Metabolomics of Bovine Respiratory Disease

Summary

Bovine respiratory disease (BRD) is important to prevent for the Australian feedlot industry. Two experiments were conducted to investigate diagnostic techniques for Bovine Respiratory Disease and it's economic impact on feedlot cattle profitability.

Experiment 1 utilised metabolomics to investigate relationships between the blood metabolome of commercial feedlot steers, conventional BRD diagnostic methods and post-mortem lung pathology. Visually sick (n=148) and visually healthy (n=152) steers were pulled from their pens by pen riders and brought to the hospital shed for clinical assessment and blood sampling for metabolomics analysis. Lung lesions indicative of BRD were scored for all trial animals upon slaughter. Nuclear magnetic resonance (NMR) spectrometry was used to identify chemical features in plasma to predict BRD status using classification and regression trees. Due to a lack of a universal gold standard for BRD diagnosis, six reference diagnosis methods were used to define an animal as sick or healthy: pen rider visual diagnosis (PD), rectal temperature diagnosis (TD; rectal temperature ≥ 40 ºC), lung auscultation diagnosis (LAD; lung auscultation score ≥ 2), hospital diagnosis (HD; sick according to either TD or LAD, or both), feedlot diagnosis (FD; sick according to both PD and HD), and lung lesion diagnosis at slaughter (LLD; lung consolidation ≥ 10% and pleurisy score of 2, or pleurisy score of 3). Multiple identified metabolites and unknown NMR features (peaks) showed high correlations with BRD status ranging from +0.68 to -0.71 (P < 0.001). The strongest correlations were found with PD and FD indicating that the blood metabolome reflected visual signs of sickness as recorded by pen riders. Tyrosine, citrate, glutamine, valine and alanine were lower in BRD case animals compared to controls.

Hydroxybutyrate, glucose chains, isoleucine, phenylalanine and creatine were higher in BRD cases compared to control animals (P < 0.001). However, the largest differences between case and control animals were found for metabolites which could not be identified. Blood NMR demonstrated high accuracy (Acc) at detecting BRD defined by PD (Acc=0.85) and FD (Acc=0.81) but were less accurate to detect animals defined as sick according to TD (Acc=0.65), HD (Acc=0.67), LAD (Acc=0.61) and LLD (Acc=0.71). All diagnosis methods required one chemical feature (peak) or metabolite but HD and LLD used 3 and 5 peaks, respectively. The lower accuracy of TD, LAD, HD, and LLD could be due to the metabolome profile reflecting clinical signs at the time when the sample was taken. However, clinical signs such as rectal temperature or lung lesions at slaughter may have peaked or developed either prior to or following blood sample collection. The results indicate that one to five metabolites in the blood of feedlot cattle are useful indicators to detect or confirm BRD in feedlots however the chemical structure of some of these need to be determined. Blood metabolomics shows great potential to aid in defining and confirming BRD cases under commercial feedlot conditions. Future work in this area should incorporate frequent blood sampling and clinical measures starting before exposure to BRD and going throughout infection to slaughter (e.g. serial slaughter of animals).

Experiment 2 determined the economic impact of BRD through analysis of feedlot records and post-mortem lung pathology. The project was undertaken at a commercial feedlot in Southern NSW, Australia, and consisted of records of steers from two separate feedlot trials (Trial 1: started in February, n=898; Trial 2: started in June, n=1,314). Animals were followed from induction to slaughter and had veterinary treatment records collected during the feeding phase and lung lesions indicative of BRD collected at slaughter. The incidence of BRD was 16.6% in late summer (Trial 1) and 9.6% (Trial 2) in autumn and winter. BRD mortalities were approximately 80% of all mortalities and between 2% (Trial 1) and 0.3% (Trial 2) of all animals inducted. Animals treated for BRD three or more times showed carcass 40 kg lighter in both trials and returned $384.97 less in Trial 1 and $284 in Trial 2 compared to healthy animals on a dead out basis. Severe lung lesions were observed in 16% (trial 1) and 6% (trial 2) of the animals slaughtered. Pleurisy score of 3 (lungs adhered to thoracic wall) was observed in 10.7% and 5.6% of animals for trial 1 and 2, respectively. In trial 1, animals with severe lung lesions or pleurisy score of 3 yielded carcases that were 14.31 kg and 18.79 kg less, respectively, compared to animals with normal lungs at slaughter. Net returns were also decreased by $91 per head for these animals. In trial 1, animals with pleurisy score of 3 returned $137 less than the other score on average. In trial 2, lung lesions and pleurisy score had no significant effects on carcass trials or net returns. These results help to gain a greater understanding of the economic cost of BRD to the Australian feedlot industry and can aid in identifying areas to address in order to reduce its impact on commercial feedlots.