Caffeine

Chapter 28
Caffeine
LAWRENCE L. SPRIET AND RICHARD A. HOWLETT
Introduction
Caffeine is a socially acceptable drug that is
widely consumed throughout the world. It is
also commonly used by athletes in their daily
lives and in preparation for athletic training
and competitions. Caffeine is a ‘controlled or
restricted drug’ in the athletic world. Urinary
caffeine levels greater than 12 mg · ml–1 following competitions are considered illegal by
the International Olympic Committee (IOC).
However, most athletes who consume caffeine beverages prior to exercise would not
approach the illegal limit following a competition. Therefore, if caffeine ingestion
enhances sports performance, it occupies a
unique position in the sports world. It is an
accepted component of the diet of many athletes,
although it has no nutritional value and would
be a ‘legal’ drug and ergogenic aid in these
situations.
Review articles in the early 1990s concluded
that the effects of caffeine ingestion on exercise
performance and metabolism were inconsistent
(Wilcox 1990; Conlee 1991). The authors stated
that many experiments had not been well
controlled and Conlee (1991) summarized the
factors which appeared to confound the caffeine
results: the exercise modality, exercise power
output, caffeine dose used in the experimental
design; the nutritional status, training status,
previous caffeine use of the subjects; and individual variation. An additional factor is the ability to
reliably measure exercise performance, which
improves with increased training frequency and
intensity.
Recent research has attempted to control these
factors and has demonstrated an ergogenic effect
of caffeine during prolonged endurance exercise
(> 40 min). Investigations examining the effects
of caffeine on exercise performance during
intense exercise lasting approximately 20 min
and shorter durations (ª 4–7 min) and sprinting
(< 90 s) have also appeared. At this point it is difficult to conclude whether caffeine is ergogenic
during exercise lasting less than 20 min (for
review, see Spriet 1997).
Caffeine appears to be taken up by all tissues
of the body, making it difficult to independently
study its effects on the central nervous
system (CNS) and the peripheral tissues (skeletal
muscle, liver and adipose tissue) in the exercising human. It is also likely that multiple and/or
different mechanisms may be responsible for
performance enhancement in different types of
exercise.
This chapter provides a brief but comprehensive review of the issues surrounding caffeine’s
ability to enhance exercise performance in
humans and the mechanisms which may explain
the ergogenic effects. The chapter does not
contain a complete list of citations but highlights
current thinking in the caffeine area and indicates where information is lacking.
379
nutrition and exercise
Caffeine and endurance
exercise performance
Early studies
The interest in caffeine as an ergogenic aid
during endurance exercise was initially stimulated by work from Costill’s laboratory in the late
1970s. Trained cyclists improved their cycle time
to exhaustion at 80% of maximal oxygen con.
sumption (Vo2max.) from 75 min in the placebo
condition to 96 min following caffeine (330 mg)
ingestion (Costill et al. 1978). A second study
demonstrated a 20% increase in the amount of
work performed in 2 h following 250 mg caffeine
(Ivy et al. 1979). These studies reported increased
venous free fatty acid (FFA) concentrations,
decreased respiratory exchange ratios (RER) and
increased fat oxidation (ª 30%) in the caffeine
trials. A third study reported that ingestion of
5 mg caffeine · kg–1 body mass spared muscle
glycogen and increased muscle triacylglycerol
(TG) use (Essig et al. 1980). In the 1980s, most
investigators examined only the effects of caffeine on metabolism and not on endurance performance. Furthermore, conclusions regarding
the metabolic effects of caffeine were equivocal
and based on changes in plasma FFA and
RER. This work has been extensively reviewed
(Wilcox 1990; Graham et al. 1994; Tarnopolsky
1994; Spriet 1995).
Recent endurance performance and
metabolic studies
Several well-controlled studies in the 1990s
examined the performance and metabolism
effects of caffeine in well-trained athletes, accustomed to exhaustive exercise and race conditions. These experiments examined the effects of
9 mg caffeine · kg–1 body mass (in capsule form)
on running and cycling time to exhaustion at
.
80–85% Vo2max. (Graham & Spriet 1991; Spriet
et al. 1992), the effects of varying doses
(3–13 mg · kg–1) of caffeine on cycling performance (Graham & Spriet 1995; Pasman et al.
1995) and the effects of a moderate caffeine dose
(5 mg · kg–1) on performance of repeated 30-min
bouts of cycling (5 min rest between bouts) at
.
85–90% Vo2max. (Trice & Haymes 1995).
Collectively, this work produced or confirmed
several important findings. Endurance performance was improved by approximately 20–50%
compared with the placebo trial (40–77 min) following ingestion of varying caffeine doses (3–13
mg · kg–1) in elite and recreationally trained athletes while running or cycling at approximately
.
80–90% Vo2max. (Figs 28.1, 28.2). Without exception, the 3, 5 and 6 mg · kg–1 doses produced an
ergogenic effect with urinary caffeine levels
below the IOC acceptable limit (Fig. 28.3). Three
of four experiments using a 9 mg · kg–1 dose
reported performance increases, while 6/22
athletes tested in these studies had urinary
caffeine at or above 12 mg · ml–1. Performance was
enhanced with a 13 mg · kg–1 dose, but 6/9 athletes had urinary caffeine well above 12 mg · ml–1
(Fig. 28.3). The side-effects of caffeine ingestion
(dizziness, headache, insomnia and gastrointestinal distress) were rare with doses at or below
6 mg · kg–1, but prevalent at higher doses (9–13
mg · kg–1) and associated with decreased perfor-
80
Mean performance time (min)
380
151%
144%
Bike
Treadmill
60
40
20
0
Fig. 28.1 Performance times for subjects running
and
. cycling to exhaustion at approximately 85%
Vo2max. after placebo (䊐) or caffeine ( ) ingestion.
Performance was significantly improved by 51%
during running and 44% while cycling. From Graham
and Spriet (1991), with permission.
caffeine
80
70
381
9 mg
Caffeine (µM)
70
60
50
60
6 mg
50
40
30
3 mg
Time (min)
20
40
10
0
30
0
6
3
Dose (mg.kg–1 body wt)
5
15
Exh.
9
Fig. 28.2 Performance
. times during running to
exhaustion at 85% Vo2max. following placebo or
caffeine ingestion (3, 6 or 9 mg · kg–1 body weight) 1 h
prior to exercise. All caffeine conditions were
significantly different from placebo. From Graham and
Spriet (1995), with permission.
