Forty-four patients with chronic cervical osteoarthritis took part in this study. Patients were treated with acupuncture, sham-acupuncture, diazepam or placebo-diazepam in randomized order. Pain was rated on visual analogue scales before, during, and after treatment. Two scales were separately used to rate the intensity (sensory component) and the unpleasantness (affective component) of pain. The results analyzed from these trials show that diazepam, placebo-diazepam, acupuncture and sham-acupuncture have a more pronounced effect on the affective than on the sensory component of pain. Acupuncture was significantly more effective than placebo-diazepam (p less than 0.05), but not significantly more effective than diazepam or sham-acupuncture.