beatadje

http://article.pubs.nrc-cnrc.gc.ca/ppv/RPViewDoc?_handler_=HandleInitialGet&journal=cjpp&volume=76&calyLang=eng&articleFile=y98-046.pdf

Immune changes induced by exercise in an
adverse environment1
Roy J. Shephard


Abstract: Both physical activity and exposure to environmental
stressors such as cold, heat, and high altitudes modify
various components of immune function: T cell counts, natural killer
(NK) cell counts, and cytolytic activity, cytokine
secretion, lymphocyte proliferation and immunoglobulin levels. Light
physical activity or a moderate level of
environmental stress stimulate the immune response, but exhausting
physical activity or more severe environmental
stress have a suppressant effect, manifested by a temporary increase
in susceptibility to viral infections. Combinations
of physical activity and environmental stress generally have at least
an additive effect. Thus, an intensity of physical
activity or of environmental stress that is beneficial in itself can
readily cause immunosuppression if the body is
challenged by the two stimuli simultaneously.
Key words: cold exposure, environmental stress, heat exposure, high
altitudes, immunosuppression.

Shephard 546
The influence of exercise upon the immune system is now
well established (Shephard 1997). Moderate physical activity
(for example, 1 h of endurance exercise at 50% of maximal
oxygen intake) enhances immune responses. In contrast,
exhausting physical activity (for example, 2 h of concentric
exercise at 65–75% of maximal oxygen intake, or shorter
periods of more intensive or eccentric effort) lead to a suppression
of immune function that persists for 2–24 h.
Less information exists regarding immune responses
when moderate or severe physical activity is performed in a
challenging environment. The physical characteristics of the
environment may influence immune function in their own
right, and indeed for some authors, the changes induced by
vigorous exercise are but a specific example of a more generalized
stress response (Pedersen et al. 1994). Immune responses
may also be modified by the subject's cognitive
appraisal of a given environment. But despite the potential
contribution of neuropsychological influences (LaPerriere
et al. 1994), the psychological effects of heat, cold, or high
altitude exposure are relatively small once subjects have become
habituated to the environmental challenge. This article
thus focuses on the physiological rather than the psychological
component of immune responses to environmental
stressors, both at rest and during vigorous exercise.
The average adult is not exposed to environmental stress
very frequently in the climate-controlled cities of North
America. In contrast, athletes may undertake vigorous exercise
in a very hostile environment, thus exacerbating normal
immune responses to physical activity and muscle injury.
Cold, dry air causes bronchospasm. It also depresses tracheal
ciliary function, and increases the viscosity of tracheal
mucus. Further, whole-body cooling may facilitate the proliferation
of some microorganisms. All of these factors tend
to increase a person's susceptibility to upper respiratory infections.
Leukocyte demargination and cell trafficking (Kappel
et al. 1991c; McCarthy and Dale 1988) are also affected by
cold-induced secretion of stress hormones. Increased norepinephrine
levels boost the circulating counts of both the first
line of viral defence (non-major histocompatibility complex
(MHC), natural killer (NK) cells, CD3–CD16+CD56+) and
MHC-restricted T cell counts. In contrast, an increase in
cortisol levels causes a migration of lymphocytes out of the
circulation, also depressing both the T cell – cytokine cascade
Can. J. Physiol. Pharmacol. 76: 539–546 (1998) © 1998 NRC Canada
Received August 1, 1997.
R.J. Shephard. Faculty of Physical Education and Health,
320 Huron St., Toronto, ON M5S 1A1, Canada
and Defence and Civil Institute of Environmental Medicine,
North York, and Health Studies Programme, Brock
University, St. Catharines, Ontario (correspondence to:
P.O. Box 521, Brackendale, BC V0N 1H0) (e-mail:
roy.shephard@mountain-inter.net).
1This paper has undergone the Journal's usual peer review.
and NK cell cytolytic function (for references, see Shephard
1997).

Despite the substantial increase of metabolism associated
with exercise, physical activity does not necessarily reduce
the severity of cold stress relative to sedentary rest. Perhaps
for this reason, most authors have found little difference of
immune responses between subjects who are passively exposed
to cold and those who undertake moderate exercise
under equivalent environmental conditions.
Overall immune responses
Local inflammatory reactions are enhanced by sustained
cold exposure. For instance, if mice are kept at 5°C, their
contact sensitivity reactions to 2-4-dinitro-1-fluorobenzene
are increased (Blecha et al. 1982). Prolonged cold exposure
also increases susceptibility to both anthrax and poliomyelitis
(for references, see Shephard 1997), although anecdotal
reports of enhanced vulnerability to upper respiratory tract
viruses are less well substantiated.

