J Appl Physiol 94: 1557–1562, 2003.
First published December 13, 2002; 10.1152/japplphysiol.00911.2002.
Effect of a divided caffeine dose on endurance cycling
performance, postexercise urinary caffeine concentration,
and plasma paraxanthine
Kylie J. Conway, Rhonda Orr, and Stephen R. Stannard
School of Exercise and Sport Science, University of Sydney, Lidcombe 1825, New South Wales, Australia
Submitted 3 October 2002; accepted in final form 2 December 2002
urine; doping
CAFFEINE IS ONE OF THE MOST widely consumed drugs in
the world and is known to be ingested by sportspeople
to augment performance. Ample laboratory-based (6, 8,
13–16, 19, 21, 25) and some field-based evidence (3, 23)
demonstrate the beneficial effects of caffeine on endurance exercise performance. However, few reports document the incidence of its use as an ergogenic aid in
the athletic population and its ideal pattern of delivery.
There is some evidence that caffeine-containing cola
drinks are commonly consumed by endurance cyclists
during their event, particularly toward the latter
stages (24). Sport or “energy” drinks and carbohydrate
gels that contain caffeine (or caffeine analogs such as
guarana) have recently become widely available. Such
products are designed for consumption both before and
during endurance exercise.
Address for reprint requests and other correspondence: R. Orr,
School of Exercise and Sport Science, University of Sydney, PO Box 170,
Lidcombe 1825, New South Wales, Australia ([email protected]).
http://www.jap.org
The majority of laboratory studies demonstrating
the ergogenic effect of caffeine ingestion on performance administer caffeine as a single dose 1 h before
exercise (3, 6, 12–15, 23, 25). This is largely to ensure
a peak plasma concentration during exercise. However, it is unknown whether this is the optimal timing
of caffeine administration to maximize its ergogenic
effect. Few studies have measured the plasma concentration of caffeine during exercise or examined its
variability in exercising subjects.
To date, five studies have investigated the effects of
repeated caffeine administration during exercise (7,
19, 21, 26, 29), and not all observed a performanceenhancing effect. Of these, only two have employed a
self-paced time trial, a protocol more reliable (20) and
representative of competition, as an endurance performance measure (7, 21). Thus there is limited available
information on the effects of the timing of caffeine
ingestion during simulated competition.
As a central nervous system stimulant, caffeine is
classed as a prohibited substance by the International
Olympic Committee (IOC) and other sporting bodies.
However, because it is present in many commonly
ingested foodstuffs, 12 ␮g/ml is the permissible postexercise urinary threshold under which no doping offense
is recorded. Performance-enhancing effects have been
observed with postexercise urinary levels well below
this threshold (14, 25), so it is uncommon for a positive
result to be registered (9). There is substantial intersubject variability in the metabolism and elimination
of caffeine (4, 21, 25), particularly during exercise (11).
Thus urinary caffeine concentration as a marker of
caffeine consumption may not accurately reflect dose or
plasma levels.
Both Cox et al. (7) and Kovacs et al. (21) measured
postexercise urinary caffeine concentrations after caffeine ingestion during exercise. The former study
showed no effect of splitting the dose on urine concentrations, but the latter observed lower concentrations
(2.5 ␮g/ml) after a split dose than previous studies [5.8
␮g/ml, 6.1 ␮g/ml (28), and 4.8 ␮g/ml (25)] with a bolus
dose. It is possible that splitting the caffeine dose
The costs of publication of this article were defrayed in part by the
payment of page charges. The article must therefore be hereby
marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
8750-7587/03 $5.00 Copyright © 2003 the American Physiological Society
1557
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 14, 2017
Conway, Kylie J., Rhonda Orr, and Stephen R. Stannard. Effect of a divided caffeine dose on endurance cycling
performance, postexercise urinary caffeine concentration,
and plasma paraxanthine. J Appl Physiol 94: 1557–1562, 2003.
First published December 13, 2002; 10.1152/japplphysiol.
