Why does the SARS-CoV-2 coronavirus impact more on some areas than others? Air pollution could be a common denominator for countries with the greatest impact – in terms of numbers and severity – of COVID-19. But it could also explain why certain areas are affected more than others within the same country. It should be noted that the Po Valley (Italy) and the Wuhan area (China) are among the most polluted areas on the planet with regard to atmospheric particulate matter. On the other hand, in the scientific literature, for years, a clear correlation has emerged between the particulate matter (PM10 and PM2.5) and the spread of some airborne viral agents, which is also confirmed by the Italian data for COVID-19. All this has important consequences for addressing the mitigation measures of the COVID-19 epidemic.
The so-called “aerobiology” plays a fundamental role in the transmission of infectious diseases. This is an area of biology that, to prevent or mitigate the transmission of infectious diseases through the air, studies the effect on this transmission of many variables, such as: particle size, type of particles, duration of stay in the air of the particles, distance traveled by the particles, meteorological and environmental factors, etc. In the case of viral respiratory diseases, however, in recent years it has been discovered that air pollution is among the most significant environmental factors.
There is a solid scientific literature that correlates the incidence of viral infection cases with concentrations of atmospheric particulate matter (e.g. PM10 and PM2.5). As observed by Setti et al. (2020), atmospheric particulate matter acts as a “carrier”, or as a transport vector, for many chemical and biological contaminants, including viruses. Viruses “stick” (with a coagulation process) to atmospheric particulate matter, consisting of solid and/or liquid particles capable of remaining in the atmosphere for hours, days or weeks, and which can spread and be transported even for long distances.
The collapse in the level of air pollution (below), compared to the previous year (above), in Wuhan, the epicenter of the great Chinese outbreak of COVID-19, following the quarantine to which almost all the population was forced.
This can, in principle, explain very well why in Italy the greatest outbreaks of COVID-19 developed in the Po Valley (or “Pianura Padana”), one of the most polluting areas in the world, while in the rest of Italy there were cases of minor infections, with outbreaks almost absent or relatively small. But above all it can explain why, while in other areas of Italy there have been observed expansion curves of the infection compatible with the epidemic models of person-person transmission, in the Po Valley there has been an accelerated spread, such as that mediated by a vehicle: the particulate matter.
Note that the World Health Organization (WHO) uses a particle diameter of 5 μm to distinguish between airborne transmission (≤5 μm) and droplet transmission (> 5 μm). Some studies suggest that particles above 6 μm tend to settle mainly in the upper airways, while particles below 2 μm settle mainly in the alveolar region. Therefore, particles less than 10 μm can penetrate deeper into the respiratory tract and are more likely to carry a virus in the lower lung region.
The areas most affected by COVID-19 and the most polluted areas
As it is known, the initial outbreak of the COVID-10 epidemic in China, the disease caused by severe “new” coronavirus (SARS-CoV-2) acute respiratory syndrome, was first identified in the city of Wuhan, the capital of the province of Hubei, in December 2019. These early cases mostly had connections with the Huanan seafood wholesale market, which also sold live animals. Wuhan is a mega-city of 11 million inhabitants which houses vast factories that manufacture cars for various global brands (General Motors, Nissan, Honda), and is home to a steel industry.
Workers install photovoltaic panels on a residential tower in Wuhan, China, in 2018. Note the high level of pollution in the city.
So it is not surprising that, although the area of Beijing and its surroundings has already resulted, from the satellite maps provided in 2004 by the Envisat satellite, the most polluted on the planet together with the Po Valley and the area around New York (see here), also the Wuhan area shows very high pollution levels, comparable to the Beijing area, as shown in the figure below, relating to tropospheric nitrogen dioxide (NO2) levels. In fact, I remember that this gas is traditionally used as a pollution tracer in satellite surveys, and shows a good correlation with particulate pollution.
