Seasonality of COVID-19, Other Coronaviruses, and Influenza

Will the incidence of COVID-19 decrease in the summer?

There is reason to hope that it will, since in temperate climates influenza and the four coronaviruses that are among the causes of the “common cold” do follow a seasonal pattern, with many fewer cases in the summer. If COVID-19 is affected by season, this would obviously be of importance for policies regarding “lockdown” and provision of health care resources. Furthermore, understanding the reasons for seasonal variation might point towards ways of controlling the spread of COVID-19 (caused by a coronavirus sometimes referred to as SARS-CoV-2, though I’ll usually ignore this pedantic distinction).

I’ll look here at the evidence for seasonality in influenza and the common cold coronaviruses, and to what extent one might expect COVID-19 to also be seasonal. I’ll consider three classes of possible reasons for seasonality — seasonal changes in virus survival and transmissibility, in human resistance to infection, and in social behaviour. I’ll then consider whether we might be able to enhance such seasonal effects, further reducing the spread of COVID-19 in summer, and also extend these effects to winter.

Viral structure and immunity

To start, here’s some of my recently-acquired amateur knowledge of virology, from various sources (eg, Fenner and White’s Medical Virology, contents currently available online):

• Viruses have a protein “capsid” enclosing their genetic material (RNA or DNA).
• In “enveloped” viruses, the capsid is itself surrounded by a membrane derived from a membrane of the host.
• The viral envelope is generally needed for infectivity.
• Enveloped viruses are generally more vulnerable to environmental influences.

The virus causing COVID-19 as well as other coronaviruses are enveloped, as is influenza virus. Of the viruses responsible for the “common cold”, rhinoviruses and respiratory adenoviruses are non-enveloped, while coronaviruses, human metapneumoviruses, human parainfluenza viruses, and respiratory syncytial viruses are enveloped.

Infection with influenza or common cold viruses typically leads to at least some degree of immunity in the survivors, but this immunity may be short lived (eg, a year or so), due either to the person’s immune response fading, or to the virus mutating to a form less recognized by the immune system. Note that the “common cold” can be caused by various unrelated viruses, so it is not uncommon to get multiple colds in a year, even if one is immune to the viruses previously contracted. There are also multiple strains of influenza virus, with immunity to one not conferring complete immunity to the others.

Seasonality of common cold and influenza viruses

The enveloped common cold viruses display seasonal patterns in temperate climates (usually, but not always, with a peak in winter/spring), while the non-enveloped common cold viruses typically do not show strong seasonality. For example, see Price, Graham, and Ramalingam, Association between viral seasonality and meteorological factors (also here). The review by Moriyama, et al is also of interest.

Seasonality can be seen in the data at CDC’s NREVSS site. For instance, here is a plot of the fraction of tests of US patients with respiratory illness who are positive for each of the four common cold coronaviruses (from here):

The plot goes from April 2018 to April 2020. The two biggest peaks are in early January of 2019 and 2020, with the first being for OC43 and the second for HKU1, which are betacoronaviruses. Peaks are also seen for NL63 and 229E, which are alphacoronaviruses. Note that COVID-19 is caused by a betacoronavirus.

Influenza (also an enveloped virus) displays seasonal patterns as well, with a winter/spring peak in temperate climates, and more complex patterns in tropical climates. For example, see Tamerius, et al. Global Influenza Seasonality: Reconciling Patterns across Temperate and Tropical Regions (also here), which has the following figure:

That enveloped respiratory viruses are typically seasonal and non-enveloped respiratory viruses are typically non-seasonal is consistent with a viral envelope being a vulnerability, with the degree of vulnerability varying with the season. The season might also affect host vulnerability, however — for example, via the effect of sunlight on vitamin D levels. Many social behaviours that could affect spread of viruses are also seasonal — for example, school schedules, sporting activities, holidays, and vacations. So the reasons for seasonality could be complex.

