key: cord-285306-leu2hygk
authors: Gallagher, J.; Watson, C.; Ledwidge, M.
title: Association of Bacille Calmette-Guerin (BCG), Adult Pneumococcal and Adult Seasonal Influenza Vaccines with Covid-19 Adjusted Mortality Rates in Level 4 European countries
date: 2020-06-05
journal: nan
DOI: 10.1101/2020.06.03.20121624
sha: 
doc_id: 285306
cord_uid: leu2hygk

Introduction Non-specific effects of vaccines have gained increasing interest during the Covid-19 pandemic. In particular, population use of BCG vaccine has been associated with improved outcomes. This study sought to determine the association of population use of BCG, adult pneumococcal and adult seasonal influenza vaccination with Covid-19 mortality when adjusted for a number of confounding variables. Methods: Using publicly available data, mortality adjusted for the timeframe of crisis, population size and population characteristics was calculated. The primary analysis was the relationship between each of the day 15 and day 30 standardised mortality rates and BCG, adult pneumococcal and influenza vaccination scores using unadjusted measures and with adjustment for population structure and case fatality rates. Secondary analyses were measures of case increases and mortality increases from day 15 to day 30 for each of the relative vaccination scores. Finally, we also analysed the peak Z score reflecting increases in total mortality from historical averages reported by EuroMOMO (Euromomo.eu), Results: Following adjustment for the effects of population size, median age, population density, the proportion of population living in an urban setting, life-expectancy, the elderly dependency ratio (or proportion over 65 years), net migration, days from day 1 to lockdown and case-fatality rate, only BCG vaccination score remained significantly associated with Covid-19 mortality at day 30. In the best fit model, BCG vaccination score was associated with a 64% reduction in log(10) mortality per 10 million population (OR 0.362 reduction [95% CI 0.188 to 0.698]), following adjustment for population size, median age, density, urbanization, elderly dependency ratio, days to lockdown, yearly migration and case fatality rate. Conclusion BCG vaccine was associated with reduced mortality rates in level 4 countries while adult pneumococcal and adult seasonal influenza vaccines were not when adjusted for a number of confounding variables. A number of trials are ongoing to determine if BCG is protective against severe Covid-19 infection.

Using publicly available mortality data from the John's Hopkins Coronavirus Research Centre and demographic information presented on worldometers.info, we constructed curves of cumulative mortality adjusted for timeframe of crisis and population size, as previously described in a rapid response to an editorial on SARS-CoV-2testing and comparative international mortality rates [11a] . In order to standardise the timeframe, we designated the baseline (day 1) as the day that mortality rates exceeded 1 per 5 million population. We calculated the day 15 and day 30 mortality rate per 10 million population for each of the following 25 level 4 European countries Austria, Belgium, Czechia, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Lithuania, Luxembourg, Netherlands, Norway, Poland, Portugal, Slovakia, Slovenia, Spain, Sweden, Switzerland, United Kingdom. We also obtained recent data on population structure including population size, the median age (years), population density (persons/Km2), the proportion of population living in an urban setting, life-expectancy (years), the elderly dependency ratio (proportion of people over 65 years divided by the proportion aged 15-65) and net migration (expressed per 10 million population). In multivariable modelling we also included the case fatality rate on day 21 to account for differences in testing rates, testing strategy and case definition as well as the relative number of days between baseline (day 1) and lockdown.

In order to standardise the estimation of the proportion of the population covered with BCG vaccination, we generated a BCG vaccination score by dividing the median life expectancy by the total number of years that BCG vaccination was available and recommended for all children. We adjusted this calculation by multiplying with a measure of vaccination programme efficacy, using the most recent vaccination coverage rate for either BCG or universal childhood vaccination rates as appropriate. In the case of two countries that do not routinely recommend BCG vaccination and where few details were available (Iceland, Luxembourg) we used referenced estimates of the total population covered by BCG vaccination. Similarly, for pneumococcal vaccination, we generated a pneumococcal vaccination score by estimating the proportion of older people covered by pneumococcal vaccination over the past decade using a variety of referenced sources in circumstances where uptake is considered relatively low in most European countries, despite recommendations for their use in over 65s and vulnerable populations. Finally, we created an influenza vaccination score as the proportion of older adults with seasonal influenza vaccination from most recent sources provided by the European Centre for Disease Control and other referenced reports.

The primary analysis of our report was the relationship between each of the day 15 and day 30 standardised mortality rates and BCG, pneumococcal and Influenza vaccination scores using unadjusted measures and with adjustment for population structure and case fatality rates.

