Introduction

Front. Vet. Sci.

Frontiers in Veterinary Science

Front. Vet. Sci.

2297-1769

Frontiers Media S.A.

10.3389/fvets.2022.894146

Veterinary Science

Original Research

Is Continuous Monitoring of Skin Surface Temperature a Reliable Proxy to Assess the Thermoregulatory Response in Endurance Horses During Field Exercise?

Verdegaal

Elisabeth-Lidwien J. M. M.

¹ ² ^* Howarth

Gordon S.

¹ ³ McWhorter

Todd J.

¹ Delesalle

Catherine J. G.

¹Equine Health and Performance Centre, University of Adelaide, Adelaide, SA, Australia ²Department of Translational Physiology, Infectiology and Public Health, Research Group of Comparative Physiology, Faculty of Veterinary Medicine, Ghent University, Merelbeke, Belgium ³Gastroenterology Department, Women's and Children's Hospital, Adelaide, SA, Australia

Edited by: Severiano Silva, Universidade de Trás-os-Montes e Alto, Portugal

Reviewed by: Mariana Almeida, University of Trás-os-Montes and Alto Douro, Portugal; Daniel Mota-Rojas, Metropolitan Autonomous University, Mexico

*Correspondence: Elisabeth-Lidwien J. M. M. Verdegaal lidwien.verdegaal@adelaide.edu.au

This article was submitted to Animal Behavior and Welfare, a section of the journal Frontiers in Veterinary Science

27 05 2022

2022

894146

11 03 2022 25 04 2022

2022

Verdegaal, Howarth, McWhorter and Delesalle

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

Hyperthermia is a performance and welfare issue for exercising horses. The thermoregulatory stressors associated with exercise have typically been estimated by responses in the laboratory. However, monitoring surface skin temperature (T_sk) coincident with core temperature (T_c) has not previously been investigated in horses exercising in the field. We investigated the suitability of monitoring surface T_sk as a metric of the thermoregulatory response, and simultaneously investigated its relationship with T_c using gastrointestinal (GI) temperature. We evaluated T_sk in 13 endurance horses competing during four endurance rides over 40 km (n = 1) or a total of 80 km (n = 12) distance. Following each 40-km loop, the horses were rested for 60 min. T_sk and T_c were continuously recorded every 15 s by an infrared thermistor sensor located in a modified belt and by telemetric GI pill, respectively, and expressed as mean ± SD. The net area under the curve (AUC) was calculated to estimate the thermoregulatory response to the thermal load of T_sk over time (°C × minutes) using the trapezoidal method. The relationship between T_sk and T_c was assessed using scatterplots, paired t-test or generalized linear model ANOVA (delta T_sk) (n = 8). Ambient temperature ranged from 6.7°C to 18.4°C. No relationship was found between T_sk and T_c profiles during exercise and recovery periods, and no significant difference between delta T_sk results was detected when comparing exercise and rest. However, time to maximum T_sk (67 min) was significantly reduced compared to T_c (139 min) (p = 0.0004) with a significantly lesser maximum T_sk (30.3°C) than T_c (39°C) (p = 0.0002) during exercise. Net AUC T_sk was 1,164 ± 1,448 and −305 ± 388°C × minutes during periods of exercise and recovery, respectively. We conclude that T_sk monitoring does not provide a reliable proxy for the thermoregulatory response and horse welfare, most probably because many factors can modulate T_sk without directly affecting T_c. Those factors, such as weather conditions, applicable to all field studies can influence the results of T_sk in endurance horses. The study also reveals important inter-individual differences in T_sk and T_c time profiles, emphasizing the importance of an individualized model of temperature monitoring.

skin surface temperature thermoregulation metabolic heat endurance exercise performance competition hyperthermia

Introduction

In the face of climate change, hyperthermia and heat stress have become increasingly challenging issues for a wide array of equine sports disciplines, especially during field competitions (1, 2). An increase in core body temperature (T_c) leading to hyperthermia may cause widespread cytotoxicity as a direct effect of heat, while indirect effects related to decreased cardiac output cause neural and intestinal ischemia. If unchecked, these systemic inflammation processes eventually lead to exertional heat illness (EHI) (3–5). The clinical manifestations of EHI include neurological signs, varying from irritability, depression, ataxia, collapse, and may further progress to exertional heat stroke (EHS) with multi-organ dysfunction and death (3–5). In human athletes, EHS is among the top three causes of sudden death, and, in summer, it is the number one cause of athlete death in the USA (3). Similarly, both EHI and EHS are problematic conditions in equine athletes (6). The prevalence of metabolic disorders in, for example, endurance horses, triggered by thermoregulatory-induced physiological feedback failure and exhaustion, ranges from 4.2% to 15% (7–14). Recently, the prevalence of EHI in racehorses has been reported: two studies from Japan state prevalence of 0.09% during summer with a clear increase over the past few years (15, 16). Moreover, a study in Eastern Australia focused on selected EHI cases post-exercise at the racetrack and suggested an EHI incidence of up to 9.5% during hot summer months (17). The latter study used the four severity levels of EHI reported by Brownlow et al. (4) and concluded that 96% of horses could be categorized as Level 1. This suggests that a low level and thus discrete EHI cases may have been overlooked in the past.

Environmental conditions are the dominant risk factor in heat stress events, and EHI cases are expected to further increase in prevalence due to global warming (1, 4, 15–17). That worrying reality drives the ongoing efforts of research groups worldwide to develop reliable approaches to monitor and safeguard thermoregulatory wellbeing in horses (2, 18–21). We have previously reported on the continuous monitoring of the thermoregulatory response in endurance and trotter horses using a telemetric gastrointestinal (GI) pill (2). Briefly, the GI temperature pill is a non-invasive method to monitor T_c during exercise in the field. The GI pill was administered the night before the endurance competition to allow recording of a large temperature data set to establish T_c profiles of exercising horses. The data were telemetrically transferred to a device located in a belt under the saddle girth and supported continuous real-time monitoring. It was demonstrated that the GI telemetry pill is a reliable method to monitor T_c and assess the individual dynamic thermal response in exercising horses in the field. The study revealed important inter-individual differences in T_c time profiles, despite the horses performing the same exercise protocol. This finding emphasizes the importance of an individualized model of temperature monitoring. It can be concluded that the continuous monitoring approach allows for intervention at the early stages of heat accumulation and the possibility to take prompt and effective preventative measures.

Another elegant approach to monitoring the thermoregulatory response in exercising horses in the field could be continuous monitoring of surface skin temperature (T_sk) as a reliable proxy for monitoring thermoregulatory wellbeing. Advantages of using tools, such as infrared thermography (IRT), to assess T_sk include the non-invasive nature and easy collection of temperature data (22–24). Notably, most exercise studies use such monitoring methods to assess T_sk only pre- or post-exercise, not during exercise. To illustrate, a recent systematic review of T_sk studies performed in human endurance athletes has reported that only a few studies involved the continuous monitoring of T_sk; only two out of the 45 exercise studies monitored T_sk every 30 s, one study every 180 s, and four studies every 5 min (24). Also, these human studies highlight the existence of essential inter-individual differences with respect to the T_sk response time profile (24). Important inter-individual T_sk differences have been reported in all mammals (23, 25). For example, an equine study comparing T_sk in 21 horses at rest revealed significant inter-individual differences (26). Most existing equine thermoregulatory exercise studies (n = 12 studies) involving T_sk have focused on single point post-exercise measurements using a handheld IRT camera, leaving both the pre-exercise and intra-exercise periods out of their scope (21, 22, 27–37). An overview of all T_sk studies on horses at rest and during exercise is presented in Supplementary Table 1.

To date, only a few equine studies have focused on continuous monitoring of T_sk during exercise and recovery (n = 7, Supplementary Table 1). All were conducted in laboratory conditions, for example, on a treadmill (37–42). Only one study has recorded T_sk continuously during a short bout of field exercise of 4.5-min duration to investigate the effect of pre-exercise cooling in 10 horses (20). The study recorded surface T_sk using a microchip (i-Button®) attached to the skin with removable glue and simultaneously monitored rectal temperature (T_re) using a rectal probe; both T_sk and T_re reduced over time (3°C and 0.3°C, respectively).

From a physiological standpoint, it is essential to appreciate that a time lag exists between exercise-induced metabolic heat (MH) output and T_c evolvement. The T_c is subsequently translated into an additional temperature time-lag evolvement, expressed at several different anatomical locations, such as the rectum, the muscular compartment, and the skin surface, whether or not additionally complicated by environmental factors, such as hot and humid weather (1, 25, 39, 43–48). Most importantly, T_re evolvement has been reported to significantly lag behind the T_c both during and after exercise (19, 38, 49), which renders the T_re less suitable as a “whistle blower” for thermoregulatory instability. In our previous studies, we demonstrated that GI temperature is a more reliable proxy for the thermoregulatory response and T_c when compared to T_re, and that continuously monitoring GI temperature evolvement demonstrated how the equine body copes with exercise, challenging the thermoregulatory system (2, 19). Endurance horses, for example, reached their mean maximum T_c (39.0 ± 0.4°C) during exercise at 75% of completion of exercise, and T_c returned to the baseline within 60 min into recovery (2). However, the mean T_c was still 38.8 ± 0.4°C at a heart rate (HR) of 60 bpm, which currently governs “fit-to-continue” competition decisions (50), thus questioning the use of HR values to make such important decisions. However, contrary to this finding, trotter horses reached a comparable mean maximum T_c (38.8 ± 0.5°C) during recovery. Moreover, in 30% of trotters, T_c was still > 39°C at the end of the recovery period (40 ± 32 min), following exercise in a cool environment, findings that may have post-exercise management implications.

To identify a reliable proxy for thermoregulatory response in the field, a solid correlation must exist between that specific proxy and T_c evolvement, despite the existence of a time lag (24, 47). However, currently, very few equine studies have involved the simultaneous continuous monitoring of T_c (either using carotid artery temperature, or a GI pill, or the T_re), together with an additional temperature monitoring device during field exercise (19, 20, 51). On the other hand, with the ongoing development of new wearables and sensors, there are an increasing number of exercise studies investigating continuous T_sk, monitoring wearables (20, 37–40, 42, 52) (Supplementary Table 1 in Supplementary Material). These devices all provide data output, although the physiological meaning of these data is not always clear.

