Modelling the persistence and control of Rift Valley fever virus in a spatially heterogeneous landscape

Warren S D Tennant,Catherine Cêtre-Sossah,Davide Colombi,Gilles Le Godais,Simon E F Spencer,Raphaëlle Métras,Onzade Charafouddine,Matt J Keeling,Eric Cardinale,Mike J Tildesley,Youssouf Moutroifi,W John Edmunds,Vittoria Colizza

doi:10.1038/s41467-021-25833-8

Abstract

The persistence mechanisms of Rift Valley fever (RVF), a zoonotic arboviral haemorrhagic fever, at both local and broader geographical scales have yet to be fully understood and rigorously quantified. We developed a mathematical metapopulation model describing RVF virus transmission in livestock across the four islands of the Comoros archipelago, accounting for island-specific environments and inter-island animal movements. By fitting our model in a Bayesian framework to 2004–2015 surveillance data, we estimated the importance of environmental drivers and animal movements on disease persistence, and tested the impact of different control scenarios on reducing disease burden throughout the archipelago. Here we report that (i) the archipelago network was able to sustain viral transmission in the absence of explicit disease introduction events after early 2007, (ii) repeated outbreaks during 2004–2020 may have gone under-detected by local surveillance, and (iii) co-ordinated within-island control measures are more effective than between-island animal movement restrictions.

Highlights

The persistence mechanisms of Rift Valley fever (RVF), a zoonotic arboviral haemorrhagic fever, at both local and broader geographical scales have yet to be fully understood and rigorously quantified
We developed a mathematical metapopulation model to describe the spread of RVF virus within and between the four islands of the Comoros archipelago (Fig. 1), and fitted this model in a Bayesian framework to livestock seroprevalence data collected from 2004 until 2015
We modelled RVF viral transmission within agestructured livestock populations within each island as a function of the Normalized Difference Vegetation Index (NDVI), and between islands through the movement of livestock

Summary

Introduction

The persistence mechanisms of Rift Valley fever (RVF), a zoonotic arboviral haemorrhagic fever, at both local and broader geographical scales have yet to be fully understood and rigorously quantified. Beyond its potential for spread to further geographical areas, a major concern is the likelihood of persistence in previously disease free regions[7,8,9,10,11] These persistence mechanisms vary between ecosystems depending on local host communities and meteorological factors allowing favourable conditions for mosquito vectors to complete their life cycle and to be capable of virus transmission[12]. We sought to (i) estimate the importance of island-specific variables and animal movements across the islands on RVF spread, (ii) assess the likelihood of RVF persistence in the system without re-introduction from mainland Africa or Madagascar, and (iii) assess the impact of livestock movement control measures on disease incidence in the Comoros archipelago. We assessed the impacts of movement restrictions and reducing within-island transmission on each island upon the total number of new infections in livestock from 2004 to 2015

Methods

Results

Conclusion