Global risk of selection and spread of histidine-rich protein 2 and 3 gene deletions.

In the thirteen years since the first report of -deleted parasites in 2010, the World Health Organization (WHO) has found that 40 of 47 countries surveyed worldwide have reported gene deletions. Due to a high prevalence of deletions causing false-negative HRP2 RDTs, in the last five years, Eritrea, Djibouti and Ethiopia have switched or started switching to using alternative RDTs, that target pan-specific-pLDH or specific-pLDH alone of in combination with HRP2. However, manufacturing of alternative RDTs has not been brought to scale and there are no WHO prequalified combination tests that use Pf-pLDH instead of HRP2 for detection. For these reasons, the continued spread of deletions represents a growing public health crisis that threatens efforts to control and eliminate malaria. National malaria control programmes, their implementing partners and test developers desperately seek deletion data that can inform their immediate and future resource allocation. In response, we use a mathematical modelling approach to evaluate the global risk posed by deletions and explore scenarios for how deletions will continue to spread in Africa. We incorporate current best estimates of the prevalence of deletions and conduct a literature review to estimate model parameters known to impact the selection of deletions for each malaria endemic country. We identify 20 countries worldwide to prioritise for surveillance and future deployment of alternative RDT, based on quickly selecting for deletions once established. In scenarios designed to explore the continued spread of deletions in Africa, we identify 10 high threat countries that are most at risk of deletions both spreading to and subsequently being rapidly selected for. If HRP2-based RDTs continue to be relied on for malaria case management, we predict that the major route for deletions to spread is south out from the current hotspot in the Horn of Africa, moving through East Africa over the next 20 years. We explore the variation in modelled timelines through an extensive parameter sensitivity analysis and despite wide uncertainties, we identify three countries that have not yet switched RDTs (Senegal, Zambia and Kenya) that are robustly identified as high risk for deletions. These results provide a refined and updated prediction model for the emergence of deletions in an effort to help guide policy and prioritise future surveillance efforts and innovation.