The purpose of this study was to determine which test is best for predicting adverse outcomes in pregnancies complicated by diabetes: the nonstress test, biophysical profile, or umbilical artery velocimetry. We evaluated 207 singleton pregnancies complicated by diabetes within 1 week of delivery using the afore-mentioned pregnancy surveillance tests. Adverse pregnancy outcome was defined as delivery before 37 weeks of gestation or the occurrence of fetal growth restriction, hypocalcemia, hypoglycemia, hyperbilirubinemia, respiratory distress syndrome, or fetal risk requiring cesarean delivery. The prognostic value of each of the three tests was assessed, after controlling for the mothers' White classification and third trimester glycosylated hemoglobin value. Among the 207 pregnancies, 75 (36.2%) had an adverse outcome. In pregnancies in which the umbilical artery systolic to diastolic ratio was > or = 3.0, the relative risk of adverse outcome was 2.6 (95% confidence interval: 1.9-3.5, P < 0.001). For those with a biophysical profile < or = 6 the relative risk was 1.7 (95% confidence interval: 0.9-2.9, P = 0.109). Patients with a nonreactive nonstress test had a relative risk of 1.7 (95% confidence interval: 1.2-2.5, P = 0.009). Umbilical artery Doppler velocimetry was superior to either the nonstress test or the biophysical profile in identifying the subgroup of pregnancies complicated by diabetes that resulted in an adverse outcome.