Keywords: BCG; COVID-19; Flu; Tuberculosis; Vaccinations

Introduction

COVID-19 is a disease that is caused by SARS-COV2, a single stranded positive RNA virus. It is responsible for the recent pandemic leading to widespread deaths and disruption in lives and economies of nations. SARS-CoV2 infects the respiratory tract causing mild symptoms or, in severe cases, causes damage to the lungs [1]. The SARS family of viruses have a similar mechanism of entry in host cells using multiple mechanisms. The spike glycoprotein present in the virus plays important roles and interacts with the receptor angiotensin converting enzyme (ACE)2 on host cells. In addition, several other viral as well as host proteins such as cellular transmembrane proteases (e.g TMPRSS2), furin proteases etc are involved in aiding the replication of SARS-CoV2 viral particles [2]. A better understating of the viral replication cycle will be the key to the development of inhibitors some of which are likely to act as drugs.

The economic damage and loss of lives due to COVID-19 has been immense with official world- wide deaths in the range of 1.88 million in 2020, 3.58 million in 2021 and 1.22 million in 2022.  Mortality due to COVID-19 is variable across different countries and there may be several reasons for this heterogeneity: population age, sex, temperature, cultural norms such as the use of masks, better health care infrastructure, comorbidities such as hypertension, obesity, etc [3-5]. Recent studies have demonstrated the critical roles of host derived autoimmune antibodies to Interferons, polymorphisms in Toll-like receptors in protection against SARS-CoV2 infections [6]. Additionally, prior exposure to infectious pathogens may boost cross-reactive immunity to pathogens. For example, H. pylori infections may lower tuberculosis incidences [7]. Also, pre-existing infections with HIV or tuberculosis can also lead to COVID-19 severity [8]. On the other hand, malaria incidences are known to lower COVID-19 [9]. It is well known that previous exposure to family members of coronaviruses that cause common cold leads to better T cell immunity against SARS-Cov2 [10]. Another important aspect is vaccinations, which are an established public safety shield as well as a highly cost-effective strategy. They are a key factor in enhancing immunity against a vast number of pathogens, especially in protecting children [11,12]. Interestingly, immunizations with oral polio vaccine, measles vaccine, hepatitis vaccine [13-15] are known to reduce COVID-19 cases and these are likely due to boost non-specific immunity, resulting in heterologous protection.

The vast differences in the number of COVID-19 deaths in various countries led us to investigate the correlation between COVID-19 incidences or deaths with the prevalence of various diseases using data available in the public domain. Previously, we had shown that countries with high amounts of deaths due to flu or tuberculosis display less COVID-19 deaths in 2020 [16]. Also, countries with high amounts of flu vaccination led to higher COVID-19 deaths. On the other hand, countries with high BCG vaccine coverage led to lower COVID-19 deaths [16]. An important question to address is whether this trend was observed as the pandemic spread in 2021 and 2022. This aspect is important as there was a rise in variants as well as an increase in vaccinations. We find that the trends observed initially in 2020 are sustained and even better now in 2021-2022 as COVID-19 incidences as well as deaths correlate well with time. The implications of our findings are discussed in detail in this study.

Methods

Study design

Countries with more than 10,000 COVID-19 deaths were selected at four time points and the data was analysed: 24^th May 2021, 31^st December 2021, 14^th April 2022 and 31^st December 2022.

