Shivering onset, metabolic response, and convective heat transfer during cold air exposure P. TIKUISIS, D. G. BELL, AND I. JACOBS Defence and Civil Institute of Environmental Medicine, North York, Ontario M3M 3B9, Canada TIKUISIS, P., D. G. BELL, AND I. JACOBS. Shivering onset, metabolic response, and convective heat transfer during cold air exposure. J. Appl. Physiol. 70(5): 1996-2002,1991.-The onset
and intensity of shivering of various musclesduring cold air exposure are quantified and related to increasesin metabolic rate and convective heat loss.Thirteen male subjectsresting in a supineposition and wearing only shorts were exposedto 10°C air (42% relative humidity and x0.4 m/s airflow) for 2 h. Measurements included surface electromyogram recordings at six muscle sites representing the trunk and limb regions of one sideof the body, temperatures and heat fluxes at the samecontralateral sites, and metabolic rate. The subjectswere grouped accordingto lean (LEAN, n = 6) and averagebody fat (NORM, n = 7) content. While the rectal temperatures fluctuated slightly but not significantly during exposure,the skin temperatures decreasedgreatly, more at the limb sitesthan at the trunk sites.Muscles of the trunk region beganto shiver soonerand at a higher intensity than those of the limbs. The intensity of shivering and its increaseover time of exposurewere consistent with the increase in the convective heat transfer coefficient calculated from skin temperatures and heat fluxes. Both the onset of shivering and the magnitude of the increasein metabolic rate due to shivering were higher for the LEAN group than for the NORM group. A regressionanalysisindicates that, for a given decreasein mean skin temperature, the increasein metabolic rate due to shivering is attenuated by the squareroot of percent body fat. Thus the LEAN group shivered at higher intensity, resulting in higher increasesin metabolic heat production and convective heat lossduring cold air exposure than did the NORM group. electromyogram; hypothermia; temperature regulation; body fatness
defense mechanism of nonexercising humans exposed to cold. It is the primary means by which humans can increase metabolic heat production and thereby reduce the rate of decrease of body temperature (3, 6, 8, 10, 21). Indeed, shivering and heat production are highly correlated (7). Studies in the past have examined various aspects of shivering, including the neurophysiology of shivering (11), the electromyographic (EMG) activity during shivering (12), and the correlation of shivering activity among different muscles (1) and to other physiological indexes such as respiration (7, 9) and skin temperature (4); a thorough review is given by Sowood (16). Yet, since 1957 when Spurr et al. (17) reported that in cold-exposed men shivering began in the trunk region and spread to the limbs, there has been little further demonstration with regard to the onset of shivering of the individual muscles. This lack of SHIVERING
IS AN IMPORTANT
information is reflected in the development of many mathematical models that predict human thermoregulatory response to cold where it is usually assumed that the muscles begin shivering simultaneously, although certain recent models (20, 24) assume an exponential time of onset of limb shivering. Other important aspects of the thermoregulatory response to cold that have not been thoroughly investigated are the relationship of the increases in metabolic rate to body fatness and the rate of heat transfer to the environment due to shivering. Buskirk et al. (3) observed that the metabolic response to cold for average-fat and obese individuals was significantly and inversely related to body fatness. A more recent study (19) found that the increase in the metabolic rate due to shivering was determined by changes in the mean skin and deep body temperatures, but this increase was attenuated by body fatness. Data were based on lean and average-fat individuals immersed in cold water under both nude and clothed conditions. No examination of this specific attenuation has been reported for exposure to cold air. Heat transfer from a body surface to the environment is dependent, in part, on the relative motion between the surface and the environment. With increased shivering, it can be expected that this heat transfer is enhanced; yet no study has examined this effect. The present study investigated these aspects in addition to the time course of the onset of shivering of various muscles with the use of human male subjects exposed to cold air. METHODS
Subjects. Experiments were conducted with 13 male volunteer subjects, 30-42 yr old. In accordance with institutional Human Ethics Committee Guidelines for experimentation, each subject was fully informed of the experimental procedure before giving his written consent. In addition, subjects were medically screened and verified to have normal cardiovascular function. Standard height and weight measurements were taken, and body fat (BF) was determined by hydrostatic weighing (2) in each subject (Table 1). Procedures. Each experiment consisted of exposing a single subject to 10°C air at ~42% relative humidity for 2 h. The experimental chamber was well vented, and the airflow around the subject varied in direction and flow but was very mild (co.4 m/s). All subjects repeated the exposure 1 wk later. Before each exposure, the subject was fitted with EMG and electrocardiographic electrodes, heat flux (HF) transducers for both HF and tem-
1996
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
ONSET
OF SHIVERING
1. Physical characteristics of subjects
TABLE
LEAN
NORM
6 173.0-183.5 (179&5.4) 60.5-84.3 (74.6k8.0) 1.72-2.05 (1.93+0.12) 31-41 (36.3k4.0) 3.7-11.1 (8.5k2.8) 54.0-75.0 (68.3-t7.9)
n
Height, cm Weight, kg Surface area, m2 Age, yr %BF LBM, kg
7 173.0-187.0 (18O.Oe6.3) 73.5-102.5 (83.3k10.9) 1.87-2.22 (2.03rtO.14) 30-42 (35.023.9) 15.2-22.4 (17.9k2.5) 61.0-87.0 (68.4rt9.1)
P
NS NS NS NS 0.001 NS
Values are ranges with means zt SD in parentheses; n, no. of subjs. LEAN subjs were ~15% body fat (BF). LBM, lean body mass.
