To compare the prevalence of conduction disturbances and ventricular arrhythmias in cases of postoperative ventricular septal defect, 100 patients (50 with repair by right atriotomy, group 1; and 50 with repair by right ventriculotomy, group 2) underwent complete evaluation including an electrocardiogram (ECG) and 24 h ambulatory ECG monitoring. The two groups were comparable except for a shorter follow-up duration (7 ± 3 versus 12.4 ± 7 years) and a younger age at evaluation (12.4 ± 5 versus 16.9 ± 7 years) in group 1. Complete right bundle branch block was less frequent in group 1 than in group 2 (20% versus 50%, p < 0.05) but three of the four patients with complete atrioventricular (AV) block detected on ambulatory monitoring were in group 1. Six patients had significant supraventricular arrhythmias, all well tolerated. Ambulatory monitoring revealed significant ventricular arrhythmias (modified town grade 2 or higher) in 39 patients, with a lower prevalence in group 1 (30% versus 48%, p = 0.05). No correlation was found between prevalence of ventricular arrhythmias and right ventricular systolic pressure, cardiopulmonary bypass duration, presence of a synthetic patch, previous pulmonary artery banding, presence of complete right bundle branch block and cardiomegaly on chest X-ray film. Prevalence of ventricular arrhythmias increased with follow-up duration, age at evaluation and age at surgery. These were always well tolerated and did not warrant treatment. Thus, right atriotomy reduces the prevalence of right bundle branch block but does not prevent late AV block. Ventricular arrhythmias are frequent after surgical closure of ventricular septal defect whatever the surgical approach and their prevalence is not statistically different from that in postoperative tetralogy of Fallot. Only studies with a longer follow-up period will be able to assess the true significance of these arrhythmias and the true benefit of the transatrial approach. © 1990.