Pharmacometric modelling of antimalarial drugs in development

Aims: Determination of minimum inhibitory concentration (MIC) of antimalarial drugs in the clinical setting may provide a valuable alternative to empirical approaches to effective dose-finding. The aim was to estimate prospectively the in vivo MIC of cipargamin1 in adults with uncomplicated P. falciparum malaria in an open-label, dose-ranging Phase IIa study.

Methods: Twenty-five Vietnamese adults with uncomplicated P. falciparum mono-infection confirmed by microscopy were allocated sequentially to treatment with a single sub-therapeutic dose of cipargamin (30, 20, 10, or 15 mg). The in vivo MIC of cipargamin was estimated using population pharmacokinetic-pharmacodynamic modelling of observed plasma cipargamin measurements and serial parasite densities assessed by microscopy and real-time quantitative polymerase chain reaction. Modelling and simulation were performed in NONMEM v7.3.

Results: The pharmacokinetic properties of cipargamin were well described by a flexible transit-absorption model followed by a one-compartment disposition model, resulting in a high predictive performance. Individual pharmacokinetic parameter estimates from the final model were imputed into the pharmacokinetic-pharmacodynamic model in order to evaluate the cipargamin-dependent parasite killing effect. No parasite growth data were available before drug administration and the parasite multiplication rate was therefore fixed to a 10-fold multiplication per parasite life-cycle (i.e. 48 hours). The initial implementation of the pharmacokinetic-pharmacodynamic model assumed a homogenous parasite population and drug-dependent killing of parasites (EMAX-model). However, population and individual parasite clearance curves showed a clear biphasic pattern of parasitaemia decline. This could suggest presence of a non-sensitive (i.e. "dormant") parasite population that is unaffected by the drug. The fraction of sensitive/dormant parasites was estimated but allowed for inter-individual variability in the same parameter. The activation ("awakening") of dormant parasites was also estimated. Higher doses, and thus higher plasma cipargamin concentrations, were associated with a significantly faster maximum parasite killing. A total of 23 out of 25 patients were characterised accurately as cured, early treatment failure, or recrudescent infections using the final model. The resulting median (IQR) MIC was 0.126 ng/mL (0.0786–0.273 ng/mL), for patients correctly predicted as recrudescent (n=12). Time to MIC is a function of dose since a lower dose, and subsequently lower drug concentrations, will reach a putative MIC value faster than higher doses, assuming similar drug half-lives and parasite characteristics. Cured patients, correctly predicted as cured (n=7), showed a maximum median (IQR) MIC value of 0·236 ng/mL (0·0803–1·47 ng/mL), defined as the cipargamin concentration when predicted parasitemia fell below a total of 10 parasites.

Conclusion: The developed pharmacokinetic-pharmacodynamic model described the relationship between cipargamin exposure and drug-dependent parasite killing succesfully. Sub-therapeutic administration of drugs in development followed by a pharmacometric analysis demonstrated a highly informative approach for determining the in vivo MIC in patients. This provides a rational framework for dose-finding in antimalarial drug development.

References: 1. Rottmann M, McNamara C, Yeung BK, et al. Spiroindolones, a potent compound class for the treatment of malaria. Science 2010;329:1175-80.