Acupuncture is increasingly used as an alternative or complementary therapy for the treatment of pain. It is well tolerated, with a low risk of serious adverse effects. Traditional and modern acupuncture techniques may result in reported improvement in pain patterns. Research on acupuncture has had a number of limitations, including: incomplete understanding of the physiologic effects of acupuncture; ineffective blinding of participants; unclear adequacy of acupuncture "dose;" difficulty in identification of suitable sham or placebo treatments; and the use of standardized treatment regimens rather than the individualized approach that characterizes most acupuncture practice. Controlled trials have been published regarding acupuncture for lumbar, shoulder, and neck pain; headache; arthritis; fibromyalgia; temporomandibular joint pain; and other pain syndromes. Enough data are available for some conditions to allow systematic evaluations or meta-analyses. Based on published evidence, acupuncture is most likely to benefit patients with low back pain, neck pain, chronic idiopathic or tension headache, migraine, and knee osteoarthritis. Promising but less definitive data exist for shoulder pain, fibromyalgia, temporomandibular joint pain, and postoperative pain. Acupuncture has not been proven to improve pain from rheumatoid arthritis. For other pain conditions, there is not enough evidence to draw conclusions.