To quantify the effects of postoperative pain relief on surgical stress response, 16 patients undergoing cholecystectomy were allocated randomly to double-blind treatment with either fentanyl by patient controlled analgesia (PCA) with the Prominject plus saline given s.c. by nurses on demand (PCA group) or saline by the infusion pump plus morphine 10 mg/70 kg s.c. by nurses on demand (control). Pain intensity (VAS) and plasma catecholamine, cortisol and glucose concentrations were measured 2-hourly for 12 h after operation. PCA improved postoperative pain intensity (P less than 0.05) and reduced plasma cortisol (P less than 0.05), but not glucose and catecholamine concentrations compared with the control group. Thus improved postoperative pain relief per se by PCA with systemic opioids had no major influence on the catabolic response to abdominal surgery.