Content uploaded by Craig Heller
Author content
All content in this area was uploaded by Craig Heller on Aug 17, 2015
Content may be subject to copyright.
ORIGINAL ARTICLE
Enhancing Thermal Exchange in Humans
and Practical Applications
H. Craig Heller and Dennis A. Grahn
Abstract
Normal human core body temperature is regulated within a narrow range. Deviations from this range can have
serious consequences in both health and disease. However, it is difficult to efficiently manipulate body heat con-
tent because of the high heat capacity of the body and the low thermal conductance of the body surface. Mammals
have evolved vascular adaptations of the nonhairy skin to enable enhanced heat loss. These include arteriovenous
anastomoses that bypass the nutritive capillary beds to shunt the blood into retia venosa which serve as radiators.
We have quantified the area-specific heat loss from glabrous skin (palms and face) and nonglabrous skin (upper
arm, back, thigh, and abdomen). Results show that the heat loss from the nonglabrous skin does not change ap-
preciably over the course of exercise in the heat, whereas the heat loss from the glabrous skin rises to values more
than five times that of the nonglabrous skin. The application of a mild vacuum increases the heat loss from the
glabrous skin by an additional 33%. The effect of cooling of these different skin areas on the heart-rate response
to a fixed exercise load was significantly greater for the glabrous than the nonglabrous skin. The intermittent ap-
plication of vacuum cooling to the palms of individuals exercising in a hot environment had the effects of lower-
ing the rate of rise of core temperature and enhancing performance. The vacuum-enhanced heat exchange via the
glabrous skin is a disruptive technology for several reasons. It forces re-formulation of the models of human ther-
moregulation that are used to design thermal protective gear. It offers an effective means of treating heat and cold
stress. It provides an insight into controversies about the effects of temperature on human athletic performance,
and offers a means of enhancing strength and work volume training responses that are more effective than per-
formance-enhancing supplements such as anabolic steroids. There are many potential applications of vacuum-
enhanced cooling of the glabrous skin in medicine, occupational health and safety, and sport.
Key words: glabrous skin; mammalian temperature regulation; performance enhancement; regional heat transfer
Introduction
Human core body temperature is usually regulated
within a narrow range. Deviations from this range (hy-
pothermia and hyperthermia) have impacts on health, cogni-
tion, and performance, but such deviations are commonly
experienced and frequently unavoidable. For example, the ca-
pacity of the large dynamic muscles of the body to produce
heat exceeds the ability of the body to dissipate that heat
and, therefore, physical activity can cause hyperthermia.
The induction of hyperthermia is more likely in a hot environ-
ment. When environmental temperature approaches and ex-
ceeds core body temperature, hyperthermia can ensue even if
the metabolic rate remains at low resting levels. Conversely,
humans exist and work in cold environments where heat
loss from the body can exceed the capacity of the body to pro-
duce metabolic heat, resulting in hypothermia. Hypothermia
can develop even in slightly cool environments when meta-
bolic heat production is compromised such as during anes-
thesia or during sleep in the aged. Countermeasures for
hyper- and hypothermia require the ability to facilitate the
movement of heat between the body and the environment.
Such countermeasures are limited by the thermal conductiv-
ity of the body surface.
The thermal conductivity of the body surface of most
mammals is reduced by fur, which is highly adaptive in
cold environments. However, the insulation of fur poses a
problem for mammals in hot environments or during intense
exercise. The evolutionary solution to this problem has been
special vascular anatomy of the nonhairy or glabrous skin.
In most mammals, these surface areas are the pads of the
feet, the tongue, and, in some species, the ears, tails, and
Department of Biology, Stanford University, Stanford, California.
DISRUPTIVE SCIENCE AND TECHNOLOGY
Volume 1, Number 1, 2012
ªMary Ann Liebert, Inc.
DOI: 10.1089/dst.2012.0004
11
parts of the face. The unique vascular structures underlying
these glabrous skin areas are arterio-venous anastomoses
(AVAs) that can bypass the nutritive capillary beds and
shunt the arterial blood into networks of veins called retia
venosa. The potential for very high blood flow through the
retia venosa makes the overlying glabrous skin a radiator of
heat carried to those surfaces from the body core in arteries,
and since the AVAs are gated by smooth muscle, these radi-
ators can be controlled to minimize heat loss in a cold envi-
ronment and to maximize heat loss in a hot environment.
Even though humans are not heavily furred, they still pos-
sess the vascular anatomy seen in the glabrous skin of other
mammals. In this article, we will demonstrate the importance
of the controlled radiator function of the glabrous skin of hu-
mans, a technology that is used for amplifying heat exchange
through these surfaces, and some of the benefits of being able
to manipulate core body heat content by this technology.
Materials and Methods
General
The basic methods of these protocols have been previously
described in detail.
1,2
Only a brief summary of the general
methods and a detailed description of the unique specifics
of these studies are presented here. The protocols for these
studies were approved by a Stanford University Institutional
Review Board (IRB). The inclusion criteria were as follows: (1)
No known medical problems, (2) not under treatment for a
medical problem, (3) not on medication, and (4) engaged in
30 or more minutes of physical exercise for a minimum of
three times a week for the previous 6 months.
Custom-built palm cooling devices consisted of a rigid
acrylic cylindrical chamber (20 cm diameter. ·21 cm length)
into which a hand was inserted through an closed cell foam
sleeve that formed an airtight seal around the forearm. Inside
the chamber, the palm rested on a water-perfused stainless
steel heat exchanger. Cool water (15C–16C) was circulated
through the heat exchanger from a circulating refrigerated
water bath (Neslabs Model RTE 17; Thermo Scientific, Wal-
tham, MA). The chamber was connected to an adjustable
pressure relief valve, a pressure gauge, and an in-house vac-
uum system that created a pressure differential within the
chamber (40 mm Hg).
