Re: A Systematic Review to Assess Comparative Effectiveness Studies in Epidural Steroid Injections for Lumbar Spinal Stenosis and to Estimate Reimbursement Amounts

To the Editor, We read with great interest the systematic review by Bresnahan et al [1] in which they assessed comparative effectiveness studies in epidural steroid injections for lumbar spinal stenosis and also estimated reimbursement amounts. They concluded that there was no evidence for epidural injections and that they were not cost effective. It appears that the authors have missed the purpose of a systematic review, which is to highlight the findings and provide clinicians and other stakeholders a coherent recommendation that uses clear-cut methodologies that are transparent and reproducible and to provide a clinical utility of the therapy in question. Although the authors claim that they evaluated the articles for the presence of bias, the final recommendations reek of personal bias. Bresnahan et al [1] seemingly used appropriate search criteria but failed to find all the manuscripts that were included by some of the authors of this letter in the past. In the search criteria through August 2012, the authors missed 3 systematic reviews [2-4] and 2 randomized, double-blind controlled trials [5,6] in which investigators assessed the effectiveness of caudal and lumbar interlaminar approaches with 1-year follow-up with comparative effectiveness of a local anesthetic with or without steroids. Both of the studies are of high quality and report positive results on a long-term basis. In addition, the authors also have missed another randomized trial by Wilson-MacDonald et al [7], who found that the use of either an epidural steroid injection or an intramuscular injection of local anesthetic and steroids showed positive results in the short-term. However, the authors did include multiple poor-quality studies with small sample sizes and short-term follow-ups. Assessment bias appears to be extremely confusing. The authors state that using the standard evidence-based criteria, 5 of the 6 randomized controlled trials were judged to have low risk of bias from insufficient blinding or loss of followup. However, it is confusing that they have quoted 3 references, 12, 15, 17, from the manuscript. In addition, the authors have not shown the table of bias assessment and the resultant scores. In multiple systematic reviews, all of the studies included in this manuscript were judged with less than high quality or low risk of bias except the manuscript