00911.2002.—This study compared the effects of a single and
divided dose of caffeine on endurance performance and on
postexercise urinary caffeine and plasma paraxanthine concentrations. Nine male cyclists and triathletes cycled for 90
min at 68% of maximal oxygen uptake, followed by a selfpaced time trial (work equivalent to 80% of maximal oxygen
uptake workload over 30 min) with three randomized, balanced, and double-blind interventions: 1) placebo 60 min
before and 45 min into exercise (PP); 2) single caffeine dose (6
mg/kg) 60 min before exercise and placebo 45 min into exercise (CP); and 3) divided caffeine dose (3 mg/kg) 60 min before
and 45 min into exercise (CC). Time trial performance was
unchanged with caffeine ingestion (P ⫽ 0.08), but it tended to
be faster in the caffeine trials (CP: 24.2 min and CC: 23.4
min) compared with placebo (PP: 28.3 min). Postexercise
urinary caffeine concentration was significantly lower in CC
(3.8 ␮g/ml) compared with CP (6.8 ␮g/ml). Plasma paraxanthine increased in a dose-dependent fashion and did not peak
during exercise. In conclusion, dividing a caffeine dose provides no ergogenic effect over a bolus dose but reduces postexercise urinary concentration.
1558
EFFECTS OF A SPLIT VS. DIVIDED DOSE OF CAFFEINE
delays the appearance in the urine but still provides a
similar ergogenic effect to a bolus dose.
Paraxanthine, the primary metabolite of caffeine,
accounts for ⬎80% of caffeine degradation (22). It is
pharmacologically active as an adenosine-receptor antagonist and thus potentially ergogenic (2, 17, 18, 27),
presenting difficulties in determining whether caffeine
alone is responsible for the effects on exercise performance. Little is known about paraxanthine kinetics
during exercise.
The purpose of this study was threefold: 1) to compare the ergogenic effects of caffeine administration
before exercise with administration before and during
simulated endurance cycling performance, 2) to test
the hypothesis that dividing the dose would reduce the
postexercise urinary concentration, and 3) to observe
the plasma kinetics of paraxanthine during exercise
after the single and divided dose.
subsequent tests. Food and training diaries were collected for
the 48 h before presentation. Written and verbal reminders
were given to ensure compliance.
At the start of each test, a catheter was inserted in the
dorsal aspect of the hand. The catheter was kept patent by
flushing with saline every 15 min. A resting blood sample
(⫺60 min) was obtained. The test dose (caffeine or placebo)
was administered, and the subject remained seated for 1 h.
Further samples were collected after 30 min of rest (⫺30
min) and just before the beginning of exercise (0 min).
Exercise consisted of a 5 min warm-up at a workload of
half that calculated to elicit an oxygen uptake of 70% V̇O2 max.
Subjects then cycled for 90 min at the 70% V̇O2 max workload.
During this first 90 min, the ergometer was set in hyperbolic
mode so that the work rates were independent of cadence. At
90 min, the ergometer automatically changed to linear mode
(workload was proportional to cadence in a relationship defined by the linear factor) and 30-min self-paced time trial
was performed. The linear factor was calculated according to
the formula
Ẇ ⫽ L䡠(rpm) 2
METHODS
J Appl Physiol • VOL
where L is the linear factor and was determined in a way
where Ẇ represented the work rate eliciting 80% V̇O2 max and
rpm was the average pedaling rate in the final stages of the
V̇O2 max test.
The time trial required the subject to complete a target
amount of work (i.e., that eliciting 80% of V̇O2 max for 30 min)
as quickly as possible. Subjects were aware of the end point
and were verbally encouraged to complete the time trial as
quickly as possible.
During all tests, the ambient temperature was 22°C and
the relative humidity varied between 50 and 60%. A constant
fan speed was provided at all times during exercise, and the
same self-selected hydration protocol was used during each
test.
Blood sampling. Further samples were obtained at 30, 45,
60, and 90 min of exercise and at the completion of performance. Blood was drawn into a 5-ml syringe and divided
between a lithium-heparinized tube (2 ml) and a fluoridated
tube (2 ml). The tubes were immediately inverted, placed on
ice, and later centrifuged (model T J-6 centrifuge, Beckman),
and the serum was aliquoted into plastic tubes and frozen at
⫺20°C.
Analyses. Serum caffeine and paraxanthine levels were
determined by using fully automated HPLC (model 510,
Waters, Milford, MA). Samples were deproteinized by mixing
100 ␮l of 0.8 M perchloric acid with 100 ␮l of serum. After
vortex mixing, the proteins were removed by centrifugation
at 14,000 g (room temperature) for 3–4 min. A 125-␮l aliquot
of the supernatant was neutralized with 10.7 ␮l of 4 M
sodium hydroxide. Samples were centrifuged for 5 min at
2,000 g and then transferred to plastic 200-␮g tubes, respun
for 5 min at 2,000 g to remove air bubbles, and placed in the
autosampler. The method involved a dilution factor of 2.17.
