Prediction of Imminent Sudden Cardiac Arrest Using a Combination of Warning Symptoms and Clinical Features

IntroductionAt least 50% of individuals who suffer sudden cardiac arrest (SCA) experience warning symptoms before their SCA. We have previously reported chest pain and dyspnea as the most common and potentially predictive symptoms. Combining with clinical features could improve sensitivity and specificity for prediction of imminent SCA (ISCA). HypothesisA combination of warning symptoms and clinical profiles can predict ISCA. MethodsFrom two community-based studies of SCA in Oregon and California, we conducted a case-control study. Cases (n=364) were survivors of SCA who had experienced warning symptoms, and control subjects (n=313) were individuals who notified emergency medical services (EMS) for similar symptoms but did not have SCA. Symptom data were obtained from interviews with study subjects and from EMS pre-hospital care records. We constructed classification and regression tree (CART) models for major symptom categories to identify clinical predictors of ISCA. We used the area under the receiver operating characteristic curve (AUC) and 5-fold cross-validation to assess model performance and stability. ResultsHeart failure (HF) and/or coronary artery disease (CAD) were predictors of ISCA and displayed important sex differences. For example, among individuals presenting with only chest pain, male sex, particularly males with HF, was an important predictor of ISCA (AUC = 0.813). Among individuals with only dyspnea, CAD and HF were important predictors (AUC = 0.745) with no sex differences identified. The 5-fold cross-validation produced consistent results. ConclusionsCombinations of warning symptoms and clinical features distinguished individuals with SCA from individuals without SCA with good accuracy (AUCs 0.728 - 0.813).