Twenty-two studies including 989 patients in 26 cohorts were included in the review. None directly compared outcomes for manipulation under anaesthesia with arthroscopic capsular release. One cohort study compared manipulation under anaesthesia with manipulation followed by arthroscopic capsular release where the surgeon felt the initial manipulation to be insufficient. Ten studies compared different types of release, assessed the effects of one treatment in different patient populations, or used a comparator treatment outside the scope of the review. Other studies were case series. Median follow-up was 35 months for arthroscopic capsular release and 24 months for manipulation under anaesthesia.
Range of motion: For abduction, median range of motion showed a 6° greater improvement with capsular release compared with manipulation under anaesthesia. A median of 8° greater improvement in external rotation at the side was also found. There was little difference in post-treatment improvement in forward elevation between the two treatments.
Patient-reported outcomes: The Constant score was the most commonly reported; the median pre-post change in score was similar (47 for manipulation versus 50 for capsular release). Mean post-treatment scores and median scores adjusted for age and gender were higher in groups of patients treated with capsular release.
Complications: The overall complication rate for either treatment was 0.5% based on 15 studies of 638 patients with three reported complications. These were one case of proximal humerus fracture for manipulation and one case each of superficial wound infection and brachial plexopathy for capsular release.