Résumé : The authors have studied the performances of the Dexamethasone Suppression Test (DST) in 107 hospitalized patients diagnosed according to the Research Diagnostic Criteria (RDC) and Feighner's criteria. The best performances of the DST are obtained for the diagnosis of primary depressed patients, suffering from a major depressive disorder. With the combination of these two diagnostic criteria, we found a sensitivity of 81%, a specificity of 81% and the diagnostic confidence of a positive test is 93%. Our study also shows 90% of abnormal DST results in schizoaffective disorder, depressed type, and no significant difference of the mean cortisol plasma levels at 4 p.m. after dexamethasone administration between depressed schizoaffective patients and major depressives. The finding of a better therapeutic response to antidepressive treatments in DST nonsuppressor patients than in suppressors is of interest for the predictive value of the DST in relation to treatment response.