Inhibitor development is the major treatment complication in children with severe hemophilia A. It is not clear whether the risk of inhibitors is higher with recombinant factor VIII or with plasma-derived factor VIII. We used multivariate analysis to compare 2 cohorts of previously untreated patients (PUPs) with severe hemophilia A: 62 patients treated with the same brand of high-purity plasma-derived FVIII (pFVIII) containing von Willebrand factor (VWF) and 86 patients treated with full-length recombinant FVIII (rFVIII). In addition to the usual end points (all inhibitors, high inhibitors), we also examined a third end point (high inhibitors and/or immune tolerance induction). The risk of inhibitor development was higher in patients treated with rFVIII than in patients treated with pFVIII, regardless of other risk factors (F8 genotype; nonwhite origin; history of inhibitors in patients with a family history of hemophilia; age at first FVIII infusion). The adjusted relative risk (RRa) for inhibitor development with rFVIII versus pFVIII was 2.4 (all inhibitors), 2.6 (high inhibitors), and 3.2 (high inhibitors and/or immune tolerance induction), respectively, depending on the end point (above). The pathophysiology of this large effect must be understood in order to improve the characteristics of recombinant products and to reduce the incidence of inhibitors to FVIII.