended the use of molecular quantities, i.e.,
substance concentration rather than mass concentration. This recommendation might be exten-
ded to a recommendation of using the excess chemical potential rather than other quantities
(activity, partial pressure, concentration of free component, etc.) which are presently
employed as substitutes for the chemical potential.

For human arterial blood, the reference range for the excess chemical potential of CO2
is: AyC02: -8.1 to -7.4 kJ/mol and the extreme pathological range is about:
, AyC02; -11.2
to -5.2 kJ/mol, (corresponding to pC02 4.3 to 5.7 kPa and 1.3 to 13.3 kPa, respectively).

7
 : ;

The partial pressure of CO2 is related to the concentration of dissolved CO2 according
to Henry *"s law:

0CO2 = aC02 • PCO2 , (17)

where aC02 is the solubility coefficient. Generally, aCOz includes the very low concentration
of H2CO3. The solubility coefficient for normal plasma is 0.306 mmol 1"^ • mmHg"^ at 37 °C. •

(or in SI units: 0.231 mmol/J) with a normal biological standard deviation of about 0.0004
mmol • 1"^ mmHg"^.
• For lipemic plasma, however, the solubility coefficient may be consid-
erably higher, up to about 0.33 mmol 1"^• mmHg-^ [6]. •

6. The Concentration of Total CO2, etC02

The extensive quantity expressing the amount of added CO2 in the system is converted to
the concentration of total CO2 in the system by division with the volume of the system:

ntC02(S)/7S = etC02(S) . (18)

The quantity has been and still is much employed for description of the acid-base status
of the blood. The reason is, most of all, that the concentration of total CO2 is easy to
measure gasometrical ly by means of the Van Slyke apparatus, or automatically by means of
Leonard Skegg's continuous flow analyzer. For clinical purposes, this quantity is generally
used as a measure of a metabolic acid-base disturbance, i.e., as an indirect measure of the
excess concentration of titratable acid or base.

The biaarbonate oonoentration of the plasma is often calculated as the concentration of

total CO2 minus physically dissolved CO2:

eHC03(P) = atC02(P) - eC02(P) . (19)

Therefore, the "bicarbonate" concentration generally includes carbonate and carbamate. The
bicarbonate concentration of the plasma is one of the components of the electrolyte scheme
(Gamble diagram) and it is used for calculation of the so-called anion deficit (undetermined
anions)

cUA" = cNA"^ - (oCr + cHCOi) . (20)

The concentration of undeterm-jned anions is used as an indicator of an accumulation of orga-

nic anions (lactate, e-hydroxybuturate, eto.) in the plasma.

7. Conclusion

In the preceding sections, four fundamental acid-base variables have been described,
relating to the two components H"*" and CO2:

(1) the excess chemical potential of H"*", AyH'*';

(2) the excess concentration of total H"*", t\otW^

(3) the excess chemical potential of CO2, AyC02; and

(4) the concentration of total CO2, etC02.

8
 Only two of the four variables are inde-pendent variables; the other two being dependent
variables

From a physiological point of view, AyC02 and AetH^ are the independent variables,
because the system is an open system in equilibrium with a gas phase (the alveolar air) with
a certain AyCOa value. In such an open system, it is possible independently to alter the
AyC02 of the gas phase or to add various amounts of strong acid or base (add or remove H+).
Clinically speaking, AyC02 is a parameter (i.e., indicator) of the alveolar ventilation, while
ActH'*' is a parameter of the non- respiratory acid-base balance.

References

[1] International Standard, ISO 31 Series, Technical Committee ISO/TC 12, Quantities, Units,
Symbols, Conversion Factors, and Conversion Tables, International Organization for
Standardization, Switzerland (available from the different National Standards Insti-
tutes) .

[2] List of Quantities in Clinical Chemistry, Recommendation 1973, lUPAC, Section on

Clinical Chemistry, Commission on Quantities and Units, and IFCC, Committee on Stan-
dards, Expert Panel on Quantities and Units, Pure and Appl. Chem. 37, 547 (1974).

[3] Manual of Symbols and Terminology for Physicochemical Quantities and Units, lUPAC,
Division of Physical Chemistry, Commission on Symbols, Terminology, and Units, Pure
and Appl. Chem. 21, 1 (1970).

[4] Quantities and Units in Clinical Chemistry, Recommendation 1973, lUPAC, Section on
Clinical Chemistry, Commission on Quantities and Units, and IFCC, Committee on Stan-
dards, Expert Panel on Quantities and Units, Pure and Appl. Chem. 37, 517 (1974).

[5] Recommendation for Use of SI in Clinical Laboratory Measurements, ICSH, IFCC, and
WARS, Z. Klin. Chem. n., 93 (1973).

[6] Siggaard-Andersen, 0., The Acid-Base Status of the Blood, 4th ed., 229 pp. (Williams
& Wilkins, Baltimore, and Munksgaard, Copenhagen, 1974).

9
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Gaithersburg, Maryland, July 7-8, 1975. Issued June 1977.

THE BUFFER VALUE OF PLASMA, ERYTHROCYTE FLUID AND WHOLE BLOOD

Ole Siggaard-Andersen, M. R^rth^ and D. A. P. Strickland^

Department of Clinical Chemistry
University of Copenhagen
Copenhagen County Hospital
Herlev, Denmark

The buffer value (6)^ for hydrogen ions in a solution (S) is generally defined as the
Slope of the titration curve at a given pH value when titrating the solution with strong
acid or base (B') [1]'+:

BH^(S)=M_|||. „)

The unit is mol/1, although in certain cases it may be advantageous to use molality {i^):

m'^iS) = 3;?2B'{S)/9pH(S), unit: mol/kg. (2)

An alternative definition of buffer value is:

= (3)
^fj§[f|,

where tC = total added component C. The SI unit is mol^ • 1"^ J~^. C may be H^ or any•

other component, e.g., Oz- Actually, the reciprocal value might be more relevant being the
first partial differential coefficient of the function:

AyC(S) = F(ctC(S), -, -, — ). (4)

The relationship between the two buffer values of eqs. (1) and (3) is:

eVcs) • <E • T ' InlO = bh'^(s). (5)

The buffer value of H"*" in a solution can be expressed as the sum of the buffer values of
the solvent (H2O) and the solutes (C). The change in concentration of added base can be
expressed as:

dcB'(S) = deOH"(S) - deH'^(S) + zdcC'(S), (6)

^Present address: Surgery Department D, Rigshospitalet, Copenhagen, Denmark,

^Department of Clinical Measurement, Westminster Hospital, London, England.
^Concerning terminology and symbols, the reader is referred to the preceding paper, e = con-
centration, y = chemical potential.
'Figures in brackets indicate the literature references at the end of this paper.

11
where C symbolizes a base group. Dividing with dpH gives:

6H^(S) = 3H"^{H20 c S) + Z6H"^(C c S). (7)

The buffer value of the water is approximately:

BH"^(H20 c S) « (cH^(S) + eOH"(S))-lnlO. (8)

Therefore the buffer value of the water can generally be ignored for 3 < pH < 11.

The buffer value for due to solute C(CH i:? H"^ + C-) is equal to:

8eC"(S) 3cCH(S)
BH^(C c S) =
(9)
8pH(S) 9pH(S)

The equality is only valid for cC = cC" + eCH = constant. If the system is an open system
where CH or C" disappears, the amount which has disappeared must be included in order to
obtain equality. This especially applies to the CO2 and NH3 buffers.

The buffer value of the solutes may be expressed as the molar buffer value (g ) or as
the specific buffer value (3,,):

6^H'^(C g S) h 6H"^(C c S)/cC(S), (10)

= S) 5 bH'^(C c S)/pC(S), (11)

where pC(S) is the mass concentration of C in S.

The buffer value is a partial differential and the other independent variables which
are maintained constant must be specified. The buffer value of a pure bicarbonate solution
[S] is widely different for a closed system with constant ctC02 and for an open system
with constant pC02. The following equations are obtained by differentiation of K . =
aH • cHC03/(atC02 - cHCOs):

K
'aak
2.303 etC02 (12)

aak

2.303 • cHC03(S). (13)

12
The maximal molar buffer value for a closed system is therefore 0.576 for pH = pXacA- For
pK = 6.1, pH = 7.4, ctCOa = 25.7 mmol/l, and CHCO3 = 24.5 mmol/l, the values are:
BH+(S, closed) = 2.7 mmol/l, and 3H'''(S, open) = 56.6 mmol/l. This illustrates the insigni-
ficance of the carbonic acid/bicarbonate buffer in a closed system at pH = 7.4, but the
great importance in an open system [2].

1. Plasma

The buffer value for H"*" in plasma can be expressed as the sum of the buffer values of
the bicarbonate buffer and the non-bicarbonate buffers. The change in the concentration of
added base can be expressed as the following sum (ignoring water, compare eq. (6)):

dcB (P) = dc?HC03(P) + dcX'(P) (14)

Dividing with dpH gives;

eH'^(P) = 3H"^(HC0: c p) + eH'^(X c P) 05)

In the following, X is used as a symbol for "nonbicarbonate buffer."

The buffer value of the nonbicarbonate buffers can be determined as the slope of the
titration curve, when titrating the plasma in the absence of CO2. The value for normal
plasma (pPr(P) = 70 g/1, wAlb(Pr) = 0.60, t = 310.15 K) determined by Siggaard-Andersen and
Rfirth (unpublished experiments), shows a maximum of 7.8 mmol/l at pH = 7.3, while the value
for pH = 7.40 is 7.7 mmol/l. The pH variation of 6 is shown in figure 1.

I I I I I I
I I I I
I I I I I I I I

I I

1 )8X(P) buffer value of non -bicarbonate buffers Human plasma (P)

(mmol/l) /)Pr(P) = 70 g/l

i^Alb(Pr) = 0 60
'80
= 37 'C

pH(P)
65

J \ L J \ L J \ L

Figure The buffer value for H due to nonbicarbonate buffers in the plasma (symbol bH
1.
(X c in the text) as a function of pH, in the absence of CO2 as well as at pC02'^45
p)
mmHg [2]. p = mass concentration, w = mass fraction. The maximum at pH 7.3 is probably
due to the imidazole groups of the albumin. The decrease in buffer value for pH > 7 in the
presence of CO2 is due to a change in pz value of the terminal amino groups from about 7.8
(Pr-NHg) to about 5 (Pr-NH-COOH).

13
 An alternative method for calculating the buffer value of the nonbicarbonate buffers
of plasma is based on CO2 titration. The reaction following CO2 titration is:

X' + + XH"^ + HCO3, 06)

where X' represents nonbicarbonate buffer base. To be more exact, HCO5 includes CO5 and
PrNHCOO" and is therefore designated tHCO^ = titratable bicarbonate. For varying pC02 we
therefore have:

/9ctHC03(P)\ /3cX'(P)\
(17)
9PH(P) \ 9PH(P)4'(P)
/,B.(p)

As a good approximation we have:

/ 3cX' (P)\ / 3cX' (P)

(18)
,3PH(P) V 3PH(P)/pC02
^.(p) = 0

The first coefficient refers to varying pC02(P) and the second to varying AcB'(P). The latter
value should be slightly higher than the former in the pH range about pH = 8 (due to buffering
by terminal amino groups), while the former should be slightly higher in the pH range about
pH = 5 (due to buffering by the carbamate groups, PrNHCOO").

The coefficient (actHCOs/apH) p, can be calculated from experimental values of

^'^
(slog pC02/3pH)^g.:

dlog pCO^ , detHCO^

= - (1 - dpZ/dpH) + ^— ' ^ . (19)
dpH 2.3 ctHCOg dpH

This equation is obtained by differentiating K = cxH^ ctHCOs/a pC02, taking a to be

• •

constant, (3log pC02(P)/3pH(P) )^gi read on the curve nomogram provided the data for the
buffer value of nonbicarbonate buffers which were plotted in figure 1 (dotted curve).

According to figure 1 the buffer value of the nonbicarbonate buffers of plasma varies
significantly with pH. As an approximation the value is generally taken to be constant in
the physiological pH range, independent of pC02 and pH, varying with the protein concentration
only:

BH^(Xcp) = pPrCP) • 0.11 mol/kg. (20)

For pPr(P) = 70 g/1 the buffer value therefore is 7.7 mmol/1.

Inorganic phosphate accounts for about 5 percent of the nonbicarbonate buffer value
in normal plasma at pH = 7.4 (calculated from an equation anaolgous to eq. (12) with a
total phosphate concentration of 1 mmol/1 and P^^^^ = 6.8).

Albumin is the major contributor to the buffer value of nonbicarbonate buffers of

plasma. The imidazoles (16 per albumin molecule) are the most important buffer groups in
the physiological pH range. Since the maximum buffer value in our experiments is at pH =
7.3, we assume that most of the imidazoles have dissociation constants of about lO-''-^.
With an albumin concentration of c?Alb(P) = 0.62 mmol/1 {-^p Alb(P) = 42 g/1) and hence an

14
imidazole concentration of 16 0,62=>9.9 mmol/l, the maximal buffer value of the imidazoles
•

at pH = 7.3 is calculated to be 0.576 9.9=>5.7 mmol/l, i.e., they account for 73 percent of
•

the nonbicarbonate buffer value of plasma of 7.8 mmol/l at pH = 7.3. The remainder, i.e., 22
percent => 1.7 mmol/l may be ascribed to the globulins, and with pGloCP) = 28 g/1 , the specific
buffer value of the globulins is calculated to be 0.06 mol/kg, while the specific buffer
value of the albumin is 0.14 mol/kg.

2. Erythrocyte Fluid

The buffer value of the non-bicarbonate buffers of erythrocyte fluid determined by

titrating erythrocyte fluid with strong base in the absence of CO2 is shown in figure 2.

^ I I I I
I
I I I I I I I I I I I

I
I I I I
"~|

/?(E)
Erythrolysate (E)
(mmol/l) ctDPG(E)
CtHb(E) = 21-0 mmol/l
I
— 80
14.2
(mmol/l)
pCO; = 0

^ = 37 °C

HbO

— 20

pHlE) -
6.5 7.0 7.5 8.0

^> L J I L J \ \ \ \ I \ \ \ \ L J I I »l

Figure 2. The buffer value for H ^in erythrocyte fluid (symbol 3H (E) in^the text) at pCOa =
0, i.e., the buffer value for H by non-bicarbonate buffers (symbol 3H (XcE) in the text),
as a function of pH at 37 °C [2]. Hb = deoxyhemoglobin, Hb02 = oxyhemoglobin. The dotted
curve represents the buffer value of oxygenated erythrocyte fluid with a three-fold increase
in 2,3-diphosphoglycerate concentration (ctDPG). The peak in the buffer value at pH = 7.1
indicates the px value of some of the acid-base groups of the DPG molecule. The difference
between the buffer value of oxygenated and deoxygenated erythrocyte fluid is due to the
fall in pz value of the so-called oxygen-linked acid-base groups following oxygenation.

The value varies with pH, with the concentration of hemoglobin, with the oxygenation of the
hemoglobin, and with the concentration of 2,3-diphosphoglycerate. The value for normal
erythrocyte fluid is about 63 mmol/l for pH = 7.2, cHb = 21 mmol/l, and ctDPG = 4 mmol/l
[2].

2,3-DPG accounts for about 7 to 8 percent of this value. This is calculated fron] pz =
7.1 of two of the phosphate groups of DPG (by means of an equation analogous to eq. (12))
^6H+(DPGcE)«4.6 mmol/l.

The remainder is mainly due to hemoglobin. The molar buffer value of the hemoglobin
therefore is about 58/21 = 2.8. Assuming that the principal buffer groups in the physiological
pH range are the imidazole groups, a maximum of about 6 imidazoles out of the 9.5 imidazoles
of the hemoglobin chains (10 for a chains, 9 for B chains) are engaged in buffering at pH «
7.2. The remainder may be engaged in salt bridge formation.

The difference between the buffer value of oxyhemoglobin and deoxyhemoglobin may be
ascribed to the change in pis: of the so-called oxygen-linked acid-base groups, i.e., acid-

15
base groups which participate in salt bridges in deoxyhemoglobin but not in oxyhemoglobin.
These salt bridges are the following, including those obtained by interchanging ai with a2
and ei with 32 [3]:

1) a-carboxyl of arginine HC3(141ai) - a-amino of valine NA2(la2)»

2) guanidinium of arginine HC3(141ai) - B-carboxyl of aspartate H9(126a2),

3) e-amino of lysine C5{40ai) - a-carboxyl of histidine HC3(14662)»

4) imidazole of histidine HC3(1466i) - 6-carboxyl of aspartate FGI(946i).

The salt bridge formation causes a rise in dk of the C-terminal imidazole groups (HC3(1463))
and of the N-terminal amino groups (NA2(la)) so that the buffer value of deoxyhemoglobin
becomes higher than that of oxygemoglobin for pH > 7.2, but lower for pH < 7.2. The pK
values are apparently shifted from the range of 7.8 to about 6.3. The greatest difference
between the molar buffer value for oxyhemoglobin and deoxyhemoglobin (about 0.4 at pH 7.3
and 6.3) is less than the maximal difference (0.576) which should be expected if one buffer
group per heme is shifted. The explanation may be that the pK values of the imidazoles and
the valines differ somewhat so that the effect is smoothed.

As a rough approximation, the buffer value of the nonbicarbonate buffers of erythrocyte

fluid may be taken to be independent of pH varying with the total hemoglobin concentration
only [2]:

BH"^(XcE) = 3.0 • cHb(E). (21)

More accurate data for the buffer value of the nonbicarbonate buffers of plasma and
erythrocyte fluid are necessary in order to improve the accuracy of the algorithms for
calculating the relationship among the acid-base variables of blood and plasma (see subse-
quent paper on acid-base algorithms).

3. Whole Blood

The buffer value is a quantity which may refer to an equilibrium system of several
phases, e.g. whole blood, provided we define the buffer value as:

8etH^(B)
bVO) = —-r , (22)
3yH^(B)

where p is the electrochemical potential: y = m + zFi, where F is the faraday constant

{F = 96487 C/mol), and $ is the inner electrical potential of the plasma. If the electrical
potential of the continuous phase, i.e., plasma, is the reference electrical potential, the
buffer value of whole blood in terms of pH is:

8eB'(B)
(23)
3H''(B)
3pH(P)

The buffer value of the nonbicarbonate buffers of whole blood can be expressed in
terms of the values for plasma and erythrocyte fluid. We have:

cX'(B) = (^P(B) . eX'(P) + <(>E(B) • cX'(E), (24)

where <j)P(B) + (})E(B) = 1. Differentiation of this equation with respect to pH, and rearrange-
ment (regarding (j)E(B) as constant) gives:

16
 dcX'(B) dcX'(P) /deX'(E) dpH(E) deX'(P)'
= + (|,E(B) ; (25)
dpH(P) dpH(P) \dpH(E) dpH(P) dpH(P)

SH'^CX c B) = eH"^CX c P) + <^E(B) •

CBH'^CX c E) • - eH'^(X CP)). (26)

The relationship between pH(P) and pH(E) is given by (see eq. (6) of the subsequent
paper) [4]:

pH(E) = 7.19 + 0.77 (pH(P) - 7.40), (27)

for oxygenated blood at 37 °C. Inserting dpH(E)/dpH(P) = 0.77 as well as BH'^(Xcp) =

7.7 mmol/1 and 3H+(XcE) = 63 mmol/1 (for cHb(E) = 21 mmol/1) into eq. (26) gives:

BH"^(XcB) = (7.7 + (^E(B) • 40.8) nimol/1, (28)

eH"^(XcB) = 1.94 • cHbCB) + 7.7 mmol/1. C29)

In analogy with eq. (17) we have:

/8ctHC0T(B)
(30)
eH (X^B) = -
9pH(P) /cB'(B)

where the concentration of bicarbonate refers to whole blood.

Often (3ctHC03(P)/3pH(P))^g, is used as an indicator of the buffer value of non-

bicarbonate buffers in the blood, where the concentration of bicarbonate refers to the
plasma phase. We have:

etHC03(B) = ctHC03(P)-cl)P(B) + etHC03(E) • <j)E(B)

= etHC03(P) -d-d - rc) • <).E(B)), (31)

where <t>P{Q) + (f>E(B) = 1 , and reHC0q(P|E) = etHC03(E)/ctHC03(P) . Regarding ())E(B) as constant,

differentiation with respect to pH(P) and rearrangement gives:

3dtHC0"(B) drc
+ (l)E(B) • etHCO"(P)-
8etHC03(P) 9pH(P) ^ dpH(P)
(32)
9pHCP) (1 - (1 -re) • 4.E(B)

By application of l/(l-x) = ^ + x + x + ... (for |£|< 1), we obtain:

/9ctHC0"(P)\ / .
dre \

"(~^1b.,B)
= ^ • Wi) •

(33,

(1 + (1-rc) • <j,E(B) + (d-rc) • <}.E(B))^ +

17
Inserting 6H^(X c B) from eq. (26) we obtain:

/ 3ctHC0:^(P)\ , o
. 2 = 6H (Xc P) + ^ . (^E(B) +A (1-re) •
{<^E{B)r
V 3pH(P) /cB'(B) , , ,
+ A . UE(B))2

(34)
where:

^ = eH^X e E) . - eH"(X e P) . ra -H etHCO-CP) •

^) (35)

We assume that the small diffusible ions are at equilibrium across the red cell membrane
and hence:

aCr(E)
ra = —
aH'^(P)
= =
aHCOZ(E)
^ . (36)
aH (E) aCr(P) aHC03(P)

Conversion of raHCOs to rcHCO^ = cHC03(E)/eHC03(P) requires a knowledge of the ratio of the

activity coefficients (y) and the mass concentrations of H2O (p):

rcHCO::(P/E) = raHCO;(P/E) •
yHCO"(P)
•
—
pHpO(E)
. (37)
yHC03(E) pH20(P)

Experimental data indicate that yC1"(E)/yC1"(P) = 0.93 [1,3]. Using this value for
YHC03(E)/YHC0i(P) and using pH20(P) = 0.94 kg/1, and pH20(E) =0.73 kg/1, we have:

rcHCO" = raHC03 • 0.835. (38)

From eq. (27) we obtain dpH(E)/dpH(P) = 0.77, and for pH = 7.40 we further obtain ra =
0.615 and dra/dpH(P) = -2.303 (1-0.77) ra = -0.326. Hence ra = 0.515 and dre/dpH(P) =
0.272. Inserting these values together with ctHC02(P) = 24.4 mmol/l into eq. (35), we
obtain: A = 37.9 mmol/l. As an approximation ignoring the variation with ctHCOatP) and
with pH(P) we therefore have:

/3cHC0Z(P)\ , , , , , , xv2
- JlJ-] / mmol/l = 7.7 + 37.9 <{.E B) + 18.4 (<J)E(B))^
\ 3pH(P /cB'(B) , , xx3 f',n\
^ ^
+8.9 (<^E(B))-^ + .... (39)

As a further approximation fitting this equation for <))E(B) = 0 and (j)E(B) = 0.45 and for
cHB(E) = 21 mmol/l we get [2]:

/9eHC0;(P)\
- ^ /(mmol/l)
'
= 7.7 + 48.4 (f>E(B) (40)
\ 3pH(P) /cB'(B)

= 7.7 + 2.3 eHb(B)/(mmol/l). (41)

This is the approximation which is often employed for the slope of the pH, ctHCOj equilibra-
tion curve of whole blood. The present derivation indicates where the principal approxima-
tions have been made: (1) assuming the buffer values of nonbicarbonate buffers of plasma
and
erythrocyte fluid to be independent of pH, (2) ignoring the variation with ctHCOa at constant

18
pH(P), (3) ignoring the variation with pH(P) at constant ctHCOs (P), and (4) assuming a
linear variation with 4)E(B).

References

[1] Van Slyke, D. D., On the measurement of buffer values and on the relationship of buffer
value to the dissociation constant of the buffer and reaction of the buffer solution, J.
Biol. Chem. 52, 525 (1922).

[2] Siggaard-Andersen, 0., The Acid-Base Status of the Bloody 229 pp. (Williams & Wilkins,
Baltimore, and Munksgaard, Copenhagen, 1974).

[3] Perutz, M. G., Stereochemistry of cooperative effects in haemoglobin, haem-haem inter-

action and the problem of allostery, the Bohr effect and combination with organic
phosphates. Nature (Lond. ) 228, 726 (1970).

[4] Funder, J. and Wieth, J. 0., Chloride and hydrogen ion distribution between human red
cells and plasma. Acta Physiol. Scand. 68, 235 (1966).

19
 National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

ACID-BASE ALGORITHMS

Ole Siggaard-Andersen
Department of Clinical Chemistry
University of Copenhagen
Copenhagen County Hospital
Herlev, Denmark

A knowledge of the relationships between the acid-base variables of the blood in

vitro is of considerable importance, because it is often required to calculate the re-
maining variables on the basis of measurement of only two of the variables. For these
calculations various nomograms have generally been employed, e.g., the Van Slyke and
Sendroy nomogram [1]^, McLean's Z-nomogram [2], the Singer and Hastings nomogram [3], and
others [4,5]. With the development of small programmable calculators, an arithmetic cal-
culation will probably in the future be more suitable. The problem will often be that one
wishes to calculate the base excess concentration of the blood, or perhaps of the average
extracellular fluid, on the basis of pH and pC02 values measured at 37 °C. Process diagrams
for two arithmetic algorithms for this purpose are shown in figures 1 and 2. The rectangu-
lar blocks are procedure symbols [6] and each of these will be described separately in the
following. The word algorithm (or algorism) is derived from al Kuwarizmi, a Persian
mathematician from the 8th century.

pC02(B) Henderson & cHC03(P)

^
Hasselbalch
equation
eq (1)

( \
PH(P) Van Slyke AcB'(B)
equation for
whole blood
eq (4)

eHb(B)

I Figure 1. Procedure diagram for calculation of the bicarbonate concentration

of the plasma, cHCOi(P) and the base excess concentration of blood or extra-
I

I
cellular fluid AcB'(B) or AcB'(Ecf) after measuring the blood pC02, the plasma
pH, and the hemoglobin concentration of the blood cHb(B). The Henderson-
Hasselbalch equation is employed for calculation of the plasma bicarbonate
concentration. The Van Slyke equation for whole blood is employed for
calculation of the base excess concentration. The various equations are
described in the text.
w
1

Figures in brackets indicate the literature references at the end of this paper.

21
 )

r
pCOaCB) Henderson and Hasselbalch eHCOi(P)
equation
eq \ \

pH(P) Van Slyke AcB' (P)

equation for
plasma
eq (2)

pH(P) - pH{E)
equation
eq (6)

cHCOiCE)
equation
eq (7)

Van Slyke
equation for
erythrocyte fluid
eq (3)

cHb(B) P-E-B AcB'CB)

equation
eq (5)

Figure 2. Alternative algorithm based on intermediary calculation of the pH

of the erythrocyte fluid (pH(E)) and the bicarbonate concentration of the
erythrocyte fluid (cHCOslE)). The Van Slyke equation for plasma respectively
erythrocyte fluid is used for calculation of the base excess concentration
of the plasma (acB'(B)) respectively erythrocyte fluid (acB'(E)). The
various equations are described in the text. Another almost identical
algorithm uses the Henderson-Hasselbalch equation for erythrocyte fluid
to calculate cHC03(E) from pH(E) and pCOzCB).

1. The Henderson-Hasselbalch Equation

The first equation of the algorithms is the Henderson-Hasselbalch equation, which can
be expressed as follows for human plasma at 37 °C:^

cHCO:
pH = pz + log (1)
aC02 •pC02

pK = 6.10.

aCOg = 0.0306 mmol/1 • mmHg (= 0.231 mmol/1 • kPa) .

Concerning terminology and symbols the reader is referred to the preceding paper on
definitions of acid-base quantities. HCO3 includes COi and carbamate throughout.

22
The value for pK varies slightly with the composition of the plasma, with a normal bio-
logical standard deviation of about 0.0015. In pathological cases with large changes in
the ionic strength the variation is considerably greater. Due to the incorporation of
carbonate and carbamate in the "bicarbonate" concentration an apparent variation of pK
with pH appears (decreasing pK with increasing pH), but for most practical purposes the
variation of pK can be ignored. The Henderson-Hasselbalch equation can be derived by
expressing the equality of the sum of the chemical potentials of CO2 and H2O and the sum
of the chemical potentials of and HCO3.

2. The Van Slyke Equation

The CO2 equilibration curve of the blood can, as a good approximation, be represented
by a straight line in a pH, log pC02 coordinate system, and this is employed in the nomo-
graphic algorithms. Another good approximation to a straight line is obtained by plotting
the CO2 equilibration line in a pH, CHCO3 coordinate system (Davenport diagram [7]). This
was originally utilized by Van Slyke [8], and I therefore find it appropriate to call the
equation for the pH, CHCO3 equilibration curve the Van Slyke equation. For plasma the Van
Slyke equation can be written:

AcHC03(P) = -bH"^(XcP) . ApH(P) + AcB'(P), (2)

where

AcHC03(P) = AcHC03(P°lP) = eHC03(P) - 24.4 mmol/1,

ApH (P) - ApH(P°|P) = pH(P) - 7.4,

AeB'(P) = AcB'(P°|P) = cB'(P) - eB'(P°).

P° is plasma titrated with strong acid or base to pH(P) = 7.40 at pCOa = 5.33 kPa and T =
310.15 K ( ^cHCOg = 24.4 mmol/1).

3H'^(X c p) is the buffer value for H"*" due to non-bicarbonate buffers in the plasma.
According to eq (20) of the preceding paper it can be expressed as:

eH'^(XcP) = pPr(P) • 0.11 mol/kg.

i.e. J, for normal plasma with pPr(P) = 70 g/1 ^ bH'^(X c p) = 7,7 mmol/1.

The Van Slyke equation is derived from:

AcB'(P) = AeHC03(P) + AcX'(P),

and for pH(P) = 7.40 we have AeX'(P) = 0 (compare eq (14) of the preceding paper).

For erythrocyte fluid the Van Slyke equation is analogous:

AeHC03(E) = - 3H'^(X c E) x ApH(E) + AeB'(E), (3)

where E° is erythrocyte fluid in equilibrium with P°, i.e., pH(E°) = 7.19 for pC02 = 5.33
kPa and 37 °C => eHC03(E°) = 12.6 mmol/1. The buffer value for H"*" due to the non-bicarbonate
buffers is

3H"^(X c E) = 3.0 • cHb(E),

23
i.e., for normal erythrocyte fluid with cHB(E) = 21 mmo1/l eH'''(X c e) = 63 mmol/1 (see eq
(21) of the preceding paper).

For whole blood the Van Slyke equation expressed in terms of the plasma bicarbonate
concentration is:

AeHC03(P) = - B-ApH(P) + AeB'(B)/Z (4)

where B = (3cHC03(P)/3pH(P))^^g, ,gv can be expressed by the following approximation (eq

(41) of the preceding paper):

B = 2.3 • eHb(B) + 7.7 mmol/1.

z is derived from eq (31) of the preceding paper:

pH(P) = 7.40

AcB'(B) = AcHC03(B)

= AeHC03(P) • (1 - (1 - re) • (|)E(B)),

i.e. Z = (1 - (1 - rc) • (t)E(B),

where rc = eHC03(E)/eHC03(P) = 0.515 (for pH(P) = 7.40). For eHB(E) = 21 mmol/1 we therefore
get:

Z = 1 - 0.023 • eHb(B)/(mmol/l).

The relationship between the base excess concentration of plasma, erythrocyte fluid,
and whole blood is the following:

AeB'(B) = P(B) • AeB'(P) + <j)E(B) • AcB'(E) (5)

= (1 - (f>E(B) • AeB'(P) + <|)E(B) • AeB'(E).

For eHb(E) = 21 mmol/1 (^E(B3 = cHB(B)/(21 mmol/1).

3. pH(E) - pH(P) Equation

The relationship between pH(E) and pH(P) measured in normal blood can be expressed as
follows (eq (27) of the preceding paper):

pH(E) = 7.19 + 0.77 • (pH(P) - 7.40). (6)

The coefficient 0.77 is due to a pH variation of the concentration of "non-diffusible"

ions {e.g., protein anions) in plasma and erythrocyte fluid. The coefficient can be
calculated approximately from the buffer value for H"*" due to non-bicarbonate buffers in
plasma and erythrocyte fluid, assuming the erythrocyte membrane is impermeable to cations
for short term changes.

24
 )

The value 7.19 is dependent on the membrane potential across the erythrocyte membrane
(normally approximately -13.0 mV, inside negative) which is due to the active transport of
Na"*" out of the erythrocytes. Inhibition of the Na"*" pump causes a diminution in the membrane
potential and a swelling of the erythrocytes (hemolysis). Increased activity of the Na'^
pump causes an increase in the membrane potential and therefore a lower pH(E) for pH(P) =
7.40.
According to eq (38) of the preceding paper we have:

re = ra • 0.835,

where re = cHC0;(E)/eHC03(P) , and ra = aHC0;(E)/aHC03(P) = aH'^(P)/aH"^(E) . By combination of

this equation and eq (6) we get:

cHC03(E) = cHC03(P) • 0.835 • antilog (1.492 - 0.23 pH(P)) (7)

4. Comparison of the Van Slyke Equation and the Alignment Nomogram

The nomographic algorithm is based on a linear representation of the CO2 equilibration

curves of the blood in a pH, log pC02 coordinate system [3,4]. This algorithm is more
difficult to express in a simple arithmetic form because the slope of the equilibration
lines varies with both the_hemoglobin concentration and the base excess concentration of
the sample. In a pH, CHCO3 coordinate system the variation of the slope with the base
excess concentration (or with pH) is small. Both algorithms are approximations. A comparison
of the results obtained with the arithmetic algorithms and the nomographic algorithm
(Alignment Nomogram) [4] is shown in table 1. The table includes values for the calculated

Table 1. The base excess concentration of the blood and the extracellular
fluid calculated (A) by means of the alignment nomogram [4],
(B) by means of the Van Slyke equation for whole blood (eq (4)),
and (C) by means of the Van Slyke equations for plasma and erythrocyte
fluid (eq (2) and (3)) together with the relationship between pH(P)
and pH(E) (eq (6)). The calculations are based on a hemoglobin
concentration of: cHb(b) = 9.0 mmol/1, and eHb(Ecf) = 3.6 mmol/1.

Measured Calculated
pCOp(B) AcB'(B) AeB'(Ecf)
pH(P)
(mmflg) (mmol/1) (mmol/1

Acute hypercapnia 7.09 100 A -4.6 ±0.0

B -2.6 +0.5
C -2.3 +1.0
Acute hypocapnia 7.70 16 A +0.5 -2.8
B +2.9 -0.1
C +3.0 -0.3
Chronic hypercapnia 7.30 80 A +8.2 +11.7
B +9.2 +11.7
C +9.4 +11.9
Chronic hypocapnia 7.44 25 A -5.2 -6.5
B -5.2 -6.4
C -5.1 -6.5
Acute base deficit 6.90 30 A -27.6 -25.0
B -26.0 -24.4
C -27.0 -24.3
(Chronic) base deficit 7.15 10 A -25.8 -24.6
B -22.3 -22.9
C -22.8 -22.9
(Chronic) base excess 7.60 55 A +26.0 +29.0
B +27.4 +29.4
C +26.7 +28.9

25
excess concentration of base in the average extracellular fluid (including blood). This
value is calculated on the assumption that the buffer value for H"*" due to non-bicarbonate
buffers in the extracellular fluid is mainly due to hemoglobin, and that the hemoglobin
concentration of the total extracellular fluid (including blood) is approximately 3.6
mmol/1 [4]. The advantage of this quantity is that it remains virtually constant during
acute changes in the pC02(B) in vivo, whereas the base excess concentration of blood and
plasma varies in opposite directions due to a redistribution of H"*" between phases of
different buffer value.

The results indicate that the base excess concentration of the blood or extracellular
fluid can be calculated with sufficient accuracy for all clinical purposes by means of the
arithmetic algorithms. In spite of the approximations made in the derivation of the Van
Slyke equation for whole blood where the variation of B with cHC03(P) at constant pH(P) is
ignored, there seems to be no reason to use the more complicated algorithm based on the
empirical relationship between pH(P) and pH(E). The major approximation undoubtedly lies
in the use of a constant buffer value for non-bicarbonate buffers in plasma and erythrocyte
fluid irrespective of a considerable pH variation.

When the pC02 value falls in the range between 25 and 80 mmHg, the alignment nomogram
is likely to provide the best approximation, because it was constructed from experimental
data in the pCOa range. When the pH is around 7.3 to 7.5 but the pC02 is very high or
low, the Van Slyke equation is likely to provide the best approximation, because the non-
bicarbonate buffer values is most accurate for this pH range.

References

[1] Van Slyke, D. D. and Sendrby, J., Jr., Studies of gas and electrolyte equilibria in
blood. XV. Line charts for graphic calculations by the Henderson-Hasselbalch
equation, and for calculating plasma carbon dioxide content from whole blood con-
tent, J. Biol. Chem. 79_, 781 (1928).

[2] McLean, F. C, Application of the law of chemical equilibrium (law of mass action) to
biological problems, Physiol. Rev. 18, 495 (1938).

[3] Singer, R. B. and Hastings, A. B. , An improved method for the estimation of distur-
bances of the acid-base balance of human blood. Medicine (Baltimore) 2J_, 223 (1948).

[4] Siggaard-Ander^en, 0., The Acid-Base Status of the Blood, 229 pp. (Williams and Wilkins,
Baltimore and Munksgaard, Copenhagen, 1974).

[5] Thews, G. (ed.), Nomogramme zum S'dure-Basen-Status des Blutes und sum Atemgastransport,
134 pp. (Springer-Vorlag, Berlin, 1971).

[6] International Standard ISO 1028, Flowchart Symbols for Information Processing,
International Organization for Standardization, Geneva, Switzerland (available from
the different National Standards Institutes).

[7] Davenport, H. W., The ABC of Acid-Base Chemistry, 119 pp. (University of Chicago Press,
Chicago, 1969).

[8] Van Slyke, D. D., Studies of acidosis: XVII. The normal and abnormal variations in
the acid-base balance of the blood, J. Biol. chem. 48, 153 (1921).

26
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

DETERMINATION OF TOTAL CO2 CONCENTRATION IN BLOOD OR PLASMA

P. Rispens, E. J. van Kampen, and W. G. Zijlstra

Laboratory of Chemical Physiology
University of Groningen
Groningen, The Netherlands
and
Laboratory of Clinical Chemistry
^ Diakonessenhuis
Groningen, The Netherlands

Carbon dioxide is present in blood as dissolved CO2, carbonic acid (H2CO3), bicarbonate
(HCO3), carbonate (C0|~) and carbamate, i.e., CO2 bound to free aminogroups of proteins
(RNHCOO"). The quantity total CO2 concentration (c^q ) is defined as the sum of the concen-
trations of all forms in which CO2 is present. In most methods [1]^ for the determination of

in blood or plasma (c^q and c^q , respectively), blood or plasma is added to an acid
reagent for conversion of bound CO2 (HCO3, CO3" and RNHCOO") into free CO2 {H2CO3 and dis-
solved CO2). As a measure for c^^q^ may then be used

1) the total amount of CO2 which can be extracted from the mixture of sample and
reagent (extraction methods);

2) the rise in P^q in a given volume of gas which is equilibrated with the sample-
reagent mixture (equilibration methods);

3) the rise in P^q^ in the sample-reagent mixture when escape of any CO2 is prevented.

In the method developed in our laboratory (the cediometer), the equilibration principle is
applied [2,3].

1. Extraction Methods

Classic extraction methods are the volumetric [4] and the manometric Van Slyke techniques
[5,6]. More recently gas chromatographic [7,8] or mass spectrometric [9] techniques have
been applied to determine the total amount of CO2 extracted from the sample-reagent mixture.
The main problem in applying these techniques is to ascertain that all CO2 has indeed been
extracted from the sample-reagent mixture. Under favorable conditions, highly reproducible
results can be obtained (coefficient of variation s=l percent).

The absolute accuracy depends on the reliability of the calibration procedure. When
gas chromatographic or mass spectrometric techniques are used, Na2C03 or NaHC03 reference

^Figures in brackets indicate the literature references at the end of this paper.

27
solutions are used for calibration. The accuracy then obviously depends on the reliability
of the reference solutions.

The volumetric or manometric techniques are usually considered as absolute methods,

i.e., methods in which calibration with reference solutions is not required. The accuracy
depends on the accuracy of the volume or pressure measurements. In the volumetric technique,

CO2 is extracted by shaking and vacuum exposure. The volume of the CO2 set free is measured
at barometric pressure. Corrections are necessary for other gases (O2, N2) set free from the
sample-reagent mixture. Because of the uncertainty of these corrections the accuracy of the
rather simple volumetric technique is dubious.

The more laborious manometric technique of Van Slyke and Neill [5] is usually considered
as the most accurate method for determining c^q . In this method, CO2 is also extracted from
the sample-reagent mixture by shaking and vacuum exposure. After all gas has been extracted,
it is reduced to a known volume v and the pressure is determined. CO2 is then absorbed using
a concentrated NaOH solution and the pressure of the remaining gas is again measured at
volume 7. The pressure difference before and after absorption of CO2 corresponds to the
pressure in the volume F of the CO2 set free from the sample, c^q of the sample is cal-
culated by multiplying the pressure difference with a factor depending on F, the sample
volume, the acid reagent volume and the temperature. This factor, which can be obtained
from reference [10], contains a correction for the amount of CO2 redissolving in the sample-
reagent mixture while the gas is reduced to volume 7. The correction has been determined
using Na2C03 reference solutions. However, there is some doubt about the magnitude of this
correction, because the amount of CO2 that redissolves depends upon the apparatus used and
upon the skill and speed of the operator. With utmost care and skill, highly reproducible
c^Q determinations are possible (coefficient of variation 0.6 percent). However, because

of the need for Na2C03 reference solutions for finding the correction for CO2 redissolving in
the sample-reagent mixture, the manometric Van Slyke technique is no absolute method either.

2. Equilibration Methods

For the equilibration of a fixed gas volumq with a fixed volume of blood or plasma
after the addition to acid reagent, a closed system containing a circulating pump is usually
employed. For the determination of the rise in v^^ of the gas mixture, photometric [11],

potentiometric [2,3], gas chromatographic [12] and mass spectrometric techniques, as well as
infrared gas analysis [13] can be used. Under favorable conditions highly reproducible
results can be obtained (coefficient of variation *1 percent). Na2C03 or NaHCOs reference

solutions are used for calibration.

In the cediometer, equilibration of the sample-reagent mixture with a fixed gas volume
is performed in a closed system consisting of a measuring chamber filled with NaHCOs-NaCl

solution, a sample chamber filled with acid reagent, and tubing connecting measuring chamber

28
 Figure 1. Schematic diagram of the closed gas-liquid circuit of the
cediometer. E = measuring chamber; S = sample chamber; T = gas tubing;
A = CO2 absorber; C = cock; P = pump.

and sample chamber (fig. 1). The gas mixture is driven through the system by means of a
pump. The rise in p^q in the gas mixture caused by adding a constant volume of blood or

plasma to the acid reagent, is determined by the simultaneous equilibration of the gas
mixture with the HCOs-NaCl solution and measuring the ensuing change of pH (ApH^) by means

of a combined glass-Ag/AgCl electrode. ApH^ is a measure of the total CO2 concentration

of the sample injected into the acid reagent. The method is easy to perform; the time
necessary for a determination is about 3 minutes. The coefficient of variation is 0.8
percent.

3. Methods Based on -Rise in p^q of Sample-Reagent Mixture

For the determination of the rise in p^q of the sample-reagent mixture, a CO2 electrode
can be used. The method is quite simple and rather reproducible results have been reported,
the coefficient of variation being «1.2 percent [14]. Na2C03 or NaHCOs reference solutions
are used for calibration.

4. Discussion

In spite of the simplicity of many available methods for determining e , common

practice in most laboratories is to measure pH and p^q and to calculate
using the Henderson-Hasselbalch equation. Yet the direct determination of e has some
CO;
distinct advantages over that of P-^ .
LU9

29
 ,

1) Calibration can easily be performed with NaaCOs or NaHCOs reference solutions, which
can be prepared from commercially available analytical grade reagents.

2) In contrast to P^q , c^q is inde-pendent of temperature. Therefore, mistakes due to

incorrect thermos tating of the measuring system are excluded.

3) c^Q^ is faj> less influenoed by metabolism of anaerobical ly stored blood than P^q .

4) c^Q is less influenoed by escape of CO2 from the sample than P^q .

5) The instrumentation required for measuring c^^ is less vulnerable than membrane
covered CO2 electrodes are.

Because of these advantages, the determination of c^q should be recommended, especially

in smaller laboratories where some time passes between sampling and analysis or between the

determination of pH and p^q .

References

[I] Rispens, P. and Zijlstra, W. G., Blood gas transport and acid-base balance, in Clinical
Biochemistry, Principles and Methods, H. CH. Curtius, and M. Roth, eds., p. 1604 (Walter
de Gruyter, Berlin, New York, 1974).

[2] Rispens, P., Van Assendelft, 0. W., Brunsting, J. R., Zijlstra, W. G., and Van Kampen,
E. J., A direct method for the determination of the HCO3 concentration as total carbon
dioxide in blood and plasma, Clin. Chim. Acta, 1_4, 760 (1966).

[3] Rispens, P., Brunsting, J. R. , Zijlstra, W. G., and Van Kampen, E. J., Determination of
total carbon dioxide in blood and plasma by means of the cediometer. Theory and experi-
mental verification, Clin. Chim. Acta, 22_, 261 (1968).

[4] Van Slyke, D. D. Studies of acidosis.

, II. A method for the determination of carbon
dioxide and carbonates in solutions, J. Biol. Chem. 30, 347 (1917).

[5] Van Slyke, D. D. and Neill, J. M., Determination of gases in blood and other solutions
by vacuum extraction and manometric measurement, J. Biol. Chem. 6J_, 523 (1924).

[6] Natelson, S., Routine use of ultramicro methods in the clinical laboratory, Amer. J.
Clin. Path. 21_, 1153 (1951 ).

[7] Ramsey, L. H., Analysis of gas in biological fluids by gas chromatography. Science,
129 900 (1959).
,

[8] Wilson, R. H., Jay, B. E., Doty, V., Pingree, H., and Higgens, E., Analysis of blood
gases with gas absorption chromatographic technique, J. Appl. Physiol. 1_6, 374 (1961 ).

[9] Lotz, P., Dahners, H., and Pichotka, J. P., Massenspektrometrische Bestimmung des O2
und C02-Gehaltes von Blut, Pflugers Arch. 3]5_, 86 (1970).

[10] Bartels, H., Biicherl E. Hertz, C. W., Rodewald, G., and Schwab, M.
, ,

Lungenfunktionsprufungen, Methoden und Beispiele klinischer Anwendung, p. 219

(Springer Verlag, Berlin, 1959).

[II] Zijlstra, W. G., Brunsting, J. R., Dorlas, J. C, and Rispens, P., Eine Methode zur
fortlaufenden COa-Messung in der Atemluft und zur schnellen Bestimmung des
Gesamtkohlensauregehaltes von Blut und Plasma, Anaesthetist, 13^, 200 (1964).

30
[12] Gimeno, F. 0. and Tammeling, G. J., A recirculation system for the determination of
blood gases by chromatography, J. Appl. Physiol. 24, 119 (1968).

[13] Gimeno, F. 0., Orie, S. A. M. , and Tammeling, G. J., Determination of carbon dioxide
content by infrared analysis, J. Appl. Physiol. 21_, 1377 (1966).

[14] Reyes, R. J. and Neville, 0. R., A rapid electrochemical technique for measuring the
total carbon dioxide content of blood, Clin. Chem. 14^, 637 (1968).

31
I

i
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

THE APPARENT OVERALL FIRST DISSOCIATION CONSTANT OF CO2 IN PLASMA

P. Rispens and W. G. Zijlstra

Laboratory of Chemical Physiology
University of Groningen
Groningen, The Netherlands

Since Hasselbalch [1]^ in 1916 put the equilibrium equation for the dissociation of
carbonic acid in the logarithmic form, many investigations have been devoted to the de-
termination of the apparent overall first dissociation constant of CO2 in plasma (table 1).

''HCO"
pH = pz] + log (1)
^ ^COa

p^i denotes the negative logarithm of the apparent overall first dissociation constant of
CO2, ^HCOs concentration of HCO3 (mmol-l'M, S the solubility coefficient^ of CO2
(mmol 'l"^' (mm Hg)"^) and P^^ the CO2 tension (mm Hg). Introduction of any new method for
UU2 ,

the determination of pH, cuCOg in redetermination of p^i.

^COa ''"variably resulted
'^^ It

proved that the value of pz{ depended on the analytical methods. To stress the experimental

nature of the constant, the symbol p^| was introduced, denoting the value calculated for
pK\ using eq. (1) when experimentally found values for pH, i^^^q-, P^q and S are used. In

the p^i determinations of the twenties and early thirties [2-7], P^g was calculated from

Ppri of the gas mixture with which the plasma had been equilibrated. pH was measured using
LU2
hydrogen electrodes. <2^qq- was calculated from the total CO2 concentration c^q as measured

with the manometric Van Slyke technique, assuming c^^ to be equal to the sum of c^^q-

and S'P^Q 'S had been determined by experiments in which acidified plasma or serum was
equilibrated with gas mixtures of known P^q and c^q measured manometrical ly . A mean value

of 6.10 was found for p^i at body temperature with no appreciable variation, even when
abnormal plasma was used.

However, with the introduction of newer methods for pH and

ix,
Pm
UU2
measurement, doubt arose
about the constancy of pzj. The introduction of glass electrodes for the determination of
pH resulted in reports in which a decrease of the value of pzj with increasing pH was

stated, Api^l/ApH ranging from -0.03 to -0.06 [8-10,12]. The introduction of CO2 electrodes
for the direct measurement of p^q resulted in investigations [13,14] in which appreciable
variations of pxj in seriously ill patients were measured (from 5.96 to 6.27). The tem-

^Figures in brackets indicate literature references at the end of this paper.

^ ' '
^^C02 dissolved ^H2C03^ ^^C02^

33
 — ^ — i — 1 — — 1 —— 11 i 1

J2 CO
--I 0 CD CO CO
0 CO CM 00 CM ==1-
0
II

c?^
Q.
II

n:
CM CM
0 CD
0 CO
0 CD
0 0
+-) a. • • •
<x> tX) to to to to to to to to to

•(->

o
\ 0
CO ^ CO
>vl- LO 0
II

0 0 to
0 0 LO
0 0
0^
CL CL 0 00 1 1
00 1 1
0 1

-(->
in
0 (J T3
CO CO
OJ M-
CO CD CM
0 0
0 0 to CO CO CO
0
c 0 0 000 00
0 0 0
0
<1 1 00 1 1
c0 1 1
0
_ <^
.

00 '

<^ 0^ to CM — CM CO
0— CO to
f

1— CM !— CM 0
0 01 10 lO <x; >X) to tc to to to to to to to to 0 to to
'r-
-l->
n3
0;
1

01

CM
CO
1

VX>
CO
1

r—
CM
1

"^1-

CD
1 1 1

LO CM
1 1 1

00 CM
1

CO
0
1

M
ro
c: 0 t— 0 0 0 1— 0 0 0 0 r— 0 1— f—
00
0 CM
—
'r-
£ to to to to to to to to to to to to
S.
<u «^
-t->

0
(U
LO CO ^X) 0^ 1
0 CM to CD CM CO CO — CM
II
1
— CM
0 C\l to
0 1—
'Jd" c^ 0^
0 CO LO
0 f—
i

CTl
0 1
0 <X3
i

• *

0. Q- to to to to to to to to to to to to to to
</)

0
•p-
00 LO
CM CM
LO CO CO 00
r—
to LO
CM
CO to CO
CM 1—
to to
1— 1—
CM 0
to
CM
>
(U
CM
>^i-
—
1

s- in CO ,

0 0
1 1

Q. p— CO LO to LO
«4-
00 1X1 to to Ln LO LO r- 00 to to 00 00
«
0 06 00 1^ 00 CO CO
(U 1 1 1 1 1 1 1 1 1 1 1 1 I 1

a> LO
1/)
4-> CO p~ CM to — LO CM LO —
0 (
,

3
Q. 10 "d-CO CO CM CO CM 00 CO '^f CM
•r—
E
CO IX) IX) 1^ r-- r-- 1^ to to to to to -P
O) to
q: 3
LO cC
Q.
E
<U
O
o
CO C_)
CO CM
CO
CO
rr\
CD
00 f—
CO CM
CO
CO
00
CO CO CO CM
CO LO
CO CM
CO
CO
00
CO
CO LO
CO CM
>>
J3

— E
1
<u
fo
• ' '
_ >
3
to ' ro i_
0 ^ , }
<u
-E m (/) to
4-* ^ _ _ ^ ro ro
<o 0
'

0 -Q
O- tn U)
E ro 0)
C/)

X
^E E
^
^ 'O
O-
'J LJ ~1
Si
ro ro
3
cz
0 E rt3 E <

3
<
cz
o>
J-:

ro (/)
to
E >
ro

S-
"O
E
S-
<u
S-
<u
B ^ t.^
E S- E E
rtS 3
S-
O)
0
-0
T5
E 1
ro B
(/>
ro 0
ro to in S- 1— I/) S- to ro CL ro Q. +J
4-> <D (0 I— <U ro
<U OJ <u lO 0 10 E ro E to E E E Q. E E CD
10 to to E E ro ro ro ro ro •r— E
s- S- S- CD CD s- E S- CD E E E E CD E E •1—
0 0 0 0 0 0 3 0 0 3 3 3 3 0 3 S- to
3: n: Q r— z: -o ^ in 0 3Z <U
+J
3
0) T3
<U
1 1 1 E +J
0
1

1
—
CO
to 0 r—
ro

3
G s_
a 1 1 CU s-
TD r—
(/)
3 E
<u
XI 00 o
^—
ro >>^
Q) ro 1
=1 E 0
CU 4^ « CD
-0
3 0) to
to E CDi 1
<i) E s-
CD >> CD 0 E i- II II II II

S- E to ro II

Z3 E Z. ro
— zc
-Q 4-> CDl CU E Q. a. CL o. Q.
S- to > cno
3 3 0 —
O ro
Q 00 00 L—
1 1 =1 3 «u 4-

34
perature dependence ApKi/^T, usually inferred from measurements of but two temperatures, was
reported to vary between -0.0032 and -0.0063 pZ] unit-(°C)"^.

After having developed a new method for the determination of c^q in blood or plasma

[15] and a tonometry set-up [16] allowing blood or plasma to be equilibrated with gas
mixtures of known composition, we decided to redetermine pK'i over the temperature range of
16-42 °C and the pH range of 6.8-7.8. Plasma of healthy volunteers was equilibrated in
horizontally rotating tonometers at 16, 20, 26, 30, 32.5, 35, 37.5 and 42.5 °C with CO2/O2
gas mixtures from cylinders, F^q ranging from 2-16 percent. F^q was measured with a
Haldane gas analysis apparatus, the standard deviation calculated from duplicate deter-
minations being 0.04 percent. After equilibrium, two 5 ml samples were taken from each
tonometer. Plasma pH was measured with capillary microelectrodes (Radiometer E5021a)
connected to a direct reading pH meter (Radiometer PHM27) or a balancing pH meter (Radio-
meter PHM4). Level and slope of the pH measuring system were controlled with two NBS

buffers, the slope daily, the level before and after each pH determination. In each sample

pH was measured in duplicate; the standard deviation was 0.006 pH unit. Plasma c^q was
measured in each sample in duplicate using a cediometer; the coefficient of variation was
0.8 percent, pi^l was calculated as

"
''CO2 '^'^C02
^
mi = pH - log ^
^ ^C02
. (2)

For S the values given by Austin et al. [17] were used.

A total of 555 duplicate determinations of p^{ in plasma equilibrated at known tem-

perature with gas mixtures of known F^^ were performed (table 2). The standard deviation
as calculated from the duplicate determinations was 0.008. The best approximation of the
relationship between p^{, T and pH was found to be [18].

p^i = .4.7416 .
MMil- + 0.015906 T - log (l +
^^?^) •
(3)

The mean difference between the experimental p^{ values and those from eq. (3) was

0.0000 ± 0.0138 (SD). Figure 1 shows a curve relating p^i to temperature as calculated with
eq. (3) at pH = 7.4. The curve is in good agreement with the experimental data presented,
as well as with those reported by other investigators. Figure 1 also demonstrates that the
estimation of the relationship between p^i and temperature on the basis of p^j determinations
at but two temperatures is subject to substantial error, as p^| does not vary linearly with
temperature.

The decrease of p^i with increasing pH reported by other investigators has been confirmed
in the present investigation and is represented by the term -log(l + 0. 020682/ lO'P^"*"^) in eq.

(3). The decrease of p^{ with increasing pH has been ascribed to a pH dependent error in the
determination of Cur-n.- [10], to an alkaline error of glass electrodes [9] or to both.

35
 .., D

Table 2. Experimental and calculated p^l at different temperatures and pH.

Temp. pH range pH mean P^i (exp.) PK{ (calc.) Difference SD of

°C mean difference

6. 8-7.3 7. 167 6.096 -0.,005 0.,0068

7 3-7 ^ 7 fi 0Q7 6. 089 +n uuo
,
n
U .
m nR
7 5-7 8 7, 542 6.083 6. 079 n nn'?n
-7
37 5 6. 8-7.3 7. 181 U
fi a 113
1 0 1 6. 110 +0.,003 0,,0124
fn = 13fi^ 7 o— / . o 7 d?2 U a 1 U 1 6. 100 +n n
7 1-7 ft 7 6.090 6. 088 , uu^ n u ,
m1
7ft
/ 0

3R n 6. 8-7.3 7. 162 fi nn 6. 119 -0.,009 0,,0102

fn = 33^ 7 0— / • 3 7 HDD fi 10"^ 6. 107 -0.,004 u u , 1 1

/ a R-7
7 ft 7 6.087 6. 097 -0,,010 n U Oo,
m1

3? R 6. 8-7. 3 7. 191 fi HQ 6. 127 -0.,008 0,.0166

fn = 3'i') 7 3-7 5 7 457 fi 113 6. 116 -0,,003 n nnfiQ
uuuy,

7 5-7 8 7 fi33 6.103 6. 104 -0.,001 n

u U Cc.
. 1

O *7 O "7
30 n 6. 8-7.3 7. 149 fi 1d2 6. 138 +0.004 0,.0131
fn = 74^ 7 3-7 ^ 7 fi 1?Q 6. 128 +0.,001 n U OH. 1

7 5-7 ft 7 6.118 6. 111 +0.,007 n

6, 8-7.3 7. 131 fi 1 Rfi 6. 156 0,.0114

fn = 7 0— / • 0 7 fi Ml 6. 145 -0,,004 n
u nnoft
7 "1-7 ft 7 6.128 6. 132 -0.,004 n 2 nn
U • o / » o 7 1 C.\J fi IQl 6. 188 +0,.003 n
fn = Q?^ 7. 3-7.5 7. 387 fi 17Q 6. 178 +0,,001 0,.0126
7. 5-7.8 7. 616 6.162 6. 164 -0.,002 0,.0127

16.0 6. 8-7.3 7. 050 6.212 6. 212 0,.0167

(n = 85) 7. 3-7.5 7. 382 6.202 6. 200 +0.,002 0,.0139
7. 5-7.8 7. 561 6.185 6. 190 -0,,005 0,.0164

Obviously, as o^qq- is calculated as (c^q - S-P^^ ) or determined by titration, C0|- and

RNHCOO" present in plasma are included. The value used for Cj^^q- in calculating p^j will be
too high and pk'i consequently lower than pK[. As the fraction of bound CO2 present as COl"
or as RNHCOO" increases with increasing pH, the error increases with increasing pH. However,

g^q2- and ^RNHCOO- ai^e too small even at higher pH to completely explain the observed
decrease of pk'i with increasing pH. An alkaline error of the glass electrode might in part
be responsible.

The scatter of the pK'i values of table 1 and of the present investigation is much
smaller than that reported for pK'i in disease when measuring Pp-, and pH in whole blood,
using membrane covered electrodes for the determination of Pm • Some changes of pZi
(and thus of pKi) may be expected to result from changes in the ionic strength of plasma.

However, the changes of the ionic strength of plasma in disease are too small to be
responsible. Because of the unequal distribution of dissolved CO2 and HCO3 over the water
and non-water phase of plasma, the value of pK'i found in plasma differs slightly from that
in simple electrolyte solutions of the same ionic strength. Changes of the distribution
ratio of CO2 and HCO3 during disease might cause a change of pK'i. However, these changes
have been calculated to be small [19]. Variations of pK[ in disease must therefore largely
be ascribed to unknown errors in the determination of pH, Pp^, or Curn' abnormal plasma.

36
 _J I I
'

^0 35 30 25 20

Figure 1. Relationship between the experimentally determined apparent first dissociation

constant of carbonic acid in human plasma and temperature. The curve represents eq. (3)
at pH = 7.4

• present investigation o Warburg [2]

e Cullen et at. [5] x Severinghaus et at. [9]
+ Siggaard-Andersen [10] a Maas [12]

The pifj values taken from the literature have been recalculated, using for s the
values reported by Austin et al. [17].

References

[1] Hasselbalch, K. A., Die Berechnung der Wasserstoffzahl des Blutes aus der freien und
gebundenen Kohlensaure desselben, und die Sauerstoffbindung des Blutes als Funktion
der Wasserstoffzahl, Biochem. Z. 78, 112 (1916).

[2] Warburg, E. J., Carbonic acid compounds and hydrogen ion activities in blood and salt
solutions, Bioohem. J. 16, 153 (1922).

[3] Cullen, G. E., Studies of acidosis. XIX. The colorimetric determination of the
hydrogen ion concentration of blood plasma, J. Biol. Chem. 52, 501 (1922).

[4] Van Slyke, D. D., Hastings, A. B., Murray, C. D., and Sendroy, J., Studies of gas and
electrolyte equilibria in blood. VIII. The distribution of hydrogen, chloride and'
bicarbonate ions in oxygenated and reduced blood, J. Biol. Chem. 65, 301 (1925).

[5] Cullen, G. E., Keeler, H. R., and Robinson, H. W., The pZ' of the Henderson-Hassel balch
equation for hydrogen concentration of serum, J. Biol. Chem. 66, 301 (1925).

[6] Hastings, A. B., Sendroy, J., and Van Slyke, D. D., Studies of gas and electrolyte
equilibria in blood. XII. The value of \>k[ in the Henderson-Hassel balch equation
for blood serum, J. Biol. Chem. 79, 183 (1928).

[7] Robinson, H. W., Price, J. W., and Cullen, G. E., The value of pZ' in the Henderson-
Hasselbalch equation for human and dog sera, determined with the Simms electrode, J.
Biol. Chem. 106, 7 (1934).

37
8] Dill, D. B., Daly, C, and Forbes, W. B., The pK{ of serum and red cells, J. Biol.
Chem. m, 569 (1937).

9] Severinghaus, J. W., Stupfel M., and Bradley, A. F., Variations of serum carbonic
,

acid pz' with pH and temperature, J. Appl. Physiol. 9^, 197 (1956).

10] Siggaard-Andersen, 0., The first dissociation exponent of carbonic acid as a function
of pH, Sound. J. Clin. Lab. Invest. T4, 587 (1962).

11] Albers, C, Kappey, F., and Thiele, P., Die Elektrolytbrilcke als Fehlerquelle bei der
pH-Messung, Klin. Wsahr. 41_, 1095 (1965).

12] Maas, A. H. J., Van Heyst, A. N. P., and Visser, B. F., The determination of the true
equilibrium constant (p^ig) and the practical equilibrium coefficient (pi^jg) for the
first ionization of carbonic acid in solutions of sodium bicarbonate, cereorospinal
fluid, plasma and serum at 25 and 38 °C, Clin. Chim. Acta, 33, 325 (1971).

13] Trenchard, D., Noble, M. I. M,, and Guz, A., Serum carbonic acid pii:{ abnormalities in
patients with acid-base disturbances, Clin. Soi. 32^, 189 (1967).

14] Sinclair, M. J., Hart, R. A., Pope, H. M., and Campbell, E. J, M., The use of the
Henderson-Hassel balch equation in routine medical practice, Clin. Chim. Aota, 19, 63
(1968).

15] Rispens, P., Brunsting, J. R., Zijlstra, W. G., and Van Kampen, E. J., Determination of
total carbon dioxide in blood and plasma by means of the cediometer. Theory and
experimental verification, Clin. Chim. Acta, 22^, 261 (1968).

,16] Rispens, P., Signifiaanae of plasma bicarbonate for the evaluation of homeostasis,
p. 97 (Van Gorcum and Comp., Assen, The Netherlands, 1970).

17] Austin, H. W., Lacombe, E., Rand, P. W., and Chatterjee, M., Solubility of carbon
dioxide in serum from 15-38 °C, J. Appl. Physiol. ]S_, 301 (1963).

:i8] Rispens, P., Dellebarre, C. W., Eleveld, D., Helder, W., and Zijlstra, W. G., The
apparent first dissociation constant of carbonic acid in plasma between 16 and 42.5 °C,
Clin. Chim. Acta, 22, 627 (1968).

19] Austin, H. W., Ferrante, V., and Anderson, C, Evaluation of whole blood pz' in the
acutely ill patient, J. Lab. Clin. Med. 11, 129 (1968).

38
 National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975; Issued June 1977.

QUANTITATIVE RELATIONSHIPS BETWEEN TOTAL CO2 CONCENTRATION

IN BLOOD AND PLASMA, PLASMA BICARBONATE CONCENTRATION,
PLASMA pH AND CARBON DIOXIDE TENSION BETWEEN 16-42 °C

P. Rispens, J. P. Zock, and W. G. Zijistra

Laboratory of Chemical Physiology
University of Groningen
Groningen, The Netherlands

In our laboratory the acid-base status of patients or experimental animals is evaluated

by determining total CO2 concentration in blood {oqq^), blood pH at 37 °C (pH^(37)),
haemoglobin concentration and oxygen saturation (502)- Fi^om these quantities plasma
pH at 37 °C (pHp(37)), total CO2 concentration in plasma i^Q^) > plasma bicarbonate concen-
tration (^QQ") si^d blood pH and carbon dioxide tension at body temperature (pH^(r^)) and
calculated [1]^.
^CO ^^Zj^^

pH (37) is calculated from pH,(37) using eq. (1)

pHp = pH^ + 0.01 . (1)

c^Q and (^QQ- ai^e practically independent of temperature in anaerobical ly stored blood and
can be calculated from the measured quantities without considering body temperature T^.
c^Q is calculated from c^q using eq. (2)

("2.244 - 0.422
5q )
•
(8.74 - pH^(37))
aP = - - • (2)
CO2 CO2
^2.244 - 0.422 S^J •
(8.74 - pH^(37)) -0.02880^^^

e^^Q- can be calculated from c^q^ and pHp(37) using the Henderson-Hasselbalch equation,
which for this purpose may be written as

402
^C03 =
, ^oP^l
- pHp(37)
1

The value to be used for p^j at 37 °C and pH (37) is calculated using eq. (4)
P

p^l = -4.7416 +
IM0J4i ^ 0.015906 T - log (1 + )(^ °K). (4)
^^_pH°(gf+ 7

^Figures in brackets indicate the literature references at the end of this paper.
I

!
39
In contrast to c^q^ and s^'"°"9^.y depend upon temperature. PH^C^^) is
-^^QOg' ^CO
calculated using eq. (5)

P%iT^) = pH^C37) + C37 - T^] -(o.OlSl + 0.0058 (pHp(37) - 7.4)) (r^in °C). (5)

Pqq^C^^) is calculated from PH^C^^) and c^q using the Henderson-Hasselbalch equation,
which for this purpose is written as

"XOz
=
Ppn (r, ) .
' (6)
( pHP (r ) \
-
p^l
S . \10 ^ +1/

The value to be used for S at is calculated from eq. (7), which has been derived from
data of Austin et at. [2]

S = - 3.24615 + ^-^y^^ + 0.0049854 T [t in °K) . (7)

The value to be used for pH^Cr^) is calculated from pH^(r^) using eq. (1), the value to be
used for p^i at and pH^Cr^) is calculated using eq. (4).

Using the eqs. (1) through (7), a computer program can be made for the rapid calculation
of tJ^QQ- and pH^Cr^) and P^q (r^) at any temperature between 16 and 42 °C. When no computer
is available, the nomogram presented in figure 1 may be used [3]. This nomogram is an
extension of the nomogram described by Brunsting [4] which is, in turn, a modification of the
Singer-Hastings nomogram [5]. The nomogram is valid for oxygenated blood only. Correction
for desaturation is possible with the aid of the data of table 1.

Obviously, eqs. (1)through (7) can also be used to calculate pH,{2',), ^cOo^^;,)' ^HCOq»
eQ02 ^""^
^COa PH^(37), Pco2(37), o^^ and Sq^ have been measured or to calculate pH^(37),
w^^" ^^^^ '^^^^ measured.
^h^h^^ ^Z^M^' ^feo; ^?0p' 4op' ^C0.(^7)' ^Hb ^02

1 . Calculation of pH .

It has been found that there is a slight difference between pH^ and pH^ [6]. This
difference is probably due to the effect of haemolysis at the interface between saturated KCl
and blood on the liquid junction potential during the measurement of blood pH [7-9].

2. Calculation of Cqq^-

Within the pH range occurring in blood in vivo, c^q in red cells (cqo2^ ^'^ smaller than
c^Q . The difference is mainly due to the impermeability of the red cell membrane for
haemoglobin, organic phosphate and plasma protein ions, causing a Gibbs-Donnan distribution

40
 n40
c"^.. (mmol/l) PHb(37)
^ 30 shift in point of
ratation in rela-
P^Q^(mm Hg)

-10

20

-30

^0

-50

60
20T(^{°C)
70

H80
90
100
110
1 20
cPHC03-(meq/1)

(g/lOO ml)

Figure 1 Nomogram to derive c^q^> temperature {p^^ {T-j^)) and pH^ at

) ^COs' XO2
body temperature (pH^(r^)) from pH^ at 37 °C (pHj^(37)),
'CO2 ^Hb h- To derive
^Q^and ^j,Q- a line is drawn through the points corresponding to the measured values
of pH, (37), c^Q^ an^l a^^. a^^^ is then read at the point where the c^q scale for
= 0 is intersected, and <2^qq- at the point where the <^qq- scale is intersected.
''Hb
37 (37)) is read at the point where the p^q scale is intersected. To
^C02
find P^Q^ {T^)
(r, ) when T. differs from 37 °C, a line has to be drawn through the point
corresponding to the value found for c?^^ (at the e^Q^
e; scale for a^^ = 0) and another
point at the central Hj^ (37) scale. The latter point may be found using the lines
representing the shift of this point with temperature for a given value of (37) in the
right hand grid. In the right hand grid the line through the measured value of pH^(37).

is followed up (or down) to the point of intersection with the line corresponding to the
actual value of T^. The point of intersection is then projected on the central pHj(37)
scale. The lines of the left hand grid represent the change of pH^ with temperature
for a given value of pH^(37).

41
 Table Correction factors for cP^q- and
p^q^ for deoxygenated blood.
1.

'
"^Hb b^
g/1 6.8 7.0 7.2 7.4 7.6 7.8

20 0.3 0.3 0.4 0.5 0.5 0.7

40 0.6 0.7 0.8 0.9 1.1 1.3
60 1.0 1.1 1.2 1.4 1.7 2.1
80 1.3 1.5 1.7 1.9 2.3 2.8
100 1.6 1.8 2.1 2.5 2.9 3.7
120 2.0 2.2 2.6 3.0 3.6 4.5
140 2.3 2.7 3.1 3.6 4.3 5.5
160 2.7 3.1 3.6 4.2 5.1 6.5
180 3.1 3.5 4.1 4.8 5.9 7.5
200 3.5 4.0 4.6 5.5 6.7 8.7

The figures given represent the percentage to be subtracted from the value read

for ^QQ- and for p^q in the nomogram of figure 1, when the blood is completely
deoxygenated. When the blood is partly oxygenated the correction factor is found
by multiplication of the figures given with (1 - S^. ).

of the diffusible ions, amongst them HCO3. a given haemoglobin concentration within the
red cells (MCHC = a'^^) the Gibbs-Donnan ratio mainly depends upon pH^ and 5q [10], whereas
at given a^^^ and c^q^ depends upon the haematocrit H,
^qq^>

Thus, the ratio c^q ^^P®"'^^ ^P°" ^ ^""^ furthermore upon pH^ and .

/'^c02 '^Hb

Therefore, it seems preferable to base empirical equations for calculating o^q from c^q
upon c^j^ (= a'^^ • h) , pH^ and 5q . Equation (2) was derived by McHardy [11] from an equation
given by Visser [12] who derived his equation from a nomogram given by Van Slyke and Sendroy
[13]. The validity of eq, (2) was tested [9] using 317 blood samples in which o^q , c^q ,

pH^(37), c^^ and Sg were measured. Of the 317 blood samples, 83 were centrifuged at 37 °C,
the remaining 234 at room temperature. Results are shown in table 2. The mean difference
between c^q /c^q calculated using eq. (2) and c^q /c^q measured directly is but small. The

Table 2. Difference between o^q /c^q derived from pH^(37), and a^^
(or H or O2 capacity), and '^q^/(^^q^ calculated directly.

Centrifuged Mean ,

at (°C) difference SD° Remarks

37 83 +0.007 0.027 equation (2)

19-25 234 +0.014 0.034 equation (2)
19-25 and 37 317 +0.012 0.032 equation (2)

number of experiments
standard deviation

42
 .

mean difference between calculated and measured values of c^q /c^q of blood samples centrifuged

at room temperature is slightly greater than that found for samples centrifuged at 37 °C.
h
However, eq. (2) may be used to calculate c^q from c^^ at lower as well as at = 37 °C

without introducing an appreciable systematic error. Figure 2 illustrates that eq. (2) is

valid over the entire c^q range encountered in blood in vivo.

Figure 2. Relationship between ct^q^ calculated from c^q^. pHj,(37), o^^ and s^^ with

eq. (2) ^c?^Q^(calc)j and c^q^ measured directly (exp)^

(^q^

3. Calculation of "^^qq-

In contrast to eq. (6) (y.i. ) an overall check of the validity of eq. (3) is not
possible, because ^^"^"^^^ measured independently of c^q or pH^. However, because
c^COs
the denominator of the quotient in eq. (3) differs but slightly from unity (in the pH range
7.8-7.0, it ranges from 1.020 to 1.126), the inaccuracy in p!^-] and pH^ has a negligible
effect on c^^q- and the accuracy with which c^rn: determined depends upon that of
HCOc 'CO.
only.

4. Calculation of pW^{Tj^)

pH^ of anaerobically stored blood increases with decreasing temperature, because the
acid reaction of both carbonic acid, haemoglobin and plasma decreases at decreasing temper-

43
ature. Theoretically [9], ApH^/AT depends upon pH^(37), c^q and a^^ •
5q has only a minor
influence. Equation (5) was derived from experiments with bovine blood, in which the influ-
ence of c^q and o^^ proved negligible. The validity of eq. (5) for human and canine blood
was checked by measuring pH^(37) and pH^(30) of 336 human and 829 canine blood samples. In

addition, pH^(37) and pH^(20) was measured in 52 human blood samples and pH^(37) and pH^(16)
in 68 human blood samples. Results are shown in table 3. An excellent agreement between

Table 3. Mean difference and standard deviation between pH^(30), pH^(20) and pH^(16)
calculated from pH,(37) and pH,(30), pH,(20) and pH,(16) measured directly.

Number of Temperature
Blood measurements (°C) pH, (meas)-pH, (calc) SD

canine 829 30 +0.001 0.016

human 336 30 -0.003 0.015
human 58 20 -0.031 0.026
human 68 16 -0.040 0.015

calculated and measured pH^(30) was found both in human and canine blood. At 20 °C and 16 °C
the calculated values are somewhat lower than the measured ones. No explanation for this

discrepancy can be given. It is in contrast with the good agreement of P^^ at 16 °C calcu-
2
b
lated from pH^(37), a^^ , e^|^ and 5q with P^^ in the gas mixture with which the blood had

been equilibrated at 16 °C {v.i.).

5. Calculation of p^q^{Tj^) .

It is clear from eq. (6), that the accuracy with which P-^ (^t,) can be calculated from
b 2 " r,
pH^(37), Cj,Q , a^^ and 5q , depends upon the reliability with which c^q is calculated from

e^Q (eq. (2)) and P^^iTj^) is calculated from pH^(37) (eqs. (1) and (5)), as well as on the

validity of eqs. (4) and (7) for calculating pk^ and 5. The overall accuracy of the determina-
tion of P^Q (t^) was evaluated by equilibrating human blood with CO2/O2/N2 mixtures of known

Table 4, Mean difference and standard deviation of Pm calculated at tonometer

temperature from pHj(37), c^q^, c^^ and , and P^q of tonometry gas mixture.

""^"9^
^C02
(mm Hg) 37 "C 30 °C 16 °C

10-30 +1.2 ± 0.9 +0.,6 + 1.2 +0.,6 + 1.1

n = 57 n 38 n 15

30-50 +0.8 ± 1.7 +0..1 + 2.5 +0.,8 + 2.3

n = 48 n 17 n 15

-0,,9 + 2.3 +0.,6 + 4.3

50-80 -0.3 ± 2.0
n = 61 n 36 n 14

80-120 -1.9 ± 3.7 -4,.4 + 4.4

n = 27 n 13

44
 1-

composition, measuring c^q , pH^(37), a^^ and , calculating p^q^{t^) from the measured
values and comparing the calculated P^q (t^) with P^q in the gas mixture used. One hundred

ninety three blood samples were equilibrated at 37 °C, 104 at 30 °C and 44 at 16 °C. Results

are shown in table 4. At P-^ = 10-80 mm Hg a fair agreement is found. The lower values

for blood with Pm > 80 mm Hg, most evident in samples equilibrated at 30 °C, may be due to

loss of CO2 in the sampling or analysis procedure.

References

[1] Rispens, P. and Zijlstra, W. G., Blood gas transport and acid-base balance, in
Clinical Biochemistry ^ Principles and Methods , H. CH. Curtius and M. Roth, eds., p.
1604 (Walter de Gruyter, Berlin, New York, 1974).

[2] Austin, H. W., Lacombe, E., Rand, P. W., and Chatterjee, M., Solubility of carbon
dioxide in serum from 15-38 °C, J. Appl. Physiol. 18, 301 (1963).

[3] Rispens, P., Brunsting, J. R., Zock, J. P., and Zijlstra, W. G., A modified Singer-
Hastings nomogram, J. Appl. Physiol. 34, 377 (1973).

[4] Brunsting, J. R., Clinical Caj>hoxymetry , p. 95 (Van Gorcum and Comp., Assen, The Nether-
lands, 1962).

[5] Singer, R. B. and Hastings, A. B., An improved clinical method for the estimation of
the acid-base balance of human blood. Medicine, 2J_, 223 (1948).

[6] Severinghaus, J. W., Stupfel , M., and Bradley, A. F., Accuracy of blood pH and PCO2
determinations, J. Appl. Physiol. 9^, 189 (1956).

[7] Siggaard-Andersen, 0., Factors affecting the liquid-junction potential in electrometric

blood pH measurement, Scand. J. Clin. Lab. Invest. J_3, 205 (1961 ).

[8] Maas, A. H. J., pH determination of body fluids with a micro glass electrode and a
saturated KCl bridge in the cell, Clin. Chim. Acta, 28, 373 (1970).

[9] Rispens, P., Significance of plasma bicarbonate for the evaluation of homeostasis^
p. 116 (Van Gorcum and Comp., Assen, The Netherlands, 1970).

[10] Hastings, A. B., Sendroy, J., Mclntosch, J. F., and Van ,Slyke, D. D., Studies of gas
and electrolyte equilibria in blood. XIII. The distribution of chloride and bicarbonate
in the blood of normal and pathological human subjects, J. Biol. Chem. 79^, 193 (1928).

[n] McHardy, G. J. R., The relationship between the differences in pressure and content of
carbon dioxide in arterial and venous blood, Clin. Sai. 32, 299 (1967).

[12] Visser, B. F., Pulmonary diffusion of carbon dioxide, Phys. Med. Biol. b_, 155 (1961 ).

[13] Van Slyke, D. D. and Sendroy, J., Line charts for graphic calculations by the Henderson-
Hasselbalch equation and for calculating carbon dioxide content from whole blood content,
J. Biol. Chem. 79, 781 (1928).

45
I

i
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

Pf.f. INDEPENDENT QUANTITIES, WHY OR WHY NOT

P. Rispens, W. G. Zijistra, and E. J. van Kampen

Laboratory of Chemical Physiology
University of Groningen
Groningen, The Netherlands

and

Laboratory of Clinical Chemistry

Diakonessenhui s
Groningen, The Netherlands

In our laboratories, disturbances of acid-base balance are assessed by plotting plasma

pH at 37 °C (pH(37)) against plasma bicarbonate concentration (cSrr>-) "in a pH -cf,rn-
p nLUq p nUU3
diagram (fig. 1). Such a diagram, first introduced by Van Slyke [Ip and popularized by

comp. comp. prim.

n.r.oc. rolk?

Figure A pH (37)-t?2pr,- diagram with normal area and the area in which pH
1. r and may
P nl^U3 HCO3
p
be expected to fall during primary and compensated respiratory and non-respiratory dis-

turbances and during mixed disturbances of acid-base balance; r.ac. = respiratory acidosis;
r.alk. = respiratory alkalosis; n.r.ac. = non-respiratory acidosis; n.r.alk. = non-respiratory
al kalosis.

^Figures in brackets indicate literature references at the end of this paper.

47
Davenport [2] is denoted a Van Slyke-Davenport diagram. In this diagram, the areas in
which pHp and c^qq- may be expected to fall in normal conditions, primary respiratory and
non-respiratory disturbances, compensated respiratory and non-respiratory disturbances and
mixed disturbances are indicated. In addition, the diagram contains CO2 isobars and CO2
buffer lines. A CO2 isobar is a set of points with equal P^q ; a CO2 buffer line is a
graph showing the change of pH^ and <^^qq- upon changes of P^q . The basic ideas underlying
the use of the diagram can be found elsewhere [1-3].

The diagram demonstrates that the diagnosis of an acid-base disturbance cannot be

based on either pH^ or c^qq- (or Pqq^) alone. As Van Slyke put in 1921 [1]:

In order to determine which one of the possible variations

exists in the blood in vivo, it is necessary to ascertain two
of the involved variables, such as the pH, [BHCO3], and
[H2CO3]. With any of two of them a point can be located in
its proper area on a diagram but with any one of them
alone it cannot be done.

The diagram obviates the use of so-called P^q independent quantities (table 1) for
the diagnosis of acid-base disturbances. For the following reason, these quantities are
nevertheless widely in use. A respiratory disturbance of acid-base balance or a ventilatory
compensation of a non-respiratory disturbance manifests itself in a change of P^q . A
non-respiratory disturbance of acid-base balance or a metabolic or renal compensation of a
respiratory disturbance manifests itself in a change of c^qq-- However, whereas a res-
piratory disturbance and a ventilatory compensation of a non-respiratory disturbance can be
assessed by comparing the actual P^q with normal, non-respiratory disturbances, metabolic
and renal compensations of respiratory disturbances cannot be assessed by simply comparing
the actual ^qq- with normal. This is because changes of P^q also cause changes of
c^^Q-. Hence, a deviation of <?^qq- from normal may be due either to a non-respiratory

disturbance or metabolic or renal compensation or to a respiratory disturbance or ven-

tilatory compensation. The basic idea underlying the so-called P^q -independent quantities
is that the respiratory deviation of ^qq- is ruled out by restoring a normal P^q^ in blood
or plasma in vitro. Figure 2 illustrates that this basic idea is not true. The course of

pH
^
with cSrn- when P-^, is restored to normal in blood or plasma in vitro differs from
p HCO3 CO2
that in vivo, i.e., the course of the buffer line of isolated plasma and of blood in vitro
both differ from that of blood in vivo. In the case of the primary respiratory acidosis of

figure 2, standard bicarbonate is lower than normal. This would suggest a non-respiratory
acidosis (complicating a prevailing respiratory acidosis). The same conclusion would be
drawn if base excess or buffer base values were calculated from pH^, c^^Qg
^C02*
Alkali reserve is higher than normal in the case of figure 2. This would suggest a
metabolic or renal compensation of the respiratory acidosis.

It should be emphasized that neither the actual ^j^qq- nor the alleged P^Q^-independent
quantities are an unequivocal measure for non-respiratory disturbances from which the
amount of acid or base to be administered for restoring pH can be calculated [4,5]. This
y

48
 1 I
1I — 1 — ' ^ 1

<0
wn se

</) CD o
—
1

4-)
1 +->

r~ o
c <u
>
•

Xre

ra I/) c: 1 1
-o 4-> •r- I +-> s. C
J3 C •r- +J i- XI re

s-
ITS
J3 +->
(O
O)
S-
o la f~ ta
c 4->
ai
to
o
o o
X M- o <u
s- CD
to
re
O)
*«-
S-
0) 0)
O)
to o CM -o
o
'r-
E T3
Ol
<u
p-t E
<o
to
to
tl_
3
XX
M- <4-
H-
to
(O
to
i-
n3
J2
o
cu
o
t/>
S.
0) O
o +J
to
s-
05
<u
o
_Q s-
(1)
J3 O) J3 (U +J 0) re X re 4_) +-> <4-
O- <u
— • i_ OJ ^ 0)
—
II

re <4-

X3
S- (0 1
M- T3 ID- <u ro -o OJ re ^
O O) 0) J3 E ,
O 4_> fa to <u to i_ O)
+J <a s- •
S- n3 n3 to re 4^ CO n:
to
(/)
E
S_
+j
la
o 1 1
C
o
o
o
cn JD O
o j3
c\l
re o r\
3 E
E
o
0) O J3 c i- (U • #t
M- PL) i- to E s-
o ta o S- +-> , O T3 O) Q. M-
X 0) +J rc CO O)
o
M- o
dJ O)

+j
4-> o
o
CNl 4->
c:
o
+->
(0
s-
S-
XI
•r-
cn
+-> —
CO
o
Ol
c
+-> 4-
^3
X)
(U
S- M-
S-
O
T3
O)
S-
05
O O) o +J +J il 3
T3
II

3 •o
c
O H-CL Q-
00 S- !-> CL •p- cs n3 c\ to
0)
> O) re
VD 03
s- <u
Z3
cr OCVJ c
-l->
o O-
s-
re
Ol
</)
re
0)
re
0)
O
o re
OJ
3
o
o Q. +-> 4-> UJ <_) O i Q- re Q- d; E S- E
a
O)
O)
c re
31
0. s-
s-
<u
+j
re
B
s-
X
+->
0
CO
0 X 0
a: M-
»l-
<4-
re
s-
o
c
'r-
s CO
E
0
0 ly
roX
3
o>
>>
x>
-r- +J
o
o CL +J
+->
re
1

0 t-> X
X 1— s- re re o>
O
3 o —
o
re
(Ml 3
+-> O)
<4-
0
c 0 <u
to 01
re

— o-
0)
o I—
— X 3:
E
i
S- re Xre
re
zn aE
"3 —
fXi I

E E 0 cu X CD E re
S- <u E +-> 01 +J s- s- E
E
0 aE3 0
re s- to 4-> <u

C
+->
<4-
E
S-
3
4J
re
01
0 re
E
-a
to
to
X
(L)
<»-
<»- •0
•I- re o re x> (J 3 E +-> 3 <u 0 (U
c: s- E re (U XI s- <*- s-
I

OmT3
O x:
0)
Q.
II

S-
CM <U
XQ- 0 1 ro S-
Q- T5
3
to
II

3
to
O O
•r-
+J E
O)
3
0CM +->
o;
re
s-
cr>
re 0
to
re
<u
re
0)
0
0 to
re
re
0)
3 4-> 0 T3 o> a cc: E E
t/i
E E S-
cu 0 CO 0)
0 CM M- i- M-
E E <*- OJ M- 0
0 re 3 O) 0 OJ
0. S- 0 X> E S-
j- 0 E <u
4->
E
(U
<4-
4-
0 cc:
XQ. »»-
0
00 re
S-
X
+->
Ol <+- re 0) 01
3
—
X cn E
0
S-
0
s_ E
0
X 0 </5
E
to re 1
o> 4-> E
re > 0 0
XI 4- +-> +-> o> E
s-
3
O"
0
1

0
1

re
E
X
X^ re
E
E
to
+->
re
0) 0) E 0 +-> to Ol <j
<4- 0 0 E 'e 3 E
M- S- 0 S- 4-
3 re +-> J3 CD re 0) O) +J
CO +-> re +-> CL +-> </J to T3
0 S- E 0) (J re cu re re 0
1— +-> 0 •0 X E
to
X +->
QJ re T3
+J 0 S E E
re
£- -0 cn
i-
0 «4- X5 (N
+->
re
I

o rnO)
<_>
ro
3
o
re
Q.
0
0
E E 1 1
O •- re
E
re
Ol
E
0
0
0 Q)
«rt
Clh:
O
e
E
m-
0) +-> +^ OD E •- to re o)
Ol XI re
i-
re 1 1
o
•r-
3
O-
re
r-
>>
i-
+->
CJ
0)
QJ
^ T3
re
O
^ to
re
S- E XI (-> CU CL O 0) S-
0)
(->
0
s_
0 S-
3 D) "O M- re
t.
s-
M-
1- s- II
^- M- +J 3 C O S- 0 S- <j_ ... CM cu
re
re
3
cr
re
S-
M-
•r-
re
E
•!-
Q.
0 0
M-
3
XI
O O
0 M-
q-
X
n:
ro E Q) re S- -o O X s- 3 ly
1

0 to
re 0 E
S-
cu
Q.
E
+J
E
O
O •-> -O
OJ
+-> 00 O
re
II

XI
P-i

s-
XI
-a
Q.
E
re
0
E
0)
4->
O)
O X 1— re
S-
E
O)
CD
la: y
^ <4-
o +->
re
E
re

E
o> —
1

C^J
E
re
E S-
0 to 1 1 1 CJ)
I. " 3 S- 0
<(- CM E (U E
E tO
0
0 Co0 -Ol Cvl to 0
•p-
05
E
0
E
re CU 0 0 +->
0 E
+->
+J
s-
3
-a re
^ ^
T3
E
-r-
+J
re
0
CU
to
O)
E
E >^ cu
(J
> CO re to
E
t- X
+-) -!->

re
E
0
s-
si
Ol rn
^C! X
31
3
s-
XI
0 M-
(/)
re
Q- £3 Q- !D Q3 E 0 a:

49
 Fc02 (fn^Hg) KO 120 100 80

Figure 2. Alkali reserve and standard hicarhonate during a severe primary respiratory
acidosis. The alkali reserve is seen to be higher than normal, the standard bicarbonate
to be lower than normal.

is because changes of ^^^qq- (and thus, changes of P^^ -independent quantities which are in
some way related to or derived from c^qq- (table 1)) are but a distant reflection of the
non-volatile acids or bases actually gained or lost. In contrast to a respiratory dis-

turbance where the change of P^^ is a direct measure of the acid gained- or lost, a change
of t?^QQ- may indicate gain or loss of organic acids in metabolic disorders, gain or loss of
inorganic acids or bicarbonate in renal disorders and gain or loss of bicarbonate or
hydrochloric acid in gastro-intestinal disorders. It has been found, that the same pH^ and

c^QQ- may occur while different amounts of non-volatile acids or bases have been gained or
lost [6].

It should be also emphasized that neither the use of a pH -c^ro- diagram nor the use of

-independent quantities releases the physician who has to treat a patient with an

acid-base disturbance, from the obligation to go deeply into the pathophysiology of this
disturbance. The following case taken from reference 7 serves to illustrate this. It

concerns a 45-year old female patient suffering from chronic obstructive pulmonary disease,
admitted to the hospital during an acute exacerbation. On analysis of an arterial blood

sample, pH^ = 7.17, -^^^Qg ^ meq-r\ P^q^ = 92 mm Hg and S^^ = 59 percent were measured.

Treatment consisted of mechanical ventilation to bring p'^q^ back to normal. A second

arterial blood sample then yielded pH = 7.59, c^rr,- = 34 meq-l'^, P-^. = 36 mm Hg and
Sf. = 96 percent. From acidemic, the patient had become alkalemic; the treatment thus

50
ll
resulted in a serious condition leading to severe neurologic disfunction. Such mismanage-
ment may easily result from the uncritical application of either a PH^-^^^qq- diagram or of
P^Q -independent quantities. Considering the patient's history, it may be assumed that
before the acute exacerbation the patient had a compensated respiratory acidosis with high
Pqq , high c^QQ- and a nearly normal pH^, and with an arterial Pq still high enough to
ensure an arterial S~, of about 90 percent. During the acute exacerbation P-^ increased
U2 LU2
and Pq decreased. The increase of Pqq would in itself cause an increase of <^qq- (and a

decrease of pH^). However, Pq^ then decreased to values corresponding to the steep part of
the O2 dissociation curve and 5q decreased considerably. 'Hence, the patient became hypoxic,

lactic acid accumulated, pH^ decreased further and c^qq- decreased to a value usually found
during a primary respiratory acidosis. However, in this patient with chronic respiratory
disease a much higher (^qq- was to be expected if her condition had not been complicated by
a non-respiratory disturbance. When hypoxia was removed by mechanical ventilation, the
accumulated lactic acid was metabolized and (^qq- increased. As P^q decreased far below
the value existing before the acute exacerbation, pH^ increased to a high value.

If the patient's history had been considered with enough insight in the pathophysiology

of acid-base disturbances, the mechanical ventilation would have been so regulated that a
nearly normal pH^ would have been obtained instead of a normal Pqq^. For such an insight,

the alleged Pqq^- independent quantities are not required.

References

[1] Van Slyke, D. D., Studies of acidosis. XVII. The normal and abnormal variations in
the acid-base balance of the blood, J. Biol. Chem. 48, 153 (1921).,

[2] Davenport, H. W., The ABC of Aaid-Base Chemistry (The University of Chicago Press,
Chicago, 111., 1958).

[3] Rispens, P., Zijlstra, W. G., and Van Kampen, E. J., Significance of bicarbonate for
the evaluation of non-respiratory disturbance of acid-base balance, Clin. Chim. Aata,
54, 335 (1974).

[4] Van Slyke, D. D., Discussion remark, in Current concept of acid-base measurement, Ann.
N.Y. Aaad. Soi. , 133, 107 (1966).

[5] Garella, S., Dana, C. L., and Chazan, J. A., Severity of metabolic acidosis as a
determinant of bicarbonate requirements. New Eng. J. Med., 289, 121 (1973).

[6] Lemann, J., Lennon, E. J., Goodman, A. D., Litzow, J. R., and Relman, A. S., The net
balance of acids in subjects given large loads of acid or alkali, J. Clin. Invest.,
44, 507 (1965).

[7] Rotherann, E. B., Jr., Safar, P., and Robin, E. D., CNS disorder during mechanical
ventilation in chronic pulmonary disease, JAMA, 189 , 992 (1964).

[8] Van Slyke, D. D. and Cullen, G. E., Studies of acidosis. I. The bicarbonate con-
centration of the blood plasma; its significance and its determination as a measure of
acidosis, J. Biol. Chem., 30, 289 (1917).

[9] Singer, R. B. and Hastings, A. B., An improved clinical method for the estimation of
disturbances of the acid-base balance of human blood. Medicine, 27, 223 (1948).

51
[10] J0rgensen, K. and Astrup, P., Standard bicarbonate, its clinical significance and a
new method for its determination, Scand. J. Clin. Lab. Invest., 9^, 122 (1957).

[11] Siggaard-Andersen, 0. and Engel , K. , An improved method for the calculation of the
relevant blood acid-base data, Soand. J. Clin. Lab. Invest., 1_2, 177 (1960).

