1 Introduction

Rotator cuff tears are among the most common causes of shoulder pain and functional impairment, with prevalence increasing with age.¹ Massive rotator cuff tears (MRCTs), classically defined as tears greater than 5 cm or involving two or more tendons, account for approximately 10–40% of all rotator cuff tears.^2,3 These tears are characterized by significant pain, weakness, restricted motion, and marked deterioration in activities of daily living.⁴

The treatment of MRCTs represents one of the most challenging areas in orthopedic surgery. The spectrum of treatment options is broad, ranging from conservative methods such as physical therapy to various surgical procedures including arthroscopic debridement, partial repair, primary anatomic repair, tendon transfers, superior capsular reconstruction, and reverse shoulder arthroplasty (RSA).^5,6 Treatment decisions depend on numerous factors including tear characteristics, muscle quality (degree of fatty infiltration), patient age, functional expectations, and glenohumeral joint status.⁷

Over the past two decades, both arthroscopic rotator cuff repair (RCR) techniques and RSA implant designs have undergone significant advances.^8,9 Arthroscopic RCR, particularly with the development of double-row and transosseous-equivalent techniques, offers the potential for biological repair even in massive tears.¹⁰ Successful repair restores the natural biomechanics of the shoulder, providing potential for superior range of motion and functional performance.¹¹ However, healing rates in MRCTs are variable, with retear rates ranging from 25% to 94%.¹² Factors such as fatty infiltration, tendon retraction, and advanced age negatively affect healing potential.¹³

Conversely, RSA has seen increasing utilization in MRCT treatment over the past 15 years.^14,15 While originally designed for irreparable tears or cuff tear arthropathy, its indications have expanded to include selected MRCT cases without glenohumeral arthritis but with high-grade fatty infiltration and chronic pseudoparalysis.^16,17 The biomechanics of RSA optimize the deltoid lever arm by medializing and distalizing the center of rotation, enabling reliable overhead elevation even when the rotator cuff is non-functional.¹⁸ Although RSA provides consistent pain relief and reliable functional improvement, range of motion, particularly internal rotation, is more limited compared to RCR.^19,20

The literature consistently demonstrates that secondary RSA following failed RCR yields inferior outcomes compared to primary RSA.²¹ These findings emphasize the critical importance of optimal primary treatment selection. However, studies directly comparing primary RCR with primary RSA in MRCT patients without glenohumeral arthritis remain limited.^22,23 Existing comparative studies typically focus on traditional functional scores (ASES, Constant, DASH) and range of motion measurements.^24,25 Patient-priority outcome domains (pain-free sleep, overhead activity, internal rotation tasks, heavy work capacity, return to sports), which evaluate performance in specific functional tasks encountered in daily life, have not been sufficiently investigated.²⁶

Given this knowledge gap, the purpose of this retrospective cohort study is to compare traditional functional outcomes, patient-priority outcome domains, range of motion, and complication rates between primary arthroscopic RCR and primary RSA in patients with massive rotator cuff tears without glenohumeral arthritis. We hypothesized that the RCR group would demonstrate superior outcomes particularly in range of motion and high-demand functional activities, while both groups would provide significant pain relief and patient satisfaction.

2 Materials and methods

2.3

2.3 Preoperative evaluation

All patients underwent detailed clinical and radiological evaluation preoperatively. Clinical assessment included pain level (Visual Analog Scale [VAS],0–10), active and passive range of motion (flexion, abduction, internal rotation, external rotation), muscle strength testing, and functional scores (ASES, Constant, DASH). Pseudoparalysis was defined as active anterior elevation less than 90° with full passive motion.²⁷

Radiological evaluation included standard anteroposterior, axillary, and scapular-Y radiographs. Glenohumeral joint degeneration was staged using the Hamada classification.²⁸ MRI images were evaluated by two independent radiologists for tear size, tendon retraction (Patte classification), and degree of fatty infiltration in rotator cuff muscles (Goutallier classification).^28,29 (Fig. 1) Acromiohumeral distance (AHD) was measured on anteroposterior radiographs.