Selection of Samples

A set of 62 countries were selected on the criteria of deaths that are more than 10,000 by 14^th April 2022 (Figure 1). These countries were arranged and tables were constructed where COVID-19 incidences/million and deaths/million were obtained from: https://www.worldometers.info/coronavirus/ and https:// ourworldindata.org/coronavirus#explore-the-global-situation. These factors were initially correlated with pathogen exposure. TB deaths/100,000 and Flu deaths/100,000 and this data was obtained from: https://www.worldlifeexpectancy.com/worldhealth-rankings. Next, correlation analysis was performed with different vaccination programs. BCG coverage for different countries were obtained from http://www.bcgatlas.org/. BCG Atlas uses three distinct methods of obtaining data: First, it uses a questionnaire method which is sent to at least two individuals in each country based on their expertise on TB research, TB control programs and public health/vaccination programs. It includes data regarding any changes that have occurred in the last 25 years. It also includes criteria such as tuberculin skin testing, effect of HIV and different vaccine strains playing a role in affecting extent of TB infection. Secondly, case reports, published papers and policy documents from the government are also included as a part of the overall database. Finally, it uses the immunization data from the World Health Organization Vaccine Preventable Diseases Monitoring System as mentioned in http://apps.who. int/immunization_monitoring/en/globalsummary/ScheduleSelect. cfm [17]. Flu vaccination coverage data was obtained from various sources: https://www.statista.com/chart/16575/global-fluimmunization-rates-vary/, https://www.oecd.org/health/graph-ofthe-month.htm and published manuscripts [18-20]. The COVID-19 vaccination coverage in each country was extracted from: https:// ourworldindata.org/covid-vaccinations. These factors were compared to COVID-19 deaths/million and incidences/million at the four time points.

COVID variant kinetics data was obtained from: https:// covariants.org/variants which helped to confirm the dominant COVID-19 strains responsible for infection and deaths in each country at mentioned time points. The early 2020 time points suggested that a dominant strain wasn’t identified and was therefore labelled as “others” (Table 1).

Table 1: Dominant COVID-19 variants in countries with high COVID-19 deaths from 17^th May 2020-31^st December 2022.

Statistical Analysis

Correlation analysis was performed by using Spearman’s correlation analysis and the Spearman’s correlation coefficient (R_s) and p values were calculated using a Spearman’s correlation calculator: https://www.socscistatistics.com/tests/spearman/default2.aspx. Correlation was considered significant if p<0.1. Bar graphs were constructed for countries grouped either by using extent of TB and Flu deaths or using BCG and Flu vaccination coverages. Statistical analysis was done using One way ANOVA to calculate the statistical significance. Statistical analysis was considered significant if p<0.1.

Results

To perform the epidemiological study, a subset of 62 countries were selected based on the criteria in which the total deaths due to COVID-19 were greater than 10000 by 14^th April 2022. To evaluate changes with time, the cluster of countries were arranged on the basis of four distinct time points (Table 2-5): 24^th May 2021, 31^st December 2021, 14^th April 2022 and 31^st December 2022 and compared to four conditions (Figure 1): Flu deaths/100,000, TB deaths/100,000, Flu vaccination coverage and BCG coverage.

Figure 1: Flow chart describing the methodology used to perform the mentioned epidemiological study.

Table 2: List of countries with COVID-19 deaths/million and COVID-19 incidences/million, BCG, Flu Vac coverage, TB and Flu deaths/100,000 on 24^th May 2021.

Table 3: List of countries with COVID-19 deaths/million and COVID-19 incidences/million, BCG, Flu Vac coverage, TB and Flu deaths/100,000 on 31^st December 2021.

Table 4: List of countries with COVID-19 deaths/million and COVID-19 incidences/million, BCG, Flu Vac coverage, TB deaths/100,000, Flu deaths/100,000 and COVID vaccination on 14^th April 2022.

Table 5: List of countries with COVID-19 deaths/million and COVID-19 incidences/million, BCG, Flu Vac coverage, TB deaths/100,000 and Flu deaths/100,000 on 31^st December 2022.

Initially, we wished to determine how COVID-19 variants and COVID-19 vaccination coverage could have affected COVID-19 infection induced deaths from the start of the pandemic until December 2022. As seen in Table 1, at the initial phase of the pandemic (early-late 2020) it was difficult to determine the dominant strain causing the global pandemic. However, by the end of 2021 to the entire 2022 the dominant strain determining COVID-19 infection and deaths was the Omicron strain (21L, 22E and 22F). Additionally, countries which received at least 2 doses of the COVID-19 vaccination by 14^th April 2022 displayed significantly reduced COVID-19 mortality compared to 24^th May 2021 when the vaccination coverage was not prevalent across the different countries. This confirms the protective effect of the vaccination with respect to COVID-19 deaths (Figure 2).

