key: cord-0776822-qwub35cd
authors: Hoogeveen, Martijn J.
title: Pollen likely seasonal factor in inhibiting flu-like epidemics. A Dutch study into the inverse relation between pollen counts, hay fever and flu-like incidence 2016–2019
date: 2020-04-17
journal: Sci Total Environ
DOI: 10.1016/j.scitotenv.2020.138543
sha: 4513491494e0513068bbbcce601f92efec1d8dbb
doc_id: 776822
cord_uid: qwub35cd

Abstract There is uncertainty if current models for the Covid-19 pandemic should already take into account seasonality. That is because current environmental factors do not provide a powerful explanation of such seasonality, especially given climate differences between countries with moderate climates. It is hypothesized that one major factor is overlooked: pollen count. Pollen are documented to invoke strong immune responses and might create an environmental factor that makes it more difficult for flu-like viruses to survive outside a host. This Dutch study confirms that there is a (highly) significant inverse correlation between pollen count and weekly changes in medical flu consults, and that there is a highly significant inverse correlation between hay fever incidence, as measured by prescribed medication revenues, and weekly flu consults. This supports the idea that pollen are a direct or indirect factor in the seasonality of flu-like epidemics. If seasonality will be observed during the covid-19 spread as well, it is not unlikely that pollen play a role.

Virologists observe already for many years that the cold, and flu-like epidemics (influenza and corona caused) "go away in May" in the Northern hemisphere, to move to the Southern hemisphere, and return next Autumn and Winter in a slightly mutated form. Further, it is observed that these strains of viruses have a lower reproduction factor in tropical countries, outside the rain season. It is observed that possible inhibiting environmental factors in these epidemics are UV light (Schuit et al., 2020) , temperature and humidity (Chong et al., 2020) and behavioral changes during the seasons (Gozzi et al., 2020) . None of these environmental factors seems to be completely satisfactory to Journal Pre-proof J o u r n a l P r e -p r o o f 2 explain the seasonality of flu-like epidemics, which is very important to understand in the light of ongoing modelling efforts for the current Covid-19 pandemic.

The relation between the seasons and the flu-like epidemics can be an indirect relation, whereby there is actually an overlooked intermediate factor at play: the natural pollen "firewall" that typically peaks in April, is still active during the remainder of Spring and Summer, to start to lose its footing in September. Pollen are observed to be at play in plant biology: plant viruses can be transported through pollen (Bhatt & Gao, 2020) . Further, pollen play a well-researched role in triggering human immune responses.

To hypothesize in what ways pollen could be an environmental factor influencing the life cycle of flulike epidemics, the objective of this quick study is to falsify the hypothesis that there are no significant inverse correlations between pollen counts, pollen induced immune responses, and flulike epidemic life cycles in The Netherlands. An interdisciplinary, environmental view might lead to new insights (Reijnders & Hoogeveen, 2001) .

To establish whether the hypothesized inverse correlations between pollen counts per m3, pollen induced immune responses, and previous flu-like incidence are significant, we used the public data J o u r n a l P r e -p r o o f 3 We formulated three statistical null hypotheses H1(0): there is no inverse correlation between pollen count and flu-like reports H2(0): there is no inverse correlation between pollen count and weekly changes in flu incidents H3(0): there is no inverse correlation between hay fever symptoms and flu-like reports It is outside of the scope of this quick research to verify the underlying data sets of CBS and RIVM by examining the raw data, and validity and reliability of data collection methods.

When inspecting the data sets regarding pollen counts and flu-like consults at primary medical care in The Netherlands, it is clear that there are a few pollen bursts, especially during Spring, and then at a lower level during Summer. These bursts typically coincide with a decline of flu-like consults.

Given the burst like phenomenon, it is not strange that we found no clear correlation between pollen count and the incidence of flu-like medical consults for our data set (correlation coefficient: 0.07).

Therefore, we cannot falsify the hypothesis that there is no relation between pollen count and the overall count of flu-like symptoms. change of medical flu consults. The P-Value is .02, thus the result is significant at p < .05. Therefore, we can falsify the hypothesis that there is no inverse correlation between the weekly pollen count and changes in flu-like consults. This inverse correlation might provide further support for the idea that the presence of (a lot of) pollen has an inhibiting effect on flu incidence, and starts to immediately influence the direction and course of the epidemic life cycle.

If we correct the correlation for incubation period by comparing pollen count with the next week's change in medical consults the correlation coefficient becomes stronger and highly significant at -0.24 (p<0.01). P-Value is .001661 at n=165. We can still reject the null-hypothesis that there is no inverse relation between pollen count and flu-like consults (corrected for incubation period). As this correlation is stronger than if not corrected for incubation period, it is a further indication of sequentiality as a causality criterion. 

The (highly) significant results show that there is an inverse relation between pollen counts and the life cycle of flu-like epidemics, and a somewhat stronger inverse relation between pollen induced immune responses (as measured by revenues for anti-hay fever medication) and flu-like medical consults. This gives support to the idea that an activated immune system makes it more difficult for flu-like viruses to penetrate a new host. At the same time, there might also be an environmental factor at play: more pollen in the air might make it more difficult for a flu-like virus to survive outside a host. Like a pollen firewall. We can also observe that the sequentiality criterion for causality is met.

It will require further research to understand whether the same inverse correlations are found in other countries with moderate climates, and whether pollen play a similar role in tropical countries as well. It is good to compare a variety of climatological diverse countries to check for meteorological variations regarding UV light, humidity and temperature. It is very interesting to understand whether also in other countries and years a PITV of around 66 pollen/m3 (+/-8) is found for moderate flu-like epidemics, and if this threshold value can be further understood in the light of the more aggressive Covid-19 pandemic. Further, it is good to understand better the interaction between aerosol pollen and flu-like viruses, and whether general immune responses are indeed a causal factor in reducing the spread of flu-like viruses. Finally, it still needs to be observed whether the covid-19 (SARS-CoV-2 virus) spread will be seasonal -"going away in May" -whereby the pollen factor could be studied as a directly or indirectly inhibiting factor, potentially.

J o u r n a l P r e -p r o o f

Weekly Change in Flu-like Consults (per 100K Dutch Citizens) vs

CBS.nl Tabel Omzet hooikoortsmedicatie en pollen in de lucht

Latitudes mediate the association between influenza activity and meteorological factors: A nationwide modelling analysis in 45 Japanese prefectures from

Towards a data-driven characterization of behavioral changes induced by the seasonal flu

Energy effects associated with e-commerce: A case-study concerning online sales of personal computers in The Netherlands

The Influence of Simulated Sunlight on the Inactivation of Influenza Virus in Aerosols