The relative importance of skin vs. core temperature for stimulating cold acclimation (CA) was examined by 5 wk of daily 1-h water immersions (20°C) in resting (RG) and exercising (EG) subjects. Rectal temperature fell (0.8°C;P < 0.05) during immersion only in RG. Skin temperature fell (P < 0.05) similarly in both groups. Physiological responses during cold-air exposure (90 min, 5°C) were assessed before and after CA. Body temperatures and metabolic heat production were similar in both groups with no change due to CA. Cardiac output was lower (P < 0.05) in both groups post-CA (10.4 ± 1.2 l/min) than pre-CA (12.2 ± 1.0 l/min), but mean arterial pressure was unchanged (pre-CA 107 ± 2 mmHg, post-CA 101 ± 2 mmHg). The increase in norepinephrine was greater (P < 0.05) post-CA (954 ± 358 pg/ml) compared with pre-CA (1,577 ± 716 pg/ml) for RG, but CA had no effect on the increase in norepinephrine for EG (pre-CA 1,288 ± 438 pg/ml, post-CA 1,074 ± 279 pg/ml). Skin temperature reduction alone may be a sufficient stimulus during CA for increased vasoconstrictor response, but core temperature reduction appears necessary to enhance sympathetic activation during cold exposure.
- cold-induced vasodilation
humans can become naturally acclimatized or experimentally acclimated to cold (25). The specific physiological adjustments that develop vary in three general patterns depending on subjects' physical characteristics (3), the severity of cold stress (15, 18, 26), and/or the duration and frequency of cold exposures (4,15, 25). Previous studies from this laboratory (26) demonstrated that an insulative pattern of cold acclimation (CA) develops after repeated daily cold-water immersions (CWIs), characterized by a more rapid and more pronounced decrease in skin temperature during cold-air exposure (CAE) after CA. In that report (26), it was speculated that the stimulus for this insulative acclimation was the repeated reduction in the subjects' core temperatures during each CWI. However, the subjects in that investigation (26) also experienced repeated reductions in skin temperature. Thus the relative importance of repeated reductions in skin as opposed to core temperature for the stimulation of CA was not established.
To address this, we compared thermoregulatory adjustments to CA produced by repeated CWI in subjects whose core temperatures were either maintained constant or allowed to decrease, whereas the concomitant decrease in skin temperature was similar in both groups. The acclimation pattern was compared by evaluating physiological responses during CAE and CWI before, and again after, the repeated CWIs. It was hypothesized that repeated reduction of core and skin temperatures would induce an insulative pattern of CA and that such acclimation would not occur when core temperature was maintained by metabolic heat production, despite repeatedly reduced skin temperature.
A secondary aim was to examine the influence of CA and core temperature on the cold-induced vasodilation (CIVD) response. Our experimental design allowed us to compare CIVD before and after CA and at both normal and reduced core temperature. No studies have longitudinally examined the effects of whole body CA on CIVD. Whole body cooling has been reported to blunt or abolish CIVD (8,16), although these studies did not actually measure core temperature. It was hypothesized that the more pronounced vasoconstrictor response observed with insulative CA (25,26) would blunt CIVD and that CIVD would be blunted when core temperature was reduced due to whole body cooling.
Fourteen healthy active men with no history of cold injury participated in this study, which was approved by the Institute Scientific and Human Use Review Committees. Each subject volunteered to participate after being informed of the purpose, experimental procedures, and known risks of the study. Investigators adhered to Army Regulation 70-25 and US Army Medical Research and Materiel Command Regulation 70-25 on the Use of Volunteers in Research. This study required 12 mo to complete, and tests were evenly distributed throughout the year. Subjects who were tested during the winter months spent most of their time indoors or dressed in cold-weather clothing while outdoors; therefore, seasonal effects were not expected to influence these experiments.
Subjects' height (Ht) and weight (Wt) were measured. Body density and residual lung volume were estimated by hydrostatic weighing and nitrogen dilution, respectively (21), and these values were used to calculate body fat content (20). Skinfold thickness was measured at 10 sites: chin, subscapular, chest, side, suprailium, abdomen, triceps, thigh, knee, and calf (10), and mean subcutaneous fat thickness was calculated (2). Body surface area was estimated by 0.007184 · Wt0.425 · Ht0.725 (6).
