Incorporating Stage Specific Drug Action into Pharmacological Modelling of Antimalarial Drug Treatment

Pharmacological modelling of anti-parasitic treatment based on a drug's pharmacokinetic and pharmacodynamic properties plays an increasingly important role in identifying optimal drug dosing regimens and predicting their potential impact in control and elimination programmes. Conventional modelling of treatment relies on methods that do not distinguish between parasites being in different developmental stages. This is problematic for malaria parasites as their sensitivity to drugs varies substantially during their 48-hour developmental cycle. We investigated four drug types (short/long half-lives with/without stage specific killing) to quantify the accuracy of the standard methodology. The treatment dynamics of three drug types were well characterised with standard modelling. The exception were short half-life drugs with stage specific killing (i.e. artemisinins) because, depending on time of treatment, parasites might be in highly drug-sensitive stages or in much less sensitive stages. We describe how to bring such drugs into pharmacological modelling by including additional variation into the drugs maximal killing rate. Finally, we show that artemisinin kill rates may have been substantially over-estimated in previous modelling studies because (i) the parasite reduction ratio (PRR) (generally estimated as 104) is based on observed changes in circulating parasite number which generally over-estimates the ‘true' PRR which should include both circulating and sequestered parasites, and (ii) the third dose of artemisinin at 48 hours targets exactly those stages initially hit at time zero, so it is incorrect to extrapolate the PRR measured over 48 hours to predict the impact of doses at times 48 hours and later.