Effects on performance
Caffeine reduces fatigue and increases concentration and alertness, and athletes regularly use it as an ergogenic aid . Caffeine-induced increases in performance have been observed in aerobic as well as anaerobic sports (for reviews, see [26, 30, 31]). Trained athletes seem to benefit from a moderate dose of 5 mg/kg , however, even lower doses of caffeine (1.0–2.0 mg/kg) may improve performance . Some groups found significantly improved time trial performance  or maximal cycling power , most likely related to a greater reliance on fat metabolism and decreased neuromuscular fatigue, respectively. Theophylline, a metabolite of caffeine, seems to be even more effective in doing so . The effect of caffeine on fat oxidation, however, may only be significant during lower exercise intensities and may be blocked at higher intensities . Spriet et al.  found that ingestion of a high dose of caffeine before exercise reduced muscle glycogenolysis in the initial 15 min of exercise by increasing free fatty acid (FFA) levels which inhibits glycolysis and spares glycogen for later use. Caffeine’s effect of inhibition of glycogen phosphorylase has also been shown in vitro  as well as its effect on increasing HSL activity . The effect of caffeine on adipose triglyceride lipase  has not been studied and warrants investigation. Following caffeine administration prior to and after the onset of cycling, Ivy et al.  found that plasma free fatty acid levels were increased 30% compared to placebo. This action might be mediated by inhibition of the enzyme phosphodiesterase, thereby yielding higher levels of cAMP, which has been identified as important molecule for glycogen metabolism and lipolysis . Phosphodiesterase inhibition has been observed only at high concentrations . When direct Fick measurements were applied, Graham et al.  did not find altered CHO or fat metabolism, at least in the monitored leg. Further research is needed to evaluate the effect of caffeine on lipolysis, especially during higher exercise intensities.
Augmented post-exercise recovery by increased rates of muscle glycogen resynthesis has been observed [18, 20]. Pedersen et al.  found higher rates of muscle glycogen accumulation after the co-ingestion of caffeine with CHO during recovery in highly trained subjects. This might, at least in part, be mediated by the activation of AMP-activated protein kinase (AMPK)  as it is involved in the translocation of glucose transporter 4 (GLUT4) to the plasma membrane. This mechanism enables the cell to take up glucose from the plasma and store it as glycogen. Not only does caffeine impact endurance, it has also been reported to benefit cognitive function and fine motor skills . While the performance enhancing effects of caffeine in moderate-to-highly trained endurance athletes are quite clear and well documented, its effects on anaerobic, high-intensity tasks are less well investigated. Whereas caffeine supplementation did not yield significant performance increases in a Wingate test in untrained subjects [46, 47], Mora-Rodriguez et al.  report that caffeine ingestion of 3 mg/kg could counter reductions in maximum dynamic strength and muscle power output on the morning (2.5–7.0%) thereby increasing muscle performance to the levels found in the afternoon. Especially with regard to anaerobic performance caffeine’s adenosine receptor blocking effect in the CNS may be important . A possible explanation for the diverging effect of caffeine on anaerobic performance is that caffeine seems to benefit trained athletes who show specific physiological adaptations whereas performance gains in untrained subjects might be lost or masked by a high variability in performance.
It has been shown that coffee, by containing phenolic compounds such as chlorogenic acids, elicits metabolic effects independent of caffeine . These compounds may have the potential to antagonize the physiological responses of caffeine. The question therefore remains whether ingesting the same amount of caffeine via a food source (e.g. energy bar or coffee) is as effective as ingesting isolated caffeine in the form of a tablet. As mentioned above, the performance enhancing effect of caffeine is very clear. Only a few studies, however, have shown a positive effect of coffee on performance. Whereas some studies found enhanced performance after coffee consumption [50–53], others did not [49, 54, 55].
One of the earlier works by Costill et al.  reported increases in time trial performance of competitive cyclists only in the coffee trial group (containing 330 mg caffeine 1 h prior to exercise) but not in the decaffeinated coffee trial.
Graham et al.  studied exercise endurance in runners after ingestions of a caffeine (4.45 mg/kg BW) or placebo capsule with water or either decaffeinated coffee, decaffeinated coffee with added caffeine or regular coffee. The authors found that only caffeine significantly improved running time to exhaustion at 75% VO2max but neither did regular coffee or decaffeinated coffee plus caffeine. Based on these results, the authors speculated that some component(s) in coffee possibly interfere with the ergogenic response of caffeine alone.
This is in opposition to Hodgson et al.  who looked at time trial performance in trained subjects after administration of caffeine (5 mg caffeine/kg BW), coffee (5 mg caffeine/kg BW), decaffeinated coffee and placebo one hour prior to exercise. The authors report similar significant increases of ~5% in time trial performance in both the caffeine and the coffee supplemented group with no effects in the decaf or placebo group. The authors conclude that coffee consumed 1 h prior to exercise, at a high caffeine dose improved performance to the same extent as caffeine.
One reason for the disparity of the two studies mentioned above might be the different performance tests used. Whereas Graham et al. used a time to exhaustion test which reportedly can exhibit a coefficient of variation as high as ~27% , Hodgson et al. used a time trial which have been shown to be more reproducible. It has also been speculated by Hodgson et al.  that due to lower statistical power, Graham et al.  were not able to detect a difference between caffeine and coffee ingestion on performance. At this point, both coffee and caffeine exhibit a performance enhancing effect. Further research will hopefully extend our understanding on this issue.
Another reason for the widespread use of caffeine within the exercise community might be its small but significant analgesic effect , possibly mediated by augmenting plasma endorphin concentrations . It is also established that caffeine reduces the rate of perceived exertion during exercise , suggesting that athletes are able to sustain higher intensities but do not perceive this effort to be different from placebo conditions.
Some studies used caffeine-naïve whereas others used caffeine-habituated subjects. There seems to be a higher increase in plasma adrenalin in caffeine-naïves compared to caffeine habituated subjects after caffeine ingestion . However, no differences between habitual caffeine intake and 1500 m running performance  or force of contraction  could be observed. For both caffeine-naïve as well as caffeine-habituated subjects, moderate to high doses of caffeine are ergogenic during prolonged moderate intensity exercise . Although there is clearly the need to study caffeine habituation further, the differences between users and non-users do not seem to be major.