perature measurements, and a rectal thermistor. Wearing only shorts, the subject then took the standard supine position (with arms and legs spread apart) on a rope mesh cot and was covered with blankets to maintain thermal comfort. During the next 0.5 h, the subject rested at a room temperature of 21-22°C. During -15 to -5 min of this period (PRE; Fig. l), skin (Tsk) and rectal temperatures (T,,), surface HF, surface resting EMG activity, and metabolic rate (MR) were measured. Then the subject was immediately wheeled into the cold room, his blankets were removed, and the 2-h exposure began. The subject was removed if his T,, dropped below 35.5”C, 2 h elapsed, or he asked to be removed because of discomfort. During the cold exposure, Tsk, T,, HF, and EMG activity were recorded continuously. MRs were recorded every minute during the following periods: 5-30,40-60, 70-90, and loo-120 min. The subject was asked not to move voluntarily during the experiment except during 1-min stretch breaks allowed every 15 min (Fig. 1). At the end of the exposure, the subject was removed from the cold room and rewarmed in a hot tub before being released. Measurements. EMG activity was monitored with pediatric electrodes (Medi Trace, Graphic Controls Canada, Gananoque, Ontario, Canada) placed 3 cm apart, center to center, on the midpoints of six muscles chosen to represent the trunk and limbs of the body, which are consid-
PRE
10°C AIR EXPOSURE Stretch break MR Tsk, Tre, HF. EMG
-15
0
15
30
45
60
75
90 105
120
Time (min) 1. Schedule of stretch breaks and measurements before and during exposure to 10°C air. MR, metabolic rate; Tgk, mean skin temperature; T,, rectal temperature; HF, heat flux. FIG.
1997
ered important sources of heat generation during shivering in models of thermoregulation (18, 20, 24). These muscles were the pectoralis major (PE), rectus abdominis (AB), biceps femoris, brachioradialis (BR), rectus femoris (FE), and gastrocnemius (GA). EMG signals were amplified X1,000 (bandwidth 8-1,000 Hz; MBS884, Moroz Biosciences System, Hamilton, Ontario, Canada) and recorded on FM tape (bandwidth O-2,500 Hz; 3968A, Hewlett-Packard, San Diego, CA). The EMG signals were played back through amplifiers and recorded (Universal Gould and Gould TA2000 recorder, respectively, Cleveland, OH) for analysis. HF transducers with integral linear thermistors (Concept Engineering, Old Saybrook, CT) were placed on the contralateral sites of the same six muscle groups monitored with EMG. In addition, seven more HF transducers were placed on the body so that a mean weighted T, (Tsk) based on a 12-point system (23) could be obtained. T, was measured with a thermistor probe inserted 15 cm into the rectum. All HFs and temperatures were recorded every 15 s with a computerized data acquisition system (23).
MR was determined from the 0, decrement and CO, increment of expired gas plus the volume ventilation measurements. During the preexposure resting period, expiratory gas was collected in a wet spirometer (129 liter Gasometer, Collins, Braintree, MA), and during the cold exposure pulmonary ventilation was measured continuously (Alpha ventilation VMM 100 modual; Interface, Irvine, CA). All collected gases were analyzed by electrochemical analyzers (Ametek Thermox Instruments, Pittsburgh, PA). A single MR was determined for the preexposure period based on a lo-min collection period, and 1-min values were obtained during cold exposure as noted above (Fig. 1). Treatment of data. The time of onset of shivering and the time of maximum shivering of each muscle were visually and independently determined by two investigators from tracings obtained from the recorded EMGs. Onset was chosen as the point where the EMG amplitude first exceeded the preexposure resting values (Fig. 2). The determination of onset by the two investigators differed by 15% body fat. Values are means k tween LEAN and NORM. FIG.
