Infusion of ATP increases leg oxygen delivery but not oxygen uptake in the initial phase of intense knee-extensor exercise in humans
Abstract
New Findings
What is the central question of this study?
In the transition from rest to exercise, skeletal muscle blood flow, oxygen delivery and extraction of oxygen from the blood increase to accommodate the need for additional oxygen in the contracting fibres. To what extent skeletal muscle blood flow and oxygen delivery limit the rise in skeletal muscle oxygen uptake in the initial phase of intense exercise remains controversial.
What is the main finding and its importance?
A marked increase in blood flow and oxygen delivery, induced by infusion of ATP, did not affect the increase in oxygen uptake. This finding suggests that oxygen delivery does not limit skeletal muscle oxygen uptake in the initial phase of intense exercise.
The present study examined whether an increase in leg blood flow and oxygen delivery at the onset of intense exercise would speed the rate of rise in leg oxygen uptake. Nine healthy men (25 ± 1 years old, mean ± SEM) performed one-leg knee-extensor exercise (62 ± 3 W, 86 ± 3% of incremental test peak power) for 4 min during a control setting (CON) and with infusion of ATP into the femoral artery in order to increase blood flow before and during exercise. In the presence of ATP, femoral arterial blood flow and O2 delivery were higher (P < 0.001) at the onset of exercise and throughout exercise (femoral arterial blood flow after 10 s, 5.1 ± 0.5 versus 2.7 ± 0.3 l min−1; after 45 s, 6.0 ± 0.5 versus 4.1 ± 0.4 l min−1; after 90 s, 6.6 ± 0.6 versus 4.5 ± 0.4 l min−1; and after 240 s, 7.0 ± 0.6 versus 5.1 ± 0.3 l min−1 in ATP and CON conditions, respectively). Leg oxygen uptake was not different in ATP and CON conditions during the first 20 s of exercise but was lower (P < 0.05) in the ATP compared with CON conditions after 30 s and until the end of exercise (30 s, 436 ± 42 versus 549 ± 45 ml min−1; and 240 s, 705 ± 31 versus 814 ± 59 ml min−1 in ATP and CON, respectively). Lactate release was lower after 60, 120 and 180 s of exercise with ATP infusion. These results suggest that O2 delivery is not limiting the rise in skeletal muscle oxygen uptake in the initial phase of intense exercise.
Introduction
The delivery of oxygen (O2) to the working muscles increases in the transition from rest to exercise in order to meet the higher metabolic demand (Andersen & Saltin, 1985; Grassi 1996; Krustrup et al. 2004a; Jones et al. 2012). O2 has the potential to regulate metabolism, and the role of O2 delivery and conditions in which O2 delivery limits O2 uptake (
) in exercising muscles during the initial phase of exercise have been a topic of interest since the pioneering work by Krogh & Lindhard (1913). In both animal (Grassi et al. 1998a,1998b) and human skeletal muscle (Grassi et al. 1996; MacDonald et al. 1998; Koga et al. 2005; Nyberg et al. 2010; Jones et al. 2012), 
kinetics have been found not to be limited by bulk O2 delivery following the onset of moderate-intensity exercise, but whether this is the case during exercise in the high-intensity domain is not settled (Rossiter, 2011; Poole & Jones, 2012).
During intense exercise, both slow- (ST) and fast-twitch (FT) fibres are recruited, as opposed to moderate-intensity exercise, where predominantly ST fibres are engaged (Krustrup et al. 2004b,2004c). In humans, FT fibres are known to have lower oxidative capacity than ST fibres (Essen-Gustavsson & Henriksson, 1984), and inhibition of ST fibre recruitment leads to a slower 
response (Krustrup et al. 2008). Furthermore, evidence from animals suggests that muscles comprised primarily of FT fibres are perfused less than muscles comprised primarily of ST fibres (Behnke et al. 2003; McDonough et al. 2005), and O2 delivery relative to leg O2 utilization appears to be reduced in an intensity-dependent manner (Nyberg et al. 2010; Jones et al.; Christensen et al. 2013).
