Core body temperatures of rats subjected to treadmill exercise to fatigue or exhaustion: The journal Temperature toolbox

ABSTRACT

This study systematically reviewed the literature reporting the changes in rats’ core body temperature (T_CORE) induced by either incremental- or constant-speed running to fatigue or exhaustion. In addition, multiple linear regression analyses were used to determine the factors contributing to the T_CORE values attained when exercise was interrupted. Four databases (EMBASE, PubMed, SPORTDiscus, and Web of Science) were searched in October 2021, and this search was updated in August 2022. Seventy-two studies (n = 1,538 rats) were included in the systematic review. These studies described heterogeneous experimental conditions; for example, the ambient temperature ranged from 5 to 40°C. The rats quit exercising with T_CORE values varying more than 8°C among studies, with the lowest and highest values corresponding to 34.9°C and 43.4°C, respectively. Multiple linear regression analyses indicated that the ambient temperature (p < 0.001), initial T_CORE (p < 0.001), distance traveled (p < 0.001; only incremental exercises), and running speed and duration (p < 0.001; only constant exercises) contributed significantly to explaining the variance in the T_CORE at the end of the exercise. In conclusion, rats subjected to treadmill running exhibit heterogeneous T_CORE when fatigued or exhausted. Moreover, it is not possible to determine a narrow range of T_CORE associated with exercise cessation in hyperthermic rats. Ambient temperature, initial T_CORE, and physical performance-related variables are the best predictors of T_CORE at fatigue or exhaustion. From a broader perspective, this systematic review provides relevant information for selecting appropriate methods in future studies designed to investigate exercise thermoregulation in rats.

KEYWORDS:

• Body mass
• environment
• heat
• hyperthermia
• performance
• physical exercise
• regression analysis
• thermoregulation

Introduction

Rats are widely used to study the physiological mechanisms underlying the regulation of physical performance, including the thermoregulatory mechanisms [1]. In particular, they represent an experimental model for investigating exercise performance and physiology when the use of human subjects is neither feasible nor desirable [2]. For example, rats are used in studies involving invasive procedures that cannot be carried out in humans for ethical reasons [2]. These invasive procedures include brain cannulation [3,4], identification of exercise-induced activation of hypothalamic areas [5,6], and surgery for denervating peripheral receptors [7,8]. Utilizing rats is also an alternative when studies involve addressing physiological parameters in human subjects throughout their lifetimes, which is often impractical [2]. From the thermal biology perspective, studies of running rats can, with certain limitations (e.g. differences in evaporative heat loss), help understand some features of exercise thermoregulation in humans [1].

Physical exercise increases whole-body metabolic rate due to augmented chemical-to-mechanical energy transformation in contracting skeletal muscles [9]. In rats, the running-induced increase in metabolic rate can attain levels 2 to 4 times greater than baseline values [10–13], leading to higher heat production. Thus, augmented heat loss to the environment is necessary to counterbalance enhanced heat production and avoid exaggerated core body hyperthermia [14,15]. Interestingly, the increase in heat production precedes the increase in cutaneous heat loss, favoring body heat accumulation and the consequent core body temperature (T_CORE) rise at the beginning of the exercise [1]. After that, the T_CORE level attained by the rats will significantly depend on the ambient temperature (T_AMB) and exercise intensity, duration, and protocol [1,16–18].

Augmented body temperature – either T_CORE or muscular temperature – is associated with physiological benefits during exercise, including higher enzymatic activity [19], reduced saturation of hemoglobin and myoglobin with O₂ molecules [20,21], augmented local blood flow [22], and attenuated increase in blood viscosity [23]. Therefore, some level of hyperthermia is required to promote adequate metabolic responses to physical exertion (e.g. adequate O₂ supply to contracting muscles). However, severe hyperthermia influences the functioning of many physiological systems (e.g. cardiovascular and gastrointestinal systems) [24,25] and is associated with the occurrence of heat-related disorders, including heatstroke [26,27].

Alongside undesirable effects on health, severe hyperthermia might also favor the occurrence of fatigue or exhaustion during prolonged exercises, thus impairing endurance [28–30]. The increase in T_CORE, particularly brain temperature, changes the electroencephalographic activity, increases perceived exertion, and ultimately impairs the ability of the central nervous system to stimulate the contraction of skeletal muscles [31,32]. Moreover, other indirect effects of augmented T_CORE may influence endurance, such as reduced cerebral blood flow [33], enhanced cardiovascular strain [34,35], altered thermal perception [36], and muscle/systemic inflammation [37].

Considering the impacts of severe hyperthermia on an animal’s physical performance and health, we considered it essential to carefully assess the extensive literature reporting the changes in T_CORE in rats exercised to fatigue or exhaustion. By systematically reviewing this literature, we expect to provide relevant insight into the factors (e.g. T_AMB, body mass, and exercise intensity and duration) that may affect the T_CORE value when rats quit exercising on a treadmill. Moreover, identifying what factors affect the increase in T_CORE during exercise will allow the development of more effective strategies to mitigate the hyperthermia-induced degradation of endurance performance and minimize the incidence of severe heat illnesses.

Therefore, we aimed to systematically review the studies that recorded T_CORE at fatigue or exhaustion in rats subjected to incremental- or constant-speed treadmill running. In addition, we analyzed the data extracted from these studies using multiple linear regression analysis to understand what factors affect the T_CORE attained by the rats when they quit exercising.

Methods

Search strategy

The present systematic review and meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [38]. The searched databases were: PubMed (U.S. National Institutes of Health), Web of Science (Clarivate), EMBASE (Elsevier), and SPORTDiscus (EBSCO). The following search terms were combined using Boolean operators: (exercise OR running) AND (thermoregulation OR hyperthermia) AND (fatigue OR exhaustion) AND rat.

The search was conducted in October 2021, with no date restrictions, and all included studies were manuscripts written in English that contained original data. An update search was conducted in August 2022. Books and book chapters, theses, dissertations, review articles, points of view, essays, editorials, and scientific meeting proceedings were not included, though their bibliographic reference lists were consulted during the screening process. We also checked the reference list of all selected manuscripts to ensure that our search strategy has not missed relevant studies.

The following inclusion criteria were set: a) the study sample should consist of rats; b) the intervention should be a running exercise on a treadmill; c) the exercise should have been performed until the rats were fatigued or exhausted; d) the study should have measured any T_CORE index (e.g. abdominal, colonic, or brain temperature), at least when rats were fatigued or exhausted.

Methodological considerations

This systematic review focused on studies involving rats subjected to physical exercise. Although rats and humans have developed distinct strategies to address environmental challenges (e.g. rats have limited evaporative heat loss compared to humans), the exercise-induced adjustments in heat production and dissipation occur in the same direction in both species. Heat production sharply increases with exercise initiation, whereas the convective cutaneous heat loss decreases until the attainment of a threshold T_CORE that triggers skin vasodilation. Thus, the increase in heat production is always faster than the increase in heat loss, raising T_CORE at the beginning of the exercise in both species [1,39].

Treadmill running was chosen because this exercise modality allows for the continuous measurement of thermoregulatory parameters, such as the T_CORE and skin temperature (T_SKIN) [1], and the precise quantification of external work performed by rats [12,40]. Moreover, we focused on the T_CORE at fatigue and exhaustion because it usually corresponds to the highest value attained during exercise and because this was the most reproducible thermoregulatory parameter when rats were subjected to repeated incremental running [41].

The terms “fatigue” and “exhaustion” are often used interchangeably in the literature. However, many authors differentiate them by stating that exhaustion is synonymous with fatigue but more intense [42]; i.e. exhaustion is considered extreme fatigue [1]. In addition, from a medical perspective, fatigue indicates a declining ability to respond to stressors, while exhaustion indicates an almost complete inability to respond to stressors [43].

In experiments with exercising rats, fatigue is usually determined as the moment when the animals cannot keep pace with the treadmill [44,45] and/or expose themselves to electrical stimulation during a predetermined time, which corresponds to 10 s in our experiments [3,46,47]. Using this criterion, we observed that fatigued rats could right themselves when placed on their backs. In contrast, exhaustion is usually confirmed by observing that exhausted rats lose their righting reflex [48–50]. Nevertheless, different criteria exist for defining exhaustion; for instance, in the study performed by Hasegawa et al. [51], “exhaustion was considered to have occurred when the rat was unable to keep pace with the treadmill and lay flat on, and stayed on the grid positioned at the back of the treadmill for a period of 30 s despite gently being pushed with sticks or breathed on.”

The T_CORE (or deep body temperature) would ideally represent the mean temperature of the thermal core [52]. In practice, T_CORE is represented by a specified temperature, such as the abdominal, arterial blood, brain, colonic, or rectal temperature in rats. Of note, the depth at which thermocouples are inserted past the anal sphincter determines whether the rectal or colonic temperature is measured. The thermal core corresponds to those inner tissues of the body whose temperatures are not altered in their relationship to each other by circulatory adjustments and changes in heat dissipation to the environment that affect the thermal shell of the body [52]. Brain temperatures have been considered a T_CORE index [52], although evidence showing the independence of brain temperatures from other indices of T_CORE is available [53]. Rats’ T_CORE can be measured through different methods, including thermocouples, thermistors [54], and telemetry sensors [55].

