We undertook a prospective randomized trial to examine whether an intravenous cytomegalovirus (CMV) immune globulin would prevent primary CMV disease in renal-transplant recipients. Fifty-nine CMV-seronegative patients who received kidneys from donors who had antibodies against CMV were assigned to receive either intravenous CMV immune globulin or no treatment. The immune globulin was administered in multiple doses over the first four months after transplantation. The incidence of virologically confirmed CMV-associated syndromes was reduced from 60 percent in controls to 21 percent in recipients of CMV immune globulin (P less than 0.01). Fungal or parasitic superinfections were not seen in globulin recipients but occurred in 20 percent of controls (P = 0.05). Only 4 percent of globulin recipients had marked leukopenia (reflecting serious CMV disease), as compared with 37 percent of the controls (P less than 0.01). There was a concomitant but not statistically significant reduction in the incidence of CMV pneumonia (17 percent of controls as compared with 4 percent of globulin recipients). A significant reduction in serious CMV-associated disease was observed even when patients were stratified according to therapy for transplant rejection (P = 0.04). We observed no effect of immune globulin on rates of viral isolation or seroconversion, suggesting that treated patients often harbored the virus but that clinically evident disease was much less likely to develop in them. We conclude that CMV immune globulin provides effective prophylaxis in renal-transplant recipients at risk for primary CMV disease.