Introduction
Working dogs are selected for high motivation and drive to perform strenuous physical and mental activities. These characteristics along with the environments in which they work put them at risk for exertional hyperthermia. Their drive can override physiologic mechanisms that signal increases in core body temperature, and continued activity can lead to heat injury or heat stroke.1,2 In dogs, most heat is dissipated through evaporative cooling via panting. The ability to decrease core body temperature is reduced in environments with > 35% humidity.3 High temperatures, combined with moderate to high humidity and high drive to work can lead to heat stress that rapidly progresses to heat injury and eventually heat stroke.4 Heat-induced injuries are the most common nontraumatic cause of death among military working dogs2 and civilian law enforcement dogs.3 Handlers must monitor their dogs, recognize signs of heat stress, and rapidly implement effective cooling strategies.4 Cooling recommendations focus on effective heat exchange, controlling the heat-generating activity, and providing a temperature-controlled environment.1 Currently, the recommended cooling strategy after exertional hyperthermia involves full or partial immersion in cool water5,6 or using a fan to increase air circulation and evaporation after wetting the skin thoroughly.4 Many working environments lack access to safe water sources, and field operations may preclude carrying large amounts of water and equipment. Effective means of field cooling with limited resources have not been extensively studied.
This study tested the effects of 4 field-applicable cooling methods on core body temperature after exercise in working dogs. Field applications rely on first aid supplies including hand towels, chemical ice packs, and a limited amount of water that handlers or dog owners can carry. This extension of our previous study6 tested 4 practical approaches to reducing core body temperature in environments where water for immersion is unavailable or unsafe. It was hypothesized that chemical ice packs applied to the neck of the dogs would result in a more rapid reduction of core body temperature than water-soaked towels applied to the neck or axillae or voluntarily dunking the head in a bucket of water due to their lower temperature.
Methods
Participants
Twelve healthy working dogs at the Penn Vet Working Dog Center (PVWDC) that were > 8 months of age, weighed at least 18 kg, had a trained “come/recall,” and were comfortable with veterinary procedures, restraint, and unfamiliar people were enrolled (Supplementary Table S1). Dogs > 8 years old or dogs with prior neurologic or musculoskeletal disease, a body condition score of 7 to 9/9, anticipated graduation from the PVWDC during the study period, or aggression or anxiety during veterinary examinations were excluded. All dogs were evaluated by a veterinarian at the start of each study day.
Sample size estimation was carried out in Glimmpse7 on the basis of a range of estimated SDs derived from a previous study we conducted.6 A Hotelling–Lawley Trace with an α of .05 and a power of 0.8 yielded a suggested sample size of 10.
Study design
In a nonanonymized randomized crossover study of exertional hyperthermia (defined as an increase in core temperature above baseline associated with physical activity; eg, > 39.2 °C), dogs were divided into 2 groups of 6 on the basis of the PVWDC training schedules. Dogs participated in each of the 4 study cooling protocols over 4 weeks (June 27, 2023, to July 24, 2023). Dogs only participated in a trial once on a given study day, and trials were separated by at least 48 hours.
Within each group, dogs were randomly assigned by means of a random number generator8 to 1 of the 4 cooling protocols for a particular day (ie, water-soaked towels under the axillae, voluntary head dunking in water, 2 chemical ice packs on the neck, or a water-soaked towel on the neck). All water was maintained at 22.2 °C. All dogs participated in all 4 protocols. No control passive cooling group was included. Although there was no passive control group in the current study, data from a prior study6 using the same protocol was sufficient to make comparisons and inform the results from the current study. The University of Pennsylvania IACUC approved the protocol (No. 807275), and dog fosters provided informed consent.
Technology
Accelerometers, heart rate monitors, and ingestible core temperature–sensing capsules for each dog were activated at the start of each study day. The activity monitors with an omnidirectional accelerometer (Respironics Actical, version 3.1; Koninklijke Philips Electronics) were calibrated to each individual dog’s weight and age and attached to flat buckle collars on the dog during the physical examination. The activity counts recorded by these monitors have been validated in dogs as a way to continuously measure the frequency, intensity, and duration of movement.9 The monitors were used to confirm that the activity during the cooling rest phase was minimal.
Heart rate was continuously monitored throughout exercise and cooling with a commercially available heart rate monitor (Polar USA).10–12 Dogs were shaved once on the left ventrolateral aspect of the chest between ribs 3 and 6 for heart monitor placement. Ultrasound gel was applied to the strap, and the monitor was secured to the skin with Elastikon at the time of the physical examination.
Core body temperature was monitored with internal temperature–sensing capsules (Anipill; Animals Monitoring) administered by mouth in a canned canine diet meatball. Every minute, readings of core body temperature were taken to ensure the dogs’ core body temperatures did not exceed the designated safety cutoff temperatures (40.6 °C for recall and 41.9 °C for the cooling period). These capsules have been used in prior studies in dogs.13
The weather was monitored using a weather app for iPhone based on the study location zip code 19146, which covers an area of 1.69 square miles. Ambient temperature, humidity percentage, wind speed, and heat index were measured when each individual dog began their sprint test to ensure the conditions were acceptable for outdoor activity and the heat index did not exceed the study-designated maximum of 33.9 °C (Supplementary Table S2).
