A decrease in pain sensitivity during acute depression has been observed in several studies, apparently related to the severity of symptomatology. However, the question remains whether this relationship can be found only in heterogeneous groups of depressive patients or also in a single diagnostic group, such as major depression. In the present study, pain thresholds were assessed in 20 patients with major depression (DSM-III-R) and in 20 healthy controls. Two threshold methods with a differing impact of reaction time on the results were used. Contact heat was applied as a natural source of pain. With both methods the pain thresholds were significantly increased in the depressive patients. No relationship was found to the various symptoms of depression assessed by psychopathometric scales. In contrast to the pain thresholds, the thresholds of skin sensitivity for nonnoxious stimuli (warmth, cold, vibration) were only slightly increased. In subsamples (N = 10 in each group), naloxone (5 mg i.v.) and placebo were administered in a double-blind design. No systematic changes in pain thresholds occurred under either treatment. Our findings suggest that the decrease in skin sensitivity in major depression is specific to pain and not due to an increased reaction time. Moreover, the decrease appears to be related neither to a naloxone-sensitive mechanism nor to symptomatology.