Projecting the end of the Zika epidemic in Latin America: A modelling analysis

Abstract

Background: Zika virus disease emerged in Latin America in early 2015, which had serious implications for population health. In 2016, the World Health Organization declared a cluster of neurological disorders and neonatal malformations associated with Zika cases a Public Health Emergency of International Concern. 2017 incidence has declined, and future disease incidence in Latin America remains uncertain due to gaps in our understanding of the natural history of infection, considerable variation in surveillance and a lack of a comprehensive collation of available data from affected countries. Methods & Materials: This analysis combines publically available data on Zika virus incidence across most Latin American countries and a spatio-temporal dynamic transmission model for Zika virus infection to determine key transmission parameters and likely future incidence in 87 cities. Seasonality was determined by spatio-temporal estimates of Aedes aegypti vector capacity. Country and state-level data are used to infer key model parameters using Monte-Carlo methods, different movement models were tested against the data and the best-fitting parameter combinations were used to estimate incidence within each city. Results: We predict that the highest incidence in 2018 will be observed in Colombia and some Brazilian States (Parana, Sao Paulo, Rio de Janiero and Minas Gerias), but the estimated number of infections will be no more than a few hundred and the incidence risk ratio will be below 0.05. There was limited transmission in 2015, but for most cities 2016 and 2017 provided a sufficient opportunity for ZIKV transmission and populations have been depleted of susceptible individuals. Conclusion: The findings suggest that much of the ZIKV-associated morbidity will be in neonates and children that are born up to the end of 2017. The findings present a challenge to planned ZIKV-specific interventions such as phase III vaccine trials, as we predict substantially lower numbers of infections in 2018.