Abstracts

Rationale: Adverse side effects (SE) are the 6th leading cause of death and are responsible for about 7% of hospital admissions. Anti-epileptic drugs (AEDs) can cause significant SE in up to 40% of patients. Quality-of-life, morbidity and economic burden can be improved by reducing SE. Levetiracetam (LEV) is a relatively new AED that is effective at reducing seizures in both focal and generalized epilepsy. While LEV was associated with few SE in early clinical trials, pragmatic experience suggests that LEV is associated with adverse psychological side effects (PSE). However, this concern to our knowledge has not been studied in large community-based cohorts. Here, we examine whether treatment with LEV increased the risk of PSE compared to other common AEDs in a population-based cohort.Methods: Persons with epilepsy (PWE) were identified in The Health Information Network (THIN) database from the UK. Patients were identified using a validated case definition. Briefly, patients must have at least one code for epilepsy syndromes or two codes for epilepsy symptoms, as well as a subsequent prescription for LEV, phenytoin, valproic acid, carbamazepine, lamotrigine, clobazam, topiramate or phenobarbital. A two-year washout period was used prior to the prescription date to ensure it was an incident prescription. Only patients older than 18 years at prescription and on monotherapy were considered. The hazard ratio of having a PSE within the 2 years following LEV prescription vs. any of the other AEDs was calculated using an accelerated time-to-failure survival model with sex and age as covariates. Psychiatric codes were identified via systematic review of the READ ontology by clinical experts.Results: We identified 34,665 PWE in THIN (49.3% female, mean age: 46.2 ± 19.2 SD). The three most common AEDs prescribed were: carbamazepine (11,769; 33.9%), valproate (10,906; 31.5%), and phenytoin (7,434; 21.4%). LEV was prescribed as the initial drug in 407 patients (1.2%). Use of LEV was a negative predictor for remaining free of PSE within 2 years following prescription (HR = 0.63; 0.47 - 0.85 CI; p<0.01) compared to prescription for any of the other seven AEDs. Older age at prescription (HR = 1.007; 1.005 - 1.008 CI; p<0.001) and male gender (HR=1.57; 1.45 – 1.69 CI; p<0.001) were significant predictors for freedom from PSE. If analysis is restricted to patients without a history for psychiatric codes before prescription, the effect of LEV diminishes but almost reaches significance (HR=0.68; 0.46 – 1.01 CI, p=0.057), while age (HR=1.005; 1.003 – 1.008 CI, p<0.001) and male gender (HR=1.41; 1.28 – 1.55 CI, p<0.001) remain predictive of freedom from PSE.Conclusions: LEV may increase the chance of PSE when used as monotherapy, particularly in individuals with pre-existing psychological conditions. Further analysis including factors such as comorbid burden, socioeconomic status, therapy change and type of PSE are required. While we examined a large epilepsy population, few people received LEV as an initial monotherapy, emphasizing the need for big data when exploring subpopulations.