Résumé
Objective: to develop and validate a clinical risk score of Sudden Unexpected Death in Epilepsy (SUDEP) for patients suffering from drug-resistant focal epilepsy. Methods: case-control study, including control patients from the French REPO2MSE cohort study, and certain or probable SUDEP cases collected by the French Sentinel Mortality Epilepsy Network. Study period: 2010-2019. Eligibility criteria: drug-resistant patients with focal epilepsy, potentially candidate to surgery. Multiple logistic regressions, coupled with multiple imputation to address missing data was performed. After internal validation, a simplified risk score was developed using estimates rounding value of selected variables.Results: of 1110 patients analyzed, 62 died from SUDEP. Main risk factors (adjusted odds ratio, 95% confidence intervals (CI)) were generalized seizure frequency (3.83, 95% CI: 1.88-7.75), nocturnal seizure (6.63, 95% CI: 4.12-11.19) and depression (2.57, 95% CI: 1.56-4.26) while the ability to alert of incoming seizure was a major protective factor of SUDEP (0.48, 95% CI: 0.27-0.83). After internal validation, a clinically usable score ranging from -1 to 8 was developed, with high discrimination capabilities (AUC= 0.85, 95% CI: 0.80-0.91). The threshold of 3 has good sensitivity (Se= 82.2%, 95% CI: 72.8-91.8) while keeping a good specificity (Sp= 81.4%, 95% CI: 79-83.7). Significance: those results outline the foremost importance of generalized and nocturnal seizure in the Occurrence of SUDEP. It is the first study to observe a protective role of the ability to alert of incoming seizure. The SUDEP-CARE score developed is promising. Further work should allow to definitely validate this clinically usable risk score.