Altered systolic left ventricular function in horses completing a long distance endurance race




Reasons for performing study: It is unknown whether or not exercise-induced cardiac fatigue (EICF), as demonstrated in human athletes performing long duration exercise, occurs in endurance horses.

Objective: To examine the effects of a long distance endurance race on left ventricular systolic function in horses.

Methods: Echocardiography was performed before, and after, a 2 or 3 star international endurance race (106–132 km) in 11 horses. Systolic (s) and diastolic (d) interventricular and left ventricular free wall thickness (IVS and LVFW, respectively), left ventricular, left atrial and aortic internal diameter (LVID, LA and Ao, respectively), fractional shortening (FS) and ejection fraction (EF) were measured by echocardiography. Heart rate (HR), peak flow velocity (Vmax), flow velocity integral (FVI), ejection time (ET), pre-ejection period (PEP), velocity of circumferential fibre shortening (Vcf), stroke volume (SV) and cardiac output (CO) were measured from aortic Doppler wave recordings.

Results: After the race, LVIDd, Ao, LA, EF, FS, FVI, SV, ET and ET indexed for HR were significantly lower and IVSd, LVFWd, HR, PEP, PEP/ET and Vcf significantly higher as compared with prerace values. Pre- to post exercise changes in those parameters were not significantly correlated with changes in HR or in LVIDd.

Conclusions: Results suggest that EICF, with a decrease in left ventricular systolic function, could occur post exercise in horses performing long duration endurance races. However, a multanecus effect of altered preload and heart rate on the studied variables cannot be discounted.


The cardiac effects of chronic endurance training is well known in man, i.e. an enlargement of the left ventricular (LV) cavity, a reduction in resting heart rate (HR) and an augmented stroke volume (SV) (Middelton et al. 2006). Increases in LV diastolic dimensions and mass with race training have also been previously demonstrated in horses (Young 1999; Buhl et al. 2005). Training-induced cardiac changes are associated with changes in the contractile and filling properties of the heart and thus confer protection against deleterious cardiac events (Middelton et al. 2006). However, despite the cardioprotective effect of training, there is growing evidence that in human healthy athletes, performing strenuous exercise for prolonged duration (i.e. more than 6 h) may result in a transient impairment of cardiac function called exercise-induced cardiac fatigue (EICF) (McGavock et al. 2002; Middelton et al. 2006; Shave et al. 2008). Exercise-induced cardiac fatigue consists of an acute post exercise significant reduction in either or both diastolic and systolic LV function (McGavock et al. 2002; Middelton et al. 2006; Oxborough et al. 2006; Leetmaa et al. 2008; Shave et al. 2008). In horses, long distance endurance races are performed on distances ranging from 120–160 km and at speed ranging from 10–20 km/h. Races usually last more than 6 h and can therefore be considered as a prolonged strenuous exercise. Little is known of the cardiac consequences of such efforts in horses.

The objective of this study was to examine the effects of a long distance endurance race on LV systolic function in healthy trained horses using Doppler echocardiography.

Materials and methods


Eleven horses performing an endurance ride recognised by the International Equestrian Federation (FEI) as a 2 or a 3 star ‘concours de raid d'endurance international’ (CEI** and CEI***, respectively) were used in the study. Inclusion criteria included normal physical examination and normal ECG and echocardiography performed the day before the race, finishing the race including the final vet gate and to be presented to the post exercise echocardiography within 1 h after finishing the race. All horses were regularly trained and considered fit before the race. There were 3 mares and 8 geldings (8 purebred Arabians, 2 crossbred Arabians and 1 Anglo-arabian), ages ranged from 7–13 years (mean ± s.d.: 10.2 ± 2.1 years old) and weight 371–511 kg (mean ± s.d.: 439 ± 41 kg).

The performed races were one of the following:

  • • Race 1: St Galmier (France), 119 km, CEI** (n = 3)
  • • Race 2: Monpazier (France), 119 km, CEI** (n = 1)
  • • Race 3: Ghlin (Belgium), 106 km, CEI** (n = 1) or 132 km, CEI*** (n = 6)


Each horse underwent a standardised echocardiographic examination the day before and as soon as possible after finishing the race. A commercially available portable ultrasound system equipped with colour flow mapping and spectral Doppler mode and a 1.5–3.6 MHz transducer1 was used. An ECG was recorded simultaneously with the echocardiographic images. All examinations were recorded digitally and analysed offline.

The horses were first imaged from the right hemithorax to record a 2-dimensional (2D) right parasternal long axis LV outflow tract view and then to record an M-mode short axis view of the left ventricle at the level of the chordae tendinae.