Despite the strong seasonality seen for many viruses, such as the influenza and common cold coronaviruses in the plots above, it is conceivable that the magnitude of the seasonal effect is actually quite small. This possibility is discussed by Dushoff, et al. in Dynamical resonance can account for seasonality of influenza epidemics. If immunity is short-lived, oscillations in the number of cases are expected as the number of susceptible individuals is reduced by a widespread infection, but then rises again as immunity fades or the virus mutates. If the period of such intrinsic oscillations is approximately a year, even a small seasonal effect on infectivity could have a synchronization effect that leads to the peaks of such oscillations always occurring in the same season, producing a seasonal effect that appears to be much larger than the actual seasonal forcing.

Possible seasonal influences

The reason(s) for seasonality of influenza and other viruses are not well established. The reviews by Tamerius, et al and Moriyama, et al consider many possibilities.

It could be that the ability of some viruses to survive outside the body, either in air or on surfaces, is affected by the season. For influenza, one popular theory is that low absolute humidity is conducive to virus survival and transmission (see Shaman and Kohn and Shaman, Goldstein, and Lipsitch, for example), though it seems possible that very high humidity is also conducive to virus transmission. Both indoor and outdoor absolute humidity is typically lower in winter than in summer. High temperature may also independently curtail virus transmission, and higher UV radiation in summer may reduce virus survival. It is possible that seasonal changes in humidity and temperature also affect host resistance to infection.

The higher UV radiation in summer will (amongst people who go outside without blocking it with clothing or lotions) also increase vitamin D production. Since vitamin D levels affect the immune system, this has been hypothesized as a reason for seasonality of influenza (eg, Cannell, et al, On the epidemiology of influenza, also here). Some subsequent randomized trials of vitamin D supplementation have supported this hypothesis (eg, Urashima, et al, looking at 334 Japanese schoolchildren), while others have not found such an effect (eg, Aloia, Islam, and Mikhail, looking at 184 older African-American women). A simulation study (Shaman, Jeon, Giovannucci, and Lipsitch) reported that the seasonality of influenza was more robustly reproduced based on absolute humidity than on vitamin D levels. One should note that sunlight has other possibly beneficial effects besides vitamin D production, such as increasing levels of nitric oxide, which reduces blood pressure.

Many social activities that could affect virus transmission vary seasonally, of which school attendance is perhaps the most impactful.

Might COVID-19 be seasonal?

Seasonality seems to have been most studied for influenza, but as seen above, common cold coronaviruses are also seasonal. Both influenza and coronaviruses are enveloped, which might lead to similar seasonally-varying environmental sensitivities. Seasonal effects from host immunity and social behaviour might also be common to influenza and coronaviruses. So it’s natural to think that COVID-19 might also be seasonal.

There are really two questions regarding seasonality: First, if in future COVID-19 becomes an established human virus, will it produce seasonal epidemics, like influenza and the four common cold coronaviruses? And second, right now, during COVID-19’s initial expansion, will the pandemic slow down with the arrival of summer? As discussed above, the answer to the first question could be “yes”, but the answer to the second could be “not much” — even a small seasonal effect on infectivity could synchronize periodic outbreaks as immunity fades so they always occur in winter, while only slightly slowing the pandemic spread in a population with no immunity.

Marc Lipsitch (co-author of two papers cited above) answers the question — Seasonality of SARS-CoV-2: Will COVID-19 go away on its own in warmer weather? — with “probably not”, and more precisely “we may expect modest declines” in contagiousness, but “it is not reasonable to expect these declines alone to slow transmission enough to make a big dent”. The degree of certainty in this last statement seems unwarranted, considering that he also says that “Predicting how a novel virus will behave based on how others behave is always speculative”, unless he means only to say that a big dent from seasonal effects is far from being guaranteed.

One should note that the possibility that large seasonal peaks can result from small seasonal effects, due to synchronization of immunity-driven oscillations, does not necessarily imply that the seasonal effects for a particular viral disease actually are small — large seasonal effects could certainly also result in seasonality! And in fact, it seems that there may be some large-magnitude seasonal effects, such as that of absolute humidity on influenza as reported by Shaman and Kohn (bottom plots in Figure 1) and by Shaman, Goldstein, and Lipsitch (Figure 1).

