Superior capsule reconstruction, partial cuff repair, graft interposition, arthroscopic debridement or balloon spacers for large and massive irreparable rotator cuff tears: a systematic review and meta-analysis

Abstract

Background

Multiple non-arthroplasty surgical techniques are described for the management of large and massive irreparable rotator cuff tears. There is currently no consensus on the best management strategy. Our aim was to compare clinical outcomes following arthroscopic debridement, arthroscopic partial cuff repair, superior capsule reconstruction, balloon spacers or graft interposition for the management of large and massive irreparable rotator cuff tears.

Methods

A comprehensive search was performed of the following databases: Medline, Embase, CINAHL and Cochrane Database of Systematic Reviews. Data were extracted from relevant studies published since January 2000 according to the pre-specified inclusion criteria. The primary outcome was the post-operative improvement in shoulder scores. Meta-analysis of the primary outcome was performed. Secondary outcomes included retear rates and complications.

Results

Eighty-two studies were included reporting the outcomes of 2790 shoulders. Fifty-one studies were included in the meta-analysis of the primary outcome. The definition of an irreparable tear varied. All procedures resulted in improved shoulder scores at early follow-up. Shoulder scores declined after 2 years following balloon spacers, arthroscopic debridement and partial cuff repair. High retear rates were seen with partial cuff repairs (45%), graft interposition (21%) and superior capsule reconstruction (21%).

Conclusions

Large initial improvements in shoulder scores were demonstrated for all techniques despite high retear rates for reconstructive procedures. Shoulder scores may decline at mid- to long-term follow-up.

Background

Rotator cuff tears are a common cause of shoulder pain and can lead to significant functional impairment [1]. A primary consideration for large and massive tears is whether the tendon can be repaired. A tendon may be considered irreparable when it cannot be mobilised to the anatomical footprint despite mobilisation techniques such as interval slide and marginal convergence [2]. Poor tendon architecture with a high degree of fatty infiltration will also reduce the likelihood of a successful repair.

Several surgical techniques have evolved for the management of large and massive irreparable tears in adult patients without significant arthritic disease. Partial cuff repair (PCR) is an option where complete repairs are not possible [3]. Reconstruction of infraspinatus and teres minor is prioritised to re-establish the transverse force couple. Further techniques have evolved including subacromial balloon spacers, superior capsule reconstruction (SCR) and the use of autografts, allografts, xenografts or synthetic grafts to bridge residual cuff defects [2].

SCR aims to prevent superior migration of the humeral head and restore normal shoulder biomechanics; it involves passing a graft from the superior tubercle of the glenoid to the greater tuberosity [4, 5]. Similarly, balloon spacers are designed to prevent superior migration by deploying biodegradable saline-filled spacers underneath the acromion. Several large medical device companies have invested in the technology, and there has been a rapid expansion in data from case series [6].

Several tendon transfer techniques have been described to restore shoulder function according to tear configuration [7]. Posterosuperior tears may be managed with latissimus dorsi or trapezius tendon transfers [7, 8]. These have been shown to be effective in select groups; they are most commonly used in younger, high-demand patients and require specialised rehabilitation programmes [7, 8]. As a consequence, the patient populations managed with tendon transfers will likely be different compared to the other procedures in this review and tendon transfers will not be considered further. Reverse total shoulder replacement can lead to significant improvement in function and pain in patients with concurrent arthritis; however, concern about the complication and revision rates has limited its application in those without advanced arthritis, particularly in younger patients [9].

Despite the significant impact on patients’ quality of life, socio-economic considerations and protracted rehabilitation periods, a consensus on the optimal strategy has not been reached. Studies have demonstrated positive patient-reported outcome measures (PROM), functional scores and reduced pain for arthroscopic debridement, arthroscopic PCR, SCR, subacromial balloon spacers and graft interposition.

Aim

The aim of this review is to answer the following research question: In adult patients with a large or massive irreparable tear of the rotator cuff does arthroscopic debridement, arthroscopic PCR, SCR, subacromial balloon spacer or graft interposition lead to the greatest improvement in shoulder scores and the lowest risk of complications? Randomised trials and observational studies were considered. Our hypothesis is that SCR, balloon spacer and graft interposition do not result in superior shoulder scores compared to arthroscopic PCR.

Methods

The review protocol was registered on the International Prospective Register of Systematic Reviews (PROSPERO) database. The review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement [10].

Search strategy

The following databases were used: Medline, EMBASE, CINAHL and the Cochrane database of systematic reviews, from inception until May 2021. Trial registries were searched to highlight ongoing work. A scoping literature review was performed prior to protocol submission to optimise the search terms and inclusion criteria. The search terms are related to the population (rotator cuff tear, supraspinatus, infraspinatus, subscapularis, teres minor tears, AND massive, large, irreparable) and the interventions (superior capsule reconstruction, partial cuff repair, balloon spacer, graft bridging, debridement). Reference lists were scrutinised for further work. The full search terms are available in the Appendix.

Eligibility criteria

The review included adult patients with large or massive rotator cuff tears that were considered irreparable. The reparability of the cuff was determined by the authors of each study based on pre-operative imaging and intra-operative assessment. Across all studies, this was defined as a tear where the tendon could not be brought back to the tendon insertion without excessive tension on the tendon.

