Updated definitions of healthy ranges for serum alanine aminotransferase levels.

Background: Serum alanine aminotransferase (ALT) activity, the variable most commonly measured to assess hepatic disease, fails to identify many patients with hepatic injury. Current standards for "normal" ALT level were defined by using populations that included persons with subclinical liver disease.
Objective: To update definitions of healthy ranges for serum ALT level.
Design: Retrospective cohort study.
Setting: A university hospital in Milan, Italy.
Participants: 6835 persons who were first-time blood donors from 1995 through 1999, were negative for anti-hepatitis C virus (HCV), and had no contraindications to donation and 209 persons who attempted to donate blood from 1990 through 1999 but were found to have anti-HCV antibodies. Of the latter group, 131 had HCV viremia.
Measurements: Univariate and multivariate analyses examined associations between clinical and laboratory factors and ALT levels. Healthy ranges for ALT were computed from the population at lowest risk for liver disease. Sensitivity and specificity of healthy ALT ranges were evaluated in the donors with HCV antibodies, of whom 133 had liver biopsy.
Results: Serum ALT activity was independently related to body mass index and to laboratory indicators of abnormal lipid or carbohydrate metabolism. Updated upper limits (for men, 500 nkat/L [30 U/L]; for women, 317 nkat/L [19 U/L]) were lower than current limits (for men, 667 nkat/L [40 U/L]; for women, 500 nkat/L [30 U/L]) and, during 6-month follow-up, showed superior sensitivity in identifying participants with HCV viremia (sensitivity, 76.3% [95 % CI, 69.1% to 83.6%] vs. 55% [CI, 46.4% to 63.5%]). The related tradeoff in specificity was acceptable (88.5% [CI, 79.2% to 94.6%] vs. 97.4% [91% to 99.7%]). The increased sensitivity targeted patients with minimal to mild histologic lesions.
Conclusion: In patients with chronic HCV infection or nonalcoholic fatty liver disease, revision of normal limits for ALT level is advisable.