30
Urinary caffeine (µg.ml–1)
0
Fig. 28.4 Plasma caffeine concentrations
during
.
exhaustive (Exh.) cycling at 80% Vo2max. following
the ingestion of placebo or 3, 6 and 9 mg caffeine · kg–1
body mass 1 h prior to exercise. Exhaustion occurred
between 50 and 62 min in all trials. From Graham and
Spriet (1995), with permission.
10
25
20
15
10
5
0
–60
Time (min)
20
0
Placebo
5
9
mance in some athletes at 9 mg · kg–1 (Graham &
Spriet 1995).
Caffeine generally produced no change in
venous plasma noradrenaline (norepinephrine)
concentration at rest or exercise, a twofold
increase in plasma adrenaline (epinephrine) concentration at rest and exercise and increased
plasma FFA concentration at rest. The elevated
FFA concentration at the onset of exercise with
caffeine was no longer present after 15–20 min of
exercise. At the lowest caffeine dose (3 mg · kg–1),
performance was increased without a significant
increase in plasma venous adrenaline and
FFA. Muscle glycogen utilization was reduced
following caffeine ingestion, but the ‘sparing’
was limited to the initial 15 min of exercise at
.
approximately 80% Vo2max..
13
Caffeine (mg.kg–1 body wt)
Fig. 28.3 Individual urine caffeine concentrations in
15 men
. following exhaustive cycling at approximately
80% Vo2max. and the ingestion of 5, 9 or 13 mg caffeine ·
kg–1 body weight. The horizontal line depicts the
acceptable level of less than 12 mg caffeine · ml–1 urine,
as outlined by the International Olympic Committee.
From Pasman et al. (1995), with permission.
Caffeine and short-term
exercise performance
There has been recent interest in the effects of caffeine ingestion on performance of short-term
exercise lasting between 30 s and 20–40 min. If
caffeine has an ergogenic effect during shortterm exercise, the mechanism will not be related
382
nutrition and exercise
to increased fat oxidation and decreased carbohydrate (CHO) oxidation, as CHO availability
does not limit performance in this type of
exercise.
Graded exercise tests: 8–20 min
Several studies reported no effect of moderate
doses of caffeine on time to exhaustion and
.
Vo2max. during graded exercise protocols lasting
8–20 min (Dodd et al. 1993). However, two
studies reported prolonged exercise times when
doses of 10–15 mg caffeine · kg–1 were given
(McNaughton 1987; Flinn et al. 1990). Unfortunately, no mechanistic information presently
exists to explain how these high caffeine doses
prolong exercise time during a graded test,
although it might be predicted that central effects
would be the most likely cause.
Intense aerobic exercise: ª 20–40 min
Competitive races lasting approximately 20–40
min require athletes to exercise at power outputs
.
of approximately 80–95% Vo2max.. Caffeine
(6 mg · kg–1) significantly reduced 1500-m swim
trial time, from 21:22 (± 38 s) to 20:59 (± 36 s)
(min:s), in trained distance swimmers (MacIntosh & Wright 1995). The authors reported lower
pre-exercise venous plasma [K+] and higher postexercise venous blood glucose concentration
with caffeine and suggested that electrolyte
balance and exogenous glucose availability may
be related to caffeine’s ergogenic effect. A second
study reported no ergogenic effect of caffeine in
mildly trained military recruits when cycling to
exhaustion (26–27 min) at approximately 80%
.
Vo2max. at sea level (Fulco et al. 1994). However,
cycle time was improved upon acute (35 vs. 23
min) and chronic (39 vs. 31 min) exposure to
altitude.
Intense aerobic exercise: ª 4–7 min
Exercise events at high power outputs (ª100–
.
110% Vo2max.) that last for approximately 4–7 min
require near-maximal or maximal rates of energy
provision from both aerobic and anaerobic
sources.
Collomp et al. (1991) reported that moderate
caffeine doses increased cycle time to exhaustion
.
at 100% Vo2max., from 5:20 with placebo to 5:49 in
one group and 5:40 in a second group, although
the increases were not statistically significant.
Wiles et al. (1992) reported that coffee ingestion (ª
150–200 mg caffeine) improved 1500-m race time
on a treadmill by 4.2 s over placebo (4:46.0 vs.
4:50.2). The runners in this study were welltrained, but clearly not elite. In a second experiment, subjects consumed coffee or placebo and
then ran for 1100 m at a predetermined pace, followed by a final 400 m where they ran as fast as
possible. The time to complete the final 400 m
was 61.25 s with coffee and 62.88 s without. Following coffee, all subjects ran faster and the
.
mean Vo2max. during the final 400 m was higher.
To document such small changes, the average
response to three trials in the caffeine and
placebo conditions was determined in both
experiments.
Jackman et al. (1996) examined the effects of
caffeine ingestion (6 mg · kg–1) on the performance and metabolic responses to three bouts of
.
cycling at 100% Vo2max.. Bouts 1 and 2 lasted 2
min and bout 3 was to exhaustion, with rest
periods of 6 min between bouts. Time to exhaustion in bout 3 was improved with caffeine (4.93 ±
0.60 min vs. placebo, 4.12 ± 0.36 min; n = 14).
Muscle and blood lactate measurements suggested a higher production of lactate in the
caffeine trial, even in bouts 1 and 2, when power
output was fixed. The glycogenolytic rate was
not different during bouts 1 and 2 and less than
50% of the muscle glycogen store was used in
either trial during the protocol. The authors
concluded that the ergogenic effect of caffeine
during short-term intense exercise was not associated with glycogen sparing and may be caused
by either a direct action on the muscle or altered
CNS function.
Sprint exercise
Sprinting is defined as exercise or sporting
caffeine
events at power outputs corresponding to
.
150–300% Vo2max. lasting less than 90 s. The
amount of energy derived from anaerobic
processes would be approximately 75–80% of the
total in the first 30 s, approximately 65–70% over
60 s and approximately 55–60% of the total
energy over 90 s.
Williams et al. (1988) reported that caffeine
ingestion had no effect on maximal power
output or muscular endurance during short,
maximal bouts of cycling. Collomp et al. (1992)
reported that 5 mg caffeine · kg–1 did not increase
peak power or total work during a 30-s Wingate
test, but the same group later reported that 250
mg caffeine produced a 7% improvement in the
maximal power output generated during a series
of 6-s sprints at varying force–velocity relationships (Anselme et al. 1992). The authors also
examined the effects of 4.3 mg caffeine · kg–1 on
two 100-m freestyle swims, separated by 20 min
(Collomp et al. 1990). In well-trained swimmers,
caffeine increased swim velocity by 2% and 4% in
the two sprints, but performance times were not
reported. Caffeine had no effect on sprint performance in untrained swimmers.
