This study was approved by the Animal Research Bioethics Committee of the Faculty of Veterinary Medicine, Corporación Universitaria Remington, Colombia, on December 7, 2022, under Act No. 11-2022. Informed consent was obtained from the owners or caretakers of all horses involved in the study.

The study was conducted from September to December 2024 at a sports training center located in a tropical region at 6º 05’00.84” N latitude and 75º 20’05.23” W longitude, at 2,150 meters above sea level, with a relative humidity of 69%, an average ambient temperature of 16°C, and an annual rainfall of 2,117 mm. The training sessions took place on a longitudinal sand track approximately 40 m long. Aquatic exercises were performed in a circular pool measuring 23.3 m in diameter and 2.9 m deep, with water maintained at around 28°C.

A total of 10 CPH of both sexes (6 females and 4 males), with an approximate weight of 350 ± 20 kg, a body condition score of 5–6/9 [12], and an average age of 10 ± 4 years, were included in this study. All horses were housed in stables and fed a diet consisting of hay (Digitaria eriantha), fresh forage (Pennisetum purpureum), a commercial concentrate (2 kg/day) (Campeón Dorado®, Solla, Itagüí, Colombia), mineralized salt (60 g/day) (Brío Salts®, Italcol, Girardota, Colombia), and had free access to water.

All animals were clinically healthy and had complete, up-to-date health management plans along with regular dental care. To participate in the study, animals had to be free of orthopedic problems, heart murmurs, arrhythmias, and significant hematological or biochemical abnormalities, and they needed to be actively involved in athletic activities.

After a 30-day acclimation period, during which the horses adapted to the training environments and standardized preparation and exercise protocols, the study was carried out using a repeated-measures design with two exercise methods: ground training and pool training with stationary water flow. The pool water was regularly maintained to ensure proper quality.

Before and after both exercise modalities, horses underwent a general clinical examination that included measuring body temperature (°C), pulse, heart rate, and respiratory rate. Additionally, heart rate (HR)max was recorded using a heart rate monitor during a maximal exercise test on land and a submaximal test in the pool, considering a maximum heart rate stipulated for the breed of 200–220 beats per minute [9].

Once the exercise modality was chosen, all animals completed the respective protocol. In this study, the ground exercise was first performed in a randomly assigned order using the RAND function in Microsoft Excel for Microsoft 365 (Microsoft Corp., Redmond, WA, USA).

Horses underwent a high-intensity training routine in ground work. Each session lasted 40 min and consisted of 10 min of warm-up, 20 min of moderate (70%–80% of HRmax) to high-intensity exercise (80%–90% of HRmax), and 10 min of cool-down walking.

In the swimming modality, the horses exercised at moderate intensity (70%–80% of HRmax) in a circular pool with a 23.3-m lap. Each session lasted 40 min, divided into 10 min of warm-up, 20 min of intermittent swimming, and 10 min of stretching exercises followed by a walk to cool-down (Figure 1). During the swimming phase, the horses performed two consecutive laps in one direction, followed by a 1-min pause before switching direction. They swam two laps to the right, paused, and then swam two laps to the left. This cycle was repeated four times, totaling 16 laps in approximately 20 min.

The training frequency was three sessions per week for swimming and five sessions per week for ground training. Finally, this study did not include a control group for any exercise modality; instead, each animal served as its own control and was evaluated before and after each modality.

The obtained data were stored in a Microsoft Excel® database (Microsoft Corp., Redmond, WA, USA) and then analyzed using SPSS® version 29.0 (IBM Corp., Armonk, NY, USA). A descriptive analysis of all hematological variables, lactate levels, and heart rate was conducted.

Data distribution was evaluated using the Shapiro–Wilk test and Q–Q plots, while homoscedasticity was checked with Levene’s test; when applicable, sphericity was assessed with Mauchly’s test. Variables meeting the normality assumption were expressed as mean ± standard deviation, whereas those with a non-normal distribution were reported as median and interquartile range.

Parametric tests, such as Student’s t-test for both paired and independent samples, were used when the data met normality assumptions. When those assumptions were not met, non-parametric tests were used, specifically the Wilcoxon test for paired data and the Mann–Whitney U test for independent data.

The sample size was based on the total number of individuals available during the study period; therefore, no a priori power calculation was conducted. To compare repeated-measures, linear mixed-effects models were used, with horse included as a random effect. When model assumptions were not satisfied, nonparametric alternatives were employed. Statistical significance was set at p < 0.05.

