Study identifies the factors that contributed to the initial spread of COVID-19

Researchers used statistical techniques to analyze data of 126 countries. The variables that intensified contagion at the start of the outbreak included low temperature, a large proportion of older people, foreign tourist arrivals, and greeting habits involving physical contact.

By Karina Toledo  |  Agência FAPESP – When the novel coronavirus SARS-CoV-2 left the Chinese province of Hubei for Europe and other Asian countries, between December 2019 and January 2020, it found particularly favorable conditions for transmission in some regions of the planet.

After analyzing data of 126 countries, researchers affiliated with the University of Campinas (UNICAMP) in Brazil and the University of Barcelona in Spain identified a set of factors that appear to explain the rapid initial spread of the virus before public policy measures were introduced to contain the pandemic.

According to the study, which was supported by FAPESP, the factors that accelerated the early transmission of the virus include lower temperatures and, hence, less exposure to ultraviolet radiation from sunlight and less vitamin D in the bloodstream; a larger proportion of older people and hence higher life expectancy; more foreign tourists; an early start to the outbreak (the countries first hit by the virus took longer to implement preventive measures); higher prevalence of lung cancer, cancer in general and chronic obstructive pulmonary disease (COPD); a larger proportion of obese men; more urbanization; higher consumption of alcohol and tobacco; and greeting habits that involve more physical contact, such as kissing, handshaking and hugging.

“As the starting point for our analysis, we chose the day on which each country recorded the thirtieth case of COVID-19. We analyzed the ensuing days [between 12 and 20 days depending on the country]. We wanted to understand what happened during the phase in which the disease spread freely, almost exponentially,” Giorgio Torrieri, a coauthor of the article on the study posted to medRxiv in a preprint version (not yet peer-reviewed), told Agência FAPESP.

According to Torrieri, a professor in UNICAMP’s Gleb Wataghin Institute of Physics (IFGW), the researchers focused on statistical techniques typically used in physics, such as simple linear regression and coefficient of determination (R²), in an effort to understand what happened at the onset of the epidemic. They pulled data from several sources, including the public repository Our World in Data.

“We were looking for answers to these questions: if nothing had been done to contain the epidemic, then how fast would the virus have spread in different countries or social groups? Do such factors as temperature, population density, urbanization and general health influence the speed of contagion?” Torrieri said.

Reliable sources

Some studies suggest that the BCG vaccine against tuberculosis may protect people against COVID-19 to some extent. The analysis conducted by the researchers at UNICAMP and the University of Barcelona pointed to a weak correlation between the BCG vaccination and SARS-CoV-2 contagion rates, although Torrieri noted that this finding may be biased by the lack of reliable data in countries where vaccinations are not compulsory.

“The correlation was weak when countries without vaccination data were excluded, but when they were included and assumed to have low vaccination rates, the correlation was stronger,” he said.

For some of the factors analyzed, such as the prevalence of anemia, hepatitis B (among women) and high blood pressure, the researchers found a negative correlation, meaning that the initial contagion rate was slower in countries in which a larger proportion of people had one or more of these diseases. “There’s probably more cardiovascular disease in these countries, which therefore have lower life expectancy,” Torrieri said.

They found no correlation at all between the initial contagion rate and the following factors: number of inhabitants, asthma and diabetes prevalence, population density, polio, diphtheria, tetanus, pertussis and hepatitis B vaccination coverage, air pollution, the number of public holidays, and the proportion of rainy days. There was a correlation with per-capita GDP, but it disappeared when poor countries with per-capita GDP of less than €5,000 were excluded.

“GDP is associated with the quality of public infrastructure. Countries with higher per-capita GDP levels tend to have better health and housing infrastructure, for example. Below €5,000, it makes no difference, probably because of low-quality infrastructure,” Torrieri said.

As the authors stress in the paper, because several variables correlate with each other they are likely to have a common interpretation and are not easily disentangled. “The correlation structure is quite rich and nontrivial, and we encourage interested readers to study the tables [in the paper] in detail,” they write.

Some of the correlations are “obvious”, they add, citing temperature, UV and vitamin D, for example. “Others are accidental, historical and sociological. For instance, social habits such as alcohol consumption and smoking are correlated with climatic variables. In a similar vein, the correlation of smoking and lung cancer is very high, and this is likely to contribute to the correlation of the latter with climate. Historical reasons also correlate climate with per-capita GDP,” they write.

While countries cannot change some of the variables studied, such as climate, life expectancy and the proportion of older people in the population, their influence on the contagion rate should be taken into account when formulating public policies with respect to testing and social isolation.

Other variables, the authors continue, can be controlled by governments, such as testing and isolating international travelers, reducing flights to and from the most affected regions, promoting social distancing habits, campaigning to increase vitamin D intake, and combating smoking and obesity.

“We also emphasize that some variables are useful to inspire and support medical research, such as correlations of contagion with lung cancer, obesity, low vitamin D levels, blood types [or] type 1 diabetes. This definitely deserves further study […] using data from patients,” the authors conclude.

The article “COVID-19 transmission risk factors” can be downloaded at medrxiv.org/content/10.1101/2020.05.08.20095083v1.full.pdf.