The deproteinized sample (100 ␮l) was injected by autoinjector and eluted isocratically with the elution buffer [potassium phosphate (3.47 g)-methanol (300 ml)] for 20 min. The
mobile phase was filtered through a 0.45-␮m filter and degassed under vacuum before use. Eluted peaks were detected
by ultraviolet absorbance at 274 nm, and peak areas were
used for quantitation by using a four-point standard curve.
The elution time for caffeine was 7.7 min and for paraxanthine was 3.3 min. The temperature was ambient (range
21–24°C), and the flow rate was constant at 1.2 ml/min.
Standards were run in duplicate at the beginning of and
halfway through each assay.
94 • APRIL 2003 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 14, 2017
Subjects. Nine healthy and well-trained cyclists and triathletes agreed to participate in the study. The subjects [age
25.5 ⫾ 5 (SE) yr, weight 76.4 ⫾ 6.9 kg, maximal oxygen
uptake (V̇O2 max) 5.5 ⫾ 0.3 l/min] were nonsmokers. Each
subject was fully informed about the experimental procedures and possible risks before giving informed, written
consent. The protocol was approved by the Human Ethics
Committee of The University of Sydney.
Preexperimental protocol. Subjects reported to the laboratory before the start of the experiment for an incremental test
to measure their submaximal oxygen uptake and V̇O2 max on
an electronically braked cycling ergometer (Lode Excalibur
Sport, Groningen, The Netherlands). All subsequent exercise
tests were done on the same ergometer. The submaximal test
required each subject to cycle in a stepwise fashion for 8 min
at four submaximal workloads (100, 175, 250, and 325 W) at
a constant, self-selected cadence. The V̇O2 max test followed,
beginning 5 min after the last submaximal workload, and
required a linear increase in power from 0 W at 1 W/s until
volitional fatigue. A relationship between oxygen consumption in the last minute of each submaximal workload and
power was developed and used to calculate a workload equivalent to both 70 and 80% of V̇O2 max.
Experimental protocol. The study was a randomized, double-blind, placebo-controlled, balanced factorial trial. Subjects were tested on three occasions where gelatine capsules
containing either lactose (placebo) or caffeine were administered 1 h before exercise and 45 min into exercise. The
treatments administered in a double-blind manner were
placebo-placebo (PP), single caffeine dose (6 mg/kg)-placebo
(CP), and caffeine (3 mg/kg)-caffeine (3 mg/kg) (CC). The
subjects were assigned randomly to one of three test sequences (PP-CP-CC, CP-CC-PP, or CC-PP-CP), each containing three subjects. After the test, subjects were asked
whether they could determine which treatment they had
received. The tests were administered on the same day of the
week, with an interval of 7 days between tests.
Subjects arrived at the laboratory after an overnight (12 h)
fast. They were encouraged to drink adequate water on the
previous evening and on arising to ensure full hydration.
Subjects were asked to refrain from ingesting products containing caffeine for 48 h before presentation and to refrain
from heavy exercise 24 h before testing. To minimize variation in preexercise muscle glycogen status, subjects were
required to repeat the same training and food selection for
EFFECTS OF A SPLIT VS. DIVIDED DOSE OF CAFFEINE
1559
Urine samples were collected at the beginning and end of
the test. Immediately after collection, urine samples were
stored at ⫺20°C for later analysis of caffeine by using the
HPLC system. Urine volume throughout the duration of the
test was also recorded.
Statistical analyses. Dependant variables were analyzed
by a repeated-measures ANOVA. Polynomial functions of
time were chosen in the contrasts to characterize outcome
variables measured at several distinct time points. Different
treatment groups were then compared by using the appropriate interaction terms in an ANOVA table. All statistical
analyses were performed using specialized statistical software (SPSS version 8.2). Statistical significance was accepted at P ⱕ 0.05. All data are expressed as means ⫾ SE
unless otherwise stated.
