Seminars in Anesthesia, Perioperative Medicine and Pain (2005) 24, 9-16

Acid-base balance revisited: Stewart and strong ions

David A. Story, BMedSci (Hons), MBBS (Hons), MD, FANZCA,a
John A. Kellum, MD, FACP, FCCMb

a
From the Anaesthesia Research, Department of Surgery, Austin Health, The University of Melbourne, Heidelberg,
Victoria, Australia; and the
b
Department of Critical Care Medicine, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania, USA.

KEYWORDS:
Acidosis; Understanding acid-base physiology is a core requirement for anesthesiologists and intensivists and
Alkalosis; other physicians responsible for the care of critically ill and injured. Despite a long history, clinical
pH; acid-base physiology remains a confusing area for many clinicians and numerous misconceptions exist.
Strong ion difference; The application of principles of physical chemistry to clinical acid-base physiology allows for a
Strong ion gap; valuable new perspective on an old problem and may provide important insights.
Lactate; © 2005 Elsevier Inc. All rights reserved.
Lactic acidosis

“If the facts don’t fit the theory, change the facts.” The use of various laboratory variables to diagnose an
—Albert Einstein acid-base disorder is analogous to the use of the electrocar-
diogram (ECG) to diagnose a myocardial infarction. How-
The concentration of hydrogen ions (H⫹) in blood
ever, neither the changes in the ECG tracing nor the distur-
plasma and various other body solutions is among the most
bances in electrical conduction that theses changes reflect
tightly regulated variables in human physiology. In health,
were ever considered to be the cause of a myocardial in-
the free hydrogen ion concentration in arterial blood rarely
farction. In contrast, changes in bicarbonate (HCO3⫺) con-
deviates from 35-45 nmol/L—pH 7.45 to 7.35. Acute
centration, for example, have been assumed to be responsi-
changes in blood pH induce powerful regulatory effects at
ble for metabolic acidosis or alkalosis. Failure to establish
the level of the cell, organ, and organism. Yet the mecha-
causation has lead to numerous incorrect notions of acid-
nisms responsible for local, regional and systemic acid-base
base physiology and has fueled years of, often heated,
balance are incompletely understood and controversy exists
debate.1,2 Thus, it is our intent to “revisit” acid-base balance
in the literature as to what methods should be used to
in light of recent advances in the understanding of physiol-
understand these mechanisms.1 Much of this controversy
ogy and clinical epidemiology.
exists only because the strict rules for causation (as opposed
to association) have not often been applied to the under-
standing of acid-base balance and methods that are useful Acid-base: A brief history
clinically have often been used to understand physiology
without being subjected to appropriate scientific rigor. In During the 1950s there was a paradigm shift in the approach
other words “the facts” about acid-base are still in dispute. acid-base physiology and pathology.3 Bicarbonate was pro-
moted from being an important factor in acid-base balance4
to being the central factor for control of the non-respiratory
Address reprint requests and correspondence: John A. Kellum, MD,
University of Pittsburgh Medical Center, Division of Critical Care Medi- (metabolic) component of acid-base balance.5-7 One reason
cine, 200 Lothrop Street, Pittsburgh, PA 15213-2582. for this shift was a desire among clinical chemists to define
E-mail address: kellumja@ccm.upmc.edu acids along the lines of the Bronsted-Lowry definition that

0277-0326/$ -see front matter © 2005 Elsevier Inc. All rights reserved.
doi:10.1053/j.sane.2004.11.008
10 Seminars in Anesthesia, Perioperative Medicine and Pain, Vol 24, No 1, March 2005
had become popular with physical chemists: an acid is a
hydrogen donor.5,8,9 Using this new approach carbonic acid,
and its conjugate base bicarbonate, assumed a central role.9
A key feature of this bicarbonate-centered approach is to
use the Henderson-Hasselbach equation (Equation 1) to
describe acid-base status of the plasma. The Henderson-
Hasselbach equation substitutes the concentration of car-
bonic acid with the partial pressure of carbon dioxide and its
solubility coefficient.
[HCO3⫺]
pH ⫽ pKa ⫹ log10 (1)
␣pCO2
Where pH is the plasma pH, pKa is the pH at which the acid
(carbonic acid) is 50% dissociated, [HCO3⫺] is the plasma
bicarbonate concentration; ␣ is solubility of carbon dioxide
in blood at 37°C, and pCO2 is the partial pressure of carbon
dioxide in blood.
The bicarbonate centered approach remains popu-
lar.8,10,11 Many who use this approach, however, use bicar-
bonate “rules of thumb” to determine primary and mixed
disorders without quantifying these disorders.8,10,12 One ad-
vantage of using base-excess is that it quantifies non-respi-
ratory changes in acid-base status.13 Following on from this,
the quantitative effects of acidifying factors such as lactate
are poorly defined. Further, phenomena such as alkalosis
associated with decreased plasma albumin concentration,
and acidosis secondary to hyperphosphatemia are poorly
explained using the bicarbonate centered approach.14
In the late 1970s and early 1980s, Peter Stewart, a Ca-
nadian-born physiologist working at Brown University,
Rhode Island began to revisit acid-base balance.15-18 Stew-
art was concerned about many issues in acid-base chemistry
and felt that the bicarbonate centered approach to clinical
acid-base chemistry was “. . .piecemeal, qualitative and
confusing. . ..”17 Stewart’s emphasis was on a quantitative
approach, as highlighted in the title of his most widely cited
paper “Modern quantitative acid-base chemistry.”16 Stewart
examined the different roles of strong ions (ions that are
completely dissociated in solutions), and weak acids that are
only partly dissociated (Equation 2):
HA H⫹ ⫹ A⫺ . (2)
Here HA is the undissociated weak acid, H⫹ is the hydrogen
ions, and A⫺ is the dissociated anion.
