Preventive Veterinary Medicine 225 (2024) 106146

Contents lists available at ScienceDirect

Preventive Veterinary Medicine

journal homepage: www.elsevier.com/locate/prevetmed

Intra- and interspecies infectious neighbourhoods as determinant

parameters for Mycobacterium bovis infection among badgers in
southwestern France
Malika Bouchez-Zacria a, b, *, Pierre Jabert c, Édouard Réveillaud d, Céline Richomme e,
Maud Marsot a
a
French Agency for Food, Environmental and Occupational Health & Safety (ANSES), Laboratory for Animal Health, Epidemiology Unit, University Paris-Est, Maisons-
Alfort, France
b
Independent researcher, Audincthun, France
c
French Ministry of Agriculture, Directorate General for Food, 75015, Paris
d
Nouvelle-Aquitaine Regional Directorate for Food, Limoges, France
e
Nancy Laboratory for Rabies and Wildlife, French Agency for Food, Environmental and Occupational Health & Safety (ANSES), Malzéville, France

A R T I C L E I N F O A B S T R A C T

Keywords: The epidemiological system for Mycobacterium bovis in France involves cattle and, in some areas, wildlife species
European badger (mainly badgers and wild boar). This multi-host aspect complicates the control and eradication prospects for
Mycobacterium bovis bovine tuberculosis in endemic areas, despite the surveillance and control measures implemented for decades in
Cattle
this officially tuberculosis-free European country. To improve control measures, and to manage spillback
Landscape epidemiology
Dirichlet tessellation
transmission from badgers to cattle, it is necessary to clarify the transmission mechanisms of M. bovis in these
epidemiological systems. We modelled a badger population from a southwestern endemic area by a Dirichlet
tessellation based on a sett census conducted by local hunters and trappers between 2013 and 2015. We then
used a logistic regression model to test the association between the infection status of setts and computed var­
iables depicting three types of transmission (intraspecific, interspecific and landscape-associated). The apparent
prevalence of infected setts was of 40.5%. Two variables were significantly associated with the probability for a
sett to be infected: the proportion of neighbouring setts that were infected (OR: 3.19 [2.04–5.17]95%) and the
presence of nearby pastures belonging to an infected farm (OR: 2.33 [1.13–4.89]95%]. While badger culling
measures have been implemented according to the national TB control plan in the study area since 2012 (in the
vicinity of infected farms and their pastures), our results clearly highlight the need to reinforce measures aimed
at reducing both intraspecific and interspecific infection pressure. For this purpose, the promising prospect of
badger vaccination could be considered, along with biosecurity measures.

1. Introduction houses and via the testing of live cattle), France obtained this status in
2001 (Michelet et al., 2020). However, some regions are still affected by
Bovine tuberculosis (TB) caused by Mycobacterium bovis is a zoonotic the disease. This is particularly the case of the Nouvelle-Aquitaine region
disease affecting cattle herds, and, in some epidemiological contexts, (south-west France) where infected cattle farms are still detected
wildlife species. In 1964, European Council Directive 64/432/EEC through the surveillance system and where total or partial culling is
(Anon., 1964) introduced the concept of the “officially tuberculosis free” necessary to contain the M. bovis pressure (Delavenne et al., 2020). The
status for European countries with a percentage of confirmed control and eradication of TB have been complicated by the discovery of
TB-infected herds not exceeding 0.1% per year for six consecutive years M. bovis among wildlife in some areas: roe and red deer (Lambert et al.,
(Reviriego Gordejo and Vermeersch, 2006). After implementing sur­ 2017; Zanella et al., 2008), wild boar (Canini et al., 2023; Réveillaud
veillance in line with this European directive (post mortem at slaughter et al., 2018; Richomme et al., 2013; Zanella et al., 2008), badger

* Corresponding author at: French Agency for Food, Environmental and Occupational Health & Safety (ANSES), Laboratory for Animal Health, Epidemiology Unit,
University Paris-Est, Maisons-Alfort, France.
E-mail address: [email protected] (M. Bouchez-Zacria).

https://doi.org/10.1016/j.prevetmed.2024.106146
Received 19 October 2023; Received in revised form 8 February 2024; Accepted 9 February 2024
Available online 15 February 2024
0167-5877/© 2024 Elsevier B.V. All rights reserved.
M. Bouchez-Zacria et al. Preventive Veterinary Medicine 225 (2024) 106146

(Bouchez-Zacria et al., 2017; Payne et al., 2013) and, more recently, red 2009; Bodin, 2005).
foxes (Michelet et al., 2018; Richomme et al., 2020). Since 2011, a In a previous study carried out in France, certain variables were
surveillance system for free-ranging wildlife populations has been found to be significantly associated with concomitant infection in bad­
implemented to assess TB infection in deer, wild boar and badgers gers and cattle in buffer areas of 500 m and 1000 m around badger setts:
(Réveillaud et al., 2018; Rivière et al., 2015). It remains challenging, for both distances, the microrelief (i.e. the topographic heterogeneity of
nevertheless, to fully understand the role of each of these wildlife spe­ the soil surface) and the mean percentage of sand in the topsoil; in 500-
cies. While deer and foxes appear to have a minor anecdotal role (Anses, m buffers, the presence of plots of cereals or rapeseed (expected to in­
2021; Lambert et al., 2017), wild boar are thought to play the role of crease badger movements); and in 1,000-m buffers, the number of
intermediary between cattle and badgers (Canini et al., 2023). Indeed, neighbouring badger groups (expected to decrease badger movements)
badgers are thought to play a definite role in some epidemiological (Bouchez-Zacria et al., 2017). Thus, environmental and population
systems (Payne et al., 2013; Réveillaud et al., 2018), without being a characteristics could influence the presence of M. bovis infection in the
maintenance host per se (Bouchez-Zacria et al., 2023). badger-cattle epidemiological system in this area. Other studies have
The social and territorial lifestyle of badgers, living in setts where focused on the cattle host of this epidemiological system, underlining
individuals of the same group are in close proximity (Roper, 2010), the role of trade (Palisson et al., 2016) or contact at pasture (Palisson
makes them an ideal host for the transmission of M. bovis. Indeed, this et al., 2017). However, the determinants of this infection among badger
bacillus is known to survive under conditions such as shade and hu­ populations in France remain rarely studied.
midity (Broughan et al., 2016; Humblet et al., 2009; Young et al., 2005) To identify variables determining M. bovis infection among badgers
and is transmitted via droplets and sputum excreted by infected in­ in one of the main TB areas in France - Pyrénées-Atlantiques – Landes -, we
dividuals through the respiratory tract, the main route of entry for first modelled the badger population using a Dirichlet tessellation. Then,
M. bovis in badgers (Allen et al., 2021; Corner et al., 2011). The trans­ using a logistic regression model, we tested variables targeting three
cutaneous route, following bites by an infectious congener, and the oral types of determinants related to badgers, cattle (without assuming the
route (with the ingestion of food soiled with aerosol particles or infected direction of transmission between hosts) and landscape characteristics.
faeces, urine, sputum or wound exudates) has also been described as
other possible transmission routes (Corner et al., 2011). Pseudo-vertical 2. Material and methods
transmission (from mothers to badgers) has also been suggested (Morris
et al., 1994; Wilesmith, 1991). In different areas of France, badgers have 2.1. Study area
been found to be infected by the same spoligotypes (a specific classifi­
cation for Mycobacterium complex members (Haddad et al., 2001)) as The 2,749-km2 study area in southwestern France straddles two
infected cattle on neighbouring farms (Bouchez-Zacria et al., 2018, neighbouring departments: Pyrénées-Altlantiques and Landes (Fig. 1). The
2017). This is the case for example for Pyrénées-Atlantiques and Landes study area is characterised by a mainly oceanic climate. It is a mixed
departments (translation of “département”, a French administrative zone used for livestock farming in addition to crop and wood produc­
unit). In a recent publication, a positive association was demonstrated tion. The relatively flat lowlands and plateaus are mainly devoted to
between a spatial index of indirect contact between cattle and badgers in crops, while the hillsides are more of a bocage type, with alternating
France, and the risk of TB outbreaks in cattle (Marsot et al., 2023), pasture and woods on the slopes, and pasture and crops in the valleys.
underlining the importance of spatial parameters in interspecies The mean altitude is of 176 m (min: 20 m; max: 400 m). Three large river
transmission. branches segment the area.
Mainly indirect contact between cattle and badgers (O’Mahony,
2014; Payne et al., 2017; Woodroffe et al., 2016) is thought to occur in 2.2. Badger data and population modelling
pastures, where a main component of the badger’s diet — earthworms
(Lumbricus terrestris) — abounds. Indeed, earthworms can also host The study area included all the communes (the smallest French
M. bovis (Barbier et al., 2016b). But several and non-exclusive hypoth­ administrative subdivision) where a census of badger setts was taken
eses have been put forward for this interspecies transmission. Indeed, between 2013 and 2015 (Fig. 1). In France, the regulatory guidelines for
infected cattle could contaminate pastures through their excreta, res­ the trapping of badgers are defined by the French Minister of Ecology
piratory droplets and nasal mucus. Slurry spreading has also been and Sustainable Development by the ministerial order of January 29th,
mentioned (the risk of contamination depending on the techniques 2007. Badgers are trapped by volunteer trappers, duly trained and
used). Infected badgers could, for their part, shed M. bovis through authorised (accreditation required), under the supervision of pest con­
respiratory exudates and mucus and their urine (on the specific paths the trol officers (“lieutenants de louveterie” historically called wolf hunters,
use) (Allen et al., 2021). Finally, the digging of faecal latrines and the being nowadays volunteer officers in charge of pest control and who
excretion of bacilli in faeces make these latrines possible harbouring supervised badger trapping). The trapping effort is not strictly stan­
sites for the bacteria and therefore for transmission (King et al., 2015). dardized, and, given the participatory nature of the Sylvatub pro­
While pastures have been easily mapped thanks to official data and farm gramme, local trappers retain the autonomy to decide on the number of
declarations for the European Common Agricultural Policy, the home traps, trapping nights, and their placement. However, trapping is carried
ranges and movements of badgers are more uncertain in France. To out in infected areas, where one of the control measures is to reduce the
define home ranges and territories, some studies have used marked baits badger population in these areas, with an encouragement to trap close to
(e.g. in England (Delahay et al., 2000)), and others camera traps (e.g. in infected farms (technical directive from the French Minister of Agri­
Italy (Balestrieri et al., 2016)). However, the large-scale deployment of culture DGAL/SDSPA/2018–708). Indeed, the Sylvatub wildlife sur­
these types of protocol remains complicated from a material, human and veillance system was set up in 2012 in France. This system was used to
financial point of view. The modelling of home ranges by polygons ob­ collect data both on badgers trapped and analysed according to the risk
tained with a Dirichlet tessellation method based on badger sett loca­ level estimated in each department (reassessed twice a year) and on
tions (Doncaster and Woodroffe, 1993) has also been proposed and road-killed badgers (Réveillaud et al., 2018). Two types of traps are
appears to yield realistic home ranges (Roper, 2010). While the terms authorized in France: stopped cable restraints and cage traps. Night
home range and territory are sometimes used interchangeably, the shooting may also be an option in some infected areas, but, with a few
former is preferable to describe the use of space when related to a group exceptions, French trappers predominantly use stopped cable restraints.
or an individual (i.e. the area used for normal activities (Anich et al., These restraints are fitted with a stop adapted to the circumference of
2009)), while territory is more appropriate when social aspects are adult badgers to reduce the likelihood of injury. We have no published
emphasised, and more specifically the defence of an area (Anich et al., data from the control programme in France but the effects of this type of

