Two daily doses of oral Harpagophytum extract WS 1531 (600 and 1200, respectively, containing 50 and 100 mg of the marker harpagoside) were compared with placebo over 4 weeks in a randomized, double-blind study in 197 patients with chronic susceptibility to back pain and current exacerbations that were producing pain worse than 5 on a 0-10 visual analogue scale. The principal outcome measure, based on pilot studies, was the number of patients who were pain free without the permitted rescue medication (tramadol) for 5 days out of the last week. The treatment and placebo groups were well matched in physical characteristics, in the severity of pain, duration, nature and accompaniments of their pain, the Arhus low back pain index and in laboratory indices of organ system function. A total of 183 patients completed the study. The numbers of pain-free patients were three, six and 10 in the placebo group (P), the Harpagophytum 600 group (H600) and the Harpagophytum 1200 group (H1200) respectively (P = 0.027, one-tailed Cochrane-Armitage test). The majority of responders' were patients who had suffered less than 42 days of pain, and subgroup analyses suggested that the effect was confined to patients with more severe and radiating pain accompanied by neurological deficit. However, subsidiary analyses, concentrating on the current pain component of the Arhus index, painted a slightly different picture, with the benefits seeming, if anything, to be greatest in the H600 group and in patients without more severe pain, radiation or neurological deficit. Patients with more pain tended to use more tramadol, but even severe and unbearable pain would not guarantee that tramadol would be used at all, and certainly not to the maximum permitted dose. There was no evidence for Harpagophytum-related side-effects, except possibly for mild and infrequent gastrointestinal symptoms.