Interval slide was first introduced in open surgery by Bigliani et al.23) in 1992, and Tauro24) popularized arthroscopic interval slide, which involves release of the torn, retracted supraspinatus tendon from the rotator interval for improved mobility.25) Lo and Burkhart26) defined the anterior interval slide as release from the rotator interval and the posterior interval slide as release of the interval between the supraspinatus and the infraspinatus tendons. They reported excellent outcomes of repair of severely contracted, massive RCTs with an interval slide technique for increased mobility of the supraspinatus tendon in nine shoulders. At a mean follow-up of 17.9 months, the mean UCLA score increased from 10.0 preoperatively to 28.3 postoperatively. Active forward flexion and active external rotation improved from a mean of 108° preoperatively to a mean of 146.1° postoperatively and from a mean of 24.4° preoperatively to a mean of 35.0° postoperatively, respectively.

However, there is some concern that interval slide techniques can cause devascularization of the supraspinatus. Kim et al.27) compared the 2-year follow-up results of complete repair by an interval slide with those of partial repair following margin convergence. The SST, American Shoulder and Elbow Surgeons (ASES) score, and UCLA score improved in both groups without showing and significance intergroup difference. In addition, 6-month MRI revealed a retear in 91% of the complete repair group. Therefore, they concluded that complete repair using an interval slide technique has no advantages over partial repair in terms of treatment outcome. Berdusco et al.28) reported a 55% retear rate using an interval slide technique in patients with massive, contracted, and immobile rotator cuff tears, followed by MRI at an average of 25.2 months after surgery.

Based on our clinical experiences, we know that interval slide does not allow for sufficient mobilization and the nonanatomical reattachment in the posterior interval slide may increase the risk of devascularization of the supraspinatus tendon. It is our understanding that mobility of the retracted rotator cuff can be sufficiently improved with a release of the coracohumeral ligament from the coracoid under surface (Fig. 2) instead of a release from the rotator interval. For these reasons, we do not use the traditional interval slide techniques at our institution.

Burkhart et al.29) and Burkhart30) suggested margin convergence (Fig. 3) as a method of reducing strain and increasing the fixations strength. In a 2-year follow-up study by Shindle et al.,31) the margin convergence technique performed in U-shaped RCTs resulted in improvement in the ASES score (from a mean of 50.0 preoperatively to a mean of 83.3 postoperatively), active forward elevation (from a mean of 156.2° preoperatively to a mean of 168° postoperatively), and active external rotation (from a mean of 54.4° preoperatively to a mean of 57.1° postoperatively). However, the 2-year follow-up ultrasonography revealed that only 46.2% of the RCTs were healed. In a biomechanical study of margin convergence, Mazzocca et al.32) reported that the technique decreased rotator cuff strain at all degrees of rotation and gap size in 20 cadaveric shoulders. In our previous comparison study of partial repair and complete repair, the shoulders were biomechanically less stable after partial repair in spite of good clinical outcomes. Furthermore, we emphasized the importance of anterior margin convergence in decreasing gap formation and restoring range of motion and posterior infraspinatus repair in restoring abnormal humeral head apex kinematics.22) However, we are doubtful about the efficacy of margin convergence for severely degenerated poorly vascularized rotator cuffs. If necessary, margin convergence performed at the muscle-tendon junction may be more conducive to biological healing of the rotator cuff.

Medialization can be considered when reattachment of a torn tendon to the anatomical footprint is not feasible in irreparable RCTs. In a biomechanical study by Liu et al.,33) 3 mm and 10 mm medial advancement of the tendon had a negligible impact on the moment arm during elevation, whereas 17 mm medial advancement negatively affected biomechanics of the shoulder by significantly reducing the moment arm in 10 fresh-frozen cadaveric shoulders. Yamamoto et al.34) reported that ≥ 10 mm medialization in 10 cadaveric shoulders resulted in significant limitation in range of motion. Kim et al.35) reported that medialized repair in 35 RCT patients improved clinical outcomes: the mean VAS score, from 6 preoperatively to 2 postoperatively; the mean active forward elevation, from 134° preoperatively to 150° postoperatively; the mean active external rotation, from 47° preoperatively to 55° postoperatively; the mean Constant score, from 53.5 preoperatively to 79 postoperatively; the mean ASES score, from 51 preoperatively to 82 postoperatively; and the mean UCLA score, from 14 preoperatively to 28 postoperatively. The retear rate at the last follow-up was 17% in the study. Therefore, we suggest that < 10 mm medialization can be a viable option for irreparable RCTs without significant negative consequences on biomechanics and limitation on the range of motion of the shoulders.

