Disturbances of cardiac wavelength and repolarization precede Torsade de Pointes and ventricular fibrillation in Langendorff perfused rabbit hearts.

Drug-induced proarrhythmia is strongly focused on Torsade de Pointes (TdP), and QT prolongation is commonly used as a surrogate for TdP. This surrogate may be of limited value, as it produces false positives and false negatives. Cardiac electrophysiological parameters were obtained from monophasic action potentials recorded from Langendorff perfused female rabbit hearts. In 3939 experiments 179 VFs and 124 TdPs occurred during testing of various chemicals. TdP occurred with prolongation of action potential duration (APD), but also with shortening. Thus, APD/QT prolongation alone cannot predict TdP. Drug-induced abnormalities, associated with TdP or VF, were identified. The 11 most predictive abnormalities were combined in a proarrhythmic score, proportional to the incidence and amplitude of the abnormalities. This score significantly increases starting at concentrations ∼100-fold below that at which TdP/VF developed (P < 0.001) and preceded all TdP and ∼98% of VF. This derived proarrhythmic score was then tested in 451 additional experiments (not used to develop the score) and it was found to predict all 15 TdPs and 18 of the 19 VFs. In conclusion, the derived proarrhythmic score has fewer false positives and false negative than the widely used surrogate (QT-prolongation). As such, it may help in preclinical development of safer medications (better rejection of false negatives) and allowing development of safe beneficial medications with QT prolongation (less rejection of false positives). Recognition of disturbances of the score in the clinical ECG may yield safer use of some medications in vulnerable patients.