A commercially available enzyme-immunometric assay for serum alpha-glutathione S-transferase (GST) was evaluated. Endogenous serum alpha-GST diluted linearly within the calibration range. However, we recommend that the sample and second antibody reagent are always added sequentially in the assay to avoid hook effect. Between-assay variability was below 7% across the calibration range and the upper limit of the reference range in adults (n = 219) was 11.4 micrograms/L. Within-individual variability in serum alpha-GST concentrations measured over a 4-6 week period in 4 healthy adults was small. Serum alpha-GST concentrations did not change significantly 6 h after a therapeutic dose of paracetamol. Studies in two patients after liver transplantation showed that serum alpha-GST is a better discriminant of acute changes in liver function than conventional tests. Serum alpha-GST concentrations were unaffected by gross muscle damage, extrahepatic inflammation, or haemolysis and thus appear to be more liver specific than transaminase activities. The effect of renal impairment on serum alpha-GST concentrations requires further investigation.