We considered the association between diabetes and risk of endometrial cancer using data from a large case-control study conducted in Italy. Cases were 752 women with incident, histologically confirmed endometrial cancer < 75 years of age (median age 60 years, range 28-74) admitted to a network of hospitals in Milan. Controls were 2,606 patients (median age 54 years, range 25-74) aged < 75 years, admitted for acute non-neoplastic, non-gynecological, non-hormone-related conditions to the same network of hospitals where cases had been identified. A total of 132 (17.6%) cases and 116 controls (4.5%) reported a history of diabetes. The corresponding multivariate odds ratio (OR) was 2.9 [95% confidence interval (CI) 2.2-3.9]. No association emerged with diabetes diagnosed under age 40 (likely to be insulin-dependent diabetes), whereas the OR of endometrial cancer was 3.1 (95% CI 2.3-4.2) for diabetes diagnosed at age > or = 40 years. The OR of endometrial cancer in women with history of diabetes was 3.0 for women with a body mass index (BMI) (QI) kg/m2 < 25, 3.6 for those with a BMI of 25-29, and 3.3 for those with a BMI > or = 30. No consistent interaction or modifying effect was observed for any other covariate. Our results confirm that non-insulin-dependent diabetes is associated with the risk of endometrial cancer. The association may be mediated through elevated oestrogen levels in diabetic women, hyperinsulinemia or insulin-like growth factor-I (IGF-I).