Comparison of these meta-analyses revealed an interesting pattern. Meta-analysis of the no-treatment controlled trials indicated significant reductions in pain intensity due to acupuncture (by 2.3) and acupressure (by 1.4) on a 0–10 scale. However, the meta-analyses for both acupuncture and acupressure were less promising when the control arm received a sham, with both pooled analyses showing no statistically significant differences selleck between groups. This suggests that the effects of acupuncture and acupressure are mainly attributable to placebo effects. It is difficult to interpret the relevance of the specific acupoints used. Seven of the 10 experimental interventions in the acupuncture

and acupressure trials used the

SP6 (Sanyinjiao) acupoint, which is located approximately 4 cm above the medial malleolus, at the posterior border of the medial aspect of the tibia.22 Most researchers select this because it is the acupoint of choice in gynaecology.26 It is also easy to locate and apply pressure to SP6 without a clinician’s assistance. Among the acupuncture trials, the same results were obtained when different acupoints were GPCR Compound Library screening used (see Figure 2), but different results were obtained when the same acupoints were used (see Figure 4). In contrast, the forest plot of the no-treatment-controlled trials of acupressure shows a range of effects achieved using four different acupoint locations (see Figure 6). It is also to difficult to interpret the relevance of the specific characteristics of the sham acupuncture. The needling regimens were similar to the active intervention, except that Ma et al3 did not use evoke De Qi (needle sensation; stimulation of Aδ fibres evoking soreness and/or a motor response ‘needle grasp’). Ma et al3 did not specify their non-acupoints, but Shi et al23 used a non-meridian acupoint located on the lateral side of lower leg. It is now recognised that needling a few cm away from the acupuncture point may not be a credible placebo.28 and 29 A recent trial investigating the reliability

of acupuncturists in acupuncture point location suggests that there was up to a 6-cm difference in acupuncture point location between the acupuncturists. Neither study used Streitberger placebo needles, which retract – giving minimal to no stimulation.30 The mean estimate of 2.3 reported in the meta-analysis of trials of acupuncture versus no treatment exceeds the clinically significant difference of 2 on the 0–10 scale.31 However, the confidence intervals around this and the other acupuncture/pressure meta-analyses extend below this threshold, so current evidence does not exclude the possibility that the true effects of these interventions – even when supplemented by placebo effects – may be clinically trivial.