Potentially malignant ventricular arrhythmias are common in chronic heart failure. The aggravation of such arrhythmias has many causes in these patients and cannot be predicted. Therefore, proarrhythmia due to PDE-inhibitors which increase cytosolic calcium levels by specific inhibition of the degradation of cyclo-AMP may be difficult to recognize. A retrospective analysis of 24-h Holter ECG was performed in 31 patients (NYHA classes III and IV) under long-term enoximone therapy. At baseline, 68% of the patients had ventricular couplets and nonsustained ventricular tachycardia. After a mean treatment period of 7 months, 10% of the patients showed a significant increase, 16% a significant decrease (greater than 90%) of ventricular couplets and salvos, and an additional 32% of the patients showed a significant decrease (greater than 70%) of single PVCs. The change of the arrhythmia profile was not related to the clinical course in these patients. Furthermore, 24-h Holter recordings were analyzed in a randomized long-term trial with captopril and enoximone that included 20 patients of NYHA class II. Despite comparable baseline findings, a reduction of cardiopulmonary exercise capacity was observed in patients treated with enoximone, but not with captopril. However, the arrhythmia profile was similar in both treatment groups. These findings suggest that, in most patients with advanced chronic heart failure, long-term enoximone therapy is not associated with an important increase of ventricular arrhythmias. According to the 24-h Holter findings of the European Enoximone Data Bank, proarrhythmia can be expected in 12% of all patients. Control of the arrhythmia profile, however, is mandatory, because the incidence of proarrhythmia cannot be predicted in the individual patient.