In the wake of Saratoga’s race meet, which made headlines for high-profile fatalities in Grade 1 races, racing fans and stakeholders have asked each other — what else can be done to improve equine safety?
As pre-race veterinary checks have increased and medication use has become more restricted, many people have shifted their focus toward breeding. Some have gravitated toward the general progression of the breed toward speed and away from distance and pedigrees, while others have focused on particular sires or sire lines they believe are responsible for greater soundness problems.
We asked Dr. Tim Parkin, head of school and professor of veterinary epidemiology at Bristol Veterinary School, about this question. Parkin is the primary epidemiologist who analyzes data from the Equine Injury Database and together with a research team has identified patterns and risk factors that increase the chances a given horse will suffer a fatal injury during a race.
Parkin’s work has shown – predictably – that fatalities are complex events with multiple factors that influence the likelihood one will occur. One factor may explain just a portion of the difference in risk between a fatally-injured horse and a living competitor from the same race, which makes his job difficult. In 2017, Parkin said his team could explain just 35 percent of the ebb and flow of fatality rates between the start of the EID’s record-keeping in 2009 and 2017, at which point the overall fatality rate was trending downward (as it is now).
As it turns out, Parkin believes that the identity of sires and dams in a horse’s pedigree probably don’t make as much of a difference in its risk level as what we may expect.
“Every time we run the multivariable analyses we put sire and mare in as what are known as random which indicates how much influence breeding might have on risk,” he said. “They always come out as having very little influence when accounting for all other variables.”
Parkin cited two peer-reviewed studies that were conducted by a former student of his on injury data from Hong Kong and Great Britain.
In the study of Hong Kong data, which was published in 2013 in The Veterinary Journal, researchers looked backward at 15 years of health data and built statistical models to tease out how much heritability influenced a horse’s likelihood of injury. The data was not limited to horses that had suffered fatal breakdowns, but rather horses who had a musculoskeletal issue identified on an examination from an official veterinarian, or horses for whom a musculoskeletal injury was the reason for their retirement from racing.
[Story Continues Below]
Researchers looked at the sire, dam, and sires in the second generation of the pedigree for each horse. Where additional pedigree information was available, they included as many as eight generations’ worth of sire information. They also factored in other environmental risk factors and basic identity data like gender.
All horses in Hong Kong are imported from elsewhere, as there are no breeding farms there. Interestingly, the researchers found that horses coming from North America had a reduced chance of a suspensory ligament injury as compared to European imports.
“The heritability estimates for fracture, osteoarthritis, and suspensory ligament injury reported here are small to moderate in size, with large standard errors,” the study’s conclusion read.
The heritability of fracture likelihood ranged from .03 to .11 on a scale of 0 to 1, depending on which statistical model the researchers used. The heritability for tendon injury was more significant, ranging between .09 and .2.
There were also genetic correlations between horses’ risk of fracture, osteoarthritis, and suspensory ligament injury, suggesting that those who were prone to one type of injury seemed to be more prone to other types of injury, too.
Support our journalism
If you appreciate our work, you can support us by subscribing to our Patreon stream. Learn more.
“Significant positive genetic correlations between fracture, osteoarthritis and suspensory ligament injury suggest that efforts could be focused upon whichever of these conditions is most easily phenotyped, and progress in the reduction of genetic risk of the correlated condition might occur concurrently,” the authors wrote.
The British study looked at horses who suffered lower limb fractures and injuries to their superficial digital flexor tendons, and examined pedigree data to find out how heritable those conditions may be in their population. Researchers concluded the heritability for lower limb fractures in the population of British horses was .21 to .37 on a scale of 0 to 1.
“SDFT injury and distal limb fracture were positively genetically correlated,” the researchers wrote in their abstract. “These findings suggest that reductions in the risk of the conditions studied could be attempted using targeted breeding strategies.”
On its face that would appear to be a good thing for those who believe the gene pool is in need of an overhaul. A heritability in that range would seem to suggest that fractures could be reduced by 30 percent if the genetic component could be eliminated.
Unfortunately, the reality isn’t that simple, says Parkin.
Parkin says he hasn’t run the same heritability analysis on North American data that the researchers did for Hong Kong and Great Britain, but his models take into account a greater array of additional risk factors besides genetics than what the heritability analyses did. His initial results from just including first-generation pedigree information in his model, however, doesn’t give him confidence.
Then there’s this business of using heritability analysis to reduce risk. Knowing a trait has heritability between .21 and .37 doesn’t mean it can be erased from the population, because the breed is still a breed with a limited gene pool.
“If (big IF) we were able to eliminate any breeding influence of fractures then theoretically a 30 percent reduction would be possible – but this is not realistic,” said Parkin. “New horses have to come from somewhere and I would guess that all sire/mare combinations contribute something to the risk – it is just that some will contribute more and others less. How much more and how much less is not possible to say from the evidence we have.”
Even if you set out to reduce the genetic risk in the whole population, Parkin said, it would be difficult to use data from individuals who share so much common ancestry to tease out specific pointers on how to practically accomplish it. Observers may wonder whether Parkin’s models could someday flag sires whose offspring carry higher risk for injury as compared to others.
“We may be able to produce an estimate of that higher chance – but it will be an average over multiple progeny and there would also be significant uncertainty around that estimate,” Parkin said. “So each sire may have a point estimate of ‘genetic risk’ but there would be significant overlap with very many other sires also.”
For now, it seems, breeders will need to make decisions about how best to improve soundness based on the horses and pedigrees in front of them.
Autor Natalie Voss