Experimental protocol
The study was conducted on prescheduled days unless there was (1) a heat index ≥ 33.9 °C and/or (2) inclement weather that could create an unsafe terrain or environment for the study participants or personnel. The experimental protocol with the exception of the cooling interventions was based on a previously described protocol.6
A study day consisted of a physical examination by a veterinarian, warm-up exercises, a recall test, the assigned cooling intervention, and the 40-minute cooling period. Each dog was assigned a consistent trained handler and trained data recorder. The study assistant (TRC) monitoring for signs of heat stress (eg, excessive panting, elongated and flattened tongue, narrowed palpebral fissure, retracted ears, shade seeking, slowing of pace) during the recall exercise was trained to recognize behavioral signs of heat stress, exhaustion, and stroke.
Physical examination
Study days began between 8:00 am and 9:00 am and finished between 11:30 am and 12:30 pm EST. Core temperature capsules were administered 30 minutes to 2 hours prior to each dog’s physical examination. Water consumption was prohibited from the start of the physical examination until the conclusion of the 40-minute cooling period. Dogs were permitted to drink following the 40-minute cool-down period. A full physical examination including core temperature, baseline heart rate, baseline respiratory rate, mucous membrane characteristics, capillary refill time, body condition score, and hydration status was performed under a tent outside the recall.
Warm-up
Following the physical examination and core body temperature reading, the dog was walked approximately 10 m to the recall area for the warm-up stretching exercises based on the PVWDC Fit To Work Foundational Fitness program.14 Briefly, the dog walked, trotted for 30 to 60 seconds, performed 3 to 5 figure eights (active lateral side bends) through the legs of the handler, performed a 10- to 15-second “paws up” stretch rising on their back legs with their front legs on the handler’s arm (to stretch hips and shoulders), and a 10- to 15-second “play bow” stretch with dog’s elbows on the ground and hips elevated (to stretch shoulders, spine, gracilis, and semimembranosus and semitendinosus).
Recall test
Recall involved the dog running in a grassy fenced-in area between 2 individuals 25 m apart for the dog’s preferred toy, as described in our previous study.6 The dog was held by 1 person, while the second person, 25 m away with a toy, called the dog. The dog ran to the second person and played tug for 15 to 30 seconds, and the process was repeated going in the opposite direction. If a dog attempted to run outside of the designated area, a study assistant would trot with the dog on the long leash until one of the criteria was met for the completion of the recall. A camera on a tripod was used to record the recall test similarly to the setup for our previous study.6
The core temperature, heart rate, and any performance comments were recorded before the sprint began (recall 0) and after each minute of running. The dog was checked for signs of heat stress every 50 m.
The recall portion of the test concluded when 1 of 3 following scenarios was reached: the core temperature exceeded 40.6 °C, the total recall duration reached 10 minutes, or ≥ 2 signs of heat stress (eg, excessive panting, elongated and flattened tongue, narrowed palpebral fissure, retracted ears, shade seeking, slowing of pace) were demonstrated by the dog. If the dog did not complete 10 minutes, their total recall duration was recorded (Supplementary Table S3).
Cooling interventions
After the recall, the dog was led to the tent area where a core temperature reading was recorded. The dogs participated in 1 of 4 cooling protocols lasting 30 seconds on each day of the trials: wet towels placed under the axillae, a voluntary head dunk for a toy or food reward, 2 chemical ice packs placed on the neck, or a wet towel placed around the neck.
For the axillae towel protocol, the dog stood while 2 towels were placed around the axillae and pulled dorsally. The towels were folded into a square, and 237 mL of 22.2 °C water was poured on each side. The towels were unfolded and then refolded in half lengthwise twice. The 2 towels were pulled into the axillae and held for 30 seconds by attached metal rings (Figure 1). Any adverse behavior such as displaying discomfort when the towels were put near the feet or trying to lick the towels was recorded.
Representative images of 4 cooling interventions evaluated for field treatment of exercise-induced hyperthermia in 12 working dogs enrolled in a randomized, crossover study conducted between June 27, 2024, and July 24, 2024, with core body temperatures monitored with the use of orally administered temperature-sensing capsules (Anipill; Animals Monitoring) activated at the start of each study day. A—A towel soaked in 22.2 °C water is secured in the axillae of the dog with mild signs of heat stress for 30 seconds. B—The dog is encouraged to dunk its head in a cooler up to eye or ear level of 22.2 °C water to retrieve a treat or toy up to 3 times in 30 seconds. In this image, the dog is initiating the head dunk. C—Chemically activated ice packs are secured on the ventrolateral aspects of the dog’s neck for 30 seconds. D—A towel soaked in 22.2 °C water is secured around the neck of the dog for 30 seconds. Note that the dog in panel D is not showing signs of heat stress and was only demonstrating the towel placement.