The MH produced during exercise needs to be dissipated from the horse to the surrounding environment through four main pathways, namely, radiation, conduction, convection, and evaporation, the last being the most essential and pivotal pathway in horses (23, 48, 53–56). Evaporation from the body surface is mainly achieved by increased blood flow, cutaneous vasodilation followed by evaporation of sweat from the skin (70–85% of the MH load) (25, 48, 54, 56–59). Heat loss by evaporation can be enhanced by cooling techniques (35, 40, 43). When focusing on T_sk as a temperature monitoring method, it is vital to keep in mind that all these pathways to dissipate MH to the environment may influence the T_sk data output.

Monitoring T_sk simultaneously with T_c using the GI temperature pill during field exercise has not yet been investigated. The relationship between T_sk and T_c is not well understood due to physiological, endocrine, or vasomotor influences on both temperatures (25, 56, 59). Some studies have tried to correlate both T_c and T_sk (39–42). The current study aimed to evaluate the usefulness of continuous monitoring of T_sk by means of a surface IR sensor device as a proxy for the thermoregulatory response. For this purpose, the T_sk relationship with T_c was investigated by simultaneous and continuous telemetric measurements during real-time field competitions under cool weather conditions. Endurance horses were equipped with several non-invasive telemetric monitoring devices—a T_sk device positioned in a girth belt, an orally administered GI pill (T_c), a global positioning system (GPS), and an HR monitor.

Materials and Methods Horses

Thirteen mainly Arabian (n = 10) endurance horses participated in the study: 7 geldings; 6 mares; age, 9.5 ± 2.8 years; body mass (BM), 479 ± 68 kg; body condition scores varied from 2 to 3 out of 5. Two cross-Arabians and one crossbred (quarter horse—thoroughbred) were also involved (Table 1). Coat color included bay (n = 3), chestnut (n = 6), and gray (n = 4), and the color was scored as follows: dark (bay and chestnut, n = 9) compared to light (gray) (Table 1). Relevant rider and horse performance history and the Bureau of Meteorology (B.O.M.) (60) station output information were recorded (Table 1; Supplementary Material S1). All the horses were deemed to be fit and healthy based on the veterinary inspection conducted before the competition and following each 40-km loop according to AERA riding rules (50). The horses were sourced on a voluntary basis through the South Australian Endurance Riders Association (S.A.E.R.A.), and all the owners signed a written consent form. The study was approved by the University of Adelaide Animal Ethics Committee (project No. S-2011-224).

Table 1

Study population characteristics and monitoring devices.

Horse number	Sex	Age (y)	Breed	Body mass (kg)	Coat color	Distance (km)	GI Pill Y/N	GPS/HR Y/N	B.O.M. (°C) (T_a), min - max	Sweating score post-exercise, 1 tot 3
1	G	11	Arab	669	Gr	80	Y	Y	13–26	2
2	G	9	Arab	484	B	80	Y	Y	13–26	1
3	M	13	Arab	426	C	80	Y	Y	6–19	2
4	M	7	QH xTB	470	C	80	Y	Y	6–19	1
5	M	11	Arab	450	C	80	Y	Y	6–19	1
6	M	8	Arab	370	Gr	80	Y	Y	6–19	2
7	M	9	Arab	450	C	80	Y	∧	3–22	2
8	G	11	Arab	470	Gr	80	Y	Y	3–22	3
9	G	7	QH	490	C	80	–	Y	7–13	3
10	G	10	Arab xTB	484	C	80	–	Y	7–13	3
11	G	5	Arab	458	B	40	–	Y	7–13	1
12	G	7	Arab	525	B	80	–	Y	3–22	1
13	M	15	Arab	480	Gr	80	–	∧∧	3–22	2

Horses 1-13: 13 endurance horses: G (n = 7), M (n = 6). Arabian, including part-Arabian horses, QH, quarter horse; TB, thoroughbred; x, crossbred; G, gelding; M, mare; Gr, grey; C, chestnut; B, bay. The riders' and horses' performance history includes: age start, indicating age (years) when the horse started competing; horse experience, indicating number of years active in competition (40 km or more); GI pill, gastrointestinal pill; GPS, global positioning system; HR, heart rate monitor (Polar); B.O.M, Bureau of Meteorology; the local station closest to the location of exercise at varying km distances from the actual event (in total, 4 endurance locations, distance ranged from 5.3 to 53 km; -, no; Y, yes; ^∧indicates HR only; ^∧∧indicates second 40 km only.

Study Design

The horses competed over distances of 40 km (n = 1), 80 km (n = 10) or 100 km (n = 2), with each 40-km loop followed by a 60-min recovery period. The endurance horses exercised at four different locations with altitudes ranging from 4 to 462 meters above sea level. Following each 40-km exercise loop, the sweating response was graded, scoring from 1 to 3 (1: mild wet and white foam areas around head, neck, saddle, and inside hindlimbs, 2: moderate dripping sweat from the body, 3: extensive dripping sweat from the body; Table 1). In addition, the horses were immediately cooled down for average an duration of 10 min by pouring buckets of tap water (estimated average, 20°C) over their bodies and subsequently scraping it off. Following each loop, a recovery period of 60 min was allowed during which inspection of the horses for “fitness to continue” was performed, including checking for the presence of an HR below 60 beats per minute (bpm) by independent endurance veterinarians under the regulations of the Australian Endurance Riding Association (50). Horses were allowed to drink water and eat hay ad libitum during the 60-min rest period in a shaded area.

Simultaneous Continuous Monitoring of Skin Temperature (T<italic><italic>SK</italic></italic>) (°C) and Core GI Temperature (T<italic><italic>C</italic></italic>) (°C)

The T_sk (°C) was continuously recorded using an infrared (IR) sensor measuring 78 × 53 mm located in the Sensor Electronics Module (Figure 1B) (SEM, EQ02 Equivital data Logger®, Hidalgo, UK), with a 0–60°C temperature range, an emissivity of 1 and ± 0.3°C accuracy according to the manufacturer's specifications. The SEM device was located ventrally in a pocket of a modified Equivital Sensor Belt® fitted around the saddle girth (Figure 1C). The GI temperature (T_c) (°C) was continuously telemetrically recorded using the ingestible GI pill (n = 8) (Figure 1A) as previously described (T_c data are to be found in the Supplementary Material S1). The T_sk and T_c data were recorded every 15 s and uploaded and processed in the Equivital Software Manager®.

Figure 1

T_sk and T_c monitoring equipment: Jonah gastrointestinal temperature pill (A); an external receiver Equivital® Sensor Electronics Module (SEM) with an infrared sensor to measure T_sk (B); a modified belt for use on horses with the GPS Garmin®Watch and Polar® electrodes (identified by the blue pointer), with the red pointer, indicating the SEM device, including the T_sk thermistor position (C).

Monitoring of Traveled Distance, Speed, and Heart Rate Over Time

For each horse, the distance traveled and speed achieved were recorded telemetrically using GPS monitoring equipment (Garmin Forerunner 910XT GPS Watch® Garmin Ltd., Schaffhausen, Switzerland) attached to the gullet of the saddle (Figure 1C). In addition, the HR was recorded by the Garmin Watch using Polar electrodes (Polar Electro®, Kempele, Finland) (Figure 1C) (61). The GPS and HR data were recorded every second and uploaded from the Garmin watch to the Garmin Connect and processed in the Garmin Training Centre.¹

Ambient Environment

Horses exercised under varying degrees of solar radiation during the Australian winter months (June–August). On each data collection day, the ambient temperature (T_a, °C) and relative humidity (RH, %) were recorded continuously every 30 s in a shaded section of the rest area using a data logger device (OnsetHOBO Pro V2 logger temp/RH U23-00®, Onset Computer Corporation, Bourne, Maine, USA). In addition, T_a data were obtained from the nearest B.O.M. weather station, presented in Table 1. The estimated wet bulb globe temperature (WBGT) was derived from a WBGT chart (B.O.M).

Data Processing

Recordings of each exercise period of 40 km and each recovery period following that exercise loop were processed. The net AUC (the baseline set at rest T_sk and T_c) was calculated using the trapezoidal method of T_sk (and T_c) over time expressed as °C × minutes. The net AUC was summated to present the cumulative T_sk – time distribution (62). The net AUC T_sk provided an estimate of the dynamic thermal response to the thermal load on the skin. This thermal load on the skin during exercise and the recovery included the T_c and T_a together with solar radiation.

Statistical Analysis

All data are presented as mean ± SD (range). Comparison and correlation analyses were performed using IBM SPSS Statistics 26.0 software or GraphPad Prism version 9.3.0 for MacOS, GraphPad Software, San Diego, California USA². Different approaches were taken to evaluate the potential of the T_sk data as a reliable proxy to assess the thermoregulatory response. The relationship between T_sk and T_c was assessed using scatterplots (8 horses each performing two subsequent 40-km loops). In addition, maximum T_sk and T_c and the time to reach maximum T_sk and T_c were compared using the paired t-test. Delta T_sk during exercise and recovery periods was compared. The association between T_sk and T_c at different points in time and the association with HR or coat color were analyzed using a general linear model ANOVA (when no significant effects of horse identity and treatment interaction were indicated and subsequently removed using backward model selection). Statistical significance was set at α < 0.05.

Results

All the horses completed their exercise trials without any adverse occurences. The Equivital belt became dislodged in Horse 1 at the end of the first 40-km loop, causing T_sk and T_c data loss. As a result, additional modifications were applied to the belt for the subsequent recordings by fitting sturdy straps sandwiched into the belt to stabilize the girth position (Figure 1C). During recovery after the first loop, T_sk was not recorded in Horses 1 and 2 due to the owners' premature removal of the belt. The sweating response varied from 1 to 3 out of a score of 3 for all the horses (Table 1). The T_c was recorded in 8 horses over 80 km (previously published, Supplementary Material S1) (2).

