To relieve confusion about the clinical correlates and prognostic implications of the dexamethasone suppression test (DST) in schizophrenia, we conducted a DST in 44 schizophrenic inpatients at drug-free baseline and approximately 4 weeks after neuroleptic treatment. Patients were rated on positive, negative, and depressive symptoms at both times. A head computed tomography (CT) scan was performed and measures of ventricle-brain ratio (VBR) obtained. Clinical improvement was monitored at four weeks, and longer-term outcome assessed at 1 year. Seventeen of the 44 patients were DST nonsuppressors at baseline, and five of these remained nonsuppressors at 4 weeks posttreatment. Postdexamethasone plasma cortisol levels were correlated with negative symptoms at baseline (r = 0.45; p < 0.01), but not after 4 weeks of neuroleptic treatment. Postdexamethasone plasma cortisols were not related to global severity, positive, or depressive symptoms at either timepoint or to VBR. Persistent nonsuppression was assocciated with poor outcome, but baseline postdexamethasone cortisol levels were unrelated to outcome at 4 weeks and 1 year. The literature on DST in schizophrenia is reviewed and attempts are made to reconcile discrepant findings and to discuss pathophysiological implications. © 1991.