Original Research

Changes in Blood Concentration of Adenosine

Triphosphate Metabolism Biomarkers During
Incremental Exercise in Highly Trained Athletes of
Different Sport Specializations
Michał Włodarczyk,1 Krzysztof Kusy,1 Ewa Słomińska,2 Zbigniew Krasiński,3 and Jacek Zieliński1
1
Human Movement Laboratory “LABTHLETICS”, Department of Athletics, Strength and Conditioning, Poznan University of Physical
Education, Poznań, Poland; 2Department of Biochemistry, Medical University of Gdansk, Gdańsk, Poland; and 3Department of
Vascular and Endovascular Surgery, Angiology and Phlebology, Poznan University of Medical Sciences, Poznań, Poland

Abstract
Włodarczyk, M, Kusy, K, Słomińska, E, Krasiński, Z, and Zieliński, J. Changes in blood concentration of adenosine triphosphate
metabolism biomarkers during incremental exercise in highly trained athletes of different sport specializations. J Strength Cond Res
XX(X): 000–000, 2019—We hypothesized that (a) high-level specialized sport training causes different adaptations that induce
specific biomarker release dynamics during exercise and recovery and (b) skeletal muscle mass affects biomarker release. Eleven
sprinters (21–30 years), 16 endurance runners (18–31 years), 12 futsal players (18–29 years), and 12 amateur runners as controls
(22–33 years) were examined. Hypoxanthine (Hx), xanthine (X), uric acid (UA), ammonia (NH3), and lactate (LA) concentrations were
determined at rest, during an incremental treadmill exercise test (every 3 minutes), and during recovery (5, 10, 15, 20, and 30
minutes after exercise). Hx, X, and UA concentration was determined from plasma, while LA and NH3 from whole blood, and muscle
mass was assessed using dual X-ray absorptiometry method. At rest, during incremental exercise, and up to 30 minutes into the
postexercise recovery period, sprinters had lowest Hx, X, and UA concentrations, and endurance athletes had lowest NH3
concentrations. For LA during exercise, the lowest concentrations were noted in endurance athletes, except when reaching
maximum intensity, where the differences between groups were not significant. There were no significant correlations observed
between skeletal muscle mass and biomarker concentration at maximal intensity and recovery in any group. In conclusion, the
magnitude of exercise-induced biomarker concentration is only related to training adaptations through specific training profile but
not to muscle mass. In addition, the results suggest that combined measuring of LA, NH3, and Hx concentration in blood is useful in
indirectly reflecting key changes in exercise- and training-induced energy status. Further research should focus on studying how
specific training sessions affect individual biomarker response in highly trained athletes.
Key Words: lactate, ammonia, hypoxanthine, xanthine, uric acid, training

Introduction exercise (10). As exercise duration and intensity increase, the

Skeletal muscles are the main organs used during exercise. Un-
fortunately, performing muscle biopsies on highly trained athletes
is unpractical, and other methods such as blood measurement
must be used to reflect muscle biomarker concentration levels. In
addition, highly trained athletes also possess different levels of
muscle mass which can affect magnitude of biomarker release.
Because muscles use adenosine triphosphate (ATP) for energy
production, measuring ATP metabolism biomarkers can help to
better understand energy system utilization during different types
of training and exercise (1,3,10,11,20,33). Mechanisms related to
depletion and replenishing of ATP stores due to increased con-
sumption or decreased provision are reflected by blood concen-
tration of 3 basic biomarkers, i.e., lactate (LA), ammonia (NH3),
and oxypurines (7), as well as decreased power output during
Address correspondence to Jacek Zieliński, jacekzielinski@wp.pl.
Supplemental digital content is available for this article. Direct URL citations appear
in the printed text and are provided in the HTML and PDF versions of this article on
the journal’s Web site (http://journals.lww.com/nsca-jscr).
Journal of Strength and Conditioning Research 00(00)/1–9
ª 2019 National Strength and Conditioning Association
adenine nucleotide (AdN) pool and the ATP/adenosine di-
phosphate ratio decreases (7), which causes specific patterns and
magnitude of accumulation of these biomarkers in blood.
Biomarkers such as LA (3,5,8,9) and NH3 (1,10,11,18,19,31–34)
have been extensively researched in difference stages of exercise
(1,11,15,18,19,32–34). New purine metabolite biomarkers have
been studied during and after exercise (13,22,30), after short-term
training lasting 1–7 weeks (27–29), during a 1-year training cycle in
highly trained athletes (36,38–40), and even considered a universal
indicator of training status (37) as well as a predictor of performance
in highly trained athletes (35). To our knowledge, there is an in-
adequate amount of research, studying ATP metabolism biomarker
efflux before, during exercise, and during the postexercise recovery
period in highly trained athletes. Sjodin and Hellsten-Westing (24)
studied only hypoxanthine (Hx) and uric acid (UA) before, during,
and after exercise only in 4 well-trained runners. Ogino et al. (19)
studied LA, NH3, and Hx before, during, and after exercise but only
in healthy male volunteers. Finally, Stathis et al. (28) studied LA and
purine derivatives in 7 active but nonspecifically trained males with
the main focus on biomarker release during recovery as opposed to
release during exercise. As shown, there is a lack of research

Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 ATP Metabolism Biomarkers During Exercise (2019) 00:00
reflecting a greater range of ATP biomarker release around exercise
(rest, exercise, and recovery) in highly trained athletes in the same
training phase. Moreover, the contribution of specific muscle fibers,
genetically predetermined or resulting from training adaptation, is of
importance. Fast-twitch muscle fibers contain more ATP and phos-
phocreatine (16), are more glycolytically active (5) suited for LA
production (8), and contain more AMP deaminase leading to greater
AMP deamination producing more NH3 (13,33) and ultimately
more Hx in muscle compared with slow-twitch muscle fibers. Fur-
thermore, in the context of exercise, there were no attempts to elu-
cidate the relationships between skeletal muscle mass (SMM) and
ATP metabolism biomarker concentration. It is unknown whether
changes in ATP metabolism biomarker concentrations are related to
the changes in muscle mass. Greater muscle mass is related to the
total amount of ATP-PCr in muscle that can be used during exercise
and increases the amount of ATP in muscle that can be produced
through anaerobic glycolysis (21). Because muscle mass levels and
muscle properties are different in athletes of diverse physiological
and training profiles, this could cause different release dynamics
throughout exercise.
The aim of this study was to determine changes in blood con-
centration of ATP metabolism biomarkers (LA, NH3, and purine
metabolites) during graded exercise and in the recovery period, in
athletes of different training profiles and differences in SMM. We
hypothesize that (a) high-level specialized sport training causes
different adaptations resulting in specific exercise-related release
dynamics of ATP metabolism biomarkers during standard exer-
cise and in the postexercise recovery period, and that (b) differ-
ences in muscle mass of athletes will affect biomarker release.
Methods
Experimental Approach to the Problem
To obtain the full-blood ATP biomarker spectrum of release
around exercise, the testing procedures included an incremental
exercise test, where blood samples were obtained at rest, whereas
subjects ran on a mechanical treadmill with increasing speed, and
up to 30 minutes after exercise. Elite athletes of different sport
specializations were chosen to represent diverse metabolic and
physiological adaptations (speed-power, endurance, and mixed)
to see whether long-term specialized training would produce
different biomarker responses. Subjects underwent body com-
position analysis to obtain SMM levels. A control group of rec-
reationally trained athletes was used to represent a group of
general training adaptations to exercise contrary to the highly
specialized athletic groups.
Subjects
Four groups of athletes participated in the study: sprinters (SP,
n 5 11) specialized in the 100 and 200 m events, aged 24.2 6 3.2
(range 21–30) years, and practicing sport for 8.5 6 2.5 years,
endurance athletes (EN, n 5 16) consisting of triathletes and long-
distance runners aged 23.4 6 3.6 (range 18–31) years, and
practicing sport for 8.7 6 1.9 years, futsal players (FU, n 5 12)
aged 24.5 6 3.8 (range 18–29) years, and practicing sport for
10.0 6 3.4 years, and amateur runners (AM, n 5 12) representing
the control group aged 27.7 6 4.1 (range 22–33) years with no
past or current competitive sport history (recreational sports ac-
tivities 3–5 times per week). A more detailed description of the
subjects is presented in Table 1. Sprinters, EN, and FU were
members of the Polish National Team and were tested at the end
Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
of the transition period of their annual training cycle (beginning
of preparatory phase). Controls, however, were not specifically
trained, recreationally performed common forms of physical ac-
tivity (e.g., jogging, swimming, and team games) in their spare
time, and did not exceed the 150 minutes of moderate-intensity
physical activity per week recommended by the World Health
Organization. The project was approved by the Ethics Committee
at the Poznan University of Medical Sciences and was performed
according to the ethical standards laid down in the Declaration of
Helsinki. Each subject was informed of the testing procedure,
purpose, and risks of the study and submitted their written con-
sent to participate.
Procedures
Subjects were instructed not to participate in any high-intensity or
long-duration training sessions at least 24–48 hours before test-
ing. Testing was performed in the morning hours, 3 hours after
a light breakfast (no coffee or tea). Before the exercise test, sub-
jects underwent body composition analysis. Afterward, an in-
cremental treadmill exercise test until volitional exhaustion was
performed. During all examinations, room temperature remained
at 20–21° C.
Anthropometric and Body Composition Analysis. The anthro-
pometric measurements taken were body mass (kg) and height
(cm) using a digital stadiometer (SECA 285; SECA, Hamburg,
Germany). Body mass index was calculated by dividing body
mass by height squared. The dual X-ray absorptiometry method,
using Lunar Prodigy Pro (GE Healthcare, Madison, WI, USA)
and enCORE v. 16 SP1 software, was used for body composition
analysis. During examination, subjects only wore their under-
garments, without jewelry and metal objects to minimize mea-
surement error. Skeletal muscle mass was calculated based on
regression models appropriate for age and sex (17).
Respiratory Parameters. An incremental exercise test until voli-
tional exhaustion on a mechanical treadmill (H/P Cosmos Pulsar,
Sports & Medical, Nussdorf-Traunstein, Germany) was per-
formed to determine the maximal oxygen uptake (V̇ O2max).
Initial speed was set at 4 km·h21 and increased after 3 minutes to
8 km·h21. After that point, treadmill speed increased pro-
gressively by 2 km·h21 every 3 minutes until volitional exhaus-
tion. Once a 10-km·h21 speed was reached, blood samples were
drawn from the participant at the end of each 3-minute stage.
Respiratory parameters were measured (breath by breath) by
an ergospirometer (Cortex Metamax 3B R2, Leipzig, Germany)
and analyzed using MetasoftStudio v. 5.1.0 Software (Cortex-
Metamax 3B R2; Cortex Biophysik, Leipzig, Germany). Heart
rate (HR, b·min21) was monitored with a Polar Bluetooth Smart
H6 (Polar Electro Oy, Kempele, Finland) HR monitor. All sub-
jects were familiar with the exercise protocol since they already
performed this test previously.
Blood Sampling. Subjects wore a catheter (BD Venflon Pro 1.3
3 32 mm; Becton Dickinson, Helsingborg, Sweden) patent
with isotonic saline (0.9% NaCl), from which blood was drawn
from one of the antecubital veins. Blood samples were collected
at rest, at the end of each 3-minute stage above 10 km·h21
treadmill speed, immediately after exercise, and 5, 10, 15,
20, and 30 minutes into the postexercise recovery period.
Later, a 2.7-ml blood sample was taken into 2 monovettes
 ATP Metabolism Biomarkers During Exercise (2019) 00:00 | www.nsca.com

