A Clinical Prediction Model for Sudden Cardiac Arrest Presenting as Pulseless Electrical Activity

BackgroundThe incidence of sudden cardiac arrest (SCA) manifesting as pulseless electrical activity (PEA) has increased, and survival remains extremely low. Methods for early identification and management of high-risk individuals are needed, but no clinical risk scores currently exist to predict PEA-SCA. Our objective was to develop and validate a clinical prediction model for PEA-SCA. MethodsFrom an ongoing prospective, population-based study of SCA in Portland, Oregon (catchment pop. {approx}1 M, 2002-2020), we identified PEA-SCA adults. Lifetime clinical records were compared with those of a control group with >50% prevalence of significant coronary disease. Prediction models were constructed using backwards stepwise logistic regression in a training dataset (67%) and evaluated in a validation dataset (33%). Model discrimination was assessed using receiver operating characteristic curves (C statistic). External validation was performed in a geographically distinct population in Ventura County, California (population {approx}850,000, 2015-2022). ResultsThe final clinical algorithm (PEA-Risk) incorporating 12 clinical, electrocardiogram and medication variables demonstrated strong discrimination in the training dataset (C statistic = 0.860 [95% CI: 0.838-0.881]) and remained robust in internal (C statistic = 0.832 [95% CI: 0.800-0.865]) and external validation datasets (C statistic = 0.704 [95% CI: 0.665-0.743]). ConclusionsWe developed and externally validated a clinical algorithm for predicting PEA-SCA. Given the low rates of successful resuscitation after PEA arrest, this risk prediction tool may enable earlier identification and prevention of PEA-SCA. Clinical PerspectiveO_ST_ABSWhat is knownC_ST_ABSO_LIThe proportion of SCA presenting as pulseless electrical activity (PEA) is increasing, and survival from these events remains extremely low. C_LIO_LIThe are no available methods for clinical risk prediction of these events. C_LI What the study addsO_LIThe present study constructs and replicates a risk score for prediction of SCA manifesting with PEA using widely available clinical and noninvasive markers. C_LIO_LIThese findings have implications for developing prevention and management strategies for individuals at high risk of PEA-SCA. C_LI