Immediate physiological responses to exercise were assessed by comparing pre- and post-exercise values at two key points: the first and last day of the training period. Tables 1 and 2 show the average values of hemo-leukogram variables, lactate concentration, and heart rate.

Regarding immediate physiological responses, most pre-exercise values were lower than post-exercise values. In the ground exercise modality, a significant post-exercise increase was observed in erythrocyte count and hematocrit (p = 0.001), as well as in lymphocytes and globulins (p = 0.031). In the pool exercise modality, there was a significant post-exercise increase in erythrocytes (p = 0.012), hemoglobin (p = 0.001), and hematocrit (p = 0.002). Lactate concentration and heart rate followed a similar trend, showing significant post-exercise increases in both modalities (p = 0.001).

Additionally, post-exercise differences in variables such as erythrocytes, leukocytes, and heart rate were observed between the two exercise modalities (Table 3). A significant decrease in erythrocyte count (p = 0.037), leukocytes (p = 0.002), neutrophils (p = 0.000), eosinophils (p = 0.022), and lymphocytes (p = 0.003) was observed, with these changes being more pronounced in the ground exercise modality. Globulins (p = 0.002) and albumin (p = 0.031) showed significant differences, with higher values in the ground exercise group. Lactate concentration exhibited a notable difference between exercise types, being significantly higher in ground exercise (p = 0.000). Finally, maximum heart rate was significantly lower during swimming.

To our knowledge, no controlled comparison has been made between the immediate and long-term hematological, lactate, and cardiovascular adaptations from ground-based and aquatic exercise in the CPH. This study is the first to combine both acute physiological responses and training-induced changes from two different exercise methods within a standardized training program. In this context, the comparison of high-impact ground exercise and low-gravity aquatic exercise uncovered modality-specific hematological patterns that have not been reported before in the CPH.

Although this study contributes to sports medicine in CPH, its results should be viewed with consideration of certain methodological limitations. The small sample size and lack of a control group restrict the ability to draw causal conclusions. However, the repeated-measures design facilitated assessing physiological changes within the same individual over time. Exercise intensity was estimated using heart rate, a practical approach in field conditions but less accurate than direct gas exchange measurements. Additionally, while cortisol was measured as a stress marker, including other physiological indicators could have provided a more detailed understanding of responses to the training environment. These findings highlight the need for future research with larger sample sizes, controlled conditions, and increased statistical power to generalize the results to other populations.

The literature on the immediate effects of ground-based and aquatic exercise in various equine breeds and sport modalities is extensive [14, 15]; however, research on long-term hematological adaptations from structured training programs remains limited [16]. Specifically, information on the CPH is scarce, despite it being considered a high-performance breed due to the physical demands of its activity, which is often conducted under empirical training schemes [8, 9]. This study provides evidence on the immediate physiological changes caused by both ground and aquatic exercise, as well as on long-term hematological adaptations in CPH subjected to a ground-based training program, followed by a four-month rest period and then an aquatic training regimen.

Regarding exercise responses, the hemo-leukogram variables evaluated in this study do not match the patterns reported in previous studies conducted in CPH [10, 11]. However, the results show that hematological changes are affected not only by exercise intensity and duration [17, 18], but also by the type or modality of exercise. Although significant hematological changes were observed in both training modalities, they were more pronounced during ground-based exercise, likely due to higher metabolic demand, which triggered a stronger hematological response.

Specifically, a significant increase in red blood cell count and hematocrit was observed, indicating notable hemoconcentration linked to plasma volume loss caused by sweating during exercise. This reflects a combined effect of physical exertion and dehydration. Additionally, these responses may have resulted from fluid redistribution among body compartments and could have been mediated by splanic contraction triggered by rapid adrenergic discharge and increased metabolic demands [2]. In this context, elevated hematocrit is a typical finding during high-intensity exercise because it causes a physiological stress response [10]. Therefore, ground-based exercise seems to demand more physiological effort compared to aquatic exercise, which benefits from buoyancy, hydrostatic pressure, and a lower cardiovascular workload.

Hemoglobin levels changed significantly between the pre- and post-exercise periods in the pool modality. The increase in hemoglobin during exercise indicates enhanced oxygen transport capacity, which is associated with predominantly aerobic metabolism typical of moderate to low-intensity exercise. This interpretation is supported by heart rate monitoring, which showed moderate intensity during pool exercise, compared to the higher demands observed during ground training. These findings suggest that aquatic exercise produces a less demanding physiological response due to the reduced mechanical load on the musculoskeletal system and the moderating effects of water on fluid loss, body temperature regulation, and consequently hemoconcentration [19, 20].

