Subscapularis repair techniques for reverse total shoulder arthroplasty: A systematic review

Importance: Repair of the subscapularis can be effective in the setting of reverse total shoulder arthroplasty (rTSA). However, there has yet to be a consensus on an optimal repair technique. Objectives: The purpose of this systematic review is to consolidate current high-quality studies comparing out- comes after rTSA with different subscapularis repair techniques. Evidence review: A comprehensive literature review was conducted according to the preferred reporting items for systematic reviews and meta-Analyses using the PubMed, Embase, Scopus and Cochrane databases for original, English-language studies observing outcomes of rTSA after subscapularis repair published between January 1, 2000 and December 31, 2020. Subscapularis management techniques were repair to (1) tendon (tendon – tendon), (2) prosthetic stem, (3) lesser tuberosity (bone tunnels) or (4) a subscapularis-preserving approach (intact). The repair technique was recorded for included studies, and clinical and functional subjective scores were extracted from text, tables and ﬁ gures. Forest plots were created to allow for qualitative comparison of the outcomes of interest between subscapularis repair techniques. Findings: Seven comprehensive studies were identi ﬁ ed, which included 367 patients. The mean age of patient at the time of surgery was 71.1 (cid:1) 2.8 years (range ¼ 47 – 87 years). Overall, 259 patients underwent tendon – tendon repair, 48 patients underwent repair to prosthetic stem, 40 patients underwent repair with bone tunnels and 20 patients ’ subscapularis remained intact. Signi ﬁ cant improvement was seen in most studies for Single Assessment Numeric Evaluation (range, Δ 42.6 – Δ 46.0 out of 3), American Shoulder and Elbow Surgeons (range, Δ 44.2 – Δ 43.6 out of 3) and Visual Analogue Scale pain scores (range Δ 4.2 – Δ 6 out of 5). Active forward elevation (range Δ 40.4 (cid:3) – Δ 57.3 (cid:3) out of 4) and active external rotation (range Δ 2.9 (cid:3) – Δ 16.0 (cid:3) out of 4) signif-icantly improved, but forward elevation varied by nearly 17 (cid:3) ( Δ 16.94 (cid:3) ), while external rotation varied by 13 (cid:3) ( Δ 13.16 (cid:3) ) among repair techniques. Complications were reported in only one study, which used a tendon-tendon technique. Conclusions and relevance: This study summarizes the current evidence regarding subscapularis repair techniques after rTSA including functional and subjective clinical outcome scores. Several different subscapularis repair techniques during rTSA appear to lend to suf ﬁ cient improvement in clinical and subjective outcomes. This in- formation can help guide future studies in this area and highlights the need for high quality studies comparing different subscapularis repair techniques. Level of evidence: III.


Introduction
Reverse total shoulder arthroplasty (rTSA) is an evolving treatment option for cuff tear arthropathy and shoulder pseudoparalysis secondary to massive rotator cuff tear, fracture, tumour, instability or revision arthroplasty [7]. Compared to anatomic total shoulder arthroplasty (TSA), rTSA provides advantages such as medializing the centre of rotation, maximizing deltoid efficiency and providing more stability [20].
Several studies have found subscapularis repair to be safe and effective for the augmentation of rTSA [3][4][5][6]. Biomechanical and cadaveric studies have demonstrated that following rTSA without subscapularis repair, less force is required by the deltoid and posterior rotator cuff muscles in order to abduct the arm [11,13]. Additionally, patients undergoing rTSA without repair of the subscapularis tendon or those with an irreparable subscapularis tendon have shown an increased risk of early dislocation [3,8].
While several studies have investigated the outcomes of rTSA with subscapularis repair, there has yet to be a review of the literature examining different repair techniques of the subscapularis tendon. In TSA, subscapularis tenotomy with tendon-to-tendon repair has traditionally been the standard of care, while other techniques such as lesser tuberosity osteotomy and subscapularis peel were developed to improve healing, the strength of repair and function [17]. The subscapularis peel technique, which repairs tendon to bone, allows the subscapularis insertion to be medialized, resulting in increased external rotation [5] and improved tendon to bone healing [1]. While there have been positive results for the peel technique, there have been some reports of subscapularis deficiency and lower functional outcomes [2].
With this purpose in mind, we performed a systematic review to summarize the existing literature on outcomes after rTSA with different subscapularis repair techniques. Based on the existing literature, we expect no difference in clinical and functional subjective scores according to the subscapularis repair technique following rTSA.

Literature search
A comprehensive literature review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) using the PubMed, Embase, Scopas and Cochrane databases for all articles. This combination search strategy employed the following keywords: ("rTSA OR "reverse total shoulder arthroplasty") AND ("subscapularis repair") AND ("plate" OR "tendon" OR "tenotomy" OR "anchor" OR "repair to stem") AND ("outcomes"). Eligible articles were deemed as those published in English between January 1, 2000 and December 31, 2020, yielding 318 unique publications that were subsequently screened for inclusion.

