Age-period-cohort analysis on tuberculosis cases in Japan, 1953-2022

Age-period-cohort (APC) analysis has been employed to differentiate long-term trends in the incidences of communicable diseases, including tuberculosis (TB), into the effects of age, birth year, and calendar period. However, no such study was hitherto conducted for Japan, which has 70 years of surveillance data. Therefore, we conducted APC analysis for TB in Japan.

The national TB data for 1953-2022 were analyzed using the log-transformed linear model of APC analysis.

Annual age-and sex-standardized notification rates of TB peaked at 599.0 per 100 000 population in 1955 and fell by 99% to 4.5 in 2022. Adjusting for the effects of the birth cohort and period, the relative age-effect risk of TB peaked at 20-29 years and went down toward 60-69 years. Regarding the birth cohort effect, the TB risk showed a turning point in approximately 1913 for the central years of birth. Another change appeared in 1963 when the decline of the risk slightly stagnated; then, it started declining again at a rate as fast as in 1923-1953. Period effects showed a hump in the late 1950s and early 1960s, then sharply declined to the late 1970s, and reached a near plateau level until 2022.

Our results highlight the continuing peak in TB disease risk for young adults and sharp decrease in disease risk in the 1960s and 70s. The introduction of anti-TB drugs in the 1950s and early 1970s had the most important impact on the epidemiology of TB in Japan.