Evaluation of serum anion gap in patients with liver cirrhosis of diverse etiologies

Mt Sinai J Med. 2004 Sep;71(4):281-4.


Background: The low serum anion gap (AG) in patients with hepatic cirrhosis is generally attributed to hypoalbuminemia. Serum immunoglobulin G (IgG) (elevated in chronic viral hepatitis) and IgA (elevated in alcoholic cirrhosis) have different isoelectric points, and thus may affect serum AG in opposite directions.

Aim: To define the normal serum AG in patients with liver cirrhosis of diverse etiologies.

Study design: We retrospectively compared serum AG of 144 stable cirrhotics and 286 control patients (consecutive hospital admissions with serum creatinine concentration < 2 mg/dL).

Results: Serum AG was significantly lower among the cirrhotics, compared to the controls (5.8 +/- 2.2 mEq/L vs. 7.0 +/- 2.2 mEq/L, respectively, p<0.005). However, when patients with serum albumin concentration < 3.5 g/dL were excluded, there was no significant difference between the cirrhotics vs. controls (6.7 +/- 1.8 mEq/L vs. 7.0 +/- 2.2 mEq/L, p=ns). Moreover, patients with liver cirrhosis secondary to chronic viral hepatitis had AG similar to that of the alcoholic cirrhotics (5.6 +/- 2.5 mEq/L vs. 6.0 +/- 1.9 mEq/L, p=ns). There was a positive correlation between serum albumin concentrations > 1.9 g/dL and serum AG, and a tendency toward an inverse correlation between serum globulin concentration and serum AG.

Conclusion: Our results support the contention that hypoalbuminemia accounts for the decreased serum AG frequently observed in patients with liver cirrhosis. We found no difference in serum AG with different causes of cirrhosis. We also suggest a lower reference range for normal serum AG.

Publication types

  • Comparative Study

MeSH terms

  • Acid-Base Equilibrium / physiology*
  • Alcoholism / blood
  • Alcoholism / complications
  • Female
  • Hepatitis, Viral, Human / blood
  • Hepatitis, Viral, Human / complications
  • Humans
  • Liver Cirrhosis / blood*
  • Liver Cirrhosis / etiology
  • Male
  • Middle Aged
  • Reference Values
  • Retrospective Studies