Therapeutic efficacy of artesunate-amodiaquine and artemether-lumefantrine and polymorphism in Plasmodium falciparum kelch13-propeller gene in Equatorial Guinea

Equatorial Guinea is located in Central Africa and is divided into two regions: the mainland area, which lies between Cameroon and Gabon, and the island region (Bioko, Annobo'n and Corisco Bay). About 75% of the country's population lives in the mainland. The study was conducted in selected health facilities in three study sites: Malabo in Bioko Island and Bata and Ebebiyin in Litoral and Kié-Ntem provinces in the mainland, respectively (Fig. 1). The health facilities in the study were Regional Hospital in Malabo, Provincial Hospital and Angokong Health Centre in Ebebiyin, Regional Hospital, Maria Rafols Health Centre and Maria Gay Health Centre in Bata. The study was a single arm cohort study that investigated the clinical therapeutic efficacy of ASAQ and AL in the treatment of uncomplicated falciparum infection. Children with uncomplicated falciparum malaria who met the study inclusion criteria were enrolled sequentially, first to ASAQ until sample size was reached and then to AL in each site, and assessed clinically and parasitologically for 28 days, based on the 2009 WHO protocol [20].

Potential study children attending study health facilities from August 2017 to July 2018 were screened and enrolled if they met the following eligibility criteria: age between 6 months and 10 years, axillary temperature ≥ 37.5 °C and/or history of fever in the past 24 h, P. falciparum monoinfection, parasitaemia of 500 to 200,000 asexual parasites/µl, willingness to comply with the study visit schedule, and informed consent from parents or guardians (Fig. 2). Children with exclusion criteria, including the presence of general danger signs or evidence of severe falciparum malaria, mixed or monoinfection with non-falciparum species, severe malnutrition, febrile conditions due to diseases other than malaria (measles, acute lower respiratory tract infection, severe diarrhea with dehydration), or known underlying chronic diseases (e.g., cardiac, renal, or liver disease, HIV/AIDS), received appropriate care and treatment according to national guidelines. In addition, children on regular medication that could have affected the pharmacokinetics of the study ACT and those with a history of hypersensitivity reactions to the medications were excluded.

Children in the ASAQ group received daily dose of ASAQ for three consecutive days based on recommended weight ranges: one tablet and two tablets of 25 mg artesunate/67.5 mg for children weighing ≥ 4.5 to < 9 kg and 9 to < 18 kg, respectively; one tablet and two tablets of 100 mg artesunate/270 mg amodiaquine for children weighing 18 to < 36 kg and ≥ 36 kg, respectively. For children treated et al. doses of AL were administered twice daily for 3 days according to recommended weight: one tablet for children weighing 5–14 kg, two tablets for 15–24 kg, and three tablets for 25–34 kg. All treatment doses were administered under direct observation by the study team and patients were observed for 30 min. If the first dose was vomited, the treatment dose was administered again. If vomited again, the patient received an artesunate injection according to national guidelines and the patient was withdrawn from the study. Study drugs were obtained from the WHO.

Enrolled children were followed for up to 28 days at scheduled visits on days 1, 2, 3, 7, 14, 21, and 28, and at unscheduled visits if symptoms worsened or recurred. Clinical and parasitological examinations were performed at each visit. If the parents/caregivers did not show up for the scheduled visit, a team member visited them at home.

A minimum sample of 73 children per drug per site was estimated, based on the assumption of a 5% treatment failure rate for each drug ASAQ and AL with a 95% confidence level and 5% precision. An additional 20% was added to allow for "lost to follow-up" and withdrawal during the 28-day follow-up period. The target sample size was 88 patients per drug per site.

