Creativedining.com

Gatorade
Sports
Science
60
Institute ®
CAFFEINE AND EXERCISE PERFORMANCE
Terry E. Graham, Ph.D.
Lawrence L. Spriet, Ph.D.
Department of Human Biology & Nutritional Sciences Member, GSSI Sports Medicine Review Board Department of Human Biology & Nutritional Sciences KEY POINTS
1. Recent, well-controlled studies have established that moderate doses of caffeine ingested 1 h prior to exercise enhance the performance of certain types of endurance exercise in the laboratory. Moderate caffeine doses produce urinary caffeine levels well below the allowable limit,as determined by the International Olympic Committee. The results are specific to well-trained elite or recreational athletes. It is not knownif these findings will improve performance in competitions because controlled field studies of the effects of caffeine are lacking.
2. The mechanisms responsible for improved exercise endurance in prolonged exercise remain elusive. A metabolic mechanism appears to con- tribute early in exercise, when caffeine ingestion increases plasma free-fatty acid concentrations and muscle triglyceride use and sparesmuscle glycogen. However, it is not clear if increased fat oxidation causes the glycogen sparing in muscle. Increases in plasma epinephrineconcentrations usually occur following caffeine ingestion but are not essential for the accompanying metabolic changes. When studying caf-feine effects in the human it is difficult to identify the primary source of the "stimulus" because caffeine usually increases epinephrine secre-tion and is also rapidly metabolized in the liver to three dimethylxanthines (paraxanthine, theophylline and theobromine). Thedimethylxanthines can remain in the circulation longer than caffeine and may be metabolic signals in their own right.
3. Caffeine appears to enhance performance during short-term, intense cycling lasting ~5 min in the laboratory and in simulated 1500 m race time. However, positive ergogenic effects of caffeine are much less frequent during sprint exercise lasting less than 90 s and in incrementalexercise tests lasting 8-20 min.
4. Potential mechanisms for improving performance during intense exercise lasting 5-20 min include direct effects of caffeine on the central nervous system and/or excitation-contraction coupling and increased anaerobic energy provision in skeletal muscle.
INTRODUCTION
Caffeine is a "controlled or restricted drug" in the athletic world, because urinary levels of greater than 12 µg/mL following competitions are considered illegal by the International Olympic Committee (IOC). However, most athletes that consume caffeine beverages prior to exercisewould never approach the illegal limit following a competition. Therefore, caffeine occupies a unique position in the sports world. It is aninherent part of the diet of many athletes although it has no nutritional value and also has the potential to be a "legal" ergogenic aid in manyexercise situations. While it is common to equate caffeine with coffee, it should be noted that rarely is coffee the vehicle of administration inresearch studies. Therefore, it may be misleading to equate the two because coffee contains hundreds of additional chemicals.
In a 1990 Sport Science Exchange article, Wilcox concluded that few well-controlled studies had examined the effects of caffeine on endurance performance and that the results were inconsistent. Since 1990, the research examining caffeine and exercise performance increasedand demonstrated the ergogenic effect of caffeine during prolonged endurance exercise (Graham & Spriet, 1991, 1995; Pasman et al., 1995). Inaddition, investigations examining the effects of caffeine on exercise performance during sprinting (<90 s), intense exercise of short (~5 min)and long duration (~20 min)(Collomp et al., 1990, 1991; Jackman et al., 1996; MacIntosh & Wright, 1995) have appeared.
There has been general improvement in the quality of the investigations because researchers have attempted to control several factors that may confound the caffeine results. Conlee summarized these factors in a 1991 review article. Three factors relate to the nature of the experi-mental design, i.e., the exercise modality, the power output, and the caffeine dose, whereas four others relate to the status of the subjects prior tothe experiment, i.e., nutritional status, training status, previous caffeine use, and individual variability. An additional factor is the ability to reli-ably measure exercise performance. This reliability is greater in highly trained subjects than it is in the less well-trained.
