Aged humans often exhibit an impaired defense of core temperature during cold stress. However, research documenting this response has typically used small subject samples and strong cold stimuli. The purpose of this study was to determine the responses of young and older subjects, matched for anthropometric characteristics, during mild cold stress. Thirty-six young (YS; 23 ± 1 years, range 18–30) and 46 older (OS; 71 ± 1 years, range 65–89) subjects underwent a slow transient cold air exposure from a thermoneutral baseline, during which esophageal (Tes) and mean skin temperatures (Tsk), O2 consumption, and skin blood flow (SkBF; laser-Doppler flowmetry) were measured. Cold exposure was terminated at the onset of visible sustained shivering. Net metabolic heat production (Mnet), heat debt, predicted change in midregion temperature (ΔTmid), and tissue insulation (It) were calculated. Cutaneous vascular conductance (CVC) was calculated as laser-Doppler flux/mean arterial pressure and expressed as percent change from baseline (ΔCVC%base). There were no baseline group differences for Tes, but OS Mnet was lower (OS: 38.0 ± 1.1; YS: 41.9 ± 1.1 W · m−2, P < 0.05). Tes was well maintained in YS but fell progressively in OS (P < 0.01 for all timepoints after 35 min). The skin vasoconstrictor response to mild cold stress was attenuated in OS (42 ± 3 vs. 53 ± 4 ΔCVC%base, P < 0.01). There were no group differences for Tsk or It, while Mnet remained lower in OS (P < 0.05). The ΔTmid did not account for the drop in Tes in OS. Healthy aged humans failed to maintain Tes; however, the mechanisms underlying this response are not clear.
- skin blood flow
- core temperature
upon exposure to cold stress, the cutaneous vasculature constricts to reduce heat loss, metabolic heat production increases and shivering begins, in an effort to maintain core temperature (Tc). However, with advancing age, these defense mechanisms may be impaired (10, 11, 14, 15, 18, 37). Epidemiological evidence indicates that ∼60% of hypothermia deaths in the United States occur in those aged >65 years (6), and it is generally recognized that older adults fail to maintain Tc during severe cold air stress (10, 11, 14, 36). Older adults may have a lower resting metabolic rate (M) (10) due to both decreased skeletal muscle mass (24) and an impaired metabolic response to cold stress (14, 38), although contrary reports exist (22, 37). Furthermore, an attenuated vasoconstrictor response has been well documented in older subjects (7, 15, 18, 30), possibly leading to greater heat loss. However, with few exceptions (11, 36–38), there is a paucity of studies examining both heat production and heat conservation mechanisms during cold exposure in young and older subjects and those that have been published suffer from one or more deficiencies.
Much of the aging and cold-stress literature has used fairly severe cold stimuli, with cold air exposures ranging from 5 to 10°C dry-bulb temperature (Tdb) (3, 10, 14, 15, 36, 37), conditions under which individuals, regardless of age, would normally employ behavioral thermoregulation, such as adjusting the temperature or donning more clothing, which they were restricted from doing during these studies. Although these studies have been valuable in examining various thermoregulatory reflexes and differences due to aging, sex, and body composition, they have limited application to the environmental stresses normally experienced by most adults in daily living. Moreover, these studies are limited by small sample sizes, as low as eight subjects (11, 14), the potential interaction between gender and body composition (36, 37), and the frequent use of an “older” subject sample that does not meet commonly accepted age criteria of “old” or “elderly” (i.e., age >65 years) (3, 10, 36, 37). The relationship between body fat content and the metabolic response to cold is well known (5) and may explain age- and gender- associated differences reported by others (37). Additionally, the use of an abrupt transition to a severe cold stimulus in these studies excludes the possibility of determining whether a less severe cold stress, which could reasonably be encountered in daily life, would elicit an impaired defense of Tc in older subjects.
