Adhesive capsulitis usually responds favorably to therapeutic exercise, but manipulation and arthroscopic capsular release are effective for refractory cases. The results of non-operative and operative treatment for adhesive capsulitis are similar, but patients with diabetes and more limited function appear to benefit the most from capsular release. Compared with older patients, younger patients respond better to non-operative treatment, but they are also more likely to undergo surgery. Adhesive capsulitis or frozen shoulder often responds to a four-quadrant stretching program but operative intervention comprising manipulation and arthroscopic capsular release is typically employed for refractory cases. Previous studies have demonstrated good outcomes with both treatment methods, but few studies have evaluated concurrently the results of both non-operative and operative treatment. We hypothesized that both treatments produce similar outcomes and that certain patient factors predict outcome and the need for surgery. Ninety-nine consecutive patients (104 shoulders) underwent non-operative treatment for adhesive capsulitis until either symptom resolution or surgery. There were 65 females and 34 males with a mean age of 52.7 years (range 34-83 years). Non-operative treatment comprised supervised four-quadrant stretching and selective intra-articular corticosteroid injections. Twenty-nine patients, including 9 diabetics, underwent manipulation and arthroscopic pan-capsular release followed by continuous passive motion and supervised stretching. Range of motion was measured upon discharge from treatment. Shoulder function was evaluated using the Simple Shoulder Test (SST) and ASES scores at minimum 24 months follow-up (mean 39 months, range 24-68 months). Four patients (5 shoulders) were deceased and 10 patients were lost to follow-up so that outcome scores could be obtained for 85 patients (89 shoulders). Forward elevation, external rotation at the side, and internal rotation to the back improved from 118±21 to 152±15, 30±18 to 46±11, and from the gluteal region to the T12 spinous process, respectively (p<0.0001). Similar improvements in mobility were observed for both operative and non-operative groups, but the improvement in forward elevation was greater for the operative group (p<0.05). Mean SST improved from 4.0±2.7 to 9.9±2.8 (p<0.0001) and mean final ASES score was 85±15. Female patients had a lower mean initial SST than male patients (3.3±3.6 compared with 5.2±2.6, p<0.005), but mean final SSTs were similar (9.7±2.9 compared with 10.3±2.6). Improvement in SST and final ASES scores did not appear to vary by gender, diabetes, or depression. Younger patients (p<0.001) and those with lower initial SST scores (p<0.05) were more likely to undergo surgery. For the entire cohort, initial SST score predicted final SST score (p<0.05) and shorter duration of symptoms predicted a higher final ASES score (p<0.05). Non-operative and operative groups demonstrated similar final SST (10.0±2.8 and 9.7±2.9) and ASES scores (86±13 and 82±12). Following non-operative treatment, patients with diabetes had a lower mean final SST score than patients without diabetes (8.9 compared with 10.6, p<0.05), but following surgery the mean scores were identical (9.7). For patients undergoing non-operative treatment, absence of diabetes (p<0.005), shorter duration of symptoms (p=0.05), and young age (p<0.01) predicted a higher final SST. Adhesive capsulitis usually responds favorably to therapeutic exercise, but manipulation and arthroscopic capsular release are effective for refractory cases. The results of non-operative and operative treatment for adhesive capsulitis are similar, but patients with diabetes and more limited function appear to benefit the most from capsular release. Compared with older patients, younger patients respond better to non-operative treatment, but they are also more likely to undergo surgery.