A minimum tear size was required; large or massive tears were included. Large and massive tears were defined as 3 cm or greater in any dimension or involving two or more tendons. The studies must report on pre-operative and post-operative shoulder scores for any of the included procedures: arthroscopic debridement, arthroscopic PCR, SCR, subacromial balloon spacers and graft interposition of incomplete repairs. Graft interposition was defined as cases where a graft was used to bridge a remaining tendon defect. It did not include cases where grafts were used in addition to a complete repair. In SCR, the graft is anchored into the superior glenoid and the greater tuberosity. Articles published no earlier than 2000 were considered. All types of graft material for SCR and graft interposition were accepted. Randomised trials and observational studies with multiple or single intervention groups of ten or more patients were included. Where trials compared one or more techniques that matched our inclusion criteria against alternative procedures, data from the relevant treatment arm were extracted.

Articles were excluded if they reported on tears associated with additional trauma including fractures or dislocations and patients with documented cuff tear arthropathy, SLAP tears, or advanced glenohumeral osteoarthritis. Small and medium tears (< 3 cm) were excluded alongside procedures that involved additional bone marrow infiltration, complete repair or augmentation of a complete repair. Animal or in vitro work, editorials, conference abstracts, case reports, case series of less than 10 patients and letters were also excluded. We included studies where patients may have undergone a previous failed rotator cuff repair. Finally, we required a minimum of 12 months of follow-up. We excluded studies with less than 12 months follow-up, where earlier outcome data could not be separated.

Outcomes

The primary outcomes were PROMs and functional scores including, but not limited to the Constant-Murley score, American Shoulder and Elbow Score (ASES), Oxford Shoulder Score (OSS), University of California in Los Angeles (UCLA) Shoulder Score, the Disabilities of Arm, Shoulder and Hand score (DASH), Subjective Shoulder Value (SSV), Simple Shoulder Test (SST) or the Japanese Orthopaedic Association (JOA) shoulder score. Secondary outcomes include retear rates and complications.

Study selection

Two authors (AD and PS) independently assessed trials for eligibility. Search results were organised using Covidence® systematic review management software. Duplicates were removed, and articles were selected according to the inclusion criteria. Conflicts were discussed between reviewers; a third reviewer (SS) was available as required to arbitrate.

Data extraction

A standardised form guided data extraction in four areas: study characteristics, pre-operative information, details of the interventions and post-operative outcomes (Table 1). Data were extracted by AD and PS, and the results were checked independently by both authors. In order to be included in the quantitative synthesis, studies must have reported PROMs as mean (± standard deviation). To avoid error, data were not extracted from graphs. To pool PROMs across studies, outcome duration was grouped into 1 year, 2 years, 3 years and 5 or more years. The durations of follow-up were frequently presented as a mean and range. To standardise the analysis, all patients within a group must have achieved the minimum follow-up duration in order to be included.

Table 1 Data extraction. OSS—Oxford Shoulder Score, JOA—Japanese Orthopaedic Association score

Quality and bias assessment

A risk-of-bias assessment was performed in all studies by two authors (AD and PS) with consensus reached by discussion where necessary (Appendix 4). The Methodological Index for Non-randomized Studies (MINORS) instrument was chosen following publication of the protocol because all except 1 study were observational and the majority were case series [11]. This instrument includes an eight-item assessment of non-comparative studies. The instrument can be extended to 12 items in order to include comparative studies.

Data analysis

Pre-operative and post-operative PROMs and functional scores were extracted. The difference between the pre-operative and post-operative scores was calculated at 1, 2, 3 and 5 years post-procedure. Direct comparison between procedures was not made given the quality of the available literature, outcomes for each procedure were pooled independently. To account for the different scores, the standardised mean difference (SMD) was calculated and weighted according to sample size. A random effects model was selected due to the variability in study design.

To facilitate interpretation of the SMD, this was re-expressed in the units of the Constant-Murley score. This was the score most commonly used across included studies. Transformation was performed according to the Cochrane handbook [12]. We used the pooled standard deviation of all studies included in the quantitative synthesis that reported the Constant-Murley score. Analysis was performed using Review Manager 5.3. Mean differences were calculated for the continuous data. Heterogeneity was assessed using the I² Chi² and Tau² statistics.

Results

The search strategy revealed 4145 results; 5 further potential studies were identified through screening of references. After removal of duplicates, 2377 articles were available for screening (Fig. 1). Screening of titles and abstracts excluded 2124 articles. One hundred and seventy-one articles were excluded on full-text review leaving 82 studies for qualitative synthesis. Fifty-one studies were included in the quantitative synthesis of the primary outcome.

Fig. 1

PRISMA study flow diagram

Study characteristics

Of the 82 eligible studies, 60 were case series, 21 were cohort studies and 1 was a randomised controlled trial (Tables 2, 3, 4, 5 and 6). The surgical management of 2790 shoulders was reported. The majority of studies were performed in a single centre; 5 involved patients from two or more institutions. Sufficient data for quantitative synthesis were provided in 51 studies. This includes 37 case series and 14 comparative studies (13 cohort studies and 1 randomised trial), where one or more of the treatment groups were included. In four cohort studies and in the randomised trial, one or more groups did not meet the inclusion criteria and data were extracted for the relevant management arm.