RESULTS
Fig. 1. Performance times to complete target amount of work for the
3 trials. PP, placebo 60 min before and 45 min into exercise; CP,
single caffeine dose (6 mg/kg) 60 min before exercise and placebo 45
min into exercise; CC, divided caffeine dose (3 mg/kg) 60 min before
and 45 min into exercise. Values are means ⫾ SE for 8 subjects.
J Appl Physiol • VOL
Fig. 2. Plasma caffeine (A) and paraxanthine (B) concentrations
(conc) during rest and exercise. Values are means ⫾ SE for 8
subjects. V̇O2 max, maximal oxygen uptake; TT, time trial. * Significantly lower compared with CP, P ⬍ 0.05. # Significantly different
from previous value, P ⬍ 0.05.
was present when placebo was ingested (Fig. 2A).
Plasma caffeine levels were significantly elevated at
⫺30 min (P ⫽ 0.000) and reached a peak within 90 min
after CP administration. With CC there was an initial
peak between 0 and 30 min followed by a slow decline
until the second caffeine dose when plasma caffeine
levels increased again and continued to rise until the
end of the time trial. Plasma caffeine concentration
was significantly greater in the CP compared with the
CC trial until ⬃60 min into exercise (P ⫽ 0.001), after
which the caffeine concentrations were similar (P ⫽
0.483). Maximum caffeine concentrations were 7.3 and
6.7 ␮g/ml for CP and CC, respectively.
Paraxanthine. The presence of paraxanthine in the
plasma was significantly higher in the CP compared
with the CC trial (P ⬍ 0.05). Paraxanthine concentration continued to increase after caffeine ingestion but
showed a different relationship to the previously described plasma caffeine levels. The increase in paraxanthine levels occurred at a slower rate to caffeine,
and no peak concentration was obtained (Fig. 2B).
Urinary caffeine content. There were no significant
differences among trials in the volume of urine produced, but it was less after exercise than before. Urine
production during the 1-h preexercise period was
228 ⫾ 21, 247 ⫾ 27, and 233 ⫾ 31 ml for PP, CP, and
CC, respectively. No caffeine was detected in urine
after PP. Mean postexercise caffeine levels for CP and
CC were 6.89 ⫾ 0.73 ␮g/ml (3.70–9.06 ␮g/ml) and
3.82 ⫾ 0.34 ␮g/ml (2.35–6.16 ␮g/ml), respectively (P ⫽
0.002; Fig. 3). Between-subject variations in postexercise urinary caffeine levels were large.
When the initial dose of caffeine (3 vs. 6 mg/kg) was
considered, there was a significant correlation between
94 • APRIL 2003 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 14, 2017
Eight of nine subjects completed all aspects of the
trial. One subject was excluded from analysis because
he was unable to complete the time trial when given
the single dose (CP) of caffeine. The subject experienced nausea and nervousness, which have both been
commonly associated with caffeine ingestion. None of
the subjects reported habitual consumption of large
amounts of caffeine (⬍250 mg caffeine/day). During the
three trials, mean body weight loss during the entire
experiment, corrected for urine output, was 1.26 ⫾ 0.6
kg, indicating no significant differences in hydration
status between trials. The mean fluid ingested during
the exercise protocol was 1,444 ⫾ 309 ml.
Time trial performance. Repeated-measures ANOVA
revealed no main effect of sequence of ingestion on
performance time (P ⫽ 0.080). However, there was a
clear trend for the time trial to be completed significantly faster after the ingestion of caffeine (average
with CP and CC of 23.8 ⫾ 2.8 min) compared with
placebo (PP: 28.3 ⫾ 3.1 min) (Fig. 1).
Plasma caffeine levels. Before each trial, subjects
showed zero levels of caffeine, confirming their compliance to abstaining from caffeine-containing products
before testing. During the experiment, the plasma caffeine levels increased in a dose-related manner after
the ingestion of the caffeine capsules, and no caffeine
1560
EFFECTS OF A SPLIT VS. DIVIDED DOSE OF CAFFEINE
Fig. 3. Postexercise urinary caffeine concentration in CP and CC
trials. Values are means ⫾ SE for 8 subjects. * Significantly lower
compared with CP, P ⬍ 0.05.
