Outcome comparison of lower trapezius tendon transfer and arthroscopic rotator cuff tear repair using muscle advancement for massive rotator cuff tear: a systematic review

Abstract

Background

This systematic review evaluates the clinical outcomes of two distinct, arthroscopic techniques for the surgical repair of massive rotator cuff tears; lower trapezius tendon (LTT) transfer and muscle advancement (MA).

Methods

Eleven studies, involving 433 patients, selected based on PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, were analyzed.

Results

The findings reveal that MA significantly improves functional outcomes, demonstrated by higher Constant-Murley scores (mean difference: 26.26 vs. 18.31, P<0.001), University of California, Los Angeles shoulder scores (14.95 vs. 8.3, P<0.001), acromiohumeral distance (AHD; 1.94 mm vs. 0.40 mm, P<0.001), and greater abduction recovery (46.48° vs. 31.86°, P=0.030). However, visual analog scale score was better reduced in the LTT transfer groups (–3.69 vs. –2.33, P<0.001), with greater external rotation improvement (25.67° vs. 7.74°, P<0.001) and lower retear rates (11.89% vs. 19.42%, P=0.031). The complication profiles differed between techniques: LTT transfer carried a higher risk of graft rupture (2.64% vs. 0%, P=0.031), while arthroscopic MA was associated with increased postoperative stiffness (2.91% vs. 0%, P=0.011).

Conclusions

Arthroscopic MA is recommended for younger, active patients with mobile residual tissue to optimize abduction and AHD restoration. In contrast, LTT transfer is better suited for cases involving massive defects requiring dynamic stabilization and external rotation recovery. These findings emphasize the importance of individualized surgical planning that considers tear severity, tissue viability, and patient functional demands. Despite limitations stemming from retrospective study designs and clinical heterogeneity, this review highlights the distinct clinical advantages and appropriate indications for both techniques.

INTRODUCTION

Massive rotator cuff tears (MRCTs) are conventionally defined as full-thickness tears involving at least two of the four rotator cuff tendons or tears greater than 5 cm in diameter in the coronal or axial plane [1,2]. Studies have shown a significant association between increased tear size and suboptimal postoperative outcomes, coupled with elevated rates of structural failure [3]. Retrospective data have reported retear rates as high as 78% following surgical repair of MRCTs [4]. Despite that considerable failure rate, primary repair persists within the current literature as a supported therapeutic strategy for MRCTs. Part of the problem is that many MRCTs present with significant tendon retraction or inadequate tendon length, precluding reattachment to the anatomical footprint and considerably complicating the surgical repair process. Consequently, a diverse spectrum of treatment strategies has been proposed for the management of MRCTs. These range from conservative approaches such as physical therapy to surgical interventions such as arthroscopic debridement, partial repair, graft interposition, tendon transfer, superior capsular reconstruction, balloon arthroplasty, and reverse total shoulder arthroplasty [5]. The comparative efficacy of these treatment modalities remains insufficiently validated through rigorous clinical evidence.

Lower trapezius tendon (LTT) transfer is a specific tendon transfer procedure used to restore external rotation (ER) and abduction capabilities in patients presenting with posterosuperior MRCTs or certain forms of nerve-related shoulder dysfunction. Initially described by Elhassan et al. [6,7] for the treatment of paralytic shoulder conditions, LTT transfer has evolved into an arthroscopically assisted technique. The procedure typically involves the use of an autograft or allograft, commonly an Achilles tendon allograft, to bridge the lower trapezius tendon to the greater tuberosity insertion site. Biomechanically, LTT functions analogously to the infraspinatus muscle, contributing to the restoration of physiological scapulohumeral rhythm and the enhancement of dynamic glenohumeral joint stability [8,9]. Clinical investigations have reported favorable short- and mid-term outcomes following LTT transfer, characterized by effective pain relief, improved range of motion (ROM), and enhanced overall shoulder function [10].

Muscle advancement (MA), specifically involving the supraspinatus, was initially described by Debeyre et al. [11] in 1965. This technique involves the elevation and subsequent lateral advancement of the supraspinatus muscle from its native fossa, often performed in conjunction with an acromial osteotomy. The main advantage afforded by this surgical technique lies in its ability to facilitate substantial lateral advancement of the supraspinatus tendon. This mobilization permits anatomical repositioning of the tendon to achieve complete footprint coverage on the greater tuberosity and culminate in a tension-free repair construct [12]. Originally executed as an open procedure, the arthroscopic MA technique underwent refinement with the advent of arthroscopic methodologies. Lafosse et al. [13] augmented the procedure by introducing arthroscopic suprascapular nerve release, and Gupta et al. [12] further progressed the technique by establishing a fully arthroscopic approach to MA. This minimally invasive adaptation not only mitigates surgical trauma and associated postoperative complications but also furnishes a novel strategic option for the management of challenging MRCTs.