Table 2 Partial cuff repair—included studies

Table 3 Graft interposition—included studies. OSS—Oxford Shoulder Score, JOA—Japanese Orthopaedic Association score

Table 4 Superior capsule reconstruction—included studies

Table 5 Arthroscopic debridement—included studies

Table 6 Balloon spacer—included studies

Twenty-six of the included studies reported shoulder scores following partial cuff repair, and 21 studies reported scores following SCR. Eighteen articles recorded results for graft interposition and outcomes for arthroscopic debridement were reported in 15 studies. However, in one cohort study comparing PCR and arthroscopic debridement, the series in the debridement arm was not included due to there being fewer than 10 patients in this treatment group. Ten studies reported shoulder scores for balloon spacers. The criteria used to define large and massive tears were documented in 58 studies. In the remaining articles, the authors described the tears as “large” or “massive” without further details. Magnetic resonance imaging (MRI) was the most favoured imaging modality; 74 papers reported the use of pre-operative MRI or magnetic resonance arthrogram in diagnosis and classification. The remaining either used ultrasound imaging or an intra-operative assessment. The age range of participants across all studies was 33–90. The Hamada classification of rotator cuff arthropathy was the most frequently used. Several studies were excluded due to lack of reporting of tear size, the majority of these were for balloon spacers.

The terminology used to describe graft interposition varied between studies and included “bridging”, “interposition” and “augmentation”. Careful review of the description of the technique was necessary to ensure the grafts were used to fill the remaining defect after the best possible repair. The most popular interposition graft was human dermal allograft (7 studies). In contrast, fascia lata autograft was preferred in superior capsule reconstruction (9 studies). Post-operative management was well reported in the majority of studies; 70 described a progressive rehabilitation programme.

The pre-operative management of patients was described in 28 studies, the majority of these documented a minimum of 6 months of failed conservative therapy. All surgical procedures were performed in the beach chair or lateral decubitus position. In the case of bridging repairs and superior capsule reconstruction, allografts, autografts, xenografts and synthetic grafts were used (Tables 2, 3, 4, 5 and 6).

In all cases of debridement and balloon spacer insertion, pendular and passive range of motion exercises were commenced within 4 weeks, with the majority initiating range of motion exercises from the first day post-op. Sling use varied, it was used for a maximum of 3–4 weeks following these two procedures. In contrast, following arthroscopic PCR, SCR or graft interposition, patients were immobilised in a sling for up to 6 weeks followed by passive and active range of motion exercises. Full strengthening work may not begin for 6 months.

Partial cuff repair

See Table 2.

Graft interposition

See Table 3.

Superior capsule reconstruction

See Table 4.

Arthroscopic debridement

See Table 5.

Balloon spacer

See Table 6.

Function scores and patient-reported outcome measures

In total, 49 studies reported pre-operative and post-operative Constant-Murley scores. The ASES was the next most reported score (32 studies) followed by the UCLA Shoulder Score (15 studies).

In all studies, there was an improvement in post-operative PROMs data compared to pre-operative scores. Synthesis of the difference in pre-operative and post-operative PROMs scores was possible across 51 studies. The pooled standardised mean difference for each implant is given in Table 7. Data were available at 1 year and 2 years for all 5 procedures. Three-year data were available for balloon spacers only and 5-year data for arthroscopic partial cuff repair and arthroscopic debridement. The number of studies contributing to the pooled SMD is shown in brackets. The greatest number was available for partial cuff repair and superior capsule reconstruction. A study by Vad et al. demonstrated a very high risk of bias and an SMD far greater than similar studies [81]. Results were pooled with and without inclusion of this study and omitted in the results given in Table 7. The full series of forest plots are available in the Appendix.

Table 7 SMDs in the difference in pre-operative and post-operative PROMs scores. Number of articles pooled in brackets

The SMD was re-expressed as a Constant-Murley score (Fig. 2); the mean standard deviation for the pre-operative Constant-Murley score was 10.9.

Fig. 2

Standardised mean difference for each procedure, transformed to the Constant-Murley score

Secondary outcomes

The quality of reporting of revision procedures, retear rates and complications was mixed (results summarised in Tables 8 and 9 and the Appendix). Post-operative imaging was performed to a different extent across the studies. This may include a routine MRI or ultrasound (US) in all cases, a proportion of cases, post-operative imaging only when the patient was symptomatic or not at all.

Table 8 Revision procedures, retears and infections following PCR, graft interposition and SCR

Table 9 Revision procedures and infections following arthroscopic debridement and balloon spacer

Six studies reporting outcomes of PCR performed an MRI or US in all shoulders. Of 170 shoulders, 77 showed evidence of a retear (45%). Following graft interposition, nine studies reported post-operative imaging in all cases and further 6 studies scanned a proportion of patients. In total, 302 shoulders were scanned and 64 tears of the graft or tendon were identified (21%). After SCR, 12 studies reported imaging for all shoulders, and a further 5 scanned a majority of shoulders for a total of 469; 100 retears were found (21%).

Revision procedures for failure or symptom recurrence were poorly documented. Some studies may have only reported on patients that achieved the documented minimum follow-up period, resulting in an underestimation of the true revision rate. The greatest proportion of revision procedures was performed after balloon spacers (7.2%). The fewest revisions were reported following graft interposition (1.9%).