The recent paper by Kissler, Tedijanto, Goldstein, Grad, and Lipsitch on Projecting the transmission dynamics of SARS-CoV-2 through the postpandemic period estimates the magnitude of the seasonal effect in the US for the two betacoronaviruses, HCoV-OC43 and HCoV-HKU1, and that are causes of the common cold, reporting: “According to the best-fit model parameters, the R₀ for HCoV-OC43 and HCoV-HKU1 varies between 1.7 in the summer and 2.2 in the winter and peaks in the 2nd week of January”. This is a substantial seasonal effect. The seasonal effect for COVID-19 might be either larger or smaller. In their paper, Kissler, et al consider scenarios in which R₀ for COVID-19 varies seasonally, and conclude that “SARS-CoV-2 can proliferate at any time of year”. However, this conclusion is based on simulations that assume R₀ in summer is reduced by at most 40% compared to winter; the actual magnitude of the seasonal effect is of course unknown.

It is possible that COVID-19 is not currently spreading amongst a population with no immunity — there may be some immunity to COVID-19 arising from infection with the common-cold coronaviruses, especially the two betacoronaviruses. This is considered by Kissler, et al, on the assumption that any cross-immunity is the same for HCov-OC43 and HCoV-HKU1. As seen in the plot above, in the US cold season just ended, the peak for HCoV-HKU1 in January 2020 was greater than for HCoV-OC43, whereas the reverse was the case in January 2019. A difference in cross immunity with COVID-19 between these two strains would affect how much COVID-19 is currently being suppressed in the US by cross immunity, which in turn would affect expectations of its strength next fall, when immunity induced by these common cold coronaviruses would have had longer to fade. There is presumably geographic variation in which coronaviruses were most common recently.

Can we enhance seasonal suppression of COVID-19 and extend it to winter?

If any of the possible seasonal influences discussed above actually apply to COVID-19, we might hope to deliberately enhance these influences, to further suppress the virus during the summer, and to extend these effects to winter as well. Some possible interventions of this sort seem to have a good chance of being quite effective, at substantially lower cost than the “lockdown” currently in effect in many countries, which is in any case not sustainable.

Higher absolute humidity is often seen as a factor for lower prevalence of influenza in summer, and may well inhibit COVID-19 as well. Indoor humidity is relatively easily controlled — in some buildings, by simply turning the control knob on an existing humidifier. Humidifiers are comparatively simple devices whose production could be scaled up quickly, at low cost compared to many current interventions.

There has been some, but really surprisingly little, work on humidification as a method for reducing respiratory infections. A 2018 study by Reiman, et al on Humidity as a non-pharmaceutical intervention for influenza A (also here) looked at the prevalence of influenza virus in air and on surfaces in classrooms with and without humidification, and found that the humidified classrooms had less virus. Moriyama, et al briefly mention maintaining indoor relative humidity at 40 to 60 percent at normal room temperature as a possible intervention to reduce transmission of respiratory viruses. This is also advocated here.

This level of relative humidity differs somewhat from current recommendations. In the US, the EPA recommends “…high humidity keeps the air moist and increases the likelihood of mold. Keep indoor humidity between 30 and 50 percent”, a recommendation echoed by the CDC, who also are concerned with mold, and make no mention of respiratory viruses. Research to clarify the relative health risks of mold from high humidity and of respiratory viruses from low humidity would clearly be helpful. In the meantime, it would seem prudent to maintain relative humidity at at least 50 percent, especially in buildings inhabited by vulnerable individuals.

An even lower cost intervention that has a substantial probability of reducing the spread of COVID-19 is to tell people to get some sunlight, or if they don’t, to at least take vitamin D supplements. As discussed above, there is some reason to believe that sunlight is a factor in the seasonality of influenza, probably through its effect on vitamin D levels, and perhaps also through other means. It may well be a factor for COVID-19 as well. As stated in this commentary by Haroon, et al:

Regular intake of standard-dose vitamin D is not associated with adverse effects such as hypercalcaemia or renal stones. This would be a safe intervention to recommend to at-risk population groups and strengthens the argument for increased vitamin D fortification at a population level. It is highly questionable to wait for the results of RCTs before this recommendation should be made.

