We evaluated outcomes after arthroscopic-assisted LD tendon transfer performed by a single surgeon between 2012 and 2016. Indications for surgery were as follows; 1) symptomatic, massive rotator cuff tear involving supraspinatus and infraspinatus tendons, and/or a partial upper subscapularis tendon tear, 2) an irreparable tear onto the medial margin of the greater tuberosity footprint after release and mobilization of retracted tendons, 3) advanced muscle atrophy (grade ≥2 by the Warner classification) and fatty infiltration (stage 3 or 4 by Goutallier’s classification) of the supraspinatus and infraspinatus by preoperative magnetic resonance imaging (MRI), 4) no or mild arthritis of the shoulder joint on simple radiographs (Hamada stage 1, 2, or 3), and 5) a physically active patient aged <65 years. During the study period, 72 consecutive patients with symptomatic, massive rotator cuff tear were surgically treated; 38 reparable tears were treated by standard arthroscopic rotator cuff repair with or without augmentation and 29 irreparable tears with advanced arthritis (Hamada classification stage 4 or higher) in elderly patients (≥65 years) were treated by reverse shoulder arthroplasty. The remaining 5 patients underwent arthroscopic-assisted LD tendon transfer and constituted the subjects of the present study. All arthroscopic-assisted LD tendon transfer candidates were counseled about possible treatment options by the treating surgeon and informed of the advantages and disadvantages of the technique. No patient that underwent this procedure was excluded from the analysis.

Surgeries were performed by a single shoulder surgeon (TKL) with the patient in the lateral decubitus position with the affected arm abducted at 70° with longitudinal traction under general anesthesia. Using standard portals (anterior, posterior, lateral and posterolateral portals), an arthroscope was introduced into the glenohumeral joint and subacromial space. Initially, the routine aspects of rotator cuff repair procedures, such as diagnostic arthroscopy, biceps tenotomy or tenodesis, and partially torn subscapularis tendon repair were performed. Subacromial decompression and acromioplasty were then performed in the subacromial space. Suture anchors used for previous rotator cuff repair, if present, were removed from the greater tuberosity, and a bone bed was prepared with slight decortication and microfracture. Tear size was measured using a calibrated probe in millimeters mediolaterally and anteroposteriorly. A massive tear was defined as a tear exceeding 50 mm in length in any direction or complete tear of the supraspinatus and infraspinatus tendons (involving two tendons posterosuperiorly). Reparability of retracted rotator cuff tears was then evaluated using a grasper to determine how much coverage of the original footprint was possible. If torn tendons were retracted to the glenoid and only approximated to the mid-humeral head level, not up to the articular margin of the greater tuberosity, retracted tendons were released and mobilized. Anteriorly, the coracohumeral ligament was released, and any adhesions of the supraspinatus tendon to the rotator interval and capsule were thoroughly removed to improve cuff mobility. Posteriorly, the interval between the supraspinatus and infraspinatus was identified, and interval tissues were released from the scapular spine (the interval slide incontinuity [10]) or sometimes, released completely through the interval. When torn tendons were not reduced to the articular margin of the greater tuberosity footprint at the completion of all releases both anteriorly and posteriorly the tear concerned was defined as irreparable. For such tears, we continued to perform LD transfer and started with LD tendon harvesting. Briefly, a curvilinear incision was made along the lateral border of the scapula and posterior to the lateral profile of the LD tendon. After identifying the LD muscle and its insertion on the proximal humeral shaft, the insertion was sharply detached from the humerus. The arm was then further abducted and internally rotated to facilitate exposure of the tendon insertion, and the tendon was then mobilized and pulled until it reached the acromion anterior edge (Fig. 1A). When this was not possible, the tendon was released proximally to the musculotendinous junction while taking care not to damage the neurovascular bundle. Krackow stitches were then placed along both borders of the tendon using nonabsorbable sutures (No. 5 Ethibond; Ethicon, Somerville, NJ, USA). As the harvested LD tendon is thin and flat, it was augmented with use of human dermal patch allograft (Megaderm; L&C BIO, Seongnam, Korea) in order to have a thicker and longer tendon. Next, we re-introduced an arthroscope into the subacromial space. Posteriorly, the interval between the deltoid and teres major was developed bluntly to facilitate tendon passage as confirmed by inserting a finger from the posterior incision and by visualization using an arthroscope in the subacromial space. The anterior and posterior suture limbs of the Krackow sutures of the LD tendon were then retrieved from the anterior portal, and the harvested tendon was passed into the subacromial space by pulling the sutures out of the anterior portal. The anterior border of the transferred tendon was then fixed to the anterior margin of the greater tuberosity footprint using two knotless anchors (4.5 mm, Bio-Swievelock; Arthrex, Naples, FL, USA [Fig. 1B]), and its medial and lateral borders were fixed to the greater tuberosity footprint in a double-row fashion using the conventional suture bridge technique. A total of 5 anchors were used to fix the transferred LD tendon. Mean surgical time was 202 ± 17 minutes), and mean anesthesia time was 250 ± 14 minutes). During the six weeks following surgery the arm is immobilized in 30˚ of abduction and neutral rotation using abduction brace. Stretching exercises are started after brace removal, and strengthening exercises from 3 months postoperatively.