Figure 2: Covid vaccination causes a significant reduction in COVID-19 deaths across different countries: 62 countries were compared with respect to COVID-19 death:7-day average in pre-vaccination (24^th May 2021) and post-vaccination (14^th April 2022) conditions. A country is considered vaccinated against COVID-19 if the population has received at least two doses of the vaccine. Student t-test was performed to calculate statistical significance, p<0.1.

To understand the importance of pathogen exposure in providing cross protection against COVID-19 infection we started our analysis by calculating the extent of deaths due to Flu and TB and normalizing the values per 100,000 people. Subsequently, Spearman’s correlation analysis was performed with both COVID-19 incidences/million and deaths/million, in which both the number of COVID-19 infection and deaths were normalized to the population of the country. We calculated the nature and significance of each correlation by using the “R_s” (Spearman’s correlation coefficient) and p value respectively. TB deaths/100,000 showed a significant negative correlation with both COVID-19 deaths (Figure 3) and incidences (Figure 4) at all four time points suggesting the strong evidence of TB infection providing cross protection against COVID-19. Interestingly, TB deaths/100,000 showed a stronger negative correlation with COVID-19

incidences/million (R_s value range: -0.65 to -0.72 and p value<0.0001) than COVID-19 deaths/million (R_s value range: -0.27 to -0.44 and p value range: 0.03 to 0.0003).

Figure 3: TB deaths/100,000 shows a significant negative correlation with COVID-19 deaths/million: TB deaths/100,000 was correlated with respect to COVID-19 deaths/million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. Spearman’s correlation was performed and the correlation coefficient value (R_S) was calculated and the respective p values were also estimated for each graph. Correlation is considered statistically significant if p<0.1.

Figure 4: TB deaths/100,000 shows a significant negative correlation with COVID-19 incidences/million: TB deaths/100,000 was correlated with respect to COVID-19 incidences/million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. Spearman’s correlation was performed and the correlation coefficient value (R_S) was calculated and the respective p values were also estimated for each graph. Correlation is considered statistically significant if p<0.1.

Figure 5: Flu deaths/100,000 negatively correlates with COVID-19 deaths/million:Correlation was performed with Flu deaths/100,000 and COVID-19 deaths/million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. Spearman’s correlation was performed. Subsequently, the correlation coefficient value (R_S) was calculated and the respective p values were also estimated for each graph. Correlation is considered statistically significant if p<0.1.

Figure 6. Flu deaths/100,000 negatively correlates with COVID-19 incidences/million at all three timepoints: Correlation was performed with Flu deaths/100,000 and COVID-19 incidences/million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. Spearman’s correlation was performed. Subsequently, the correlation coefficient value (R_S) was calculated and the respective p values were also estimated for each graph. Correlation is considered statistically significant if p<0.1.

Also, correlation analysis with Flu deaths/100,000 showed a similar trend where there was a significant negative correlation both with COVID-19 deaths (Figure 5) and incidences (Figure 6) which strengthened the importance of cross protection by two types of respiratory pathogens in determining the variation of COVID-19 infection and deaths in different countries. As there was a significant negative correlation with TB and Flu infection induced deaths it was important to perform correlation studies with BCG and Flu vaccination. BCG vaccination showed a significant negative correlation with COVID-19 deaths/million at three time points except 31^st December 2021 (R_s = -0.20 and p = 0.11) (Figure 7). It is possible that this is due to variation in COVID-19 deaths across different countries over different timepoints. Considering there was still significant correlation in the remaining time points, the data suggested that BCG was still a major contributing factor in determining extent of COVID-19 infection in different countries. Additionally, there was significant negative correlation of BCG with COVID-19 incidences/million across all four time points further solidified the importance of BCG coverage playing a definitive role (Figure 9).Subsequently, Flu vaccination coverage positively correlated with both COVID-19 deaths (Figure 8)and incidences/million (Figure 10) across all the four timepoints. Therefore, countries with higher Flu vaccination coverage would naturally have less exposure to the infection thus making the population sensitive towards COVID-19 infection.

Figure 7: BCG coverage shows a negative correlation with COVID-19 deaths/million: Correlation graphs were plotted with respect to BCG coverage and COVID-19 deaths/million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. Spearman’s correlation was performed and the correlation coefficient value (R_S) was calculated along with the respective p values. Correlation is considered statistically significant if p<0.1.

