Acupuncture for knee pain


Below is a table comparing the differing effect sizes of a variety of treatment strategies for osteoarthritic knee pain (OANP).

Effect Size Comparisons for Several Interventions on Osteoarthritis


Effect size, 95%


Patient education, self-management vs. pragmatic comparator 0.06 (0.02–0.10)
Weight loss vs. pragmatic comparator 0.20 (0.00–0.39)
Acetaminophen vs. placebo


0.14 (0.05–0.23)

0.21 (0.02–0.41)

NSAIDs vs. placebo 0.29 (0.22–0.35)

0.37 (0.26–0.49)

COX-2 inhibitors vs. placebo 0.44 (0.33–0.55)
Topical NSAIDs vs. placebo 0.40 (0.16–0.66)
Opiates vs. placebo 0.78 (0.59–0.98)
Strengthening exercise vs. pragmatic comparator 0.32 (0.23–0.42)
Aerobic exercise vs. pragmatic comparator 0.52 (0.34–0.70)
Water-based exercise vs. pragmatic comparator 0.19 (0.04–0.35)
Acupuncture vs.

sham acupuncture


0.35 (0.15–0.55)

0.29 (0.10–0.48)

0.26 (0.17–0.34)

including outlier

0.16 (0.07–0.25)

excluding outlier

Acupuncture vs. pragmatic comparator 0.57 (0.50–0.64)

acupuncture for knee pain


A few very interesting points can be drawn from this table. Perhaps most interestingly is that the most effective treatments for OANP are opioid based medicines, acupuncture, and aerobic exercise (which varies more considerably in its benefits). Of these three treatments, acupuncture stands out as the only one that is safe and physically possible for all. Some people are just not able to engage in aerobic exercise and we all know how dangerous and addictive and dangerous opiates can be.

The rest of the treatments, the first line and follow-up strategies of the NHS, are all below 0.5 effect size. Interestingly any effect size below 0.3 is considered small. All of the initial treatments are under this level. The next level of treatment, NSAID’s and COX-2 inhibitors are more effective, however they are also more dangerous. In 2000, 2000 people died in the UK from normal doses of NSAID’s and additional drugs often have to be prescribed to mitigate the side-effects these drugs have.

It is also interesting to note the difference between acupuncture being compared by a pragmatic comparator and sham acupuncture. The reason this effect is so pronounced is that sham acupuncture is therapeutically active. To get the effect size, the effect of the comparison is taken from the study treatment. Therefore, if the comparison is therapeutically active it reduces the comparative effect of the study treatment. This is one reason acupuncture can be shown, in some studies, to have limited effectiveness. In other words, it’s not that acupuncture isn’t effective, it is that sham acupuncture is too effective to be a fair placebo control. Which means, in many cases, acupuncture is far more effective than we have been led to believe!


Birch et al. paper