[12] Brunsting, J. R., Clinical Cajcboxymetry , p. 95 (Van Gorcum and Comp. , Assen, The
Netherlands, 1962).

52
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

BASE EXCESS. WHY REOPEN THE ACID-BASE DEBATE?

P. J. N. Howorth
Department of Chemical Pathology
King's College Hospital Medical School
London, U.K. SE5 8RX

Base excess (BE) [1]^ was introduced as a formal mathematical concept with the Astrup
system (AS) [2] in 1960, although the descriptive terms base excess and base deficit were
used by Peters and Van Slyke in the 1930' s [3]. By 1963, many criticisms had been leveled
against it. Astrup himself [4] proposed a conference to discuss the many problems in acid-
base understanding. The outcome was the 1964 New York Academy of Science Conference whose
contents were published in 1966 [5]. We must take this conference as a baseline to see what
progress has been made in the last decade. The Scandinavians conceded that BE was not a
completely objective description of the whole body's acid-base state. Astrup indeed never
claimed that it was; he held that it was an excellent way of describing changes in the
blood, not the whole body [4]. However, some of his supporters still tried to apply the
system to the whole body. In 1966, came the "corrected base excess" which was a formula to
modify the BE in the light of knowledge about the way the in vivo CO2 titration curve differed
from the in vitro one [6]. This corrected base excess did not achieve wide popularity. As
the decade passed the old arguments became half forgotten and peace reigned once more. In
1971, BE was modified to allow for buffering in EC fluid (SBE) and this was incorporated into
the Sfggaard-Andersen chart nomogram without too much fuss [7]. Then more recently, rumbles
of the old acid-base debate have been heard again. Why is this and what do the rumblers
hope to achieve?

Table 1. Recent milestones in acid-base history.

1954 Mark I Astrup equipment

1957 Standard bicarbonate
1960 Astrup trolley & system: base excess
1962 RIpH Moran Campbell
1963 Schwarz & Relman critique of Astrup system
1964 NYAS Conference: peace for a decade
1975/5 SI units. New analysers with old parameters

Table 2. Measurements of the non-respiratory component.

Plasma HCOi : Van Slyke 1917, 1931

Non-respiratory pH: Hasselbalch 1916
Whitehead 1964
Stoker 1972
Buffer base: Singer & Hastings 1948
Standard HCO3 Astrup 1957
:

Base excess: Siggaard-Andersen 1960

Corrected BE: Prys Roberts 1966
Standard BE: Siggaard-Andersen 1971

^Figures in brackets indicate the literature references at the end of this paper.
 :

The first and foremost point is that the instrument manufacturers have themselves
brought various skeletons out of the closet on their current blood gas analysers. They
appear increasingly reluctant to make simple machines that measure only pH, PO2 and PCO2.
What they like to sell at greater profitability are larger machines with accessory elec-
tronic circuits to calculate and print-out or display many of the old derived parameters
discussed at the 1964 Conference. The only modern parameter seems to the SBE on the Radi-
ometer ABLl and of course this has been attacked on the familiar arguments that there is no
way of knowing what the ECF buffers are in any particular patient, especially one with an
acute on chronic acid-base disturbance.

Table 3. Parameter measured on some current instruments.

Blood Gas Analysers

Name: Radiometer ABL Corning Eel 165 IL 413 Sandoz AVL 937

Measures Hb cone. All measure pH, PCO2 & PO2

Calculates; P-HCOi P-HCO; P-HCO; P-HCO;

Std HCO3 Std HCO3

total CO2 total CO2 total CO2 total CO2

base excess base excess base excess base excess

Std base excess buffer base

O2 saturation O2 saturation

The second major reason for the re-opening of the debate is the growing use of SI units
in chemical pathology. This process is gaining momentum in Europe; for example, SI units
will be in widespread use in the UK by the end of 1975 [8]. With the introduction of the SI,
it is convenient to ask ourselves whether H+ concentration can be more conveniently reported
in an arithmetic SI unit, i.e., nanomole/1 iter , as compared with the logarithmic pH notation.
Of course these arguments are anything but new. It is worth pointing out that the pH nota-
tion was introduced by S0rensen as a shorthand for the otherwise unwieldy H+' concentration
scale [9]. Arguments about what electrodes actually measure, and activity versus concentra-
tion arose soon after. For many years H"*" activity was reported by the special symbol paH
and it was only after Peters and Van Slyke, in their classic work published in 1931, pro-
nounced in favor of the use of pH to cover both H"^ activity and concentration that this
special symbol was finally abandoned [3]. In more recent years, many individual workers have
made use of the simpler arithmetic scale. One classic plea was Moran Campbell's RIpH in
1962 [10]. The upheaval caused by the introduction of the SI provides us with a very con-
venient opportunity to make this change. The arguments will remain, but at least those who
wish to use the arithmetic SI unit will find it easier to get their colleagues to accept the
change. Already some gas analysers have partial pressures in kilopascals and no doubt soon
we will be able to get an arithemtic [H"*"] output by means of an antilog circuit board as
used with ion-selective electrodes for ions other than the H'''.

Further reasons for the fresh debate are a little more nonspecific. It is partly due
to a rejection of the rather Teutonic heavy-handed approach of the Astrup system with its
formulae and nomograms by a younger generation of clinicians. They are often unaware of the
emotions generated in the past by the debates and they prefer a more flexible approach to
acid base assessment. Some points are:

A. A growing suspicion of nomograms because the historic need for them has largely
gone. They were needed in the past because laboratory instrumentation lagged behind theory
of acid-base understanding for many years. Thus, nomograms were introduced to extrapolate
the scanty laboratory data, e.g., to predict pH from CO2 measurements. With the availability
of ion-selective electrodes the need for nomograms has gone.

54
 Table 4. Why there used to be a need for nomograms.

(1) Need to estimate blood H"*" concentration, e.g.,

from plasma CO2 and HCO3

(2) Assumed that physio-chemical constants apply to

biological systems:
CO2 solubility coefficient = 0.510
PKl = 6.10 ± 0.01.

B. The modern approach is thus that laboratories should only report what they actually
measure and leave it to the individual clinician to carry around the Radiometer blood gas
calculator [11] if he wishes to derive parameters from the primary data. At a higher level
of sophistication we are of course still in trouble. A CO2 electrode does not really
measure PCO2 but a pH change related, we hope, to it. What do osmometers really measure?
How are they calibrated?

Table 5. Philosophy of the modern approach.

(1) Clinical chemists should report only:

what is actually measured
what is usefully measured, e.g.,
blood [H+], PgCOa, actual HCO3.

(2) They should not report derived parameters,

e.g., PCO2 at 33° when measured at 38°.

(3) Use of nomograms should be restrained.

C. Thirdly, there is a greater interest in the interrelationships between O2 delivery

to the tissues and H"^ concentration. Hence, more measurements are needed in acid-base work
than the big three (pH, PO2, PCO2). When we talk of the P50 we enter another nomogram
area because it is still a relatively new topic. Since the O2 dissociation curve is pH-
dependent the P50 would be best measured at the in vivo pH. One group at NIH calculates
the so-called in vivo P50 from the standard pH 7.40 P50 using the MCHC, DP6, BE, pH and
temperature [12]. So far we have restricted ourselves to measure red cell DPG as ymol/g Hb
(not per liter red cells because the cells alter their shape in acidosis and alkalosis).

Table 6. Summary of current acid-base tests.

1974 1975

pH H^ cone, nmol/1

PCO2 mmHg PCO2 kPa

P-HCO3 meq/1 P-HCO3 mmol/1

Std P-HCO3 meq/1 PO2 & P50 kPa

Base Excess Mkl-III 2:3-DPG

PO2 mmHg

References

[1] Siggaard-Andersen, 0. and Engel , K. , A new acid-base nomogram. An improved method for
the calculation of the relevant blood acid-base data, Saand. J. Clin. Lab. invest.
12, 177 (1960).

55
[2] Astrup, P., J0rgensen, K. , Siggaard-Andersen, 0. and Engel , K. , The acid-base metabolism.
A new approach. Lancet, j_, 1035 (1960).

[3] Peters, J. P. and Van Slyke, D. D., Quantitative Clinical Chemistry, I. Interpretations,
(Bailliere, Tindall and Cox, London, 1931).

[4] Astrup, P., Acid-base disorders, New. Eng. J. Med. 269 , 817 (1963).

[5] Nahas, F. F., ed.. Current concepts of acid-base measurements, Ann. N.Y. Acad. Sci.
133 , 274 (1966).

[6] Prys-Roberts, C, Kelman, G. R. and Nunn, J. F. , Determination of the in vivo CO2 ti-
,

tration curve of anesthetised man, Brit. J. Anaesth. 38, 500 (1966).

[7] Siggaard-Andersen, 0., An acid-base chart for arterial blood with normal and patho-
physiological reference areas, Soand. J. Clin. Lab. Invest. 2J_, 239 (1971).

[8] Brit. Med. J. Editorial., S. I. units, Brit. Med. J. 4, 490 (1974).

[9] Lancet Committee report. Acid-base terminology. Lancet 11^, 1010 (1965).

[10] Campbell, E. J. M. , RIpH, Lancet L , 681 (1962).

[11] Severinghaus, J. W., Blood gas calculator, J. Appl. Fhys. 21_, 1108 (1966).

[12] Bellingham, A. J., Detter, J. C, and Lenfant, C, Regulatory mechanisms of haemoglobin

oxygen affinity in acidosis and aldalosis, J. Clin, invest. 50, 700 (1971).

56
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

USE OF m VIVO CO2 TITRATION CURVES IN THE PHYSIOLOGICAL ASSESSMENT

OF ACID-BASE BALANCE

P. J. N. Howorth
Department of Chemical Pathology
King's College Hospital Medical School
Denmark Hill
London, SE5, 8RX, United Kingdom

In their 1963 critique of the Astrup system, Schwartz and Relman [1]^ recommended
instead the use of a physiologic approach to the clinical assessment of acid-base balance in
individual patients. This approach was based upon the interpretation of clinical and lab-
oratory data in the light of the known behavior patterns in respiratory and metabolic dis-
ease. This approach can be updated by making use of the in vivo CO2 titration curves in man.
Assessment can then be subdivided into three broad areas:

1. Primary acid-base data

a. The respiratory component - arterial blood PCO2.

b. The acidity of blood--H^ nmol/1 (or pH) which gives direct information about the
accumulation of non-volatile acid or base and indirectly of changes in PCO2.

2. Other cl inicopathological information

a. Clinical data such as the duration and severity of the patient's present illness,
its antecedent pathological conditions (if any), and treatment which might modify
acid-base status.

b. Laboratory data such as plasma actual HCO3 and other appropriate measurement, e.g.,
arterial blood P02, P50 (partial pressure of O2 when Hb is 50 percent saturated
with O2), red cell DPG, K"*" and concentration of other electrolytes.

3. Knowledge of the following:

The usual compensatory changes to altered arterial blood H"*" and PCO2:

i. The ventilatory response to metabolic acidosis and alkalosis.

ii. The renal response to respiratory acidosis and alkalosis.

In the physiological assessment of acid-base balance, it is important to know how the

^
body reacts to acute and chronic changes in the two primary measurements, arterial blood [H ]
and PCO2. The relative changes in [H^] and PCO2 are different in metabolic and respiratory
disturbances since the ventilatory response to metabolic acidosis and alkalosis is rapid
whereas the renal response to respiratory failure (acidosis) is slower, taking 3-4 days at
least to become maximal.

1. The Ventilatory Response to Changes in [H^] in Metabolic

Acidosis and Alkalosis

When the [H'^] in the blood changes abruptly the PCO2 may not alter, but within a few
hours there is a ventilatory response to changes in blood [H ] which may take 24 hours or

Figures in brackets indicate the literature references at the end of this paper.

57
more for the compensatory change in PCO2 to be complete [2]. In compensated metabolic
acidosis, it is rare for the arterial blood Pcoa to fall below 2.0 kPa (15 mmHg) and the PCO2
never falls below 1.33 kPa (10 mmHg) as the work of breathing limits the maximum ventilation
volume.

In metabolic alkalosis, the suppression of normal ventilation is similarly a compen-

sating mechanism. Observations of patients with arterial blood [H+] varying from 30 to
130 nmol/1 (pH 7.55-6.88) show that there is a hyperbolic relationship between the [H"*"] and
PCO2 [3,4].

2. The Renal Response to Changes in PCO2; Respiratory

Acidosis and Alkalosis

When the arterial blood PCO2 is acutely altered either by changes in ventilation or
breathing artificially high CO2 mixtures, the resulting acid-base values are known as the
whole body C02-titration curves. They have been used for a number of years in the assessment
of acute respiratory acidosis [5]. The blood [H"*"] values normally found as the blood Pcoa
changes from 2.0 to 12.0 kPa (15-90 mmHg) are shown in figure 1 [6,7]. The body reacts to a

•
90 12

80 1 1
. PCO2 / /
(l)Melabolic alkalosis
iO
•
70
9 My (2)Melabolic acidosis
-
60 8
(3)Respiralory acidosi s
7
50
6 (4)Re5piratory alkalos: s
• 40 5

30 4
3
20
2 (2)
-
10 1

1 1 1 1 1 t, 1 1 1 1

mmHg kPa 30 50 70 90 110 I30

Blood H^concenlralion nanomol / litre

Figure 1. Blood H concentration and PCO2 chart showing the range of observed values
during primary compensated metabolic alkalosis and acidosis and acute uncompensated
primary respiratory alkalosis and acidosis. During primary metabolic changes,
there is a superimposed respiratory response. (Data from Bone, et at. [3] and
Fulop, et aZ. [4] for the metabolic changes, and from Arbus, et al. 17} and
Brackett, et al. [6] for the respiratory changes.)

raised PCO2 by the renal retention of HCO3 ^""^ excretion of H"*". This process is significant
during the first 24-48 hours and becomes maximal after 3-4 days [8]. If the arterial blood
PCO2 remains below about 8 kPa (60 mmHg), the blood [H'''] may remain near normal, but renal
compensation begins to fail when the PCO2 rises above 9.5-10 kPa (71-75 mmHg) [9,10]. An
acute rise in arterial blood PCO2 is invariably accompanied by a fall in arterial blood P02
which may be severe enough to cause lactic acidosis. In respiratory failure without assisted
ventilation, it is rare for the PCO2 to rise above 10.5-12 kPa (79-90 mmHg) as the associated
hypoxemia (P02 4-2.6 kPa) (30-20 mmHg) endangers life [11].

Hyperventilation may cause an acute fall in arterial blood PCO2, the plasma HCO3 "^sy
not change significantly, but the [H"*"] may become dangerously low. The changes found during
inhalation of low O2 gas mixtures and during adaptation to high altitudes are similar; the
renal adaptation to a low arterial blood PCO2 is slow and may take several weeks for com-
pletion [12].

58
 / .

If a high arterial blood PCO2 is reduced by artificial ventilation, the restoration of

normal acid-base balance is slow because of the time taken for the renal compensating process
to work in the opposite direction. The kidneys may continue to excrete an acid urine for
some time, thereby producing a dangerous acid-base state indistinguishable from primary
uncompensated metabolic alkalosis [13].

3. The Whole-Body in vivo COa-Titration Curves

When patients with different initial levels of [H"*"] (pre-existing metabolic acidosis or
alkalosis) are ventilated, the resulting in vivo C02-titration curves are almost parallel to
the normal response. Hence, it has been suggested that the in vivo whole-body C02-titration
curves be used as an aid to the interpretation of acid-base problems [14-17]. Figure 2 shows
whole-body in vivo COa-titration figures from the work of Stoker and his colleagues [14]
recalculated and plotted out in SI units [18].

(2) (1)

90
-
li
•
80
10 • PC02(l)kPa //
- 70 9 •
(2)mmHg /V T /
60 8

7
50
-
6
40 5 -

30 4 -

-
3
20
-
2
- 10 1
-

10 20 30 40 50 60 70 BO 90 lOO 20 40 60 80
Blood H concenlralion nanomol/litre Blood H'^nmol /I

Figure 2, Whole-body CO2 titration lines Figure 3. Whole body CO2 titration lines
in vivo in man at different values of in vivo in man as shown in figure 2 with
nonrespiratory H Concentration (blood points omitted to show the effects on
H+ at PC02 5.3 kPa (40 mmHg)). (Data PCO2 of acute changes in H+ concentra-
from Stoker, et al. [14].) tion (rise and fall in H"*") and on H"*"
concentration of acute changes in PCO2
(rise or fal 1 in PCO2)

When a patient is ventilated or retains CO2 acutely, the points representing [H"*"] and
PCO2 move along a straight line parallel to those in figure 3 in which the individual points
have been omitted for clarity. An acute metabolic rise or fall in blood [H"*"] causes a
horizontal movement to the right or left. Mixed and compensated acid-base disturbances show
movement resulting from both vectors.

4. Examples of the Use of the in vivo Chart

The use of the in vivo chart will be illustrated by some examples. It should be noted
that the chart is not a nomogram in that it is not used to calculate derived parameters. The
purpose of the chart is to simplify analysis of acid-base data by comparing the behavior of
the individual patient with the known behavior in metabolic and respiratory change.

59
 A. Acute metabolic acidosis during chronic respiratory acidosis

Even apparently simple acid-base disturbances can have a "mixed" component. Figure 4
shows how an acute exacerbation of hypoxemia in chronic respiratory failure during which
period the blood PCO2 did not alter caused a "type A" lactic acidosis [19]. There was a
horizontal "shift to the right" of H"*" and the HCOi fell from 38 to 30 mmol/1. This severe

20 40 60 80
Blood H'^nmol/l

Figure 4. Acute metabolic acidosis due to worsening of hypoxia during chronic

respiratory failure. The arterial blood P02 fell from 4.7 kPa (35 mmHg) at
point (1) to 2.7 kPa (20 mmHg at point (2) and the bicarbonate fell from 38
mmol/1, point (1) to 30 mmol/1 at point (2) without change in PCO2 (9.3 kPa
(70 mmHg)). The primary acid-base results at point (2) could equally have
arisen from an acute respiratiry acidosis with early renal compensation in
a previously normal individual (point la to 2), or from an exacerbation of
CO2 retention in a patient with moderate chronic respiratory acidosis (point
lb to 2).

chronic respiratory acidosis, with maximal renal compensation, and the additional acute
metabolic acidosis cannot be distinguished from inspection of the blood acid-base results
from an acute severe respiratory acidosis with early renal compensation in a previously
normal individual (point la) or an acute exacerbation of CO2 retention in a patient with
moderate chronic respiratory acidosis (point lb).

B. Acute on chronic respiratory failure

Figure 5 shows the changes over 48 hours in a patient with an acute exacerbation of
chronic bronchitis. On admission (1), the arterial blood PCO2 was acutely elevated. Despite
antibiotics and controlled O2 administration, there was further CO2 retention. Intermittent
positive pressure respiration for 24 hours reduced the arterial blood PCO2 to 8.0 kPa (60
mmHg) which was similar to the patient's normal PCO2.

C. Acute respiratory acidosis with concealed metabolic alkalosis

A woman aged 60 years was admitted in acute respiratory failure; she was confused,
making a history impossible, restless and cyanosed. Intermittent positive pressure ven-
tilation was instituted and after a few hours, ventilation results were at point (2) (see
fig. 6). At this point, it is now clear that the patient probably had a mixed disturbance
since, with the correction of the respiratory failure by intermittent positive pressure

60
 J 1 L

20 40 60 80
Blood H'^nmol/I

Figure 5. Acute exacerbation of chronic bronchitis in a man aged 65 years. On

admission (1), the patient was hypoxemic, P02 4.0 kPa (30 mmHg), plasma HCO3
32 mmol/1; (2), further increase in CO2 retention during early treatment, plasma
HCO3 36 mmol/1. Intermittent positive-pressure ventilation was instituted
which rapidly lowered the PCO2 (3), plasma HCO3 38 mmol/1, but the patient
remained in compensated respiratory failure.

20 40 60 80
Blood H'*'nmol/l

Figure 6. Acute respiratory failure with concealed metabolic alkalosis in a

woman aged 60 years. On admission, the patient was cyanosed and comatose;
P02 6.4 kPa (48 mmHg), plasma Na+ 137, r
5.8, CI" 83, HCO3 34, urea 3.3 mmol/1
(20 mg/dl). Intermittent positive pressure respiration quickly reduced the PCO2
and revealed the metabolic alkalosis (point 2) which was subsequently found to be
associated with long-term diuretic therapy for heart failure due to chronic
hypoxemia with bronchitis. At point (3), electrolyte results were plasma
Na+ 140, K+ 2.7, Ca2+ 1.85 (7.4 mg/dl), HCO3 45 mmol/1 and the patient began
to have tetanic convulsions due to the untreated metabolic alkalosis from which
she died 48 hours later.

61
respiration, her results at point (2) lie in the area of metabolic alkalosis shown pre-
viously. One can thus easily predict that more ventilation would only lower still further
the already dangerously low [H+], Unfortunately, ventilation was continued and the patient
died after 48 hours from tetanic convulsions. The full history later obtained from relatives
revealed that the patient was on long-term diuretic therapy for congestive cardiac failure
due to chronic hypoxemia with air way obstruction. The typical hypokalemia was concealed
initially by the respiratory acidemia. Thus, the patient had two separate reasons to
develop a metabolic alkalosis: (a) urinary K+ loss due to diuretic therapy, and (b)
paradoxical aciduria after lowering the high PCO2 (see above).

In retrospect, it is evident that this patient's only chance was for energetic treat-
ment of the metabolic alkalosis at point (2). Intravenous HCl is direct and logical [20].
If given stronger than 0.09 mol it may cause hemolysis and intravenous sodium phosphate,
,

ammonium chloride, arginine hydrochloride, and even dialysis have been advocated.

There has been current interest in metabolic alkalosis because its incidence has been
reported to be rising and the body tolerates alkalosis much less well than acidosis [21].
Whereas patients have survived with blood [H+] 170 nmol/1 (pH 6.78), the mortality is high
in metabolic alkalosis with blood [H'''] 28 nmol/1 (pH 7.55) and survival is rare below blood
[H+] 23 nmol/1 (pH 7.65). Alkalosis may occur not only from potassium loss and in hepatic
failure, but also after surgery, e.g., after bypass operations, and in trauma, especially
associated with sepsis and peritonitis. The causes are varied and may include: loss of H"*"
through gastric suction; over-use of alkaline liquids in early therapy, or urinary potassium
loss through adrenal cortical hormone activity [21].

Table 1. Metabolic alkalosis.

Mortality rises with fall in [H^]:

moderate < 28 nmol/1 pH 7.55

high H < 23 nmol .1 7.65

Occurrence:

Medical: drugs; e.g., diuretics

acute liver failure
chronic bronchitics on ventilation

Surgical: major trauma, by-pass operations

sepsis, hemorrhage, peritonitis

Aetiological Factors:

loss of gastric - suction

alkaline fluids - ACD blood, NaHCOg
paradoxical aciduria after fall in PCO2
K"*" loss in urine - adrenal hormones

Treatment

0.09 mol HCl in NaCl

iv NH4CI
iv Na phosphate, pH 7.0
iv arginine, HCl

In view of the large iatrogenic component and the graver risk to life in metabolic
alkalosis as compared with metabolic acidosis, it is clear that physicians and anesthetists
working in intensive care units need to exercise special care to prevent the unnecessary
development of metabolic alkalosis.

62
 D. Compensated metabolic acidosis

Figure 7 shows the acid-base changes over 12 hours during the resuscitation of a
patient with acute diabetic ketoacidosis. On admission (1), there was a marked metabolic
acidosis with compensatory low PCO2. Therapy with insulin, intravenous NaCl and potassium
supplements was enough to bring the patient to near normal acid-base balance. Intravenous
NaHC03 was not given.

20 40 60 80
Blood H^'ntnol/I

Figure 7. Acute diabetic ketoacidosis in a man aged 30. On admission, the plasma
glucose was 20 mmol/1 (360 mg/dl ) plasma HCO3 less than 5 mmol/1.
, The
patient was resuscitated with intravenous NaCl with KCl supplements and insulin.
After 6 hours (2), the plasma HCO3 was 8 mmol/1. The patient had recovered
by 12 hours (3), although the acid-base state was not completely normal, plasma
HCO3 18 mmol/1.

E. Compensated metabolic alkalosis and acute respiratory acidosis

A further example of a complex acid-base disturbance was provided recently (figs. 8

and 9). A 45-year-old workman (stone mason) was admitted in a semi-comatose state, pe-
ripheral cyanosis, vomiting and tetany. From his past history of suspected alcoholic
excesses, he was thought initially to have possibly Wernicke or hepatic pre-coma. His
LFTs, blood EtOH and red cell TK did not support this view so he was urgently re-assessed.

Date 14 15 16 17 21 22 April 1975

H+ 23 36 36 37 35 nmol/1
Na+ 132 140 137 145 138 134 mmol/1
K+ 2.6 3.1 3.9 3.1 3.4 2.7 mmol/1
HCO3 >50 >40 35 35 22 mmol/1
Urea 15.5 27.4 25.0 15.6 4.5 4.2 mmol/1
PCO2 10.0 6.7' 6.1 6.9 5.1 kPa
PO2 5.9 10.3 8.1 12.6 9.8 kPa
DPG 31.9 25.3 13.9 14.1 15.6 ymol/g Hb
14th semi-coma, cyanosis, ?li ver/EtOH Figure 8. CI inico-pathological data (table)
15th intubated and ventilated until 21st on a man aged 45 with combined chronic
16th pneumothorax:chest drained metabolic alkalosis due to ingestion of
A(l) chronic DU. Alkalosis from oral NaHCOs NaHCOs and acute respiratory failure due
(2) RLLpneumonia pneumonia and pneumothorax.

63
 Figure 9. Primary acid-base results show '

days after admission in brackets. i

20 40 60 80
[H"^] nanomole/liire i

A gross metabolic alkalosis, right lobar collapse and pneumothorax were found and he was i

intubated and ventilated. The alkalosis was due to excess ingestion of NaHCOs for the pain
j

of a chronic DU. The initial blood gas results are thus consistent with a profound com- '

pensated metabolic alkalosis and the high red cell DPG results support this. '

The acid-base disturbance was fairly rapidly restored to near-normality, but then his
acid-base point moved up and down the in vivo line a bit as is often found in patients on
ventilation. There was thus a chronic condition with an acute respiratory episode which
led to the patient's hospitalization.

F. Combined metabolic and respiratory acidosis

In cardiac arrest, there is a combination of acute respiratory acidosis and hypoxic

metabolic acidosis. Figure 10 shows the sequence of events in the attempted resuscitation
of a man aged 55 who arrested after a myocardial infarct. Blood taken 15 minutes after the
arrest showed a profound combined acidosis. The PCO2 was reduced by ventilation (2) and
the metabolic component was improved by 100 mmol of intravenous NaHCOs (3). At this point,
the airway became obstructed and after fruitless attempts to improve ventilation the
patient died (5) about 2 hours after the arrest.

It is reasonable to consider reducing a high metabolic [H^] in life-threatening

situations by administering NaHCOs, although it is likely that death in severe acidosis is
due to hypoxic damage rather that to H''' ions -per se. The arguments against NaHCOs therapy
are:

1. Most acute metabolic acidosis is due to oxidizable acids:

a. Lactic/Pyruvic acid in acute hypoxia.

b. Free fatty, acetoacetic and e-hydroxybutyric acids in acute diabetic

ketoacidosis.

2. O2 release to the tissues is partly controlled by red cell 2,3-diphosphoglycerate

(DPG) [22,23] which modifies the shape of the Oa-dissociation curve. For blood
with normal O2 affinity (P50 = 3.6 kPa (27 mmHg)), a "shift to the right" of 0.4
kPa (3 mmHg) may deliver over 20 percent more O2 to the tissues. Hypoxia and
[H'''] are the two major physiological factors which, with plasma inorganic phos-

phorus, determine the red cell content of DPG.

Acidosis both improves O2 release in the tissues (Bohr effect) [24] and depletes red
cell DPG. Restoration of red cell DPG is a slow process (T1/2 about 11 hours) and too
hasty correction of acidosis allows the DPG deficiency to predominate by shifting the O2-

64
 ,

20 40 60 80
Blood H'''nmol/l

Figure 10. Attempted resuscitation of a man aged 55 years with cardiac arrest
due to cardiac arrhythmia after a myocardial infarct. Results at point (1)
taken on blood 15 minutes after the arrest showed a profound combined
respiratory and metabolic acidosis, plasma HCO3 18 mmol/1. The PCO2 was reduced
by ventilation (2) and the metabolic acidosis was improved after infusion of
100 mmol NaHCOs intravenously (3), plasma HCO3 22 mmol/1. Further acute CO2
retention supervened (4) and the patient died within 2 hours (5).

dissociation curve to the left, which impairs O2 delivery to the tissues. Some workers
have suggested correction of acidosis only to about [H"*"] 55 nmol/1 (pH 7.26) [25] while
others have advocated intravenous sodium phosphate so as to increase the availability of
phosphorus for DPG production [26]. Equally, methods for the enzymatic determination of
red cell DPG are improving [27] so that it is now possible to measure DPG routinely in
patients with acute acid-base disturbance.

I am obliged to the editor of the British Journal of Chest Diseases for permission to make
use of material published previously [17].

References

[1] Schwartz, W. B. and Relman, A. S., A critique of the parameters used in the evaluation
of acid-base disorders. New Eng. J. Med. 268, 1382 (1963).

[2] Pierce, N. F., Fedson, D. S., Brigham, K. L., Mitra, R. C, Sack, R. B., and Mondel
A., The ventilatory response to acute acid-base deficit in humans, Ann. Intern. Med.
72, 633 (1970).

[3] Bone, J. M., Cowie, J., Lambie, A. T., and Robson, J. S., The relationship between
arterial PCO2 and hydrogen ion concentration in chronic metabolic acidosis and
alkalosis, Clin. Sci. Moleo. Med. 46, 113 (1974).

[4] Fulop, M., Dreyer, N., and Tannenbaum, M., The ventilatory response in diabetic
ketoacidosis, Clin. Sai. Molea. Med. 46, 539 (1974).

65
[5] Van Ypersele de Strihou, C. and Frans, A., Les desordres acido-basique au cours de
1
insuff icance respiratoire:
' Un nouveau diagramme destine a leur interpretation,
Presse Med. 75, 1797 (1967).

[6] Brackett, N, C, Cohen, J. J., and Schwartz, W. B., Carbon dioxide titration curve of
normal man: Effect of increasing degrees of acute hypercapnia on acid-base equilibrium.
New Eng. J. Med. Zn, 6 (1965).

[7] Arbus, G. S,, Hebert, L. A., Levesque, P. R., Etsten, B. E., and Schwartz, W. B.,
Characterization and clinical application of the 'significance band' for acute respiratory
alkalosis. New Eng. J. Med. 280 117 (1969).
,

[8] Sullivan, W. J. and Dorman, P. J., The renal response to chronic respiratory acidosis,
J. Clin. Invest. 34, 268 (1955).

[9] Refsum, H. E., Arterial blood gases in respiratory insufficiency. Acta Chir. Sound.,
Suppl. 253, 175 (1960).

[10] Refsum, H. E., Acid-base status in patients with chronic hypercapnia and hypoxaemia,
Clin. Sai. 27, 407 (1964).

[11] Refsum, H. E., Relationship between state of consciousness and arterial hypoxaemia 1n
patients with pulmonary insufficiency, breathing air, Clin. Sai. 25, 361 0963).

[12] Chiodi, H., Respiratory adaptation to high altitude, in The Regulation of Human
Respiration, D. J. C. Cunningham and B. B. Lloyd, eds., p. 363 (Blackwell, Oxford,
1963).

[13] Refsum, H. E., Hypokalaemic alkalosis with paradoxical aciduria during artificial
ventilation of patients with pulmonary insufficiency and high plasma bicarbonate
concentration, Sound. J. Clin. Lab. Invest. 481 (1961 ).

[14] Stoker, J. B., Kappagoda, C. T., Grimshaw, V. A., and Linden, R. J., A new method for
assessing states of acute acidaemia in man, Clin. Soi. 42, 455 (1972).

[15] Stoker, J. B., Kappagoda, C. T., Snow, H. M., and Linden, R. J., The assessment of
acid-base disturbance in man by the use of carbon dioxide titration curves, Clin. Soi.
and Moleo. Med. 48, 133 (1975).

[16] Acids, bases and nomograms, Lunoet (editorial), 814 (1974).

[17] Howorth, P. J. N., The physiological assessment of .acid-base balance, Brit. J. Dis.
Chest. 69, 75 (1975).

[18] Howorth, P. J. N., RIpH revisited, Lunoet, U 253 (1974).

[19] Lactic acidosis, Lanoet (editorial), n_' 27 (1973).

[20] Hydrochloric acid for metabolic alkalosis, Lunoet (editorial), l_, 720 (1974).

[21] Wilson, R. E., Gibson, D., Percinel, A. K., Ali, M. A., Baker, G., LeBlanc, L. P., and
Lucas, C, Severe alkalosis in critically ill surgical patient, Arohs. Surg., Chiougo,
105 197 (1972).
,

[22] Benesch, R. and Benesch, R. E., The effect of organic phosphates from the human eryth-
rocyte on the allosteric properties of haemoglobin, Bioohem. Biophys. Res. Comm. 26^,
162 (1967).

[23] Chanutin, A. and Cornish, R. R., Effect of organic and inorganic phosphates on the
oxygen equilibrium of human erythrocytes, Arohs. Bioohem. Biophys. 1 21 96 (1967). ,

66
[24] Bohr,C, Hasselbalch, K., and Krogh, A., Uber einer in Biologischer Beziehung Wichtigen
Einfiuss, den die Kohlensaurespannung des Blutes auf dessen Sauerstoff Bindung, Ubt,
Soand. Arch. Physiol. 16, 402 (1904).

[25] Zimmet, P. Z., Taft, P., Ennis, G. C, and Sheath, J., Acid production in diabetic
acidosis; a more rational approach to alkali replacement, Br>. Med. J. 3^, 610 (1970).

[26] Alberti, K. G. M. M., Darley, J. H., Emerson, P. M., and Hockaday, T. D. R., 2,3-
diphosphoglycerate and tissue oxygenation in uncontrolled diabetes Mellitus, Lancet,
Jl, 391 (1972).

[27] Teunissen, A. J., De Leeuw, R. J. M., Boink, A. B. T. J., Hamelink, M. L., and Maas, A.
H. J., Comparison of five methods for determination of 2,3-diphosphoglycerate in blood,
Clin. Chem. 20, 649 (1974).

67
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

RELATIVE ACTIVITY AND SI UNITS

B. F. Visser
Department of Anaesthesiology
University Hospital
Nijmegen, The Netherlands

A. H. J. Maas
Department of Cardiology
University Hospital
Utrecht, The Netherlands

1. Sl-Units

The name International System (Systeme International, SI) of Units was adopted by the
11th Conference G^nerale des Poids et Mesures (CGPM) in 1960 [1]^- The system is now based
on seven base units for the corresponding basic kinds of quantities (table 1).

Table 1. The seven base units of the International System of Units.

(Symbols for quantities are printed in italics , symbols for
units in Roman type).

Quantity Symbol Dimension Unit Symbol of

length I L metre m

mass m M kilogram kg

time t T second s

electric current I I ampere A

thermodynamic temperature T 0 kelvin K

luminous intensity J candela cd

^V
amount of substance n N mole mol

All other SI units are derived from base units by multiplication or division without
the introduction of numerical factors; for example the derived, coherent unit of volume is
the cubic metre (m^) and the derived, coherent unit of (substance) concentration is the
mole per cubic metre (mol/m^ or mol'm"^). Several derived SI units have been given special
names, some of them are to be found in table 2.

The litre (1) is a non-coherent unit of volume. The litre has been redefined [2] and
is now exactly equal to the cubic decimetre (dm^). All other units containing the litre in
either the numerator or denominator are non-coherent, e.g.^ mol /I as the unit of substance
concentration.

Figures in brackets indicate the literature references at the end of this paper.

69
 Table 2. Some derived SI units with special names.
(A full stop (•) indicates multiplication
of units. A solidus (/) indicates division
of units; alternatively the denominator
may be expressed with a negative exponent.)

Quantity Symbol Unit Name Symbol

force F I<g«m/s2 newton N

pressure V N/m2 pascal Pa

work, energy W N-m joule J

Pa.m3

kPa-1

An extremely important feature of the International System is the fact that there is
only one unit of energy, the joule (1 J = 1 N-m = 1 kg^m^/s^). Work is mechanical energy
and, therefore, the unit of work is also the joule. The product of pressure (p) and volume
(f) has the dimension of work; coherent units are pascal (Pa = N/m^), cubic metre (m^) and
Pa-m^ = N-m = J. When the litre is chosen as the non-coherent unit of volume, the kilopas-
cal (kPa) must be the noncoherent unit of pressure [3].

1 kPa-1 = 103 Pa. 10-3 ni3 = 1 Pa.m3 = 1 N-m = 1 J (1)

2. Reference Quantities

The International Organization for Standardization defines a reference quantity in the

following way [4]:

"Physical quantities are concepts used for qualitative and quantitative descriptions
of physical phenomena. Such quantities may be classified into categories, each category
containing only quantities which are mutually comparable. If one of the quantities in such
a category is chosen as a reference quantity, called the unit, any other quantity in this
category can be expressed as a product of this unit and a number, called the numerical value
of the quantity. For a quantity symbolized by A, this relationship may be expressed in the
form

4 = U) X U] (2)

where U] is here used to symbolize the unit chosen for the quantity A, and [a) to symbolize
the numerical value of the quantity A when expressed in the unit U]."

The simplest case is present, when the unit is unity (proposed symbol I, not yet accepted
by CQUCC and EPQU [5]). Examples are: volume fraction {^) and m.ole fraction {x). The SI
units for plane angle and solid angle, the radian (rad) and the steradian (sr), respectively,
are called supplementary units in the International System of Units.

When a quantity has to be made dimensionless , two ways are followed.

a) The quantity A is divided by the unit U], thus giving the numerical value {a)

70
This procedure is recommended by Siggaard-Andersen [6,7].

b) The quantity A is divided by a quantity .4^=1, representative for the standard state

U)- 'A
= (A) . (4)
^0 (!)• A_

This convention is advocated by Bates [8].

In both cases the result of the operation is a numerical value depending on the unit
chosen. This is in contrast with true dimensionless quantities, where the value is indepen-
dent of the system of units, the unit being unity.

c) Theoretical ly--and why not practical ly--the solution is to refer to a standard

quantity

In this way a ratio of numerical values is obtained, this ratio being independent of the
system of units.

3. Relative Activities

There are three different activity scales:

a) on a concentration (c) basis

a = yo lim a /e = 1 (6)
^ c=o ^

b) on a molality (m) basis

a = ym lim a /m = 1 (7]

c) on a mole fraction {x) basis

a
X
= f'x Mm a /x =
X
] (8)

Activities on the molality and concentration scales are related by the equation [8]:

a /a = p„ (9)

where is the density of the solvent. When the solvent is water, the value of
close to unity, and for this reason no distinction is made between a and a " [61.
cm "is very

71
 The value of Pq is close to unity only if the litre is chosen as the non-coherent unit
of volume. The coherent SI unit of density is kg/m^, the value of the density of water becomes
approximately 1000. Therefore, the simplification is not justified.

When activities have to be made dimensionless , the suggestion of section 2.c. may be
followed. Foreach scale there will be one standard quantity, but the relationship between
these quantities will be known.

References

[1 ] Comptes rendus des seanaes de la Heme Conferenoe Generate des Poids et Mesures, Paris,
1960 (Gauthier-Vallers, Paris, 1961).

[2] Comptes rendus des seanaes de la 12eme Conference denerale des Voids et Mesures, Paris,
1964 (Gauthier-Villars, Paris, 1964).

[3] Visser, B. F., Quantities and units in respiratory physiology. Bull. Physiopath. Hesp.
9, 513 (1973).

[4] International Standard ISO 21 series. Part 0. General principles concerning quantities,
units and symbols, 13 pp. (Internationpil Organization for Standardization, Geneva,
1974). Part 8. Physical chemistry and molecular physics, 13 pp. (International Organi-
zation for Standardization, Geneva, 1973).

[5] Minutes of the meeting of CQUCC (Commission on Quantities and Units in Clinical Chemistry,
lUPAC section on Clinical Chemistry) and EPQU (Expert Panel on Quantities and Units,
IFCC Committee on Standards), Wageningen (April 18-21, 1975).

[6] Siggaard-Andersen, 0., The Aaid-Base Status of the Blood, 4th revised ed., 229 pp.
(Munksgaard, Copenhagen, 1974).

[7] Siggaard-Andersen, 0., Clin. Chem. 20, 727 (1974).

[8] Bates, R. G., Determinations ofpH. Theory and Practice, 2nd ed. , 435 pp. (John
Wiley & Sons, New York, 1973).

72
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

SEMI-EMPIRICAL ACID-BASE PROGRAM

B, F. Visser
Department of Anaesthesiology
University Hospital
Nijmegen, The Netherlands

A. J, Hoelen, J. A. Kreuger, and A. H. J. Maas

Department of Cardiology
University Hospital
Utrecht, The Netherlands

In spite of many efforts {e.g., [1-3]^), mathematical models of the acid-base behavior
of human blood have not yet led to useful programs for calculating derived acid-base data
from measured primary data. The calculations of Scheffner and Martin [4] are restricted to
the interpolation method for pC02 and the determination of actual bicarbonate, total CO2 and
standard bicarbonate.

On the other hand, empirical approaches have been quite successful. Siggaard-Andersen
[5] distinguishes between programs utilizing linear pH, log pC02 equilibration curves and
those utilizing linear pH, ctHCOa equilibration curves. We confine ourselves to the first
category.

Dell and Winters [6] store the coordinates of the base excess and buffer base curves in
the computer. Hardt [7] uses three different empirical equations for base excess, each
working in its own pH area. Recently the same system was followed by Knoll et al. [8].
Vallbona et al. [9] apply a simplified equation for the calculation of the base excess
concentration. Englesson et al. [10] make use of an assumed "mathematical center," the
point of convergence for the pH-log pCOa lines.

We call our program a semi -empirical one, because the basic equations have a sound
theoretical background, but are empirically adjusted to measured data by the introduction of
second-order correction terms or factors.

The slopes of the equilibration lines of separated human blood plasma in the pH-log
PCO2 diagram could be predicted [11]. The same type of equation was used to describe the
slopes of the equilibration lines of whole blood as a function of plasma protein, hemoglobin
(Hb) and standard bicarbonate (SB) concentrations. Adding a term with Hb^ led to the
necessary accuracy for the calculation of Hb from a given slope M

M = - 1.0285 + 0.0051 Hb - (3.5242 + 0.6105 Hb + 0.009 Hb2)/SB

Base excess (BE) is calculated from Hb content and Van Slyke standard bicarbonate (VB)

BE = (1 + D X Hb)-(VB - 24) (1)

Figures in brackets indicate the literature references at the end of this paper.

73
with

D = - 0.0233(BE + 20)/(34.232 + BE) (2)

instead of taking a constant value of D [5].

For further details of the program refer to [12].

The semi -empirical approach combines high accuracy with low computer power and could be
used as a means of averaging measured data.

References

[I] Visser, B. F., Zuur-base eigenschappen van bloed bij constante temperatuur en constante
verzadiging, Utrecht (1966) 26 pp. Cited by [3] and [13].

[2] Lloyd, B. B. and Michel, C. C, A theoretical treatment of the carbon dioxide dissocia-
tion curve of true plasma in vitro, Resp. Physiol. 107 (1966).

[3] Brodda, K., On the theory of base excess curve in the Siggaard-Andersen nomogram.
Respiration 32, 378 (1975).

[4] Scheffner, D. and Martin, H., Vergleich zwischen elektronisch errechneten und aus dem
Nomogramm nach Singer und Hastings bestimmten Werten des Saurebasenhaushalts, Clin.
Chim. Acta, J_6, 297 (1967).

[5] Siggaard-Andersen, 0., The Aoid-Base Status of the Blood, 4th revised ed., 299 pp.
(Munksgaard, Copenhagen, 1974).

[6] Dell, R. B. and Winters, R. W., A computer program for the blood pH-log pCOa nomogram,
Scand. J. Clin. Lab. Invest. 19, 29 (1967).

[7] Hardt, J., A computer program for calculating blood acid-base parameters on an Olivetti
"Programma 101" desk computer, Clin. Chem. 18, 658 (1972).

[8] Knoll, E., Wisser, H. and Dettmer, K., Berechung der Parameter des Saure^Basen-Haushaltes
mit Hilfe eines Kleincomputers, Z. Klin. Chem. Klin. Bioohem. 13, 37 (1975).

[9] Vallbona, C, Pevny, E. and McMath, F., Computer analysis of blood gases and of acid-
base status, Comp. Biomed. Res. 4^, 623 (1971 ).

[10] Englesson, S., Grevsten, S. and Olin, A., Some numerical methods of estimating acid-
base variables in normal human blood with a haemoglobin concentration of 5 g/100 cm^,
Soand. J. Clin. Lab. Invest. 32, 289 (1973).

[II] Visser, B. F. and Maas, A. H. J., The pH-log pC02 diagram of separated human blood
plasma, Clin. Chim. Acta, S_, 850 (1960).

[12] Maas, A. H. J., Kreuger, A. J., Hoelen, A. J. and Visser, B. F., A computer program for
calculating the acid-base parameters in samples of blood using a mini-computer, (Pflugers)
Arah. Ges. Physiol. 334, 264 (1972).

[13] Siggaard-Andersen, 0., The Aoid-Base Status of the Blood, 2nd ed., 134 pp. (Munksgaard,
Copenhagen, 1964).

74
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

"ACID-BASE SEMANTICS"— A CENTURY OF THE TOWER OF BABEL^

Harry F. Weisberg
Division of Biochemistry
Mount Sinai Medical Center
Milwaukee, Wisconsin 53201, USA

Almost 100 years have passed since 1877 when Walter [3]^ originated the concept of
"alkaline reserve" which was estimated by the oarhon dioxide oontent of the plasma. It was
assumed that the alkali of the blood existed mainly in the form of carbonates (or carbonates
and bi carbonates as the terms are used today) and that the quantity of carbon dioxide was
essentially proportional to the quantity of alkali contained in the blood. Walter [3]
stated that "...the carbon dioxide content of the blood will .. .permit of conclusions as to
the quantity of alkali that has been withdrawn from the body as a whole."

Subsequently, alkali (ne) reserve has been used not only to designate the total carbon
dioxide content of blood but also the carbon dioxide capacity, the plasma bicarbonate
(determined by the carbon dioxide combining capacity or carbon dioxide combining power),
the amount of sodium combined with the plasma bicarbonate, and the total cation ("base")
concentration. In 1931, Cummer [4] wrote that the alkali reserve "...by means of which
acids are neutralized, is made up of the bicarbonates , small quantities of phosphates, and
alkaline protein compounds."

The four "CO2" determinations mentioned above are compared to each other and to the
AutoAnalyzerf^ technic in table 1. Skeggs placed the AutoAnalyzer trays in a box to equili-
brate the plasma [5] or serum [6] specimens with carbon dioxide to obtain the CO2 combining
power since it was difficult to achieve the necessary anaerobic conditions during collection
and analysis required for the CO2 content. A modification was described in which each cup
was equilibrated three times just prior to aspiration into the apparatus [7]. The present
technic for the AutoAnalyzer CO2 is termed a "content" since no equilibration is utilized
but, in reality, the result is an approximation of a bicarbonate concentration.

If the PCO2 is 40 mm Hg, the relationship among the determinations is shown in eq. (1)
and (2) for arterial and venous blood, respectively.

CT,^^° = [CY^°°- 0.6] = [CC + 1.2] = [CP + 1.2] (1)

a

CTy = CCY^" - 0.4] = [CC + 1.4] = [CP + 1.4] (2)

If the PCO2 is above 40 mm Hg, the relationship is as in eq. (3);

CT > CC > CY < CP (3)

^Based upon and expanded from Weisberg [1,2] with permission of publishers, Williams and
Wilkins [1] and ASCP [2].
^Figures in brackets indicate the literature references at the end of this paper.

75
 , «

Q QJ
N _
QJ

>^ c \ Q
+->
CM
CO
CM CM Q. CO
o
<_>
O
o 03 0)
(/I >>
o
c
O
e
o
c

o>
~r~

<;,..
E
_Q o CO E
(_j 1 ^ ^
Q O Q~ • ^ 5 (J
C_j Q_ _Q
00 VII VII
Csl C\| CL CO
Q o_ D_
O 03 CO >> c CM 1

0
"O

I/)
CO
Q
>^ " *
•M CO
CM C\J
1 1

CD CO C\J dJ
1 CM
O Q •

^
CD CNl 'N
CD
CO
00 0
CU
C/)
Q o
C\J 1 CM
LO
CM
_ CO

O
-
if)

>^
4_j _^ ^ 00 CTi
CM OJ CM
(_) fO (_) 1 1

LO 4_)
Q_ o _Q CM CM (tJ
4— Q
Q (_) CM
cu CD QJ 00
c~ CM CM CO CO d
o
^_;>
j_
_j_>
cu o Q
CM CM Q.
Q (

csl
Q_ _
Q
Q
s_

c^
^ ^ 0^
p- r— CM CM t/)
^0 (_) 1
4-' CO LO
Q _Q CM CM CO d)
Q —
4_}
^ (J
j
Q ~ 4_)
'
i_

CM
QJ
CO CO
cu (U
c CM 0 to
o o C <D
>^
QJ
>^
Q Q LD
CM CM
fC
4J
(/)
2^ d>
0
^
•r-
•

"O CM
<D O
<->
CO
zr. CO
cn 0^
E CO E •JZ CO CD
&.
CO CM E Q d)
OJ Q E Q
•!-> *+- E 4_) CO
OJ 0
CO Q 13
^
„

|cn + Q
cr
Q
CU
4_) Q
SI
o
o (J
(/I
0
(N
Q_
(J
d) 4_>

CO
-M
CM
o Q t|H
0
Q.
CM
d)
J_
i-
d>
E
TO CM U Q
i-
+-> £Z 0. ^
o CU E CM
_ fO (J C;
CO dJ
_ ' * f ^ S- :3 0.
CU 0 C/) CO
CM +-> c
Q o o
o oQ
cu
• r— i-
i-
QJ
CU Q S- cu
dJ d)
ry.
>> cr O 0 S-
CO 1— (jU +-> <_) 4- -Q 0

76
 —
and if the PCO2 is below 40 mm Hg, the relationship is as in eq. (4).

CY > CC > CT < CP . (4)

In 1883, Stadelmann [8] had formulated the concept of "acidosis" which resulted from
excess acid production Which could then result in alkali deficit. Naunyn [9] expounded on
the use of the term acidosis due to endogenous production of hydtroxyhutyric acid in diabetes,
which was then followed by the secondary alkali deficit or hypoalkalinity or alkalipenia.
Acidosis was used to designate qualitative and/or quantitative changes of acid in the blood
[10,11] but not the "reaction" of the blood [10].

From 1909, however, the consensus [12-15] seemed to revert to the position of Walter
and of Stadelmann, the term "acidosis" being used to indicate any type of acid which could
alter the acid-base balance, resulting in a lowering of the alkaline reserve. If the
respiratory mechanism could decrease the carbonic acid concentration to maintain the
"normal" bi carbonate rcarboni c acid ratio, it was called compensated acidosis^ but if the
carbonic acid could not be reduced sufficiently, with an accompanying increase of hydro-
gen ion (pH decreased), Hasselbalch and Gammeltoft [16] called it uncompensated acidosis.

In 1931, Peters and Van Slyke [17] used the terms 1° alkali deficit and 1° alkali
excess to denote changes in the blood. These can be traced back to 1877 (Walter [3]), and
to 1883 (Stadelmann [8]). Changes in body content were called "base" deficit and "base"
excess [17]. In 1947 (first edition), Davenport [18] modified the terms to describe blood
changes as extra' fixed acid (or acid excess) and extra fixed base (or base excess). Singer
and Hastings [19,20] used combinations of these terms fixed acid excess (or base deficit)
and fixed acid deficit (or base excess); Roos and Thomas [21] also use the terms fixed acid
excess and fixed acid deficit for metabolic acidosis and metabolic alkalosis, respectively.

In 1948, Singer and Hastings [19,20] introduced the determination of buffer base which
describes the buffer system of whole blood. It is the cation equivalent to the sum of
buffer (or "labile") anions ( "base")— bicarbonate, proteinate, and phosphate in the plasma,
and the oxyhemoglobin and other erythrocyte buffer anions [e.g., phosphate) [22,23]. The
term "buffer base" was used by Stadie, et al. [23a] in 1925 to represent the bicarbonate and
protein concentration.

The "non-bicarbonate" buffer base was designated as [Buf"]; the non-buffer (or "fixed")
anions consist of chloride, sulfate, and the organic acids. The pH and total carbon dioxide
content are determined in whole blood, which is corrected for the hematocrit or hemoglobin
content, and the partial pressure of carbon dioxide and "buffer base" calculated from a
nomogram [20] or diagram [19].

The normal buffer base of whole blood of normal hematocrit is 49 mEq per liter (range
46 to 52) and, for plasma, 42 mEq per liter (range 39 to 45). The difference is due to the
much higher buffer (oxyhemoglobin) content in the erythrocytes. Note the similarity of the
definition of alkaline reserve of plasma by Cummer [4] and the buffer base of plasma (omitting
hemoglobin factor for whole blood buffer base). The buffer base is decreased in metabolic
acidosis and increased in metabolic alkalosis. Recently [24], buffer base has been equated
with total bicarbonate.

Buffer base, originally represented as B5+, was modified by Elkinton and Danowski [25]
to BB to signify the body buffers which were equal to the total buffer cations (value is
equivalent to the buffer anions, also termed Buf). The buffer base was the difference
between the fixed (total) cations and the "fixed" anions, necessary to maintain electro-
neutrality [26].

The symbol for buffer base, as used by Owen, et al. [27], is B~; Kintner and Gambino
[28,29] used B for "base". Roos and Thomas [21] use BBB to represent blood buffer base and
suggested the term standard blood buffer base or SBBB to designate the buffer base which
would be obtained if mixed venous blood were brought to PCO2 40 mm Hg in vivo. They utilize
the Singer-Hastings nomogram [20] to derive the conventional BBB and the SBBB; both BBB and
SBBB include the carbamino-COa [21].

77
 .

At a symposium on pH and blood gas measurement, Astrup [72] introduced the term A acid-
base to represent the total amount of surplus acid or surplus base in one liter of blood at
a given concentration of standard bicarbonate [50,51] {vide infra). One could calculate the
amount of sodium bicarbonate or ammonium chloride needed to correct a surplus acid or base,
respectively, by multiplying the a acid-base per liter by the body weight in kilograms times
0.3 (the extracellular space) [31,72].

In 1960, Astrup and his colleagues [30-32] changed acid excess and base excess as
descriptive terms for the nonrespiratory factors to base excess ±. Base excess is defined
as the deviation of the buffer base (aBB) from the "normal" buffer base. Instead of the
actual titration of the blood to pH 7.40 at 37 °C at a PCO2 of 40 mm Hg, it is much easier
to calculate BE [33] or to derive it from the curve nomogram [34], alignment nomogram [35],
or slide rules [36,37].

To avoid the conceptual ambiguity of "negative base excess" Lyons and Moore [38] use
buffer base deviation; Kintner and Gambino [28,29] use base deviation. Suero and Woolf [39]
use the term nonrespiratory acid excess.

Davenport [18] uses the terms base excess (as previously) and base deficit (for his
previous acid excess) for blood changes; these were the terms used by Peters and Van Slyke
[17] in 1931 for changes in body content. Davenport [18] utilized the bicarbonate-pH
diagram for graphic representation of his terms extra fixed acid (or acid excess or base
deficit)--point below the buffer slope of carbon dioxide--and extra fixed base (or base
excess)--point above the line. I have used similar graphic representations [40] on a
modified CO2 content-pH diagram first described by Cullen and Jonas [41] and Hastings et al.
[42]. The descriptive term for the (metabolic) acid-base alteration is delta CO content
^
(ACT), eq. (5).

A CT = Actual CO2 CT - Theoretical CO2 CT (5)

The theoretical CO2 content can be calculated for arterial and venous blood, utilizing a
buffer slope or molar buffer value of 28, from the intercept of actual pH with the CO2
buffer slope or from eqs. (6) and (7).

Theoretical CO2 CT,

a
=25+28 (7.40 - pH) (6)

Theoretical CO2 CT^ =27+28 (7.37 - pH) (7)

The values for arterial and venous theoretical CO2 content also may be obtained from
the table on the Weisberg TRI-SLIDE™ calculator for the Henderson-Hasselbalch equation.
A "negative" delta content (ACT) signifies a bicarbonate (combined CO2) deficit or metabolic
acidosis and is equivalent to the "base excess negative" or "base deficit" whereas a
positive ACT signifies bicarbonate (combined CO2) excess or metabolic alkalosis and is
equivalent to the "base excess positive" or "base excess". Changes in the bicarbonate
concentration may be substituted for the CO2 content changes but only when the PCO2 value is
normal

Since the non-bicarbonate buffers [Buf~] are not directly measurable, Filley [43]
utilized the corrected a[HC03] as the base excess. However, the actual A[HC03]--the dif-
ference between the actual plasma bicarbonate and 24 mEq per liter— seldom differs by more
than 3 mEq per liter from the "base excess" [43]. Collier et al. [44] have applied the
concept of base excess to the extracellular fluid (BE^rp) rather than the blood. It can be
calculated from their bicarbonate-pH diagram (with buffer line of extracellular fluid rather
than that of plasma) or from their formula;

BE^(,p = [HC03]p + e^(,p (pH - 7.4) - 24 (8)

78
in which [HCOslp is the actual plasma bicarbonate and 3ecf "^^ ^"^^^ apparent buffer

value (approximately 12). One can transform this concept to that of "ACT" by altering eq.
(5) to eq. (9).

ECF^^^ = Actual CO2 Content - Theoretical ECF(,q (9)

Similarly, eqs. (6) and (7) are transformed to eqs. (10) and (11).

Theoretical ECF^q = 25 + 12 (7.40 - pH) (10)

Theoretical ECF^q^ = 27 + 12 (7.37 - pH) (11)

Use of eqs. (9), (10), and (11) is much easier than the necessary plotting on diagrams and
finding the vertical displacement, etc.; special slide rules [37,45,46] are easier to use
than the mathematical calculations with formulas.

In 1930, Henderson, et at. [47] introduced the term, T^q? to represent the millimoles
of total carhonic acid per liter of blood when the PCO2 equals 40 mm Hg. The carbon dioxide
content had also been standardized to a pH of 7.40. The formula of Peters and Van Slyke
[17] for the standardized carbon dioxide content is

[C02CT]^u = [CO2CT] + (pH - 7.40) (8.2 + 2.6 Hb) - 0.36 Hb02 (12)
P"7.40

in which the carbon dioxide content, hemoglobin (as oxygen capacity), and oxygen content are
expressed as millimoles per liter. Peters and Van Slyke [17] also had a formula for the
standardized bicarbonate, adjusted to pH 7.40.

[HC03]„M = [HCO^] + (pH - 7.40) (8.2 + 2.3 Hb) (13)

The hemoglobin was expressed in terms of millimoles per liter which can be obtained by
dividing grams percent by 1.67. This equation can be rewritten in terms of grams percent
hemoglobin.

[HC03]p,^^ = [HCO3] + (pH - 7.40) (8.2 + [1.37 x g% Hb]) (14)

It was assumed that the total plasma protein concentration was constant at 7 g per 100 ml.
In 1952, Bunker et al. [48] reported a modified formula for the corrected bicarbonate.

[HC03]^n = tHCO;] + (pH - 7.40) (8.6 + [1.4 x g% Hb]) (15)

Eichenholz, et al. [49] determined the standardized bicarbonate using the intersection at pH
7.40 of a line parallel to the normal carbon dioxide absorption slope.

79
 In 1957, Astrup and his associates [50,51] introduced the term standard bicarbonate,
which is defined as the concentration of bicarbonate in plasma separated from the cells with
the hemoglobin completely oxygenated at a PCO2 of 40 mm Hg and at a temperature of 38 °C.
They use the standard bicarbonate to characterize nonrespiratory ("metabolic") disturbances
because the carbon dioxide content will vary with the PCO2 and P02 present in the blood
sample. The standard bicarbonate is approximately 1.2 mmol per liter less than the carbon
dioxide capacity determined at a PCO2 of 40 mm Hg but on arterial ("completely" oxygenated)
blood at 38 °C (instead of the usual procedure utilizing venous blood at 20 °C). It is not
as useful as changes in buffer base or of base excess [30]. The original [52] and modified
[34] Siggaard-Andersen curve nomogram yields values for standard bicarbonate, etc. In
1967, however, the standard bicarbonate on the nomogram (distributed by Radiometer A/S) was
changed to plasma bicarbonate (at PCO2 40 mm Hg) with the supporting descriptive statement
that this "plasma bicarbonate" is the standard bicarbonate.

Armstrong, et al. [53] prefer T^q bicarbonate (named after L. J. Henderson [47] who
based many of his studies on blood at a CO2 tension (PCO2') of 40 mm Hg)--also equilibrated
to PCO2 40 mm Hg, but as if it occurred in the subject [in vivo) rather than after the
sample is removed {in vitro). The T^g bicarbonate differs from the standard bicarbonate by
the amount of dilution by the interstitial fluid of the bicarbonate generated during hyper-
capnia. As explained above, the standard bicarbonate can be described as a "modification"
of the carbon dioxide capacity; similarly the T^g bicarbonate is similar to the carbon
dioxide combining capacity, the bicarbonate concentration at a PCO2 of 40 mm Hg (determined
on separated plasma without the hemoglobin buffering capacity). It can be calculated from
the Singer-Hastings nomogram [20] (using an "effective" (extracellular) hematocrit of 0.09)
or the Astrup log PCO2/PH diagram [30,52] (using an "effective" hemoglobin of 3 g per 100
ml); such calculations utilize a "dilution factor" of five to obtain the effective hemato-
crit or hemoglobin (based on ratio of blood volume to extracellular fluid and ratio of
buffering capacity of blood and interstitial fluid). In addition, the Singer-Hastings
nomogram and the Astrup diagram have been modified [44], allowing direct estimation of the
T40 bicarbonate without hematocrit or hemoglobin values, respectively. The Ti+q bicarbonate
can be estimated mentally if the PCO2 is known. For "hypercarbia" , eq. (16) is used [53],
whereas for "hypocarbia" , eq. (17) is used [54].

PCO2 - 40
^"'^
[HCO3] l-^^'-'sl
-oivo vitro 15

PCO2 - 40
[HCO^] . . = [HCO;] .
in
^ ^n vt.vo ^ xn v%tro ID

The in vivo buffer curve for carbon dioxide (expressed as [HCO3]) is curvilinear
rather than rectilinear [18,55]. Davenport [18] describes the relationship of bicarbonate
for specific values of PCO2 as eq. (18).

PCO2
[HCO;] = 31.39 X (18)
PCO2 + 12.95

The same curvilinear relationship holds for the carbon dioxide titration curve when expressed
as carbon dioxide content. It is best therefore to estimate ACT or base excess/deficit from
the in vivo buffer data when the pH is below 7.40 ("acidosis") and from the in vitro buffer
data when the pH is above 7.40 ("alkalosis").

In similar fashion, Severinghaus and Bradley [56] state that the blood in vivo behaves
as if the hemoglobin concentration were one-third (e.g., 5 g per 100 ml) of the actual value
since the extracellular fluid is about twice the blood volume and the bicarbonate, generated
from carbon dioxide, diffuses from the blood into the extracellular fluid [sic). With the
Siggaard-Andersen alignment nomogram [35] they use the one-third hemoglobin value on the

80
original line as a fulcrum to connect with PCO2 40 mm Hg (rather than the "original" base
excess with PCO2 40) and thus derive the in vivo standard (plasma) bicarbonate and base
excess (latter read at the actual hemoglobin concentration).

An in vivo carbon dioxide titration curve, based on a buffer capacity of about 5 g per
100 ml hemoglobin was reported by Siggaard-Andersen in 1967 [57] and expanded upon in 1971
[58], giving values for base excess and base deficit, and for bicarbonate at PCO2 40 mm Hg.

Roos and Thomas [21] discuss the "error'" of the conventional standard bicarbonate (SB
in vitro minus SB in vivo) which should be subtracted from the in vitro value to obtain the
estimated in vivo value of standard bicarbonate; the "error" varies with the pH, hematocrit,
and PCO2.

Owen et at. [27] proposed the term appropriate hicarhonate that bicarbonate appropriate
,

to the actual Pao2 present in the patient. The difference between the appropriate and the
actual (observed) bicarbonate values is a measure of the change in plasma buffers.

A Plasma buffers = [HCOs]^^^^^^ - [HCOi] . . (19)

In 1916, Hasselbalch [11] saturated blood with carbon dioxide at 37 °C under 40 mm

tension and determined the "reduced hydrogen ion concentration" (or reduced pH or nonrespi-
ratory pH). Under such conditions the carbonic acid concentration is fixed and the hydrogen
ion concentration must vary inversely as the concentration of (sodium) bicarbonate; thus the
"reduced hydrogen ion concentration" is a measure of the blood bicarbonate [14,14a].

In 1959, Peirce and his colleagues [59,60] determined the metabolic and respiratory pH
factors by a "double pH" method--measuring the actual pH and the euoapnic pH (blood equili-
brated to PCO2 40 mm Hg). If the actual and eucapnic pH values were the "same", a pure
"metabolic" imbalance could be present; a low pH indicated metabolic acidosis and high pH
indicated metabolic alkalosis. If the eucapnic (but not the actual) pH equalled "7.40", a
pure "respiratory" imbalance was present. If the actual pH was lower than the eucapnic pH,
respiratory acidosis was present; conversely, if the actual pH was higher than the eucapnic
pH, respiratory alkalosis was present.

In 1965, Whitehead [61,62] described changes in acid-base balance in terms of hydrogen

ion units, expressed as nanoequi valents (or nanomoles) per liter. The pH values must be
converted into hydrogen ion concentration; the Astrup diagram is used to determine the
hydrogen ion concentration at PCO2 40 mm Hg (with hemoglobin fully saturated). The total
change is the sum of the changes in respiratory and nonrespiratory factors, a[H''"] = R + NR.

The complete equation is.

in which [H ]tot is the actual hydrogen ion concentration for the determined pH at the
actual PCO2 in the patient and [H"*"] PCO2 40 is the hydrogen ion concentration for the pH at
PCO2 of 40 mm Hg. Each term has a "normal" range of ± 4 nmol ; a negative sum means a deficit
of hydrogen ions (alkalosis) and a positive sum, an excess of hydrogen ions or acidosis.

Kintner [63,64] describes acid-base changes in terms of A and B, respectively, the

deviation of PCO2 ("acid") and carbon dioxide content ("base") from their respective normal
values. Subsequently, the "base" portion was equated with changes in buffer base (base
excess and base deficit) and the Siggaard-Andersen alignment nomogram was used to determine
the A and B factors; in addition, the term delta [H"*"] was used to quantitate the acid-base
imbalance [28,29].

81
 The molar buffer value was used by Van Slyke to describe buffering capacity (or buffer
slope [43])--the number of moles of acid needed to decrease pH by one unit in one liter of a
molar solution. In 1965, Woodbury [65] introduced the term Van Slyke or slyke (abbreviated
si) for the molar buffer value, defined as the ratio of the change in base [65] or bicarbo-
nate [43] to the change in pH as the measure of "buffer capacity" of blood due to the pre-
sence of hemoglobin and plasma proteins. Hemoglobin has a buffer value of 3 sl/mmol and
plasma protein has a value of 0.1 sl/g. Instead of converting hemoglobin as grams per
deciliter to millimoles per liter by dividing g/dl by 1.67 or multiplying g/dl by 0.621, the
entire relationship can be described in eq. (21),

sl/1 = (1.86 X Hb g/dl) + [TP g/dl x (1 - Hct)] (21)

in which the Hb is hemoglobin, TP total protein, and Hct hematocrit.

Substitution of "average" values of 14 g/dl and 6 g/dl for hemoglobin and total protein,
respectively, with a hematocrit of 0.45 gives 29.3 sl/1 of blood. Filley [43] gives 29
sl/1 as a normal value for blood of pH 7.3 to 7.5. For "in vivo" conditions as exemplified
in a 70 kg individual with a blood volume of 7.1 percent of body weight ("average" of 7.6
percent for average male and 6.6 percent for average female), there are 146.5 si in the 5 1
of blood. The "interstitial fluid" volume in both sexes is about 12 percent of body weight
and contains an average of 0.66 g/dl protein (with a hematocrit of "zero"); therefore 0.66
g/dl X 10 X 0.1 x 8.4 1 yields 5.5 si. The grand total of 152 si divided by the 13.4 1 of
extracellular fluid yields a concentration of about 11.3 sl/1. Factors of 28 si and 12 si
are utilized in eqs. (6) and (7), and (10) and (11), respectively.

Data for PCO2, [HCO3], and [H"*"] changes have also been presented as index values (sample
value/median normal value), permitting calculation of the compensation (and buffering)
present as a percentage of the (calculated) "uncompensated" increase of hydrogen ion [66].

Table 2 summarizes some schemes for determining and describing acid-base alterations of
the blood. The "old" system consisted only of measuring the alkaline reserve (carbon dioxide
content), a decreased value establishing a diagnosis of acidosis [4]; the possibility that
alkalosis could exist was denied [13].

There are no significant differences among the scheme that I use, or Siggaard-Andersen's
[33] or the report (not recommendations, as originally intended) from the New York Academy
of Sciences [67]. It was agreed that the preferential order for describing an acid-base
alteration is to (a) use the specific numerical data for the test result, (b) describe the
test values as "high, low or normal", or (c) use the general "descriptive" terms.

The ad hoc committee [67] suggested that acidemia and alkalemia be used to describe
alterations of blood pH and to utilize acidosis and alkalosis to describe the overall phy-
siological process or condition which tends to cause a deviation in pH without being depen-
dent upon deviation of the pH per se.

My criticism of acidemia-alkalemia is that these terms refer to blood and leave the pH
of the interstitial fluid (in equilibrium with the intravascular fluid) in limbo. When pH
changes of the intracellular compartment are reported (primarily research), it is so desig-
nated since the normal values differ from that of the "extracellular" fluid. These comments
are even more relevant today with the emphasis on the in vivo titration curve, etc. [44,56-
58]. A possible compromise may be to use eupUemia and eupHuria to represent a normal pH of
blood and urine, respectively, since these terms "are compact, easily spelled, and easily
pronounced" [68]. In a similar vein one could use the terms hypopHemia and hyperpUemia to
designate decreased pH (acidosis and/or acidemia) and increased pH (alkalosis and/or alka-
lemia), respectively.

All agree that the respiratory changes are determined by measuring PCO2. I have
designated the test as PCO2 or [H2CO3]; the latter (used by Peters and Van Slyke) is equal
to 0.03 x PCO2 and emphasizes the carbonic acid term, similar to that of the Henderson-
Hasselbalch equation, avoiding confusion with the various "CO2" determinations. The descrip-

82
 ^- — 1

E
cu
to
o. fO (U
a. E E Q. X3 to
o «3 •>- s- O 10 S- »0
I I
j- cu T3 i- O cu XI
(J 4-1 0) o Q. o
to O a. Q. >> Q.
o cj Q)
Q >> >)
Q. C
(U cu
i- s~

<U 4-
> C o
•I- o
+-> O CU CU to
rO +-> 4-> (/) X3 to
(-> O) fO (O 1- CU
< C
T3
O
C
<U >-
c
O
X5
X3
S. •!-
OXU
fO

cu
O CM
5- cu XI
4-
(O O)
O O M- 01
o •r- 3 4-> ro

O
Q D. X3 X) to X>

>1 >>
CU >> s. s-
> S- i- o o
•1—
+J —
-
to
•»~
10
"1
O
+->
to
•!-
o
+->
-l-> 4->
to to (O
CO
Q. E to —
1 to (0 to (O •!- S- to S- 00
s.
CO
•1- i. o (0 O S- o to •r- 'r— •p- 'r-
i- 0) +-> D +-> X) r- -O •I- O cl to to
o +-> o to O •!- ca.T- Q. (U to O t/l o
c (/) 4->
^— XI
rO 4-> O tO <J to to CU cu cu -o
<u «3 CU ro 1. to S- •!-
<u o S- ^ £=
O
(tj C CJ
to XI O CO
s-
O)
c
X3 oc
cC
I
c
T3
s-
o (/)
« l-> to
n3 0}
a> u
CD •r- X
,
E -<- -> -«- -> + <u 1

S-
i. <u CM (U
U 4-> o (/)

CU 3: o lO
0)
o Q Q. XI

T3
C
a T3 4->
a) 'p- CM CM C
(O >
•r-
to
to
o CJ
ta «
cu
4->
o
O cu
4-) C_)
o CU
4->
D1 4-) V) •r- CM CM ra
C Ql E O to o o
O o o C T3
(O
c -a
oo
O
•E" - O o
•1- I. •r- CU cu 4->
S- CU (0 T3 c to XI c to X5 E -r-
O 4-> o cu to o cu o S- •- to 1- T- (J
1/5
CU
O
fO
X3
s-
lU
s-
cu
o
J3
S_
OJ
S-
._
C
CO
O E
-Q CU (O
o E
X3 •!-
"^-
o
o
s- Q (O
o ^
M- X
O) Si:.
•r-o X
XI O cu
(J
•1-o
XI o
cu
-o
XI
(L)
—
1

ro
C
o
o
o +J
4J
CM c
n: cu
1 1 4-"
c -f- ->-
+ c 1

i- o
i
S-
o o
<u CM CM
+->
(U nz
o
o
o
o<
Q Q.
cu
4->
>, >^ £=
s- S- = o
O)
>
o
4-> to
= O
4-»
to
•--
U iA O
•r- to
o
(O -t- CO </) CO
+->
D. E
S- CO
•r- O
s- o O o
X3
O
XI
to
O
o
C_)
•1- S- CL-O O. ro ro ro ro T3
S- <U to •!— to -li 4-> -ii 4J
O 4-> 0) CJ CU I— CU CU O
<u
_S_ fO S_ ro E ro B ro

o
o
O
cu

CU
4J o
a: cu c c_>
Q. c cuo
•r- > o
r—
ro
S-
CU CM
o
o
I—
to
CU
O
o <
<: s

83
 —
[HCO;] + [Prot"] + [HPO"] + [cr] + [so^] +
[Org Acids"]

Combined CO2 [HCO3]

CO2 combining capacity
CO2 combining power
T
Alkaline resemje^
available anion Fixed anions
available "acid" fixed (non-carbonic) "acids"
gaseous "acid" (anion) non-gaseous "acids" (anions)
volatile "acid" (anion) non-volatile "acids" (anions)
"Base" bicarbonate I

y
Y V
[HCOi] Ibuf"^

Buffer base Non-bicarbonate buffer base

diffusible buffer Non-diffusible buffer base (P")'
base

V V
Alkaline resevve (Cummer)^
Labile anions Fixed anions
Buffer anions (also Buf~) non-buffer anions
Buffer base (Bg+) (B") (B) (BBB)*^
Equivalent to: buffer cations
body buffers (BB)

V
TOTAL ANIONS^

Total "acids"

Additional "definitions" of alkaline reserve:

Total CO2 content: combined CO2 [HCO3] + free CO2 [H2CO3]
CO2 capacity: (combined CO2 + free CO2) at PCO2 40 mm Hg and 20 °C
Portion of [Na"*"]: "base" bicarbonate or bicarbonate-bound "base" or available alkali
or available "base" or available cation.
. Total cations or total "base" or fixed cations or fixed "base."
A Non-diffusible buffer base also includes hemoglobin proteins.
Buffer base of plasma also defined as bicarbonate plus protein. Buffer base of whole
, blood also includes buffer activity of hemoglobin, etc. of erythrocytes.
Total anions minus those anions which have been determined equals RESIDUAL ANIONS or
Residual "acids."

Figure 1. Terms applied to various fraction of total plasma anions modified from Weisberg [2];
duplicated terms in italics.

84
 —
tive terms of the committee for respiratory alterations do not meet with my support. For
conformity with acidemia and alkalemia, they should use hyper- and hypo-capnemia. There is
easy confusion between the words -capnia (Greek root kapnos meaning smoke or vapor) and
-capnea (Greek root pnoe meaning breathing). Some authors have used the terms hyper- and
hypo-carbia (Latin root aarbo meaning coal), which are "barbarisms"— combinations of Greek
prefixes and Latin nouns [69].

It is obvious that the major disagreements are in relation to the diagnosis of the
"metabolic" al terations--especiany which test to use. Utilizing the Br0nsted concept, all
anions are conjugate bases and could fit into the descriptive terms (hyper- or hypo-basemia)
of the committee report, but only the bicarbonate anion and organic acid anions have a
primary effect on the hydrogen ion concentration. Bicarbonate is considered to be part of
the labile "buffer" anions (equivalent to buffer base) whereas the organic acids are part of
the fixed "non-buffer" anions.

Although not intended to confuse the reader, figure 1 shows the interrelationships of
some of the various terms applied to different components of the total anions; some dupli-
cations are self-evident. Though all the anions are "conjugate bases" and can act as a base
by accepting a proton, the "buffer" anions are the major buffers since their pK' values are
closer to the pH of blood (the pK' values of the "fixed" anion systems are lower).

Figure 2 is a simple flow chart for the differential diagnosis of acid-balance imbalance.
Of potential thirteen diagnoses, seven can be distinguished by use of pH and PCO2. To
distinguish between a pure respiratory or a mixed respiratory and metabolic imbalance,
the actual bicarbonate concentration or CO2 content is utilized; and to distinguish between
an acute or chronic respiratory imbalance, the base excess/deficit or delta CO2 content is
necessary.

pH

~\ r
PCO2 and/or [H2CO3] I N
3 4
I 1 I

[CO2 CT] or [HCO^]

mixed vs
8 (9 or 10) n (12 or 13)

respiratory

ACT or BE/D N
acute vs chronic 10 12 13
respiratory

Figure 2. Flow chart for differential diagnosis of acid-base imbalance.

Key to numbers in chart:

1) normal acid-base 8) mixed resp acid and

2) resp acid, comp met acid
(met alk, comp) 9) resp acid, acute
3) resp alk, comp 10) resp acid, chronic or
(met acid, comp) partial comp
4) met acid, acute 11 ) mixed resp alk and
5) met acid, partial comp met alk
6) met alk, acute 12) resp alk, acute
7) met alk, partial comp 13) resp alk, chronic or
partial comp

It is ironic that we have come full circle in the course of a century. As stated by
Filley [43], "Ambiguity in the meaning of words, a major curse of specialization, arises

85
both because words often outlive the concepts they originally stood for and because the same
word is used by different experts to mean different things, i.e., abused." Additional
commentaries are by two poets--ex-Senator Eugene McCarthy [70] said, "If the language is
debased or misused, if the meaning of words is obscured, the basis for common judgment is
undermined, if not destroyed"; and W. H. Auden [71] said "... there is only one political
duty, and that is to defend one's language from corruption. When it's corrupted, people
lose faith in what they hear, and this leads to violence."

If one can communicate without misunderstanding, it does not make much difference which
"terms" are usedl One cannot, however, take the position described by Lewis Carroll in
Through the Looking Glass— "When I use a word, Humpty Dumpty said, in a rather scornful
tone, it means what I choose it to mean. Neither more nor less."

References

[I] Weisberg, H. F., f/ater^ Eleotrolye, and Acid-Base Balance; Normal and Fathologia
Physiology as a Basis for Therapy, 2nd ed. 533 pp. (Third edition in preparation.)
,

(Williams and Wilkins, Baltimore, MD, 1962).

[2] Weisberg, H. F., Antics with "Acid-Base" semantics. Lab. Med. 3, 11, 32 (1972).

[3] Walter, F., Untersuchungen ilber die Wirkung der Sauren auf den thierischen Organismus,
Arch. Exp. Path. Pharmakol. 7, 148 (1877). Quoted by Woodyatt [15].

[4] Cummer, C. I. A Manual of Clinical Laboratory Methods, 3rd ed.

, (Lea and Febiger,
Philadelphia, PA, 1931).

[5] Skeggs, L. T., An automatic method for the determination of carbon dioxide in blood
plasma. Tech. Bull. Registry Med. Teohnol. 30, 1 (1960).

[6] Skeggs, L. T., The determination of carbon dioxide in blood serum, Ann. N.Y. Acad. Sci.
87, 650 (1960).

[7] Marsters, R. W. , An equilibrating device for use in conjunction with the automatic
determination of carbon dioxide-combining power, Clin. Chem. 8_, 91 (1962).

[8] Stadelmann, E., Uber die Ursachen der pathologischen Ammonia-kausscheidung beim Diabetes
mellitus und des Coma diabeticum. Arch. Exp. Path. Pharmakol. ]]_, 443 (1883). Quoted
by Woodyatt [15].

[9] Naunyn, B. , Der Diabetes Mellitus, 2nd ed. (Holder, Vienna, 1906).

[10] Barcroft, J., The Respiratory Function of the Blood (University Press, Cambridge,
1914).

[II] Hasselbalch, K. A., Die "reduzierte" und die "regulierte" Wasserstoffzahl des Blutes,
Biochem. Z. 74, 56 (1916).

[12] Henderson, L. J., Acidosis, Trans. Assoc. Amer. Physicians (1916). Quoted by Woodyatt
[15].

[13] Sel lards, A. W., The Principles of Acidosis and Clinical Methods for Its Study
(Harvard University Press, Cambridge, 1917).

[14] Van Slyke, D. D. and Cullen, G. E. , Studies of acidosis, I. The bicarbonate concen-
tration of the blood plasma; its significance, and its determination as a measure of
acidosis, J. Biol. chem. 30, 289 (1917).

[14a] Frumin, M. J., A simplified approach to acid-base analysis, Anesthesia Rounds 3^,

(5), 1 (1967).

[15] Woodyatt, R. T., Foundations of the conception of acidosis, Proo. Amer. Diabetes
Assoc. 8, 17 (1948).

86
[16] Hasselbalch, K. A. and Gammeltoft, S. A., Die Neutralitatsregulation des graviden
Organismus, Bioahem. Z. 68, 206 (1915).

[17] Peters, J. P. and Van Slyke, D. D. , Quantitative Clinical Chemistry^ Vol. 2, Interpre-
tations, (Williams and Wilkins, Baltimore, MD, 1931).

[18] Davenport, H. W., The ABC of Acid-Base Chemistry^ 6th ed., (University Chicago Press,
Chicago, IL, 1974).

[19] Singer, R. B., A new diagram for the visualization and interpretation of acid-base
changes, Amer. J. Med. Sci. 221_, 199 (1951).

[20] Singer, R. B. and Hastings, A. B., Improved clinical method for estimation of distur-
bances of acid-base balance of human blood. Medicine^ Z7_, 223 (1948).

[21] Roos, A. and Thomas, L. J., Jr., The in-vitro and in-vivo carbon dioxide dissociation
curves of true plasma. Anesthesiology, 28, 1048 (1967).

[22] Singer, R. B., Acid-Base Balance, in Biology Data Book, P. L. Altman and P. S. Dittmer,
eds., pp. 262-263 (Federation of American Societies for Experimental Biology, Washington
D. C, 1964).

[23] Yeomans, A. and Stueck, G. H., Jr., Clinical-chemical studies of acid-base abnormalities
changes in acid-base balance observed in renal and respiratory disease, Amer. J. Med.
13, 183 (1952).

[23a] Stadie, W. C, Austin, J. H., and Robinson, H. W., The effect of temperature on the
acid-base-protein equilibrium and its influence on the CO2 absorption curve of whole
blood, true and separated serum, J. Biol. Chem. 66, 901 (1925).

[24] Bessmann, A. N. and Tsao, J. M. , Calculation of anion gap as a therapeutic tool in

diabetic acidosis, Clin. Res. 19^, 347 (1971).

[25] Elkinton, J. R. and Danowski , T. S., The Body Fluids: Basic Physiology and Practical
Therapeutics (Williams and Wilkins, Baltimore, MD, 1955).

[26] Barker, E. S. and Elkinton, J. R., Hydrogen ions and buffer base, Amer. J. Med. 25,
1 (1958).

[27] Owen, J. A., Dudley, H. A. F., and Masterton, J. P., Acid-base status assessed from
measurement of hydrogen-ion concentration and PCO2, Lancet, 2^, 660 (1965).

[28] Gambino, S. R. , Water, Electrolytes, Acid-Base, and Oxygen, in Todd- Sanford's Clinical
Diagnosis by Laboratory Methods, I. Davidsohn, and J. B. Henry, eds., 14th ed.,
pp 645-672, (W. B. Saunders, Philadelphia, PA, 1969).

[29] Kintner, E. P. and Gambino, S. R. , Acid-Base Balance: A New Look, ASCP Clinical
Chemistry "Check Sample" No. CC-44 (April 1967).

[30] Astrup, P., J0rgensen, K. , Siggaard-Andersen, 0., and Engel , K. , The acid-base metabo-
lism; a new approach. Lancet, 1_> 1035 (1960).

[31] Mellemgaard, K. and Astrup, P., The quantitative determination of surplus amounts of
acid or base in the human body, Scand. J. Clin. Lab. Invest. 1_2, 187 (1960).

[32] Siggaard-Andersen, 0., Engel, K. , J0rgensen, K. , and Astrup, P., A micro method for
determination of pH, carbon dioxide tension, base excess and standard bicarbonate in
capillary blood, Scand. J. Clin. Lab. Invest. 12^, 172 (1960).

[33] Siggaard-Andersen, 0., The Acid-Base Status of the Blood, 2nd ed. (Williams and Wilkins,
Baltimore, MD, 1964).

87
[34] Siggaard-Andersen, 0., The pH-log PCO2 blood acid-base nomogram revised, Soand. J.
Clin. Lab. Invest. U, 598 (1962).

[35] Siggaard-Andersen, 0., Blood acid-base alignment nomogram. Scales for pH, PCO2, base
excess of whole blood of different hemoglobin concentrations, plasma bicarbonate and
plasma total CO2, Scand. J. Clin. Lab. Invest. 15, 211 (1963).

[36] Rattenborg, C. C, Blood Gas Evaluator, (Graphic Calculator Co., Barrington, 111.,
1970) .

[37] Severinghaus , J. W. , Blood gas calculator, J. Appl. Physiol. 21_, 1108 (1966).

[38] Lyons, J. H. , Jr. and Moore, F. D. , Post-traumatic alkalosis: incidence and pathophy-
siology of alkalosis in surgery. Surgery^ 60, 93 (1966).

[39] Suero, J. T. and Woolf, C. R., An equation for calculating "derived" acid-base para-
meters. Can. J. Physiol. Pharmacol. 45, 891 (1967).

[40] Weisberg, H. F., A better understanding of anion-cation ("acid-base") balance, scienti-

fic exhibit presented at annual meeting of 111. Med. Soc, Chicago, 111., May, 1956;
and Surg. Clin. N. Amer. 39, 93 (1959).

[41] Cullen, G. E. and Jonas, L., The effect of insulin treatment on the hydrogen ion
concentration and alkali reserve of the blood in diabetic acidosis, J. Biol. Chem. 57,
541 (1923).

[42] Hastings, A. B., Neill, J. M., Morgan, H. J., and Binger, C. A. L., Blood reaction and
blood gases in pneumonia, J. Clin. Invest. 1_, 25 (1924-25).

[43] Filley, G. D. , Aoid-Base and Blood Gas Regulation (Lea and Febiger, Philadelphia, PA,
1971) .

[44] Collier, C. R. Hackney, J. D. , and Mohler, J. G., Use of extracellular base excess in
,

diagnosis of acid-base disorders: a conceptual approach, Chest, 61^, 6S-12S C1972).

[45] Weisberg, H.F., Calculator for the Henderson-Hasselbalch equation, in Manual for
Procedures for the Applied Seminar on Clinical Pathology of Respiratory Diseases ^
F. W. Sunderman, ed., pp. 239-245 (Institute for Clinical Science, Philadelphia,
PA, 1972).

[46] Weisberg, H. F., Water and electrolytes, acid-base, and oxygen, in Todd-Sanford's
Clinical Diagnosis by Laboratory Medicine, I. Davidsohn, and J. B. Henry, eds. pp.
772-803 (W. B. Saunders, Philadelphia, PA, 1974).

[47] Henderson, L. J., Bock, A. V., Dill, D. B., and Edwards, H. T. , Blood as a physicochemi-
cal system; IX. The carbon dioxide dissociation curves of oxygenated human blood, J.
Biol. Chem. 87, 181 (1930).

[48] Bunker, J.P., Brewster, W. R. Smith, R. M. , and Beecher, H. K. , Metabolic effects of

,

anesthesia in man; III. Acid-base balance in infants and children during anesthesia,
J. Appl. Physiol. 233 (1952-53).

[49] Eichenholz, A., Mulhausen, R. 0., Anderson, W. E., and MacDonald, F. M., Primary
hypocapnia--a cause of metabolic acidosis, J. Appl. Physiol. V7> 283 (1962).

[50] Astrup, P., A simple electrometric technique for the determination of carbon dioxide
tension in blood and plasma, total content of carbon dioxide in plasma, and bicarbonate
content in "separated" plasma at a fixed carbon dioxide tension (40 mm Hg), Scand. J.
Clin. Lab. Invest. 3, 33 (1956).

[51] J0rgensen, K. and Astrup, P., Standard bicarbonate, its clinical significance, and a
new method for its determination, Scand. J. Clin. Lab. Invest. 9_, 122 (1957).

88
 . ,

[52] Siggaard-Andersen, 0., and Engel , K. , A new acid-base nomogram; an improved method for
the calculation of the relevant blood acid-base data, Sound. J. Clin. Lab. Invest. 12^,
177 {I960).

[53] Armstrong, B. W., Mohler, J. G,, Jung, R. C. , and Remmers, J., The in vivo carbon-
dioxide titration curve, Ldncet 1_> 759 (1966).

[54] Mohler, J. G., personal communication (1972).

[55] Rispens, P., Zijlstra, W. G., and Van Kampen, E. J. Significance of bicarbonate for the
evaluation of nonrespiratory disturbances of acid-base balance, Clin. chim. Acta 54,
335 (1974).

[56] Severinghaus , J. W. and Bradley, A. F., Blood Gas Electrodes or What the Instructions
Didn't Say (1969)

[57] Siggaard-Andersen, 0., Therapeutic aspects of acid-base disorders, in Modem Trends in

Anesthesia^ Vol. Z, Aspects of Metaboliem and Pulmonary Ventilation, F. T. Evans, and
T. C. Gray, eds., pp. 99-131 (Butterworths London, 1967).
,

[58] Siggaard-Andersen, 0., An acid-base chart for arterial blood with normal and path-
ophysiological reference areas, Soand. J., Clin. Lab. Invest. 2J_, 239 (1971 ).

[59] Peirce, E. C, II, Further development of a simplified method for determining metabolic
and respiratory pH factors. Trans. Amer. Soc. Artif. Intern. Organs, 6^, 240 (I960).

[60] Peirce, E. C, II, Effects of 2-Amino-2-hydroxymethyl-l ,3-propanediol (Tris) during

cardiac bypass procedures, Ann. N.I. Acad. Sci. 92^, 765 (1961).

[61] Whitehead, T. P., Ph.D. Thesis, U. of Birmingham (1964).

[62] Whitehead, T. P., Acid-base status, pH and PCO2 (letter to editor). Lancet, 2_, 1015
(1965).

[63] Kintner, E. P., The A/B ratio: a new approach to acid-base balance, Amer. J. Clin.
Path. 47, 614 (1967).

[64] Kintner, E. P., Calculation of the A/B ratio: a new alignment nomogram. Tech. Bull.
Registry Med. Teohnol. 38, 254 (1968).

[65] Woodbury, J. W., Regulation of pH, in Physiology and Biophysics, T. C. Ruch, and
H. D. Patton, eds., 19th ed., pp. 899-934 (W. B. Saunders, Philadelphia, PA, 1965).

[66] Quintero-Atencio, J., Index values in acid-base characterization, Clin. Res.

(1971).
—
19, 545

[67] Report of ad hoc committee on acid-base terminology, in Current Concepts of Acid-Base

Measurement, C. G. Nahas , ed., Ann. N.Y. Acad. Sci. 133_, 251 (1966).

[68] Schmidt, J. E., Medical lexicographer; normal pH, Mod. Med. p 119 (June 29, 1970).

[69] Lamont, A., "Hypercapnia" versus "Hypercarbia" , Anesthesiology, 22, 324 (1961).

[70] McCarthy, E., quoted in Hechinger, G., The insidious pollution of language. The Wall
street Journal (October 27, 1971).

[71] Auden, W. H., quoted in Time (November 1, 1971).

[72] Astrup, P., Ultra-micro-method for determining pH, PCO2 and standard bicarbonate in
capillary blood, in A Symposium on pH and Blood Gas Measurement, R. F. Woolmer, ed.
pp. 81-93 (Little, Brown & Co., Boston, MA, 1959).

89
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

TRI-SLIDE™ CALCULATOR FOR HENDERSON-HASSELBALCH EQUATION

AND C02RREC°t-02-SLIDE™ FOR TEMPERATURE CORRECTIONS
OF pH, Pc02, AND P02

Harry F. Weisberg
Division of Biochemistry
Mount Sinai Medical Center
Milwaukee, Wisconsin 53201, USA

The general form of the Henderson-Hassel balch equation

may be rewritten for the bicarbonate:carbonic acid system with the usual apparent pK' of
6.10 (in whole blood or plasma at 37 °C body temperature) as eq (2).

[HCOi]
pH = 6.1 + log
[H2CO3]
The total carbon dioxide content (CO2 CT) is essentially the sum of the bicarbonate {oomhined
CO2) and carbonic acid {free CO2); the latter is proportional to the partial pressure of
carbon dioxide (PCO2 symbol as used by respiratory physiologists) and includes the dissolved
(free) carbon dioxide.

[CO2 CT] = [HCO3] + [H2CO3] (3)

[HCOi] = [CO2 CT] - [H2CO3] (4)

[H2CO3] =aPc02 = 0.03 PCO2 (5)

Since [HCU3] and [H2CO3] are not directly determined, eq (4) and (5) are substituted
into eq (2) giving rise to eq (6).

[CO2 CT] - 0.03 PCO2

pH = 6.1 + log (6)
0.03 PCO2

1. TRI-SLIDE

If any two of the three unknown quantities in eq (6) are known, the third can be
calculated [1-3]^. The TRI-SLIDE Calculator (fig. 1) (first designed in 1963 with revisions
in 1965, 1969, and 1971) has been redesigned; it reduces all calculations to one setting
of the special slide rule [4-6].

The calculator. has a table summarizing the three possible settings. If the pH and PCO2
and/or carbonic acid values are known, setting the PCO2 value on the left hand movable scale
(or the [H2CO3] on right hand movable scale) opposite the pH value on center section allows
one to read, opposite the red arrow (at 7.61 on pH scale), the [CO2 CT] on the left hand

Figures in brackets indicate the literature references at the end of this paper.

91
scale. If the pH and [CO2 CT] values are known, the PCO2 and/or [H2CO3] values are obtained
by setting the [CO2 CT] on left scale opposite red arrow and reading the PCO2 and/or [H2CO3]
opposite the pH. Finally, the pH can be determined if the [CO2 CT] and PCO2 and/or [H2CO3]
values are known; the [CO2 CT] is set opposite the red arrow and the pH read opposite the
PCO2 and/or [H2CO3].

