The primary aim of this study was to investigate the possible relationship between coagulation factor level and bleeding frequency during prophylactic treatment of haemophilia after stratification of the patients according to joint scores. The secondary aim was to obtain a systematic overview of the doses of coagulation factors prescribed for prophylaxis at the Malmo haemophilia treatment centre during a 6-year period. A retrospective survey of medical records for the years 1997-2002 and pharmacokinetic study results from the 1990s was complemented by collection of blood samples for coagulation factor assay when needed. Information on the dosing and plasma levels of factor VIII or factor IX, joint scores and incidence of bleedings (joint bleeds and 'other bleeds') was compiled. The patients were stratified by age (0-6, 7-12, 13-18, 19-36 and >36 years) and joint score (0, 1-6 and >6). Individual pharmacokinetic parameters of plasma coagulation factor activities (FVIII:C and FIX:C) were estimated. Trough levels during the treatment were calculated, as well as the number of hours per week of treatment during which plasma FVIII:C/FIX:C fell below a 1, 2 or 3% target level. Fifty-one patients with haemophilia A (two moderate, 49 severe) and 13 with haemophilia B (all severe) were included, yielding data for 364 patient-years of treatment. There was a wide range of dosing schedules, the most common ones being three times a week or every other day for FVIII and twice a week or every third day for FIX. The overall relationship between FVIII:C/FIX:C levels and incidence of joint bleeding was very weak, even after stratification of the patients according to joint score. There was no relationship between coagulation factor level and incidence of other bleeds. In this cohort of patients on high-dose prophylactic treatment, dosing was based more on clinical outcome in terms of bleeding frequency than on the aim to maintain a 1% target level of FVIII:C/FIX:C. Some patients did not bleed in spite of a trough level of <1% and others did in spite of trough levels >3%. The practical implication of our findings is that dosing in prophylactic treatment of haemophilia should be individualized. Thus, proposed standard regimens should be implemented only after careful clinical consideration, with a high readiness for re-assessment and individual dose tailoring.