The maintenance of Tc and therefore heat balance is represented in its simplest form as heat production equaling heat loss. Data from young subjects indicate that the thermometric determination of changes in body heat content may not be valid compared with estimation by partitional calorimetry (19, 35). It has been proposed that the temperature change of a “mid-region” (Tmid), distinct from the core and skin regions, may account for the discrepancy between thermometric and calorimetric determinations of heat debt (33). Calculation of predicted Tmid has not been conducted in older subjects or during mild cold stress.
Thus the purpose of the present investigation was to determine the influence of primary human aging in the absence of overt pathology, in a relatively large subject population, on the defense of Tc during a mild cold transient. We hypothesized that older subjects, with similar group anthropometric characteristics compared with young subjects, would have an impaired defense of Tc because of impaired cutaneous vasoconstriction and/or a lower metabolic rate. An additional purpose was to determine the applicability of the midregion concept in young and older subjects undergoing mild cold stress.
MATERIALS AND METHODS
Thirty-six young subjects (YS; 18–30 yr; 16 men, 20 women) and 46 older subjects (OS; 65–89 yr; 24 men, 22 women) subjects participated in this study. All subjects were normotensive, nonsmokers, and not taking any medications that might alter the cardiovascular or thermoregulatory responses to cooling. Young women were eumennorheic, not taking oral contraceptives, and were tested in the early follicular phase of the menstrual cycle. All older women were postmenopausal and not taking hormone replacement therapy. All subjects abstained from alcohol and caffeine for 12 h before reporting to the laboratory on the day of the experiment. Verbal and written informed consent was obtained from each subject before participation, and the protocol was approved in advance by The Pennsylvania State University Institutional Review Board.
Subjects underwent a standardized medical screening, including a resting electrocardiogram, blood chemistry analysis (CHEM-24, complete blood count, and thyroid hormone analysis, Quest Diagnostics), and a physical exam. Body composition (%fat) was determined via dual-energy X-ray absorptiometry (DXA; model QDR 4500W, Hologic, Waltham, MA). Appendicular skeletal muscle mass (ASMM) was taken as the sum of the arm and leg lean masses determined via DXA and expressed relative to height in square meters (kg/m2) as proposed by Baumgartner (2). Skinfold thickness was measured at the chest, midaxillary, tricep, subscapular, abdominal, suprailiac, and midthigh sites by the same investigator as an estimate of subcutaneous adiposity. Body surface area (Ad) was estimated, according to Dubois and Dubois (8) and the surface area-to-mass ratio (Ad/mass) was calculated.
Instrumentation and measurements.
Subjects arrived at the laboratory between 0800 and 0900. A copper-constantin thermocouple sealed in a pediatric feeding tube was inserted through the naris a distance equal to 1/4 of the subjects standing height for measurement of esophageal temperature (Tes) (39). The subject then entered the environmental chamber (Tdb = 26.5°C) and was positioned in a semirecumbent position, dressed only in shorts (men) or shorts and a sports bra (women). Skin temperatures were measured using copper-constantin thermocouples affixed to the skin at eight sites. Mean skin temperature (Tsk) was calculated as follows: 0.07Thead + 0.175Tchest + 0.175Tback + 0.07Tupper arm + 0.07Tforearm + 0.05Thand + 0.19Tthigh + 0.20Tlowerleg (12, 21).
Skin blood flow (SkBF) was measured via integrated laser-Doppler flowmetry (LDF; DRT4, Moor Instruments, Devon, England). LDF probes were placed on the ventral surface of the right forearm, taking care to avoid any surface blood vessels, and data were recorded continuously throughout the experiment. Arterial blood pressure was measured every 10 min via brachial auscultation and mean arterial pressure (MAP) was calculated as 1/3(pulse pressure) + diastolic blood pressure. Cutaneous vascular conductance (CVC) was calculated as laser-Doppler flux/MAP and expressed as percent change from baseline values (ΔCVC%base). Forearm blood flow (FBF) was measured by venous occlusion plethysmography (Hokanson EC4, Bellevue, WA) using a mercury-in-Silastic strain gauge and forearm vascular conductance (FVC) was calculated as FBF/MAP. Oxygen consumption (V̇o2) and carbon dioxide production were measured via open-circuit spirometry for 3 min every 10 min (TrueOne 2400 Metabolic Measurement System, ParvoMedics, Salt Lake City, UT). The first expired air sampling and FBF measurements were performed after 5 min of baseline had elapsed and were repeated every 10 min thereafter. Thermal sensation was assessed every 10 min using a 0–8 scale, in which 0 = unbearably cold, 4 = thermoneutral, and 8 = unbearably hot (41). All temperature and SkBF data were recorded and stored as 1-min averages using computer software (LabView) and a data-acquisition system (National Instruments, Austin, TX).
