Jennifer C. Seida, MPH; Claire LeBlanc, MD; Janine R. Schouten, BSc; Shima S. Mousavi, MD; Lisa Hartling, PhD; Ben Vandermeer, MSc; Lisa Tjosvold, MLIS; David M. Sheps, MD, MSc
Disclaimer: The authors of this report are responsible for its content. Statements in the report should not be construed as endorsement by the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services.
Grant Support: By the Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services (contract 290-02-0023).
Potential Conflicts of Interest: Ms. Seida, Dr. LeBlanc, Ms. Schouten, Dr. Mousavi, Mr. Vandermeer, and Dr. Sheps: Grants received (money to institution): Agency for Healthcare Research and Quality. Dr. LeBlanc: Consultancy: University of Alberta. Dr. Hartling: Other (money to institution): Agency for Healthcare Research and Quality. Dr. Sheps: Consultancy: Agency for Healthcare Research and Quality. Disclosures can also be viewed at www.acponline.org/authors/icmje/ConflictOfInterestForms.do?msNum=M10-0825.
Corresponding Author: Jennifer C. Seida, MPH, Alberta Research Centre for Health Evidence, University of Alberta, 11402 University Avenue, Edmonton, Alberta T6G 2J3, Canada; e-mail, email@example.com.
Current Author Addresses: Ms. Seida, Ms. Schouten, Drs. Mousavi and Hartling, and Mr. Vandermeer: Alberta Research Centre for Health Evidence, University of Alberta, Aberhart Center, 11402 University Avenue, Edmonton, Alberta T6G 2J3, Canada.
Dr. LeBlanc: University of Alberta, Aberhart Center, 11402 University Avenue, Edmonton, Alberta T6G 2J3, Canada.
Ms. Tjosvold: 2K3.28 Walter C. Mackenzie Health Sciences Center, University of Alberta, Edmonton, Alberta T6G 2R7, Canada.
Dr. Sheps: 10839 124th Street, Edmonton, Alberta T5M 0H4, Canada.
Author Contributions: Conception and design: J.C. Seida, C. LeBlanc, L. Hartling, D.M. Sheps.
Analysis and interpretation of the data: J.C. Seida, C. LeBlanc, J.R. Schouten, L. Hartling, B. Vandermeer, D.M. Sheps.
Drafting of the article: J.C. Seida, J.R. Schouten, S.S. Mousavi, L. Hartling, B. Vandermeer, L. Tjosvold, D.M. Sheps.
Critical revision of the article for important intellectual content: J.C. Seida, C. LeBlanc, S.S. Mousavi, L. Hartling, B. Vandermeer.
Final approval of the article: J.C. Seida, C. LeBlanc, L. Hartling, L. Tjosvold.
Statistical expertise: B. Vandermeer.
Obtaining of funding: L. Hartling.
Administrative, technical, or logistic support: J.C. Seida, L. Tjosvold.
Collection and assembly of data: J.C. Seida, J.R. Schouten, S.S. Mousavi, L. Hartling, L. Tjosvold.
Seida J., LeBlanc C., Schouten J., Mousavi S., Hartling L., Vandermeer B., Tjosvold L., Sheps D.; Systematic Review: Nonoperative and Operative Treatments for Rotator Cuff Tears. Ann Intern Med. 2010;153:246-255. doi: 10.7326/0003-4819-153-4-201008170-00263
Download citation file:
Published: Ann Intern Med. 2010;153(4):246-255.
Many approaches exist for managing rotator cuff tears.
To compare the benefits and harms of nonoperative and operative interventions on clinically important outcomes in adults with rotator cuff tears.
12 electronic databases (1990 to September 2009), gray literature, trial registries, and reference lists were searched.
Controlled and uncontrolled studies that assessed nonoperative or operative treatments or postoperative rehabilitation for adults with confirmed rotator cuff tears were included. Operative studies in English-language publications and nonoperative and postoperative rehabilitation studies in English, French, or German were considered. Studies were assessed in duplicate.
2 reviewers assessed risk for bias by using the Cochrane Risk of Bias tool and the Newcastleâ€“Ottawa Scale. One reviewer rated the evidence by using a modified GRADE (Grading of Recommendations Assessment, Development, and Evaluation) approach. Data were extracted by one reviewer and verified by another.
