Better Short-Term Outcomes After Rotator Cuff Repair in Studies With Poorer Mean Shoulder Scores and Predominantly...

To perform a meta-analysis to explore factors associated with clinical and structural short-term outcomes in randomized and nonrandomized prospective studies of rotator cuff repair.

Medline, clinicaltrials.gov, and Mendeley were searched for literature published from January 2000 to December 2020 to identify randomized controlled trials (RCT) and nonrandomized prospective cohort studies (PCS) describing the outcome of surgical repair of full-thickness rotator cuff tears. Study quality was assessed by two independent reviewers. We calculated standardized mean difference (SMD) from baseline to follow-up in each trial arm, preferably at 12 months follow-up. Between-study heterogeneity of outcomes, small-study effects and rates of retear were assessed. Meta-regression was performed to estimate associations between prespecified variables and clinical and structural outcomes.

Outcomes in 64 RCT and 19 PCS trial arms were analyzed. Median age was 59 years. There was substantial between-study heterogeneity in clinical outcomes (SMD range: .42 to 6.44; I² = 93% in RCT, 88% in PCS) and summary estimates were not calculated. On the basis of multivariate analysis, better clinical outcome was associated with lower (worse) mean outcome value at baseline, smaller tear size, and lower proportion of large-massive tears (R² = 56 and 44%, respectively). Overall retear rate at median 13-month follow-up was 19.9% (interquartile range: 10-30). Higher mean age together with larger tear size and higher proportion of large-massive tears were associated with increased retear rates (R² = 33% and 58%, respectively). Clinical outcome was not significantly related to rate of retear.

Studies with lower mean outcome values at baseline and predominantly small- to medium-sized tears reported better clinical outcomes. Studies with higher mean age and a predominance of large-massive tears had significantly increased retear rates, but retear rates were not associated with clinical outcome.

Level II, meta-analysis of level I and II studies.