The science of soundness in Thoroughbreds
For as long as I can remember – and admittedly it is not as long as some – there has been a continual question over the soundness of the Thoroughbred racehorse. While, in recent history, the death of Eight Belles at the Kentucky Derby caused the North American industry to, at least for a moment, to confront the issue, it is one that in reality has been present, if not always front and center, for a far longer period of time.
In recent years there has been a significant push from both scientific circles and racing regulators to put the question of soundness of the thoroughbred into more scientific terms with original research. Rather than letting the lunatic fringe blame the presence of "too much Raise a Native" or "too much Phalaris", theories which are far from being based on reality, the move has been towards well constructed scientific studies of a complex issue.
At the Jockey Club Round Table Conference this year, Dr. Hiram C. Polk Jr. a member of their Thoroughbred Safety Committee presented preliminary data from a study that the Jockey Club have undertaken generated as an attempt to answer the question previously posed by Jockey Club Chairman Mr. Stuart Janney, that is, "(if) the prolonged use of furosemide (Lasix) potentially weakens the breed."
The data presented and its interpretation raised some questions in my mind, several of which I suspect that Dr. Polk himself might also share. In Dr. Polk’s defense, the Jockey Club Round Table is hardly a proper venue for presenting this type of data in the form in which it was offered. Dr. Polk also made it clear that the data he presented was a "work in progress" and that a paper to be published later would better qualify their findings. Still, it is worth reviewing the transcript and watching the video to see how the study was constructed, and how it sought to answer the question posed by by Mr. Janney.
The study was based on runners in Australia, and embraced 14,502 horses that had made a total of 230,541 starts. Australia racehorses were selected for the study because there is no race day medication allowed in that country; there is a large pool of racehorses; shuttle stallions from the Northern Hemisphere have been quite active in recent years; and there has also been significant importation of fillies and mares from the Northern Hemisphere. These horses were then divided into three groups based on their pedigree:-
Group 1 - 239 horses that had 100% of their pedigree (defined as the ancestors within 3 generations) as having raced exclusively in North America
Group 2 - 469 horses that had at least 50% of their pedigree as having raced exclusively in North America
Group 3 - 13,794 horses that had 0% of their pedigree as having raced exclusively in North America.
Based on these groups and an analysis of their race careers, the preliminary conclusions were that Group 1 - those horse with 100% North American pedigrees - raced more frequently and had a greater number of starts when compared to the other groups. At face value (and I would again reiterate again, that I am guarded in my criticism given the venue and timeframe that Dr. Polk had to work within), it is compelling and it will be interesting to see what finally comes out of the study when papers are published later in the year. However, I do have some concerns.
Firstly, the a priori assumption of the study is with regard to horse with 100% North American pedigrees, and the assumption that all 14 ancestors raced on Lasix and/or in some way bled or was predisposed to bleeding. That is not necessarily true. New York was the last state to allow Lasix, prompted, ironically by Summer Squall ( who at the time was owned by Cot Campbell and subsequently retired to one-time Jockey Club Chairman Will Farish’s Lane’s End Farm), missing the Belmont Stakes, as he was a bleeder that required Lasix to compete (read a great story here by Andy Beyer back in 1990 – there was a study even back then that said Lasix was performance enhancing….eeesshh).
The generational interval in thoroughbreds is roughly 10 years (its shorter on the sire side than it is on the dam side but it averages out at 10 years) and this study looked at all foals born in Australia since 1995 (presumably through to 2005 as they would need these horses to have complete racing careers to make assessments). This period is not nearly enough to suggest that all ancestors in the pedigrees of these horses raced on Lasix to start with. Parents, yes, grandparents, maybe, great grandparents, no. The grandparents of horses born in the period of 1995 to 2005 would have probably been born and racing around the late 1960's, 1970's and early 1980's. Some of these ancestors would have raced on Lasix for all of their career, some for some of their career but a lot wouldn't have ever seen the drug during their race career.
Secondly, the study doesn't really attempt tackle the heritability aspect of soundness. All that is being said, is that for the 239 horses with 100% North American bloodlines (a number I would argue isn't statistically significant - you need at least 5,000-8,000 horses to truly represent a population in statistical terms) is that they had more total starts, and started more frequently, than the other populations. What in effect is being said, is that the North American-bred thoroughbred can perform without Lasix. I'm not sure I 100% agree with that either, but we will get to that a little later.
The data presented indicated that the horses with 100% North American pedigrees were making, on average, 17.5 starts in their lifetime, which compares favorably to the group with 50% North American-raced ancestors (16.8 lifetime starts), and the group whose pedigrees were devoid of North American-raced ancestors (15.8 lifetime starts). That, however, is a little bit of an apples to oranges comparison, and it also doesn’t really answer the question posed by Mr Janney.
What needs to be established is whether the immediate ancestors (sires, dams, grandparents) of the 100% group had, on average, more starts than their offspring. If the 100% North American group had averaged 17.5 starts in their lifetime, it would be telling a different story than the one taken up by the media, if we find that the ancestors for this group of horses averaged, say 25 lifetime starts. In this case, it would suggest us that by taking this group the medicated environment in which their ancestors performed, they have been less sound. To make it fair, and more descriptive, you'd need to do the same for the other groups. What if the ancestors of the 0% group (all Australian) had an average lifetime starts of 11 and their offspring are now 15.8? That may show that horses whose parents have not been exposed to widespread race day medication can actually be bred to improve average starts in a career, something that would also satisfy the Jockey Club’s narrative, particularly in an era of declining foal numbers. Of course to do that, you’d have to acknowledge that horses in North America have lower lifetime starts as a collective group racing on dirt when compared to horses racing in Australia on turf, and to properly account for that in the study, which is probably why the Jockey Club study never got into the true heritability of soundness in the first place (at least based on the data that they did present).
