Purpose: To compare subcutaneous PCA tramadol with subcutaneous PCA morphine for postoperative pain relief after major orthopaedic surgery and for the incidence of side-effects. Methods: In a double-blind randomised controlled study 40 patients (20 in each group) self-administered either tramadol or morphine for 72 hr after surgery via sc PCA. The following variables were recorded at various time intervals: (i) pain score by means of a visual analogue scale, (ii) drug consumption and total PCA demands, (iii) vital signs (blood pressure and heart rate), (iv) oxygen saturation and respiratory rate, and (v) side-effects (sedation, nausea/vomiting, pruritus, urinary retention and constipation). Results: Both drugs provided effective analgesia. The mean consumption in the first 24 hr was 792 ± 90 mg tramadol and 42 ± 4 mg morphine. Thereafter, consumption of both drugs declined markedly. Moderate haemodynamic changes were observed in both the tramadol and morphine groups (with a maximum 20% decrease in mean blood pressure and a maximum 17% increase in heart rate) during the 72 hr period. Both tramadol and morphine were associated with a clinically and statistically significant (P < 0.001) decrease in oxygen saturation, but without changes in respiratory rates. Desaturation was less marked with tramadol. Tramadol appeared to cause more nausea and vomiting than morphine. Sedation was mild and only seen during the first few hours after surgery in both groups. Conclusion: Tramadol is an effective analgesic agent for the relief of acute postoperative pain when administered by PCA via the subcutaneous route. Under these conditions tramadol behaves much like morphine with a similar side-effect profile.