Performance horses as a model for exercise-associated cardiac arrhythmias and sudden cardiac death

Abstract

This paper reviews the myocardial substrate of horses relative to that of humans and discusses the utility of performance horses as a model of exercise-associated cardiac arrhythmias and sudden cardiac death in athletes. The coronary circulation is similar between the species while coronary artery anomalies and myocardial bridging appear to only be associated with athletic mortality in human athletes and not in performance horses. There are subtle differences in the histology of the sinus and atrioventricular nodes, of unknown clinical significance, while the His bundle is more highly innervated in horses. The equine Purkinje network is much more extensive, contributing to a difference in the mean electrical axis between horses and humans. Differences in ion channel expression have been reported, although they are poorly characterized, and are of unknown clinical significance. However, horses may be a particularly good model to investigate the function of Kv1.5 due to its spontaneous ventricular expression, which is lacking in human ventricles. Similarities in cardiac structure, coronary vasculature, and ability to exercise at high levels makes performance horses a good model to investigate exercise-associated cardiac arrhythmias and sudden cardiac death in athletes. However, differences in myocardial substrate should be taken into consideration when designing studies and interpreting results.

Graphical abstract

1. Introduction

Sudden cardiac death (SCD) is a leading cause of death in human athletes and is 2.8 times more likely to occur in athletes than non-athletes [1]. SCD is often attributed to an unrecognized underlying predisposition to life-threatening arrhythmias which is unmasked by athletic activity [1]. Some of the underlying cardiac abnormalities associated with SCD in human athletes are shown in Fig. 1. Despite this established link between human athletes, cardiac arrhythmias, and SCD, the relationship remains poorly understood, making identification of athletes at increased risk problematic.

Fig. 1.

Comparison of the causes of sudden death of human athletes (top row) and equine athletes (bottom row) identified at post-mortem examination. In humans, there is a larger percentage of athletes who have underlying cardiac abnormalities (all shades of red combined), compared with horses where a large percentage are necropsy-negative (grey). The “Cardiac – Miscellaneous” category in Dennis et al. 2018 is notably larger than other studies due to inclusion of adult (>35 years of age) athletes, and therefore a greater percentage of coronary artery disease which is less common in younger populations. “Cardiac – Miscellaneous” in equine studies include mild, focal areas of fibrosis, of unknown clinical significance [1,8,10,11,59,60]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Similar to human athletes, SCD is also a leading cause of death in performance horses. SCD in racehorses is estimated to be over 200 times more likely than in human athletes [2]. However, unlike human athletes where underlying cardiac abnormalities are frequently identified, performance horses rarely have any cardiac abnormalities identified on post-mortem examination, leaving the cause uncertain (Fig. 1). A direct comparison of risk between human and equine athletes is complicated because published equine studies consist of 2 types: epidemiological studies examining factors associated with fatality [[3], [4], [5], [6], [7]]; and those reporting post-mortem findings [[8], [9], [10], [11]]. Although the studies are well conducted and informative, it is not possible to directly compare the risk of SCD between horses and humans without equine studies that include both pieces of information. Nevertheless, SCD in racehorses is well documented [[10], [11], [12]]. Additionally, the burden of proof varies between (and potentially even within) studies where some report the most significant finding as the cause of death, even though it may be difficult to determine what was the immediate cause of death, whereas others report presumptive or unexplained diagnoses [8,9].

Horses are natural athletes, having evolved over many years for the athletic abilities required to avoid predation. Racehorses perform at extremely high exercise intensities with maximal heart rates exceeding those of human athletes [13,14]. Horses, in comparison to humans, have a much greater heart rate range (at least 28–240 bpm); roughly double the maximal aerobic capacity; and double the heart-size to body-weight ratio [15]. There are also important differences in the ability of equine and human athletes to achieve the highest levels of performance. Racehorses are purpose-bred and given every opportunity to show maximal potential, whereas human athletes must first choose to partake, choose the type of athletic activity, and overcome numerous physical, mental, and socioeconomic barriers to athletic activities. Performance horses, therefore, possess and express a cellular, metabolic, and molecular phenotype that supports elite athletic performance. The molecular elements of the heart that permit horses to perform at a high level and that may predispose to necropsy-negative SCD are not well understood. However, some suspect characteristics are known, and ongoing investigations offer opportunities to compare cardiovascular function and disease risk between human and equine athletes.