Figure 8: Flu vaccination coverage shows a positive correlation with COVID-19 deaths/million: Correlation graphs were plotted with respect to Flu vac coverage and COVID-19 deaths/million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. Spearman’s correlation was performed to calculate the correlation coefficient value (R_S) along with the respective p values. Correlation is considered statistically significant if p<0.1. (Note Fig 8 has been replotted and changed, 31 Dec 2022 data - no change in interpretation of data)

Figure 9: BCG coverage shows a significant negative correlation with COVID-19 incidences/million: Correlation graphs were plotted with respect to BCG coverage and COVID-19 incidences/million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and 31^st December 2022. Spearman’s correlation was performed and the correlation coefficient value (R_S) was calculated along with the respective p values. Correlation is considered statistically significant if p<0.1.

Figure 10: Flu vaccination coverage shows a strong positive correlation with COVID-19 incidences/million: Correlation graphs were plotted with respect to Flu vac coverage and COVID-19 incidences/million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. Spearman’s correlation was performed to calculate the correlation coefficient value (R_S) along with the respective p values. Correlation is considered statistically significant if p<0.1.

As pathogen exposure was effectively correlating with COVID-19 infections and deaths, we wished to understand whether the heterogeneity in countries with TB and Flu infections affected COVID-19 infections and deaths. To test this, we segregated the 62 countries based on extent of deaths due to TB and Flu. Deaths were considered high (“hi”) if deaths were >= 25 deaths/100,000 and low (“lo”) when deaths were <=25 deaths/100,000. Using these demarcations, we were able to segregate the countries into three basic groups: TB^hiFlu^hi, TB^loFlu^hi and TB^loFlu^{lo 16}. It was observed that countries with both high TB and Flu deaths were the most protected against COVID-19 incidences (Figure 12)and deaths/million (Figure 13). On the contrary, countries with both low TB and Flu deaths were the most sensitive against COVID-19 infection. Interestingly, countries with low TB deaths but high Flu deaths displayed less COVID-19 incidences compared to countries with low TB and Flu deaths (Figure 12).

Figure 11: Countries with high BCG coverage and TB deaths have better protection from COVID-19 deaths: Countries with high BCG coverage were divided using TB deaths as the differentiating criteria.  BCG^hiTB^hi and BCG^hiTB^lo groups were then compared for COVID-19 deaths/million for four mentioned timepoints: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. Student’s t-test was used to calculate statistical significance. Data is considered statistically significant if p<0.1.

Figure 12: Countries with high TB and Flu deaths have the lowest COVID-19 incidences/million: Countries were grouped based on deaths due to TB and flu infection. The countries were then compared with respect to COVID-19 incidences/million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. One-way ANOVA was performed to calculate statistical significance, p<0.1.

Figure 13: Countries with high TB and Flu deaths are most protected from deaths due to COVID-19: Countries were grouped based on deaths due to TB and flu infection. The countries were then compared with respect to COVID-19 deaths/million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. One-way ANOVA was performed to calculate statistical significance, p<0.1.

Considering both TB deaths and BCG coverage had a strong negative correlation with both COVID-19 incidences/million and deaths/million, we asked the subsequent question: whether the extent of COVID-19 infection was dependent on a preferred factor out of the two. We thus selected countries which had high BCG coverage and divided the countries based on high and low TB deaths. We thus had two groups: BCG^hiTB^hi and BCG^hiTB^lo. Figure 11 and Figure 14 shows that countries which had high BCG coverage and high TB deaths had significantly less COVID-19 cases and deaths than countries which only had high BCG coverage. Hence, high BCG coverage and TB exposure in the former have an additive effect in promoting the best protection against COVID-19 infection and deaths compared to latter which gets protection only due to high BCG coverage.

Figure 14: Countries with high BCG coverage and TB deaths have better COVID-19 protection: Countries with high BCG coverage were divided into two main groups using TB deaths as the differentiating criteria.  BCG^hiTB^hi and BCG^hiTB^lo groups were compared for COVID-19 incidences/million for four timepoints: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. Student’s t-test was used to calculate statistical significance. Data is considered statistically significant if p<0.1.

