INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

OUTDOOR WORKERS are often required to perform manual dexterity tasks in cold environments. This frequently necessitates the removal of protective mitts in favor of working with gloved or even bare hands. Exposure of the hands to such conditions can result in the rapid cooling of the extremities, a loss of manual dexterity (4, 11, 20, 27, 31, 34), a loss of tactile sensitivity (24, 26, 29), and an increased risk of cold injury. If the gloves or mitts are not removed, a compromise must be made between the total insulation provided by the gloves or mitts and their effect on manual dexterity. Van Dilla et al. (36) have shown that the maximum glove thickness consistent with reasonable finger dexterity (i.e., 64-mm-thick fabric) is not enough to protect the hands from a cold environment (15°C). To counteract this problem, thin electrically heated gloves were developed over the years since World War II to allow individuals to maintain both a good deal of hand dexterity and hand comfort during extended periods of cold exposure.

The use of electrically heated gloves during cold exposure can present some difficulties. For example, electrically heated gloves may increase the chances of insidious hypothermia (28, 37). Van Someran et al. (37) state that the hands are normally colder than the rest of the body during exposure to a cold ambient environment, but during hand heating the hands are at the same temperature as the rest of the body. This results in a situation in which the human thermoregulatory system is not normally adjusted to respond effectively. Van Someran et al. suggest that when heat is applied solely to the hands the reflexes responsible for shivering may be lost because the hands play a particularly important role in the whole body thermoregulatory response to cold. As a result, hand heating may cause a nonsymptomatic decrease in core temperature. Other problems with electrically heated gloves include the following: the heating elements increase the stiffness of the gloves, thereby hampering dexterity; the robustness of the heating elements to repeated flexing in the cold can be a problem; and electrically heated gloves do not always have a uniform heat distribution.

Even with the improvements in heat distribution that have occurred over the years in the design of electrically heated gloves, "...the thermal lagging of the hand with different mass-to-surface areas (such as the fifth finger and the thumb web) makes maintenance of a uniform extremity temperature difficult" (13). Either some parts of the hand will be cold while others are comfortable, or "hot spots" will arise if those parts of the hand that cool at a faster rate are kept comfortable (13). Hot spots with auxiliary hand or foot heating can also arise when pressure enhances contact with the heating system. For example, Hickey et al. (17) found that when subjects were seated during cold exposure the electrically heated insoles worn by the subjects were thermally comfortable. However, the insoles were too hot when the subject applied pressure to the insoles by standing. Moreover, direct hand heating is an inefficient process because much of the added energy is lost to the environment through the thin insulation of the gloves.

In addition to direct auxiliary heating of the hands, past studies on hand comfort have also looked at the effects of applying local auxiliary heat to various parts of the body (1, 2, 5, 6, 12, 18, 21, 22, 25, 32, 35). However, most of these studies were done in a thermoneutral or cool environment (at temperatures of no less than 11°C). To our knowledge, only two studies have looked at the effects of auxiliary heat on hand comfort while subjects are barehanded during cold exposure. Auxiliary heating of the forearms while exposed to a 18°C ambient environment has been found to be unsuccessful in maintaining hand comfort (27). Rapaport et al. (30), on the other hand, found that bare hands could be kept comfortable (at a hand-skin temperature above 21°C) for a 1-h period during exposure to an ambient temperature of 34°C with the use of a full body, air-heated suit. However, the heating source for the suit was designed for laboratory use only; it did not have the mobility necessary for it to be practical in the field. Rapaport et al. also found that, in general, the hands were kept comfortable whenever there was a net body heat gain (i.e., body heat gain is greater than body heat loss), but, in some cases, the bare hands were kept comfortable for 1 h even when there was a slight negative net body heat balance.

Goldman (13) also conducted a similar experiment in which subjects were exposed to a 40°C environment while heat was provided to the torso region by using an air-heated vest. However, despite the fact that subjects wore Arctic mitts and a complete Arctic ensemble (4.3 clo), and were in a positive-heat-balance state (i.e., there was a net body heat gain), extremity comfort was not achieved. Goldman concluded that torso heating is an ineffective means of warming the extremities during cold exposure.

