Background/purpose COVID-19 mRNA vaccines have been reported to reduce severe COVID-19 outcomes in high-risk populations. In Japan, booster vaccinations up to the seventh dose are publicly funded; however, evidence regarding the safety and potential benefits of repeated booster doses in younger adults remains limited. This exploratory observational study aimed to describe age-specific patterns of all-cause mortality according to the number of COVID-19 vaccine doses using municipal administrative registry data. Methods Municipal registry data obtained through information disclosure requests from Hamamatsu City (2021–2024) and Matsudo City (2021–2025) were analyzed. The dataset included age group or year of birth (5-year increments), sex, date of residence, COVID-19 vaccination records (date and lot number), and date of death. Individuals aged 20–89 years were categorized into three age groups (20–49, 50–64, and 65–89 years). All-cause mortality rates were calculated per 100,000 person-years according to the number of vaccine doses. Exploratory comparisons between dose groups were conducted using Poisson tests. Because the registry dataset lacked information on comorbidities, healthcare utilization, socioeconomic status, and cause of death, multivariable adjustment and causal inference were not feasible. Results Age-specific differences in all-cause mortality rates were observed across vaccine dose groups. Among individuals aged 65–89 years, higher numbers of vaccine doses were associated with lower all-cause mortality rates. In contrast, among those aged 20–49 years, all-cause mortality rates were higher among individuals who had received ≥5 doses than among those who had received fewer doses. Among individuals aged 50–64 years, the all-cause mortality rate was higher among those who had received six doses than among those who had received 1, 3, or 4 doses. Conclusions This exploratory analysis identified age-specific differences in all-cause mortality across COVID-19 vaccine dose groups. However, the dataset available so far lacked important clinical covariates and information on causes of death, limiting causal interpretation. These findings highlight the need for constructing in Japan nationally linked datasets that allow adjustment for comorbidities, healthcare utilization, and other potential confounders.
IWAMOTO et al. (Mon,) studied this question.