On the Surveillance and Spread of Drug Resistant Malaria in Sub-Saharan Africa

Date of Award

Spring 2022

Document Type


Degree Name

Doctor of Philosophy (PhD)


Public Health

First Advisor

Parikh, Sunil


Artemisinin-based combination therapies (ACTs) are among the most important tools in our arsenal to treat and control malaria. ACTs and other antimalarial drugs have contributed to significant declines in global malaria-attributed morbidity and mortality over the past few decades. However, progress is the effort to eradicate the disease has not been equally distributed: sub-Saharan Africa continues to experiences the vast majority of the burden of the disease. Recent reports of ACT resistance in Rwanda and Uganda confirm longstanding expectations that artemisinin-resistant parasites will eventually emerge in Africa. To delay the spread of artemisinin resistance across the malaria endemic sub-continent, it is essential to protect the efficacy of the ACT partner drugs that provide additional mechanisms of parasite clearance. Drug sensitivity can be measured using standardized clinical, parasitological, and molecular criteria. The assessment of molecular markers can provide rapid evidence of partner drug sensitivity at the population level, helping to monitor the geospatial distribution of resistant parasites. This dissertation begins with a scoping review to identify the state of molecular surveillance efforts related to ACT partner drug resistance. Through a systematic search, we identified over 500 surveys assessing the prevalence of four molecular markers in two key transporter genes (pfcrt 76T and pfmdr1 86Y, 184F, and 1246Y) in sub-Saharan Africa from 2004-2018, corresponding to the uptake and widespread usage of ACTs. We observed a median time lag of over three years from final sample acquisition to publication. Nearly half of all countries conducted, on average, less than one study every three years. Surveys were more likely to be conducted in urban settings and resourced by institutions external to malaria endemic countries. Overall, surveillance efforts likely failed to capture the heterogenous, spatially-dependent nature of partner drug resistance. In Chapter 2, molecular data from Chapter 1 were collated to estimate the spatial distribution of key partner drug-associated molecular markers across the sub-continent. To predict the prevalence of markers in regions without any sampling data, we fit hierarchical Bayesian spatial models to molecular data aggregated from these surveys. From the interpolated maps, we observed significant increases in the prevalence of pfcrt K76, pfmdr1 N86 and D1246 over the study period (in over 90% of study regions), suggesting widespread decreased susceptibility to lumefantrine, the partner drug component of AL. Rainfall seasonality was the strongest predictor of the prevalence of wild-type genotypes, with other covariates, including first-line drug policy and transmission intensity, more weakly associated. The models also enabled identification of regions with potential decreased susceptibility to local first-line ACTs that may be prioritized for enhanced surveillance. Our results may be used to infer the degree of molecular resistance and magnitude of change in regions without survey data. Surveillance typically relies on invasive blood samples collected from infected individuals, yet such studies are limited by costly infrastructure, regulatory oversight, and significant reporting delays. In Chapter 3, we hypothesized that mosquito blood meals could be used to efficiently and economically monitor markers of resistance in parasite populations. We conducted a series of cross-sectional surveys in southwest Burkina Faso to compare molecular markers in humans and blood-fed mosquitos. We found that infection rates in mosquito blood meals remained constant over time despite fluctuations in human prevalence. The frequency of molecular markers in humans and mosquitos were significantly impacted by multiclonal infections. We found that rates of molecular markers in humans and mosquitos were comparable, particularly for markers circulating at lower mutation frequencies. Xenosurveillance may therefore be a useful, supplementary surveillance tool for less prevalent or emergent drug resistance mutations.

This document is currently not available here.