See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/23669680

Caffeine and sport performance

Article  in  Applied Physiology Nutrition and Metabolism · January 2009

DOI: 10.1139/H08-130 · Source: PubMed

CITATIONS READS
244 14,185

1 author:

Louise Mary Burke

Australian Institute of Sport (AIS)
336 PUBLICATIONS   13,838 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

periodisation of the CHO intake View project

The Stay Healthy Project View project

All content following this page was uploaded by Louise Mary Burke on 31 May 2014.

The user has requested enhancement of the downloaded file.

 1319

SYMPOSIUM / SYMPOSIUM

Caffeine and sports performance

Louise M. Burke

Abstract: Athletes are among the groups of people who are interested in the effects of caffeine on endurance and exercise
capacity. Although many studies have investigated the effect of caffeine ingestion on exercise, not all are suited to draw
conclusions regarding caffeine and sports performance. Characteristics of studies that can better explore the issues of ath-
letes include the use of well-trained subjects, conditions that reflect actual practices in sport, and exercise protocols that
simulate real-life events. There is a scarcity of field-based studies and investigations involving elite performers. Research-
ers are encouraged to use statistical analyses that consider the magnitude of changes, and to establish whether these are
meaningful to the outcome of sport. The available literature that follows such guidelines suggests that performance benefits
can be seen with moderate amounts (~3 mg
kg–1 body mass) of caffeine. Furthermore, these benefits are likely to occur
across a range of sports, including endurance events, stop-and-go events (e.g., team and racquet sports), and sports involv-
ing sustained high-intensity activity lasting from 1–60 min (e.g., swimming, rowing, and middle and distance running
races). The direct effects on single events involving strength and power, such as lifts, throws, and sprints, are unclear. Fur-
ther studies are needed to better elucidate the range of protocols (timing and amount of doses) that produce benefits and
the range of sports to which these may apply. Individual responses, the politics of sport, and the effects of caffeine on
other goals, such as sleep, hydration, and refuelling, also need to be considered.
Key words: ergogenic aid, sports performance, doping.
Résumé : Les athlètes font partie des gens concernés par les effets de la caféine sur l’endurance et la capacité physique.
Même si de nombreuses études ont porté sur les effets de la consommation de la caféine sur l’exercice physique, elles ne
permettent pas toutes de tirer des conclusions au sujet des effets de la caféine sur la performance sportive. Pour analyser
de tels effets, il faut des études incluant des sujets bien entraı̂nés, des conditions reflétant les pratiques sportives en cours
et des protocoles expérimentaux simulant des conditions réelles. Il y a très peu d’études réalisées sur le terrain qui incluent
des athlètes d’élite. On invite les chercheurs à utiliser des outils statistiques mesurant l’importance des variations notam-
ment sur le plan de la pertinence dans la pratique sportive. Les études scientifiques qui prennent en compte ces directives
rapportent qu’une quantité modérée de café (~3 mg
kg–1 de masse corporelle) suscite des gains sur le plan de la perfor-
mance. De plus, ces gains devraient se manifester dans un large spectre d’activités sportives dont les activités d’endurance,
les activités constituées d’arrêts et de départs tels les sports d’équipe et de raquette et les activités demandant une forte in-
tensité soutenue de 1 min à 60 min comme la natation, l’aviron, la course de demi-fond et de fond. Les effets directs de la
consommation de caféine dans les activités de force et de puissance tels les levers, les lancers et les sprints ne sont pas
bien établis. Il faut faire d’autres études pour bien déterminer les variétés de protocoles admissibles (moment de l’année,
quantité consommée) qui suscitent des gains et qui identifient les sports pouvant en bénéficier. Il faut aussi faire d’autres
études sur les réponses individuelles, les politiques du sport et sur les effets de la caféine sur d’autres facteurs tels le som-
meil, l’hydratation et la recharge d’énergie.
Mots-clés : facteur ergogène, performance sportive, dopage.
[Traduit par la Rédaction]

Introduction caffeine in reducing fatigue and increasing wakefulness and

alertness have been recognised for many centuries. These
Caffeine is a drug that enjoys social acceptance and wide- properties have been targeted by shiftworkers, long-haul
spread use around the world, with about 90% of adults con- truck drivers, members of the military forces, athletes, and
suming it in their everyday eating patterns. The effects of other populations who need to fight fatigue or prolong their
capacity to undertake their occupational activities. Indeed,
Received 26 July 2008. Accepted 14 October 2008. Published on the availability, in many countries, of nonprescription medi-
the NRC Research Press Web site at apnm.nrc.ca on cations, energy drinks, confectionary and sports foods, and
6 December 2008. (or) supplements that contain caffeine or guarana (Table 1)
L.M. Burke. Department of Sports Nutrition, Australian has increased the opportunities for people to specifically
Institute of Sport, P.O. Box 176, Belconnen, ACT, Canberra consume caffeine as an ergogenic (work-enhancing) aid.
2616, Australia (e-mail: louise.burke@ausport.gov.au). The past 30 years has seen the publication of a substantial

Appl. Physiol. Nutr. Metab. 33: 1319–1334 (2008) doi:10.1139/H08-130 # 2008 NRC Canada
1320 Appl. Physiol. Nutr. Metab. Vol. 33, 2008

Table 1. Caffeine content of common foods, drinks, and nonprescription preparations.

Food or drink Serving Caffeine, mg*

Instant coffee 250 mL (8 oz) cup 60.(12–169){
Brewed coffee 250 mL (8 oz) cup 80.(40–110){
Short black coffee or espresso 1 standard serving 107.(25–214){
Starbucks Breakfast Blend brewed coffee (Venti size) 600 mL (20 oz) 415.(300–564)§
Iced coffee (commercial brands) 500 mL bottle (16 oz) 30.–200
Frappuccino 375 mL (12 oz) cup 90.
Tea 250 mL (8 oz) cup 27.(9–51){
Iced tea 600 mL (20 oz) bottle 20.–40
Hot chocolate 250 mL (8 oz) cup 5.–10
Chocolate milk 60 g 5.–15
Dark chocolate 60 g 10.–50
Viking chocolate bar 60 g 58.
Coca-Cola 375 mL (12 oz) can 49.
Pepsi cola 1375 mL (12 oz) can 40.
Jolt soft drink 1375 mL (12 oz) can 75.
Red Bull energy drink 250 mL (8 oz) can 80.
Red Eye Power energy drink 250 mL (8 oz) can 50.
V Energy drink 250 mL (8 oz) can 50.
Smart Drink - Brain fuel 250 mL (8 oz) can 80.
Lift Plus energy drink 250 mL (8 oz) can 36.
Lipovitan energy drink 250 mL (8 oz) can 50.
Black Stallion energy drink 250 mL (8 oz) can 80.
AMP Energy tallboy 500 mL (16 oz) can 143.
Spike Shotgun energy drink 500 mL (16 oz) can 350.
Fixx energy drink 600 mL (20 oz) can 500.
Ammo energy shot 30 g (1 oz) 170.
Jolt endurance shot 60 g (2 oz) 150.
PowerBar caffeinated sports gel 40 g sachet 25.
PowerBar double caffeinated sports gel 40 g sachet 50.
GU caffeinated sport gel 32 g sachet 20.
Carboshotz caffeinated sports gel 50 g sachet 80.
PB Speed sports gel 35 g sachet 40.
PowerBar Acticaf Performance bar 65 g bar 50.
Jolt caffeinated gum 1 stick 33.
No-Doz (Australia) 1 tablet 100.
No-Doz (U.S.) 1 tablet 200.
Extra Etrength Excedrin 1 tablet 65.
*These values were gathered from a variety of sources, including manufacturers’ information and nutrition databases (Centre
for Science in the Public Interest (available at http://www.cspinet.org/new/cafchart.htm), and USDA National Nutrient Database
(available at http://www.nal.usda.gov/fnic/foodcomp/search/). Note that commercial brands may vary slightly from country to
country.
{
The caffeine content of tea and coffee varies widely, depending on the brand, the way the beverage is made, and the size of the
mug or cup.
{
Commercial samples bought from a variety of outlets (Desbrow et al. 2007).
§
Commercial samples bought from the same Starbucks outlet (McCusker et al. 2003).

number of studies of caffeine supplementation and exercise pressed in this paper only apply to adult athletes who
or physical activity. Table 2 provides a summary of our cur- already consume caffeine within their normal dietary practi-
rent knowledge about the effects of caffeine on exercise ca- ces. This author believes it is inappropriate and unnecessary
pacity or performance from this robust literature. It is for children and young adults to consume caffeine as an er-
beyond the scope of this paper to examine the mechanisms gogenic aid, and notes that younger populations have the po-
by which caffeine exerts its ergogenic effects related to ex- tential for greater performance enhancement through
ercise on the body; readers are referred to several extensive maturation and experience in their sport. Caffeine use in all
reviews for this information (Graham 2001a, 2001b, 2008; populations should be seen against the background of its ef-
Jones 2008; Keisler and Armsey 2006; Tarnopolsky 2008). fects on human health, where it has been suggested that, in
Instead, the aim of this paper is to discuss the available in- healthy adult populations, moderate daily caffeine intakes of
formation on caffeine and exercise from the perspective of up to 400 mg
d–1 or ~6 mg
kg–1 are not associated with ad-
sports performance. It should be noted that the views ex- verse effects, whereas children aged 12 or under should
# 2008 NRC Canada
Burke 1321

Table 2. A summary of our current understanding of the effect of caffeine supplementation on exercise capacity or performance.