Thick and thin blood slides were stained with Giemsa and examined microscopically to detect malaria parasites and determine level of parasitaemia based on the WHO procedure [20]. The number of asexual parasites was counted against 200 white blood cells (or per 500 if the count was < 100 parasites/200 white blood cells). A blood slide was confirmed negative if no parasite was seen after counting 1000 white blood cells. Parasite density, defined as parasites per µl, was calculated assuming a leukocyte count of 6000/µl of blood. All blood slides were examined by two independent microscopists. A third microscopist re-examined the slides with discordant results in terms of species diagnosis, parasite density of > 50% or presence of parasites. Final parasite density was calculated by averaging the two closest counts.

Filter paper blood samples were collected from each patient on day-0 and in case of recurrence, on the day of parasite recurrence (from day-7). Samples were dried and stored in individual plastic bags containing desiccant. Each dried blood spot was punched out with a sterile puncher, and the spots were placed in numerical order in a 96-well plate. Parasite DNA was extracted using QIAamp DNA Blood Mini Kit (Qiagen). The DNA samples (day-0 and day of recurrence) were analysed for genotyping of the highly polymorphic regions msp1, msp2 (merozoite surface proteins 1 and 2) and glurp (glutamate-rich protein) loci, as described elsewhere [21]. The genotypic profiles of the parasites at day-0 and day of recurrence were compared to determine whether the recurent infections were a recrudescence (same strain) or a new infection (different strain), according to the current WHO-recommended algorithm [22]. As an explanatory endpoint, reinfection and recrudescence were also determined by the newly proposed two out of three (2/3) algorithm [23]. In this strategy, the classification of recurrent failures is based on a consensus result of msp1 and msp2 and disparate results are resolved by glurp. Such an analysis demands concommittant results from at least two markers for classification of reinfection or recrudescence compared to three with the standard WHO methodology. If results for only 2 markers were available and results for a third marker were missing, the PCR correction was classified as undertermined. Samples were analysed at Institut Pasteur in Paris, France.

DNA extracted from day-0 dried blood spots were analysed to detect the presence of mutations in the propeller domain of Pfkelch13 gene (PF3D7_1343700) previously described to be associated with artemisinin resistance [4]. PCR amplification (codons 440–680, 720 bp) was performed using the method described by Ariey et al. [4]. For the inner round, five μl DNA was amplified with 0.25 μM each primer, 0.2 mM dNTP, 2.5 mM MgCl2, and 1.25 U Taq DNA polymerase (Solis Biodyne, Estonia), in 25 μl volume using the following cycling program: 15 min at 95 °C, then 35 cycles of 30 s at 95 °C, 2 min at 58 °C, 2 min at 72 °C, and final extension 10 min at 72 °C. For the outer PCR round, 5 μl of primary PCR products were amplified under the same conditions, except for annealing and extension (1 min). PCR products were detected using capillary gel electrophoresis (Fragment Analyzer, Agilent, France). Double strand sequencing of PCR products was performed by Eurofins (Germany). Sequence polymorphisms were identified with the CLC Main Workbench 20 software (Qiagen) by using the 3D7 strain of P. falciparum (PF3D7_1343700) as a reference sequence. Electropherograms with mixed alleles were considered as mutant for the purpose of mutation frequency estimation. The quality control was assessed by including three blinded quality-control samples (wild type, C580Y and R539T) in each 96-well sequencing plate.

Treatment response was classified as early treatment failure (ETF), late clinical failure (LCF), late parasitological failure (LPF) and adequate clinical and parasitological response (ACPR) before and after PCR correction using the WHO protocol [18]. The primary study endpoint was PCR-corrected adequate clinical and parasitological response. Secondary endpoints included parasitaemia at day-3 and PCR-uncorrected treatment failure.

The study was approved by the Ministry of Health and Social Welfare of Equatorial Guinea (No. 731–150), the Ethics Committee of Spanish National Health Institute, Carlos III (CEI PI 57_2016-v3) and the WHO Research Ethics Review Committee (ERC.000286). Parents or guardians were informed about the study procedure, its benefits and potential risks, and their consent to enroll their children was obtained in writing before enrollment. They gave their written consent before enrolling their children in the study. If a patient, parent, or guardian was illiterate, he or she chose a witness to co-sign the consent form.

