Objectives. Effective analgesia after total knee arthroplasty (TKA) improves patient satisfaction, mobility and expedites discharge. This study assessed whether continuous femoral nerve infusion (CFNI) was superior to a single-shot femoral nerve block in primary TKA surgery completed under subarachnoid blockade including morphine. Methods. We performed an adequately powered, prospective, randomised, placebo-controlled trial comparing CFNI of 0.125% bupivacaine versus normal saline following a single-shot femoral nerve block and subarachnoid anaesthesia with intrathecal morphine for primary TKA. Patients were randomised to either treatment (CFNI 0 ml to 10 ml/h 0.125% bupivacaine) or placebo (CFNI 0 ml to 10 ml/h normal saline). Both groups received a single-shot femoral nerve block (0.25% 20 ml bupivacaine) prior to placement of femoral nerve catheter and subarachnoid anaesthesia with intrathecal morphine. All patients had a standardised analgesic protocol. The primary end point was post-operative visual analogue scale (VAS) pain score over 72 hours post-surgery. Secondary outcomes were morphine equivalent dose, range of movement, side effects, and length of stay. Results. A total of 86 patients were recruited. Treatment and placebo groups were comparable. No significant difference was found in VAS pain scores, total morphine equivalent requirements, side effects, range of movement, motor block, or length of hospital stay. Conclusion. No significant advantage was found for CFNI over a single-shot femoral block and subarachnoid anaesthesia after TKA. Cite this article: Bone Joint Res 2015;4:11–16