Safety data were reviewed from several controlled clinical trials of ibutilide, a new class III antiarrhythmic drug recently approved for the acute interruption of atrial fibrillation and flutter. Noncardiovascular adverse effects of ibutilide were similar in frequency to those with placebo. Cardiovascular adverse effects occurred in 24.9% of 586 ibutilide-treated patients as compared with 22.2% of 108 sotalol-treated patients, and 7.1% of 127 patients who received placebo. Polymorphous ventricular tachycardia, diagnosed as torsades de pointes, was more common with ibutilide than with placebo or sotalol treatment. It occurred in 4.3% of patients, including 1.7% whose torsades de pointes was sustained and required cardioversion. In the ibutilide group, 4.9% of patients had nonsustained monomorphic ventricular tachycardia compared with 3.7% of patients who received sotalol and 0.8% of patients who received placebo. All of the sustained arrhythmias except 1 occurred within 1 hour of the end of ibutilide infusion, and all were successfully terminated without sequelae. In a multiple logistic regression analysis, bradycardia, low body weight, and history of congestive heart failure were predictive of the occurrence of torsades de pointes. Hypotension, conduction block, bradycardia, and all other cardiovascular adverse effects all occurred at similar rates in the ibutilide- and placebo-treated groups. For patients who failed to convert while receiving ibutilide, there was no decrease in the efficiency of cardioversion, nor was there an increase in the mean energy requirements for subsequent electrical cardioversion. Analysis of a 3-month follow-up study showed that patients receiving ibutilide had similar outcomes compared with patients receiving placebo. One placebo-treated patient died. Other than torsades de pointes, ibutilide has a very good safety profile. Under the proper clinical conditions, this complication of ibutilide therapy can be rapidly diagnosed and effectively treated.