Exercício_10_Artigo_Ensaio Clínico

(1)

Tafenoquine plus chloroquine for the treatment and relapse

prevention of Plasmodium vivax malaria (DETECTIVE):

a multicentre, double-blind, randomised, phase 2b

dose-selection study

Alejandro Llanos-Cuentas, Marcus V Lacerda, Ronnatrai Rueangweerayut, Srivicha Krudsood, Sandeep K Gupta, Sanjay K Kochar, Preetam Arthur, Nuttagarn Chuenchom, Jörg J Möhrle, Stephan Duparc, Cletus Ugwuegbulam, Jörg-Peter Kleim, Nick Carter, Justin A Green, Lynda Kellam

Summary

Background Clinical eﬀ ectiveness of previous regimens to treat Plasmodium vivax infection have been hampered by compliance. We aimed to assess the dose–response, safety, and tolerability of single-dose tafenoquine plus 3-day chloroquine for P vivax malaria radical cure.

Methods In this double-blind, randomised, dose-ranging phase 2b study, men and women (aged ≥16 years) with microscopically confi rmed P vivax monoinfection (parasite density >100 to <100 000 per µL blood) were enrolled from community health centres and hospitals across seven sites in Brazil, Peru, India, and Thailand. Patients with glucose-6-phosphate dehydrogenase enzyme activity of less than 70% were excluded. Eligible patients received chloroquine (days 1−3) and were randomly assigned (1:1:1:1:1:1) by a computer-generated randomisation schedule to receive single-dose tafenoquine 50 mg, 100 mg, 300 mg, or 600 mg, primaquine 15 mg for 14 days, or chloroquine alone. Randomisation was stratifi ed by baseline parasite count (≤7500 and >7500 per µL blood). The primary eﬃ cacy endpoint was relapse-free eﬃ cacy at 6 months from initial dose (ie, clearance of initial infection without subsequent microscopically confi rmed infection), analysed by intention to treat. This study is registered with ClinicalTrials.gov, number NCT01376167.

Findings Between Sept 19, 2011, and March 25, 2013, 329 patients were randomly assigned to a treatment group (chloroquine plus tafenoquine 50 mg [n=55], 100 mg [n=57], 300 mg [n=57], 600 mg [n=56]; or to chloroquine plus primaquine [n=50]; or chloroquine alone [n=54]). Relapse-free effi cacy at 6 months was 57·7% (95% CI 43−70) with tafenoquine 50 mg, 54·1% (40−66) with tafenoquine 100 mg, 89·2% (77−95) with tafenoquine 300 mg, 91·9% (80−97) with tafenoquine 600 mg, 77·3% (63−87) with primaquine, and 37·5% (23−52) with chloroquine alone. Tafenoquine 300 mg and 600 mg had better effi cacy than chloroquine alone (treatment diff erences 51·7% [95% CI 35−69], p<0·0001, with tafenoquine 300 mg and 54·5% [38−71], p<0·0001, with tafenoquine 600 mg), as did primaquine (treatment diff erence 39·9% [21−59], p=0·0004). Adverse events were similar between treatments. 29 serious adverse events occurred in 26 (8%) of 329 patients; QT prolongation was the most common serious adverse event (11 [3%] of 329), occurring in fi ve (2%) of 225 patients receiving tafenoquine, four (8%) of 50 patients receiving primaquine, and two (4%) of 54 patients receiving chloroquine alone, with no evidence of an additional eff ect on QT of chloroquine plus tafenoquine coadministration.

Interpretation Single-dose tafenoquine 300 mg coadministered with chloroquine for P vivax malaria relapse prevention was more eﬃ cacious than chloroquine alone, with a similar safety profi le. As a result, it has been selected for further clinical assessment in phase 3.

Funding GlaxoSmithKline, Medicines for Malaria Venture.

Introduction

Until recently, Plasmodium vivax was a relatively neglected pathogen, with previous eﬀ orts mainly focused on reducing the mortality associated with Plasmodium

falciparum infection in Africa.1,2_{However, across Asia and}

Latin America, where P vivax infection is most prevalent, the infection is a major health and economic burden.3,4

Unlike P falciparum, P vivax forms hypnozoites that persist in the human liver for extended periods, which can cause many clinical relapses for months or years after infection from just one mosquito bite. Additionally,

eﬀ orts to eliminate malaria will need to address P vivax infection, in particular the hypnozoite reservoir.5

For 60 years, primaquine has been the only drug licensed for P vivax hypnozoite eradication.6_{The standard}

regimen for P vivax radical cure involves treatment of the blood-stage infection, usually with chloroquine for 3 days, followed by primaquine for up to 14 days. Patients often feel clinically well within a few days of starting chloroquine; thus, compliance with primaquine treat-ment is a substantial obstacle to achieving adequate clinical eﬀ ectiveness with this regimen.7

Published Online December 19, 2013 http://dx.doi.org/10.1016/ S0140-6736(13)62568-4 See Online/Comment http://dx.doi.org/10.1016/ S0140-6736(13)62672-0 Instituto de Medicina Tropical, Alexander von Humboldt, Universidad Peruana Cayetano Heredia, Lima, Peru (Prof A Llanos-Cuentas MD); Fundação de Medicina Tropical Doutor Heitor Vieira Dourado, Manaus, Amazonas, Brazil (Prof M V Lacerda MD); Mae Sot Hospital, Mae Sot, Tak Province, Thailand (R Rueangweerayut MD, N Chuenchom MD); Faculty of Tropical Medicine, Mahidol University, Bangkok, Thailand (Prof S Krudsood MD); MV Hospital and Research Centre, Lucknow, Uttar Pradesh, India (S K Gupta MD); Sardar Patel Medical College, Bikaner, Rajasthan, India (S K Kochar MD); Sri Ramchandra Medical College and Research Institute, Department of Medicine, Porur, Chennai, Tamil Nadu, India (P Arthur MD); Medicines for Malaria Venture, Geneva, Switzerland (J J Möhrle PhD, S Duparc MD); and GlaxoSmithKline, Stockley Park West, Uxbridge, Middlesex, UK (C Ugwuegbulam PhD, J-P Kleim PhD, N Carter MSc, J A Green MD, L Kellam MSc) Correspondence to: Ms Lynda Kellam, GlaxoSmithKline, Stockley Park West, Uxbridge, Middlesex UB11 1BT, UK [email protected]

