Shoulder pain after stroke is a very common, causing significant morbidity disease. Subacromial and subdeltoid (SASD) bursitis are common causes of pain or disability of the shoulder joint in stroke patients. Traditional therapeutic approaches for the shoulder pain therapy including pharmacotherapy, injection therapy, physical therapy, and behavioural modification. Unfortunately, these therapy methods may not be effective in many patients and long term benefit after treatment is transient, the outcomes may also be incomplete or non-existent. Botulinum toxin A (BoNT-A) is a neurotoxin that can inhibit not only the acetylcholine at the neuromuscular junctions but also other neurotransmitters such as glutamate, substance P and calcitonin gene related peptide, all of which have been indicated in pain transmission. Despite the therapeutic benefit of BTX in alleviating painful muscle spasms, its efficacy in SASD bursitis conditions is less clear. So we perform this study to examine the efficacy of ultrasound guided SASD injection with BoNT-A in reducing refractory shoulder pain after stroke.

Inclusion Criteria:

1. Stroke patients with pain around the shoulder or lateral deltoid area and deteriorated during active or passive overhead activity;
2. Neer and/or Hawkins tests (+);
3. NRS>5 at rest;
4. Symptoms lasted for at least for 2 months and were unresponsive to analgesic medication or physical therapy for 1 month.
5. Subjects voluntarily sign the informed consent.
6. Age between 18 and 80 years old. -

Exclusion Criteria:

1. Received earlier subacromial injections of corticosteroids or botulinuim toxin in the last 6 months;

2. Shoulder fracture, glenohumeral osteoarthritis, bone tumors or osteonecrosis in plain radiographs.

3. Known allergy or sensitivity to study medication or its components.

4. Infection or dermatological condition at the injection sites.

5. Any medical condition that may put the subject at increased risk with exposure , including diagnosed myasthenia gravis, Eaton-Lambert syndrome, amyotrophic lateral sclerosis, or any other disorder that might have interfered with neuromuscular function.

6. QTc criteria: QTc ≥ 450 millisecond (msec) or≥480msec for subjects with Bundle Branch Block-values based on either single electrocardiogram (ECG) values or triplicate ECG averaged QTc values obtained over a brief recording period

7. Liver function tests: aspartate aminotransferase (AST) and alanine aminotransferase (ALT) ≥2xULN; alkaline phosphatase and bilirubin >1.5xULN (isolated bilirubin >1.5ULN is acceptable if bilirubin is fractionated and direct bilirubin <35%).

8. Concurrent use of aminoglycoside antibiotics or other agents that might interfere with neuromuscular function.

9. Patients with severe cognitive impairment or neurological diseases affecting the implementation or evaluation of the test, and drug-dependent patients.

10. Presence of clinically unstable severe cardiovascular, renal or respiratory disease

11. Researchers believe there are other factors unfit to participate in this study of patients.

    -