Stewart developed an acid-base framework that can be
applied to any physiological solution.18 However, in com-
mon with clinical acid-base work, he focused on plasma. In
plasma, Stewart’s three principal independent factors are:
(1) the difference between the concentrations of strong
cations and anions (particularly sodium and chloride); (2)
the concentration of weak-acids (particularly albumin); and
(3) the partial pressure of carbon dioxide.16,17 Bicarbonate,
while still a clinical marker, is but a minor electrolyte,
dependent on the three independent variables.17 A fourth
component is the dissociation constants of weak-acids and
water. The hydrogen ion concentration of plasma (or any
physiological solution) will increase if there is an increase
in the partial pressure of carbon dioxide, and increase in the
concentration of weak acids, or a decrease in the strong-
ion-difference.16
Water: A unique substance
While water is the solvent for physiological solutions, water
also dissociates to produce hydrogen (H⫹) ions and hy-
droxyl (OH⫺) ions.19 Under the Stewart approach, water is
the principal source of hydrogen ions in physiological so-
lutions.16
As temperature increases, the dissociation of water in-
creases, which in turn increases the H⫹ concentration. The
dissociation of water can be expressed in terms of the law of
mass action:
[H⫹] ⫻ [OH⫺]
Kw ⫽ (3)
[H2O]
Where Kw is the dissociation constant of water, [H⫹] the
concentration of hydrogen ions, [OH⫺] the concentration of
hydroxyl ions, and [H2O] the concentration of water. The
concentration of “water” in water is 55 mol/l. The Stewart
approach assumes that because the concentration of water is
thousands of times the concentrations of H⫹ and OH⫺, it
can be used as a constant. Further, because the dissociation
of water is temperature dependent the dissociation constant
used in calculations should be temperature corrected. A
modified dissociation constant for water (K’w) is calcu-
lated.16
K
w ⫽ [H⫹] ⫻ [OH⫺] (4)
This equation simplifies the understanding of Stewart’s ap-
proach—that increased water dissociation increases the hy-
drogen ion concentration. Further, Equation 4 simplifies the
underlying mathematics.16 Importantly, while the hydrogen
ion concentration is important for physiological mecha-
nisms8 from a quantitative perspective, hydrogen ions are
unimportant because compared with the quantitatively sig-
nificant ions in solutions (such as plasma), that are in mil-
limolar concentrations, hydrogen ions are in nanomolar
concentrations.16
Strong and weak acids
One question is how to determine when an acid is weak or
strong. Stewart16 noted that in general, lactic acid is re-
garded as a weak acid by chemists, compared to the many
strong acids available in the laboratory. Stewart also noted
that whether an acid is weak or strong is relative to the
surrounding acid-base environment. At pH 7.40 lactic acid
is more than 99.9% dissociated and can be thought of as a
strong acid; that is, all the lactic acid is assumed to be in the
lactate anion form. The operational pH range for a weak
Story and Kellum Acid-base Balance Revisited
acid is a pH environment that is within one pH unit of the
pKa for the weak acid.19 This is one approach to divide
weak acids from strong acids. At pH 7.40 an acid with a
pKa of 6.4 will be 90% dissociated. A pH range of two pH
units from the pKa would mean that at pH 7.4 the acid
would more than 99% in the anionic form. For physiological
applications an acid could be described as weak acid if it has
a pKa greater than 5.4 at 37oC, or a strong acid if the pKa
is less than 5.4. With this approach, acetate and gluconate
can be defined as strong anions.
A base can be defined as a substance that when added to
a solution decreases the hydrogen ion concentration. Am-
monia is a base and has a pKb (association constant of the
base) of 9.00 at 37°C when it forms ammonium (NH4⫹).
