INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MORAN AND COLLEAGUES (9) recently developed a cold strain index (CSI) that allows quantification of physiological cold strain in real time. CSI is calculated as a weighted average of core temperature (T_core) and mean skin temperature (_sk) differences from initial baseline values and rated on a universal scale of 1-10. The CSI was developed using data from three studies (4, 10, 17) that exposed resting seminude subjects to either cold air or water. During these types of experimental conditions, skin temperature (T_sk) and T_core typically decrease. However, during exercise in the cold, T_core and T_sk may respond differently. For example, depending on exercise intensity and clothing insulation, T_core might increase, remain constant, or decrease. Likewise, T_sk values are influenced by the insulation and exercise intensity. Whether CSI can provide a useful/meaningful estimate of physiological strain in people exposed to cold-wet conditions and exercising is unknown.

An alternative approach to assessing physiological strain in humans exposed to cold might be to obtain subjective evaluations from volunteers. Little is known about subjective perception of effort and thermal cues during exercise-cold stress. Toner et al. (13) demonstrated that during exercise in cold water, rated perceived exertion (RPE) was moderately correlated with heart rate (HR), ventilation, and oxygen uptake (O₂) but not associated with rectal temperature (T_re) or T_sk. Conversely, thermal sensation was associated with T_re and T_sk but not with cardiopulmonary cues (13). Exercise in cold air may lead to perceptual ratings that rely on different cues because T_core will tend to rise at the onset of exercise, and HR during prolonged exercise-cold stress does not usually exhibit cardiovascular drift (3, 15).

The purpose of this study was to determine whether meaningful quantification of physiological strain could be obtained by applying the CSI to body temperatures measured during cold-wet exposure in clothed, exercising men. In addition, two subjective comfort scales (RPE and thermal sensation) were compared with CSI. Experiments were performed on subjects before and after completing 3 days of exhaustive exercise that resulted in a reduced ability to thermoregulate (3). It was hypothesized that the CSI would reflect the increased cold strain across time during cold-wet exposure, but it would not be sensitive enough to discriminate the reduced tissue insulation that occurs after 3 days of exhaustive exercise (3). Furthermore, it was hypothesized that during exercise-cold stress, RPE and thermal sensation would be positively correlated with both cardiopulmonary and thermal cues and with CSI.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Ten male volunteers participated in this study, which was approved by the appropriate institutional review boards. The subjects provided written consent after being fully informed of the requirements and risks associated with participating in the research. Subject characteristics were (means ± SE) age, 24 ± 1 yr; height, 177 ± 2 cm; weight, 82.8 ± 3.6 kg; %fat, 16.4 ± 1.9%; peak O₂ (O_2 peak), 56.0 ± 1.8 ml · kg¹ · min¹; and body surface area, 1.99 ± 0.05 m².

Preliminary testing. Body composition, O_2 peak, and one-repetition maximum of the upright row, chest press, latissimus dorsi pull down, and biceps curl were determined as previously described (3).

Experimental design. After preliminary testing, subjects completed two experimental cold, wet walks (CW). Both CW took place from ~1330 to 2000 with the first CW completed when they were well rested before beginning the heavy exercise regimen (D0) and the second CW completed after 3 consecutive days of exhaustive exercise (D3; see Ref. 3 for details of exhaustive exercise regimen). On D3, ~2.5 h (140-170 min) elapsed between the end of the last daily exercise session and the subsequent CW. The CW was modified from an experimental protocol described by Weller et al. (15). Briefly, CW consisted of 360 min of intermittent treadmill walking (6 cycles of 10 min of standing rest in the rain, 45 min of walking, and 5 min for transition between rest and walking) in an environmental chamber with air temperature set at 5°C. During the rain, the subjects stood still for 10 min (except for the initial cycle of rain, during which they sat) and were wetted at a rate of 5.41 cm/h under a sprinkler designed to simulate rainfall. After each rest/rain period, subjects walked at 1.34 m/s (3 miles/h) at 0% grade on a motor-driven treadmill. Wind velocity was 1.1 m/s (2.5 miles/h) during the 10-min rain and 5.4 m/s (12 miles/h) while walking. The CW for each subject was terminated if T_re was <35°C or if the subject asked to stop.

2

1

Measurements and calculations. CSI was calculated according to Moran et al. (9) as follows: CSI = 6.67(T_core t  T_core 0) · (35  T_core 0)¹ + 3.33(_{sk t  sk 0}) · (20  _sk 0)¹, where T_core 0 and _sk 0 are the baseline measurements, and T_core t and _sk t are simultaneous measurements taken at any time t during the CW; when T_core t > T_core 0, then T_core t  T_core 0 = 0. During the experiments, T_re (i.e., T_core) was measured every minute using a thermistor inserted 10 cm past the anal sphincter. T_sk was measured using thermistor disk sensors (Concept Engineering, Old Saybrook, CT) attached on the skin surface at five sites (ventral aspect of forearm, triceps, subscapula, anterior thigh, and calf). _sk was calculated as _sk = 0.28T_subscapular + 0.14T_forearm + 0.08T_triceps + 0.22T_calf + 0.28T_thigh. O₂, carbon dioxide output, and minute ventilation were measured by open-circuit spirometry before CW (sitting) and during the 25th to 27th minutes of walking during each exercise portion of the rest-walking cycle. Percent oxygen (Applied Electrochemistry S-3A), carbon dioxide (Beckman LB-2), and volume (Tissot Spirometer, Collins) measurements were collected as previously described (3). HR was measured near the end of each walking portion of the CW from three chest electrodes (CM-5 configuration) and radiotelemetered to an oscilloscope-cardiotachometer (Hewlett-Packard). Thermal sensation and RPE were evaluated using a 17-point category scale (0 to 8.0, 0.5 increments, lower number corresponds to higher cold sensation) (18) and the Borg scale (2), respectively.