Upon entering the environmental chamber, a 20-min thermoneutral (Tdb 26.5°C) baseline period began. After the baseline period, Tdb was steadily reduced at a rate of 0.25°C/min for 20 min, followed by a decrease in Tdb at a rate of 0.05°C/min for the remainder of the protocol. The protocol was terminated when sustained involuntary shivering was reported by the subject and/or observed by the investigators.
Metabolic rate, M, was calculated as: M (W·m−2) = 352*(0.23*RER+0.77)*(V̇o2/Ad) (12). M was corrected for respiratory evaporative and convective heat loss and expressed as net metabolic heat production, or Mnet (12). Heat storage (S) was calculated as S (W/m) = Mnet − (R+C) − Ediff, where (R+C) represents heat loss due to radiation and convection and Ediff represents heat loss due to water vapor diffusion through the skin (23). R+C was calculated as h(Tsk−Tdb), where h is the combined radiative and convective heat transfer coefficient, calculated according to Parsons (23). Ediff was calculated as fw*16.5*hc*(Psk − Pa) (31); fw is minimum skin wettedness, assumed as 0.06, hc is the convective heat transfer coefficient, Psk was determined by Antoine's equation (23), and Pa was obtained from a standard psychrometric chart. Heat debt (HD) was calculated as HD (kJ/m2) = Σ[Mnet-(R+C)−Ediff]*Δt, where Δt is the time interval. By convention, HD is usually expressed in kilojoules, where 1 W·h = 3.6 kJ. Whole body tissue insulation (It;°C·m2·W−1) was calculated as (Tes − Tsk)/(Mnet − S).
The midregion is assumed to encompass all parts of the body not included in the core and skin, specifically fat, connective tissue, muscle and bone. The calculation of Tmid is based on the principle of conservation of heat, which implies that HD must equal the sum of the products of the temperature changes in all body compartments, accounting for their heat capacities and masses. The specific heats of skin, core, fat, and midregion were 1.02, 1.00, 0.64, and 0.91 W·h·°C−1·kg−1, respectively (33). The mass fractions were calculated as follows: ffat = %fat/100, fsk = 0.062*(1 − ffat), fc = 0.159*(1 − ffat), and fmid = 0.779*(1 − ffat) (33). The change in Tmid was calculated as (−HD/wt-fsk*csk*ΔTsk−fc*cc*ΔTc)/(ffat*cfat+fmid*cmid), where wt is body mass in kilograms and f and c are the whole body fraction and specific heat of each region, respectively.
Data were analyzed using Student's t-test for subject characteristics and repeated measures ANOVA and post hoc t-tests with Bonferroni correction for multiple analyses where appropriate (SAS statistical software, version 9.1, SAS Institute, Cary, NC). Because of the possible interaction of age, gender, and %fat, multiple regression analysis was performed to determine if any of these variables predicted the change in Tes. Age and sex were included in the regression model as dummy variables, where 0 = male, 1 = female and 0 = young, 1 = older. Correlation analysis was performed to test for interrelationships among significant predictors identified by regression analysis. Time to onset of shivering was analyzed using survival curve analysis (Minitab statistical software, version 14, Minitab, State College, PA). Statistical significance was set at α = 0.05, and data are expressed as means ± SE unless otherwise noted.
