When used to treat pain, acupuncture offers only limited relief that may not be clinically relevant.

So say Danish researchers who examined data from 13 acupuncture pain studies that included more than 3,000 patients.

The studies compared real acupuncture, placebo acupuncture and no acupuncture for a wide range of painful conditions such as knee osteoarthritis, migraine, low back pain and postoperative pain.

Compared to placebo acupuncture, real acupuncture offered only a small amount of relief (about 4 millimeters on a 100-mm pain scoring scale), according to the review authors. A 10-mm change on this scale is classified as "minimal" or "little change," which means the apparent relief offered by acupuncture seems to be below clinically relevant improvement.

The findings, published online Jan. 28 in the British Medical Journal, support a number of previous reviews that found no clear evidence that acupuncture offers effective pain relief.

Future studies should focus on reducing bias and trying to separate the physiological effect of using a needle and the psychological impact of the treatment ritual, said the researchers at the Nordic Cochrane Center in Copenhagen.

The overall effect of acupuncture in relation to usual care may not be large, but it may be clinically relevant for musculoskeletal conditions due to the limited treatment options and acupuncture's safety record and patient preference, Dr. Adrian White and Dr. Mike Cummings of the British Medical Acupuncture Society wrote in an accompanying editorial.

Future research should focus on a comparison of acupuncture with the best existing treatments for different conditions, they suggested.