Normal average values (and ranges) for adult males and females for arterial and venous
blood are given in a table on the calculator (fig. 1); in addition, "normal" ranges are
bracketed in red and blue (for arterial and venous blood, respectively) on the various slide

[H2C03]

7.32 7.31 7.35- 7.34- X CD

7 7.45) 7.44)
741)
42)
SX mo gi

CD
If*
S OT y 0
1-5
39 42 3 2 sill*
52; 55]
32-42) 35-45)

n z
0 §? fa 3
V: A:

1.11
> C/3
-ml
1.29

(i.i;
138(1

(0.96
1.20(1.0!
|]z <
>
2(
) i-1.35)

M.561
1-1.26) C
m
1.65J
v> A a% ^
q - a> 'J
<: a* > S aw r
0 > Sd3 g.

1 fS
1 p TO
>
?.

•"E
Z
m §5
<: » It
3.?
CO
-d
^E -HE
0
-2 .0
e
TO
u
+2)
1?

KNOW SET OPPOSITE OPPOSITE

pH and Pco? RCO2 or
pH [CO2 cri
or[H2C03] [H2CO3]
PCO2 or
pH and [COj CT] [CO2 CT] pH
[H2CO3]

[CO2 CT] and PCO2 [CO, CD PCO2 or pH

[H2CO3]
[H2CO3] = 0.03 X PCO2 [HCOsT = [CO2 Cri - [H2CO3]

wvimti fflVlVD

Iheo.COjCI ItiM.COjCT
(fflniol/l) [« + ] [Hcr] (mmol/D

(m\/\) pH [H2CO3]

3€i 37.4 100 7.00 8.0/1 29.8 31.4

Figure 1. TRI-SLIDE calculator. = 25 + = 25 + :2 (7 40-pH)

1, Theor. COj CT 28 (7.40-pH) 3. Theor.ECFMzCTirt
2 TlieorC02CT„„ = 27 + 28 (7 37-pH) 4. Theor. ECFc02(rr,„ = 27 + 12 (7-37-pH)

6UID£ TO THERAPY
The following formula (based on extracellular volume of 25% of body weight) is used
fluid
as a "guMe" to the parenteral therapy of the Bicarbonate Deficit or Bicarbonate Excess com-
pofKnt. It is best to administer hall of the calculated volume and reevaluate the pttitirt before
therapy is continued.
Negative answer denotes: ml 1/6 molar Sodium lactate (or bicarbonate,
acetate, citrate, gluconate, etc.) for Bicarbonate Deficit ("metabolic acido-
sis" or "Base Deficit").
1.5 X kg X JCT = OR
Positiveanswer denotes; ml 1/6 molar Ammonium chloride for Bicarbonate
Excess ("metabolic alkalosis" or "Base Excess").
iCT("BE/D") = Actual CO2 CT - Theoretical CO2CT

SOLUTION mmol/l FACTOR SOLUTION mmol/l FACTOR

"1/6 molar" 166 1.5 4.2% NaHCOs 500 0.5
5.0% NaHCOj 595 0.4
Gastric or «3 70 3.5 7.5% NaHCOa
(for NH^n (50 ml "ampoul") 892 0.28

2.14% NH,CI 400 0.6 8.4% NaHCOs

(40 ml "ampoul") 1000 0.25

C1I974 AMERICAN Slid*-Charl Cor WhMlon, III. «0I87. frinl*d In U.S.A.

92
rule scales. The carbonic acid, bicarbonate, CO2 content, and ACT (defined below) are
expressed as millimoles per liter (mmol/1). The values for PCO2 are given in terms of
pressure (P) as millimeters of mercury (mmHg) and have been expressed as "torr" units; they
also can be expressed (in keeping with the recommendation of the lUPAC and IFCC) in terms of
pressure (p) as kilopascals (kPa).

One pascal (Pa) equals 1 newton (N) per square meter (N/m^) ; 1 bar equals 100 000 N/m^
or 100 kPa; 1 mbar equals 0.1 kPa or 0.75 mmHg whereas 1 mmHg (or 1 "torr") equals 1.333 22
mbar or 0.133 322 kPa; and 1 kPa equals 7.5 mmHg or 10 mbar. Equation (5) gave the concen-
tration of carbonic acid utilizing the factor of 0.03 (at 37 °C) and the partial pressure of
carbon dioxide (PCO2) expressed in mmHg; for the "proposed" SI notation, eq (5) can be
rewritten as

[H2CO3] = 0.23 PCO2 (7)

in which the partial pressure of carbon dioxide (symbols pC02 or PCO2) is expressed in kPa.
Table 1 compares the gas values in terms of mmHg and kPa.

The delta content (ACT) has the same connotation as Base Excess/Deficit.

ACT (or "BE/D") = Actual [CO2 CT] - Theoretical [CO2 CT] (8)

Table 1. Comparison of "gas" values.

PRESENT (mmHg) "PROPOSED" (kPa)

(mmHg = 7.5 x kPa) (kPa = 0.133 X mmHg)

A: 40 (35-45)M A: 5.3 (4.7-6.0)M

37 (32-42)F 4.9 (4.3-5.6)F
PC02
V: 46 (42-55)M V: 6.1 (5.6-7.3)M
43 (39-52)F 5.7 (5.2-6.9)F

A: 95 (75-100) A: 12.7 (10-13.3)

P02
V: 40 (30-50) V: 5.3 (4.0-6.7)

P50 (T50) "26.6" "3.5"

A positive delta content (ACT) signifies Bicarbonate or Combined CO2 Excess (metabolic
alkalosis) and is equivalent to Base Excess (BE) or "Base Excess Positive." A negative
delta content (ACT) signifies Bicarbonate or Combined CO2 Deficit (metabolic acidosis) and
is equivalent to Base Deficit (BD) or "Base Excess Negative."

The "theoretical" [CO2 CT] for arterial or venous blood [in vitro and in vivo or
"extracellular") can be obtained from (a) the window on the calculator (which also gives the
equivalent hydrogen ion concentration as nanomoles per liter (nmol/1) and the ratio of
bicarbonate.-carbonic acid for various pH values); (b) e.g..

Theoretical CO2 CT 25 + 28 (7.40 pH) (9)

art
Theoretical CO 2 CT 27 + 28 (7.37 pH) (10)
ven
Theoretical ECF 25 + 12 (7.40 pH) (11)
CO 2 CT
art
Theoretical ECF 27 + 12 (7.37 pH); (12)
CO 2 CT
ven

93
or (c) the Weisberg Acid-Base Balance Evaluation Diagram (fig. 2) from the intersection of
the actual pH of the specimen and the in vitro or in vivo respiratory buffer lines.

6.8 6.9 7.0 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8

0 '

I
M t I
I
I I I I
I
I I I 1^1 I I I
I
I I I I
; I I 1 I I I 1 I I
I
I 1 iTl I I I I
I
I I I I 1

pH 6.a *.9 7.0 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8
[H ^ ] nmol/l 159 126 100 80 63 50 40 32 25 20 16
IMCOj- S *^ JL i2 'iJ i5 20 2S 32 40 SO
(H,CO,l
l

T I 1 1 1 1 I I I 1 1

ACID-BASE BALANCE EVALUATION DIAGRAM (1974 REVISIONS

Figure 2. Acid-base balance evaluation diagram.

|

The buffering capacity--the number of moles of acid needed to decrease pH by one unit |

in one liter of a molar solution— was termed the molar buffer values by Van Slyke. Wood-
j

bury [7] introduced the term Van Slyke or slyke (abbreviated si) for the molar buffer ,

value, defined as the ratio of the change in base [7] or bicarbonate [8] to the change in pH i

as the measure of "buffer capacity" of blood due to the presence of hemoglobin and plasma I

proteins. Hemoglobin has a buffer value of 3 sl/mmol (or 1.86 si per g/dl) and plasma
protein has a value of 0.1 sl/g. The entire relationship can be described in eq (13), |

sl/1 = (1.86 X Hb g/dl) + [TP g/dl x (1 - Hct)] (13) j;

in which the Hb is hemoglobin, TP total protein, and Hct hematocrit. |

Substitution of "average" values of 14 g/dl and 6 g/dl for hemoglobin and total protein,
respectively, with a hematocrit of 0,45 gives 29.3 sl/1 of blood. Filley [8] gives 29 sl/1 i

94 :

I
as a normal value for blood of pH 7.3 to 7.5. For "in vivo" conditions as exemplified in a
70 kg individual with a blood volume of 7.1 percent of body weight ("average" of 7.6 percent
for average male and 6.6 percent for average female), there are 146.5 si in the 5 1 of
blood. The "interstitial fluid" volume in both sexes is about 12 percent of body weight and
contains an average of 0.66 g/dl protein (with a hematocrit of "zero"); therefore 0.66 g/dl
X 10 X 0.1 X 8.4 1 yields 5.5 si. The grand total of 152 si divided by the 13.4 1 of extra-
cellular fluid yields a concentration of about 11.3 sl/1. Factors of 28 si and 12 si are
utilized in eqs (9) and (10), and (11) and (12), respectively, and for the theoretical CO2
content values on the TRI-SLIDE table (fig. 1) and in the in vitro and in vivo buffer lines
on the Weisberg Acid-Base Balance Evaluation Diagram (fig. 2).

The in vivo buffer curve for carbon dioxide (expressed as [HCO3]) is curvilinear rather
than rectilinear [9-11] and expressed as eq (14) by Davenport [9] to give the value for
bicarbonate at specific values of PCO2 (which fall along the buffer curve).

PCO;
[HCO3] = 31.39 X (14)
PCO2 + 12.95

The same curvilinear relationship holds for the carbon dioxide titration curve when expressed
as [CO2 CT]. It is best to estimate the ACT (or BE/D) from the in vivo buffer data when the
pH is below "7 AO" ("acidosis") and from the in vitro buffer data when the pH is above
"7.40" ("alkalosis").

2. Diagnosis

Seven out of 13 possible conditions of acid-base imbalance can be diagnosed (fig. 3)

from the pH and PCO2 and/or [H2CO3]. The CO2 content or the actual bicarbonate value is
used to distinguish a "mixed" imbalance from respiratory conditions; and the ACT or base
excess/deficit (BE/D) is needed to differentiate between acute and chronic (partial compen-
sated) respiratory conditions. The ACT or BE/D is also used in calculating the amount and
type of fluid therapy required.

pH

PCO2 and/or [H2CO3]

[CO2 CT] or [HCOi]

Ni
mixed vs respiratory 11 (12 or 13)
8 (9 or 10)
ACT or BE/D
t N
acute vs chronic respiratory
9 10 12 13

Figure 3. Flow chart for differential diagnosis of acid-base imbalance (see fig. 4)

Key to numbers in chart:

1 normal acid-base
2 resp acid, comp (met alk, comp)
3 resp alk, comp (met acid, comp)
4 met acid, acute
5 met acid, partial comp
6 met alk, acute
7 met alk, partial comp
8 mixed resp acid and met acid
9 resp acid, acute
10 resp acid, chronic or partial comp
11 mixed resp alk and met alk
12 resp alk, acute
13 resp alk, chronic or partial comp

95
 The same diagnostic area numbers in figure 3 are used for diagnosis with the evaluation
diagram as exemplified in figure 4. The "average line" of compensation is shown rather than
overlapping "areas."

6 8 6 9 7 0 7 1 7 2 7 3 7 4 7.5 7 6 7 7 7 8
'
I I I I I I I I I
I
I I I I I I I I J I I I I I I L I I I I I I I I I 1 I I I I III I I I I

I
M I I
I
M F I
I
I I ! 1^1 I I I
I
M M M I
I I
I
I I M I
I I IT| I I I I
|
I I I

PH 6.8 6.9 7.0 7 1 7.2 7.3 7 4 7.5 7 6 7 7 7.8

[H ) nmol/l 159 63
IHCO, 12 i
Ralio
IM,CO,l 1

ACID-BASE BALANCE EVALUATION DIAGRAM (1974 REVISIONl

Figure 4. Acid-base diagram with lines for diagnosis and "confidence limits" (see fig. 3)

3. Therapy

The physician should utilize the laboratory data in the management of a patient with
acid-base imbalance. Equation (15), based on an extracellular fluid volume of 25 percent of
body weight, is used as a "guide" to the parenteral therapy of the Bicarbonate Deficit or
Bicarbonate Excess component.
Negative answer denotes: ml 1/6
molar sodium lactate (or bicarbon-
ate, acetate, citrate, gluconate,
etc) for Bicarbonate Deficit
(metabolic acidosis or "Base
Deficit")
1.5 X kg X ACT (or "BE/D") = <! OR (15)

Fositive answer denotes: ml 1/6

molar ammonium chloride for Bicar-
bonate Excess (metabolic alkalosis
or "Base Excess")

The 1.5 "therapy factor" applies to 1/6 molar solutions— about 166 mmol of cations and of
anions (see table 2 for factors for other solutions)! It is best for the physician to
administer half of the calculated volume and reevaluate the patient before therapy is
continued [1 ,12].

96
 1

Table 2. Factors for parenteral fluid therapy.

9(11 IITTflN
OULU 1 Uli 1 lltlllu 1 / 1
mmn
IMMKJ 1/1
/ 1 1

"1/6 molar" 166 1.5 4 2% NaHCOs 500 0.5

5 0% NaHCOg 595 0.4
Gastric or #3 70 3.5 7. 5% NaHCOa 892 0.28.
(for ml) (50 ml "ampoul")
2.14% NH4CI 400 0.6 8. 4% NaHCOa 1000 0.25
(40 ml "ampoul")

4. C02rrec°t-02-Slide

The pH, Pco2, and P02 values are usually determined at 37 °C with the instruments
available in the clinical laboratory. This slide rule (fig. 5) is used to determine the
respective values at 37 °C if they were determined at other temperatures, e.g., water bath
inaccurate or under special conditions. It, is also used to determine the values "corrected"
to the temperature of the patient.

37»C
40
TEMPERATURE °C
C02RREC*1-02-SLIDr -
lihlililililijH|li|ili|ilili|ilililili{ihlilili|iliTililil)lilililil
[I l|l'
Designed 1974 By Harry F. Weisbcrg, M D.
I I I I
lll|lll[|llll||lll|llf
I I I I I I I

18 20 25 30 35 40 45 50 60 70 80 On respective scale, set deter-

6 90 100
mined value opposite tempera-
lure of test (usually 37 C),
37°C opposite temperature ol pa-
45 40-W-3S 30 25 20 15 10 tient, read "corrected" value
TEMPERATURE CONVERSION TEMPERATURE "C for PcO; or pH,

75 80 85 ii|iiii|iiii]iln|iili|lili|liil|ihl|i(ii|iili(li(i|iiilp^
°F 50 55
i
lilili|i
60
i|ililil
65 70
lilililiKlilil|i|i'lilili ih
^
(WHOLE BLOOD) 7 40 7.50 7.60 7.70 7 80 7 90 8.00
| | | ,

"C 10 15 20

"F 85 90 95 100 105 110 115 °F -25 0 +25 *J^Set PO; from -c^ opposite ACT or BE/D; opposite :ero, read
ACT or BE/D "final" corrected Poj.
kii|i||if^
nil liii|ini|iii{|iii

® P02 ("FINAL"
CORRECTION) ;o 80 90 100
Delermine
value Xl-
% 0; saturation on iT; from ~ "corrected" Po,

37°C
SET 10 15 20 25 30 3^''*^ *° *^
C 1
_
"7
I CONDITIONS
FOR Poi
I lilllllilllllilllilililil Jil|{ili{llllili|lii|ili|lll|ll||ll|lli|lil|^
|iiii|iiil|iiil|nil|iiil|iiii|iiii|iiil|iiilp^
(
J DHERMINATION
6.6 6.7 6.8 6.9 7.0 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 8.0
Poi CORRECTED 25 30 35 40 45 50 60 70 80 90 100 150 200 250 300 350 400 500 600
I I
I I
I I
I I I IIm!||I|I|II|! I I 1 I ll Illllllllllltl|llll1lllllnillllllllllll
I I I I llllll I I I I I I I I I I II II I IliiiliiiilnilliiiJiiN
I I I

I I I I I I I I
I
I
I
I lll|llli|ini|llii|iH>|
I
, I li|>lll|llli|lnl|liil|i<ii|Mil|.lil|Mii|ini|i|,,i|i||| I , , I I lll|llll|illi|liii|.ili|iltl|iiM|
Poj DETERMINED I I

25 30 35 40 45 5 0 60 70 80 90 10 0 150 200 250 300 350 400 500 600 mmHg

P02»T37°,7.«I,OBE/D
I tjjn ll j|ij|llilli|i|^l M|ilii'iilii|iljiiiLijij|imjiiiiliii|l J l|li|il^ I I I I I I I I 1 1 1 1 1 1 liiiiliiiiliiijliiiiliiiij
f ll I M ll|llll|l l lf|f I

1
1 1 1
I
I
j
[
j

% 0; SATURATION
10
[ I

15 20 25 30 40 50 60 95 96 97 98 98.5 99.5 99.6 99.7 99.8 99.9 %

Figure 5. C02rrec°t-02-Slide.

It is imperative to correct pH, PCO2, and P02 values to the patient's temperature
rather than assuming 37 °C as for the thermostated water bath. Random patients were tested
with an IVAC Model 811 electronic thermometer (°F) with the oral probe placed into the
sublingual pocket for about 15 seconds. The accuracy of these thermometers is 0.15 °F in
contrast to 0.5 °F for glass (mercury) thermometers. Figure 6 shows the range of temperatures
found in 383 patients--94. 5 to 101.7 °F (or 34.8 - 38.5 °C); the average temperature was
97.6 °F or 36.5 °C. The "discrepancy" will be greater with patients subjected to hypothermia
or those with high fevers.

Equation (16) is utilized on the C02rrec°t-02-sl ide to calculate the pH of whole blood at
the patient's temperature (pHo^.) from the pH value as it was determined, usually at 37 °C
^
(pH o) [13].
37

WB pHo^ = pH o + 0.0147 (37 - °t) .

(16)

97
 0

Since plasma or serum [14] and cerebrospinal fluid contain less protein, the factors are
different and the respective formulas are given in eqs (17) and (18).

P/S pHo^ = pH o + 0.012 (37 - °t) (17)

^ 37

CSF pHo^. = pH c + 0.003 (37 - °t) (18)

t 37
90"

80"

70 —

60-

50-

40-

30^

20 —

10
Figure 6. Range of temperatures
found in 383 patients.

94.1.'94.6- 95. 1^95.6-' 96.1 -'96..6-*97.1-'97.6-'98.1-'98.6-' 99. 1-' 997^^*100. l-'lOO.eilOI. 1^101 .1'
94.5 95.0 95.5 96.0 96.5 97.0 97.5 98.0 98.5 99.0 99.5 100.0 100.5 101.0 101.5 102.
•F

In addition, the actual P02 of the patient can be determined if the percentage oxygen
saturation ("corrected") and the pH at the actual temperature of the patient are known.
Finally, the C02rrec°t-02-sl ide allows one to correct the P02 to the temperature of the
patient, and to correct P02 values to 37 °C, 7.40, and zero ACT (BE/D) for entrance into
the oxyhemoglobin dissociation curve. The Po2/% O2 saturation scale (F) on the slide rule
applies to "adult" hemoglobin A; the oxyhemoglobin dissociation curve of the newborn is
different [15]! Other factors (2,3-DPG, type of hemoglobin, etc.) and various disease
states can cause a shift of the position of the oxyhemoglobin dissociation curve.

The A (temperature) and B (pH) scales are used to set the conditions at which the test
was done or to conditions desired for calculated results. The C scale shows P02 correated
to 37 °C and 7.40, whereas D scale shows determined P02; the E scale is used to correct P02
for existing changes in ACT or BE/D. The F scale is the "normalized" (37 °C, 7.40, and
zero ACT or BE/D) oxyhemoglobin dissociation curve (Po2/% O2 saturation) up to 600 mmHg and
99.9 percent oxygen saturation.

Table 3 compares temperature corrections for various hypothetical conditions as calculated

by different nomograms or slide rules.

98
 1

Table 3. Comparison of corrections of pH, PCO2,

and P02 to temperature of patient.

I, If patient's temperature is

A. pH 7.40^ 20°C 30°C 35 C 40°C 45°C

C02rrec°t-02-slide 7.65 7.50 7.43 7.36 7.28
Severinghaus Calculator [16] , 7.67 7.51 7.43 7.36 7.29
Siggaard-Andersen Nomogram [17] 7.66 7.50 7.43 7.35 -

Kelman & Nunn Nomogram [18] 7.65 7.50 7.43 7.36 -

B. Pco? 40 mmHg^

C02rrec°t-02-slide 17.5 28.5 36.5 46.5 59

Severinghaus Calculator [16] 17.2 28.8 36.7 45.5 53.3
Siggaard-Andersen Nomogram [17] 18 29 36.5 45
Severinghaus Factors [19] 29.7 36.7 45.5 56.3
Kelman & Nunn Nomogram [18] 19.2 29.6 36.8 45.6 -

Greenburg & Moulder Nomogram [20] 18 29 36.5 45 56

Rattenborg [25] 19.5 30 37 46 57

C. P0 7 90 mmHg^

C02rrec°t-02-slide 00 D 1 ou 1U/
Severinghaus Calculator [16] LI . C / 0 111 ^
Severinghaus-Astrup Nomogram [19,21,22]' 34 60 80 105 140
Severinghaus Factors [19] 27.5 54.9 78.3 111.8 158
Severinghaus Factors [23] , 1 53.2 79.5 105.8 132
Kelman & Nunn Nomogram [18] 36 61.2 81 103.5
Kelman & Nunn Factors [18] 60.8 80.7 105.9
Rattenborg [25] 27 55 79 -

Instrumentation Laboratory Nomogram [26] - 59 79 105 140

II. For PCO2 val ues^

If Patient at 34°C 30 40 50 60 70

C02rrec°t-02-slide 26 34.6 43.3 52 60.5

Severinghaus Calculator [16] 26.2 35 43.6 52.5 61
Siggaard-Andersen Nomogram [17]^ 27 35 43 51 59
Severinghaus Factors [19] 26.4 35.2 44 52.8 61.6
Kelman & Nunn Nomogram [18] 26. 34.8 43.5 52.2 60.9
Greenburg & Moulder Nomogram [20] 26 34 44 52 61
Rattenborg [25] 26.5 35.5 44 53 61

III. For pH values^

A. If Patient at 34°C 00 7. 10 7.20 7. 30 7.50 7.60

C02rrec°t-02-slide 045 7. 145 7.245 7. 345 7.545 7.645
Severinghaus Calculator [16] , 040 7, 142 7.243 7. 348 7.550 7.650
Siggaard-Andersen Nomogram [17] 045 7. 140 7.245 7. 340 7.540 7.640
Kelman & Nunn Nomogram [18] 045 7. 145 7.245 7. 345 7.545 7.645

B. If Patient at 34°C (P02 = 90 mmHg)

C02rrec°t-02-slide 113 101 91 82 65 58

Severinghaus Calculator [16] 113 102 91 81.5 65.5 58.3
Severinghaus-Astrup Nomogram [19,21,22]^ 120 106 95 86 68 60
Kelman & Nunn Factors [18] 115 102,5 92 82.5 66 59
Instrumentat ion Laboratory Nomogram [26] 112 100 90 80 65 58
^ As determined on laboratory instrumentation
'

Modified from 38° to 37°C by adding 2 mmHg.

at 37°C. '
Using 96% saturation line.
^ Modified from 38° to 37°C by substracting '

Radiometer Chart No. 984-204.

0.02 pH.

99
 ,

By definition, the P^o (or T50) value is the partial pressure of oxygen (corrected to
37 °C, pH 7.40, and BE/D (or ACT) of zero) corresponding to 50 percent oxygen saturation;
normal value is about 26.6 mmHg or 3.5 kPa. The methods to obtain the P50 value are time-
consuming. Canizaro, et al. [24] have developed a nomogram which allows estimation of the
P50 value from a single venous blood sample on which the P02 [oorreated to patient's body
temperatuve andpH) and the percent oxygen saturation are determined. The estimated P50
values compare favorably (r = 0.92) with the actual P50 values if the oxygen saturation
(venous blood) is between 25 and 75 percent.

References

[1] Weisberq, H. F., A better understanding of anion-cation ("acid-base") balance,

Surg. Clinics N. Amer. 39, 93 0959).

[2] Weisberg, H. F., The Henderson-Hasselbalch equation is simple. Scientific Exhibit,

ASCP and CAP Annual Meeting, Seattle, Washington (1961).

[3] Weisberg, H. F., Water, Eteatrolyte, and Acid-Base Balance; Eormal and Pathologic
Physiology as a Basis for Therapy, 2nd ed. , 533 pp. (Williams & Wilkins, Baltimore,
MD 1962).

[4] Weisberg, H. F., pH and CO2 content; electrolyte and acid-base balance. Commission
on Continuing Education, Check Sample CC-36, (Dec, Amer. Soc. Clin. Path., 1-33 (1965).

[5] Weisberg, H. F., Calculator for the Henderson-Hasselbalch equation, in Manual for
Procedures for the Applied Seminar on Clinical Pathology of Respiratory Diseases
F. W. Sunderman, ed. , pp. 239-245 (Institute for Clinical Science, Inc., Philadelphia,
PA 1972).

[6] Weisberg, H. F., Water, electrolyte, acid-base, and oxygen, in Todd- Sanford's Clinical
Diagnosis by Laboratory Methods, I. Davidsohn, and J. B. Henry, eds., 15th ed., pp. 772-
803 (W. B. Saunders Co., Philadelphia, PA 1965).

[7] Woodbury, J. W. , Regulation of pH, in Physiology and Biophysics, T. C. Ruch, and

H. D. Patton, eds., 19th ed., pp. 899-934 (W. B. Saunders, Philadelphia, PA 1965).

[8] Filley, G. D., Acid-Base and Blood Gas Regulation, (Lea and Febiger, Philadelphia,
PA 1971).

[9] Davenport, H. W., The ABC of Acid-Base Chemistry, 6th ed., Rev. (U. Chicago Press,
Chicago, IL 1974).

[10] Rispens, P., Significance of Plasma Bicarbonate for the Evaluation of H^ Homeostasis,
Thesis, U. of Groningen, Netherlands (1970).

[11] Rispens, P., Zijlstra, W. G., and Van Kampen, E. J., Significance of bicarbonate for the
evaluation of non-respiratory disturbances of acid-base balance, Clin. Chim. Acta, 54,
335 (1974).

[12] Weisberg, H. F., Parenteral fluid therapy in adults, in Current Therapy 1975,
H. F. Conn, ed. , pp. 421-433 CW. B, Saunders Co., Philadelphia, PA 1975).

[13] Rosenthal, T. B., The effect of temperature on the pH of blood and plasma in vitro,
J. Biol. Chem. 173 , 25 (1948).

[14] Siggaard-Andersen, 0., The Acid-Base Status of the Blood, 2nd ed., (Williams and Wilkins,
Baltimore, MD 1964).

[15] Nelson, N. M., et al., A further extension of the in vivo oxygen-dissociation curve for
the blood of the newborn infant, J. Clin. Invest. 43, 606 (1964).

100
[16] Severinghaus, J. W., Blood gas calculator, J. AppZ. Physiol. Z\_, 1108 (1966).

N.Y. Aaad. Sai. 133 ,

[17] Siggaard-Andersen, 0., Titratable acid or base of body fluids, Ann.
41 (1966).

[18] Kelman, G. R. and Nunn, J. F., Nomograms for correction of blood P02, PCO2, pH and base
excess for time and temperature, J. Appl. Physiol. 21_, 1484 (1966).

[19] Severinghaus, J. W., Blood gas concentrations, in Handbook of Physiology, Section 3^,
Respiration, 2, 1475 (1965).

[20] Greenburg, A. G. and Moulder, P. V., Temperature coefficients for PCO2 and pH in whole
blood. Arch: Surg. 91, 867 (1965).

[21] Astrup, P., Engel , K., Severinghaus, J. W., and Munson, E., The influence of temperature
and pH on the dissociation curve of oxyhemoglobin of human blood, Sound. J. Clin. Lab.
Invest. ]]_, 515 (1965).

[22] Severinghaus, J. W., Oxyhemoglobin dissociation curve correction for temperature and pH
variation in human blood, J. Appl. Physiol. J_2, 485 (1958).

[23] Severinghaus, J. W. , Use of single-line nomograms, in Biology Data Book, 2nd ed., 3^,

1875 (1974).

[24] Canizaro, P. C. et al. , A technique for estimating the position of the oxygen-hemoglobin
dissociation curve, Ann. Surg. 180 , 364 (1974).

[25] Rattenburg, C. C, Blood Gas Evaluator (Graphic Calculator Co., Barrington, IL, 1970).

[26] Instrumentation Laboratory, Inc., Instruction Manual for Model 182 CO-Oximeter, revised
edition (1972).

101
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

AIDS FOR EVALUATION OF ACID-BASE IMBALANCE-DIAGRAMS,

NOMOGRAMS, AND SLIDE-RULES

Harry F. Weisberg
Division of Biochemistry
Mount Sinai Medical Center
Milwaukee, WI 53201, USA

In 1962 I had accumulated 12 acid-base diagrams spanning the period, 1921-1959 (table
1). Since then I have been able to add 7 time charts (1917-1974), 17 nomograms (1923-1974)
[45,46,162,166-180]^ 8 mathematical factors (1957-1967), and 12 slide rules (1958-1974)
,

[181-192]. The original 12 diagrams have expanded to 39 for the period 1921-1959.

Table 1. Evolution of acid base diagrams. Reproduced with permission

of the publisher, Williams and Wilkins [1].

Vertical Horizontal
coordinate coordinate Other
(ordinate) (abstcissa) coordinates

Author year Value Log Value Log Value Log

scale? scale? scale?

Van Slyke 1921 CT^ PC02 PH^

Peters 1923 CT Yes PC02 Yes pH
Peters & Van Slyke 1931 HCOs"'^ . PH
or CT
Shock & Hastings 1934 HCO3" Yes pH PC02^ Yes
Shock & Hastings 1935
Davenport 1947 HCO3" pH PCO2 Yes
Weisberg 1950
Westerfeld 1953^
CT pH PCO2 Yes
Weisberg 1956'
Winters et at. 1958 CT Yes pH PCO2 Yes
Weisberg 1961 CT [H^]^ PCO2 Yes
Singer 1951 BB^ PCO2 pH and Yes
CT
Peirce et al. 1959 PCO2 pH BB Yes
Brewin et al. 1955^
PCO2 Yes pH
Astrup 1956/
Astrup et al. 1959 PCO2 Yes pH BB, BE,?
and SB
CT--carbon dioxide content; pH-- reciprocal of hydrogen ion concentration;
HC03--bi carbonate; dpco2--partial pressure of carbon dioxide; ^[H+]--hydrogen
ion concentration; fBB--buffer base; %E—
base excess; and 'iSB--standard bicarbonate.

Figures in brackets indicate literature references at the end of this paper.

103
 . 2

Figure 1 illustrates the fecundity of the "sixties" for the production of acid-base
aids. The earliest diagram I have been able to unearth was published in 1914 by Christiansen,
Douglas and Haldane [3]. Table 2 is the latest accumulation of 105 diagrams (1914-1973) used
in the evaluation of acid-base imbalance. The references listed in the 1° column are the
originals I have been able to find, whereas those references in the 2° column are corrections,
modifications, and variations based on the originals.

50-1
Diagrams D 105
Nomograms 17

40-
Slide Rules ^ 1

30-

20-

im.
1911-20 1921-30 1931-40 1941-50 1951-60 1961-70 1971-

Ftgure 1. Histogram of acid-base aids. Modified from Weisberg [2].

Table 2. Comparison of 105 acid-base diagrams. Modified

from Weisberg [1 ,2]

Key to Symbols :

Triaxial coordinates
aH^ Hydrogen ion activity
A Acid, expressed as ± Pco2 (Kintner)
Alk. Res. Alkaline reserve (McClendon)
B Base, expressed as ± base (=G=BE) (Kintner)
BB Buffer base
BE Base excess
BEj^g Base excess, whole body (Russell st at.)

BE/D^^P Base excess/deficit of extracellular fluid (S.iggaard-Andersen)

BE/Dy. c Base excess/deficit for extracellular fluid with hemoglobin at
5 g/dl CRooth)

C^7 Hydrogen ion concentration (OpH)

CSF Cerebrospinal fluid (McClendon)
CT Total carbon dioxide (as mmol/1 or vol %, etc.)
CY CO2 capacity (Goldberger)
aA-B a Acid-base (Astrup)

104
 Table 2. Comparison of 105 acid-base diagrams (continued).

ECF Extracellular fluid

[H+] Hydrogen ion concentration
h\o Hydrogen at 40 mmHg (Frumin)
[HC03~] Bicarbonate ion concentration
[HCOs'l^o Bicarbonate at 40 mmHg (Siggaard-Andersen) ; as SB
[H2CO3] Carbonic acid (and dissolved CO2) concentration
LA Lactic acid
N Normality of h"*" (McClendon)
"N" Normality of titration of alkaline reserve (McClendon)
EO2] Concentration of (dissolved) O2
O2 CT Total oxygen (vol %)

PCO2 Partial pressure of carbon dioxide

"PCO2" Assumed from % CO2 in a alveolar air (McClendon)
pHi^o pH at 40 mmHg (Kappagoda et at.)
PR Serum protein (Stadie et at.)
R Ventilation ratio (Kim et al.)
S Solubility coefficient for CO2 in plasma at 38 °C
SB Standard bicarbonate at PCO2 40 mmHg, 38 °C, Hb oxygenated
SBy^ito Standard bicarbonate at pH 7.40 (Van Slyke)
t" Temperature (Stadie et al.) (Lenfant)
Ti+o Bicarbonate at 40 mmHg in vivo (Collier et al.)
TB Total buffer (Coats)
VR Ventilation ratio
WB Whole body

Reference Values Plotted On

Year Author(s) 1° 2° Ordinate Log? Abscissa Log? Others Log?

1914 Christiansen et al. 3 4,5 CT PC02

1916 Lewis et al. 6 7 pH CT

1916 McClendon 8 PCO2 pH

CT

1917 McClendon et al. 9 PCO2 Yes

Eh+] Yes

1917 McClendon 10 pH PC02 Yes [HCOi] Yes

rH+] Yes

1917 McClendon 10 CT Alk. Res. Yes

1917 McClendon et al. 11 PCO2 Yes Yes Alk. Res. Yes

pH

1917 Parsons 12 13 pH PC02

105
 Table 2. Comparison of 105 acid-base diagrams (continued).

Reference Values Plotted On

Year Author(s) 1° 2° Ordinate Log? Abscissa Log? Others Log?

1917 Parsons 12 13-15 PC02

1918 McClendon 16 "PC02" Yes [H^] Yes CSF "N" Yes

pH

1918 Straub & Meier 17 CT PCO2 pH Yes

[H2CO3]

1919 Haggard & 18 CT PC02 Yes

Y. Henderson [H2CO3]

1920 Henderson 19 EHCO3] [H^]

1921 Henderson 20 [HCO3] pH [H2CO3]

[C1-]
P02
Hb02

1921 Van Slyke 21 22-24 CT PC02 pH Yes

[H2CO3]

1921 Van Slyke 21 [HCO3] pH

1922 Van Slyke et al. 25 CT

LHCO3J pH

1922 Doisy et al. 26 pH [HCOi]

1923 Peters et al. 27 28 [HCOi] Yes [H2CO3] Yes

1923 Peters 29 CT Yes PC02 Yes pH Yes

1923 Peters 29 24 CT Yes PC02 Yes pH

[H2CO3] Yes

1923 Cullen & Jonas 30

^
32,33,193 CT pH PC02 Yes
1924- 25 Hastings et al. j 31

1924 Bock et al. 28 23 [HC03"J pH

1924 Bock et al. 28 CT pH

1925 Murray & Hastings 34 [HCO3] Yes pH [H2P0^] Yes

[HP04^"] Yes
[PO^^"] Yes

[Ca2+] Yes

[COs^"] Yes

1925 Austin & Cullen 35 [HCOi] PC02 [H2CO3]

1925 Austin & Cullen 35 36 CT pH Pco2 Yes

106
 Table 2. Comparison of 105 acid-base diagrams (continued).

Reference Values Plotted On

Year Author ( s 1° 2° Ordi na fp na?
Luy nn?
1 : LOg
1
: u uners Log f

1925 Stadie et at. 37 [HC03] pH

r K

4- 0
OI t pH [HCOi] Yes
[H2CO3] Yes
PC02 Yes

1927 Eisenman 38 CT Yes PCO2 Yes

1927 Dill et al. 39 CT Yes PCO2 Yes

1928 Henderson 36 CT PCO2 -

[H2CO3J
LO2J

1928 Henderson 36 PCO2 pH CT Yes

1931 Peters & Van Slyke 23 40 CT Yes PCO2 Yes pH

nastings & /n 4^-46

1 ))o 1
PCO2 Yes pH [HCOi] Yes
Steinhaus°

1932 Douglas & Havard 47 pH PCO2 -

CT
L 1

McCl endon° 48 pH rC02 Yes [HCOi] Yes

1947 Davenport 49 45,50-54 [HCOi] pH PC0 2 Yes

1948 Clark 55 56-61 ,165 [HCO3] PCO2 pH Yes

1 yb 1 Singer BB PCO2 CT Yes

pH Yes

1954 Astrup 1 63,69,77, PCO2 Yes pH -

164

1955 Cranston et al. 64 [H2CO3] -

1956 Roberts et at. 65 PCO2 CT -

pH

1 yoD roppei et au. 00 0/ CO

C7 ,00 rC02 L

\ yoD MS irup 03 rC02 Yes pH SB Yes

1958 Winters et al. 71 72 CT Yes pH PCO2 Yes

1958 J0rgensen 73 SB Yes PCO2 Yes pH

1959 Go! dberger 74 CT pH

1
CY

1959 Peirce 75 PCO2 pH BB Yes

107
 Table 2. Comparison of 105 acid-base diagrams (continued).

Reference Values Plotted On

Year Author(s) 1° 2° Ordinate Log? Abscissa Log? Others Log?

1959 Astrup 76 PC02 Yes pH BB Yes

SB Yes
AA D
AM-b I es

1960 Astrup et at. ] 77 46,52, PC02 Yes pH BB Yes

1you Si ggaard-Andersen 7Q
/o 7Q QQ
/ts-oy CD Vqc
1es
]

BE Yes

1 ybu de la Huerga on
yu, L. 1 pH Vac rC02
91

CT rH''"T PCO2 Yes

1961 Wei s berg 92
pH Voc
I cb

1961 Lenfant 93 [H2CO3] Yes pH CT Yes

1 951 Lenfant 9o ru rr\ 1

LH2CU3J Vat-
Yes PC02 t

1961 We is berg 92 1 CT pH PCO2 Yes

[HCO3]
Yes
LH2CO3J

1962 Campbel 1 94 95,96 PCO2 [HCO3] Yes

Ln2'-U3J pn 1Co

1 ytD£ Campbel yf Q7
y/ rCU2 1 Co EHCOg] 1 CO un 1Co
[H2CO3] Yes [H+] Yes
pH

1962 Nunn 98 [HCO'a] PC02 pH Yes

1963 Si ggaard-Andersen 81 99 PC02 Yes pH [HCOi] Yes

1963 Weisberg (revised) 100 101-103 CT dH PC02 Yes

[H+] Yes
[HCO3]
Yes
LH2CO3J

1963 Robin 104 PH PC02 [HCOi] Yes

i ybo rink a iNanas t Ub PC02 Jnous J Vqc

ICS
pH Yes VR

964 Wni tehead rC02 Yes Vqc

I es
1 106,
107

1964 Coats 108 [HCOi] [H2CO3] pH Yes

TR

1964 Da r row 109 [HCO^] PC02 pH Yes

[H2CO3J [HCO3]
Yes
[H2CO3]

1965 Woodbury 110 15,52, [HCOi] PH^ PC02 Yes

111-114 [H+J Yes

108
 ))

Table 2. Comparison of 105 acid-base diagrams (continued).

Reference Values Plotted On

Year Author(s) 1° 2° Ordinate Log? Abscissa Log? Others Log?

1965 Schwartz {et al. 115, 13,117, [HCOi] PC02

116 118

1965 Schwartz (et al. 115, 117-119 PC02

116 pH Yes

1965 Mithoefer et al. 120 121 [HCOi] Yes PC02 Yes [H^] Yes
pH

1965 Owen et al. 122 PC02 [H^] EHCO3]

pH Yes

1965 El kinton 123 124 [HCO'g] Yes PCO2 Yes pH

1965 Siesjo & Ponten 125 PC02 Yes [HCOs] SB Yes

S-PC02

1966 Instrumentation 126 PC02 Yes pH CT Yes

Lab'

1966 Lennon & Lemann 14 [HCO3]

1966 Lennon & Lemann Id 1 c. / , 1 CO , PC02 iHCOi] -

140

1966 Kim et al. 129 CT PC02 pH Yes

R
O2 CT

1966 Stinebaugh & 130 131,132 PC02 CT pH Yes

Austin

1966 Young 133 [HCO3] Yes pH PC02 Yes

[H+] Yes

1967 Cohen 134 135-138 [H+] PCO2 [HCOi] Yes

pH Yes

-
1967 Albert et al. 127 PCO2 BE

-
1967 Albert et al. 127 139 PC02

1 / n Uci1 L c: u CLL' . 127 PC02 nH

pn

1967 Siggaard-Andersen 85 89,141 PC02 Yes pH Yes

^^(ECF)
Yes [HCOi] 40 Yes

1967 Kintner 142 CT pH PC02 Yes

[H+] Yes A Yes
B Yes

1967 Winters 57 59 PC02 [HCOi] pH Yes

1967 Winters 57 59,143 PC02 BE pH Yes

1967 Winters 57 59 BE PC02 pH Yes

109
 Table 2. Comparison of 105 acid-base diagrams (continued).

Reference Values Plotted On

Year Author(s) 1° 2° Ordinate Log? Abscissa Log? Others Log?

1967 Srouji 144 BE PC02 _

1968 Engel et at. 13 BE PC02

1968 Goldring al. 14(1

1"tO pH EHCO3J PCO2 Yes

1969 Li & Holder 147 pH SB PC02 Yes

1969 Nunn 112 PC02 Yes pH

[H+] Yes
'

1969 Weisberg 148 149-153 CT PCO2 Yes

Yes [H2CO3] Yes
[HCO3]
Yes
[H2CO3J

PC02 -
1970 Gilbert &
Auchincloss

1970 Heisler & Schorer 155 89 [HCOi] pH PCO2 Yes

BB Yes
BE Yes

1970 Kappagoda et at. 156 PC02 Yes pH pHitO

1970 Rooth 157 PC02 pH Yes

^^Hbs
^^Hbs

1970 Rooth 157 PC0 2 pH Yes

8°Hb5
-
1970 Blair 158 PC02 pH
BE
LA

1971 Stephens 159 CT pH Yes PCO2 Yes

1971 Slonim & Hamilton 160 [HCO^J PCO2 pH

[H2CO3]
VR Yes

1972 Kintner 161 [HCOi] pH PCO2 Yes

BB

1972 Russel et al. 162 PC02 Yes pH Yes

^^WB

1973 Visser 163 PC02 Yes pH Yes

110
 References

[1] Weisberg, f/ater^ Electrolyte and Acid-Base Balance, Znd edition, 533 pp. (Williams and
Wilkins Co., Baltimore, 1962).

[2] Weisberg, InteTyretation of Electrolyte, Acid-Base, and Oxygen Imbalance, 4th edition,
3rd revision (Private mimeographed, 1975).

[3] Christiansen, Douglas and Haldane, J. Physiol. 44, 244 (1914).

[4] Chambers and Kleinschmidt, J. Biol. Chem. 55, 257 (1923).

[5] Eisenman, J. Biol. Chem. 71, 611 (1927).

[6] Lewis et al. , Brit. Med. J. 2, 517 (1916).

[7] Astrup, Ugeskr. f. Laeg. 116 , 758 (1954).

[8] McClendon, J. Biol. Chem. 28, 135 (1916-17).

[9] McClendon et al. , Carnegie Inst. Washington Vl_, 23 (1917).

DO McClendon et al. , J. Biol. Chem. 30 265 (1917).

[11 McClendon et al. , J. Biol. Chem. 31 519 (1917).

[12 Parsons, J. Physiol. 51 440 (1917).

[13 Engel et al., J. Appl. Physiol. 24-, 288 (1968).

[14 Lennon and Lemann, Ann. Int. Med. 65, 265 (1966).

[15 Ichiyanagi et al., Resp. Physiol. 7, 310 (1969).

[16 McClendon, J. Amer. Med. Assoc. 70, 977 (1918).

[17 Straub and Meier, Bioahem. Z. 89, 156 (1918).

[18 Haggard and Henderson, Y., j. Biol. chem. 39, 163 (1919).

[19 Henderson, J. Biol. Chem. 41, 401 (1920).

[20 Henderson, J. Biol. Chem. 46, 411 (1921).

[21 Van Slyke, J. Biol. Chem. 48, 153 (1921).

[22 Austin et al. , J. Biol. Chem. 54, 121 (1922).

[23 Peters and Van Slyke, Quantitative Clinical Chemistry, Vol. I, Interpretations
(Williams and Wilkins Co., Baltimore, 1931).
'

[24 Van Slyke, Bull. N.Y. Acad. Soi. ]0_, 103 (1934).

[25 Van Slyke et al. , J. Biol Chem. 53, 277 (1922).

[26 Doisy et al., J. Biol. Chem. 54 305 (1922).

[27 Peters et al. , J. Biol. Chem. 55, 709 (1923).

[28 Bock et al., J. Biol. Chem. 59, 353 (1924).

[29 Peters, J. Biol. Chem. 56, 745 (1923).

Ill
[30] Cullen and Jonas, J. Biol. Chem. 57, 941 (1923).

[31] Hastings et al. , J. Clin. Invest. 1, 25 (1924-25).

[32] Westerfeld, Fundamentals of Aoid-Base and Electrolyte Bioohemistry (Syracuse, N

mimeographed edition, 1953).

[33] Weisberg (1956).

[34] Murray and Hastings, J. Biol. Chem. 65, 265 (1925).

[35] Austin and Cullen, Medicine 4, 275 (1925).

[36] Henderson, Blood (Yale Univ. Press, New Haven, 1928).

[37] Stadie et al., J. Biol. Chem. 66, 901 (1925).

[38] Eisenman, J. Biol. Chem. T\_, 587 (1927).

[39] Dill et al., J. Biol. Chem. 73, 251 (1927).

[40] Miller, Acta Med. Scand. 165, Suppl . 348 (1959).

[41] Hastings and Steinhaus, Amer. J. Physiol. 96, 538 (1931).

[42] Shock and Hastings, J. Biol. Chem. 104, 585 (1934).

[43] Shock and Hastings, J. Biol. Chem. ri_2, 239 (1935).

[44] Hartmann, Pediatrics 2, 584 (1948).

[45] Rispens, Thesis, U. of Groningen (1970).

[46] Mohler (1971), personal communication.

[47] Douglas and Havard, J. Physiol. 74, 471 (1932).

[48] McClendon, Science 81, 569 (1935).

[49] Davenport, The ABC of Aoid-Base Chemistry, 4th edition (Univ. of Chicago Press,
Chicago, 1958).

[50] Weisberg, Bull. Mt. Sinai Hosp. (Chicago) 2, 85 (1950).

[51] Jurado-Garcia (1960), personal communication.

[52] Davenport, The ABC of Acid-Base Chemistry, 5th edition (Univ. of Chicago Press,
Chicago, 1969).

[53] Nelson and Riegel, Pediatrics 43^, 821 (1969).

[54] Refsum, Scand. J. Clin. Lab. Inv. 27, (Suppl. 118) 8 (1971).

[55] Clark, Topics in Physical Chemistry (Williams and Wilkins Co., Baltimore, 1948)

[56] Filley, John Hopkins Med. J. 120 , 355 (1967).

[57] Winters, Pediatrics 39, 700 (1967).

[58] Hamilton, in Stefanini, ed.. Progress in Clinical Pathology 284 (1969).

[59] Winters et al., Acid-Base Physiology in Medicine, The London Company (1967).

112
Arbus, Canadian Med. Asen. J. 109, 291 (1973).

Ayres, i?esp. Care 19, 280 (1974).

Singer, Mer. J. Med. Soi. 221 , 199 (1951).

Brewin et at., Guy's Hoep. Reports 104, 177 (1955).

Cranston et at., J. Physiol. 127 . 380 (1955).

Roberts et al., J. Clin. Invest. 35, 261 (1956).

Poppell et al. , J. Lab. Clin. Med. 47, 885 (1956).

Bane et al. , J. Appl. Physiol. 10^, 151 (1957).

Austin, Amer. Heart J. 69, 691 (1965).

Astrup, Scand. J. Clin. Lab. Invest. 8, 33 (1956).

Visser and Maas, Clin. Chim. Acta 5_, 850 (19G'.).

Winters et al. , J. Clin. Invest. 37_, 640 (1958).

Done and Temple, Mod. Treatment 8, 528 (1971).

J0rgensen, Ugeskr. f. Laeg. ]2^, 993 (1958).

Goldberger, A Primer of Water j Electrolyte and Acid-Base Syndromes (Lea and Febiger,
Philadelphia, 1959).

Peirce, Trans. Amer. Soo. Artif. Int. Orgs. 6^, 240 (1960).

Astrup, in Woolmer, ed., pH and Blood Gas Measurement (Little, Brown and Company,
Boston, 1959).

Astrup et al.. Lancet 1035 (1960).

Siggaard-Andersen, Scand. J. Clin. Lab. Invest. 1^, 311 (1960).

Rosevear (1960), personal communication.

Siggaard-Andersen, Scand. J. Clin. Lab. Invest. U, 598 (1962).

Siggaard-Andersen, Scand. J. Lab. Clin. Med. |5^, Suppl . 70 (1963).

Bates and Christie, Respiratory Function in disease (Saunders, Philadelphia, 1964).

Goldberger, A Primer of Water, Electrolyte and Acid-Base Syndromes, 3rd edition (Lea
and Febiger, Philadelphia, 1965).

Bendixen et al. Respiratory Care, (C. V. Mosby Co., St. Louis, 1965).

Siggaard-Andersen, in Evans and Gray, eds.. Modem Trends in Anesthesia 99 (1967).

Kelman, Clin. Chim. Acta 22, 111 (1968).

Thompson and Mullen, J. Surg. Res. 1_0, 117 (1970).

Mason, Fluid Electrolyte and Nutrient Therapy in Surgery (Lea and Febiger, Philadelphia,
1974).
Siggaard-Andersen, The Acid-Base Status of the Blood, 4th edition (William and Wilkins
Co., Baltimore, 1974).

113
^90] de la Huerga (1960), personal communication.

'91] de la Huerga, in Hoffman, Biochemistry of Clinical Medicine, 4th edition (Year Book
Med. Publ., Chicago, 1970).

:92] Weisberg ,(1961).

;93] Lenfant, (ADI 6810, Library of Congress), J. Appl. Physiol. 16 909 (1961).

;94] Campbell, Lancet 2, 154 (1962).

;95] McNicol and Campbell, Lancet 336 (1965).

;96] Makoff, Geriatrics 26, 107 (1971).

]97] Campbell, Dickinson and Slater, Clinical Physiology, 2nd edition (Blackwell Scientifi
Publications, 1963).

'98] Nunn, in Evans and Gray, eds.. Modern Trends in Anaesthesia 2, Butterworths , London
(1962).

;99] Fink, Proc. Royal Sac. Med. 57, 693 (1964).

:iOO] Weisberg (1963).

:i01] Weisberg (1965).

:i02] Weisberg (1967).

il03] Lang (1966), personal communication.

;i04] Robin, New Eng. J. Med. 268, 917 (1963).

;i05] Fink and Nahas, Ann. N.Y. Acad. Soi. 109 . 804 (1963).

;i06] Whitehead, Ph.D. Thesis, Univ. of Birmingham (1964).

;i07] Whitehead, Lancet 2, 1015 (1965).

^108] Coats (1964), personal communication.

'109] Darrow, A Guide to Learning Fluid Therapy, (Charles C. Thomas, Springfield, IL, 1964)

^110] Woodbury, in Ruch and Patton, eds.. Physiology and Biophysics , 19th edition (W. B.
Saunders Co., Philadelphia, 1965).

;ill] Refsum, Scand. J. Clin. Lab. Invest. V7, 508 (1965).

112] Nunn, Applied Respiratory Physiology (Appleton-Century-Crofts 1969).

113] Collier, Hackney and Mohler, Chest 61, 6S (1972).

114] Woodbury, in Ruch and Patton, eds.. Physiology and Biophysics, 20th edition (W. B.
Saunders Co., Philadelphia, 1974).

115] Brackett et al. , New Eng. J. Med. 272, 6 (1965).

116] Schwartz, J. Clin. Invest. 44, 291 (1965).

117] Arbus et al. , New Eng. J. Med. 280, 117 (1969).

118] Levesque (1972), personal communication.

119] Van Ypersele de Strihou and Frans, Presse Med. 75, 1797 (1967).

114
 . ,

120] Mithoefer et at.. New Eng. J. Med. 272, 1200 (1965).

121] Elkinton, Med. Clin. N. Amer. 50, 1325 (1966).

122] Owen et al.. Lancet Z, 660 (1965).

123] Elkinton, Ann. Int. Med. §2, 893 (1965).

124] Filley, Acid-Base and Blood Gas Regulation (Lea and Febiger, Philadelphia 1971).

125] oiesjo ana runicn, Ann. n.i. Roaa. ocu. loo, lou ^isoay.

126] 1 ns LrumciHa l i uri i_a uu r a lu ry, iiic. ^laooy.

127] Ml Deri ev au. } Ann. mv. Mea. dd, \\yo/i.

128] riciLc aiy. ^ /inn. mv. lyi&a. uoo \ i y/u j

129] i\i[n ev au. , v . Hpp u . rnyst-o u. ^i, looo ^i^dd^.

130] otmcDauyri dnu nusiin, iiew arig. d. Mea. c/o, o/o ^lyooy.

131] oiineuaugn ana Hustin, Avon. mv. Mea. iiy, \oc. yxvoi).

132] rieiscner, Mnnuai nesting, AmeT. ooc. Luvn. tavn \ \v\>ii.

133] loung ^lyoDj, personal conimum cation.

134] Lonen, /in?^. int. Mea. bo, loy ^lyO'J-

135] rieniey. Lancet 1, yb 1 ^19/ I

J.

136] Masoro and Siege! , Acid-Base Regulation (W. B. Saunders Co. Philadelphia 1971).

137] aoiuDerg au. ^ d. Ameip. Mea. assoc. ccj, co)) ^ly/o^.

138] Gardner and Cannon, et al. , personal communication.

139] Howorth, Brit. J. Bis. Chest 69, 75 (1975).

140] Van Ypersele de Strihou and Frans, Avch. Inter. Physiol. Biochimie 76, 547 (1968)

141] Siggaard-Andersen, Soand. J. Lab. Clin. Invest. 21_, 239 (1971).

142] Kinter, Amer. J. Clin. Path 47, 614 (1967).

143] Monti and Rooth, Scand. J. Clin. Lab. Invest. 26, 381 (1970)

144] Srouji, Surgery 62, 958 (1967).

145] Goldring et al., J. Clin. Invest. 47, 188 (1968).

146] Michel, Respiration Physiol. 4, 283 (1968).

147] Li and Holder, Anesth. and Analgesia 381 (1969).

148] Weisberg (1969).

149] Weisberg (1970).

150] Weisberg (1972).

151] Weisberg (1974).

115
'152] Bronk (1971), personal communication.

'153] Corning Scientific Instruments.

;i54] Gilbert and Auchincloss, Arch. Int. Med. 125 , 820 (1970).

]155] Heisler and Schorer, Der Anaesthetist 19^, 93 (1970).

;i56] Kappagoda et at., Clin Soi. 39, 169 (1970).

;i57] Rooth, Clinical Acid-Base and Electrolyte Balance, Studentl itteratur (1970).

;i58] Blair, Am. J. Surg. 119 , 433 (1970).

^159] Stephens, in Hardy, ed.. Critical Surgical Illness, quoted by Randall (W. B.
Saunders, Philadelphia, 1971).

"160] Slonim and Hamilton, Respiratory Physiology , 2nd edition (C. V. Mosby, St. Lou
1971).

'161] Kintner, in Stefanini, ed.. Progress in Clinical Pathology 4, 143 (1972).

;i62] Russell et al. , Amer. J. Physiol. 223, 689 (1972).

!163] Visser (1973), personal communication.

;i64] Fulop et al., Clin. Sci. Mol. Med. 46, 539 (1974).

:i65] King (1972).

J 66] Van Slyke et al., J. Biol. Chem. 56, 765 (1923).

;i67] Henderson et al., J. Biol. Chem. 59, 379 (1924).

;i68] Van Slyke and Sendroy, J. Biol. Chem. 79_, 781 (1928).

;i69] Hastings and Shock, J. Biol. Chem. 104, 575 (1934).

;i70] Bill et al., J. Biol. Chem. 118 , 635 (1937).

;i71] McLean, Physiol. Rev. 18, 495 (1938).

;i72] Singer and Hastings, Medicine ?J_, 111 (1948).

il73] Siggaard-Andersen, Scand. J. Clin. Lab. Invest. J_5, 211 (1963).

;i74] Kinter, Amer. J. Clin. Path 614 (1967).

;i75] Thews, Pflugers Arch. ges. Physiol. 296, 212 (1967).

;i76] Kintner, Tech. Bull. 38, 254 (1968).

il77] Rattenborg (1968), personal communication.

;i78] Suero, Clin. Biochem. 2, 177 (1969).

;i79] Rispens, J. Appl. Physiol. 34, 377 (1973).

;i80] Rdher et al. , Prakt. Andeth. 9, 40 (1974).

;i81] Joseph, Science 128, 1207 (1958).

;i82] Searcy et al., Amer. J. Med. Tech. 26_, 183 (1960).

116
[183] Weisberg (1963).

[184] Weisberg (1965).

[185] Abbott Laboratories

[186] Severinghaus (1965).

[187] Instrumentation Laboratories (1967).

[188] Kintner and Gambino (1967).

[189] Rattenborg et al. (1968).

[190] Burrows, Mev. J. Clin. Path 52, 585 (1969).

[191] Instrumentation Laboratories (1970).

[192] Weisberg (1974).

[193] Toro and Ackermann, eds., Pvaotioal Clinical Chemistry (Little Brown and Co., Boston,
1975).

117
I

i
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
Issued June 1977
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975.

THE OVERALL FIRST IONIZATION EQUATION OF CARBONIC ACID AS

RELATED TO CO2 IN GAS PHASE: A NEW pK

A. H. J: Maas
Department of Cardiology and Thoracic Surgery
University Hospital
Utrecht, The Netherlands

B. F. Visser
Department of Anesthesiology
University Hospital
Nijmegen, The Netherlands

For the evaluation of acid-base disturbances in biological fluids, the carbonic

acid/bicarbonate buffer system is in common use. In all studies this system is described by
the 1st ionization equation of carbonic acid as related to the physically dissolved carbon
dioxide in the solution: the well-known Henderson-Hasselbalch equation. As a result of
various applications, originally for calculation of the pH of plasma, later on the pCOa and
nowadays the actual bicarbonate concentration, standard bicarbonate concentration or total-
CO2 concentration, this equation is written in different forms. However, little emphasis
has been placed on an exact formulation which accounts for the non-ideality of the system.
By virtue of the thermodynamic theory of equilibrium, we have shown that the 1st ionization
equilibrium of carbonic acid, which relates the bicarbonate concentration and the partial
pressure of CO2 to the pH of the solution, is preferable to the conventional Henderson-
Hasselbalch equation [1]^.

1 . Theory

If there is equilibrium between CO2 in the gas phase and that in the liquid phase, we
can consider the overall equilibrium:

CO2 (gas) + H2O J HCO3 + H"^ (1)

of which the equilibrium constant K^^ is defined as:

aH^ • aHCOa i/HCOg aH"^ • cHCOg

=
^ig = X . (2)
aH20 • fCOz aHzO • gCOz pCOz

Suffix g in Kin refers to CO2 in gas, a, f, y, g, a and p symbolize respectively, activity,

fugacity, activity coefficient, fugacity coefficient, concentration, and partial pressure.

For practical purpose we define a practical coefficient z{q:

aH20 • ^C02 aW^ • cHCOi

K{n = X Kiq = (3)
J/HCOg PCO2

^Figures in brackets indicate the literature references at the end of this paper.

119
 )

Usually one considers the 1st ionization equilibrium as related to the sum of freely dissolved
CO2 and the hydrated CO2. The starting-point in that case is the overall 1st ionization
equilibrium, related to the dissolved carbon dioxide:

COaCliquid) + H2O j H"^ + HCO3 (4)

of which, the equilibrium constant is defined as follows:

aH"*" • aHCOg yHCOi qH"^ • CHCO3

X = = X .
(5)
aHaO • aC02 aH20 • 1/CO2 ^^^^z

The suffix 1 in Ki-, refers to CO2 in the liquid phase. Now the CO2 concentration is equal
to the solubility s times pC02 minus the concentration of H2CO3:

0CO2 = S ' pCOz - CH2CO3 (6)

which equation by means of the hydration constant K^^ can be written as:

aH20 • 1/C02 • K.
0CO2 = S • pC02/(l + -) (7)
2yH2C03

or in a simplified form:

<3C02 = S . pC02/(l + a) . (8)

If we replace eC02 by S • pC02/(l + a) in equation (5) then Ki-, is changed into the conven-
tional form:

ytiCOl • (1 + a) aH"^ • CHCO3

jC = X .
(9)
aHzO • 1/C02 S •
PCO2

For practice we again define a coefficient easy in operation:

aH20 • Z/C02 ' CHCO3

K{. = X = .
(10)
'
yHCOi • (1 + A) ' S •
PCO2

To be able to choose between the two practical equations we first write them in the usual
form. We take the negative logarithm of these equations and apply the notation -log = p.
This gives us for equation (10):

+ aH20 •
yC02 oHCOl CHCO3
paH = pKi. - log + log =
p^i-, + log (11
z/HCOs • (1 + a) S '
PCO2 S • PCO2

120
which equation represents the exactly defined expression of what is now known as the Henderson-
Hasselbalch equation and for equation (3):

^ aHjO • gC02 CHCO3 CHCO3

paH = - log + log = pi^ig + log (12)
p^ig
yHCOs pCOa pCOa

which equation we propose to call the modified Henderson-Hasselbalch equation. Since this
equation contains one variable less, namely the solubility coefficient 5, this modification
of the Henderson-Hasselbalch equation is the most suitable equation for description of the
carbonic acid/bicarbonate buffer system. Assuming g-COa equals 1 (see ref [1]) and a equals
zero, it will be clear that the variability of p^ig is determined by the ratio aH20/y\]C0l
and the variability of pKi-. overmore by yC02, which coefficient strongly depends on lipid
content of the solution.

2. Experimental Section

In order to be able to apply the 1st ionization equilibrium of carbonic acid as related
to CO2 in gas,j«/e established the values of the ionization constant (pi^ig) and practical
coefficient (p?ig) over a large range of application [2].

A. pZig

The true ionization constant pKig has been determined in water using (NaHCOs + NaCl)
solutions of ionic strength 0.01 < I < 0.500 in equilibrium with (CO2 + H2) gas mixtures
applying electromotive force measurements with a hydrogen electrode cell without a salt
bridge from which the exact pKig value results. We found:

pZig (25°C) = 7.720 pZig (38°C) = 7.810 .

pZig, a thermodynamic constant, is independent of solution composition and varies with tem-
perature only. For the temperature range 25-40 °C, a useful equation could be derived sub-
tracting the equation

= 3374J62 ^
pZi^ Q_Q3272y . ^^jqj (13)

from Shedlovsky and Mclnnes [3] and

log 5q = ^-^^y^^ + 0.0148483T - 13.670 (14)

from data of Bartels [4] and Austin [5] for pure water resulting in the relationship:

pZig = pZi^ - log = 0.017877 - 1.037 (15)

where T is the absolute temperature. Our measured data agree within 0.005.

121
 B. pXjg

The practical coefficient

log =
p^ig = pXiq - pH - log CHCO3 + log pCOa (16)

was determined in (NaHCOs + NaCl) solutions, cerebrospinal fluid (CSF), plasma and serum,
measuring pH, c?HC03 and pC02. We use the symbol ~ to indicate that pH in consequence ofjthe
residual liquid junction potential in the measurement is not exactly identical to paH'^.pK-^n
and pK{g are related to p^ig and pK{g according to the following equation

+
pXlg - pKig = pK[g - \)K{g = pH - paH . (17)

3. Bicarbonate Solutions

Using a hydrogen electrode cell including a salt bridge, we measured the practical
coefficient for various ionic strengths and found by extrapolating to infinite dilution an
approximate value of the ionization constant p^ig, which differed airaa n.n05 from pKig
indicating that the residual liquid junction is very small. From these measurements, we
also derived an empirical equation for yHCOs, which in the form as recalculated by Siggaard-
Andersen [6] reads:

log yHCOg = -(0.50 + O.OOlAr) /f + 0.131 (18)

where at = r - 288.15 K.

As a first approximation, log aH20 decreases linearly with I according to the

equation [2]:

log aH20 = -0.0151 (19)

Combining equations (15), (18), (19) and adding 0.005 for the residual liquid junction
potential give a valid approximation of pZ|q as a function of ionic strength at temperatures
^
15-40 °C.

pK{g = 102|-404 ^ 0.017877 - 1.042 - (0.50 + O.OOIat) /i + 0.1457 .

(20)

Our measured results deviate less than 0.005. For the activity coefficient yH^ , we recalcu-
lated our data and found the relationship:

log yti^ = -(0.50 + 0.001A7) /f + 0.81 . (21)

Apart from the problem of the residual liquid junction potential, the usual glass
electrode deviates from the fundamental hydrogen electrode in the pH determination [7].

122
 ^
Further, the bicarbonate determinations in body fluids are subject to error of analysis,
because carbonate and carbamate are enclosed in the present titrimetric and gasometric
analysis [8]. To be able to understand quantitatively the influence of the disturbing
"bicarbonate" fract^ions and the error of the glass electrode, we have determined the practi-
cal coefficients pz{g of a bicarbonate solution isotonic with CSF and plasma, composition
0.025 mol/1 NaHCOs + 0.135 mol/1 NaCl over the range 6-8 pH and performed simultaneously the
required pH measurements with the hydrogen electrode cell as well as with the glass electrode
cell at 25 and 38 °C (fig. 1).

25° 25^

7.55 7.55-

m &-.-fi..p.

— 7.54-
7.54- ffl A'

^ A A
* » -^A. A
7.53 7.53- \
\
\

7.52- \ 7.52- b

7.51 7.51-
pl<'g(cb,gl)---
Pi<^g(cb)— O pK^g(gi) •••••
7.50-

1
P^ig
-pH
— 7.50
pKjg —A
pH
—I 1 1
1
—
6.0 6.5 7.0 7.5 8.0 6.0 6.5 7.0 7.5 8.0

38°

7.64 7.64
S.^j»

t. <
7.63 7.63 A ^ 0\p-..

7.62' 7.62

7.61 7.61

7.60- 7.60-
pKjg(cb,gl)-'
••
pKjgCgl)
7.59- 7.59'
pK'

pH

6.0 6.5 7.0 7.5 8.0 eTo eTi tTo tT? 8.0

Figure 1. The uncorrected and corrected practical coefficient pZjg of (NaHCOs +

NaCl) solution (j = 0.16), determined with the hydrogen electrode cell and
with the glass electrode cell as a function of pH (see text).

123
 In the two graphs on the left-hand side, the overall 1st ionization coefficient, deter-
mined with the hydrogen electrode cell.^is plotted as a function of the pH. The broken line
indicates the uncorrected coefficient p^ig Ccb) , and the drawn line the coefficient pK'ig
corrected for the carbonate error. In the graphs on the right-hand side, the coefficient,
determined with a cell with the glass electrode js plotted as a function of the pH. The
broken line gives the uncorrected coefficient pK'iQ (cb, gl), the dotted line the coefficient
pK'ig (gl) corrected for the carbonate error and tne straight line the coefficient pK[q
corrected for the carbonate error and the glass electrode. These experiments affirm that
the pK'ig does not vary with the pH, as is established^by definition. On the other hand.
Si ggaard-Andersen [9J explains the variation of the pK'ig with the pH he found, by the
presence of non-dissociated NaCOs besides COl~ ions. His experiments have been carried out
with a glass electrode of the type used by us. By applying the right correction, however,
we have come to the conclusion that the variation of pZig (gl) with the pH should be
explained from the property of the glass electrode itself.

4. CSF, Plasma and Serum

Altogether, of 15 samples of CSF and 12 samples of plasma and sera the coefficient
pFig (cb, gl) was determined at 25 °C &nd 38 °C.

CPS 25 °C

7.55-

•
7.54
•

7.53
•
•V
7.52 •V •
* v

pKig(cb,gl)
7.51

1 pH •
7.50

6.2 6.6 7.0 7.4

Figure 2. The practical coefficient pK[g (cb, gl ) of CSF and plasma (serum) as a
function of pH (see text).

124
group of p^ig (cb, gl) we have taken the mean value, indicated by a. Through it, we have
drawn a line at sight. At pH smaller than 7, p^ig (cb, gl) is constant, and at pH higher
than 7, pK{q (cb, gl) decreases. This decrease is the same as that of the discussed bicar-
bonate solution. From this, we may conclude that the dependence of this practical coeffi-
cient on the pH in the physiologic pH range must be due to the influence of the 2nd dissocia-
tion of carbonic acid on the bicarbonate determination and the non-ideal cell with the glass
electrode for the pH determination.

The p^ig (cb, gl) values of CSF (7.527 at 25 °C and 7.617 at 38 °C) and plasma (7.529
at 25 °C and 7.624 at 38 °C) do not appear to differ significantly, indicating that the
ionic strength of both fluids is equal. So it is also misleading to use the old values
-pK[-\ (cb, gl) = 6.13 for CSF and pK{-\ (cb, gl) = 6.10 for plasma (at 38 °C) , suggesting that
the ionic strength of plasma is higher than that of spinal fluid. For simplicity we propose
to take a mean quantitity for CSF and plasma, to be designated by the asterisk, pK*g.
In the graphs pK*^ is drawn as a function of the pH (dotted line). Finally, we deduced
from eq. (15) and the formula of mentioned dotted line [2] an approximation for pxtn valid
valid in CSF and plasma over the pH range of 6.0-7.8 pH at temperature 15-40 °C.

.I02|^, 0.017877

- 1.21813 - 0.0012(pH-7.0) - 0.0406(pH-7.0)2

The precision of the pK*g value obtained for CSF or plasma of normal subjects, is better
than 0.01 p^ig. Extreme pathological variation in the ionic strength of the plasma, for
example J = 0.160 ± 0.030, causes a variation in pKi^ of about ± 0.015 as verified with
eq. (20).

5. Conclusion

On theoretical and practical grounds, we recommend the use of the practical equation of
the overall 1st ionization equilibrium of carbonic acid as related to the gas phase for
calculations of the bicarbonate concentration (or total CO2 content) of CSF or plasma. At
pH = 7.40 and t = 37 °C the pK*n value is equal to 7.615 for CSF, plasma and serum.

References

[1] Maas, A. H. J., Visser, B. F. van Leeuwen, A. M. and Overbeek, J. T. G.

, An overall
,

equation for the first ionization equilibrium of carbonic acid, Pflugers Arch. 304,
20 (1968).

[2] Maas, A. H. J.^van Heyst, A. N. P. and Visser, B. F., The practical equilibrium
coefficient (p2'ig) for the first ionization of carbonic acid in solutions of sodium
bicarbonate, cerebrospinal fluid, plasma and serum at 25 and 38 °C, Clin. chim. Acta
33, 325 (1971).

[3] Shedlovsky, Th. and Mclnnes, D. A., The first ionization constant of carbonic acid,
0-38 °C from conductance measurements, J. Amer. Chem. Soo. 57, 1705 (1935).

[4] Bartels, H., and Wrbitzky, R. , Bestimmung des C02-Absorptions-koeffizienten zwischen

15 and 38 °C in Wasser and Plasma, Pflugers Arah. 27]_, 162 (1960).

[5] Austin, W. H., Lacombe, E. Rand, P. W. and Chatterjee, M.

, Solubility of carbon
,

dioxide in serum from 15 to 38 °C, J. Appl. Physiol. 18, 301 (1963).

[6] Siggaard-Andersen, 0., ed. The Acid-Base Status of the Blood, 4th revised edition,
,

p. 34 (Munksgaard, Copenhagen, 1974).

[7] Maas, A. H. J., pH determination of body fluids with a micro glass electrode and a
saturated KCl bridge in the cell, Clin. Chim. Acta, 28, 373 (1970a).

125
 determination of actual bicarbonate in
Maas, A. H. J., A Titrimetric method for the
cerebrospinal fluid and plasma or serum, Clin. Chim. Acta, 29, 567 il970b).

carbonic acid as a function

Siggaard-Andersen, 0., The first dissociation exponent of
of pH, Scand. J. Clin. Lab. Invest. U, 587 (1962b).

126
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

TOWARDS A PHYSIOLOGIC NOMENCLATURE FOR IE VIVO DISTURBANCES

OF ACID-BASE BALANCE

Jordan J. Cohen
New England Medical Center Hospital
Boston, Massachusetts 02111, USA

In considering disturbances of acid-base equilibrium from a clinical vantage point, it

would seem axiomatic that a fundamentally physiological frame of reference must dominate
any conceptual scheme offered as an aid to diagnosis and therapy. It is abundantly clear
that both normal physiologic regulation as well as patho-physiologic disturbances of acid-
base equilibrium in the living organism involve a large array of complexly interacting,
dynamic processes. Although the simplest of these processes, that of buffering, can for
some purposes be conceived of in strictly chemical terms, a thorough analysis of even this
process must incorporate many features which are strictly within the domain of physiology
and which are not adequately encompassed by purely chemical notions. I refer here to such
features as compartmental ization recruitment, endocrine influences, "slow equilibration,"
,

membrane transport, and the like. Add to this the physiologic "regulatory" functions of
the respiratory and renal systems, each of which is impinged upon by a veritable welter of
influences, and one can scarcely deny that true comprehension of this system poses an
enormous challenge.

As has been amply discussed, this challenge cannot be circumvented by systems of acid-
base analysis based exclusively on the in vitro behavior of blood [1]^. There would appear,
in fact, to be no alternative to using the empiric behavior of the intact living organism
itself, as the foundation for any systematic and satisfactory analysis of acid-base metabo-
lism.

Beginning with this assumption, with which I would hope we could all agree, our task
would appear to be to develop a nomenclature commensurate with the physiologic complexities
that are known to exist. As a point of departure for developing such a nomenclature, there
are compelling reasons for utilizing the time-honored, carbonic acid/bicarbonate buffer
pair. As we all know, this buffer pair occupies a unique position in the living organism
since it provides the biologic linchpin connecting the chemical process of buffering, on
the one hand, to the physiologic process of acidity regulation, on the other. This unique
property stems, of course, from the fact that carbonic acid and bicarbonate ion concentra-
tions are regulated by semi -independent physiologic control systems. Moreover, extrapolating
directly from the isohydric principle, one can state that alterations in body fluid acidity
can only occur as a consequence of alterations either in carbonic acid or in bicarbonate
concentrations (or both). Disturbances of acidity can be classified with little ambiguity,
therefore, into "respiratory" and "metabolic"^ subgroups.

Considerably more ambiguity is encountered, however, in attempting to develop terms

which connote the directional changes in acidity that are produced by these two classes of
disturbances. In the first place, there is the difficulty encountered with respect to the
secondary changes in bicarbonate and carbonic acid concentration which are evoked by
primary disturbances in the countervening variable. For example, should the secondary
hyperventil ization induced by an initial reduction in bicarbonate concentration be given a
separate term such as "respiratory alkalosis" when it is part and parcel of the organisms

1 Figures in brackets indicate the literature references at the end of this paper.
^Although the term "metabolic" is somewhat narrow in scope in that it fails to communicate
adequately the frequently overriding. influence of renal mechanisms on bicarbonate con-
centration, it is far too well engrained to be easily supplanted.

127
inherent response to this challenge and when, in fact, the patient remains acidemic. Even
more troublesome is the difficulty encountered in certain mixed acid-base disturbances in
which the interaction of more than one primary disturbance, and their respective physiologic
responses, produces truly paradoxical deviations in acidity. For example, we have recently
described an experimental situation in which chronic hyperventilation causes a significant
elevation in hydrogen ion concentration [2].

Based on the considerations that I have all too briefly summarized above, I would
offer the following set of recommendations for a physiologic nomenclature adequate to deal
with in vivo disturbances of acid-base balance.

A. The acid-base status of an individual at a given moment in time should be char-

acterized only by:

1) a direct measure of the "respiratory" component, i.e., carbonic acid con-

centration (H2CO3) or carbon dioxide tension (PCO2) and

2) a direct measurement of the "metabolic" component, i.e., bicarbonate concentration

(HCO3).

B. The level of blood acidity, which is determined by the levels of the co-existing
respiratory and metabolic components, should be quantitated in terms of the actual hydrogen
ion activity, symbolized by pH or [H"*"] or a|^+.

C. Deviations from normal acidity, irrespective of cause, should be termed "acidemia"

and "alkalemia," respectively,

D. Deviations from the normal level of the respiratory component, irrespective of

cause, should be termed "hypocapnia" and "hypercapnia ," respectively.

E. Deviations from the normal level of the metabolic components should be termed
"reduced bicarbonate concentration" (? hypobicarbia) and "elevated bicarbonate concentration"
(? hyperbicarbia) respectively.
,

F. Disturbance of acidity initiated by changes in the respiratory component should

be termed "respiratory alkalosis" and "respiratory acidosis," respectively. Changes in the
respiratory component occurring in response to primary changes in the metabolic component
should be termed "secondary hyperventilation" and "secondary hypoventilation," respectively,
and should not be termed respiratory alkalosis and respiratory acidosis.

G. Disturbances of acidity initiated by changes in the metabolic component should be

termed "metabolic acidosis" and "metabolic alkalosis," respectively. Changes in the metabo-
lic component occurring in response to primary changes in the respiratory component should
be termed "secondary elevations in bicarbonate concentration" and "secondary reductions in
bicarbonate concentration," respectively, and should not be termed metabolic alkalosis and
metabolic acidosis.

H. For the respiratory disturbances, the adjectives "hyper-acute," "acute,"

"transient," and "chronic" should be used to designate the duration of the acid-base
disturbance with respect to the secondary physiologic adjustments in bicarbonate concen-
tration.

1) "Hyper-acute"--the virtually instantaneous, and largely theoretical, interval

prior to the titration of tissue buffers.

2) "Acute"--the interval following the immediate titration of tissue buffers and

prior to a measurable renal contribution.

3) "Transient"--the period during which the renal contribution is accumulating.

4) "Chronic"--the indefinite period following completion of the renal contribution

to the new- steady state.

I. The adjective "uncomplicated" should be used to designate an abnormal state of

acid-base equilibrium produced by a single primary disturbance, including the influence of

128
the secondary physiologic response appropriate to the specific degree and duration of the
primary change.

J. The adjective "mixed" should be used to designate a state of acid-base equilibrium

produced by the simultaneous presence of more than one, independent disturbance.

K. The terms "correction" and "repair" should be used to designate the restoration
to normal levels of the component responsible for initiating an acid-base disturbance.

It should be noted that the terms "compensated" and "uncompensated" have been avoided
in this scheme since these terms are commonly misconstrued to mean success and failure,
respectively, in returning the level of acidity to normal; it is now abundantly clear that
unimpeded physiologic responses evoked by initiating disturbances of acid-base equilibrium
should not be expected to restore hydrogen ion concentration to control levels.

REFERENCES

[1] Schwartz, W. B. and Relman, A. S., A critique of the parameters used in the evaluation
of acid-base disorders. New Eng. J. Med. 268 , 1382 (1963).

[2] Cohen, J. J., Madias, N. E., Wolf, C. J., and Schwartz, W. B., Evidence that the renal
response to hypocapnia is not homeostatic in origin. Clin. Res. 23^, 430A (1975).

129
I

I
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

MINIMAL ACCEPTANCE CRITERIA FOR ACID-BASE NOMOGRAMS

Jordan J. Cohen
New England Medical Center Hospital
Boston, Massachusetts 02111, USA

Acid-base nomograms can be classified into two major subgroups: (a) those providing a
graphic representation of the chemical equilibrium of the carbonic acid-bicarbonate and/or
other buffer systems, and (b) those concerned with delimiting the range of physiologic
responses to be anticipated, on empiric grounds, in response to primary disturbances of
acid-base equilibrium of a specified magnitude.

The first class of nomogram has proven useful as a convenient means for specifying the
value of one (unmeasured) variable when other components of the equilibrium reaction have
been independently analyzed. These nomograms are also helpful in the teaching of acid-base
physiology since they can often be used as a backdrop for the visual representation of the
various pathways of disturbances. Such nomograms have also, on occasion, been modified and
extended to permit evaluation of certain "derived" parameters which have been offered as
aids to the analysis of clinical disturbances of acid-base equilibrium. To the extent that
such nomograms reflect accurately the actual chemical equilibrium upon which they are based,
one can certainly have no quarrel with their use as handy, calculational aids. It is only
when such nomograms make available certain derivative acid-base parameters that are sub-
sequently misused or misinterpreted by the unsophisticated that problems arise. Clearly,
this is not an indictment of such nomograms themselves, but does raise certain risk-benefit
considerations that might provide a useful focus for our discussion. To put my own bias
squarely on the table, let me say that I can find no compelling reason for developing
nomograms or algorithms that permit the calculation of so-called derived parameters since an
adequate and straightforward analysis of acid-base disturbances can be based directly on the
actual concentrations of bicarbonate and carbonic acid. To state the issue in a slightly
different way, the information content required to analyze a clinical acid-base disturbance
is no less when using derived parameters than when using the actual values for bicarbonate
and carbonic acid (PCO2) themselves.

The second class of nomograms, that is concerned with delimiting the empiric physiologic
ranges for the anticipated responses to a given acid-base disturbance, has gained wide
popularity in recent years. This type of nomogram has, of course, been made possible owing
to the availability of experimental data which serve to characterize the response of other-
wise normal living organisms to graded degrees of both respiratory and metabolic acid-base
disturbances. A typical nomogram of this type was published recently by Arbus [1]^. In
order for such nomograms to be used appropriately as a clinical aid, the delimited zones
must be based upon steady-state observations in individuals known to have uncomplicated
acid-base disturbances. Moreover, to the extent that animal data are utilized for such
purposes, it must be clearly understood that extrapolation to human beings is tentative and
speculative. Although the applicability of this type of nomogram to the clinical evaluation
of acid-base disturbances has been widely acknowledged, care must be taken to avoid over
interpretation and over simplification. For example, a common misconception is that values
that fall within the zone ("confidence band") of a given uncomplicated acid-base disturbance
necessarily indicate the presence of an uncomplicated disturbance. In fact, it can be shown
that the co-existence of dual or triple ("mixed") disturbances may cause acid-base parameters
to fall well within a given confidence band [2].

Figures in brackets indicate literature references at the end of this paper.

131
 For this and other reasons, it must be emphasized that such graphic aids cannot replace
the intelligent synthesis of all relevant clinical and laboratory data in achieving a
rational diagnosis and therapeutic plan.

REFERENCES

[1] Arbus, 6. S., An in vivo acid-base nomogram for clinical use. Can. Med. Assoc. J.
109 , 291 (1973).

[2] Arbus, G, S., Hebert, M. D., Levesque, P. S., Etsten, B. E., and Schwartz, W. B,,
Characterization and clinical applications of the "significance band" for acute
respiratory alkalosis. New Eng. J. Med. 280 , 117 (1969).

132
National Bureau of Standards Special Publication 450. Prcc£tci1ngs of a Workshop on pH
Issued June 1977.
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975.

A PHYSIOLOGICAL APPROACH TO ACID-BASE DIAGNOSTICS

Poul Kildeberg, and Knud Engel

Departments of Clinical Chemistry and Pediatrics
Odense University
DK 500 Odense, Denmark

The organism may be conceived as a pluricompartmental buffer solution through which

titratable acid and base, because of a relatively high buffer value (b) and the presence of
specific control mechanisms, are transported at varying rates at a nearly constant extra-
cellular pH. Correspondingly, the clinical acid-base physiologist is concerned with acid
and base rather than with acids and &asee--being more often than not in no position to imply
specific molecular moieties: his principal task is to evaluate extent and control of
hydrogen ion donation in the body, i.e.^ to assess sources of gain and loss, fluxes,
distribution, concentrations, turnover, and exchange of titratable acid or base. The acid-
base physiologist, then, is primarily concerned with the three key variables of buffer
chemistry: pH, B, and titratable acid (TA). In terms of Van Slyke's definition [1],^

pH = - /e'^dTA + pHq (1)

where pHg is the titration end-point.

Amounts of hydrogen ion transfer taking place under circumstances

defined by end-point specification are stated in terms of the acid-base
chemical quantity, TA. Magnitudes of that quantity are expressed as
the product of a dimensionless number (numeric value) and a unit. The
basic unit concerned is (milli) mole of titratable hydrogen ion, conven-
tionally named (milli) equivalent. The use of this unit makes it possible
to describe the total titratable acidity of any known solution as the
algebraic sum of fractional titratable values contributed by individual
specific Br^nsted acids (and bases) making up the system. Although the
milli-equivalent (meq) is being abandoned in current clinical chemical
usage [2], it remains particularly convenient for the buffer chemist.
The meq is derived from the mmol by qualification with respect to
component (H"*") and expressed by multiplying the basic unit by a dimension-
less quantity, N, the value of which depends on the acid properties
(strength) of the molecular moiety concerned. For a given nonionic
acid, meq = mmol titratable H"*" = N-mmol of acid. Accordingly, TA (meq)
= C'N'mmol of acid where C is the numerical factor of the millimolar
concentration. N is the average number of negative charges conferred
per molecule of the nonionic conjugate acid at the neutral pH (pHo). In
general , N is equal to

pH(.-pK, 2pHn-pK,-pK2 i-pH^-pKn ... -pK-

^
10 '
+ 2-10 ^ ' ^ ... + i-lO " '

pHf.-pK, 2pHn-pK,-pK„ i-pHn-pKn ... -pK-

^
1 + 10 ^ ' + 10 '
... + 10 " '

^ Figures in brackets indicate literature references at the end of this paper.

133
 where pK-j is the pK^ value corresponding to the i'th dissociation step of
a pluribasic acid. For an arbitrary solution of an acid and its salt(s),
we get

TA(meq) = N-C^^^^^ - [b""^] - 2[b"-2].... i[B"-'"]

where C^otal ''s the millimolar buffer total concentration and b"~^ are
conjugate bases of the acid HB". At very low or very high levels of pH,
the equation must be extended by the terms +[H''"] and -[OH"].

To the extent that different categories of Br0nsted acids and bases occurring in the
body exibit physiological specificity in the sense that they conform with characteristic
patterns of origin, transport, and elimination, the acid-base physiologist must evaluate
titratable contributions by such categories separately. Therefore, a very fundamental
question is this: Do we know categories of acids and bases the unspecific net titratable
contributions by which are subject to organ-specific control and elimination? Viewing the
organism as a "pH-stat" which within limits of the capacities of the organ systems involved
ensures constancy of the extracellular pH, we may rephrase the question: Do we know
categories of titratable acid and base which are subject to independent control in relation
to a common extracellular pH?

Whereas the total TA of the body fluids (as determined by closed system titration) is
of limited physiological interest, the unique role of cavhonic acid in acid-base metabolism
was early recognized; and the development by Van Slyke [3,4] of methods suitable for CO2
measurement in biological fluids heralded fifty years of acid-base physiological research.
However, as an immediate consequence of the introduction of quantitative CO2 measurements
in biological chemistry, the operational concept of non-oarbonic (non-volatile) acid arose
which has been a constant source of controversy. By means of various modifications of the
CO2 technique, clinical physiologists and clinical chemists have endeavored to express— in
semiquantitative machine language: "CO2 capacity" [5], "CO2 combining power" [6], "alkaline
reserve capacity" [7], "standard bicarbonate" [8], etc. --concentrations of non-carbonic
acid without ever questioning the relevance of this concept. It is characteristic of the
situation that when Astrup, et al. [9,10] finally succeeded in measuring (by CO2 titration)
exact concentrations of titratable non-carbonic acid in blood, an extensive debate evolved
which did not establish any such relevance [11-18]. The difficulties were compounded by a
persistent demand by clinicians for diagnoetioally (physiologically) specific acid-base
measurements.

Measured concentrations of non-carbonic acid (base) in biological fluids include: (1)

titratable equivalents originating in reversible processes of intermediary metabolism or
represented by exogenous (dietary) metabolizable "organic" acid, and (2) titratable equiv-
alents originating in irreversible reactions or gastro-intestinal absorption of non-metab-
olizable acid and base. The kidney contributes to acid-base homeostasis by exclusively
influencing the extracellular concentration of acid (base) belonging to the second category
which in turn cannot be changed by extrarenal mechanisms unless the capacity of the kidney
in this respect is overcharged. Correspondingly, changes in the level of renal function
cannot per se bring about variations in extracellular concentrations of carbon dioxide or
metabolical ly reversible (eliminable) equivalents which depend on pulmonary function and
metabolic processes, respectively. Physiological coupling of variations in extracellular
concentrations of titratable carbonic acid (CA = 0.9522'[C02]tQtal ) metabolizable and
non-metabolizable, non-carbonic acid takes place by the operation of renal, metabolic, and
respiratory control mechanisms in relation to a common "set-point" pH. For example, a
primary rise in the extracellular concentration of non-metabolizable base will elicit a
corrective response by the kidney and compensatory responses by the respiratory centers
[18a] and soft tissues [19-22].

1. Definitions
The chemical concept of non-carbonic acid is devoid of physiological specificity:
implicit in current knowledge exist three rather than two relevant categories of (titratable)
acid and base [22],

134
 Titratable acid (TA) (chemical language)
'

I 1

MA NA CA (physiological language)

(soft (kidney) (lung)

tissue)

where MA, NA {"net aoid") , and CA represent titratable values of metabol izable acid, non-
metabolizable, non-carbonic acid, and carbonic acid, respectively. Each of these three
categories of titratable acid may be given an explicit operational definition:

Net acid (NA) s titratable acid on titration to an end-point at pH 7.40, temperature,

37 °C, ionic strength of normal blood plasma, and CA = MA = zero meq/^.

Metabol izable acid (MA) e titratable acid on titration to an end-point at pH 7.40, tem-
perature 37 °C, ionic strength of normal plasma, and NA = CA = zero meg/£.

Carbonic acid (CA) = titratable acid on titration to an end-point at pH 7.40 temperature

37 °C, ionic strength of normal plasma, and NA = MA = zero meq/£.

Obviously primary definitions are required for two of these quantities. Carbonic acid is
quantitatively defined as the titratable value of the stoichiometrical CO2 content/con-
centration of a given system. Disregarding carbamination and "CO2 fixation" we have (pK^ =

6.11).

1.0502-CA =
(C02)tQtal
" + H2CO3 + HCO^ + CO3

where CA is equal to the millimolar content of bicarbonate plus twice the millimolar content
of carbonate at pH 7,40. Also, for each system or organism concerned, a decision must be
made with respect to the distinction between rsversihle and non-reversible equivalents.

2. Methodology

Concentration of CA are measured a.m. Van Slyke or by the equilibration method [10].
Concentrations of NA may be determined by direct titration following in vitro ashing [23,24].
This method, however, assumes an identity between (the extent of) in vitro and in vivo
processes of oxidation which does not always obtain, particularly not in the growing organism
where neutral sulfur is stored in new protein. A more convenient approach, which has the
further advantages of allowing separation of "addition disturbances" from "subtraction
disturbances" and of statements of titratable values at specified levels of ionic strength,
is based on the principle of electroneutrality. Thus, for any sample the concentration of
NA is given by

- ^^^An-]^^ - EKjCat^]^^ (2)

where [] are stoichiometrical millimolar concentrations. Kg and Kc denote the average

number of negative and positive charges, respectively, per molecule at pH 7.40; and the
^
subscript "nm" signifies non-metabolizable ionic species (in biological fluids largely Na ,

K+, Ca++, Mg"^"^, CI", SOi;, and phosphate). Accordingly, the balance of net acid [22,25-31]
is given by

NAB = 1.8.Bp + B^^. - B^j^, - B^, - 2 • B^^,, - 2 • B^^,,

(3)

135
where "B" are millimolar balances representing differences between oral and parenteral
intake and renal and gastrointestinal losses. It will be noted that eq. (3) assumes iden-
tity between endogenous production (gain) and renal excretion (loss) of sulfate ion. It is
apparent that the concept of NAB is deeply embodied in the older medical acid-base litera-
ture [30] and that no genuine discrepancy exists between the older "anion-cation termino-
logy" and the acid-base chemical concepts of "acid" and "TA".

Finally, MA is obtained as the difference between NCA, determined by titration at a

P-ri of zero mm Hg, and NA.

3. Normal Acid-Base Metabolism

As a result of endogenous metabolic activity, very large quantities of CA and MA are

subject to exchange and cyclic turnover, respectively, at an extracellular pool size (CA
+ MA) of about 525 meq in the normal adult. Because the pH of extracellular water is
approximately 7.42, an extracellular pool of net base of the same magnitude must be
maintained. The kidney accomplishes this by continuously releasing hydroxyl ions (ac-
companied by non-metabolizable cation, i.e., net base) to the peritubular venous return at
a controlled rate, in the adult about 6300 meq/24 hours. Because in the physiological
range of pH free hydroxyl ions exist only in negligible concentrations, OH" ions generated
by the renal tubules are instantaneously and veversibly neutralized by endogenous CA and
MA. Extracellular net base, therefore, is represented by bicarbonate and "metabolizable
organic anion" (MOA"). Of the net base concentration of normal plasma (41 meq/£), ap-
proximately 17 meq/Ji or 41 percent are represented by MOA" (partly protein). Net base
absorbed from the gastrointestinal tract and net acid (sulfuric acid, uric acid, bilirubin,
etc.) and net base (urea, creatinine, etc.) originating in irrevevsihle metabolic reactions
as well as net base communicated by bone contribute to the extracellular pool. The net
effect of such processes is normally given by the rate of renal acid excretion, equal to
the difference between the rates of tubular net base generation and glomerular net base
loss [22].

Urinary net acid is partly represented by the metabolizable cation, NH^.

By generating ammonia, the kidney does net ultimately remove hydrogen
ions from the body fluids. It does, however, replace extracellular NA
by (somewhat larger quantities of) MA which can be eliminated by extrarenal
mechanisms.
12365
A

H2CO3 MOA

OH +H2CO3 + MOA
-10
V 7365 w
OH" [h 003] + [mo A"]
(24) , (17)

6310 5000 1300 [hco3]+[moa]

(27) (8)
75

-75
NH.

-10

Figure 1. A quantitative outline of acid-base metabolism. P = plasma

compartment, g - glomerular membrane, F = glomerular ul traf i 1 trate.

136
 The acid-base homeostatic "capacity" of the kidney is reflected by the fact that
renal net base exchange (-20 - 0 meq/24 hours) amounts to only a fraction of a percent of
the rate of renal net base turnover. An outline of the gross quantitative aspects of
acid-base metabolism is shown in figure 1.

4. Acid-Base Status: Clinical Evaluation

The term "acid-base status" implies a set of relevant variables which adequately
characterize the acid-base composition of a given sample medium. The acid-base status of
biological fluids (blood, plasma, urine, cerebrospinal fluid, milk, gastric juice, etc.)
lould always be expressed in the same chemically unambiguous terms.

CA [B]
NA TA pH = pKg + log (4)
MA [A]

where the right-hand equation refers to selected relevant acid-base pairs, usually the
HCO3/H2CO3 system. Representative values for some biological fluids are listed in table L
The relationship between the quantities recommended here and some commonly used plasma
acid-base variables is given by

TA - CA = NCA = NA + MA = NBB - BB - K = -BE - K (5)

where NCA is titratable non-carbonic acid, NBB is "normal buffer base" [10], BB is "buffer
base" [32], BE is "base excess" [10], and K is a constant, approximately 23.8 meq/£.