Caffeine appears to be taken up by of all tissues of the body, making it difficult to independently study the effects of caffeine on the central nervous system, the muscles, and fat tissue in the exercising human. It is also apparent that different mechanisms are probably responsible forperformance enhancement in different types of exercise. However, it is important to note that the mechanism(s) may not be entirely due to caf-feine. For example, caffeine ingestion usually increases the plasma concentration of epinephrine, a hormone with widespread effects, and the liver rapidly metabolizes the caffeine, a trimethylxanthine, into three dimethylxan- (VO2max) (Costill et al., 1978). The trained thines, i.e., paraxanthine, theophylline, and cyclists improved performance from 75 min were rare with doses at or below 6 mg/kg, but prevalent at higher doses (9-13 mg/kg).
metabolites increase in the plasma as the lowing caffeine ingestion. A second study caffeine concentration declines, and parax- decreased performance in some athletes at potential metabolic stimuli. Thus, it is diffi- amount of work performed in 2 h (Ivy et al., cult to resolve which tissues are directly or 1979). These studies suggested that utiliza- increase in venous plasma EPI at rest and indirectly affected by which compound.
tion of fat for energy increased by ~30% in the caffeine trials. A third study examined at rest. The elevated FFA with caffeine was "caffeine" is used in this report, the reader that ingestion of 5 mg of caffeine/kg body mg/kg), performance was increased without significant increases in plasma venous EPI al., 1980). In the 1980's, few investigations THEORIES OF ERGOGENICITY
actually tested the ergogenic effects of caf- reduced following caffeine ingestion, but the "sparing" was limited to the initial 15 ergogenic effect of caffeine during exercise.
The first theory suggests a direct effect on affected by caffeine. Furthermore, conclu- some portion of the central nervous system sions regarding the metabolic effects of caf- There is little information on the perfor- that affects the perception of effort and/or feine were often based on indirect indicators mance and metabolic effects of caffeine in the neural activation of muscle contraction.
of fat metabolism, i.e., increases in plasma recreationally active or untrained subjects The second theory proposes a direct effect free-fatty acids (FFA) and/or decreases in because performance in these groups is dif- of caffeine on skeletal muscle performance.
the ratio of carbon dioxide production to ficult to measure accurately. Chesley et al This may involve ion transport (including (1994) reported a variable glycogen sparing Ca2+ transport) and direct effects on key reg- response to a high caffeine dose (9 mg/kg) ulatory enzymes, including those control- these suggestions is largely derived from in 1995; Tarnopolsky, 1994, Wilcox, 1990).
vitro investigations in which high pharma- results suggest that the metabolic responses to caffeine ingestion in untrained subjects used to demonstrate effects. If these "test- RECENT STUDIES OF CAFFEINE
are more variable than in aerobically trained tube" results have any relevance during EFFECTS ON ENDURANCE
exercise, the most likely candidates for con- PERFORMANCE AND METABOLISM
tributing to an ergogenic effect of caffeine are changes in calcium activity and in the effects of caffeine (4.5 mg/kg) in "pure" effects of caffeine in well-trained athletes tablet form to the same amount of caffeine from the extracellular fluid to the interior of who are accustomed to exhaustive exercise in a coffee beverage (~two mugs of strong the muscle fibers; caffeine levels during and race conditions. Most confounding fac- coffee ingested in 10 min). Caffeine as a exercise are similar to the lowest concentra- tors were well controlled, and performance tablet resulted in the usual metabolic and tions of caffeine used in vitro that can affect performance effects, but when ingested as petitive conditions. The studies examined a beverage there was less of a response in The third theory is the classic or "meta- the effects of a caffeine dose of 9 mg/kg plasma epinephrine and little or no effect bolic" explanation that involves an increase in fat oxidation and a reduction in carbohy- exhaustion at 80-85% VO2max (Graham & caffeine concentrations were identical.
