Italians, air pollution and the pathogen SARS-CoV-2

What, you think, another blog post about the SARS-CoV-2 virus, the famous coronavirus. But that's all we're reading these days.

Yeah, one more. But let's be honest, it was hard to cut through. Our current research project is called "People, Pollution and Pathogens". So when I heard about this Italian study, it was impossible to resist. All the ingredients are there.

People. People first. It's a study of the Italian population between February 10 and March 3, 2020.

Pollution. The pollution that has been examined is air pollution, air pollution, and more specifically the events for which the concentration of PM10 exceeded the legal limit.

PM10? But what is it? PM stands for particulate matter. It is the fine particles, the particles suspended in the air (or in water). There are several kinds of PM, named according to their size. Simply put, PM10 is less than 10 µm in size. To give you an order of magnitude, a human hair has a diameter of about 60 µm. PM10 is therefore 6 times smaller than a hair, but it is the largest PM. PM comes from a variety of sources, from natural phenomena such as erosion or volcanic emissions, and from human activities such as chimney fires, car engines or industrial processes.

According to the WHO, air pollution was responsible for 3.7 million deaths worldwide in 2012, accounting for as much as 6.7% of all deaths. PM can have an irritating mechanical effect on the lungs. They may also contain toxic substances. The fine particles enter deep into the lungs and can be absorbed by macrophages (immune cells) and endothelial cells (which line the inside of blood vessels and the heart). They can even pass through the alveolar wall and be carried by the blood to all organs. They thus cause respiratory and pulmonary diseases (asthma, allergies, inflammations) but also cardiovascular diseases.

Pathogen. The pathogen of interest is SARS-Cov-2, responsible for covid-19 disease, you’ve likely heard about it.

What does the Italian study say?

The Italian study applies the principles of epidemiology (explained to you there, post on John Snow). It relates maps of the SARS-CoV-2 epidemic to maps of PM10 pollution in Italy. It is based on previous studies suggesting that viruses, such as avian influenza, could be transported in the air, on particles, over long distances. Indeed, virus DNA or RNA has been found in wind-borne dust. But beware, when viral genetic material is found, it does not necessarily mean that the virus is still "alive" and, above all, infectious.

By comparing the covid-19 pollution and disease maps, the Italian study found a correlation between the number of people reported to be infected with the SARS-CoV-2 virus at a given time and exceeding the PM10 limit 14 days earlier (as the infection is on average diagnosed 14 days after infection). In the Po Valley and in the city of Rome, where air pollution is higher, an abnormally high number of sick people was detected. The authors of the study say that their results suggest a spread of the virus by particles in the air, in addition to a person-to-person spread.

Why is a correlation no evidence of causation?

In science, a correlation is simply a correlation. It does not necessarily imply causation. In fact, the Italian study does not allow us to conclude with certainty that the virus was spread through the air on PMs. Moreover, the authors are honest and cautious. In order to prove the mechanism, the airborne transport on PMs, further studies would have to be conducted. Experimental studies.

If we take John Snow's study during the cholera epidemic in London. John Snow first conducted a map study. He superimposed the map of cholera cases on the map of the city's drinking water system and found that there were more cases around a certain water pump. It's a correlative study. For cholera, the proof that there was causation, that cholera was carried by drinking water, was definitively proven by making the contaminated pump unusable. The arm of the pump was removed and cholera cases decreased and then disappeared. This was the experimental study.

Let's go back to our Italian study. A correlation is simply a correlation. And a correlation could be just due to chance. By chance? Yes. In fact, there's a whole website that shows some very nice correlations that are totally due to chance (https://www.tylervigen.com/spurious-correlations). For example, there is a correlation between the number of people drowned in swimming pools and the number of films in which Nicolas Cage appears at that time. If one wanted to find causation, one could suggest that Nicolas Cage's films lead people to drown in their swimming pools (but we don’t). In short, just because there's a correlation doesn't mean there's really a link.

The third factor – the stork effect

Sometimes, when we make a correlation, we do not necessarily know it, but it can happen that the two elements measured, A and B, are not completely independent of each other. And that's why they are correlated. The correlation may be due to a third factor (factor C), which is not measured, and which induces the relationship. It's called the stork effect. Let's imagine that factor C causes A and factor C also causes B, A and C are correlated, B and C are correlated, so A and B are correlated.

What I'm telling you is not very clear? Let's take an example. To check whether it is storks that bring babies, let us correlate the number of storks' nests on the roofs of houses with the number of children living there, in Alsace. We find a correlation! The more stork nests there are, the more children there are! It's unbelievable. One might be tempted to conclude that storks bring babies. But we'd be poking our noses in, of course. Actually, the reason is that the houses with the most children are also bigger. These houses have more chimneys and therefore can accommodate more stork nests. The size of the house is the third factor.

In the case of the Italian study, it is not the size of the house that was forgotten, but the density of the population. It is reasonable to assume that a more densely populated area will have more cases of covid-19 because the transmission of the disease will be accelerated by human contact. Such an area is also likely to be more polluted.

Another factor that was not considered in the Italian study is the pre-existing link between fine particle pollution and the health status of people. It has been known for decades that PM affects human health through various mechanisms, including damage to the lungs. This is why smokers have worse lungs on average than non-smokers. It has also been known for even longer that people in poorer health have poorer immune defenses and are therefore more susceptible to viral infections. In one study, people were injected with an inactivated influenza virus (vaccination). Before that, half of the people were exposed to pollution from a diesel engine for two hours, the other half breathed unpolluted air. People who were exposed to polluted air had a diminished immune response to the inactivated virus compared to the other group. This proves that pollution, in addition to mechanically damaging our lungs, decreases our immune protection. It is quite likely that people whose health and lungs are damaged by pollution are more susceptible to CoV-2-SARS. There is then a link between PM pollution and the number of cases of covid-19, without PM carrying the virus.

Obviously, the correlation found by our Italian colleagues could just as well be explained by the fact that the SARS-CoV-2 virus is carried by PM10, as the authors of the study seem to think. This is one explanation among others. At this stage, with a correlation, we cannot really conclude in favour of one explanation or another.

The Italian study is here:

http://www.simaonlus.it/wpsima/wp-content/uploads/2020/03/COVID19_Position-Paper_Relazione-circa-l%E2%80%99effetto-dell%E2%80%99inquinamento-da-particolato-atmosferico-e-la-diffusione-di-virus-nella-popolazione.pdf

For those who (like me) don't understand a word of Italian, I translated the study with the help of the translator available on the internet DeepL.com/Translator