AB
FE
GA
Group of PE, BI, BR, AB, FE, and GA LEAN, ~15% body fat; NORM, SE. *Significant differences be-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
ONSET
OF
SHIVERING
37.2
36
v v -
G LIc! 37.0
PELEAN PENORM GALEAN GANORM
36.8
36.6
36.4 -30-M
0
15
30
Time
45
60
75
90
105120
-30-15
0
15
(min)
45
Time
4. Rectal temperature (T,, mean t SE) before and during exposure to 10°C air. Values before exposure decreased significantly, whereas values during exposure, averaged over 6-min intervals (see text), did not change significantly and were not significantly different between LEAN and NORM. FIG.
by upper and lower body regions and from trunk to extremities) of the two groups. In five of the seven NORM subjects, EMG activity was not detected in one or both of the extreme peripheral muscles, i.e., BR and GA. In these cases, a conservative estimate of onset time of 120 min was assigned for display purposes. Shivering appears to begin in the trunk and upper leg regions of the LEAN
g"UPand'n'h'uPPe"~nk'f'heNORMgr'uP~
30
andit
propagates toward the extremities, in general agreement with the findings of Spurr et al. (17). Figure 3 also indicates that the onset of shivering of the leg muscles for the LEAN group is significantly sooner than for the NORM group. No difference was found for the mean occurrence of maximum shivering, which was between 75 and 90 min for all muscles of both groups except the BR and GA of the NORM group, where EMG activity was not detected in some subjects. Body temperatures and HFs. Before exposure, T, is seen in Fig. 4 to decrease slightly but significantly for both BF groups. However, the change during exposure was not significant, where T, increased slightly during the first half and then decreased slightly during the latter half, but not below its starting value. Rapid and significant decreases in T,, and increases in HF were measured on initial exposure to cold at all six muscle sites. In all limb muscle sites, the changes in T,, and HF decayed exponentially, whereas the magnitude of the change was greater for T,, than for HF. This is in contrast to the PE and AB muscle sites, which showed initial changes similar to the other muscles but slight reversals about 1 h later. To show these trends, mean temperatures and HFs are shown in Figs. 5 and 6 for only the PE and GA muscles, which represent the most active and least active muscles, respectively. The small increase in T,, of the PE muscle site before exposure (between -15 and -5 min) was significant for both BF groups. No significant changes in T,, were found for the other muscle sites before exposure, nor were changes in HF significant before exposure for all muscle sites. The site of the most active shivering muscle, PE, showed the smallest changes in temperature, whereas
60
75
90
105120
(min)
5. Skin temperature (Tsk, mean t sites before and during exposure to 10°C increased slightly but significantly for PE exposure, averaged over 6-min intervals cantly over time and were significantlydifferent NORM for PE but not for GA muscle site. FIG.
SE) at PE and GA muscle air. Values before exposure muscle site. Values during (see text), changed signifibetween LEAN and
that of the least active muscle, GA, showed the opposite. This is not surprising because increased shivering activity generates heat to offset the cooling effect of the environment, thus maintaining a warmer skin temperature. Also, changes in T,, and HF reversed slightly for the PE muscle site, whereas they continued steadily for the GA muscle site. Another difference between the two sites is a
separation of t-e changes in T,, and HF between the
LEAN and NORM groups for the PE muscle but not for the GA muscle. Although skin temperatures and HFs have similar starting values for both groups of subjects during cold exposure, they are maintained significantly higher for the LEAN group at the PE muscle site but are indistinguishable at the GA muscle site. Equation 1 was applied to the mean data shown in Figs. 5 and 6 and its prediction of h is shown in Fig. 7 for both the PE and GA muscle sites. Although values are higher for the LEAN group than the NORM group at the PE muscle site, these differences are not significant. The interesting feature for both muscle sites is the significant 200
150
NI E gu 100 v -
% 50
0 -30-15
0
15
PEL PENORM GALEAN
30
Time
45
60
75
90
105120
(min)
6. Heat flux (HF, mean t SE) at PE and GA muscle sites before and during exposure to 10°C air. Values during exposure, averaged over 6-min intervals (see text), changed significantly over time and were significantly different between LEAN and NORM for PE but not for GA muscle site. FIG.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
2000
ONSET I -
11 7”
OF SHIVERING
PELEAN PENORM GALEAN GANofw
. . .
0.0 24
25
26
27
28
29
30
31
Tsk ( “C) 9. Relationship of mean skin temperature (Y&J to increase in metabolic rate (AMR) during exposure to 10°C air. Values are normalized to lean body mass (LBM). Regression lines AMR/LB,M = 7.26 0.23T,, (R2 = 0.65) and 5.17 - O.l7T,, (R2 = 0.62) for LEAN and NORM, respectively, are significantly different (P < 0.001). FIG.