Reporting of additional complications was sparse; these are documented in the Appendix. Polacek et al. reported three cases of acute porcine dermal xenograft rejection which required urgent revision [71]. Two cases of donor site morbidity were reported in two studies where fascia lata autograft was used for a SCR, and one required revision surgery at the donor site [72, 73]. Twenty studies reported VAS scores. In all studies, there was improvement in the post-operative pain scores. The limited data on pain scores preclude meaningful comparison between techniques.

Risk of bias

A risk-of-bias assessment of the included studies was performed according to the MINORS instrument and is summarised in Table 10 [11]. The cohort studies were generally of low quality. In order to maximise the opportunity for comparison between procedures, we extracted each relevant arm of the comparison separately and treated them as individual series. Consequently, the abbreviated 7-point MINORS criteria for non-comparative studies were calculated for all included studies. The results are summarised in Table 10. The full tables, including the results of the extended MINORS instrument for comparative studies, are available in the Appendix. For each procedure the mean risk of bias score across all seven domains was calculated.  

Table 10 Risk-of-bias scores

The greatest risk of bias was seen in studies on balloon spacers. These frequently described significant numbers lost to follow-up, and similar to all studies in this review, outcome scores were only presented for patients that completed follow-up.

Discussion

This systematic review and meta-analysis compared common surgical techniques for the management of large and massive irreparable rotator cuff tears. Management of this pathology in the absence of glenohumeral arthritis is challenging. This work demonstrates that arthroscopic debridement, arthroscopic PCR, SCR, subacromial balloon spacers and graft interposition all lead to improved shoulder scores at early- to mid-term follow-up. It is encouraging for surgeons and patients that there are several potential surgical strategies which lead to improvements in shoulder function. However, we show that clinical improvement may decline over time for arthroscopic debridement, PCR and balloon spacers, and this is most apparent for balloon spacers. Across the included case series, the improvement in shoulder scores for PCR was comparable with graft interposition and SCR. Despite the additional cost of graft augmentation and the additional expertise required for SCR, they may not provide further clinical benefit in the early to mid-term period compared to PCR.

The number of studies and participants available for synthesis at each time point varied. Three hundred and ninety participants from 10 studies were included in the synthesis of SCR 12-month outcomes compared to 50 in two studies for arthroscopic debridement at 5 years. We required a minimum post-operative duration of all patients within a series to be included in the quantitative synthesis. The greatest volume of evidence was available for graft interposition, partial cuff repair and superior capsule reconstruction. The pooled PROMs scores for each of these procedures were similar up to 2 years.

The results of arthroscopic debridement and subacromial spacers should be interpreted with caution. The trajectory of improvement in the post-operative scores for arthroscopic debridement shown in Fig. 2 demonstrates a decline in scores at a minimum of 24 months followed by an improvement at 5 years. In the 5-year group, there were two studies with a moderate to high risk of bias and a total of only 52 patients. The studies reporting on subacromial spacers had a higher failure rate and loss to follow-up. A randomised trial comparing balloon spacer to arthroscopic debridement demonstrated superior shoulder scores following arthroscopic debridement at 12 months [94]. A further trial comparing subacromial spacers to PCR demonstrated non-inferiority of subacromial spacers at 12 months [95]. However, despite improvement in the first 12–24 months after PCR and subacromial spacers, our work shows a decline in PROMs after this point, which is most prominent following subacromial spacers.

Reverse shoulder arthroplasty (RSA) was not included in this review. The use of RSA is rapidly expanding. However, they are performed infrequently in the absence of glenohumeral osteoarthritis, particularly in younger patients. RSA can be effective at mid- to long-term follow-up in patients with massive tears, with or without arthritic disease [96,97,98]. Ernstbrunner et al. reported that improvement in shoulder scores may be sustained over 10 years in the presence of massive tears; however, survivorship was 85% at 5 years and 70% at 15 years [97]. Other series demonstrate a sustained improvement in shoulder scores and a lower incidence of complications in the absence of glenohumeral osteoarthritis; however, these remain a concern [98, 99]. The National Joint Registry of England, Wales and Northern Ireland report a risk of revision of 3.85% at 8 years in all patients who receive a RSA, with a median patient age of 76 years [100]. Current evidence would suggest caution and shared decision-making when considering RSA in younger patients without osteoarthritis.

Latissimus dorsi and trapezius tendon transfers show encouraging results in specific patient groups. Recent reviews report large increases in shoulder scores across several series [16, 101, 102]; improved shoulder function may persist over 10 years[103]. A comparative study by Cavalier demonstrated an inferior Constant score at 12 months with latissimus dorsi tendon transfer compared to PCR and RSA, and a randomised trial of 42 patients reported inferior functional scores for LDTT compared to SCR [16, 68].

This review included a broad and comprehensive search strategy and included a large number of patients not included in earlier work. Previous systematic reviews have attempted to compare procedures for irreparable or massive rotator cuff tears [104, 105]. We have captured multiple studies not included in previous reviews. This is the first study to include a quantitative synthesis across a large number of studies accounting for different outcome scores.