Below pH 7.0, ammonium can be regarded as a strong
cation. The major functional acid-base role for ammonium
is to provide a cation, other than sodium or potassium, to
maintain electroneutrality with urine chloride anions. This
role for ammonia allows chloride excretion in urine without
also losing sodium cations.12,17 Therefore, in the presence
of acidosis, increased ammonia excretion will allow an
increase in the plasma strong-ion-difference which will
counter the acidosis. On the other hand, THAM (tris(hy-
droxymethyl)aminoethane) is also a base but because
THAM has a pKb of 7.80 it acts in plasma as a weak base.20
Therefore as the concentration of THAM increases the hy-
drogen ion concentration will decrease but the change will
also depend on the pKb of THAM, which is also tempera-
ture-dependent.21
Anecdotal evidence suggests that one problem for
greater acceptance of the Stewart approach by clinicians is
the lack of obvious connection between a decrease in the
strong-ion-difference (SID) and an increase in the hydrogen
ion concentration. In the first step of his mathematical ex-
position of his approach,16,17 Stewart used the example of a
solution of sodium chloride in water. His first equation
included the dissociation constant of water that slightly
clouded the relation ship between hydrogen ion concentra-
tion and SID. This first equation can be simplified by ex-
cluding the dissociation constant for water. If we consider a
solution of sodium chloride in water, the strong ions will be
sodium cations and chloride anions. The SID, in this case
will equal the sodium minus chloride. The remaining elec-
trolytes will be hydrogen cations (H⫹) and hydroxyl anions
(OH⫺) from the dissociation of water. Given the require-
ment for no overall charge (the principle of electroneutral-
ity), the sum of the difference between hydrogen and hy-
droxyl ions and the difference between the strong ions must
equal zero.9,16 That is:
0 ⫽ H⫹ ⫺ OH⫺ ⫹ SID. (5)
Therefore, if we solve for hydrogen ions:
⫹ ⫺
H ⫽ OH ⫺ SID. (6)
As the SID becomes smaller the hydrogen ion concentration
will increase. It is important to note that by the time we
11
reach a mathematical summary for plasma considering all
the components of the three independent variables Stewart
used as a quadratic equation with the three independent
variables and four constants.16
While Stewart’s approach seems a departure from con-
ventional thinking, Stewart’s work in fact returns to ele-
ments of earlier work, particularly that of Van Slyke and
co-workers.4 In the 1930s, Peters and Van Slyke felt that
bicarbonate was a marker of acid-base status and com-
mented “. . .the most important (facts) concern the roles
played by the individual anions other than bicarbonate, and
by the cations, in determining the acid-base balance and the
bicarbonate content of the blood.” Stewart also used a def-
inition of acids that is more useful in the physiological
setting than the Bronsted-Lowry definition: an acid is a
substance that when added to a solution increases the hy-
drogen ion concentration.17 This definition is similar to the
generalized form of the definition that Arrhenius developed
in the 1880s.3,22 One advantage of Stewart’s definition is
that carbon dioxide can be defined as an acid directly rather
than having to consider carbonic acid. Stewart also used
Van Slyke’s approach4,9 of considering chloride as an acid
and sodium as a base. Therefore, under the Stewart ap-
proach, the important alkalinizing component in sodium
bicarbonate therapy is not the bicarbonate but the sodium
(being administered without a strong anion).20,23,24
Albumin and phosphate
Stewart did not go far in delineating which were the impor-
tant weak acids in plasma but as early as 1908 Henderson25
recognized a role for phosphate and speculated on the role
of plasma proteins. By the 1930s Van Slyke4 had confirmed
the role of plasma proteins. Over the last 15 years research-
ers have tried to identify which plasma proteins are impor-
tant and to quantify the anionic role of the proteins and
phosphate. The major group in this area since Van Slyke has
been Figge’s group.26,27
Figge et al. showed that the most important protein weak
acid is albumin.26,27 They analyzed the 16 imidazole groups
on albumin molecules and derived dissociation constants for
each. They reached their conclusion that albumin is the
dominant source of weak acid by comparing plasma results
with a model derived from albumin only solutions and the
16 dissociation constants. These data allowed them to de-
velop equations to calculate the anionic effect of albumin.
Albumin anions (meq ⁄ l) ⫽ [albumin]g ⁄ l
⫻ (0.123 ⫻ pH ⫺ 0.631) (7)
They also developed an equation for the anionic effect of
phosphate.
Phosphate anions (med ⁄ l) ⫽ [phosphate] mmol ⁄ l
⫻ (0.309 ⫻ pH ⫺ 0.469) (8)
12 Seminars in Anesthesia, Perioperative Medicine and Pain, Vol 24, No 1, March 2005

For clinical situations, a simple formula can be used.

[A⫺] ⫽ 0.25X[albumin]g ⁄ l
⫹ 1.3X[phosphate]mmol ⁄ l23,24 (9)
where [A⫺] is the weak acid concentration in plasma.
Although there have not been extensive surveys of the
acid-base status of different patient groups, several studies
have found that many critically ill patients have a decreased
plasma albumin concentration. McAuliffe and colleagues28
studied eight patients with a normal SID and a normal
anion-gap (corrected for albumin). These patients all had a
metabolic alkalosis as defined by base-excess and a low
plasma albumin concentration. Using the Stewart approach
they concluded that the reduced plasma albumin, causing
reduced total-weak-acid concentration, had caused the alka-
losis. The finding that a decreased [A⫺] results in alkalosis
has been supported by others.14,29 The association of alka-
losis and decreased albumin can be easily explained by the
Stewart approach, a reduced total-weak-acid concentra-
tion.30 However, controversy remains about the interpreta- Figure 1 Gamblegram of human plasma. This diagram shows
tion.30,31 For example, we have argued that a reduced the equal amounts of cations and anions (electroneutrality) in
strong-ion-difference in the face of a reduced [A⫺], for plasma at equilibrium in a healthy human. SID, strong-ion-differ-
example, secondary to hypoalbuminema, is not an acid-base ence; A⫺, weak acid anions; other, other strong anions. The
disorder if the pH is normal (and no respiratory compo- Gamblegram shows the dominant influence of sodium and chloride
nent).23 The base-excess will also remain zero in this ex- on the strong-ion-difference. Further, it shows that if all strong
ample. This view is supported by preliminary observations ions are known the strong-ion-difference is equal to the sum of the
in an analbuminemic strain of rats (Nagase), which at rest bicarbonate anions and the anions from weak acids, notably albu-
min and phosphate.
