A population-based post mortem study of sudden unexpected death in epilepsy

Damian Clark, Kate Riney

The aim of this study was to review population autopsy data on epilepsy-related deaths (ERD) in Queensland, Australia, to establish the incidence of autopsy-confirmed sudden unexpected death in epilepsy (SUDEP), explore factors associated with SUDEP, and determine if complete autopsy examinations of SUDEP were performed. All autopsy reports for a 5year period in Queensland were electronically searched for the terms 'epilepsy' or 'seizure'. The identified reports were reviewed, and data were extracted for all ERD. In the study period, 175 ERD were identified from autopsy records (123 SUDEP, 34 accident-related, 3 due to status epilepticus). From data available on the prevalence of epilepsy in Queensland (National Health Survey), the incidence of autopsy-confirmed SUDEP was 0.7 per 1000 person years (95% confidence interval 0.5-1.2 per 1000 person years). The factors associated with SUDEP were male sex (for those >18 years) and subtherapeutic anticonvulsant medication levels (found in 55%). Where recorded, the majority of deaths happened in the person's usual residence (90%), were overnight (70%) and unwitnessed (87%), with the person found prone (74%), in or adjacent to their bed (49%) and with signs of proximate seizure (60%). A complete autopsy was undertaken for only 59% of cases, the majority in urban locations. This study provides support for an unwitnessed overnight seizure being a key factor in autopsy-confirmed SUDEP in Queensland.