SynopsisHypothalamic–pituitary–adrenal axis function was investigated in 72 patients with primary depression. Forty-four per cent of the patients demonstrated abnormal suppression of their cortisol levels after a 1 mg overnight dexamethasone suppression test. Patients with abnormal suppression (‘non-suppressors’) were not clearly distinguished from ‘suppressors’ by the commonly used diagnostic classifications. They did not appear to be more severely depressed, but they were more likely than the ‘suppressors’ to be in-patients. Multivariate analysis of the data suggested that two clinical features were independently associated with non-suppression: the PSE syndromes of Slowness and General Anxiety. However, the association of these syndromes with non-suppression was relatively weak, indicating that the clinical significance of the dexamethasone suppression test is, as yet, unclear. The results raise doubts about the validity of using the dexamethasone suppression test as a diagnostic marker for a specific depressive syndrome.