INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MANY OCCUPATIONS (e.g., firefighters, soldiers, astronauts, explosive ordinance, toxic waste clean-up) require workers to wear protective clothing with characteristic high insulation and low moisture permeability properties. These conditions impose uncompensable heat stress (required evaporative cooling exceeds evaporative cooling capacity of environment) that results in rapid heat storage and a reduction in work capabilities (2, 6, 8, 16). Microclimate cooling (MCC) systems are designed to remove heat from the skin by using ice-packet vests, cooled air, or by circulating cooled liquid in tubes. Each of these methods is effective in reducing heat strain and extending work performance (12). For most military, space, and firefighting applications, liquid-cooled systems have several advantages over other MCC approaches, including reduced logistical requirements and sustainable high cooling capacities (10, 20).

Engineering approaches for developing optimal liquid MCC systems have focused on enhancing power (battery capacity) to increase cooling capabilities (>300 W) (20). Unfortunately, traditional engineering approaches for improving cooling effectiveness (i.e., reducing the perfusate temperature and increasing flow) may increase power requirements. From a physiological perspective, these engineering approaches are potentially self defeating. Skin cooling can produce cutaneous vascular constriction that decreases convective heat transfer from the body core to the periphery (13, 14, 23). Superficial shell insulation approaches near maximal values at skin temperatures (T_sk) of 30°C, with the onset of vasoconstriction occurring between 32 and 33°C (23). Thus the heat loss advantage obtained by widening the core-to-T_sk gradient (T_re-T_sk) is progressively reduced by increased superficial shell insulation as T_sk drops below 32°C (5, 23). Attempts to circumvent MCC limitations are presently limited to cooling of body regions with low vasomotor tone (head) (11) and regional cooling of the skin overlying the active skeletal muscle (25) due to the greater heat flux from active muscle during exercise (15).

No study has compared the effects of intermittent regional cooling (IRC) to constant cooling (CC) of the skin for improving MCC effectiveness. We theorized that the heat flux benefits from periodically applying cooling to warm vasodilated skin might offset the potential for vasoconstriction that occurs with continuous skin cooling. In support of this belief, Constable et al. (3) observed that the first few minutes of liquid cooling are the most effective. In addition, if IRC improves heat flux, issues related to both the optimal body surface area (BSA) and the proper perfusion periods for cooling must also be resolved. Therefore, the purpose of this experiment was to test the effectiveness of four different IRC regimens by manipulating the BSA cooled and the duration that perfusion was applied. Our hypothesis was that IRC might allow for improved skin heat flux, thereby reducing the BSA and perfusion period necessary to achieve a given reduction in exercise heat strain when compared with CC paradigms.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Five healthy men served as volunteers for this study (mean ± SD; age 26 ± 7 yr, height 180 ± 4 cm, weight 82 ± 8 kg, BSA 2.0 ± 0.1 m²). All subjects were medically cleared and provided written, informed consent regarding the risks and requirements of participation. The appropriate Institutional Review Boards approved the study, and the investigators adhered to AR 70-25 and US Army Research Institute of Environmental Medicine 70-25 on Use of Volunteers in Research.

Experimental procedures. One week before experimental testing, subjects performed a 5-day heat acclimation protocol. Briefly, subjects walked on a motor-driven treadmill at 1.36 m/s while at a 2% grade for 100 min (2 repeats of 50-min exercise with a 10-min rest between) in a hot-wet environment (35°C, 50% relative humidity, wind speed 2.2 m/s) while wearing a t-shirt, shorts, socks, and athletic shoes. Water intake was provided ad libitum during exercise, and subjects were rehydrated after exercise to within 1% of initial body weight. Experimental testing began after completion of the heat acclimation program.

