Abstracts

Rationale: Switching between FDA-approved and bioequivalent anti-epileptic drugs (AED) remains of some concern to physicians and patients because of possible untoward effects owing to a narrow therapeutic range. Recent research has shown that the risk of seizure-related events was elevated in epilepsy patients who had a recent change in medication. We sought to investigate the association between switching A-rated AED and seizure-related events in a large medical database. Methods: A retrospective case-control analysis was performed using the Thompson Healthcare MarketScan(R) commercial database (16 million covered lives) utilizing dates of service between 2005-2006. Cases were identified as patients that experienced an ambulance ride, emergency room visit or inpatient hospitalization with the primary diagnosis of epilepsy in 2006. Controls were those with an epilepsy (primary diagnosis) outpatient office visit in 2006 and were matched three to one on age and epilepsy diagnosis (ICD-9 codes). Cases and controls were limited to: ≥12 and <65 years old, those with 6 months continuous eligibility, no seizure related event in 6 months prior to index, and ≥145 days of AED prescription claims in the 6 months prior to index. Pharmacy claims were evaluated for an A-rated switch in the 6 months prior to index. We estimated the relative risk of seizure-related events using a matched odds ratio (discordant pairs) and conditional logistic regression adjusting for matching, gender, and total number of AED prescriptions filled in the 6 months prior to index. Results: Cases (N=757) were younger and had a different distribution of seizure diagnosis compared to unmatched controls (N=9707) (p-values <0.0001), but had similar gender and geographic distributions. A total of 84 out of 757 (11.1%) cases and 147 out of 2271 matched controls (6.5%) experienced an A-rated switch during the six months prior to index. Of those with a switch, 8 out of 84 cases and 9 out of 147 matched controls experienced more than one switch during the six months prior to index. The odds of a seizure-related event was 1.78 (95% CI 1.35 to 2.36) fold higher for those that switched and adjusted for gender and total number of AED prescriptions filled was 1.57 (95% CI 1.17, 2.10). The time from last switch to the index date was 13.1 days (95% CI 1.2, 25.0) shorter for cases. The majority of switches occurred in phenytoin, clonazepam or zonisamide. Conclusions: Switching between A-rated formulations of the same AED is associated with an increased risk of seizure-related events when compared to patients without an A-rated switch. Care should be taken to both advise and monitor patients when switching between such formulations.