Clin Orthop Relat Res. 2026 Jun 5. doi: 10.1097/CORR.0000000000004010. Online ahead of print.
ABSTRACT
BACKGROUND: Rotator cuff tears with concomitant preoperative shoulder stiffness present a clinical challenge because the rehabilitation needed to restore motion may conflict with the protection typically used to support tendon healing after rotator cuff repair. Although recent studies suggest that preoperative stiffness may not be associated with inferior long-term outcomes, whether it is associated with a different pattern of postoperative improvement and distinct inflammatory features in the torn tendon remains unclear.
QUESTIONS/PURPOSES: (1) Do a greater proportion of patients with preoperative shoulder stiffness achieve improvements in functional outcomes that exceed the minimum clinically important difference (MCID) at 3, 6, 12, and 24 months? (2) Do supraspinatus tendons from patients with shoulder stiffness show molecular and histologic features consistent with greater inflammation than tendons from patients without stiffness?
METHODS: This prospective, observational cohort study included patients scheduled to undergo arthroscopic rotator cuff repair for unilateral rotator cuff tears with medium-sized, full-thickness supraspinatus tears, with or without concomitant preoperative shoulder stiffness. A medium-sized tear was defined as a supraspinatus tendon tear measuring 3 cm or less on preoperative coronal MRI. Preoperative shoulder stiffness was defined as passive shoulder ROM of less than 120° in forward flexion, less than 30° in external rotation at the side, or internal rotation lower than the L3 spinal level. Patients who met any one of these three criteria were assigned to the stiffness group. Between March 2020 and December 2023, 474 patients from both urban and rural areas who were scheduled to undergo repair for rotator cuff tears were evaluated at our shoulder pain specialty clinic. The inclusion criteria were: (1) full-thickness unilateral supraspinatus tendon tears measuring 3 cm or less in the AP dimension, confirmed on T2-weighted MRI, and (2) radiographs and MRI demonstrating no other shoulder pathology, including fracture, arthritis, dislocation, or glenoid labral tear. After strict eligibility assessment, 39% (186 of 474) of patients met all inclusion criteria and were enrolled into either the stiffness group (92 patients) or the without stiffness group (94 patients) based on predefined criteria. Specifically, 28% (135 of 474) of patients were excluded because symptom duration exceeded 12 months, 1% (2 of 474) were excluded because of missing MRI data, and 4% (20 of 474) were excluded because they had received more than three glucocorticoid injections. In addition, 18% (84 of 474) of patients were excluded because of systemic comorbidities or a history of surgery on the affected shoulder: rheumatoid arthritis in 1% (2 of 474), thyroid disorders in 3% (12 of 474), diabetes in 9% (43 of 474), gout in 1% (1 of 474), osteoporosis in 1% (5 of 474), hyperlipidemia in 4% (17 of 474), previous shoulder surgery in 1% (3 of 474), and concomitant upper limb fracture in 1% (1 of 474). A preoperative MRI evaluation resulted in the exclusion of patients with labral lesions (1% [5 of 474]), calcific tendinitis (1% [3 of 474]), massive tears (4% [21 of 474]), and partial-thickness tears (4% [18 of 474]). The enrolled cohort comprised 186 patients, with a mean age of 62 ± 7 years and a mean BMI of 24.1 ± 2.8 kg/m2. Women accounted for 68% (127 of 186) of the cohort. Of the 186 enrolled patients, 4% (8 of 186) were lost to follow-up by 12 months and an additional 15% (27 of 186) by 24 months, leaving 81% (151 of 186) for the final analysis; there were 75 patients with stiffness and 76 without stiffness. Clinical outcomes, including ROM, the VAS, Oxford shoulder score (OSS), and Constant-Murley Score (CMS), were assessed preoperatively and at 3, 6, 12, and 24 months postoperatively. MRI was performed preoperatively and at 24 months. Retears were assessed using the Sugaya classification at 24-month follow-up, with Types IV and V defined as postoperative retears. Intraoperative supraspinatus tendon samples were collected and analyzed using histology and bulk RNA sequencing. To address our first study question, we evaluated postoperative recovery at two levels. First, we applied anchor-based MCID thresholds reported in previous studies: 10 points for the CMS, 6 points for the OSS, and 1.4 points for the VAS. We then compared the proportions of patients in the two groups who achieved clinically meaningful improvement at each postoperative follow-up time point (3, 6, 12, and 24 months) to determine whether patients with preoperative stiffness were more likely to achieve meaningful improvement at scheduled follow-up visits. Second, we used repeated-measures ANOVA and mixed-effects models to analyze the longitudinal trajectories of the CMS and OSS across follow-up and to evaluate whether the postoperative pattern of improvement differed between groups. To address our second study question, supraspinatus tendon tissue was collected intraoperatively for histologic and bulk RNA sequencing. Histologic differences were assessed using the revised Bonar scoring system, and transcriptomic differences were evaluated using differential gene expression and functional enrichment analyses.