(2)

Tafenoquine is an 8-aminoquinoline antihypnozoite drug being developed for P vivax radical cure, and is coadministered with the standard chloroquine regimen. Results of studies in rhesus monkeys suggest that, compared with tafenoquine monotherapy, elimination of parasite asexual stages by cotreatment with chloro-quine or artemether-lumefantrine or pretreatment with quinine increases tafenoquine antihypnozoite eﬃ cacy by ten times.8_{Tafenoquine has a half-life of 2–3 weeks,}

raising the possibility of single-dose treatment and directly observed therapy.9,10_{Results of previous clinical}

studies have shown the potential of tafenoquine plus chloroquine in P vivax radical cure, with signifi cant eﬃ cacy and good tolerability at low tafenoquine single doses of 500 and 600 mg.10,11_{However, a key safety issue}

for both primaquine and tafenoquine is the capacity to induce haemolysis in glucose-6-phosphate (G6PD)-defi cient patients.12

The Dose and Eﬃ cacy Trial Evaluating Chloroquine and Tafenoquine In Vivax Elimination (DETECTIVE) was a phase 2b clinical study of an integrated phase 2b/3 programme. It aimed to characterise the dose–response of standard chloroquine plus tafenoquine doses between 50 mg and 600 mg for P vivax relapse prevention and to assess the safety and tolerability of the regimen.

Methods

Study design and patients

In this double-blind, randomised, dose-ranging study, patients were recruited from community health centres and hospitals across seven sites in Peru, Brazil, Thailand, and India. Male and female patients aged 16 years or older who presented with malarial symptoms were screened for eligibility. Inclusion criteria were microscopically confi rmed uncomplicated P vivax monoinfection with an asexual parasite density of greater than 100 per µL blood and less than 100 000 per µL blood. Main exclusion criteria were antimalarial treat-ment within the past 30 days, severe malaria, any clinically signifi cant concurrent illness, severe vomiting, or a haemoglobin concentration of less than 70 g/L. Pregnant or lactating female patients were excluded. Patients were excluded if their G6PD enzyme activity was less than 70% of the derived site median, as assessed with a spectrophotometric semiquantitative assay (Trinity Biotech, Bray, County Wicklow, Ireland, or Pointe Scientifi c, Canton, MI, USA). Female patients with G6PD enzyme activity of less than 90% of the site median were required to have a haemoglobin con-centration of 100 g/L or greater. The site median of G6PD enzyme activity was established at each site before study start and was based on results from 36 healthy men who had been tested using the same assay as for enrolment.

The study conformed to Good Clinical Practice and the Declaration of Helsinki (2000) and met all applicable

regulatory requirements. Protocol approval was obtained from each study site’s ethics committee or institutional review board. Written informed consent was obtained from all participants aged 18 years or older and from the parents or guardians of those aged 16–17 years with the additional requirement of patient assent. The protocol underwent two amendments during the study; neither pertained to study treatments or outcomes, details are in the fi nal protocol available from the corresponding author.

Randomisation and masking

Eligible patients were randomly assigned in equal blocks of 12 to six treatment groups in a blind, double-dummy design. A computer-generated randomisation schedule, stratifi ed by baseline parasite count (≤7500 and >7500 per µL blood), was provided to each site by GlaxoSmithKline (Harlow, UK). Patients, study staﬀ , and GlaxoSmithKline personnel were masked to study treat-ment allocation. Patients were allocated an individual numbered treatment pack on the basis of their random-isation number. Matched tafenoquine and primaquine placebos were presented in identical packaging.

Procedures

Study treatments were supplied as chloroquine tablets containing 300 mg chloroquine free base (Aralen, Sanofi , Bridgewater, NJ, USA), tafenoquine formulated as 50 mg, 100 mg, and 150 mg hard gelatine capsules of identical appearance (GlaxoSmithKline, Harlow, UK), and overencapsulated primaquine tablets containing 15 mg primaquine free base (Sanofi ).

All patients were given once-daily oral chloroquine on day 1 (600 mg), day 2 (600 mg), and day 3 (300 mg). The six treatment groups were: additional therapy with tafenoquine (given as a single oral dose with food on day 1 or day 2) at doses of 50 mg, 100 mg, 300 mg, or 600 mg; additional therapy with once-daily oral prima-quine (15 mg for 14 days starting on day 2), or no additional therapy (ie, chloroquine alone). Tafenoquine and primaquine placebos were administered to main-tain masking. Patients received clinic-based directly observed therapy for days 1−3 (ie, comprising all doses of chloroquine and tafenoquine, and the fi rst two doses of primaquine). Subsequently, patients were followed up as outpatients on days 5, 8, 11, 15, 22, 29, 60, 90, 120, and 180. Pill counts and reminders were used to reinforce primaquine compliance until day 15. Phys-ical examination was done at screening and on days 2, 3, 8, 15, and all subsequent follow-up visits. A 12-lead electrocardiograph was taken at screening and on days 2, 3, 29, and 180 (twice daily on days 1–3). Blood samples were taken for haematological testing and clinical chemistry tests and samples were obtained for urinalysis at screening and on days 3, 5, 8, 11, 15, 22, 29, 60, 90, and 180. G6PD genotyping (Emory Genetic Laboratory, Decatur, GA, USA) was done for all female

(3)

patients and any male patient with a substantial decrease in haemoglobin (decrease of >25 g/L or ≥25% decline from baseline). A range of ophthalmic safety assessments of the lens and retina were done at three study sites (Manaus, Brazil; Mae Sot, Thailand; and Bikaner, India) at screening and on days 29 and 90, and on day 180 if day 90 observations were abnormal.

P vivax asexual parasitaemia and gametocytaemia were

ascertained from duplicate Giemsa-stained thick blood smears (10 µL fi nger prick) using established methods. Thin blood fi lms were also taken for species deter-mination. Parasite counts were obtained at screening, after treatment on day 1, and on days 2 and 3 twice daily every 6−12 h until parasite clearance was confi rmed in two consecutive readings. Follow-up blood smears were obtained at days 8, 15, 22, 29, 60, 90, 120, and 180. Parasite count was calculated relative to the patient’s white blood cell count (up to a maximum of 500 white blood cells). Mean values obtained independently from two micro-scopists were used. Any discrepancies were resolved by comparing the individual values obtained by the fi rst two microscopists with the mean value obtained by a third microscopist; the closest value to this being recorded. Quality control was done centrally (FHI 360, Bangkok, Thailand) on all baseline slides, a proportion of

postbaseline slides from each centre, and all slides from patients with parasite recurrence.