The laboratory-based experimental trials were conducted
in a 2.4 ·3.3 ·2.4 m (width, length, height) temperature-
controlled environmental chamber. The ambient conditions
inside the environmental chamber were 35.5C–0.5C, at a
relative humidity of 20%–35%. The environmental chamber
housed two treadmills (model SC7000; SciFit, Tulsa, OK).
V
O2max
tests were conducted in a 23C room on a treadmill
using a respiratory gases/metabolic analysis system (Parvo-
medics, Salt Lake City, UT).
Heart rates that were measured using portable heart rate
monitors/data loggers (model S810; Polar Electro Oy, Kem-
pele, Finland). Core temperatures (T
core
)—either esophageal
or tympanic membrane temperatures (T
es
and T
ty
)—were
measured in some subjects. T
es
was measured using a com-
mercially available general-purpose thermocouple probe
(Mon-a-Therm # 503-0028; Mallinckrodt Medical, Inc.,
St. Louis, MO) self-inserted through the nose or mouth to a
depth of 38–39 cm and held in place by a loop of surgical
tape. T
ty
was measured using a commercially available tym-
panic thermocouple probe (Mon-a-Therm # 503-0013) self-
inserted into an external auditory meatus to the point of sub-
jective discomfort and taped in place. The thermocouple
probes were connected to a laptop-based thermocouple trans-
ducer/data collection system (GEC instruments, Gainesville,
FL) that recorded temperature data at 1-sec intervals. At the
conclusion of each trial, heart rate and temperature data
were downloaded to a central spreadsheet (Microsoft Office
Excel for Windows2003; Microsoft, Inc., Redmond, WA) for
subsequent off-line analysis.
All experimental trials on individual subjects were con-
ducted at the same time of day and separated by a minimum
of 48 h.
Determining regional heat transfer
Heat transfer across the nonglabrous skin regions was
measured at five sites (the upper back, lower back, abdomen,
thigh, and upper arm) and across the glabrous skin at two
sites (the face and the palm of the hand). Urethane-coated
nylon water perfusion pads (Plas-tech, Corona, CA) were
used as the skin surface interface; a 35.6 ·21.6 cm rectangular
pad for the nonglabrous skin regions, a 12.7 ·25.4 cm rectan-
gular pad for the palmar surface of the hand, and a form-
fitting mask for the face. The larger water-perfusion pads
and face mask were backed on one side with matching
pieces of 1.3 cm-thick closed-cell neoprene foam. Mounting har-
nesses—fashioned from lengths of neoprene, elastic strapping,
and Velcro fasteners—securely abutted the water-perfusion
pads to the targeted skin surfaces. For the palm of the
hand, the subjects lightly grasped the smaller water perfusion
pad wrapped around an 8 cm-diameter ·13 cm-long neo-
prene foam cylinder. Cool water (15C–16C) was circulated
through the water-perfusion pads from a circulating refriger-
ated water bath.
Heat transfer calculations were based on the water temper-
ature differential between the inlet to and the outlet from the
water-perfusion pads and the flow rate of the water stream.
Water temperatures were measured using thermocouples
mounted in the inlet and outlet tubing lines immediately ad-
jacent to the water perfusion pads. The thermocouple junc-
tions were threaded through the male pipe thread (MPT)
openings of barbed hose T-fittings (1/4 barbed ·1/4 barbed ·
1/8 mpt) so that the tip of the thermocouple was at the center
of the throughpath. The thermocouples placement in the T-
fitting was secured by filling the MPT port with epoxy. The
barbed T-fittings were inserted into the flexible plastic tubing
lines directly adjacent to the water perfusion pads. These
thermocouples were connected to the thermocouple trans-
ducer/data collection system, and data were recorded at 1-
sec intervals. Flow rates were measured using an in-line tur-
bine flow meter with analog output (model FLR 1005; Omega
Engineering, Stamford, CT) at a 1 Hz sampling rate. Flow rate
data were stored along with the temperature data. The heat
transferred was calculated using the formula,
q=mðDTÞCp,
where q =heat transfer (cal/sec), m =mass flow (g H
2
O/sec),
DT=the change in temperature (inlet—outlet water tempera-
ture [C]), and Cp =the specific heat of water (1 cal/ g ·C).
Heat transfer was converted from calories/sec to Watts
(joules/sec) using the conversion factor 4.184 J =1cal. Baseline
12 HELLER AND GRAHN
heat loss through the perfusion pads (parasitic heat loss) was
determined for each trial by measuring the heat transfer of
the pads alone (without contact with a skin surface) for a min-
imum of 10 min immediately before and immediately after
each experimental trial. The heat transferred across the skin
surface was determined by subtracting the mean parasitic
heat transfer from each experimental heat transfer data
point. The data collection method was validated by directly
measuring the temperature differential between the inlet and
outlet water streams using a two-junction type-T thermocou-
ple circuit inserted into the inlet and outlet lines adjacent to
the single thermocouples. The voltage of the contact potential
between the two thermocouple junctions was amplified
10,000 ·(Instrumentation amplifier, model EI-1040; Elec-
tronics Innovations Corp., Lakewood, CO) and fed into a desk-
top computer, where the data were collected at a 1 Hz
sampling rate. The heat transfer was calculated as described.
The skin surface areas of the palms of the hands were de-
termined by tracing the outline of the hand on a sheet of
paper of a known weight. The outline of the hand was cut
out and weighed. The surface area was determined by the
ratio of the weight of the cutout to the weight of the sheet
of paper. Face surface area was determined in a similar man-
ner: The face of the subject was covered with a thin coating of
mineral oil, and a paper cutout of the water-perfused face
mask was placed on the face. The paper cutout was removed
from the face, flattened out, and the outlines of the oil marks
were transferred to a clean sheet of paper. The transferred oil
mark outlines were then cut out and weighed. The treatment
surface areas of the nonglabrous skin were 35.6 ·21.6 cm.