Issue Supporting evidence

There is sound evidence that caffeine enhances endurance and provides a small but Jones (2008); Keisler and Armsey (2006); Doh-
worthwhile enhancement of performance over a range of exercise protocols, with erty and Smith (2004); Graham (2001a, 2001b)
the traditional protocol involving a caffeine dose of ~6 mg
kg–1 body mass taken
1 h pre-exercise.
Recent studies show that beneficial effects from caffeine occur at very modest levels Bridge and Jones (2006); Cox et al. (2002);
of intake (1–3 mg
kg–1 body mass, or 70–150 mg caffeine). Kovacs et al. (1998); Graham and Spriet (1995)
Several studies suggest there is no dose–response relationship between caffeine Anderson et al. (2000); Bruce et al. (2000); Cox
intake and benefits to endurance exercise or, if a dose–response exists, there is a et al. (2002); Kovacs et al. (1998); Pasman et al.
plateau at ~3 mg
kg–1. (1995); Graham and Spriet (1995)
A variety of protocols of caffeine intake, including doses before or during exercise, Cox et al. (2002); Kovacs et al. (1998)
or after the onset of fatigue, may be beneficial for exercise capacity.
Some tissues become tolerant to repeated caffeine use, while others do not. Since the Graham (2001b)
mechanisms by which caffeine exerts effects on performance are not fully known,
it is unclear whether an athlete should withdraw from caffeine prior to competition.
Some studies show that there is no difference in the performance response between
nonusers and users of caffeine. Withdrawal can be achieved by 24–48 h of
cessation of caffeine use. However, athletes should be aware of the side effects of
withdrawal, such as headaches, and the greater potential for negative side effects
from subsequent caffeine exposure.
The effects of caffeine can be long lasting. Although there is some evidence that the Bell and McLellan (2002, 2003)
benefits do not persist up to 6 h, people who ingest caffeine to enhance a morning
exercise task may still receive benefits during a session undertaken later in the day.
Coffee may not be a good source of caffeine for exercise enhancement; it is difficult Graham et al. (1998); Graham (2001a, 2001b)
to know the dose of caffeine in any serving of coffee. Some studies have found
that while caffeine alone was ergogenic for a given exercise task, caffeine
consumed in a caffeine medium did not enhance performance of the same protocol.
Coffee may contain other ingredients that counteract the benefits of caffeine.
Nevertheless it appears that caffeine is ergogenic when coffee is consumed with a
prerace meal.
There is individual variability in the changes in exercise capacity in response to Graham and Spriet (1995)
caffeine. Some people are nonresponders to caffeine.
Higher doses of caffeine (>6–9 mg
kg–1) may be associated with side effects, such as Graham and Spriet (1995)
jitters, increased heart rate, and performance impairment.

limit their caffeine intake to <2.5 mg
kg–1
d–1 because of only apply to the population and the situation that was in-
their increased risk of behavioural side effects (Nawrot et vestigated. Therefore, Table 3 provides a list of the charac-
al. 2003). teristics that should be included in studies specifically
designed to investigate the effect of caffeine on sports per-
Special issues related to study design for formance. It is likely that there is a spectrum of athletes
who are interested in the outcomes of such research. The re-
examining caffeine and sports performance wards for achieving success in elite-level sports are highly
Research design reflects a number of scientific and practi- visible and offer a clear incentive to search for strategies
cal concerns, including the primary question (hypothesis) of that can enhance performance by even a small margin.
interest, level of funding, the availability and limitations of Theoretically, the potential for detecting small but worth-
equipment and subjects, interest in examining the mecha- while changes in performance should be greater among elite
nisms underpinning outcomes, approval from ethics panels, athletes. After all, elite athletes are usually highly reliable at
and the requirements or expectations of participants in the performing the tasks for which they have trained; a small
peer-review process. Given the range of potential uses for coefficient of variation in performance increases the preci-
any beneficial effects of caffeine on physical and occupa- sion of the separation of the true effects of an intervention
tional activities, and the diverse interests of scientists who (the signal) from everyday differences in performance (the
have undertaken studies of caffeine and exercise, we might noise).
expect that a variety of research protocols have been under- Ironically, few studies on caffeine and elite athletes are
taken. While many studies may have been able to address available. This is understandable because, by definition,
issues related to caffeine and exercise per se, not all were they are few and special. It is usually difficult to achieve a
well suited to address special issues of sports performance. large sample size of elite athletes for statistical rigour or to
Table 3 summarizes some of the characteristics of the impose the conditions or invasive techniques of research on
methodologies used in many studies of caffeine supplemen- their training and competition schedules. Therefore, most
tation and exercise, and contrasts these with the features of studies of caffeine and sports performance have been under-
real-life sport. taken on athletes at recreational to highly trained but sub-
One of the key tenets of research is that the results really elite levels (Tables 4–7). It is unclear whether the results of
# 2008 NRC Canada
 Table 3. The characteristics of traditional laboratory-based research vs. research focussed on elite sports performance.
1322

Characteristics of traditional studies Characteristics of elite and serious sport Comments on the ideal characteristics of studies on sports performance
Subjects are often drawn from available populations, Competitors are highly trained in their sport and event. Characteristics Subjects should reflect the population to which the results of the study
such as college students or recreationally and include reliability in repeating a given performance task, and are intended to apply; subjects should be familiar with and reliable in
moderately trained subjects. specific adaptations achieved though natural selection of sport and undertaking the performance protocol; and studies that involve elite
the conditioning effects of training. or highly trained athletes are underrepresented and should be
encouraged.
Exercise protocols typically measure endurance or Sports performance typically includes completing a task in the fastest Protocols should involve a close-looped task (i.e., completing a task in
exercise capacity (the ability to sustain a given possible time (pace judgement is important), executing skills and the fastest time possible), which involves pacing rather than simply
exercise task for as long as possible). The task is making complex decisions while undertaking exercise, and exercising to fatigue; protocols should simulate, as much as possible,
terminated when the subject is fatigued and unable to executing a single task as well as possible. Training situations may a real-life event; and field studies are underrepresented and should be
continue at the prescribed output. better represent the execution of an exercise task to fatigue. encouraged.
Protocols are often undertaken with baseline metabolic Athletes undertake other nutrition strategies that provide additional Protocols should allow subjects to follow nutrition strategies that
conditions (subjects fast overnight and consume only support for performance, including eating a carbohydrate-rich pre- optimize performance and reflect the real-life practices of athletes;
water during exercise) and without the confounding event meal consuming carbohydrates during prolonged events, and and studies should investigate the interaction between caffeine intake
effects of other nutritional strategies. using other scientifically supported ergogenic aids (e.g., bicarbonate, and other nutrition strategies or supplements, and their combined
creatine). effect on performance.
Differences (which must reach a level of >5% The margins between winning and losing, or between the ‘‘podium’’ Differences in performance should be assessed using magnitude-based
probability to be considered significant) in athletes (first-, second-, and third-place winners) and the rest of the statistics: the likely range of the true difference in performance
performance between the control treatment and the field, can often be measured in hundredths of seconds and meters. should be compared with the coefficient of variation of performance
active treatments(s) are assessed using probability for individuals undertaking that protocol.
statistics.
Protocols are conducted with consideration to the Sports are conducted within the regulations of their governing body, Projects that are focussed on high-level sports should be conducted
concerns of the ethics committee overseeing the and many athletes compete within an anti-doping code that may not within the logistical and anti-doping rules that govern athletes in that
project. permit the use of otherwise legal products or strategies. sport.
individual situation.

scenario in real life.

small changes that could be important.

#
of the results and, therefore, the likelihood of detecting

creatine (Doherty and Smith 2004) and bicarbonate

which tight control has allowed the effects on metabolism

(Pruscino et al. 2008), are important, since this is a likely

performance trials, and the choice of conditions that mimic
current research apply to the true elite, since top competitors

come of an event (Batterham and Hopkins 2006). In fact,

sporting activities. Table 3 lists many such factors, including

carbohydrate intake during prolonged events are helpful

ment and the athlete. This is likely to reduce the precision
the environment and course over the event, and the effect

(Cox et al. 2002). Similarly, studies examining the inter-

this will mean that top class athletes will need to extrapolate

well as selecting an appropriate subject pool, there are fea-

idity of performance measures, since they can include fea-
more studies of elite athletes, the practicalities that prevent
sport best apply is far larger than the elite athlete popula-

of competition and other extrinsic motivating factors. The

et al. 2005). The advantages of such studies relate to the val-
ies of well-trained athletes, and to test what works for their

decimal points. An emerging area in sports science is the

quired subjects to be well fuelled from the days and the
control or standardization of training and diet leading into
the choice of performance tasks that are related to sport, the
reduction in the control that is possible over the environ-
bests. As such, they will be highly interested in utilizing the
rewards are simply the satisfaction of achieving personal
els of competition can also be highly motivated, even if the

use of magnitude-based statistics, which look at the range

need to be undertaken with sympathy for effect sizes that
The analysis and interpretation of the results of studies
cent studies of caffeine and endurance sports that have re-
what happens in the world of sport. For example, some re-

the differences in performance that might change the out-

would be worthwhile to an athlete. In many sports, the mar-

of the likely true effect of an intervention in comparison to

gins between winning and losing can be measured to several
hance the reliability and validity of the results in relation to
tures that can be built into laboratory (or field) studies to en-
caffeine exists because of the laboratory-based studies in
Therefore, it is important to recognise that our interest in
negatives of field and real-life settings typically include a
pacing strategies (which are often stochastic), changes in
tures such as real airflow and ground resistance, real-life
ably interested in investigations that have been carried out
results of caffeine studies. Although it would be good to see
tion. It should also be remembered that athletes at lower lev-

meal leading into a performance trial, and have provided

Hemmingsson 1982; Cohen et al. 1996; Van Nieuwenhoven
in a field setting or in a situation of sport (Berglund and
Sports scientists who work with athletes are understand-
the information on protocols for caffeine in sport from stud-
conditioning, that have made them the best. However, we

2008 NRC Canada

action of caffeine with other proven ergogenic aids, such as
can be sure that the population base to which the results of
may have unusual characteristics, as a result of genetics or

currently available studies of caffeine and exercise and (or)

and exercise capacity or performance to be detected. As

Appl. Physiol. Nutr. Metab. Vol. 33, 2008
 Burke
Table 4. Crossover-designed studies of caffeine supplementation related to an endurance sport (>60 min).

