Objective - The present study was conducted to determine the excess of mortality in long-term psychiatric inpatients. Setting - Psychiatric hospital, Bologna, Italy. Design – Retrospective study on a cohort of 264 chronic psychiatric inpatients, followed up for ten years since January 1981. Mortality measures were computed by psychiatric diagnosis, length of stay and «still resident at follow-up/discharged» condition. Main measures – Standardized mortality ratios were computed with reference to the mortality by age of the population resident in Bologna in the years 1981- 1991. Presence in the cohort was expressed in person/years. Results – The excess of deaths in the cohort was 2.89 (CI 95%: 2.12-3.86) for males and 2.38 (CI: 1.66-3.31) for females. A remarkable excess of deaths for both sexes was observed in the class of age 15/44 (18.33 for males, CI 9.14-32.81; 60 for females, CI: 21.91-130.6 ) and in association with a diagnosis of personality disorder/alcohol abuse (11.25 for males, CI: 5.13-21.36; 5 for females, CI: 1.35-12.8). Length of stay and age of patients were inversely associated with SMRs. Conclusion – The present study confirms that the excess of mortality is high in psychiatric long-term inpatients and that it may be concentrated in younger classes of age, first years after admission and in subjects with diagnoses of personality disorder/alcohol abuse.