Statistical analyses. Data were analyzed using a two-factor (time × experimental trial) repeated-measures ANOVA. When significant F ratios were calculated, paired comparisons were made post hoc using Fisher's least significant difference test. Because exposure duration varied for each subject between the trials, statistical analysis was performed on complete data sets for both trials. Therefore data from minute 0 to minute 180 were analyzed using n = 10, and data from minute 190 to minute 360 were analyzed using n = 4. Relationships between variables were determined using Pearson product-moment correlations. Unless otherwise specified, the level of significance for differences reported is P < 0.05. Data are presented as means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

CSI and body temperatures. Figure 1 presents the CSI values during exercise in both trials. CSI increased (P < 0.05) from ~2.5 U during minutes 30-60 to 4.2-4.6 U during minutes 180-360. There were no differences in CSI during exercise-cold exposure between D0 and D3.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Cold strain index (CSI) vs. time during exercise-cold exposure before (D0) and after 3 days of exhaustive exercise (D3). Data from 0 to 180 min are from 10 subjects, and data from 190 to 360 min are from 4 subjects.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Change in rectal temperature (Tre; A) and change in mean skin temperature (sk; B) vs. time during cold exposure before (D0) and after 3 days of exhaustive exercise (D3). Data from 0 to 180 min are from 10 subjects, and data from 190 to 360 min are from 4 subjects. * D3 significantly (P < 0.05) different from D0. Fisher's least significant difference was used for post hoc analysis. The absolute initial Tre was 37.30 ± 0.08 and 37.44 ± 0.05°C for D0 and D3, respectively. The absolute initial sk was 33.84 ± 0.18 and 34.39 ± 0.13°C for D0 and D3, respectively.

RPE and thermal sensation. Figure 3 presents the RPE and thermal sensation data for both trials. RPE was not significantly different between trial days. Thermal sensation values were significantly higher on D3 vs. D0, before and during the 1st hour of cold exposure. Table 1 shows the relationship between the physiological cues and RPE and thermal sensation. Both experimental conditions were combined for these correlational analyses. There were moderate correlations between RPE and HR, ventilation, metabolic heat production, and _sk. Thermal sensation was highly correlated with ventilation, metabolic heat production, and _sk, whereas there was a moderate correlation with HR and a small correlation with T_re. There were no significant correlations between CSI and thermal sensation (r = 0.20) and between CSI and RPE (r = 0.07).

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Rating of perceived exertion (A) and thermal sensation (B) vs. time during cold exposure before (D0) and after 3 days of exhaustive exercise (D3). Data from 0 to 180 min are from 10 subjects, and data from 190 to 360 min are from 4 subjects. * D3 significantly (P < 0.05) different from D0. Fisher's least significant difference was used for post hoc analysis.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

CSI integrates T_core and T_sk on the basis of their relative contributions to thermoregulatory effector responses (1, 5, 12, 14). This study's principal finding was that CSI was not sensitive to degraded thermoregulatory responses and increased cold strain after several days of exhaustive exercise (3). Compared with exercise-cold stress when subjects were well rested, T_re fell more and T_sk was elevated when exhausted, due to a reduced ability to maintain tissue insulation, but the CSI could not differentiate between experimental trials.

A criticism of the CSI is that there is no universally agreed-on index of cold strain or independent measure to validate against, and as a consequence it cannot be confirmed that the new index really measures strain. In fact, what CSI is actually measuring is not known. Strain is defined in Webster's Third New International Dictionary as "to exert to the utmost" or "to put to great effort." However, during cold exposure, achieving lower T_core and/or T_sk may not always mean that the body is experiencing greater cold strain. For example, lower body temperatures may produce less strain on the body because elevated metabolic requirements are not required for maintaining higher body temperatures and the body therefore does not have to "strain" as much to maintain homeostasis. This is what is observed in cold habituation (16), and physiologists would not argue that cold habituation induces greater cold strain.

