Symptomatic changes and influencing factors following manipulation under anaesthesia in patients with adhesive capsulitis of shoulder: a protocol for a prospective cohort study.

Adhesive capsulitis, also known as frozen shoulder, is characterised by spontaneous shoulder pain and progressive restriction of both active and passive motion. 1 Multiple factors may influence its onset and progression. With an ageing population and increasingly sedentary lifestyles, its incidence has risen to 2%–5%. 2 This disease notably impacts patients’ daily lives. Many still endure recurrent pain 6 years after onset, 3 and some may suffer from permanent symptoms. 4 6 Mostly, it occurs without specific triggers and is regarded as idiopathic. Epidemiological studies reveal that it often affects middle-aged women, especially those aged 50–70. And those with metabolic disorders like diabetes and thyroid diseases have an even higher risk. 6 7 At the microscopic level, chronic low-grade inflammation plays a key role in disease progression. Hyperglycaemia can alter the collagen matrix of the joint capsule, leading to fibrosis and inflammatory changes typical of the disease. 8 9 Inflammatory mediators activate nociceptive nerve endings and promote central sensitisation, reflecting the close interaction among the nervous, endocrine and immune systems. Dysregulation of any of these systems may contribute to disease development. Beyond endocrine factors, nerve compression, especially of the C5 root innervating the shoulder girdle, may result in shoulder pain and dysfunction. Clinically, many patients also present with cervical spine disorders, such as spinal stenosis. 10 11 Chronic shoulder pain may itself induce central sensitisation, amplifying pain transmission and perception. 12 13 When accompanied by neck pain, this sensitisation can intensify discomfort and further limit shoulder function. Patients may experience pain even without obvious triggers, a phenomenon common in those with chronic pain and limited mobility. The link between central sensitisation and adhesive capsulitis is still poorly understood. In addition, psychological, lifestyle and social factors can affect disease progression and recovery. 14 Considering these factors helps formulate better prevention and treatment strategies. Treatments for adhesive capsulitis can be categorised into non-surgical and surgical approaches. Non-surgical interventions include exercise therapy, oral or injected anti-inflammatory medications and hydrodilation, which aim to alleviate pain, expand joint mobility and address adhesions. When these conservative measures fail, surgical treatment, such as manipulation under anaesthesia (MUA), may be considered. This involves passive shoulder movements under anaesthesia to relieve pain and restore motion. It can rapidly improve patients’ function; many patients experience significant pain relief and mobility gains shortly after the procedure. 15 17 Studies have demonstrated that MUA is more effective than conventional rehabilitation in broader populations, with minimal safety differences from physical therapies and better cost-effectiveness compared with arthroscopic release. 17 The UK Frozen Shoulder Trial also supports its cost-effectiveness, with no significant differences in adverse event rates, 18 and over 80% of patients showing improvement post-procedure. 19 23 Although MUA is widely used for the treatment of adhesive capsulitis, postoperative recovery varies considerably, and the specific prognostic factors underlying these differences remain unclear. Few studies have systematically evaluated the influence of demographic and socioeconomic factors on functional recovery, the impact of metabolic and cervical-spine–related comorbidities on the resolution of inflammation and pain, the predictive value of baseline shoulder range of motion for recovery or the role of baseline central sensitisation levels in shaping postoperative pain trajectories. This knowledge gap makes it challenging for clinicians to make informed decisions. To address this knowledge gap, our study aims to track the changing trajectory of patients’ prognosis and identify the factors that influence recovery patterns. We hypothesise that baseline metabolic comorbidities and central sensitisation features are associated with slower functional recovery trajectories after MUA. This study aims to observe the progression of symptoms in patients with adhesive capsulitis, evaluate the timeline of improvement following MUA and explore potential factors influencing patient prognosis.

This study has been approved by the Institutional Review Board of Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine (approval number: 2025-1707-047-02). Written informed consent will be obtained from all participants prior to study enrolment. The consent form, in its approved version (dated 20 March 2025), will be provided to participants, and a trained study investigator will explain the study purpose, procedures, risks and benefits. Participants will have the opportunity to ask questions and will sign the consent form if they agree to participate. To ensure data accuracy and confidentiality, all study materials will be securely stored at the study site. Paper files will be kept in locked cabinets with restricted access. All data will be anonymised before analysis to ensure participant privacy. The findings of this protocol will be disseminated through publication in international peer-reviewed journals and presentation at international conferences. Participants, clinicians and relevant research staff will be informed of the study results.

