The prognostic accuracy of different QT interval measures

Background: The QT intervals accuracy for predicting arrhythmic death varies between studies, possibly due to differences in the selection of the lead used for measurement of the QT interval. The purpose of this study was to analyze the prognostic accuracy of all known ways to select the lead. Methods and Results: Three institutions that used different methods for measuring QT intervals provided their QT databases. They included more than 3500 twelve-lead surface ECGs. The data represented low- and high-risk patients of the normal population (survivors vs dead from cardiovascular causes), acute myocardial infarction (survivors versus death from all causes) and remote myocardial infarction (with vs without a history of ventricular arrhythmia). The prognostic accuracy was defined as the area under the Receiver Operator Curve (ROC-area). The most accurate standard leads were I and aVL and the least accurate was AVR. The most accurate precordial lead was V,. The prognostic accuracy of the longest QT interval was higher than for any standard lead. The prognostic accuracy of the mean of the three longest QT intervals was equal to or slightly lower than for the longest QT interval. Conclusions: The highest prognostic accuracy is obtained with the longest QT interval. The accuracies of the lead selection methods are so different that it can explain a substantial part of the differences between otherwise similar studies in the literature. We recommend the use of the mean value of the three longest QT intervals.