Outcomes

For all eﬃ cacy endpoints, no distinction was made between P vivax recurrence caused by recrudescence, re-infection, or relapse through hypnozoite activation. The primary endpoint was the proportion of patients who had relapse-free eﬃ cacy at 6 months, which was defi ned as clearance of the initial malaria infection without sub-sequent microscopically confi rmed recurrence during the 6 month follow-up.

Secondary effi cacy endpoints were relapse-free effi cacy at 4 months, time to relapse, parasite clearance time, and fever clearance time (defi ned as time of fi rst dose of study medication to aparasitaemia or apyrexia in two consecutive samples obtained 6−12 h apart). Additional effi cacy endpoints were time to gametocyte clearance, defi ned as for parasite clearance time above, and the incidence of P falciparum malaria risk.

Key protocol-defi ned safety outcomes were clinically relevant declines in haemoglobin (decrease of >25 g/L or ≥25% decline from baseline) and changes in methaemo-globin. Additional safety outcomes were the incidence and severity of adverse events, and abnormal fi ndings

Figure 1: Trial profi le

G6PD=glucose-6-phosphate dehydrogenase. QTcF=QT corrected using Fredericia’s formula. P vivax=Plasmodium vivax. *Patients might have had more than one reason for study ineligibility. 424 patients screened for eligibility

319 patients completed the study 329 patients randomly assigned to

treatment 136 Peru, Iquitos 37 Brazil, Manaus 99 Thailand 57 Bangkok 42 Mae Sot 57 India 24 Lucknow 29 Bikaner 4 Chennai 329 patients analysed (intention-to-treat population) 55 patients received chloroquine plus tafenoquine 50 mg

1 patient was lost to follow-up

57 patients received chloroquine plus tafenoquine 100 mg

3 patients were lost to follow-up

1 patient was lost to follow-up

50 patients received chloroquine plus primaquine

54 patients received chloroquine only

0 patients were lost to follow-up 95 ineligible*

21 raised aspartate aminotransferase 18 G6PD entry criteria not met 17 no P vivax parasites 14 QTcF >450 ms

8 low P vivax parasite count 5 concomitant illness 4 no consent

3 childbearing criteria not met 3 drug unavailable 2 mixed malaria infection 2 haemoglobin entry criteria

not met 1 severe malaria 1 adverse event before

(4)

from clinical chemistry and haematology laboratory assessments (defi ned according to Common Terminology Criteria for Adverse Events version 4.0), 12-lead electro-cardiographs, and ophthalmic assessments.

Statistical methods

The study was designed to test superiority of any tafenoquine dose (plus chloroquine) versus chloroquine alone in the prevention of P vivax malaria relapse. Assuming a 60% effi cacy rate for chloroquine alone and a 10% dropout rate, the planned sample size was 324 (54 per treatment group). This provided 90% power to detect a treatment diff erence of 30% between each tafenoquine dose and chloroquine alone for the primary effi cacy endpoint. We used a step-down testing procedure to adjust for the multiple primary comparisons of each tafenoquine dose versus chloroquine, with p values generated only where the effi cacy of the next higher dose was statistically signifi cant at the 5% level. We did not plan a comparison between the tafenoquine and prima-quine groups, although we did compare chloro prima-quine plus primaquine with chloroquine alone.

The primary effi cacy analysis was done on all randomly assigned patients (intention-to-treat population). The primary effi cacy endpoint, relapse-free effi cacy at 6 months, was analysed using Kaplan-Meier estimates. A two-sided log-rank test was done over 6 months using a 5% signifi cance level. Patients that were relapse-free at day 180 were censored at day 180. Censoring before day 180 occurred if patients did not have P vivax at baseline, or took a drug with antimalarial action despite the absence of malaria parasites, or did not have a 6 month assessment within the defi ned time window (day 170–190). Relapse-free effi cacy at 4 months was similarly established. Clearance times for parasites, fever, and gametocytes and time to P vivax relapse were estimated using Kaplan-Meier methods. All other effi cacy and safety endpoints were analysed with descriptive statistics.

This study is registered with ClinicalTrials.gov, number NCT01376167.

Role of the funding source

The draft protocol was developed by the study sponsors and reviewed or agreed by the site principal investigators.

Chloroquine plus tafenoquine Chloroquine plus

primaquine (n=50) Chloroquine alone (n=54) 50 mg (n=55) 100 mg (n=57) 300 mg (n=57) 600 mg (n=56) Women 18 (33%) 13 (23%) 14 (25%) 11 (20%) 15 (30%) 15 (28%) Ethnic origin American Indian 27 (49%) 28 (49%) 29 (51%) 29 (52%) 25 (50%) 27 (50%) Asian, central/south 11 (20%) 11 (19%) 9 (16%) 10 (18%) 6 (12%) 10 (19%) Asian, southeast 16 (29%) 16 (28%) 19 (33%) 16 (29%) 16 (32%) 16 (30%) Mixed 1 (2%) 2 (4%) 0 1 (2%) 3 (6%) 1 (2%) Age, years 36·3 (17−68; 13·3) 34·6 (16−74; 14·1) 36·2 (16−64; 13·5) 35·7 (17−68; 15·1) 36·0 (16−72; 13·9) 33·6 (16−68; 14·2) Weight, kg 59·9 (37−91; 11·2) 59·4 (44−95; 10·6) 59·4 (43−84; 9·8) 62·2 (42−106; 13·6) 60·0 (40−99; 12·6) 59·3 (34−101; 13·8) Temperature, °C 37·4 (36–39; 0·9) 37·6 (35–40; 1·0) 37·4 (35–41; 1·1) 37·6 (35–41; 1·1) 37·5 (35–40; 1·0) 37·5 (36–40; 1·0)

Patients with fever 20 (36%) 24 (42%) 19 (33%) 21 (38%) 14 (28%) 21 (39%)

