Abstract
ing on deep sand is commonly employed in endurance horses, but its physiological adaptation remains poorly characterized. This study aimed to characterize locomotor adaptations during a 7 km controlled-speed canter on deep sand in eighteen endurance horses, to identify heart rate variability (HRV) components, and to investigate changes in hematological variables before and after exercise. Stride frequency (SF) and stride length (SL), HRV, and hematological profiles were recorded during exercise and recovery with a fitness tracker. Associations between maximum speed and locomotor parameters were assessed by linear regression, while Pearson’s correlation assessed HRV relationships, also with physiological parameters. Hematological parameters were assessed with paired t-test before and after training. SL percentage change was the strongest predictor of speed (β= 0.677). HRV analysis revealed delayed parasympathetic reactivation; the parasym- pathetic recovery index (PNS REC) was correlated with mean RR interval on the ECG (r = 0.968) and heart rate (r =−0.964) during recovery. Post-exercise rectal tempera- ture showed correlations with HRV recovery indices. Hematological evaluation revealed Vet. Sci.2025,12, 1028 https://doi.org/10.3390/vetsci12111028
Vet. Sci.2025,12, 1028 2 of 13 post-exercise increases in red blood cell count, hematocrit, hemoglobin, and corpuscular indices. SL plays a predominant role in achieving higher speeds on deep sand, while PNS REC emerges as a practical and accessible marker of autonomic recovery and fatigue. Horses with enhanced thermoregulation recover better. Hematological results confirm a physiological stress response that may optimize oxygen delivery. Integrating locomotor, cardiovascular, and hematological monitoring may improve management and welfare in endurance training. Keywords:equine; athlete; exercise; autonomic regulation; fitness tracker 1. Introduction Endurance horses represent a unique model of the equine athlete, as they are required to sustain prolonged aerobic effort in the Federation Equestre Internationale (FEI) competi- tions over distances ranging from 100 to 160 km, often under challenging environmental conditions [1]. Performance in this discipline depends on a delicate balance between cardiovascular efficiency, metabolic resilience, thermoregulation, and musculoskeletal integrity [2,3]. Unlike speed-oriented equestrian disciplines, endurance riding places greater emphasis on fatigue resistance, recovery capacity, and long-term soundness [4]. For these features, endurance riding is one of the most energetically demanding equestrian discipline, making the evaluation of fitness essential. Training programs must be care- fully designed to enhance aerobic capacity while minimizing the risk of overtraining and injury [5,6]. In this context, the standardization and objective monitoring of training ses- sions are of paramount importance. By continuously assessing locomotor patterns, heart rate dynamics, and autonomic regulation, trainers and veterinarians can better tailor con- ditioning strategies, identifying early signs of fatigue or maladaptation, and optimizing the welfare and competitive longevity of the endurance horse [7]. Traditionally, exercise capacity in horses is assessed using the heart rate (HR) response, which is influenced by the balance between sympathetic and parasympathetic efferent controls [8,9]. In human medicine, the more promising method to monitor individual adaptation to training in- volves the monitoring of the cardiac autonomic nervous system through the measurement of exercise [10] and recovery heart rate variability (HRV) [11]. Negative adaptation to training is thought to be associated with post-exercise sympathetic nervous system hy- peractivity that may lead to ischemic heart disease, ventricular arrhythmias, and sudden cardiac death
promising method to monitor individual adaptation to training in- volves the monitoring of the cardiac autonomic nervous system through the measurement of exercise [10] and recovery heart rate variability (HRV) [11]. Negative adaptation to training is thought to be associated with post-exercise sympathetic nervous system hy- peractivity that may lead to ischemic heart disease, ventricular arrhythmias, and sudden cardiac death [12,13]. In the equestrian endurance discipline, rapid cardiac recovery is a key indicator of fitness and success and serves as a reliable marker of fatigue during com- petition [14], which highlights the importance of evaluating post-exercise parasympathetic reactivation by using vagal-related indices of HRV. In equines, cardiac activity is influ- enced by various factors, including individual level of training [15], age [14], duration of exercise [5], and environmental conditions [1], but the effect of track surface on cardio- vascular physiology is unknown. Studies in human sports medicine demonstrated that training on soft and compliant surface, like deep sand, increases propulsive muscular effort, resulting in a higher HR at lower speed [16]. This would allow training of the cardiores- piratory system without reaching the maximal speed, decreasing stress on osteoarticular, tendinous, and ligamentous structures. In recent years, equine monitoring systems have undergone rapid technological development. The Equimetre ® 2.0 is a wearable device de- signed for monitoring the training and performance of horses. The device integrates surface ECG electrodes, inertial sensors (accelerometer and gyroscope), and multi-constellation GNSS to record heart rate, heart rate variability, stride parameters, and speed during field