Unlike many potential animal models, performance horses participate in athletic activities that imitate human activities, ranging from occasional/recreational exercise, to daily/intense conditioning. Like human athletes, horses also frequently undergo periods of detraining due to injuries and therefore provide numerous opportunities to investigate the impact of repetitive exercise on possible cellular and molecular mechanisms of arrhythmias and SCD. The occurrence of fatal cardiac arrhythmias in athletes of both species may allow horses to serve as a model for studying SCD, while simultaneously offering a model to further our understanding of the physiological response to exercise. While SCD is the focus of this review, both horse and human athletes also experience exercise-associated atrial fibrillation, which has been reviewed recently [16,17]. Exercise-associated atrial fibrillation and SCD share many associations based on current research, such as return of parasympathetic activity, and high exercise load [17]. Therefore, research using the horse as a model of either SCD or atrial fibrillation (AF) should consider comparative differences in the underlying myocardial substrate. While there are practical limitations to producing and maintaining horses for research purposes, numerous opportunities exist for knowledge integration and interdisciplinary approaches, benefitting both human and equine athletes.

2. Myocardial blood supply

The organization of the coronary arteries is subject to variation both within and between species [18,19]. These variations may have important implications for oxygen delivery, particularly during maximal intensity exercise. Numerous types of coronary artery anomalies (CAA) have been described in humans and few, including anomalous pulmonary or aortic origin, have been associated with SCD [20]. However, evidence of causation is lacking and attempts at surgical correction of anomalous vessels has not been demonstrated to eliminate the risk of SCD [20]. In contrast, although CAA are reported in horses, a predisposition to SCD has not been ascribed either due to the type of abnormality or the clinical history [10,21]. It is unlikely that CAA as a cause of SCD are being missed in horses owing to the rigorous post-mortem examination protocols employed in the majority of horseracing jurisdictions [9,22]. Therefore, comparative investigations may be fundamental in characterizing the mechanisms of SCD related to CAA and/or determining whether the clinical significance of certain anomalies in human athletes has been overstated.

The myocardial blood supply overall is very similar between horses and humans with both primarily possessing a left and right coronary artery that originate from respective coronary sinuses of the aortic root. Right coronary dominance is predominant in both horses (98.3 %) and humans (85 %) [18,19], and both horses and humans show roughly equal contribution from the left and right coronary arteries for perfusion of the interventricular septum [23].

One additional potentially important difference between humans and horses is the presence of myocardial bridging (MB), which is identified in humans but not horses. MB, defined as portions of coronary arteries congenitally located underneath working myocardium, has sometimes been associated with SCD in humans [24] and has been reported in 7–86 % of human hearts at autopsy [25], but has not been identified in horses [19,26]. The significance of different configurations of MB and the tunneling arteries is debated and often a diagnosis of exclusion [27]. However, it is interesting that despite the apparent absence of this phenomenon in horses, horses are thought to experience higher rates of exercise-associated SCD than human athletes. This absence of MB in horses may provide insight into the significance of this finding in cases of human SCD.

Overall, the fundamental similarities between horses and humans in myocardial circulation support the use of horses as a large animal model for studies into human exercise physiology.

3. Conduction system

Much of the conduction system anatomy including the sinus node, Bachmann's bundle, atrioventricular (AV) node and His bundle is conserved between horses and humans. However, there are some important differences, in particular the Purkinje network.