The thermoregulation of small animal species, including rats, depends largely on environmental conditions [56], and, therefore, these conditions are another methodological aspect deserving attention. In this sense, it is crucial to introduce the thermoneutral zone (TNZ) concept. There are many definitions of the TNZ (for a detailed review, see ref [57]). According to the third Glossary of Terms for Thermal Physiology, the TNZ is defined as “the range of ambient temperature at which temperature regulation is achieved only by control of sensible heat loss, i.e. without regulatory changes in metabolic heat production or evaporative heat loss” [52], and this definition is currently prevalent [57]. Using this definition, Romanovsky et al. [58] developed a practical way to determine the TNZ of rats by measuring the T_SKIN of their principal “heat-loss organ,” the tail, and then calculating the heat loss index (HLI). Briefly, the HLI is a ratio of two temperature gradients (T_SKIN – T_AMB and T_CORE – T_AMB) [59] and varies from 0 (maximal vasoconstriction) to 1 (maximal vasodilation). Within the TNZ, tail-skin vasoconstriction constantly changes for tail-skin vasodilation, and, consequently, high magnitude changes in tail T_SKIN and HLI are observed [58].

When Wistar rats are resting in the treadmill setup at T_AMBs below 24°C, their tails are vasoconstricted [46,60], and the HLI stabilizes at levels below < 0.1 (sub-neutral conditions). On the other hand, we observed HLI values ranging from 0.2 to 0.3 in resting rats maintained inside the chamber that contained the treadmill belt at a local temperature of 24–26°C [61–63], thus suggesting this T_AMB range corresponds to the lower end of the TNZ of rats inside the treadmill setup. Moreover, at 30°C, the HLI varies largely between 0.2 to 0.4, indicating thermoneutral conditions [60]. Finally, at 35°C, rats’ tails are fully vasodilated [61,64], with HLI values consistently above 0.5 (supra-neutral conditions). Therefore, the TNZ of Wistar rats resting inside the treadmill setup is located between 23°C and 35°C, most likely close to 30°C.

It is noteworthy that exercise-induced T_CORE increase is exaggerated at 31–32°C, as compared to 23–24°C [65–67], and no steady-state T_CORE is attained while running under these warm environmental conditions (i.e. 31–32°C). Indeed, no previous study tried to determine a range of T_AMB at which T_CORE regulation in an exercising rat is achieved only by control of sensible heat loss. Even when T_CORE remained unaltered throughout a running session in the cold, rats’ tails were clearly vasoconstricted [46]. Moreover, environmental factors other than T_AMB may influence the TNZ inside the treadmill chamber, such as the wind speed (e.g. whether the equipment has an electrical fan in front of the treadmill belt) or relative humidity (e.g. rats increase evaporative heat loss from the respiratory tract in proportion to exercise intensity [17]). However, because wind speed and relative humidity were not reported in most experiments we reviewed, the T_AMB was the only environmental factor included as an independent variable in the regression analyses.

Study selection

Studies were searched and inserted into the Rayyan web application (https://rayyan.qcri.org). The selection was carried out in two stages that involved reading the titles/abstracts and full texts. Next, studies were screened for inclusion by three researchers (MA, NB, and SW), with all cases of disagreement discussed until consensus was reached. Finally, inclusion and exclusion decisions were duly labeled in the Rayyan web application [68].

The search process retrieved a total of 329 articles. One hundred and eighty-four of these articles were excluded for being duplicates. Then, the title and abstract of the 145 remaining articles were read, and 53 were excluded at this stage. Next, 92 articles were selected for full reading, with 20 of these articles being excluded at this final stage; the reasons for exclusions are presented in Figure 1. In the end, 72 studies published between 1968 and 2022 were included in the systematic review (Figure 1).

Figure 1. Flowchart of the process of identification, selection, eligibility, and inclusion of studies investigating the core body temperature in rats subjected to treadmill running until they were fatigued or exhausted. The screening process is based on the PRISMA flow diagram. Briefly, inadequate publication type included review articles and proceedings abstracts, whereas inadequate study design included other exercise types (i.e. swimming and running wheel), exercise with predetermined duration (not to fatigue or exhaustion), experiments with isolated tissues, and using heat-acclimated rats.

Data extraction

After selection, the following information was extracted from each article included in the systematic review: author, year of publication, rat strain, sex, body mass, age, exercise protocol, T_AMB, daytime/period when experiments were carried out, T_CORE index measured, time to fatigue/exhaustion, distance traveled, initial T_CORE, and T_CORE at fatigue or exhaustion. The treadmill speed was also extracted from the manuscripts involving constant-speed exercises. We always extracted the information as provided by the authors, even though literature provides different definitions for fatigue and exhaustion and for the site of T_CORE measurement (e.g. rectal vs. colonic).

The data were mainly extracted from the control trials. We did not use data from rats subjected to pharmacological and behavioral interventions (e.g. manipulation of brain neurotransmission and sleep deprivation, respectively) or from trained or heat-acclimated rats because these conditions affect exercise thermoregulation. For example, aerobic training improves thermoregulatory efficiency during treadmill running, as trained rats run farther distances to attain a similar T_CORE at fatigue compared to untrained rats [67,69]. Experiments with fasted rats were included in the systematic review if they were not fasted for more than 24 h. As reported earlier, fasting was used to standardize the metabolic and hormonal parameters before exercise and minimize their influence on performance [70]. Moreover, 24-h fasting did not influence endurance in low-intensity exercise under environmental heat stress [71].

When not available in the text or tables, the temperature and performance data were extracted from figures using the WebPlotDigitizer (version 4.5, https://automeris.io/WebPlotDigitizer; Pacifica, CA, USA) or were obtained through contact with the correspondent authors via electronic message.

The extracted data were divided into two tables: data from constant-speed (Table 1) and incremental-speed (Table 2) exercises, both to fatigue/exhaustion. The exercises with at least three different speed stages were considered incremental exercises. For example, Lubbe et al. [72] subjected the rats to an exercise starting at a speed of 10 m/min for 5 min, after which the speed was suddenly increased to 30 m/min; therefore, we considered the initial 5 min as a warm-up period and running at 30 m/min as a constant-speed exercise to exhaustion. We separated the data from these two protocols due to the marked differences in the evolution of running speed, which impacts the evolution of exercise intensity and, ultimately, the mechanisms underpinning fatigue/exhaustion. In this sense, metabolic factors are likely more critical for regulating fatigue than thermoregulatory factors during incremental exercises [18].

Table 1. Studies that investigated the core body temperature at fatigue/exhaustion in rats subjected to incremental-speed exercises on a treadmill.

