Effectiveness of Combined Program of Manual Therapy and Exercise Vs Exercise Only in Patients With Rotator Cuff-related Shoulder Pain: A Systematic Review and Meta-analysis

Abstract

Context:

Therapeutic exercise is considered the mainstay in the management of rotator cuff-related shoulder pain (RCRSP). Manual therapy (MT) interventions have also shown to be effective in RCRSP. However, the benefits of adding MT along with exercise interventions for the management of RCRSP remain unknown.

Objective:

To evaluate the additional benefits of MT with exercise compared with exercise in isolation for the management of RCRSP.

Data Sources:

A search of PubMed, Scopus, PEDro, and EBSCO from the inception date of each database through April 20, 2022, was conducted for randomized trials comparing the additional effects of MT in exercise interventions compared with exercise alone for pain management and function in patients with RCRSP. Standardized mean differences (SMDs) and 95% CIs were calculated using a random-effects inverse variance model according to the outcome of interest and comparison group. Methodological quality was assessed with PEDro and quality of evidence with the grading of recommendations assessment, development and evaluation approach.

Study Design:

Meta-analysis of randomized controlled trials.

Level of Evidence:

Level 2.

Results:

Twelve articles were found eligible and 8 of them demonstrated high methodological quality. Eleven articles were included for quantitative analysis. Pain with movement was not significantly different between MT and exercise versus exercise alone (SMD [95% CI] = −0.15 [−0.41 to 0.12]; I² = 0%), whereas pain at rest was significantly improved in the groups that used exercise only with a moderate effect size (SMD [95% CI] = 0.47 [0.04 to 0.89]; I² = 75%). Furthermore, shoulder function was not significantly different between MT and exercise versus exercise alone in the short term (SMD [95% CI] = 0.23 [−0.22 to 0.69]; I² = 88%) or the long term (SMD [95% CI] = −0.02 [−0.21 to 0.16]; I² = 2%).

Conclusion:

Adding MT to exercise interventions for the management of RCRSP is not more effective than exercise alone for pain and function in adult patients.

Shoulder pain is one of the most common musculoskeletal symptoms, with an estimated prevalence of 15% to 30% of the population at any one time. ⁴⁹ Furthermore, rotator cuff-related shoulder pain (RCRSP) is evaluated in approximately 70% of the patients who experience shoulder pain. ⁴⁹ RCRSP is an over-arching term that encompasses a spectrum of shoulder conditions including subacromial pain, (impingement) syndrome, rotator cuff tendinopathy, and symptomatic partial and full-thickness rotator cuff tears. ³² It has been argued that the term RCRSP is more appropriate than more traditional diagnoses that were related to pathoanatomic and structural pathologies because it remains almost impossible to identify a specific structure as the primary contributor to a patient’s shoulder pain.^32,41

Conservative management of RCRSP entails strengthening exercise, stretching, and mobility exercise of the shoulder, the thoracic and cervical spine. Exercise is considered the mainstay of management of shoulder pathology, with several systematic reviews highlighting the statistical and clinical superiority of exercise interventions on pain and function.^1,31,51 However, previous research trials have shown beneficial effects from the addition of manual therapy (MT) in exercise interventions for patients with RCRSP.^3,5,13,35,53 It has been proposed that MT may induce hypoalgesia^40,50 or restore normal biomechanics by increasing the range of motion in the shoulder region.^33,55

The benefits of MT in RCRSP have been mostly investigated in isolation in previous systematic reviews.^6,7,17 Moreover, the results of the aforementioned reviews suggested that there was a lack of evidence concerning the efficacy of MT when used alone and that evidence regarding the efficacy of the addition of MT to exercise for the treatment of rotator cuff tendinopathy was not conclusive. However, the possible adjunctive benefits of MT with exercise interventions when compared with exercise alone have not previously been investigated in a systematic review. Furthermore, since then, new studies have been published that investigated the additional benefits of MT with exercise compared with exercise alone and not only in rotator cuff tendinopathies but also in a variety of clinical populations that are now under the umbrella term RCRSP. Thus, this systematic review aimed to perform an updated review of the evidence regarding the efficacy of MT with exercise, compared with exercise alone, for the treatment of RCRSP.

Methods

This systematic review was conducted by the criteria set out in the preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement, and it was prospectively registered on PROSPERO (CRD42021288224).

