Recombinant human FIX (rFIX) was evaluated in 28 subjects, including 26 with mild, moderate, or severe haemophilia B and two haemophilia B carriers undergoing 36 surgical procedures. Preoperative rFIX dose was highly correlated with postinfusion FIX activity, r=0.61, P=0.0158. Peri- and post-operative estimated blood loss was similar to that expected in non-haemophilic individuals, and haemostasis was rated as excellent or good in 34 of 35 (97.1%) of the operative procedures. Transfusions were required in five of 36 (13.9%) procedures, including one liver transplantation, and three knee and one hip arthroplasties. Adverse events occurred in 15 of 28 (53.6%) subjects, but there were no perioperative haemorrhages, thromboembolic events, coagulation activation, viral transmission, or inhibitor formation. A transient low-responding FIX inhibitor developed in one subject preoperatively, but required no change in treatment and resolved 15 months later. Thus, rFIX was found to be safe and effective in achieving haemostasis in subjects with FIX deficiency undergoing surgery.