Along these lines, a ‘tipping point’ with regard to O2 delivery has been suggested to exist, beyond which 
kinetics become progressively slowed with further reductions in O2 delivery (Poole et al. 2008). Such a point could lie within the more intense domain, where a large proportion of FT fibres are contracting, and skeletal muscle blood flow and O2 delivery could, therefore, limit skeletal muscle 
. In agreement, during intense knee-extensor exercise in humans, the amplitude and kinetics of blood flow to the exercising leg have been suggested to be closely linked to the primary 
kinetics (Paterson et al. 2005). Furthermore, increasing O2 delivery to the exercising forearm has been shown to result in a faster 
response in the initial phase of intense exercise (Perrey et al. 2001; Faisal et al. 2010), and increasing convective O2 delivery to electrically stimulated contractions corresponding to peak 
resulted in faster 
kinetics in canine skeletal muscle (Grassi et al. 2000). In addition, in subjects breathing a hyperoxic gas mixture, which increases arterial O2 pressure, the mean response time was reported to be reduced (Macdonald et al. 1997; Wilkerson et al. 2006). In contrast, a recent study demonstrated that reduced arterial O2 pressure and O2 delivery did not affect leg 
in the initial phase of intense knee-extensor exercise (Christensen et al. 2012). Evidently, the role of O2 delivery in the control of skeletal muscle 
during the initial phase of intense exercise remains to be clarified.
If O2 delivery is limiting skeletal muscle 
during the initial phase of intense exercise, it would be expected that 
kinetics are speeded when O2 delivery to the contracting muscle is increased. Although data from an animal model during these conditions have been presented (Grassi et al. 2000), there is a paucity of direct measurements of human limb 
during intense exercise in conditions of increased blood flow and O2 delivery. Arterial infusion of the vasoactive substance ATP leads to vasodilatation and increased blood flow (González-Alonso et al. 2002), and this procedure may, therefore, be used to increase O2 delivery to the exercising limb in the initial phase of exercise.
The aim of the present study, therefore, was to examine whether an increased O2 delivery induced by femoral arterial ATP infusion in humans would speed leg 
in the initial phase of intense exercise. We hypothesized that the 
response would not be affected by an increase in O2 delivery.
Methods
Subjects
Nine healthy male subjects with a mean (±SEM) age of 25 ± 1 years, body weight of 83 ± 3 kg and height of 186 ± 3 cm participated in the study. All subjects were engaged three or four times per week in regular physical activity, such as soccer, racquet sports, cycling and running. The purpose, nature and potential risks were explained to the subjects before they gave their informed, written consent to participate in the study. The study was approved by the Ethics committee of Copenhagen and Frederiksberg and conducted in accordance with the guidelines of the Declaration of Helsinki. Subject characteristics and data from the control exercise condition for seven of the nine subjects have previously been published (Christensen et al. 2013).
The subjects were informed to maintain their involvement in physical activity throughout the study period, but to abstain from exercise for 48 h and from caffeine and alcohol for 24 h prior to the experimental days. In addition, subjects were informed to consume a light meal 2 h prior to meeting in the laboratory.
Experimental protocol
Preliminary trials
Prior to the main experimental day, the subjects visited the laboratory on three or four occasions in order to become familiar with the single-leg knee-extensor model used. All subjects used the right leg, and they were sitting upright with a hip angle of ∼120 deg. The preliminary trials were also used to establish an exercise intensity that would elicit task failure after 7–10 min, which was defined as a kicking frequency below 55 revolutions min−1 for more than 3 s from the target cadence of 60 revolutions min−1 that was used throughout all trials. During the first visit, the subjects carried out an incremental test, commencing from a baseline of 12 W performed for 10 min, followed by increases in workload of 6 W min−1 until task failure. Incremental test peak power output (iPPO) was calculated at task failure as the sum of the power output (in watts) at the last completed step and the duration (in seconds) at the step leading to task failure/Mean iPPO was 69 ± 3 W.
On the next two or three visits, subjects performed constant-load exercise with a load of ∼85% iPPO for 7 min followed by increases in workload of 6 W min−1 until task failure. If task failure was markedly less or more than the desired 7–10 min, the constant load during the first 7 min of exercise was decreased or increased, respectively, during the next visit. The average work load for the constant-load exercise in the main experiment was 62 ± 3 W (86 ± 3% iPPO).
