One hundred and thirty-seven patients were studied to assess whether the timing of a dose of ketorolac affected cumulative morphine requirements during the first 12 post-operative hours. Pain, sedation and nausea scores, respiratory rate and degree of operative blood loss were also recorded. Thirty-six patients (group A) were given placebo injections pre- and intra-operatively. Thirty-one patients (group B) received placebo pre-operatively and ketorolac 30 mg intra-operatively. Thirty-six patients (group C) received ketorolac 10 mg pre-operatively and ketorolac 20 mg intra-operatively and thirty-four patients (group D) were given ketorolac 30 mg pre-operatively and placebo intra-operatively. Post-operative analgesia was with intravenous (i.v.) morphine administered using a patient controlled analgesia (PCA) device. Analysis of variance revealed a significant difference in morphine consumption at 1, 2, 4, 8 and 12 h post-operatively (P < 0.05) between group A (no ketorolac) and groups B, C and D (ketorolac). However, there were no significant differences between groups B, C and D during the study period. Thus, the timing of ketorolac administration made no difference to overall morphine consumption. Pain, nausea, sedation and respiratory rate scores were similar in all four groups. There was a significantly greater blood loss in patients receiving ketorolac (groups B, C and D) compared with those receiving placebo alone (group A).