japan contact: kamono.mana@ gmail.c om · 4/17/2020 · d ph. ara, h zu asa a masak 1 1, 5 9 1 0-0...
TRANSCRIPT
-
The effect of BCG vaccination on COVID-19 examined by a statistical
approach: no positive results from the Diamond Princess and cross-national
differences previously reported by world-wide comparisons are flawed in
several ways
Masakazu Asahara, Ph. D.1 1Division of Liberal Arts and Sciences, Aichi Gakuin University, Nisshin, Aichi 470-0195,
Japan
Contact: [email protected]
Abstract
Recently, the controversial hypothesis that past BCG (Bacillus
Calmette–Guérin) vaccination reduces infection or severity of COVID-19 has
been proposed. The present study examined this hypothesis using statistical
approaches based on the public data. Three approaches were utilized: 1)
comparing the infection and mortality ratio of people on the cruise ship
Diamond Princess, 2) comparing the number of mortalities among nations,
and 3) comparing the maximum daily increase rate of total mortalities
among nations. The result of 1) showed that there is no significant difference
in infection per person onboard or mortality-infection between Japanese
citizens vs. US citizens and BCG obligatory nations vs. non-BCG obligatory
nations on the Diamond Princess. The result of 2) showed that the number of
mortalities among nations is similar to the previous studies, but this
analysis also considered the timing of COVID-19 arrival in each nation. After
correcting for arrival time, previously reported effect of BCG vaccination on
decreasing total mortality disappeared. This is because nations that lack
BCG vaccination are concentrated in Western Europe, which is near an
epicenter of COVID-19. Therefore some previous reports are now considered
to be affected by this artifact; the result may have been flawed by dispersal
from an epicenter. However, some results showed weakly significant
differences in the number of deaths at a particular time among BCG
obligatory and non-BCG nations (especially the use of Japanese BCG strain
Tokyo 172). However, these results are affected by the results of three
countries and the effect of BCG vaccination remains inconclusive. The result
of 3) showed that the maximum daily increasing rate in death among nations
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
showed no significant difference among BCG vaccination policies. In the
present study, although some results showed statistically significant
differences among BCG vaccination policies, they may be affected by the
impact of various other factors, such as national infection-control policies,
social distancing, behavioral changes of people, possible previous local
epidemics of closely related viruses, or inter-population differences in ACE2
or other genetic polymorphism. Further research is needed to better
understand the underlying cause of the observed differences in infection and
mortality of the disease among nations. Nevertheless, our results show that
the effect of past BCG vaccination, if any, can be masked by many other
factors. Therefore, the possible effect might be relatively small. In fact, in
Japan, where almost all citizens have been vaccinated, COVID-19 cases are
constantly increasing. Given the importance of people’s behavior in
preventing viral propagation, the spread of optimism triggered by this
hypothesis would be harmful to BCG vaccination nations.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
Introduction
In April 2020, the global pandemic of coronavirus disease-2019
(COVID-19) caused by severe acute respiratory coronavirus 2 (SARS-CoV-2)
is a serious problem all over the world. Recently, several authors have
suggested the hypothesis that past BCG (Bacillus
Calmette-Guérinvaccination) vaccination reduces the morbidity and
mortality of COVID-19. Several authors reported the correlation between
national BCG vaccination policy and infection or mortality rate (or its
increasing rate) in those nations (Akiyama and Ishida 2020; Berg et al. 2020;
Dolgikh 2020; Goswami et al. 2020; Miller et al. 2020; Sala and Miyakawa
2020; Shet et al. 2020), and some have argued that the hypothesis is
supported. However, several authors failed to account for how infectious
diseases spread from one place to another. That is, several authors simply
compared the total number of deaths in each nation without considering the
timing disease arrival in each nation (e.g. Miller et al. 2020; Sala and
Miyakawa 2020), suggesting that past BCG vaccination changes the number
of infection and death by double or triple digits (e.g. they simply compared
16523 deaths until April 6 in Italy which BCG- vs. 47 deaths until April 6 in
Russia which BCG+, where the timing of the pandemic enter to the nation is
different). This shortcoming would produce critically important
misunderstandings of the effect of BCG vaccination on COVID-19. In fact,
mandatory BCG vaccination is discontinued in Western Europe, which is
near a COVID-19 epicenter. This should affect the conclusions of previous
studies. Therefore, the present study attempted to examine the effect of BCG
vaccination while accounting for these biases.
Three approaches were used in the present study. The first approach is a
test in the cruise ship Diamond Princess. In this ship, most patients were
infected before they were aware that the disease was spreading in the ship.
Therefore, the cultural effect of countries (such as wearing masks) or
national policy (such as the “cluster buster” policy in Japan) can be excluded.
The strain of virus on the ship is the same, and the spread occurred
simultaneously on board (Sekizuka et al. 2020). The second approach is a
similar comparison as in many previous studies (e.g. Miller et al. 2020; Sala
and Miyakawa 2020). That is, the number of mortalities in each nation was
compared, but with consideration for the timing of disease arrival. The third
approach is also similar to a previous study (Akiyama and Ishida 2020). That
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
is, the rate of increase of the number of mortalities in each nation was
compared.
Materials and Methods
Data was obtained from previous publications or public databases. The
number of cases, deaths, and passengers’ nationality on the Diamond
princess was obtained from public data (some districts within the nation
were treated separately) (Ministry of Health, Labour and Welfare, Japan,
2020a; b) and a previous report (Moriarty et al. 2020). BCG vaccination
policies were obtained from The BCG World Atlas (Zwerling et al. 2011;
http://www.bcgatlas.org/). The nationalities of dead passengers were
obtained from each news report.
Data on cases and mortality from each nation was obtained from a public
database (some districts within the nation were treated separately) (Dong et
al. 2020; Johns Hopkins University). The number of international tourist'
arrivals in 2017 was obtained from a previously published report (World
Tourism Organization 2019). Some data (BCG vaccination policy, population,
and life expectancy) were obtained from the supplementary material of a
previous study (Sala and Miyakawa 2020). Some BCG strain information
was obtained from other literature (Akiyama and Ishida 2020; Joung and
Ryoo, 2013).
In the analysis of the Diamond Princess, the effects of BCG policy and
nationality were examined using the Fisher’s exact test. If the BCG
vaccination changes the infection or mortality rate by two or three digits as
previous studies suggested, a power analysis indicated that the sample size
for the analysis of infection rate may be sufficient for 90% power on the
analysis (e.g., n=62 is sufficient to detect a difference between the incidence
rates of 0.2 and 0.02 in each group). With respect to the mortality rate, the
sample size may meet 50 to 70% of the power on the analysis (e.g., 70%:
n=124 is sufficient for 70% power to detect a difference between the
incidence rates of 0.05 and 0.0005; 60%: n=99, 0.05 and 0.0005; 50%: n=394,
0.01 and 0.0001; n=78, 0.05 and 0.0005, respectively).
To compare the cases and deaths among nations, the timing of the fifth
death was used to align the timing of disease entry to the nation. Although
Shet et al. (2020) used the timing of the 100th case, this alignment is
considered better because the specific number of the case depends on the
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
national policy of examination. In addition, the present study focused on the
total number of deaths in each nation, rather than death per population,
because the pandemic was severe in particular cities or particular areas. In
other words, it does not make sense to compensate for the spread of infection
in Hubei province with the population of Beijing and Shanghai. This
relationship is the same in Moscow and Siberia, and the same in Tokyo and
Iwate. Moreover, national political decisions to prevent disease spread,
which is one of the most important factors of disease spread, should be
affected by the total number of victims. Therefore, aligning to the total
national population is irrelevant. Moreover, almost all people are still
considered susceptible to the disease (even in small nations such as San
Marino). In some nations, such as China, the increase in the number of
deaths was stopped or flattened. Therefore, cumulative deaths 30 days after
the date fifth death (this treatment is based on Akiyama and Ishida 2020)
was used. The general linear model was used to test the effect of past BCG
vaccination, the timing of pandemic entry to the nation, and region (e.g.,
East Asia, Europe, etc.). Numbers of death were log-transformed. The region
may serve as rough indexes of social custom, different ratios of genetic
polymorphisms, or the distribution of wildlife. For some results, nations
using the Japanese BCG strain (Tokyo 172) were separated from other BCG
strains because some reports have suggested the Japanese strain is more
effective (Akiyama and Ishida 2020).
