The Efficacy of Manual Therapy for Rotator Cuff Tendinopathy: A Systematic Review and Meta-Analysis

Transcription

1 The Efficacy of Manual Therapy for Rotator Cuff Tendinopathy: A Systematic Review and Meta-Analysis Ariel Desjardins-Charbonneau, PT, MSc 1 Jean-Sébastien Roy, PT, PhD 2,3 Clermont E. Dionne, OT, PhD 2,4 Pierre Frémont, MD, PhD 2,5 Joy C. MacDermid, PT, PhD 6 François Desmeules, PT, PhD 1,7 1 Maisonneuve-Rosemont Hospital Research Center, University of Montreal Affiliated Research Center, Montreal, Quebec, Canada 2 Department of Rehabilitation, Faculty of Medicine, Laval University, Quebec City, Quebec, Canada 3 Center for Interdisciplinary Research in Rehabilitation and Social Integration, Quebec City, Quebec, Canada 4 Population Health Research Unit (URESP), Laval University Hospital (CHU) Research Center, Quebec City, Quebec, Canada 5 Laval University Hospital (CHU) Research Center, Quebec City, Quebec, Canada 6 School of Rehabilitation Science, McMaster University, Hamilton, Ontario, Canada 7 School of Rehabilitation, Faculty of Medicine, University of Montreal, Montreal, Quebec, Canada The authors declare that they have no competing interests. Financial support has been provided by the Institut de Recherche en Santé et Sécurité au Travail (IRSST) and the Réseau Provincial de Recherche en Adaptation-Réadaptation (REPAR). Ariel Desjardins-Charbonneau is supported by a fellowship of the Ordre de la Physiothérapie du Québec (OPPQ). Address for correspondence and reprints request: François Desmeules, PT, PhD Unité de recherche clinique en orthopédie/ Orthopaedic clinical research unit Centre de recherche de l Hôpital Maisonneuve-Rosemont (CRHMR) 5415 Blvd L'Assomption, Pav. Rachel Tourigny, bureau/office 4163 Montréal, QC H1T 2M4 Téléphone / Phone: , # 5607 Télécopieur/ Fax: f.desmeules@umontreal.ca! 1

2 1! 2! 3! 4! 5! 6! 7! 8! 9! 10! 11! 12! 13! 14! 15! 16! 17! 18! 19! 20! 21! 22! 23! Abstract Study Design: Systematic review and meta-analysis. Objectives: To evaluate the efficacy of manual therapy (MT) for patients with rotator cuff (RC) tendinopathy. Background: RC tendinopathy is a highly prevalent musculoskeletal disorder. MT is a common intervention used by physiotherapists for this condition although evidence regarding its efficacy is inconclusive. Methods: A literature search using terms related to shoulder, RC tendinopathy, and MT was conducted in 4 databases to identify randomized controlled trials (RCTs) that compared MT to any other type of intervention to treat RC tendinopathy. RCTs were assessed with the Cochrane Risk of Bias Tool. Meta-analyses or qualitative synthesis of evidence were performed. Results: Twenty-one studies were included. The majority had a high risk of bias with only 5 studies having a score of 69% or greater, indicative of moderate to low risk of bias. A small but statistically significant overall effect for pain reduction of MT compared with a placebo or in addition to another intervention was observed (n=406) which may or may not be clinically important given a mean difference of 1.1(95% confidence interval [CI]: 0.6, 1.6) on a 10 cm visual analog scale (VAS). Adding manual therapy to an exercise program (n=226) significantly decreased pain and this effect may or may not be clinically important (mean difference of 1.0; 95%CI: 0.7, 1.4 on a 10 cm VAS). Based on qualitative analyses it is unclear that MT used alone or added to exercises improves function. Conclusions: For patients with RC tendinopathy, based on low to moderate quality evidence, MT may decrease pain but it is unclear if it could improve function. More methodologically sound studies are needed before more definitive conclusions can be made.! 2

3 24! 25! Level of evidence: Therapy, level 1a-, J Orthop Sports Phys Ther 2015;xxxxx Key words: mobilization, physiotherapy, shoulder impingement syndrome, shoulder pain! 3

4 26! 27! 28! 29! 30! 31! 32! 33! 34! 35! 36! 37! 38! 39! 40! 41! 42! 43! 44! 45! 46! 47! 48! INTRODUCTION! Shoulder pain is highly prevalent and among musculoskeletal disorders, it is the third most common reason for visiting a primary care physician. 17, 35, 47 As many as two thirds of people who have shoulder complaints receive a diagnosis of rotator cuff (RC) tendinopathy. 47 RC tendinopathy often leads to decreased function, 36 lower health-related quality of life, 36 poor sleep quality, 48 and work absenteeism. 41 RC tendinopathy is a broad diagnosis and mounting evidence suggest that diagnoses such as shoulder impingement syndrome, RC tendinitis/tendinosis, as well as subacromial bursitis may be considered as the same clinical entity. 20 Conservative treatment of RC tendinopathy generally includes rest, non-steroidal antiinflammatory drugs, and rehabilitation interventions such as exercise. 20 High level evidence supports exercise as an effective treatment. 21 In conjunction with exercise, physiotherapists often add manual therapy (MT) interventions to address impairments potentially associated with RC tendinopathy. 51 MT interventions have been defined by the International Federation of Orthopaedic Manipulative Physical Therapist (IFOMPT) as skilled hand movements performed by a therapist. 27 Systematic reviews on the efficacy of MT and exercises for the treatment of RC tendinopathy have recently been published 7, 8 with the authors concluding that there is 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 RC tendinopathy was inconclusive. 7, 8 However, in these reviews, only qualitative synthesis of results was performed without pooling of results into meta-analyses. Moreover, many relevant clinical trials were excluded from these reviews 1, 3, 5, 11, 13, 39, 45, 49 and, since the publication of these reviews, new! 4

5 49! 50! 51! trials have also been published. 23, 28, 30 Thus, the aim of this systematic review and meta-analysis was to perform an updated review of the evidence regarding the efficacy of MT, used either alone or in combination with other interventions, for the treatment of RC tendinopathy.! 5

6 52! 53! 54! 55! 56! 57! 58! 59! 60! 61! 62! 63! 64! 65! 66! 67! 68! 69! 70! 71! 72! 73! 74! METHODS Literature search An electronic bibliographical search was conducted in MEDLINE, EMBASE, the Physiotherapy Evidence Database (PEDro), and the Cumulative Index to Nursing & Allied Health Literature (CINAHL), from their date of inception to June A combination of MESH terms and text words was used to identify relevant articles. In addition, a hand search was performed of the reference lists of included articles and previous published reviews (FIGURE 1). Study selection Two reviewers independently reviewed titles and abstracts to identify articles of interest. Consensus of the 2 reviewers was needed to include the studies in the literature review. A third reviewer was available for final determination if consensus was not achieved by the pair of initial reviewers. Articles were included if they met the following inclusion criteria: participants were diagnosed with RC tendinopathy/tendinitis, shoulder impingement syndrome, or subacromial bursitis; participants were adults; a MT intervention was compared with another type of treatment including other MT interventions; the study design was a randomized controlled trial (RCT); the article was published in English or French (FIGURE 1). MT interventions that were considered for inclusion had to be specific hands-on interventions. 27 Included interventions meeting this definition were joint mobilisations, manipulations, specific soft tissue massage techniques, neurodynamic interventions, and mobilizations with movement (MWM) of the shoulder girdle or spine. All types of outcome measures were considered for inclusion. Studies that included! 6

7 75! 76! 77! 78! 79! 80! 81! 82! 83! 84! 85! 86! 87! 88! 89! 90! 91! 92! 93! 94! 95! 96! 97! participants with shoulder pain without further diagnostic information were included for review if it was possible to determine that the majority of participants had RC tendinopathy. Studies were excluded if participants had RC full-thickness tear, calcific tendinopathy, or presented with a postsurgical condition. Data extraction Data and results from the included studies were extracted using a standardized form that documented: characteristics of the participants, diagnostic criteria, interventions, follow-up period, outcome measures, and results (TABLES 1-3). When results were missing or not fully reported, efforts were made to contact the contributing authors to retrieve missing data. Risk of bias appraisal tool The internal validity of the included studies was assessed with the Cochrane Risk of Bias Tool. 25 This tool appraises 5 different biases (selection, performance, attrition, detection, reporting) using 8 different domains (methodological items). A score on a 3-point scale was attributed to each judgement of risk of bias, (0 for a high risk of bias, 1 for an unclear risk of bias, and 2 for a low risk of bias), thus a maximum of 16 points was possible if the study presented a low risk of bias. Two reviewers independently assessed the risk of bias of each study and met to compare ratings. Disagreement was resolved by consensus. If no consensus was reached a third reviewer was available to make final determination. Data analyses A pre-consensus intraclass correlation coefficient (ICC) was calculated on the total score of the! 7

