During factor VIII prophylaxis for severe hemophilia A, bleeding risk increases with time when factor VIII activity is below 1%. Maintaining trough activity above 1% does not protect all patients from bleeding. The relationship between factor VIII activity during prophylaxis and bleeding risk has not been thoroughly studied. We investigated factor VIII activity and annualized bleeding rate for spontaneous bleeds during prophylaxis. A population pharmacokinetic model derived from three clinical trials was combined with dosing data and bleed information from patient diaries. Each patients' time on prophylaxis was divided into five categories of predicted activity (0-1%, >1-5%, >5-15%, >15-50%, and >50%). Exposure time, mean factor VIII activity, and bleed number (from patient diaries) were calculated for each activity category, and annualized bleeding rates estimated using negative binomial regression and a parametric model. Relationships between these bleeding rates and factor VIII activity were evaluated by trial phase (pivotal vs. extension) and age (adults/adolescents ≥12 years vs. children 0-<12 years). In total (N=187; 815 patient-years' exposure), factor VIII activity was predicted to reach >1% for 85.64% of the time. Annualized bleeding rate decreased as factor VIII activity increased in each trial phase and age group. However, for a given activity level, bleeding rate differed substantially by trial phase, and age. This suggests that bleeding risk can change over time and is influenced by factors independent of factor VIII pharmacokinetics and trough levels. Target trough and prophylactic regimen should consider patient age, joint disease activity, and other bleeding risk determinants.