To compare the increasing rate of death in each nation, daily increases in
the number of total deaths were calculated. The present study used two
weeks average of the daily rate: Average rates were calculated every two
weeks and the highest value was used as the maximum increase rate per day.
Comparisons were performed using nations where the total number of
deaths was over 25. The effect of BCG policy and region was examined using
the ANOVA-Tukey test. A posthoc power analysis indicated that if the
between-group variance was over 0.0075 (Japanese strain unseparated) and
0.021 (Japanese strain separated), the sample size is sufficient for 90%
power on the analysis. These variances are comparable to within-group
variances. The data of Serbia may not be completed, it was excluded from
this analysis. In this study, all statistical analyses were performed by
Minitab 18 (Minitab Inc, USA) except for the power analyses which were
performed by R (R Core Team, Vienna, Austria). The data and calculations in
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
the spreadsheet are shown in the supplementary data.
Results and Discussion
The Diamond Princess shows no positive perspective for a personal-level
effect of BCG vaccination
The results of the Diamond Princess are shown in Table 1 and 2. The
case-fatality ratio was not significantly different between people from BCG
mandatory or not mandatory nations even when the analysis was limited to
the elderly (Table 1). The infection and mortality ratio were not significantly
different between Japanese citizens (BCG+) and US citizens (BCG-) (Table 2).
The mortality rate was not significantly different between passengers from
several countries (Table 2).
To discuss the result, Japanese citizens under age 98 should have
received BCG vaccination; in 1951 all infants and citizens under 29 were
inoculated BCG vaccine if their tuberculin test was negative (Ministry of
Health, Labour and Welfare, Japan, 2016). According to the estimated rate of
tuberculosis infection, 12.0% of five year old children and 24.6% of ten year
old children were infected with tuberculosis in 1950 (Omori 2009). Therefore,
80-90% of 70-79 years old people were considered to be inoculated with the
BCG vaccine in Japan. While most Japanese victims on the Diamond
Princess were over 70 years old (except for two victims whose age has not
been announced), most had been inoculated with the BCG vaccination.
However, the infection rate and mortality rate were not different
significantly from that of other people onboard from BCG negative nations.
It should be noted that there was 1045 staff among the 3711 people on the
Diamond Princess (National Institute of Infectious Diseases, 2020). Staff
may be younger than passengers, and the mortality rate should be relatively
low. However, when the analyses focused on older people or passengers only,
the conclusion was not changed (Table 1 and 2). In addition, at that time, the
Diamond Princess traveled from Hong Kong to Japan. Therefore, East Asian
people (from BCG+ countries) could have easily boarded the cruise, even if
they had pre-existing diseases. However, passengers from Western Europe,
the United States, and Australia may have been restricted to individuals
without pre-existing diseases due to the long flight required to reach the
cruise. In addition, 8 people remain in the hospital as of 2020 April 27
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
(Ministry of Health, Labour and Welfare, Japan, 2020a). These factors may
affect the result. However, it should be emphasized that the Diamond
Princess presented no positive support for the hypothesis that past BCG
vaccination reduces the infection and mortality of COVID-19.
Cross-national comparison showed that previous reports were flawed due to
the timing of spread, yet there is still a weakly significant result in
particular comparisons
The bivariate plots show the results of the cross-national comparison
(Table 3 and 4; Fig. 1, 2, 3, 4, and 5). The result indicated that the number of
deaths is seriously affected by when the disease first spread to the nation (i.e.
date of the fifth death) (Table 3 and 4). In addition, it is also significantly
affected by the region which is used as an index of social custom, different
ratios of genetic polymorphisms, or the distribution of wildlife. For the total
number of deaths, when the period after pandemic entry was fixed, the BCG
effect disappeared (Analysis A and C: Table 3). In contrast, when analyzing
the Japanese BCG strain separately, BCG policy was not significant in
analysis B, but was significant in analysis D (Table 3). This result may be
due to three nations (Japan, Iraq, and South Korea) which showed relatively
lower mortality and use the Japanese BCG strain. Therefore, unique
characteristic effects of those nations may affect the result. For example,
public security problems in Iraq might force people to stay home. The South
Korean government examined many patients and succeeded in controlling
the pandemic. The Japanese government adopted a unique strategy of
“cluster buster” based on the previous result that the “super-spreader” (an
infected person who spreads the virus to many other people) is restricted in
SARS-CoV-2 (Nishiura et al. 2020). In addition, governmental requests for
social distancing may work in Japan (on April 11, the number of train
passengers in Shibuya station, Tokyo reduced 98% from one year ago, except
for those using commuter’s ticket: TBS News 2020). The same metrics
showed 66-89% of passenger reduction at other major stations in Tokyo that
weekend (Cavinet Secretariat, Japan 2020). It should be noted that South
Korea had been using the Danish BCG strain, but later changed to the
Japanese strain Tokyo 172 (Akiyama and Ishida 2020). Among other nations
examined, China and the Philippines, which are BCG+, showed relatively
fewer deaths (Fig. 1 and 2). These nations are known to choose strong
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
policies to prevent transfer among cities. Australia, which is a BCG- nation,
also showed a similar pattern to Japan, Iraq, and South Korea (Fig. 1 and 2)
although the cause of this is not clear. The pattern of San Marino may be
explained by its smaller population. All nations possess particular factors
that affect their results.
The timing of disease entry seems to correlate to number of tourists (Fig.
3). In addition, political and economic relationships to China might affect
disease entry such as in Iran and South Korea. Therefore, the international
flow of people might affect the timing of pandemic entry and the number of
deaths. However, focusing on nations whose deaths are relatively small,
there seems to be less relationship between the number of tourists and total
deaths (Fig. 5). Some nations using the Japanese BCG strain such as
Thailand, seemed to succeed in preventing transmission despite the number
of tourists. However, some BCG- nations, such as Australia and New Zealand,
also seem successful in preventing the pandemic. In addition, most nations
where the disease spread has so far been prevented are BCG+ nations. These
nations are located outside of Europe, which might flaw previous studies (Fig.
5).
Although previous reports suggested that past BCG vaccination reduces
infection and mortality (e.g. Miller et al. 2020; Sala and Miyakawa 2020), the
present study suggests that these previous results were biased by the route
of pandemic spread and the timing of pandemic entry to the nation. In other
words, the fact that some nations in Western Europe became an epicenter of
the pandemic as a stochastic event is an important factor. This is clearer in
the analysis using only European nations (Table 4; Fig. 4).
The present study did not consider GDP, latitude, or temperature, etc.
However, the results indicate that even if past BCG vaccination may have an
effect on COVID-19, it can be easily masked by other factors. In fact, the
timing of pandemic entry had much larger effect on the observed
cross-national difference than the BCG policy did (Table 3 and 4). Therefore,
the author considers that the effect of BCG, if any, may not be potent as
previous reports suggested; the result does not support that the past BCG
vaccination reduces the number of infection and death changes by double or
triple digits (e.g. Miller et al. 2020; Sala and Miyakawa 2020).
BCG may possess little or no effect on the increase rate of deaths in countries
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
where the disease spreading
The result of the increase rate of deaths is shown in Fig. 6. The real
increase graph is shown in Fig. 7. There is no significant difference between
BCG+ or – nations (Fig. 6). However, when analyzing the Japanese BCG
strain separately, the result was weakly significant (Fig. 6). The result may
be affected by sample size (the comparison was performed on 55 nations).