Citation: Journal of the American Veterinary Medical Association 262, 12; 10.2460/javma.24.06.0368
Prior to the study, all dogs were trained for 1 week to retrieve a toy or food from the bottom of a bucket of water (“head dunking”; Figure 1). The dog voluntarily immersed most of their head into a cooler with 19.72 L of 22.2 °C water to retrieve a toy or food reward. The water temperature was monitored with a temperature probe (WD-20250-01 Type-K/J single-input thermocouple thermometer, Traceable Products; or GoodCook touch instant read thermometer, Bradshaw Home), and ice was added as needed. Each dog had fresh water in the cooler for their trial. Prior to dunking, collars were removed to prevent interference with the head dunk. A study assistant first teased the dog with the toy or food while the handler held them back from the cooler. Once engaged, the handler released the dog and allowed them to approach the cooler and try to retrieve the reward. The study assistant held the reward near the bottom of the cooler to ensure the dog would submerge their head to retrieve the reward. The dog had 30 seconds to complete a maximum of 3 head dunks. A successful head dunk occurred when the water reached the level of the eyes or ears. No dog was physically forced to dunk their head. Dogs were discouraged from drinking the water as much as possible. Any adverse behavior such as reluctance to head dunk or only wanting to drink the water was recorded.
The ice pack protocol used 2 chemical ice packs secured with Velcro on a collar around the dog’s neck. Once the dog completed the recall exercise, the ice packs were activated by cracking the ice pack and shaking it for 1 minute. The dog’s collars were removed and replaced by the collar with the ice packs. Ice packs were positioned on the cranial aspect of the neck covering the jugular veins lengthwise (Figure 1) for 30 seconds, and any adverse behavior, such as reluctance for the ice packs to be placed, was recorded.
The neck towel protocol consisted of 1 wet towel placed circumferentially on the cranial aspect of the neck and secured by a collar. The towel was wet in the same manner as the axillae towel protocol, but only 1 towel was needed and folded in half widthwise only once. The dog’s collars were removed prior to the towel being placed (Figure 1). The towel and collar remained on the dog for 30 seconds, and any adverse behavior, such as reluctance to the towel being placed, was recorded.
After completion of the protocol, core temperature was recorded, the accelerometer collar was replaced if it had been removed, and the dog was taken to a shaded cooling area with individual wire crates and monitored for 40 minutes. Heart rate and core temperature were recorded every minute. In the wire crate, the dog was allowed to sit, lay, or stand. If the dog’s temperature reached 41.9 °C at any time during the study, the dog was immediately placed into the rescue protocol consisting of water immersion in a wading pool and controlled access to drinking water. If the dog’s temperature remained above 39.4 °C after the 40 minutes of passive cooling, the dog was actively cooled by water immersion in a pool until the core temperature was < 39.4 °C.
Additional data collection
Five of the 12 dogs did not dunk their heads in their initial head dunk trial day. One of these dogs’ core temperature was 41.9 °C; therefore, the dog was placed into the cooling rescue protocol and head dunk was not attempted during that trial. One dog lacked the motivation to perform the head dunk and was not retested. The other 4 dogs participated in a retest of the trial to attempt the head dunk protocol again. Two dogs successfully head dunked and their data was included in the analysis, making the total number of dogs that successfully performed the head dunk 9 of 12. It was not possible to quantify the volume of water ingested during the head dunk; however, the number of times the dogs lapped water during the head dunk protocol was obtained from video review.
Statistical analysis
Descriptive statistics were generated in SigmaPlot (version 14.5; Grafiti LLC). Normality was tested by the Shapiro-Wilk test, and normally distributed data were reported as mean and SD. Data that failed normality were reported as median and IQR. All statistical analyses were run in R version 4.2.3 (The R Project for Statistical Computing).15
Model 1: effect of cooling protocol and time point on temperature
A piecewise approach with linear mixed-effects models was used to examine the effect of time after exertion and cooling protocol on temperature with a random intercept of dog. The models were generated with the lme4 package in R.16 A model was fit for each of 2 time periods: initial rapid cooling and return to baseline. Initial rapid cooling occurred between time points 0 and 5 minutes. Return to baseline occurred between time points 6 and 40 minutes after cooling. Time after cooling was included as a continuous time variable in both models. These time periods were chosen to better represent the data and to address (1) the extent to which each protocol rapidly lowered dogs’ temperatures after exertion and (2) the extent to which protocols continued cooling to a baseline temperature.
This model met all assumptions for analysis. Post hoc comparisons of estimated marginal means were performed with a Tukey correction to compare cooling protocols in different time periods with the emmeans package in R.17
In this model, cooling protocol represented 4 different cooling interventions. When including categorical variables in a regression model, 1 category is selected as the reference category. The coefficients for the other categories represent the difference in the response variable (temperature) relative to this reference category. Ice pack was chosen because of the hypothesis that the ice pack would be the superior cooling method. Since we planned to conduct post hoc tests to further examine comparisons between groups, the specific selection of the reference variable was less important.
Model 2: assessment of return to baseline temperature
A repeated-measures ANOVA was used to assess whether the dogs’ temperature following cooling protocols returned to baseline, as reflected by the temperatures of the dogs taken prior to exercise. This model was generated with the car package in R.18 This model used pre– or post–cooling protocol temperature and cooling protocol as within-subjects factors and temperature as an outcome variable, with dog as a random factor. This model met assumptions of normality and sphericity necessary for analysis.