Six male and seven female volunteers (19–22 years of age)
participated in this study. Before the experimental trials, each
subject participated in order: a V
O2max
test in a thermoneutral
environment, three hot environment baseline treadmill assess-
ment trials (walking at 5.63 km/h with the slope increased by
2% increments at 3 min intervals), and three training trials.
During the training trials, the treadmill slopes were adjusted
so that the subjects would reach the temperature (39C),
heart rate (95% of maximum heart rate during the V
O2max
test), or subjective fatigue stop criteria within 30–40min of
treadmill exercise. Treadmill speed and slopes remained con-
stant for each subject throughout the experimental trials.
The daily procedure for the experimental trials was as fol-
lows: (1) attach the monitoring equipment and rest quietly for
5 min in a thermoneutral room (T
a
=22C–24C); (2) rest for
10 min in the hot environment; (3) don the heat transfer
equipment and commence the treadmill exercise; (4) walk
on the treadmill to a stop criterion; (5) rest for 10 min in the
hot environment; and (6) remove the heat transfer equipment
and monitoring equipment.
Heat transfer, heart rate, and T
core
data for each trial were
plotted against time. The heat transfer data were compiled on
a central spread sheet and sorted by skin surface area and
subject. Mean and maximum heat transfer, maximum heart
rate, and core temperatures (at 5 min intervals) were calcu-
lated for each trial and sorted by subject and treatment area.
Temperature and performance during
repeated high-intensity exercises: a field trial
We coordinated with the coaching staff of the Stanford
University Football Program to test the effects of palm cooling
on performance during a training exercise: ‘‘Getting yards’’—
a repeated high-intensity sprint drill. The six subjects were
varsity athletes selected by the coaching staff. The exercises
consisted of a series of 2-min sprints (back and forth across
a 91.44 m-long practice field) separated by 3-min rest inter-
vals. The getting yards drills were managed by members of
the coaching staff. A stopwatch was used to time the exercise
starts and stops. The exercise start and stop times were commu-
nicated to the subjects by the coach’s whistle. The final 10 sec of
each exercise bout was counted down by the coach wielding the
stopwatch, and time intervals during the rest periods were also
periodically announced (e.g., ‘‘2 min,’’ ‘‘1 min,’’ 30 sec,’’ ‘‘10 sec’’).
During the rest intervals, the subjects walked back to the start-
ing line area, received treatment, and prepared for the next
sprint. Treatment entailed placing one hand in a palm-cooling
device or rest only. The subjects participated in four trials:
paired palm cooling and rest-only trials at two ambient temper-
atures (35Cand22C, 30% relative humidity).
The distance traveled during each sprint was recorded
manually, and the data were subsequently transferred to a
spread sheet. The sprint distances were tabulated for each
trial. The data were sorted by subject and trial conditions (am-
bient temperature and treatment).
Temperature and repeated sprint exercises
in the laboratory
Twelve subjects (6 men and 6 women, 18–21 years of age)
participated in these trials. The basic protocol was as just de-
scribed in the regional heat transfer section with the following
variations: (1) the exercises were six repeated 5-min exercise/
rest cycles of 110% V
O2max
work loads on a treadmill housed
in the 35C environmental chamber; (2) treatment (palm cool-
ing or no palm cooling) was applied during the rest phases of
the exercise/rest cycles; (3) the subjects were clad in track
warm-up suits; and (4) the stop criterion for exercise was sub-
jective fatigue (i.e., the subjects hit the treadmill stop button to
end the exercise). Each subject participated in two sets of
paired trials: In one set of trials, the subjects rested in a
35C environment and in the other set of trials, the subjects
rested in a 22C environment.
Total sprint durations and mean sprint durations were cal-
culated for each trial. The rate of change of core temperature
(DT
core
) was determined by subjecting individual T
ty
and, if
available, T
es
versus time plots (from initial exercise onset
to final exercise offset) to a regression analysis (Microsoft
Office Excel for Windows 2003 data analysis tools). Perform-
ance and core temperature data were grouped by subject,
treatment, and ambient conditions. To determine the effects
of treatment on DT
core
and performance, the value of the de-
pendent variable in the palm-cooling trial was divided by the
value of the same variable in the paired control (rest only)
trial. The 24 data sets (12 subjects, 2 paired trials each) were
sorted based on the influence of the palm cooling on the
rates of the rise of core temperature and divided them into
two equal groups based on the rank order of the treatment
effect.
Statistical analysis
Descriptive statistics were calculated for the data sets, and
the data sets were subjected to analysis of variance (ANOVA)
and appropriate post-hoc tests (Microsoft Office Excel for
EFFICIENT THERMAL EXCHANGE 13
Windows2003 data analysis tools). The specifics of each sta-
tistical analysis are presented in the results section and the
figure captions.
Results
Heat loss through the glabrous skin is more variable
and can reach higher values than heat loss through
the nonglabrous skin
Figure 1 shows the data from one subject on the area-
specific heat loss from different body surfaces, glabrous and
nonglabrous, during treadmill exercise in a hot environment.