The above-mentioned two studies by Rapaport et al. (30) and Goldman (13) provide contrasting results as to whether extremity comfort is dependent on the body's state of heat balance. Therefore, the present study was done to provide more information about the effect of the body's state of heat balance on bare-hand comfort and body heat transfer during exposure to 15°C air. The use of active torso heating may result in a net body heat gain large enough to trigger increased circulation of blood to the extremities to dissipate the extra heat. In turn, warming of the extremities may be achieved. The relationship between extremity comfort and body heat balance obtained in the present study may provide an explanation for the contrasting results obtained by Rapaport et al. and Goldman.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Glossary

BTC	Body thermal comfort (scale 1-13)
Ctorso,h	Heated torso coefficient
Ctorso,uh	Unheated torso coefficient
CTre	Rectal temperature coefficient
CTre,h	Heated body rectal temperature coefficient
CTre,uh	Unheated body rectal temperature coefficient
Csk	Mean body skin temperature coefficient
	Mean heated body skin temperature coefficient
	Mean unheated body skin temperature coefficient
FTC	Finger thermal comfort (scale 1-13)
	Mean body heat flow (W)
	Mean heated body heat flow (W)
	Mean unheated body heat flow (W)
vest	Mean vest heater power (W)
	Mean middle-finger blood flow (PU)
	Rate of body heat storage (W)
body	Mean body temperature (°C)
body,h	Mean heated body temperature (°C)
body,uh	Mean unheated body temperature (°C)
fing	Mean middle-finger skin temperature (°C)
toe	Mean large-toe skin temperature (°C)
sk	Mean body skin temperature (°C)
sk,uh	Mean unheated body skin temperature (°C)
	Mean heated body skin temperature (°C)

Subjects

Experimental Protocol

THT. During THT, the subjects were asked to sit on a wooden bench in a cold chamber maintained at 15°C (wind ~2 km/h) for 3 h. Subjects entered the chamber barehanded; their mean finger temperature (_fing; average of middle-finger temperatures of left and right hand) was monitored until it reached 15°C (cooling phase). A _fing of 15°C was used as the criterion for the initiation of rewarming because it has been found that finger dexterity is decreased when _fing falls below 15°C (11). At this point, an attempt was made to indirectly rewarm the fingers by heating the torso to 42 ± 0.5°C with an electrically heated vest for the remainder of the test (rewarming phase). During the test, the subjects wore the first two layers of the new Canadian Forces (CF) Arctic clothing ensemble, which consisted of a fleece garment (jacket and pants) and an uninsulated jacket and pants. A thin pair of long cotton underwear were also worn under the fleece pants. Standard CF mukluks, woolen socks, and a balaclava were also worn. The clothing (excluding the long cotton underwear) provided 2.6 clo (0.4 m² · k¹ · W¹) of insulation.

CT. The CT was identical to the THT except for the rewarming phase. During the CT rewarming phase, CF Arctic mitts were donned after _fing reached 15°C. No heat was delivered to the torso via the electrically heated vest.

THT familiarization run. The subjects were familiarized with the equipment and procedures that were used during THT. In particular, the approximate torso heater power required to maintain the skin temperature underneath the heating vest at 42°C during THT was established. The subjects were fully dressed with the sensors, heaters, and clothing to be used during the test, then exposed to a 1-h exposure to 15°C air. The standard procedures of THT were followed.

Physiological Parameters Measured

2

The "heated region of torso coefficient" (C_torso,h) for each subject was calculated by dividing the vest area (0.366 m²) by the entire body surface area (m²). The insulated 0.366-m² vest area was used to represent the actively heated region of the torso instead of the 0.266-m² heater area because it was assumed that the skin temperature within 1-2 cm of the edge of the heaters was very close to the skin temperature directly underneath the heaters because of the reflective Mylar and Thinsulate insulation (covering an area of 0.366 m²) that covered the heaters. Once C_torso,h was calculated, the front and back "unheated region of torso coefficients" (C_torso,uh) for each subject were calculated in the following manner: C_torso,uh = (0.35  C_torso,h)/2, where 0.35 is the coefficient Hardy and DuBois (14) used to represent the torso area.