2
M. Bouchez-Zacria et al. Preventive Veterinary Medicine 225 (2024) 106146

Fig. 1. Study area (A: Location in metropolitan France; B: tessellations (delimited by grey lines) of setts censused in the study area (grey points: censused setts not
having been analysed; red points: setts where at least one badger was found to be infected between 2014 and 2020; green points: setts where at least two badgers
were analysed and found not to be infected, in addition to no badgers being found to be infected between 2014 and 2020); C: setts included in the model and their
corresponding tessellations (areas delimited by grey lines) (grey areas: border of edge tessellations and setts, excluded from the analysis).

trapping system were investigated by Byrne et al. (2015) and showed sensitivities of these tests have been estimated (through scenario trees
very low rates of moderate/severe injury (Byrne et al., 2015). Once representing the steps of Sylvatub programme) to be respectively 0.51
caught, the badgers are shot to death (shot to the head). The animals are [0.47–0.54]95% and 0.74 [0.70–0.78]95% (Rivière, 2016). Traps were set
double-wrapped in the freezer of the nearest collection centre, then close to setts, on paths that originate from sett entrances. We therefore
transported frozen to the laboratory, where the carcasses are kept frozen assumed that the association of a badger with a sett would be reliable in
until necropsy and post-mortem analysis. All the badgers were tested for terms of trapping. The surveillance results we used (trapping or collec­
M. bovis infection on pooled lymph nodes with or without visible lesions, tion date, attributed sett and infection status) were provided by Syl­
with culture until 2015 and by PCR detection thereafter. The vatub. The ANSES national reference laboratory (NRL) at Maisons-Alfort

3
M. Bouchez-Zacria et al. Preventive Veterinary Medicine 225 (2024) 106146

(France) provided the spoligotype (Hauer et al., 2019) results of M. bovis 2.3. Cattle data
strains. The study period relating to badgers was from January 2014 to
December 2020. The surveillance and control protocols implemented between 2010
To reflect a more realistic badger metapopulation (a set of pop­ and 2021 have been previously described and consisted in annual,
ulations of the same species, interconnected mainly through dispersion biannual or triannual herd testing (using single intradermal tuberculin
movements), we grouped together setts that were within 400 m of each tests [SITT] or single comparative cervical tuberculin tests [SICCT])
other into sett clusters. Indeed, two arguments led us to this choice: i) depending on which commune and department were involved (Bou­
individuals from setts less than 400 m apart have been proved to be chez-Zacria et al., 2018, 2017). M. bovis infection in a cattle herd was
genetically related (through spatial autocorrelation) and ii) badger ge­ confirmed by polymerase chain reaction (PCR) and/or bacterial culture
netic structure became non-significant between 400 m and 600 m either following a skin test or meat inspection. This confirmation was
(Jacquier et al., 2020). Setts within 400 m of other setts were linked followed by the slaughter of all the cattle on the farm between 2010 and
through a network using an iterative approach. If setts were isolated, 2016. Partial depopulation has been made possible from 2016 onwards,
they were considered separate clusters and we stopped the computation. with a test-and-cull protocol introduced in 40% of infected farms
If not, the largest component of the network (setts linked either directly (Bouchez-Zacria et al., 2023). The TB surveillance data (date and farm
or indirectly, i.e. one or more setts linking them) was considered a po­ identification number) were provided by the French Ministry of Agri­
tential cluster. Although the maximum distance between pairs of setts in culture (FMA) and spoligotyping data were provided by the NRL. The
this cluster was greater than 400 m, we looked for the most central sett spoligotype was determined for all the isolates from each infected farm.
in the cluster (i.e. the one with the most neighbours) and excluded setts
more than 400 m away from it. The remaining setts constituted a cluster
and they were excluded from the computation; the previous steps were 2.4. Landscape data
then repeated until there were no more setts to consider.
. For each cluster, we first defined the sett closest to the centroid. Cattle-grazed plots (pastures owned by farmers and used for cattle
Then we computed a Dirichlet tessellation around all these setts (Don­ grazing) were obtained from the 2013 national graphical land survey
caster and Woodroffe, 1993; Roper, 2010) and computed the intersec­ (Registre Parcellaire Graphique, RPG). This dataset has been previously
tion between each tile we obtained and a 1,000-m buffer area drawn described (Bouchez-Zacria et al., 2018, 2017). Plots of maize were also
around these reference locations, to obtain realistic proxy for the area provided by the 2013 RPG.
used by badger social groups, i.e. proxy for social group home ranges We used the BD TOPO VEGETATION® database (2015, vector data)
(Bouchez-Zacria et al., 2018) (Fig. 2). provided by the French National Institute of Geographical and Forest
Information (IGN) to extract data on vegetation types. Several classes of
vegetation were available: forests (closed forests — pine, deciduous and

Fig. 2. Area used by badger social groups modelling steps (A: censused setts [grey border points]; B: main setts in clusters of setts within 400 m [red points: main
setts]; C: Dirichlet tessellations [grey border areas] around main setts; D: 1,000-m buffers [dark grey dotted border areas] around main setts; E: intersection between
buffers and Dirichlet tessellations [dark grey border areas]; F: badger tessellations [dark grey border areas]).

4
M. Bouchez-Zacria et al. Preventive Veterinary Medicine 225 (2024) 106146

mixed — and open forests), heathlands, poplar plantations, wooded Table 1

areas, hedgerows, orchards and vineyards. Data about water courses Explanatory variables (n=12) computed for each sett for the analysis and their
were extracted from the BD TOPO HYDROGRAPHY® database (2015, expected impact on the infection status of the sett (↑: in favour of M. bovis
vector data) also provided by the IGN. We extracted soil data from the transmission and higher values associated with infected setts; ↓: detrimental to
topsoil database provided by the European Soil Data Centre (ESDAC), M. bovis transmission and lower values associated with infected setts).
stored in a raster file of 500×500 m pixels describing soil composition Variable Source Computation and Units M. bovis
(20 cm topsoil), and built upon the Land Use and Cover Area frame classes transmission

Statistical survey (LUCAS) (Ballabio et al., 2016; Bouchez-Zacria et al., Badgers

