191 patients were randomized to compare the effect of corticosteroid injections at differing sites of the shoulder in treating adhesive capsulitis. Participants were allocated to one of four groups: one received injection to the subacromial bursa, another received injection to the glenohumeral joint, and another received injection at both sites. The fourth group was treated with non-steroidal anti-inflammatory drugs. The results at 24 week follow-up up indicated that improvement in pain, shoulder range of motion and shoulder function were similar among the three groups treated with corticosteroid injections, regardless of injection site.