Vet. Sci.2025,12, 1028 3 of 13 exercise. It introduced significant improvements, including real-time data transmission via 4G network, enhanced mechanical robustness, and a more compact design, thereby increasing reliability and expanding field applications [17]. At the same time, researchers have explored the use of inertial measurement units (IMU) to assess locomotor symmetry and enable early detection of lameness, often supported by artificial intelligence algorithms and convolutional neural networks, with promising diagnostic accuracy [18,19]. Other innovative approaches include the analysis of respiratory patterns during exercise using mi- crophones combined with deep learning models, capable of identifying breathing dynamics with high precision [20]. These advances highlight how the integration of cardiovascular, respiratory, locomotor, and behavioral data may pave the way toward equine deep phe- notyping, in which tools such as Equimetre ® are positioned within a broader ecosystem of digital technologies applied to performance monitoring and athlete welfare. Moreover, Equimetre ® , has been validated to accurately record HR in conjunction with telemetric electrocardiogram (ECG) during training sessions, providing information on HRV [21]. Its accuracy for HR and HRV measurement, as well as for arrhythmia detection, has been validated against telemetric ECG recordings in controlled and field conditions [9,21]. This technology can monitor cardiovascular response to exercise and recovery, characteristic of locomotory function and speed. To date, no studies have been conducted to report such parameters for endurance horses trained in deep sand. Moreover, the respiratory system is the primary factor for oxygen (O2) delivery, aerobic metabolism, and athletic performance in the horse [22]; indeed, during exercise, ventilation must increase in response to the metabolic demand of exercising muscle. Since training in deep sand requires elevated muscular activity, and since during cantering there is a tight locomotor-respiratory cou- pling, locomotor adaptation to sand needs to be investigated. The aims of this study were: (i) to characterize the locomotor adaptations to deep sand by analyzing stride frequency (SF) and stride length (SL) during canter; (ii) to identify HRV components in endurance horses trained in deep sand; (iii) to investigate the changes in some hematological vari- ables before and after training on deep sand.
sand needs to be investigated. The aims of this study were: (i) to characterize the locomotor adaptations to deep sand by analyzing stride frequency (SF) and stride length (SL) during canter; (ii) to identify HRV components in endurance horses trained in deep sand; (iii) to investigate the changes in some hematological vari- ables before and after training on deep sand. Hematological evaluation, indeed, provides complementary insight into the physiological adaptations to exercise. In equine athletes, complete blood count (CBC) is routinely performed to assess health and training status, as it provides information on splenic contraction, plasma volume shifts, and oxygen transport efficiency. In this study, these classical hematological parameters were analyzed because they are part of the standard CBC routinely used in athletic horses and allow non-invasive assessment of physiological adaptation. The innovative application is the integration of these hematological indicators with real-time physiological and locomotor data collected by a wearable fitness tracker. This approach enables a better interpretation of blood changes in the context of autonomic and mechanical workload and suggests that, in the future, continuous digital monitoring could complement or even replace routine hematological testing for evaluating training adaptation and fitness in endurance horses. The authors hypothesized that the primary locomotor strategy for acceleration relies on increasing SL, a mechanism representing the first mechanism of acceleration in longer- distance performers, thereby contributing to performance efficiency. Furthermore, the authors proposed that superior training adaptation in endurance horses is reflected by greater HRV amplitude, indicating enhanced parasympathetic activity. Lastly, the authors hypothesized that a relatively short training session on deep sand could also result in changes in the hematological variables, resulting in mobilization of splenic erythrocytes and, therefore, increasing the oxygen transport capacity.
