Printer Friendly

Stability of the strong ion gap versus the anion gap over extremes of PC[O.sub.2] and pH.


The strong ion gap (SIG) is under evaluation as a scanning tool for unmeasured ions. SIG is calculated by subtracting [buffer base], which is ([[A.sup.-]]+[HC[O.sup.-.sub.3]]), from the apparent strong ion difference, which is ([[Na.sup.+]+ [[K.sup.+]]+[[Ca.sup.++]]+[[Mg.sup.++]]-[[Cl.sup.-]]-[L-lactate]). [A.sup.-] is the negative charge on albumin and phosphate. We compared the pH stability of the SIG with that of the anion gap (AG).

In normal and hypoalbuminaemic hyperlactaemic blood, PC[O.sub.2] was reduced stepwise in vitro from >200 mmHg to <20 mmHg, with serial blood gas and electrolyte analyses, and [albumin] and [phosphate] measurement on completion.

Respective [haemoglobin], [albumin], [phosphate] and [lactate] in normal blood were 156 (0.9) g/l, 44 (2) g/l, 1.14 (0.06) mmol/l and 1.7 (0.8) mEg/l, and in hypoalbuminaemic blood 116 (0.9) g/l, 24 (2) g/l, 0.78 (0.06) mmol/l and 8.5 (0.5) mEq/l. pH increased from <685 to >755, causing significant falls in [[Na.sup.+]] and elevations in [[Cl.sup.-]]. Initial and final SIG values did not differ, showing no correlation with pH. Mean SIG was 0.5 [+ or -] 1.5 mEg/l. AG values were directly correlated with pH (normal: [R.sup.2]=0.51, hypoalbuminaemic: [R.sup.2]=0.65). Final AG values significantly exceeded initial values (normal blood. 15.9 (1.7) mEg/l versus 8.9 (1.8) mEg/l, P<0.01; hypoalbuminaemic blood: 16.5 (0.8) mEg/l versus 11.8 (20) mEq/l, P<0.05).

We conclude that, unlike the AG, the SIG is not affected by severe respiratory acidosis and alkalosis, enhancing its utility in acid-base disturbances.

Key Words: anion gap, pH, strong ion gap, variability


In 1990 Jones proposed a new tool to scan for unmeasured plasma ions (1), based on Stewart's physical chemical approach to acid-base (2,3). An almost identical suggestion appeared subsequently in a paper by Figge and co-workers (4). In 1995, Kellum and colleagues described the tool as the 'strong ion gap' (SIG), an advance on its predecessor, the anion gap (AG) (5). This is a slight misnomer, since the unmeasured ions creating the 'gap' can be either strong or weak, but the term has persisted.

The SIG concept exploits the fact that the plasma strong ion difference (SID), which is the net charge in mEq/1 of all fully dissociated plasma anions and cations (i.e. strong ions), can be estimated either as the 'apparent' SID (SIDa), or as the 'effective' SID (SIDe) (6). SIDa is [[Na.sup.+]]+[[K.sup.+]]+[[Ca.sup.++]]+[[Mg.sup.++]-[[Cl.sup.-]][L-lactate]. SIDa is normally about 42 mEq/1. SIDe takes the opposite tack by calculating the total plasma concentration of weak negative ions [[A.sup.-]] + [HC[O.sub.3.sup.-]], where [A.sup.-] is the ionised component of nonvolatile weak acid (albumin and phosphate). [A.sup.-] and HC[O.sub.3.sup.-] are the 'buffer base' ions (7). Their negative charge counters and therefore equals the net positive charge of the strong ions (in other words the SID), by the principle of electrical neutrality.

SIG is SIDa-SIDe. In the absence of measurement error, SIG should be zero unless there are unmeasured ions. 'Unmeasured ions' in this context means charged entities participating in overall plasma electrical neutrality but not incorporated in the gap equation. Unmeasured strong anions, for example high concentrations of keto-anions, increase SIG by introducing positive bias into SIDa. Unmeasured weak anions (for example myeloma paraproteins) increase SIG by causing negative bias in SIDe. Conversely, unmeasured strong cations such as lithium reduce SIG by introducing negative bias into SIDa, whereas unmeasured weak cations (TRAM administration) create positive bias in SIDe, again reducing SIG (8).

Despite these multiple influences, the SIG signal should be subject to less noise than that of the AG (9), since the list of unmeasured ions that can contribute to the AG is much longer'. Specifically, the AG but not the SIG should be altered by abnormal concentrations of L-lactate, [Ca.sup.++], [Mg.sup.++], albumin or phosphate.