3.1. Sinus node

The sinus node, located within the crista terminalis at the junction between the cranial/anterior vena cava and the right atrial appendage, is overall quite similar between horses and humans. The sinus node in both species contains specialized cells (P cells) and receives its nutrient supply from a central nodal artery [28,29]. In horses, the sinus node is identified by cellular morphology due to the absence of a fibrous capsule [28]. There is a lack of comparable studies with which to comment on the relative sizes of the sinus node in horses and humans, however, the overall shape is described as oblong in horses, and crescentic in humans [28,29]. Changes in the P wave morphology during sinus rhythm, referred to as wandering sinus node, are frequent in horses and have been previously attributed to the size and shape of the sinus node in horses [30]. However, the absence of a fibrous capsule surrounding the nodal tissue in horses could explain/contribute to this phenomenon due to a lack of electrical isolation of the node from the surrounding atrial muscle. However, studies documenting the frequency of wandering sinus node in equine athletes are lacking. The term wandering sinus node is used here rather than wandering pacemaker to refer to changes in the origin of the electrical activity and to differentiate between this and physical dislodgement of an implanted device. It is also possible the increase in fibrous tissue in humans vs. horses represents a general increase in connective tissue associated with aging if the human patients examined are in adulthood, while the horses used in many studies are not yet fully mature.

The human sinus node has been described as having an abundant nervous input both within and surrounding the nodal tissue [29], whereas horses only have nervous tissue surrounding the node and not within it [28]. This finding is interesting when the frequency of arrhythmias arising from the sinus node is considered. Despite the dominant parasympathetic control of the heart in horses, sinus arrhythmia, sinus bradycardia, sinus blocks/arrest are all uncommon in young performance horses [31], especially when compared to the frequency of atrioventricular block. Humans may be more likely to have electrocardiographic changes at the level of the sinus node with alterations in autonomic activity as a result of the increased nervous supply. Sinus node innervation and dysfunction are not currently considered in the context of exercise-associated SCD and cardiac arrhythmias in athletes.

3.2. Atrioventricular node

The location of the AV node, although less consistent than that of the sinus node, remains similar between horses and humans: within the interatrial septum, between the non-coronary sinus and septal attachment of the tricuspid valve. The area adjacent to the AV node has been found to contain cartilage (described as cartilaginous metaplasia) in 77.8 % of horses but not in humans [32]. This finding is associated with sudden death in human neonates due to presumed arrhythmia [33]. It is unclear whether the underlying mechanism reflects impaired conduction through the AV node, or re-entrant arrhythmia due to the disruption of normal electrical activity. However, the fact it was commonly found in asymptomatic horses questions its significance in the absence of electrocardiographic abnormalities [32]. There is, however, a lack of studies describing the presence or absence of cartilage adjacent to the AV node in horses with known clinical histories, creating uncertainty about its significance. Whether cartilage within the conduction system predisposes to; is the result of; or is incidental to arrhythmia remains to be determined. Since athletes undergo structural and electrical changes in response to repetitive bouts of exercise, whether cartilage metaplasia is an irreversible consequence of cardiac adaptions should be investigated.

There are possible implications of cartilaginous metaplasia in the area of the AV node in the context of species differences in AF. Horses in AF generally have a normal resting heart rate, indicating the equine AV node is less responsive to overstimulation. Conversely, humans are often tachycardic and therefore the AV node may be more responsive. However, cartilage in this location is present in 60 % of dogs [32] yet they are similar to humans in generally being tachycardic with AF, and therefore there may be no significance of this finding.

3.3. His bundle

A potentially important difference between horses and humans in the His bundle is the presence of nerve fibers. Gómez-Torres & Ruíz-Sauri identified many nerve fibers either within or at the periphery of the His bundle in horses, while none were identified in humans [34]. This finding, although reported in a small number of animals (5 each of humans and horses), may indicate there is greater autonomic nervous input to the His bundle in horses [31]. In contrast to what occurs at the level of the sinus node, AV block associated with parasympathetic stimulation is common in horses [31] and this increase in innervation offers a mechanistic explanation. Additionally, an increase in autonomic nervous system input in the His bundle of horses may increase the risk of triggered junctional activity, which is, in fact, commonly reported in performance horses [35]. The frequency of AV block is higher in racehorses than other types of horse [36], consistent with a positive association between increased parasympathetic activity at the level of the AV node/His bundle and physical fitness. Therefore, it is plausible that the junctional premature complexes observed are a result of the alterations in parasympathetic activity in response to exercise. As such, horses may serve as a useful model for specific investigation of the role of conditioning and deconditioning on ectopic activity arising from this location, which could be an important cause of exercise-associated arrhythmias in both horses and humans.