Note: If 25 mmol of CO2 (pKi 6.11), 17.2 mmol of lactic acid (or 103 g
of albumin, TAgib " 0.167 meq/g) and 41 mmol of sodium hydroxide are
added to one liter of water in a closed system, the resulting solution
will have a pH of 7.40, a P-^ of 40 mm Hg, and a BE of zero meq/£.
LU2

Table 1. Representative values for acid-base status

of some biological fluids.

CA MA NA TA pH
(meq/L)

Blood plasma^ 24.0 17.0 -41 .0 0 7 4

Whole blood^ 21.7 1.0 -17.4 5. 3 7 2^

Normal acid urine 1.2 0^ 5 6. 2 5 8

Human milk 4.9 25.0 -24.0 5 9 6 6

Cow's milk 4.1 66.0 -60.0 10. 1 6 5

arterial ized whole blood, hemoglobin concentration 15 g/100 ml

filtered metabol izable acid minus secreted metabol izable base (NH3);
50 - 50 = 0 meq/l
hemolyzed blood

137
 The pH-stat concept implies three mutually non-exclusive causes of deviations in
extracellular pH, applicable to each regulatory organ system involved: set-point de-
viation, functional errors, and overcharging by excessive loads. Because of the com-
plexity of the biologic pH-stat careful operational definitions of these are mandatory.
-eve-vekarg-inct- Vie- may undeics-tand "loading with acid.,.or-Jias€-r €xogenous -wiit. respect to the
regulatory organ considered and sufficient to produce a stable change in extraclTllilar pH.
Malfunction may be taken to imply intrinsic disease of the organ concerned; and--by
exc^us^ on- -deviations in set-point may be indicated by functional changes in the absence
of (local) disease and acid-base loading. In such terms, diabetic keto-acidosis represents
malfunction of soft tissues and overcharging of kidney and lung; obstructive or restrictive
lung disease with chronic hypercapnia and a normal extracellular pH represents malfunction
of the brain-lung system; and steriod alkalosis and acute salicylism involve set-point
deviations for kidney and lung, respectively.

Any such combination of causes may lead to a primary change in the TA of extra-
cellular water, the associated change in pH setting off secondary processes of correction
and compensation. The resulting ATA is the algebraic sum of concurrent changes in the
concentrations of CA, MA, and NA. Each such change is brought about by a combination of
processes of retention and redistribution and changes in the volume of solvent, due in
turn to osmotic redistribution of water across the cell membrane or to changes in water
balance. Balance changes may be accounted for by identifying several sources of input and
output. This approach to systematic evaluation of clinical acid-base disturbances is
shown schematically in figure 2. As an example, figure 3 shows the suggested steps in the

set mol-
point funct load CAUSES
1
(etiology)

sec prim sec

corr A\A comp
1 1

1
prim sec. changes result
result Z\NA AHA
pH AT A
OBSERV. solv
distr bal vol
(din. chem.

input urine bal distr

T
r
acid base ure-
MECHANISM gain loss RTA mica
out (physiology)
input put bal distr
rate 'grad NH3
Mm' lim' def?

Figure 2. Interpretative steps in the Figure 3. Evaluation of a primary

evaluation of the acid-base status. rise in the concentration of
extracellular net acid.

diagnostic interpretation of an observed rise in extracellular NA, which on the basis of

the pattern of the acid-base status of blood, the total clinical situation of the patient,
and existing physiological knowledge is considered to reflect primary accumulation of NA
in the extracellular compartment. It may be pointed out that a normal acid-base status of

138
blood does not preclude a significant disturbance of acid-base metabolism. For example,
in hyperparathyroidism primary redistribution of body net base (release of skeletal base)
may lead to negative net base balance in the absence of significant change in the acid-
base status. Because a primary rise in extracellular NA rarely if ever is brought about
by processes" of" redistribution, such changes may be evaluated in terms_of the -balance of
net acid and changes in the volume of solvent ("contraction alkalosis" and "dilution
acidosis", respectively) [33,34]. Contraction alkalosis due to negative water balance and
a secondary readjustment of the renal tubular set-point contributes significantly to the
acid-base status in "gastric alkalosis" [25]; and extracellular dilution acidosis due to
osmotic redistribution of water across the cell membrane can be produced experimentally by
infusion of hypertonic solutions of aprotes such as sodium chloride and glucose [35]. A
positive balance of net acid is a frequent cause of primary increases in extracellular NA.
Defining the net acid output as urinary net acid [22,31], and output disturbance leading
to NA accumulation is by definition a renal acidosis of which two types are generally
recognized: the acidosis of uremia and renal tubular acidosis (RTA). Renal tubular
acidosis in turn may be due to failure to establish an adequate trans-tubular pH gradient
("distal" or "gradient limited" RTA [36]), to a relatively low rate of (proximal) tubular
net base generation ("proximal" or "rate limited" RTA [37]), or possibly to impaired
ammonia excretion [38]. Finally, input disturbances involve extrarenal sources of net
acid, in particular oral intake of NA ("addition acidosis") and gastro-intestinal losses
of net base ("subtraction acidosis").

References

[I] Van Slyke, D. D. On the measurement of buffer value and on the relationship of buffer
,

value to the dissociation constant of the buffer and the concentration and reaction of
the buffer solution, J. Biol. Chem. 52, 525 (1922).

[2] Dybkaer, R. and J0rgensen, K. Quantities and Units in Clinioal Chemistry^ including
,

recommendation 1966 (Munksgaard, Copenhagen, 1967).

[3] Van Slyke, D. D. Studies of acidosis.

, II. A method for the determination of carbon
dioxide and carbonates in solution, J. Biol. Chem. 30, 347 (1917).

[4] Van Slyke, D. D. and Neill, J. M., The determination of gases in blood and other
solutions by vacuum extraction and manometric measurement, J. Biol. Chem. 61_, 523
(1924).

[5] Henderson, Y. and Haggard, H. W. , Respiratory regulation of the CO2 capacity of the
blood, J. Biol. Chem. 33, 333 (1918).

[6] Van Slyke, D. D. Studies of acidosis.

, XVII. The normal and abnormal variations in
the acid-base balance of the blood, J. Biol. Chem. 48, 153 (1921).

[7] Van Slyke, D. D. and Cullen, G. E., Studies of acidosis, J. Biol. Chem. 30, 289 (1917).

[8] J0rgensen, K. and Astrup, P., Standard bicarbonate, its clinical significance, and a
new method for its determination, Saand. J. Clin. Lab. Invest. 9_, 122 (1957).

[9] Astrup, P., J0rgensen, K. Siggaard-Andersen

, , 0., and Engel , K, , The acid-base metabolism.
A new approach, Lanoet, !_> 1035 (1960).

[10] Siggaard-Andersen, 0., The Aaid-Base Status of the Bloody 4th revised edition (Munksgaard,
Copenhagen, 1974).

[II] Astrup, P., Acid-base disorders. New Eng. J. Med. 269, 817 (1963).

[12] Bunker, J.
(1965).
P., The great trans-Atlantic acid-base debate. Anesthesiology, ~
26, 591

[13] Report of the ad hoo Committee on Acid-Base Terminology, the New York Academy of
Sciences Conference, Nov. 23-24, 1964, Ann. N.I. Acad. Sci. 133, 251 (1966).

[14] Crawford, J. S. and Holaday, D. A., Acid/base disturbances. Lancet, 1, 834 (1964).

139
[15] Schwartz, W. B. and Relman, A. S., A critique of the parameters used in the
evaluation of acid-base disorders. "Whole-blood buffer base" and "standard bicarbonate"
compared with blood pH and plasma bicarbonate concentration. New Eng. J. Med. 268 ,
1382 (1963).

Siggaard-Andersen, 0., Acid/base disturbances, Lancet, 1104 (1964).

Siggaard-Andersen, 0. and Engel , K. , Acid-base debate, Anesthesiology, 2J_, 202 (1966).

Winters, R. W. , Terminology of acid-base disorders, Ann. Int. Med. 63, 873 (1965).

[18a] Kildeberg, P., Respiratory compensation in metabolic alkalosis. Acta Med. Soand. 174 ,

515 (1963).

Berry, M. N. and Scheuer, J., Splanchnic lactic acid metabolism in hyperventilation,

metabolic alkalosis and shock. Metabolism, 16^, 537 (1967).

Eldridge, F. and Salzer, J., Effect of respiratory alkalosis on blood lactate and
pyruvate in humans, J. Appl. Physiol. 71, 461 (1967).

Haldi, J., Lactic acid in blood and tissues following intravenous injection of sodium
bicarbonate, Am. J. Physiol. 106, 134 (1933).

Kildeberg, P., Quantitative acid-base physiology. System physiology and pathophysiology

of the hydroxyl ion, in preparation (1975).

Camien, M. N; and Gonick, H. C, Relationship of renal "net acid" excretion to titratable

ash-acidity (ashrTA) in diet and feces, Proc. Soo. Exper. Biol. Med. (N.Y.), 126 ,

45 (1967).

Gonick, H. S., Goldberg, G., and Mulcare, D., Reexamination of the acid-ash content of
several diets, Amer. J. Clin. Nutr. 2]_, 898 (1968).

Degn, J. K. , Wamberg, S., Engel K., and Kildeberg, P., Metabolic alkalosis in obstructive
vomiting. Volume depletion and balance of net acid. Acta Paediat. Soand. 63^, 537 (1974).

Goodman, A. D., Lemann, J., Jr., Lennon, E. J., and Relman, A. S., Production, excretion
and net balance of fixed acid in patients with renal acidosis, J. Clin. Invest. 44,
495 (1965).

Kildeberg, P., Engel, K. and Winters, R. W., Balance of net acid in growing infants.
,

Endogenous and trans-intestinal aspects. Acta Paediat. Scand. 58, 321 (1969).

Lemann, J., Jr., Lennon, E. J., Goodman, A. D., Litzow, J. R., and Relman, A. S., The
net balance of acid in subjects given large loads of acid or alkali, J. Clin. Invest.
44, 507 (1965).

Lennon, E. J., Lemann, J., Jr., and Litzow, J, R., The effects of diet and stool
composition on the net external acid balance of normal subjects, J. Clin. Invest. 45,
1601 (1966).

Shohl , A. T. and Sato,. A., Acid-base metabolism. I. Determination of base balance,

J. Biol. Chem. 58, 235 (1924).

Wamberg, S. T., Engel, K., and Kildeberg, P., Balance of net base in the rat.
Reference values in relation to growth rate, Biol. Neonat. 28_, 171 [1976).

Singer, R. B. and Hastings, A. B., An improved clinical method for the estimation of
disturbances of the acid-base balance of human blood. Medicine, 27, 223 (1948).

Cannon, P. J., Heinemann, H. 0., Albert, M. S., Laragh, J. H., and Winters, R. W.,
"Contraction" alkalosis after diuresis of edematous patients with ethacrynic acid,
Ann. Int. Med. 62, 979 (1965).

140
< [34] Winters, R. W. Scaglione, P. R., Nahas, G. G., and Verosky, M., The mechanism of
,

acidosis produced by hyperosmotic infusions, J. Clin. Invest. 43, 647 (1964).

[35] Sotos, J. F., Dodge, P. R., and Talbot, N. B., Studies in experimental hypertonicity.
II. Hypertonicity of body fluids as a cause of acidosis, Pediatvios, _30, 180 (1962).

[36] Morris, R. C, Jr., Renal tubular acidosis. Mechanisms, classification and impli-
cations, mew Eng. J. Med. 281, ^405 (1969).

[37] Rodriguez-Soriano, J. and Edelmann, C. M., Jr., Renal tubular acidosis, Ann. Rev. Med.
29, 363 (1969).

[38] McCance, R. A., Matheson, W. J., Gresham, G. A., and Elkinton, J. R., The cerebro-
ocular-renal dystrophies: a new variant, Arch. Dis. Childh. 2S_, 240 (1960).

141
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

CO2 SOLUBILITY, pK' AND RELATED FACTORS IN ACID-BASE BALANCE

William H, Austin
Maine Medical Center
Portland, Maine, U.S.A.

1. The Role of pK', CO2 Solubility, and Temperature

in Acid-Base Measurement

An appreciation of the problems of pK' and CO2 solubility necessitates an understanding

of some of the features of the Henderson-Hasselbalch equation that are not always emphasized.
With this in mind, a brief restatement of this equilibrium reaction, which provides a
system for the examination of acid-base balance, is made here. The thermodynamic equation
may be derived from the equilibrium reaction, starting at the partial pressure of CO2 to
the dissociated hydrogen ion activity.

PCO2 ^ dissolved CO2 ^ H2CO3 H"^ + HCO3 (1)

The first consideration, usually not made a part of the familiar constant, is the solubility
of CO2 itself.

The first step involves the uptake of CO2 in the plasma. The constant which is neces-
sary for the understanding of this phenomenon is the solubility coefficient a., (Bunsen's
coefficient) [1]^, or "S", the solubility factor [2]. Neither appears in the collected
constant (pK'), though it is mathematically possible [3,4]. Some consideration has been
given to this, but classically it is excluded from the "overall" constant. The latter,
often called the first dissociation constant, or apparent pK' is in reality a collection of
,

the ionization and hydration constants and will be treated later. CO2 solubility may be
given as Bunsen's solubility coefficient a, (ml C02/ml plasma), often anglicized as "a", and
frequently mistaken for the CO2 factor "S", which is millimoles of COa/mm HgPC02.

Some use a, "a", and "S" interchangeably, as noted recently [5]. This point should
be clarified. The original solubility of 0.541 ml/ml [6] was determined by Bohr in 1905,
and has an interesting history. Van Slyke [7], in 1928, restudied the solubility, using
apparently four samples: two of ox serum, one of normal human serum, and one of normal
human oxylated plasma. He obtained values of 0.509, 0.511, 0.506, and 0.511 ml C02/cm3
for these respective samples. The more familiar figure of 0.510, or "S" = 0.0301, is there-
fore in reality an average of solubilities on ox and human plasma and serum with a water
content of 0.925 g/cm^ at 38 °C. Eisenman, et at. [8], Danowski and Gilmore [9], and we
[10] found different water contents, and used 0.935 in our solubility studies. The recalcu-
lation of Van Slyke's data, using a higher water content, interestingly enough yields values
almost identical to ours, as referred to in the "Handbook of Physiology" [11], and the Ad
Hog Committee on Methodology of the New York Academy of Science in 1966 [12]. This value is
"S" = 0.0306.^

^Figures in brackets indicate the literature references at the end of this paper.
2
Van Slyke astutely recognized [7] that the actual value of a or S is irrelevant if it is
used with the appropriate pK' Note that "ml" or "cc" is used exactly as cited in the ori-
.

ginal papers.

143
 For years, the variation of CO2 solubility in plasma was assumed to CO2 parallel that
of water. This was found not to be the case (fig. 1) [10,13], and Severinghaus and others
made appropriate corrections which were reflected in pK's, especially at the lower tempera-
tures [10,13]. The variations at 37 °C are ^mall by clinical standards, and should not pose-
any problems-.

Water--Bartels et ai.
"' Extrapolated serum--Dill & Forbes 1.00
Extrapolated plasma--Severinghaus
et al .
. .95
Human serum--author
Ox plasma--Bartels et ai.
.90

.85
c
0
. .80 +->
3
sol

ml

s.
.70 <u
0.
CN
.65 0
0
•
.60

.55

. .50

"20"
15 25 30 35 40
temperature °C

Figure 1. Solubility of carbon dioxide in water, serum and

plasma from 15 °C to 38 °C. Solid line indicates CO2 solu-
bility in plasma does not bear a constant relationship to
water (top line), hence to extrapolated values for plasma
CO2 (Severinghaus). (Reproduced, with permission, from the
J. Appl. Physiol. 19, 893-96, 1964).

Attendant alterations in disease states, blood constituents, viscosity, and ionic

strength could conceivably alter CO2 solubility. These have been considered [14-19], and
if the solubility factor is assumed, for the sake of discussion, to be fixed, then these
changes, if any, should be reflected in the pK' This is the constant for the balance of
.

the equilibrium reaction:

Dissolved CO2 HpCOc Hco; (2)

To interrelate the solubility and these reflected variations in pK' , the origin of this
constant must be examined. From this "overall" reaction, a pK' may be derived.

The "overall" equilibrium constant falls into two sets of reactions:

1. The Hydration reaction:

dissolved CO2 =^ H2CO3 which may be quantitatively expressed as (C02)K. =
(H2CO3) with as the hydration constant.

2. The Ionization reaction:

H2CO3 ^ H + HCO3 - quantitatively expressed as (H2C03)K.
with K. as the ionization constant.
= (H )(HC03)

144
These terms may be expressed as concentration or activity. In actual practice, _H (or pH)
is an activity, since this is what the electrode measures, and CO2 content (HCO3) is concen-
tration, usually from gasometric measurements. Titration values for HCO3 are less commonly
used. For this presentation, the practically measured pH (or H+) will be an activity, and
bicarbonate-related entities will be concentrations. Since both dissolved CO2 and (H+) +
(HCO3) are in equilibrium with (H2CO3), the hydration and ionization reactions may be equated,
and constants collected.

(K^ X K. = K'), resulting in (H"^) = K' x (C02)/(HC03) (3)

Although this derived equation is theoretically correct, it does not correspond to the
experimental operations involved in the determination of the "overall" equilibrium constant
(K' or pK'). One experimentally fixes the CO2 concentration by equilibrating blood, plasma
or serum with a given partial pressure of carbon dioxide gas. This operation fixes the
concentration of dissolved carbon dioxide plus the true carbonic acid in proportions deter-
mined by the hydration coefficient K^. This being the case, the numerator "CO2" of the
equation must be CO2 (dissolved) plus H2CO3 (actual), as this is what the analytic determi-
nation produces, rather than what the thermodynamic determination alone gives. There are
a few bicarbonate ions^ which arise, but they arc insignificant. The "overall" reaction
must be rewritten as:

(H^) = K' . (CO2) + (H2CO3) /(HCO3) (4)

Since it has been recognized that this relationship is more properly expressed logarithmical-
ly, one may rewrite the equation as:

(HCO^)
pH = pK' + log (5)
(CO2) + (H2CO3)

The term (CO2) + (H2CO3) is often referred to as H2CO3, creating the false impression that
all of the denominator of the Henderson-Hasselbalch equation is carbonic acid as such. The
dissolved CO2 may be potential H2CO3, but not actual carbonic acid. This has been a point
of confusion to many students of carbonic acid chemistry.

The use of pK' , rather than K' , has been subjected to the same criticism as the use of
pH and H"^ concentration; however, the logic is the same. The most cogent, and sometimes
less appreciated, argument has been cited by Davis [20]. Essentially, he notes that a more
meaningful description of the equilibria constants is given in their logarithmic form. This
is not an accident, nor is it pure convenience nor perpetuation of a misconception. The
thermodynamic behavior of a substance is directly related to a potential, the logarithmic
function of activity--in this case the log (negative) of the constant K', or pK'

Davis has expressed it well, "...the most meaningful property describing dissociation
equilibria of acids, whether in chemical or physiological systems, is not the dissociation
constant itself, but the logarithm of this dissociation constant, pK' , which is directly
proportional to free energy of dissociation and clearly a measure of the strength of an
acid in aqueous solution." Buffering capacity, employing the constant being discussed, is
related to the free energy of dissociation, and is not coincidental to the units used [20,21].

^Virtually all of the bicarbonate which appears in actual practice arises from the salt of
the weak acid, in this case H2CO3. The origin of the bicarbonate does not affect the logic
of the derivation; however, the large quantity of these ions from sodium bicarbonate is very
important to the buffering capability of the bicarbonate buffer system.

145
 Reasoning in terms of the reactants may be a more appropriate approach. Chemical and
physiological activities are more precisely related to the potential, which in turn is a
direct function of pH. Therefore, participation of hydrogen ions in physical and physiologi-
cal reactions is determined by ionic potentials, which are logarithmic expressions of ionic
concentrations. (This may be extended to include the other reactants. The Nerst equation,
which deals with potentials, applies for similar entities, chemically determined, which
react physiologically to the logarithm of molarity.)

Digressing again to the example of pH, which may be more easily visualized, but appli-
cable to pK, one finds hydrogen ion concentration can never really be derived accurately
from activity. In theory, there may be pH changes with no H""" change, if there is an indepen-
dent change in activity coefficients [20]. pH, hence H+, is an apparent entity. pH results
from emf with an electrode calibrated with buffers of "assigned" values. Though assigned,
these values are probably very accurate; however, the pH one gets when blood is measured is
relative. The blood or plasma is a biological fluid, differing greatly from the buffer.
The pH "number" is obviously only relative. There must be distortion 'also of the "real" pH
through inherent errors in the measurement itself, apart from the buffer. Studies of liquid-
junction potentials show this [22], so how, especially in view of the other points, can one
arrive at a meaningful H"*" concentration?

The argument of putting acid-base measurement in familiar terms is a pitfall. One

takes this entity, calls it a concentration or equivalent, and immediately assumes that the
physiologist or clinician can now understand it and can relate this value to other parameters
The end result is like putting a planet and a grain of sand side by side and saying, "Now we
are dealing with mass, and all is clear." The student, through working with concentration,
has to be constantly aware of the diversity of the magnitude of concentration and, not as
has been the case, think of H"*" in the same scope as Na+ or HCO3 concentrations. It is easy
to see how the dimensions of H+ can be thrown out of proportion after thinking of Na"*",
HCO3, CI", K+, etc. It is too easy to lose track of the diminutive nature of hydrogen ion
concentration, especially since it is the intensity of acidity that we are seeking. A
change in using the inexact concentration concept can only foster this, whereas pH has the
dimensionless nature which would diminish this tendency without the need for clarifying that
H''' is not a true concentration, and is calculated or arrived at with the many reservations
mentioned above, keeping in mind the magnitude of the quantity.

Also, conversion to concentration does not have the advantage of keeping all of the
parameters in the same dimension. In addition to the gigantic difference in magnitude, we
are now accustomed to using PCO2, a far more easily acceptable entity than H2CO3, which is
not without its conceptual problems, (i.e.^ H2CO3 is really dissolved CO2 and true HjCOg—
a point of much misunderstanding). It makes no sense to say that H"^ concentration is similar
to bicarbonate concentration (HCO3), and then relate these two to a pressure (PCO2), espe-
cially if this is done in the name of clarity. Filley [23] has stated, with references, an
excellent analysis of this controversy.

Since it is more useful to use the partial pressure of carbon dioxide (PCO2) in the
equation, the solubility factor must be introduced to make (CO2) + (H2CO3) = PCO2. This
gives to the familiar:

(HCO^)
pH = pK' + log (6)
PC02- S

or, for pK' determinations made in the studies discussed:

CO2 content
pK' = pH - log ( 1) (7)
S • PCO2

'CO2 content determined gasometrically in most studies cited above.

146
The original pk"s determined by Severinghaus in 1956 [16] have been widely used. A value
for pK' has been determined in many studies, and has been said to be a function of pH [16,24]
and temperature [10], and are affected by disease states [17] and biochemical abnormalities
[151. Re-examination of pK' strongly indicates to us that neither pathologic conditions
[26] nor abnormal blood constituents [15], exclusive of temperature, alter the pK' in a
practical way. In fact, we could not find significant variations with pH within the clini-
cal range [27]. This is actually similar to the classical studies of Severinghaus [16], and
is confirmed by Nunn [28]. By significant, we would mean less than 0.01 between 7.10 and
7.70 Cfig. 2). This is not to say that electrodes and other factors described by Maas [4],

pK' = 6.132 -0.0042 pH± O.OII

•
• • -
• •
•

rr^.
•
— • • • •

pK' • •

pH
1 1 1
1 1 1

7.10 7.20 7.30 7.40 7.50 7.60 7.70

Figure 2. pK variations with pH in human blood. (Reproduced, with permis-

sion, from the Archives of Internal Medicine , 126 , 699 1970).

Severinghaus [16], and Siggaard-Andersen [24] do not give rise to pH/pK variations, but
that for all practical purposes, a pK' of 6.10 on whole blood pH (not corrected for "suspen-
sion", liquid-junction effect) plasma or serum CO2 content, gasometrical ly determined, and
fixed, tonometered CO2 pressures, is suitable for clinical, and even research, purposes at
37 °C [27]. It is of interest that pK' determinations made using the CO2 electrode revealed
the same pK', but with a much wider standard deviation. The pK's are termed "operational",
since no attempt is made to estimate or calculate pHs of plasma because only precise values
for whole blood pH were known, and "correction" could not be satisfactorily made despite
estimated differences of 0.01 by some [2]. CO2 content may be determined accurately by the
Van Slyke method, though all of the CO2 is not from HCO3 or dissolved CO2. Carbonate and
carbamate certainly figure into the picture, but are of questionable significance in the
overall picture.

The PCO2 electrode, although not used as the primary method of classical assessment of
partial pressure of CO2 for pK' work, has shown limitations which, though small, might be
considered by some to be important.

The magnitude of the variation in PCO2 measured by electrode may be seen in an indirect
way by comparing pK and S.D., using actual PCO2 measurements and electrode-measured PCO2.
This reflection in the S.D. of the pK is from ±0.013 to ±0.032 [26]. In terms of actual
partial pressure, the standard estimate of the error for actual and electrode PCO2, measured
from approximately 20 mmHg to 50 mmHg, is ±3.5 mmHg. Other studies over a larger range,
with 142 samples, give slightly smaller but significant differences of ±2.11 mmHg [29].

147
 Figure 3 breaks down values for pK' as determined in patients and in normal samples by
tonometering high and low CO2 concentrations. The significance is in the similarity. Rigid
techniques were observed, as in previous studies [26], and significant differences were not
present. This is very consistent with another fifty determinations, using randomly selected
patient blood (fig. 4). Whole blood pH and electrode PCO2 measurements were used with Van

PK'
TOTAL (56)
5% CO2 10% CO2
NORM, and ABNORM
6.101 ± Oil
ABNORM
NORM.and
6.098± .010 6.104+ .011

ABNORMAL
6.102
/A
±
(28)
.013
NORMAL
6.101 ± .009
(28)

(14) I0%C02 (14) 5% CO2

6.104+ .013

(14)
/
5% CO (14)
\ I0%C02
6.097 ± .009

6.101 ± .013 6.104 ± .008

Figure 3. pK's derived from blood of normal and ill patients by tonometry
with gas mixtures of approximately 5 percent and 10 percent. No apparent
differences noted under various circumstances.

6.03 6.04 6.05 6.06 6.07 6.08 -6.09 6.10 6.11 6.12 6.13 6.14 6.15

PK'
Figure 4. pK' determinations derived from tonometered whole blood with
fixed partial pressures of CO2 and with PCO2 determined by PCO2 deter-
mined by PCO2 electrode. Hatched boxes represent two samples, one for
each category. Note the narrow S.D. with, the tonometered method.

148
Slyke CO2 contents as above. The mean was 6.099. Again, the only difference is in the
S.D, of ±0.030, presumably due to CO2 electrode variations. When twenty-eight of these
samples were examined with the tonometry method, the mean was 6.012 and the S.D. was ±0.013.

Perhaps the most cogent point arising from the pK' studies relates to the calculated
CO2 contents' relationship to actual, directly determined Van Slyke C02's. This variable (or
HCO3) is most frequently used at present, and it is significant in that 54 of 59 samples
(most referred to in studies presented here) are within 3 millimoles/liter of the directly
determined figure, with the maximum spread of 5 millimoles/liter occurring only once. This
is seen with the simple uncorrected pK' of 6.100. Using pK's corrected to pH [10,11] changes
the calculated values by less than 1.0 millimole in all instances.

A relatively unexplored, but not totally unrealistic, concept would involve the inclu-
sion of the solubility factor into the pK' itself. Maas [4] has explored this, and Van
Slyke [7] himself alluded to this interrelationship many years ago. Simple, logarithmic
conversion of "S" (0.0306 to 1.50) added to 6.10 would yield 7.60 as the pK' , the only
constant in the entire equation. In effect, this is what we actually do {i.e., use an
appropriate pK' for a given "S"); however, conceptual and logistical factors make this claim
of simplification debatable.

Measurement, by agreement, may be made at 37 °C. Buffers and calibrating gases will
not be treated here, but the current practice of using pH/C factors of -0.0148 (0.0150)
[30-32] and temperature/PC02 factors [33,34] can persist and yield universally acceptable
values. Since Nunn [28] and Severinghaus [33] have good evidence of the stability of CO2
content at different temperatures, this allows determinations of pH and PCO2 at 37 °C and
calculations of CO2 content (or bicarbonate) which are still valid where pH and PCO2 are
corrected. The other route of correcting pH and PCO2 to body temperature then, with appro-
priate solubility factors and pK's calculated to the third variable [11,32] (HCO3 or CO2
content), is also valid but not necessary, and it is cumbersome. Tables [36], slide rules
[35], and diagrams [37,38] will simplify the calculations, and whichever is preferred may be
acceptable, as long as it yields the same answers for the calculated third variable. This
matter, again, is under scrutiny; here standardization should be achievable.

2. Summary

We must use the most refined techniques: stable, sensitive electrodes and electrometers,
accurate buffers, good CO2 electrodes and calibration, and careful sampling and storing--all
of which are being discussed, along with the best possible pK's and CO2 solubilities; however,
practicality must prevail. It is our contention that a pK' of 6.100 and an "S" of 0.0306
at 37 °C would not be unreasonable [10-12], and that whole blood pHs be used. Slight variations
of pK' with pH would not be unacceptable, but are not essential. The changes in pH with
temperatures of -0.015/°C (0.0148) [30,32], as well as PCO2 (a log PCO2 = F-T at 37 °C; F =
0.019) have been well studied and seem to be reasonable standards [28,34].

The misconception which may arise from pK' studies is that the value determined is the
appropriate one for the given sample. Because it is the result of two or three determina-
tions, this makes it subject to the accumulation of all the errors inherent in each measure-
ment. This is apparent from electrode and tonometer groups. For example, on one occasion
an electrode PCO2 sample gave a value of 6.030, but 6.102 when determined by tonometry.
Another sample gave 6.147 and 6.097, respectively. This suggests to me that there is no
"normal" for a given laboratory. pK' may be affected by various factors as noted, but the
mean arrived at from the many studies is the operational one for practical use. Considering
all the variables, true deviation of the calculated "third factor" is very hard to syste-
matically quantitate, but must be extremely small even using these standard factors outlined
here.

References

[1] Bohr, C. and Bock, J., Bertimmung der absorption Einiger Gase un wasser Beiden tempera-
ture zwischen 0 und 100°, Weid. Ann. Physik. U. Chem. 44, 318 (1891).

149
[2] Severinghaus J. W.
, Stupfel , M,
, and Bradley, A. F., Accuracy of blood pH and PCO2
,

determinations, J. Appl. Physiol. 9_, 189 (1956).

[3] Austin, W. H. and Stinebaugh, B. J., The "not so apparent" pK' , J. Me. Med. Assoc. 60,
~"
10 (1969).

[4] Maas, A. H. J., Eet Koolzvurevenwioht in Natriumbiaarbonaatoplossingen, Liquor Cere-

brospinaliis en Menselijk Bloedplasma (Bronder-Offset, Rotterdam, 1967).

[5] Austin, W. H., Notations in acid-base balance. Am. Heart Jour. 70, 574 (1965).

[6] Bohr, C, Absorptionscoefficienten des blutes und des blut plasmas for gase, Skand.
Arch. Physiol. ]!_, 104 (1905).

[7] Van Slyke, D., Sendroy, J., Hastings, A., and Neill, J., Studies of gas and electrolyte
equilibria in blood, J. Biol. Chem. 78, 765 (1928).

[8] Eisenman, A. J., MacKenzie, L. B., and Peters, J. B., Protein and water of serum and
cells of human blood, with a note on the measurement of red blood cell volume, J.
Biol. Chem. 116^, 33 (1936).

[9] Danowski, T. S. and Gilmore, G. H. Relationships between serum water or weight and
,

protein concentration, j. Lab. Clin. Med. 35, 67 (1950).

[10] Austin, W. H. , Lacombe, E., Rand, P. W. , and Chatterjee, M., Solubility of CO2 in serum
from 15-38 °C., J. Appl. Physiol. 18, 301 (1963).

[11] Fenn, W. and Rahn, H., eds.. The Handbook of Physiology: Respiration, 699 pp. (Williams
and Wilkins Co., Baltimore, Md., 1965).

[12] Nahas, G. G. Ed., Report of the Ad Hoc Committee on Methodology, Current Concepts of
,

Acid-Base Measurment, Annals of the New York Academy of Science, New York City, 274 pp.
(Edward Bros., Inc., Ann Arbor, MI., 1966).

[13] Bartels, H. and Wrbitzky, R. , Determination of carbon dioxide absorption coefficients

between 15 and 18 degrees C in water and plasma, Pflueger Aech Ges Physiol. 2T\_, 162
(1960).

[14] Ludbrook, J., Estimation of PCO2 by means of the Henderson-Hasselbalch equation, in pH

and Blood Gas Measurements, R. F. Woolmer, ed. , p 34 (Little, Brown and Co., Boston,
MA, 1959).

[15] Austin, W. H., Ferrante, V., and Ritchie, R. , Effect of abnormal plasma constituents on
the pK' of whole blood. Am. Jour. Clin. Path. 51_, 799 (1969).

[16] Severinghaus, J. W., Stupfel , M. , and Bradley, A. F., Variations of serum carbonic acid
pK' with pH and temperature, J. Appl. Physiol. 9^, 197 (1956).

[17] Trenchard, D. , Noble, M. , and Guz, A., Serum carbonic acid pK' abnormalities in patients
with acid-base disturbances, J. Clin. Sci. 32, 189 (1967).

[18] Sinclair, M. J., Hart, R. A., Pope, H. M. , and Campbell, E. J. M., The use of the Hen-
derson-Hasselbalch equation in routine medical practice, Clin. Chim. Acta. 1_9, 63 (1968).

[19] Hastings, A. B. and Sendroy, J., Jr., The effect of variations in ionic strength on the
apparent first and second dissociation constants of carbonic acid, J. Biol. Chem. 65,
445 (1925).

[20] Davis, R. P., Logland: A Gibbsian view of acid-base balance. Am. J. Med. 42, 159
(1967).

[21] Huckabee, W. E., Henderson vs. Hasselbalch, Clin. Res. 9, 116 (1957).

150
[22] Austin, W. H. and Littlefield, S., The difference in apparent pH of blood and buffer
caused by raising the liquid junction from room temperature to 37.5 °C, J. Lab. and
Clin. Med. 67, 516 (1966).

[23] Filley, G. F., Acid-Base and Blood Gas Regulation, 213 pp. (Lea and Febiger, Philadel-
phia, PA, 1971).

[24] Siggaard-Anderson, 0., The first dissociation exponent of carbonic acid as a function
of pH, Soand. J. Clin. & Lab. Invest. U, 487 (1962).

[25] Thornton, J. A. and Nunn, J.F. , Accuracy of determination of PCO2 by the indirect
method, Guy's Hospital Report, 109, 203 (1960).

[26] Austin, W. H. Ferrante, V., and Anderson, C, Evaluation of whole blood pK' in the
,

acutely ill patient, J. Lab. Clin. Med. 12^, 129 (1968).

[27] Austin, W. H. and Ferrante, V., The operational value of whole blood pK' to pH, Avoh.
Int. Med. ]2^, 699 (1970).

[28] Nunn, J. F., Bergman, N. A., Bunatyan, A., and Coleman, A. J., Temperature coefficients
for PCO2 and PO2 of blood in vitro, J. Appl. Physiol. 20, 23 (1965).

[29] De Raedt, M. , Vandenbergh, E., and van de Woestijne, K. P., Direct and indirect determi-
nation of partial pressure of CO2 in the arterial blood of patients with respiratory
insufficiency, Clin. Soi. 35, 347 (1968).

[30] Rosenthal, T. B., The effect of temperature on the pH of blood and plasma in vitro, J.
Biol. Chem. 173, 25 (1948).

[31] Greenburg, A. 6. and Moulder, P. V., Temperature coefficients for PCO2 and pH in whole
blood, Aroh. Surg. 91, 867 (1965).

[32] Austin, W. H., Lacombe, E. H., and Rand, P. W., pH-temperature conversion factors and
PCO2 factors for hypothermia, J. Appl. Physiol. 19, 893 (1964).

[33] Kelman, G. R. and Nunn, 0. F., Nomograms for correction of blood PO2, PCO2 and base
excess for time and temperature, J. Appl. Physiol. 2]_, 1484 (1966).

[34] Bradley, A. F., Stupfel , M., and Severinghaus , J. W., Effect of temperature on PCO2 and
PO2 of blood in vitro, J. Appl. Physiol. 9, 201 (1956).

[35] Severinghaus, J. W., Blood gas calculator, J. Appl. Physiol. 2]_, 3 (1966).

[36] Olszowka, A., Rahn, H., and Farki , L., eds.. Blood Gases: Hemoglobin, Base Excess and
Maldistribution, 170 pp. (Lea and Febiger, Philadelphia, PA, 1973).

[37] Stinebaugh, B. and Austin, W. H., Acid-base balance, a common sense approach. Arch.
Int. Med. 219, 182 (1967).

[38] Fleischer, W. R. and Gambino, R. S., eds.. Blood pH, PO^ and Oxygen Saturation, 223 pp.
(Am. Soc. Clin. Path., Chicago, 1972).

151
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

DEFINITION OF OXYGEN SATURATION AND

CHARACTERIZATION OF OXYGEN-HEMOGLOBIN AFFINITY

Arthur L. Malenfant
Instrumentation Laboratory Inc.
113 Hartwell Avenue
Lexington, Massachusetts 02173, USA

1. Introduction

The function of hemoglobin is to transport oxygen from the lungs to the tissues. In

order to carry out this function optimally, the hemoglobin must take up a maximal amount of
oxygen during circulation through the lungs, and then yield this oxygen to the tissues perfused
during circulation through the body. The ability of the blood to carry out both these
functions is determined by the affinity of the hemoglobin for oxygen. Specific definition
of oxygen saturation is needed.

The characterization of the relationship between the P^, and percent oxyhemoglobin
U2
saturation--the oxyhemoglobin dissociation curve--has been found to be increasingly complex.
It is now known that there are a number of factors which affect this relationship, resulting
in shifts in the oxyhemoglobin dissociation curve. The significance of these shifts centers
around the fact that the amount of oxygen available to tissues at a given Pq^ is changed.

Factors known to affect the oxyhemoglobin dissociation curve are pH, temperature, P^q
and 2,3-DPG (fig. 1). Increase in pH or decrease in the value of any of the other factors
shifts the dissociation curve to the left [1,2]^. When all these have a "normal value" the
oxyhemoglobin dissociation curve has a P50 (Pq for 50 percent saturation of the hemoglobin
with oxygen ) of 26.7 mm. In addition to the normally observed factors which influence the
oxygen-hemoglobin affinity, carbon monoxide, present because of environmental pollution and
in cigarette smoke, also effectively shifts the dissociation curve to the left [1,7].

Both acute shifts in P50 [3] and what can be considered to be adaptive shifts in P50
[4] have been observed.

2. Measurement Techniques

The techniques employed in characterizing oxygen saturation--and more importantly, the

hemoglobin-oxygen affinity--of blood must necessarily include measurement of a number of
parameters. Direct photometric measurement of oxygen saturation of blood is straightforward

Figures in brackets indicate the literature references at the end of this paper.

153
 Figure 1. The normal dissociation curve for whole blood
at pH 7.4 — , dissociation curves for hemoglobin
solution at pH 7.22 under varying conditions of P^q
and 2,3-DPG/Hb ratios.

as long as factors such as carboxyhemoglobin and methemoglobin are taken into consideration

[5,6]. The typical concentrations of methemoglobin, however, are low enough to not signi-
ficantly affect accuracy. Failing to characterize the presence of these species can lead to
erroneously high values for saturation, however.

Other methods include direct manometric measurement of oxygen content of a particular

sample compared with the oxygen content of a sample saturated with oxygen by appropriate
means [8]. Oxygen content can also be determined using a fuel cell to consume the oxygen
liberated from blood samples, otherwise treated in essentially the same manner as the
manometric technique above. Comparing oxygen content with oxygen capacity yields oxygen
saturation defined as a percentage of "available hemoglobin"--hemoglobin available for
combination with oxygen [9]. Comparing oxygen content with that oxygen capacity theoreti-
cally available through measurement of total hemoglobin concentration yields an oxygen
saturation defined as a percentage .of total theoretical oxygen capacity [6]. The molecular

154
weight of hemoglobin has been accepted as 64,458 [10]. Each molecule of hemoglobin combines

with four molecules of oxygen. Therefore, in traditional terminology, 1.39 ml of oxygen at

STP will combine with 1 gram of hemoglobin. Hemoglobin concentration expressed in grams
(g/100 ml) multiplied by 1.39 ml of oxygen yields the theoretical oxygen capacity of the
blood sample. The relationship in more acceptable terms is 4 moles of oxygen per mole of
hemoglobin. Hemoglobin concentration of 2.327 millimoles (15g%) theoretically binds 9.308
millimoles of oxygen (20.85 ml) when complete saturation is achieved.

3. P50 Nomogram

Use of measurements typically employed in blood gas analysis, i.e., pH, Pq^ and P^q^
combined with the measurement of 2,3-DPG will also yield measurement of oxygen saturation
which is accurate [11].

For normal hemoglobins the influence of pH on the oxyhemoglobin dissociation curve has
been well characterized and has been reconfirmed; the effect of P^q at fixed pH has been

established. The characterization of oxygen hemoglobin affinity when changing the ratio
between 2,3-DPG and hemoglobin concentrations has been completed as well (fig. 2). Avail-
ability of all of these relationships has enabled construction of an empirical nomogram
which predicts oxygen saturation which results at 37 °C. This nomogram relates the logarithm
of the Pp. required to give 50 percent saturation to changes in pH, Pp^, , and 2,3-DPG/Hb
U2 LU2
ratio (fig. 3). The nomogram can be used to calculate any of the parameters when the other
three are known, and could be used as a means of indirectly measuring 2,3-DPG if that were
of interest.

Figure 2. Values of log P50 of human blood vevsus 2,3 DPG/Hb ratio at zero

Ppp, and Ppp, = 41 mmHg.

155
 I I J \ I I I I U-
7.0 7.1 7.2 7.3 7.4 7.5 7.6 7.7
pH

Figure 3. Values of log P50 as

a function of pH and P^^ a) normal :

human blood at 37 °C; b) human blood with no 2,3 -DPG;^c) human

blood with 2,3-DPG/Hb ratio = 0.75; d) human blood at P^q = 0 and
2,3-DPG/Hb ratio = 2.3 and Pp^ = 41 mmHg and 2,3-DPG/Hb ratio = 2.5.

The work which led to the data which supports this nomogram was carried out by a group
comprised of researchers from Instrumentation Laboratory, Inc., the University of Milan and
University of California at Santa Barbara. The aim of the work was to extend studies which
have been performed on the effects of these hemoglobin effecters from diluted and purified
hemoglobin solutions to whole human blood under near physiological conditions of temperature,
Ppp, , pH, etc.

The oxygen dissociation curve (ODC) for a concentrated (ea. 25 percent) sample of
human hemoglobin in 0.1 M KCl serves as a basis for determining the effect of the individual
species. These ODC's were obtained using the "open" tonometric technique at a constant pH
of 7.22. This pH was chosen since it has been demonstrated [15] that an extracellular pH of
7.4 closely corresponds to an intracellular pH of 7.22. Thus, at the same intracellular
pH only two cof actors, CO2 and 2,3-DPG are required to make the ODC of a hemoglobin solution

correspond to the ODC of whole blood.

156
 Table 1. Values of log P50 as a function of pH and
Pp-, interpolated through the data.

2,3-DPG P^« pH: 7.0 7.2 7.4 7.6

CO2
Hb

0 1 .260 1 .148 1 .038 0,.926

0 20.874 1,.285 1.820 1,.154 1,.088

41.535 1,.328 1.274 1 .220 1,.166

57.794 1 .342 1.305 1 .272 1,.232

0.75 0 1,.528 1.432 1 .336 1,.240

41.535 1,, .545 1.468 1 .396 1,.316

0 1,.575 1 .484 1 .394 1,.306

20.874 1,.584 1.496 1 .412 1,.330

1

41.535 1 .592 1.512 1 .436 1,.356

57.794 1,.604 1 .530 1 .456 1,.380

2.3 0 1 .716 1 .628 1 .540 1,.452

2.5 41.535 1 .756 1.672 1 .592 1 .512

The position of the ODC along the P^, axis will be described in terms of its P50 (Pn
U2 U2
at which the hemoglobin is 50 percent saturated with oxygen). Plotting log P50 versus
pH yields straight-line relationships which are a function of P^g and 2,3-DPG. Table 1

indicates the values for log P50 interpolated from the data generated at varying levels of
P^Q^ and 2,3-DPG.

These data can be either represented in a nomogram or mathematically derived from an

empirical formula. Both the nomogram and the empirical function allow calculation of pH,
Pqq » log P50 or 2,3-DPG/Hb ratio when the other three are known. The final monogram
(fig. 4) has been completed by adding a scale for the P^, and a corresponding scale for S^, ,
U2 U2
assuming that the relation between Pq and Sq conforms to the classical Hill's equation [12].

- = log K + n log Pp
100
U2

At 50 percent saturation log K = -n log P50. The value of n was assumed to be 2.7. Clearly
this equation can only be assumed to give a rough estimate of the relation between S^. and P^
U2 U2
Other workers have also found that n varies for normal human blood when CO2 and 2,3-DPG are
varied [13]. With these limitations in mind, it is possible to obtain the function Sq^ vs. Pq^

derived from Hill's equation at a definite value of log K = (-n log P50).

157
 7
1

46-

42 -
40 -1.6

38 -
36- 1 0 -

3 4-
SO2 (•/.)

3 2- -1.5 -- 1 0

30

2 8--
3 0 20-
2 6- 4 0

.-1.4 5 0 --7. 8
2 4- 6 0
® REF Pc02 3 0-
7 0
2-' -7.
2
--80
40 -
20--1.3 9 0 -7.6
50-
9 5
18-. 6 0' 7.5

70-
1 6-_-1.2 80- -7.4
•
9 0

100-
-7.3
12 0-
1 40- -7.2

1 2 - 16 0-
180- -7.
20 0-
REF 2.3 PPG ^
1 0 - -10 Hb -7.0
P50 P02
log Pso PH
(mmHg) ( mmHg)

Figure 4. The oxygen dissociation curve and P50 nomogram.

Use of the P50 nomogram is relatively straightforward (fig. 5). If it is required to

find log P50 for values of pH = 7.4, P^q = 40 mm, G = 1.0 the steps are as follow: (1)-

Line 1, connecting P^q with Ref. P^q is traced; 2) G = 1.0 and Ref. 2,3-DPG/Hb are connected
by line 2; (3) A straight line connecting pH = 7.4 with the log P50 scale is traced through
point of intersection of lines 1 and 2. The value of P50 = 26.7 mm is read on the appropri-
ate scale. As shown in figure 6, the value of G may be established by knowing P50, P^q ,

and Pn .
U2

Figure 7 indicates the means by which an GDC may be- constructed by knowing P50, P^q ,

and Pf. .
U2

By inserting values of G, pH and P^q into the mathematical expression which is

expressed below, values of log P50 can be directly calculated.

158
 4 6-
1,1.-

42 -
40 -16
3 8-
3 6-

3 4

3 2- -1.5
30 -

2 8--
2 0-
2 6-
-1.4 7. 8
2 4-
30-
2- -7.7
2

40 -
20 •1.3 7.6
5 0-
8 -
1
60 •7.5

-v^O-l
6--1.2 \
1
8 0 •7.4.

90-
100- 7.3
12 0-
1 40 7.2
1 2 160-
180- -7.
200

1 0 - -1.0 Hb -7.0
Pso log Pso P02
PH
(mmHg) ( mmHg)

Figure 5. Use of nomogram to determine P50 from measurement

of 2,3-DPG/Hb ratio, P^q and pH (see text).

log P50 = 1.9729 x 10-3G2pH P^q - 0.6002 x 10-3G pH P^^q^ + 3.9274 x 10-3

pH P^Q^ + 206.5 X 10-3G2pH - 247.79 x 10-3G pH - 410.31 x 10-3

pH + 15.85 X 10-3G2 P
LU2
+ 1.157 x 10-3G P„ - 25.948 x 10-3
1,1)2

P(>Q^ - 1.6337 G2 -h 2.2427 G + 4.125.

Evaluation of errors which will occur in values for P50 at pH = 7.0, 7.2, 7.4 and
7.6; Pj,Q = 0, 20, 41 and 57; and G = 0, 0.75, 1, 2.3 and 2.5 has been carried out by use

159
 Figure 6. Use of nomogram to determine 2,3-DPG/Hb ratio from
measurement of P50, Pr-n. and pH (see text).

of the mathematical expression. Table 2 gives the results obtained for log P50 using the
mathematical expression and can be compared with the results interpolated from the empirical
data. In the range of values which encompasses most physiological and pathological condi-
tions, the maximum discrepancy of log P50 is 0.037 at a value of log P50 = 1.632 which is
about 2 percent. In the range of G from 0.1 to 1.0 the maximum deviation is 0.011 for log
P50 = 1.328. Overall, it is likely that the error in the determination of log P50 or G in
the range of the variables indicated in the nomogram does not exceed 3 percent of the
desired quantity.

160
 4 6-
4* -
42 -
40 1.6
3 8-
36- 1 0 -

3 4-
SO2
3 2- -1.5
30 -
-20
2 8- 0--
-- 3 0 2

26 - 4 0

5 0

2 4- 60
® REF PCO2 30-
7 0

2 2-
+ 80
40 •

20--1.3 9 0
5 0-
9 5
1 8 -

70-
' «--1.2 80-
90-
10 0-

12 0-

1 40-
1 2 - 1 60-
180-
200-

1 0 - -10 -7.0
Pso log Pso PH
(mmHg) ( mmHg)

Figure 7. Use of nomogram to construct oxygen dissociation curve

from measurement of P50 (see text).

This work is by no means final, and use of the nomograms can safely be applied only to
blood with normal hemoglobin concentration in the absence of pathological hemoglobins. We

have no data available at this time, for instance, on blood from individuals who are anemic.
In these individuals G is altered (mainly by decrease in hemoglobin concentration) from its
normal value [14].

The purpose here is to demonstrate the requirements for an accurate assessment of

oxygen saturation of blood samples. The several species which will affect the relative
hemoglobin-oxygen affinity must be accounted for in order to accurately make an assessment.

Direct photometric measurement of oxygen saturation, combined with measurement of

blood gas parameters pH, P^q , and Pq yields the required information for correct assess-
ment of hemoglobin-oxygen affinity.

161
 b ,
.

Table 2. Values of the function log P50 = f(pH, Pp^, , G)

derived from equation. 2

L )
0— Ur p nH 7 n 7 ? 7 &. 7
^C02
Hb

0 1 ^ 71
.1/1
1 1
1 1 . ucjy 1 . UU/
0 20.87 1,.285 1 .219 1 .154 1 .088
41.53 1,.317 1 .267 1 .218 1.169
57.79 1,.;342 1 .305 1.269 1.232

0.75 0 1,.528 1 .432 1 .336 1 .240

41.535 1,,545 1 .469 1.392 1 .316

0 ] 573 1 1 . o^C
20.87 1 ,584 1 .499 1 .415 1 .330
1

41 .53 1,,595 1 .516 1 .437 1.358

57 7Q 1
1 1 1
1 /ICC

2.3 0 1.,426 1 .448 1 .471 1.493

2.5 41.53 1.,625 1 .595 1.565 1 .535

References

[I] Shappel, S. D., Lenfant, C. J. M., Anesthesiology, 37, 127 (1972).

[2] Fallon, K. D. et at.. Analyzer, 4, 11 (1974).

[3] Valeri, C. R., Collins, F. B., X){l\lth Scientific Meeting of the Blood Research Institute,
p. 397.

[4] Metcalfe, J. et al. , Circulation Research, 25, 47 (1969).

[5] Siggaard-Andersen, 0., et al. y Clin. Chim. Acta, A2, 85 (1972).

[6] Malenfant, A. L., et al., 20th National Meeting, AACC (August 19, 1968)

[7] Astrup, P., Forsvarsmedicin, S_, 199 (1969).

[8] Van Slyke, D. D., Neil, J. M., J. Biol. Chem. 61, 523 (1924).

[9] Clauvpl, M., Schwartz, K., Clin. Chem. ]±, 3, 253 (1968).

[10] Eilers, R. M., Amer. J. Clin. Path. 47, 212 (1967).

[II] Muletti, A. et al.^ Physiological Basis of Anaesthesiology , M. M. Mushin, et al., ed.

p. 95 (Piccin Medical Books, Padua, 1975).

[12] Hill, A. v., Proc. Physiol. Soc. J3., 4 (1910).

[13] Garby, L., et al. , Acta Physiol. Scand. 84, 482 (1972).

[14] Astrup, P., Rorth, M., J. Clin. Lab. Invest. 31, 311 (1973).

[15] Agostoni, A. et al. , Science 182 . 300 (1973).

162
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

PROBLEMS ASSOCIATED WITH THE DEFINITION OF MEASURED AND CALCULATED

QUANTITIES IN BLOOD pH AND GAS ANALYSIS

Robert W. Burnett
Clinical Chemistry Laboratory
Hartford Hospital
Hartford, Connecticut 06115, USA

For most students, the subject of acid-base balance has the reputation of being a
singularly complex and confusing one. This seems to me to be due in large part to the
multitude of methods which have been used and quantities which have been defined to express
the acid-base status of an individual. Much confusion has been caused by the existence of
several methods, that might all be in common use, and that measure similar, but not exactly
the same, quantities. Another obstacle to understanding is the existence of several terms
to denote a single quantity, or group of quantities that differ from each other only in a
trivial way. Moreover, many modern and widely used textbooks contribute to the problem by
devoting a disproportionate amount of space to a discussion of methods and quantities that
are generally agreed to be outmoded, and giving inadequate space to a discussion of legit-
imate differences of opinion or differences in usage which exist in this field today.

My objective in this presentation is to outline some of the problems that exist in the
area of definition of quantities used in blood pH and gas analysis. I think it will be
clear that many non-trivial questions exist that will not, of course, be answered at this
meeting, but that do deserve our careful consideration.

The quantities that perhaps cause the fewest problems, in terms of their meaning,
happen to be the three most commonly measured, namely pH, P^q and Pq This is not to
.

imply that no questions exist in the definition of these. The pH of an unknown solution is
defined and determined in an operational way by comparison with the pH of an accepted
standard solution. Systematic biases in blood pH measurements between laboratories often
occur however, even when both are using the same pH standard. These can be the result from
differences in the type of liquid junction, differences in bridge solution composition,
different temperatures of the salt bridge, or junction potential differences due to the
so-called "cell effect" seen with whole blood. While it would be worthwhile to work
toward eliminating such inter-laboratory variations, it must also be remembered that ac-
curacy as a concept in pH measurements with whole blood is much more difficult to deal with
than with most other measurements. Quoting from Bates' monograph [1]^,

"...the experimental pH can never be an exact measure of either

the concentration or the activity of hydrogen ion It is
safe to say that no quantitative interpretation of measured
pH values should be attempted unless the medium can be clas-
sified as a dilute aqueous solution of simple solutes."

The need, therefore, is to establish a convention for the measurement of pH in blood, so

that consistent and reproducible numbers can be obtained in different laboratories and with
different instruments.

With regard to the gas tensions, P^q and Pq , definition is not a problem, but again.

^Figures in brackets indicate the literature references at the end of this paper.

163
consistent measurement of these quantities is often a problem. Differences in electrode
response time at different gas tensions and differences in measured gas tensions between
gas and liquid standards are two common reasons for inconsistent and inaccurate measurements.
These problems are actually ones of standardization, and perhaps of instrument specifications,
and should be discussed later in this symposium.

Another measurement which is routine in most clinical chemistry laboratories is CO2

content, but this term has been used in connection with many different methods, not all of
which give equivalent results. The determination has been made by measuring the volume of
CO2 released from acidified serum or plasma or measuring the color change in a weakly
buffered indicator solution on reaction with an aliquot of this CO2, for which different
indicators and buffers have been used. The measurement can be made with or without the
addition of base as a preservative, with or without equilibration of the sample to a known
CO2 tension, and after centrifugation of the blood at either body temperature or room
temperature. Clearly, there is a need to adopt a single definition for this quantity, and
then to determine the extent of systematic bias that is present in the several methods that
are available.

Some other quantities that are often measured directly, relate to the oxygen transport
status of an individual. These too can have more than one meaning, depending on the par-
ticular method of measurement. Hemoglobin may be measured as oxyhemoglobin alone, or in
combination with reduced hemoglobin, carboxyhemoglobin or methemoglobin. Likewise, oxygen
saturation, which is actually a derived quantity, can be calculated with or without terms to
include the effect of carboxyhemoglobin or methemoglobin. Spectrophotometric methods with
measurements at one, two, three or four wavelengths have all been used, and will give
different results with the same blood sample if carboxyhemoglobin or methemoglobin is
present. To add to the confusion, there is also the very common practice of calculating
oxygen saturation from a measured pH and P« The equations or nomograms used for this
.
U2
purpose are constructed from a standardized oxyhemoglobin dissociation curve--not necessarily
the same one--which is only valid under a single set of conditions, including CO2 tension,
2,3-diphosphoglycerate concentration and concentrations of carboxyhemoglobin and methemoglobin.
A systematic bias is always present when all these conditions are not met by the specimen of
blood being analyzed. Moreover, it is quite common to find that two of these nomograms give
significantly different answers, particularly at high oxygen saturations.

Increased interest in the actual oxyhemoglobin dissociation curve which is obtained

in a given sample of blood has led to the more common calculation of P50, the oxygen tension
which corresponds to 50 percent saturation. Here too, careful definition and consistent
conventions are needed in order to avoid inter-laboratory differences. First, the meaning
of oxygen saturation must be established, and second, agreement must be reached on the
values for constants to be used in the Hill equation, Adair equation, or other relationship
used to calculate P50.

A final quantity of interest related to oxygen transport is oxygen content. This may
be measured directly by either the manometric Van Slyke method or by a technique in which
all the oxygen in the sample is reduced electrochemically . Alternatively, oxygen content
can be calculated as (K x total hemoglobin x oxygen saturation). This calculation will be
meaningful only if two conditions are met. First, total hemoglobin and oxygen saturation
must be expressed in units that are consistent with each other, and second, a value for K
must be agreed upon. The theoretical value for K can be calculated from the molecular
weight of hemoglobin A and the molar volume of oxygen at standard temperature and pressure,
and is equal to 1.39 ml O2 per gram of hemoglobin. This value is not always considered the
best to use however, due to reports of an unexplained systematic bias between calculated and
measured oxygen contents. Again, it is clear that a convention needs to be adopted in order
to make it possible to obtain values that are meaningful and consistent among laboratories.

Returning to the definition of acid-base status, I have described some problems having
to do with measured quantities and I would like to discuss now some factors related to
derived quantities. These are sources of confusion and inconsistency, again because a given
quantity is often calculated using different definitions or different values for constants.
First, consider the quantities that may be calculated using the Henderson-Hassel balch
equation. If we assume that accurate values of pH and P^q are available, the Henderson-
Hasselbalch equation allows calculation of plasma bicarbonate and CO2 content. The condition

164
that must be met in order to insure consistent results among laboratories is, of course,
that a single value for the dissociation constant of carbonic acid be adopted and a single
number for the solubility coefficient of CO2 in plasma be used. Careful experimental work
has resulted in apparently accurate values for these constants, so this should not present a
serious problem.

A more difficult quantity to reach a consensus on in the past has been base excess.
This reflects the fact that base excess can be defined in relation to various fluid com-
partments in the body, including plasma, whole blood, interstitial fluid, or some combination
of these. The parameter that is necessary to define base excess mathematically is the
buffer capacity of the fluid. Although an early definition of base excess was based on the
buffer capacity, of whole blood, the most widely accepted definition at present is based on
the buffer capacity of extracellular fluid (i.e., whole blood together with interstitial
fluid) which has been determined in vivo by several groups. Unfortunately, although most
workers who are active in the field of blood pH and gas analysis seem to agree on the
advantages of the so-called extracellular fluid base excess, several widely used textbooks,
published within the last five years, do not mention this quantity at all, but persist in
using the older definition. Clearly, when there is indeed consensus among experts, then the
conventions that have been agreed upon must be effectively publicized if they are to influence
current teaching.

Finally, reference should be made to the many other terms that have been used in the
past and in some cases continue to be used to describe various aspects of the acid-base
status of an individual. This list includes alkali reserve, bicarbonate reserve, buffer
base, buffer anions, CO2 combining power, acid-base ratio, standard bicarbonate and CO2
capacity. Some of these terms are redundant; some are simply no longer useful. All of
these terms may be found in recently published textbooks. The need for derived quantities
can perhaps be placed in perspective by keeping in mind that acid-base status is completely
determined if only two quantities are known, for example, pH and P^q The justification for
.

continuing to use other quantities must rest on a real practical advantage in the understanding
or treatment of acid-base abnormalities.

In summary, present usage of both measured and derived quantities for blood pH and gas
analysis is often inconsistent and confusing. This area of clinical chemistry would benefit
greatly from international agreement on the mathematical definition of quantities and on
values for appropriate constants. It would then also be necessary for the areas of agreement
to be widely publicized. One change which I feel would be a significant improvement would
be for scientific journals to adopt uniform criteria for nomenclature and definition of
terms in this area. This has been done successfully by other journals in the areas of
spectrophotometry and statistics.

The most important reasons for precise definition of terms and agreement on conventions
are that they will make possible measurements whose results are reproducible among lab-
oratories, it will greatly facilitate the interpretation and comparison of data generated in
different laboratories, and it will make the job of those who apply the results of these
measurements to patient care less complex. The desirability of achieving these goals is
something that can certainly be immediately agreed upon.

Reference

[1] Bates, R. G. , Determination of pH, 2nd edition, p. 103 (John Wiley & Sons, New York,
1973).

165
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

ARTERIAL AND VENOUS BLOOD SAMPLES IN ACID-BASE BALANCE

William H. Austin
Department of Research
Maine Medical Center
Portland, Maine 04122 USA

The field of pH and blood gas measurement has stood as the stepchild of anesthesiology,
pulmonary medicine, nephrology, endocrinology, and research itself. It has never gained the
status of a full-fledged discipline. Because of these multi-faceted relationships, different
concepts concerning sampling, measuring, calibrating, reporting, and interpreting have
arisen.

A simple example of this is seen in the use of PCO2 by the respiratory physiologists
and PCO2 by others. In this simple example, the symbol "P" is used to mean pressure, whereas
it is also used in the designation of a constant, such as pK' The respiratory physiologists
.

have achieved some measure of uniformity by agreeing upon symbols and nomenclature for
international use. However, "a", the solubility factor (ml C02/ml plasma) and "S", the
coefficient (mm PC02/ml plasma) are still used interchangeably.

This could be logically extended to the entire field of acid-base balance, as it is

used in the various disciplines mentioned above. The logical way to achieve consistency and
uniformity is to make pH and acid-base balance a specialty in itself. This would be done
with the understanding that this area would not create isolation.

In doing this, there must be general agreement in the many areas of this subject. This
workshop, I believe, has addressed itself primarily to this point. Many of these areas are
being discussed here, and no single paper can encompass the entire field. One area in which
there has been debate is that of the site of the sample taken to be measured [1-4]^. It is
obvious that arterial blood is essential when dealing in matters of pulmonary physiology;
however, there has been no universal agreement regarding the role of venous blood in the
other physiologic areas. Gambino [5-7] has been an outspoken advocate of venous blood, and
has used venous blood where pH and PCO2 have been of primary concern. For others, the
addition of PO2 is essential, and dictates the origin of the sample. I am not aware of any
clear definition of the role of venous blood measurements, despite consideration of mixed
venous (brachial and hand) blood. Much has been written about "arterial izing" venous blood,
and using various techniques to ensure the uniformity of venous samples. There is evidence
of the similarity of venous and arterial blood when drawn from the brachial vessels under
controlled conditions; however, we have encountered some dissimilarities despite rigid
sampling and measuring techniques. At times, there may be very little difference, but it
has been difficult to determine the conditions which produce A-V dissimilarities.

Under controlled conditions, the PCO2 of arterial blood will change dramatically, rela-
tive to venous PCO2, with alteration in ventilation. Figure 1 indicates this point. In
the studies illustrated here, blood samples were taken from the inferior vena cava and distal
aorta in each of eight dogs, before and following a twofold increase in minute ventilation
[8], The dogs were intubated, and had the appropriate vessels cannulated, with a one-hour
steady state during which the PCO2 measured by electrode was kept within ±3-4 mm. The
figures show changes which occurred at various intervals in both the arterial and venous
systems. The rapidity of the fall of PCO2 in arterial blood, and the lag in decrement for
venous blood illustrates just one facet of this problem. The lowering of the arterial
partial pressure of carbon dioxide is consistent with the change in ventilation, and the
venous pressures reflect changing carbon dioxide levels at the tissue level.

1
Figures in brackets indicate literature references at the end of this paper.

167
 Figure 1. The average and mean changes
in venous {•) and arterial (l) PCO2
resulting from a two-fold increase in
minute ventilation. (Reproduced with
permission of The Journal of the Maine
Medical Association.
0 2 4 6 8 10 12
MINUTES

Each value is appropriate for the situation in the unsteady state, and there is no
"right" or "wrong" value. There is, then, the logical extension of this simple maneuver to
the patient. There is no axiom which states that venous and arterial blood should have a
fixed difference in either sample, or will represent one physiologic set of circumstances.
The misconception occurs when one tries to correct arterial for venous, or venous for arte-
rial, blood sample values. The major issue is, "which sample of blood reflects the clinical
state, and does it provide the information necessary for the appropriate treatment of the
patient?"

Table 1 illustrates the A-V differences which were measured in fifty consecutive pa-
tients with various disease states. Numbers 32,35,36,37,39,42, and 46 suggest abnormalities
if the venous blood is observed. This is not the case. No statistical analysis can be made
here, but neither the nature of the illness nor its severity appears to correlate with the
differences observed. That there are differences is no surprise to some, but may be chal-
lenged by others. Yet the facts must speak for themselves. Although there is no systematic,
objective proof of the matter, the arterial values appear most consistent with the patient's
clinical state. In those instances where one would not anticipate abnormalities, venous
blood measurements suggest that there are. All this is done using the criteria set forth
for drawing and measuring pH and blood gases [6]. The reasons for these differences are not
apparent, and no obvious technical errors were made in the determinations. The conditions
for sampling and measuring have been made elsewhere, and strict adherence to these has been
used. Reliability of calculated variables has been stated [1-5]. Transient changes in
venous blood flow may account for some of these differences; however, it was impossible to
predict when the wide A-V differences would occur. It could be argued that the venous
measurements in reality reflect the tissue state, which is perhaps desirable, but this may
only be representative of the area from which the sample is taken. Here a mixed venous
blood sample would be the logical site, but not practical to obtain.

I have been an advocate of the use of venous blood, but over the years I must concede
that I find myself relying more and more on arterial values, even in disease states which
are not primarily of respiratory origin. Now technicians can easily obtain samples, which
obviates a previous objection, and frequent punctures have not been a problem. The circum-
stances are such that the calculated CO2 content, or bicarbonate, made when the pH and PCO2
are measured, is generally quite similar, or differs by a consistently small amount. This
means to me that the Henderson-Hasselbalch equation is giving correct CO2 contents, or bicar-
bonates, but that the pH and PCO2 values may be misleading. On other occasions, abnormal
pH's and PC02's were found on venous blood, when no disorder existed after subsequent arterial
samples were taken. I have become more and more convinced of the reliability of arterial pH
and PCO2 as representative of the severity of the disease state, even though venous CO2 con-
tents are usually consistent. This appears to apply more to PCO2 measurements which we have
found to be extremely variable in venous blood.

168
 Table 1. A-V blood gas differences in fifty patients.

pH PCO2 TCO2

A V A A V A A V A

c
1
1 • 7 46 7 43 03 41 . 46.6 0 . 1 29. 6 31 0 1 .4
9 r r
c • 1 41
/ • *T 1
7 37 n4 43.5 49.0 0. b 27. 7 28. 5 0.8
0• 7 4fi 7 43 •uo 40.4 44. 3. 9 28. 8 30. 0 1 .2
4. 7.41 7.38 .03 42.0 45!o 3. 0 26! 7 26! 8 o!i
5. 7.44 7.41 .03 36. 41 .5 4. 9 25. 0 26. 4 1 .4
6. 7.46 7.43 .03 37.7 41.5 3. 8 26. 9 27. 6 0.7
7. l.Zl 7.35 .02 53.4 58.0 4. 6 31. 1 32. 1 1.0
8. 7.46 7.43 .03 29.8 31.3 2. 5 21 3 20. 8 0.5
9. 7.52 7.46 .06 38.5 42.5 4. 0 31. 6 30. 2 1.4
10. 7.43 7.39 .06 42.5 48.5 6. 0 28. 2 29. 5 1.3

1 1 • 7 dl 7 35 06 36 7 41 n 5. 3 23 3 23. 2 0.1
9
1\
c. 7 /in 7 3Q m *+VJ . u 47 n 6 4 28 5 3.3
0•
1
7 4^ 7 3Q 04 41 4 4ft n 6. 6 77 5 29. 2 1 .7
14. 7.41 7.37 .04 49.2 56.0 6. 8 31. 3 32! 6 r.3
15. 7.44 7.42 .02 42.0 46. 5 4. 5 28. 5 30. 3 1 .8
16. 7.42 7.39 .03 48.0 51.0 3. 0 31. 2 31. 0 0.2
17. 7.49 7.44 .05 46.0 52.6 6. 6 35. 2 35. 4 0.2
18. 7.40 7.40 .00 49.0 50.0 1. 0 31. 0 31 1 0.1
19. 7.38 7.35 .03 74.0 79.0 5. 0 44. 2 44. 2 0.0
20. 7.41 7.36 .05 40.5 52.5 12. 0 25, 8 30. 0 4.2

21 7 4"? 7 38 05 47. 5 57. 10.0 29 2 34 3 5.1

22. 7 ^R 7 34 04 41 . 50.0 8. 5 24 7 27 1 2.4
23 7 43 7 39 04 46. 5 56.0 9. 5 28 6 34 1 5.5
24. 7.41 7.35 .06 37^4 46!4 9. 0 23 8 25 7 1.9
25. 7.49 7.45 .04 30.5 39.0 8. 5 23 4 27 1 3.7
26. 7.42 7.37 .05 47.0 57.0 10.0 30 6 33 1 2.5
27. 7.40 7.37 .03 59.0 70.0 11. 0 36 8 40 8 4.0
28. 7.48 7.36 .12 35.5 48.6 13. 1 26 2 27 8 1.6
29. 7.37 7.32 .05 60.0 74.0 14. 0 35 1 38 3 3.2
30. 7.42 7.36 .06 36.5 54.0 17. 5 23 8 31 9 8.1

31 7. 53 7.40 . 1 36.0 52.0 16. 0 30.2 32. 5 2.3

32. 7.40 7.33 07 43. 62. 19 0 27. 4 32. 8 5.4
33 7.45 7. 38 •07 33.5 47.0 13 5 23. 4 28. 1 4.7
34. 7.46 7^42 !o4 36! 3 48!4 12 1 25. 9 31! 5 5.6
35. 7.43 7.29 .14 36.8 60.0 23 2 34. 2 29 3 4.9
36. 7.53 7.45 .08 34.0 51.0 17 0 28 6 35 6 7.0
37. 7.39 7.34 .05 43.7 61.0 17 3 26 6 34 5 7.9
38. 7.17 7.14 .03 77.8 102.0 24 2 30 0 36 0 2.7
39. 7.37 7.33 .05 42.0 57.0 15 0 27 8 32 0 4.2
40. 7.25 7.17 .08 51.5 74.0 22 5 23 5 28.0 4.5

41 7. 51 7 45 06 48 8 60 0 11 2 39 8 44 0 4.2
42. 7.40 7 34 •06
\J\J 39 4 57 0 17 .6 25 4 32 0 7.5
43 7 43 7 37 06 36 6 59 0 22 4 25 4 35 0 9 6
44. 7.49 7.40 .09 36.4 58.0 21 .6 28 8 36 0 7.2
45. 7.43 7.38 .05 50.5 64.0 13 .5 34 5 39 0 3.5
46. 7.41 7.34 .07 37.1 53.9 16 .8 24 0 30 6 6.6
47. 7.38 7.25 .13 23.3 35.7 12 .4 14 2 16 0 1 .8
48. 7.29 7.29 .00 87.0 98.0 11 .0 41 3 48 0 6.7
49. 7.34 7.31 .03 29.1 41.1 12 .0 16 2 22 0 5.8
50. 7.37 7.24 .13 28.5 51.4 22 .9 17 0 23 3 6.0

169
 On the basis of the small amount of data presented here, clinical experience, and
information derived from the literature, I would generally favor arterial blood as a reference
point for all acid-base disturbances, even those of metabolic origin. This is not to exclude
the use of serial venous pH and PCO2 measurements made with the reservations mentioned
above. I believe that the door should be kept open on this point, and other experienced
opinions obtained. Again, this may be a matter of judgement--reinforcing my original
concept of the field of pH and blood gas measurement.

REFERENCES

[1] Forster, H. V., Dempsey, J. A., Thomson, J., Virduk, E., and doPico, 6. A., Estimation
of arterial PO2, PCO2, pH and lactate from arterialized venous blood, J. Appl. Physiol.
32, 134 (1972).

[2] Samet, P., Linhart, J., Barold, S. and Hildner, F. , Reliability of mixed venous blood
,

for the measurement of blood gas parameters, J. Thor. Card. Surg. 58, 131 (1969).

[3] Sutton, R. N., Wilson, R. F., and Walt, A. J., Differences in acid-base levels and
oxygen-saturation between central venous and arterial blood. Lancet, 2^, 748 (1967).

[4] Paine, E. G. , Boutwell, J. H., and Soloff, L. A., The reliability of "arterialized" venous
blood for measuring arterial pH and PCO2, Am. J. Med. Soi. 2M, 431 (1961).

[5] Fleischer, W. R. and Gambino, R. S., eds., Blood pH, PO2 and Oxygen Saturation, 223 pp.
(Am. Soc. Clin. Path., Chicago, 1972).

[6] Gambino, S. R. , Comparisons of pH in human arterial, venous, and capillary blood.

Am. J. Clin. Path. 12, 298 (1959).

[7] Nahas, G. G., ed. Report of The Ad Hoc Committee on Methodology, Current Concepts
,

of Acid-Base Measurement, Annals of the New York Academy of Science, New York City
274 pp. (Edward Bros., Inc., Ann Arbor, Mich., 1966).

[8] Austin, W. H., The use of arterial or venous blood in acid-base balance, J. Me.
Med. Assoc. 236 (1970).

170
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Gaithersburg, Maryland, July 7-8, 1975. Issued June 1977.

BLOOD SAMPLING AND HANDLING IN THE DETERMINATION OF BLOOD

pH AND BLOOD GASES

Arthur H. Richards
Division of Laboratories and Research
New York State Department of Health
Albany, New York 12001 USA ,

The collection and handling of blood for the determination of blood pH and blood gases
require that the source (arterial, capillary or venous), sampling site, collection materials
and storage conditions be precisely and clearly defined in order to ensure that a proper
sample is presented to the analytical unit.

Although the source of the blood must sometimes be selected on a purely pragmatic
basis, arterial blood is the preferred type of sample and will reflect the true status of
the pulmonary system with respect to pH and blood gas content. Collection of the sample,
usually from the brachial, femoral or radial arteries, requires a high degree of expertise.
The patient must be relaxed and the blood drawn with a minimum of trauma. Multiple arterial
samples are not recommended, but, when several samples might be required, a Cournand needle
or Seddinger catheter can be employed.

Blood collected via a vasodilated capillary bed has been reported to be virtually
equivalent to arterial blood [1-3]^. The proper technique for collection of a capillary
sample (usually from the earlobe, finger or heel) is to warm the site thoroughly by immersion
in warm water, then allow the blood to flow into a capillary tube without impairing or
forcing the blood flow. The results on arterialized capillary blood have been reported to
show good agreement with samples taken from a true arterial source, although a notable
exception has been reported: P02 arterial versus capillary measurements were reported to be
in poor agreement [5].

Venous blood can be employed for the routine measurement of CO2 content. However, it
is well documented that oxygen tension can be considerably different than in arterial blood
[10-12,14]. Some reports have stated that reasonably satisfactory samples can be obtained
by the application of heat to the source to create "arterialization. " One monograph recommends
that blood be taken from a vein on the back of a hand that has been subjected to 46-47 °C
heat for 10-15 minutes [11]. The use of a light tourniquet is acceptable, but it should be-
released during the drawing period. It has been reported that, when these conditions are
closely followed, good approximation of arterial blood is achieved for pH and PCO2, although
P02 results can be significantly different.

The collection device most frequently recommended is the glass- type syringe with matching
barrel and plunger. A light coating of heparin anticoagulant solution should be used for
lubrication as well as for preventing clot formation. Samples are always drawn under
anaerobic conditions. Plastic syringes cannot be used. as the porosity of the plastic creates
a potential for CO2 and O2 to diffuse into the plastic. Vacutainers containing sodium
heparinate have been shown to be acceptable [8]. Other anticoagulants are unacceptable, and
oil or lubricants must never be used.

It is recommended that the blood sample be placed in ice water if analysis cannot be
performed within 20 minutes [2]. It is reported that blood pH will not change more than
0.01 pH units when kept at 4 °C for up to four hours. PCO2 is a more labile determination.

Figures in brackets indicate the literature references at the end of this paper.

171
for which it is recommended that the sample stored on ice be analyzed within 30 minutes.
The sample must be kept from exposure to room air by inserting the collection needle into a
cork or rubber stopper or by using any appropriate covering device. The sample must be
remixed just prior to proceeding with the determinations. Mixing is best accomplished by
adding a few drops of mercury to the sample, both to assist in completely mixing the
anticoagulant and remix to the sample at a later time.

Numerous studies have compared the results of tests in which arterial blood and venous
and capillary bloods, arterialized by a variety of techniques, were used. Paine, et at. [4]
investigated a group of 29 hospitalized patients by making a comparison between arterial
samples, venous samples and arterialized venous samples. The arterialized venous samples
were obtained by prewarming a vein on the dorsum of the hand for 15-20 minutes and were
collected with a tourniquet applied to the wrist. The pH values were on the average 0.022
units lower than arterial blood, while the PCO2 was 2.86 mm Hg lower. It was concluded that
this arterialized sample is satisfactory for the estimation of arterial pH.

In a separate study, Jung, et al. [14] made comparisons of arterial and venous bloods
collected simultaneously with both syringes and vacutainers. Capillary blood was also used
in some of these studies. The pH, PCO2, bicarbonate, total buffer base and base excess were
all satisfactory on capillary blood, but P02 was not acceptable. The study stated that
blood obtained by Vacutainer was not adequate for any of the determinations, but haparinized
venous blood could be substituted for arterial blood except for the P02 determination.

A study by Spock, et al. [13] showed that P02 could be accurately determined on capillary
samples if a proper technique was employed. An excellent correlation with arterial samples
was shown in this study. These samples were obtained by thumb prick after the thumb had
been warmed at 40 °C. The samples were analyzed promptly and were representative of a number
of disease states.

In yet another study, Koch [11] has shown that capillary blood from the middle finger,
when arterialized by warming at 45 °C, was comparable to arterial blood for measuring pH,
PCO2 and bicarbonate concentration.

One of the favored collection sites for blood pH and blood gas determinations is from
the earlobe. Laughlin, et al. [6] have shown that by warming the earlobe with an electric
heater, P02 values could be obtained quite comparable with P02 values in arterial blood. In
the same study samples from finger punctures gave less satisfactory P02 values than those
from earlobe punctures.

Langlands, et al. [7] compared P02, PCO2 and pH values brachial artery and earlobe
blood. Their results showed that values in arterialized earlobe blood corresponded very
closely to arterial blood.

Gambino [8] has utilized the earlobe as a site for collection of capillary blood. He
described a rubber-cup collection system and compared his results with arterial values. No
significant differences were observed between the capillary blood and arterial blood values
for pH, CO2 content, PCO2 and oxygen saturation.

Siggaard-Anderson, et al. [9] have done studies on capillary blood and state that pH
shows a standard deviation of 0.006 pH units and PCO2 accuracy of ± 2 percent.

In summary, it would appear that although the ideal situation is to obtain arterial
blood, venous or capillary blood can serve as a valid substitute providing the collection
site is carefully chosen and prewarmed and the specimen is carefully collected and properly
cared for prior to analysis.

References

[1] Ibott, F. A., LaSanga, T. S. , Gin, J. B., and Inkpen, J. A., in Clinical Chemistry
Principles and Techniques, R. J. Henry, D. C. Cannon, and J. W. Winkelman, eds. (Harper
and Row, Hagerstown, MD, 1974).

172
[2] Fleischer, W. R. and Gambino, S. R., in Blood pH, P02, and Oxygen Saturation (American
Society of Clinical Pathologists, Commission on Continuing Education, Chicago, IL,
1972).

[3] Gambino, S. R. , Astrup, P., Bates, R. G., Campbell, E. J. M., Chinard, F. P., Nahas, G.
G. , Siggaard-Anderson, 0., and Winters, R. , Report of ad hoc Committee on Acid-Base
Methodology. Amer. J. Clin. Pathol. 46, 376 (1966).

[4] Paine, E. G., Boutwell, J. H., and Soloff, L. A., The reliability of "arterial ized"
venous blood for measuring arterial pH and PCO2, Amer. J. Med. Sci. 242 , 431 (1961).

[5] Adams, A. P., Morgan-Hughes, J. 0., and Sykes, M. K. , pH and blood-gas analysis. Anesthesia,
22, 575 (1968).

[6] Laughlin, D., McDonald, J., and Bedell, G., A microtechnique for measurement of P02 in
"arterial ized" earlobe blood, J. Lab. Clin. Med. 64, 330 (1964).

[7] Langlands, J. and Wallace, W., Small blood samples from earlobe puncture. Lancet, 1_,
315 (1965).

[8] Gambino, S., Collection of capillary blood for simultaneous determinations of arterial
pH, CO2 content, PCO2 and oxygen saturation, Amer. J. Clin. Path. 35, 175 (1961).

[9] Siggaard-Anderson, 0., Engle, K., Jorgensen, K. ,and Astrup, P., A micro method for
determination of pH, carbon dioxide tension, base excess and standard bicarbonate in
capillary blood, Scand. J. Clin, and Lab. Invest. 12^, 172 (1960).

[10] Adams, A., Morgan-Hughes, J., and Sykes, M., pH and blood-gas analysis. Anesthesia, 22,
47 (1968).

[11] Kock, G. , Comparison of carbon dioxide tension, pH and standard bicarbonate in capillary
blood and in arterial blood, Scand. J. Clin, and Lab. Invest. 17^, 223 (1965).

[12] Koch, G. , The validity of P02 measurement in capillary blood as a substitute for arterial
P02, Scand. J. Clin, and Lab. Invest. 2]_, 9 (1968).

[13] Spock, A., Lewis, M., and Albertson, T. , A microtechnique for measurement of oxygen
tension in capillary blood, J. Pediatrics, 68, 987 (1966).

[14] Jung, R. , Balchum, 0., and Massey, F., The accuracy of venous and capillary blood for
the prediction of areterial pH, PCO2 and P02 measurements, Amer. J. Clin. Pathol. 45,
129 (1966).

173
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

NON-ANALYTICAL SOURCES OF LABORATORY ERROR IN pH AND BLOOD GAS ANALYSIS

Jack H. Ladenson
Division of Laboratory Medicine
Departments of Pathology and Medicine
Washington University School of Medicine and Barnes Hospital
St. Louis, Missouri 63110, USA

Laboratory error includes more than just the analytical error associated with instrumen-
tal measurement. Properly, it should be defined as any error from the ordering of a test
procedure until the interpretation of results. In our experience, the non-analytical sources
of error are more severe than the instrumental and harder to document and eliminate [1]^.
The purpose of this presentation is to review some of these non-analytical sources of labora-
tory error in pH/blood gas measurements in an effort to alert both the laboratorian and
clinician to these problems.

1. Choice of Collection Site

A. Arterial versus venous blood

The differences between arterial and venous blood found by various studies are shown
in table 1. Comparisons of arterial and capillary blood obtained without any attempt to
arterialize the capillary collection site are also shown. It is evident that venous blood
obtained from an arm vein has a lower pH than arterial blood. The magnitude of this
difference varied from study to study but appears to be '^0.05 pH unit. When capillary
blood is compared to arterial blood, discrepancies between the conclusions of the studies
are found. This is probably due to differences between the patients studied as will be
discussed in the next section. Only small differences between arterial and capillary
blood were found in studies of patients under general anesthesia [6,10,11], and were
attributed to vasodilation and high cutaneous blood flow caused by anesthesia. These
phenomena have been reported to help to arterialize the capillary blood [8]. In newborns,
very poor correlations were found between arterial blood and blood obtained from an un-
warmed heel [15].

The Pco2 of venous blood is higher than arterial blood. The exact magnitude of the
mean difference varied from study to study (table 1) but appears to be between 5 and 10 mm
Hg. Again capillary blood agreed more closely with arterial blood than did blood from an
arm vein. An important exception to this finding was the poor correlation of capillary
(unwarmed heel) and arterial blood for PCO2 in newborns [15]. Very large differences in
P02 and O2 saturation between arterial and venous blood have been observed [4,7] with
smaller differences between arterial and capillary blood [12,14].

The data indicates that for most purposes, substitution of venous blood for arterial
blood will not cause clinically significant differences for pH, may cause clinically
significant differences for PCO2, and will invalidate measurements of oxygen status.

B. Arterial versus arterialized blood

Because of the possible risk and discomfort of arterial puncture [16,17], many
attempts to arterialize venous blood by means of warming or vasodilation have been

Figures in brackets indicate the literature references at the end of this paper.

175
 Table 1. Differences between arterial and venous blood. ^'

Venous PCO2 P02 O2 Saturation Number and type

Year site pH (mmHg) (mmHg) {%) of subjects Reference

1963 Vena Cava 005 +2.4 4 patients with 2

suspected car-
diac defects

1974 Central + 041 501 critically or 3

Venous seriously ill
patients on
100 percent O2
Central -7.5 328 critically or 3
Venous seriously ill
patients on
100 percent O2

1942 Int. Jug. + 053 -10 31 .7 50 normals 4

1959 Arm + 013 30 resting patients 5

1964 Antecubital + 06 +7.4 30 patients under 6

anesthes i a

1966 Cephal ic + 033 -4.7 +27.3 42-43 patients 7

1959 Hand or Wrist + 051 -8.3 14 patients 8

Hand or Wrist + 01 -1.2 7 patients with 8

warm extremities
Hand or Wrist + 002 -1.1 10 patients under 8
anesthesia

1961 Hand or Wrist 044 -3.4 14 patients 9

Hand or Wrist -1.6 10 patients under 10

anesthesia
(13 samples)

1962 Capi 1 lary ,

001 +0.4 31 patients under 11
earlobe anesthesia

1964 Capillary + 01 -1.0 15 patients under 6

earlobe anesthesia
Capillary , +9 8 patients 12
earlobe

1962 Capi1 lary ,

- 004 +0.2 20 patients under 11
finger anesthesia

1963 Capillary - 04 +6 4 patients with 2

finger suspected car-
diac defects

1964 Capillary ,
+ 03 -7.3 15 patients under 6
finger anesthesia

1965 Capillary . + 009 -2.8 13 patients 13

finger

1968 Capillary, +8,5 14 patients 14

thumb

1964 Capillary, .15(max) -48(max) 23 newborns 15

heel very poor very poor
correlation correlation

^ All differences are expressed as' arterial -venous

^ All capillary blood was non-arterial i zed

176
attempted. These results are summarized in table 2, Direct comparisons of arterial,
arterial ized earlobe, and arterial ized thumb pulp have been performed for P02 measure-
ments. Good agreement between arterial blood and blood prepared by massaging the earlobe
with nicotinic acid (Trafuril^) was obtained. Blood obtained from the earlobe prepared
by massage without nicotinic acid, the untreated thumb pulp, or the heated hand was found
to differ significantly from arterial blood [14]. Other direct comparisons also found
arterialized blood from the earlobe to be superior to that from the finger [12,27].

Table 2. Differences between arterial and "arterialized" blood.

Sample PCO2 P02 O2 Saturation Number and type

Year site PH (mmHg) (mmHg) (%) of subjects Referc

1 959 Dorsal hand + .008 -1.9 14 patients 8

vein

1961 Dorsal hand + ,018

, -0.1 14 patients 9
vein

1972 Dorsal hand + ,005

, -1.1 16 8 patients 18
vein (P02>70) and
c
.
0
y 5 normals
(Po2<70)

1944 Earlobe 0.5 n patients 19

1961 Earlobe -,,002 22 patients 20

Earl obe +0.5 10 patients under 10

anesthesia (21
samples)

1964 Earlobe 3 33 patients 12

L. a [ 1 u uc Cnnri
agreement 19 patients 21

Earl obe 0.1 20 pati ents 22

1965 Earlobe 0.3 21 patients 23

Earlobe ,006 + 1.0 0.6 14 patients and 24
2 normals

1966 Earlobe r = 0.93 16 patients 25

1967 Earlobe -..001 + 0.3 17 patients (respi- 26

ratory insuffi-
ciency)

1968 Earlobe 0.2 42 patients 27

(P02, 34-103)
Earlobe 90 12 patients 27
(P02, 222-603)
Earl obe 0. 22 patients during £.1

(P02, 40-91) exercise

1968 Ffl rl nhp 0. 5 14 patients 14

1972 Earlobe ,002 -0.8 0.4 25 normals 28

1973 Earlobe + .006

, -1.4 0.2 84 patients, pre- 29
operative, opera
tive, and post-
operative

1961 Ear or + ,003 0.2 0.1 13 patients 30

finger
Ear or ,003 8 patients after 30
finger one minute
exercise

177
 Table 2. Differences between arterial and "arterial ized" blood (continued)

Sampl PCO2 P02 O2 Saturation Number and type

Year site pH (mmHg) (mmHg) (%) of subjects Refere

Ear or -0.2 5 patients after 30

finger one minute
exercise

1961 Finger 6.4 43 patients 26

1963 Finger -.011 +2.2 47 patients 2

1965 Finger + .034 -0.3 30 patients 13

1966 Finger -.004 0.2 0 20 patients 31

1962 Finger and 1.0 0.2 22 patients 32

heel

1964 Heel + .027 -5.9 20 newborns 15

0-1 hour

+ m -1.0 20 newborns 15
1-3 hours
Heel + .006 -0.7 34 newborns 15
over 3 hours

1967 Heel + .026 -6.2 34 newborns 33

30 minutes to
24 hours
Heel 42 newborns 33
2-6
f— \j davs
\j uy0

Heel 9.5 5 hours to 33

6 days

1969 Heel + .007 -8.3 6.3 45 newborns 34

3-95 hours

1973 Heel r = 0.99 -(10-15 ) r = 0.86 37 npwhnm<i nldpr 35

than 3.5 days

1965 Scalp Smal 1 Smal 1 Smal 1 48 patients 36

^All differences expressed as arterial -"arterial i zed".

As table 2 indicates, in almost all studies of adults the capillary method gave
results equivalent to arterial blood. In newborns, however, differences have been found
and it appears that this technique is not valid in very young children (less than a day)
and questionable even in somewhat older children. The lack of agreement between arterial
and arterial ized capillary blood in newborns has generally been ascribed to poor circu-
lation and warnings about using arterialized capillary blood in any patient with circula-
tory shock have been presented [12,25,27,37]. Observations by Banister [34], however,
have suggested the presence of edema of the foot, rather than circulatory status, as the
important factor in obtaining valid results from- capillary blood in children.

Prior publications [38,39] have presented conclusions which are in keeping with the
data reviewed here. These conclusions were that arterialized capillary blood is equiv-
alent to arterial blood except for the very young, newborns with respiratory distress
syndrome, and patients in shock. Unfortunately, these categories represent a number of
the patients in whom a valid capillary sample would be highly desirable.

2. Collection Containers

Three types of collection containers for macro pH/blood gas determination have been
suggested; glass syringe, plastic syringe, or vacuum tube. In an early study, no dif-

178
ference between the pH of whole blood obtained from a vacuum tube and that sampled di-
rectly from the patient was noted. In the same study, no differences in plasma CO2
content between sedimented blood from a syringe and centrifuged vacuum tubes was observed.
The O2 saturation of arterial blood collected in a syringe or in a vacuum tube was also
found to be the same [40]. Fleisher and Schwartz [41] noted a poor correlation between
the usual vacuum tubes and glass syringes for arterial blood gas determinations and
designed a nitrogen filled vacuum tube. Comparison of this tube and glass syringes,
showed excellent agreement for PCO2 and P02, and only a +0.01 difference for pH [41].
These nitrogen filled tubes were also tested by Lang, et al. [42], who found them to be
severely contaminated with oxygen. For blood samples, good correlations for pH and PCO2
were observed, but significantly higher P02 values were found with the vacuum tube. Further
work by these authors showed the unsuitability of the presently available vacuum tubes for
measurement of'Po2 and oxygen saturation [43].

A. Glass versus plastic

Saline equilibrated to a high P02 (^^700 mm Hg) showed no decline in measured P02
when kept in glass syringes, but showed a decline of 21.6 percent after 3 hours and 67
percent after 48 hours when stored in plastic syringes. Blood tonometered to high P02
values ('^^600 mm Hg) also showed a greater decay in plastic syringes as compared to glass.
The differences in decay rate were apparent in fifteen minutes [14]. The diffusion of
oxygen and carbon dioxide from plastic syringes has been studied. The percent change in
gas content per hour varied considerably (over 100 percent) between the two types of
plastic syringes tested. For the better syringe type, the changes at room temperature (in
percent change of gas content/h) were 0.034 and 0.082 for O2 and CO2, respectively. When
the syringes were kept on ice these changes were 0.128 and 0.154. These authors expected that
the changes would be considerably slower when the syringe content was liquid and did not
feel that the decay would be clinically significant [44]. Measuring the P02 of deoxyge-
nated water gave higher results for samples in plastic syringes than for glass (37 mm Hg
compared to 13 mm Hg) [42]. Laver and Seifen [45] have noted that plastic syringes maintain
a high P02 as well as glass syringes for up to 2 hours, but recommend glass syringes if
delay in analysis is expected.

Scott, et al. [46], found that when water tonometered to 96 percent O2 was sampled in
6 different types of plastic syringes, large decreases were found when the results were
compared to the same sample in glass syringes. Similarly, large increases were found when
sampling water at 0 percent O2. Syringes made of polystyrene showed considerably greater
changes than those made of polypropylene or S.A.N, co-polymer. Tonometered blood samples
showed the following changes (in mm Hg) after 11.5 minutes, -68, 0.4, and 0.0 at an initial
P02 of 680, 102, and 68, respectively. These authors presented evidence that the loss or
gain in O2 was due to diffusion into the walls of the syringe rather than diffusion through
the syringe. These authors estimated the magnitude of errors expected under various
conditions. This data is presented in table 3, and extimates the influence of initial P02,
temperature of storage, pH, hemoglobin and barometric pressure on the error introduced by
the use of plastic syringes. It is evident that the error introduced will depend on a
number of factors and the acceptibil ity of plastic syringes will depend on the application.
The above study also noted that interchangeable glass syringes were not as effective as
glass syringes with matched barrel and plunger.

