Effectiveness of Acupuncture for Low Back Pain: A Systematic Review

Jing Yuan, PhD; Nithima Purepong, MSc; Daniel Paul Kerr, PhD; Jongbae Park, KMD, PhD; Ian Bradbury, PhD; Suzanne McDonough, PhD


Spine. 2008;33(23):E887-E900. 

In This Article


This review has provided strong evidence that there is no significant difference between acupuncture and sham acupuncture (superficial needle insertion at nonacupoints), for short-term and intermediate pain relief and functional improvement, which updates the previous evidence that favored acupuncture over sham acupuncture.[15,16] For other comparisons, the addition of the 6 RCTs[11,15,16,17,18,19,20,21] either strengthened or confirmed the previous conclusions, by providing moderate evidence favoring acupuncture over no treatment, strong evidence favoring acupuncture as an adjunctive therapy over conventional therapy alone, and conflicting evidence for acupuncture versus conventional therapy.

Given that our review has shown no difference between acupuncture and sham acupuncture, it is worth exploring the reasons for this result in more detail. Our review included additional studies published after the search dates of the earlier reviews,[15,16] 4 of which we classified as high quality and held significant weight in our qualitative analysis.[11,18,20,21] Another important difference was the fact when the studies were pooled,[15,16] over half were sham TENS studies (all of which we defined as unreliable[52,57,62,63]) and only 3 studies compared acupuncture to sham acupuncture alone or as an adjunct to some form of conventional care.[55,58,69] In our qualitative synthesis, we separated out these 2 latter comparisons to show strong evidence that acupuncture alone is not significantly different from sham acupuncture alone (based on the addition of 2 new trials,[18,21]), whereas the findings for acupuncture/sham acupuncture as an adjunct to conventional care[55,58] provide conflicting evidence.

This lack of difference between sham and real acupuncture raises a debate about how appropriate controls can be chosen. Four of the included studies used superficial needling outside meridians,[18,21,55,58] which has been argued to be as effective as deep needling at specific acupoints[45,58,70,71,72] and considered of therapeutic benefit in traditional acupuncture practice.[73,74] The recently developed nonpenetrating sham needles have been advocated as more appropriate controls.[75,76,77] Indeed, in this review, the only 1 study favoring real over sham acupuncture used a nonpenetrating needling as the control;[19] however, it is worth noting that in other clinical areas, studies using such controls have provided conflicting results.[78,79,80,81]

We were able to strengthen other comparisons, for example, acupuncture was superior to no treatment and as an adjunct to conventional care. We included an additional large high-quality trial[18] to the 2 small low-quality trials used by Furlan to support the superiority of acupuncture to no (acupuncture) treatment. In terms of acupuncture as an adjunct to conventional care, we were able to include 3 new RCTs (2 of which were large high quality trials and used standard medical care as the conventional care comparator[11,20]) with small to large effect sizes. It is of interest to note that in general, the most potent effect sizes in terms of pain and functional disability were observed in the comparison of acupuncture versus no treatment, or acupuncture as an adjunct to conventional therapy, from discharge to intermediate term follow-up. Whereas much smaller effect sizes were observed, in general, when making comparisons to sham acupuncture.

Given the plethora of treatments for LBP, it is important to contextualize the results of the current review with respect to current guidelines such as the European Guidelines.[22] The effects of acupuncture are equivalent to the effects sizes for treatments currently advocated (exercise, pain relief e.g., NSAIDS, behavioral treatments).[82] Although the current review is unable to answer the question about acupuncture versus a completely inert and indistinguishable placebo control as in medication studies,[83] this is also the case for manipulation, which has a smaller effect size[82] and is advocated in the guidelines.[22]

There are some limitations to this review. Firstly, although it was carried out in nonspecific LBP, a few studies on mixed/unclear type of LBP were included,[68] and only studies on specific LBP, such as sciatica, were excluded. Secondly, it was limited to English studies only. However, many of the non-English articles e.g., 29 RCTs in Chinese would have been excluded in our review because of the lack of valid/reliable or objective outcome measures. Finally, the measure of clinical effectiveness for pain in our review was set at 2 points (or 30% relative to baseline), which correlates with a patient global improvement rating of much improved or very much improved.[84,85] It has been suggested that a cut off of 50% would be more stringent, but as pointed out in the editorial by Rowbotham,[86] a 50% reduction in pain intensity corresponds to the highest level of patient impression of improvement. Given the accompanying lack of side effects of acupuncture for pain relief[87,88] and the consensus in LBP around 2 points (or 30%) as a indicator of real change from the patients perspective,[29,32,89] we feel that a choice of 2 points is a valid cut off for meaningful clinical change.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.