Authors(year)	Number, sex, and strain	Body mass(g)	Time to fatigue (min)	Distance traveled (m)	Ambient temperature (°C)	Time of the day(h)	TCORE measured	TCORE at exercise initiation (°C)	TCORE at fatigue(°C)
Gollnick & Ianuzzo (1968)	5 maleSprague-Dawley	332–343	–	–	24 ± 1	13 to 16	Colonic	39.4 ± 0.2	41.8 ± 0.3
Fruth & Gisolfi (1983)	20 male Sprague-Dawley (survivors)	250–400	68.0 ± 5.3	1521	27.1 (from 23 to 35)	NR	Colonic	37.8	41.4 ± 0.1
	14 male Sprague-Dawley(fatalities)	250–400	79.9 ± 2.8	1842	28.2 (from 23 to 35)	NR	Colonic	37.9	42.4 ± 0.2
Ardévol et al. (1999)	6 female lean Zucker	175–220	9.1 ± 0.4	255	27.8	NR	Aortic	36.7 ± 0.1	37.4 ± 0.1
Balthazar et al. (2009)	6 male Wistar	250–300	17 ± 3	210	22 ± 1	13 to 17	Abdominal	37.7 ± 0.2	39.1
Balthazar et al. (2010)	6 male Wistar	250–300	22.0 ± 3.1	290	22 ± 1	13 to 17	Abdominal	37.6 ± 0.4	38.5
Fonseca et al. (2014)	8 male Wistar	250–365	42.9 ± 1.5	868	12	13 to 18	Abdominal	37.4 ± 0.1	37.3 ± 0.2
	8 male Wistar	250–365	33.2 ± 3.6	591	25	13 to 18	Abdominal	37.3 ± 0.1	38.6 ± 0.2
Kunstetter et al. (2014)	7 male Wistar	250–350	59.5 ± 4.4	1156	25.2 ± 0.2	10 to 16	Brain (cortical)	37.2 ± 0.1	39.3 ± 0.3
Machado et al. (2015)	11 male Wistar	297 ± 5	52.1 ± 2.5	948	23	6	Abdominal	37.0 ± 0.1	40.0 ± 0.1
	11 male Wistar	297 ± 5	47.0 ± 2.0	815	23	20	Abdominal	38.4 ± 0.1	40.3 ± 0.2
Drummond et al. (2016)	8 male Wistar	383 ± 11	41.0 ± 3.4	670	25	7 to 13	Abdominal/Brain (cortical)	36.8 ± 0.1 /36.9 ± 0.1	39.4 ± 0.1 /38.6 ± 0.2
	8 male Wistar	383 ± 11	28.5 ± 1.6	407	32	7 to 13	Abdominal/Brain (cortical)	37.0 ± 0.1 /36.9 ± 0.1	40.0 ± 0.1 /39.5 ± 0.2
Machado et al. (2016)	11 male Wistar	342 ± 11	46.5 ± 2.3	803	24 ± 1	6	Abdominal	37.1 ± 0.2	39.8 ± 0.3
	11 male Wistar	340 ± 10	40.5 ± 2.3	659	24 ± 1	20	Abdominal	38.2 ± 0.2	40.1 ± 0.2
Morozova et al. (2016)	6 maleSprague-Dawley	300 ± 20	11.9	163	24	10 to 16	Abdominal	38.3	39.4
	6 maleSprague-Dawley	300 ± 20	14.8 ± 0.8	235	24	10 to 16	Abdominal	38.3	39.9 ± 0.2
Santiago et al. (2016)	8 male Wistar	350–370	38.6 ± 1.9	615	23 ± 1	NR	Abdominal	37.6 ± 0.1	38.7 ± 0.1
Authors(year)	Number, sex, and strain	Body mass(g)	Time to fatigue (min)	Distance traveled (m)	Ambient temperature (°C)	Time of the day(h)	TCORE measured	TCORE at exercise initiation (°C)	TCORE at fatigue(°C)
Kunstetter et al. (2018)	6 male Wistar	344 ± 18	43.8 ± 2.4	736	23 to 24	NR	Abdominal	37.0 ± 0.2	39.7 ± 0.2
	6 male Wistar	340 ± 16	45.2 ± 2.7	770	23 to 24	NR	Abdominal	38.0 ± 0.3	40.1 ± 0.2
Bittencourt et al. (2020)	8 male Wistar	287 ± 7	45.9 ± 2.9	788	32.1 ± 0.0	NR	Colonic	37.3 ± 0.1	42.1 ± 0.2
	8 male Wistar	287 ± 7	47.3 ± 2.4	823	32.1 ± 0.0	NR	Colonic	37.3 ± 0.1	41.9 ± 0.3
	8 male Wistar	287 ± 7	49.5 ± 4.3	879	32.1 ± 0.1	NR	Colonic	37.0 ± 0.1	41.6 ± 0.3
Nunan et al. (2021)	12 male Wistar	270–300 g	37.9 ± 1.9	600	25	Light phase	Abdominal	37.0 ± 0.1	39.4 ± 0.2
Shang et al. (2021)	9 male Wistar	250–350	41.6 ± 2.6	889	24	19 to 21	Colonic	37.9 ± 0.1	40.3 ± 0.3
	9 male Wistar	250–350	43.6 ± 2.5	951	24	19 to 21	Colonic	38.0 ± 0.1	40.4 ± 0.2
	9 male Wistar	250–350	32.7 ± 2.3	636	31	19 to 21	Colonic	38.1 ± 0.2	41.7 ± 0.3
	9 male Wistar	250–350	39.1 ± 2.4	814	24	19 to 21	Colonic	38.1 ± 0.2	40.4 ± 0.1
	9 male Wistar	250–350	31.8 ± 1.8	612	31	19 to 21	Colonic	38.3 ± 0.1	41.8 ± 0.2
Teixeira-Coelho et al. (2021)	8 male Wistar	220–250 g	39.3 ± 2.8	630	23	NR	Abdominal	37.1 ± 0.1	39.5 ± 0.2
	8 male Wistar	220–250 g	33.6 ± 2.5	508	32	NR	Abdominal	37.1 ± 0.1	40.4 ± 0.3
Andrade et al. (2022)	20 male Wistar	266.1 ± 19.8	55.1 ± 8.3	1041	24	12:30 to 17:30	Abdominal	36.7 ± 0.4	39.6 ± 0.8
	20 male Wistar	272.0 ± 18.8	57.1 ± 8.6	1099	24	12:30 to 17:30	Abdominal	36.9 ± 0.5	39.9 ± 0.8
	8 male Wistar	339.4 ± 30.0	57.4 ± 5.5	1099	24	12:30 to 17:30	Abdominal	37.0 ± 0.3	40.1 ± 1.0
	8 male Wistar	340.9 ± 29.0	47.6 ± 7.1	838	31	12:30 to 17:30	Abdominal	36.9 ± 0.3	40.9 ± 0.7

Data are expressed as means ± standard deviations or standard errors. Legend: NR = not reported. The data in italics were kindly provided by the corresponding/first authors, correspond to estimated values calculated from data included in the manuscripts, or were extracted from figures using the WebPlotDigitizer software.

Table 2. Studies that investigated the core body temperature at fatigue/exhaustion in rats subjected to constant-speed exercises on a treadmill.

