Chloroquine Versus Dihydroartemisinin-Piperaquine With Standard High-dose Primaquine Given Either for 7 Days or 14 Days in Plasmodium vivax Malaria

Chloroquine or dihydroartemisinin-piperaquine with either 7- or 14-day primaquine regimens provided highly effective radical cure of vivax malaria on the Thailand-Myanmar border. Short course higher dose primaquine regimens for vivax malaria radical cure may improve efficacy by improving adherence.

Plasmodium vivax is the most geographically dispersed of the human malarias, once extending to the Arctic circle and now as far north as the Korean peninsula [1]. Chloroquine has been the first-line treatment for over 70 years [2]. In recent years low grade chloroquine resistance, manifested by earlier appearance of relapses, has been reported increasingly, but high grade resistance remains confined to Oceania and Indonesia [3]. Piperaquine (a slowly eliminated bisquinoline) combined with dihydroartemisinin is well tolerated and highly effective against chloroquine resistant vivax malaria and has replaced chloroquine as first-line treatment in Indonesia [4].
The majority of P. vivax infections in East Asia and Oceania result from relapses [5,6]. These occur usually at 5-9 week intervals following slowly eliminated antimalarial treatments. Relapses are a major cause of morbidity, particularly in children. The only currently available antirelapse drug is the 8-aminoquinoline primaquine. The most widely used radical cure regimen is 0.25 mg base/kg/day over 14 days [2]. Some studies report poor adherence to the 14-day regimen [7][8][9][10]. Shorter 5-day regimens at lower total doses have been recommended in the past [11,12], but they are ineffective [13,14]. Nearly all South American countries now recommend a 7-day regimen with the same total dose (3.5mg/kg: adult dose 210 mg) as the 14-day regimen. This is considered to improve adherence and thus effectiveness. In Southeast Asia and Oceania higher doses (7mg base/kg divided over 14 days: adult dose 420 mg) are required for radical cure. Preliminary trial data suggest this total dose can also be given safely over 7 days [15,16].
This open 2-way randomized controlled trial compared the radical curative efficacy of short course 7-day high-dose primaquine with the standard 14-day high-dose regimen. Patients were also randomized either to chloroquine or dihydroartemisinin-piperaquine for treatment of the blood stage P. vivax infection.

Ethics Approval
This study was approved by both the Mahidol University Faculty of Tropical Medicine Ethics Committee (MUTM 2011-043, TMEC 11-008) and the Oxford Tropical Research Ethics Committee (OXTREC  and was registered at ClinicalTrials.gov (NCT01640574).

Setting
This study was conducted by the Shoklo Malaria Research Unit (SMRU), which operates clinics along the Thailand-Myanmar border. This is an area of hill forest and low seasonal malaria transmission. The patient population comprised migrant workers and displaced persons of Burman and Karen ethnicities [17].

Participants and Study Procedures
Patients ≥6 months and ≥7 kg with uncomplicated P. vivax monoinfections were included. Patients were excluded if glucose-6-phosphate dehydrogenase (G6PD) deficient by the fluorescent spot test (FST), were pregnant or breastfeeding an infant ≤ 6 months, had a hematocrit ≤25%, or received a blood transfusion within 3 months. Written informed consent was obtained from patients or from parents or guardians of children less than 18 years old.

Randomization
Randomization was computer generated in blocks of 20. Patients were randomized to either: Food and drink were given before drug administration. All doses were supervised.

Enrolment Procedures
At enrolment (day 0) a medical history and vital signs were recorded, and a physical examination was performed. A malaria smear, hematocrit, complete blood count (CBC), and urine β-human chorionic gonadotropin (β-hCG) pregnancy test were performed. Parasite cultures were taken when densities exceeded 300 per 500 white blood cells (WBC) (approximately 4500/µL). Three blood spots on filter paper (Whatman 3MM) were collected for human and parasite genotyping.

