Abstracts

Rationale: A recent AAN/AES practice guideline (Harden et al., 2017) identified risk factors for Sudden Unexpected Death in Epilepsy (SUDEP). The strongest level of evidence is for the risk factor of having at least one generalized tonic-clonic (GTC) seizure in the last 12 months, but there remain gaps in the field. Amassing enough cases of SUDEP to study prospectively and conduct multivariate analyses to develop robust prediction algorithms is a challenge. Individuals and families who use apps to track seizures may be an enriched population for the study of SUDEP and its risk factors because they are likely to have sufficient number of seizures. We evaluated a cohort of intractable epilepsy patients to determine the range of risk across SUDEP risk factors. Our hypothesis is that app users are likely to be at greater risk than the general epilepsy population. Methods: Risk factors were identified through a literature search including a focus on the AAN/AES practice guideline on SUDEP and the SUDEP-7 which is the only known SUDEP risk factor inventory (Novak et al., 2015). Several risk factors have been proposed or investigated; however a subset of factors (see Table 1) were selected based on the strength of the evidence, if there was conflicting evidence, or had high relevance for pediatric patients and if the information was available in our cohort. Our cohort consisted of 26,785 registered users of an online tool (SeizureTracker™) self-reporting over 2 million seizures since 2007. Information was de-identified and descriptive analysis across the factors was conducted. We also reported data from a 1-year period to conduct a binomial test to determine if our cohort differs from historical controls from a meta-analysis of 4 case-control studies from four countries (Hesdorffer et al., 2011). Results: Data from the full SeizureTracker cohort used information logged by users between November 2007-March 2019. Users were balanced across sex (52.0% female, 4.5% unreported) and age (45% <18years). Of all seizures logged,15.9% were GTCs (n=322,204), which includes secondarily generalized events. Forty percent of GTCs occurred between 8pm-6am indicative of nocturnal seizures. Length was under 5 minutes for 96% of GTCs with 57% under 1 minute; 0.05% lasted >30 minutes. Over a 1-year period (March 2018-March 2019), 289,224 seizures were logged by 821 users; 39,074 (14%) were GTCs. Of the 821 users with GTCs, 20.4% had only one GTC; 61.1% had >3 GTCs. The number of people with >3 GTCs was greater in the SeizureTracker population compared with historical controls (11.9%, p<.000001), even compared to those who had died from SUDEP (48.0%, p<.000001). Conclusions: This cohort is at high risk of SUDEP with over 60% at a 15-fold increase based on the GTC risk factor alone. It is estimated that in this high-risk group there are 18 deaths per 1,000 per year. We hope this cohort might allow us to predict better and improve risk assessment which will improve care by being able to inform and educate families with better precision. This is a first step to developing a SUDEP risk algorithm. This cohort could also provide an enriched clinical population to trial SUDEP preventative strategies such as nocturnal supervision (e.g., co-sleeping, seizure watch, etc.). Limitations include potential data collection bias as one is relying on parent/care provider- or self-report (e.g., GTCs generally last >