Association of BCG vaccination policy with prevalence and mortality of COVID-19

Giovanni Sala; Tsuyoshi Miyakawa

Nevertheless, most people in Japan, with the notable exception of Tokyo metropolitan, are currently working as usual. Surprisingly, despite such a liberal measure taken and relatively low number of PCR tests conducted, Japan is ranked at 117 th and 69 th among 199 countries and regions 1 , as of March 28th, regarding the number of cases per 1 million population and the number of deaths per 1 million population, respectively. Also, there seem to be visible differences in COVID-19 prevalence and mortality between countries in Western Europe and Eastern Europe. There are many potential geographical, social, and biological factors, such as temperature, humidity, life expectancy, average income, social norms, and ethnical genetic background, that could potentially explain such variabilities among countries, but it remained fully explored. In case they are successfully identified, such information could potentially contribute to the mitigation of the spread of the virus. Some vaccines (e.g., measles and Bacillus Calmette-Guérin (BCG) vaccines) are associated with a lower risk of illness and death from other disorders [2] [3] [4] . A systematic review of epidemiological studies provided evidence for the non-specific beneficial effects of BCG vaccine on all-cause mortality 5, 6 . Induction of cytokines associated with trained immunity is proposed to underly such protection against non-related viral infections 7 . Researchers in four countries will soon start a clinical trial of BCG vaccine on this disease, based on the proposed beneficial effect of BCG 4 . Interestingly, a blogger noticed that there might be a correlation between the administration of BCG vaccine in each country and the extent to which the spread of COVID-19 occurs 8 . In this study, we evaluated the hypothesis that BCG vaccination has some protective effect against COVID-19 by utilizing publicly available datasets through linear regression modeling. Among the potentially confounding factors, we assessed two factors, life expectancy and the average . CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.