Since intravenous prophylactic anti-inflammatory agents have been suggested to reduce or even replace opiates in postoperative pain therapy, we studied the demand for morphine in 45 patients recovering from abdominal surgery who had received a baseline infusion of either indomethacin, morphine or saline placebo. When extubated after inhalational anaesthesia, each patient received an i.v. bolus of either 0.5 mg.kg-1 indomethacin, 0.07 mg.kg-1 morphine or saline placebo. Thereafter a 20-h infusion of the same test analgesic was started, either 0.1 mg.kg-1.h-1 indomethacin, 0.03 mg.kg-1.h-1 morphine or saline placebo. For additional analgesia, a patient-controlled analgesia device (PCA) delivering 5-mg boluses of morphine was used. For the first 5 postoperative hours, significantly more (P less than 0.05) PCA morphine was needed in the indomethacin group (35 mg) than in the morphine group (24 mg), while the placebo group demanded mean 30 mg. For equal analgesia (measured by VAS and VRS) between 5-20 h, similar amounts (mean 23 and 19 mg) of PCA morphine were required in the indomethacin and morphine groups, in contrast to the placebo group (mean 40 mg) (P less than 0.001). Morphine infusion increased the total consumption of morphine by 25% as compared to placebo. We conclude that, following abdominal surgery, the analgesic effect of indomethacin infusion became apparent after the first 5 postoperative hours, thereafter reducing the demand for PCA morphine by about 40%. Continuous morphine infusion diminishes the postoperative demand for PCA morphine, but also increases the total morphine consumption.