The use of Doppler color flow imaging and axial contrast angiography in the preoperative detection of additional ventricular septal defects (in the setting of a known large defect) were compared in a prospective fashion. One hundred seventy-nine infants with two ventricles (each of at least normal size) and a large, nonrestrictive ventricular septal defect underwent reparative surgery before 2 years of age. The reference standard for the presence of additional defects was intraoperative verification or (in cases in which the surgeon did not visualize any additional defect) subsequent identification at postoperative angiography, postoperative color Doppler examination or reoperation. Only six patients (3%) had additional ventricular septal defects confirmed at the time of repair; an additional five (3%) had defects found only postoperatively. The negative predictive value of Doppler color flow imaging and angiography was 0.95 (168 of 176) and 0.97 (168 of 174), respectively. The sensitivity was 0.27 (3 of 11) and 0.45 (5 of 11), respectively. For certain malformations with a very low prevalence of additional muscular defects (such as perimembranous ventricular septal defect with normally aligned great arteries), a clinical trial of reparative surgery without prior invasive study appears reasonable.