The increase rate decreases from BCG –, BCG – (past +), BCG +, and BCG +
(Japanese strain) nations. However, there are several outliers and large
regional differences. In addition, few nations affected the result.
Furthermore, the timing of the pandemic entry may affect because people’s
behavior may have gradually changed. Therefore, the effect of BCG
vaccination is not clear. It should be noted that another study reported a
highly significant correlation between BCG vaccination and the increasing
rate (Akiyama and Ishida 2020). This difference may be attributed to the
previous study focusing on the average increase rate whereas the present
study is focused on the maximum increase rate. The author considers that
the maximum increase rate indicates potential of disease increase and may
be more critical to pandemic control.
Possible factors affecting the results and the remaining possibility of the
BCG hypothesis
The hypothesis that past BCG vaccination affects COVID-19 can be
separated into three. The first hypothesis is regarding reducing personal risk
of infection or mortality. The present study partially tested this using the
data of 1) the Diamond Princess and did not show positive result. Another
study also reported a negative result by comparing infection rate of BCG+
and BCG- age people in several nations (Fukui et al. 2020). The second
hypothesis is regarding reducing the total morbidity and mortality in the
population (e.g. Miller et al. 2020; Sala and Miyakawa 2020). The present
study did not show a positive perspective regarding this aspect, as seen by
the result of 2) the cross-national comparisons. The third hypothesis is about
reducing the rate of propagation in the population (e.g. Akiyama and Ishida
2020). The present study did not show clear results about this. As discussed
above, although the possibility of undetectable effect cannot be ruled out,
none of the hypothetical effects of past BCG vaccination on COVID-19 are
supported in this study.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
The hypotheses of BCG have emerged from previous reports that BCG
has non-specific provocation of natural immunity (termed “trained
immunity”; reviewed by e.g. Covián et al. 2019; Angelidou et al. 2020). This
mechanism is very attractive, although the present study did not show a
positive result. However, if this mechanism is effective, then the effects of
BCG on COVID-19 may not differ in high magnitude among BCG strains.
T-cell epitopes of BCG have been estimated previously (Zhang et al. 2013).
According to the previous study (Zhang et al. 2013), the Japanese BCG
strain possesses more epitopes than other strains. The author preliminary
examined whether SARS-CoV-2 possesses homologous amino acid sequences
of these BCG epitopes using BLAST. However, no homologous sequence was
observed (Personal Observations; Fig. 8). If the Japanese strain had a larger
effect on the disease, it would not be due to affinity to T-cell receptors. A
possible mechanism could be that a domain or constituent unique to the
Japanese strain has a high affinity to Toll-like receptors or some other
receptors for innate immunity and leads to strong trained immunity (Fig. 8).
However, according to the results of the Diamond Princess, this is rather
unlikely.
The cross-national difference of morbidity and mortality can be
influenced by many factors. A schematic illustration is shown in Fig. 8. A
possibility explained above is that several nations have unique reasons that
reduce disease propagation. Each nation’s policy in response to the pandemic,
cultural differences, or lifestyle differences should be considered.
Additionally, as our cross-national comparisons showed the region
significantly affected the result (Table 3), there can be other hypotheses.
The author could suggest another possible hypothesis that several
nations already experienced local epidemics of similar viruses. Wild bats and
civets possess viruses related to SARS-CoV-2 (e.g., bat/ civet SARS
coronaviruses) (e.g. Guan et al. 2003; Lau et al. 2005; Li et al. 2005; Song et
al. 2005; Wu et al. 2016; Luk et al. 2019; Lam et al. 2020). Epitopes of
SARS-CoV-2 were suggested by a previous study (Ahmed et al. 2020). The
author preliminarily examined whether SARS-related viruses possess
homologous amino acid sequences of these SARS-CoV-2 epitopes using
BLAST. There are many homologous sequences in bat and civet SARS
coronaviruses (Personal Observations). Rhinolophus bats are one of the wild
hosts of these viruses and distribute in South Europe, Middle East, South
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
Asia, South East Asia, East Asia, and Oceania. Another host is the masked
palm civet (Paguma larvata) distributes in East Asia. If there were local
epidemics of these related viruses manifested as a slight cold in several
particular nations, many people may have acquired immunity to these
viruses, hence explaining the different morbidity and mortality patterns
between nations. A large scale antibody test may address these hypotheses.
In addition, cross-immunity to normal colds should also be considered. In
Japan, there is a culture that people go to work even if they have a cold.
Therefore, a high percentage of the working generation may have already
contracted the common cold. This may have an impact on the situation in
Japan. However, this effect will not work for the second wave predicted in
the near future, as immunity of normal cold does not last long.
The author could suggest one more possible hypothesis being that human
genetic variation caused the difference between nations. As other SARS-like
coronaviruses, SARS-CoV-2 may use the angiotensin-converting enzyme II
(ACE2) protein (Ge et al., 2013; Hoffman et al. 2020) and the
transmembrane serine protease (TMPRSS2) protein (Matsuyama et al. 2010;
Glowacka et al. 2011; Shulla et al. 2011) for infection. Some reports have
suggested that ACE2 variants (Elisa et al. 2020) or TMPRSS2 variants
(Asselta et al. 2020) can affect the severity of COVID-19. A previous report
suggested that the proportion of ACE2 polymorphisms is different among
populations; the ratio of some allele variant gradually decreases from Europe
> America > Africa, South Asia> East Asia (including China) (Cao et al. 2020).
Geographic variation of MHC class I (HLA) gene whose susceptibility the
COVID-19 may be different was also reported (Nguyen et al. 2020). These
patterns may explain the relatively low mortality and increase rate in East
Asian nations. Perhaps this geographic pattern may have been caused
through natural selection by a possible historical epidemic of SARS related
viruses in East Asia or other regions. Given that historical contact with
livestock may have been important to resistance to some kind of infectious
disease (Diamond 1997), increased resistance to another kind of zoonosis can
be observed in areas where are close to biodiversity hotspots. Perhaps the
resistance could be a cultural or political propensity.
The author could suggest another possible hypothesis of polymorphisms
affecting the genetic ability of disease spread. Although many studies focus
on susceptibility of the disease, ability of disease transmission to others
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
would be necessary to examine as a factor on the regional difference. In fact,
several results suggest relatively low difference of susceptibility among
human populations (Table 1 and 2). Previous studies have suggested that
almost all infected persons do not spread COVID-19, but rather only a
selected few people spread the disease (Nishiura et al. 2020; Endo 2020). In
addition, a previous study reported that viral load in the saliva correlated
with age (To et al., 2020), but their result seems to show high variance
among individuals. If the ability of disease spread varies among individuals
and the proportion of “super-spreaders” is different among nations, the
previously reported spread pattern (Nishiura et al. 2020) and the different
increase rates of victims among nations can be explained. The author
suggests two patterns of hypothetical causation. The first is that the degree
of expression of ACE2 or other related genes in the salivary gland or
intraoral organs varies. If the expression is high, the number of viruses
increases in the saliva. COVID-19 may infect through droplets (Lai et al.
2020) possibly including those blown out during conversation. Therefore, this
hypothesis is feasible. In addition, the amount or viscidity of saliva, and
morphological character of oral organs might affect disease spread. However,
to prove this hypothesis, an intensive examination of the route of infection
and disease spread would be needed.
As discussed above, there can be many hypotheses explaining the
differences of COVID-19 morbidity, mortality, and how those increase among
nations. Given that the effect of the region was significant in analysis 2,
simple comparisons among nations may fail to address these
problems/hypotheses.
