Strain diversity and disease risk – implications for malaria elimination

Background/Question/Methods

The goal of malaria elimination is to reduce the reservoir of malaria infection in humans to zero.  There is increasing evidence of an alarmingly large diversity of strains circulating in local populations of hosts, as defined by repertoires of antigen encoding genes (the var genes).   The consequences of this diversity for epidemiological dynamics, especially disease control, are not yet fully understood, although previous theory by Gupta and Day (1994) has proposed that ignoring strain multiplicity may inflate estimations of the basic reproductive rate of the disease and therefore underestimate the effectiveness of intervention measures.  With an agent-based modeling framework, tailored to this multi-copy family of antigen encoding genes, we have previously shown that even under high recombination a population structure of distinct strains emerges, with no simple relationship between antigenic diversity and repertoire diversity.  Furthermore, control strategies that reduce transmission, do not necessarily reduce epidemiologically relevant parasite genetic diversity.

Results/Conclusions

Consideration of parasite diversity in compartmental models of malaria transmission is typically limited to a low number of fixed pathogen strains. We construct an alternative framework with multiple strains (an S-I_n-R model) to analytically calculate the risk of infection to different strains as a function of diversity.   Based on this risk and numerical simulations, we revisit the argument relating the age of first infection to the proportion of required vaccination to achieve elimination.  Through the population dynamics of the S-I_n-R model, we evaluate the persistence of the disease after vaccination.  It is well known that because of the very young age at which children experience their first P. falciparum malaria infection, the risk of infection is assumed to be very high in endemic areas, and so is the corresponding fraction of the population that would need to be vaccinated, when a single strain is assumed.   In accordance with Gupta and Day (1994), we demonstrate that these estimates do not apply in the presence of high antigenic diversity.  We further show that the likeliness of mass vaccination success rapidly increases with diversity, as evaluated through disease persistence.  We compare the results of this tractable SI_nR model, for which strains are fixed and given, to those of simulations of our original stochastic and individual-based model in which strains are an emergent feature resulting from competitive interactions and immune selection.