[Estimation of the excess death associated with influenza pandemics and epidemics in Japan after world war II: relation with pandemics and the vaccination system]

Abstract

Objectives: To estimate the excess death associated with influenza pandemics and epidemics in Japan after World War II, and to reexamine the relationship between the excess death and the vaccination system in Japan.

Methods: Using the Japanese national vital statistics data for 1952-2009, we specified months with influenza epidemics, monthly mortality rates and the seasonal index for 1952-74 and for 1975-2009. Then we calculated excess deaths of each month from the observed number of deaths and the 95% range of expected deaths. Lastly we calculated age-adjusted excess death rates using the 1985 model population of Japan.

Results: The total number of excess deaths for 1952-2009 was 687,279 (95% range, 384,149-970,468), 12,058 (95% range, 6,739-17,026) per year. The total number of excess deaths in 6 pandemic years of 1957-58, 58-59, 1968-69, 69-70, 77-78 and 78-79, was 95,904, while that in 51 'non-pandemic' years was 591,376, 6.17 fold larger than pandemic years. The average number of excess deaths for pandemic years was 23,976, nearly equal to that for 'non-pandemic' years, 23,655. At the beginning of pandemics, 1957-58, 1968-69, 1969-70, the proportion of those aged <65 years in excess deaths rose compared with 'non-pandemic' years. In the 1970s and 1980s, when the vaccination program for schoolchildren was mandatory in Japan on the basis of the "Fukumi thesis", age-adjusted average excess mortality rates were relatively low, with an average of 6.17 per hundred thousand. In the 1990s, when group vaccination was discontinued, age-adjusted excess mortality rose up to 9.42, only to drop again to 2.04 when influenza vaccination was made available to the elderly in the 2000s, suggesting that the vaccination of Japanese children prevented excess deaths from influenza pandemics and epidemics. Moreover, in the age group under 65, average excess mortality rates were low in the 1970s and 1980s rather than in the 2000s, which shows that the "Social Defensive" schoolchildren vaccination program in the 1970s and 1980s was more effective than the "Individual Defensive" vaccination program in the 2000s.

Conclusion: Excess deaths were observed continually, and not limited to pandemic years. We must not slight public health interventions for 'non-pandemic' influenza as well as pandemic influenza. We should also re-examine the importance of "Social Defenses", including preventative vaccination, for public health policy.