Although there are many things to do before proving the BCG hypothesis,
many mass media outlets are reporting the hypothesis, causing people to
gradually consider the hypothesis to be true (especially in BCG+ nations
such as Japan). Some people even suggested that no special protective
measure is needed in Japan. The author seriously worries about this
problem because the change in people’s behavior is the most important
avenue to prevent viral spread (Flaxman et al. 2020; Ministry of Health,
Labour and Welfare, Japan, 2020a; Zhang et al. 2020). Optimism caused by
the spread of this hypothesis is a serious problem. We should be cautious
that in Japan, even though almost all citizens received BCG vaccination
(using the Japanese strain, which may be most effective), cases and deaths
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
are constantly increasing at the beginning of April, 2020.
The effect of past and recent vaccination should not be confused
The present study focuses on past vaccination (such as that in many
decades ago) as in the previous cross-national comparative studies. In the
meantime, clinical studies examining the effects of recent vaccination to
adults are currently being conducted (e.g. reported by Vrieze, 2020). However,
it is dangerous to confuse the effects of recent vaccination with the past
vaccination (Asahara, submitted). Even if the cross-national comparison
results were positive, it cannot prove the effect of recent vaccination because
inoculation to infants may be necessary for effectiveness (e.g. low efficacy of
adult inoculation for tuberculosis resistance: Mangtani et al. 2014). Of course,
the cross-national comparisons have too many factors to consider, and that
alone is not sufficient to prove the BCG effect. On the other hand, even if the
results of a clinical study are positive, it cannot prove that the effects will
last for decades. In other words, the result of the clinical trials cannot prove
that past BCG vaccination is responsible for the cross-national differences in
cases and deaths. Other studies would be required such as examining the
previous vaccination history of infected people in a country where
vaccination is optional. However, clinical studies would provide more reliable
data on the effect of BCG vaccination on the COVID-19 than cross-national
comparisons.
Conclusions
The hypothesis that BCG vaccination reduces the infection and mortality
of COVID-19 is attractive. However, previous international comparative
reports may not prove the hypothesis because many other possibilities can
explain the observed pattern. In addition, the present study did not show a
positive result. Clinical researches are currently being conducted (e.g.
reported by Vrieze, 2020). These ongoing clinical studies should provide a
better understanding of this hypothesis. Until then, we should be careful
about patterns observed in the statistical data. Moreover, spreading the
optimistic view triggered by the hypothesis is rather harmful.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
Acknowledgement
The author thanks anonymous readers who commented on the
preprint version of the manuscript (published in April 22, 2020). English of
the some part of the manuscript have been corrected by courtesy of Editage
service.
References
Ahmed S. F. et al. 2020. Preliminary identification of potential vaccine targets for the
COVID-19 coronavirus (SARS-CoV-2) based on SARS-CoV immunological studies.
Viruses 12: 254.
doi:10.3390/v12030254
Akiyama Y. and Ishida T. 2020. Relationship between COVID-19 death toll doubling
time
and national BCG vaccination policy. (preprint)
http://www.bi.cs.titech.ac.jp/COVID-19/Death_vs_BCGpolicy.html?fbclid=IwAR1-3au
eH73F0HIBZeWLsJXXNVKQ56P7DaX5a20iO1eCNbjPBWoHC7q%E2%80%A6
Angelidou A. et al. 2020. BCG as a Case study for precision vaccine
development: lessons from vaccine heterogeneity, trained immunity, and
immune ontogeny. Frontiers in Immunology 11: 332.
doi: 10.3389/fmicb.2020.00332
Asselta R. et al. 2020. ACE2 and TMPRSS2 variants and expression as candidates to
sex and
country differences in COVID-19 severity in Italy.
medRxiv preprint doi: https://doi.org/10.1101/2020.03.30.20047878
Berg M. K. et al. 2020. Mandated Bacillus Calmette-Guérin (BCG) vaccination predicts
flattened curves for the spread of COVID-19.
MedRxiv preprint doi: https://doi.org/10.1101/2020.04.05.20054163.
Cabinet Secretariat. 2020. 駅の改札通過�数の推移(対前年�)[in Japanese]
https://corona.go.jp/toppage/pdf/area-transition/20200414_station.pdf
Cao Y. et al. 2020. Comparative genetic analysis of the novel coronavirus
(2019-nCoV/SARS-CoV-2) receptor ACE2 in different populations. Cell
Descovery 6: 11.
https://doi.org/10.1038/s41421-020-0147-1
Covián C. et al. 2019. BCG-induced cross-protection and development of
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
trained immunity: implication for vaccine design. Frontiers in
Immunology 10: 2806.
doi: 10.3389/fimmu.2019.02806
Diamond J. 1997. Guns, Germs, and Steel. W. W. Norton & Company inc., New York.
Dolgikh S. 2020. Further Evidence of a Possible Correlation Between the Severity of
Covid-19 and BCG Immunization.
medRxiv preprint doi: https://doi.org/10.1101/2020.04.07.20056994
Dong E., Du H., Gardner L. 2020. An interactive web-based dashboard to track
COVID-19 in real time. The Lancet.
https://doi.org/10.1016/S1473-3099(20)30120-1
Endo A. 2020. Estimating the overdispersion in COVID-19 transmission using outbreak
sizes outside China. Wellcome Open Research 5:67. (preprint)
https://doi.org/10.12688/wellcomeopenres.15842.1
Elisa B. et al. 2020. ACE2 variants underlie interindividual variability and
susceptibility to
COVID-19 in Italian population.
medRxiv preprint doi: https://doi.org/10.1101/2020.04.03.20047977
Flaxman S. et al. 2020. Estimating the number of infections and the impact of
nonpharmaceutical interventions on COVID-19 in 11 European countries. Imperial
College London (30-03-2020)
doi: https://doi.org/10.25561/77731.
Fukui M. et al. Does TB vaccination reduce COVID-19 infection? No evidence from a
regression discontinuity analysis. JEL I18: I15. (preprint)
Ge X.-Y. et al. 2013. Isolation and characterization of a bat SARS-like coronavirus that
uses the ACE2 receptor. Nature 503: 535-538.
Glowacka I. et al. 2011. Evidence that TMPRSS2 activates the severe acute respiratory
syndrome coronavirus spike protein for membrane fusion and reduces viral control by
the humoral immune response. Journal of Virology 85: 4122-4134.
Goswami R. P. et al. 2020. Interaction between malarial transmission and BCG
vaccination with COVID-19 incidence in the world map: A cross-sectional study.
medRxiv preprint doi: https://doi.org/10.1101/2020.04.03.20052563
Guan Y. et al. 2003. Isolation and characterization of viruses related to the SARS
coronavirus from animals in southern China. Science 302: 276-278.
Hoffmann M. et al. 2020. The novel coronavirus 2019 (2019-nCoV) uses the
SARS-coronavirus receptor ACE2 and the cellular protease TMPRSS2 for entry into
target cells.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
bioRxiv preprint doi: https://doi.org/10.1101/2020.01.31.929042
Lai C.-C. et al. 2020. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)
and coronavirus disease-2019 (COVID-19): The epidemic and the challenges.
International Journal of Antimicrobial Agents 55: 105924.
https://doi.org/10.1016/j.ijantimicag.2020.105924
Lam T. T.-Y. et al. 2020. Identifying SARS-CoV-2 related coronaviruses in Malayan
pangolins. Nature (online)
https://doi.org/10.1038/s41586-020-2169-0
Lau S. K. P. et al. 2005. Severe acute respiratory syndrome coronavirus-like virus in
Chinese horseshoe bats. PNAS 102: 14040-14045.
Li W. et al. 2005. Bats are natural reservoirs of SARS-like coronaviruses. Science 310:
676-679.
Luk H. K. H. 2019. Molecular epidemiology, evolution and phylogeny of SARS
coronavirus. Infection, Genetics and Evolution 71: 21-30.
