Acupuncture is an ancient Chinese technique of using a fine needle to stimulate points along theoretical meridians of energy to correct imbalances thought to be responsible for specific disease states. In the United States, acupuncture is often used for the treatment of painful conditions. The 1997 NIH Consensus Conference concluded that there was adequate evidence of efficacy in an acute dental pain model and in nausea. In chronic pain, most studies were too small, poorly designed, poorly executed, or improperly controlled to adequately demonstrate that needle acupuncture worked better than sham acupuncture, placebo, standard medical therapy, or even no treatment. Osteoarthritis (OA) of the knee has been proposed as a good model to test the efficacy of acupuncture in a chronic pain condition because it is an extremely common, well defined, and disabling condition with well established outcome measures for symptoms and functional status. There is clinical trial evidence of efficacy for the standard treatments of acetaminophen and NSAIDs, and exercise physical therapy (EPT), which is usually added when the patient develops functional limitations. One high quality study of acupuncture for knee OA, demonstrated moderate benefit in an unblinded comparison to a usual care control group. As such, a major question remains about whether acupuncture, used in addition to exercise therapy, will provide a clinically meaningful improvement in pain and function. Since pain can be the primary limiting factor in improved exercise capacity, if acupuncture has any efficacy in reducing the pain of knee OA, then the combination with an EPT program should be substantially more effective than EPT alone. Another major concern is that the effect of the acupuncture may be predominantly mediated by non- specific placebo effects rather than the specific effects of the placement of a needle. Another important component of this proposal is our use of a validated blinded placebo needle instead of sham acupuncture points. Therefore, the primary goal of this proposal is to use a properly designed randomized blinded clinical trial, using American College of Rheumatology (ACR) criteria and Food and Drug Administration (FDA) recommended outcome measures, to determine whether the addition of acupuncture to standard EPT provides an overall clinically important benefit to patients with symptomatic knee OA compared to placebo acupuncture. As a secondary goal, we will use the clinical trial data to develop prognostic and etiologic models for the patients that are most likely to respond to acupuncture. If a clinically important benefit for acupuncture is found, a broader application of this technique would be justified. However, if the results are negative, then the addition of acupuncture to EPT should be generally curtailed.