The usefulness of the dexamethasone suppression test (DST) in the prediction of treatment response to selective antidepressants was tested in 56 major depressive inpatients. Following DST, patients were randomly assigned to treatment by either nomifensine, a catecholaminergic antidepressant, or zimeldine, a serotonergic antidepressant during a 3-week period and assessed by means of the second part of the Clinical Global Impressions (CGI-2). No significant difference was present between the 27 DST suppressor and the 29 DST non-suppressor patients in their overall clinical outcome. Moreover, no preferential response to nomifensine or zimeldine was noted in any of the two groups defined according to DST status. Therefore, these results do not support the usefulness of the DST in the prediction of the treatment response to antidepressants in general and to selective antidepressants in particular.