Main experiment
On the day of the main experiment, the subjects arrived at the laboratory at 08.30 h after a light breakfast. After 30 min in the supine position, catheters were placed into the femoral artery and vein of the experimental leg and the femoral artery of the non-experimental leg under local anaesthesia (lidocaine, 20 mg ml−1). After 30 min of additional rest, the subjects were placed in the knee-extensor model and rested for an extra 15 min. Thereafter, the leg was moved passively for 1 min followed by 4 min of high-intensity constant-load exercise (62 ± 3 W) in the upright position in two conditions, i.e. without (control; CON conditions) or with infusion of ATP into the femoral artery of the working leg in order to induce vasodilatation of the leg vasculature (ATP conditions). The order of CON and ATP conditions was randomized and separated by 90 min. To minimize any influence from performing prior exercise, a 4 min warm-up was performed 30 min prior to both CON and ATP, consisting of 2 min at 35 ± 1 W (50 ± 2% iPPO) and 2 min at 49 ± 2 W (70 ± 2% iPPO). The ATP (32 μmol min−1) was infused into the femoral artery for 1 min prior to and during the 1 min of passive movement of the leg and throughout exercise. To test any potential performance-enhancing effect of ATP infusion, two of the subjects performed the exercise until task failure in a randomized order (these trials were separated by 120 min). Blood samples (1–5 ml) were drawn from the femoral vein of the experimental leg at rest, during the passive movement of the leg 15 s before active exercise and after 5, 10, 15, 20, 30, 45, 60, 90, 120, 180 and 240 s of exercise. Arterial blood was drawn from the non-experimental leg ∼5 s prior to the venous samples at rest, 30 s before active exercise and after 10, 40, 55, 85, 115, 175 and 235 s of exercise. Femoral arterial blood flow was measured continuously at the same time as blood samples were drawn.
Data acquisition and analyses
Femoral arterial blood flow was measured with an ultrasound machine (Logic E9; GE Healthcare, Copenhagen, Denmark) equipped with a linear probe operating with an imaging frequency of 9 MHz and Doppler frequency of 3.1 MHz. The site of blood velocity measurements in the common femoral artery was distal to the inguinal ligament but above the bifurcation into the superficial and deep femoral branches to avoid turbulence from the bifurcation. All recordings were obtained at the lowest possible insonation angle and always below 60 deg. The sample volume maximized according to the width of the vessel and kept clear of the vessel walls. A low-velocity filter (velocities <1.8 m s−1) rejected noises caused by turbulence at the vascular wall. Doppler traces and B-mode images were recorded continuously, and Doppler traces were averaged over eight heart cycles at the time of blood sampling. Arterial diameter measures were assessed during systole from arterial B-mode images with the vessel parallel to the transducer. Measurements obtained with this method are reproducible and have the same order of magnitude and similar accuracy to thermodilution venous outflow measurements (Rådegran, 1997) and can be used to measure arterial inflow to dynamically contracting muscle in the transition from rest to exercise at different intensities up to levels approaching peak power output (Rådegran & Saltin, 1998). Arterial pressure (in millimetres of mercury) was monitored with transducers positioned at the level of the heart (Pressure Monitoring Kit; Baxter, Deerfield, IL, USA). Arterial and venous blood samples were immediately analysed for 
, 
, pH, O2 saturation, haemoglobin, lactate and potassium (ABL725; Radiometer, Copenhagen, Denmark).
Calculations
Leg O2 delivery was calculated as femoral arterial blood flow multiplied by the arterial O2 content, and leg 
was calculated as femoral arterial blood flow multiplied by the femoral arteriovenous O2 difference. To determine O2 extraction and 
at the capillary level, corrections were made for the transit times from the capillaries to the collection points in the femoral artery and vein based on the mean transit time from artery to vein measured in various phases of intense exercise, with approximately one-third of the time representing the time from artery to capillary and approximately two-thirds of the time representing the time from capillary to vein (Bangsbo et al. 2000). All blood variables are presented in relation to mean time at the capillary level. Lactate release was calculated as femoral arterial blood flow multiplied by the venous–arterial difference. Potassium release was calculated as femoral arterial blood flow multiplied by the venous–arterial difference (adjusted for changes in plasma fraction based on hematocrit measurements).
