Twenty-one psychiatric inpatients with prominent depressive symptoms underwent dexamethasone suppression tests and assessment with observer-rated and self-rated anxiety, depression, and somatic symptom inventories. This was done to test the hypothesis that anxiety, more than depression, was related to cortisol nonsuppression seen in psychiatric patients including those diagnosed as having major depressive disorders. Nonsuppressors were significantly more depressed but not more anxious on the symptom inventories. In addition, it was noted retrospectively that the depression symptom inventory scores predicted nonsuppression. Several individual items from the symptom scales which correlated with post-dexamethasone cortisol levels were also identified.