Follow-up
Vital signs, concomitant medications, and adverse events were assessed at each follow-up visit. A malaria smear was taken daily until negative. Fever clearance was defined as the interval to an aural temperature <37.5° C on 2 consecutive daily measurements. A venous plasma blood sample for antimalarial drug levels was taken on day 6. Methemoglobin was measured using a transcutaneous pulse oximeter (Masimo® Radical-7) on days 0, 3, 6, and 13 and additionally on day 10 in the primaquine 14-day groups. Follow-up visits continued at weeks 2 and 4 and then every 4 weeks until 52 weeks. At each visit, a malaria smear and CBC were taken. Urine β-HCG pregnancy testing was performed in females of childbearing potential.

Laboratory Investigations
Malaria blood films were stained with Giemsa, and parasites were counted per 1000 red cells or per 500 WBC. In vitro susceptibility assays were performed when >80% ring stages were present. Hematocrits were measured with a Hawksley Micro-Haematocrit reader. The CBC was performed using a CeltacF MEK-8222K hematology analyzer (Nihon Kohden, Japan). The G6PD FST (R&D Diagnostic, Greece) was performed as described previously [18]. Chloroquine/desethylchloroquine, piperaquine, and primaquine/ carboxyprimaquine blood concentrations were measured using 3 validated liquid chromatography (LC)-tandem mass spectrometry (MS/MS) methods. The lower limits of quantification (LLOQ) were set to 1.13, 1.20, 0.912 and 3.90 ng/mL for chloroquine/desethylchloroquine, piperaquine, primaquine, and carboxyprimaquine, respectively. All 3 quantification methods were validated according to regulatory standards, and 3 levels of quality control samples were analyzed in triplicate within each batch of clinical samples. Total imprecision (ie, relative standard deviation) for all quality control samples was below 10% during drug quantification.

Malaria Infections Recurring During Follow-up
Enrolment study procedures were repeated for P. vivax recurrences. Blood was taken again for chloroquine or piperaquine levels. Standard high-dose primaquine (0.5 mg/kg/ day for 14 days) and chloroquine were then given. Follow-up was restarted as if newly recruited. The total study duration remained 52 weeks from enrolment. For P. falciparum infections, standard treatment with mefloquine and artesunate or dihydroartemisinin-piperaquine [2] was given and follow-up continued without interruption.

Pregnancy During Follow-up
Patients with a positive urine β-HCG pregnancy test during follow-up were referred to the antenatal clinic, and study follow-up was continued. If any malaria infection occurred during pregnancy, the patient was censored from the study and managed in the antenatal clinic.

Adverse Events and Serious Adverse Events
Adverse event data were collected until day 42. Patients were treated with hematinics for hematocrit <30% or <34% if under 2 years old. The blood transfusion threshold was a hematocrit <18% or if the patient was symptomatic. Inpatient observation was performed for symptomatic methemoglobin elevation or if ≥15% without symptoms. Serious adverse events were reported to a Data Safety Monitoring Board within 24 hours and the local ethics committee within 7 days.

Sample Size
Based on earlier experience at SMRU, the annual P. vivax recurrence rate in the 14-day groups was anticipated to be ~18%. With a 1-sided alpha of 0.025, power 80%, and up to 20% loss to follow-up, a sample size of 680 (170 subjects per arm) was necessary to show noninferiority of the 7-day primaquine groups with a delta of 10% compared to the 14-day groups [19,20].

Statistical Analysis
The primary endpoint was the cumulative risk of P. vivax recurrence by 52 weeks. Secondary endpoints included the cumulative risk of P. vivax recurrence by 8 weeks, adverse events within 42 days, and the relationship of day 6 drug concentrations with the risk of recurrence. A per protocol timeto-event (ie, time to first recurrence) analysis was performed using Cox proportional hazards model, with premature discontinuations right-censored. A second, mixed-effects Cox model was fitted to all data, with individual random effects for multiple recurrences. Both models included sex, treatment arm, and the log 10 day primaquine and carboxy-primaquine concentrations as independent variables. A Cox proportional hazards model was also used to assess the risk factors associated with premature discontinuation. Groups were compared using Pearson χ 2 test, nonparametric K-sample test, multivariable linear or logistic regression as appropriate. Thirtyfour G6PD heterozygous females (determined by G6PD spectrophotometry and genotyping) were excluded from the hematocrit analyses (except for enrolment data). These data were reported previously [21]. Stata 15.1 (StataCorp, College Station, TX, USA) and R version 3.4.3 (open source) were used for the data analysis.

