Joint-preserving surgical treatment options for irreparable posterosuperior rotator cuff tear

The goal of arthroscopic partial repair is to restore the force couple consisting of the remaining subscapularis and infraspinatus tendons [5]. Complete anatomic reconstruction of the rotator cuff is no longer possible because of the advanced retraction. The upper edge of the infraspinatus tendon is reattached as far as possible to the original insertion site. The biceps tendon is either tenodesized or tenodized. It is known and has been published by Boileau et al. that treatment of the biceps alone significantly improves shoulder function and satisfaction in patients with massive irreparable RCTs [3]. In patients with irreparable posterosuperior rotator cuff tears (IPRCT) and concomitant shoulder stiffness, the treatment results of tendon transfer are poor; accordingly, from the authors’ point of view, shoulder stiffness is a relative contraindication for tendon transfers or SCR. In young patients with a painful and at the same time stiffened shoulder, APR can be considered as a treatment option and should be combined with arthroscopic capsular release. Postoperative mobilization begins passively from the first postoperative day, with actively assisted range-of-motion exercises from the 4th–6th postoperative week. Gentle strengthening to a maximum of 5 kg begins between the 10th and 12th postoperative weeks. Full weightbearing is allowed from the 16th postoperative week.

The clinical results of APR in short- and mid-term follow-up show significantly improved shoulder function and patient satisfaction [8, 20, 27, 33]. Interestingly, the 5‑year outcomes of patients treated with APR for irreparable massive rotator cuff tear are comparable to the outcomes of patients who underwent anatomic arthroscopic reconstruction of reparable massive RCTs [21]. The integrity of the teres minor was found to be an important predictive factor [1, 29]. Radiologically, re-rupture of the partial repair was described in 49% [23]. No clinical outcome differences were seen between patients after APR versus arthroscopically assisted latissimus dorsi tendon transfer (LDT) during a short-term follow-up period [1]. A randomized trial demonstrated no significant difference in clinical treatment outcomes 2 years after APR vs. SCR in 41 patients [19].

According to SCR inventor T. Mihata, implantation of an auto- or allograft membrane that replaces the superior capsule and the irreparable rotator cuff leads to improved vertical stability and thus reduces subacromial impingement [25]. Furthermore, the implanted membrane acts as an anchor for the remaining subscapularis and infraspinatus tendons, leading to an increase in their tendon tension and thereby restoring the force couple [26].

The procedure is performed arthroscopically in a beach chair position. At the authors’ institutions, a 3-mm dermal allograft is used. After thorough subacromial debridement and arthroscopic biceps tenotomy or tenodesis, partial repair of the remaining rotator cuff is performed if possible. The superior glenoid is debrided and exposed to fix the allograft with two to three anchors medially between the 10 and 12 o’clock positions. Fixation on the lateral aspect of the greater tuberosity is achieved with a double-row suture bridge technique with the arm held at 45° of abduction. Postoperatively, patients are immobilized in a 45-degree abduction pillow for a total of 6 weeks, starting with passive range of motion between 4 and 6 weeks and active assistive range of motion between 6 and 8 weeks. Light weightbearing is allowed after 10 to 12 weeks. Unlimited weight bearing is allowed 16 weeks postoperatively.

Clinical outcomes from the authors’ clinic (Massachusetts General Hospital) were published in the Journal of Shoulder and Elbow Surgery in 2019 [34]. The results were exceedingly modest. In total, SCR surgery was performed in 65 patients by 6 different fellowship-trained shoulder surgeons within the period from January 2015 to November 2017; however, 31 patients had to be excluded as they were not available for 12-month follow-up. Of the 34 patients included, 65% were dissatisfied with the treatment outcome at a mean follow-up of 12 months. The 1‑ and 2‑year failure-free survival rates were 34 and 16%, respectively. The 1‑ and 2‑year “free of reoperation rate” was 64 and 44%, respectively. There was no significant improvement in shoulder function or pain with SCR surgery. Risk factors for a poor treatment outcome were the extent of fatty infiltration in the infraspinatus muscles and the number of SCR operations performed by the operating surgeon (< 10). It should be emphasized that numerous level III and IV studies exist that show good to excellent short- and mid-term results, even in patients with preoperative pseudoparalysis [4, 6, 24].

Nevertheless, the poor treatment results at the authors’ institution (Massachusetts General Hospital) have led to a drastic reduction of the indication for SCR, which is now only considered in patients with strictly isolated supraspinatus tendon tears or type A lesions. The majority of patients—especially those with advanced fatty degeneration of the posterosuperior rotator cuff muscles (infraspinatus and teres minor)—are treated according to the treatment algorithm with tendon transfer.

LTT was originally developed to treat patients with brachial plexus palsy [2, 11]. These patients suffer in particular from a lack of external rotation due to failure of the external rotator muscles (infraspinatus and teres minor). In addition, cervical nerve root 5 is frequently involved, which causes dysfunction of the deltoid muscle. Prior to development of LTT, these patients with deltoid dysfunction were treated with LDT. A common complication in these patients was posterior (sub)dislocation, which led to the search for an alternative tendon transfer to the LDT [2, 11]. In this regard, LTT was developed by Elhassan to address these patients with less complications [13]. It was shown biomechanically that LTT transfer, leaving the latissimus dorsi tendon intact, did not present posterior instability problems in these patients with deltoid dysfunction [12].

