IFHA Global Summit on Equine Safety and Technology: Reducing the risk of Exercise Associated Sudden Death

In June 2024, an international multi-disciplinary group of researchers and clinicians with an interest in Exercise Associated Sudden Death (EASD) gathered at Woodbine Racecourse, Toronto. The aim was to discuss current evidence in the field, identify knowledge gaps, and suggest potential pathways to solve those gaps. A critical goal was to discuss how new and evolving knowledge and technology can be harnessed to provide tangible and practical improvements to equine safety and welfare. The workshop was included in the International Federation of Horseracing Authorities (IFHA) Global Summit on Equine Safety & Technology, an event sponsored by The Hong Kong Jockey Club Equine Welfare Research Foundation, Cornell University's Harry M Zweig Memorial Fund for Equine Research and Woodbine Entertainment Group. The discussions aimed to form a multi-disciplinary group of experts that could act as advisors and critical friends to racing, providing a unique opportunity to forge collaborations and open conversations that will lead to practical, actionable items for future implementation.

In the face of changing societal perspectives on the use of animals in sport, racing's social licence to operate is truly under threat. Consequently, the industry and governing bodies need to demonstrate to the public that equine welfare is being taken seriously and forge a path to ensure continued acceptance of racing. The research community must gather the data needed to support the ongoing development and adoption of evidence-based strategies to reduce fatality rates. This editorial serves to highlight the key areas of discussion along with outcomes to be actioned by the group, most notably the need to determine what is ‘normal’ in terms of cardiopulmonary physiology and the identification of risk factors for EASD.

A workable and standardised definition of EASD is vital to draw comparisons between jurisdictions and ensure data collection and research has a consistent and relevant direction. For the purposes of global audit and the discussions that follow, the group recommends EASD be defined as: ‘a fatal collapse in a closely monitored and previously presumed clinically healthy horse that occurs during exercise or within approximately one hour after exercise’.^1-3 Although lesions sustained during racing or training may lead to death at any time after exercise, this definition focuses on acute, idiopathic cases, which are also most societally impactful.

When looking globally across racing jurisdictions, EASD's are low probability events, accounting for 10%–25% of racecourse deaths,³ with an incidence of 1 to 3 per 10 000 starts.^4-9 Compared with the incidence of sudden athletic death in human athletes, the problem in racehorses appears to be much greater.^{1, 10, 11} EASD cases are high impact events that warrant dedicated attention to reduce incidence due to the increased risk of fatality in horses compared to humans, the fact that cases often occur in the public eye, and because EASD massively increases the risk of jockey injury.¹² This workshop focused on cardiopulmonary issues as the most commonly identified causes, but future consideration should also be given to vascular rupture, brain and spinal cord lesions and upper respiratory tract failure; all of which can lead to sudden death during exercise.^{3, 13, 14}

At present, there is limited understanding of what can be considered ‘normal’ in terms of equine cardiac physiology, arrhythmogenic potential, the myocardial response to exercise, and equine ion channel expression (‘channelomics’). To further the understanding of EASD, many study designs rely on the comparison of cases with healthy controls. Accurate phenotyping of EASD case horses is essential, and there is a risk that supposed healthy controls are, in fact, ‘cases in waiting’. A key issue is the need to differentiate normal biological variation from that predisposing to collapse and from potentially pathological changes that could act as early warning signs; something that will only be achieved through extensive data collection and analysis.