drate oxidation. In this scheme, caffeine Spriet, 1991: Spriet et al., 1992), the effects directly enhances the activity of enzymes of varying doses (3-13 mg/kg) of caffeine that break down fat into fatty acids or caf- on cycling performance (Graham & Spriet, feine increases circulating levels of epi- 1995; Pasman et al., 1995) and the effects MECHANISMS FOR
IMPROVED ENDURANCE
stores in fat or muscle tissue. The increased cycling (5 min rest between bouts) at 85- fatty acid availability increases muscle fat 90% VO2max (Trice & Haymes, 1995).
oxidation and reduces carbohydrate oxida- tion, thereby improving the performance of important findings. Endurance performance exercise that becomes exhausting when car- except at the low caffeine doses for which placebo trial ~20-50% following ingestion this hypothesis has not been fully examined.
The increased FFA at the onset of exercise, of varying doses of caffeine (3-13 mg/kg) in the glycogen sparing in the initial 15 min, elite and recreationally trained athletes who and the report of increased intramuscular varying types of exercise that are catego- TG use during the first 30 min of exercise rized according to power output and time to exception, the 3, 5, and 6 mg/kg doses pro- suggest a greater role for fat metabolism exhaustion or to completion of a race.
duced an ergogenic effect with urinary caf-feine levels that were below the IOC early in exercise following caffeine doses of acceptable limit. Three of four experiments CAFFEINE AND ENDURANCE
using a 9 mg/kg dose reported performance bolic findings do not preclude other factors PERFORMANCE
increases, while 6/22 athletes tested in these contributing to enhanced endurance perfor-mance. For example, caffeine appears to studies had urinary caffeine levels at or stimulate transport of potassium into inac- ergogenic aid was initially stimulated by tive tissues, leading to an attenuation of the work from Costill's laboratory. They exam- ined the effect of ingesting 330 mg of caf- athletes had urinary caffeine levels well during exercise. It has been postulated that feine 1 h prior to cycling to exhaustion at above 12 µg/mL. The side effects of caf- the lower plasma potassium helps maintain CAFFEINE AND PERFORMANCE OF
the excitability of the cell membranes in SHORT-TERM INTENSE EXERCISE
events lasting less than 90 s. The amount contracting muscles and contributes to caf- feine's ergogenic effect during endurance effects of caffeine on performance during processes would be ~75-80% of the total in the first 30 s, ~65-70% over 60 s, and ~55- lasting ~5 min; near-maximal provision of Williams et al. (1988) reported that caf- changes occurring with caffeine ingestion.
sources is required for such activities.
power output or muscular endurance during short, maximal bouts of cycling. Collomp et al. (1992) found that ingestion of caffeine at exercise. In addition, an infusion of EPI with placebo to 5:49, although the increase that was designed to produce exercise EPI was not significant. A third trial, in which power or total work completed in six sub- concentrations similar to those induced by subjects received 250 mg caffeine daily for jects performing a 30-s Wingate test.
5 d also increased exhaustion time non-sig- However, the same group later reported that 250 mg of caffeine produced a significant (Chesley et al., 1995). Also, Van Soeren et time on a treadmill in well-trained runners by 4.2 s compared to placebo (4:46.0 vs.
Therefore, the known alterations in muscle 4:50.2). In a second protocol subjects drank 1992)). The same authors also examined the metabolism alone cannot presently explain either coffee or a placebo, ran for 1100 m at effects of 250 mg of caffeine on two 100-m freestyle swims that were separated by 20 as fast as possible. The average speed of the min (Collomp et al., 1990). In well-trained swimmers, the velocity during the first and CAFFEINE AND PERFORMANCE
OF GRADED EXERCISE TESTS
Therefore, given the present information, examined the effects of caffeine ingestion it is not possible to conclude whether or not caffeine has an ergogenic effect on sprint graded exercise protocols lasting 8-20 min bolic responses to repeated bouts of cycling performance. The brief and intense nature of sprint exercise makes it very difficult to studies from the same laboratory reported study and to demonstrate significant effects intervening rest periods of 6 min each. The of caffeine were given (Flinn et al., 1990; first two cycling bouts at a controlled power McNaughton, 1987). The first study used 10 output lasted 2 min, and the third continued FIELD STUDIES
and 15 mg/kg caffeine doses and reported a to exhaustion. Cycle time to exhaustion was small, significant increase in performance.
improved with caffeine (4.93 + 0.60 min vs.