Search Methods

International electronic databases were used for the literature search, including PubMed, Scopus, PEDro, and EBSCO from inception up to April 20, 2022. We consulted with experts in the field, manually reviewed the reference lists of articles that fulfilled the eligibility criteria, and searched the gray literature for eligible articles. The keywords that were used for the search were shoulder impingement, subacromial pain, tendinopathy, rotator cuff tears, manual therapy, mobilization, manipulation, exercise, function, strength, range of motion, and similar combinations of words. All studies were downloaded into Endnote X8 for screening. A detailed explanation of our search strategy for each database is listed in Appendix Table A1 (available in the online version of this article).

Study Selection

Articles were initially screened for eligibility by title and abstract and then by full text. The search and full inclusion process was performed by 2 independent reviewers using the PICOS framework (P = participants; I = interventions; C = comparison; O = outcomes, S = study design). Full texts of potentially relevant articles were retrieved for final evaluation. The selection process for the selected studies was conducted by consensus, and when a consensus was not achieved, a third reviewer was available to assist the process for a final judgment.

Eligibility Criteria Based on the PICOS Framework

Participants

Studies were eligible for inclusion if they recruited adult participants (≥18 years). Only studies that recruited symptomatic patients diagnosed with RCRSP, including subacromial pain (impingement) syndrome, rotator cuff tendinopathy, and symptomatic partial, full-thickness rotator cuff tears, subacromial bursitis or disorders that affect tissues and result in shoulder pain, were included in the final selection. In addition, studies with nonspecific shoulder pain were included. Only articles that clearly stated that diagnosis was made based on physical examination were included. Studies that evaluated patients only through diagnostic imaging were excluded. Studies that recruited patients with a history of traumatic injury, surgery or systemic diseases, or diseases related to other areas such as the cervical spine were excluded.

Intervention

Studies that evaluated the effectiveness of MT and exercise versus exercise alone were eligible. We considered MT the following interventions: massage therapy, mobilization, and manipulation. Additionally, we considered as exercise the following interventions: mobility and strengthening exercises, aerobic exercises, and motor control exercises performed on the ground, on machines, or in water, with or without supervision.

Comparison Groups

The comparator was groups of patients who received only exercise. We considered as exercise the following: mobility and strengthening exercises, aerobic exercises, and motor control exercises performed on the ground, on machines, or in water with or without supervision.

Outcome Measures

Studies were considered eligible if they analyzed at least 1 of the following outcome measures at baseline and final follow-up assessment: pain with a subjective measurement, function with a shoulder-specific questionnaire (ie, shoulder pain and disability index; SPADI), range of motion, and strength. The primary outcomes considered were pain and function, whereas all the others were considered secondary outcomes.

Study Design

English-written articles and randomized controlled trials were considered eligible for this review. Articles that were not full text and pilot studies were excluded.

Methodological Quality Assessment

The methodological quality and risk of bias of the included studies was assessed with the PEDro scale, which consists of 11 items related to the validity of the articles assessed. ¹⁶ The first item concerned the external validity; however, this was not included in the final score. The remaining items concerned internal validity. These items were related to the random allocation, allocation concealment, baseline comparability, blinding of therapists, patients, and raters, experimental mortality, intention-to-treat analysis, statistical comparisons and point measures, and measures of variability. These items could assist the readers to identify studies that were likely to be internally valid (items 2-9) and studies with sufficient statistical information to make their results interpretable (items 10 and 11). ⁵² The final score of the PEDro may have ranged from 0 (low quality) to 10 (high quality).

Previously indexed studies in the PEDro database maintained their scores, whereas nonindexed studies were assessed by 2 independent reviewers, and with the assistance of a third reviewer, when needed, to resolve any discrepancies. Based on the protocol of previous studies, studies that scored ≥7 in the PEDro were classified as studies with high methodological quality, 5 or 6 with moderate quality, and ≤4 as poor.

Data Extraction

Data were extracted independently by 2 reviewers using a standardized form that collected information regarding participant characteristics, study design, follow-up, interventions (type, duration, and the number of sessions), comparison group characteristics, and outcomes.