The subject rested in the 42C room for 15 min before the
onset of exercise, so he was in a state of thermal vasodilation
when exercise commenced. During the 35 min of exercise, his
core body temperature (esophageal) increased linearly by
about 2C. At the onset of exercise, the heat loss (W/cm
2
)
from the glabrous skin of the face and palm was two to
three times higher than the heat loss from the various nongla-
brous skin areas. The rates of heat loss from the nonglabrous
skin did not increase significantly during the exercise and
onset of hyperthermia, whereas heat loss from the glabrous
skin increased steadily so that by the end of exercise, the
area-specific heat loss from the glabrous skin areas was ap-
proximately five times higher than the area-specific heat
loss from the nonglabrous skin areas. Figure 2a (left panel)
shows the heat loss data from 13 subjects undergoing the
same experimental procedures. The open bars indicate heat
loss values at the beginning of exercise, and the shaded
bars are the values at the end of exercise. The initial values
for the glabrous skin areas are more than twice those for
the nonglabrous skin areas. Most nonglabrous skin areas
show no significant increase in heat loss over the exercise
bout, but the glabrous skin areas show a more than doubling
of heat loss during the exercise. The two nonglabrous areas
that show some increase during exercise are the upper back
and the thigh. Notably, these areas overlie muscles that are
actively engaged in the exercise effort.
Heat exchange across the glabrous skin can be
amplified by the application of mild negative pressure
Figure 2b (right panel) shows that the rate of heat loss
across the glabrous skin of the palm can be significantly in-
creased (33%) by the application of negative pressure. The
values shown are the mean rates of area-specific heat
loss across the 35 min bouts of treadmill exercise for the 13
subjects.
End-of-exercise heart rates reflect greater efficiency
of heat extraction from glabrous skin
Figure 3 shows the mean heart rates at the end of the
35 min exercise bouts for the 13 subjects when cooling is ap-
plied to different body surfaces. Cooling of the nonglabrous
skin areas was administered with cooling pads that were
722 cm
2
, whereas the cooling pads applied to the face aver-
aged only 288 cm
2
, and the cooling pads applied to the
palm averaged about 152 cm
2
. Exercise intensity and duration
for each subject were the same in all trials. Treatment (skin
surface regions cooled: thigh, lower back, upper arm, abdo-
men, upper back, face, palm, and palm with vacuum) had a
significant effect on maximum heart rate (two-factor
ANOVA without replication (subject [13] ·treatment [8]
p<0.05). There were no significant differences in the end-of-
exercise heart rates between the runs in which the different re-
gions of nonglabrous skin were treated (two-factor ANOVA
without replication (subject [13] ·treatment [5] p=0.20), and
there were no significant differences between the runs in
which the different regions of the glabrous skin were treated
(two-factor ANOVA without replication (subject [13] ·treat-
ment [3] p<0.26). However, end-of-exercise heart rates
trended lower for trials in which the glabrous skin of the
face or palm was treated in spite of the fact that the cooling
was applied over a smaller surface area (multiple paired
t-tests. prange 0.06–0.2). This trend was significant when
comparing the palm cooling with vacuum trials to the cooling
trials of the thigh ( p<0.01), lower back ( p<0.02), upper arm
(p<0.04), or abdomen ( p<0.03). The fact that the palm-
cooling area was only about 1/5 the cooling area on the non-
glabrous skin makes these data more impressive.
Vacuum-enhanced heat extraction from the glabrous
skin reduces rate of core temperature rise during heat
exposure and exercise and improves performance
An experiment to address whether heat extraction in a sit-
uation of repeated bouts of maximal metabolic effort could
improve performance was carried out on the Stanford foot-
ball team engaged in a practice routine called ‘‘getting
yards.’’ In this routine, the subjects started at a goal line
and did repeated 2 min sprints up and down the field.
Between the sprints, there was a 3 min interval during
which the subjects returned to the goal line and rested. We
inserted episodes of palm cooling that averaged no more
0
0.05
0.1
0.15
0.2
0.25
0102030
Time (min)
Heat Transfer
(w/sq cm)
T
LB
UA
UB
AB
F
P
36.5
38.5
Tes (C)
FIG. 1. Local heat loss across various regions of the body
surface during fixed-load treadmill exercise in a 42C envi-
ronment: Individual subject data. Top panel: T
es
versus
time. Lower panel: heat transfer/unit surface area versus
time. Closed symbols: glabrous skin regions. Open symbols:
nonglabrous skin regions. T, thigh; LB, lower back; UA,
upper arm; AB, abdomen; UB, upper back; F, face; P, palm.
14 HELLER AND GRAHN
than 1.5 min during those rests. These experiments were car-
ried out on hot days (T
a
averaged 35C) and on cooler days
(T
a
averaged 22C). The dependent variable was the total dis-
tance gained during the sprints (Fig. 4). The total yards
gained for each successive sprint decreased under all condi-
tions. On the warm days, the distance gained in the first
sprints were lower than they were on the cool days, and in
the runs without palm cooling, the rate of drop off of distance
in the successive sprints was greater than in the cool days or
on the hot days with palm cooling. There was no effect of
palm cooling on the cool days. The effect of palm cooing on
the hot days was to decrease the drop off in the distance
gained so that the performance was similar to the perfor-
mance on the cool days. As a result of the effect of palm cool-
ing on the hot days, the distance gained in the last three
sprints and in the overall distance gained was significantly
greater with palm cooling than without it. Thus, the experi-
ment showed that the applications of palm cooling in situa-
tions where heat stress is an issue can lead to improved
performance. However, these field experiments did not
make the connection between the treatment, core tempera-
ture, and performance. We, therefore, designed a laboratory
experiment to replicate the conditions of the ‘‘getting yards’’
routine.
In this laboratory experiment on the effects of palm cooling
on the rate of rise of core temperature (T
ty
and/or T
es
) and
performance, 12 subjects engaged in six sprints to subjective
fatigue at 5 min intervals on a treadmill set to a speed that re-
quired an effort of 110% VO
2max
. The inter-sprint rest condi-
tions in these experiments were a cool ambient (22C) and
a warm ambient (35C) with and without palm cooling.
