PURPOSE Rotator cuff repair may result in severe postoperative pain. We compared a continuous intra-articular infusion to a continuous interscalene block with ropivacaine for patients undergoing outpatient rotator cuff repair. METHODS Seventeen patients were randomized to one of two groups: 1) interscalene block with 0.5% ropivacaine (40 mL) followed by a postoperative intra-articular infusion or; 2) interscalene block with 0.5% ropivacaine (40 mL) followed by a postoperative continuous interscalene infusion. Infusions were 0.2% ropivacaine at 10 mL x hr(-1) for both groups. Infusions were maintained for 48 hr. Patients were discharged on the day of surgery. Verbal analogue pain scores (VAS) and postoperative oxycodone consumption were measured for 48 hr. RESULTS Eight patients (47%; four in each group) had side effects or logistical problems complicating care. The mean VAS scores at rest and with movement in the postanesthesia care unit and at 12 hr, 24 hr, and 48 hr were not different (P > 0.1). Inadequate analgesia was reported in 50-75% of all study patients. Time until first oxycodone use was similar between groups 829 min +/- 432 (interscalene) and 999 min +/- 823 (intra-articular; P = 0.6). Total oxycodone consumption was also similar 49 mg +/- 48 and 59 mg +/- 51 (P = 0.7), respectively. CONCLUSIONS This study demonstrates the difficulties of ambulatory interscalene and intra-articular infusion for rotator cuff surgery. The high VAS scores and need for additional medical care suggest that intra-articular administration may not be reasonable for this magnitude of surgery. Further refinement of the perineural local anesthetic infusion is necessary to consistently provide analgesia after ambulatory rotator cuff surgery.