Although species have been categorized as having either more (e.g., horses and pigs) or less (e.g., humans and dogs) collagen fibers present within the His bundle, the percent area of fibrous tissue is similar in humans and horses (36.9 % vs. 43.2 %, respectively), compared to pigs (84.7 %) [34]. Further work is needed to draw any conclusions about the significance of these differences.

3.4. Purkinje network

Purkinje fibers are essential for the rapid conduction of electrical activity throughout the ventricles; however, there is also a known arrhythmic predisposition at the junction of these fast-conducting fibers and cardiomyocytes [37]. Understanding the differences between the Purkinje fiber distribution in horses and humans is therefore critical to assessing the use of horses as a model for human cardiac dysfunction.

Purkinje fibers are most abundant in the sub-endocardium of humans and other non-hooved mammals [28], whereas horses possess both sub-endocardial and intramural Purkinje fibers organized in an extensive network (Fig. 2) [38,39]. Additionally, Purkinje fiber cells in horses are larger and contain more abundant desmosomes and larger gap junctions, thereby being more distinct from contractile cardiomyocytes [40]. In comparison, the Purkinje cells of non-hooved mammals are difficult to distinguish histologically (Fig. 2C). This difference between a sub-endocardial and extensive myocardial network of Purkinje fibers is thought to be responsible for differences in the mean electrical axis [41]. Therefore, comparative electrocardiographic studies should be aware of and acknowledge this important difference.

Fig. 2.

Purkinje cells stained with hematoxylin and eosin within the myocardium of a 7-month-old foal (A, B) and adult cat (C) for comparison. (A) Full thickness slice (4× magnification) of the left ventricular free wall. Arrowheads show the sub-endocardial location of Purkinje cells while black arrows denote the location of an intramural Purkinje fiber. (B) Enlargement of A showing the Purkinje cells located between the black arrows in the sub-endocardium (20× magnification). (C) Purkinje cells located in the sub-endocardium of an adult cat (black arrow, 40× magnification). Key: End – endocardium; Epi – epicardium.

Images courtesy of Dr. Jeff Caswell.

The impact of Purkinje fiber distribution on arrhythmia risk, particularly in athletes, is largely speculative. The theoretical benefit of a more extensive network of fast conducting fibers in equine athletes is to improve synchronicity of electrical activity during exercise, which could protect against ventricular re-entry arrhythmias. However, the fact that horses experience higher rates of SCD even with a more extensive conducting network suggests that triggered activity at the junction of Purkinje cells and cardiomyocytes may be more significant as an arrhythmia risk than any protective effects of improved synchronicity.

4. Electrophysiology

A primary difference between equine and human cardiac electrophysiology is the mean electrical axis (Fig. 3A, B). There are 2 factors to consider: ventricular activation pattern and physical orientation of the heart in the thorax. Horses, along with other hooved mammals, have a ventricular activation pattern that primarily runs from heart apex to base, reflecting the extensive Purkinje network [41] and a heart that is oriented with the long axis almost perpendicular to the long axis of the body. The mean electrical axis of the horse is directed from the sternum dorsally towards the spine. In contrast, humans have a ventricular activation pattern from heart base to apex and orientation with the long axis of the heart on the same plane as the long axis of the body in a right-superior to left-inferior orientation. Therefore, an ECG obtained with a negative electrode towards the heart base and a positive electrode towards the heart apex will be primarily positive in humans and negative in horses (Fig. 3C, D). In addition to differences in the primary QRS vector, the extensive Purkinje network in horses also results in electrical activity that often moves in opposing directions at the same time, resulting in a cancellation of vectors and no contribution to a body surface recording. As such, ventricular hypertrophy and bundle branch blocks are not readily recognized on the body surface ECG of horses [31,41].

Fig. 3.

