Structural equation models to estimate risk of infection and tolerance to bovine mastitis

• Published in: Genetics selection evolution (Paris) Vol. 45; no. 1; p. 6
• Main Authors: Detilleux, Johann, Theron, Léonard, Duprez, Jean-Noël, Reding, Edouard, Humblet, Marie-France, Planchon, Viviane, Delfosse, Camille, Bertozzi, Carlo, Mainil, Jacques, Hanzen, Christian
• Format: Journal Article Web Resource
• Language: English
• Published: France Springer-Verlag 06.03.2013; BioMed Central Ltd; BioMed Central; EDP Sciences
• Subjects: Animal Husbandry; Animal welfare; Animals; Bacteria - isolation & purification; Bacterial diseases; Bacterial infections; bovine mastitis; Cattle; Dairy cattle; dairy cows; Data Interpretation, Statistical; Data mining; Disease Resistance; Disinfection; Drug resistance in microorganisms; Economic models; Epidemiology; Farms; Female; Health aspects; Health risks; herds; Infection; Infections; Infectious diseases; Influence; Laboratories; Life sciences; Mastitis; Mastitis, Bovine - epidemiology; Mastitis, Bovine - etiology; Mastitis, Bovine - microbiology; Medical research; Medicine, Experimental; Methods; Milk; Milk - microbiology; Milking; Models, Statistical; Multivariate statistical analysis; Médecine vétérinaire & santé animale; Parasites; Pathogens; resistance mechanisms; risk; Risk analysis; risk estimate; Risk Factors; Sciences du vivant; Structural equation modeling; Studies; surveys; Teat dips; Teats; Urea; Variables; Veterinary medicine & animal health; Belgium
• ISSN: 1297-9686; 0999-193X; 1297-9686
• DOI: 10.1186/1297-9686-45-6

BACKGROUND: One method to improve durably animal welfare is to select, as reproducers, animals with the highest ability to resist or tolerate infection. To do so, it is necessary to distinguish direct and indirect mechanisms of resistance and tolerance because selection on these traits is believed to have different epidemiological and evolutionary consequences. METHODS: We propose structural equation models with latent variables (1) to quantify the latent risk of infection and to identify, among the many potential mediators of infection, the few ones that influence it significantly and (2) to estimate direct and indirect levels of tolerance of animals infected naturally with pathogens. We applied the method to two surveys of bovine mastitis in the Walloon region of Belgium, in which we recorded herd management practices, mastitis frequency, and results of bacteriological analyses of milk samples. RESULTS AND DISCUSSION: Structural equation models suggested that, among more than 35 surveyed herd characteristics, only nine (age, addition of urea in the rations, treatment of subclinical mastitis, presence of dirty liner, cows with hyperkeratotic teats, machine stripping, pre- and post-milking teat disinfection, and housing of milking cows in cubicles) were directly and significantly related to a latent measure of bovine mastitis, and that treatment of subclinical mastitis was involved in the pathway between post-milking teat disinfection and latent mastitis. These models also allowed the separation of direct and indirect effects of bacterial infection on milk productivity. Results suggested that infected cows were tolerant but not resistant to mastitis pathogens. CONCLUSIONS: We revealed the advantages of structural equation models, compared to classical models, for dissecting measurements of resistance and tolerance to infectious diseases, here bovine mastitis. Using our method, we identified nine major risk factors that were directly associated with an increased risk of mastitis and suggested that cows were tolerant but not resistant to mastitis. Selection should aim at improved resistance to infection by mastitis pathogens, although further investigations are needed due to the limitations of the data used in this study.