Data Synthesis and Analysis

Interrater agreement for reviewers’ assessments of study eligibility was calculated with Cohen’s kappa (κ) coefficient. A κ of 0 to 0.2 represented slight agreement; 0.21 to 0.40, fair agreement; 0.41 to 0.60, moderate agreement; and 0.61 to 0.80, substantial agreement. ¹⁷ A value greater than 0.80 was considered as complete agreement. Interrater agreement for assessments of methodological quality was calculated with the intraclass correlation coefficient (ICC). The κ and ICC were calculated using SPSS software version 26 (IBM Corp).

The review manager software (RevMan 5.3) was used to summarize the effects of MT and exercise versus exercise alone when sufficient common outcomes were available for a meta-analysis. Outcome data were transformed to 0 to 100 point scales for the meta-analysis. Subgroup analysis was performed for each outcome based on follow-up duration. The length of follow-up was divided into short-term (≤3 months), intermediate (4 and 5 months), and long-term (≥6 months) follow-up. Mean differences for the same outcome or standardized mean differences (SMDs) (when necessary) with 95% CIs for continuous data were calculated. Due to clinical and methodological heterogeneity between studies, a random-effects model was selected to pool studies’ outcomes. According to Cohen’s criteria, SMD values were classified as small (≤0.20), moderate (between 0.21 and 0.79), and large (≥0.80). ¹² In addition, heterogeneity was assessed using the I² statistic and we interpreted I² values >50% as significant heterogeneity. ²⁷ Whenever possible, the overall treatment effect was compared with its minimum clinically important difference (MCID). Sensitivity analysis was conducted by excluding studies with ‘low’ or ‘moderate quality’ (PEDro score <7) or large treatment effect sizes.

Quality of Evidence

To evaluate the overall quality of evidence and strength of recommendations, we applied the grading of recommendations assessment, development and evaluation (GRADE) in the meta-analysis for each outcome. ²³ Meta-analyses of studies were initially categorized as high-quality evidence and were downgraded by 1 level for each serious flaw present in the following overall domains: risk of bias (we downgraded quality of evidence when >25% of the studies in each comparison was considered as low quality), inconsistency (we downgraded quality of evidence when significant heterogeneity [I² value >50%], minimal or no overlap of confidence intervals and wide variance of point estimates across studies was detected), indirectness (we downgraded quality of evidence when participants, interventions, or outcomes measures from included studies were essentially different), imprecision (we downgraded quality of evidence when the pooled sample was <400 participants and/or wide confidence intervals were evident indicating no benefit from the intervention under study), publication bias (we downgraded quality of evidence in meta-analyses that included >10 studies that presented with asymmetrical funnel plots, sponsored studies or the authors reported conflict of interest). Based on the above criteria, the quality of evidence was classified as very low, low, moderate, or high. ²⁷

Results

Study Selection

We identified 42,295 trials after removal of duplicates that were potentially relevant and after reading the titles and abstracts 20 articles were found as potentially eligible for review. Full texts of the 20 articles were scrutinized for eligibility based on our inclusion and exclusion criteria and 8 articles were excluded. The reasons for the removal of the studies can be found in Figure 1, and these were related to the pathology of the participants,^9,20,37 or the interventions that were compared.^{14,15,18,22,28,54} Finally, 12 articles were found appropriate for a qualitative synthesis.^{2,8,10,19,24,29,30,36,45–47,55} After excluding 1 article ⁴⁷ due to a lack of relevant data available in the full text and even after a request from the primary authors, the remaining 11 were found appropriate for quantitative synthesis. There was almost complete agreement between the 2 reviewers at the title/abstract screening stage (κ = 0.80) and at the full-text screening stage (κ = 0.82).

Figure 1.

Flow chart of literature search results.

Participants

The characteristics of the participants from the included studies are listed in Appendix Table A2 (available online). The final sample in our qualitative synthesis was 574 participants after 52 dropouts (9%). The vast majority of the studies included samples with mean ages that ranged from 42 to 65 years, except for 2 studies that recruited younger patients between the ages of 21 and 35 years.^8,46 Also, the majority of the studies recruited more women than men; however, 1 study ⁴⁵ did not mention the female:male ratio, and thus it was impossible to estimate the pooled sample of women separately. All the studies assessed patients with physical examination, and the majority of the studies used common orthopaedic special tests to identify signs of impingement, except for 2 studies that recruited patients with painful shoulder during physiological and accessory movements in the shoulder region.^10,55

