We explore whether racial differences in a United States population influence disease prevalence and perinatal outcome in gestational diabetes mellitus (GDM). The data presented are based on 3744 consecutive patients who underwent universal screening at 24-28 wk gestation; those with a 1-h plasma glucose greater than or equal to 7.2 mM underwent a 100-g 3-h oral glucose tolerance test (OGTT). The overall prevalence of GDM was 3.5 cases/100 with the standard O'Sullivan-Mahan diagnostic criteria derived for plasma, whereas use of the Carpenter-Coustan modification of the O'Sullivan-Mahan criteria yielded a prevalence of 5.5. The population was 39.1% white, 37.7% black, 19.8% Hispanic, and 3.4% Oriental/other. For those patients with a nondiagnostic test, mean plasma glucose at each time point of the OGTT was similar for all racial groups. Because of demographic and phenotypic heterogeneity between different racial groups, the influence of these different variables on the prevalence of GDM was tested by multiple logistic regression. Black and Hispanic race, maternal age, and percentage ideal body weight were found to have significant independent effects on the prevalence of GDM (P less than 0.05, 0.001, 0.001, and 0.001, respectively). The adjusted relative risk of GDM was significantly higher in black (1.81, 95% confidence interval [CI] 1.13-2.89, P less than 0.05) and Hispanic (2.45, 95% CI 1.48-4.04, P less than 0.001) patients compared with whites. The influence of race on infant birth weight was examined in the 92 patients with GDM controlled with diet.(ABSTRACT TRUNCATED AT 250 WORDS)