Strong ion gap: a methodology for exploring unexplained anions

J Crit Care. 1995 Jun;10(2):51-5. doi: 10.1016/0883-9441(95)90016-0.


Purpose: This paper describes the calculation of the strong ion gap (SIG), a physical chemical methodology similar to the anion gap (AG), as a measure of the anion/cation balance exclusive of sodium, potassium, chloride, and bicarbonate. We compared the SIG and AG methodologies in three groups of subjects with and without unexplained anions. These groups were (1) healthy volunteers with hyperlacticemia during exercise; (2) intensive care unit (ICU) patients with sepsis; and (3) ICU patients with severe liver disease.

Methods: The SIG, AG, and corrected AG (AGc) were calculated for each group from data available in the original reports (groups 1 and 2) and by retrospective chart review (group 3).

Results: The SIG correlated poorly with the AG in group 2, whereas no correlation was seen in groups 1 and 3. The AGc correlated with SIG in all three groups (r = .99, .93, and .91 respectively; P < .01 for each group). Although the AG was similar, the SIG differed for each group. Group 1 had levels of SIG near zero, and groups 2 and 3 had mean SIG's of 4.80 +/- 4.67 mEq/L and 9.60 +/- 6.43 mEq/L respectively. The composition of the anion gap differed markedly among subject types.

Conclusions: The SIG correlates with the AG once corrected for all known anions. The SIG technique can detect unknown anions in a patient population known to have them and does not detect unknown anions in healthy volunteers during exercise. This test detects large amounts of unknown anions in some patients with sepsis or liver disease. Therefore, the test is both sensitive and specific in characterizing metabolic acidosis.

Publication types

  • Comparative Study

MeSH terms

  • Acid-Base Equilibrium*
  • Acidosis, Lactic / blood*
  • Chemistry, Physical / methods
  • Data Interpretation, Statistical
  • Humans
  • Liver Diseases / blood*
  • Reproducibility of Results
  • Retrospective Studies
  • Sepsis / blood*