Basal and postdexamethasone concentrations of cortisol and prolactin were studied in three groups of male patients: chronic pain patients with no psychiatric diagnosis (n = 12), chronic pain patients with coexisting major depression by Research Diagnostic Criteria (RDC) (n = 24), and pain-free psychiatric patients meeting RDC criteria for major depression (n = 28). Basal cortisol concentrations were significantly higher in pain-major depression and psychiatric-major depression patients compared to pain patients without psychiatric illness. The frequency of cortisol nonsuppression after dexamethasone was significantly greater in pain patients with major depression (41.7%) compared to pain patients without psychiatric disorder (8.3%), and was comparable to that of psychiatric patients (21.4%). Prolactin concentrations, but not cortisol levels, were significantly correlated with observer-rated severity of depression in pain patients. These findings suggest that cortisol and prolactin abnormalities in chronic pain may be related to psychiatric disorder rather than to pain per se, at least in male patients, and may indicate a role for cholinergic mechanisms in the interface of pain and depression.