Environmental Field Conditions

The T_a and RH were successfully recorded between 5.00 a.m. and 3.00 p.m. on all occasions. The T_a was relatively cool with a mean minimum of 6.7 ±.4°C and mean maximum of 18.4 ± 2.9°C (B.O.M.) (Table 1). More specifically, the T_a on the four separate days of endurance exercise showed a minimum value of 13.4, 6.3, 2.8, and 6.6°C, respectively, and a maximum value of 26.3, 19.0, 22.0, and 18.8°C, respectively (HOBO data). The minimum RH ranged from 47.1% to 61.7% to a maximum of 84.8–100% value. Overall mean calculated values were 15.3°C (T_a) and 75.6% (RH), respectively, while the approximate WBGT was <20°C. In summary, all the endurance horses competed in a cool environment.

Speed and Heart Rate Data

All the horses executed their endurance competition at a mean speed of 14.0 ± 1.4 km h⁻¹ over the first 40 km (n = 11) and 14.2 ± 2.1 km h⁻¹ over the second 40-km (n = 11) loop, with a mean HR of 114 ± 13 bpm. An overview of recorded speeds and HR data for individual horses can be found in Supplementary Material S1.

Individual T<italic><italic>sk</italic></italic> and T<italic><italic>c</italic></italic> Recordings During Endurance Exercise Over Time

An overview of the simultaneously recorded individual T_sk (°C) and T_c (°C) time profiles is provided for all the horses in Figure 2. All individual T_sk parameters, their respective descriptive analysis, and specific T_sk points in time during the 40-km endurance loops are presented in Table 2.

Figure 2

Scatterplots of T_sk, skin temperature (°C, identified as blue) and T_c, gastrointestinal temperature (°C, identified as black) (left y-axis) per subsequent exercise loop of 40 km (gray blocks) (h, hours, x-axis) in endurance horses, Horses 1–13; dotted blocks identify blocks of no data recording; after each exercise loop, the horses were cooled down by pouring buckets of water over their bodies, followed by scraping the water off the bodies for a period of ~10 min.

Table 2

T_sk (°C) parameters during 40-km exercise (or extra 20 km in 2 horses exercising over a total of 100 km) and recovery.

Horse	Distance 40 km – 1^st or 2^nd	Net AUC T_sk (°C x min) exercise 40 km	Net AUC T_sk (°C x min) recovery 40 km	Base T_sk (°C) at start of exercise	Mean ±SD T_sk (°C) exercise	Min - max T_sk (°C) 40 km exercise	Delta T_sk (°C) exercise	T_sk (°C) at start of recovery (end exercise)	Delta T_sk (°C) first 10 min recovery	Delta T_sk (°C) recovery	Delta T_sk (°C) end exercise to end recovery/min	Mean ±SD T_sk (°C) recovery	Min –max T_sk (°C) recovery	Time to max T_sk exercise (min)	Time to min T_sk in recovery (min)	T_sk (°C) at end of recovery	T_sk (°C) at end of recovery
1	1st	−743.7	–	28.5	20.7 ± 2.4	17.8–28.4	10.6	21.9	–	–	–	–	–	0	–	23.2	23.2
1	2nd	587.3	−0.78	23.2	26.4 ± 2.6	21.2–31	10.8	24	−0.3	3.7	0.06	23.7 ± 1.1	22.1–25.8	74	21	26.4	22.8
2	1st	−1,328.6	–	26.4	20.9 ± 7.2	15.3–36.2	20.9	24.2	−1	6.6	–	23.3 ± 1.7	–	45	–	24.7	24.3
2	2nd	−515.3	54.9	24.3	22.6 ± 1.4	20.4–26.1	5.7	22.3	−1	5.1	0.09	23.3 ± 1.3	21–26.1	76	20	26	25.9
3	1st	838.1	−624.6	29.4	33.5 ± 0.8	29.4–34.2	4.8	32.5	−9.3	13.9	0.23	21.8 ± 3.5	18.8–32.7	39	29	22.1	22.5
3	2nd	2,820.5	−597.3	22.5	33.9 ± 2.0	22.5–35.4	12.9	34.8	−6.4	11.5	0.19	26.4 ± 2.7	23.4–34.9	72	23	24.6	24.4
4	1st	1,382.4	−85.5	27.7	21.0 ± 1.3	27.8–35.4	7.7	34.2	−12.3	18.4	0.31	21.8 ± 3.5	18.8–32.7	110	28	18.2	18.2
4	2nd	1,422.9	169.2	18.2	24.0 ± 3.2	15.8–30.5	14.7	24.7	−0.6	7.5	0.13	22.1 ± 1.8	19.3–26.5	79	16	26.8	26.8
4	Extra 20	(1,051.1)	(–468.1)	26.9	34.4 ± 1.6	26.9–35.5	8.5	34.7	–8.6	17.2	0.29	22.8 ± 4.7	17.6–34.7	–	27	19.5	19.6
5	1st	570.8	−635.5	27.6	30.7 ± 1.5	26.1–32.5	6.4	28.3	−10	15.4	0.26	16.7 ± 3.5	14.3–29.7	56	26	16.2	16.1
5	2nd	2965.8	−997.3	16.1	31.9 ± 3.8	15.8–34	18.2	33.5	−12.9	16.9	0.28	19.2 ± 3.3	16.6–33.4	79	16	18.2	18.2
6	1st	2,736.5	−408.7	13.7	29.0 ± 4.0	12.4–32	19.6	29.7	−0.8	14.5	0.24	22.8 ± 5.3	15.3–29.5	74	12	19.3	19.5
6	2nd	995.4	251.0	19.5	26.2 ± 2.8	19–30.5	11.5	24.6	0.8	6.5	0.11	23.2 ± 2.2	20.1–26.6	55	19	20.1	20.1
6	Extra 20	(809.8)	(–427.0)	20.1	29.0 ± 4.1	18.1–32.7	14.5	31	1.2	15.1	0.25	23.8 ± 5.9	17.2–32.3	36	46	20.1	20.1
7	1st	960.6	−3	11	17.7 ± 6.3	10–32	22	13.3	−0.7	15	0.25	16.4 ± 3.9	12.1–27.1	26	10	18.9	19.2
7	2nd	80.4	97.5	19.2	20.4 ± 1.0	18.2–23.9	5.7	19.7	0.8	9.5	0.16	20.8 ± 1.2	15.8–25.3	64	18	20.2	20.2
8	1st	4,170.6	−547.4	10	30.4 ± 4.1	10–34.5	24.5	25.9	1	11.4	0.19	19.3 ± 3.1	15.8–34.5	75	10	20.3	20.2
8	2nd	884.2	110.8	20.2	24.8 ± 2.1	19.2–29.4	10.2	26.7	−1	12.2	0.20	23.8 ± 3.1	15–27.2	150	19	18.3	18.8
9	1st	535.8	133.2	18.8	19.8 ± 1.7	15.6–23.5	7.8	23.4	−2.8	8.9	0.15	18.0 ± 2.6	14.6–23.4	179	16	22.6	22.6
9	2nd	−68.3	−172.4	22.6	22.3 ± 0.9	18.7–24.4	4.7	23.4	−3.2	10.4	0.17	17.6 ± 2.8	13–32.3	141	20	14.3	14.5
10	1st	934	−628.4	29.8	34.2 ± 1.1	29.9–36	6	34.7	0.3	17	0.28	23.9 ± 5.9	17.7–35	44	30	26.1	26.2
10	2nd	1,825.3	−852.4	26.2	34.7 ± 1.5	26.2–36.1	10	34.8	−15.3	17	0.28	20.4 ± 2.3	17.5–32.3	63	26	21.2	20.9
11	1st	1,691.6	−944.0	26.5	34.3 ± 1.4	26.6–35.5	9	35.4	−15.9	21	0.35	18.5 ± 5.7	14.5–35.4	179	27	17.7	17.7
12	1st	4,748.6	−597.2	32.3	30.9 ± 7.6	9.5–37.5	29	35.3	−10.4	19	0.32	22.8 ± 3.5	17.9–35	152	26	28.4	28.7
12	2nd	1,356.4	−524.7	28.7	37.4 ± 1.7	28.7–38.7	10	38.1	−19.4	22.2	0.37	19.6 ± 4.9	16.3–38.2	90	29	16.8	16.8
13	1st	−204.5	−169.3	29.5	24.2 ± 4.4	17.9–33.8	16	25.5	4.6	17	0.28	22.6 ± 4.6	15.5–32.8	84	18	26.8	26.4
13	2nd	453.9	−45.1	26.4	28.6 ± 2.5	21.6–35.2	14	27.9	2.6	11	0.18	27.1 ± 2.2	22.6–33.4	196	22	24.1	26.2

Data are presented as mean ± SD. T_sk, skin temperature; AUC, area under the curve; min, minutes; min-max, minimum to maximum; italic indicates an extra loop of 20-km exercise (total, 100 km, n = 2 horses); delta (°C change), T_sk change during exercise, and recovery periods, including the first 10-min recovery period; – indicates no data collected/not available.

Overall T<italic><italic>sk</italic></italic> Profiles and Comparison to T<italic><italic>c</italic></italic>

The overall T_sk profiles during endurance exercise and recovery and their associated parameters are presented in Table 3, showing a mean time to maximum T_sk of 88 ± 51 min (n = 13). The mean maximum T_sk during exercise was 32.4 ± 4.3°C, and the mean minimum T_sk during recovery was 17.3 ± 3.1°C (n = 13). The mean overall response of T_sk was 1,164 ± 1,448°C × minutes for each 40-km exercise period. During recovery, the T_sk response was −305 ± 388°C × minutes (Table 3). The T_sk and T_c profiles over time were compared in the 8 horses, and no relationship was found (Figure 2).

Table 3

Overall T_sk and T_c variables during exercise and recovery of endurance exercise in a cool environment.

