System Complexity in Influenza Infection and Vaccination: Effects upon Excess Winter Mortality

Rodney P. Jones; Andriy Ponomarenko

doi:10.3390/idr14030035

Jones, Rodney P.;Ponomarenko, Andriy

2022-04-21 00:00:00

Article System Complexity in Inﬂuenza Infection and Vaccination: Effects upon Excess Winter Mortality 1 , 2 Rodney P. Jones * and Andriy Ponomarenko Healthcare Analysis & Forecasting, Wantage OX12 0NE, UK Department of Biophysics, Informatics and Medical Instrumentation, Odessa National Medical University, Valikhovsky Lane 2, 65082 Odessa, Ukraine; aponom@hotmail.com * Correspondence: hcaf_rod@yahoo.co.uk Abstract: Unexpected outcomes are usually associated with interventions in complex systems. Ex- cess winter mortality (EWM) is a measure of the net effect of all competing forces operating each winter, including inﬂuenza(s) and non-inﬂuenza pathogens. In this study over 2400 data points from 97 countries are used to look at the net effect of inﬂuenza vaccination rates in the elderly aged 65+ against excess winter mortality (EWM) each year from the winter of 1980/81 through to 2019/20. The observed international net effect of inﬂuenza vaccination ranges from a 7.8% reduction in EWM estimated at 100% elderly vaccination for the winter of 1989/90 down to a 9.3% increase in EWM for the winter of 2018/19. The average was only a 0.3% reduction in EWM for a 100% vaccinated elderly population. Such outcomes do not contradict the known protective effect of inﬂuenza vaccination against inﬂuenza mortality per se—they merely indicate that multiple complex interactions lie behind the observed net effect against all-causes (including all pathogen causes) of winter mortality. This range from net beneﬁt to net disbeneﬁt is proposed to arise from system complexity which includes environmental conditions (weather, solar cycles), the antigenic distance between constantly emerg- ing circulating inﬂuenza clades and the inﬂuenza vaccine makeup, vaccination timing, pathogen interference, and human immune diversity (including individual history of host-virus, host-antigen interactions and immunosenescence) all interacting to give the observed outcomes each year. We Citation: Jones, R.P.; Ponomarenko, propose that a narrow focus on inﬂuenza vaccine effectiveness misses the far wider complexity of A. System Complexity in Inﬂuenza winter mortality. Inﬂuenza vaccines may need to be formulated in different ways, and perhaps Infection and Vaccination: Effects administered over a shorter timeframe to avoid the unanticipated adverse net outcomes seen in upon Excess Winter Mortality. Infect. around 40% of years. Dis. Rep. 2022, 14, 287–309. https:// doi.org/10.3390/idr14030035 Keywords: inﬂuenza; vaccination; pathogen interference; immune diversity; antigenic distance; Academic Editor: Joan Puig-Barberà winter mortality Received: 25 March 2022 Accepted: 18 April 2022 Published: 21 April 2022 1. The Excess Winter Mortality (EWM) Calculation Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in This and previous studies use a rolling/moving EWM calculation which shows deaths published maps and institutional afﬁl- in the four ‘winter ’ months as a percentage difference to the preceding eight non-winter iations. months. Since winter infectious outbreaks can occur early or late, and that ‘winter ’ is more objective near the equator the calculation is performed as a rolling or moving percentage difference. Hence, we start at the ﬁrst 12-months data, where the EWM calculation is: EWM = average deaths (September to December) average deaths (January to August). Copyright: © 2022 by the authors. Move forward one month and recalculate, etc. EWM for that winter is the maximum Licensee MDPI, Basel, Switzerland. value. In the northern hemisphere temperate zone, the EWM most commonly reaches a This article is an open access article maximum at the 12-months ending in March. The EWM calculation is very reliable and distributed under the terms and the only instances when it will give an answer lower than actual is when there is a highly conditions of the Creative Commons unusual summer heat wave or when the winter of the preceding year occurs very late, and Attribution (CC BY) license (https:// the current winter occurs early. Both can be overcome by retrospective adjustment. creativecommons.org/licenses/by/ 4.0/). Infect. Dis. Rep. 2022, 14, 287–309. https://doi.org/10.3390/idr14030035 https://www.mdpi.com/journal/idr Infect. Dis. Rep. 2022, 14 288 2. Introduction During the past 70 years both inﬂuenza epidemics and vaccination have been largely viewed from a narrow single pathogen perspective. From this point of view, efﬁcient epidemic control for an antigenically variable pathogen, such as inﬂuenza, is achieved by regular immunization of most of the human population—within the constraints of cost beneﬁt [1]. However, more recently it has become apparent that inﬂuenza outbreaks, inﬂuenza vaccination and the observed excess winter (all-cause) mortality operate within a complex system of: 1. human immune variability which includes gender, chronological and immune age, individual history of host-virus and host-antigen interactions, ethnicity, persistent pathogens, genetic mutations, epigenetic factors, psychological stress, and metabolic health [2–8], 2. the role of meteorological variables on inﬂuenza (and other respiratory pathogen) survival and transmission [9–14], 3. inﬂuenza virus evolution [15,16], 4. the variable spatiotemporal spread and distribution of inﬂuenza strains and mutations (clades) each year [17,18]. 5. the pathogenicity of inﬂuenza being the result of a complex system of interactions between the inﬂuenza viruses, other viruses, the host, anthropogenic interventions, and secondary infections [19–21]. 6. the totality of winter pathogen-induced deaths which is a composite of (co)infection by multiple pathogens [22–25]. All these factors combine to give remarkably high inter- and intra-national variation in excess winter mortality (EWM) during each inﬂuenza season, along with highly complex long-term trends [26], and equally remarkable variations in vaccine effectiveness between seasons [27]. A recent study has suggested that the long-term average for the net effect of inﬂuenza vaccination upon EWM was undetectable [26], because the whole system is far more complex than just inﬂuenza and inﬂuenza vaccination. This same observation has also been noted in two other large studies where, during a time of rapidly rising inﬂuenza vaccination in the elderly, no net reduction in EWM could be discerned [28,29]. As an example of the shift to a more complex system view of inﬂuenza epidemics and inﬂuenza vaccination, Table 1 shows the results of a search using Google Scholar regarding the number of hits for a variety of inﬂuenza-related complex system queries. Clearly some of these hits may not be relevant or be duplicates, nevertheless they indicate a general trend toward system complexity thinking. Table 1. Searches on inﬂuenza system complexity using Google Scholar. Search conducted on 6th October 2021. Search String Documents Identiﬁed Inﬂuenza epidemics “complex systems” 94,800 Inﬂuenza and “systems biology” 22,000 Complex system dynamics pandemic inﬂuenza 18,800 Interactions inﬂuenza and “other pathogens” 16,200 Inﬂuenza and “pathogen interactions” 14,600 Inﬂuenza and “complex system” 10,900 Inﬂuenza vaccination and “complex system” 4520 Inﬂuenza and “pertussis complex relationship” 570 Key features of complex systems are unexpected dynamic and unexpected outcomes of interventions, called ‘emergent behavior ’, bifurcation (or tipping) points where a division into branches or sub-groups occurs, i.e., fractal behavior, and unrealized multiple equilibria or steady states [30–36]. The population dynamics of pathogens and pathogen-host interac- Infect. Dis. Rep. 2022, 14 289 tions depend on multiple factors, including natural and anthropogenic factors, along with other hidden factors, and are often underestimated. Regarding the multiple equilibria, the immune system does indeed exist in multiple steady states [37–42]. Such immune-endocrine steady states can correspond to certain illnesses, such as Gulf War illness and chronic fatigue syndrome [38]. Infectious disease models, likewise, show multiple equilibria [43–45]. In such a complex system, inﬂuenza vaccination may yield unexpected outcomes, i.e., may beneﬁt one group, have no effect on another or cause disbeneﬁt in another. Hence, assessing overall, or net, vaccine beneﬁts in such a complex system may be less than straightforward. One study which used a systems biology approach screened multiple morbidities, anthropometric measurements, and biochemical parameters and concluded that only relative lymphopenia (decreased percent of lymphocytes in WBC differential), OR 0.94, (95% CI 0.88–0.99); vitamin B12 deﬁciency OR 0.99, (0.99–1.00); and hyperhomocysteinaemia OR 1.15 (0.99–1.32) showed potential to predict an inﬂuenza vaccine response (as antibody production) for the 2003/04 trivalent vaccine [46]. For more accurate evaluation of inﬂuenza vaccine efﬁcacy parameters of cellular immunity and their interaction with other factors should also be measured. However, this seemingly indicated that biochemical health (howsoever determined) may be a neglected key parameter in determining antibody production (but not necessarily vaccine efﬁcacy). Hence factors such as obesity and multi-morbidity are indirect measures of their effects upon the individual’s biochemical balance. The above issues are neatly summarized in a recent study, which demonstrated that all-cause excess winter mortality (EWM) is the output of an exceedingly complex system which exhibits long-term undulations in EWM—and therefore implies the existence of potential hidden and unexpected ‘emergent’ outcomes [26]. The methodology behind the calculation of EWM has been extensively discussed in two previous articles [26,47]. In summary, it calculates the percentage of excess winter deaths for the four winter months relative to the eight non-winter months. The calculation is performed on a running/moving basis to detect which four-month period gives the maximum difference. This then allows for years in which inﬂuenza outbreaks may occur very early or late and allows for winter in the southern hemisphere. This study contains several parts. In the ﬁrst is an overview of the international trends in EWM, especially focusing on high inter- and intra-national spatiotemporal granularity in each year and what this may imply regarding the complexity of each winter. We then investigate if there is a relationship between international EWM and proportion of those aged 65+ who receive inﬂuenza vaccinations. This is achieved using two data sets, namely, age 65+ vaccinated data and doses of inﬂuenza vaccine distributed. The latter is then converted into an age 65+ vaccinated equivalent. Both are previously described [26]. Rather than conduct this analysis over a longitudinal scale, as was done previously [26], the analysis focuses on each winter and, speciﬁcally, on the differences in EWM as a function of the differences in elderly inﬂuenza vaccination rates between world countries. The emphasis is on the detection of unexpected or emergent outcomes which complexity theory indicates should exist. EWM is a key tool, because it measures the net effects inherent in each winter and can thereby detect unexpected or emergent behavior. 3. Materials and Methods 3.1. Sources of the Data Monthly deaths and rolling/moving EWM calculations for a range of countries were taken from a previous study [26]. Proportion of persons aged 65+ vaccinated in each country over time was also taken from the previous study [26]. Data relating to vaccine effectiveness in those aged 65+ in the USA was from the Center for Disease Control and Prevention (CDC) [11]. Annual estimates of adult obesity since the 1980s for world countries was obtained from the World Health Organization (WHO) [48] and the Global Obesity Observatory [49]. Infect. Dis. Rep. 2022, 14 290 3.2. Adjusting EWM for Each Country to a US-Equivalent The USA has the most available data for rates of vaccination in those aged 65+, plus EWM data [26]. It therefore makes sense to adjust the EWM of all other countries to a US-equivalent. This was achieved by adjusting the data from all countries using the median EWM for each country compared to that of the USA. EWM data for each country was adjusted such that the adjusted EWM has a median equal to that seen in the USA as detailed in the previous study [26]. 3.3. Method for Excluding Outlying EWM Values For smaller countries with lower deaths per annum there can occasionally be statis- tically high/low values for EWM. For countries lying close to 0% vaccination, adjusted values of EWM lower than 5% and higher than 20% were trimmed. For countries with higher rates of vaccination a different rule was applied, such that values were only ex- cluded from the study if they were markedly higher/lower than all other countries. This sometimes occurs for data from smaller countries where Poisson randomness becomes more signiﬁcant. Exclusion is required to avoid the undue effect of outlying values on linear regression based on the least-squares methodology. The distance squared means that outlying values are unduly weighted in the regression. Future studies on this topic could use weighted regression without trimming; however, this is unlikely to make a material change to the conclusions. 3.4. Adjustment of EWM for Obesity Relative to the USA As in the previous study EWM data for each year was adjusted to give the equivalent to that in the USA [26]. Obesity data for world countries in 2016 was plotted against the median EWM for each country over the period 1990 to 2020. This gave a slope of 0.2, i.e., for each percentage point increase in obesity the median EWM increases by 0.2% (See Figure A1 in the Appendix A). This was higher than that observed in an earlier study [26], and so the effect of the slope upon the relationship between EWM and inﬂuenza vaccination was evaluated for values of the slope between 0.02 and 0.3. (Table S1 in the Supplementary Material). The R-squared for this relationship reached a maximum at a value of the slope equal to 0.12 (Figure A2 in the Appendix A). Since all countries in this study had a level of adult obesity less than the USA the adjustment factor for EWM was then as follows: Obesity Adjusted EWM = Raw EWM + [adult obesity in USA (%) adult obesity in coun- try A (%)] 0.12. This calculation is repeated for each year. 3.5. EWM in US States since 2008 Monthly deaths have been available for US states since January 2008 [26]. The median EWM for each state was calculated up to the winter of 2019/20 and adjusted EWM was calculated as per Section 3.2. The proportion of persons aged 65+ vaccinated for inﬂuenza for each state was estimated by multiplying the US average by the ratio of nursing home residents vaccinated in each state relative to the US average [50]. 