Performance of cardiac troponins within the HEART score in predicting major adverse cardiac events at the emergency department.

BACKGROUND: This study compared the performance of a single blood draw of high-sensitivity troponin T (hsTnT), high-sensitivity troponin I (hsTnI) and conventional troponin I (cTnI) within a modified HEART score for predicting 30-day MACE at Emergency Department (ED) presentation, and established local reference norms for all three assays by determining the cut-off point which yielded the highest sensitivity and negative predictive value for acute myocardial infarction and 30-day MACE. METHODS: This single-center prospective cohort study recruited chest pain patients at the ED, whose hsTnT, hsTnI and cTnI were taken on admission. Subjects were classified into low and non-low risk group according to their modified HEART score, with MACE as the primary endpoint. Receiver-operating characteristic (ROC) curves were generated, area under the curves (AUCs) were calculated; the performance characteristics were determined. RESULTS: The performance of modified HEART scores was comparable among the three assays for 30-day MACE (84.9-87.0% sensitivity, 95.6-96.0% NPV, 95%CI) and none of these had very high AUC and specificity (AUC 0.70-0.71, 53.7-56.7% specificity, 95% CI). The modified HEART score using a single blood draw of either hsTnT (3.9ng/L), hsTnI (0.9ng/L) or cTnI (0.0ng/L) at presentation yielded a sensitivity of 100% for 30-day MACE. CONCLUSION: The modified HEART score using a single blood draw of either hsTnT, hsTnI or cTnI was equally effective in risk-stratifying chest pain patients for safe discharge. The theoretical cut-off points yielding 100% sensitivity are potentially useful (when achieved) for safely discharging low risk patients with undifferentiated chest pain in the ED.

Application of ensemble machine learning algorithms on lifestyle factors and wearables for cardiovascular risk prediction.

This study looked at novel data sources for cardiovascular risk prediction including detailed lifestyle questionnaire and continuous blood pressure monitoring, using ensemble machine learning algorithms (MLAs). The reference conventional risk score compared against was the Framingham Risk Score (FRS). The outcome variables were low or high risk based on calcium score 0 or calcium score 100 and above. Ensemble MLAs were built based on naive bayes, random forest and support vector classifier for low risk and generalized linear regression, support vector regressor and stochastic gradient descent regressor for high risk categories. MLAs were trained on 600 Southeast Asians aged 21 to 69 years free of cardiovascular disease. All MLAs outperformed the FRS for low and high-risk categories. MLA based on lifestyle questionnaire only achieved AUC of 0.715 (95% CI 0.681, 0.750) and 0.710 (95% CI 0.653, 0.766) for low and high risk respectively. Combining all groups of risk factors (lifestyle survey questionnaires, clinical blood tests, 24-h ambulatory blood pressure and heart rate monitoring) along with feature selection, prediction of low and high CVD risk groups were further enhanced to 0.791 (95% CI 0.759, 0.822) and 0.790 (95% CI 0.745, 0.836). Besides conventional predictors, self-reported physical activity, average daily heart rate, awake blood pressure variability and percentage time in diastolic hypertension were important contributors to CVD risk classification.