This PRISMA-compliant systematic review and network meta-analysis of 65 RCTs (4,097 participants) asks which treatments for frozen shoulder — injections, physiotherapy, manipulation, surgery, or combinations — produce clinically meaningful improvements in pain, function, and external rotation ROM. The key question is not just what is statistically significant, but what crosses the threshold for minimal clinically relevant difference.
Management of frozen shoulder has historically been inconsistent, with wide variation from primary to secondary care and no clear hierarchy among available treatments. Multiple options showed statistical superiority in prior trials, but whether those differences were large enough to matter clinically was rarely addressed.
When a patient presents with frozen shoulder of less than one year's duration, offer an intra-articular corticosteroid injection at first contact. It is the only intervention with proven short-term superiority over both placebo and physiotherapy for pain and function.
Pair the injection with a structured home exercise program. Physiotherapy alone adds early external rotation ROM but does not close the gap with injection for pain or function. The home exercise program is what extends injection benefit into the mid-term.
Do not upgrade to arthrographic distension, acupuncture, or ESWT based on statistical significance alone. None of these crossed the minimal clinically relevant difference threshold in this analysis, and that threshold — 1 VAS point for pain, 10° for external rotation ROM. Is the practical standard for recommending a treatment.
This PRISMA-compliant systematic review and network meta-analysis of 65 RCTs (4,097 participants) asks which treatments for frozen shoulder — injections, physiotherapy, manipulation, surgery, or combinations — produce clinically meaningful improvements in pain, function, and external rotation ROM. The key question is not just what is statistically significant, but what crosses the threshold for minimal clinically relevant difference.
Management of frozen shoulder has historically been inconsistent, with wide variation from primary to secondary care and no clear hierarchy among available treatments. Multiple options showed statistical superiority in prior trials, but whether those differences were large enough to matter clinically was rarely addressed.
When a patient presents with frozen shoulder of less than one year's duration, offer an intra-articular corticosteroid injection at first contact. It is the only intervention with proven short-term superiority over both placebo and physiotherapy for pain and function.
Pair the injection with a structured home exercise program. Physiotherapy alone adds early external rotation ROM but does not close the gap with injection for pain or function. The home exercise program is what extends injection benefit into the mid-term.
Do not upgrade to arthrographic distension, acupuncture, or ESWT based on statistical significance alone. None of these crossed the minimal clinically relevant difference threshold in this analysis, and that threshold — 1 VAS point for pain, 10° for external rotation ROM. Is the practical standard for recommending a treatment.