3.1 Growth rate of new cases and deaths
Table 1 shows the descriptive statistics of meteorological data, air pollution indicators, and the growth rate cases and deaths of coronavirus disease (COVID-19). During the observation period (December 31, 2019, to October 12, 2020), the number of COVID-19-related cases and deaths were 649,260 and 16,249,266 in 210 countries, respectively. The average growth rate of COVID-19 cases, including incidence and deaths, was approximately 0.32 and 0.77 among 210 countries, respectively. The lowest (zero) average growth rate of COVID-19 incidence was related to Anguilla, Grenada, Montserrat, and the highest (2.99) to Bosnia and Herzegovina. The highest average growth rate of deaths was observed in Chile (1.69), Yemen (1.55), and Kyrgyzstan (1.52) [Figure 1].
Table 1
Summary of COVID-19 average growth rate case and death, meteorological data
variables | Daily measures |
Mean ± SD | Min | P 25 | Median | P 75 | Max |
Mean growth rate cases | .77 ± 0.49 | .00 | .35 | .81 | 1.12 | 2.99 |
Mean growth rate death | .32 ± .38 | .00 | .005 | .12 | .63 | 1.69 |
Meteorological factors | Temperature (°C) | 24.66 ± 6.16 | 1 | 21.00 | 26.00 | 28.00 | 43 |
Humidity (%) | 69.14 ± 22.65 | 6 | 54.00 | 74.00 | 86.75 | 100 |
Pressure(mbar) | 1000.04 ± 43.67 | 723 | 1005.00 | 1011.00 | 1015.00 | 1027 |
Wind(ms) | 8.41 ± 11.46 | 0 | 2.25 | 5.00 | 11.00 | 136 |
AQI | 58.14 ± 66.07 | 4 | 25.00 | 45.00 | 71.50 | 184 |
COVID-19, Coronavirus disease; SD, Standard Deviance; Min, Minimum; P25, 25th percentile; P75, 75th percentile; Max, Maximum |
Concerning temperature, New Zealand had the lowest temperature (1 ° C) and Qatar had the highest temperature (43 ° C) in Asia. Tunisia reported the lowest humidity (6%), and Bhutan and Mexico reported the lowest (723) and highest (1027) pressure (mbar), respectively. As to the air pollution, Belarus had the best air quality (AQI = 4) while Uganda had the worst air quality (AQI = 184). In terms of wind (mph), New Zealand registered the fastest wind speed of 136 mph.
The COVID-19 average growth rate of daily cases showed no correlation with humidity (r = -0.08, P = 0.151). Also, the average growth rate of daily deaths of COVID-19 was not associated with humidity (r= -0.09, P = 0.207). Instead, there was a negative correlation between the average growth rate of daily COVID-19 cases and deaths with the Air Quality Index (AQI) and wind (r= -0.25, P = 0.04). Humidity was not correlated with AQI (r= -0.15, P = 0.061) but with temperature (r= -0.31, P ≤ 0.001). Furthermore, there was a strong positive correlation between PM10 and the average growth rate of daily COVID-19 cases (r = 0.33, P ≤ 0.001) and a trend toward a significant correlation with deaths (r = 0.13, P ≤ 0.056) [Figures 2].
According to the model, AQI had a nonlinear effect on COVID-19 average growth rate of daily cases (P = 0.018 and edf = 1.751), pressure (P = 0.047, edf = 1.00) and wind (P = 0.00029, edf = 1.953) [Figure 4]. On the other hand, COVID-19 average growth rate of daily deaths was affected by AQI (P = 0.0162, edf = 1.00) and wind (P ≤ 0.001, edf = 1.997) [Figure 5].
3.2 Adjusted Cumulative incidence of daily cases and deaths
The highest cumulative incidence rate of daily COVID-19 cases in the first sixty-day interval was related to Kyrgyzstan with 132,032.36 per 100,000 people, in the second sixty-day interval, Bahrain with 1647.43 per 100,000 people, and in the third sixty-day interval, Qatar with 2647.81 per 100,000 people. In the fourth sixty-day interval, the number was close to zero for all countries, but it was shown that in the last sixty-day interval, Tajikistan registered 600,000 cumulative incidence rate of daily cases per 100,000 people. The daily cases based on the population size of countries have been reported. The highest cumulative rate of daily deaths in the first sixty consecutive days belongs to Peru with 46.80 per 100,000 people, in the second sixty consecutive days to San Marino with 104.49 per 100,000 people and in the third sixty consecutive-days to Andorra with 40.69 per 100,000 people.
