1 Introduction

The management of posterosuperior irreparable rotator cuff tears (PSIRCTs) continues to pose a substantial clinical challenge.^1,2 These tears often result in persistent pain, muscle weakness, and limited range of motion (ROM), and conventional repair techniques frequently fail to achieve durable restoration of function.^3,4 Tendon transfer procedures, particularly lower trapezius transfer and latissimus dorsi (LD) transfer, have gained recognition as effective reconstructive options to restore range of motion and to improve shoulder kinematics in patients with PSIRCTs.^4–10 Among these, posterior LD transfer has demonstrated favorable clinical and radiologic outcomes, offering a reliable, joint-preserving alternative for patients without advanced glenohumeral arthritis.^6,9–11.

Numerous studies on arthroscopic repair of massive cuff tears have established that revision surgeries typically result in inferior outcomes compared with primary repairs, largely due to factors such as poor tendon quality, peritendinous fibrosis, and diminished vascularity, which together impede both biological healing and mechanical stability.^12–16 Likewise, secondary tendon transfer performed after a failed rotator cuff repair or other shoulder procedures may be affected by these compromised biological conditions.¹⁷ The presence of scar tissue, altered deltoid mechanics, and disrupted anatomic planes can further complicate tendon mobilization and fixation, increasing the risk of inadequate healing and diminished functional recovery.^12,15,16

Despite increasing clinical use, not many studies have directly compared outcomes between primary and secondary posterior LD transfers. The present study aims to evaluate and compare the clinical and radiologic outcomes of primary LD transfer (in shoulders without prior surgery) and secondary LD transfer (after failed rotator cuff repair) in patients with PSIRCTs. The hypothesis is that secondary LD transfer would yield inferior functional outcomes and higher re-tear rates compared with primary transfer due to scar formation, reduced healing capacity, compromised tendon footprint, and disrupted anatomy resulting from previous surgery.

2 Materials and methods

This study was approved by the Institutional Review Board by the authors’ affiliated institutions, and the requirement for informed consent was waived owing to the retrospective design of the study and the lack of additional harm to the patients.

3 Patient selection

This retrospective cohort study included patients who underwent arthroscopically-assisted posterior LD tendon transfer for PSIRCTs between April 2012 and June 2020. Patients were categorized into two groups: the Primary group (without any prior shoulder surgery), and the Secondary group (with a history of failed rotator cuff repair). The indications for surgery were as follows: (1) persistent pain and functional limitation despite failed conservative treatment; (2) absence or minimal glenohumeral arthritis (Hamada¹⁸ grade ≤2); (3) intact or repairable subscapularis tendon (Lafosse¹⁹ grade ≤ III); (4) high-grade fatty degeneration of the supraspinatus and infraspinatus muscles (Goutallier²⁰ grade 3 or 4); and (5) intraoperative confirmation that the retracted supraspinatus and infraspinatus tendons could not be mobilized to their original footprint despite standard release techniques. Patients were excluded if they were lost to follow-up or lacked postoperative clinical or radiologic data.

4 Surgical procedure

All procedures were performed by a single senior surgeon using a standardized arthroscopic-assisted posterior LD transfer technique previously described in the literature. Patients were positioned in the lateral decubitus position.¹¹ The procedure began with diagnostic arthroscopy to evaluate the long head of the biceps (LHB) tendon. Pathologic findings of the LHB were managed with tenodesis (Primary group, n = 9; Secondary group, n = 5) or tenotomy (Primary group, n = 12; Secondary group, n = 5), and some patients presented with auto-tenotomy status (Primary group, n = 2; Secondary group, n = 2). Concomitant subscapularis tears were repaired arthroscopically when present (Primary group, n = 9; Secondary group, n = 4). The supraspinatus and infraspinatus tendons were then assessed, and when these tendons could not be mobilized to their anatomic footprint despite standard release techniques, LDTM transfer was performed.