This study has limitations. The quality of included studies varied, and reliable conclusions cannot be made about the comparative effectiveness of each procedure. Although further well-constructed and appropriately powered randomised trails comparing these techniques are warranted, they are practically challenging given the varied nature of tendon reparability and tendon quality in patients with large and massive tears. Despite an accepted definition of an irreparable tendon as one that cannot be brought back to the tendon footprint without undue tension, intra-operative assessments may differ and the techniques used to bring the tendon back to the footprint, such as interval slide and marginal convergence, are not routinely described—the decision of whether a tendon can be brought to the greater tuberosity may influence outcomes in borderline cases. We did not classify studies according to pre-operative management. This was poorly reported; dedicated anterior deltoid strengthening work can improve shoulder function in patients with irreparable cuff tears [106]. We did investigate the change in shoulder scores, rather than post-operative score alone which would account for any large differences in pre-operative function.

Conclusion

In adult patients with large and massive irreparable rotator cuff tears arthroscopic debridement, arthroscopic PCR, SCR, subacromial balloon spacers and graft interposition lead to improved early- and mid-term patient-reported outcomes and functional scores. Retear rates are 21% for SCR, 21% for graft interposition, and 45% for arthroscopic PCR. Mid- to long-term follow-up is necessary to further investigate whether there is a significant decline in shoulder function with time.

Appendix

Complete results tables: shoulder scores

Partial cuff repair

See Table 11.

Table 11 Partial cuff repair—included studies

Graft interposition

See Table 12.

Table 12 Graft interposition—included studies

Superior capsule reconstruction

See Table 13.

Table 13 Superior capsule reconstruction—included studies

Arthroscopic debridement

See Table 14.

Table 14 Arthroscopic debridement—included studies

Balloon spacer

See Table 15.

Table 15 Balloon spacer—included studies

Forest plots

PCR 12 months

PCR 24 months

PCR 5 years

Graft interposition 12 months

Graft interposition 24 months

SCR 12 months

SCR 24 months

Arthroscopic debridement 12 months

Arthroscopic debridement 24 months

Arthroscopic debridement 5 years

Balloon 12 months

Balloon 24 months

Balloon 3 years

Complications

Partial cuff repair

Author and Date	Number of shoulders in PCR group for which f/u data available	No. of revision procedures for failure procedure	Post-op imaging performed—no. of patients	Number retears on imaging	Infections	Additional complications
Baverel [13]	26	0	0	n/a	0	0
Berth [14]	21	1	21	11	0	0
Besnard [15]	20	4	0	n/a	n/r	n/r
Cavalier [16]	61	0	0	n/a	1	1 anchor migration
Chen [17]	37	0	Unknown	Unknown	n/r	n/r
Cuff [18]	28	3	0	n/a	0	0
DiBenedetto [19]	31	0	0	n/a	n/r	n/r
Farazdaghi [20]	14	4	0	n/a	0	0
Franceschi [21]	34	n/r	0	n/a	n/r	n/r
Galasso [22]	95	8	0	n/a	n/r	n/r
Greiner [23]	20	0	0	n/a	0	0
Heuberer [24]	22	1	Unknown	Unknown	2 (washouts)	0
Holtby [25]	73	n/r	0	n/a	n/r	n/r
Jeong [26]	33	0	33	28	n/r	n/r
Kim [27]	27	0	0	n/r	n/r	n/r
Kim [28]	19	0	0	n/r	n/r	n/r
Lee [29]	42	0	42	10	0	0
Malahias [30]	16	0	0	n/a	1	0
Mori [31]	24	0	24	10	0	0
Moser [32]	11	0	0	n/a	0	0
Pandey [33]	13	0	13	4	0	0
Park [34]	37	0	37	14	0	0
Paribelli [35]	20	0	0	n/a	n/r	n/r
Porcellini [36]	67	0	0	n/a	0	2 stiffness (capsular release)
Shon [37]	31	0	0	n/a	n/r	n/r
Wellmann [38]	38	0	0	n/a	n/r	n/r

Graft interposition

Author and Date	Number of shoulders in graft group for which f/u data available	No. of revision procedures	Post-op imaging performed—no. of patients	Number of retears on imaging	Infections	Additional complications
Audernaert [39]	39	0	39	4	0	0
Badhe [40]	10	0	10	2	0	0
Bond [41]	16	0	16	3	0	0
Dukan [42]	23	5	18	9	0	2 intraop partial glenoid fractures
Gupta [43]	24	0	19	5	0	0
Gupta [44]	27	0	22	6	0	0
Kim [45]	24	0	24	5	0	0
Kokkalis [46]	21	0	0	n/a	0	0
Modi [47]	61	1	14	2	1 (washout)	0
Mori [31]	24	0	24	5	0	0
Nada [48]	21	1	21	0	0	0
Pandey [33]	13	0	13	1	0	0
Petrie [49]	31	2	0	n/a	0	0
Rhee [50]	31	1	14	5	0	0
Rhee [51]	24	0	24	13	n/r	n/r
Sano [52]	14	0	14	1	0	0
Dimitrios [53]	68	0	30	3	0	0
Wong [54]	45	0	0	n/a	1 (washout)	0

