The study by Buchbinder et al¹ suffers from a major drawback to any study dealing with frozen shoulder when the pathophysiology has passed the acute phase. In their study the mean duration of symptoms was 25.5 weeks in the active group with a standard deviation of 13.3 weeks, the mean therefore being approximately 6 months. This is well over the time course one would expect in the inflammatory phase of frozen shoulder, and therefore it is not surprising that prednisolone provided some benefit. I suspect the benefit provided related to improvement in myalgia and wellbeing that occurred with the prescription of prednisolone, but which was quickly lost when prednisolone was reduced and stopped.

Although the authors quote a study co-authored by Buchbinder on a standardised protocol for the measurement of shoulder movement,² I have concerns that the definition of frozen shoulder, as restricted passive movement by <30° in two or …