To assess aerobic fitness, peak oxygen uptake (V˙o 2 peak) was determined by using a continuous-effort, progressive-intensity cycle ergometer protocol. After a brief warm-up, the subject pedaled for 2 min at 60 rpm against zero resistance; thereafter, the power output was increased 30 W every 2 min until the subject was unable to continue despite verbal encouragement. During the exercise, oxygen uptake (V˙o 2) and carbon dioxide output (V˙co 2) were measured continuously.V˙o 2 peak was defined as the highestV˙o 2 achieved. An electrocardiogram, obtained from chest electrodes (CM-5 placement) and sent by radiotelemetry to an oscilloscope-cardiotachometer, was used to record heart rate (HR) at the end of each stage.
As subjects volunteered for this study they were assigned to one of two groups, cold-water exercise (EG) or cold-water resting (RG), on the basis of their physical characteristics such that over the course of the study the groups remained of similar number and did not differ in average fitness or anthropometric characteristics. The physical characteristics of both groups are presented in Table1.
Before subjects began CA, CAE and CWI tests were conducted on separate days. A cold finger immersion (CFI) test was conducted twice: immediately after CWI and again on a separate day. Within 3 days of completing these tests, subjects began CA. To prevent RG from “detraining” and experiencing a decline in aerobic capacity, these subjects participated in a fitness-maintenance program consisting of 60 min of cycling at 50% V˙o 2 peak on land in ambient air twice a week during the acclimation period. The sessions for fitness maintenance were completed 2–5 h after the immersion sessions.
Within 4 days after the last CA session, CAE, CWI, and the two CFI tests were again conducted. All tests were conducted at the same time of day for each subject, between 0700 and 1100, to avoid the confounding effects of circadian rhythms. During any tests or CA sessions, if rectal temperature (Tre) reached 35°C, the test or session was terminated. Subjects were instructed to fast for 12 h before all tests.
After arriving at the laboratory, subjects voided their bladders, and their nude weights were measured. For the experiment, subjects dressed in shorts and socks with holes cut over the top of the foot. They inserted a thermistor 12 cm beyond the anal sphincter for measurement of Tre and a thermocouple into the esophagus at a depth of one-fourth of each subject's height for measurement of esophageal temperature (Tes). A catheter was placed in an antecubital vein for collection of blood samples. Thermocouples were placed on the skin surface of the foot, medial thigh, chest, and back of the hand. Additional skin temperature and heat flow (HF) measurements were obtained by placing HF-thermistor disks (Concept Engineering, Old Saybrook, CT) on the skin surface of the back, triceps, forearm, lateral thigh, and calf. Mean weighed skin temperature (T̄sk) was calculated as T̄sk = 0.06 · Tfoot + 0.17 · Tcalf + 0.14 · Tmedial thigh + 0.14 · Tlateral thigh + 0.14 · Tchest + 0.07 · Ttriceps + 0.07 · forearm temperature (Tfa)+ 0.14 · Tback + 0.07 · Thand(22). HF was calculated as HF = 0.28 HFback + 0.14 HFforearm + 0.08 HFtriceps + 0.22 HFcalf + 0.28 HFthigh (22). Thermal gradient (TG) was calculated as TG = [(Tre + Tes)/2 − T̄sk]. Tissue insulation (Itissue) was calculated as Itissue = TG/HF (23). Electrocardiogram electrodes were attached for measurement of HR.
Subjects rested in a recumbent position for a 40-min baseline period at 26°C and then moved into a chamber where they resumed the recumbent position for a 90-min exposure to 5°C air. Body temperature and HF data were recorded every minute. HR was recorded every 2 min. Metabolic rate (M˙) normalized to body surface area was estimated fromV˙o 2 and the respiratory exchange ratio, which were measured during baseline and for 10 min each 0.5 h of exposure (17). Blood pressure was measured before eachM˙ measurement. Mean arterial pressure (MAP) was calculated from the diastolic and systolic pressures [MAP = diastolic pressure + (systolic pressure − diastolic pressure)]. After each M˙ measurement, cardiac output (Q˙) was estimated in duplicate by using the Collier CO2-rebreathing technique and calculated according to the Fick principle (7). Total peripheral resistance (TPR) was calculated as MAP/Q˙. Blood samples were taken at the end of both baseline and exposure periods for analysis of norepinephrine (NE) by high-performance liquid chromatography with electrochemical detection.