Subject characteristics are presented in Table 1. The two age groups were well matched for all variables, except for a statistically significant difference in plasma thyroid stimulating hormone. The wide range of adiposity in each group was representative of the general population, ranging from the 10th to the 85th percentile compared with published reference data (1). Multiple regression indicated that age and %fat were significant predictors of ΔTes (P < 0.01 for each), but sex was not. Additionally, sex and %fat were significantly correlated (r = 0.68, P < 0.001), while age and %fat were not; therefore, data for men and women were not analyzed separately in the ANOVA model.
During baseline, Tsk was 32.6 ± 0.1 and 32.6 ± 0.1°C (P = 0.68), and thermal sensation was 4.2 ± 0.1 and 4.1 ± 0.1 (P = 0.50), in YS and OS, respectively, indicating sensory thermal neutrality. There was a trend toward a lower Tes in OS at baseline (YS: 37.09 ± 0.05, OS: 36.98 ± 0.04°C; P = 0.07). The median time to shivering was 81 and 79 min for YS and OS, respectively (P = 0.87), and there was no age difference for Tsk (YS: 29.4 ± 0.3°C, OS: 29.3 ± 0.2°C, P = 0.70) or Tdb (YS: 20.7 ± 0.2°C, OS: 20.9 ± 0.2°C, P = 0.38) at the onset of shivering. During cooling, OS failed to maintain Tes, decreasing by ∼0.2°C (Fig. 1B). Conversely, YS demonstrated a slight increase in Tes (P < 0.01 from 45 to 85 min). The difference in Tes between groups was significant after 35 min (Tdb = 23.3 ± 0.1°C), and the gap widened over time. There was no difference in Tsk between groups throughout the experiment (P = 0.33; Fig. 1C).
Figure 2 depicts the blood flow responses to cooling. CVC was higher in OS throughout the protocol (P < 0.001; Fig. 2A). When expressed as ΔCVC%base OS had an attenuated vasoconstrictor response to cold compared with YS (P < 0.01 from 35 min onward), with a maximal vasoconstrictor response of −42 ± 3 ΔCVC%base in OS and −53 ± 4 ΔCVC%base in YS (P < 0.01). There was no age difference for FVC throughout the protocol (P = 0.25). While OS baseline MAP was ∼7 mmHg higher than YS (92 ± 1 vs. 85 ± 1 mmHg, P < 0.01), the increase in MAP during cooling was similar between groups, so that the ∼7 mmHg difference remained throughout cooling.
OS had a lower Mnet throughout the protocol (P < 0.05 vs. YS; Fig. 3A). The lack of significant difference toward the end of the protocol was likely due to reduced statistical power due to subject attrition, rather than a physiological difference. Heat loss by either water vapor diffusion through the skin (Ediff) or radiation and convection (R+C) were not different between age groups (age effect P = 0.82 and P = 0.37 for Ediff and R+C, respectively). There were no age differences for HD or predicted ΔTmid (Fig. 3, B and C).
Baseline It was 0.062 ± 0.002 and 0.063 ± 0.002°C·m2·W−1 and peak calculated values were 0.092 ± 0.003 and 0.094 ± 0.002°C·m2·W−1 for YS and OS, respectively (P = 0.95 for age effect). Thermal sensation did not differ between groups (P = 0.26), with both groups rating baseline as “comfortable” and reaching “very cold” by cold exposure termination.
The purpose of this study was to investigate the mechanisms underlying age-related differences in the defense of Tc during mild passive cooling in a large cohort of young and older subjects. An additional purpose was to extend the midregion temperature concept to older subjects and to mild cold stress. Two subject groups, well matched for relevant anthropometric characteristics but differing in age by almost 50 years, were compared. The principle finding of this study was that OS failed to defend Tes during even a mild cold transient. However, a biophysics analysis of the data did not reveal any age-associated differences. Heat production, while lower in the older subjects during cooling, was also lower at baseline. Heat loss was similar between age groups, even though the older subjects demonstrated an attenuated vasoconstrictor response. The predicted change in midregion temperature did not help explain the decrease in esophageal temperature in spite of equal heat loss and heat production. However, it has been reported by others (26, 33) that ΔTc and HD show a poor relationship to one another. It is only when the rate of change of HD is relatively slow (i.e., during heat balance) that these variables correspond well. Additionally, numerous studies have documented a paradoxical increase in Tc during the first ∼30 min of cold exposure, when HD is increasing rapidly.
