Exercise therapy is effective for improvement in range of motion, function, and pain in patients with frozen shoulder: a systematic review and meta-analysis

Key points

1. 01Frozen shoulder is a self-limiting but painful condition of capsular inflammation and fibrosis that restricts both active and passive shoulder range of motion (ROM).
2. 02Across 33 included studies (19 pooled), exercises both alone and within a program improved ROM, function, and pain.
3. 03Most head-to-head program formats showed little to no difference for ROM and pain, and the effect on function was uncertain.
4. 04Adding exercises to a program (vs a program without exercises) improved active ROM, but adding physical modalities (vs sham) gave no benefit for ROM or pain.
5. 05Muscle energy techniques (eg proprioceptive neuromuscular facilitation) improved function more than other exercise types, while most evidence was rated low quality by GRADE.

How it was conducted

Design

Systematic review and meta-analysis of randomized controlled trials, certainty rated by GRADE

Search

PubMed, Web of Science, and Cochrane CENTRAL plus hand search; last search April 2019, updated May 2021

Participants

33 studies of patients with frozen shoulder (adhesive capsulitis), 19 included in the meta-analysis

Intervention

Exercise therapy alone or within multimodal programs vs solely exercises, programs without exercises, sham, or different exercise types

Outcomes

Passive and active ROM, function/disability (eg SPADI, Constant Murley Score), pain, and muscle strength

What they found

• 33 studies were included qualitatively and 19 in the meta-analysis; most studies were rated low quality by GRADE.
• Exercises within a multimodal program vs solely exercises: uncertain effect on function (SMD -0.04, 95% CI -0.56 to 0.64) and no clear pain reduction (MD -1.13, 95% CI -2.61 to 0.35).
• Adding exercises vs a program without exercises increased active ROM (eg flexion MD 16.00, 95% CI 14.07 to 17.93), but function (SMD -0.78, 95% CI -2.06 to 0.49) and pain (SMD -0.06, 95% CI -0.42 to 0.30) effects were uncertain.
• Muscle energy techniques vs other exercises: likely improved function/disability (SMD -0.62, 95% CI -1.28 to 0.04) with similar ROM and uncertain pain (SMD -0.36, 95% CI -1.24 to 0.52).
• Physical modalities plus exercises vs sham plus exercises gave no improvement in ROM (MD 1.51, 95% CI -4.14 to 7.16) or pain (MD 0.10, 95% CI -0.26 to 0.46); preliminary evidence favored supervised over home exercise (ROM MD 10.96 degrees, function CMS 16.10 points).

Limitations

• Heterogeneity of exercise types and a shortage of trials on solely exercise prevented some planned meta-analyses.
• Most included studies were low quality (GRADE), largely due to poor reporting of adherence and lack of blinding.
• Only one reviewer performed the GRADE assessment, and only English and Dutch studies were included with no gray literature.
• No conclusions could be drawn about long-term outcomes or the most effective exercise dose.

Why it matters

For patients

Exercises help restore movement, function, and reduce pain in a frozen shoulder, and supervised sessions may add a bit over a home-only program, but added machines or modalities do not help.

For clinicians

Prescribe exercise (muscle energy techniques are a good choice for function), but do not expect added physical modalities to help, and recognize the best program format and dose are unsettled.

For readers

Exercise works for frozen shoulder, yet the evidence cannot say which exercise format or dose is best, and adding modalities adds nothing.