This single-centre prospective cohort reporting will follow the Strengthening the Reporting of Observational Studies in Epidemiology 24 guidelines for observational cohort studies. Prognostic objectives and analysis will be guided by the Prognosis Research Strategy (PROGRESS) framework Type 2. 25 Patients and the public will not be involved in the design, conduct, reporting or dissemination plans of this research. This longitudinal prospective cohort study will be conducted in inpatient departments of Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine in China. Preoperative data collection will include the following: basic demographic information (age, gender, body mass index (BMI), socioeconomic status scale), metabolic diseases (eg, diabetes and thyroid disorders), neck pain, duration of adhesive capsulitis onset, shoulder range of motion assessment (flexion, abduction, internal rotation, external rotation) and central sensitisation evaluation using the Central Sensitisation Inventory (CSI) and pressure pain threshold (PPT). PPT will be measured using a handheld electronic algometer (SBMEDIC Electronics, Algometer Type II; Solna, Sweden). Postoperative assessments will be conducted at 1, 2, 4, 6, 9 and 12 months, including shoulder function evaluation (Oxford Shoulder Score, OSS), quality of life assessment (EuroQol Five Dimensions Five Levels, EQ-5D-5L) and pain assessment (Numeric Rating Scale, NRS). Patient recruitment is scheduled to begin in May 2025. Follow-up will continue for 12 months after the last participant is enrolled, and all data collection is planned to be completed by September 2026. The study procedures are illustrated in the flowchart ( figure 1 ). The study population includes adult patients with adhesive capsulitis who are scheduled to undergo MUA as part of routine clinical care. To reduce selection bias, data will be collected consecutively from all eligible patient consultations. The diagnostic criteria follow the clinical guidelines established by the Orthopaedic Section of the American Physical Therapy Association: The diagnosis of adhesive capsulitis is based on medical history and physical examination. (1) Patients typically present with gradually progressive pain, which may worsen at night and with end-of-range movement. (2) The affected shoulder exhibits a restriction of range of motion greater than 25% in at least two planes, with passive external rotation loss greater than 50% compared with the unaffected shoulder, or the external rotation less than 30°. 1 Other conditions, such as rotator cuff tears, will be excluded based on radiographic evaluation. Participants will receive regular reminders to attend follow-up visits. Those who discontinue the intervention will be encouraged to complete all scheduled assessments. Outcome data will be collected whenever possible. Inclusion criteria: Adults aged 30 to 75 years. Diagnosed with adhesive capsulitis according to clinical diagnostic standards. Exclusion criteria: History of prior shoulder surgery. Pregnant women. Patients’ intolerance to manipulation treatment. Presence of local skin lesions or infection risk. Severe osteoporosis or shoulder joint deformities. Comorbidities that could affect study outcomes (eg, full-thickness cuff tears, cervical radiculopathy, neuropathies, prior fractures or instability). Dropout criteria: Patients can request to withdraw from the intervention at any time, without providing a reason. The patient will be placed in the supine position with the head tilted toward the contralateral side. The affected side received a brachial plexus block with 10 mL of 2% lidocaine. The surgeon stood on the same side. Under anaesthesia, the affected arm will be maintained in a forward flexed position. The shoulder will be gently brought into 180° forward flexion three times, during which tearing of the fibrotic capsule could usually be palpated or heard. Flex the affected shoulder joint to 90° in forward flexion, bend the elbow and internally rotate the shoulder to 45°, repeating the manoeuvre three times. Then, rotate the affected shoulder sequentially in the neutral position, the 45° abducted position, the 90° abducted position, the 135° abducted position and finally with the glenohumeral joint extended and the elbow straight, repeating each movement three times. Next, position the patient on the healthy side, with the affected shoulder extended and internally rotated, and the elbow flexed so that the fingers can reach the healthy scapula; repeat this manoeuvre three times (see online supplemental material, appendix 1 ). If during manipulation a hard end-feel or abnormal resistance is encountered, new neurological symptoms arise (eg, numbness or weakness), iatrogenic injury is suspected (eg, fracture or dislocation) or any other safety concern emerges, the procedure will be stopped immediately. The shoulder will then be re-evaluated, neurovascular status checked and the limb protected. If structural injury is suspected, immediate imaging (eg, bedside or standard radiographs) will be performed, and appropriate clinical management initiated. All intra-procedural adverse events will be documented in the study records. After the procedure, the affected limb is immobilised and protected. No routine regional analgesia or intra-articular injections are administered; pain is mainly managed with non-opioid analgesics, with additional injections provided only if clinically indicated and recorded. Standard post-procedure radiographs of the affected shoulder (anteroposterior and axillary or scapular-Y views) are obtained to rule out iatrogenic fracture or dislocation. Any abnormal findings prompt further evaluation and management as required. Rehabilitation exercises begin on the day of MUA and are performed daily during hospitalisation, followed by 3–5 sessions per week for the first 6 weeks and 2–3 sessions per week thereafter until 12 weeks. The programme includes passive and active-assisted range of motion exercises in flexion, abduction and internal and external rotation, as well as functional stretching, with emphasis on scapular mobilisation to support glenohumeral movement. Non-opioid analgesics or ice therapy is applied as needed to manage pain and ensure safe exercise performance. At each follow-up time point, all additional interventions—including medications, injections or physical therapy sessions—are recorded to account for their potential impact on recovery outcomes. The outcome measures will be evaluated from three dimensions: function, quality of life and pain. The primary outcome measure will be the OSS, and the secondary outcome measures will be the EQ-5D-5L and NRS. For detailed information, refer to online supplemental table 1 . Potential predictors after MUA include the following: Basic demographic and socioeconomic information (age, sex, BMI, Socioeconomic Status Scale). Medical history (diabetes, hypothyroidism, cervical spine pain, onset time of adhesive capsulitis). Range of motion. Central sensitisation (CSI, PPT). An overview of the potential prognostic factors is provided in online supplemental table 2 including a detailed description of all measurement tools and testing methods. We will use linear mixed-effects models to evaluate the influence of baseline prognostic factors (eg, age, BMI and diabetes) on longitudinal OSS. In this framework, time will be modelled as a fixed effect, while repeated measures within individuals will be accounted for by random effects. Prognostic factors and their interactions with time will be included as fixed effects, enabling the identification of variables that are associated with overall OSS levels and those that modify recovery trajectories over time. The OSS will also be divided into four levels (OSS categories: 0–19, 20–29, 30–39, 40–48). 