2017). The soil wetness index (SWI) tiles (8×8 km) for our study area B1: Number of Sylvatub Number of setts in setts ↓
neighbouring adjoining
were extracted from the Météo France database (https://www.data.
setts tessellations (3
gouv.fr/fr/datasets/donnees-mensuelles-dindice-dh classes: [2–6];]
umidite-des-sols-pour-le-dispositif-catastroph 6–9];]9–18])
es-naturelles/#resources). This index synthetises the water status of the B2: Proportion Sylvatub, Number of infected setts ↑
surface soil. At a depth of about two metres, it represents the state of the of infected NRL1 setts out of all setts
neighbouring in adjoining
soil’s water reserve in relation to the useful reserve (water available for
setts tessellations (3
plant nutrition): an SWI equal to zero indicates a very dry soil, while an classes: 0;]0–0.2];]
SWI equal to one indicates a water-saturated soil having reached its 0.2–0.67])
useful reserve (https://donneespubliques.meteofrance.fr/client/doc B3: Proportion Sylvatub, Proportion of setts ↑
of infected NRL1 infected setts out of
ument/doc_swi_catnat_268.pdf).
setts* all setts (binary: 0;
1)
2.5. Computation of variables Cattle
C1: Proportion FMA2, Number of pastures pastures ↑
Our variable of interest was the infection status of each sett analysed of pastures NRL1, belonging to
belonging to Sylvatub infected farms out
between 2014 and 2020. To classify each sett as infected or non-infected infected of all pastures (3
(binary variable), we first aggregated all the badgers analysed and farms* classes: 0;]
associated with the same sett. If at least one analysed badger was found 0–0.19];]
to be infected during the study period, the sett was considered as 0.19–0.85])
C2: Surface BD TOPO, Total surface area m2
infected. If at least two analysed badgers were found to be non-infected
↑
area of water FMA2, of rivers and/or
(and no analysed badger was found to be infected) in a sett, it was in pastures of NRL1 bodies of water in
considered as non-infected. Setts with one non-infected badger were infected pastures of infected
excluded from the study, as we assumed this single analysis was insuf­ farms* farms (binary: 0;
ficient to allow us to draw a conclusion on the status of infection. 1])
C3: Length of BD TOPO, Total length of the km
For each sett, we computed the area (km2) of the dedicated tessel­
↑
pasture edge* FMA2 edge between
lation and considered it as an adjustment variable. Fourteen explanatory pastures and
variables were also computed (Table 1), corresponding to three types of woodland (3
risk for M. bovis transmission to badgers, related to: i) badgers (denoted classes: 0;]
0–4.55];]
Bi), ii) cattle (denoted Ci) and iii) landscape features (e.g. due to M. bovis
4.55–17.24])
surviving in the soil (Barbier et al., 2016a; Humblet et al., 2009) and C4: Length of BD TOPO, Total length of the km ↑
denoted Li). Three variables concerned badger-related transmission: the pasture edge FMA2, edge between
number of neighbouring setts (B1) was the total number of setts included belonging to NRL1 pastures of infected
in tessellations intersecting with a 1-m buffer around the sett tessella­ infected farms and
farms* woodland (binary:
tion. We assumed that an elevated number of neighbouring groups could 0; >0)
limit badger movements and thus the risk of transmission (Bouchez-­ Landscape features
Zacria et al., 2017). The following two variables were chosen to capture L1: Surface FMA2, Total surface area km2 ↑
the pressure of infected badgers in the tessellation and in the neigh­ area of maize IGN1 of maize plots (3
* classes: [0–0.10];]
bourhood. The proportion of infected neighbouring setts (B2) was the
0.10–0.31];]
total number of setts found to be infected during the study period and 0.31–0.94])
belonging to other tessellations intersecting with a 1-m buffer around L2: Surface FMA2, Total surface area km2 ↑
the tessellation, divided by the number of setts in these neighbouring area of IGN1 of pastures
tessellations. Indeed, in a previous study carried out in Republic of pastures* belonging to any
farms (3 classes:
Ireland, a badger was found more likely to test positive if caught in a sett [0–0.09];]
close to other infected setts (Byrne et al., 2015). The proportion of 0.09–0.43];]
infected setts in the tessellation (B3) was the number of infected setts (i. 0.43–1.30])
e. with at least one badger found to be infected) during the study period L3: Sett type* BD TOPO, Wood or pasture - ↓↑
Sylvatub edge/hedgerow/
divided by the total number of setts in the tessellation. Four variables
open area/other
concerned the cattle-related risk of M. bovis transmission to badgers, as (binary: wood or
the circulation of M. bovis in this two-species system has been widely pasture edge/other)
studied (Akhmetova et al., 2023; Crispell et al., 2019; Milne et al., 2020; L4: Percentage ESDAC3 Mean percentage of % ↑
Silk et al., 2018): the proportion of pastures belonging to infected farms of clay* clay in topsoil (3
classes:
in the tessellation (C1) was the number of pastures (included fully or [7.22–11.14];]
partially) belonging to farms detected as infected between the year n-4 11.14–15.23];]
(with n the earliest year of badger trapping within the sett) and the year 15.23–22.88])
n’+1 (with n’ the last year of badger trapping within the sett), divided by L5: Soil wetness Météo Mean SWI (3 - ↑
index (SWI)* France classes:
the total number of pastures (included fully or partially) in the tessel­
[0.8–0.84];]
lation. If the sett was infected, only pastures of infected farms with the
(continued on next page)
same spoligotype or with partial M. bovis (i.e. identified but incomplete

5
M. Bouchez-Zacria et al. Preventive Veterinary Medicine 225 (2024) 106146

Table 1 (continued ) types of landscapes: vegetation (hedgerows, forests and heathlands),

Variable Source Computation and Units M. bovis pastures, cropped land (all plots that were not pastures) and background
classes transmission type for all other types of landscapes. We wanted to capture the impact
0.84–0.88];]
on badger movements on M. bovis transmission with these variables
0.88–1.04]) depicting landscape fragmentation and composition. We assumed that a
L6: Simpson IGN1 Fragstats (Fragstats, - ↑ higher fragmentation would enhance the risk of transmission by
diversity 2012) (3 classes: increasing badger movements (specifically when looking for food).
index* [0.011–0.51];]
0.51–0.63];]
0.63–0.66]) 2.6. Statistical analysis
L7: Contagion IGN1 Fragstats (Fragstats, - ↑
index* 2012) (3 classes: The probability of a sett being infected by M. bovis was modelled
[40.23–45.73];]
using a multivariable logistic regression model. Continuous variables
45.73–56.07];]
56.07–97.16])
were first transformed into categorical variables, either with three
*
classes (absence/low, middle, and high) or, when transformed into bi­
In the tessellation; 1: National Geographical Institute; 2: French Ministry of nary variables (absence/presence), according to their distribution (bi­
Agriculture; 3: European Soil Data Centre
nary variables were used when more than one third of observations had
the same value, Table 1). We first checked correlations among variables
profile) were counted as infected. If the sett was non-infected, all the using Spearman’s coefficient (ρ). If ρ > 0.80, one of the two correlated
pastures of farms found to be infected during the n-4 to n’+1 period were variables was excluded from the model, to avoid collinearity (the vari­
counted as infected, regardless of spoligotype. The surface area of water able with the lowest p value was retained, using a univariate analysis).
in the pastures of infected farms (C2) was the total surface area of rivers We excluded from the analysis variables with a single value in more than
and bodies of water in the pastures of infected farms in the tessellation 90% of the observations. We checked the linearity of categorical vari­
(m2). We assumed that humid areas might enhance both M. bovis sur­ ables included in the model using a graph of predicted values of the logit
vival (Barbier et al., 2016a; Humblet et al., 2009) and interspecies univariate model versus the variable values. We then checked the
transmission. The length of the edge with adjoining pastures in the sett absence of multicollinearity in the dataset based on the variance infla­
tessellation (C3) was calculated by i) computing the intersection be­ tion factor (VIF), using a threshold of 10 (Dohoo et al., 2009). To obtain
tween woodlands (including forests, poplar plantations and wooded the most parsimonious final model, we used a backward variable se­
areas) and pastures, ii) transforming the resulting polygons into lines lection, according to the AIC. We computed odds ratios (ORs) and their
and iii) computing the total length of these lines intersecting the associated 95% confidence intervals. We also computed the area under
tessellation (km). The length of the edge adjoining pastures belonging to the ROC curve (AUC) and the associated 95% confidence interval
infected farms in the sett tessellation (C4) was computed in the same (Dohoo et al., 2009) to assess the model’s quality of fit (an AUC between
way, but only for pastures belonging to infected farms (with the same 0.7 and 0.8 indicates acceptable discrimination by the model, while an
selection as for the proportion of infected farms, with respect to the sett AUC greater than 0.8 signals excellent discrimination (Dohoo et al.,
of interest). Finally, seven variables concerned the landscape features 2010)). We performed an internal validation using a leave-one-out
that could affect the risk of M. bovis transmission to badgers: the surface approach. For that purpose, we fitted the model repeatedly (as many
area of maize (L1), an attractive component of the badger’s diet, was times as there are variables of interest), with one observation left out at
calculated using i) the intersection between all plots of maize included in each iteration (Dohoo et al., 2009). We obtained an average AUC and
the tessellation and ii) the total surface area of the resulting polygons compared it with the AUC obtained with the full dataset (close values
(km2). As badgers also eat a large quantity of earthworms (Barbier et al., signalling the model’s reliability).
2016b; Roper, 2010) that they could easily find in pastures, we We used QGIS® software version 3.4.7 (QGIS Development Team,
computed the surface area of pastures in each sett tessellation (L2) 2016) and R® software version 4.2.1 (R Development Core Team, 2016).
similarly but with plots of pastures. The sett type (L3) was determined Spatial computations in R were done with package sf (Pebesma, 2018),
by i) computing the intersection between a 10-m buffer around the sett the contagion index was computed using Fragstats® software version
and wooded areas, ii) computing the intersection between a 3-m buffer 4.2 (Fragstats, 2012). We used MASS (Venables and Ripley, 2002), car
around the sett and hedgerows, and iii) comparing the surface areas. (Fox and Weisberg, 2019) and pROC (Robin et al., 2011) packages for
Four types of setts were thus created: “wooded or pasture edge” (setts for the statistical analysis.
which the surface area of the 10-m buffer was equal to the surface area of
its intersection with wooded areas or setts for which the surface area of 3. Results
the intersection with wooded areas was smaller than the surface area of
the 10-m buffer, without a 3-m buffer intersection with hedgerows), 3.1. Badgers
“hedgerow” (setts without wooded areas in the 10-m buffer but hedge­
rows included in the 3-m buffer), “open area” (setts without wooded In the study area, 3449 badgers were analysed in the framework of
areas in the 10-m buffer and without hedgerows in the 3-m buffer) and Sylvatub between 2014 and 2021. After checking information on trap­
“other” (setts with wooded areas in the 10-m buffer and hedgerows in ping locations (incoherent or incomplete GPS locations) and excluding
the 3-m buffer) setts. The purpose of computing this variable was to find those analysed in 2021, 1021 badgers were finally included to complete
a type of sett environment that potentially enhanced (or reduced) sett census data. We excluded 223 badgers out of the 1021 with
M. bovis transmission. The percentage of clay (L4) (which facilitates the exploitable locations, because of the lack of information on analysis
survival of M. bovis in the soil (Barbier et al., 2017)) in each sett results. These excluded badgers belonged to 464 distinct setts: 234 with
tessellation was determined through i) the intersection of the sett one badger analysed and not infected [class 1], 22 with no conclusive
tessellation with tiles of clay and ii) computing the mean percentage of analysis [class 2], 125 with at least two badgers analysed and not
clay in the tessellation. The soil wetness index (SWI) of each sett infected [class 3], and 83 with at least one badger found to be infected
tessellation (L5) was determined (as M. bovis survives longer in wet soils during the study period [class 4]. The vast majority (90%) of badgers
(Humblet et al., 2009; Young et al., 2005)) in the same way as L4, with included in our study came from traps (some were hunted badgers − 4%-
intersections between the tiles of SWI and the sett tessellation. The last and road-killed badgers − 6%). For setts, 4% of setts were associated
two variables were the Simpson diversity (L6) and contagion (L7) in­ only with road-killed badgers or wildlife mortality surveillance (3% of
dexes (Fragstats, 2012), computed for each sett tessellation using four infected setts and 1% of non-infected setts) and 1.5% were associated

