Cooling Capacity of Transpulmonary Cooling and Cold-Water Immersion After Exercise-Induced Hyperthermia

Twelve recreationally active individuals (42% female, n = 5; Table 1) volunteered to participate in this randomized, crossover design study (ClinicalTrials.gov Identifier NCT0419026). Participants completed 3 experimental exercise and cooling trials: exercise followed by PASS (control trial), exercise followed by transpulmonary cooling using a microenvironmental air-chiller thermal rehabilitation machine (Polar Breeze [PB]; Statim Technologies LLC), and exercise followed by cooling with CWI. Before the trials, each participant completed a medical history questionnaire to ensure the following criteria were met: (1) no chronic health problems; (2) no fever or illness at the time of testing; (3) no history of cardiovascular, metabolic, or respiratory disease; (4) no current musculoskeletal injury that limited the ability to exercise; (5) no history of exertional heat illness within the 3 years before the study; and (6) self-reported engagement in moderate- to vigorous-intensity exercise 3 to 4 d·wk⁻¹ for at least 30 minutes per session. Menstrual status was not measured or used as an exclusionary factor. All participants provided written informed consent, and the study was approved by our university's institutional review board.

Before the start of the exercise trials, participants arrived at the laboratory to become familiarized with the experimental procedures. Each person provided a measure of height, nude body mass (NBM), and body fat percentage. We measured NBM to the nearest 0.1 kg using a digital scale (model WB-800S Plus; Tanita Corp) and height to the nearest 0.1 cm using a wall-mounted stadiometer (model 216; Seca). Body density was measured using the methods established by Jackson and Pollock²⁰  and Jackson et al.²¹  The average of 2 skinfold measures (Lange Skinfold Caliper; Beta Technology Inc) taken at 3 sites on the right side of the body were used; measures were obtained at the chest, abdomen, and thigh sites for men²⁰  and suprailiac, triceps, and thigh sites for women.²¹  Body fat percentage was estimated using the method established by Siri.²²  Participants were also familiarized with the perceptual scales and the use of the PB. After being instructed in the use of the PB, participants performed a graded exercise test on a motorized treadmill (model T150; h/p/cosmos) so that we could assess peak oxygen uptake (VO₂peak). Oxygen uptake and related gas exchange were captured using open-circuit spirometry (model TrueOne 22400; Parvo Medics Inc). The resulting VO₂peak measure and corresponding running speed at which this was achieved were used to identify the participant's relative exercise intensities for each exercise trial.

To ensure that participants were properly hydrated, they were instructed to consume an additional 500 mL of water before bed the night before and the morning of their scheduled exercise trial. Upon arrival at the laboratory, participants provided a urine sample for assessment of urine specific gravity (USG) to ensure euhydration (USG <1.020).²³  If the USG was >1.020, he or she was instructed to consume 500 mL of water before the start of the exercise trial to ensure euhydration. They then provided a pre-exercise NBM (PRE_NBM), inserted a flexible thermistor (model 401AC; Measurement Specialties) 15 cm past the anal sphincter for assessment of T_REC,²⁴  and donned electrodes (model EL503; Biopac Systems Inc) for assessment of heart rate (HR). For all exercise trials, participants wore shorts, T-shirts, and running shoes. Female participants also wore a sport bra.

Participants entered a climatic chamber (model CES-5-43; CANTROL International Inc) set at 40°C and 40% relative humidity and rested passively for 10 minutes to become acclimated to the environmental conditions. After the equilibration period, baseline measures of T_REC, HR, rating of perceived exertion (RPE),²⁵  thermal sensation,²⁶  and fatigue were taken. Fatigue was measured using an 11-point scale with words anchored at the following numbers: 0, no fatigue at all; 1, very small amount of fatigue; 2, small amount of fatigue; 3, moderately fatigued; 4, somewhat fatigued; 5, fatigued; 7, very fatigued; 9, extremely fatigued; and 10, completely fatigued. On a motorized treadmill, participants then began exercise that consisted of 3 sets of 20 minutes of exercise (a 5-minute walk on a 5% incline at a speed equal to 30% of their VO₂peak, followed by a 15-minute jog on a 1% incline at a speed equal to 70% of their VO₂peak). Exercise continued until (1) T_REC equaled 39.99°C, (2) the participant asked to stop, (3) the participant exhibited an altered or uneven gait, (4) HR was >5 beats/min from the participant's maximum for 5 minutes, or (5) 60 minutes of exercise had elapsed. Throughout exercise, T_REC and HR were measured continuously, and perceptual scales were assessed periodically.