The above data indicates that under most conditions, the interchange of glass or
plastic syringes will not affect the clinical utility of the oxygen measurement [44-46].
The excellent work of Scott (table 3), however, does show that in some circumstances
inaccurate results will be obtained when plastic syringes are utilized. An example of
this situation may be the severely anemic patient with alkalosis who is on a respirator.
There are advantages to the laboratory in using plastic syringes such as cost, safety, and
seal reliability. When mixing the blood, the plunger of more than one glass syringe has
fallen out if care is not taken. The area of collection containers appears to be in need
of some fresh approaches which would retain the accuracy of glass syringes but with the
convenience of plastic ones. Regardless of whether glass or plastic syringes are used,
the addition of mercury to facilitate mixing the blood is not recommended due to the
danger of peripheral embolism [47].

179
 Table 3. Computed oxygen exchange of blood stored in plastic syringes [46].

Change of tension Change

Storage Ini tial of
time tension content
(min) (mmHg) (mmHg) (% initial {7o initial)

"Normal" 650 -106 3 -16 4 -1 4

30-60
condi tions 300 -21 3 -7 1 -0 3
100 5 8 5 8 0 4
33 0 3 0 9 0 9

Effect of storage time 650 -30 3 -4 7 -0 4

Ta
650 37 -129. 1 -19. 9 -1. 7
Effect of ambient
650 4 -83. 3 -12. 8 -1. 1
temperature during 30-60
100 37 3. 3 3. 3 0. 2
storage (Ta)
100 4 7. 6 7. 6 0. 5

Hb
650 15 -106.3 -16.4 -1.4
650 5 -107.5 -16.5 -3.6
Effect of Hb 30-60
100 15 5.8 5.8 0.4
100 5 12.7 12.7 1 .1

pH
7.7 -24.3 -8.1 -0.4
Effect of pH 30-60 300
6.8 -10.8 -3.6 -0.2

Pb
780 6.1 0.4
Effect of barometric
30-60 100 710 5.2 0.3
pressure- (Pb)
600 3.7 0.2

Notes: Syringes: 5 ml or 2 ml , initially in equilibrium with ambient air.

Exchange factor: 8 x 10''* (ml/lOO m)/mmHg for 30-60 minutes storage and
2 X lo"** for 2 minutes.
Normal conditions: Hb = 15 g/lOO m, pH = 7.4, Pb = 760 mmHg, initial
blood temperature = 37 °C, ambient temperature = 22.5 °C.
Reprinted with the permission of Brit. Med. J.

3. Effect of Anticoagulants

While serum and plasma give comparable results for pH [40] and CO2 content [48],
^
whole blood is necessary for oxygen analysis. The use of citrate [40,49], oxalate (NH^
and K"*") [40], or EDTA [7,40] as anticoagulants has been found to cause low results for pH.
The use of potassium oxalate alone as anticoagulant caused an increase in pH [40,50].
This increase in pH was shown to be related to chelation of calcium [50]. Heparin has
almost no effect on pH (1 mg/ml = -0.003). This small effect was the same in lipemic
blood but no data concerning the storage of such specimens was presented [51]. In an
effort to inhibit glycolysis, heparin-fluoride mixtures have been utilized. In one study,
no effect of fluoride on oxygen consumption was found [52] while in others a decrease in
decay of P02 of 1/4 to 1/3 was observed [53,54]. Paradoxically, fluoride completely eli-
minated the rise in PCO2 with storage even though the P02 was only partially affected [53].
No differences in the pH of four blood specimens were found when heparin-fluoride was
utilized [49]. The addition of fluoride in amounts sufficient to inhibit glycolysis
lowers the pK' of blood by 0.01 units due to a salt, effect [55]. Fluoride (5 mg/ml)
caused an increase in pH of 0.03 and a fall in PCO2 of 6 mm Hg. This effect was found to
be concentration dependent [51]. Upon storage, blood containing fluoride reacted differently
than blood without fluoride. When stored at 38 °C, the change in the pH of blood reached a
minimum at 1/2-1 hour (0.01-0.02) and a maximum at 3-4 hours (0.17-0.19). This phenomena
was concentration dependent, delayed at 22 °C, and was not due to leukocytes [51]. The
addition of fluoride causes a rise in potassium which invalidates its use for potassium
measurement [56]. In our laboratory, potassium analysis is being requested on the same
sample of blood as pH/blood gas with greater frequency. The rise in potassium and the above
data on changes in pH during storage suggest that fluoride not be used as an additive to
samples for pH/blood gas measurement. It is therefore recommended that heparin be used as
the sole anticoagulant for pH/blood gas analysis.

180
 4. Stability In Vitro

In this section the stability of pH, PCO2, CO2 content, and P02 in blood under varying
conditions is reviewed. Since anticoagulants other than heparin can have a deleterious
effect on the measured values of these quantities, only data obtained with this anticoagu-
lant will be considered. As will be discussed later, the temperature of measurement can
affect results and therefore some early data where the temperature control is suspect [57]
will not be considered.

A. pH

The average decay rates for pH (-pH/h) found in various studies are shown in table 4.
This decay rate has been shown to be independent of hemoglobin concentration but very
dependent on the number of leukocytes [51]. It is evident that the decay in pH is quite
temperature dependent and this dependence has been shown to follow the Arrhenius relation-
ship [62]. Exposure to air by dropping 50 yl of blood onto a glass slide resulted in a
rise of only 0.01 pH after 2 minutes [51]. The small decay rate for pH should not cause
problems in routine handling of blood specimens if they are placed in ice water after
being obtained. Such treatment has been stated to result in decay rates of no more than
-0.008 pH/h even in patients with leukemia [63],

B. Pcoa

As shown in table 4, PCO2 in blood has been found to increase during storage. For
blood stored at 4 °C the average increase is less than 0.7 mm Hg/h. The magnitude of this
increase has been found to be related to the number of white blood cells present but not
to the number of reticulocytes [53]. This small increase will not usually be clinically

Table 4. Stability of pH, PCO2, CO2 content and P02.

Storage Rate of Number of

Parameter Year temperature change experiments Comments Reference

1956 0. 060 (-pH/h) ? Data not given 55

1961 38 °C 0 062 (-pH/h) 10 Linear for 3 hrs and 51

dependent on leuko-
cytes but not hemo-
globin
22-24 °C 0 024 (-pH/h) 5 Linear for 3 hrs and
dependent on leuko-
cytes but not hemo-
globin
0-4 °C 0 006 (-pH/h) 4 Linear for 3 hrs and
dependent on leuko-
cytes but not hemo-
globin

1965 2-4 °c 0 006 (-pH/h) 17 Based on 4 hours 13

1961 38 °c 4 8 (mmHg/h) 10 Linear for 3 hours 51

22-24 °c 2 5 (mmHg/h) 5

0-4 °c 0 6 (mmHg/h) 4

1965 37 °c 6 6 (mmHg/h) ? 52
4 °c 0 7 (mmHg/h) ?

2-4 °c .
0 2 (mmHg/h) 5 13

181
 .
.

Table 4. Stability of pH, Pcoa, CO2 content and P02 (continued)

Storage Rate of Number of

Parameter Year temperature change experiments Comments Reference

1959 room? 0 ? No effect after 2 40

hours

1971 room? 0 1 No effect after 2 58

hours

PO; 1961 37 °C 175 (-mmHg/h) 5 Initial P02 , over 400 52

20 (-mmHg/h) 5 Initial P02 , 'v, 100

1965 37 °C 3. 4 (-mmHg/h) ? Initial P02 , 51-93 59

22-24 °C 1. 8 (-mmHg/h) ? Rate was not constant,
and was calculated
from 1 hr data
1 °C 1 7 (-mmHg/h) ? Rate was not constant,
and was calculated
from 1 hr data
37 °C 121 (-mmHg/h) ? Initial P02, 360-670,
rate was not constant
and was calculated from
1 hr data

22-24 °C 16 (-mmHg/h) ? Initial P02, 360-670,

rate was not constant
and was calculated
from 1 hr data
1 °C 1 7 (-mmHg/h) 7 Initial P02, 360-670,
rate was not constant
and was calculated
from 1 hr data
37 °C 156 (-mmHg/h) 55 High initial P02 53

4 °C 30 (-mmHg/h) 10 High initial P02

07 op 24 ( -mmHn / h ^ 5 Initial P02, 60-100 53

37 °C 48 (-mmHg/h) 16 Initial P02, 125-150

37 °C 162 (-mmHg/h) 55 Initial P02, 440-610

1966 4 °C 10 (-mmHg/h) 7 31

room 450 (-mmHg/h) 89 samples Breathing 100% O2. 60

5 normals Rate is projected based
on first 10 minutes

1968 37 °C 180 (-mmHg/h) 7 Initial P02, 680 14

24 °C 120 (-mmHg/h) 7 Initial P02 , 680

4 °C 30 (-mmHg/h) 7 Initial P02 , 680

37 °C 50 (-mmHg/h) 7 Initial P02, 143

24 °C 20 ( -mmHg/h) 7 Initial P02 , 143

4 °C 10 (-mmHg/h) 7 Initial P02, 143

7 180 (-mmHg/h) 7 Higher if leukocyto- 61
sis present

1968 0-4 °C -0 5 (-mmHg/h) 10 P02, 52-92 27

significant but extensive information on patients with high levels of white blood cells,
e.g., leukemia, is not available.

C. CO2 content

As expected from the data concerning pH and PCO2, little or no change in blood CO2
content has been found when the blood is kept unopened to air (table 4). The temperature

182
of centrif ugation has been suggested as a possible small cause of error in CO2 content
measurements because of the different temperature coefficients of plasma and whole blood
[64]. In two studies, however, no effect of the temperature of centrifugation (5 °C or
37.5 °C) was found [40,55]. Since the temperature of a centrifuge operated in an ambient
temperature room is over 30 °C, this potential error source can be ignored.

A more significant stability problem occurs when plasma or serum is placed in the
plastic sampling cups which are so commonly used. The loss of CO2 in sample cups open to
the air has been shown to average 3.5 meq/1 for 15 samples after 1 hour. In one patient
with a PCO2 of 75 mm Hg, a 7 meq/1 loss after 3 hours occurred [65]. Four techniques to
minimize CO2 loss to the atmosphere have been described. The oldest of these techniques
is the use of oil to eliminate contact of the sample with the atmosphere. This technique
has been criticized for two reasons: (1) since blood collected with and without oil gives
the same results [48], its use during collection is superfluous; and (2) CO2 is quite solu-
ble in oil (see annotated bibliography in reference 66) and significant absorption of CO2
occurs if the sample is mixed with the oil. Such mixing can occur during collection
and/or centrifugation [67]. While covering the sample cups with oil does retard CO2 loss
to the atmosphere [68], the potential CO2 loss to the oil makes its use unadvisable.

The second method consists of placing a plastic disk in the sample cup so that the
sample is isolated from air. The sample probe then offsets the disk at the time of analysis.
This system has resulted in no loss of CO2 in 3 hours [69]. The third method is the use
of added alkali (1 drop of 1 N_ NHi+OH) to decrease the PCO2. This technique was found to
stabilize the CO2 content for up to 4 hours and was felt to be superior to the plastic
disk method [65]. The fourth technique is the use of a metal or plastic plate to cover
all the cups in a sample tray. Such equipment is commercially available but the author is
unaware of any published studies on its effectiveness in retarding CO2 loss.

The loss of CO2 to the atmosphere during handling is certainly capable of affecting
the clinical interpretation of CO2 content measurements. Of the methods noted above for
minimizing this phenomena, the alkali method of Gambino [65] appears to be best but further
validation by other workers would be helpful.

D. P02

The stability of P02 in shed blood under various conditions is shown in table 4. It
is evident that the stability of P02 is quite temperature dependent [14,25,53,54,59,70,71]
and rapid chilling of all blood samples is strongly suggested. It is also evident that
the rate of decay of P02 is highly dependent on the initial P02 [14,25,52, 53,59,71,72].
This has been attributed to the shape of the oxygen dissociation curve [53]. The O2 decay
rate in shed blood is not affected by total hemolysis of the blood sample [53].

The source of the oxygen consumption in blood specimens has been the subject of
study. In one study the oxygen consumption of normal mature erythrocytes was found to be
negligible but blood with abnormal numbers of reticulocytes had an oxygen consumption
proportional to the percentage of reticulocytes present [73]. This increase in oxygen
consumption with increase in percentage of reticulocytes has been confirmed by others
[53,74]. The rate of oxygen consumption has also been related to the white cells. Figure
1 shows the data of Lenfant and Aucutt [53] relating oxygen consumption to the white blood
count. Normal and abnormal white cells are included in this data but no comparison of
their oxygen consumption was made. Higher oxygen consumption in blood from leukemic
subjects has been noted by another worker as well [61]. Because of the dependency of O2
consumption on the cellular components of the blood, correction factors for storage time
are not recommended and prompt analysis of chilled specimens appear the best way of
handling blood for accurate P02 measurements.

Exposure of blood to air for short periods (10-30 seconds) results in significant
changes in P02 toward the P02 of room air [25]. The presence of air bubbles (up to 10
percent of the total syringe volume) has little effect on P02 until the surface area is
increased by mixing [71]. The presence of air bubbles in blood samples has been blamed
for discrepancies in interlaboratory surveys for P02 [75]. Since whole blood is mixed
prior to pH/blood gas analysis to avoid suspension of the cells, great care should be
exerted to exclude air bubbles at the time the sample is obtained.

183
 30
PO2 decay, mm Hg/min

25

20 •

15

10"

-
5

\ •• ••.
• • • •
White Blood Cells Count
per mm^

10,000 20,000 50,000 100,000 200,000 500,000 1,000,000

Figure 1. P02 decay as a function of the white cell count. Black circles are for samples
containing NaF. Arrows indicate the P02 decay of samples with normal white cell counts
without f-^) and with ( —
>) NaF. (From Lenfant and Aucutt [53] with permission of
J. Appl. Physiol. )

5. Effect of Measurement Temperature

pH, P02, and PCO2 values are all dependent on the measurement temperature. Since
virtually all commercial pH/blood gas equipment measures these quantities at 37 °C, an
error is introduced when the temperature of the patient at the time the sample is obtained
deviates from this value. The work of various authors has been reviewed and a simplified
table derived for temperature correction of pH, P02, and PCO2 [76], This table is shown
as table 5 and is recommended as a guide to temperature effects on pH and blood gas measure-
ments.

6. Effects of Red Cells on pH Measurement

No effects of hematocrit on the measurement of P02 or PCO2 have been reported. However,
differences in measured pH between plasma and whole blood have been observed (plasma-whole
blood =0.01) [55,77]. The relationship between pH and hematocrit (0-100) in 5 blood
samples has been elucidated and the cause of the small pH discrepancies was felt to be due
to differences in residual liquid junction potentials between plasma and whole blood [77].
The magnitude of this error is quite small and except in highly unusual circumstances will
not be clinically significant.

7. Effects of Dilution

A 12-13 percent dilution of blood with physiological saline produced no significant

change in pH but resulted in a fall in PCO2 of 16 percent and a fall in plasma bicarbonate
of 15 percent [51]. Another study showed an increase in pH of 0.023 pH unit after a 20
percent dilution with physiological saline. The addition of calcium (5 meq/1) to the
saline eliminated this dilution error [78]. A 33 percent dilution of blood with isotonic
saline caused no change in pH, a decrease of 3.8 mm Hg (4 percent) in P02 and a decrease of
9.2 mm Hg (24 percent) in PCO2 [24]. The above data indicate that pH and P62 exhibit little

184
 Table 5. Temperature- correction factors for blood pH and gas measurements [76].^

atient's temperature pH
^C02
{%) (%)

°F °C (Add to observed values)

110 43 -.09 +22 +33

109 42.5 -.08 +21 +32
108 42 -.07 +19 +30

107 41 .5 -.07 +17 +27

106 41 -.06 +16 +25
105 40. 5 -.05 +14 +22

104 40 -.04 +12 +19

103 39.5 -.04 +10 +16
102 39 -.03 +8 +13

101 38. 5 -.02 +6 +10

100 38 -.01 +4 +7
98-99 37 None None None

97 36 + .01 -4 -7
96 35. 5 + .02 -6 -10
95 35 + .03 -8 -13

94 34. 5 + .04 -10 -16

93 34 + .04 -12 -19
91 33 + .06 -16 -25

90 32 + .07 -19 -30

88 31 + .09 -22 -35
86 30 + .10 -26 -39

84 29 + .12 -29 -43

82 28 + .13 -32 -47
81 27 + .15 -34 -51

79 26 + .16 -37 -54

77 25 + .18 -40 -57
75 24 + .19 -43 -60

73 23 + .21 -4S -63

72 22 + .22 -48 -65
70 21 + .24 -50 -67

68 20 + .25 -53 -70

^Reprinted with permission of Cliniaal Chemietyy.

change, even with high dilution, but PCO2 and CO2 content change in proportion to the degree
of dilution. Unless a large volume of anticoagulant is used, this should not introduce a
significant error. In sampling from indwelling catheters, care should be taken to remove
the fluid filling the catheter before obtaining the blood specimen if accurate assessment of
CO2 status is desired.

8. Effect of Stasis

No difference in pH between venous blood obtained without a tourniquet and blood

obtained with the tourniquet in place for 2 minutes was found in 5 subjects. It was
emphasized that the tourniquet should be left in place since the stasis produced is at the
capillary level and is not reflected in the vein until the tourniquet is released [79].
These results have been confirmed and it was noted that the tourniquet can be left in
place for 5 minutes without exercise (hand pumping) but that after tourniquet release the
pH decreased precipitously [5].

185
 .

9. Miscellaneous Effects

Arterial pH was found to decrease by over 0.2 pH units after exhausive exercise and
took over 30 minutes to recover. Pcoa decreased by approximately 5 mm Hg as well [80].
Continuous monitoring of blood pH has shown small fluctuations with each group of respira-
tions and has shown that hyperventilation results in a large increase in pH [81]. Warnings
concerning increases in pH and decreases in PCO2 during hyperventilation due to the pain
or anxiety of arterial puncture have been presented [82]. PCO2 values obtained in the
sitting or standing position are 3-4 mm Hg lower than those obtained while recumbent [83].
The above data suggest that pH/blood gas studies should be performed on subjects known to
be at rest unless special investigations are being performed. The injestion of food has
been shown to cause an increase in arterial pH of 0.015 to 0.03 with no change in PCO2.
The PCO2 was found to rise with sleep (5 mm Hg maximum) and the pH to decrease [84].
These changes associated with sleep and meals are small but should be thought of when it
is decided when to obtain a sample for pH and blood gas analysis. Capnography (recording
of expiratory COa-curve) did not alter the measured values for P02 or PCO2 in patients
with chronic lung diseases [85]. Halothane caused a slow response of some P02 electrodes
and could cause erroneous P02 results with such electrodes [86]. Further work on this
problem with more modern P02 electrodes has not yet been reported.

10. Conclusion

Advances in instrumentation have resulted in reliable, commercially available systems

for the analysis of pH and blood gases. The future will probably see even more reliable
instruments and these determinations will become more commonly performed than they are
today. As the analytical component of laboratory error decreases it will be even more
important for the clinician and laboratorian to appreciate the myriad of other factors
which can lead to "laboratory error". In reviewing this information, it was obvious that
for proper interpretation of laboratory data more knowledge needs to be obtained regarding
these factors. Factors such as composition and design of collection devices, changes in
values when storing blood with abnormal cells, and laboratory sample handling for CO2
content require further careful investigation.

References

[1] Ladenson, J. H., Patients as their own controls: Use of the computer to identify
"laboratory error", Clin. Chem. Z\_, 1648 (1975).

[2] Searcy, R. L., Gordon, G. F., and Simms, N. M. , Choice of blood for acid-base studies.
Lanaet 2_, 1232 (1963)

[3] Wilson, R. F. Wilson, J. A., Gibson, D. B., and Lucas, C. E., Arterial -central venous
,

differences in critically ill and injured patients, J. Trauma. 1_4, 924 (1974).

[4] Gibbs, E. L., Lennox, W. G. Nims, L. F., and Gibbs, F. A., Arterial and cerebral
,

venous blood-arterial-venous differences in man, J. Biol. Chem. 144 325 (1942). ,

[5] Gambino, S. R. , Comparisons of pH in human arterial, venous, and capillary blood.

Amev. J. Clin. Path. 32, 298 (1959).

[6] Howland, W. S., Schweizer, 0. and Murphy, T. W. Estimation of acid-base of venous

,

and arterial blood from capillary samples. Acta. Anaesth. Saand. 8, 191 (1964).

[7] Jung, R. C, Balchum, 0. J. and Massey, F. J., The accuracy of venous and capillary
blood for the prediction of arterial pH, PCO2 and P02 measurements, Amer. J. Clin.
Path. 45^, 129 (1966).

[8] Brooks, D. and Wynn, V.,'Use of venous blood for pH and carbon-dioxide studies especially
in respiratory failure and during anaesthesia. Lancet 1_, 227 (1959).

[9] Paine, E. 6., Boutwell, J. H, and Soloff, L. A., The reliability of "arterial ized"
venous blood for measuring arterial pH and PCO2, Amer. J. Med. Soi. 2A2^, 431 (1961 ).

186
[10] Cooper, E. A. and Smith, H., Indirect estimation of arterial PCO2, Anaesthesia j_6,
445 (1961).

[11] Knudsen, E. J. and Hansen, P., Carbon dioxide tensions in non-arterialised capillary
and arterial blood during anaesthesia, Acta. Anaesth. Soand. 29 (1962).

[12] Laughlin,D. E., McDonald, J. S., and Bedell, G. N., A microtechnique for measurement
of P02 in "arterial i zed" earlobe blood, J. Lab. Clin. Med. 64, 330 (1964).

[13] Koch, G. , Comparison of carbon dioxide tension, pH and standard bicarbonate in

capillary blood and in arterial blood: with special respect to relations in patients
with impaired cardiovascular and pulmonary function and during exercise, Scand.
J. Clin. Lab. Invest. ]]_, 223 (1965).

[14] Maclntyre, 0., Norman, J. N., and Smith, G. , Use of capillary b'lood in measurement of
arterial P02, Brit. Med. J. 3_, 640 (1968).

[15] Gandy, G. Grann, L., Cunningham, N., Adamsons, K.

, Jr., and James, L. S., The
,

validity of pH and PCO2 measurements in capillary samples in sick and healthy newborn
infants, Pediatrias 34, 192 (1964).

[16] Mortensen, J. D. , Clinical sequelae from arterial needle puncture, cunnulation, and
incision, Ciraulation 35, 1118 (1967).,

[17] Letters to the Editor, Brit. Med. J. 2, 310, 510, 511, 758, 819, 877, 1072 (1964).

[1*8] Forster, H. V., Dempsey, J. A., Thomson, J., Vidruk, E., and DoPico, G. A., Estimation
of arterial P02, PCO2, pH and lactate from arterialized venous blood, J. Appl.
Physiol. 32, 134 (1972).

[19] Lilienthal, J. L., Jr. and Riley, R. L., On the determination of arterial oxygen
saturations from samples of "capillary" blood, J. Clin, invest. 23, 904 (1944).

[20] Maas, A. H. J. and van Heijst, A. N. P., A comparison of the pH of arterial blood
with arterial ised blood from the earlobe with Astrup's micro glass electrode. Clin.
Chim. Acta. 6, 31 (1961).

[21] Gambino, S. R. Goldberg, H. E., and Polanyi, M. L., A micro method for determination
,

of oxygen saturation by means of reflection spectrophotometry, Amer. J. Clin. Path.

42, 364 (1964).

[22] Maas, A. H. J., Zuijdgeest, L. P. W. A., and Kreukniet, J., Microspectrophotometric

determination of the haemoglobin oxygen saturation in haemolyzed arterialized capillary
blood, Clin. Chim. Aota. 9_, 236 (1964).

[23] Torjussen, W., Oxygen saturation in capillary blood, Soand. J. Clin. Lab. invest.
17, 505 (1965).

[24] Langlands, J. H. M. and Wallace, W. F. M. Small blood-samples from earlobe puncture:

,

a substitute for arterial puncture. Lancet 2^, 315 (1965).

[25] Johnstone, J. H., Measurement of arterial and capillary blood oxygen tension. J.
Clin. Path. 19, 357 (1966).

[26] Torjussen, W. and Nitter-Hauge, S., pH and PCO2 in capillary blood from the earlobe,
Soand. J. Clin. Lab. Invest. 19_, 79 (1967).

[27] Koch, G. The validity of P02 measurement in capillary blood as

, a substitute for
arterial P02, Soand. J. Clin. Lab. 2]_, 10 (1968).

[28] Kanarek, D. J., Goldman, H. I., and Zwi , S., Arterial oxygen tension values in normal
adults at an altitude of 1763 metres, S. Afr. Med. J. 46, 315 (1972).

187
Sadove, M. S., Thomason, R. D., and Jobgen, E., Capillary versus arterial blood gases.
Anesthesia and Analgesia 52, 724 (1973).

Gambino, S. R. , Collection of capillary blood for simultaneous determinations of

arterial pH, CO2 content, PC02, and oxygen saturation, Amer. J. Clin. Path. 35,
175 (1961).

Spock, A,, Hinton, M. L. and Albertson, T. H., A microtechnique for measurement of

oxygen tension in capillary blood, J. Pediatrics 68, 987 (1966).

Bates, G. D. and Oliver, T. K. , Jr., A micromodif ication of the bubble method for
direct determination of blood gas tensions, J. Appl. Physiol. ]7_, 743 (1962).

Koch, G. and Wendel H., Comparison of pH, carbon dioxide tension, standard bi-
,

carbonate and oxygen tension in capillary blood and in arterial blood during the
neonatal period. Acta. Ped. Soand. 10 (1967).

Banister, A., Comparison of arterial and arterialized capillary blood in infants with
respiratory distress, Aroh. Dis. Childh. H, 726 (1969).

Hunt, C. E., Capillary blood sampling in the infant: usefulness and limitations of
two methods of sampling, compared with arterial blood, Pediatrics 51_, 501 (1973).

Desai, S. D. , Holloway, R., and Thambiran, A. K. , Blood-samples from earlobe puncture.

Lancet 2_, 1126 (1965).

Worth, G., Blood samples from earlobe puncture. Lancet 2^, 907 (1965).

Siggaard-Andersen, 0., Acid-base and blood gas parameters: arterial or capillary

blood? Scand. J. Clin. Lab. Invest. 21_, 289 (1968).

Questions and answers. Sources of blood samples for blood gas studies in critically
ill patients, J. A.M. A. 216., 1042 (1971).

Gambino, S. R. , Heparinized vacuum tubes for determination of plasma pH, plasma CO2
content, and blood oxygen saturation, Amer. J. Clin. Path. 32, 285 (1959).

Fleisher, M. and Schwartz, M. K. , Use of evacuated collection tubes for routine

determination of arterial blood gases and pH, Clin. Chem. V7, 610 (1971).

Lang, G. E., Mueller, R. G., and Hunt, P. K. Possible error resulting from use of
,

"nitrogen-filled" vacutainers for blood-gas determinations, Clin. Chem. ]9_, 559

(1973).

Mueller, R. G. and Lang, G. E. , Phase equilibrium of oxygen in "nitrogen-filled

vacutainers", Clin. Chem. 19, 1198 (1973).

Rosenberg, E. and Price, M. Diffusion of CO2 and 0? through the walls of plastic
,

syringes. Can. J. Physiol. Pharmacol. 51_, 608 (1973).

Laver, M. B. and Seifen, A., Measurement of blood oxygen tension in anesthesia.

Anesthesiology 26_, 73 (1965).

Scott, P. v., Horton, J. N., and Mapleson, W. W. , Leakage of oxygen from blood and
water samples stored in plastic and glass syringes, Brit. Med. J. 3_, 512 (1971).

Lathem, W. Lesser, G. T. , Messinger, W. J., and Galdston, M.

,
Peripheral embolism by
,

metallic mercury during arterial blood sampling. Arch. int. Med. 93, 550 (1954).

Paulsen, L., Comparison between total C02 content ("total CO2") in plasma/serum from
blood collected with or without paraffine oil, Scand. J. Clin. Lab. Invest. 9_,
402 (1957).

Graig, F. A., Lange, K. Oberman, J., and Carson, S., A simple, accurate method of
,

blood pH determinations for clinical use, Arch. Bioohem. Biophys. 38, 357 (1952).

188
Rodkey, F. L., Spectrophotometric determination of blood pH, J. Biol. Chem. 236 ,

1589 (1961).

Siggaard-Andersen, 0., Sampling and storing of blood for determination of acid-

base status, Saand. J. Clin. Lab. Invest. 13, 196 (1961).

Asmussen, E. and Neilsen, M. , The disappearance of oxygen from shed blood at body
temperatures, Soand. J. Clin. Lab. Invest. 297 (1961).

Lenfant, C. and Aucutt, C. Oxygen uptake and change in carbon dioxide tension in
,

human blood stored at 37 °C, J. Appl. Physiol. 20, 503 (1965).

Mellemgaard, K. , Disappearance of oxygen from shed blood at various temperatures,

Saand. J. Clin. Lab. Invest. 18, 380 (1966).

Severinghaus J. W. Stupfel M.
, , and Bradley, A.
, , F. , Accuracy of blood pH and Pcoa
determinations, J. Appl. Physiol. 9^, 189 (1956).

Ladenson, J. H., Tsai L-M. B., Michael, J. M. , Kessler, G. , and Joist, J. H., Serum
,

versus heparinized plasma for eighteen common chemistry tests. Is serum the appro-
priate specimen? AmeT. J. Clin. Path. 62, 545 (1974).

Piatt, B. S., The so-called "first acid change" in freshly shed blood, J. Lab. Clin.
Med. 22, 1115 (1937).

Spring, B. L., Simultaneous determination of carbon dioxide and chloride by the

autoanalyzer, Amer. J. Med. Teoh. 37_, 361 (1971).

Eldridge, F. and Fretwell, L. K. , Change in oxygen tension of shed blood at various

temperatures, J. Appl. Physiol. 20^, 790 (1965).

Fletcher, G. and Barber, J. L., Effect of sampling technique on the determination of

Pa^^ during oxygen breathing, J. Appl. Physiol. 21, 463 (1966).

Severinghaus, J. W. , Measurements of blood gases: P02 and Pcoa, Ann. n.y. Acad.
Soi. 148, 115 (1968).

Kelman, G. R. and Nunn, J. F., Nomograms for correction of blood P02, PCO2, pH, and
base excess for time and temperature, J. Appl. Physiol. 21, 1484 (1966).

Gambino, S. R. Astrup, P., Bates, R. G., Campbell, E. J. M. , Chinard, F, P., Nahas,

,

G. G. , Siggaard-Andersen, 0., and Winters, R. , Report of the Ad Hoc Committee on acid-

base methodology, Amer. J. Clin. Path. 46^, 376 (1966).

Astrup, P. and Siggaard-Andersen, 0., Micromethods for measuring acid-base values of

blood. Adv. Clin. Chem. 6^, 1 (1963).

Gambino, S. R. and Schreiber, H., The measurement of CO2 content with the auto-
analyzer: A comparison with 3 standard methods and a description of a new method
(alkalinization) for preventing loss of CO2 from open cups, Amer. J. Clin. Path. 45,
406 (1966).

Gambino, S.
(1961).
R. , Mineral oil and carbon dioxide, Amer. J. Clin. Path. ~
35, 268

Geubelle, F. Ultramicromethode de mesure du CO2 total du plasma et du sang cutane,

,

Chez le nourrison, Clin. Chim. Acta 2_, 442 (1957).

Skeggs, L. T. , An automatic method for the determination of carbon dioxide in blood

plasma, Amer. J. Clin. Path. 33, 181 (1960).

Friedner, S. and Philipson, A., A method for retaining CO2 in serum in the auto-
analyzer, Soand. J. Clin. Lab. Invest. 17, 185 (1965).

189
Rhodes, P. G. and Moser, K. M. , Sources of error in oxygen tension measurement, J.
Appl. Physiol. 2T_, 729 (1966).

Ishikawa, S., Fornier, A., Borst, C, and Segal, M. S., The effects of air bubbles and
time delay on blood gas analysis, Ann. Allergy 33^, 72 (1974).

Woolf, C, R. , The in vitro and in vivo use of an oxygen microelectrode, J. Lab.

Clin. Med. 69, 853 (1967).

Harrop, G. A., The oxygen consumption of human erythrocytes, Arch. Int. Med. 23,
745 (1919).

Jalavisto, E., Oxygen consumption of blood and plasma and the percentage of reti-
culocytes, Aota Physiol. Soand. 46, 244 (1959).

Runck, A. H., Proper technic in measurement of blood gases, n.e.j.m. 291_, 476 (1974).

Burnett, R. W. and Noonan, D. C, Calculations and correction factors used in de-

termination of blood pH and blood gases, Clin. Chem. 20, 1499 (1974).

Maas, A. H. J., pH determination of body fluids with micro glass electrode and a
saturated KCl bridge in the cell, CUn. Chim. Acta 28, 373 (1970).

Ware, A. G. , Nowack, J., and Westover, L., Capillary blood pH by a dilution technic,
Clin. Chem. 9, 340 (1963).

Strumfjord, J. V., Jr., Determination of blood pH, Std. Methods Clin. Chem. 2^,
107 (1958).

Coester, N., Elliott, J. C, and Luft, U. C, Plasma electrolytes, pH, and ECG during
and after exhaustive exercise, J. Appl. Physiol. 34^, 677 (1973).

King, C. E. and Benz, E. W., Scientific apparatus and laboratory methods: a method
for measuring and recording continuously the pH of the circulating blood. Science
92, 409 (1940).

Golden, G. T. , Litwiller, R. W., and Heironimus, T. W., The interpretation of arterial

blood gases: A concise guide for clinicians, South. Med. J. 66, 1051 (1973).

Siggaard-Andersen, 0., The Acid-Base Status of the Blood, 2nd edition, p. 27 (Williams
and Wilkins Co., Baltimore, 1964).

Rune, S. J. and Lassen, N. A., Diurnal variations in the acid-base balance of blood,
Saand. J. Clin. Lab. Invest. 22, 151 (1968).

Kokkola, K. Relevance of arterial blood gas values obtained during capnography,

,

Soand. J. Clin. Lab. Invest. 34, 43 (1974).

Severinghaus J. W. Weiskopf, R. B., Nishimura, M. and Bradley, A. F.

, , Oxygen ,

electrode errors due to polarographic reduction of halothane, J. Appl. Physiol. 31_,

640 (1971).

190
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975.
Issued June 1977.

INSTRUMENT SPECIFICATIONS

S. Raymond Gambino
Columbia Presbyterian Medical Center
New York, New York 10032, USA

Before I give you my specifications for laboratory instruments I want to show you what
has happened to the number of blood gas assays performed at our hospital from 1970 through
1974 (table 1).

Table 1. Blood gas tallies.

Central Peripheral

Day Evening Night

1970 5,639 9,824

1971 7,034 13,593

1972 8,464 11,794 3,083

1973 7,505 8,812 4,756

1974 8,292 9,732 5,053 13,356

The data in table 1 show that in 1970, we performed a little over 15,000 blood gas
analyses whereas in 1974 we performed 36,433 such assays. More important than the short
doubling rate is the distribution of the assays. They do not concentrate on the day shift
or on weekdays. Instead, the load is rather uniform throughout a 24-hour period, 7 days a
week. In addition, table 1 shows preliminary data for assays performed outside of the main
laboratory. In 1974, we had 2 peripheral blood gas machines located in intensive care units.
Recorded assays performed on these 2 instruments by non-laboratory staff came to 13,356 in
1974. However, the actual number of such assays was probably more than 3 times that number
since most emergency assays are not logged. In 1975, with the availability of instruments
which incorporate automatic washout. Dr. Philip Altman and I have expanded peripheral cov-
erage on the adult service to four locations utilizing Instrumentation Laboratory's 513
series

Given the above data, and the obvious need for "bedside" measurements, the specifica-
tions for a practical clinical blood gas instrument become more obvious. Such an instrument
should:

- Have automatic calibration.

- Have one simple sampling step.
- Be self cleaning.
- Require as small a sample size as possible, preferably less than 200 yl
- Be as small as possible in physical size.
- Consume as little gas and reagent as possible, yet maintain constant calibration and
be ready for instant use at any time.
- Be operable by non- laboratory personnel, including non-technical personnel such as
clerks and nurses' aides.

191
 I have outlined the functional characteristics of my ideal instrument. What about the
required accuracy? What are the maximum error limits tolerable in a peripheral blood gas
unit serving critically ill patients? The following "rough cut" figures (table 2) are based
on my observations of what doctors actually do with the data provided.

Table 2. Tolerable error limits (± 2 standard deviations).

PH + 0.02

PCOg + 3 mmHg at 40 mmHg

POg + 5 mmHg at 100 mmHg

+ 3 mmHg at 60 mmHg

Base + 2 mM/1

Do current instruments meet these minimum standards for accuracy and precision? Do
current operators of instruments meet these standards? The following tables show that there
is more lab-to-lab variation in precision, utilizing the same instrument-model, than there
is between models from different manufacturers.

Table 3 shows data for pH. Note the striking differences in precision for the same
instrument in different laboratories. These data were obtained by Dr. M. Miller of General
Diagnostics utilizing a new ampouled blood gas control which I describe in greater detail in
my other paper at this symposium.

Table 3. pH triplicates on 16 days, different labs different machines.

instrument N Mean %RSD

165 a 48 7.366 5.1

165 b 48 7.389 2.0

BMS 3 48 7.386 2.8

ABLl a 48 7.374 5.5

ABLl b 48 7.336 1 .3

313 a 48 7.361 8.5

313 b 15 7.380 6.0

513 48 7.374 2.2

Labs ' mean 7.376

Manufacturer 7.371 1.4

(RSD calculated on H ion cone. not pH)

192
 Table 4 shows the data for PCO2. We see again the striking differences in precision
(expressed as %RSD or percent relative standard deviation, i.e. S.D. expressed as a percent
of the mean)

Table 4. PCO2 triplicates on 16 days, different labs different machines.

nstrument N Mean %RSD

165 a 48 37 6 7.4

165 b 48 39 5 2.8

BMS 3 48 33 8 6.4

ABLl a 48 35 2 11.5

ABLl b 48 43 8 3.5

313 a 48 39 0 13.2

3.3 b 15 40 1 1.0

513 48 41 3 3.2

Labs ' mean 38 6

Manufacturer 39 9 1 .8

Table 5 shows data for P02. The same story is again evident. There is greater varia-
tion between labs utilizing the same model of an instrument than there is between different
models made by different manufacturers.

Table 5. P02 triplicates on 16 days, different labs different machines.

Instrument N Mean %RSD

165 a 48 120.4 11.3

165 b 48 102.6 3.0

BMS 3 48 109.3 3.2

ABLl a 48 .
126.8 14.9

ABLl b 12 120.8 4.4

313 a 48 109.9 7.9

313 b 15 102.8 1.6

513 48 100.0 1 .6

Labs ' mean 107.5

Manufacturer 104.9 1.4

These data indicate to me that our problem is more a problem of quality control at
individual sites than it is a problem with manufacturers, or with primary standards for pH
and gases.

193
 National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

EFFECTS OF THE LIQUID JUNCTION ON pH MEASUREMENT IN BLOOD;

THE 0.160 MOL/L SODIUM CHLORIDE BRIDGE

A. H. J. Maas

Department of Cardiology and Thoracic Surgery

University Hospital
Cathari jnesingel 101
Utrecht, The Netherlands

1 . Saturated KC1 Bridge

The practical pH-cell is built up of an electrode sensitive to hydrogen ions (usually

the glass electrode) and a reference electrode independent of the test solution (usually a
calomel electrode) with a salt bridge of saturated KC1 which means that buffers and test
,

solutions are applied in cells with a diffusion potential. Consequently the pH of a solution
X is defined operationally by the relationship:

pH(X) = pH(S) + ^
RT £n 10
^ (1)

In most cases the liquid junction potential during measurement of the standard solutions
is not equal to that existing during measurement of the test sample, that is, the residual
liquid junction potential will be unequal to zero, with the result that the measured pH of
the sample will hardly ever be exactly on the conventional pH(S) scale of the National
Bureau of Standards (NBS). The degree of error can be calculated semiquantitatively from
the molar conductivity and the concentration of the ions using Henderson's equation for the
liquid junction potential between two solutions of different composition. With this formula
we calculated some diffusion potentials of the junction: test soln. satd. KCl at 25 °C [1]^. ||

Table 1. Calculated liquid junction potentials and residual liquid junction

potentials of junctions with the saturated KCl bridge (see text).

Liquid junction Residual liquid

potential junction potential

No. Solution X I mV pH diffe- mV pH

units rence units

1. equimolal phosphate buffer 0.100 2.1 0.035 1-2 0.1 0.002

2. blood phosphate buffer 0.100 2.0 0.033 2-3 0.9 0.015

3. 0.160 mol/1 NaCl 0.160 1.1 0.018 2-4 0.8 0.013

4. 0.025 mol/1 NaHCOg + 0.135 mol/1 NaCl 0.160 1.2 0.020 2-5 0.8 0.013

5. 1 iquor cerebrospinal is (CSF) 0.160 1.2 0.020 2-6 0.5 0.008

6. plasma 0.160 1.5 0.025

^Figures in brackets indicate the literature references at the end of this paper.

195
j

1
 : .

This table shows that:

a. A saturated KCl bridge provides junction potentials that are small: those of the
phosphate buffers of NBS are nearly equal as well as those of spinal fluid (CSF) and plasma.

b. The residual liquid junction potential in the pH determination of CSF and plasma
after calibration with the phosphate buffers is about 0.01 pH unit.

Various factors influence the liquid junction potential [1,2]:

a. Composition Saturated KCl solution is mostly used as the bridge solution. Atten-
.

tion ought to be paid to dilution of the bridge solution, which gives lower results. So the
pH of plasma is measured 0.003 pH lower with a 3.5 mol/1 KCl bridge as recommended by Bates
[3]. With a 0.15 mol/1 KCl or 0.16 mol/1 NaCl bridge, as we will discuss further on, values
more than 0.1 pH lower than those using a saturated KCl bridge in the cell are obtained.
Therefore, it is important to recommend the pH determination with the saturated KCl bridge as
a reference method.

b. Structure As demonstrated with the micro glass capillary electrode for over
.

fifteen years, a stable and reproducible liquid junction is obtained when the test solution
in a small plastic tube is dipped into the saturated KCl bridge solution, -i.e., the high
density KCl solution is below the test solution. The opposite arrangement is unstable as
the KCl tends to flow down. Allowance has to be made for this too when permeable membranes,
separating the test solution from the KCl solution, are used at the boundary. As a conse-
quence to such a bad construction, it is necessary to refill the reference electrode in the
Corning M 165 blood gas instrument every day or two.

c. Temperature We determined the change in the residual liquid junction potential

.

with the temperature from the pH difference: pH (ref. 25 °C) - pH (ref. 38 °C) where pH
(ref. 25 °C) and pH (ref. 38 °C) mean the values obtained with the same micro glass electrode
at a constant temperature of 38 °C and two saturated calomel electrodes with a salt bridge,
each at the indicated temperature. Both cells were calibrated with the equimolal phosphate
buffer (pH = 6.840 at 38 °C). For blood, plasma, cerebrospinal fluid and a 0.025 mol/1
sodium bicarbonate + 0.135 mol/1 sodium chloride solution a difference of 0.011 pH was
found; so, we calculate for the temperature coefficient of the residual liquid junction
potential

ApH/At = 0.011/13 = 0.00085 pH/°C.

The fact that this coefficient is equal for the mentioned solutions indicates that dif-
ferences in activity and mobility of the potassium and chloride ions at 25 °C and at 38 °C
are the most important factors. To reduce the influence of this thermal diffusion on the pH
readings, the whole pH cell should be thermostated at one temperature, i.e., 37 °C.

d. Effect of Blood Cells To be able to understand in which order blood cells in-
.

fluence the diffusion potential of the salt bridge, we determined directly the liquid junction
potential differences:

E. (plasma ||
satd. KCl) -E. (plasma with cells ||
satd. KCl),

which effect will from now on be called "cell effect."

We used the following cell:

Cal I, satd. KCl ||

plasma |
suspension of cells in plasma ||
satd. KCl, Cal II

E . E . E
J.P J,pc j,c

If we put ^. equal to zero, the emf difference will give

^ the cell effect:
j,pc ^

E = E TT + E. - E. - E , (E. E 0).
calT II J ,c J ,p calT I ^
J ,pc

196
We determined this effect at 25 °C and 38 °C. Figure 1 shows the cell effect in pH units of
10 experiments plotted as a function of the hematocrit.

0.005+ 0.005+
0
0.000 o" > 0.000
0
0.005- 0.005-
0 20 40 60 80 100

Figure 1. The cell effect at, respectively, the junction with saturated KCl (la and lb)
and with 0.160 NaCl (Ila and lib) as a function of the hematocrit. The different
symbols indicate different blood samples.

Graphs la and lb show the cell effect increasing with the cell concentration. At lower
pH, the cell effect is greater than at higher pH. At pH = 7.4 and normal hematocrit value
(45%), the cell effect is 0.008 pH ± 0.002 (38 °C) and 0.012 pH + 0.002 (25 "C). To be able
to establish whether the cell effect could be due to the negative charge of the erythrocytes
(a so-called suspension effect), we determined in an analogous way the difference in dif-
fusion potential of the bridge with an isotonic solution.

Graphs Ila and lib show the cell effect of the 0.160 mol/1 NaCl bridge plotted as a
function of the hematocrit. We observe that even with hematocrit values of 90 percent, no
cell effect exists. From the experiments we may conclude that the cell effect of the
saturated KCl bridge is not caused by intact blood cells but most likely due to precipitation
of proteins.

2. 0.160 mol/1 NaCl Bridge

In order to avoid this disturbing effect of blood cells on the diffusion potential at
the saturated KCl junction, we investigated whether the saturated KCl calomel electrode
could be substituted by a calomel electrode with an isotonic salt bridge, notably the 0.160
mol/1 NaCl bridge [4]. To that end, we made comparative pH measurements with the glass
electrode-saturated KCl calomel electrode cell and the glass electrode - 0.160 mol/1 NaCl
calomel electrode cell; we calibrated with equimolal phosphate buffer.

197
 The differences pH (satd. KCl ) - pH (0.160 mol/1 NaCl ) were as follows:

Table 2. pH difference of values obtained with the glass

electrode - saturated KCl calomel electrode cell and the glass
electrode - 0.160 mol/1 NaCl calomel electrode cell.

Solution X pH range pH (satd. KCl) - pH

(0.160 mol/1 NaCl)
Blood phosphate buffer 7.391 + 0.005

0.025 mol/1 NaHCOs + 0.135 mol/1 NaCl 7.267 + 0.153

Cerebrospinal fluid ("pooled") 7.486 + 0.163

Plasma 7.453-7.481 + 0.125 ± 0.005

Blood 7.311-7.370 + 0.106 ± 0.005

This table shows that the pH of the blood phosphate buffer is influenced little. On the
other hand, the pH values of body fluids measured with a cell with a 0.160 mol/1 NaCl
bridge, dependent on the test solution, are more than 0.1 pH lower than those determined
using a cell with a saturated KCl bridge. Since the glass electrode, calomel electrode and
standardization for both cells were the same, the pH differences must be interpreted as
differences in residual diffusion potential.

This result will be more clear when we compare some diffusion potentials of the junc-
tion: solution X 0.160 mol/1 NaCl calculated with Henderson's formula.
I
These values are
given in the following table 3:

Table 3. Calculated and measured residual liquid junction

potentials of junctions with 0.160 mol/1 NaCl bridge

Liquid junction Residual liquid-

potential junction potential

calculated measured

No. Solution X I mV pH diffe- mV pH pH

units rence units units

1 equimolal phosphage buffer 0.100 8.1 0.137 1-2 1.1 0.019 0.005

2 blood phosphate buffer 0.100 7.0 0.118 1-3 7.1 0.120 0.153

3 0.025 mol/1 NaHCOs + 0.135 mol/1 NaCl 0.160 1.0 0.017 1-4 6.7 0.114 0.163

4 liquor cerebrospinal is 0.160 1.4 0.023 1-5 4.9 0.085 0.125

5 plasma 0.160 3.2 0.053

This table shows that diffusion potentials of about 8 mV are found with the two phos-
phate buffers and much smaller values with body fluids and bicarbonate solution of I = 0.160.
Further it appears that the calculated residual liquid junction potentials are of the same
size as the measured differences. Just applying a pH cell with an isotonic salt bridge thus
means, in the case of the 0.160 mol/1 NaCl bridge, a shift in the pH scale of more than 0.1
pH with regard to the cell with a saturated KCl bridge, depending on the test solution.
Both cells have to have the same pH scale if they will be useful. So new buffers had to be
developed with NaCl added to limit the liquid junction potentials with regard to 0.160 mol/1
NaCl. The pH value of these NaCl containing buffers can be determined in cells with a
saturated KCl bridge.

198
 Calibration buffers were prepared from the two standard phosphate buffers of the NBS.
So much sodium chloride was added that the pH of solution X, measured with the cell con-
taining the 0.160 mol/1 NaCl bridge became identical with the pH measured with the cell
containing the saturated KCl bridge. In our case, plasma was the solution to be measured.
The following buffers appear to meet these conditions:

Table 4. pH(Ss,l.j.) values of phosphate buffers with NaCl at 25 °C and 38 °C.

Composition of buffers pH{Ss,l.j.) at 25 °C pH(Ss,l.j.) at 38 °C

with NaCl

0.025 mol/1 KH2PO4

0.025 mol/1 NazHPO^ 6.686 6.660
0.130 mol/1 NaCl

0.008695 mol/1 KH2PO4

0.03043 mol/1 NaaHPOi^ 7.255 7.229
0.120 mol/1 NaCl

To agree with the pH(S) scale of NBS, we have attached an operational pH value (pH(Ss,l j ). .

to the standard buffers, which was measured with the hydrogen electrode-saturated KCl calo-
mel electrode cell with regard to one of the NBS buffers. Redetermination by Drinker et al.
[5] resulted in 0.01 pH lower values.

ter

pH, CORNING, MAAS MODE

For checking purposes, the pH's of a number of blood samples were measured and compared
by means of the glass electrode-saturated calomel electrode and the glass electrode - 0.160 M
NaCl calomel electrode cells. From the hematocrit value of the samples that pH determination
with the cell with saturated KCl bridge was corrected for the cell effect. The average
difference between the pH values of 21 blood samples obtained with both cells was -0.001 pH
which, experimentally, was not significant. A glass electrode system with an isotonic salt
bridge as used in this investigation, is completely reliable, and because of its special way
of standardization perfectly equivalent to the conventional cells with a saturated KCl

199
 .

bridge. Further, the pH cell with a 0.160 mol/l NaCl bridge offers the advantages that the
liquid junction potential at the boundary between blood and 0.160 mol/l NaCl is independent
of the concentration of blood cells and that it can be used without danger for continuous pH
measurement in patients. Until now, this bridge was successfully applied by Drinker et al.
using the Corning Model 165 instrument [5].

References

[1] Maas, A. H. J., pH determination of body fluids with a micro glass electrode and a
saturated KCl bridge in the cell, Clin. Chim. Acta 28, 373 (1970a).

[2] Siggard-Andersen , 0., The Aaid-Base Status of the Blood, 4th revised edition, p. 155
(Munksgaard, Copenhagen, 1974).

[3] Bates, R. G., Determination of pH^ Theory and Practice, p. 312 (John Wiley and Sons,
New York, 1973)

[4] Maas, A. H. J., ph determination in body fluids using a cell with an isotonic sodium
chloride bridge, J. Appl. Physiol. 30, 248 (1971).

[5] Drinker, P. A., Noonan, D. C, Ramanai.ah, N., and Tole, J. R., Use of a sodium chloride
phosphate buffer for pH standardization in a new blood gas analyzer with an isotonic
sodium chloride bridge, Clin. Chem. ^9, 1243 (1973).

200
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

THE CRITICAL CARE LABORATORY: A 10-YEAR PERSPECTIVE

Myron B. Laver and Domenic R. Misiano

Anethesia Laboratories
Harvard Medical School
Massachusetts General Hospital
Boston, Massachusetts 02114, USA

Fifteen years have passed since funds were first made available from the National
Institutes of Health to the Department of Anesthesia at Massachusetts General Hospital for
the evaluation of arterial oxygen partial pressure measurements as a clinical method for
early detection of impaired postoperative blood-gas exchange. Progress achieved after a
few years of endless frustration with methodology has been exceedingly rewarding. Although
cause and effect are difficult to establish, one is tempted to conjecture that the birth of
intensive care as a separate entity followed shortly after the arrival of the P02 electrode.
The progress achieved from these early days characterized by endless frustration with
methodology, have been exceedingly rewarding. Acute respiratory failure in the patient
with previously normal lungs once considered a medical curiosity is now recognized as a
national health problem and the routine measurement of arterial blood gases, particularly
Pq^, is an indispensable tool for early recognition, prevention, and therapy of abnormal
lung function. An intensive care unit without access to blood gas measurement seems
inconceivable today. Certainly, if an attempt is made to support patients with mechanical
ventilation, lack of this facility may prove to be disastrous. An understanding of the
growth and present-day problems attendant upon blood gas analyses can be gained by re-
viewing our experience for the past decade at the Massachusetts General Hospital.

In 1965, the importance of access to blood gases and pH for acute care was recognized
by a greater number of our professional colleagues which, combined with the pressures
imposed by growing clinical demands on a research facility, prompted formation of the
Anesthesia Blood Gas Laboratory.

Remarkable foresight was exercised by the hospital administration and director of the
Chemistry Laboratory (Dr. S. Rieder) when they insisted that the nature of the laboratory
demanded supervision and control by physicians and/or personnel participating closely in
the care of critically ill patients.

Figure 1 indicates how the demands upon the laboratory have evolved since its incep-
tion. By late 1975, we will have performed a grand total of 700,000 blood gas analyses,
each one including Pq^, PcO?» P'^- ^^^^ additional responsibilities listed in table
1, the laboratory now provides an average of 1049 analyses daily for 82 patients, all
categorized as being critically ill.

These figures are not presented to impress. Rather, they go to the very heart of the
problem we have gathered to discuss: (1) what are valid criteria for quality control, and
(2) how is standardization of sensors and their calibration for every day use most appro-
priate?

Separation of ours from the main hospital laboratory was considered mandatory if a
quick turn-around time was to be achieved. It was our original contention that a maximum
period of 10 minutes be allowed for the time between arrival of the sample to the lab-
oratory and a return telephone call to the patient's bedside with the answers. I regret to
say that although we are not far away from the original limit, few requests find their way
back within the ideal period. In the early days, problems with electrode technology were

201
 1 20 -
B Total Determinations (Blood Gases)

1965 1966 1967 1968 1969 1970 1971 1972 1973 1974

Figure 1. Yearly number of blood gas analyses performed in the Critical Care
Laboratory of the Massachusetts General Hospital since its inception in 1965.

Table 1. Daily load in the Acute Care Laboratory.

Average number of patients studied/day = 82

Average number of individual analyses/day = 1049

Analyses : PO2 + PCO2 + pH Serum Osmolarity

O2 content Urine Osmolarity

Serum [K"^], [Na"^] Total Protein

Serum [Ca^^] Cardiac Output

Urine [rt, [Na"^]

Hematocrit

Total determinations: 10

Average number of analyses/open heart patients/day = 38

the principle reason. Today, we are overwhelmed by sheer numbers and delay is usually
caused by our administrative inability to process the numerous samples pre- and post-
analysis.
V
The laboratory is located adjacent to the Operating Room, the Recovery Room, the
Respiratory Unit and the Surgical Intensive Care area. Unfortunately, the demand from
other parts of the hospital is increasing as the incidence of acute respiratory distur-
bances is being recognized with greater frequency. Personnel coverage is shown in table 2.
Twenty percent of the total technician time is spent in handling administrative detail
unrelated to the measurements. The sensors have changed little while improvement in
calibration, sample handling, and display of data has been significant (fig. 2).

202
 Table 2. Personnel deployment.

Number of Technician-
Hours Technicians Hours/Day

Monday 7-3 5 40

to 3-11 4 32

Friday n-7 3 24
96

Saturday 7-3 24

and 3-11 24

Sunday 11-7 24
72

Total technician-hours/week = 624

Average number samples/week = 1049 x 7 = 7343

Average number samples/technician-hours = 7343 =11

A technician performs 1 analysis every 5.1 minutes.

AUTOMATED SUBSTATION I

PRINT-OUT
CORE
AUTOMATED SUBSTATION II AUTOMATED SUBSTATION III

LABORATORY
Blood Determinations Technology:
DATA Mass spectrometer
Na*, K*, Ca* COLLECTION Membranes
Pol/electrolyte
Electrodes
Hct

Routine Preventative Maintenance

At 2-4 Hour Intervals

Figure 2. Potential plan for a decentralized automated critical care laboratory.

The substations will be tied in with a core laboratory where the data are
collected and stored. Errors and diagnostics are recorded continuously in the
core laboratory from the substations.

Present problems include: (1) lesser control over individual technician performance,
(2) prolongation of the turn-around time, and (3) excessive loss of personnel hours when
samples are hand carried from distant parts of the hospital.^ Points 2 and 3 suggest that

^ We have calculated that approximately 6 miles are logged daily to hand carry samples
from the bedside or every labyrinth of the hospital to the laboratory. If one includes
the return trip, then the daily total is closer to 12 miles.

203
decentralization within this system must be considered. Together with a diminishing supply
of motivated personnel these problems have emphasized the need for automation which will
permit a short turn-around time assuming no loss of quality control. In order to improve
our perspective for standards of accuracy we will review briefly the logistics that go into
determining performance of an acute care laboratory. They apply equally well when attempting
to improve an existing service (see table 3).

Table 3. Logistics to be considered in establishing a critical care laboratory.

1. Geography

Where is the laboratory to be located in relation to ICU, recovery

room and operating room?

Will the demands of other parts of the hospital be met?

2. Turn-around time

How long after a blood sample is drawn does the answer return to
the bedside?

3. Quality control

("Are the results trustworthy?")

Instrumentation:

1. How often are the delays prompted by equipment

fail ure?

2. Is the service available from the manufacturer

adequate?

Personnel:

1. Is training adequate to recognize and check for

"bizarre" values?

2. Is there sufficient back-up to cover sickness

and vacation?

4. Cost

How much and how often should the patient be charged?

1. Geography

Geography, i.e., location in relation to the areas that need the service most. In the
smaller community hospital, where separation of services is less than extreme, a decision
as to location may be made with ease. In the larger institution such as a regional center
where building plans generally perpetuate architectural and administrative obsolescence,
centralization may be undesirable, if not impossible.

2. Turn-Around Time

Turn-around time, i.e. , the time elapsed between withdrawal of the blood sample and
return of the results to the bedside. If the samples are hand carried by non-physician
and/or non-committed personnel, the possibility of breakage and loss must be considered.
If the laboratory facilities are strained {e.g., numerous analyses requested at one time),
then a delay is inevitable even if the sample arrives promptly in the laboratory.

204
 3. Quality Control

Quality control, i.e., consistent "trustworthiness" is the most difficult goal of all
to achieve. Electrode technology is still an art, not science. In fact, the weak link is
still the membrane-covered O2 cathode. Although we have experienced significant qualitative
improvement in recent years, the overall caliber depends on the experience of the technicians
and their willingness to provide the extra effort necessary for frequent and appropriate
calibration. Unfortunately, the gap between corporate profit motive and desire for ex-
cellence is still wide. Opinion is far from unanimous on the manner of calibration and
many laboratories will modify the manufacturer's instructions to suit local needs. There
is little reason left to continue Pq^ electrode calibration with liquids that provide
readings equal to blood. The error introduced by using gas-equilibrated water, and in some
systems, gas alone introduces errors that are of little clinical importance. Few will
argue that tor clinical purposes, a error of 5 percent above 150 torr is unacceptable
although improvement on this figure is most desirable. Because of the importance of
obtaining a prompt answer, improvement in performance cannot ignore speed. Any device that
does not allow a technician to process 15 samples per hour is unlikely to prove useful.

4. Cost

Cost has received little attention in the past. Our economy no longer permits such
luxuries. Most laboratories have followed the pattern established for other services: one
determination, one charge. If we agree that the acute phase of respiratory failure requires
multiple blood gas analyses at short intervals, then the cost to the critically ill individual
may reach astronomical proportions. One must consider the possibility that excessive cost
will encourage a reduction in tests ordered. In order to prevent this trend, our laboratory
initially instituted the "three unit charge" principle, or a maximum 24-hour charge per
patient calculated on the basis of the cost for 3 analyses. All tests in excess of 3 were
free of charge and physicians were encouraged to take samples as often as required by the
patient's condition. In fact, it is not unusual for a patient in severe respiratory
failure to have arterial blood samples drawn and analyzed at 15 to 30 minute intervals.
Clinical experience suggests that this approach can go a long way toward prevention of
catastrophies.

Because of the growing complexities in accounting, we have now established a fixed per
patient per 24-hour cost with no limitation on the number of analyses performed. It may
appear surprising, but effective administration makes this service quite equitable.

Finally, a comment on in vitro versus the in vivo blood gas monitoring.

When our laboratory first became a hospital service, a commitment had to be made
regarding the type of monitoring most suitable for a large hospital population subject to a
high incidence of respiratory failure. The intermittent route was chosen for four reasons.
First, it permitted better quality control. Second, it allowed all services to take
advantage of this facility. Thus, we did not restrict our services to the intensive care
areas. Third, maintenance and fail-safe technology remained the responsibility of a small
core of trained technicians. Fourth, it was more economical. Nothing that has transpired
so far has given reason for a change in attitude.

5. Future

Reponsibil ities of the Critical Care Laboratory have grown. As a result, its effi-
ciency has suffered. It is fair to say that we now face two problems, both related to
stability: the membrane-covered Pq^ electrode and personnel. The latter can be partially
solved by a reduction in numbers, but not without increased automation. In fact, increased
automation and decentralization are the directions for the future. Ultimately, automated
substations will be located at several sites in the hospital. The problems there will be
the same as we face today and faced ten years ago: formation of clot and electrode drift.
Ancillary methodology requires critical' evaluation. Mass spectrometry, modifications of

205
gas chromatography, dielectric dispersion spectrometry^ deserve serious consideration to
overcome the electrode drift problem. In the meanwhile, blood gas electrodes will be with
us for years to come. The core laboratory will provide around-the-clock preventive
maintenance service with the option of assuming responsibility for analyses if the sub-
station is down. Details of substation design must incorporate the "building block"
approach to facilitate replacement of malfunctioning parts. Although this may appear as an
enormous task, I regard it as most feasible and even deserving of grant support.

It would be helpful to predict that our ability to prevent acute respiratory failure
has advanced sufficiently to obviate an investment into a critical care laboratory. There
is nothing in the natural history of the disease to suggest such a course. In fact, trauma
and advanced heart disease are responsible for the most frequent causes of acute respi-
ratory failure that require the services of a critical care laboratory. Unfortunately,
"inadequate gas exchange" is easy to recognize while "adequate gas exchange" is a difficult
diagnosis to make. Until such time that this statement will be obsolete, it is not likely
that we will be out of business.

Dr. Henry K. Beecher, Professor of Anesthesia, Emeritus, gave the senior author
unflinching support in the early days when blood gas electrodes and respiratory care were
in their infancy; Miss Anna Murphy proved to be a loyal and critical assistant whose
administrative devotion made the Critical Care Laboratory a reality. Finally, a .laboratory
working under pressure can only survive if both interest and morale are maintained at high
level. We have been fortunate to receive superb support from numerous technicians.
Although not usually the recipients of accolades, it is they who have made this endeavor a
success.

2 Dr. A. Michaels, Alza Research, Palo Alto, California, personal communication.

206
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

A THEORETICAL AND PRACTICAL ANALYSIS OF MICROELECTRODE BEHAVIOR:

THE THREE-SHELL MODEL

R. G. Buckles^, H. Heitmann^, and M. B. Laver^

A study was undertaken to examine the behavior of membrane-covered microelectrodes as

predicted by a three-dimensional diffusion model. The following parameters have been
found to be of importance for accurate prediction of electrode performance: (1) cathode
diameter, (2) electrolyte layer thickness and permeability, (3) membrane thickness and
permeability, (4) temperature, and (5) sample description. The model does predict the
previously observed variations in electrode sensitivity, its dependence upon temperature,
and the discrepancy we have noted in readings between gas, water, and blood samples.
The influence of sample viscosity is also predicted accurately. The changes that occur
within the membrane including their influence on overall electrode stability and sensitivity
have been considered. The signal decay observed with the electrode in a microcuvette has
been discussed as an extension of the information provided by the theoretical model. This
has allowed us to consider also the manner in which cuvette design influences electrode
performance.

1. Introduction

Polarographic oxygen electrodes which incorporate a small metal cathode have found
wide applications for biophysical research and are now used routinely in the measurement
of blood gases. When used in vivo, the electrodes allow for the measurement of PO2
in flowing blood, in different tissues, and even at the cellular level [1-5]'*. In vitro
applications include the analysis of oxygen partial pressure in the gas phase, liquids,
and whole blood [2,3,6-8], the measurement of oxygen reaction rates in the Hartridge-
Roughton fast-reaction apparatus [9], and finally, evaluation of tissue metabolic rates
[10]. These applications are possible because the microelectrode consumes very little
oxygen and membranes with high diffusivity for oxygen can be used to cover the electrode
surface. The original purpose in using a covering membrane was to obviate cathode poisoning
due to protein deposition [2]; however, an important function of these membranes is to
isolate the oxygen diffusion field surrounding the cathode [3,4,11].

Proper performance of microelectrodes is dependent upon a stable and reproducible

diffusion field [3,4]. Such a field is generally more difficult to attain then the
potential fields required for thermodynamic electrodes, making the polarographic electrode
more sensitive to artifact. The sensitivity (amps output/mm Hg Pq ) is subject to constant
drift and is not independent of the sample characteristics [12], i.e., variations in
hydrostatic pressure applied during injection of a sample and temperature, cause rapid and
unpredictable variations in sensitivity. For example, the sensitivity for gas will be
different than for liquid samples with the same oxygen tension (P02) [6,11] stirred samples
read differently from stagnant ones [12], and sample viscosity causes variations in
sensitivity [1,13]. As a result of these complications, frequent calibration is mandatory
and, depending on the accuracy desired, the use of tonometered fluids of the same con-

^Present address: ALZA Research, Palo Alto, CA 94304

^Present address: Pius-hospital, 29 Oldenburg, Federal Republic of Germany.
^Present address: Anesthesia Laboratories, Harvard Medical School, Massachusetts General
Hospital, Boston, MA 02114.
"^Figures in brackets indicate "literature references at the end of this paper.

207
sistency as the sample has been recommended [12]. Because the latter approach is im-
possible with implanted electrodes, in vivo results are difficult to quantify [3].

Several mathematical models have been proposed to describe microelectrode behavior

[3,4,11,14-16]. However, none describe the behavior of the electrode even in a well-
characterized in vitro environment. If a model is to be useful, it must predict quantit-
atively electrode sensitivity, the temperature dependence of sensitivity, and the influence
of different membrane characteristics on cathode performance. Both the simple spherical
[15] and linear diffusion models [3] predict a much lower sensitivity [11,12] and a greater
temperature and membrane thickness dependence than observed experimentally [12].

The literature also indicates the electrode behavior depends on the environment in
which it is used. Some authors have described a linear relationship between temperature
and sensitivity [9]; according to others, this relationship is exponential [12].

There are reports that the sensitivity is greater in water than in air [11] and vice
versa [6]. In small cuvettes, the sensitivity in liquid samples decreases with time and
anomalous signals may be recorded during sample injection [12].

Because of inconsistencies in the literature, we reported earlier our studies with

several commercially available cathodes maintained in a wel 1 -characterized environment in
vitro [12]. We now describe a theoretical analysis of electrode behavior with numerical
solution of the three-dimensional diffusion process. The latter is compared to the idealized
models often used in biophysical modeling. The microelectrode serves as a prototype for a
study of transmembrane transport problems and thereby exemplifies the influence of spatial
geometry on the behavior of mathematical models dealing with diffusion.

A. Mathematical analysis

Figure 1 is a schematic presentation of an electrically isolated microelectrode. It

consists of a fine wire (diameter < 0.050 mm) imbedded in an inert matrix (glass or epoxy).

SAMPLE

//////////
/>1EMBRANE// / /
V///////A
ELECTROLYTE
1'.
' y/^ y/^
*
}^
?S<;iNERT<VVV 1

•\/ MATRIX nAXX^ CATHODE

1

i
Figure 1. Simplified O2 cathode geometry, r^ = cathode radius; r = lateral, radial dis-
tance away from cathode center; X = vertical distance from cathode surface; te + t^^
= sum of electrolyte and membrane thicknesses.

while its end remains exposed. The surface of the inert material may or may not be smooth
as shown in the figure; most often it has a finite surface roughness that may reach a
depth of 20 to 30 micrometers. This assures a continuous electrolyte bridge. Surface of

208
the cathode is also rounded but its radius of curvature is so much greater than the
cathode diameter (the latter defines the size of the diffusional field) that the surface
can be considered flat as in the vicinity of the cathode. If we accept the geometric
simplification shown in figure 1, we can characterize the diffusion of oxygen from sampl
to cathode in terms of a cylindrical coordinate system. The sample whose is being
evaluated may be either a stagnant liquid or a well-defined segment of tissue. If the
fluid is stirred, diffusion occurs only through a thin layer between sample bulk and
membrane, where it is assumed that the liquid flows in laminar fashion over the membrane
surface.

The diffusion of oxygen to the cathode is defined by the cylindrical coordinate

system, using the center of the cathode surface as the origin [17]. For a stagnant
sample, the differential equations describing the diffusion of oxygen from sample to
cathode can be expressed as follows:

r 8r
8

\
—
/r^P\
3r
i ^—± -A^M
92p

SX^
=
9P

9t
/

Dm • -
r
• —
9r
9 /9P \

\ 9r/
— s \
+
,
92p

X^
s
9P
m-
9t ^"
(t < X <
"
(t + t)
^

1 9 /r9P \ 92p 9P
D
s \ ,
^=-^e
s
(OlX<tJ
^
r 3r \ 9r / dX^ 9t

where D = diffusivity (cm^/sec)

P = P02 (mm Hg)

M = O2 metabolic rate divided by the O2 solubility coefficient (mm Hg/sec)

e = electrolyte
m = membrane
s = sample
X = vertical distance from cathode

At a large distance away from the cathode, the oxygen gradient disappears {i.e.,
(9P/9r) = (9P/9X) = 0); the gradient at the other boundaries defines the coupling of eqs
(1) through (3). Therefore,

0 (X ^ -)
9X

9P 9P
m e
(r -> 00)

9r 9r

mam I
V
—
9P

gx
]
/
X = t^ + t^
m s

P = P
s m

D
mam
/
—^m
9X/
\ _
= n
^
/

\
— 9X/
e
X = t^
'

P = P
m e

209
 kP D (X - 0, r < rc) (8)
e
( 9X )

e
0 (X = 0, r > r^) (9)

where k = O2 reduction rate at the cathode

r^ = radial distance from cathode center (fig. 4)

a = Bunsen solubility coefficient for O2.

For the initial conditions we may assume that a is a constant concentration of oxygen
present throughout the three regions, and at time zero the electrode is polarized. This
can be analyzed with eqs. (1) through (9) to give the electrode current (i) as a function
of time:

(10)

X = 0

where n = number of electrons generated per O2 molecules reduced

= cathode surface area
F = Faraday's constant

It is commonly assumed that the reaction rate for oxygen at the cathode is infinitely
fast when compared with the diffusion rate (in eq. (8), k «). Under these conditions,
P (x = 0) = 0.0 and the O2 reduction rate is limited by O2 diffusion to the cathode [15].
However, the polarograms shown in figure 2, by the absence of a definite plateau, do not
indicate any region of complete diffusional limitation, except perhaps at low oxygen
tensions. Therefore, the rate of reaction is included for completeness. The equations
presented here do not include initial conditions nor do they consider the fate of oxygen
once it has reacted at the cathode.

If the electrode is polarized at time zero, the recorded current must include also
the current due to ion migration in the electric field; the time required to establish a
stable ionic electric field is greater than the diffusional response time and involves far
greater currents than are attributable to oxygen reduction. Therefore, it is best if we
assume a fully polarized and stabilized electrode whenever eqs. (1) through (9) are used
to characterize cathode performance. The solution to the initial condition of constant
PO2 does indicate the response time to a change in P02 of the polarized electrode.

The "n" term in eq. (10) is normally assumed equal to 4.0. However, there is evidence
that oxygen reduction to hydroxyl ions is not complete on a stoichiometric basis.
Satterfield [18] has shown that H2O2 can diffuse away from the cathode and be converted
back to oxygen. Under these conditions one cannot be certain that exactly 4 electrons
correspond to one mole of oxygen reduction at the cathode (eq. (10)).

Although eqs. (1) through (3) are linear, the analytic solution is not easily obtained.
Therefore, we have utilized the Method of Zones numerical technique for their solution
[19]. Figure 3 presents the results of these computations, where cathode current and
sensitivity are given as a function of time. We have assumed that the electrolyte,
membrane, and sample are in equilibrium at a Pq^ of 100 mm Hg and the electrode reaction
rate is infinitely fast. The sample volume is taken to be 1 x 10-^° ml^ with an oxygen solu

^This sample corresponds to a 20 ym layer (unstirred) with the same radial dimensions as
selected for the membrane and electrolyte (see fig. 1). A larger sample volume is more
complex, mathematically speaking, but would not alter materially the results.

210
bility coefficient of 0.0238 ml {STP)/m1 -atm. The membrane is polyethylene, 25 micrometers
thick with the permeation properties shown in figure 3 [20]. The electrolyte layer
thickness is assumed to be 20 ym (see Sec. 2, Results) and with the same O2 permeation
properties as water,
45r
Instrumentation
Lab (x 10-12)

1 40
\

35-

30
Radiometer I (xlQ-'^)

Radiometer n (xlQ-'^)
25

1
Radiometer m (xlQ-'^)
^ 20 Beckman (xlQ-'')
Radiometer J2 (xlQ-'z)

15

10
0.4 0.6 0.8
\/OLTS

Figure 2. Effect of polarizing voltage on standard sensitivity ^0

(^"'^STAND^ ^
for ambient air of various commercially available oxygen electrodes. Sample:
stirred water at 38 °C. Membrane: polypropylene 0.015 mm.

5.0

NUMERICAL SOLUTIONS, S
o 4.0 -

» 3.0
••STEADY STATE RESULT
I
E
6
OF THREE-SHELL MODEL
CURRENT, I'
/
CONDITIONS
rc = IO^ te =20;i Pe = Ps = 1.115x10"'
tm = 25/i ae =08 = 0238 atm -I

Pm=5.8 X IO'' °m1(STP)-cm om= .0502atm-l

1.0
cm'^ -cm Hg-sec

10 20 30 40 50
TIME, SEC.

Figure 3. Electrode sensitivity (S), or the ratio of current (I) to sample Pq, was plotted
as a function of time. The sample P02 is the average value of the four-sample zones
indicated in figure 4. For symbols see Appendix.

211
 The current decreases rapidly for about 10 seconds, then levels off to decrease at a
constant rate. During the rapidly decreasing portion of the curve, oxygen dissolved in
the electrolyte and membrane (both at P02 = 100 mm Hg) is being consumed. Since oxygen
from both sites diffuses a shorter distance than oxygen from the sample, the rate of
arrival at the cathode surface is greatly increased. As this O2 is consumed, the P02
near the cathode decreases, thereby establishing a gradient from the sample to the cathode.
The current. continues to decrease linearly because the sample size is too small to allow
for a sufficient fraction of dissolved O2 to diffuse to the cathode. Since this rate of
decrease is linear, we must infer that the overall electrode response time (a complex
function of the diffusive properties of electrolyte, membrane, and geometry of the system)
is much less than the rate of oxygen depletion from the sample. Thus, although the sample
PO2 is decreasing continuously the concentration profile within the electrolyte and membrane
readjusts so, rapidly that the ratio of P02 remains constant with time.^ Presence of this
condition is supported by the graph which expresses electrode sensitivity (S) as a function
of time (fig. 3). After the initial transient period, sensitivity appears to be independent
of time (and therefore, Po2)- ^""^s constant for the appearance of a stable sensitivity
is 2.8 seconds, while the predicted time constant for linear diffusion through the membrane
(i.e., t2/Dni) is 7.1 seconds. The fact that the sensitivity does pass through a transient
period is contrary to the predictions of Hudson [14]. The steady state isobars for the
electrode which we described in figure 3, are shown in figure 4. The lines are seen to

Figure 4. Pq, isobars surrounding the cathode under steady state conditions. These
values are By numerical solution of eqs. (1) to (9). This consists of dividing each
zone (electrolyte, membrane, and sample) into four radially-located zones and several
zones in the X direction. 62 flux equations for each zone are derived from eqs. (1) to
(3) and solved simultaneously by using the boundary conditions described in eqs. (4)
and (9). Note in these considerations, the sample P^ has decreased from 100 to 73.8 mm
U2
Hg.

form an approximately spherical surface around the cathode. P02 in the electrolyte at a
radial distance of 90 m, or 9 times the cathode radius (r^ = 10 pm) is less than the
sample Pq while Po„ at the membrane-sample interface is witnin 0.6 percent of the sample
value. computing the P02 gradient in the electrolyte and membrane from these isobars,
one can see that over 98 percent of the oxygen flux comes by diffusion in a radial direction
within the electrolyte {i.e., it arrives by crossing the membrane in the region where
r > r^). If the electrolyte layer were thinner, this flux would not be significant.

^For a more detailed analysis of concentration-relaxation problems, see reference [21],

212
However, because of the large radial flux term, the amount of oxygen reaching the edge of
the cathode is greater than that which reaches its middle. Thus, the electrolyte layer
thickness (ELT) is an important determinant of total flux because an increase in ELT
permits oxygen transport across a larger membrane area and results in a higher measured
electrode sensitivity.

From these studies of diffusion patterns in the peri-electrode space, it became apparent
that a geometrically simpler model would provide a useful tool in our attempt to account for
the influence of ELT on electrode performance. In the spherical model shown in figure 5,

P02-'"

<
re
'
rm

Figure 5. Idealized 3-zone model of the oxygen microelectrode, electrolyte, and membrane
complex. The radial P02 gradient under steady state conditions assumes a significant
resistance in each zone. For symbols see Appendix.

the cathode is treated as a hemisphere surrounded by hemispherical shells of electrolyte

and membrane. The actual cathode radius is used but the current or

is defined in terms of a spherical area on the assumption that oxygen flux density is
uniform over this cathode. An increase in area compensates for the increased flux
density near the outer edge of the cathode as determined in the previous model and, by
increasing the ELT, the effective membrane is also increased, thereby increasing sensi-
tivity. We have used the model only to predict the steady-state electrode sensitivity
because this is most readily shown for a spherical model [15]. The oxygen transfer rate
from sample to membrane surface is characterized by a transfer coefficient which can be
expressed as a function of stirring rate. The same boundary equations apply as given in
eqs. (4), (5), (7), and (8). While the boundary conditions expressed by eq. (6) are
replaced by:

D
mam (— I

y ^^1
= k(P,
^ s
- P„)
m'
r = r^.
m

213
 The mass transfer coefficient, k, is used to allow for the effect of stirring. Also,
since convection is never absent in the cuvette chamber (particularly when one is using
whole blood), use of eq. (6) provides only an approximation. As there are several empiri-
cal correlations available for predicting k, its incorporation into the model does not
seriously limit the model's utility. (See Appendix for a description of one method for
predicting k.)

The steady-state solution for sensitivity (S) in the spherical 3-shell model is given
by:
i 2^nF / r 1 t^ t^
S = (12)
22,400 kr^ P r r P r r
m m m e e e c J

where Pm is the product of D and a, the so-called permeability coefficient. Using the
same constants as were used in the numerical solution presented in figures 3 and 4, the 3-
shell model predicts a sensitivity (S) of 3.38 x 10"^^ amps/mm Hg, while the steady-state
value from figure 3 is 3.45 x 10"^^ amps/mm Hg. If one uses the simpler spherical model
(without an electrolyte layer), the predicted sensitivity will be 8.8 x 10"^^ amps/mm Hg,
while the linear diffusion model [3] predicts a sensitivity of 3.15 x 10-^^ amps/mm Hg.
The 3-shen model appears to agree favorably with the numerical solution, while the simpler
models predict a much lower cathode sensitivity.

Because of the reasonably good agreement we have found between the numerical solution
and the 3-shell model, we elected to use this more easily manipulated model for our evalua-
tion of electrode behavior. Equation (11) indicates the properties of sample, membrane
and electrolyte required to evaluate the sensitivity of the electrode. Information required
include cathode size, membrane thickness, ELT, temperature, membrane and electrolyte
permeability, and finally, sample characteristics. Since ELT is usually not known, all
other information must be available in order to infer its value. The results to be re-
ported indicate that all these parameters are indeed significant.

2. Results

Using data published by others and our own [12], we have calculated: (a) electrode
sensitivity, (b) temperature-dependence of the sensitivity, and (c) dependence of sensi-
tivity on sample type. A technique for evaluation of the influence of electrolyte layer
thickness is also described.

A. Electrode sensitivity

Prediction of electrode sensitivity requires all the information described in eq.

(12). Unfortunately, there are no reliable methods at present for determining the electro-
lyte layer thickness directly, therefore its value must be inferred. Failure to include
the electrolyte layer thickness has resulted in models that do not satisfy the observed
electrode behavior. Figures 6 and 7 indicate the dependence of electrode sensitivity on
electrolyte layer thickness for two different membranes and several membrane thicknesses.
The sensitivity increases with increasing electrolyte layer thickness because the high
electrolyte permeability (table 1) does not limit oxygen diffusion so as to offset the
increased area

^'^ -
f (^CATH ^ ^e ' *m)'^

on the outside of the membrane. For each membrane thickness, several curves are drawn,
the lower one for stagnant water and the upper curve for a gas sample.

In figure 6 we have plotted our experimental data for two values of membrane thickness
on lines calculated according to eq. (12). These lines indicate the values for sensitivity
when the electrodes record P02 in air, stirred water, and stagnant water. For each membrane

214
 0.015 0020 0.025
ELECTROLYTE LAYER THICKNESS (mm)
Figure 6. Electrode sensitivity as a function of electrolyte layer thickness for poly-
propylene membranes of different thicknesses. The curves are calculated with the 3-zone
model, i.e.y eq. (12), based on the following conditions: T = 38 °C; cathode diameter,
0.020 mm; sample, water equilibrated with ambient air.

thickness, the three data points obtained with a microelectrode predict the same electrolyte
layer thickness, irrespective of sample type. The electrolyte thickness so determined for
several electrodes covered with polypropylene fell in the range of 0.020 to 0.030 mm. It is
important to note that a change in sensitivity caused by a change in electrolyte layer thick-
ness does not alter the difference between the readings obtained with gas and stagnant liquid
sample.

Similar data have plotted for two different thicknesses of Teflon FEP membranes at
38 and 25 °C (fig. 7). The derived range of electrolyte thickness is 0.0055 to 0.007 mm.
In figure 8 we have indicated the sensitivity dependence on cathode size for a 0.015 mm
polypropylene membrane at 38 °C using the derived electrolyte thickness of 0.025 mm.

In an attempt to measure directly the influence of electrolyte layer thickness on

electrode sensitivity, we constructed the device (i.e., electrode holder) shown in figure 9.
The micrometer pe'^mits adjustment of the cathode tip distance (any commercially available
microelectrode can be used) relative to the fixed membrane in steps of several micro-
meters. The hydrostatic pressure of the electrolyte is maintained constant by an overflow
hole. A sequence of runs were made by suspending the electrode in a thermostatted bottle
filled with water, stirred by a magnetic bar at 300 rpm, and through which we bubbled a
gas of known oxygen concentration. The cathode was positioned at about 0.75 mm above the
membrane and a steady state reading obtained. The micrometer was then turned to the next
unit (0.032 mm/mark) and the new sensitivity determined. This procedure was continued
until the membrane appeared to be visibly distended.

Figure 10 presents the results of these measurements. Two extreme cases may be
estimated by simplified models and the theoretical maximum and minimum sensitivity com-
puted. Maximum sensitivity (S|vj/\x) must occur at an electrolyte layer thickness where the
membrane no longer influences the diffusion field. In this case, the electrolyte will be
equilibrated with the sample P02 and the spherical-diffusion-in-a-homogeneous-medium model

215
 0 0.005 0.010 0.015 0.020
ELECTROLYTE LAYER THICKNESS (mm)

Figure 7. Electrode sensitivity as a function of electrolyte layer thickness for dif-

ferent thicknesses of Teflon FEP membranes. The curves calculated with the 3-zone model
based on the following conditions: T = 38 or 25 °C, as indicated; cathode diameter,
0.020 mm; sample, water in equilibrium with ambient air, stirred or stagnant; membrane
thicknesses from 0.0125 to 0.050 mm as indicated. Experimental data were obtained from
Heitmann et at. [12]. Note that Teflon FEP (0.050 mm) exhibited the smallest theoreti-
cal difference between stirred and stagnant water; experimentally no significant dif-
ference was found between these two samples.

I
6.0

4.0

te =0.025 mm
2.0
T = sa-c
tn)=O.OI5mm (Polypropylene)
Sample: Gas
/ 1 1 1 1

0.005 0.010 0.015 0.020 0.025

CATHODE RAD/ US (r^J./nm

Figure 8. Dependence of electrode sensitivity on cathode size under the conditions stated.

should apply [15]. This value is plotted in figure 10 and is about the same as the maxi-
mum we have obtained experimentally. Minimum sensitivity (S^in) will be obtained when the
spherical diffusion flux is totally restricted to the membrane; i.e., when the electrolyte

216
Figure Cuvette-electrode assembly which permits a change in electrolyte layer thick-
9.
ness. Electrode-micrometer assembly which allows for an accurate displacement of the
A.
electrode away or toward the membrane. B. Housing. The electrolyte is contained
within the lower half of the plexiglas chamber. C. Membrane holder plate. D. Assembled
electrode.

0 minimum, m m+160 m+320 m + 480 oo

ELECTROLYTE LAYER THICKNESS, fg (microns)

Figure 10. Influence of electrolyte layer thickness (ELT)(tg) on electrode sensitivity

(S). Maximum sensitivity (Sm/\x) is recorded at right (®) at an electrolyte layer
thickness which is large enough so that the membrane no longer influences the diffusion
field. Point at left (®) indicates the theoretical minimum sensitivity (S^in). At
Sm/\x» behavior of the electrode is membrane- independent and ELT dependent; conversely,
at Smii\| electrode behavior is limited principally by membrane diffusivity and solubility
for O2 and less by these properties in the electrolyte.

217
layer thickness is zero. The theoretical value, also plotted in figure 10, is much lower
than the actually determined minimum sensitivity. The asymptotic approach to the experi-
mental minimum sensitivity suggests that one can never reach the level of a zero electrolyte
layer thickness. These results suggest that the electrolyte layer thickness in respon-
sible for the changes in sensitivity of microelectrodes over the span of theoretical
values used.

B. Temperature dependence of electrode sensitivity

According to eq. (12), three temperature-dependent parameters determine electrode

sensitivity. They are: permeability of the electrolyte, membrane, and sample. It can be
shown that one need only consider behavior of the electrolyte layer and membrane in order
to determine the true electrode temperature sensitivity. Both of these parameters are
given by the general form:

p T ^ p
-AE/RT (13)
i e

where P and Ep are obtained from table 1. Within the limits of 0 to 50 °C, one can also
express the temperature dependence of electrode sensitivity in the following form:

6/RT
(14)

where Pi is obtained from table 1. Within the limits of 0 to 50 °C, one can also
tionality. The rate of change of sT with respect to the temperature, T, is not constant;
rather, it decreases with increasing temperature. Thus, it is preferable to report the
temperature sensitivity in terms of the constant B.

Table 1. Oxygen permeability data.^

Manufacturer and Permeabil ity

Membrane Description Reference
(Pi)

Polypropyl ene Union Carbide (50% S./xlO-Vl^^OO/f^T 26

crystal 1 ine)

Polyethyl ene Dupont (branched) 8.9xlO-3e-9900/RT 26

Polyethylene Dupont (linear) 2.3xl0-4e-8S0°/«T 26

Polypropyl ene Vorschein Fol ien-Fabrik^ 1.7X10-V8090/RT 21

(W. Germany)

Teflon FEP Dupont 9.5xl0-^e-5700/RT 29

Teflon TFE Dilectrix 14.6x10"^° (38 °C) 4

Sil icone rubber Dow Corning 21

Water 4.7xlO-^^e^3140/RT 30

Mylar Mylar 5.78x10"^^ (37 °C) 7

^Because of manufacturing variations, one should consult the manufacturer for accurate
permeability data, if not using any of the sources listed.
^Expressed as (cm^ (STP) • cm)/cm2 . ggc • cm Hg).

^Marked with Radiometer P^ electrodes.

U2

218
 The influence of electrolyte layer thickness on the ratio of sensitivities at 25 and
38 °C is shown in figure 11. The inferred electrolyte thickness for polpropylene is 0.021

0.9

Teflon FEP
tm O.OI25mm
=
/
0.8
o
/
k.
/
Exp. Range*

0.7 /
/
Exp. Range

«0
I
0.6 Polypropylene
tfn=0.020 mm

Tcoth " 0 010 mm

0.5- *Heitmann et al, 1967

0 0.010 0.020 0.030

ELECTROLYTE LAYER THICKNESS, mm
Figure 11. Influence of electrolyte layer thickness (ELT) on electrode temperature de-
pendence. The experimental ranges indicated were obtained from Heitmann et at. [12].
The two curves were calcualted from eq. (12) using the data presented in table 1.

to 0.029 mm and that for Teflon FEP is 0.007 to 0.008 mm. These ranges agree remarkably
well with those inferred on the basis of absolute electrode sensitivity.

Some authors have reported a much higher electrode temperature dependence, often
approaching that of the membrane [22]. However, the cathodes used were large (macro-
electrodes), in which case diffusion is mainly linear. In such cases, the effect of
electrolyte temperature-dependence is not significant, but does not imply that the electro-
lyte layer is any thinner.

C. Influence of sample condition and characteristics on electrode sensitivity

As indicated by figures 6 and 7, the influence of sample stirring and the differences
in electrode sensitivity between gas and liquid samples were predicted satisfactorily with
only one value for electrolyte layer thickness. However, this is not a very sensitive
test of the 3-zone model. Measurement of in a stagnant liquid sample within a cuvette
results in the greatest reduction of sensitivity of any sample media; let us examine this
case further.

Variations in sample viscosity influence both the diffusion coefficient and the
solubility coefficient of the sample [23]. By taking these influences into account, and a
particular set of electrode conditions, one should be able to predict the influence of

219
sample viscosity on electrode sensitivity. Figure 12 presents the results of two such
calculations, one for 0.00625 mm Teflon FEP and the other for 0.020 mm polypropylene.

.00

""cath
• 0.010 mm
T =3800
Stagnant Sample
0.96 —
to
Polypropylene
tm = 0.020 mm
0.92 te = 0.025 mm

Teflon FEP
J: 0.89 tm = 0.00|2 mm
te =0.0065 mm

0.84

O 1.0 2.0 3.0 4.0

VISCOSITY OF GL YCEROL - WA TER
SOLUTION, cps
Figure 12. The influence of sample viscosity on electrode sensitivity. The experimental
data*, taken from Heitman et at. [12], while the two curves (solid and dashed lines)
were calculated from the data of table 1.

These two curves agree within 1 percent of the experimentally determined points. When the
sample is whole blood, many other parameters besides viscosity must be considered {e.g.,
variations in "permeability," convective settling, etc.). For this reason, calibration of
the electrode with a solution of viscosity identical to the blood sample may reduce the
discrepancy between blood and gas but not improve the consistency of the results. Rather,
one must empirically determine, for a given electrode and cuvette, the solution whose
viscosity results in the same relative reading as blood of a given hematocrit.

3. Discussion

We have presented a theoretical 3-zone model which predicts O2 cathode behavior by

taking into account membrane characteristics, electrolyte layer thickness, and finally,
cathode diameter. Variation in sensitivity with cathode size has been reported to be
nearly linear by Fatt [11], but this author assumed, on the basis of his model, that the
findings represented anomalous electrode behavior. This is not at variance with our data
(fig. 8). At a higher cathode size, the dependence does become of second order as pre-
dicted by linear diffusion models. As a matter of fact, any model that does not take into
account the electrolyte layer thickness will predict a second order relationship between
sensitivity and cathode size.

Staub [9] reported a microelectrode sensitivity of about 0.6 x lO"'^^ amp/mm Hg for a
0.050 mm diameter cathode, a value much lower than predicted by the 3-zone model for tlji^
thinnest available Teflon TFE membrane (0.00625 mm; sensitivity approximately 10 x 10"
amp/mm Hg). This implies that when »
t^ linear diffusion limits the electrode sensi-
tivity.

The possible influence of other parameters on the behavior of microelectrodes does

require consideration.

220
 A. Membrane characteristics

The membrane serves two functions: (a) it is the medium in which the major oxygen
gradient is established, and (b) it is the electrical isolator for the cathode. Poly-
propylene is the most commonly used membrane for static measurements. Others include
Teflon'' (both FEP and TFE), silicone rubber, and Mylar^. Ideally, the oxygen cathode-
membrane system should exhibit no change in sensitivity with time in order to reduce the
requirements for frequent calibration. Unfortunately, none of the membranes presently
available fulfill these specifications, and we must consider their properties in more
detail in order to anticipate their influence on cathode characteristics.

The above mentioned membranes are not truly homogeneous. Silicone rubber is always
reinforced with silica particles (approximately 25 percent by weight). The other mem-
branes have both crystalline and amorphous regions. Michaels and Bixler [24] have devel-
oped a model for permeation of gases through such membranes based on the concept that
crystalline regions are impermeable to gases and act as an obstruction in an otherwise
unmodified membrane. On this basis, permeability of the membrane can be expressed by the
following equation:

= -^^P/'^T (15)
p
'^m ^ e

where (t)Pg = P from eq. (13).

The factor is generally a linear function of percent crystal 1 inity.

cj) The latter
depends on the manufacturing process, thermal treatment, and degree of stretching. Recent
studies with polypropylene have shown a three-fold variation in permeability induced by
mechanical stress [25]. These same studies indicated that stretched membranes will gradu-
ally "relax", resulting in a gradual increase in permeability subsequent to the initial
stretching.

This phenomenon probably explains the cathode "aging" patterns which have been observed.
The initial stretch which results when the membrane is first applied reduces permeability,
but the subsequent stress-relaxation increases permeability, and therefore increases
sensitivity with aging. However, there is an alternative explanation. Teflon-covered
electrodes show a greater increase in sensitivity with "aging" than polypropylene [12].
Since stress-relaxation is accomplished by an extended deformation of the membrane, the
increased sensitivity may be the result of a thicker electrolyte layer. To check this
hypothesis, we examined the electrode temperature coefficient, 3, on three succeeding
days. As the sensitivity increased, g decreased, indicating a thicker electrolyte layer.
Both explanations depend on the stress-relaxation phenomena, common to all the presently
used membranes. At this time, there are insufficient data to decide between these two
mechanisms.

The observation that polypropylene membranes produce electrolyte thicknesses of 0.020

to 0.030 mm has two important effects on electrode performance: (1) hydrodynamic drag can
certainly be expected to effect the electrolyte thickness, and (2) the temperature sensiti-
vity cannot be accurately predicted for a given electrode. Although this second fact is
of great importance when one is considering the use of thermistors for automatic tempera-
ture compensation in oxygen electrodes, it is of little consequence in thermostated systems
used routinely for static blood-gas measurements.

