(321) Therapeutic strategies after discontinuation of valproate in clinical practice in women with epilepsy: a cohort study in the United Kingdom and France databases

Background: Valproate (VPA) indicated to treat epilepsy and bipolar disorder can be teratogenic to unborn children if taken during pregnancy. In 2018 European authorities recommended strong restrictions on the use of VPA in women of childbearing potential (WCBP) and pregnant women, and set-up a pregnancy prevention program. Identifying and evaluating the real-life practices for epilepsy therapeutic management leading up to a successful switch after VPA discontinuation were therefore needed.

Objectives: To determine the clusters of patients the most likely to reflect a success in epilepsy management after VPA discontinuation, and to identify their associated factors.

Methods: WCBP who had been using VPA for epilepsy and discontinued its use between 1 Jan. 2014 and 31 Dec. 2017 (index date) were identified in the French SNDS and UK CPRD databases and followed-up for 1 year. Clusters of women that most likely reflected a success in epilepsy management after VPA discontinuation were identified based on treatment patterns. Success was defined based on No VPA reintroduction in the follow-up period, contextualised according to clinical relapse, hospitalisation, polypharmacy. Baseline factors associated with successful/unsuccessful clusters were also assessed in SNDS (not feasible in CPRD due to small sample size).

Results: In total 7,345 (SNDS) and 358 (CPRD) WCBP with an epilepsy diagnosis were included of whom 67.3% (SNDS) and 49.4% (CPRD) in successful clusters. After VPA discontinuation, 27.7%, 25.5%, 10% and 4% in SNDS had a trend to use 0, 1, 2 and ≥3 combined antiepileptic drug (AED); 20.9%, 20.7%, 7.8% in CPRD had a trend to use 0, 1 and 2 AEDs. In SNDS, factors most strongly associated with no reintroduction of VPA were: more specific care (neurology imaging exam or neurologist visits) in the 90 days prior index date (OR= 2.30), VPA dose-tapering phase in the 1-year pre-index (OR=2.40), pregnancy at index (OR=1.96), levetiracetam or lamotrigine prescription in the 90 days prior index (OR=1.81, OR=1.54). Factors most strongly associated with reintroduction of VPA were: older age (OR=0.49, 0.68, 0.82 for [40-49], [30-39], [20-29] year old vs [13-29]), no exposure to an AED within 3 months after index (OR=0.59), more dispensing of other nervous system treatments during the 90-days prior index (OR=0.58, 0.73 for >5 and [4-5] vs < 4 treatments), longer history of epilepsy (OR=0.63, 0.71 for ≥5 and ]4-5[ vs < 1 year of history).

Conclusions: Discontinuing VPA was maintained in half of the WCBP with epilepsy, especially if young, with a stabilised disease. Treatments used after discontinuation were consistent with the experts’ consensus. VPA was mostly reintroduced in older women with a more advanced disease and a resurgence of clinical relapses, probably to control their disease.