Considering the postoperative biomechanical stability demonstrated in previous clinical and biomechanical studies, complete repair should be attempted whenever possible. The above-described techniques including interval slide, margin convergence, and medialization can aid in reducing tension on the tendon, minimizing gap formation and restoring normal kinematics of the humeral head for complete repair.

Latissimus dorsi transfer can be used to replicate posterior force couple of the infraspinatus and the teres minor muscle. During the procedure, the insertion of the latissimus dorsi on the lesser tuberosity is transferred to the greater tuberosity of the humerus, which converts this internal rotator of the shoulder to an external rotator. In the presence of a posterior RCT, the relatively high muscle strength of the anterior rotator cuff results in increased internal rotation and a significant external rotation deficit. Under such circumstances, transferring an internal rotator into an external rotator was suggested as a method to maintain force couples and restore normal kinematics in irreparable RCTs. However, it should be noted that after a latissimus dorsi transfer, the line of pull is more vertical than the force vector of the posterior rotator cuff muscles. In a biomechanical research, Oh et al.52) showed that latissimus dorsi transfer is advantageous for restoring internal/external rotational range of motion and balance of the glenohumeral joint. At the same time, they pointed out that limited excursion at 60° of abduction may cause an overcompensation phenomenon and increased contact pressure, further deteriorating normal biomechanics of the shoulder. In addition, abnormal kinematics of the humeral head can result in persisting pain and glenohumeral osteoarthritis. In a study by Aoki et al.,53) active forward flexion improved after a mean of 35 months of latissimus dorsi transfer, but osteoarthritic changes were observed in 41%. Gerber et al.54) also reported that range of motion, muscle strength, and function scores improved after latissimus dorsi transfer (n = 69), but osteoarthritic changes progressed in 30%. In addition, they found that insufficiency of the subscapularis was associated with unfavorable postoperative outcomes. In a biomechanical study by Werner et al.,55) the relationship between the presence of a subscapularis tendon tear and inferior results of latissimus dorsi transfer was demonstrated in a cadaveric model.

Recently, arthroscopically-assisted latissimus dorsi transfer has been shown to provide favorable clinical outcomes.56) The long-term clinical outcome of the procedure was also promising in a recent study: successful functional improvement was observed at a mean of 9.3 years with only 10% of clinical failure.57) As Henseler et al.58) suggested, synergistic muscle activity of the latissimus dorsi during arm abduction and external rotation confirmed at 1-year electromyography may indicate that restoration of active external rotation is achieved by active muscle contraction of the transferred latissimus dorsi rather than passive tenodesis effect.

It has yet to be elucidated whether favorable clinical outcome after latissimus dorsi transfer is due to active muscle contraction rather than passive tenodesis effect. However, based on the 10-year follow-up results and anatomical biomechanical research, we believe latissimus dorsi transfer can be used with success to improve the outcome of RCT repair when performed exclusively in relatively young patients without osteoarthritis or subscapularis insufficiency and with sufficient latissimus dorsi excursion. Since the procedure demands a long rehabilitation process, it is difficult to accurately predict the extent of recovery.