Animal studies
Responses of the immune system to cold exposure depend
on the severity of the environmental challenge. A very brief
exposure (3 min of swimming in cold water on each of
5 days) had a stimulatory effect on the immune system of
the rat, increasing CD4+/CD8+ ratios, NK cell cytolytic activity,
interleukin (IL)-2 production, and lymphocyte proliferative
responses to concanavalin A (ConA) and lipopolysaccharide
(LPS) (Shu et al. 1993). However, when the cold exposure
was increased to three 3-min exposures on a single day,
CD4+ and CD8+ percentages, IL-2 production, and lymphocyte
proliferative responses were all decreased.
Other investigators have noted that cold exposure is followed
by a variety of changes that could increase susceptibility
to infection, including a reduction in NK cell numbers
and cytotoxic activity (Aarstad et al. 1983), and a reduced
proliferative response to both LPS and ConA (Regnier and
Kelley 1981).

Human studies
A number of human studies have reported a
catecholamine-related enhancement of immune function
with moderate cold exposure.
Lackovic et al. (1988) exposed subjects to 4°C air for
30 min. This stimulus led to a release of noradrenaline and
(presumably because of a release into the circulation of
marginated NK cells) NK cell cytolytic activity was increased.
Jansky et al. (1996) exposed healthy men to 14°C
water for 1 h three times per week for 6 weeks. This stimulus
induced a considerable increase in plasma catecholamine
levels, with associated increases in leukocyte count. On day
1, changes of immune function were minimal, but as exposure
was repeated, the immune system showed signs of progressive
activation: increases in the proportion of circulating
monocytes and CD25+ lymphocytes, increased counts of
CD3+, CD4+, and CD8+ cells, an increased proportion of activated
T and B lymphocytes (HLA–DR+), and increased
plasma levels of the proinflammatory cytokines IL-6 and tumor
necrosis factor a (TNF)-a.
If exposure to cold air precipitates bronchospasm, levels
of the immunoglobulin associated with anaphylactic reactions
(IgE) may also be increased (Parker 1991).
In contrast with these reports, Cross et al. (1996) found
that the circulating leukocyte count, granulocyte count, and
monocyte count were all negatively correlated with rectal
temperatures when subjects sat in 23°C water for 40 min, a
time sufficient to reduce core temperature to about 36.3°C.
Personal fitness influences acute responses to cold in several
differing ways. Subcutaneous fat is important to insulation,
but well-developed muscles can also offer thermal
protection. Moreover, subjects who are physically fit are
better able to sustain shivering, because they have greater
initial glycogen reserves. It is as yet unclear how interindividual
differences in fitness and cold acclimatization affect
immune responses in a cold environment.
A distinction must be drawn between responses to a fever
(where endogenous pyrogens such as IL-1, IL-6, interferon
(IFN), and TNF increase the temperature set-point of the
body, Chang 1993; Kappel et al. 1991c) and passive heat exposure
or vigorous physical activity (where the set-point remains
normal, but body temperature rises because of
problems in heat dissipation).
Depending on ambient conditions, the core temperature of
runners may increase by 2–4°C over as little as 30 min of
running (Simon 1993). Factors that modify white cell counts
during an episode of hyperthermia include increases in cardiac
output and plasma catecholamine levels (both of which
cause cell demargination; Kappel et al. 1991b; McCarthy
and Dale 1988), and the secretion of cortisol (which induces
a migration of neutrophils out of bone marrow and into the
tissues; Kappel et al. 1991b). Parallels can be drawn between
certain manifestations of severe heat stress (for example,
hepatic and renal failure) and the multiple organ distress
syndrome of severe sepsis, although mechanisms probably
differ.