Asexual parasites, per μL blood

3736 (134–51 175) 3027 (0–43 600) 5796 (200–78 880) 5429 (160–56 100) 4322 (120–46 460) 5091 (30–38 630) Patients with

gametocytes 41 (75%) 44 (77%) 47 (82%) 44 (79%) 36 (72%) 41 (76%)

Previous malaria episode

Yes 35 (64%) 36 (63%) 28 (49%) 31 (55%) 31 (62%) 33 (61%) No 20 (36%) 20 (35%) 27 (47%) 25 (45%) 18 (36%) 21 (39%) Unknown 0 1 (2%) 2 (4%) 0 1 (2%) 0 Haemoglobin, g/L 121 (75–157; 19·9) 128 (99–164; 15·6) 125 (76–165; 18·2) 125 (74–156; 19·3) 122 (76–148; 15·8) 128 (91–165; 14·5) G6PD enzyme activity, IUg/Hb 9·9 (6−18; 3·0) 9·4 (6−19; 2·9) 9·2 (4−15; 2·4) 9·4 (5−18; 2·7) 9·5 (5−16; 2·6) 9·2 (5−17; 2·5) G6PD enzyme activity, % of site median 116·0 (71−246; 34·5) 110·7 (74−194; 23·7) 107·3 (74−178; 20·3) 113·0 (79−190; 24·6) 114·3 (71−208; 27·7) 108·7 (72−172; 19·3) Total chloroquine dose,

mg/kg 25·9 (16·4–40·3; 4·8) 25·9 (15·8–34·5; 3·9) 25·9 (17·9–35·0; 4·2) 25·1 (14·2–35·7; 4·9) 26·0 (15·2–37·5; 5·2) 26·5 (14·9–44·0; 5·7) Total tafenoquine or

primaquine dose, mg/kg 0·86 (0·55–1·34; 0·16) 1·73 (1·05–2·30; 0·26) 5·19 (3·57–7·01; 0·85) 10·10 (5·69–14·29; 1·95) 3·64 (2·13–5·25; 0·72) NA Data are number (%) or mean (range; SD), or median (range). NA=not applicable.

(5)

Data collection, statistical analysis, and clinical report development were done by GlaxoSmithKline. All authors had access to the primary data and are able to take responsibility for the accuracy and completeness of the results. All authors had fi nal responsibility for the decision to submit for publication.

Results

Between Sept 19, 2011, and March 25, 2013, 329 patients were randomly assigned to a treatment group (fi gure 1). All 329 patients were included in the safety and intention-to-treat populations. Ten patients did not complete the study; all were lost to follow-up. Baseline characteristics were similar across the treatment groups (table 1).

Relapse-free effi cacy at 6 months was highest in the two tafenoquine groups with the highest doses (89·2% [95% CI 77−95] in the 300 mg group and 91·9% [80−97] in the 600 mg group; table 2). Compared with chloro-quine alone, effi cacy was signifi cantly higher with the addition of tafenoquine 300 mg (treatment diff erence 51·7% [95% CI 35 −69], p<0·0001) or tafenoquine 600 mg (treatment diff erence 54·5% [38−71], p<0·0001). A post-hoc analysis requested by The Lancet comparing tafeno quine 300 mg with chloroquine alone gave an odds ratio of 8·4 (95% CI 3·4–20·6) for relapse-free effi cacy. The improvement in effi cacy for chloroquine plus prima quine compared with chloroquine alone was smaller than that observed for tafenoquine 300 mg and 600 mg, but was statistically signifi cant (treatment diff erence 39·9% [95% CI 21−59], p=0·0004). Time to

P vivax relapse is shown in fi gure 2 and shows the

diﬀ erentiation between the two higher and two lower tafenoquine doses tested. Analysis of the primary outcome by country is shown in table 3. Generally, the results were similar to the overall analysis, except for India where relapse-free eﬃ cacy at 6 months was 80% or higher for all study groups.

Similarly, relapse-free effi cacy at 4 months, when compared with that in the chloroquine alone group, was signifi cantly higher in the chloroquine plus tafenoquine 300 mg group (treatment diff erence 42·9% [95% CI 26−60], p<0·0001), tafenoquine 600 mg group (treatment diff erence 51·6% [37−66], p<0·0001), and primaquine group (treatment diff erence 32·0% [13−50], p=0·002). Relapse-free effi cacy at 4 months was not signifi cantly

diﬀ erent between the tafenoquine 50 mg or 100 mg doses and chloroquine alone (data not shown). Absolute relapse-free eﬃ cacy rates at 4 months were higher than at 6 months (data not shown). Only four patients had early treatment failure (before day 33), one each in the tafenoquine 100 mg and primaquine groups and two in the chloroquine alone group.

We noted no trends across treatment groups in median parasite, fever, or gametocyte clearance times (table 4). Only four patients had P falciparum malaria in the study, two in the chloroquine alone group and one each in the tafenoquine 100 mg and primaquine groups.

Overall, 227 (69%) of 329 patients reported adverse events (table 5), the most common being associated with episodes of malaria recurrence—ie, headache (91 [28%] of 329), pyrexia (79 [24%] of 329), and chills (76 [23%]

Figure 2: Kaplan-Meier estimates of time to Plasmodium vivax relapse

The fi nal day 180 assessment visit could have been done between day 170 and day 190, and a relapse was noted in the chloroquine plus primaquine group at day 186.

Number at risk Chloroquine +tafenoquine 50 mg +tafenoquine 100 mg +tafenoquine 300 mg +tafenoquine 600 mg +primaquine Chloroquine alone 0 30 54 55 56 55 49 52 60 46 46 56 53 44 39 90 43 38 53 50 37 31 120 34 33 50 49 37 26 150 32 30 50 46 37 24 180 8 7 18 11 11 11 190 Study day 0 0·2 0·4 0·6 0·8 1·0

Estimated survival function

Chloroquine+tafenoquine 50 mg Chloroquine+tafenoquine 100 mg Chloroquine+tafenoquine 300 mg Chloroquine+tafenoquine 600 mg Chloroquine+primaquine Chloroquine alone

primaquine (n=50) Chloroquine alone(n=54) 50 mg (n=55) 100 mg (n=57) 300 mg (n=57) 600 mg (n=56)

Eﬃ cacy, % (95% CI) 57·7% (43−70) 54·1% (40−66) 89·2% (77−95) 91·9% (80−97) 77·3% (63−87) 37·5% (23−52) Diﬀ erence versus chloroquine only,

% (95% CI) 20·3% (0−40) 16·6% (−3 to 36) 51·7% (35−69) 54·5% (38−71) 39·9% (21−59) ··

Log-rank test, p value ND 0·158 <0·0001 <0·0001 0·0004 ··

ND=not determined because comparison of higher dose was not signifi cant.