Pectoralis major transfer is used for anterosuperior RCTs. It was first described by Wirth and Rockwood in 1997 in patients with irreparable RCTs.59) Resch et al.59) reported that pain relief and improvement in the Constant score were observed at a mean of 28 months after the procedure in 12 patients. In a study by Gavriilidis et al.,60) in spite of the improved pain and Constant score, there was no increase in the range of motion in 15 patients at a mean of 37 months. Thrirteen of the total patients were available for MRI during the follow-up: the transferred pectoralis major muscle was intact in 70% of them and ruptured in 15%. Jost et al.61) performed 30 pectoralis major transfers and followed them for a mean of 32 months. They reported improved pain, range of motion, and Constant score but suggested a subscapularis tear combined with a supraspinatus tear was associated with unfavorable results.

Pectoralis major transfer can be a viable option for relatively young patients with isolated irreparable subscapularis tears without arthritis. However, it is a technically demanding procedure, and care should be taken to avoid injury to the musculocutaneous nerve.

Hemiarthroplasty can be a viable option in patients with ≥ 90° active forward elevation when pseudoparalysis is not present. It can be done in patients who have consistent pain associated with glenohumeral osteoarthritis after conservative treatment with sparing of the glenoid articular surface, and patients who have all types of glenohumeral arthritis with inadequate glenoid bone stock are also indicated for the procedure. Goldberg et al.72) reported that 76% of the 31 patients (34 shoulders) who underwent hemiarthroplasty for the treatment of RCT arthropathy obtained excellent clinical outcomes and improvement in range of motion at a mean of 3.7 years after surgery. The improvement was more noticeable in the long-term in patients who had active forward elevation of ≥ 90° before surgery. Zuckerman et al.73) also observed pain relief and improvement in forward elevation and external rotation in 87% of 15 hemiarthroplasties performed in elderly patients with advanced RCT arthropathy for a mean follow-up of 28 months. Visotsky et al.74) reported improvement in 60 patients who underwent hemiarthroplasty for the treatment of RCT arthropathy with a mean follow-up of 2 years: active forward flexion, from 56° preoperatively to 116° postoperatively; and active external rotation, from 8° preoperatively to 30° postoperatively. In a study by Sanchez-Sotelo et al.,75) approximately 67% of 30 patients (33 shoulders) who underwent hemiarthroplasty for RCTs showed pain relief and increased range of motion during a mean follow-up of 5 years. However, anterosuperior instability occurred in seven cases that had a history of iatrogenic injury to the coracoacromial arch associated with subacromial decompression. Hemiarthroplasty may also be effective in improving pain and range of motion in patients with irreparable RCTs accompanied by glenohumeral joint arthritis. However, it may result in bone loss and instability of the humeral head, which is significantly affected by the integrity of the coracoacromial arch. Labral degeneration may also have a significant impact on clinical outcome of hemiarthroplasty. In addition, the procedure is not considered beneficial for patients with chronic pseudoparalysis.

Since Paul Grammont introduced the biomechanical concept of medialization of the center of rotation and inferior translation of the humeral head in 1985, RTSA has continuously evolved into an alternative to total shoulder arthroplasty for various shoulder disorders that cannot be managed with the traditional arthroscopic procedure.76) In the procedure, medialization of the glenohumeral center of rotation increases the moment arm of the deltoid muscle and inferior translation of the rotation center lengthens the lever arm to optimize the efficiency of the deltoid muscle. RCTs with superior migration of the humeral head in the presence of osteoarthritis are standard indications for RTSA for optimal clinical outcome. In particular, it can be a solution for painful pseudoparalysis. Sirveaux et al.77) assessed the efficacy of RTSA in 80 shoulders with massive rupture of the cuff with a mean follow-up of 44 months: the Constant score improved from 22.6 preoperatively to 65.6 postoperatively; active forward elevation improved from 73° preoperatively to 138° postoperatively; and 96% of the patients had pain relief. Frankle et al.78) followed up 60 patients who had RTSA for the treatment of rotator cuff deficiency for a mean of 2 years: the mean active forward elevation increased from 55° preoperatively to 105.1° postoperatively; the mean external rotation increased from 41° preoperatively to 101° postoperatively; the mean VAS for pain improved from 6.3 preoperatively to 2.2 postoperatively. In a longterm follow-up multicenter study, Favard et al.79) reported the 10-year survivorship of RTSA in 484 patients with rotator cuff arthropathy as 89%. However, they additionally mentioned that the Constant score decreased over time, and the 10-year survivorship fell to 72% when the endpoint was set as the Constant score of 30 points. By contrast, in a recent study, the 10-year survivorship was close to 90% in 46 cases of RTSA performed in relatively young patients (< 65 years) with a mean follow-up of 93 months. The patients also obtained good clinical outcomes, such as improvement in shoulder function scores and pain relief.80) However, given that RTSA is considered as the last resort, it should be carried out when all other salvage procedures have failed. Furthermore, if possible, it should be an option exclusively for elderly patients (≥ 65 years).

