Congestive heart failure (CHF) is often associated with atrial fibrillation. The safety of many antiarrhythmic drugs in CHF is limited by proarrhythmic effects. We aimed to assess the safety of a novel atrial-selective K(+)-channel blocker AVE0118 in CHF compared to a selective (dofetilide) and a non-selective IKr blocker (terfenadine). For the induction of CHF, rabbits (n = 12) underwent rapid right ventricular pacing (330-380 bpm for 30 days). AVE0118 (1 mg/kg) dofetilide (0.02 mg/kg) and terfenadine (2 mg/kg) were administered in baseline (BL) and CHF. A six-lead ECG was continuously recorded digitally for 30 min after each drug administration. At BL, dofetilide and terfenadine significantly prolonged QTc interval (218 +/- 30 ms vs 155 +/- 8 ms, p = 0.001 and 178 +/- 23 ms vs. 153 +/- 12 ms, p = 0.01, respectively) while QTc intervals were constant after administration of AVE0118 (p = n.s.). In CHF, dofetilide and terfenadine caused torsades de pointes and symptomatic bradycardia, respectively, and prolonged QTc interval (178 +/- 30 ms vs. 153 +/- 14 ms, p = 0.02 and 157 +/- 7 ms vs. 147 +/- 10 ms, p = 0.02, respectively) even at reduced dosages, whereas no QTc-prolongation or arrhythmia was observed after full-dose administration of AVE0118. In conclusion, atrial-selective K(+)-channel blockade by AVE0118 appears safe in experimental CHF.