have a reduced weak acid concentration and a reduced
strong-ion-difference with a pH of 7.4 and a pCO2 of 40 mm
Hg.32
albumin using the assumption that albumin has both a weak-
The findings of Wilkes14 in critically ill patients also
acid component and a strong anion component. While using
appear to support this assertion. Wilkes demonstrated that
different methodology, the authors derived similar results to
the “set point” for the SID to achieve a normal pH given a
Figge.27 This study complements Constable’s remodeling of
normal partial pressure of carbon dioxide changes with
Stewart’s mathematics that may provide an important tool
changes in albumin. Furthermore, although the loss of
in physiological studies.37
weak-acid from the plasma space is an alkalinizing process,
Some of the earlier criticism of the Stewart approach
there is no evidence that the body regulates albumin to
came from those who developed base-excess as an acid-
maintain acid-base balance and there is no evidence that
base variable.31 Wooten,36 however, has demonstrated that
clinicians should treat hypoalbuminemia as an acid-base
the Stewart approach can be analyzed mathematically in a
disorder.
similar way to base-excess. In applying the Stewart ap-
The overall quantitative physical chemistry of a physio-
proach to clinical work, base-excess remains an important
logical solution such as plasma can be summarized with the
clinical marker. As shown previously by Kellum and col-
column diagrams (Figure 1) called Gamblegrams after the
leagues,38 the change in the SID is virtually identical to the
American pediatrician James Gamble who popularized
change in the standard base-excess.
these diagrams.33,34 The Gamblegram for plasma highlights
the point that while sodium and chloride are the dominant
variables in the SID, hydrogen ions are quantitatively un-
important.
Clinical application
Stewart’s approach to acid-base disorders has growing pop-
New views on stewart ularity among clinicians, particularly in anesthesia and crit-
ical care.12,23,24,30 Recently, using Stewart’s work, there
Several papers have used complex mathematical analyses to have been advances in quantifying non-respiratory acid-
examine detailed physiology.35,36 In plasma, albumin is the base changes, particularly in quantifying the role of unmea-
principal weak-acid with a smaller effect from phosphate.38 sured ions. Until recently the anion-gap has been the prin-
Staempfli and Constable35 studied the acid-base effects of cipal approach to determining the presence of unmeasured
Story and Kellum Acid-base Balance Revisited
ions, including lactate.12,39 The Stewart approach highlights
the effect of albumin on the plasma bicarbonate concentra-
tion.16 As albumin falls, the bicarbonate will increase as a
direct effect for the same partial pressure of carbon dioxide.
Decreased albumin is common among critically ill pa-
tients14,24,30 Figge and colleagues demonstrated that 0.25⫻
(the change in albumin, g/l) can be used to correct the
anion-gap in mmol/l.40 A more precise estimate of the
unmeasured anions can be obtained by comparing the ap-
parent SID (SIDa) calculated from the electrolytes:
SIDa ⫽ [Na⫹] ⫹ [K⫹] ⫹ [Mg⫹⫹] ⫹ [Ca⫹⫹]
⫺ [Cl⫺] ⫺ [lactate anions] (10)
with the SID calculated from the opposing effects of CO2,
albumin (from Equation 7) and phosphate (from Equation
8). This later “version” of SID has been termed the effective
SID (SIDe).
SIDe ⫽ 2.46 ⫻ 10⫺8 X pCO2 ⁄ 10⫺pH
⫹ [albumin]g ⁄ l ⫻ (0.213 ⫻ pH ⫺ 0.631)
⫹ [phosphate]mmol ⁄ l ⫻ (0.309 ⫻ pH ⫺ 0.469)
(11)
When SIDa ⬎ SIDe unmeasured anions must be present.
This difference or “gap” has been termed the strong ion gap
(SIG) to distinguish it from the anion gap.
SIG ⫽ SIDa ⫺ SIDe (12)
36
The SIG is normally very low (⬍2 mEeq) but may be
elevated in critically ill patients from ketones and renal
failure (see below) and perhaps through other mechanisms
not yet understood.23
Intravenous fluids
Crystalloid solutions are a mainstay of perioperative and
critical care medicine. Discussion of fluids for intravenous
use illuminates a number of the issues in the methodology
of the Stewart approach16 to acid-base physiology and
pathophysiology. Water can be used clinically in the treat-
ment of hypernatremia in critically ill patients. What is not
obvious is that the administration of water may aggravate
metabolic acidosis. Post infusion acidosis is usually associ-
ated with administration of normal saline.41
It will probably come as a surprise to many that the
infusion of sterile water or dextrose will be acidifying. As
water is administered to a patient, the plasma concentrations
of sodium and chloride decrease together. However, the
decrease in sodium concentration is faster than the decrease
in chloride concentration. Thus, the difference between
plasma sodium and chloride will decrease resulting in a
smaller SID and hence, acidosis. The physiological acid-
base effect of an infusion of 5% Dextrose will be an acidosis
due to the effective free water infusion associated with
13
dextrose use. The acidosis will be accompanied by various
effects on glucose metabolism under different physiological
and pathophysiological conditions. Acidosis from mannitol
solution is also due to a similar decrease in SID.