(6)

of 329). The incidence of diarrhoea was higher in the tafenoquine 600 mg group (nine [16%] of 56) than in the other groups (ranging from one [2%] of 57 in the tafenoquine 100 mg group to four [8%] of 50 in the primaquine group), although only one case in the tafenoquine 600 mg groupwas thought by the investi-gator to be drug related. Asthenia occurred only in patients receiving tafenoquine, although no obvious dose eﬀ ect was found and this event was not thought to be drug related. No adverse events leading to study withdrawal occurred, nor did any deaths. Three adverse events led to study drug discontinuation; all were QT prolongation (one case each in the tafenoquine 50 mg, the primaquine, and the chloroquine alone groups).

Overall, 29 serious adverse events occurred in 26 patients (table 6), the most common being asymptomatic QT prolongation (occurring in 11 [3%] of 329 patients). Of these 11 patients, four were in the primaquine group, two each in the tafenoquine 50 mg, tafenoquine 100 mg, and chloroquine alone groups, and one in the tafenoquine 300 mg group. Thus we found no evidence of a dose eﬀ ect of tafenoquine on QT prolongation, and tafenoquine did not seem to have an additional eﬀ ect on the known QT prolongation associated with chloroquine. In no cases was the QTcF (QT corrected using Fredericia’s formula) more than 500 ms.

No blood transfusions were done. Six patients had declines in haemoglobin of more than 25 g/L or at least a 25% decline compared with baseline (four assigned to tafenoquine, one to primaquine, and one to chloroquine

alone; table 7). In fi ve of these six patients the decline in haemoglobin occurred before day 14 and was recorded as a serious adverse event (two events were defi ned as anaemia and three as haemoglobin declines; table 6). In all cases, treatment was continued and haemoglobin recovered to normal concentrations with supportive care. Investigation of these cases showed no evidence of drug-induced haemolysis; the reduction in haemo-globin occurred shortly after initiation of antimalarial therapy and was attributed to rehydration. Genotyping showed that all six patients had no G6PD variants associated with enzyme defi ciency. G6PD genotyping of all female patients identifi ed three that were hetero-zygous for known G6PD-defi cient variants: Aures (chloro quine alone), Mahidol (chloroquine plus tafeno-quine 300 mg), and Kalyan (chlorotafeno-quine alone). None had an adverse event related to haemolysis.

No patients were reported to have keratopathy or clinically important changes in any ophthalmic assess-ment, with the exception of seven (11%) of 61 patients receiving tafenoquine and one each in the primaquine (one [7%] of 15), and chloroquine alone groups (one [6%] of 17) who had postbaseline transient changes in the results of their Humphrey visual fi eld test, all of which had resolved by day 180.

Discussion

This study identifi ed two tafenoquine doses (300 mg and 600 mg) that, when coadministered with chloroquine, had signifi cantly improved relapse-free eﬃ cacy at 6 months

primaquine (n=50) Chloroquine alone (n=54) 50 mg (n=55) 100 mg (n=57) 300 mg (n=57) 600 mg (n=56)

Parasite clearance time, h 45·0 (42−49) 43·0 (38−50) 42·0 (40−45) 47·0 (43−59) 44·5 (39−48) 42·5 (38−48) Fever clearance time, h 11·0 (4−22) 6·5 (4−19) 15·0 (5−21) 19·0 (5−26) 19·5 (4−37) 8·0 (5−16) Gametocyte clearance time, days 2·0 (1·0–2·0) 1·0 (1·0–2·0) 1·0 (1·0–2·0) 1·0 (1·0–2·0) 2·0 (1·0–2·0) 2·0 (1·0–2·0) Results are median (95% CI).

Table 4: Median parasite and fever and gametocyte clearance times

Peru, n 22 24 23 23 22 22

Eﬃ cacy, % (95% CI) 45·5% (23–66) 39·5% (20–58) 81·1% (57–92) 84·0% (58–95) 58·7% (36–76) 12·2% (2–31)

Brazil, n 6 6 6 7 6 6

Eﬃ cacy, % (95% CI) 33·3% (5–68) 33·3% (5–68) 83·3% (27–97) 85·7% (33–98) 83·3% (27–97) 16·7% (1–52)

Thailand, n 16 16 19 16 16 16

Eﬃ cacy, % (95% CI) 60·0% (32–80) 67·3% (38–85) 94·7% (68–99) 100% (100–100) 92·9% (59–99) 56·3% (30–76)

India, n 11 11 9 10 6 10

Eﬃ cacy, % (95% CI) 90·9% (51–99) 80·0% (41–95) 100% (100–100) 100% (100–100) 100% (100–100) 90·0% (47–99) No statistical comparisons were made at the country level.

(7)

Any adverse event 37 (67%) 42 (74%) 38 (67%) 37 (66%) 32 (64%) 41 (76%)

Headache 14 (25%) 17 (30%) 10 (18%) 16 (29%) 14 (28%) 20 (37%)

Chills 16 (29%) 16 (28%) 5 (9%) 9 (16%) 10 (20%) 20 (37%)

Diarrhoea 4 (7%) 1 (2%) 3 (5%) 9 (16%) 4 (8%) 4 (7%)

Pyrexia 18 (33%) 16 (28%) 5 (9%) 7 (13%) 12 (24%) 21 (39%)

Upper abdominal pain 6 (11%) 5 (9%) 6 (11%) 6 (11%) 7 (14%) 5 (9%)