8 98! 99! 100! Cochrane Risk of Bias Tool and interrater agreement was calculated for the 8 risk of bias domains with the kappa ( ) statistic. Statistical analyses were performed with the Statistical Package for Social Sciences (IBM SPSS Statistics 21, Chicago, USA). 101! 102! 103! 104! 105! 106! 107! 108! 109! 110! 111! 112! 113! 114! 115! 116! Results from studies with similar comparators or outcome measures were pooled into metaanalyses. Only meta-analyses without a significant degree of heterogeneity ( 2 P>.10 and I 2 < 60%) were retained for reporting. 24 When quantitative pooling was not performed, results were qualitatively synthesized. The primary analysis compared the overall efficacy of manual therapy to a placebo or to another intervention. Secondary analyses included the comparisons of: 1) MT alone to a placebo, 2) MT with an exercise program to an exercise program, 3) MT alone to another intervention, 4) MT combined with other types of interventions to a placebo or a multimodal intervention, 5) Different types of MT. Mean differences (MD) or standardised mean differences (SMD) with 95% confidence intervals (CI) were calculated using Review Manager (version 5.2). 25 Because the overall number of studies included in the meta-analyses was small and true effect sizes varied between studies, random effects models were used. Alpha level was set at 0.05 to test for overall effect. Funnel plots were not generated because of the small number of trials included in each meta-analysis. 25! 8

9 117! 118! 119! RESULTS From the 32 potentially relevant articles identified after title and abstract review, 21 studies met 1-6, 11-13, 23, 28-30, 37, 39, 43-45, 49, 53, 54 the eligibility criteria following full text review (FIGURE 1). 120! 121! 122! 123! 124! 125! 126! 127! 128! 129! 130! 131! 132! 133! 134! 135! 136! 137! 138! 139! Risk of bias appraisal The mean ± standard deviation (SD) score on the Cochrane Risk of Bias Tool across all studies was 52.7% ± Five studies 5, 13, 29, 30, 49 had a score of at least 69% (11/16), indicating a moderate to low risk of bias. The other 16 studies 1-4, 6, 11, 12, 23, 28, 37, 39, 43-45, 53, 54 scored less than 69%, indicating a high risk of bias. Reviewers agreement on the overall risk of bias score was excellent with an intraclass correlation coefficient of 0.93 (95% CI: 0.84, 0.97). Pre-consensus interrater agreement for individual items of the Cochrane Risk of Bias tool ranged from fair to perfect ( = 0.49, 1.0). 33 Consensus was always achieved between the pair of initial reviewers. None of the studies had a low risk of bias for all the 8 methodological items. Due to the nature of the intervention, blinding of the participant was rarely possible and it was achieved for only 6 of the studies. 5, 13, 28, 29, 39, 49 Blinding of the provider was also impossible because of the nature of 2, 5, 12, 13, 23, 29, 30, 37, 45, the intervention. Blinding of assessors was adequately reported in 11 studies. 49, 54 Risk of reporting bias was present in 18 of the 21 studies because no protocol or trial registration number was provided (FIGURES 2 and 3). Outcome measures Fourteen studies 1-3, 5, 6, 11, 12, 29, 30, 37, 39, used a visual analog scale (VAS) or a numeric pain 4, 5, 11, 23, 28-30, 43, 45 rating scale (NPRS) to measure a variety of pain related outcomes. Nine trials! 9

10 140! 141! 142! 143! 144! 145! 146! 147! 148! 149! 150! 151! 152! 153! 154! 155! 156! 157! 158! 159! 160! 161! 162! used validated functional outcome measures. Shoulder active range of motion (ROM) in flexion or abduction was also commonly used as an outcome measure. 1, 3, 6, 11-13, 23, 29, 37, 43-45, 49 Seven studies did not report treatment effect estimates, only reporting conclusions of statistical testing or results in a graphical form. 1, 2, 6, 11, 12, 44, 53 Efforts made to contact contributing authors to obtain additional results were all unsuccessful. Primary analyses Overall efficacy of manual therapy compared with a placebo or in addition to another intervention Thirteen RCTs assessed the effect of MT compared to a sham treatment or assessed the effect of adding a MT intervention to another intervention such as: exercises, electrotherapy, or education. 1, 2, 4, 6, 12, 29, 30, 37, 39, 43-45, 49 Eleven RCTs 1, 2, 6, 12, 29, 30, 37, 39, assessed treatment effect using pain as an outcome measure and 5 trials used functional changes. 2, 4, 29, 30, 45 1, 2, 6, Ten 12, 29, 30, 37, 39, 43, 45 of the 11 RCTs (n=406) provided results and were pooled to evaluate the efficacy of MT on pain relief. Pooled results demonstrated a significant effect in favor of the MT intervention either when used alone or when used in conjunction with another intervention (10 cm VAS MD: 1.1; 95% CI: 0.6, 1.6); this effect is small but could be considered clinically important (FIGURE 4). 46 For the 5 RCTs (n=206) using functional outcome measures, pooling of the results was attempted but significant heterogeneity was present ( 2 P< and I 2 =93%). Therefore, the results from these trials are presented in the following sections depending on the MT intervention 2, 4, 29, 30, 45 under study and comparators.! 10

11 163! 164! 165! 166! 167! 168! 169! 170! 171! 172! 173! 174! 175! 176! 177! 178! 179! 180! 181! 182! 183! 184! 185! Secondary analyses MT alone compared with a placebo Four clinical trials assessed the efficacy of MT compared with a placebo to address pain. 1, 37, 39, 45 Interventions consisted of shoulder girdle and cervical spine mobilization and manipulations. Results of these 4 trials (n=195) were pooled into a metaanalysis and revealed a significant effect in favor of MT (10 cm VAS MD: 1.0; 95% CI: 0.6, 1.4) (FIGURE 5). The magnitude of the treatment effect is small but could be considered clinically important. 46 Only 1 RCT (n=39) measured function as an outcome measure. Significant differences were observed in change on the Constant-Murley Score (CMS) in favor of the MT group compared to no intervention (CMS MD ± SD: 6.1% ± 2.9; P<.016) and to sham MT (CMS MD ± SD: 1.9% ± 0.92; P<.016). 45 Three trials (n=123) assessed the efficacy of MT on active shoulder ROM, but results were not pooled into a meta-analysis because of high heterogeneity ( 2 P=.010 and I 2 =74%). 1, 45, 49 Results from the trial by Teys et al 49 revealed that MT is effective in increasing pain free abduction in the scapular plane immediately after the intervention (MWM) compared with sham MT (MD: 9.0 o ; 95% CI: 4.3 o, 15.6 o ). Another trial observed a significant immediate increase for active shoulder flexion and abduction in the MT group (flexion MD ± SD: 7.6 o ± 2.5 and abduction MD ± SD: 6.1 o ± 2.9; P<.016) compared with sham MT. 45 However, the third trial by Atkinson et al 1 reported no significant differences in ROM changes between the MT and the placebo at 2 weeks (flexion MD ± SD: 3.2 o ± 1.5 and abduction MD ± SD: 11.3 o ± 1.9 ; P.05).! 11

12 186! 187! 188! 189! 190! 191! 192! 193! 194! 195! 196! 197! 198! 199! 200! 201! 202! 203! 204! 205! 206! 207! 208! Efficacy of adding a MT intervention to an exercise program or to a multimodal rehabilitation program with exercises Seven trials observed effects on pain of MT and exercises compared with exercises alone. Five trials 2, 6, 29, 30, 43 provided result that were pooled into a meta-analysis. The trial by Conroy et Hayes 12 was removed from theses analyses because of significant statistical heterogeneity ( 2 P=.02 and I 2 =64%) and the results of the trial by Senbursa et al 44 were not pooled with the other studies because of missing data. For the remaining 5 studies (n=226) a significant difference was observed for the addition of MT to exercises for overall pain reduction at 4 weeks (10 cm VAS MD: 1.0; 95%CI: 0.7, 1.4) (FIGURE 6). Although the treatment effect is small, it could be clinically important. 46 Outcomes of the 2 trials that were not pooled reported conflicting results. Senbursa et al 44 reported no significant differences when MWM was added to an exercise program for overall pain at 4 and 12 weeks (exact results not reported, P.05). The RCT by Conroy and Hayes 12 compared standard rehabilitation intervention (hot pack, soft tissue mobilization, patient education, and stretching and strengthening exercises) with and without the addition of MT (glenohumeral joint mobilizations). The authors observed statistically and clinically important differences in maximum pain at 4 weeks (10 cm VAS MD: 3.2; P=.008; SD not reported) and in pain on the subacromial compression test (10 cm VAS MD: 2.2; P=.032; SD not reported) favoring the MT group. 46 Six trials (n=173) observed long-term functional outcomes of adding MT to an exercise program but trials could not be pooled into a meta-analyses because of high heterogeneity ( 2 P< and I 2 =95%). Four trials 2, 4, 43, 44 concluded that there was a significant effect in favor of adding! 12

