Malaria transmission, infection, and disease at three sites with varied transmission intensity in Uganda: implications for malaria control

Abstract

The intensification of control interventions has led to marked reductions in malaria burden in some settings, but not others. To provide a comprehensive description of malaria epidemiology in Uganda, we conducted surveillance studies over 24 months in 100 houses randomly selected from each of three subcounties: Walukuba (peri-urban), Kihihi (rural), and Nagongera (rural). Annual entomological inoculation rate (aEIR) was estimated from monthly Centers for Disease Control and Prevention (CDC) light trap mosquito collections. Children aged 0.5-10 years were provided long-lasting insecticidal nets (LLINs) and followed for measures of parasite prevalence, anemia and malaria incidence. Estimates of aEIR were 2.8, 32.0, and 310 infectious bites per year, and estimates of parasite prevalence 7.4%, 9.3%, and 28.7% for Walukuba, Kihihi, and Nagongera, respectively. Over the 2-year study, malaria incidence per person-years decreased in Walukuba (0.51 versus 0.31, P = 0.001) and increased in Kihihi (0.97 versus 1.93, P < 0.001) and Nagongera (2.33 versus 3.30, P < 0.001). Of 2,582 episodes of malaria, only 8 (0.3%) met criteria for severe disease. The prevalence of anemia was low and not associated with transmission intensity. In our cohorts, where LLINs and prompt effective treatment were provided, the risk of complicated malaria and anemia was extremely low. However, malaria incidence was high and increased over time at the two rural sites, suggesting improved community-wide coverage of LLIN and additional malaria control interventions are needed in Uganda.

Figure 1.

Map of Uganda showing study sites.

Figure 2.

Enrolment and follow-up of study participant at the three study sites.

Figure 3.

Parasite prevalence by age at the three study sites with varying malaria transmission intensity. Curves for each site were fitted using locally weighted least squares (lowess) regression, with standard errors (shaded region) estimated with 1,000 bootstrapping replicates, with resampling at the level of the subject.

Figure 4.

Malaria incidence by age at the three study sites with varying malaria transmission intensity. Curves for each site were fitted using locally weighted least squares (lowess) regression, with standard errors (shaded region) estimated with 1,000 bootstrapping replicates, with resampling at the level of the subject.

Figure 5.

Temporal trends in measures of malaria infection, transmission, and disease at the three study sites. Entomological inoculation rates (dashed lines) for each site were estimated using the product of human biting rates (averaged monthly) and sporozoite rates (averaged yearly). Malaria incidence (solid lines) for each site was estimated using mean monthly incidence; prevalence (dots) for each site was estimated from active surveillance occurring every 3 months. Estimates of average monthly rainfall were obtained from the NASA Tropical Rainfall Measuring Mission Project.