Results
Descriptive statistics
The median core temperature of dogs prior to exercise (while outside) was 39.1 °C (IQR, 38.8 to 39.6 °C); after 40 minutes of cooling, the median temperature was 39.4 °C (IQR, 39.0 to 39.7 °C) (Figure 2). There were 19 instances in which dogs’ temperatures were > 39.4 °C after the 40-minute outdoor cooling period, and those dogs were all cooled in the pool prior to returning to the climate-controlled kennel. Seven were in the ice pack protocol, 5 in the axillae towel protocol, 4 in the head dunk protocol, and 3 in the neck towel protocol. In 8 of the 19 instances, dogs’ temperatures were above 39.4 °C prior to exercise. Video recordings of each individual dog’s recall test, cooling protocol, and cooling period were reviewed. Each dog trotted or sprinted during the recall; the mean activity counts during the recall test were 3,392 ± 517 counts/min. During rest, the median activity counts were 73 counts/min (IQR, 14 to 220 counts/min), consistent with rest.9 The median heart rates during activity and rest were 144 beats/min (bpm; IQR, 123 to 170 bpm) and 113 bpm (IQR, 100 to 131 bpm), respectively. The most common signs of heat stress were retracted ears (n = 25) and elongated tongue (17; Supplementary Table S4). All dogs were tolerant of the cooling protocols, although 3 dogs failed to perform the head dunk. All dogs lapped some water during the head dunk protocol. In review of the head dunk videos detailing 9 of 9 dogs that successfully dunked, dogs lapped water between 2 and approximately 32 times. On the basis of Adolph19 and, more recently, Gart et al,20 the expected volume per lap is 1 to 2.5 mL; therefore, assuming 2 mL/lap, it is estimated that dogs ingested between 0.27 and 2.8 mL/kg, with a median of 0.55 mL/kg. Of the 5 dogs that failed to head dunk on the initial protocol, 1 dog was panting excessively, whereas 4 were considered training failures (one of these was not retested due to lack of motivation). In repeat testing of 4 dogs, 2 dogs successfully dunked. The one dog that was withdrawn from the initial protocol failed the retest due to excessive panting, whereas the other dog was considered a training failure for the second time. Overall, between the test and retest, 2 dogs failed due to panting excessively and 3 dogs failed due to training issues (1 dog failed twice).
Mean core body temperature for the dogs described in Figure 1, grouped by cooling intervention of axillae towel (black square; n = 12), voluntary head dunk (gray circle; 9), chemical ice pack on the neck (gray triangle; 12), or neck towel (gray diamond; 12) during the cool-down period. The dotted line represents the median core temperature, and the shaded region is the IQR at the time of the outdoor physical examination prior to exercise.
Citation: Journal of the American Veterinary Medical Association 262, 12; 10.2460/javma.24.06.0368
Model 1: effect of cooling protocol on temperature in different time periods after cooling
Post hoc comparisons were performed on the linear mixed-effects models to examine differences in temperature by cooling protocol in different time periods after cooling (initial rapid cooling and return to baseline).
During the period of initial rapid cooling (from immediately after cooling intervention to 5 minutes after cooling intervention), head dunk led to overall lower temperatures than all other conditions (Figure 2) (head dunk vs ice pack, t = –8.78 and P < .0001; head dunk vs axillae towel, t = –8.48 and P < .0001; head dunk vs neck towel, t = –8.51 and P < .0001). The temperature in other protocols did not differ from each other in this time period (ice pack vs axillae towel, t = 0.32 and P = .99; ice pack vs neck towel, t = 0.29 and P = .99; axillae towel vs neck towel, t = –0.03 and P = 1.00). This suggests that, in the crucial initial period of time after exercise, head dunks cool dogs more effectively than the other interventions.
During the period of return to baseline (from 6 to 40 minutes after cooling intervention) (Figure 2), head dunk continued to lead to overall lower temperatures than all other conditions (head dunk versus ice pack, t = –18.89 and P < .0001; head dunk vs axillae towel, t = –17.09 and P < .0001; head dunk vs neck towel, t = –16.14 and P < .0001). Neck towel led to overall lower temperatures than ice pack in this time period (t = 3.36; P = .005). The temperature in other protocols did not differ from each other in this time period (ice pack vs axillae towel, t = 2.32 and P = .09; axillae towel vs neck towel, t = 1.05 and P = .72).
Model 2: assessment of return to baseline temperature
There was no effect of pre– or post–cooling protocol temperature (Figure 2), suggesting that there was no significant difference in the dogs’ baseline temperatures and postcooling temperatures (F = 3.31; P = .10). This means that, within 40 minutes, dogs overall were cooled back to their baseline temperature after all cooling protocols. There was no effect of protocol such that protocol led to different temperatures when baseline and postcooling temperatures were pooled together (F = 0.5; P = .68). There was no interaction between pre- and postcooling temperatures and protocol, suggesting that across protocols there was no difference in baseline temperature and postcooling temperature (F = 1.08; P = .37).