13 209! 210! 211! 212! 213! 214! 215! 216! 217! 218! 219! 220! 221! 222! 223! 224! 225! 226! 227! 228! 229! 230! 231! MT to an exercise program and 2 trials 29, 30 did not observed such benefits. In all trials, soft tissue and joint mobilizations of the shoulder girdle was added to an exercise program. The trials by Senbursa et al 43, 44 observed a significant between- difference at 12 weeks on the Neer Functional Score (exact results not reported, P<.05) and on the modified American Shoulder and Elbow Surgery (mases) score at 4 and 12 weeks (exact results not reported, P<.05). In another trial, authors observed a significant difference at 2 months on a functional questionnaire (MD ± SD [/45]: 2.1 ± 0.48 P=.005). 2 Barbosa et al 4 also reported a significant and clinically important difference on the Disabilities of the Arm, Shoulder and Hand (DASH) outcome measure (MD ± SD: 20.1% ± 18.1; P=.021), and also a significant difference in the CMS scores (MD ± SD: 7.1% ± 15.4; P=.004) between. 42 Finally, no significant differences between on the SPADI scores were observed after 6 weeks in the trial by Kachingwe et al 29 (P.05) or in the trial of Kromer et al 30 either at 5 weeks (SPADI MD: 1.8%; 95%CI: -5.7%, 9.2%) or at 12 weeks (SPADI MD: 0.4%; 95%CI: -5.1%, 6.0%). Four trials (n=151) investigated changes in shoulder ROM when MT was added to an exercise program, pooling of data was only possible for 2 trials 6, 12 because results of 2 trials were incompletely reported. 43, 44 Pooled results demonstrated that there were no significant differences between when MT was added to an exercise program for overall ROM (MD: %CI: -20.6, 8.4 ) (FIGURE 7). The 2 remaining trials, compared changes in ROM (flexion, abduction, external rotation) when manipulation and mobilization of the shoulder girdle and 43, 44 spine was added to an exercise program consisting of shoulder stretching and strengthening. While the first trial demonstrated a significant intergroup difference at 4 weeks in favor of the MT group (exact results not reported, P<.05), the second study did not report such a difference at! 13

14 232! 4 and 12 weeks (exact results not reported, P.05). 233! 234! Efficacy of MT alone compared with another intervention One trial compared MT 235! 236! 237! 238! 239! 240! 241! 242! 243! 244! 245! 246! 247! 248! 249! 250! 251! 252! 253! (mobilizations and manipulations of the shoulder girdle) to corticosteroid injections and to standard care (exercises, massage, and physical agents applications). 53, 54 The mean duration of symptoms in the MT group was significantly longer than in the corticosteroid injection group (exact results not reported, P<.001), but significantly shorter compared with the standard care group (exact results not reported, P<.001). 54 At 2 to 3 years, there were no differences between in terms of symptoms, functional limitations, and perception of being cured. 53 Another RCT compared supraspinatus deep friction massage to therapeutic ultrasound. 3 Results showed that while both improved at 10 days, the MT group had significantly lower pain at rest (10 cm VAS MD: 0.85; P=.014, SD not reported) and significantly greater active abduction ROM than the ultrasound group (MD: 6.7 o ; P=.023, SD not reported), however those changes were not clinically important. Efficacy of MT combined with other types of interventions compared with a placebo or a multimodal intervention Four trials (n=202) assessed the efficacy of MT combined with other interventions compared with a different multimodal intervention for treating RC tendinopathy. 5, 11, 13, 23 High heterogeneity between trials prevented pooling of data for pain ( 2 P=.06 and I 2 >72%) or for function ( 2 P=.02 and I 2 >76%), therefore only a qualitative analysis was performed.! 14

15 254! 255! 256! 257! 258! 259! 260! 261! 262! 263! 264! 265! 266! 267! 268! 269! 270! 271! 272! 273! 274! 275! 276! Bennell and colleagues 5 compared MT (soft tissue massage and mobilizations of the shoulder girdle and spine) in conjunction with an exercise program to a sham ultrasound. These authors observed no significant differences at 11 weeks (10 cm VAS MD: %CI: -0.10, 1.5) or at 22 weeks (10 cm VAS MD: %CI: to 1.7) for pain on movement. Citaker et al 11 also observed no significant changes in pain during movement when comparing shoulder mobilizations and standard care (hot pack and strengthening exercises) to proprioceptive neuromuscular facilitation (PNF) and the same standard care program (10 cm VAS MD ± SD: ± 4.8; P.05). In the trial by Bennell et al, 5 between-group differences in functional changes as evaluated with the SPADI were not significant at 11 weeks (MD: 3.6%; 95%CI: -2.1, 9.4), but were significant at 22 weeks in favor of the MT group (MD: 7.1%; 95%CI: 0.3, 13.9). Results from Citaker et al 11 showed no significant differences in the UCLA shoulder rating scale (P.05). Heredia-Rizo et al 23 found no significant differences in terms of function (DASH MD ± SD: 8.1% ± 18.3; P=.184) at 3 weeks when comparing MT (joint and soft tissue mobilizations of the shoulder girdle and spine) combined with electrotherapy compared with a group receiving electrotherapy and exercises. Citaker et al 11 and Heredia-Rizo et al 23 measured active ROM in abduction and their results demonstrated no difference between both (P.05 and P=.12). However, Djordjevic and colleagues 13 observed significant differences after 10 days on pain-free flexion and abduction in favor of the MT intervention (MWM) combined with shoulder kinesiotaping compared with a supervised exercise program (P<.001).! 15

16 277! 278! 279! 280! 281! 282! 283! 284! 285! 286! 287! 288! 289! 290! Effect of different types of MT A subgroup analysis of the efficacy of different types of MT for pain was also performed. Seven studies 1, 2, 11, 29, 30, 37, 39 in which interventions were only joint (capsular) mobilisations or manipulations (n=329) compared with a placebo or in addition to another intervention were pooled into a meta-analysis. The resulting pooled effect showed a significant difference in favor of MT (10 cm VAS MD: 0.7; 95%CI: 0.2, 1.2) but this effect may not be clinically important (FIGURE 8). 46 Significant heterogeneity prevented us from pooling clinical trials using other types of MT interventions ( 2 P<.10 and I 2 >60%). Effect of adding a thoracic manipulation to MT and exercises One exploratory trial not included in any of the current meta-analyses, that included only 9 participants assessed the efficacy of adding a thoracic spinal manipulation to active and passive glenohumeral joint mobilizations, transverse frictions to the RC tendons, and shoulder strengthening exercises. 28 No statistical analyses were performed within and between, but the thoracic manipulation group tended to require fewer sessions to feel cured.! 16

17 291! 292! 293! DISCUSSION The aim of this systematic review was to assess the efficacy of MT for RC tendinopathy. Twenty-one RCTs were included, with risk of bias established as high to moderate. 294! 295! 296! 297! 298! 299! 300! 301! 302! 303! 304! 305! 306! 307! 308! 309! 310! 311! 312! 313! Primary findings Based on our primary meta-analysis of 10 RCTs, there is low to moderate evidence, that overall, MT either alone or in conjunction with other modalities, may be effective in reducing pain. While pooled results demonstrate a statistically significant difference of 1.1 (95%CI: 0.6, 1.6) on a 10 cm scale, the point estimate is slightly below the previously reported minimal clinically important difference (MCID) of 1.4 cm. 46 However, this MCID is within the 95% CI and therefore we can not exclude the possibility that MT may have a clinically important therapeutic effect. 46 It is also important to highlight that this MCID has been defined for change occurring after 6 weeks of treatment and this does not reflect the exact time period when outcomes were measured in the studies included in our meta-analysis. Secondary findings Secondary analyses include research designs that separate out the effects attributable to manual therapy, which provides evidence of efficacy. Conclusions from those secondary analyses were consistent with those of our primary analysis and lead to similar conclusions. Compared with a placebo, MT alone significantly reduces overall pain. Similar to the results of the primary analysis, while the point estimate of the pooled treatment effect is slightly below the MCID, the MCID is within the limits of the 95% CI of the treatment effect. 46 Therefore we cannot rule out the possibility that although the treatment effect is small, it may be clinically important.! 17

18 314! 315! 316! 317! 318! 319! 320! 321! 322! 323! 324! 325! 326! 327! 328! 329! 330! 331! 332! 333! 334! 335! 336! One study with moderate risk of bias concluded that MT significantly increases function as measured with the CMS compared with placebo; however, the difference between the was small (mean ± SD: 2.0% ± 12) and not considered clinically important. 45 Two trials assessed the effect of MT compared with a placebo for changes in shoulder ROM with only 1 of the trials presenting a significant difference favoring MT, but again the differences were small (3 to 7 ); a third trial assessed pain free shoulder ROM and found a significant difference favoring the MT group which could be clinically important. 1, 45, 49 Therefore, it is unclear whether or not MT results in greater improvement in ROM when compared with a placebo, but it seems that it could however be superior to improve pain-free ROM. There is low evidence that adding MT to an exercise program could improve pain but may not lead to any additional improvement in function. Results from 7 trials 2, 6, 12, 29, 30, 43, 44 of low to moderate quality in which 5 were pooled into a meta-analysis 2, 6, 29, 30, 43 suggest a statistically significant decrease in pain that could be considered clinically important. 46 From a qualitative analysis of 6 low to moderate quality RCTs that assessed functional changes, a positive trend in functional improvement was observed. Results from 4 of the 6 RCTs reported a significant difference in favor of MT but again for 3 of these RCTs the magnitude of the treatment effect was generally small and it is unclear if the magnitude of change is clinically important. Because in these studies, the results were only partially reported or the functional outcome measures used were not formally validated, it limits the strength of the recommendation we can formulate. Changes in shoulder ROM were assessed in 4 trials, results from pooled data of 2 RCTs considered together with those of a third RCT demonstrate no difference between ; thus! 18