Superior capsule reconstruction

Author and Date	Number of shoulders in SCR group for which f/u data available	No. of revision procedures	Post-op imaging performed—no. of patients	Number of retears on imaging	Infections	Additional complications
Alarcon [55]	31	0	29	4	0	1 haematoma
Barth [56]	24	0	24	2	n/r	n/r
Burkhart [57]	41	2 (post-falls)	26	4	0	1 stroke 1 tear biceps tenodesis
Denard [58]	59	0	20	11	1	1 biceps pain
Ferrando [59]	56	4	56	14	n/r	n/r
Greiner [23]	20	0	20	1	n/r	n/r
Kim [60]	45	0	45	6	0	1 skin dehiscence, 2 Popeye sign
Kocaoglu [61]	26	0	12	2	n/r	n/r
LaBelle [62]	35	4	21	13	n/r	n/r
Lacheta [63]	22	1	21	5	n/r	n/r
Lee and Min [64]	36	0	36	13	n/r	n/r
Lim [65]	31	0	31	9	0	1 PE
Mihata [5]	24	0	24	4	0	0
Okamura [66]	35	0	35	2	0	1 synovitis
Ohta [67]	35	1	35	7	0	1 severe synovitis 1 anchor dislodgement
Ozturk [68]	20	0	20	1	0	1 significant stiffness
Pashuck [69]	14	1	14	2	0	0
Pennington [70]	88	1	Only with Sx	-	n/r	n/r
Polacek [71]	20	5	Only with Sx	-	0	3 acute graft rejection
Polacek [72]	24	3 incl. 1 fascia lata site revision	Only with Sx	-	0	1 fascia lata muscle prolapse. 1 haematoma
Takayama [73]	46	0	Only with Sx	-		1 swelling at donor site 1 shoulder swelling

Debridement

Author and Date	Number of shoulders in debridement group for which f/u data available	No. of revision procedures	Infections	Additional complications
Berth [14]	21	1	0	0
Blanke [74]	59	0	0	0
Boileau [75]	72	3	1 (washout)	2 persistent pain. 1 shoulder stiffness, 2 GHJ OA, 3 pseudo paralysis
Cavalier [16]	26	0	0	0
Franceschi [21]	34	n/r	n/r	n/r
Heuberer [24]	23	1	0	0
Klinger [76]	41	0	0	3 pseudoparalysis 2 pain 1 stiffness. 2 GHJ OA. 1 humeral head migration
Lee [77]	32	4	n/r	n/r
Mirzaee [78]	12	0	0	2 GHJ OA
Park [79]	16	1	0	0
Scheibel [80]	23	1	0	1 haematoma
Vad [81]	32	0	n/r	n/r
Veado [82]	15	0	n/r	n/r
Verhelst [83]	32	2	n/r	n/r

Balloon spacer

Author and Date	Number of shoulders in balloon spacer group for which f/u data available	No. of revision procedures	Infections	Additional complications
Basat [84]	12	0	n/r	n/r
Bakti [85]	26	1	0	0
Deranlot [86]	39	1	0	0
Familiari [87]	51	6	0	0
Gervasi [88]	15	1	0	0
Gervasi [89]	40	3	0	0
Iban [90]	16	5	0	0
Lorente 2017 [91]	15	3	n/r	n/r
Maman [92]	42	0	n/r	n/r
Senekovic [93]	20	0	0	0

Risk of bias

All arms of relevant non-comparative and comparative studies were treated as individual series. The risk-of-bias results for the abbreviated 7-point MINORS instrument for non-comparative studies were applied to all articles and presented here. The results of the extended MINORS instrument for comparative studies are presented in the final section of this document.

PCR

Author and Date	Clearly state aim	Inclusion of consecutive patients	Prospective data collection	Unbiased endpoint assessment	Follow-up appropriate	Loss to f/u < 5%	Prospective calculation of sample size
Baverel [13]	2	2	2	2	2	2	0
Berth [14]	2	1	2	2	2	2	0
Besnard [15]	2	2	1	2	2	0	0
Cavalier [16]	2	2	2	2	2	2	0
Chen [17]	2	1	1	2	2	1	0
Cuff [18]	2	1	1	2	2	1	0
DiBenedetto [19]	2	2	1	2	2	2	0
Farazdaghi [20]	2	1	1	2	2	1	2
Franceschi [21]	2	1	2	2	2	2	2
Galasso [22]	2	2	2	2	2	0	0
Greiner [23]	2	2	2	2	2	2	2
Heuberer [24]	2	1	2	2	2	0	0
Holtby [25]	2	2	2	2	2	2	0
Jeong [26]	2	1	1	2	2	1	0
Kim [27]	2	2	1	2	2	2	0
Kim [28]	2	2	1	2	2	2	0
Lee [29]	2	1	1	2	2	0	0
Malahias [30]	2	1	2	2	2	2	0
Mori [31]	2	2	2	2	2	2	0
Moser [32]	2	1	2	2	2	2	0
Pandey [33]	2	2	2	2	2	2	0
Park [34]	2	1	2	2	2	2	0
Paribelli [35]	2	1	2	2	2	2	0
Porcellini [36]	2	2	2	2	2	1	0
Shon [37]	2	1	1	2	2	2	0
Wellmann [38]	2	1	1	2	2	2	0