There has for many years been much debate on the health benefits of sunlight and vitamin D supplementation, with some evidence implicating low levels of vitamin D in many health conditions, while other studies find no effect. There is, however, no reason to think that taking 1000 IU of vitamin D per day has any harmful effects, nor that moderate exposure to the sun is bad. Nevertheless at the Health Canada page on Sun safety basics we see exhortations to “Protect against UV rays all year round”; “Limit your time in the sun”; “When the UV index is 3 or higher, wear protective clothing, sunglasses, and sunscreen”. There is no recommendation to take vitamin D supplements. This advice can only be characterized as dangerous and irresponsible.

Another possible driver of seasonality for respiratory infections is social behaviour, with school attendance possibly being a major factor. Reducing this factor could be a considerable help in avoiding a resurgence of infection in the fall.

An obvious way to reduce virus transmission at school is to encourage home schooling. Trying to pressure families into homeschooling who would otherwise be disinclined to do so would not be appropriate, but for families who are considering this option, COVID-19 provided an additional reason to do so, and for governments to be supportive of this choice.

For children who are not home schooled, viral transmission will be reduced if they are in a smaller class. Cutting class sizes in half (say, from 20 students to 10 students) seems likely to substantial reduce transmission. (Schools that already have small classes of around 10 students would stay the same.)

This may seem infeasible — how do we quickly get twice as many teachers, and twice as many classrooms? There’s a simple solution: just cut the length of the school day in half. What used to be a class of 20 students becomes two classes of 10 students, one of which meets in the morning, the other in the afternoon, sharing the same teacher and classroom. As a bonus, the classroom designed for 20 students can now hold 10 students using desks that are further apart.

But won’t the amount children learn also be cut in half? No. First, it is widely accepted that students learn better in smaller classes. Beyond that, anyone who has experienced conventional primary education, or has children who have attended primary school, will be aware that less than half the day is devoted to learning in any case, and that the remainder can be cut with little loss, or indeed to good effect. For high school, some time could be switched to online learning, while in-person instruction, at a more individualized level, remains for those aspects of learning where it is most beneficial.

Where the students spend time outside their now-shorter school time is an obvious issue. The benefit is lost if they go to a daycare with 20 kids. Note that the school day is already shorter than the typical work day, so parents already have some scheme for child care. In light of COVID-19, in-person work days may well be shorter, with more work from home, and this could interact well with a shorter school day.

References

Aloia, Islam, and Mikhail, Vitamin D and Acute Respiratory Infections—The PODA Trial (also here).

Burrell, Howard, and Murphy, Fenner and White’s Medical Virology.

Cannell, et al, On the epidemiology of influenza (also here).

Dushoff, et al, Dynamical resonance can account for seasonality of influenza epidemics.

Haroon, et al, Should vitamin D supplementation be recommended to prevent COVID-19?.

Kissler, Tedijanto, Goldstein, Grad, and Lipsitch, Projecting the transmission dynamics of SARS-CoV-2 through the postpandemic period.

Lipsitch, Seasonality of SARS-CoV-2: Will COVID-19 go away on its own in warmer weather?.

Moriyama, et al, Seasonality of Respiratory Viral Infections.

Price, Graham, and Ramalingam, Association between viral seasonality and meteorological factors (also here).

Reiman, et al,  Humidity as a non-pharmaceutical intervention for influenza A (also here).

Shaman, Goldstein, and Lipsitch, Absolute Humidity and Pandemic Versus Epidemic Influenza.

Shaman, Jeon, Giovannucci, and Lipsitch, Shortcomings of Vitamin D-Based Model Simulations of Seasonal Influenza (also here).

Shaman and Kohn, Absolute humidity modulates influenza survival, transmission, and seasonality.

Tamerius, et al, Global Influenza Seasonality: Reconciling Patterns across Temperate and Tropical Regions (also here).

Urashima, et al, Randomized trial of vitamin D supplementation to prevent seasonal influenza A in schoolchildren.