Significant Efficacy of a Single Low Dose of Primaquine Compared to Stand-Alone Artemisinin Combination Therapy in Reducing Gametocyte Carriage in Cambodian Patients with Uncomplicated Multidrug-Resistant Plasmodium falciparum Malaria

Since 2012, a single low dose of primaquine (SLDPQ; 0.25 mg/kg of body weight) with artemisinin-based combination therapies has been recommended as the first-line treatment of acute uncomplicated Plasmodium falciparum malaria to interrupt its transmission, especially in low-transmission settings of multidrug resistance, including artemisinin resistance. Policy makers in Cambodia have been reluctant to implement this recommendation due to primaquine safety concerns and a lack of data on its efficacy.

aquine (the KEL1/PLA1 lineage) in the eastern Greater Mekong Subregion (GMS) threaten this remarkable global achievement (2)(3)(4)(5)(6)(7). Eliminating rapidly multidrugresistant (MDR) Plasmodium falciparum is therefore the top priority for countries of the GMS. Implementation of novel tools and strategies or repurposing existing tools, like primaquine, that specifically aim at interrupting malaria transmission are essential to reach this goal.
Although artemisinin derivatives are the more potent compounds in artemisininbased combination therapies (ACTs) and are active against early gametocyte stages (8-10) and reduce transmission (11,12), the only currently available drug effective against the transmissible mature stage V gametocytes is primaquine. In 2012, the WHO recommended the addition of a single low dose of primaquine (SLDPQ; target dose, 0.25 mg/kg of body weight) to ACTs to interrupt malaria transmission, primarily in areas with low rates of transmission of MDR P. falciparum, like GMS (13).
To date, several clinical studies conducted in Africa, Colombia, and Cambodia have assessed the safety and efficacy of SLDPQ (14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24), mainly by measuring gametocyte carriage (by microscopy or TaqMan reverse transcriptase PCR [RT-PCR]) as a surrogate marker of its transmission-blocking efficacy. This is due to the challenging logistical requirements of carrying out mosquito infectivity studies. These studies showed that SLDPQ, dosed from 0.20 to 0.75 mg/kg, clearly reduced gametocyte carriage and increased gametocyte clearance in a dose-dependent manner (15-17, 19, 21-25). A small number of studies have investigated human-to-mosquito transmission, based on infectivity measures, with variable results being found due to low infectivity before and/or after SLDPQ treatment (16,17,19,22,24). Only two studies from Mali demonstrated conclusively the transmission-blocking efficacy of SLDPQ, which produced a 92.6% to 100% within-person reduction in infectivity at day 2 versus that at baseline following treatment with 0.25 mg/kg of SLDPQ (16,19).
In Cambodia, glucose-6-phosphate dehydrogenase deficiency (G6PDd) is a common X-linked disorder of the red blood cells, with frequencies ranging from 10.8% to 29.6% in males (26) and with a prevalence rate in malaria patients seen at health centers of 13.9% (27). This is the main reason why Cambodian policy makers have been reluctant to deploy SLDPQ without local data on the safety and the efficacy of SLDPQ. Therefore, we conducted a trial assessing the tolerability and the safety of SLDPQ (0.25 mg/kg) in Cambodia in 2015 and 2016 (28), which followed on from a safety trial of 0.75 mg/kg/ week of primaquine (PQ) in vivax malaria patients (29). Here, we present the results of our investigations on the transmission-blocking efficacy of this treatment through the evaluation of gametocyte prevalence dynamics as well as infectivity measures for a subset of patients using membrane feeding assays with Anopheles minimus mosquitoes, one of the main malaria vectors in Southeast Asia.
Gametocyte dynamics. Examining the dynamics of TaqMan RT-PCR gametocyte carriage to day 28, drug arm, baseline gametocytemia, recrudescence status, the baseline hemoglobin concentration, and day of follow-up were all significant factors for gametocytemia in the univariate analysis ( Table 2).