Therefore, given the present information, it is
not possible to conclude whether caffeine has an
ergogenic effect on sprint performance. The brief
and intense nature of sprint exercise makes it
difficult to study and demonstrate significant
differences.
Field studies
Exercise performance in most laboratory studies
is measured as the time taken to reach exhaustion
at a given power output or the amount of work
that can be performed in a given amount of time.
However, in the field, performance is usually
measured as the time taken to complete a certain
distance. Consequently, extrapolations from
the laboratory to field settings may not be valid.
Occasionally, laboratory studies simulate race
conditions and other studies measure performance in the field (track, swimming pool) in time
trial settings without actual race conditions.
However, these studies still do not simulate real
383
competitions. In field studies that do simulate
race conditions, it is often impossible to employ the controls required to generate conclusive
results. For example, Berglund and Hemmingsson (1982) reported that caffeine increased crosscountry ski performance by 1–2.5 min with a
control race lasting 1–1.5 h. This improvement
occurred at altitude but not at sea level. Unfortunately, the weather and snow conditions were
variable in both locations, requiring normalization of the performance times in order to
compare results. A recent field study reported
that ingesting 0, 5 or 9 mg caffeine · kg–1 had no
effect on 21-km road-race performance in hot
and humid environments (Cohen et al. 1996).
While subjects acted as their own controls, no
subjects received the placebo treatment in all
three races to assess whether between race environmental differences affected race performance,
independent of caffeine.
The problems associated with field trials
raise questions about the validity of the results
and indicate how difficult it is to perform
well-controlled and meaningful field trials.
However, there is clearly a need for more field
studies.
Theories of ergogenicity
The mechanisms that may contribute to the
ergogenic effects of caffeine are categorized
into three general theories. The first theory is
the classic or ‘metabolic’ explanation for the
ergogenic effects of caffeine during endurance
exercise involving an increase in fat oxidation
and reduction in CHO oxidation. The metabolic
category also includes factors which may affect
muscle metabolism and performance in a direct
manner, including inhibition of phosphodiesterase, leading to an elevated cyclic adenosine
monophosphate (AMP) concentration, and
direct effects on key enzymes such as glycogen
phosphorylase (PHOS). The second theory proposes a direct effect of caffeine on skeletal muscle
performance via ion handling, including Na+–
K+-ATPase activity and Ca2+ kinetics. The third
theory suggests that caffeine exerts a direct effect
384
nutrition and exercise
on portions of the CNS that alter the perception
of effort and/or motor unit recruitment.
Metabolic mechanisms for improved
exercise performance
Presently, it seems that metabolic mechanisms
are part of the explanation for the improvement
in endurance performance following caffeine
ingestion (5–13 mg · kg–1), except at low caffeine
doses (2–4 mg · kg–1) where this has not been fully
examined. The increased plasma FFA concentration at the onset of exercise, the glycogen sparing
in the initial 15 min of exercise and increased
intramuscular TG use during the first 30 min of
exercise suggest a greater role for fat metabolism
early in exercise following caffeine doses of at
least 5 mg · kg–1. However, there are currently no
definitive measurements of increased plasma
FFA use following caffeine ingestion. Also, these
metabolic findings do not preclude other factors
contributing to enhanced endurance performance as discussed below.
It has been suggested that the increased fat oxidation and decreased glycogen use in muscle following caffeine ingestion could be explained by
the classic glucose–fatty acid cycle proposed by
Randle and colleagues (Spriet & Dyck 1996). In
this scheme, elevated FFA availability to the
muscle produced increases in muscle citrate
and acetyl-coenzyme A, which were believed
to inhibit the enzymes phosphofructokinase
and pyruvate dehydrogenase. The subsequent
decrease in glycolytic activity increased glucose
6-phosphate content, leading to inhibition of
hexokinase and ultimately decreased muscle
glucose uptake and oxidation. However, these
mechanisms were not involved in the CHO
.
sparing during exercise at 85% Vo2max. with caffeine ingestion or increased fat availability
(Spriet et al. 1992; Dyck et al. 1993). Instead, the
mechanism for muscle glycogen sparing following caffeine ingestion appeared related to the
regulation of glycogen PHOS activity via the
energy status of the cell (Chesley et al. 1998). Subjects who spared muscle glycogen had smaller
decreases in muscle phosphocreatine and
smaller increases in free AMP during exercise in
the caffeine vs. placebo trials. The resultant lower
free inorganic phosphate and AMP concentrations decreased the flux through the more active
a form of PHOS. There were no differences in
these metabolites between trials in subjects who
did not spare muscle glycogen. It is not presently
clear how caffeine defends the energy state of the
cell at the onset of intense exercise, but it may be
related to the availability of fat (Chesley et al.
1998).
It also appears that adrenaline does not contribute to the metabolic changes which lead to
enhanced endurance performance following
caffeine ingestion. First, performance was
enhanced with 3 mg caffeine · kg–1 without significant increases in plasma adrenaline and FFA,
although FFA were increased twofold at rest
(Graham & Spriet 1995). Second, an infusion
of adrenaline, designed to produce resting and
exercise adrenaline concentrations similar to
those induced by caffeine had no effect on
plasma FFA concentration or muscle glycogenolysis during exercise (Chesley et al. 1995). Third,
Van Soeren et al. (1996) gave caffeine to spinalcord injured subjects and reported an increased
plasma FFA concentration without changes in
adrenaline concentration. These findings suggest
that caffeine ingestion affects the mobilization of
fat by antagonizing the adenosine receptors in
adipose tissue.
Therefore, while it is clear that metabolic
changes contribute to the ergogenic effect of caffeine during endurance exercise, aspects of the
metabolic contribution have not been adequately
examined in all situations. Measurements of
muscle glycogen and TG use and plasma FFA
turnover are required to determine the magnitude of the metabolic link to improved performance at all caffeine doses and endurance
exercise situations.
There is some evidence that caffeine has an
ergogenic effect on short-term intense exercise.
The mechanism will not be related to increased
fat oxidation and decreased CHO oxidation, as
CHO availability does not limit performance in
this situation. It is possible that increased anaero-
caffeine
bic energy provision from glycogen breakdown
and the glycolytic pathway may contribute to the
improvement in performance during repeated
.
bouts of intense exercise (100% Vo2max.) lasting
2–5 min (Jackman et al. 1996). If this occurred, it
would likely be the result of a direct effect of caffeine or a caffeine metabolite.
