A strategy to assess spillover risk of bat SARS-related coronaviruses in Southeast Asia

Abstract

Emerging diseases caused by coronaviruses of likely bat origin (e.g., SARS, MERS, SADS, COVID-19) have disrupted global health and economies for two decades. Evidence suggests that some bat SARS-related coronaviruses (SARSr-CoVs) could infect people directly, and that their spillover is more frequent than previously recognized. Each zoonotic spillover of a novel virus represents an opportunity for evolutionary adaptation and further spread; therefore, quantifying the extent of this spillover may help target prevention programs. We derive current range distributions for known bat SARSr-CoV hosts and quantify their overlap with human populations. We then use probabilistic risk assessment and data on human-bat contact, human viral seroprevalence, and antibody duration to estimate that a median of 66,280 people (95% CI: 65,351-67,131) are infected with SARSr-CoVs annually in Southeast Asia. These data on the geography and scale of spillover can be used to target surveillance and prevention programs for potential future bat-CoV emergence.

Fig. 1. Relationships among area of habitat (AOH) size, number of people, and habitat proportions.

a Scatterplot showing the total number of people living in each AOH versus the total area, for each SARSr-CoV bat host species. A best-fit line was fit through the origin, for which the R² is displayed. The four unlabeled species at the bottom left corner are (left to right): R. hipposideros, N. leisleri, T. teniotis, and R. creaghi. b Proportion of each habitat type (by area) within species AOHs. c People living in the AOH of each species, separated by habitat type. For b and c, species are listed in order of increasing AOH size, and the species order is the same in both panels. Genus abbreviations in all panels: A. = Aselliscus, C. = Chaerephon, H. = Hipposideros, N. = Nyctalus, R. = Rhinolophus, T. = Tadarida.

Fig. 2. Hotspots of SARSr-CoV bat host species richness and human overlap in Southeast Asia.

a Species richness of SARSr-CoV bat host species in Southeast Asia, created by overlaying area of habitat maps for all 26 SARSr-CoV bat host species known for this region. b Relative bat-human overlap: bat host species richness multiplied by human population count. Values were ln(x + 1) transformed and then normalized to a 0–1 scale. For both panels, redder colors indicate larger values and bluer colors indicate smaller values.

Fig. 3. Spillover simulations and sensitivity analyses.

a Density plots of the estimated total number of people in Southeast Asia infected with SARSr-CoVs by bats each year. Line colors represent scenarios exploring how estimated spillover changes if one or both of two parameter distributions (P_contact and P_detect) are refit after excluding the highest estimates of these parameters gathered via literature searches (see Methods for details). Note that the x-axis is on a log₁₀ scale. b Estimated total number of people in Southeast Asia infected with SARSr-CoVs by bats each year, plotted as a function of four input variables. Values correspond to the original scenario (i.e., no adjusted parameters). c Sobol sensitivity indices, indicating the amount of variance in the outcome due to each input on its own (first-order index) and the amount of variance in the outcome due to each input including interactions with other inputs (total-order index). Sobol indices were calculated with two random samples of 200,000 points each. Red dots represent calculated index values and error bars represent 95% confidence intervals. Values correspond to the original scenario (i.e., no adjusted parameters).

Fig. 4. Key data inputs and future research needs to improve estimates of SARSr-CoVs spillover from bats to humans.

Additional data inputs are organized according to steps in our probabilistic risk assessment to improve our understanding of: a wildlife reservoir host (bat) distribution, b people overlapping with bat hosts, c probability of human-bat contact, d probability of detecting antibodies given contact, and e dynamics of antibody response in individuals.