Table 1 presents the available permeability data for membranes commonly used with
oxygen cathodes. The data from Buckles [21] were evaluated with polypropylene membranes
available with one type cathode (Radiometer) and are distinctly different from the data of
Myers, Stannett, and Szwarc [26]. This difference is undoubtedly due to variations in
crystalline content [27].

Stiffness of the membrane determines its fit on the cathode surface. The flexible
materials {i.e.^ Teflon FEP) can be expected to conform closely to the shape of the
cathode tip and to fold more smoothly over the outer cathode edges. The degree of stress-

Trademarks of E. I. DuPont de Nemour Company.

221
relaxation, or "creep", is a function of the initial stretch; because of its lower "tough-
ness". Teflon will exhibit the greater amount of stress-relaxation (for more detailed
discussion of the mechanical properties of polymers, see ref. [3]).

B. Cathode geometry

Shape of the cathode tip has a pervasive influence on cathode sensitivity and stability.
Since we are considering cathodes with diameters less than 0.1 mm, the radius of curvature
is always much greater than the cathode radius. Yet in a previous section, we have shown
that the surface need not be sharply pointed to result in a near-spherical diffusion
pattern. Attempts at reducing the radius of curvature of the electrode tip to increase
"spherical" diffusion is virtually impossible with microelectrodes. Therefore, even
pointed electrodes are still "flat" in the region where diffusion occurs. A pointed
electrode tip will result in an increased stress to the applied membrane; as a result,
cathode instability is prominent.

C. Cuvette design

The microelectrode has been developed to eliminate the mechanical inconvenience of

stirring and because it is desirable to use small blood samples without depleting its
oxygen electrodes and their cuvettes have varied considerably and, as expected, the
results have been inconsistent. According to Fatt [11], water flowing through a cuvette
resulted in a signal deviation of 20 percent as compared with the stagnant sample. On the
other hand, stirring did not result in any variation of the signal from that of the
stagnant sample. Explanation for this disagreement was not given. If a sample flows past
an electrode in one direction, hydrodynamic drag occurs in addition to surface renewal of
sample. The hydrodynamic drag is not present when the sample is stirred symmetrically
within the cuvette. Hydrodynamic drag is mainly dependent on the linear velocity across
the membrane surface and sample viscosity (the linear velocity can be approximated by:
electrode diameter x sample flow rate, ml /sec/cuvette volume in ml); it influences the
electrode signal by modifying the membrane position relative to the cathode and by al-
tering the electrolyte layer thickness.

The cuvette design also influences the long-term stability of the signal from a
stagnant sample. Addition of a sample to the cuvette, previously equilibrated with a
sample of lower P02 generally results in the current rising swiftly to a peak signal, then
slowly dropping off. If the cuvette is a thin, dish-shaped space, the capillary forces
will tend to hold the sample rigidly so that oxygen transfer from sample to membrane
occurs only by diffusion. Thus, once the oxygen is consumed around the cathode, oxygen
will have to diffuse farther to reach the cathode, resulting in a lower signal. In terms
of the 3-zone model described above, one may say that

t/Y
k = k„e (16)
0

where t = time in seconds.

The time constant, y, will depend on the cuvette design, being largest when the
spacing is greatest. When the sample is not contained by capillary forces, natural con-
vection will enhance simple diffusion of oxygen.

4. Appendix

A,- area normal to flux, j^, at a surface i (cm^)

C(r) oxygen concentration, [ml (STP)/ml] at a radius r

D, oxygen diffusion coefficient of the material i (cm^/sec)

F Faraday units, (96,500 amp-sec/mol)

P, oxygen permeability of material i, expressed as

222
 ml (STP) - cm
(17)
cm^-sec-mm Hg

R = gas content (cal/g mol-K)

= electrode sensitivity i.e.,

[current for P^, of ambient air] - [current when P^, = 0]

U2 U2

Pn (mm Hg) for ambient air

U2

T = temperature (kelvin)

U^. and V^. = constants of integration

j-j
- oxygen flux rate, ml (STP)/sec, across surface i

T = PA U2-I
•^i (18)

"i

k* = sample mass transfer coefficient, [ml (STP)/cm2 - mm Hg - sec]

*The mass transfer coefficient to a spherical particle of radius can be estimated

as a function of the dimensionless velocity in a variety of ways [28]. However, at
low Peclet Numbers (below 1.0 (see ref. [24]), the dimensionless transfer coefficient,
"k"/D, reaches a limiting value of 1.0. Thus,

k = D a /r^ X 760. (19)

m m m

When stirring occurs, one generally finds that the ratio of the sensitivity with
stirring to the sensitivity in a stagnant sample for an electrode immersed in electro-
lyte without a membrane, is 2:1 [12]. Thus, we may safely say that the k to be used
when complete
^
stirring^ occurs is k = 2P /r„ x 760.
m m
"k" = mass transfer coefficient to spherical particles, mol/mol sec
n = number of electrons that react to reduce one mol of O2
n^. = vector normal to surface A^.

r = radius (cm)

r„ = r^ + t„
e c e
r = r + t + t
m c e m
t = thickness
a = Bunsen solubility coefficient [ml (STP)/ml-atm]
e = effective activation energy for oxygen diffusion in a micro-Clark-type oxygen
electrode (cal/g mol
aE = activation energy of permeation (cal/g mol)

Y = time constant for oxygen diffusion in a stagnant sample

(J)
= permeation factor; reflects dependence of permeation on crystalline content

Subscripts:
m = membrane
e = electrolyte
s = sample
c = cathode
O2 = oxygen

223
 This investigation was supported by U.S. Public Health Service Grant GM 15904-08.

References

[I] Charlton, G., A microelectrode for determination of dissolved oxygen in tissue, J.

Appl. Physiol. 16_, 729 (1961 ).

[2] Clark, L. C, Jr., Monitor and control of blood and tissue O2 tensions. Trans. Amer.
Soo. Art. Internal Organs 2^, 41 (1956).

[3] Davies, R. W. The oxygen cathode, in Physical Technique in Biological Research^ Vol
,

p. 137 (Academic Press, New York, 1962).

[4] Grangsjo, F., and Ulfendahl, H. R., Factors influencing the properties of electrodes
for the continuous measurement of oxygen tension in tissues. Acta Soa. Med. Upsal.
67, 107 (1962).

[5] Kreuzer, F., Harris, E. 0., Jr., and Nessler, C. G., Jr., A method for continuous
recording in vivo of blood oxygen tension, J. Appl. Physiol. ]5^, 77 (1960).

[6] Laver, M. B,, and Seifen, A., Measurement of blood oxygen tension in anesthesia.
Anesthesiology 26 73 (1965).
,

[7] Polgar, G., and Forster, R. E., Measurement of O2 tension in unstirred blood with a
platinum electrode, J. Appl. Physiol. jl_5, 706 C1960).

[8] Silver, I. A., A simple micro-cathode for measuring ?q in gas or fluid, Med. Elec.
Biol. Eng. 547 0963).

[9] Staub, N. C, A simple small oxygen electrode, J. Appl. Physiol. 16, 192 0961).

[10] Hospodka, J., Caslansky, Z., Beran, K., and Stross, F., Continued Cultivation of
Microorganisms, 2nd Symposium, Prague, p. 353 (1962).

[II] Fatt, I., An ultramicro oxygen electrode, J. Appl. Physiol. 1^, 326 (1964).

[12] Heitmann, H., Buckles, R. G., and Laver, M. B., Blood Pq^ measurements: performance
of microelectrodes, Resp. Physiol. _3, 380 (1967).

[13] Tsao, M. v., and Vadnay, A., An electrode for continuous measurement of transient
blood P02 in the vessel, J. Appl. Physiol. 15^, 712 C1960).

[14] Hudson, J. A., Measurement of oxygen tension by the oxygen cathode, Med. and Biol.
Engng. 5, 207 (1967).

[15] Kolthoff, I. M., and Lingane, J. J., Polarography , Vols. I and II (New York,
Interscience, 1952).

[16] Schuler, F., and Kreuzer, F., Properties and performance of membrane-covered rapid
polarographic oxygen catheter electrodes for continuous oxygen recording in vivo.
Prog. Resp. Res. 64 (1969).

[17] Sherwood, T. and Satterfield, C, The Role of Diffusion in Catalysis (Addison-

K. ,

Wesley Publishing Co., Reading, Mass., 1963).

[18] Satterfield, C, Supersaturation of oxygen in aqueous hydrogen peroxide solution, J.

Chem. Eng. Data 504 (1961).

[19] Strong, P. F., and Emslie, A. G., The method of zones for the calculation of
tempera
ture distributions, report to JPL, Arthur D. Little, Inc., April 1963.

224
[20] Bixler, H. J., Solution and Flow of Gases in Polyethylene, Sci . D. Thesis, Massachusetts
Institute of Technology, 1959.

[21] Buckles, R. G., Ph.D. Thesis, Department of Chemical Engineering, Massachusetts

Institute of Technology, 1966.

[22] Carey, F. G., and Teal, J. M., Responses of oxygen electrodes to variables in
construction, assembly, and use, J. Appl. Physiol. 20, 1074 (1965).

[23] Reid, R. C, and Sherwood, T. K. , The Properties of Gases and Liquids (McGraw-Hill
Book Co., Inc., New York, 1958).

[24] Michaels, A., S., and Bixler, H. J., Solubility of gases in polyethylene, J. Polymer
Sci. 50, 393 (1961).

[25] Krewinghaus, B., Sc. D. Thesis, Department of Chemical Engineering, Massachusetts

Institute of Technology, 1966.

[26] Myers, A. W., Stannett, V., and Szwarc, M., The permeability of polyethylene to gases
and vapors, J. Polymer Sci. 35^, 285 (1959).

[27] Conner, W.P., and Scherts, 6. L., Effect of density and orientation on the per-
meability of polypropylene film, SPE Transaations 2^, 186 (1964).

[28] Levitch, V. G. , Physioaoahemiaal Hydrodynamics [Pi^entice-Hal 1 , Inc., Englewood Cliffs,

N.J., 1962).

[29] DuPont Bulletin T-3A, Teflon FER— Chemical Properties.

[30] Hodgman, C. D., Ed., Handbook of Chemistry and Physios (The Chemical Rubber Publishing
Co., Cleveland, 1963).

225
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

DYNAMIC RESPONSE OF A pCOa ELECTRODE

S. J. Pace and M. J. D. Brand^

Technicon Instruments Corporation
Tarrytown, NY 10591, USA

1. Introduction

Gas sensing membrane electrodes are almost universally used for the measurement of the
partial pressures of oxygen and carbon dioxide in blood. The basic operating principles
have remained unchanged since the introduction of the PO2 electrode by Clark [l]^ and the
PCO2 by Severinghaus and Bradley [2], In each electrode the sample is separated from an
electrolyte by a gas permeable membrane which allows transport of O2 and/or CO2. The dif-
fusing gas is sensed in the electrolyte solution by: (a) reduction at a platinum or other
noble metal cathode for O2, or (b) measuring the change in pH in a buffer solution with a
glass electrode for CO2.

It is remarkable that despite the widespread use of these electrodes for nearly 20
years there have been few attempts to describe their response theoretically. Lucero [3] has
considered some factors which affect the design of PO2 electrodes from a practical stand-
point. It was not until 1974 that an analysis was made of the oxygen profiles around a
membrane covered PO2 electrode [4]. Theory of the PCO2 electrode was limited to the approxi-
mate equilibrium model given by Severinghaus [2] until a contemporary treatment by Ross et
al. [5]. This model considered only the mass transport through the membrane. In the parti-
cular case of PCO2 electrode, the kinetics of the reaction between CO2 and water in the
electrolyte solution can be rate limiting.

In this paper, a model for a PCO2 electrode is proposed which considers mass transport
in the membrane and in the electrolyte solution as well as chemical kinetics.

2. Mass Transport of CO2

This model applies to a gas sensing electrode having the geometrical configuration
shown in figure 1. The response of this electrode is determined by the concentration of gas
(CO2) as a function of time and distance from the ion detector. Once the concentration
profile is established through both the membrane and the internal electrolyte phases, the
CO2 concentration may then be applied to the rate law governing the reaction:

fast
C02(g) + H20(ii)<=^ H2C03(£) ^ H + HCOi

^a

The homogeneous reaction is assumed to be rate-limiting and even though the molecu-
larity of reaction (1) is second prder, the solvent mole fraction is very nearly one {i.e.,
C^^Q>>C^Q^) resulting in an overall pseudo-first order reaction. The experimental basis for
this assumption is the reported [6] decreased response time of the CO2 gas-sensing electrode
when carbonic anhydrase is incorporated in the internal electrolyte {i.e., the enzyme catalyzes

^Present address: Department of Chemistry, Texas A&M University, College Station, TX 77843^
^Figures in brackets indicate the literature references at the end of this paper.

227
 EMF

E— pH Electrode Assembly
R — Internal Electrolyte Compartment
M — Membrane or Air Gap Compartment
S — Sample

Ref Electrode

pH Electrode-
0

Gas Permeable
M Membrane

Sample Out

Sample In

Figure 1. Geometric configuration of gas sensing electrode.

reaction (1)). Furthermore, the high sample concentration limit rarely exceeds 10"^ molar,
representing a solvent to solute mole ratio of several orders of magnitude. The concentra-
tion of CO2 in the sample is assumed to be invariant; this is equivalent to assuming that
the sample is buffered in CO2 and is stirred.

Consider the concentration gradient through the membrane and electrolyte media shown
in figure 2. A steady state concentration of gas A (CO2) is maintained in the sample
compartment S, The membrane (M) in this case is simply an air gap of mostly nitrogen gas
(B). The air gap may be maintained by a thin gas-permeable membrane delineating the
sample/air gap interface (although the entire compartment M may be composed of a gas-
permeable membrane, e.g., silicone rubber). At the gas-liquid interphase S/M, which is
denoted as X = 0, and through the membrane phase M bound at X = L, the gas phase concentration
(C/\) is expressed in moles/liter or as a mole fraction N/\. Na may, in turn, be defined as
the partial pressure of A divided by the total pressure, PA/Ptot' provided A and B form an
ideal gas mixture. Gas B is inert to both the sample liquid and gas A and the partial
pressures of both gases are in equilibrium with the liquid S,

At X m corresponding
... ^ ^ to the gas-internal electrolyte interphase
, (M/R), the con- ,

"
centration is" Zp^^. "ntire system is controlled at constant temperature and pressure
The entire pressur
The interface R/E at X L separates the internal electrolyte from the pH glass electrode
surface.
The boundary conditions for this mass transport problem are

X = 0 c, = C,^

X = m "
^A ^A

\3x
/x^L

The condition at X = L states that no H3O diffuses through the pH glass membrane. This is

228
(a)

t
Ca

(a) Low-to-high
concentration change
M
Gas A (b) High-to-low
concentration change
Gas B
S — Sample
m L M — Membrane and Air Gap
R — Internal Electrolyte
E — pH sensing Electrode
(b)
t
Ca
Ca.

Gas A M
Figure 2. Concentration profiles
Gas B at electrode-sample interface.
Ca;

m L X

a reasonable assumption since the diffusion coefficient of HsO"*" through glass is much
smaller than the diffusion coefficients of A in B (Dy^g) and A in R (D^l)-

The electrode is subjected to an instantaneous step concentration change from C^' to .

0 0
The resulting concentration gradient is the driving force for molecular diffusion of A
into M and subsequently into R. Solving the mass balance equation through region M,
one obtains the following equations:

(1)
\ m 0 ' c

or

x/m

(Ptot - Pa
J \Ptot - Pa^

The flux at the gas-liquid interface M/R is:

1 - P,
°AB Ptot Dab/^^
In
(^A)x=m = -
IT (\ -
^A^) m 1 - P,
(2)

As gas A diffuses through medium R, it is chemically transformed to compound Q (H2CO3) by

229
reaction (1). The total mass transport equation including homogeneous chemical kinetics is:

- (ki + - k2 (C^^ - = 0 (3)

CqJ

where C„ and Cp, are the initial concentrations of A and Q. The solution to the above equation
% ^0
for the above boundary conditions defines the concentration profile through region R:

- _ m_

where

ki + k2 k2

and

a\l ^ a^^ (2L - m)

Y = e + e

The flux at X = m is;

(5)

From eqs. (2) and (5) an expression for C. is derived;

m
C

m ( p- ) a ? - 1

where

(2L
(6)

The removal of geometrical dependence on the ion sensing element in the internal elec-
trolyte can be advantageous for practical applications. This could be accomplished by

230
designing a sensor so as to make it responsive to an average ion concentration as, for example,
a cylindrical pH sensor enveloping the internal electrolyte compartment. The concentration
of A may be averaged throughout the entire volume of compartment R resulting in the following
relationship:

dx
avg
m m

(7)
avg.
(L - m)a^ (L - m)

3. Time Response

Equations (1), (2), (4), and (5) accurately describe within the confines of the model
the steady state concentration of CO2 (gas A) and the rate of mass transport. However, no
information about the response time of the electrode can be derived from the above treatment
thus far. Since the CO2 gas sensing electrode is slow responding due to a rate limiting
chemical reaction, eq, (4) may be applied for the appropriate distance X (at X = L in this
case) to the rate law for reaction (1). An expression for the time response is obtained
by solving the first order rate equation for reaction Cl)-

dC,
(8)

The solution of eq. (8) yields an expression for the time required to reach the equili-
brium concentration C. at distance X = L:

r C
0 e
In (ki + k2)t

The following mass action law can then be applied:

+ l<i[C02]
(9)
kzCHCOg]

to describe the time to reach a given pH change as;

t = (ki + k2)"^ In
X% ' ^
(% ^(HCOP - ^H^ ^(HCOj)
J (10)
^b(V -
^Hp,^ ^(HCOs)^ "
^H^ ^(HCOi)^

where Cg is the concentration of internal electrolyte (usually NaHCOs) and C^^q- is that

231
formed from CO2. Equation (10) reveals that the time response is inversely proportional to
the rate constants and logarithmically dependent on ion concentration changes although in
a complex manner. In the high concentration limit of detection, the square terms become
insignificant reducing the logarithmic term to the simple ratio of (Cu+ - Cu^•^ - fC,,+ - C,.+^

The time required to reach equilibrium via steady state mass transport may also be
determined. The flux at, i.e., X = m may be equated to d/C. Wdt;

dC,
'^m

^ m /

Integration of the above equation yields the following expression for time:

t = [DAL(ki k,)]-^Mn (11

j
j

where ? is the reduced concentration parameter fC„ - and ^. is defined as the fractional
reduced concentration. In terms of the average concentration change in compartment R:

t = [d^l^^i ' k,)]-"^|^ln [(L - m) C^^^ - - In [(L - m) C^^^ - . c^gj^ . (12)

3]
j |^ |
j

The reverse process considering a high to low step concentration change is similarly
treated, resulting in virtually identical relationships with the exception that
e' = [l<i/(ki + k2)](^CQ - ) substitutes the expression for 6.

4. Discussion

The sensitivity and detection limit of a CO2 gas sensing electrode are determined by
the equilibrium reaction (1) and by eq. (9). The detection limit is also a function of the
resolution of the pH electrode; for a limit of 0.01 pH unit the CO2 concentration detection
limit can be calculated from eq. (9) to be approximately 10" M. The equilibrium response
of the electrode is not of prime consideration here, but rather the time-dependent response
to changes in CO2 concentration.

Equations (1) and (4) give the steady state CO2 concentration profiles through regions
M and R following a step change in sample concentration. Figures 3 and 4 show numerical
solutions to these equations under different conditions. In figure 3, conditions correspond
to a near ideal CO2 sensing electrode. Diffusion of CO2 across the membrane region M is
almost unimpeded, i.e., the membrane is an "air gap." However, it must be noted that
diffusion of CC2 into the electrolyte region R is of much greater importance, even though
the thickness of layer R is much less than M. The curve shown in figure 4 corresponds to
a more general case in which layer M represents a diffusion barrier, i.e., it is a gas-
permeable plastic membrane.

232
 M—IK
0.10
m — 0.01 cm

L 0.012 cm
0.08
Dal 1.1 X 10^ cm^ sec

Dab z.'t X lU cm sec

0.06
Cone.
— 10'^ M
Molar
10"^ M
0.04
0.1 sec'^

^2 10^ sec"^
0.02

.001 .002 0.010 0.011 0.012

X (cm)

Figure 3. Concentration profile for CO2 electrode where the membrane M is an

gap.

0.1

0.001

m = 1.0 cm
L = 1.2 cm
= 10"^ cm^ see
Dal
Cone.
= 10"^ em^ sec
Molar Dab
10'^ M
^a:
= 10'^ M
1
a = 10^ (cm^)

^
= 10'^ M
cm

0.0001

0
0.1 0.2 0.90 1.00 1.10 1.20
X (cm)

Figure 4. Concentration profile for a generalized gas sensing CO2 electrode

where the membrane M is a gas permeable plastic.

233
 The time response of the electrode is given by eq. (10). Table 1 shows some cal-
culated values for 99 percent response times for various concentration changes. It must be
noted that the time response for a high-low concentration change is greater than for the
reverse. The magnitude of the differences in these response times is a function of the
homogeneous chemical kinetic rate constants and the diffusion coefficient of CO2 in the
electrolyte region R. In the case of the diffusion rate limited process, the thickness of
regions R and M effect the response time equation.

Table 1. Time required to achieve 99 percent of concentration change .

Initial Final t [Dn. (ki + ks)]"' eq. (11)

concentration concentration . /. , , s
/t^n
^ ^
(moles/liter) (moles/liter)

IQ-^ 10"^ 4.61

10"^ 10"^ 4.61
10"^ 10"^ 4.61
10"^ 10"^ 9.21
10"^ 10"^ 11.51
10"^ 10"5 13.82

The results obtained using this model of a CO2 sensing electrode agree in principle with
the known experimentally determined properties. It is hoped that the model will lead to a
clearer understanding of those factors which contribute to the electrode response and to im-
provements in electrode performance.

References

[1] Clark, L. C, Jr., Monitor and control of blood and tissue oxygen tensions. Trans.
Amer. Soo. Artificial Internal Organs, 2^, 41 (1956).

[2] Severinghaus, J. W. and Bradley, A. F., Electrodes for blood PO2 and pCOa determination,
J. Appl. Physiol. U, 515, (1958).

[3] Lucero, D. P., Design of membrane-covered polarographic gas detectors, Anal. Chem. , 41_,
613 (1968).

[4] Gutherman, H. E. and Goldstick, T. K., Numerical analysis of oxygen profiles around a
Clark oxygen electrode. Proceedings of the 27th Annual Conference on Engineering in
Medicine and Biology, Vol. 16, p. 399 (The Alliance for Engineering in Medicine and
Biology, Chevy Chase, Maryland, 1974).

[5] Ross, J. W., Riseman, J. H. and Kreuger, J. A., Potentiometric gas sensing electrodes,
in International Symposium on Selective Ion-Sensitive Electrodes, G. J. Moody, ed.,
p. 473, (Butterworth & Co., Ltd., London 1973).

[6] Severinghaus, J. W., Measurement of blood gases: PO2 and pC02j Annals of the New York
Academy of Sciences, 148 , 115 (1968).

234
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

MONITORING OF OXYGEN PRESSURE IN HUMAN AND ANIMAL BLOOD

H. P. Kimmich and F. Kreuzer

Department of Physiology
University of Nijmegen
Nijmegen, The Netherlands

The surveillance of arterial oxygen pressure (Pa02) in patients with pulmonary or

cardiac failure by frequent sampling of blood has several obvious disadvantages. Con-
tinuous monitoring of Pa02, including recording of variations in Pa02 with respiration and
cardiac activity, provides more useful information, but this sophisticated monitoring by
use of catheter electrodes is still affected by numerous problems. The size of the sensor
as well as stability, flow dependency and transient response are crucial factors. Low
flow dependency and fast response oppose each other [1] but these requirements calling
,

for an optimal compromise in any individual situation have received relatively little
attention since a fast response has often been considered to be unimportant, particularly
in clinical monitoring. It cannot be ignored, however, that a fast response is often
necessary or at least desirable and may provide interesting accessory information [2-4],
We have been paying particular attention to this aspect for a number of years.

1. General Problems and Requirements of PO2 Electrodes

in Continuous Monitoring

The problems and requirements of PO2 electrodes in continuous monitoring in vivo include
technical as well as biological aspects. For the evaluation and appraisal of any such
system the following check list may be useful:

A. Sensitivity and accuracy including linearity, reproducibility, short-term stability,

long-term stability, shelf life, transient response, zero current (value and drift), flow
dependency, temperature dependency, sensitivity to CO2, sensitivity to other gases, sensiti-
vity to pressure, to acceleration, and to water vapor (water deposition).

B. Design and constructive lay-out including miniaturization, safety, and complexity

of servicing and operation.
C. Sterilization.

D. In vivo parameters including biocompatibil ity (toxicity, blood changes, tissue

reactions, blood coagulation, etc.), mechanical and electrical compatibility (flow distortion,
tissue changes, electroshock, membrane potential changes), influences of the medium on the
sensor (protein deposition, blood coagulation at electrode, mechanical stress, galvanic
reactions)
Intensive research during the past years has led to the solution of most problems
connected with the technical parameters sensitivity, accuracy, constructive lay-out, and
sterilization. For human applications, however, improvement of the constructive lay-out
remains desirable, particularly with respect to the size of the transducer. For prolonged
monitoring of human Pa02 the main issue, however, is the solution of the problems connected
with the in vivo parameters, particularly the biocompatibil ity of the catheter electrode and
the influences of the medium on the sensor. But also the mechanical compatibility and the
positioning of the transducer are important in certain circumstances. Biocompatibility of a
catheter electrode is never complete but can be improved by a careful construction of the
sensor and by use of proper materials including coating the tip of the catheter with sili-
cone. Protein and blood cell deposition as well as blood coagulation at the tip of the

Figures in brackets indicate the literature references at the end of this paper.

235
 •

electrode affect fast electrodes much more than slow ones. Since these phenomena also
involve a certain risk for the patient, it is advantageous that their effects on the fast
electrodes become noticeable quite readily and thus provide a warning at an early stage.

Trials of monitoring PaOz in the human radial artery using our ring electrode [5]
revealed the necessity of a catheter probe with a diameter smaller than 2 mm. First, this
problem was circumvented by measuring Pa02 in an arterio-venous shunt (as described below),
but simultaneously a new electrode with a diameter of 1.2 mm and essentially similar charac-
teristics as previous designs has been developed. These two lines of approach will now be
discussed.

2. Monitoring of Human Pa02

Jank, et at. [6] applied our previously developed type of ring electrode [5] to patients
by inserting it into a specially made shunt between the radial artery and the antecubital
vein. A slight disadvantage of this system is a possible reduction of the peripheral circu-
lation due to the shunt. On the other hand, this system offers several important advantages.
The markedly increased blood flow in the shunt (as compared with normal radial arterial
flow) decreases the risk of blood coagulation (which is further minimized by coating with
silicone and continuous instillation of heparin through the arterial end of the shunt),
provides thermostating of the metal block holding the electrode by the blood itself, and
secures a well-defined flow situation which excludes artifacts due to the remaining flow
sensitivity of the electrode. Further advantages of this shunt arrangement are that calibra-
tion can be checked at every desired moment, the transducer can be replaced at any moment if
necessary, and other parameters may be measured simultaneously. Cathode and Teflon membrane
are sterilized at 55 °C with ethylene oxide, the Silastic shunt tube by germicide. The
system, as tested in vitro, is stable (variation less than 0.5 percent in 24 hours), linear
and precise (± 0.2 percent) in a broad range of oxygen pressures (from about 10 to more than
700 mm Hg); its response time for 95 percent deflection is 0.4 second.

Continuous recording of Pa02 was achieved for periods of 6 to 24 hours in more than 100
patients, mainly with respiratory failure. The PO2 of blood samples taken from the shunt of
each patient was compared with the continuous Pa02 readings. The values tallied well up to
110 mm Hg. A systematically lower PO2 measured in the sample from 110 mm Hg upwards is
likely due to leakage from the syringe and oxygen consumption by the erythrocytes during
transport and analysis of the blood samples. This system may be reliably used for monitoring
Pa02 in patients up to 24 hours without occurrence of any cutaneous or vascular reactions.

3. New Miniature PO2 Electrode

Our new electrode has a lay-out similar to our previous design [5] and also a response
time for 95 percent deflection of 0.4 second, but a diameter decreased from 2 to 1.2 mm
[7,8]. The cathode again is a platinum ring of 0.3 mm diameter and approximately 3 ym
thickness which makes possible a very low flow dependency in the presence of a relatively
high current output. In order to guarantee a fast sufficient electrolyte exchange in the
space between the membrane and the cathode area (which is important for the stability and
small CO2 sensitivity of the electrode) the ratio of the unsupported to total membrane area
had to be increased from 4 to 6 percent. Since this might lead to an increased sensitivity
to pressure, the static and dynamic sensitivity to pressure was investigated with particular
care. The most important characteristics of the probe are (with a 6 ym Teflon membrane at
37 °C):

sensitivity = about 0.6 nA/mm Hg,

linearity = ± 0.2% (from 10 to 760 mm Hg),
reproducibility = better than 0.1 mm Hg,
short-term stability (2 h) = ± 0.25%,
long-term stability (24 h) = ± 1%,
response time for 95% deflection = 0.4 s,
zero current < 15 mm Hg (drift 0.5 mm Hg/24 h),
flow dependency (7-100 cm/s) = ± 1%,
temperature dependency (nonlinear) = 2-3%/l °C,
sensitivity to physiological CO2 < 2 mm Hg,
sensitivity to pressure: static ^ 0.2%,
dynamic < 1%,
sensitivity to acceleration (0 to 10 g) < 0.1%.

236
 This electrode so far has been applied to physiological studies in the dog and cat as
well as to continuous monitoring of human Pa02 in the radial artery where it was introduced
through a Teflon needle. Figure 1 shows a tracing in a cat. In the animal experiments
where the catheter tip was protected by silicone oil, no blood coagulation occurred at the
tip for several hours.

Ventilatory 40/ 20/ .

mi n ^0/min
; 1 i
1

-—
Frequency !

p i

1
i.o -h-hT
[10 -H
APaQ (torr)
i

1
•
Tt
!

-4

-J

. .
1

Respiratory p500

% (torr) 0

r60
Respiratory

(torr) -
'C02

Integrated b
Phrenic
Nerve
Activity
fmmmmi mMmmm smb?mm
r200
Systemic
Blood
Pressure (torr)

10 s

Figure 1. Recording of arterial oxygen pressure (PaOa; top tracing) in the

carotid artery of an artificially ventilated cat during an experiment con-
cerning the regulation of breathing. The respiratory frequency was changed
from 40/min to 10/min, while end-expiratory PO2 and PCO2 were kept constant
by adjusting the inspiratory concentrations. Note the respiratory fluc-
tuations of PaOa (APaOa), increasing with decreasing respiratoryf requency,
which can be followed faithfully by these fast PO2 electrodes (for proof,
see reference 4)

4. References

[1] Schuler, R. and Kreuzer, F., Rapid polarographic in vivo oxygen catheter electrodes,
Eespir. Physiol. 3, 90 (1967).

[2] Yokota, H., Respiratory fluctuations of oxygen pressure in alveolar gas and blood of the
dog, thesis, Nijmegen, 94 pp. (1973).

237
Kreuzer, F. Respiratory fluctuations of oxygen pressure in alveolar air and arterial
,

blood, in Oxygen Measurements in Biology and Medicine, J. P. Payne and D. W. Hill, eds.
p. 139 (Butterworths , London and Boston, 1975).

Kreuzer, F. , Transmission of alveolar oxygen pressure oscillations, in Morphology and

Mechanisms of Chemoreceptors A. S. Paintal ed., p. 176 (Navchetan Press, 1976).
, ,

Kimmich, H. P., and Kreuzer, F. , Catheter PO2 electrode with low flow dependency and
fast response, in Oxygen Pressure Recording in Gases, Fluids, and Tissues, F. Kreuzer
ed.; Progr. Resp. Res. 3^, H. Herzog, ed., p. 100-110, (Karger, Basel, 1969).

Jank, K. Hemptinne, J., Swietochowski

, , A., and Demeester, M., Continuous in vivo
measurement of arterial PO2 in humans, J. Appl. Physiol. 38, 730 (1975).

Kimmich, H. P., Kreuzer, F., Spaan, J. G., Jank, K. , de Hemptinne, J, and Demeester, M.,
Monitoring of PO2 in human blood, in Oxygen Transport to Tissue - II. Adv. Exper. Med.
Biol., Vol. 75, J. Grote, D. Reneau and G. Thews, eds., p. 33 (Plenum Press, New York,
1976).

Kreuzer, and Kimmich, H. P., Recent developments in oxygen polarography as applied

F.
to physiology, in Measurement of Oxygen, H. Degn I. Balslev and R. Brook, eds., p. 123
,

(Elsevier, Amsterdam, 1976).

238
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Gaithersburg, Maryland, July 7-8, 1975. Issued June 1977.

ALTERNATIVE METHODS OF CO2 MEASUREMENT, WITH PARTICULAR

REFERENCE TO CONTINUOUS RECORDING

L. H. J. van Kempen and F. Kreuzer

Department of Physiology
University of Nijmegen
Geert Grooteplein Noord 21a
Nijmegen, The Netherlands

Continuous in vivo recording of CO2 pressure (PCO2) in the gas phase has long been
common but continuous in vivo monitoring in fluids, particularly blood, has been impeded by
the relatively large dimensions and the slow response of the available devices. A method
of measuring PCO2 in liquids accurately and rapidly by a miniaturized instrument analogous
to the oxygen electrodes developed by Kreuzer and his coworkers [1-3]^ presumably might help
to solve many problems in physiology and reaction kinetics, as well as in clinical medicine.

The conventional CO2 electrode according to Stow et al. [4,5] has been widely used but
remains limited to a response time of the order of one minute in most cases, particularly
with miniaturized types which furthermore are apt to be less stable and have a shorter life
time. The recently introduced General Electric catheter CO2 electrode [6], e.g.^ still has
a response time of one minute. A recent calorimetric CO2 electrode [7] also has a response
time of 2 minutes for 99 percent deflection. We therefore had decided to investigate the
possibility of applying other principles to the measurement of CO2. In the present paper,
we will discuss the results of our experience with the quinhydrone electrode and with the
conductivity electrode.

1. The Quinhydrone Electrode

The quinhydrone system was previously applied to pH measurement and has now been
modified for the determination of PCO2 [8]. The principle resides on the oxidation and
reduction of hydroquinone and qui none, respectively; the oxidation-reduction reaction and
the corresponding electrode potential are described by:

Q + 2H'^ + 2e" ^ HgQ (1)

F
t
-
- F
L„
0
+
+
2F
,
In -—
^^H^^
a^^Q
-
= r
0
.
RT ,
E„ + or In -^^^
2F a^^Q
^
F
—
RT ,
+ —f In an+
, ^
H
(2)

where Q = quinone; aq, s,^ a^+ = ionic activities of quinone, hydroquinone and hydrogen
g,
ions, respectively; E = electrode potential; E = standard electrode potential = 700 mV at
25 °C; other symbols as usual.

If the activities of quinone and hydroquinone are the same, the electrode potential
becomes independent of any of these species and is a function of the hydrogen ion activity
only. This condition prevails when using quinhydrone, an equimolar compound of quinone and
hydroquinone.

Figures in brackets indicate the literature references at the end of this paper.

239
 A membrane- covered electrode was constructed with either a calomel or a silver-silver
chloride electrode as a reference, either of which is suited. Figure 1 shows the con-
struction of this electrode. A platinum ring (o.d. 4.9, i.d. 4.5 mm) was mounted around
the liquid junction of a calomel electrode, or a platinum plate (surface area 8 mm^)
together with a silver ring (o.d. 6.8, i.d. 5.8 mm) was melted into glass (not shown here).
A membrane (Silastic 25 ym or Teflon 6 ym) was mounted and the electrode housing was filled
with a solution_of 2 ml quinhydrone (10"^ mol/1) and KCl (0.1 mol/1) with or without sodium
bicarbonate (10'^ mol/1). Lens paper was used as a spacer to stabilize the fluid layer.

1 ucite support

rubber ring
covering KCI inlet

tefl on
brass ring

calomel electrode-

1 uci te
stainless steel
glass chamber —
circulating water —
0-ring supporting
calomel liquid junction
the membrane

rubber ring

1 ucite housing
supporting the
membrane
|-1 ucite
.platinum
ring

teflon membrane P]atin cal omel

[
ring liquid junction
I

I
I

lenspaper —
Figure 1. Construction of C02-quinhydrone electrode [8]
(with permission of Respir. Physiol.).

The pK of hydroquinone is 10.32 ± 0.08. The plot of E vs pH is linear between pH 4

and 9, with a slope -AE/ApH = 58.2 mV/pH at 25 °C, the same with or without CO2. The plot
of E vs log PCO2 with 0.1 mol/1 KCl is linear between PCO2I5 and 80 mm Hg, with a slope
-AE/Alog PCO2 = 55.8 mV/log PCO2 at 25 °C in the presence of bicarbonate. This results in
a -ApH/Alog PCO2 = 0.96, in good agreement with the conventional CO2 electrode.

A number of factors such as chloride concentration, oxygen concentration, light, time,

acid-base reactions of hydroquinone and quinone, semiquinones autoxidation and salts
, ,

influence the performance of this electrode, but it could be shown that their effect in 0.1
mol/1 KCl is negligible in physiological conditions [9].

The most important characteristics of performance of this electrode, to be compared

with the conventional CO2 electrode, are listed in table 1. Calibration lines are shown in
figure 2 in bovine blood at 37 °C with or without bicarbonate added to the electrode solution.

240
 Table 1. Synopsis of properties of various CO^ electrodes
for continuous monitoring in -vivo.

General Electric
Property Stow catheter probe Quinhydrone electrode Conductivity Cal or i metric
electrode (Stow principle) macro micro electrode electrode

Accuracy 2% ±2-3 mm Hq similar to Stow electrode <2 mm Hg SE = ±5 mm Hg

Cal ibration 1 inear 1 inear 1 inear 1 inear 1 inear 1 inear

Response 1 min 67 s 2 min 1 min 4-7 s 2 min

time for '99. 5% for 90% for 95% on for 95% for 90% for 99%
{^'tn vivo)
3 min
for 95% off

Stability 0.2 mV/h ±3 mm Hg <0.2 mV/h 4 mV/h +5 mm Hg

during 6 h (platinized) during 1 week

"NaHCOg ^H20
(mV) (mV)
-IOOt r+20

0.025 10
-80- -+40

60- +60

40- -+80

'NaHCOc 57.2

20 ^HoO 29.7
h+100

—r— —I 1
T 1
'^C02(mmHg)
10 20 40 60 80 100

Figure 2. Calibration lines of COa-quinhydrone electrode with

bovine blood equilibrated with various CO2 concentrations at
37 °C with or without bicarbonate added to electrode solution
[8] (with permission of Respir. Physiol.).

The performance of this electrode is fully comparable to that of the conventional CO2
electrode. Compared with the latter, it has the advantage of being an inexpensive low-
impedance electrode which may be miniaturized more easily and therefore might be adapted to
in vivo PCO2 estimation.

241
 2. A Single-Unit Electrode for CO2 and O2

We also miniaturized the C02-quinhydrone electrode and combined it with a polarographic

oxygen electrode of the CI ark- type, yielding a single-unit CO2-O2 sensing microelectrode
system. The design of this electrode is similar to that of the oxygen electrode constructed
by Schuler and Kreuzer [1] and Kimmich and Kreuzer [2]; a slightly larger version is shown
in figure 3 [10].

insul ation fluid compartment

silver-silver chloride electrode Araldit
trovidur lead platinum electrode
I

si 1 ver support silver ring

catheter Silastic membrane

8. 7 mm 3.5 mm

Figure 3. Construction of C02-quinhydrone microelectrode system [10]

(with permission of Re spiv, Physiol.).

The cylindrical probe has an external diameter of 3.5 mm and a length of 8.7 mm. It
consists of a central platinum electrode with a surface area of about 2.5 mm^ sealed with
Araldite into a holder of polyvinyl chloride (Trovidur) and surrounded by a tubular Ag-AgCl
reference electrode. The tip of the probe is covered by a thin membrane (Silastic 25 pm or
Teflon 6 ym) fixed to the reference electrode by a silver ring. A quinhydrone solution of
10 3 mol/l in 0.1 mol/1 KCl with or without bicarbonate added, is introduced into the
probe.

The performance data of this probe used as a CO2 electrode are listed in table 1.
Figure 4 shows linear calibration lines for blank and platinized platinum electrodes. The
line of the blank probe lies somewhat above that of the platinized probe. The slopes with
0.1 mol/1 KCl are similar and lower than in the macroelectrode described above. The
sensitivity -ApH/Alog PCO2 thus is about 0.42. Stability is better in the platinized probe
but considerably inferior to that of the macroelectrode described above. Response time,
however, is superior to that of the macroelectrode.

The properties of this probe used as an oxygen -electrode are fully comparable with
those of the oxygen sensors described previously by Kreuzer and his group, particularly
also concerning the fast response, the response time for 95 percent deflection being about
0.4 second at 20 °C, independent of the electrolyte composition. The polarogram is not
affected by addition of quinhydrone with or without bicarbonate.

The responses of this unit used as a CO2 electrode to oxygen and used as an oxygen
electrode to CO2 are mutually independent in physiological conditions, at least as long as
the succession is not too fast. In view of a response time for 95 percent deflection of
about one minute for CO2 and about 0.4 second for oxygen, it is possible to take successive
recordings of oxygen and CO2 in vivo after elapse of the respective response times.

There is a lower limit for the surface area of the electrode concerning CO2, being
reached in the blank platinum electrode at a surface area of about 1 mm^, which is im-
portant for further miniaturization. The platinum surface of this electrode is too large
for the determination of oxygen in fluids since the flow dependency is considerable. Any
compromise in this respect may be difficult without membranes combining high permeation
rate for CO2 (in order to get a sufficiently fast response) with poor permeability to oxygen.
.

242
 3. The Conductivity Electrode

In a further attempt to reduce the response time which is still too long in the types
of quinhydrone electrode just described, an entirely different principle was adopted in the
device to be presented now, in which the principle of establishing a CO2 equilibrium across
the electrode membrane has been abandoned. Electrolyte conductivity has proven to be a
fast, sensitive and accurate method for the determination of CO2 in liquids. The micro-
electrode system presented here is a first attempt to apply conductometry to continuous
monitoring of PCO2 in vivo [11].

A stainless steel tip is attached to one end of an x-ray double-lumen polyethylene

catheter of 60 cm length, 2.7 mm external diameter and 0.8 mm lumen diameter. A Silastic
membrane of 25 pm is mounted on this tip by a device specially designed for this purpose
and fixed by a stainless steel ring. Each lumen is separately connected to a conductivity
cell located at some distance from the membrane. The whole electrode system is flushed
continuously with bidistilled water at a constant flow rate. Electrolyte conductivity of
the water is measured before entering the catheter and after having passed the membrane.
The difference in conductivity between inflowing and outflowing water is related to the
PCO2 of the medium to which the electrode is exposed.

The sensitivity of the electrode system, which is the change in specific conductivity
(in S'cm"^) of the carrier to a change in PCO2 of the medium, is determined, among other
factors, by the efficiency of CO2 transfer from the medium across the membrane to the
carrier. It depends on membrane permeability and surface area, contact time of the carrier
with the membrane (contact length/flow velocity), and carrier volume per unit membrane
surface area (carrier layer thickness). In this electrode prototype, a membrane surface
area of 3.5 mm^ and a layer thickness of 250 m were chosen. Carrier flow rate was within
a range of 1 to 7 ml/min, thus giving contact times between 0.05 and 0.007 second.

The choice of the conductivity cell depends on the specific conductivity to be measured.
Since at a specific conductivity of the order of yS-cm"^, variations of a few percent have

243
to be determined, a conductivity cell with a_1ow cell constant (= distance between the two
plates/surface area of one cell plate, in cm"^) is required. The cell volume should be
close to that of the carrier in contact with the membrane in order to maintain the upper
limit of the frequency at which PCO2 fluctuations of the medium can be followed by the
electrode. A series of 6 types of cell were investigated from which the type shown in
figure 5 was most suited. It consists of two platinum wires (0.2 x 0.05 mm) coated by
Araldite and wound at a distance of 0.2 mm to a coil of 0.6 mm i.d. and a length of 8 mm
(the interior of the coil being stripped of insulating Araldite), fitting into the lumen of
the catheter tip. The main advantage of this location is the reduction of the response
time and the convenience of temperature control. Between uses, the cells were filled with
Hibitane 0.2 percent to prevent microbial contamination.

Figure 5'. Stainless steel conductivity electrode tip covered by a C02-per-

meable membrane provided with a spiral conductivity cell and mounted on
a flexible catheter [11] (with permission of Respir. Physiol.).

Electrolyte conductivity was determined with a commercially available direct-reading

conductivity meter (Radiometer COM 2, frequency 70 Hz, test voltage 0.25 V) connected to a
two-channel recorder. Conductivity meter and recorder were calibrated by precision resistance
boxes. In this way, it was possible to estimate variations in specific conductivity of up
to 0.02 yS'cm"^. A further improvement was achieved by applying a conductivity bridge
which includes both measuring and reference cell. This makes it possible to directly
detect relative variations in impedance due to CO2, reduces the difficulties of eliminating
A-C interference (capacitance effects), and minimizes the influence of temperature changes.

A plot of the specific conductivity against PCO2 in water is curvilinear and convex to
the abscissa. However, in the low range used in the operation of the CO2 electrode (specific
conduc tivity about 1-3 yS-cm"^) the curve is practically lineajr, as shown in figure 6.
Position and slope of this straight line depend on the flow rate. Increasing flow rate
lowers the specific conductivity at a constant PCO2 as well as the sensitivity (slope).
The reduction of specific conductivity at PCO2 = 0 is due to the decreasing conductivity of
the carrier at higher flow rates, probably originating from boundary effects on the surface
of the conductivity cell. The sensitivity for CO2 is diminished at higher flow rates
because both contact time of the carrier with CO2 at the membrane and time available for
hydration of CO2 are shortened. The dependencies of sensitivity and response time on the
flow rate are both curvilinear and convex against the coordinates. Thus, high sensitivity
and fast response oppose each other. The cell type discussed here at a distance of 10 mm
from the membrane provides, when exposed to a step change of PCO2 between 0 and 5 percent
CO2 and with a flow rate of 0.7 ml/min at 18 °C, a response time of 4 seconds on and 7
seconds off. The sensitivity of this cell in gas or in a stirred liquid phase is similar.
The performance of this cell is again summarized in table 1.

Theoretically, the performance of this electrode will depend on the conditions of

membrane diffusion, convection, and kinetics of CO2. In practical terms, the conclusion is
important that both flow rate and distance between membrane and cell determine the sensitivity
of the system. At constant flow rate, the sensitivity increases with this distance until
chemical equilibrium is reached. At constant distance and varying flow rate, the sensitivity
is determined by the initial CO2 concentration in the carrier which is inversely proportional
to the flow rate, and by the time required to attain a certain fraction of the equilibrium

244
 0.55 ml/min

1 .75 ml/min

"o
3.35 ml/min

Figure 6. Specific conductivity

of the carrier at three different
flow rates (0.55, 1.75, and
3.35 ml/min) after exposure of
1 .80- the electrode to a gas phase
of varying CO2 partial pres-
sure at 10 °C [11] (with per-
mission of Respir. Physiol.)

20 40 60 100
(mmHg)

concentration and hence on the flow rate as well. When a fast response is required, the
conductivity cell should be located close to the membrane. If at the same time high
sensitivity and low CO2 consumption are needed (in order to reduce dependency on the flow
rate of the medium), the carrier flow rate may optimally be chosen in the range where
turnover of CO2 into bicarbonate is high. The flow dependency in liquid media might be
minimized by using membranes with smaller surface area and/or lower permeability to CO2.
The resulting reduction in sensitivity should then be improved by other means, e.g., by
replacement of bidistilled water by highly purified water (conductivity water) which would
increase carrier resistance and thus the relative change in specific conductivity. At
these lower conductivities, however, balancing of the bridge becomes more difficult which
might be partially avoided by using cells of a smaller cell constant. Along the lines of
this reasoning, a further miniaturization of the electrode might be expected to increase
sensitivity and reduce CO2 consumption (flow dependency) at the same time.

4. Accurate Estimation of Total CO2 in Fluids

Finally, it might be mentioned that we have achieved an accurate estimation of total

CO2 in fluids of low CO2 concentration (0.05-0.5 mmol/1) where the conventional Van Slyke
technique with an accuracy of 0.1 mmol/1 is likely to produce errors ranging from 20 to 100
percent. The essence of this method [9,12,13] is the use of an infrared CO2 analyzer (URAS
2, Hartmann and Braun; cell of 250 mm length, corresponding to a full-scale deflection of
0.01 percent CO2) surrounded by nitrogen rather than by air as usual. The inlet and outlet
of the analyzer were connected to a gas extraction chamber provided with a sample inlet
system. All connections were made of CO2 impermeable Viton tubing. The circulating gas
flow of 120 ml/min was regulated by a membrane pump and a flowmeter. In order to expel all
CO2 from the sample, 3 ml of 0.1 mol/1 lactic acid with antifoam were added to the extrac-
tion chamber. Before the injection of the sample, the circuit was flushed with C02-free
nitrogen and then closed by turning a stopcock. The analyzer was calibrated by mixing
various amounts of air (containing 0.03 percent CO2) with C02-free nitrogen in a gas mixing
pump. The calibration curve thus obtained is not linear. At these low CO2 concentrations,
water vapor in the closed circuit interferes with the CO2 readings. This effect of cross-
sensitivity could be eliminated by inserting a suitable filter supplied by the firm. In

245
this way, total CO2 concentrations of 0.1 mmol/1 can be estimated in a 0.5 ml sample with an
error of less than 2 percent. By changing the closed circuit volume or the length of the
cell, CO2 concentrations ranging from 0.05 to 100 mmol/1 can be estimated without any con-
siderable loss of accuracy. Thiele and van Kempen [14,15] applied this method to the mea-
surement of CO2 release by human skin and were able to determine CO2 amounts of ppm/10 cm^
mi n

References

[I] Schuler, R. and Kreuzer, F., Rapid polarographic in vivo oxygen catheter electrodes,
Res-piv. Physiol. 3_. 90 (1967).

[2] Kimmich, H. P. and Kreuzer, F., Catheter PO2 electrode with low flow dependency and
fast response, in Oxygen Pressure Recording in Gases^ Fluids^ and Tissues, F. Kreuzer,
ed.; Progr. Resp. Res. 3, H. Herzog, ed., p. 100 (Karger, Basel, 1969).

[3] Kimmich, H. P., Spaan, J. G., Kreuzer, F. Jank, K. , de Hemptinne, J. and Demeester,
M., Monitoring of PO2 in human blood, in Oxygen transport toTissue - '

II. Adv. Exper

Med. Biol. , Vol 75, J. Grote. D. Reneau and G. Thews, eds. , p. 33 (Plenum Press,
.
r ^ \

New York, 1976).

[4] Stow, R. W. and Randall, B. F., Electrical measurement of the PCO2 of blood, Amer.
J. Physio. 179, 678 (1954). {

[5] Stow, R. W., Baer, R. F., and Randall, B. F., Rapid measurement of the tension of I

carbon dioxide in blood. Arch. Phys. Med. Rehab. 38, 646 (1957). I

Schulz, v., Erdmann, W., Ulmer, H.-V., Kunke, S. , Baum, P., and Frey, R., Zur kontinuierl icherl
Messung des arteriellen Kohlensauredrucks mit Katheterelektroden und Moglichkeiten
ihres klinischen Einsatzes, Anaesthesist , 22, 416 (1972).
|

Vurek, G. G. and Kolobow, T., Continuous blood-gas measurement during prolonged I

extracorporeal circulation, in Oxygen Measurements in Biology and Medicine, J. P. Payne !

and D. W. Hill, eds., p. 285 (Butterworths, London and Boston, 1975). i

van Kempen, L. H. J., Deurenberg, H., and Kreuzer, F., The C02-quinhydrone electrode.
A new method to measure partial CO2 pressure in gases and liquids, Respir. Physiol. \

14, 366 (1972). I

[9] van Kempen, L. H. J., Estimation of Free and Hemoglobin-Bound CO2, Thesis, Nijmegen, i

100 pages (1972). i

[10] van Kempen, L. H. J. and Kreuzer, F., A single-unit carbon dioxide-oxygen sensing \

microelectrode system, Respir. Physiol. 23_, 371 (1975).

1

[II] van Kempen, L. H. J. and Kreuzer, F., The CO? conductivity electrode, a fast responding I

'

CO2 microelectrode, Respir. Physiol. 2A_, 89 (1975).

'

[12] Ortega, F. G., Orie, S. A. M., and Tammeling, G. J., Determination of carbon dioxide j

content of blood by infrared analysis, J. Appl. Physiol. 21_, 1377 (1966). I

[13] van Kempen, L. H. J. and Kreuzer, F., Accurate estimation of low CO2 content of fluids
by infrared analysis, in Oxygen Affinity of Hemoglobin and Red Cell Acid-Base |

Status, M. R(iirth and P. Astrup, eds., p. 247 (Munksgaard, Copenhagen, 1972). I

[14] Thiele, F. A. J. and van Kempen, L. H. J., Die Porositat der menschlichen Haut: CO2- i

Abgabe durch Haut oder Schweissdriisen? (Al kal ineutral isation) , Fette, Seifen '

Anstrichmittel, 2_, 190 (1971 ).

r

[15] Thiele, F. A. J. and van Kempen, L. H. J., A micro method for measuring the carbon dioxide
release by small skin areas, Brit. J. Dermatol. 86_, 463 (1972). i

246

I
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

NBS STANDARDS FOR pH AND ION ACTIVITY

MEASUREMENTS IN BIOLOGICAL FLUIDS

Richard A. Durst
Analytical Chemistry Division
National Bureau of Standards
Washington, D.C., 20234, USA

and

Roger G. Bates
Department of Chemistry
University of Florida
Gainesville, Florida, 32611, USA

1 . pH Standards

It has been about fifteen years since the National Bureau of Standards, in response to
the needs of clinical chemists and physiologists, certified a special composition of the
phosphate pH buffer for use in the biological range of interest between pH 7.3 and 7.5 [1]^.
This buffer, with a ph value of 7.392 at 37 °C, has been widely accepted and used in clinical
laboratories as the primary pH standard. The proximity of the pH value of this standard to
that of blood minimizes errors due to faulty response of the glass electrode or nonlinearity
in the meter response. Of course, calibration with a second pH standard such as the 1:1
phosphate buffer (pH = 6.839 at 37 °C) is necessary to detect gross deviations from Nernstian
behavior which would necessitate replacement of the glass electrode.

In 1972, NBS reported the availability of a new buffer for use in the physiologic pH
range (pH = 7.382 at 37 °C) [2]. This buffer, tris(hydroxymethyl )aminomethane ("Tris") and
its hydrochloride salt mixed in a ratio of 1:3 at a tris-HCl molality of 0.05, was selected
because it exhibited several advantages over the phosphate buffer. These included stability
and compatibility with biological fluids and a temperature coefficient which more closely
approximates that of whole blood than does the phosphate buffer [2]. Since preliminary
experiments with cells containing a saturated KCl bridge solution and liquid junction
indicated residual liquid junction errors (i.e.j differences between the operational pH and
the corresponding pa^ values) of up to 0.041 pH unit at 25 °C when compared to the 1:1
phosphate buffer [3], the original Tris buffer was issued on a "provisional" certificate
until consistency with the pH scale could be checked more thoroughly. This buffer should be
considered a secondary standard until more extensive studies of the liquid junction problem
are completed.

Ladenson, et at. have also noted an inconsistency in the operational pH values when
Tris buffers are measured on pH/blood gas analyzers which had been standardized with the
phosphate buffers [4]. While their data indicate liquid junction problems with these
instruments, the interpretation of their findings is not unequivocal, since they also find
pH errors when a physiologic phosphate buffer is measured. The pH errors with the Tris
buffers are somewhat larger but also exhibit both positive and negative deviations from the
expected values. When directly compared to the NBS phosphate buffer, the Tris buffer gave
a pH value which was about 0.012 unit below the assigned value. While the Tris buffer may
not be suitable for use as a primary standard, Ladenson, et at. found it is eminently
suited for comparative measurement with blood pH systems. It also proved to be convenient
and effective in monitoring the performance of these analyzers.

Figures in brackets indicate literature references at the end of this paper.

247
 In order to better approximate the colligative properties of blood plasma and serum,
preliminary investigations have been made into the certification of the Tris buffer in
isotonic saline solution [2]. While it was hoped that this expedient would reduce the
uncompensated residual liquid junction potential, initial studies at NBS [3] again indicate
operational pH values that are low by about 0.05 pH unit at 25 °C when measured in a cell
with a saturated KCl liquid junction. It is obvious that much more extensive research must
be carried out to elucidate the nature and magnitude of this problem.

Whereas the Tris buffer may eventually prove to be unsuitable as a pH standard, there
are many other possible materials for use as clinical pH buffers, such as the materials in
the list published in 1966 by Good, et at. [5]. One of these, tris(hydroxymethyl )methylglycine
("Tricine"), was the subject of a recent study by Bates and coworkers [6] who found that a
solution consisting of 0.06 m Tricine + 0.02 m sodium Tricinate (pa^ = 7.407 at 37 °C) had
a residual liquid junction error of about 0.01 pH unit which would not exclude the Tricine
buffer from consideration as a primary standard for pH measurements. Studies in isotonic
saline solution of two other zwitterionic buffers, HEPES [N-2-hydroxyethylpiperazine-N'-2-
ethanesulfonic acid] and TES [N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid], with
pH values between 7 and 8 have recently been completed by C. A. Vega at the University of
Florida.

2. Ion Activity Standards

Ion-selective electrodes, like the pH glass electrode, are finding growing use in
various biomedical studies for determining or monitoring the activities of ions such as
Ca'^''', Na+, K"*", Cl~, F", etc. One very important area of development is the use of these
electrodes in automatic, multi -electrolyte clinical analyzers. As in the case of the pH
electrode, standards are also required for the reliable application of these sensors [7].
Investigations at NBS have been primarily concerned with aqueous standards and will progress
in the future to mixed electrolyte systems and synthetic biological fluids. To date, NBS
has certified standards for Na"*", K"^, CI", and F", and a standard for Ca"*"*" is in progress.

In considering the certification of ion activity standards, one important difference

exists in regard to the measurement process. For pH certification, an electrode system of
proven reversibility exists, i.e., the hydrogen gas electrode/silver-silver chloride electrode,
whereas in the case of ion-selective electrodes, the ideality of behavior has not been
unequivocably demonstrated. This consideration requires that more reliance be placed on the
activity scales and theoretical conventions used in calculating the single-ion activities.
The NBS has adopted the convention for single ionic activities proposed by Bates, et at. [8]
which is based on the hydration theory of Stokes and Robinson [9]. The details of this ion
activity scale have been reviewed by Bates [10] and will not be repeated here. Suffice it
to say that the single ion activity coefficients are calculated from experimental values of
the mean molal activity coefficient (yMx)» the osmotic coefficient (0), and a hydration
number (h) according to the equations for univalent electrolytes:

log = log ^ 0-00782 (h^ - \\^)m 0

Y[vix

and

log Yx" = log

Ymx
+ 0.00782 (h^ - \)m 0.

As in the case of the pH scale, it is necessary to make certain extra-thermodynamic assumptions.

The "conventional" step in the hydration treatment is the assignment of hydration numbers to
individual ionic species. The hydration numbers are calculated from smoothed values of Y|^x
and 0 using the Stokes-Robinson hydration theory. The hydration number for the electrolyte
is then "split" between the cation and anion by assigning a hydration value to one of the
ions. For the present scale, the hydration number for the chloride ion was taken to be
zero, and the other ionic hydration numbers are referred to this convention [10]. Ionic
activity for unassociated electrolytes derived on the basis of this hydration convention
have been found to be consistent with the observed responses of ion-selective electrodes.

248
 Whereas the ion activity standards certified at NBS to date have been single salts in-
tended for use in relatively simple electrolyte solutions, such standards are only reliable
over a concentration range of three or four decades, due to ion association problems at
higher concentrations and contamination and/or adsorption problems at lower concentrations.
One way to avoid this latter difficulty is to prepare ionic activity buffers which can
extend the low activity limit by several orders of magnitude.

Early studies at NBS demonstrated that the linear response of the silver ion-selective
electrode could be extended down to approximately 10"^.^ mol/1 with solutions consisting of
mixed silver halides and sulfide [11]. Not only was the response linear and approximately
Nernstian over 25 decades of concentration, but whereas the response time of the electrode
in unbuffered silver ion solutions increased with decreasing concentration, in the silver-
buffered solutions, the electrode response was almost immediate, i.e.j less than 15 seconds,
even at the lowest levels of silver ion activity.

While the concept of "metal buffers" is not new, their use in the calibration of ion-
selective electrodes has been studied primarily by workers in Denmark and Hungary [12-17].
In Denmark, Blum and Fog developed water-soluble buffers for copper which avoid the slowness
of solution equilibration inherent in precipitate- type buffers which involve heterogeneous
equilibria [12]. This concept has been extended by Hansen, Ruzicka and coworkers [13-15].
They calibrated ion-selective electrodes using buffers for CU++, Cd"*"^, and Ca''"'' based on
soluble EDTA and NTA complexes.

In Hungary, Havas, Kaszas, and Varsanyi have reported [16] the development of mixed
precipitate buffers for the halides (including fluoride) and silver ion. More recently.
Bailey and Pungor have suggested that, instead of serial dilutions of standards, the standards
in the range from 10"'+ to 10"^ mol/1 should be prepared by electrolytic generation of the
appropriate ion [17]. They demonstrated their technique with iodide and silver and also
indicated that sulfide, fluoride and thiocyanate could be generated electrochemically
Although the solutions are not buffered, an advantage of such a procedure is that it could
be automated and is more reproducible than the serial dilution method.

Great care must be exercised, however, in applying either the "pure" electrolyte or
buffered standards to the calibration of ion-selective electrodes below the point where the
pure electrolyte response begins to deviate from linearity. The emf vs. ion activity response
will be entirely different for the two types of standards, and the interpretation of the
sample data will depend upon whether the sample system is buffered or not. It is clear that
much more work is required to develop ion activity standards which will be suitable for most
applications and acceptable to everyone.

3. Standards for the Future

As the requirements of clinical laboratories and available instrumentation become more

refined, data for salt effects and medium effects on electrolytes and gases are needed to
provide the basis for the development and certification of the required standards and quality
control materials.

We plan to obtain data for ionic equilibrium processes in saline media, to develop
accurate methods for determining ion concentrations and gas tensions, and to develop multi-
component standards for calibration of the clinical instruments used for the determination
of pH, dissolved gases, and electrolytes in various biological fluids.

Ultimately such standards may contain all of the major ionic constituents, complexing
agents, and non-electrolytes at levels approximating those in the real samples. Thus, a
synthetic blood standard would consist of an isotonic electrolyte solution containing, for
example, Na+, K*, Ca"*"'", Mg"*"^, O2, HCO3, CI", HPO^, and polyelectrolytes in physiological
amounts. Of course, the problems encountered in developing such a multi component standard
would be very great, but the availability of a single standard for the calibration of
instrumentation for the determination of multiple parameters is vitally needed for the
future development of this field. Only when such standards are available will it be feasible
to calibrate multi-parameter clinical instruments conveniently and with some assurance of
the consistency and comparability of data obtained in different laboratories. As far as the
major ionic constituents are concerned, the first step in this direction has already been
taken [18].

249
 We are optimistic that given a high enough priority and sufficient time, the problems
will be solved. The continued support of organizations such as those sponsoring this work-
shop will help insure the success of this goal.

Addendum
NBS Facilities for pH Certification Measurements

Since there was no time during the Workshop on pH and Blood Gases for the participants
to tour the facilities used for electrolyte studies and pH certification, a section from a
recent NBS Special Publication [19] on this subject is being reprinted here.

In 1972, the NBS emf measurement system was automated in order to increase the reliability
of the data acquisition and to reduce the time required for a pH certification run from
about five 8-hour days to two 24-hour days with minimal operator attention. The primary
purpose of this system is to cause a constant temperature water bath to cycle through a
preprogrammed set of temperatures and to record the data when the bath temperature and
potentials read from the measurement cells pass certain stability and control requirements.

The temperature-controlled emf measurement system is diagrammed in figure 1. A centrally

located computer, a Univac Series 60, model 6135, is time-shared with about a dozen other
systems in the Analytical Chemistry Division. The data communications system consists of
four main parts: (1) the digital data bus, (2) the computer interface, (3) the laboratory
logic box interface, and (4) the laboratory control console. The computer interface connects
the computer to the laboratory via the digital data bus and laboratory interface. The
laboratory instrument console provides control and the capability of entering auxiliary
data for use by the computer to service and/or control the interfaced system. Two-party
line digital data busses have been installed in the Chemistry Building with outlets for each

UNIVAC Series 60
Model 6135 Computer

Console Line Mag Tape

Computer Interface Disk
Typewriter Printer Archive

Digital
Data Bus

Lab Interface

Teletype Lab Control Console

Temp Set Point

Selector Digital Barometer Fluke 8400A DVM

Temp Proportional Temp Standard

Sensor Controller Resistor

Input
Multiplexer
Platinum Resistance
Refrigeration Heater Thermometer
Unit Unit

Electrochemical
Constant Temp Electrometer Eppley
Cells
Water Bath Standard Cells
(Electrodes)

Figure 1. Diagram of the temperature-controlled emf measurement system.

250
 '

laboratory. The communication between the laboratory and the computer interfaces is accomplished
by a multi-level time-sharing scheme. The laboratory interface contains the logic circuitry
j
necessary to connect as many as four experimental systems to the digital data bus. Data
that are routinely collected by the system are stored on a random-access disc for a period
: of several days. If space on the disc becomes scarce, the data files are read onto a
master archive magnetic tape for storage.

! In performing a pH certification, the first requirement is to prepare the buffer solutions

and measurement cells with utmost care. As specified in the pH certificates, the carefully
I

dried and weighed salts are dissolved in distilled water of sufficient purity to have a
conductivity of less than 2 x 10" 6 S'cm~i (ohm"i.cm"^) at 25 °C. For buffers in the neutral
and basic regions, carbon dioxide must also be removed from the water prior to dissolution
of the buffer salts. The hydrogen gas is purified of oxygen by passage through a catalytic
reduction tube or through a palladium purifier.

j
The specially designed emf measurement cell shown in figure 2 consists of a hydrogen
i
gas electrode compartment, a silver/silver chloride electrode compartment, and a series of
gas-dispersion compartments for humidifying the incoming hydrogen gas by passage through the
i

I
buffer solution. Details of the preparation of the platinized (or palladized) platinum and
the silver/silver chloride electrodes can be found in references [20] and [21].

Figure 2. Emf measurement cell used for the certification of pH buffer solutions.

251
 The prepared measurement cells, usually consisting of two cells containing each of the
three levels of added chloride, are placed in the controlled temperature water bath as shown
in figure 3. Also shown in this figure are the coiled copper tubing hydrogen gas inlets and
the two platinum resistance thermometers.

Figure 3. Emf measurement cells in position in the control led-temperature water bath.

Figure 4. View of the measurement facility including the water bath, instrument rack and
laboratory logic box interface (open on rear wall).

252
 The complete measurement system, with the laboratory logic box interface (open) on the
rear wall, is shown in figure 4. Another view, figure 5, shows the entire electronics rack
and the thermostated standard cells on the laboratory bench top. Briefly, the equipment in
the electronics rack consists of (from the top): A digital barometer which is automatically

Figure 5. Close-up view of the

instrument rack with thermo-
stated standard cells on the
laboratory bench to the right.

monitored during the certification measurements is used to correct the partial pressure of
hydrogen for variations in the atmospheric pressure; the autoranging and autofunctioning
digital voltmeter which serves as the analog-to-digital converter for both resistance and
emf measurements; the proportional temperature controller which, in conjunction with the
temperature set point selector (next panel down) and the bath refrigeration unit, maintains
preset bath temperatures to better than ± 0.01 °C by means of immersion heaters; and the DVM
input multiplexer which, under computer control, switches the various measurement parameters
into the DVM for A-to-D conversion prior to transmission to the computer for acquisition and
storage. The functions multiplexed include: temperature (platinum resistance), standard
resistor (calibration), standard emf cells (calibration), and up to ten measurement cells.
The control console permits communication with the computer via a series of thumbwheel
switches which enter all of the required input parameters, e.g., temperature sequence,
number of cells, time delays, and program identification numbers. The control console also
includes a set of pushbuttons which are used to set up, start (send), and terminate the
experiment, as well as indicator lights which signal operations (data acquisition, operate,
auto, ete.) or problems (error, reject). A vibrating-reed electrometer is located below the
control console and, in conjunction with the three-position pH switch, provides a high-
impedance input to the DVM for measurements with glass and high-resistance ion-selective
electrodes. The remainder of the equipment in the rack constitutes the Mueller bridge
system for manual checking of the water bath temperature via the second platinum resistance
thermometer shown in figure 3.

253
 )

In operation, computer control begins by automaticany positioning the temperature set

point selector at the starting sequence number. When the bath temperature achieves the
nominal value, within certain preset control limits as indicated by the platinum resistance
thermometer, the input multiplexer automatically switches through the readout positions.
After a programmed time delay, the multiplexer recycles through the positions and compares
the values to the previous sets of data. After three cycles, if the data agree within the
requisite control limits, the computer sets the temperature control to the next sequence
value and the operation is repeated. If any of the values exceed the control limits, an
out-of-range message is printed next to the incorrect value and, following two more cycles,
the computer proceeds to the next sequence temperature after again flagging the out-of-range
value. Depending on the temperature value or the direction of the temperature change, the
computer turns on the refrigeration unit or an auxiliary heater. After the temperature
sequence is completed, the routine is automatically terminated.

Reduction of the acquired data is performed by batch operation on the NBS central
computer, a Univac 1108. The data reduction consists of a series of operations including
correcting the experimental emf values to the standard partial pressure of hydrogen, calculating
the acidity functions, extrapolating to zero molality of added chloride to obtain the limiting
acidity functions, evaluating the chloride ion activity coefficient using the Bates-Guggenheim
pH convention, and finally calculating the pa^ values. These experimental values are then
smoothed with respect to temperature by the method of least squares to give the certified
pH(S) values.

References

[I] Bower, V. E. , Paabo, M., and Bates, R. G., Standard for the measurement of the pH of
blood and other physiological media, J. Res. Nat. Bur. Stand. 65A, 267 (1961).

[2] Durst, R. A. and Staples, B. R. Tris/Tris -HCl

, A standard buffer for use in the
:

physiologic pH range, Clin. Chem. 18^, 206 (1972).

[3] Paabo, M., Acidimetric Measurements, Chapter 2 in Eleatroohemioal Analysis Section:

Summary of Activities July 1969 to June 1970, R. A. Durst, ed. , NBS Technical Note
543 (U. S. Government Printing Office, Washington, D. C, Nov. 1970).

[4] Ladenson, J. H., Smith, C. H. Dietzler, D. N., and Davis, J. E., Use of tris buffers
for quality control of blood pH, Clin. Chem. 20, 1337 (1974).

[5] Good, N. E., Winget, G. D. Winter, W.

, Connolly, T. N., Izawa, S., and Singh, R. M. M.,
,

Hydrogen ion buffers for biological research, Biochem. 5^, 467 (1966).

[6] Bates, R. G., Roy, R. N., and Robinson, R. A., Buffer standards of tris(hydroxymethyl
methyl gyl cine ("Tricine") for the physiological range pH 7.2 to 8.5, Anal. Chem. 45,
1663 (1973).

[7] Durst, R. A., Staples, B. R. , and Paabo, M., Activity standards for ion-selective
electrodes, in Biological Aspects of Electrochemistry, Experientia Suppl. 1_8, 275
(1971).

[8] Bates, R. G., Staples, B. R. , and Robinson, R. A., Ionic hydration and single ion
activities in unassociated chlorides at high ionic strengths. Anal. chem. 4^, 867
(1970).

[9] Stokes, R. H. and Robinson, R. A., J. Amer. Chem. Soc. 70, 1870 (1948).

[10] Bates, R. G. , Ion activity scales for use with selective ion-sensitive electrodes. Pure
Appl. Chem. 36^, 407 (1973).

[II] Durst, R. A., Analytical techniques and applications. Chapter 11 in lon-Seleotive

Electrodes, R. A. Durst ed., NBS Special Publication 314 (U. S. Government Printing
Office, Washington, D. C, 1969).

254
[12] Blum, R. and Fog, H. M., Metal buffers as standards in direct potentiometric determination
of metal ion activities, J. Eleotroanal. Chem. 34, 485 (1972).

[13] Hansen, E. H., Lamm, C. G., and Ruzicka, J., Selectrode - the universal ion-selective
solid-state electrode: Part II, Anal. Chim. Acta, 59, 403 (1972),

[14] Ruzicka, J, and Hansen, E, H., Selectrode - the universal ion-selective electrode: Part
IV, Anal. Chim. Acta, 63, 115 (1973).

[15] Ruzicka, J., Hansen, E. H. and Tjell , J. C. , Selectrode - the universal ion-selective
electrode: Part VI, Anal. Chim. Acta, 67, 155 (1973).

[16] Havas, J., Kaszas, M. , and Varsanyi , M. , Radelkis ion buffer solutions. Bung. Soi.
Instr. 25, 23 (1972).

[17] Bailey, P. L. and Pungor, E., The calibration and response of ion-selective electrodes
at low concentrations of primary ions. Anal. Chim. Acta, 64, 423 (1973).

[18] Mohan, M. S. and Bates, R. G. , Calibration of ion-selective electrodes for use in

biological fluids, Clin. Chem. 21_, 864 (1975).

[19] Durst, R. A., Standard Beferenoe Materials: Standardization of pH Measurements, NBS

Special Publication 260-53 (U. S. Government Printing Office, Washington, D. C, Dec.
1975).

[20] Bates, R. G. , Determination of pH, Theory and Practice, 2nd edition (John Wiley and
Sons, New York, 1973).

[21] Ives, D. J. G. and Janz, G. J. eds.. Reference Electrodes (Academic Press, New York,
1961).

255
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

USE OF CARBON DIOXIDE- AND OXYGEN TONOMETERED PHOSPHATE-BICARBONATE-

CHLORIDE-GLYCEROL-WATER MIXTURES FOR CALIBRATION AND CONTROL
OF pH, pCOz, AND pOz ELECTRODE SYSTEMS

A. H. J. Maas, A. H. Veefkind,! R. A. M. Van den Camp,

'A. J. Teunissen, A. B. T. J. Boink, and T. J. C. Ruigrok
Department of Cardiology and Thoracic Surgery
University Hospital
Cathari jnesingel 101
Utrecht, The Netherlands

The desire for simultaneous calibration of pH, pC02, and p02 electrode systems of
modern blood gas analyzers and the need for adequate quality control for this equipment
stimulated this investigation. The most important difficulties to overcome were: (1) estab-
lishing the pH, pCOa-relationship of the calibration solution, and (2) systematic bias between
the response of gas electrodes, in different media.

1. Calibration Solutions

In this study (for details see reference [1]^), we started from phosphate-bicarbonate-
chloride solutions equilibrated with carbon dioxide gas for which we were able to calculate
a pH, log pCOa-relationship based on the law of electroneutral ity and the definition of the
ionic strength, using the practical ionization coefficients of phosphoric acid from Bates [2],
and of carbonic acid from Maas [3]. For the NBS equimolal phosphate buffer (0.025 mol/1) to
which NaHCOa and NaCl were added in a total concentration of 0.060 mol/1, we found in the
pH range 6-8, at 37 °C the following equation set:

0.185 - I
pH = 7.1834 - 1.34207 /f, + 0.84361^ + a - log
^ _ q 135
a

pH = 10.806 - 0.4662 /f^ + a - log pCOa + log (0.160 - + y) (2)

in which is the ionic strength on a concentration basis, y the added amount of NaHCOa
and a the factor wherein the deviation of the practical ionization coefficients defined
in other ionic and medium systems is discounted.

Subtracting eq (2) from eq (1) results, after rearrangement, in a relationship between

pC02 and I^:

(0-185 - I^)(0.160 - + y)
.0 + (3.6226 + 0.8758/f - 0.84361 ). (3)
pCU2 + X 10 a o
(I - 0.135)
a

This means: by equilibration of buffer solutions of this kind, with gas mixtures of known
composition, that is to say a given pC02, the ionic strength I is fixed, and could be
estimated by eq. (3). In this manner, pH-log pC02 equil ibratiSn lines of buffer solutions

^Department of Medical Physics, Free University, Amsterdam, The Netherlands.

^Figures in brackets indicate literature references at the end of this paper.

257
with different sodium bicarbonate concentration can be predicted from eqs. (1) and (2),
on the condition that a is known. The factor a could be determined from eq. (1) or (2)
by pH measurements of equilibrated buffer solutions at a given pC02. From figures 1 and
2, it is demonstrated evidently that introduction of an empirical factor a leads to a good
agreement.

6.6 6.8 7.0 7.2 7.4 7.6 7.8 8.0; pH

Figure 1. pH-log pC02 equilibration lines of four buffer solutions with 0, 15, 30 and 45
mmol/1 sodium bicarbonate, respectively, calculated by means of eqs. (1) and (2) without
correction factor a; the dots are the measured pH values of the same buffer solutions
equilibrated with gas mixtures of 2, 4, 8 and 15 percent CO2, respectively.

Figure 2. pH-log pC02 equilibration lines, differing from each other by 5 mmol of sodium
bicarbonate per liter, calculated from eqs. (1) and (2) by substituting a. The dots are
the same measuring points as in figure 1.

It is known from the literature [4] that whole blood and glycerol /water (3/7 by
volume) mixtures at the same oxygen tension produce nearly equal diffusion currents as a
consequence of their similar viscosity. In order to minimize differences between blood
samples and calibration solutions, we added glycerol to the buffer solutions. In aqueous
solutions the correction term (a) was dependent on y:

a = 0.0331 + 1.513y - 8.44y2 - 118. 5y3. (4)

In the glycerol /water (3/7 by volume) solutions, the correction term (b) was dependent on

258
I as well, caused by the medium effect:

6 = (-5.667 + 290.999y - 3113. 55y2 - 4573y3)j

(5)
+ 0.8700 - 39.343y + 317.56y2 - 2459. 3y 3.

By means of eq. (5) substituted in eqs. (1) and (2) (read a = 3), pH-log pC02 equilibration
lines of glycerol /water (3/7 by volume) buffer solutions with different sodium bicarbonate
concentration can be predicted.

So we were able to prepare suitable calibration solutions of desired pH, pC02 and
p02 values, by equilibrating a solution of 25 mmol of the phosphates, 30 mmol sodium
bicarbonate and 30 mmol sodium chloride per liter glycerol/water (3/7 by volume) mixture
with 4 percent CO2 in air and 8 percent CO2 in nitrogen, respectively. Various pH, pC02,
and PO2 values, depending on the barometric pressures, are given in the next calibration
table 1.

Table 1. Calibration values obtained by equilibration of above mentioned

buffer solution gas mixtures 1 and 2 at 37 °C over the
barometric pressure (B) range 740-780 mmHg.

Gas mixture 1 Gas mixture 2

4% CO2 and 20.09% O2 8% CO2 and 0% O2

B pH pH
PCO2 PO2 PCO2 PO2

mmHg mmHg mmHg mmHg mmHg

740 7 420 27.7 139.2 7 226 55.4 0.0

745 7 418 27.9 140.2 7 224 55.8 0.0

750 7 416 28.1 141.2 7 222 56.2 0.0

755 7 414 28.3 142.2 7 220 56.6 0.0

760 7 412 28.5 143.2 7 218 57.0 0.0

765 7 410 28.7 144.3 7 216 57.4 0.0

770 7 408 28.9 145.3 7 214 57.8 0.0

775 7 406 29.1 146.3 7 212 58.2 0.0

780 7 404 29.3 147.3 7 210 58.6 0.0

The usefulness of this calibration buffer system was checked by measuring pH, pCOa
and PO2 values of tonometered whole blood on the Corning M 165 blood-gas analyzer, both by
the conventional calibration methods (buffer and gases) and by using the equilibrated
solutions. The measurements of pH and pC02 were at least comparable, whereas p02 measure-
ments were substantially improved by minimizing the gas-blood difference using calibration
liquid of broadly the same viscosity as blood (see fig. 3).

The accuracy of this calibration method depends on the accuracy of the phosphate
buffers, against which the pH values of the calibration solution is operationally derived
(±0.005 pH), and the accuracy of the composition of the gas mixtures (±1% of the pC02
and PO2 value).

259
 Figure 3. Relationship between measured pOz and calculated pOz- Continuous lines are the
calculated regression lines for y on x and the broken lines are the lines of identity;
(a) gas-calibration method; (b) buffer-calibration method.

2. Approach to Quality Control

The proposed buffer system not only opens new perspectives for simultaneous calibration
of pH, pC02, and pOa electrodes, but it also is a step in the direction of quality control,
as indicated by Noonan and Burnett [5]. Realization can be thought by the following line:

(a) Intralaboratory comparison of blood gas equipment can easily be realized by

offering samples from the same thermostated tonometer setup, to the different apparatus.

(b) Interlaboratory comparison would be possible in principal by sending around

ampoules containing a control solution for simultaneous control of pH, pC02, and p02.

In our routine laboratory, we have at our disposal three blood-gas apparatus: from
Radiometer, Models BMS2-MK2, and ABL-1; from Instrumentation Laboratory, Model IL-413.
For quality control of the BMS2-MK2, we measure a few times a week the pH of buffer solu-
tions, marked A III and C III (composition: 0.025 mol/liter Na2HP0[,, 0.025 mol/liter
KH2PO1+, 0.030 mol/liter NaHCOs and 0.030 mol/liter NaCl using, respectively, water (A III)
and glycerol/water (3/7 by volume) mixture (C III) as solvent) tonometered with gas mixtures
of 4 percent CO2 and 8 percent CO2 in air. In figure 4, the fluctuation of the bias be-
tween the readings and the stated values over a two-month period is presented.

We also obtained initial experience with quality control of the ABL-1 and IL-413 with
the same buffer solutions equilibrated in a special setup consisting of one gas mixing
pump, delivering two gas mixtures, and a thermostated tonometer vessel. The results of
twelve measurements performed on various days are summarized in table 2.

The pH differences [ca. 0.012 pH) could be explained partly by a difference of 0.008
pH between the ABL-1 and IL-413 calibration buffers and the Radiometer precision buffer
(pH = 7.383), and partly by the type of salt bridge: ABL-1, 20 percent KCl ; IL-413,
saturated KCl with a membrane instead of an open junction.

The PCO2 differences (table 2) are equal for both instruments and do not deviate
significantly from the stated value for the lower pC02 value in contrast to the measured
high PCO2 values which are ca. 3-5 percent lower. Comparison of results obtained with
both instruments using tonometered blood samples revealed a relationship between the
measured and calculated pC02 values as presented in figures 5a and 5b. The regression
lines cross the identity lines near the lower calibration point. The main cause of de-
viation from identity seems due to the "memory effect" of pC02 electrodes as recently
described by Berkenbosch [6], Cramption Smith [7], and Hahn [8]. They show also that this
effect is reproducible and corrections can be made for it.

260
 CO
O o
o CTl ID CO

LO
o CO o
CO
Ln I

CM
O ID
O
o II II

o o
o CM vo o 00 CO
O o o CO o vo
CM
o
Q.

(M Ln
o CD
3: o o
o 1-^ o CO I—
00
CO o E o O O CM I— Ln
Lf) ID
OO
1^ CO
CO
o II

o o
o CO cri o o CO
o o o o o o LD o
Q.
oQ. oQ.

o o
o un I 00
C-J
E o CO 1— 00
CM
I—
CM 1

CO I

o
O II II o
o o
o VD o
o CM
CM
LD
Q- I

r—
1—
o o
>vl-

o p- 1^ CM 00 LD un 00
CO
Ln
O C3 O I— LD Cvj ^ I—
00
00 +
CM
Ln
1^ CO
o o
oCM II
o o CO CTl CM CM cyi Ln
o o o o o CD I—" CM o i-^
o
a. I

•4->

fO
C
CM o o i/> cn
OJ
+J p B
r— B CD
JO
"o
«/)
o
Q.
> o
vo
(/)
3
E
E T3 r-. +->
s- o QJ ro
o oCM
cu
H-
4-
+->
S-
II S-
(O
<_)
OCM
C_) Oa.CM Q oD.CM Q
to
n3
fO
!->
S-
ro
Q.
a. Q- Q
OO
Q.
<
Q
OO
CL
<1
Q
00 00 < OO
CQ CD 00 CQ

261
 a ;

BMS 73-131. AOO-5357-3690

0.01- A pH
\ /
+o X / °
\ A HI 4

A > A
^A —
/\ /

—A — A^

0.01-

0.01 —I —25— —
I
V I
i-i-i ——
I I- -I ————— — ——
I I
I I
c

I
8

I
I — ———— ———
I 1
1
I I
- I I I

24 26 27 1 2 7 9 11 15 16 17 22 24 25 Ti
28 1 6 12 13 U 15 21 23 27
march I april mai

Figure 4. Tonometry results for pH, with BMS2-MK2 (see text). The mean differences and
standard deviations are: A III 4: 0.0054 ± 0.0045; A III 8: 0.0030 ± 0.0044; C III 4;
0.0026 ± 0.0046; C III 8: 0.0040 + 0.0049 which means that the precision is very high.

Figure5. Relationship between pCOs of blood and gas: (a) measurements with ABL-1
y = 0.9115X + 2.1091; (b) measurements with IL-413; y = 0.8822x + 4.2134. Each dot
is the mean of ten measurements.

Concerning the p02 measurements (table 2), we found that the zero point could not
adequately be established by a single filling of the cuvette probably caused by mixing
with contaminants remaining behind in the chamber after a measurement and oxygen loss from
the electrode. Values of aa. 30 mm Hg were read.

262
 Using A III solution, the high p02 values agree with the stated value for the ABL-1
and were significantly lower for the IL-413; using C III solution, the p02 readings are
lower for both instruments. Also we evaluated the p02 electrode systems of both instruments
using tonometered blood samples. The results are illustrated in figures 6a and 6b. There
is a tendency for the blood-gas regression lines to have slopes lower than those due to
gas for both instruments. In the physiological range the difference between the value for
gas and tonometered blood is a few percent, but at higher oxygen tension readings are
significantly too low. We feel that there is a need to improve the method of cali'^ration
of blood-gas analyzers.

7oo-p02 measured
mmHg

600..

HO- p02 measured

mmHg
120

p02 calculated mmHg

100 200 300 iOO 500 600 700 20 40 60 80 100 120 UO

700- p02 measured

mmHg

600-
UO- p02 measured
mmHg
500-
120-

100

300

200
L-413

100-

p02 calculated mmHg

-I- -I
100 200 300 400 500 600 700 20 40 60 80 100 120 140

Figure 6. Relationship between p02 of blood and gas: (a) measurements with ABL-1;
y = 0.9682 + 2.2981; (b) measurements with IL-413; y = 0.9804 - 1.0731. Each dot
is the mean of ten measurements.

263
 Introductory in two trials, we have sent around 100 ml buffer solutions in plastic
bottles to approximately ten laboratories with the suggestion to equilibrate these solutions
with gas mixtures of CO2 content and to measure pH. So this method was restricted to
setups of the indirect pC02 method according to Astrup and Siggaard-Andersen. In figure 7
the results of these trials are reflected.

0.03 ApH
CIII - TRIAL I
0.02

0.01

"o.oi

0.02-

0.03

0.04

0.05
pH
7.0 7.1 7.2 7.3 7.4 7.5 pH 7.6
0.03- ApH

0.02 CIII - TRIAL II

0.01
4^
0

0.01

0.02-

0.03-

.pH
7.2 7.3 7.4 7.5 7.6
0.03- ApH

0.02- AIII - TRIAL II

0.01
+ ^
0

0.01

0.02-

0.03

~' 1 7 ! 2 7 ! 3 7:4 7 7: 6

Figure 7. Two trial results of pH determined in CO2 equilibrated A III and C III buffer
solutions. Different symbols are from different laboratories. ^

264
 In reference to these figures, we may make the following remarks:

(1) Rather good agreement {ca. 0.01 pH) in the physiological pH range is found by most
of the participants.
(2) Improvement of the results at the second trial, probably caused by experience and
sending aroun-d a precision phosphate buffer for calibration as well. Dramatic was the dis-
covery of pH deviation of 0.011 unit between the precision buffer sent along, and a buffer
from the same firm and with the same batch number, present in one of the laboratories. Up to
now the explanatory suggestions are still unproved.
(3) Differences up to 0.02 pH unit for equipment in the same laboratory. See the
closed and open squares in figure 7a.
(4) Possibility of signaling a wrong analysis of a gas cylinder. See open square at
the bottom of figure 7b.
(50 Possibility of detecting small difference in gas mixing pumps. See open and closed
triangles in figure 7b.

In an attempt to eliminate differences in the equilibration setup, and to simplify

the procedure, we are engaged in an investigation into the possibility of additives to the
buffer solution, which are able to absorb oxygen. To such a solution, with enhanced
capacity, the desired carbon dioxide and oxygen tension can be introduced. The only thing
to be done, prior to pH, pC02 and p02 measurement, is to bring the solution to 37 °C
before opening the sample holder. Preliminary results with a fl uorocarbon because of its
,

high dissolving capacity for gases, showed usefulness for some equipment and uselessness
for others.

Most likely there will still be many obstacles on the way to quality control, however,
the purpose is worthwhile continuing this way.

References

[1] Veefkind, A. H., Van den Camp, R. A. M. and Maas, A. H. J., Use of carbon dioxide - and
,

oxygen tonometered phosphate-bicarbonate-chloride-glycerol -water mixtures for calibration

and control of pH, pC02 and p02 electrode systems, Clin. Chem. 2J[, 685 (1975).

[2] Bates, R. G., and Acree, S. G. pH of aqueous mixtures of potassium dihydrogen

,

phosphate and disodium hydrogen phosphate at 0 °C to 60 °C, J. Res. Nat. Bur. Stand.
34, 373 (1945).

[3] Maas, A. H. J., van Heyst, A. N. P., and Visser, B. F. , The determination of the true
equilibrium constant (pi^ig) and the practical equilibrium coefficient (pA'jg) for the
first ionization of carbonic acid in solutions of sodium bicarbonate, cereBrospinal
fluid, plasma and serum at 25 and 38 °C, Clin. Chim. Acta 33^, 325 (1971 ).

[4] Heitman, H., Buckles, R. G. , and Laver, M. B., Blood p02 measurements: performance of
micro-electrodes, Respir. Physiol. 3_, 380 (1967).

[5] Noonan, D. C. and Burnett, R. W.

, Quality-control system for blood pH and gas measure-
,

ments, with use of a tonometered bicarbonate-chloride solution and duplicate samples of

whole blood, Clin. Chem. 20, 660 (1974).

[6] Berkenbosch, A., Measurement of carbon dioxide tension in poorly buffered solutions
with the pC02 electrode, Pflugers Arch. 318, 217 (1970).

[7] Crampton Smith, A., and Hahn, C. E. W., Studies with the "Severinghaus" pCOo electrode
I: Electrode stability; memory and S plots, Brit. J. Anaesth 47^, 553 (1975)'.

[8] Hahn, C. E. W. , and Crampton Smith, A., Studies with the "Severinghaus" pC02 electrode II:
CO2 measurement using a single control analyzer, Brit. J. Anaesth 47, 559 (1975).

265
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

CALIBRATION OF BLOOD GAS ANALYZERS

Alan H. Runck
Corning Scientific Instruments
Medfield, MA 02052, USA

Blood gas measurements are unique in the field of medical laboratory analysis. They
are frequently requested at a point when the subject patient is in a life-threatening
situation which will be dealt with according to the results obtained from these measure-
ments. The techniques that are used by experienced blood gas technicians for collection
and storage of samples and carrying out of measurements are complex but highly standardized:
anaerobic sample handling, "double insertion" of the sample into the cuvette, monitoring of
cuvette temperature and waiting for the proper end-point to be achieved are good examples
of the precise operator technique required. The equipment used by virtually every medical
laboratory to perform blood gas analysis is basically identical (glass electrode system for
pH, Severinghaus electrode for PCO2 and Clark electrode for PO2), with variations in the
volume of blood sample required for analysis, design of supporting hardware and degree of
automation. Yet, for all the similarities in the methodology for measuring blood gas, the
means for calibration of blood gas analyzers are not uniform.

Classically, blood gas measurements have been calibrated using dilute aqueous phosphate
buffers for pH [1]^ and analyzed gases for PCO2 and PO2 [2]. Many modifications of cal-
ibrating materials however, have been advocated in the literature and are in limited
current use. For pH measurements, calibrating buffers containing added sodium chloride
(saline buffers) have been described [3], as well as bicarbonate-based buffers [4]. Gases
used for calibration of PCO2 and PO2 measurements are not of standard composition, but vary
from laboratory to laboratory, especially with respect to oxygen; some laboratories even
use a solution such as sodium sulphite [5] in lieu of an oxygen-free gas for zero cal-
ibration of PO2. The calibrating gases themselves may not, depending on the method of
formulation, contain the composition of gas that is stated on the accompanying label.
Finally, use of tonometered liquids for calibration of PCO2 and PO2 measurements has been
repeatedly reported [4,6,7] and is in routine use in a growing number of laboratories [3].

Why such diversity in calibration of instruments when basic design of the instru-
mentation is the same? The common reason in the literature seems to be: (1) a search for
more reliable and accurate performance, and (2) a search for convenience. Modification of
pH calibrating buffers has occurred for both reasons. Use of saline buffers coupled with
use of a sodium chloride salt bridge is claimed to improve accuracy by reducing the sodium
response of different pH electrodes, and improve reliability by reducing depletion of the
salt bridge solution, as well as improving the stability of the liquid junction potential
[3]. Use of the bicarbonate buffering system is more convenient, since one aqueous solution
win serve as a pH, PCO2 and PO2 calibration point. Change in the manner of calibrating
PCO2 and PO2 measurements has been attempted for the purpose of improving accuracy and
reliability of these measurements. Differences in the PCO2 and PO2 readings between gases
and liquids having identical gas tensions have been reported; for this reason, use of
tonometered liquids having known values for PCO2 and PO2 has been investigated by others as
an alternative to calibration using gases alone.

A major area of controversy in blood gas has been the question of whether changing the
method of calibration will improve performance. PO2 measurements have been singled out as
being most in need of improvement, because of the tendency of the PO2 sensor to read lower
on liquid than on gas samples. Maas and coworkers [8] summarized the concern about the

Figures in brackets indicate the literature references at the end of this paper.

267
accuracy of PO2 measurements by presenting data concerning the average PO2 readings obtained
when blood samples having known PO2 values were read on conventional blood gas instrumen-
tation (table 1). The results showed a bias, the sign and magnitude of which were dependent

Tablp 1 Rlnnri/na*^ di'f'fprpnrp in

III Pn^ mpfl irpmpnl"
rijp IflCUOUi CMldIL
<^ 1 (?\fff^r
^ai M;^;ic;
Lci riudO) A
M ,
H
n. u.
.1 V\kz\ Lcllb 9 r . U • 9

Clin. Chim. Acta, 28, 443 (1970)).

O2 PO2 PO2 Correction

(%) (mm) (mm) (%)

5 JO JO 0 . - 1 . /

10 / u /I.I -1.0
20 140 132.8 5,1
40 280 269 3.9
60 420 399 5.0
80 560 521 7.0

on the level of PO2 in the sample being measured. Prior to Maas' studies, the bias was
assumed to be a constant percentage of the sample PO2; it was thought to be caused by the
difference in diffusion coefficient in gas calibrant vs. liquid samples and, therefore,
able to be corrected for in the final result. Maas' data, however, refuted this.

