This was a multicenter, double-blind, balanced randomization (1:1), parallel-group study conducted at two orthopaedic surgery facilities from June 2003 to February 2011 (including all follow-up examinations to twenty-four months after surgery). Approval was granted by research ethics boards at both sites prior to enrollment. This study was registered with ClinicalTrials.gov (ClinicalTrials.gov identifier NCT00290888). Potential participants were identified by one of two participating surgeons (P.M. and P.L.) on the basis of the preoperative consultation. Patients screened for inclusion were referred to a surgeon because of a rotator cuff tear that had undergone unsuccessful conservative management (e.g., activity modification, administration of analgesic or anti-inflammatory medication, and/or physiotherapy). If individuals agreed to be contacted, they were approached by one of two research assistants who guided them through the informed consent process.
The criteria for inclusion were patients eighteen years of age or older with a diagnosis of a full-thickness rotator cuff tear by clinical and imaging criteria, including either ultrasound or magnetic resonance imaging. Tears of ≤4 cm in size of one or more tendons were included. All patients had persistent pain and functional disability for at least six months and had six months of conservative treatment without success. Exclusion criteria were evidence of substantial osteoarthritis or an articular cartilage pathological condition in the shoulder, evidence of glenohumeral instability (i.e., full labral tears superiorly or complete labral tears anteriorly or posteriorly [Bankart lesions]), previous surgical procedures on the affected shoulder, evidence of major joint trauma, an infection or osteonecrosis in the shoulder, partial-thickness tears of the rotator cuff, an inability to provide informed consent because of a language barrier or mental status, a major medical condition that would affect quality of life, and patients with a Workers’ Compensation claim or an unwillingness to be followed for the duration of the study.
The primary outcome measure was a validated disease-specific quality-of-life measurement for rotator cuff disease, the WORC index14,15. The WORC questionnaire consists of twenty-one questions that delve into the domains of physical symptoms, social well-being, work, sports and recreation, and emotional well-being. Secondary outcome measures included the validated ASES standardized form for the assessment of shoulder function16 and shoulder range of motion. Patients completed both outcome measures preoperatively and at three, six, twelve, eighteen, and twenty-four months after surgery. The primary follow-up interval, on which our hypothesis, sample size calculation, and conclusions were based, was twenty-four months after surgery. To allow for subgroup analysis on the basis of acromial morphology, acromion type was determined by the surgeon using preoperative radiographs (supraspinatus outlet view) done in a standardized fashion, on the basis of the classification system of Bigliani et al.1. Type 1 refers to an acromion with a flat undersurface; Type 2, an acromion with a curved undersurface; and Type 3, a hooked acromion.
Final eligibility of the participants for the study occurred following intraoperative visual inspection of the rotator cuff tear and following determination of repairability. A full-thickness tear was classified as repairable if the tendon could be restored to a point where the insertion footprint was covered when traction was applied without undue tension, which would include signs of tendon and suture interface failure. All surgical procedures were performed by one of two fellowship-trained shoulder surgeons (P.M. or P.L.). Patients were allocated to undergo either an arthroscopic rotator cuff repair with acromioplasty (ACR-A) or an arthroscopic cuff repair alone (ACR). In the current study, acromioplasty was considered to include release of the coracoacromial ligament and excision of the anteroinferior surface of the acromion until flat. Allocation of patients was conducted with use of a series of opaque envelopes containing group assignment based on a computer-generated randomization list. Patient assignment to the study groups was concealed from the researcher who enrolled and assessed the patients. The randomization envelope was opened following intraoperative inspection of the shoulder. The surgeon was not blinded to the treatment allocation, but the patient and the research assistant performing follow-up evaluations were blinded.
For both study groups, surgery was performed with the patient in the lateral decubitus position and under general anesthesia. The arthroscope was inserted through a standard posterior portal. An anterior portal was established anterior to the acromioclavicular joint, and a lateral portal was made slightly inferior to the lateral border of the acromion. Instrumentation was done through these two portals. Assessment of the glenohumeral joint was performed with special attention paid to the labrum, biceps tendon, humeral head, and glenohumeral ligaments. Any frayed tissue in these regions was debrided, and the arthroscope was then redirected to the subacromial space. Any bursal tissue that obscured visualization of the rotator cuff was removed. A tear in the tendon was visualized and adhesions removed from retracted tendons until full excursion was possible. The characteristics of the tear and its suitability for repair and inclusion in the study were assessed.
For the patients assigned to the ACR-A group, acromioplasty was then performed by first releasing the coracoacromial ligament off the anterior undersurface of the acromion by use of an arthroscopic soft-tissue shaving resector and then thinning the inferior surface of the acromion, with a motorized burr, until it was flat (the acromion type was determined preoperatively from radiographs with use of the Bigliani outlet view)1. A motorized arthroscopic burr was then again used to smooth the acromial undersurface and check for ridges, and any inferior osteophytes were resected. Those in the ACR group did not undergo either the division of the coracoacromial ligament or partial resection of the acromion.