Interventions

In all studies, exercise programs consisted of strengthening and mobility exercises for the shoulder, whereas 4 provided supervision^2,8,30,36 and 1 provided extra advice on how to avoid or minimize painful shoulder movements during activities of daily living. ⁵⁵ Furthermore, 4 studies incorporated exercises that targeted the scapula^2,36,46,55 and motor control exercises.^36,46,55 As for the MT techniques, 4 studies used techniques that targeted the lower cervical or the thoracic spine,^2,8,24,36 5 studies targeted the shoulder regions, including the glenohumeral, acromioclavicular, and sternoclavicular joints,^2,8,10,45,55 1 the scapula directly ⁴⁵ and 1 study targeted both the spine (thoracic and cervical) and the glenohumeral joint. ²⁹ Of the included studies, 2 used grade III and IV mobilizations,^2,8 1 study used high-velocity mobilizations, ³⁶ 2 studies used low velocity mobilizations,^10,55 1 study used 1 nonthrust mobilization and 3 different thrust manipulation techniques directed at the thoracic spine. ²⁴ One study did not clearly define the force of the mobilization but commented only on the direction of the mobilization. ³⁰

Outcome Measures

Pain with activity in the short term was assessed in 4 studies^2,8,45,55 and the long term only in 1. ⁵⁵ Pain at rest in the short term was assessed in 6 studies^{2,8,24,30,36,45} and in the long term only in 1. ³⁶ Shoulder function was assessed with the SPADI in the short term in 5 studies,^{10,24,30,36,46,55} and with the Constant Murley score, ¹⁹ the disabilities of the arm, shoulder and hand (DASH) questionnaire, ⁸ and the functional assessment score. ² Shoulder function was also assessed in the long term in 3 studies.^10,36,55 A pooled estimate of effect was generated for the aforementioned outcomes. Furthermore, 2 studies assessed pain levels with the SPADI in the long term.^30,55 One study measured shoulder strength ² and 2 measured shoulder range of motion,^10,45 although the values were available only in 1 of them. ¹⁰ In outcomes that were used only once, a pooled estimate of effect was not possible. These were the self-rated questionnaire for symptom change, ⁵⁵ scapular kinematic measurement and mechanical sensitivity, ⁸ the QuickDASH, ³⁶ the global rating of change, ³⁶ and the patient acceptable symptom state. ³⁶

Risk of Bias

A summary of the risk of bias across the included studies is listed in Table 1. Based on the PEDro criteria, 67% of the studies were classified as studies with high methodological quality. The scoring of studies for the risk of bias ranged from 4 to 8. Blinding of therapists was not feasible in any trial, while almost all of them did not satisfy the fifth criterion that is related to blinding of individuals. A significant proportion of the eligible studies (42%) did not include an intention-to-treat analysis.

Table 1.

Methodological quality assessment using the PEDro scale

Studies	1	2	3	4	5	6	7	8	9	10	11	Total Score
Bang and Deyle (2000) 2	Y	Y	N	Y	N	N	Y	Y	N	Y	Y	6
Senbursa et al (2007) 45	Y	Y	N	Y	N	N	N	N	N	Y	Y	4
Chen et al (2009) 10	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	8
Yiasemides et al (2011) 55	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	8
Kromer et al (2013) 29	Y	Y	Y	Y	N	N	N	Y	Y	Y	Y	7
Kromer et al (2014) 30	N	Y	Y	Y	N	N	N	Y	Y	Y	Y	7
Camargo et al (2015) 8	N	Y	Y	Y	N	N	Y	Y	Y	Y	Y	8
Mintken et al (2016) 36	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	8
Haider et al (2018) 24	N	Y	N	Y	N	N	N	Y	N	Y	Y	5
Sharma et al (2021) 46	Y	Y	Y	Y	Y	N	N	Y	N	Y	Y	7
Sharma et al (2021) 47	Y	Y	N	Y	N	N	Y	Y	N	Y	Y	6
Eliason et al (2021) 19	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	8
≥7/10												8/12

N, no; Y, yes.

^1.Eligibility criteria; 2. Random allocation; 3. Concealed allocation; 4. Baseline comparability; 5. Blinding of individuals; 6. Blinding of therapists; 7. Blinding of assessors; 8. Adequate follow-up; 9. Intention-to-treat analysis; 10. Between-group comparisons; 11. Point estimates and variability.

Item 1 (eligibility criteria) does not contribute to the total score.

