Abstract 4369161: Use of AI in young cases of sudden cardiac arrest: Implications for treatments and neuropsychiatric pathologies.

Abstract

Background: Sudden Cardiac Arrest (SCA) in young adults is a significant public health issue, with survival rates often less than 15%. In young adults, most SCA arise from patients for whom there are no well-identified risk factors. Strengthening our understanding of SCA in young adults is therefore essential for identifying those at risk and developing interventions to reduce its incidence. Methods: We conducted a case-control study using electronic health records (EHRs) from 89,040 individuals in one city and its suburbs (2011–2020). We included 4,244 prospectively collected patients aged ≤ 40 years and 84,796 aged > 40 years, with a 1:3 case-controls matching on age, sex, and residence. We trained Explainable Boosting Machine (EBM) models on medical diagnoses and prescriptions from the five years preceding SCA. Model performance was assessed using AUC, sensitivity, specificity, and feature importance, stratified by age groups. Results: For patients aged ≤40 years, neuropsychiatric factors, including antipsychotics, anxiolytics, and epilepsy emerged as key predictors of SCA, whereas cardiovascular factors were the primary predictors in the older group. The younger cohort’s model achieved an AUC of 0.74 (95% CI: 0.70–0.78). The older cohort’s model had an AUC of 0.80 (95% CI: 0.79–0.81). Both models demonstrated robust calibration, and sensitivity analyses confirmed consistent findings across various feature subsets and hyperparameter settings. Conclusion: These findings underscore the crucial role of neuropsychiatric factors in risk stratification among younger individuals. By highlighting risk markers such as epilepsy and the use of antipsychotics or anxiolytics, our results call for greater interdisciplinary collaboration to expand current SCA conception beyond the strict cardiac domain. Future research should investigate non-cardiac markers to refine SCA risk stratification and explore non-cardiac contributors to better understand young and undiagnosed cases.