Elsevier

Acta Tropica

Volume 162, October 2016, Pages 114-124
Acta Tropica

Geographic potential of disease caused by Ebola and Marburg viruses in Africa

https://doi.org/10.1016/j.actatropica.2016.06.012Get rights and content

Highlights

  • Developed detailed database of likely transfer events of filoviruses from reservoir to humans or other primates.

  • Transmission risk maps developed for Ebola and Marburg viruses.

  • Updates and corrects results of previous mapping efforts.

  • Presents maps of uncertainty for each risk map.

Abstract

Filoviruses represent a significant public health threat worldwide. West Africa recently experienced the largest-scale and most complex filovirus outbreak yet known, which underlines the need for a predictive understanding of the geographic distribution and potential for transmission to humans of these viruses. Here, we used ecological niche modeling techniques to understand the relationship between known filovirus occurrences and environmental characteristics. Our study derived a picture of the potential transmission geography of Ebola virus species and Marburg, paired with views of the spatial uncertainty associated with model-to-model variation in our predictions. We found that filovirus species have diverged ecologically, but only three species are sufficiently well known that models could be developed with significant predictive power. We quantified uncertainty in predictions, assessed potential for outbreaks outside of known transmission areas, and highlighted the Ethiopian Highlands and scattered areas across East Africa as additional potentially unrecognized transmission areas.

Introduction

Since 1976, scattered human cases and outbreaks have been documented of hemorrhagic fever caused by viruses of the family Filoviridae, caused by the five known Ebola species and the related Marburg viruses (Groseth et al., 2007), which are non-segmented, negative-stranded RNA viruses. These viruses are believed to be hosted in the long term by fruit bats (family Pteropodidae), although full clarity in this issue is largely lacking: that is, solid causal evidence is accumulating regarding the bat Rousettus aegyptiacus as a reservoir for Marburg virus (Amman et al., 2012, Amman et al., 2015, Towner et al., 2009), and a tie of infections in humans to exposure to mines and caves is clear (Peterson et al., 2006). However, evidence regarding the reservoir of Ebola virus is less clear (Pigott et al., 2015). Although compelling temporal and anecdotal links have been pointed out (Leroy et al., 2009), serological evidence has painted a more complex picture, with detections of viruses in multiple bat species, and in regions and under ecological conditions where particular viruses have never been documented (Hayman et al., 2012, Pourrut et al., 2009). Most recently, Ebola surprised the world community with an emergence in Guinea—quite apart from the magnitude of the outbreak, Zaire ebolavirus was unknown in West Africa, as the virus that would have been expected in Guinea was Taï Forest ebolavirus (Bausch and Schwarz, 2014).

The West African Zaire ebolavirus outbreak underlines the need for a predictive understanding of the geographic distribution of these viruses and their potential for transmission to humans across Africa. Although studies on filoviruses almost invariably include a table (e.g., Chippaux, 2014) and/or map (e.g., Polonsky et al., 2014) of known outbreaks, only four studies have gone beyond the occurrences to interpolate or estimate a full potential distribution: early analyses that explored the basic idea (Peterson et al., 2004, Peterson et al., 2006) and a recent pair of assessments that took advantage of an additional decade of accumulation of occurrence data (Pigott et al., 2014, Pigott et al., 2015). The older analyses are rather dated, in terms of both the occurrence information and the quality of the environmental data and tools for analysis; the new studies, on the other hand, have a number of shortcomings in methodology, which are treated in detail in the Discussion.

This contribution aims to present a more comprehensive view of the geographic potential of Ebola and Marburg viruses known to infect humans, to offer an up-to-date and rigorous view of where these viruses may be found. Specifically, we (1) tested for niche divergence between Ebola species, and consequently treated Ebola species separately in modeling efforts; (2) we developed analyses for Marburg virus as well as Ebola; (3) we assessed and addressed pseudoreplication and its effects on model predictions; and (4) we addressed uncertainty in our model outputs. In our model calibration efforts, we took care to control for accessibility of areas, but we also assessed the possibility of long-distance jumps—effectively ‘surprises’ akin to Zaire ebolavirus appearing in West Africa in 2014 (Bausch and Schwarz, 2014).