Close

Fig. 1 MRI appearance of a rotator cuff tear. (a) Tendon retraction of the torn rotator cuff. (b) Fatty infiltration of the rotator cuff musculature.

Export to PPT

2.4

2.4 Surgical technique

Arthroscopic Rotator Cuff Repair (RCR): All procedures were performed by a single experienced upper extremity surgeon. Standard posterior, anterior, and lateral portals were established. Following diagnostic arthroscopy and systematic evaluation, limited acromioplasty, debridement, and adhesiolysis were performed as needed. Long head biceps tendon underwent tenotomy or tenodesis. Repair utilized margin-convergence, double-row, or transosseous-equivalent suture-bridge techniques based on tear configuration (Fig. 2).

Close

Fig. 2 Arthroscopic view of rotator cuff repair.

Export to PPT

Reverse Shoulder Arthroplasty (RSA): RSA procedures were performed by the same surgeon via deltopectoral approach. A 36 mm baseplate was fixed with central and peripheral compression screws, with a 36 mm glenosphere. The humeral canal was prepared and a press-fit or cemented stem implanted. Appropriate polyethylene insert (typically 39–41 mm, +3 or +6 offset) was selected. The subscapularis was repaired when possible (Fig. 3).

Close

Fig. 3 Postoperative anteroposterior radiograph of the shoulder following reverse shoulder arthroplasty.

Export to PPT

3 Results

3.2

3.2 Functional outcomes and range of motion

Both groups achieved significant improvements in all functional scores and pain levels. At final follow-up, RCR demonstrated higher ASES scores (79.2 ± 23.4 vs. 69.8 ± 26.1, p = 0.042), lower DASH scores (22.3 ± 16.8 vs. 31.8 ± 20.4, p = 0.023), and superior range of motion. Active flexion was significantly better in RCR (146.3° ± 24.8° vs. 117.2° ± 27.3°, p < 0.001), as was internal rotation (L3 vs. S1 vertebral level, p = 0.001). VAS pain improved comparably in both groups (Table 3,Fig. 4).

Table 3 Clinical outcomes at final follow-up.

Outcome Measure	RCR Group (n = 36)	RSA Group (n = 34)	P value
Functional Scores
VAS Pain (0-10)	1.6 ± 1.9	2.3 ± 2.2	0.128
ASES (0–100)	79.2 ± 23.4	69.8 ± 26.1	0.042∗
Constant (0–100)	68.5 ± 18.7	64.1 ± 19.3	0.276
DASH (0–100)	22.3 ± 16.8	31.8 ± 20.4	0.023∗
Range of Motion
Active flexion (°)	146.3 ± 24.8	117.2 ± 27.3	<0.001∗
Active abduction (°)	141.8 ± 26.2	114.6 ± 29.1	<0.001∗
External rotation (°)	48.6 ± 15.3	42.3 ± 18.7	0.094
Internal rotation (level)	L3 (L2-L4)	S1 (L5-S2)	0.001∗
Patient-Priority Outcomes (0–4)
Pain-free sleep	3.4 ± 0.8	3.2 ± 0.9	0.288
Overhead activity	3.2 ± 0.8	2.6 ± 0.9	0.002∗
Internal rotation tasks	3.3 ± 0.7	2.2 ± 1.0	<0.001∗

Close

Fig. 4 Comprehensive range of motion analysis. (A) Active flexion, abduction, and external rotation with standard deviation. (B) Internal rotation distribution by vertebral level. (C) Flexion recovery from preoperative assessment to final follow-up. (D) Abduction recovery over time. (E) Proportion of patients reaching functional ROM thresholds (≥120° and ≥150° flexion, ≥120° abduction, ≥40° external rotation). (F) Overall ROM gain from baseline. Green dashed line indicates normal reference values. RCR shows superior ROM across all parameters (∗∗∗p < 0.001).

Export to PPT

3.4

3.4 Complications and revisions

Overall complication rates were comparable (RCR: 11.1%, RSA: 8.8%, p = 0.734). In the RCR group, 44.4% showed retears on imaging, but only 11.1% required revision to RSA due to persistent symptoms. RSA complications included one dislocation (2.9%), one aseptic loosening (2.9%), and minimal scapular notching (2.9%). Patient satisfaction was high in both groups (RCR: 86.1%, RSA: 91.2%, p = 0.481) (Figs. 5 and 6).