Statistical analysis
Changes in the different variables during the entire period of exercise (passive and active) were examined with a two-way ANOVA using time and intervention (CON or ATP) as factors. If a significant main effect was found, a Student–Newman–Keuls post hoc test was performed to locate the differences. All values are means ± SEM.
Results
Cardiovascular response
At rest, femoral arterial blood flow was ∼13-fold higher (P < 0.001) in ATP compared with CON (4.9 ± 0.3 versus 0.4 ± 0.0 l min−1, respectively; Fig. 1A). Femoral arterial blood flow was higher (P < 0.001) in ATP compared with CON during passive movement of the leg and throughout exercise (10 s, 5.1 ± 0.5 versus 2.7 ± 0.3 l min−1; 45 s, 6.0 ± 0.5 versus 4.1 ± 0.4 l min−1; 90 s, 6.6 ± 0.6 versus 4.5 ± 0.4 l min−1; and 240 s, 7.0 ± 0.6 versus 5.1 ± 0.3 l min−1 in ATP and CON, respectively).
In this and all subsequent figures, data are expressed as means ± SEM. Significant difference from CON: *P < 0.05, **P < 0.01 and ***P < 0.001.
Mean arterial blood pressure at rest was and 90 ± 6 in ATP and 97 ± 2 mmHg in CON (Fig. 1B). During passive movement of the leg, mean arterial blood pressure was lower (P < 0.05) in ATP than in CON (92 ± 6 versus 101 ± 1 mmHg) and during exercise, mean arterial blood pressure increased to ∼135 mmHg in ATP and CON, with no difference between the two trials at any time point.
Leg vascular conductance was ∼15-fold and fivefold higher (P < 0.001) at rest and during passive movement of the leg in ATP compared with CON (56.5 ± 3.5 versus 3.9 ± 0.5 and 54.1 ± 6.0 0 versus 11.6 ± 2.0 ml min−1 mmHg−1; Fig. 1C). Leg vascular conductance remained ∼40–150% higher in ATP until 180 s of exercise (P < 0.001; Fig. 1C).
Leg oxygen delivery, oxygen uptake and arteriovenous difference
At rest, leg O2 delivery was ∼14-fold higher (P < 0.001) in ATP compared with CON (996 ± 86 versus 73 ± 9 ml min−1; Fig. 2A). Leg O2 delivery was higher (P < 0.001, first 180 s; P < 0.01, 240 s; Fig. 2A) in ATP compared with CON during passive movement of the leg and throughout exercise (10 s, 1038 ± 113 versus 536 ± 58 ml min−1; 45 s, 1246 ± 125 versus 852 ± 88 ml min−1; 90 s, 1384 ± 149 versus 950 ± 96 ml min−1; and 240 s, 1406 ± 95 versus 1072 ± 74 ml min−1 in ATP and CON, respectively).
Significant difference from CON: *P < 0.05, **P < 0.01 and ***P < 0.001.
Leg 
was not different between CON and ATP at rest, during passive movement of the leg and during the first 20 s of exercise, but leg 
was lower (P < 0.05 and P < 0.01) after 30 s and until end of exercise in ATP compared with CON (Fig. 2B).
The difference between leg O2 delivery and leg 
was higher (P < 0.001) in ATP compared with CON at rest as well as during passive movement of the leg and during exercise (10 s, 887 ± 117 versus 334 ± 52 ml min−1; and 45 s, 762 ± 106 versus 254 ± 37 ml l−1; Fig. 2C).
At rest, leg arteriovenous O2 difference was lower (P < 0.001) in ATP compared with CON (27 ± 6 versus 63 ± 7 ml l−1) and remained lower (P < 0.001) throughout passive movement of the leg and during exercise (10 s, 32 ± 5 versus 77 ± 7 ml l−1; and 45 s, 82 ± 8 versus 146 ± 6 ml l−1; Fig. 2D).