Both LTT transfer and arthroscopic MA have independently demonstrated positive clinical outcomes in the therapeutic management of MRCTs. In spite of their individual successes, however, the relative efficacy of these distinct techniques continues to be a subject of ongoing investigation [7,10,14-16]. Therefore, this review offers a comprehensive evaluation and comparative analysis of the clinical outcomes, complication profiles, and failure rates associated with LTT transfer and arthroscopic MA.

METHODS

Study Selection

In this study, we used a network meta-analysis and systematic review to evaluate LTT transfer and arthroscopic MA as surgical interventions for MRCTs. The literature search and evaluation processes were meticulously conducted by all authors and adhered to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines to ensure a transparent and rigorous review. Any disagreements during the review were resolved by the senior author. Prior to data retrieval, the systematic review was registered in the PROSPERO database (CRD42024532236) to establish a clear research protocol [17]. The review process commenced with an initial screening of titles and abstracts to identify relevant studies, followed by full-text reviews to confirm their alignment with the predefined inclusion criteria. References from the included studies were thoroughly examined to capture any relevant research that might have been overlooked during the initial selection process. This systematic approach ensured a comprehensive evaluation of the available evidence and contributed to a robust comparison of the efficacy and safety of two surgical techniques for managing MRCTs.

Database Screening and Search Strategy for a Systematic Review of Surgical Techniques for MRCTs

In October 2024, an extensive search of the Medline, Embase, and Cochrane Library databases was undertaken to identify studies pertinent to a systematic review comparing surgical techniques for MRCTs. Reference lists from prior systematic reviews were thoroughly examined to ensure the inclusion of all relevant correlational studies. The comprehensive search strategy used specific Boolean operators and keywords to refine the search and include pertinent studies. The following search terms were used: (“lower trapezius” OR “lower trapezius tendon transfer”) AND “rotator cuff,” and (“muscle advancement” OR “muscular advancement”) AND “rotator cuff.”

The search terms were formulated to return a comprehensive array of studies about the application of the LTT transfer and arthroscopic MA techniques for the management of rotator cuff injuries. This meticulous approach was intended to ensure the inclusion of all relevant literature, thereby providing a solid foundation for subsequent network meta-analyses and systematic reviews.

Eligibility Criteria

For this systematic review comparing LTT transfer and arthroscopic MA for the treatment of MRCTs, studies were included based on the following rigorous criteria to ensure the reliability and relevance of the findings. (1) Retraction of a posterior superior cuff tear reaching at least the glenoid level on magnetic resonance imaging (MRI; corresponding to Patte stage 3) and significant fatty degeneration (classified as Goutallier stage 3 or 4). (2) Only studies involving patients who underwent either arthroscopically assisted or all-arthroscopic LTT transfer or arthroscopic MA procedures were considered, with a minimum follow-up period of 12 months postoperatively to capture long-term outcomes. (3) To ensure sufficient data for a meaningful analysis, we required that the studies have a minimum sample size of 10 patients. (4) We also mandated that the included studies be published in peer-reviewed journals, with the full text available in English, thereby guaranteeing the quality and accessibility of the research. (5) Functional outcomes in the selected studies were assessed using at least one of the following validated measurement tools: the Constant-Murley score (CMS), University of California, Los Angeles (UCLA) shoulder score, American Shoulder and Elbow Surgeons (ASES) shoulder score, visual analog scale (VAS) pain score, acromiohumeral distance (AHD), and ROM. The inclusion criteria were carefully designed to ensure that the review incorporated high-quality studies that provided relevant and robust data to compare the efficacy and outcomes of LTT transfer and arthroscopic MA in the management of MRCTs.

The exclusion criteria for this systematic review of LTT transfer and arthroscopic MA in MRCTs were established to maintain the focus and quality of the included studies. (1) Studies were excluded if they provided insufficient information regarding surgical techniques, rendering proper assessment impossible. (2) Additionally, studies that involved duplicate patient populations with shorter follow-up periods were excluded to prevent redundancy and ensure the inclusion of the most comprehensive data available. (3) Nonclinical studies, such as those conducted on cadavers or animals and those focused on biomechanical analyses, were also excluded because they did not provide clinical outcomes directly relevant to patient care. (4) Several specific clinical exclusion criteria were applied to prevent the inclusion of studies in which certain arthroplasty options would be contraindicated, such as those involving patients with glenohumeral osteoarthritis or advanced rotator cuff tear arthropathy (Hamada grade ≥4). (5) Furthermore, we excluded case reports, reviews, systematic reviews (with or without meta-analyses), pilot studies, unpublished manuscripts, commentaries, editorials, lectures, meeting abstracts, book chapters, conference proceedings, and dissertations. These exclusion criteria were designed to ensure that only high-quality relevant clinical studies were included so that this review could provide a reliable comparison of the outcomes of LTT transfer and arthroscopic MA in the treatment of MRCTs.

