Background: The electrocardiographic QT interval is used to identify drugs with potential harmful effects on cardiac repolarization in drug trials, but the variability of the measurement can mask drug-induced ECG changes. The use of complementary electrocardiographic indices of abnormal repolarization is therefore warranted. Most drugs associated with risk are inhibitors of the rapidly activating delayed rectifier potassium current (I(Kr)). This current is also inhibited in the congenital type 2 form of the long QT syndrome (LQT2). It is therefore possible that electrocardiographic LQT2 patterns might be used to identify abnormal repolarization patterns induced by drugs.
Objective: To develop distinct T-wave morphology parameters typical of LQT2 and investigate their use as a composite measure for identification of d,l-sotalol (sotalol)-induced changes in T-wave morphology.
Methods: Three independent study groups were included: a group of 917 healthy subjects and a group of 30 LQT2 carriers were used for the development of T-wave morphology measures. The computerized measure for T-wave morphology (morphology combination score, MCS) was based on asymmetry, flatness and notching, which are typical ECG patterns in LQT2. Blinded to labels, the new morphology measures were tested in a third group of 39 healthy subjects receiving sotalol. Over 3 days the sotalol group received 0, 160 and 320 mg doses, respectively, and a 12-lead Holter ECG was recorded for 22.5 hours each day. Drug-induced prolongation of the heart rate corrected QT interval (QTcF) was compared with changes in the computerized measure for T-wave morphology. Effect sizes for QTcF and MCS were calculated at the time of maximum plasma concentrations and for maximum change from baseline. Accuracy for separating baseline from sotalol recordings was evaluated by area under the receiver operating characteristic curves (AUCs) using all recordings from the time immediately post-dose to maximum change.
Results: MCS separated baseline recordings from sotalol treatment with higher accuracy than QTcF for the 160 mg dose: (AUC) 84% versus 72% and for the 320 mg dose: (AUC) 94% versus 87%, p < 0.001. At maximum serum-plasma concentrations and at maximum individual change from baseline, the effect sizes for QTcF were less than half the effect sizes for MCS, p < 0.001. Effect sizes at peak changes of the mean were up to 3-fold higher for MCS compared with QTcF, p < 0.001. In subjects receiving sotalol, T-wave morphology reached similarity to LQT2, whereas QTcF did not.
Conclusion: Distinct ECG patterns in LQT2 carriers effectively quantified repolarization changes induced by sotalol. Further studies are needed to validate whether this measure has general validity for the identification of drug-induced disturbed repolarization.