A number of epidemiological studies have identified certain groups of horses as being at higher risk of EASD and specific risk factors have varied between studies. Various case definitions have been applied limiting comparisons between studies and generally the outcome measure of interest is fatalities of all causes and epidemiological studies looking at specific pathological conditions which can lead to EASD are lacking. The horse, race distance, type and going, season, purse, field size and a horse's recent exercise and lay-up history, have all been identified as risk factors for fatality in racehorses.^{4, 15, 16} In Ontario's Thoroughbred population, deaths occurring within 60 days of a race or trial entry were examined suggesting rapid accumulation of workload in animals early in their preparation was likely to be damaging and the fatality rate fell toward the end of a season and for horses with a long career history of successful performance.¹⁵ Recent work from Australia which included cases from training as well as racing, has shown that sudden cardiac death was more prevalent in training and affected individuals early in their career.¹⁷ Increased age has been shown to increase risk of fatality in some^{4, 15, 16} but not all¹⁷ epidemiological studies. Use of furosemide was identified as a risk factor for sudden death in one study.¹⁶ However, the definition of sudden death in this study included some horses' death as long as 3 days after racing and it is worth noting that the magnitude of increased risk was small. Although the exact cause of death was not established, the results led to speculation that this may relate to electrolyte imbalance as a trigger for fatal arrhythmia. It is also possible that the observation reflects exercise associated pulmonary haemorrhage (EIPH) as a comorbidity in these cases. Although these studies provide useful guidance for future research and understanding of pathophysiology, the small magnitude of the increased risk for any of these categories and the inability to establish the causative basis of the association between the risk factors and outcome limit the current practical application of this information in prevention programmes. Likewise, studies on heritability and genomics of equine EASD are currently sparse and there is a lack of clear evidence for inherited conditions contributing to EASD.^{2, 18-20} The heritability of pedigree based sudden death appears to be low but individual stallions may be more likely to produce affected progeny.² In the same study, one dam produced two affected progeny of 10 foals suggesting that inherited factors may be involved in some mechanisms of EASD.² To-date, there have been no heritability studies performed using genetic variant based approaches.

Fatal cardiac arrhythmia is believed to be an important factor in EASD cases. However, since large numbers of apparently healthy horses exhibit non-fatal arrhythmia during exercise including official racing,^21-24 this belief remains unsubstantiated. Following death, there is currently no way to determine whether an arrhythmic event preceded it.

A trio of factors are required for formation of a clinically significant and sustained arrhythmia: (i) a substrate within the myocardium, possibly with a structural change such as fibrosis, inflammation or hypertrophy, and with altered electrophysiological properties affecting action potential duration, refractory period or conduction velocity, (ii) a trigger(s), an ectopic rhythm, due to automaticity, triggered activity or reentry and (iii) modulator(s) such as stretch, hypoxia, alterations in acid–base or electrolyte balance, or autonomic tone. A horse could have a substrate for its entire life and never develop an arrhythmia until it is exposed to a trigger. A fatal arrhythmia and EASD could thus result from a complex combination of factors at a single point in time.

There is some evidence in horses suggesting both structural^25-27 and electrical²⁸ cardiac remodelling occurs in response to training. Racehorses in training have been reported to display exercise-induced cardiac hypertrophy²⁶ and may also show increased myocardial fibrosis.²⁷ In human endurance athletes, exercise induces cardiac structural remodelling, particularly of the right ventricle.²⁹ Whilst patchy fibrosis affecting the interventricular septum and right ventricular insertion point are seen in these athletes,³⁰ even in humans, the role of exercise in inducing such lesions, and association with arrhythmia development remains contentious.^31-33 If these changes can be proven to be a substrate for arrhythmia development in horses, there may be possible advice for racehorse trainers on how to modify exercise regimes to minimise or reverse changes that might predispose an individual horse to adverse cardiac events.

In terms of electrical remodelling and resulting changes to the ECG, training tends to lower resting heart rate, lengthen the P wave, QRS duration and PR interval, and increase the likelihood of second-degree atrioventricular block.^{28, 34} Further understanding of these potential changes, and in particular of the level of training required to induce them, is required before the information can be used to reduce the risk of arrhythmogenesis in racehorses.