Authors(year)	Number, sex, and strain	Body mass(g)	Time to fatigue (min)	Distance traveled (m)	Ambient temperature (°C)		Time of the day(h)	TCORE measured	TCORE at exercise initiation (°C)	TCORE at fatigue(°C)
Hubbard et al. (1976)	30 male Sprague-Dawley	484 ± 15	100 ± 30	1,100	5		NR	Rectal	38.4 ± 0.6	39.7 ± 0.7
	22 male Sprague-Dawley	488 ± 17	57 ± 23	627	20, 23 or 26		NR	Rectal	38.1 ± 0.7	41.5 ± 0.2
	41 male Sprague-Dawley	486 ± 15	55 ± 22	605	20, 23 or 26		NR	Rectal	38.4 ± 0.6	41.9 ± 0.4
Hubbard et al. (1977)	21 male Sprague-Dawley	485–545	57 ± 30	627	5, 20, 23 or 26		NR	Rectal	NR	41.1 ± 0.4
	16 male Sprague-Dawley	485–545	58 ± 19	638	5, 20, 23 or 26		NR	Rectal	NR	41.5 ± 0.3
	19 male Sprague-Dawley	485–545	50 ± 16	550	5, 20, 23 or 26		NR	Rectal	NR	41.8 ± 0.3
	22 male Sprague-Dawley	485–545	50 ± 12	550	5, 20, 23 or 26		NR	Rectal	NR	42.0 ± 0.4
	40 male Sprague-Dawley	485–545	65 ± 23	715	5, 20, 23 or 26		NR	Rectal	NR	42.3 ± 0.5
Francesconi & Mager (1979a)	6 male Sprague-Dawley	280–350	25.8 ± 1.4	236	35 ± 1		NR	Rectal	38.2	42.9 ± 0.1
	6 male Sprague-Dawley	280–350	25.1 ± 1.1	229	35 ± 1		NR	Rectal	37.4	42.7 ± 0.1
Francesconi & Mager (1979b)	27 male Sprague-Dawley	300–400	NR	NR	34–35		NR	Rectal	NR	42.7 ± 0.1
Hubbard et al. (1979)	13 male Sprague-Dawley	485 ± 12	145.3	1,598	5		8	Rectal	NR	38.2 ± 1.1
	57 male Sprague-Dawley	485 ± 12	68.1	749	20, 26 or 30		8	Rectal	NR	41.5 ± 1.0
Francesconi & Mager (1980)	13 male Sprague-Dawley	280–340	32.9 ± 1.7	301	35		NR	Rectal	37.0	42.6 ± 0.1
Francesconi & Mager (1981a)	13 male Sprague-Dawley	250–350	32.9 ± 1.7	301	35 ± 0.5		NR	Rectal	37.6	42.6 ± 0.1
Francesconi & Mager (1981b)	16 male Sprague-Dawley	250–350	30.5 ± 2.3	279	35 ± 0.5		NR	Rectal	38.4	42.9 ± 0.3
Francesconi & Mager (1981c)	12 male Sprague-Dawley	275–350	32.3	245	35 ± 0.5		NR	Rectal	37.4	42.41
Hubbard et al. (1981)	12 male Sprague-Dawley	250–300	113.3	1,246	15		NR	Rectal	37.9 ± 0.7	39.9 ± 0.5
Authors(year)	Number, sex, and strain	Body mass(g)	Time to fatigue (min)	Distance traveled (m)	Ambient temperature (°C)		Time of the day(h)	TCORE measured	TCORE at exercise initiation (°C)	TCORE at fatigue(°C)
Hubbard et al. (1981)	12 male Sprague-Dawley	250–300	117.7	1,295	20		NR	Rectal	37.9 ± 0.7	41.1 ± 0.5
Francesconi et al. (1983a)	12 male Sprague-Dawley	310–325	37.0	338	35		NR	Rectal	NR	43.0
Francesconi et al. (1983b)	22 males	300–325	34.7 ± 1.4	317	35 ± 0.5		NR	Rectal	38.7	43.4
Francesconi & Hubbard (1985)	16 male Sprague-Dawley	322.7 ± 4.5	31.6	289	35.5		NR	Rectal	38.3	43.0
	16 male Sprague-Dawley	297.3 ± 3.7	35.1	321	35.5		NR	Rectal	38.5	42.1
Durkot et al. (1986)	16 male Sprague-Dawley	237 ± 3.0	221 ± 26	2,431	26		6 to 12	Rectal	37.5	41.7 ± 0.2
	16 male Sprague-Dawley	256 ± 3.0	108 ± 10	1,188	26		6 to 12	Rectal	38.0	42.3 ± 0.1
	16 male Sprague-Dawley	503 ± 3.0	67 ± 7	737	26		6 to 12	Rectal	38.2	42.0 ± 0.1
	16 male Sprague-Dawley	504 ± 3.0	54 ± 4	594	26		6 to 12	Rectal	38.6	41.9 ± 0.1
Francesconi & Hubbard (1986)	16 male Sprague-Dawley	320 ± 4.7	31.9 ± 1.1	292	35.5 ± 0.5		NR	Rectal	38.4 ± 0.1	43.0
Matthew et al. (1987)	10 male Sprague-Dawley	510–530	53	583	26		NR	Rectal	38.6 ± 0.2	41.6 ± 0.2
Matthew et al. (1990)	8 male Sprague-Dawley	510–530	62 ± 8	682	26		NR	NR	38.8 ± 0.2	42.1
Caputa & Kamari (1991)	12 male/female Wistar	NR	62.1 ± 5.4	1,552	18 → 5		NR	Rectal	37.2	37.7
	12 male/female Wistar	NR	62.1 ± 5.4	1,552	18 → 5		NR	Brain	37.5	37.8
	12 male/female Wistar	NR	16.2 ± 1.2	405	45 → 27		NR	Rectal	40.4	41.6
	12 male/female Wistar	NR	16.2 ± 1.2	405	45 → 27		NR	Brain	40.3	41.4
Durkot et al. (1992)	12 male Sprague-Dawley	362 ± 7	70.3 ± 3.3	773	30		NR	Rectal	37.2	42.1
Matthew et al. (1992)	12 male Sprague-Dawley	510–530	67 ± 6	737	26		NR	Rectal	38.8 ± 0.1	41.8 ± 0.2
Matthew et al. (1993)	12 male Sprague-Dawley	510–530	80.1	881	10		NR	Rectal	38.8 ± 0.1	39.4 ± 0.3
Authors(year)	Number, sex, and strain	Body mass(g)	Time to fatigue (min)	Distance traveled (m)	Ambient temperature (°C)		Time of the day(h)	TCORE measured	TCORE at exercise initiation (°C)	TCORE at fatigue(°C)
Matthew et al. (1993)	12 male Sprague-Dawley	510–530	90.4	994	15		NR	Rectal	39.0 ± 0.1	40.5 ± 0.3
	12 male Sprague-Dawley	510–530	67.6	744	26		NR	Rectal	38.8 ± 0.1	41.8 ± 0.2
	12 male Sprague-Dawley	510–530	30.0	330	30		NR	Rectal	39.0 ± 0.1	42.8 ± 0.1
Durkot et al. (1995)	20 male Sprague-Dawley	300–400	101.3 ± 5.0	1,114	26		NR	Rectal	38.5 ± 0.1	41.3 ± 0.2
Fuller et al. (1998)	14 male Sprague-Dawley	350–450	29.4 ± 5.9	441	33		15.5	Abdominal	37.7 ± 0.6	39.9 ± 0.3
	14 male Sprague-Dawley	350–450	29.4 ± 5.9	441	33		15.5	Brain	37.8 ± 0.6	40.2 ± 0.4
	14 male Sprague-Dawley	350–450	22.1 ± 3.7	332	38		15.5	Abdominal	37.8 ± 0.5	39.9 ± 0.3
	14 male Sprague-Dawley	350–450	22.1 ± 3.7	332	38		15.5	Brain	37.9 ± 0.6	40.2 ± 0.4
	14 male Sprague-Dawley	350–450	14.3 ± 2.9	215	38		15.5	Abdominal	38.2 ± 0.2	39.8 ± 0.3
	14 male Sprague-Dawley	350–450	14.3 ± 2.9	215	38		15.5	Brain	38.3 ± 0.2	40.1 ± 0.4
Moran et al. (1999)	8 male Zabar	300–350	14 ± 1	350	40		NR	Colonic	NR	40.9 ± 0.2
Durkot et al. (2000)	20 male Sprague-Dawley	350	41 ± 4	451	30		NR	Rectal	38.1 ± 0.1	41.3 ± 0.2
Walters et al. (2000)	11 male Sprague-Dawley	380–390	38.1 ± 5.5	686	35 ± 1		12 to 13	Rectal	39.5 ± 0.1	42.4 ± 0.1
	11 male Sprague-Dawley	380–390	38.1 ± 5.5	686	35 ± 1		12 to 13	Brain	39.4 ± 0.1	42.2 ± 0.2
	11 male Sprague-Dawley	380–390	35.8 ± 3.9	644	35 ± 1		12 to 13	Rectal	39.0 ± 0.2	42.5 ± 0.1
	11 male Sprague-Dawley	380–390	35.8 ± 3.9	644	35 ± 1		12 to 13	Brain	39.0 ± 0.2	42.2 ± 0.2
Rodrigues et al. (2003)	6 male Wistar	260–360	57.7 ± 8.6	1,212	22		NR	Abdominal	37.2 ± 0.2	39.3 ± 0.3
	6 male Wistar	260–360	41.6 ± 8.1	874	28		NR	Abdominal	37.6 ± 0.2	39.9 ± 0.3
	6 male Wistar	260–360	28.3 ± 5.4	594	35		NR	Abdominal	37.4 ± 0.1	41.3 ± 0.2
	6 male Wistar	260–360	38.3 ± 5.8	919	22		NR	Abdominal	37.2 ± 0.1	39.0 ± 0.2
	6 male Wistar	260–360	25.8 ± 3.2	619	28		NR	Abdominal	37.7 ± 0.2	40.2 ± 0.2
Authors(year)	Number, sex, and strain	Body mass(g)	Time to fatigue (min)	Distance traveled (m)	Ambient temperature (°C)		Time of the day(h)	TCORE measured	TCORE at exercise initiation (°C)	TCORE at fatigue(°C)
Rodrigues et al. (2003)	6 male Wistar	260–360	17.2 ± 2.2	413	35		NR	Abdominal	37.3 ± 0.2	40.5 ± 0.1
Rodrigues et al. (2004)	8 male Wistar	250–300	52.5 ± 1.8	1,050	23 ± 2		10 to 14	Colonic	37.3 ± 0.1	39.1 ± 0.1
Soares et al. (2004)	6 male Wistar	270 ± 20	40 ± 3	720	23 ± 2		10 to 14	Abdominal	37.5 ± 0.2	38.0 ± 0.3
Lacerda et al. (2005)	6 male Wistar	250–340	92.7	1,667	21 ± 2		13 to 17	Colonic	37.3	39.2 ± 0.1
Leite et al. (2006)	15 male Wistar	240–330	52.5 ± 6.7	945	22 ± 2		10 to 14	Abdominal	37.6	38.7 ± 0.1
Pires et al. (2007)	8 male Wistar	250–300	27.8 ± 5.4	667	26 ± 1		9 to 16	Abdominal	37.5 ± 0.2	38.5 ± 0.4
Wanner et al. (2007)	8 male Wistar	250 − 300	33.5 ± 3.4	804	23 ± 1		NR	Abdominal	36.8 ± 0.4	38.0
	4 male Wistar	250 − 300	23.7 ± 2.5	569	23 ± 1		NR	Abdominal	36.8 ± 0.2	37.8
Hasegawa et al. (2008)	8 male Wistar	300–350	143.6 ± 21	3,734	18		NR	Abdominal	37.8 ± 0.8	39.6 ± 0.6
	8 male Wistar	300–350	143.6 ± 21	3,734	18		NR	Brain	37.9	39.1 ± 0.8
	8 male Wistar	300–350	65.8 ± 13	1,711	30		NR	Abdominal	37.6 ± 0.8	41.0 ± 0.7
	8 male Wistar	300–350	65.8 ± 13	1,711	30		NR	Brain	37.8	40.5 ± 0.7
Rodrigues et al. (2008)	7 male Wistar	300 ± 20	53.2 ± 2.5	1,064	23 ± 2		10 to 14	Colonic	37.3 ± 0.1	38.9 ± 0.1
Leite et al. (2010)	6 male Wistar	290 ± 10	70.7 ± 6.6	1,273	22 ± 2		10 to 14	Abdominal	37.3	38.7 ± 0.1
Pires et al. (2010)	23 male Wistar	250–350	51 ± 4	918	26		9 to 16	Abdominal	38.0 ± 0.1	38.8 ± 0.1
	5 male Wistar	250–350	45.8	824	26		9 to 16	Abdominal	38.0	38.8
Guimarães et al. (2013)	9 male Wistar	250–350	74.7 ± 12.8	1,494	8		NR	Abdominal	37.2 ± 0.4	36.5 ± 0.4
	9 male Wistar	250–350	63.9 ± 7.1	1,278	12		NR	Abdominal	37.2 ± 0.3	37.1 ± 0.4
	9 male Wistar	250–350	61.3 ± 6.8	1,226	15		NR	Abdominal	37.0 ± 0.3	37.6 ± 0.4
	7 male Wistar	250–350	53.7 ± 6.7	1,074	5		NR	Abdominal	36.0 ± 0.3	34.9
	7 male Wistar	250–350	49.3 ± 3.7	986	12		NR	Abdominal	37.1 ± 0.3	37.3 ± 0.3
	6 male Wistar	250–350	50.7 ± 4.6	1,014	15		NR	Abdominal	36.8 ± 0.2	37.4
Pires et al. (2013)	6 male Wistar	280–350	66 ± 11	1,188	25		8 to 16	Abdominal	37.5	38.5
	6 male Wistar	280–350	23 ± 2.0	414	35		8 to 16	Abdominal	37.7	40.9 ± 0.1
Campos et al. (2014)	8 male Wistar	350–450	47.0 ± 7.1	630	25		7 to 12	Abdominal	36.9	38.9
	8 male Wistar	350–450	15.8 ± 1.0	211	32		7 to 12	Abdominal	37.1	39.1
Authors(year)	Number, sex, and strain	Body mass(g)	Time to fatigue (min)	Distance traveled (m)	Ambient temperature (°C)	Time of the day(h)		TCORE measured	TCORE at exercise initiation (°C)	TCORE at fatigue(°C)
Cordeiro et al. (2014)	6 male Wistar	250–300	70.7 ± 20.8	1,273	23	NR		Abdominal	37.5 ± 0.1	38.5 ± 0.1
Fonseca et al. (2014)	6 male Wistar	250–365	67.1 ± 13.7	1,342	12	13 to 18		Abdominal	37.3 ± 0.1	37.7 ± 0.2
	6 male Wistar	250–365	67.1 ± 13.7	1,342	12	13 to 18		Brain	37.4 ± 0.3	37.7 ± 0.2
	7 male Wistar	250–365	75.8 ± 13.2	1,516	12	13 to 18		Abdominal	37.2	37.4
	7 male Wistar	250–365	75.8 ± 13.2	1,516	12	13 to 18		Brain	37.4	37.5
	7 male Wistar	250–365	31.5 ± 5.6	630	25	13 to 18		Abdominal	37.4 ± 0.1	38.8 ± 0.2
	7 male Wistar	250–365	31.5 ± 5.6	630	25	13 to 18		Brain	37.7 ± 0.1	39.2 ± 0.1
	7 male Wistar	250–365	38.4 ± 6.1	768	25	13 to 18		Abdominal	37.5	39.0
	7 male Wistar	250–365	38.4 ± 6.1	768	25	13 to 18		Brain	37.7	39.2
Kunstetter et al. (2014)	9 male Wistar	250–350	207 ± 15	3,726	25.2 ± 0.2	10 to 16		Brain	37.2 ± 0.1	40.5 ± 0.3
	9 male Wistar	250–350	149 ± 14	3,129	25.2 ± 0.2	10 to 16		Brain	37.3 ± 0.1	40.3 ± 0.3
	9 male Wistar	250–350	62 ± 11	1,488	25.2 ± 0.2	10 to 16		Brain	37.2 ± 0.1	40.1 ± 0.2
Lima et al. (2014)	10 male Wistar	240–330	61.2 ± 8.4	1,102	23 ± 1	7 to 12		Abdominal	37.4 ± 0.1	39.5 ± 0.2
Zheng et al. (2014)	10 male Wistar	300–350	104.4 ± 30.9	1,879	23	NR		Abdominal	37.6 ± 0.2	39.8 ± 0.6
Damasceno et al. (2015)	6 male Wistar	250–350	113.3 ± 25.0	2,379	24 ± 1	7 to 11		Abdominal	36.9	39.5
	6 male Wistar	250–350	113.3 ± 25.0	2,379	24 ± 1	7 to 11		Brain	37.6	39.6
Wanner et al. (2015)	14 male Wistar	250–280	27.2 ± 1.8	490	24.1 ± 0.4	NR		Abdominal	37.9 ± 0.1	38.9 ± 0.1
Drummond et al. (2016)	8 male Wistar	383 ± 11	177.5 ± 8.6	2,485	25	7 to 13		Abdominal	36.8	38.9 ± 0.2
	8 male Wistar	383 ± 11	177.5 ± 8.6	2,485	25	7 to 13		Brain	36.8	38.4 ± 0.2
	8 male Wistar	383 ± 11	78.8 ± 11.8	1,103	32	7 to 13		Abdominal	36.9	40.7
	8 male Wistar	383 ± 11	78.8 ± 11.8	1,103	32	7 to 13		Brain	36.9	40.2
Zheng et al. (2016)	10 male Wistar	300–350	91 ± 14.1	1,638	23	NR		Abdominal	37.8	39.7 ± 0.5
Malheiros-Lima et al. (2018)	14 male Wistar	307	39 ± 3	702	25	NR		Abdominal	37.1	38.7 ± 0.2
Zaretsky et al. (2018)	7 male Sprague-Dawley	300 ± 20	15.2 ± 1.8	274	32 ± 1	10 to 16		Abdominal	37.6 ± 0.1	40.0 ± 0.4
Gomes et al. (2019)	8 male Wistar	390–463	21.9 ± 1.91	350	25	NR		Abdominal	36.4	39.8
Lima et al. (2019)	7 male Wistar	250–350	287.0 ± 14.7	4,908	23 ± 1	8 to 12		Abdominal	37.0 ± 0.5	38.6 ± 0.3
Rabelo et al. (2019)	7 male Wistar	240–280	49.0 ± 8.4	490	24 ± 1	8 to 12		Abdominal	36.2 ± 0.1	38.1 ± 0.3
	6 male Wistar	240–280	109.4 ± 8.3	1,422	24 ± 1	8 to 12		Abdominal	36.3 ± 0.1	38.1 ± 0.5
	7 male Wistar	240–280	224.6 ± 26.7	4,267	24 ± 1	8 to 12		Abdominal	36.6 ± 0.4	39.1 ± 0.2
Wang et al. (2019)	8 male Wistar	270 ± 30	68.9 ± 10.0	1,446	38 to 40	NR		Rectal	37.1	40.8 ± 0.4
Rodovalho et al. (2020)	7 male Wistar	280–330	37.9 ± 10.0	682	NR	NR		Abdominal	37.9	38.9
Lubbe et al. (2021)	12 male Sprague-Dawley	150 ± 10	5.59	168	NR	NR		Rectal	33.5 ± 0.5	35.4 ± 1.0
	12 female Sprague-Dawley	150 ± 10	8.83	265	NR	NR		Rectal	34.8 ± 1.0	37.1 ± 0.8