Diagrammatic representation of the orientation of the heart and genesis of a base-apex ECG in horses and humans. Approximate location of the heart (red outline) in the chest and mean electrical axis (arrow) in humans (A) and horses (B). Additionally, the placement of positive and negative electrodes to capture the main electrical vector of each species is shown with the corresponding ECG for humans (C) and horses (D). Note that despite placement of the electrodes being in the same positions relative to the orientation of the heart, the ECG signals are inverted due to differences in the pattern of ventricular activation. An approximate scale bar shows the relative duration of the average QT interval between horses and horses at approximately resting heart rate. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

The QT interval on the body surface ECG is both a function of the ventricular action potential, and the direction and speed with which an impulse spreads throughout the myocardium. Electrical activity that moves in opposing directions at the same time is equalized and therefore is not visible on a surface-level recording. Despite the increased size of the equine myocardium relative to humans, QT interval duration at rest is quite similar between horses and humans at roughly 450–560 ms and 360–450 ms, respectively, with sex differences described in both species [42,43]. Additionally, analysis of the relationship between electrical systole and diastole, termed restitution, has been conducted in both horses and humans [44,45], albeit not specifically in athletes. Cardiac restitution (also known as ECG restitution) has been described using a ratio of QT/TQ with a ratio of >1 considered pro-arrhythmic and has been used as a predictor of arrhythmia risk in humans at rest [46]. Differences in cardiac restitution have been shown in humans with active myocarditis [47] and other cardiac abnormalities [44]. Cardiac restitution analysis may have the ability to detect underlying cardiac abnormalities which predispose athletes to arrhythmia and SCD. Preliminary work has described cardiac restitution in equine athletes [45], however further investigation is needed to determine whether cardiac restitution is a useful predictor of risk in horses. Work performed to date does not examine what occurs during maximal effort in either species, but investigations using exercising ECGs in horses will provide insight into the feasibility of large-scale investigations in athletes.

5. Cardiomyocyte protein expression

Despite some clear differences in cardiac anatomy, a recent proteomic analysis revealed closer similarity among ‘large mammals’ like pigs and horses than between ‘large mammals’ and ‘small mammals’ like rats and mice [48], which supports the use of horses as a model for human cardiac disease. However, these results were based on biopsy samples of a small number of non-athlete animals. There are many other potential factors, particularly at the cellular and molecular levels, that should be taken into consideration, rather than just animal size and gross anatomy of the cardiovascular system.

5.1. Ion channels

Ion channels are essential to the generation and conduction of electrical activity throughout the heart. The cardiac action potential is the summation of all ion movements across the cell membrane. Therefore the expression level and functionality of specific ion channels determines the characteristics of the cardiac action potential, and muscle activation.

Monophasic action potentials have been characterized in the right ventricle of awake humans and horses [49,50] with differences shown in Fig. 4. Of note, the duration of the ventricular action potential differs due to differences in the resting heart rate between the species (~70 bpm in humans, ~40 bpm in horses). Interestingly, the APD at 10 % of the action potential amplitude (APD90) in horses paced at 100 bpm was still ~25 % longer than humans at rest (322 ms vs. 243 ms) [49,50]. Notable interspecies differences in the shape of the ventricular monophasic action potential include a more prominent plateau phase in horses during which the membrane potential increases back up to the level of the initial depolarization, and a more rapid repolarization following the elongated plateau (Fig. 4). It appears that the features that differentiate the ventricular action potential from that of the atria – prolonged plateau phase at a higher membrane potential and more rapid repolarization, are accentuated in horses compared with humans. The differences in the equine ventricular action potential, therefore, may be to optimize contractility and reflect the superior athleticism of horses compared with humans.

Fig. 4.