Data Extraction and Quality Assessment

Two researchers (JL and VKM) reviewed the studies and extracted data on authorship, publication year, study quality, surgery type, demographics, and outcomes, including the UCLA shoulder score, CMS, ASES, AHD, forward flexion (FF), abduction, and ER. Quality was rated using the Newcastle-Ottawa scale (NOS) for cohort studies [18]. We discussed each discrepancy, referring to the original manuscript and predefined coding rules. Persistent disagreements were adjudicated by the senior author.

Data Analysis

Descriptive and clinical outcome data are expressed as the mean and standard deviation (SD). Weighted means and corresponding SDs were determined across studies using inverse-variance weighting. In the analysis of preoperative and postoperative scores, the mean difference between time points was calculated to assess the effectiveness of each intervention [19]. For studies with known preoperative and follow-up SDs, a correlation coefficient of 0.5 was applied to adjust variance estimation in longitudinal comparisons [20]. This correlation adjustment enhanced the precision of the pooled SD calculations when integrating repeated measures data. Continuous variables were analyzed using Student t-test, with Welch's correction applied for unequal variances when necessary. Categorical variables were evaluated using chi-square tests, supplemented by Fisher's exact test for small cell counts (expected frequencies <5). Statistical significance was defined as a two-tailed P<0.05. All analyses were conducted using R version 4.4.3 (R Foundation for Statistical Computing).

RESULTS

Study Selection

After removing duplicates, our comprehensive search strategy identified 954 studies (Fig. 1). Through the initial screening of titles and abstracts, we narrowed this number to 41 studies for a more detailed examination. Following the application of our predefined inclusion and exclusion criteria, 30 studies were excluded from the further analysis. Consequently, 11 studies were deemed appropriate for inclusion.

Fig. 1.

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) study selection flow diagram. MA: mus­cle advancement, LTT: lower trapezius tendon.

Description of Studies

Among the 11 research articles examined, six concentrated on LTT transfer, comprising three case series and three cohort studies [21-26]. All six studies used arthroscopically assisted techniques. Four studies on arthroscopic MA were cohort studies, and one was a case series, and they used either arthroscopically assisted surgery or exclusively arthroscopic methods [15,16,27-29]. Further details and specific characteristics of these studies, including the study designs, intervention specifics, and outcome measures, are provided in Tables 1 and 2, which offer a comprehensive overview of the methodologies and findings relevant to this systematic review.

Table 1.

Study details

Study	Year	Level of evidence/study type	Surgical approach	Type of graft	No. of shoulders	Age (yr), mean±SD (range)	Follow-up (mo), mean±SD (range)	Sex (male:female)
Arthroscopic MA
Gupta et al. [29]	2024	4/Retrospective case series	All-arthroscopic	No	43	56.2±6.2 (42–70)	18.8±11.3 (12–55)	29:9
Oh et al. [15]	2024	3/Cohort study	Arthroscopic and open	No	29	62.6±6.9	12±0	12:17
Yokoya et al. (study group) [28]	2020	3/Cohort study	Arthroscopic and open	PGA	47	68.3±8.1	24.2±1	27:20
Yokoya et al. (control group) [28]	2020	3/Cohort study	Arthroscopic and open	No	27	69.4±7.1	24.3±0.7	12:15
Yokoya et al. [16]	2019	3/Retrospective cohort study	Arthroscopic and open	No	26	68.6±8.8	28.3±7.3	13:13
Morihara et al. [27]	2018	NR	Arthroscopic and open	No	34	64.8±6.9 (46–73)	32.5 (24–70)	21:12
LTT transfer
Baek et al. [21]	2024	NR/Retrospective case series	Arthroscopic and open	Autograft	22	63.5±4.9 (51–71)	35.9±15.9 (24–64)	17:5
Chopra et al. [25]	2024	4/Treatment study	Arthroscopic and open	Allograft	19	56.7±10 (29–72)	14.6±10 (6–45)	17:2
Baek et al. [26]	2024	4/Treatment study	Arthroscopic and open	Allograft	36	63.4±5.4 (51–73)	58.2±5.3 (48–70)	24:12
Marigi et al. [24]	2023	3/Retrospective cohort study	Arthroscopic and open	Allograft	72	56.8±7.9	37.2±16.8	58:14
Baek et al. [23]	2022	3/Cohort study	Arthroscopic and open	Allograft	36	63.1±5.5	37.6±9.8	22:14
Baek et al. [22]	2022	3/Retrospective cohort study	Arthroscopic and open	Allograft	42	63.3±5.8	24±0	27:15

SD: standard deviation, MA: muscle advancement, PGA: polyglycolic acid, NR: not reported, LTT: lower trapezius tendon.

Table 2.

Preoperative study characteristics

Variable	Arthroscopic MA	LTT transfer	P-value
Age (yr)	64.58±8.80	60.69±7.44	<0.001***
Follow-up (mo)	19.68±8.56	36.13±16.85	<0.001***
Sex (male:female)	114:86	165:62	<0.001***
No. of shoulders	206	227	-

Values are presented as mean±standard deviation.

MA: muscle advancement, LTT: lower trapezius tendon.

^***P<0.001; indicates statistical significance.