Discussion
Building on the recommendation from our previous study6 and prior literature5,21 for partial water immersion to reduce core body temperature after exercise-induced hyperthermia, this study tested field-applicable cooling strategies to reduce core body temperature in dogs after exertional hyperthermia. Four different cooling protocols (chemical ice packs placed on the neck, a wet towel placed on the neck, wet towels placed on both axillae, and trained voluntary head dunking into a bucket of water for a reward) were compared. Core body temperature returned to baseline over 40 minutes in all protocols. The head dunk protocol, which included limited water ingestion, produced the lowest mean temperature after cooling, followed by neck towel, axillae towel, and, lastly, ice packs. Of greatest clinical relevance, the head dunk with limited water ingestion was the only protocol to immediately decrease core body temperature in the first 30 seconds of cooling. No other protocol prevented the initial rise in core body temperature after exercise. Additionally, the head dunk protocol led to the lowest temperatures in the initial postexercise time period, an important factor in treating heat injury. Proper training and implementation of the head dunk is critical, as 3 of the 12 dogs failed to perform the procedure during the protocol or retest. Two of those 3 were considered training failures.
The neck and axillae towel protocols provided cooling by conduction at the application site and circulatory convection. The resulting decrease in core body temperature over time was anticipated on the basis of similar studies of humans.22,23 When human athletes wore an ice vest between exercises and after exercise, skin temperature was reduced and heat loss was improved within the first 5 minutes of recovery.22 The use of an ice towel on the neck, thighs, and head significantly decreased core body temperature and thermal strain after a simulated tennis match.23 It is possible, however, that use of a wet towel may have hindered a larger reduction in core body temperature. Moisture trapped in the towel24 and subsequently left on the dog may have impaired adequate evaporation and limited cooling by trapping the heat produced by the dogs and not allowing it to transfer to the ambient environment.25 The dogs were in open wire crates, but limited ambient air movement and high humidity may have further hindered evaporative cooling.21
The initial hypothesis that the ice pack protocol would be the most effective due to its low temperature and subsequent increased conductive properties was not proven. In human studies, ice packs have been shown to reduce the physiologic effects of heat strain by conductive cooling and increase the amount of evaporative cooling.26 Additionally, the use of ice packs following orthopedic surgery in dogs significantly decreases intra-articular temperature.27 There are a few reasons that our results may not have aligned with these studies. First, the small surface area of the ice packs may not have been sufficient to provide adequate conductive cooling. A more likely reason is that fur on the dogs’ necks could have acted as an insulating layer preventing direct skin contact with the ice packs.
The head dunk protocol with limited water ingestion caused the largest decrease in the mean core body temperature and cooled most in the initial 5 minutes after exercise when compared to the other protocols. During heat stress, arterioles in skin vascular beds dilate and arteriovenous anastomoses open in the muzzle, ears, and limbs to increase blood flow and heat delivery to the periphery to dissipate into the environment by convection and radiation.28 Conduction can also decrease core body temperature if the dog is in contact with a cooler surface. The head has thinner skin, and dogs lose heat through their muzzle and ears readily through vasodilation and convection. The carotid rete, a network of anastomosing vessels, conducts and distributes blood draining from the nasal cavity and reduces the blood temperature prior to reaching the brain.29 With increased exercise, panting helps to cool the blood draining from the nasal cavity and further helps protect the brain from elevated core body temperatures.28
In previous studies5,6,21 of exercise-induced hyperthermia in dogs, partial water immersion was effective in reducing core body temperature. The large surface area allows for evaporative cooling and convective cooling; however, without sufficient airflow or with excess humidity, the cooling effect might be limited.21 Furthermore, insulation from fur or body fat can restrict convective cooling. An additional limitation of partial water immersion is that a safe cool-water source may not be readily available in all settings. In humans, during work- or exercise-induced hyperthermia, cooling the head improved perception of cooling and performance but did not lower body temperature.30–32 In this study, head immersion in cool water dramatically and rapidly reduced core temperature and prevented the postexercise rise in temperature in dogs. The difference in response to head cooling between humans and dogs may be due to the contribution of conductive cooling from the head and ears of the dog, whereas humans rely on greater body surface area and sweating to eliminate heat. The dramatic decrease in core temperature in dogs following head dunk suggests that direct translation of cooling studies from humans to dogs may not be possible.
Although the head represents a fraction of total body surface area in the dog, perfusion during hyperthermia is preferentially directed to superficial regions (ie, ears, tongue, nasal turbinates) of the head.30,33,34 When the dog’s head is submerged in water up to the eyes or ears, the blood perfusing the cranium, especially in the nasal mucosa and tongue,33,34 is cooled via conduction. This cool blood then circulates to the body to reduce core body temperature. The cool water also affects the temperature of the blood draining from the nasal cavity, thus potentially decreasing the brain temperature.
Another aspect of the head dunk protocol that may have contributed to the cooling effect was the limited amount of water the dogs were able to drink when retrieving their reward. Consumption of cold liquids reduces core body temperature by conductive cooling of the intestinal tract35 and, in human studies, decreases tympanic membrane temperature.36,37 Although it is recognized that core capsules are subject to artifact and transient temperature decreases following water consumption,36 the limited volume of water ingested was unlikely to have had a major or prolonged effect on capsules, although further investigation is warranted.
