Gestational diabetes mellitus (GDM) is associated with defects in insulin secretion and insulin action, and women with a history of GDM carry a high risk for the development of non-insulin-dependent diabetes mellitus (NIDDM). Assessment of subjects with a history of GDM who are currently normoglycemic should help elucidate some of the underlying defects in insulin secretion or action in the evolution of NIDDM. We have studied 14 women with normal oral glucose tolerance who had a history of GDM. They were compared with a group of control subjects who were matched for both body mass index (BMI) and waist-to-hip ratio (WHR). All subjects underwent tests for the determination of oral glucose tolerance, ultradian oscillations in insulin secretion during a 28-h glucose infusion, insulin secretion in response to intravenous glucose, glucose disappearance after intravenous glucose (K_g), and insulin sensitivity (S_I) as measured by the Bergman minimal model method. The BMI in the post-GDM women was similar to that in the control subjects (24.9 ± 1.2 vs. 25.4 ± 1.4 kg/m², respectively), as was the WHR ratio (0.80 ± 0.01 vs. 0.76 ± 0.01, respectively). The post-GDM women were slightly older (35.2 ± 0.9 vs. 32.1 ± 1.4 years, P = 0.04). The fasting plasma glucose levels were significantly higher in the post-GDM group than in the control group (4.9 ± 0.1 vs. 4.4 ± 0.1 mmol/l, respectively, P < 0.001) and remained higher at each of the subsequent determinations during the oral glucose tolerance test, although none had a result indicative of either diabetes or impaired glucose tolerance. All measures of ultradian insulin secretory oscillations in post-GDM subjects were indistinguishable from those in the control subjects. The first-phase insulin release to intravenous glucose was lower in the post-GDM group. S_I was also impaired in the post-GDM group compared with the control subjects (4.6 ± 0.5 vs. 6.8 ± 1.0·10⁻⁴·min⁻¹· ³U−1·ml, respectively, P < 0.05). K_g was reduced in the post-GDM women compared with the control subjects (1.3 ± 0.1 vs. 2.7 ± 0.4%, P < 0.01). When the subjects were divided according to their BMI, lean post-GDM subjects (<24.2, n = 8) were more insulin resistant than the lean control subjects: S_I 5.3 ± 0.6 vs. 8.8 ± 1 · 1·10⁻⁴ min ⁻¹· ³U⁻¹·ml, P = 0.02, whereas obese post-GDM (>24.2 kg/m², n = 6) and control subjects had a lower S_I than the lean subjects, but they were not different from each other (3.6 ± 0.7 vs. 4.2 ± 1.2· 10⁻⁴· min⁻¹ · ³U⁻¹ · ml, respectively, P = 0.67). The acute insulin responses to glucose (0–10 min) within these groups showed that the lean post-GDM group had a significantly lower insulin response compared with control subjects (1,205 ± 179 vs. 2,404 ± 416 pmol· 1⁻¹ min, respectively, P = 0.007), whereas the obese groups had similar responses (2,777 ± 1,112 vs. 3,114 ± 847 pmol ·1⁻·min, post-GDM vs. control subjects, P = 0.8). We have found defects in insulin secretion and action in post-GDM subjects who are at high risk for the development of NIDDM at a time that oral glucose tolerance is normal. These defects are present in the absence of obesity. Ultradian insulin secretory oscillations during constant glucose infusion are normal in these post-GDM subjects predisposed to NIDDM. We conclude that defects in both insulin secretion and insulin action are present before the development of hyperglycemia in women with a history of GDM.