6-
. 5 0
.~I..l~.l.‘~.‘~ll~~~~~~~ 15 30 45
60
75
90
105
120
Time (min) FIG. 7. Heat transfer coefficient (h, mean t SE) at PE and GA muscle sites during exposure to 10°C air. Values increased significantly over time but were not significantly different between LEAN and NORM for either muscle site.
increase in h with increasing exposure time. Since the ambient air conditions (temperature, humidity, and circulation) did not change over the exposure period, the increase in h must be attributed to a decrease in the boundary air layers at the muscle sites caused by an increase in movement of the muscle site through shivering. This is consistent with the increase in shivering intensity observed from the EMG recordings. iMetaboLicrate. The MR rose rapidly for both groups on initial exposure to the cold and then increased slowly but significantly over time, as shown in Fig. 8. This pattern is consistent with the findings of Girling (6) for subjects exposed to 10°C air and Glickman et al. (7) for clothed subjects exposed to -3OOC air. The high values of MR between 75 and 90 min are consistent with the time of maximum shivering noted earlier. The average peak 200
1
180 160
value of about 156 W for MR reported by Girling (6), which occurred at the end of a 90-min exposure to 10°C air, is very close to that found for the NORM group in the present study. However, MR of the LEAN group in the present study was significantly higher than that of the NORM group. The decrease in MR of the LEAN group after 90 min is not significant because two subjects were removed at this time and mean MR of those remaining was lower than that indicated at 86 min. From previous work (18, 19) it can be expected that AMR depends on decreases in both the core and skin temperatures. Although the change in Tsk is large during cold exposure in the present study, the change in T, is not (mean maximal changes in T,, were +0.15 and +O.l6”C for the LEAN and NORM groups, respectively; Fig. 4). Therefore, T, cannot be considered to contribute to the shivering stimulus (the small decrease in T, before exposure is discussed further below). Attention is instead focused on the relationship between AMR and changes in ‘I‘,,. First, the AMR values during the cold exposure were divided by the lean body mass [LBM = body mass X (100 - %BF)/lOO] of the groups to normalize for muscle content, although the difference in LBM between the two groups is not significant (Table 1). Figure 9 shows a tendency for the LEAN group to respond to a similar cold stress with a higher AMR (based on the same decrease in T,J than the NORM group. A simple regression of these results (Fig. 9) indicates a significant difference (P < 0.001) between the two groups of subjects. DISCUSSION
80 60
-15
-
LEAN
-
Nofwl
1 0
15
30
45
Time
60
75
90
105 120
(min)
8. Metabolic rate (MR, mean & SE) before and during exposure to 10°C air. Values during exposure increased significantly over time, and LEAN values were significantly higher than NORM values. FIG.
In general, this study reaffirms the finding of Spurr et al. (17) that shivering begins in the trunk region and spreads to the limbs. Spurr et al. reported that the average time to the first indication of shivering (in either the neck, pectoral, or abdominal region) was 6.4 min. In the present study, the average time to shivering of both BF groups was the same for the pectoral region, i.e., 4.1 min, but differed greatly for the abdominal region (5.4 and 21.4 min for the LEAN and NORM groups, respectively). This order in onset times, in fact, was observed for all
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
ONSET
OF
muscle groups except PE, suggesting that the LEAN subjects respond sooner to cold exposure than do the NORM subjects. Another interesting observation is that the time of onset of shivering of FE coincides approximately to the onset time of AB, which may be due to the proximity of FE to the trunk region and/or related to muscle mass (i.e., onset may be sooner with increased muscle mass). The observation of increasing h values with time of exposure (Fig. 7) can be explained by a decrease in the thickness of the air boundary layer that occurs with increasing relative air velocity (15). This can occur with an increase in air velocity and/or movement of the surface itself. The fact that no increase in room airflow was detected during the exposures suggests that increased movement of the skin surface by way of increased shivering leads to the increase in h over time. In addition, the h values of the PE muscle site are higher than those for the GA muscle site, in contrast to the theoretical prediction of a decreasing h, with decreasing surface curvature. Although it is possible that the trunk was in contact with a higher air velocity than were the limbs during the cold exposure, it is quite unlikely that the air velocity was almost an order of magnitude higher, as required by theory (15) to explain the different h values. Hence, the higher h values observed for the PE muscle site are probably due to a much higher shivering intensity, which is consistent with the visual observation discussed earlier. This finding suggests that predictions of heat loss that are based on fixed values of h derived from nonshivering conditions underpredict the actual heat loss during shivering. Despite attempts to ensure a steady-state thermal condition at the start of the cold exposure, a slight but significant decrease in T,, (