Normal saline (0.9%) has a sodium chloride concentra-
tion of 154 mmol/l in some countries, including the United
States, and 150 mmol/l in other countries, including Aus-
tralia. Infusion of 0.9% saline, particularly in large volumes,
produces marked acidosis in both animal models and peri-
operative patients.41-49 The mechanism for this acidosis,
like water, is a reduction in the plasma SID. However, with
saline, the change in sodium and chloride concentrations is
upward with the rate of increase in chloride concentration
being greater than the rate of increase in sodium concentra-
tion. Previously, this acidosis has been referred to as dilu-
tional.41 The Stewart approach provides for a more logical
explanation. The infusion of a solution with a SID less than
that of plasma will be acidifying, while one with a SID
greater than plasma will be alkalinizing. Note that unless the
solution also contains weak acids or bases (eg, albumin), it
is the SID of the solution, not its pH that will determine the
effect it has on the recipients’ plasma pH.50
In an influential review from the 1970s, Garella and
colleagues51 concluded that the acidifying effect of isotonic
saline was minimal. This conclusion was drawn from ani-
mal work. Prough and Bidani52 concluded that the contrast
between the acidemia seen in the clinical study of Schein-
graber53 and minimal acidifying effects of saline infusion in
previous animal work51 is difficult to explain. They argue,
however, that interspecies variation partly explains the dif-
ference in findings. Williams and colleagues46 provide fur-
ther support for the concept that moderate isotonic saline
infusion is associated with detectable acidosis. Their study
demonstrated a mean pH fall of 0.04 pH units following 50
ml/kg infusion of 0.9% saline in healthy volunteers. The
rate and extent of acidification following saline infusion will
depend on the rate and amount of saline administered, renal
handling of sodium and chloride, and transmembrane move-
ment of strong ions, and associated blood loss. This finding
has been supported by others48 who went as far as to call
0.9% saline “(ab)normal saline.”
Intravenous fluids with anions other than
chloride
Several randomized clinical studies47,49,53 have shown that
perioperative use of solutions containing anions other than
chloride results in less acidosis than 0.9% saline. Commer-
cially available fluids include: Hartmann’s, Ringer’s Lac-
tate, and Plasmalyte (Table 1). Scheingraber’s group49 ex-
amined acid-base changes in patients undergoing
gynecologic surgery. Patients received 30ml/kg/hr of 0.9%
saline or Ringers Lactate. After 120 minutes of infusion the
Ringer’s Lactate group had a mean pH of 7.41 and the saline
group had a mean pH of 7.28. The Stewart analysis of this
14 Seminars in Anesthesia, Perioperative Medicine and Pain, Vol 24, No 1, March 2005
Table 1 Electrolyte content (mmol/l) in crystalloid solutions for intravenous use
0.9% saline Hartmann’s
Osmolality 308 274
Sodium 154 129
Chloride 154 109
Potassium 0 5
Calcium 0 2
Magnesium 0 0
Other anions 0 29 (lactate)
phenomenon has several features. Because the anions used
instead of chloride are completely dissociated at pH 7.40,
the anions, lactate, acetate, and gluconate (Table 1), can be
treated as strong ions.16 Intracellular uptake of these anions
is rapid (minutes) and usually complete. The uptake of these
anions results in a larger extracellular SID and less acidosis.
Further alkalosis follows increases in extracellular SID as-
sociated with metabolism of the anions. Previously this
alkalosis has been explained by bicarbonate production.54 It
can now be explained by release of strong cations, particu-
larly sodium, without a strong anion. The sodium release
will increase the extracellular SID. In Scheingraber’s
study,49 both groups had an initial SID of about 38 meq/l.
The SID in the Ringer’s Lactate group fell to 33 meq/l and
in the saline group to 28 meq/l. A further effect of all
crystalloid solutions is to reduce the concentration of
plasma albumin49,50 which will reduce the plasma weak
acid concentration which will be mildly alkalinizing—thus
explaining the pH of 7.41 in this study despite the decrease
in SID to 33.
The clinical importance of decreased post-infusion aci-
dosis by electrolyte solutions containing anions like lactate,
acetate, and gluconate is unclear. Using a rat model, Healey
and colleagues55 found that 50% survived after massive
saline infusion and 100% survived after comparable Ring-
er’s Lactate (Table 1) resuscitation. Traverso and co-work-
ers44 used a swine model to examine mortality following
life threatening hemorrhage and subsequent resuscitation
using different crystalloid solutions. Following hemorrhage
over 15 minutes, all groups had metabolic acidosis that was
initially aggravated by crystalloid infusion. Some of the
clinically important differences in acid-base change and
survival were not statistically significant; however, multiple
comparisons may have limited the power of the study. The
highest 24-hour survival was in the Ringer’s Lactate Group
(67%) followed by 0.9% saline (50%). Plasmalyte had the
lowest survival of 30%. Magnesium and acetate ions are the
only constituents present in Plasmalyte but not Ringer’s
Lactate. The authors speculate that these electrolytes may
explain a worse survival for the Plasmalyte group.
Some groups56,57 have used the fluid prime of cardiopul-
monary bypass circuits to examine the acid-base effects of
the different components of fluids used for intravenous
resuscitation and maintenance. There were two main find-
Plasmalyte 148 Ringer’s lactate
294 274
140 130
98 109
5 4
0 3
1.5 0
27 (acetate) 28 (lactate)
23 (gluconate)
ings. First, they found that succinalyated gelatin in colloid
solutions was acidifying and increased the concentration of
unmeasured anions. Second, they found that exogenous
lactate has an alkalinizing effect because it is removed from
the plasma while the accompanying sodium is retained thus
increasing the SID.