Table 1
Basic characteristics of the athletic groups and controls.*†
Futsal players Endurance athletes Sprinters Control group One-way Effect
(n 5 12) (n 5 16) (n 5 11) (n 5 12) ANOVA size
Age (y) 24.50 6 3.8 (22.1–26.9) 23.38 6 3.6 (21.5–25.3) 24.18 6 3.2 (22.0–26.3) 27.70 6 4.1‡ (25.1–30.3) 0.025 0.18
Training experience (y) 10.00 6 3.4 (7.8–12.2) 8.69 6 1.9 (7.7–9.7) 8.45 6 2.5 (6.8–10.1) — 0.299 0.06
Height at examination (cm) 181.38 6 5.5 (177.9–184.9) 181.80 6 6.1 (178.6–185.0) 186.19 6 4.6 (183.1–189.3) 180.44 6 5.6 (176.9–184.0) 0.080 0.13
Mass at examination (kg) 77.12 6 7.2 (72.6–81.7) 73.41 6 7.1 (69.6–77.2) 81.61 6 5.5‡ (77.9–85.3) 76.72 6 7.7 (71.8–81.6) 0.039 0.16
BMI (kg·m22) 23.47 6 2.2 (22.1–24.8) 22.23 6 2.1 (21.1–23.4) 23.53 6 1.0 (22.9–24.2) 23.68 6 2.6 (22.0–25.3) 0.221 0.09
SMM (kg) 33.65 6 3.2{ (31.6–35.7) 32.40 6 3.2# (30.7–34.1) 39.08 6 3.7 (36.6–41.6) 33.12 6 3.1{ (31.2–35.1) ,0.001 0.39
HRVT (b·min21) 146.33 6 14.7 (137.0–155.7) 156.31 6 13.8 (149.0–163.7) 147.09 6 15.4 (136.8–157.4) 151.50 6 13.6 (142.8–160.2) 0.244 0.08
VO2VT (L·kg21) 2.97 6 0.4 (2.7–3.2) 3.26 6 0.5 (3.0–3.5) 2.70 6 0.5‡ (2.3–3.1) 3.02 6 0.4 (2.8–3.3) 0.034 0.17
VO2VT (ml·kg21·min21) 38.45 6 3.9‡ (35.9–40.9) 44.52 6 6.1 (41.3–47.8) 33.25 6 6.9‖ (28.6–37.9) 39.44 6 3.9 (37.0–41.9) ,0.001 0.38
21
VO2VT (ml·kg smm · 88.16 6 9.4{ (82.2–94.1) 100.88 6 14.1# (93.3–108.4) 69.61 6 14.6 (59.8–79.5) 91.14 6 6.8{ (86.8–95.5) ,0.001 0.49
min21)
HRRCP (b·min21) 174.83 6 11.3 (167.6–182.0) 181.61 6 9.5 (176.6–186.7) 179.09 6 8.4 (173.5–184.7) 179.17 6 10.2 (172.7–185.6) 0.366 0.06
VO2RCP (L·kg21) 3.73 6 0.5‡ (3.4–4.0) 4.29 6 0.5 (4.0–4.6) 3.84 6 0.4 (3.6–4.1) 4.05 6 0.6 (3.7–4.4) 0.025 0.18
21 21
VO2RCP (ml·kg ·min ) 48.36 6 4.3‖ (45.6–51.1) 58.55 6 5.6 (55.6–61.5) 47.08 6 4.4‖ (44.1–50.1) 52.82 6 4.6‡ (49.9–55.8) ,0.001 0.50
VO2RCP (ml·kg 110.93 6 10.9‖ (104.0–117.9) 132.82 6 14.6 (125.0–140.6) 98.60 6 10.2‖ (91.8–105.4) 122.41 6 12.8# ,0.001 0.54
smm21·min21) (114.3–130.5)
HRmax (b·min21) 187.50 6 10.0 (181.1–193.9) 191.94 6 8.5 (187.4–196.5) 188.27 6 9.9 (181.6–195.0) 187.33 6 7.7 (182.4–192.2) 0.482 0.05
VȮ 2max (L·kg )
21
4.44 6 0.5 (4.2–4.7) 4.91 6 0.6 (4.6–5.2) 4.34 6 0.4‡ (4.1–4.6) 4.31 6 0.4‡ (4.1–4.6) 0.004 0.24
VȮ 2max (ml·kg21· 57.53 6 1.1‖ (56.8–58.2) 66.86 6 5.0 (64.2–69.5) 53.27 6 3.6‖ (50.8–55.7) 56.34 6 2.7‖ (54.6–58.1) ,0.001 0.71
min21)
VȮ 2max (ml·kg 132.01 6 6.4§ (128.0–136.1) 151.65 6 13.5 (144.5–158.8) 111.54 6 8.8‖ (105.6–117.5) 130.46 6 7.0‖# ,0.001 0.70
smm21·min21) (126.0–134.9)
Time to complete test 17.14 6 1.4§ (16.3–18.0) 19.11 6 1.2 (18.5–19.8) 16.24 6 1.2‖ (15.4–17.0) 16.51 6 1.2‖ (15.7–17.3) ,0.001 0.50
(min)

*ANOVA 5 analysis of variance; BMI 5 body mass index; SMM 5 skeletal muscle mass; VO2VT 5 oxygen consumption at ventilatory threshold; VO2RCP 5 oxygen consumption at respiratory compensation
point; VȮ 2max 5 maximal oxygen uptake.
†Values are mean values 6 SDs (confidence intervals 95%).
‡Post hoc tests: p , 0.05, significantly different from endurance athletes.
§p , 0.01, significantly different from endurance athletes.
‖p , 0.001, significantly different from endurance athletes.
{p , 0.01, significantly different from sprinters.
#p , 0.001, significantly different from sprinters.

