Background Data on the optimal location of the ECG leads for the diagnosis of drug-induced long QT syndrome (diLQTS) with Torsades de Pointes (TdP) are lacking. Methods We systematically reviewed the literature for ECGs of patients with diLQTS and subsequent TdP. We assessed T-wave morphology in each lead and measured the longest QT interval in the limb and chest leads in a standardized fashion. Results Of 84 patients, 61.9% were female and mean age was 58.8 years. QTc was significantly longer in chest versus limb leads (mean (standard deviation) 671 (102) vs 655 (97) ms, p=0.02). Using only limb leads for QT interpretation, 18 (21.4%) ECGs were non-interpretable: 10 (11.9%) due to too flat T-waves, 7 (8.3%) due to frequent, early PVCs and 1 (1.2%) due to too low ECG recording quality. In the chest leads, ECGs were non-interpretable in 9 (10.7%) patients: 6 (7.1%) due to frequent, early PVCs, 1 (1.2%) due to insufficient ECG quality, 2 (2.4%) due to missing chest leads but none due to too flat T-waves. The most common T-wave morphologies in the limb leads were flat (51.0%), broad (14.3%) and late peaking (12.6%) T-waves. Corresponding chest lead morphologies were inverted (35.5%), flat (19.6%) and biphasic (15.2%) T-waves. Conclusions Our results indicate that QT evaluation by limb leads only underestimates the incidence of diLQTS experiencing TdP and favors the screening using both limb and chest lead ECG.