Risk assessment of future myocardial infarction from automated serial ECG analysis

From 1980 to 1990 a cohort of 7,542 employees took part in the MARISK study. This study was designed in order to explore the utility of computerized electrocardiographic (ECG) measurements for predicting future myocardial infarction. After an initial period of annual examinations, follow-up information was collected by mailed questionnaires and obtained from 89.3% of all study participants with a median follow-up time of more than 7 years. Eighty-three participants suffered from myocardial infarction during the course of the study. Multivariate logistic regression analysis revealed that automated ECG measurements contribute significant additional information to risk estimates obtained from classical risk factors. Serial ECG analysis yields stronger risk indicators than single measurements. Results are represented via receiver operating characteristic (ROC) curves and regression coefficients, which can be used in independent applications. The results of this study indicate that it is useful to include automated serial ECG analysis in cardiovascular screening programs. It still has to be evaluated whether successful preventive measures may be based on such screening programs.