In order to investigate the role of serotonergic mechanisms in depressive disorders, the fenfluramine challenge test was performed in 31 patients suffering from different types of depression. The strategy was to select a simple method (i.e., easier to perform than CSF studies for instance) to be applied to a wide range of patients, as close as possible to everyday cases in a clinical setting (i.e., not only to such severe or highly selected groups as is normally the case in biological research in psychiatry). The neuroendocrine test (which consisted of the measurement of variations in the secretion of prolactin, growth hormone and Cortisol after the administration of 60 mg dl-fenfluramine p. o.) did not correlate with symptoms of behavior patterns previously identified with a "serotonin deficit" (i. e., suicidal behavior or attempts, lowering of the control of impulses, sleep disturbances) but only with the severity of the diagnosis (in the DSM-III hierarchical scale) or with indexes of endogeneity (Newcastle scale). This fact could be explained by methodological artifacts (i. e., dlfenfluramine is not a clean probe, showing influence in the dopamine and noradrenaline metabolism; the absorption of fenfluramine was not controlled) or by the fact that the involvement of serotonin in affective disorders is not a selective, isolated dysfunction, but is integrated in more complex interrelationships. Nevertheless, our preliminary findings (even without the results of the comparison with a control group and the evaluation of a few more data and cases) do coincide with the absence of predictors or the lack of specific patterns of response of symptoms with new selective re-uptake blockers of serotonin antidepressants.