Patient Characteristics
Between February 2012 and July 2014, 680 patients with uncomplicated acute vivax malaria were enrolled ( Figure 1). Patient characteristics were similar between treatment arms (Tables S1  and S2). Splenomegaly was more common in the 5-15 year age group (6%) (Supplementary Table S3 Table S6).

Post-treatment Suppression of Recurrent Infections
At this location, the majority of P. vivax recurrences (mainly relapses) without primaquine occur within 8 weeks, and over 90% occur within 4 months [22] (Figure 2). In this study the earliest recurrence was on day 37. First recurrences were slightly earlier in the chloroquine groups: between weeks 6 and 8 there were 10 recurrences of 361 at risk in the chloroquine arms, and 4 of 326 in the piperaquine arm (all in week 8) (Figure 2).

Radical Cure Effectiveness and the Relationship With In Vivo Drug Concentrations and In Vitro Susceptibility
The majority of patients completed their primaquine treatment (Supplementary Table S7). There was no association between the day 6 total chloroquine or piperaquine concentrations and the interval to first P. vivax recurrence within 8 weeks. No patients had "therapeutic" blood chloroquine (≥10 ng/mL [23]) or piperaquine (≥30 ng/mL [24]) concentrations at week 8. In both the Cox proportional hazards model of time to first recurrence and the mixed-effects Cox model (all recurrences) higher day 6 concentrations of carboxyprimaquine (P = .04, both models) but not primaquine were associated with a lower risk of subsequent recurrence (Figure 4). The geometric mean (95% CI) of chloroquine IC 50 in the 36 successful in vitro tests was 17.9 ng/ mL (11.5 to 27.7). Six isolates (16.7%) had an IC 50 > 50 ng/mL, suggesting resistance. Figure 1. Study diagram. a One patient randomized to the CP7 arm was given DP7 b Two patients randomized to the CP14 arm were given CP7 c Two patients vomited the drug treatment twice and one patient had received a blood transfusion in the prior 3 months. d A total of 26 patient were excluded from the analyses because they had not completed the schizonticidal treatment. The three patients with incorrect randomization were included in the analysis. Abbreviations: CP7, Chloroquine + Primaquine 1 mg/kg/day for 7 days; CP14, Chloroquine + Primaquine 0.5 mg/kg/day for 14 days; DP7, Dihydroartemisinin-piperaquine + Primaquine 1 mg/kg/day for 7 days; DP14, Dihydroartemisininpiperaquine + Primaquine 0.5 mg/kg/day for 14 days; G6PD, glucose-6-phosphate dehydrogenase.
Except for abdominal pain, there were no significant differences in reported adverse events between the 4 groups ( Table 2) Vertical notches denote right censored observations. 95% confidence intervals are shown by the shaded areas. The dashed black line shows the proportion recurring when not given radical cure in recent (two years previous) historical controls from the same study site [22].   Table S8).
All resolved without further sequelae, and none had a subsequent P. vivax infection. There were 30 serious adverse events reported; most common were methaemoglobinaemia (n = 10), haemolysis (n = 3), and presumed bacterial infection (n = 10). Four deaths occurred. None was considered related to the study drugs.  Relapse of vivax malaria along the Thailand-Myanmar border was very effectively prevented by both the conventional 14-day high-dose regimen and the 7-day high-dose primaquine regimen. This is consistent with earlier studies with shorter follow-up periods [15,16,25,26]. Only 14 (2%) of 654 patients had recurrent P. vivax infections within 8 weeks, whereas without primaquine, approximately 50% of patients experience relapse within 8 weeks of chloroquine treatment [27,28] in this area ( Figure 2). Without primaquine, 90% of all recurrences (mainly relapses) occur within 4 months [22]. However, with radical cure in this study, recurrences showed no periodicity, so they were most likely newly acquired infections after the post-treatment prophylaxis provided by chloroquine or piperaquine had waned. Radical cure efficacy is thought to depend mainly on the total exposure to the bioactive metabolites