The optimal site for wound delivery of local anaesthetic after total knee arthroplasty is undetermined. Sixty patients having total knee arthroplasty received intra-operative infiltration analgesia with ropivacaine 0.2% and were then were randomly assigned to receive either intracapsular or intra-articular catheters with 20 ml ropivacaine 0.5% given at 6 h and again at 24 h, postoperatively. Analgesic efficacy was assessed for 3 h after each injection, using a visual analogue score, where 0 = no pain and 100 = worst pain. There was no statistically significant difference between groups. Maximum pain relief (median (IQR [range])) at rest observed in the 3 h after the 6 and 24 h postoperative injections was 17 (7-31 [0-80]) and 10 (4-27 [0-50]) p = 0.27 for 6-9 h; and 17 (7-33 [0-100]) and 13 (3-25 [0-72]) p = 0.28 for 24-27 h, for intracapsular and intra-articular, respectively. Intracapsular local anaesthetic has similar analgesic efficacy to intra-articular after total knee arthroplasty.