3.6. Data Manipulation All data was manipulated using Microsoft Excel. Linear regression was performed using the “Add Trendline” function. 4. Results 4.1. EWM Shows Extreme Spatiotemporal Volatility Excess winter mortality (EWM) varies considerably from one year to the next and Figure 1 shows this variation using a rolling/moving EWM calculation for up to 143 countries and states/provinces. In Figure 1 the EWM for each country has been adjusted up/down by the ratio of the median EWM for the USA divided by the median EWM for each coun- Infect. Dis. Rep. 2022, 14 291 try [26]. Note that the inter-quartile range only covers the 50% of countries closest to the international median for each winter. Figure 1. Upper and lower quartile for a rolling EWM calculation for 134 countries and 34 states/ provinces (Australia, Canada, Germany). Due to data availability, there is a maximum of 143 coun- tries/states for each winter. The variation is illustrated by showing the international upper and lower quartile. As can be seen, EWM reached a minimum in the winters of 2000/01 and 2013/14 with a median = 8.8% (for clarity the line for the median is not shown) and a maximum in the winters of 1999/00 (median = 17.2%) and 2014/15 (median = 16.5%). However, the inter quartile range (IQR) reached its maximum extent of 16.2% for the winter of 1989/90 and its minimum extent of 6.9% for the winter of 2019/20, just before the COVID-19 pandemic. A high IQR indicates extreme differences around the world. The sharpness of each winter peak measures the differences in timing between countries. Note the cluster of 3 high years between 2014/15 and 2017/18 and 4 high years between 1995/96 and 1999/00. Inﬂuenza vaccination is therefore being applied into a system showing high intrinsic international variation, clustering of high EWM, and year-to-year volatility. To determine if a wide IQR is speciﬁc to world countries Figure 2 shows an identical rolling EWM analysis to that in Figure 1 using 417 local government areas (LGA) within the UK (2001 to 2021). Figure 2 also contains the ﬁrst and second wave of COVID-19 as an illustration of an infectious outbreak with high spatiotemporal variation. The key point is that the range for the upper and lower quartile within the UK is very close to that for world countries, even though world countries range from near the equator to close to the poles, i.e., even the within-country variation in EWM is profoundly high. The IQR in Figure 2 is not an artefact of LGA size since the median size (as deaths per annum) in the two tails is not greatly different from the middle 50% of EWM values (the IQR) and is at least 3-times higher than the minimum size threshold (400 deaths per annum) applied to the international data. Indeed 50% of UK LGA have over 1500 deaths per annum and 75% are higher than 1000 deaths per annum. The second point is that both the upper and lower quartile in Figure 2 is made up from unique winter behavior, which is indicative of differing spatial spread of the causative agents—as observed for the two COVID-19 waves (last two peaks). Note that the events in the winters of 2014/15 and 2017/18 have an upper quartile equal in magnitude to the Infect. Dis. Rep. 2022, 14 292 Infect. Dis. Rep. 2022, 14, FOR PEER REVIEW 6 second COVID-19 winter. It is proposed that it is the spatiotemporal spread of pathogens within the UK which drives the variation, of which inﬂuenza made a signiﬁcant (but not exclusive) contribution prior to the arrival of COVID-19. Figure 2. Upper and lower quartile for a rolling/moving EWM calculation covering 417 local Figure 2. Upper and lower quartile for a rolling/moving EWM calculation covering 417 local government areas (LGA) across the UK each with fewer than 5000 deaths per annum. EWM for each government areas (LGA) across the UK each with fewer than 5000 deaths per annum. EWM for each LGA LGA has not has not been been adjusted adjusted.. The winter of 2013/14 in the UK, which had the lowest EWM, was remarkably mild The key point is that the range for the upper and lower quartile within the UK is very and wet [14], which seemingly led to low levels of inﬂuenza (and other winter pathogen) close to that for world countries, even though world countries range from near the equator activity and mortality [15]. This is consistent with cold-dry conditions favoring inﬂuenza to close to the poles, i.e., even the within-country variation in EWM is profoundly high. spread in temperate countries [13]. Hence, low EWM is associated with low levels of winter The IQR in Figure 2 is not an artefact of LGA size since the median size (as deaths pathogens while high EWM is associated with high levels of pathogens (as per COVID-19). per annum) in the two tails is not greatly different from the middle 50% of EWM values While high spatiotemporal variation in infectious outbreaks is well known to epidemi- (the IQR) and is at least 3-times higher than the minimum size threshold (400 deaths per ologists, the implications to inherent complexity and unexpected or emergent behavior annum) applied to the international data. Indeed 50% of UK LGA have over 1500 deaths may have been largely overlooked. per annum and 75% are higher than 1000 deaths per annum. The second point is that both the upper and lower quartile in Figure 2 is made up 4.2. Inﬂuenza Vaccination in the Elderly from unique winter behavior, which is indicative of differing spatial spread of the Rates of inﬂuenza vaccination vary widely between world countries. The median causative agents—as observed for the two COVID-19 waves (last two peaks). Note that for vaccination rates between countries in those aged 65+ ranges from 4% in 1988/89 the events in the winters of 2014/15 and 2017/18 have an upper quartile equal in (maximum 45%) to 48% in 2019/20 (maximum 85%). Countries with highest vaccination magnitude to the second COVID-19 winter. It is proposed that it is the spatiotemporal rates for age 65+ changes over time with the Netherlands highest between 2000/01 to spread of pathogens within the UK which drives the variation, of which influenza made 2008/09 (range 76% to 83%), Mexico was the highest in 2009/10 during the Swine ﬂu a significant (but not exclusive) contribution prior to the arrival of COVID-19. pandemic (88.2%), and brieﬂy highest between 2013/14 and 2014/15 (79% to 82%), while The winter of 2013/14 in the UK, which had the lowest EWM, was remarkably mild South Korea was highest in 2011/12 and 2012/13, and from 2015/16 onward (up to 86% and wet [14], which seemingly led to low levels of influenza (and other winter pathogen) vaccinated) [51]. Hence there is a sufﬁciently wide range in vaccination rates for every year activity and mortality [15]. This is consistent with cold-dry conditions favoring influenza during the study to enable evaluation of the role of vaccination on EWM. spread in temperate countries [13]. Hence, low EWM is associated with low levels of If inﬂuenza vaccination has a net protective effect the slope of the relationship be- winter pathogens while high EWM is associated with high levels of pathogens (as per tween EWM and proportion aged 65+ vaccinated should have a negative slope. Figure 3 COVID-19). gives one example of such analysis for the winter of 2017/18 where the resulting slope is While high spatiotemporal variation in infectious outbreaks is well known to positive (disbeneﬁt) rather than negative. The R-squared for Figure 3 was 0.156. Such low epidemiologists, the implications to inherent complexity and unexpected or emergent values of R-squared are typical for each year and arise as a direct consequence of the high behavior may have been largely overlooked. international variation demonstrated in Figures 1 and 2. 4.2. Influenza Vaccination in the Elderly Rates of influenza vaccination vary widely between world countries. The median for vaccination rates between countries in those aged 65+ ranges from 4% in 1988/89 Infect. Dis. Rep. 2022, 14, FOR PEER REVIEW 7 (maximum 45%) to 48% in 2019/20 (maximum 85%). Countries with highest vaccination rates for age 65+ changes over time with the Netherlands highest between 2000/01 to 2008/09 (range 76% to 83%), Mexico was the highest in 2009/10 during the Swine flu pandemic (88.2%), and briefly highest between 2013/14 and 2014/15 (79% to 82%), while South Korea was highest in 2011/12 and 2012/13, and from 2015/16 onward (up to 86% vaccinated) [51]. Hence there is a sufficiently wide range in vaccination rates for every year during the study to enable evaluation of the role of vaccination on EWM. If influenza vaccination has a net protective effect the slope of the relationship between EWM and proportion aged 65+ vaccinated should have a negative slope. Figure 3 gives one example of such analysis for the winter of 2017/18 where the resulting slope is positive (disbenefit) rather than negative. The R-squared for Figure 3 was 0.156. Such low Infect. Dis. Rep. 2022, 14 293 values of R-squared are typical for each year and arise as a direct consequence of the high international variation demonstrated in Figures 1 and 2. 35% y = 0.073x + 0.1369 30% 25% 20% 15% 10% 5% 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Proportion aged 65+ vaccinated Figure 3. Slope for the relationship between obesity adjusted EWM and proportion aged 65+ vacci- Figure 3. Slope for the relationship between obesity adjusted EWM and proportion aged 65+ vaccinated for 80 world countries during the winter of 2017/18. Linear regression as the dotted line. nated for 80 world countries during the winter of 2017/18. Linear regression as the dotted line. Note that the winter of 2017/18 in both Figures 1 and 2 shows unusually high EWM. Note that the winter of 2017/18 in both Figures 1 and 2 shows unusually high EWM. A slope of 0.073 in Figure 3 implies that a population with 100% vaccinated elderly persons A slope of 0.073 in Figure 3 implies that a population with 100% vaccinated elderly will have an EWM (at the US equivalent) which is 7.3% higher than if there had been no persons will have an EWM (at the US equivalent) which is 7.3% higher than if there had vaccination, i.e., an adverse outcome. In Figure 3 raw EWM for each country was ﬁrst been no vaccination, i.e., an adverse outcome. In Figure 3 raw EWM for each country was adjusted to the equivalent to the USA using the median EWM and then further adjusted to first adjusted to the equivalent to the USA using the median EWM and then further match US levels of obesity via the effect of obesity on international EWM (as per #4 below). adjusted to match US levels of obesity via the effect of obesity on international EWM (as Similar analysis to Figure 3 was conducted each year from the winter of 1987/88 per #4 below). through to 2019/20. Four alternative scenarios for each year were performed, namely: Similar analysis to Figure 3 was conducted each year from the winter of 1987/88 1. Data from all available countries through to 2019/20. Four alternative scenarios for each year were performed, namely: 2. The 50 countries with the highest number of years of available data 1. Data from all available countries 3. #1 plus data from US states (available for 2007/08 onward) [26] 2. The 50 countries with the highest number of years of available data 4. #1 plus additional adjustment of each country for difference in obesity relative to the 3. #1 plus data from US states (available for 2007/08 onward) [26] USA [48,49] 4. #1 plus additional adjustment of each country for difference in obesity relative to the These four scenarios were performed to demonstrate that the resulting slope and USA [48,49] intercept are robust. The resulting values for each scenario are given in Figure 4. These four scenarios were performed to demonstrate that the resulting slope and The year shows the winter ending in that year, hence, 1989 = 1988/89 through to intercept are robust. The resulting values for each scenario are given in Figure 4. 2020 = 2019/20. EWM for 2020 was calculated at the end of March to avoid distortion due to the COVID-19 pandemic. In this study complete or partial data was available for 97 countries. The minimum available data pertained to 50 countries in 1987/88 through to a maximum of 85 in 2013/14 and 2014/15. Countries were ranked by years of available data. The top 50 group was an arbitrary division. The most complete data were for members of the European Union, Australia, New Zealand, USA, and Canada. Table S2 in the Supplementary Material shows the number of available countries for all countries and the top 50 countries. The top 50 countries contain ﬁve small countries (Greenland, Malta, Iceland, Liechtenstein, and Luxembourg) where Poisson randomness leads to occasional instances of EWM values which were excluded. Hence, count of available data for the top 50 range from 37 in 1987/88 through to 50, median is 46. For the 47 other countries available data ranges from 11 to 36, with a median of 27. From Figure 4 the intercept for the data with additional obesity adjustment is slightly higher than the other scenarios. This is because all countries have lower levels of adult obesity compared with the USA. The gap between obesity in the USA and other countries rises with time. The maximum gap was a 12.9% difference in 1980 rising to a 34.7% Median-adjusted EWM plus obesity adjustment Infect. Dis. Rep. 2022, 14 294 Infect. Dis. Rep. 2022, 14, FOR PEER REVIEW 8 (percentage point) difference between Japan and the USA in 2019. Hence their EWM is adjusted upward by a maximum of between 1.5% (in 1980/81) and 4.2% in 2019/20. A higher intercept is therefore to be expected. 