Risk of Bias

All 11 studies included in the analysis were evaluated using the NOS to assess their methodological quality. The seven cohort studies were scored using the NOS for cohort studies. The remaining four studies, classified as case series, were assessed using the NOS for case-control studies [18]. The scoring outcomes, as detailed in Table 3, indicated that all studies met the criteria for high quality.

Table 3.

Quality assessment of the included studies according to the modified Newcastle-Ottawa scale

Study	Year	Level of evidence/study type	Selection				Comparability	Outcome			Quality/total score
			Representativeness of the exposed cohort	Selection of the nonexposed cohort	Ascertainment of exposure	Demonstration that outcome of interest was not present at start of study	Comparability of cohorts on the basis of the design or analysis	Assessment of outcome	Follow-up was long enough for outcomes to occur	Adequacy of follow-up of cohorts
Cohort study
MA
Oh et al. [15]	2024	3/Cohort study	√	√	√	√	√	√	√	√	8
Yokoya et al. [28]	2020	3/Cohort study	√	√	√	√	√	√	√	√	8
Yokoya et al. [16]	2019	3/Retrospective cohort study	√	√	√	√	√	√	√	√	8
Morihara et al. [27]	2018	NR	√	√	√	√	√	√	√	√	8
LTT Transfer
Marigi et al. [24]	2023	3/Retrospective cohort study	√	√	√	√	√	√	√	√	8
Baek et al. [23]	2022	3/Cohort study	√	√	√	√	√	√	√	√	8
Baek et al. [22]	2022	3/Retrospective cohort comparison treatment study	√	√	√	√	√	√	√	√	8
								Exposure

Case-control study	Year	Level of evidence/study type	Selection				Comparability	Exposure			Quality/total score
			The case definition is adequate	Representativeness of the cases	Selection of controls	Definition of controls	Comparability of cohorts on the basis of the design or analysis	Assessment of outcome	Same method of assessment for cases and controls	Adequacy of cohort follow-up
Case-control study			The case definition is adequate	Representativeness of the cases	Selection of controls	Definition of controls	Comparability of cohorts on the basis of the design or analysis	Assessment of outcome	Same method of assessment for cases and controls	Adequacy of cohort follow-up
MA
Gupta et al. [29]	2024	4/Retrospective case series	√	√	√	√	√	√	√	√	8
LTT transfer
Chopra et al. [25]	2024	4/Case series; treatment study	√	√	√	√	√	√	√	√	8
Baek et al. [26]	2024	4/Case series; treatment study	√	√	√	√	√	√	√	√	8
Baek et al. [21]	2024	Retrospective case series	√	√	√	√	√	√	√	√	8

MA: muscle advancement, NR: not reported, LTT: lower trapezius tendon.

The quality of the studies included in this systematic review was classified based on the total score obtained from the modified NOS. According to the scoring system, studies were categorized as "low" quality if they scored between 0 and 3 points, "moderate" quality for scores between 4 and 6 points, and "high" quality for scores between 7 and 9 points. Each study was evaluated based on specific criteria within categories, with a maximum of 1 or 2 points awarded for each criterion, as outlined by the modified NOS rules. The scoring allowed for a standardized assessment of study quality, ensuring that only robust and reliable evidence was included in the analysis. The consistent high-quality ratings across studies underscore the validity of our review's conclusions about the efficacy of LTT transfer and arthroscopic MA in the treatment of MRCTs [18].

Study Characteristics

The procedural types, namely LTT transfer or arthroscopic MA, and surgical approaches used in the studies are outlined in Table 1, and the preoperative characteristics of the patients are presented in Table 2. The six studies in the LTT transfer group reported on 227 shoulders, and the five studies in the arthroscopic MA group reported on 206 shoulders. The mean age of patients at the time of surgery was 60.69±7.44 years for the LTT transfer group and 64.58±8.80 years for the arthroscopic MA group, with a statistically significant difference between the groups (P<0.001). Furthermore, 165 patients in the LTT transfer group were male, compared with 114 patients in the arthroscopic MA group; this difference was also statistically significant (P<0.001). The LTT transfer group exhibited a longer follow-up period than the arthroscopic MA group, averaging 36.13 months versus 19.68 months, respectively, with the difference being statistically significant (P<0.001).

Surgical Techniques

All six studies in the LTT transfer group concentrated on arthroscopy-assisted surgical techniques. Concerning the selection of graft material, five of these studies used an allogeneic Achilles tendon as the graft [22-26], and the remaining study used autologous fascia lata [21]. In the arthroscopic MA group, one study used an all-arthroscopic approach [29], and four studies used arthroscopy-assisted surgical methods. Because arthroscopic MA surgery involves freeing and advancing muscles to reduce tension and enhance repair, the procedure generally does not require the use of a graft [15,16,27,28]. However, in one cohort study, a polyglycolic acid sheet was used as an adjunct to support rotator cuff repair and promote tissue regeneration [28].