Study selection
Articles captured in the initial search were excluded from the review if they were published before the year 2000 or in a non-English language. The titles and abstracts of all resulting articles were independently investigated by three authors (A.M.P., T.M. and A.D.L.) to determine relevancy to rTSA outcomes with and without subscapularis repair. Following the initial title and abstract screening, the remaining studies were meticulously reviewed by each author utilizing the following inclusion criteria: clinical trials, RCTs, and cohort studies that examined quantifiable outcomes of rTSA with/without subscapularis repair, peerreviewed clinical studies with evidence level I-IV and that were at least 6-months follow-up after the index rTSA. Articles that met each of these criteria were included in the systematic review. Articles were excluded if they included fracture repair, revision surgery, case reports, systemic reviews, surveys, animal studies or cadaveric studies. Additionally, the references of each included study were reviewed for additional publications that were not captured in the initial database search.
Patient demographics, subscapularis repair status, surgical technique, outcome measures, complication rates and a total number of revisions were obtained from each article.

Study risk of bias assessment
Screening at each level and for each study was performed by two reviewers in a blinded fashion in order to eliminate inter-observer bias. A third reviewer was used to settle any conflicts. Grading for levels of evidence was performed in the same manner to ensure minimal bias. Levels I through IV were included during the review to minimize the risk of bias from any studies of lesser quality.

Data analysis
Due to a limited number of studies eligible for inclusion, a statistical comparison of outcomes was not performed. Forest plots were created to allow for qualitative comparison of the outcomes of interest. Plots were utilized for outcomes when three or more studies reported results (

Search results
The initial search generated 318 articles that were reviewed by the lead investigators (A.M.P., T.M. and A.D.L.) in which 28 abstracts were relevant to our review. Seven publications [4,9,10,15,16,19,23] (four retrospective case-control level III and three case series level IV) were eligible for our inclusion-exclusion criteria and were included in the analysis of our final review. A summary of the flowchart of our literature search according to PRISMA guidelines can be found in Fig. 1.

Demographics
Of the 7 studies included, there were a total of 367 patients. The mean age of patient at the time of surgery was 71.1 AE 2.8 years (range ¼ 47-87 years). Mean patient follow-up was reported for six studies [4,9,10,15,16,19,23] (86%) and was 13.54 AE 11.2 months (range ¼ 6-91 months), while mean percent follow-up was 81%. Indications for surgery included rotator cuff (RC) arthropathy, glenohumeral osteoarthritis, irreparable massive RC tear, migration of the humeral head, inflammatory arthritis with RC tear and failed prior RC repair. Patient demographics can be viewed in Table 1.