A few additional metabolic mechanisms have
been suggested to contribute to the ergogenic
effects of caffeine. It is commonly stated that caffeine inhibits phosphodiesterase, leading to an
increase in cyclic AMP concentration and
muscle glycogen PHOS activation. However, the
support for these conclusions is from in vitro or
‘test tube’ studies that used pharmacological caffeine levels and it is now generally accepted that
these effects would not be present at physiological caffeine concentrations (for review, see
Tarnopolsky 1994; Spriet 1995). Vergauwen et al.
(1994) recently reported that adenosine receptors
mediate the stimulation of glucose uptake and
transport by insulin and contractions in rat skeletal muscle. Caffeine, as an adenosine receptor
antagonist and at a physiological level (77 mm),
decreased glucose uptake during contractions.
This may be an additional mechanism whereby
CHO use is spared following caffeine ingestion and replaced by increased fat oxidation.
However, there have been no definitive reports
demonstrating that adenosine receptors exist in
human skeletal muscle.
Ion handling in skeletal muscle
Caffeine may alter the handling of ions in skeletal muscle and contribute to an ergogenic effect
during exercise. Most of the supporting evidence
has come from in vitro experiments using pharmacological doses of methylxanthines. The candidates that have been suggested to contribute to
an ergogenic effect in a physiological environment are increased Ca2+ release during the latter
stages of exercise and increased Na+–K+-ATPase
activity, which may help maintain the membrane
potential during exercise. These are the most
likely candidates since the lowest methylxanthine concentration used to show these effects in
385
the in vitro experiments approached the actual
methylxanthine concentrations that have been
shown to be ergogenic in vivo (Lindinger et al.
1993; Tarnopolsky 1994).
It has been demonstrated in vitro that pharmacological levels of methylxanthine affect several
steps in skeletal muscle excitation–contraction
coupling:
1 increasing the release of Ca2+ from the sarcoplasmic reticulum;
2 enhancing troponin/myosin Ca2+ sensitivity;
and
3 decreasing the reuptake of Ca2+ by the sarcoplasmic reticulum (Tarnopolsky 1994).
Methylxanthines also stimulate Na+–K+ATPase activity in inactive skeletal muscle
leading to increased rates of K+ uptake and Na+
efflux. This attenuates the rise in plasma [K+]
with exercise, which may help maintain the
membrane potential in contracting muscle and
contribute to caffeine’s ergogenic effect during
exercise (Lindinger et al. 1993, 1996). Any of these
changes could produce increases in skeletal
muscle force production. However, at the present
time, it is not clear if these potential ion-handling
effects of caffeine contribute to an ergogenic
effect, given the physiological or in vivo
methylxanthine concentration normally found in
humans.
Central effects of caffeine
While it is almost universally accepted that some
of the ergogenic effects of caffeine are manifested
through effects on the CNS, it is almost impossible to quantify how much of caffeine’s ability
to delay fatigue is due to central or peripheral
effects. Complicating the problem is the fact that
it is not clear how caffeine exerts its actions on
the CNS. Caffeine is certainly a CNS stimulant,
causing increased wakefulness and vigilance
(Van Handel 1983; Nehlig et al. 1992; Daly 1993).
Some have attributed the increased performance
derived from caffeine simply to this increased
alertness or improved mood (Nehlig & Debry
1994). However, the ability of caffeine to delay
fatigue points to more complex mechanisms than
386
nutrition and exercise
simply heightened arousal. Because they are also
related to peripheral metabolic effects, the following topics are of special interest in a discussion of caffeine’s central effects: adenosine
receptor antagonism, lowered perceived exertion and the central fatigue hypothesis.
Adenosine receptor antagonism
Since caffeine can freely pass through the
blood–brain barrier (Nehlig et al. 1992), its concentration in the brain and CNS increases rapidly
following ingestion, in concert with changes
in other body tissues (Daly 1993). Caffeine
increases brain neurotransmitter concentration,
causing increases in spontaneous locomotor
activity and neuronal firing in animals (Nehlig et
al. 1992). It is generally accepted that the mechanism for neurotransmitter increases is adenosine
receptor antagonism and high adenosine receptor levels in the brain support this hypothesis
(Fernstrom & Fernstrom 1984; Snyder 1984; Daly
1993; Fredholm 1995).
Adenosine is both a neurotransmitter and neuromodulator, capable of affecting the release of
other neurotransmitters (Fernstrom & Fernstrom
1984). Adenosine and adenosine analogues generally cause lowered motor activity, decreased
wakefulness and vigilance, and decreases in
other neurotransmitter concentrations. Caffeine
and adenosine receptor antagonists have the
opposite effect by blocking the adenosine receptors. It is generally believed that the inhibition
(adenosine) or stimulation (caffeine) of neurotransmitter release is presynaptic (Snyder 1984;
Fredholm 1995). It has been demonstrated that
caffeine increases the concentration, synthesis
and/or turnover of all major neurotransmitters,
including serotonin, dopamine, acetylcholine,
noradrenaline and glutamate. These neurotransmitters are all inhibited by adenosine. The exact
consequences of these changes in neurotransmitters with regards to performance is currently
not known. Both dopamine and serotonin levels
have been implicated in the central effects of caffeine on fatigue and behaviour (Fernstrom &
Fernstrom 1984; Daly 1993), and in the develop-
ment of central fatigue exclusive of caffeine
ingestion (Davis & Bailey 1997). It has been suggested that an increase in excitatory neurotransmitters could lead to decreases in motorneurone
threshold, resulting in greater motor unit recruitment (Waldeck 1973) and subsequently lower
perceived exertion for a given power output
(Nehlig & Debry 1994; Cole et al. 1996). However,
this theory has not been demonstrated during
exercise, although it continues to be cited as a
potential mechanism (Nehlig & Debry 1994; Cole
et al. 1996).