Figure 5 shows the effects of palm cooling on the rates of
the rise of esophageal and tympanic temperatures from a sin-
gle subject performing the sprint routine in the hot environ-
ment. Clearly, the palm cooling had an effect. The effects of
treatment on the two-core temperature measures (T
es
and
T
ty
) were indistinguishable. Figure 6a (left panel) shows the
core temperature (T
ty
) data from all 12 subjects in all 4 rest
conditions and also shows the performance measures. The
left four bars show the effects of the thermal environment
and palm cooling on performance (summed sprint dura-
tions). The first bar is the optimal situation of a cool environ-
ment with palm cooling, whereas the fourth bar is the worse
0
0.1
0.2
0.3
Non-glabrous
Heat Transfer
(w/sq cm)
exercise onset
exercise offset
0
0.05
0.1
0.15
0.2
TLBUAABUBF P W/O W
Pressure Differential
Mean Heat Transfer
(w/sq cm)
***
**
**
Glab.
FIG. 2. Local heat transfer across various regions of the body surface at the onset and offset of 35 min of fixed load exercise in
a42C environment: Group data (mean –s. d., n=14). Left panel: heat transfer/unit surface area. A two-way ANOVA with
treatment (skin regions) and time (exercise onset and offset) as the variables and time as the repeated measure determined
significant treatment and time effects ( p<0.0001). *F and P onset values significantly different from other regions ( p<0.01,
paired t-tests) and offset values different from onset values ( p<0.0001). **UB, T, and UA exercise offset values different
from exercise onset values ( p<0.01, paired t-tests). Abbreviations are as in Figure 1. Right panel: the effect of applying a
40 mm Hg pressure differential to the hand on heat transfer (n=24, p<0.02, paired t-test). ANOVA, analysis of variance;
s. d., standard deviation; W, with; W/O, without.
160
170
180
190
T
LB
UA
AB
UB
F
P
WV
Body Re
g
ion
Max. HR (bpm)
*
FIG. 3. The effect of regional cooling on maximum heart
rate during fixed-load exercise in a 42C environment.
Open bars: cooling of nonglabrous skin regions. Closed
bars: cooling of glabrous skin regions. Abbreviations as in
Figure 1 and WV, Palm w/vac (40 mmHg pressure differen-
tial applied to the hand along with cooling). A one-way
ANOVA determined a significant treatment (skin region) ef-
fect ( p<0.01).*Different from T, LB, UA, and AB ( p<0.05,
paired t-tests).
EFFICIENT THERMAL EXCHANGE 15
situation of a hot environment with no palm heat extraction.
In spite of a great deal of variability in the responses of the
subjects, these two sets of data are significantly different
(p=0.0004, paired t-test). The two center bars (not signifi-
cantly different) represent trials in which there was either
an environmental cooling factor or a palm cooling factor.
The performances in these two groups of trials were interme-
diate between the two sets of runs at extreme conditions.
The four bars at the right of Figure 6a represent the effects
of the environment and treatment on core body temperature.
Under conditions of cool ambient and palm cooling, the rate
of the rise of core temperature was the least, and under con-
ditions of a warm environment without palm cooling, the rate
of core temperature rise was the greatest ( p=0.02). Intermedi-
ate values were seen in the cool environment without palm
cooling and in the warm environment with palm cooling.
Thus, in terms of effects on the performance and effects on
the rate of rise of core temperature, the application of palm
cooling was the equivalent of a 13C drop in the ambient tem-
perature.
The high variance in these data sets led us to look at how
different subjects were responding to the palm-cooling treat-
ments. We sorted the 24-paired data sets (12 subjects, 2 paired
trials each) based on the influence of the palm cooling on the
rates of the rise of core temperature and divided them into
two equal groups based on the rank order of the treatment ef-
fect (12 highest vs. 12 lowest). If there was no effect of palm
cooling, then the ratio of the rates of the rise of core temper-
ature for cooling and control runs was 1, and if there was
an effect of palm cooling, then this ratio was less than 1.
For the two groups, we then plotted the ratios of the core tem-
perature effects against the ratios for the performance differ-
ences (Fig. 6 right panel). The plot clearly shows that the
subjects with the greatest effects of palm cooling on the rate
of the rise of core temperature showed no performance bene-
fits. However, the subjects with no apparent effect of palm
cooling on the rate of the rise of core temperature showed
large performance improvements. Our conclusion is that
the two groups represented pacers and pushers. The pacers
maintained the same performance level, but suffered a lesser
rate of rise in core temperature. The pushers converted the
cooling effect into a performance gain and, therefore, showed
the same rate of core temperature rise in the experimental and
control conditions.
Discussion
The realization of the importance of the vascular structures
underlying the glabrous skin for the dissipation of excess heat
from the body explains why various models of human heat
exchange have been more successful in modeling thermoreg-
ulation in a cold environment than in a hot environment.
3,4
Most models have not included the hands and feet, and,
therefore, failed to take into account the capacity of the gla-
brous skin for enhancing heat loss. This omission would be
less serious for modeling responses to cold environments,
as under those conditions, the AVAs would be closed, and
the heat loss from the hands and feet would be minimal.