Heat sensitivity of microorganisms
The growth and replication of certain bacteria, viruses,
and fungi are impaired by high local temperatures
(Mackowiak 1981). One defence mechanism of poikilothermic
animals is thus a relocation to a warm environment
following infection, and homiotherms can increase their
body temperatures by production of IL-1 and other pyrogens
during infection or inflammation (Downing and Taylor 1987).
When assessing the influence of heat on susceptibility to
infection, account must thus be taken not only of any changes
of in vitro immune responses, but also of temperatureinduced
changes in vulnerability of the target organisms.
Whole-body responses to heat exposure
A single 60 min of heating to a core temperature of 37°C
decreases the delayed hypersensitivity reaction of mice to
percutaneous 1-chloro-2-4-dinitrobenzene. In contrast,
hypersensitivity
reactions are enhanced by repeated or sustained
heat exposure. Thus, the caging of animals at 35°C
increases both contact reactions to 2,4-dinitro-fluorobenzene
© 1998 NRC Canada
540 Can. J. Physiol. Pharmacol. Vol. 76, 1998
© 1998 NRC Canada
Shephard 541
and delayed reactions to injected sheep red cells (Blecha
et al. 1982).
In humans, there are claims that regular sauna bathing enhances
an individual's overall resistance to upper respiratory
tract infections (Ernst et al. 1990). In addition to the immediate
deleterious effects of heat on invading microorganisms
(above), regular bouts of moderate heat exposure could have
a priming effect on the immune system, so that it responds
more readily to a viral threat. Nevertheless, it would be surprising
if major beneficial effects were induced by a single
sauna session per week. Other possible explanations of the
observed gains from sauna treatment include a Hawthorne
effect, a reduction in stress-mediated inhibition of immune
function (LaPerriere et al. 1994), or an improved drainage of
nasal sinuses.


Cellular responses
Lambs show a progressive leukocytosis if the environmental
temperature is held at 35°C for 24 h (Minton and
Blecha 1990).
A brief, moderate increase of core temperature has little
effect on leukocyte numbers in human subjects. Cross et al.
(1996) immersed young men to mid-chest for 80 min. When
they were in water at 39°C, the rectal temperature rose to
near 38°C, but little increase of white cell count was seen.
Likewise, Severs et al. (1996) reported little increase of leukocyte
numbers with 3 h of seated rest at an air temperature
of 40°C, 30% relative humidity.
Nevertheless, like animals, humans do develop a leukocytosis
in response to more severe heat stress. Presumably,
this reflects cell trafficking in response to an increased cardiac
output and the secretion of stress hormones. The effect
of an increase in cardiac output seems dominant, since
Stephenson et al. (1985) noted that 30 min of exposure to an
80°C sauna gave a 16% leukocytosis in dehydrated subjects,
but the magnitude of the response increased to 40% if subjects
were rehydrated.


Subset responses
Most of any heat-induced leukocytosis is attributable to
neutrophils. Hyperthermia also increases the proportion of
immature neutrophils in the circulation. Liburdy (1980) suggested
that during body heating, lymphocytes migrated from
the lungs to the spleen, liver, and bone marrow. Changes in
the relative proportions of the other leukocyte subsets have
varied from one study to another. This may reflect either
technical problems (counting technology, the timing of
blood samples, and allowances made for changes in plasma
volume), or inter-trial differences in severity of the heat
stress.
The immediate response to body heating is usually a
decrease in CD3+ and CD4+ counts, a decrease in the
CD4+/CD8+ ratio, and an increase in CD8+ and NK cell
counts. An increase of CD14+ count 2 h following hyperthermia
may cause a late boosting of immune function
(Pedersen et al. 1994).
Repeated heat exposure upregulates immune function, increasing
the resting percentage of CD4+ cells (Kiecolt-
Glaser et al. 1986).


Lymphocyte proliferation
A moderate, passive increase of core temperature (0.7°C
for 3 h) has no effect on the proliferative response of
PBMCs to various mitogens (Severs et al. 1996). A larger
increase of rectal temperature initially increases proliferative
responses, but if the exposure is prolonged or severe, proliferation
is decreased (Downing and Taylor 1987; Regnier and
Kelley 1981; Szmigielski et al. 1991).
NK cells
Heat exposure may either increase or decrease NK cell activity,
depending on the magnitude of the increase in temperature,
the duration of exposure to the hot environment, and
the intensity and duration of any associated physical activity.
In vitro studies
When compared with other cytolytic cells, NK cells seem
particularly sensitive to high incubation temperatures. Thus
IL-2 enhanced cytotoxic activity (a figure that reflects
largely the response of LAK and CD8+ cells) is unaffected
by 3 h of incubation at 42.5°C (Spagnoli et al. 1983), but a
treatment of this severity greatly reduces NK cell cytolytic
function (Dinarello et al. 1986).