Complicating the effects of caffeine on adenosine antagonism is the existence of two main
classes of adenosine receptors, A1 and A2 (Snyder
1984; Graham et al. 1994), each having differing
affinities for endogenous adenosine and xanthines, and affecting the release of different neurotransmitters (Daly 1993) Likewise, antagonism
of these receptors is dependent on the caffeine
concentration, which will either inhibit (A1) or
stimulate (A2) adenylate cyclase, leading to
differential effects and possibly explaining the
biphasic response to caffeine. Increasing caffeine
doses are stimulatory, but very high physiological doses are depressant (Snyder 1984). As well,
some adenosine antagonists display the same
affinity as xanthines for adenosine receptors, but
do not cause the same effects (Snyder 1984; Daly
1993). Finally, the binding of caffeine to benzodiazepine receptors and the relationship to
gamma-aminobutyric acid (GABA) and excitatory amino acids is currently being explored
(Nehlig et al. 1992; Daly 1993). Some authors
assert that adenosine receptor antagonism, while
likely the primary mechanism, cannot account
for all of caffeine’s actions on the CNS (Graham
et al. 1994).
Ratings of perceived exertion
One quantifiable aspect of caffeine’s central
effects is a lower rating of perceived exertion
(RPE) during exercise. Several studies have
demonstrated that (i) RPE at a standard power
output was lower in subjects following caffeine
ingestion than in controls (Costill et al. 1978), and
caffeine
(ii) subjects accomplished a greater amount of
work following caffeine ingestion than controls
when RPE was held constant (Ivy et al. 1979; Cole
et al. 1996). This significant decline in experimental RPE is certainly supported by anecdotal evidence. It has been speculated that the lowered
RPE with caffeine is due to a decrease in the
firing threshold of motorneurones (Nehlig &
Debry 1994; Cole et al. 1996) or changes in muscle
contraction force (Tarnopolsky 1994). Both
mechanisms would result in lowered afferent
feedback from the working muscle and a
lowered RPE, the first mechanism because more
motor units would be recruited for a given task
and the second because the force for a given stimulus would be greater. However, the ability of
physiological caffeine concentration to alter contractile function is equivocal as discussed earlier
(Graham et al. 1994; Tarnopolsky 1994). Another
hypothesis is that caffeine directly affects the
release of b-endorphins and other hormones that
modulate the feelings of discomfort and pain
associated with exhaustive exercise (Nehlig et al.
1992). A final explanation for the reduced RPE
may involve the central fatigue hypothesis
(Tarnopolsky 1994).
387
delayed the onset of CNS fatigue via serotonin
levels, then it must lower 5-HT levels or inhibit
the rise in 5-HT. However, the effects of caffeine
on the CNS and peripheral metabolism appear to
counter this process for two reasons. First, acute
caffeine ingestion has been shown to significantly increase brain 5-HT levels, most likely due
to increases in brain free TRP levels (Fernstrom &
Fernstrom 1984; Nehlig et al. 1992). Second, caffeine ingestion prior to exercise elevates plasma
FFA concentration at the onset of exercise, which
should increase free TRP, due to competition for
albumin binding, and hasten fatigue. It is possible that the rise in 5-HT at the onset of exercise
is overridden by other factors, such as
increased sympathetic drive, or favourable metabolic factors. Similarly, since it has been postulated that the ratio of 5-HT to dopamine is a
larger determinant in fatigue than the [5-HT]
alone (Davis & Bailey 1997), the caffeine-induced
rise in both neurotransmitters could offset each
other.
In summary, the caffeine-induced mechanism(s) that may delay central fatigue are still
undiscovered, but the link between caffeine
and the central fatigue hypothesis remains
intriguing.
c e n t r a l fat i gue h y p o t h e si s
Given that caffeine affects the CNS, it is appealing to link it to one proposed mechanism of
fatigue currently being investigated, the central
fatigue hypothesis (see Chapter 12). Briefly, this
hypothesis argues that the central component of
fatigue caused by exhaustive exercise is mediated by elevated levels of serotonin (5-HT) in the
brain, caused by an increase in its precursor,
tryptophan (TRP) (Blomstrand & Newsholme
1996). Tryptophan is the only amino acid that is
transported in plasma bound to albumin and it
competes for transport into the brain with
branched-chain amino acids (BCAA). Evidence
for the central fatigue theory includes increased
levels of brain 5-HT at fatigue, increased plasma
free TRP at fatigue caused by high FFAs, and
decreased fatigue with BCAA supplementation
(Blomstrand & Newsholme 1996). If caffeine
Complications of studying caffeine,
exercise performance and metabolism
It is important to note in a discussion of the performance, metabolic and central effects of caffeine ingestion that the mechanism(s) of action
may not be entirely due to the primary effects of
caffeine. Caffeine is a trimethylxanthine compound, which is rapidly metabolized in the liver
to three dimethylxanthines, paraxanthine, theophylline and theobromine. These are released
into the plasma as the caffeine concentration
declines and remain in the circulation longer.
While the plasma dimethylxanthine concentrations are not large, paraxanthine and theophylline are potential adenosine antagonists and
metabolic stimuli. Therefore, as caffeine and its
metabolites are often present at the same time, it
is difficult to resolve which tissues are directly or
388
nutrition and exercise
indirectly affected by which compound (Fig.
28.5). Due to this uncertainty, the reader should
note that when the term ‘caffeine’ is used in this
chapter, it could be any of the methylxanthines.
Another complication of studying caffeine
ingestion is the variability of individual
responses, affecting central, metabolic and exercise performance responses to caffeine. This
problem affects all categories of subjects, but is a
larger problem with less aerobically fit individuals. Chesley et al. (1998) reported a variable
glycogen sparing response to a high caffeine
dose (9 mg · kg–1) in untrained men. Only 6/12
subjects demonstrated glycogen sparing during
.
15 min of cycling at approximately 85% Vo2max.,
whereas the sparing response was more uniform
in a group of trained men (Spriet et al. 1992). Variability is also present in all groups of caffeine
users, including mild and heavy users, users
withdrawn from caffeine and non-users. Therefore, while mean results in groups of subjects and
athletes predict improved athletic performance,
predictions that a given person will improve are
less certain.
There has been a recent report comparing the
effects of 4.5 mg caffeine · kg–1, given in ‘pure’
3.5
6 mg
Paraxanthine (µM)
3.0
9 mg
2.5
2.0
3 mg
1.5
1.0
Placebo
0.5
0
–60
0
5
15
Exh.