However, the control trial always preceded the caffeine trials, leading to the possibility as the time taken to reach exhaustion at a blood lactate measurements throughout the of an order effect. The second study used a given power output. However, in the field, protocol suggested a higher production of 10 mg/kg caffeine dose 3 h prior to cycling performance is measured as the time taken lactate in the caffeine trial, even in the ini- exercise and reported an increased time to tial two bouts when power output was con- exhaustion. The subjects completed control, Consequently, extrapolations from the labo- trolled. The net rate of glycogen breakdown placebo and caffeine trials with the control ratory to field settings may not be valid.
trial always first, and the remaining two Occasionally, laboratory studies simulate trials randomized. It appears that the high race conditions by allowing the subject to caffeine dose is the most likely factor that control speed on a treadmill or cadence and during the protocol. The authors concluded separates these positive findings from the resistance on a cycle ergometer in order to that the ergogenic effect of caffeine during studies reporting no effect. Unfortunately, short-term intense exercise was not associ- no mechanistic information presently exists work in the shortest possible time. Other muscle or by altered function of the central actual race conditions but in time trials.
CAFFEINE AND PERFORMANCE OF
However, these studies still do not entirely INTENSE AEROBIC EXERCISE
require athletes to exercise at power outputs conditions, it is often impossible to employ the controls required to generate conclusive energy provision, direct effects of caffeine on the transport of ions in muscles, and cen- effect of 6 mg caffeine/kg on performance tral nervous system effects on the sensation in 1500 m swim trials in trained distance of effort and/or activation of muscle con- ingestion on performance during endurance exercise. Cross-country ski performance in (min:sec). The authors reported lower pre- a race lasting 1-1.5 h was improved by 1- exercise plasma potassium levels and higher CAFFEINE AND SPRINT
2.5 min compared to a control condition.
post-exercise blood glucose concentrations PERFORMANCE
Oddly, this improvement occurred during a with caffeine and suggested that electrolyte race at high altitude but not at sea level.
fatiguing exercise at power outputs 1.5- to Unfortunately, the weather and snow condi- related to the ergogenic effects of caffeine. 3-fold greater than that required to elicit requiring mathematical "normalization" of ingest small and large amounts of caffeine, al., 1990; Gordon et al., 1982). A recent tions about the validity of the results and glycogen sparing following caffeine inges- indicate how difficult it is to perform well- tion is greater in samples of untrained males controlled and meaningful field trials.
than in trained males (Chesley et al., 1994; There is a tremendous need for more field Spriet et al., 1992). Few females have been Ethical Considerations
studied to determine if the variability in response to caffeine ingestion is similar to that in males. Menstrual status needs to be been reported with doses of 3-6 mg/kg, it is easy for endurance athletes to enhance per- PRACTICAL CONSIDERATIONS OF
estrogen may affect the half-life of caffeine.