19 337! 338! 6, 12, 43, adding MT to an exercises program does not seem to help further improve shoulder ROM ! 340! 341! 342! 343! 344! 345! 346! 347! 348! 349! 350! 351! 352! 353! 354! 355! 356! 357! 358! 359! Six trials compared MT with other intervention to another multimodal intervention or to a placebo 3, 5, 11, 13, 23, 54 and results from these studies generally suggest that adding MT to a multimodal intervention does not improve pain, function, or shoulder ROM. However, because of the heterogeneous nature of the studies, we are unable to make formal conclusions on this topic. In our review, MT interventions were varied, and therefore, it could be argued that the effect may differ depending on the technique used. A secondary analysis including only clinical trials using joint (capsular) mobilisations and manipulations resulted in slightly different results than those from our primary meta-analysis; suggesting that while a statistically significant effect was observed, that effect may not be clinically important. MT interventions in the present review did also vary in terms of intensity and duration. This is certainly a factor that may potentially affect the overall expected treatment effect, but we could not account for this factor in our analyses because the details of the MT interventions, for examples, intensity and duration, were in general poorly described. Another factor that may limit the efficacy of MT is the clinical appropriateness of techniques used. The use of MT may need to be tailored to specific impairments observed in patients with RC tendinopathy. Two important sub of patients that may benefit the most from mobilization and manipulation techniques are patients who have posteroinferior capsular tightness of the glenohumeral joint or decreased cervicothoracic extension. Several studies suggest that in these patients, shoulder kinematics are altered in a manner to promote RC! 19

20 360! 361! 362! 363! 364! 365! 366! 367! 368! 369! 370! 371! 372! 373! 374! 375! 376! 377! 378! 379! 380! 381! 382! tendinopathy. 15, 18, 19, 22, 40, 51, 52 In the included studies, the type of MT intervention was not based on the participants impairments, therefore potentially limiting the overall treatment effect observed. Future trials may want to take into account participants impairment and tailor the MT intervention accordingly. Our conclusions are somewhat different from other published systematic reviews on the same topic. 7, 8, 14, 16, 31, 38, 50 The 2 most recent systematic reviews on this topic stated that there were inconclusive or conflicting results for the efficacy of MT used alone. 9, 26 Our review, which included 5 additional clinical trials, leads to the conclusion that MT alone or in conjunction with another intervention significantly decreases pain by a small but statistically significant amount, although it is unclear whether this pain reduction is clinically important. However it is unclear if MT used alone can improve function. Ho et al 26 specifically concluded that there is no clear evidence of an additional benefit when adding MT to another intervention while we concluded that adding manual therapy to an exercise program results in statistically significant pain decreases, however these changes were small and may or may not be clinically significant. However, until more methodologically sound studies are published on MT, widely accepted interventions such as exercises, which have been proven effective for treating RC tendinopathy, 10, 21, 32, 34 should be preferred. Strengths and limitations of the review Strengths of this review are the use of 4 important databases, the use of the validated Cochrane Risk of Bias Tool, and the strong interrater agreement for the evaluation of the risk of bias. We! 20

21 383! 384! 385! are able to make recommendations on the efficacy of MT for RC tendinopathy, based on quantitative and qualitative analyses. Funnel plots were not generated and thus the risk of publication bias was not assessed and this limits the strength of our conclusions. 386! 387! 388! 389! 390! 391! 392! 393! 394! Unanswered questions and future research Before more definitive conclusions on the efficacy of MT used alone or combined with exercises can be made, RCTs with high methodological quality and larger sample sizes are required. To be more representative of current clinical practice, RCTs that compare specific MT interventions tailored for sub of patients defined by their impairments are also required. To facilitate data synthesis and meta-analyses, we also suggest that standardized and validated diagnostic tests and outcome measures should be used in future trials.! 21

22 395! 396! 397! 398! 399! 400! 401! 402! 403! CONCLUSIONS Based on low to moderate level evidence, MT either used alone or in conjunction with other interventions significantly reduce pain in individuals with RC tendinopathy and this effect may or may not be clinically important. Based on low-level evidence, MT intervention added to an exercise program significantly reduce pain in individuals with RC tendinopathy and this effect may or may not be clinically important. Based on low-level evidence, it is unclear if MT can improve function in the treatment of RC tendinopathy. Future methodologically sound studies could possibly modify the present conclusions.! 22

23 404! 405! KEY POINTS 406! 407! 408! 409! 410! 411! 412! 413! 414! Findings Overall effects of MT and the addition of MT to exercises significantly reduce pain in individuals with RC tendinopathy and this effect may or may not be clinically important. Implication Clinicians should not rely solely on MT to treat RC tendinopathy. RCTs that compare MT interventions for specific sub of patients with RC tendinopathy are required. Caution More methodologically sound studies are needed to further support the present conclusions.! 23

24 415! 416! 417! 418! 419! 420! 421! 422! 423! 424! 425! 426! 427! 428! 429! 430! 431! 432! 433! 434! 435! 436! 437! 438! 439! 440! 441! 442! 443! 444! 445! 446! 447! 448! 449! 450! 451! 452! 453! 454! 455! 456! 457! 458! References 1.! Atkinson!M,!Mathews!R,!Brantingham!JW,!et!al.!A!randomized!controlled!trial!to! assess!the!efficacy!of!shoulder!manipulation!versus!placebo!in!the!treatment!of! shoulder!pain!due!to!rotator!cuff!tendinopathy.!journal(of(the(american(chiropractic( 2.! Bang!MD,!Deyle!GD.!Comparison!of!supervised!exercise!with!and!without!manual! physical!therapy!for!patients!with!shoulder!impingement!syndrome.!the(journal(of( 3.! Bansal!K,!Padamkumar!S.!A!comparative!study!between!the!efficacy!of!therapeutic! ultrasound!and!soft!tissue!massage!(deep!friction!massage)!in!supraspinatus! 4.! Barbosa!R,!Goes!R,!Mazzer!N,!Fonseca!M.!The!influence!of!joint!mobilization!on! tendinopathy!of!the!biceps!brachii!and!supraspinatus!muscles.!brazilian(journal(of( Physical(Therapy.!2008;12:298Z303.! 5.! Bennell!K,!Wee!E,!Coburn!S,!et!al.!Efficacy!of!standardised!manual!therapy!and!home! exercise!programme!for!chronic!rotator!cuff!disease:!randomised!placebo!controlled! trial.!bmj.!2010;340:c2756zc2756.! 6.! Bialoszewski!D,!Zaborowski!G.!Usefulness!of!Manual!Therapy!in!the!Rehabilitation! of!patients!vvith!chronic!rotator!cuff!injuries.!preliminary!report.!ortopedia( 7.! Braun!C,!Bularczyk!M,!Heintsch!J,!Hanchard!NCA.!Manual!therapy!and!exercises!for! shoulder!impingement!revisited.!physical(therapy(reviews.!2013;18:263z284.! 8.! Braun!C,!Hanchard!NCA.!Manual!therapy!and!exercise!for!impingementrelated! shoulder!pain.!physical(therapy(reviews.!2010;15:62z83.! 9.! Bronfort!G,!Haas!M,!Evans!R,!Leininger!B,!Triano!J.!Effectiveness!of!manual! therapies:!the!uk!evidence!report.!chiropractic(&(osteopathy.!2010;18:3.! 10.! Bruce!A.!The!Diagnosis!and!Management!of!Soft!Tissue!Shoulder!InjuriesZ!Best! Practice!EvidenceZBased!Guideline.!New!Zealand!Guidelines!Group;!2004.! 11.! Citaker!S,!Taskiran!H,!Akdur!H,!Onel!Arabaci!U,!Ekici!G.!Comparison!of!the! mobilization!and!proprioceptive!neuromuscular!facilitation!methods!in!the! 202.!2005;! 12.! Conroy!DE,!Hayes!KW.!The!effect!of!joint!mobilization!as!a!component!of! comprehensive!treatment!for!primary!shoulder!impingement!syndrome.!the(journal( 13.! Djordjevic!OC,!Vukicevic!D,!Katunac!L,!Jovic!S.!Mobilization!with!movement!and! kinesiotaping!compared!with!a!supervised!exercise!program!for!painful!shoulder:! results!of!a!clinical!trial.!journal(of(manipulative(and(physiological(therapeutics.! 2012;35:454Z463.! 14.! Faber!E,!Kuiper!JI,!Burdorf!A,!Miedema!HS,!Verhaar!JA.!Treatment!of!impingement! syndrome:!a!systematic!review!of!the!effects!on!functional!limitations!and!return!to! work.!journal(of(occupational(rehabilitation.!2006;16:7z25.!! 24