and intensity of shivering of various musclesduring cold air exposure are quantified and related to increasesin metabolic rate and convective heat loss.Thirteen male subjectsresting in a supineposition and wearing only shorts were exposedto 10°C air (42% relative humidity and x0.4 m/s airflow) for 2 h. Measurements included surface electromyogram recordings at six muscle sites representing the trunk and limb regions of one sideof the body, temperatures and heat fluxes at the samecontralateral sites, and metabolic rate. The subjectswere grouped accordingto lean (LEAN, n = 6) and averagebody fat (NORM, n = 7) content. While the rectal temperatures fluctuated slightly but not significantly during exposure,the skin temperatures decreasedgreatly, more at the limb sitesthan at the trunk sites.Muscles of the trunk region beganto shiver soonerand at a higher intensity than those of the limbs. The intensity of shivering and its increaseover time of exposurewere consistent with the increase in the convective heat transfer coefficient calculated from skin temperatures and heat fluxes. Both the onset of shivering and the magnitude of the increasein metabolic rate due to shivering were higher for the LEAN group than for the NORM group. A regressionanalysisindicates that, for a given decreasein mean skin temperature, the increasein metabolic rate due to shivering is attenuated by the squareroot of percent body fat. Thus the LEAN group shivered at higher intensity, resulting in higher increasesin metabolic heat production and convective heat lossduring cold air exposure than did the NORM group. electromyogram; hypothermia; temperature regulation; body fatness
defense mechanism of nonexercising humans exposed to cold. It is the primary means by which humans can increase metabolic heat production and thereby reduce the rate of decrease of body temperature (3, 6, 8, 10, 21). Indeed, shivering and heat production are highly correlated (7). Studies in the past have examined various aspects of shivering, including the neurophysiology of shivering (11), the electromyographic (EMG) activity during shivering (12), and the correlation of shivering activity among different muscles (1) and to other physiological indexes such as respiration (7, 9) and skin temperature (4); a thorough review is given by Sowood (16). Yet, since 1957 when Spurr et al. (17) reported that in cold-exposed men shivering began in the trunk region and spread to the limbs, there has been little further demonstration with regard to the onset of shivering of the individual muscles. This lack of SHIVERING
IS AN IMPORTANT
information is reflected in the development of many mathematical models that predict human thermoregulatory response to cold where it is usually assumed that the muscles begin shivering simultaneously, although certain recent models (20, 24) assume an exponential time of onset of limb shivering. Other important aspects of the thermoregulatory response to cold that have not been thoroughly investigated are the relationship of the increases in metabolic rate to body fatness and the rate of heat transfer to the environment due to shivering. Buskirk et al. (3) observed that the metabolic response to cold for average-fat and obese individuals was significantly and inversely related to body fatness. A more recent study (19) found that the increase in the metabolic rate due to shivering was determined by changes in the mean skin and deep body temperatures, but this increase was attenuated by body fatness. Data were based on lean and average-fat individuals immersed in cold water under both nude and clothed conditions. No examination of this specific attenuation has been reported for exposure to cold air. Heat transfer from a body surface to the environment is dependent, in part, on the relative motion between the surface and the environment. With increased shivering, it can be expected that this heat transfer is enhanced; yet no study has examined this effect. The present study investigated these aspects in addition to the time course of the onset of shivering of various muscles with the use of human male subjects exposed to cold air. METHODS
Subjects. Experiments were conducted with 13 male volunteer subjects, 30-42 yr old. In accordance with institutional Human Ethics Committee Guidelines for experimentation, each subject was fully informed of the experimental procedure before giving his written consent. In addition, subjects were medically screened and verified to have normal cardiovascular function. Standard height and weight measurements were taken, and body fat (BF) was determined by hydrostatic weighing (2) in each subject (Table 1). Procedures. Each experiment consisted of exposing a single subject to 10°C air at ~42% relative humidity for 2 h. The experimental chamber was well vented, and the airflow around the subject varied in direction and flow but was very mild (co.4 m/s). All subjects repeated the exposure 1 wk later. Before each exposure, the subject was fitted with EMG and electrocardiographic electrodes, heat flux (HF) transducers for both HF and tem-
1996
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
ONSET
OF SHIVERING
1. Physical characteristics of subjects
TABLE
LEAN
NORM
6 173.0-183.5 (179&5.4) 60.5-84.3 (74.6k8.0) 1.72-2.05 (1.93+0.12) 31-41 (36.3k4.0) 3.7-11.1 (8.5k2.8) 54.0-75.0 (68.3-t7.9)
n
Height, cm Weight, kg Surface area, m2 Age, yr %BF LBM, kg
7 173.0-187.0 (18O.Oe6.3) 73.5-102.5 (83.3k10.9) 1.87-2.22 (2.03rtO.14) 30-42 (35.023.9) 15.2-22.4 (17.9k2.5) 61.0-87.0 (68.4rt9.1)
P
NS NS NS NS 0.001 NS
Values are ranges with means zt SD in parentheses; n, no. of subjs. LEAN subjs were ~15% body fat (BF). LBM, lean body mass.