Diagrammatic representation of the right ventricular monophasic action potential in humans (A), and horses (B). Ion channels known (humans) or thought to be (horses) responsible for the primary ion currents in each phase of the action potential are listed. Action potential of the horse is based on work by De Clercq et al. [50]. Ion channels in blue font are hypothesized based on mRNA studies and have not been confirmed in horses. Additionally, there are potentially other unidentified channels contributing to the action potential in horses. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Despite publicly available comparative proteomic analysis [48], there is limited information on the differences in ion channel expression profiles between horses and humans that might explain these differences in the shape of the monophasic action potential. Predicted expression based on mixed tissue mRNA models showed key differences in potassium, sodium, and calcium channel isoform expression [51]. In this study, however, the predicted expression patterns were based on stored RNA sequencing data of tissues collected from unknown anatomic locations (atria vs. ventricles); under a variety of conditions; and with low confidence. Interestingly, some of the channels with the highest expected expression (K_ir6.2, for example) [51] were not detected in the proteomic analysis (Table 1) [48]. Additionally, some of the potassium channels which are known to be important in humans (e.g., transient outward K_v4.3) were not detected in either human or horse samples [48]. Details on tissue collection are lacking in some studies [48], and it is not unusual for collection of equine myocardial samples to require a prolonged time, potentially resulting in protein degradation and accounting for the lack of detection of some proteins which are expected to be present.

Table 1.

Selected ion channels and ion handling proteins. Expression levels in horses and humans generated from the average of 3 left ventricular samples from online proteomic analysis (available from: https://atlas.cardiacproteomics.com/Search [35]).

Gene/channel	Phase of action potential contribution	Function	Human expression (A.U.)	Horse expression (average of LV samples, A.U.)
KCND2/Kv4.2	Phase 1	Fast transient outward potassium current	No evidence	No evidence
KCND3/Kv4.3	Phase 1	Fast transient outward potassium current	No evidence	No evidence
SCN5A/Nav1.5	Phase 0	Depolarization	187,667	146,667
SCN4A/Nav1.4	Phase 0	Depolarization (primarily in skeletal muscle)	No evidence	4333
CACNA1C/Cav1.2	Phase 2	Voltage-gated calcium channel	9,700,000	129,667
CACNA1S/Cav1.1	Phase 2	Voltage-gated calcium channel	No evidence	No evidence
KCNQ1/Kv7.1	Phase 2	Slow delayed rectifier	No evidence	21,000,000
KCNH2/Kv11.1	Phase 2	Rapid delayed rectifier	No evidence	No evidence
KCNA5/Kv1.5	Phase 2	Ultra-rapid delayed rectifier	No evidence	3,300,000
KCNJ11/Kir6.2	?	ATP-sensitive inward rectifier	No evidence	No evidence
RYR1/RYR1	n/a	Calcium release from sarcoplasmic reticulum (primarily in skeletal muscle)	1,266,667	36,833,333
RYR2/RYR2	n/a	Calcium release from sarcoplasmic reticulum	12,800,000,000	25,000,000,000
CASQ1/CASQ1	n/a	Sarcoplasmic reticulum Ca2+ binding protein (primarily in skeletal muscle)	No evidence	3,066,667
CASQ2/CASQ2	n/a	Sarcoplasmic reticulum Ca2+ binding protein	24,333,333	20,333,333
ATP2A1/SERCA1	n/a	Calcium sequestration into sarcoplasmic reticulum (primarily in skeletal muscle)	11,333,333	806,666,667
ATP2A2/SERCA2	n/a	Calcium sequestration into sarcoplasmic reticulum	60,000,000,000	69,333,333,333

A.U. – arbitrary units.

The ATP-sensitive potassium channel, K_ir6.2/KCNJ11 is of particular interest in athletes due to its role in matching cellular metabolism with membrane excitability, as well as its upregulation in response to exercise [52]. The contribution of K_ir6.2 to the cardiac action potential is not well described since it is primarily closed at rest [53], however it likely plays a role in shortening of the cardiac action potential with increases in heart rates during exercise. It is possible, therefore, that there is an optimal amount of K_ir6.2 activity whereby animals who express lower levels and have longer action potentials at higher heart rates are at higher risk of re-entrant arrhythmias (as occurs during myocardial ischemia [53]). Thus, if the higher level of K_ir6.2 predicted in horses is manifest at the protein level, this unique expression pattern could explain the higher risk of SCD reported in equine athletes, relative to humans.