If the CSI has value and can be modified for applicability to exercise-cold stress, what changes should be incorporated? Perhaps the weighting factors for the components of CSI (T_core and T_skin) need to be reevaluated. Even in resting conditions, thermoregulatory responses are elicited early on, when T_core is near or above baseline values but T_sk is decreasing (7, 17). Frank et al. (6) suggest that the changes in T_sk may make it the dominant afferent input to the cold response, at least initially. Thus it is possible that once someone is exposed to cold, the percent contribution from peripheral and central sources is likely to be dynamic, not static, with peripheral and central inputs important during the early and latter stages, respectively, of cold exposure. Weightings of T_core and T_sk when calculating CSI (or any index of cold strain) may need to reflect these dynamic changes during cold exposure. Further evidence for more weighting given to peripheral temperature is suggested by the cutaneous contribution to thermal comfort. Frank et al. (6) found that T_core and T_sk equally contributed (unlike autonomic responses) to thermal comfort. Because the CSI is highly correlated (r² = 94%) to thermal sensation at rest (9), it may be that greater weighting needs to be given to T_sk throughout cold exposure for calculating CSI. However, one problem with this approach is that in the present study, thermal sensation was not related to the CSI during exercise-cold stress. Therefore, there may have to be different weightings given to T_core and T_sk during exercise-cold stress compared with resting cold exposures.

Another possible limitation of the CSI is its reliance on changes in T_core and T_sk from time 0. The CSI was developed using baseline values obtained in thermoneutral conditions. If thermoneutral baseline values are not available, then the CSI will be less reflective of the actual physiological stress/strain. For example, if a person begins a cold exposure at T_core and T_sk of 37 and 34°C, respectively, and these temperatures decrease over 90 min to 35.5 and 21°C, respectively, then the CSI will be calculated as 8.1. However, if someone begins a cold exposure at T_core and T_sk of 36.4 and 30°C, respectively, and these fall to 35.5 and 21°C, respectively, over 90 min, then the CSI will be calculated as 7.3, even though the physiological stress (same T_core and T_sk) is the same at the end of 90 min in both cases. The need for a true baseline value limits the applicability of CSI for field settings where comparisons between two environmental conditions may be confounded by different initial temperatures before each cold exposure, such as might occur over a day with multiple cold exposures separated by periods of rewarming (4). One solution to the problem caused by different baseline values is to calculate CSI using a standard "initial" temperature (i.e., T_core 0 = 37°C). Thus CSI will always have the same reference point, and comparisons between field conditions or experimental studies can be more easily compared.

We found that RPE during exercise-cold stress was moderately associated with HR, ventilation, and metabolic heat production (index of O₂), but not with CSI. These findings are in agreement with much of the exercise literature linking RPE to cardiopulmonary cues. However, we also demonstrated that RPE was correlated with _sk, a finding not observed by Toner et al. (13). One potential reason may be the different environmental conditions employed in each study. The present study was conducted with clothed, wet subjects in cold, windy conditions. In contrast, the experiments of Toner et al. (13) were conducted during cold water immersion, which quickly clamps T_sk at ~0.5-1°C above the water temperature and thus provides a fairly constant afferent stimulus during exposure.

We observed that thermal sensation was correlated with cardiopulmonary cues, but not with CSI. Our findings of relationships between thermal sensation and cardiopulmonary indexes agree with Kamon et al. (8), who reported associations between thermal sensation and HR and O₂ during exercise-heat stress. In the one study that fully examined thermal sensation during exercise-cold stress, Toner et al. (13) did not observe a relationship between thermal sensation and cardiopulmonary cues. Thermal cues were also related to thermal sensation, but unlike previous findings, only 6.3% of the variance in thermal sensation was accounted for by T_re. What accounts for the differences in the relationship between thermal sensation and T_core between the present study and Toner et al. (13)? Even though the exercise intensity was comparable and the changes in T_core (0.55 to 0.7°C) were similar between this study and Toner et al. (13), the rate of fall was much greater in Toner et al. (13), who observed a 0.7°C decline in T_core over a 45-min period. In the present study, T_core was elevated above the initial temperature for the 1st hour of cold exposure, and it was not until the 3rd and 4th hour of cold exposure that T_re decreased 0.4-0.5°C. Thus it may be that one of the physiological cues for thermal sensation may not be T_core per se, but the rate of temperature change. Also, the previous study (13) employed arm exercise, and this was included with leg exercise and arm and leg exercise in their correlational analysis. As suggested by Pandolf et al. (11), perceptual sensitivity for the processing of physiological information may be enhanced during small muscle exercise. Perhaps that is why thermal sensation is more highly correlated with T_core in Toner et al. (13), but not in the present investigation. The lack of a correlation between CSI and thermal sensation is different from that observed by Moran et al. (9). Our findings suggest that during exercise-cold stress, subjective ratings are not interchangeable with a physiologically based cold index.

In summary, CSI increased during prolonged exercise-cold exposure but did not discriminate between rested and fatigued conditions, in which T_core was lower and T_sk was higher after multiple days of exhaustive exercise. CSI does indicate greater cold strain across time, but its utility as a marker of strain between different treatments or studies is uncertain. Its utility during the initial stages of exercise-cold exposure is also limited because T_core is typically greater than that observed during baseline measurements. The CSI may need to account for this by weighting T_sk to a greater extent during the initial stages of cold exposure or by using a standard initial T_core (i.e, 37°C). Also, besides cardiopulmonary cues being related to ratings of perceived exertion, T_sk may also provide physiological input to perception of effort during prolonged exercise-cold air stress.