Assessing the Risk of Occurrence of Bluetongue in Senegal

Abstract

Bluetongue is a non-contagious viral disease affecting small ruminants and cattle that can cause severe economic losses in the livestock sector. The virus is transmitted by certain species of the genus Culicoides and consequently, understanding their distribution is essential to enable the identification of high-risk transmission areas. In this work we use bioclimatic and environmental variables to predict vector abundance, and estimate spatial variations in the basic reproductive ratio &nbsp;R0. The resulting estimates were combined with livestock mobility and serological data to assess the risk of Bluetongue outbreaks in Senegal. The results show an increasing abundance of C. imicola, C. oxystoma, C. enderleini, and C. miombo from north to south. R0 < 1 for most areas of Senegal, whilst southern (Casamance) and southeastern (Kedougou and part of Tambacounda) agro-pastoral areas have the highest risk of outbreak (R0 = 2.7 and 2.9, respectively). The next higher risk areas are in the Senegal River Valley (R0 = 1.07), and the Atlantic coast zones. Seroprevalence rates, shown by cELISA, weren't positively correlated with outbreak probability. Future works should include follow-up studies of competent vector abundancies and serological surveys based on the results of the risk analysis conducted here to optimize the national epidemiological surveillance system.

Keywords: basic reproduction number; compartmental model; risk map; vector abundance model.

Figure 1

Seroprevalence, entomological, and demographic data. The circles show regional BT seroprevalence based on data collected during the national sero-survey in 2018. The squares show the distribution of the four Culicoides species of interest; the square’s size is proportional to the logarithm of the number of Culicoides collected during the national entomological survey in 2012. The background colors (red and blue) delimit the two pastoral zones with different birth and mortality rates. Dashed lines delimit the six agro-ecological zones in Senegal (Ferlo, Senegal River Valley, Eastern Senegal, Casamance, Groundnut Basin, Niayes), while the thin solids lines identify the administrative regions.

Figure 2

Mobility data. On top (a) representation of livestock mobility network, each line corresponds to movements existing between the department, and the color is relative to the number of animals exchanged. The bottom part, departments are colored based on: (b) The incoming volume (number of animals arriving in the department); (c) outgoing volume (number of animals leaving the department), na: data not available.

Figure 3

Pictorial representation of the mathematical model. Host dynamics are at the top of the figure, vector dynamics at the bottom.

Figure 4

Maps of abundance of the four main Bluetongue virus (BTV) vector species in Senegal: (a) C. imicola; (b) C. oxystoma; (c) C. enderleini, and (d) C. miombo.

Figure 5

Transmission maps ($R_0$) (a) realized using temperature-independent parameters; (b) using temperature-dependent parameters.

Figure 6

Tornado plot of sensitivity analysis in the 2 cases when parameters used in the model are temperature-independent (a) or temperature-dependent (b). The name of the epidemiological parameters that are varied on shown on the y-axis. The variation in the number of pixels at risk is shown on the x-axis. For each parameter, the bars represent the variation in the number of at risk pixels following a decrease or increase in a parameter value. The color of the bar indicates if the variation in the number of pixels at risk is due to an increase (red) or a decrease (blue) in the corresponding parameter value. The length of the bar is proportional to the variation in the number of pixels at risk (expressed as a percentage).

Figure 7

Probability of outbreaks due to the introduction of infected animals. (a) using R₀ temperature-independent values, (b) using R₀ temperature-dependent values. Livestock movements from infected areas could explain why areas with a low value of R₀ could have a high risk of outbreak occurrence. White areas correspond to areas where the outbreak probability has not been estimated because of missing information.