Immediately after exercise, participants stepped off the treadmill and took a seated position while remaining in the climatic chamber. Participants were randomly assigned to the PASS, PB, or CWI intervention. Cooling ceased after T_REC reached 38.25°C to provide an adequate cooling curve for calculation of cooling rates. For PASS, participants sat in a comfortable position in a chair in the climatic chamber and relied on their natural ability to dissipate stored body heat. For PB, participants donned a hood with a volume of 22 240 cm³ that permitted cooled air (22.2°C ± 1.0°C) to flow into the hood from the thermal rehabilitation machine per the manufacturer's instructions (Figure 1). Ambient air (40°C and 40% relative humidity) inside the climatic chamber was drawn into the intake of the PB, where it passed through a series of thermocouples to cool the air. After being cooled, the air was pumped through the output hose and into the hood for participants to inhale. For CWI, participants sat in a tub of ice water (75.7 L of water, 75.7 L of ice, temperature = 1.45°C ± 2.3°C) and were immersed approximately to the level of the xiphoid process. During cooling, the ice water was continuously circulated to maximize convective cooling.

After cooling ended, we obtained the final physiologic and perceptual measures, and participants exited the climatic chamber. They dried off, removed the rectal thermistor, and provided a final NBM (POST_NBM) so that we could calculate the extent of body mass loss due to dehydration: [(POST_NBM – PRE_NBM)/PRE_NBM] × 100.

One-way analyses of variances (ANOVAs) were performed to assess differences among pre-exercise USG, percentage of body mass loss, and exercise time for each trial. A condition × time repeated-measures ANOVA was used to assess differences in NBM for each trial. To determine differences in the independent variables (cooling rates and time to cool to the established temperature thresholds between trials), we conducted separate 1-way ANOVAs. When we found a significant difference between trials, Tukey post hoc analyses were used to identify where the differences occurred. The magnitudes of differences were measured using η² , with effect sizes interpreted as small (η² < 0.01), medium (0.01 < η² > 0.06), or large (η² > 0.14). Mean differences (MDs) and 95% the RPE CIs depicted the magnitude of differences. A post hoc power analysis (G*Power 3.1; Heinrich-Heine Universität, Düsseldorf) comparing the means of the cooling rates for CWI, PASS, and PB (primary outcomes) confirmed that the study was sufficiently powered (0.97). The α level was set a priori at .05. Statistical analyses were performed using SPSS (version 25; IBM Corp, Armonk, NY).

Participant characteristics across all 3 trials are presented in Table 2. The percentage of body mass loss was lower in CWI than PASS (MD [95% CI]; −0.7% [−1.2%, 0.2%]; P = .02). We observed no differences among trials for USG, NBM, T_REC, or exercise time (P > .05; Table 2).

Comparisons of average cooling rates across time for each cooling trial revealed that cooling rates were greater during CWI (η² = 0.862) than during PASS (0.16°C·min⁻¹ [0.13°C·min⁻¹, 0.19°C·min⁻¹]; P < .001) and PB (0.15°C·min⁻¹ [0.12°C·min⁻¹, 0.16°C·min⁻¹]; P < .001; Figure 2). The time to reach T_REC of 38.25°C was also shorter with CWI (η² = 0.711) than with PASS (−58.1 minutes [−77.1, −39.9 minutes]; P < .001) and PB (−46.8 minutes [−65.5, −28.2 minutes]; P < .001; Figure 3). Cooling rates and the time to reach a T_REC of 38.25°C between PASS and PB were not different (P > .05).

Changes in RPE, thermal sensation, and fatigue during cooling are depicted in Table 3. RPE, thermal sensation, and fatigue were lower across all trials at postcooling than at precooling (P < .05). Thermal sensation was lower during PB (P = .03) and CWI (P < .001) than during PASS at postcooling. We observed no differences among trials for RPE and fatigue (P > .05).

The purpose of our study was to examine the efficacy of transpulmonary cooling in reducing body temperature after exercise-induced hyperthermia. Our primary findings showed that PB cooled the body more slowly (0.03°C·min⁻¹) than CWI (0.18°C·min⁻¹). Furthermore, the cooling rate of PB was not different from that of PASS (MD [95% CI]; 0.01°C·min⁻¹ [−0.03°C·min⁻¹, 0.04°C·min⁻¹]; P > .05), providing further evidence that the inhalation of cooled air afforded by PB had no added cooling benefit beyond the body's ability to cool passively.