Time (min)
Fig. 28.5 Plasma paraxanthine concentrations
during
.
exhaustive (Exh.) cycling at 80% Vo2max. following
the ingestion of placebo or 3, 6 and 9 mg caffeine · kg–1
body mass 1 h prior to exercise. Exhaustion occurred
between 50 and 62 min in all trials. From Graham and
Spriet (1995), with permission.
capsule form or in two mugs of strong coffee
(Graham et al. 1998). Caffeine in capsule form
resulted in the usual metabolic and performance
effects, but the ingested coffee produced less of a
response in plasma adrenaline concentration and
little or no effect on performance, even though
the plasma caffeine concentrations were identical. It appears that the hundreds of additional
chemicals in coffee negated the usual ergogenic
benefit. On the other hand, there have been
reports where caffeine administration in coffee
produced strong ergogenic performance effects
(Wiles et al. 1992). Therefore, while it is common
to equate caffeine with coffee, it should be noted
that rarely is coffee the method of administration
in research studies and it may be misleading to
equate the two.
The study of caffeine ingestion and exercise
performance has been generally limited to male
subjects. There has been little systematic study of
the response of females to caffeine ingestion at
rest and during exercise. It will be important to
control for menstrual status in future studies, as
oestrogen may affect the half-life of caffeine.
Other considerations of
ingesting caffeine
Caffeine dose
Caffeine is a ‘controlled or restricted substance’
with respect to the IOC. Athletes are permitted
up to 12 mg caffeine · ml–1 urine before it is considered illegal. This allows athletes who normally
consume caffeine in their diet to continue
this practice prior to competition. An athlete can
consume a very large amount of caffeine before
reaching the ‘illegal limit’. A 70-kg person could
drink three or four mugs or six regular-size
cups of drip-percolated coffee approximately 1 h
before exercise, exercise for 1–1.5 h, and a subsequent urine sample would only approach the
urinary caffeine limit. A caffeine level above 12
mg · ml–1 suggests that a person has deliberately
taken caffeine in capsule or tablet form or as suppositories, in an attempt to improve performance. Not surprisingly, only a few athletes have
caffeine
been caught with illegal levels during competitions, although formal reports of the frequency of
caffeine abuse are rare. One study reported that
26/775 cyclists had illegal urinary caffeine levels
when tested following competition (Delbecke &
Debachere 1984).
Urinary caffeine and doping
The use of urinary caffeine levels to determine
caffeine abuse in sport has been criticized
(Duthel et al. 1991). Only 0.5–3% of orally
ingested caffeine actually reaches the urine as the
majority is metabolized in the liver. The excreted
caffeine by-products are not measured in doping
tests. Other factors also affect the amount of caffeine that reaches the urine, including body
weight, gender and hydration status of the
athlete. The time elapsed between caffeine ingestion and urine collection is also important and
affected by the exercise duration and environmental conditions. Sport governing bodies may
not regard these concerns as problems since most
people caught with illegal levels of caffeine
will have used the drug in a doping manner.
However, it is possible that someone who metabolizes caffeine slowly or who excretes 3% of the
ingested dose rather than 0.5% could have illegal
urine levels following a moderate dose.
Habitual caffeine consumption
An athlete’s normal caffeine intake habits may
affect whether acute caffeine ingestion improves
performance. Many investigators ask users to
refrain from caffeine consumption for 2–3 days
prior to experiments. Caffeine metabolism is not
increased by use, but the effects of caffeine may
be altered by habitual use via alterations in
adenosine receptor populations. As reviewed by
Graham et al. (1994), several studies suggest that
chronic caffeine use dampens the adrenaline
response to exercise and caffeine, but does not
affect plasma FFA concentration or exercise
RER (Bangsbo et al. 1992; Van Soeren et al.
1993). However, these changes do not appear to
dampen the ergogenic effect of 9 mg caffeine ·
389
kg–1. Endurance performance increased in all
subjects when both caffeine users and non-users
were examined and users abstained from caffeine for 48–72 h prior to experiments (Graham &
Spriet 1991; Spriet et al. 1992). However, the performance results were more variable in a subsequent study with more non-users (Graham &
Spriet 1995). In addition, Van Soeren and
Graham (1998) reported no effect of up to 4 days
of caffeine withdrawal on exercise hormonal
and metabolic responses to doses of 6 or 9 mg
caffeine · kg–1 in recreational cyclists. Time to
.
exhaustion at 80–85% Vo2max. improved with
caffeine and was unaffected by 0–4 days of
withdrawal.
Caffeine and high carbohydrate diets
An early investigation suggested that a high
CHO diet and a prerace CHO meal negated the
expected increase in plasma FFA concentration
following caffeine ingestion during 2 h of exer.
cise at approximately 75% Vo2max. (Weir et al.
1987). These results implied that high CHO
diets negated the ergogenic effects of caffeine,
although performance was not measured.
However, a high CHO diet and a pretrial CHO
meal did not prevent caffeine-induced increases
in performance in a number of recent studies
using well-trained/recreational runners and
cyclists (Spriet 1995).
Diuretic effect of caffeine
Because caffeine is a diuretic, it has been suggested that caffeine ingestion may lead to poor
hydration status prior to and during exercise.
However, no changes in core temperature, sweat
loss or plasma volume were reported during
exercise following caffeine ingestion (Gordon
et al. 1982; Falk et al. 1990). It has also been
demonstrated that urine flow rate, decreases in
plasma volume, sweat rate and heart rate were
unaffected by caffeine (ª 600 mg), ingested in a
CHO electrolyte drink (ª 2.5 l) during 1 h at rest
.
and 3 h of cycling at 60% Vo2max. (Wemple et al.
1997).
390
nutrition and exercise
Conclusion
Caffeine ingestion (3–13 mg · kg–1 body mass)
prior to exercise increases performance during
prolonged endurance cycling and running in the
laboratory. Caffeine doses below 9 mg · kg–1 generally produce urine caffeine levels below the
IOC allowable limit of 12 mg · ml–1. Moderate
caffeine doses (5–6 mg · kg–1) may also increase
short-term intense cycling performance (ª 4–7
min) in the laboratory and decrease 1500-m
swim time (ª 20 min). These results are generally
reported in well-trained or recreational athletes,
but field studies are lacking to confirm the
ergogenic effects of caffeine in the athletic world.
The mechanisms for the improved endurance
have not been clearly established. Caffeine ingestion generally increases resting venous plasma
FFA concentration and reduces muscle glycogen
use and increases muscle TG use early during
endurance exercise, suggesting greater fat oxidation and reduced CHO oxidation in the working
muscles. However, a single metabolic explanation for the ergogenic effect of caffeine is unlikely,
especially at low caffeine doses that do not cause
major metabolic changes. All human performance studies have been unable to separate the
central effects of caffeine from peripheral effects.
Therefore, a central contribution to the enhancement of endurance exercise performance following caffeine ingestion is a strong possibility.