Enhanced
Publication Subjects Caffeine intake Sports performance performance Comments
Cycling
Jenkins et al. (2008) Trained cyclists 1 mg
kg–1, 2 mg
kg–1, or 3 mg
kg–1 15 min cycling (60% Yes for 2 and Work done during the 15 min TT was
(13 males) (60 min pre-exercise) VO2 max) + 15 min TT 3 mg
kg–1 increased by 4% (1–6.8) with 2 mg
kg–1
doses; no for of caffeine and by 3% (–0.4%–6.8%)
1 mg
kg–1 dose with 3 mg
kg–1; improvement varied in
magnitude between individual cyclists
Cureton et al. Well-trained cyclists Total, 5.3 mg
kg–1; 1.2 mg
kg–1 120 min cycling (60% Yes Higher exercise intensity in 15 min TT
(2007) (16 males) pre-exercise + 0.6 mg
kg–1 every and 75% VO2 max) + with caffeine and CHO (90±11%
15 min during exercise 15 min TT; CHO-fed VO2 max) vs. CHO alone (79±14 VO2 max)
during cycling
Conway et al. Trained cyclists and 6 mg
kg–1 60 min pre-exercise; 90 min cycling at 68% Perhaps Trend to better performance in TT with
(2003) triathletes (9 males) 3 mg
kg–1 pre-exercise; and VO2 max + TT (~30 min) caffeine trials (~24.2 and 23.4 min) vs.
3 mg
kg–1 at 45 min during placebo (28.3 min) (p = 0.08); urinary
exercise caffeine concentrations lower with split
dose
Hunter et al. (2002) Highly trained 6 mg
kg–1 60 min before exercise + 100 km cycling TT, No No difference between trials with respect to
cyclists (8 males) 0.33 mg
kg–1 every 15 min including 5  1 km and total 100 km time or time to complete
4  4 km high-intensity high-intensity efforts; no difference
efforts; CHO-fed during between trials in EMG characteristics,
cycling although differences within trial
attributable to workload
Cox et al. (2002) Well-trained cyclists 6 mg
kg–1 60 min pre-exercise; 6  2 h cycling at 70% Yes at all doses Compared with placebo, caffeine in large
and triathletes 1 mg every 20 min during VO2 max + 7 kJ
kg–1 TT dose (6 mg
kg–1) provided 3%
(12 males) exercise; 10 mL
kg–1 Coca-Cola (~30 min); CHO-fed performance benefit in TT, regardless of
in last 50 min (~1–1.5 mg
kg–1 during cycling timing of intake; commercial cola drink
caffeine) consumed late in exercise (~1 mg
kg–1
caffeine) produced effects of equal
magnitude; urinary caffeine levels
~4–5 mg
mL–1 for large dose of caffeine
and <1 mg
mL–1 for cola drink
Cox et al. (2002) Well-trained cyclists Sports drink replaced during last 2 h cycling at 70% Yes Commercial cola drink consumed late in
and triathletes 70 min with 15 mL
kg–1 of a cola VO2 max + 7 kJ
kg–1 TT exercise produced 3% performance
(8 males) drink (caffeine dose ~1.5 mg
kg–1): (~30 min); CHO-fed benefit in TT compared with cola-
6% CHO; 11% CHO; 6% CHO + during cycling flavoured placebo drink. Benefits
130 mg
mL–1 caffeine; or 11% attributable to caffeine content (~2%) and
CHO + 130 mg
mL–1 caffeine* increased CHO intake (~1%)
Jacobson et al. Trained cyclists 6 mg
kg–1 (60 min pre-exercise) 2 h cycling at 70% No TT performance similar in caffeine + CHO
(2001) (8 males) VO2 max + 7 kJ
kg–1 TT trial (29.12 min) and CHO trial
(~30 min); CHO-fed (30.12 min), with both trials better than
during cycling placebo trial
#
2008 NRC Canada

Ivy et al. (1979) Trained cyclists Total dose, 500 mg; 250 mg at 2 h isokinetic cycling at Yes 7% increase in total work, compared with
(9 males + females) 60 min pre-exercise + 7 doses 80 r
min–1 placebo trial; RPE same, despite
during exercise increased work

1323
 1324
Table 4 (concluded).

Enhanced
Publication Subjects Caffeine intake Sports performance performance Comments
Kovacs et al. (1998) Well-trained cyclists 2.1 mg
kg–1, 3.2 mg
kg–1, and Cycling TT of about Yes at all doses Addition of caffeine to CHO–electrolyte
(15 males) 4.5 mg
kg–1 doses; 75 min ~1 h; CHO-fed during drinks improved 60 min TT performance;
pre-exercise and at 20 and 40 min cycling improvement with 3.2 and 4.5 mg
kg–1
during TT caffeine doses equal to and greater than
improvement with 2.1 mg
kg–1; urinary
caffeine levels related to total dose, but
all below 12 mg
mL–1
Cross-country Skiing
Berglund and Well-trained cross- 6 mg
kg–1 (prerace) 21 km cross-country ski Perhaps at low Race times were normalized to account for
Hemmingsson country skiers race (field study) at low altitude; yes at differences in weather (individual times
(1982) (14 males) and high altitudes high altitude expressed as % of mean race time); at
low altitudes, at half way, the race time
with caffeine was decreased by 0.9% of
the mean time (~33 s), compared with
placebo (p < 0.05); at full distance,
decrease was 1.7% of the mean time
(~59 s) (p < 0.1); at high altitudes, the
race time was significantly faster with
caffeine than with placebo (p < 0.001)
both after 1 lap (2.2% or ~101 s) and 2
laps (3.2% or ~152 s)
Distance running
Cohen et al. (1996) Trained runners 5 mg
kg–1, 9 mg
kg–1 (prerace) 21 km half-marathon No No effects on RPE or performance at either
(5 males + 2 (field study) dose, compared with placebo
females)
Van Nieuwenhoven Trained to well- ~1.3 mg
kg–1 in 7% CHO sport 18 km road running race No No differences in performance of whole
et al. (2005) trained runners drink vs. CHO sport drink alone + (field study); CHO-fed group between caffeinated sport drink
(90 males + 8 water (pre-exercise and at 4.5, 9, during some trials (78:03±8:42 min:s), sport drink
females) and 13.5 km during race) (78:23±8:47 min:s), or water (78:03±8:30
min:s), or for 10 fastest runners (63:41,

Appl. Physiol. Nutr. Metab. Vol. 33, 2008

63:54, and 63:50 min:s for caffeine sport
drink, sport drink, and water,
respectively)
Note: CHO, carbohydrate; EMG, electromyographical; RPE, rate of perceived exertion; TT, time trials; VO2 max, maximal oxygen consumption.
*Caffeine content equivalent to Coca-Cola.
#
2008 NRC Canada
 Burke
Table 5. Crossover-designed studies of caffeine supplementation and performance of sustained high-intensity sports (1–60 min).