Harvesting of the LDTM was conducted through an mini-open posterior axillary approach. A skin incision was made along the anterior border of the LD with the shoulder flexed, internally rotated, and abducted. The LD tendons was identified and carefully isolated, without separating them, with protection of the anterior humeral circumflex vessels (“three sisters”) and the radial nerve. Blunt dissection was used to mobilize the neurovascular bundle safely, and the LD muscle was released from proximal to distal. The LD tendon was prepared with #2 nonabsorbable sutures using a continuous running locking stitch technique.

After harvesting the LD tendon, the supraspinatus and infraspinatus footprint was arthroscopically debrided and decorticated to promote tendon-to-bone healing. In cases of secondary surgery, any retained suture material from previous failed repairs was removed as necessary. Under arthroscopic guidance, the prepared LD tendon was passed through the interval between the external rotators and the deltoid. Two triple-loaded suture anchors were inserted on the medial side of the prepared footprint, and the sutures were passed through the LD tendon. The tendon was secured to the supraspinatus footprint, just posterior to the bicipital groove, using a double-row suture-bridge technique with three knotless anchors to achieve stable fixation.

5 Postoperative Therapy protocol

Following surgery, patients were immobilized in a 30° abduction sling with the shoulder in neutral rotation for the first four weeks. During this period, self-assisted passive ROM exercises were performed to maintain joint mobility while protecting the tendon repair. Between weeks 4 and 12, rehabilitation progressed to gentle active-assisted and active exercises, focusing on gradual restoration of shoulder kinematics. After three months, a structured strengthening program targeting the rotator cuff and periscapular muscles was initiated to improve dynamic stability and functional recovery. Patients were advised to avoid heavy lifting or labor-intensive activities for at least six months postoperatively.

6 Clinical and radiologic assessment

Demographic and clinical data, including age, sex, body mass index (BMI), smoking status, symptom duration, and comorbidities, were systematically recorded. Pain intensity was assessed using the visual analog scale (VAS). Shoulder function was evaluated preoperatively and at final follow-up using validated outcome measures: Single Assessment Numeric Evaluation (SANE), Constant score, and American Shoulder and Elbow Surgeons (ASES) score. The Minimal Clinically Important Difference (MCID) was determined using a 0.5 standard deviation distribution-based method.²¹ Active ROM was measured with a goniometer and included forward elevation, abduction, external rotation at the side, and internal rotation behind the back, graded according to the highest vertebral level reached (0 = greater trochanter to 10 = T8). Isometric muscle strength was assessed using a handheld dynamometer (Hoggan Health Industries, UT, USA). Radiologic assessment included pre- and postoperative plain radiographs (anteroposterior, axillary, and lateral views) to evaluate acromiohumeral distance (AHD) and glenohumeral joint degeneration according to the Hamada classification. Postoperative tendon integrity was assessed via magnetic resonance imaging (MRI) at final follow-up. All imaging studies were independently reviewed by a musculoskeletal radiologist blinded to clinical outcomes. Perioperative and postoperative complications were prospectively documented at each follow-up visit.

7 Statistical analysis

Continuous variables were presented as mean ± standard deviation, and categorical variables were reported as frequencies and percentages. Within-group comparisons between preoperative and postoperative measures were performed using paired t-tests for normally distributed data or Wilcoxon signed-rank tests for non-normally distributed data. Categorical variables were analyzed with Chi-square or Fisher's exact tests, as appropriate. Between-group differences were assessed using independent t-tests for continuous variables and Chi-square tests for categorical variables, with Bonferroni correction applied for multiple comparisons. Statistical analyses were conducted using SPSS software (version 20.0; IBM Corp., Armonk, NY, USA), and a two-sided p-value <0.05 was considered statistically significant.

8 Results

Seven patients were excluded, including two deaths and 5 lost to follow-up or with incomplete data, leaving 51 patients for analysis (Primary group, n = 34; Secondary group, n = 16). Baseline demographics—including age, sex, BMI, comorbidities, smoking status, symptom duration, and dominant arm involvement—were comparable (Table I). Preoperative shoulder conditions, including fatty infiltration and irreducible retracted tendons were similar across groups. The Secondary group included 16 patients with failed supraspinatus, with or without infraspinatus repair, and 2 patients who had failed repairs twice, with an average time to surgery of approximately 1 year and 6 months.