Ion channel dysfunction (‘channelopathies’) is a major focus in the investigation for causes of sudden cardiac death in the absence of structural heart disease in humans, particularly in young athletes.^35-37 One third of young athletes with sudden cardiac death in the absence of structural heart disease have variants determined to be pathogenic in ion channel genes, and many of the remaining two thirds have variants of as yet uncertain significance in these genes.^{38, 39} Long Q-T syndrome and catecholaminergic polymorphic ventricular tachycardia are the most intensively studied channelopathies in this context, but variants in cardiomyopathy-associated genes are also increasingly recognised.⁴⁰ In horses, channelomic research is still in its infancy. The transcriptional profiles of the most important cardiac sodium, calcium and potassium channels are similar to those in humans^{41, 42} and the currents generated by the most common LQTS genes in humans, KCNQ1 and KCNH2, are functional in horses and largely comparable to their human counterparts.⁴³ Horses therefore appear to have the molecular basis for LQTS.⁴⁴ Normal ranges for Q-T intervals have been defined in a range of breeds,^{45, 46} and whilst acquired LQTS-like rhythms have been observed in horses⁴⁷ inherited equine LQTS has not yet been documented. Early work on equine cardiac ion channel expression predict that it may differ from humans,⁴⁸ but full characterisation of the four chamber channelome remains an important area for future research.

Molecular autopsy has been revolutionary in human cardiology in both clinical and research arenas. For this DNA is extracted from samples collected at autopsy and used for investigation of the genetic basis of disease; allowing proactive clinical and family screening to identify family members at risk, and to facilitate intervention prior to clinical manifestation.^{40, 49, 50} If links between specific ion channels or genetic variants and sudden death are identified, then in the future, a molecular autopsy could be achieved for racehorses with EASD. Given the rarity of these events, coordinated and standardised sampling of EASD cases across jurisdictions is essential to accumulate knowledge.

EASD cases are a heterogenous group of animals with many different pathologies leading to a single end point of sudden death. Historically a definitive or presumptive cause of death is not established at post mortem examination in approximately 50% of EASD cases.³ When this is combined with the subjectivity of, and complicated logistics surrounding, necropsy examination, cardiac failure is often only presumed through the ruling out of other causes, and the lack of specific changes identified. With the need to more accurately phenotype EASD cases to further future research, this is an area where direct reformative action needs to be taken. This must start with agreement on not only a definition of EASD by its presentation, but also on what is NOT an EASD case and on the appropriate criteria, at this stage of our knowledge, for modifying our definitions. These decisions must be made on an objective, scientific, evidence-guided basis.

EASDiR group discussion identified the need for a more standardised cardiac pathology necropsy protocol that could be implemented across jurisdictions. This must state what is not present as well as what is. The post-mortem examination protocol used in California provides a foundation¹³ and refinements might incorporate steps such as: implementation of quantitative measures including heart and lung weight in reporting, use of semi-quantitative scoring systems such as percentage areas affected, quantification of fibrosis and affected regions of the heart (i.e., atrial, ventricular or nodal).²⁷ Additionally, description of pulmonary haemorrhage and oedema⁵¹ as well as a standardised sampling protocols are required to ensure more consistent reporting across different pathologists and jurisdictions.

Pulmonary haemorrhage is frequently identified at necropsy examination in horses succumbing to sudden cardiac death⁵¹ but it remains unclear whether this haemorrhage is the primary cause of death, or a consequence that has occurred secondary to cardiac arrhythmia. Horses with exercise-associated fatal pulmonary haemorrhage are phenotypically different from horses in which EIPH is found as an incidental observation in horses dying of other causes: there is less vascular remodelling and other long-term pulmonary abnormalities than are typically seen with EIPH.⁵¹ Atrial fibrillation is a risk factor for pulmonary haemorrhage,^{52, 53} and clinically is seen in association with ventricular rhythms supporting a potential mechanistic relationship between pulmonary haemorrhage and arrhythmia. Nevertheless, it is possible that the pathway starts with haemorrhage compromising respiratory function sufficiently to induce fatal arrhythmia or that there is a direct link from haemorrhage to fatal arrythmia. Indeed, arrhythmia is commonly seen in mares with severe intra-abdominal haemorrhage⁵⁴ and in horses with acute haemorrhage.⁵⁵