Meta-analysis

Pain

The adjunctive effects of MT with exercise versus exercise alone on pain during movement were examined by 4 studies^2,8,45,55 in the short term (≤3 months) from a pooled sample of 224 subjects (Figure 2). Overall, on completion of the interventions, pain with movement was not significantly different between MT and exercise versus exercise alone (SMD [95% CI] = −0.15 [−0.41 to 0.12]; I² = 0%), with no substantial heterogeneity, based on moderate-quality evidence (Appendix Table A3 [available online]). Furthermore, the adjunctive effects of MT with exercise versus exercise alone on pain at rest were examined by 6 studies^{2,8,24,30,36,45} in the short term (≤3 months) from a pooled sample of 394 subjects. Overall, on completion of the interventions, pain at rest was significantly improved in the groups that used exercise only when compared with the groups that also used MT, with a moderate effect size (SMD [95% CI] = 0.47 [0.04 to 0.89]; I² = 75%) and substantial heterogeneity, based on low-quality evidence (Appendix Table A3 [available online]).

Figure 2.

Forest plot showing the effects of exercise compared with manual therapy (MT) and exercise for pain in patients with rotator cuff-related shoulder pain. Data are depicted according to measurement conditions. IV, inverse variance.

Shoulder Function

The adjunctive effects of MT with exercise versus exercise alone on function were assessed with the SPADI,^{10,24,30,36,46,55} the Constant-Murley score, ¹⁹ the DASH, ⁸ and the functional assessment score. ² The results in the short term (≤3 months) were examined by 9 studies^{2,8,10,19,24,30,36,46,55} from a pooled sample of 699 subjects (Figure 3). Overall, on completion of the interventions, shoulder function was not significantly different between MT and exercise versus exercise alone (SMD [95% CI] = 0.23 [-0.22 to 0.69]; I² = 88%), with substantial heterogeneity, based on moderate-quality evidence (Appendix Table A4 [available online]). The adjunctive effects of MT with exercise versus exercise alone on function with the SPADI in the long term (≥ 6 months) were examined by 5 studies^{10,19,30,36,55} from a pooled sample of 482 subjects. Overall, on completion of the interventions, shoulder function was not significantly different between MT and exercise versus exercise alone (SMD [95% CI] = −0.02 [−0.21 to 0.16]; I² = 2%), with no substantial heterogeneity, based on high-quality evidence (Appendix Table A4 [available online]). Sensitivity analysis demonstrated that the inclusion of studies that used only the SPADI did not change the significance of the overall result, either in the short term (MD [95% CI] = 5.21 [−0.73 to 11.15]; I² = 82%) or the long term (MD [95% CI] = 2.24 [−2.19 to 6.68]; I² = 0%).

Figure 3.

Forest plot showing the effects of exercise compared with manual therapy (MT) and exercise for shoulder function in patients with rotator cuff-related shoulder pain. Data are depicted according to follow-up time. IV, inverse variance.

SPADI Pain Scale

The pain scale of the SPADI was used separately on 2 occasions^30,55 to examine the effects of MT with exercise against exercise only, in the long term from a pooled sample of 175 subjects (Figure 4). Overall, pain was not significantly different between MT and exercise versus exercise alone (SMD [95% CI] = −0.14 [−0.44 to 0.16]; I² = 0%), with no substantial heterogeneity, based on low-quality evidence (Appendix Table A5 [available online]).

Figure 4.

Forest plot showing the effects of exercise compared with manual therapy and exercise for pain with the shoulder pain and disability index in patients with rotator cuff-related shoulder pain. Data are depicted according to measurement conditions. IV, inverse variance.

Discussion

The purpose of this systematic review was to determine the additional benefits of MT with exercise in patients with RCRSP when compared with exercise only for pain, function, range of motion, and strength. From our literature search, we found enough studies for a meta-analysis, only for our primary outcomes of pain and function. The results of this systematic review cannot support the superiority of a combined intervention that includes MT and exercise for the management of RCRSP.