perature measurements, and a rectal thermistor. Wearing only shorts, the subject then took the standard supine position (with arms and legs spread apart) on a rope mesh cot and was covered with blankets to maintain thermal comfort. During the next 0.5 h, the subject rested at a room temperature of 21-22°C. During -15 to -5 min of this period (PRE; Fig. l), skin (Tsk) and rectal temperatures (T,,), surface HF, surface resting EMG activity, and metabolic rate (MR) were measured. Then the subject was immediately wheeled into the cold room, his blankets were removed, and the 2-h exposure began. The subject was removed if his T,, dropped below 35.5”C, 2 h elapsed, or he asked to be removed because of discomfort. During the cold exposure, Tsk, T,, HF, and EMG activity were recorded continuously. MRs were recorded every minute during the following periods: 5-30,40-60, 70-90, and loo-120 min. The subject was asked not to move voluntarily during the experiment except during 1-min stretch breaks allowed every 15 min (Fig. 1). At the end of the exposure, the subject was removed from the cold room and rewarmed in a hot tub before being released. Measurements. EMG activity was monitored with pediatric electrodes (Medi Trace, Graphic Controls Canada, Gananoque, Ontario, Canada) placed 3 cm apart, center to center, on the midpoints of six muscles chosen to represent the trunk and limbs of the body, which are consid-
PRE
10°C AIR EXPOSURE Stretch break MR Tsk, Tre, HF. EMG
-15
0
15
30
45
60
75
90 105
120
Time (min) 1. Schedule of stretch breaks and measurements before and during exposure to 10°C air. MR, metabolic rate; Tgk, mean skin temperature; T,, rectal temperature; HF, heat flux. FIG.
1997
ered important sources of heat generation during shivering in models of thermoregulation (18, 20, 24). These muscles were the pectoralis major (PE), rectus abdominis (AB), biceps femoris, brachioradialis (BR), rectus femoris (FE), and gastrocnemius (GA). EMG signals were amplified X1,000 (bandwidth 8-1,000 Hz; MBS884, Moroz Biosciences System, Hamilton, Ontario, Canada) and recorded on FM tape (bandwidth O-2,500 Hz; 3968A, Hewlett-Packard, San Diego, CA). The EMG signals were played back through amplifiers and recorded (Universal Gould and Gould TA2000 recorder, respectively, Cleveland, OH) for analysis. HF transducers with integral linear thermistors (Concept Engineering, Old Saybrook, CT) were placed on the contralateral sites of the same six muscle groups monitored with EMG. In addition, seven more HF transducers were placed on the body so that a mean weighted T, (Tsk) based on a 12-point system (23) could be obtained. T, was measured with a thermistor probe inserted 15 cm into the rectum. All HFs and temperatures were recorded every 15 s with a computerized data acquisition system (23).
MR was determined from the 0, decrement and CO, increment of expired gas plus the volume ventilation measurements. During the preexposure resting period, expiratory gas was collected in a wet spirometer (129 liter Gasometer, Collins, Braintree, MA), and during the cold exposure pulmonary ventilation was measured continuously (Alpha ventilation VMM 100 modual; Interface, Irvine, CA). All collected gases were analyzed by electrochemical analyzers (Ametek Thermox Instruments, Pittsburgh, PA). A single MR was determined for the preexposure period based on a lo-min collection period, and 1-min values were obtained during cold exposure as noted above (Fig. 1). Treatment of data. The time of onset of shivering and the time of maximum shivering of each muscle were visually and independently determined by two investigators from tracings obtained from the recorded EMGs. Onset was chosen as the point where the EMG amplitude first exceeded the preexposure resting values (Fig. 2). The determination of onset by the two investigators differed by 15% body fat. Values are means k tween LEAN and NORM. FIG.