24% Intercept All Slope All Intercept top 50 Slope top 50 Intercept incl US states Slope incl US states 20% Intercept Obesity adjusted Slope Obesity adjusted 16% 12% 8% 4% 0% -4% -8% Figure 4. Slope and intercept of the relationship between adjusted excess winter mortality (EWM) and Figure 4. Slope and intercept of the relationship between adjusted excess winter mortality (EWM) proportion aged 65+ vaccinated. A negative slope implies that the net effects of inﬂuenza vaccination and proportion aged 65+ vaccinated. A negative slope implies that the net effects of influenza are beneﬁcial while a positive slope implies the opposite. vaccination are beneficial while a positive slope implies the opposite. The slope of the relationship after obesity adjustment is highly correlated with the The year shows the winter ending in that year, hence, 1989 = 1988/89 through to 2020 slope before obesity adjustment (R-squared = 0.9798) but values of the adjusted EWM are = 2019/20. EWM for 2020 was calculated at the end of March to avoid distortion due to the 0.37% higher than the unadjusted slope (see Figure A2 in the Appendix A) because the COVID-19 pandemic. In this study complete or partial data was available for 97 countries. intercept has been increased. This is consistent with a slightly higher intercept leading to a The minimum available data pertained to 50 countries in 1987/88 through to a maximum 1.45% reduction in the slope (Table S1 in the Supplementary Material). of 85 in 2013/14 and 2014/15. Countries were ranked by years of available data. The top 50 The slope for the top 50 countries shows highest divergence mainly because there group was an arbitrary division. The most complete data were for members of the are fewer data points (as discussed above) and hence the uncertainty in the slope will be European Union, Australia, New Zealand, USA, and Canada. Table S2 in the higher. The scenario including US states has the highest number of data points each year. Supplementary Material shows the number of available countries for all countries and the The intercept and slope for each year is also shown in Figure A3 in the Appendix A. A top 5 summary 0 countri of the es. The top 5 available data 0 countries conta is given in Table in S2 five insmall co the Supplementary untries (Green Material land, M which alta, Iceland also includes , Liechtenstein, an estimate anof d Luxembour the standard g) wh deviation ere Poof isson ran the slope domn by comparing ess leads to o methods ccasional #1, #3 and #4 above. instances of EWM values which were excluded. Hence, count of available data for the top 50 ran As ge fro can also m 37 bein 1 seen, 987 the /88 t slope hrouof gh t theo 50 relationship , median is ranges 46. For t fromh e 4 6.7% 7 otfor her the count winter ries of 2003/04 (a net beneﬁcial effect) up to +7.2% for the winters of 2014/15 and 2017/18 available data ranges from 11 to 36, with a median of 27. (net disbeneﬁt). In addition, cyclic behavior is also apparent with the ﬁrst cycle rising to From Figure 4 the intercept for the data with additional obesity adjustment is slightly a maximum in the winter of 1989/99. After 1998/99 there is a trend down to 2003/04, higher than the other scenarios. This is because all countries have lower levels of adult another trend up to a plateau, and a period of instability beyond 2012/13. Roughly half the obesity compared with the USA. The gap between obesity in the USA and other countries data lie above/below a slope of 0%, i.e., the point of no net effect. rises with time. The maximum gap was a 12.9% difference in 1980 rising to a 34.7% (percentage point) difference between Japan and the USA in 2019. Hence their EWM is 4.3. Comparison with a Previous Study adjusted upward by a maximum of between 1.5% (in 1980/81) and 4.2% in 2019/20. A A previous study gave an apparent zero slope for the relationship between adjusted higher intercept is therefore to be expected. EWM, and proportion elderly vaccinated using data over a 30-year period [26]. The slope of the relationship after obesity adjustment is highly correlated with the slope before obesity adjustment (R-squared = 0.9798) but values of the adjusted EWM are 0.37% higher than the unadjusted slope (see Figure A2 in the Appendix A) because the Intercept or slope 2020 Infect. Dis. Rep. 2022, 14, FOR PEER REVIEW 9 intercept has been increased. This is consistent with a slightly higher intercept leading to a 1.45% reduction in the slope (Table S1 in the Supplementary Material). The slope for the top 50 countries shows highest divergence mainly because there are fewer data points (as discussed above) and hence the uncertainty in the slope will be higher. The scenario including US states has the highest number of data points each year. The intercept and slope for each year is also shown in Figure A3 in the Appendix A. A summary of the available data is given in Table S2 in the Supplementary Material which also includes an estimate of the standard deviation of the slope by comparing methods #1, #3 and #4 above. As can also be seen, the slope of the relationship ranges from −6.7% for the winter of 2003/04 (a net beneficial effect) up to +7.2% for the winters of 2014/15 and 2017/18 (net disbenefit). In addition, cyclic behavior is also apparent with the first cycle rising to a maximum in the winter of 1989/99. After 1998/99 there is a trend down to 2003/04, another trend up to a plateau, and a period of instability beyond 2012/13. Roughly half the data lie above/below a slope of 0%, i.e., the point of no net effect. 4.3. Comparison with a Previous Study Infect. Dis. Rep. 2022, 14 295 A previous study gave an apparent zero slope for the relationship between adjusted EWM, and proportion elderly vaccinated using data over a 30-year period [26]. As was pointed out in the previous study [26], it would be highly unlikely for As was pointed out in the previous study [26], it would be highly unlikely for inﬂuenza influenza vaccination to have zero net effect on EWM in every single year and to this end vaccination to have zero net effect on EWM in every single year and to this end a cumulative a cumulative sum of differences (CUSUM) is relevant. The CUSUM of the slope over time sum of differences (CUSUM) is relevant. The CUSUM of the slope over time is given is given in Figure 5. A CUSUM is a useful tool to reveal when the behavior shows a sudden in Figure 5. A CUSUM is a useful tool to reveal when the behavior shows a sudden transition [52], which leads to a change in slope in the CUSUM. Over this 39-year period transition [52], which leads to a change in slope in the CUSUM. Over this 39-year period there are two extended periods of net benefit, namely, 1986/87 to 1994/95 and 2000/01 to there are two extended periods of net beneﬁt, namely, 1986/87 to 1994/95 and 2000/01 2006/07, and two periods of net dis-benefit, namely, 1995/96 to 1999/00 and from 2008/09 to 2006/07, and two periods of net dis-beneﬁt, namely, 1995/96 to 1999/00 and from onwards. 2008/09 onwards. 0% -5% -10% -15% -20% -25% -30% Figure 5. CUSUM of the annual value of the slope. Figure 5. CUSUM of the annual value of the slope. There are two periods of high instability, namely 1980/81 to 1985/86 and 2013/14 to 2019/20. Over the entire 39 years the overall average net beneﬁt is only 0.4% (percentage point) reduction in EWM per annum, at a theoretical 100% of elderly vaccinated, i.e., a CUSUM of 15% divided by 39 years. In comparison, for the 14-year period ending 1994/95 the average net beneﬁt is a 2% (percentage point) reduction in EWM per annum at 100% elderly vaccination. The perception of the net beneﬁt of inﬂuenza vaccination depends entirely upon when the study is conducted. Indeed, most studies only cover a limited number of years. 4.4. Further Validation of the Results The previous study [26] also included a second large data set where EWM was plotted against total vaccine doses per 1000 total population (all-age) which covered the winters 1980/81 through to 2012/13. Does this data behave in the same way as that used in Figure 4? Figure 6 shows the output from such analysis where the slope of the EWM versus doses per 1000 population data is plotted alongside the slope for proportion elderly vaccinated data. The vaccine doses distributed data has ﬁrst been adjusted for the fact that this method always gives a greater value than that from the elderly vaccinated data. This relationship is shown in Figure 7 where the slope from doses distributed must ﬁrst be multiplied by 0.4936 to give an equivalent slope to that from the proportion aged 65+ study. From the comparison of the two data sets a further period of instability operates between 1980/81 and 1986/87. However, the point has been established that both data sets mirror each other. CUSUM net benefit at 100% vaccination 1980/81 1982/83 1984/85 1986/87 1988/89 1990/91 1992/93 1994/95 1996/97 1998/99 2000/01 2002/03 2004/05 2006/07 2008/09 2010/11 2012/13 2014/15 2016/17 2018/19 Infect. Dis. Rep. 2022, 14, FOR PEER REVIEW 10 There are two periods of high instability, namely 1980/81 to 1985/86 and 2013/14 to 2019/20. Over the entire 39 years the overall average net benefit is only 0.4% (percentage point) reduction in EWM per annum, at a theoretical 100% of elderly vaccinated, i.e., a CUSUM of −15% divided by 39 years. In comparison, for the 14-year period ending 1994/95 the average net benefit is a 2% (percentage point) reduction in EWM per annum at 100% elderly vaccination. The perception of the net benefit of influenza vaccination depends entirely upon when the study is conducted. Indeed, most studies only cover a limited number of years. 4.4. Further Validation of the Results The previous study [26] also included a second large data set where EWM was plotted against total vaccine doses per 1000 total population (all-age) which covered the winters 1980/81 through to 2012/13. Does this data behave in the same way as that used in Figure 4? Figure 6 shows the output from such analysis where the slope of the EWM versus doses per 1000 population data is plotted alongside the slope for proportion elderly vaccinated data. The vaccine doses distributed data has first been adjusted for the fact that this method always gives a greater value than that from the elderly vaccinated data. This relationship is shown in Figure 7 where the slope from doses distributed must first be multiplied by 0.4936 to give an equivalent slope to that from the proportion aged 65+ study. From the comparison of Infect. Dis. Rep. 2022, 14 296 the two data sets a further period of instability operates between 1980/81 and 1986/87. However, the point has been established that both data sets mirror each other. 10% Elderly vaccinated - top 50 >67 countries, >33% Elderly vaccinated - all maximum vaccinated 8% Doses adjusted >50 countries, 6% >12% maximum vaccinated 4% 2% 0% -2% -4% -6% Infect. Dis. Rep. 2022, 14, FOR PEER REVIEW 11 -8% Figure 6. Comparison of the slope for the two large data sets. Figure 6. Comparison of the slope for the two large data sets. 6% y = 0.4936x R² = 0.8694 4% 2% 0% -2% -4% -6% -10% -8% -6% -4% -2% 0% 2% 4% 6% 8% 10% 12% Slope from vaccine doses distributed Figure 7. Comparison of calculated slope in the relationship between adjusted EWM and proportion Figure 7. Comparison of calculated slope in the relationship between adjusted EWM and proportion vaccinated using either proportion elderly aged 65+ vaccinated or vaccine doses distributed per 1000 vaccinated using either proportion elderly aged 65+ vaccinated or vaccine doses distributed per population, in the overlap years 1988/89 to 2012/13. 1000 population, in the overlap years 1988/89 to 2012/13. 4.5. The Values for the Slope Follow an Extreme Value Distribution 4.5. The Values for the Slope Follow an Extreme Value Distribution Using the 40 years of available data, and ignoring the fact that the trend may have Using the 40 years of available data, and ignoring the fact that the trend may have cyclic cyclic e elements, lements, allows allows an analysis alysis of the of the fr fr equency equency distributi distribution on for for th thee slope. The slope. The average average value of the slope for each year was determined from #1, #2, #4 (plus 0.66% to account for value of the slope for each year was determined from #1, #2, #4 (plus 0.66% to account for the difference in the obesity adjusted slope identiﬁed in Section 4.2) above plus available the difference in the obesity adjusted slope identified in Section 4.2) above plus available data from the vaccine doses distributed data after adjustment as in Figure 7. Data was data from the vaccine doses distributed data after adjustment as in Figure 7. Data was aggregated into 1% increments in the value of the average slope, and this is presented aggregated into 1% increments in the value of the average slope, and this is presented in Figure 8. The average value for the slope is −0.3%, the median value is −1.2% and a slope of −2% to −3% represents the most frequent value (the mode). The distribution is right skewed, and a negative slope occurs on 63% of occasions, while 58% of the values lie in the range 0% to −5%. The best description is that the shape of the distribution resembles an extreme value distribution, or possibly the outcome of two or more extreme value distributions. The implications of an extreme value distribution will be covered in Section 5.7 of the Discussion. Slope from age 65+ vaccinated Change in EWM at 100% vaccination 1980/81 1982/83 1984/85 1986/87 1988/89 1990/91 1992/93 1994/95 1996/97 1998/99 2000/01 2002/03 2004/05 2006/07 2008/09 2010/11 2012/13 2014/15 2016/17 2018/19 Infect. Dis. Rep. 2022, 14 297 in Figure 8. The average value for the slope is 0.3%, the median value is 1.2% and a slope of 2% to 3% represents the most frequent value (the mode). The distribution is right skewed, and a negative slope occurs on 63% of occasions, while 58% of the values lie in the range 0% to 5%. The best description is that the shape of the distribution Infect. Dis. Rep. 2022, 14, FOR PEER REVIEW 12 resembles an extreme value distribution, or possibly the outcome of two or more extreme value distributions. The implications of an extreme value distribution will be covered in Section 5.7 of the Discussion. 