Sixty-four patients with a Research Diagnostic Criteria (RDC) diagnosis of major depressive disorder were categorized into three groups based on their number of depressive episodes (DE): Gr 1 (1 DE), n = 16, Gr II (2-4 DE), n = 25; and Gr III (5 or more DE), n = 23. All patients were nonsuppressors after 1 mg dexamethasone suppression test (DST) prior to the start of treatment. Patients were monitored during the course of their treatment using serial Hamilton Depression scores and post-DST plasma cortisol levels. A proportionately equal number of patients in the three groups had a favorable outcome, i.e., the number of depressive episodes did not predict recovery. Despite favorable clinical outcome, patients with higher numbers of depressive episodes had significantly higher post-DST plasma cortisol levels that were above the supressive range (greater than 5 μg/dl). Patients with a higher number of depressive episodes had a significantly shorter duration of index episodes and were younger at first depressive episode than patients in the other two groups. These results, however, were confounded with polarity, with a higher number of bipolars in Gr III than in the other two groups. Results are discussed in light of phenomenological and psychoendocrine findings of earlier studies. © 1990.