Subsequently, this technique was also used in patients with irreparable rotator cuff rupture. Biomechanically, the force vector of the inferior trapezius portion was found to be similar to that of the infraspinatus and teres minor muscles. Another advantage of the lower trapezius muscle is its synergistic action with the infraspinatus and teres minor, as the muscle contracts during glenohumeral external rotation. A disadvantage of the lower trapezius muscle is the reduced excursion of the muscle as well as the short tendon length, which makes direct fixation to the greater tuberosity impossible. Therefore, interpositional tenodesis with an allograft or autograft is necessary.

The technique is described very clearly in the publication by Wagner et al. [14]. Here, the authors wish to point out the most important steps and possible sources of error, which are crucial for the outcome of the operation in their opinion.

In the study by Elhassan et al., a significant improvement in function, pain, and subjective shoulder score was observed in 32 of 33 patients (mean age 53 years) after a mean follow-up duration of 47 months [13]. The better the active and passive mobility preoperatively, the more likely a good treatment outcome can be assumed. Accordingly, from the authors’ point of view, shoulder stiffness is a (relative) contraindication for tendon transfers. Another study by Elhassan demonstrated significant improvement in shoulder function and patient satisfaction 12 months postoperatively in 90% of a total of 41 patients [14]. True pseudoparalysis was present in 46% (19 patients) preoperatively. In 18 of the 19 pseudoparalytic patients, the pseudoparalysis resolved and a significant improvement in shoulder function and patient satisfaction was achieved.

The pioneering description of LDT for chronic posterosuperior RCTs was presented by Gerber in 1988 [17].

The latissimus dorsi acts as a powerful internal rotator, adductor, and extensor of the humerus in the glenohumeral joint. The force vector is significantly more vertical than the lower trapezius portion due to its natural course. However, the latissimus dorsi muscle exhibits such good excursion that the tendon can be inserted directly at the superolateral facet of the greater tuberosity. The transfer restores the force couple of the rotator cuff, allowing centering of the humeral head during deltoid activation.

The procedure was first described as an open procedure in a lateral decubitus position [15]. Nowadays, the majority of the procedure is arthroscopically assisted, with the patient in the beach chair position and the tendon being harvested openly through a skin incision along the posterior axillary fold [18]. The all-arthroscopic technique has also been described, although it has not yet become widely performed [9]. The advantages of the arthroscopically assisted approach are the ability to supply the subscapularis tendon, perform a partial repair of the posterosuperior rotator cuff tendons if possible, sparing of the deltoid muscle, and less scarring.

There are several descriptions of arthroscopically assisted LDT. A particularly illustrative presentation of the single surgical steps can be viewed on VuMedi. link: https://​www.​vumedi.​com/​share/​a5eb4a01-32d6-4aac-a300-3b099041e440/​.

Herein, the authors focus on the pearls and pitfalls of the technique:

A major strength of LDT as a treatment option is that it is the only therapeutic option for the treatment of IPRCT for which long-term results are available [10, 16]. In the study by Gerber et al., with a mean follow-up of 12 years (minimum follow-up of 10 years), it was shown that shoulder function (Constant score % pre- vs. postoperative: 56 vs 80%) and patient satisfaction (subjective shoulder value from preoperative 29 to postoperative 70%) of 44 patients with 46 operated shoulders were sustainably improved over this long period [16]. Mid-term results of the newer arthroscopically assisted technique are also available and do not differ significantly from the open technique [32]. Long-term failure rates range from 10% [10] to 14% [32] and 30% [16].

Over the years, the following risk factors for poor outcome have been identified: fatty atrophy of the teres minor muscle, preoperative pseudoparalysis, excessively high critical shoulder angle, irreparable subscapularis tear, shoulder stiffness, and previous rotator cuff procedures.

The balloon spacer is made of biodegradable copolymer (poly-lactide and ε‑caprolactone) that is completely broken down by the body within 1 year. The injected physiological saline solution starts to deflate already after 3 months. The main biomechanical effect propagated is the avoidance or even revision of superior decentration of the humeral head [28].

Another potential indication is protection of an arthroscopic rotator cuff reconstruction using the balloon spacer on top of the repair to reduce the head in the glenohumeral joint during the rotator cuff reconstruction healing period [31].

The evidence of treatment outcomes is weak, with only mostly level IV and few level III studies showing good short- to mid-term results but no superiority compared to patients treated with APR [22]. In the only level I study by Verma N et al., no significant differences in clinical outcomes and patient satisfaction were seen between the balloon spacer group (n = 93) and the partial repair group (n = 91). This study was presented during the ASES annual meeting in October 2020. The study serves as an FDA approval study. Therefore, the study patients were allowed to be treated with a subacromial balloon spacer in the USA. The approval process is still pending, and it remains unclear whether the balloon spacer will be used in the US.

In a systematic review by Stewart RK et al., a total of 10 level IV and 2 level III studies with a mean follow-up of 23 (range 12–52) months were investigated [30]. It was shown that the Constant score was improved between 19 and 50 points by Stewart et al. [30]. The complication rate of 2.1% was relatively low and included superficial and deep wound infection and balloon spacer dislocation. In 3 of the 12 studies, surgical time was collected, albeit inconsistently, with implantation time alone lasting an average of 10 min (range 2–30 min).

A comparison study demonstrated no statistically significant differences in pain and clinical outcomes 1 year postoperatively between APR alone (n = 16) and APR and balloon spacer (n = 16) patients [22].