Forty-one RCTs (2,672 patients, range 21 to 331) were included in the review. Study quality ranged between 7 and 13 out of 13.
Corticosteroid injections for lateral epicondylalgia (12 RCTs, 1,171 patients):
Corticosteroids showed statistically significant improvements on all outcomes in short-term treatment of lateral epicondylalgia compared to no intervention: pain (SMD 1.44, 95% CI 1.17 to 1.71; three RCTs, Ι²=0%), function (SMD 1.50, 95% CI 1.22 to 1.77; three RCTs; Ι²=0%) and overall improvement (RR 3.47, 95% CI 2.11 to 5.69; two RCTs; Ι²=50%).
Individual trials showed that compared to physiotherapy, corticosteroids statistically significantly improved: pain (three RCTs that could not be pooled due to significant statistical heterogeneity, Ι²=68%), function (SMD 1.29, 95% CI 1.03 to 1.55; three RCTs; Ι²=0%) and overall improvement (RR 2.37, 95% CI 1.75 to 3.21; three RCTs; Ι²=43%).
By contrast, significant negative effects in favour of no intervention and physiotherapy groups were reported at intermediate and long-term follow-up. There were no statistically significant differences between corticosteroids and placebo for any outcome in the short and intermediate term.
Corticosteroid injections for rotator-cuff tendinopathy (10 RCTs, 780 patients): Corticosteroids showed statistically significant short-term improvements compared with placebo for pain (SMD 0.68, 95% CI 0.35 to 1.01; three RCTs; Ι²=0%) and function (SMD 0.62, 95% CI 0.29 to 0.95; three RCTs; Ι²=0%). No statistically significant differences were found for other comparisons.
Non-corticosteroid injections (15 RCTs; results reported for 12 RCTs): Three RCTs of patients with lateral epicondylalgia showed statistically significant improvements with non-corticosteroids versus placebo: sodium hyaluronate in the short, intermediate and long term for overall improvement and pain (one RCT); botulinum toxin for pain in the short term (one RCT); and prolotherapy for pain in the intermediate but not short term.
Other results for comparisons that included only one RCT were reported in the review.
A statistically significantly greater risk of atrophy was reported in one RCT using corticosteroid injection. Two RCTs reported statistically significantly greater risk of itching or burning with aprotinin. One RCT reported greater overall risk of adverse events with botulinum toxin.