Case control study of SUDEP (Sudden unexpected death in epilepsy)

Sudden unexpected death in people with epilepsy (SUDEP) is an under-appreciated tragedy. It has devastating effects on families, friends and colleagues of those who die, and the causes of SUDEP remain unknown. We will conduct an international, multicentre, prospective, case-control study of SUDEP, recruiting participants over four years. Each centre will define a cohort from which cases and controls will be prospectively identified. The cohort will comprise people with epilepsy who have been seen at the centre since a specific date (no earlier than 01/01/2015). Cases will be people with epilepsy from these pre-specified cohorts who have died from definite or probable SUDEP (with or without other co-morbidity), or resuscitated (near) SUDEP if the patient died within 72 hours of the initial collapse. Controls will be people with epilepsy who will be individually-matched by age (±2 years), sex, centre, and enrolled in the cohort at the time of death of the case (with four controls for each case). For each case, three controls will be randomly selected from all patients in the same cohort who meet the above matching criteria, using a random number generator. A fourth control will be a proxy control, who will be a spouse, relative or close friend nominated by a true control, who is randomly selected from the true-control triplet. Use of the proxy controls will allow any measurement error associated with proxy data to be quantified and corrected for. The families of cases, the control patients, and the families of controls patients (i.e. proxy controls), will be interviewed over the phone by a research coordinator or an epilepsy fellow, using a structured questionnaire. Data from the interviews will be recorded in the EpiNet database. This will include demographic data, socio-economic factors, epilepsy factors, usual sleeping arrangements, epilepsy treatments, co-morbidities, medications, alcohol, caffeine, and cigarette use, plus the use of seizure monitors, nocturnal supervision, and/or anti-suffocation pillows. For cases, relatives will be asked about the deceased's sleeping arrangements for the night of death or for the night immediately prior to death, and whether this was different from the typical sleep arrangements. Controls will be asked questions relating to their sleep arrangements on a specific nominated night's sleep (chosen randomly to be in the two weeks prior to the interview). Data neurologists, pathologists, and coroners regarding circumstances around the death, the cause of death, epilepsy and treatment factors will also be recorded if available. The data will be analysed to identify risk factors for SUDEP. Data cleaning will be performed prior to analysis. Odds ratios will be calculated using the Mantel-Haenszel method and logistic regression to control for covariates.