Clothing. Subjects were dressed the same for each trial (spandex shorts, cotton socks, and athletic shoes) before being fitted with a three-piece liquid cooling garment (LCG) to include coverage of the head (hood), torso (vest), and legs (pants). The LCG was stored at ambient test conditions to allow the water inside to reach equilibrium with room temperature before the garment was donned. The LCG design consisted of cotton or Nomex aramid fabric woven or laminated around small-diameter Tygon tubing [2.5-mm internal diameter (ID)] divided into multiple parallel circuits. Total estimated tubing length for the ensemble was ~108 m. Total BSA covered was 72% (head = 6%, torso = 22%, and legs = 44%) as measured with a Cyberware three-dimensional head and whole body scanner (Cyberware, Monterey, CA). After LCG fitting, subjects were dressed in a chemical protective clothing system that included a charcoal-impregnated overgarment (top and bottom), cotton glove liners, butyl gloves, and M-40 chemical-biological field mask with hood. Based on copper manikin studies in still air, this modified mission-oriented protective posture (MOPP-3) configuration provided approximate insulative and vapor permeability characteristics of 2.1 and 0.32, respectively (6). Fully clothed body mass was recorded before the start of exercise. At the completion of exercise, body mass was again recorded both fully clothed and semi-nude.

Design. All subjects completed six experimental trials (Fig. 1) in a warm, dry environment (dry bulb temperature = 29.8 ± 0.4°C; dewpoint temperature = 10.9 ± 0.1°C, equivalent to 30% relative humidity). Wind speed averaged 0.7 m/s at the back and 0.2 m/s at the chest. Treadmill exercise (1.36 m/s, 2% grade) was performed at 224 ± 5 W/m² for 80 min during each of six trials. Figure 1 illustrates the paradigms as CC to four body regions covering 72% of BSA (CC), IRC to two body regions covering a mean 36% of BSA with on-to-off perfusion ratios of 2:2 min (IRC₁) or 4:4 min (IRC₂), IRC to two body regions covering a mean 18% of BSA with 1:3 min (IRC₃) or 2:6 min (IRC₄) perfusion ratios, and NC. All paradigms were tested in a counterbalanced order. Inlet water temperature was maintained at 21°C to approximate the most common self-selected temperature for thermal comfort (17).

View larger version (52K): [in this window] [in a new window]   Fig. 1.   Summary of the 6 intermittent regional cooling (IRC) paradigms. ×, active cooling of the region. BSA, body surface area; CC, constant cooling; NC, no cooling.

Calculations. Mean T_sk were calculated from the formula T_sk = 0.07T_head + 0.10T_{upper back} + 0.10T_{lower back} + 0.10T_chest + 0.10T_abdomen + 0.14T_forearm + 0.19T_thigh + 0.20T_calf (4), where the subscript tells the location of where the temperature was taken. The T_re  T_sk gradient was calculated as the difference between T_re and T_sk. Mean body temperature (T_b) was calculated from T_re and T_sk as T_b = xT_re + (1  x)T_sk, where x is the appropriate weighting coefficient (0.8) for environments ranging from cold (22) to thermoneutral (1). Whole body sweating rate was determined from change in semi-nude body mass corrected for respiratory water loss and CO₂-O₂ exchange (7). The ratio (%) of evaporative to nonevaporative sweat loss was calculated as (semi-nude mass  clothing mass/semi-nude body mass) × 100. The physiological strain index was calculated by using the originally published formula (9).

Statistical analysis. Data were analyzed by using commercial software (SigmaStat, SPSS Science, Chicago, IL). Both one-way (trial) and two-way (trial × time) analyses of variance for repeated measures were performed. Tukey's honestly significant different post hoc test was applied when significant main or interaction effects were found. Five subjects were calculated a priori to provide adequate power to detect a 1.5-fold difference (± 0.3°C) in T_re among trials by using a standard deviation of 0.2°C,  = 0.05, and  = 0.20. Statistical significance was set at P < 0.05. All data are reported as means ± SD.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All six trials were experimentally the same except for the cooling strategy employed. No differences were observed among these trials for environmental conditions (P > 0.05). Although metabolic rate ranged narrowly from 217-235 W/m², the difference between the extremes (paradigms CC and NC) was statistically significant (Table 1). Complete thermoregulatory data was obtained for 30 of 30 trials.