These results raised the question that if drinking water contributed to reduced core body temperature, why not allocate the limited water supply to drinking in times of exercise-induced hyperthermia? Although hyperthermia is not a reported risk factor for gastric dilatation and volvulus, concerns of rapid consumption of large volumes of cool water include aerophagia, panting, and the gulping of water38,39 after strenuous exercise that may contribute to the risk of gastric dilatation and volvulus. One option to reduce core body temperature would incorporate limited water consumption with the head dunk.
The American College of Veterinary Emergency and Critical Care Committee on Trauma recommends “cool first, transport second,”3 which is supported by several clinical studies.21,40–42 The rapid cooling and prevention of postexercise increase in core temperature makes the head dunk protocol along with limited ingestion of cool water an attractive recommendation for handlers in the field. However, dogs should never be forced to dunk their heads, and dogs with signs of heat stroke, collapse, or abnormal mentation should not participate in head dunking. Training canine athletes and working dogs to immerse their heads in water for a reward can provide a low-stress approach to cooling with the added benefit of providing a cooperative approach to flushing eyes and nasal passages.
A major limitation of this study was the lack of a passive cooling group. On the basis of our historical data from a similar study, passive cooling was not effective (Supplementary Figure S1).6 If the head dunk had not been so different from the other 3 cooling interventions, this oversight would have prevented us from clearly identifying a preferred cooling method. One of the biggest limitations was the core monitoring procedure. Although well validated, water consumption will artificially, albeit briefly, lower pill temperature in the gastrointestinal tract.43,44 However, drinking ice water is a recommended cooling method for humans and leads to core cooling.35,36 The limited amounts of water consumed in this study was at 22 °C; therefore, the overall effect on both the pill and core cooling was likely small. The head dunking protocol was an anticipated behavioral challenge for this study. One week prior to the study, dogs were trained with positive reinforcement to immerse their heads to retrieve a toy or treat in the cooler. Teasing the dog with the reward and using multiple rewards (toys, different foods) helped motivate the dogs to dunk their heads. Three dogs refused to dunk their heads, and 2 failed even after a repeat trial; one failed due to excessive panting during both trials, and the other 2 dogs were deemed training/motivation failures. This may have reflected a need for a longer head dunking training period or brief precooling prior to head dunking. Head dunking could be incorporated into the regular training of dogs at risk of exertional hyperthermia, such as working or sporting dogs specifically training in the field environment after progressively more intense exercise. Additionally, validation of the efficacy and tolerance of water poured onto the head should be investigated.
When the towels were placed for the axillae protocol, some water dripped onto the platform where the dogs were standing, creating a small puddle. Dogs’ feet were moved to ensure they were not standing directly in the water, but there could have been evaporative cooling of the paw pads similar to the effect of standing on a towel with isopropyl alcohol.6 The variation in coat length may have impacted the efficacy of the cooling with the ice packs and warrants further investigation.
This study of cooling after exercise-induced hyperthermia in dogs extended our previous findings6 that partial water immersion reduced core body temperature by investigating cooling tactics for field use. Although all protocols (ie, ice packs placed on the neck, a wet towel placed on the neck, wet towels placed on both axillae, and head dunking into a bucket of water for a reward) produced some cooling, the voluntary head dunk protocol with limited water ingestion resulted in the lowest mean core body temperature and led to the lowest temperatures in the initial 5 minutes after exercise, which are crucial factors in preventing serious heat injury.
While voluntarily dunking in cool water will rapidly decrease core body temperature, some dogs with exertional hyperthermia that were previously trained to dunk may refuse to do so if they are unable to interrupt panting. In addition, if the dog is showing signs of heat stroke, the head dunk should not be attempted. If the dog refuses to dunk, it may be effective to allow them to drink a small amount, then pour water over the dog’s head and neck or reattempt a head dunk. The head dunk may be a lifesaving approach to exertional hyperthermia, and additionally implementing cooperative care helps preserve the human-animal bond. This approach is not intended for use in cases of heat stroke; if the dog is not mentally appropriate, emergency cooling measures and rapid transportation to veterinary care are required.
Although the head dunk has a dramatic effect on cooling, future field studies could further advance care by investigating the effect of an air source (eg, a fan), the effect of intermittent head dunk during pauses in exercise, the role of different water temperatures, and use of core temperature monitoring that would not be influenced by water consumption. Additionally, incorporating wet bulb globe temperature would allow for the effects of direct sun on heating and cooling to be evaluated. Investigating the effects of cooling technique on biomarkers associated with heat injury would provide additional valuable information.
Heat injury is a preventable cause of morbidity and mortality in dogs. Having a variety of proven options for field cooling will facilitate implementation in areas with limited supplies and benefit canine health and welfare.
Supplementary Materials
Supplementary materials are posted online at the journal website: avmajournals.avma.org.
Acknowledgments
The authors would like to thank the Penn Vet Working Dog Center staff and volunteers for their help with the study and extend special thanks to Darko Stefanovski, PhD, for statistical review.
Disclosures
The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.
Funding
Funding was provided by the American Kennel Club Canine Health Foundation Grant 03077-A.