Graft interposition

Author and Date	Clearly state aim	Inclusion of consecutive patients	Prospective data collection	Unbiased endpoint assessment	Follow-up appropriate	Loss to f/u < 5%	Prospective calculation of sample size
Audernaert [39]	2	2	2	2	2	2	0
Badhe [40]	2	2	2	2	2	2	0
Bond [41]	2	2	2	2	2	2	0
Dukan [42]	2	2	2	2	2	1	0
Gupta [43]	1	2	2	2	2	2	0
Gupta [44]	2	2	2	2	2	2	0
Kim [45]	2	2	2	2	2	2	0
Kokkalis [46]	1	2	1	2	2	2	0
Modi [47]	2	2	2	2	2	2	0
Mori [31]	2	2	2	2	2	2	0
Nada [48]	1	0	2	2	2	2	0
Pandey [33]	2	2	2	2	2	2	0
Petrie [49]	2	2	2	2	2	1	0
Rhee [50]	2	1	2	2	2	1	0
Rhee [51]	1	2	2	2	2	2	0
Sano [52]	2	1	0	2	2	2	0
Varvitsiotis [53]	2	2	2	2	2	2	0
Wong [54]	1	2	2	2	2	2	0

Superior capsule reconstruction

Author and Date	Clearly state aim	Inclusion of consecutive patients	Prospective data collection	Unbiased endpoint assessment	Follow-up appropriate	Loss to f/u < 5%	Prospective calculation of sample size
Alarcon [55]	2	2	1	2	2	2	0
Barth [56]	2	2	1	2	2	1	0
Burkhart [57]	2	2	1	2	2	1	0
Denard [58]	2	1	2	2	2	1	0
Ferrando [59]	2	2	2	2	2	2	0
Greiner [23]	2	2	2	2	2	2	2
Kim [60]	2	2	1	2	2	1	0
Kocaoglu [61]	2	2	1	2	2	2	0
LaBelle [62]	2	2	1	2	2	1	0
Lacheta [63]	2	2	2	2	2	2	0
Lee and Min [64]	2	2	1	2	2	1	0
Lim [65]	2	2	2	2	2	2	0
Mihata [5]	2	2	2	2	2	2	0
Okamura [66]	2	2	2	2	2	2	2
Ohta [67]	2	1	1	2	2	1	0
Ozturk [68]	2	2	2	2	2	2	0
Pashuck [69]	2	1	2	2	2	2	0
Pennington [70]	2	2	2	2	2	2	0
Polacek [71]	2	2	2	2	2	2	0
Polacek [72]	2	2	2	2	2	2	0
Takayama [73]	2	2	2	2	2	2	0

Balloon spacer

Author and Date	Clearly state aim	Inclusion of consecutive patients	Prospective data collection	Unbiased endpoint assessment	Follow-up appropriate	Loss to f/u < 5%	Prospective calculation of sample size
Basat [84]	2	1	2	2	2	2	0
Batki [85]	2	1	2	2	2	0	0
Deranlot [86]	2	2	2	2	2	2	0
Familiari [87]	2	1	2	2	2	0	0
Gervasi [88]	2	1	2	2	2	0	0
Gervasi [89]	2	1	2	2	2	1	0
Iban [90]	2	2	2	2	2	0	0
Lorente [91]	2	2	2	2	2	1	0
Maman [92]	2	1	2	2	2	0	0
Senekovic [93]	2	1	2	2	2	1	0

Debridement

Author and Date	Clearly state aim	Inclusion of consecutive patients	Prospective data collection	Unbiased endpoint assessment	Follow-up appropriate	Loss to f/u < 5%	Prospective calculation of sample size
Berth [14]	2	1	2	2	2	2	0
Blanke [74]	2	1	1	2	2	2	0
Boileau [75]	2	1	1	2	2	2	0
Cavalier [16]	2	2	2	2	2	2	0
Franceschi [21]	2	1	2	2	2	2	0
Heuberer [24]	2	1	2	2	2	0	0
Klinger [76]	2	2	2	2	2	2	0
Lee [77]	2	2	2	2	2	2	0
Mirzaee [78]	2	2	2	2	2	2	0
Park [79]	2	1	1	2	2	1	0
Scheibel [80]	2	1	2	2	2	2	0
Vad [81]	2	0	0	2	2	0	0
Veado [82]	2	0	2	2	2	1	0
Verhelst [83]	2	1	2	2	2	1	0

