This study investigated the effects of non-steroidal anti-inflammatory drugs (NSAIDs) on peri-operative blood loss during elective total hip replacement. Patients were randomized to receive enteric-coated diclofenac 50 mg (n = 18), rofecoxib 12.5 mg (n = 17) or placebo (n = 16) administered orally three times daily for 2 weeks prior to surgery. Severe adverse effects resulting in discontinuation of trial participation occurred in six patients in the diclofenac group, five patients in the rofecoxib group and two patients in the placebo group; all drop-outs occurred at various times after surgery. Compared with placebo, peri-operative blood loss increased by 32% in the diclofenac group and by 7% in the rofecoxib group. Total mean +/- SD blood loss was 1040 +/- 136 ml in the diclofenac group, 844 +/- 83 ml in the rofecoxib group and 789 +/- 82 ml in the placebo group. Thus, administering a non-selective NSAID 2 weeks prior to elective total hip replacement significantly increases peri-operative blood loss.