This study prospectively evaluated the clinical efficacy, the incidence of torsades de pointes, and the presumable risk factors for torsades de pointes in patients treated with d,l-sotalol for sustained ventricular tachyarrhythmias. Eighty-one consecutive patients (54 with coronary artery disease, and 20 with dilated cardiomyopathy) with inducible sustained ventricular tachycardia or ventricular fibrillation received oral d,l-sotalol to prevent induction of the ventricular tachyarrhythmia. During oral loading with d,l-sotalol, continuous electrocardiographic (ECG) monitoring was performed. Those patients in whom d,l-sotalol prevented induction of ventricular tachycardia or ventricular fibrillation were discharged with the drug and followed up on an outpatient basis for 21 +/- 18 months. Induction of the ventricular tachyarrhythmia was prevented by oral d,l-sotalol in 35 (43%) patients; the ventricular tachyarrhythmia remained inducible in 40 (49%) patients; and two (2.5%) patients did not tolerate even 40 mg of d,l-sotalol once daily. Four (5%) patients had from torsades de pointes during the initial oral treatment with d,l-sotalol. Neither ECG [sinus-cycle length (SCL), QT or QTc interval, or U wave] nor clinical parameters identified patients at risk for torsades de pointes. However, the oral dose of d,l-sotalol was significantly lower in patients with torsades de pointes (200 +/- 46 vs. 328 +/- 53 mg/day; p = 0.0017). Risk factors associated with the development of torsades de pointes were the appearance of an U wave (p = 0.049), female gender (p = 0.015), and significant dose-corrected changes of SCL, QT interval, and QTc interval (p < 0.05). During follow-up, seven (20%) patients had a nonfatal ventricular tachycardia recurrence, and two (6%) patients died suddenly. One female patient with stable cardiac disease had recurrent torsades de pointes after 2 years of successful treatment with d,l-sotalol. Torsades de pointes occurred early during treatment even with low doses of oral d,l-sotalol. Pronounced changes in the surface ECG (cycle length, QT, and QTc) in relation to the dose of oral d,l-sotalol might identify a subgroup of patients with an increased risk for torsades de pointes. Other ECG parameters before the application of d,l-sotalol did not identify patients at increased risk for torsades de pointes. Recurrence rates of ventricular tachyarrhythmias are high despite complete suppression of the arrhythmia during programmed stimulation. Therefore programmed electrical stimulation in the case of d,l-sotalol seems to be of limited prognostic value.