The _sk calculation during THT was heavily influenced by the skin thermistors under the heating vest; therefore, any significant differences observed in _sk between CT and THT may not exist if the torso is not included in the calculation. Because of this possibility, a mean unheated body skin temperature (_sk,uh) was also calculated. _sk,uh did not include any skin thermistor sites that were on the torso. For the same reason, because the calculation is heavily influenced by the HFTs under the heating vest, a mean unheated body heat loss () was calculated.

Mean heated body heat flux () in W · m² and mean heated body skin temperature (_sk,h) were measured by placing a HFT embedded with a thermistor [HA-13-18-10-P(C), Concept Engineering] under each of the heaters that make up the vest. The average skin temperatures under each of the five pairs of heaters (chest, abdomen, side, shoulder, and back) were calculated, and the five values were averaged to establish _sk,h. (in W · m²) was multiplied by the surface area of the vest (0.366 m²) to determine mean heated body heat flow (, in W). provides a measurement of the heat delivered to the torso area directly underneath the vest, whereas the mean vest heater power (_vest, in W; voltage × current of the 5 power supplies) is the power needed to deliver the desired . _vest is greater than because a portion of the heat from the heaters is lost to the ambient environment. Body heat storage (S; in kJ) was calculated during both CT and THT by using the thermometric method described by Burton (3). S was calculated as follows where 3.47 is the specific heat of the tissues, m_body is the body mass of the individual, and _body is the change in mean body temperature at time t from the initial mean body temperature at time 0. This was calculated as follows _body was calculated as the weighted sum of T_re and _sk. In the past, the weighting coefficients have been adjusted to reflect the ambient temperature. That is, the core is given more weighting in hotter environments and less weighting in cooler environments. For example, the _body of an individual exposed to cool conditions (nude, 23°C ambient environment) under steady-state conditions may be estimated by the following relationship

(3)

whereas the _body of an individual exposed to a thermoneutral environment (nude, 28-30°C environment) and a hot environment (nude, 33-36°C ambient environment) under steady-state conditions may be estimated by the following relationships (Ref. 15 and Refs. 9 and 15, respectively), respectively The above equations are appropriate to steady-state conditions. In the present experiment, the subjects were either cooling over time (during CT) or subjected to heat being delivered to the torso after a short period of cooling (during THT; i.e., the subjects were not likely in a steady state). Hence the use of any one equation was deemed inappropriate. Livingstone (23) found that, under non-steady-state conditions, the weighting coefficients for T_re and _sk should be adjusted not only for ambient temperature but also for time. In the present study, these adjustments were made after the data were collected.

During CT, the change in T_re coefficient (C_Tre) over time was determined every minute during the 3-h exposure by using a two-step approach. C_Tre was first determined every 10 min by using the following relationship where f and i are final and initial, respectively.

_body in the above equation was not derived from thermometry calculations but rather was directly measured during a 3-h whole body air calorimetry experiment (unpublished observations, personal communication, J. W. Snellen, 1996), in which T_re and _sk responses of the subjects were very similar to the T_re and _sk responses observed in the present experiment. Therefore, the cooling curve from Snellen's experiment was applied to the present data to determine C_Tre. When C_Tre was plotted as a function of time, an exponential decrease in C_Tre was observed. An exponential curve fit was calculated, and the equation of the curve was determined (cooling curve). The cooling curve was then applied to each set of T_re and _sk data for each subject for the calculation of _body. During CT, C_Tre decreased exponentially from 0.8 to 0.66. A C_Tre of 0.8 was chosen as the starting point because the _sk of the subjects (on entering the chamber) in the present study was similar to the _sk of nude subjects exposed to a thermoneutral environment (28°C ambient temperature; personal communication, J. W. Snellen, 1996). It is interesting to note that the exponential decrease in C_Tre was almost identical to the exponential decrease in _sk. Therefore, it appears as though the _sk response may be used to establish the rate of change in C_Tre.