Variables	Endurance 40 km n = 13(n = 8^*)
Duration (minutes) exercise	198 ± 63
Duration (minutes) recovery	60
T_sk (°C) overall	27.8 ± 5.6 (17.71–37.37)
Base T_sk (°C) (at start-of-exercise)	23.1 ± 6.1 (10–32.3)
Min T_sk (°C) exercise	19.8 ± 6.2 (9.5–29.9)
Max T_sk (°C) exercise	32.3 ± 4.3 (23.5–38.7); 31.6 ± 3.5°C^*
T_sk (°C) exercise	27.2 ± 5.7 (17.7–37.4)
Time to max T_sk exercise (minutes)^*	88 ± 51 (0–196); 67 ± 34^*
Delta T_sk (°C) exercise	12.5 ± 6.6 (4.7–29); 12.9 ± 6.4^*
Net AUC T_sk exercise (°C × minutes)	1,164 ± 1,448 (−1,329 to 4,749); 1,114 ± 1,469^*
T_sk (°C) at end-of-exercise	28.0 ± 6.1 (13.3–38.1)
Min T_sk (°C) recovery	17.3 ± 3.1 (12.1–23.4); 17.7 ± 3.3^*
Max T_sk (°C) recovery	30.9 ± 4.1 (23.4–38.2)
T_sk (°C) recovery	21.5 ± 2.8 (16.4–27.1
Delta T_sk (°C) recovery	13.0 ± 5.1 (3.7–22.2)
Delta T_sk (°C) first 10 min recovery	−4.7 ± 6.7 (−19.4 to 4.6)
Delta T_sk first 10 min recovery/minute (°C/min)	−0.5 ± 0.7 (−1.9 to 0.5)
Number of horses T_sk > 39°C^**	None
T_sk (°C) at end-of-recovery 40 km (n = 25)	21.8 ± 3.8 (14.5–28.7)
Number 40 km periods T_sk returned to base T_sk at the end of 60 min recovery	14/25
Net AUC T_sk recovery (°C × minutes)	−305 ± 388 (−997 to 251);−230 ± 392^*

Data are presented as overall mean ± SD. T_sk, skin temperature; AUC, area under the curve; T_c, GI pill temperature; max T_c or T_sk, maximum T_c or T_sk; n, number identified only if different 40-km exercise periods.

Indicates a total of 8 horses (comparison T_sk to T_c in 8 horses; total, 16-x-40-km periods);

T_sk > than 39°C based on (21).

Different T_sk and T_c points in time were compared to assess associations. Interestingly, the only significant correlation found was between the T_sk (°C) at the end-of-exercise period and the T_c (°C) at the end-of-recovery period (F_1,14 = 5.519, p = 0.034). More precisely, a higher T_sk at the end-of-exercise period was associated with a lower T_c at the end-of-recovery period. The additional analyses revealed no significant correlations between T_sk (°C) and T_c (°C), including no correlation between T_sk at the start-of-exercise period (baseline T_sk) and the maximum T_c (F_1,14 = 0.127, p = 0.727). The study could not identify a significant effect of time to maximum T_sk (67 min) during exercise on the maximum T_c (39°C) (F_1,14 = 0.001, p = 0.978, n = 8). On all occasions, peak T_c values (39°C) were significantly greater than peak T_sk values (32°C) (p = 0.0002) (Figure 3A). In addition, in all cases, there was a significantly shorter time to maximum T_sk (88 min) compared to the time to maximum T_c (139 min) (p = 0.0004) (Figure 3B).

Figure 3

A boxplot diagram depicting T_c and T_sk, respectively, in endurance horses. Mean (a horizontal line) ± SD (error bars) of individual maximum T_c (°C, n = 8) and maximum T_sk (°C, n = 8) values. There was no significant association between T_sk and T_c; no significant difference between delta T_sk exercise and recovery (p = 0.41); a significantly greater maximum T_c when compared to maximum T_sk (p = 0.0002) (A); and a significantly increased time to maximum T_c than T_sk (p = 0.0004) (B); *indicates a significant difference.

Delta T_sk data were compared, and, when comparing exercise and recovery periods, no significant difference could be found between the two delta T_sk (°C) (mean 13°C and 11°C, respectively) (p = 0.41) (Figure 3A). The delta T_sk was greater when compared to delta T_c on all occasions (Figure 3A). Also, delta T_sk and T_c points in time were compared to assess associations. A correlation between the delta T_sk during cooling in the first 10 min of the recovery period and the T_c at the end of the recovery period was not identified (F_1,13 = 1.348, p = 0.266).

Additionally, there were no significant effects of coat color on the maximum T_c (F_1,14 = 0.010, p = 0.920) nor on the maximum T_sk (F_1,22 = 0.015, p = 0.904). Furthermore, coat color was not associated either with delta T_sk during exercise (12.5 ± 6.6) (F_1,22 = 1.098, p = 0.306) or delta T_sk during recovery (13.0 ± 5.1) (F_1,21 = 1.246, p = 0.277).

Evaluation of Heart Rate in Relation to T<italic><italic>sk</italic></italic>

Additional analysis to evaluate the relationship between the T_sk and HR < 60 bpm revealed no significant correlation between the T_sk at the end of exercise and the duration to HR < 60 bpm (F_1,13 = 4.019, p = 0.066). In conclusion, the study did not identify an association between HR recovery and the recorded T_sk during endurance exercise and recovery.

Discussion

This is the first study to simultaneously monitor both T_sk and T_c continuously by means of several telemetric temperature recording devices on exercising horses in the field. The thermal sensors functioned correctly throughout the study, hence, there was little to no data loss. Consistent with our previous field study (2), the current study confirmed a substantial inter-individual variability in the T_sk time profiles despite execution of the same exercise protocol. Similar findings have been reported in human athlete studies and underline the physiological complexity of the mammalian thermoregulatory response, which is governed by a plethora of individually intrinsic variables (63–66).

With respect to T_sk and T_c monitoring in the current study, there was a lack of correlation between continuous monitoring of T_c and T_sk. Nevertheless, the association of a higher single point T_sk at the end-of-exercise period with a lower T_c at the end-of-recovery period is an interesting finding in the current study.

Up until now, a limited number of studies, almost all of which were treadmill based, monitored the T_sk continuously in the horses during the exercise and compared the T_sk to T_c (38–40, 42, 52). Only one recent field study has been performed, monitoring T_sk and T_re (20). The treadmill studies confirmed the lack of correlation between T_sk and T_c. For example, two submaximal exercise studies using arterial blood temperature compared the effect of different environments on thorax surface T_sk and showed T_sk was different from T_c (38, 42). Two high-intensity studies reported tail surface T_sk responses to cooling methods and acclimation, respectively, and both studies showed T_sk recordings were different from the T_c recordings (40, 41). Apart from those studies, only one laboratory-based high-intensity exercise study, comparing exercise in four horses in a cool vs. hot environment, suggested that the tail T_sk evolvement pattern seemed to follow the T_c pattern (using arterial blood temperature), although a statistical correlation was not investigated (39). A recent laboratory equine exercise study using the implantation of microchips, measuring muscle temperature (defined as “outer shell temperature” in that study) which may be extrapolated to field exercise in the future (67). There was a good correlation between central venous temperature (CVT) evolvement and outer shell temperature during a short bout of exercise (8–11.5 min) until CVT reached 41°C, although the outer shell temperature was reported to lag behind CVT during the recovery phase. While most of these experimental studies did not reveal a statistical correlation between T_sk and T_c, efforts to further investigate T_sk continue as the technique could easily be employed in the field. Consequently, wearable thermo-sensor techniques are being upgraded at an ever-increasing pace.

Continuous T<italic><italic>sk</italic></italic> Sensor Recording During Field Exercise

Ongoing efforts to identify a reliable proxy for continuously monitoring the thermoregulatory response in horses during field exercise are not always successful or practical. A more invasive method with thermistors placed in blood and brain was used in three horses during free field exercise and reported a good correlation; however, for obvious reasons, this is not easily applicable in practice (51). Several field studies have investigated less-invasive continuously monitoring approaches, such as the intra-uterine temperature (two mares) or the GI temperature; both approaches (intra-uterine and GI) showed a good correlation with the T_c (18, 19, 68). A recent study has continuously monitored surface T_sk using i-Button® and simultaneous T_re during two canter bouts of 4.5 min of field exercise at a speed ranging from 6.7 to 7.5 meters per second (20). The T_sk was continuously monitored at the level of the rump and shoulder regions in 10 eventing horses using a cross-over study design. The study showed that pre-exercise cooling resulted in a delta T_sk ranging from −2.3 to −3.3°C and a reduced median T_re of 0.3°C, which peaked at 9 min into recovery, compared to the control group (20). Although the study did not investigate correlations between T_sk and T_c, the effect of lower T_sk pre-exercise on a reduced T_re is of interest, and consistent with previous human sports studies (69, 70).

In brief, the current reliance on continuous T_sk sensor recordings during field exercise has been proven to be inconsistent and unreliable as a proxy for the thermoregulatory response. This is consistent with a study comparing the effects of precooling in 10 human athletes to estimate the T_c (using a GI pill) (71).

Comparing T<italic><italic>sk</italic></italic> With T<italic><italic>c</italic></italic>

When comparing the delta T_c with delta T_sk in the current study, the delta T_sk was greater during endurance exercise as depicted in Figure 3A. However, a laboratory-based high-intensity equine exercise study using thermocouples attached to the skin with tape and located in pulmonary artery blood to continuously monitor T_sk and T_c in six horses revealed a delta T_sk of 2.5°C (42) similar to the delta T_c value in the current study (3°C). Associations between T_sk and T_c were not evaluated in that study; however, the difference in the exercise duration may indicate a difference in thermoregulatory activity over time, namely, 200 min of endurance exercise in the current study vs. average of 40 min in the former study evidenced by the end-of-exercise T_c > 41°C (42).

Another interesting finding was the significantly higher time to maximum T_c when compared to the time to maximum T_sk (Figure 3B)_. This finding suggests that the endurance horses in the current study performing in a cool environment were efficiently thermoregulating during exercise without the development of hyperthermia (T_c > 39°C) as has been documented previously (2). On the other hand, a short duration of high-intensity exercise in more challenging environmental conditions may trigger T_c > 41°C and, consequently, requires dissipation of excess MH to occur post-exercise (38, 42, 53).

Despite all these ongoing efforts to practically incorporate surface T_sk monitoring into thermoregulation and wellbeing monitoring in the field, researchers should always keep in mind the possible factors that challenge a potential correlation between T_sk and T_c. In addition, the monitoring device used must be able to correctly function and cope with the practical conditions under which horses exercise and compete. Important factors are environmental variables; among which are weather conditions, and whether or not additional cooling is applied. Furthermore, the type of temperature sensor equipment and the anatomical site at which the T_sk equipment is placed have their influence, together with individually intrinsic horse-related factors.

