Background Valproate is associated with teratogenic and neurodevelopmental effects. Several agencies have restricted the conditions of its prescription in bipolar disorders (BD). We aimed to assess the evolution of valproate prescription and the clinical profile of BD women of childbearing age receiving valproate. Methods Based on a large national cohort, we included all BD women 16–50 years old. Sociodemographic, clinical and pharmacological data were recorded. Logistic regression analyses were used to describe variables associated with valproate prescription. Results Of the 1018 included women 16–50 years old, 26.9% were treated with valproate with a mean daily dosage of 968 mg. The prevalence of BD women using valproate was 32.6% before May 2015 and 17.3% after May 2015 (p<0.001), the date of French regulatory publication of restriction of valproate prescription. The multivariate analysis revealed that the inclusion period after May 2015 (OR=0.54, CI 95% 0.37–0.78, p=0.001), the age lower than 40 years (OR=0.65, CI 95% 0.43–0.98, p=0.040) and the number of lifetime mood episodes (OR=0.98, CI 95% 0.95–0.99, p=0.040) were the variables negatively associated with the use of valproate. Limitations Study could be underpowered to determine a clinical profile associated with valproate prescription. Conclusions The regulatory change in BD women of childbearing age had a significant impact on valproate prescription, even if the prescription rate remains high. Important efforts are needed to help clinicians and patients to improve quality of care in BD women of childbearing age.< Réduire