Epidemiology of bovine respiratory disease in Australian feedlot cattle

Abstract

Bovine respiratory disease (BRD) is the most important cause of morbidity and mortality in feedlot cattle, both in Australia and worldwide. BRD broadly describes a complex of diseases involving the respiratory system in cattle. It has a multifactorial aetiology involving a combination of susceptible animals, infectious agents and stressors. Most published BRD research is from studies conducted in North America, but there are important differences between Australian and North American beef cattle production and feedlot management practices. The relative importance of particular risk factors at the population level would also be expected to differ between Australia and North America because animal entry characteristics, management practices and environmental exposures differ. In addition, there are conflicting results from the literature relating to associations between some risk factors and BRD.To address these issues a nationwide cohort study was conducted in Australian feedlot cattle to quantify strengths of associations between numerous putative risk factors and BRD, and to determine the population-level impact of relevant risk factors. The main cohort study population comprised 35,131 animals nested within 1,077 groups nested within 170 cohorts (feedlot pens) nested within 14 feedlots. In addition, a subset of 7,450 animals was selected for inclusion in a nested case-control study. The objectives of this study were to: (i) describe the seroprevalences of antibodies to four viruses at induction (i.e. processing at entry) (ii) describe changes in serostatuses six weeks after induction, and (iii) investigate associations between serological risk factors and BRD occurrence. Data relating to numerous putative risk factors were collected from several sources during the course of the study. Novel use of lifetime animal movement data obtained from a nationwide database allowed detailed analysis of putative risk factors describing each animal's prior mixing history, group dynamics, lifetime saleyard exposure and timing of the animal's move to the feedlot. Laboratory analysis of serum samples and nasal swabs allowed the differentiation of animals persistently or transiently infected with bovine viral diarrhoea virus (BVDV). Hence the effects of exposure to BVDV on BRD incidence in the main cohort study population could be assessed.Causal diagrams were used to inform model building by considering a priori biologically plausible pathways. Multilevel Bayesian logistic models were utilised to estimate the effects of putative risk factors. In addition, a parsimonious model was built and used to determine the partitioning of outcome (i.e. BRD) variance at different hierarchical levels; this was used for identifying the most appropriate level for interventions and further research. Population-level effects of important risk factors were calculated and used to rank risk factors and identify management strategies with the largest potential overall effects in reducing BRD risk in Australian feedlot cattle.Several management-related risk factors were identified as having a marked effect on BRD risk at both the animal level and the population level. Factors related to the animal's lifetime mixing history, feedlot move timing and the numbers of animals in groups established at least two weeks before feedlot entry were all very important. The practice of sharing water troughs between feedlot pens had a very large effect; this previously unreported risk factor is readily amenable to intervention. Exposure to BVDV had a moderate population-level effect, providing a measure of the expected impact in feedlot populations if effective programs to prevent BVDV entry into feedlots were implemented. Animal factors (breed, sex and weight) and broad non-specific factors (feedlot region and season of induction) had modest to large population-level effects.Animals that were seropositive to any of four viruses at induction were generally at reduced risk of BRD compared to those that were seronegative, although those with low antibody levels to BVDV appeared to be at increased risk. Animals that were seropositive to increasing numbers of viruses at induction were at reduced risk of BRD compared to those seropositive to fewer viruses. Seroconversion or seroincrease to any of four viruses during the first six weeks on feed was associated with increased risk of BRD. Animals that seroincreased to one virus were at increased risk, and animals that seroincreased to two or more viruses were at markedly increased risk of BRD compared to animals whose serological status did not change.The studies described in this thesis have identified several important management-related risk factors that are amenable to interventions with the potential to markedly reduce BRD incidence in Australian feedlot populations.