Plasmodium falciparum genetic polymorphisms that mediate altered drug sensitivity may impact upon virulence. In a cross-sectional study, Ugandan children with infections mutant at pfcrt K76T, pfmdr1 N86Y, or pfmdr1 D1246Y had about one-fourth the odds of symptomatic malaria compared to those with infections with wild type (WT) sequences. However, results may have been confounded by greater likelihood in those with symptomatic disease of higher density mixed infections and/or recent prior treatment that selected for WT alleles. Polymorphisms in samples from paired episodes of asymptomatic and symptomatic parasitaemia in 114 subjects aged 4-11 years were followed longitudinally in Tororo District, Uganda. Paired episodes occurred within 3-12 months of each other and had no treatment for malaria in the prior 60 days. The prevalence of WT, mixed, and mutant alleles was determined using multiplex ligase detection reaction-fluorescent microsphere assays. Considering paired episodes in the same subject, the odds of symptomatic malaria were lower for infections with mutant compared to WT or mixed sequence at N86Y (OR 0.26, 95% CI 0.09-0.79, p = 0.018), but not K76T or D1246Y. However, symptomatic episodes (which had higher densities) were more likely than asymptomatic to be mixed (for N86Y OR 2.0, 95% CI 1.04-4.0, p = 0.036). Excluding mixed infections, the odds of symptomatic malaria were lower for infections with mutant compared to WT sequence at N86Y (OR 0.33, 95% CI 0.11-0.98, p = 0.046), but not the other alleles. However, if mixed genotypes were grouped with mutants in this analysis or assuming that mixed infections consisted of 50% WT and 50% mutant genotypes, the odds of symptomatic infection did not differ between infections that were mutant or WT at the studied alleles. Although infections with only the mutant pfmdr1 86Y genotype were associated with symptomatic infection, this association could primarily be explained by greater parasite densities and therefore greater prevalence of mixed infections in symptomatic children. These results indicate limited association between the tested polymorphisms and risk of symptomatic disease and highlight the value of longitudinal studies for assessing associations between parasite factors and clinical outcomes.