(S-Monovette 2.7 ml KE; Sarstedt, Nümbrecht, Germany) one
with a lithium anticoagulant (heparin) and another containing
an anticoagulant (EDTA).
Lactate and Ammonia. Biosen C-line (EKF diagnostic GmbH,
Barleben, Germany) was used to measure lactate level using
whole blood. To determine lactate level, 20 ml of whole blood
was placed into a capillary. The apparatus measurement accu-
racy (CV) was 1.5% for 12 mmol·L21 concentration. To de-
termine ammonia level, PocketChem BA PA-4140 (Arkay,
Kyoto, Japan) was used. To perform a measurement on a testing
strip (Ammonia Test Kit II; Arkay), 20 ml of blood was placed
using a pipette. The apparatus measuring range is 8–285
mmol·L21. The apparatus measurement accuracy (CV)
was 2.3%.
Purine Metabolites (Hx, X, UA). The vial (Eppendorf,
Wesseling-Berzdorf, Germany) containing the blood mixture
with EDTA was centrifuged (Universal 320R; Hettich Lab
Technology, Tuttlingen, Germany) at 4° C for 30 seconds at
14,000 rpm (using short spin function). After centrifugation, 0.2
ml of plasma was pipetted into 1.5 ml of vials (Eppendorf,
Wesseling-Berzdorf), in duplicate and immediately frozen in
liquid nitrogen. All samples were then stored in a freezer (280°
C). At the time of extraction, 0.2 ml of 1.2 mol·L21 HClO4 was
added to frozen plasma for deproteinization. After centrifuga-
tion, to remove the protein pellet an acid supernatant was
neutralized with 3 mol·L21 K3PO4, centrifuged to remove pre-
cipitated KClO4, and stored in 280° C.
Purines in plasma were determined by high liquid performance
chromatography (HPLC) with UV detection as described earlier
(26). The analyses were performed using 1100 HPLC system
(HPLC System; Agilent, Santa Clara, California). Separation was
achieved with analytical column BDS Hypersil C18 (150 3 4.6 mm
3 3 mm; Thermo ScientificTM, Waltham, Massachusetts) main-
tained at 18° C, protected by SecurityGuard precolumn (Secur-
ityGuard precolumn; Phenomenex, Torrance, California). The
mobile phase consisted of A: 122 mM KH2PO4, 150 mM KCL,
and 28 mM K2HPO4 and B: 15% (vol/vol) acetonitrile in A. The
percentage of B changed from 0 to 100% in 8 minutes, was
maintained at 100% for 1 minute, and then returned to 0% for re-
equilibration. The separation time with re-equilibration was 13.5
minutes and was conducted at a flow rate of 0.9 ml·min21. The
sample injection volume was 20 ml. The quantitative analyses were
performed based on external calibration of the signal at 254 or 280
nm. Data acquisition and processing was managed by the Chem-
station software (Chemstation Software; Agilent, Santa Clara,
California). The within-run/between-run %CVs were 3.1/4.1, 3.3/
4.4, and 2.7/3.2% for Hx, X, and UA, respectively.
Statistical Analyses
The sample size was a priori estimated based on the assumption
that effect size will be at least medium. Using an a-level of 0.05,
a power (1 2 b) of 0.80, it was calculated that at least 8 par-
ticipants would be needed to detect a significant change or dif-
ferences in lactate and plasma purine metabolite concentration
(G*Power; Heinrich-Heine-Universitat Dusseldorf, Dusseldorf,

Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 ATP Metabolism Biomarkers During Exercise (2019) 00:00
Germany). A one-way repeated-measures analysis of variance
(ANOVA) was performed to test the changes in LA, NH3, and
purine metabolite (Hx, X, and UA) concentration over time
(exercise and recovery measurement points) within each group
of participants. A one-way repeated-measures ANOVA was also
used to test differences between groups at the same measurement
times. A post hoc Scheffe test was conducted if a significant
difference was found (p , 0.05). All effect sizes for ANOVA
were calculated using h (2) and defined as small (0.01), medium
(0.06), and large (0.14) while confidence intervals (CI 95%)
were also calculated. Pearson correlation coefficients (r) and
effects sizes were used to describe the relationship between
biomarker concentrations within each group at maximal exer-
cise and recovery compared with total-body SMM and defined
as small (0.1), medium (0.3), or large (0.5). All statistical anal-
yses were performed using STATISTICA 13.0 software (Stat-
Soft, Tulsa, OK, USA). Significance level for all statistical
analysis was set at p # 0.05. All values were presented as mean
6 SD.
Results
Descriptive Characteristics
Sprinters had a significantly greater amount of SMM compared
with the remaining groups and significantly greater weight at ex-
amination than EN. EN differed significantly from the SP and AM
groups in maximal absolute (L·min21) and relative (ml·kg21·min21
and ml·kg smm21·min21) V̇ O2max. The AM group was signifi-
cantly older than the other groups. Other baseline anthropometric
and physical characteristics were similar in all groups and are
presented in Table 1.
Biomarker Concentration Changes
All detailed mean values with SD, effects sizes, and confidence
intervals are presented as supplementary digital content (see
Tables 1 and 2, Supplemental Digital Content 1 and 2, http://
links.lww.com/JSCR/A128 and http://links.lww.com/JSCR/
A129). Absolute LA concentration at rest, at the point of ex-
haustion, and during recovery did not differ between groups.
During consecutive stages of exercise, sprinters showed
a rapid increase in LA from the initial stage of the test, whereas
in endurance runners, the increase was noticeably delayed
(Figure 1A). After converting LA per 1 kg muscle mass, the
groups also significantly differed at exhaustion and during
recovery. Highest LASMM values were obtained by endurance
athletes and lowest by sprinters (Figure 1B).
For absolute NH3 values, significant between-group differ-
ences were observed at rest, during exercise, and recovery but not
at maximum exertion. During exercise, sprinters had a higher
concentration and most rapid increase in NH3 and endurance
athletes the lowest values compared with other groups, whereas
during recovery, controls showed highest and endurance athletes
the lowest NH3 levels (Figure 1C). Converting the values relative
to SMM resulted in lowest values in sprinters during recovery
(Figure 1D).
Across the entire measurement spectrum (rest, exercise, ex-
haustion, and recovery), the groups significantly differed in both
absolute and relative (per 1 kg muscle mass) values of Hx, X,
and UA. Regardless of metabolite and measure unit, highest
values were consistently obtained by controls and lowest values
by sprinters (Figure 2A‒F). Unlike LA and NH3, maximum Hx
Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
and X concentration was shifted by 5–10 minutes (even 15
minutes in controls) into recovery period (after exhaustion). In
case of UA, its concentration in all groups still increased after 30
minutes of recovery, and thus actual maximal values could not
be measured.
Correlations Between Biomarker Concentration and Skeletal
Muscle Mass
There were no significant correlations observed between SMM
and biomarker concentration at maximal intensity and recovery
within groups. Pearson’s correlation coefficients (r) during exer-
cise and recovery for LA were 20.02 to 0.28, 20.13 to 0.15,
0.10–0.22, and 20.13 to 0.17 (FU, EN, SP, and AM, re-
spectively), for NH3 were 20.2 to 0.27, 0.15–0.46, 0.16–0.30,
and 20.05 to 0.44 (FU, EN, SP, and AM, respectively), for Hx
were 20.13 to 0.18, 20.47 to 20.44, 20.45 to 20.03, and 0.25
to 0.46 (FU, EN, SP, and AM, respectively), for X were 20.12 to
0.18, 20.47 to 20.42, 20.45 to 20.03, and 0.21 to 0.37 (FU,
EN, SP, and AM, respectively), and for UA were 20.43 to 20.27,
20.11 to 0.19, 0.16–0.40, and 0.06–0.36 (FU, EN, SP, and AM,
respectively).
Discussion
This is the first study to measure ATP metabolism biomarker (LA,
NH3, and purine) concentration in blood at rest, during in-
cremental exercise, and in the postexercise recovery period with
many sampling points in highly trained athletes. Each group of
our elite athletes represented a distinct physiological and meta-
bolic training profile dictated by sport discipline. Sprinters rep-
resented speed/power athletes engaged in very high-intensity
exercise requiring mostly anaerobically dominant energy contri-
bution (6). Triathletes and distance runners represented endur-
ance athletes performing predominantly aerobic exercise in their
training (4). Futsal players represented the team sport group
characterized by mixed energy system contribution from both
aerobic and anaerobic pathways (2).
Our main finding is that the pattern and magnitude of bio-
marker concentration depends on the specific training profile of
highly trained athletes in distinct sport disciplines. At rest, during
incremental exercise, and up to 30 minutes into the postexercise
recovery period sprinters had lowest purine metabolism bio-
marker concentrations, and endurance athletes had lowest am-
monia concentrations. For LA during exercise, the lowest
concentrations were noted in endurance athletes, except when
reaching maximum intensity, where the differences between
groups were not significant. In this study, sprinters had signifi-
cantly greater absolute SMM than all remaining groups. Finally,
there were no correlations between SMM and biomarker con-
centrations at maximal intensity and during recovery within each
group.
We observed that groups differed in regard to specific bio-
marker concentration. Sprinters had overall lower purine bio-
marker concentration than all groups. Sprint training is mainly
composed of short, high-intensity exercise bouts with complete or
incomplete rest depending on training goal (6,36). This type of
exercise causes a decrease in the muscle adenine nucleotide pool
(including ATP), through degradation of inosine monophosphate
(IMP) into Hx and X, which further effluxes into the bloodstream
and forms UA, later being excreted with urine (13,14). Zieliński
et al. (36–40) have demonstrated that short-term and long-term
 ATP Metabolism Biomarkers During Exercise (2019) 00:00 | www.nsca.com