Section snippets

Materials and methods

This paper presents a variety of niche model-based analyses of filovirus (i.e., Ebola and Marburg) potential distributions across Africa. We excluded two filovirus taxa from consideration for lack of sufficient occurrence information (Cuevavirus is known from one site in Spain only) or any occurrence information whatsoever (transmission of Reston ebolavirus from its natural reservoir is not known from any definable location, such that no geographic information is available to us). We are fully

Results

We first assessed ecological niche similarity between Zaire ebolavirus and Sudan ebolavirus. The two null (background similarity) distributions (i.e., Zaire ebolavirus compared to background models for Sudan ebolavirus, and vice versa) had ranges of similarity values of 0.76–0.92 and 0.84–0.98, whereas the observed value was 0.48. As a consequence, we rejected the null hypothesis of niche similarity, in favor of an alternative hypothesis of niche difference. An important implication of this

Discussion

This study updates the current understanding of filovirus geography in Africa. In the 10 years since our original analysis, the number of known outbreaks has approximately doubled, but most have occurred under quite predictable circumstances; however, with augmented input data, as well as improved analytical frameworks, new maps have been developed that should both anticipate future outbreak events and highlight areas of uncertain predictions. Our careful consideration of uncertainty and how

Acknowledgements

We thank Lindsay Campbell for help with analyses and comments on an early draft of the manuscript, and the Egyptian Fulbright Mission Program (EFMP) for support of AMS during development of this contribution.

References (50)

  • K. Changula et al.

    Ebola and Marburg virus diseases in Africa: increased risk of outbreaks in previously unaffected areas?

    Microbiol. Immunol.

    (2014)
  • J.-P. Chippaux

    Outbreaks of Ebola virus disease in Africa: the beginnings of a tragic saga

    J. Venom. Anim. Toxins Incl. Trop. Dis.

    (2014)
  • J.L. Conrad et al.

    Epidemiologic investigation of Marburg virus disease, southern Africa, 1975

    Am. J. Trop. Med. Hyg.

    (1978)
  • R Development Core Team, 2013. R: A Language and Environment for Statistical Computing http://www.R-project.org...
  • E. Fichet-Calvet et al.

    Risk maps of Lassa fever in west Africa

    PLoS Negl. Trop. Dis.

    (2009)
  • P. Formenty et al.

    Ebola virus outbreak among wild chimpanzees living in a rain forest of Cote d'Ivoire

    J. Infect. Dis.

    (1999)
  • D.T. Hayman et al.

    Ebola virus antibodies in fruit bats, Ghana, West Africa

    Emerg. Infect. Dis.

    (2012)
  • M. James et al.

    The Pathfinder AVHRR land data set: an improved coarse resolution data set for terrestrial monitoring

    Int. J. Remote Sens.

    (1994)
  • E.D. Johnson et al.

    Characterization of a new Marburg virus isolated from a 1987 fatal case in Kenya

    Arch. Virol.

    (1996)
  • E.M. Leroy et al.

    Human Ebola outbreak resulting from direct exposure to fruit bats in Luebo, Democratic Republic of Congo, 2007

    Vector-borne Zoonotic Dis.

    (2009)
  • A. Mylne et al.

    A comprehensive database of the geographic spread of past human Ebola outbreaks

    Sci. Data

    (2014)
  • Y. Nakazawa et al.

    Ecological niche modeling of Francisella tularensis subspecies and clades in the United States

    Am. J. Trop. Med. Hyg.

    (2010)
  • A. Negredo et al.

    Discovery of an Ebolavirus-like filovirus in Europe

    PLoS Pathog.

    (2011)
  • A.T. Peterson et al.

    Phylogenetic assessment of filoviruses: how many lineages of Marburgvirus?

    Ecol. Evol.

    (2012)
  • A.T. Peterson et al.

    Geographic evaluation of conservation status of African forest squirrels (Sciuridae) considering land use change and climate change: the importance of point data

    Biodivers. Conserv.

    (2007)
  • Cited by (0)

    View full text