Vet. Sci.2025,12, 1028 4 of 13 2. Materials and Methods 2.1. Horses Eighteen elite endurance horses were included in this study. All horses were stabled in the same training center (Rabdan Endurance Stable, Dubai, United Arab Emirates) and were subjected to the same management in terms of training and nutrition. The study population included 11 Arabian and 7 Anglo-Arabian endurance horses, aged from 8 to 20 years (median 13 yo), consisting of 6 mares, 11 geldings, and 1 stallion. Information on each horse’s training and racing profile (such as categories raced, total distance covered during races, number of races, and weekly training schedules) was obtained from the FEI database or orally obtained by the trainer. The level of training of each horse was determined based on the highest Concours d’Endurance Internationale (CEI) category in which the animal had competed during the 2024–2025 season; 2 horses competed in CEI*, 11 horses in CEI**, and 5 in CEI***. 2.2. Training Session Data were collected during a training session on deep sand in the Dubai desert. All horses underwent a dynamic examination at trot in straight line and a short clinical examination (HR, respiratory rate, rectal temperature) before the start of the training to assess that they were fit to perform. Before the beginning of the training session, each horse was equipped with a fitness tracker (Equimetre ® , Arioneo, Paris, France) fitted as previously described [21]. Throughout all training sessions, the ambient temperature ( ◦ C) and humidity (%) were recorded using a portable device (GPS Temperature and Humidity Data Logger GSP-6, Trans Instruments Ltd., Leicester, UK) held within 4 km from the deep sand training surface. The training exercise consisted of 15 km total distance. The warm-up was performed on compact sand and included 2 km of walk, followed by 2 km of trot at unregulated speed. This was followed by 7 km of canter on an oval deep-sand track at a controlled speed of 21 km/h (5.8 m/s), regulated using a GPS monitoring device (Garmin ® Forerunner 945 watch, Garmin Ltd., Olathe, KS, USA) held by the rider. The return
sand and included 2 km of walk, followed by 2 km of trot at unregulated speed. This was followed by 7 km of canter on an oval deep-sand track at a controlled speed of 21 km/h (5.8 m/s), regulated using a GPS monitoring device (Garmin ® Forerunner 945 watch, Garmin Ltd., Olathe, KS, USA) held by the rider. The return to the stables was performed on compact sand and included 2 km of trot at unregulated speed and 2 km of canter at unregulated speed. Upon arrival, rectal temperature was collected and horses were unsaddled, cooled down with cold water, and underwent a dynamic examination at trot in straight line to assess for any signs of lameness following the training. Data recorded from the Equimetre ® were downloaded via a wireless connection and automatically uploaded to the online platform for analysis. 2.3. Equimetre ® Data Analysis Data on locomotor performance during the 7 km of canter on deep sand (HIGH_EXERCISE) and data on cardiac activity during HIGH_EXERCISE and dur- ing the first 5 min of 2 km recovery trot (REC_EXERCISE) were collected using Equimetre ® . Speed- and stride-related variables were obtained by a combination of GNSS (GPS + GLONASS + Galileo), which provided real-time positioning data used to calculate velocity, and accelerometers and gyroscopes which measured the acceleration patterns of the horse’s movement. From inertial sensors, SF and SL were derived. HR data were measured using surface electrodes embedded in the device’s girth, which captured the ECG signal throughout the session (Figure). All raw and processed parameters were auto- matically transmitted via Bluetooth and cellular connection to the Equimetre ® cloud-based platform immediately after exercise. The platform stored the data in the individual horse’s profile, where they were exported through the online dashboard for analysis. Locomotor parameters included maximal speed (Smax), stride frequency (SFmax) and stride length (SLmax) at maximal speed, and stride frequency (SF20) and stride length (SL20) at 20 km/h.
through the online dashboard for analysis. Locomotor parameters included maximal speed (Smax), stride frequency (SFmax) and stride length (SLmax) at maximal speed, and stride frequency (SF20) and stride length (SL20) at 20 km/h.