formance "legally" with caffeine. We sug- INGESTING CAFFEINE
Therefore, although mean results in a group gested on the basis of our work that caffeine Caffeine Dose
of athletes may predict an improved athletic should be banned prior to competitions inendurance athletes. This would ensure that Caffeine is a "controlled or restricted performance, it is impossible to reliably no athlete had an unfair advantage on race substance" with respect to the IOC. Athletes predict that the performance of a given indi- day but would not prevent caffeine use in are allowed up to 12 µg caffeine/mL urine before it is considered illegal. This permits Habitual Caffeine Consumption
from caffeine ~48-72 h prior to competition athletes who normally consume caffeine in to achieve this goal. However, in the pre- their diets to continue this practice prior to several recent studies suggest that chronic competition. An athlete can consume a very large amount of caffeine before reaching the exercise and to caffeine but does not affect amounts to make sure they are not missing "illegal limit". A 70 kg person could drink indirect markers of fat metabolism during about three or four mugs or six regular size exercise (Bangsbo et al., 1992; Van Soeren cups of drip-percolated coffee ~1 h before et al., 1993). However, these changes do not exercise, exercise for 1-1.5 h and produce a appear to dampen the ergogenic effect of 9 point of view may be popular because caf- feine use is prevalent in society, and ath- approach the urinary caffeine limit. It is not increased in all subjects in two studies in letes will not have "illegal" amounts in their easy to reach the limit by ingesting coffee.
which both users and non-users of caffeine A caffeine level above 12 µg/mL suggests moderation is a trivial issue; other drugs that an individual has deliberately taken caf- feine in the form of tablets or suppositories (Graham & Spriet, 1991; Spriet et al., greater attention. Nevertheless, the potential in an attempt to improve performance. Not ergogenic effect of caffeine is impressive.
surprisingly, only a few athletes have been caught with illegal caffeine levels during with more non-users (Graham & Spriet, use counteracts the "win-at-all-costs" men- competitions, although formal reports of the 1995). In addition, Van Soeren et al. (1993) tality and sets the proper example for youth.
frequency of caffeine abuse are rare. One recently reported that prior caffeine with- older study reported that 26/775 cyclists had drawal for up to 4 d did not affect exercise- illegal urinary caffeine levels when tested aged 11-18 reported using caffeine in the prior year to help them do better in sports.
ingested caffeine doses of 6 or 9 mg/kg.
Urinary Caffeine and Doping
determine caffeine abuse in sport has been criticized. Only 0.5-3% of orally ingested caffeine actually reaches the urine because most of the caffeine is metabolized in the Caffeine and High Carbohydrate Diets
weight) prior to exercise often increases excreted are not measured in doping tests.
cycling and running in a laboratory setting.
Other factors also affect the amount of caf- feine that reaches the urine, including body FFA following caffeine ingestion during 2 h weight, gender, and hydration status of the athlete. The time elapsed between caffeine increase short-term intense cycling (~5 min) important and will be affected by the exer- negate the ergogenic effects of caffeine, cise duration and environmental conditions.
generally reported in well-trained elite or diet and a pre-trial carbohydrate meal did recreational athletes, but field studies are people caught with illegal levels of caffeine not prevent caffeine-induced increases in lacking to confirm the ergogenic effects of will have used caffeine in a doping manner.
performance in a number of recent studies caffeine in the athletic world. The mecha- using well-trained/recreational runners and nisms for the improved endurance have not metabolizes caffeine slowly or who excretes could produce IOC-illegal amounts of uri- Diuretic Effect of Caffeine
effects of caffeine on muscles and/or on the nary caffeine following ingestion of a mod- been suggested that caffeine ingestion may Variability of Caffeine Responses
lead to poor hydration status prior to andduring exercise. However, two studies metabolic responses to caffeine is large.
sweat loss, or plasma volume during exer- cise following caffeine ingestion (Falk et References
Anselme, F., K. Collomp, B. Mercier, S. Ahmaidi, and C. Prefaut. (1992). Caffeine increases maximal anaerobic power and blood lactate concentration. Eur. J. Appl.
Physiol. 65:188-191.
Bangsbo, J., K. Jacobsen, N. Nordberg, N.J. Christensen, and T. Graham. (1992). Acute and habitual caffeine ingestion and metabolic responses to steady-state exer-cise. J. Appl. Physiol. 72:1297-1303.
