The study was conducted from September to November 2005 in the research zone of the Centre de Recherche en Santé de Nouna (CRSN) in Nouna Health District, north-western Burkina Faso. The Nouna area is a dry orchard savannah, populated mainly by subsistence farmers of different ethnic groups. Malaria is holoendemic but highly seasonal with most cases occurring during or briefly after the rainy season which lasts from June until October 39. Formal health services are limited to a few rural health centres and the district hospital in Nouna town (25,000 population). CQ as well as SP were shown to be still quite effective in the rural, but not in the urban CRSN study area 3040. Despite an official policy change from CQ to ACT as national malaria first-line therapy in early 2005, CQ has remained the main drug used for fever treatment in the area until today 37.

The study was designed as an uncontrolled efficacy trial in two different areas (urban and rural) of Nouna Health District. The study was open label, with blinding only for the laboratory technicians being responsible for parasite determination in blood smears.

The primary objective of the study was to determine and compare the day 14 efficacy of AQ in the treatment of young children with uncomplicated falciparum malaria in the rural and urban CRSN study area. Secondary objectives were to determine and compare the day 28 efficacy of AQ in the rural and urban CRSN study area and to determine the safety of AQ in the treatment of young children with uncomplicated falciparum malaria in Burkina Faso.

After information of the study communities, mothers with febrile young children were invited for examination by the study team to specific fever measurement points (public buildings). Out of 420 children screened, a total of 117 children were enrolled, 62 from two purposely (good accessibility) chosen villages in 10 and 20 km distance from Nouna town respectively, and 55 from two randomly chosen city quarters of Nouna town. It was initially planned to include 120 children, 60 from the urban and 60 from the rural study area. This sample size would allow for a 20% loss to follow-up during the determination of an overall 40% failure rate (95% confidence level) at day 28 at a precision of 10%, as well as the detection of a major difference in AQ resistance between the rural and urban study areas.

Inclusion criteria were: age 6–59 months, uncomplicated falciparum malaria (axillary temperature ≥ 37.5°C and ≥ 2,000 P. falciparum asexual parasites per μl blood), and written informed consent given by the parents/caretakers. Exclusion criteria were complicated or severe malaria (repeated vomiting, seizures or other neurological impairment, haemoglobin < 7 g/dl or haematocrit < 21%), any apparent significant disease (e. g. pneumonia, meningitis, hepatitis, severe diarrhoea, measles, severe malnutrition), and malaria treatment with western drugs and/or antibiotics with antimalarial potency during last seven days except chloroquine.

Study children received a total dose of 25 mg/kg oral AQ divided over three days (first and second day: 10 mg/kg, third day: 5 mg/kg; AQ tablets or syrup, essential drug stock, Burkina Faso Ministry of Health). All patients were directly observed for at least 30 min following each treatment. If vomiting occurred once, treatment was repeated. In case of repeated vomiting, the participant was withdrawn from the study. Children having fever ≥ 38.5°C received a standard dose of paracetamol tablets (= acetaminophen; 10 mg/kg every six hours) until symptoms subsided. Follow-up of study children was for 28 days using a slightly simplified version of the latest WHO protocol on drug efficacy testing 41: A finger-prick blood sample was taken on days 0, 2, 14 and 28. From this, malaria parasitaemia (thin and thick blood smears) as well as haematocrit values were determined using standard CRSN procedures 39. All malaria slides were examined by two experienced laboratory technicians supervised by one of the investigators (BC). Asexual parasites were counted on thick blood films against 200 white blood cells (WBCs) and parasite density was calculated assuming an average WBC count of 10.000/μL. Slides were declared negative if no parasites were seen in 400 fields on the thick film. For quality control a 10% random sample of blood films is regularly cross checked at the laboratory of the Heidelberg School of Tropical Medicine 39. Follow-up procedures were carried out through trained field workers who visited study children at follow-up days in their homes. Self-reported adverse events were registered through a standard questionnaire during follow-up visits.

