A group of 89 patients admitted to the Psychiatric Unit of a General Hospital, with diagnosis of depression according to ICD-9 criteria and randomly chosen were studied. 26 of them (29.2%) were resistant depressions and 63 (7.82%) responded to treatment. Both groups were composed retrospectively in order to analyse resistance or lack of response to a first treatment with tricyclic or tetracyclic antidepressants in effective dosages. We considered: the personality type; the associated somatic pathology; prolonged social stress; period of evolution of the disorder; previous treatments; type of disorder uni or bipolar; familial morbidity; diagnosis reconsideration; analysis of therapeutic compliance; side effects and intolerance. We observed a statistically significant difference with regard to prophylactic treatment (38.4% resistant versus 98.3% non-resistant, p less than 0.0001); a greater suspicion of psycho-organicity in the resistant group (p less than 0.05); a longer interval of time for the resistant group between the beginning of the disorder and the beginning of treatment in our service (F: 1.45, t: 2.58, p less than 0.01). The alternatives used with our patients are analysed and we propose schedules for tackling this kind of problem.