Study design and setting
This study is an extension of a longitudinal study that took place in the malaria endemic district of Iganga- Mayuge in eastern Uganda . From September 2008 to October 2008, a team of well-trained home visitors approached households to systematically recruit children into the baseline study. Eligible children were enrolled into the baseline study and followed up for a period of one year from November 2008 to November 2009. The study cohort was recruited from a community living within six villages of Iganga district that are in close proximity to the malaria study clinic located at Makerere University Iganga/Mayuge Demographic Surveillance Site (MaK-DSS). No interventional studies were undertaken in this study area at the time this study cohort was assembled. Inclusion criteria of the cohort study was as described in an earlier study  and followed; 1) age six months to nine years; 2) agreement to come to study clinic for any febrile episode or illness; 3) agreement to avoid medications administered outside the study; 4) agreement to remain in study area during the twelve months follow up; 5) absence of known chronic disease and 6) written informed consent provided by parent or guardian. Severely malnourished children (below -3z scores of the median World Health Organization (WHO) growth standards)  were excluded. Follow-up started when children fulfilled all of the selection criteria and were free of symptomatic malaria.
Active case detection and estimation of malaria incidence
Study villages were divided by convenience into active (nearby) villages and passive (more remote) villages. Study personnel sought for verbal consent from parents/guardians of the children to participate in a brief demographic survey, and written informed consent was obtained before enrolment into the study. Using a standardized questionnaire, demographics and malaria indicator information was collected. After the baseline survey, parents or guardians were instructed to bring their children to the malaria clinic based at Iganga Hospital whenever they felt unwell. Follow-up started when children fulfilled all of the selection criteria and were free of symptomatic malaria.
Children were visited twice a week by the study field workers at convenient times of day. A standardized questionnaire was administered for collecting information regarding illnesses that had occurred since the last visit, use of health care facilities and medications used. At each visit, the tympanic temperature was recorded using a digital thermometer. When fever (tympanic temperature of ≥37.5°C) or history of recent fever (within the last 24 hours) was observed or reported for any study child, a rapid diagnostic test (RDT, OPTIMAL®) and microscopy of a stained blood smear were performed to screen for malaria and confirm the presence of malaria parasites, respectively. Uncomplicated malaria was confirmed using the WHO criteria that includes having any P. falciparum parasitaemia plus fever or a history of fever (within the past 24 hours) . Children found with asexual malaria parasitaemia were administered artemisinin combination therapy (ACT) at the study clinic following Uganda national treatment guidelines  or were referred to Iganga Hospital acute care unit in case of severe illness.
Follow-up started for children who were free of symptomatic malaria after treatment with artemisinin combination therapy. The time at risk for new infection was defined as the duration of study participation excluding 14 days after each ACT treated episode of malaria. The incidence of malaria was determined by calculating the number of malaria episodes/child over the one year of active follow up.
Sample preparation and DNA extraction
Whole blood samples were obtained from all study children for subsequent DNA analyses. Blood samples were drawn into Ethylene Diamine Tetra Acetic acid (EDTA) anticoagulant tubes. Buffy coats were prepared from 1 to 2 ml of whole blood by differential sedimentation using phosphate buffer saline (PBS) and 2% fetal bovine serum (FBS). Genomic DNA was extracted from blood leukocytes using E.Z.N.A Blood DNA kit following the manufacturer’s protocol (Omega Bio-tek, USA). DNA samples were stored at -20oC for subsequent genomic analysis.
Haptoglobin genotypes were determined by allele-specific polymerase chain reaction amplification as described before , using primers sets listed in Table 1. This was based on determination of the polymorphic alpha (α) chain alleles, Hp1 and Hp2. Hp1 encodes for the α1-1S (slow form) and 1F (fast form) chains, while Hp2 encodes for the Hp α2 polypeptide chain. Primers F3 and C42 were used to amplify a 935bp fragment of the Hp2 allele. To amplify the 1.2kb DNA fragments in the Hp1S allele, primers C51 and S2 were used (reaction S). While for amplification of the Hp1F allele, primers F3 and C72 were used in reaction F and this combination amplified 1.4kb DNA fragments.
Polymerase chain reaction (PCR)
PCR amplification was carried out in a 25µL reaction containing 10mM Tris-HCl (pH 9.0), 50mM KCl, and 0.1% Triton X-100, 2.5mM MgCl2, 200µM of each dNTP, 0.2 µM of each primer and 1.5 units of Taq polymerase supplied with buffer (Thermo Scientific, Inc). For each reaction, approximately 1µL of the DNA sample was used. After preheating at 95o C for 3 min, PCR was performed with 35 cycles of heating at 94 o C for 40 sec, at 58 o C for 1 min and annealing at 72 o C for 2 min. PCR products were separated by electrophoresis on 1.2% agarose gels stained with ethidium bromide. Hp genotypes were determined by observing the amplified DNA fragments under ultra violet light.
Data management and analysis
Data were cleaned, coded and entered into Microsoft Office Access TM 2007. Descriptive statistics, Chi-square tests and multivariate analysis were carried out using Stata 12.0 (Stata Corp, College Station, Texas, USA). Allele and genotype frequencies were calculated. The association between haptoglobin genotype and incidence of uncomplicated malaria was estimated using a multivariate negative binomial regression model. From a previous analysis in the same children’s cohort, age, malaria history and insecticide bed net (ITN) were identified as independent predictors of malaria incidence . These factors were treated in the final multivariate analysis to determine the extent to which haptoglobin genotypes affected the incidence of uncomplicated malaria in the children’s cohort. Adjusted incidence rate ratios (aIRRs), P values and 95% confidence intervals were calculated. All statistical tests were two-tailed and P-values less than 0.05 were considered significant.