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% Figure 8. Frequency distribution for the calculated slope (as 1% increments) in the relationship be- Figure 8. Frequency distribution for the calculated slope (as 1% increments) in the tween adjusted EWM and proportion vaccinated using either proportion elderly aged 65+ vaccinated relationsh or vaccineip between adj doses distributed ust per ed EW 1000 M population, and prop over ortion va the 40-year ccinaperiod ted using ei 1980/81 ther proporti to 2019/20. on elderly aged 65+ vaccinated or vaccine doses distributed per 1000 population, over the 5. Discussion 40-year period 1980/81 to 2019/20. This study does not in any way seek to claim that inﬂuenza vaccination does not offer a measure of protection against inﬂuenza induced death per se. We merely highlight that 5. Discussion winter is a multi-pathogen complex system, and that unexpected or emergent outcomes This study does not in any way seek to claim that influenza vaccination does not offer should be expected as ‘normal’. Figures 1 and 2 illustrate system complexity which seems a measure of protection against influenza induced death per se. We merely highlight that far higher than could arise from the action and spread of a single pathogen, i.e., inﬂuenza. winter is a multi-pathogen complex system, and that unexpected or emergent outcomes should be expected as ‘normal’. Figures 1 and 2 illustrate system complexity which seems 5.1. What Is the “Real” Long-Term Effect? far higher than could arise from the action and spread of a single pathogen, i.e., influenza. Our earlier study suggested that higher rates of inﬂuenza vaccination appeared to make no effect on the long-term trend in EWM [26]. We proposed that this may be due 5.1. What Is the “Real” Long-Term Effect? to increasing (multi) morbidity in many countries acting to mask the effects of inﬂuenza Our earlier study suggested that higher rates of influenza vaccination appeared to vaccination. However, Figure 5 gives an alternative explanation in that the apparent slope make no effect on the long-term trend in EWM [26]. We proposed that this may be due to of the relationship will depend on the time-period. The periods of beneﬁt/disbeneﬁt also increasing (multi) morbidity in many countries acting to mask the effects of influenza help to explain the high variation associated with the proportion of age 65+ vaccinated in vaccination. However, Figure 5 gives an alternative explanation in that the apparent slope the earlier study. Recall that in the earlier study levels of vaccination increased over time. of the relationship will depend on the time-period. The periods of benefit/disbenefit also Using the data behind Figure 4 and applying a 12-year rolling median/average (as help to exp an example lain the high v of a randomly ariation chosen assoc period) iated the with the apparproportion of age ent median/average 65+ vacc slopeinated between in the earlier study. Recall that in the earlier study levels of vaccination increased over time. 1996/97 to 2007/08 would be 1.5%/0.5% respectively (net beneﬁt), while the apparent slope Usin between g the d 2008/09 ata behin and d Fi2019/20 gure 4 and would applbe ying +1.2%/+1.4% a 12-year rol rling espectively median/ (net aver disbeneﬁt). age (as an example of Hence,a random over the longer ly chosen period) th term, the years ine which appare inﬂuenza nt median/ vaccination average slope has a net between beneﬁt 19 is9cancelled 6/97 to 20out 07/08 wo by theuld years be in−1.5 which %/−0.5% respecti there is net dis-beneﬁt. vely (net be The nefit), while the apparent rolling 12-year average in this study (using 12-years as a random example) goes from a net zero effect up to the slope between 2008/09 and 2019/20 would be +1.2%/+1.4% respectively (net disbenefit). 12-years ending 2008/09, reaches a maximum net beneﬁt of 1.3% for the 12-years ending Hence, over the longer term, the years in which influenza vaccination has a net 2011/12 and then shows maximum net disbeneﬁt of +1.4% for the 12-years ending 2019/20. benefit is cancelled out by the years in which there is net dis-benefit. The rolling 12-year average in this study (using 12-years as a random example) goes from a net zero effect up to the 12-years ending 2008/09, reaches a maximum net benefit of −1.3% for the 12-years ending 2011/12 and then shows maximum net disbenefit of +1.4% for the 12-years ending 2019/20. Hence the conclusion from this, and the previous study [26] that the “real” long-term slope is close to zero, is likely to be the best estimate, since the observed medium-term slope shows undulations over time. Clearly any effect due to the increasing proportion of persons aged 65+ vaccinated is being overwhelmed by other specific annual factors. Frequency −5 to −6% −4 to −5% −3 to −4% −2 to −3% −1 to −2% 0 to − 1% 0 to 1% 1 to 2% 2 to 3% 3 to 4% 4 to 5% 5 to 6% 6 to 7% 7 to 8% Infect. Dis. Rep. 2022, 14 298 Hence the conclusion from this, and the previous study [26] that the “real” long-term slope is close to zero, is likely to be the best estimate, since the observed medium-term slope shows undulations over time. Clearly any effect due to the increasing proportion of persons aged 65+ vaccinated is being overwhelmed by other speciﬁc annual factors. 5.2. Limitations of Our Earlier Hypothesis In our preceding paper we proposed that the beneﬁts of increased inﬂuenza vacci- nation were being counterbalanced by rising levels of obesity and other (multi) morbidi- ties [26]. The USA was used as a worst-case scenario. EWM in the USA was increasing at just 0.02% (percentage points) per annum [26]. A 0.8% (percentage point) increase in 40 years. However, the study of Simonsen et al. [28], which covered the somewhat shorter period of very rapid expansion in elderly inﬂuenza vaccination in the USA between 1987 to 1996 (a jump from 25% to 62% elderly vaccinated in just 9 years), was unable to detect any measurable effect on EWM. Obesity and other (multi) morbidities only increase slowly over decades and would be totally unable to overwhelm the beneﬁt of such a large and rapid expansion in elderly inﬂuenza vaccination. The same was observed to occur in Italy [29]. In Figure 6, 1987 to 1996 encompasses a 7-year period of moderate net beneﬁt followed by a 4-year period of rapidly escalating net disbeneﬁt (also illustrated in the CUSUM in Figure 5). It is this switch from net beneﬁt to net disbeneﬁt which confounded the above- mentioned studies [28,39], rather than any small increment in obesity and (multi) morbidi- ties. Indeed, as this study demonstrates, adjusting world countries to the equivalent US obesity level has little effect on the observed slope of the relationship each winter. Hence, while we concede that increasing obesity and (multi) morbidities may act slowly over decades to erode the beneﬁts of increasing elderly vaccination, it is likely that the more powerful annual effects far outweigh such long-term trends in human health status. 5.3. Adjustment for Obesity The adjustment for obesity in Section 4.2 is an example of a single parameter model. As such, it is highly likely that obesity may be acting as a proxy for the wider morbidity issues discussed in the previous study [26], which are also increasing with time. The relatively low slope for the seeming effect of ‘obesity’, i.e., a 0.12% increase in EWM for each 1% increase in obesity seems to add weight to the proposal that rising levels of morbidities are not the cause of the apparent lack of effect of inﬂuenza vaccination observed over a 40-year period in the earlier study. The real reason lies in the annual effects reported in this study. 5.4. Implications of High International Variation The high inter- and intra-national variation observed in Figures 1 and 2, along with the high scatter around the trend line in Figure 3, leads to a low R-squared. An R-squared of 0.156 was quoted for the winter of 2017/18 (Figure 3) with a similarly low value for 2014/15 of 0.1336 (as a wider example). A low R-squared implies that the principal variable, i.e., proportion of persons age 65+ vaccinated, is only explaining 13% to 16% of the observed variation in EWM. This will partly be because inﬂuenza vaccine effectiveness (VE) is itself highly variable [27]. However, the low R-squared is probably more to do with the fact that winter is a multi-pathogen complex system. Hence inﬂuenza vaccination per se is unable to exert much control over the variation in EWM. This concurs with the sometimes-unexpected results presented in Figures 4 and 6. 5.5. 2014/15 as an Example of Poor Vaccine Matching As can be seen in Figures 4 and 6 there are only 2 years with a very high net protective effect from inﬂuenza vaccination (1988/89 and 2000/01), but four years with a very high net disbeneﬁt (1998/99, 1999/2000, 2014/15, 2017/18) and all the net disbeneﬁt years Infect. Dis. Rep. 2022, 14 299 correspond with very high EWM. Hence it is difﬁcult to claim that inﬂuenza alone was responsible for the higher deaths in these four high years. The winter of 2014/15 can be used as an example since in Figure 1 it is characterized as having the highest upper quartile for world countries, as also seen in Figure 2 for UK LGAs. The Public Health England summary report covering the whole UK noted that levels of inﬂuenza-like-illness (ILI) were barely above baseline [53], which was entirely insufﬁcient to explain the unusually high mortality. Inﬂuenza A predominated early in late 2014 to early 2015 while B predominated after week 10 in 2015. Most excess deaths occurred in 2015 [53]. However, a series of antigen mismatches in both inﬂuenza(s) A and B between the vaccine used that year and the strains and variants which circulated were noted [53]. The report also noted that “A portion of 2014 to 2015 inﬂuenza A(H3N2) viruses did not grow sufﬁciently for antigenic characterization” [53]. Hence, additional hidden antigenic complexity may be involved. Vaccine uptake for those aged 65+ across the UK ranged from 68% to 76% for the four countries in the union. A mid-season estimate of vaccine effectiveness (VE) for inﬂuenza A was only 2.3% (range 48.5% to +36.1%). Inﬂuenza A(H3N2) had a VE of only 0.6%. No VE for inﬂuenza B was given in this report [53]. VE in Canada and several other countries went negative, and unusual patterns of small area deaths were noted across England and Wales [54]. Respiratory syncytial virus (RSV) was also active [53]. It should be noted that VE in the UK is determined on (ambulatory) General practi- tioner (GP) visits and hence may overestimate VE relating to deaths. For example, in a Swiss study persons aged 65+ admitted to hospital with ILI were 7.5-times more prevalent than the ILI visits to a GP surgery and the GP sample contained 5.8-times more aged 5–14, and 4.5-times more aged 15–29 [55]. Presentation at the hospital also commenced earlier than in the community [55]. Inﬂuenza activity and excess deaths during the earlier 2014 winter in Australia (south- ern hemisphere) were unremarkable [56], hence, the emerging strains/variants which contributed to high EWM in the northern hemisphere likely became more prevalent after September of 2014. It is unknown when and where they originated. We propose that low and possibly negative VE (for death) is seemingly associated with the unusually high winter deaths seen in both the UK and other northern hemisphere world countries during 2014/15. This has partly contributed to the observed net disbeneﬁt due to inﬂuenza vaccination that year. 5.6. Roles for Pathogen Interference Winter is a multi-pathogen event [22,57–59], and multiple pathogens cause inﬂuenza- like-illness (ILI), and death [59]. Interaction between pathogens is very common and is termed ‘pathogen interference’ [60,61]. Pathogen interference in coinfections can diminish or augment infection by other pathogens and has direct clinical consequences [62]. Since pathogen interference is not a widely appreciated phenomenon, it has been claimed that the imposition of lockdowns during the COVID-19 pandemic were responsible for the early decline in inﬂuenza activity during the winter of 2019/20 [63,64]. However, close inspection of weekly inﬂuenza activity ﬁgures in the UK show very clearly that inﬂuenza activity had dropped to baseline levels by week 3 of 2020 and had declined to zero during week 12 [65]. Lockdown in the UK legally came into force on Thursday 26th March 2020 [66] which is just at the point when inﬂuenza activity had already dropped to zero. In the UK, lockdowns cannot in any way be said to have contributed to the fall in inﬂuenza activity which commenced its rapid decline much earlier in the year when COVID-19 spread was gaining momentum [65]. In Canada during the 20-week period after week 11 of 2020 compared to the pervious 148 weeks a 70% decline in inﬂuenza prevalence was observed. However, respiratory syncytial virus (RSV) only declined by 54%, parainﬂuenza virus (PIV) declined by 60%, but coronaviruses (hCoVs) (excluding COVID-19s) increased by 80%, metapneumoviruses (HMPV) increased by +45%, and entero/rhino viruses (hERV) by +40% [64]. These results Infect. Dis. Rep. 2022, 14 300 indicate that while protective measures may have played a limited role, additional virus- speciﬁc factors were speciﬁcally involved. Of even greater relevance to pathogen interference is the virtual extinction of inﬂuenza B/Yamagata during the COVID-19 pandemic [67]. In Israel, pneumococcal disease in young children radically reduced during the ﬁrst year of COVID-19, mainly due to suppression of RSV, inﬂuenza viruses, and hMPV. However, hERV and PIV activities were within or above expected levels [68]. In the USA inﬂuenza and RSV activity were initially suppressed by COVID-19, How- ever, RSV then underwent an unusual resurgence during the summer of 2021 [69]. This study also demonstrated that there was considerable variation in the reduction in inﬂuenza activity between US states, with a 79% reduction in Texas through to a 28% increase in Idaho [69]. Coinfection with inﬂuenza and COVID-19 occurs at low frequency, although coinfection appears to occur more often in Asia than the USA [70]. To explain all the above requires that pathogen interference between COVID-19 and other viruses is the predominant explanatory force. We propose that pathogen interference, which has been active for many centuries, has a major role in the observed long-term cycles in EWM detailed in the previous study [26] and during the COVID-19 era. 5.7. Could Vaccine Effectiveness Be an Illusion Created by Pathogen Interference? The introduction of PCR-conﬁrmed ‘test negative’ inﬂuenza VE commenced around the early 2000s and is well recognized to rely on the assumption that the levels (and pathogenicity) of non-inﬂuenza pathogens is identical in both groups [71]. However, earlier studies consistently reported lower net VE. For example, the study of Fireman et al. [72] found that inﬂuenza vaccination only reduced mortality by 4.6% over 9 ﬂu seasons. Note that the design of this study is such that this is a net reduction, i.e., the net effect in a multi-pathogen complex system. Several studies do exist which suggest that pathogen interference is active after in- ﬂuenza vaccination in children [73,74], during pregnancy [75] and in the elderly [76]. Such observations question the fundamental assumptions behind the calculation of VE and indicate a shift to higher infection by non-inﬂuenza pathogens. As an aside, pregnancy is an example of a temporary immune steady state [77]. Indeed, the immune response to inﬂuenza vaccination is recognized to exhibit variation between individuals [78] and has been proposed to alter the balance of pathogen interference and affect the optimum timing of vaccination [79]. In light of the ﬁndings in this study this area requires far greater investigation. 