In isolates obtained at recruitment and at treatment failure, the P. falciparum crt K76T mutation associated with resistance to CQ, and limitedly AQ 424344454647, was determined at the Institute of Tropical Medicine and International Health in Berlin (Germany) by PCR-RFLP 42. Isolates comprising both, mutant and wildtype alleles were considered mutant. Differentiation between recrudescence and re-infection took place through determination of msp1 and msp2 genotypes in paired filter paper blood samples from day 0 and day 14 as well as day 28 and from any other day after day 7 when a child had developed malaria according to standard methodology 48.

Treatment outcome was categorized according to the latest WHO definitions for early treatment failure (ETF), late clinical failure (LCF), late parasitological failure (LPF) and adequate clinical and parasitological response (ACPR) 41. Losses to follow-up and dropouts due to other reasons were considered as ETF or LCF depending on the time of dropout in an intention-to-treat manner.

The primary analysis parameter was the rate of clinical failures on day 14. Secondary parameters were the rate of clinical failures on day 28, the rate of early clinical failures, the rate of late parasitological failures (day 14 and day 28), and the rate of self-reported adverse events. Efficacy results were compared before and after PCR correction for reinfections. Data were analysed in the overall group of study children and for rural (n = 62) and urban (n = 55) study children separately.

The Chi square test (Chi) was used to compare rates, and exact binominal 95% confidence intervals (95%CIs) were calculated. The non parametric Wilcoxon-Mann-Whitney test (WMW) was applied to compare metric or ordinal data. The McNemar test was used to compare the proportions of pfcrt K76T at recruitment and at treatment failure. The statistical calculation was performed with SAS release 8.02 (SAS^® Institute Inc, Cary, NC, USA).

The study protocol was approved by the local Ethics Committee in Burkina Faso.

117 children with uncomplicated malaria were enrolled, 62 children from the rural and 55 from the urban study area, respectively (Table 1). The analysis is based on these 117 children. This is the a priori defined full analysis set (FAS) for the intention to treat analysis. There was no difference in the distribution of sex between the children from rural and urban area but children from the urban area were younger compared to children from the rural area (means, 26.4 months vs. 33.8 months, P < 0.001). Roughly one quarter of children had been treated for the current disease episode with CQ prior to inclusion, with no major differences between children of urban and rural residence. More children of the urban (63%) compared to the rural area (46%) carried parasites with the pfcrt K76T mutation (P = 0.06). There were six losses to follow-up, all but one were from the urban study area. The reason for the five urban cases was withdrawal by the parents who were dissatisfied with repeated blood taking. The reason for the rural case was a treated fever episode during follow-up without malaria parasitaemia.

There was no serious adverse event and no death over the 28 days follow-up period. Reported minor adverse events were 16 overall, eight in the rural and eight in the urban study children. Gastroenteritis was the main complaint in seven of the study children.

There was no ETF. The PCR-uncorrected rate of ACPR was 103/117 (88%) in the overall study group by day 14. This figure decreased to 28/117 (24%) by day 28. After PCR correction for re-infections, rates of ACPR were 108/117 (92%) and 71/117 (61%) by day 14 and day 28 respectively (Table 2 and Table 3). There were no significant differences in efficacy parameters between urban and rural study children.

Pfcrt K76T did not predict AQ treatment failure: children harbouring parasites with this mutation revealed a PCR-uncorrected (day 28) rate of ACPR of 26% (15/57) as compared to 25% (13/59) among children infected with wild-type parasites (P = 0.88; withdrawals excluded from analysis). Correction for re-infections did not change this finding (70%, 40/57 vs. 60% 31/52, P = 0.25). However, all (77/77) parasites isolated from children at AQ treatment failure exhibited the pfcrt K76T mutation as compared to 52% (40/77) at baseline (P < 0.0001). This was true for recrudescent infections (n = 33) and re-infections (n = 41; both, P = 0.0001).