Irreparable RCTs unaccompanied by osteoarthritis can also be a primary indication for RTSA. Boileau et al.81) evaluated 42 shoulders that were managed by RTSA after failed rotator cuff surgery with a mean follow-up of 50 months. In pseudoparalytic shoulders, the mean active forward elevation increased from 56° preoperatively to 123° postoperatively, and complications occurred in 12%. Mulieri et al.82) also performed RTSA for the treatment of 72 irreparable RCTs without glenohumeral arthritis. During a mean 52 months of follow-up, shoulder function scores improved, pain was alleviated, and the complication rate was 20%. Zumstein et al.83) emphasized the need to resolve the ambiguity with regard to the definition of complications and reoperations among studies. They suggested to differentiate a problem as an event that does not affect the patient's final outcome from a complication that influences the outcome. We reviewed 21 different papers published between 1995 and 2008 to assess complications based on their definition. Among 782 shoulders, complications occurred in 188 shoulders (24%). The most common complication was instability (4.7%), followed by infection (3.8%) and glenoid loosening (3.5%). Although scapular notching (35%) was the most common event, it was classified as a problem according to the definition. However, considering that scapular notching could potentially be a complication, more long-term follow-up studies are needed to determine the validity of the definition of complication.

RTSA is a good option for a multitude of shoulder disorders. However, considering the various complications described in the literature, a cautious approach is advised when determining the procedure for the treatment of irreparable RCTs with pseudoparalysis unaccompanied by osteoarthritis. Oh et al.84) compared the outcomes of arthroscopic repair of large-to-massive RCTs between patients with and without pseudoparalysis using propensity score matching. In both patient groups, range of motion improved postoperatively. More significant improvement in active forward elevation was observed in the pseudoparalytic group. The function scores showed no significant difference between the groups, and pseudoparalysis was reversed in 73% of the pseudoparalytic patients after repair. Thus, they argued that pseudoparalysis should not be considered the absolute indication for RTSA. They suggested that considering the high complications rates of RTSA, rotator cuff repair should the first-line treatment option even in the case of nonarthritic large-to-massive RCTs with pseudoparalysis. Denard et al.8586) retrospectively reviewed arthroscopic rotator cuff repair of massive RCTs in pseudoparalytic patients who underwent either primary rotator cuff repair (group I, n = 39) or revision repair (group II, n = 14). Group I showed improvement in forward elevation (from a mean of 49° preoperatively to a mean of 155° postoperatively); the mean duration of pseudoparalysis was 4.6 months; and pseudoparalysis was reversed in 90% of patients. In group II, the mean forward elevation improved from 43° preoperatively to 109° postoperatively, the mean duration of pseudoparalysis was 20.8 months, and pseudoparalysis was reversed in 43%. In summary, the procedure led to reversal of preoperative pseudoparalysis in 90% of patients who had no previous surgery and the duration of pseudoparalysis was significantly shorter in them. In another prospective multicenter study, reversal of pseudoparalysis was observed in 95% of 56 massive RCT patients with pseudoparalysis. There was no difference in the rate of reversal of pseudoparalysis between patients with an AI of ≤ 7 mm and those with > 7-mm AI and between patients with fatty infiltration of grade 3 or higher and those with less than grade 3.86) Similarly, Miyazaki et al.87) reported a pseudoparalysis reversal rate of 97.4%. Based on these studies demonstrating a 90% likelihood of reversal of pseudoparalysis, we believe that rotator cuff repair is a more reasonable strategy than RTSA as the first line of treatment even in massive RCT patients with pseudoparalysis.