Figure 1. Venous plasma LA (A), LASMM (B), NH3 (C), and NH3SMM (D), concentrations at rest, during incremental
treadmill exercise test until exhaustion, and after exercise in futsal players (FU), endurance athletes (EN), sprinters
(SP), and control group (AM). Data are mean values 6 SDs. Vertical arrows indicate point of volitional exhaustion
(V̇O2max). Horizontal arrows indicate first significant difference from blood sampling points at rest and maximal
exercise within groups. Significant differences between blood sampling points: ***p , 0.001, **p , 0.01. Significant
differences between groups at same sampling point: #p , 0.001, ‡p , 0.01, §p , 0.05.

training decreases postexercise Hx levels, which coincides with
increased activity of the enzyme hypoxanthine-guanine phos-
phoribosyltransferase (HGPRT) that catalyzes the conversion of
Hx to IMP. This Hx decrease occurs during periods of training
with high-intensity intermittent-type sessions, especially during
the special preparatory and competitive period. Also, this is more
evident in highly trained athletes, since they engage in more
metabolically challenging training sessions producing greater
ATP precursor loss (36,37,39). Thus, high-intensity intermittent
training results in higher erythrocyte HGPRT activity (36–40)
causing more efficient utilization of the salvage pathway in re-
covering derivatives of adenine nucleotide breakdown, limiting
purine efflux from skeletal muscle, and, consequently, decreasing
postexercise plasma Hx concentrations (12,14,28,36–40). Futsal
players also had relatively lower purine concentrations at rest,
during exercise, and after exercise compared with endurance
athletes and controls. Since gameplay during matches requires the
ability to repeat sprints with incomplete rest while continuously
running, team sport training also encompasses higher-intensity
intermittent training interspersed with aerobic training sessions
(25,27). Plasma purine concentration relative to SMM did not
change differences between groups compared with absolute val-
ues. This demonstrates that blood purine metabolite concentra-
tion is not sensitive to differences in SMM but only to specific
metabolic adaptations.
In terms of NH3, endurance athletes overall had lower blood
concentrations than all groups. Endurance training is mostly
composed of high-volume, low- to moderate-intensity-type ex-
ercise bouts with some higher-intensity training sessions in-
corporated into certain phases of the annual training plan
(4,36,38). Endurance training has been shown to reduce ammo-
nia production during submaximal exercise suggesting that en-
durance training does not decrease ammonia production capacity
but instead delays it (33). Endurance training also increases the
mitochondrial content and blood flow capacity of muscles
resulting in less adenosine monophosphate (AMP) deaminase
activity and decreased ammonia production. This most likely
causes greater utilization of the aerobic system during this type of
incremental exercise, thus decreasing overall ATP breakdown
through AMP degradation that is indicative of cellular energetic
stress and cellular energy status (7). Endurance athletes also have
a smaller amount of fast-twitch (FT) muscle fibers, which contain

Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 ATP Metabolism Biomarkers During Exercise (2019) 00:00

Figure 2. Venous plasma Hx (A), HxSMM (B), X (C), XSMM (D), UA (E), and UASMM (F) concentrations at rest, during
incremental treadmill exercise test until exhaustion and after exercise in futsal players (FU), endurance athletes
(EN), sprinters (SP), and control group (AM). Data are mean values 6 SDs. Vertical arrows indicate point of
volitional exhaustion (V̇ O2max). Horizontal arrows indicate first significant difference from blood sampling points at
rest and maximal exercise within groups. Significant differences between blood sampling points: ***p , 0.001, **p
, 0.01, *p , 0.05. Significant differences between groups at same sampling point: #p , 0.001, ‡p , 0.01. UA 5
uric acid.