26 Using the significant prognostic factors obtained from the linear mixed-effects models as the only exposure variables, the probability of transitioning between OSS levels from baseline to follow-up will be calculated (along with the 95% CI). Latent Growth Curve Modelling (LGCM) will be used to analyse OSS at different time points to observe the recovery trajectory of patients. The model fit will be evaluated through χ² tests and fit indices (Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), Root Mean Square Error of Approximation (RMSEA)). Maximum likelihood estimation will be used to handle missing data. Missing data will be handled using multiple imputation to account for the uncertainty associated with incomplete observations. Multiple complete datasets will be generated and analysed separately, with results subsequently pooled to provide more reliable and robust estimates. Potential confounding variables will be systematically accounted for in the analyses. Any co-interventions such as medications, injections or physical therapy will be included as fixed-effect covariates in the linear mixed-effects models and latent growth curve models. The above statistical analyses will be conducted using SPSS Statistics 27.01. This study is designed as a prognostic factor investigation, adhering to the PROGRESS. 25 According to PROGRESS, prognosis studies are classified into four interrelated types, with this study falling under Type II. This type focuses on examining the relationship between individual factors and a single outcome. In our multivariable model, we aim to include at least 10 events per variable to ensure robust statistical power. 27 28 Accounting for an anticipated dropout rate of 10%, we determined that a minimum sample size of 145 participants is required to achieve reliable results. To ensure accurate, complete and reliable data, all study-related information will be securely stored at the study site. Participant information will be kept in locked cabinets with limited access, and a coded ID number will be used on reports, data collection and administrative forms to maintain confidentiality. Given that MUA is a well-established and widely used clinical procedure with a low incidence of serious complications, and the study involves no investigational drugs or devices, a formal independent data monitoring committee has not been established. Instead, safety will be overseen by the principal investigator and the clinical research team, with periodic review by the Institutional Review Board. This approach is consistent with the low-risk profile of MUA. Adverse events (AEs) are defined as any unfavourable and unintended signs, symptoms or diseases temporally associated with the MUA procedure, regardless of whether they are considered related to the intervention, such as temporary chest infection, injury to adjacent structures or neuropathic symptoms. Serious adverse events (SAEs) are AEs that result in death, are life-threatening, require inpatient hospitalisation or prolongation of existing hospitalisation, lead to persistent or significant disability/incapacity or cause a fracture, nerve injury or shoulder joint instability attributable to MUA. 16 All AEs and SAEs will be monitored by MUA through 12 months. Participants will report new symptoms at scheduled visits or anytime between visits. Events will be documented and assessed for severity and relation to MUA by the principal investigator and senior surgeons. Non-serious AEs will be recorded for analysis; serious AEs—including death, hospitalisation, persistent disability, fracture, nerve injury or shoulder instability—will be reported to the Institutional Review Board within 24 hours and managed clinically. Patients or members of the public will not be involved in the design, conduct, reporting or dissemination plans of this research.

Adhesive capsulitis is a prevalent musculoskeletal condition characterised by pain and restricted shoulder mobility. 5 When conservative treatments fail to provide adequate relief, MUA is often employed to rapidly improve shoulder function. Extensive evidence supports the efficacy of MUA in alleviating pain and restoring function, with outcomes comparable to those of arthroscopic capsular release, while also demonstrating favourable cost-effectiveness. 16 29 However, a subset of patients may experience suboptimal recovery following the procedure, and identifying the specific patient characteristics that predict successful outcomes remains a challenge. This underscores the need for further research to elucidate the indications and efficacy of MUA. Prior research on prognostic factors in adhesive capsulitis recovery has highlighted the roles of pain catastrophising and anxiety, diabetes and hypothyroidism in predicting treatment outcomes. 30 32 Building on these findings, our research will investigate the impact of initial range of motion and central sensitisation on long-term functional recovery. Although our study has several strengths, it is important to recognise its limitations. First, the duration of follow-up, while adequate for evaluating short-term outcomes, may not fully reflect long-term recovery patterns. Future research should consider extending the follow-up period to better understand the long-term durability of treatment effects. Second, while our study comprehensively assessed central sensitisation, future investigations should incorporate more advanced neurophysiological evaluations to enhance the precision of the findings. Third, the use of advanced trajectory modelling techniques, such as growth mixture modelling, could help identify distinct subgroups with unique recovery trajectories in future studies, thereby supporting the development of more personalised treatment approaches.