Accurate case phenotyping and clinical characterisation and comparison to healthy controls and horses that succumb due to causes other than EASD are required to further our understanding of the pathology behind EASD. Storage of cardiac and other tissue samples from EASD cases for retrospective analysis, combined with accurate and protocol-driven post-mortem examination reporting and availability of the medical records and clinical histories of affected horses, would all pave the way to greater pathological understanding, and possibly the ability to predict and prevent future EASD cases. Formation of a network of interlinked regional biobanks, using a standardised sampling and storage protocol and with samples stored at multiple locations, could facilitate resource sharing for international research efforts. Although such a network was of unanimously agreed importance, biobank formation is not without significant intellectual property, confidentiality, international movement and cost considerations, which would need to be resolved prior to its creation.

Currently, arrhythmia diagnosis relies on individual assessment of ECGs at rest and/or exercise. ECGs are technically difficult to perform at exercise due to motion artefact degrading the quality of the trace. With manual interrogation the only presently available analysis method, interpretation is time consuming, subject to inter- and intra-observer variability and difficult to achieve on a large scale. Additionally, there are no universal guidelines on electrode placement nor consensus on the interpretation of results; there are also important questions to answer on how many premature beats, what timing or which type of complex arrhythmias indicate a negative cardiovascular and performance impact, versus normal biological variation. Also, with individual horses also inconsistently displaying arrhythmias when examined at multiple time points, examination on a single occasion may have limited clinical relevance.

Work is in progress on artificial intelligence (AI) to allow automation of ECG interpretation.⁵⁶ The use of complexity and restitution analysis has the potential to provide insight regarding an individual's electrophysiological substrate and arrhythmia risk and has been successfully applied for prediction of atrial fibrillation.^57-59 Both AI approaches could be applied to larger scale analysis of equine ECGs. Again, a lack of knowledge of normal variations hampers our ability to differentiate healthy individuals from those at increased risk of EASD. No study exists in which electrocardiograms at rest or during exercise are described from horses that later developed EASD. Therefore, the type of abnormalities that are associated with an increased risk of EASD can only be intuited from electrophysiological first principles. What became clear from the workshop discussion, was the need to start collecting ECG data from large numbers of horses over significant periods of time in a consistent manner; only then will we gain enough information to refine the delineation between normal and pathological. Implantable loop recorders have been used successfully in horses for diagnostic purposes,^60-65 however, there are limitations to the use of these devices in horses in active training and racing. Wearable devices can provide basic single-lead ECG data^66-68 and are an immediately actionable area that could help to facilitate data collection. It is plausible that large datasets of one-lead exercising electrocardiograms using wearables combined with longer recordings and the recent and upcoming developments in 12-lead and novel multi-lead analysis can be combined to gain more knowledge of arrhythmogenic risk.^{69, 70}

Importantly, it is critical to recognise that given the relatively high prevalence of rhythm disturbances in equine athletes,^{21-24, 71} it is unlikely that identification of pre-existing rhythm disturbances alone will prove to be a robust predictor of EASD. Indeed, studies examining poorly performing racehorses in which arrhythmia had been detected indicated that future racing outcomes were not impacted⁷² and ventricular rhythms have been documented in race conditions in apparently healthy Standardbred racehorses²¹ and Thoroughbreds in Chuckwagon races.^{24, 71}

There are certain barriers that must be navigated before such widespread application of ECG data collection via wearables is achievable. First, the devices to be used must be validated to recognised standards if we are to make regulatory decisions based on the data generated by them. There are issues relating to confidentiality, data accessibility, and repercussions when clinically silent ECG abnormalities are detected. The same issues arise around the use of alternative biomarkers such as genetic tools and indices of electrophysiological substrates. There is a notable ethical dilemma with regard to research data collection from horses in race-training. If a putative abnormality is detected during a study should this be notified to the horse's owner and trainer and to regulators? Should this horse be allowed to race again? What are the potential consequences if such a horse continues to race and goes on to suffer an EASD event? Conversely, if researchers intervene and the horse is taken out of training based on hypotheses rather than robust and validated evidence, it will never be possible to determine the prognostic value of potential biomarkers. Where the biomarker in question is a molecule detectable in tissues, such as a genetic variant, such research can be performed retrospectively avoiding this ethical minefield. Where, by its nature, the biomarker has to be derived pre-mortem as is the case with ECG data or plasma cardiac troponin concentrations, only longitudinal designs are applicable.