AB
FE
GA
Group of PE, BI, BR, AB, FE, and GA LEAN, ~15% body fat; NORM, SE. *Significant differences be-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
ONSET
OF
SHIVERING
37.2
36
v v -
G LIc! 37.0
PELEAN PENORM GALEAN GANORM
36.8
36.6
36.4 -30-M
0
15
30
Time
45
60
75
90
105120
-30-15
0
15
(min)
45
Time
4. Rectal temperature (T,, mean t SE) before and during exposure to 10°C air. Values before exposure decreased significantly, whereas values during exposure, averaged over 6-min intervals (see text), did not change significantly and were not significantly different between LEAN and NORM. FIG.
by upper and lower body regions and from trunk to extremities) of the two groups. In five of the seven NORM subjects, EMG activity was not detected in one or both of the extreme peripheral muscles, i.e., BR and GA. In these cases, a conservative estimate of onset time of 120 min was assigned for display purposes. Shivering appears to begin in the trunk and upper leg regions of the LEAN
g"UPand'n'h'uPPe"~nk'f'heNORMgr'uP~
30
andit
propagates toward the extremities, in general agreement with the findings of Spurr et al. (17). Figure 3 also indicates that the onset of shivering of the leg muscles for the LEAN group is significantly sooner than for the NORM group. No difference was found for the mean occurrence of maximum shivering, which was between 75 and 90 min for all muscles of both groups except the BR and GA of the NORM group, where EMG activity was not detected in some subjects. Body temperatures and HFs. Before exposure, T, is seen in Fig. 4 to decrease slightly but significantly for both BF groups. However, the change during exposure was not significant, where T, increased slightly during the first half and then decreased slightly during the latter half, but not below its starting value. Rapid and significant decreases in T,, and increases in HF were measured on initial exposure to cold at all six muscle sites. In all limb muscle sites, the changes in T,, and HF decayed exponentially, whereas the magnitude of the change was greater for T,, than for HF. This is in contrast to the PE and AB muscle sites, which showed initial changes similar to the other muscles but slight reversals about 1 h later. To show these trends, mean temperatures and HFs are shown in Figs. 5 and 6 for only the PE and GA muscles, which represent the most active and least active muscles, respectively. The small increase in T,, of the PE muscle site before exposure (between -15 and -5 min) was significant for both BF groups. No significant changes in T,, were found for the other muscle sites before exposure, nor were changes in HF significant before exposure for all muscle sites. The site of the most active shivering muscle, PE, showed the smallest changes in temperature, whereas
60
75
90
105120
(min)
5. Skin temperature (Tsk, mean t sites before and during exposure to 10°C increased slightly but significantly for PE exposure, averaged over 6-min intervals cantly over time and were significantlydifferent NORM for PE but not for GA muscle site. FIG.
SE) at PE and GA muscle air. Values before exposure muscle site. Values during (see text), changed signifibetween LEAN and
that of the least active muscle, GA, showed the opposite. This is not surprising because increased shivering activity generates heat to offset the cooling effect of the environment, thus maintaining a warmer skin temperature. Also, changes in T,, and HF reversed slightly for the PE muscle site, whereas they continued steadily for the GA muscle site. Another difference between the two sites is a
separation of t-e changes in T,, and HF between the
LEAN and NORM groups for the PE muscle but not for the GA muscle. Although skin temperatures and HFs have similar starting values for both groups of subjects during cold exposure, they are maintained significantly higher for the LEAN group at the PE muscle site but are indistinguishable at the GA muscle site. Equation 1 was applied to the mean data shown in Figs. 5 and 6 and its prediction of h is shown in Fig. 7 for both the PE and GA muscle sites. Although values are higher for the LEAN group than the NORM group at the PE muscle site, these differences are not significant. The interesting feature for both muscle sites is the significant 200
150
NI E gu 100 v -
% 50
0 -30-15
0
15
PEL PENORM GALEAN
30
Time
45
60
75
90
105120
(min)
6. Heat flux (HF, mean t SE) at PE and GA muscle sites before and during exposure to 10°C air. Values during exposure, averaged over 6-min intervals (see text), changed significantly over time and were significantly different between LEAN and NORM for PE but not for GA muscle site. FIG.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
2000
ONSET I -
11 7”
OF SHIVERING
PELEAN PENORM GALEAN GANofw
. . .
0.0 24
25
26
27
28
29
30
31
Tsk ( “C) 9. Relationship of mean skin temperature (Y&J to increase in metabolic rate (AMR) during exposure to 10°C air. Values are normalized to lean body mass (LBM). Regression lines AMR/LB,M = 7.26 0.23T,, (R2 = 0.65) and 5.17 - O.l7T,, (R2 = 0.62) for LEAN and NORM, respectively, are significantly different (P < 0.001). FIG.
