From December 1971 to December 1989, 62 patients (pts) 42 males, 20 females, mean age 66 years (yr) 6 months (mth) (range 52-80) were operated upon for post-infarction ventricular septal defect (VSD), (anterior 34, inferior 28). Eight pts (13%), group (G) 1 presented with cardiogenic shock, 19 pts (30.5%), G2 with severe congestive heart failure (CHF); 31 pts (50%), G3 were stable with mild CHF and 4 pts (6.5%), G4 without CHF. Preoperative intra-aortic balloon pumping (IABP) was used in 49 pts (79%). One transplanted pt was excluded from this study. The VSD was closed from 1 day (d) to 5 mth (mean 13 d) after its occurrence. Hospital death (HD) occurred in 23 pts (37.7% +/- 6%). Of 44 incremental risk factors (RF) for HD studied, the preoperative status (PS) was the most significant (P less than 0.01). G1: 87% +/- 12%, G2: 42% +/- 12%, G3: 25.8% +/- 8%, G4: 0%. [table: see text] Non-survivors had a shorter mean delay between VSD occurrence and surgery than survivors: 5.6 +/- 3.7 d vs 18.2 +/- 30 d (P less than 0.05), but this delay was correlated to PS. The follow-up of the 38 early survivors ranges from 2 mth to 14 yr (mean 3 yr, 11 mth); 11 pts died between 45 d and 14 yr. No RF was identified for premature late death. HD included, the actuarial survival rate at 1, 5, 10 yr is: 57% (+/- 7%), 44% (+/- 8%), 30% (+/- 10%), respectively.