Environmental Factors Factors Influencing T<italic><italic>sk</italic></italic> and T<italic><italic>c</italic></italic> Evolvement and How They Relate to Each Other

The T_sk at any site on the skin surface reflects a balance between heat being delivered to the skin by arterial blood, body, and local skin metabolism, and heat exchange with the environment by convection, radiation, and evaporation. Any factors that interfere with this balance can change the T_sk. Many factors that modulate T_c evolvement during exercise simultaneously influence T_sk, such as a plethora of performance capacity parameters, as well as environmental conditions. Environmental factors can easily and quickly change T_sk without directly affecting T_c (1, 47, 59, 72–74). These factors include the T_a, solar radiation, soil radiation, humidity, shade, and wind speed (air movement). For example, a varying T_a ranging from 20 to 30°C was directly related to the onset of skin vasodilation and sweat evaporation (72), while, on the other hand, a low T_a was shown to induce a lower sensitivity (50%) of percutaneous T_sk microchips in 52 foals and 30 adult horses to identify fever compared to measuring T_sk in a hotter T_a (29°C) (73). That would mean that, in case of fever (also known as an increased T_c set-point), a cool sunless environment renders T_sk monitoring using microchips less representative for T_c monitoring. In addition, Holcomb et al. (74) demonstrated that T_sk and T_re were highest at the peak solar radiation during the day. The T_sk sensor in the current study was located ventrally on the chest of the horse covered by the belt and the girth, thus avoiding T_a effects, such as solar radiation.

It is common practice to cool down endurance horses during the recovery period between subsequent exercise loops. Cooling down was also applied in the current study design in a real-life competition context. The goal was to challenge the temperature monitoring devices with real circumstances in which they would be required to function. With respect to cooling down approaches, the mechanism by which the thermoregulatory systems are challenged greatly depends on how the loss of heat counteracts MH production through non-evaporative pathways as well as evaporative methods. The evaporative exchange of heat of the skin with the environment depends on the thermal gradient between T_sk by local skin perfusion and its immediate environment, including vapor pressure, airflow, and solar radiation, especially during field exercise (17, 49, 75). At the end-of-exercise period, the cooling of sport horses is standard, especially in endurance and 3-day eventing competitions. However, post-race cooling methods are not standardized in the racing industry. The duration of cooling endurance horses in the field is, on average, 10 min based on each owner's judgment, which could include HR monitoring. The mean end-of-exercise T_sk in our study with endurance horses exercising in a cool environment (mean, 15.3°C) was 28°C, and no horses developed a T_sk higher than 39°C.

On the other hand, exercise studies in warmer environments documented a post-exercise T_sk higher than 39°C. For example, a recent report has revealed that 28 out of 38 horses exercising in a hot, dry environment (mean T_a, 38.8°C), and 6 out of 37 horses exercising in a warm, humid environment (mean T_a, 31.1°C) showed a post-exercise IRT T_sk higher than 39°C. These researchers suggested horses recording T_sk higher than 39°C were at risk of developing heat stress and EHI and used this T_sk response as an indicator for racehorses requiring cooling (21). The association between T_sk and EHI risk could be physiologically explained by a low T_c-to-T_sk gradient, therefore decreased capability to transfer MH to the skin, thus compromising the dissipation of MH by evaporation (21, 56).

A similar mean IRT T_sk of 40°C was recorded at the end of exercise in a recent study, evaluating cooling methods in racehorses in a warm environment (mean T_a, 31.8°C) (35). A T_sk higher than 39°C is consistent with earlier laboratory-based studies in a warm T_a (29.1°C and 31.1°C, respectively) (40, 45). In retrospect, scraping off water from the horses during cooling down was not the most optimal approach since Takahashi et al. (35) favored continuous application of cold water without subsequently scraping it off.

T<italic><italic>sk</italic></italic> Equipment-Related Features and Location Equipment to Measure T<italic><italic>sk</italic></italic>

Within the rapidly expanding wearable digital device industry, surface T_sk monitoring devices are constantly being upgraded to provide data output. However, in that respect, the critical question remains: how should we interpret those data? Overall, three different types of temperature sensor equipment are reported: thermistors (such as microchips), thermocouples, and IRT devices, with IRT being the most studied device recently in horses (21, 22, 26, 32–37, 76–80). It is essential to understand that those sensors use different physical processes to obtain data, which may result in significant differences in data output. These sensor surface T_sk differences due to the type of equipment may show only a minor bias, which may prove to be clinically meaningful (75). For example, a study comparing IRT and thermocouples at single pre-exercise, intra-exercise and post-exercise points in 12 human athletes revealed a poor Bland Altman agreement and low reliability between the different methods (81).

To produce IRT imaging to picture surface T_sk of different parts or the whole body, a remote IRT camera positioned at 30-cm proximity to the skin surface has been recently evaluated with varying results (26, 37, 82). For example, a study compared IRT T_sk to T_re in 40 adult horses and concluded T_sk was not an accurate method to determine the T_c (82). The remote position of the IRT held far from the skin has the advantage of not interfering with the local T_sk balance, although the remote T_sk measurement will be partly affected by the adjacent environment surrounding the skin (22, 23, 75). The temperature sensors that were used in the current study were in direct contact with the skin and covered by a belt. This belt might have interfered with the local thermal conductivity and the local evaporative cooling capacity and thus might have delayed equilibration of the local T_sk with the surrounding skin. On the other hand, an adequate and essential sensor-to-skin contact was ensured by the position of the sensor in the belt. Furthermore, the skin surface covered by the sensor was small enough to prevent causing local skin changes (22, 75).

We were unable to calibrate and validate the IR sensor prior to the study; however, studies of different T_sk recording methods and comparisons with a certified thermocouple in a thermo-statically controlled water bath are extremely rare. One study evaluated sensor systems in human athletes during rest and exercise in a hot environment and revealed a good agreement for employing a telemetric thermistor system when compared to the standard hard-wired thermistor system and a poor agreement for using a thermal camera (83).

In summary, IRT techniques differ widely in human and equine medicine, including positioning of the camera and environmental control measures (22, 24, 75). A consensus guideline has been developed only recently, addressing the multiple data collection methods of the human T_sk using IRT (84), while Soroko and Howell (22) described a protocol using IRT in equine medicine.

T<italic><italic>sk</italic></italic> Equipment Location

The anatomical location of the sensor on the horse to record T_sk measurements has been shown to influence T_sk results (22, 23, 26, 29, 37, 47). For example, remote IRT was used to evaluate differences between 10 locations on the body during two seasons in the year with the highest T_sk recorded at the level of the chest (22.5°C) and shoulders (20.4°C) in horses at rest in a cool T_a (mean, 6.7°C) (26). In another study, the IRT T_sk was greatest at the shoulder area when compared to three other T_sk locations measured at the start and the end of 20-min exercise (32.3°C and 34.2°C, respectively) in a moderate T_a (mean, 23°C) (37). The results of our study in a cool T_a (mean, 15.3°C) revealed a mean T_sk measured at the lower chest area of 23.1°C during an average of 200-min exercise and a mean T_sk of 21.8°C during recovery. The different T_sk values between the current study and Soroko et al. (37) illustrate the effect of exercise intensity (submaximal vs. maximal) and duration (long vs. short). The differences in T_sk over several body areas may relate to the varying networks of blood vessels in those body regions and their vasodilation to exchange thermal heat with the proximal environment (23, 55, 56, 59). It is essential to note that both monitoring methods share some vasomotor or endocrine mechanisms, although they present differences depending on the degree of heat dissipation or retention that the organism needs. Consequently, monitoring of T_sk in the current study revealed a physiological response of the local T_sk to the changes of T_c during endurance exercise over time, although the responses were not correlated.

While IRT cameras are increasingly used in equine sports medicine, this method involves a single point in time measurement. One exception is the study by Soroko et al. (37), who reported dynamic IRT monitoring every 15 s during treadmill exercise. To be precise, a review of the use of IRT in human endurance athletes reported that 25 of the 45 studies were conducted over the last 5 years (2017–2021), but, up until now, only five real-life field endurance studies have been performed (24). The latter review concluded that further analysis is required to assess whether T_sk could be used as a reliable proxy to describe real-time thermoregulation (24). Another important relevant finding is that surface T_sk may be low in human athletes with EHI and, hence, provide misleading information (85). A different IRT method approach is measuring eye surface temperature; that study revealed no relationship with T_c in horses (86).

Individual Horse-Related Factors

Horse-related factors include breed, body condition score, age, character (such as nervousness), and skin-related properties, such as sweat rate, skin thickness, blood vessel density, hair coat properties, clipping, and coat color (23, 27, 30, 36, 55, 77, 78, 80, 87). The sweat loss in the current study was subjectively scored from 1 to 3 by E-L.V., and, in retrospect, more accurate sweating scoring based on objective specific phenotypic descriptions would have been a better approach (88). The effect of breed on T_sk relates to the ratio of BM to the body surface area—the higher the body surface area in relation to the BM, the higher the heat dissipation (72). The low-surface-area-to-mass ratio of the horse results in greater demands being imposed on the thermoregulatory system during long-term submaximal exercise (30, 49, 53). Our study included mainly Arabian horses, known to have a lower BM and, hence, a higher-surface-area-to-mass ratio.

The hair coat length in the current study was similar (all clipped), which is essential as clipping the winter coat resulted in improved heat dissipation during and after exercise, resulting in decreased T_sk and T_re, as reported in previous studies (27, 30). One of those previous studies used a thermistor probe to evaluate the effect of coat clipping in three horses on both the surface T_sk and T_re. That study reported no effect of clipping on post-exercise T_sk, while T_re was ~0.2°C higher in unclipped horses (30). Indeed, a longer haircoat length limited the thermal imaging in a study assessing T_sk in mares (77). Furthermore, coat colors may be relevant (87); however, the current study of 13 horses revealed that light or dark coat color had no significant effect on T_sk, which is consistent with a previous study (73). Individual horse-related character differences may exist, such as nervousness that triggers sympathetic nerve activity associated with vasoconstriction of skin blood vessels. This neurophysiological response may explain varying reduced local T_sks, decreased heat loss, and hyperthermia (23, 55, 78, 89).

