A prospective population-based study of gestational diabetes mellitus was done with 2272 patients to determine perinatal and maternal outcomes. A large data base was collected on all patients. Patients with gestational diabetes mellitus were older, shorter, heavier, and had more children than did the control group. The higher cesarean section rate in the patients with gestational diabetes mellitus was explained by their increased rate of repeat cesarean section compared with control patients. This was associated with increased infectious complications. Other maternal complication rates were similar in the two groups. Acceptable glucose control did not normalize birth weight percentiles in patients with gestational diabetes mellitus. Maternal weight at delivery was the only significant predictor of birth weight percentile in the group with gestational diabetes mellitus. Plasma glucose levels were a poor predictor of birth weight percentile. Factors associated with maternal obesity in well-controlled gestational diabetes mellitus may be more significant than glucose control in the development of large-for-gestational-age infants.