Data are expressed as means ± standard deviations or standard errors. Legend: NR = not reported. The data in italics were kindly provided by the corresponding/first authors, correspond to estimated values calculated from data included in the manuscripts, or were extracted from figures using the WebPlotDigitizer software.

Analysis of the studies’ quality

The quality of the articles included in the systematic review was determined using a qualitative assessment consisting of 13 questions, with answer categories of “no” and “yes,” scoring 0 and 1, respectively. Some questions also had an intermediate answer category (i.e. “partial”), corresponding to a score of 0.5. All questions and the assessment criteria are described in detail in Supplementary file 1. This quality assessment was elaborated by the first (MA) and the last (SW) authors and then improved by the other collaborators; it is worth noting that some of the current study’s authors have extensive experience studying exercise thermoregulation and performance in rats. Pearson’s correlation coefficient was used to assess the strength of the association between the years of publication and the studies’ quality.

Multiple linear regression analysis

This analysis was performed to understand the contribution of different parameters on the T_CORE attained by rats at fatigue or exhaustion during treadmill running. Variables related to the regression model (adjusted R² and standard error of the estimate), the regression coefficients, and the beta weights are reported in the Results section. It is noteworthy that the regression coefficients cannot be used to establish the relative importance of specific variables within a regression equation because they are based on different units of measurement [73]. Therefore, the standardized regression coefficients (i.e. converted to z-scores), also called beta weights, were presented to provide the reader with a complete and practical interpretation of the observed relationships [73]. The analyses were performed using the IBM SPSS Statistics software (version 19.0, International Business Machines Corporation, Armonk – NY, USA). The significance level was set at p < 0.05.

