Study site, design and ethics
The study was conducted in June 2006 in Fukayosi Primary Health Care Centre, Bagamoyo District, Tanzania. The study site is located in a rural area, where P. falciparum is the predominant malaria species and Anopheles gambiae complex the main vector. At the time of the trial malaria transmission was high, i.e. with an overall slide positivity rate of approximately 40%, and perennial with peaks related to the seasonal rainfalls. The timing of the trial was chosen to coincide with the local yearly peak in malaria transmission following the main rainy season in March-May.
By the time of the trial sulphadoxine-pyrimethamine was the first- and amodiaquine second-line treatment for uncomplicated malaria in the study area. No artemisinin-containing therapies were available through the government health care services in the district during the conduct of the trial.
The primary objective of the study was to explore parasite population dynamics in children with acute uncomplicated malaria during 72 h after diagnosis and initiation of anti-malarial drug treatment. Secondary objectives included determining parasite clearance according to microscopy and PCR. The primary end-point was to compare the total number of P. falciparum genotypes detected pre-treatment with the accumulated number of genotypes identified per patient in all blood samples collected during the trial. The study was considered as exploratory, which precludes a sample size calculation. A sample of 50 children was predefined.
The trial was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice. It was granted ethics clearance by the National Institute for Medical Research, Dar es Salaam, Tanzania, and the Regional Ethics Committee, Stockholm, Sweden. Informed consent was obtained from parents/guardians of all children enrolled. The study is registered at Clinical Trials [
12], with identifier NCT00336375.
Patients and procedures
Patients were recruited among children presenting at the study site with symptoms and signs compatible with acute uncomplicated malaria. Inclusion criteria were: a finger-prick screening blood slide with presence of P. falciparum asexual parasites with a density of 2,000-200,000/μL; age 1-10 years; axillary temperature ≥ 37.5°C or a history of fever in the preceding 24 h; haemoglobin (Hb) ≥ 70 g/L; informed consent; and a successful insertion of an intra-venous cannula in an ante-brachial vein. Exclusion criteria were: any cause of fever other than malaria; symptoms/signs of severe malaria; any danger signs; any serious underlying disease; intake of anti-malarial drugs within the last 14 days; known allergy to artemether-lumefantrine; and daily maximum number of five enrolments completed due to limited number of beds available in the study site.
The inclusion criteria used in this trial divert slightly from the general standards used in efficacy trials of anti-malarial drugs conducted in Africa, which usually enrol children <5 years of age with Hb ≥ 50 g/L [
13]. However, the adjustments in inclusion criteria made with regards to age and Hb in this study were justified by the frequent blood sampling and the need of insertion of an intra-venous cannula.
All children excluded during the screening phase were referred to the clinician in charge of the health facility, who treated the patients according to standard practice and national guidelines.
The enrolled patients were hospitalized and randomly assigned treatment with artemether-lumefantrine (Coartem; Novartis), either accompanied with or without intake of 200 ml of full cream milk together with each drug dose. The purpose of the randomization procedure was for a pharmacokinetic and pharmacodynamic study, which has been reported elsewhere [
13], where all patients participated in parallel. In the present report all enrolled patients have been analysed as a single group.
All drug doses were administered by a study nurse under supervision of the study physician at 0, 8, 24, 36, 48 and 60 h. Children weighing 5-14 kg, 15-24 kg and 25-34 kg were given one, two and three tablets per dose, respectively, each tablet containing 20 mg artemether and 120 mg lumefantrine. If a child vomited within 30 min after ingestion of the study drug, the dose was repeated.
In-patient follow-up duration was 72 h. All clinical data were recorded in a Case Record Form (CRF). Clinical assessment and blood sampling were performed in total 11 times for each participant, i.e. at the time of diagnosis (-2 h) and 0, 2, 4, 8, 16, 24, 36, 48, 60 and 72 h post-treatment. The blood sampling always included both a thick blood smear for microscopy and collection of circa 100 μL of blood on a filter paper (3 MM; Whatman) for genotyping. The screening blood sample (-2 h) was collected from a finger-prick (capillary blood). To reduce the discomfort due to the multiple blood sampling during the trial an intra-venous cannula was inserted in an ante-brachial vein just prior to enrolment and initiation of treatment (0 h), from where the remaining blood samples were to be collected (venous blood). Thus, the two pre-treatment samples (-2 and 0 h) arose from different sampling routes, capillary and venous, respectively, whereas the remaining nine blood samples collected after initiation of treatment arose from the venous route.
Giemsa-stained thick blood slides were examined directly in the field by an experienced microscopist by counting parasites against 200 white blood cells (WBC). Parasite density was quantified (parasites/μL) by assuming an average of 8,000 leucocytes per μL of blood. A negative blood slide was defined as the absence of any asexual parasite after examining 200 oil-immersion fields. Quality control was performed for all blood slides centrally (Ifakara Health Research and Development Centre, Bagamoyo Research and Training Unit) by an independent microscopists blinded to the initial blood slide results.
All blood samples for genotyping were collected on filter papers, thoroughly dried and put in individual zipper plastic bags. After completion of the field trial they were transported to Karolinska Institutet, Sweden, where the genotyping was performed.
At discharge from the study all participants received an insecticide treated net free of charge.
Molecular analyses
DNA was extracted from all blood samples collected on filter paper using the ABI PRISM
® 6100 Nucleic Acid PrepStation (Applied Biosystems, Fresno, CA, USA) as previously described [
14]. Nested PCR genotyping was performed both for the
msp1 and
msp2 marker, considered to be the two most informative genetic markers for assessment of multiplicity of infection [
15,
16]. The respective initial amplification was followed by individual nested PCR reactions using family specific primers for
msp1 (KI, MAD20 and R033), and
msp2 (FC27 and IC), according to previously described standard protocols [
17]. The
msp1 and
msp2 PCR products were loaded on 3% and 2% agarose gels, respectively, stained with ethidium-bromide, separated by electrophoresis and visualized under UV trans-illumination (GelDoc
®, Biorad, Hercules, Ca, USA). All blood samples collected from the same patient were processed in parallel during all aspects of the laboratory work, from DNA extraction to separation of PCR products on agarose gels. At least 2 positive controls from different
P. falciparum laboratory strains as well as 1 negative control were incorporated in each PCR run.
An adequate PCR outcome was defined for each patient, as the presence of amplified DNA in at least one genetic marker in both pre-treatment samples (-2 and 0 h), together with detectable PCR products in the positive controls. For patients not fulfilling this definition a repeat PCR run including all blood samples was performed with an increased DNA template in the initial amplification only, i.e. 3 μL instead of 1 μL in a total reaction volume of 20 μL. If an adequate PCR outcome could not be retrieved after this second PCR attempt, DNA was re-extracted and the PCR analysis repeated according to the initial protocol.
Analyses of number of genotypes and size polymorphism were digitalized using Quantity One® software (Biorad, Hercules, Ca, USA). Individual genotypes were defined by binning 20 base-pairs intervals together. Initially, the median genotype for each family of the respective genetic markers was identified. The absolute size of the identified median band +/- 10 base-pairs formed the initial bin. Thereafter, each 20 base pair interval below and above the median band were defined as representing a distinct genotype.