6
M. Bouchez-Zacria et al. Preventive Veterinary Medicine 225 (2024) 106146

with shooting (0.5% of infected setts and 1% of non-infected setts). To in 1,000-m buffer areas around setts detected as being infected (μ=2.6,
the first 2668 setts censused between 2013 and 2015, 22 setts could be min=1, max = 8) and setts detected as non-infected (μ=2.7, min=1,
added through new locations corresponding to badger trapping. Thus, max=10) was not significant (p=1 with the Kolmogorov-Smirnov test,
2690 setts were considered for grouping into clusters for setts within distributions being non-normal (significant Shapiro test, respectively,
400 m of each other. Finally, 1974 clusters were found, with a pre­ p=1.3.10− 7 and p=9.1.10− 11)).
dominance of isolated sett clusters (Fig. 3). This allowed us to compute The presence of cattle pastures in 1,000-m buffer areas around
the same number as tessellations to model the badgers’ area use (Fig. 2, infected and non-infected setts was similar, with a non-significant dif­
Fig. 1B). We excluded setts belonging to classes 1 and 2, as well as those ference between the number of pastures around infected (μ=16.1,
belonging to border tessellations (n=13 setts belonging to 10 distinct min=3, max=34) and non-infected setts (μ=16.9, min=6, max=32)
tessellations) because of an absence of information on the badgers’ (Student test: p=0.78; non-significant Shapiro test, respectively: p= 0.49
neighbourhood. Finally, we included 119 non-infected and 81 infected and p=0.14). There was no significant difference between the number of
setts (Fig. 1C) (n=200 setts (Fig. 3A)) belonging to, respectively, 115 pastures belonging to infected farms in the 1,000-m buffer areas around
and 77 tessellations (n=188 tessellations, four of them hosting both infected setts (μ=4.1, min=1, max=13) and non-infected setts (μ=3.6,
infected and non-infected setts (Fig. 1C, Fig. 3B)). The average area of a min=1, max=13) (p=0.36 with the Kolmogorov-Smirnov test), the two
tessellation was of 11.9 ha (min: 3.2; max: 28.1; SD: 4.8) when consid­ distributions being not normal (Shapiro tests, respectively: p=1.2.10− 5
ering all tessellations, of 13.2 ha (min: 4.1; max: 28.1; SD: 5.5) for tes­ and p=1.3.10− 7).
sellations associated with infected setts and of 11.0 ha (min: 3.2; max: The distribution of variables differed according to the infection sta­
24.6; SD: 4) for tessellations associated with non-infected setts. tus of the sett (Fig. 5, Table 2). According to their distributions (Sup­
In infected setts, the mean number of infected badgers was 1.15 plementary Materials Figure S1), variables B1, B2, C1, C3, L1, L2, L4, L5,
(maximum three infected badgers) (Fig. 4A) and the mean number of L6 and L7 were transformed into three classes of categorical variables.
non-infected badgers was 0.88 (maximum seven non-infected badgers) Variable B3 (proportion of infected setts in a tessellation) was excluded
(Fig. 4B). In non-infected setts, the mean number of non-infected bad­ as 95% of its values were equal to zero. Variables C2 and L3 were
gers was 2.76 (maximum nine non-infected badgers) (Fig. 4B). converted into binary variables (presence of water in pastures belonging
to infected farms in the tessellation versus absence for C2; wooded/
3.2. Cattle pasture edge sett type versus other type for L3) as, respectively, 79% and
67% of their values were identical. Variable C4 was also transformed
The study area included 6415 pastures intersecting all or part of the into a binary variable as 56% of its values were equal to zero and no
tessellations delimiting the area used by badger social groups. Among pertinent classes could be defined in the rest of the distribution (pres­
them, 405 belonged to 153 farms detected to be infected between 2010 ence of an edge adjoining pastures of infected farms versus absence) (the
and 2021. Nine spoligotypes were found, but a great proportion of these different classes of variables are summarised in Table 1).
infected farms were contaminated with M. bovis spoligotypes shared The correlation matrix (Supplementary Materials Figure S2) pointed
with badgers (60% SB0821 and 31% SB0832, the only two spoligotypes out two correlations between variables i) C4 and C1 (ρ=0.93) and ii) L6
detected in badgers during this period (Bouchez-Zacria et al., 2018, and L7 (ρ=0.89). The univariate analysis of these four variables led us to
2017), while 3% corresponded to partial spoligotypes, 2% were infected retain C4 (p=0.0019; C4: p=0.0020) and L6 (p=0.348; L7: p=0.416). No
with SB0928, 1% with SB0295 and <1% each for five other spoligo­ multicollinearity was found, and eleven variables were included in the
types: SB0120, SB0823, SB0851,SB0121 and SB2539). complete model (B1, B2, C2, C3, C4, L1, L2, L3, L4, L5, L6). We checked
the linearity of categorical variables with the variable of interest (Sup­
3.3. Statistical analysis plementary Materials Figure S3). The best model according to the AIC
(AIC= 226) included (in addition to the area of the sett tessellation, the
The numbers of setts trapped and analysed in 1,000-m buffer areas adjustment variable) variables B2 (proportion of infected neighbouring
around the pastures of infected farms (μ=4.1, min=1, max=23) and setts), C4 (presence of an edge adjoining pastures of infected farms in the
around the pastures of non-infected farms (μ=2.7, min=1, max=25) sett tessellation), L2 (total surface area of pastures in each sett tessel­
were significantly different (p=7.02.10− 12 with the Kolmogorov- lation) and L3 (wooded/pasture edge sett type or not). The proportion of
Smirnov test, the two distributions being not normal (significant Sha­ infected neighbouring setts (B2) and the presence of an edge adjoining
piro test: p<2.2.10− 16)). The comparison of the number of setts analysed pastures of infected farms in the sett tessellation (C4) were significantly

Fig. 3. Distribution of the number of setts (A) and of clusters (B) (A, B: white bars and black statistics: within the total population of setts in the study area; A: blue bars
and blue statistics: setts included in the analysis; B: green bars and green statistics: clusters included in the study analysis).

7
M. Bouchez-Zacria et al. Preventive Veterinary Medicine 225 (2024) 106146

Fig. 4. Distribution of the number of infected and non-infected badgers in infected (A) and non-infected setts (B).

Fig. 5. Distribution of continuous variables considered for the logistic regression model before their transformation into categorical variables.

The discrimination capacity of the model was good (AUC=0.79

Table 2
[0.73–0.86]95%) and the internal validation (n=200 iterations) sug­
Distribution of variable L3 considered for the logistic regression model.
gested that the model was reliable, with mean AUC values from the
Sett status (Y) Sett type (L3) leave-one-out models (0.79) close to the model based on the full dataset.
Wooded/ Hedgerow Open area Total
Pasture edge 4. Discussion
Non-infected 101 2 16 119
Infected 65 2 14 81 We investigated the determinants of badgers’ M. bovis infection in
Total 166 4 30 200 one of the main TB areas in South West France, in the departments of
Pyrénées-Atlantiques and Landes. Using Dirichlet tessellation based on
associated with the probability of being an infected sett (respectively 2690 censused setts, we modelled 1974 badger area use (mean area =
p<0.0001; OR: 3.19 [2.04–5.17]95% and p=0.023; OR: 2.33 11.9 ha; SD: 4.8). Among all the censused setts, the apparent prevalence
[1.13–4.89]95%) (Table 3). of infected setts was of 40.5%, with 81 setts which could be classified as