At final follow-up, both groups demonstrated significant improvements in pain, patient-reported outcome measures (PROMs), and active ROM. VAS scores decreased from 5.1 ± 1.1 to 2.0 ± 1.2 in the Primary group and from 4.9 ± 0.9 to 2.8 ± 1.7 in the Secondary group. SANE, Constant, UCLA, and ASES scores also improved significantly in both groups (Table II). Active forward elevation, abduction, and external rotation increased significantly in the Primary group, whereas improvements in the Secondary group were not statistically significant; internal rotation remained largely unchanged. Direct comparison revealed significantly higher postoperative SANE (75.3 ± 11.6 vs. 63.0 ± 12.8, p = 0.001), Constant scores (70.4 ± 12.6 vs. 58.0 ± 12.6, p = 0.002), and ASES score (77.1 ± 14.0 vs. 63.3 ± 11.7, p = 0.001) isn the Primary group. Forward elevation, abduction, and external rotation tended to be higher in the Primary group, though not statistically significant. Muscle strength improved significantly for forward elevation, abduction, and external rotation in both groups, with no gains in internal rotation; comparisons showed significantly greater strength in the Primary group. Radiologic evaluation demonstrated preserved AHD and minimal progression of glenohumeral arthritis in both groups (Table III). The Primary group achieved higher rates of MCID across all PROMs (Table IV). The re-tear rate was significantly higher in the Secondary group than in the Primary group (31.2% vs. 11.7%, p = 0.034). Two patients in each group underwent conversion to lower trapezius tendon transfer due to symptomatic re-tear. Progression of severe arthritis (Hamada ≥3) occurred in one patient from each group, both of whom had re-tears and subsequently underwent conversion to reverse shoulder arthroplasty approximately three years postoperatively. Other complications, including revision surgery and infection, were infrequent and showed no significant difference between groups (Table V).

9 Discussion

This study demonstrates that arthroscopically-assisted posterior LD tendon transfer leads to significant improvements in pain, shoulder function, and ROM in patients with PSIRCTs. Both Primary and Secondary groups achieved significant improvement in SANE, Constant, and ASES scores, as well as restoration of active forward elevation and abduction. However, direct comparison showed that the Primary group attained significantly higher postoperative SANE, Constant, and ASES scores and greater strength in forward elevation, abduction, and external rotation. Although improvements in active ROM were not statistically significant between groups, the Primary group showed a trend toward higher postoperative ROM. Re-tear rates were significantly higher in the Secondary group, suggesting that prior rotator cuff repair may modestly impair tendon healing and certain aspects of functional recovery.

Many studies on arthroscopic repair of massive rotator cuff tears has consistently shown that revision procedures yield less favorable outcomes than primary repairs, largely due to compromised tendon quality, peritendinous fibrosis, and reduced vascularity, which limit both biological healing and mechanical durability.^12–16 Alternative procedures such as superior capsular reconstruction or reverse shoulder arthroplasty may be used for failed repairs, but prior surgical interventions are associated with higher complication rates and less predictable outcomes.^22–24 In contrast, Baek et al.²⁵ reported that lower trapezius tendon transfer outcomes may not differ significantly between patients with and without prior rotator cuff surgery. Secondary lower trapezius tendon transfer following failed rotator cuff repair still represents an effective surgical option for managing PSIRCTs, yielding meaningful improvements in pain, function, and radiologic outcomes.²⁵ Nonetheless, the current study's findings indicate that a failed rotator cuff repair can modestly affect certain outcomes of posterior LD transfer, including tendon healing, postoperative strength, and patient-reported function.

