Predicting the antimalarial effect of artesunate-mefloquine combination therapy using a mechanistic pharmacokinetic-pharmacodynamic model

Background: Malaria is a mosquito-borne infectious disease, with the most severe form caused by the parasite Plasmodium falciparum. The first-line defence against uncomplicated falciparum malaria is artemisinin-based combination therapy (ACT), involving multiple administrations of two or more antimalarial drugs. ACTs are highly effective worldwide, however, recently decreased efficacy of the ACT artesunate-mefloquine has been reported in South-East Asia (Dondorp A et al, 2009).

Objectives: To explore how variation in (1) the pharmacokinetic profile of artesunate and (2) parasite pfmdr1 gene copy number affects the therapeutic response to artesunate-mefloquine.

Methods: Parasite-time profiles for artesunate-mefloquine were simulated according to a mechanism-based pharmacokinetic-pharmacodynamic model. The model included parameters to represent the concentration-time profiles of the two drugs, the parasite killing rates of the two drugs, and parasite multiplication due to asexual reproduction of the parasite. Simulations were performed using R, with the differential equation describing the change in parasite burden over time integrated numerically using the odesolve library. The starting parasite burden was 1.137 x 10^11 parasites per patient.

(1) Between subject variability in dihydroartemisinin (the biologically active metabolite of artesunate) pharmacokinetic profiles results in 2-3% of patients failing to eliminate all parasites (known as radical cure). The parasite count for these patients typically falls below the limit of detection (known as parasite clearance time) around 3 days following treatment initiation, rising back above this limit about day 55.
(2) The proportion of patients who fail to achieve radical cure increases substantially as the population prevalence of increased pfmdr1 copy number increases from 0% to 50%. However, the distribution of parasite clearance times among the patient populations increases only slightly. For patients who fail to clear all parasites, parasite recrudescence above the limit of detection typically occurs around 25 days after treatment initiation.

Conclusions: The increased parasite clearance times reported in Cambodia by Dondorp et al cannot be explained by between-subject variation in dihydroartemisinin pharmacokinetics or by increased prevalence of increased pfmdr1 copy number. The long periods for which parasite counts can remain below the limit of detection without radical cure occurring emphasizes the need to follow up patients treated with this therapy for the recommended 63 days. We plan to extend our pharmacokinetic-pharmacodynamic model in the future by taking into account the following factors: the killing rates of artesunate and mefloquine vary throughout the 48-hour life cycle of the parasite; the distribution of the age of parasites differs across patients; and the patient’s background immunity to infection.

Reference: Dondorp AM, Nosten F, Yi P, Das D, Phyo AP, Tarning J, et al. Artemisinin resistance in Plasmodium falciparum malaria.[see comment]. N Engl J Med. 2009;361(5):455-467.