Enhanced
Publication Subjects Caffeine intake Sports performance performance Comments
Middle distance and distance running
Bridge and Jones Distance runners 3 mg
kg–1 (60 min pre-exercise) 8 km race on track Yes Relative to the mean time of the control and
(2006) (8 males) placebo trials, caffeine supplementation
resulted in a 23.8 s or 1.2% improvement in
run time (p < 0.05), with individual
improvements ranging from 10 to 61 s; heart
rate was significantly higher in caffeine trial,
with trend toward lower RPE, despite faster
running speed
Wiles et al. (1992) Well-trained runners 3 g of coffee (150–200 mg of 1500 m race on treadmill Yes Mean time improved by ~4.2 s (p < 0.05) with
(18 males) caffeine) 60 min pre-exercise caffeine, compared with placebo
Wiles et al. (1992) Well-trained runners 3 g of coffee (150–200 mg 1500 m race: 1100 m at Yes Caffeine enhanced speed of 1 min final burst by
(10 males) caffeine) 60 min pre-exercise constant speed and 1 min ~0.6 km
h–1, equivalent to 10 m (p < 0.05)
final burst at self-selected
speed
Rowing
Bruce et al. (2000) Well-trained rowers 6 mg
kg–1 or 9 g
kg–1 60 min 2000 m ergometer row Yes for both Caffeine enhanced performance by a mean of
(8 males) prerace doses 1.3% and 1% for 6 mg
kg–1 and 9 g
kg–1
doses, respectively, compared with placebo
(p < 0.05); some participants had urinary
caffeine concentrations >12 ng
mL–1 with
higher caffeine dose, but participants were
unable to identify caffeine trials, suggesting
that effect is subtle
Anderson et al. Well-trained rowers 6 mg
kg–1 or 9 g
kg–1 60 min 2000 m ergometer row Yes for both Caffeine enhanced performance by a mean of
(2000) (8 females) pre-exercise doses 0.7% and 1.3% for 6 mg
kg–1 and 9 g
kg–1
doses, respectively, compared with placebo
(p < 0.05); performance improvement achieve
primarily by enhancing the first 500 m
Swimming
Burke et al. Elite and highly 2 mg
kg–1 60 min prerace 100 m race (best stroke) No, but No difference in reaction time, 50 m split, or
(unpublished ob- trained swimmers lower RPE 100 m race time between trials, but ratings of
servations) (15 males + perceived exertion was lower in the caffeine
females) trial (16.6 vs. 17.1; p = 0.01); self-reports of
sleeping patterns following the trial found that
caffeine supplementation was associated with
an increase in time taken to fall asleep and a
reduction in quality of sleep
MacIntosh and Well-trained 6 mg
kg–1 60 min prerace 1500 m freestyle race Yes 23 s improvement in swimming time with
#

Wright (1995) swimmers caffeine (p < 0.05); caffeine affected substrate

2008 NRC Canada

(11 males + and electrolyte balance

females)

1325
1326 Appl. Physiol. Nutr. Metab. Vol. 33, 2008

both 100 m races (p < 0.01), and prevented the

these performance differences are not the hundredths of sec-

Caffeine enhanced mean swimming velocity in

which also achieved practical significance in

context of real-life 1 km track cycling event
onds or millimetres that often separate competitors or a win-

Performance improved by a mean of 2.4 s or

decrease in velocity otherwise seen in the

3.1% (95% CI, 0.7% to 5.6%; p < 0.05),

ning shot from a miss in the memorable moments of sport.

second swim with the placebo treatment

Rather, the modelling of athletic performances has shown
that they are related to the coefficient of variation of per-
formers in the event (Hopkins et al. 1999). Readers are di-
rected to the work of Will Hopkins and colleagues, which
describes this new approach (Batterham and Hopkins 2006;
Hopkins et al. 1999) and provides resources with which to
undertake it (www.sportsci.org).
It is sometimes difficult to convince the reviewers of jour-
nals to accept this new approach, and some researchers try
to combine traditional probability-based statistics with a
Comments

more athlete-friendly interpretation. For example, Wiles and

colleagues (2006) undertook a laboratory-based study simu-
lating the 1 km cycling time trial in track cycling. They
found that caffeine supplementation enhanced the perform-
ance of trained cyclists by a mean of 2.4 s, or 3.1%, which
performance

achieved statistical significance. However, in highlighting

Enhanced

the relevance of these results, they noted that the 95% con-
Yes

Yes

fidence limits of this effect (the range of likely true effects

in a similar cycling population) showed a decrease in 1 km
time, ranging from 0.7% to 5.6%. To put this into context,
2  100 m swimming races,

at the mens 1 km event at the 2004 Summer Olympic

Games in Athens, the difference between the gold and silver
separated by 20 min

medal performances was 0.0185 s, or 0.3%, while the differ-

Sports performance

ence between first and tenth place was 2.39 s (Wiles et al.
1 km cycling TT

2006).
Finally, the issue of the blinding of caffeine supplementa-
tion is important to consider. A double-blinded application
of an intervention, in which neither the subject nor the re-
searcher know which treatment has been received, is consid-
ered a benchmark of study design. However, because
caffeine has effects on various body functions, subjects are
5 mg
kg–1 75 min pre-exercise

sometimes able to detect whether or not they have received

250 mg (~4 mg
kg–1) 60 min

an active treatment from clues such as changes in heart rate

or arousal. The placebo effect has been well documented in
sport. In fact, studies in which trained subjects were told
that they were receiving caffeine showed a dose-dependent
improvement in cycling time-trial performance (small im-
Caffeine intake

pre-exercise

provement when they thought they were receiving a small

caffeine dose, and larger improvement when they thought
they were receiving a large caffeine dose), even when they
actually received an inert substance on all occasions (Beedie
et al. 2006). Since the perceived benefit of a treatment may
actually allow subjects to perform better, researchers should
consider strategies to minimize or at least recognise the pos-
Trained swimmers

8 trained cyclists

sibility that the placebo effect occurs in response to the un-

(14 males +

masking of a blinded treatment. Asking subjects to rank the

order of their treatments and their performance results may
females)
Subjects

uncover whether the placebo effect could be, at least in

part, responsible for performance outcomes.

Studies of caffeine and sports performance

Table 5 (concluded).

Wiles et al. (2006)

So what does the current scientific literature say about the

Track cycling
Collomp et al.

effect of caffeine on sports performance? A summary has

Publication

been prepared that includes studies that meet as many of

(1992)

the criteria outlined in Table 3 as possible, but with particu-

lar focus on the use of trained subjects, and an exercise pro-
tocol involving a close-looped outcome rather than time to
# 2008 NRC Canada
 Burke
Table 6. Crossover-designed studies of caffeine supplementation and stop-and-go sports (team and racquet sports).

Enhanced
Publication Subjects Caffeine intake Sports performance performance Comments
Team sports
Stuart et al. (2005) Rugby union players 6 mg
kg–1 2  40 min circuits Study involved probability statistics rather than
(9 males) (70 min pre-exercise) (simulated rugby union testing of null hypothesis. Interpretation
protocol), involving included change in fatigue with caffeine,
repetitions of: compared with placebo. Mean improvements of
 20 m sprint speed Possible 0.5%–3% in performance of sprint tasks, with
 30 m sprint speed Very likely greater improvement in second half. Suggests
 Offensive sprint Likely caffeine effect achieved by reduction in fatigue.
 Defensive sprint Likely Improvement (10%) in ability to pass ball
accurately because of enhancement of arousal
 Drive 1 power Likely
or attention.
 Drive 2 power No, harm
possible
 Tackle speed Likely
 Passing ability Likely
Schneiker et al. (2006) Team athletes 6 mg
kg–1 2  36 min cycle protocol, Total work during sprints in first half was 8.5%
(10 males) (60 min pre-exercise) each involving: greater in caffeine trial than placebo, and work
18  4 s sprint with Yes in second half was 7.6% greater in caffeine trial
2 min recovery (p < 0.05 for both). Mean peak power score
achieved during sprints in first and second
halves were 7% and 6.6% greater, respectively,
in caffeine trial than in placebo trial (p < 0.05
for both).
Paton et al. (2001) Team athletes 6 mg
kg–1 10  20 m sprints on No Negligible difference between caffeine and
(16 males) (60 min pre-exercise) interval of 10 s placebo trials for time to complete 10 sprints
and decay in performance over 10 sprints.
Racquet sports
Strecker et al. (2007) Collegiate tennis 3 mg
kg–1 Skill test performed pre- Caffeine trial showed better performance of both
players (10 males) (90 min pre-exercise) exercise, 30 min, 60 min, forehand shots across the 90 min of simulated
and 90 min during tennis play. There was no difference in skill in
simulated tennis play backhand shots between trials.
against a ball machine (15
ground strokes in all 4
directions; 60 shots total):
 Forehand cross-court Yes
 Forehand up the line Yes
 Backhand cross-court No
 Backhand up the line No
#
2008 NRC Canada

1327
1328 Appl. Physiol. Nutr. Metab. Vol. 33, 2008

accuracy or success of games played during 4 h

only be expected if protocol had caused fatigue,
trial resulted in deterioration of these aspects of

different than placebo trial. No effect on hitting

fatigue. There are relatively few investigations that meet

further enhance post-trial performance. Authors

quality and shuttle run speed, whereas placebo

No effect of caffeine supplementation on tennis-

and this did not occur because of CHO intake.

players had greater success during tennis play

performance. Caffeine added to CHO did not
these criteria in the larger literature on caffeine and exercise

However, it is also possible that effect could

CHO trial resulted in maintenance of stroke (Doherty and Smith 2004); these have been grouped accord-

with male participants. However, female

specific running speed; caffeine trial not
suggest that caffeine dose was too high.
ing to the characteristics of endurance (Table 4), sustained
high-intensity (Table 5), stop-and-go (Table 6), and strength
and (or) power (Table 7) sports.
The majority of the studies of caffeine and athletic per-

on caffeine than on placebo.

formance concern endurance sports, including running, cy-
cling, and cross-country skiing events. These provide
reasonable but not unanimous support that caffeine use can
be beneficial for these activities. A variety of protocols ap-
pear to be useful, including intake before and during the
event, and investigations using relatively low doses of
Comments

caffeine (2–3 mg
kg–1) have been more prevalent in these

sports. There is also evidence that caffeine can enhance
performance of sustained high-intensity events (lasting 1–
20 min) in running, cycling, swimming, and rowing. It is
more difficult to find clear support for benefits to the per-
performance

males, no
formance of work patterns and skills activities within team
Enhanced

Females,

and racquet sports. Whether the effect of caffeine is smaller

yes;

or absent in these sports, or whether the current studies are

No

No

No

confounded by problems in the reliability or validity of pro-

tocols, requires further examination. Finally, there is a
dearth of studies involving strength or power, such as true
 70 m shuttle run (CHO

 6  15 min sprint with

Test) measuring stroke

and speed undertaken at

success during games

sprints, lifts, and throws; therefore, there is currently a lack

Tests undertaken pre and

 Hitting accuracy and

150 min); tests of skill
Tennis Performance

4 h singles tennis (with

consumed in some
 Skills test (Leuven
post 2 h match play:

of information on the effect of caffeine on such events.