After arriving at the laboratory, subjects voided their bladders, and their nude weights were measured. For the experiment, subjects were dressed in swim shorts and wore aquatic sandals to protect their feet. Instrumentation and measurements were the same as indicated for CAE (see CAE above), with the exception of blood pressure, which was only measured preimmersion. During immersion, subjects sat on a chair constructed of stainless steel, perforated to facilitate water movement around the subject and minimize the formation of a still boundary layer. The chair was mounted on a platform that could be raised and lowered in and out of a 36,000-liter pool in which water was continuously circulated and temperature maintained within ±0.1°C of the desired temperature. The water level was adjusted to cover the subjects' shoulders while they were immersed. Subjects sat quietly for a 40-min baseline period in 26°C air, after which the platform was lowered into the water (20°C) for 60 min. Immediately after the CWI test, subjects were removed from the water, dried, wrapped in blankets, and instrumented for the CFI described in CFIbelow.
On the day CFI was conducted as a separate test [normothermic condition (NC)], subjects arrived in the laboratory and voided their bladder and a rectal temperature probe was inserted. Subjects completed the NC CFI wearing shirts, shorts, and shoes. On the day CFI was conducted directly after CWI [hypothermic condition (HC)], subjects were wrapped in blankets, and instrumentation that was not necessary for CFI was removed. For HC, 11.2 ± 0.5 min elapsed between termination of the CWI and beginning of CFI. A thermocouple was attached to the right anterior forearm and to the right middle finger tip along the nail bed by using ∼1 cm2 of Hy-Tape (Hy-Tape Surgical Products, Yonkers, NY). Subjects sat quietly in ambient air (26.2 ± 0.4°C) for 30 min during NC and for 2 min during HC and then immersed their middle finger to the middle phalanx into 4°C water for 30 min by using a refrigerated water bath (model RTE-111, Neslab Instruments, Newington, NH). HR was recorded every 5 min. Body temperatures were recorded every 2 min during preimmersion and every 10 s during immersion. A plastic cover supported the hand over the water bath.
During CA, each subject was immersed to the shoulder in 20°C water for 60 min, 5 times a week, for 5 wk. During immersion, RG subjects sat quietly, whereas EG subjects performed leg exercise as described below. During alternate sessions of CA, Tre and HR were monitored and recorded before and every 10 min during immersion. In addition, V˙o 2 andV˙co 2 were measured at minutes 10–20 for EG, and at minutes 15–25 and45–55 for RG.
Subjects in the EG cycled at 50% ofV˙o 2 peak during immersion on a semirecumbent cycle ergometer modified for use in water. Each subject performed several short bouts (8–15 min) of leg exercise while immersed in 30°C water to determine the resistance that elicited 50% of V˙o 2 peak. Flywheel resistance was adjusted by attaching fins of varying length to the flywheel (19). Subjects wore aquatic sport sandals, open at the top, toes, and sides, to protect the bottom of the feet during pedaling.
Data were analyzed by using analysis of variance for repeated measures. Differences were evaluated by a three-way analysis among groups, trials, and times of measurement. Tukey's honestly significant difference post hoc test was used to make pairwise comparisons when a significant F ratio was found. Statistical significance was set at P < 0.05, and, unless otherwise indicated, differences reported are significant at that level. Data reported are means ± SE. The statistical analysis program (SAS Institute, Cary, NC) used a general linear model such that missing values did not eliminate observations from the analysis.
A low-pass filter (SigmaPlot 3.03, Jandel Scientific, San Rafael, CA) was used on finger temperature (Tf) during CFI to reduce data noise. A rise or fall in Tf of at least 0.5°C was considered to represent a viable CIVD. Data were examined to identify the time and temperature at the first nadir, the first apex, and the second nadir temperature. For descriptive purposes, subsequent nadir and apex temperatures were also determined. The means of these data were used to construct an “average” CIVD for each of four conditions for statistical analysis: preacclimation normothermic, preacclimation hypothermic, postacclimation normothermic, and postacclimation hypothermic. These data were evaluated by a two-way analysis of variance among groups and conditions.
All subjects completed the entire 60 min during each of the daily immersions of CA, with the exception of one subject in RG who withdrew early on his first two immersion sessions because of a fall in Tre to 35°C, the medical safety limit. Mean daily Tre immediately before immersion was similar for both groups (37.20 ± 0.06°C). At the end of each immersion, Tre in RG had fallen to 36.35 ± 0.15°C, with an average daily change of −0.83 ± 0.16°C. In EG, Treremained unchanged during CA sessions (37.22 ± 0.15°C).