Thermoregulatory strain. The efficacy of CC and IRC paradigms on reducing the T_re response to NC are represented in Fig. 2 and Table 2. All cooling paradigms produced a significantly lower T_re between 50 and 75 min of exercise when compared with the NC experiment, with no additional differences among trials. Each cooling paradigm produced a smaller (P < 0.05) change in T_re than NC, whereas IRC₁ was also lower than IRC₄ (Table 2). Figure 3 depicts changes in local thigh temperatures produced by altering LCG perfusion ratios. Reproducible temperature fluctuations of ~1°C were observed for paradigms IRC₁ and IRC₃. When the off period of perfusion was >3 min, fluctuations of ~2°C occurred as shown for IRC₂ and IRC₄. The area-weighted mean of eight local skin sites is shown across time in Fig. 4A. At all measured time points during exercise (10-75 min), T_sk was significantly higher for paradigm NC and lower for paradigm CC when compared with all other trials (P < 0.05). T_sk was lower (P < 0.05) for paradigms IRC₁ and IRC₂ when compared with IRC₃ and IRC₄ during the final 45 min of exercise. T_re  T_sk gradient across time (Fig. 4B) illustrates the significantly smaller difference for paradigm NC and larger difference for paradigm CC when compared with all other trials (P < 0.05). Gradients for paradigms IRC₁ and IRC₂ were larger than IRC₃ and IRC₄ throughout the final 25 min of exercise. I_t was highest for paradigm CC (0.026 ± 0.005°C · W¹ · m²), followed by 36% BSA cooling trials (IRC₁ and IRC₂) > 18% BSA cooling trials (IRC₃ and IRC₄) > NC (Fig. 5A). I_t and T_sk were highly correlated (r = 0.92, P < 0.05; Fig. 5B).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Influence of cooling paradigm on rectal temperature (Tre) over time. Values and bars are means ± SE. Some error bars have been removed for clarity. * NC paradigm significantly different from all others at 50-75 min.

View larger version (28K): [in this window] [in a new window]   Fig. 3.   Effect of 6 IRC paradigms on variability in local thigh temperature. Data represent the mean for all subjects measured continuously. Shaded areas represents temperature bands for optimal heat transfer (5, 13, 14, 23).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Influence of cooling paradigm on mean weighted skin temperature (Tsk; A) and core-to-Tsk (Tre-Tsk) gradient (B) over time. A: * paradigms IRC1 and IRC2 are significantly different from paradigms IRC3 and IRC4 at 30-75 min. Paradigms CC and NC are significantly different from all others at 10-75 min. B: * paradigms IRC1 and IRC2 are significantly different from paradigms IRC3 and IRC4 at 50-75 min. Paradigms CC and NC are significantly different from all others at 10-75 min. Some error bars have been removed for clarity. Shaded areas represent temperature bands for optimal heat transfer (5, 13, 14, 23).

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Influence of cooling paradigm on mean total body insulation (It). A: * significantly different from all other paradigms. ** Significantly different from IRC4 and NC. *** Significantly different from NC. B: relationship between mean It and Tsk (r = 0.92, P < 0.05). Dotted lines show Tsk that approach the asymptotes for maximal and minimal It (23). Shaded area represents temperature band for optimal heat transfer (5, 13, 14, 23).

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Comparison of cooling efficiency among experimental IRC paradigms. Dotted line represents 100% efficiency (paradigm CC). Efficiency calculations presented in text. * Significantly different from paradigm IRC4.

Cardiovascular strain. The influence of the six experimental trials on HR are depicted in Fig. 7. All cooling paradigms produced a significantly smaller rise in HR between 30 and 75 min of exercise when compared with the NC experiment (P < 0.05). Smaller, but significant, differences (9-12 beats/min) were also observed between paradigm CC and IRC₄ during the final 25 min of exercise. All cooling trials resulted in smaller HR responses when compared with NC (Table 2), whereas HR for paradigm CC was also significantly less than IRC₃ and IRC₄ (P < 0.05). As shown in Fig. 7B, the influence of T_sk on HR was curvilinear. When T_sk rose from 32 to 35°C (3°C), HR increased 26 beats/min. However, this cardiovascular penalty nearly doubled (49 beats/min) when T_sk was increased 4°C from 32 to 36°C. Although statistical differences were observed between some cooling trials on a combined measure of HR and T_re strain (physiological strain index) (Table 2), all were within the same strain category (low) except for trial NC (high) (9).

