[Psychiatric diagnosis and cause of death in a hospitalized population].

Psychosomatic theories suggest that psychiatric pathology modulates mortality rates; some diagnoses could be considered as risk factors and others as protective factors regarding the different causes of death. The results obtained so far are controversial. The present study aimed at determining the associations between psychiatric diagnosis and cause of death on the basis of 1698 deaths which occurred in a regional Swiss psychiatric hospital during the 1945-1989 period. The diagnoses were divided into four groups (psycho-organic syndromes/dysthymia/schizophrenia/others) and the causes of death into five groups (cardiovascular/lung diseases/cancer/suicide/others). As data were lacking for the entire group of hospitalized patients, associations were analyzed by way of a proportional approach (PMR). The proportional variation factors attributed to a given cause were studied by logistic regression (case-control design). The results mainly reveal the specificity of dysthymic patients which were characterized by a decreasing proportion of deaths due to pulmonary diseases and an increasing proportion of deaths due to cardio-vascular disease. Suicide was proportionally more frequent in both dysthymic and schizophrenic patients. The classic hypothesis that schizophrenics are protected against cancer was not confirmed. The limits of the proportional approach (PMR) are discussed.