In vivo studies
In vivo studies have shown disparate responses (constant,
increased, or decreased NK and LAK cell activity) as body
temperatures have been increased. In addition to inter-trial
differences in the average increase of core temperature and
in the times of blood sampling, some data have been collected
on healthy young adults, and others on patients with
cancer.
Brenner et al. (1996) increased the core temperature of
their subjects by 0.7°C during 3 h of seated rest in a climatic
chamber. They observed no change in either NK cell count
or cytolytic activity relative to a similar period of seated rest
under thermo-neutral conditions. However, somewhat larger
increases in core body temperatures have generally enhanced
cytotoxicity. Downing and Taylor (1987) found a
53% increase of NK cell cytolytic activity when a rectal
temperature of 39°C was induced by water-bath immersion.
Kappel et al. (1991c) increased the core temperature to
39.5°C, also by water-bath immersion. Their subjects
showed increases in CD16+ numbers in response to heating;
moreover, both the spontaneous and the IL-2 enhanced cytotoxic
activity were increased on a per-cell basis. Lackovic
et al. (1988) also found an increase of NK cell activity in
subjects who had been immersed in a 39°C water-bath for
30 min, linking this to a release of somatotropin.
Very high core temperatures have sometimes been induced
in cancer therapy. Here, the interpretation of findings
is complicated by abnormalities in resting data (Brenner
et al. 1995).
Information on the long-term effects of heat exposure is
limited mainly to animals, which have sometimes been exposed
to very high temperatures. Such conditions depress
NK cell cytolytic activity (Johnston et al. 1986; Yoshioka
et al. 1990).


Mechanisms modifying NK activity
In some studies, changes in NK cell count are sufficient to
account for the alterations in NK cell activity observed during
body heating. Catecholamines that are secreted in response
to heat stress contribute to the increase in circulating
NK cell numbers, probably by causing a demargination of
NK cells. Increased concentrations of various components of
the cytokine cascade, including IL-1, IL-2, and IFN-g likely
explain any increase of cytotoxic activity per NK cell with
moderate increases of body temperature (Downing and
Taylor 1987; Neville and Sauder 1988), while increased
concentrations of prostaglandin E2 or cortisol account for
any downregulation of NK cell function with more severe
thermal stress (Gatti et al. 1987; Shephard 1997; Yang et al.
1992). Hyperthermia does not seem to influence either the
viability of NK cells, or their ability to bind to appropriate
targets (Johnston et al. 1986), but the release of cytotoxic
granules is compromised, possibly because of the release of
prostaglandins (Yang et al. 1992), or a disruption of the
microtubular system in the NK cells (Yoshioka et al. 1990).
Phagocytes
In vitro data suggest that a brief increase of incubation
temperature (to 39°C) enhances the rate of neutrophil migration
(Nahas et al. 1971). In contrast, a sustained elevation of
temperature (38.5°C for 72 h) inhibits the movement of
phagocytes (Roberts and Sandberg 1979).
Despite reports that a moderate increase of incubation
temperature (to 39°C) enhances the bactericidal activity of
phagocytes, Kappel et al. (1993) found little change in either
chemiluminescence or superoxide production per cell, if
blood was sampled when the core body temperature had
been increased to 39.5°C. The in vitro rate and extent of
phagocytosis diminishes if the temperature is allowed to rise
further, to 41°C (Mandell 1975; Peterson et al. 1976).


Soluble factors
Cytokines
Taylor et al. (1984) reported that 2 h of exposure to an incubation
temperature of 40.7–42.7°C had no effect on the in
vitro production of IFN-g, but after 6 h at this temperature
the production of both IFN-a and IFN-g was enhanced.
Downing and Taylor (1987) also found an increased secretion
of IL-2 and IFN-g in lymphocytes that had been drawn
from hyperthermic subjects. Prolonged (1–2 days) incubation
at 39°C suppressed the in vitro production of "inflammatory"
cytokines (IL-1-b, IL-6, and IFN-g), but not the
production of IL-1-a or TNF-a (Kappel et al. 1991a).
Increased plasma concentrations of inflammatory cytokines
have generally been observed during both moderate and severe
body heating. A 2°C increase of core temperature increased
plasma IFN-a concentrations (Downing and Taylor
1987), and IL-1 concentrations were significantly increased
16–20 h following an experiment where core temperature
had been increased by 4°C for 60 min (Neville and Saunders
1988). Plasma levels of IL-1, IL-6, TNF-a, and LPS are also
known to be increased in heat stroke (Bouchama et al. 1991;
Chang 1993).


