COMPARATIVE EFFECTIVENESS OF INJECTION THERAPIES IN LATERAL EPICONDYLITIS: A SYSTEMATIC REVIEW AND NETWORK META-ANALYSIS OF RANDOMIZED CONTROLLED TRIALS

Background Injection therapy with glucocorticoids has been used since the 1950s as a treatment strategy for lateral epicondylitis (tennis elbow). Lately several other injection therapies have become available. Objectives Using a systematic review and network meta-analysis, our objective was to assess the comparative effectiveness of different injection therapies used for pain reduction. Methods Structured literature searches in Medline, Embase, Cinahl, Web of Science, and the Cochrane Clinical Trials Database. Selected studies were randomized controlled trials comparing different injection therapies and/or placebo injections for lateral epicondylitis. Risk of bias was assessed in each of the included trials. The outcome measures were change in pain intensity at the end of the masked trial period (i.e., endpoint) and adverse events. Network (random effects) meta-analysis was applied to combine direct and indirect evidence within and across trial data. Following the arm-based network meta-analysis standardized mean differences (SMDs) were used as the effect size based on differences between means divided by the pooled standard deviation. Results Seventeen trials (1381 patients) allocating individuals to injection with glucocorticoid, botulinum toxin, platelet-rich plasma, autologous blood, hyaluronic acid, prolotherapy, polidocanol, glycosaminoglycan polysulfate, and/or placebo, fulfilled the inclusion criteria. Only 3 (18%) of the trials were judged to be at low risk of bias: one trial of glucocorticoid compared to platelet-rich plasma, and placebo-controlled trials of prolotherapy and polidocanol. Pooled results from all included studies showed that glucocorticoid injection (investigated in 12 treatment arms with 310 patients) was no more effective than placebo with respect to improvement in pain (SMD -0.04 [95% CI -0.45 to 0.35], P=0.81); Botulinum toxin injection (4 treatment arms with 130 patients) was statistically significant but clinically only marginally more effective than placebo (-0.50 [-0.91 to -0.08], P=0.023); whereas autologous blood injection investigated in three treatment arms with 96 patients (-1.43 [-2.15 to -0.71], P =0.001) and platelet-rich plasma with two treatment arms and 114 patients (-1.13 [-1.77 to -0.49], P =0.002) were both clearly more efficacious than placebo. Neither polidocanol (one treatment arm, 18 patients) nor glycosaminoglycan polysulfate (one treatment arm, 32 patients) were found to be superior to placebo, while prolotherapy (one treatment arm, 10 patients) and hyaluronic acid (one treatment arm, 165 patients) were both found to be much more efficacious than placebo (-2.71 [-4.60 to -0.82], P =0.009, and -5.58 [-6.35 to -4.82], P <0.0001 respectively). Conclusions Our study found that most trials that have assessed the value of injection therapy for lateral epicondylitis are at high risk of bias and that their results therefore need to be interpreted with caution. Further high quality trials are needed in this area.