RESULTS: At 3 months, a greater proportion of patients in the stiffness group achieved the MCID in the CMS (43% [32 of 75] versus 13% [10 of 76], OR 4.9 [95% CI 2.2 to 11.0]; p < 0.001), OSS (59% [44 of 75] versus 21% [16 of 76], OR 5.3 [95% CI 2.6 to 10.9]; p < 0.001), and VAS-movement (33% [25 of 75] versus 17% [13 of 76], OR 2.4 [95% CI 1.1 to 5.2]; p = 0.02). Similar patterns were observed at 6 and 12 months, with higher proportions of patients in the stiffness group achieving clinically meaningful improvement across the CMS, OSS, and VAS-movement. By 24 months, the patients with stiffness still had higher MCID achievement proportions for the CMS (100% [75 of 75] versus 83% [63 of 76], OR 32.1 [95% CI 1.9 to 550.8]; p < 0.001), OSS (99% [74 of 75] versus 88% [67 of 76], OR 9.9 [95% CI 1.2 to 80.6]; p = 0.02), and VAS-movement (83% [62 of 75] versus 68% [52 of 76], OR 2.2 [95% CI 1.0 to 4.7]; p = 0.04), whereas we found no differences between groups for VAS-rest or VAS-sleep at any follow-up timepoint. Longitudinal analyses showed different postoperative improvement patterns for the CMS and OSS between groups (group-by-time interaction: CMS, ges = 0.121; p < 0.001; OSS, ges = 0.075; p < 0.001), although these differences were no longer observed after adjustment for baseline score (CMS: β 0.1 [95% CI -0.1 to 0.2]; p = 0.63). With respect to tendon biology, patients with and without preoperative stiffness showed different molecular and histologic features. The patients with stiffness had 254 upregulated genes and 554 downregulated genes, were enriched in the inflammatory tendinopathy subtype (NES 2.6; adjusted p < 0.01), and showed more inflammatory cell infiltration and higher revised Bonar inflammation scores. By contrast, the patients without stiffness were enriched in the hypoxia-white subtype (NES -2.8; adjusted p < 0.01) and in metabolic, muscle-development, and neurodevelopment pathways. Immune deconvolution also supported a more immune-activated tendon microenvironment in the patients with stiffness.
CONCLUSION: In patients with medium-sized rotator cuff tears, preoperative shoulder stiffness was not associated with worse long-term outcomes after arthroscopic rotator cuff repair. Although patients with stiffness started with worse shoulder function, they were more likely to achieve clinically meaningful improvement during the early postoperative period, a pattern that likely reflects lower baseline function and greater room for improvement rather than superior final recovery. In addition, inflammatory tendon features from patients with shoulder stiffness reflected a distinct biologic context. For surgeons, these findings may help when counseling patients with rotator cuff tears that satisfactory recovery is achievable even with concomitant preoperative shoulder stiffness. Future work should focus on whether this inflammatory tendon feature can be validated in broader patient populations and can be used to refine biologic classification and postoperative management.
LEVEL OF EVIDENCE: Level II, therapeutic study.
PMID:42248213 | DOI:10.1097/CORR.0000000000004010