Clinical Outcomes

Constant-Murley score

CMS was reported in two studies involving 58 shoulders [21,26] in the LTT transfer group and all five studies involving 206 shoulders in the arthroscopic MA group [15,16,27-29]. Our analysis revealed that the preoperative CMS in the LTT transfer group was significantly higher than that in the arthroscopic MA group (52.37±13.48 vs. 45.96±18.79, P=0.004). However, no significant difference in postoperative scores was observed between the LTT transfer and arthroscopic MA groups (72.58±15.02 vs. 70.68±9.95, P=0.258). Additionally, the mean improvement in CMS was significantly greater in the arthroscopic MA group than the LTT transfer group (26.26 vs. 18.31, P<0.001) (Table 4).

Table 4.

Preoperative and postoperative clinical and radiological outcomes

Study	No. of shoulders	Preop CMS	Postop CMS	Mean difference	Preop UCLA score	Postop UCLA score	Mean difference	Preop AHD	Postop AHD	Mean difference	Preop VAS pain score	Postop VAS pain score	Mean difference	Preop ASES score	Postop ASES score	Mean difference
MA
Gupta et al. [29]	43	55.7±20.1	84.1±11.4		16.7±4.8	31.9±4.2					3.8±2.8	0.5±1		54.4±20	93.3±10.5
Oh et al. [15]	29	59.4±15.5	70.1±10.8	10.7	18.6±6.6	27±6.3	8.4				3.8±2.3	2.9±2.3	–0.9	60.6±21.0	74.5±20.9	13.9
Yokoya et al. (study group) [28]	47	41.2±16.1	70.4±18		14.6±5.2	29.6±6.4		6.9±2.8	9.1±2.7
Yokoya et al. (control group) [28]	27	42.7±15	69.3±14.3		12.6±3.1	29±6.4		7.4±2	9.3±2.3
Yokoya et al. [16]	26	43±15	71.2±12.7		12.5±3.1	28.5±6.3		7.9±2.2	9.4±2.9
Morihara et al. [27]	34	33.6±15.5	66.8±12.1		13.5±4.1	31.7±3.5
Total	206	45.96±18.79	72.58±15.02	26.62	14.89±5.16	29.84±5.83	14.95	7.30±2.49	9.23±2.66	1.94	3.80±2.61	1.47±2.03	–2.33	56.90±20.63	85.73±18.08	28.83
LTT transfer
Baek et al. [21]	22	51.5±13.2	71.3±13.1					7.9±2.5	7.5±2.1		4.3±1.4	1.4±0.7		54.8±14.5	76.5±15
Chopra et al. [25]	19										5.9±2	1.8±2		44.6±18	71.2±24
Baek et al. [26]	36	52.9±13.8	70.3±7.6		19.7±4.7	28±2.9		7.9±0.8	7.5±1		4.5±1.2	1.4±0.7		57.1±13.4	83.7±8.2
Marigi et al. [24]	72										6.1±2.5	1.1±2
Baek et al. [23]	36							7.1±2.5	7.5±1.2		4.5±1.8	1.3±1		47.2±17.3	84.8±7.6
Baek et al. [22]	42							7.2±2.1	8.7±2.3		4.3±1.6	1.7±1.3		55.8±16.7	82.7±8.6
Total	227	52.37±13.48	70.68±9.95	18.31	19.7±4.7	28±2.9	8.3	7.47±2.06	7.87±1.80	0.40	5.07±2.08	1.38±1.49	–3.69	52.59±16.51	81.13±12.91	28.54
P-value		0.004**	0.258	<0.001***	<0.001***	0.004**	<0.001***	0.562	<0.001***	<0.001***	<0.001***	0.732	<0.001***	0.124	0.056	0.919

Values are presented as mean±standard deviation.

Preop: preoperative, Postop: postoperative, CMS: Constant-Murley score, UCLA: University of California, Los Angeles shoulder, AHD: acromiohumeral distance, VAS: visual analog scale, ASES: American Shoulder and Elbow Surgeons, MA: muscle advancement, LTT: lower trapezius tendon.

^**P<0.01,

^***P<0.001; indicate statistical significance.

UCLA score

The UCLA shoulder score was reported in one study [26] in the LTT transfer group involving 36 shoulders and in all five studies in the arthroscopic MA group involving 206 shoulders [15,16,27-29]. Our analysis revealed that the preoperative UCLA score in the LTT transfer group was significantly higher than that in the arthroscopic MA group (14.89±5.16 vs. 19.7±4.7, P<0.001). Postoperatively, the UCLA score in the arthroscopic MA group was significantly higher than that in the LTT transfer group (29.84±5.83 vs. 28±2.9, P=0.004). Furthermore, the improvement in the UCLA score was significantly greater in the arthroscopic MA group than the LTT transfer group (14.95 vs. 8.3, P<0.001) (Table 4).

