The diminishing returns of atovaquone-proguanil for elimination of Plasmodium falciparum malaria: modelling mass drug administration and treatment.
Maude RJ., Nguon C., Dondorp AM., White LJ., White NJ.
<h4>Background</h4>Artemisinin resistance is a major threat to current efforts to eliminate Plasmodium falciparum malaria which rely heavily on the continuing efficacy of artemisinin combination therapy (ACT). It has been suggested that ACT should not be used in mass drug administration (MDA) in areas where artemisinin-resistant P. falciparum is prevalent, and that atovaquone-proguanil (A-P) might be a preferable alternative. However, a single point mutation in the cytochrome b gene confers high level resistance to atovaquone, and such mutant parasites arise frequently during treatment making A-P a vulnerable tool for elimination.<h4>Methods</h4>A deterministic, population level, mathematical model was developed based on data from Cambodia to explore the possible effects of large-scale use of A-P compared to dihydroartemisinin-piperaquine ACT for mass drug administration and/or treatment of P. falciparum malaria, with and without adjunctive primaquine (PQ) and long-lasting insecticide-treated bed nets (LLIN). The aim was local elimination.<h4>Results</h4>The model showed the initial efficacy of ACT and A-P for MDA to be similar. However, each round of A-P MDA resulted in rapid acquisition and spread of atovaquone resistance. Even a single round of MDA could compromise efficacy sufficient to preclude its use for treatment or prophylaxis. A switch to A-P for treatment of symptomatic episodes resulted in a complete loss of efficacy in the population within four to five years of its introduction. The impact of MDA was temporary and a combination of maintained high coverage with ACT treatment for symptomatic individuals and LLIN was necessary for elimination.<h4>Conclusion</h4>For malaria elimination, A-P for MDA or treatment of symptomatic cases should be avoided. A combined strategy of high coverage with ACT for treatment of symptomatic episodes, LLIN and ACT + P MDA would be preferable.