25 459! 460! 461! 462! 463! 464! 465! 466! 467! 468! 469! 470! 471! 472! 473! 474! 475! 476! 477! 478! 479! 480! 481! 482! 483! 484! 485! 486! 487! 488! 489! 490! 491! 492! 493! 494! 495! 496! 497! 498! 499! 500! 501! 502! 503! 504! 15.! Gates!JJ,!Gupta!A,!McGarry!MH,!Tibone!JE,!Lee!TQ.!The!effect!of!glenohumeral! internal!rotation!deficit!due!to!posterior!capsular!contracture!on!passive! glenohumeral!joint!motion.!the(american(journal(of(sports(medicine.!2012;40:2794z 2800.! 16.! Green!S,!Buchbinder!R,!Hetrick!Sarah!E.!Physiotherapy(interventions(for(shoulder( pain.!john!wiley!&!sons,!ltd;!2003.! 17.! Greving!K,!Dorrestijn!O,!Winters!JC,!et!al.!Incidence,!prevalence,!and!consultation! rates!of!shoulder!complaints!in!general!practice.!scandinavian(journal(of( rheumatology.!2012;41:150z155.! 18.! Grossman!MG,!Tibone!JE,!McGarry!MH,!Schneider!DJ,!Veneziani!S,!Lee!TQ.!A! cadaveric!model!of!the!throwing!shoulder:!a!possible!etiology!of!superior!labrum! anteriorztozposterior!lesions.!the(journal(of(bone(and(joint(surgery.(american(volume.! 2005;87:824Z831.! 19.! Gumina!S,!Di!Giorgio!G,!Postacchini!F,!Postacchini!R.!Subacromial!space!in!adult! patients!with!thoracic!hyperkyphosis!and!in!healthy!volunteers.!la(chirurgia(degli( organi(di(movimento.!2008;91:93z96.! 20.! diagnosis,(assessment(and(physiotherapy(management(of(shoulder(impingement( syndrone.!london,!uk:!chartered!society!of!physiotherapy;!2004.! 21.! Hanratty!CE,!McVeigh!JG,!Kerr!DP,!et!al.!The!Effectiveness!of!Physiotherapy! Exercises!in!Subacromial!Impingement!Syndrome:!A!Systematic!Review!and!MetaZ Analysis.!Seminars(in(Arthritis(and(Rheumatism.!2012;42:297Z316.! 22.! Harryman!DT,!2nd,!Sidles!JA,!Clark!JM,!McQuade!KJ,!Gibb!TD,!Matsen!FA,!3rd.! Translation!of!the!humeral!head!on!the!glenoid!with!passive!glenohumeral!motion.! The(Journal(of(bone(and(joint(surgery.(American(volume.!1990;72:1334Z1343.! 23.! HerediaZRizo!AM,!LópezZHervás!A,!HerreraZMonge!P,!GutiérrezZLeonard!A,!PiñaZ Pozo!F.!Shoulder!functionality!after!manual!therapy!in!subjects!with!shoulder! impingement!syndrome:!a!case!series.!journal(of(bodywork(and(movement( Therapies.!2012;! 24.! Higgins!JP,!Thompson!SG,!Deeks!JJ,!Altman!DG.!Measuring!inconsistency!in!metaZ analyses.!bmj((clinical(research(ed.).!2003;327:557z560.! 25.! Higgins!JPT,!Green!S.!Cochrane(Handbook(for(Systematic(Reviews(of(Interventions( Version(5.1.0([updated(March(2011].!The!Cochrane!Collaboration;!2011.! 26.! Ho!CC,!Sole!G,!Munn!J.!The!effectiveness!of!manual!therapy!in!the!management!of! musculoskeletal!disorders!of!the!shoulder:!a!systematic!review.!manual(therapy.! 2009;14:463Z474.! 27.! International!Federation!of!Orthopaedics!Manipulative!Physical!Therapist.! Standards!Committee!Glossary.!Available!at:! February!22th!2014,!! 28.! Janse,!Atkins!E.!Does!thoracic!manipulation!increase!shoulder!range!of!movement!in! patients!with!subacromial!impingement!syndrome?!a!pilot!study.!international( Musculoskeletal(Medicine.!2012;34:101Z107.! 29.! Kachingwe!AF,!Phillips!B,!Sletten!E,!Plunkett!SW.!Comparison!of!manual!therapy! techniques!with!therapeutic!exercise!in!the!treatment!of!shoulder!impingement:!a! randomized!controlled!pilot!clinical!trial.!j(man(manip(ther.!2008;16:238z247.!! 25

26 505! 506! 507! 508! 509! 510! 511! 512! 513! 514! 515! 516! 517! 518! 519! 520! 521! 522! 523! 524! 525! 526! 527! 528! 529! 530! 531! 532! 533! 534! 535! 536! 537! 538! 539! 540! 541! 542! 543! 544! 545! 546! 547! 548! 549! 30.! Kromer!TO,!de!Bie!RA,!Bastiaenen!CH.!Physiotherapy!in!patients!with!clinical!signs! of!shoulder!impingement!syndrome:!a!randomized!controlled!trial.!journal(of( rehabilitation(medicine(:(official(journal(of(the(uems(european(board(of(physical(and( Rehabilitation(Medicine.!2013;45:488Z497.! 31.! Kromer!TO,!Tautenhahn!UG,!de!Bie!RA,!Staal!JB,!Bastiaenen!CH.!Effects!of! physiotherapy!in!patients!with!shoulder!impingement!syndrome:!a!systematic! review!of!the!literature.!journal(of(rehabilitation(medicine(:(official(journal(of(the( UEMS(European(Board(of(Physical(and(Rehabilitation(Medicine.!2009;41:870Z880.! 32.! Kuhn!JE.!Exercise!in!the!treatment!of!rotator!cuff!impingement:!a!systematic!review! and!a!synthesized!evidencezbased!rehabilitation!protocol.!journal(of(shoulder(and( 33.! Landis!JR,!Koch!GG.!The!measurement!of!observer!agreement!for!categorical!data.! Biometrics.!1977;33:159Z174.! 34.! Littlewood!C,!Ashton!J,!ChanceZLarsen!K,!May!S,!Sturrock!B.!Exercise!for!rotator!cuff! 35.! Luime!JJ,!Koes!BW,!Hendriksen!IJ,!et!al.!Prevalence!and!incidence!of!shoulder!pain!in! the!general!population;!a!systematic!review.!scandinavian(journal(of(rheumatology.! 2004;33:73Z81.! 36.! MacDermid!JC,!Ramos!J,!Drosdowech!D,!Faber!K,!Patterson!S.!The!impact!of!rotator! cuff!pathology!on!isometric!and!isokinetic!strength,!function,!and!quality!of!life.! Journal(of(shoulder(and(elbow(surgery(/(American(Shoulder(and(Elbow(Surgeons(...([et( al.].!2004;13:593z598.! 37.! McClatchie!L,!Laprade!J,!Martin!S,!Jaglal!SB,!Richardson!D,!Agur!A.!Mobilizations!of! the!asymptomatic!cervical!spine!can!reduce!signs!of!shoulder!dysfunction!in!adults.! Manual(therapy.!2009;14:369Z374.! 38.! Michener!LA,!Walsworth!MK,!Burnet!EN.!Effectiveness!of!rehabilitation!for!patients! with!subacromial!impingement!syndrome:!a!systematic!review.!journal(of(hand( therapy(:(official(journal(of(the(american(society(of(hand(therapists.!2004;17:152z 164.! 39.! Munday!SL,!Jones!A,!Brantingham!JW,!Globe!G,!Jensen!M,!Price!JL.!A!randomized,! singlezblinded,!placebozcontrolled!clinical!trial!to!evaluate!the!efficacy!of! chiropractic!shoulder!girdle!adjustment!in!the!treatment!of!shoulder!impingement! 2007;! 40.! Myers!JB,!Laudner!KG,!Pasquale!MR,!Bradley!JP,!Lephart!SM.!Glenohumeral!range!of! motion!deficits!and!posterior!shoulder!tightness!in!throwers!with!pathologic! internal!impingement.!the(american(journal(of(sports(medicine.!2006;34:385z391.! 41.! Osteras!H,!Arild!Torstensen!T,!Arntzen!G,!B!SO.!A!comparison!of!work!absence! periods!and!the!associated!costs!for!two!different!modes!of!exercise!therapies!for! patients!with!longstanding!subacromial!pain.!journal(of(medical(economics.! 2008;11:371Z381.! 42.! Roy!JZS,!Esculier!JZF.!Psychometric!evidence!for!clinical!outcome!measures!assessing! shoulder!disorders.!physical(therapy(reviews.!2011;16:331z346.! 43.! Senbursa!G,!Baltacı!G,!Atay!A.!Comparison!of!conservative!treatment!with!and! without!manual!physical!therapy!for!patients!with!shoulder!impingement!! 26