Berglund, B., and P. Hemingsson. (1982). Effects of caffeine ingestion on exercise performance at low and high altitudes in cross country skiers. Int. J. Sports Med.
3:234-236.
Chesley, A., E. Hultman, and L.L. Spriet. (1995). Effects of epinephrine infusion on muscle glycogenolysis during intense aerobic exercise. Am. J. Physiol. 268(Endocrinol. Met.):E127-E134.
Chesley, A., E. Hultman, and L.L. Spriet. (1994). Variable effects of caffeine on muscle glycogenolysis in recreationally active subjects during intense aerobic exer-cise. Can. J. Appl. Physiol. 19:10P, 1994. (Abstract).
Collomp, K., C. Caillaud, M. Audran, J.-L. Chanal, and C. Prefaut. (1990). Influence of acute and chronic bouts of caffeine on performance and catecholamines in thecourse of maximal exercise. C.R. Soc. Biol. 184:87-92.
Collomp. K., S. Ahmaidi, M. Audran, J.-L. Chanal, and C. Prefaut. (1991). Effects of caffeine ingestion on performance and anaerobic metabolism during theWingate test. Int. J. Sports Med. 12:439-443.
Collomp, K., S. Ahmaidi, J.C. Chatard, M. Audran, and C. Prefaut. (1992). Benefits of caffeine ingestion on sprint performance in trained and untrained swimmers.
Eur. J. Appl. Physiol. 64:377-380.
Conlee, R.K. (1991). Amphetamine, caffeine and cocaine. In: D.R. Lamb and M.H. Williams (Eds.) Ergogenics: Enhancement of Performance in Exercise and Sport.
Indianapolis: Brown and Benchmark, pp. 285-330.
Costill, D.L., G. Dalsky, and W. Fink. (1978). Effects of caffeine ingestion on metabolism and exercise performance. Med. Sci. Sports 10: 155-158.
Delbecke, F.T., and M. Debachere. (1984). Caffeine: use and abuse in sports. Int. J. Sports Med. 5:179-182.
Dodd, S.L., R.A. Herb, and S.K. Powers. (1993). Caffeine and endurance performance: An update. Sports Med. 15:14-23.
Essig, D., D.L. Costill, and P.J. VanHandel. (1980). Effects of caffeine ingestion on utilization of muscle glycogen and lipid during leg ergometer cycling. Int. J.
Sports Med. 1:86-90.
Falk, B., R. Burstein, J. Rosenblum, Y. Shapiro, E. Zylber-Katz, and N. Bashan. (1990). Effects of caffeine ingestion on body fluid balance and thermoregulationduring exercise. Can. J. Physiol. Pharmacol. 68:889-892.
Flinn, S., J. Gregory, L.R. Mcnaughton, S. Tristram, and P. Davies. (1990). Caffeine ingestion prior to incremental cycling to exhaustion in recreational cyclists. Int. J.
Sports Med. 11:188-193.
Gordon, N.F., J.L. Myburgh, P.E. Kruger, P.G. Kempff, J.F. Cilliers, J. Moolman, and H.C. Grobler. (1982). Effects of caffeine on thermoregulatory and myocardialfunction during endurance performance. S. Afr. Med. J. 62:644-647.
Graham, T.E., and L.L. Spriet. (1991). Performance and metabolic responses to a high caffeine dose during prolonged exercise. J. Appl. Physiol. 71:2292-2298.
Graham, T.E., and L.L. Spriet. (1995). Metabolic, catecholamine and exercise performance responses to varying doses of caffeine. J. Appl. Physiol. 78:867-874.
Graham, T.E., J.W.E. Rush, and M.H. VanSoeren. (1994). Caffeine and exercise: metabolism and performance. Can. J. Appl. Physiol. 2:111-138.
Graham, T.E., E. Hibbert, and P. Sathasivam. (1995). Caffeine Vs. coffee: coffee isn't an effective ergogenic aid. Med. Sci. Sports Exerc. 27:S224. (Abstract).