However, Werner et al.88) reported that clinical outcomes of RTSA were better when it was a primary treatment than a revision of a failed repair. In the study, of the total 58 irreparable RCT patients with pseudoparalysis, the procedure was the primary treatment in 17 and a revision in the remaining 41. At a mean follow-up of 38 months, the former group of patients had a lower revision rate and higher Constant scores. All these studies underscore the importance of a comprehensive decision-making process where a treatment decision is not based on the consideration of a few clinical presentations but on various factors, such as patient's age, level of activity, and comorbidities.

Factors affecting the outcome of RTSA surgery are still controversial. By medialization of the center of rotation, RTSA minimizes the torque of the glenoid components and recruits more fibers of deltoid muscle to act as shoulder abductors. However, there are some problems in medialization of the center of rotation, such as scapular notching, restriction of external rotation of the shoulder, and instability due to cam effect. Recently, lateralized glenoid components have been used to reduce scapular notching and improve passive internal rotation and external rotation. However, when using a lateralized glenoid component, there is a concern that the torque at the baseplate-glenoid interface increases and the shear force increases at the glenoid fixation site, thereby increasing the failure rate. The 135° neck shaft angle and the 155° neck shaft angle also have advantages and disadvantages. The 135° neck shaft angle is more anatomical, less scapular notching, and favorable for passive internal rotation and external rotation. The 155° neck shaft angle has good joint stability and is effective for lengthening deltoid muscle and improving ability of forward elevation. Repair of subscapularis with RTSA is controversial. Vourazeris et al.89) reported that primary RTSAs with or without subscapularis have similar clinical outcome scores, range of motion, strength, and rates of complications, including dislocations at 3 years of follow-up. Friedman et al.90) reported a significant increase in internal rotation in the subscapularis repair group compared to the non-subscapularis repair group. In addition, active abduction and passive external rotation were significantly increased in the non-subscapularis repair group. And no difference was noted in the complication or scapular notching rates between cohorts. Werner et al.91) reported that ASES scores were significantly less improved in patients who underwent subscapularis repair and glenosphere lateralization. The version of the humeral component plays an important role in range of motion and impingement in RTSA. Anteversion can significantly reduce the amount of external rotation that can be achieved after RTSA. Rhee et al.92) reported that range of motion after RTSA was not significantly different between 0° and 20° of humeral component retroversion angle and most daily activities did not differ between the two groups, but the activity score related to the internal rotation was better in the 0° retroversion angle group. The anatomical differences of the shoulders between Asian and Western populations should be considered: the Korean population has a smaller humerus length and a humerus head diameter than the Western population; the humerus neck shaft angle is similar between the two populations; the humerus retroversion is greater in the Korean population than in the Western population; and the Korean population has a smaller glenoid diameter and a larger lateral extension of acromion than the Western population.93) Because of these anatomical differences, it seems reasonable for Koreans to consider a lateralized prosthesis. Considering such differences between Asian and Western people as well as individual differences, it is necessary to implement an individualized approach.

Arthroplasty can be considered a first-line treatment for irreparable RCTs with advanced osteoarthritis. Hemiarthroplasty may provide satisfying outcomes in rotator cuff arthropathy patients without pseudoparalysis when the coracoacromial arch is intact. RTSA is the optimal solution among elderly patients with RCTs accompanied by pseudoparalysis and osteoarthritis for alleviation of pain and restoration of active elevation function. Factors that should be considered in arthroplasty include the patient's age and desired activity level, and presence of pseudoparalysis and osteoarthritis. For arthritic irreparable RCTs without pseudoparalysis, other procedures enabling joint preservation should be attempted first. In a treatment decision process, it is important to consider various relevant factors and have sufficient communication with patients to determine the most beneficial treatment strategy for them.