The second aspect, I would question is with regard to control of variables. Again, in his defense Dr. Polk did mention this as an issue, but it is apparent that the 100% group of just 239 horses would have, by way of commercial considerations alone, received greater opportunity to achieve more starts, higher earnings and the like when compared to the other groups. If a mare is imported from North America she is going to be a well bred and/or performed mare because it costs you $20,000 to get her there in the first place. The barrier for entry is reasonably high for the mare to start with which probably puts them at an unnatural advantage.
Presume then that she is mated to More Than Ready as an example, so the resultant foal is 100% North American bred. It is a pretty good bet that the foal is going to go to a better trainer, be trained at a better track, receive better veterinary care, and so on, than the average member of the Australian Thoroughbred population which they are comparing them against. Failure to correct for these variables on a population basis looks to potentially be a major shortfall for this study. Equally, in a study by Brandon Velie and colleagues at the University of Sydney on 117,088 horses (now that is a genuine population study) racing in Australia, they found that geldings had significantly longer careers than females and intact males, and females had significantly longer careers than intact males. With such a small group of 239 horses in the 100% North American group, you'd only need a slightly higher proportion of them than normal to be geldings, say 10 extra horses, to make the numbers generated by this study pretty skewed.
The findings that were presented at the Round Table are also at considerable divergence to an earlier population study, again by Brandon Velie and his colleagues at the University of Sydney. That study looked at the soundness and career starts for horses that raced in Hong Kong, a highly controlled environment with the best drug testing on the planet. In that study of 4,950 horses, their key finding was that there was a significant influence of region of origin on the hazard rate for Hong Kong racing retirement. They found that horses originating from Europe had a significantly lower hazard rate for Hong Kong racing retirement. Conversely, they found that horses imported from North America had a greater hazard rate for racing retirement in Hong Kong than those from the other racing regions. The region with the second highest hazard rate for retirement was South America, a region that up to a few years ago was allowing race day administration of Lasix. Given this peer-reviewed study basically found that North American bred horses are less sound in Hong Kong, as pure a racing jurisdiction that can be found in the world, it will be interesting to see how the Jockey Club's study reconciles with this (less publicized) finding in their upcoming paper.
Finally in regards to bleeding/EIPH, which, returning to Mr Janney’s original proposition, is the reason cited by horseman for the use of Lasix: the statistical data on this is unequivocal – bleeding is highly heritable. The original South African study on 63,146 racehorses showed that heritability estimates for bleeding in an animal and sire models were 0.23 and 0.40, respectively, which indicated that epistaxis has a strong genetic basis in South African racehorses. Equally a study on the Australian racing population by Hamilton and colleagues, in 2012 found that using a sire model, heritability for incurring one episode of epistaxis associated with EIPH was estimated at 0.51 suggesting there is a genetic basis to the disease in Australian Thoroughbreds. Any trait with a heritability over 0.20 can be selected for (or against), so we are talking about a deleterious disease that is highly heritable.
Interestingly the study by Hamilton and colleagues found that the heritability for sires ranged from –1.12 to 2.10, indicating a 25–fold difference in the odds of the disease. Some sires were negatively associated in siring bleeders (they didn’t get many), while other sires were strongly positively associated (they got a lot). Hamilton, et al didn’t name names in their study which is in contrast to the South African study who flat out named Al Mufti, an American bred stallion by Roberto out of Lassie Dear, as a major source of bleeding, despite being a leading sire (there in itself is a conflict for breeders to consider). Now I’m no genius but I do know two things – Summer Squall, the horse that needed to race on Lasix to win a Preakness, is out of a daughter of Lassie Dear and Roberto is by Hail to Reason who is the grandsire of another stallion named Southern Halo, who was a bad bleeder when in training with Wayne Lukas and whose successful introduction to the South American breeding population as a leading sire also coincided with (caused???) the widespread use of Lasix in Argentina, a position they have since reversed, having seen its impact on their breed and export opportunities.
Knowing all of this, I fall short of completely agreeing with what the underlying narrative of the Jockey Club study seems to be - North American bred horses are not genetically tainted and can race without Lasix. The more complete and well constructed study by Velie, et al in Hong Kong points to underlying differences in career lengths with North American horses coming off poorly, with a higher hazard rate for their careers in Hong Kong. That study of course relates to all soundness, so the source of unsoundness, or less race starts, may not be related specifically or individually to EIPH. Additionally, if the day comes that the North American thoroughbred is taken off Lasix, there are clearly going to be stallions that are exposed for being sires of bleeders. There is no escaping that. That is going to create a whole new set of issues for breeders as while it is heritable (and thus can be selected against) it is possible, as was the case with Al Mufti and Southern Halo, breeders will elect to use stallions that sire a high proportion of bleeders because of their genetic prowess, even in the knowledge that bleeding comes with the territory.