Johns Hopkins University. Mapping 2019-nCoV.
https://github.com/CSSEGISandData/COVID-19
Mangtani P. et al. 2014. Protection by BCG vaccine against tuberculosis: a systematic
review of randomized controlled trials. Clinical Infectious Diseases 58: 470-480.
https://doi.org/10.1093/cid/cit790
Matsuyama S. et al. 2010. Efficient activation of the severe acute respiratory syndrome
coronavirus spike protein by the transmembrane protease TMPRSS2. Journal of
Virology 84: 12658-12664.
Moriarty L. F. et al. 2020. Public Health Responses to COVID-19 outbreakes of cruise
ships — Worldwide, February–March 2020. Morbidity and Mortality Weekly Report
69: 347-352.
https://www.cdc.gov/mmwr/volumes/69/wr/pdfs/mm6912e3-H.pdf
Miller A. et al. 2020. Correlation between universal BCG vaccination policy and reduced
morbidity and mortality for COVID-19: an epidemiological study.
medRxiv preprint doi: https://doi.org/10.1101/2020.03.24.20042937
Ministry of Health, Labour and Welfare, Japan, 2020a. About the new coronavirus
infection. (accessed in 2020 April 27)
https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/0000164708_00001.html
Ministry of Health, Labour and Welfare, Japan, 2020b. Diamond Princess Cases.
(accessed in 2020 April 27)
https://www.mhlw.go.jp/content/000598727.pdf
Ministry of Health, Labour and Welfare, Japan, 2016. Kekkaku ni kausuru tokutei
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
kansensyo yobo shishin ni Tsuite (結核に関する特定感染症予防指針について). [in
Japanese]
https://www.mhlw.go.jp/stf/shingi2/0000110066.html
National Institute of Infectious Diseases. 2020. Field Briefing: Diamond
Princess COVID-19 Cases.
https://www.niid.go.jp/niid/en/2019-ncov-e/9407-covid-dp-fe-01.html
Nguyen A. et al. 2020. Human leukocyte antigen susceptibility map for
SARS-CoV-2. Journal of Virology
DOI: 10.1128/JVI.00510-20
Nishiura H. et al. 2020. Closed environments facilitate secondary
transmission of coronavirus disease 2019 (COVID-19).
medRxiv preprint doi: https://doi.org/10.1101/2020.02.28.20029272.
Joung S. M. and Ryoo S. 2013. BCG vaccine in Korea. Clinical and
Experimental Vaccine Research 2: 83.
http://dx.doi.org/10.7774/cevr.2013.2.2.83
Omori M., 2009. Kekkaku kikansensya suu no suikei (結核既感染者数の推計).
The Research Institute of Tuverculosis, Japan. [in Japanese]
http://www.jata.or.jp/rit/ekigaku/
Sala G. and Miyakawa T. 2020. Association of BCG vaccination policy with prevalence
and mortality of COVID-19.
medRxiv preprint doi: https://doi.org/10.1101/2020.03.30.20048165.
Sekizuka T et al. 2020. Haplotype networks of SARS-CoV-2 infections in the Diamond
Princess cruise ship outbreak. MedRxiv preprint
https://doi.org/10.1101/2020.03.23.20041970.
Shet A. et al. 2020. Differential COVID-19-attributable mortality and BCG vaccine use
in countries.
medRxiv preprint doi: https://doi.org/10.1101/2020.04.01.20049478
Shulla A. et al. 2011. A transmembrane serine protease is linked to the severe acute
respiratory syndrome coronavirus receptor and activates virus entry. Journal of
Virology 85: 873-882.
TBS News. 2002. Passengers 98% reduction in Shibuya station on the first Saturday
after the announcement of state emergency. (「緊急事態宣言」後の土曜日、JR渋谷駅
利用客98%減) [in Japanese]
https://news.tbs.co.jp/newseye/tbs_newseye3955068.html
To et al. 2020. Temporal profiles of viral load in posterior oropharyngeal saliva samples
and serum antibody responses during infection by SARS-CoV-2: an observational
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
cohort study. The Lancet Infectious Diseases (published online)
https://doi.org/10.1016/S1473-3099(20)30196-1
Wourld Tourism Organization. 2019. International Tourism Highlights.
https://www.e-unwto.org/doi/book/10.18111/9789284421152
Wu Z. et al. 2016. Deciphering the bat virome catalog to better understand the
ecological diversity of bat viruses and the bat origin of emerging infectious diseases.
The ISME Journal 10: 609-620.
de Vrieze J. 2020. Can a century-old TB vaccine steel the immune system against the
new coronavirus? Science (online) doi:10.1126/science.abb8297
Zhang W. et al. 2013. Genome sequencing and analysis of BCG vaccine strains. PLoS
ONE 8: e71243.
doi:10.1371/journal.pone.0071243
Zhang J. et al. 2020. Changes in contact patterns shape the dynamics of the COVID-19
outbreak in China. Science. DOI: 10.1126/science.abb8001
Zhou P. et al. 2020. A pneumonia outbreak associated with a new coronavirus of
probable bat origin. Nature 579: 270-273.
https://doi.org/10.1038/s41586-020-2012-7
Zwerling A. et al. 2011. The BCG World atlas: a database of global BCG vaccination
policies and practices. Plos Medicine 8: e1001012.
doi:10.1371/journal.pmed.1001012
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
Table and Figure Legends
Table 1. Case fatality ratio between the nationality of different BCG vaccination policies
in the Diamond Princess (April 27, 2020).
Both past and current implementing of vaccination are categorized to +
Current implementing of vaccination is categorized to +
*There are still 6 people in the hospital whose nationality is undisclosed.
Table 2. Infection and mortality ratio between citizens of Japan and other countries in
the Diamond Princess (April 27, 2020)
Table 3. The result of general linear model testing for the effect of BCG vaccination and
other factors using all available nations around the world (Fig. 1 and 2)
Table 4. The result of general linear model testing for the effect of BCG vaccination and
other factors using European nations (Fig. 4)
Fig. 1. Bivariate plots of the Date of fifth Death (as an index of the timing of pandemic
entry into the nation) vs. Log (Total deaths up to April 6). Most nations showed
exponential increases in the number. However, some nations plotted lower than other
nations. It should be noted that BCG – countries are mainly located in Europe, there is
an epicenter of the pandemic.
Fig. 2. Bivariate plots of Days from fifth Death to April 6 or 30 (as an index of the time
from the entry of the pandemic to the nation/the time until the number of deaths has
flattened) vs. Log (Total deaths or sum of deaths at 30 days after the fifth death, i.e. the
number of deaths during the time X axis). Most nations showed exponential growth in
the number. However, some nations plotted lower than the other nations.
Fig. 3. Bivariate plots of international tourist arrivals in 2017 vs. Date fifth Death.
Pandemic propagation may be affected by tourist arrivals.
Fig. 4. Bivariate plots using only European nations. The plots clearly show that high
number of deaths are affected by when the pandemic enter the nation. The pandemic
explosion occurred in Western Europe where BCG – nations are located, and then
spread to the peripheral of Europe where BCG + nations are located.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
Fig. 5. Bivariate plots of International tourist arrivals in 2017 vs. Total deaths up to
April 6 using non-European nations with total deaths less than 200. Most nations are
BCG +, but some BCG - nations seem to succeed in suppressing COVID-19. Most
non-European nations are BCG + and are located where COVID-19 has not yet entered
at full scale. This fact would flaw the results of previous studies.
Fig. 6. Box plots of maximum increase rate per day (two week average). Nations where
deaths are more than 25 were illustrated. Boxes indicate quartiles, central-lateral bars
indicate medians, vertical bars indicate the range of the specimens, and asterisks
indicate outliers. The increase rate seems to higher in BCG – nations. However, only
one pair was weakly statistically significant, but most are not (ANOVA/ Tukey’s test).
Fig. 7. Cumulative deaths of nations in which total deaths over 25. Although the
increase rate in Iraq, Korea, and Japan seem to low, at the beginning of the increase,
the rate is not particularly low. Malaysia which uses Japanese strain shows a similar
pattern to the other nations.
