A retrospective review was conducted of patients who underwent arthroscopic-assisted LTT transfer using an Achilles tendon allograft following failed SCR between January 2018 and May 2020. Patients were included if they met the following criteria: (1) structural failure of SCR, defined as a graft retear confirmed by magnetic resonance imaging (MRI) or functional failure of SCR, characterized by persistent pain and functional limitations despite an intact graft on MRI; (2) minimal glenohumeral joint arthritis, classified as Hamada grade 1 or 2; (3) intraoperative arthroscopic findings indicating irreparable supraspinatus, infraspinatus with/without teres minor tendons that could not be reduced to their anatomical footprint; and (4) torn supraspinatus and infraspinatus with/without teres minor associated with advanced fatty degeneration, corresponding to Goutallier grade 3 or higher, and severe tendon retraction to the glenoid level, as defined by Patte stage 3, as assessed by MRI. Exclusion criteria included insufficient clinical data, and loss to follow-up ( Fig 1 ).

Flowchart showing patient selection for this study. LTT, lower trapezius tendon; SCR, superior capsular reconstruction; MRI, magnetic resonance imaging; f/u, follow-up

Surgery was performed at Mayo Clinic (Rochester, MN) by a single experienced shoulder surgeon (B.E.) using a standardized technique, as described in a previously published paper, with modifications tailored to revision surgery following failed SCR. ^, Under general anesthesia, patients were positioned in the beach-chair position. Subscapularis tendon tears (n = 5; 27.8%) were addressed with concomitant repair at the time of LTT transfer. A 4–5-cm straight incision was made just inferior to the scapular spine to harvest the LTT. Dissection was carried through the subcutaneous fat to expose the obliquely oriented fibers of the lower trapezius. The tendon was carefully released from the underlying infraspinatus fascia, with meticulous attention to avoid injury to the spinal accessory nerve by limiting dissection near the medial border of the scapula. Once isolated, the tendon was secured with a traction suture at the end.

Arthroscopic evaluation and preparation of the glenohumeral and subacromial spaces were then performed. All remnants of the previously implanted SCR graft, including residual suture materials, were thoroughly debrided to expose the native footprint on the greater tuberosity. The footprint was prepared with a burr and shaver to enhance graft healing. An Achilles tendon allograft was used to bridge the harvested LTT to the humeral footprint. After removing the bony section of the Achilles tendon allograft, the calcaneal end of the graft was prepared with nonabsorbable sutures at each edge. A fascial slit was created in the infraspinatus region, and the graft was introduced into the subacromial space using an arthroscopic grasper. The allograft was positioned over the prepared footprint and secured to the infraspinatus footprint using multiple knotless anchors to ensure stable fixation. With the arm maintained in 60° to 90° of abduction and full external rotation, the proximal end of the allograft was split and attached to the native LTT using the Pulvertaft weave technique, ensuring no overtensioning. Excess graft tissue was excised. The procedure concluded with the arm positioned in adduction and external rotation. A custom external rotation brace was then applied before the emergence from anesthesia.

Postoperative rehabilitation followed a protocol adapted from previously published guidelines. ^, Patients were immobilized in a custom external rotation brace (gunslinger type) for 6 to 8 weeks, maintaining the shoulder in 40° to 60° of external rotation to protect the repair and promote graft healing. From weeks 8 to 16, passive- and active-assisted ROM exercises were initiated, with internal rotation restricted to minimize tension on the graft. After 16 weeks, internal rotation restrictions were lifted, and patients resumed daily activities within a pain-free range. A progressive strengthening program targeting the periscapular and rotator cuff muscles began at 24 weeks. Most patients were cleared for full, unrestricted activity by 6 months, depending on functional recovery.

Clinical evaluations were performed both preoperatively and postoperatively, encompassing assessments of pain, ROM, and standardized functional outcome measures. Pain intensity was measured using the visual analog scale (VAS). Functional outcomes were evaluated using patient-reported outcome measures (PROMs), including the Subjective Shoulder Value (SSV), Constant score, and the American Shoulder and Elbow Surgeons (ASES) score. ROM was assessed with a hand-held goniometer by an experienced clinician to maintain measurement consistency.

Radiologic assessment included standardized preoperative and postoperative shoulder radiographs, as well as preoperative MRI to evaluate the integrity of the previously implanted SCR graft. Postoperative integrity of the transferred LTT was evaluated both clinically and radiologically at the final follow-up. Clinically, the extended LTT–Achilles tendon construct was visible and palpable beneath the skin, indicating graft continuity. This finding was further confirmed by using musculoskeletal ultrasonography to assess the structural integrity of the reconstruction. The data were analyzed with sex disaggregation to assess its impact on patient-reported outcome measures, range of motion, complications, and demographic characteristics. Patients who underwent conversion from LTT transfer to RSA were included in the outcome analysis up to their final follow-up assessment prior to the RSA procedure.

Statistical analyses were performed using SPSS software for Windows (version 11.0; IBM Corp., Armonk, NY). A power analysis was carried out to ensure that the current study was adequately powered to detect the primary variable (ASES) of the study using G∗Power software (version 3.1.9; Heinrich Heine University), with an α level of 0.05 and statistical power of 0.9. The required sample size was calculated to be at least 4 participants. Continuous variables were analyzed using paired t -tests or Wilcoxon signed-rank tests, depending on the normality of distribution. Categorical variables were compared using Chi-square or Fisher’s exact tests, as appropriate. For minimal clinically important difference (MCID) was determined using the 0.5 SD distribution-based method. A P value <.05 was considered statistically significant.