Insulin resistance characterizes non-insulin dependent diabetes (NIDDM). Insulin resistance may coexist in clinical syndromes with hyperestrogenism and hyperandrogenism, suggesting that the ovary may be sensitive to effects of insulin. In addition, insulin-like growth factor-I receptors, which are capable of binding insulin, have been identified in ovarian cancer tissue and are proposed to regulate cell growth. We evaluated the association between a history of diabetes mellitus and ovarian cancer in a case-control study in seven counties in Washington and in Utah (United States) during the years 1975-87. Cases included women newly diagnosed with ovarian cancer over a five-year period who were identified through population-based cancer reporting. Controls similar to cases with regard to age and county of residence were identified via household surveys or random digit dialing. The study included 595 cases and 1,587 controls. Twenty-seven cases (4.5 percent) and 72 controls (4.5 percent) reported a history of diabetes. Logistic regression analysis of the association between diabetes and ovarian cancer controlling for age, body mass index, and race resulted in an odds ratio (OR) of 0.9 (95 percent confidence interval [CI] = 0.6-1.5). The OR was not changed with further controlling for prior oral contraceptive use or prior pregnancy. None of the 20 women with nonepithelial tumors (15 of which were stromal tumors) had a history of diabetes (upper CI = 4.0). These results, together with findings of two earlier cohort studies, do not support the hypothesis that diabetes is a risk factor for epithelial ovarian cancer.