Concern with Meta-Analysis of Acupuncture for Neck Pain

Dear Editor:

In the review “Randomized Controlled Trials of Acupuncture for Neck Pain: Systematic Review and Meta-Analysis,” Dr. Fu and colleagues found 29 randomized trials on acupuncture treatment for neck pain, and determined that only 14 of those trials were eligible for their review. ¹ The authors then performed nine meta-analyses under various objectives to assess the effects of acupuncture on neck pain. In their meta-analysis examining the effect of acupuncture for treatment of cervical radiculopathy, only one of the three studies used was included in the original systematic review. The other two studies had been excluded because their full reports were not available in English, and so only the data published in their abstracts were pooled. ^2,3 Based on the results from this meta-analysis, the authors concluded that acupuncture was effective for cervical nerve root pain from the statistical standpoint.

Aside from limitations with using data from abstracts alone in meta-analyses, a bigger concern that warrants our attention is that these two studies did not meet the inclusion criteria stated in the systematic review. One of the inclusion criteria established by the review authors was that subjects participating in the study had neck pain for at least a month. However, upon reviewing the abstracts of these studies, we found that neither of them explicitly stated the duration of neck pain subjects had experienced prior to their participation in the study. Thus, these two studies should not have been used in the meta-analysis, and the conclusion drawn from it must be interpreted with caution.