5.8. Heliobiology and Additional Hidden Complexity As can be seen in Figures 4–6 the data seems to become more volatile/unstable during two periods from 1980/81 to 1986/87 and 2000/01 onwards. We propose a potential relationship with ﬂuctuations in solar radiation or, more correctly, coronal mass ejections (CME) [80]. Solar output of electromagnetic radiation (7% X-ray, gamma-ray and ultraviolet, 44% visible, 49% microwave, infrared and radio wave) is surprisingly volatile even at the level of seconds and minutes [81]. These ﬂuctuating emissions are due to coronal mass ejections (CMEs), which also include high energy protons, and tend to occur more often (but not always) at periods when solar ﬂares are most active [80]. One of the observable effects of these solar storms (CMEs) are electrical power grid anomalies (power surges and electrical transformer failures) which arise from geomagnetically induced currents [82]. CMEs and resulting electromagnetic levels have been linked to short-term ﬂuctuations in human health, immune function, morbidity, and mortality called heliobiology [83]. One review concluded that 10–15% of the population are predisposed to the adverse effects of geomagnetic variations [84]. Patients with multiple sclerosis show enhanced hospital admission during periods of geomagnetic disturbance [85]. Obscure phenomena, such as sudden infant deaths, appear to rise with sunspot activity (by implication CMEs) [86]. Infect. Dis. Rep. 2022, 14 301 Geomagnetic ﬁeld ﬂuctuations have been observed to alter gene expression [87]. Several studies have suggested that inﬂuenza pandemics are aligned with the solar cycle [88,89]. COVID-19 and other pathogen outbreaks all appear to fall into the same pattern [90]. We propose that CMEs add a hidden layer to the already complex behavior observed in the previous study [26] and this study. We offer the following tentative observations; namely, the points at which the CUSUM in Figure 5 changes to a positive slope around 1994/95 and 2007/08 occur as sunspot cycles 22 and 23 approach their minima. The two periods of instability both correspond to very intense instances of sunspots at the peaks of sunspot cycles 22 and 24. The CUSUM slope goes negative after the intense parts of the peaks in sunspot cycles 22 and 23, see chart in reference [91]. We stress this is tentative evidence, since CME magnitude and timing does not exactly follow sunspot cycles. However, a body of evidence appears to be accumulating. 5.9. Implications of an Extreme Value Distribution for the Slope Figure 8 demonstrated that the slope of the annual relationship appeared to be an example of an extreme value distribution. Extreme value distribution is commonly used to describe natural events such as temperature variation, rainfall, river ﬂow, ﬂooding, and stock market volatility [92]. The implication is that the volatility in the slope is subject to natural world complexity in which the minimum value of the slope, i.e., inﬂuenza vaccination is net protective, i.e., has a lower boundary, while the upper boundary can exhibit extreme values, i.e., inﬂuenza vaccination promotes net disbeneﬁt. Roles for CMEs (Section 5.6) and other potential contributory factors need to be further explored. It is fundamentally important to understand which factors trigger the unexpected adverse net effects of inﬂuenza vaccination. In practice, CMEs are very difﬁcult to quantify (apart from directly measuring the electromagnetic ﬂux at different points on the Earth’s surface) and are highly likely to show extreme spatiotemporal variation. 5.10. Minimum Value of the Slope In the previous study, a minimum possible slope of 10% was assumed [26] and this corresponds to 6% at 60% VE. A VE of 60% is the highest VE for persons aged 65+ ever reported in the USA [27]. A slope of 6% is demonstrated in this study to only occur once in 40 years. This study therefore questions the preliminary suggestion made in the earlier study that obesity and other (multi) morbidities may be masking the effects of inﬂuenza vaccination [26]. This is especially relevant in that adjustment of annual data for the effects of obesity in Figure 4 made little effect on the slope of the relationship (also discussed in Section 4.5). Recall that in Figure 4 the effect of obesity is most likely to be serving as a proxy for wider time-related changes in multiple morbidities [26]. 5.11. Biochemical and Immune Health One study has implicated roles for biochemical health in the response to vaccina- tion [46]. A large study is relevant to this concept. In this study the results from common biochemical tests were combined into a composite score [92]. The interesting observation was that humans had a wide range for the composite score, which was, however, relatively stable over time for each person. The population average for this score (biochemical health) only showed a small decline with age but showed a rapid decline in the weeks and months preceding death [93]. This is consistent with the nearness-to-death effect [94], where frailty, cognitive function, perceived physical health and mental wellbeing, etc., only show a rapid change as death approaches [95–98]. The suspicion is that nearness-to-death, howsoever determined, is a completely neglected variable and may imply that birth cohort effects play an additional role in long-term trends and vaccine effectiveness [99–101]. This point is raised in the context of additional hidden system complexity. The immune system consists of specialized cell populations that communicate with each other to achieve systemic immune responses. Analysis of various immune cell popu- Infect. Dis. Rep. 2022, 14 302 lation frequencies in healthy humans and their responses to diverse stimuli showed that human immune variation is continuous in nature, rather than characterized by discrete groups of similar individuals (as observed for the composite biochemical score study above) [102]. Three combinations of immune cell population frequencies were observed to deﬁne an individual’s immunotype and predict a set of functional responses to cytokine stimulation. Even though inter-individual variations in speciﬁc cell population frequencies can be large, unrelated individuals of younger age had more homogeneous immunotypes than older individuals. Across age groups, cytomegalovirus seropositive individuals dis- played immunotypes characteristic of older individuals. The conceptual framework for deﬁning immunotypes suggests the development of better therapies that appropriately modulate collective immunotypes, rather than individual immune components [102]. The above suggests that certain individuals may be more susceptible to the unintended adverse effects of inﬂuenza vaccination. This possibility requires further investigation. Indeed, do persons in the terminal decline phase of life, which occurs in the last year of life, beneﬁt equally from inﬂuenza vaccination? There are gaps in our understanding, which may be relevant to the unintended net effects of inﬂuenza vaccination. 5.12. A Potential Basis for Extreme Variation in the Net Effects of Inﬂuenza Vaccination The thrust of this paper has been that well intended interventions into a highly complex system are likely to generate unexpected outcomes. This is supported by wider research in complexity theory [30–33]. We have highlighted instances of immune and biochemical health, and of heliobiology, where differences exist among individuals within a population. We would also like to point out that the immune manipulating persistent virus, cytomegalovirus, has a major reservoir of infection in the lung [103]. This virus has been proposed to interact with inﬂuenza in the lung; however, CMV, likewise, seems to affect some individuals more so than others. All of this is then within the context of pathogen interference and the potential unintended effects of inﬂuenza vaccination upon pathogen balance and the immune response of different people. 6. Pragmatic Implications to Health Care Services Inﬂuenza vaccination is widely recommended by public health agencies as a route to reducing health service winter pressures. One of the contributing factors to this study was the observation that increasing inﬂuenza vaccination rates did not seem to be making a net contribution to the reduction in hospital winter capacity pressures [104]. This seemed to contradict the known ability of inﬂuenza vaccination to reduce inﬂuenza-related hospital admissions and death. This study conﬁrms this earlier observation that deaths, and the associated acute care prior to death, are showing unexpected outcomes. 7. Implications to Inﬂuenza Policy The economic rationale for inﬂuenza vaccination partly relies on the assumption that it has a net beneﬁcial effect against deaths [105]. The implications of the earlier study [26] and this study question this assumption. Furthermore, two large regression discontinuity studies have demonstrated that at the age 65 boundary, where inﬂuenza vaccination is widely recommended, there is no statistically detectable net beneﬁt against hospital admission and deaths [106,107]. The possibility exists that estimating inﬂuenza VE for the age 65+ group—an age range which is far too wide—is concealing further complexity, in that age is acting as a poor proxy for nearness-to-death. Policy must be based on facts and not upon ﬂawed single pathogen, simple system behavior assumptions. 8. Limitations and Future Research This study is limited by the availability of monthly data. The majority of Africa has no data and data from Asia and South America is limited. Countries with larger states/provinces/regions should conﬁrm the results of this study using sub-national data. Brazil is an ideal example, since it spans the equator. Total proportion vaccinated (all-age) Infect. Dis. Rep. 2022, 14 303 or age 65+ vaccinated can be used depending on data availability. Potential roles for other winter pathogens need to be clariﬁed. The adjustment factor based on median EWM may need to be reﬁned given the long-term cycles which seemingly characterize the trends in EWM [26]. It is unknown if these cycles are country-speciﬁc or are driven by other factors, such as heliobiology. The exact role of obesity, other morbidities and polypharmacy remains to be accurately quantiﬁed—although they represent long-term trends. Regarding the risk of death due to COVID-19, it has been noted that “polypharmacy may represent a marker of vulnerability, especially for younger groups of older adults” [108]. Japan, South Korea, and Singapore can serve as low obesity benchmarks. Future studies on this topic could use weighted regression without trimming; however, this is unlikely to make a material change to the conclusions. However, even after assuming a moderately high contribution for obesity upon EWM no signiﬁcant effect could be demonstrated on the slope of the annual data. 9. Conclusions This study has demonstrated that unexpected or emergent behavior is indeed oc- curring as an unintended effect of widespread inﬂuenza vaccination. Adverse outcomes regarding net winter mortality after inﬂuenza vaccination occur in roughly 40% of years. However, the exact relationship appears to follow long-term cycles. The existence of such cycles was demonstrated in a previous study [26]. This study appears to conﬁrm the predic- tions made in the 2010 study of Berencsi et al. [79] that vaccination has the potential to alter pathogen balance. One of the points made in their study was that the timing of vaccination may need to be modiﬁed to account for time-based prevalence of other pathogens [78]. This is a testable hypothesis, given that the date of vaccination for individuals is gener- ally readily available and that many countries also have data on the prevalence of other common winter pathogens. Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/idr14030035/s1, Table S1: Relationship between the obesity factor and the resulting correlation between All Country slope and Obesity adjusted slope. Table S2: Summary of data used in the study including inﬂuenza vaccine doses distributed and proportion aged 65+ vaccinated versus adjusted EWM. Author Contributions: Conceptualization, R.P.J.; methodology, R.P.J.; software, R.P.J.; validation, R.P.J.; formal analysis, R.P.J.; investigation, R.P.J.; resources, R.P.J.; data curation, R.P.J.; writing— original draft preparation, R.P.J.; writing—review and editing, A.P.; visualization, R.P.J. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: All data is publicly available. Copies of various data tables can be obtained from Rodney Jones on request, email: hcaf_rod@yahoo.co.uk. Conﬂicts of Interest: The authors declare no conﬂict of interest. Infect. Dis. Rep. 2022, 14, FOR PEER REVIEW 19 Infect. Dis. Rep. 2022, 14 304 Appendix A Appendix A 22% 20% y = 0.211x + 0.0748 18% 16% 14% 12% 10% 8% 6% 4% 0% 5% 10% 15% 20% 25% 30% 35% 40% Proportion of obese adults in 2016 Figure A1. Proportion of obese adults in 2016 versus the median excess winter mortality (EWM) for Figure A1. Proportion of obese adults in 2016 versus the median excess winter mortality (EWM) for 65 world countries. Footnote: Median EWM data has not been adjusted for latitude and the resulting 65 world countries. Footnote: Median EWM data has not been adjusted for latitude and the resulting slope is likely to be an over-estimate. In a previous study regarding the effects of elderly obesity on slope is likely to be an over-estimate. In a previous study regarding the effects of elderly obesity on latitude adjusted median EWM for US states the slope of the relationship was 0.07 [26]. Hence the latitude adjusted median EWM for US states the slope of the relationship was 0.07 [26]. Hence the true slope is probably somewhere between these two values. The most likely value of the slope was determined by applying values of the slope between 0.02 and 0.3 to the study data. The R-squared true slope is probably somewhere between these two values. The most likely value of the slope was value reached a maximum at a slope of 0.12 (as per Table S1 in the Supplementary Material). Note determined by applying values of the slope between 0.02 and 0.3 to the study data. The R-squared that the proportion of obese elderly aged 65+ is very similar to overall adult obesity. As this study Infect. Dis. Rep. 2022, 14, FOR PEER REVIEW 20 value reached a maximum at a slope of 0.12 (as per Table S1 in the Supplementary Material). Note demonstrates adjusting for the obesity differential to the USA makes no effect on the outcome. that the proportion of obese elderly aged 65+ is very similar to overall adult obesity. As this study demonstrates adjusting for the obesity differential to the USA makes no effect on the outcome. 