EXERCISE THERAPY FOR CHRONIC SHOULDER PAIN: A NETWORK META-ANALYSIS

A. Silveira¹, C. Lima¹, L. Beaupre², J. Chepeha³, A. Jones²

¹University of Alberta, School of Public Health, Edmonton, Canada, ²University of Alberta, Faculty of Rehabilitation Sciences, Edmonton, Canada, ³University of Alberta, Collaborative Orthopaedic Research, Edmonton, Canada

Background: Exercise therapy (ET) is usually the first treatment choice when treating shoulder pain, yet evidence on optimal ET strategies to expedite recovery has yet to be defined. Moreover, the value of adding adjunct therapies (i.e. manual therapy, electrotherapy, medications, and injections) to exercise therapy is currently inconsistent.

Purpose: This network meta-analysis (NMA) synthesized and simultaneously combined both direct and indirect evidence across studies on the effectiveness of ET to treat adults with chronic shoulder pain.

Methods: Randomized or quasi-randomized control trials comparing ET plus or minus adjunct therapies were included. Outcomes of interested included pain, shoulder range of motion (ROM), and health-related quality of life (HRQL) measures. We systematically searched the following databases up to 22-May-2022: MEDLINE, Embase, CINAHL, Sportdiscus, CENTRAL, Conference Proceedings Citation Index- Science, clinicaltrials.gov, and association websites. Data analysis combined direct and indirect comparisons in a Frequentist hierarchical model. CINeMA tool assessed the confidence in the results.

Results: 52 studies met the study inclusion criteria comprising 3,893 participants. The mean age was 51.26 (SD=7.55), 52.7% were females and the mean intervention duration was 7.09 weeks (SD=3.67). ET alone (Mean difference (MD)= -2.1 and 95% confidence interval (CI)= -3.5 to -0.7) as well as in combination with electrotherapy (MD= -2.5 and 95% confidence interval (CI)= -4.2 to -0.7), injections (MD= -2.4 and 95% confidence interval (CI)= -3.9 to -1.04) and manual therapy (MD= -2.3 and 95% confidence interval (CI)= -3.7 to -0.8) decreased pain significantly when compared to usual medical care (consult with general physicians, advice on shoulder condition, use of pain medications or no treatment). Pain relief was clinically important (over 20% difference). A trend towards improvements of shoulder ROM and HRQL scores was seen; however, none of the interventions achieved statistical or clinically important significance. Sensitivity analysis including only moderate to low risk of bias studies showed the similar results, with exception of injections that did not reach significance (MD= -1.3 and 95% confidence interval (CI)= -4.3 to 1.7).

Conclusions: Based on findings from this NMA, ET alone or in combination with adjunct therapies is effective in providing pain relief among patients with chronic shoulder pain (pain for more than 3 months). No other differences in ROM and HRQL were identified. Results need be interpreted with caution, given the quality of evidence.

Implications: Clinically, findings from this NMA concluded that ET reduces shoulder pain as compared to usual care. Although evidence was based on studies that met strict inclusion/exclusion criteria, the quality of most included studies had low to moderate risk of bias. ET interventions need to be better defined in terms of terminology, frequency, delivery mode, treatment compliance and intervention duration. Moreover, studies with more strict methodology are needed to decrease the interventions variability and increase the quality of the evidence.

Funding acknowledgements: Alberta Strategy for Patient-Oriented Research (SPOR) SUPPORT Unit, Knowledge Translation Platform, University of Alberta

Keywords:
Exercise
Shoulder
Pain

Topics:
Musculoskeletal: upper limb
Orthopaedics

Did this work require ethics approval? No

Reason: There is no need for ethics approval for network meta-analysis.