ASES score

ASES scores were reported in five studies [21-23,25,26] involving 155 shoulders in the LTT transfer group and two studies [15,29] involving 72 shoulders in the arthroscopic MA group. The analysis revealed no statistically significant difference in preoperative ASES scores between the groups (56.90±20.63 vs. 52.59±16.51, P=0.123). Similarly, no significant difference was observed in the postoperative scores between the groups (85.73±18.08 vs. 81.13±12.91, P=0.055). Furthermore, the changes in scores from the preoperative to postoperative assessments did not differ significantly between the groups (28.83 vs. 28.54, P=0.919) (Table 4).

VAS pain score

Pre- and postoperative VAS pain scores were measured in six studies, involving 227 cases in the LTT transfer group [21-26] and 72 cases in the arthroscopic MA group [15,29]. Our analysis revealed that the preoperative VAS score for the LTT transfer group was significantly higher than that in the arthroscopic MA group (5.07±2.08 vs. 3.80±2.61, P<0.001). However, the postoperative VAS did not differ significantly between the groups (1.38±1.49 vs. 1.47±2.03, P=0.732). Additionally, the VAS scores before and after surgery differed significantly between the groups with the LTT transfer group showing a greater reduction than the arthroscopic MA group (–3.69 vs. –2.33, P<0.001) (Table 4).

Radiological evaluation

Pre- and postoperative AHD was measured in four studies, involving 136 cases in the LTT transfer group [21,24-26] and 100 cases in the arthroscopic MA group [16,27,28]. Our analysis found no statistically significant difference in the preoperative AHD between the groups (7.30±2.49 vs. 7.47±2.06, P=0.562). However, the arthroscopic MA group had significantly greater postoperative AHD than the LTT transfer group (9.32±2.66 vs. 7.87±1.80, P<0.001). Furthermore, the changes in AHD before and after surgery differed significantly between the groups (1.94 vs. 0.40, P<0.001) (Table 4).

Range of motion

FF was measured in all six studies in the LTT transfer group involving 227 cases [21-26] and 177 cases in the arthroscopic MA group [16,27-29]. Abduction was measured in 58 patients in the LTT transfer group [21,26] and 77 cases in the arthroscopic MA group [27,29]. ER was measured in all 227 cases in the LTT transfer group [21-26] and all 206 cases in the arthroscopic MA group [15,16,27-29]. Preoperative FF did not differ significantly between the LTT transfer and arthroscopic MA groups (118.69°±43.51° vs. 124.83°±42.06°, P=0.154). Similarly, postoperative FF did not differ significantly between the groups (152.37°±24.42° vs. 153.99°±28.45°, P=0.539), and the improvement in postoperative ROM for FF was comparable between the groups (33.68° vs. 29.16°, P=0.230) (Table 5).

Table 5.

Preoperative and postoperative ROM

Study	No. of shoulders	Preop ROM abduction (°)	Postop ROM abduction (°)	Mean difference (°)	Preop ROM FF (°)	Postop ROM FF (°)	Mean difference (°)	Preop ROM ER (°)	Postop ROM ER (°)	Mean difference (°)
MA
Gupta et al. [29]	43	126±42	161±21		139±47	170±22		47±21	62±12
Oh et al. [15]	29							36.4±20.3	49.3±22	12.9
Yokoya et al. (study group) [28]	47				111±43	140±26		39.8±29.3	42.7±22.3
Yokoya et al. (control group) [28]	27				122±37	148±21		51.4±22.3	50.2±13.6
Yokoya et al. [16]	26				120±35	148±18		49.3±23	51.7±13.6
Morihara et al. [27]	34	91±45 (20–170)	152±23 (40–170)		100±39 (40–170)	154±18 (120–170)		37±17 (–40 to 70)	49±21 (15–80)
Total	206	110.55±46.71	157.03±22.36	46.48	118.69±43.51	152.37±24.42	33.68	43.08±23.61	50.82±19.34	7.74
LTT transfer
Baek et al. [21]	22	110±38	140±32		126±39	155±29		26±15	39±17
Chopra et al. [25]	19				139.5±26	147.4±29		10.5±17	40.5±13
Baek et al. [26]	36	116±41	149±32		134±42	160±30		25±13	43±12
Marigi et al. [24]	72				111±41	143±31		22±26	51±15
Baek et al. [23]	36				134.2±53.6	165.7±22.3		27.5±14.3	51.7±10.9
Baek et al. [22]	42				125.4±34.9	160.1±19.5		22.3±10.9	54.8±9.8
Total	227	113.72±39.66	145.59±32.02	31.86	124.83±42.06	153.99±28.45	29.16	22.83±18.81	48.50±14.11	25.67
P-value		0.670	0.022*	0.030*	0.154	0.539	0.230	<0.001***	0.160	<0.001***

Values are presented as mean±standard deviation (SD) or mean±SD (range).

ROM: range of motion, Preop: preoperative, Postop: postoperative, FF: forward flexion, ER: external rotation, MA: muscle advancement, LTT: lower trapezius tendon.

^*P<0.05; indicates statistical significance.