DISCUSSION
There is ample evidence in the literature that caffeine, ingested before exercise, improves endurance
exercise performance (6, 7, 8, 13–16, 19, 25). Although
the present study does not mirror these previous findings, there was a strong trend (P ⫽ 0.080) toward an
ergogenic effect of caffeine. The mean effect size of the
caffeine treatments (0.56) indicates a worthwhile ergogenic effect of caffeine ingestion compared with placebo. However, low subject numbers and the exclusion
of data from one subject because of the possible adverse
effects of caffeine reduced the statistical power (1 ⫺
␤ ⫽ 0.442). Lack of a significant effect of caffeine on
time trial performance in the present study was thus
likely due to a type II error.
To our knowledge, only one previous study (7) has
reported the effect of caffeine given as a single dose
compared with the same amount given as a divided
dose. Others (19, 21, 29) have administered caffeine at
intervals throughout exercise, but most were not primarily concerned with examining a divided dose of
caffeine on performance and therefore did not include a
single dose for comparison. These studies also used
concomitant carbohydrate supplementation, making
comparison between the findings of the present study
with these investigations difficult. The results of the
present study, like those of Cox et al. (7), indicate that
dividing a caffeine dose for ingestion during exercise
provides no ergogenic effect on endurance cycling performance over a bolus dose given before exercise. It is
possible that the maximal ergogenic effect of caffeine
occurs with doses ⱕ3 mg/kg (7, 15). Therefore, ingestion of more caffeine before or during exercise will have
no further ergogenic effect.
J Appl Physiol • VOL
Fig. 4. Relationship between initial caffeine dose and final urinary
caffeine concentration (r ⫽ 0.735, P ⫽ 0.001; n ⫽ 8 subjects).
94 • APRIL 2003 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 14, 2017
initial caffeine intake and postexercise urinary caffeine
concentration (r ⫽ 0.735, P ⫽ 0.001; Fig. 4). A similar
relationship was observed between plasma and urinary caffeine concentrations. No relationship was
present between urine and plasma caffeine levels at
the end (120 min), whereas there was a significant
relationship between plasma caffeine levels at 45 min
and urinary caffeine levels at the completion of exercise (r ⫽ 0.774, P ⬍ 0.001).
Caffeine was rapidly absorbed into the blood, and
peak concentration occurred within 90 min after ingestion. Similar findings have been previously reported
whereby peak plasma concentration was observed to
occur ⬃75 min after caffeine ingestion (2, 4, 5). The
mean peak plasma concentration of 7.5 ␮g/ml was also
in agreement with earlier studies (2, 4, 5).
Despite a relationship between plasma caffeine levels and initial caffeine dose, we were surprised that
there was no noticeable difference in endurance performance between single and split dosing. Type II error
is the unlikely cause because of the small difference
between the two caffeine trials, and others (7) have had
similar findings.
Paraxanthine exhibits similar pharmacological and
ergogenic properties to caffeine (2, 18). Its effects on
skeletal muscle have been demonstrated in vivo to
reduce plasma potassium concentrations (27) and in
vitro to increase calcium concentration transiently to
subcontracture levels in resting skeletal muscle (17).
Paraxanthine concentrations were significantly higher
throughout the test after the single dose compared
with the divided caffeine dose; maximum levels were
1.4 and 1.1 ␮g/ml, respectively. However, with both
treatments, paraxanthine continued to increase
throughout exercise such that no peak was obtained.
Therefore, when plasma caffeine levels were decreasing, paraxanthine was still rising. These results are
consistent with the findings of earlier studies. Caffeine
doses of 2 and 4 mg/kg gave peak paraxanthine concentrations of 1.4 ␮g/ml at 300 min and 1.6 ␮g/ml at
540 min, respectively (2). In six spinal cord-injured
subjects, paraxanthine concentration increased gradually throughout 180 min without plateauing, after administration of 6 mg/kg caffeine (27). Similarly, a constant elevation of paraxanthine after caffeine ingestion
(4 mg/kg) to a peak concentration of 1.15 ␮g/ml was
measured at 180 min with no indication of a plateau
(18).
The relative contributions of caffeine and paraxanthine to performance enhancement remain unclear. It
is has been suggested that paraxanthine formation
may be quantitatively more important in having pharmacological effects than previously believed (22).
Nonetheless, no conclusions can be made to determine
EFFECTS OF A SPLIT VS. DIVIDED DOSE OF CAFFEINE
J Appl Physiol • VOL
the caffeine dose resulted in substantially lower postexercise urinary caffeine levels. This latter effect may
present a potential loophole to accurately determining
an athlete’s caffeine intake via urine analysis.
