Among patients surgically treated by ARCR at our hospital, between September 2011 and June 2018, there were 53 shoulders (52 patients) presenting large or massive rotator cuff tears according to the DeOrio and Cofield classification with delamination. Inclusion criteria were large or massive rotator cuff tears undergoing ARCR and a minimum follow-up of 2 years. Exclusion criteria were patients underwent revision rotator cuff repair, and clinical findings of instability. Then, the subjects were classified into three groups for the treatment methods. The superficial and deep layers were fixed by en bloc suturing using suture anchors in the EMSB group, the superficial and deep layers were individually fixed by sutures similar to the method reported by Mochizuki et al. in the DLSB group [11, 14]. When the rotator cuff was unable to be covered up to the footprint even though it was sufficiently mobilized, the modified DP procedure  was added for fixation of the stump after DLSB following the method reported by Morihara et al., being designated as the DLSB + DP group. The EMSB group comprised 18 shoulders of 18 patients (12 males and 6 females at mean age of 69 years (range: 58–78 years)), the DLSB group comprised 24 shoulders of 23 patients (11 males and 12 females at mean age of 69.6 years (range: 49–87 years)), and the DLSB + DP group comprised 11 shoulders of 11 patients (4 males and 7 females at mean age of 72.9 years (range: 67–87 years)). The mean postoperative follow-up observation duration was 31.5 months (range: 24–67 months) in the EMSB group, 34.9 months (range: 24–72 months) in the DLSB group, and 26.7 months (range: 24–36 months) in the DLSB + DP group (Table 1).
The study was approved by an institutional review board, and the enrolled patients provided informed consent for participation in the study.
ARCR was performed in the beach chair position under general anesthesia in all patients. For retention of the upper limb, an arm holder (TRIMANO Arm Holder, Arthrex, Naples, FL, USA) was used. Five portals were mainly used for the operation (posterior, anterior, anterolateral, posterolateral, and upper anchor portals). For cases of subacromial decompression, subacromial synovectomy and subacromial osteophyte resection were mainly performed. Prior to repair, the rotator cuff was sufficiently mobilized. For the rotator cuff repair method, 1–2 medial anchors and 2–3 lateral anchors were set corresponding to the tear size using suture anchors (HEALIX, DePuy Synethes Mitek Sports Medicine, Raynham, MA, USA). In EMSB, the superficial and deep layers were fixed en bloc by passing the medial anchor thread. In DLSB, first, the medial anchor thread was passed through the deep layer to fix and suture by adding sliding knots and 2 half-hitches, as described by Mochizuki et al. , then all threads were passed through the superficial layer without cutting the anchor thread and fixed using the lateral anchor. In cases unable to be covered up to the footprint despite the rotator cuff being sufficiently mobilized, an approximately 3-cm transverse skin incision was made outward from the medial border of the spine of the scapula following the method reported by Morihara et al. . The modified DP procedure was added dissecting the supraspinatus and infraspinatus muscles, while maintaining continuity to the fascia of the rhomboideus muscle. Then, suture fixation of the stump was performed by DLSB.
In the EMSB or DLSB group, a postoperative shoulder abduction brace was attached for 3–4 weeks corresponding to the tear size. In the DLSB + DP group, the abduction brace was attached for 6 weeks after surgery. Passive ROM training was initiated on the day following surgery, and active ROM training was initiated after detaching the abduction brace. Resistance exercise was initiated 12 weeks after surgery.
We assessed the pre- and postoperative ROM (flexion, abduction, external rotation, internal rotation), Constant score as clinical outcomes, and global fatty degeneration index (GFDI) representing the preoperative degree of fatty infiltration into the rotator cuff observed on MRI as the mean of that of the supraspinatus muscle, infraspinatus muscle, and subscapularis muscle following the Goutallier classification [20, 21]. In addition, postoperative cuff integrity was evaluated using the Sugaya’s classification, and types IV and V were regarded as retear .
For statistical analysis, EZR (Ver. 1.50, Saitama Medical Center, Jichii Medical University, Saitama, Japan) was used. The pre- and postoperative ROM, and the constant scores in each group were analyzed using the Mann-Whitney U-test, the Kruskal-Wallis was used for pre- and postoperative comparisons among the 3 groups, and Fisher’s test was used for comparison of the incidence of retear among the 3 groups, setting the significance level to below 5%.