Next, a soft-tissue shaver or electrical ablation device was used to remove scar tissue between the supraspinatus tendon and the acromion or deltoid fascia. If additional release was necessary, the superior articular capsule was incised along the glenoid margin to allow increased excursion of the supraspinatus. Interval releases were performed as necessary but were limited to the rotator interval. This involved a release of the superior glenohumeral ligament and/or coracohumeral ligament to the glenoid rim to facilitate reapproximation of the tendon to the insertion site. A cancellous osseous bed was prepared at the site of the proposed attachment of the tendon between the articular cartilage of the head of the humerus and the greater tuberosity. A burr was used to remove a thin layer of cortical bone, and at least one bone anchor loaded with braided, nonabsorbable suture was placed lateral to the cancellous bone surface in a single-row configuration. Sutures were passed through the supraspinatus tendon approximately 5 mm from the site of the tear. The number of suture anchors varied with the length of the tear. Traction was placed on each suture in the margin of the tendon to reduce the tendon to its repair site and to allow tying of the suture without excessive tension with the upper extremity in the 0° adducted position. If this could not be accomplished, the tendon was repaired by attaching it medial to its anatomical location. After all of the repair sutures were tied, the traction suture was removed and the incisions were closed with absorbable sutures and adhesive tape.
Patients were discharged on the same day as the surgery. Passive or active-assisted shoulder range-of-motion exercises began one week after surgery. Active shoulder motion began at eight weeks postoperatively with strengthening exercises and reintegration into normal activities at twelve weeks postoperatively. Patients returned to the clinic at two weeks and at six to eight weeks postoperatively for wound check only, and at three, six, twelve, eighteen, and twenty-four months postoperatively for study follow-up. All follow-up evaluations were conducted by the research assistants in the clinic setting.
Eighty-six patients were enrolled in this study. Sample size calculation was based on estimates of the expected mean and standard deviation of the WORC score determined in a pilot study, 63% and 16%, respectively, and a difference in WORC score of 25% was considered clinically relevant. To achieve 80% power to detect a significant difference, with alpha at 0.05, seventy-four patients were required in the trial. The sample size was then inflated to eighty-six patients to account for an expected loss to follow-up of 15%.
Statistical Analysis
Statistical analysis was performed with SPSS statistical software (version 16; SPSS, Chicago, Illinois). Repeated-measures analysis of variance (ANOVA) was used to determine the effect of surgery on WORC and ASES scores for each group from baseline to twenty-four months after surgery. At each follow-up time point, comparisons of WORC and ASES scores between the two groups were made with use of t tests for two independent groups. Subanalyses within each group were carried out, with use of one-way ANOVA, to examine differences in outcome score based on acromion type. Post hoc analysis comparing the number of revision surgeries performed between groups was performed with use of the Fisher exact test. Results for all analyses were considered significant at p < 0.05.
Source of Funding
The funding source for this study was the Alexander Gibson Fund from the University of Manitoba. It supported the expenses for a study coordinator, statistical support, and supplies (e.g., postage, courier, patient files, and questionnaires). The funding source did not play a role in this investigation.
The hypothesis that arthroscopic repair of full-thickness rotator cuff tears with associated arthroscopic acromioplasty results in similar outcomes in patient self-assessed shoulder comfort and function compared with the same surgery without acromioplasty was accepted. However, the reoperation rate in the group that did not undergo an acromioplasty was substantially greater than that in the study group in which no reoperations occurred.
The concept of impingement was introduced by Neer in 197217. The procedure of reshaping the acromion to relieve mechanical pressure on the rotator cuff was subsequently widely adopted in open rotator cuff repair. This approach was then advanced with the establishment of a classification system for describing acromial morphology1. The adaptation of the acromioplasty procedure to be performed arthroscopically was first described by Ellman in 198718. The benefits and disadvantages of open surgery compared with the arthroscopic approach to acromioplasty have been identified in a series of studies. As summarized by Spangehl et al., the open procedure is technically simpler to perform and requires less surgeon expertise19. The advantages of the arthroscopic approach include improved cosmetic appearance of the surgical scar, preservation of the deltoid muscle, and faster recovery. Studies comparing the outcome of open and arthroscopic acromioplasty have generally found no outcomes differences between these approaches19-21. None of these prior studies examined differences in outcome on the basis of acromial morphology.
The potential negative aspects of the acromioplasty do not lie with the approach, but with the fundamental components of the procedure that are said to compromise the stability of the coracoacromial arch. A small number of biomechanical cadaveric studies have examined the effect of acromioplasty on glenohumeral stability. To varying degrees, those studies found a substantial increase in superior and anterosuperior translation of the humeral head with resection of the coracoacromial ligament22,23. Due to the complexity of shoulder mechanics and the lack of dynamic stability in cadaveric studies, the clinical relevance of this increased instability in cadavers has yet to be determined22.