Sentinel 5P satellite images of China, taken from early January to late February 2020, when the coronavirus was at its peak. There was a sharp and dramatic drop in nitrogen dioxide levels during the national quarantine. (source: ESA)
In Italy, the initial outbreak of the COVID-19 epidemic occurred in the Codogno area, a town of 15,000 inhabitants located in the province of Lodi, in the Po Valley. It then expanded to other regions and provinces (Bergamo, Brescia, Cremona, Milan, etc.), but developed with particular impetus in the more industrialized provinces of the Po Valley, as clearly illustrated by the figure below. And, as shown in 2017 by the Sentinel-5P satellite, the Po Valley is by far the most polluted area in Europe, as well as among the most polluted in the world, confirming what was observed 13 years earlier by Envisat.
The trend of the COVID-19 epidemic in the Italian provinces with the most contagions. Evolution has been particularly accelerated in those of the Po Valley. (source: Ministry of Health)
The relationship between the transmission of the virus and air quality
In literature, the possible relationship between environmental factors and the transmission of respiratory viruses (and therefore the development of respiratory diseases) has already been studied by some groups of researchers around the world. For example, it has been found that inadequate airflow and ventilation can play a role in the infectivity and transmission of a respiratory virus. As regards the effect of air quality, we know from the literature that atmospheric particulate matter (PM10, PM2.5) constitutes an effective vector for the transport, spread and proliferation of viral infections.
For example, Chen et al. (2010) demonstrated that avian influenza can be spread over long distances through dust storms carrying the virus, and found that there is an exponential correlation between the quantities of cases of infection (Overall Cumulative Relative Risk RR) and the concentrations of PM10 and PM2.5. More recently, Zhao et al. (2019) suggested that most of the 2015 avian influenza positive cases in the United States may have received the airborne virus, transported by fine particulates, from infected farms even very distant.
The correlation between PM2.5 (left) and PM10 (right) concentration and the number of cases of bird flu infection in China. (source: Setti et al., 2020)
Very recently Peng et al. (2020), collecting daily data on measles cases, air pollutants and meteorological data from 2005 to 2009 in the city of Lanzhou (China) found that air pollutants and meteorological factors had statistically significant associations with daily measles cases: the PM10 and the maximum temperature with a delay of 5 days, nitrogen dioxide (NO2) with a delay of 6 days, the average air pressure with a delay of 4 days, visibility and wind speed with a delay of 3 days . The effect of PM10 was greater in the summer.
Ye et al. (2016) investigated the association between respiratory syncytial virus (RSV) infection in children and air pollutants and ambient temperature in Hangzhou, China. They found a positive correlation between the infection rate and particulate matter PM2.5 (r = 0.446, P <0.001), PM10 (r = 0.397, P <0.001), SO2 (r = 0.389, P <0.001), NO2 (r = 0.365, P <0.001) and CO (r = 0.532, P <0.001). Instead, a negative correlation was found between room temperature and RSV infection, and was stronger with the minimum room temperature (r = -0.8080, P <0.001).
The relationship between the rate of spread of the human respiratory syncytial virus (right) and the levels of some air pollutants, including PM10 (in red) and PM2.5 (in blue). (source: Ye et al., 2016)
The link between pollution and the number of COVID-19 cases
In a “position paper” (link here) released in March 2020 by the Italian Society of Environmental Medicine (SIMA), a group of various university researchers led by the chemist Leonardo Setti, of the University of Bologna, assessed the possible correlation between the spread of COVID-19 in Italy and the levels of atmospheric particulate matter (PM10) provided by the ARPA control units. In particular, the correlation between the number of cases infected with COVID-19 and the number of exceedances of the Italian legal limit (50 μg/mc) for the daily concentration of PM10 was analyzed, compared to the number of active control units per Province.
The analysis carried out by Setti et al. highlighted the existence of a strong correlation (R2 = 0.9794) – and therefore of a probable direct relationship – between these two variables (see the graph below), and specifically between the exceedances of the concentration limits of PM10 recorded in the period 10 February-29 February and the number of cases infected with COVID-19 updated to 3 March (considering an intermediate time delay relating to the period 10-29 February of 14 days, approximately equal to the incubation time of the virus up to identification of the contracted infection).
