Patients who came to the Department of Rehabilitation Medicine with the primary complaints of shoulder pain from January 2011 to June 2012 were selected. Those who met the following inclusion criteria were finally considered: 1) who had a clinical sign of a painful arc and positive in Hawkins test or Neer impingement sign; 2) who had a precise rotation cuff injury including partial or full rotation cuff tears, or subacromial bursitis under ultrasonography; and 3) who showed limited active ROM and stiffness to the extent not to be included in the exclusion criteria of adhesive capsulitis. The exclusion criteria were as follows: 1) who was diagnosed with adhesive capsulitis-normal findings on X-ray; <100° of active or passive elevation of arm, when raising the arm above the head to a maximum; and passive external rotation of glenohumeral joint being 50% less than the unaffected side when measured with the examiner holding the wrist of the subject and bending his/her elbow to 90°; 2) who underwent shoulder surgery; 3) who had corticosteroids or hyaluronic acid injection into the shoulder due to same etiology; 4) who had hemiplegic shoulder pain; and 5) who showed any suspected fracture on X-ray. Eligible subjects were randomized into three groups A, B and C. The participants were assigned into the groups with the random block assignment method, and patients were lined up to be assigned. Six patient-blocks were further subdivided into groups A, B and C. For every sixth participant, one method was randomly selected and assigned to that individual.

We performed physical examination and evaluation of shoulder pain and function and used ultrasonography to confirm the lesion. Then, the subject made his/her arms to internal rotation and hypertension and put the forearms close to the back. The target area was disinfected, and ultrasound-guided subacromial bursa injection was carried out. All subjects of the three groups had an injection once a week for two consecutive weeks with different dosage: group A, triamcinolone acetonide 20 mg (0.5 mL) and 0.5% lidocaine 5.5 mL, total 6 mL; group B, triamcinolone acetonide 20 mg (0.5 mL), H-lase (Kuhnil Pharm., Seoul, Korea) 1,500 IU (1 mL) and 0.5% lidocaine 4.5 mL, total 6 mL; and group C, triamcinolone acetonide 40 mg (1 mL) and 0.5% lidocaine 5 mL, total 6 mL. They were asked to discontinue any medication including painkiller and anti-inflammatory drug 1 week before the injection. The patients were educated on perishoulder stabilization and stretching exercises to be performed at home. At each visit, the patients were also educated and recommended to perform other self-exercises including the scapular stabilization exercise. There was no additional physiotherapy or medication.

Subjects were evaluated three times: before the 1st injection, 1 week after the 1st injection, and 1 week after the 2nd injection. The examinations were conducted by one examiner, who had evaluated the subjects before the 1st injection. The examiner asked the subjects to rate the average degree of pain for 24 hours before the evaluation, on the visual analogue scale (VAS, 0 to 10).

The examiner then measured active ROM of the shoulder but did not limit its movement. To measure shoulder flexion and abduction, the examiner measured the range-of-movement of the arms in sagittal and coronal planes while the subject was asked to sit and extend the elbow joint. For external and internal rotation, the subject was sat and externally rotated the shoulder to 90°. The normal range of shoulder flexion and abduction was defined as 180°, while that of external and internal rotation was 90°. We adapted the shoulder disability questionnaire (SDQ) to evaluate the shoulder function of subjects. SDQ is a tool whose reliability and validity have been verified for evaluating the degree of shoulder disability that a patient feels. It is composed of 22 self-reported questions (0, no disability; up to 22, maximum disability) [8].

SPSS ver. 12.0 (SPSS Inc., Chicago, IL, USA) for Windows was used for all analyses. All values were measured for the groups before 1st injection, 1 week after 1st injection, and 1 week after 2nd injection. Differences in sex, age, and shoulder-affected as well as pre-injection scores between groups A, B, and C were analyzed using Kruskal-Wallis test with Mann-Whitney U test. Differences in VAS, SDQ, and ROM between groups A and B and between groups A and C were analyzed using Mann-Whitney U test. The degree of VAS, SDQ, and ROM before and after injection in each group was analyzed by using Wilcoxon signed-rank test. Statistical significance was assumed at p<0.05.

Thirty subjects were randomly assigned to groups A, B, and C (10 subjects in each group). There was no significant difference in sex, age, and shoulder-affected as well as in pre-injection scores between the groups (Table 1). No significant difference was observed in pre-injection ultrasonic findings between the groups (Table 2).

All groups showed statistically significant improvement after 1st injection in VAS, compared to pre-injection. The significant improvement was also observed after 2nd injection, compared to 1st injection (Table 3). When comparing the degree of improvement after 2nd injection to that before injection, there was statistically significant difference between groups C and A or B, but not between groups A and B (Table 4).

In group B, SDQ was significantly improved after 1st injection, compared to pre-injection. SDQ was statistically significantly improved after 2nd injection in all groups, compared to 1st injection, but when compared to pre-injection, only in groups B and C (Table 3). There was statistically significant difference in improvement after 2nd injection between groups C and A or B, compared to pre-injection, but not between groups A and B (Table 4).

When comparing the 1st injection results to pre-injection, there were statistically significant improvements in abduction (groups B and C) and flexion (group C only). In comparison of the 2nd injection results to 1st injection, statistically significant improvements in abduction and external rotation were observed in group B only. When comparing the 2nd injection results to pre-injection, flexion and abduction were statistically significantly improved in all groups and additionally external rotation in groups B and C (Table 3). In regard to the degree of improvement after 2nd injection, compared to pre-injection, there was no statistically significant difference between the groups (Table 4).