Overmortality among psychiatric patients has been a regular observation from the XIXth century to nowadays. If the rates of mortality have decreased in these last fifty years, they still remain higher than in the general population Standardized Mortality Ratio (SMR) (two or three times greater than normal). The authors reviewed 53 recent articles from european, north american, japanese and israelian researches on this question. Most of them proceed from crossings between psychiatric case register and death-register, and concern inpatients only (Brook, Giel, Saugstad, Mortensen, Herman, Haugland, Rorsman, Sturt, Winokur, Zilber). Some take into account outpatients as well (Eastwood, Koranyi, Martin, Ribourdouille). SMR are calculated and comparisons are made between different groups by sex, age, diagnosis and cause of death. Those are usually divided into 2 categories: natural deaths somatic diseases) and unnatural deaths (suicides-accidents). RESULTS Among the patients, mortality rates are higher for men than women, but SMR are higher for women. The highest mortality relative risk is observed between 20 and 40 years of age. Except for three authors, unnatural deaths are not sufficient to explain overmortality. SMR for suicides and accidental deaths are decreasing with age; the relative risk is more important for outpatients, men, some specific diagnoses (affective disorders, acute schizophrenia) and during the first two years of the course of the illness. Suicide rates have been increasing among patients these last twenty years. Natural death is more frequent among patients with organic brain syndromes but is also in excess for the other patients. Cardiovascular diseases represent the first cause of mortality but infections (pneumonia-influenza) and metabolic diseases are over-represented.(ABSTRACT TRUNCATED AT 250 WORDS)