Nausea 7 (13%) 3 (5%) 5 (9%) 5 (9%) 4 (8%) 3 (6%) Asthenia 5 (9%) 4 (7%) 1 (2%) 5 (9%) 0 0 Dizziness 7 (13%) 2 (4%) 5 (9%) 4 (7%) 5 (10%) 5 (9%) Parasitic gastroenteritis 5 (9%) 5 (9%) 3 (5%) 4 (7%) 2 (4%) 2 (4%) Back pain 2 (4%) 0 5 (9%) 4 (7%) 2 (4%) 2 (4%) Insomnia 2 (4%) 3 (5%) 5 (9%) 3 (5%) 3 (6%) 1 (2%) Myalgia 3 (5%) 4 (7%) 1 (2%) 3 (5%) 2 (4%) 3 (6%) Arthralgia 3 (5%) 2 (4%) 2 (4%) 3 (5%) 1 (2%) 1 (2%) Vomiting 3 (5%) 2 (4%) 2 (4%) 3 (5%) 5 (10%) 0

Increased ALT concentration 1 (2%) 1 (2%) 4 (7%) 3 (5%) 4 (8%) 1 (2%)

Pruritus 4 (7%) 8 (14%) 8 (14%) 2 (4%) 3 (6%) 7 (13%)

Urinary tract infection 5 (9%) 3 (5%) 4 (7%) 1 (2%) 3 (6%) 4 (7%)

Cough 3 (5%) 3 (5%) 4 (7%) 1 (2%) 4 (8%) 4 (7%)

QT prolongation 3 (5%) 2 (4%) 3 (5%) 1 (2%) 5 (10%) 4 (7%)

Data are number of patients (% of patients). Data shown are for any adverse events occurring in at least 4% of all patients receiving tafenoquine (n=225), listed in order of the most frequently occurring event in the highest tafenoquine dose group. ALT=alanine aminotransferase.

Table 5: Most frequent adverse events of any cause reported in the safety (intention-to-treat) population

Any serious adverse event, n patients 2 (4%) 6 (11%) 2 (4%) 5 (9%) 7 (14%) 4 (7%)

QT prolongation 2 (4%) 2 (4%) 1 (2%) 0 4 (8%) 2 (4%) Anaemia 0 1 (2%) 1 (2%) 0 0 0 Decreased haemoglobin 0 0 0 1 (2%) 1 (2%) 1 (2%) Haemarthrosis 0 0 0 1 (2%) 0 0 Pyelonephritis 0 0 0 1 (2%) 0 0 Scrub typhus 0 0 0 1 (2%) 0 0 Depressed mood 0 0 0 1 (2%)* 0 0 Nausea 0 0 0 1 (2%)* 0 0 Epigastralgia 0 0 0 1 (2%)* 0 0 Diarrhoea 0 0 0 1 (2%)* 0 0 Pyrexia 0 1 (2%) 0 0 0 0 Dehydration 0 1 (2%) 0 0 0 0

Abortion induced (elective) 0 1 (2%) 0 0 0 0

Acute hepatitis 0 0 0 0 1 (2%) 0

Methaemoglobinaemia 0 0 0 0 1 (2%) 0

Increased ALT concentration 0 0 0 0 0 1 (2%)

Data are number of patients (% of patients). ALT = alanine aminotransferase. *Four serious adverse events occurred in one patient who had a history of depression with concomitant diazepam 10 mg.

(8)

compared with chloroquine alone. At the primaquine dose currently recommended by national treatment guidelines and WHO,13_{the eﬃ cacy rate was numerically}

lower than that of tafenoquine 300 mg and 600 mg. However, even with close monitoring and regular, early follow-up, we are uncertain as to whether all patients were compliant with the 14 day primaquine regimen, and non-compliance might have adversely aﬀ ected eﬃ cacy. Also, the range of total primaquine dose by bodyweight was 2·13–5·25 mg/kg, hence we acknowledge that some patients received less than the 0·25 mg/kg per day (3·5 mg/kg total dose) that they would have received if their dose had been weight adjusted.

On the basis of the results of a study assessing haemolytic risk in G6PD defi ciency,14_{and the lack of}

evidence for any increase in eﬃ cacy between tafenoquine 300 mg and tafenoquine 600 mg, tafenoquine 300 mg has been selected for further clinical studies. Notably, the study was done in a resource-restricted setting, but still achieved high rates of patient retention and protocol adherence during the 6 month follow-up. This study was done only in adults aged 16 years or older; the safety and eﬃ cacy of tafenoquine in children has yet to be assessed.

The study endpoint of relapse-free effi cacy did not distinguish between recrudescence of the initial infection, reinfection, and hypnozoite-induced relapse, and cases of asymptomatic malaria that might have occurred between follow-up dates could have been missed. However, recru-descence is aff ected mainly by the effi cacy of chloroquine, and reinfection rate or asymptomatic carriage by the infective mosquito biting rate, so these factors would have been balanced across all treatment groups. This study included countries with generally low P vivax transmission rates in which chloroquine remains eff ective against

P vivax. Thus, in this setting, hypnozoite-induced relapse

is most probably the major contributor to clinical disease, and the diff erences in relapse-free effi cacy for tafenoquine and primaquine (accrued by hypnozoite eradication) when compared with chloroquine alone were sustained for at least 6 months. Tafenoquine safety and effi cacy in regions of high P vivax transmission, and its use in combination with blood schizonticides other than chloroquine, such as artemisinin-based combination therapy, would need to be investigated in appropriate studies. A study in Papua New Guinea, where chloroquine resistance and reinfection rates are thought to be much higher than in the regions we assessed, reported that a signifi cant reduction in P vivax relapse with artesunate plus primaquine compared with artesunate alone was only maintained for the fi rst 3 months after treatment.15

However, even in this case, hyponozoites represented a substantial burden of clinical disease.15

The study was not powered to show individual country diff erences in relapse rates for tafenoquine or primaquine compared with the chloroquine alone group. However, the relapse rates in the absence of antihypnozoite therapy (ie, in the chloroquine alone group) were notably diff erent across the three regions studied; we observed the highest relapse rates in the two Latin American countries and the lowest in the Indian sites. The small number of patients recruited in Thailand suggest a relapse rate lower than that previously reported for southeast Asia, but still adequate to show the predefi ned level of treatment diff erence.16_{The low relapse rate in India in the}

chloro-quine alone group is consistent with the extended latency for P vivax relapse reported in subtropical and temperate

Panel: Research in context Systematic review

We added to data available on fi le by searching PubMed for articles published between January, 1966, and October, 2013, using the terms “tafenoquine”, “primaquine”, and “Plasmodium

vivax relapses” for all references, which we then reviewed for

relevance. We did not apply any language restrictions.