The cumulative incidence rate of COVID-19 daily cases over the third sixty-day interval was not correlated with humidity (first 60days: r=-0.07, p = 0.289, second 60 days: r=-0.19, p = 0.027; third 60 days: r=-0.19, p = 0.153). The cumulative rate of COVID-19 daily deaths during the third sixty- days period showed no correlation with humidity (first 60 days: r = 0.02, p = 0.808, second 60 days: r=-0.04, p = 0.634; third 60 days: r=-0.02, p = 0.882). The cumulative incidence rate of COVID-19 daily cases within the first and second sixty-day intervals was not correlated with Air Quality Index (AQI) (first 60 days: r=-0.02, p = 0.264, second 60 days: r = 0.19, p = 0.185; third sixty-days: r = 0.10, p = 0.041). Finally, the cumulative rate of COVID-19 daily deaths over the first and second sixty-day intervals was not correlated with Air Quality Index (AQI) (first sixty-days: r = 0.17, p = 0.260, second sixty-days: r=-0.04, p = 0.268, third sixty-days: r=-0.07, p = 0.024). Furthermore, the cumulative rate of COVID-19 daily cases and deaths from the first to the third sixty-day intervals was correlated with PM10 (daily cases: first 60 days: r = 0.34, P ≤ 0.001, second 60 days: r = 0.22, P ≤ 0.001, third 60 days: r = 0.21, P ≤ 0.001; deaths: first sixty- day interval: r = 0.51, P ≤ 0.001, second sixty-day interval: r = 0.58, P ≤ 0.001, third sixty-day interval: r = 0.65, P ≤ 0.001) [see Fig. 3].
According to the model AQI, temperature, pressure, humidity, as well as wind, had no non-linear effect on the cumulative incidence rate of daily COVID-19 cases in the first sixty-day period (edf = 1for five variables). The cumulative incidence rate of daily COVID-19 cases over the second sixty-day period was affected by temperature (P = 0.004, edf = 3.25) [see Fig. 6]. The cumulative incidence rate of daily COVID-19 cases during the third sixty-day interval was affected by temperature (P = 0.0002, edf = 8.73) and humidity (P < 0.001, edf = 8.71) [see Fig. 7]. The cumulative incidence rate of daily COVID-19 cases in the fourth sixty-day interval was zero. Cumulative incidence rate of daily COVID-19 cases in the fifth 60 days was affected by temperature (P = 0.025, edf = 1.39), humidity (P = 0.023, edf = 6.17) and pressure (P < 0.001, edf = 8.83) [see Fig. 8].
The cumulative rate of COVID-19 daily deaths in the first sixty-day interval was affected by AQI (P = 0.043, edf = 1.00) and temperature (P = 0.001, edf = 1.00). Besides, the cumulative rate of daily COVID-19 deaths in the second sixty- day interval was affected by pressure (P = 0.012, edf = 7.32), and the cumulative rate of daily COVID-19 deaths in the third, fourth, and fifth sixty-day intervals was not affected by meteorological factors and air pollution indicators.
In this study, we developed a generalized additive model to investigate the relationship of meteorological factors and air pollution indicators with COVID-19 average growth rate of daily cases and deaths as well as the cumulative incidence rate of daily COVID-19 cases and deaths based on the population size of 210 countries spread over five consecutive sixty-day intervals from December 31, 2019, to October 12, 2020.
Our results revealed that there was a negative, non-significant correlation between humidity and COVID-19 average growth rate of daily cases and deaths, but based on the population size of countries, daily COVID-19 cases and deaths from the first to the third sixty-day interval did not correlate with meteorological measurements and air pollution indicators but with PM10.
Most importantly, this study rejects the emerging hypothesis that absolute humidity, even after temperature control, has been observed to be an important determinant of COVID-19 results. These results weaken the recent laboratory findings on guinea pigs and state-level epidemiologic evidence (Shaman, Jeffrey and Pitzer, Virginia E and Viboud, cile and Grenfell, Bryan T and Lipsitch, 2010; Shaman and Kohn, 2009), but our key conclusion is that there is a non-linear relationship between AQI and COVID-19 average growth rate of daily cases and deaths.
Moreover, the cumulative incidence rate of COVID-19 daily cases and deaths, within the second sixty-day interval was affected by temperature, during the third sixty-day interval by temperature and humidity, and over the fifth sixty-day period by temperature, humidity, and pressure. The cumulative rate of COVID-19 deaths in the first 60 days was influenced by AQI and temperature and in the second 60 days by pressure. It appears that meteorological factors and air pollution indicators affect different epidemiological features and each of these variables has a significant impact on the disease in the first sixty-day period.