The average duration of CAE was 90 ± 0 and 78 ± 7 min pre-CA and was 90 ± 0 and 87 ± 2 min post-CA for RG and EG, respectively. Three of the subjects in EG withdrew before completing the entire 90 min of CAE pre-CA, and two of these withdrew before completing the entire 90 min of CAE post-CA. All subjects withdrawing from the exposure before completion complained of severe finger pain.
Figure 1 shows Tre,T̄sk, TG, Itissue, and HF for CAE. After an increase on initial exposure to cold, Tre decreased by 90 min CAE. The T̄sk decreased throughout CAE, and TG increased. After an initial decrease on cold exposure, Itissue returned to preexposure values by the end of CAE. On exposure to cold, HF increased but then decreased as CAE continued. There were no differences between RG and EG, or between pre-CA and post-CA, for Tre, T̄sk, Itissue, or HF during CAE. There were no differences between pre-CA and post-CA for TG, but post-CA TG was higher in RG than EG. A similar increase in M˙ occurred during CAE for RG and EG, both before and after acclimation, from preexposure (52.1 ± 1.6 W/m2) to cold exposure (132.3 ± 4.0 W/m2).
The percent changes in Q˙, MAP, and TPR induced by CAE are shown in Fig. 2. There were no group differences in any of these measurements; therefore, data are shown for all subjects combined. The increase in Q˙ induced by CAE was smaller after CA, compared with before. The increase in MAP during CAE was the same before and after CA. The reduction in TPR during CAE was smaller after CA, compared with before. After CA, the increase in NE (Fig. 3) during CAE was greater for RG postacclimation compared with preacclimation, but CA had no effect on the NE response to CAE in EG.
The average duration of CWI was 59 ± 3 and 58 ± 5 min pre-CA and was 60 ± 0 and 60 ± 1 min post-CA for RG and EG, respectively. Two subjects in RG and one subject in EG did not complete the entire 60 min of CWI pre-CA; the subject from EG also was withdrawn after 57 min post-CA. All withdrawals from the immersion were mandated by core temperature falling to the medical safety limit (35°C).
Figure 4 shows Tre,T̄sk, TG, Itissue, and HF for CWI. During CWI, Tre, T̄sk, and TG decreased. After an initial decrease on immersion, Itissue increased by the end of CWI. The HF initially increased on CWI and then decreased as CWI continued. CA had no effect on Tre,T̄sk, TG, or M˙ during CWI for either RG or EG. At the end of CWI, Itissue in RG was lower after acclimation, compared with preacclimation, and higher than EG, both pre- and postacclimation. A lower HF was observed in EG postacclimation compared with preacclimation, but CA had no effect on HF in RG. A similar increase in M˙ occurred during CWI for RG and EG, both before and after CA, from preimmersion (55.6 ± 1.6 W/m2) to CWI (135.7 ± 4.2 W/m2). Both RG and EG exhibited a similar increase in Q˙ from 5.83 ± 0.22 l/min before CWI to 10.01 ± 0.36 l/min during CWI, and CA had no effect on Q˙ during CWI. There was an increase in NE (Fig. 3) during CWI compared with preimmersion, but CA had no effect on the NE response to CWI.
Table 2 presents Tre, Tf, and Tfa responses of RG and EG during the four CFI tests. The Tre was lower at the beginning of CFI during HC than NC, as intended by the study design; Tfa was lower throughout HC, compared with NC; and the initial Tfwas lower than NC. During HC, Tf at the first nadir and the first apex was lower than NC. In addition, time to the first nadir and first apex was longer during HC than NC. After CA, Tf at the second nadir was lower in all subjects, compared with pre-CA. Figure 5 shows Tf for all subjects during CFI tests before CA under NC and HC (A) and under NC pre- and post-CA (B).