View larger version (15K): [in this window] [in a new window]   Fig. 7.   Influence of cooling paradigm on heart rate. A: * paradigm CC significantly different from paradigm IRC4 at 50-75 min. Paradigm NC significantly different from all others at 30-75 min. Some error bars have been removed for clarity. B: curvilinear relationship between change in heart rate (HR) and Tsk [y = 3,668 + 225.8(Tsk) + 3.499(Tsk)2]. Shaded area represents temperature band for optimal heat transfer (5, 13, 14, 23).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study examined the efficacy of IRC to reduce thermoregulatory and cardiovascular strain when compared with a traditional CC paradigm. Our hypothesis was that cycling coolant to different body regions on an intermittent basis would reduce the vasomotor response to cooling and maintain heat flux from the core to the skin surface. The principal finding of this study was that IRC to a smaller BSA was equally effective in reducing exercise heat strain when compared with CC of a larger BSA.

IRC maintained T_sk and the T_re-T_sk gradient within ranges (~34-35°C and ~3-4°C; Fig. 4) associated with greater cutaneous vasodilation and a larger heat flux from the core to the skin (5, 13, 14, 23). T_sk was maintained between the T_sk approaching the asymptotes for minimal (35°C) or maximal (32°C) I_t, respectively (23) (Fig. 5B). In contrast, CC reduced T_sk to ~32°C, increased the T_re-T_sk gradient to ~5°C (Fig. 4), and produced T_sk associated with near maximal I_t (Fig. 5B) (23). Because relatively low T_sk are associated with the onset of vasoconstriction, increased I_t (23), and reduced total body skin blood flow (5), IRC resulted in more optimal conditions for cutaneous heat flux when compared with CC. These physiological data help explain the derived values of K (Table 1) that show no differences in Newtonian cooling between CC over 72% BSA and IRC₁ or IRC₂ over 36% BSA. All IRC paradigms maintained T_re to <50% of NC (range = 30-46%), and all T_re were kept well below a temperature increase of 1°C (Table 2). As a result, IRC may provide an engineering power advantage by reducing the relative perfusion area and absolute perfusion time necessary for reducing exercise heat strain.

All four IRC trials were more efficient (164-215%) than CC (Fig. 6). This finding is in direct contrast with efficiencies reported for individual regions perfused constantly (19) but is explained by the changes in heat flux produced by IRC (see above). The "efficiency" of different LCG configurations has been reported by using total (19) or net (18, 19) BSA covered to reduce body heat storage by a fraction (%) of that observed without cooling. In our calculation of CE, we chose to use total BSA because 1) net surface area is generally only an estimate of the percentage of tubing actually in contact with the skin (18, 19), and 2) we were able to very accurately measure the BSA covered by our ensemble by using a three-dimensional whole body and head scanner (Cyberware). Although the actual BSA in contact with the LCG worn was the same for all trials, regional perfusion and the duration of perfusion were altered (Fig. 1). IRC paradigms represented mean regional perfusions of 50 or 25% of the BSA perfused during CC. CE was calculated as the ratio of cooling provided (T_reNC  T_rex) per unit surface area (%) covered, which is similar to previous methods (19). Therefore, IRC paradigms offer an advantage over traditional CC by providing a comparable level of cooling with a 50-75% reduction in BSA cooled.

The higher T_sk afforded by IRC increased heat flux over a smaller BSA with only a small cardiovascular penalty. The HR that resulted from an increase in T_sk from 32 to 34°C (2°C) was 10 beats/min (Fig. 7B). From 32 to 35°C (3°C), HR increased to 26 beats/min but then nearly doubled to 49 beats/min as T_sk rose from 32 to 36°C (4°C). The universal scale for the physiological strain index (9) suggests that the combination of thermal and cardiovascular stress was categorically the same (low = 2-4) for all cooling paradigms but distinguishably high (8) for NC. When combined, Figs. 5B and 7B support that the maintenance of T_sk between 33 and 35°C results in increased heat flux without significant additional cardiovascular strain.