Potential mechanisms for improved performance during short-term intense exercise
include direct caffeine effects on the CNS and/or
ion handling in skeletal muscle and increased
anaerobic energy provision in muscle. Definitive
research into the mechanisms of the ergogenic
effects of caffeine in exercising humans is hampered by the ability of this drug and its byproducts to affect both central and peripheral
processes.
References
Anselme, F., Collomp, K., Mercier, B., Ahmaidi, S. &
Prefaut, C. (1992) Caffeine increases maximal anaerobic power and blood lactate concentration. European Journal of Applied Physiology 65, 188–191.
Bangsbo, J., Jacobsen, K., Nordberg, N., Christensen,
N.J. & Graham, T. (1992) Acute and habitual caffeine
ingestion and metabolic responses to steady-state
exercise. Journal of Applied Physiology 72, 1297–1303.
Berglund, B. & Hemmingsson, P. (1982) Effects of caffeine ingestion on exercise performance at low and
high altitudes in cross country skiers. International
Journal of Sports Medicine 3, 234–236.
Blomstrand, E.E. & Newsholme, E.A. (1996) Glucosefatty acid cycle and fatigue involving 5hydroxytryptamine. In Biochemistry of Exercise IX (ed.
R.J. Maughan & S.M. Shirreffs), pp. 185–195. Human
Kinetics, Champaign, IL.
Chesley, A., Hultman, E. & Spriet, L.L. (1995) Effects of
epinephrine infusion on muscle glycogenolysis
during intense aerobic exercise. American Journal
of Physiology 268 (Endocrinology Metabolism),
E127–E134.
Chesley, A., Howlett, R.A., Heigenhauser, G.J.F.,
Hultman, E. & Spriet, L.L. (1998) Regulation of
muscle glycogenolytic flux during intense aerobic
exercise after caffeine ingestion. American Journal of
Physiology 275, R596–603.
Cohen, B.S., Nelson, A.G., Prevost, M.C., Thompson,
G.D., Marx, B.D. & Morris, G.S. (1996) Effects of caffeine ingestion on endurance racing in heat and
humidity. European Journal of Applied Physiology 73,
358–363.
Cole, K.J., Costill, D.L., Starling, R.D., Goodpaster,
B.H., Trappe, S.W. & Fink, W.J. (1996) Effect of caffeine ingestion on perception of effort and subsequent work production. International Journal of Sport
Nutrition 6, 14–23.
Collomp, K., Caillaud, C., Audran, M., Chanal, J.-L. &
Prefaut, C. (1990) Influence of acute and chronic
bouts of caffeine on performance and catecholamines in the course of maximal exercise.
Comptes Rendues des Seances de la Societe de Biologie
184, 87–92.
Collomp., K., Ahmaidi, S., Audran, M., Chanal, J.-L. &
Prefaut, C. (1991) Effects of caffeine ingestion on performance and anaerobic metabolism during the
Wingate test. International Journal of Sports Medicine
12, 439–443.
Collomp, K., Ahmaidi, S., Chatard, J.C., Audran, M. &
Prefaut, C. (1992) Benefits of caffeine ingestion on
sprint performance in trained and untrained swimmers. European Journal of Applied Physiology 64,
377–380.
Conlee, R.K. (1991) Amphetamine, caffeine and
cocaine. In Ergogenics: Enhancement of Performance in
Exercise and Sport (ed. D.R. Lamb & M.H. Williams),
pp. 285–330. Brown and Benchmark, Indianapolis,
IN.
Costill, D.L., Dalsky, G.P. & Fink, W.J. (1978) Effects of
caffeine on metabolism and exercise performance.
Medicine and Science in Sports 10, 155–158.
caffeine
Daly, J.W. (1993) Mechanism of action of caffeine. In
Caffeine, Coffee, and Health (ed. S. Garatttini), pp.
97–150. Raven Press, New York.
Davis, J.M. & Bailey, S.P. (1997) Possible mechanisms of
central nervous system fatigue during exercise.
Medicine and Science in Sports and Exercise 29, 45–57.
Delbecke, F.T. & Debachere, M. (1984) Caffeine: use and
abuse in sports. International Journal of Sports Medicine 5, 179–182.
Dodd, S.L., Herb, R.A. & Powers, S.K. (1993) Caffeine
and exercise performance: an update. Sports Medicine
15, 14–23.
Duthel, J.M., Vallon, J.J., Martin, G., Ferret, J.M.,
Mathieu, R. & Videman, R. (1991) Caffeine and sport:
role of physical exercise upon elimination. Medicine
and Science in Sports and Exercise 23, 980–985.
Dyck, D.J., Putman, C.T., Heigenhauser, G.J.F.,
Hultman, E. & Spriet, L.L. (1993) Regulation of
fat–carbohydrate interaction in skeletal muscle
during intense aerobic cycling. American Journal of
Physiology 265, E852–859.
Essig, D., Costill, D.L. & VanHandel, P.J. (1980) Effects
of caffeine ingestion on utilization of muscle glycogen and lipid during leg ergometer cycling. International Journal of Sports Medicine 1, 86–90.
Falk, B., Burstein, R., Rosenblum, J., Shapiro, Y.,
Zylber-Katz, E. & Bashan, N. (1990) Effects of caffeine ingestion on body fluid balance and thermoregulation during exercise. Canadian Journal of
Physiology and Pharmacology 68, 889–892.
Fernstrom, J.D. & Fernstrom, M.H. (1984) Effects of caffeine on monamine neurotransmitters in the central
and peripheral nervous system. In Caffeine (ed. P.B.
Dews), pp. 107–118. Springer-Verlag, Berlin.
Flinn, S., Gregory, J., McNaughton, L.R., Tristram, S. &
Davies, P. (1990) Caffeine ingestion prior to incremental cycling to exhaustion in recreational cyclists.
International Journal of Sports Medicine 11, 188–193.
Fredholm, B.B. (1995) Adenosine, adenosine receptors
and the actions of caffeine. Pharmacology and
Toxicology 76, 93–101.
Fulco, C.S., Rock, P.B., Trad, L.A. et al. (1994) Effect of
caffeine on submaximal exercise performance at altitude. Aviation Space and Environmental Medicine 65,
539–545.
Gordon, N.F., Myburgh, J.L., Kruger, P.E. et al. (1982)
Effects of caffeine on thermoregulatory and myocardial function during endurance performance. South
African Medical Journal 62, 644–647.