8% 6% 4% y = 0.9855x - 0.0037 R² = 0.9799 2% 0% -2% -4% -6% -8% -8% -6% -4% -2% 0% 2% 4% 6% 8% Slope before obesity adjustment Figure A2. Relationship between the slope of the relationship between EWM and proportion aged Figure A2. Relationship between the slope of the relationship between EWM and proportion aged 65+ vaccinated before and after obesity adjustment. Footnote: The obesity adjustment factor is a 65+ vaccinated before and after obesity adjustment. Footnote: The obesity adjustment factor is a 0.12% 0.12% increase in EWM for a 1% increase in adult obesity as per Table S1 in the Supplementary increase in EWM for a 1% increase in adult obesity as per Table S1 in the Supplementary Material. Material. 21% 19% 17% 15% 13% 11% 9% 7% 5% 0% 20% 40% 60% 80% 100% Percentage aged 65+ vaccinated 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Figure A3. Slope and intercept of the linear regression between international EWM values and percentage elderly vaccinated over all the years included in this study. Footnote: The slope and intercept are not correlated (R-squared = 0.022). The standard deviation between the lines of best fit reaches a minimum at 0% to 15% vaccination. The intercept at 0% vaccination ranges from 16.6% in 2016/17 down to 9.6% in 2019/20. As can be seen there are only 2 years with a very high net protective effect (1988/89 and 2000/01), but four years with a very high net disbenefit (1998/99, 1999/2000, 2014/15, 2017/18) and all these years correspond with very high EWM. Hence it is difficult to claim that influenza alone was responsible for the higher deaths in these four years. International EWM Median EWM Slope after obesity adjustment Infect. Dis. Rep. 2022, 14, FOR PEER REVIEW 20 8% 6% 4% y = 0.9855x - 0.0037 R² = 0.9799 2% 0% -2% -4% -6% -8% -8% -6% -4% -2% 0% 2% 4% 6% 8% Slope before obesity adjustment Figure A2. Relationship between the slope of the relationship between EWM and proportion aged 65+ vaccinated before and after obesity adjustment. Footnote: The obesity adjustment factor is a Infect. Dis. Rep. 2022, 14 305 0.12% increase in EWM for a 1% increase in adult obesity as per Table S1 in the Supplementary Material. 21% 19% 17% 15% 13% 11% 9% 7% 5% 0% 20% 40% 60% 80% 100% Percentage aged 65+ vaccinated 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Figure A3. Slope and intercept of the linear regression between international EWM values and Figure A3. Slope and intercept of the linear regression between international EWM values and percentage elderly vaccinated over all the years included in this study. Footnote: The slope and percentage elderly vaccinated over all the years included in this study. Footnote: The slope and intercept are not corr intercept are not correlated (R-squared = 0.02 elated (R-squared = 0.022). The standard deviation 2). The standar betwee dn dev the iation bet lines of best ween the lines ﬁt of best fit reaches a minimum at 0% to 15% vaccination. The intercept at 0% vaccination ranges from 16.6% in reaches a minimum at 0% to 15% vaccination. The intercept at 0% vaccination ranges from 16.6% in 2016/17 down to 9.6% in 2019/20. As can be seen there are only 2 years with a very high net protective 2016/17 down to 9.6% in 2019/20. As can be seen there are only 2 years with a very high net protective effect (1988/89 and 2000/01), but four years with a very high net disbenefit (1998/99, 1999/2000, effect (1988/89 and 2000/01), but four years with a very high net disbeneﬁt (1998/99, 1999/2000, 2014/15, 2017/18) and all these years correspond with very high EWM. Hence it is difficult to claim 2014/15, 2017/18) and all these years correspond with very high EWM. Hence it is difﬁcult to claim that influenza alone was responsible for the higher deaths in these four years. that inﬂuenza alone was responsible for the higher deaths in these four years. References 1. D’Angiolella, L.S.; Lafranconi, A.; Cortesi, P.A.; Rota, S.; Cesana, G.; Mantovani, L.G. Costs and effectiveness of inﬂuenza vaccination: A systematic review. Ann. Ist. Super. Sanita 2018, 54, 49–57. 2. Boahen, C.; Joosten, L.; Netea, M.; Kumar, V. Conceptualization of population-speciﬁc human functional immune-genomics projects to identify factors that contribute to variability in immune and infectious diseases. Heliyon 2021, 7, e06755. [CrossRef] [PubMed] 3. Burel, J.G.; Qian, Y.; Arlehamn, C.L.; Weiskopf, D.; Zapardiel-Gonzalo, J.; Taplitz, R.; Gilman, R.H.; Saito, M.; De Silva, A.D.; Vijayanand, P.; et al. An integrated workﬂow to assess technical and biological variability of cell population frequencies in human peripheral blood by ﬂow cytometry. J. Immunol. 2017, 198, 1748–1758. [CrossRef] [PubMed] 4. Brodin, P.; Davis, M. Human immune system variation. Nat. Rev. Immunol. 2017, 17, 21–29. [CrossRef] [PubMed] 5. Lakshmikanth, T.; Muhammad, S.A.; Olin, A.; Chen, Y.; Mikes, J.; Fagerberg, L.; Gummesson, A.; Bergström, G.; Uhlen, M.; Brodin, P. Human immune system variation during 1 year. Cell Rep. 2020, 32, 107923. [CrossRef] 6. Alpert, A.; Pickman, Y.; Leipold, M.; Rosenberg-Hasson, Y.; Ji, X.; Gaujoux, R.; Rabani, H.; Starosvetsky, E.; Kveler, K.; Schaffert, S.; et al. A clinically meaningful metric of immune age derived from high-dimensional longitudinal monitoring. Nat. Med. 2019, 25, 487–495. [CrossRef] 7. Jones, E.; Sheng, J.; Carlson, J.; Wang, S. Aging-induced fragility of the immune system. J. Theor. Biol. 2021, 510, 110473. [CrossRef] 8. Glaser, R.; Kiecolt-Glaser, J. Stress-induced immune dysfunction, implications for health. Nat. Rev. Immunol. 2005, 5, 243–251. [CrossRef] 9. Chen, C.; Zhang, X.; Jiang, D.; Yan, D.; Guan, Z.; Zhou, Y.; Liu, X.; Huang, C.; Ding, C.; Lan, L.; et al. Associations between temperature and inﬂuenza activity, A national time series study in China. Int. J. Environ. Res. Public Health 2021, 18, 10846. [CrossRef] 10. Lowen, A.C.; Mubareka, S.; Steel, J.; Palese, P. Inﬂuenza virus transmission is dependent on relative humidity and temperature. PLoS Pathog. 2007, 3, e151. [CrossRef] International EWM Slope after obesity adjustment Infect. Dis. Rep. 2022, 14 306 11. Roussel, M.; Pontier, D.; Cohen, J.-M.; Lina, B.; Fouchet, D. Quantifying the role of weather on seasonal inﬂuenza. BMC Public Health 2016, 16, 441. [CrossRef] [PubMed] 12. Su, W.; Liu, T.; Geng, X.; Yang, G. Seasonal pattern of inﬂuenza and the association with meteorological factors based on wavelet analysis in Jinan City, Eastern China, 2013–2016. Peerj 2020, 8, e8626. [CrossRef] [PubMed] 13. Peci, A.; Winter, A.-L.; Li, Y.; Gnaneshan, S.; Liu, J.; Mubareka, S.; Gubbay, J.B. Effects of absolute humidity, relative humidity, temperature, and wind speed on inﬂuenza activity in Toronto, Ontario, Canada. Appl. Environ. Microbiol. 2019, 85, e02426-18. [CrossRef] [PubMed] 14. Kendon, M.; McCarthy, M. The UK’s wet and stormy winter of 2013/2014. Weather 2015, 70, 40–47. [CrossRef] 15. Public Health England. Surveillance of Inﬂuenza and Other Respiratory Viruses in the United Kingdom, Winter 2013/14. June 2014. Available online: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/ﬁle/ 325203/Flu_annual_report_June_2014.pdf (accessed on 6 December 2021). 16. Jiang, T. Modeling inﬂuenza sequence evolution for vaccination. BMC Proc. 2012, 6, P44. [CrossRef] 17. Lam, E.K.S.; Morris, D.H.; Hurt, A.C.; Barr, I.G.; Russell, C.A. The impact of climate and antigenic evolution on seasonal inﬂuenza virus epidemics in Australia. Nat. Commun. 2020, 11, 2741. [CrossRef] 18. Zhuang, Q.; Wang, S.; Liu, S.; Hou, G.; Li, J.; Jiang, W.; Wang, K.; Peng, C.; Liu, D.; Guo, A.; et al. Diversity and distribution of type A inﬂuenza viruses, an updated panorama analysis based on protein sequences. Virol. J. 2019, 16, 85. [CrossRef] 19. Xue, K.S.; Stevens-Ayers, T.; Campbell, A.P.; Englund, J.A.; Pergam, S.A.; Boeckh, M.; Bloom, J.D. Parallel evolution of inﬂuenza across multiple spatiotemporal scales. eLife 2017, 6, e26875. [CrossRef] 20. Jansen, A.J.G.; Spaan, T.; Low, H.Z.; Di Iorio, D.; Brand, J.V.D.; Tieke, M.; Barendrecht, A.; Rohn, K.; Van Amerongen, G.; Stittelaar, K.; et al. Inﬂuenza-induced thrombocytopenia is dependent on the subtype and sialoglycan receptor and increases with virus pathogenicity. Blood Adv. 2020, 4, 2967–2978. [CrossRef] 21. Hu, J.; Ma, C.; Liu, X. PA-X, a key regulator of inﬂuenza A virus pathogenicity and host immune responses. Med. Microbiol. Immunol. 2018, 207, 255–269. [CrossRef] [PubMed] 22. van Asten, L.; van den Wijngaard, C.; van Pelt, W.; van De Kassteele, J.; Meijer, A.; van Der Hoek, W.; Kretzschmar, M.; Koopmans, M. Mortality attributable to 9 common infections, signiﬁcant effect of inﬂuenza A, respiratory syncytial virus, inﬂuenza B, norovirus, and parainﬂuenza in elderly persons. J. Infect. Dis. 2012, 206, 628–639. [CrossRef] [PubMed] 23. Jung, J.; Seo, E.; Yoo, R.N.; Sung, H.; Lee, J. Clinical signiﬁcance of viral-bacterial codetection among young children with respiratory tract infections, Findings of RSV, inﬂuenza, adenoviral infections. Medicine 2020, 99, e18504. [CrossRef] [PubMed] 24. Hassan, A.; Blanchard, N. Microbial (co)infections, Powerful immune inﬂuencers. PLoS Pathog. 2022, 18, e1010212. [CrossRef] [PubMed] 25. Kumagai, S.; Ishida, T.; Tachibana, H.; Ito, Y.; Ito, A.; Hashimoto, T. Impact of bacterial coinfection on clinical outcomes in pneumococcal pneumonia. Eur. J. Clin. Microbiol. Infect. Dis. 2015, 34, 1839–1847. [CrossRef] [PubMed] 26. Jones, R.P.; Ponomarenko, A. Trends in excess winter mortality (EWM) from 1900/01 to 2019/20—Evidence for a complex system of multiple long-term trends. Int. J. Environ. Res. Public Health 2022, 19, 3407. [CrossRef] [PubMed] 27. Centers for Disease Control and Prevention. CDC Seasonal Flu Vaccine Effectiveness Studies. Available online: https://www.cdc. gov/ﬂu/vaccines-work/effectiveness-studies.htm (accessed on 6 February 2021). 28. Simonsen, L.; Reichert, T.A.; Viboud, C.; Blackwelder, W.C.; Taylor, R.J.; Miller, M.A. Impact of inﬂuenza vaccination on seasonal mortality in the US elderly population. Arch. Intern. Med. 2005, 165, 265–272. 29. Rizzo, C.; Viboud, C.; Montomoli, E.; Simonsen, L.; Miller, M.A. Inﬂuenza–related mortality in the Italian elderly, no decline associated with increasing vaccination coverage. Vaccine 2006, 24, 6468–6475. [CrossRef] 30. Siegenfeld, A.; Bar-Yam, Y. An introduction to complex systems science and its applications. Complexity 2020, 2020, 6105872. [CrossRef] 31. Rutter, H.; Savona, N.; Glonti, K.; Bibby, J.; Cummins, S.; Finegood, D.T.; Greaves, F.; Harper, L.; Hawe, P.; Moore, L.; et al. The need for a complex systems model of evidence for public health. Lancet 2017, 390, 2602–2604. [CrossRef] 32. Serpa, C.; Forouharfar, A. Fractalization of chaos and complexity, Proposition of a new method in the study of complex systems. In Chaos, Complexity and Leadership 2020; Erçetin, S. ¸ S., ¸ Açıkalın, S.N., ¸ Vajzovic, ´ E., Eds.; Springer Proceedings in Complexity; Springer: Cham, Switzerland, 2021. 33. Adak, D.; Bairagi, N. Bifurcation analysis of a multidelayed HIV model in presence of immune response and understanding of in-host viral dynamics. Math. Methods Appl. Sci. 2019, 42, 4256–4272. [CrossRef] 34. McKenzie, D.R.; Hart, R.; Bah, N.; Ushakov, D.S.; Muñoz-Ruiz, M.; Feederle, R.; Hayday, A.C. Normality sensing licenses local T cells for innate-like tissue surveillance. Nat. Immunol. 2022, 23, 411–422. [CrossRef] [PubMed] 35. Rickenbach, C.; Gericke, C. Speciﬁcity of Adaptive Immune Responses in Central Nervous System Health, Aging and Diseases. Front. Neurosci. 2022, 15, 806260. [CrossRef] [PubMed] 36. Wani, S.A.; Sahu, A.R.; Saxena, S.; Hussain, S.; Pandey, A.; Kanchan, S.; Sahoo, A.P.; Mishra, B.; Tiwari, A.K.; Mishra, B.P.; et al. Systems biology approach, Panacea for unravelling host-virus interactions and dynamics of vaccine induced immune response. Gene Rep. 2016, 5, 23–29. [CrossRef] [PubMed] 37. Banks, H.T.; Davidian, M.; Hu, S.; Kepler, G.M.; Rosenberg, E.S. Modelling HIV immune response and validation with clinical data. J. Biol. Dyn. 2008, 2, 357–385. [CrossRef] Infect. Dis. Rep. 2022, 14 307 38. Craddock, T.J.; Fritsch, P.; Rice, M.A., Jr.; del Rosario, R.M.; Miller, D.B.; Fletcher, M.A.; Klimas, N.G.; Broderick, G. A Role for Homeostatic Drive in the Perpetuation of Complex Chronic Illness, Gulf War Illness and Chronic Fatigue Syndrome. PLoS ONE 2014, 9, e84839. [CrossRef] 39. Waters, R.S.; Perry, J.S.A.; Han, S.; Bielekova, B.; Gedeon, T. The effects of interleukin-2 on immune response regulation. Math. Med. Biol. 2018, 35, 79–119. [CrossRef] 40. El Karkri, J.; Boudchich, F.; Volpert, V.; Aboulaich, R. Stability Analysis of a Delayed Immune Response Model to Viral Infection. Differ. Equ. Dyn. Syst. 2022. [CrossRef] 41. Elaiw, A.; AlShamrani, N. Stability of an adaptive immunity pathogen dynamics model with latency and multiple delays. Math. Methods Appl. Sci. 2018, 41, 6645–6672. [CrossRef] 42. Bocharov, G.; Meyerhans, A.; Bessonov, N.; Troﬁmchuk, S.; Volpert, V. Modelling the dynamics of virus infection and immune response in space and time. Int. J. Parallel Emergent Distrib. Syst. 2017, 34, 341–355. [CrossRef] 43. Wang, A.; Xiao, Y.; Smith, R. Multiple Equilibria in a Non-smooth Epidemic Model with Medical-Resource Constraints. Bull. Math. Biol. 2019, 81, 963–994. [CrossRef] 44. Reluga, T.; Medlock, J.; Perelson, A. Backward bifurcations and multiple equilibria in epidemic models with structured immunity. J. Theor. Biol. 2008, 252, 155–165. [CrossRef] [PubMed] 45. Jain, S.; Kumar, S. Dynamic analysis of the role of innate immunity in SEIS epidemic model. Eur. Phys. J. Plus 2021, 136, 439. [CrossRef] [PubMed] 46. Majnaric-T ´ rtica, L.; Vitale, B. Systems biology as a conceptual framework for research in family medicine, use in predicting response to inﬂuenza vaccination. Prim. Health Care Res. Dev. 2011, 12, 310–321. [CrossRef] [PubMed] 47. Jones, R. Excess winter mortality (EWM) as a dynamic forensic tool, Where, when, which conditions, gender, ethnicity, and age. Int. J. Environ. Res. Public Health 2021, 18, 2161. [CrossRef] 48. WHO. Prevalence of Obesity among Adults. Available online: https://www.who.int/data/gho/indicator-metadata-registry/ imr-details/2389 (accessed on 3 March 2022). 