Evaluating dengue transmission dynamics in an endemic urban setting in Viet Nam: an index-household contact cohort study.

Abstract

Current methods for controlling the spread of dengue in endemic settings typically rely on vector control in and around the homes of reporting cases, with varying levels of success. We aimed to evaluate the role of the household as a source of infection risk and assess the spatial scale of dengue virus transmission to understand why targeted vector control might not work effectively. In this index-household contact cohort study, dengue virus NS1-positive index cases reporting to a polyclinic or Tropical Medicine Hospital, together with their household contacts (aged >18 years), were enrolled, and followed up, in Nha Trang, Viet Nam between Oct 1, 2016, and May 15, 2019. Epidemiological (case reporting dates, questionnaires, and geographical locations), serological (ELISA and plaque reduction neutralisation tests), and molecular (serotyping PCR, quantitative PCR, and whole viral genome sequencing) data were collected, where possible, from all enrolled individuals. We then implemented an analysis framework, using viral pairwise genetic distance and modelling of the serial interval (duration between symptom onsets) between infected individuals, to identify individuals in chains of transmissions within the cohort, and define the transmission routes between humans via Aedes spp mosquitoes. During the study period, 838 index cases and 1241 household contacts were enrolled. Within the same household, we found evidence that only nine (29·0%) of the 31 pairs of concurrently infected individuals could be identified as plausible transmissions via the same infected mosquito. Using our genomic and time-modelling framework, we showed limited clustering of transmission chains at the mosquito flight range (150 m), using the home address location of infected individuals. Infections with identical, or near-identical genomes, consistent with our definition of plausible transmission, are identified in this dataset with a median home address distance of 1274 m (IQR 543-2016). Our results emphasise human movement as a major driver of dengue transmission in an urbanised setting, underscoring the importance of understanding mobility patterns to locate the sites of infected mosquitoes. Human daytime activity locations and social contacts should be examined more closely to better understand transmission networks, rather than relying solely on reactive, reporting index household-directed vector control. Johnson & Johnson, London School of Hygiene and Tropical Medicine, Japan Program for Infectious Diseases Research and Infrastructure, and Japan Agency for Medical Research and Development.