Each patient underwent a preoperative clinical evaluation the day before surgery by an independent observer and scheduled for regular follow-up visits at 2 weeks, 6 weeks, 3 months, 6 months, 1 year, and 2 years after surgery. Preoperative and postoperative clinical outcomes were assessed using the visual analogue scale (VAS) pain scale, American Shoulder and Elbow Surgeon’s (ASES) scores, the University of California Los Angeles (UCLA) scale, and measured ranges of motion. Radiographic assessments included preoperative standard radiographs and MRI of the shoulder joint. At 6 months postoperatively, transferred LD tendon repair integrities were evaluated by MRI or ultrasound. All radiographic images were reviewed by the treating surgeon and an orthopedic resident.

The paired t-test and Wilcoxon’s signed-rank test were used to compare preoperative VAS, ASES, and UCLA scores and ranges of motion with those obtained at final follow-up. For all analyses, p-values of less than 0.05 were considered significant (p<0.05).

Mean age of the five patients was 55 years (range, 48–61 years). There were two males and right shoulders were affected in three patients. Two patients had a trauma history; one during ice hockey and the other by a slip-down while washing clothes. The three female patients were housewives and complained of shoulder pain and weakness during daily housework, especially difficulty in keeping their hands in space. One male patient worked in a market and his work involved heavy lifting. The other male patient was paraplegic as a result of spinal cord injury and was wheelchair-ridden. His shoulder was weight-bearing to move from bed to wheelchair, and he was a member of the national disabled basketball team. Therefore, all five patients were functionally demanding with regard to shoulder activities. Three of the five had previously undergone one or more previous rotator cuff repair surgeries.

Mean postoperative follow-up duration was 20 months (range, 12.0–27.2 months) (Table 1, 2) Mean VAS improved significantly from 6.6 ± 2.6 (range, 2–8) preoperatively to 1.8 ± 2.5 (range, 0–6) postoperatively (p=0.009). Mean ASES score increased from 67.6 ± 9.2 (range, 61–80) to 84.6 ± 15.1 (range, 57–93), and mean UCLA score from 18.0 ± 1.4 (range, 17–20) to 28.8 ± 8.5 (range, 13–34) (all p<0.001) (Table 2, 3). According to UCLA scores 2 patients had an excellent, 2 patients a good, and 1 patient a poor outcome. Active forward elevation improved from a mean of 162° ± 40° (range, 100°–180°) to 172° ± 13° (range, 150°–180°), active external rotation from 14° ± 26° (range, 0°–40°) to 31° ± 24° (range, 0°–50°), and internal rotation from L2–L3 level to T7–T8; however, no significant differences were found (p=0.128, 0.763, and 0.339, respectively; Table 3). No complication, such as hematoma or infection, occurred at donor sites, and no patient required revision surgery.

Postoperative imaging of the transferred LD tendon showed intact repair in 4 patients and rupture in the fifth (Fig. 2). Outcomes of the four patients with intact transfer were rated as excellent or good based on UCLA scores, and outcome in the fifth was rated as poor.