49. Global Obesity Observatory. World Obesity Federation Global Obesity Observatory. Available online: https://data.worldobesity. org/ (accessed on 3 March 2022). 50. CDC. Vaccine Coverage among Nursing Home Residents. Available online: https://www.cdc.gov/ﬂu/ﬂuvaxview/interactive- nursing-home.htm (accessed on 3 March 2022). 51. OECD Data. Inﬂuenza Vaccination Rates. Available online: https://data.oecd.org/healthcare/inﬂuenza-vaccination-rates.htm (accessed on 12 March 2022). 52. Wachs, S. What Is a CUSUM Chart and When Should I Use One? Available online: https://accendoreliability.com/cusum-chart- use-one/ (accessed on 12 March 2022). 53. Public Health England. Surveillance of Inﬂuenza and Other Respiratory Viruses in the United Kingdom, Winter 2014 to 2015. May 2015. Available online: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/ ﬁle/429617/Annualreport_March2015_ver4.pdf (accessed on 25 February 2022). 54. Jones, R. Year-to-year variation in deaths in English Output Areas (OA), and the interaction between a presumed infectious agent and inﬂuenza in 2015. SMU Med. J. 2017, 4, 37–69. Available online: http//smu.edu.in/content/dam/manipal/smu/smims/ Volume4No2July2017/SMU%20Med%20J%20(July%202017)%20-%204.pdf (accessed on 25 March 2022). 55. Gonçalves, A.R.; Iten, A.; Suter-Boquete, P.; Schibler, M.; Kaiser, L.; Cordey, S. Hospital surveillance of inﬂuenza strains: A concordant image of viruses identiﬁed by the Swiss Sentinel system? Inﬂuenza Other Resp Viruses 2017, 11, 41–47. [CrossRef] 56. Australian Government, Department of Health. Australian Inﬂuenza Surveillance Report. Available online: https://www1 .health.gov.au/internet/main/publishing.nsf/Content/cda-surveil-ozﬂu-ﬂucurr.htm (accessed on 12 March 2022). 57. Simonsen, L.; Viboud, C. The Art of Modeling the Mortality Impact of Winter-Seasonal Pathogens. J. Infect. Dis. 2012, 206, 625–627. [CrossRef] 58. National Foundation for Infectious Diseases. Respiratory Syncytial Virus in Older Adults, A Hidden Annual Epidemic. Available online: https://www.nﬁd.org/wp-content/uploads/2019/08/rsv-report.pdf (accessed on 12 March 2022). 59. Taubenberger, J.K.; Morens, D.M. The pathology of inﬂuenza virus infections. Annu. Rev. Pathol. 2008, 3, 499–522. [CrossRef] 60. Opatowski, L.; Baguelin, M.; Eggo, R.M. Inﬂuenza interaction with cocirculating pathogens and its impact on surveillance, pathogenesis, and epidemic proﬁle, A key role for mathematical modelling. PLoS Pathog. 2018, 14, e1006770. [CrossRef] 61. Zheng, X.; Song, Z.; Li, Y.; Zhang, J.; Wang, X.-L. Possible interference between seasonal epidemics of inﬂuenza and other respiratory viruses in Hong Kong, 2014–2017. BMC Infect. Dis. 2017, 17, 772. [CrossRef] 62. Hedberg, P.; Johansson, N.; Ternhag, A.; Abdel-Halim, L.; Hedlund, J.; Nauclér, P. Bacterial co-infections in community-acquired pneumonia caused by SARS-CoV-2, inﬂuenza virus and respiratory syncytial virus. BMC Infect. Dis. 2022, 22, 108. [CrossRef] [PubMed] 63. Novella, S. COVID-19 Lockdown and the Flu. Neurologica Blog, 15 June 2020. Available online: https://theness.com/ neurologicablog/index.php/covid-19-lockdown-and-the-ﬂu/ (accessed on 11 March 2022). 64. Doroshenko, A.; Lee, N.; MacDonald, C.; Zelyas, N.; Asadi, L.; Kanji, J.N. Decline of Inﬂuenza and Respiratory Viruses with COVID-19 Public Health Measures, Alberta, Canada. Mayo Clin. Proc. 2021, 96, 3042–3052. [CrossRef] [PubMed] Infect. Dis. Rep. 2022, 14 308 65. PHE. National Inﬂuenza Report. 1 October 2020—Week 40 Report. Available online: https://assets.publishing.service.gov.uk/ government/uploads/system/uploads/attachment_data/ﬁle/923246/National_Inﬂuenza_report_1_October_2020_week_40 .pdf (accessed on 12 March 2022). 66. Institute for Government. Timeline for UK Coronavirus Lockdowns. Available online: https://www.instituteforgovernment.org. uk/charts/uk-government-coronavirus-lockdowns (accessed on 1 January 2022). 67. Koutsakos, M.; Wheatley, A.K.; Laurie, K.; Kent, S.J.; Rockman, S. Inﬂuenza lineage extinction during the COVID-19 pandemic? Nat. Rev. Microbiol. 2021, 19, 741–742. [CrossRef] [PubMed] 68. Danino, D.; Ben-Shimol, S.; van der Beek, B.A.; Givon-Lavi, N.; Avni, Y.S.; Greenberg, D.; Weinberger, D.M.; Dagan, R. Decline in Pneumococcal Disease in Young Children during the COVID-19 Pandemic in Israel Associated with Suppression of seasonal Respiratory Viruses, despite Persistent Pneumococcal Carriage, A Prospective Cohort Study. Clin. Infect. Dis. 2021, ciab1014. 69. Mondal, P.; Sinharoy, A.; Gope, S. The Inﬂuence of COVID-19 on Inﬂuenza and Respiratory Syncytial Virus Activities. Infect. Dis. Rep. 2022, 14, 17. [CrossRef] 70. Dadashi, M.; Khaleghnejad, S.; Elkhichi, P.A.; Goudarzi, M.; Goudarzi, H.; Taghavi, A.; Vaezjalali, M.; Hajikhani, B. COVID-19 and Inﬂuenza Co-infection, A Systematic Review and Meta-Analysis. Front. Med. 2021, 8, 681469. [CrossRef] 71. Jackson, M.L.; Nelson, J.C. The test-negative design for estimating inﬂuenza vaccine effectiveness. Vaccine 2013, 31, 2165–2168. [CrossRef] 72. Fireman, B.; Lee, J.; Lewis, N.; Bembom, O.; van der Laan, M.; Baxter, R. Inﬂuenza vaccination and mortality, differentiating vaccine effects from bias. Am. J. Epidemiol. 2009, 170, 650–656. [CrossRef] 73. Cowling, B.J.; Fang, V.J.; Nishiura, H.; Chan, K.-H.; Ng, S.; Ip, D.K.M.; Chiu, S.S.; Leung, G.M.; Peiris, J.S.M. Increased risk of noninﬂuenza respiratory virus infections associated with receipt of inactivated inﬂuenza vaccine. Clin. Infect. Dis. 2012, 54, 1778–1783. [CrossRef] 74. Rikin, S.; Jia, H.; Vargas, C.Y.; de Belliard, Y.C.; Reed, C.; LaRussa, P.; Larson, E.L.; Saiman, L.; Stockwell, M.S. Assessment of temporally–related acute respiratory illness following inﬂuenza vaccination. Vaccine 2018, 36, 1958–1964. [CrossRef] 75. Hansen, K.P.; Benn, C.S.; Aamand, T.; Buus, M.; da Silva, I.; Aaby, P.; Fisker, A.B.; Thysen, S.M. Does Inﬂuenza Vaccination during Pregnancy Have Effects on Non-Inﬂuenza Infectious Morbidity? A Systematic Review and Meta-Analysis of Randomised Controlled Trials. Vaccines 2021, 9, 1452. [CrossRef] [PubMed] 76. Van Beek, J.; Veenhoven, R.H.; Bruin, J.P.; Boxtel, R.A.J.V.; De Lange, M.M.A.; Meijer, A.; Sanders, E.A.M.; Rots, N.Y.; Luytjes, W. Inﬂuenza-like Illness Incidence Is Not Reduced by Inﬂuenza Vaccination in a Cohort of Older Adults, Despite Effectively Reducing Laboratory-Conﬁrmed Inﬂuenza Virus Infections. J. Infect. Dis. 2017, 216, 415–424. [CrossRef] [PubMed] 77. Lauzon-Joset, J.; Scott, N.; Mincham, K.; Stumbles, P.; Holt, P.; Strickland, D. Pregnancy Induces a Steady-State Shift in Alveolar Macrophage M1/M2 Phenotype That Is Associated with a Heightened Severity of Inﬂuenza Virus Infection, Mechanistic Insight Using Mouse Models. J Infect. Dis. 2019, 219, 1823–1831. [CrossRef] [PubMed] 78. Jenkins, B.N.; Hunter, J.F.; Cross, M.P.; Acevedo, A.M.; Pressman, S.D. When is affect variability bad for health? The association between affect variability and immunee response to the inﬂuenza vaccination. J. Psychosom. Res. 2018, 104, 41–47. [CrossRef] 79. Berencsi, G.; Kapusinszky, B.; Rigó, Z.; Szomor, K. Interference among viruses circulating and administered in Hungary from 1931 to 2008. Acta Microbiol. Immunol. Hung. 2010, 57, 73–86. [CrossRef] 80. NOAA. Coronal Mass Ejections. Available online: https://www.swpc.noaa.gov/phenomena/coronal-mass-ejections (accessed on 7 March 2022). 81. Madanchi, A.; Absalan, M.; Lohmann, G.M.; Anvari, M.; Tabar, M.R.R. Strong short-term non-linearity of solar irradiance ﬂuctuations. Sol. Energy 2017, 144, 1–9. [CrossRef] 82. Výbošt’oková, T.; Švanda, M. Statistical analysis of the correlation between anomalies in the Czech electric power grid and geomagnetic activity. Space Weather 2019, 17, 1208–1218. [CrossRef] 83. Zenchenko, T.A.; Breus, T.K. The Possible Effect of Space Weather Factors on Various Physiological Systems of the Human Organism. Atmosphere 2021, 12, 346. [CrossRef] 84. Palmer, S.J.; Rycroft, M.J.; Cermack, M. Solar and geomagnetic activity, extremely low frequency magnetic and electric ﬁelds and human health at the Earth’s surface. Surv. Geophys. 2006, 27, 557–595. [CrossRef] 85. Papathanasopoulos, P.; Preka-Papadema, P.; Gkotsinas, A.; Dimisianos, N.; Hillaris, A.; Katsavrias, C.; Antonakopoulos, G.; Moussas, X.; Andreadou, E.; Georgiou, V.; et al. The possible effects of the solar and geomagnetic activity on multiple sclerosis. Clin. Neurol. Neurosurg. 2016, 146, 82–89. [CrossRef] 86. Goldwater, P.N.; Oberg, E.O. Infection, Celestial Inﬂuences, and Sudden Infant Death Syndrome: A New Paradigm. Cureus 2021, 13, 17449. [CrossRef] [PubMed] 87. Zaporozhan, V.; Ponomarenko, A. Mechanisms of geomagnetic ﬁeld inﬂuence on gene expression using inﬂuenza as a model system, basics of physical epidemiology. Int. J. Environ. Res. Public Health 2010, 7, 938–965. [CrossRef] [PubMed] 88. Yeung, J.W. A hypothesis: Sunspot cycles may detect pandemic inﬂuenza A in 1700-2000 A.D. Med. Hypotheses 2006, 67, 1016–1022. [CrossRef] [PubMed] 89. Hayes, D.P. Inﬂuenza pandemics, solar activity cycles, and vitamin D. Med. Hypotheses 2010, 74, 831–834. [CrossRef] [PubMed] 90. Nasirpour, M.H.; Shariﬁ, A.; Ahmadi, M.; Jafarzadeh Ghoushchi, S. Revealing the relationship between solar activity and COVID-19 and forecasting of possible future viruses using multi-step autoregression (MSAR). Environ. Sci. Pollut. Res. Int. 2021, 28, 38074–38084. [CrossRef] Infect. Dis. Rep. 2022, 14 309 91. Principia Scientiﬁc International. Weakening Trend in Solar Cycles Since SC21. Available online: https://nextgrandminimum. wordpress.com/2019/06/20/weakening-trend-in-solar-cycle-since-sc21/ (accessed on 11 April 2022). 92. Ramos, P.; Louzada, F.; Ramos, E.; Dey, S. The Fréchet distribution: Estimation and application—An overview. J. Stat. Manag. Syst. 2020, 23, 549–578. [CrossRef] 93. Jones, R.; Sleet, G.; Pearce, O.; Wetherill, M. Complex changes in blood biochemistry revealed by a composite score derived from Principal Component Analysis: Effects of age, patient acuity, end of life, day-of week, and potential insights into the issues surrounding the ‘Weekend’ effect in hospital mortality. J. Adv. Med. Med. Res. 2016, 18, 1–28. [CrossRef] 94. von Wyl, V. Proximity to death and health care expenditure increase revisited: A 15-year panel analysis of elderly persons. Health Econ. Rev. 2019, 9, 9–16. [CrossRef] 95. Vogel, N.; Schilling, O.K.; Wahl, H.-W.; Beekman, A.T.F.; Penninx, B.W.J.H. Time-to-death-related change in positive and negative affect among older adults approaching the end of life. Psychol. Aging 2013, 28, 128–141. [CrossRef] 96. White, N.; Cunningham, W. Is Terminal Drop Pervasive or Speciﬁc? J. Gerontol. 1988, 43, P141–P144. [CrossRef] 97. Bäckman, L.; Macdonald, S.W. Death and Cognition. Eur. Psychol. 2006, 11, 224–235. [CrossRef] 98. Stanaway, F.F.; Gnjidic, D.; Blyth, F.M.; Le Couteur, D.; Naganathan, V.; Waite, L.; Seibel, M.; Handelsman, D.J.; Sambrook, P.N.; Cumming, R. How fast does the Grim Reaper walk? Receiver operating characteristics curve analysis in healthy men aged 70 and over. BMJ 2011, 343, d7679. [CrossRef] [PubMed] 99. Flannery, B.; Smith, C.; Garten, R.J.; Levine, M.Z.; Chung, J.R.; Jackson, M.L.; Jackson, L.A.; Monto, A.S.; Martin, E.T.; Belongia, E.A.; et al. Inﬂuence of Birth Cohort on Effectiveness of 2015–2016 Inﬂuenza Vaccine Against Medically Attended Illness Due to 2009 Pandemic Inﬂuenza A(H1N1) Virus in the United States. J. Infect. Dis. 2018, 218, 189–196. [CrossRef] [PubMed] 100. Kissling, E.; Pozo, F.; Buda, S.; Vilcu, A.M.; Gherasim, A.; Brytting, M.; Domegan, L.; Gómezet, V.; Meijer, A.; Lazar, M.; et al. Low 2018/19 vaccine effectiveness against inﬂuenza A(H3N2) among 15-64-year-olds in Europe: Exploration by birth cohort. Eurosurveillance 2019, 24, 1900604. [CrossRef] 101. Sanniklova, T.; Romanyukha, A.; Barbi, E.; Caselli, G.; Franceschi, C.; Yashin, A. Modeling of Immunosenescence and Risk of Death from Respiratory Infections: Evaluation of the Role of Antigenic Load and Population Heterogeneity. Math. Model. Nat. Phenom. 2017, 12, 48–62. [CrossRef] 102. Kaczorowski, K.J.; Shekhar, K.; Nkulikiyimfura, D.; Dekker, C.L.; Maecker, H.; Davis, M.M.; Chakraborty, A.K.; Brodin, P. Continuous immunotypes describe human immune variation and predict diverse responses. Proc. Natl. Acad. Sci. USA 2017, 114, E6097–E6106. [CrossRef] 103. Jones, R. A Study of an unexplained and large increase in respiratory deaths in England and Wales: Is the pattern of diagnoses consistent with the potential involvement of Cytomegalovirus? J. Adv. Med. Med. Res. 2014, 4, 5179–5192. Available online: https://journaljammr.com/index.php/JAMMR/article/view/15226 (accessed on 1 January 2022). [CrossRef] 104. Jones, R. Multidisciplinary insights into health care ﬁnancial risk and hospital surge capacity, Part 3: Outbreaks of a new type or kind of disease create unique risk patterns and confounds traditional trend analysis. J. Health Care Financ. 2021, 47, 1–57. Available online: http://healthﬁnancejournal.com/index.php/johcf/article/view/242 (accessed on 1 January 2022). 105. Franklin, B.; Hochlaf, D. An Economic Analysis of Flu Vaccination. The International Longevity Centre—UK. Available online: An-economic-analysis-of-ﬂu-vaccination-18(1).pdf (accessed on 20 March 2022). 106. Anderson, M.L.; Dobkin, C.; Gorry, D. The Effect of Inﬂuenza Vaccination for the Elderly on Hospitalization and Mortality: An Observational Study with a Regression Discontinuity Design. Ann. Intern. Med. 2020, 172, 445–452. [CrossRef] 107. Van Ourti, T.; Bouckaert, N. The Dutch inﬂuenza vaccination policy and medication use, outpatient visits, hospitalization and mortality at age 65. Eur. J. Public Health 2020, 30, 275–280. [CrossRef] 108. Sirois, C.; Boiteau, V.; Chiu, Y.; Gilca, R.; Simard, M. Exploring the associations between polypharmacy and COVID-19-related hospitalisations and deaths: A population-based cohort study among older adults in Quebec, Canada. BMJ Open 2022, 12, e060295. [CrossRef] [PubMed]

