The description ‘seasonal infection’ is often used to specify infectious diseases that cause outbreaks at particular times of the year. Factors such as temperature and humidity, and human behaviours during the summer and winter months, can interact with changing levels of population immunity to influence how common a virus infection is within the community at different times of the year. The transmission of some respiratory viruses, such as influenza, varies seasonally with outbreaks generally occurring during the winter months. This seasonality is most pronounced in temperate regions of the world, with more complicated trends observed in the tropics, although exceptions do occur. For example, during the 2009 H1N1 influenza pandemic there was a summer wave of infections as the virus spread through highly susceptible northern hemisphere populations.
Because SARS-CoV-2 has emerged only recently as a human pathogen, there is still great uncertainty as to whether its transmission will vary seasonally in different parts of the world. However, we do know that it is transmitted between people via direct and indirect contact with respiratory droplets, similarly to other respiratory viruses such as the four human coronaviruses that have circulated in the human population for decades. These coronaviruses, usually associated with mild cold symptoms, tend to spread alongside influenza during the winter in temperate regions.
What can we learn from other seasonal human coronavirus infections?
Four human coronaviruses are commonly found in human populations: HKU1, HCoV-229E, HCoV-NL63 and HCoV-OC43. In the United Kingdom and in other temperate regions of the world, these viruses circulate during the winter months, usually peaking between January and March. A detailed study of their prevalence and peak time of infection was conducted in Scotland between 2005 and 2017. Patients that presented at their GP surgery or hospital with respiratory illness were screened against a panel of viruses that cause respiratory infections, including seasonal human coronaviruses. These coronaviruses varied in prevalence, with HKU1 having such a low prevalence that screening for this virus was discontinued in 2012; HCoV-OC43 had the highest prevalence in this Scottish population. The study also found increased spread during the winter but revealed differences in the year to year spread of each type of coronavirus. During the 13 years over which data were collected, HCoV-229E peaked every second year, whereas HCoV-OC43 and HCoV-NL61 infections peaked every year, though the patterns were not always consistent. One explanation for these observations is that infection with one coronavirus confers a degree of protective immunity, not only to that coronavirus but also cross-protective immunity against other coronaviruses, but further research is needed to confirm this. It is also not yet known whether there might be cross-protection between the four well-known coronaviruses and the new coronavirus SARS-CoV-2 and, if so, for how long.
Data from the seasonal spread of the four human coronavirus outbreaks have been used in a mathematical model to study the spread of SARS-CoV-2 (and COVID-19). The model predicts that, in the short-term, the spread of SARS-CoV-2 would be strongly influenced by the duration of immunity that a person builds following an infection, the degree of cross-immunity to other coronaviruses, and by the nature and duration of non-pharmaceutical interventions such as physical distancing. The model predicted that SARS-CoV-2 outbreaks could occur at any time of year. However, outbreaks emerging in late winter or spring are likely to have less severe peaks than outbreaks occurring in autumn or early winter. These conclusions depend on the similarity of SARS-CoV-2 transmission to that of the common human coronaviruses, including seasonality. If there are similarities, and if SARS-CoV-2 establishes itself permanently in the human population, then it is likely that outbreaks will occur most frequently during the winter months.
Can the weather influence the spread of SARS-CoV-2?
A number of studies have explored the possible link between weather and the number of COVID-19 infections, with no clear consensus as yet. Many of these studies are currently available from pre-print servers and have not yet been peer-reviewed, meaning that their findings should be interpreted cautiously and critically. One approach is to compare the average temperature and humidity with the rate of viral transmission across different geographical areas. A study estimating the basic case reproduction number (R0, a measure of the rate of infection) across provinces in China found no correlation with temperature or humidity. This finding is consistent with a further study of 224 cities across China, which observed no clear link between ambient temperature and rates of new infections. On the other hand, a separate study of 429 Chinese cities found a negative relationship between the rate of spread of the virus and local temperature.
So far, therefore, there is only weak evidence for seasonal variation in SARS-CoV-2 transmission, with the suggestion of stronger spread during winter months in temperate regions. Further research is needed to confirm this seasonal variation and then to understand whether it results from differences in virus survival outside of the human host, or seasonal differences in human social behaviours. If SARS-CoV-2 does interact, through cross-immunity, with other seasonal coronaviruses already circulating in human populations then these interactions might indirectly cause seasonal changes in the prevalence of COVID-19.