Analysis of the correlation between the variables illustrated in the text, carried out by grouping the Provinces into 5 classes based on the number of infected cases (on a logarithmic scale: infected logs), in relation to exceedances of the PM10 concentration limit for each of the 5 classes of Provinces (average by class: average no. of exceedances of PM10 / no. of Prov. pollution measurement units). (source: Setti at al., 2020)
In a completely independent and brilliant way, the Italian theoretical physicist Gianluca Malato has quantitatively analyzed the correlation between the diffusion in Italy of COVID-19 and air pollution, reaching similar conclusions: the existence of a relationship between the two variables. For this analysis (that you find here), he concentrated on the Italian Provinces: for each municipality in Italy, he obtained historical data on air pollution as the average concentration of PM10 in a given period, which can be obtained directly from here. The concentration in question was then averaged over all cities in a Province.
The concentration of starting PM10 per Municipality (left) and the average per Province obtained (right). This is true for all Municipalities and Provinces of Italy.
Every city in Italy belongs to a Province, so Malato downloaded the city-province mapping table from ISTAT, which is the Italian national statistical institute, from here. After that, he calculated a measure of the virus infection rate that describes the rate of infection in that province. Air pollution and infection rate were then plotted together and analyzed with Spearman’s correlation coefficient and Fisher’s exact test to see if there is a correlation between them. Everything was done in Python. The file with all the calculations is available here on GitHub.
The data on the infection used by the patient are those provided by the Civil Protection, and you can find them here. For each province, the infection rate was estimated as the average difference in the cumulative number of infected people in a city between two consecutive days. At this point, Malato reported the PM10 concentration and infection rate for COVID-19 on a graph (see below). As you can see, “low values for PM10 concentration are always related to low infection rates. As long as the concentration of PM10 increases, the tendency of the infection rate increases and reaches higher values ”.
The concentration of PM10 and the infection rate for COVID-19. Low values for PM10 concentration are always related to low infection rates.
Malato then calculated Spearman’s correlation coefficient in order to evaluate whether what we are seeing is random or not: “The value of the coefficient is 0.42 and the p-value of a statistical test whose null hypothesis states that the absence of a correlation is 0.00002. Since the p-value is less than 5%, we can reject the null hypothesis that there is no correlation. This is sufficient to say that there is statistical evidence of a positive correlation between air pollution and the COVID-19 infection rate. If the level of air pollution is high, the infection spreads more quickly”.
The figure below, created by me, instead shows the number of monthly exceedances of the Italian legal limit for the daily concentration of PM10 for various Italian cities and, by comparison, with the city of Wuhan, the outbreak of the Chinese epidemic. It can be seen that in the cities of the Po Valley most affected by COVID-19 the number of days of exceeding was comparable to that which occurred in Wuhan, while in the large metropolises of Rome and Naples particulate pollution was, in comparison, very reduced. This would explain why outbreaks did not occur in Rome, despite the fact that the virus had been present there since January.
The number of days of exceeding the Italian legal limit for PM10 in the period from 1 January to 18 March 2020, for some cities in Northern Italy particularly affected by COVID-19 and, by comparison, with the city of Wuhan and with some big cities in the rest of Italy. (data source: ARPA, AQICN)
Even more interesting is the fact that today, 23 March 2020, the third country in the world most affected by the epidemic is the United States (see the figure below), where 45.4% of those infected (15168 out of 33404 in total) located in the New York area. And here “the circle closes”. In fact, only a few “insiders” – including myself, who wrote in 2018 the aforementioned article that you find here – know that since 2004, thanks to satellite images, it has resulted that the three most polluted areas of the planet are a large area of China, the Po Valley and the New York area, coincidentally those most affected by COVID-19. Coincidence? Hard to believe it.
The three countries (and within them the three areas) most affected by COVID-19 are, essentially, the three most polluted areas of our planet, as revealed in 2004 by the Envisat satellite.
The link between pollution and severity of COVID-19
There is still much we do not know about COVID-19. However, a clear link has been found many times in the literature between exposure to long-term air pollution and other lung-related diseases, such as chronic obstructive pulmonary disease, or asthma, which is an increasing disease in cities where extremely high pollution levels are observed. The European Public Health Alliance has warned that “diseases caused by polluted air could reduce the chances of survival of COVID-19 patients”.
