The study was conducted in the district of Kolokani, Mali. The district of Kolokani is an administrative subdivision of the region of Koulikoro. The town of Kolokani is located approximately 140 km north of Bamako. The district covers 14,380 km2 and is divided into 22 health sub-districts. Each health sub-district is composed of several villages and has a health care center under the responsibility of a physician or chief-nurse and is staffed with at least two or three midwives and vaccinators. Three physicians ensure the coordination of health activities at the district level. The total population was 214,509 inhabitants. The children under one year of age represent about 4% of the total population. Malaria transmission is seasonal (July to November) with parasite prevalence in children 0-5 years of 45% at the end of the no-transmission season and above 70% at the end of the transmission season. First-line treatment of uncomplicated malaria was artesunate-amodiaquine or artemether-lumefantrine. Quinine was recommended for the treatment of severe malaria and SP was recommended prophylaxis for pregnant women.
Study participants and design
The study was a cluster-randomized trial. Twenty-two health sub-districts were randomized into two groups in a 1:1 ratio and IPTi was implemented in 11 health sub-districts (intervention zone) while the other 11 health sub-districts served as the control (non-intervention zone). The implementation of IPTi started in December 2006 and consisted of the administration of ½ tablet of SP alongside the DTP2, DTP3 and Measles routine immunizations. Afridoxine® (manufactured by Fourrts Laboratories Pvt. Ltd, Tamil Nadu, India), a breakable and dispersible form of SP was used after quality controls at the Laboratoire National de Santé in Bamako, Mali. Two cross-sectional surveys were performed. The first survey was performed in May 2007 a few months after the start of the IPTi-implementation, but before the start of the transmission season to collect baseline prevalence of the molecular markers of SP resistance. The second cross-sectional survey was conducted after one year of implementation in December 2007 after the end of the transmission season, to assess the impact of IPTi on the selection of SP-resistant dhfr and dhps mutations in P. falciparum. Children aged 0-5 years randomly selected in each of the health sub-districts were screened for malaria parasitaemia using thick smears obtained by finger pricking. Filter paper blood dots were also collected at the same time for the analysis of SP resistance molecular markers.
Thick smears were Giemsa-stained and examined by microscopy. Parasite density was determined by counting parasites against 300 leucocytes using a conversion factor of 25, based on an assumed leukocyte count of 7,500/μl. Filter papers samples were air-dried and stored at room temperature in labeled individual envelopes. Preliminary genotyping of one hundred samples demonstrated that samples with a parasitaemia less than 500 parasites/μl did not consistently yield PCR product. Therefore, only filter paper samples from children with P. falciparum mono-infection parasitaemia greater or equal to 500 asexual stages per μl of blood were analysed by PCR for dhfr and dhps point mutations. Because of a higher than expected number P. falciparum mono-infection positive samples, 427 samples of the baseline survey and 923 samples of end of one year implementation survey were randomly selected and tested by PCR. Of the 427 baseline survey samples, 274 samples (150 in the intervention zone and 124 in the non-intervention zone) were successfully genotyped for all three dhfr and two dhps mutations, while the remaining were successfully tested for one, two, three or four mutations. Of the 923 randomly selected samples of end of one-year implementation survey, 742 samples (379 in the intervention zone and 363 in the non-intervention zone) were successfully genotyped for all three dhfr and two dhps mutations, and the remaining for one, two, three or four mutations.
A simple DNA extraction method for filter paper strips was used . Briefly, approximately 1 × 2 mm piece of blood-soaked filter paper was placed in 50 μl of methanol for 15 min. The methanol was poured off and the paper heated at 95-100°C in 50 μl of water for 10 minutes. Five microliters of the resulting solution was used as PCR template for the first round of the Nested PCR. One microliter of the product of the first PCR was amplified in the second round of PCR. Samples that failed to yield PCR amplification with the methanol method were re-extracted using an alternative chelex-based method . Dhfr and dhps genotypes were determined using nested mutation-specific PCR and/or PCR-RFLP according to published methods [16, 23]. Dhfr mutations at codons 51, 59 and 108 and dhps mutations at codons 437 and 540 were analysed. One third of the PCR gels were examined by a senior scientist external to the study. The conformity of the gels with standard criteria for PCR gel validation and the fidelity of the transcription of lab notebook (source documents) with electronic files were checked. Error cases were corrected and PCR analyses repeated as needed.
The primary endpoint was the frequency of quadruple mutant of molecular markers (dhfr 51, 59 and 108 and dhps 437) after one year of IPTi implementation. Secondary endpoints were the frequencies of single mutations as well as the triple mutants (dhfr 51 + 59 + 108) determined by PCR at the end of one year of implementation and frequencies of these individual and multiples dhfr and dhps mutations at baseline. The frequency of a particular mutant was calculated as the proportion of the specific mutant samples among the total number of samples successfully analysed for this mutation. Similarly, the frequencies of triple, quadruple and quintuple mutants were determined as the proportion of subjects with the three, four and five mutations among the total numbers of samples tested for the each.
Assuming a proportion of the primary endpoint (quadruple mutants, dhfr 51, 59 and 108 and dhps 437) of 8% in control zone and 14% in IPTi intervention zone, one side p-value 0.025, a power of 80%, and inter class correlation using 0.01, a sample size of 380 children with positive parasitaemia in each group with a minimum of 22 subjects and average of 35 subjects per cluster were needed. Assuming a prevalence of P. falciparum parasitaemia of 35% and about 10% lost or unsuccessful samples, about 110 subjects were need to be screened per cluster resulting in a total of 2,420 children for the two arms.
Cases of mixed infection (wild type and mutant) were categorized as mutant throughout the analysis. Prevalence of single mutations as well as the triple mutant (dhfr 51 + 59 + 108) and quadruple mutant (triple mutant + dhps 437) genotypes were determined and compared between the IPTi implementation and control zones with 95% confidence intervals around the odds ratios adjusted for non-independence within health sub-districts using cluster option in Stata (version 9, Houston Texas USA). Proportions of categorical variables were compared using Pearson's Chi square test. Student t test was used for the comparison of mean ages and Mann-Whitney test for the comparison of parasite densities between the two zones.
Ethical clearance was obtained from the Ethical Committee of the Faculty of Medicine Pharmacy and Dentistry of the University of Bamako, Mali, and written informed consent from a parent or legal guardian was obtained prior to the enrollment of the child in the survey.