137 studies met eligibility criteria. All trials had high risk for bias. Cohort and uncontrolled studies were of moderate quality. Reported functional outcomes did not differ between open versus mini-open repair, mini-open versus arthroscopic repair, arthroscopic repair with versus without acromioplasty, or single-row versus double-row fixation. Earlier return to work was reported for mini-open repair versus open repair and for continuous passive motion with physical therapy versus physical therapy alone. Open repairs showed greater improvement in function than did arthroscopic debridement. Complication rates were low across all interventions.
Limited evidence, which was often of low quality, precluded conclusions for most comparisons. Language restrictions may have excluded some relevant studies, and selective outcome reporting may have introduced bias.
Evidence on the comparative effectiveness and harms of various operative and nonoperative treatments for rotator cuff tears is limited and inconclusive.
Agency for Healthcare Research and Quality.
Multiple therapies are available to treat rotator cuff tears.
This systematic review of 137 studies found evidence that patients with rotator cuff tears experienced improvements in function after undergoing any of several operative procedures or nonoperative therapy. Few complications were seen with the surgical procedures. Sparse comparative data precluded recommendations for a superior treatment approach.
The amount and strength of available evidence was low for many therapies.
Several therapies might benefit patients with rotator cuff tears and are associated with a reasonably low risk for harm, but we do not yet know the most beneficial therapy.
The rotator cuff comprises 4 muscle tendon units that stabilize the humeral head within the shoulder joint and aid in moving the upper extremity (1). “Rotator cuff tear” refers to partial or full discontinuation of 1 or more of these units due to traumatic injury or degeneration. Incidence increases with age. Approximately 54% of adults older than 60 years have a partial or complete rotator cuff tear, compared with only 4% of those aged 40 to 60 years (2). Symptoms include pain, weakness, and limitation of motion (1).
Both nonoperative and operative treatments are used to relieve pain and restore movement and function of the shoulder (3). Most patients initially undergo 6 weeks to 3 months of nonoperative treatment, which may consist of combinations of oral medications and injections; rest from activity; passive and active exercise; and such therapies as heat, cold, or ultrasonography. Failing nonoperative treatment, the cuff may be surgically repaired by using an open, mini-open, or arthroscopic approach. Various postoperative rehabilitation programs can help restore range of motion, strength, and function.
Earlier operative repair may result in better patient outcomes, earlier return to work, and decreased costs (4, 5). Patients and clinicians struggle with when to abort nonoperative treatment in favor of surgery. The relative effectiveness of the various nonoperative and operative treatment options for patients with rotator cuff tears remains uncertain. Our comparative effectiveness review examines the relative effectiveness and safety of all nonoperative and operative treatments for rotator cuff tears in adults.
We prospectively developed and followed a research protocol. An external panel of content experts provided input in formulating the key questions and identifying interventions and outcomes of interest. The full evidence report, including search strategies and evidence tables, is available at www.effectivehealthcare.ahrq.gov/reports/final.cfm.
We investigated early versus late surgical repair, operative interventions, nonoperative interventions, postoperative rehabilitation, and operative versus nonoperative treatments. We also examined complications and prognostic factors in all included studies.
We systematically searched MEDLINE, EMBASE, Evidence-Based Medicine Reviews, the Cochrane Library, AMED, CINAHL, SPORTDiscus with Full Text, Academic Search Elite, Health Source, Science Citation Index Expanded (via Web of Science), Scopus, BIOSIS Previews, CRISP, Current Controlled Trials, ClinicalTrials.gov, and the Nederlands Trial Register from January 1990 to September 2009. Appendix Table 1 lists our MEDLINE search strategy. We also hand-searched abstracts from recent scientific meetings and reference lists of relevant reviews to identify additional studies. No language restrictions were applied.