Fatal racetrack arrhythmia/collapse events are difficult to manage due to a combination of their distressing nature, the variable exposure of attending racecourse veterinarians to this eventuality, and the limited equipment and treatment options available. Confirming and characterising any arrhythmia present can be challenging. However, these events actually represent a golden opportunity to obtain valuable information that could be used to prevent future cases, and there are two elements that need to be considered. As a first priority, it needs to be determined if the horse can be saved and, if so, which is the best and safest way in which to help it. Currently, and especially without a diagnosis and limited ability to achieve one sufficiently promptly, we do not know the best therapeutic approach for these cases, but the group identified the value of having a protocol to follow in such a pressurised situation. Implicit procedural memory allows for easier recall and it was indicated that formation of a CPR protocol for the collapsed horse was an achievable outcome which could help to improve and standardise care across racing jurisdictions. When fatalities occur, consideration needs to be given to the potential to collect samples and carry out a post mortem.

Where horses stabilise and recover after an on-track arrhythmic episode, the majority are likely to be cases of atrial fibrillation^{53, 73} although post-race ventricular tachycardia has also been documented.²³ Atrial fibrillation is a recurrent condition but duration between detected episodes is highly variable, usually measured in months to years and horses can race successfully in the interim.⁷³ Regulators' policies for assessing horses which have been identified as having repeated racecourse arrhythmic events are inconsistent: broadly, the UK adopts a case-by-case approach where each individual horse is evaluated and a specific plan made for that animal, whilst Hong Kong and Australia operate an incidence-based approach where a case triggers a stand-down for a fixed duration, which in Australia includes standardised protocols for diagnostic investigations which must be completed before the horse returns to the racetrack. Further information is needed to elucidate which is the best approach but, again, collaboration to ensure minimum levels and standards of diagnostic data are collected across jurisdictions would serve to both generate usable data for research and improve racehorse welfare and care.

It is important that this work is carried out within a background of trust, transparency and free sharing of standardised data between racing jurisdictions, researchers and regulators; something, with their extensive connections, the IFHA are in a unique position to broker. Of particular importance is encouraging all racing jurisdictions to carry out uniform fatality reporting, using the outlined EASD definition, combined with individual review and phenotyping of all fatalities in training and racing, using the SOP's formulated by the EASDiR group. Only then can true fatality rates be calculated; an important baseline against which future improvements can be benchmarked. The standardisation of reporting should additionally allow for global auditing to be carried out.

Although there are currently huge and important knowledge gaps concerning the causes of EASD, their pathogenesis and prediction, the progress that has already been made in the characterisation and understanding of EASD should not be underestimated. However, if we are to truly make tangible improvements to equine welfare in racing, the research groups working in this area need sustainability in terms of personnel, financial support, easy access to data and collaborative work options. This workshop has forged collaborations and sparked conversations that must be maintained. It is intended that the EASDiR group immediately starts working on the identified actionable items, but has fluid membership and regular meetings to ensure progress and accountability.

Victoria Anne Colgate: Conceptualization; writing – original draft. EASDir Working group: Conceptualization; writing – review and editing.

The International Federation of Horseracing Authorities (IFHA) Global Summit on Equine Safety & Technology was sponsored by The Hong Kong Jockey Club Equine Welfare Research Foundation, Cornell University's Harry M Zweig Memorial Fund for Equine Research and Woodbine Entertainment Group.