While an abnormal oral glucose tolerance test (GTT) is known to be associated with an increased risk of pregnancy complications, the impact of one abnormal value is not clear. In 1986 we screened 4618 pregnant women for gestational diabetes at 24-28 weeks' gestation. Eighty-seven percent had normal results; of the 13% with abnormal screening tests, 139 had one abnormal value on the subsequent 3-hour oral GTT. These women were then compared with 725 randomly selected patients with a normal screening test. The frequency of chronic hypertension, cesarean section, 5-minute Apgar score below 7, preterm delivery, shoulder dystocia, congenital malformations, and perinatal mortality did not differ significantly between the groups. The incidence of macrosomia (birth weight above 4000 g) was significantly greater in the study group (18.0%) than in the control group (6.6%) (odds ratio 2.18; 95% confidence interval 1.77-5.37), a relationship that persisted after controlling for confounding risk factors by logistic regression modeling (odds ratio 2.55; 95% confidence interval 1.44-4.52). The incidence of preeclampsia/eclampsia was significantly greater in the study group (7.9%) than in the control group (3.3%) (odds ratio 2.51; 95% confidence interval 1.14-5.52), which also persisted after controlling for confounding risk factors using logistic regression modeling (odds ratio 2.81; 95% confidence interval 1.26-6.28). Our results suggest that patients with one abnormal value on an oral GTT during pregnancy are at risk for delivering macrosomic infants and developing preeclampsia/eclampsia.