Abstract
BACKGROUND: Both bipolar disorder (BD) and epilepsy (ES) have been linked to polycystic ovary syndrome (PCOS) that is one of the most common endocrine disorders in women of reproductive age. The antiseizures medication valproate is widely used in the treatment of both disorders but has been suspected to increase the risk of PCOS. Previous studies have been limited by small sample sizes and heterogeneous definitions. We aimed to investigate the association between valproate exposure and incident PCOS in females with BS and ES. METHODS: We conducted a register-based cohort study including all females in Denmark with a first diagnosis of BD (ICD-10: F30.x-F31.x) or ES (ICD-10: G40.x) between January 1, 2000, and July 31, 2022. Women with BD, ES, valproate exposure, or PCOS prior to January 1, 2000, were excluded. Exposure to valproate was primarily modeled as current cumulative exposure. We also included a never/ever analysis and an overall cumulative analysis, accumulating dosages over the entire study period. The outcome was incident PCOS (ICD-10: E28.2). Cox regression models adjusted for age at diagnosis and calendar year were applied. RESULTS: The cohort comprised of 20,967 women, 8,003 diagnosed with BD and 12,964 diagnosed with ES. In total, 266 females developed PCOS during follow-up, of whom 160 had been exposed to valproate. In the main analysis, current cumulative exposure was strongly associated with PCOS, with HRRs rising from 4.43, 95%CI:3.42-5.73 (0-90 DDDs) to 7.08, 95%CI:3.85-13.03 (> 365 DDDs), P < 0.001. In the never/ever analysis, valproate exposure was also associated with increased PCOS risk (HRR 1.55, 95%CI:1.20-2.00). By contrast, overall cumulative exposure showed a less consistent pattern, with risk most clearly elevated in the highest dosage category (> 365 DDDs, HRR 2.04, 95%:CI 1.28-3.20), p < 0.01. CONCLUSIONS: Valproate exposure was associated with an increased risk of PCOS. The risk was especially pronounced during current and cumulative exposure, whereas overall cumulative exposure suggested increased risk at higher thresholds. These findings suggest that PCOS risk may be driven by acute pharmacological effects, although long-term cumulative use may also contribute. The results reinforce recommendations to avoid valproate in women of reproductive age when possible.