Examination of the weighting coefficients for core, mid-, and skin regions may provide an explanation for the lack of a difference in midregion temperature between groups. Recalling that the relative mass of the midregion is almost 5 times as large as the core region (0.779 vs. 0.159), 1/5th of the change in Tes is all that would be required in the midregion. Considering that the Tes difference between YS and OS was ∼0.35°C, the difference in predicted ΔTmid between groups would only be ∼0.07°C. We believe that this difference is statistically undetectable, even in a large subject population. Accurate prediction of ΔTmid may require a relatively large change in Tes, and empirical measurements of Tmid are needed to validate the model. Additionally, as the weighting coefficients were derived from young subjects' data, different coefficients may be required in older subjects.
Previous studies showing similar decrements in Tc used more extreme cold stress than in the present study, such as cold air exposure ranging from 5 to 10°C (3, 10, 14, 15, 36, 37). Furthermore, exposure to these conditions was usually sudden, that is, the subject was rapidly introduced to the experimental conditions. Two limitations arise from previous experiment designs: the cold stress has limited application to the conditions normally experienced by adults during every-day living, and the temperature at which groups begin to differ in their responses is unknown. By using a milder cold stress that incorporated a gradual transition from thermoneutral to cool conditions, we attempted to overcome these limitations. When Tdb was 23.3°C, ∼3°C lower than thermoneutral baseline conditions, the OS's Tes was significantly lower than that of the YS, and OS had already begun to show an attenuated cutaneous vasoconstrictor response. These findings indicate that a mild cold stress can elicit an attenuated thermoregulatory response and failure to defend Tes in OS. The magnitude of the difference between subject groups continued to increase as Tdb decreased further.
An attenuated vasoconstrictor response to cold in an aged population has been well documented, despite differences in the cold stimulus and methodology used (3, 11, 15, 18, 25, 29, 30, 36). Kenney and Armstrong (15), using venous occlusion plethysmography as an index of SkBF, were the first to control for confounding anthropometric characteristics that could affect thermoregulatory responses. Recent research indicates that reflex vasoconstriction is mediated by norepinephrine (NE) and an unknown cotransmitter (27, 30) in young subjects. Our laboratory has recently demonstrated that a loss of cotransmission (30) and decreased sensitivity to NE (29) both contribute to the attenuated vasoconstrictor response in aged healthy human subjects. When the vasoconstrictor effects of NE were blocked, young subjects retained ∼40% of their vasoconstrictor capacity, while vasoconstriction was abolished in OSs, indicating loss of functional cotransmission in these subjects (30). Administration of exogenous NE revealed blunted vasoconstriction in older subjects, indicating that decreased sensitivity to NE contributes to these subjects' attenuated vasoconstrictor response (29). However, local cooling responses are maintained in aged subjects (28), suggesting that the attenuated vasoconstrictor response in our subject population is due to reflex rather than local effects.
A failure of older subjects to increase metabolic rate to the same extent as young subjects during a cold stress has been known for many years (14), however, most studies have used a severe cold stress to induce shivering. The present investigation differs in that the exposure was terminated when shivering began; therefore, our data represent nonshivering conditions throughout the exposure. An increase in Mnet was not expected, as nonshivering thermogenesis contributes little, if at all, to heat production during cold stress in adult humans (9). Despite the fact that our subject groups were matched for ASMM, baseline Mnet was lower in the older subjects and remained so throughout the experiment. There were no differences for plasma thyroxine or triiodothyronine in our subjects, suggesting that other factors may mediate the difference in Mnet. Lean body mass (LBM) has been reported to account for over half of the variability in metabolic rate (40), while the factors contributing to the remaining variability have yet to be elucidated. Proposed mechanisms include lower fat oxidation, decreased skeletal muscle protein turnover, and lower Na+-K+ ATPase activity (40), which may contribute to the difference in resting metabolic rate between young and older subjects. Our It results support those of Budd and colleagues (3), who reported no significant relationship between age and It in subjects ranging from 26 to 52 years old. Recalling that It = (Tes − Tsk)/(Mnet − S), age-related differences in any of these variables would determine differences in tissue insulation. Our findings indicate that the lower Tes and Mnet offset each other, resulting in no age-related differences for It. The influence of reduced LBM on It within an aged population warrants further attention (16, 34).