27 550! 551! 552! 553! 554! 555! 556! 557! 558! 559! 560! 561! 562! 563! 564! 565! 566! 567! 568! 569! 570! 571! 572! 573! 574! 575! 576! 577! 578! 579! 580! 581! 582! 583! 584! 585! syndrome:!a!prospective,!randomized!clinical!trial.!knee(surg(sports(traumatol( Arthr.!2007;15:915Z921.! 44.! Senbursa!G,!Baltacı!G,!Atay!A.!The!effectiveness!of!manual!therapy!in!supraspinatus! tendinopathy.!acta(orthopaedica(et(traumatologica(turcica.!2011;45:162z167.! 45.! Surenkok!O,!Aytar!A,!Baltaci!G.!Acute!effects!of!scapular!mobilization!in!shoulder! dysfunction:!a!doublezblind!randomized!placebozcontrolled!trial.!journal(of(sport( rehabilitation.!2009;18:493z501.! 46.! Tashjian!RZ,!Deloach!J,!Porucznik!CA,!Powell!AP.!Minimal!clinically!important! differences!(mcid)!and!patient!acceptable!symptomatic!state!(pass)!for!visual! analog!scales!(vas)!measuring!pain!in!patients!treated!for!rotator!cuff!disease.! Journal(of(Shoulder(and(Elbow(Surgery.!2009;18:927Z932.! 47.! Tekavec!E,!Joud!A,!Rittner!R,!et!al.!PopulationZbased!consultation!patterns!in! patients!with!shoulder!pain!diagnoses.!bmc(musculoskeletal(disorders.!2012;13:238.! 48.! Tekeoglu!I,!Ediz!L,!Hiz!O,!Toprak!M,!Yazmalar!L,!Karaaslan!G.!The!relationship! between!shoulder!impingement!syndrome!and!sleep!quality.!european(review(for( medical(and(pharmacological(sciences.!2013;17:370z374.! 49.! Teys!P,!Bisset!L,!Vicenzino!B.!The!initial!effects!of!a!Mulligan's!mobilization!with! movement!technique!on!range!of!movement!and!pressure!pain!threshold!in!painz limited!shoulders.!manual(therapy.!2008;13:37z42.! 50.! Trampas!A,!Kitsios!A.!Exercise!and!manual!therapy!for!the!treatment!of! impingement!syndrome!of!the!shoulder:!a!systematic!review.!physical(therapy( 51.! Tyler!TF,!Nicholas!SJ,!Lee!SJ,!Mullaney!M,!McHugh!MP.!Correction!of!posterior! shoulder!tightness!is!associated!with!symptom!resolution!in!patients!with!internal! impingement.!the(american(journal(of(sports(medicine.!2010;38:114z119.! 52.! Tyler!TF,!Nicholas!SJ,!Roy!T,!Gleim!GW.!Quantification!of!posterior!capsule!tightness! and!motion!loss!in!patients!with!shoulder!impingement.!the(american(journal(of( sports(medicine.!2000;28:668z673.! 53.! Winters!JC,!Jorritsma!W,!Groenier!KH,!Sobel!JS,!MeyboomZde!Jong!B,!Arendzen!HJ.! Treatment!of!shoulder!complaints!in!general!practice:!long!term!results!of!a! randomised,!single!blind!study!comparing!physiotherapy,!manipulation,!and! corticosteroid!injection.!bmj.!1999;318:1395z1396.! 54.! Winters!JC,!Sobel!JS,!Groenier!KH,!Arendzen!HJ,!Jong!BMZd.!Comparison!of! physiotherapy,!manipulation,!and!corticosteroid!injection!for!treating!shoulder! complaints!in!general!practice:!randomised,!single!blind!study.!bmj.!1997;314:1320.!! 27

28 586!! 587! TABLE 1. Characteristics of the included studies for the efficacy of MT alone Study Participants Interventions Number of participants Follow up period Atkinson et al 2008 Bansal and Deyle 2010 McClatchie et al 2009 n=60 Gender Male: 43 Female: 17 Mean age: 41.8 Diagnosis: 3 out of 4 positive tests : pain on palpation of the greater tuberosity or the anterior acromion, painful arc in abduction, empty can test n=40 Gender Male: 21 Female: 19 Mean age: 30.6 ± 5.4 Diagnosis: pain on palpation of the greater tuberosity, painful resisted abduction, positive empty can test n=21 Gender Male: 7 Female: 14 Mean age: 49.8 ± 9.8 Diagnosis: painful arc in shoulder abduction No improvement following previous physiotherapy treatments Group 1: MT (manipulation of the glenohumeral or acromioclavicular joint) Group 2: Placebo (Detuned laser) Group 1: MT (Supraspinatus transverse frictions) Group 2: Ultrasound Cross-over trial Group 1: MT (lateral glides of C5, C6, and C7) Group 2: Sham MT (hands on patient, no movement) n=30 n=30 n=20 n=20 n=21 n=21 Outcome measure (units) 2 weeks NPRS (/100 mm) ROM in flexion ( 0 ) ROM in abduction ( 0 ) Pain pressure threshold on the anterior shoulder (kg/cm 2 ) 10 days Pain at rest (10 cm VAS) same day Active ROM in abduction ( 0 ) Pain on active abduction (10 cm VAS) Total range of the painful arc for abduction ( 0 ) Main results 1: 19.9 ± 13.9, P=.000 2: 10.6 ± 14.7, P=.002 Difference between : 9.3 ± 3.69, P= : 14.5 ± 5.9, P=.000 2: 11.2 ± 6.0, P=.002 Difference between : 3.2 ± 1.5, P.05 1: 18.4 ± 8.0, P=.000 2: 7.1 ± 6.8, P=.047 Difference between : 11.3 ± 1.9, P.05 Statistically significant difference between favoring group 1, P<.001 1: 4.4, P<.001 2: 3.5, P<.001 Difference between : 0.85, P=.014 1: 32.6, P<.001 2: 25.9, P<.001 Difference between : 6.7, P=.023 1: 1.3 ± 1.1, P<.001 2: 0.2 ± 0.6, P=.078 Difference between : 1.1, P= : 12.5 ± 15.6, P=.002 2: 8.8 ± 12.7, P=.005 No statistical test was performed between Risk of bias score 8/16 6/16 10/16! 28

29 Munday et al 2012 Surenkok et al 2009 n=30 Gender Male: 16 Female: 14 Mean age: 22.5 Diagnosis: 3 out of 5: pain on palpation of the greater tuberosity or the anterior acromion, painful arc in abduction, positive Neer impingement sign or Hawkins Kennedy test n=39 Gender Male: 17 Female: 22 Mean age: 54.3 ± 14.1 Diagnosis: limited painful abduction in scapular plane Individual diagnoses: impingement n=22 tenosynovitis n=10 frozen shoulder n=7 Group 1: MT (manipulation of the glenohumeral joint, acromioclavicular joint, ribs, and scapula) Group 2: Placebo (detuned ultrasound) Group 1: MT (scapula mobilization) Group 2: Sham MT Group 3: No intervention n=15 n=15 n=13 n=13 n=13 4 weeks Pain (10 cm VAS) Immediately after treatment Pain pressure threshold on the anterior shoulder (kg/cm 2 ) Short form McGill Pain questionnaire Pain at rest and activity (100 mm VAS) CMS (%) ROM in active flexion ( 0 ) ROM in active abduction ( 0 ) 1: 2.7 ± 2.4, P<.05 2: 1.9 ± 2.3, P<.05 Difference between : 0.80 ± 0.86, P=.019 1: 1.9 ± 2.1, P<.05 2: 0.6 ± 2.6, P<.05 Difference between : 1.3 ± 3.34, P=.014 1: 10.7 ± 22.1, P<.05 2: 24.1 ± 16.6, P<.05 Difference between : ± 27.6, P<.005 1: 0.46 ± 9.2 2: 0.16 ± : 0.92 ± 8.76 No statistically significant differences between P : 2.2 ± 2.2 2: 0.31 ± 2.3 3: 3.77 ± Statistically significant differences between post intervention favoring group 1 over group 2 (1.92 ± 0.92) and 3 (6.1 ± 3.1), P<.016 1: 7.8 ± 6.3 2: 0.23 ± 6.6 3: 1.7 ± 8.5 Statistically significant differences between favoring group 1 over group 2 (7.6 ± 2.5) and 3 (6.1 ± 2.9), P<.016 1: 5.6 ± 4.6 2: 0.52 ± 4.2 3: 1.8 ± 1.3 Statistically significant differences between favoring group 1 over group 2 (5.1 ± 1.7) and 3 (3.8 ± 1.3), P< /16 8/16! 29

30 Scapula upward rotation at rest ( 0 ) Statistically significant differences between favoring group 1 over group 2 and 3, P<.016 Teys et al 2008 n=24 Gender Male: 11 Female: 13 Mean age: 46.1 ± 9.8 Diagnosis: limited painful shoulder abduction Cross-over trial Group 1: MT (mobilization with movement) Group 2: Sham MT (sham mobilization with movement) Group 3: No intervention n=24 n=24 n=24 Immediately after treatment Pressure pain threshold (kpa) Pain free abduction in scapular plane ( 0 ) 1: 62.6, 95%CI 33.6, : 25.9, 95%CI 0.2, : 20.0, 95%CI -1.5, 41.5 Difference between 1 and 2: 36.7, 95%CI -0.01, and 3: 42.6, 95%CI 8.3, : 15.6, 95%CI 10.1, : 3.9, 95%CI -0.1, 7.9 3: 0.27, 95%CI -0.24, 3 Difference between 1 and 2: 11.7, 95%CI 5.3, and 3:15.3, 95%CI 9.5, 21.1 Abbreviations: CMS, Constant-Murley score, pain and function questionnaire, isometric abduction strength, active ROM in flexion, abduction, internal and external rotation, higher scores indicate a greater level of function; MD, mean difference; MT, manual therapy; NPRS, numerical pain rating scale; ROM, range of motion; VAS, visual analog scale 12/16! 30

