The inclusion criteria were as follows: (1) patients with large (3.0–5.0 cm) or massive (>5.0 cm) RCTs who had undergone arthroscopic repair, (2) those who were available for magnetic resonance imaging (MRI) preoperatively, (3) those who underwent the appointed postoperative rehabilitation program, and (4) those who were available for a postoperative follow-up after 1 year. However, the exclusion criteria were as follows: (1) patients who had a successful primary repair during surgery, (2) those who had irreparable SSC tears (> Lafosse classification type 2), (3) those who had other orthopedics-associated or systemic diseases, and (4) those who could not be followed-up.

From 2017 to 2020, 179 patients with RCTs underwent arthroscopic rotator cuff repair. According to our criteria, those with small/medium-sized tears (n=100), those who experienced primarily repaired large or massive tears (n=79), those who had a successful primary repair during surgery (n=37), those with RCTs that had subscapularis tendon involvement (>Lafosse classification type 2, n=5), patients with other orthopedics-associated or systemic diseases (n=2), and those who could not be followed-up (n=1) were excluded. Consequently, 34 patients who met the inclusion criteria were examined and then included in the present study. There were 3 large and 31 massive tears; their mean width and length were 50.9 mm and 50.6 mm, respectively. The mean age at the time of surgery was 68.4 years (range, 57–76 years), with a mean follow-up period of 12.6 months (range, 12–18 months). Further details are shown in Table 1.

After the induction of general anesthesia, the patients were placed in the beach chair position. The tendon stumps in all patients were proximally retracted beyond the glenoid edge (Fig. 1A). Despite thorough mobilization, including capsular/coracohumeral ligament release, these stumps did not reach the footprint of the greater tuberosity (Fig. 1B). Next, an approximately 4-cm straight incision was made on the scapula spina. The interface between the subcutaneous tissue and the infraspinatus (ISP) was released digitally. Using fingers and metal instruments, the margin of the ISP was gently separated from the medial border of the scapula, the posterior margin of the glenoid, the superior portion of the teres minor, and the scapula body (Fig. 2A). These procedures allowed the tendon stumps to easily reach the superior facet area near the bicipital groove (Fig. 2B). Finally, the approximated tendon was fixed using the suture bridge technique with two anchors in the medial row (Healix Advance; DePuy Synthes, Raynham, MA, USA) and two anchors in the lateral row (SwiveLock; Arthrex, Naples, FL, USA) (Fig. 1C).

In cases with osteophytes in the subacromial space, acromioplasty was performed (75.7%). Tenotomy of the long head biceps tendon was carried out when a tear of half or more of the width was present (45.9%). Partial tears of the upper subscapularis tendon were treated by shaving, not by repair; in our series, no full-thickness tears of this tendon were detected. The mean operative time was 115±23.2 minutes. Further details are shown in Table 1.

After surgery, the patients were placed in an immobilization sling in a neutral position for 8–10 weeks. All patients underwent postoperative regimens under the strict supervision of a physical therapist. We began passive elevation exercises and external/internal rotations in a supine position one week postoperatively. Active exercises were subsequently initiated in the sitting or standing position at 8–10 weeks postoperatively. Strength exercises were started at 4–5 months postoperatively, and the patients resumed their previous work activities 6 months postoperatively.

The functional outcomes were assessed by the Constant-Murley Score (Constant), University of California at Los Angeles (UCLA) Shoulder Score, and Japanese Orthopaedic Association (JOA) scores. Patient-reported outcomes were assessed using the Shoulder Index of the American Shoulder and Elbow Surgeons (ASES), Western Ontario Rotator Cuff Index (WORC), and JOA Shoulder 36 score version 1.3 (Shoulder 36). Range of motion (ROM) was evaluated before the surgery and then again at 12 months postoperatively using the active elevation, external rotation, and internal rotation. The internal rotation was measured as the highest vertebral body that the patient could reach with the thumb of the affected arm. The degree of pain (in motion and at night) was assessed using a visual analog scale (0–100 mm). Muscle strength was evaluated using a handheld dynamometer (MicroFET2; Hoggan Health Industries, Draper, UT, USA) in a sitting position with the hips and knees at 90° flexion. Measurements of the strength of the 40° abductor and the 90° abductor were performed in the shoulder joint’s internal and external rotation positions, respectively, along with the external rotation and internal rotation strength in the 0° abduction position. Each measurement was taken three times, and the average value was calculated. These functional evaluations were performed both before and at 12 months after the surgery.

Preoperatively, the tear size was measured by MRI as the maximum anteroposterior tear width on T2-weighted sagittal images and the maximum mediolateral tear length on T2-weighted oblique-coronal images [16]. Preoperative fatty infiltration of the rotator cuff muscles was assessed using the Goutallier Classification [17]. The postoperative cuff integrity was evaluated with MRI 12 months after surgery; images with Sugaya classification Types 4 or 5 were considered re-tears [18]. Subsequently, two observers blinded to this study independently assessed the structural outcomes. Reproducibility between these two observers showed “good” interrater reliability (ICC [3, 1]=0.85) and “excellent” intrarater reliability (ICC [1, 2]=0.92).

We used statistical software (R version 2.8.1; R Foundation for Statistical Computing, Vienna, Austria) to analyze the data. After confirming a normal distribution using the Shapiro-Wilk test, a paired t-test or Wilcoxon rank-sum test was utilized. The Wilcoxon rank-sum test was used to compare the Constant, UCLA, JOA, ASES, WORC, and Shoulder 36 scores, ROM, pain level, and muscle strength both before and after surgery. A p-value <0.05 was considered statistically significant.

Compared with the preoperative scores, the mean scores significantly improved 12 months postoperatively as follows: Constant (53.3±21.1 to 76.8±10.5), UCLA (15.6±3.6 to 27.8±6.7), ASES (51.8±18.3 to 89.1±13.5), and WORC (925.0±436.8 to 480.3±373.2) (all p<0.001). Although the range of motion of active elevation significantly improved after surgery (95.6°±51.1 to 146.9°±14.5; p<0.001), these improvements were not observed for external or internal rotation (Table 2). Additionally, while 44.1% of all patients exhibited pseudoparalysis before the surgery, they all had improved 1 year after surgery.

This study included 15 patients with preoperative pseudoparalysis (15/34, 44.1%) who experienced significant improvements in their functional and patient-reported outcomes after surgery. We observed no re-tears in these 15 patients. The results also showed that the preoperative pain level while in motion and at night significantly improved after surgery (41.4±25.2 to 6.7±14.9 mm and 21.8±24.8 to 4.3±10.3 mm; p<0.001), respectively (Table 2).

Additionally, the 40° abductor muscle strength during external rotation improved from 53.0±24.4 to 63.3±19.5 N (p=0.028), while the internal rotation improved from 60.9±24.0 to 70.7±20.4 N (p=0.042); however, these improvements were not observed at the other positions. These data are summarized in Table 2.

At the 1-year postoperative evaluation, the Sugaya’s classifications were type I in 8 patients, type II in 15 patients, type III in 9 patients, type IV in 1 patient, and type V in 1 patient (Table 2). Consequently, postoperative re-tears were noted in two shoulders (5.9%) that were types IV and V.