Surgical techniques
Two studies performed only a tenotomy of the subscapularis, one of which repaired tendon-to-tendon [15] (tendon-tendon), and the other repaired the subscapularis to the prosthetic stem [9]. One study used a subscapularis-preserving approach in which the subscapularis remained intact during the entirety of the procedure [16] (intact). Three studies directly compared a tenotomy of the subscapularis with subsequent repair using the tendon-tendon approach versus no repair [4,10,23]. One study directly compared a tenotomy of the subscapularis with subsequent attachment to the lesser tuberosity bone tunnels) versus no repair [19] ( Table 2). Overall, 259 patients underwent tendon-tendon repair, 48 patients underwent repair to prosthetic stem, 40 patients underwent bone tunnel repair and 20 patients' subscapularis remained intact (see Table 3).
Among the four retrospective studies that directly compared the subscapularis repair group to a non-repair cohort, three [4,19,23] (tendon-tendon repair or bone tunnels) reported no significant difference between pre-and post-operative clinical or patient-reported outcome scores. Interestingly, Franceschetti et al. [10] found that patients who underwent subscapularis repair using a tendon-tendon approach had significantly better post-operative internal rotation (p ¼ 0.02), while patients who did not undergo subscapularis repair had significantly better abduction (p ¼ 0.04).
Complications were reported in only one study [17], which used a tendon-tendon technique. A total of 6% (7 out of 116 patients) experienced complications that included postoperative infection, base plate failure, humeral component loosening and dislocation.
The findings of the current study suggest that subjective postoperative scores, including ASES and SANE, are comparable among intact, prosthetic stem, tendon-tendon and bone tunnel subscapularis repair techniques performed during rTSA. These observations are consistent with a case control study by Hartline et al. [14] that examined differences in postoperative outcomes between patients that underwent a tendon-tendon, prosthetic stem and no-repair subscapularis management techniques during primary rTSA. The study found no significant difference in subjective outcome scores (ASES), complications or reoperations (p > 0.05). Interestingly, Hartline et al. [14] observed that subscapularis repair-regardless of technique-resulted in significantly greater postoperative pain improvement (VAS) than the non-repair group. These findings are consistent with the current review that observed homogenously significant improvement in VAS among all subscapularis repair techniques. While the current review does not directly compare repair techniques to a non-repair cohort, Godin et al. [12] suggest that the sufficiency of the subscapularis tendon is an important factor resulting in better subjective outcome scores. These results suggest that patients with healthier subscapularis preoperatively have less shoulder pathology potentially resulting in independently better postoperative subjective outcomes scores. Thus, the competency of the subscapularis preoperatively is a notable confounding variable that may attribute to improved patient-reported outcomes.
Variable outcomes in postoperative ROM depending on subscapularis management technique warrants further investigation. The current review found that improvement in postoperative FE varied by nearly 17 , while ER varied by 13 . A tendon-tendon repair technique was used in both studies that reported the greatest increase in postoperative FE [14] and ER [13]. Franceschetti et al. [10] found patients undergoing subscapularis repair using a tendon-tendon approach and had significantly better post-operative internal rotation versus patients who did not Fig. 2. Forest plot of mean pre-to postoperative reported SANE scores with 95% Confidence Intervals. Significant improvement in SANE was observed in three cases (range, Δ 42.6-Δ 46.0), of which included tendon-tendon repair [14] (p < 0.001), repair to prosthetic stem [12] (p < 0.001), and an intact approach [15] (p < 0.001). Fig. 3. Forest plot of mean pre-to postoperative reported ASES scores with 95% Confidence Intervals. Significant improvement was observed in ASES scores in three cases (range Δ44.2-Δ43.6), including repair to stem [12] (p < 0.001), tendon-tendon repair [14] (p < 0.001) and bone tunnelling [16] (p ¼ 0.025).  undergo subscapularis repair had significantly better abduction. These results may suggest an altered force vector by repairing the subscapularis non-anatomically.
While studies have found that subscapularis repair in rTSA can affect postoperative ROM due to antagonistic forces of the subscapularis exerted on the posterior RC [13], variability is observed among types of subscapularis repair techniques. This difference could be explained by a plethora of potential confounding factors including patient population, age and surgeon experience. The results of the current review are consistent with the findings of Hartline et al. [14], which observed a significantly greater improvement in ER strength among patients who underwent a tendon-tendon subscapularis repair than a bone-tendon repair. However, Khazzam et al. [15], who also used a tendon-tendon repair technique, reported only a 2.8 improvement in ER --11.5 less than the second lowest pre-to postoperative improvement in ER. While the differences in ER and FE outcomes among repair techniques remain uncertain, we speculate this observed difference could relate to the length of the repaired tendon. In a cadaveric study by Van den Berghe et al. 22 that examines the effect of subscapularis repair techniques on ROM in TSA, it was found that tendon-tendon repairs shortened the subscapularis by 5.3 mm while a tendon-bone repair lengthened the tendon by an average of 2.6 mm [22]. One centimetre of subscapularis shortening may cause a 20 loss of ER [18], thus shortening the subscapularis tendon may have considerable effects on ROM depending on the repair technique.
At present, the paucity of literature comparing lesser tuberosity osteotomy (LTO) to tenotomy or peel technique in rTSA leaves this review incomplete. Comparing these techniques in patients undergoing TSA, a meta-analysis by Del Core et al. [6] found that LTO had superior healing and subscapularis specific strength testing. However, there was no difference in ROM between peel, tenotomy and LTO with all groups showing improvement in patient-reported outcomes. These findings regarding improvements in patient-reported outcomes using the peel and tenotomy techniques are consistent with their utilization in rTSA in our study. Nevertheless, there is little literature on LTO utilization in rTSA. While there are different techniques, several studies have shown that LTO is biomechanically superior and has favourable healing rates compared with the peel or tenotomy techniques [21]. It could be speculated that the patient populations undergoing rTSA are older, with weaker bones for which surgeons are less likely to utilize LTO. Further study is required if there is any potential utility for LTO in rTSA.
There are several notable limitations of the current study. Systematic reviews are inherently limited by the relevant data reported in the literature. To date, there has been no level one or level two studies regarding rTSA and subscapularis repair. As such, the current study carries the same biases as those of non-randomized and retrospective analyses. Furthermore, the tendon-to-tendon repair technique was overrepresented as it was performed in five of the seven studies. Within subscapularis repair, there are many other variables that may affect outcomes including preoperative ROM, implant characteristics and preoperative subscapularis competency which make these results more difficult to interpret independently. Surgeon experience, patient age andpatient populations are also impossible to control for in comparisons made in a review article. Additionally, the heterogeneity of shoulder outcome scores utilized made the analysis more challenging. There were limited studies to compare these techniques, preventing a deeper statistical analysis.

Conclusion
With the sparsity of high-quality literature, further study is required to accurately compare subscapularis repair techniques after rTSA. While tendon-tendon repair was overrepresented in this study, all repair techniques examined in this study appear to achieve successful functional and subjective clinical outcome scores. Although a comparison was not made to a non-repair group, this review displays that subscapularis repair techniques have little impact on both subjective and objective clinical outcomes. This information can help guide future studies in this area.