The rationale for not including hands and feet in human
heat exchange models was that these appendages were diffi-
cult to describe geometrically and they comprised only a
small surface area. In fact, some experiments that tested
400
500
600
123456
Sprint Cycle
Distance (m)
35 PC
35 R
22 PC
22 R
**
FIG. 4. Field trial: Getting Yards drill (repeated 2 min
sprints with a 3 min rest between exercises) under 2 ambient
conditions and with and without palm cooling. Distance trav-
eled in a 2 min sprint versus sprint number (mean –s. d.,
n=6). Conditions: 22 PCT
a
=22C with palm cooling, 22
RT
a
=22C with rest only, 35 PCT
a
=35C with palm cool-
ing, and 35 RT
a
=35C with rest only. A two-way ANOVA
with repeated measures for treatments (4 combinations of palm
cooling and ambient temperature) and sprint cycles (six) deter-
mined a significant treatment and sprint cycle effect ( p<0.0001).
*35 R different from other conditions ( p<0.05, paired t-tests). T
a
,
ambient temperature; PC, palm cooling; R, rest only.
36
37
38
Tes (C)
PC
R
36
37
38
Tty (C)
100
150
200
0102030
Time (min)
HR (bpm)
FIG. 5. Core temperatures (T
es
and T
ty
) and heart rate dur-
ing repeated 5 min high-intensity work/rest cycles in a 35C
environment: Individual subject. Open symbols: rest only be-
tween exercises. Closed symbols: palm cooling during rest.
HR, heart rate.
16 HELLER AND GRAHN
skin blood flow placed tourniquets on the wrists and ankles
to eliminate a potential source of error,
5
but in reality, they
eliminated a highly significant thermoregulatory effector
mechanism. Understanding the thermoregulatory role played
by the glabrous skin will make it possible to improve human
thermoregulatory models and their use in the design of ther-
mally protective gear.
Exploiting the thermoregulatory functions of the glabrous
skin makes it possible to advance the understanding of the ef-
fects of core body temperature on various aspects of human
health and performance and to intervene in those effects.
For example, we have shown how vacuum-enhanced heat
extraction from only one hand can decrease the rate of rise
of core body temperature during aerobic exercise and
thereby extend endurance by a large amount.
1
Similarly, the
same treatment can be used in heat-stressed subjects to dra-
matically speed recovery.
6
Every year, young athletes suffer
from heat illness and even occasional deaths due to exercise-
induced hyperthermia during summer and fall practices and
competition. Such tragedies have also occurred in professional
athletics. The fact that heat stroke is one of the leading causes
of sudden death in sport stimulated the founding of the Korey
Stringer Institute at the University of Connecticut dedicated to
reducing the occurrences and consequences of heat illness in
athletes. Even though whole-body immersion in cold water
is still the gold standard for treatment of hyperthermia, that
treatment option is not always immediately available. When
a cold water bath is not available or until it is available, vacu-
um-enhanced heat extraction from a palm is a safe and efficient
alternative or an additional means of avoiding morbidity and
mortality due to hyperthermia.
Two studies have claimed no effect of palm cooling on per-
formance.
7,8
In both cases, the experimental protocol pre-
cluded positive results, but for opposite reasons. The
Amorim et al. study involved a level of thermal stress that
far exceeded the heat loss capacity of a single palmar surface.
7
The subjects walked on a treadmill (5.4–6.7 Km$h
1
,0%to4%
grade) in a 42C environment wearing military battledress
uniforms, body armor, and a weighted backpack. The work-
load was designed to be at 50% of the subjects VO
2max
. and
averaged 537 W. Given the level of heat production and the
insulation of the uniform plus body armor, the authors con-
clude that the maximal possible heat extraction from a single
palmar surface was insufficient to have a substantial effect on
the rate of the rise of core body temperature. A similar finding
was reported by Grahn et al.
1
in a comparison of the effective-
ness of palmar cooling at different workloads. That study
demonstrated that the effect of cooling a single palm on exer-
cise endurance was exponentially related to workload.
Endurance at a workload that produced fatigue in controls
at only 15 to 20 min. was minimally improved by palmar
cooling, but endurance at a workload that produced fatigue
in 60 min. in controls was doubled by palmar cooling. Both
these studies applied the cooling device to only one palmar
surface. It should be kept in mind that the effects of cooling
multiple glabrous skin areas is additive.
6
The Walker et al. study
8
purported to test whether palmar
cooling ‘‘.during active rest periods of multiple set training
is an effective means to increase performance.’’ The protocol
was for the subjects to do 8, 30 s, self-paced treadmill sprints
interspersed with 1.5 min walk/jog rests. During the rests,
they used the palm-cooling device, rested a hand on a gel
pack at the same temperature, or rested the hand on a shelf.
The data showed no significant drop off of the distances cov-
ered in the sprints under any of the test conditions. The only
claim made for acute use of palmar cooling is that it mitigates
heat-related impairment of performance. Therefore, since
there was no impairment of performance, it was impossible
to show a mitigation of impairment with palmar cooling.
Vacuum-enhanced heat exchange across the palm has also
been used in reverse to insert heat into the core of hypother-
mic individuals recovering from anesthesia.
9
The standard
0
0.02
0.04
0.06
0.08
22
PC
22
R
35
PC
35
R
Ta - Treatment
Delta Tcore (C/min)
7
9
11
13
22
PC
22
R
35
PC
35
R
Performance (min)
0.5
1
1.5
0.9 1.3
Performance
Tcore
Statistical analysis: p values
Time Tcore
ANOVA 0.005 0.008
22 PC vs. 22 R 0.07 0.15
22 PC vs. 35 PC 0.04 0.03
22 PC vs. 35 R 0.0004 0.02
22 R vs. 35 PC 0.04 0.02
22 R vs. 35 R 0.69 0.62
35 PC vs. 35 R 0.06 0.07
FIG. 6. The effects of the rest environment and palm cooling during rest between sets of high-intensity sprint exercises
(T
a
=35C) (mean –s. d., n=12). Rest conditions: 22 PCT
a
=22C with palm cooling, 22 RT
a
=22C with rest only, 35
PCT
a
=35C with palm cooling, and 35 RT
a
=35C with rest only. Left panel: performance (summed sprint times) and
rates of change of T
ty
versus rest conditions. Upper right; effects of palm cooling on core temperature versus effects on perfor-
mance: value during cooling trial/value during control trial. Lower right: Statistical analysis for data graphed in left panels (A
one-way ANOVA and paired t-tests).
