Simultaneous multiple automated analyses of liver function can be performed quickly and cheaply, but their usefulness in mass screening is questionable. Reference intervals are frequently applied without regard to race and sex, despite the fact that reported values may vary considerably in relation to these factors. Serum analyte results for greater than 5000 black and white men and women in the CARDIA Study showed clinically and statistically significant differences by race and sex for values of aspartate aminotransferase, gamma-glutamyltransferase, alkaline phosphatase, total bilirubin, total protein, and albumin; these differences were not explained by differences in age, body mass, reported ethanol intake, smoking, or oral contraceptive use. Results for at least one of these six tests were out of range in 38% of the men and 19% of the women. Sex- and race-specific reference intervals are recommended to decrease the frequency of values reported as abnormal in otherwise healthy young adults.