Abstracts

Rationale: Some commonly prescribed AEDs such as carbamazepine and phenytoin are known to induce hepatic enzymatic activity. This can lead to alterations in endogenous metabolic pathways and concomitant medication levels, contributing to development of comorbidities (e.g. osteoporosis, cardiovascular risk). Higher long-term healthcare costs with the use of EIAEDs compared with nEAAEDs were previously reported for adult patients in the UK. We have performed the same analysis in patients ≥65 years of age; a population likely to have considerable co-medication and comorbidity burdens. Methods: The Clinical Practice Research Datalink of UK primary care medical records, linked to Hospital Episode Statistics data, was used to select patients ≥65 years with epilepsy, prescribed an EIAED or nEAAED between 2001 and 2010 (index), after ≥1 year without AEDs (baseline). Patients were followed until time to treatment failure or censoring (Fig 1). Propensity score matching was used to reduce the effects of confounding factors between the two cohorts. Key outcomes were time to index AED treatment failure (end of index AED or start of new AED), and direct healthcare costs (drugs; primary care consultations; accident and emergency visits; outpatient referrals; procedures; and hospitalizations) in 2014 £GBP values. Median monthly costs were compared using the Mann-Whitney U test. Categorical and continuous variables were summarized descriptively. Time to index AED treatment failure was analyzed using Kaplan-Meier (K-M) survival methods. Results: The initial elderly population consisted of 1,425 patients; 964 with EIAEDs and 461 with nEAAEDs. At baseline, the EIAED cohort was older (mean age of 76.2 vs 75.1 years) and more likely to be male. Baseline direct healthcare costs were similar. After matching, the 2 cohorts (n=210 each) were similar on all selected confounders. The most commonly prescribed EIAED and nEAAED were carbamazepine (52.9%) and lamotrigine (51.9%), respectively. The estimated median time to index AED treatment failure for the EIAED cohort was 807 days vs 910 days for the nEAAED cohort. Over the entire follow-up period, median monthly direct healthcare costs were significantly higher for patients on EIAEDs than nEAAEDs; £403 vs £317 (p=0.0150). Costs were higher for patients remaining in the EIAED cohort at each year during follow-up (Fig 2). Conclusions: In the UK, during the period studied, newly treated patients (≥65 years old) with epilepsy were more likely to be prescribed EIAEDs than nEAAEDs, despite the known risks for drug interactions and comorbidities that EIAEDs may pose. In matched cohorts, time to index AED treatment failure was shorter for elderly patients taking EIAEDs than those taking nEAAEDs. Although the drug acquisition costs for EIAEDs are generally lower, the average total direct healthcare cost was larger for elderly patients on EIAEDs than on nEAAEDs. Changing treatment practices might have potential to improve patient care and reduce costs. Funding: UCB Pharma-sponsored.