The reliable assessment of proarrhythmic side effects of drugs in development is essential but remains elusive. Recently, the short-term beat-to-beat variability of the QT interval (QTV) has been suggested as a novel parameter to predict ventricular arrhythmias due to repolarisation abnormalities. Here we studied the incidence of Torsades de Pointes (TdP) arrhythmias and its correlation with the frequency corrected QT (QTc) interval and the QTV following combined pharmacological block of IKs and IKr, IK1 and IKr, and IK1 and IKs in anaesthetised rabbits. ECGs were recorded before and after the administration of the IKr blocker, dofetilide, the IKs blocker, HMR-1556, and the IK1 blocker, BaCl2, and their combination, intravenously. Dofetilide significantly increased the QTc interval and induced TdP in 28% of animals; HMR-1556 or BaCl2 alone had little effect on QTc and did not cause TdP. The combination of dofetilide and HMR-1556, BaCl2 and dofetilide significantly increased the QTc interval, the QTV and led to a high incidence of TdP (82% and 63%). In conclusion, the QTV and the incidence of TdP increased markedly following the combined pharmacological block of IKs and IKr as well as after combined block of IK1 and IKr. Our results suggest that QTV may be a better predictor of subsequent TdP development than the prolongation of the QTc interval alone. The current model, where repolarisation reserve is impaired by pharmacological block of one or more potassium currents, may be utilised for more reliable testing of drugs in development for their possible proarrhythmic side effects.