Interpretation

Our study is the fi rst to test a single tafenoquine dose below 500 mg in the prevention of P vivax relapse. Our results are consistent with the known antihypnozoite activity of tafenoquine, and safety outcomes were as expected from previous studies assessing higher tafenoquine doses. Our results show that tafenoquine has high eﬃ cacy in the prevention of P vivax relapse in a well controlled, comparative clinical trial and support the decision for further development of the 300 mg dose in pivotal clinical trials.

≤15 g/L 48 (87%) 37 (65%) 38 (67%) 33 (59%) 34 (68%) 38 (70%)

15–25 g/L 7 (13%) 19 (33%) 17 (30%) 22 (39%) 15 (30%) 15 (28%)

>25 g/L or ≥25% decline 0 1 (2%) 2 (4%) 1 (2%) 1 (2%) 1 (2%)

Percentage change in haemoglobin

from baseline at day 8, g/L –0·1 (–5·7 to 5·2) –1·0 (–6·3 to 4·3) –0·6 (–8·9 to 4·4) –0·4 (–7·2 to 3·2) –1·4 (–5·9 to 0·8) –2·2 (–6·4 to 2·1) Data are number (%) or mean (IQR).

Table 7: Maximum haemoglobin declines between days 1 and 29 compared with baseline haemoglobin and percentage change in haemoglobin at day 8 compared with baseline haemoglobin

(9)

regions.17_{Modelling of P vivax on the basis of}

surveil-lance data from northwest India provided estimates of 7·1 months for the mean latency and 31% for the relapse rate.18_{Thus, the 6 month assessment was inadequate to}

capture these extended relapse latencies. As expected, excluding the data from India, the diﬀ erences between chloroquine alone and the addition of tafeno quine 300 mg, tafenoquine 600 mg, or primaquine become more pronounced (data not shown).

In previous studies, tafenoquine has been tested for the prevention of P vivax relapse at doses up to 300 mg per day for 7 days and 600 mg per day for 3 days.10,11_The

much lower tafenoquine doses tested in this study did not produce any emergent safety signals of concern.

No adverse events of methaemoglobinaemia with tafenoquine were reported, which is consistent with previous studies at total doses of 600 mg or less.10,11_An

earlier study in soldiers receiving high-dose, long-term tafenoquine prophylaxis reported reversible vortex keratopathy.19_{In our trial, no evidence of keratopathy was}

reported, and although the potential ophthalmic risks of a tafenoquine 300 mg are likely to be small, further confi rmation will be needed from phase 3 studies.

The haemolytic potential of tafenoquine in G6PD defi ciency has been assessed in a separate study, the results of which suggested that at tafenoquine doses of 100 mg or greater, phenotypic G6PD testing is needed before treat-ment.14,20_{Patients with phenotypic G6PD defi ciency were}

excluded from the current study and there was no evidence of haemolysis. This fi nding suggests that when patients who are not at risk of haemolysis caused by G6PD defi ciency can be identifi ed, they can be treated safely with tafenoquine. However, the relative risk of tafenoquine with G6PD testing versus primaquine with or without G6PD testing has yet to be examined.

Coadministration of tafenoquine and chloroquine was not shown to exacerbate the QT prolongation known to occur with chloroquine. This result is consistent with fi ndings from a drug interaction study of tafenoquine and chloroquine that found no evidence of an additional eﬀ ect of tafenoquine on QT.21_{Tafenoquine has been}

further assessed in a thorough QT study.22

As well as generating a large proportion of the clinical disease attributed to P vivax, the persistence of P vivax hypnozoites presents a substantial obstacle for malaria elimination,23_{because relapses also produce gametocytes.}

A recent assessment in Thailand showed that the ratio of gametocytaemia to asexual parasitaemia was similar between acute and recurrent infections.24_Moreover,

unlike P falciparum, P vivax gametocytes appear before the development of malaria symptoms, so patients are infectious before they receive antimalarial treatment. Thus, prevention of the characteristic multiple relapses of P vivax infection is likely to be a key strategy for interrupting transmission of this parasite and reducing the overall burden of malaria, with the ultimate aim of malaria elimination.24

Tafenoquine as a single-dose treatment coadministered with chloroquine eff ectively prevents P vivax relapse, and is a potential candidate for fi rst-choice treatment of uncomplicated P vivax malaria, assuming that safety and effi cacy at least equivalent to primaquine can be shown (panel). However, in a public health context, an eff ective single-dose treatment for P vivax hypnozoite eradication could profoundly aff ect the way that we approach malaria elimination.

Contributors

All authors were involved in development of the study protocol, critically reviewed this manuscript, and approved the fi nal version for

submission. AL-C, MVL, SK, RR, NCh, SKG, SKK, and PA were principal investigators and involved in data acquisition. JJM, SD, JAG, J-PK, LK, and CU provided study oversight. JAG, NCa, and LK contributed to the analysis and interpretation of the data. Confl icts of interest

JJM and SD are employees of Medicines for Malaria Venture. CU, J-PK, NCa, JAG, and LK are employees of GlaxoSmithKline and stockholders in GlaxoSmithKline. All other authors declare that they have no confl icts of interest.

Acknowledgments

We thank the patients and their caregivers for their consent to participate. We also acknowledge the following teams and individuals for their contributions and critical review during the development of this manuscript: the clinical research staﬀ at the Instituto de Medicina Tropical Alexander von Humboldt, Universidad Peruana Cayetano Heredia, Lima, Peru; the Fundação de Medicina Tropical Doutor Heitor Vieira Dourado, Manaus, Amazonas, Brazil; the Faculty of Tropical Medicine, Mahidol University, Bangkok, and Mae Sot Hospital, Mae Sot, Thailand; the MV Hospital and Research Centre, Lucknow, Uttar Pradesh, and the Sarder Patel Medical College, Bikaner, Rajasthan, India. We also thank Sowmya Gopalan and Ramadurai Srinivasan (Sri Ramchandra Medical College and Research Institute, Department of Medicine, Porur, Chennai, Tamil Nadu, India); Raul Chuquiyauri (Universidad Peruana Cayetano Heredia, Lima, Peru); Martín Casapía (Asociación Civil Selva Amazónica, Iquitos, Peru); Dhanpat K Kochar (Sarder Patel Medical College, and Rajasthan University of Health Sciences, Jaipur, India) for providing a basis for estimating the normal range of serum porphyrin level in healthy volunteers, as well as review of the manuscript; and Vijay K Tundwal, Jai K Meel, Anju Kochar, and Gaurav Garg (Sarder Patel Medical College) for data acquisition. At GlaxoSmithKline we thank Sushma Narayan, Viviana Garay, Kim Fletcher, and the TAF112582 study teams in the Peru, Brazil, Thailand, and India GlaxoSmithKline local oﬃ ces. We also thank Naomi Richardson of Magenta Communications (funded by GlaxoSmithKline) for providing assistance in writing the fi rst draft of this publication and editorial assistance in the collation of author contributions.