Close

Fig. 5 Integrated clinical outcomes dashboard. (A) Patient-priority outcome domains (0–4 Likert scale), with RCR showing better overhead activity, internal rotation, heavy work, and return to sport. (B) Comparison of complication and adverse event rates (retear/loosening, revision, stiffness, instability). (C) Prognostic factors for RCR success (ASES >80) presented as odds ratios (95% CI); favorable factors include age <65 years, Goutallier ≤2, no pseudoparalysis, Patte ≤2, and acromiohumeral distance >6 mm. (D) Patient distribution by age and fatty infiltration grade. (E) Time to return to different activity levels. (F) Overall success profile (radar plot) comparing pain, ROM, function, satisfaction, and return to activities. OR, odds ratio; CI, confidence interval. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Export to PPT

Close

Fig. 6 Correlation matrix of clinical outcomes and patient factors. Heat map of Pearson correlation coefficients between functional scores (ASES, Constant, DASH, VAS), range of motion, and patient variables (age, Goutallier grade, Patte stage, follow-up duration). Colors range from −1 (strong negative, blue) to +1 (strong positive, red). Strong positive correlations are seen between ASES and Constant (r = 0.78), with inverse relationships between DASH and functional measures. Lower Goutallier grade (≤2), younger age (<65 years), and absence of preoperative pseudoparalysis are independent predictors of functional success (ASES >80) in the RCR group.

Export to PPT

4 Discussion

This retrospective cohort study compared primary arthroscopic rotator cuff repair with primary reverse shoulder arthroplasty in patients with massive rotator cuff tears without glenohumeral arthritis. Our findings demonstrate that both procedures provide significant pain relief and functional improvement, but serve different patient populations and achieve different clinical outcomes. RCR, particularly in younger, active patients with adequate tissue quality, provides superior range of motion and patient-centered functional outcomes, while RSA emerges as a reliable treatment option for elderly patients with high-grade fatty infiltration and chronic pseudoparalysis.

Our results clearly demonstrate that RCR and RSA patient selection is based on different clinical profiles. The RSA group was on average 7.3 years older, had longer symptom duration, and showed markedly higher pseudoparalysis incidence. These findings are consistent with other comparative studies in the literature.^30–32 Age is one of the most important factors affecting tendon healing potential, as collagen quality, vascularization, and osteogenic potential decline with advancing age.³³ Preoperative imaging findings highlight the pathological differences between the two groups. The RSA group had 94.4% high-grade fatty infiltration (≥2 tendons Goutallier ≥3), 100% advanced tendon retraction, and significantly shorter acromiohumeral distance. Fatty infiltration is the strongest prognostic factor indicating rotator cuff pathology chronicity and repair failure risk.³⁴ Goutallier Grade 3–4 fatty infiltration is known to be irreversible, and muscle function remains limited even after successful repair.³⁵ Therefore, selecting RSA for patients with this degree of fatty infiltration is based on biological rationale (Fig. 7).

Close

Fig. 7 Clinical decision algorithm for choosing rotator cuff repair (RCR) versus reverse shoulder arthroplasty (RSA) in massive rotator cuff tears. The flowchart integrates glenohumeral arthritis (Hamada grade), age and activity level, tissue quality (Goutallier, Patte), pseudoparalysis, and acromiohumeral distance (AHD ≥6 mm favoring RCR). Gray zones indicate cases requiring shared decision-making. Abbreviations: RCR, rotator cuff repair; RSA, reverse shoulder arthroplasty; AHD, acromiohumeral distance; ROM, range of motion; IR, internal rotation; ER, external rotation; US, ultrasound; MRI, magnetic resonance imaging.

