Background/Purpose:

SubAcromial Impingement (SAI) accounts for approximately 50% of all shoulder pain. The most commonly used treatments for SAI are glucocorticoid injection, exercise therapy or a combination of both. However the effect of exercise added to glucocorticoid injections is inadequately described in the peer reviewed literature. The aim of this study was to investigate the effectiveness of a standardized combination of exercise added to glucocorticoid injection compared to the glucocorticoid alone as the therapy for treatment of SAI.

Methods:

Patients with unilateral SAI of minimum 4 weeks duration and thickened subacromial bursa (>2mm assessed by ultrasound) were included (NCT: 01506804). At baseline patients were randomized to two steroid injections into the painful shoulder separated by one week with subsequent 10 weeks exercise therapy of the involved shoulder (intervention group, IG), or two steroid injections into the painful shoulder separated by one week with subsequent 10 weeks exercise therapy of the un-involved shoulder (defined as sham-intervention, control group; CG). The patients were re-examined after the exercise program (at week 12) and again at week 26. Primary outcomes were change from baseline in shoulder pain at rest and during active shoulder abduction analyzed using intention-to-treat population with a non-responder imputation for missing data (i.e., baseline observation carried forward).

Results:

99 patients (58 females) with an average age of 49 years were randomised (49 IG/50 CG). In the IG, 17 participants were lost to follow-up; in the CG 18 were lost.  

At the 12 week follow-up there were no statistically significant group difference in pain at rest (MD= 1.677 [95% CI -3.644 to 6.001], P=0.533); and during abduction (MD=2.230 [95% CI -6.462 to 10.921], P=0.612); both groups had decreased pain. . At the second follow up (26 weeks) the positive effect on pain at rest was maintained in the CG but decreased in the IG group; there was a trend  towards a difference between groups  (MD= 5.613 [95% -0.867 to 12.093], P=0.088). The beneficial effect on pain during active abduction was not different between groups at week 26 (MD= 2.234 [95% CI -6.754 to 11.224], P=0.623).   

Conclusion:

We found no beneficial effects of combined glucocorticoid injection and exercise therapy of the involved shoulder compared to injections combined with exercise of the uninvolved shoulder. In fact, the CG experienced a prolonged effect on pain at rest compared to the IG. This finding is in contradiction to other studies showing a positive effect of exercise in patients who have received glucocorticoid injection prior to exercise therapy.