Contraception and the risk of type 2 diabetes mellitus in Latina women with prior gestational diabetes mellitus

Context.-Effective contraception is essential in women with prior gestational diabetes mellitus (GDM) but should not increase their already substantial risk of developing type 2 diabetes. Objective.-To determine whether exposure to low-dose oral contraceptives increases the risk of developing type 2 diabetes mellitus in women with recent GDM. Design.-Retrospective cohort study of 904 Latinas with GDM who gave birth between January 1987 and March 1994, in whom postpartum diabetes was excluded at 4 to 16 weeks post partum. Interventions.-At their initial postpartum visit, 443 women selected a nonhormonal form of contraception, 383 received a low-dose, estrogen-progestin combination oral contraceptive (OC), and 78 breast-feeding women received the progestin-only OC. When breast-feeding ended, patients initially taking progestin-only OCs were switched to combination OCs, Patients were followed up periodically with oral glucose tolerance tests for up to 7 1/2 years. Main Outcome Measures.-Person time was used to compute unadjusted average annual incidence rates of developing diabetes mellitus, as defined by the National Diabetes Data Group Criteria. Survival analysis was used to compute the unadjusted cumulative incidence rates and adjusted relative risks of diabetes mellitus. Results.-The unadjusted average annual incidence rates of type 2 diabetes mellitus were 8.7%, 10.4%, and 26.5%, respectively, for patients using nonhormonal forms of contraception, combination OCs, and progestin-only OCs. Cumulative incidence rates were virtually identical for patients with uninterrupted use of combination OCs and nonhormonal forms of contraception, but patients using progestin-only OCs developed diabetes mellitus more rapidly during the first 2 years of use. After adjustment for potential confounding factors, the use of progestin-only OCs almost tripled the risk of type 2 diabetes mellitus compared with equivalent use of low-dose combination OCs (adjusted relative risk, 2.87; 95% confidence interval, 1.57-5.27). The magnitude of this risk increased with duration of uninterrupted use. Conclusion.-Progestin-only OCs were associated with an increased risk of diabetes in breast-feeding Latinas with recent GDM and probably should be prescribed with caution, if at all, in these women. Long-term use of low-dose combination OCs did not increase the risk of type 2 diabetes compared with use of nonhormonal contraception. Thus, combination OCs do not appear to increase the risk of diabetes in non-breast-feeding women with recent GDM.