Serum aspartate aminotransferase (AST) activity was measured by the methods recommended by the Scandinavian Committee on Enzymes (SCE) and by the International Federation of Clinical Chemistry (IFCC) with pyridoxal phosphate (PLP) and without (-PLP) in one laboratory at 37 degrees C with the Abbott ABA-100 and in another at 30 degrees C with the IL Multistat III. Reference ranges were determined on 195 healthy hospital staff. Sera from 102 patients with suspected hepatobiliary disease (HBD) and 104 with suspected myocardial infarction (MI) were assayed at both laboratories by all three methods. Based on the above reference ranges, all assays with each method at both hospitals were abnormal in 59 of 67 cases with HBD and 53 of 55 with MI. In aggregate, all three methods yielded comparable rates of misclassification (20-23). The SCE method gave highest false negatives (18) and lowest false positives (5); the IFCC method gave lowest false negatives (1) and highest false positives (20); intermediate values of 8 false positives and 12 false negatives were given by the IFCC (-PLP) method. Using receiver operating characteristic (ROC) curves, the SCE method was clearly superior at 30 degrees C, and the IFCC (-PLP) method was marginally superior at 37 degrees C. However, when the decision threshold corresponded with a 2.5% false positive rate in the non-HBD, non-MI patients, the SCE method gave the lowest false negatives at both temperatures and, on the basis of the present data, must be considered to be the method of choice for AST activity determinations.