Umeå University's logo

Objective: To assess the risk of type 2 diabetes (T2D) in women with polycystic ovary syndrome (PCOS) in relation to body mass index (BMI) and the hyperandrogenic (HA) PCOS phenotype.

Design: Population-based cohort study.

Setting: Data from six Swedish national registers, with participants being followed for a maximum of 19 years.

Patient(s): All women with an International Statistical Classification of Diseases and Related Health Problems, version 10, diagnosis of PCOS, androgen excess, or anovulatory infertility born between 1950 and 1999 (n = 52,535) were identified in the Patient Register. The HA PCOS phenotype was defined by two filled prescriptions for anti-androgenic drugs. For each woman with PCOS, five control women (n = 254,624) were randomly chosen from the Total Population Register, matched for age and geographic area.

Intervention(s): No interventions were performed.

Main Outcome Measure(s): International Statistical Classification of Diseases and Related Health Problems, version 10, diagnosis of T2D or prescription of antidiabetic treatment other than metformin.

Result(s): The cumulative incidence rates of T2D were 1.3%, 4.4%, and 14.2% in controls (non-PCOS women) and women with normoandrogenic (NA) and HA PCOS, respectively. After adjustment for BMI, women with PCOS had a twofold higher rate of T2D than non-PCOS women (adjusted hazard ratio, 2.52 [95% confidence interval, 2.15–2.96]). Women with HA PCOS had a higher rate of T2D than those with NA PCOS (adjusted hazard ratio, 3.86 [95% confidence interval, 3.16–4.72]).

Conclusion(s): Polycystic ovary syndrome is an independent risk factor for T2D, even after adjustment for BMI. Women with the HA PCOS phenotype face an even higher risk of T2D than those with the NA PCOS phenotype.