Morgan and co-workers studied the effects of multiple
fluids with varying the strong-ion-difference on the pH of
whole blood ex vivo.50 Their findings were that given the
simultaneous dilution of weak acids and addition of strong
ions, a fluid with a SID of 25 meq/l resulted in no change in
standard base-excess when added to whole blood. Using an
animal model of septic shock, Kellum58 showed that resus-
citation with a solution containing heta-starch and an effec-
tive SID of 28 meq/L resulted in less acidosis and longer
survival time compared to 0.9% saline despite resuscitation
to the same arterial pressure.
Renal failure in the ICU
Two recent papers, from the same research group, have
examined the clinical chemistry of patients with acute renal
failure of critical illness requiring renal-replacement therapy
in the ICU.59,60 The first study compared the patients with
acute renal failure to control groups of critically patients
without acute renal failure.59 Patients with renal failure had
a mean base-excess of ⫺7.5 mmol/l largely due to increased
concentrations of unmeasured anions and phosphate. This
pattern was different from the controls. The patients with
acute renal failure then received continuous veno-venous
hemofiltration.60 The researchers found that hemofiltration
corrected metabolic acidosis by decreasing the strong-ion-
gap including the phosphate weak acid effect. Further the
SID was increased principally through a decrease in chlo-
ride. These decreases coupled with the persisting alkaliniz-
ing effect of hypoalbuminemia, lead to a metabolic alkalosis
developing after 72 hours of treatment.
Predicting outcome
In critically ill children, Balsubramanyan and colleagues61
estimated the SIG from the standard base-excess and found
Story and Kellum Acid-base Balance Revisited
that standard base-excess ⬍⫺5 mmol/l, when due to un-
measured anions was an important predictor of mortality.
This finding was not supported by a recent study of criti-
cally ill adults that found that the neither SIG nor the
base-excess effect from unmeasured ions were strong pre-
dictors of mortality but were good predictors of a blood
lactate ⬎5 mmol/l.59 This study59 also found that lactate
itself was only a moderate predictor of mortality. However,
others found lactate to be an important predictor of mortal-
ity in children62 and adults.63 Therapy in the intensive care
unit may obscure the predictive value of the SIG, particu-
larly when fluids containing unmeasured anions are used,
such as gelatins.64 Kaplan and Kellum have recently shown
that in victims of major trauma, prior to any therapy, acid-
base values and especially the SIG, are powerful predictors
of hospital mortality. Indeed the SIG outperformed injury
severity score, serum lactate, and pH.65
Measurement error
One challenge to all clinical chemistry is the reliability of
the results obtained from the laboratory, whether a central
laboratory or point-of-care. Morimatsu and colleagues66 ex-
amined the agreement in measuring biochemical variables
at two sites in one hospital. One site was the central hospital
laboratory the other was an adult intensive care unit point-
of-care machine. The measured biochemical variables in-
cluded sodium and chloride, important for calculating both
the strong-ion-difference and the anion-gap. For these two
related calculations the cumulative difference was consid-
erable. For the anion-gap the agreement was poor: 3 ⫾ 6
mmol/l and similar poor agreement for the strong-ion-dif-
ference. The authors suggest that clinicians need to know
the relative relationships of biochemical variables not only
between hospitals but also within hospitals. This study com-
plements a previous study that found similar poor agree-
ment between two different estimates of plasma bicarbonate
in critically ill patients; the agreement was 1.6 ⫾ 4 mmol/
l.67 Bicarbonate estimates are used in their own right to
determine the base-excess,68 the anion gap,40 and the strong
ion gap.69 Further work is needed to clarify the most ap-
propriate assays for estimating clinical chemistry variables,
particularly in the critically ill.
Conclusions
Stewart’s work continues to grow in popularity. The phys-
iological underpinning of the approach is becoming increas-
ingly sophisticated35 while the clinical application is be-
coming more straightforward and the clinical utility more
obvious.24,30 Clinical application of approaches such as the
modified anion-gap40 and the SIG69 are being assisted by
the availability of lab coat sized personal computers. Chal-
lenges include relating the acid-base variables to outcome
15
and optimizing the reliability of the clinical chemistry. Im-
portantly, the concept of independent and dependent vari-
ables in the analysis of acid-base and the importance of
electrolytes in determining blood pH are now, more than
ever, understood to be fundamental to clinical acid-base
chemistry. For this alone, we owe a substantial debt to the
late Peter Stewart.
