Objectives： This study aims to compare the efficacy of transcatheter arterial microembolization (TAME) and corticosteroid injection for treatment of frozen shoulder.

Background： Frozen shoulder (adhesive capsulitis) is a common condition that leads to significant functional impairment and reduced quality of life. Current treatment options are diverse, including pharmacological therapy, physical rehabilitation, corticosteroid injection, hydrodilatation, and the more recently developed transcatheter arterial microembolization (TAME). Corticosteroid injection, due to its anti-inflammatory effects, has been shown in numerous clinical studies to provide short-term pain relief and improved range of motion; however, its benefits are often limited to several weeks or months, and some patients continue to experience persistent symptoms. In contrast, TAME has emerged as a novel treatment for musculoskeletal disorders by targeting abnormal neovascularization and inflammatory processes in the affected tissues, there by achieving pain reduction. Multiple studies from both international and domestic medical centers have demonstrated its safety and efficacy, though most reports focus on case reports or case series observational outcomes. To date, there is a lack of clinical trials directly comparing the therapeutic outcomes of TAME versus corticosteroid injection in patients with frozen shoulder.

Study Design： This study is a single-center, prospective, randomized controlled trial (RCT).

Methods： A total of 94 patients diagnosed with frozen shoulder in the freeing stage will be recruited from the Department of Physical Medicine and Rehabilitation at Shin Kong Hospital. Eligible patients will be randomly assigned to either the transcatheter arterial microembolization (TAME) group and the corticosteroid injection (CS) group.

The arterial microvascular embolization group will receive one treatment, while the corticosteroid injection group will receive two corticosteroid injections into the shoulder joint and subacromial bursa, with a two-week interval between the two injections. Additionally, all patients in both groups will undergo 4 weeks of physical therapy.

Clinical assessments will be conducted at baseline and at 1, 2, 4, and 6 months after initiation of treatment, including Shoulder Pain and Disability Index (SPADI), the Visual Analogue Scale (VAS) for pain, Shoulder Disability Questionnaire (SDQ), active and passive range of motion (ROM), Short Form-36 (SF-36) for quality of life, and patient-reported global efficacy. Patients will be followed for 6 months after completion of treatment, and statistical analyses will be performed to evaluate the therapeutic outcomes.

Effect： Efficacy comparison: Both the TAME and corticosteroid injection are expected to effectively improve pain, function, and quality of life. However, the TAME group is anticipated to demonstrate greater pain relief and better recovery of shoulder function compared with the corticosteroid injection group, with longer-lasting therapeutic effects.

Safety: Both treatments are considered safe. TAME may lead to temporary local pain, transient radial artery spasm, or low-grade fever, all of which are self-limiting or manageable with supportive care. Corticosteroid injection may be associated with transient hyperglycemia, local injection site pain, or, rarely, infection. No permanent or severe adverse events are expected.

Exploration of prognostic factors: By comparing treatment responses between the two groups and incorporating clinical characteristics, imaging findings, and treatment outcomes, this study aims to identify potential prognostic factors that may predict therapeutic response, thereby guiding future clinical decision-making.

Intermediate follow-up results: At 1 to 3 months post-treatment, both groups are expected to show improvement in the clinical outcomes, but the TAME group is expected to demonstrate superior effect compared with that of the corticosteroid injection group.

Frozen shoulder, or adhesive capsulitis of the shoulder, is a common clinical problem characterized by pain, loss of both passive and active range of motion (ROM) of glenohumeral joint (GHJ), leads to functional limitation, and reduces quality of life. Primary frozen shoulder occurs without any obvious connection to any other condition, and secondary frozen shoulder is related to some medical problems, such as an injury, inflammatory, immunological, endocrine conditions, and surgery of the shoulder. The prevalence of frozen shoulder is 2%-5% in general population, usually between 40 to 60 years old, and more in female.2 Diabetes patients have more chance of developing frozen shoulder than general population and the prevalence was 13.4%.

Frozen shoulder is clinically divided into 3 overlapping phases. The painful freezing phase has a duration of 10 to 36 weeks and is characterized by pain and stiffness around the shoulder, which may worsen at night. The frozen phase is characterized by restricted ROM with a gradual relief of pain, which occurs at 4 to 12 months. The thawing phase with spontaneous improvement in the ROM takes 12 to 42 months. Although frozen shoulder is a self-limited condition, the recovery may be slow and incomplete.

Treatment of frozen shoulder consists of physical therapy (PT), oral medication (nonsteroidal anti-inflammatory drugs or corticosteroid), intra-articular corticosteroid injection, hydrodilatation, suprascapular nerve block, mobilization, manipulation under anesthesia, and operative intervention (arthroscopic release or open release). Because of the uncertainty of the efficacy and risk of surgical treatment, nonsurgical treatments are more likely chosen by patients. Among them, intra-articular steroid injection and PT are commonly used nonsurgical treatments and have shown some benefits. Because the clinical picture of frozen shoulder may be similar to, or combining with subacromial bursitis, especially in the freezing stage, concomitant subacromial/subdeltoid (SASD) bursa injection may be needed for treatment of frozen shoulder. In addition, the 3 stages of frozen shoulder often overlap and the clinical symptoms of patients are complex, adjuvant therapy is often needed throughout the course of treatment. Even after corticosteroid injection and PT, mild to moderate contracture, especially external rotation, abduction, and internal rotation may still be present.

TAME (TransArterial MicroEmbolization) has emerged in recent years as a novel treatment modality for musculoskeletal disorders, especially frozen shoulder. In patients with frozen shoulder, numerous studies have shown that TAME can significantly improve pain and functional mobility. By targeting and occluding abnormal neovascularization and associated inflammatory activity at the pathological site, TAME aims to reduce pain and inflammation. Multiple studies from both international and domestic institutions, including National Taiwan University Hospital, Tri-Service General Hospital, National Cheng Kung University Hospital, and Chung Shan Medical University Hospital have demonstrated its safety and effectiveness. Nonetheless, most published reports to date remain retrospective observational studies focusing on single-treatment cohorts.