The authors thank Patricia Ruell from the School of Exercise and
Sport Science for assistance in the analysis of urinary caffeine,
plasma caffeine, and plasma paraxanthine concentrations.
REFERENCES
1. Arnaud MJ. Metabolism of caffeine and other components of
coffee. In: Caffeine, Coffee, and Health, edited by Garattini S.
New York: Raven, 1993, p. 43–143.
2. Benowitz NL. Sympathomimetic effects of paraxanthine and
caffeine in humans. Clin Pharmacol Ther 58: 684–691, 1995.
3. Berglund B and Hemmingsson P. Effects of caffeine ingestion
on exercise performance at low and high altitudes in crosscountry skiers. Int J Sports Med 3: 234–236, 1982.
4. Blanchard J. The absolute bioavailability of caffeine in man.
Eur J Clin Pharmacol 24: 93–98, 1983.
5. Collomp K. Effects of moderate exercise on the pharmacokinetics of caffeine. Eur J Clin Pharmacol 40: 279–282, 1991.
6. Costill DL, Dalsky GP, and Fink WJ. Effects of caffeine
ingestion on metabolism and exercise performance. Med Sci
Sports 10: 155–158, 1978.
7. Cox GR, Desbrow B, Montgomery PG, Anderson ME,
Bruce CR, Macrides TA, Martin DT, Moquin A, Roberts A,
Hawley JA, and Burke LM. Effect of different protocols of
caffeine intake on metabolism and endurance performance.
J Appl Physiol 93: 990–999, 2002.
8. Coyle EF. Ergogenic aids. Clin Sports Med 3: 731–742, 1984.
9. Delbeke FT. Doping in cyclism: results of unannounced controls
in Flanders (1987–1994). Int J Sports Med 17: 434–438, 1996.
10. Denaro CP. Validation of urine caffeine metabolite ratios with
use of stable isotope-labeled caffeine clearance. Clin Pharmacol
Ther 59: 284–296, 1996.
11. Duthel JM. Caffeine and sport: role of physical exercise upon
elimination. Med Sci Sports Exerc 23: 980–985, 1991.
12. Graham TE, Helge JW, MacLean DA, Kiens B, and Richter
EA. Caffeine ingestion does not alter carbohydrate or fat metabolism in human skeletal muscle during exercise. J Physiol 529:
837–847, 2000.
13. Graham TE, Hibbert E, and Sathasivam P. Metabolic and
exercise endurance effects of coffee and caffeine ingestion. J Appl
Physiol 85: 883–889, 1998.
14. Graham TE and Spriet LL. Performance and metabolic responses to a high caffeine dose during prolonged exercise. J Appl
Physiol 71: 2292–2298, 1991.
15. Graham TE and Spriet LL. Metabolic, catecholamine, and
exercise performance responses to various doses of caffeine.
J Appl Physiol 78: 867–874, 1995.
16. Greer F, Friars D, and Graham TE. Comparison of caffeine
and theophylline ingestion: exercise metabolism and endurance.
J Appl Physiol 89: 1837–1844, 2000.
17. Hawke TJ, Allen DG, and Lindinger MI. Paraxanthine, a
caffeine metabolite, dose dependently increases [Ca2⫹]i in skeletal muscle. J Appl Physiol 89: 2312–2317, 2000.
18. Hetzler RK. Effect of paraxanthine on FFA mobilization after
intravenous caffeine administration in humans. J Appl Physiol
68: 44–47, 1990.
19. Ivy JL, Costill DL, Fink WJ, and Lower RW. Influence of
caffeine and carbohydrate feedings on endurance performance.
Med Sci Sports 11: 6–11, 1979.
20. Jeukendrup A, Saris WH, Brouns F, and Kester AD. A new
validated endurance performance test. Med Sci Sports Exerc 28:
266–270, 1996.
21. Kovacs EM, Stegen J, and Brouns F. Effect of caffeinated
drinks on substrate metabolism, caffeine excretion, and performance. J Appl Physiol 85: 709–715, 1998.
22. Lelo A. Quantitative assessment of caffeine partial clearances
in man. Br J Clin Pharmacol 22: 183–186, 1986.
94 • APRIL 2003 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 14, 2017
the impact of paraxanthine during exercise in this
study.
