Wish I knew what CV-19 was up to next. How seasonal is it? Sneaky little bug. Looking at a number of countries, it appears that when something like 20% or so of the population has an infection, transmission slows noticeably. The uncertainty in how many people actually have been infected makes it difficult to ascertain potential dynamics. Several recent papers explored the course of the epidemic. I present all of these with my usual cautions about modeling-based studies.
The first paper looked at the role of seasonality and second waves. (Medrxiv Paper) The authors note that seasonality is a complex concept which includes many aspects of weather and human behavior. In general, many respiratory infections appear to have a pattern of stronger prevalence in the winter months and lower in the summer. The massive lockdowns around the world may have a significant impact on any seasonality effect. Taking everything into account, the paper projects that around September 15, we will see a pattern reversal in the northern hemisphere, with a succeeding reversal the following April 15. I think we have already seen very complex dynamics in the US alone, which aren’t consistent with this simplistic picture. The Northeast had a very bad epidemic in spring, but much of the upper Midwest, at the same latitude, did not. Then the South and Southwest had a strong wave, right in the middle of summer. Other areas, like Europe, have similar complexity. So I am withholding judgment on seasonality effects.
The next study was a more sophisticated version of modeling, with the goal of assessing whether a second wave might occur. (Medrxiv Paper) The authors are from the United Kingdom and looked at the specific projections made by the standard epidemic model of a larger second wave in that country. In contrast to those standard models, they incorporate assumptions of variability in exposure to CV-19, susceptibility to infection and susceptibility to being infectious, as have several other papers in recent weeks. They test these assumptions against data from a number of countries and find that a model with these variability characteristics is more consistent with that data. And they conclude that any second wave in the UK, and most countries, would be very limited. There is some sophisticated math here, which I will avoid attempting to describe, but the essence is that it is designed to test a large number of model variations quickly and assess which model matches reality best. I would suggest that the models which incorporate this heterogeneity in critical parameters have defined the course of the epidemic more accurately than those that ignore this variability.
This paper examined Bangladesh’s epidemic experience to determine if population immunity was close or feasible. (Medrxiv Paper) The researchers used an adjusted SIRD (Susceptible, Infected, Recovered, Death) model, which focused on age-based clustering of cases and contacts to find that transmission would slow to a low level at around a 30% population infection rate. Due to population clustering in that country around ages 20 to 40, where people are working and have a high level of contacts, but tend to have non-severe illness, if 90% of that age group became infected, there would be limited transmission to older, more vulnerable groups. Again, this paper indicates that transmission may already be substantially slowed in a number of countries, and incorporates notions of heterogeneity in key parameters.
And this paper looks at when CV-19 may shift from an acute epidemic phase to an endemic, or present, but at a low steady level, pathogen. (Medrxiv Paper) The researchers describe CV-19 as causing significant disease severity because it attacks a frail elderly population with an unprepared immune system. The authors build a model based on three types of immunity that progress along with an epidemic. One is often referred to as “sterilizing” immunity–the pathogen is eliminated before it has any chance to replicate, another is immunity that attenuates disease upon reinfection and the third is immunity that reduces infectiousness. I think this is more likely a spectrum of response which is linked and the effects of which may be dependent on the dose of the virus encountered. Looking at the existing seasonal coronavirus, they cite data that most children have been infected and some form of immune response by age 15, but that the immune response may wane rapidly, although subsequent reinfections will be milder. The overall prediction of the paper is optimistic: that CV-19 will become a mild, endemic strain, similar to the current seasonal coronaviruses. So far this seems consistent with the relatively short course of the epidemic.