Comparison of the probability of four anticonvulsant mood stabilizers to facilitate polycystic ovary syndrome in women with epilepsies or bipolar disorder—A systematic review and meta-analysis

BACKGROUND: Patients treated with anticonvulsant mood stabilizers have a higher incidence of polycystic ovary syndrome (PCOS). However, there is no comparison between different anticonvulsant mood stabilizers. The purpose of this study was to systematically evaluate the prevalence of PCOS in women taking anticonvulsant mood stabilizers and compare the probability of PCOS caused by different anticonvulsant mood stabilizers. METHODS: Five databases, namely PubMed, Embase, Web of Science, Cochrane Library, and Clinical Trials, were searched for literature on anticonvulsant mood stabilizers and PCOS published up to October 28, 2022. This meta-analysis was performed using Revman 5.4, Stata 14.0, and R4.1.0, and effect size pooling was performed in fixed- or random-effects models based on the results of I(2) and Q-test, and the surface under the cumulative ranking curve (SUCRA) was used for analysis to assess the cumulative probability of drug-induced PCOS. Publication bias was assessed by funnel plot Egger's test and meta regression. RESULTS: Twenty studies with a total of 1,524 patients were included in a single-arm analysis, which showed a combined effect size (95% CI) of 0.21 (0.15–0.28) for PCOS in patients taking anticonvulsant mood stabilizers. Nine controlled studies, including 500 patients taking medication and 457 healthy controls, were included in a meta-analysis, which showed OR = 3.23 and 95% CI = 2.19–4.76 for PCOS in women taking anticonvulsant mood stabilizers. Sixteen studies with a total of 1416 patients were included in a network meta-analysis involving four drugs, valproate (VPA), carbamazepine (CBZ), oxcarbazepine (OXC), and lamotrigine (LTG), and the results of the network meta-analysis showed that VPA (OR = 6.86, 95% CI = 2.92–24.07), CBZ (OR = 3.28, 95% CI = 0.99–12.64), OXC (OR = 4.30, 95% CI = 0.40–49.49), and LTG (OR = 1.99, 95% CI = 0.16–10.30), with cumulative probabilities ranked as VPA (90.1%), OXC (63.9%), CBZ (50.1%), and LTG (44.0%). CONCLUSION: The incidence of PCOS was higher in female patients treated with anticonvulsant mood stabilizers than in the healthy population, with VPA having the highest likelihood of causing PCOS. The most recommended medication when considering PCOS factors is LTG. SYSTEMATIC REVIEW REGISTRATION: identifier: CRD42022380927