Vet. Sci.2025,12, 1028 5 of 13 Cardiac parameters considered for HIGH_EXERCISE and REC_EXERCISE included mean, maximum, and minimum HR. ECG data collected by the device during HIGH_EXERCISE and REC_EXERCISE were exported as csv files from the online dashboard and imported into Kubios ® HRV Standard (Kubios HRV software [version 3.0.2], Biomedical Signal Anal- ysis Group, Department of Applied Physics, University of Kuopio, Finland) for analysis of time-domain variables to determine HRV. An automatic identification of the RR was per- formed with a strong artifact correction filter applied to reduce error across the sample [23]. Occasional ectopic beats were visually identified and manually replaced with interpolated adjacent RR interval values. The mean RR interval, standard deviation of the normal RR (SDNN), and root mean square of the standard deviation (RMSSD) were calculated for HIGH_EXERCISE and REC_EXERCISE. Poincaré plot analysis identified the shape of the ellipse made by plotting each RR interval as a function of the previous RR interval. The ellipse’s width (SD1) and length (SD2), together with parasympathetic (PNS) and sympathetic (SNS) indices, were collected during HIGH_EXERCISE and REC_EXERCISE. All the data were exported in text file for statistical analysis. Figure 1.Example of Equimetre ® physiological and biomechanical parameters recorded during the training session. The upper panel shows the heart rate (HR) (red line, bpm; lefty-axis) and speed (blue line, km/h; right y-axis) as a function of distance (x-axis, km). HR increases progressively after the initial warm-up, stabilizes during steady-state running on deep sand, and decreases toward the end of the session. Speed follows a similar pattern. The lower panel displays stride frequency (SF) (orange line, stride/s; lefty-axis) and stride length (SL) (green line, m; righty-axis) as a function of distance (x-axis, km). Both parameters show initial fluctuations during the warm-up, stabilization during the main running phase, and a decline toward the end of the session. The dotted vertical line at ~11 km indicates the end of the 7 km canter on deep sand. 2.4. Blood Parameters Blood samples were collected by the attending stable veterinarian via venipuncture of the external jugular vein, using a vacutainer system with EDTA collection
warm-up, stabilization during the main running phase, and a decline toward the end of the session. The dotted vertical line at ~11 km indicates the end of the 7 km canter on deep sand. 2.4. Blood Parameters Blood samples were collected by the attending stable veterinarian via venipuncture of the external jugular vein, using a vacutainer system with EDTA collection tubes. These samples were obtained as part of routine clinical monitoring to assess the horses’ health status during training and ensure their fitness prior to training. The complete blood count, including red and white cell parameters, was therefore performed for clinical purposes. The data were subsequently analyzed for research purposes to investigate hematological changes associated with exercise on deep sand.
Vet. Sci.2025,12, 1028 6 of 13 2.5. Statistical Analyses Data from Equimetre ® were analyzed using JASP free statistical software (version 0.18.3, JASP Team, Amsterdam, The Netherlands). Numerical variables were presented as mean±standard deviation or median and range, as appropriate; nominal data were presented as frequencies and percentages. Descriptive statistic was performed for rectal temperature before and after exercise, and environmental conditions; in addition, descriptive statistic was performed for blood analysis results, locomotor parameters, and HRV analysis stratified by HIGH_EXERCISE and REC_EXERCISE. Numerical data were assessed for homoscedasticity using Shapiro–Wilk test, for normality, and Levene test, for homogeneity of the variance, and tests were applied as appropriate. Pearson’s correlation coefficients were calculated among Smax, SFmax, SLmax, SF20, and SL20. For each horse, the difference between the measurements at maximum velocity and those at 20 km/h was calculated as∆SF and∆SL. The percentage variation was calculated as: ∆%SF = ((SFmax−SF20)/SF20))×100 ∆%SL = ((SLmax−SL20)/SL20)×100 A linear regression model was built to estimate the relationship between Smax and percentage of changes (∆%SF and∆%SL). Regression coefficients (standardized) were interpreted to assess the relative contribution of each variable to maximum speed. To assess relationships among all HRV variables and physiological parameters, Pearson’s correlation was performed. Correlation coefficients were interpreted according to thresholds previously described in the literature, where the strength of the relationship is classified as poor (r < 0.20), weak (0.20≤r < 0.50), moderate (0.51≤r < 0.80), strong (0.81≤r < 0.99), and perfect (r≥0.99) [24]. Hematological parameters were measured in horses before (PRE) and after (POST) training session, and paired-samplet-tests were performed for each parameter to compare PRE and POST values. The statistical analyses and boxplots were performed in R, version 4.3.0 (R Core Team, Vienna, Austria) using the tidyverse 2.0.0, ggplot2 3.5.2, and ggpubr 0.6.1 packages. Significance was set atp< 0.05. Given the limited number of elite endurance horses available, no a priori sample size calculation was performed. Instead, a post hoc sensitivity analysis was conducted (α= 0.05, power = 0.80) to estimate the minimum detectable effect sizes with the available sample (N= 18) and presented in Supplementary Table S1. This sensitivity analysis indicated
Description
This study investigates locomotor adaptations and physiological responses in endurance horses trained on deep sand.