The prevalence of polycystic ovaries, according to ultrasonography, and associated clinical, endocrine, and metabolic features were investigated in women with previous gestational diabetes mellitus (GDM). Thirty-four women with GDM 3-5 yr before the investigation and 36 controls with uncomplicated pregnancies, selected for similar age, parity, and date of delivery, were investigated. The women with previous GDM showed a higher prevalence of polycystic ovaries [14 of 34 (41%) vs. 1 of 36 (3%); P < 0.0001], hirsutism (P < 0.01), irregular menstrual cycles (P < 0.01), and a higher body mass index (BMI; P < 0.001) than the controls. Five women (15%) with previous GDM had developed manifest diabetes (excluded in comparisons of metabolic variables). After dividing the women with previous GDM into subgroups according to ovarian appearance, the 2 subgroups showed similar glucose tolerance and prevalence of diabetes, whereas the women with polycystic ovaries were younger (mean +/- SD, 33.3 +/- 1.4 vs. 38.2 +/- 1.1; P < 0.01), had higher truncal-abdominal/femoral fat ratio according to skin folds (P < 0.05), had higher concentrations of androstenedione (P < 0.01) and testosterone (P < 0.01), and had a higher LH/FSH ratio (P < 0.01), lower levels of GH (P < 0.01), higher levels of triglycerides (P < 0.05) and cholesterol (P < 0.05) in very low density lipoprotein, all independent of age and BMI, and had a higher prevalence of pregnancy-induced hypertension (50% vs. 15%; P < 0.05) during the index pregnancy compared with the women with normal ovaries. The group of women with GDM showed a lower early insulin release after glucose (i.v. glucose tolerance test) for their degree of insulin resistance (euglycemic hyperinsulinemic clamp) compared with controls (P < 0.05). In the two subgroups, insulin sensitivity was lower in the polycystic ovaries group, independent of BMI (P < 0.05), than in the group with normal ovaries. In conclusion, ultrasonographic, clinical and endocrine signs of polycystic ovary syndrome were much increased in women with a history of GDM. Compared with the women with normal ovaries and previous GDM, those with polycystic ovaries formed a distinct subgroup that may be more prone to develop various features of the insulin resistance syndrome. Both groups showed a similarly disturbed balance between beta-cell activity and insulin sensitivity, but in women with polycystic ovaries, insulin resistance may be the dominant component.