| Methods |
Randomised controlled trial
Blinding: patients and outcome assessors blinded
Loss to follow‐up: 15 pateints withdrew ‐ 9(16%) in steroid group and 6(13%) in placebo group.
Appropriate statistical analysis: unclear if completers analysis only |
| Participants |
101 patients 1) pain on resisted abduction and/ or external rotation and/or
2) loss of passive movement of glenohumeral joint
Exclusion criteria: polymyalgia rheumatica, biceps tendonitis, as judged by pain on resisted forearm supination, polyarthritis with shoulder involvement, abnormal neurological signs or shoulder/hand syndrome, arthritis acromioclavicualr joint as judges by joint tenderness |
| Interventions |
Group 1(54 patients): Intra‐articular (1 ml) and subdeltoid bursa (1ml) steroid injection of prednisolone acetate (single skin puncture at baseline and 2 weeks.
Group 2(47 patients): saline injection as per steroid group. Arthrogram checked placement and indicated correct placement in subacromial bursa but not joint |
| Outcomes |
Assessed at baseline, 2 and 4 weeks
1) Pain
2) Night pain
3) Pain on resisted abduction or external rotation
4) Loss of passive abduction external, and/or internal rotation.
5) At 2 and 6 weeks, degree of improvement on 5‐point scale (1=worse, 2=no change, 3=slight improvement, 4=definite improvement, 5=complete recovery) for pain, night pain, pain with resisted abduction or external rotation (mean score) and loss of passive abduction, external and internal rotation (mean score) |
| Notes |
Met inclusion criteria for review, but no means or standard deviations reported.
Reported percent of patients with definite improvement or complete recovery (scores of 4 or 5) at 2 and 6 weeks but unclear if all patients included in analysis or just completers so not included in meta‐analysis. Reported a trend toward steroid injections being more effective than placebo. |
| Risk of bias |
| Bias |
Authors' judgement |
Support for judgement |
| Allocation concealment? |
High risk |
C ‐ Inadequate |
