Thursday, September 22, 2011

Electroacupuncture Ineffective for Neuropathy

Acupuncture A recently reported clinical trial examined electroacupuncture — needles inserted at acupoints with electric current applied — for chronic, peripheral neuropathic pain. This therapy afforded no significant pain relief or other benefits; however, the study also serves as an example of poor-quality evidence garnering more attention in a pain journal than it probably deserves.

Writing online ahead of publication in the journal Pain Medicine, researchers in Milan, Italy, report a double-blind, placebo-controlled, cross-over trial to investigate the analgesic efficacy of electroacupuncture (EA) in patients with axonal polyneuropathy [Penza et al. 2011]. They included 16 patients with a diagnosis of chronic neuropathic pain in their lower limbs — Visual Analog Scale [VAS] score > 4.0 for at least 6 months — and taking analgesic medications for at least 3 months at stable doses. There were 9 females and 7 males enrolled; mean age 64.9 years (range 43-75 years).

Subjects were randomized into either EA or Sham-EA (placebo) groups and treated weekly. The protocol involved 6 weeks of treatment, 12 weeks free of treatment, and then cross-over to the alternate group for a further 6 weeks of treatment. Therefore, by the end of 24 weeks all patients had received both EA and Sham-EA, making a total of 16 subjects per group. The primary outcome measure was the number of patients treated with EA achieving at least a 50% improvement from baseline in pain relief, as compared with placebo. Secondary outcomes were improvements in patients’ global impressions of change, depression and anxiety, and quality of life.

Results indicated that only one (1) patient in each group (EA and Sham-EA) reported 50% pain relief at the end of treatment compared with their pain intensity at baseline. Changes in pain intensity were not statistically significant and did not substantially differ between EA (Mean ± SD on VAS: 5.7 ± 2.3 at baseline and 4.97 ± 3.23 after treatment) and Sham-EA (4.9 ± 1.9 at baseline and 4.18 ± 2.69 after treatment). There also were no changes in the secondary outcomes; however, no adverse effects of therapy were recorded for either group.

The authors conclude that their results do not support the use of EA in this population of patients with painful neuropathy, and further trials in larger groups of patients are warranted to confirm this observation. However, this study may not justify any conclusion regarding EA one way or the other.

COMMENTARY: Acupuncture is an ancient modality that is a somewhat popular complementary or alternative therapy for the management of select pain conditions; however, it also is controversial. In some circles, there is considerable skepticism as to whether acupuncture is a science, a pseudoscience, or a hoax [eg, see SBM blogpost here].

The term acupuncture pertains to procedures that stimulate precise anatomical locations by various techniques to produce clinical effects. Besides techniques described in Traditional Chinese Medicine (TCM), variations now include Nogier auricular (ear) acupuncture, Koryo hand acupuncture, Yamamoto scalp acupuncture, and Po abdominal acupuncture. Acupuncture points can be stimulated by surface pressure (acupressure), insertion of needles without manipulation, insertion of needles with manipulation (twisting and/or thrusting), heating of acupuncture needles through radiant heat or moxibustion, laser stimulation of acupuncture points, and electrical stimulation of inserted needles, or electroacupuncture, as in the study by Penza et al. described above.

In a review last April 2011, a practitioner from the Mayo Clinic wrote quite favorably of acupuncture as potentially benefitting a diversity of chronic pain conditions [Dorsher 2011]. At about that same time, Ernst and colleagues reported an overview of 57 qualifying reviews — published between 2000 and 2009, incorporating nearly 700 trials — and concluded there is little convincing evidence that acupuncture is effective in reducing many types of pain and there also should be concerns about its safety [see UPDATE here]. In other Pain-Topics UPDATES [eg, here, here, and here], we have commented on conflicting, and sometimes confusing, evidence of acupuncture’s efficacy.

This present study by Penza and colleagues does little, if anything, to clarify the uncertainties surrounding acupuncture or, in particular, electroacupuncture. For example…

  • Relatively few data are included for interpretation in the article, overall, and the authors do not report having conducted a power analysis to justify the small number of subjects enrolled. Our own calculation finds that, at 80% power and with relatively small standard deviations (ie, high precision) a total of 16 subjects per group could be sufficient to demonstrate a statistically significant large effect size >1.0 (50% absolute reduction in pain) if it does exist. However, the design is weakened by the cross-over design — in effect having each subject serve as their own placebo control — and such small numbers of subjects would not have external validity even if a very large effect size as a result of active EA were demonstrated [see UPDATE discussing statistical power here].

    Note: In fairness, the authors do report checking for a cross-over interaction, and found there was no difference in VAS scores between patients who received EA as their first treatment and patients initially treated with placebo. But, again, with so few subjects in these analyses it is difficult to know if the statistics can be trusted as being valid or relevant for a larger population.

  • Mean baseline VAS scores in the two groups (EA=5.7, Sham-EA=4.9), suggest that these were patients with only moderate neuropathic pain at the outset and large reductions in pain could be more difficult to achieve than in patients having severe pain [see UPDATE here]. So, hypothesizing a possible 50% reduction in VAS scores as a result of EA, and with such a small number of subjects having only moderate pain to demonstrate this, might have been unrealistic and “stacked the deck” so to speak against a significantly successful outcome.

  • It seems unusual that there was such a diminutive response to active EA and that even a typical placebo response to Sham-EA also was suppressed. Responsiveness to both active treatment and placebo might have been confounded by the unspecified analgesics that patients were taking pretrial and were asked to continue throughout the study.

  • Furthermore, throughout the research literature on acupuncture, questions have been raised about the validity of sham or placebo approaches, and this study is no exception. The researchers report that “pseudo” or Sham-EA was performed by applying needles in neutral anatomical points, close to true acupoints, and delivering the same electrical stimulation as in the active EA condition; that is, alternating electric current at low voltage (6 mV) and low frequency (2-10 Hz). In some research, even minimal needling of skin over acupuncture and non-acupuncture point locations has produced physiological effects that muddle differences between placebo responses and therapeutic efficacy. And, the subcutaneous administration of electrical current at any location might possibly affect nerve transmission to alter neuropathic pain perception.

    Therefore, the mere act of inserting needles, the positioning of the needles, and the application of electrical currents at the specified voltage/frequency may have been confounding factors that combined to cancel-out rather than facilitate responses to either active or placebo therapy. Or, electroacupuncture may simply be a largely inert and ineffective therapy for neuropathic pain — we do not really have any definitive answers from this trial, nor is there a solid direction for designing larger, confirmatory trials.

Probably the most puzzling aspect of this low-quality study is why it was published in the (supposedly) peer-reviewed journal Pain Medicine from the American Academy of Pain Medicine. It only narrowly qualifies as a pilot study and barely has adequate substance to be a non-juried poster presentation at a pain conference; so, why is it published as a 5-page journal article?! Studies of this low caliber being published in pain journals is becoming overly common these days and, of further concern, the article now will be indexed in Medline and similar services, cited by other authors as scientific “proof” against electroacupuncture, and included in future data meta-analyses on the subject.

While we have consistently neither endorsed nor eschewed acupuncture, or any of its variations, studies such as this do nothing to help confirm or deny the validity of acupuncture as a modality for better pain management. The field deserves, and reputable journals should insist upon, much better research.

> Dorsher PT. Acupuncture for chronic pain. Tech Regional Anesth Pain Managmt. 2011(April);15(2):55-63 [
abstract here].
> Penza P, Bricchi M, Scola A, et al. Electroacupuncture Is Not Effective in Chronic Painful Neuropathies. Pain Medicine. 2011(Sep 14), published online ahead of print [
abstract here].