The horses were then examined from the left hemithorax to record a 2D left parasternal long axis angled view of the left ventricle at the level of the largest diameter of the left atrium. From this view, the beam was then rotated anticlockwise until the aortic root was visible to obtain a 2D left parasternal apical 5 chamber view. In this view, the gate of the Doppler pulsed-wave mode was placed just distal to the aortic valve for recording the aortic outflow velocity spectrum.

Echocardiographic measurements

All measurements were the mean for 3–5 cardiac cycles. Measured cardiac cycles were not necessarily consecutive. The interventricular septal and LV free wall thicknesses (IVS and LVFW, respectively) and LV internal diameter (LVID) were measured at end-diastole (d) and at end-systole (s) from the M-mode right parasternal short-axis view of the left ventricle. From LVID, end-diastolic (EDV) and end-systolic (ESV) volumes were calculated using the Teicholz method (Boon 1998). The LV fractional shortening and ejection fraction (FS and EF, respectively) were calculated using the classical formulae (Boon 1998). The aortic diameter (Ao) and left atrium internal diameter (LA) were measured from the right parasternal 2D long axis LV outflow tract view and from the 2D left parasternal long axis 4-chamber angled view, respectively (Boon 1998).

From the pulsed-wave Doppler tracing of the aortic flow, peak flow velocity (PFV), flow velocity integral (FVI), ejection time (ET), pre-ejection period (PEP), and HR were measured. The SV and cardiac output (CO) derived from FVI, Ao and HR measurements were calculated using the classical formula (Boon 1998). The ratio PEP/ET was also calculated. The indexed ET (ETI) was calculated as ET + (0.55 × HR) (Boon 1998). The Velocity of circumferential shortening (Vcf) was calculated as (LVIDs-LVDd)/(LVIDd X ET), and the indexed Vcf (VcfI) was calculated as Vcf X 100/HR (Boon 1998).

Statistical analysis

The various traits were analysed using a mixed linear model, fitting the horse as random effect and the measurement time (before or after the race) as a fixed effect (Searle et al. 1992).

Post hoc comparisons of measurement times were made using least square means and P values below 0.05 were considered as significant.

Moreover, correlations between pre- to post exercise changes in functional echocardiographic, Doppler flow and tissue Doppler parameters and pre- to post exercise changes in HR and in LVIDd were assessed via a Pearson's product-moment analysis.


The studied horses finished the race at the first to the 23rd place. The mean run distance was 124.9 ± 8.9 km (ranging from 106–132 km). The mean duration of the course was 7 h 37 min ± 61 min (ranging from 6 h 22 min to 8 h 43 min) and mean speed was 16.3 ± 1.3 km/h (ranging from 14.6–18.5 km/h).

Echocardiography was performed 36 ± 11 min (ranging from 27–55 min) after finishing the race. Results of echocardiography are shown in Table 1 and results of aortic Doppler flow are given in Table 2. Diastolic LV wall thicknesses, PEP, PEP/ET, HR and Vcf were significantly higher after than before the race and LVIDd, Ao, LA, EDV, ESV, FS, EF, FVI, ET, ETI and SV were significantly lower after than before the race. The evolution of SV, FS and PEP/ET before and after the race in each horse is shown in Figure 1.

Table 1. Mean ± s.d. values of echocardiographic parameters measured before and after completion of a 106–132 km endurance race in 11 healthy adult horses
 PreracePost race
  • *

    d and s, diastolic and systolic measurement, respectively; IVS and LVFW, thickness of the interventricular septum and of the left ventricular free wall, respectively; LVID, internal diameter of left ventricle; Ao and LA, internal diameter of the aorta and of the left atrium, respectively; FS, fractional shortening of the left ventricle; EVD and ESV, end-diastolic and end-systolic left ventricular volume, respectively; EF, ejection fraction. Significantly different from the mean value measured before the race, mixed linear model, P≤0.05.