References
1. Severinghaus JW: Siggaard-Andersen and the “Great Trans-Atlantic
Acid-Base Debate.” Scand J Clin Lab Invest Suppl 214:99-104, 1993
2. Worthley L: Strong ion difference: a new paradigm or new clothes for
the acid-base emperor. Crit Care Resusc 1:211-214, 1999
3. Story DA: Bench-to-bedside review: a brief history of clinical acid-
base. Crit Care 8:253-258, 2004
4. Peters J, Van Slyke D: Quantitative Clinical Chemistry. Baltimore,
MD, Williams and Wilkins, 1931
5. Christensen: Anions versus cations. Am J Med 27:163-165, 1957
6. Frazer S, Stewart C: Acidosis and alkalosis: a modern view. J Clin
Path 12:195-206, 1959
7. Relman A: What are “acids” and “bases?” Am J Med 17:435-437,
1954
8. Abelow B: Understanding Acid-Base. Baltimore, MD, Williams and
Wilkins, 1998
9. Siggard-Anderson O: The Acid-Base Status of the Blood (ed 4).
Copenhagen, Munksgaard, 1974
10. Worthley LIG: Acid-base balance and disorders, in Bersten AD, Soni
N (eds): Oh’s Intensive Care Manual. Sydney, Butterworth-Heinmann,
2003, pp 873-883
11. Whittier WL, Rutecki GW: Primer on clinical acid-base problem
solving. Dis Mon 50:122-162, 2004
12. Sirker AA, Rhodes A, Grounds RM, et al: Acid-base physiology: the
“traditional” and the “modern” approaches. Anaesthesia 57:348-356,
2002
13. Schlichtig R, Grogono AW, Severinghaus JW: Human PaCO2 and
standard base excess compensation for acid-base imbalance. Crit Care
Med 26:1173-1179, 1998
14. Wilkes P: Hypoproteinemia, strong-ion difference, and acid-base sta-
tus in critically ill patients. J Appl Physiol 84:1740-1748, 1998
15. Jones NL: Our debt to Peter Stewart. Can J Appl Physiol 20:326-332,
1995
16. Stewart PA: Modern quantitative acid-base chemistry. Can J Physiol
Pharmacol 61:1444-1461, 1983
17. Stewart PA: How to Understand Acid-Base. New York, NY, Elsevier,
1981
18. Stewart PA: Independent and dependent variables of acid-base control.
Respir Physiol 33:9-26, 1978
19. Morris JG: A Biologist’s Physical Chemistry. London, Edward Ar-
nold, 1974
20. Gehlbach BK, Schmidt GA: Bench-to-bedside review: treating acid-
base abnormalities in the intensive care unit: the role of buffers. Crit
Care 8:259-265, 2004
21. Nahas GG: The clinical pharmacology of THAM (tris(hydroxymeth-
yl)aminoethane). Clin Pharmacol Ther 4:784-803, 1963
22. Dorland’s Illustrated Medical Dictionary (ed 30). Philadelphia, PA,
Saunders, 2003
23. Kellum JA: Determinants of blood pH in health and disease. Crit Care
4:6-14, 2000
24. Story DA, Morimatsu H, Bellomo R: Strong ions, weak acids and base
excess: a simplified Fencl-Stewart approach to clinical acid-base dis-
orders. Br J Anaesth 92:54-60, 2004
25. Henderson LJ: The theory of neutrality regulation in the animal or-
ganism. Am J Physiol 21:427-448, 1908
16 Seminars in Anesthesia, Perioperative Medicine and Pain, Vol 24, No 1, March 2005
26. Figge J, Rossing TH, Fencl V: The role of serum proteins in acid-base
equilibria. J Lab Clin Med 117:453-467, 1991
27. Figge J, Rossing T, Fencl V: Serum proteins and acid-base equilibria:
A follow-up. J Lab Clin Med 120:713-719, 1992
28. McAuliffe JJ, Lind LJ, Leith DE, et al: Hypoproteinemic alkalosis.
Am J Med 81:86-90, 1986
29. Story DA, Poustie S, Bellomo R: Quantitative physical chemistry
analysis of acid-base disorders in critically ill patients. Anaesthesia
56:530-533, 2001
30. Fencl V, Jabor A, Kazda A, et al: Diagnosis of metabolic acid-base
disturbances in critically ill patients. Am J Respir Crit Care Med
162:2246-2251, 2000
31. Siggaard-Andersen O, Fogh-Andersen N: Base excess or buffer base
(strong ion difference) as a measure of a non-respiratory acid-base
disturbance. Acta Anesthesiol Scand 39:123-128, 1995 (suppl 107)
32. Gomez JL, Roberts G, Schmigel J, et al: Effects of hypercapnia on
blood pressure in Nagase analbuminemic rats. Crit Care Med 31:A31,
2003 (suppl)
