The current study shows that 400 mg ATP per day was effective in improving leg muscle low peak torque in set 2 (p < 0.01), and tended to decrease leg muscle fatigue in set 3 (p < 0.10) of three successive sets of knee extension exercises. However, the improvement in low peak torque and decreased fatigue were not sufficient to translate into improvements in leg muscle power or work performed. These observations lead us to speculate that supplemental ATP may provide cumulative benefits in strenuous, repetitive, and exhaustive exercise activities, which could lead to improved strength and lean body mass gains.
There is limited human data related to the potential for oral ATP to manifest physiologic modifications that would improve skeletal muscle efficiency or work performed . As muscle undergoes prolonged work, ATP synthesis increases in an attempt to keep up with energy demand . To accomplish this, the muscle needs substrates, such as oxygen and glucose, supplied from the peripheral circulation. Endogenous muscle stores of ATP are limited and support maximal work for only a fraction of, or at most 1–2 seconds and is replenished by the supply of intercellular phosphocreatine for only an additional 2–7 seconds . Muscle performing exhaustive exercise then relies primarily on anaerobic glycolysis for regeneration of ATP which results in the production of lactate and H+. The associated decrease in intracellular pH is a factor leading to muscle fatigue [23, 24]. Therefore, during maximal exertion blood flow is needed not only for oxygen supply to support continued oxidative phosphorylation, but also for H+ removal for muscle pH regulation. It would seem that exogenous ATP would likely have a greater impact on the muscles’ ability to perform fatiguing exercise by increasing substrate availability to the muscle and/or facilitating waste product removal through increased blood flow through the muscle tissues.
Both ATP and adenosine can act through purinergic receptors in endothelial smooth muscle of the vascular system resulting in vasodilation and increased blood flow [14, 15, 25]. A study by Gonzalez-Alonso showed that intra-arterial infusion of ATP was associated with vasodilation and increased blood flow with a significant reduction in venous ATP levels in the non-exercising limb suggesting utilization of ATP or metabolites . These observations were confirmed by Calbet et al. who hypothesized that increased delivery of ATP would affect non-exercising vasoconstrictive muscle tissue . These are most likely due to activation of purinergic receptors affecting blood flow . Furthermore, exogenous adenosine administration results in vasodilation  and increased glucose and O2 uptake by muscle which provide for an increased substrate pool . The ATP used in the present study was not enterically coated and was fed encapsulated as the disodium salt. The sodium salt would have provided buffering of the ATP through the stomach and the ATP itself should have been metabolically available as soon as it reached the proximal duodenum, which has been shown to be the most biologically active site for ATP metabolism and/or absorption . In France, this chemical form of ATP is approved as a drug for lower back pain [27, 28]. One proposed mechanism of action is through improved oxygenation of the muscle, which could be of similar benefit during exhaustive exercise.
Other effects of ATP or its metabolites could also indirectly impact work performance as ATP has immunomodulatory effects , and inotropic effects on cardiac muscle [30, 31]. Oral administration of ATP to rabbits for 14 days results in systemic effects through a reduction in peripheral vascular resistance, improvement of cardiac output, reduction of lung resistance, and increased arterial PaO2. A study in humans demonstrated that interstitial infusion of adenosine in muscle induced nitric oxide formation in skeletal muscle and nitric oxide and prostacyclin formation in microvascular endothelial cells . Alternatively, the effects of cbvexogenously administered ATP may also be due to the associated increase in plasma uric acid, which has been proposed to act as an anti-oxidant [33, 34]. Increased plasma uric acid has been demonstrated with ATP supplementation [17, 18]. These studies may indicate further metabolism of adenosine before becoming bioavailable and warrant further investigation. These effects of ATP and adenosine could account, at least in part, for the improvements in low peak torque and torque fatigue we observed. The current study tested the hypothesis that oral ATP could improve performance during high intensity exercise. While we have shown this may be possible, the current study did not utilize methodologies to investigate the potential mechanism for the effects we observed. Further studies should incorporate measures of ATP and metabolites in blood components, should include measures of blood oxygenation and muscle blood flow, and also should investigate the extracellular role of ATP on the neuromuscular junction via Ca2+ mediated effects  as indicators of the potential mechanism by which oral ATP affects the ability to perform strenuous exercise.
Our study, like others in the literature, has limitations. The number of participants in the present study (n=16), while higher than that (n=9) previously studied by Jordan et al. , may not be sufficient to answer all the questions needed to validate the findings. Another limitation may relate to the timing of the last dose of oral ATP (or placebo) given. In our study the last dose was consumed 12 hours prior to testing. This contrasts with the study by Jordan et al. who studied participants after 14 days of supplementation and 3 hours post supplement dosing, and found ATP increased within group set 1 repetitions and total lifting volume . Another potential limitation is that the study involved eumenorrheic females who were not differentiated based upon phase of the menstrual cycle. Other potential limitations include participants’ potential variation in physical activity or diet before testing. However, participants did commit to maintaining their physical activity level for the duration of the study and to exercise restrictions for 3 days prior to testing which within a crossover design should have minimized the effect of activity on the results. Additionally, participants were required to repeat a similar dietary intake 24 h before each testing period and the testing was performed after an overnight fast which should have minimized any acute dietary effect on testing results.