IVSd (cm)2.71 ± 0.123.11 ± 0.12*
IVSs (cm)4.34 ± 0.134.30 ± 0.13
LVIDd (cm)11.28 ± 0.209.91 ± 0.20*
LVIDs (cm)6.52 ± 0.206.05 ± 0.20
LVFWd (cm)2.37 ± 0.102.62 ± 0.10*
LVFWs (cm)4.08 ± 0.174.16 ± 0.17
Ao (cm)7.6 ± 0.127.3 ± 0.12*
LA (cm)11.7 ± 0.210.5 ± 0.2*
FS (%)42.3 ± 1.538.5 ± 1.5*
EDV (ml)737.2 ± 26.3556.9 ± 26.3*
ESV (ml)221.6 ± 14.9193.8 ± 14.9
EF (%)70.4 ± 1.866.0 ± 1.8*
Table 2. Mean ± s.d. values of aortic Doppler flow parameters measured before and after completion of a 106 to 132 km endurance race in 11 healthy adult horses
 PreracePost race
  • *

    PFV, peak velocity of the systolic aortic flow; FVI, flow velocity integral of the systolic aortic flow; PEP and ET, pre-ejection period and ejection time as evaluated by the aortic Doppler flow, respectively; HR, heart rate; SV, stroke volume; CO, cardiac output; Vcf, peak velocity of circumferential fibre shortening; VcfI, indexed Vcf. Significantly different from the mean value measured before the race, mixed linear model, P≤0.05.

Aortic flow
PFV (m/s)0.84 ± 0.030.76 ± 0.02
FVI (cm)27.15 ± 0.9519.00 ± 0.91*
PEP (ms)108.6 ± 4.5139.4 ± 4.3*
ET (ms)445.6 ± 13.3358.0 ± 12.7*
ETI (ms)418.6 ± 8.2374.5 ± 7.9*
PEP/ET0.24 ± 0.020.40 ± 0.02*
HR36.9 ± 2.150.2 ± 2.0*
SV (ml/beat)1245 ± 61839 ± 63*
CO (l/min)45.1 ± 2.440.4 ± 2.5
Vcf (mm/s)1.64 ± 0.112.03 ± 0.11*
VcfI (mm/beat)4.41 ± 0.264.09 ± 0.25
Figure 1.

Evolution of the stroke volume (SV), fractional shortening (FS) and the ratio pre-ejection period to ejection time (PEP/ET) before (PRE) and after (POST) completion of a 106–132 km endurance race in 11 healthy adult horses.

Pre- to post exercise changes in functional echocardiographic and Doppler variables were not significantly correlated with corresponding changes in HR and LVIDd.


It is pertinent to first mention the limitations of this study. As with most complex field-based studies, only a limited number of horses were included, making general conclusions more difficult. The number of investigated individuals in most human studies on EICF was similarly limited (Leetmaa et al. 2008; Shave et al. 2008). To increase the number of studied horses in this study, it was performed at 3 different competitions and the run distances and environmental conditions were, therefore, not standardised. However, all studied horses performed distance around 100 km at a speed ranging from 14–18 km/h. Consequently, the exercise was strenuous for all investigated horses and had a total duration of 6 h 30 to 8 h 45, which is comparable to humans performing long to ultra LDE, such as ironman triathlon or ultra marathon.

The phenomenon of EICF has been studied in depth in human athletes: several hundred studies including review papers and meta-analyses have been performed on this subject in the last few decades. Conclusions drawn from those studies remain somewhat equivocal (Shave et al. 2004). However, there are considerable differences in the methods and research designs of EICF studies: both laboratory or field studies have been completed, the nature and duration of exercise widely varied and the athletes studied have been heterogeneous in terms of age, sex and training status. The disparities between those studies could explain some of the disparities in the drawn conclusions (Shave et al. 2008). Indeed, numerous confounding variables have been suggested in the aetiology of EICF (Shave et al. 2004; Vidotto et al. 2005). Amongst them, one of the most important is the duration of exercise and demonstrated to be particularly important is exercise-induced systolic dysfunction: this occurs in most athletes only after at least 6 h of strenuous exercise (McGavock et al. 2002; Middelton et al. 2006). In contrast, diastolic dysfunction, not studied in the present work, is less dependent on exercise duration and may occur as soon as after 1 h of strenuous exercise (Shave et al. 2004, 2008; Oxborough et al. 2006, Middelton et al. 2006). The training status also appears to be a key determinant of EICF that clearly occurs earlier in untrained than in trained individuals. Other factors such as nature and intensity of the exercise, the athlete's age or sex and adverse environmental conditions such as altitude, heat, or humidity, have also been incriminated in the aetiology of EICF (Shave et al. 2004; Vidotto et al. 2005).