33. Astrup P, Severinghaus JW: The History of Blood Gasses, Acids and
Bases. Copenhagen, Munksgaard, 1986
34. Gamble J: Chemical Anatomy, Physiology and Pathology of Extracel-
lular Fluid. A Lecture Syllabus. Cambridge, Harvard University Press,
1954
35. Staempfli HR, Constable PD: Experimental determination of net pro-
tein charge and A(tot) and K(a) of nonvolatile buffers in human
plasma. J Appl Physiol 95:620-630, 2003
36. Wooten EW: Calculation of physiological acid-base parameters in
multicompartment systems with application to human blood. J Appl
Physiol 95:2333-2344, 2003
37. Constable PD: Total weak acid concentration and effective dissocia-
tion constant of nonvolatile buffers in human plasma. J Appl Physiol
91:1364-1371, 2001
38. Kellum JA, Bellomo R, Kramer DJ, et al: Splanchnic buffering of
metabolic acid during early endotoxemia. J Crit Care 12:7-12, 1997
39. Story D, Poustie S, Bellomo R: Estimating unmeasured anions in
critically ill patients: anion-gap, base-deficit, and strong-ion-gap. An-
aesthesia 57:1109-1114, 2002
40. Figge J, Jabor A, Kazda A, et al: Anion gap and hypoalbuminemia.
Crit Care Med 26:1807-1810, 1998
41. Mathes DD, Morell RC, Rohr MS: Dilutional acidosis: is it a real
clinical entity? Anesthesiology 86:501-503, 1997
42. Kellum JA, Bellomo R, Kramer DJ, et al: Etiology of metabolic
acidosis during saline resuscitation in endotoxemia. Shock 9:364-368,
1998
43. Odaira T: Influence of some neutral salt solutions, intravenously ad-
ministered, on the alkalai reserve of the blood. Tohoku J Exp Med
4:523-528, 1923
44. Traverso LW, Lee WP, Langford MJ: Fluid resuscitation after an
otherwise fatal hemorrhage. I. Crystalloid solutions. J Trauma 26:168-
175, 1986
45. Shires GT, Holman J: Dilution acidosis. Ann Int Med 28:557-559,
1948
46. Williams EL, Hildebrand KL, McCormick SA, et al: The effect of
intravenous lactated Ringer’s solution versus 0.9% sodium chloride
solution on serum osmolality in human volunteers. Anesth Analg
88:999-1003, 1999
47. McFarlane C, Lee A: A comparison of Plasmalyte 148 and 0.9% saline
for intra-operative fluid replacement. Anaesthesia 49:779-781, 1994
48. Reid F, Lobo DN, Williams RN, et al: (Ab)normal saline and physi-
ological Hartmann’s solution: a randomized double-blind crossover
study. Clin Sci (Lond) 104:17-24, 2003
49. Scheingraber S, Rehm M, Sehmisch C, et al: Rapid saline infusion
produces hyperchloremic acidosis in patients undergoing gynecologic
surgery. Anesthesiology 90:1265-1270, 1999
50. Morgan TJ, Venkatesh B, Hall J: Crystalloid strong ion difference
determines metabolic acid-base change during in vitro hemodilution.
Crit Care Med 30:157-160, 2002
51. Garella S, Chang BS, Kahn SI: Dilution acidosis and contraction
alkalosis: review of a concept. Kidney Int 8:279-283, 1975
52. Prough DS, Bidani A: Hyperchloremic metabolic acidosis is a predict-
able consequence of intraoperative infusion of 0.9% saline. Anesthe-
siology 90:1247-1249, 1999
53. Waters JH, Gottlieb A, Schoenwald P, et al: Normal saline versus
lactated Ringer’s solution for intraoperative fluid management in pa-
tients undergoing abdominal aortic aneurysm repair: an outcome
study. Anesth Analg 93:817-822, 2001
54. White SA, Goldhill DR: is Hartmann’s the solution? Anaesthesia
52:422-427, 1997
55. Healey MA, Davis RE, Liu FC, et al: Lactated ringer’s is superior to
normal saline in a model of massive hemorrhage and resuscitation.
J Trauma 45:894-899, 1998
56. Liskaser F, Story D: The acid-base physiology of colloid solutions.
Curr Opin Crit Care 5:440-442, 1999
57. Himpe D, Neels H, De Hert S, et al: Adding lactate to the prime
solution during hypothermic cardiopulmonary bypass: a quantitative
acid-base analysis. Br J Anaesth 90:440-445, 2003
58. Kellum JA: Fluid resuscitation and hyperchloremic acidosis in exper-
imental sepsis: improved short-term survival and acid-base balance
with Hextend compared with saline. Crit Care Med 30:300-305,
2002
59. Rocktaeschel J, Morimatsu H, Uchino S, et al: Unmeasured anions in
critically ill patients: can they predict mortality? Crit Care Med 31:
2131-2136, 2003
60. Rocktaeschel J, Morimatsu H, Uchino S, et al: Acid-base status of
critically ill patients with acute renal failure: analysis based on Stew-
art-Figge methodology. Crit Care 7:R60-R66, 2003
61. Balasubramanyan N, Havens PL, Hoffman GM: Unmeasured anions
identified by the Fencl-Stewart method predict mortality better than
base excess, anion gap, and lactate in patients in the pediatric intensive
care unit. Crit Care Med 27:1577-1581, 1999
62. Hatherill M, Waggie Z, Purves L, et al: Mortality and the nature of
metabolic acidosis in children with shock. Intensive Care Med 29:286-
291, 2003
63. Husain FA, Martin MJ, Mullenix PS, et al: Serum lactate and base
deficit as predictors of mortality and morbidity. Am J Surg 185:485-
491, 2003
64. Hayhoe M, Bellomo R, Liu G, et al: The aetiology and pathogenesis of
cardiopulmonary bypass-associated metabolic acidosis using polyge-
line pump prime. Intensive Care Med 25:680-685, 1999
65. Kaplan LJ, Kellum JA: Initial pH, base deficit, lactate, anion gap,
strong ion difference, and strong ion gap predict outcome from major
vascular injury. Crit Care Med 32:1120-1124, 2004
66. Morimatsu H, Rocktaschel J, Bellomo R, et al: Comparison of point-
of-care versus central laboratory measurement of electrolyte concen-
trations on calculations of the anion gap and the strong ion difference.
Anesthesiology 98:1077-1084, 2003
67. Story DA, Poustie S: Agreement between two plasma bicarbonate
assays in critically ill patients. Anaesth Intensive Care 28:399-402,
2000
68. Kofstad J: Base excess: a historical review: has the calculation of base
excess been more standardised the last 20 years? Clin Chim Acta
307:193-195, 2001
69. Kellum JA, Kramer DJ, Pinsky MR: Strong ion gap: a methodology
for exploring unexplained anions. J Crit Care 10:51-55, 1995