The DP recipients were more likely to be gametocytemic than the DP-SLDPQ recipients during the follow-up, which was seen more clearly on day 7 and day 14 ( Table 2; Fig. 1), as were patients with baseline gametocytemia ( Table 2; Fig. 2) and recrudescent versus cured patients ( Table 2).
The baseline hemoglobin concentration and the day of follow-up were inversely associated with the probability of being gametocytemic during the follow-up (Table 2).
When the aforementioned factors were tested in a multivariable model, all except drug arm remained significant (Table 3, model 1). However, when removing recrudescence from the multivariable model, drug arm was again a significant explanatory factor (Table 3, model 2).
Post hoc calculation showed that there was a 99.1% power to detect a difference between the DP and DP-SLDPQ treatments with the study sample sizes (n ϭ 49 and n ϭ 56 on day 7 for the DP and DP-SLDPQ treatment groups, respectively).
Mosquito infectivity. A total of 55 patients were included in the direct membrane feeding assays (DMFAs). Among the 292 feeding assays, only 113 of 14,444 (0.78%) dissected mosquitoes were infected (Table 4). Overall, 3 out of 48 (6.3% Ϯ 6.8%) individuals infected at least one mosquito at baseline, one individual (treated with DP) was infectious on day 14, and two individuals with recrudescent infections (Tables 4  and 5) who were repeat DP treatment failures were infectious (i) at presentation on day 124 (DP-SLDPQ) and (ii) day 35 (7 days after DP was started on day 28). The three  individuals infectious on day 0 were all gametocytemic by microscopy, received DP-SLDPQ, and were not infectious on day 1 or day 3 posttreatment (Table 5).

DISCUSSION
In 2012, the WHO recommended adding 0.25-mg/kg SLDPQ to first-line treatments (ACTs) to kill mature gametocytes and block human-to-mosquito transmission (13). As no data were available at that time in Cambodia, we investigated the transmissionblocking efficacy of SLDPQ when given with DP (the recommended first-line standard ACT in 2015) by evaluating gametocyte prevalence dynamics (by TaqMan RT-PCR) and Anopheles minimus infectivity by DMFA.
Our results showed that SLDPQ, when given on the first day of the 3-day standard ACT treatment, significantly decreased gametocyte carriage over time, starting on day 3, and that gametocyte carriage was the lowest on day 7, resulting in a 5.59-fold reduced risk relative to that with treatment with DP alone over 28 days in the univariate analysis. This finding is consistent with recent reports on the efficacy of PQ in reducing gametocyte carriage (15-17, 19, 21-25, 31) and earlier reports showing a lag between reduced gametocyte carriage and rapid mosquito infectivity. The multivariable analysis revealed the significant effect of recrudescent (resistant) infections on gametocyte carriage, with a 3-fold reduction in gametocyte carriage in the absence of resistance being seen.
Our patients had a high rate of baseline gametocytemia (19% by microscopy, just under 45% by TaqMan RT-PCR), which is similar to that reported previously in 2010 in patients with artemisinin-resistant P. falciparum infections in western Cambodia (ϳ19%) but which is 3-fold higher than the 6% reported earlier in Ratanakiri Province,  Cambodia (32). Given the high prevalence of PfKelch13 mutant P. falciparum parasites (63%) in our study and the very high recrudescent rate (~28%), our high gametocyte carriage rate is consistent with the presence of artemisinin-resistant P. falciparum, which has also led to piperaquine resistance (7). This is consistent with another study in Cambodia which showed that DP-treated patients from Oddar Meanchey province (northern Cambodia) also had prolonged gametocyte carriage that was seen in patients with recrudescent infections but was independent of the slow asexual parasite clearance time due to artemisinin-resistant P. falciparum (22,35). However, prolonged gametocyte carriage has also been seen in DP-treated African patients without artemisinin-resistant P. falciparum infection who, nevertheless, had low rates of human-to-mosquito transmission (16,32,33). Moreover, several comparative studies have found that gametocyte carriage is higher in DP-treated patients than in artemether-lumefantrine-treated patients (33)(34)(35)(36), suggesting a reduced gametocytocidal activity of piperaquine compared to lumefantrine.
