An optimised age-based dosing regimen for single low-dose primaquine for blocking malaria transmission in Cambodia

This is the first, optimised, age-based regimen for SLDPQ and was developed using a substantial database of some 28,000 Cambodians across the age, sex, urban, and rural spectra. It has four dosing bands that could be adopted by the Centre National de Malariologie and, if suitable, adopted also by other malaria control programmes in the GMS as part of their urgent efforts to control and eliminate ARPf.

The data were obtained from several sources and included data from the two-stage, cluster-randomised nationwide DHS survey, dog bite victims (likely also to be a random sample in Cambodia), village surveys in malaria-endemic areas (some using PCR), and malaria patients. The data from patients with suspected CNS infections are likely to be the most biased samples. However, all of the assembled data overlapped substantially and the difference in weight between urban and rural dwellers was small and not significant.

Symptomatic malaria is a rural disease in Cambodia and affects mostly young adult males who work in forests. However, there are also pools of asymptomatic, submicroscopic falciparum malaria cases (detectable by PCR) affecting all ages and both sexes across many rural regions of Cambodia [82–85]. Our weight-for-age database reflected that epidemiology, containing approximately 80 % rural dwellers, malaria patients, and asymptomatic parasite-infected individuals. Thus, our dosing regimen is applicable to malaria patients and, if MDA becomes policy, asymptomatic malaria parasite carriers as well.

Although our data are unsuitable as reference data to construct a weight-for-age “growth chart” for Cambodia as a whole, we did construct “growth curves” using Box Cox modelling, which enabled us to calculate the mg base/kg dose of PQ that individuals would receive. Our weight-for-age curve is similar to those of Cambodia, Laos, and Myanmar constructed by Hayes et al. [22]. These countries are also affected by ARPf and share similar G6PDd variants with Cambodia [86]. Thus, our SLDPQ regimen is generalisable to two of Cambodia’s GMS neighbours.

There are some limitations to our study. Children aged??5 years and women of child-bearing age predominated because most of these data came from the DHS survey. Previous databases have suffered from having fewer individuals aged 5–15 years [21] and anthropometric data from several of the databases used by Hayes et al. [22] were also somewhat patchy. Our database had relatively fewer data in the 5–15 years age group, with individual ages numbering between 200 and 300. In common with other malaria-endemic countries, individuals often do not know their exact dates of birth but are usually able to estimate their ages reasonably well. This is more important for children and help with age determination can be obtained from “Road to Health” charts and, in some countries, national ID cards.

Although we believe our SLDPQ regimen is accurate based on current knowledge of PQ’s PK and PD characteristics, new knowledge could result in its revision. Several significant knowledge gaps exist, notably, the lack of PK OM data, dose-response curves for mosquito infectivity and red cell haemolysis, and important PK parameters such as the maximum concentration or area under the concentration-time curve. There are no PQ PK data in children aged??5 years. Yet, these data are crucial to establish whether the target PQ dose of 0.25 mg base/kg applies to this vulnerable group.

Our knowledge of PQ drug interactions is also limited. Increased awareness of drugs that could potentiate the haemolytic toxicity of SLDPQ is especially important either by a metabolic interaction at the CYP level or by direct red blood cell toxicity. Once SLDPQ is deployed, good pharmacovigilance systems will need to be in place so confidence in SLDPQ by the populace remains high. This is especially so when SLDPQ is used as MDA because healthy individuals will take a drug with the intention of a community rather than individual benefit, but which could result in toxicity for individuals. Moreover, accidental exposure to PQ in pregnancy can be expected and is another critical area of pharmacovigilance.

The effect over time of the widespread use of SLDPQ on gametocyte sensitivity and P. vivax liver hypnozoites is unknown. This would need to be studied to enable malaria control programmes to adapt their strategies if PQ resistance were to develop.

Currently, in Cambodia, approximately 30 % of all malaria treatments are given by village malaria workers. They will need to be trained on how to administer SLDPQ, what advice to give to recipients regarding PQ toxicity, and how to manage visibly malnourished or overweight children. As countries move towards malaria elimination and adopt MDA with SLDPQ, it is clear that the efficient dosing of large numbers of individuals is best achieved using an age-based regimen, much in the same way as anthelminthic MDA is done. Coordination between different MDA programmes would be essential.

Few manufacturers produce PQ to international Good Manufacturing Practice. Current tablet strengths are limited to 7.5 and 15 mg of PQ base, which is quite inadequate. Moreover, there are no PQ formulations for children and PQ has a very bitter taste. Options to increase palatability include sweetened dispersible tablets, as has been developed for AL [87], granules, mini tablets, and gel coating [88, 89]. These are issues of great importance that could substantially affect access to PQ and the success of malaria elimination. If regional- or country-specific, age-based dosing regimens could be designed that use a limited number of the same tablet strengths, this would be a strong argument in convincing pharmaceutical companies to engage positively in malaria elimination.

South American populations weigh more than their African counterparts, who in turn weigh more than South East Asian/Western Pacific populations [22]. Therefore, our SLDPQ regimen cannot be generalised to South America and Africa. Caution is also needed for the Middle East and parts of west Asia where anthropometric characteristics need further study and where the most severe G6PDd variant – Mediterranean – is not uncommon [33].

Our SLDPQ regimen has four age-dosing groups and we have proposed a weight-based regimen with four (five if very heavy individuals are separated off) weight bands. Neither regimen fits in neatly with commonly used ACTs. Artesunate-mefloquine, artesunate-amodiaquine (not used currently in Cambodia), and DHAPP have four, four, and six age-based dosing regimens, respectively, and AL and artesunate-pyronaridine both have four weight-based regimens. This underlies the need for good training when the age-based PQ regimen is eventually deployed [20].