References

1 Price RN, Tjitra E, Guerra CA, Yeung S, White NJ, Anstey NM. Vivax malaria: neglected and not benign. Am J Trop Med Hyg 2007; 77 (6 suppl): 79–87.

2 Mendis K, Sina BJ, Marchesini P, Carter R. The neglected burden of

Plasmodium vivax malaria. Am J Trop Med Hyg 2001;

64 (1–2 suppl): 97–106.

3 Douglas NM, Anstey NM, Buﬀ et PA, et al. The anaemia of

Plasmodium vivax malaria. Malar J 2012; 11: 135.

4 Battle KE, Gething PW, Elyazar IR,et al. The global public health signifi cance of Plasmodium vivax. Adv Parasitol 2012; 80: 1–111. 5 Wells TN, Burrows JN, Baird JK. Targeting the hypnozoite reservoir

of Plasmodium vivax: the hidden obstacle to malaria elimination.

Trends Parasitol 2010; 26: 145–51.

6 Vale N, Moreira R, Gomes P. Primaquine revisited six decades after its discovery. Eur J Med Chem 2009; 44: 937–53.

7 John GK, Douglas NM, von Seidlein L, et al. Primaquine radical cure of Plasmodium vivax: a critical review of the literature. Malar J 2012; 11: 280.

(10)

8 Dow GS, Gettayacamin M, Hansukjariya P, et al. Radical curative eﬃ cacy of tafenoquine combination regimens in

Plasmodium cynomolgi-infected Rhesus monkeys (Macaca mulatta). Malar J 2011; 10: 212.

9 Brueckner RP, Lasseter KC, Lin ET, Schuster BG.

First-time-in-humans safety and pharmacokinetics of WR 238605, a new antimalarial. Am J Trop Med Hyg 1998; 58: 645–49.

10 Walsh DS, Looareesuwan S, Wilairatana P, et al. Randomized dose-ranging study of the safety and eﬃ cacy of WR 238605 (Tafenoquine) in the prevention of relapse of Plasmodium vivax malaria in Thailand. J Infect Dis 1999; 180: 1282–87.

11 Walsh DS, Wilairatana P, Tang DB, et al. Randomized trial of 3-dose regimens of tafenoquine (WR238605) versus low-dose primaquine for preventing Plasmodium vivax malaria relapse. Clin Infect Dis 2004; 39: 1095–103.

12 Alving AS, Carson PE, Flanagan CL, Ickes CE. Enzymatic defi ciency in primaquine-sensitive erythrocytes. Science 1956; 124: 484–85. 13 WHO. Guidelines for the treatment of malaria, 2nd edn. 2010. http://whqlibdoc.who.int/publications/2010/9789241547925_eng. pdf (accessed Oct 19, 2013).

14 Rueangweerayut R, Bancone G, Beelan A, et al. A phase I study to investigate the haemolytic potential of tafenoquine in healthy subjects with glucose-6-phosphate dehydrogenase defi ciency (TAF110027). 61st annual meeting of the American Society of Tropical Medicine and Hygiene; Atlanta, GA; Nov 11–15, 2012. Abstr 435. 15 Betuela I, Rosanas-Urgell A, Kiniboro B, et al. Relapses contribute

signifi cantly to the risk of Plasmodium vivax infection and disease in Papua New Guinean children 1–5 years of age. J Infect Dis 2012; 206: 1771–80.

16 Luxemburger C, van Vugt M, Jonathan S, et al. Treatment of vivax malaria on the western border of Thailand.

Trans R Soc Trop Med Hyg 1999; 93: 433–38.

17 White NJ. Determinants of relapse periodicity in Plasmodium vivax malaria. Malar J 2011; 10: 297.

18 Roy M, Bouma MJ, Ionides EL, Dhiman RC, Pascual M. The potential elimination of Plasmodium vivax malaria by relapse treatment: insights from a transmission model and surveillance data from NW India. PLoS Negl Trop Dis 2013; 7: e1979. 19 Nasveld PE, Edstein MD, Reid M, et al. Randomized, double-blind

study of the safety, tolerability, and eﬃ cacy of tafenoquine versus mefl oquine for malaria prophylaxis in nonimmune subjects.

Antimicrob Agents Chemother 2010; 54: 792–98.

20 Green J, Harrell E, Narayan S, et al. A clinical summary of investigations to determine the haemolytic potential of tafenoquine in G6PD-defi cient subjects. 62nd annual meeting of the American Society of Tropical Medicine and Hygiene; Washington, DC; Nov 13–17, 2013. Abstr 42B.

21 Miller AK, Harrell E, Ye L, et al. Pharmacokinetic interactions and safety evaluations of coadministered tafenoquine and chloroquine in healthy subjects. Br J Clin Pharmacol 2013; 76: 858–67. 22 Green J, Patel A, Hussaini A, et al. A randomized, placebo-controlled

study to evaluate the eﬀ ect of tafenoquine on the electrocardiogram (ECG), with focus on cardiac repolarization (QTc duration) in healthy subjects. 61st annual meeting of the American Society of Tropical Medicine and Hygiene; Atlanta, GA; Nov 11–15, 2012. Abstr 150. 23 Aguas R, Ferreira MU, Gomes MG. Modeling the eﬀ ects of relapse

in the transmission dynamics of malaria parasites. J Parasitol Res 2012; 2012: 921715.

24 Douglas NM, Simpson JA, Phyo AP, et al. Gametocyte dynamics and the role of drugs in reducing the transmission potential of