8
M. Bouchez-Zacria et al. Preventive Veterinary Medicine 225 (2024) 106146

Table 3 through the proportion of infected neighbouring setts (B2) variable. In

Variables included in the final logistic regression model and results (CI: Confi­ the future, studies exploiting features such as total precipitation, soil
dence Interval). temperature (Humblet et al., 2009; Young et al., 2005) or net surface
Variables Odds 95% CI p solar radiation (Broughan et al., 2016) would be useful to describe,
ratio understand and specify the transmission of M. bovis among local badger
Tessellation surface area (km2) 4.18 1.97–9.3 0.0003 populations. Moreover, weather conditions could influence badger
B2: Proportion of neighbouring infected setts 3.19 2.04 – <0.0001 populations estimated abundance: i) badger captures have been found to
(3 classes: 0;]0–0.2];]0.2–0.67]) 5.17 be highest in drizzle, rain and heavy rain weather conditions and when
C4: Presence of an edge adjoining pastures of 2.33 1.13–4.89 0.023
minimum values of temperature ranged from 3 to 8◦ C (Martin et al.,
infected farms (binary: 0; 1)
L2: Surface area of pastures 0.68 0.39–1.15 0.152 2017) and ii) estimation of badger population abundance based on
(3 classes: [0–0.09];]0.09–0.43];] trapping data could be affected by weather conditions scenario (with
0.43–1.30]) over or under-estimation depending on the scenario), and thus might be
L3: Sett type (wooded/ pasture edge sett type 0.46 0.19–1.11 0.082 misrepresentative of the true population abundance (Noonan et al.,
or not)
2015). Therefore, it is not only the association between infectious status
and environmental conditions that could be affected by weather con­
infected and 119 as non-infected. The logistic regression model high­ ditions (via the survival time of the bacteria in the soil), but also the
lighted two determinants associated with the TB infection status of setts, number of neighbouring setts (linked to neighbouring badgers) and
namely 1) the proportion of infected setts in the neighbourhood of the therefore indirectly to the B2 variable.
sett tessellation (variable B2), with an OR of 3.19 (95% CI: 2.04–5.17) A first variable significantly associated with the probability of a sett
and 2) the presence of an edge adjoining pastures belonging to infected being determined as infected by M. bovis during the study period was
farms in the sett tessellation (variable C4) with an OR of 2.33 (95% CI: variable C4 (presence of an edge adjoining pastures belonging to
1.13 – 4.89). The identification of two variables related to both intra- infected farms in the sett tessellation). This variable was related to
and interspecies transmission is in line with previous studies in this area interspecies transmission between cattle and badgers. Indeed, we
(Bouchez-Zacria et al., 2018, 2017). assumed that the presence of such edges in the tessellation would have
The mean tessellation area we obtained (around 12 ha) was lower enabled interspecies contact representing the area used by badgers for
than estimated in two previous studies elsewhere in France (both using foraging and intragroup social interactions around their relative sett, the
the kernel method from radio-telemetry monitoring of badgers): in presence of edges of farms detected as infected during the study period
Ardennes (north), group home ranges were estimated at between 90 and would have increased the risk of M. bovis presence in the tessellation
200 ha (mean values depending on the season) and mean individual environment. This result is consistent with a previous study carried out
home ranges at 23 ha (Bodin, 2005). In Côte-d’Or (north-east), indi­ in the Republic of Ireland, which found a positive and significant asso­
vidual home ranges were estimated at 123.6 ha (mean value) (Payne, ciation between cattle prevalence and badger test-status at a spatial
2014). Another recent Irish study used the same approach to model scale of 1 km around setts (Byrne et al., 2015). This comparison is also
badger territories, with Dirichlet tessellations and a maximum distance all the more relevant as our spatial scale was also limited to 1 km within
from the main sett to the tessellation edge of 2 km (Byrne et al., 2019). our tessellations. This result is consistent with another study conducted
Our constraint condition (which was of 1 km) was less restrictive than in the same study area over the 2007–2019 period (Bouchez-Zacria
last value, but allowed to be closer to home ranges previously described et al., 2023). Using dynamic modelling to represent M. bovis trans­
by Bodin (Bodin, 2005) and Payne (Payne, 2014). In the present study, mission through the cattle-badger system, the authors found that
not only did we group analysed badgers by sett, but we also grouped transmission from cattle to badgers was faster than from badgers to
setts less than 400 m apart before calculating tessellations, as we cattle. Moreover, in the badger groups-cattle farms network, the number
assumed that these ranges were those of groups associated with the sett of transmission edges from cattle (farms) to badger (groups) was greater
of interest. In addition, the studies carried out in Ardennes and Côte d’Or than the number of transmission edges from badgers to cattle (Bou­
were conducted in smaller study areas and may not reflect the possible chez-Zacria et al., 2023). In Northern Ireland, a recent study based on
variability in space used by badger groups. Moreover, environmental genome sequencing and phylogenetic analyses (Akhmetova et al.,
disparities could explain the differences we observed. However, if the 2023), also pointed out that transmission of M. bovis was more frequent
badger tessellations that we depicted and used for calculations in our from cattle (herds) to badgers (individuals) than from badgers to cattle.
model were underestimated, this could have led us to underestimate In England, badger and cattle data collected during the Randomised
variables related to the tessellation area and subsequently the role of Badger Culling Trial (1998–2005) have been analysed through a com­
some area-dependent variables (e.g. C2, C3, C4, L1, L2), though the bination of whole genome sequencing and the identification of trans­
difference was not significant according to the logistic regression model. mission clusters (Tonder et al., 2021). The results showed, among
In addition, understanding spatial distribution is one of the key elements others, that badger-to-cattle transmission was twice as high as
in setting up a rational cull aimed at limiting the spread of a pathogen cattle-to-badger transmission. However, the intraspecific transmissions
(including, among other things, the assessment of population size, or the (cattle to cattle and badger to badger) were more frequent than inter­
identification of individuals to be culled) (Miguel et al., 2020). A badger specific transmissions (intraspecific transmission among cattle being
tracking study in the Pyrénées-Atlantiques – Landes area based on a GPS more common than intraspecific transmission among badgers). Thus,
would be therefore useful in the future to validate or invalidate the the direction of transmission appears to vary according to local contexts
extent of the tessellations as modelled and to improve local control (Akhmetova et al., 2023; Tonder et al., 2021). Thus in United Kingdom,
measures. badger density has been highlighted as a contribution to the spatial
In the study area, no fine-scale meteorological data could be heterogeneity of bTB transmission (Chang et al., 2023). In a low risk bTB
exploited (9 km2 data coverage [https://cds.climate.copernicus.eu/cdsa area, without previous record of persistent infection in cattle and
pp#!/dataset/reanalysis-era5-land-monthly-means], which is not suffi­ without any infection in wildlife, the analysis of phylodynamic trees of
ciently discriminating regarding sett tessellation areas) even though an outbreak revealed a most likely within species transmission, and an
climatic conditions (Broughan et al., 2016; Humblet et al., 2009; Young inter-species transmission dominated by cattle-to-badger events (Rossi
et al., 2005) are known to influence the survival time of M. bovis in the et al., 2022). Nevertheless, in our study area, badgers also play a role in
environment. It would doubtless have been beneficial to be able to the persistence of M. bovis in the epidemiological system, with a pro­
explore smaller-scale data as the role of indirect contact was highlighted longed infection period in badger groups (2.9 – 5.7 years) (Bouchez-­
Zacria et al., 2023).