The following parameters were evaluated when analyzing the incremental exercises: body mass, T_AMB, distance traveled, and initial T_CORE. When the manuscript reported a range for rats’ body mass, the highest value was used in the regression analysis because most studies exercised the rats at the end of the experimental design. The model did not include daytime and time to fatigue because these parameters directly affect the initial T_CORE and distance traveled, respectively. In addition, strain, sex, exercise protocol, and T_CORE index were not included because these are categorical parameters. We performed an additional analysis that considered only the most common conditions observed in the systematic review to ensure that some categorical parameters (e.g. rat strain and T_CORE index) have not influenced our results. Therefore, the additional analysis included male Wistar rats with their abdominal temperature measured while being subjected to an incremental exercise with an initial speed of 10 m/min and increases of 1 m/min every 3 min.

When analyzing the constant-speed exercises, we included treadmill speed and exercise duration instead of distance traveled in the multiple linear regression analysis. This is because faster treadmill speeds are commonly associated with shorter durations and lower distances traveled during constant exercises, whereas faster speeds are usually a prerequisite to traveling farther distances during incremental exercises. In addition, although the rats ran at 18 m/min in 16 studies (please see the Results section), the experimental conditions were highly variable between these investigations (e.g. rat strain, treadmill incline, and measured T_CORE). This heterogeneity precluded us from performing an additional analysis with sufficient trials and considering only the most common conditions in the studies involving constant-speed exercises.

Results

Incremental-speed exercise

Eighteen studies, including 35 experimental trials, investigated the T_CORE of rats subjected to incremental exercises until they were fatigued/exhausted (Table 1). Among them, seven studies provided data from one experimental trial [18,69,74–78], eight studies from two trials [47,67,79–84], and one study from three [85], four [41] or five trials [66].

Data from 324 rats were included in this analysis, with an average of 9 ± 4 (mean ± SD) rats per experimental trial. The animals subjected to exercise were Wistar male rats in most investigations, except for three studies that used Sprague-Dawley male rats [74,79,83] and one study that used lean Zucker female rats [75]. Of note, the latter study [75] was the only one investigating the thermoregulatory responses of female rats subjected to an incremental exercise to fatigue/exhaustion. The average body mass corresponded to 328 ± 45 g, but it is not always clear whether these values were recorded on the day rats exercised (Table 3).

Table 3. Data extracted from studies assessing the core body temperatures of rats subjected to an exercise session to fatigue or exhaustion.

	Incremental exercises		Constant exercises
Parameter	Mean ± SD	Min. – Max.	Mean ± SD	Min. – Max.
Body mass (g)	328 ± 45	220–400	375 ± 83	150–545
Time to fatigue (min)	40.8 ± 14.9	9.1–79.9	63.7 ± 47.8	5.6–287.0
Running speed (m/min)	NA	NA	17 ± 5	9–30
Distance traveled (m)	749 ± 344	163–1,842	1,037 ± 869	168–4,908
TAMB (°C)	25.7 ± 4.2	12.0–32.1	25.4 ± 8.2	5.0–40.0
Initial TCORE (°C)	37.5 ± 0.6	36.7–39.4	37.6 ± 0.9	33.5–40.4
TCORE at fatigue/exh. (°C)	40.0 ± 1.2	37.3–42.4	40.0 ± 1.8	34.9–43.4

Legend: exh. = exhaustion; Max. = maximum value; Min. = minimum value; NA = not applicable; SD = standard deviation; T_AMB = ambient temperature; T_CORE = core temperature. n = 35 and 101 trials for the incremental and constant exercises, respectively.

Concerning the exercise protocol, 12 studies reported incremental treadmill running with an initial speed of 10 m/min and 1 m/min increments every 3 min (Supplementary Table 1). However, Fonseca et al. [47] and Shang et al. [66] used slightly different protocols. In the first manuscript, the initial speed and the increments were the same; however, the speed was increased every 2 min [47]. In the second manuscript, the rats were subjected to a treadmill running with an initial speed of 13 m/min and 1.3 m/min increments every 3 min [66]. Finally, Gollnick & Ianuzzo [74] did not precisely describe the time to exhaustion and the duration of the running stages, which corresponded in their experiments to the time elapsed until the first and second plateau of T_CORE. Therefore, this study was not included in the multiple linear regression analysis.

The rats ran on average for 40.8 ± 14.9 min and traveled a distance of 749 ± 344 m. Fruth & Gisolfi [79] and Ardévol et al. [75] reported the longest and shortest exercise duration (79.9 and 9.1 min), respectively. The farthest distance traveled corresponded to 1,842 m and was observed by Fruth & Gisolfi [79], whereas rats covered the shortest distance (i.e. 163 m) in the study by Morozova et al. [83]. The dry T_AMB ranged from 12°C [47] to 32°C [67,85], with the average temperature corresponding to 25.7 ± 4.2°C.

The abdominal temperature was the T_CORE index in most studies included in this systematic review. In contrast, Gollnick & Ianuzzo [74], Fruth & Gisolfi [79], Bittencourt et al. [85], and Shang et al. [66] measured the colonic temperature, whereas Ardévol et al. [75] and Kunstetter et al. [18] measured, respectively, the aortic and brain cortex temperatures. Interestingly, Drummond et al. [82] simultaneously measured the abdominal and brain cortex temperatures. The initial temperature (i.e. at the beginning of the exercise) ranged from 36.7°C [41] to 39.4°C [74], with average values of 37.5 ± 0.6°C. Finally, the T_CORE at fatigue or exhaustion corresponded to 40.0 ± 1.2°C, being 37.3°C [47] and 42.4°C [79], the lowest and highest temperature values recorded (Figure 2a).

Figure 2. Histograms showing the distribution of the core body temperature (T_CORE) attained at fatigue or exhaustion in rats subjected to incremental-speed (A) or constant-speed (B) exercises. The numeric values on the top indicate the mean ± SD for T_CORE at fatigue/exhaustion in each exercise condition.

We performed a multiple regression analysis to understand the determinants of T_CORE attained by rats at fatigue and exhaustion during incremental exercises (n = 34 trials). Three of the four parameters included in the analysis contributed significantly to explaining the variance in the T_CORE at fatigue/exhaustion: distance traveled (t = 6.863; p < 0.001), T_AMB (t = 9.401; p < 0.001), and initial temperature (t = 6.192; p < 0.001); the contribution of the body mass (t = −1.790; p = 0.083) was close to reaching statistical significance. This analysis led to the following regression equation, with an adjusted R² = 0.813 and standard error of the estimate = 0.521:

T_CORE at fatigue/exhaustion = −4.631 – (0.004 × Body mass) + (0.002 × Distance traveled) + (0.198 × T_AMB) + (1.051 × Initial T_CORE)

Standardized beta coefficients corresponded to −0.144, 0.540, 0.693, and 0.469 for the body mass, distance traveled, T_AMB, and initial T_CORE, respectively (Figure 3a). These data indicate that T_AMB and body mass were, respectively, the variables with the best and worst predictive values used in the model. Moreover, these beta values mean that T_CORE at fatigue/exhaustion increases by 0.693 (in standard deviations) when T_AMB increases by one standard deviation – assuming other variables in the model are held constant. Finally, the negative values for the t-value and standardized beta coefficient regarding the body mass indicate an inverse association between this parameter and T_CORE.

Figure 3. Standardized beta coefficients for the different factors included in the multiple linear regression analysis regarding the core body temperature (T_CORE) of rats subjected to incremental-speed exercises to fatigue or exhaustion. Panel A shows the results from all studies included in the analysis. In contrast, panel B shows the results from studies with common conditions: the measurement of abdominal temperature in male Wistar rats subjected to an incremental exercise to fatigue, with an initial speed of 10 m/min and increases of 1 m/min every 3 min. The bars in red mean negative coefficients (i.e. inverse effects), while the bars in dark green mean positive coefficients (i.e. direct effects). In addition, the variables with greater coefficients are the most important for predicting T_CORE at fatigue or exhaustion.

An additional analysis, considering only the 18 trials using the most common conditions observed in the systematic review, was conducted to ensure that excluding strain, exercise protocol, and T_CORE index from the original regression analysis has not influenced our results. These most common conditions included measuring abdominal temperature in male Wistar rats subjected to an incremental exercise to fatigue (not exhaustion), with an initial speed of 10 m/min and increases of 1 m/min every 3 min. In this case, the same three parameters contributed significantly to explaining the variance in the T_CORE at fatigue – distance traveled (t = 5.143; p < 0.001), T_AMB (t = 5.762; p < 0.001), and initial T_CORE (t = 3.292; p = 0.006) – whereas the body mass did not (t = −0.859; p = 0.406). This analysis led to the following regression equation, with an adjusted R² = 0.727 and standard error of the estimate = 0.308:

T_CORE at fatigue = 14.064 – (0.002 × Body mass) + (0.002 × Distance traveled) + (0.147 × T_AMB) + (0.572 × Initial T_CORE)

Standardized beta coefficients corresponded to −0.111, 0.709, 0.808, and 0.484 for the body mass, distance traveled, T_AMB, and initial T_CORE, respectively (Figure 3b). Again, these data indicate that T_AMB and body mass were, respectively, the variables with the best and worst predictive values used in the model.

