Hypnosis and acupuncture can alleviate experimentally induced pain but the mechanism of analgesia remains unclear for both techniques. Experimental pain was induced by cold pressor test (CPT) in 8 male volunteers. Analgesic effect of hypnosis (HA) and acupuncture (AA) was assessed before and after double-blind administration of placebo or naloxone, in a prospective, cross-over study. We found that pain intensity was significantly lower with HA as compared with AA, both with naloxone (P less than 0.001) and placebo (P less than 0.001). Within HA or AA groups, pain scores did not differ significantly when naloxone or placebo was administered. During AA, however, pain scores were similar to control values when naloxone was given (P = 0.05) but decreased significantly with placebo (P less than 0.002). Analog scales for pain intensity and pain relief showed a good correlation (r = 0.94). Plasma levels of beta-endorphins did not change significantly in any combination. Heart rate, peripheral arterial blood pressure and skin conductance were very insensitive indices to assess pain intensity or relief, as well as intensity of acupuncture stimulation or depth of hypnotic trance. We conclude: (1) HA and AA can significantly reduce pain from CPT, and HA is more effective than AA: (2) HA and AA are not primarily mediated by the opiate endorphin system; and (3) plasmatic levels of beta-endorphins are not significantly affected by either HA or AA nor by naloxone or placebo administration.