The amount of caffeine excreted is consistent with
values reported in other studies that used caffeine
doses up to 9 mg/kg, which observed significant improvements in endurance performance with caffeine
doses that produce urinary levels well below the IOC
threshold of 12 ␮g/ml (14, 15, 21, 28). We believe this is
the first study to report a lower postexercise urinary
caffeine concentration of divided-dose caffeine administration as opposed to single-dose administration. The
mean caffeine level present in the urine was almost
twofold greater in the single-dose trial compared with
divided-dose trial. These results support those of Kovacs et al. (21) that lower urinary caffeine levels were
due to spreading the dose throughout the duration of
exercise.
Because total caffeine intake was identical in both
trials, the timing of administration must have influenced urine concentration. This is reflected by the
significant positive correlation between initial caffeine
dose and the final urine concentration (Fig. 4), suggesting that the second dose of caffeine did not influence
the urinary caffeine levels at the time of collection. A
comparison of previous observations support these
findings. Split administration of a 4.5 mg/kg dose (21)
achieved markedly lower urinary caffeine concentrations (2.5 ␮g/ml) than a similar (5 mg/kg) single dose
[4.8 ␮g/ml (25) and 5.8 and 6.1 ␮g/ml (28)].
Two important implications can be drawn from the
urine results. First, larger ergogenic doses of caffeine
may not exceed the IOC limit if the dose were divided
throughout exercise. Second, the timing of caffeine
intake and urinary sampling has a large influence on
the subsequent caffeine concentration in the urine.
Therefore, the caffeine levels present in the urine are
not always an accurate indication of total caffeine
intake. It is possible that an individual could manipulate the urinary caffeine concentration at the completion of competition, by altering the timing of caffeine
ingestion.
Our results suggest that urine is not the best method
for doping analysis of caffeine. Current urine analysis
is based solely on the 1–3% excreted as unmetabolized
caffeine and does not account for the metabolites
cleared by the kidney. Paraxanthine, like caffeine, is a
potent competitive adenosine antagonist that can contribute to the ergogenic effects of caffeine (2, 17, 27).
Furthermore, paraxanthine concentrations continue to
rise after completion of exercise (2). Although the paraxanthine concentration can be measured in the urine,
its excretion is highly variable (1) and, like caffeine,
could be influenced by event duration, individual metabolism, and environmental conditions. Other more
accurate methods, such as analysis of caffeine metabolite ratios in the urine, have been suggested (10).
In conclusion, caffeine ingestion did not significantly
alter endurance cycling performance compared with a
placebo, although it tended to have an ergogenic effect.
Furthermore, there was no significant effect of timing
of caffeine delivery on performance; however, dividing
1561
1562
EFFECTS OF A SPLIT VS. DIVIDED DOSE OF CAFFEINE
23. MacIntosh BR and Wright BM. Caffeine ingestion and performance of a 1,500-metre swim. Can J Appl Physiol 20: 168–
177, 1995.
24. Martin DT, Roussos S, Perry C, and Salzwedel H. Coca-Cola
preferred by top endurance cyclists. [Online]. Sports Science News
http://www.sportsci.org/news9711/martin.html. November-December, 1997.
25. Pasman WJ, van Baak MA, Jeukendrup AE, and de Haan
A. The effect of different dosages of caffeine on endurance performance time. Int J Sports Med 16: 225–230, 1995.
26. Sasaki H, Takaoka I, and Ishiko T. Effects of sucrose or
caffeine ingestion on running performance and biochemical re-
sponses to endurance running. Int J Sports Med 8: 203–207,
1987.
27. Van Soeren M, Mohr T, Kjaer M, and Graham TE. Acute
effects of caffeine ingestion at rest in humans with impaired
epinephrine responses. J Appl Physiol 80: 999–1005, 1996.
28. Van Soeren MH, Sathasivam P, Spriet LL, and Graham
TE. Caffeine metabolism and epinephrine responses during
exercise in users and nonusers. J Appl Physiol 75: 805–812,
1993.
29. Wemple RD, Lamb DR, and McKeever KH. Caffeine vs.
caffeine-free sports drinks: effects on urine production at rest
and during prolonged exercise. Int J Sports Med 18: 40–46, 1997.
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 14, 2017
J Appl Physiol • VOL
94 • APRIL 2003 •
www.jap.org