Expression of predominant delayed rectifier potassium channels K_v7.1/KCNQ1, K_v11.1/KCNH2 and accessory proteins encoded by genes KCNE1 and KCNE3 have been confirmed with immunoblotting in equine myocardium [54]. However, K_v7.1 and K_v11.1, which are known to be important delayed rectifiers, were not identified in human tissue in a proteomic analysis, and only K_v7.1 was found in horses [48]. Additionally, equine KCNQ1, KCNH2, and KCNE2 have been sequenced and investigated in functional expression studies, which revealed some electrophysiological changes that are presently of unknown clinical significance [55,56].

There is, however, some consistency across studies with regards to K_v1.5/KCNA5. K_v1.5 was predicted [51] and its expression confirmed [48] in equine ventricles, but it is primarily expressed in human atria compared to ventricles [57]. KCNA5 encodes an ultrarapid delayed rectifier potassium channel and its induced expression in isolated cardiomyocytes has showed decreased action potential durations resulting in more rapid and irregular automaticity [58]. This unique ventricular K_v1.5 expression offers a plausible explanation for the more rapid repolarization demonstrated in horses (Fig. 4B). Therefore, its expression in horses may present a unique opportunity to understand its significance, particularly with reference to action potential duration in response to changes in heart rate and risk of exercise-associated cardiac arrhythmias. Mutations in KCNA5 that lead to slower repolarization could be a predisposition to arrhythmias and SCD in horses that is not present in humans.

The sodium/potassium and sodium/calcium exchangers on the cell surface have not been described in horses and are not well described in the absence of heart failure or myocardial infarctions in humans. Despite not contributing an ion current to the cardiac action potential, the exchangers restore ionic gradients during and between successive action potentials. Therefore, characterizing their expression and functionality in horse and human athletes may provide insight into the ability of the myocardium to tolerate sustained increases and dramatic changes in firing rates, as observed in athletic activity. This is particularly true in horses due to the prominent plateau phase of the action potential; the presumably higher influx of calcium likely demands a unique calcium-handling system.

Due to the inconsistencies in data identifying the expression of numerous ion channels, further work is necessary to create a fundamental understanding of the molecular elements that contribute to cardiac action potentials in horses, and to refine our understanding and provide further insight on mechanisms that may contribute to exercise-associated SCD in both horse and human athletes. Additionally, other myocardial proteins, including those found in the myofilament complex; cell-cell junctions and intercalated discs; and calcium handling proteins have limited research in horses and therefore further work is needed to comment on the utility of horses as a model for humans specifically in those areas.

6. Limitations

Much of the relevant literature in horses does not report the breed or use of the horses studied [28,32,34,40,48] and extrapolations from these populations to the performance horse or general horse populations should be undertaken with caution. In future, reporting such details as the source of the horse (veterinary patient vs. slaughter plant), athletic use, age, and breed would improve our ability to interpret findings. Additionally, much of the literature cited herein was performed using small sample sizes from a single research group, and therefore there remain many unanswered questions about the degree to which findings can be generalized.

7. Conclusions

Performance horses are a good model to investigate exercise-associated cardiac arrhythmias and SCD in athletes, and they offer specific opportunities to investigate the cellular and molecular characteristics of a heart adapted for maximal athletic performance. Much of the myocardial blood supply and conduction systems are similar to humans. However, there are important differences in the presence of myocardial bridging and distribution of Purkinje fibers. Comparative studies offer a particularly good opportunity to investigate the significance of CAA, the role of autonomic input on triggered activity in the His bundle, and the role of K_v1.5 in the response of the action potential to changes in heart rate. Further work is needed to fully understand any differences in the ion channelome, especially with reference to characteristics such as sex, age, and athletic use.

CRediT authorship contribution statement

Amanda Avison: Writing – review & editing, Writing – original draft, Visualization, Conceptualization. Peter W. Physick-Sheard: Writing – review & editing. W. Glen Pyle: Writing – review & editing, Conceptualization.

Funding sources

AA, PWPS, and WGP receive funding from Equine Guelph for research on sudden cardiac death in horses.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.