Acromioplasty is still commonly performed, though there are few controlled clinical studies that support such widespread use. Gartsman and O'Connor conducted a prospective randomized study comparing the outcomes of arthroscopic rotator cuff repair with and without subacromial decompression in ninety-three patients at one year after surgery12. The study included only patients with a Type-2 acromion, and the outcome measure used was the ASES shoulder score, a validated general functional shoulder outcome measure. The study concluded that arthroscopic subacromial decompression does not appear to change functional outcome following arthroscopic rotator cuff repair. Another randomized controlled trial conducted by Milano et al. examined eighty patients with Type-2 and Type-3 acromions at a two-year follow-up interval13. On the basis of another upper-extremity measure, the Disabilities of the Arm, Shoulder and Hand (DASH), no significant differences were found between those who did or did not undergo subacromial decompression as part of the rotator cuff repair. No report of reoperation rates was included in those studies.
The findings of the current study are consistent with those of previous reports12,13. One difference in study design that distinguishes our approach is the inclusion of patients with all types of acromial shapes. If subacromial space were a definite contributor to clinical outcome, one might have expected that those in the acromioplasty group with a Type-2 or Type-3 acromion would have shown more benefit from undergoing the additional procedure than those with an already flattened acromion (Type 1). Conversely, those in the group only undergoing rotator cuff repair with a Type-2 or Type-3 acromion should have had lower outcome scores as they did not receive the acromioplasty. The fact that neither of these outcomes was found provides further evidence that there may be limited benefit of acromioplasty with rotator cuff repair.
Another difference in our study is the examination of reoperation rates as a secondary outcome measure. The finding that four patients, particularly those with a Type-2 or Type-3 acromion, who did not undergo initial acromioplasty later required acromioplasty within the first two years after surgery is notable. The decision to offer additional surgical intervention was based on patient symptoms and not on postoperative imaging. Although unlikely, it is possible that clinical improvement may have occurred with further nonoperative therapy, given more time. It is also possible that the higher reoperation rate, including rotator cuff retears, occurred as a result of the unaltered acromion at the time of the initial surgery. Further study that includes follow-up imaging and patient-reported outcomes over a longer follow-up period is needed.
The strengths of this study include the prospective randomized design involving a single procedure performed by two shoulder surgeons at two separate centers for the correction of a single specific disease entity. The demographics of the study group reflect data identified in epidemiologic studies on the prevalence of rotator cuff tears in the general population and are comparable with other studies8,12,13,24, which enhances the generalizability of the findings in the present study to a broad population. Previous studies in this area have utilized either upper-extremity questionnaires or joint-specific functional outcome measures. Another strength of the current study involves the use of a disease-specific measure (the WORC index score) as the primary outcome, which enhances our ability to detect differences between groups. The WORC index questionnaire has been found to be a more responsive measure of change in patients with rotator cuff disorders than other shoulder-specific questionnaires25.
There are limitations to the study. First, some variation in surgical technique and postoperative rehabilitation may have occurred between the two study sites, although these variations might increase the generalizability of the findings. Second, there was greater loss to follow-up at the primary end point, twenty-four months after surgery, than was expected; eighteen (21%) of eighty-six participants were lost, which was more than the thirteen (15%) expected. Third, we do not have confirmation of rotator cuff healing status nor evaluation of the amount of bone resection as no imaging was obtained postoperatively. Previous studies have not included these evaluations, and the reliability of radiographs in determining the amount of bone resection is questionable. We focused on the use of validated functional outcome scores and quality-of-life indices as patient-centered outcomes that may be considered a more important measure of surgical success than imaging studies alone. Finally, our assessment of acromial morphology was a subjective measure that was based solely on preoperative radiographs. Therefore, subgroup analyses may have been affected by variability in the classification of acromion type within and between surgeons26.
Our final follow-up period was twenty-four months, but additional follow-up to observe longer-term outcomes could lead to a greater understanding of the role of acromioplasty in improving patient outcome and function. It is reasonable to expect that a proportion of patients may show a decline in function that may not be evident at twenty-four months. Future research with sufficient power to determine the relationship among acromion type, acromioplasty, and surgical outcome of arthroscopic rotator cuff repair is needed. This could lead to a better understanding of whether our observation of a higher reoperation rate in the group managed without acromioplasty was a true finding or whether it represented a type-I error.
In conclusion, we observed significant improvements in quality-of-life and functional outcome scores in patients who underwent rotator cuff repair with an acromioplasty and those who underwent a repair only. We did not observe any significant differences in patient-reported pain, function, and quality-of-life scores between the groups up to two years after surgery. The findings of this study do not support the routine use of acromioplasty as an adjunct to arthroscopic cuff repair. A higher reoperation rate was observed in patients who had the rotator cuff repair without acromioplasty. Further study that includes follow-up imaging and patient-reported outcomes over an extended period is needed.
Note: The authors thank Dr. Ahmed Abdoh, University of Manitoba, for his statistical expertise. They also thank Kimberly Bell for her contribution as study coordinator at the Ottawa Hospital, and the University of Manitoba Gibson Fund and Pan Am Clinic Foundation for their support of this project.