Acute-phase proteins
Hietala et al. (1982) noted increases of a-1-antitrypsin
and transferrin following 1 week in which subjects had received
twice-daily 30-min sauna exposures that brought core
body temperatures to 38.5°C. There have been other reports
of abnormal overall complement levels following occupational
heat stress, but plasma concentrations of C3 and C4
apparently remain unchanged (Hietala et al. 1982).


Immunoglobulins
Moderate heat exposure usually has little effect upon either
plasma immunoglobulin levels (Hietala et al. 1982;
Stephenson et al. 1985) or the in vitro production of immunoglobulins
(Severs et al. 1996). However, repeated heat exposure
augments serum levels of IgA, IgG, and IgM (Green
et al. 1988; Jasnoski and Kluger 1987).

Heat shock proteins
If a core temperature of 42°C is sustained for 20–60 min,
this stimulates the production of heat shock proteins (Currie
and White 1983; Löcke et al. 1990). In contrast, a bout of
exercise that raises core temperature to 40°C or higher reduces
the subsequent secretion of the 70-kD heat stress protein
by leukocytes if they are incubated at 41°C (Ryan et al.
1991).


Exercise plus heat stress
Exercise exacerbates the rise in core body temperature in
any given environment, so that an external temperature that
boosts the immune function of a sedentary person can have
an excessive or a suppressant effect on the individual who is
active. Groups such as marathon runners and North American
football players frequently develop rectal temperatures
of 40–41°C. The industrial worker in a toxic environment or
the soldier who is wearing NBC protective clothing also
face problems of heat elimination, although occupational
physicians commonly limit the permitted increase of core
body temperature to 39.0–39.5°C.
The inflammatory-type immune responses observed during
passive heating are generally increased if the subject
also engages in steady moderate exercise (Bouchama et al.
1991). Nevertheless, it remains unclear whether the immune
changes observed during and immediately following a bout
of vigorous physical activity are attributable simply to the
rise in core body temperature, or whether physical activity
exacerbates the response for a given amount of body heating.
Very heavy physical activity can in itself give rise to a
moderate inflammatory response, and if exercise is taken out
of doors, the effects of the rise in core temperature, muscle
metabolism, and subclinical injuries may be further complicated
by an ultraviolet-induced increase in the formation of
free radicals (Hersey et al. 1983).
Heat exposure alone usually offers no physical stimulus
that would not have been encountered during vigorous physical
activity. Nevertheless, there may be differences of cognitive
appraisal between passively induced thermal stress
and the rise of temperature that is caused by exercise, and
these differences have the potential to modify immune responses.
Heavy physical activity introduces other factors
that are either absent or much less marked during passive
heat exposure (Table 1). There is a very substantial increase
of cardiac output. Plasma levels of catecholamines and
cortisol also increase greatly. Impulses arising from the motor
centers and from peripheral chemoreceptors modulate
© 1998 NRC Canada
542 Can. J. Physiol. Pharmacol. Vol. 76, 1998
autonomic activity, and in competitive events, an athlete
may face considerable psychological stress. Finally, if physical
activity is prolonged, immune function may be compromised
by a depletion of plasma amino acid levels, an
ac***ulation of tissue injuries, and the release of prostaglandins
and reactive species.