Trials, cohort studies, and prospective uncontrolled studies that evaluated any nonoperative or operative treatment or postoperative rehabilitation for adults with confirmed rotator cuff tears were eligible for inclusion. We defined uncontrolled studies as single-group studies that reported baseline and follow-up data. For the purposes of this review, we also considered cohort studies that compared the effectiveness of only 1 intervention across 2 patient populations (such as open repair in older versus younger patients) to be uncontrolled studies. Confirmed tears were defined as partial- or full-thickness lesions diagnosed by imaging or intraoperative findings. Studies were required to enroll a minimum of 11 participants and report at least 1 of the following outcomes: quality of life, function, time to return to work or activity, pain, range of motion, or strength. Operative studies were required to follow participants for at least 12 months; no follow-up criteria were set for nonoperative or postoperative rehabilitation studies. We included operative studies published in English only because of a lack of translation resources. We considered English-, German-, and French-language publications for studies that examined nonoperative treatments and postoperative rehabilitation, because the literature on these interventions was sparse. One reviewer screened titles, keywords, and abstracts for broad relevance. Two independent reviewers assessed the full publication of potentially relevant studies, and discrepancies were resolved by consensus.
One reviewer extracted data by using a standardized form, and a second reviewer verified the data for accuracy and completeness. Reviewers resolved discrepancies by consensus or through a third party.
Two reviewers independently assessed the methodological quality of included studies. We evaluated trials by using the Cochrane Risk of Bias tool (6) and observational analytic studies by using a modified Newcastle–Ottawa Scale (7). Uncontrolled studies were assessed for consecutive enrollment, complete outcome data, and standardized or independent approach to outcome assessment. The funding sources were recorded for all studies.
One reviewer (Dr. Hartling) graded the strength of evidence according to published guidelines (8, 9) for the 4 key outcomes of quality of life, functional outcomes, time to return to work, and cuff integrity. Four domains were assessed: risk for bias (low, medium, or high), consistency (no inconsistency, inconsistency present, unknown, or not applicable), directness (direct or indirect), and precision (precise or imprecise).
We summarized the included studies qualitatively. Controlled studies were combined by using meta-analysis if the study design, study population, interventions being compared, and outcomes were sufficiently similar. For continuous outcomes measured on different scales across studies, we calculated a standardized mean difference for the pooled estimate. Results were combined by using random-effects models, and statistical heterogeneity was quantified by using the I2 statistic (10, 11).
The Agency for Healthcare Research and Quality funded this review and provided feedback on the question formulation and decision to submit for publication but was not involved in the searches, selection, data extraction, data analysis, or interpretation of the findings.
We identified 5677 citations and included 137 studies in our review (Figure). Appendix Table 2 lists included and excluded studies, along with the reasons for exclusion.
Table 1 summarizes general study characteristics. Number of study participants ranged from 12 to 224 (median, 55; interquartile range [IQR], 33 to 93). Mean age ranged from 41.2 to 80 years. Appendix Table 3 provides further study characteristics.
All randomized, controlled trials (RCTs) and controlled clinical trials had a high risk for bias. The most common sources of potential bias were inadequate blinding, inadequate allocation concealment, and incomplete outcome data. The methodological quality of the cohort studies was moderate, with a median score of 5 out of 8 stars (IQR, 4 to 6 stars). Limitations in study design included lack of independent, blinded outcome assessment and failure to adequately control for potential confounders. Uncontrolled studies generally had moderate quality, with consecutive enrollment, adequate follow-up, and standardized outcome assessment reported in 63%, 77%, and 44% of studies, respectively. Funding sources were reported for only 49 studies (36%). Appendix Table 4 details the methodological quality of each study.
One RCT (12), which compared early versus late surgical repair after failed nonoperative treatment, found superior mean functional outcome scores with early repair but did not report the statistical significance of this difference. The groups did not significantly differ in cuff integrity. Overall, the evidence was too limited to make a conclusion. Table 2 summarizes the strength of the evidence for all comparisons.
One hundred thirteen studies examined operative interventions, and 11 examined postoperative rehabilitation. Studies that assessed surgery were categorized as comparing an operative approach (such as open, mini-open, or arthroscopic), a technique (such as suture, anchor type, or configuration), or augmentation. Table 2 summarizes the strength of the evidence for these comparisons.
Operative approaches were evaluated in 32 controlled and 58 uncontrolled studies. The evidence on the specific comparisons from the trials and cohort studies is examined here. The uncontrolled studies consistently reported functional improvement from preoperative to postoperative scores, regardless of the approach (open, mini-open, or arthroscopic), sample size, or outcome measure.