Constant-speed exercise

Fifty-seven studies, including 101 experimental trials, investigated the T_CORE of rats subjected to constant exercise to fatigue or exhaustion (Table 2). Among them, 34 studies provided data from one experimental trial [4,6,44,45,62,63,86–113], 13 studies from two trials [3,8,51,64,70–72,82,114–118], four studies from three trials [13,18,48,119], three studies from four trials [47,120,121], one study from five trials [122], and two studies from six trials [46,123].

Data from 1,214 rats were included in this analysis, with an average of 12 ± 8 (mean ± SD) rats per experimental trial. In most investigations, the animals subjected to exercise were Wistar and Sprague-Dawley rats, except Moran et al. [99], which used Zabar rats, and Francesconi et al. [92], which did not report the strain. Only two studies investigated the thermoregulatory responses of female rats subjected to constant exercise [72,116]. The average body mass corresponded to 375 ± 83 g (Table 3).

Concerning the exercise protocol, 16 of the 57 studies reported constant treadmill running at 18 m/min (Supplementary Table 2). This was the most common protocol found in the literature, possibly because 18 m/min corresponds to approximately 65–80% of the maximum aerobic speed attained by untrained rats [41,124] and thus allows elevated metabolic heat production for long periods. The fastest running speed corresponded to 30 m/min [72], whereas the slowest speed was 9.14 m/min, used in a series of studies by Hubbard’s group in the 1970s and 1980s [48,70,115,122].

The rats ran on average for 63.7 ± 47.8 min and traveled a distance of 1,037 ± 869 m. Lima et al. [6] and Lubbe et al. [72] reported the longest and shortest exercise duration (287.0 and 5.6 min), respectively. Therefore, the farthest distance traveled corresponded to 4,908 m and was observed by Lima et al. [6], whereas rats covered the shortest distance (i.e. 168 m) in the study by Lubbe et al. [72]. The dry T_AMB ranged from 5°C [46] to 40°C [99], with the average temperature corresponding to 25.4 ± 8.2°C. Of note, Caputa & Kamari [116] initiated an experimental trial at a T_AMB of 45°C but gradually decreased it to 27°C; this strategy aimed to maintain hypothalamic temperature around 41°C while rats were running.

In most studies included in this systematic review, the rectal and abdominal temperature were the T_CORE index measured. Seven studies simultaneously recorded the rectal/abdominal and brain temperatures in the same rats [47,51,82,109,116,117,119]. The initial pre-exercise temperature ranged from 33.5°C [72] to 40.4°C [116], with average values of 37.6 ± 0.9°C. Finally, the T_CORE at fatigue or exhaustion corresponded to 40.0 ± 1.8°C, being 34.9°C [46] and 43.4°C [92], the lowest and highest temperature values recorded (Figure 2b).

Multiple linear regression analysis was done to understand the determinants of T_CORE values when rats quit running during constant exercises. Four of the five parameters included in the analysis contributed significantly to explaining the variance in the T_CORE at fatigue/exhaustion: running speed (t = −5.610; p < 0.001), time to fatigue or exhaustion (t = 2.728; p = 0.008), T_AMB (t = 10.022; p < 0.001), and initial T_CORE (t = 7.756; p < 0.001). In contrast, the body mass (t = −0.714; p = 0.477) did not contribute significantly to explaining the variance of T_CORE. This analysis led to the following regression equation, with an adjusted R² = 0.786 and standard error of the estimate = 0.808:

T_CORE at fatigue/exhaustion = −2.627 – (0.001 × Body mass) – (0.105 × Running speed) + (0.005 × Time to fatigue) + (0.120 × T_AMB) + (1.097 × Initial T_CORE)

Standardized beta coefficients corresponded to −0.040, −0.294, 0.139, 0.535, and 0.443 for the body mass, running speed, time to fatigue/exhaustion, T_AMB, and initial T_CORE, respectively (Figure 4). Again, T_AMB and body mass were the variables with the best and worst predictive values used in the model. Finally, the negative values for the t-value and standardized beta coefficient regarding the body mass and running speed indicate the existence of inverse associations between these parameters and T_CORE.

Figure 4. Standardized beta coefficients for the different factors included in the multiple linear regression analysis regarding the core body temperature (T_CORE) of rats subjected to constant-speed exercises to fatigue or exhaustion. The bars in red mean negative coefficients (i.e. inverse effects), while the bars in dark green mean positive coefficients (i.e. direct effects). In addition, the variables with greater coefficients are the most important for predicting T_CORE at fatigue or exhaustion.

Quality assessment

The quality assessment of the studies in which rats were subjected to incremental and constant-speed exercises is presented in Table 4.

Table 4. Quality assessment results of the studies in which the rats were subjected to incremental- or constant-speed exercises to fatigue or exhaustion.

Item	Incremental	Constant
1- – Was the purpose of the study clearly stated?	1.0 ± 0.0	1.0 ± 0.2
2- Was the information on the rats’ sex/strain and origin provided?	0.7 ± 0.2	0.7 ± 0.3
3- Was the information on the age and body mass of the rats provided?	0.9 ± 0.2	0.9 ± 0.2
4- Were the rats familiarized with running on a treadmill prior to the experiments?	0.9 ± 0.3	0.6 ± 0.5
5- Was the method used to measure core body temperature described in detail?	1.0 ± 0.0	0.9 ± 0.2
6- Was the exercise protocol described in detail?	0.9 ± 0.2	0.7 ± 0.3
7- Was the core body temperature continuously measured during exercise?	1.0 ± 0.1	0.9 ± 0.2
8- Were the criteria used to determine fatigue/exhaustion described in detail?	0.8 ± 0.4	0.6 ± 0.4
9- Were the environmental conditions described in detail?	0.5 ± 0.1	0.7 ± 0.3
10- Was the daytime/period when experiments were carried out described?	0.5 ± 0.5	0.3 ± 0.4
11- Were the exercise-induced changes in core temperature analyzed with appropriate statistical tests?	0.7 ± 0.3	0.6 ± 0.2
12- Was a limitation paragraph included in the study?	0.2 ± 0.4	0.4 ± 0.5
13- Were the conclusions appropriate considering the objectives and methods of the study?	1.0 ± 0.0	0.9 ± 0.3
TOTAL	10.1 ± 1.1	9.3 ± 1.5

Concerning the studies with incremental exercises, the average score was 10.1 ± 1.1 (mean ± SD) in 13.0 maximum possible points. All manuscripts attained the maximum score in the questions related to the objective, method for measuring T_CORE, and conclusions. In contrast, the lowest scores were related to the lack of a limitations paragraph and inadequate information on the environmental conditions and daytime when experiments were conducted (more details in Supplementary Table 3). Interestingly, the publication date was positively associated with the quality score given to a manuscript (R² = 0.300, p = 0.019; Pearson’s coefficient), thus indicating that the more recent studies were evaluated better than older ones.

With respect to the studies with constant-speed exercises, their average score was 9.3 ± 1.5 points (mean ± SD). The highest scores were obtained in the questions regarding the study objectives, information on age and body mass, method for measuring T_CORE, continuous measurement of T_CORE during the exercise (i.e. more than two time points), and conclusions. In contrast, the lowest scores were related to the lack of a limitations paragraph and insufficient information on daytime when experiments were conducted (more details in Supplementary Table 4). Again, the publication date was positively associated with the quality score given to a manuscript (R² = 0.426, p < 0.001), thus indicating that the more recent manuscripts received better scores than older ones.

Discussion

The present systematic review provides extensive information regarding the T_CORE attained by rats when exercising to fatigue or exhaustion and relevant information on the factors modulating the running-induced T_CORE increase. The studies mainly investigated exercise thermoregulation in male Wistar or Sprague-Dawley rats and described heterogeneous experimental conditions; for example, T_AMB ranged from 5 to 40°C, whereas the initial (pre-exercise) T_CORE ranged from 33.5 to 40.4°C. Quite fascinating, T_CORE values reported at exercise cessation varied approximately 8°C among the studies reviewed. The present systematic review also revealed that higher values of T_AMB, initial T_CORE, and physical performance-related variables (e.g. exercise duration for constant running or distance traveled for incremental running) are associated with higher T_CORE at fatigue or exhaustion.

The lowest T_CORE value (i.e. 34.9°C) at fatigue/exhaustion was observed in rats running in a cold environment [46]. Although this value is similar to hypothermic states induced experimentally by causing severe aseptic inflammation [125,126], Guimarães et al. [46] did not report fatalities following treadmill running, suggesting that hypothermia was reversed after the exercise. Interestingly, rats seem to have greater endurance under cold than temperate conditions [47,127]. On the other hand, the highest T_CORE values were generally associated with fatalities. For example, almost every sedentary (i.e. non-trained) rat attaining more than 43.0°C did not survive [79]. Under these conditions, animal death results from heatstroke, a life-treating condition associated with uncontrolled systemic inflammatory response syndrome and disseminated intravascular coagulation, which combine to induce multi-organ system dysfunction and failure [27,128].