Finally, we addressed the roles of BCG and Flu vaccination as contrasting correlation patterns were observed. The segregation pattern was as follows: countries were segregated based on the presence or absence of a particular vaccination which was denoted as “+” or “-“ respectively. For countries with both vaccination programs, the segregation pattern was based on the extent of vaccination coverage. Countries with <50% coverage of a particular vaccination was considered as low coverage (“lo”). On the other hand, >50% coverage was considered high coverage (“hi”). Using these parameters countries were segregated in four distinct groups: BCG⁺FluVac^-, BCGFluVac⁺, BCG^hiFluVac^hi and BCG^hiFluVac^{lo 16}. These four groups were first compared using COVID-19 incidences/million where it was evident that countries with only BCG vaccination had the lowest COVID-19 incidences/million whereas countries with only Flu vaccination had the highest (Figure 15). However, this observation was not seen when analysing COVID-19 deaths/million (Figure 16). This could possibly be because countries belonging to the group BCG⁺FluVac^- had more variation in the extent of deaths at these later time points since countries such as India, South Africa etc. have now higher deaths due to COVID-19 over time which has caused the overall significant differences to disappear.

Figure 15: BCG⁺FluVac^- countries are the most protected from COVID-19 infection: Countries were segregated based on extent of BCG and Flu vaccination coverage and then compared with respect to COVID-19 incidences/million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. One-way ANOVA was performed to calculate statistical significance, p<0.1.

Figure 16: BCG⁺FluVac^- countries are more protected with respect to deaths due to COVID -19 than countries which are BCG^hiFluVac^lo: Countries were segregated based on extent of BCG and Flu vaccination coverage and then compared with respect to COVID-19 deaths/ million at four different time points: A. 24^th May 2021, B. 31^st December 2021, C. 14^th April 2022 and D. 31^st December 2022. One-way ANOVA was performed to calculate statistical significance, p<0.1.

Discussion

Pre-existing infections may modulate responses to newer infections by pathogens [7,9,10]. In this study, we followed the COVID-19 pandemic and showed that countries with high flu and tuberculosis deaths showed less COVID-19 incidences as well as deaths (Figures 1, 3-6). Also, there are opposing roles of vaccine. Countries with high flu vaccinations show greater COVID-19 incidences and deaths (Figures 8 and 10). On the other hand, countries with high BCG vaccination show lower COVID-19 incidences and deaths (Figures 7 and 9). In our previous study based on 2020 data, correlations were seen only with respect to COVID- 19 deaths [16]. However, in the current study as the pandemic has progressed correlations are seen with both COVID-19 incidences as well as deaths.

There are several points that are notable in this study. First, this study involves a large number of countries and is global in nature. Second, the number of cases as well as deaths have been extracted from various official sites.  Third, fluctuations in cases and deaths may vary in different countries across various time points; however, this does not appear to affect the overall trends.

For example, the number of cases and deaths during the second wave in India in 2021 were very high. Despite this observation, India seems to have weathered the effects of the COVID-19 pandemic well. From the point of immune responses, it is possible that countries with high exposure of pathogens lead to higher nonspecific immunity. Indeed, flu vaccinations have been shown to protect against COVID-19 infections [21,22], perhaps due to an increase in non-specific immunity. However, our epidemiological study shows the opposite effect of flu infections, perhaps due to long-term effects. Similar to our observation, another study has shown a negative correlation between COVID-19 deaths/100,000 and flu-pneumonia deaths per year per 100,000 in various countries [23]. In fact, flu infections are known to increase non-specific immunity [24,25]. Also, infections by some viruses may interfere with the replication of other viruses, a phenomenon known as viral interference. Flu infections are known to affect the replication of other viruses [26,27]. In fact, coinfections studies have shown that flu virus infection lowers the replication of the SARS-Cov2 virus due to an early burst of Type I Interferon [28]. Interestingly, the chances of testing positive for SARS-Cov2 is lower in patients suffering from influenza although those infected with both viruses had a higher risk of mortality compared to those infected with either virus [29]. Clearly, these studies highlight the importance of virus-virus interactions that may modulate immune responses in the respiratory track.