Fig. 8. Schematic illustration of how cross-national statistical analysis (“ecological
study”) can be influenced. Many factors influence the results. The author’s idea of the
remaining possibility of BCG vaccination is also illustrated. The map is modified based
on the map provided by the Geospatial Information Authority of Japan.
https://maps.gsi.go.jp/development/ichiran.html
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
Both past and current implementing of vaccination are categorized to +BCG Cases Deaths Mortaliry rate (%)
+ 493 13 2.64- 141 1 0.71
No significant differencep=0.325 (Fisher's exact test)
Current implementing of vaccination is categorized to +BCG Cases Deaths Mortaliry rate (%)
+ 422 11 2.61- 212 3 1.42
No significant differencep=0.405 (Fisher's exact test)
Analysis on people over 60 years old (age-indeterminate deaths included)Both past and current implementing of vaccination are categorized to +
BCG Cases Deaths Mortaliry rate (%)+ 350 13 3.71- 126 1 0.79
No significant differencep=0.127 (Fisher's exact test)
Current implementing of vaccination is categorized to +BCG Cases Deaths Mortaliry rate (%)
+ 301 11 3.65- 175 3 1.71
No significant differencep=0.273 (Fisher's exact test)
Analysis on people over 70 years old (age-indeterminate deaths included)Both past and current implementing of vaccination are categorized to +
BCG Cases Deaths Mortaliry rate (%)+ 227 13 5.73- 71 1 1.41
No significant differencep=0.200 (Fisher's exact test)
Current implementing of vaccination is categorized to +BCG Cases Deaths Mortaliry rate (%)
+ 198 11 5.56- 100 3 3.00
No significant differencep=0.398 (Fisher's exact test)
*There are still 8 people in hospital whose nationality is undisclosed
Table 1. Case fatality ratio between nationality of different BCG vaccinationpolicies in the Diamond Princess (April 27, 2020)
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
Infection rate in the Diamond Princess (*2)Nationality Number on board Cases Infection rate (%)
Japan (BCG +) 1341 270 20.13USA (BCG -) 428 107 25.00
No significant differencep=0.1017 (Fisher's exact test)
Mortality rate in the Diamond Princess (*2)Nationality Number on board Deaths Mortality rate (%)
Japan (BCG +) 1341 9 0.67USA (BCG -) 428 0 0.00
No significant differencep=0.1246 (Fisher's exact test)
Mortality rate in the Diamond Princess passengers (*3)Nationality Number on board Deaths Mortality rate (%)
Japan (BCG +) 1281 9 0.70Hong Kong (BCG +) 260 2 0.77
USA (BCG -) 416 0 0.00Canada (BCG -) 251 1 0.40
Australia (BCG - / past+) 223 1 0.45UK (BCG - / past+) 57 1 1.75
No significant differenceJapan + Hong Kong vs Others; Others vs USA + Canada: p > 0.05 (Fisher's exact test)
*1 There are still 8 people in hospital whose nationality is undisclosed*2 US data was obrained from Moriarty et al. (2020)*3 Passengers data was obrained from Moriarty et al. (2020)
Table 2. Infection and mortality ratio between citizens of Japan and other countries in theDiamond Princess (April 27, 2020)
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
Table 3. The result of general linear model testing for the effect of BCG vaccination and other factors using all available nations aroud the world (Fig. 1 and 2)
Analysis A Analysis B
Adj SS Adj MS F P Adj SS Adj MS F PBCG policy 0.4857 0.2429 1.730 0.184 BCG policy (Japanese strain separated) 0.4895 0.1632 1.150 0.336Date 5th death 29.1332 29.1332 207.920 0.000 Date 5th death 29.006 29.006 204.210 0.000Region 8.1498 1.0187 7.270 0.000 Region 7.0906 0.8863 6.240 0.000
R^2=0.8367 R^2=0.8091Analysis C Analysis D
Adj SS Adj MS F P Adj SS Adj MS F PBCG policy 0.0821 0.0411 0.280 0.760 BCG policy (Japanese strain separated) 1.5887 0.5296 4.110 0.010Days from 5th death 25.005 25.005 168.160 0.000 Days from 5th death 26.353 26.353 204.720 0.000Region 3.6858 0.4607 3.100 0.005 Region 2.4454 0.3057 2.370 0.026
R^2=0.8080 R^2=0.8362*Period Death: Total deaths or sum of deaths 30th days from 5th death
Log Total Death = BCG polocy + Date 5th death + Region Log Total Death = BCG polocy (J strain separated) + Date 5th death + Region
Log Period Death = BCG polocy + Days from 5th death + Region Log Period Death = BCG polocy (J strain separated) + Days from 5th death + Region
. C
C-B
Y-N
C-N
D 4.0 International license
It is made available under a
is the author/funder, who has granted m
edRxiv a license to display the preprint in perpetuity.
(wh
ich w
as no
t certified b
y peer review
)T
he copyright holder for this preprint this version posted M
ay 22, 2020. ;
https://doi.org/10.1101/2020.04.17.20068601doi:
medR
xiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
Analysis E
Adj SS Adj MS F PBCG policy 0.7688 0.3844 2.850 0.072Date 5th death 17.3209 17.3209 128.450 0.000
R^2=0.8582Analysis F
Adj SS Adj MS F PBCG policy 0.36 0.18 1.270 0.295Days from 5th death 16.118 16.118 113.400 0.000
R^2=0.8427*Period Death: Total deaths or sum of deaths 30th days from 5th death
Log Period Death = BCG polocy + Days from 5th death
Table 4. The result of general linear model testing for the effect of BCGvaccination and other factors using European nations (Fig. 4)
Log Total Death = BCG polocy + Date 5th death
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
2020/04/012020/03/012020/02/01
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0 Venezuela
US
Uruguay
United Kingdom
Channel Islands
United Arab Emirates
Ukraine
Turkey
Tunisia
Trinidad and Tobago
Thailand
Switzerland
Sweden
Spain
South Africa
Slovenia
Singapore
Serbia
Saudi ArabiaSan Marino
Russia
Romania
Portugal
Poland
Philippines
Peru
PanamaPakistan
Norway
North Macedonia
Niger
Netherlands
Sint Maarten
Morocco
Moldova
Mexico
Mauritius
Malaysia
Luxembourg
LithuaniaLebanon
Korea, South
KenyaKazakhstanJordan
Japan
Italy
Israel
Ireland
Iraq
Iran
Indonesia
India
Iceland
Hungary
Honduras
Greece
Germany
France
Guadeloupe
Finland
Estonia
Egypt
Ecuador
Dominican Republic
Denmark
Czechia
CyprusCuba
CroatiaCongo (Kinshasa)
Colombia
China
Chile
Canada
Cameroon
Burkina FasoBulgaria
Brazil
Bosnia and Herzegovina
Bolivia
Belgium
BelarusBangladesh
Azerbaijan
Austria
Australia
Armenia
Argentina
Andorra
Algeria
Albania
Afghanistan
BCG +
BCG - (past BCG +)BCG -
BCG + (using Japanese strain)
Date 5th Death
Date 5th Death
Log (Total deaths to April 6)
Log (Total deaths to April 6)
A
B
Fig. 1. Bivariate plots of the Date of fifth Death (as an index of the timing of pandemic entry into the nation) vs. Log (Total deaths up
to April 6). Most nations showed exponential increases in the number. However, some nations plotted lower than other nations. It
should be noted that BCG – countries are mainly located in Europe, there is an epicenter of the pandemic.