more AMP deaminase associated with greater NH3 accumulation
in blood (33). Oxidative capacity of muscle has shown to be
sensitive to training, and at the same exercise intensities, aerobi-
cally trained muscle will have a smaller increase in glycolytic rate,
suggesting that AMP deaminase is less active in FT fibers of aer-
obically trained muscles (33). Sprint-trained athletes, on the other
hand, possess higher amounts of fast-twitch (FT) muscle fibers,
containing more AMP deaminase, leading to greater NH3 accu-
mulation in blood (33). Blood NH3 concentration relative to
SMM (NH3SMM) at maximum intensity and recovery was lowest
in sprinters compared with absolute values. This is most likely due
to greater absolute SMM in sprinters.
For LA, during exercise, the lowest concentrations were noted
in endurance athletes, except when reaching maximum intensity,

Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 ATP Metabolism Biomarkers During Exercise (2019) 00:00
where endurance athletes had the highest concentrations. Blood
lactate response to incremental and constant intensity tests dif-
ferentiates 3 intensity zones, which reflect 3 distinct metabolic
responses (3): (a) exercise intensity that does not require an in-
crease in glycolytic ATP resynthesis (intensity below “lactate
threshold”), (b) exercise intensity that causes an increase in gly-
colytic rate leading to blood lactate concentration steady state,
matched by pyruvate utilization and therefore leading to intensity
between lactate threshold and maximal lactate steady state
(MLSS), and (c) exercise intensity leading to a glycolytic rate that
cannot be matched by aerobic pyruvate utilization indicating
need for ATP generation through anaerobic glycolysis (intensity
above MLSS). Training of highly trained endurance athletes is
mostly dedicated to the first intensity domain (70–90% of
training volume) and less to the second and third intensity
domains (1). This most likely leads to LA concentrations being
very low during initial stages of an incremental exercise test. Also,
endurance athletes’ LA thresholds are relatively high, which
allows them to exercise at a higher relative intensity (%V̇ O2max)
for longer duration. In endurance athletes, glycolytic rate is
matched by pyruvate utilization at much higher intensities than in
untrained and recreationally trained subjects (3). This may in-
dicate that lactate shuttling to other tissues that use lactate as fuel
in the body may be higher in these athletes, which is reflected by
decreased LA blood concentration. Endurance athletes also have
more slow-twitch (ST) muscle fibers, which produce less LA
through anaerobic glycolysis. Sprinters have lower LA thresholds,
a greater percentage of fast-twitch (FT) muscle fibers therefore use
anaerobic glycolysis to a higher degree leading to greater LA
production and release. In addition, sprinters have greater levels
of PCr that can be used during exercise. Utilization of PCr in turn
is coupled closely with glycolysis; depending on intensity and
duration of exercise, greater depletion of PCr will lead to in-
creased glycolysis (21). Finally, futsal athletes had the second
largest LA concentration during incremental exercise (sprinters
the most). Since futsal athletes represent mixed energy system
contribution, their LA concentration increase could be caused by
greater anaerobic glycolytic activity than endurance athletes and
controls, which relied more on aerobic contribution. For LAmax,
sprinters had the lowest concentrations relative to SMM
(LASMM). In contrast to absolute LA concentration, LASMM
concentration during exercise was lowest in sprinters compared
with other groups most likely due to greater absolute SMM.
LA, NH3, and Hx are all valuable for monitoring training
status in highly trained athletes, and each demonstrates different
exercise-induced changes in skeletal muscle. Gorostiaga et al. (11)
found that LA concentration in muscle correlated with plasma LA
concentration in blood, and that ATP concentration in muscle
correlated with plasma NH3 concentration in blood. This indi-
cates that by measuring plasma concentration of LA or NH3,
energy status of skeletal muscle can be indirectly determined.
Purine concentration, on the other hand, mirrors adenine nucle-
otide derivative loss in muscle. Stathis et al. (29) and Sahlin et al.
(22) demonstrated that subjects with a high blood and muscle
purine concentration exhibited a decreased muscle total adenine
nucleotide (TAN) pool or muscle ATP level. Overall, plasma LA
concentration reflects muscle LA concentration that further
reflects anaerobic glycolysis utilization, plasma NH3 concentra-
tion reflects muscle ATP content, and plasma purine concentra-
tion reflects TAN pool depletion. However, calculating
biomarker concentration relative to muscle mass (per 1 kg of
SMM) does not present a clear and accurate picture and does not
have diagnostic value. Athletes having larger amounts of muscle
Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
| www.nsca.com
mass will seem to have lower biomarker levels at the same exercise
intensity, which could falsely imply better training level.
Determining LA and NH3 plasma concentration is more fa-
vorable during exercise because both increase proportionately to
exercise intensity and achieve peak concentration at maximum
intensity. Determining purine metabolite (Hx, X, and UA) con-
centration is more useful after exercise because purine efflux is
indicative of adenine nucleotide loss and consequently reflects en-
ergy loss. Determining LA threshold is often used in endurance
athletes to prescribe appropriate training loads below, at, or above
this threshold (23). However, because endurance athletes have LA
thresholds at relatively high intensities (percentage of their
V̇ O2max), and because most engage in low- to moderate-intensity
exercise below this threshold, it leaves a large unquantifiable area
to prescribe training intensity. Therefore, when quantifying train-
ing load at intensities below LA threshold, NH3 is more intensity-
sensitive. LA is a more adequate marker to assess exercise intensity
in sprint-trained and game sport athletes, and mainly during
higher-intensity interval–type sessions (HIIT), since it is an in-
dicator of anaerobic glycolysis utilization during exercise (7).
NH3 may be used as an indirect indicator of muscle ATP loss
during exercise (11) because it increases proportionately to ex-
ercise intensity and is more intensity-sensitive at lower intensities
below lactate threshold. This verifies the study by Banister et al.
(1), where NH3 blood concentration was elevated even at work