Preoperative abduction did not differ significantly between the LTT transfer and arthroscopic MA groups (110.55°±46.70° vs. 113.72°±39.66°, P=0.670). Postoperative abduction was better in the MA group than the LTT transfer group, with a significant difference between them (157.03°±22.36° vs. 145.59°±32.02°, P=0.022). The improvement in abduction was also significantly greater in the arthroscopic MA group than the LTT transfer group (46.48° vs. 31.86°, P=0.030) (Table 5).

On the other hand, significant differences were observed in preoperative ER, with the LTT transfer group showing lower values than the arthroscopic MA group (22.83°±18.81° vs. 43.08°±23.61°, P<0.001). Despite that initial difference, postoperative ER did not differ significantly between the groups (50.82°±19.34° vs. 48.50° ±14.11°, P=0.160). Interestingly, the improvement in postoperative ROM for ER was significantly greater in the LTT transfer group than the arthroscopic MA group (25.67° vs. 7.74°, P<0.001) (Table 5).

Complications

In the arthroscopic MA group, 40 instances of failure/retear/reoperation occurred, constituting 19.42% of the complications, along with three cases of infection (1.46%), 6 cases of stiffness (2.91%), and two cases of nerve palsy (0.97%). The reported nerve palsy involved numbness of the fingers post-surgery, and the infection was identified as a superficial skin infection. In the LTT transfer group, 27 instances of failure/retear/reoperation occurred, accounting for 11.89% of complications, along with six cases of infection (2.64%), one case of persistent discomfort (0.44%), and six cases of complete graft rupture (2.64%). Overall, the incidence of complications did not differ significantly between the LTT transfer and arthroscopic MA groups (51/206 vs. 40/227, P=0.069). The complications of failure/retear/reoperation and stiffness were more prevalent in the arthroscopic MA group than the LTT transfer group, with a significant difference observed (failure/retear/reoperations, P=0.031; stiffness, P=0.011). The LTT transfer group exhibited a higher incidence of complete graft ruptures than the arthroscopic MA group (P=0.031), and no significant difference was found between the groups in other complications (infection, P=0.508; nerve palsy, P=0.226; persistent discomfort, P=1.000) (Table 6).

Table 6.

Complications

Study	No.	Total	Infection	Nerve palsy	Failure/retear/reoperation	Complete graft ruptures	Stiffness	Consistent discomfort
MA
Gupta et al. [29]	43	12	2 (4.65)		4 (9.30)		6 (13.95)
Oh et al. [15]	29	9			9 (31.03)
Yokoya et al. (study group) [28]	47	6			6 (12.77)
Yokoya et al. (control group) [28]	27	7			7 (25.93)
Yokoya et al. [16]	26	6			6 (23.08)
Morihara et al. [27]	34	11	1 (2.94)	2 (5.88)	8 (23.53)
Total	206	51	3 (1.46)	2 (0.97)	40 (19.42)	0	6 (2.91)	0
LTT transfer
Baek et al. [21]	22	4	1 (4.55)		2 (9.09)			1 (4.55)
Chopra et al. [25]	19	3	1 (5.26)		2 (10.53)
Baek et al. [26]	36	9	2 (5.56)		7 (19.44)
Marigi et al. [24]	72	13			7 (9.72)	6 (8.30)
Baek et al. [23]	36	3			3 (8.33)
Baek et al. [22]	42	8	2 (4.76)		6 (14.29)
Total	227	40	6 (2.64)	0	27 (11.89)	6 (2.64)	0	1 (0.44)
P-value		0.069	0.508	0.226	0.031*	0.031*	0.011*	1.000

Values are presented as number (%).

MA: muscle advancement, LTT: lower trapezius tendon.

^*P<0.05; indicates statistical significance.

DISCUSSION

This systematic review included 11 studies, encompassing 433 cases of rotator cuff injuries treated with either LTT transfer or arthroscopic MA. The LTT transfer group comprised 227 cases from six studies [21-26], and the arthroscopic MA group consisted of 206 cases from five studies [15,16,27-29]. The arthroscopic MA group exhibited significantly superior functional improvement, as measured by the CMS (P<0.001) and UCLA shoulder score (P<0.001), in addition to enhanced structural restoration (P<0.001) and greater recovery of abduction (P=0.022). Conversely, the LTT transfer group demonstrated significantly greater pain reduction (P<0.001), superior improvement in external rotation (P<0.001), and lower incidence of overall failure/retear (P=0.031) and postoperative stiffness (P=0.011). Nonetheless, the LTT transfer procedure was associated with a higher occurrence of graft-specific complications, including complete graft rupture (P=0.031), highlighting the risks inherent with grafts.

The LTT transfer procedure involves relocating the lower trapezius muscle to the greater tuberosity of the humerus to replace or enhance the function of the damaged rotator cuff, particularly targeting the infraspinatus and teres minor muscles. The lower trapezius originates from the spinous processes of the thoracic vertebrae, and following the transfer, it requires the re-establishment of a mechanical connection to the humerus [10,30]. Conversely, arthroscopic MA repairs address tears by advancing the remaining rotator cuff tissue, such as the supraspinatus, with the objective of restoring the original anatomical configuration and mechanical equilibrium of the rotator cuff in a tension-free repair. The primary distinction between these two surgical techniques is whether they reconstruct or restore the original anatomical structure [11,12].