Graham, T.E. & Spriet, L.L. (1991) Performance and
metabolic responses to a high caffeine dose during
prolonged exercise. Journal of Applied Physiology 71,
2292–2298.
Graham, T.E. & Spriet, L.L. (1995) Metabolic, catecholamine and exercise performance responses
to varying doses of caffeine. Journal of Applied
Physiology 78, 867–874.
391
Graham, T.E., Rush, J.W.E. & van Soeren, M.H. (1994)
Caffeine and exercise: metabolism and performance.
Canadian Journal of Applied Physiology 19, 111–138.
Graham, T.E., Hibbert, E. & Sathasivam, P. (1998)
Metabolic and exercise endurance effects of coffee
and caffeine ingestion. Journal of Applied Physiology
85, 883–889.
Ivy, J.L., Costill, D.L., Fink, W.J. & Lower, R.W. (1979)
Influence of caffeine and carbohydrate feedings on
endurance performance. Medicine and Science in
Sports 11, 6–11.
Jackman, M., Wendling, P., Friars, D. & Graham, T.E.
(1996) Metabolic, catecholamine and endurance
responses to caffeine during intense exercise. Journal
of Applied Physiology 81, 1658–1663.
Lindinger, M.I., Graham, T.E. & Spriet, L.L. (1993)
Caffeine attenuates the exercise-induced increase in
plasma [K+] in humans. Journal of Applied Physiology
74, 1149–1155.
Lindinger, M.I., Willmets, R.G. & Hawke, T.J. (1996)
Stimulation of Na+, K+-pump activity in skeletal
muscle by methylxanthines: evidence and proposed
mechanisms. Acta Physiologica Scandinavica 156,
347–353.
MacIntosh, B.R. & Wright, B.M. (1995) Caffeine ingestion and performance of a 1500-metre swim. Canadian Journal of Applied Physiology 20, 168–177.
McNaughton, L. (1987) Two levels of caffeine ingestion
on blood lactate and free fatty acid responses during
incremental exercise. Research Quarterly of Exercise
and Sport 58, 255–259.
Nehlig, A. & Debry, G. (1994) Caffeine and sports
activity: a review. International Journal of Sports
Medicine 15, 215–223.
Nehlig, A., Daval, J.-L. & Debry, G. (1992) Caffeine and
the central nervous system: mechanisms of action,
biochemical, metabolic, and psychostimulant effects.
Brain Research Reviews 17, 139–170.
Pasman, W.J., VanBaak, M.A., Jeukendrup, A.E. &
DeHaan, A. (1995) The effect of different dosages of
caffeine on endurance performance time. International Journal of Sports Medicine 16, 225–230.
Snyder, S.H. (1984) Adenosine as a mediator of the
behavioral effects of xanthines. In Caffeine (ed. P.B.
Dews), pp. 129–141. Springer-Verlag, Berlin.
Spriet, L.L. (1995) Caffeine and performance. International Journal of Sports Nutrition 5, S84–S99.
Spriet, L.L. (1997) Ergogenic aids: recent advances and
retreats. In Recent Advances in the Science and Medicine
of Sport, Vol. 10 (ed. D.R. Lamb & R. Murray), pp.
186–238. Cooper Publishing, Carmel.
Spriet, L.L. & Dyck, D.J. (1996) The glucose-fatty acid
cycle in skeletal muscle at rest and during exercise. In
Biochemistry of Exercise IX (ed. R.J. Maughan &
S.M. Shirreffs), pp. 127–155. Human Kinetics,
Champaign, IL.
Spriet, L.L., MacLean, D.A., Dyck, D.J., Hultman, E.,
392
nutrition and exercise
Cederblad, G. & Graham, T.E. (1992) Caffeine ingestion and muscle metabolism during prolonged exercise in humans. American Journal of Physiology 262
(Endocrinology Metabolism), E891–E898.
Tarnopolsky, M.A. (1994) Caffeine and endurance performance. Sports Medicine 18, 109–125.
Trice, I. & Haymes, E.M. (1995) Effects of caffeine
ingestion on exercise-induced changes during highintensity, intermittent exercise. International Journal
of Sports Nutrition 5, 37–44.
Van Handel, P. (1983) Caffeine. In Ergogenic Aids in
Sport (ed. M.H. Williams), pp. 128–163. Human
Kinetics, Champaign, IL.
Van Soeren, M.H. & Graham, T.E. (1998) Effect of caffeine on metabolism, exercise endurance, and catecholamine responses after withdrawal. Journal of
Applied Physiology 85, 1493–1501.
Van Soeren, M.H., Sathasivam, P., Spriet, L.L. &
Graham, T.E. (1993) Caffeine metabolism and epinephrine responses during exercise in users and
non-users. Journal of Applied Physiology 75, 805–812.
Van Soeren, M.H., Mohr, T., Kjaer, M. & Graham, T.E.
(1996) Acute effects of caffeine ingestion at rest in
humans with impaired epinephrine responses.
Journal of Applied Physiology 80, 999–1005.
Vergauwen, L., Hespel, P. & Richter, E.A. (1994) Adeno-
sine receptors mediate synergistic stimulation of
glucose uptake and transport by insulin and by
contractions in rat skeletal muscle. Journal of Clinical
Investigation 93, 974–981.
Waldeck, B. (1973) Sensitization by caffeine of central
catecholamine receptors. Journal of Neural
Transmission 34, 61–72.
Weir, J., Noakes, T.D., Myburgh, K. & Adams, B. (1987)
A high carbohydrate diet negates the metabolic effect
of caffeine during exercise. Medicine and Science in
Sports Exercise 19, 100–105.
Wemple, R.D., Lamb, D.R. & McKeever, K.H. (1997)
Caffeine vs. caffeine-free sports drinks: effects on
urine production at rest and during prolonged exercise. International Journal of Sports Medicine 18, 40–46.
Wilcox, A.R. (1990) Caffeine and endurance performance. In Sports Science Exchange, pp. 1–5. Gatorade
Sports Science Institute, Barrington, IL.
Wiles, J.D., Bird, S.R., Hopkins, J. & Riley, M. (1992)
Effect of caffeinated coffee on running speed, respiratory factors, blood lactate and perceived exertion
during 1500-m treadmill running. British Journal of
Sports Medicine 26, 116–120.
Williams, J.H., Signoille, J.F., Barnes, W.S. & Henrich,
T.W. (1988) Caffeine, maximal power output and
fatigue. British Journal of Sports Medicine 229, 132–134.