One RCT (13) and 2 retrospective cohort studies (14, 15) compared open with mini-open repair. Quality of life (13), function (13–15), cuff integrity (14, 15), and range of motion (13, 14) did not significantly differ between the groups. The cohort studies demonstrated statistically significant benefit from mini-open repairs, with patients returning to work or activity approximately 1 month earlier (mean difference, 1.08 months [95% CI, 0.63 to 1.52 months]) and having greater abduction strength (14).
Ten studies (1 controlled clinical trial , 2 prospective cohort studies [17, 18], and 7 retrospective cohort studies [19–25]) compared mini-open with arthroscopic repair. Function did not significantly differ, either for the controlled clinical trial or the pooled estimate of 9 cohort studies (standardized mean difference, −0.11 [CI, −0.28 to 0.06]). Cuff integrity (20, 23), pain (16, 18, 23, 24), range of motion (16, 18, 20, 23, 24), and strength (16, 24) also did not differ.
One prospective cohort study (26) and 2 retrospective cohort studies (27, 28) compared open versus arthroscopic rotator cuff repair. A pooled estimate showed no differences in function (standardized mean difference, −0.49 [CI, −1.12 to 0.13]); however, we found significant statistical heterogeneity among the 3 studies (I2 = 83%). Patient age, type of tear, and tear size did not seem to differ among the studies; the heterogeneity may be due to differences in the study design (prospective vs. retrospective) or length of follow-up. One study (28) found no difference in pain or cuff integrity but statistically significant differences in favor of arthroscopic repair for external rotation range of motion and strength and supraspinatus strength.
Two prospective cohort studies that compared open or mini-open repair with arthroscopic repair found no difference in function (29, 30) or cuff integrity (29). Arthroscopic repair was favored for pain relief in one study (30), whereas open or mini-open repair was favored for external rotation in the other (29).
Two controlled clinical trials (31, 32) and 2 retrospective cohort studies (33, 34) compared open repair with debridement. Improvement in function was statistically significant for the repair groups (standardized mean difference for trials, 0.59 [CI, 0.15 to 1.03]; for cohort studies, 1.00 [CI, 0.11 to 1.90]); however, we found substantial heterogeneity among the cohort studies (I2 = 79%). The statistical heterogeneity among studies may be explained by the different study designs (prospective vs. retrospective) or tear sizes (small or medium tears vs. massive tears). The magnitude of the difference varied across studies from an absolute difference of 2.2 on a 35-point scale (32) to 11.5 on an 83-point scale (33). One cohort study (33) showed a statistically significant shorter time to maximum range of motion with arthroscopic debridement (3.2 vs. 6.8 months).
Two RCTs (35, 36) compared arthroscopic rotator cuff repair with acromioplasty versus repair without acromioplasty. One prospective cohort study (37) compared arthroscopic repair with acromioplasty alone. Function did not differ between the groups.
Seven studies compared different operative approaches: biceps tenotomy versus tenodesis (38), rotator cuff repair versus palliative treatment (39), arthroscopic rotator cuff repair plus superior labral anterior-to-posterior lesion repair versus arthroscopic rotator cuff repair plus biceps tenotomy (40), arthroscopic rotator cuff repair plus tenodesis with versus without proximal biceps detachment (41), arthroscopic debridement with versus without tenotomy (42), complete open repair versus partial open repair versus debridement (43), and open repair plus classic acromioplasty versus repair plus modified acromioplasty (44). We found few clinically important differences between groups across all studies. No differences in function were observed for 5 of the comparisons (38, 41–44). One study (39) found a statistically significant difference in function that favored rotator cuff repair more than palliative treatment. Another (40) showed greater functional improvement with arthroscopic rotator cuff repair with biceps tenotomy than with arthroscopic rotator cuff repair plus superior labral anterior-to-posterior lesion repair. However, the absolute difference of 4 points on the 35-point scale is of questionable clinical importance. Range of motion (38, 41, 43) and strength (43) did not significantly differ.
