In July 1990, the Recombinate Study Group initiated a prospective, open-labeled investigation of recombinant factor VIII (r-FVIII) to assess its safety and efficacy and to characterize the natural history of inhibitor development in previously untreated patients (PUPs) with hemophilia A. All study subjects have severe FVIII deficiency (baseline FVIII level < or = 2% of normal) and no history of blood product exposure before study entry. Following the first r-FVIII infusion, plasma was screened for inhibitors once every 3 months, and plasma recovery of r-FVIII at 30 minutes and 24 hours postinfusion was assayed at least once every 6 months. As of May 1993, 73 of 79 patients originally enrolled in the trial continue to participate. The median number of r-FVIII exposure-days for the 71 subjects who have received at least one r-FVIII infusion is 11. A total of 1,785 infusions have been administered to treat 810 bleeding events. Ninety-two percent of bleeding events responded as anticipated to one or two infusions. Two, nonrecurring, acute adverse reactions occurred coincident with r-FVIII infusion, one of which was unrelated and the other, possibly related to the infusion. Seventeen (23.9%) subjects have developed inhibitors: five with peak titers more than 10 Bethesda units (BU) and 12 with peak titers < or = 10 BU (range, 0.5 to 10). Survival analysis showed that the probability of remaining inhibitor-free in this group of patients with severe hemophilia A is 88.4% after 8, 73.6% after 10, and 61.6% after 25 r-FVIII exposure-days. Inhibitors disappeared in five (29.4%) subjects on retesting 2 to 16 months after the last positive inhibitor assay. r-FVIII is safe and effective in the treatment of hemophilia A-related bleeding. To date, the inhibitor risk associated with its use is comparable to that in patients treated with plasma-derived concentrates. The majority of inhibitors identified are low in titer and do not preclude continued on-demand therapy with r-FVIII.