This discrepancy appears due to different inclusion criteria allowing different trials to be included.11 included a sham-controlled, no treatment-controlled or pharmacological- or non-pharmacological-controlled trials. Their review had a trial where acupressure was compared to ibuprofen and a sham-controlled trial published in Farsi.

Meta-analysis of the two trials of spinal manipulation did not identify a significant effect on pain overall. One of the two trials did achieve a statistically significant benefit, but as the interventions applied in both trials were similar and both used sham manipulation as a control, it is difficult to attribute this to anything other than random variation. Therefore, the result of the meta-analysis provides the best answer: if there is any effect, it is clinically trivial. A similar result Selleck SCH-900776 was reported by Proctor et al,10 although that review also allowed the inclusion of data about the chiropractic Toftness adjustment technique. Heat caused a significant reduction in pain, although this result was derived from only one trial with 40 participants.19 This was achieved with a 180-cm2 heat patch capable of supplying 38.9 °C heat for 12 hours per day for 3 days. As noted in

Table 2, both groups also A-1210477 nmr received a placebo tablet (because other participants in the trial received ibuprofen). Therefore, even if participants science recognised that their patch was unheated, the placebo

tablet may have helped to control for placebo effects. The reduction in pain of 1.8 is close to the clinically worthwhile threshold of 2,31 so further data in this area would be helpful in narrowing the 95% CI, which currently extends up to a clinically worthwhile 2.7 and down to a clinically trivial 0.9 on the 0–10 scale. The evidence about TENS had similarities to the evidence about heat. It was derived from one small trial; the best estimate of the effect (ie, 2.3) was similar to the clinically worthwhile threshold; and the 95% CI extended well above and below this threshold. This result contradicts that of Proctor et al,9 who pooled the results of three studies and concluded that TENS had no statistically significant effect, although their analysis was based on the odds of obtained threshold pain reduction. To achieve the result observed in our review, Neighbors et al2 delivered TENS at a rate of 1 pulse per second with pulse width 40 μs for 30 minutes. Low-rate TENS delivered at a frequency of 2 Hz is believed to induce analgesic effect through an endorphin-mediated mechanism.32 The yoga intervention assessed a set of three simple postures (cobra, cat, and fish) executed in a 20-minute session daily during the luteal phase. The mean reduction in pain (3.2) and the 95% CI limits (2.2 to 4.2) were all above the clinically worthwhile threshold of 2.