From January 1980 through September 1990, 130 children underwent surgical closure of isolated multiple ventricular septal defects (mean age 14 ± 18 months, mean weight 7.0 ± 4.4 kg). Sixty-one were less than 1 year of age. Sixty-one children had pulmonary protection, 51 had pulmonary artery banding, and 10 had pulmonary valve stenosis. All other patients had severe pulmonary hypertension (mean systolic pressure 75.7 ± 20.5 mm Hg and already disabling heart failure (New York Heart Association classes III and IV). The surgical management was based on the location of the defects and the ventricular dominance that were assessed preoperatively and intraoperatively. Midtrabecular ventricular septal defects were always centered by the moderator band and were therefore divided into low trabecular, midtrabecular, and high trabecular defects. The perimembranous septum was involved in 102 patients, the trabecular in 121, the inlet septum in 12, and the infundibular septum in 9. Fifty patients had the 'Swiss cheese' form of the lesion. Closure of the ventricular septal defects included Dacron patch and mattress sutures. They were always first approached through a right atriotomy, which was sufficient for complete repair in 82 patients. In midtrabecular ventricular septal defects, section of the moderator band (n = 24) allowed closure of all the defects with a single Dacron patch. In 48 patients a right atriotomy and a right (n = 32) or left (n = 14) (particularly for low trabecular ventricular septal defects) or both right and left (n = 2) ventriculotomies were necessary to secure the repair. The hospital mortality rate was 7.7% (10 patients). The causes of deaths were residual ventricular septal defect (n = 5), pulmonary hypertension (n = 2), hypoplastic right ventricle (n = 1) and left ventricle (n = 1), and myocardial infarction (n = 1). Among eighteen survivors with residual ventricular septal defect, six were reoperated on; there were two deaths. A permanent pacemaker was necessary in four patients. Low trabecular ventricular septal defects and left ventriculotomy were significant risk factors for morbidity (death, residual ventricular septal defect), p < 0.01. At 7 years of follow-up, 90% of survivors were in New York Heart Association class I. Actuarial survival and freedom from reoperation at 7 years were 89.6% and 87.5%, respectively.