6-
. 5 0
.~I..l~.l.‘~.‘~ll~~~~~~~ 15 30 45
60
75
90
105
120
Time (min) FIG. 7. Heat transfer coefficient (h, mean t SE) at PE and GA muscle sites during exposure to 10°C air. Values increased significantly over time but were not significantly different between LEAN and NORM for either muscle site.
increase in h with increasing exposure time. Since the ambient air conditions (temperature, humidity, and circulation) did not change over the exposure period, the increase in h must be attributed to a decrease in the boundary air layers at the muscle sites caused by an increase in movement of the muscle site through shivering. This is consistent with the increase in shivering intensity observed from the EMG recordings. iMetaboLicrate. The MR rose rapidly for both groups on initial exposure to the cold and then increased slowly but significantly over time, as shown in Fig. 8. This pattern is consistent with the findings of Girling (6) for subjects exposed to 10°C air and Glickman et al. (7) for clothed subjects exposed to -3OOC air. The high values of MR between 75 and 90 min are consistent with the time of maximum shivering noted earlier. The average peak 200
1
180 160
value of about 156 W for MR reported by Girling (6), which occurred at the end of a 90-min exposure to 10°C air, is very close to that found for the NORM group in the present study. However, MR of the LEAN group in the present study was significantly higher than that of the NORM group. The decrease in MR of the LEAN group after 90 min is not significant because two subjects were removed at this time and mean MR of those remaining was lower than that indicated at 86 min. From previous work (18, 19) it can be expected that AMR depends on decreases in both the core and skin temperatures. Although the change in Tsk is large during cold exposure in the present study, the change in T, is not (mean maximal changes in T,, were +0.15 and +O.l6”C for the LEAN and NORM groups, respectively; Fig. 4). Therefore, T, cannot be considered to contribute to the shivering stimulus (the small decrease in T, before exposure is discussed further below). Attention is instead focused on the relationship between AMR and changes in ‘I‘,,. First, the AMR values during the cold exposure were divided by the lean body mass [LBM = body mass X (100 - %BF)/lOO] of the groups to normalize for muscle content, although the difference in LBM between the two groups is not significant (Table 1). Figure 9 shows a tendency for the LEAN group to respond to a similar cold stress with a higher AMR (based on the same decrease in T,J than the NORM group. A simple regression of these results (Fig. 9) indicates a significant difference (P < 0.001) between the two groups of subjects. DISCUSSION
80 60
-15
-
LEAN
-
Nofwl
1 0
15
30
45
Time
60
75
90
105 120
(min)
8. Metabolic rate (MR, mean & SE) before and during exposure to 10°C air. Values during exposure increased significantly over time, and LEAN values were significantly higher than NORM values. FIG.
In general, this study reaffirms the finding of Spurr et al. (17) that shivering begins in the trunk region and spreads to the limbs. Spurr et al. reported that the average time to the first indication of shivering (in either the neck, pectoral, or abdominal region) was 6.4 min. In the present study, the average time to shivering of both BF groups was the same for the pectoral region, i.e., 4.1 min, but differed greatly for the abdominal region (5.4 and 21.4 min for the LEAN and NORM groups, respectively). This order in onset times, in fact, was observed for all
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 20, 2018. Copyright © 1991 American Physiological Society. All rights reserved.
ONSET
OF
muscle groups except PE, suggesting that the LEAN subjects respond sooner to cold exposure than do the NORM subjects. Another interesting observation is that the time of onset of shivering of FE coincides approximately to the onset time of AB, which may be due to the proximity of FE to the trunk region and/or related to muscle mass (i.e., onset may be sooner with increased muscle mass). The observation of increasing h values with time of exposure (Fig. 7) can be explained by a decrease in the thickness of the air boundary layer that occurs with increasing relative air velocity (15). This can occur with an increase in air velocity and/or movement of the surface itself. The fact that no increase in room airflow was detected during the exposures suggests that increased movement of the skin surface by way of increased shivering leads to the increase in h over time. In addition, the h values of the PE muscle site are higher than those for the GA muscle site, in contrast to the theoretical prediction of a decreasing h, with decreasing surface curvature. Although it is possible that the trunk was in contact with a higher air velocity than were the limbs during the cold exposure, it is quite unlikely that the air velocity was almost an order of magnitude higher, as required by theory (15) to explain the different h values. Hence, the higher h values observed for the PE muscle site are probably due to a much higher shivering intensity, which is consistent with the visual observation discussed earlier. This finding suggests that predictions of heat loss that are based on fixed values of h derived from nonshivering conditions underpredict the actual heat loss during shivering. Despite attempts to ensure a steady-state thermal condition at the start of the cold exposure, a slight but significant decrease in T,, (
