After completion of a placebo-controlled trial of intravenous immunoglobulin (IVIG) infection prophylaxis, patients were offered open label IVIG and optional participation in a follow-up study. The purpose of the follow-up study was to evaluate the IVIG effect in original placebo recipients and longevity of IVIG benefit in original IVIG recipients. Of 212 human immunodeficiency virus-infected children on study at trial closure, 148 (67 of 98 (68%) placebo and 81 of 114 (71%) IVIG patients) received open label IVIG for a mean of 16 months. When open label IVIG was begun, 45% were receiving trimethoprim-sulfamethoxazole prophylaxis for Pneumocystis carinii pneumonia (43% of placebo and 47% of IVIG patients) and 54% were receiving zidovudine (55% of placebo and 53% of IVIG patients). In patients who received placebo during the original study, the rate of serious bacterial infections was significantly lower after change to open label IVIG (estimated 15.8 fewer episodes/100 patient years; 95% confidence interval, 3.2 to 28.5; P = 0.014). Similar findings were observed for minor bacterial infections (estimated 61.2 fewer/100 patient years; 95% confidence interval, 29.2 to 93.3; P < 0.001) and hospitalizations (estimated 43.7 fewer/100 patient years; 95% confidence interval, 27.7 to 59.6; P < 0.001). Decreases were observed whether or not trimethoprim-sulfamethoxazole prophylaxis was being given at the time open label IVIG was begun. In patients who received IVIG during the original study, no significant difference was seen in infections or hospitalizations after change to open label IVIG.(ABSTRACT TRUNCATED AT 250 WORDS)