The dexamethasone suppression test (DST) has been widely used in psychiatry as a laboratory aid for the diagnosis of endogenous depression; failure to suppress serum cortisol levels is interpreted as confirming a clinical diagnosis of endogenous depression. We found that serum dexamethasone concentrations in this test vary widely and are determinants of the DST response: non-suppression of serum cortisol levels is associated with low serum dexamethasone concentrations, and suppression is associated with high concentrations.