Blood gases and pH
At rest, femoral arterial 
was higher (P < 0.01) in ATP compared with CON (110 ± 3 versus 101 ± 1 mmHg; Fig. 3A). No difference in femoral arterial 
was detected during passive movement of the leg and during exercise. Femoral venous 
was higher (P < 0.001) in ATP compared with CON at rest as well as during passive movement of the leg and during exercise.
Significant difference from CON: **P < 0.01 and ***P < 0.001.
Femoral arterial 
remained at ∼40 mmHg throughout exercise in both ATP and CON (Fig. 3B). Femoral venous 
was lower (P < 0.001) in ATP compared with CON after 45 s of exercise (49.4 ± 1.9 versus 59.6 ± 1.5 mmHg) and remained lower (P < 0.001) until the end of exercise.
Femoral arterial pH was ∼7.40 throughout exercise, with no difference between CON and ATP (Fig. 3C). At rest, femoral venous pH was higher (P < 0.01) in ATP compared with CON and remained higher (P < 0.001) throughout exercise.
Lactate and potassium release
At rest, femoral arterial lactate was ∼1.3 mmol l−1 in both CON and ATP, and it increased to a similar extent until after 120 s of exercise, when femoral arterial lactate was higher (P < 0.01) in ATP compared with CON until the end of exercise (Fig. 4A). Femoral venous lactate was lower (P < 0.05) in ATP compared with CON after 60 s of exercise but not at any other time point. At rest, leg lactate release was −1.4 ± 0.6 and −0.1 ± 0.0 mmol min−1 in ATP and CON, respectively, and increased to ∼10 mmol min−1 at the end of exercise in both ATP and CON (Fig. 4B). Leg lactate release was lower (P < 0.05) after 60, 120 and 180 s of exercise in ATP compared with CON.
Significant difference from CON: *P < 0.05 and **P < 0.01.
At rest, femoral arterial potassium was ∼3.8 mmol l−1 in both CON and ATP and it increased to a similar extent to ∼4.8 mmol l−1 at the end of exercise (Fig. 5A). Femoral venous potassium was lower (10–30 and 60 s, P < 0.05; 45 s, P < 0.001) in ATP compared with CON (10 s, 4.2 ± 0.1 versus 4.5 ± 0.1 mmol l−1). Leg potassium release was 0.5 ± 0.4 and 0.0 ± 0.0 mmol min−1 at rest in ATP and CON, respectively, and increased to a similar extent to 0.9 ± 0.5 and 1.2 ± 0.4 mmol min−1 at the end of exercise in ATP and CON, respectively (Fig. 5B).
Significant difference from CON: *P < 0.05 and ***P < 0.001.
Discussion
The primary finding of the present study was that an ATP-induced increase in femoral arterial blood flow and O2 delivery in the initial phase of intense knee-extensor exercise did not affect the rise in leg 
during the first 20 s. Furthermore, leg 
was lower during ATP infusion after ∼30 s of exercise and until the end of exercise, with lactate release being lower after 60, 120 and 180 s of exercise with ATP infusion.
In the present experimental set-up, leg 
in the initial phase of exercise was not affected despite an increase in blood flow and O2 delivery by ∼75–130% to the exercising limb. These findings suggest that the rise in leg 
during intense exercise with a small muscle mass is not limited by O2 delivery and is set by other (intracellular) factors, such as an inertia of one or more of the enzymes of oxidative metabolism (Poole et al. 2008) or temporal energy buffering by creatine kinase (Grassi et al. 2011). In agreement, increasing the arterial O2 pressure by breathing a hyperoxic gas does not seem to speed skeletal muscle 
kinetics during the initial phase of intense cycling (Wilkerson et al. 2006), and increasing blood O2-carrying capacity through treatment with recombinant human erythropoietin had no influence on pulmonary 
kinetics during intense cycling exercise (Wilkerson et al. 2005). Furthermore, lowering arterial O2 pressure and O2 delivery did not affect leg 
during intense exercise (Christensen et al. 2012).