Additional full risk-of-bias table for comparative studies

Author and Date	Comparisons	Clearly state aim	Inclusion of consecutive patients	Prospective data collection	Unbiased endpoint assessment	Follow-up appropriate	Loss to f/u < 5%	Prospective calculation of sample size	Adequate control group	Contemporary groups	Baseline equivalence of groups	Adequate statistical analyses
Baverel [13]	PCR vs LDTT	2	2	2	2	2	2	0	2	2	2	2
Berth [14]	PCR vs debridement	2	1	2	2	2	2	0	2	2	2	2
Besnard [15]	PCR vs complete	2	2	1	2	2	0	0	2	2	0	2
Cavalier [16]	PCR vs debridement	2	2	2	2	2	2	0	2	2	1	2
Farazdaghi [20]	PCR vs complete	2	1	1	2	2	1	2	2	2	0	2
Franceschi [21]	PCR vs debridement	2	1	2	2	2	2	2	2	2	2	2
Greiner [23]	PCR vs SCR	2	2	2	2	2	2	2	2	2	2	2
Heuberer [24]	PCR vs debridement	2	1	2	2	2	0	0	2	2	1	2
Holtby [25]	PCR vs complete	2	2	2	2	2	2	0	2	2	1	2
Jeong [26]	PCR vs complete	2	1	1	2	2	1	0	2	2	1	2
Kim [28]	PCR vs complete	2	2	2	2	2	2	0	2	2	2	2
Malahias [30]	PCR vs PCR and balloon	2	1	2	2	2	2	0	2	2	0	2
Mori [31]	PCR vs graft interposition	2	2	2	2	2	2	0	2	2	2	2
Moser [32] 2007	PCR vs complete vs debridement	2	1	2	2	2	2	0	2	2	2	2
Pandey [33]	PCR vs graft interposition	2	2	2	2	2	2	0	2	2	0	2
Park [34]	PCR vs augmentation	2	1	2	2	2	2	0	2	2	2	2
Paribelli [35]	PCR vs LDTT	2	1	2	2	2	2	0	2	2	2	2
Barth [56]	SCR vs complete vs augmentation	2	2	1	2	2	1	0	2	2	2	2
Kocaoglu [61]	SCR vs SCR and PCR	2	2	1	2	2	2	0	2	2	2	2
Okamura [66]	SCR 1 layer graft vs 3 layers	2	2	2	2	2	2	2	2	2	2	2
Ozturk [68]	SCR vs LDTT	2	2	2	2	2	2	0	2	2	2	2

Search strategy

• 1 Rotator cuff/

• 2 Rotator Cuff Injuries/

• 3 1 or 2

• 4 (massive or large or irreparable).ti,ab.

• 5 3 and 4

• 6 ((massive or large or irreparable) adj3 (rotator cuff* adj3 (tear* or rupture* or injur*))).ti,ab.

• 7 ((massive or large or irreparable) adj3 (supraspinatus adj3 (tear* or rupture* or injur*))).ti,ab.

• 8 ((massive or large or irreparable) adj3 (infraspinatus adj3 (tear* or rupture* or injur*))).ti,ab.

• 9 ((massive or large or irreparable) adj3 (subscapularis adj3 (tear* or rupture* or injur*))).ti,ab.

• 10 ((massive or large or irreparable) adj3 (teres minor adj3 (tear* or rupture* or injur*))).ti,ab.

• 11 ((massive or large or irreparable) adj3 (posterosuperior adj3 (tear* or rupture* or injur*))).ti,ab.

• 12 6 or 7 or 8 or 9 or 10 or 11

• 13 5 or 12

• 14 Arthroscopy/

• 15 Debridement/

• 16 arthroscop*.ti,ab.

• 17 debridement*.ti,ab.

• 18 Superior capsul* reconstruction*.ti,ab.

• 19 capsul* reconstruction*.ti,ab.

• 20 Reconstructive Surgical Procedures/

• 21 Reconstruct*.ti,ab.

• 22 Repair*.ti,ab.

• 23 Tenotomy/

• 24 Tenodesis/

• 25 tenotom*.ti,ab.

• 26 tenodesis.ti,ab.

• 27 partial*.ti,ab.

• 28 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27

• 29 balloon*.ti,ab.

• 30 spacer*.ti,ab.

• 31 29 or 30

• 32 graft*.ti,ab.

• 33 graftjacket.ti,ab.

• 34 patch*.ti,ab.

• 35 allografts/

• 36 autografts/

• 37 surgical mesh/

• 38 bioprosthesis/

• 39 tissue scaffolds/

• 40 extracellular matrix/

• 41 acellular dermis/

• 42 allograft*.ti,ab.

• 43 autograft*.ti,ab.

• 44 surgical mesh.ti,ab.

• 45 bioprosthe*.ti,ab.

• 46 tissue scaffold*.ti,ab.

• 47 extracellular matrix.ti,ab.

• 48 acellular dermal matrix.ti,ab.

• 49 bioartificial tendon*.ti,ab.

• 50 augment*.ti,ab.

• 51 "Zimmer Collagen Repair Patch".ti,ab.

• 52 "Permacol".ti,ab.

• 53 "TissueMend".ti,ab.

• 54 "BioBlanket".ti,ab.

• 55 "Conexa".ti,ab.

• 56 "AlloPatch".ti,ab.

• 57 "Shelhigh Encuff Patch".ti,ab.

• 58 "OrthADAPT".ti,ab.

• 59 "Restore".ti,ab.

• 60 "CuffPatch".ti,ab.

• 61 "Polytape".ti,ab.

• 62 "SportMesh".ti,ab.

• 63 "Arthelon".ti,ab.

• 64 "Gore-tex".ti,ab.

• 65 "BioFiber".ti,ab.

• 66 "STR Grafts".ti,ab.

• 67 "Lars Ligament".ti,ab.

• 68 "X-repair".ti,ab.

• 69 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46 or 47 or 48 or 49 or 50 or 51 or 52 or 53 or 54 or 55 or 56 or 57 or 58 or 59 or 60 or 61 or 62 or 63 or 64 or 65 or 66 or 67 or 68

• 70 13 and 28

• 71 13 and 31

• 72 13 and 69

• 73 70 or 71 or 72

• 74 exp animals/ not humans.sh.

• 75 73 not 7472 13 and 69

73 70 or 71 or 72

74 exp animals/ not humans.sh.

75 73 not 74