Serial 1-mg dexamethasone suppression tests with concurrent plasma dexamethasone assessments were conducted in 58 patients with endogenous depression treated with electroconvulsive therapy (ECT). Plasma cortisol levels decreased significantly from pretreatment to immediately posttreatment, and they declined further during the first week after the ECT course, when patients remained drug free. Plasma dexamethasone levels showed an opposite pattern of progressive increases over these three time points. The progressive changes in plasma dexamethasone and cortisol levels seen during the week after ECT indicate that alterations in the bioavailability of dexamethasone and in hypothalamic-pituitary-adrenal axis function may be incomplete immediately after the ECT course. This may partly account for previous inconsistencies in serial dexamethasone suppression test findings with this treatment modality. The major finding was that clinical response was associated with increased plasma dexamethasone levels, whereas changes in cortisol levels were independent of clinical outcome. With ECT, changes in plasma dexamethasone levels may be more related to changes in clinical state than changes in postdexamethasone cortisol levels. The extent to which clinical recovery with other treatments in depression is associated with altered bioavailability of dexamethasone and perhaps other compounds is unknown and in need of investigation.