As explained by Sara De Matteis, lung disease expert at the University of Cagliari, Italy, and member of the environmental health committee of the European Respiratory Society, “chronic lung and heart disease caused or worsened by long-term exposure to air pollution are less able to fight lung infections. It is therefore likely that this will also happen with COVID-19. By lowering air pollution levels, we can help the most vulnerable people in their fight against this and all possible future pandemics”.
Scientific evidence suggests that poor air quality may have increased the death toll from a previous coronavirus outbreak: the 2003 SARS pandemic. A study of patients with SARS (Cue et al., 2003), in fact, found that people living in regions with moderate amounts of air pollution were 84% more likely to die than those in regions with cleaner air. People who have developed chronic respiratory problems, cancer and various pathologies for polluted air are therefore more vulnerable to a more severe impact of COVID-19.
The scientific article by Cue et al. (2003) cited in the text.
Here is a summary of the Abstract and conclusions from the article just quoted: “Severe acute respiratory syndrome (SARS) has caused 349 deaths with 5,327 probable cases reported in mainland China since November 2002. Case mortality from SARS has varied across geographic areas , which could be partially explained by the level of air pollution. [..] Our studies have shown a positive association between air pollution and SARS accident mortality in the Chinese population using publicly available data on SARS statistics and air pollution indices”.
The analysis made by the authors of the aforementioned research (Cue et al., 2003), conducted between 5 regions with 100 or more cases of SARS, showed that the mortality rate of the case increased with the increase in the pollution index of the air (API) derived from concentrations of particulate matter, sulfur dioxide, nitrogen dioxide, carbon monoxide and ozone at ground level: case mortality = – 0.063 + 0.001 * API. SARS patients from high API regions were twice as likely to die from SARS than those from low API regions (RR = 2.18, 95% CI: 1.31-3.65).
The correlation and association between short-term exposure to environmental air pollution and mortality from SARS cases in the People’s Republic of China. (source Cue et al., 2003)
So, the air pollution and the coronavirus that causes COVID-19 have a close relationship. Polluted air breathing is linked to hypertension, diabetes and respiratory diseases, conditions that doctors are beginning to associate with higher mortality rates for COVID-19. Doctors say people with these chronic conditions may be less able to fight infections and are more likely to die of the disease. This effect therefore adds to that of physical “carrier” of the infection, probably more important, exercised by the pollution.
Because it attacks the lungs, the coronavirus that causes COVID-19 could be a particularly serious threat to those who smoke tobacco or marijuana or use the electronic cigarette. In China, 52.9 percent of men smoke, in contrast to only 2.4 percent of women; further analysis of emerging data on COVID-19 from China could help determine whether this disparity contributed to the higher mortality observed in men than women, as reported by the China CDC. Even in Italy smokers are more numerous among men than among women.
The impact on the future course of the epidemic
The trend of the concentration of particulates in the Po Valley shows a strongly seasonal trend (see for example the figure below relating to PM10). The Po Valley, in fact, is a particular area of Italy characterized by the famous “fog in Valpadana” – frequent and persistent especially in the winter period – due to the phenomenon of the so-called “thermal inversion” on the ground, which helps to compress the lower layers of the atmosphere by creating a “cover” that does not let escape the pollutants produced by the various sources in the area.
Trend of the monthly average concentrations of PM10 in Milan (Italy) in 2017 (above) and in Bergamo (Italy) from 2001 to 2012 (below).
The persistence of high pressure and the absence of ventilation often guarantee, during the winter, the presence of a dense blanket of fog – and a consequent heavy pollution, not surprisingly often associated with the blocking of traffic in large cities – for several days. Therefore, the various harmful substances (dioxins, fine dust, fine and ultrafine particulates, toxic gases, etc.) produced by the various sources are retained on the ground, and tend to stagnate and accumulate progressively reaching completely anomalous concentrations, unlike what it happens in other areas of Italy.
Therefore, also thanks to the end of the winter heating of the buildings (which is one of the major sources of atmospheric pollutants), it is reasonable to expect that from April to September the concentration of particulates in the Po Valley will reach the minimum annual levels, favoring the mitigation of the epidemic, especially if accompanied by the additional measures that we will illustrate in the next section. Therefore, even the models used to predict the future trend of the epidemic should take into account the pollution factor, at least as regards the Po Valley.