9
M. Bouchez-Zacria et al. Preventive Veterinary Medicine 225 (2024) 106146

In the study area, the recommendations made to trappers were to 5. Conclusion

capture for analysis those badgers found close to the pastures of infected
farms. Our results confirmed this effort as the numbers of setts trapped Our study highlighted the role of intraspecific and interspecific
and analysed in 1,000-m buffer areas around the pastures of infected transmission of M. bovis in the infection of badgers in a TB-infected area
farms and around the pastures of non-infected farms were significantly in South West France. This result confirms the multifactorial aspect of
different. It would be pertinent to consider whether these recommen­ M. bovis transmission and is consistent with previous knowledge on
dations led to differences in sampling. However, the comparison of the badger behaviour, which favours both direct transmission (via social
number of setts analysed in 1,000-m buffer areas around setts detected and territorial contacts) and indirect transmission (via latrines or in
as being infected and setts detected as non-infected was not significant. setts). The survival of M. bovis is also a determining factor in indirect
This result was consistent with a similar trapping pressure around transmission. Our study would benefit from further investigations into
infected and non-infected setts. Similarly, the presence of cattle pastures the local spatial organisation of badgers, as well as meteorological data,
in 1,000-m buffer areas around infected and non-infected setts was possibly related to the survival time of M. bovis in the soil. We none­
similar, with a non-significant difference between the number of pas­ theless confirmed the need to implement measures to limit intraspecific
tures around infected and non-infected setts. transmission and, in this respect, the vaccination of badgers would be
A final comparison of interest was the absence of any significant beneficial in order to reduce infectious pressure on this population.
difference between the number of pastures belonging to infected farms
in the 1,000-m buffer areas around infected setts and non-infected setts. CRediT authorship contribution statement
This result seems counter-intuitive when compared with the results of
our model, in which variable C4 (presence of an edge adjoining a pasture Bouchez-Zacria Malika: Writing – review & editing, Writing –
belonging to an infected farm) was significantly related to the infected original draft, Software, Formal analysis, Data curation, Conceptuali­
status of the sett. However, it should be remembered that the tessellation zation. Jabert Pierre: Writing – original draft, Validation, Formal
in the model’s spatial approach was more precise than simply consid­ analysis, Conceptualization. Réveillaud Edouard: Writing – review &
ering a buffer area of 1000 m, and it also emphasised land occupation by editing, Validation. Richomme Céline: Writing – review & editing,
the social group. In addition, the edge variable specified cattle-badger Validation, Supervision, Conceptualization. Marsot Maud: Writing –
interactions: the linear aspect of the interface might play a more review & editing, Validation, Supervision, Project administration,
important role in interspecific transmission than its length. Formal analysis, Data curation, Conceptualization.
A second variable significantly associated with the probability of a
sett being found to be infected by M. bovis during the study period, was Declaration of Competing Interest
variable B2 (proportion of infected neighbouring setts), linked to the
badger population and accounting for intraspecies infectious pressure. All authors declare that they have no conflicts of interest.
In other words, this result would suggest that, between 2014 and 2020,
within the badger population we studied, some of the M. bovis trans­ Acknowledgements
mission was intraspecific. This result is in line with the territorial
behaviour of badgers, and the risk of transmission through indirect The authors would like to thank the DDETSPP40, the DDPP64, the
contact (Drewe et al., 2013), especially at boundary latrines (Roper departmental federation of hunters from Pyrénées-Atlantiques (FDC64),
et al., 1993), or direct contacts especially in the case of males, who all the hunters in the Pyrénées-Atlantiques and Landes departments,
behave aggressively to win reproduction rights and defend their terri­ badger trappers and departmental trapping associations, as well as wolf-
tory (Corner et al., 2011). Our result is also consistent with a similar hunting lieutenants for the census of badger setts and the collection of
study carried out in Ireland at the badger level: analysis of factors badgers. They are also grateful to the departmental directorates for so­
influencing infection in culled badgers using a mutivariable model cial cohesion and population protection (DDETSPP 40) and more spe­
showed that badgers were more likely to be test positive if they were cifically Agnès Lequy (DDPP64) and Lionel Lesterlin (DDPP64 and
trapped in a sett close to other infected setts (Byrne et al., 2015). To DDETSPP40) for the programme’s organisation and logistics. Our final
control M. bovis infection in badgers, an effective measure would thanks go to Julie Rivière and Stéphanie Desvaux, national coordinator
therefore be to reduce this pressure by reducing the number of infected of Sylvatub.
badgers in order to avoid a domino effect from one home range to
another. This is already the objective of the TB control programme in Appendix A. Supporting information
France, which regulates the trapping and culling of badgers in
TB-infected areas (Arrêté du 7 décembre, 2016). However, this measure Supplementary data associated with this article can be found in the
has been implemented since 2013 in the study area (Pyrénées-Atlantiques online version at doi:10.1016/j.prevetmed.2024.106146.
– Landes), but has not yet enabled us to emerge from the endemic phase
(Delavenne et al., 2020). The disruptive effect, not yet studied in France, References
could not be excluded (Macdonald et al., 2006; Prentice et al., 2019).
Indeed, depending on local context, capturing badgers could create Akhmetova, A., Guerrero, J., McAdam, P., Salvador, L.C.M., Crispell, J., Lavery, J.,
some vacuum effect with the dispersion of individuals seeking for new Presho, E., Kao, R.R., Biek, R., Menzies, F., Trimble, N., Harwood, R., Pepler, P.T.,
Oravcova, K., Graham, J., Skuce, R., du Plessis, L., Thompson, S., Wright, L.,
home ranges or territorial disruption with an increase of badger move­
Byrne, A.W., Allen, A.R., 2023. Genomic epidemiology of Mycobacterium bovis
ments between groups and even increased ranging behaviour (Carter infection in sympatric badger and cattle populations in Northern Ireland. Micro
et al., 2007). In addition, stress-induced immunosuppression could not Genom. 9 https://doi.org/10.1099/mgen.0.001023.
Allen, A.R., Ford, T., Skuce, R.A., 2021. Does Mycobacterium tuberculosis var. bovis
be excluded, leading to greater spread of M. bovis in the population
Survival in the Environment Confound Bovine Tuberculosis Control and
(Gallagher and Clifton-Hadley, 2000). The prospect of vaccination Eradication? A Literature Review. Vet. Med. Int. 2021, e8812898 https://doi.org/
therefore appears to be a promising and recommended alternative 10.1155/2021/8812898.
and/or complementary measure (Beltrán-Beck et al., 2014; Chambers Anich, N.M., Benson, T.J., Bednarz, J.C., 2009. Estimating territory and home-range
sizes: do singing locations alone provide an accurate estimate of space use? Auk 126,
et al., 2011; Gormley et al., 2017; Lesellier et al., 2009). Such a solution 626–634. https://doi.org/10.1525/auk.2009.08219.
is currently the focus of ongoing studies in France (Anses, 2019; Payne Anon., 1964. Council Directive 64/432/EEC of 26 June 1964 on animal health problems
et al., 2022). affecting intra-Community trade in bovine animals and swine, DD I.
Anses, 2021. Avis et rapport relatifs à une demande d’évaluation du rôle
épidémiologique du renard dans la transmission de la tuberculose bovine (No.
Saisine n◦ 2018-SA-0179). Maisons-Alfort, France.