Modeling Using T<italic><italic>sk</italic></italic>

While, generally, T_sk can be easily monitored, the T_sk in the current study did not provide data suitable for extrapolating to similar changes in the T_c. Consequently, the development of integrative models using T_sk to determine the heat balance during exercise has been investigated in human studies and in one equine study (87, 90, 91). However, no regression model could predict physiological stress load using single-point IRT T_sk in 17 human marathon runners in the field (92). A recent approach in human exercise research has investigated the application of models and algorithms using data and variables, such as HR and HR variability, to successfully estimate T_c (47, 90, 93–95). Physiologically, HR reflects the blood flow rate to the muscles (MH production) and blood flow to the skin (heat loss). For example, recent studies have concluded that combining continuous insulated T_sk and HR monitoring in 13 and 8 human athletes in a hot (35°C) and warm (25°C) environment, respectively, could provide a predictive model of T_re or T_c (using GI pills) (90, 94). In contrast in the current study, HR recovery in the endurance horses was not directly related to T_sk. Further investigation is required into the potential association of T_sk and HR for accurate predictive modeling of T_c in equine athletes.

Association Between Single-Point T<italic><italic>sk</italic></italic> at the End-of-Exercise Period Compared With the T<italic><italic>c</italic></italic> at the End-of-Recovery Period

An interesting finding of the current study performed with the endurance horses was the association of a greater T_sk at the end of exercise with a significantly lesser T_c at the end of recovery (60 min). Several theories could be considered to explain this association between T_sk and T_c: firstly, the raised T_sk indicates the launch of an active thermoregulatory response to anticipate the increased T_c, and, once the MH is successfully dissipated, the T_c decreases. This argument can be coupled with the effect of cooling post-exercise, which may be more prominent when T_sk is greater and, ultimately, results in higher dissipation of MH and a reduced T_c. Several other field exercise studies in horses have investigated correlations between single-point T_sk and other variables (21, 32). For example, a recent equine study involving 8 endurance horses has investigated the association between endurance training intensity (1 h at 19 km/h vs. 2 h at 16 km/h vs. 3 h at 20 km/h) and T_sk using an IRT camera measured at different locations and at different time points. The study identified that the T_sk at the coronary band increased with training intensity unlike the maximum T_sk (32).

Aside from the variance in hot vs. cool T_a in these studies the differences in exercise intensity could explain the dissimilarity between the racehorse study results of Brownlow and Mizzi (5) and the current study involving endurance horses. For racehorses undertaking high-intensity, short-duration exercise, the dissipation of MH occurs post-exercise as opposed to endurance horses, which manage their MH throughout their submaximal long-duration exercise (2, 5, 21, 56). For example, the T_sk and its evolvement pattern can be related to acute blood flow variances associated with a different type of exercise intensity (25, 95). Overall, in our study monitoring endurance horses conducting exercise during cooler months, the mean T_sk at the end of exercise was 28°C, while none of these horses had a T_sk higher than 39°C.

The end-of-recovery period T_sk showed a considerable individual variation (range, 14.5–28.7°C) despite the application of a uniform cooling protocol. The T_sk during the 60-min recovery period revealed that the T_sk returned to the baseline only in over 50% of the 40-km recovery periods. This is in contrast to other studies, which found that after 20 min of treadmill exercise in a hot (32–34°C) and dry T_a condition, all T_sk values returned to baseline T_sk after 60 min, and, after 45 min in a T_a of 20°C (without cooling), respectively (28, 42). The main difference between the current study and other laboratory-based studies was the continuous T_sk monitoring during a field exercise in an uncontrolled T_a.

Limitations

As in any study, there are several limitations that should be considered. Obviously, throughout this “in-the-field” study, not all research conditions could be controlled for 100% of the time, such as weather conditions involving T_a and the degree of solar radiation, the training, and the dietary management of participating horses. These factors may have affected the individual T_sk and T_c time profiles. However, this applies to all “in-the-field” competition studies and, under ideal conditions, should not interfere with the reliability of a solid thermoregulatory monitoring proxy suitable for assuring the thermoregulatory wellbeing of competition horses in the field (74). Endeavors to assess the thermal environmental variables were limited to BOM and HOBO recordings of the T_a and the RH, with the HOBO device placed at one location. Other essential external variables, such as wind speed, were not included in the T_a measurements (17). The current study involved only one type and location of the wearable T_sk sensor based on IR technology. In the future, other thermo-physical measuring approaches will prove to be more robust. However, on all occasions, the involvement of a validated “gold standard” against which the performance of new individual monitoring devices is set should be an essential part of future studies (75).

Conclusion

While the method of monitoring T_sk may be non-invasive and straightforward, our results have clearly shown that T_sk monitoring on its own does not reliably estimate the T_c evolvement during a field exercise in endurance horses since a correlation between T_c and T_sk could not be identified. Notably, a high T_sk at a single point during a field exercise in a cool T_a did not identify the endurance horses with an increased T_c. Further research into T_c monitoring in different equine sports and under differing weather conditions must be undertaken to create a baseline for further fine-tuning hot weather policies. Accordingly, veterinarians, trainers, and owners can be advised to continuously monitor T_c to ensure the health and welfare of all horses.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics Statement

The animal study was reviewed and approved by University of Adelaide Animal Ethics Committee. Written informed consent was obtained from the owners for the participation of their animals in this study.

Author Contributions

E-LV prepared and carried out the study design, data collection, preparation and creation of the database, descriptive and part of statistical analysis, interpretation of the data, and writing of the manuscript. CD was involved in study design, preparation and creation of the database, analysis and interpretation of the data, and writing of the manuscript. GH contributed to drafting and revising the manuscript. TM contributed to reviewing the manuscript and statistical analysis. The final manuscript was read and approved by all authors.

Funding

This study was partly financially supported by a grant from the University of Adelaide for the cost of monitoring equipment.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

The authors are deeply grateful to all the participating endurance horse owners and the SAERA organization for their welcome and logistic support. The authors would like to acknowledge Honours Research Student Stacey E. Jonas for assistance with data collection.

Supplementary Material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fvets.2022.894146/full#supplementary-material

References 1. Raymond

Matthews

Horton

. The emergence of heat and humidity too severe for human tolerance. Sci Advances. (2020) 6:eaaw1838. 10.1126/sciadv.aaw1838

32494693

2. Verdegaal

ELJMM

Howarth

Mcwhorter

Boshuizen

Franklin

Vidal Moreno De Vega

. Continuous monitoring of the thermoregulatory response in endurance horses and trotter horses during field exercise: Baselining for future hot weather studies. Front Physiol. (2021) 12:708737. 10.3389/fphys.2021.708737

34512382

3. Casa

Demartini

Bergeron

Csillan

Eichner

Lopez

. National athletic trainers' association position statement: exertional heat illnesses. J Athl Train. (2015) 50:986–1000. 10.4085/1062-6050-50.9.07

26381473

4. Brownlow

Dart

Jeffcott

. Exertional heat illness: a review of the syndrome affecting racing Thoroughbreds in hot and humid climates. Aust Vet J. (2016) 94:240–7. 10.1111/avj.12454

27349884

5. Brownlow

Mizzi

. Exertional heat illness in thoroughbred racehorses – pathophysiology, case definition and treatment rationale. Equine Vet Ed. (2022) 34,5:259–71. 10.1111/eve.13459 6. Geor

McCutcheon

. Hydration effects on physiological strain of horses during exercise-heat stress. J Appl Physiol. (1998) 84:2042–51. 10.1152/jappl.1998.84.6.2042

9609799

7. Barnes

Kingston

Beetson

Kuiper

. Endurance veterinarians detect physiologically compromised horses in a 160 km ride. Equine Vet J Suppl. (2010) 42:6–11. 10.1111/j.2042-3306.2010.00225.x

21058975

8. Nagy

Murray

Dyson

. Elimination from elite endurance rides in nine countries: a preliminary study. Equine Vet J Suppl. (2010) 42:637–43. 10.1111/j.2042-3306.2010.00220.x

21059073

9. Fielding

Meier

Balch

Kass

. Risk factors for the elimination of endurance horses from competition. J Am Vet Med Assoc. (2011) 239:493–8. 10.2460/javma.239.4.493

21838587

10. Nagy

Murray

Dyson

. Descriptive epidemiology and risk factors for eliminations from Federation Equestre Internationale endurance rides due to lameness and metabolic reasons (2008-2011). Equine Vet J. (2014) 46:38–44. 10.1111/evj.12069

23574513

11. Younes

Robert

Cottin

Barrey

. Speed and cardiac recovery variables predict the probability of elimination in equine endurance events. PLoS ONE. (2015) 10:e0137013. 10.1371/journal.pone.0137013

26322506

12. Fielding

Meier

Fellers

Magdesian

. Ability of clinicopathologic variables and clinical examination findings to predict race elimination in endurance horses. Am J Vet Res. (2017) 78:50–6. 10.2460/ajvr.78.1.50

28029292

13. Bennet

Parkin

TDH

. Federation equestre internationale endurance events: Risk factors for failure to qualify outcomes at the level of the horse, ride and rider (2010-2015). Vet J. (2018) 236:44–8. 10.1016/j.tvjl.2018.04.011

29871749

14. Legg

Weston

Gee

Bolwell

Bridges

Rogers

. Characteristics of endurance competitions and risk factors for elimination in new zealand during six seasons of competition (2010/11-2015/16). Animals. (2019) 9:611. 10.3390/ani9090611

31461901

15. Nomura

Shiose

Ishikawa

Mizobe

Sakai

Kusano

. Prevalence of post-race exertional heat illness in Thoroughbred racehorses and climate conditions at racecourses in Japan. J Equine Sci. (2019) 30:17–23. 10.1294/jes.30.17

31285689

16. Takahashi

Takahashi

. Risk factors for exertional heat illness in Thoroughbred racehorses in flat races in Japan (2005-2016). Equine Vet J. (2020) 52:364–8. 10.1111/evj.13179

31505059

17. Brownlow

Brotherhood

. An investigation into environmental variables influencing post-race exertional heat illness in Thoroughbred racehorses in temperate eastern Australia. Aus Vet J. (2021) 99:473–81. 10.1111/avj.13108