Plasma protein and albumin concentrations did not show significant changes, despite expectations of a post-exercise increase, especially during ground-based exercise where sweating is more intense, as reported in endurance-tested horses and ground-trained CPH [10, 21]. The type of exercise may have influenced this outcome due to changes in plasma volume, which directly affect these parameters. Moreover, plasma protein concentration is a multifactorial variable influenced by both intrinsic and extrinsic factors, including renal filtration rate, metabolic demand, neuroendocrine regulation, nutritional status, hydration balance, and animal health—factors that contribute to its homeostatic regulation [22].

In the hemo-leukogram, greater changes were observed after exercise in the ground modality, shown by a general increase in leukocyte count, although no significant differences were seen in leukocyte subtypes. In contrast, the pool modality did not show notable changes in this parameter. These findings suggest that exercise intensity affects the leukocyte response, as previously described, with muscle fatigue associated with marked leukocytosis driven by neutrophilia [23]. However, as noted in racehorses undergoing repeated training sessions [24], neutrophilia was not seen in this study, nor were fatigue levels enough to indicate a state of exhaustion. Therefore, the neutrophil-to-lymphocyte ratio (10:1), which has been proposed as an indicator of stress, overtraining, or physiological exhaustion in horses [25], was not measured.

The post-exercise increase in lymphocyte count has been extensively studied [26, 27] and is associated with the immune system’s response to physical activity. After exercise, leukocytosis indicates complex immune activation, with possible effects on health and athletic performance [28]. This change was observed only in the ground modality, which involved higher intensity, and, along with the observed rise in globulins, suggests an exercise-induced immune and inflammatory response. This interpretation is further supported by the detection of lipopolysaccharides in the blood of some horses in this group (data not yet published). However, some studies have reported no significant changes in lymphocyte or monocyte counts in horses undergoing intense treadmill exercise [23], indicating that these responses can vary depending on the type, duration, and intensity of exercise, as well as individual factors.

Many hemo-leukogram variables in both exercise types showed changes between pre- and post-exercise periods without reaching statistical significance. This may be due to the small number of animals and differences in the intensity and duration of protocols used in previous studies that did demonstrate significant changes [10]. Therefore, further research is needed to develop scientifically grounded training protocols tailored to the specific needs of CPH, as reported in other athletic breeds [29].

Comparison of post-exercise parameters between modalities revealed significant differences in most variables, confirming greater intensity and physiological effort during ground exercise. Meanwhile, other parameters showed no differences in this comparison but did differ from pre-exercise values within each modality, suggesting similar intensity between modalities at this stage, with differences in metabolism and cardiovascular impact, as indicated by increased lactate concentration and heart rate, markers of exercise intensity [1, 30–32]. More recently, the behavior of these two variables has been shown to be useful for estimating aerobic fitness in CPH without speed tests or invasive procedures [33]. These changes were also reported in a recent study comparing treadmills with and without water [34].

Regarding long-term adaptations from the training protocols, this study showed a significant decline in resting heart rate after both programs, indicating enhanced cardiovascular efficiency in horses trained with either method. In the ground training protocol, resting heart rate dropped from 87 bpm to 71 bpm, while in the pool protocol, it decreased from 96 bpm to 83 bpm. This decline aligns with the physiological changes typical of aerobic conditioning, where increased vagal tone and better myocardial function lead to lower basal heart rate, as noted in other horse breeds [35–37].

Although both training protocols were effective, the horses exhibited higher resting heart rates before pool training, possibly because of the 4-month detraining period, which suggests a partial loss of previously acquired cardiovascular adaptations. Nonetheless, their ability to respond to a new training stimulus remained intact, as shown by the significant decrease observed at the end of the second protocol. This underscores the value of aquatic exercise as an effective method for restoring cardiovascular fitness in horses that are either previously trained or detrained, without causing musculoskeletal overload.

Moreover, it should be noted that pre-exercise heart rate may be part of the anticipatory response to effort, which is potentially influenced by stress. In this context, the higher pre-exercise heart rate observed in the pool could be linked to the increased alertness caused by water entry, a potentially more stressful stimulus for the horse [2], despite the adaptation period these animals had. However, no heart rate monitoring or electrocardiograms were conducted during swimming, which could have provided additional insights into underwater cardiac function [37]. Additionally, although the shape of the pool may affect exercise development [38], this factor should also be examined in future research.