Study population
Patients with uncomplicated vivax malaria were included in the study, after giving informed consent, if they met the following inclusion criteria: age between 18 and 70 years; weight between 50 and 90 kg, P. vivax mono-infection confirmed by microscopy, asexual parasite count > 250/μL, axillary temperature ≥ 37.5 °C or a history of fever during the past 48 h, and haemoglobin > 7.0 g/dL. Exclusion criteria were: malaria treatment in the previous 63 days; signs of severe malaria; concurrent other febrile conditions or chronic disease (such as severe cardiac, hepatic or renal disorders or HIV); the use of any medication known to interfere with anti-malarial pharmacokinetics; previous history of intolerance to any study drug; known glucose-6-phosphate deficiency; pregnancy confirmed by urinary human chorionic gonadotropin (hCG) testing; and breastfeeding.
Study design and drug administration procedures
This prospective, randomized, open label, three-arm efficacy study of uncomplicated vivax malaria was conducted according to World Health Organization (WHO) methods for surveillance of anti-malarial drug efficacy [
18] at two centres in the Amazon Region of Brazil: the Tropical Medicine Research Centre (CEPEM) in Rondônia and Tropical Medicine Foundation Dr Heitor Vieira Dourado (FMT-HVD) in Manaus.
A 90% cure rate or greater is considered by the WHO to be sufficient evidence of efficacy to support the choice of a specific regimen by National Malaria Control Programmes in their treatment guidelines [
18]. The sample size was calculated with an expected failure rate of 5%. 88 patients were included in each study arm to achieve a precision of 5% and allowing for 20% loss to follow-up, leading to a total of 264 patients.
A randomization list using blocks of six and allocation rate 1:1:1 was generated by software (Etcetera, version 2.72). Sequentially numbered (0–176 CEPEM and 177–264 FMT) opaque sealed envelopes were provided to the local clinical coordinators and used to randomize patients. Differences in dosing schedules and the difficulty of dummy blinding meant that neither patients nor healthcare workers were blinded, but microscopists were not aware of treatment allocation. The statistician was blind to the treatment allocation until the database had been locked.
Eligible patients were allocated to one of the following three treatment groups: (a) chloroquine (CQ); (b) fixed dose combination of artesunate and mefloquine (ASMQ) and; (c) fixed dose combination of artemether and lumefantrine (AL). All three arms received the same primaquine regimen.
Group A received chloroquine (Farmanguinhos—Fiocruz, Batch Numbers 12080940 and 14060467) 600 mg on day 1, and 450 mg on days 2 and 3. This is the Brazilian Ministry of Health current recommendation for uncomplicated malaria vivax treatment [
13].
Group B received two tablets daily for 3 days of a fixed dose combination of 100 mg + 200 mg artesunate and mefloquine (ASMQ) tablets (Farmanguinhos—Fiocruz Batch Numbers 11100680 and 13040348) in a total of six tablets.
Group C received four tablets twice a day for 3 days of a fixed dose combination of 20 mg + 120 mg artemether and lumefantrine (AL) tablets (Coartem®—Novartis, Batch Numbers F2618 and K30711) in a total of 24 AL tablets.
All groups also received two tablets of 15 mg primaquine (Pq) (Farmanguinhos—Fiocruz, Batch numbers 12010038 and 13030282) for 7, 8 or 9 days (according to three weight ranges; ≥ 50–69; 70–79; 80–90 kg, respectively), as recommended by National Malaria Control Programme treatment guidelines [
13]. Patients received a total primaquine dose between 3.0 and 4.2 mg/kg.
The first 3 days’ treatment were supervised for Group A and B. For Group C, only the first daily AL dose was directly observed. Patients in the AL arm were asked about their adherence to the previous second daily dose after the morning supervised dose. For all groups, the first dose was administered after diagnosis, and all subsequent supervised doses were taken between 8 and 10 a.m. If the patient vomited within 30 min after a dose, the same dose was administered again. At D7 follow-up visit, patients were inquired about their treatment adherence to unsupervised primaquine on days 4 through 7. The use of other treatments was recorded at every visit.