rates at 40–50% V̇ O2max, whereas LA blood concentration in-
creased much more slowly during exercise. Itoh (15) observed
that peak ammonia concentration did not differ between 15-, 30-,
and 45-second sprint bouts, but LA concentration did. This
demonstrates that ammonia is more intensity-sensitive because
each sprint was performed at maximal volitional exertion leading
to similar ammonia concentrations (depletion of TAN pool after
15 seconds in all cases). LA concentrations increased because of
increases in utilization of anaerobic glycolytic pathways in longer
sprint bouts. Determining plasma NH3 concentration can be used
to evaluate, monitor, and prescribe exercise intensity during
clinical exercise tests, conditioning programs, occupational tasks,
and athletic performance.
Purine metabolites, especially Hx can be used during exercise
because Hx increases proportionately to exercise intensity, but it
is worthwhile to measure Hx concentration in the postexercise
recovery period to determine total purine derivative loss. This can
help assess recovery needs of athletes after high-intensity exercise
sessions. Hx level also indicates adaptation to high-intensity ex-
ercise because of reduced purine efflux from skeletal muscle and
can be viewed as a marker of anaerobic metabolism (14,36). Hx
can be used in highly trained athletes as an indicator of training
status in different training phases of a 1-year training cycle re-
gardless of age and sport discipline (37). Using Hx is particularly
valuable in highly trained athletes because commonly used
measures such as V̇ O2max, “anaerobic threshold” or LA con-
centration are not sensitive to changes in training status and
performance of highly trained athletes. Training prescription
should be revised based on recent knowledge regarding purine
metabolism adaptations to fit training programs of highly trained
athletes. Hx can also be used during tapering periods to determine
preparedness and readiness to competition (37). Notably, UA
plasma concentration did not reach peak values during exercise
and even after 30 minutes after exercise. This questions its use-
fulness as a biomarker of muscle status since it is difficult to
achieve a peak value in the specified time frame.
Finally, there were also no correlations observed between
SMM and biomarker concentrations at maximal intensity and
 ATP Metabolism Biomarkers During Exercise (2019) 00:00
recovery in all groups. This indicates that muscle mass levels are
not significantly related to magnitude of biomarker release in
blood. Because we did not perform muscle biopsies, we were not
able to determine biomarker concentrations in muscle and their
correlations with blood biomarker levels. However, this may in-
dicate that training level and especially training profile as a result
of metabolic adaptations in muscle dictates pattern of biomarker
release in blood, not muscle mass itself. Although more muscle
mass may indeed produce greater absolute amounts of these
biomarkers in muscle, the blood concentration is the combined
result of many factors (production and utilization in various
metabolic processes).
A limitation in this study was that subjects could not perform an
additional familiarization session with V̇ O2max testing a few days
before the actual exercise (with blood sampling). Because most
athletes are part of the Polish National Team, it is difficult to ask
them to perform additional tests since they have a very rigorous
training cycle, and additional loading in the form of this exercise
test would not be accepted by their respective coaches. Nonetheless,
all subjects performed these tests regularly throughout their annual
training plan in our laboratory to monitor training, so they were
accustomed to such a test. Another limitation in this study is the
differences in the average time it took participants to complete the
test. The EN group, being the most endurance-trained group, took
the longest to perform the exercise test. However, the aim of using
such a test was to reach V̇ O2max that would standardize intensity
for all groups although duration was greater in the EN group.
Practical Applications
The magnitude and exact time course of exercise-induced
biomarker concentration during a standard progressive ex-
ercise until exhaustion is related to training adaptations
through specific training profile. Combined measuring of LA,
NH3, and Hx concentration in blood is useful in indirectly
reflecting key changes in exercise- and training-induced en-
ergy status without the need to use invasive methods such as
muscle biopsies. It is essential to note that the 3 metabolites
play different roles. (a) Determining LA concentration in
blood is useful for athletes when performing training sessions
using anaerobic glycolysis to a high degree and for athletes in
disciplines where LA threshold determination is needed to
divide training-intensity zones (3). (b) NH3 concentration is
helpful for monitoring exercise at lower exercise intensities
and can be applied for monitoring ATP loss during exercise,
for training prescription and evaluation (33), and can be used
by endurance athletes to more precisely monitor exercise at
low intensities. (c) Purine, especially Hx concentration in
blood, is a valuable tool to determine postexercise purine
derivative loss. Hx is particularly important in highly trained
athletes (36–40). Hx can indicate adaptations to anaerobic
exercise to check training status in different phases of a 1-year
training cycle, especially during tapering periods, regardless of
sport discipline. The differences in plasma purine concentra-
tion of metabolites and NH3 between specifically adapted
athletic groups are independent from SMM. Skeletal muscle
mass level does not significantly affect magnitude of bio-
marker release, and thus sport practitioners do not have to
consider muscle mass levels when measuring biomarker con-
centration in blood. Further research should focus on studying
how specific training sessions affect individual biomarker re-
sponse in highly trained athletes.
Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
Acknowledgments
The authors do not have professional relations with any
companies or manufacturers who will benefit from the results
of this study. The results of this study do not constitute
endorsement of the product by the authors or the NSCA. The
authors thank the coaches, athletes, and volunteers for their full
participation in the study. The authors declare that they have no
conflicts of interest. This work was funded by the Polish Ministry
of Science and Higher Education from financial resources of the