Previously, our laboratory (26) observed that subjects immersed in 18°C water for 90 min/day, 5 days/wk, for 5 wk developed characteristics of an insulative-type of acclimation. Specifically, after CA, CAE elicited a faster and more pronounced decrease in skin temperature and a greater increase in blood NE compared with the nonacclimated state (26). We interpreted this as indicating that CA had enhanced the peripheral vasoconstrictor response and sympathetic nervous response to cold (26). The present investigation aimed to test our speculation that the physiological stimulus for development of this type of CA was the repeated core temperature reduction experienced by our subjects during CA. Our experimental approach was to compare the effects of repeatedly reducing both core and skin temperature in resting subjects immersed in cold water with the effects of repeatedly reducing only skin temperature in subjects who exercised during CWI to maintain core temperature constant. The findings indicate that repeated core temperature reduction is necessary to stimulate the enhanced sympathetic nervous response to cold but may not be required for the development of a more pronounced vasoconstrictor response to cold.
As is usually observed (18, 26), cold exposure (both in air and water) caused circulating NE concentrations to increase in our subjects, which is typically attributed to increased sympathetic nervous activity. In agreement with previous observations from our laboratory (26), subjects in this investigation who acclimated by resting in cold water and experienced repeated reduction of both core and skin temperature exhibited a more pronounced increment in circulating NE during CAE after CA. In contrast, subjects who acclimated by exercising in cold water and only experienced the repeated reduction of skin temperature without any decrease in core temperature did not exhibit any change in the increment of circulating NE elicited by CAE. Thus development of an enhanced sympathetic response to cold appears to require repeated core temperature reduction during CA, and repeated skin temperature reduction alone may be an insufficient stimulus for that adjustment.
Alternatively, development of training effects in the exercising group may have modulated or prevented acclimation effects from being manifested. During the acclimation period, the EG completed 60 min of cycle exercise, 5 days/wk, whereas the RG only exercised 60 min, 2 days/wk. Our intent was that the two groups remain closely matched with respect to their maximal aerobic capacity and body composition throughout the study, and this was accomplished. However, other training effects may not have been successfully controlled between groups. For example, aerobic physical training attenuates the sympathetic nervous response to acute aerobic exercise (14). It may be that this training effect was induced in the EG but not the RG because of the additional training by the EG, and perhaps the diminished sympathetic responsiveness to exercise was also manifested in the sympathetic response to cold stress. Furthermore, Weller et al. (24) have demonstrated that cold exposure has a much smaller effect on circulating NE during strenuous than mild exercise. Thus, in our EG, daily cold exposure may have produced less sympathetic stimulation than in the RG in whom the daily cold exposure had a pronounced stimulatory effect on the sympathetic nervous system and circulating NE.
A more pronounced vasoconstrictor response to cold exposure is considered the hallmark of the insulative pattern of CA (25). The findings from this study suggest that this adaptation, in contrast to the enhanced sympathetic nervous response to cold exposure, develops in response to repeated reductions in skin temperature alone and does not require repeated reduction in core temperature during acclimation. This conclusion is demonstrated by our observation that both the EG and RG exhibited a smaller decrease in peripheral vascular resistance during CAE after CA, which is indicative of more pronounced vasoconstriction. TPR reflects resistance in all the body's vascular beds, but those most likely involved in this adaptation are the vascular beds within the peripheral skeletal muscles and skin that constrict in response to cold exposure. Despite the apparent augmentation of the vasoconstrictor response to cold that developed with CA, the subjects in the present investigation, in contrast to those of the earlier study from our laboratory (26), did not exhibit any change in skin temperature response to CAE, nor was there evidence that Itissue (i.e., resistance to HF) was improved.
It is not clear why the more pronounced vasoconstrictor response to cold induced by CA did not also manifest as a greater and more rapid decline in skin temperature or other signs of improved Itissue during CAE as previously observed by our laboratory (26). Possibly the subjects of this investigation had not become acclimated to the same degree as the subjects in the earlier study from our laboratory (26). In the present study, we chose to use a 20°C water temperature because we anticipated that the exercise intensity required to maintain body temperature during immersion at 18°C, as in the previous study from our laboratory (26), would be difficult for most subjects to sustain. The subjects in this study who acclimated by resting in cold water experienced a mean daily core temperature reduction of ∼0.8°C, which was comparable to the reduction experienced by subjects in the earlier study (26). Furthermore, the slight water temperature difference would be expected to account for only a small (∼2°C) difference in the skin temperature during acclimation. Therefore, the magnitude of environmental stress used to stimulate acclimation seems comparable between the two studies. On the other hand, subjects in the earlier study spent 50% more time acclimating; daily immersions in that study lasted 90 min (26) as opposed to 60 min in the present study. Although subjects in both experiments had similar changes in body heat content at the end of immersion, the longer duration of the earlier study means those subjects sustained greater overall heat loss during immersion. Thus exposure duration may be an important determinant for the physiological adjustments produced during CA.