According to the results, there was no relationship between temperature and COVID-19 average growth rate of daily cases and deaths. A linear relationship but not a significant one between ambient temperature and COVID-19 average growth rate of daily cases and deaths was observed. There was also a relationship between temperature and cumulative incidence of daily COVID-19 cases and deaths. Previous studies have shown that temperature was also an important factor in the survival and transmission of other viruses, such as SARS-CoV and MERS-CoV (Bi et al., 2007; Casanova et al., 2010; Chan et al., 2011; Tan et al., 2005; Van Doremalen et al., 2013). Bi et al. (2007) (Bi et al., 2007) reported that temperature was negatively related to SARS transmission in Hong Kong and Beijing in 2003. A laboratory study was carried out with virus substitutes. This study investigated the effect of temperature on the survival of coronavirus on surfaces. The results showed that viruses were inactivated more rapidly at 20°C than 4°C (Casanova et al., 2010). Another laboratory study found that coronavirus has been stable on flat surfaces for more than five days at temperatures between 22°C to 25°C. Also, the results of the study demonstrated that the virus was rapidly killed at high temperatures, such as 38°C (Chan et al., 2011). Van Doremalen et al. (2013) also found that MERS-CoV is more stable at high temperatures than at lower temperatures. This issue revealed that COVID-19 may not be eliminated without public health interventions in warmer weather. Therefore, people and governments cannot expect to eradicate this new virus at high temperatures. Several studies have reported that death from respiratory diseases increases with decreasing temperature, and it is strongly associated with low temperatures (Dadbakhsh et al., 2017; Fallah Ghalhari and Mayvaneh, 2016; Gómez-Acebo et al., 2013; Macfarlane, 1977). Whereas, another study revealed that both cold and hot weather may have adverse effects on death and respiratory diseases (Li et al., 2019). In the case of exposure to cold and immune system function, it has been reported that low temperatures may suppress the function of the immune system (Shephard and Shek, 1998).In particular, previous findings showed that the phagocytic function of the tested pulmonary alveolar macrophages decreased under laboratory conditions and cold stress (Luo et al., 2017). Breathing cold air can lead to bronchial contraction which, in turn, can increase susceptibility to lung infection (Martens and medicine, 1998). In summary, most studies have shown that the optimum temperature for killing the virus is high and that high temperatures are detrimental to its survival (Van Doremalen et al., 2013). Moreover, Some other studies have shown that temperature and humidity correlated with underlying diseases such as heart and respiratory diseases (Hajat and Haines, 2002; Lin et al., 2009; Schwartz et al., 2004; Tian et al., 2012), which in turn could increase the number of COVID-19 deaths exacerbated by such underlying diseases (Chen et al., 2020; Masetti et al., 2020). However, this paper has not found the effect of high temperature on the growth rate of COVID-19 infection. One possible reason may be related to the ecological constraints of the study and diurnal temperature variation. In this regard, other laboratory studies are required to determine the underlying mechanism.
The study did not show any relationship between the growth rate of COVID-19 with temperature and humidity but there was a relationship between the cumulative incidence rate of daily COVID-19 cases and deaths with temperature and humidity. Several researchers have confirmed that respiratory infections have increased in unusual conditions of cold weather and low humidity and low humidity may also be a major risk factor for respiratory diseases (Davis, Robert E, and McGregor, Glenn R and Enfield, 2016; Davis et al., 2016). A 25-year study found that humidity is an important factor in the number of deaths and low humidity may greatly increase death rates, potentially by influenza-related mechanisms (Barreca, 2012). This result is similar to a study conducted in the United States (Barreca and Shimshack, 2012). Breathing dry can either damage the epithelium or reduce mucus secretion, resulting in respiratory viral infection (Lowen et al., 2007). Also, the transmission of the pandemic influenza virus is an effective factor in cold and dry conditions (Steel et al., 2011) and the survival rate of the influenza virus has significantly increased due to the reduction in absolute humidity. This situation holds about coronavirus (Shaman and Kohn, 2009) but our results indicate that the COVID-19 average growth rate of new cases and deaths had no association with absolute humidity. Tosepu et al. (2020) reported that COVID-19 prevalence is significantly dependent upon the average temperature. However, it is claimed that COVID-19 prevalence is not significantly dependent upon the minimum and maximum temperatures, rainfall, and humidity (Barreca, 2012). In another study, Chan et al. (2011) claimed that the higher temperatures and higher relative humidity (e.g., 38˚C and relative humidity above 95%) make the virus stable (Barreca and Shimshack, 2012). Our results also show the effect of wind on the COVID-19 average growth rate of daily cases and deaths. Mehmet Şahin et al. (2020) indicated that wind speed and population density are correlated with COVID-19 (Şahin, 2020). The spread of the virus can be affected by wind conditions. This finding that the COVID-19 virus, which is not visible in the air, spreads more in the air, suggests that the spread of the virus through the air is a threat to humans, especially when the wind speed increases airflow (Coşkun et al., 2021). An important question that arises here is that can the virus spread through infectious dust or not? The answer to this question is out of the scope of the present research and more laboratory studies and modeling should be done in this field.
