2.1 Final research question

In adults with degenerative rotator cuff tears, how does conservative treatment compare to surgical repair in terms of long-term clinical outcomes (e.g., Constant-Murley Score, VAS for pain), and how is it connected to cost-effectiveness? This is clearly indicated in Table 1.

2.2 Secondary research question

How novel therapies such as extracorporeal shockwave therapy (ESWT) and platelet-rich plasma (PRP) have changed the course of conservative treatment?

The eligibility criteria for the studies are listed in Table 2.

2.3 Inclusion exclusion criteria

2.4 Source information and search strategy

The literature search strategy was developed in accordance with the PRISMA 2020 reporting guidelines to ensure a comprehensive and reproducible identification of relevant studies. A systematic search was conducted across three major databases: PubMed, Embase, Scopus, and the Cochrane Library, covering studies published from January 1, 2010, to June 30, 2025. The search aimed to identify peer-reviewed articles focusing on conservative (non-surgical) treatments for degenerative rotator cuff tears, with an emphasis on long-term clinical outcomes and cost-effectiveness. Controlled vocabulary and free-text keywords were used in combination to maximize sensitivity and precision. Boolean operators (AND, OR) were applied as appropriate. The search strategy was designed collaboratively and refined iteratively based on preliminary results to reduce noise and enhance relevance as listed in Table 3.

2.5 Study Selection Process

The study selection process adhered to PRISMA 2020 guidelines. A total of 332 records were retrieved through systematic searches across PubMed (n = 73), Cochrane Library (n = 26), and Embase (n = 14), and Scopus (n = 119). After removing 19 duplicates using Zotero, 213 titles and abstracts were screened. 121 irrelevant or ineligible studies were excluded. Of the 92 full-text articles assessed for eligibility, 76 were excluded for reasons including: not degenerative rotator cuff tears related, no outcomes of interest reported, case reports, editorials, conference abstracts without full text, non-comparative studies, reviews and meta-analyses. Ultimately, 12 ^14–25 studies (all RCTs or cohort studies) were included in the final analysis, Fig. 1. Although Ranebo et al. (15) evaluated acute traumatic tears rather than purely degenerative lesions, the study was included for its long-term comparative data on surgical vs. conservative strategies. Findings were interpreted with caution. The studies by Moosmayer et al. ^17–20 and Kukkonen et al. ^21–23 represent a single study each with multiple follow-ups and were considered as one. All these studies are included in Table 4. For the evaluation of novel therapies 4 studies (3 RCT and 1 cohort) were identified ^26–29. These are included in Table 6.

Close

Fig. 1 Prisma flow diagram for the study selection process.

Export to PPT

2.6 Data extraction

Data extraction was performed by two reviewers using a standardized, pre-piloted data extraction form to ensure consistency and reduce bias. For each study included in the review, the following information was extracted: Study ID (first author, year of publication), study design, sample size, intervention, comparator, outcomes measured, key findings and follow up duration. Clinical outcome data were extracted for both groups, including the Constant-Murley Score CMS and VAS for pain (see Table 4).

2.7 Quality assessment

In this systematic review, risk of bias was assessed using tools tailored to the study design. Randomized controlled trials (RCTs) were evaluated with the Cochrane Risk of Bias tool version 2 (RoB 2), which is the current standard for assessing methodological quality in randomized trials. RoB 2 specifically addresses domains relevant to randomization, deviations from intended interventions, completeness of outcome data, accuracy of outcome measurement, and selective reporting. Its structured framework allows transparent and reproducible judgments about the internal validity of RCT evidence.

For non-randomized studies of interventions (NRSIs), we applied the ROBINS-I tool. Unlike RoB 2, ROBINS-I is designed to capture additional threats to validity inherent in observational designs, particularly bias due to confounding and selection of participants. These domains are critical when evaluating comparative cohort studies, where treatment allocation is not randomized and may be influenced by prognostic factors. ROBINS-I also evaluates classification of interventions, deviations from intended interventions, missing data, outcome measurement, and selective reporting, aligning conceptually with the domains of RoB 2 but adapted for the observational context (see Table 5 and Fig. 2), as not all studies selected for the review were randomized controlled trials. This tool evaluates five key domains: bias arising from the randomization process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in measurement of the outcome, and bias in selection of the reported result. Each study was independently assessed by two reviewers, and any discrepancies were resolved through discussion and consensus. The overall risk of bias judgment was assigned as “low risk,” “some concerns,” or “high risk” based on the domain-level evaluations. This standardized approach ensured a consistent and transparent appraisal of methodological quality across all included studies.

Close

Fig. 2 Quality assessment using RoB2 & RO.