Export to PPT

Traditional functional scores revealed that both groups showed significant improvement compared to their preoperative status. Pain reduction was highly effective in both procedures, with VAS scores decreasing by 68.6% in the RCR group and 58.9% in the RSA group. This finding is consistent with the literature consensus that both RCR and RSA are primarily pain-relieving procedures.³⁶ ASES, Constant, and DASH score improvements were also similar in both groups, indicating that both approaches provide symptomatic benefit. However, in between-group comparisons, RCR offered distinct advantages. Final ASES score was 9.4 points higher in the RCR group (79.2 vs. 69.8, p = 0.042). DASH score also favored RCR (22.3 vs. 31.8, p = 0.023). Although these differences may be below the minimal clinically important difference (MCID) thresholds, differences in range of motion and patient-priority outcomes are clinically much more significant.

One of our most striking findings is the clear superiority of the RCR group in active flexion (146.3° vs. 117.2°, difference 29.1°) and internal rotation (L3 vs. S1 level). These findings are consistent with numerous comparative studies.⁹ This superiority of RCR stems from anatomic repair capacity to restore the natural biomechanics of the shoulder. Successful repair re-establishes the rotator cuff force couple, working synergistically with the deltoid to provide full motion potential in superior, anterior, and posterior directions.³⁷ Conversely, the non-anatomic biomechanics of RSA, due to the medialized and distalized center of rotation, impose limitations especially for extreme ranges.³⁸ In RSA, the deltoid muscle becomes the primary elevator, but without the rotator cuff, fine motor control and internal rotation capacity are limited. Internal rotation limitation is the most well-known functional disadvantage of RSA.³⁹ In our study, RSA patients achieved an average S1 level while RCR patients reached L3 level. This difference is critically important for basic daily activities such as toilet hygiene, reaching behind the back, and using the back pocket. The literature reports that 20–40% of RSA patients experience difficulty in such activities due to internal rotation insufficiency.⁴⁰

Assessment of patient-priority outcome domains revealed specific functional limitations that traditional scores cannot capture. While both groups achieved similar and high scores in pain-free sleep ability (3.4 vs. 3.2, p = 0.288), confirming the effectiveness of both procedures in pain relief, RCR showed clear superiority in overhead activity, internal rotation tasks, heavy work capacity, and return to sports. The effect size was “large” (Cohen d = 1.26, p < 0.001) particularly for internal rotation tasks. While 61.8% of RSA group patients experienced moderate-severe difficulty in internal rotation tasks, this rate was only 11.1% in the RCR group. This finding demonstrates the concrete impact of RSA known internal rotation limitation on daily life. RCR was also advantageous in return to sports and heavy work capacity. Complete strength restoration is a challenging goal with both methods for massive rotator cuff tears, but RCR offers a higher strength ceiling by restoring anatomy. Lifetime lifting restrictions (typically 5 kg limit) applied to RSA patients are necessary to preserve implant longevity but limit functional independence.⁴¹

The clinical implication of these findings is significant: in younger, active patients, especially where physical work or sports are important, RCR should be the priority choice if tissue quality is appropriate. RSA provides reliable pain relief and basic function restoration but may not be the best option for high-demand activities. Total complication rates were low and similar in both groups (RCR: 11.1%, RSA: 8.8%, p = 0.734). The most common “complication” in the RCR group was retears, seen at a 44.4% rate. This rate is in the middle of the 25–94% range reported in the literature for massive tears.⁴² Importantly, more than half of patients with detected retears were asymptomatic or minimally symptomatic. Only 11.1% of patients required conversion to RSA due to persistent symptoms. This finding is consistent with literature showing that despite the absence of biological healing, a successful repair can still provide biomechanical stability and pain relief.⁴³ RSA complications were more heterogeneous: one dislocation, one aseptic loosening, and one minimal scapular notching case. Our dislocation rate (2.9%) is consistent with the 1–7% range reported for RSA in the literature.⁴⁴ Subscapularis repair integrity is critical in preventing instability.⁴⁵ Our aseptic loosening case occurred in the early period and may be related to implant positioning error. Our scapular notching incidence is very low, reflecting that modern glenosphere designs and proper positioning reduce notching risk.⁴⁶