The ability to cool the body quickly after exercise-induced hyperthermia, especially when exertional heat illness is suspected, is paramount to ensuring survival.^1,27  Researchers have suggested that the optimal cooling rate to ensure survival from EHS exceeds 0.15°C·min⁻¹,^{1,4,7,8,27,28}  with the goal of reducing the body temperature to 38.89°C within 30 minutes from the moment of collapse.^2,3,27  We confirmed that CWI exceeded the optimal cooling rate recommended for EHS treatment and reduced T_REC to our cooling threshold cutoff of 38.25°C in 9.71 ± 3.30 minutes. In comparison, the PASS and PB experimental trials averaged 68.22 ± 27.60 and 56.56 ± 25.52 minutes, respectively, to cool to 38.25°C (Figure 3).

When the cooling threshold cutoff was adjusted to 38.89°C, which aligns with the threshold to cease cooling during EHS care, CWI (6.85 ± 2.48 minutes) remained the best cooling option compared with PASS (38.87 ± 13.13 minutes) and PB (25.08 ± 10.82 minutes). Although PB exhibited an overall cooling time of <30 minutes, T_REC at the onset of cooling in PB averaged 39.70°C, which was less than the diagnostic criteria (40°C) for EHS. If these findings were extrapolated to a patient with EHS and an initial T_REC of 41.44°C⁴  was assumed, 85 minutes (SD range = 63.75–127.5 minutes) would have been needed to cool the patient, which is well past the 30-minute period for ensuring survival.

Mariak et al¹³  found that when patients inhaled cool air (22°C) after passive hyperthermia, they exhibited immediate declines in brain and esophageal temperature, providing evidence that the lungs can effectively cool the brain and a valid measure of body temperature used during exercise (ie, esophageal temperature). However, the rate of decline in T_REC did not match the former measures. Conversely, Kumar et al¹²  noted that inhalation of cooled heliox (70% He, 30% O₂; 0°C ± 2°C) reduced T_REC in a swine model at a rate similar to that of esophageal, pulmonary artery, and tympanic membrane temperatures. Our results showed that transpulmonary cooling had no added cooling benefit for T_REC compared with passive cooling. An argument could be posed that because we assessed T_REC, we would not have been able to capture the acute changes in body temperature that may have occurred in the esophagus or tympanic membrane, as reported by earlier authors. Although this argument may have merit, given the medical literature on the pathophysiological cascade emanating from the gastrointestinal tract with the infiltration of endotoxins into the systemic circulation during EHS^29–32  and the current best medical practices regarding the recognition of EHS,^2,3,27  assessing the effect of transpulmonary cooling on T_REC is imperative.

We observed that after exercise-induced hyperthermia, RPE, thermal sensation, and fatigue declined at the postcooling timepoint. Furthermore, thermal sensation was less with both the CWI and PB trials than with the PASS trial at postcooling, which is supported by evidence suggesting that reductions in skin temperature after exercise alleviate thermal discomfort.^33,34  Based on the differences in thermal sensation among trials at postcooling that stemmed from afferent feedback of the thermoreceptors on the skin, we speculated that thermal sensation during PB may have been lower throughout cooling than during PASS, despite no difference in cooling rates. The latter could create a scenario in which thermal sensation may be reduced (ie, a sustained elevated internal body temperature), and another bout of physical activity (eg, following thermal rehabilitation in occupational settings, halftime during sport) may increase the risk profile of exertional heat illness.

Our study had a number of strengths. First, our randomized crossover clinical trial design allowed participants to act as their own controls when we compared the experimental trials. Second, the comparison of the PB with CWI in both men and women after exercise-induced hyperthermia (>39.75°C) allows these results to be directly translated to clinical practice. Third, using a controlled laboratory setting allowed us to control factors that may influence cooling capacity, such as environmental conditions and hydration status.

However, this work was not without limitations. We examined only the utility of the PB on postexercise cooling, which did not provide any evidence on the ability of the PB to attenuate the rise in body temperature from repeated bouts of exercise in a hot environment, such as during firefighting operations or at halftime in an athletic competition. We were also unable to reduce the air temperature from the PB to <22.2°C ± 1.0°C because of the machine constraints in cooling the air from the external environment. Further reducing the temperature of the inspired air may offer added benefits to the body's ability to cool. We were also unable to discern the effect of transpulmonary cooling with the PB given that the cooled air was inspired and then passed over the face and torso of participants.

After exercise-induced hyperthermia, transpulmonary cooling using the PB did not exhibit cooling rates similar to those of CWI. For the treatment of EHS, whole-body CWI remains the criterion standard.

We acknowledge Statim Technologies for providing funding for this study.