OBJECTIVE: To compare the utility of intra-articular analgesia (IA) to that of a continuous interscalene block (CIB) by evaluating the quality of postoperative analgesia, a satisfaction index, and the incidence of complications. PATIENTS AND METHOD: A randomized controlled trial enrolling patients classified as ASA 1 or 2. The IA group received 25 mL of ropivacaine 0.2% plus 2 mg of morphine. The CIB group received a 7-mL/h infusion of bupivacaine 0.0625% plus 1 microg/mL of sufentanil. Postoperative pain was expressed on a visual analog scale (VAS) at 2, 4, 6, 12, 24, and 48 hours. Intravenous morphine was used as a rescue analgesic. Morphine use, incidence of adverse effects, and level of patient satisfaction after 48 hours were recorded. RESULTS: Twenty-three patients were randomized to the IA group and 24 to the CIB group. There were no between-group differences in patient characteristics. VAS scores and morphine use were similar in the 2 groups in the first 12 hours but were lower in the CIB group at 24 and 48 hours. The level of patient satisfaction was higher in the CIB group. There was a higher incidence of nausea and/or vomiting in the IA group at 24 and 48 hours. CONCLUSIONS: Postoperative pain in the first 12 hours after shoulder surgery can be adequately managed with either IA or CIB. CIB is more effective than IA between 12 and 48 hours after surgery.