Fifteen controlled studies examined operative techniques. Six studies (4 RCTs [45–48] and 2 cohort studies [49, 50]) compared single-row suture anchor repairs with double-row repairs. For function, the pooled estimate showed statistically significant improvement that favored double-row fixation (standardized mean difference for trials, 0.55 [CI, 0.02 to 1.07]; for cohort studies, 0.78 [CI, 0.46 to 1.11]). We found considerable heterogeneity among the trials (P = 0.008; I2 = 75%) but not among the cohort studies. The studies were similar in design, patient age, and tear type; the heterogeneity may be attributable to tear size, which ranged from small in one study (45) to massive in another (47). Although the meta-analysis showed statistical significance, the absolute differences in the change scores were small (5 points on a 100-point scale) (50) and therefore of questionable clinical relevance. For cuff integrity, the pooled risk ratio from 3 trials (45–47) showed no difference between groups (risk ratio, 1.20 [CI, 0.86 to 1.68]). However, 1 cohort study (50) found a statistically significant difference that favored double-row fixation. Measures of health-related quality of life (45), return to work (46), range of motion (47), and strength (45, 48, 49) did not differ across techniques.
Of the 2 studies that compared the effectiveness of mattress stitch versus simple stitch, 1 controlled clinical trial (51) favored mattress stitch for functional outcomes. A prospective cohort study showed no difference between groups (52). Cuff integrity (51, 52), pain (51, 52), and range of motion (51) did not differ.
Each of the 7 remaining studies examined different technique comparisons, including bioabsorbable tacks versus suture tying (53), side-to-side versus tendon-to-bone fixation (54), nonabsorbable versus absorbable suture (55), headed bioabsorbable corkscrew versus metal suture anchor (56), mattress versus single transosseous suture (57), ultrasonic suture welding versus hand-tied knots (58), and staple fixation versus side-to-side suture (59). The overall level of evidence was low for these techniques. All studies assessed function; 3 found statistically significant differences (12 to 15 points on a 100-point scale) between the groups examined (53, 54, 56). We noted a statistically significant difference in cuff integrity in 1 study (55) but could not assess the difference in 2 others because of incomplete data reporting (57, 59). Individual studies found statistically significant differences for range of motion (56) and strength (57) but not for pain (53, 55).
Three small controlled and 5 uncontrolled studies assessed surgical augmentation. One RCT (60) and 1 retrospective cohort study (61) compared porcine small intestine submucosa xenografts with no augmentation. Function (60) and cuff integrity (60, 61) did not significantly differ. One study (61) found a slower rate of activity-related pain resolution, an almost global loss of strength, and less sport participation in the augmentation group. One retrospective cohort study (62) compared patch graft with no augmentation and found no statistically significant difference in function. Range of motion for abduction was significantly improved with patch augmentation (40° difference between groups). Five uncontrolled studies evaluated different types of augmentations, and all showed improvement in functional scores.
Ten controlled studies and 1 uncontrolled study evaluated postoperative rehabilitation. Three RCTs (63–65) studied the addition of continuous passive motion to physical therapy. Overall, moderate evidence showed no difference in function or pain (pooled standardized mean differences, 0.08 [CI, −0.37 to 0.52] and −0.12 [CI, −1.08 to 0.83], respectively). One study (63) found no statistically significant difference between groups for range of motion or strength. Another study (64) found that the time to return to work favored continuous passive motion with physical therapy over physical therapy alone (absolute difference of 12 and 21 days, respectively). Similarly, time to 90° abduction favored continuous passive motion. Continuous passive motion may improve recovery over the short term.
Seven additional studies each reported on different postoperative rehabilitation comparisons: a land-based program with or without aquatic therapy (66), inpatient versus outpatient rehabilitation (67), individualized physical therapy plus home exercise versus home exercise alone (68), rehabilitation with progressive loading versus traditional loading (69), inpatient rehabilitation versus private outpatient rehabilitation with the Concept Global d'Epaule method (70), standardized versus nonstandardized physical therapy (71), and videotape-based versus therapist-based home exercise instruction (72). One study (71) demonstrated that patients who received standardized physical therapy showed statistically significant improvement in function compared with those who received nonstandardized treatment. Progressive loading showed a statistically significant reduction in pain versus traditional loading (69). Outpatient therapy with the Concept Global d'Epaule method reduced pain better than inpatient rehabilitation (70). Health-related quality of life (66), function (68, 70, 72), pain (67, 71), range of motion (66–68), and strength (68, 70) did not significantly differ across the remaining studies. We could not evaluate 1 study because the investigators did not report levels of significance for most outcomes (69). One uncontrolled study (73) demonstrated significant improvements in health-related quality of life and function after rehabilitation.