2020/04/012020/03/012020/02/01
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0 South Asia
Africa
America
Central AsiaEast Asia
Europe
Middle EastPacific
SE AsiaVenezuela
US
Uruguay
United Kingdom
Channel Islands
United Arab Emirates
Ukraine
Turkey
Tunisia
Trinidad and Tobago
Thailand
Switzerland
Sweden
Spain
South Africa
Slovenia
Singapore
Serbia
Saudi ArabiaSan Marino
Russia
Romania
Portugal
Poland
Philippines
Peru
PanamaPakistan
Norway
North Macedonia
Niger
Netherlands
Sint Maarten
Morocco
Moldova
Mexico
Mauritius
Malaysia
Luxembourg
LithuaniaLebanon
South Korea
KenyaKazakhstanJordan
Japan
Italy
Israel
Ireland
Iraq
Iran
Indonesia
India
Iceland
Hungary
Honduras
Greece
Germany
France
Guadeloupe
Finland
Estonia
Egypt
Ecuador
Dominican Republic
Denmark
Czech
CyprusCuba
CroatiaCongo (Kinshasa)
Colombia
China
Chile
Canada
Cameroon
Burkina FasoBulgaria
Brazil
Bosnia and Herzegovina
Bolivia
Belgium
BelarusBangladesh
Azerbaijan
Austria
Australia
Armenia
Argentina
Andorra
Algeria
Albania
Afghanistan
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
302520151050
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0 Venezuela
USUnited Kingdom
Channel Islands
United Arab Emirates
Ukraine
Turkey
Tunisia
Trinidad and Tobago
Thailand
Switzerland
Sweden
Spain
South Africa
Slovenia
Singapore
Serbia
Saudi ArabiaSan Marino
Russia
Romania
Portugal
Poland
Philippines
Peru
PanamaPakistan
Norway
North Macedonia
Niger
Netherlands
Morocco
Moldova
Mexico
Mauritius
Malaysia
Luxembourg
LithuaniaLebanon
Korea, South
Kazakhstan
Japan
Italy
Israel
Ireland
Iraq
Iran
Indonesia
India
Hungary
Honduras
Greece
Germany
France
Guadeloupe
Finland
Estonia
Egypt
Ecuador
Dominican Republic
Denmark
Czechia
CyprusCuba
CroatiaCongo (Kinshasa)
Colombia
China
Chile
Canada
Cameroon
Burkina FasoBulgaria
Brazil
Bosnia and Herzegovina
Bolivia
Belgium
Belarus Bangladesh
Azerbaijan
Austria
Australia
Armenia
Argentina
Andorra
Algeria
Albania
Afghanistan
BCG +
BCG - (past BCG +)BCG -
BCG + (using Japanese strain)
Days from 5th death to April 6 (-30)
Days from 5th death to April 6 (-30)Log (Total deaths or sum of deaths 30th days from 5th death)
Log (Total deaths or sum of deaths 30th days from 5th death)
A
B
Fig. 2. Bivariate plots of Days from fifth Death to April 6 or 30 (as an index of the time from the entry of the pandemic to the
nation/the time until the number of deaths has flattened) vs. Log (Total deaths or sum of deaths at 30 days after the fifth death, i.e.
the number of deaths during the time X axis). Most nations showed exponential growth in the number. However, some nations
plotted lower than the other nations.
302520151050
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
South Asia
Africa
America
Central AsiaEast Asia
Europe
Middle EastPacific
SE Asia
Venezuela
USUnited Kingdom
Channel Islands
United Arab Emirates
Ukraine
Turkey
Tunisia
Trinidad and Tobago
Thailand
Switzerland
Sweden
Spain
South Africa
Slovenia
Singapore
Serbia
Saudi ArabiaSan Marino
Russia
Romania
Portugal
Poland
Philippines
Peru
PanamaPakistan
Norway
North Macedonia
Niger
Netherlands
Morocco
Moldova
Mexico
Mauritius
Malaysia
Luxembourg
LithuaniaLebanon
South Korea
Kazakhstan
Japan
Italy
Israel
Ireland
Iraq
Iran
Indonesia
India
Hungary
Honduras
Greece
Germany
France
Guadeloupe
Finland
Estonia
Egypt
Ecuador
Dominican Republic
Denmark
Czech
CyprusCuba
CroatiaCongo (Kinshasa)
Colombia
China
Chile
Canada
Cameroon
Burkina FasoBulgaria
Brazil
Bosnia and Herzegovina
Bolivia
Belgium
Belarus Bangladesh
Azerbaijan
Austria
Australia
Armenia
Argentina
Andorra
Algeria
Albania
Afghanistan
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
100000100001000100
2020/04/01
2020/03/01
2020/02/01
Venezuela
US
Uruguay
United Kingdom
United Arab Emirates
Ukraine
Turkey
TunisiaThailand
Switzerland
Sweden
Spain
Slovenia
Singapore
Saudi Arabia
San Marino
Russia
Romania
PortugalPoland
Philippines
PeruPanama
Norway
North Macedonia
Niger
Netherlands
Morocco
Moldova
Mexico
Mauritius
MalaysiaLuxembourg
Lithuania
Lebanon
Korea, South
Kenya Jordan
Japan
Italy
Israel
Ireland
Iran
Indonesia
IndiaHungary
Greece
Germany
France
Finland
Estonia
Egypt
Ecuador
Dominican Republic
Denmark
Czechia
CubaCroatia
Colombia
China
Chile
Canada
Burkina FasoBulgaria
Brazil
Bosnia and HerzegovinaBolivia
Belgium
Belarus
Bangladesh
Azerbaijan
AustriaAustralia
Armenia
Argentina
Andorra
Algeria
Albania
BCG +
BCG - (past BCG +)BCG -
BCG + (using Japanese strain)
Date 5th Death
International tourist arrivals 2017
Date 5th Death
International tourist arrivals 2017
A
B
Fig. 3. Bivariate plots of international tourist arrivals in 2017 vs. Date fifth Death. Pandemic propagation may be affected by tourist
arrivals.
