Sudden Unexpected Death in Epilepsy: A Personalized Prediction Tool

• Published in: Neurology Vol. 96; no. 21; p. e2627
• Main Authors: Jha, Ashwani, Oh, Cheongeun, Hesdorffer, Dale, Diehl, Beate, Devore, Sasha, Brodie, Martin J, Tomson, Torbjörn, Sander, Josemir W, Walczak, Thaddeus S, Devinsky, Orrin
• Format: Journal Article
• Language: English
• Published: United States 25.05.2021
• Subjects: Adolescent; Adult; Aged; Bayes Theorem; Epilepsy; Female; Humans; Male; Middle Aged; Risk Factors; Sudden Unexpected Death in Epilepsy; Young Adult
• DOI: 10.1212/WNL.0000000000011849

To develop and validate a tool for individualized prediction of sudden unexpected death in epilepsy (SUDEP) risk, we reanalyzed data from 1 cohort and 3 case-control studies undertaken from 1980 through 2005. We entered 1,273 epilepsy cases (287 SUDEP, 986 controls) and 22 clinical predictor variables into a Bayesian logistic regression model. Cross-validated individualized model predictions were superior to baseline models developed from only average population risk or from generalized tonic-clonic seizure frequency (pairwise difference in leave-one-subject-out expected log posterior density = 35.9, SEM ± 12.5, and 22.9, SEM ± 11.0, respectively). The mean cross-validated (95% bootstrap confidence interval) area under the receiver operating curve was 0.71 (0.68-0.74) for our model vs 0.38 (0.33-0.42) and 0.63 (0.59-0.67) for the baseline average and generalized tonic-clonic seizure frequency models, respectively. Model performance was weaker when applied to nonrepresented populations. Prognostic factors included generalized tonic-clonic and focal-onset seizure frequency, alcohol excess, younger age at epilepsy onset, and family history of epilepsy. Antiseizure medication adherence was associated with lower risk. Even when generalized to unseen data, model predictions are more accurate than population-based estimates of SUDEP. Our tool can enable risk-based stratification for biomarker discovery and interventional trials. With further validation in unrepresented populations, it may be suitable for routine individualized clinical decision-making. Clinicians should consider assessment of multiple risk factors, and not focus only on the frequency of convulsions.