Glucose-6-phosphate dehydrogenase (G6PD) deficiency screening was not performed: this is not routinely performed in Brazil and it is not required by the national treatment guidelines. At enrolment, patients were asked about adverse events during previous primaquine use.
Efficacy and safety evaluations
Patients were assessed on the day of enrolment and on days 1, 2, 3, 7, 14, 21, 28, 42 and 63 days after study inclusion. The scheduled study procedures comprised a full history, physical examination, and urinalysis at enrolment and assessment for clinical signs and adverse events (AE) at every follow-up visit. Blood samples were collected for parasite counts at 0, 3, 7, 14, 21, 28, 42 and 63 days, and haemoglobin was measured at 0, 14, 28, 42 and 63 days. An additional blood smear was also collected whenever treatment failed. Data was double entered using an electronic clinical record form (OpenClinica Community, version 3.1.3.1). Analysis was performed using R (version 3.2.5).
The primary efficacy endpoint was the proportion of the population with an adequate clinical and parasitological response (ACPR) at 63 days. Early treatment failure or late treatment failure were classified in line with standard WHO methodology [
18]. The primary analysis was per protocol (PP); an intention-to-treat (ITT) analysis was also performed. The PP population excluded any participant with a protocol violation. Patients who missed the 28 or 42-day visit, but who had an ACPR at the subsequent follow visit was considered as a success at the previous visit. In the ITT population, protocol violations or losses to follow-up were considered as parasitological and clinical failure.
The secondary efficacy endpoints included the success rate at day 3 (72 h after first drug administration) as well as gametocyte clearance, fever clearance, and the cumulative success rate at days 28, 42 and 63.
The cumulative success rate by day 63, i.e. the probability of remaining parasite-free at day 63, was calculated using a Kaplan–Meier survival curve. Categorical variables were summarized using frequencies and percentages, while for quantitative variables, means, standard deviations (SD) and maximum–minimum values were used. Parasite counts were presented using geometric means.
The study was not designed to compare outcomes between treatment arms. However, exploratory analyses were conducted to explore the impact of the treatment arm on the safety and efficacy outcomes. The proportions of categorical variables were compared using Pearson’s Chi squared test with Yates’ continuity correction at significance level of 5%. Nonparametric tests of Wilcoxon and Kruskal–Wallis were used for continuous variables. Additionally, generalized linear models (binominal and Poisson distribution, respectively) were used to estimate the effect of predictors (baseline characteristics and the use of other medications) on treatment success at day 63 and the numbers of AEs that were possibly, probably or highly probably related to the treatment.
A safety analysis was conducted in the ITT population describing frequency, causality, and severity of AEs in each treatment arm. AEs reports were subdivided into: serious AEs, AEs leading to treatment suspension, and AEs which were described as possibly, probably or highly probably related. The mean haemoglobin results at baseline and follow-up are also presented.
Patients were encouraged to seek unscheduled assessments if any AE was suspected. All clinical or laboratory abnormalities were categorized as Grade I to IV according to the Common Terminology Criteria for Adverse Events (CTCAE) of the National Cancer Institute [
21]. Any suspected serious AE (standard definitions) was reported to the sponsor and the Ethical Review Committee. Recognized drug-related events were recorded as an AE, even if it could be related to malaria.
Parasitological densities were estimated using Giemsa-stained blood slides at a magnification of 1000× using WHO recommended methods [
18]. Two trained microscopists read slides independently. The final density was calculated as the mean of the two readings. A third microscopist examined slides if the two readers disagreed over whether there were parasites present, the species, or the parasite density (more than 50% difference). In such cases, the final density was considered as the average of the two closest counts. A slide was considered as negative only after examining 1000 leucocytes in microscopic fields. Gametocyte presence was also recorded.