Propofol, an intravenous general anaesthetic, has been reported to relieve some forms of pruritus at subhypnotic doses. We assessed its effectiveness in 32 patients with several kinds of non-malignant chronic pain, in a placebo-controlled, double-blind study. We found that central pain, but not neuropathic pain, is at least partially controlled by propofol at subhypnotic doses, without major side-effects. In particular, allodynia associated with central, but no neuropathic, pain has been completely controlled. Propofol analgesia leads to renormalization of brain metabolism as seen on single photon emission computed tomography. We conclude that propofol may help in the diagnosis of central pain, particularly in unclear cases, and also in treatment. Possible mechanisms of action are discussed.