30 min break after
Sports performance

30 min rest

Other effects of caffeine related to sports

end of 4 h:

nutrition goals
quality

trials)

The previous section examined the effects of specific sup-

plementation with caffeine on the performance of a single
bout of exercise, with the most immediate application being
the outcome of a competitive event. However, any summary
2 h; separate trials for
0.75 mg
kg–1
h–1 over

of the effects of caffeine for an athlete should be widened to

260 mg for females
1 h pre-exercise +

consider issues of training and recovery. It is beyond the

(~4–4.5 mg
kg–1)
caffeine + CHO,

364 mg for males,

CHO only, and

scope of this paper, and indeed the scientific literature, to

Caffeine intake

consider the effect of caffeine on repeated bouts of exercise

within a single session or from day to day (i.e., the training
5 mg
kg–1

placebo

scenario). Nevertheless, it is reasonable to speculate that the

habitual social intake of caffeine, or its specific use in rela-
tion to a training session, may have some benefits in
promoting endurance — that is, prolonging the athlete’s ca-
players (8 males + 8

pacity to undertake the physical and mental components of

players (13 males)
Well-trained tennis

Competitive tennis

their workout. Therefore, caffeine use may indirectly en-

hance competition performance by allowing the athlete to
train hard.
females)

There are several other effects of caffeine that may make

Subjects

an indirect contribution to or may impair sports perform-

ance. First, the effect of caffeine on promoting wakefulness
or interfering with sleep must be considered in the practical
context of sport. In many sports, the outcome of a competi-
tion is decided through a series of races or games spread
Vergauwen et al. (1998)

over days — for example, heats and semi-finals before the

Ferrauti et al. (1997)
Table 6 (concluded).

final event, or the schedule of matches in a tournament. In

studies of exercise in the military context, caffeine has been
shown to combat the effects of sleep deprivation on the per-
Publication

formance of mentally and physically challenging tasks.

However, there are no studies that satisfactorily consider
that caffeine use by athletes in competition could potentially
contribute to sleep deprivation after the first event and im-
# 2008 NRC Canada
Burke 1329

Table 7. Studies of caffeine supplementation and performance of power events (throws, lifts, sprints <20 s).

Enhanced
Publication Subjects Caffeine intake Sports performance performance Comments
Astorino et al. 22 resistance- 6 mg
kg–1 1 RM bench press No No changes in strength of lower or
(2008) trained males (60 min pre- upper body with caffeine
exercise)
1 RM leg press No
Beck et al. 37 resistance- 6 mg
kg–1 1 RM bench press Yes Caffeine supplement group showed
(2006) trained males; (60 min pre- a 2% (2 kg) increase in upper
parallel-group exercise) body strength (1 RM bench press)
design (trials following treatment, but no
48 h apart) change in placebo group; there
were no differences in lower body
strength in either group
1 RM leg extension No
Note: RM, repetition maximum.

pair the performance of subsequent events. It would be val- strategies to accommodate caffeine use or to avoid otherwise
uable if future studies of caffeine and performance of an ex- successful caffeine supplementation strategies in hot weather
ercise task examined whether caffeine doses that are found or other dehydrating environments.
to be ergogenic also affect the quality and duration of sleep Finally, caffeine is known to have a range of apparently
during the night following the exercise task. More sophisti- contradictory effects on carbohydrate metabolism, including
cated studies are needed to measure the carryover effect of short-term impairment of insulin-mediated glucose disposal
caffeine-related impairment of sleep on subsequent perform- in response to an acute dose at rest, along with an ap-
ance. This would be an important issue to study because parently protective effect (at least for coffee consumption)
there is at least anecdotal evidence that some athletes resort on the development of type II diabetes (for review, see van
to a cycle of caffeine supplementation followed by use of Dam and Hu 2005). One outcome of an effect of caffeine on
sleeping agents during a multiday competition. Until such glucose disposal would be to impair the synthesis of muscle
studies can be undertaken, it would seem prudent for scien- glycogen, a key element of recovery after prolonged or
tists and athletes to look for the smallest dose of caffeine high-intensity exercise. However, Battram and colleagues
that is ergogenic for sports performance. (2004) found that the intake of 6 mg
kg–1 of caffeine before
Another indirect way in which caffeine supplementation and during glycogen-depleting exercise did not affect the
can affect sports performance is through its impact on hy- rate of glycogen synthesis during the 5 h of recovery when
dration status during exercise or in the recovery between ex- adequate amounts of carbohydrate were consumed. There
ercise bouts. Acute intake of caffeine is known to have a has been recent attention directed to the results of a study
diuretic effect — that is, to increase urine excretion. Indeed, reporting enhanced muscle glycogen resynthesis following
common education messages regarding caffeine include ad- glycogen-depleting exercise in well-trained subjects. In that
vice to limit caffeine intake in situations in which hydration study, the intake of large amounts of caffeine after exercise
is challenged (e.g., air travel) or to consume extra fluid in (8 mg
kg–1) were found to enhance the rates of muscle gly-
combination with the intake of caffeine. There are a few cogen synthesis over 4 h of recovery, by 66%, when co-
studies involving exercise and hydration that show that caf- ingested with carbohydrate (Pedersen et al. 2008). In fact,
feine can have a numeric effect on fluid losses. For exam- the rates of sustained postexercise glycogen synthesis over
ple, during recovery from exercise, the intake of caffeine this time period were among the highest reported in the lit-
from cola beverages has been shown to cause a small but erature. However, the intake of such high doses of caffeine
statistically significant increase in urine production, com- may cause side effects in some subjects or may interfere
pared with hydration with caffeine-free fluids (Gonzalez- with other aspects of recovery, such as the quality of sleep
Alonso et al. 1992). However, a recent review of caffeine (see earlier). As such, they may be impractical for use in
and hydration status found that there is little scientific evi- sport. Further studies are needed to examine whether this ef-
dence that caffeine intake impairs overall fluid status fect is seen at lower levels of caffeine intake.
(Armstrong 2002). That report concluded that the effect of
caffeine on diuresis is overstated and may be minimal in
The politics of caffeine in sport
people who are habitual caffeine users. In fact, many studies
that have examined caffeine supplementation and fluid bal- The 1984 Summer Olympic Games in Los Angeles saw
ance have found that doses of caffeine that are within the the introduction of an anti-doping program by the Interna-
range proven to be ergogenic do not alter sweat rates, urine tional Olympic Committee, involving the testing of a single
losses, or indices of hydration status during exercise urine sample collected after an event for the absence or
(Millard-Stafford et al. 2007; Wemple et al. 1997). Chronic presence of items described on a list of prohibited substan-
daily intakes of caffeine, or a sudden increase in caffeine in- ces. Caffeine was included on that list, with the definition
take, have also been shown not to impair body fluid balance of a doping offence being a urinary caffeine exceeding a
(Armstrong et al. 2005; Fiala et al. 2004). Therefore, it cutoff of 15 mg
mL–1. This threshold was reduced in 1985
seems that athletes do not need to alter their fluid intake to 12 mg
mL–1. The cutoff value was chosen to exclude nor-
# 2008 NRC Canada
1330 Appl. Physiol. Nutr. Metab. Vol. 33, 2008