Three controlled and 7 uncontrolled studies examined nonoperative interventions. One RCT (74) compared sodium hyaluronate with dexamethasone injections; however, no head-to-head comparison regarding the relative efficacy of these interventions was reported. A retrospective cohort study (75) compared rehabilitation that focused on protecting the cuff through reliance on other muscles (deltoid, pectoralis major, and latissimus dorsi) with no rehabilitation. Differences in function that favored the rehabilitation group were statistically significant and clinically important (absolute difference, 26.9 points on a 100-point scale). A second retrospective cohort study (33) compared administration of steroid injection with no steroid injection among participants undergoing physical therapy (treatment components not specified) and receiving oral medications (not specified). Function (absolute difference, 11 on an 83-point scale) and time to maximum range of motion (absolute difference, 4 months) significantly improved. For the uncontrolled studies, the degree of improvement in functional outcome scores varied considerably. The strength of evidence was too low to make conclusions for any of the nonoperative interventions (Table 2).
Five studies compared nonoperative with operative treatments. Four studies (12, 33, 76, 77) included either physical therapy (treatment components not specified) or stretching and strengthening exercises, with or without the addition of steroid injections, oral medications, activity modification, or manual therapy. One study (78) examined the use of shock-wave therapy. Nonoperative treatments were compared with either open or mini-open repair. One study included a third comparison group that received arthroscopic debridement (33). All groups showed statistically significant improvements regardless of the intervention. All but 1 study (76) showed statistically significant differences in function that favored operative repair. One study (33) showed that patients who had arthroscopic debridement had a statistically significant shorter time to maximum range of motion (3.2 months) than did those in the nonoperative and open repair groups (6.8 months each). In general, the evidence was too limited to make conclusions regarding comparative effectiveness (Table 2).
Sixty-four studies provided data on 34 different complications, and an additional 21 studies reported no complications during follow-up. Five complications were identified a priori to be most clinically relevant. Generally, complications were uncommon (Appendix Table 5). Recurrent tears occurred in a median of 3% of patients (IQR, 2% to 7%) in operative studies and 4% (IQR, 0% to 5%) in postoperative rehabilitation studies. In most operative studies, no patients had infections (IQR, 0% to 2%), whereas a median of 0.5% of patients (IQR, 1.3% to 6%) in postoperative rehabilitation studies reported infections. Stiffness occurred in a median of 2% of patients (IQR, 0% to 3%) in operative studies, whereas 2 nonoperative studies reported events in 4% and 7% of patients. Two percent of patients (IQR, 1.8% to 4.5%) in operative studies developed reflex sympathetic dystrophy, versus 0% and 7% in 2 postoperative rehabilitation studies. Neurologic injury was very rare (median, 0% [IQR, 0% to 0.1%]). Because of the low event rates, the benefit of receiving treatment for rotator cuff tears seems to outweigh the risk for associated harms.
Seventy-two of the 137 studies assessed the effect of known prognostic factors on patient outcomes. The number of prognostic factors examined across many different outcome measures and the inconsistency among the investigators' conclusions limited our ability to identify predictors of good outcome for nonoperative and operative treatments of rotator cuff tears. Overall, older age, increased tear size, and greater preoperative symptoms were repeatedly found to be associated with recurrent tears. Sex, worker's compensation board status, and duration of symptoms were not found to be associated with poorer outcomes in most studies that examined these variables.
Our comparative effectiveness review provides a current synthesis of the state of the evidence on nonoperative and operative interventions for rotator cuff tears. Sparse data are available for most comparisons, which precludes firm conclusions for a single approach or the optimal overall management of this condition. Patients experienced substantial improvements across all interventions. We found few clinically important differences when comparisons between interventions were available. Complications were rare, and we considered few to be clinically important. The benefit of receiving treatment for rotator cuff tears seems to outweigh the risk for associated harms.
The strength of evidence was low for most interventions. This low grade was driven by the high risk for bias within individual studies and the lack of consistency and precision across studies. Many studies used weak study designs, failed to control for important sources of bias, and lacked an independent comparison group. Trials had a high risk for bias and frequently had inadequate blinding, insufficient allocation concealment, and incomplete outcome data. Although blinding is not always feasible because of the nature of the intervention, adequate allocation concealment is always possible in an RCT. Selective outcome reporting may also have introduced bias. The methodological quality was moderate for the cohort and uncontrolled studies. Many studies did not control for important potential confounders in their design or analysis. We could not assess the possibility of publication bias because of the small number of studies for each comparison.