During THT, the calculation of _body by using thermometry is complicated by the fact that some parts of the body are cooling while others (i.e., the area covered by the vest) are being heated. This problem was dealt with by treating the body as two separate regions (i.e., the heated and the unheated region of the body), so that where _body,uh is the weighted sum of _sk,uh and T_re, and _body,h is the weighted sum of _sk,h and T_re. C_Tre is different for the heated and unheated regions of the body. The changes in C_Tre over time for the two regions will now be examined in detail. For the unheated region of the body, the CT cooling curve described earlier was used to represent the change in C_Tre over time because _sk,uh during THT was not significantly different (P  0.05) from that during CT. The C_Tre for the unheated region of the body was defined as C_Tre,uh. For the heated region of the body, the CT cooling curve was applied up to the point at which the heating vest was turned on. At that point, the change in C_Tre over time followed a heating curve. The heating curve is based on the observation that during CT, the rate of decrease in _sk closely mimics the rate of decrease in C_Tre. Therefore, it was assumed that the rate of increase in C_Tre for the heated region of the body closely mimics the exponential increase in skin temperature under the heating vest (i.e., _sk,h). The C_Tre for the heated region of the body was defined as C_Tre,h.

Therefore, by using the above information, _body,uh and _body,h during THT were calculated in the following manner (by using 0.366 m² to represent the heated area of the torso) where C_Tre,uh, , and _sk,uh are the C_Tre, _sk coefficient, and _sk for the unheated region of the body, respectively; A_D is the Hardy and DuBois (14) body surface area (in m²); C_Tre,h, , and _sk,h are the C_Tre, _sk coefficient, and _sk for the heated region of the body, respectively.

After S was calculated by using the above procedures to calculate _body, the rate of body heat storage (, in W) during the exposure was calculated at every minute Measurements of _fing, _toe, T_re, _sk, _sk,uh, and were made five times per minute over the course of 3 h by using a data-acquisition system (model 3497A data-acquisition/control unit; Hewlett-Packard). An average value was printed out each minute.

Statistical Analyses

10

10

To reduce the data set to a more manageable size, average values over 5 min were calculated for each measured variable and were assigned to the middle time point of each 5-min interval (e.g., t₂, t₇, and t₁₂..., are average values for the intervals t₀ to t₄, t₅ to t₉, and t₁₀ to t₁₄,..., and so on).

A two-way ANOVA for repeated measures (SuperAnova, Abacus Concepts) with heating condition and time as the independent variables was used to compare conditions CT and THT for the dependent measurements _fing, _toe, T_re, _sk, _sk,uh, , , BTC, FTC, and . Results were considered statistically significant at P < 0.05 (by using the Greenhouse-Geisser adjustment for repeated measures).

If a significant treatment × time interaction existed in the two-way ANOVA for the repeated-measures test, a contrast test was performed every 5 min from t₁₀ to t₁₂₄ (using n = 6) and from t₁₀ to t₁₅₁ (using n = 5) to determine the time interval at which a significant difference existed between CT and THT.

A two-way ANOVA for the repeated-measures test was used to determine whether significant differences existed in left and right T_fing and T_toe responses from t₁₀ to t₁₂₄ (using n = 6) and from t₁₀ to t₁₅₁ (using n = 5) during CT and THT.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

The variables described below were measured during both CT and THT. The data are presented as means ± SE (n = 6) for each minute during the time period t₁₀ to t₁₅₁ (see Figs. 1-9, which identify where n is reduced from 6 to 5). is presented as 5-min averages during the time period t₃ to t₁₄₇. _fing, , _toe, _sk, BTC, FTC, and were all significantly higher (P < 0.05) during most of the THT rewarming phase from t₀ to t₁₅₁. was significantly lower (P < 0.05) during the THT rewarming phase. There were no significant differences (P  0.05) in T_re, _sk,uh, and between CT and THT during the rewarming phase. Tests for statistically significant differences between CT and THT were not done for data past t₁₅₁ because of the decreasing n value that resulted from time-shifting the data. There were no significant differences (P  0.05) between the CT and THT conditions for the 10-min period before t₀ for any of the measured variables. There were also no significant differences (P  0.05) in the _fing and _toe responses (for n = 6) between the left and right finger and/or toe temperatures during both CT and THT. Therefore, the left and right finger and/or toe temperatures were averaged for all six subjects.

