Sources of persistent malaria transmission in a setting with effective malaria control in eastern Uganda: a longitudinal, observational cohort study

BACKGROUND: Symptomatic malaria cases reflect only a small proportion of all Plasmodium spp infections. Many infected individuals are asymptomatic, and persistent asymptomatic Plasmodium falciparum infections are common in endemic settings. We aimed to quantify the contribution of symptomatic and asymptomatic infections to P falciparum transmission in Tororo, Uganda. METHODS: We did a longitudinal, observational cohort study in Tororo district, Uganda. We recruited participants of all ages from randomly selected households within this district. Participants were eligible if the selected household had no more than nine permanent residents and at least two members younger than 10 years, and the household was their primary residence, and they agreed to come to the study clinic for any fever episode and avoid antimalarial medications outside the study. Participants were followed-up by continuous passive surveillance for the incidence of symptomatic infections; routine assessments (ie, standardised clinical evaluation and blood samples) were done at baseline and at routine visits every 4 weeks for 2 years. P falciparum parasite density, gametocyte density, and genetic composition were determined molecularly using quantitative PCR (qPCR), quantitative reverse transcriptase PCR (qRT-PCR), and amplicon deep sequencing, respectively. Membrane feeding assays were also done to assess infectivity to mosquitoes. The contribution of different populations to the infectious reservoir was estimated for symptomatic infections, asymptomatic but microscopically detected infections, and asymptomatic but qPCR-detected infections; and for age groups younger than 5 years, 5–15 years, and 16 years or older. FINDINGS: Between Oct 4, 2017, and Oct 31, 2019, 531 individuals were enrolled from 80 randomly selected households and were followed-up for 2 years. At baseline, P falciparum was detected in 28 (5·3%) of 531 participants by microscopy and an additional 64 (12·1%) by qPCR and declined thereafter. In 538 mosquito feeding experiments on 107 individuals, 446 (1·2%) of 37 404 mosquitoes became infected, with mosquito infection rates being strongly associated with gametocyte densities (β=2·11, 95% CI 1·62–2·67; p<0·0001). Considering both transmissibility of infections and their relative frequency, the estimated human infectious reservoir consisted primarily of asymptomatic microscopy-detected infections (83·8%), followed by asymptomatic submicroscopic infections (15·6%), and symptomatic infections (0·6%). Children aged 5–15 years accounted for more than half of the infectious reservoir (58·7%); individuals younger than 5 years (25·8%) and those 16 years or older (15·6%) contributed less. Samples from four children contribued to 279 (62·6%) of 446 infected mosquitoes after multiple mosquito-feeding assays. INTERPRETATION: Individuals with asymptomatic infections were important drivers of malaria transmission. School-aged children contributed to more than half of all mosquito infections, with a small minority of asymptomatic children being highly infectious. Demographically targeted interventions, aimed at school-aged children, could further reduce transmission in areas under effective vector control. FUNDING: US National Institutes of Health, Bill & Melinda Gates Foundation, and the European Research Council.