Export to PPT

2.8 Data synthesis

Data synthesis was conducted using a narrative approach due to clinical and methodological heterogeneity among the included studies. Continuous outcome measures, including the CMS and VAS for pain, were summarized descriptively, with mean values and ranges reported where available. Statistical pooling was not performed because of variability in study designs, interventions, outcome measures, and insufficient reporting across studies. Instead, key findings were synthesized thematically to highlight patterns, similarities, and differences.

3.1 Quality assessment

The dual approach of quality assessment depending on the type of study ensured each study was judged with the most appropriate validated framework, enhancing the transparency and reliability of our synthesis (see Table 5 and Fig. 2). The assessment revealed that all RCTs show a consistent pattern for randomization and adherence to protocol with some concerns for outcome measurement. In the non-randomized study, Ramme et al. showed some mild confounding bias. By contrast, Kijima et al. carries serious risk of bias, driven by confounding and participant selection, with additional concerns for deviations, missing data, and outcome measurements, making it not high risk for driven pool data.

For the novel therapies included an appropriate risk of bias assessment was done (Table 7, Fig. 3). Overall, the two RCTs (Kim 2019; Alian 2020) show high scientific rigor yielding some concerns for outcome measurement. Moretti et al. low risk in randomization, deviations, and missing data, with some concerns for outcome measurement, yielding an overall “some concerns” judgment. Among the observational studies, Moretti 2023 shows serious confounding bias so its findings should be interpreted cautiously and not drive pooled effects. Kuo 2024 is moderate risk, mainly due to residual confounding/selection bias.

Close

Fig. 3 Quality assessment using RoB2 (PRP + ESWT).

Export to PPT

3.2 Results

In adults with degenerative rotator cuff tears, the choice between conservative treatment and surgical repair requires careful consideration of long-term clinical outcomes and cost-effectiveness. Degenerative rotator cuff disease (RCD), encompassing a spectrum from tendinopathy to full-thickness tears, is a leading cause of shoulder pain and functional impairment in adults. Current evidence suggests that the decision to operate or manage conservatively should depend on tear characteristics, symptom persistence, and patient-specific factors. Non-surgical management, typically involving physiotherapy, is often the first-line approach, particularly in cases without full-thickness tendon tears.

One of the most consistently used outcome measures in evaluating rotator cuff interventions is the Constant-Murley Score (CMS), which assesses shoulder function through pain, daily activities, strength, and range of motion. Several studies provide insight into the comparative benefits of both treatment pathways. According to Cederqvist et al. (16) which performed a pragmatic RCT with 187 patients, both surgical and non-surgical interventions improve pain and function significantly, with non-surgical treatment yielding comparable results to surgery for partial-thickness or tendinopathic presentations. They reported that there was no significant difference in CMS between surgical and non-surgical treatments in patients without full-thickness tears. Similarly, the reduction in pain (Visual Analogue Scale) showed no difference (pain: mean difference 4, p = 0.25). However, among those with full-thickness tears, surgical treatment resulted in a significantly greater CMS improvement (mean increase of 20.0 versus 13.0), and pain reduction (greater VAS reduction by 13 points, p = 0.002) suggesting better functional restoration through surgery. At the 2-year follow-up, overall changes across most domains of the RAND-36 (physical function, general health, vitality, role limitations—physical and emotional, social functioning, and bodily pain) were not significantly different between the surgery and non-surgery groups. The only difference was bodily pain for the subset of patients with full-thickness RCT, the surgery group experienced a 13-point greater increase (95% CI: 3 to 23; p = 0.011) compared to the non-surgery group. This reinforces the recommendation of initial conservative management in the absence of complete tendon rupture. However, for full-thickness tears, outcomes increasingly favor surgical repair over time, particularly regarding sustained shoulder function, pain reduction, and patient satisfaction.¹⁶

Heerspink et al.¹⁴ found no significant difference in CMS at 12 months in a general cohort of degenerative full-thickness tear patients (81.9 for surgery vs. 73.7 for conservative, p = 0.08), but a subgroup analysis showed that patients without re-tears following surgery had markedly better outcomes (CMS 88.5), indicating that surgical success depends on tendon integrity post-repair. However, they questioned the clinical significance of the observed 10.1-point CMS improvement favoring surgery, as it did not exceed the minimum clinically important difference (MCID) of 10.4 points.