A critical point is that secondary RSA following failed RCR yields inferior outcomes compared to primary RSA. In Welch et al. meta-analysis, RSA after prior RCR showed 8.3 points lower ASES, 6.7° less flexion, and higher pain scores compared to primary RSA.²¹ This finding emphasizes the critical importance of initial surgical choice: an RCR attempt with high failure risk can also negatively affect subsequent RSA outcomes. Therefore, in the presence of poor prognostic factors such as high-grade fatty infiltration, advanced age, and chronic pseudoparalysis, primary RSA may be a more prudent choice. Our findings and literature together suggest an algorithm for treatment selection in massive rotator cuff tears: RCR is ideal for patients aged <65–70 years, with high functional demand, Goutallier Grade ≤2, Patte retraction Stage ≤2, acromiohumeral distance >6 mm, Hamada Grade ≤2, and acute-subacute pseudoparalysis (<6 months). RSA is ideal for patients aged >70 years, with low-moderate functional demand, Goutallier Grade ≥3 (≥2 tendon involvement), Patte retraction Stage 3–4, acromiohumeral distance <6 mm, Hamada Grade ≥3, chronic pseudoparalysis (>6 months), and presence of glenohumeral arthritis. For the “gray zone” (age 65–70 years, mixed Grade 2–3 fatty infiltration, moderate functional demand), preoperative counseling is critical; advantages-disadvantages of both options should be discussed in detail, patient expectations should be set realistically, and shared decision-making should be emphasized.

Our study has several limitations. The retrospective design carries inherent selection bias and confounding factor risks. Groups were not randomized, and treatment selection was based on surgeon judgment and patient preference. The RSA group being older and having more advanced pathology reflects real-world clinical practice but may affect outcome comparisons. Follow-up duration (mean 24.3 months) is relatively short for long-term RSA outcomes; longer-term studies are needed for implant survival and late complications. Retear assessment was performed using ultrasound and/or MRI, but not all patients received MRI; this may have caused underestimation of retears in some cases. Patient-priority outcome domains, while reflecting functional performance, were measured using a simple Likert scale; more comprehensive patient-reported outcome measures could provide additional information. Finally, being a single-center study conducted by a single surgeon may limit generalizability; multicenter studies with different surgeons are needed. Despite these limitations, our study's strengths include comparative evaluation of two fundamentally different treatment approaches, comprehensive assessment of patient-priority outcome domains, adequate sample size and statistical power, and use of validated functional outcome measures.

5 Conclusion

In conclusion, primary arthroscopic rotator cuff repair in younger, active patients with massive rotator cuff tears provides better range of motion and patient-priority functional outcomes when fatty infiltration grade is appropriate. In advanced age, high-grade fatty infiltration, and presence of chronic pseudoparalysis, reverse shoulder arthroplasty emerges as a reliable treatment option. Careful evaluation of these factors in patient selection is critically important for optimal outcomes. Surgeons should comprehensively evaluate patient age, activity level, tissue quality, and functional expectations when making surgical decisions and conduct detailed preoperative counseling with patients. Future prospective randomized controlled trials and long-term follow-up studies will contribute to our understanding of the optimal treatment algorithm.

Patient consent

Not required due to the retrospective design and full anonymization of all patient data.

CRediT authorship contribution statement

Fatih Emre Topsakal: Conceptualization; Formal analysis; Data curation; Software; Investigation; Project administration; Writing – original draft; Writing – review & editing; Visualization; Validation.

Ekrem Özdemir: Methodology; Formal analysis; Data curation; Investigation; Supervision; Writing – original draft; Writing – review & editing; Visualization; Validation.

Nasuhi Altay: Writing – review & editing; Writing – original draft; Visualization; Supervision; Software; Investigation; Conceptualization.

Ethics approval and consent to participate

The experimental procedures were all in accordance with the guideline of the Ethics Committee of Erzurum Medical Faculty Scientific Research for Clinical Research and has approved by the Ethics Committee of Erzurum Medical Faculty Scientific Research for Clinical Research (Approval No: 2025/12–307, Date: 10/12/2025). This study complies with the Declaration of Helsinki. A signed written informed consent was obtained from each patient.

Consent for publication

Not applicable.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.