The increase in temperatures should also help in the mitigation of the epidemic, as suggested by the aforementioned works by Ye et al. (2016) and Peng et al. (2020), an effect that would also benefit all other areas of Italy. This also happened with SARS, which appeared in November 2002 and which disappeared almost completely already in early summer 2003 (see figure). In the case of the SARS-CoV-2 that causes COVID-19, however, there is a growing consensus among epidemiologists that it can recur in the autumn, as the virus is now widespread also in the southern hemisphere.
The trend of the SARS epidemic, which appeared in November 2002 and ended with the virus “completely disappeared” from circulation in the summer of 2003.
Confirming the plausibility of the “help” against coronavirus provided by the increase in temperatures, comes the study done by scientists from the University of Maryland who belong to the Global Virus network, who have discovered that COVID-19 is essentially developing in the so-called “coronavirus belt”, the green band indicated in the figure below, an area between 30 and 50 degrees of latitude and characterized, in the weeks in which the virus spread, by average temperatures between 5 and 11 °C and a humidity between 47 and 79%. In this area, the disease has exploded more severely.
The “coronavirus belt” identified by researchers from the University of Maryland.
Additional mitigation measures needed
In light of what has been illustrated so far, it can be seen that the extraordinary measures taken by the Italian Government are not sufficient, on their own, for a rapid and efficient mitigation of the epidemic in the Po Valley area: not surprisingly, in China a relatively rapid arrest of the epidemic was obtained through the coercive imposition, in particular on the entire population of Hubei, of not moving from their homes for several weeks. Therefore, I would like to suggest a series of additional interventions, in my opinion, necessary in the Po Valley as of now and in any “relapses” in the autumn period:
1) The reduction of particulate emissions, which are normally produced by winter heating systems, industries, vehicular traffic (see figure), and which travel for a few kilometers in the case of PM10 and heavier dust, but for tens or hundreds of kilometers in the case of fine and ultrafine particulates; it is therefore clear that the use of motor vehicles must be prohibited as much as possible and all non-essential polluting plants must be blocked, starting with biogas and biomass plants, which in the Po Valley have grown exponentially in the last 15 years due to the government incentives.
The various contributions to PM10 particulate emissions in Lombardy, Italy, in 2015. Winter heating (pellets, biomass, etc.) is the dominant component, while vehicular traffic contributes only a quarter of the total. (source: ISPRA data)
2) The increase in air supply and exhaust ventilation in commercial establishments and in work and domestic environments. This is a key measure for COVID-19 mitigation in the advice provided by the Federation of European Heating, Ventilation and Air Conditioning Associations (REHVA). In fact, I remember that indoor particulate pollution is correlated with atmospheric pollution when the latter is high and there are no adequate filtration systems, as happens for example. in homes, where the level of particulate matter can be even higher when cooking, smoking, etc.
3) Control of heating, ventilation and air conditioning (HVAC) systems. The current common denominator that influences the transmission and/or reduction of the transmission of airborne particles in a building is its HVAC system (Fernstrom, 2013), which – in addition to particle filtering – implements the control of three central variables known in the aerial transmission of infectious particles: temperature, relative humidity and drafts. These systems can spread, especially in high-occupancy buildings, viral vectors introduced from the outside or by infected people (“positive” in the case of COVID-19).
4) The generalized use of filter masks by the population. Even in the spring, especially in certain areas, there can be a high concentration of pollen and spores, while throughout the year the areas around the polluting plants are characterized by high concentrations of PM10 in a radius of several kilometers. It should be noted, in fact, that Italian ARPA control units for pollution measurement are always located well away from the numerous powerful point sources existing (such as industries, biogas plants, etc.) and therefore only provide an indication of the background level in relatively less polluted areas.
The ARPA control units are deliberately always placed far away from the polluting plants, therefore the graphics on the particulates supplied by them give to non-experts the illusion that within a few kilometers of these plants the concentration of particulates is low in spring and summer, while it is not so!