mal or social coffee drinking (Delbeke and Debackere 1984) ces and Methods (see http://www.wada-ama.org/en/dynamic.
and to target the doses of caffeine that were being found to ch2?pageCategory.id=267). However, the code has evolved
be ergogenic in the studies of the time. Indeed, with caffeine to include the possibility of ‘‘nonanalytical violations.’’ Ath-
supplementation of up to 5–6 mg
kg–1, positive urinary caf- letes (and their support staff) can be found guilty of a dop-
feine levels are unlikely (Conway et al. 2003; Cox et al. ing offence without a positive urine or blood test. Other
2002; Kovacs et al. 1998; Pasman et al. 1995); a substantial offences include the possession or admitted use of these pro-
risk of urinary caffeine values greater than 12 mg
mL–1 does hibited substances or methods.
not occur until intakes are greater than 9 mg
kg–1 (Pasman et The code that immediately preceded the institution of
al. 1995). It is unclear whether this ban was primarily re- the first WADA Code, the 2003 Olympic Movement Anti-
lated to safety concerns over intakes of very large doses of Doping Code, included caffeine within the category of
caffeine or the ethics of achieving performance advantages stimulants banned in competition, with an explanatory com-
through caffeine use. In any case, there were relatively few ment that ‘‘for caffeine the definition of a positive is a con-
cases of positive doping outcomes for caffeine use among centration in urine greater than 12 mg/mL.’’ (International
elite athletes over the subsequent decade. Olympic Committee 2003) However, there are several inter-
In the new millennium, the landscape of caffeine in sport pretations of the wording of this code. It could mean that
has changed markedly. First, there is greater awareness of caffeine is a prohibited substance with the collary. Further-
the frailty of urinary caffeine concentrations as a marker of more, it could mean that a urinary caffeine concentration >
caffeine use. Urinary concentration reflects the small amount 12 mg
mL–1 could serve as a reporting limit, and that all ob-
(~1%) of plasma caffeine that escapes metabolism and is ex- served or admitted uses of caffeine would constitute a
creted unchanged. Metabolic clearance of caffeine varies doping offence. Alternatively, it could mean that caffeine is
widely among athletes and among different occasions of use permitted at doses that produce urinary caffeine concentra-
by the same athlete (Birkett and Miners 1991). Urinary caf- tions < 12 mg
mL–1. These interpretations have widely dif-
feine levels are determined by a variety of factors, including ferent and far-reaching outcomes. Indeed, there are a range
the size of caffeine dose, the metabolic clearance of caf- of different issues related to the different positions that caf-
feine, and the timing of the urine sample in relation to the feine could have in an anti-doping code, many of which
caffeine dose. Since there is huge variation in urinary caf- would create considerable practical challenges if imple-
feine content for the same caffeine dose, and neither the mented (Table 8).
standardization of the time between caffeine intake and In fact, caffeine was removed from WADA’s List of Pro-
urine collection nor the prevention of opportunities to uri- hibited Substances and Methods that came into effect on
nate during or after an event, we now recognise that urinary 1 January 2004, meaning that athletes could consume
caffeine levels have no practical utility as markers of a par- caffeine, either in their background diets or for the specific
ticular use of caffeine. purposes of performance enhancement, without fear of
Second, the emerging evidence from studies over the past sanctions. However, it is currently still on the list of banned
decade is that performance benefits can be found with very drug classes of the National Collegiate Athletic Association,
modest caffeine intakes (e.g., 2–3 mg
kg–1 body mass, or the body governing college sport in the United States (http://
~100–200 mg caffeine) when caffeine is taken before and www1.ncaa.org/membership/ed_outreach/health-safety/
(or) during exercise. Furthermore, there is no evidence of a drug_testing/banned_drug_classes.pdf). Furthermore, caffeine
dose–response relationship to caffeine beyond this level of is part of the WADA monitoring program, meaning that caf-
intake — that is, performance benefits do not increase with feine concentrations are still measured in urine samples as a
increases in the caffeine dose. One of the practical outcomes means of detecting patterns of misuse in sport. This has al-
of these newer findings is that athletes no longer need to lowed some examination of the impact of the removal of caf-
practise controlled doping (i.e., finding the largest dose of feine from the prohibited list on caffeine use patterns by
caffeine that can be taken while keeping urinary caffeine athletes. Some recent studies have found a high prevalence of
levels below 12 mg
mL–1). Instead, performance benefits can caffeine use for perceived ergogenic effects among select
be found with caffeine intakes that are well within, or even groups of athletes, such as Ironman triathletes (Desbrow and
below, normal social uses. Such intakes of caffeine are Leveritt 2006) and British track and field athletes and cyclists
likely to be associated with very low urinary caffeine levels (Chester and Wojek 2008). However, measurement of over
in most athletes. In essence, there is no longer a distinction 4600 urine samples undertaken for doping control across
between normal (social) caffeine intake and caffeine intake 56 sports by a single laboratory in 2004 found no increase
that enhances performance. in the mean caffeine concentration, compared with results
Finally, there have been changes in the methods and in- from 1993–2002 (Van Thuyne and Delbeke 2006). The
tentions of the major anti-doping programs. The World mean caffeine concentration in samples in 2004 was
Anti-Doping Agency (WADA) was created in 1999 as an in- 1.12 mg
mL–1, in comparison to a finding of 1.22 mg
mL–1
dependent international organization that promotes, coordi- from over 11 000 samples collected in 1993–2002 (Van
nates, and monitors the fight against doping in sport in all Thuyne et al. 2005). The 2004 study noted differences in
its forms (www.wada-ama.org/en/index.ch2). Following caffeine use among sports, with an increased average con-
work to harmonise anti-doping policies and rules among centration and a larger percentage of higher urinary caf-
sports and authorities, it took over the anti-doping work of feine concentrations (defined as >4 mg
mL–1) in cycling
the International Olympic Committee and instituted its first and strength and (or) power sports than in other sports.
code and international standards on 1 January 2004. The Cycling showed an apparent increase in the percentage of
WADA Code is still based on a List of Prohibited Substan- higher urinary caffeine concentrations in 2004, while there
# 2008 NRC Canada
Burke 1331

Table 8. Examples of potential rulings regarding caffeine use in sport (Burke 2001).

Ruling Implications and issues

1. Caffeine is a prohibited substance in competition in absolute terms
Underpinning rationale: (i) caffeine  Athletes would not be able to consume any tea, coffee, cola drinks, chocolate, etc., prior to and
is a stimulant; (ii) caffeine intake during competition.
enhances sports performance;  Strong education messages would be needed to convey this message and its implications to
(iii) caffeine is neither a nutrient sport.
nor a necessary part of the diet.  The limit of urinary caffeine content would need to be set at very low levels. Although some
athletes would be able to consume caffeine and remain below this limit, a positive doping
offence would also be deemed to occur if the athlete was observed or admitted to consuming a
caffeine-containing product during the competition period.
 Presumably, the difficulty in removing all caffeine from the normal diet would result in a large
number of positive doping offences.
 Issues of possession and trafficking would need to consider the sale of tea, coffee, cola, and
chocolate at sporting venues, and the sponsorship of athletes, events, and sporting organizations
by companies that manufacture these products. These activities would need to be banned to be
consistent with the anti-doping code.
 Even though caffeine intake would be banned in competition, athletes would still be able to
consume caffeine during training in a manner that provides a benefit to their performance and
adaptation to the training program.
2. Caffeine is a prohibited substance only when consumed in competition settings in doses that produce urinary caffeine levels
above a certain limit (to be determined)
Underpinning rationale: aurinary  It is impossible to find a limit that distinguishes between social and intentional use of caffeine.
caffeine limit can be set that Caffeine intakes that produce a performance enhancement are indistinguishable from the
discriminates between social and caffeine intakes reported by the normal population.
intentional use of caffeine, or at  Urinary caffeine levels vary among and within individuals, and there is no standardization of the
least only picks up a few cases of collection of urine samples with regard to the timing between caffeine intake and sampling.
high caffeine use. Therefore, urinary caffeine limits do not treat caffeine use equally.
 Education messages to athletes could contain information about levels of caffeine intake that are
unlikely to produce a urinary caffeine level above this limit.
 Athletes could not investigate their urinary caffeine concentrations in relation to various levels
of intake (trying to find how much caffeine they can take without producing a positive result for
caffeine doping), since this would be regarded as controlled doping.
 Issues related to possession and trafficking of caffeine (as outlined in ruling 1) could still apply.
3. Caffeine is a prohibited substance only when intentionally consumed in competition settings in doses that produce urinary
caffeine levels above a certain limit (to be determined)
Underpinning rationale: a model As in ruling 2, except that since urinary caffeine levels vary among individuals, and there is no
similar to drinking and driving standardization of the collection of urine samples with regard to the timing between caffeine
laws will prevent accidental cases intake and sampling, athletes should be allowed (or encouraged) to investigate what intakes of
of high caffeine levels. caffeine can be tolerated without producing a positive test, without any penalty or prejudice. The
concept of controlled doping would not be applicable.
4. The prohibition on caffeine use in sport is removed
Underpinning factors: (i) caffeine is  Practical solution to overcome challenging situation.
so entrenched in the normal diet  Avoids tainting common foods and (or) drinks or common transactions in sport (e.g.,
that it is not practical to try to sponsorship of companies that produce these products) with the odium of cheating.
achieve a ban on intake; (ii) there  Research should target the smallest dose of caffeine that can produce an ergogenic benefit.
is no unfair advantage if the Education messages to athletes could promote the message that, if they want to use caffeine
majority of athletes already (socially or intentionally), small intakes produce maximal effects. Athletes would be encouraged
consume caffeine, and choose to to reduce rather than increase their intake of caffeine, thus minimizing the health implications
consume it for social reasons; and cost of using special products.
(iii) there are no health  The market for specialized sports products containing caffeine might increase, unless education
disadvantages to the intake of messages (as above) remove the perceived benefits of large doses of caffeine or the need for
small amounts of caffeine; (iv) the special food products.
ergogenic benefits of caffeine on
performance, although worthwhile,
are small — they are similar in
magnitude to the effects of
consuming CHO during an
endurance event, but less than the
beneficial effects of consuming
fluid to minimize dehydration
Note: CHO, carbohydrate.