Overall, we observed a biphasic pattern of gametocyte prevalence, with a drop on  a DP, dihydroartemisinin piperaquine; SLDPQ, a single low dose of primaquine; Rec1 and Rec2, the first and second recrudescences, respectively; NA, data not available. b Positive isolates were isolates which were able to infect at least one mosquito from a direct membrane feeding assay (DFMA). day 7 and a rebound on day 14, followed by a decrease on day 21 and day 28. However, this effect was mainly driven by the DP-treated patients, as their gametocyte carriage rate was similar over the 28-day follow-up, whereas it decreased markedly in the DP-SLDPQ-treated patients (Fig. 1). We were not able to assess the transmission-blocking efficacy of SLDPQ due to the very low infectivity rate observed in only three patients (ϳ6%) at baseline and three infectious events during follow-up. This is a major limitation of DFMAs, as previously reported by Lin et al. (22) at baseline, using Anopheles dirus mosquitoes, and on day 7, without assessing the baseline, by Okebe et al. (24). In another study in African children, despite a higher infectivity rate at enrollment after microscopy-based selection of gametocyte-positive asymptomatic children (38%), Gonçalves et al. (17) were also unable to detect the added value of SLDPQ due to very low infectivity posttreatments in all arms. As our clinical trial aimed primarily at assessing SLDPQ tolerability and safety, we did not particularly select patent or high-density gametocyte carriers and thus would have needed substantially more individuals to assess transmission-blocking efficacy. Only two studies, which were conducted in Mali, have statistically confirmed the high anti-infectivity efficacy (Ͼ90%) at 48 h of SLDPQ, when dosed at baseline, after selecting gametocyte-positive participants by microscopy (16,19). However, we observed that the three SLDPQ-treated infectious patients on day 0 did not infect mosquitoes on subsequent days, despite the presence of gametocytes. Several reasons may explain this finding. Following primaquine treatment, the anti-infectivity effect is rapid (Յ24 h) and precedes the decline in gametocytemia (37). Male gametocytes seem to be more sensitive to primaquine than female gametocytes (38), and the P. falciparum sex ratio is biased toward females, potentially resulting in delayed gametocyte clearance. The relationship between gametocytemia and infectivity is nonlinear and is, thus, an indirect measure of infectivity (39)(40)(41). To conclude, our study shows that the addition of SLDPQ to ACT treatment for symptomatic falciparum malaria decreases substantially gametocyte carriage in patients with MDR P. falciparum infections. Although we were unable to reconfirm the findings of the elegant studies of Dicko et al. (16,19), SLDPQ, dosed on the first day of treatment, is likely to further decrease the transmission potential of P. falciparum patients in our setting when combined with an effective ACT.
Owing to the high failure rate of DP, Cambodia has switched to artesunatemefloquine treatment as the first-line treatment for uncomplicated P. falciparum and in January 2018 deployed SLDPQ in the whole country, partly on the basis of the findings of the trial described here and the good tolerability of SLDPQ in G6PDd patients (28). An age-based dosing regimen of SLDPQ, designed for Cambodia, will aid elimination efforts in areas where weighing scales are unavailable (42).

Study design.
This open-label randomized controlled trial assessing the tolerability and the safety of SLDPQ was carried out in Banlung, Rattanakiri Province, in northeastern Cambodia in 2015 and 2016 (ClinicalTrials.gov registration number NCT02434952), as described by Dysoley et al. (28). Briefly, 109 nonpregnant, non-breast-feeding patients aged Ն1 year with acute uncomplicated falciparum malaria (Ն1 asexual parasites/500 white blood cells, equivalent to 16 asexual parasites/l) and a hemoglobin concentration of Ն6 g/dl were recruited to receive either dihydroartemisinin-piperaquine (DP; dihydroartemisinin at 40 mg and piperaquine at 320 mg; Duo-Cotecxin; Zhejiang Holley Nanhu Pharmaceutical Co. Ltd., Jiaxing, Zhejiang, China) alone or combined with SLDPQ (0.25 mg/kg given with the first dose of DP, 15 mg of primaquine base; Thai Government Pharmaceutical Organization). Clinical and laboratory assessments were performed on days 0, 1, 2, 3, 7, 14, 21, and 28. Anopheles minimus mosquitoes were fed on blood samples collected from a subset of 55 patients and on the days of recurrent (Drec) falciparum parasitemia (defined by the WHO as late treatment failures), depending on mosquito availability, using direct membrane feeding assays (DMFAs).