Data by Hulands and coworkers [9] showed another aspect of PO2 performance that was
equally disturbing (fig. 1). They showed that the magnitude of the PO2 bias varied not

1.20

_ 1.15

d
CO
+1

1 1. 10

J
T3
O
m 1.05

1.00
14 18 22 26 30 2 6 10 14 18 22
June July
Figure 1. Day-to-day variation in PO2 bias (from Hulands et at. [9]).

only according to sample PO2, but also varied on a day-to-day basis that was unpredictable.
Hulands felt that this was caused by the gas/liquid correction factor for the PO2 electrode
changing from day to day; he proposed to remedy this by calibrating the electrode directly
with tonometered 30 percent glycerol solutions (fig. 2).

Our studies have centered on evaluating the clinical results obtained when gases vs.
liquids are used for calibration of commercial blood gas analyzers. It seemed that one
variable that was not able to be controlled in the studies of Maas and Hulands was that of
sample preparation; blood tonometry and sample handling at high levels of oxygen tension is

268
 1.20 t \
^
\
^
r
Glycerol <^ = 0.897 (blood <^) + 0.115
r = 0.902

1.15

CD

1.05
• Air Figure 2. Paired measurements
o of 30 percent glycerol
High PO2
correction factor and whole
X
Low PO2 blood correction factor (from
Hulands et al. [9]).
1.00 I

.00 1.05 1. 10 1.15 .20

Whole blood
difficult. Therefore, we decided to look directly at differences between instruments
calibrated with gases and those calibrated with liquids. The two types of systems were set
up side-by-side in the Blood Gas Laboratory at Peter Bent Brigham Hospital and used to
measure blood gas values on patients' samples. The results, shown in figure 3, indicate
that the variable bias described by both Maas and Hulands for PO2 is not observed. Compar-
ison data, where both instruments were calibrated with tonometered liquids, are shown in
figure 4. The PCO2 measurements that were observed in this study are shown in figure 5.

500
X P02(gas col.) = 0.98xP02( liq.col.)

E S.D. = 5.7mnnHg

n = 55 samples
£=

a 400

8
CO
a
P 300
13

/
CD /
£
Instrunnent
13
CO
200
O o A (June 73)
cu
E J- • B Feb 74)
(

cvj
Figure 3. Comparison of gas
A C Feb 74)
o
Q_
(
vs. liquid calibration for
blood PO2 measurement.
100
100 200 300 400 500
PO2 measurements using liquid calibration (mmHg)

269
 -

500

P02.A = 1.01xP02,B
S.D. = 12.5mnnHg
X 400 n = 99 samples
/
E .

E
<
CD
300
E
3

200
c
X5
O
q;

CM
o
a. 100 Figure 4. Comparison of liquid
vs. liquid calibration for
blood PO2 measurement.

100 200 300 400 500

PO2 reading, instrument B (mmHg)

PCO2 (GAS CAL)-: 1,01 XPCO2 (LIQ CAU

REGRESSION STD. DEV. • 1.6mm

r = 0.99 2 Figure 5. Comparison of gas vs.

liquid calibration for blood
n' 45
PCO2 measurement.

30 40 50 60 70 80
PCO^ reading, using liquid calibration
(mm Hg)
The data indicate that gas calibration of blood gas analyzers will yield PO2 and PCO2
values that are as accurate and reliable as those obtained when liquids are used for cal-
ibration. These results have been confirmed by more recent data in our laboratories on
different types of equipment, and by Maas, who recently compared results obtained with
instruments calibrated with gases and instruments calibrated with tonometered glycerol
water solutions [4].

270
 It is intriguing to ponder why earlier data showed such a pronounced variation in
results, both in terms of PO2 dependent bias and in terms of a day-to-day variation in
accuracy. We addressed this question by measuring the PO2 of tonometered water solutions
using a gas calibrated analyzer, and comparing these measurements with those obtained
indirectly from oxygen content determinations on the same samples, using the Lex-Oa-Con
oxygen analyzer. The results, shown in figure 6, are interesting: the oxygen content

500 n 71
• PO2 O2

400
calculated from

o PO2 measured
content

^
/
§ 7.6%

X 300
E 8.6%
E
/
g 200

P 4.1 %
100
1.0%

0 1
0 2 3 4
O2 content (m£/dl)

Figure 6. Comparison of measured PO2 of tonometered water vs. PO2 derived from oxygen
content measurements on the same samples.

measurements on the samples at all PO2 levels show that the tonometer performed as expected,
with each sample giving the anticipated PO2 value. However, PO2 measurments on the same
samples, using a blood gas analyzer, show the very type of P02-dependent bias as was reported
by Maas. The conclusion from these data is that the variable bias reported by Maas is a
function of atmospheric contamination, either during sample handling or instrument manip-
ulation. Such contamination has been observed for every instrument studied and is man-
ifested when recovery studies on tonometered blood or water are attempted. These conclusions
should not imply that a gas/liquid correction for PO2 measurements does not exist, but
rather, that the correction factor remains a constant over a period of time. In our experience,
the correction factor for present day blood gas analyzers is 5 percent, and varies no more
than ± 1 percent. What is most important, is for manufacturers of blood gas instrumentation
to explicitly state the accuracy of their instrumentation at all levels of PO2 and PCO2 so
that users can effectively evaluate the performance of the total system without placing all
the blame for poor recovery of PO2 on a "high gas/liquid correction."

Calibration of blood gas instrumentation should be as simple and error free as possible,
without compromising performance. Our studies, and those reported by Maas [4], indicate
that gas calibration for PCO2 and PO2 measurements provides results which are comparable to
those obtained when liquid calibration is used, when accuracy is used as the guideline.

References

[1] Bower, V. E., Paabo, M., and Bates, R. G., A standard for the measurement of the pH
of blood and other physiological media, J. Res. NBS, 65A , 267 (1961).

[2] Severinghaus, J. W., and Bradley, A. F., Electrodes for blood PO2 and PCO2 deter-
minations, J. App. Physiol. 1_3, 515 (1958).

271
Drinker, P. A., Noonan, D. C, Ramanaiah, N., and Tole, J. R., Use of a sodium
chloride-phosphate buffer for pH standardization in a new blood gas analyzer with
an isotonic sodium chloride bridge, Clin. Chem. 1_9, 1243 (1973).

Veefkind, A. H., Van den Camp, R. A. M., and Maas, A. H. J., Use of carbon dioxide
and oxygen tonometered phosphate-bicarbonate-chloride-glycerol -water mixtures for
calibration and control of pH, PCO2 and PO2 electrode systems, Clin, Chem. 21_,
685 (1975).

Heitmann, H., Buckles, R. G., and Laver, M. B., Blood PO2 measurements: performance
of microelectrodes, Resp. Physiol. 3^, 380 (1967).

LeFevre, M. E., Calibration of dark oxygen for use in aqueous solutions, J. App.
Physiol. 26, 844 (1969).

Gleder, R. L., and Neville, J. F., Commentary on PO? and PCO2 electrode calibrations,
Amer. J. Clin. Path. 55, 325 (1971).

Maas, A. H. J., and Mertens, P. J., The measurement of the PCO2 and PO2 of blood with
electrodes in an open cuvette system, Clin. Chim. Acta, 28, 443 (1970).

Hulands, G. H., Nunn, J. F., and Paterson, G. M., Calibration of polarographic

electrodes with glycerol water mixtures, Brit. J. Anaesth. 42, 9 (1970).

272
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

QUALITY CONTROL

S. Raymond Gambino
Columbia Presbyterian Medical Center
New York, New York 10032, USA

In my discussion of instrument specifications I indicated that better quality control

(Q.C.) at individual sites was needed more than quality control of manufacturers. In fact,
I can state categorically that without some type of on-site Q.C. program it is impossible
to provide physicians with reliable blood gas measurements. Too many things can go wrong.

Attending physicians provide some degree of Q.C, but it is insufficient. If the data
agree with what the attending physician expects, then he or she is satisfied. If the data
provided by the blood gas instrument is not consi-tent with his or her expectations, then
the physician does not believe the data. In this situation, some type of readily available
referee sample is essential. Since blood gas measurements are made 24 hours a day, 7 days a
week, the ideal Q.C. material would be available for immediate use at all times. Tonometered
blood is, therefore, inadequate as an "instant" control for peripheral and central blood
gas instruments around the clock. Tonometered blood, on the other hand, is the proper
sample for use as a secondary standard.

More than 10 years ago. Dr. Arthur Babson of General Diagnostics developed a formula
for alyophilized serum control for acid-base assays. This lyophilized control is marketed
as "Versatol Acid-Base." In 1970, when we first began our high volume blood gas service, I
received many complaints from physicians during the first month of operation. At that time
we did not use controls on every shift. The supervisor of the chemistry service, Ms. I.
Fonseca, suggested that we run the lyophilized controls on every shift Qvery day of the
week. Once we began to utilize controls, at 3 different levels on every shift, the number
of complaints fell to nearly zero. Now problems are discovered in the laboratory before
results were reported. But the lyophilized controls are not practical for the distant
units operated by non- laboratory personnel. Furthermore, the lyophilized controls lack
assigned values for P02, although a single, but somewhat variable, P02 value for all 3
controls can be obtained by assay.

Following the introduction of our first peripheral blood gas unit it became obvious
that a lyophilized control, or a whole blood tonometered control, was not adequate. What
was required was an ampouled single-use control. Such a control has just been developed by
Dr. James Turner and Dr. Arthur Babson of General Diagnostics. It is marketed as "Blood
G.A.S. Control." Pilot lots of the ampouled controls have been available to investigators
for more than a year. In fact, it was the availability of these unit-use controls for pH,
PCO2, and P02, at 3 different levels, that permitted us to expand our peripheral blood gas
units in 1975.

The new ampouled controls have revolutionized our operation. We are able to leave a
peripheral unit unattended and unvisited by any laboratory personnel for several days as
long as ampouled controls are available at the peripheral site.

What happens when we open an ampoule? Table 1 shows that in 4 minutes there is little
change in pH or PCO2. The P02 data in this experiment cannot be interpreted since the P02
in the ampoule is so close to the P02 of room air.

How reproducible are the values for pH, PCO2 and P02 when repeat measurements are made
on 2 different instruments that are calibrated with the same gases and buffers? Table 2
shows the data for pH. The pH values are very reproducible. Table 3 shows the data for
PCO2. The PCO2 is reproducible, but there is an obvious systematic difference between the
two instruments even though they were calibrated with the same gases. Table 4 shows the data
for P02. P02 is the least reproducible of the measurements at low values. This has to do
with the particular instrument used. The IL 313 introduces a bolus of room air in front of

273
 s . . . . s

Table 1. Effect of exposure to air, Table 2. pH comparison of two IL 313'

mean of 6 ampoules in each (same buffer)
of 3 different laboratories.

Ampoule No. 1 No. 2

Time: zero 30 s 60 s 240 s

1
1
-7
/
/1 00 -7
/ ,
/inn
4 y
1

pH 7.628 7,,629 7.,630 7.634

2 7.,421 7.,416

PC02 66.0 66..2 65.,9 65.7 3 7,,419 7..420

PO2 163.5 163.,8 164,.5 163.2 4 7..417 7..419

5 7,.417 7,.424

Mean 7,.419 7,.420

Table 4. P02 comparison of two IL 313'

Table 3. PCO2 comparison of two IL 313'
(same gas)
(same gas).

Ampoule No. 1 No. 2 Ampoule No. 1 No. 2

1 61 58 1 60 62
2 62 58 2 59 61

3 61 58 3 59 63
4 61 '
58 4 59 60
5 61 59 5 58 59

Mean 61.2 58.2 Mean 59 61

each sample. The controls do not have any buffering capacity for P02, therefore they are
quite sensitive to the presence of oxygen in the instrument at any pressure other than the
pressure in the control.

Are these controls useful when used routinely? Decidedly yes. Table 5 shows typical
actual problems revealed by the control and not revealed by standard gas and buffer cal-
ibration.

Table 5. Typical problems revealed by controls.

Expected Found Problem

PCO2 42 46 Needed new membrane

PCO2 58 62 Needed slope adjustment
P02 100 80 Mold in wash solution
P02 100 Slow Bubble behind membrane
PCO2 58 Slow Space between electrode
and membrane too thick

274
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

QUALITY CONTROL AND STANDARDS

Daniel C. Noonan
Port Huron Hospital
Port Huron, Michigan, 48060, USA

We have heard many different points of view on many aspects of blood gas measurements
at this workshop. There is one point, however, upon which we all probably agree. We all
feel that our own respective laboratories provide blood gas results which are of the highest
quality available with modern technology. While this is undoubtedly true, if we were to
survey our laboratories, we would find a wide range of methods for the basic standardization
and quality control of these measurements in daily use.

This workshop should have several objectives in this area of standards and controls.
First, we should strive to identify the methods and approaches which seem to be commonly
agreed upon to produce accurate standardization and useful control. Secondly, .for those
areas where disagreement exists, we need to discuss the kind of experimental approaches
which might resolve these questions. Lastly, we need to look into our own crystal ball and
make some predictions on how standardization and control should be done in the future.

With these points in mind, I would like to express some of my own opinions on this
subject and also describe some of niy recent experiences with single phase control material.

Standardization should have optimally several characteristics. First, the material

used should be precisely definable. By that, I mean it should be describable in terms of
exact quantities of materials treated in a specific way to obtain the standard value.
Second, the material should be as close in characteristics to the sample as possible. When
dealing with whole blood as a sample, this is often difficult to achieve while maintaining
definability. Lastly, the standardizing measurement should be done in the same manner as a
sample measurement. Often the particular instrument used requires different sample appli-
cation procedures for sample and standard.

Standardization of pH measurements meets these criteria reasonably well. Gas stan-

dardization, however, is another question. By far, the most widely used technique for gas
standardization utilizes mixtures of pure gases flowing through a sample chamber. This
technique violates two of the three basic principles of standardization. First, it is a gas
while the sample is a liquid, and second, it is flowing, whereas the sample is static. The
reason for its widespread utilization is that this material is easily handled and the
resulting gas tensions are quite reproducible. As long as the flow rate of the gas is
reasonably constant and humidif ication of the gas with water is accomplished at a constant
temperature, good day-to-day reproducibility is attainable. Biases associated with the
media and flow differences are accepted as the price of reproducibility. Thus, there is a
gas-to-liquid correction factor associated with oxygen measurements which stems from the
choice of a gas as the primary calibration medium £1]^. Likewise, there is a difference in
response characteristics of the primary sensors, such that the response time in the flowing
gas is invariably shorter than the response time in a static liquid. Fortunately, both of
these effects tend to be constant and day-to-day reproducibility is generally attainable.
However, there is good evidence to suggest that during the operating life of these electrodes
and prior to obvious malfunction, changes occur in these characteristics which are only
evident by measuring liquids of known gas tensions.

Figures in brackets indicate the literature references at the end of this paper.

275
 .

Methods for correcting for these effects have included calculation of error factors,
variation of sampling techniques [2], and equilibration of liquids with known gas compositions
[3]. This latter effort was equally directed at defining a single phase control material.
Sufficient progress has been made in this area to allow the recommendation of an all liquid
system for pH and blood gas calibration [4,5]. Commercial equipment is now available to
perform the equilibration in an accurate and simple manner so as to warrant utilization of
these liquids as a standardization device.

The Radiometer Company has recently made a major advance in instrument design by
incorporation of liquid calibration into their latest blood gas system. The ABLl incorporates
into its design a calibration system which utilizes various buffers equilibrated with
mixtures of carbon dioxide and atmospheric air. It is important that we encourage Radiometer
and other manufacturers to continue in this sound technical direction for future instrumentation.
As modern equipment becomes more automatic, we are often required to use the calibration
scheme chosen by the manufacturer even though better calibration methods might be available.

Moving on to the subject of quality control, I feel methods of control should have two
additional characteristics over and above those mentioned already. First, they must be
simple, because a tedious and complicated procedure just does not get utilized in a busy
laboratory. And secondly, they should respond predictably to a developing malfunction
before it becomes obvious without a control.

Many attempts have been made to prepare control systems with a whole blood matrix [6].
While these appear to have usefulness within a given laboratory, specimen variability and
handling difficulties prevent universal application of tonometered whole blood. On the
other hand, protein based controls for pH and PCO2 are commercially available [7]. There
has been a good deal of discussion as to the necessity of a protein based solution for
quality control, the feeling being that a protein solution most clearly acts as whole blood
would act. My attitude is that whatever advantages a protein solution might possess (and I
am somewhat dubious that it has any), it is basically a nondefinable medium. I feel the
reproducibility and predictability inherent in a tonometered pure aqueous solution makes it
the media of choice for blood gas control.

Dr. Burnett and I reported on a particularly simple and also accurate device for the
preparation of liquid controls [5]. Prior to the availability of this system, it was very
difficult to get more than a subjective indication of the long term precision and accuracy
of blood gas results. We have had now nearly two years experience with this approach.
Table 1 shows cumulative daily statistics collected over a six month period at Hartford

Table 1. Blood gas quality control with tonometered bicarbonate (Hartford Hospital, 1974). \

pH •
PCO2 PO2 1

(calc. val = 7.35) (calc. val. = 86 mm Hg) (calc. val .

= 149 mm Hg)

Mean S.D. Mean S.D. Mean S.D. !

Dec 7.34 .016 81 5 142 6 !

Jan 7.35 .012 79 4 145 3

1

Feb 7.34 .011 79 3 144 2

Mar 7.34 .022 84 7 146 3 !

April 7.36 .014 80 5 144 2 ?

May 7.35 .012 83 4 145

Port Huron 7.35 .013 80 3

276
 Hospital and similar data for a one month period at Port Huron Hospital. These data show
the consistency of results obtainable over a long period of time with tonometered aqueous
solutions. It also shows that inter! aboratory comparison is also achievable with this
material. Use of material such as this will allow meaningful comparisons of interl aboratory
biases and errors. We have been in contact with several laboratories which have experienced
initial difficulties in matching our expected values. After some consultation, we were able
to identify se'rious unexpected system difficulties in some of the blood gas laboratories.

From the standpoint of commercially available materials, the General Diagnostics Company
has developed a series of aqueous, gas-equilibrated liquids packaged in glass ampoules with
assigned values of pH, PO2, and PCO2. This represents a particularly simple approach to
avoid the complexity of tonometry. Table 2 shows some data which we obtained using this
material in our laboratory. Notice that the precision obtainable with this material for pH
and carbon dioxide is as good as that obtained with our own temperature-controlled tonometer.
The oxygen results are somewhat influenced by variations in room temperature. However, a
more important factor in the case of oxygen was that we had difficulties establishing a
consistent method for transferring our specimens to a syringe for insertion into our instrument.
It is expected that direct sampling will improve this precision. The real beauty of this
product is that it makes blood gas control a very simple experimental process. Undoubtedly,
this technological achievement indicates that we will soon see liquid standards available,
packaged in a similar manner, perhaps even certified by NBS.

Table 2. Blood gas quality control with liquid-filled ampoules (Port Huron Hospital
February 1975)

pH PCO2 PO2

Mean S.D. Mean S.D. Mean S.D.

GD#1 7.06 .021 23 1.3 172 7.1

GD#2 7.39 .010 40 1.2 121 6.5

GD#3 7.64 .021 65 1.3 82 12.4

In closing my remarks here, I think we now have available the methods and materials to
begin to resolve in an objective manner some of the nagging questions in blood gas technology.
In our own laboratory we are beginning a study comparing the methods of quality control over
a long period of time in regard to maintenance factors. I look forward to other activities
directed toward the resolution of the problems in blood gas standardization and control.

References

of the
[1] Adams, A. P. and Morgan-Hughes, J. 0., Determination of the blood-gas factor
oxygen electrode using a new tonometer, Brit. J. Anaesth. 39, 107 (1967).
Medfield,
[2] Instruction Manual for the Model 165 Blood Gas Analyzer (Corning Glass Works,
MA, 02052).

'[3] Holmes, P. L. Green, H. E. , and Majano-Lopez , V. L., Evaluation of methods for calibration
,

of O2 and CO2 electrodes, Amer. J. Clin. Path. 54, 566 (1970).

I

[4] Veefkind, A. H., Van den Camp, A. M. , and Maas, A. H. J., Use of carbon dioxide and
oxygen-tonometered phosphate-bicarbonate-chloride-glycerol-water mixtures for calibration
and control of pH, PCO2 and PO2 electrode systems, Clin. Chem. 21_, 685 (1975).

gas measurements,
[5] Noonan, D. C. and Burnett, R. W., Quality control system for blood pH and
with use of a tonometered bicarbonate-chloride solution and duplicate samples of whole
blood, Clin. Chem. 20, 660 (1974).

277
 P'-oficiency testing and quality
[6] Kenny. M. A., Delaney, C. J., and Raisys, V. A., A
control program for blood gas analysis, Clin. Chem. }3_, b4y 6).

Warner Lambert Company, Morris

[7] versatol Acid Base (General Diagnostics, Division of
Plains, NJ, 07950).

278
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

DEVELOPMENT OF REFERENCE METHODS: BLOOD GAS ANALYSIS

Arthur L. Malenfant and Kevin D. Fallan

Instrumentation Laboratory Inc.
113 Hartwell Avenue
Lexington, Massachusetts 02173, USA

1. Introduction

The nature of the parameters measured in blood gas analysis leads to unique problems
in attempting to establish reference methods which will serve as an accepted norm against
which manufacturers of instrumentation and practitioners of the method may compare the
results which they obtain. The species are extremely labile both because of changes induced
within the sample and also because of the ease of contamination of the sample by atmo-
sphere. Metabolism of the sample results in lowering of Pq and elevation of ^qq^, while
mixing with air will reverse this situation.

Since the analysis of blood for pH, P^q and Pq must be performed in an environment
which may easily lead to errors resulting from contamination, reference methods must pay
strict attention to both operator dependent factors and instrument dependent factors.

The primary purpose for establishing reference methods is to establish the precision
and accuracy of the procedure used to obtain blood gas information in the clinical lab-
oratory. As in any analysis, accuracy is defined as the expected- or absolute-value. The
nature of the measuring system requires that the reference methods be established to define
the accuracy and precision at the level where the Pq and/or P^q values are being measured.
It has been prevously documented that calibration for P^ should vary with the level
U2
measured [1,2]^. Instruments calibrated to achieve Pn measurements which are within 2
U2
percent of the expected value at 50 mm Hg must be calibrated differently to achieve the
same performance at Pq 's of 300 mm Hg. The measurement of Pqq in the clinical laboratory
is over a much narrower range, however, and this requirement is not evident for that
measurement.

While accuracy of results is to be sought as a worthwhile goal, the importance of

precision ought not to be underestimated. Monitoring of the clinical state of a patient
while following the results of sequential measurements over several hours or days provides
a clinically useful tool. Therefore, while results might even be in error in absolute
terms, good precision of results will permit successful care of the patient.

^Figures in brackets indicate literature references at the end of this paper.

279
 2. Operational Requirements

The number of factors which affect the accuracy of blood gas is significant. In order
to organize them, they will be described as operator dependent factors and instrument
dependent factors.

Operator dependent factors which affect results begin with the drawing of the blood
sample for the analysis. There will be no attempt to deal with the actual technique used
by the clinician in drawing either arterial or venous samples. The attempt will be made,
however, to delineate aspects which can affect blood gas results by actions taken both
before and after the sample is drawn.

The syringe used to draw the blood will have an effect on the results achieved. There
is literature which indicates that only glass syringes permit accurate results [3] while

other literature indicates that disposable plastic syringes are suitable [4]. Our ex-
perience has been that as long as the blood is not stored for extended periods between
drawing and analysis, no significant bias is introduced. If samples are stored under iced

conditions for several hours, changes consistent with increases in Pq and decreases in
Pj,Q caused by diffusion through the walls of the plastic syringe are observed when samples
stored in plastic are compared with samples stored in glass.

The amount and nature of the anticoagulant used also affects the blood gas results
achieved for pH and P^q . The anticoagulant of choice for blood gas analysis is sodium
heparin [5]. At the present time, however, standardization of the number of units of
heparin per ml of sample is not adequately defined. As a result, variation in the volume

of anticoagulant used affects results through dilution of the sample, as well as through pH
changes induced by using other anticoagulants such as ammonium heparin [6].

For the measurement of P^ and for the direct photometric measurement of oxygen
saturation, the choice of anticoagulant is not so critical since the only effect on these
measurements would be through dilution of the sample by the oxygen contained in the anti-
coagulant.

Once the operator has a properly prepared sample, the next factor under the control of
the operator is the proper calibration of the blood gas analyzer. Accuracy of calibration
of Pq and P^q is dependent on the accuracy of the calibrating media, whether liquid or
gas, and on correction for barometric pressure changes. Manufacturers of calibrating gases
standardize their values through either the Scholander technique or gas chromatography.
Within a given geographic region, day-to-day fluctuations in barometric pressure will be
small, but many, if not most locations will have a standard barometric pressure lower than
the 760 mm Hg standard at sea level. At 3300 feet elevation the barometric pressure would
average about 674 mm Hg.

280
 The accuracy of blood gas results, even when each of the necessary calibration steps
is carried out properly, is also dependent on the operator. The instrument must be given
proper maintenance. At present, totally blood compatible materials are not available for

construction of blood gas analyzers. Therefore, regular routine cleaning of the system is
necessary to prevent buildup of protein residues from previous analyses. The advent of
automation in blood gas analysis insures that samples will each be handled similarly once
they are inducted into the instrument, but without proper maintenance, even these systems
will not yield results which are accurate in the long term.

Instrument dependent factors present in achieving accurate, precise results in blood

gas analysis deal primarily with the proper calibration of the sensors and introduction of
the blood sample into the measuring chamber of the system.

In the near term, establishment of standards for blood gas calibrators is required.
At the present time, the use of phosphate buffers serves to establish the calibration of

the glass electrode for monitoring pH of blood [7]. Calibration of the gas sensors,
however, continues to be a source of some controversy. The use of standardized, humidified
calibrating gases flowing past the tip of the sensors has been the calibration method of
choice for many systems. However, the question of the gas/liquid membrane correction
factor [8] for oxygen measurements has yet to be adequately resolved. Proposals have been
made for alternatives such as tonometered glycerol -water solutions containing phosphate,
bicarbonate and chloride [9] to serve as the calibration medium. Proposals for other
fluids such as tonometered blood [10] and sodium bicarbonate--sodium chloride solutions
equilibrated with known gas mixtures [11] have also been made, the latter being proposed as
a quality control procedure.

3. Methodology Currently Employed

Currently available control materials for characterizing the reliability of blood gas
results fall into two categories. The first category includes serum-based controls that
are reconstituted by addition of an appropriate solvent. The resulting solution, once
reconstituted, is then inducted into a calibrated blood gas analyzer and the results are
compared with the nominal values to be expected. These require little time to reconstitute
and are relatively simple to use. Disadvantages in attempting to follow this approach,
however, to develop a standard reference material include the requirement for a high degree
of care in mixing with the solvent to prevent undue contamination with air before sampling,
the fact that the values are established by the same technique as the method being tested,
the temperature coefficient is different from that for blood, and the material must be
handled differently than a blood sample would be handled. This last would lead to in-
correct assumption either of problems or lack of them.

Another control material which might be considered is the aqueous based control material
which is supplied in liquid form with known values of pH, P^q and Pq . In this case the

mixing step is completely eliminated. The solution need only be inducted into the blood gas
analyzer and the results recorded. The disadvantages here, however, are similar to the
serum-based material above with respect to a standard reference material. First of all, it

281
 .

is simply an aqueous solution, not blood. Also, the range of values available is limited of
necessity and may not include some of the more extreme ranges where blood gas analysis may
be performed.

In addition, neither of the above approaches tests the operator's ability to properly
handle blood samples under controlled conditions. As noted previously, this is an important

factor in achieving accurate blood gas results.

4. Tonometry of Whole Blood

It is proposed that the method of choice for developing a reference method for blood
gas analysis be the tonometry of whole blood. This approach has a number of advantages.
First, the method being used is the primary method. Second, all controlled parameters are
independent of the measuring system, i.e. established without reference to a blood gas
analyzer. Equilibration of the whole blood can be made with gases having accurately known
Pq 's and P^q 's. Instrument response will be identical to that when clinical samples are
being measured. The operator is handling whole blood in the same way as a clinical sample
once the blood is withdrawn from the tonometer. Any deviations observed will be identical
to those to be experienced with clinical samples. Temperature coefficients are identical,
for example. The pH of whole blood may be adjusted by addition of bicarbonate so that it is
both known and predictable under specific tonometry conditions to be used to establish the
Pq and P^q values desired.

Below is shown the results' achieved using the approach described. A volume of whole
blood was standardized to a known pH by addition of bicarbonate. This was then sealed into
containers and kept refrigerated. Over a thirty-day period these blood samples were to-
nometered with gases of known composition, and the pH, P^q and Pq determined after to-
nometry. The results are tabulated as shown. The analyses were performed both at Instru-
mentation Laboratory, as well as at another institution. The gases used for tonometry were:

Location No. 1 Location No. 2

Gas No. SD SD

1^ pH: 7 370 0.006 pH: 1 360 0.007

P 49 5 mm Hg 0.6 ^^002
46 5 mm Hg 2.0

v = 51 7 mm Hg 1.0 51. 4 mm Hg 1.95

2^ pH: 7 429 0.007 pH: 7 420 0.010

35 8 mm Hg 0.44 33 9 nni Hg 0.77
^C02' ''C02"

139 4 mm Hg 1.4 137 5 mm Hg 2.1

3^ pH: 7 511 0.006 pH: 7 500 0.013

21 7 mm Hg 0.4 P 20 7 mm Hg 0.58
^C02"
75 1 mm Hg 1.3 75 00 mm Hg 2.8

^Gas No. 1: 7.0 percent O2, 7.0 percent CO2, balance N2

"^Gas No. 2: 20.0 percent O2, 5.0 percent CO2, balance N2
^Gas No. 3: 10.0 percent O2, 3.0 percent CO2, balance N2

282
 Thus, it can be seen that consistent results may be achieved over a period of time
using tonometry with gases of known composition to bring whole blood samples into gas
equilibrium. Although the range shown here is relatively limited, it could easily be
extended to, for instance, high Pp. 's by use of a gas having a P^ of 300 or 400 mm.
U2 U2

References

[I] Key, A., Biom. Eng. 9, 1954 (1974).

[2] Reppeto, N. P. and Moser, K. M. , NSCPT Analyzer, 4 (2), 13 (1974).

[3] Sackner, M. A., Med. Times, 95, 79 (1967).

[4] Winkler, J. B., et al.. Chest, 66, 518 (1974).

[5] Gambino, S. R. , et al.. Annals of N.Y. Acad. Science, 133, 259 (1966).

[6] Shapiro, B. A., Chicago Year Book, Med. Pub., 167 (1973).

[7] Bates, R. G., J. Res. Nat. Bur. Stand. 66A, 179 (1962).

[8] Reppetto, N. P. , ihid.

[9] Veerkind, A. H,, et al., Clin. Chem. 21_ (6), 685 (1975).

[10] Adams, A. P. and Morgan Hughes, J. 0., Br. J. Anesth. 39, 107 (1967).

[II] Noonan, D. C. and Burnett, R. W. , Clin. Chem. 20_, 660 (1974).

283
I

I
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975, Issued June 1977.

QUALITY CONTROL AND STANDARDS

S0ren K. S0rensen
Radiometer A/S
Copenhagen, Denmark

The term "quality control" can be interpreted in various ways. In a narrower sense, it
can thus be taken to mean no more than a quality check-up of a measuring system. In a
wider sense however, it aims at establishing whether the results of tests performed on a
sample actually do represent the bodily condition of a patient. It is, of course, this
latter objective of quality control that is the ideal one, but unfortunately it is not
always attainable.

The flow diagram for a sample is seen in figure 1. A proper quality control is
performed by letting a known sample run through the whole system— in other words, it is the
reference person we are looking for. It has been attempted [1]^ as a tool for controlling
the "daily mean" to take blood samples from the staff everyday and then use the results as
control material. It is quite natural to use blood in quality control as it is the sample
material, but blood is a very special liquid and it is difficult to imitate the physical
and chemical behavior of blood in other media. When trying to simulate blood, it is
necessary to get the correct buffer capacity for pH, PCO2, and P02, where the oxygen
buffer is the most difficult to establish. It is necessary to simulate what is called the
suspension effect from the red blood cells when measuring pH. It is necessary to simulate
dissolved protein and the viscosity of the blood. A list of specifications can thus be
made, from which a lot of practical "artificial blood references" can be produced.

the the assessment

the the
blood measure- of
sampling transport ment result the result

Figure 1. Flow diagram for running a sample.

In some types of analysis, it is possible to get very close to the ideal condition,
e.g., serum reference materials used in quality control of the inorganic salts are excellent
synthetic substitutes for the genuine sample. But as far as pH and blood gases are concerned,
it is not possible to produce the stable reference material that is a necessary condition
for carrying out quality control [2].

Quality control is performed as spot tests, but no test can improve a weak system.
The spot tests are therefore unable to solve quality problems; they can only indicate that
something is going wrong. This means that one cannot merely concentrate on finding the
best of all reference materials. It is equally important to ensure that inbuilt security
in the system is provided. Internationally accepted rules for the application of reference
methods, units, formulas for derived parameters, choice of derived parameters and sample
handling are some of the indispensable conditions on which the quality control must be
based. Internal quality control is much to be the preferred, but technical limitations,
e.g.. Stability problems, and legislation requirements are some of the reasons why external
quality control is utilized. Western European countries, and especially the Medical Society

Figures in brackets indicate the literature references at the end of this paper.

285
of West Germany have shown a great interest in external quality control. Their guidelines
do not deal with blood gases, however, as they have based their quality control on test
sera. External control may be a useful addition to laboratory control, but the time delay
involved is reason enough for rendering it unsuitable as a primary control basis in blood
gas analyses. As a first step to quality control of blood gases, collaborative control
within small geographical regions is a possibility. This could be a valuable, although
expensive, way to build up a control system.

This author feels that it is important to establish that both choice of derived para-
meters and units are part of a quality control program. One of the most important things in
quality control is the log book. Without recording the trends and the change in scatter of
the results, it is impossible to foresee the coming break-down of the system, which always
occurs at the most inconvenient time in the day. To reduce the shut-down period, the
quality control log book is thus an indispensable tool.

It is very difficult to solve the entire quality control problem for pH and blood
gases, but even a partial solution will be a great improvement over the present situation.
Before starting a control program, it is natural to ask what limits of accuracy and precision
are necessary to give the control a meaning. Manufacturers usually refer to the instru-
mentation specifications. However, these specifications are a result of technical lim-
itations and, therefore, are not necessarily in agreement with the clinical requirements.
The best plan is to let clinical needs, nbt the technical ability, determine the limits of
accuracy and precision. This point is very important.

Some authors [3] are of the opinion that the methodological inaccuracy must not be
greater than one SD of the normal range. The normal range is then 4 SD, which incorporates
both the biological and the methodological variation. For the blood pH, this means an
accuracy around 15 mpH, which should be obtainable. In Pco2 measurements, the same range
is 3.5 mm Hg, which can be easily reached.

As regards the separate parameters, pH is the best established. There is a definition

of pH--whether good or bad--and there are buffers which can be used to calibrate the electrode.
Pco2 can be determined directly with a gas-sensing electrode, or it can be determined
indirectly by means of the equilibration function of the sample. In either case, an analyzed
gas mixture is the basis for evaluating the quality of the measurement. P02 is determined
directly with a P02 electrode, and here again gas mixtures form the basis of calibrations.

The lack of a reference material is not the only hindrance to careful control. Blood-
gas measurements are typical stat measurements of one to two samples at a time, which means
a Q.C. every second sample is desirable, and that is, in practice, impossible. A complete
calibration of the system is therefore preferable just before the measurement is performed,
but in practice even this is normally not possible. A compromise involving frequent
calibrations during the day can, however, improve the quality of the measurement.

For calibrating purposes, gases are satisfactory, but as a medium in quality control,
a liquid is much to be preferred. A minimum requirement is that quality control shall be
performed on a liquid with known partial pressures of carbon dioxide and oxygen— in other
words--a tonometered test solution. Blood is, of course, the preferred liquid for that
purpose. The preferable control procedures are listed in figure 2.

pH: Buffer, Ton. bicarbonate solution

PC02' Blood, Ton. bicarbonate solution

P02= Blood, (Ton. aqueous solution)

Figure 2. Liquids for quality control.

It is not a sophisticated way of performing quality control, but, on the other hand,
many laboratories do not even reach this level, and those that do are satisfied. That leads
to another problem, namely the gap in the technical standard between the leading laboratories

286
and the ordinary ones, even in relatively homogeneous countries. It is very important that
this gap in technical standard does not grow even wider, thereby creating a situation where
quality control procedures become unattainable for lesser laboratories so that there is a
risk that they will refrain from using them. Quite likely, such laboratories would be the
ones that stand in greatest need of quality control.

The development problem has other aspects. If the laboratories are unable to prove
their willingness to document the quality of an analysis, some other authority will take
the initiative and create rules which, in the worst case, are impracticable, and in the best
case, impractical. The same demands must be made with respect to the industry that produces
equipment and diagnostic products for the laboratory, as a quality control program will
only be a success if complete confidence exists between authorities, laboratories, and the
industry involved.

The tonometer technique will no doubt give a lot of problems owing to inadequate
training and experience. The only possibility is to promote the tonometer technique and
then to learn from experience whether it is usable for control purposes. The manufacturers
of blood-gas equipment must therefore be prepared to take the great responsibility for
producing tonometers of a reliable and simple technical construction. Today, only a few
companies produce a commercial tonometer. A tonometer, however, is only the tool used to
produce gas equilibrium. The standard is the gas mixture which is used to equilibrate the
blood.

Gas mixtures can be obtained everywhere but they have to be analyzed before they are
usable as standards in tonometry. That is a point where cooperation on an international
plane could produce fast results. Determination of the compositions of gases is purely a
measuring problem, but internationally certified gas mixtures could remove one uncertainty
from tonometry and also from the calibration procedure.

From a technical point of view, tonometered buffers are easier to handle than blood,
and it might be worthwhile to try to evaluate in practice all the differences that can
occur between blood and aqueous solutions. This compromise using buffers instead of blood,
can give the unskilled laboratory a chance to start a control program without being completely
lost from the beginning. Because blood-gas analyses are so complicated, the need for good
quality control is immense, but I believe that progress here is more a matter of perspiration
than of inspiration.

References

[1] Whitehead, T. P., Methods of quality control, Ann. Clin. Bioohem. 6_, 94 (1969).

[2] Broughton, P. M. G., The future of quality control, Ann. Clin. Bioohem. 6^, 147
(1969).

[3] Gowenlock, A. H. , The influence of accuracy and precision on the normal range,
Ann. Clin. Bioohem. 6_, 3 (1969).

287
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

DEVELOPMENT OF REFERENCE METHODS

S0ren K. S0rensen
Radiometer A/S
Copenhagen, Denmark

The precision of a measurement can be determined without having a reference method,

whereas the accuracy may elude the determination without an established and well -documented
reference method. The reference method is the most reliable method irrespective of analysis
speed and cost, which implies that the reference method is not necessarily identical with
the routine method.

A routine analysis is the practical procedure adopted by a particular laboratory. A

routine analysis is, of course, tested against the reference method to expose the limitations
of the former. The routine method, which may not be the reference method, is classified in
the group of methods containing bias--known or unknown. It allows one to select just that
method of analysis that meets the prevailing conditions {viz.,, local legislation, technical
training, instrumentation, and future development).

A proposal for categorizing analytical principles has been made by IFCC. According to
this definition, a reference method is a method with an accuracy = 0 ± 6, where 6 is
negligible in comparison with interlaboratory imprecision. The ultimate method, called
"definitive method", has no known source of inaccuracy.

In order to obtain the total concept, a definitive method must be available. For pH it
is not possible to set up a definitive method, as I will explain later, but for both Pcoa
and P02 it is--at least in theory--possible to measure the partial pressure in a system by
known, established techniques. What I have in mind is to try to investigate a method for
measuring the total amount of carbon dioxide or oxygen in an infinitely small quantity of
gas. The problem with blood is that the total content of carbon dioxide or oxygen does not
give any information on the partial pressure of the gases in the blood. In other media it
is, however, possible to get such information due to simple dissolving processes in these
media. By equilibrating, e.g.^, a small piece of plastic in the blood and measuring the
total gas content, the partial pressure can be calculated. The only limitation to such a
method is the technical problems. With the advanced technical skill of today, it should,
however, be possible to develop a "definitive method" based on this principle, but as far as
I know this has not yet been attempted in practice.

The very restrictive definition of reference methods is difficult to meet especially

for such complex systems as pH and blood gases. The day when such reference methods are
available lies somewhere out in the future, and in the light of this definition even pH is a
doubtful parameter. pH determination on blood is based on an operational definition, using
a hydrogen electrode and a saturated KCl liquid- junction. Especially the residual liquid-
junction potential gives a more or less unknown bias on the measurement. Nevertheless, no
one will deny that pH in blood is a valuable and important parameter, and there are, in fact,
no means of eliminating the liquid-junction problem as long as electrochemical cells are
used to measure the hydrogen ion activity. The normal way of measuring pH is, of course, to
use a glass electrode instead of a hydrogen electrode. The junction potential can be changed
by using other types of reference electrode systems, but it cannot be eliminated and it is
unlikely that anyone has tried to measure blood pH with a hydrogen electrode! What can be
done is to accept the pH definition as it is or, better still, to improve it with a more
precise description of the liquid-junction type of salt bridge and the geometry of the
junction.

289
 pH is a good example of a practically defined parameter which has shown its worth, even
though several standardizing committees have tried to change the units for pH. It is my
opinion that we still have the best of all proposed solutions to pH in the operational pH
definition, which requires only small changes to lead to what we may call an "assigned
method."

Turning to the PCO2, matters are far more complicated. It may be possible, as already
mentioned, to think out a definitive method which, no doubt, would be useless in practice.
There are two established methods for determining PCO2, one measures the CO2 partial pres-
sure with the Severinghaus-type carbon dioxide electrode. The other is the equilibration
method, known as the Astrup technique, developed by Astrup and Siggaard-Andersen. Neither of
these techniques is a direct measurement of CO2 pressure or carbon dioxide activity. It may
seem obvious to favor the electrode technique, as one can argue that an electrode gives a
more direct measurement than the equilibration technique.

As all other reference methods, the CO2 reference methods require that certain condi-
tions are defined, for example, fixed and known temperature, oxygen saturation of the sample
and hemoglobin concentration. However, a lot of problems connected with the electrode are
known, and can be classified as interferences. Leakage of hydroxide ions from the pH glass
surface, different osmotic pressure in the sample and the electrode electrolyte, and loss of
gas into the bulk electrolyte, are some of the problems which are not easily overcome.

Using the Astrup technique as reference method is not the ideal solution, but for
several reasons I feel that this method offers certain advantages. By tonometer ing the same
sample as used for measurement, the equilibration curve is established directly from the
sample. The tonometry process will offer a fixed saturation. When using an electrode,
PCO2 is converted to pH in an electrolyte layer after diffusion across a thin plastic mem-
brane, and a calculation is necessary. In the equilibration technique, on the other hand,
the electrode--indeed, one of the most reliable ones--is used for both calibration and mea-
surement. The Astrup method thus eliminates the Henderson-Hasselbalch equation. One disad-
vantage of the equilibration is the need for defined or known oxygen saturation of the blood
sample. The Astrup technique will, of course, not oust the gas electrode technique, but it
is my opinion that we have a better theoretical background when using an equilibration techni-
que as reference method.

The P02 analysis is, no doubt, the most complicated one for which to propose a reference
method. Compared with PCO2, for which two well-established methods are available, P02
offers only one method, namely the P02 electrode itself. Admittedly, an electrode which in
practice has shown its validity, but whose theoretical basis is very weak, has a lot of
known interferences and side-reactions. To found a reference method on the P02 electrode
itself will hardly lead to success.

Here, too, there is the possibility of using an equilibration technique. If one has a
precise knowledge of the oxygen dissociation curve, then the electrode technique can be
converted into an optical determination of the oxygen saturation. Experience with the
optical saturation measurements are fairly good, but the other side of the problem--prior
knowledge of the oxygen dissociation curve--may kill the idea. Several parameters change the
dissociation curve. Examples are pH, PCO2, and enzymes (2,3-DP&-). I feel that optical
analysis based on a carefully selected set of definitions in order to lock the dissociation
curve in position would provide a better reference method than any system founded on the
polarographic oxygen electrode. If the outlined methods are discovered to be useless, only
the definitive methods are left. As a matter of form, it may be recalled that potentiometri-
cally working oxygen electrodes based on redox-potential measurements do exist, but so far,
problems connected with the irreversibility of the material have limited its use. Someday,
a good, reversible, oxygen-binding substance may lead to workable electrodes.

pH none pH glass electrode

'C02 determination CO2 Astrup method

in"microbubble"
'O2 saturation
determination equilibration
in "microbubble" technique

Figure 1. Definitive methods. Figure 2. Reference methods.

290
 The reason for choosing a method based on equilibration both for carbon dioxide and for
oxygen is that, from a scientific point of view, the processes are better known. From a
practical point of view, the suggested methods are more debatable, and a simpler way to
solve the reference method problem is not to base the measurements on the gas sensing system,
but to use the tonometer as the reference method. That means to develop a method which is
able to produce a reference material --the tonometered blood sample.

I have now tried to establish a system of methods which could serve as an inspiration

in our endeavors to develop a suitable reference method. No doubt, it will be clear to

everybody that, so far, these methods are carried out only as suppositions, but they do at
least expose the fields which in my opinion ought to be studied.

291
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

STANDARDIZATION OF ION-SELECTIVE ELECTRODES FOR SERUM ANALYSIS

M. S. Mohan and Roger G. Bates

Department of Chemistry
University of Florida
Gainesville, Florida 32611, USA

Electromotive- force measurements at 37 °C in synthetic electrolyte mixtures simulating

serum provide evidence that ion-selective electrodes respond in a near-Nerstian manner to
changes in concentrations of the species to which they are selective. The relatively
minor effects of changes in ionic strength on the electrode potentials are almost entirely
accounted for by changes in the activity coefficients. It is suggested that similar
electrolyte mixtures may serve as calibrating standards in clinical monitoring with ion-
selective electrodes. The lanthanum fluoride electrode shows promise as a useful reference
electrode in media of this sort.

1. Introduction

In the decade which has passed since their introduction, ion-selective electrodes have
found increasing use in a variety of biological applications [1,2].^ In addition to electrodes
designed to measure electrolytes in body fluids, novel devices based on the "building block"
principle [3] have been constructed to determine other clinically important entities such as
urea [4] and glucose [5]. Changes in the configuration of existing electrodes, such as
immobilizing the exchanger in a polymer matrix, have resulted in highly reliable and rugged
devices [6].

In the analysis of electrolytes, an important advantage of ion-selective electrodes

over existing clinical methods such as photometry is that they respond to fvee ionic activi-
ties or concentrations rather than to total concentrations. If one could obtain free ionic
concentrations (which are the medically significant quantities) from measured ion-selective
electrode potentials, the utility of ion-selective electrodes would be very significantly
enhanced. Such a procedure is dependent upon the development of reliable calibration proce-
dures, particularly in the complex mixtures of electrolytes found in biological fluids. Two
major problems must be faced in establishing suitable standards. First is the experimental
limitation imposed by the imperfect selectivity of the electrode; second is the theoretical
difficulty attending a conversion of activity to concentration.

2. Selectivity

The empirical Nicolsky-Eisenman equation [7] is useful in describing the response of an

electrode selective for an ion in the presence of an interferent j:

=
.'''^'^j
Eg^ Constant ± log |^a. + Zk-^.a (1)

where Eg] is the electrode potential, a^ and aj are the activities of the primary ion and
the interferent, respectively, z-j and zj are the corresponding electrical charges, and k^j

1
Figures in brackets indicate the literature references at the end of this paper.

293
is the selectivity ratio. This equation holds fairly well at low interference levels but,
generally speaking, the selectivity ratio is not a constant and depends on the activities of
i and j. Since the relative concentrations of ionic constituents in body fluids are likely
to vary under disease conditions, it becomes necessary to examine the extent to which measured
potential differences are affected by selectivity parameters.

3. Activities vs Concentrations

The development of an operational pH scale was facilitated by the availability of

buffer substances spanning a wide range of pH values at comparatively low ionic strengths
between 0 and 0.1. This not only enables one to separate conventionally the thermodynamically
measurable mean ionic activity coefficients into single ion activity coefficients but assures
minimal residual liquid-junction effects in practical measurements. In the calibration of
ion-selective electrodes, used to determine the clinically important ions, the ionic strengths
of the solutions may exceed 0.1 and may vary rather widely, spanning a wide range of activities
[8]. This would mean not only that the simplest formulas for the conventional separation of
mean ionic coefficients into individual ionic contributions would no longer be valid but
that the residual liquid-junction potentials might be large enough to cause appreciable
errors in the measurements.

Metal-ion buffers offer a possible solution. Although buffers for calcium ion, based
on the EDTA and NTA complexes, have been proposed [9], the measured pCa values depend heavily
on the accuracy of the conditional formation constants. Furthermore, ligands complexing
with Na"*", K"*", and other alkali metal ions of sufficient stability in aqueous media are not
available.

Despite the success of the convention based on hydration numbers [10] in certain special
cases, the evaluation of single ion activity coefficients in multi component electrolyte
mixtures continues to pose considerable difficulties. On the other hand, the ionic strengths
of body fluids are sufficiently low (usually less than 0.17 mol 1"^) that the actual single
ionic activity coefficients are likely to be reasonably close to the values derived from the
conventions that have proved useful at ionic strengths less than 0.1. Taking these factors
into consideration, suitable standards for calibrating ion-selective electrodes for use in
biological fluids are provided by a set of synthetic electrolyte mixtures whose ionic
concentrations span the range normally encountered in clinical analysis. It has been shown
[11] that the electrode potential responds in a near-Nernstian manner to changes in the
aoncentvation Of the ions of interest at the normal ionic strength of serum, although the
slopes S = Eel/log C may depart slightly from the values given by theory. In addition, it
appears possible to apply corrections for changes in ionic strength if these occur due to
disease conditions.

4. Emf Measurements in Standards of Constant Ionic Strength

Measurement of electromotive force at 37 °C in cells with and without liquid junction

were made with a series of calibration standards in which the concentrations of Na^, K"*",
Ca^"*", H"^, and CI" were varied simultaneously between the lower limits encountered under
clinical conditions and their normal values. The indicator electrodes used in the study
were: sodium glass electrode (Corning), potassium electrode (valinomycin exchanger in
diphenyl ether), calcium electrode (Simon lipophilic Ca^"*" exchanger [12] immobilized in
polyvinyl chloride), and pH glass electrode (Corning). Tris(hydroxymethy)aminomethane
("Tris") was used as the buffer substance. The buffer ratios in the reference solutions
were varied and pH derived from pK'bh+ = 7.907 in isotonic saline at 37 °C; this apparent
dissociation constant for protonated Tris was calculated from data given by Durst and
Staples [13]. Nitric acid and sodium nitrate were used to vary the buffer ratio and to
maintain the ionic strength constant at 0.16 mol 1"^. Details of the preparation of the
standard solutions and other experimental procedures have been given elsewhere [11].

Compositions of the standards of constant ionic strength are given in table 1. As

already mentioned, it is important to eliminate or minimize liquid-junction potentials in
order that the measured potentials truly reflect changes in ionic activities or concentrations.
For this reason, particular attention was given to the use of cells without liquid junction,
with reference electrodes immersed directly in the standard solutions. The silver-silver

294
 Table 1. Compositions of calibration standards.^ (All solutions have
an ionic strength of 0.16 and contain 1 mmol NaF per liter.)

Standard PH
Number
Sa ^Ca ^Cl

1 100 1 .0 0.30 66.6 7.90

2 105 1 .4 0.38 70.0 7.80
3 110 1 .8 0.46 75.0 7.70
4 115 2.2 0.54 80.0 7.60

5 120 2.6 0.62 85.0 7.55

6 125 3.0 0.70 90.0 7.50
7 130 3.4 0.78 95.0 7.45
8 135 3.8 0.90 100.0 7.40
9 140 4.0 1.00 106.0 7.36

^C^ in mmol 1

chloride electrode is of limited usefulness here, since the chloride ion is one of the
constituents being monitored. The reliability of the lanthanum fluoride electrode as a
reference was investigated. Accordingly, small concentrations of fluoride ion were added to
each of the standard solutions. Possible side reactions such as precipitation of calicum
fluoride or adsorption of the fluoride ion on the walls of the containing vessel were shown
to be of little concern [11]. The emf of three types of cell, namely

SCE II
Bridge Solution ||
Standard ]
ISE A

Ag;AgCl |
Standard |
ISE B
and

LaFg I
Standard ]
ISE C

was measured. Steady potentials (to ±0.1 mV) were usually obtained after about 45 minutes.

In order to obtain the indicator electrode slopes from data for cells of type B, the
contribution to the total cell emf made by the changing chloride ion concentration was
allowed for by assuming Nernstian response of the Ag;AgCl electrode. It is evident from the
least squares regression parameters for different electrode pairs given in table 2 that the
electrodes responded in a near-Nernstian manner to changes in the concentrations of the ions
being sensed when the ionic strength was constant.

This conclusion can be confirmed by examining the variation of the activity coefficients
in individual cases. For example, the emf of cells of type B can be written as

E = E°' + S log C^^ + 61.54 log C^^y_ (2)

where Cm and Cq] are the concentrations of the cation being sensed and the chloride ion,
respectively, and y+ and y_ are the corresponding activity coefficients on the molar scale;
S and 61.54 are the Nernst slopes (in mV) at 37 °C for the indicator electrode and the
Ag;AgCl electrode, respectively. On rearrangement, we obtain

E°" E E°' + S log y^ + 61.54 log y_ (3)

= E - S log C„ - 61 .54 log Cp, .

295
 1

Table 2. Parameters for the calibration of ion-selective

electrodes in standards of ionic strength = 0.16.

Indicated ion Reference E_,,(mV at 37 °C) S D.

ce 1

(i) electrode

SCE 435.1 - 60.9 pH 0 15

LaFg 437.0 - 61.15 pH 0 32

AgjAgCl 0.68 - 62.20 pH 0 11

SCE 63.3 + 57.3 loq C„ 0 16

LaFg 58.14 + 55.4 log C^ 0 40

Ag;AgCl 0.38 + 58.9 log C^ 0 16

SCE 36.58 + 54.04 loq

luy C,
y.^^ 0 06

LaFa 37.06 + 54.9 log C.. 0 27

Na
Ag;AgCl 0.35 + 56.4 log C.. 0 21
IN a

Ca^-' SCE 124.9 + 27.8 loq 0 10

LaFa 122.9 + 27.3 log C^^ 0 29

Ag;AgCl 0.25 + 29.8 log C^^ 0 14

^Standard deviation for regression, in mV.

Constancy of the right-hand side of eq. (3) will confirm the invariance of the activity
coefficients. From the low standard deviations for the values of E°" (table 3), it is
evident that the activity coefficients remain substantially constant throughout the range of
calibration.

Table 3. Constancy of activity coefficients in calibration

standards of ionic strength = 0.16.

Indicated ion Range of C. E°" S.D.

(i) (mmol r^y (mV) (mV)

Na"^ 100 to 140 49.00 0 18

1.8 to 4.0 77.80 0 20

Ca^-^ 0.3 to 1 .0 140.00 0 14

In measurements of cell A with the SCE as reference, the possible variation of the
junction potentials with the composition of the standards must be considered. Taking as
example the cell of type A with a sodium indicator electrode and SCE reference, one can
write

E°' + E. = E - S log C^g (4)

where Ej is the algebraic sum of all the junction potentials. The right-hand side of eq.
(4) was found to remain constant throughout the calibration range with a standard deviation
of only 0.2 mV. This result is important in that it attests to the effectiveness of the
constant ionic strength in stabilizing the potentials at the liquid junctions.

296
 5. Emf Measurements at Varying Ionic Strength

Under disease conditions, the ionic strength of serum may vary markedly from its normal
value near 0.16 mol 1"^. It thus becomes necessary to examine the effect of this variation
on the measured electrode potentials. It is then possible to identify and evaluate the
errors involved when the electrodes calibrated at one ionic strength are used in the widest
range of ionic strengths encountered in the analysis of serum.

The effect of changing ionic strength was studied by preparing three additional sets of
solutions. In all_sets, the concentrations of K+ and Ca^"*" were maintained at their normal
levels of 4.0 x 10"^ and 1.0 x 10"^ mol 1"^, respectively. The compositions are summarized
in table 4. In Set 1, the ionic strength was varied between 0.12 and 0.2 by changing the
buffer concentration while the Na"^ and CI" concentrations were kept constant at 0.1 mol 1"^.
In the solutions of Set 2, the sodium ion concentration was constant at its normal value of
0.14 mol 1"^ and the ionic strength was varied from 0.16 to 0.2 by altering the buffer
concentration. In the third set, the buffer concentration was maintained constant and the
ionic strength was varied from 0.12 to 0.2 by changing the concentration of sodium chloride.

Table 4. Compositions of solutions of ionic strengths

from 0.12 to 0.2a.

Set No.
^Na '^Cl 4 ^Ca
I

1 100 100 4.0 1.0 0 12 - 0 2

2 140 100 4.0 1 .0 0 16 - 0 2

3 100-180 106-186 4.0 1 .0 0 12 - 0 2

C-j in mmol 1'^ pH = 7.4 at I = 0. 16. All solutions contained NaF

;

at a concentration of 0.1 mmol 1"-^.

At these ionic strengths, single ion activity coefficients could be calculated with
sufficient accuracy for the present purpose by the Debye-Huckel equation in the form

-log y, - 'IL^ (5)

l+Ba/r
o
with the following values applicable at 37 °C: A = 0.5232, B = 0.3316, and a (the ion-size
parameter) = 4.6A.

The emf of cells of the types

Ag;AgCl ]
Solution |
(ISE) D

Ag;AgCl |
Solution |
Ca^"^ (ISE) E

and

LaFg 1
Solution |
K"^ (ISE) F

LaFg 1
Solution |
Ca^"^ (ISE) G

was measured for the solutions of varying ionic strength and examined in terms of changes
in activity coefficients to be expected from eq. (5). The results indicated that the
changes in emf agreed within 0.3 mV with expectations based on the changes in ionic strength
and activity coefficients.

297
 In addition, a comparison of data for solutions of Set 1 with those for similar solutions
of the same ionic strength in Set 2 showed that replacement of part of the Tris-H''' by Na"*"
resulted in changes in emf no greater than 0.3 mV. Similarly, a comparison of data for Sets
1 and 3 at the same ionic strengths showed that the emf of cells F and G were likewise
unchanged at all but the highest ionic strength [11]. Thus there is strong evidence that
electrode selectivity parameters have little influence on the potassium and calcium measure-
ments under conditions where a change in ionic strength of serum results from sodium depletion
or excess. This conclusion is important in measurements of serum calcium, for many commercial
calcium electrodes now in common use are subject to a significantly greater sodium error
than is the Simon electrode used in our study.

A further observation is of interest. In the measurements with the SCE reference

electrode it was found that the potential at the junction SCE| jSolution changes with ionic
strength in such a way as to counteract the activity effect. Since most practical measure-
ments are made with calomel reference electrodes, this would tend to reduce the error in the
measured concentrations of ions below what would be expected from activity effects alone.
Probable errors caused by changes in ionic strength are indicated in" table 5. A positive
error signifies that the value obtained from the electrode measurement is too high.

Table 5. Probable errors from differences in ionic strength

in cells with SCE standardized at ionic strength = 0.16.

I Na" Ca"
error, % error, % error, %

0.12 +1.2 -1.1 +5.2

0.14 +0.8 -0.4 +3.0

0.16 0 0 0

0.18 +0.8 +0.4 -1.3

0.20 +2.0 +1.1 -2.1

6. Conclusions

While the evaluation of a single ion activities in biological fluids remains a problem
of considerable complexity, measurements in synthetic electrolyte mixtures simulating serum
indicate that ion-selective electrodes respond in a near-Nernstian manner to changes in the
concentrations of the species for which they are selective. At a constant ionic strength of
0.16 mol 1'^, the activity coefficients and 1 i quid- junction potentials are reasonably con-
stant. Effects of changes in ionic strength on the measured cell emf are relatively minor
and almost entirely accounted for by alterations in the activity coefficients. These results
lead one to expect that standards similiar to the ones described here may be well suited for
the calibration, on a concentration basis, of ion-selective electrodes for use in biological
or clinical media. The lanthanum fluoride electrode appears to be promising as a reference
electrode, provided account is taken of the effect of ionic strength on the activity
coefficient of the fluoride ion.

References

[1] Moore, E.W. ,Studies With Ion-Exchange Calcium Electrodes in Biological Fluids: Some
Applications in Biomedical Research and Clinical Medicine, in lon-Selective Electrodes^
Durst, R. A., ed., p. 215, NBS Spec. Publ 314 (1969).
.

[2] Lustgarten, J. A., Wenk, R. E., Byrd, C, and Hall, B. , Clin. Chem. 20, 1217 (1974).

[3] Rechnitz, G. A., Chem. & Eng. News, p. 29 (January 27, 1975).

298
 [4] Guilbault, G. 6., and Montalvo, J. G., J. Amer. Chem. Soo. 92, 2533 (1970).

[5] Llenado, R. A., and Rechnitz, G. A., Anal. Chem. 45, 2165 (1973).

[6] Baum, 6., and Lynn, M. , Anal. Chim. Acta 65, 393 (1973).

[7] Eisenman, G. , The Electrochemistry of Cation-Sensitive Glass Electrodes, in Advanees

in Analytical Chemistvy and Instrumentation ^ C. N. Reilley, ed., Vol. 4, p. 213
(John Wiley & Sons, Inc., New York, 1965).

[8] Bates, R. G., Pure Appl. Chem. 36, 407 (1973).

[9] Ruzicka, J., Hansen, E. H., and Tjell, J. C, Anal. Chim. Acta, 67, 155 (1973).

[10] Robinson, R. A., and Bates, R. G., Anal. chem. 45, 1684 (1973).

[11] Mohan, M. S., and Bates, R. G. , Clin. Chem. 864 (1975).

[12] Morf, W. E., Ammann, D., Pretsch, E., and Simon, W., Pure Appl. Chem. 36, 421
(1973).

[13] Durst, R. A., and Staples, B. R. , Clin. chem. 206 (1972).

299
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

ELECTROLYTE ACTIVITIES IN HUMAN BLOOD PLASMA

M. J. D. Brand! gnd W. J. Scott

Technicon Instruments Corporation
Tarry town, N.Y. 10591, USA

1. Introduction

Assessment of a patient's acid-base and electrolyte status is an essential diagnostic

procedure in many acute and non-acute diseased states. Diseases of the renal and respira-
tory systems primarily result in acid-base and electrolyte imbalance although other
regulatory mechanisms exist [1]^. Care of the critically ill patient may require frequent
measurement of blood gases and electrolytes during potentially life-threatening episodes,
e.g. 3 in cardiac surgery. The use of serum or plasma electrolyte measurements to evaluate
the condition of acutely ill patients reflects the fundamental importance of extracellular
electrolyte concentrations in basic physiological processes.

From a physico-chemical standpoint blood plasma is an extremely complex liquid. It

is moderately concentrated in inorganic salts and contains appreciable levels of poly-
electrolytes (proteins). Uncharged species, e.g.^, CO2, glucose and urea, are present and
have important effects on metabolic processes. Because of the complexity of this chemical
system, the physician bases his clinical judgments on measurements of total electrolyte
concentration. Indeed, he has had little choice as the analytical methods in routine use
in clinical laboratories provide only concentration measurements for Na"*", K"*", CI", and
HCO3. The total concentrations of these species give little information on the true
composition of plasma, which is influenced by specific ion-ion interactions (metal-1 igand
complex formation) and non-specific interactions characteristic of non-ideal solutions.
The physical chemist prefers to define for each species present an "activity" which may
crudely be regarded as a concentration term, corrected for the various specific and non-
specific interactions.

The complexity of human plasma electrolyte composition has been recognized most
clearly in the measurement of ionized calcium [2]. The preferred practical method of
making this measurement uses an ion-selective membrane electrode [3]. Such electrodes are
unique among analytical sensors in that they have a response determined by ionic activity,
rather than total concentration. Routine serum electrolyte measurements in many clinical
laboratories are now made with ion-selective electrodes used in the Technicon Instrument
Corporation systems SMAC and Stat- Ion. These systems dilute the serum sample so that
ionic interactions are controlled, and the results are obtained as concentrations.
There have been several reported applications of ion-selective electrodes in the direct
measurement of electrolytes other than ionized calcium in whole blood, plasma and serum.
These have included the determination of sodium and potassium [4-7] as well as descriptions
of more complete systems for acid-base and electrolyte measurements [8-11].

The use of ion-selective electrodes in whole blood, serum or plasma as opposed to

dilute serum presents numerous problems, not least of which is an understanding of what is
actually being measured. Rational interpretation of the results of electrode measure-
ments requires a knowledge of the true nature of each ion species measured. Previously,
Walser [12] and Robertson [13] have considered ion interactions in plasma ultrafi Urates.

^Present address: Department of Chemistry, Texas A&M University, College Station,

Texas 77843.
2
Figures in brackets indicate literature references at the end of this paper.

301
Dahms et at. [14] have discussed the activities of sodium, potassium and chloride in human
serum as determined by electrode. In this paper a model is developed for electrolyte ion
activities in human plasma. A complete model would include acid-base, metal-ligand
complex, and redox equilibria as well as those involving gas and solid phases. However in
order to contain the complexity of the problem, only a limited range of acid-base and
complex equilibria will be considered.

2. Development of the Model

This model is intended to represent the extracellular fluid in equilibrium with

erythrocytes and other blood cellular components, i.e., the plasma fraction of blood. The
term plasma implies in vivo plasma rather than the in vitro counterpart which contains an
anticoagulant. However, because many of the equilibria constants required have only been
measured at 25 °C, numerical results are obtained at this temperature; this does not
detract from the accuracy of the model but only from its interpretation.
A. Water content of plasma

The solids, particularly proteins dissolved in plasma occupy a significant fraction

of the total plasma volume, i.e., the partial molar volume of proteins is not negligible.
An alternative statement is that the activity of plasma water is less than one. Activi-
ties may be given in two different scales; molar (mol/1 of solution) and molal (mol/kg of
solution). The molar scale is adopted here mainly because it is in agreement with normal
clinical chemistry practice (meq/1, mg/dl , etc.).

McLean and Hastings [15] have given the water content of plasma as a function of the
protein concentration (in g/dl):

H^O/100 ml plasma = (100-1) - 0.75 P (1)

Accordingly, the total concentration of all non-protein species is corrected from mol/1 of
plasma to mol/1 of plasma water using the factor given by eq (1).

B. Specific ionic interactions

is restricted to consideration of the clinically significant cations H"*",

This model
Na"*", K"*", and Mg^"*" and their interactions with various anions; metal -protein inter-
Ca^"*",
actions will be considered below. It is necessary to restrict the range of anionic species
considered; table 1 shows the anions which are thought to represent the majority of those
present in plasma. A great many others, e.g., amino acids, ascorbic acid, etc, could have
been included but were omitted to simplify the model. As a further simplification, not
more than one acid-base dissociation was considered for each acid as shown in table 1;
this is a reasonable restriction, from comparison of the acid pK with physiological pH
values. In general terms, the model considers the formation of a 1:1 complex between each
metal and each anion. No polynuclear complexes are considered, e.g., of the type M2L, nor
are mixed complexes, e.g., M1M2L. Not all of the possible 1:1 complexes were included as
for some the stability constant is too small to measure, i.e., the complex is completely
dissociated; table 2 lists the complexes included.

For each ligand:

4^4 (2)

M. . L^^M.L^ K (3)

^ ^4— ^4 K,
2i " [M,-]
[4]
(4)

302
 Table 1. Plasma anions and ionization constants.

CT
CO2 + HCO3 + H"^ 6.35

H2PO4 HPO4 + H"^ 7.21

HSO4 SO4 + 1.99

HCitrate" Citrate"" + H"*" 6.49

Lactic acid Lactate" + 3..86

Pyruvic acid Pyruvate" + 2.49

Numerical values from [16].

'chloride is assumed to be completely ionized.

Table 2. Complexes and dissociation constants pK.^

HCO3 ^2^°4 HCit^" Cit^" Lactate" Pyruvate"

Na^ -1.11 -0.72 -0.93

K"^ -1.00 -0.96 -1.10

Ca^^ -1.26 -1.08 -2.70 -2.31 -3.09 -4.85 -1.42 -1.08

Mg^^ -1.16 -l'' -2.50 -2.25 -2.46 -4.57 -1.37 -1 (2)

^Numerical values from [13] and [16].

'^Estimated values.

where [ ] denotes activity. The total concentration of the ligand is:

^ot = ' W/^2 h4]/^M.L^ ' ? [^•4]/^M.L2 (5)

where y represents an activity coefficient. From (2) - (5):

n -1 = lliot (6)

303
 [4] =
mhv^ (7)

[M.4] = K^,[M.][LJ (8)

[M.4] = 4i[M,.][L2] (9)

Equations (6) to (9) can be solved given L^^^, H, M^. and the activity coefficients. The
values of L^^^ and H {i.e. pH) are assumed known, as are the total concentrations of each
metal ion. The mass balance condition for each metal ion is given by:

Solution of equations (6) to (9) is obtained iteratively using an initial estimate for
M^-and calculated activity coefficients to give an improved estimate for M^. from (10).
The method of calculating activity coefficients and details of the computer program used
to obtain the solution are given below.

C. Protein- ion interactions.

This model is restricted to a consideration of the interaction of calcium and mag-

nesium with albumin and globulin. The effect of chloride binding to albumin, although
included in an earlier treatment [14] is not considered here as the magnitude of such
interactions has been shown to be negligible at physiological pH [17,18].

There have been many studies of calcium binding to albumin, the results of which have
not always been in agreement. Recently, Pedersen [19] has shown that under physiological
conditions, the concentration of bound calcium is given by

)
(CaAlb) "^l^^^

'^^^ot "
il H- K,(Ca2+) .
10-PH
K 1
ni)
I a

where = 10 »Ki = 241 1/mol, and ( ) represents concentration. This equation is

used in the model. Pedersen [19] also studied the binding of calcium to y-globulin.
Although far less data is available for globulin binding, it was shown that the binding
was less and of smaller pH dependence. A linear regression to the data of table 2 [19]
gives the equation.

(CaGlob)/Glob. .
= 0.564 pH - 1.79 (12)

at a mean ionized calcium concentration of 3.8 mmol/1 and a globulin concentration of 0.13
mmol/1. Pedersen has given insufficient data to estimate a reliable equilibrium constant
for this reaction and, therefore, the approximation is made here that under physiological
conditions the concentration of globulin bound calcium is linearly proportional to the
free calcium ion concentration.

Relatively few studies have been made of magnesium binding to albumin. A recent
treatment by Frye et at. [20] shows that the equilibrium follows a pH dependent mass law

304
relationship. The data of table 2 [20] can be reduced by linear regression to show the pH
dependence of the Mg-albumin complex stability constant:

= 142.65 pH - 952.79. (13)

Assuming six ion binding sites per molecule of albumin,

6 K,(Mg2"')Alb
(MgAlb) = '

. (14)
1 + (Mg^^)

No data has been found on magnesium binding to globulins and, therefore, the assump-
tion is made here that calcium and magnesium bind identically to y-globulin.

The treatment used in this model assumes that calcium and magnesium do not compete
for the same binding sites on albumin or globulin. Data on the competitive binding of
these two metal ions to proteins does not appear tn be available. However, it must be
noted that under normal physiological conditions not all binding sites are occupied and
therefore the assumption of independent bindjng of calcium and magnesium may not be un-
reasonable.

D. Activity coefficients

Activity coefficients for all uncharged species were assumed to be 1, as were the
activity coefficients of proteins and metal -protein complexes. The activity coefficients
for all other ionic species were calculated by means of the Davies equation [21]:

- logy. = 0.509 nj
^
M— + /r
0.31 (15)

where n-j is the ion charge and I the ionic strength. This equation takes no account of
the individuality of ionic activities nor is it particularly accurate for trivalent species.
However, only one such ion (citrate) was included, and at low concentrations, so the equa-
tion was considered adequate.

Following the convention of Scatchard [18], the ionic strength was considered to be a
function only of non-protein ions but was otherwise defined conventionally;

1-0.5 En? - C. (16)

where C^. is the ion concentration.

3. Computer Simulation of the Model

Figure 1 shows a flow diagram of the computer program used for numerical simulation
of the model. The program is written in Tymshare Fortran IV; full details may be obtained
from the authors.

Input data to the program includes: pH; Na, K, Ca, Mg, total concentrations; albumin
and globulin total concentrations; the total concentrations of each of the seven anions
given in table 1. Initial values for the ionic strength (equal to the sodium ion concen-
tration) and free metal ion activities are set. The estimates for the monovalent cations
are obtained from the products of the concentrations and the activity coefficients derived
from the estimated ionic strength. Estimates for divalent ion activities were obtained by
trial and error; for normal plasma these values were set equal to half the product of
concentration and the estimated divalent activity constant.

305
 INPUT
DATA

SET INITIAL
CONDITIONS

ACTIVITY
COEFFICIENTS

FREE
LIGANDS
FLOW CHART OF
METAL - PROGRAM 'PANIC
PROTEIN
COMPLEXES
METAL-
LIGAND
COMPLEXES

Figure 1. Flow diagram of computer program used to simulate model.

The program then proceeds to solve equations (6) to (9) to calculate the free ligand
and metal -ligand complex activities. Next, the concentration of protein bound metals are
calculated using equations (11) to (14). At this point it is possible to make a better
estimate for the free metal ion activities using equation (10). The program now tests for
convergence by comparing the initial estimates. If these differ by more than 1 percent,
the free metal ion activities are reset to the mean of the initial and new estimates, and
the calculations repeated. In this iterative fashion a solution is usually quickly ob-
tained to the set of equations.

4. Results and Discussion

A real advantage of a model of a system as complex as blood plasma is that it allows

the effects of variables to be investigated when these may be experimentally inaccessible.
Thus, it is possible to examine the properties of plasma when all the components take
extreme high or low values; clinically such a situation is extremely unlikely. Table 3
shows the calculated ionic strengths and activity coefficients in such hypothetical plasma
samples compared to normal values. Monovalent activity coefficients are in the range 0.76
± 0.02 and divalent coefficients in the range 0.33 ± 0.03. The constancy of the activity
coefficients over the extreme physiological range indicates that either an activity or a
concentration scale may be used to obtain medically significant results. In the parti-
cular case of ion-selective electrode measurement of electrolytes, it is possible to
calibrate the electrodes on a thermodynamic activity scale or on a concentration scale by
appropriate choice of standards at a physiological ionic strength [22]. It must be noted
that significant differences exist between concentrations expressed in terms of plasma

306
 Table 3. Comparison of normal and abnormal plasma.

Input data (total concentration) Low Normal High

pH 7.0 7 4 8.0
Na meq/1 120 140 170
K meq/1 2 4 8

Ca mg/dl 5 10 15
Mg mg/dl 1 2 5 5

Albumin g/dl 2 4 7 8
Globulin g/dl 1 2 5 5

CI meq/1 70 100 130

CO2 meq/1 10 30 40
-
P mg/dl 1 3 7 10
S mg/dl 0.5 n
u Q 1.5
Citrate mg/dl 2.4 2. 4 2.4
Lactate mg/dl 10 10 10

Pyruvate mg/dl 0.6 0. 6 0.6

Calculated values

Ionic strength 0.108 0. 152 0.204

Monovalent activity coefficient 0.778 0. 759 0.746
Divalent activity coefficient 0.365 0. 332 0.310

volume and plasma water volume. Table 4 shows the effect for the four commonly determined
electrolytes. Calibration of an electrode with aqueous protein-free standards results in
plasma measurements on the plasma water volume scale.

The model predicts the activities of 46 components of plasma; table 5 shows the
calculated distribution of these in normal plasma. It is also possible to investigate the
effect of one variable on others, e.g.^ the pH dependence of ionized calcium and magnesium
(fig. 2). These results indicate possible applications of the model; however, it is not
intended to explore these in detail here.

Table 4. Comparison of normal plasma electrolyte values

on different scales.

Total concentration Ion concentration Ion activity

meq/1 plasma meq/1 plasma HgO meq/1 plasma H2O

Na 140 149 113

K 4 4.3 3.2

01 100 107 81

CO. 30 30^ 23^

HCO3

307
 - —

un 1^ ID CO LT)
1

o O 1

o 1

o 1

o o 1

o 1

o 1

o O 1

o
X X X X X X
X
in CO
X
(Ti
CM
X
CO
00
CM oX o
X
CO
tn
un
<£< CO CM CM in

+ -=3-

+ CO O +-> + +
o O
O Q- O
CM Q- O o s-
>>
CD
CD
CD
n: ic 31
C7> CD
00 DC
C7>
o C7>
_J
CD cn
s: SI

oCM
co ID in
(/>
E
(O
1

o 1

o 1

o
1
1

o
—
1
1

o 1

o 1

o o
r—
c
X X X X
Q- ^X X X
vo
X X
CM
X X
in
o O
1 1

CM CM CM >=1- Ln CO CM
CM >5l- 00

o
1-
oCO oCM o
Q. o
1

+->
+
u
"3
s-
>>
J
n3
ct
fO
CJ3
fO
<J>
zr.
10
:jz
ro
nz
10 fO
O
to
_J
-l->
O O C_) o O o O o o C_)
a;

o
Q-
E
O
o CO U3
10 O o I

O
E
(/) X
r—
X X
CM
CM CT> m

o
o. o o
o
4-)
in
s-
!-> o
—
r
o
cy> X X X
o
E
o
o
o
in
o I

o >=1-

J3
ro n3
O

30 CO CM
o ID CO
o in
1

O 1

o 1

o 1

o 1

o I

o 1

o o I 1

o 1

O 1

o 1

o 1

o o 1 1

o
X X X X X X X X X X X X X X X
«^
CM
CM
00
to
in
o o CD
CTl CO
CO
o
en
CTl
CM
in in
in
00 to
CM
Ln o CM CM CO CM CM CM

E
E
3
Z3
CO o
^ 1 II
+-> o s-
ja o
oCM o
o
1

D- o
CM Q.
II "S-
o
00
II

O >=d-
<_)
III
4-> <0
_l
o >>
Q.
's-

5 5 o o n: nr oo JZ O zc _J a: a.

308
 U.4 Figure 2. Calculated pH dependence of ionized
calcium and magnesium in normal plasma.
n9

0 7.0 7.2 7.4 7.6 7.8 8.0

pH

It may be argued that the model is incomplete in that it does not require electro-
neutral ity--a condition which is always satisfied for an actual plasma sample. The ab-
sence of this condition is intentional and, in fact, it is not required of the model.
Electroneutrality is obtained at a given pH by charge balance between the proteins and
other ions. Perhaps the most serious criticism of the model is that it relies entirely on
published equilibrium constants, the accuracy of which is not always beyond question. If
it serves no other purpose the model emphasizes the need for reliable thermodynamic measure-
ments of these constants at various temperatures and at physiological ionic strength. It
is to be hoped that this treatment of ionic equilibria will lead to a clearer understanding
of the species present in blood plasma and their role in metabolic processes.

References

[1] Collins, R. D. , Illustrated Manual of Laboratory Diagnosis, Second Edition, Chapter

4, p. 35 (J. B. Lippincott Company, Philadelphia, 1968).

[2] Moore, E. W. Ionized calcium in normal serum, ultrafiltrates and whole blood determined
,

by ion-exchange electrodes, J. Clin, invest. 49^, 318 (1970).

[3] Weissman, N. and Pileggi, V. J., Inorganic ions, in Clinical Chemistry Principles and
Techniques^ 2nd Edition, R. J. Henry, D. C. Cannon, and J. W. Winkelman, eds.,
p. 662 (Harper and Row Publishers, Maryland, 1974).

[4] Moore, E. W. and Wilson, D. W. The determination of sodium in body fluids by the
,

glass electrode, J. Clin, invest. 42^, 293 (1963).

[5] Friedman, S. M. , Wong, S., and Walton, J. H., Glass electrode measurements of blood
sodium and potassium in man, J. Applied Physiol. ]%_, 950 (1963).

[6] Portnoy, H. D. and Gurdjian, E. S., Glass electrode determination of blood sodium and
potassium. Am. J. Clin. Path. 45^, 283 (1966).

[7] Miyada, D. S., Inanu, K. and Matsuyama, G., Direct potentiometric determination of
potassium and sodium in blood, plasma and serum, Clin. Chem. ]7_, 27 (1971 ).

[8] Dahms, H., Automated potentiometric determination of inorganic blood constituents

(Na+, K+, H+, CI), Clin. Chem. U, 437 (1967).

309
[9] Neff, G. W., Radke, W. A., Sambucetti, C. J. and Widdowson, G. M., A computer-assisted
electrode system for measuring blood pH, PO2, PCO2, sodium and potassium, Clin. Chem.
1_6, 566 (1970).

[10] Orion Research Incorporated, Automated Potentiometrio Eleotrolyte Analysis System,

Final Report, Contract NAS9-12117 (National Aeronautics and Space Administration,
Houston, Texas, December 1973).

[11] Steele, R. E., Walker, W. E., McWhorter, M. M., Lewis, K. A., Robertson, C. R.,
and Maffly, R. H., Critical care blood gas, pH and electrolyte analyser, Pvooeedings
of the 27t'h Annual Conference on Engineering in Medicine and Biology, Vol. 16, p. 401
(The Alliance for Engineering in Medicine and Biology, Chevy Chase, Maryland, 1974).

[12] Walser, M., Ion association VI. Interactions between calcium, magnesium, inorganic
~
phosphate, citrate and protein in normal human plasma, J. Clin. Invest. 40, 723
(1961).

[13] Robertson, W. G. , Measurement of ionized calcium in biological fluids, Clin. Chim.

Acta, 2A, 149 (1969).

[14] Dahms, H., Rock, R. and Seligson, D. , Ionic activities of sodium, potassium and
chloride in human serum, Clin. Chem. 1_4,859 (1968).

[15] McLean, F. C. and Hastings, A. B., The state of calcium in the fluids of the body.
I. The conditions affecting the ionization of calcium, J. Biol. Chem., 108, 285
(1935).

[16] Sillen, L. G. and Martell, A. E., Stability Constants of Metal- Ion Complexes, Chemical
Society Special Publication No. 17 (Metcalf and Cooper Ltd., London, 1964).

[17] Carr, C. W., Determination of ionic activity in protein solutions with collodion
membrane electrodes, in Electrochemistry in Biology and Medicine, T. Shedlovsky
ed., p. 266 (Wiley and Sons, Inc., New York, 1955).

[18] Scatchard, G. and Yap, W. T. , Physical chemistry of protein solutions. The effects
of temperature and hydroxide ion on the binding of small anions to human serum albumin,
J. Am. Chem. Soc, 86, 3434 (1964).

[19] Pedersen, K. 0., Binding of calcium to serum albumin II. Effect of pH via competitive
hydrogen and calcium ion binding to the imidazole groups of albumin. Scand. J. Clin.
Lab. Invest. 29, 75 (1972).

[20] Frye, R. M., Lees, H. and Rechnitz, G. A., Magnesium-albumin binding measurements using
ion-selective membrane electrodes, Clin. Biochem. 7_, 258 (1974).

[21] Davies, C. W., Ion Association (Butterworths , London, 1962).

[22] Bates, R. G. and Mohan, M. S., Studies in the calibration of ion-selective electrodes
for use in biological fluids, Clin. Chem. 21_, 864 (1975).

310
National Bureau of Standards Special Publication 450. Proceedings of a Workshop on pH
and Blood Gases held at NBS, Galthersburg, Maryland, July 7-8, 1975. Issued June 1977.

THE KING'S COLLEGE HOSPITAL ION-SELECTIVE ELECTRODE SERUM

ELECTROLYTE ANALYZER

A. D. Hirst, P. Gay, P. Richardson, and P. J. N. Howorth

Department of Chemical Pathology
King's College Hospital Medical School
London, U.K., SE58RX

In recent years there has been a great increase in the availability of ion-selective
electrodes, and these are finding wider use in routine clinical chemistry laboratories
(table 1). The obvious difficulty is that such electrodes measure activity which must be
somehow related to concentration by calibration and conversion factors. Early reports,
e.g., with the Na electrode suggested that the conversion factor might vary in disease
states when abnormal concentrations of protein were present [1]^.

Table 1. Some ion-selective electrodes.

Cations Anions

h"" (pH) F" CI"

NH3 (NHt) Br" r

Na" k" CN" NO3

Ca2"^ Cd2"^

Cu2"^ Pb2"^

We felt that it would be an interesting and potentially useful experiment to construct

a multi-channel continuous-flow serum electrolyte analyzer based entirely on ion-selective
electrodes. This would dispense with the need for chemical reactions to first make a colored
solution, pass light through it, measure what emerges and then by complex electronics, convert
it to the concentration of the substance being measured.

Table 2 shows the general principles adopted in the analyser. The Na electrode is
probably the most critical part of the system. The Na electrode is based upon an alkaline

Table 2. General outline of KCH analyzer.

Ne" electrode: acid-serum dialyzed into high-osmolar

alkaline buffer

k" electrode : Na" activity backed-off

HCO3 : pH electrode CO2 from acidified serum

diffuses into Na2C03

Glucose : O2 electrode. High concentration glucose

oxidase causes quick fall in PO2

Urea : NH3 electrode. Urease NH3 from urea

Figures in brackets indicate the literature references at the end of this paper.

317
error in a glass pH electrode. If serum is made alkaline with buffer, then protein inter-
ference can be a problem. By dialyzing an acidified serum sample into an alkaline solution
of high molarity, a linear response to changes in Na concentration can be obtained. Figure
1 shows the flow- diagram for the analyzer.

s ample line

waste
dialyzer ai r

t 0.2 mol HaSOit

Na"*" electrode TT 1 mol NHijOH/CI
air

K"*" electrode 1 mol triethanol amine

air HCl pH 6.7

waste
dialyzer ai r
1 0.2 mol HaSOit
H"*" felectrode weak Na2C03'
ai r

^0.5 mol NaOH

NH3 electrode 4 ureas e/EDTA

.air pH 6

gl ucose oxi dase

I

O2 electrode q- 0.15 mol NaCl

-ai r

Figure 1. Flow-diagram for KCH Ion-selective electrode serum electrolyte analyzer.

The K-selective glass electrode is responsive to an appreciable extent (activity

ratio about 10:1) to Na ions, but fortunately the Na response is linear and additive.
Accordingly, K is measured after Na by means of time-delay coils and the Na contribution
from the Na electrode is electronically backed-off from the total ion activity. The serum
HCO5 measurement* is very simple. About 50 percent of the CO2 liberated from acidified
serum can be dialyzed across a standard membrane into a Na carbonate buffer whose pH
alters with varying CO2 content.

The measurement of glucose by an O2 electrode gave trouble at first owing to the O2

content in the air-segmentation bubbles in the mixing coils. It was not found necessary,
however, to use inert gases for this purpose. By using a high concentration of glucose
oxidase, the rapid fall in PO2 found in the first 30 seconds was found to be proportional
to the glucose concentration. The urea method is potentially as accurate a method as it
is possible to devise and should be superior to the non-specific diacetyl monoxime reaction.

312
The urease reaction takes place at a slightly acid pH after which the sample is made
highly alkaline (pH 12) before being fed through the NH3 electrode.

The electrolyte analyzer has been assembled at King's College Hospital from commer-
cially bought electrodes, sampler, pump, dialyzers, mixing coils, heated water bath,
recorder and electronic components. Unfortunately at the time of writing, it is not yet
in routine operation although all the individual channels have been worked up prior to the
final design of the machine. Figures 2 and 3 show general views of the analyzer.

Figure 2. View of KCH analyzer.

Figure 3. Close-up view of analyzer.

313
 All the arguments that have raged for 60 years about the measurement of H concentra-
tion or activity apply in some measure to ion-selective electrodes. However, since no one
in practical terms really wants to measure or report urea as "PNH3" or K"^ as "pK" [2],
then we will all make the necessary effort to adapt in a reasonable manner the activity
measurements to concentration units. All our routine methods have inherent errors or
compromises of one sort or another. We really do not see any good reason at all why we
cannot also make an effort with regard to the pH-glass electrode and use a single unified
set of units such as the SI provides for all our common electrolyte and blood gas analyses.

References

[1] Dahms, H. Rock, R., and Seligson, D. , Ionic activities of sodium, potassium, and
,

chloride in human serum, Clin. Chem. ^4, 859 (1968).

[2] Howorth, P. J. N. and Hirst, A. D. , SI units, pH and cH"^, J. Clin. Path. 28, 423 (1975).

314
 WORKSHOP DISCUSSION

The transcription of selected segments of the discussion sessions is included in this

volume because it supplements much of the material in the preceding chapters. In this chap-
ter will be found additional comments on practical problems, instrumentation, methodologies,
and concepts on the subject of blood pH, gases and electrolytes.

The workshop was arranged to provide time for discussion at the conclusion of each
series of lectures on a particular subject. The following discussions were transcribed with
a minimum of editing and represent the spontaneous exchange of remarks of the various parti-
cipants as they were recorded at the workshop. The speakers have had an opportunity to
correct the transcription of their remarks for technical accuracy.

Bates :

Before this session began, several of us were reminiscing about the meeting in 1964 at
the New York Academy of Sciences and I recall in 1964 there was a great deal of discussion
about two things, one of them being the validity of so-called "unlogging" the pH value, and
the other one was concentration versus activity, in other words, concentration versus excess
free energy. It seems rather remarkable that in eleven years the utility of one of these
hasn't proved itself. We still have the same problem of which to use, and it occurs to me
that this might be due to the fact that the experimental methods for determining pH, for
example, do depend more closely on an activity than on concentration, and I wonder if this
is the reason that this dilemma still faces us.

Cohen :

My view of this is that this is not a dilemma but a sematics difference. When hydrogen
ion concentration is spoken of, at least in the parlance of clinicians and physiologists, I
think what is meant is "the negative antilogarithm of the pH" which is, as you point out, an
activity measurement. This is not to say that one is ignoring or brushing aside the dif-
ference between concentration and activity, but that one is simply using a shorthand notation
to designate measured quantity in terms that are more readily conceptualized. And I am
wondering whether you would view this as an acceptable "derivation" from the measurement of
pH or whether we ought employ a different term for it.

Bates :

Well, it seems to me that we are really begging the question of whether it's an activity
or a concentration that is crucial in biochemical and biological processes, aren't we?

Cohen :

Well, that is what I mean to imply. That is, I don't think anybody really quarrels with
the notion that it is the activity that is fundamentally at issue, that biological systems
and chemical systems must be responding to the chemical activity. That being the case then,
I think it's a matter of arbitrary definition as to what one choses to call the unit of
measure as long as it serves the need for unambiguous communication.

Austin :

I think that perhaps it may be the fact that pH is essentially a "dimensionless" factor
that is important. pH serves as an indicator in a given situation, and it avoids the issue
of applying other than a unit dimension. Is it a nano-equivalent? How does that fit with
a milli-equivalent? I think that is one important factor, and it speaks for the continued
use of pH. We are dealing with rather gigantic differences when we're talking about milli-
equivalents versus nanomoles. One can not argue that we should then speak of the log of the
sodium or potassium concentration, since we administer these in their actual concentrations,
even though their biological activity is what is important. Though we give bicarbonate in
mEq to correct hydrogen ion concentrations/activities, no one corrects acidosis, for example,
by hanging a bottle of H+.

315
We is berg :

I have two thoughts on pH versus the hydrogen ion concentration. Physicians have de-
graded themselves by saying they can't understand pH, whereas chemists as a group use pH,
We're not going to be able to change the rest of the chemical field by saying we're going to
use hydrogen ion concentration. Secondly, I find that those who speak for the hydrogen ion
concentration are really deluding themselves because a change of 10 nanomoles in a hydrogen
ion concentration has a completely different interpretation when it's on the highly alkalotic
side versus the highly acidotic side. So you're going to give a greater importance to one
nanomole of hydrogen ion where the effect on the body enzyme activity, and so on, is com-
pletely different. So my own feeling is that I give both pH and hydrogen ion concentration
in my report on the diagram; you can use whichever you will. But if you realize, as most
of us do, that a pH unit change is a tenth or ten times going the other way, we can get that
concentration effect the same way. We're dealing with numbers, but we're giving undue impor-
tance to a number of the hydrogen ion concentration.

Cohen :

Well, I think one could find as many examples of biologic relationships that would be
linear with respect to hydrogen ion concentration as you could with respect to pH. I don't

really think the argument that enzyme systems or other things are responding linearly to pH
rather than to hydrogen ion concentration or activity is a very compelling one. I think the
only compelling argument, at least in my mind, for retaining, or perhaps substituting, the
notation, "hydrogen ion concentration", (with the caveat that I mentioned earlier) is that
it simplifies the mathematics that describe the bicarbonate-carbonic acid buffer system.
That is the only justification that I can see, but it seems to me that is one that can't be
minimized. It has some very useful pedagogical and practical advantages.

Engel :

I understand Dr. Bates' surprise because it has taken a long time since we started these
discussions at the meetings at New York Academy of Sciences. I think also that you should
realize that the people working in this field of acid-base physiology are a very inhomogeneous
group of people. We have pure chemists, we have physiologists, we have medical doctors in
the ward, etc, and the interests of the different groups of people are very different and,
therefore, the emphasis is sometimes put on one side of the aspect. The other thing I would
like to comment on is the old questions around the "great trans-Atlantic acid-base debate"
which was brought up again. I don't think it has ever stopped even though it sort of dampened
for a few years. I think, however, that we should agree to divide our problems in three dif-
ferent groups. First, we have some basic definitions which apply to clinical chemistry as
well as to pure and applied chemistry. I think it's very feasible to use these very well
systematized and well defined concepts. Second, as clinical physiologists and as acid-base
physiologists, we might be interested in some derivatives of these well-defined and measurable
variables and, of course, we would like to define some variables or derived variables which
make physiological sense. And thirdly, we have the problems related to the clinicians with
pathophysiological interpretation of these either derived or basically measured values. I

think that we have been very good in the past of mixing all three of these groups together in
one melting pot, and that lots of the questions which were discussed during the "great trans-
Atlantic acid-base debate" came about because of this maxture of what our different goals
and definitions were.

Howorth :

I'd like to go back to something that people were talking earlier on about pH as opposed
to so-called concentration units. I think many physicians in practical terms need some sort
of concentration unit. I mean if a surgeon is aspirating gastric juice, he likes to have
some measure of how much acid is being removed from that patient. You also need a quantita-
tive measure with renal excretion of H+ and renal function tests. In everyday practical med-
icine we need to know how much acid a patient is putting out in his urine on a quantitative
basis. This is where the mole as the amount of substance is superior to the dimensionless
pH concept.

316
Cohen:

Is it not true that in other systems where the measurement is done electrochemically,
for example, calcium, we report the substance in concentration units when we recognize that
we are measuring it electrochemically? I don't think that poses a problem.

Noonan :

Well, I think a fundamental concept in analytical chemistry is the idea of recoverabil-

ity. I can say that I am measuring glucose quantitatively if I can recover a certain amount
of pure glucose which has been added to it. The glucose technique is directly measuring the
concentration. However, I cannot take a certain quantity of hydrogen ion and add it to a
solution and measure it with an electrode and say that it is concentration because concen-
tration is not being measured. In terms of ion-selective electrodes for sodium and potassium,
and indeed pH, you can get some constant correlation between the activity and the concentra-
tion that will work in many, many cases but will not work in all cases. If you really want
to be precise, there should be a difference between a sodium measurement that's made with an
electrode and a sodium measurement that's made with a flame photometer.