In contrast to the findings above, increasing O2 delivery has been shown to be associated with a small increase in the skeletal muscle 
response in electrically stimulated canine skeletal muscle (Grassi et al. 2000), which to some extent may be related to species differences. Furthermore, elevating O2 delivery to the exercising forearm also resulted in a faster 
response in the initial phase of intense exercise when the subjects were in a supine position with the contracting muscles positioned at the level of the heart (Perrey et al. 2001; Faisal et al. 2010). Likewise, both blood flow and 
have been reported to be slower in the supine relative to the upright position (MacDonald et al. 1998), indicating that O2 delivery could be limiting skeletal muscle 
in the supine position due to a lower perfusion pressure, hence lower blood flow.
The present findings suggest that skeletal muscle 
in the initial phase of intense exercise is not limited by O2 delivery during exercise with a small muscle mass in an upright position, but it does not seem to be the case when O2 delivery is markedly lowered. We have recently shown that a reduction in leg O2 delivery by ∼25–50% in the initial phase of intense knee-extensor exercise led to a slowing of the rise in leg 
(Christensen et al. 2013). Thus, with this substantial reduction in blood flow, leg O2 delivery was also found to be very close to leg 
, suggesting that 
in some of the contracting fibres was limited by O2 delivery. By controlling the blood flow on-kinetics mean response time during electrically stimulated contractions, it was shown that canine skeletal muscle 
on-kinetics response was slowed in proportion to the reduction in convective O2 delivery when mean response time was increased by 125 and 250% (Goodwin et al. 2012). Although the contractions elicited a 
corresponding to only 50–70% of peak 
, this finding, along with that in humans (Christensen et al. 2013), suggest that a marked reduction in O2 delivery can affect skeletal muscle 
in the initial phase of both moderate-intensity and intense exercise.
Notably, as the exercise intensity in the present study was ∼86% of incremental test peak power output, future investigations measuring limb 
should focus on the effect of increasing limb O2 delivery during even more intense contractions, because leg blood flow and O2 delivery relative to leg O2 utilization appear to be reduced in an intensity-dependent manner (Nyberg et al. 2010; Jones et al. 2012; Christensen et al. 2013).
An interesting finding in the present study was that leg 
after 30 s and until the end of exercise was lower during infusion of ATP compared with CON. This effect of ATP infusion is in agreement with a previous study, where infusion of ATP into the femoral artery during maximal cycling was found to reduce leg 
by 6% (Calbet et al. 2006). A lower leg 
in the ATP trial could be related to the ATP-induced change in blood pH and CO2 levels. Accordingly, femoral venous pH and CO2 levels were found to be higher and lower, respectively, which increases the affinity for O2 in the red blood cell, thereby decreasing the off-loading of O2 (Jensen, 2004). It should be noted, however, that the femoral vein is draining both active and inactive tissue, and the higher pH and lower CO2 levels may have been caused by an increased perfusion of non-contracting tissues. Thus, the femoral venous pH and CO2 levels may not reflect the microvascular conditions in the active skeletal muscle. The observation that lactate release was lower with ATP infusion may reflect that the reduced leg 
was not compensated for by an increased anaerobic metabolism, which would entail that mechanical efficiency of the muscle was improved with ATP infusion. As quantification of the utilization of creatine phosphate and accumulation of lactate in the active skeletal muscle was not performed, such an effect of ATP on mechanical efficiency remains to be determined. However, an improvement in mechanical efficiency of 15% seems improbable. Given that leg 
is the product of femoral arterial blood flow and the leg arteriovenous O2 difference, a lower leg 
in the ATP trial could be the result of an underestimation of blood flow caused by the augmented exercise hyperaemia. However, measurements of blood flow obtained with thermodilution and ultrasound are closely correlated irrespective of the magnitude of blood flow (Rådegran, 1997), suggesting that the lower leg 
in the ATP trial was not related to an underestimation of blood flow.
Whatever the physiological mechanism underlying the lower leg 
in the ATP trial after 30 s of exercise may have been, this could also potentially have been acting during the first part of the exercise transition. In this scenario, leg 
during the first 20 s of exercise would be underestimated, but more evidence is needed to substantiate such an effect of ATP infusion on leg 
during this part of exercise.
A much greater difference in blood flow and vascular conductance was observed between conditions at rest compared with exercise due to a larger increase in these variables with exercise in the control setting. This is most likely to be a result of the active fibres in the quadriceps muscle being highly perfused already before exercise in the ATP trial and, therefore, the contraction-induced increase in blood flow would be less. In accordance, the vasodilator effect of ATP infusion during near-maximal cycling exercise was found to be less than that during rest (Calbet et al. 2006), suggesting that parts of the musculature were highly perfused and therefore less responsive to the vasoactive effects of ATP.
Addition of ATP has been shown to increase the Na+–K+ pump activity and excitability of depolarized skeletal muscles (Broch-Lips et al. 2010), an effect that is likely to be mediated via P2Y receptors on the sarcolemma (Walas & Juel, 2012). Given the putative role of extracellular potassium for fatigue development during intense exercise (McKenna et al. 2008), infusion of ATP could potentially have influenced muscle excitability, which would have been likely to improve muscle performance. However, two of the subjects performed both exercise trials until task failure and, in both instances, they fatigued earlier during the ATP infusion (219 ± 66 versus 312 ± 88 s, 30 ± 2% difference). Although it is not possible to make valid conclusions based on observations on two subjects, the pronounced impairment in both subjects does indicate that ATP infusion could be detrimental to performance. The reason for this may be that the infused ATP did not reach the contracting skeletal muscle cells, because ATP infusion of a similar dose during resting conditions has previously been reported not to increase interstitial ATP concentrations (Mortensen et al. 2009).
Interestingly, venous levels of potassium were lower during the first 60 s of exercise with ATP, whereas leg potassium release was found to be unaltered with ATP infusion throughout exercise. If ATP did reach the contracting muscle fibres and increased Na+–K+ pump activity, it would be expected that venous levels were lower as a result of reduced interstitial potassium levels; however, potassium release would also be expected to be lower. With regard to the potassium release, the increased perfusion of inactive tissue with ATP infusion during exercise would cause an uptake of potassium in these regions, because the arterial concentrations of potassium were higher than the levels known to exist in resting skeletal muscle (∼4.2 mmol l−1; Juel et al. 2000). In this scenario, the unaltered net leg release of potassium with ATP infusion would have to be the result of an increased release of potassium from the active fibres, most probably as an effect of higher interstitial potassium levels. Such an effect of ATP infusion on interstitial potassium levels could explain the more rapid development of fatigue (McKenna et al. 2008). The lower working capacity with ATP infusion could, therefore, be linked to this effect of ATP on interstitial potassium and/or the reduced leg 
.
Summary
Increasing the blood flow and O2 delivery to the leg in the initial phase of intense knee-extensor exercise by infusion of ATP did not affect leg 
, suggesting that O2 delivery is not limiting the rise in skeletal muscle 
when a small muscle mass is engaged. The mechanism underlying the lower leg 
after 30 s and until the end of exercise with ATP infusion remains to be determined.
References
peak and leg O2 extraction during maximal exercise in humans
on-kinetics in isolated canine muscle in situ during slowed convective O2 delivery
on-kinetics in isolated in situ canine muscle
on-kinetics in isolated in situ canine muscle
on-kinetics in canine muscle contracting at peak 
kinetics in heavy submaximal exercise by hyperoxia and prior high-intensity exercise
and femoral artery blood flow during heavy-intensity, knee-extension exerciseAdditional Information
Competing interests
None declared.
Author contributions
M.N., P.M.C., Y.H. and J.B. conceived and designed the research; M.N., P.M.C., S.P.M. and J.B. performed experiments; M.N., P.M.C. and S.P.M. analysed data; M.N., P.M.C., and J.B. prepared figures; M.N., P.M.C., S.P.M., Y.H. and J.B. interpreted results of experiments, drafted the manuscript, edited and revised the manuscript and approved the final version of manuscript.
Funding
The study was supported by the Novo Nordisk Foundation.
