Summary. This paper emphasizes the importance of distinguishing between the prevalence, incidence and cumulative incidence of inhibitors in haemophilia A. Incidence and cumulative incidence data will include patients with transient inhibitors or whose inhibitors have been eliminated by treatment. As these will not be included in prevalence data, prevalence studies will tend to give rise to lower figures than incidence studies. As a result, the most accurate estimates of the true risk of inhibitor development comes from prospective studies of newly diagnosed haemophiliacs who are tested regularly for the presence of inhibitors. This paper reports a systematic review of the best available evidence relating to the epidemiology of inhibitors in haemophilia A. Cohort studies, registry data reporting incidence or prevalence of inhibitors in patients with haemophilia A, and prospective studies of factor VIII (FVIII) in the treatment of previously untreated patients which reported the development of inhibitors as an outcome, were included in the review. The overall prevalence of inhibitors in unselected haemophiliac populations was found to be 5–7%. The cumulative risk of inhibitor development varied (0–39%). Incidence and prevalence were substantially higher in patients with severe haemophilia. Studies of patients using a single plasma-derived FVIII (pdFVIII) preparation reported lower inhibitor incidence than those using multiple pdFVIII preparations or single recombinant FVIII preparations. Incidence data should be used to estimate the likely demand for treatments aimed at eliminating inhibitors, whereas the best estimates of the overall burden to the National Health Service (NHS) of treating bleeding episodes in patients with continuing inhibitors will come from prevalence studies.