The AG also varies with pH and PC[O.sub.2], primarily as a response to changing negative charges on albumin and phosphate. The so-called 'albumin corrected' AG has no pH adjustment (10). Here again, the SIG calculation should be immune, since it incorporates pH and PC[O.sub.2] in its SIDe component. However, inaccuracies in the SIDe pH adjustment, if present, are potentially magnified in severe acid-base disturbances, situations when the detection of unmeasured ions is important (8). Figge's simplified quantification of weak acid dissociation, normally used to calculate SIDe, is therefore vital (4).

The SIG may be a useful addition to the clinical diagnostic armamentarium (8), but requires evaluation before it becomes part of standard biochemical reporting. Because pH stability is desirable in any proposed scanning tool for unmeasured ions, we tested the robustness of the SIG calculation across a wide spectrum of in vitro respiratory acid-base equilibria, varying from extreme respiratory acidosis to extreme alkalosis. We used normal blood and, in order to stress the SIG algorithm further, we also tested blood with hypoalbuminaemia and lactic acidosis. For comparison we performed simultaneous AG calculations.


The need for ethics approval was waived by the Mater Health Services Human Research Ethics Committee. Blood was collected on separate days from each of the four investigators and tested both unmodified and after dilution with Compound Sodium Lactate (Hartmann's Solution, Baxter Healthcare, Sydney, Australia), a balanced salt solution containing L-lactate 29 mmol/1.

A single 18 ml sample of venous blood was collected from each investigator in a sodium heparin coated syringe and divided equally between two 50 ml syringes. Into one, 3 ml of Hartmann's solution was aspirated. Specimens were equilibrated with 100% C[O.sub.2] aspirated into the syringes (BOC Gases Australia Ltd) by gentle agitation for five minutes, after which the C[O.sub.2] was expelled and the first blood gas and electrolyte measurement performed.

Each specimen was then subjected to a series of two minute equilibrations with air, using a repeated sequence of air aspiration and gentle agitation, followed by air expulsion and blood gas, electrolyte and co-oximetry measurement. The cycle was continued until a PC[O.sub.2] target of <20 mmHg was achieved, a process taking approximately 20 minutes. No attempt was made to control specimen temperatures during this time, although all analyser measurements were at 37[degrees]C. Each set of measurements recorded included [haemoglobin], plasma pH, PC[O.sub.2], [[Na.sup.+]], [[K.sup.+]], [[Ca.sup.++]] (ionised), [[Cl.sup.-]] and [L-lactate] (ABL 700, Radiometer, Copenhagen, Denmark). After the final equilibration, plasma [albumin] and [phosphate] were determined by multi-channel analysis (Vitros Chemistry System, Ortho-Clinical Diagnostics, Rochester, U.S.A.).


AG=[[Na.sup.+]]+[[K.sup.+]]-[[Cl.sup.-]]-[HC[O.sub.3.sup.-]] with all measurements in mEq/1, and [HC[O.sub.3.sup.-]] determined as 0.0301 x PC[O.sub.2] x [10.sup.(pH - 6.1)], where PC[O.sub.2]=the plasma carbon dioxide tension in mmHg.

SIDa=[[Na.sup.+]]+[[K.sup.+]]+[[Ca.sup.++]]-[[Cl.sup.-]]-[L-lactate], with all measurements in mEq/1. [[Mg.sup.++]] (ionised) was not available.

[[A.sup.-]], the plasma concentration (mEq/1) of nonvolatile weak anion, was determined using the equation of Figge and co-authors (4):

[[A.sup.-]]=[albumin] (1.123 pH-0.631) +[phosphate] (0.309 pH-0.469)

with [albumin] expressed in g/1 and [phosphate] in mmol/1.



Statistical analysis

Correlations between variables were analysed by linear regression. Paired t-tests were used to compare initial and final measured and derived variables. Significance was accepted at P [less than or equal to] 50.05. Data are expressed as mean (SD).


On the four undiluted blood specimens, C[O.sub.2] equilibration and stepwise C[O.sub.2] elimination produced 11, 10, 9 and 7 blood gas and electrolyte measurement episodes respectively, each set of measurements (along with plasma [albumin] and [phosphate]) allowing simultaneous calculations of SIG and AG (Figure 1). The lowest pH achieved was 6.8 and 7.57 was the highest. In the diluted specimens, there were 11, 8, 9 and 7 sets of measurements respectively with simultaneous SIG and AG calculations, across a pH range of 6.64 to 7.71.


Mean [haemoglobin], plasma [albumin] and [phosphate] in the undiluted specimens were 156 (0.9) g/l, 44 (2) g/l and 1.14 (0.06) mmol/l respectively. In the diluted specimens these values were 116 (0.9) g/l, 24 (2) g/l and 0.78 (0.06) mmol/l respectively. Results from the first and last measurements are set out in Table 1.

