Abstract Body: Background :
Sudden cardiac arrest (SCA) in young adults, with survival rates below 15%, represents a major public health concern. Most cases occur in individuals without identifiable cardiac disease. Current prevention strategies, focused on cardiovascular risk factors, may therefore overlook key non-cardiac contributors.

Methods :
We conducted a population-based case–control study linking the exhaustive Paris Cardiac Arrest Center registry (2011–2020) to national electronic health records, including 1,061 SCA cases and 3,183 matched controls aged ≤40 years, and 21,199 cases with 63,597 controls aged >40 years. All prospectively collected medical diagnoses and prescriptions from the preceding five years were analyzed using an interpretable machine learning model (Explainable Boosting Machine). Age-specific risk markers were ranked by their relative contribution to model performance, and findings were externally validated in the Seattle–King County registry (2013–2021).

Results :
In young adults, neuropsychiatric drugs accounted for three of the top five predictors of SCA (antiepileptics, antipsychotics, and anxiolytics), while beta-blockers were the only cardiovascular marker among the top ten. In older adults, cardiovascular diseases and related treatments predominated. Model performance was robust (AUC = 0.74 [95% CI 0.70–0.78] in ≤40 years; 0.80 [0.79–0.81] in >40 years). The overrepresentation of neuropsychiatric factors in the young was confirmed in the U.S. validation cohort.

Conclusion :
Neuropsychiatric disorders and treatments emerged as major risk markers for SCA in adults under 40, underscoring the need to integrate them with cardiovascular factors and to promote interdisciplinary prevention strategies.