# 2008 NRC Canada

1332 Appl. Physiol. Nutr. Metab. Vol. 33, 2008

was a decrease in this outcome in swimming and basket- gestion. Int. J. Sport Nutr. Exerc. Metab. 10: 464–475.
ball. Overall, there was a decrease in the percentage of PMID:11099373.
urine samples showing caffeine concentrations below the Armstrong, L.E. 2002. Caffeine, body fluid-electrolyte balance,
detectable range, and no increase in the percentage of sam- and exercise performance. Int. J. Sport Nutr. Exerc. Metab. 12:
ples with concentrations above 12 mg
mL–1. In fact, only 6 189–206. PMID:12187618.
samples were found with a concentration above the former Armstrong, L.E., Pumerantz, A.C., Roti, M.W., Judelson, D.A.,
cutoff level (Van Thuyne and Delbeke 2006). Watson, G., Dias, J.C., et al. 2005. Fluid, electrolyte, and renal
Further monitoring needs to take place before firm con- indices of hydration during 11 days of controlled caffeine con-
sumption. Int. J. Sport Nutr. Exerc. Metab. 15: 252–265.
clusions can be made. Nevertheless, it seems that there is lit-
PMID:16131696.
tle evidence of systematic increases in the use or misuse of
Astorino, T.A., Rohmann, R.L., and Firth, K. 2008. Effect of caffeine
caffeine at the highest levels of sport. One outcome of the ingestion on one-repetition maximum muscular strength. Eur. J.
removal of caffeine from the prohibited list is the potential Appl. Physiol. 102: 127–132. doi:10.1007/s00421-007-0557-x.
for an increase in research activities and transparent educa- PMID:17851681.
tion about the benefits and disadvantages of caffeine use in Batterham, A.M., and Hopkins, W.G. 2006. Making meaningful infer-
sport. Greater dissemination of the emerging information ences about magnitudes. Int. J. Sports Physiol. Perform. 1: 50–57.
that the benefits of caffeine occur at small to moderate Battram, D.S., Shearer, J., Robinson, D., and Graham, T.E. 2004.
doses, and of the presence of individual variability and po- Caffeine ingestion does not impede the resynthesis of proglycogen
tential side effects in response to caffeine intake, may ac- and macroglycogen after prolonged exercise and carbohydrate
tually lead to a reduction in caffeine use by athletes (e.g., supplementation in humans. J. Appl. Physiol. 96: 943–950.
lower doses being taken on fewer occasions). There is inad- doi:10.1152/japplphysiol.00745.2003. PMID:14617526.
equate information about the total caffeine intakes of ath- Beedie, C.J., Stuart, E.M., Coleman, D.A., and Foad, A.J. 2006.
letes and recreational sportspeople and their patterns of Placebo effects of caffeine on cycling performance. Med. Sci.
caffeine use. However, by choosing to withdraw caffeine in- Sports Exerc. 38: 2159–2164. doi:10.1249/01.mss.0000233805.
take before key events and by using the minimum ergogenic 56315.a9. PMID:17146324.
doses during competition or training sessions, it is possible Beck, T.W., Housh, T.J., Schmidt, R.J., Johnson, G.O., Housh,
that athletes will not consume more caffeine than the gen- D.J., Coburn, J.W.,et al. 2006. The acute effects of a caffeine-
eral population, but rather, consume it in a more targeted containing supplement on strength, muscular endurance, and
manner in relation to their sporting activities. anaerobic capabilities. J. Strength Cond. Res. 20: 506–510.
PMID:18714218.
Bell, D.G., and McLellan, T.M. 2002. Exercise endurance 1, 3, and
Summary 6 h after caffeine ingestion in caffeine users and nonusers. J.
Appl. Physiol. 93: 1227–1234. PMID:12235019.
Caffeine is widely consumed from a variety of sources as Bell, D.G., and McLellan, T.M. 2003. Effect of repeated caffeine
part of a normal diet, as well as in specialized sports foods ingestion on repeated exhaustive exercise endurance. Med. Sci.
and supplements that may be used by athletes during train- Sports Exerc. 35: 1348–1354. doi:10.1249/01.MSS.0000079071.
ing and competition. There is clear evidence that caffeine is 92647.F2. PMID:12900689.
an ergogenic aid for a variety of types of sports, although Berglund, B., and Hemmingsson, P. 1982. Effects of caffeine in-
studies involving elite athletes and field situations are lack- gestion on exercise performance at low and high altitudes in
ing. Further research is needed to define the range of caf- cross-country skiers. Int. J. Sports Med. 3: 234–236. PMID:
feine protocols and sports activities that show evidence of 7152771.
performance enhancement, as well as the benefits or harm Birkett, D.J., and Miners, J.O. 1991. Caffeine renal clearance and
to other issues underpinning recovery after exercise or prep- urine caffeine concentrations during steady state dosing Implica-
aration for an event. Newer evidence suggests, at least in en- tions for monitoring caffeine intake during sports events. Br. J.
durance sports, that the maximal benefits of caffeine are Clin. Pharmacol. 31: 405–408. PMID:2049248.
seen at small to moderate caffeine doses (2–3 mg
kg–1), Bridge, C.A., and Jones, M.A. 2006. The effect of caffeine ingestion
which are well within the normal daily caffeine intakes of on 8 km run performance in a field setting. J. Sports Sci. 24:
the general population. This makes the recent decision to re- 433–439. doi:10.1080/02640410500231496. PMID:16492607.
move caffeine from the list of prohibited substances in Bruce, C.R., Anderson, M.E., Fraser, S.F., Stepto, N.K., Klein, R.,
sports a pragmatic choice. To date, there is little evidence Hopkins, W.G., and Hawley, J.A. 2000. Enhancement of 2000-m
that this change has increased the use or misuse of caffeine rowing performance after caffeine ingestion. Med. Sci. Sports
by athletes, at least within the levels of elite and subelite Exerc. 32: 1958–1963. doi:10.1097/00005768-200011000-00021.
sport, where anti-doping codes apply. Caffeine use may also PMID:11079528.
enhance the performance of sport in recreational athletes, Burke, L.M. 2001. Report from workshop on caffeine in sport. De-
but it is inappropriate and unnecessary for use by young cember 2001. Australian Institute of Sport, Canberra, Australia.
Chester, N., and Wojek, N. 2008. Caffeine consumption amongst
adults. There is a need for strong and transparent education
British athletes following changes to the 2004 WADA prohib-
to ensure that the correct messages about caffeine in sport
ited list. Int. J. Sports Med. 29: 524–528. doi:10.1055/s-2007-
are provided to all athletes.
989231. PMID:18027309.
Cohen, B.S., Nelson, A.G., Prevost, M.C., Thompson, G.D., Marx,
References B.D., and Morris, G.S. 1996. Effects of caffeine ingestion on en-
Anderson, M.E., Bruce, C.R., Fraser, S.F., Stepto, N.K., Klein, R., durance racing in heat and humidity. Eur. J. Appl. Physiol. Occup.
Hopkins, W.G., and Hawley, J.A. 2000. Improved 2000-meter Physiol. 73: 358–363. doi:10.1007/BF02425499. PMID:8781869.
rowing performance in competitive oarswomen after caffeine in- Collomp, K., Ahmaidi, S., Chatard, J.C., Audran, M., and Prefaut,
# 2008 NRC Canada
Burke 1333