EFFICIENT THERMAL EXCHANGE 17
methods of body surface warming used in the recovery room
to raise the core temperatures of patients to the normal range
and stop tremors took hours. In fact, one study of a method
of forced warming that is still a standard of care showed
that it had no significant effect compared with blankets on
speeding the rise of core temperature.
10
In contrast, the use
of a vacuum-enhanced palm-heating device reduced the
time for core temperature rewarming and tremor cessation
to 10 min or less. The ability to insert significant amounts of
heat into the body core through the glabrous skin opens up
the possibility of designing devices that would improve the
functionality and survival of humans working in extremely
cold environments.
Experiments on palm heat extraction from subjects engaged
in high levels of large dynamic muscle exercise led to the very
interesting observation that muscle fatigue may be largely due
to the rise in the temperature of the muscle.
2
The work capacity
of muscles was greatly enhanced by palm cooling, and when
this benefit was applied to strength-conditioning regimes, the
rates of gain were dramatic, exceeding what has been ob-
served through the use of anabolic steroids.
2,11
Given the
facts that steroid use is prohibited by most collegiate and pro-
fessional sports associations, and that steroid use carries very
significant health risks, the option of a safe alternative for
those who wish to improve their strength and work capacity
could greatly decrease the use of performance-enhancing drugs
and supplements.
Knowledge of the thermoregulatory functions of the AVAs
and retia venosa has helped resolve some physiological quan-
dries. In the 1960s, experiments were conducted on humans
at rest in a hot environment for the purpose of understanding
the effects of heat on cardiovascular performance.
12
The car-
diac output was seen to increase by more than a factor of 2,
from 6 to 14 L/min. Where did this blood go? The researchers
found that blood flow to the internal organs and skeletal mus-
cles actually decreased. The only large rise in blood flow, they
observed, was in the forearm. They concluded that the in-
creased cardiac output had to be accommodated by a large in-
crease in cutaneous blood flow due to thermoregulatory
vasodilation. However, increases in blood flow through non-
glabrous skin have since been demonstrated to be only mod-
est.
13
It did not occur to the investigators that the great
increase in forearm blood flow and in cardiac output was
due to the opening of the AVAs and the subsequent increased
flow through the retia venosa in the glabrous skin.
The fact that heat extraction can mitigate muscle fatigue in
both endurance events and anaerobic metabolism events in-
dicates that heat per se is a cause of muscle fatigue.
2
What
could be the mechanism for this dramatic effect of tempera-
ture on muscle function? In vitro studies of rabbit skeletal
muscle have shown that pyruvate kinase, a critical enzyme
for the flow of substrate into oxidative phosphorylation, is
highly temperature sensitive.
14–16
Its activity is greatest
around 35C–38C and drops off rapidly as the temperature
rises due to conformational changes. If the rise in muscle tem-
perature during exercise compromises the activity of pyru-
vate kinase, then the flow of the substrate into the Krebs
cycle and, therefore, the rate of production of ATP will be re-
duced. Continued glycolysis, however, will result in a build
up of lactate. These considerations of the possible role of
the temperature sensitivity of pyruvate kinase in muscle fa-
tigue cast a new light on the role of lactate in fatigue. Lactate
accumulation may be a consequence of muscle fatigue rather
than a cause.
There are many medical conditions that could potentially
be influenced by a technology that rapidly and efficiently in-
fluences the heat balance of the human body. One that we
have already demonstrated is the mitigation of heat exacerba-
tion of symptoms in multiple sclerosis. The vast majority of
MS patients are highly temperature sensitive such that a
small increase in ambient temperature or body temperature
can incapacitate them. Being able to efficiently counter an ex-
ercise or an environmentally induced rise in core temperature
can prevent the exacerbation of symptoms and preserve func-
tionality.
17
Examples of other conditions that are targets for
investigation of the effects of this heat exchange technology
are as follows: induction of hyperthermia to augment
chemo- or radiation therapy for cancer, maintenance of core
body temperature perioperatively, peripheral neuropathy,
and quality of life issues such as menopausal hot flashes
and insomnia. This simple new tool can have far-reaching
impacts.
Disruptive Science and Technology
The temperature of the body core is a crucial factor in
health and disease. However, it is difficult to manipulate
body core temperature efficiently due to the large heat capac-
ity of the body and the thermal conductance of the body
surface. Our work has revealed the thermoregulatory signifi-
cance of a general mammalian vascular feature of the non-
hairy skin—AVAs and retia venosa—that has previously
been ignored in models of human thermal exchange. Incorpo-
rating these thermoregulatory effectors into such models will
improve their accuracy and usefulness in the design of protec-
tive thermal gear. We have developed a technology that
amplifies the heat exchange capacity of these structures.
Experiments using that technology have demonstrated its
disruptive potential in several areas. In medicine, it chal-
lenges the current standard of care for perioperative temper-
ature management. In sport and fitness, it challenges the
dogma of the cause of fatigue and introduces a means of
improving performance that far exceeds the effects of the
use of anabolic steroids and other performance enhancers.
In occupational health and safety, it can lead to the develop-
ment of devices that can reduce the incidence of heat illness
due to working in hot environments, in fire-fighting, or in
chemical/biological protective gear. Many other applications
in medicine and in emergency medicine are possible.
