Fig. 1

Forest plot of multivariable logistic regression results. Presence of parasitaemia was applied as the dichotomous response variable (more than or equal to 20 parasites per microlitre of whole blood were considered parasite positive). Included possible predictors: continuous parameters (age, hepcidin, cytokines, haemoglobin) and categorical variables (sex, haemoglobin genotype groups, malnutrition sub-categories). Colors group the covariates by subcategory (black: demographics, orange: continuous biomarkers, blue: haemoglobin genotypes, green: malnutrition). Significance codes: ***0.001; **0.01; *0.05; -a/aa with AA = single deletion of 3.7 kb or 4.2 kb, aa/aa with AC = haemoglobin C (heterozygous), -a/aa with AC = single deletion of 3.7 kb with haemoglobin C (heterozygous), aa/aa with AS = haemoglobin S (heterozygous), -a/aa with AS = single deletion of 3.7 kb with haemoglobin S (heterozygous); rarer traits = other genotype (details in Additional file 1: Tables S2, S3)