Importantly, we did find a relationship between COVID-19 average growth rate of daily cases and deaths and AQI in our study. This may be due to the impact of AQI itself, the increase in traffic within the city, and the reduction in social distance. Other studies have reported an association between air pollution and decline in lung function, increased hospitalization, increased respiratory symptoms, and higher use of asthma medications (Jalaludin et al., 2004; Simoni et al., 2015). Several studies have examined the effects of meteorology and air pollution on acute viral respiratory infections and bronchiolitis (it is a disease associated with seasonal changes in respiratory viruses) during the early years of life (Nenna et al., 2017; Ségala et al., 2008; Vandini et al., 2015, 2013). The composition of environmental particles or particulate matter (PM) varies geographically and seasonally. This situation is due to the combination of resources at each location and time. A large body of literature shows the short-term effects of air pollution on health, but long-term air pollution also affects morbidity in the long run (Lowen et al., 2007).
Furthermore, the findings of our study showed that PM10 was positively associated with the cumulative rate of deaths in the first, second, and third sixty-day intervals. The study by Yao et al. (2020) showed that there was a positive correlation between PM10 and Case Fatality Rate (CFR) from COVID-19 in Wuhan, China. The findings of studies conducted by (Conticini et al., 2020; Piazzalunga-Expert, 2020; Travaglio et al., 2020; Wu, Xiao, and Nethery, Rachel C and Sabath, Benjamin M and Braun, Danielle and Dominici, 2020; Yongjian et al., 2020) were consistent with the result obtained by this study. One explanation for the geographical differences in the number of cases is that high levels of pollutants may transmit viral diseases and increase their persistence in the environment (Frontera et al., 2020). A recent study suggests that the virus may survive in the aerosol for several hours (Van Doremalen et al, 2020). Air pollution may also play a stronger role when susceptibility to infection and the pulmonary defense mechanisms that lead to more severe forms of the disease increases (Frontera et al., 2020). Other noteworthy issues include prolonged exposure to air pollution that impairs lung function (Chauhan and Johnston, 2003; Götschi et al., 2008) and may reduce resistance to viral infections (Yang et al., 2020) and prolonged traffic-related pollution which is an important source of exposure. Typical inactivation of viruses is showed by phagocytosis in which macrophages prevent the virus from replicating and kill the virus-containing cells. Cytotoxic T-lymphocytes (CTLs) destroy infected cells. Such functions may likely be disrupted by exposure to air pollutants (Chauhan and Johnston, 2003) Therefore, air pollution as well as PM10 can be among the factors affecting the epidemiological characteristics of COVID-19.
The strengths of this study are:
-
It was conducted across 210 countries affected by the COVID-19 disease. It is the most comprehensive study on the relationship between environmental factors and COVID-19 in which a large number of countries and multiple variables were included.
-
It was examined the cumulative incidence of daily cases and death COVID-19 based on the population size of incidence and death in these 210 countries, which can help decision-makers control the emergence of daily cases and deaths.
However, there are some limitations with the present study that should be acknowledged. The first limitation of the study is that other important factors may affect COVID-19, including the number of confirmed cases, the number of deaths, and the number of recovered cases such as government intervention, medical resources, and so on. Therefore, these issues should be considered in future studies. The second limitation is the design of the ecological studies which is the case in the present study. This issue may have some ecological fallacy. Besides, although the outdoor and indoor air conditions are similar, collecting meteorological and air pollution data is challenging at an individual scale. This similarity may be due to the problem of shutting down the air conditioning system and keeping the window open for 24 hours in hospitals during the COVID-19 treatment.
The findings of this study contribute to the existing research about the COVID-19. To better understand the COVID-19 mechanism, however, it is essential to investigate more laboratory research. Besides, it is necessary to provide more observational and ecological evidence for understanding the COVID-19 consequences in the complex world of human behavior but despite all these limitations, our results provide important implications for health and public policies. Estimates can be used to predict the place and time of death from COVID-19 in the future. Also, the prominent non-linear relationship between AQI and COVID-19 average growth rate of daily cases and deaths may particularly be useful for understanding regional changes in COVID-19 results. This could be a result of climate change. Moreover, this research demonstrates that future explorations regarding the role of AQI in COVID-19 will yield important results, but more laboratory research is needed to determine the effects of humidity, evaporation as well as temperature on viral protein structure and survival.