Cochrane in CORR^®: Manual Therapy and Exercise for Rotator Cuff Disease

Importance of the Topic

Rotator cuff disease accounts for more than 70% of shoulder complaints and is one of the most common musculoskeletal disorders in the adult population [11]. Among the general population, its prevalence ranges from 7% to 26% [8]. The term is broad and encompasses acute and chronic pathology affecting the shoulder joint, including rotator cuff tears, tendonitis, and other similar pathologies. Individuals affected by this condition commonly describe debilitating pain with movement, particularly during overhead activities, as well as pain being worse at night. In the United States, rotator cuff tendinopathy accounts for approximately 4.5 million annual physician visits, with treatment and management reaching an estimated USD 3 billion annually. If we include indirect costs, such as lost time from work, this number is even larger [6].

Generally, initial treatment options include manual therapy including joint mobilization and manipulation, as well as specific exercise regimens and/or anti-inflammatory medications. The goal of physical therapy is to increase ROM, promote healing, strengthen periscapular musculature, and improve the stabilizing function of the rotator cuff [4, 7]. When nonsurgical approaches are unsuccessful, surgical treatment includes rotator cuff débridement or repair and often subacromial decompression This Cochrane review of randomized control trials evaluated the efficacy of exercise and/or manual therapy in the management of rotator cuff disease.

Upon Closer Inspection

Despite identifying more than 60 trials for inclusion, only one trial directly compared both exercise and manual therapy to placebo [2], which suggested a slight statistical improvement in function and little or no improvement in pain relief. When evaluating the results of such self-reported patient-reported outcomes from clinical trials, clinicians often focus on findings of statistical significance and reported p values, without consideration of effect sizes, confidence intervals, or whether the results have meaning in the clinical context [3, 10]. Confidence intervals allow for an assessment of the variability in the magnitude and direction of effect, while judgments about clinical relevance, typically informed by a metric called the minimum clinically important difference (MCID), help clinicians to consider whether the intervention produces improvements in health that are large enough for the patient to care about.

In the case of the single trial in this review that evaluated exercise and manual therapy compared to placebo, patients who received the active treatment had an improvement in physical function greater, on average, than patients who received placebo; however, the effect size was below the MCID. Despite the statistical finding, the actual impact of the intervention was so small that patients are unlikely to perceive that exercise and manual therapy makes a noticeable improvement in their overall shoulder pain or function [2].

Although two high-quality trials [1, 2] with small sample sizes were included in the review, the remainder of the studies included in the review were at risk of potential bias, particularly with regards to blinding, where 88% of included trials were either unblinded or blinding was unclear. In addition, selective outcome reporting was a concern; 80% of trial protocols were either not registered or outcome data were incompletely reported.

The sample sizes of the included trials in this review ranged from nine to 207 patients, which is particularly problematic when trying to examine the effect of an intervention or the additive effect of an intervention with a similar control (for example, exercise and manual therapy versus exercise alone). Small sample sizes make it difficult to detect whether true differences between treatment groups exist [12]. Several trials compared interventions to placebo or sham treatments. Although, this is an ideal approach to evaluate the efficacy of an intervention, it is important to recognize that not all placebo controls are equal and some do not have a true null treatment effect [5]. For example, differing effects may be seen with the use of mobilization, manipulation, ultrasound or laser as placebo comparators. Additionally, some trials did not distinguish between acute onset of pain and functional limits after traumatic and chronic nontraumatic etiology introducing a confounding variable of acute versus chronic rotator cuff tear management. In the former case, it is generally the approach of surgeons to operative, whereas it is usually the case in the latter group to treat with physiotherapy.

Take-home Messages

This Cochrane review evaluated the efficacy of exercise or manual therapy in the setting of rotator cuff disease, and concluded that no clinical benefit exists for such interventions over placebo or other treatments. The conclusions of this review are consistent with the American Academy of Orthopaedic Surgeons Clinical Practice Guideline Summary, which was unable to recommend for or against exercise programs for patients with rotator cuff tears due to inconclusive evidence [9].

More evidence from high-quality and adequately powered trials may produce results that differ from the conclusions of this review. Further research is required to identify which patients might benefit from such treatments given the heterogeneous nature of rotator cuff disease.

Given the limitations of the available evidence, surgeon experience and patient motivation should play an important role in determining whether exercise and/or manual therapy is attempted for individuals presenting with rotator cuff disease.

Appendix