COVID-19 and microbiome diversity in sub-Saharan Africa - Author's reply.

We thank Wim Van Damme and colleagues for their interest in our work and their Correspondence. Their comments highlight unique aspects of how the COVID-19 pandemic has unfolded in sub-Saharan Africa. These include the difficulty of estimating the true number of deaths for countries without excess mortality data, the effects of comorbidities on COVID-19 disease severity, and the distinctive immunological profile of populations in sub-Saharan Africa. We agree with the main thrust of their Correspondence and believe that clarifying some details about the scope and implementation of our analysis could be helpful.

Regarding excess mortality, we used observed excess mortality rates to estimate the true number of COVID-19 deaths in a population, a concept we call total COVID-19 mortality, as part of a statistical model that included many covariates.1, 2 Among others, the set of covariates included: prevalence of diabetes for year 2019, prevalence of smoking for year 2019, and prevalence of hypertension for year 2019 (described in appendix 1 [p 10] in the Article). Out-of-sample predictions therefore did take into consideration comorbidities that modify the severity of COVID-19 disease and are not merely an average of existing data.

Regarding within-population differences in the infection–fatality ratio, we agree that characterising heterogeneity by wealth and urbanicity would be helpful. Most data sources do not take these factors into account, however, making this difficult to do directly.

Finally, we agree that a population's immunological characteristics should be considered alongside other aetiological explanations for variation in the infection–fatality ratio. Research linking microbiome diversity to COVID-19 severity (ie, the hygiene hypothesis) is methodologically challenging but important. We did briefly discuss co-infection as a potential explanation for mortality patterns in sub-Saharan Africa. The topic deserves a deeper discussion, however, given research (eg, a study by Wolday and colleagues) indicating that previous exposure to certain pathogens can serve as an effect modifier on the relationship between SARS-CoV-2 infection and COVID-19 health outcomes.