Plasmodium falciparum isolates obtained from children participating in two related, randomized IPTi trials conducted in Tamale, hyperendemic northern Ghana, and in Afigya Sekyere district, holoendemic southern Ghana, were examined. The attempt of identical examinations in the IPTi trial from Lambaréné, Gabon , was abandoned because of scarce availability of positive samples precluding a meaningful analysis. The study protocols were reviewed and approved by the Ethics Committees, University for Development Studies, Tamale, Kwame Nkrumah University of Science and Technology, Kumasi, and International Foundation of the Albert Schweitzer Hospital, Lambaréné, and informed consent was obtained from the children's parents. Details of the site characteristics and the analysed two trials have been presented elsewhere. The trial in Tamale was performed from March 2003 through July 2005, and the one in Afigya Sekyere from January 2003 through September 2005 [21, 22]. Briefly, while Tamale is a city of some 350,000 inhabitants but of rather rural character, the study in Afigya Sekyere comprised nine villages of 1,500 to 13,000 inhabitants. Malaria transmission is perennial with seasonal variation at both sites; entomologic data are not available for Tamale but in Afigya Sekyere, approximately 400 infective bites per person-year were observed in 2003-2005 (unpublished data). Coverage with insecticide-treated bed nets was low at ≤3% in both sites in 2003-2005 [21, 22]. In the first year of life, the incidence of malaria has been estimated as 0.95 and 1.27 episodes per person-years at risk in Tamale and Afigya Sekyere, respectively . In the two IPTi trials, 2270 children were recruited at three months of age and followed up actively in three-monthly (Tamale) or monthly (Afigya Sekyere) intervals, and passively, until 24 months of age. IPTi-SP or placebo was administered at ages 3, 9, and 15 months. Active follow-up included clinical examination, regular blood sampling, and parasitological and haematological examinations. During passive follow-up visits, i.e. when patients presented independently from regular visits, identical procedures were done as detailed elsewhere [21, 22]. Episodes of malaria were treated with artesunate (4 mg/kg, double dose on first day, for five days, Tamale) or amodiaquine-artesunate (10/4 mg/kg for three days, Afigya Sekyere). A sample size of ≥1,070 infants in each trial was estimated to provide 80% power for detecting a 20% reduction of hazards of developing malaria in the SP group, compared with the placebo group [21, 22].
On the day of IPTi dose-3, i.e., six months following IPTi dose-2 and at approximately 15 months of age, venous blood samples were collected. These were stabilized by adding an equal volume of DNA-stabilizing buffer (AS1, Qiagen, Germany), stored at +4°C, and DNA was extracted by commercial kits (QIAmp, Qiagen, Germany). All samples with microscopically confirmed P. falciparum parasitaemia were subjected to assessing MOI. For that, sequences corresponding to the allelic families of P. falciparum msp1 block 2 (K1, Mad20, Ro33) and of msp2 block 3 (FC27, IC) were amplified in five separate nested PCR assays . These alleles are characterized by conserved regions flanked by repeat sequences of variable length. Therefore, size variation within the alleles can be used to discriminate different parasite clones by PCR fragment length polymorphism. Fragments were separated on 3% GTG®-agarose gels (Biozym, Germany) and analysed using GeneSnap software (SynGene, UK). In case of negative or inconclusive PCR results, assays were repeated maximally twice. MOI was calculated as the highest number of fragments for either msp1 or msp2. Of a total of 348 microscopically positive samples collected at 15 months of age, MOI could be assessed in 277 (79.6%). For the Tamale trial, all samples collected at 15 months of age were examined by Plasmodium-specific nested PCR assays  and msp1/2 genotyping was extended to additional 142 samples identified to have submicroscopic parasitaemia of which 112 yielded a result for MOI (typing efficiency, 78.9%). Also, for the Tamale cohort, a random sample of microscopically positive samples (if available, each 25 for placebo and IPTi-SP groups) collected at 3, 6, 9, and 12 months of age was genotyped to illustrate the development of MOI with age (typing efficiency, 72.2%). For all samples, processing and DNA extraction were performed in a standardized manner, and all samples were typed at the Institute of Tropical Medicine in Berlin, Germany.
Malaria was defined by the presence of microscopically visible parasitaemia plus fever (axillary temperature, ≥37.5°C, or, Afigya Sekyere, ≥38.0, <12 months, rectally; ≥12 months, tympanically) or fever during the preceding 48 hours reported by mothers without being asked. For incidence estimates, children were regarded as not being at risk for 21 days following an episode of malaria and person-time-at-risk was reduced accordingly. Polyclonal infections had a MOI of >1. Rainy season (> 10 days or >90 mm of precipitation/month) was defined as before [5, 22].
Continuous parameters were compared between groups by the non-parametric Mann-Whitney U-test or Kruskal-Wallis test. These tests were also applied to compare MOI between groups but results are primarily presented as arithmetic means since this measure of average more selectively displays (the direction of) differences than medians, which are also shown, nevertheless. Proportions of polyclonal infections were analysed by χ2 tests. In multivariate analysis, known or potential confounders of the proportion of polyclonal infections and MOI were adjusted for by including these in logistic regression models and by applying non-parametric multiple ordinal regression, respectively.