Both the physiological (T_sk, T_re, HR) and biophysical (K) data support the use of IRC as a novel approach to MCC. Constable et al. (3) observed that, when subjects were actively cooled only during rest periods, the first few minutes of cooling were the most effective. However, this "heat sink" phenomenon decays over time as T_sk declines and progressively narrows toward tubing temperature (T_sk  T_t) during active cooling. To overcome this limitation, IRC was used to allow T_sk to fluctuate systematically (Fig. 3). Perfusion cycles (on/off) maintained T_sk and T_re-T_sk within a narrow operational band (Fig. 4) commonly associated with optimal heat flux (5, 13, 14, 23). The off period of perfusion also allowed time for water inside the LCG, held without flow against the skin, to continually absorb some heat as T_sk increased. Thus, when rechilled coolant was again perfused through the LCG in the on period, the heat absorbed by the LCG (heat sink) was removed and T_sk was gradually lowered to begin another cycle. With the use of thigh temperatures as an example (Fig. 3), an increase in the outlet-inlet gradient equal only to the difference in T_sk between CC and IRC trials (~2°C) would produce a 60% increase in heat transfer for an equal amount of cooling time once pumps were turned from off to on, in accordance with the formula [heat transfer (J/s or W) = flow of coolant (ml/s) × specific gravity of coolant (g/ml) × specific heat of coolant (J/g) × (outlet-inlet water temperature gradient) (°C)] (19, 21).

The on-to-off ratios within like paradigms (IRC₁ vs. IRC₂ or IRC₃ vs. IRC₄) had no measurable impact on cooling so long as T_sk and T_re-T_sk were maintained within an optimal band (Fig. 4). Although T_sk fluctuations were larger for paradigm IRC₂ when compared with IRC₁ (36% BSA trials), no differences in K were observed. Longer on periods of perfusion in IRC₂ drove local thigh temperature to absolute lower levels, but the longer corresponding off period produced temperature peaks similar to IRC₁ (Fig. 3). In IRC₃ and IRC₄, T_sk was maintained within the optimal temperature band longer than IRC₁ and IRC₂, but no greater cooling was observed, probably because of the smaller BSA cooled (18%). The design of this experiment does not allow an "ideal" BSA or on-to-off perfusion ratio to be identified. However, the data support that the local temperature fluctuations observed for IRC₃ and IRC₄ (1-2°C within the 33-35°C temperature range), spread out over 36% BSA instead of 18%, could represent the best potential combination of perfusion and SA cooling.

The substantial cooling benefit observed by simply wearing the LCG without perfusion (paradigm NC) was an unexpected finding (Table 1) and a limitation to this experiment. Although the high specific heat characteristics of water compliment IRC, the small volume of water in the LCG (<400 ml) would itself only absorb a fraction of the heat calculated in Table 1, paradigm NC (97 W/m²) (<2% with the use of c × mass × change in temperature, assuming a temperature change of 7°C). Heat must have also been conducted from the LCG to the outer garment and was probably lost through radiation to the cooler surrounding environment. This proposition would alter (R + C) and K calculations in Table 1; however, the actual garment and LCG fluid temperatures required to distinguish between these specific heat loss avenues were not measured.

We conclude that IRC is a significant and comparable countermeasure to thermoregulatory and cardiovascular strain when compared with CC paradigms when used in combination with protective clothing. Furthermore, IRC represents a more efficient means of removing body heat and has implications for reducing MCC power requirements. No ideal BSA or perfusion ratio could be identified as optimal. However, because different body surface regions will both fluctuate in temperature and respond to cooling differently based on the mode of activity and cutaneous vascularity (11, 15, 17, 25), a promising next generation of IRC experiments is the development of an IRC system that maintains a small range in T_sk and T_re-T_sk through automated control by physiologically driven (T_sk) feedback loops.