The role of the subscapularis following reverse shoulder arthroplasty (RSA) remains controversial as repair could restore adequate internal rotation and shoulder stability, but might compromise external rotation. The purpose of this multi-centre study, on a large cohort of RSAs performed using the deltopectoral approach, was to determine the effect of subscapularis preoperative status and intraoperative repair on range of motion, clinical scores and rates of complications at a minimum follow-up of 2 years.

Repair of a functional subscapularis grants better internal rotation and stability without compromising external rotation.

We retrospectively reviewed records of all RSAs (n=916) performed by 14 surgeons that participated in a large national society symposium, and excluded 234 operated by the anterosuperior approach, 42 operated for rheumatoid arthritis, fractures, instability or post traumatic sequalae, and eight that had adjuvant latissimus dorsi tendon transfer (LDTT). This left 632 RSAs, in which the subscapularis was detached in 594, and repaired after detachment in 495. Patients completed pre- and postoperative Constant Score and postoperative Subjective Shoulder Value (SSV). Active forward elevation, active external rotation (ER), and active internal rotation (IR) were recorded pre- and postoperatively. Complications following RSA that required conservative treatment, reoperation without implant removal, as well as reoperation with implant removal were recorded.

Of the initial cohort of 632 shoulders, 120 (19%) were lost to follow-up, 12 (2%) died due to causes unrelated to RSA, and 26 (4%) were revised with implant removal. Of the remaining 474 shoulders, 259 (55%) had a functional repaired subscapularis (Group FR), 59 (12%) had a non-functional repaired subscapularis (Group nFR), whereas 68 (14%) had a not repaired subscapularis (Group nR). Comparison of patient demographics revealed no significant differences among the three groups. Postoperative Constant Score (p=0.031) and SSV (p=0.016) were significantly different among the three groups, but effect sizes were small and not clinically relevant. Differences in postoperative ER as well as postoperative IR were not significant among the three groups: 91 (35%) gained functional IR in Group FR, 11 (19%) in Group nFR, and 13 (19%) in Group nR. Three shoulders (1.2%) reported subjective instability and 1 (0.4%) dislocated in Group FR, but none in either Groups nFR or nR.

Despite statistically significant differences in Constant Score and SSV among the three groups, the clinical relevance of these differences is negligible, so repair of functional subscapularis has little or no influence on functional outcomes. During RSA by deltopectoral approach, repair of a functional subscapularis could slightly improve internal rotation in some patients but does not compromise external rotation, forward elevation or clinical scores, while repair of a non-functional subscapularis, compared to non-repair, did not improve range of motion or clinical scores.

III; retrospective comparative.