Manipulation under anaesthetic (MUA) for the treatment of frozen shoulder is well established and effective however timing of surgery remains controversial. Intervention before 9 months has previously been shown to be associated with improved outcome. We test this theory by measuring Oxford Shoulder Score (OSS), re-MUA and subsequent surgery rate.

A retrospective review of a prospectively collected, single surgeon, consecutive patient series revealed 244 primary frozen shoulders treated by MUA within 4 weeks of presentation. The mean duration of antecedent symptoms was 28 weeks (95% CI 4-44 weeks) and time to follow up was 26 days (95% CI 11-41 days). The mean OSS improved by 16 points (2-tailed t test p< 0.001) with a mean follow up OSS of 43 (95% CI 38-48). 195 shoulders were manipulated before 38 weeks (9 months) and had the same mean change in OSS (16) as the 49 shoulders manipulated after 38 weeks. 48 shoulders, including 15 diabetic shoulders required further MUA. 8 shoulders had subsequent surgery. These events were also independent of antecedent symptom duration.

Early MUA does not appear to produce improved outcomes when compared to later intervention but we note does result in an earlier return to function.