The findings of the present study partially align with those reported by Kany et al.,¹⁷ who compared isolated LD transfer performed for primary massive rotator cuff tears and for failed rotator cuff repairs. Similar to the present results, their cohort demonstrated superior functional outcomes in patients with primary massive tears, including higher ASES, ADLER, and subjective shoulder value scores, and lower VAS pain scores, whereas Constant scores and ROM were comparable between groups. Complication rates, including re-tears (8% vs. 10%) and conversions to reverse shoulder arthroplasty (3% in both groups), were also similar. Consistent with these findings, the Primary group in the current study showed superior postoperative PROMs, with significantly higher SANE, Constant, and ASES scores than the Secondary group. However, unlike the results reported by Kany et al.,¹⁷ the Primary group in this study also demonstrated a trend toward greater ROM improvements, while the Secondary group exhibited significantly higher re-tear rates. The discrepancy in re-tear rates between the two studies for the Secondary group (10% vs 31.2%) may be attributed to differences in patient cohorts, fixation techniques, and imaging modalities used to evaluate postoperative tendon integrity. Kany et al.¹⁷ assessed tendon integrity using radiographs with metallic markers, whereas postoperative MRI was employed at the final follow-up in the present study, allowing for a more detailed evaluation of soft-tissue healing. Moreover, Kany et al.¹⁷ included both all-arthroscopic and arthroscopically assisted procedures, while the current study exclusively utilized an arthroscopically assisted approach. Their fixation method involved transosseous fixation to the humeral head at either the supraspinatus or infraspinatus footprint, whereas the technique applied in the present study used an onlay suture-bridge configuration with suture anchors at the supraspinatus footprint. These methodological and technical differences likely contributed to the variations observed in tendon healing and re-tear rates. Further comparative studies are warranted to clarify how differences in fixation methods, tendon combinations, and imaging modalities influence tendon integrity and clinical outcomes after transfer.

Several limitations should be noted. The retrospective, single-center design may introduce selection bias, and the relatively small number of patients in the Secondary group limits statistical power. While the follow-up duration was sufficient to evaluate short-to mid-term outcomes, long-term graft durability, structural integrity, and progression of glenohumeral arthritis require further investigation. The absence of randomization and the lack of a control group limit direct comparison with other surgical techniques or treatment modalities. Additionally, radiologic evaluation was based solely on MRI at the final follow-up without dynamic or serial assessments, which may not fully capture the temporal changes in tendon healing. Functional outcomes were evaluated using patient-reported measures and clinical scores, which, despite their validity, may be subject to response bias. Lastly, the findings may not be generalizable to other centers or surgeons due to differences in technical expertise and postoperative rehabilitation protocols.

10 Conclusion

Posterior LD transfer provides significant improvements in pain relief, function, and shoulder motion in patients with PSIRCTs. Although both groups demonstrated overall improvements in PROM, ROM, and radiologic findings, the re-tear rate was higher in the Secondary group. Furthermore, the Primary group achieved superior postoperative functional scores and greater muscle strength improvement compared with the Secondary group.

Guardian/patient's consent

Regarding the manuscript “Middle Trapezius Tendon Transfer for Isolated Irreparable Supraspinatus Tears Shows Favorable Outcomes Despite Subscapularis Tear”as submitted to the Journal of Orthopedics, collectively declares that the requirement for informed consent was waived owing to the retrospective design of the study and the lack of additional harm to the patients.

Credit author statement

Chang Hee Baek: conceptualization, investigation, methodology, validation, project administration, resources, supervision, validation.

Bo Taek Kim: conceptualization, data curation, formal analysis, investigation, validation, methodology, validation, writing original draft, review/editing final draft.

Jung Gon Kim: conceptualization, formal analysis, investigation, methodology, validation.

Chaemoon Lim: conceptualization, formal analysis, investigation, methodology, validation.

Seung Jin Kim: data curation, formal analysis, investigation, methodology.

IRB information

Approved by the Institutional Review Board (No. P01-2025-01-004)

Ethical statement

This study was conducted in accordance with the Code of Ethics of the World Medical.

Association (Declaration of Helsinki).

Funding source

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.