Austin :

I agree, sodium concentration is important when you are replacing it, but with pH, when
considering dissociation equilibria (Davis,' R. P., Amer. J. Med. 42^, 159-162, 1967), the quan-
tity of hydrogen ion that you're measuring is so much smaller that the quantity of sodium.
They are of an entirely different magnitude. This may create some difficulties with measuring
diminutive quantities of hydrogen ion where we're measuring an entirely different number of
sodium and potassium ions. How does this help you? Apart from concentrations, pH denotes
a more meaningful description of the acid-base equilibrium reaction. If we are talking
about acid-base balance that's one thing; ionic concentrations or activity are another matter.
Recall in the Henderson-Hassel balch equation that we do take the log of the bicarbonate con-
centration, even though we give bicarbonate in its sodium form for practical purposes.

Gambino :

I think I'd like to say that I agree basically with Siggaard that it is preferable to go
to some type of measurement or some type of indication of the amount of energy and the quan-
tity of substance, but I'd like to have some measure of the energy. I think it's important
to point out that we cannot utilize the suggestion that the way we make the measurement
should determine the unit. For example, in temperature we estimate the average kinetic
energy, we measure the movement of a mercury column and we measure current in a system. But
we don't attempt to say that because we measure current that we should then define temperature
as an estimate of kinetic energy in terms of current. We don't do that. We go back to some
fundamental quantity such as the joule. I think we should try to get a quantity into these
fundamental units.

Durst :

There's something that I picked up from our discussions that I might just mention now.
That is that we're using the word "derived" in two different ways. One is in reference to
the SI system where derived means units expressed as products or ratios of the base units
without numerical factors as opposed to what we are calling derived units such as bicarbonate,
that is to say, a parameter calculated from a measured parameter or parameters. Again, this
in one more example of a semantic problem which could be significant.

Weisberq :

May I add one more semantic problem. You're using the word nomogram erroneously. I

have collected 105 diagrams, 17 nomograms, and 12 slide rules. The first diagram goes back
to 1914, and I think we are really referring to diagrams rather than nomograms. Dr. Malenfant
had a true nomogram. A Cartesian nomogram and a diagram are two completely different things.
A diagram does have benefits for a practitioner to see, especially with continuing sets of
data. One set of data means nothing; a point in time of compensation or uncompensation. But

317
if you can tell where the patient has been or where the patient is going and thus follow that
on a flow chart. Dr. Siggaard-Andersen had, of course, the time chart going along with that.
So I think we ought to be clear that we are speaking of diagrams most of the time for physi-
cians to look at. The nomogram is a laboratory device to save ourselves some mathematics.

Howorth :

I think a nomogram is something that one uses to derive a figure from primarily observed
data. A parameter isn't something which is directly measured, it's something that's indi-
rectly measured.

Vi sser :

Well, there will be no confusion in this case between "derived" and "derived", because
base excess is a derived quantity, but maybe another derived quantity can be expressed in a
basic unit. So we must distinguish between the quantity and the unit.

Cohen :

I'd just like to comment on Dr. Engel's presentation. I had two commentaries to make;
one trivial, really, and the other perhaps more substantive. The trivial one being that the
use of the term "net acid" in his scheme, if I understood it, is somewhat at variance with
modern definitions of acids and bases; it doesn't seem to conform to a proton donor. The
second is the issue of simplicity. I think there is nothing internally inconsistent about
that system for describing, in phenomonological or empiric terms, what happens to living
organisms when they are challenged. The only quarrel I have with it is that it seems to me
to be unnecessarily compl icated. One of the issues that surfaces in my mind, as a dominant
theme, relates to the fact the system with which we are dealing is inherently complicated,
simply by virtue of its physiology being so multifactorial. We are, therefore, better served
by the conceptual framework for discussing that system's behavior which is as simple as we
can make it. I don't see any compelling justification or reason for a system, such as
described, in contradistinction to one which deals just with the carbonic acid/bicarbonate
equilibrium and describes what happens to the organism in those very simple terms.

Engel :

If I can take the last comment first, then if you did total body studies of acid-base
metabolism, you would very quickly run into the same problems which we have been running
into. That is, you can't describe them with the set of variables we have been using so far.
When you define the body as a system, I think you should make it just as complicated as is
needed in order to completely describe the system. This is exactly why we wanted divide ti-
tratable acid up in three components: Carbonic acid which is regulated by the lung-brain
system, metabol izable acid which is regulated by the metabolism, and the net acid for which
the only route of escape is the kidney, and this makes physiological sense. I don't know

whether the scheme for the breakdown was too complicated or not, but anyway we feel that this
is the only way where you can go in and put a finger on the physiological mechanisms and
causes which are involved in an acid-base disturbance. One thing more is that you may mea-
sure, as I said, a normal acid-base status in the blood and yet have a very pronounced acid-
base disturbance in terms of balance studies.

Cohen :

I don't think there is any question about this latter point. Clearly, we can have a
very markedly disturbed external balance without having the deviation.

Engel :

Right, but you won't find out unless you carry out the balance measurements.

Cohen :

Of course, but let me just take those two issues in turn. One is that I think it is
possible to designate external balance of acids and bases without utilizing that breakdown

318
factor of acidity in the classic context of the net acid balance. In other words, there is
a certain rate of net acid excretion, by the kidney, which we all recognize as ammonia plus
TA minus the bicarbonate that is excreted. In the steady state, this balances the rate of
endogenous acid production. Now, there is no satisfactory way, nor do I think your system
provides me, for quantitating directly the rate of endogenous acid production. Given a
diet, one can't use that scheme, for example, to specify what its net contribution of hydro-
gen ion or base will be to the body fluids, any more than one can take a diet and analyze
it for this purpose in any other way. So one is left with a conundrum that's been ^n issue
in physiology from the beginning of time; that is, unless one constructs a purely art.-^icial
diet which can be demonstrated to be complete metabolized, a la Relman, external hydrogen ion
balance must be estimated indirectly since we have no handle on the rate of endogenous acid
production. For this purpose, we use the rate of net acid excretion by the kidney, measured
in terms of hydrogen ions which are gaining access to the urine, as a fix on that parameter.

Engel :

Well, I think that you can assess the endogenous net acid production. But there is a
discrepancy between what we and you call net acid. Net acid in our terms is defined as the
titratable value (TA) to an endpoint at pH 7.40 and temp. = 37 when carbonic acid (CA)
and metabol izable acid (MA) concentrations are zero. This is the operational definition
and it gives you values which are quite different from the net acid values you are talking
about. NA can be measured in food, you can measure them in feces, and you can measure them
in the urine, etc.

Cohen ;

I certainly don't quarrel with the fact that it can be measured, but the issue is to
what end, what use can be made of the information once available. Let me, by way of illus-
trating that, ask you to analyze the in vivo acid-base disturbance that I illustrated in
my talk in terms of the components of your "titratable acidity" and to reconstruct the se-
quence of events that led to that situation.

Gambino :

That's just the question that I was going to ask. My analysis of what you said, based
on what Dr. Engel said, gives a different interpretation of it. You said your paper described
the effects of hypocapnia. But I saw it as a secondary response to using your model to dif-
ferent types of acid load. You had an oral acid load, increase in hydrogen ion load absor-
bancy, and you had 9 percent oxygen, which I would suspect to be a primary anoxic acid load
with lactic acid.

Cohen :

No, Dr. Gambino, there was no lactic acid production in that experiment. That degree
of hypoxia produces no increase in lactate or unmeasured anion concentration at all.

Gambino :

How do you explain the bicarbonate?

Cohen :

That was the consequence of the renal response to the hypocapnia; the reduction in
bicarbonate reabsorption at the renal level resulted in a steady-state reduction in bicarbon-
ate concentration, unaccompanied by any change in unmeasured anion concentration, no measured
lactate accumulation. In other words, the decrement in bicarbonate was offset entirely by an
increase in chloride concentration and the change in bicarbonate being accounted for by a
reduction, by my terminology, of net acid excretion. There was an accumulated change in acid
excretion by the kidney, that led to retention of endogenous acids that reduced bicarbonate
to that extent. So what one is seeing in that experiment is the physiologic response of the
kidney to two independent variables. One, the oral hydrochloric acid load, and two, the
effect of the hypocapnia on renal bicarbonate reabsorption.

319
Weisberg :

If what you say is true, and I'm not doubting that, if you have that with a normal dog
without having the previous hydrochloric acidosis and have the same delta that you had with
the decrease in the PCO2 with the hypocapnia, why shouldn't you have the same effect in the
kidney where the loss of the bicarbonate is leading to that acidosis?

Cohen :

We find the same retention of hydrogen ion in the two situations, whether one goes from
normal to the chronic hypocapnic state or from the acid-fed to the chronic hypocapnic state.
The difference is in the initial level of plasma bicarbonate. In one, it is normal, and
in the other it is low due to the acid feeding, but we are superimposing the same delta bi-
carbonate on these two different baseline levels; as a result, we see the divergent responses
of hydrogen ion. There is nothing magic about this. Given the change in PCO2 and the
physiologic response of the organism thereto, the change in acidity is predictable from the
relationships described by the Henderson equation.

Siggaard-Andersen :

It should be kept in mind that the same end result is obtained if (1) the initial dis-
turbance is an acute respiratory alkalosis which is subsequently modified by metabolic com-
pensation ending in a chronic respiratory alkalosis, and if (2) a metabolic acidosis is
initially induced before the respiratory alkalosis develops.

Cohen :

I certainly agree with that. I don't think the sequence in which these two maneuvers
are employed is crucial. One could reverse them and end up in the same situation, I would
predict. But we haven't done that experiment. But I still think that the issue is how does
one unravel such complex situations. What is the conceptual scheme one utilizes to dissect
out the two components which are in that acid-base disturbance; it would be instructive to
see how Dr. Engel would do that.

Engel :

Yes, I should first like to draw a parallel to that. We were at one point measuring
acid-base balance on patients with pyloric stenosis and, as you know, the interpretation of
the concurrent alkalosis has always been that, they developed alkalosis, because they are
losing hydrochloric acid in the vomit from the stomach. Now making, for five or six days,
balance studies on patients with pyloric stenosis showed that the alkalosis only partly was
due to a change in the net acid balance of the patients, and that most of the deviation in
Base Excess could be related to the fact that the patient at the same time lost large volumes
of water. Thus, it was rather a "contraction" alkalosis. I think that if we want to bring
about new ways of assessing acid-base metabolism in practice, and if you want to learn new
things, you cannot avoid measurement of external acid-base balances.

Siggaard-Andersen :

One of the problems is that the lactate ion behaves chemically as a neutral substance
when we try to titrate it to pH = 7.4. Nevertheless, we prefer to consider it a base from
a cl inical -physiological point of view. In Engel 's terminology, lactate is called a metabo-
lic organic acid (MOA) with a negative sign. I prefer to call it a metabolic organic base,
in other words to distinguish between metabolic acids and metabolic bases, in order to avoid
talking about a component with a negative sign. The negative sign should not refer to the
component but to the measured quantity: the change in the amount of substance of the com-
ponent in the organism or the body fluid.

I would like to use this opportunity to make a plea for the SI system. It is important
that the SI system, which has been advocated quite generally for science and technology, is
also adopted by the clinical sciences. It is important that those of us who are involved in
the measurement of physical and chemical quantities know the terminology of physics and
physical chemistry which constitutes the basis of clinical chemistry and clinical physiology.

320
However, we cannot expect all physicians to remember all their physics and chemistry, and
therefore it is our responsibility to translate the information contained in the clinical
chemical data into common language.

The acid-base data including electrolytes represent a complicated set of data, perhaps
even including redundant data {e.g.^ total CO2, bicarbonate, base excess, buffer base). Much
of this information may be contained in "diagnoses" such as: partly compensated metabolic
acidosis, hypokalmeic alkalosis, etc. We may have to make slightly longer statements but
anyway this set of terms (acidosis, alkalosis, etc.) is very useful as a supplement to the
mere list of measured quantities.

Engel :

I agree completely with Siggaard, and I think that there is one pitfall we should avoid,
and this is to let our instruments mix interpretation into the measured quantities or into
the derived quantities. I think the manufacturers should stay far away from that. Interpre-
tation is the business of the clinical physiologist or the clinician dealing with the patient

Bates :

I'd like to ask Dr. Austin a question. Was your pK' a concentration constant or was it
the same as Dr. Maas' constants?

Austin :

It was the pK arrived at by determining the pH of whole blood, fixing the PCO2 by
known gas concentrations and measuring the CO2 content by the Van Slyke manometric technique.

Bates :

In other words, the hydrogen ion is supposedly an activity but the others are partial
pressures.

Austin :

Right.

Bates :

They are remarkably constant.

Cohen :

Harkening back to Dr. Bates' first question from this morning, I would like to ask
Dr. Maas if he would have any objection to simply taking the negative antilogarithm of the
equation which you put up, I've forgotten what the numbers were, I think 16 or 17 or some-
thing of that series. In any event, would he be willing to express the hydrogen ion con-
centration, in quotation marks, as being equivalent to the ratio of PCO2 and bicarbonate
ion concentration multiplied by the negative antilogarithm of your pK value. In this way,
simply converting into linear units of activity or concentration, the logarithmic units of
pH, in order to specify the level of acidity.

Maas :

You can consider an equilibrium between gaseous CO2 and bicarbonate directly without
looking at the H2CO3 in between. So the answer to your question is, yes, the ratio PCO2 and
bicarbonate.

Noonan :

I wonder if we can get some discussion on this whole idea of talking about the plasma

pH, plasma PCO2 versus the similar quantities in whole blood. Does anyone have any feelings
about which it should be, or should we have one, or should we always talk about plasma pH in
measuring blood pH.

321
Austin:

The difference is about 0.01.

Weisberg :

The difference is that we now have convenience. The machines usually use whole blood;
remember, we used to use plasma. Plasma is wonderful because you never clog the capillary
electrodes. When you have whole blood, you have all the problems. But we're lazy. Machines
are made that way, so it's easier to put in whole blood. And, of course, it's easier mechani-
cally, but technically, as Dr. Austin pointed out, we're measuring the same thing. So it
makes no difference technically. Clinically, that 0.01 difference between plasma pH and
whole blood pH is not going to make any difference. And there's no major difference between
the PCO2 in plasma and whole blood as far as clinical purposes are concerned. I don't see
you point, but it is interesting.

Gambino :

We don't have thermostated separation systems. So you have serious temperature errors
in pH but even more so in oxygen and PCO2. The other factor will come out tomorrow. These
data are utilized by anesthesiologists and ventilation therapists, etc., for the immediate
treatment of patients within minutes of having the results. Whole blood is essential. It's
not just a matter of being lazy.

We were discussing pH and base and this is a workshop on blood gases and most of my
blood measurements are performed because there is a need to know CO2 gas pressure as a mea-
sure of ventilation and the partial pressure of oxygen as a measure of diffusion. These mea-
surements are used to change ventilation therapy. And that's the most common cause for the
high volume testing, and the acid-base part of it is important but not as dominant in the
minute by minute therapy in the intensive treating. So I think we will have to also consider
the fact that there is a need for gas measurements for ventilation therapy.

Laver :

I prefer not to address myself to that problem. Dr. Rispens, to be sure I understood
you correctly, did you say that the ratio of the CO2 content of plasma to that for whole
blood was constant over a specific temperature range? Did you measure the CO2 content of
whole blood?

Ri spens :

In anaerobically stored blood the ratio of total CG2 in plasma to total CO2 in blood is
constant. I measured total CO2 both in plasma and in blood. I emphasized that it was not

influenced by the temperature at which centrifugation of the blood took place.

Laver :

Did you say whole blood? In other words, the CO2 content of plasma plus and red cells?

Rispens :

Yes.

Laver :

So, in other words, the ratio of the CO2 content in plasma to the CO2 content of whole
blood was constant over a large temperature range.

Rispens :

I emphasized that it was not influenced by the temperature at which centrifugation of

the blood took place.

322
Laver:

Let me start over again because I think it important we do not misunderstand. The
patient is at a low body temperature. CO2 content of whole blood rises more rapidly than
the CO2 content of plasma. If blood is equilibrated at a low temperature, let us say 25 °C,
and we now measure the CO2 content of whole blood and plasma, what happens to the ratio be-
tween the two as compared with equilibration at 37 °C?

Rispens :

Well, we measure the pH at 37 °C and put that in the equation [i.e., equation (2) in the
paper "Quantitative relationships ..."] to calculate the ratio between total CO2 in plasma
and total CO2 in blood.

Laver :

Let us assume we measure it by your technique, that of Van Slyke, or with the pH and
PCO2 electrode. I still keep coming back to the plasma versus whole l?lood CO2 content.

Austin :

CO2 content is constant regardless of the temperature. Dr. Severinghaus and Dr. Nunn
have shown that.

Laver :

I am sorry to be persistent, but you have not understood my question. If you equili-
brate whole blood at a certain PCO2, for example 40 mm Hg at low temperature (25 °C), either
in a patient or in a tonometer, measure the CO2 content of whole blood as well as the CO2
content of plasma and compare that to the content for each when blood is equilibrated at
37 °C, what happens to the ratio between the two?

Rispens :

I said that in anaerobically stored blood the ratio between CO2 in plasma and CO2 in
blood is constant. Thus if you manage to get in the blood the same pH at 37 °C by equili-
brating at 37 °C with a higher PCO2 than you used at 25 °C, the total CO2 concentration will
be the same and so the ratio.

Laver :

Yes, you keep saying total CO2. Yet CO2 content in the red cell and CO2 content in
plasma refers to two different things.

Gambino :

You've come back to something reported many years ago, namely, that whole blood with 15
grams of hemoglobin and with a changed temperature, you don't get any change in pH with
changing temperature. Let me draw it on the board. This is invariant with the temperature
when it is whole blood, but its not invariant with the temperature when it's plasma. I think
that's what you're driving at, right? If the temperature changes in a whole blood system,
the solubility of CO2 gas causes changes in one direction and the ionization of the proteins,
including hemoglobin, changes in the opposite direction so that the net pH remains constant.

Rispens :

That is certainly true, but that is not the point I wanted to make. I took a blood

sample, centrifuged it at 20 °C and 37 °C and found that the ratio of CO2 in plasma and
CO2 in blood was not affected.

Howorth:

It wasn't entirely clear about the method for measuring total CO2. Do you always use
plasma or do you sometimes use anaerobic blood?

323
Rispens :

I use blood.

Howorth:

One of the common techniques used in England is to separate the serum or plasma and then
measure total CO2 by acidification extraction using the Technicon AutoAnalyzer. You can get
serious losses of CO2 from the analyzer cups prior to analysis.
Weisberg :

When Skeggs first described the AutoAnalyzer in 1957, that was a fine thing. Then in
1960, he designed the technique for the CO2 combining power, and, you may recall, he put the
entire tray into a box which was equilibrated with CO2 gas. Two years after that, about
1962, Masters had a different device. Some of you may remember the AutoAnalyzer when
they used to do hemoglobins and they added a little stirrer just before it was aspirated, to
bring up the hemoglobin. He had a crook for the CO2 gas going in three times for each cup
and so he didn't have to put the entire tray into a box, but they did a CO2 combining power,
and therefore, representative of bicarbonate. Dr. Cohen and I were just talking here quietly
and he mentioned Dr. Gambino's technique with the al kal inization. Dr. Gambino and I had many
discussions on this. There is no way as far as I'm concerned, by definition, that the
AutoAnalyzer can give you a CO2 content, because it's not anaerobic and you don't have the
entire blood as a true plasma specimen. Whether you do the al kal inization or not, it's not
going to be, by definition, and that's it. Dr. Whitehead in England showed that very defi-
nitely.

Austin :

I think what we're saying is that there is a difference between the CO2 content of whole
blood and plasma. Van Slyke showed this many years ago.

Ladenson :

I think the point is that when most of us measure the CO2 content by the AutoAnalyzer,
we are off by perhaps as much as 3 or 4 milli-equivalents per liter due to CO2 losses to the
atmosphere in these little sample cups.

Weisberg :

You're really getting the bicarbonate.

Gambino :

That's something that Harry Weisberg and I argued about for a long time. I think you're

absolutely wrong. In the al kal inization technique, we drive the PCO2 to about that of the
atmosphere and when the CO2 of the sample is at atmosphere, in a 15 minute period, you don't
get any significant exchange. You don't measure bicarbonate. When you lose CO2, you have
a loss of carbonic acid and bicarbonate and you come down to a PCO2 of about 4 or 5 mmHg.
We've done a lot of work in this area and we do 400 or 500 6/60 's a day and several hundred
blood gases. Many of the patients have the same measurements and physicians are looking at
the calculated bicarbonate and the measured bicarbonate, and if they disagree by more than
2 milli-equivalents, we hear about it right away.

Weisberg :

You just said bicarbonate, not total CO2.

Gambino :

No. We give them a calculated total CO2, which you can get from the pH and PCO2.
Austin :

I agree with Dr. Gambino. There is very little difference between the calculated CO2
content of plasma and the AutoAnalyzer CO2 content.

324
Gambino :

When the sample is handled properly.

Cohen :

May I add my signature to that document. I agree. We have done thousands by the
Gambino technique and they are extremely close to the Van Slyke...

Wei s berg :

You put them into the equation that Dr. Rispens and Dr. Maas were talking about for
research purposes, or were you talking clinical?

Gambino :

This in cl inical

We is berg :

OK. Fine. There's the difference. By definition it is not a CO2 because it is not
completely anaerobic. If we do say that it's a true CO2 content, we have to change the
original definition.

Gambino :

That's a different story.

We is berg :

That's what I'm talking about.

Gambino :

I'm not recommending it as a primary standard method. If you don't alkalinize, you get
disasterous results.

Weisberg :

What do you do with a patient who has a high PCO2 of 100 and you alkalinize? You're
changing it down to room conditions, and you're having to change what was existing in that
patient at that time at his temperature. Under most normal circumstances I would agree with
you. But when you have extreme changes in PCO2, high or low in a patient, you're bringing
him up or you're bringing him down to those room conditions and, therefore, you're not having
the same conditions for your patient.

Rispens :

I hope there is no misunderstanding on our method of measuring total CO2. There is no

escape of CO2 possible before the sample is added to the acid reagent.

Engel :

I should like to return to Dr. Rispens paper in which he mentioned five reasons not to
use base excess or standard bicarbonate. They are not required he says. But what is re-
quired? You can use the same five reasons against actual bicarbonate concentration. If you
have enough information to define the system you can of course calculate or derive the remain
ing variables.

Ri spens :

After measuring pH and bicarbonate you can, of course, calculate base excess with any
nomogram, but what I stress is that you do not need it for assessing disturbances in patients

325
Engel

But we have a very convenient method to determine base excess. Why shouldn't we use it
then.

Ri spens :

Because you do not need it.

Engel :

But you don't need bicarbonate. I don't need bicarbonate if I measure base excess, of
course. Why should I need bicarbonate?

Ri spens :

Actually you need it for calculating base excess. Calculating base excess is in fact
always based on quantities related to bicarbonate through the Henderson-Hassel bal ch equation.
Deriving a new quantity from the same measurement results can not produce any information
which is not already contained in measured values.

Engel :

If you go to the second reason I would like to say, and so what? This is again the
business about the interpretation of results and not measurements of data. An interpretation
always has to be done with regard to the patient and the ongoing disease and the history of
the patient, etc.

Ri spens :

Now that's what I have been saying continuously. My objectives against base excess and
standard bicarbonate is that they are presented as a quantitative measure for the metabolic
component of an acid-base disturbance. Base excess does not always mean excess of base. In
the education of medical students we teach them in the second year how to assess acid-base
disturbances from pH, PCO2 and bicarbonate. We feel that they then clearly understand what
is going on. In the fifth year they go to the clinic and, coming back in the afternoons to
the laboratory, they appear quite confused by all those derived quantities such as standard
bicarbonate and base excess.

Engel :

I can tell you that in our hospital it's exactly the opposite, because we taught them
from the very beginning about base excess and they think they do understand base excess.

Weisberg :

Can I alleviate this heated discussion between Ray and myself in the sense of what we use,
and I'll have a different track when I talk a little bit later. Mark Twain said, "To do good
is noble; to tell others to do good is also noble, but much easier."

Ladenson :

I was very pleased to hear the separation of diagram and nomogram made because I think
that the two have completely separate uses. If I understand correctly, you use the diagram
as a clinical aid and therefore it has to be evaluated on that basis. But I would submit
that the nonogram is very close to a cop-out; because all you're doing with a nomogram is
trying to calculate a parameter that you either cannot measure or choose not to measure. In
many cases, I think this can be dangerous. In tommorrow's session perhaps we can comment on
McLain-Hastings calcium nomogram, but even today Dr. Mai enfant commented on a nomogram for
oxygen saturation which cannot be used at high levels of carboxyhemoglobin or methemoglobin.
In our institution, these are the situations when we often need an oxygen saturation esti-
mate, when we are analyzing the blood of firemen or burn patients. Also we have seen patients
following an industrial chemical exp.losion with severe methemoglobinemia in whom oxygen

326
saturation measurements would have been useful. This limitation may not be common but I
think that what is really needed is a better method which would be accurate under any and all
circumstances rather than a nomogram. I think it really is misleading and very dangerous
to depend on a nomogram.

Meisberg :

Not quite. Jack. Nomograms are fine because they're based upon data derived in the
laboratory. They are not theoretical; they may be based upon a formula based upon facts.
But what you're talking about is that they are not taking into account the methemoglobin and
carboxyhemoglobin. They don't have nomograms that way. And that's where the mistake is,
that you don't realize that you don't have it all as oxyhemoglobin.

Ladenson :

But the problem in that, Harry, is that new compounds have to be continually added to
correct the nomogram. I think if you look at some of the current ones you'll see that you
have to measure six parameters to get at one. Perhaps we should have been spending the time
getting a better method to measure that one.

Weisberg :

Agreed.

Runck :

This question is directed to Dr. Weisberg, Dr. Visser, and Dr. Siggaard-Andersen. Is
it possible for the three of you to agree on an equation for base excess? I've heard three
different versions of the base excess equation being given.

Weisberg :

I'll agree with Dr. Visser, that you need a big computer to take all these things in.
It is impossible to do it otherwise.

Vi sser :

You see that when you change the pH of blood, there is a water shift from the plasma to
the cells. You could call that buffering. Because of the displacement of water from the
plasma to the cells when you increase the PCO2, you decrease the water content, the bicarbonate
concentration is increased and so the resulting difference in plasma pH is smaller than you
expected. So now you have to define the base excess including the water shift. That makes
it extremely complicated. You introduce the Donnan equilibrium, you have to introduce
osmotic pressure, you have to explain why the erythrocytes don't burst, and then we need a
big computer to do a small job. You may have a practical definition of base excess which is
accurate within one unit. That's another thing you can aim at, but if you want to have an
exact definition of base excess, an exact calculation, you're trying to shoot a fly with a
cannon.

Weisberg :

The other way you're trying to shoot an elephant with a popgun.

Siggaard-Andersen :

Which is the best equation or algorithm for base excess? All equations or nomograms are
based on empirical data for titration curves of blood and plasma, or buffer values for hemo-
globin and plasma proteins. All approaches involve certain approximations and it is not pos-
sible to say generally that one algorithm is better than the other. The reference value will
always be the value measured directly by titration.

Cohen :

I would just like to ask Dr. Engel and Dr. Siggaard-Andersen whether or not they would
agree with the assertion that I made, that if one is beginning from the vantage point of

327
wanting to develop a system that adequately described the acid-base status of a living
organism, whether or not they believe that there is any less information required in deve-
loping such an analysis utilizing the concept of base excess or standard bicarbonate than
is necessary utilizing the bicarbonate and PCO2 itself?

Siggaard-Andeirsen :

For the evaluation of the patient we actually do not need any of the derived parameters
(bicarbonate, standard bicarbonate, base excess, buffer base). All we need is the pH and the
PCO2 and a chart (like the one Dr. Weisberg showed). By plotting pH and PCO2 in the chart
we can evaluate the degree of metabolic acidosis or alkalosis as the distance from the
reference line or the reference area for acute respiratory disturbances.

Those who are critical of calculating the excess concentration of base in the extracel-
lular fluid as a measure of a metabolic acid-base disorder, but prefer to estimate the devia-
tion from the reference line in the chart more loosely, are actually deceiving themselves
because estimating the distance from the reference line actually involves estimating the
excess concentration of base in the extracellular fluid.

Cohen :

I'm sure you understand that I didn't' mean to imply that I thought there was anything
less informative about the base excess or standard bicarbonate. What I'm trying to under-
stand is whether or not your position is that there is more information contained in these
parameters, i.e., do you require less physiologic understanding of the behavior of the orga-
nism in order to unravel a complex situation when using these parameters as opposed to the
pH and PCO2 and/or bicarbonate.

Siggaard-Andersen :

It requires the same amount of physiologic understanding to interpret the base excess
value as the interpretation of the bicarbonate. However, there is more relevant information
contained in the base excess value as exemplified by the direct proportionality between the
change in the extracellular base excess and the rise in the lactate concentration during
muscular exercise.

Engel :

Well, I agree with Siggaard. I think that in most cases, maybe 80 percent or more of
the cases, it doesn't matter whether you are using one or the other measure, but for certain
specific cases, for example the exercise case, I also think that you get more information out
of using the base excess than you do using the carbonic acid-bicarbonate system. But we
want to go one step further, and think that if you in the exercise experiment had divided the
base excess, and preferably the base excess measured at a PCO2 of zero, up into the net
acid part and the metabol izable acid part, then you would have had one further piece of
information; namely, that what you were dealing with was an organic acidosis as opposed to a
net acid acidosis.

Cohen :

The difficulty, of course, is that when we are in real life, using various systems for
analyzing acid-base disturbances, we are presented with an unknown. We don't know that
exercise and/or anoxia and/or salicylate intoxication or whatever is at issue or what
combination of those factors is at issue. So it seems to me that the burden of the system
is whether or not it does, in fact, facilitate our understanding of these clinical distur-
bances. And in that context, I still remain to be convinced that I know anything more about
that individual when I know his "standard bicarbonate" or "base excess" than I already have
in my head when I know his bicarbonate and PCO2. I don't see that my understanding has been

improved by those additional parameters. This is an old argument. I know that we've discussed
it many times. But it seems to me that if, in fact, there is nothing more that I've learned
from these additional parameters, then I have to ask myself: Have I given up anything; is
there a price or a risk in using these parameters? I'm not saying that it's not as good.
Certainly, one can develop an internally consistent system that describes nature perfectly

328
well using base excess and standard bicarbonate. There is no quarrel with that. But, the
question is, if I'm not gaining anything, am I losing something? And if there is anything
lost, then it seems to me that the risk-benefit ratio would argue strongly against its
adoption. And I think there is a potent loss or risk. To the extent that the term "base
excess" or "standard bicarbonate" implies to the unsophisticated some information vis-a-vis
diagnosis or therapy, there is a potential for misuse. This is obviously not, as I said in
my presentation, an indictment of the system but is something that has to be considered as a
liability in introducing such terms. I would submit that this is not an issue when one uses
bicarbonate and PCO2 directly for diagnostic purposes.

Weisberg :

Just one comment in reference to Dr. Cohen. Unfortunately, many people, especially in
the United States, think that the M.D. after our name stands for Million Dollars. That is
not so, I'll guarantee that on a Bible. But also, on the other hand, I don't think the M.D.
should stand for Mentally Deficient. And, unfortunately, we have a problem that we can give
all the laboratory data, those of us working in the lab, and if you're on the other side, the
practicing clinician, we can give all the laboratory data to the clinician and, if he's not
going to use it properly, he's going to have a big problem. Let me give you an example. If
we have a patient who has 25 mill i-equivalents per liter of keto-acids, calculated, what
would your guess be for the diagnosis on an acid-base evaluation on that patient?

Cohen :

I don't have enough information.

Mei sberg :

That's exactly the point. Because you find more keto-acids with metabolic alkalosis due
to perniciou^; vomiting then you do with keto-acidosis due to diabetic acidosis. You gave me
the answer and we agreed; you need more information which you get from the bedside. You can
utilize the laboratory data very well whether you use the base excess concept or the PH-PCO2
concept. The average clinician is not like that, unfortunately, and therefore that's the
difficulty that we, who are servicing that type individual with the laboratory data, have.

Co hen :

I'm not so sure from what you've said, what side of the question you've come down on.

Weisberg :

I'm on the side of the question saying that they agree; all you need is a pH and a
PCO2 presented as numbers to distinguish seven diagnoses of acid-base imbalance. You need
a bicarbonate (or total CO2 content) for two additional ones--to distinguish between a
mixed and a respiratory imbalance. The only time I use a base excess (or delta content) is
to distinguish between the acute and chronic respiratory conditions. Even with that, I still
would have to see the patient to make a definitive diagnosis. I'm not hedging and straddling,
but that's a statement of the fact of the case that you cannot make a diagnosis without
knowing the condition of the patient.

Engel :

Well, I think I agree very well with Dr. Weisberg, and maybe one of the major problems
in understanding acid-base is that most people don't know enough physiology.

Layer :

Well, you remind me of a comment made by Joseph Barcroft 50 years ago, namely that the
art of successful advertising consists of saying something which is true but totally irrele-
vant. At the bedside, nothing categorizes the critically ill patient better than the un-
steady state. To make much out of derived values in the unsteady state is a potentially
dangerous proposition. Despite my experience with critically ill patients, acid-base balance,
and gas exchange, I would not know how to use these numbers, and I would be loathe to react
upon them. Although we do not report derived data in our institution, they are available

329
upon request. I like to think that it has not made much difference in the care of the crit-
ically ill patient. What is the value of such interpretation when the derived data, based
on hemoglobin concentration, are received at the bedside after the patient has had an infusion
of two units of packed red cells and the hemoglobin has increased from 10 to 15 g/100 ml? In
exercise, a trained athlete may have a blood lactate concentration of 40 mM at a particular
pH and PCO2, and yet a patient with exactly the same values for lactate, pH, and PCO2 may
be moribund. How do we interpret that? Derivation is important when applied to the steady
state. Its relevance to the critically ill, characterized by the unsteady state, is highly
questionable.

Gambino :

I have just a few comments. One, regarding plastic syringes versus glass and the data
that you showed. We've done some studies in our own lab and find that if the blood gas
comes to the laboratory 50 minutes after it's drawn, in my opinion, it's practically useless,
and what we do are two things. You'll hear tomorrow that we have peripheral laboratories so
that we decrease tremendously this transport problem that blood gases should be measured as
close as possible to if not in the site where they are to be used, to eliminate that problem.
Secondly, we use ice not so much to prevent change but the abundant ice melts and spills and
the person who is given the specimen to carry, sees the melting ice and knows he has to
hurry. The third thing we do is that the person who brings the specimen returns the report,
because, as mentioned, the actual measurement can be done very rapidly. The specimen comes
to a central area and the person who brings the specimen returns the report. That way, the
person who sent the specimen to us knows what's happening because he expects that report to
come back with the messenger and is less likely to be lost because he knows he must come
back.

Regarding the syringes, the dominant reason for measuring PO2 frequently, is to monitor
oxygen therapy so you don't get above 100 mm and below 60 mm, approximately. And most of
the specimens that are critical are in that 60 to 100 range and are coming within 15 minutes.
In my experience, there is no practical problem with plastic syringes versus glass, if the
blood gases are done rapidly. There is a problem with dilution. I am concerned about the
amount of heparin in the syringe and too much heparin will introduce PO2 errors in a sample.
We've had problems with glass syringes when the barrels are loose. There is a need for an
improved collection system that would be impervious to gas and would not have any gas stored
in it and would have a sealed barrel that moves very, very easily. Al kal inization that we
perform for the CO2 content, done routinely, is performed when the Vacutainer arrives, with
the stopper on. It is centrifuged sealed, so there is no change in total CO2 content. As
soon as the stopper is removed, that's the al kal inization point right at that moment before
any transfer takes place, because as soon as you transfer you lose CO2, much as you do in
pouring a Coke or champagne. You're going to lose CO2 when you move the specimen. Finally,
another unpublished study, done at the Presbyterian Hospital in San Francisco by Dr. Burns
and the cardiovascular group there, has shown that the a venous-arterial pH and PCO2 is very
small as long as the cardiac index is above three liters. As soon as you start getting
decreased cardiac output, you increase that spread and, of course, this is the basic problem
with venous samples and capillary samples. In very, very sick patients in shock, the patient
with a serious problem that requires frequent blood gas measurements, those are the patients
that may give you the largest errors. So I think that for the critically ill patient, we
must get that arterial specimen.

Ladenson :

The use of mercury in the syringe is very dangerous as there have been two documented
cases of embolism due to its use.

Unidentified :

I think that this was intended to be after the specimen was taken.

Gambino :

Even then it's not needed. We used it a long time ago; mercury in the syringe, but it
is dangerous to the laboratory personnel as well. But I think it's been shown that you get
adequate mixing by shaking and I think that is the most important. It is essential that the

330
sample presented to the instrument be mixed whole blood and not have any separation. We do
see samples coming into the lab that have been iced and you get extraordinary errors in PO2
if you don't mix that specimen uniformly. So I think that's a critical aspect of presenting
the sample to a blood gas machine.

Ladenson :

What is your median delivery time using your peripheral laboratories?

Gambino :

The peripheral labs are in the intensive care units and there are the four intensive
care units in the medical center. There are four peripheral labs, and they service those
patients in the intensive care unit. So it is as fast as you get the blood and walk to the
machine. I would say the mean time is three, four, or five minutes. Now the mean time to
the central laboratory will vary from ten to twenty minutes, because of the elevator and
walking time, but never longer than 30 minutes.

Weisberg :

Do you have a special runner going up to get the blood?

Gambino :

No, the bloods are drawn at the site, not by the laboratory.

Weisberg :

You mean the house staff. .

Gambino :

Or whoever; the inhalation therapist, the anesthesiologist, the nurse, the nurse prac-
titioner, a variety of people draw the specimen. Because radial artery puncture with a 23
gauge needle and no anesthetic is in many ways easier than drawing venous blood. They do
multiple punctures every day. And in the 6 years I've been at Columbia, there has been
absolutely no serious morbidity, none, from the radial artery. Not true of the femoral or
brachial

S^rensen :

I should like to comment on those plastic containers because we have started looking
at these in the last few months. I feel the problem is not so much the plastic as it is the
plunger that causes the problem. Because in the plastic syringe, the plunger friction is
much larger than in a glass syringe. If you can avoid this by using radial arterial puncture
with small needles and having air bleed in the plunger you get a much better arterial blood
sampl e.

Austin :

I've been an advocate of venous blood gases, but I have seen recently a number of patients
where the A-V PCO2 difference has been quite wide. We followed the instructions religiously,
the instrumentation I think has been OK, and I just cannot explain it when a patient has not
been in shock, but it just seems to be cropping up in about 5 or 10 percent of the patients
where the spread is greater than the 10 mm Hg. I just have no answer for it. It disturbs me
and I've searched continuously and I cannot find it.

Burnett :

How do you do tha- determination? Do you draw venous and arterial specimens simultane-
ously using the same collection techiques?

Austin :

Yes, same collection technique using glass syringes.

331
 :

Gambi no

These are people at rest?

Austin :

People at rest. Patients in hospitals. Warm patients. I'm on one side, someone's on
the other side and we're doing arterial and venous puncture. It doesn't usually change the
diagnosis or the clinical evaluation, but the spread of PCO2 does seem to be wider, and I
just have no explanation for it. I wonder if anybody has seen that sort of thing.

Gambi no :

I have no recent data since we're doing almost exclusive arterial blood work now.

Runck :

Two comments. First of all. Jack Ladenson commented about the O2 decay. If we're
referring to PO2 decay in a blood gas analyzer, this, in fact, is probably due to a leak
in the system. If the blood has been presented as a properly iced sample, whenever PO2 decay
is observed, it can be corrected by really looking at the fit of the electrodes around the
system. Hanson and Neville in Syracuse actually determined this when they put their blood
gas analyzer in a great big plastic bag and filled the bag with 100 percent oxygen and, lo
and behold, the PO2 decay suddenly disappeared. The second comment concerns use of plastic
versus glass syringes. I would agree with Dr. Gambino, you should not discount the use of
plastic syringes. There still are people around using glass syringes who allow the bottom
of the syringe to fall out when the syringe is full. I think we have to recognize that not
everybody is as proficient as we'd like in terms of handling samples; use of plastic syringes
make sample handling easier and more reliable.

Howorth :

I have a comment about Dr. Ladenson's paper. This is in connection with the plastic
cover used on the AutoAnalyzer. I understood that the reason one has to have an orange
plastic cover is because the bilirubin is extremely labile to fluorescent lighting.

Ladenson :

There are two types of plastic covers, one of which is amber and the other clear,

Howorth :

Well, OK. It's essential to stop bilirubin from disappearing due to photolysis prior
to analysis. Then the second comment is about the rate of fall in vitro in blood PO2 in
leukemia. I understood that primitive white blood cells utilize anaerobic metabolism to a
greater extent than normal white cells.

Ladenson :

I have no evidence of my own. The one study that I have found concerning PO2 decay
and white blood cell count did not distinguish as to the cause of the elevated white cell
count.

Gambino :

I think it's important that the plate cover did no good in the AutoAnalyzer. One very
important reason, when you use the little disk that was described for the cup, what's hap-
pening there, of course, is that the disk is effectively reducing the surface area of the
sample exposed. The flow rate or volume of CO2 escaping per unit of time from a particular
surface area is decreased. But when you have the cup, like this, you have a big plastic
thing, there are all kinds of cracks in it. It doesn't mean a thing, and so the surface
area exposed to the atmosphere is the same as it is right out in the open room. This does
absolutely no good at all. That's why we never reported on it. It's absolutely useless.

332
Now as far as the bilirubin is concerned, it is equally useless. If you come to my lab you'll
see that we have a stainless steel cover that's opaque to light because the bilirubin will
deteriorate through that semi-colored cover. So if you want to get the light effect out, we
use a metal cover over the AutoAnalyzer tray, and it is an important factor if you have bright
1 ights or sunshine.

Weisberg :

That kind of cover must be used on the AutoAnalyzer. However, if you allow the cups to
rotate for an hour, you will find that if you have it uncovered, the dehydration of the
proteins will be increased by about 8 percent. So that's an absolute technique for getting
to the one with the acid-base blood gases.

Gambino :

That's not as effective as it should be either, because you still have the leaks.

Weisberg :

Yes, but it's going to reduce it. Without it you're going to have an 8 percent change.

Siggaard-Andersen :

The papers by Dr. Gambino and Dr. Laver were very interesting, but it occurred to me
that neither of them stated directly that another worthwhile effort might be to try to reduce
the number of requests. I think Dr. Laver hinted yesterday that we produce all these numbers
but what do the doctors do with them? There were a number of redundancies among the tests
you mentioned, e.g., serum osmolality and serum sodium.

Gambino :

I think your point is well made, I'm very aware of it, and I believe it can be approached
in several ways. First, the peripheral ization of the laboratory as I conceive of it and as
Dr. Laver was suggesting, when the physician himself or the people taking care of the patients
actually do the measurements, there is direct control. You see, the laboratory is simply
an extension of physical diagnosis. It's a more refined physical diagnosis. Now, the phy-
sician, himself, when he performs a physical exam, knows he has only 24 hours in a day and
then he has so much time to give to the patient, so he rations his time accordingly. When
you have a large central laboratory, and all you have to do is write an order on a chart,
that is conceived of as an infinite resource, much as air and light and water. There's
plenty of air and plenty of water, but we know it's a finite resource, and a laboratory is
a finite resource. So, one of the ways of avoiding abuse of this finite resource is to put
the resource directly into the hands of the person who is using it so that he, himself,
experiences that. And that's what happens with the peripheral unit. In other words, if they
start doing more blood gases, they'll just buy the blood gas machine, set it by the patient,
so that providing laboratory services at the bedside is one way and having the person who is
utilizing the lab tests become more active as part of that, either by actually doing it or by
actually bringing it to the laboratory. In the case of stats in our hospital, a telephone
call is required and a stat voucher is made. We do not pay any attention to written stats.
There must be a phone call by the physician, it must be an action so that there is involve-
ment. I mentioned yesterday about the specimen being brought to the laboratory. When it
comes to the central lab, the person who brings the specimen must wait to bring the report
back. That closes this chain. And so we are very, very concerned about it. The second
thing we do to eliminate tests, and we just last month no longer do single SGOT's as
isolated tests. We don't do HBDH. When we introduce a new test, we always look to see if
we can eliminate another test. So just as the physician only has an hour or a half an
hour--or whatever the time--to do his physical exam, I only have a certain amount of time
and resource and total energy in the laboratory to expend on these results, so I must monitor
what's happening and how useful it is. But I agree that it is easy for this to take off
and be abused. But when the person is doing it himself, it is much less likely. So that the
second thing is that I don't think there is going to be any slackening in the lab work because
the lab is such an essential component of physical diagnosis. You don't want to wait until
the patient is blue to diagnose anoxia. And we don't want to wait until the patient has

333
 . :

cardiac arrest to diagnose digoxin toxicity. Therefore, the physician requires these tech-
niques and, as I told Dr. Laver last night, if you deny a physician laboratory tests, you're
effectively denying him access to his patient. It's as if you locked the door on the room
and you denied access to the patient. I think this is extremely important for the clinical
chemist to appreciate and the people at NBS and the regulatory agencies to realize, that
laboratory testing is such a direct continuance from physical diagnosis that you are inter-
acting very directly with physicians' activities. And the physician is not as dumb as
sometimes has been alluded to--that they don't know what they're doing. They, indeed,
have a pretty good idea of what's happening, and I think that the key thing is for labora-
tories to provide those tests that are physiologically and biochemically correct. When
we understand the physiology and we understand the biochemistry of the disease process, then
the testing becomes important. You will find that as our knowledge of any disease process
increases, testing volume will increase because the test becomes critical. Now that we
understand pulmonary physiology better, the tests are extremely critical. When we don't
understand a disease, such as diabetes, testing is very vague and diffuse and not very
useful

Austin :

I think maybe this has been referred to, but it's my contention that pieces of equipment
need a mother, so to speak, and we have found that when errors develop and when deviations
occur it's because too many people have access to a piece of equipment. They really need to
be, in my opinion, supervised by a very few people and I think when you get too many people
using it, too many handling it, that's where it breaks down. That's my theory.

Laver :

In regard to Dr. Siggaard-Andersen s comments--yes

'
, you are quite right about osmo-
larity. We introduced it a long time ago when we did not have immediate access to a blood
sugar measurement and, of course, the blood urea nitrogen. Now, we receive few requests
for serum osmolarity. However, matters are quite different when one is dealing with blood
gases. These patients are usually in the early phase of acute respiratory failure. It
was our original intent to provide continuous measurements by repeated analyses and a short
turnaround time. Depending on severity of disease and the number of patients in need of
ventilator support, we perform an average of 400 blood gas determinations daily in some 80
patients. Fortunately, the number of patients in need of our services has not increased, and
the average number of samples per patient has not changed. The overall demand is not likely
to be altered until we have reliable, continuous blood gas measurement devices.

Richards :

I direct this to Dr. Gambino. One of the figures I jotted down was a relative standard
deviation of 11.3 prcent for PO2. If you took that at the 2 standard deviation level, it
would mean about 22 percent.

Gambino

Yes. It's not tolerable.

Richa rds :

It seems to me, if this kind of quality figure is being generated, it might be better
for a laboratory to cut back on the work in order to provide a better quality statement.
For those of us who participate in proficiency testing, 22 percent is really very high.
This is the sort of thing you see for some of the enzymes.

Gambino :

I didn't want to give the impression that I found that acceptable. I'm telling you that
this is what was seen when a particular study was performed in different laboratories. In
other words, across laboratories, and of course you have to realize that indicates that there
are problems with instruments in the field.

334
Kreuzer:

I would like to return to Dr. Laver's remark. It is also my contention that the future
is continuous recording, and I would like to draw the attention again to the group in Brus-
sels which is led by Dr. De Meester. This group has embarked on a rather large program to
apply these continuous recordings. Of course, I am well aware of all the problems, but I
think it's worth the effort. My second remark also refers to Dr. Laver's paper. I wonder
whether anybody has had experience with the microporous membranes as applied to measuring
problems e.g., in polarography, because microporous membranes might be just the material
we are looking for. This is due to the fact that we are concerned here with a suspended gas
phase in essense; a large part of the membrane is essentially gas phase. You have high
diffusivity and practically no solubility. Has anybody had experience with these membranes?

Brand :

I tried using such a membrane. You refer I take it, to something like the microporous
Teflon membrane that Millipore makes, that kind of material. In our experience, it clogs
with red cells very quickly. It works very well for about half an hour and then clogs up
irreversibly.

Engel :

I would like to return to the problem of the decentralized laboratories and I agree with
Siggaard that when you get up to a size as we have in our laboratory, with more than 200 em-
ployees, it is comparable to a power station where the requests come in just like when you
turn on a switch. But I also agree with Dr. Gambino that establishment of peripheral labo-
ratories probably will reduce the number of tests which are not really necessary in order to
handle the patient. We have a very good example, when we made a peripheral laboratory in our
pediatric department, the number of tests from that department went down by more than 50 per-
cent.

Vi sser :

I think that the need for blood gas analysis can be enormously decreased by using con-

tinuous CO2 monitoring in the expired air. In Europe, this has been in use for about 20
years. Twenty years ago I had to build my own infrared analyzer because I had no money to
buy one. It is now everywhere in use, hundreds of analyzers, and especially for respirator
treatment of patients with normal lungs, you don't need the blood gas analysis because, in
that case, it is especially the CO2 and the pH that is of importance. And the O2 doesn't
change so rapidly. So this is also decentralized, and you can do it as continuous monitoring
or intermittently, if you don't want to buy so many instruments. Of course, you can't do
that for oxygen. There is a big difference between Europe and the United States, in that
respect.

Laver :

I think that comment requires a vigorous rebuttal, particularly from someone whose
experience has indicated that end tidal PCO2 measurements are worthless in the critically
ill. First, because the difference between end tidal and arterial PCO2 can be substantial.
Second, the blood-gas exchange problem in patients with acute respiratory failure (ARF) is
oxygenation, not CO2 removal. The latter is apparent as a terminal event in ARF and common
with patients in chronic respiratory failure. To my knowledge, there appears to be little
difference of opinion on the subject. Monitoring of end-tidal CO2 may be of value to follow
performance of the ventilator but not for evaluation of blood-gas exchange in the patient
with acute respiratory failure and previously normal lungs. I find no justification to con-
done this return to a questionable practice.

Visser:

That is not what I am proposing. I am proposing a decrease in the number of blood gas

analyses by replacing a number of blood gas analyses by measurement of alveolar or end tidal
CO2, taking into account the difference between end tidal CO2 and arterial CO2. If you have
a patient's arterial and alveolar PCO2, you can follow changes of PCO2 by measuring expired
gas instead of blood. So the number of blood gas analyses is decreased but not replaced by .,

335
Laver:

Unfortunately, things do not work out that way in practice. The number of blood gases
is not reduced because one must still follow the level of arterial oxygenation. Occasionally,
samples must be drawn every 10 to 20 minutes because repeated evaluation of arterial PO2 is
necessary. However, such determinations include the PCO2 and pH. Addition of acid-base
evaluation does not complicate or lengthen the procedure by more than a few seconds.

Kreuzer :

I'm inclined to agree with Dr. Laver, and I might add that the electrodes which we have
described are equally applicable to the gas phase, to the continuous recording of PO2 in the
respiratory gas. We have done hundreds, even thousands of experiments over many hours, with
recordings in all kinds of situations. We have even developed a telemetric method. I might
add here, since I couldn't go into this for lack of time, that we succeeded to telemeter,
over distances, of one to two kilometers, the oxygen uptake of a subject by means of, among
other parameters, the continuous recording of PO2 in the respiratory air.

Si ggaard-Andersen :

Has anybody had experience with the use of the isotonic bridge described by Maas some
years ago? The isotonic bridge eliminates the effect of varying hemoglobin concentration on
the liquid-junction potential but it does not eliminate the effect of varying ionic strength
in the sample. The latter effect can possibly be calculated by means of the Henderson equa-
tion. In dehydrated or over-hydrated patients the ionic strength of the plasma may vary
considerably which might cause a considerable variation in liquid-junction potential. This
would be of special importance with the calcium electrode where the electrode sensitivity is
only half of that for a monovalent ion.

Austin :

I'm not sure I'm answering your question, but I think saturated liquid-junctions give
more stable readings in my experience. I'm not sure that's what you're asking, but that's
my experience.

Bates :

I'd just like to point out, in connection with that figure of Dr. Maas, that strictly
speaking, these standards, which have an ionic strength of 0.16, are not on the NBS scale
exactly because the NBS convention was limited to 0.1 ionic strength. However, I think the
excellent agreement that you find experimently might be justification for relaxing this re-
striction.

Weisberg :

May I change this subject before Dr. Laver leaves and I'd like to have this on the
record. Dr. Kreuzer could answer this too, because back in 1971, Severinghaus and his group
reported an error on the oxygen electrode due to the reduction of the halogenated hydrocarbons,
especially halothane. This has been in the literature and I've asked several anesthesiolo-
gists and it seems to be a puzzle. I've discussed this with Dr. Laver but I'd like him to
put it in his own words for the record and perhaps Dr. Kreuzer could answer too. Does halo-
thane have an effect on the polarographic electrode for oxygen?

Laver :

The answer is no. We have run polarograms with whole blood in equilibrium with up to
4 percent halothane, a variety of electrodes, a variety of membranes and buffers (pH 7 and pH
10). We have found no discrepancy attributable to halothane. This holds true for methoxy-
flurane, fluoxene and other halogenated anesthetics, irrespective of oxygen concentration. The
effect of halogenated anesthetics on the readings of the PO2 electrode need not be of concern
to clinical laboratories.

336
Kreuzer:

First of an, I would like to confirm this, but maybe one word about CO2 should be added.
You keep reading statements in the literature about CO2 sensitivity of the electrode. We
have investigated this very thoroughly, too, and in our experience, if an electrode is markedly
CO2 sensitive, then it is apt to be a bad electrode. If the electrode is constructed properly,
{i.e.:, if there is sufficiently fast exchange of ions between the electrolyte covering the
cathode and the bulk of the solution) and if the bias voltage is chosen correctly it is prac-
tically not sensitive to CO2.

Austin :

Just getting back for a moment to the liquid-junction subject, I would be interested
in Dr. Bates' feelings about the molarity of the solution for a liquid-junction.

Bates ;

Well, I really don't have any special information beyond that presented by Dr. Maas. I

think everyone realizes that the early work indicated that, in systems which are homogeneous,
the stronger the concentration of KCl the more reduction in the liquid-junction potential
that ensues. But I think that with a system such as blood where you have the danger of cre-
nation of blood cells and precipitation of proteins, etc., you may have a special problem.
Perhaps that outweighs the matter of the actual magnitude of the liquid-junction potential
that you would expect in completely homogeneous systems.

Runck :

I have a comment dealing with Dr. Gambino's statement on the performance of the Versatol
Controls and Dr. Noonan's statement on the Coke machine concept. I think Dr. Gambino's data
on the Versatol Controls presented two aspects of performance. The first aspect is the
precision, or repeatability, of replicate measurements. The second aspect wasn't mentioned
and that is the accuracy, i.e., the bias, seen by the laboratories reporting the values.
Our experience has been that the precision problem is actually an operator problem. And that
can be corrected very quickly by going in and retraining people, showing them how to do it
properly. The bias problem is more of a systematic problem, and that has to be looked at in
instrumentation. Now the Coke machine type instruments, the automated instruments, tend to
deal with the precision problem very well. We find very good precision performance with
automated type instruments.

Ladenson :

My question is actually in the form of a request to Dr. Durst and the Bureau of Stan-
dards. I hope that future certified pH standards would be checked with systems that have
liquid-junctions. For example, when we analyzed the certified Tris buffer against the certi-
fied phosphate buffers we got a value that was 0.01 to 0.02 lower than the certified value
{Clin. Chem. y 20^, 1337, 1974). This means that the two certified buffers will give different
results on the type of equipment that is used in virtually all laboratories.

Durst:

Preliminary studies made at NBS several years ago indicated that this a real effect,
but the magnitude of the problem is still under study. Unfortunately, we do not have clini-
cal pH instruments available for our testing and must rely on classical liquid-junction
measurements.

Bates:

Didn't you designate that as a secondary standard?

Durst :

It was supposed to have been issued as a provisional standard until such time as we
demonstrated consistency with the pH scale for cells with liquid-junctions. But at the

337
 : .

point where we certified it, the Office of Standard Reference Materials decided that they
weren't going to issue provisional standards anymore, and it went out as a final certificate,
without any indication that it was to be considered a secondary standard. It was an internal
communications problem in that respect. It went out essentially as a primary standard, when
it really was provisional as far as we were concerned. The certificate has since been revised
to reflect these considerations and also carries the secondary-standard designation.

Kreuzer

I would like to make two remarks with particular reference to the PO2 measurement; one,
the gas-liquid factor and two, the tonometer. As far as the first point goes, what I often
miss in these graphs relating gas PO2 to fluid or blood PO2 is the exact description of the
apparatus. In other words, the blood liquid-gas factor is nothing else than a stirring fac-
tor or flow factor, or a flow dependency as we call it in our continuous system, and it has
only meaning if we know exactly what the electrode is, that is to say, if we know the diam-
eter of the cathode as well as the thickness and the material of the membrane. And this is
not always self-evident. It may be self-evident for the user, but not for the listener.
This factor may be quite different depending on the type of electrode. Radiometer Company,
e.g., has used different types of electrodes over the years with different diameters of the
cathode. Concerning the second point, the tonometer, it has been described already decades
ago that in tonometers there is apt to be a separation between plasma and erythrocytes. If
you draw a sample from the tonometer you may easily have blood that has a higher hematocrit
than native blood and this, of course, as Dr. Runck showed in one of his graphs, will affect
the gas-liquid factor. That is to say, you have to know when you are tonometering what
the situation is in terms of the hematocrit.

Gambino :

I wanted to ask Dr. Maas about his laboratory. You presented data on various instruments
in your laboratory in regard to your aqueous quality control calibration system. Which junc-
tion do you use in routine practice? Do you use your sodium chloride junction or, when you
use the instruments, do you use the junction material recommended by the manufacturer?

Maas :

We have used the junction recommended by the manufacturer.

Gambino :

So that you are not in actual practice, in your hospital, carrying out on patient samples,
the isotonic junction?

Maas :

No, we don't use that type.

Gambino :

Do you think you will? In other words, we woul<l like to know, to help us to decide which
way to go.

Maas :

I feel that in practice by changes of the concentration of the isotonic solutions, for
example, by diffusion of phosphate, poorer results are obtained than using saturated KCl

Gambino :

So, at the present time, you would not favor recommending moving in that direction.
That clarifies it then.

338
Sigqaard-Andersen :

Concerning reference methods, I think it might be possible to develop a reference method

for blood pH on the basis of the operational definition given by lUPAC, using a capillary
electrode, standardized saturated KCl liquid-junction, etc.

Concerning PCO2 and PO2, the reference method would probably have to be based on gasome-
tric analysis of a small gas bubble in equilibrium with the blood, i.e., the Schola ^er tech-
nique. This techinque was too complicated as a routine method but with new techniques """t may
prove useful as a reference method.

Ladenson :

There have been some statements made that this reference material would have to be blood.
If it were blood, how could we standardize its use? I cannot see any way of doing this
because everyone will tonometer it differently. Are we going to have a standard tonometer?
From a practical standpoint, the problem is that all instruments do not give the same results.
To those of us in the field, the problem is which one is correct, and how can we ascertain it.
To the manufacturer, I am sure the problem is similar when it comes to instrument design. The
aqueous material (General Diagnostics) described previously, apparently parallels the dif-
ferences one sees with blood. This material which was apparently shipped all over the
country under all kinds of conditions has very similar results in various laboratories. It
appears like we now have a good start on a reference material and we need a central unit
to provide values for it. I think we are now sitting in the place that can do this better than
anyone and can do it independent of any particular manufacturer's electrodes. I believe the
National Bureau of Standards should play a major role in ascertaining whether there is a
material that is stable and whether values for pH and blood gases can be assigned to it based
on rational measurements independent of a particular commercial instrument.

Durst :

I would like to respond to that briefly insofar as we are being put on the spot. This
was one of the purposes, as far as I was concerned, of the workshop; to try to get ideas of
what could be used as a good standard material for calibrating pH and blood gases. We were
thinking along the lines of one of our buffers in isotonic saline and coming up independently
with a gas mixture of certified CO2 and O2 with the balance in nitrogen. Then one would have
to develop a reference tonometer to put them together and come up with a reference solution.
But it seems from the discussion of the materials that Dr. Gambino described that there
may be a material already available that we could use in this way.

Gambino :

I want to clarify that. I think it's essential that the National Bureau of Standards
provide a reference material for gases, in other words, a certified CO2-O2-N2 mixture as a
reference sample. So I think that's step number one. That can be done. Your mass spec can
help you there. You've got the technology to do that. That will fix a reference point. Now
then I think the material such as I showed, and the material that Dr. Maas has developed in
Holland, are secondary materials and will be dependent upon the primary gas standards. If
you don't go to blood, you can define the PO2 of those systems, and you can also measure,
you can recover, you can do gas chromotography , you can analyze that solution which is a
simple solution, a single aqueous sterile solution, and quantitate the amount of gas. And
from the solubility coefficient, etc., you'll know what PO2 to expect. But I would see it as
kind of a secondary reference material, not something that the NBS could successfully do
immediately, or should do. I think you should definitely do the gas as soon as possible.

Austin :

The same thing should hold true for buffers.

Gambino :

Well, they've done that. Along those lines, I think the NBS should consider doing more
work at 37 °C on other buffers. I think that would be essential for more physiologic research.

339
Me is berg :

What about bracketing the normal pH with something above 7.4?

Gambino :

Exactly, something higher, a buffer certified at 37 degrees that would go above 7.4.

Ri spens :

I think as far as PCO2 is concerned, it is much easier. You do not need gases. You can
make a mixture of a bicarbonate solution and a phosphate buffer which has a definite PCO2
after mixing.

Gambino :

Do you supply gases for any other type of standard?

Durst :

Yes, for example, we have certified air-pollution gases which are mixtures of CO2 and
nitrogen.

Runck :

Blood gas measurements shouldn't be any different than any other clinical chemistry mea-
surement. We need both a standard reference material, as suggested by Dr. Noonan, which is
traceable to NBS, but in addition, we must have a standard reference method for the determi-
nation of blood gases, just as we have atomic absorption for calcium and hexokinase for
gl ucose.

Ladenson :

The only problem I see in stopping there, is that some manufacturers have designed in-
struments that would not be compatible in routine use with a primary gas standard, e.g., the
Radiometer ABLl system. For this reason, I believe that we need something that will be
applicable to all instruments. I wonder, and I address this to the manufacturers, would
such a non-blood material be acceptable to you as a starting point to try to get these values
together.

Runck :

Absolutely.

Ladenson :

Then I think that is what has to be done.

Austin :

Everything is relevant to that.

Ladenson :

Well, we can all agree that with such an NBS certified material we can begin to standard-
ize the results obtained with different instruments.

Gambino :

It would be possible with an NBS gas in a defined tonometer system to reproduce ...

340
Ladenson:

What I'm saying is that the material has to be made or purchased in large single lots
in the same way as all the other Standard Reference Materials. NBS writes the specifications,
analyzes the material, and provides a certified value for a material which you can ship
around the world.

Unidentified :

I don't know how other people feel about this, but I found using a wide variety of dif-
ferent types of tonometers, made in different ways, that it's not that difficult to tono-
eter a solution accurately.

Runck :

That depends, in our experience, on the level to which you tonometer your blood samples.
Most of the data we have seen are at normal values for PCO2 and PO2. Those don't stress the
instrument. It would probably be advisable to have, in addition to normal ranges for PCO2
and PO2, a control material or a reference material at the extreme values as well.

Gambino :

I would disagree with you, at least clinically, maybe in physiologic research you might
want it, but clinically we don't want patients to go above 110 millimeters at all.

We is berg :

What if you have a patient on therapy?

Gambino :

That is why we do so many oxygen tension measurements and nobody is supposed to go over
100 millimeters, period. That's why we do it. You keep lowering the inspired oxygen and
that's why, in my particular laboratory, I couldn't care about accuracy at very high P02's.

Weisberg :

You get levels of 300 and 400 in cardiac bypass.

Gambino :

So what? It doesn't make any difference.

Weisberg :

The point being that they want those with accuracy, and they're trying to make a
decision on it.

Gambino :

No they don't. They don't want it at all.

Runck :

I think there's controversy on that. There are two points on the table right now.
Number one, can you tonometer accurately? Number two, if you can, what are the correct
levels to use? Each has to be treated separately.

Gambino :

There may be some special physiologic situations in a hyperbaric situation where some-
body's trying to maintain a high PO2. In a hypothermic situation, they want a high amount
of oxygen physically dissolved. Those are special cases. I think when you see the problems

341
that are occurring, however, the critical decision levels are at 100 and at 60. Those are
the dominant decision levels. It's essential that we be accurate and precise at those i

levels. think you get into tremendous sample handling problems with a high PO2. At high
'

PO2, you have to go right from your tonometer or your patient directly into the instrument. 1

If you have any type of sampling handling, you're not going to get the right value. !

Siggaard-Andersen :

Just a comment on the technical problems of tonometry. If you have very small bubbles,
the pressure inside the bubbles is higher than atmospheric pressure. If you have a very
tiny outlet from the tonometer, you may also have a higher pressure. It is not always easy
to obtain exactly the same temperature in the tonometer as in the electrode, so if we want ii

accuracies of 0.1 mmHg it is not going to be easy.

'

Runck :
I

My comment relates to Dr. Gambino's comment on the levels of PCO2 and PO2 at which the
operator should test the instrument. I think we must keep in mind that there are two purposes
in testing an instrument. First, is to assure yourself that patient treatment is being
carried out on good numbers. Second, is to give the operator early warning of impending
instrument problems. In the case of CO2 measurements, if we choose 35 millimeters as our !

checkpoint for CO2, the carbon dioxide electrode could actually turn to stone and you would
still get very nice readings at 35 millimeters because the electrode is always kept in the
area of 35 millimeters. It is very important that a CO2 level other than 35 millimeters be
tested. And the same thing with PO2. The early warning of a PO2 problem witli an instrument
is going to be very apparent at a high PO2. If you have excessive atmospheric contamination
or a leak in the sample chamber, giving relatively small problems at 86 or 90 millimeters, you
will see this in exaggerated fashion at 200 or 300 millimeters. This is why manufacturers
should state performance characteristics at these levels as well.

Gambino :

I agree with that. You also see it if it is very low.

Cali :

Would it be interesting to the group to find out how we would handle the question of
reference materials here within the Bureau, in order to arrive at a reference material?

Group : I

Yes. I

Cali :

I've heard you talking about accuracy and precision. Where it is possible, we believe |

that all standardization should be based upon accuracy. I won't go into the complete logic
\

of that proposition but if you are trying to achieve compatabil ity in measurement and by ;

that I mean simply that everyone measuring the same sample gets the same result, then that 1

goal can be achieved directly if all concerned are making measurements that are accurate. So, |

we start off by asking what is the basis for achieving accuracy? Well, accuracy is related j

to the true value, but let's not get into a philosophical fight on what we mean by "true ,

value." But you can say in an operational sense, that somehow you've got to have access to
|

the base and derived units of measurement, which is now the SI system. We have access to
the SI by what we call here at NBS definitive methods. Now, definitive methods do not
exist, in which case as an alternate, you then use two or more reliable methods. What |

constitutes a reliable method is a matter of expert subjective and, hopefully, objective j

opinion. The preferred way, however, is via the definitive method. A definitive method [

must fulfill several criteria. You must have a well-established theory and usually it 'j

should be expressable in mathematical terms, usually via appropriate equations. When you [

make such a definitive measurement, using the equation, the derived end result is expressed
j

in terms of one or more measurement parameters. Call them A, B, C, and so on. You must |

then make your measurements in such a way that each of these parameters. A, let's say is I

342
mass, has to be directly traceable, in a well defined series of steps, to the base system in
units. And similarly, if B is temperature, you have to know what systematic errors are in-
volved in the making of that measurement, and so on. When you are finally done, you can then
say that your measurement is within some error bounds in terms of accuracy, or inaccuracy,
to be more correct. Using such a proven definitive method, you are now in a position to
take a material and to build into that material the accurate values of whatever property or
properties are under consideration. From this we arrive at what we call here at NBS a
Standard Reference Material. Now the rest of the world does not like this word "standard",
in the phrase "Standard Reference Material". It's a word very difficult to translate into
other languages. The accepted term now would be reference materials, in general, and what
we make at NBS would be called certified vefevenae materials. So we've got certified
reference materials (CRM's) and, more generally, reference materials. At this point, because
CRM's are directly related to the SI by a one-step process, we call these primary reference
materials. Now, this methodology is usually very difficult to apply in the typical clinical
laboratory situation. For example, we may use such techniques as isotope-dilution mass
spectrometry which involves instrumentation running well over $150,000 or $200,000 plus very
highly trained technicians and scientists and so on. Therefore, we want to transfer this
accuracy in some way that is of a more practical nature. So now, we go from definitive
methods via the CRM to reference methods, which is what you've been discussing. Given then
a CRM and a reference method we now have the possibility of the manufacturing community
producing accurately known secondary materials, in a matrix that is really of interest.
Such secondary materials can be used for two purposes: (1) quality control on a daily basis
through these secondary materials, and (2) through the use the reference method and the
primary reference material to check in turn the accuracy of what I call the field or routine
methods used on a daily basis. This whole process has, of course, variations, depending
upon the particular scientific or technological field. In clinical chemistry, I have just
described the logical way to proceed toward achieving measurement compatibility. This
system is, however, just in its initial stages. Many more definitive methods, reference
methods, and CRM's need to be developed before major advances can come about.

Have I heard correctly that, although you may not all be in exact agreement, what you
would like from NBS would be a series of gases at the appropriate partial pressures of
oxygen and carbon dioxide in nitrogen as a first CRM. As a second CRM, perhaps a buffer
that's equilibrated with appropriate partial pressures of those two gases, and then prefer-
ably, blood or blood-like matrices prepared the same way.

Engel :

Right.

Cal i :

Have I got it about right?

Several voices :

Yes.

Cali:

You'll have it tomorrow!

Ladenson :

I'd like to ask Dr. Bates a question. You said you used Tris buffer to control pH. Did
you do any studies with Simon's calcium exchanger on the effect of Tris buffer? Both the
Orion and the Ruzicka exchangers will show apparent calcium binding to Tris buffer.

Bates :

We used Simon calcium exchanger, and our studies suggested that there is no significant
association between calcium and Tris buffer.

343
 ^

Ladenson

relationship.

Bates:
^i.^ the linpar relationship between the
concentration of calc,
;tn'poss?b,r"tSrr4.f:n^iraSor?h,t Lresponsible for the slope being ,

than theoretical
1li

No 0 nan:

z.,^

^mmmm
J- '"'-i-nreJ
anything but ;%Sinr^s^:rorrbe\oi:?rt°b:t^:^^e
since then, 'E.f I th nk some or zu^ V
extensive studies i,

activity and concentration is a question

correlate them to some normal ratio between

s?r?:srdiusi "nnLS^:d=^r"bui levels.

i^i--','-:-i-,,s^oi:s%"hrd
these ^
chloride measurement. Looking at them with
'

levels bv a total ^^^^^^f

normality. And likewise variations in tota pr^^^^^^^^
'deviatfonsVrom average
since they would directly affect the
to make electrolytes look differently,
activity coefficient.

Cohen :

How do you interpret that? fyoi

Gambino:
I think it's important to consi
Yes what could a physician do with that data?
measurements, but don't think enough work has been don^Je
activity and the electrode I
In (

ies
Noonan :
Jxpe

So that you could s

When you look at the total, the correlation is excellent.
'ita

on your flame photometer, but once

we're qoinq to give you 140 on electrode and 140
i it;

going to give you 130, or its going to give you 150. The d the
while that electrode is -
the usefulness of this approach.
studies have just never been done to demonstrate
!

ions relating to the thernj

does seem to me that the physiological activity of these
basically a function of activity, and not q
of ion movement across cells, etc., are
tion.

Weisberg :

Seligson did that about 10 or 12 years ago with Dahms and Rock, and they had ihi;,
publication in Cliniaal Chemistry but it's just hanging in limbo because you can'tli'jto
your number to concentration. As Ray pointed out, the clinician is thinking in coi
for electrolytes, and therefore, it may be a very good method, but you have now go
factor to correlate it to the concentration.
i|5 re,

Cohen : i "itei

.
) Did

The issue is whether or not these are really telling us something that is goti/hyi

the patient. trei

344
 I

is old data suggesting that chloride is measureably bound to albumin.

hat is true. One of the things that was missing from my model, of course, was
actions with albumin. I specifically left out chloride because I don't believe
ignificant binding at physiological pH. If you look back, Dahms used chloride
explain some anomalous chloride results that he got. If you look back to the
rk of Scatchard, and it's been confirmed on a couple of occasions, then it shows
lloride binding of albumin is markedly pH dependent, and at physiological pH, it's
igible. So I think that there is little justification for including chloride-
ding. In terms of CO2, of course, there's plasma carbamates which ought to have
d in the model that I presented, but I just didn't because the data wasn't readily
me.

Ik about chloride binding in the physiologic range, what magnitude are you

s far as I can remember, the thing that's reported is number of ions of chloride
|nd per molecule of albumin, and the number in the physiological range is less
here is less than one chloride ion bound per molecule of albumin. When you get
of around three (or perhaps five, I don't remember exactly) it goes up to about
ions per albumin molecule.

ersen:

the report by Dr. Brand very interesting but I wonder how you have estimated the
the protein polyelectrolytes on the ionic strength? Just taking the net protein
tration and assuming the influence equivalent to monovalent ions is questionable,
lated the ionic strength of normal plasma on the basis of the pK value for carbonic
fna as compared to pure aqueous solutions and found a value of about 0.17 mol/kg.
lyou have any comments to that.

nsi

e ionic strengths that

I show are excluding protein. They are electrolyte ionic
other words, I am following the convention of Scatchard, that the activities
In
tes in protein solutions are unaffected by the proteins. Now I don't think that's
experimentally by anyone other than Scatchard, so I can't comment, but that's a
sr^ntative source. The model also, I should perhaps have commented, does not force
fility. It does not require that the system is in electroneutrality and it simply
the excess charge, whichever way, is made up by the protein. I think a model
?1oped forcing electroneutrality. It would be an interesting thing to do but I
lerni it, and it wasn't included in the model presented.
)t c

tuggested that, as a final item, we have summary made of what's happened over
'days, and I didn't want to do this because I am very subjective about the con-
i[disqualified myself. Since I wanted to get a disinterested person, I asked
n't g to provide us with a summary.
cof
I
go'

e really disinterested, I wouldn't have come. But there's an old Army saying,
nited States and I am sure in the army of every country, "never volunteer." I

Dick a little earlier asking if a summary would be given and you've heard his
goi why he isn't giving it. I had suggested some names and some of those named
t reverted back to me again, you see; so here I am.

I
345
 I apologize because if I had known I was going to do this, I would have listened mt
better to you people. If I leave anyone's name out, please accept my apologies, and I cl'

ask you please to contribute until the time of your leaving for the airport, if I have n|;

any errors or if you object to what I say. I will honestly try to be objective and lea\/'
out my personal bias. I have to tell you what Dr. Laver said yesterday when I was kiddij'
him, having known him for some years, why he didn't contribute as much as I did by inter
ting, and he said he learned his lesson. He was at a conference several months ago and i
spoke so much that he now has 15 pages to proof on his added comments; so I am stuck aga;
on that. >

Ithink that we can have some subheadings for Dr. Rubin and the idea for the Expert ;are
Panel. Number one would be the groups of the Definition of Quantities and Concepts to iK
combined with the Recommendation of Nomenclature and Physiological Terminology and SymboipJ
As mentioned earlier, there is a phoenix that has arisen from 1964, and obviously we stij
have the same problems we had at that time. We have the problem of overlapping laborato
and clinical terminology. We have the problem of which parameters to use in reference tif
the diagnosis of acid-base imbalance. Here, of course, I'm mentioning names. Dr. Cohen
introduced some additional terms regarding the duration and compensation of acid-base '

disturbances, and Dr. Engel introduced additional parameters, especially regarding the I'^^y^
titratable acidity. idiis

lefa

Ithink the entire problem is more complicated in comparison to 1964, especially by! :liet

the introduction of the SI units. The United States is lagging far behind the other I
lion,

countries, even though the other countries are still not quite uniform; for instance in off i

England, they are going to use hemoglobin as grams per deciliter because of the confusio'
that would result with the use of millimoles. We have been lagging for the following
reasons: number one, we're not in the metric system and, in reality, in order to get a ifis ai

metric system and SI units, we should be doing this with our school children rather than' fi

talking ex cathedra from this level to go down. We are really going to have our problem' acy
in the United States. In addition to this, we have a problem that we can accept it in Vf^^
laboratory; I calculated that there are nine practicing clinical chemists in the audienc
We can do it, but the fact is that the physician is not going to do it we

i'tie r

In addition to the problem of SIunits, we have the problem of in vivo and in vitro
diagnosis. The question of extra-cellular fluid regarding pH, especially, and blood gas
Several years ago when there wasn't much distinction, they had a controversy. I think m^"^''
and more today we're going over to the in vivo type of consideration and the clinicians ft''
ato
going to have to learn to adjust.
»urs

Our third area is the question of temperature correction. To quote two good areas- ved
Dr. Laver is not here now and we can say that his is a good area, Massachusetts General, ed
the Mayo Clinic. They are routinely doing temperature correction for pH and blood gases^'^ys
all routine determinations as well as for the cardiac by-pass type of surgery. I think F'ist
gave my stand yesterday on the plea that it should be done. The question of us today isr^e
what advantages would there be? We do not question as to what the normal would be at thi"'
patient's body temperature. We get the situation, such as Dr. Gambino has written aboutj
his monthly newsletter, of the patient coming in who has a PO2 of over 150 breathing roof
air; the physician says it's impossible, and when you check back, of course, that patienf
a low temperature. And finally, another area of controversy is the question of pH versujj
hydrogen ion concentration; and I won't make any comments about that.

Possible areas to be considered in greater depth by the subcommittees in the future

perhaps would be P50, or as our European friends call it T50, oxygen content, and 2,3-DPi
But there has been some agreement. Number one, we agree whether pH or hydrogen ion wouli
be the area of determination for acidosis/alkalosis or acidemia/al kalemia. The PCO2, or
if you wish, the carbonic acid concentration or total CO2 content by calculation would b^Jssi '

reported for respiratory factors. We still have our century of the Tower of Babel in the
we still lack agreement on which metabolic term and which metabolic test to be used. 1

Whether the values should be reported per se, to use calculated values or those based on'M
the determined ones, should be looked into to see what can be done. As a subgroup of th 01
subcommittee A, is the question of the Evaluation of Nomograms and Algorithms. Here it
was agreed, at least I think it was agreed, that there is a distinction between a diagrai tan
which would be a graphic representation for education and possible diagnostic help to the'|cti '

physician and the nomogram which is a means of calculating additional parameters for thef

346
 — •'^ J personnel in
yratory •• order
— — -w report
to — i^w to
vvy the physician.
^1 ^ Now it is not necessary to use
'I
a •
I
• . • >, V.I Ijr I V

liagram or a nomogram in the better centers, but I think it does have some area area" of us(
use
ivijhe boondocks, if you would, where they don't have ve quite the expertise as this group
today.

A subcommittee B would be entitled, with our subsections today, Blood Sampling,

and Storage.
jjling, I think this is important enough to be a subsection of the proposed
rt Panel and it may need a restatement of the state-of-the-art rather than any further
^tigation. But you heard the comments and the- difficulties; as Dr. Ladenson showed,
b are indefinite or insufficient data but there have to be more definitive areas of
we're doing.
1
I think here again, with my own personal bias, that for an evaluation
I2, you must have arterial blood, for the routine acid-base evaluation where you don't

til a PO2, no matter how small that percentage, that venous blood would be perfectly
factory. This has been substantiated time and time again. Dr. Gambino, of course,
tdhe of the leaders in that area.

A subcommittee C would combine, I think, the Establishment of Reference Values,

ity Control and Standards, and the Development of Reference Methods. And, as is
Dus from this afternoon, this was the area of greatest interest, contributed no doubt
le fact that Mr. Cali gave his beautiful exposition of what the NBS does and so on. I
here again, the problems came up with tonometry. Should we use blood, an aqueous
ion, or a glycerol solution as advocated by Dr. Maas? I think this requires a great
of expert input, and we would have to have that subcommittee do that. The question
Vpe of tonometer, and I apologize to Dan ahead of time, the simple Noonan tonometer,
simple Noonan, but the simple Noonan-tonometer, versus the more complex ones. I think
s agreed that the NBS should work in the areas of certifying gas mixtures for standardi-
n for tonometry and for standardization of the instruments with a probable ultimate
euifacy within one percent. It was also agreed, I believe, that the NBS should work in
t|area of certifying buffers or buffer compounds for buffers of pH. I think the sugges-
'DCfl
that was made is extremely important that these buffers be certified at 37 °C and
we bracket the 7.40. It is crazy to have a pH in the low sixes and under 6.8, and so
ve really should be bracketing and my own preference would be somewhere about 7, 7.4
7.6 or a 7.7 in that area. It was agreed again to use an operational definition for
jjslAs far as PCO2 and PO2 is concerned, there was some discussion for mass spectrometry,
iner this is a definitive method or is this really a secondary method for accuracy. A
IS
,inetric Scholander-Riley technique would be the one to be utilized in the reference
i)'

atory. The problem came up of the question of normal levels for PCO2 and PO2 and,
)urse, that evolved itself with a give and take, but we won't mention the people
ved. I believe, that an abnormal level, as it seemed to be the consensus, should
;ed primarily to check out instruments rather than for the beneficial knowledge to
)hysician for diagnostic levels. Dr. Gambino reported on some data using General
lostic's new aqueous acid-base control which contains PO2 values, and this would
lately replace the one that they have now which does not have a PO2 and is a lyophilized
'ial, which I think has a great many problems attacked to it. If the liquid material
out well, as I think it will, it should be substantiated and supported.
;!
My own
iistion to Dr. Durst is that I think there have been enough people working with it,
iambi no, Mr. Komjathy, and myself, that this should be included in the final mono-
1, because you don't expect a company to publish somebody's in-house, out-of-the-house
of work but with the monograph I think it will give some good data to show on that
of support.
jre

-DPI
The next area of section D is the Instrument Specification and here, of course, we
'the best available and then beyond that. I think the consensus that everybody would
5 to is that we need better service manuals for maintenance and repair at all hours,
or

i\i
ksible. But what we're really saying is we need a VIP instrument; VIP instrument
tliJ
hs for Very Idiot Proof. I think the manufacturers should be listening, as they were
especially to that Section C with the Reference Methods and Standards, because even
0(1
ph they are in the business to make money, we also have to have some of the accuracies
tl
on. The future problem, as emphasized by Prof. Kreuzer, is the question of the in
it
monitoring. Here, of course, with a PO2 electrode that he described, micro and so
gramli
an be extremely convenient. It was extremely interesting to see this new PCO2
tlie
ictivity electrode because this is now opening some new parameters whereby we may make
advances.

347
 And finally, we have the Electrolytes. And this I find a little bit good and bad
because it's sort of an afterthought. We had 2 days on the other groups and had only 2
hours for electrolytes. This is like the New Orleans type of talk of lagniappe, something
extra for nothing. It came out very well because this is leading into my conclusion. And
by taking it out of context, Dr. Brand gave a very nice talk and his comment that reporting
the values as per liter of plasma water is extremely important. I think this has been
suggested many times in the past. I was very pleased to hear Dr. Bates talk about sea
water. Number one, because he is the other bald-headed, white-haired man in the audience,
and number two, the sea water study that he is working with now takes me back to 1916 when
the diagram (number 3 in my list) by McClendon also worked on sea water. So really we've
gone full cycle now for another 60 years.

I wish to thank Dr. Durst for the honor of summing up the deliberations and, on
behalf of all the participants, to congratulate him on being the honorary obstetrician for
the birth and/or delivery of this conference. I hope that the monograph will be as success-
ful as the one on Ion-Selective Electrodes.

348
 )

I!BS-1 14A (REV. 7-73)

U.S. DEPT. OF COMM. 1 . PU B L IC ATION OR R E PORT N 0 . 2. Gov't Accession 3. Recipient's Accession No.
1
BIBLIOGRAPHIC DATA No.
1
SHEET NBS-SP 45 0
i. TITLE AND SUBTITLE 5. Publication Date
Blood pH, Gases and Electrolytes
June 1977
j

j
Proceedings of the Workshop on pH and Blood Gases held 6. Performing Organization Code

I
at the National Bureau of Standards 7/7-8/75

. AUTHOR(S) 8. Performing Organ. Report No.

1 Richard A. Durst, Editor
PERFORMING ORGANIZATION NAME AND ADDRESS 10. Project/Task/Worlc Unit No.

NATIONAL BUREAU OF STANDARDS

DEPARTMENT OF COMMERCE 11. Contract/Grant No.
WASHINGTON, D.C. 20234

2. Sponsoring Organization Naipe and ComDlete Address (Street^ City, State, ZIP) 13. Type of Report & Period
American Association for Clinical Chemistry- Covered
American Society of Clinical Pathologists
Final
International Federation of Clinical Chemistry
14. Sponsoring Agency Code
National Committee for Clinical Laboratoiry Standards
TT,S. Department of Commerce, National Bureau of Standards
5. bOFPLEMENTARY NOTES

Library of Congress Catalog Card Number: 76-608179

6. ABSTRACT (A 200-word or less [actual summary o( most significant information. If document includes a significant
bibliography or literature survey, mention it here.)

On July 7-8, 1975, a workshop was held at the National Bureau of Standards to
discuss the status and needs of this very important area of clinical measurement,
'

A major goal of this workshop was the initiation of cooperative efforts on an inter-
national level toward the standardization of pH and blood gas measurements and the
various quantities and terms used in this field.

To this end, the first technical session was concerned with the acid-base status
of blood and included the topics: Definitions of Quantities and Concepts; Recommen-
dations of Nomenclature, Physiological Terminology and Symbols; Reference Values;
and the Evaluation of Nomograms and Algorithms. The second session addressed
itself to the more practical aspects of this subject and included the topics: Blood
Sampling, Handling, and Storage; Instrument Specifications; Quality Control and
Standards; and the Development of Reference Methods. Finally, a brief session on
the newer topic of the electrometric measurement of blood electrolytes.

This volume contains all of the papers invited for presentation at the workshop by
some of the leading clinical and medical authorities on this subject and also includes
a transcription of the extensive discussion sessions.

17. KEY WORDS (six to twelve entries; alphabetical order; capitalize only the first letter of the first key word unless a proper
name; separated by semicolons
Acid-base status; blood electrolytes; blood gases; blood pH; calcium; carbon dioxide;
hydrogen ion concentration; nomograms; oxygen; ^ qq pH; Pq potassium; sodium. > ;

18. AVAILABILITY [X^ Unlimited

19. SECURITY CLASS 21. NO. OF PAGES
(THIS REPORT)

1
'
For Official Distribution. Do Not Release to NTIS 338
UNCL ASSIFIED
ly 1
Order From Sup. of Doc, U.S. Government Printing Office 20. SECURITY CLASS 22. Price
$4.75
Washington. D.C. 20402. SD Cat. No. C13 , 10:45 0 (THIS PAGE)

1 1
Order From National Technical Information Service (NTIS)
Springfield, Virginia 22151 UNCLASSIFIED
USCOMM.DC 29042-P74
t}.U.S. GOVERNMENT PRINTING OFFICE: 1977-2 40-848/1 94
 //
/ y ' ' monthly
*

X y^^m^^ f
/ magazine of the Nation-
There's ^^^^ X
^^^^hi^"//^^ a' Bureau of Standards.

anew
featured are special ar- Still

^^^^^^^pV/ tides of general interest on

y current topics such as consum-
_ \>»^
^^^^"^ / ^/
C"" product safety and building

technology. In addition^ new sec-

look tions are designed to . . . PROVIDE

SCIENTISTS with illustrated discussions
of recent technical developments and

to... work in progress . . . INFORM INDUSTRIAL

MANAGERS of technology transfer activities in
Federal and private labs. . . DESCRIBE TO MAN-
UFACTURERS advances in the field of voluntary and
mandatory standards. The new DIMENSIONS/NBS also
carries complete listings of upcoming conferences to be
held at NBS and reports on all the latest NBS publications,
with information on how to order. Finally, each issue carries
a page of News Briefs, aimed at keeping scientist and consum-
alike up to date on major developments at the Nation's physi-
cal sciences and measurement laboratory.

(please detach here)

SUBSCRIPTIO^l ORDER FORM

Enter my Subscription To DIMENSIONS/NBS at $12.50. Add $3.15 for foreign mailing. No additional Remittance Enclosed
postage is required for mailing within the United States or its possessions. Domestic remittances (Make checks payable
should be made either by postal money order, express money order, or check. Foreign remittances to Superintendent of
should be made either by international money order, draft on an American bank, or by UNESCO Documents)
coupons.

Send Subscription to:

D Charge to my Deposit
Account No.
NAME-FIRST, LAST

1 I I I I I I I I I I I I I M I I I I I I M I I I I I

COMPANY NAME OR ADDITIONAL ADDRESS LINE

MAIL ORDER FORM TO:

STREET ADDRESS Superintendent of Documents
I I I I I I I I I I I I I I I I I I I I I I I I I I I I
Government Printing Office
Washington, D.C. 20402
CODE

MM
STATE ZIP

I I I I I I I I I I I I I I I
1

PLEASE PRINT
 NBS TECHNICAL PUBLICATIONS
PERIODICALS NOTE: At present the principal publication outlet for
these data is the Journal of Physical and Chemical
JOURNAL OF RESEARCH reports National Bureau Reference Data (JPCRD) published quarterly for NBS
of Standards research and development in physics, by the American Chemical Society (ACS) and the Amer-
mathematics, and chemistry. It is published in two ican Institute of Physics (AIP). Subscriptions, reprints,
sections, available separately: and supplements available from ACS, 1155 Sixteenth
• Physics and Chemistry (Section A) <^ St. N.W., Wash. D. C. 20056.
Papers of interest primarily to scier* ^v*^ orking in —
these fields. This sectipn covers a br ^
.ige of physi-
cal and chemical research, wit> ^^'^ ^r emphasis on
Building Science Series Disseminates technical infor-
mation developed at the Bureau on building materials,
components, systems, and whole structures. The series
standards of physical measu'- ^ fundamental con-
,
presents research results, test methods, and perform-
stants, and properties of m' ^<v*'.dsued six times a year. ance criteria related to the structural and environmental
Annual subscription: D- ^^-^o"*.:, $17.00; Foreign, $21.25. functions and the durability and safety characteristics
• Mathematical Sci'- ^,{^'^ ^Section B) of building elements and systems.
Studies and com*- ^* .is designed mainly for the math-
ematician anH Topics in mathemat-
.Q<v*<;tical physicist.
—
Technical Notes Studies or reports which are complete
in themselves but restrictive in their treatment of a
ical statis^'^^j;!?^.ieory of experiment design, numerical
subject. Analogous to monographs but not so compre-
analysi'" jretical physics and chemistry, logical de- hensive in scope or definitive in treatment of the sub-
sign ^
programming of computers and computer sys-
Q^jnort numerical tables. Issued quarterly. Annual
ject area. Often serve as a vehicle for final reports of
t'"
work performed at NBS under the sponsorship of other
siL Domestic, $9.00; Foreign, $11.25.
.jcription: government agencies.
DIMENSIONS/NBS (formerly Technical News Bulle-
Voluntary Product Standards— Developed under proce-
tin) —
This monthly magazine is published to inform
dures published by the Department of Commerce in Part
scientists, engineers, businessmen, industry, teachers,
10, Title 15, of the Code of Federal Regulations. The
students, and consumers of the latest advances in
purpose of the standards is to establish nationally rec-
science and technology, with primary emphasis on the
ognized requirements for products, and to provide all
work at NBS. The magazine highlights and reviews concerned interests with a basis for common under-
such issues as energy research, fire protection, building standing of the characteristics of the products. NBS
technology, metric conversion, pollution abatement,
administers this program as a supplement to the activi-
health and safety, and consumer product performance.
ties of the private sector standardizing organizations.
it reports the results of Bureau programs
In addition,
inmeasurement standards and techniques, properties of Consumer Information Series — Practical information,
matter and materials, engineering standards and serv-
based on NBS research and experience, covering areas
of interest to the consumer. Easily understandable lang-
ices, instrumentation, and automatic data processing.
uage and illustrations provide useful background knowl-
Annual subscription: Domestic, $12.50; Foreign, $15.65. edge for shopping in today's technological marketplace.
NONPERIODICALS Order above NBS publications from: Superintendent
of Documents, Government Printing Office, Washington,
Monographs —
Major contributions to the technical liter- B.C. 20J,02.
ature on various subjects related to the Bureau's scien-
tific and technical activities.
—
Order following NBS publications NBSIR's and FIPS
from the Natio)ial Technical Information Services,
Handbooks — Recommended codes of engineering and Springfield, Va. 22161.
industrial practice (including safety codes) developed
Federal Information Processing Standards Publications
in cooperation with interested industries, professional
organizations, and regulatory bodies.
—
(FIPS PUBS) Publications in this series collectively
constitute the Federal Information Processing Stand-
Special Publications — Include proceedings of conferences ards Register. Register serves as the official source of
sponsored by NBS, NBS annual reports, and other information in the Federal Government regarding stand-
special publications appropriate to this grouping such
ards issued by NBS pursuant to the Federal Property
as wall charts, pocket cards, and bibliographies.
and Administrative Services Act of 1949 as amended,
Applied Mathematics Series —
Mathematical tables, man- Public Law 89-306 (79 Stat. 1127), and as implemented
uals, and studies of special interest to physicists, engi- by Executive Order 11717 (38 FR 12315, dated May 11,
neers, chemists, biologists, mathematicians, com- 1973) and Part 6 of Title 15 CFR (Code of Federal
puter programmers, and others engaged in scientific Regulations).
and technical work. NBS —
Interagency Reports (NBSIR) A special series of
National Standard Reference Data Series Provides — interim or final reports on work performed by NBS for
quantitative data on the physical and chemical proper- outside sponsors (both government and non-govern-
ties of materials, compiled from the world's literature ment). In general, initial distribution is handled by the
and critically evaluated. Developed under a world-wide sponsor; public distribution is by the National Techni-
program coordinated by NBS. Program under authority cal Information Services (Springfield, Va. 22161) in
of National Standard Data Act (Public Law 90-396). paper copy or microfiche form.

BIBLIOGRAPHIC SUBSCRIPTION SERVICES

The following current-awareness and literature-survey Superconducting Devices and Materials. A literature
bibliographies are issued periodically by the Bureau: survey issued quarterly. Annual subscription: $30.00'.
Cryogenic Data Center Current Awareness Service. A
Send subscription orders and remittances for the pre-
literature survey issued biweekly. Annual subscrip-
tion: Domestic, $25.00 Foreign, $30.00i. ceding bibliographic services to National Bureau of
;

Liquified Natural Gas. A

literature survey issued quar- Standards, Cryogenic Data Center (275.02) Boulder,
terly. Annual subscription: $20.00. Colorado 80302.
U.S. DEPARTMENT OF COMMERCE
National Buraau of Standards
Washington, D.C. 20834
POSTAGE AND FEES PAID
U.S. DEPARTMENT OF COMMERCE
OFFICIAL BUSINESS
COM-215
Penalty for Private Use. S300
SPECIAL FOURTH-CLASS RATE
BOOK

\
6 6 29
I