Despite the added features of heavy physical activity, the
overall immune response to moderate body heating seems
remarkably similar whether a person is resting or is exercising
vigorously (Brenner et al. 1996; Severs et al. 1996). The
most notable difference between exercise and passive heating
is that whereas moderate passive heating enhances lymphocyte
proliferation (Hietala et al. 1982), an equivalent
exercise-induced increase of core body temperature decreases
lymphocyte proliferation (Brenner et al. 1995). The
discrepancy can probably be explained by differences in the
CD4+/CD8+ cell ratio and in the concentrations of catecholamines
and cortisol reached during the two forms of stress.
Cross et al. (1996) had subjects exercise on an electrically
isolated cycle ergometer while they were immersed in water.
By varying the water temperature, the core body temperature
was either allowed to rise, or was clamped at normal resting
values. Their findings suggested that as much as a half of
the normally observed exercise-induced leukocytosis was attributable
to a rise in core temperature. Increases in lymphocyte
and granulocyte counts were largely checked by
thermal clamping, but the exercise-induced increase in
monocyte count persisted. Multiple regression ****yses suggested
that in addition to core temperature, changes in
cortisol and growth hormone concentrations contributed to
the changes in leukocyte sub-set counts that were observed
under the various experimental conditions.
Schectman (1987) compared the response to 1 h of exercise
at 60% of maximal oxygen intake with the response of
the same subjects when they were heated passively to an
equivalent rectal temperature (38.5°C). In their study, IL-2
and IFN-g levels were higher during passive heating than
when equivalent body heating was accompanied by vigorous
exercise. Our group, also, has noted the ability of a bout of
exercise to reduce IL-2 production (Rhind et al. 1996).
Brenner et al. (1996) used a randomized block design to
expose subjects to four conditions (control, passive heating,
two 30-min bouts of exercise at 50% of maximal oxygen intake,
and the same amount of exercise performed in a warm
environment). The NK cell count increased in both exercise
conditions, with parallel increases in NK cytotoxic activity.
Both responses were slightly larger in the hot than in the
thermally neutral environment, but any difference between
the two experiments could reflect no more than the larger increase
of rectal temperature in the hot environment (a 1.5 vs.
a 1.1°C increment). Multiple regression ****yses examined
the relationship of changes in NK cytotoxic function to
cortisol, epinephrine, and norepinephrine concentrations.
Perhaps because the intensity of activity was only moderate,
changes in NK function conformed most closely to changes
in norepinephrine concentrations.


Severs et al. (1996) studied other aspects of immune function
in the same group of subjects. Exercise-induced increases in
the counts for total leukocytes, granulocytes, lymphocytes,
CD3+ cells, CD4+ cells, CD8+ cells, and monocytes, and the
in vitro pokeweed mitogen (PWM)-stimulated production of
IgM were all greater in the hot than in the thermally neutral
environment. However, the greater response could again reflect
a higher core temperature, rather than a specific effect
of the added exercise.
Housh et al. (1991) found no differences in salivary IgA
either immediately or 1 h following a 1-h bout of exercise at
80% of maximal oxygen intake at any environmental temperature
from 6 to 34°C. Likewise, Sawka et al. (1989) saw
little change in the intravascular mass of serum
immunoglobulins when exercise was performed in the heat
rather than in a cool, environment, although there was a
translocation of C3 to the intravascular space as subjects became
hypohydrated.
Taken together, these observations suggest that heat exposure
may exacerbate both the early inflammatory reaction and
the subsequent counter-regulatory trend to immunosuppression
© 1998 NRC Canada
Shephard 543
Variable
Effect of
increased core
temperature
Effect of
physical
activity
Total leukocyte count * *
Granulocytes * *
Neutrophils * *
Eosinophils *, «,¯ ¯
Basophils « «
Monocytes *,« *
Lymphocytes
%CD3+ ¯ «, *
%CD4+ ¯ ¯
%CD8+ * *, «
%CDI9+ ¯,« ¯
%NK * *
Lymphocyte function
Phagocytosis * (¯ if 41.0°C) *, «, ¯
NK cell activiy * (¯ if 41.6°C) *
LAK activiy * *
Proliferation rate * (¯ if 40.0°C) «, ¯
Soluble factors
Acute-phase proteins * *
Complement (overall) * *
C3, C4 « «, ¯
Plasma immunoglobulins *,« *
Salivary IgA ¯ ¯
Cytokules
IL-1 * *
IL-2 * ¯, *
IL-6 * *
Interferons
IFNa * *
IFNg *
Others
CSF *
TNF * *
Note: *, increase; ¯ decrease; «, no significant change; NK, natural
killer; LAK, lymphokine-activated killer; IFN, interferon; CSF, colony
stimulating factor; TNF, tumor necrosis factor.
Table 1. A comparison of immunological responses to increased
body temperature and to physical activity. (Data taken largely
from Brenner et al. (1995, p.70)).