A retrospective study of the autopsy rate of the clinico-pathological correlations was made in a group of inpatients. The trend of autopsy rate was observed in all inpatients died in a university hospital during a nine year period (from 1975 to 1983). The agreement between the clinical and pathological death diagnosis was retrospectively controlled in 294 consecutive patients died and submitted to autopsy in the same hospital during a 6 month period (from January to June 1983). The results show that: In the oldest (greater than 60 yrs) patients, the autopsy rate trend was significantly reduced from 1975 to 1983, whereas the hospital admissions and the mortality rate increased. In the youngest subjects (less than 60 yrs), the autopsy rate trend, the hospital admissions and the mortality rate did not significantly change from 1975 to 1983. In all the years considered, the autopsy rate was significantly reduced in the oldest class (p less than 0.001). The agreement between the clinical and pathological diagnosis was observed in 83% of cases in the less than 60 years class and in 63% of cases in the greater than 60 yrs class (p less than 0.001). The association of main disease with other diseases was significantly more frequent in the greater than 60 yrs class with respect to the less than 60 yrs class (p less than 0.01). It has been concluded that the autopsy, especially in the elderly, is absolutely necessary for a better quality control of the clinical diagnosis and of the medical care.