of primaquine. In this study, higher day 6 concentrations of the more slowly eliminated biologically inactive metabolite carboxyprimaquine were associated with lower recurrence rates ( Figure 4). This may reflect greater exposure to the active metabolites. The health benefits of preventing relapse are substantial. Where supervised treatment is not possible, the 7-day high-dose regimen should have an adherence advantage while providing similar radical curative efficacy to the standard 14-day regimen. But it is slightly less well tolerated with a higher incidence of abdominal discomfort and greater methemoglobinemia. More concerning was clinically significant hemolytic anemia in G6PD Mahidol female heterozygotes [21]. This is potentially a major limitation in areas where G6PD deficiency is prevalent and quantitative G6PD assessments are unavailable. Otherwise, the hematological consequences were mild. Although G6PD testing is recommended when radical cure is provided, it is usually unavailable in most endemic areas. Current qualitative tests identify only a minority of female heterozygotes. Higher daily primaquine dose regimens carry higher hemolytic risks. Quantitative G6PD tests are being developed, which would identify the vulnerable heterozygous females [29] allowing the 7-day high-dose primaquine regimen to be used safely. Further studies are in progress to assess the safety and effectiveness of the 7-day high-dose primaquine regimen in areas where other G6PD deficiency variants are prevalent [30].
Low grade chloroquine resistance in P. vivax is prevalent along the Thailand-Myanmar border, as evidenced by recent slowing of fever and parasite clearance, an increasing proportion of recurrences within 28 days of starting chloroquine Figure 5. Median methemoglobin measurements during primaquine administration. Evolution of methemoglobin concentrations over time within the four study groups. Data from PMQ7 arm are shown in light and dark red for CQ and DP, respectively. Data from PMQ14 are shown in light and dark blue for CQ and DP, respectively. The thick brown and aqua lines correspond to the median trends for the pooled 7-day and 14-day primaquine groups, respectively. Abbreviations: CQ, chloroquine; DP, dihydroartemisinin-piperaquine, PMQ7, primaquine 1 mg/kg/day for 7 days; PMQ14, primaquine 0.5 mg/kg/day for 14 days. alone and declining in vitro susceptibility [4]. Despite this, combination treatment with primaquine, which also has significant asexual stage activity against P. vivax [31], remains efficacious. As reported previously, dihydroartemisinin-piperaquine is a well-tolerated and highly efficacious alternative to chloroquine, which can be considered as first-line treatment, thereby providing a unified treatment for all malarias [2]. Dihydroartemisinin-piperaquine has the advantage of giving faster therapeutic responses (attributed to the artemisinin component), with shorter times to fever clearance, and slightly longer period of post-treatment suppression of recurrent infections. There was also a trend to lower recurrence rates with dihydroartemisinin-piperaquine ( Figure 2), but this was not statistically significant.
A limitation of this study was the high rate of premature discontinuation, which was predominantly due to migration. These patients may be at higher risk for P. vivax malaria, so recurrence rates in this study could be underestimated.
In this region, both chloroquine and dihydroartemisinin-piperaquine combined with primaquine (given over 14 days or in the same total dose over 7 days) provide highly effective radical cure regimens in P. vivax malaria. Radical cure needs to be deployed more widely. Quantitative G6PD testing would ensure the safe use of 7-day high-dose primaquine regimens in G6PD heterozygous females.

Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author. Fatigue, n (%) 11 (7%) 9 (6%) 6 (4%) 10 (6%) .668 a χ 2 used for significance testing unless otherwise noted.

Notes
b Abdominal pain was significantly less in the DP14 arm compared to the other arms; logistic regression was used to compare differences between groups.