Arthroscopic MA can more directly restore the depressing effect of the rotator cuff on the humeral head, thereby increasing AHD [28]. After arthroscopic MA repair, the mechanical support of the rotator cuff is restored, effectively counteracting the superior pull of the deltoid muscle and preventing superior migration of the humeral head. LTT transfer enhances the dynamic stability of the shoulder joint by transferring the lower trapezius, but its direct repair effect on the rotator cuff is limited, failing to fully restore the depressing effect of the rotator cuff on the humeral head, so it is less effective than arthroscopic MA in counteracting the superior pull of the deltoid muscle [14]. Although it can also restore AHD, its improvement is limited compared with that of arthroscopic MA.

MA can improve the abduction of the shoulder joint by directly repairing the supraspinatus [29], the primary muscle for abduction, and its repair directly affects ROM during abduction. Additionally, restoring the mechanical balance of the rotator cuff and recovering AHD can also improve abduction function. LTT transfer primarily replaces the external rotation of the infraspinatus and teres minor while enhancing the overall stability of the shoulder joint [31]. However, its effectiveness in restoring flexion and abduction is limited.

MA relies on the quality and advancement potential of the residual rotator cuff tissue, which depends on the patient's original tendon condition [32]. In patients with massive tears or poor tissue quality, the repair effect can be limited, and the risk of repair failure or retear increased. LTT transfer requires transferring the lower trapezius to the greater tuberosity of the humerus, a process involving complex anatomical dissection and tendon fixation, which can result in a higher incidence of complications such as graft rupture or infection [33].

Our analysis highlights distinct indications for LTT transfer and arthroscopic MA, necessitating a detailed comparison of patient selection criteria. LTT transfer is prioritized for extensive posterosuperior tears with preserved scapulothoracic motion, whereas arthroscopic MA is reserved for mobile tears where anatomical repair is feasible. Younger, high-demand patients with repairable tissue can benefit from MA’s structural restoration, whereas older patients and those with compromised tissue might favor LTT ’s dynamic stabilization despite the graft-related risks. Notably, the postoperative rehabilitation protocols for the two procedures differ substantially: LTT transfer requires prolonged immobilization (6–8 weeks) to protect the transferred tendon [22], followed by gradual passive-to-active motion, whereas arthroscopic MA allows for earlier mobilization (4–6 weeks) due to direct tendon-to-bone healing [15]. These protocols likely influence functional recovery timelines and complication rates, but comparative data remain sparse.

The studies included in this review did not address cost-effectiveness or technical feasibility, but those factors merit consideration. LTT transfer incurs higher costs (allograft procurement, extended operative time) and technical complexity (graft fixation), whereas arthroscopic MA relies on native tissue repair, reducing the resource burden. However, MA’s higher retear rates might offset the initial savings through revision surgeries. In a study with a mean follow-up duration of 58.2 months, Baek et al. [26] reported that LTT transfer resulted in significant improvements in both clinical and radiographic outcomes at final follow-up, whereas no studies have yet demonstrated the long-term clinical outcomes of arthroscopic MA. Future studies should integrate cost analyses, standardized rehabilitation reporting, and extended follow-up times to refine surgical algorithms and optimize value-based care.

This review was constrained by the lack of high-quality randomized controlled trials (RCTs), with most included studies being retrospective cohorts or case series. Those designs introduce risks of selection bias and unmeasured confounders and limit generalizability. Additionally, inconsistent reporting of long-term outcomes could affect the validity of our conclusions. Our exclusion of patients with Hamada grade ≥4—who are typically candidates for reverse shoulder arthroplasty rather than LTT transfer or arthroscopic MA—enhanced the internal validity of our analysis by ensuring cohort homogeneity and prioritizing clinically relevant comparisons for the target population. Nonetheless, heterogeneity persisted across studies due to variability in patient demographics, differences in follow-up duration, and the limitations inherent in non-randomized designs. Further methodological inconsistencies arose from variability in measurement protocols, such as AHD assessments (radiographic versus MRI-based techniques, discrepancies in patient positioning). To address those issues, we standardized the arthroscopic techniques, applied inverse-variance weighting for baseline-adjusted comparisons, and included exclusively high-quality studies (NOS-rated ≥7). Despite those efforts, residual variability from differences in graft/fixation, unmeasured confounders, and inherent selection bias underscores the need for future RCTs with unified imaging guidelines to resolve interpretative uncertainties and enhance generalizability.

CONCLUSIONS

LTT transfer provides superior pain relief and enhances external rotation, whereas arthroscopic MA contributes to functional recovery and structural restoration. LTT transfer is particularly suitable for addressing massive tears that require dynamic stability, and arthroscopic MA is advantageous for younger patients with mobile residual tissue. The choice of surgical intervention should be tailored to the specific needs of each patient.