Summary. Total knee replacement (TKR) is a well recognized treatment for haemophilic arthropathy. Successful haemostasis can be achieved by bolus doses or continuous infusion (CI) using either recombinant (r) or plasma-derived (pd) factor IX (FIX). We retrospectively analysed our experience of factor replacement to cover TKR in haemophilia B patients and explored factors related to FIX use during surgery. Between 2000 and 2010, 13 primary TKRs were performed in 11 haemophilia B patients. Operations were performed by the same surgeon using standard techniques. Median age was 58 years (42–79). An adjusted CI protocol was used for 5 days followed by bolus doses. FIX:C was maintained at 100 IU dL−1 in the immediate postoperative period. There was no excess haemorrhage. There was no evidence of thrombosis or infection. All patients received mechanical thromboprophylaxis and only one chemical. CI was used in seven cases. Ten patients received pdFIX. Median hospital stay was 14 days (8–17). Median factor usage was 999 IU kg−1 (768–1248). During CI, factor consumption was 695 IU kg−1, 691 IU kg−1 and 495 IU kg−1 for BeneFix®, Replenine® and Haemonine, respectively. Clearance of both pdFIX and rFIX reduced during CI. All operations were uncomplicated. The decreased clearance in the CI setting reduced the amount of FIX required to maintain a therapeutic level. This reduction was greater with pdFIX and may be related to pharmacokinetic differences between pdFIX and rFIX. Given the excellent safety profile of the pdFIX products, CI of FIX and particularly pdFIX is safe, efficacious and convenient.