The transition from respiratory acidosis to alkalosis generated significant decreases in plasma [[Na.sup.+]] and [[Ca.sup.++]] and increases in [[Cl.sup.-]] and [L-lactate]. The increase in [[Cl.sup.-]] was particularly marked, more so in the undiluted specimens. These changes caused large decreases in SIDa, again more so in the undiluted specimens. All SIDa decreases were accompanied by near identical changes in SIDe, so that final SIG values did not differ significantly from initial values. In both undiluted and diluted specimens, SIG values were close to zero for the entire pH range, so that mean SIG for the complete data-set was 0.5 [+ or -] 1.5 mEq/l.

There was no correlation between pH and SIG values ([R.sup.2]=0.01 and 0.08 for undiluted and diluted specimens respectively). In contrast, AG values were directly correlated with pH ([R.sup.2]=0.51 and 0.65 for undiluted and diluted specimens respectively), with final values significantly exceeding initial values in undiluted and diluted specimens.


We found that the SIG algorithm was robust in the setting of severe pH stress (<6.85 to >7.55), in both normal blood and in blood diluted to produce hypoalbuminaemia co-existent with hyperlactaemia. These findings reaffirm the accuracy of the Figge formula (4), in which plasma protein weak acid activity is attributed to albumin alone, and in which linear approximations are used to calculate the negative charge on albumin and phosphate at any given pH. Our data indicate that these approximations do not hinder the ability of SIDe to track SIDa across a wide pH range, despite considerable SID variation from ionic shifts between erythrocytes and plasma. As a result SIG, which is the difference between SIDa and SIDe, is likely to retain its ability to reveal unmeasured ions during severe acid-base disturbances.

In contrast, we found that the AG increased steadily as PC[O.sub.2] fell and pH rose, ultimately almost doubling in the undiluted specimens. The AG increase with pH is well known, largely attributable to altered albumin and phosphate dissociation (11). This is one likely factor in the poor performance of the AG as a scanning tool for unmeasured ions (12,11), one that according to our findings is not shared by the SIG.

Although the pH stability of the SIG signal is encouraging, a remaining concern is the number of analyte concentrations in the algorithm. Each brings additional imprecision. SIG confidence intervals due to summated measurement variability could be as wide as [+ or -] 7 mEq/l (14). Our inability to include ionised magnesium measurements replicates the current situation in most laboratories. This will have a small effect on SIG reference ranges, but should otherwise be of little consequence. Finally, one correctable source of error in SIDe is over-estimation of [albumin] in severe hypoalbuminaemia. Immunoassay should be instituted at the appropriate [albumin] threshold.

We conclude that the pH stability of the SIG in vitro is superior to that of the AG, an advantage in any potential scanning tool for unmeasured ions. It would be a simple matter to calculate and report the SIG whenever blood gas analysis coincides with a biochemistry profile--a common occurrence in critical illness. Before this practice is adopted, we recommend a direct evaluation of SIG sensitivity and specificity in scanning for specific ions of known concentration, both in vitro and ideally during in vivo assessment of acid-base disorders.

Accepted for publication on December 7, 2006.


(1.) Jones NL. A quantitative physicochemical approach to acid-base physiology. Clin Biochem 1990; 23:189-195.

(2.) Stewart PA. Modern quantitative acid-base chemistry. Can J Physiol Pharmacol 1983; 61:1444-1461.

(3.) Stewart PA. How to understand acid-base. In: Stewart PA ed. A quantitative acid-base primer for biology and medicine. New York: Elsevier; 1981:1-286.

(4.) Figge J, Mydosh T, Fencl V Serum proteins and acid-base equilibria: a follow-up. J Lab Clin Med 1992; 120:713-719.

(5.) Kellum JA, Kramer DJ, Pinsky MR. Strong ion gap: a methodology for exploring unexplained anions. J Crit Care 1995; 10:51-55.

(6.) Schlichtig R, Grogono AW, Severinghaus JW Current status of acid-base quantitation in physiology and medicine. Int Anesthesiol Clin 1998; 16:211-213.

(7.) Singer RB, Hastings AB. An improved clinical method for the estimation of disturbances of the acid-base balance of human blood. Medicine 1948; 27:223-242.

(8.) Emmet M, Narins RG. Clinical use of the anion gap. Medicine 1977; 56:38-54.

(9.) Morgan TJ. What exactly is the strong ion gap, and does anybody care? Crit Care Resusc 2004; 6:155-159.

(10.) Figge J, Jabor A, Kazda A, Fencl V Anion gap and hypoalbuminemia. Crit Care Med 1998; 26:1807-1810.

(11.) Madias NE, Ayus JC, Adrogue HJ. Increased anion gap in metabolic alkalosis: the role of plasma protein equivalency. N Engl J Med 1979; 300:1421-1423.

(12.) Salem MM, Mujais SK. Gaps in the anion gap. Arch Intern Med 1992; 152:1625-1629.