Barriers to Be Overcome
Human thermal exchanges have been studied and mod-
eled for more than 50 years without consideration of the spe-
cial adaptations of the glabrous skin for heat loss. Thus, there
is an enormous amount of literature and practice that should
be re-examined, so resistance to these new views is expected.
Practical applications of technologies that enhance heat ex-
change across the glabrous skin have been demonstrated,
but technological challenges exist for the design of devices
for optimal effectiveness in different applications. For exam-
ple, wearable (rather than episodic use) devices that have a
minimal impact on mobility or manual dexterity are desirable
for applications such as firefighting and construction or in-
dustrial work. Clinical trials will be necessary to show the
18 HELLER AND GRAHN
effectiveness of the technology in potential medical applica-
tions such as cancer therapy, peripheral neuropathy, and
perioperative temperature management. Large-scale epide-
miological studies will be necessary to support the use of
vacuum-cooling devices for the immediate treatment of
heat illness in venues such as high-school sports programs.
None of these barriers are insurmountable, but they will
require considerable resources.
Acknowledgments
Vinh Cao was instrumental in the execution of these proto-
cols. This project was supported in part by a grant from the
U.S. Department of Defense Advanced Research Projects
Agency (DARPA).
Author Disclosure Statement
Patents have been issued for the vacuum-enhanced heat
transfer technology discussed in this article (D. Grahn and
H.C. Heller [Inventors]; Stanford University [Assignee]),
and Stanford University has entered into a licensing agree-
ment with AVAcore Technologies, Inc., for the commerciali-
zation of the technology. Included in the license is a royalty
agreement that grants Stanford University a percentage of
the net sales of the technology, which will be shared by the
University and the inventors. D. Grahn and H.C. Heller are
founders of AVAcore Technologies but receive no ongoing
compensation from the company, and AVAcore Technologies
provided no financial support for the research. No AVAcore
equipment was used in this research. To assure that potential
conflicts of interest did not influence the outcome of the re-
search, Stanford University required that Grahn and Heller
had no participation in the recruitment of subjects, the con-
duct of the experimental trials, or the initial analysis of the
data in studies that employed the technology.
References
1. Grahn DA, Cao VH, HC Heller, Heat extraction through the
palm of one hand improves aerobic exercise endurance in a
hot environment. J Appl Physiol. 2005;99:972–978.
2. Grahn DA, Cao VH, Nguyen CM, et al. Work volume and
strength training responses to resistive exercise improve
with periodic heat extraction from the palm. J Strength
Cond Res. 2011 [Epub ahead of print].
3. Ruxton GD, Wilkinson DM. Thermoregulation and endur-
ance running in extinct hominins: Wheeler’s models revis-
ited. J Hum Evol. 2011;61:169–175.
4. van Beek JH, Supandi F, Gavai AK, et al. Simulating the
physiology of athletes during endurance sports events: Mod-
elling human energy conversion and metabolism. Philos
Trans A Math Phys Eng Sci. 2011;369:4295–4315.
5. Brothers RM, Wingo JE, Hubing KA, et al. Methodological
assessment of skin and limb blood flows in the human fore-
arm during thermal and baroreceptor provocations. J Appl
Physiol. 2010;109:895–900.
6. Grahn, DA, Dillon JL, Heller HC. Heat loss through the gla-
brous skin surfaces of heavily insulated, heat-stressed indi-
viduals. J Biomech Eng. 2009;131:071005.
7. Amorim FT, Yamada PM, Robergs RA, et al. Palm cooling
does not reduce heat strain during exercise in a hot, dry en-
vironment. Appl Physiol Nutr Metab. 2010;35:1–10.
8. Walker TB, Zupan MF, McGregor N, et al. Is performance of
intermittent intense exercise enhanced by use of a commer-
cial palm cooling device? J Strength Cond Res. 2009;23:
2666–2672.
9. Grahn D, Brock-Utne JG, Watenpaugh DE, et al. Recovery
from mild hypothermia can be accelerated by mechanically
distending blood vessels in the hand. J Appl Physiol.
1998;85:1643–1648.
10. Ereth MH, Lennon RL, Sessler DI. Limited heat transfer
between thermal compartments during rewarming in vaso-
constricted patients. Aviat Space Environ Med. 1992;63:
1065–1069.
11. Kwon YS, Robergs RA, Kravits LR, et al. Palm cooling delays
fatigue during high-intensity bench press exercise. Med Sci
Sports Exerc. 2010;42:1557–1565.
12. Rowell LB. Human cardiovascular adjustments to exercise
and thermal stress. Physiol Rev. 1974;54:75–159.
13. Yamazaki F. Vasomotor responses in glabrous and nongla-
brous skin during sinusoidal exercise. Med Sci Sports Exerc.
2002;34:767–772.
14. Herman P, Lee JC. Functional energetic landscape in the allo-
steric regulation of muscle pyruvate kinase. 1. Calorimetric
study. Biochemistry. 2009;48:9448–9455.
15. Herman P, Lee JC. Functional energetic landscape in the allo-
steric regulation of muscle pyruvate kinase. 2. Fluorescence
study. Biochemistry. 2009;48:9456–9465.
16. Herman P, Lee JC. Functional energetic landscape in the allo-
steric regulation of muscle pyruvate kinase. 3. Mechanism.
Biochemistry. 2009;48:9466–9470.
17. Grahn DA, Murray JV, Heller HC. Cooling via one hand im-
proves physical performance in heat-sensitive individuals
with multiple sclerosis: a preliminary study. BMC Neurol.
2008;8:14.
Address correspondence to:
H. Craig Heller
Department of biology
Stanford University
371 Serra Mall
Stanford, CA 94305-5020
E-mail: hcheller@stanford.edu
EFFICIENT THERMAL EXCHANGE 19