Simultaneous measurement of Tc, Mnet, and SkBF in young and older subjects has been reported in only four studies (11, 36–38). Wagner and colleagues (38) did not show any difference in limb blood flow, Mnet, or Tc between young (20–29 yr) and middle aged (46–67 yr) men. In a later study, Wagner and Horvath indicated that older subjects failed to maintain Tc but did not demonstrate any age-related attenuation in cutaneous blood flow during the cold stress (36, 37). However, it should be noted that the subject groups in that study differed significantly with respect to body fat, leading the authors to conclude that the changes in Tc were due to body fat differences. Using a cold saline infusion model, Frank and colleagues (11) reported decreased fingertip blood flow when assessed via LDF, an attenuated metabolic response, and a greater drop in tympanic temperature during the cold challenge in older subjects. Cold saline infusion creates a condition in which the core is cooled faster than the skin, which is opposite of the normal physiological response to cold, limiting the generalizability of these results. The present study appears to be the first to simultaneously measure Tc, Mnet, and SkBF during environmental cooling in young and older subjects matched for relevant anthropometric characteristics.
During exposure to cold, the stimulus for shivering is usually peripheral (Tsk) rather than central (Tc), although shivering has been induced during thermoneutral Tsk conditions (4), and direct application of cold to the hypothalamus in experimental animals (13). It should be noted that the present study was not designed a priori to elucidate time to shivering differences. One of our primary goals was to model the vasoconstrictor response to mild cooling; therefore, we choose shivering as our end-point as it is a good indicator that maximal vasoconstriction had been achieved. Inspection of individual CVC curves for each subject indicated that a nadir and plateau were reached before the onset of shivering. Therefore, we were liberal in our assessment that shivering had begun, and EMG analysis may have indicated a longer time to shivering. Also, while we observed shivering, there was no increase in metabolic rate, which has been used as an indicator that shivering began (17). Kenny and colleagues (17) reported a Tsk shivering threshold ∼0.4°C lower than in the present investigation, using a similar cold transient protocol in YS. We are unaware of any rigorously controlled studies examining the Tsk threshold for shivering in YS vs. OS.
We chose not to subgroup the age groups by sex, as others have noted that apparent sex differences during cold stress are actually due to body composition differences (5, 20, 32, 37), which were not present in our study population. Additionally, sex was not a predictor of the change in Tes. The quantity and distribution of body fat are related to the metabolic response and It (5), and to changes in Tsk and Tes (37). However, other anthropometric characteristics may be relevant, as Ad/mass may account for observed sex differences (20, 32). McArdle and colleagues concluded that higher %fat in the women did not appear to provide any protective benefit for maintaining Tc, and differences in LBM may also account for some of the gender difference (20).
In summary, older men and women, when exposed to mild cold transients, demonstrated an attenuated vasoconstrictor response and failed to maintain Tes. These results appear related to chronological age per se, as the groups were well matched for relevant anthropometric characteristics. A biophysical analysis of heat balance failed to yield an explanation for the failure of older subjects to defend Tes during mild cold stress.
This study was supported by National Institutes of Health Grants R01-AG-07004–15 (to W. L. Kenney) and M01-RR-10732 (General Clinical Research Center) and National Institutes of Aging predoctoral training grant T32-AG-000048–28 (Penn State University Gerontology Center).
The authors thank the subjects for their participation in this study, and the General Clinical Research Center for medical screenings and consultations. The data collection assistance of Lacy Holowatz, Caitlin Thompson, and Jane Pierzga is greatly appreciated. The expertise and advice of Peter Tikuisis concerning the calculating of midregion temperature is appreciated as well.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society