Re:"Determining the level of evidence for the effectiveness of spinal manipulation in the upper limb: A systematic review meta-analysis"

We recently read a very interesting article that is in press in Manual Therapy titled, “Determining the level of evidence for the effectiveness of spinal manipulation in the upper limb: A systematic review and meta-analysis”. (Aoyagi et al., in press) The authors should be commended for their work. (Aoyagi et al., in press) cite a study by (Riley et al., 2015) titled, “Short-term effects of thoracic spinal manipulations and message conveyed by clinicians to patients with musculoskeletal shoulder symptoms: a randomized clinical trial” (Riley et al., 2015) In their review and metaanalysis, (Aoyagi et al., in press) use the PEDro scale and the Risk of bias summary to evaluate the research studies. On both of these scales, the authors report that the (Riley et al., 2015) study did not meet the criteria of either blinding of patients or a blind assessor. Since none of the studies in their reviewmet the criteria of having a blind assessor, we are unsure how (Aoyagi et al., in press) graded this criteria. In Section 2.6 of their paper they report, “As the primary outcome in this study was pain intensity, which is a subjectivemeasure, blinding of outcome assessors were unsatisfied unless patients were blinded.” (Aoyagi et al., in press) This is consistent with the PEDro scoring guidelines, (PEDro scale, 2015) which state “In trials in which key outcomes are self-reported (eg, visual analogue scale, pain diary), the assessor is considered to be blind if the subject was blind”. However, the following sentence is found in the methods section of (Riley et al., 2015) manuscript on page 2: “The therapist performing the examination was blinded to group allocation, and performed all measures before the interventions, immediately following the interventions and at a short-term follow-up.” (Riley et al., 2015). In the case of Riley et al., 2015, subjects were randomly assigned to one of two similar treatment arms, specifically to address the concern of patient blinding. Indeed, in their discussion (Aoyagi et al., in press) state, “Riley et al. (in press) employed scapular manipulation as a control intervention. However, due to the close relationship between the scapula and thoracic spine, as the authors stated, it may have resulted in inducing similar impulses as SM.” While it is difficult to implement a convincing sham manipulation, if these two interventions were in fact so