Variation in Antiepileptic Drug Adherence Among Older Patients with New-Onset Epilepsy

Abstract

Background: Medication adherence among patients with epilepsy remains a significant challenge, even for patients prescribed newer antiepileptic drugs (AEDs), leading to increased risks of seizures, hospitalizations, and higher treatment costs. Despite substantial research identifying numerous risk factors, the role of specific medications has been neglected. Objective: To analyze adherence to 9 different AEDs in a national clinical sample of elderly patients with new-onset epilepsy. Methods: Patients over age 66 receiving care in the Veterans Health Administration were eligible if they met criteria for new-onset epilepsy with AED monotherapy of at least 3 months. A cross-sectional study design was used to assess adherence as defined by the medication possession ratio (MPR) and a 90-day or longer medication gap. Multivariable logistic regression modeled each dichotomous adherence outcome as a function of clinical and demographic measures. Results: The sample (N = 6373) was primarily male (98%), white (79%), and exempt from medication copayments due to disability status; nearly 40% had a prior psychiatric or dementia diagnosis. Nearly half of the patients were poorly adherent, with rates ranging from 42% to 63% across AEDs. In multivariable models, patients on phenobarbital, valproate, and gabapentin were significantly less likely to be adherent on both outcomes, while lamotrigine and levetiracetam were positively associated with adherence per the MPR. Conclusions: Adherence difficulty in this elderly cohort is attributable to several factors, yet specific AEDs substantially increased this risk. Drugs that produce adverse effects such as cognitive difficulty or weight gain may prove detrimental to maintaining appropriate adherence early in the treatment course. Given comparable efficacy among AEDs, providers should be aware that certain medications impart differential risks of poor adherence in older patients with epilepsy.