In contrast to previous systematic reviews^6,7,17,51 that examined MT and exercise interventions for the management of shoulder pathology, this review provides more evidence for the lack of statistically important effectiveness of MT, in addition to exercise for RCRSP. Although when looking at the results from the meta-analysis, MT and exercise resulted in a superior reduction of pain at rest in the short term, and the results were not statistically significant. Also, previous research has supported that a minimum reduction of 14 of 100 for patients with shoulder pain is a clinically significant reduction. ⁴³ Although all studies except 1 of moderate quality ⁸ demonstrated a clinically important reduction in pain for both groups that received MT with exercise or exercise alone, the between-group difference in all studies was not clinically important (<14 MCID). Previous research has shown that a minimum change of 13.1 of 100 points in function for the SPADI, for patients with shoulder pain, is clinically significant. ⁴² Although all studies demonstrated a clinically important improvement in function with the SPADI, only 1 moderate quality study showed clinically superior functional improvements between groups in the short term, when MT was used as an adjunctive intervention with exercise (MCID >13.1). ⁴⁶

Previous reviews, including a Cochrane review, have proposed that MT with exercise may be more effective than exercise alone in the short term for patients with shoulder impingement.^6,7,38,51 However, this is the first systematic review with a meta-analysis that examined only studies that directly compared the effectiveness of adding MT in exercise interventions against exercise only for a variety of disorders that fall under the umbrella of RCRSP. Interestingly, this meta-analysis failed to show the superiority of MT in pain or function, by including studies that incorporated a variety of MT techniques, including high- or low-velocity mobilizations in the shoulder region or the spine. Joint-based MT has been shown to activate pain inhibitory cortical systems, ⁴⁸ and thus a reduction of pain, at least in the short term, should be expected. However, the results of our meta-analysis showed that exercise only, was statically more effective than exercise with MT in the short term, with a moderate effect size. It should be noted though, that the results did not demonstrate the clinical superiority of exercise in isolation.

Moreover, combined interventions were not more effective than single interventions. The addition of MT in exercise interventions has failed to show superiority in previous systematic reviews for patients with hip osteoarthiritis ⁴⁴ and neck pain. ²¹ Combined interventions may result in antagonistic interactions between treatments, but also patients who receive MT and exercise may spend less time on each exercise, which could affect the overall outcome of the treatment. Furthermore, the variety of different MT techniques that were used in the trials may have compromised the effectiveness of the interventions. For example, several different MT techniques were used based on clinical judgment, although the symptomatology of the subjects was similar in all studies. Using clinical judgment of course in clinical practice is a necessity; however, in research, it is difficult to draw any conclusions. Moreover, some studies recruited treating therapists with a variety of experience and qualifications, which may have affected the proper delivery of the interventions in nonpharmacological interventions. ⁴ More importantly, the majority of the included studies used more than 1 therapist for the delivery of the MT techniques, which could also have compromised the results. Finally, MT dosage was under-reported on many occasions^{2,8,10,29,45,47} in terms of duration. As limited research exists on the optimal dosage of MT for shoulder pathology, research is necessary to establish the minimally effective dose of MT before examining its effectiveness.

Another factor that needs to be considered when evaluating the adjunctive benefits of MT in exercise interventions for patients with RCRSP is the psychology of patients. Psychological factors may influence the perpetuation of pain intensity and disability in patients with shoulder pain. ³⁴ Also, personal and work-related factors may contribute to the development of shoulder pain. ¹¹ Thus, the exercise interventions used and the therapeutic alliance that is created between patients and therapists during the rehabilitation process might result in important changes in shoulder pathology and patients’ behavior that are together effective enough. ³⁹ As has been highlighted previously ²¹ adding MT is not expected to attribute to the patient-therapist relationship leading to no additional benefits for the patient. Clinicians may consider the findings of this study in light of treatment costs, which are bound to increase when more interventions are provided without evidence that proves efficacy.

Limitations

Our investigation should be interpreted in light of its limitations. Initially, this study included a limited number of studies, and thus publication bias could not be assessed. It is likely that eligible studies were not located because we included only English articles. Also, we considered MT interventions as several techniques such as mobilization, manipulations, and soft tissue massage based on previous research. ²⁶ However, there is currently no ideal classification of MT techniques as MT is broad by nature. ²⁵ Finally, only a limited number of studies assessed pain in the long term. Although there are limitations in this review, an important strength of this study was that we reduced the confounding effects of other interventions by including only studies that compared MT and exercise with exercise in isolation. Furthermore, the findings of this review are based on moderate- and high-quality studies, except for 1 low-quality study.

Conclusion

Based on high- to low-quality evidence, MT and exercise are not more effective than exercise in isolation for the management of shoulder pain and function in adult patients with RCRSP.