ORCID
C. M. Otto https://orcid.org/0000-0003-0846-2114
References
- 1.↑
Cobb ML, Otto CM, Fine AH. The animal welfare science of working dogs: current perspectives on recent advances and future directions. Front Vet Sci. 2021;8:666898. doi:10.3389/fvets.2021.666898
- 2.↑
Baker JL, Hollier PJ, Miller L, Lacy WA. Rethinking heat injury in the SOF multipurpose canine: a critical review. J Spec Oper Med. 2012;12(2):8-15. doi:10.55460/Y0AS-S4Y3
- 3.↑
Hanel RM, Palmer L, Baker J, et al. Best practice recommendations for prehospital veterinary care of dogs and cats. J Vet Emerg Crit Care (San Antonio). 2016;26(2):166-233. doi:10.1111/vec.12455
- 4.↑
Working dog handler medical care manual. Department of Homeland Security Office of Health Affairs. Accessed August 16, 2023. users.neo.registeredsite.com/1/2/1/13151121/assets/DHS_Handler_Working_Dog_Manual_2017.pdf
- 5.↑
Davis MS, Marcellin-Little DJ, O’Connor E. Comparison of postexercise cooling methods in working dogs. J Spec Oper Med. 2019;19(1):56-60. doi:10.55460/2ATZ-TMQ7
- 6.↑
Parnes SC, Mallikarjun A, Ramos MT, Stone TA, Otto CM. A randomized cross-over study comparing cooling methods for exercise-induced hyperthermia in working dogs in training. Animals (Basel). 2023;13(23):3673. doi:10.3390/ani13233673
- 7.↑
Kreidler SM, Muller KE, Grunwald GK, et al. GLIMMPSE: online power computation for linear models with and without a baseline covariate. J Stat Softw. 2013;54(10):i10. doi:10.18637/jss.v054.i10
- 9.↑
Michel KE, Brown DC. Determination and application of cut points for accelerometer-based activity counts of activities with differing intensity in pet dogs. Am J Vet Res. 2011;72(7):866-870. doi:10.2460/ajvr.72.7.866
- 10.↑
Boonhoh W, Wongtawan T. The application of human heart rate monitor in dogs: a preliminary study. In: Proceedings of the 23rd Khonkaen Veterinary Annual International Conference. Khonkaen University;2022. doi:10.13140/RG.2.2.24896.00007
- 11.
Harvie H, Rodrigo A, Briggs C, Thiessen S, Kelly DM. Does stress run through the leash? An examination of stress transmission between owners and dogs during a walk. Anim Cogn. 2021;24(2):239-250. doi:10.1007/s10071-020-01460-6
- 12.↑
Shull SA, Rich SK, Gillette RL, Manfredi JM. Heart rate changes before, during, and after treadmill walking exercise in normal dogs. Front Vet Sci. 2021;8:641871. doi:10.3389/fvets.2021.641871
- 13.↑
Schulze LS, Heuwieser W, Arlt SP. Body temperature of bitches in the first week after parturition measured by ingestible loggers. Reprod Domest Anim. 2018;53(suppl 3):63-69. doi:10.1111/rda.13330
- 14.↑
Farr BD, Ramos MT, Otto CM. The Penn Vet Working Dog Center Fit to Work Program: a formalized method for assessing and developing foundational canine physical fitness. Front Vet Sci. 2020;7:470. doi:10.3389/fvets.2020.00470
- 15.↑
R Core Team. R: A language and environment for statistical computing. Published online 2021. Accessed July 25, 2024. https://www.r-project.org/
- 16.↑
Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Softw. Published online June 23, 2014. doi:10.18637/jss.v067.i01
- 17.↑
Lenth RV, Bolker B, Buerkner P, et al. emmeans: Estimated Marginal Means, aka Least-Squares Means. Published July 1, 2024. Accessed July 25, 2024. https://cran.r-project.org/web/packages/emmeans/index.html
- 18.↑
Fox J, Weisberg S. An R Companion to Applied Regression. Published 2019. Accessed July 25, 2024. https://www.john-fox.ca/Companion/index.html
- 19.↑
Adolph EF. Measurements of water drinking in dogs. Am J Physiol. 1938;125(1):75-86. doi:10.1152/ajplegacy.1938.125.1.75
- 20.↑
Gart S, Socha JJ, Vlachos PP, Jung S. Dogs lap using acceleration-driven open pumping. Proc Natl Acad Sci USA. 2015;112(52):15798-15802. doi:10.1073/pnas.1514842112
- 21.↑
Carter AJ, Hall EJ, Bradbury J, et al. Post-exercise management of exertional hyperthermia in dogs participating in dog sport (canicross) events in the UK. J Therm Biol. 2024;121:103827. doi:10.1016/j.jtherbio.2024.103827
- 22.↑
Seeley AD, Sherman RA. An ice vest, but not single-hand cooling, is effective at reducing thermo-physiological strain during exercise recovery in the heat. Front Sports Act Living. 2021;3:660910. doi:10.3389/fspor.2021.660910
- 23.↑