100000100001000100
2020/04/01
2020/03/01
2020/02/01
South Asia
Africa
America
Central AsiaEast Asia
Europe
Middle EastPacific
SE Asia
Venezuela
US
Uruguay
United Kingdom
United Arab Emirates
Ukraine
Turkey
TunisiaThailand
Switzerland
Sweden
Spain
Slovenia
Singapore
Saudi Arabia
San Marino
Russia
Romania
PortugalPoland
Philippines
PeruPanama
Norway
North Macedonia
Niger
Netherlands
Morocco
Moldova
Mexico
Mauritius
MalaysiaLuxembourg
Lithuania
Lebanon
South Korea
Kenya Jordan
Japan
Italy
Israel
Ireland
Iran
Indonesia
IndiaHungary
Greece
Germany
France
Finland
Estonia
Egypt
Ecuador
Dominican Republic
Denmark
Czech
CubaCroatia
Colombia
China
Chile
Canada
Burkina FasoBulgaria
Brazil
Bosnia and HerzegovinaBolivia
Belgium
Belarus
Bangladesh
Azerbaijan
AustriaAustralia
Armenia
Argentina
Andorra
Algeria
Albania
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
302520151050
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
United Kingdom
Ukraine
TurkeySwitzerland
Sweden
Spain
Slovenia San Marino
Russia
Romania
Portugal
Poland
Norway
North Macedonia
Netherlands
Moldova
Luxembourg
Lithuania
Italy
Ireland
Hungary
Greece
Germany
France
Finland
Estonia
Denmark
Czechia
Croatia
BulgariaBosnia and Herzegovina
Belgium
Belarus
Azerbaijan
Austria
Armenia
Andorra Albania
2020
/04/
07
2020
/04/
01
2020
/03/
25
2020
/03/
19
2020
/03/
13
2020
/03/
07
2020
/03/
01
2020
/02/
25
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
United Kingdom
Ukraine
TurkeySwitzerland
Sweden
Spain
SloveniaSan Marino
Russia
Romania
Portugal
Poland
Norway
North Macedonia
Netherlands
Moldova
Luxembourg
Lithuania
Italy
Ireland
Hungary
Greece
Germany
France
Finland
Estonia
Denmark
Czechia
Croatia
BulgariaBosnia and Herzegovina
Belgium
Belarus
Azerbaijan
Austria
Armenia
AndorraAlbania
BCG +
BCG - (past BCG +)BCG -
BCG + (using Japanese strain)
Log (Total deaths to April 6)
Days from 5th death to April 6 (-30)
Date 5th Death
Log (Total deaths or sum of deaths 30th days from 5th death)
A
B
Number of total deathsmay reflect the route of propagation
Fig. 4. Bivariate plots using only European nations. The plots clearly show that high number of deaths are affected by when the
pandemic enter the nation. The pandemic explosion occurred in Western Europe where BCG – nations are located, and then spread
to the peripheral of Europe where BCG + nations are located.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
400003000020000100000
10
8
6
4
2
0
BCG +
BCG - (past BCG +)BCG -
BCG + (using Japanese strain)
VietnamFijiChadBhutan
Benin
Costa RicaBarbados
Bahrain
UgandaSeychellesNamibiaMadagascar
ZimbabweZambiaTanzaniaGambia
SudanSenegalEthiopiaAngola
Togo
GhanaCongo (Brazzaville)
Kenya
Mauritius
Niger
Saint Vincent and the Grenadines
NicaraguaHaiti
JamaicaGuatemala
GuyanaEl Salvador
Paraguay
Uruguay
Venezuela
Cuba
Uzbekistan
Mongolia
Hong Kong
Taiwan
Oman
Qatar
Jordan
Papua New Guinea
New Zealand
VietnamCambodia
Singapore
NepalMaldives
Sri Lanka
400003000020000100000
200
150
100
50
0VietnamFijiChadBhutanBenin
Costa R icaBarbadosBahrain
UgandaSeychellesNamibiaMadagascar ZimbabweZambiaTanzaniaGambiaSudanSenegalEthiopiaAngola
TogoGhanaCongo (Brazzaville)
KenyaMauritiusNiger
Burkina Faso
Tunisia
Morocco
Egypt
A lgeria
Saint V incent and the GrenadinesNicaraguaHaitiJamaicaGuatemalaGuyanaEl Salvador
ParaguayUruguayVenezuela
CubaBolivia
Chile
ColombiaArgentina
Panama
Dominican Republic
PeruMexico
Ecuador
UzbekistanMongolia
Hong KongTaiwan
Japan
Korea, South
OmanQatar
Jordan
United Arab Emirates
Lebanon
Saudi Arabia
Israel
Papua New Guinea New Zealand
Australia
V ietnamCambodia
Singapore
Thailand
Malaysia
Philippines
NepalMaldives
Sri Lanka
Bangladesh
India
Total death to April 6
Total death to April 6
International tourist arrivals 2017 (1000)
Fig. 5. Bivariate plots of International tourist arrivals in 2017 vs. Total deaths up to April 6 using non-European nations with total
deaths less than 200. Most nations are BCG +, but some BCG - nations seem to succeed in suppressing COVID-19. Most
non-European nations are BCG + and are located where COVID-19 has not yet entered at full scale. This fact would flaw the results
of previous studies.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
1.6
1.5
1.4
1.3
1.2
1.1
Region
1.6
1.5
1.4
1.3
1.2
1.1
+ (Japanese strain)
-- (past +)
+
1.6
1.5
1.4
1.3
1.2
1.1
-- (past +)
+
Nations where deaths are more than 25
Maximum increase rate per day (two week average)
BCG policy
BCG policy (Japanese strain separated)
No significant difference (P = 0.128)
No significant difference (P = 0.055)
Weakly significant difference (P = 0.037)
South AsiaSE AsiaPacificMiddle EastEuropeEast AsiaAmericaAfrica
(P = 0.043)
Fig. 6. Box plots of maximum increase rate per day (two week average). Nations where deaths are more than 100 and 25 were
separately illustrated. Boxes indicate quartiles, central-lateral bars indicate medians, vertical bars indicate the range of the
specimens, and asterisks indicate outliers. The increase rate seems to higher in BCG – nations. However, only one pair was weakly
statistically significant, but most are not (ANOVA/ Tukey’ s test).
N = 33
N = 18 N = 4
N = 27N = 18 N = 4
N = 6
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
2020
/1/2
220
20/1
/23
2020
/1/2
420
20/1
/25
2020
/1/2
620
20/1
/27
2020
/1/2
820
20/1
/29
2020
/1/3
020
20/1
/31
2020
/2/1
2020
/2/2
2020
/2/3
2020
/2/4
2020
/2/5
2020
/2/6
2020
/2/7
2020
/2/8
2020
/2/9
2020
/2/1
020
20/2
/11
2020
/2/1
220
20/2
/13
2020
/2/1
420
20/2
/15
2020
/2/1
620
20/2
/17
2020
/2/1
820
20/2
/19
2020
/2/2
020
20/2
/21
2020
/2/2
220
20/2
/23
2020
/2/2
420
20/2
/25
2020
/2/2
620
20/2
/27
2020
/2/2
820
20/2
/29
2020
/3/1
2020
/3/2
2020
/3/3
2020
/3/4
2020
/3/5
2 020
/3/6
2020
/3/7
2020
/3/8
2020
/3/9
2020
/3/1
020
20/3
/11
2020
/3/1
220
20/3
/13
2020
/3/1
420
20/3
/15
2020
/3/1
620
20/3
/17
2020
/3/1
820
20/3
/19
2020
/3/2
020
20/3
/21
2020
/3/2
220
20/3
/23
2020
/3/2
420
20/3
/25
2020
/3/2
620
20/3
/27
2020
/3/2
820
20/3
/29
2020
/3/3
020
20/3
/31
2020
/4/1
2020
/4/2
2020
/4/3
2020
/4/4
2020
/4/5
Italy
Spain
US
France
United Kingdom
Iran
China
Netherlands
Germany
Belgium
Switzerland
Turkey
Brazil
Sweden
Portugal
Canada
Austria
Indonesia
Korea, South
Ecuador
Denmark
Ireland
Algeria
Philippines
Romania
India
Poland
Peru
Dominican Republic
Mexico
Egypt
Japan
Greece
Norway
Morocco
Czechia
Iraq
Malaysia
Serbia
Log (Cumulative Deaths)
Date
BCG -
BCG - (previously +)
BCG +
BCG + (Japanese strain)
Fig. 7. Cumulative deaths of nations in which total deaths over 25. Although the increase rate in Iraq, Korea, and Japan seem to low, at the beginning of the increase, the rate is not particularly
low. Malaysia which uses Japanese strain shows a similar pattern to the other nations.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/
-
Toll-likereceptor
Epitopes
Specific domain of the Japanese strain
No homologous amino acid sequences
High affinity?
SARS-CoV-2
A protein of BCG
ACE2
Polymorphism?
Difference of ACE2 expression in salivary gland?
Polymorphism of salivaor oral morphology?
Difference in susceptibility?
Culture, national politicy etc.
Rhinolophus bat distribution
Paguma larveta distribution
Geographic variation inratio of the variant?
Epidemic?
Cross-national difference in morbidity and mortality of COVID-19
BCG vaccination?
Malaria mosquito distribution
Fig. 8. Schematic illustration of how cross-national statistical analysis ( “ecological study” ) can be influenced. Many factors
influence the results. The author’ s idea of the remaining possibility of BCG vaccination is also illustrated. The map is modified
based on the map provided by the Geospatial Information Authority of Japan. https://maps.gsi.go.jp/development/ichiran.html
. CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 22, 2020. ; https://doi.org/10.1101/2020.04.17.20068601doi: medRxiv preprint
https://doi.org/10.1101/2020.04.17.20068601http://creativecommons.org/licenses/by-nc-nd/4.0/