The P6 Acupuncture Point and Postoperative Nausea and Vomiting
Author: Kimball C. Atwood IV, MD
Abstract:
Several recent reviews state that stimulation of the P6 acupuncture point is effective for preventing postoperative
nausea and vomiting. Each review cites the same 1999 meta-analysis. This article argues that the sources for that
meta-analysis are insufficiently reliable to confirm such an effect. Trials performed since 1999 have been mixed, but
the largest and best of these was negative.
INTRODUCTION
Several recent reviews have asserted that stimulation of the "pericardium 6" (P6) acupuncture point, on the ventral aspect of the wrist, is effective for preventing postoperative nausea and vomiting (PONV). [1-4] Each review offers, as its only evidence, a 1999 meta-analysis by Lee and Done. [5] Table 1 in that meta-analysis reported 19 trials of nonpharmacologic techniques in PONV, each of which involved some form of stimulation of the P6 point-"manual rotation of needles, semipermanent needles, electrical stimulation of needles, transcutaneous electrical stimulation, and acupressure"-in the preoperative, intraoperative, or postoperative setting. [5] Eleven trials were reported as positive and 8 as negative. All 4 trials in children were negative. Lee and Done concluded that these techniques are effective for adults but not for children. They maintained that opinion in the most recent update of the Cochrane Review of the subject, of which they are also the authors. [6]
ANALYSIS
That conclusion is unlikely to be valid, for the following reasons:
1. Prior to the publication of the meta-analysis, other investigators had reported that during the period 1965-1995 (the entire period investigated) certain countries- China, Japan, Hong Kong, Taiwan, and Russia/Soviet Union-produced exclusively positive studies of acupuncture and overwhelmingly positive studies of interventions other than acupuncture. [7] By comparison, 53% and 60% of acupuncture studies in the United States and the UK, respectively, and 75% of nonacupuncture studies in the UK, were positive. The investigators concluded that publication bias in some countries was the most likely explanation. There may also have been other types of bias, but the authors' conclusions were limited by a lack of additional information.
In the Lee and Done meta-analysis, 4 of the positive trials and none of the negative trials were from Taiwan. Among these was the only trial that Lee and Done awarded their highest possible methodology rating, but that trial reported an improbable result: In the intervention group only 1/30 postop caesarian patients had nausea within 24 hours and 0/30 had vomiting (the placebo results were typical for this setting: 43% experienced nausea and 27% vomited). [8] Two of the other reports from Taiwan were given the lowest methodology rating by Lee and Done (see below). Thus it is reasonable to regard all 4 reports from Taiwan as unreliable.
2. Of the 7 remaining positive trials, Lee and Done gave a methodology rating of 1 point (out of a possible 5) to 4. None of the negative trials received that low a rating. According to Lee and Done, a rating of 1 meant either that the study had described itself as randomized but not blinded-with no explanation of the method of randomization- or that the study had reported the number and reason for withdrawals. Thus, to be included in the metaanalysis, a trial needn't have been blinded, appropriately randomized, or randomized at all. It is therefore reasonable to dismiss all trials rated 1.
3. Of the 3 remaining positive trials, 1 reported an anomalous result: The incidence of postdischarge vomiting in the intervention group was 4 times higher than the incidence of postdischarge nausea in the same group. [9] The investigators offered no methodological or statistical explanation for this (such as having intentionally counted nausea and vomiting as vomiting only). Vomiting without nausea in the postoperative setting is unlikely and was not reported by any of the other trials. Thus it is reasonable to regard this report as unreliable.
4. What remain are 2 positive trials and 8 negative trials. Even if we accept Lee and Done's decision to segregate the pediatric from the adult trials-questionable in itself- there are, in adults, 4 negative trials involving 348 patients and 2 positive trials involving 177 patients. [5]
5. Lee and Done concluded that the effects of P6 stimulation were modest: they calculated the number needed to treat (NNT) to prevent early postoperative nausea in a single patient to be 5 (CI = 3-8) and the NNT for early postoperative vomiting to be 5 (CI = 4-8). [5] These numbers are similar to those found for notoriously ineffective pharmacologic antiemetics.1 As Lee and Done mentioned in their Methods but not in their Discussion, interventions for prevention of PONV have been proposed to be clinically useful only if the NNT is equal or less than 5. [10] Thus the data for P6 stimulation barely achieved clinical relevance even without considering the objections above.
6. Nausea is measured subjectively, and its prevention is probably psychogenic in some cases. [11] It is difficult to perform strictly double-blinded, placebo-controlled acupuncture trials. These facts increase the likelihood of interpretive errors.
7. Trials subsequent to Lee and Done's meta-analysis have been mixed. They continue to be vulnerable to serious criticisms, including financial support by the manufacturer of a device that was being tested. [12] A recent randomized, double-blind, placebo-controlled trial found no difference in PONV between 109 women receiving P6 acupuncture and 111 women receiving placebo acupuncture. [13] This is the most methodologically sound trial to date and, as such, holds more weight than a meta-analysis of lesser studies. [14] Lee and Done did not consider this trial in their Cochrane Review update of April 2004 because their search was limited to trials published before February 2003. [6]
COMMENT AND CONCLUSION
Investigations of acupuncture to date have not demonstrated effects on the natural history of any disease. [2,3,15] Results have also been disappointing in acute and chronic pain management, at one time thought the most promising role for acupuncture. [2,3,15,16] At least one recent review considers P6 stimulation for PONV to be the only convincing case for the efficacy of acupuncture, [2] and another would add to that only nausea from chemotherapy or pregnancy, and dental pain. [3] Thus P6 stimulation for PONV is key to any current argument that acupuncture has specific, clinically useful effects. But neither the evidence reviewed by Lee and Done nor subsequent evidence confirms a specific antinausea or antiemetic effect for perioperative stimulation of the P6 acupuncture point.