In this study, as in previous studies (1, 8), we observed that the CIVD response was blunted by whole body cooling. One hour of CWI resulted in lower core temperatures before and throughout our HC CIVD assessment compared with the NC trial. CWI also reduced T̄sk values, and these lower skin temperatures persisted, as evidenced by the lower Tfa values, throughout the HC CIVD test. Thus neither this study nor the earlier ones (1, 8) provide conclusive evidence to document whether the blunted CIVD after whole body cooling is an effect of reduced core temperature, reduced skin temperature, or a combined effect of both. Nevertheless, a reasonable argument can be offered to suggest that the reduction in skin temperature is the principal factor accounting for the blunted CIVD after whole body cooling.
First, core temperature was probably not reduced in those earlier studies reporting that whole body cooling blunted CIVD, leaving a skin temperature reduction as the only factor accounting for the effect on CIVD. Although core temperature was not measured in those studies (1, 8), cold exposures of similar magnitude (1 h at 5–7°C) in our laboratory typically reduce skin temperature but have no effect on core temperature (11). Additionally, during the HC in our study, all subjects exhibited a delay in the onset of CIVD and a lower nadir temperature that was similar in magnitude to the effect observed by others, whose experimental conditions probably had no effect on core temperature (1). Furthermore, although there was a considerable variation in core temperatures (35.0–36.9°C) after whole body cooling in our HC trial, there was no correlation between the reduction in core temperature and the magnitude of the corresponding effects on CIVD. Lastly, in a more recent investigation attempting to demonstrate independent effects of altering core and skin temperatures on CIVD (5), a delay in the onset of CIVD occurred during whole body cold exposure (15°C air), compared with neutral conditions, regardless of whether the subjects' core temperature had been increased by drinking hot fluids or decreased by drinking cold fluids, presumably because mean skin temperature was reduced in both conditions. Collectively, these observations suggest that reduced core temperature has little additional effect on the CIVD relative to the effect of reduced skin temperature.
This investigation is the first to attempt a definitive demonstration of how whole body cold acclimation affects CIVD. Experimental observations from studies of persons who have developed a natural acclimatization to cold suggest that repeated or chronic whole body cold exposure may diminish CIVD, possibly because of central nervous system adaptations (9, 12). For example, studies conducted in the 1960s demonstrated that, compared with nonadapted control subjects, Korean diving women who repeatedly dove in cold (7–10°C) water wearing only cotton swimsuits exhibited augmented vasoconstrictor responses during whole body cold exposure, and an absence of CIVD during immersion of the hand in cold water (12). In 1977, these diving women began wearing wet suits, and studies conducted subsequently demonstrated that their core temperature declined much less while they worked compared with their counterparts studied earlier (13). These modern divers no longer exhibited the enhanced vasoconstrictor response to cold, but they did exhibit a CIVD response (13). On the basis of these observations, we hypothesized that subjects acclimated by resting in cold water would develop an enhanced vasoconstrictor response to cold, which would in turn blunt CIVD. Our data demonstrate a lower Tf at the second nadir in both groups of subjects, suggesting that the repeated reduction in skin temperature was sufficient to enhance the vasoconstrictor response and blunt CIVD.
In conclusion, several new findings are apparent from the results of this study. A decrease in core temperature during acclimation sessions appears to be a necessary stimulus for the development of the increased sympathetic response to cold that accompanies insulative adaptation. The duration of the reduction in core temperature may be an important factor for inducing insulative acclimation. A decrease in skin temperature during acclimation sessions is sufficient stimulus for an increased vasoconstrictor response to cold. Future research should consider the benefit of periodic evaluations throughout the course of acclimation to better characterize the time course of the development of the adaptation.
We sincerely thank the volunteers for their effort. The technical assistance of L. Blanchard, J. Bogart, H. Jackson, M. LeQuesne, C. MacLellan, D. McCook, A. Rouse, K. Spear, and R. Weber and the statistical assistance of R. Wallace are gratefully acknowledged.
This study was conducted while D. T. Lee was a National Research Council Fellow/US Army Medical Research and Materiel Command Resident Research Associate.
Address for reprint requests and other correspondence: C. O'Brien, USARIEM-TMD, Kansas St., Natick, MA 01760-5007 (E-mail:).
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. §1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society