Longitudinal data from Moosmayer et al. ^17–20 provide further clarity, particularly in small and medium-sized tears. He performed a RCT with 103 patients allocated in two groups (primary surgical repair vs physiotherapy). At 1-year, surgical repair demonstrated a 13-point CMS advantage (p = 0.002), exceeding the MCID threshold. This benefit persisted at 5 years, although the difference narrowed to 5.3 points and was potentially below the threshold for clinical importance.¹⁹ Importantly, at 10 years, the CMS difference widened to 9.6 points (80.5 vs. 71.8), and while slightly below the 10.4-point threshold, it was considered clinically meaningful due to the trend of functional stability in the surgical group versus decline in the physiotherapy group.¹⁷ At the 15-year follow-up 15/51 patients of the physiotherapy group had crossed over to secondary surgery. Results from primary tendon repair were superior by a mean difference of 11.8 points for the Constant score (p = 0.001), 13.9 points for the ASES score (p < 0.001), 1.8 cm on a 10-cm visual analog scale for pain (p < 0.001), and 16.2 and 22.4, respectively, for pain-free abduction and flexion (p = 0.04 and .001), highlighting the superior long-term outcomes of tendon repair compared to physiotherapy. Similarly, Kukkonen et al. in his multicenter study compared three treatments: physiotherapy alone, acromioplasty plus physiotherapy, and rotator cuff repair with acromioplasty and physiotherapy in patients over 55 with non-traumatic supraspinatus tears.²¹ At 1-year follow-up, Constant scores across all groups were similar (74.1, 77.2, and 77.9, respectively), reinforcing the notion that in select populations, conservative treatment may suffice, especially when the surgical burden and recovery are weighed. Patient satisfaction at one-year of follow-up was 87% in group 1, 96% in group 2 and 95% in group 3 (p = 0.14). Similarly, the groups' improvements in Constant scores at the two-year follow-up were 18.4 points for physiotherapy, 20.5 points for acromioplasty, and 22.6 points for repair (p = 0.38). No statistically significant differences were observed. In a similar vein, neither patient satisfaction nor VAS pain scores varied significantly. Nonetheless, the surgical repair group's mean tear size progression was significantly lower (4.2 mm) than that of the other groups (11.0mm; p < 0.01). The expenses of surgical procedures were much greater than those of physiotherapy alone. However, the mean tear size progression was significantly lower in the surgical repair group (4.2 mm) compared to the other groups (11.0 mm; p < 0.01). Surgical interventions incurred significantly higher costs than physiotherapy alone.²³

Ramme et al. analyzed 214 patients using propensity score matching to compare outcomes of rotator cuff repair (n = 107) with nonoperative treatment (n = 107) in patients with full-thickness tears confirmed by MRI. At a minimum of 2 years follow-up, surgically treated patients demonstrated significantly greater improvements in CMS (79.4 vs. 69.0; p < 0.05), ASES scores (89.1 vs. 76.5; p < 0.05), and patient satisfaction (92% vs. 70%). Effect sizes were moderate, and regression analysis showed better outcomes were independently associated with surgical repair, younger age, and shorter symptom duration.²⁴

In contrast, Ranebo et al.¹⁵ focused on small, acute traumatic supraspinatus tears in a randomized controlled trial of 58 patients (mean age 59.7 years), comparing surgical repair (n = 32) to physiotherapy without repair (n = 26). At 12-month follow-up, no statistically significant differences were found in the Constant-Murley score (median 83 vs. 78; p = 0.68), WORC index (median 91% vs. 86%; p = 0.62), pain scores, or quality-of-life measures. However, imaging at follow-up revealed that 29.2% of nonoperatively treated patients experienced a tear progression of more than 5 mm, while only 6.5% in the surgical group had a retear.¹⁵ The repair group had the 12-month median Western Ontario Rotator Cuff Index (WORC0) of 91% (QR, 24) and the physiotherapy group 86% (QR, 24),a nonsignificant between-group difference of 5.0 (4 to 9,95% CI; P ¼ .62)

Besides CMS and VAS, other outcome measures including American Shoulder and Elbow Surgeons (ASES) score and patient satisfaction also provides additional insight to long-term benefits of surgery. Moosmayer et al. ^17–20 found always a significantly higher ASES in the surgical group at all time points: 16.1 points at 1 year, 9.0 points at 5 years as well as 15.7 points at 10 years with p < 0.001 (17, 19, 20). Such results were reflected in patient satisfaction where a 1-year surgical treatment was associated with higher scores (9.0 vs. 7.2) compared to 5 years (9.2 vs. 8.3) and 10 years (9.2 vs. 8.2) on a 10-point scale. The metrics measure a long-lived patient-reported advantage of surgery, which is especially pertinent where the treatments choice is not only based on objective functioning but also the subjective quality of life.

Tear progression and re-tear rates are also crucial considerations. Moosmayer et al. reported that 37% of conservatively managed tears enlarged by more than 5 mm over 5 years, which correlated with worse outcomes.¹⁷ By 10 years, 41% of tears had enlarged by 10 mm or more, and these patients had significantly lower CMS scores (Moosmayer et al., 2019). Although surgical groups experienced re-tears (e.g., 73% at 1 year in Heerspink et al., and 34% at 10 years in Moosmayer et al.), the functional outcomes remained stable. This suggests that even with structural failure, surgical repair might delay or mitigate clinical deterioration, contrasting with the progressive functional decline seen in untreated enlarging tears.

