Abstracts

Rationale: Patients with poststroke epilepsy (PSE) constitute a cardiovascular high-risk group with increased mortality. The choice of antiepileptic drug (AED) treatment seems to alter the risk of cardiovascular events and several studies have linked AEDs to abnormalities in lipid levels. Furthermore, enzyme-inducing AEDs leads to alterations in serum levels of many commonly prescribed drugs that are considered cornerstones in secondary prevention after stroke. The extent to which AED treatment contributes to mortality in this patient group is unknown. In this study, high-quality national registers were used to investigate if the choice of AED monotherapy influences the risk of cardiovascular and all-cause death. Methods: Patients with a stroke during the years 2005-2010 were identified using the Swedish stroke register (n=131330). The extracted data was cross-linked to three other national records for information on epilepsy, date and cause of death, and dispensed drugs. 31/12/2014 was declared the final point of data capture. 7790 patients met our criteria for PSE, i.e., a detected ICD-10-code for epilepsy, seizures, or status epilepticus (G40, G41, or R56.8) more than seven days after stroke. Having a record of AED dispensations before PSE (n=1530), dispensations of multiple AEDs (n=1580), being under 18 years of age at stroke onset (n=2) or having survived less than two months after stroke (n=52) resulted in exclusion, leaving 2926 patients with continuous AED monotherapy for survival analyses. Death from cardiovascular disease (CVD) was defined as having an ICD-10-code of I0-I99 as the underlying cause of death. When analysing cardiovascular mortality, Fine and Gray’s competing risk regression model was used with carbamazepine as reference. For all-cause mortality, we used Cox proportional hazards regression. The hazard ratios (HR) were adjusted for age, sex, stroke severity and comorbidities, including hypertension, smoking, diabetes and atrial fibrillation. Results: 1801 events of death occurred, and CVD was considered the underlying cause in 1131 of them. The adjusted HR (95%CI) of cardiovascular death compared to carbamazepine (CBZ) was 0.73 (0.58-0.93) for lamotrigine (LTG), 0.81 (0.62-1.06) for levetiracetam (LEV), 1.36 (1.15-1.62) for valproic acid (VPA), 1.00 (0.68-1.49) for phenytoin (PHT) and 0.72 (0.39-1.32) for oxcarbazepine (OXC). The adjusted HR (95%CI) of all-cause death compared to CBZ was 0.75 (0.62-0.91) for LTG, 1.08 (0.89-1.32) for LEV, 1.35 (1.18-1.56) for VPA, 1.22 (0.92-1.63) for PHT and 1.38 (0.93-2.05) for OXC. Conclusions: Patients with PSE who are treated with LTG in monotherapy seem to have significantly lower mortality compared to CBZ. The opposite applies for patients treated with VPA, who seem to have a higher risk of both cardiovascular and all-cause death compared to CBZ and LTG. The analyses are adjusted for several confounders, but further studies are needed to establish a causal relationship. Funding: Swedish society for medical research, Swedish society of medicine, Gothenburg society of medicine, Magnus Bergvall foundation, Felix Neuberg foundation, Region Västra Götaland.