The lack of consistency and precision of results across the studies was primarily due to varied comparisons made across this body of literature; relatively few studies compared the same interventions. In addition, variation in the pathologic presentation of rotator cuff disease contributed to inconsistency among the studies. Although most patients had full-thickness tears, the size and configuration of the tears, degree of fatty infiltration, and number and type of comorbid conditions varied widely across the included studies. Both outcome measures and timing of measurements varied considerably across studies, which made comparisons difficult. The most common outcome was function, but 21 different tools were used for this purpose, and multiple tools were often used in the same study.
The quality of reporting was poor and inconsistent, particularly in studies that evaluated nonoperative interventions. Physical therapy was often labeled as an intervention, with no description of treatment components or delivery. Studies rarely described the timing and frequency of intervention components, the training and experience of the staff who implemented the interventions, or any co-interventions, which leads to interpretive challenges.
Our review differs from previously published reviews because it assesses the full range of nonoperative and operative treatment options, includes a broad range of study designs, and considers only studies that confirmed the presence of rotator cuff tears. One previous review (79), published in 2004, similarly examined all treatment methods in controlled and uncontrolled studies, but we have included the considerable literature published since then. One limitation of our review is the language restriction, which we applied because of a lack of translation resources. In addition, we could not retrieve 29 potentially relevant references. It is unlikely that these would have affected our results, because we identified many of these citations through searches of gray literature.
Priorities for future research include studies that compare the effectiveness of early versus delayed surgery. Patients generally have surgery after several months of failed conservative treatment. Evidence is needed to determine whether surgery should be delayed and, if so, for how long and for whom. Evidence that compared the relative effectiveness of operative versus nonoperative treatments or compared the various nonoperative treatment options was sparse. Although most of the studies we reviewed focused on the comparative effectiveness of operative treatments, evidence for most individual treatment comparisons was sparse, which leaves many unanswered questions. For future research, investigators should use a streamlined approach when evaluating operative treatments that begins with assessing broad treatment questions before focusing on detailed procedures. They should use a comparative design and ensure that the diagnosis of rotator cuff tears is confirmed appropriately. Consensus on clinically important and patient-important outcomes, as well as choice of measurement tools, is needed to ensure consistency and comparability across studies. Detailed reporting of study methods and interventions is needed to allow appropriate interpretation of results and replication of treatments. The CONSORT (Consolidated Standards of Reporting Trials) (80) and STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) (81) statements provide guidance for minimizing the risk for bias when designing RCTs and cohort studies, respectively.
We found moderate evidence for some interventions, but the data were too limited to make definite conclusions for most interventions examined. Few differences of clinical importance were evident when we compared the relative effectiveness of the various treatments. Future studies, of high methodological quality, are needed to explore the relative effectiveness of early versus delayed surgery and of nonoperative versus operative treatment, as well as among the various nonoperative and operative interventions.
The In the Clinic® slide sets are owned and copyrighted by the American College of Physicians (ACP). All text, graphics, trademarks, and other intellectual property incorporated into the slide sets remain the sole and exclusive property of the ACP. The slide sets may be used only by the person who downloads or purchases them and only for the purpose of presenting them during not-for-profit educational activities. Users may incorporate the entire slide set or selected individual slides into their own teaching presentations but may not alter the content of the slides in any way or remove the ACP copyright notice. Users may make print copies for use as hand-outs for the audience the user is personally addressing but may not otherwise reproduce or distribute the slides by any means or media, including but not limited to sending them as e-mail attachments, posting them on Internet or Intranet sites, publishing them in meeting proceedings, or making them available for sale or distribution in any unauthorized form, without the express written permission of the ACP. Unauthorized use of the In the Clinic slide sets will constitute copyright infringement.
Healthcare Delivery and Policy, Prevention/Screening.
Results provided by:
Copyright © 2016 American College of Physicians. All Rights Reserved.
Print ISSN: 0003-4819 | Online ISSN: 1539-3704
Conditions of Use
This PDF is available to Subscribers Only