C. 1992. Benefits of caffeine ingestion on sprint performance in prohibited methods 2003. International Olympic Committee,
trained and untrained swimmers. Eur. J. Appl. Physiol. Occup. Lausanne, Switzerland.
Physiol. 64: 377–380. doi:10.1007/BF00636227. PMID:1592065. Ivy, J.L., Costill, D.L., Fink, W.J., and Lower, R.W. 1979. Influ-
Conway, K.J., Orr, R., and Stannard, S.R. 2003. Effect of a divided ence of caffeine and carbohydrate feedings on endurance perfor-
dose of endurance cycling performance, postexercise urinary mance. Med. Sci. Sports, 11: 6–11. PMID:481158.
caffeine concentration and plasma paraxanthine. J. Appl. Phy- Jacobson, T.L., Febbraio, M.A., Arkinstall, M.J., and Hawley, J.A.
siol. 94: 1557–1562. doi:10.1063/1.1589600. PMID:12482764. 2001. Effect of caffeine co-ingested with carbohydrate or fat on
Cox, G.R., Desbrow, B., Montgomery, P.G., Anderson, M.E., metabolism and performance in endurance-trained men. Exp.
Bruce, C.R., Macrides, T.A., et al. 2002. Effect of different pro- Physiol. 86: 137–144. doi:10.1113/eph8602072. PMID:11429627.
tocols of caffeine intake on metabolism and endurance perfor- Jenkins, N.T., Trilk, J.L., Singhal, A., O’Connor, P.J., and Cureton,
mance. J. Appl. Physiol. 93: 990–999. PMID:12183495. K.J. 2008. Ergogenic effects of low doses of caffeine on cycling
Cureton, K.J., Warren, G.L., Millard-Stafford, M.L., Wingo, J.E., performance. Int. J. Sport Nutr. Exerc. Metab. 18: 328–342.
Trilk, J., and Buyckx, M. 2007. Caffeinated sports drink: ergo- PMID:18562777.
genic effects and possible mechanisms. Int. J. Sport Nutr. Exerc. Jones, G. 2008. Caffeine and other sympathomimetic stimulants:
Metab. 17: 35–55. PMID:17460332. modes of action and effects on sports performance. Essays Bio-
Delbeke, F.T., and Debackere, M. 1984. Caffeine: use and abuse in chem. 44: 109–123. doi:10.1042/BSE0440109. PMID:18384286.
sports. Int. J. Sports Med. 5: 179–182. PMID:6480201. Keisler, B.D., and Armsey, T.D. 2006. Caffeine as an ergogenic
Desbrow, B., and Leveritt, M. 2006. Awareness and use of caffeine aid. Curr. Sports Med. Rep. 5: 215–219. PMID:16822345.
by athletes competing at the 2005 Ironman Triathlon World Kovacs, E.M.R., Stegen, J.H.C.H., and Brouns, F. 1998. Effect of
Championships. Int. J. Sport Nutr. Exerc. Metab. 16: 545–558. caffeinated drinks on substrate metabolism, caffeine excretion,
PMID:17240785. and performance. J. Appl. Physiol. 85: 709–715. PMID:9688750.
Desbrow, B., Hughes, R., Leveritt, M., and Scheelings, P. 2007. An MacIntosh, B.R., and Wright, B.M. 1995. Caffeine ingestion and
examination of consumer exposure to caffeine from retail coffee performance of a 1,500-metre swim. Can. J. Appl. Physiol. 20:
outlets. Food Chem. Toxicol. 45: 1588–1592. doi:10.1016/j.fct. 168–177. PMID:7640644.
2007.02.020. PMID:17412475. McCusker, R.R., Goldberger, B.A., and Cone, E.J. 2003. Caffeine
Doherty, M., and Smith, P.M. 2004. Effects of caffeine ingestion on content of specialty coffees. J. Anal. Toxicol. 27: 520–522.
exercise testing. Int. J. Sport Nutr. Exerc. Metab. 14: 626–646. PMID:14607010.
PMID:15657469. Millard-Stafford, M.L., Cureton, K.J., Wingo, J.E., Trilk, J.,
Ferrauti, A., Weber, K., and Struder, H.K. 1997. Metabolic and er- Warren, G.L., and Buyckx, M. 2007. Hydration during exercise
gogenic effects of carbohydrate and caffeine beverages in tennis. in warm, humid conditions: effect of caffeinated sports drink.
J. Sports Med. Phys. Fitness, 37: 258–266. PMID:9509824. Int. J. Sport Nutr. Exerc. Metab. 17: 163–177. PMID:17507741.
Fiala, K.A., Casa, D.J., and Roti, M.W. 2004. Rehydration with a Nawrot, P., Jordan, S., Eastwood, J., Rotstein, J., Hugenholtz, A.,
caffeinated beverage during the nonexercise periods of 3 conse- and Feeley, M. 2003. Effects of caffeine on human health. Food
cutive days of 2-a-day practices. Int. J. Sport Nutr. Exerc. Me- Addit. Contam. 20: 1–30. doi:10.1080/0265203021000007840.
tab. 14: 419–429. PMID:15467100. PMID:12519715.
Gonzalez-Alonso, J., Heaps, C.L., and Coyle, E.F. 1992. Rehydra- Pasman, W.J., van Baak, M.A., Jeukendrup, A.E., and de Haan, A.
tion after exercise with common beverages and water. Int. J. 1995. The effect of different dosages of caffeine on endurance
Sports Med. 13: 399–406. PMID:1521958. performance time. Int. J. Sports Med. 16: 225–230. doi:10.
Graham, T.E. 2001a. Caffeine, coffee and ephedrine: impact on ex- 1055/s-2007-972996. PMID:7657415.
ercise performance and metabolism. Can. J. Appl. Physiol. 26: Paton, C.D., Hopkins, W.G., and Vollebregt, L. 2001. Little effect
S103–S109. PMID:11897887. of caffeine ingestion on repeated sprints in team-sport athletes.
Graham, T.E. 2001b. Caffeine and exercise: metabolism, endurance Med. Sci. Sports Exerc. 33: 822–825. doi:10.1097/00005768-
and performance. Sports Med. 31: 785–807. doi:10.2165/ 200105001-00964. PMID:11323555.
00007256-200131110-00002. PMID:11583104. Pedersen, D.J., Lessard, S.J., Coffey, V., Churchley, E.G., Wootton,
Graham, T.E., and Spriet, L.L. 1995. Metabolic, catecholamine, A.M., Ng, T., et al. 2008. High rates of muscle glycogen re-
and exercise performance responses to various doses of caffeine. synthesis after exhaustive exercise when carbohydrate is co-
J. Appl. Physiol. 78: 867–874. PMID:7775331. ingested with caffeine. J. Appl. Physiol. 105: 7–13. doi:10.
Graham, T.E., Hibbert, E., and Sathasivam, P. 1998. Metabolic and 1152/japplphysiol.01121.2007. PMID:18467543.
exercise endurance effects of coffee and caffeine ingestion. J. Pruscino, C.L., Ross, M.L., Gregory, J.R., Savage, B., and
Appl. Physiol. 85: 883–889. PMID:9729561. Flanagan, T.R. 2008. The effects of sodium bicarbonate, caf-
Graham, T.E., Battram, D.S., Dela, F., El-Sohemy, A., and Thong, feine and their combination on repeated 200-m freestyle perfor-
F.S.L. 2008. Does caffeine alter muscle carbohydrate and fat mance. Int. J. Sport Nutr. Exerc. Metab. 18: 116–130.
metabolism during exercise? Appl. Physiol. Nutr. Metab. 33: PMID:18458356.
this issue. Schneiker, K.T., Bishop, D., Dawson, B., and Hackett, L.P. 2006.
Hopkins, W.G., Hawley, J.A., and Burke, L.M. 1999. Design and Effects of caffeine on prolonged intermittent-sprint ability in
analysis of research on sport performance enhancement. Med. team-sport athletes. Med. Sci. Sports Exerc. 38: 578–585.
Sci. Sports Exerc. 31: 472–485. doi:10.1097/00005768-1999 doi:10.1249/01.mss.0000188449.18968.62. PMID:16540848.
03000-00018. PMID:10188754. Strecker, E., Foster, E.B., Taylor, K., Bell, L., and Pascoe, D.D.
Hunter, A.M., St Clair Gibson, A., Collins, M., Lambert, M., and 2007. The effect of caffeine and ingestion on tennis skill perfor-
Noakes, T.D. 2002. Caffeine ingestion does not alter perfor- mance and hydration status. Med. Sci. Sports Exerc. 39: S43.
mance during a 100-km cycling time-trial performance. Int. J. Stuart, G.R., Hopkins, W.G., Cook, C., and Cairns, S.P. 2005. Mul-
Sport Nutr. Exerc. Metab. 12: 438–452. PMID:12500987. tiple effects of caffeine on simulated high-intensity team sport
International Olympic Committee. 2003. Olympic Movement Anti- performance. Med. Sci. Sports Exerc. 37: 1998–2005. doi:10.
doping Code. Appendix A. Prohibited classes of substances and 1249/01.mss.0000177216.21847.8a. PMID:16286872.
# 2008 NRC Canada
1334 Appl. Physiol. Nutr. Metab. Vol. 33, 2008

Tarnopolsky, M.A. 2008. Effect of caffeine on the neuromuscular Vergauwen, L., Brouns, F., and Hespel, P. 1998. Carbohydrate sup-
system — potential as an ergogenic aid. Appl. Physiol. Nutr. plementation improves stroke performance in tennis. Med. Sci.
Metab. 33: this issue. Sports Exerc. 30: 1289–1295. doi:10.1097/00005768-199808000-
van Dam, R.M., and Hu, F.B. 2005. Coffee consumption and risk 00017. PMID:9710871.
of type 2 diabetes: a systematic review. JAMA, 294: 97–104. Wemple, R.D., Lamb, D.R., and McKeever, K.H. 1997. Caffeine
doi:10.1001/jama.294.1.97. PMID:15998896. vs. caffeine-free sports drinks: effects on urine production at rest
Van Nieuwenhoven, M.A., Brouns, F., and Kovacs, E.M.R. 2005. The and during prolonged exercise. Int. J. Sports Med. 18: 40–46.
effect of two sports drinks and water on GI complaints and perfor- doi:10.1055/s-2007-972593. PMID:9059904.
mance during an 18-km run. Int. J. Sports Med. 26: 281–285. Wiles, J.D., Bird, S.R., Hopkins, J., and Riley, M. 1992. Effect of
doi:10.1055/s-2004-820931. PMID:15795812. caffeinated coffee on running speed, respiratory factors, blood
Van Thuyne, W., and Delbeke, F.T. 2006. Distribution of caffeine lactate and perceived exertion during 1500-m treadmill running.
levels in sport. Int. J. Sports Med. 27: 745–750. doi:10.1055/s- Br. J. Sports Med. 26: 116–120. PMID:1623356.
2005-872921. PMID:16586337. Wiles, J.D., Coleman, D., Tegerdine, M., and Swaine, I.L. 2006. The
Van Thuyne, W., Roels, K., and Delbeke, F.T. 2005. Distribution effects of caffeine ingestion on performance time, speed and
of caffeine levels in urine in different sports in relation to dop- power during a laboratory-based 1 km cycling time-trial. J. Sports
ing control. Int. J. Sports Med. 26: 714–718. doi:10.1055/s- Sci. 24: 1165–1171. doi:10.1080/02640410500457687. PMID:
2005-837437. PMID:16237615. 17035165.

# 2008 NRC Canada

View publication stats