An in vitro model for assessment of drug-induced torsade de pointes arrhythmia : effects of haloperidol and dofetilide on potential duration, repolarization inhomogeneities, and torsade de pointes arrhythmia

Abstract

Torsade de pointes (TdP) is a serious side effect of many drugs. We aimed to establish an in vitro TdP model for drug testing, which includes typical risk factors, such as female gender, hypokalemia, low magnesium levels, and bradycardia. Isolated, spontaneously beating rabbit hearts (female White New Zealand rabbits) were perfused according to the Langendorff technique and submitted to conditions known as risk factors for TdP, i.e., [K(+)](e) = 2.5 mM and [Mg(++)](e) = 0.5 mM, with 10(-8) M noradrenaline and 10(-7) M carbachol. Thereafter, cumulative concentration-response curves for haloperidol (10, 100, 200, 1,000, and 2,000 nM) and dofetilide (1, 10, 20, 100, and 200 nM) were performed, while cardiac activation and repolarization was measured at 256 ventricular sites (unipolar extracellular potentials). We found in three of six hearts under haloperidol TdP arrhythmias in supratherapeutic concentrations > or =100 nM. Dofetilide also induced TdP (three of seven) in concentrations > or =20 nM. The TdP showed a complex pattern being initiated in one region by an early R-on-T ventricular extrasystole, when in the other regions high activation-recovery interval (ARI) dispersion occurred, then spreading in complex beat-to-beat changing patterns until self-termination. Dofetilide and haloperidol significantly prolonged ARI and QTc. Haloperidol significantly increased dispersion predominantly at the right wall and prolonged basic cycle length. Dofetilide also increased dispersion and slowed basic cycle length. Haloperidol (> or =100 nM) and dofetilide (> or =20 nM) can induce TdP by prolongation of ARI, slowing of heart rate, and increasing repolarization inhomogeneities. The linear combination of the independent variables QTc, BCL and dispersion could highly significantly predict TaP (adjusted R(2): 0.896, p < 0.001) The model seems suitable to identify a pharmacological risk for TdP in vitro within a limited number of animals.