Strong Ion Gap


Published on

A critique of the strong ion gap as a predictor of mortality

Edward M. Omron MD, MPH
Pulmonary, Critical Care Medicine
Morgan Hill, CA 95037

Published in: Health & Medicine
1 Comment
  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Strong Ion Gap

  1. 1. Letters to the EditorStrong Ion Gap recognized risk factor for poor outcome survivor group and suggest that our re- in the acutely ill (5). sults are inconsistent with the publishedTo the Editor: It would be of interest to know literature in healthy adults or acutely ill We read with interest the recent study whether nonsurvivors in this study suf- patients. Their only reference is a publi-by Drs. Kaplan and Kellum (1) investigat- fered significantly more renal impair- cation by Fencl et al. (1), which did noting the discriminative power of the strong ment as a result of their injury. If this examine acutely ill patients and onlyion gap calculation in separating survivors were the case, since serum creatinine lev- studied nine healthy controls along withfrom nonsurvivors after major vascular els are a lagging indicator of acute renal 152 critically ill patients already in thetrauma. We congratulate the authors on a failure, it may turn out that the incre- ICU. Importantly, Fencl and colleaguesmost interesting and provocative research ment in the strong ion gap may represent did not calculate SIG, but it can be cal-study. Although the magnitude of the a surrogate, early, and sensitive indicator culated from the data provided in thestrong ion gap value in the nonsurvivor of acute renal failure in the acutely ill. manuscript. Drs. Omron and Gilbertgroup was discussed at length, the more might have misinterpreted the [XA-] vari-interesting clinical finding may be the low Edward M. Omron, MD, MPH, Russell able reported by Fencl et al. as the SIG,strong ion gap value in the survivor group C. Gilbert, MD, Pulmonary and Criti- but the two are not the same. As dis-(2.4 1.8 mEq/L). The authors state that a cal Care Division, National Naval Med- cussed in our report (2), and as publishednormal plasma strong ion gap value is ical Center, Bethesda, MD previously (3, 4), SIG SIDa SIDe, 2 mEq/L. In what patient population are where SIDe (the effective strong ion dif-they referring to? This cutoff value is not REFERENCES ference) is calculated from the CO2, albu-consistent with the strong ion gap calcula- min, and Pi concentrations exactly as per 1. Kaplan LJ, Kellum JA: Initial pH, base deficit,tion in either healthy adults or the acutely Fencl et al. (SIDe is labeled simply “SID” lactate, anion gap, strong ion difference, andill in the published literature (2). The in- in their report, and SIDa [the apparent strong ion gap predict outcome from majorcrement or decrement in the strong ion vascular injury. Crit Care Med 2004; 32: strong ion difference] is calculated fromgap is a relative value since normal and 1120 –1124 the Na , K , Ca2 , Mg2 , Cl , and no-abnormal values have not been established 2. Fencl V, Jabor A, Kazda A, et al: Diagnosis of tably, lactate and urate (2– 4). Since uratein critical care populations. The magnitude metabolic acid-base disturbances in critically is typically not an important factor ( 0.2of the strong ion gap value in nonsurvivors ill patients. Am J Respir Crit Care Med 2000; mEq/L), it is usually omitted. However,in Kaplan and Kellum’s study is similar to 162:2246 –2251 lactate is not so trivial and, thus at aother recently published articles on this 3. Cusack RJ, Rhodes A, Lochhead P, et al: The minimum, [XA-] would be expected totopic in adult medical and surgical inten- strong ion gap does not have prognostic value in critically ill patients in a mixed medical/ equal SIG lactate. Since lactate is notsive care units (3, 4). Their retrospective reported in the publication by Fencl et surgical adult ICU. Intensive Care Med 2002;study stands apart from the previous stud- al., we can only speculate that it was 28:864 – 869ies only with respect to the lower strong ion 4. Rocktaeschel J, Morimatsu H, Uchino S, et al: normal in the healthy controls and,gap value in survivors. In an unpublished Unmeasured anions in critically ill patients: therefore, could not have accounted forretrospective study of 31 consecutive adult, Can they predict mortality? Crit Care Med more than about 2 mEq/L.blunt trauma patients who survived to dis- 2003; 31:2131–2135 However, there is a further, and verycharge, we found on presentation to the 5. Rocktaeschel J, Morimatsu H, Uchino S, et al: significant, problem with the calculationintensive care unit the strong ion gap Acid-base status of critically ill patients with of [XA-]. This calculation used total,1.8 0.6 mEq/L, the strong ion difference acute renal failure: Analysis based on Stewart- Figge methodology. Crit Care 2003; 7:R60 –R66 rather than ionized, Ca2 and Mg2 .effective 33.6 0.8, and the strong ion Bound Ca2 and Mg2 do not behave asdifference apparent 35.4 0.8 mEq/L. DOI: 10.1097/01.CCM.0000148174.18460.9D strong ions and only ionized Ca2 andTen of the 31 patients manifested a nega- Mg2 should be considered in the calcu-tive strong ion gap calculation suggestive of The author replies: lation of SIDa. While ionized, Mg2 isunmeasured plasma cations. One might We appreciate the insightful com- typically only about 30% lower than totalspeculate that the decrement in the strong ments by Drs. Omron and Gilbert regard- Mg2 (5), using total instead of ionizedion gap (strong ion gap 5 mEq/L) before ing our recent work that applies Stewart’s Ca2 would result in a significant over-volume resuscitation in survivors of major physical-chemical approach to acid-base estimate of SIDa. Using the data fromvascular trauma is the more important dynamics to a selected group of trauma Table 2 of the Fencl et al. study (1), weprognostic finding and that unmeasured patients with major vascular injury. We can calculate SIG (assuming a lactate 2plasma cations play a role. agree that the results are indeed provoc- mmol/L) at 2.3 mEq/L. Although this is Acute renal failure causes an in-crease in the strong ion gap as well as ative. As intensivists and traumatologists still higher than the 0.3 0.6 mEq/Lthe more familiar anion gap and is a search for better means of risk stratifica- calculated from published data on exer- tion for critically ill and injured patients, cising humans in the original report of an elevated strong ion gap (SIG) is SIG (3) or the 0.5 1.8 mEq/L found in emerging as a potentially valuable tool. healthy volunteers reported in abstract Copyright © 2005 by the Society of Critical Care However, Drs. Omron and Gilbert form by Gunnerson et al. (6), it is cer-Medicine and Lippincott Williams & Wilkins have focused on the “low” SIG in the tainly much closer. Finally, to have 8266 Crit Care Med 2005 Vol. 33, No. 1