31 TABLE 2. Characteristics of the included studies for the efficacy of adding a MT intervention to exercises or to a multimodal rehabilitation program that includes exercise. Study Participants Interventions Number of participants Bang et al 2000 Barbosa et al 2008 Bialoszewski et al 2011 Conroy and Hayes 1998 n=52 Gender Male: 30 Female: 22 Mean age: 43.4 ± 9.1 Diagnosis: Positive Neer or Hawkins Kennedy test and painful active abduction or painful resisted movement (abduction, or internal or external rotation) n=14 Gender Male: 5 Female: 9 Mean age: 46.1 ± 7.6 Diagnosis: pain on palpation and at least 1 positive test : Speed, Yeargason, Jobe n=30 Gender Male: 18 Female :22 Mean age: 51.3 Diagnosis: positive Jobe's relocation test and painful arc in abduction n=14 Gender Male: 8 Female: 6 Mean age: 52.8 ± 19.3 Diagnosis: Pain on the Group 1: MT (mobilization and manipulation of the upper quarter) and supervised exercises Group 2: Supervised exercises (flexibility and strengthening of the shoulder) Group 1: MT (Glenohumeral mobilization) with ultrasound and exercises Group 2: Ultrasound and eccentric exercises Group 1: MT (anteroposterior and rollglide glenohumeral mobilization, mobilization with movement, deep transverse friction) and standard care Group 2: Standard care (TENS, ultrasound, strengthening, and passive exercises) Group 1: MT (anterior, posterior, inferior mobilization or traction of the glenohumeral joint) and standard care Group 2: Standard care (hot packs, active ROM, stretching and n=27 n=23 n=7 n=7 n=15 n=15 n=7 n=7 Follow up period Outcome measure (units) 6-7 weeks Pain during activities (10 cm VAS) Isometric strength (abduction and internal and external rotation) (Newton) 2 months Functional questionnaire (/45) Main results 1: 4.1 ± : 2.0 ± 0.77 Difference between : 2.1 ± 0.19, P=.0017 Statistically significant difference between favoring group 1, P<.001 1: 9.8 ± 1.3 2: 4.7 ± 2.0 Difference between : 5.1 ± 0.9, P= weeks DASH (%) 1: 40.5 ± 13.6, P<.001 2: 19.9 ± 11.9, P<.001 Differences between : 20.0 ± 18.1, P=.021 CMS (%) 1: 21.5 ± 11.6, P<.001 2: 14.5 ± 10.0, P<.001 Differences between : 7.0 ± 15.3, P=.004 not reported Pain on functional tests (10 cm VAS) Active ROM in abduction ( ) 3 weeks Maximum pain intensity in the last 24 hours (100 mm VAS) Maximum pain intensity with subacromial 1: 5.3 ± 2.9 2: 3.2 ± 1.3 Difference between : 2.1 ± 3.18, P<.005 1: 19.3 ± : 8.3 ± 5.2 Difference between : 11.0 ± 13.9, P<.007 1: 37.0 ± 11.8, P=.005 2: 2.2 ± 14.9, P=.823 Difference between : 34.9 ± 7.0, P=.008 1: 27.0 ± 10.5, P=.003 2: 11.1 ± 15.0, P=.842 Risk of bias score 8/16 6/16 6/16 9/16! 31

32 Kachingwe et al 2008 Kromer et al 2013 superolateral shoulder and 1 of the following: decreased active ROM in flexion, painful subacromial compression test, or limited functional movement patterns in an elevated position n=33 Gender Male:17 Female: 16 Mean age: 46.4 Diagnosis: superolateral shoulder pain and 2 out of 4 positive tests: Neer, Hawkins Kennedy, painful limitation of active shoulder elevation, pain or limitation with the functional movement patterns n=90 Gender Male: 44 Female: 46 Mean age: 51.8 ± 11.2 Diagnosis: 1 positive test: Neer, Hawkins Kennedy, or painful arc, and pain on 1 resisted test: external rotation, internal rotation, abduction, or flexion strengthening shoulder exercises, soft tissue mobilization, and patient education) Group 1: MT (posterior, anterior, inferior, or traction mobilizations of the glenohumeral joint) and supervised exercises Group 2: MT (mobilization with movement) and supervised exercises Group 3: supervised exercises (muscle strengthening, capsule stretching, postural correction) Group 4: Unsupervised exercises, advice Group 1: MT (mobilization and manipulation of the shoulder girdle, upper thoracic and cervical spine) for 5 weeks and individually adapted exercises Group 2: individually adapted exercises (shoulder and neck stretches, strengthening of the shoulder) n=9 n=9 n=8 n=7 n=46 n=44 compression test (100 mm VAS) Active ROM in abduction, flexion, abduction in scapular plane, internal and external rotation ( ) 6 weeks Change in maximum pain in the last 24 hours (%) Change in SPADI (%) Change in active pain free abduction in scapular plane (%) 5 weeks Mean weekly pain score (11 points VNRS) SPADI (%) Patient global impression of change 12 weeks Mean weekly pain score (11 points VNRS) SPADI (%) Difference between : 15.9 ± 6.9, P=.032 No difference between P Difference within 1: 4.4 ± 3.8, P<.001 2: 5.5 ± 3.1, P<.001 3: 2.0 ± 11.2, P<.001 4: 1.1 ± 11.9, P<.001 No differences between, P.05 Difference within 1: 56.7 ± 29.8, P<.001 2: 55.5 ± 20.1, P<.001 3: 61.6 ± 35.9, P<.001 4: 34.2 ± 58.9, P<.001 No differences between, P.05 Difference within 1: 2.5 ± 88.8, P<.001 2: 66.5 ± 28.1, P<.001 3: 19.8 ± 70.3, P<.001 4: 29.8 ± 49.0, P<.001 No differences between, P.05 1: 2.3, 95%CI 1.7, 2.8 2: 1.6, 95%CI 1.0, 2.3 Differences between : 0.6, 95%CI -0.2, 1.5 1: 16.2, 95%CI 10.8, : 14.4, 95%CI 9.2, 19.6 Differences between : 1.8, 95%CI -5.7, 9.2 Relative risk of slightly and much better overall change: 1.06, 95%CI 0.93, : 0.6, 95%CI 0.1, 1.0 2: 1.1, 95%CI 0.5, 1.5 Differences between : -0.4, 95%CI -1.1, 0.2 1: 7.5, 95%CI 3.7, : 7.0, 95%CI 2.8, 11.2 Differences between : 11/16 12/16! 32

33 0.4, 95%CI -5.1, 6 Senbursa et al 2007 Senbursa et al 2011 n=30 Gender proportion not reported Mean age: 48.8 ± 10.8 Diagnosis: Positive Neer impingement sign n=77 Gender proportion not reported Mean age: 48.8 Diagnosis: Positive Neer impingement sign or Hawkins Kennedy test Group 1: MT (glenohumeral and scapula mobilization, deep friction massage to supraspinatus tendon, radial nerve stretch, PNF) and exercises Group 2: Exercises (stretching and strengthening of the shoulder) Group 1: MT (glenohumeral and scapula mobilization, deep friction massage to supraspinatus tendon, radial nerve stretch, PNF) and supervised exercises Group 2: Supervised exercises (stretching and strengthening) Group 3:Home based rehabilitation program (same exercises as supervised exercises) n=15 n=15 n=22 n=25 n=30 4 weeks Global pain (at rest, at night, with movement) (10 cm VAS) ROM (flexion, abduction, external rotation) ( ) 13 weeks Neer functional assessment questionnaire 4 and 12 weeks Global pain (at rest, at night, with movement) (10 cm VAS) MASES score (%) ROM in flexion, abduction, external rotation ( ) Isometric strength in flexion, abduction, internal and external rotation (5 points 1: 4.7 ± 0.83, P<.05 2: 3.6 ± 0.58, P<.05 Difference between : 1.1 ± 2.9, P<.05 1: Statistically significant differences, P<.05 2: No statistically significant differences, P.05 Statistically significant difference between favoring group 1, P<.05 1: Statistically significant differences, P<.05 2: No statistically significant differences, P.05 Statistically significant difference between favoring group 1, P<.05 Statistically significant differences within all, P<.05 No statistically significant differences between, P.05 At week 4 Statistically significant differences within all, P<.05 Statistically significant differences between favoring group 1 over group 2 and 3, P<.05 At week 12 Statistically significant differences within group 1, P<.05 No statistically significant differences within 2 and 3, P<.05 Statistically significant differences between favoring group 1 over group 2 and 3, P<.05 Statistically significant differences within all, P<.05 No statistically significant differences between, P.05 Statistically significant differences within all, P<.05 No statistically significant differences between, P.05 5/16 4/16! 33

34 manual muscle testing) Abbreviation: DASH, Disability of the arm shoulder and hand, self reported disability questionnaire, higher scores indicate a greater level of disability; CMS, Constant-Murley score, pain and function questionnaire, isometric abduction strength, active ROM in flexion, abduction, internal and external rotation, higher scores indicate a greater level of function; MASES, Modified American Shoulder and Elbow Surgeons, self-assessment of function and pain, higher scores indicate a greater level of disability; MD, mean difference; MT, manual therapy; NPRS : numerical pain rating scale; PNF, proprioceptive neuromuscular facilitation; ROM, range of motion; SPADI : Shoulder Pain and Disability Index, self-assessment of symptoms and function of the shoulder, higher scores indicate a greater level of disability; VAS, visual analog scale; VNRS, visual numeric rating scale! 34

