Serological Variation and the COVID-19 Death Toll
I have written before about the upside-down burden of the COVID-19 pandemic caused by SARS-COV-2 that originated in Wuhan, China. Quite unlike any other disease burden, COVID-19 is more prevalent in places that are richer, live longer, are taller, and have generally higher health status. This is deeply perplexing — although the medico-biological scientific community is too deep into figuring out the etiology of the pandemic to pay attention to the overall pattern. Basically, everything we expected going into the pandemic in February turned out to be wrong. Instead of being negatively correlated with per capita income, socio-development index, mean adult height, life expectancy, protein intake per capita, calorie intake per capita, and absolute latitude, the correlations turned out to be unexpectedly positive. The next figure shows my estimates from the end of May.
I documented back then that phylogenetic distance predicts the COVID-19 toll, even after controlling for geodesic distance. This raised the specter that exposure to the pandemic is a function of population genetics. Fortunately, as it turned out, populations at the losing end of global polarization are less exposed to COVID-19 than otherwise luckier northern populations.
These were obviously first-pass estimates, subject to revision in light of further data and more careful analysis. Now, it turns out that I was right on the money. Part of the explanation of the upside-down toll of the novel coronavirus turns out to be serological variation. I have documented variation in ABO blood group frequencies before in the context of the intellectual history of serological racialism. That turns out to be very handy in understanding the cross-section of the COVID-19 burden.
In a paper published in The New England Journal of Medicine, by an army of geneticists, it has been shown than genetic markers related to ABO blood group are a significant risk factor for the severity of COVID-19. As we shall see, this goes some way towards explaining the upside-down pattern of the COVID-19 death toll.
We show that the frequency of blood group O is positively associated with COVID-19 death rates, while the frequency of blood group B and the heterozygous blood group AB is negatively associated with COVID-19 death rates. The latter makes sense in light of the general pattern of the beneficial effects of heterozygosity, which captures genetic variation in a population. Similarly, the deleterious effect of O makes sense because it is homozygous. The protective effect of B on the other hand, suggests specific genetic factors at play in exposure to COVID-19. Note that we do not adjust for age. Researchers intrigued by these results should try to control for that.
The frequency of the homozygous blood group O is positively associated with COVID-19 deaths per capita in the international cross-section. Recall that the frequency of O is a measure of homozygosity or lack of genetic variation. This goes some way towards explaining the higher burden in America.
The frequency of blood group B is negatively associated with COVID-19 deaths per capita. Recall that the frequency of B increases as we move east on the Eurasian continent. This explains why the death toll in South Asia is so low compared to Europe.
Finally, the frequency of the heterozygous blood group AB, is negatively associated with COVID-19 deaths per capita. This makes perfect sense in light of the general protective effect of heterozygosity. Since Africa has the highest amount of heterozygosity and the New World the lowest, this explains the extraordinarily low COVID-19 toll in Africa, relative to Europe and the Anglo-Saxon settler colonies.
Genetic reductionism should never be our first resort. But especially in epidemiological contexts, it cannot be ruled out and must be on the table. Our initial results had suggested that genetics may play a role. Now it seems that genetics, and classical genetic markers at that, can help explain the upside-down toll of the novel coronavirus. Exciting stuff!