A molecular strategy to trace the origins of malaria cases and map transmission potential in countries approaching elimination
To implement a sustainable, field-friendly system for assessing and monitoring the microepidemiology of malaria transmission throughout Swaziland and in a district of northern Namibia.
Introduction: A critical need for countries attempting to eliminate malaria is the ability to distinguish local from imported infections. Combating local transmission requires local, focused interventions while combating imported transmission requires control at a distant focus of infection or at international boundaries. In addition, quantifying local transmission is critical in assessing progress toward elimination, efficacy of interventions, and eventual confirmation of elimination. Our aim was to add molecular genotyping of Plasmodium falciparum infections to existing surveillance systems to infer malaria transmission trees, thereby determining the origin of infections and quantifying local transmission potential.
Methods: collected dried blood spot samples and epidemiologic data, including location of residence and travel history, from individuals in Swaziland identified as being infected with P. falciparum via passive or reactive case detection beginning in 2012. Samples from 115 individuals with confirmed infection detected between 2012 and 2013 were successfully genotyped. Genotyping consisted of 29 microsatellites located in 10 regions of the genome, creating short haplotypes of 2-5 microsatellites within 10kb of each other at each locus. Transmission networks were estimated from genetic and temporal data using a Bayesian Metropolis coupled Markov chain Monte Carlo (MC3) approach. Essentially, the most likely ancestor(s) for any given infection were estimated via calculation of likelihoods which incorporated the probability of each allele being propagated or lost during transmission, potential genotyping error, and the possibility of intermediate, unobserved hosts, which may introduce additional alleles. Numerous simulated “true” networks were created over a range of epidemiologic and genomic parameters to evaluate the accuracy of network estimation.
Results: Comprehensive evaluation of network estimation on simulated data sets is ongoing, but so far shows excellent performance for network accuracy and estimation of key parameters ranging from epidemiologic settings with many or few imported cases. Networks analyzed from genetic and temporal data collected in Swaziland reveal that the majority of transmission is comprised of short, terminal chains, consistent with excellent malaria control in most of the country, but with a limited number of longer, persistent chains of transmission indicating significant ongoing transmission in some areas.
Discussion: In this project, we have developed a series of laboratory and analytical tools to, for the first time to our knowledge, formally estimate malaria transmission networks using parasite genetic and temporal data. The approach has been developed to handle the unique challenges previously making such estimation difficult – namely the frequent occurrence of superinfection leading to polyclonal infections, and parasite genetic variability primarily generated via sexual recombination. Broader application of these methods may allow assessment of the contribution of people movement within and beyond country borders to transmission, targeting of local foci of transmission, evaluation of the efficacy of interventions, and eventual confirmation of elimination.