Kijima et al.²⁵ reported a 13-year follow-up after conservative treatment of rotator cuff tears. Of 100 patients (103 shoulders) eligible, 43 shoulders were evaluable at final follow-up (60 excluded). Mean pain was 25.4/30 and mean ADL 9.4/10; 88% reported no or only slight pain (none with severe or frequent night pain), 72% had no ADL limitation, and the combined score averaged 34.8/40. The authors concluded that structured conservative care can yield durable pain relief and functional maintenance—particularly in older adults; however, given the study's major limitations (no control group, no randomization, and substantial attrition), its findings should be interpreted cautiously and should not drive pooled effect estimates.

3.3 PRP and ESWT

There were few studies focusing on the results of newer therapeutical methods such as PRP and ESWT as listed in Table 6.

PRP vs Non-Operative Standard Care: In a 2019 RCT²⁸ of degenerative partial tears, a single PRP injection provided superior long-term outcomes compared to exercise alone. By 6 months the PRP group had higher Constant–Murley scores (exceeding the MCID) and reduced tendon thickness on ultrasound (indicating tendon healing), whereas the exercise group showed tendon swelling and smaller functional gains. PRP also offered earlier pain relief (significantly lower VAS at 6 weeks). These findings suggest PRP's biological effects can enhance tendon quality and function beyond what exercise therapy achieves in partial tears.

PRP vs ESWT (Head-to-Head): Two studies comparing PRP injections to shockwave therapy in partial tears found both treatments to be effective, with some divergence in durability. Alian et al.²⁷ reported that while ESWT and PRP both improved pain, ROM, and disability by 4 weeks, PRP's benefits were better sustained at 3 months. The ESWT group experienced relapse of pain and loss of initial ROM gains by 12 weeks, resulting in significantly worse outcomes than PRP (e.g. VAS ∼6.7 vs 3.3; abduction 100° vs 151°). In contrast, a larger 2023 cohort study²⁹ found no significant difference between PRP and ESWT by 6 months: both groups had comparable pain reduction and Constant score improvements (≈+20 points). That study formally concluded PRP was non-inferior to ESWT. Notably, PRP-treated tendons showed evidence of improved tendon structure on ultrasound (reduced degeneration and vascularity), aligning with PRP's role in enhancing biological healing. Across studies, PRP and ESWT each achieved clinically meaningful improvements in pain (often 50–60% VAS reductions) and function in patients with partial tears. However, PRP may confer more sustained or thorough recovery in some cases, whereas ESWT outcomes can be equivalent by 6+ months. No major complications were reported for either therapy.

Combined ESWT + PRP vs PRP Alone: A 2024 prospective cohort²⁶ directly examined adding ESWT to PRP for partial-thickness tears. Both modalities together produced faster early gains in shoulder mobility than PRP alone – at 1 month the combination group had significantly greater forward flexion and abduction and over 60% of combo patients achieved MCID in CMS (versus 30% with PRP). The combined approach also transiently lowered inflammatory biomarker levels (S100A8/A9) relative to PRP-only, indicating an enhanced early anti-inflammatory effect. By 6 months, however, pain relief and functional scores equalized between groups. Both achieved excellent outcomes (Constant ≈80, VAS ∼2) and similar proportions of patients reaching CMS ≥80 and low pain levels. In short, adding ESWT to PRP did not improve long-term pain or shoulder function beyond PRP's effect, but it did yield some additional short-term benefits in range-of-motion recovery. This combined therapy was safe and well-tolerated, with no increase in adverse events. It illustrates that mechanical stimulation (ESWT) can complement PRP's biological repair mechanisms to accelerate early improvements, even if final outcomes converge with PRP alone.

All included studies reported statistically significant improvements from baseline in pain (VAS/NRS) and shoulder function scores (Constant, DASH, SPADI or ASES) for PRP and/or ESWT interventions. Where reported, these changes often met or exceeded established MCID thresholds (e.g. ∼10 points for Constant, 1.5–2 for VAS) and a substantial subset of patients achieved PASS criteria (e.g. Constant ≥80) by final follow-up. Particularly, PRP injections tended to yield sustained functional gains and tendon healing signs, while ESWT provided marked pain relief and functional improvement that in some cases plateaued or required continued therapy.

It should be noted that these findings pertain mostly to partial-thickness degenerative tears. There is little published evidence on PRP/ESWT for full-thickness tears without surgery (likely due to the structural discontinuity in full tears). The implication for practice is that, in patients with partial degenerative cuff tears, PRP and ESWT are both viable conservative options that can significantly reduce pain and improve shoulder function. PRP may have an edge in promoting tendon tissue quality (as seen on imaging and molecular markers), whereas ESWT can provide mechanical stimulus for pain modulation and tendon remodeling.