T_AMB was the primary parameter determining the T_CORE attained at fatigue or exhaustion during incremental and constant exercises, with augmented environmental heat stress inducing more significant hyperthermia. This finding confirms previous observations, including experiments when different T_AMBs were used to induce different levels of hyperthermia [46,47,65]. Indeed, the apparent dependence of exercise hyperthermia on T_AMB reinforces the notion that the regulation of T_CORE in small mammals (e.g. rats and mice), which have a higher body mass-to-body surface ratio than bigger mammals, is susceptible to slight variations in T_AMB [56,125,129,130]. Therefore, high T_AMB limits the ability of rats to dissipate heat through dry pathways (e.g. convection and radiation), thus favoring marked increases in T_CORE. Interestingly, while running in the cold, the facilitated heat exchange with the environment due to the high body mass-to-body surface ratio is not compensated by increased metabolism, and, therefore, rats become hypothermic [46].

The initial T_CORE also determined its level when exercise was interrupted; the rats starting the treadmill running with higher T_CORE were the ones that became fatigued or exhausted with more severe hyperthermia. Indirectly, this observation confirms the experimental evidence showing a circadian influence [80,81] or the influence of prior exposure to stress [84] or hot environments [119] on thermoregulation during exercise. The prominent role played by the initial T_CORE also highlights the importance of developing rigorous designs that control the daytime at which experiments are conducted and ensure that rats are tested under stress-free conditions.

The distance traveled also determined T_CORE at fatigue or exhaustion; therefore, the rats that ran farther attained higher temperatures during incremental exercise. In the case of this protocol, the distance traveled is a performance index that provides information on exercise duration but primarily on intensity. As speed gradually increases, the rats will cover longer distances within an exact time interval. Interestingly, the prominent role played by exercise intensity in increasing T_CORE corroborates with data obtained from athletes. For example, the women cyclists winning gold medals in a team time trial (i.e. the ones with the highest power output) under environmental heat stress were among the athletes with higher hyperthermia levels (up to 41.5°C) during this competition [131]. Taken together, our current data in rats and data obtained from athletes suggest that the intensity of physical exertion determines the hyperthermia level attained during incremental exercise in rats or self-paced intermittent exercise in humans.

During constant-speed exercises, the T_CORE at exercise interruption was positively associated with running duration while negatively associated with treadmill speed. Although a faster treadmill speed produces more intense increases in metabolic rate, tolerance to exercise under these conditions is very restricted. For example, male rats could run for less than 6 min at 30 m/min; this speed is greater than the maximum aerobic speed attained by untrained rats reported in some studies included herein [82,111,118]. On the other hand, several studies that subjected rats to hyperthermic exhaustion consisted of slow treadmill speeds (i.e. 9.14 and 11 m/min) during prolonged periods under environmental heat stress [48,114,120].

Finally, the body mass was the factor analyzed that less predicted the T_CORE at fatigue or exhaustion, regardless of the exercise protocol. Moreover, the associations between the body mass and T_CORE level were consistently negative, with lighter rats more susceptive to heating up. In this sense, lighter animals will heat up faster than heavier animals, provided the same amount of heat is stored in the core body region, mainly if heat originates from external sources. In contrast, it is expected that a lower body mass would exert a protective thermoregulatory role, as previously shown in humans. For instance, treadmill walking at a fixed external workload elicits a lower rate of metabolic heat production in lighter than heavier individuals [132]. Despite the divergent effects observed in rat and human experiments, body mass is a concern for interpreting data in weight-bearing exercises. Therefore, caution should be exercised in interpreting thermoregulatory data in rat experiments when body mass is different between experimental groups, such as the lower body mass in spontaneously hypertensive rats than in normotensive Wistar rats [82] and the reduction in body mass caused by aerobic training in rats subjected to a high-fat diet [133]. Lastly, whether the body mass indeed contributes inversely to determining the rats’ T_CORE at fatigue or exhaustion is an issue to be investigated carefully in future studies.

This systematic review included 72 articles published between 1968 and 2022. Among these studies, only three subjected female rats to treadmill running [72,75,116]. The lack of data on female rats reproduces the underrepresentation of women in research dealing with exercise thermoregulation. While Hutchins et al. [134] reported that women accounted for 30% of the human subjects in the exercise thermoregulation research published in 2019, we observed that females represented only 2% of the rats investigated since 1968. The underrepresentation of women/females is undesirable because it significantly limits the generalization ability of the findings obtained only in men/males. This issue must be resolved soon, considering the increased number of women involved in physically demanding work activities [135] and their growing participation in elite sports [136,137].

The manuscripts included in the present review had their quality assessed as recommended by the PRISMA guidelines [38]. Although the average score was higher than 9 in 13 maximum possible points for the studies involving incremental- or constant-speed exercises, some relevant points should be highlighted. First, the authors of the current review elaborated themselves the checklist for quality assessment; therefore, some criteria seen as necessary by other research groups may have been left aside from the assessment. Second, positive and significant correlations were observed between the publication date and quality score, indicating that the more recent manuscripts received better scores than older ones. This observation is expected because the requirements for scientific publishing have become more stringent over time. Indeed, the recent manuscripts have method sections, particularly the statistical analyses, which are much more extensive and detailed than the older ones.

Despite the criticisms presented earlier, the quality assessment provided some crucial information for enhancing the quality of future experiments on the topic, including the necessity of more comprehensive control of methodological issues, such as the daytime when experiments are performed and the environmental factors other than T_AMB. The current findings also encourage inserting the studies’ limitations at the end of the discussion section. Furthermore, it is worth noting that while the more recent studies received better scores in most criteria (e.g. description/use of familiarization sessions with treadmill running), this statement is not valid for all criteria. For example, the description of environmental conditions followed the opposite way, suggesting that the authors are not controlling these factors (e.g. relative humidity and artificial airflow inside the treadmill) as effectively as in the past. Indeed, the airflow generated by electric fans may facilitate convective cutaneous heat loss [2], and rats can evaporate water from their respiratory tract while running [17].

The present systematic review is not free of limitations. First, the data used in the multiple linear regression analysis consisted of average data extracted from the manuscripts included in the review. This is not an ideal procedure because average T_CORE may be associated with different levels of heterogeneity or result from different sample sizes in the studies we reviewed. Ideally, data from individual rats should be used to run the regression analysis; however, these data are not available because several manuscripts were published more than 30 years ago. Second, data related to cutaneous heat loss and metabolic heat production were not extracted from the manuscripts included in the review. Because the changes in T_CORE result from imbalances between the rates of heat production and dissipation [1,39], these aspects should be considered in future investigations.

Third, body mass data were reported as a range in several studies, and we did not have access to the body mass measured on the day of the treadmill exercise. In this case, we considered the greatest mass, which may have added imprecision to our analysis. Fourth, we included data from different T_CORE indices, namely the abdominal, brain cortex, and colonic temperatures. The temperatures measured in specific body compartments are not homogeneous and do not respond in a similar way (particularly regarding their time course) to several arousing stimuli [138] and physical exercise [139]. Nonetheless, a sub-analysis using only the abdominal temperature measured during incremental exercises reproduced most findings of the analysis using the three T_CORE indices, thus suggesting that the site of T_CORE measurement was possibly not a confounding factor in the outcomes of the multiple linear regression analyses. Despite all these limitations, essential patterns have emerged from regression analyses, consistent with some existing ideas in the literature, as discussed earlier.

In conclusion, when fatigued or exhausted, rats subjected to treadmill running exhibit heterogeneous T_CORE values. Moreover, it is not possible to determine a single T_CORE or a narrow range of T_CORE associated with exercise cessation in hyperthermic rats. More importantly, the present systematic review helps understand the parameters that determine the T_CORE values attained at fatigue/exhaustion in two physical exercise protocols, with a particular reference to T_AMB, initial T_CORE, and physical performance-related parameters (i.e. distance traveled in the incremental exercises and duration in the constant exercises). In contrast, among the factors analyzed, the body mass was the one that least predicted the level of T_CORE attained in both exercise protocols. Finally, from a broader perspective, this systematic review provides relevant information for selecting appropriate methods in future studies designed to investigate exercise thermoregulation in rats.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/23328940.2022.2115274

Additional information

Funding

This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. NHSB was the recipient of post-graduate fellowship from CAPES, whereas KNOG received a post-doctoral fellowship from CAPES/Print. SPW receives a fellowship from the Conselho Nacional de Desenvolvimento Científico e Tecnologico (CNPq) for being a productive researcher (grant number 315199/2021–0). DAPG is currently funded by the Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG, grant number APQ-01268-21). DDS, DAPG, TTM, and SPW are researchers of the MEDIANTAR group, which is supported by CNPq/MCTIC/CAPES/FNDCT/PROANTAR (grant number 442645/2018-0).