Use of Epidemiological Analytic Tools and Molecular Characterization Techniques to Inform Bovine Tuberculosis Eradication Programs

PhD Thesis defense by Pilar Pozo Piñol at the VISAVET Centre of the Complutense University of Madrid

December 17^th, 2020

Bovine tuberculosis (bTB) is an infectious disease with a complex epidemiology at the human-animal-environment interface which, despite the application of extensive control measures, is still prevalent in livestock and wildlife in several regions of Spain. The disease, caused by Mycobacterium bovis (and to a lesser extent M. caprae), remains a concern due to its negative impact on animal health and welfare, together with the important consequences for production and animal trade that its presence imposes. Eradication programs of bTB in cattle in Spain are based on a combination of regular test and slaughter policies using mostly the single intradermal test (SIT), the interferon-gamma (IFN-γ) assay as an ancillary test in infected herds, and abattoir post-mortem inspection. Disease eradication at the herd level is challenging due to several factors including the limitations in the sensitivity of available diagnostic techniques, the risk of introduction of undetected infected animals, the proximity to positive farms, and the contact with infected wildlife reservoirs. This thesis is based on the application of different analytic tools such as social network analysis, epidemiological models and whole genome sequencing (WGS) to quantify the relative importance of risk factors impairing disease control in Spain using a multidisciplinary approach.

Chapter 1 includes the characterization of the cattle movement network in Castilla y Leon, a high bTB prevalence area in Spain, and the assessment of the possible role of between-farm movements on the observed pattern of bTB breakdowns between 2010 and 2015 in the region. A high degree of heterogeneity in the contact patterns was observed, as most units had a small number of connections, but a small proportion of units were highly connected, what translates into an also heterogeneous risk of contributing to bTB spread if infected. The observed pattern of bTB cases in Castilla y Leon revealed that bTB-positive farms were not more connected than would be expected by chance with other positive farms through direct movements. However, positive farms were more connected to other units and moved more cattle than negative farms, suggesting an association between cattle movements and infection. Additionally, positive farms were clustered spatially, with positive farms being more likely to be within 10 km of other positive farms than expected. This study emphasizes the potential for units with certain characteristics and located at pivotal positions in the network to act as sources of bTB if movements of infected cattle occur before its detection.

Chapter 2 comprises the study of the factors contributing to disease persistence in bTB-positive herds between 2010 and 2017 in Castilla y Leon. This study investigates the impact of farm characteristics and bTB breakdown-specific variables on the risk of experiencing prolonged bTB breakdowns and becoming chronically infected herds (outbreaks with duration ≥784 days) using survival analysis and multivariable models. Province of origin and municipality bTB herd prevalence, production type, herd size, entry of animals in the last three years, and number of skin test reactors and culture-positive animals in the disclosing test were associated with increased breakdown duration. In contrast, production type was not associated with experiencing a chronic breakdown once the farm is infected. This study provides useful information on the characteristics that influence the persistence of bTB in highly prevalent areas and can be used to identify infected herds that may experience chronic outbreaks at an early stage.

Chapter 3 aims to evaluate the efficiency of slaughterhouse surveillance in detecting bTB and subsequently confirming the disease through laboratory tests between 2010 and 2017 in abattoirs in Castilla y Leon. Two multivariable mixed models were fitted to calculate the risk of lesion detection and laboratory confirmation per abattoir in animals that were positive or negative/with no information in ante-mortem tests while accounting for the effect of potential confounding variables. The probability of disclosure of bTB-like lesions among reactors was associated with the type of source unit, the breed and age of the animal, and the season of slaughter, whereas for the negative animals or those with no information on ante-mortem tests the latter was not. A substantial heterogeneity in the detection and laboratory confirmation rates of the abattoirs in Castilla y Leon was detected, especially regarding the risk of bTB-suspect lesion detection in reactors. However, animal and herd level characteristics did not contribute to this variability, indicating that the practices applied in detecting bTB-like lesions were not uniform among abattoirs. Results obtained in this study may help to quantify the baseline detection rates of bTB-like lesions in different animal populations and the observed variability in detection between abattoirs, along with the identification of certain animal- and farm-related factors.

Chapter 4 includes the characterization of the genomic diversity of M. bovis isolates recovered from different animal species and locations of Spain that belong to the third most prevalent spoligotype in Spain (SB0339). The reconstruction of the phylogenetic relationships between the isolates revealed a variable genetic diversity for isolates recovered from the same and different host species. However, this genetic heterogeneity was geographic rather than host species-specific, as isolates recovered from both cattle and wildlife sharing recent common ancestors were more closely related within same provinces compared to isolates from the same species but geographically distant. Limited within-herd genetic diversity between isolates recovered up to ten years apart in five out of ten chronically infected herds was observed. Additionally, closely related isolates coming from wildlife and cattle were also found in certain instances, demonstrating the potential for between-species transmission and the role that wildlife may play in certain settings for disease maintenance in endemic areas in Spain, and highlighting the need to develop improved strategies for controlling bTB in both cattle and wildlife populations.

The combination of epidemiological and molecular tools applied here may help to elucidate the effect of the risk factors hampering bTB control in Spain, especially in areas where disease is endemic. The findings of this thesis demonstrate the usefulness of a multidisciplinary approach to gain insights on the critical points driving bTB occurrence in livestock (and wildlife) in Spain.