In an examination of the figures presented in this paper, a discontinuity is often observed at t₁₂₄ during CT and THT. This is because of a sudden increase in the overall mean value of the measured variable that resulted when one subject ended the cold exposure early because of an uncomfortable _fing (i.e., the discontinuity is because of a decrease in the n value from 6 to 5). It was later observed that this subject had an abnormally high susceptibility to finger cooling. The following subsections provide a descriptive summary of each measured variable from t₁₀ to t₁₂₄ for CT and THT (during which time n = 6) even though the n = 5 data are included in the figures. The n = 5 data were included to illustrate that the general trend for all measured parameters remained the same up to t₁₅₁ for the five subjects who remained in the experiment.

_vest and

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Mean vest heat flow [; control test (CT) and torso-heating test (THT)] and mean vest heater power (vest; THT) from time t10 to t151 at 15°C. Values are means ± SE; n = 6 unless otherwise noted. * Time of 1st significant difference between THT and CT (P < 0.05).

_fing

10

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Mean middle-finger temperature from t10 to t151 at 15°C during CT and THT. Values are means ± SE; n = 6 unless otherwise noted. * Time of 1st significant difference between THT and CT. Mean middle-finger temperature was significantly higher (P < 0.05) during THT from * until t151.

View larger version (40K): [in this window] [in a new window]   Fig. 3.   Mean middle-finger blood flow from t10 to t151 at 15°C during CT and THT. Values are means ± SE; n = 6 unless otherwise noted. * Time of 1st significant difference between THT and CT. Mean middle-finger blood flow was significantly higher (P < 0.05) during THT from * until t151. Unit of measurement used to represent skin blood flow is perfusion unit (PU), which is a relative unit of blood flow.

_toe

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Mean large-toe temperature from t10 to t151 at 15°C during CT and THT. Values are means ± SE; n = 6 unless otherwise noted. * Time of 1st significant difference between THT and CT. Mean large-toe temperature was significantly higher (P < 0.05) during THT from * until t151.

T_re

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Rectal temperature from t10 to t151 at 15°C during CT and THT. Values are means ± SE; n = 6 unless otherwise noted. There was no significant difference (P  0.05) in rectal temperature between CT and THT from t0 to t151.

_sk and _sk,uh

View larger version (32K): [in this window] [in a new window]   Fig. 6.   Mean body skin temperature (sk) and mean unheated body skin temperature (sk,uh) from t10 to t151 at 15°C during CT and THT. Values are means ± SE; n = 6 unless otherwise noted. * Time of 1st significant difference between THT and CT for mean body skin temperature. Mean body skin temperature was significantly higher (P < 0.05) during THT from * until t151. There was no significant difference (P  0.05) in mean unheated body skin temperature between CT and THT from t0 to t151.

,

View larger version (34K): [in this window] [in a new window]   Fig. 7.   Mean body heat flow () and mean unheated body heat flow () for t10 to t151 at 15°C during CT and THT. Values are means ± SE; n = 6 unless otherwise noted. * Time of 1st significant difference between THT and CT. Mean body heat flow was significantly lower (P < 0.05) during THT from * until t151. There was no significant difference (P  0.05) in mean unheated body heat flow between CT and THT from t0 to t151.

BTC

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Body thermal comfort for t0 to t151 at 15°C during CT and THT. Values are means ± SE; n = 6 unless otherwise noted. * Time of 1st significant difference between THT and CT. Body thermal comfort was significantly higher (P < 0.05) during THT from * until t151.

Mean FTC

View larger version (27K): [in this window] [in a new window]   Fig. 9.   Body heat storage for t3 to t147 at 15°C during CT and THT. Values are means ± SE; n = 6 unless otherwise noted. Body heat storage was significantly higher (P < 0.05) during THT from t7 to t147. Body heat storage for CT and THT was significantly different (P < 0.05) from 0 W.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The results indicate _fing, , _toe, _sk, BTC, FTC, and were all significantly higher (P < 0.05) during the THT rewarming phase than during the CT rewarming phase. was significantly (P < 0.05) lower during the THT rewarming phase. In addition, there were no significant differences (P  0.05) in T_re, _sk,uh, and between the CT and THT rewarming phases. Overall, torso heating can keep the body, fingers, and toes comfortable for an extended period of time during exposure to 15°C air.

