Abrams 2014.

Methods	Design: single‐centre, parallel, two‐arm, randomised controlled trial Setting: one orthopedic centre in the USA Timing: October 2007 and January 2011 Interventions: rotator cuff repair with acromioplasty versus rotator cuff repair without acromioplasty Sample size: a post hoc power analysis was conducted before enrolment closure based on previous study, which reported ASES scores at 2‐year follow‐up in similar comparison and found a 95% confidence interval of ‐13.0 to 3.2 points for the score difference. Using the reported SD and MID of 10 in the ASES score, and group size of 45 patients each, we would be powered at 81.1% to detect a difference. The authors set a goal of 50 patients in each of the groups for the current investigation to account for attrition. Analysis: per protocol analysis (the authors excluded participants after randomisation if they did not have full‐thickness tears in the operation)
Participants	Number of participants 156 assessed for eligibility 42 excluded; 21 for not meeting inclusion criteria, 15 refused to participate, 6 for other reasons 114 randomised (65 to repair and subacromial decompression (SAD) and 49 to repair) At 6, 12 and 24 months data available for 52 in repair and SAD and for 43 in repair group Inclusion criteria A full‐thickness tear of the superior rotator cuff At least 18 years of age Exclusion criteria Isolated subscapularis tear Partial tears Irreparable tears Partial repair Previous repair (revision surgery) Baseline data Rotator cuff repair with acromioplasty Mean (SD) age 59.6 (8.2) Number (%) females 15 (29) Number (%) smokers 2 (4) Number (%) diabetes 5 (10) Number (%) workers compensation 8 (15) Number (%) acute tears 27 (52) Number (SD) of involved tendons 1.4 (0.6) Mean (SD) size of tear, mm 25.8 (10.8) Mean (SD) retraction of tendon, mm 12.3 (11.9) Mean (SD) ASES score 48.8 (18.2) Mean (SD) SST score 5.2 (2.6) Mean (SD) UCLA score 10.7 (2.9) Mean (SD) VAS for pain 4.4 (2.3) Mean (SD) Constant score 51.9 (17.2) Rotator cuff repair without acromioplasty Mean (SD) age 58.0 (8.0) Number (%) females 16 (37) Number (%) smokers 4 (9) Number (%) diabetes 3 (7) Number (%) workers compensation 7 (16.3) Number (%) acute tears 24 (56) Number (SD) of involved tendons 1.3 (0.5) Mean (SD) size of tear, mm 25.8 (8.5) Mean (SD) retraction of tendon, mm 12.5 (10.4) Mean (SD) ASES score 55.1 (19.1) Mean (SD) SST score 5.1 (3.0) Mean (SD) UCLA score 11.8 (2.8) Mean (SD) VAS for pain 3.8 (2.5) Mean (SD) Constant score 48.3 (17.1)
Interventions	All repairs were performed arthroscopically by four trained surgeons. Repair with acromioplasty group Preferred suture passing techniques typically included use of a curved shuttling device through the posterior portal while viewing from lateral. Details including type and number of anchors used as well as repair configuration were recorded. Those in the acromioplasty group underwent release of the coracoacromial ligament and flattening of the anterior inferior surface of the acromion. This was performed with a combination of shaver and electrocautery use to remove bursal tissue and define the lateral border and undersurface of the acromion. A motorised bur was then used to remove bone until the undersurface of the acromion was flat when viewed from the lateral portal using a posterior cutting block technique. Rotator cuff repair was performed in standard fashion by use of a combination of suture passing devices. Extensive releases were not performed with the exception of rotator interval releases to assist with reduction of retracted tears. Repair without acromioplasty group Repair was performed similarly as in repair and SAD group. Extension of bursectomy not described Both groups All patients were discharged on the day of the surgery. Physical therapy was standardised for both groups; it was instituted approximately 1 to 2 weeks after surgery, after the first postoperative visit, and focused on passive motion only. Sling immobilisation when patients were not performing physical therapy or a home exercise program was continued for 6 weeks after surgery. Active range of motion was begun at 6 weeks, and strengthening was deferred until 12 weeks postoperatively.
Outcomes	Outcomes were assessed at 6 months, 12 months and two years. Outcomes: Constant score (0 to 100, higher score indicates better function) ASES score (0 to 100, higher score indicates better function) Simple shoulder test (0 to 12 higher indicates better) UCLA score (0 to 35 higher indicates better function) VAS (0 to 10 higher score indicates worse pain) SF‐12 Number with recurrent tear Outcomes used in this review Mean pain; pain in VAS Mean function; Constant score Treatment failure; number of participants with recurrent tear.
Source of funding	Several authors report receiving funding from medical industry but the authors do not report any source of funding for this particular work.
Notes	Trial registration: N/A Data analysis: we used data from long‐term (mean of 92 months) follow‐up in repair failure analysis (Analysis 2.3). Withdrawals: in repair and SAD group, 13/65 (20%) were lost to follow‐up. In repair‐only group 6/49 (12%) were lost to follow‐up Re‐operations: in repair and SAD group, 1 re‐repair. In repair‐only group 3 re‐repairs and 1 capsular release and biceps tenotomy. Adverse events: the authors did not report if there were any adverse events besides re‐repairs. Serious adverse events: the authors did not report if there were any serious adverse events or not.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Random sequence generation not described.
Allocation concealment (selection bias)	Unclear risk	Quote: "Patients were randomised to acromioplasty or non‐acromioplasty groups via a sealed envelope" No further safe guards were reported.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	It was not reported if the hospital staff or the study personnel were blinded to the intervention.
Blinding of outcome assessment for self‐reported outcomes including pain, function and global assessment (detection bias)	Low risk	Quote: "Assignments were not disclosed to the patient". The postoperative treatment and data collection were similar in both groups.
Blinding of outcome assessment for incidence of full‐thickness tears at follow‐up (detection bias)	Unclear risk	Trial authors did not report this outcome.
Incomplete outcome data (attrition bias) All outcomes	Low risk	6/49 (12%) in repair versus 13/65 (20%) in repair with acromioplasty group were lost to follow‐up with reasons not reported. Not likely to bias the results significantly
Selective reporting (reporting bias)	High risk	Protocol not available, adverse events incompletely reported. SF‐12 measured but not reported.
Other bias	High risk	There was imbalance in the treatment received. Quote "An increased number of patients who were enrolled in the non‐acromioplasty group were examined intraoperatively and excluded for a lack of full‐thickness rotator cuff tear" It is unclear if this could bias the outcomes in favour of acromioplasty.