Because the effect of TAME against frozen shoulder has not been well established, we aim to compare the effectiveness of TAME with corticosteroid injection, in addition to routine PT, for treatment of frozen shoulder? We hypothesize that TAME may be more effective in the treatment of frozen shoulder.

Method Study Design This is a prospective single-blinded, and assessor-blinded randomized controlled trial. We plan to enroll 94 (first year 60, second year 34) participants with primary frozen shoulder from the outpatient clinic of the Department of Physical Medicine and Rehabilitation and pain clinic in our hospital. The study protocol and consent form have been approved by our ethics committee. After detailed explanation about the content of this study, participant is asked to sign an informed consent. The patients will be randomly divided into two groups: 1). TAME+PT, TAME group; 2). corticosteroid injection+PT, CS group. Randomization will be assigned according to a random number from a computer program by a statistician. The patients and the evaluator are unknown which group the patients are allocated, but the clinicians who perform TAME or corticosteroid injection know which group the patients are assigned.

Intervention. TAME group. Participants allocated to the TAME group will undergo transarterial microvascular embolization following a standardized protocol. All procedures were performed by a single interventional radiologist with experience of TAME in over 500 cases. All procedures will be performed under local anesthesia. Percutaneous access to the ipsilateral radial artery will be established using a 4-F introducer sheath. Subsequently, subclavian digital subtraction angiography (DSA) will be carried out using a 4-F diagnostic catheter (JR4, Terumo, Tokyo, Japan). A total of 15 mL of iodinated contrast medium (Xenetix 350, Guerbet, Villepinte, France) will be administered at an injection rate of 3 mL/s.

Anatomically, the suprascapular artery originates from the subclavian artery, whereas the remaining four arteries and one branch supplying the target region arise from the axillary artery. Selective angiography of these vessels will be performed using a 4-F catheter (JR4, Terumo, Tokyo, Japan; or RIM, Cook Medical, Bloomington, IN, USA) positioned at each arterial orifice. A 1.98-F microcatheter (Masters Parkway Soft, Asahi Intecc, Aichi, Japan) will then be advanced distally to segments corresponding to or adjacent to the patient's symptomatic area.

Angiography will be performed via slow manual injection of 1-2 mL of contrast material. When abnormal vascular staining is identified or when the patient reports evoked pain-defined as reproducible shoulder pain or a heat-like sensation that matches their typical symptoms-embolization will be initiated.

The embolic material consists of imipenem/cilastatin sodium (IPM/CS) (Facta Farmaceutici S.p.A., Italy). Based on our previous papers, these agents exhibit limited water solubility and form crystalline particles when suspended in contrast medium, enabling transient microvascular embolization. A suspension will be prepared by mixing 0.5 g of IPM/CS with 5-10 mL of contrast medium using a dual-syringe pumping technique.

The embolization endpoint is defined as the elimination or substantial reduction of abnormal vascularity (Fig.3,4). Following infusion of the IPM/CS suspension, the disappearance or reduction of evoked pain will be recorded as an immediate procedural response.

CS injection group. The participants will receive corticosteroid (20mg triamcinolone with 3mL 1% lidocaine in a 5 mL syringe) posterior GHJ and SASD bursa injection for 2 times at 2 weekly interval.

Posterior GHJ (glenohumeral joint) injection: The patient sits upright on an examination chair with the involved arm adducted and hand on the opposite shoulder to open the posterior shoulder joint. The ultrasound transducer (L18-4, Philips EPI Q5) is aligned in the long axis to the musculotendinous junction of the infraspinatus muscle, just inferior to the scapular spine, with the posterior glenoid rim and posterior glenohumeral joint centered in the field of view. The needle is inserted laterally to medially into the posterior joint capsule, between the posterior humeral head and labrum via an in-plane approach and 2/3 of the injectate is injected into the posterior humeral joint SASD bursa injection: After the GHJ injection, the involved arm relaxed at the side and the transducer is moved to the lateral side of the shoulder. The remaining 1/3 of the injectate is then injected into the subdeltoid bursa via in-plane approach.

Physical therapy: The physical therapy program includes physical modalities (heat therapy and electric therapy) and therapeutic exercise (stretching, ROM exercise, strengthening and mobilization exercise), three times a week, and will be continued for 4 weeks.

Inclusion Criteria:

• 1). age between 18 to 80 years old;
• 2). shoulder pain for ≥ 1 month;
• 3). > 30% loss of passive range of motion (ROM) of the affected shoulder in external rotation and/or abduction, comparing with the sound side;
• 4). pain VAS on maximal passive external rotation or abduction > 4;
• 5). showing willing to receive TAME, or shoulder joint injection, and attend regular physical therapy programs for 4 weeks.
• 6). an empty or soft end feel during passive external rotation or abduction and/or night or rest pain.

Exclusion Criteria:

• 1). severe systemic disorders including cancer, stroke, or cardiopulmonary diseases;
• 2). uncontrolled DM;
• 3). rotator cuff tear or calcification of the affected shoulder;
• 4). fracture, dislocation, or arthritis of the shoulder due to rheumatic disorders; - 5). a history of allergy to local anesthetics, corticosteroids, imipenem/cilastatin sodium or contrast medium;
• 6). pregnancy;
• 7). corticosteroid joint or subdeltoid bursa injection of the affected shoulder during the preceding three months;
• 8). receiving TAME of the affected shoulder during the preceding six months.