34355381

18. Smith

Barnes

Maloney

. A nonsurgical method allowing continuous core temperature monitoring in mares for extended periods, including during endurance exercise. Equine Vet J. (2006) 36:65–9. 10.1111/j.2042-3306.2006.tb05515.x

17402394

19. Verdegaal

EJMM

Delesalle

Caraguel

CGB

Folwell

Mcwhorter

Howarth

. Evaluation of a telemetric gastrointestinal pill for continuous monitoring of gastrointestinal temperature in horses at rest and during exercise. Am J Vet Res. (2017) 78:778–84. 10.2460/ajvr.78.7.778

28650242

20. Klous

Siegers

Van Den Broek

Folkerts

Gerrett

Van Oldruitenborgh-Oosterbaan

. Effects of pre-cooling on thermophysiological responses in elite eventing horses. Animals. (2020) 10:1664. 10.3390/ani10091664

32947831

21. Brownlow

Smith

. The use of the hand-held infrared thermometer as an early detection tool for Exertional Heat Illness in Thoroughbred racehorses: a study at racetracks in eastern Australia. Equine Vet Ed. (2021) 33:296–305. 10.1111/eve.13299 22. Soroko

Howell

. Infrared thermography: Current applications in equine medicine. J Equine Vet Sci. (2018) 60:90–6. 10.1016/j.jevs.2016.11.002 23. Mota-Rojas

Pereira

AMF

Wang

Martínez-Burnes

Ghezzi

Hernández-Avalos

. Clinical applications and factors involved in validating thermal windows used in infrared thermography in cattle and river buffalo to assess health and productivity. Animals. (2021) 11:2247. 10.3390/ani11082247

34438705

24. Rojas-Valverde

Tomás-Carús

Timón

Batalha

Sánchez-Ureña

Gutiérrez-Vargas

. Short-term skin temperature responses to endurance exercise: A systematic review of methods and future challenges in the use of infrared thermography. Life. (2021) 11:1286. 10.3390/life11121286

34947817

25. Mota-Rojas

Titto

Orihuela

Martínez-Burnes

Gómez-Prado

Torres-Bernal

. Physiological and behavioral mechanisms of thermoregulation in mammals. Animals. (2021) 11:1733. 10.3390/ani11061733

34200650

26. Meisfjord Jorgensen

Mejdell

Boe

. Effects of hair coat characteristics on radiant surface temperature in horses. J Therm Biol. (2020) 87:102474. 10.1016/j.jtherbio.2019.102474

31999605

27. Morgan

Funkquist

Nyman

. The effect of coat clipping on thermoregulation during intense exercise in trotters. Equine Vet J Suppl. (2002) 34:564–7. 10.1111/j.2042-3306.2002.tb05484.x

12405752

28. Simon

Gaughan

Epp

Spire

. Influence of exercise on thermographically determined surface temperatures of thoracic and pelvic limbs in horses. J Am Vet Med Assoc. (2006) 229:1940–4. 10.2460/javma.229.12.1940

17173534

29. Jodkowska

Oblacinska

Tabak

Radiukiewicz

. Differences in dietary patterns between overweight and normal-weight adolescents. Med Wieku Rozwoj. (2011) 15,3:266–73. 22006481 30. Wallsten

Olsson

Dahlborn

. Temperature regulation in horses during exercise and recovery in a cool environment. Acta Vet Scand. (2012) 54:42. 10.1186/1751-0147-54-42

22805591

31. Yarnell

Fleming

Stratton

Brassington

. Monitoring changes in skin temperature associated with exercise in horses on a water treadmill by use of infrared thermography. J Therm Biol. (2014) 45:110–6. 10.1016/j.jtherbio.2014.08.003

25436959

32. Redaelli

Luzi

Mazzola

Bariffi

Zappaterra

Nanni Costa

. The use of infrared thermography (IRT) as stress indicator in horses trained for endurance: a pilot study. Animals. (2019) 9:84. 10.3390/ani9030084

30866503

33. Soroko

Spitalniak-Bajerska

Zaborski

Pozniak

Dudek

Janczarek

. Exercise-induced changes in skin temperature and blood parameters in horses. Arch Anim Breed. (2019) 62:205–13. 10.5194/aab-62-205-2019

31807631

34. Soroko

Zaborski

Dudek

Yarnell

Gorniak

Vardasca

. Evaluation of thermal pattern distributions in racehorse saddles using infrared thermography. PLoS ONE. (2019) 14:e0221622. 10.1371/journal.pone.0221622

31449556

35. Takahashi

Ohmura

Mukai

Shiose

Takahashi

. A comparison of five cooling methods in hot and humid environments in Thoroughbred horses. J Equine Vet Sc. (2020) 91:103130. 10.1016/j.jevs.2020.103130

32684268

36. Wilk

Wnuk-Pawlak

Janczarek

Kaczmarek

Dybczynska

Przetacznik

. Distribution of superficial body temperature in horses ridden by two riders with varied body weights. Animals. (2020) 10:340. 10.3390/ani10020340

32098105

37. Soroko

Howell

Dudek

Wilk

Zastrzezyńska

Janczarek

. A pilot study into the utility of dynamic infrared thermography for measuring body surface temperature changes during treadmill exercise in horses. J Equine Vet Sci. (2018) 62:44–6. 10.1016/j.jevs.2017.12.010 38. Geor

McCutcheon

Ecker

Lindinger

. Thermal and cardiorespiratory responses of horses to submaximal exercise under hot and humid conditions. Equine Vet J Suppl. (1995) 20:125–32. 10.1111/j.2042-3306.1995.tb05018.x

8933095

39. Marlin

Scott

Schroter

Mills

Harris

. Physiological responses in nonheat acclimated horses performing treadmill exercise in cool (20(C/40%RH), hot dry (30(C/40%RH) and hot humid (30 (C/80%RH) conditions. Equine Vet J Suppl. (1996) 22:70–84. 10.1111/j.2042-3306.1996.tb05034.x

8894553

40. Marlin

Scott

Roberts

Casas

Holah

Schroter

. Post exercise changes in compartmental body temperature accompanying intermittent cold water cooling in the hyperthermic horse. Equine Vet J. (1998) 30:28–34. 10.1111/j.2042-3306.1998.tb04085.x

9458396

41. Marlin

Scott

Schroter

Harris

Roberts

. Physiological responses of horses to a treadmill simulated speed and endurance test in high heat and humidity before and after humid heat acclimation. Equine Vet J. (1999) 31:31–42. 10.1111/j.2042-3306.1999.tb03788.x

9952327

42. Geor

McCutcheon

Ecker

Lindinger

. Heat storage in horses during submaximal exercise before and after humid heat acclimation. J Appl Physiol. (2000) 89:2283–93. 10.1152/jappl.2000.89.6.2283

11090580

43. Kohn

Hinchcliff

. Physiological responses to the endurance test of a 3-day-event during hot and cool weather. Equine Vet J Suppl. (1995) 31–6. 10.1111/j.2042-3306.1995.tb05005.x

8933082

44. Jeffcott

Kohn

. Contributions of equine exercise physiology research to the success of the 1996 Equestrian Olympic games: a review. Equine Vet J Suppl. (1999) 30:347–55. 10.1111/j.2042-3306.1999.tb05247.x

10659281

45. Kohn

Hinchcliff

Mckeever

. Effect of ambient temperature and humidity on pulmonary artery temperature of exercising horses. Equine Vet J Suppl. (1999) 404–11. 10.1111/j.2042-3306.1999.tb05256.x

10659290

46. Jeffcott

Leung

Riggs

. Managing the effects of the weather on the equestrian events of the 2008 Beijing Olympic games. Vet J. (2009) 182:412–29. 10.1016/j.tvjl.2009.07.037

19716734

47. Wartzek

Muhlsteff

Imhoff

. Temperature measurement. Biomed Tech. (2011) 56:241–57. 10.1515/BMT.2011.108

21988157

48. Hodgson

. Thermoregulation. In: The Athletic Horse, Principles and Practice of Equine Sport Medicine. 2nd ed. St. Louis: Saunders/Elsevier (2014). p. 108–124. 49. Hodgson

McCutcheon

Byrd

Brown

Bayly

Brengelmann

. Dissipation of metabolic heat in the horse during exercise. J Appl Physiol. (1993) 74:1161–70. 10.1152/jappl.1993.74.3.1161

8482654

50. AERA. AERA Rulebook. Australian Endurance Riding Association (2020). Available online at: http://aera.asn.au/national-rules/ (accessed December 20, 2020). 51. Mitchell

Fuller

Maloney

Rump

Mitchell

. Guttural pouches, brain temperature and exercise in horses. Biol Letters. (2006) 2:475–7. 10.1098/rsbl.2006.0469

17148434

52. Marlin

Schroter

Scott

White

Nyrop

Maykuth

. Sweating and skin temperature responses of normal and anhidrotic horses to intravenous adrenaline. Equine Vet J Suppl. (1999) 30:362–9. 10.1111/j.2042-3306.1999.tb05249.x

10659283

53. Hodgson

Davis

Mcconaghy

. Thermoregulation in the horse in response to exercise. Br Vet J. (1994) 150:219–35. 10.1016/S0007-1935(05)80003-X

8044664

54. McCutcheon

Geor

. Thermoregulation and exercise-associated heat stress. In: Equine Exerc. Physiol., The Science of Exercise in the Athletic Horse, Hinchcliff

Geor

Kaneps

, editors. (2008). p. 382–396. 55. Bertoni

Mota-Rojas

Álvarez-Macias

Mora-Medina

Guerrero-Legarreta

Morales-Canela

. Scientific findings related to changes in vascular microcirculation using infrared thermography in the river buffalo. J An Behav Biometeorology. (2020) 8:288–97. 10.31893/jabb.20038 56. Brownlow

Mizzi

. Thermoregulatory capacity of the Thoroughbred racehorse and its relationship to the pathogenesis of exertional heat illness. Equine Vet Ed. (2022) 34:214–21. 10.1111/eve.13433 57. Jones

Carlson

. Estimation of metabolic energy cost and heat production during a 3-day-event. Equine Vet J Suppl. (1995) 27,S20:23–30. 10.1111/j.2042-3306.1995.tb05004.x