(13.) Iberti TJ, Leibowitz AB, Papadakos PJ, Fischer EP Low sensitivity of the anion gap as a screen to detect hyperlactaemia in critically ill patients. Crit Care Med 1991; 19:130-131.

(14.) Story DA, Poustie S, Bellomo R. Comparison of three methods to estimate bicarbonate in critically ill patients: Henderson-Hasselbalch, enzymatic, and strong ion gap. Anaesth Intensive Care 2001; 29:585-590.

T. J. MORGAN *, D. M. COWLEY ([dagger]), S. L. WEIER ([double dagger]), B. VENKATESH ([section])

Department of Chemical Pathology, Mater Health Services, Brisbane, Queensland, Australia

* M.B., B.S., F.J.F.I.C.M., Senior Specialist, Adult Intensive Care Unit, Mater Health Services,

([dagger]) M.B., Ch.B., B.Sc. (Hons), F.R.C.P.A., F.H.G.S.A., Director of Chemical Pathology, Mater Health Services Pathology.

([double dagger]) M.Sc., Senior Scientist, Mater Health Services Pathology.

([section]) F.R.C.A., F.J.F.I.C.M., M.D., Professor of Intensive Care, Intensive Care Units, Princess Alexandra Hospital, Wesley Hospital and University of Queensland.

Address for reprints: Dr T.J. Morgan, Adult Intensive Care, Mater Misericordiae Health Services, Raymond Terrace, South Brisbane, Qld. 4101.
Initial and final measured and derived variables in normal blood and
in blood diluted with Compound Sodium Lactate solution. Values are
presented as mean (SD).

 Initial (normal) Final (normal)

PC[O.sub.2] (mmHg) 264 (19) 19 (2) (c)
pH 6.84 (0.03) 7.59 (0.03) (c)
[[Na.sup.+]] (mEq/l) 144 (3) 137 (3) (c)
[[K.sup.+]] (mEq/l) 4.4 (0.4) 4.2 (0.4)
[[Ca.sup.++]] (ionised) (mEq/l) 2.64 (0.14) 2.04 (0.10)
[[Cl.sup.-]] (mEq/l) 96 (l) 108 (2) (c)
[L-Lactate] (mEq/l) 1.7 (0.8) 2.3 (0.8) (a)
[HC[O.sub.3.sup.-]] (mEq/l) 43.3 (0.7) 17.1 (1.0) (c)
[[A.sup.-]] (mEq/l) 11.0 (0.4) 15.3 (0.5) (c)
SIDa (mEq/l) 53.1 (1.1) 32.8 (0.7) (c)
SIDe (mEq/l) 54.3 (0.5) 32.4 (1.3) (c)
SIG (mEq/l) -1.2 (l.3) 0.4 (1.5)
AG (mEq/l) 8.9 (1.8) 15.9 (1.7) (b)

 Initial (diluted) Final (diluted)

PC[O.sub.2] (mmHg) 232 (18) 15 (5) (c)
pH 6.77 (0.03) 7.57 (0.11) (c)
[[Na.sup.+]] (mEq/l) 139 (2) 133 (2) (b)
[[K.sup.+]] (mEq/l) 4.7 (0.2) 4.6 (0.3)
[[Ca.sup.++]] (ionised) (mEq/l) 2.68 (0.10) 2.24 (0.14) (c)
[[Cl.sup.-]] (mEq/l) 100 (1) 108 (1) (b)
[L-Lactate] (mEq/l) 8.5 (0.5) 9.3 (0.5) (b)
[HC[O.sub.3.sup.-]] (mEq/l) 32.4 (l.5) 13.1 (1.3) (c)
[[A.sup.-]] (mEq/l) 6.0 (0.3) 8.6 (0.2) (c)
SIDa (mEq/l) 38.3 (l.1) 22.5 (1.5) (c)
SIDe (mEq/l) 38.4 (l.8) 21.7 (1.5) (c)
SIG (mEq/l) -0.1 (1.9) 0.8 (0.8)
AG (mEq/l) 11.8 (2.0) 16.5 (0.8) (a)

(a) Differs from initial values P<0.05.

(b) Differs from initial values P<0.01.

(c) Differs from initial values P<0.001.
COPYRIGHT 2007 Australian Society of Anaesthetists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2007 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Morgan, T.J.; Cowley, D.M.; Weier, S.L.; Venkatesh, B.
Publication:Anaesthesia and Intensive Care
Article Type:Clinical report
Geographic Code:8AUST
Date:Jun 1, 2007
Previous Article:Mycoplasma pneumoniae associated with Stevens Johnson syndrome.
Next Article:High thoracic epidural analgesia for cardiac surgery: an audit of 874 cases.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters