Publications

Objective: To calculate, using the Stewart approach to acid-base disorders, the strong anion gap as an estimate for the contribution of unmeasured plasma anions other than lactate to the metabolic acidosis that characterizes severe falciparum malaria and to assess its relative prognostic significance. Design: Cohort study. Setting: The intensive care unit of an infectious diseases hospital in southern Vietnam. Patients: Consecutive adult patients (n = 268) with severe falciparum malaria. Interventions: The intervention was clinical management in a dedicated unit. We measured baseline venous lactate, electrolytes, biochemical variables, admission arterial blood pH, and gas tensions for calculation of the strong anion gap. Measurements and Main Results: The mean (95% confidence interval) admission strong anion gap was 11.1 (10.4–11.9) mEq/L, compared with lactate (geometric mean, 95% confidence interval) at 2.9 (2.7–3.2) mmol/L. Strong anion gap had a high predictive value for mortality (area under the receiver operating characteristic curve 0.73 (95% confidence interval, 0.65–0.82), which was independent of plasma lactate and creatinine concentrations. Renal failure and hepatic dysfunction were both associated with, but were not the sole determinants of, high levels of strong anion gap. Conclusions: In severe malaria, unidentified anions other than lactate are the most important contributors to metabolic acidosis, a major cause of death. The strong anion gap is a powerful prognostic indicator in patients with severe malaria.

As part of a targeted malaria elimination project, mass drug administrations (MDAs) were conducted in Vietnam. The impact of MDAs on malaria transmission depends largely on the efficacy of the anti-malarial drug regimen, the malaria epidemiology in the site and the population coverage. To explore why some people participate in MDAs and others do not, a quantitative survey of the villagers' perceptions was undertaken in Vietnam. In 2013/2014 MDAs were conducted in a village in Binh Phuoc province and a village in Ninh Thuan province. Within three months of the drug administration, 59 respondents in a village in Binh Phuoc and 79 respondents in a village in Ninh Thuan were randomly selected and interviewed. Comprehension of the purpose of the intervention was of paramount importance for participation in the intervention. Respondents aware that the intervention aims to protect against malaria were significantly more likely to participate than respondents who were unaware of the MDA's purpose. Secondly, how and by whom villagers were informed was critical for participation. There was a strong association between sensitization by an informant such as a member of the local health team with participation in the intervention. The study suggests several approaches to increase participation in mass drug administration campaigns. Training trustworthy informants to sensitize the study population is critical to maximize village participation in this setting. To achieve high coverage the entire community must understand and agree with the intervention.

Among human parasitic diseases, malaria remains the world's most important cause of morbidity and mortality. The impact of malaria on health and on development is huge and, because of the spread of drug resistance in P. falciparum, it is worsening. In all endemic regions, children and pregnant women are the two groups mostly at risk of severe disease and death. The true magnitude of the burden of maternal malaria is unknown, but recent reports indicate that it may be much higher than previously thought. For example recently, falciparum malaria was identified as the leading cause of maternal death in a hospital study in Mozambique (Granja et al. 2001) and in a population in Zimbabwe (Rutgers 2001). Falciparum malaria is particularly lethal in pregnant women during epidemics. Through the adverse effects of malaria on birthweight and prematurity, maternal infection also results in excess infant mortality (Luxemburger et al. 2001; Steketee et al. 2001). In areas of high stable transmission, the most visible impact of malaria on maternal health is severe anaemia, an important contributor to maternal mortality (Brabin et al. 2001). Prevention strategies include vector control (impregnated bed nets) and chemoprophylaxis (where still effective) or intermittent preventive treatment (IPT). In areas of low or unstable transmission, severe disease is more common and early detection and prompt treatment are essential to reduce mortality. In all situations antimalarial drugs play the central role. It is recommended that pregnant women living in malaria-endemic areas receive antimalarial chemoprophyl-axis – but with what drug? Chloroquine, still the most widely recommended and used antimalarial for pregnant women in Africa, has become ineffective against P. falciparum in most endemic countries. There is no universally effective drug, with a similarly good safety profile in pregnancy, to replace it. This is partly the result of the systematic exclusion of pregnant women from any drug development programme, because of the fear of potential adverse effects on the unborn child. In a world increasingly driven by economic forces, the spectre of liability looms over the pharmaceutical industry and also over international institutions. The debate on antimalarials for pregnant women has shifted from 'Which drug has the most favourable risk-benefit ratio for the mother?' to 'Who is responsible if an unwanted side effect of the drug on the foetus is discovered?' In endemic areas of Africa, most of the pregnant women remain unprotected against malaria. Those who are treated, often receive ineffective drugs. These are difficult times for public health authorities in many malaria-affected countries, as they come under increasing pressure to change antimalarial drug policies. In this issue of TMIH, Newman and colleagues present a comprehensive analysis of available antimalarials, with the aim of helping maternal malaria prevention programmes in sub-Saharan Africa. Importantly, the authors take into account not only the efficacy of various drug regimens (using the data on children as a proxy of the efficacy in pregnancy, because there are usually no data for pregnant women) but also their 'effectiveness': a combination of efficacy with availability, acceptability, affordability, and deliverability. It was concluded that, the best possible option to date is to use intermittent presumptive therapy (IPT) with sulfadoxine-pyrimethamine (SP). Five studies have been conducted, all in high transmission settings in East Africa (two in Malawi and three in Kenya). These have shown that two treatment doses of SP given to HIV-negative mothers during pregnancy, can reduce maternal anaemia (significant in three of the four studies in which this was evaluated) and low birth weight (two of four studies) (Schultz et al. 1994; Parise et al. 1998; Verhoeff et al. 1998; Shulman et al. 1999). Three doses are needed for HIV-positive mothers. On the basis of these results, IPT is now recommended for pregnant women by WHO (2000) and more recently, has been proposed for infants and children (Pasha et al. 2003). But resistance to SP has taken hold in East Africa, where 28-day failure rates range between 30 and 60% and it is spreading rapidly across the continent; for example SP failed to cure more than 50% of the treated children in a recent study from Liberia (Checchi et al. 2002). Newman et al. describe the therapeutic efficacy of SP as being 'fair to good' in most of Africa, but this may be over-optimistic. Since the four pivotal studies on the use of SP-based IPT were conducted in Kenya and Malawi 5 years ago, resistance to SP has worsened significantly in East Africa. Although, therapeutic responses in mothers may be better than in children, because of their greater immunity, SP resistance is likely to have compromised the undoubted benefits seen in these trials already. Mutations conferring resistance to sulfadoxine and pyrimethamine are spreading fast, encouraged by the increased gametocyte production observed in patients treated with SP. There is no doubt we are losing SP; the only question is 'How fast?' If adopted widely, the recommendation of Newman et al. to use SP may soon have to be changed because of resistance. But implementing changes in drug policies is a lengthy and painful process, and it is notoriously difficult to achieve. The deployment of SP across Africa is part of a wider debate, on the use of antimalarials. The main WHO–RBM contributors (USAID and DFID) have been encouraging the use of SP across Africa for all cases of malaria (not just in pregnancy) as a temporary response to the final demise of chloroquine and the resulting serious increase in morbidity and mortality. This approach is based – as in the paper of Newman et al. – on the claim that SP has the best 'efficacy-effectiveness' score. There are serious objections to this, notably from the African malaria control programmes themselves and from WHO and malaria researchers (McNeil 2002). As emphasized rightly by Newman and colleagues, the 'practice' of treating patients with ineffective drugs must be abandoned. Efficacy is essential for effectiveness. Combination therapies containing artemisinin derivatives are the most efficacious antimalarial treatments (WHO 2001). In all trials in Africa, adding artesunate to SP has resulted in consistently faster clinical recovery and higher parasitological cure rates. But pregnant women have been excluded from systematic studies because of concerns over the safety of artemisinin derivatives in pregnancy. The paradox is that SP is not free of toxicity to pregnant animals (Phillips-Howard & Wood 1996; Uche-Nwachi 1998) and fewer than 2000 exposed pregnancies have been documented properly, although SP is certainly widely used and is generally regarded as safe. On the other hand, the artemisinin derivatives are being denied to pregnant women, in programmatic terms, despite no evidence of adverse effects in more than 600 well-documented exposures (McGready et al. 2001). The artemisinin derivatives have been known for many years to induce foetal resorption in early animal pregnancy, and so does SP. A recent WHO consensus meeting concluded after detailed review of all available evidence that artemisinins could be used in the second and third trimesters but, pending further information, should not be used in the first trimester (WHO 2002). Malaria control programmes urgently need to know if artemisinin combinations (ACT) can replace the failing SP for IPT in pregnancy. We strongly support the plea of Newman and colleagues for urgent evaluation of ACTs in pregnancy. If we are to achieve the Abuja declaration objectives of providing pregnant women with effective malaria prevention and treatment by the year 2005, action is needed. New strategies of prevention and treatment (and the means to support them) must be developed and must be adapted to the various epidemiological situations within and outside Africa. Further data on the safety and efficacy of ACT treatment must be obtained urgently in pregnant women. Large community studies of their use in IPT should start in high transmission areas where SP resistance is established. Large-scale deployment of ACTs should be accompanied by a surveillance system, to monitor their positive and any negative impact in pregnancy. There is a pressing need to evaluate the true burden of maternal malaria: to improve detection, prevention and treatment in high transmission settings, to revisit the dogma that pregnant women in hyperendemic malarious areas remain asymptomatic, and to try and measure the malaria attributable direct and indirect maternal mortality. ACTs require administration over 3 days for the treatment of malaria. Efforts will be needed to improve prescribing practices and packaging to improve adherence, particularly if 3-day regimines are used as IPTs in otherwise healthy women. More information is needed on the relationship between levels of resistance, and effects of IPT on maternal anaemia and birth weight, and also the relative effectiveness of insecticide-impregnated bed nets and IPT, and the interaction between the two, at differing levels of transmission intensity. Only then can a risk-benefit analysis for the mother be performed, based on evidence. This will be achievable if the 'partnership' between the governments, the pharmaceutical industry, international institutions and the research community is more than a trendy, but often empty word. In sub-Saharan Africa and elsewhere, maternal malaria is a potentially life-threatening disease. Pregnant women and their babies need protection by truly effective antimalarial drugs. It is certainly time for action, and time is running out. Authors Dr François Nosten (corresponding author) and Dr Rose McGready Shoklo Malaria Research Unit, P.O. Box 46, 63110 Mae Sot, Thailand. E-mail: shoklo@cscoms.comProf. Sornchai Looareesuwan and Prof. Nicholas J. White Faculty of Tropical Medicine, Mahidol University, 121–126 Rajvithi Road, 10400 Bangkok, Thailand E-mail: fnnjw@diamond.mahidol.ac.th

Background Molecular genotyping in Plasmodium serves many aims including providing tools for studying parasite population genetics and distinguishing recrudescence from reinfection. Microsatellite typing, insertion-deletion (INDEL) and single nucleotide polymorphisms is used for genotyping, but only limited information is available for P. malariae, an important human malaria species. This study aimed to provide a set of genetic markers to facilitate the study of P. malariae population genetics. Methods Markers for microsatellite genotyping and merozoite surface protein 1 (pmmsp1) gene polymorphisms were developed and validated in symptomatic P. malariae field isolates from Myanmar (N=37). Fragment analysis was used to determine allele sizes at each locus to calculate multiplicity of infections (MOI), linkage disequilibrium, genetic richness index, heterozygosity and construct dendrograms. Nucleotide diversity (π), number of haplotypes, and genetic diversity (Hd) were assessed and a phylogenetic tree was constructed. Genome-wide microsatellite maps with annotated regions of newly identified markers were constructed.Results Six microsatellite markers were developed and tested in 37 P. malariae isolates which showed sufficient heterozygosity (0.530-0.922), genetic richness index (0.050-0.250) and absence of linkage disequilibrium (IAS = 0.03, p-value > 0.05) (N=37). In addition, a tandem repeat (VNTR)-based pmmsp1 INDEL polymorphisms marker was developed and assessed in 27 P. malariae isolates showing a nucleotide diversity of 0.092, haplotype gene diversity of 0.698 and identified 14 unique variants. The size of VNTR consensus repeat unit adopted as allele was 27 base pairs. The markers Pm12_426 and pmmsp1 showed greatest diversity with heterozygosity scores of 0.920 and 0.835, respectively. Using six microsatellites markers, the likelihood that any two parasite strains would have the same microsatellite genotypes was 8.46 × 10-4 and was further reduced to 1.66 × 10-4 when pmmsp1 polymorphisms were included.Conclusions Six novel microsatellites genotyping markers and a set of pmmsp1 VNTR-based INDEL polymorphisms markers for P. malariae were developed and validated. Each marker could be independently or in combination employed to access genotyping of the parasite. The newly developed markers may serve as a useful tool for investigating parasite diversity, population genetics, molecular epidemiology and for distinguishing recrudescence from reinfection in drug efficacy studies.

Dihydroartemisinin-piperaquine has shown excellent efficacy and tolerability in malaria treatment. However, concerns have been raised of potentially harmful cardiotoxic effects associated with piperaquine. The population pharmacokinetics and cardiac effects of piperaquine were evaluated in 1,000 patients, mostly children enrolled in a multicenter trial from 10 sites in Africa. A linear relationship described the QTc-prolonging effect of piperaquine, estimating a 5.90-ms mean QTc prolongation per 100-ng/ml increase in piperaquine concentration. The effect of piperaquine on absolute QTc interval estimated a mean maximum QTc interval of 456 ms (50% effective concentration of 209 ng/ml). Simulations from the pharmacokinetic-pharmacodynamic models predicted 1.98 to 2.46% risk of having QTc prolongation of >60 ms in all treatment settings. Although piperaquine administration resulted in QTc prolongation, no cardiovascular adverse events were found in these patients. Thus, the use of dihydroartemisinin-piperaquine should not be limited by this concern. (This study has been registered at ClinicalTrials.gov under identifier NCT02199951.).

Limited microbiology services impede adequate diagnosis and treatment of common infections such as pyelonephritis in resource-limited settings. Febrile pregnant women attending antenatal clinics at Shoklo Malaria Research Unit were offered urine dipstick, sediment microscopy, urine culture, and a 5-mL blood culture. The incidence of pyelonephritis was 11/1,000 deliveries (N = 53 in 4,819 pregnancies) between January 7, 2004 and May 17, 2006. Pyelonephritis accounted for 20.2% (41/203) of fever cases in pregnancy. Escherichia coli was the most commonly isolated pathogen: 87.5% (28/32) of organisms cultured. Susceptibility of E. coli to ampicillin (14%), cotrimoxazole (21%), and amoxicillin-clavulanic acid (48%) was very low. E. coli was susceptible to ceftriaxone and ciprofloxacin. The rate of extended spectrum β-lactamase (4.2%; 95% confidence interval = 0.7-19.5) was low. The rate and causes of pyelonephritis in pregnant refugee and migrant women were comparable with those described in developed countries. Diagnostic innovation in microbiology that permits affordable access is a high priority for resource-poor settings.

Objectives: To assess impact of serial lumbar punctures on association between cerebrospinal fluid (CSF) opening pressure and prognosis in HIV-associated cryptococcal meningitis; to explore time course and relationship of opening pressure with neurological findings, CSF fungal burden, immune response, and CD4 cell count. Design: Evaluation of 163 HIV-positive ART-naive patients enrolled in three trials of amphotericin B-based therapy for cryptococcal meningitis in Thailand and South Africa. Methods: Study protocols required four lumbar punctures with measurements of opening pressure over the first 2 weeks of treatment and additional lumbar punctures if opening pressure raised. Fungal burden and clearance, CSF immune parameters, CD4 cell count, neurological symptoms and signs, and outcome at 2 and 10 weeks were compared between groups categorized by opening pressure at cryptococcal meningitis diagnosis. Results: Patients with higher baseline fungal burden had higher baseline opening pressure. High fungal burden appeared necessary but not sufficient for development of high pressure. Baseline opening pressure was not associated with CD4 cell count, CSF pro-inflammatory cytokines, or altered mental status. Day 14 opening pressure was associated with day 14 fungal burden. Overall mortality was 12% (20/162) at 2 weeks and 26% (42/160) at 10 weeks, with no significant differences between opening pressure groups. Conclusion: Studies are needed to define factors, in addition to fungal burden, associated with raised opening pressure. Aggressive management of raised opening pressure through repeated CSF drainage appeared to prevent any adverse impact of raised opening pressure on outcome in patients with cryptococcal meningitis. The results support increasing access to manometers in resource-poor settings and routine management of opening pressure in patients with cryptococcal meningitis.

Malaria is endemic in the southern plain of Nepal which shares a porous border with India. More than 80% cases of malaria in Nepal are caused by Plasmodium vivax . The main objective of this study was to review the epidemiology of P. vivax malaria infections as recorded by the national malaria control program of Nepal between 1963 and 2016. National malaria data were retrieved from the National Malaria program in the Ministry of Health, Government of Nepal. The epidemiological trends and malariometric indicators were analyzed. Vivax malaria has predominated over falciparum malaria in the past 53 years, with P. vivax malaria comprising 70–95% of the annual malaria infections. In 1985, a malaria epidemic occurred with 42,321 cases (82% P. vivax and 17% Plasmodium falciparum ). Nepal had experienced further outbreaks of malaria in 1991 and 2002. Plasmodium falciparum cases increased from 2005 to 2010 but since then declined. Analyzing the overall trend between 2002 (12,786 cases) until 2016 (1,009 cases) shows a case reduction by 92%. The proportion of imported malaria cases has increased from 18% of cases in 2001 to 50% in 2016. The current trends of malariometric indices indicate that Nepal is making a significant progress toward achieving the goal of malaria elimination by 2025. Most of the cases are caused by P. vivax with imported malaria comprising an increasing proportion of cases. The malaria control program in Nepal needs to counter importation of malaria at high risk areas with collaborative cross border malaria control activities.

The clinical response to ceftriaxone in patients with typhoid fever is significantly slower than with ofloxacin, despite infection with Salmonella enterica serovar Typhi (S. Typhi) isolates with similar susceptibilities (MIC 0.03-0.12 mg/L). The response to ofloxacin is slower if the isolate has intermediate susceptibility (MIC 0.25-1.0 mg/L).To determine the bactericidal activity and post-antibiotic effect (PAE) of ceftriaxone and ofloxacin against S. Typhi.The mean time to reach a 99.9% reduction in log10 count (bactericidal activity) and PAE of ceftriaxone and ofloxacin were determined for 18 clinical isolates of S. Typhi in time-kill experiments (MIC range for ofloxacin 0.06-1.0 mg/L and for ceftriaxone 0.03-0.12 mg/L).The mean (SD) bactericidal activity of ofloxacin was 33.1 (15.2) min and 384.4 (60) min for ceftriaxone. After a 30 min exposure to ofloxacin, the mean (SD) duration of PAE was 154.7 (52.6) min. There was no detectable PAE after 1 h of exposure to ceftriaxone. For ofloxacin, bactericidal activity and PAE did not significantly differ between isolates with full or intermediate susceptibility provided ofloxacin concentrations were maintained at 4×MIC.Infections with S. Typhi with intermediate ofloxacin susceptibility may respond to doses that maintain ofloxacin concentrations at 4×MIC at the site of infection. The slow bactericidal activity of ceftriaxone and absent PAE may explain the slow clinical response in typhoid.

Background: In recent years, substantial progress in malaria control has been made in Myanmar which has 85% of the malaria in the Greater Mekong subregion. The residual burden of disease now remains primarily in remote areas where currently recommended malaria elimination approaches are generally not feasible.Method:Medical Action Myanmar introduced community health workers (CHWs) to deliver early diagnosis and treatment for malaria in hard-to-reach communities in Mon state, East Myanmar. We conducted a retrospective analysis to assess the impact of this intervention. The decline in malaria incidence and prevalence in febrile patients was assessed using negative binomial regression mixed effects models.Results: Between 2011 and 2018, a network of 172 CHWs serving a population of 236,340 carried out 260,201 malaria rapid diagnostic tests (RDTs) to investigate acute febrile illness. The median (IQR) blood examination rate was 1.33% (0.38–3.48%) per month. The incidence of P. falciparum malaria (including mixed infections) declined by 70% (95% CI: 65–75%) each year of CHW operation. P. vivax malaria incidence declined by 56% (95% CI: 50–62%) per year. Malaria RDT positivity rates for P. falciparum and P vivax declined by 69% (95% CI: 62–75%) and 53% (95% CI: 47–59%) per year respectively. Between 2017 and 2018 only one imported P. falciparum case was detected in 54,961 RDTs.Interpretation: The introduction of CHWs providing community-based malaria diagnosis and treatment, integrated in basic health care services, in remote communities in Mon state led to a dramatic reduction in malaria. Within six years, P. falciparum was eliminated.Funding: The malaria project was supported by The Global Fund to fight AIDS, Tuberculosis and Malaria and Medical Action Myanmar. This retrospective analysis has been conducted with the support of The Wellcome Trust.Declaration of Interest: We declare no competing interests.Ethical Approval: This is not a trial or a description of prospective clinical research. It did not require ethical approval. It is a description of how a malaria control implementation in Myanmar was conducted in cooperation with the Myanmar National Malaria Control Programme, and what impact that had. The implementation was in accordance with, and part of, the National Malaria Control programme (who subsequently took over parts of the project). It is a description of best practice. We have published about this project earlier. ERB approval was also not required.

Artemisinin resistance in Plasmodium falciparum extends across Southeast Asia where it is associated with worsening partner drug resistance and a decline in the efficacy of frontline artemisinin-based combination therapy. Dihydroartemisinin-piperaquine (DP) is an essential component of preventive and curative treatment in the region, but its therapeutic efficacy has fallen in Cambodia. A prospective clinical and parasitological evaluation of DP was conducted at two sites in Upper Myanmar between August 2013 and December 2014, enrolling 116 patients with acute uncomplicated falciparum malaria. Patients received DP orally for 3 days together with primaquine 0.25 mg/kg on admission. Parasite clearance half-lives based on 6 hourly blood smears, and day 42 therapeutic responses were assessed as well as parasite K13 genotypes. Median parasite clearance half-life was prolonged, and clearance half-life was greater than 5 h in 21 % of patients. Delayed parasite clearance was significantly associated with mutations in the propeller region of the parasite k13 gene. The k13 F446I mutation was found in 25.4 % of infections and was associated with a median clearance half-life of 4.7 h compared with 2.7 h for infections without k13 mutations (p < 0.001). There were no failures after 42 days of follow-up, although 18 % of patients had persistent parasitaemia on day 3. The dominant k13 mutation observed in Upper Myanmar, F446I, appears to be associated with an intermediate rate of parasite clearance compared to other common mutations described elsewhere in the Greater Mekong Subregion. Discerning this phenotype requires relatively detailed clearance measurements, highlighting the importance of methodology in assessing artemisinin resistance.

In all experimental mammals tested (rats, dogs, primates), intramuscular injections of the oil-soluble antimalarial artemisinin derivatives artemether and arteether have produced an unusual pattern of selective damage to brain stem centers predominantly involved in auditory processing and vestibular reflexes. Artesunate, the most widely used of these compounds, is a water soluble hemisuccinate derivative given parenterally either by intravenous or intramuscular injection. The neurotoxic potential of parenteral artesunate and artemether was compared in a murine model. Adult Swiss albino mice were assigned randomly to 28-day regimens of intramuscular artemether or artesunate in doses ranging from 30 to 100 mg/kg/day. At 30 mg/kg/day, no abnormalities were detected with either drug. At 50 mg/kg/day, abnormalities were observed in six of 12 artemether recipients and two of 12 artesunate recipients. These were reversible in all but one (artemether) mouse. At 100 mg/kg/day, eight of 36 artemether recipients, two of 36 artesunate recipients, and one of 18 control mice died. All but four surviving mice in the artemether group (86%) showed obvious and usually irreversible abnormalities of balance and equilibrium, whereas only four artesunate recipients (11%) exhibited abnormalities, and these were reversible in each case (P &lt; 0.001). At this dose the relative risk (95% confidence interval) for death or disability was 5.3 (2.6-11.2) for artemether recipients. Intramuscular artemether is significantly more neurotoxic than intramuscular artesunate in this murine model.

Background Primaquine is the only available drug that clears mature Plasmodium falciparum gametocytes in infected human hosts, thereby preventing transmission of malaria to mosquitoes. However, concerns about dose-dependent haemolysis in people with glucose-6-phosphate dehydrogenase (G6PD) deficiencies have limited its use. We assessed the dose-response association of single-dose primaquine for gametocyte clearance and for safety in P falciparum malaria. Methods We undertook this randomised, double-blind, placebo-controlled trial with four parallel groups in Jinja district, eastern Uganda. We randomly allocated Ugandan children aged 1–10 years with uncomplicated falciparum malaria and normal G6PD enzyme function to receive artemether–lumefantrine, combined with either placebo or with 0·1 mg/kg, 0·4 mg/kg, or 0·75 mg/kg (WHO reference dose) primaquine base. Randomisation was done with computer-generated four-digit treatment assignment codes allocated to random dose groups in block sizes of 16. Study staff who provided care or assessed outcomes and the participants remained masked to the intervention group after assignment. The primary efficacy endpoint was the non-inferiority of the mean duration of gametocyte carriage in the test doses compared with the reference group of 0·75 mg primaquine per kg, with a non-inferiority margin of 2·5 days. The primary safety endpoint was the superiority of the arithmetic mean maximum decrease in haemoglobin concentration from enrolment to day 28 of follow-up in the primaquine treatment groups compared with placebo, with use of significance testing of pairwise comparisons with a cutoff of p=0·05. The trial is registered with ClinicalTrials.gov, number NCT01365598. Findings We randomly allocated 468 participants to receive artemether–lumefantrine combined with placebo (119 children) or with 0·1 mg/kg (116), 0·4 mg/kg (116), or 0·75 mg/kg (117) primaquine base. The mean duration of gametocyte carriage was 6·6 days (95% CI 5·3–7·8) in the 0·75 mg/kg reference group, 6·3 days (5·1–7·5) in the 0·4 mg/kg primaquine group (p=0·74), 8·0 days (6·6–9·4) in the 0·1 mg/kg primaquine group (p=0·14), and 12·4 days (9·9–15·0) in the placebo group (p<0·0001). No children showed evidence of treatment-related haemolysis, and the mean maximum decrease in haemoglobin concentration was not associated with the dose of primaquine received—it did not differ significantly compared with placebo (10·7 g/L, SD 11·1) in the 0·1 mg/kg (11·4 g/L, 9·4; p=0·61), 0·4 mg/kg (11·3 g/L, 10·0; p=0·67), or 0·75 mg/kg (12·7 g/L, 8·2; p=0·11) primaquine groups. Interpretation We conclude that 0·4 mg/kg primaquine has similar gametocytocidal efficacy to the reference 0·75 mg/kg primaquine dose, but a dose of 0·1 mg/kg was inconclusive for non-inferiority. Our findings call for the prioritisation of further trials into the efficacy and safety of doses of primaquine between 0·1 mg/kg and 0·4 mg/kg (including the dose of 0·25 mg/kg recently recommended by WHO), in view of the potential for widespread use of the drug to block malaria transmission. Funding Wellcome Trust and the Bill & Melinda Gates Foundation.

Dihydroartemisinin-piperaquine (DP) is a fixed-dose artemisinin-based combination treatment. Field pharmacokinetic studies would be simplified and facilitated by being able to use small volume capillary assays rather than venous blood. The aim of this study was to describe the relationship between piperaquine concentrations measured in capillary blood, venous blood and venous plasma. Samples of plasma, whole blood obtained by venesection and capillary blood were taken simultaneously from patients with uncomplicated Plasmodium falciparum malaria treated with DP between 0 and 9 weeks after treatment. Piperaquine concentrations in venous and capillary samples were measured using solid phase extraction and analysis by liquid chromatography with ultraviolet detection. A total of 161 sets of the three measures were obtained from 54 patients. Piperaquine concentrations in the venous blood samples were approximately twofold higher and those in the capillary blood samples were threefold higher than the corresponding venous plasma concentrations. Capillary blood piperaquine concentrations were approximately 1.7-fold higher than venous blood concentrations, and this difference also increased with time. Differences in whole blood and plasma levels of piperaquine suggest compartmentalisation of the drug within blood cells, as also occurs with the structurally related quinoline chloroquine. The relationship between piperaquine concentrations in the venous plasma, venous blood and capillary blood is variable and unpredictable at low concentrations. However, within the range of concentrations usually present in patients between 3 and 21 days after treatment with currently recommended doses, the relationship between capillary and venous whole blood is predictable; consequently, capillary blood sampling can be used in field assessments.