Population: boys with severe HA (FVIII activity <0.01 IU/mL) first treated with rFVIII and treated with Products A, C, D, or E within the first 75 EDs (n = 287). Products B and F were used only in 10 and 10 boys and 331 and 483 EDs, respectively, so their effect on inhibitor development was not studied. Main studied factor: type of rFVIII product received during the first 75 EDs (time-varying factor). Statistical method: Cox proportional hazards model with ED as observational time. aHR took into account the following cofactors: 4 fixed factors (F8 gene defect, family history of hemophilia and inhibitor, ethnic origin, and age at first infusion of rFVIII) and 7 time-varying factors (calendar period, regular prophylaxis initiation, interval between EDs calculated over the last 5 EDs, mean dose of rFVIII product calculated over the last 5 EDs, history of peak treatment episodes ≥5 consecutive EDs, history of peak treatment episodes ≥10 consecutive EDs, and history of severe bleeding episodes).

P value for global test.