Schranner D, Scherer L, Lynch GP, et al. In-play cooling interventions for simulated match-play tennis in hot/humid conditions. Med Sci Sports Exerc. 2017;49(5):991-998. doi:10.1249/MSS.0000000000001183
- 24.↑
Keiser C, Becker C, Rossi RM. Moisture transport and absorption in multilayer protective clothing fabrics. Text Res J. 2008;78(7):604-613. doi:10.1177/0040517507081309
- 25.↑
Mandal S, Mazumder NUS, Agnew RJ, Song G, Li R. Characterization and modeling of thermal protective and thermo-physiological comfort performance of polymeric textile materials: a review. Materials (Basel). 2021;14(9):2397. doi:10.3390/ma14092397
- 26.↑
Jay O, Capon A, Berry P, et al. Reducing the health effects of hot weather and heat extremes: from personal cooling strategies to green cities. Lancet. 2021;398(10301):709-724. doi:10.1016/S0140-6736(21)01209-5
- 27.↑
Heinrichs K. Superficial thermal modalities. In: Canine Rehabilitation and Physical Therapy. Elsevier; 2004:277-288. doi:10.1016/B978-0-7216-9555-6.50020-6
- 28.↑
Klein BG. Thermoregulation. In: Cunningham’s Textbook of Veterinary Physiology. 6th Edition. Elsevier. W.B. Saunders; 2020:596-607. doi:10.1016/B978-0-323-55227-1.00053-3
- 29.↑
Gillilan LA. Extra- and intra-cranial blood supply to brains of dog and cat. Am J Anat. 1976;146(3):237-253. doi:10.1002/aja.1001460303
- 30.↑
Watanuki S. Effects of head cooling on cardiovascular and body temperature responses during submaximal exercise. Ann Physiol Anthropol. 1993;12(6):327-333. doi:10.2114/ahs1983.12.327
- 31.
Walters P, Thom N, Libby K, et al. The effect of intermittent head cooling on aerobic performance in the heat. J Sports Sci Med. 2017;16(1):77-83.
- 32.↑
Fujii RK, Horie S, Tsutsui T, Nagano C. Effectiveness of a head wash cooling protocol using non-refrigerated water in reducing heat stress. J Occup Health. 2008;50(3):251-261. doi:10.1539/joh.l7097
- 33.↑
Hales JRS. Thermoregulatory requirements for circulatory adjustments to promote heat loss in animals. J Therm Biol. 1983;8(1):219-224. doi:10.1016/0306-4565(83)90108-0
- 34.↑
Hales JRS, Dampney RAL. The redistribution of cardiac output in the dog during heat stress. J Therm Biol. 1975;1(1):29-34. doi:10.1016/0306-4565(75)90008-X
- 35.↑
Tan PMS, Lee JKW. The role of fluid temperature and form on endurance performance in the heat. Scand J Med Sci Sports. 2015;25(suppl 1):39-51. doi:10.1111/sms.12366
- 36.↑
Miwa C, Shimasaki H, Deguchi A, et al. Effects of temperature of drinking water on regulation of body temperature in humans. J Jpn Soc Balneology Climatol Phys Med. 2019;82(2):78-85. doi:10.11390/onki.2324
- 37.↑
Bongers CCWG, Hopman MT, Eijsvogels TM. Cooling interventions for athletes: an overview of effectiveness, physiological mechanisms, and practical considerations. Temperature (Austin). 2017;4(1):60-78. doi:10.1080/23328940.2016.1277003
- 38.↑
Broome CJ, Walsh VP. Gastric dilatation-volvulus in dogs. N Z Vet J. 2003;51(6):275-283. doi:10.1080/00480169.2003.36381
- 39.↑
Elwood CM. Risk factors for gastric dilatation in Irish Setter dogs. J Small Anim Pract. 1998;39(4):185-190. doi:10.1111/j.1748-5827.1998.tb03627.x
- 40.↑
Hall EJ, Carter AJ, Bradbury J, et al. Cooling methods used to manage heat-related illness in dogs presented to primary care veterinary practices during 2016-2018 in the UK. Vet Sci. 2023;10(7):465. doi:10.3390/vetsci10070465
- 41.
Bruchim Y, Klement E, Saragusty J, Finkeilstein E, Kass P, Aroch I. Heat stroke in dogs: a retrospective study of 54 cases (1999-2004) and analysis of risk factors for death. J Vet Intern Med. 2006;20(1):38-46. doi:10.1892/0891-6640(2006)20[38:hsidar]2.0.co;2
- 42.↑
Flournoy WS, Macintire DK, Wohl JS. Heatstroke in dogs: clinical signs, treatment, prognosis, and prevention. Compend Contin Educ Pract Vet. 2003;25(6):422-431.
- 43.↑
Wimer GS, Lamb DR, Sherman WM, Swanson SC. Temperature of ingested water and thermoregulation during moderate-intensity exercise. Can J Appl Physiol. 1997;22(5):479-493. doi:10.1139/h97-031
- 44.↑
Wilkinson DM, Carter JM, Richmond VL, Blacker SD, Rayson MP. The effect of cool water ingestion on gastrointestinal pill temperature. Med Sci Sports Exerc. 2008;40(3):523-528. doi:10.1249/MSS.0b013e31815cc43e