Secondary analyses were measures of case increases and mortality increases from day 15 to day 30 for each of the relative vaccination scores. Finally, we also analysed the peak Z score reflecting increases in total mortality from historical averages reported by EuroMOMO (Euromomo.eu), a European mortality monitoring activity, aiming to detect and measure excess deaths related to seasonal influenza, pandemics and other public health threats, supported by the European Centre for Disease Prevention and Control (ECDC) and the World Health Organization (WHO).

Primary and secondary outcomes were analysed using generalized linear modeling with a Poisson outcome distribution for adjusted mortality rates. Analyses were performed both with and without adjustment for the effects of population size, the median age, population density, the proportion of population living in an urban setting, life-expectancy, the elderly dependency ratio (or proportion over 65 years), net migration, days from day 1 to lockdown and case-fatality rate. Variables such as day 15 and day 30 mortality rates, case fatality rates, population size, population density and life expectancy, were not normally distributed across the countries and were log-transformed before inclusion in the generalized linear model. Nominal statistical significance was set at P<0.05 with Bonferroni correction for the 3 pairwise correlations involved in the primary analyses (three vaccination scores linked to day 30 mortality rates), meaning a P value of 0.05/3=0.0167 was considered statistically significant. Descriptive data are presented as either mean ± SD or median (25th: 75th percentile) for normally and non-normally distributed continuous variables, respectively.

Frequencies and percentages (in parentheses) summarize categorical variables. Graphs were presented using GraphPad Prism v6 and statistical analysis was performed using the R statistical software package (Version 3.6.2, 2019-12-12, Copyright 2019 The R Foundation).

The population structure of the 25 countries are presented in Table 1 . The date of exceeding 1 SARS-CoV-2death per 5 million population, date of lockdown and cases, deaths and case fatality rates (CFR) on days 15 and 30 are presented in Table 2 .

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.03.20121624 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2020. .

The unadjusted linear regression between log of deaths per 10 million on day 30 versus BCG, pneumococcal and influenza vaccination scores are presented in Figure 2A -C. Following adjustment for the effects of population size, median age, population density, the proportion of population living in an urban setting, life-expectancy, the elderly dependency ratio (or proportion over 65 years), net migration, days from day 1 to lockdown and case-fatality rate, BCG vaccination score remained significantly associated with Covid 19 mortality at day 30. In the best fit model, BCG vaccination score was associated with a 64% reduction in log (10) Similarly, as shown in Figure 3A . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2020. . . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2020. Studies suggesting that countries providing universal BCG vaccine coverage have decreased mortality with COVID 19(6-9) have been the source of wide media interest. However, there have been significant concerns about confounding factors and the lack of randomised controlled data to determine if there is a true causation between BCG vaccination and reduced mortality with COVID 19 (10). While this ecological study may be the first to evaluate BCG vaccination in models with standardisation of mortality rates for timeframe of the outbreak as well as population size, as well as incorporating extensive covariate adjustment related to population structure, the major limitation is that it demonstrates correlation but cannot definitively assign causation. There may be many other confounding variables, which could account for the relationship between BCG vaccination and mortality. Furthermore, it is well accepted that differences in recording mortality between countries have been reported which may affect numbers. Finally, the BCG vaccination score is an estimate of the population coverage with BCG and may not accurately reflect the true population vaccination.

Conversely, the extensive covariate adjustment as well as differential relationship between BCG and influenza/pneumococcal vaccination scores, may support the hypothesis that BCG vaccination is associated with protection from the severest consequences of Covid 19. There is a persistent relationship between BCG vaccination and a variety of mortality and morbidity measures, including peak standardised Z-scores reported for the pandemic. The present study builds on a growing body of evidence that BCG vaccination may be associated with decreased illness severity unrelated to mycobacterium infection. Many have been epidemiological studies looking at overall mortality rates . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2020. . https://doi.org/10.1101/2020.06.03.20121624 doi: medRxiv preprint but others have looked at specific unrelated causes such as parasites (11) , fungi (12) and bacteria (13) in humans and animals. A recent systematic review concluded that there is evidence with low to moderate risk of bias that BCG vaccination prevents respiratory infections (pneumonia and influenza) in children and the elderly(14) A number of trials are ongoing at present to elucidate the role of BCG in preventing severe COVID 19 with three reported in clinicaltrials.gov (15) (16) (17) .

To date, the nature of non-specific effects with BCG has not been fully elucidated, yet there are a number of possible mechanisms of causation to explain the correlations described in our study.

