Time series analysis of age-sex specific death rates from aplastic anemia and the trend in production amount of chloramphenicol.

Abstract

To explore the epidemiological features ofaplastic anemia, the mortality statistics (1958-1978) in Japan were examined by age and sex. Mortality trends by age were fitted by the regression curves derived from Akaike Information Criteria Method and correlated to the trend'in annual per capita production amounts of chloramphenicol (CP), taking into consideration the time-lag between the production of CP and the aplastic anemia deaths. A gradual decrease in age-specific death rates was apparent from 1960 onwards in males aged less than 45 years. In males aged 45 years and over, the year in which the mortality started to decrease apparentiy shifted to much more recent years with advancing age groups. Quite similar mortality trends by age were observed in females, though the year in which the mortality began to decrease shifted more quickly to recent years in females than in males. Annual per capital production amount ofCP had a negative correlation in the younger age groups, particularly in males, but a significantly positive correlation in the advanced aged population, though the largest correlation coefficients were obtained with different time-lags in different age groups. The time-series analysis suggested that CP appeared to be hazardous to the aged population, but this was not fully evidenced by the present analysis.