The investigators aim to compare the effect of ultrasound-guided suprascapular nerve block and intra-articular corticosteroid injection for frozen shoulder in pain control, improvement of range of motion, quality of life, and function.

The inclusion criteria is: 1. unilateral side involvement; 2. ≥50% loss of passive ROM (external rotation or abduction) in the glenohumeral joint comparing to the unaffected side; 3. duration of symptoms ≥3 months. 4. age≥ 20year old.

The exclusion criteria include: 1. ever received manipulation of the affected shoulder with/without anesthesia; 2. systemic disease, severe degeneration, or trauma involving the shoulder (ie, rheumatoid arthritis, osteoarthritis, history of injury to the labrum or articular cartilage or malignancies in the shoulder region, etc.); 3. neurologic diseases such as stroke or peripheral nerve neuropathy that have already affect the activity of shoulder; 4.pain or disorders of the cervical spine, elbow, wrist, or hand; 5.a history of drug allergy to local or corticosteroids; 6. Pregnancy or lactation; and 7. Received corticosteroids, or hyaluronic acid intra-articular injection into the affected shoulder during the preceding 4 weeks 70 participants will be randomly divided to SCNB group and IACI group. Each subject in either group will receive 2 times of injection in a two-weekly interval. After injection, each subject in either group will receive the same protocol of physiotherapy program for 2 months. Outcome measures include the SPADI, the SDQ, active and passive ROM of the affected shoulder, and SF-36. Evaluation will be performed at baseline and at 4 weeks and 12 weeks after the beginning of the treatment. Statistics will be performed after completing the patients' treatment and evaluations.

A Shapiro-Wilk test will be applied to ensure that all the calculated variables followed a normal distribution. Subsequently, a 2-by-3, 2-way mixed-model analysis of variance, which has 1 between-subject factor (group: SCNB and IACI) and 1 within-subject factor (evaluation time: pretreatment, 4 weeks after beginning treatment, and 12 weeks after the beginning of the treatment), will be performed. Pair-wise comparisons between groups will be performed using an independent t test when a significant interaction is found; otherwise, the main effects are reported. When a time effect is found, a post hoc analysis will be performed using a polynomial test to determine the trend (linear or quadratic). All significance levels are set at alpha=.05, and SPSS version 15.0a will be used for all the statistical analyses.

Inclusion Criteria:

1. unilateral side involvement.
2. ≥50% loss of passive ROM (external rotation or abduction) in the glenohumeral joint comparing to the unaffected side.
3. duration of symptoms ≥3 months.
4. age≥ 20year old.

Exclusion Criteria:

1. ever received manipulation of the affected shoulder with/without anesthesia.
2. systemic disease, severe degeneration, or trauma involving the shoulder. (ie, rheumatoid arthritis, osteoarthritis, history of injury to the labrum or articular cartilage or malignancies in the shoulder region, etc.)
3. neurologic diseases such as stroke or peripheral nerve neuropathy that have already affect the activity of shoulder.
4. pain or disorders of the cervical spine, elbow, wrist, or hand.
5. a history of drug allergy to local or corticosteroids.
6. Pregnancy or lactation.
7. Received corticosteroids, or hyaluronic acid intra-articular injection into the affected shoulder during the preceding 4 weeks.