Ivy, J.L., D.L. Costill, W.J. Fink, and R.W. Lower. (1979). Influence of caffeine and carbohydrate feedings on endurance performance. Med. Sci. Sports 11:6-11.
Jackman, M., P. Wendling, D. Friars, and T.E. Graham. (1996). Metabolic, catecholamine and endurance responses to caffeine during intense exercise. J. Appl.
Physiol. 80: In press.
Lindinger, M.I., T.E. Graham, and L.L. Spriet. (1993). Caffeine attenuates the exercise-induced increase in plasma [K+] in humans. J. Appl. Physiol. 74:1149-1155.
MacIntosh, B.R., and B.M. Wright. (1995). Caffeine ingestion and performance of a 1500-metre swim. Can. J. Appl. Physiol. 20:168-177.
McNaughton, L. (1987). Two levels of caffeine ingestion on blood lactate and free fatty acid responses during incremental exercise. Res. Q. Exerc. Sport 58:255-259.
Pasman, W.J., M.A. VanBaak, A.E. Jeukendrup, and A. DeHaan. (1995). The effect of different dosages of caffeine on endurance performance time. Int. J. SportsMed. 16:225-230.
Spriet, L.L., D.A. MacLean, D.J. Dyck, E. Hultman, G. Cederblad, and T.E. Graham. (1992). Caffeine ingestion and muscle metabolism during prolonged exercise inhumans. Am. J. Physiol. 262 (Endocrinol. Metab.):E891-E898.
Spriet, L.L. (1995). Caffeine and performance. Int. J. Sports Nutr. 5:S84-S99.
Tarnopolsky, M.A. (1994). Caffeine and endurance performance. Sports Med. 18:109-125.
Trice, I., and E.M. Haymes. (1995). Effects of caffeine ingestion on exercise-induced changes during high-intensity, intermittent exercise. Int. J. Sports. Nutr. 5:37-44.
VanSoeren, M.H., P. Sathasivam, L.L. Spriet, and T.E. Graham. (1993). Caffeine metabolism and epinephrine responses during exercise in users and non-users.
J. Appl. Physiol. 75:805-812.
VanSoeren, M.H., P. Sathasivam, L.L. Spriet, and T.E. Graham. (1993). Short term withdrawal does not alter caffeine-induced metabolic changes during intensiveexercise. FASEB J. 7:A518. (Abstract). VanSoeren, M.H., T. Mohr, M. Kjaer, and T.E. Graham. (1996). Acute effects of caffeine ingestion at rest in humans with impaired epimephrine responses. J. Appl.
Physiol. 80: 999-1005, 1996.
Weir, J., T.D. Noakes, K. Myburgh, and B. Adams. (1987). A high carbohydrate diet negates the metabolic effect of caffeine during exercise. Med. Sci. Sports Exerc.
19:100-105.
Wemple, R.D., D.R. Lamb, and A.C. Blostein. (1994). Caffeine ingested in a fluid replacement beverage during prolonged exercise does not cause diuresis. Med. Sci.
Sports Exerc. 26:S204. (Abstract).
Wilcox, A.R. (1990). Caffeine and endurance performance. In: Sports Science Exchange. Barringtron, IL: Gatorade Sports Science Institute. 3:1-5.
Wiles, J.D., S.R. Bird, J. Hopkins, and M. Riley. (1992). Effect of caffeinated coffee on running speed, respiratory factors, blood lactate and perceived exertion during1500-m treadmill running. Br. J. Sports Med. 26:166-120.
Williams, J.H., J.F. Signoille, W.S. Barnes, and T.W. Henrich. (1988). Caffeine, maximal power output and fatigue. Br. J. Sports Med. 229:132-134.

Source: http://www.creativedining.com/pages/NI4U/caffeine_and_exercise.pdf