The impact of EICF is unknown. Because it has been shown to disappear within 24 h after completion of the exercise and has been shown not to be associated with a large increase in myocardial injury markers, its clinical impact is probably minimal (Shave et al. 2004, 2008; Middelton et al. 2006). However, EICF could have an affect on performance, as does metabolic problems or skeletal muscle alterations (Middelton et al. 2006). Exercise-induced cardiac fatigue in athletes performing several bouts of prolonged exercise on successive days has been studied less and could be more deleterious to cardiac function than a single bout of strenuous exercise (McGavock et al. 2002; Middelton et al. 2006; Shave et al. 2008). The same could be true in equine endurance races performed on several days. Moreover, significant individual differences in the pattern of EICF have been shown in human athletes (Shave et al. 2008) and long duration strenuous exercise may be dangerous for cardiac function in some athletes. Exercise-induced cardiac fatigue has even been mentioned as potentially responsible for exercise-induced sudden cardiac death in athletes without any history of previous occult cardiovascular disease (McGavock et al. 2002).

The results of this study demonstrate significant changes in most of the parameters evaluating the systolic LV function (FS, EF, FVI, SV, Vcf and systolic time intervals) after completion of a strenuous prolonged endurance exercise in healthy horses, suggesting the occurrence of exercise-induced systolic LV dysfunction. However, these findings should be interpreted with caution because these changes could be explained by the simultaneous increase in HR and decrease in LV preload (as suggested by the decrease in LVIDd, EDV and LA after the race). Such changes in loading conditions of the heart are known to affect most Doppler and echocardiographic parameters (Boon 1998). Similar preload reduction has been reported in athletes performing LDE and has been attributed to dehydration-induced decrements in central blood volume and redistribution of blood to the periphery (Hart et al. 2007). Afterload is another factor that is known to affect echocardiographic parameters (Boon 1998). Most studies performed in human athletes suggested a tendency to a reduced afterload after LDE as evaluated by a reduction in the LV meridional wall stress, incorporating systolic blood pressure and M-mode measures of the left ventricle (McGavock et al. 2002). In this study, the afterload was not evaluated because of technical limitations in accurately measuring the absolute value of arterial blood pressure in conscious horses noninvasively (Corley 2008). However, it could be hypothesised that the same tendency in man may be observed in horses and potential reduction in afterload could have acted in the opposite direction than the reduction in preload.

In human athletes, the role of the reduced preload or the increased HR on exercise-induced Doppler and echocardiographic parameters changes has been studied using different approaches. First, a decrease in cardiac function under unaltered changes in LV loading conditions has been previously reported (Shave et al. 2004; Middelton et al. 2006). Alternative techniques to conventional and Doppler echocardiographic techniques, such as the use of parameters issued from pulmonary vein flow velocities or from TDI (including mitral annulus velocity and myocardial strain or strain rate), have been used as less HR- or preload-dependent measures of cardiac function (Middelton et al. 2006; Hart et al. 2007). Studies using these techniques further supported a reduced LV function after LDE (George et al. 2005; Shave et al. 2008). However, actual preload independence of those techniques has been questioned (Hsiao et al. 2005). Finally, in a meta-analysis performed on an overall sample of 413 individuals, exercise-induced changes in EF were shown to be correlated with changes in LVIDd, and thus it was concluded that the reduction in systolic function in those athletes was partly due to an altered cardiac loading (Middelton et al. 2006). In the present study, such correlation analysis, even if the results are to be interpreted with caution because of the limited number of data, suggests that changes of most of the measured parameters were not significantly correlated with changes in LVIDd or in HR, which could support the hypothesis of a decreased systolic function after a long distance endurance race in horses.

The significant decrease in FS, EF, FVI and ET and significant increase in Vcf and PEP suggest systolic dysfunction. All those parameters are known to be highly load- and HR-dependent (Boon 1998). Indeed, when corrected for HR (VcfI), Vcf was no more significantly increased post exercise. On the contrary, even after correction of ET for HR, ETI was significantly reduced post exercise, as was PEP/ET, a parameter known to be less HR dependent than the other systolic intervals (Boon 1998; Anderson 2007).

In conclusion, the results of this study suggest that in healthy horses performing a long distance endurance race, a systolic LV dysfunction could occur directly after the race, but those changes are probably partly, but not entirely, associated with an increased HR and a decreased preload.


The authors thank all the horse's owners and riders (P. Auffret, J. Begaud, M. Bertoli, A. Chanvin, J. Goachet, J.M. Grimal, L. Houassin, B. Lamoriniere, P.M. Morvan, R. Muller and J. Van Cauter) who participated in the study.

This work was supported by the ‘Comité d'Orientation Scientifique et Technique’ (COST) from the ‘Haras nationaux’ of France and by the ‘Fonds spéciaux pour la recherche – Crédits classiques 2008’ from the University of Liège.

Conflicts of interest

The authors have not declared any conflicts.

Manufacturer's address

1 Vivid I, General Electric Healthcare Europe GMBH, Diegem, Belgium.