Vaccines with non-specific effects would induce reprogramming of the innate immune responses, a mechanism that clearly differs from the adaptive immunity induced by the antigen-specific responses to the vaccine. Also, since BCG is generally administered to children, any putative innate immune cell "priming" following BCG would, necessarily, be retained many years following vaccination, suggesting an epigenetic link. Interestingly, BCG vaccination has been shown to cause epigenetic changes in innate immune cells via histone-3-lysine-4 (H3K4) methylation [13] , which can can augment human IFN and ISG responses to non-mycobacterium and even fungal infections. Therefore, the BCG "primed" innate immune system could counter the SARS-like coronavirus early IFN and ISG suppression, providing a more robust initial innate immune response to limit replication of SARS-CoV-2virus. [18] Interestingly The search for a vaccine to manage the SARS-CoV-2pandemic is a global healthcare priority because of high transmissibility and Covid19 mortality rates. However, following a small, non-human primate . CC-BY-NC-ND 4.0 International license It 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 June 5, 2020. . study of the Oxford RNA vaccine, the world's largest producer of vaccines is in negotiation to dedicate its production capacity to this developmental vaccine for SARS-CoV-2. [23,24] However, this could limit normal childhood vaccination programmes despite risks of using novel technology platforms to treat the novel SARS-CoV-2 virus. While we await the results of prospective, randomised, controlled clinical studies on novel vaccines specific for SARS-CoV-2, recent clinical progress with known drugs such as remdesivir highlight the potential of repurposing known therapies to combat Covid-19 (18) . The present study suggests that ongoing, prospective clinical studies to determine if BCG vaccination reduces the severity of Covid-19 in vulnerable patients could provide important information on the potential for innate immune priming with BCG as an adjunct to other clinical strategies for protection against the worst consequences of SARS-Cov2.

The association of BCG and reduced mortality from Covid-19 was seen even after standardising mortality between level 4 European countries and adjusting for confounding variables relating to population structure. A similar association was not seen with adult influenza and pneumococcal vaccination. The results of prospective clinical studies on BCG vaccination as an adjunct to usual care in the setting of Covid-19 are awaited.

Abbreviations: BCG Bacille Calmette-Guérin

Vaccination Coverate Rate

Pneumococcal Vaccination; Influenza, Influenza; UK United Kingdom

using DPT1/3 as proxy for National Childhood VCR; 3 No data available, replaced with average of reported data for 25 countries

Vaccine Coverage of Pre-School Age Children In France in 2000

BCG vaccination reported mandatory from 1947 to 1995 referencing "Tuberkulosevaksinasjon -veileder for helsepersonell

European Centre for Disease Prevention and Control. Seasonal influenza vaccination and antiviral use in EU/EEA Member States -Overview of vaccine recommendations for 2017-2018 and vaccination coverage rates for 2015-2016 and 2016-2017 influenza seasons

The Effect of Oral Polio Vaccine at Birth on Infant Mortality: A Randomized Trial

Non-specific effects of standard measles vaccine at 4.5 and 9 months of age on childhood mortality: randomised controlled trial

Heterologous) Protection of Neonatal BCG Vaccination Against Hospitalization Due to Respiratory Infection and Sepsis

Non-specific Effect of Vaccines: Immediate Protection against Respiratory Syncytial Virus Infection by a Live Attenuated Influenza Vaccine

The Vaccine Trialist Group. A role for Streptococcus pneumoniae in virusassociated pneumonia

Correlation between universal BCG vaccination policy and reduced morbidity and mortality for COVID-19: an epidemiological study

Connecting BCG Vaccination and COVID-19: Additional Data

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

Differential COVID-19-attributable mortality and BCG vaccine use in countries

Bacille Calmette-Guérin (BCG) vaccination and COVID-19 Scientific Brief

BCG vaccination is associated with reduced malaria prevalence in children under the age of 5 years in sub-Saharan Africa

The Role of BCG/PPD-Activated Macrophages in Resistance against Systemic Candidiasis in Mice

Bacille Calmette-Guerin induces NOD2-dependent nonspecific protection from reinfection via epigenetic reprogramming of monocytes

Does BCG vaccination protect against acute respiratory infections and COVID 19? A rapid review of current evidence

Does BCG vaccination protect against acute respiratory infections and COVID-19? A rapid review of current evidence

Reducing Health Care Workers Absenteeism in Covid-19 Pandemic Through BCG Vaccine (BCG-CORONA)

BCG Vaccination to Protect Healthcare Workers Against COVID-19 (BRACE)

BCG Vaccine for Health Care Workers as Defense Against COVID 19 (BADAS)

Remdesivir for the Treatment of Covid-19 -Preliminary Report

No funding was received for this study