Data were double entered, cleaned, and analysed using the software programme WHO excel (http://​www.​who.​int/​malaria/​publications/​atoz/​9789241597531/​en/​). Enrolled patients who were lost to follow-up or withdrawn from the study, had recurrent parasitaemia with new infection, or unknown PCR (indeterminate or missing) were excluded from per-protocol analysis. However, these cases were included in the Kaplan–Meier analysis up to the day of loss or withdrawal from the study. Descriptive statistics including percentages, mean, standard deviation, and range were used. Student's t-test was used for analysis of continuous variables (parasite density and age) and Fisher's exact test was used for categorical data. A p-value of < 0.05 was considered significant.

Of the 1725 children screened, 490 were enrolled in the study: 264 were treated with ASAQ (88 in Malabo, 88 in Bata and 88 in Ebibeyin) and 226 with AL (50 in Malabo, 88 in Bata and 88 in Ebibeyin). The target sample size per drug and per site (n  =  88) was achieved except for the group treated with AL in Malabo site due to low malaria transmission. Baseline characteristics (age, axillary temperature, parasitaemia) of the study children in the different sites and treatment groups were comparable (Table 1).

Table 2 shows the treatment outcomes as per-protocol and Kaplan Meier analysis before PCR correction. For patients treated on ASAQ, per-protocol analysis of uncorrected PCR data showed ACPR of 100% (95% CI 94.6–100%), 96.3% (95% CI 89.6–99.2%) and 88.8% (95% CI 79.7–94.7%) in Malabo, Bata and Ebebiyin sites, respectively. Among patients treated with AL, per-protocol analysis of uncorrected PCR data revealed ACPR of 92.9% (95% CI 80.5–98.5%), 95.3% (95% CI 88.4–98.7%) and 73.5% (95% CI 62.7–82.6%) in Malabo, Bata and Ebebiyin, respectively. For the 41 paired samples, msp1, msp2 and glurp results were available for 95.1%, 93.9% and 95.1%, respectively. Per-protocol analysis of PCR corrected data (using the WHO protocol) showed ACPR of 100% (95% CI 94.6–100%), 100% (95% CI 95.4–100%) and 98.6% (95% CI 92.5–100%) in patients treated with ASAQ in Malabo, Bata and Ebebiyin, respectively (Table 3). In children treated with AL, PCR corrected ACPR of 95.1% (95% CI 83.5–99.4%), 100% (95% CI 95.5–100%) and 92.4% (95% CI 83.2–97.5%) were in Malabo, Bata and Ebebiyin, respectively. Based on Kaplan Meier survival analysis, PCR-corrected cumulative cure rates varied from 98.7% to 100% for ASAQ and from 93.7% to 100% for AL (Table 3). Study children at the Ebebiyin site had a higher rate of new infections (9.1% and 19.3% in the ASAQ and AL groups, respectively), compared to less than 5% at the other sites (Table 3). However, the difference was significant only for the Ebebiyin group AL (Fisher's exact test: p  =  0.01). Children in both treatment groups were parasite free on day-2 except one case in the ASAQ group in Bata (1.2%), and all of them cleared parasitaemia by day-3.

Using the 2/3 algorythm, PCR correction remained the same in 22/41 (53.7%) recurrences, changed to recrudescence in 12/41 (29.2%) recurrences, and was undertermined for 7/41 (17.1%) recurrences (Table 3). The failure rate increased significantly in Ebibeyin for both treatment arms (5.2% for ASAQ and 8.8% for AL). No significant changes were observed for the other two sites.

Among the 490 patients, three samples were missing and 11 gave non-interpretable results. Most of the samples with interpretable results (98.1%), carried Pfkelch13 wild type allele and only three (0.6%) had non-synonymous mutations (two carried A578S and one with E433D, Table 4). None were previously linked with artemisinin resistance. Of the 8 patients with recrudescent parasite, only four samples were successfully sequenced, and all carried Pfkelch13 wild type allele.