5) The ban on motor and non-essential activities carried out outdoors. Since, in the presence of high concentrations of particulate matter in the atmosphere, the latter acts as a vector for respiratory viruses, outdoor activities of a recreational and sporting nature, which would also entail the emission into the air of high quantities of viruses from any positive subjects. For the same reasons, flash-mobs should be banned from buildings, where even just singing by looking at each other, droplets can be created that can be transmitted both to people on the floors below and to those far away.
As we have illustrated, the transmission of respiratory viruses can take place not only through droplets, which are relatively heavy, but also through the much lighter particulate particles present in the air in indoor and outdoor environments. Therefore, the rule of social distance of 1 meter is not enough to protect the population from this second transmission channel, and the adoption of the most stringent measures suggested in this article is necessary, as happened in China, while it is now completely ignored by the restrictive measures put in place by the Italian government.
The use of filter masks is a fundamental measure that should be imposed on the entire Italian population until we have eliminated the threat of COVID-19.
Professionals and those responsible for infection control in indoor environments are still forced to use technologies that are not optimal for the purpose, as well as dated, to try to contain the transmission of infections in the air (e.g. HEPA filtration systems were developed over the years ’40). High efficiency air filtration systems are expensive to use and easily prone to leaks and problems that compromise the overall effectiveness of the system. Therefore, systems with inefficient filters – including air conditioning systems – could promote the transmission of respiratory viruses.
As more links are found between air (and indoor) pollution and the spread and severity of COVID-19, it is likely that the air we breathe will gain even more attention in the months and years to come. People all over the world are already noticing significant improvements in air quality as fewer vehicles are on the roads and many activities are stopped, as demonstrated by the comparison of air quality pre- and post-lock-down following the coronavirus emergency both in the whole of China and in the Po Valley area.
The decline in Italy of air pollution, in particular nitrogen dioxide emissions, from the levels of early January 2020 (above) to those of 11 March 2020 (below), with the country in quarantine. (source: ESA images provided by the Sentinel 5P satellite)
Some hope that an increase in awareness levels will help increase measures and pressure to reduce air pollution in the future to help the most vulnerable in their fight against this and any possible future pandemic. Others predict that, upon returning to normal activities, we could rekindle the spread of the virus, if certain thresholds in the levels of particulate pollution are exceeded locally – in time and/or in space. In any case, the link between pollution and the impact of COVID-19 should be taken into consideration from now on.
Dr. Mario Menichella – Physicist (formerly Italian National Institute of Nuclear Physics) – firstname.lastname@example.org
- Chen P-S. et al., “Ambient Influenza and Avian Influenza Virus during Dust Storm Days and Background Days”, Environmental Health Perspectives, 2010. Link
- Ciencewicki J. et al., “Air Pollution and Respiratory Viral Infection”, Inhalation Toxicology, 2007. Link
- Cui Y. et al., “Air pollution and case fatality of SARS in the People’s Republic of China: an ecologic study”, Environmental Health, 2003. Link
- Fernstrom A., Goldblatt M., “Aerobiology and Its Role in the Transmission of Infectious Diseases”, Journal of Pathogens, 2013. Link
- Malato G., “Covid-19 infection in Italy and air pollution. Is there a correlation?”, Data Science Reporter, March 2020. Link
- Menichella M., “Inquinamento globale dell’aria: le mappe satellitari”, Inquinamento Italia, 2018. Link
- Menichella M., “Inquinamento interno (indoor) ed esterno (outdoor): la connessione”, Inquinamento Italia, 2018. Link
- Peng L. et al., “The effects of air pollution and meteorological factors on measles cases in Lanzhou, China”, Environmental Science and Pollution Research, 2020. Link
- Setti L. et al., “Relazione circa l’effetto dell’inquinamento da particolato atmosferico e la diffusione di virus nella popolazione” (Report on the effect of air pollution and the spread of viruses in the population), Italian Society of Environmental Medicine (SIMA), marzo 2020. Link
- Ye Q. et al., “Haze is a risk factor contributing to the rapid spread of respiratory syncytial virus in children”, Environ. Science and Pollution Research, 2016. Link
- Zhao Y. et al., “Airborne transmission may have played a role in the spread of 2015 highly pathogenic avian influenza outbreaks in the United States”, Scientific Reports, 2019. Link