The sample size calculation was based on the primary outcome of a reduction in the mean day 7 hemoglobin (Hb) concentration of 1 g/dl in the DP-SLDPQ arm with G6PDd patients versus that in the DP-SLDPQ arm with patients with normal G6PD levels. Assuming a mean Ϯ standard deviation Hb concentration of 11.27 Ϯ 1.74 g/dl (calculated from a Southeast Asian database of~6,800 ACT-treated, P. falciparum-infected patients of all ages [W. R. Taylor, unpublished data]), a two-sided alpha value of 0.05, and a power of 80%, the sample size was 48 patients per arm, and this number was rounded up to 50. G6PD status was initially diagnosed in the field using a fluorescent spot test to allow for the allocation of SLDPQ. The results of the qualitative Carestart (AccessBio, Somerset, NJ) rapid diagnostic test were assessed in parallel.
Mosquito infection. DMFAs were carried out to assess individual malaria infectivity as described previously (30). Briefly, 3-to 5-day-old A. minimus female mosquitoes were fed through membranes on the patients' blood. The mosquitoes were starved for 24 h before being provided a blood meal. Venous blood samples were collected in heparinized tubes, and 400 l of blood was made available in membrane feeders maintained at 37°C. Females were fed only once on freshly drawn blood. Postfeeding, unfed females were discarded and fed females were kept in cages (20 by 20 by 20 cm) with constant access to a 10% sucrose solution. Patient infectivity was determined by assessing infection prevalence (i.e., the proportion of infected females) and infection intensity (i.e., the number of Plasmodium oocysts in infected females). Midguts were dissected in a 1% mercurochrome stain, and the presence and the number of oocysts were determined under a microscope (ϫ20 magnification). Dissections were performed on 7 day after the blood meal. The mean number of dissected females was 49 (range, 15 to 85; median, 50).
Biological investigations. Parasite RNA was extracted from TRIzol reagent (Life Technologies Holdings Pte. Ltd., Singapore)-conserved whole-blood samples using a QIAamp RNA blood minikit (Qiagen, Germany), following the protocol recommended by the manufacturer. A two-step semiquantitative real-time PCR was performed to detect malaria parasites, as previously described (43). Following PCR amplification, P. falciparum-positive samples were analyzed for the presence of gametocytes by a TaqMan RT-PCR, using primers spanning an exon-exon junction and targeting the Plasmodium falciparum meiotic recombination protein DMC1-like protein gene (GenBank accession number AF356553), as described previously (44). Gametocyte dilution series based on an in vitro-cultured local strain were used to estimate gametocyte blood concentrations.
Ethical statement. Ethical approvals were obtained from the Cambodian National Ethics Committee for Health Research (approval number 0370NECHR), and the trial is registered at ClinicalTrials.gov (ClinicalTrials.gov registration number NCT02434952). The protocols conformed to the Helsinki Declaration on ethical principles for medical research involving human subjects (version 2002), and informed written consent was obtained from all volunteers.
Statistical analyses. Initially, univariate generalized linear mixed models (GLMMs) with a binomial distribution were fitted to model the PCR-measured gametocyte prevalence rates. In these models, follow-up days (days 1, 2, 3, 7, 14, 21, and 28), treatment (DP versus DP-SLDPQ), the presence of gametocytes prior to treatment on day 0, recrudescence status (recrudescent versus cured patient), and the hemoglobin concentration prior to treatment on day 0 were included in the model as fixed factors.