The optimal dose of intrathecal morphine for postoperative analgesia after major surgery is a matter of debate, with some uncertainty concerning the therapeutic potential and safety of intrathecal morphine in the dose range 0.3-1.0 mg. This randomised double-blind study compared the efficacy and side-effect profile of 0.2 mg and 0.5 mg intrathecal morphine in 70 patients undergoing knee replacement surgery. The primary endpoint was the number of patients requiring rescue analgesia (tramadol) during the first 24 h postoperatively. Secondary endpoints included consumption of tramadol and the incidence of adverse effects. Fewer patients in the 0.5-mg group required rescue analgesia in the first 24 h than in the 0.2-mg group (16 (48%) vs 28 (85%), respectively; p = 0.003). Median (IQR [range]) tramadol consumption was lower in the 0.5-mg group than in the 0.2-mg group (0 (0-100 [0-350]) mg vs 100 (50-100 [0-350]) mg, respectively; p = 0.02). The incidence of adverse effects was similar in both groups. This study has demonstrated that 0.5 mg intrathecal morphine produces better analgesia than 0.2 mg after knee replacement without any increase in side-effects.