Lamotrigine vs levetiracetam in female patients of childbearing age with juvenile absence epilepsy: A Bayesian reanalysis.

ABSTRACT

OBJECTIVE: Women of childbearing age with juvenile absence epilepsy (JAE) face treatment challenges due to limited access to safe and effective anti-seizure medications (ASMs). In a previous study we compared the effectiveness of levetiracetam (LEV) and lamotrigine (LTG) in women with idiopathic generalized epilepsy (IGE), highlighting a superiority of LEV in juvenile myoclonic epilepsy. In this study, we specifically reanalyzed, through a Bayesian approach and by expanding the previously published cohort, the comparative effectiveness of these ASMs as initial monotherapy in JAE. METHODS: We conducted a multicenter, retrospective, comparative effectiveness study on women of childbearing age diagnosed with JAE and prescribed LEV or LTG as the initial ASM. Inverse probability treatment weighting (IPTW) Bayesian Cox proportional hazard models were employed to evaluate treatment failure (TF) due to ineffectiveness and ASM retention. The patients' center of provenance and year of prescription were considered as random effect factors. Posterior probabilities and relative log-risk distribution were computed, and the distribution of posterior draws was analyzed to assess the evidence supporting LTG superiority over LEV. RESULTS: Of 123 patients, those treated with LTG (n = 67) demonstrated lower TF and higher ASM retention than those treated with LEV (n = 56), with the IPTW-weighted Bayesian Cox proportional hazards model showing a 99.2% posterior probability of LTG being superior on TF and a 99.5% probability on ASM retention. Additional analyses on ≥50% and ≥75% seizure reduction through IPTW-weighted Bayesian logistic regression largely confirmed these findings, whereas the two ASMs did not show evident differences in terms of seizure freedom. The two ASMs showed comparable safety profiles, with only a minority of patients discontinuing treatment due to side effects. SIGNIFICANCE: Bayesian reanalysis supports LTG as first-line monotherapy for JAE in women of childbearing age, emphasizing the importance of individualized treatment strategies in women with IGE. This study underscores the value of Bayesian methods in refining clinical research and treatment decisions.