Accurate detection of drug-induced delayed ventricular repolarization with a suitable correction formula in Langendorff guinea pig heart.

The aims of this study were to determine a suitable method to correct the ventricular repolarization period against the RR interval in isolated perfused Langendorff guinea pig heart and to clarify the reliability of this model using several drugs. QT and RR intervals from an electrocardiogram and the epicardial monophasic action potential duration (MAP(90)) were measured. Two drugs clinically known to be QT-prolonging (E-4031, moxifloxacin) and two known to be non-QT-prolonging (verapamil, zatebradine) were used for the study. To determine a method of correcting the ventricular repolarization period against RR interval, heart rates were slowed with 0.3 mu M zatebradine, a specific bradycardiac agent, and then accelerated with atrial pacing to obtain a wide range of MAP(90)/RR relationships. An exponential rate-correction model elicited the most appropriate algorithm for the relationship among the four models tested. Based on linear regression analysis, the exponential showed superior dissociation of corrected MAP(90)s against RR intervals than generic Bazett's and Fridericia's formulae. E-4031 and moxifloxacin prolonged the corrected QT (QTc) intervals and MAP(90) under atrial pacing at a cycle length of 0.25 sec (MAP(90(pacing))) dose-dependently; verapamil and zatebradine failed to prolong them, indicating that the reliability of this model was excellent. MAP(90)(pacing) prolongation by moxifloxacin, the positive compound in the clinical "Thorough QT/QTc Study", was seen at around QTc-prolonging concentrations in clinic, suggesting that the sensitivity would be appropriate for QT evaluation. We therefore concluded that the isolated guinea pig heart model is sufficiently sensitive and useful for assessing the potential QT prolongation of drugs.