Pre-treatment and posttreatment dexamethasone suppression test (DST) results in physically healthy elderly major depressives without dementia demonstrated an association between treatment and DST normalization. Sixty percent of subjects were nonsuppressors at baseline compared to 17% after intensive treatment. DST results normalized in 75% of initial nonsuppressors; none of the initial suppressors converted to nonsuppression. A strong correlation between clinical improvement and decreases in afternoon cortisol levels was identified. Initial suppression status did not influence this association. There was a nonsignificant trend for very high (> 15 micrograms/dl) afternoon cortisol levels to be associated with delusional depression. The advantage of using continuous rather than categorical measures to assess the relationship between reversal of depression and DST results is discussed.