35 TABLE 3. Characteristics of the included studies for the efficacy of MT combined with other types of interventions compared with multimodal interventions Study Participants Interventions Number of participant s Follow up period Outcome measure (units) Main results Risk of bias score Bennell et al 2010 Citaker et al 2005 n=120 Gender Male: 30 Female: 22 Mean age: 60.1± 16.0 Diagnosis: Positive Neer impingement sign and pain on active external rotation or abduction n=40 Gender proportion not reported Mean age: 54.1 ± 13.3 Diagnosis: patient with prior diagnosis of RC tendinopathy made clinically and radiologically Group 1: MT and exercises (anteroposterior and inferior mobilization of the glenohumeral joint, posteroanterior mobilization of C5 to T1, soft tissue massage of the shoulder, posture correction with tape, stretching and strengthening of the shoulder) Group 2: Placebo (sham ultrasound) Group 1: MT (mobilization) and standard care Group 2: Proprioceptive neuro-muscular facilitation and standard care (hot packs and strengthening and pendulum exercises) n=59 n=61 n=20 n=20 11 weeks Average pain on movement (11 point NPRS) SPADI (%) SF-36 (%) Participant self perceived overall change 22 weeks Average pain on movement (11 point NPRS) not reported SPADI (%) SF-36 (%) Participant self perceived overall change Pain (at night and day with and without motion) (10 cm VAS) ROM (flexion, abduction) ( ) UCLA shoulder score (/35) 1: 2.1 ± 2.6, P<.001 2: 1.3 ± 2.2, P<.001 Difference between : 0.7, 95%CI -0.1, 1.5 1: 16.6 ± 17.7, P<.001 2: 12.7 ± 16.3, P<.001 Difference between : 3.6, 95%CI -2.1, 9.4 1: 11.7 ± 26.5, P<.001 2: 6.1 ± 17.4, P<.001 Difference between : 5.7, 95%CI -2.1, 13.6 Relative risk of much better overall change favoring group 1: 1.43, 95%CI 0.87, : 2.6 ± 2.9, P<.001 2: 1.6 ± 2.4, P<.001 Difference between : 0.91, 95%CI -0.03, 1.7 1: 22.4 ± 22.0, P<.001 2: 15.6 ± 17.8, P<.001 Difference between : 7.1, 95%CI 0.3, : 10.8 ± 25.0, P<.001 2: 4.7 ± 22.3, P<.001 Difference between : 6.3, 95%CI -2.0, 14.5 Relative risk of much better overall change favoring group 1: 1.39, 95%CI 0.94, 2.03 Statistically significant differences within both, P<.005 No statistically significant differences between, P.05 Statistically significant differences within both, P<.001 No statistically significant differences between, P.05 1: 16.6, P= : 14.1, P=.0001 No statistically significant 14/16 6/16! 35

36 differences between, P.05 Djordjevic et al 2012 Heredia-Rizo et al 2012 Janse et al 2010 Winters et al. 1997, 1999 n=20 Gender Male: 7 Female: 13 Mean age: 53.0 ± 8.6 Diagnosis: combination of positive impingement tests: Neer, Hawkins Kennedy, Speed, Empty can n=22 Gender Male: 13 Female: 9 Mean age: 58 ±10.8 Diagnosis: at least 2 positive impingement tests out of 3: Neer, Jobe, Yeargason n=9 Gender Male: 6 Female: 3 Mean age: 57.6 Diagnosis: 2 of 3 positive tests: Hawkins Kennedy, Neer, painful arc test n=172 Gender Male: 78 Female: 96 Group 1: MT (mobilization with movement) and kinesiotaping Group 2: Supervised exercise program (painfree active ROM and strengthening of the shoulder) Group 1: MT (cervical, thoracic and glenohumeral mobilization, soft tissue mobilization of the shoulder girdle) and electrotherapy Group 2: Electrotherapy (laser, ultrasound, TENS) and exercises (passive, active, active assisted mobilization, PNF, proprioceptive exercises, pendular movement) Group 1: MT (manipulation of the thoracic spine plus regular MT at the shoulder girdle) and exercises Group 2: Regular MT at the shoulder girdle and exercises (active or passive mobilization of glenohumeral joint and transverse frictions to the rotator cuff, strengthening of the rotator cuff and lower trapezius) Group 1: MT (mobilization and manipulation cervical spine, upper thoracic spine, upper ribs, n=10 n=10 n=11 n=11 n=6 n=3 10 days Painfree flexion ( ) Painfree abduction ( ) 1: ± : 17.0 ± 26.3 Difference between : 94.0 ± 43.9, P=.000 1: ± : 14.5 ± 21.2 Difference between : ± 34.8, P= weeks DASH (%) 1: 18.9, 95%CI 10.9, 26.9 P=.001 2: 10.9, 95%CI 0.04, 21.7 P=.06 Active ROM in abduction ( ) Differences between : 8.07 ± 18.31, P=.184 1: 32.5, 95%CI 16.6, : 10.9, 95%CI 9.7, 28.0 Differences between 21.6 ± 25.1, P= weeks DASH (%) 1: 22.4 ± 1.9 2: 21.4 ± 5.5 Difference between : 1.0 ± 5.8 No statistical test was performed within and between Active ROM in abduction ( ) 1: 43.1 ± : 52.5 ± 17.6 Difference between group: ± 26.2 No statistical test was performed within and between Active ROM in flexion ( ) n=61 11 weeks Shoulder pain questionnaire (/28) 1: 30.5 ± : 25.0 ± 21.2 Difference between : ± 26.8 No statistical test was performed within and between Statistically significant differences within all, P<.001 No statistical test was performed between 11/16 10/16 7/16 5/16! 36

37 Mean age: 49.3 ± 12.9 Diagnosis: Shoulder pain, limited shoulder ROM n=130 Group 1 acromioclavicular joint and glenohumeral joint) Group 2: "Classic physiotherapy" (exercises, massage, and physical agents applications) Group 3: Corticosteroid injections Group 2 Group 3 n=64 n=47 n=44 n=48 n=38 2 to 3 years Perception of being cured Duration of complaints (weeks) Recurrence of complaints by week 11 (% of patient) Current or previous complaints Limitation in activities of daily living Perception of being cured Statistically significant differences within all, P<.001 No statistical test was performed between Statistically significant differences between favoring group 3 over group 1 and 2, P<.001 Statistically significant differences between favoring group 1 over group 2, P<.001 Group 1:8.6% Group 2:14.3% Group 3:17.9% No statistical test was performed between No statistically significant differences between, P.05 No statistically significant differences between, P.05 No statistically significant differences between, P.05 Abbreviation: DASH, Disability of the arm shoulder and hand, self reported disability questionnaire, higher scores indicate a greater level of disability; MD, mean difference; MT, manual therapy; NPRS : numerical pain rating scale; PNF, proprioceptive neuromuscular facilitation; ROM, range of motion; SPADI : Shoulder Pain and Disability Index, self-assessment of symptoms and function of the shoulder, higher scores indicate a greater level of disability;sf-36, The Short Form (36) Health Survey, patient-reported survey of patient health, higher scores indicate a greater level of function; UCLA shoulder score, University of California at Los Angeles shoulder score, pain, function, active flexion, flexion strength, patient satisfaction, higher scores indicate a greater level of function; VAS, visual analog scale! 37

38 TABLE 4. Summary of evidence for the efficacy of manual therapy Treatment Number of studies Total number of participants Outcome measures and pooled effect Conclusions Quality of evidence Overall effect of MT either alone or in conjunction with another intervention compared with placebo or another intervention MT alone compared with a placebo Adding a MT intervention to exercises or to a multimodal rehabilitation program with exercises MT combined with other types of interventions compared with multimodal interventions Pain (10 cm VAS) pooled effect: 1.2, 95%CI 0.8, 1.6 favoring MT Pain (10 cm VAS) pooled effect: 1.0, 95%CI 0.6, 1.4 favoring MT Significant effect that could be clinically important Significant effect that could be clinically important 1 39 Function Unclear if MT has an effect on function 2 99 Strength and ROM Contradictory results Pain (10 cm VAS) pooled effect: 1.0, 95%CI 0.7, 1.4 favoring MT added to exercises 2 91 Pain (10 cm VAS) Significant effect that could be clinically important Function Unclear if MT has an effect on function 2 88 ROM pooled effect: -6.1, 95%CI -20.6, 8.4 MT does not improve ROM Pain, function, ROM Contradictory results Low to moderate Low Low Low Abbreviations : MT, manual therapy ; ROM, range of motion ; SPADI, Shoulder Pain and Disability Index, self-assessment of symptoms and function of the shoulder, higher scores indicate a greater level of disability ; VAS, visual analog scale! 38

39 FIGURE 1. Schematic breakdown of literature search results! Included( Eligibility( Screening( Identification(!! Search strategy: 1. Search terms relevant to shoulder: shoulder, glenohumeral, subacromial, rotator cuff 2. Search terms relevant to impingement: impingement, tendinitis, tendonitis, tendinopathy, pain, bursitis, syndrome, rupture, tear NOT capsulitis/slap 3. Search terms relevant to manual therapy: manual therapy, manipulation, mobilization, physical therapy, nerve/neurodynamic, deep/transverse/cross friction, and massage Types: randomized controlled trial Langage: english or french Dates: date of inception to June 2014 Records identified through database searching (n=287) Records after duplicates removed (n=278) Records screened (n=278) Full-text articles assessed for eligibility (n=32) Included studies (n=21) Additional records identified through hand searches of previous systematic reviews and retrieved study reference lists (n=6) Records excluded (n=246) -Did not meet inclusion criteria (n=244) -Reviews (n=2) Full-text articles excluded (n=11) -Other pathologies (n=4) -Manual therapy was not the primary treatment (n=4) -Not a randomized controlled trial (n=2) -Same group different follow up time (n=1)! 39

40 FIGURE 2. Detailed methodological assessment of included studies using the Cochrane Risk of Bias Tool Green, low risk of bias; Red, high risk of bias ; Yellow, unclear or unknown risk of bias! 40

41 FIGURE 3.Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.! 41