Scanning the globe for the best BRD diagnostic tools

When it comes to Bovine Respiratory Disease (BRD), MLA’s latest research has taken a global approach, partnering with universities, research institutes and technology providers from across the world to find the most effective methods for diagnosis in feedlots.  

To support the current method of diagnosis which is based on clinical indicators used by pen riders and livestock staff, non-visual modalities such as wearable technologies, physical tests and laboratory tests are currently being explored.  

The project, led by Dr Tony Batterham from Apiam Animal Health, tested and compared these modalities in a commercial feedlot to help identify opportunities to improve the detection, diagnosis and management of BRD in Australian cattle.

This could be a game changer for animal health, welfare and productivity in Australian feedlots.

The modalities

Remote monitoring systems

Two commercially available remote monitoring systems were tested, both capturing data every five minutes.

• The REDI (Remote Early Disease Identification) system, a type of tracking device, was fitted to wearable collars. For each animal, the REDI system collected information on time at the feed bunk and water trough, time spent in a group or isolated and distance travelled.
  These movement, social and eating/drinking behaviours were associated with health status.
  
  The pro’s: regular information and potential for early detection; based on algorithms that could be continually updated and improved based on machine learning
  The con’s: the next generation of infrastructure is still under development prior to commercialisation of algorithms by MLA and Precision Animal Solutions.
  
  
• A reticular rumen bolus which is ingested by cattle, was used to measure core body temperature. Patterns and changes to core body temperature were associated with BRD cases.
  
  

The pro’s: regular information and potential for early detection; based on algorithms that could be continually updated and improved based on machine learning
The con’s: cost; retrieving the technology from the animal post-slaughter

Physical (cattle crush side) tests

Electronic stethoscope and ultrasound were also used. These both function the same way as used in humans.

• The stethoscope, which transforms breath sound into wave forms and sonograms that can be viewed on a computer, was used to detect and grade issues.

The pro’s: can be performed by feedlot staff after some training
The con’s: inaccurate in this project

• Ultrasound presents an image to identify abnormalities in the lungs.

The pro’s: can be performed by feedlot staff after some training; could be deployed immediately; more cost-effective although still a significant investment
The con’s: can be time consuming (around 10-30 mins per animal); can pose a safety risk to staff working closely with livestock; crush needs to be set up correctly for access so modifications may be required; can only be used for confirmation, rather than detection.

Lab tests

Blood tests were taken to test for metabolomics and haptoglobin. Deep nasal swabs were also taken to look at microbiota. The analysis of these tests is not yet complete due to COVID interruptions but results will soon be added to the dataset. 

The pro's: likely to pick up health issues; can be used to support further research
The con's: wait times between taking samples and receiving test results; can only be used for confirmation, rather than detection.

The research

A rigorous process was used to identify livestock for testing, with pens being checked twice daily by livestock staff and vets. Any animal identified three times as a presumptive case and presenting with a high rectal temperature was tested with all diagnostic tools. Every positive case detected was paired with a control (an animal presenting with no clinical signs nor a high temperature) that was also tested with a full set of the diagnostics. All livestock were fitted with both wearable devices.

The diagnostic accuracy of the different modalities was assessed based on sensitivity (the ability to correctly identify livestock with the disease) and specificity (the ability to correctly identify livestock without it).

The best and worst performers

Of the modalities that have been analysed so far, the ultrasound was the best performer in terms of accuracy, detecting 10 out of 10 of each of the cases and controls.

“For overall accuracy, the ultrasound was the best. For practical purposes, [it] diagnosed all cases correctly. It captured all of the cases and correctly ruled out all of the controls,” said Dr Batterham.

“Next after that was the REDI system, followed by the bolus and then the stethoscope.

“Ballpark figures – we saw 95% accuracy in the ultrasound, 80% in the REDI system, about 75% in the bolus, and closer to 60% in the stethoscope.”

Dr Batterham notes the added advantages of the remote monitoring systems for early detection of subclinical cases, and the opportunities to improve these over time.

“We’ve still got to do the analysis on the subclinical animals from the remote monitors… because what they will have as an advantage over the crush-side options is that you could potentially be using those monitors to identify cattle that are never observed by staff.

“The REDI system, whenever it diagnosed an issue, it did so days in advance of the human observer practically every time... that lends itself to much better treatment outcomes.”

Where to next?

With improvements to the algorithms, Dr Batterham sees a lot of potential in the remote monitors. 

“What you could do to improve both remote monitors over time would be to run the experiment again using the ultrasound to confirm cases.

“Use the remote monitors to make calls on healthy and sick animals, bring them up to the chute, do the validation by ultrasound and lab methods, then continually feed that validation data into the model to improve the technology.

“Not only is this good for the health of the animal, the thing that’s front and centre here is antimicrobial stewardship. If you have a better diagnostic test for animals that truly have BRD, you’re using antibiotics on truly diseased animals,” said Dr Batterham.

“If the remote monitors are detecting animals early, we have a much better chance of response to therapy and recovery using antibiotics.

“It could also diagnose other conditions – for example, digestive conditions like acidosis that could be picked up through eating behaviours and movement – or something like lameness.

“With these tools it’s possible you could also build an animal welfare index, there will be a set of variables that are not just consistent with animal health, but also animal welfare. You could be getting a dual benefit– health status + welfare status.”