10
M. Bouchez-Zacria et al. Preventive Veterinary Medicine 225 (2024) 106146

Anses, 2019. Gestion de la tuberculose bovine et des blaireaux - Avis de l’Anses - Rapport Trewby, H., Wilson, D.J., Zadoks, R.N., Delahay, R.J., Kao, R.R., 2019. Combining
d’expertise collective révisé. genomics and epidemiology to analyse bi-directional transmission of Mycobacterium
Arrêté du 7 décembre 2016 relatif à certaines mesures de surveillance et de lutte contre bovis in a multi-host system. Elife 8. https://doi.org/10.7554/eLife.45833.
la tuberculose lors de la mise en évidence de cette maladie dans la faune sauvage Delahay, R.J., Brown, J.A., Mallinson, P.J., Spyvee, P.D., Handoll, D., Rogers, L.M.,
(NOR: AGRG1635531A), 2016. Cheeseman, C.L., 2000. The use of marked bait in studies of the territorial
Balestrieri, A., Cardarelli, E., Pandini, M., Remonti, L., Saino, N., Prigioni, C., 2016. organization of the European Badger (Meles meles). Mammal. Rev. 30, 73–87.
Spatial organisation of European badger (Meles meles) in northern Italy as assessed https://doi.org/10.1046/j.1365-2907.2000.00058.x.
by camera-trapping. Eur. J. Wildl. Res. 62, 219–226. https://doi.org/10.1007/ Delavenne, C., Pandolfi, F., Girard, S., Réveillaud, E., Jabert, P., Boschiroli, M.-L.,
s10344-016-0996-y. Dommergues, L., Garapin, F., Keck, N., Martin, F., Moussu, M., Philizot, S., Rivière,
Ballabio, C., Panagos, P., Monatanarella, L., 2016. Mapping topsoil physical properties at J., Tourette, I., Calavas, D., Dupuy, C., Dufour, B., Chevalier, F., 2020. Tuberculose
European scale using the LUCAS database. Geoderma 261, 110–123. https://doi. bovine: bilan et évolution de la situation épidémiologique entre 2015 et 2017 en
org/10.1016/j.geoderma.2015.07.006. France métropolitaine. Bulletin épidémiologique 91, 12.
Barbier, E., Boschiroli, M. l, Gueneau, E., Rochelet, M., Payne, A., de Cruz, K., Blieux, A. Dohoo, I., Martin, W., Stryhn, H., 2009. Veterinary epidemiologic research, Second
l, Fossot, C., Hartmann, A., 2016a. First molecular detection of Mycobacterium bovis edition. ed. VER Inc., Charlottetown, Prince Edward Island, Canada.
in environmental samples from a French region with endemic bovine tuberculosis. Dohoo, I.R., Martin, S.W., Stryhn, H., 2010. Measures of association, Veterinary
J. Appl. Microbiol 120, 1193–1207. https://doi.org/10.1111/jam.13090. Epidemiologic Research. AVC Inc., Charlottetown, Prince Edward Island, canada, pp.
Barbier, E., Chantemesse, B., Rochelet, M., Fayolle, L., Bollache, L., Boschiroli, M.L., 135-150.
Hartmann, A., 2016b. Rapid dissemination of M. bovis from cattle dung to soil by the Doncaster, C.P., Woodroffe, R., 1993. Den site can determine shape and size of badger
earthworm Lumbricus terrestris. Vet. Microbiol. https://doi.org/10.1016/j. territories: implications for group-living. Oikos 66, 88–93. https://doi.org/10.2307/
vetmic.2016.01.025. 3545199.
Barbier, E., Rochelet, M., Gal, L., Boschiroli, M.L., Hartmann, A., 2017. Impact of Drewe, J.A., O’connor, H.M., Weber, N., McDonald, R.A., Delahay, R.J., 2013. Patterns of
temperature and soil type on Mycobacterium bovis survival in the environment. PLOS direct and indirect contact between cattle and badgers naturally infected with
ONE 12, e0176315. https://doi.org/10.1371/journal.pone.0176315. tuberculosis. Epidemiol. Infect. 141, 1467–1475. https://doi.org/10.1017/
Beltrán-Beck, B., de la Fuente, J., Garrido, J.M., Aranaz, A., Sevilla, I., Villar, M., S0950268813000691.
Boadella, M., Galindo, R.C., Pérez de la Lastra, J.M., Moreno-Cid, J.A., Fernández de Fox, J., Weisberg, S., 2019. An R Companion to Applied Regression, Third Edition. Sage,,
Mera, I.G., Alberdi, P., Santos, G., Ballesteros, C., Lyashchenko, K.P., Minguijón, E., Thousand Oaks CA (URL: https://socialsciences.mcmaster.ca/jfox/Books/
Romero, B., de Juan, L., Domínguez, L., Juste, R., Gortázar, C., 2014. Oral Companion).
vaccination with heat inactivated Mycobacterium bovis activates the complement Fragstats, 2012. McGarigal, K., SA Cushman, and E Ene. 2012. FRAGSTATS v4: Spatial
system to protect against tuberculosis. PLoS ONE 9, e98048. https://doi.org/ pattern analysis program for categorical and continuous maps. Computer software
10.1371/journal.pone.0098048. program produced by the authors at the University of Massachusetts, Amherst. 〈http:
Bodin, C., 2005. Partage de l’espace et relation de voisinage dans une population de //www.umass.edu/landeco/research/fragstats/fragstats.html〉.
blaireaux européens (Meles meles) (Thèse universitaire). Université Montpellier II, Gallagher, J., Clifton-Hadley, R.S., 2000. Tuberculosis in badgers; a review of the disease
Montpellier. and its significance for other animals. Res. Vet. Sci. 69, 203–217. https://doi.org/
Bouchez-Zacria, M., Courcoul, A., Durand, B., 2018. The distribution of bovine 10.1053/rvsc.2000.0422.
tuberculosis in cattle farms is linked to cattle trade and badger-mediated contact Gormley, E., Bhuachalla, D.N., O’Keeffe, J., Murphy, D., Aldwell, F.E., Fitzsimons, T.,
networks in south-western France, 2007–2015. Front. Vet. Sci. 5 https://doi.org/ Stanley, P., Tratalos, J.A., McGrath, G., Fogarty, N., Kenny, K., More, S.J.,
10.3389/fvets.2018.00173. Messam, L.L.M., Corner, L.A.L., 2017. Oral Vaccination of Free-Living Badgers (Meles
Bouchez-Zacria, M., Courcoul, A., Jabert, P., Richomme, C., Durand, B., 2017. meles) with Bacille Calmette Guérin (BCG) Vaccine Confers Protection against
Environmental determinants of the Mycobacterium bovis concomitant infection in Tuberculosis. PLOS ONE 12, e0168851. https://doi.org/10.1371/journal.
cattle and badgers in France. Eur. J. Wildl. Res. 63, 74. https://doi.org/10.1007/ pone.0168851.
s10344-017-1131-4. Haddad, N., Ostyn, A., Karoui, C., Masselot, M., Thorel, M.F., Hughes, S.L., Inwald, J.,
Bouchez-Zacria, M., Ruette, S., Richomme, C., Lesellier, S., Payne, A., Boschiroli, M.-L., Hewinson, R.G., Durand, B., 2001. Spoligotype Diversity of Mycobacterium bovis
Courcoul, A., Durand, B., 2023. Analysis of a multi-type resurgence of Mycobacterium Strains Isolated in France from 1979 to 2000. J. Clin. Microbiol. 39, 3623–3632.
bovis in cattle and badgers in Southwest France, 2007-2019. Vet. Res 54, 41. https:// https://doi.org/10.1128/JCM.39.10.3623-3632.2001.
doi.org/10.1186/s13567-023-01168-8. Hauer, A., Michelet, L., Cochard, T., Branger, M., Nunez, J., Boschiroli, M.-L., Biet, F.,
Broughan, J.M., Judge, J., Ely, E., Delahay, R.J., Wilson, G., Clifton-Hadley, R.S., 2019. Accurate Phylogenetic Relationships Among Mycobacterium bovis Strains
Goodchild, A.V., Bishop, H., Parry, J.E., Downs, S.H., 2016. A review of risk factors Circulating in France Based on Whole Genome Sequencing and Single Nucleotide
for bovine tuberculosis infection in cattle in the UK and Ireland. Epidemiol. Infect. Polymorphism Analysis. Front. Microbiol. 10.
144, 2899–2926. https://doi.org/10.1017/S095026881600131X. Humblet, M.-F., Boschiroli, M.L., Saegerman, C., 2009. Classification of worldwide
Byrne, A.W., Kenny, K., Fogarty, U., O’Keeffe, J.J., More, S.J., McGrath, G., Teeling, M., bovine tuberculosis risk factors in cattle: a stratified approach. Vet. Res 40, 1–24.
Martin, S.W., Dohoo, I.R., 2015. Spatial and temporal analyses of metrics of https://doi.org/10.1051/vetres/2009033.
tuberculosis infection in badgers (Meles meles) from the Republic of Ireland: Trends Jacquier, M., Vandel, J.-M., Léger, F., Duhayer, J., Pardonnet, S., Queney, G., Kaerle, C.,
in apparent prevalence. Prev. Vet. Med. https://doi.org/10.1016/j. Say, L., Ruette, S., Devillard, S., 2020. Population genetic structures at multiple
prevetmed.2015.10.013. spatial scales: importance of social groups in European badgers. J. Mammal. 101,
Byrne, A.W., O’Keeffe, J., Buesching, C.D., Newman, C., 2019. Push and pull factors 1380–1391. https://doi.org/10.1093/jmammal/gyaa090.
driving movement in a social mammal: context dependent behavioral plasticity at King, H.C., Murphy, A., James, P., Travis, E., Porter, D., Hung, Y.-J., Sawyer, J., Cork, J.,
the landscape scale. Curr. Zool. 65, 517–525. https://doi.org/10.1093/cz/zoy081. Delahay, R.J., Gaze, W., Courtenay, O., Wellington, E.M., 2015. The variability and
Byrne, A.W., O’Keeffe, J., Fogarty, U., Rooney, P., Martin, S.W., 2015. Monitoring trap- seasonality of the environmental reservoir of Mycobacterium bovis shed by wild
related injury status during large-scale wildlife management programmes: an European badgers. Sci. Rep. 5, 12318 https://doi.org/10.1038/srep12318.
adaptive management approach. Eur. J. Wildl. Res. 61 (3), 445–455. https://doi. Lambert, S., Hars, J., Réveillaud, E., Moyen, J.-L., Gares, H., Rambaud, T., Gueneau, E.,
org/10.1007/s10344-015-0916-6. Faure, E., Boschiroli, M.-L., Richomme, C., 2017. Host status of wild roe deer in
Canini, L., Modenesi, G., Courcoul, A., Boschiroli, M.-L., Durand, B., Michelet, L., 2023. bovine tuberculosis endemic areas. Eur. J. Wildl. Res 63, 15. https://doi.org/
Deciphering the role of host species for two Mycobacterium bovis genotypes from the 10.1007/s10344-016-1071-4.
European 3 clonal complex circulation within a cattle-badger-wild boar multihost Lesellier, S., Corner, L., Costello, E., Lyashchenko, K., Greenwald, R., Esfandiari, J.,
system. MicrobiologyOpen 12, e1331. https://doi.org/10.1002/mbo3.1331. Singh, M., Hewinson, R.G., Chambers, M., Gormley, E., 2009. Immunological
Carter, S.P., Delahay, R.J., Smith, G.C., Macdonald, D.W., Riordan, P., Etherington, T.R., responses and protective immunity in BCG vaccinated badgers following
Pimley, E.R., Walker, N.J., Cheeseman, C.L., 2007. Culling-induced social endobronchial infection with Mycobacterium bovis. Vaccine 27, 402–409. https://doi.
perturbation in Eurasian badgers Meles meles and the management of TB in cattle: an org/10.1016/j.vaccine.2008.10.068.
analysis of a critical problem in applied ecology. Proc. R. Soc. Lond. B: Biol. Sci. 274, Macdonald, D.W., Riordan, P., Mathews, F., 2006. Biological hurdles to the control of TB
2769–2777. https://doi.org/10.1098/rspb.2007.0998. in cattle: A test of two hypotheses concerning wildlife to explain the failure of
Chambers, M.A., Rogers, F., Delahay, R.J., Lesellier, S., Ashford, R., Dalley, D., control. Biol. Conserv., Infect. Dis. Mamm. Conserv. 131, 268–286. https://doi.org/
Gowtage, S., Davé, D., Palmer, S., Brewer, J., Crawshaw, T., Clifton-Hadley, R., 10.1016/j.biocon.2006.05.006.
Carter, S., Cheeseman, C., Hanks, C., Murray, A., Palphramand, K., Pietravalle, S., Marsot, M., Bernard, C., Payne, A., Rossi, S., Ruette, S., Desvaux, S., Richomme, C.,
Smith, G.C., Tomlinson, A., Walker, N.J., Wilson, G.J., Corner, L.A.L., Rushton, S.P., Bouchez-Zacria, M., Durand, B., 2023. BACACIX”, a spatial index combining proxies
Shirley, M.D.F., Gettinby, G., McDonald, R.A., Hewinson, R.G., 2011. Bacillus of bovine and badger space use associated with extended Mycobacterium bovis
Calmette-Guérin vaccination reduces the severity and progression of tuberculosis in circulation in France. Prev. Vet. Med. 211, 105817 https://doi.org/10.1016/j.
badgers. Proc. Biol. Sci. 278, 1913–1920. https://doi.org/10.1098/rspb.2010.1953. prevetmed.2022.105817.
Chang, Y., Hartemink, N., Byrne, A.W., Gormley, E., McGrath, G., Tratalos, J.A., Martin, L.E.R., Byrne, A.W., O’Keeffe, J., Miller, M.A., Olea-Popelka, F.J., 2017. Weather
Breslin, P., More, S.J., de Jong, M.C.M., 2023. Inferring bovine tuberculosis influences trapping success for tuberculosis management in European badgers (Meles
transmission between cattle and badgers via the environment and risk mapping. meles). Eur. J. Wildl. Res 63, 30. https://doi.org/10.1007/s10344-017-1089-2.
Front. Vet. Sci. 10. Michelet, L., De Cruz, K., Hénault, S., Tambosco, J., Richomme, C., Réveillaud, É.,
Corner, L.A.L., Murphy, D., Gormley, E., 2011. Mycobacterium bovis infection in the Gares, H., Moyen, J.-L., Boschiroli, M.L., 2018. Mycobacterium bovis infection of red
Eurasian badger (Meles meles): the disease, pathogenesis, epidemiology and control. fox, France. Emerg. Infect. Dis. 24, 1150–1153. https://doi.org/10.3201/
J. Comp. Pathol. 144, 1–24. https://doi.org/10.1016/j.jcpa.2010.10.003. eid2406.180094.
Crispell, J., Benton, C.H., Balaz, D., De Maio, N., Ahkmetova, A., Allen, A., Biek, R., Michelet, L., Durand, B., Boschiroli, M.L., 2020. Tuberculose bovine: Bilan génotypique
Presho, E.L., Dale, J., Hewinson, G., Lycett, S.J., Nunez-Garcia, J., Skuce, R.A., de M. bovis à l’origine des foyers bovins entre 2015 et 2017 en France

11
M. Bouchez-Zacria et al. Preventive Veterinary Medicine 225 (2024) 106146

métropolitaine. Bull. Epidémiologique 91, 1–4 https://be.anses.fr/sites/default/ Reviriego Gordejo, F.J., Vermeersch, J.P., 2006. Towards eradication of bovine
files/O-035_2021-08-04_Tub-genotyp-Michelet_MaqF.pdf. tuberculosis in the European Union. Vet. Microbiol., 4th Int. Conf. Mycobacterium
Miguel, E., Grosbois, V., Caron, A., Pople, D., Roche, B., Donnelly, C.A., 2020. A systemic bovis 112, 101–109. https://doi.org/10.1016/j.vetmic.2005.11.034.
approach to assess the potential and risks of wildlife culling for infectious disease Richomme, C., Boadella, M., Courcoul, A., Durand, B., Drapeau, A., Corde, Y., Hars, J.,
control. Commun. Biol. 3, 1–14. https://doi.org/10.1038/s42003-020-1032-z. Payne, A., Fediaevsky, A., Boschiroli, M.L., 2013. Exposure of wild boar to
Milne, G., Allen, A., Graham, J., Kirke, R., McCormick, C., Presho, E., Skuce, R., Byrne, A. Mycobacterium tuberculosis complex in France since 2000 is consistent with the
W., 2020. Mycobacterium bovis Population Structure in Cattle and Local Badgers: Co- distribution of bovine tuberculosis outbreaks in cattle. PLoS ONE 8, e77842. https://
Localisation and Variation by Farm Type. Pathogens 9, 592. https://doi.org/ doi.org/10.1371/journal.pone.0077842.
10.3390/pathogens9070592. Richomme, C., Réveillaud, E., Moyen, J.-L., Sabatier, P., De Cruz, K., Michelet, L.,
Morris, R.S., Pfeiffer, D.U., Jackson, R., 1994. The epidemiology of Mycobacterium bovis Boschiroli, M.L., 2020. Mycobacterium bovis Infection in Red Foxes in Four Animal
infections. Vet. Microbiol. 40, 153–177. https://doi.org/10.1016/0378-1135(94) Tuberculosis Endemic Areas in France. Microorganisms 8, 1070. https://doi.org/
90053-1. 10.3390/microorganisms8071070.
Noonan, M.J., Rahman, M.A., Newman, C., Buesching, C.D., Macdonald, D.W., 2015. Rivière, J., 2016. Evaluation de dispositif de surveillance de la tuberculose bovine dans la
Avoiding verisimilitude when modelling ecological responses to climate change: the faune sauvage en France à l’aide de méthodes épidémiologique, économique et
influence of weather conditions on trapping efficiency in European badgers (Meles sociologique. Université Paris-Saclay.
meles). Glob. Chang Biol. 21, 3575–3585. https://doi.org/10.1111/gcb.12942. Rivière, J., Strat, Y.L., Dufour, B., Hendrikx, P., 2015. Sensitivity of Bovine Tuberculosis
O’Mahony, D.T., 2014. Badger-cattle interactions in the rural environment - implications Surveillance in Wildlife in France: A Scenario Tree Approach. PLOS ONE 10,
for bovine tuberculosis transmission. TB & Brucellosis Policy Branch, Department of e0141884. https://doi.org/10.1371/journal.pone.0141884.
Agriculture and Rural Development, Northern Ireland. Robin, X., Turck, N., Hainard, A., Tiberti, N., Lisacek, F., Sanchez, J.-C., Müller, M., 2011.
Palisson, A., Courcoul, A., Durand, B., 2017. Analysis of the spatial organization of pROC: an open-source package for R and S+ to analyze and compare ROC curves.
pastures as a contact network, implications for potential disease spread and BMC Bioinformatics, 12, p. 77. DOI: 10.1186/1471-2105-12-77 <http://www.
biosecurity in livestock, France, 2010. PLOS ONE 12, e0169881. https://doi.org/ biomedcentral.com/1471-2105/12/77/>.
10.1371/journal.pone.0169881. Roper, T.J., 2010. Badger, The new naturalist library. Collins, London.
Palisson, A., Courcoul, A., Durand, B., 2016. Role of cattle movements in bovine Roper, T.J., Conradt, L., Butler, J., Christian, S.E., Ostler, J., Schmid, T.K., 1993.
tuberculosis spread in France between 2005 and 2014. PLoS ONE 11, e0152578. Territorial Marking With Faeces in Badgers (Meles meles): a Comparison of Boundary
https://doi.org/10.1371/journal.pone.0152578. and Hinterland Latrine Use. Behaviour 127, 289–307. https://doi.org/10.1163/
Payne, A., 2014. Rôle de la faune sauvage dans le système multi-hôtes de Mycobacterium 156853993×00074.
bovis et risque de transmission entre faune sauvage et bovins. Etude expérimentale Rossi, G., Crispell, J., Brough, T., Lycett, S.J., White, P.C.L., Allen, A., Ellis, R.J.,
en Côte d’Or. Université de Lyon. Gordon, S.V., Harwood, R., Palkopoulou, E., Presho, E.L., Skuce, R., Smith, G.C.,
Payne, A., Boschiroli, M.L., Gueneau, Eric, Moyen, J.-L., Rambaud, T., Dufour, B., Gillot- Kao, R.R., 2022. Phylodynamic analysis of an emergent Mycobacterium bovis
Fromont, E., Hars, J., 2013. Bovine tuberculosis in “Eurasian” badgers (Meles meles) outbreak in an area with no previously known wildlife infections. J. Appl. Ecol. 59,
in France. Eur. J. Wildl. Res. 59, 331–339. https://doi.org/10.1007/s10344-012- 210–222. https://doi.org/10.1111/1365-2664.14046.
0678-3. Silk, M.J., Drewe, J.A., Delahay, R.J., Weber, N., Steward, L.C., Wilson-Aggarwal, J.,
Payne, A., Philipon, S., Hars, J., Dufour, B., Gilot-Fromont, E., 2017. Wildlife interactions Boots, M., Hodgson, D.J., Croft, D.P., McDonald, R.A., 2018. Quantifying direct and
on baited places and waterholes in a French area infected by bovine tuberculosis. indirect contacts for the potential transmission of infection between species using a
Front. Vet. Sci. 3 https://doi.org/10.3389/fvets.2016.00122. multilayer contact network. Behaviour 155, 731–757. https://doi.org/10.1163/
Payne, A., Ruette, S., Jacquier, M., Richomme, C., Lesellier, S., Middleton, S., 1568539X-00003493.
Duhayer, J., Rossi, S., 2022. Estimation of Bait Uptake by Badgers, Using Non- Tonder, A.J., van, Thornton, M.J., Conlan, A.J.K., Jolley, K.A., Goolding, L., Mitchell, A.
invasive Methods, in the Perspective of Oral Vaccination Against Bovine P., Dale, J., Palkopoulou, E., Hogarth, P.J., Hewinson, R.G., Wood, J.L.N.,
Tuberculosis in a French Infected Area. Front. Vet. Sci. 9. Parkhill, J., 2021. Inferring Mycobacterium bovis transmission between cattle and
Pebesma, E., 2018. Simple Features for R: Standardized Support for Spatial Vector Data. badgers using isolates from the Randomised Badger Culling Trial. PLOS Pathog. 17,
R. J. 10 (1), 439–446. https://doi.org/10.32614/RJ-2018-009. e1010075 https://doi.org/10.1371/journal.ppat.1010075.
Prentice, J.C., Fox, N.J., Hutchings, M.R., White, P.C.L., Davidson, R.S., Marion, G., Venables, W., Ripley, B., 2002. Modern Applied Statistics with S.
2019. When to kill a cull: factors affecting the success of culling wildlife for disease Wilesmith, J.W. (Central V.L., 1991. Ecological and epidemiological findings from a
control. J. R. Soc. Interface 16, 20180901. https://doi.org/10.1098/rsif.2018.0901. prospective study of a naturally infected badger population. Publication - Veterinary
QGIS Development Team, 2016. QGIS Geographic Information System. Open Source Continuing Education, Massey University (New Zealand).
Geospatial Foundation Project. 〈http://www.qgis.org/〉. Woodroffe, R., Donnelly, C.A., Ham, C., Jackson, S.Y.B., Moyes, K., Chapman, K.,
R Development Core Team, 2016. R: A language and environment for statistical Stratton, N.G., Cartwright, S.J., 2016. Badgers prefer cattle pasture but avoid cattle:
computing. R Foundation for Statistical Computing, Vienna, Austria. 〈http://www. implications for bovine tuberculosis control. Ecol. Lett. 19, 1201–1208. https://doi.
R-project.org〉. org/10.1111/ele.12654.
Réveillaud, É., Desvaux, S., Boschiroli, M.-L., Hars, J., Faure, É., Fediaevsky, A., Young, J.S., Gormley, E., Wellington, E.M.H., 2005. Molecular detection of
Cavalerie, L., Chevalier, F., Jabert, P., Poliak, S., Tourette, I., Hendrikx, P., Mycobacterium bovis and Mycobacterium bovis BCG (Pasteur) in soil. Appl. Environ.
Richomme, C., 2018. Infection of wildlife by Mycobacterium bovis in France Microbiol 71, 1946–1952. https://doi.org/10.1128/AEM.71.4.1946-1952.2005.
Assessment through a national surveillance system, Sylvatub. Front. Vet. Sci. 5 Zanella, G., Duvauchelle, A., Hars, J., Moutou, F., Boschiroli, M.L., Durand, B., 2008.
https://doi.org/10.3389/fvets.2018.00262. Patterns of lesions of bovine tuberculosis in wild red deer and wild boar. Vet. Rec.
163, 43–47.

12