http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png

Infectious Disease Reports Multidisciplinary Digital Publishing Institute

http://www.deepdyve.com/lp/multidisciplinary-digital-publishing-institute/system-complexity-in-influenza-infection-and-vaccination-effects-upon-qKZBUE5pLZ

System Complexity in Influenza Infection and Vaccination: Effects upon Excess Winter Mortality

Loading next page...

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher: Multidisciplinary Digital Publishing Institute
Copyright: © 1996-2022 MDPI (Basel, Switzerland) unless otherwise stated Disclaimer The statements, opinions and data contained in the journals are solely those of the individual authors and contributors and not of the publisher and the editor(s). MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Terms and Conditions Privacy Policy
ISSN: 2036-7449
DOI: 10.3390/idr14030035
Publisher site: See Article on Publisher Site

Abstract

Journal

Infectious Disease Reports – Multidisciplinary Digital Publishing Institute

Published: Apr 21, 2022

Keywords: influenza; vaccination; pathogen interference; immune diversity; antigenic distance; winter mortality

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

System Complexity in Influenza Infection and Vaccination: Effects upon Excess Winter Mortality

System Complexity in Influenza Infection and Vaccination: Effects upon Excess Winter Mortality

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

System Complexity in Influenza Infection and Vaccination: Effects upon Excess Winter Mortality

System Complexity in Influenza Infection and Vaccination: Effects upon Excess Winter Mortality

References

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies