Acupuncture can bring some relief to people with knee arthritis, but the benefits may be at least partly from a placebo effect, a new research review suggests.

In an analysis of nine clinical trials from the past 15 years, researchers found that acupuncture generally seemed to improve knee arthritis sufferers' pain and stiffness in the short term. The patients had osteoarthritis, a degenerative joint disease associated with age, as opposed to arthritis associated with an autoimmune disorder.

However, a closer look showed that the benefits were limited to trials that compared acupuncture with doing nothing or with "usual care," such as anti-inflammatory medication.

In trials that compared acupuncture with "sham" acupuncture, on the other hand, there was no clear evidence that the real therapy was more effective.

Sham acupuncture is accomplished by using non-penetrating needles, or inserting needles only into the superficial layer of skin, at random sites rather than the specific points used in real acupuncture. In studies that evaluated electro-acupuncture, the sham version involved phony electrodes and "mock" electrical stimulation of acupuncture points.

The point is to keep study participants from knowing whether they were receiving the real or the placebo treatment. This helps separate the specific effects of a therapy from any placebo effects -- where people feel better simply because they believe they've been treated.

The new findings suggest that the benefits of acupuncture for knee arthritis are at least partly due to patients' expectations, the study authors report in the Annals of Internal Medicine.

However, that doesn't mean acupuncture is not worthwhile, according to the researchers, led by Eric Manheimer of the University of Maryland School of Medicine in Baltimore.

Research does suggest that acupuncture has a "genuine biological effect," and there was evidence in some studies that the real therapy resulted in somewhat better short-term effects than sham acupuncture, the researchers note.

For their study, Manheimer and his colleagues combined the results of nine clinical trials conducted in Europe, the U.S. and Thailand. The trials included a total of more than 3,500 subjects.

Each trial included a patient group that received acupuncture for knee arthritis, as well as a "control" group. In some studies, control patients were placed on a waiting list for acupuncture, while in others they received some standard therapy that acupuncture patients did not. Control patients in other studies received sham acupuncture.

In general, the Manheimer's team found, only studies that pitted acupuncture against doing nothing, or against standard care, showed clear benefits. The results of the sham-controlled trials were too mixed to show any benefits, according to the researchers.

The investigators do not, however, dismiss the potential benefits of acupuncture for knee arthritis. Indeed, they note, a possible explanation for the mixed results is that sham acupuncture had some actual biological effects.

Given the overall safety of acupuncture, the researchers conclude, patients can still consider it as one option in a "multidisciplinary approach" to treating knee arthritis.