First, the fluctuations themselves most likely do not provide a true representation of the actual changes that occur within the body because it is unlikely that such fluctuations occur so frequently and to such a great extent (see Fig. 9). Second, when the thermometric method is used, it is generally assumed that C_Tre and of the subject do not change during the course of the experiment. In the present experiment, subjects wore a heating vest and the first two layers of the CF Arctic clothing ensemble. When the subjects entered the cold chamber, no torso heating was applied during THT for the first 10-29 min of the exposure. Therefore, the core-to-shell weighting could well have changed during the course of the experiment. Hence, in the present study, the weighting coefficients were adjusted over time by allowing the C_Tre to follow the _sk responses in an attempt to more properly represent the _body over time. Despite this improvement in the thermometric method, the true _body were not directly verified in the present experiment. Therefore, there is a chance that these adjustments are not representative of the true _body. Last, an inaccuracy in the estimation of during THT can be used to explain the finding that, despite a of 165 W during CT and a of 68 W during THT, differed by only 19 W between the two conditions (i.e., 67 W during CT vs. 48 W during THT).

The thermometric method used in the present study to calculate can be compared with a theoretical partitional calorimetry calculation of . For example, assuming that the metabolic rate was ~100 W during each condition, calculating this value into the heat balance equation would result in an of about 65 W during CT (not taking into account any evaporative losses from the body). This theoretical value of is very close to the value obtained by thermometry (67 W). However, during THT, the measured value of (48 W) is not close to the theoretical value of (+32 W). Because the calculated by using thermometry is most likely a crude estimation of the actual , further research, using better methods of determining body heat balance (e.g., using partitional calorimetry or direct calorimetry), is needed before any conclusions relating extremity comfort to body heat balance can be made.

The two primary measured variables used in the calculation of are _sk and T_re. Hence the significantly (P < 0.05) higher observed during THT, compared with CT, can be attributable mainly to the significantly (P < 0.05) higher _sk maintained during THT because there was no significant (P  0.05) difference in the T_re response between CT and THT.

It is worthwhile noting that Rapaport et al. (30) used partitional calorimetry to calculate , whereas Goldman (13) used thermometry to calculate . Rapaport et al. found that, in general, extremity comfort was maintained for subjects with an   0. Goldman, on the other hand, found that extremity comfort could not be achieved despite a net body heat gain of 84 W (based on a 2-m² body surface area). If extremity comfort is directly related to a body heat balance  0, this would suggest that Goldman's thermometry calculations may have overestimated whereas, in the present study, may have been underestimated. It should also be noted that Rapaport et al. (30) found that, in one experiment, bare-hand comfort could be maintained (average hand-skin temperature between 25 and 29°C) for 1 h during exposure to 34°C air, even when there was a of 23 W. The data of Rapaport et al. support our finding that the fingers and toes could be kept comfortable during torso heating even when is negative.

T_re

On average, there was no noticeable increase in T_re just before an increase in peripheral blood flow or peripheral skin temperature as was observed in some past studies (33, 35, 39). Instead, T_re remained stable, on average, during the entire cold exposure. Perhaps an increase in T_re was not observed because, during the cold exposure, the rate of heat dissipation from the hands and the rest of the area not covered by the vest was greater than the heat dissipated from the hands of subjects exposed to much warmer temperatures in other studies. The larger temperature gradient between the air and the hand surface temperature of the subject, for example, may have evoked a greater rate of heat loss. As a result, a potential increase in core temperature was quickly offset by a large heat loss from the hands, and, therefore, no increase in T_re was observed. It is also possible that an increase in T_re was not observed because it is not a rapidly responding measurement site. Perhaps another core temperature site, such as the esophagus, would have shown an increase.

Finger and Toe Temperature