8933081

58. McCutcheon

Geor

Hare

Ecker

Lindinger

. Sweating rate and sweat composition during exercise and recovery in ambient heat and humidity. Equine Vet J Suppl. (1995) 20:153–7. 10.1111/j.2042-3306.1995.tb05022.x

8933099

59. Mota-Rojas

Wang

Titto

Gómez-Prado

Carvajal-de la Fuente

Ghezzi

. Pathophysiology of fever and application of infrared thermography (IRT) in the detection of sick domestic animals: recent advances. Animals. (2021) 11:2316. 10.3390/ani11082316

34438772

60. B.O.M. Bureau for Meteorology. (2020). Available online at: http://www.bom.gov.au/climate/data/ (accessed December 2, 2020). 61. Parker

Goodwin

Eager

Redhead

Marlin

. Comparison of polar® heart rate interval data with simultaneously recorded ecg signals in horses. Comp Exercise Physiol. (2009) 6:137–42. 10.1017/S1755254010000024 62. Datta

Marder

Datta

Meister

Puric

Stutz

. Quantification of thermal dose in moderate clinical hyperthermia with radiotherapy: a relook using temperature-time Area Under the Curve (AUC). Int J Hyperthermia. (2021) 38:296–307. 10.1080/02656736.2021.1875060

33627018

63. Fenemor

Gill

Sims

Beaven

Driller

. Validity of a tympanic thermometer and thermal imaging camera for measuring core and skin temperature during exercise in the heat. Measurement Phys Ed Exerc Sci. (2020) 24:49–55. 10.1080/1091367X.2019.1667361 64. Foster

Hodder

Lloyd

Havenith

. Individual responses to heat stress: Implications for hyperthermia and physical work capacity. Front Physiol. (2020) 11:541483. 10.3389/fphys.2020.541483

33013476

65. De Korte

Bongers

CCWG

Hopman

MTE

Eijsvogels

TMH

. Exercise performance and thermoregulatory responses of elite athletes exercising in the heat: outcomes of the thermo Tokyo study. Sports Med. (2021) 51:2423–36. 10.1007/s40279-021-01530-w

34396493

66. Westwood

Fallowfield

Delves

Nunns

Ogden

Layden

. Individual risk factors associated with exertional heat illness: a systematic review. Exp Physiol. (2021) 106:191–9. 10.1113/EP088458

32249985

67. Kang

Zsoldos

Woldeyohannes

Gaughan

Sole Guitart

. The use of percutaneous thermal sensing microchips for body temperature measurements in horses prior to, during and after treadmill exercise. Animals. (2020) 10:2274. 10.3390/ani10122274

33276500

68. Green

Gates

Lawrence

. Measurement of horse core body temperature. J Therm Biol. (2005) 30:370–7. 10.1016/j.jtherbio.2005.03.003

11202378

69. Bongers

CCWG

Hopman

MTE

Eijsvogels

TMH

. Cooling interventions for athletes: an overview of effectiveness, physiological mechanisms, and practical considerations. Temperature. (2017) 4:60–78. 10.1080/23328940.2016.1277003

28349095

70. Racinais

Ihsan

Taylor

Cardinale

Adami

Alonso

. Hydration and cooling in elite athletes: Relationship with performance, body mass loss and body temperatures during the doha 2019 iaaf world athletics championships. Br J Sports Med. (2021) 55:1335–41. 10.1136/bjsports-2020-103613

33579722

71. Faulkner

Hupperets

Hodder

Havenith

. Conductive and evaporative precooling lowers mean skin temperature and improves time trial performance in the heat: conductive and evaporative precooling enhances performance. Scan J Med Sci Sports. (2015) 25:183–9. 10.1111/sms.12373

25943669

72. Morgan

EKM

. Climatic energy demand of horses. Equine Vet J. (1995) 27:396–9. 10.1111/j.2042-3306.1995.tb04960.x 73. Robinson

Hussey

Hill

Heckendorf

Stricklin

Traub-Dargatz

. Comparison of temperature readings from a percutaneous thermal sensing microchip with temperature readings from a digital rectal thermometer in equids. J Am Vet Med Ass. (2008) 233:613–7. 10.2460/javma.233.4.613

18710319

74. Holcomb

Tucker

Stull

. Physiological, behavioral, and serological responses of horses to shaded or unshaded pens in a hot, sunny environment1. J An Sci. (2013) 91:5926–36. 10.2527/jas.2013-6497

24126269

75. MacRae

Annaheim

Spengler

Rossi

. Skin temperature measurement using contact thermometry: a systematic review of setup variables and their effects on measured values. Front Physiol. (2018) 9:29–29. 10.3389/fphys.2018.00029

29441024

76. Giannetto

Arfuso

Giudice

Gianesella

Fazio

Panzera

. Infrared methodologies for the assessment of skin temperature daily rhythm in two domestic mammalian species. J Therm Biol. (2020) 92:102677. 10.1016/j.jtherbio.2020.102677

32888574

77. Maśko

Witkowska-Piłaszewicz

Jasiński

Domino

. Thermal features, ambient temperature and hair coat lengths: Limitations of infrared imaging in pregnant primitive breed mares within a year. Repro Dom Animals. (2021) 56:1315–28. 10.1111/rda.13994

34310786

78. Mota-Rojas

Titto

de Mira Geraldo

Martínez-Burnes

Gómez

Hernández-Ávalos

. Efficacy and function of feathers, hair, and glabrous skin in the thermoregulation strategies of domestic animals. Animals. (2021) 11:3472. 10.3390/ani11123472

34944249

79. Zielińska

Soroko

Howell

Godlewska

Hildebrand

Dudek

. Comparison of the effect of high-intensity laser therapy (hilt) on skin surface temperature and vein diameter in pigmented and non-pigmented skin in healthy racehorses. Animals. (2021) 11:1965. 10.3390/ani11071965

34209183

80. Domino

Borowska

Trojakowska

Kozłowska

Zdrojkowski

Jasiński

. The effect of rider: horse bodyweight ratio on the superficial body temperature of horse's thoracolumbar region evaluated by advanced thermal image processing. Animals. (2022) 12:195. 10.3390/ani12020195

35049815

81. Fernandes

ADA

Amorim

PRDS

Brito

De Moura

Moreira

Costa

CMA

. Measuring skin temperature before, during and after exercise: a comparison of thermocouples and infrared thermography. Physiol Measure. (2014) 35:189–203. 10.1088/0967-3334/35/2/189

24398429

82. Ramey

Bachmann

Lee

. A comparative study of non-contact infrared and digital rectal thermometer measurements of body temperature in the horse. J Equine Vet Sci. (2011) 31:191–3. 10.1016/j.jevs.2011.02.009 83. James

Richardson

Watt

Maxwell

. Reliability and validity of skin temperature measurement by telemetry thermistors and a thermal camera during exercise in the heat. J Therm Biol. (2014) 45:141–9. 10.1016/j.jtherbio.2014.08.010

25436963

84. Moreira

Costello

Brito

Adamczyk

Ammer

Bach

AJE

. Thermographic imaging in sports and exercise medicine: a delphi study and consensus statement on the measurement of human skin temperature. J Therm Biol. (2017) 69:155–62. 10.1016/j.jtherbio.2017.07.006

29037377

85. Belval

Armstrong

. Comparative physiology of thermoregulation. In: Sport and Physical Activity in the Heat. Casa

, editor (Springer IntPubl Switzerland) (2018). p. 3–14. 86. Soroko

Howell

Zwyrzykowska

Dudek

Zielińska

Kupczyński

. Maximum eye temperature in the assessment of training in racehorses: correlations with salivary cortisol concentration, rectal temperature, and heart rate. J Equine Vet Sci. (2016) 45:39–45. 10.1016/j.jevs.2016.06.005 87. Mostert

Lund

Guthrie

Cilliers

. Integrative model for predicting thermal balance in exercising horses. Equine Vet J Suppl. (1996) 22:7–15. 10.1111/j.2042-3306.1996.tb05026.x

8894545

88. Zeyner

Romanowski

Vernunft

Harris

Kienzle

. Scoring of sweat losses in exercised horses - a pilot study. J An Physiol An Nutr. (2014) 98:246–50. 10.1111/jpn.12073

23534876

89. Hetem

Mitchell

de Witt

Fick

Meyer

Maloney

. Cheetah do not abandon hunts because they overheat. Biol Lett. (2013) 9:20130472. 10.1098/rsbl.2013.0472

23883578

90. Eggenberger

MacRae

Kemp

Burgisser

Rossi

Annaheim

. Prediction of core body temperature based on skin temperature, heat flux, and heart rate under different exercise and clothing conditions in the heat in young adult males. Front Physiol. (2018) 9:1780. 10.3389/fphys.2018.01780

30618795

91. Tanda

. A simplified approach to describe the mean skin temperature variations during prolonged running exercise. J Therm Biol. (2021) 99:103005. 10.1016/j.jtherbio.2021.103005

34420635

92. Pérez-Guarner

Priego-Quesada

Oficial-Casado

Cibrián Ortiz De Anda

Carpes

Palmer

. Association between physiological stress and skin temperature response after a half marathon. Physiol Measure. (2019) 40:034009. 10.1088/1361-6579/ab0fdc

30870816

93. Cuddy

Buller

Hailes

Ruby

. Skin temperature and heart rate can be used to estimate physiological strain during exercise in the heat in a cohort of fit and unfit males. Mil Med. (2013) 178:e841–7. 10.7205/MILMED-D-12-00524

23820362

94. Welles

Buller

Looney

Rumpler

Gribok

Hoyt

. Estimation of metabolic energy expenditure from core temperature using a human thermoregulatory model. J Therm Biol. (2018) 72:44–52. 10.1016/j.jtherbio.2017.12.007

29496014

95. Hillen

Pfirrmann

Nägele

Simon

. Infrared thermography in exercise physiology: the dawning of exercise radiomics. Sports Med. (2019) 50:263–82. 10.1007/s40279-019-01210-w

31734882

Abbreviations

Gastrointestinal

heart rate

bpm

beats per minute

B.O.M.

Bureau of Meteorology

GPS

Global Positioning System

A.E.R.A.

Australian Endurance Riders Association

SAERA

South Australian Endurance Riders Association.

¹https://connect.garmin.com/

²www.graphpad.com