National Science Center within the OPUS 5 program (application
and grant number: 2013/09/B/NZ7/02556).
References
1. Banister EW, Allen ME, Mekjavic IB, Singh AK, Legge B, and Mutch
BJC. The time course of ammonia and lactate accumulation in blood
during bicycle exercise. Eur J Appl Physiol Occup Physiol 51: 195–202,
1983.
2. Beato M, Coratella G, Schena F, and Hulton AT. Evaluation of the ex-
ternal and internal workload in female futsal players. Biol Sport 34:
227–231, 2017.
3. Beneke R, Leithäuser RM, and Ochentel O. Blood lactate diagnostics in
exercise testing and training. Int J Sports Physiol Perform 6: 8–24, 2011.
4. Billat VL, Demarle A, Slawinski J, Paiva M, and Koralsztein JP. Physical
and training characteristics of top-class marathon runners. Med Sci Sports
Exerc 33: 2089–2097, 2001.
5. Brooks GA. Lactate: Link between glycolytic and oxidative metabolism.
Sports Med 37: 341–343, 2007.
6. Duffield R, Dawson B, and Goodman C. Energy system contribution to
100-m and 200-m track running events. J Sci Med Sport 7: 302–313,
2004.
7. Finsterer J. Biomarkers of peripheral muscle fatigue during exercise. BMC
Musculoskelet Disord 13: 218, 2012.
8. Gladden LB. Lactate metabolism: A new paradigm for the third millen-
nium. J Physiol 558: 5–30, 2004.
9. Goodwin ML, Harris JE, Hernández A, and Gladden LB. Blood lactate
measurements and analysis during exercise: A guide for clinicians.
J Diabetes Sci Technol 1: 558–569, 2007.
10. Gorostiaga EM, Navarro-Amezqueta I, Calbet JAL, Hellsten Y, Cusso R,
Guerrero M, et al. Energy metabolism during repeated sets of leg press
exercise leading to failure or not. Plos One 7: e40621, 2012.
11. Gorostiaga EM, Navarro-Amezqueta I, Calbet JAL, Sanchez-Medina L,
Cusso R, Guerrero M, et al. Blood ammonia and lactate as markers of
muscle metabolites during leg press exercise. J Strength Conditioning Res
28: 2775–2785, 2014.
12. Hellsten-Westing Y, Norman B, Balsom PD, and Sjodin B. Decreased
resting levels of adenine-nucleotides in human skeletal-muscle after high-
intensity training. J Appl Physiol 74: 2523–2528, 1993.
13. Hellsten Y, Richter EA, Kiens B, and Bangsbo J. AMP deamination and
purine exchange in human skeletal muscle during and after intense exer-
cise. J Physiology-London 520: 909–920, 1999.
14. Ipata P and Pesi R. Metabolic interaction between purine nucleotide cycle
and oxypurine cycle during skeletal muscle contraction of different in-
tensities: A biochemical reappraisal. Metabolomics 14: 42, 2018.
15. Itoh H and Ohkuwa T. Ammonia and lactate in the blood after short-term
sprint exercise. Eur J Appl Physiol Occup Physiol 62: 22–25, 1991.
16. Karatzaferi C, de Haan A, van Mechelen W, and Sargeant AJ. Metabolic
changes in single human muscle fibres during brief maximal exercise. Exp
Physiol 86: 411–415, 2001.
17. Kim J, Wang ZM, Heymsfield SB, Baumgartner RN, and Gallagher D.
Total-body skeletal muscle mass: Estimation by a new dual-energy X-ray
absorptiometry method. Am J Clin Nutr 76: 378–383, 2002.
18. Mutch BJ and Banister EW. Ammonia metabolism in exercise and fatigue:
A review. Med Sci Sports Exerc 15: 41–50, 1983.
19. Ogino K, Kinugawa T, Osaki S, Kato M, Endoh A, Furuse Y, et al. Am-
monia response to constant exercise: Differences to the lactate response.
Clin Exp Pharmacol Physiol 2000;27: 612–617.
20. Pedersen BK. IL-6 signalling in exercise and disease. Biochem Soc Trans
35: 1295–1297, 2007.
21. Sahlin K. Muscle energetics during explosive activities and potential effects
of nutrition and training. Sports Med 44(Suppl 2): S167–S173, 2014.
 ATP Metabolism Biomarkers During Exercise (2019) 00:00
22. Sahlin K, Tonkonogi M, and Söderlund K. Plasma hypoxanthine and
ammonia in humans during prolonged exercise. Eur J Appl Physiol Occup
Physiol 80: 417–422, 1999.
23. Seiler KS and Kjerland GO. Quantifying training intensity distribution in
elite endurance athletes: Is there evidence for an “optimal” distribution?
Scand J Med Sci Sports 2006;16: 49–56.
24. Sjodin B and Westing YH. Changes in plasma-concentration of hypo-
xanthine and uric-acid in man with short-distance running at various
intensities. Int J Sports Med 11: 493–495, 1990.
25. Slattery KM, Wallace LK, Bentley DJ, and Coutts AJ. Effect of training
load on simulated team sport match performance. Appl Physiol Nutr
Metab 37: 315–322, 2012.
26. Smolenski RT and Yacoub MH. Liquid-chromatographic evaluation of
purine production in the donor human heart during transplantation.
Biomed Chromatogr 7: 189–195, 1993.
27. Spencer M, Bishop D, and Lawrence S. Longitudinal assessment of the
effects of field-hockey training on repeated sprint ability. J Sci Med Sport
7: 323–334, 2004.
28. Stathis CG, Carey MF, Hayes A, Garnham AP, and Snow RJ. Sprint
training reduces urinary purine loss following intense exercise in humans.
Appl Physiol Nutr Metab 31: 702–708, 2006.
29. Stathis CG, Febbraio MA, Carey MF, and Snow RJ. Influence of sprint
training on human skeletal muscle purine nucleotide metabolism. J Appl
Physiol 76: 1802–1809, 1994.
30. Stathis CG, Zhao S, Carey MF, and Snow RJ. Purine loss after repeated
sprint bouts in humans. J Appl Physiol 87: 2037–2042, 1999.
31. Wilkinson DJ, Smeeton NJ, and Watt PW. Ammonia metabolism, the brain
and fatigue; revisiting the link. Prog Neurobiol 91: 200–219, 2010.
Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
| www.nsca.com
32. Yuan Y and Chan KM. A longitudinal study on the ammonia threshold in
junior cyclists. Br J Sports Med 38: 115–119, 2004.
33. Yuan Y and Chan KM. A review of the literature on the application of
blood ammonia measurement in sports science. Res Q Exerc Sport 71:
145–151, 2000.
34. Yuan Y, So R, Wong S, and Chan KM. Ammonia threshold—
Comparison to lactate threshold, correlation to other physiological
parameters and response to training. Scand J Med Sci Sports 2002;12:
358–364.
35. Zielinski J, Krasinska B, and Kusy K. Hypoxanthine as a predictor of
performance in highly trained athletes. Int J Sports Med 34: 1079–1086,
2013.
36. Zielinski J and Kusy K. Training-induced adaptation in purine metabo-
lism in high-level sprinters vs. triathletes. J Appl Physiol 112: 542–551,
2012.
37. Zielinski J and Kusy K. Hypoxanthine: A universal metabolic indicator of
training status in competitive sports. Exerc Sport Sci Rev 43: 214–221,
2015.
38. Zielinski J, Kusy K, and Rychlewski T. Effect of training load structure on
purine metabolism in middle-distance runners. Med Sci Sports Exerc 43:
1798–1807, 2011.
39. Zielinski J, Kusy K, and Slominska E. Alterations in purine metabolism in
middle-aged elite, amateur, and recreational runners across a 1-year
training cycle. Eur J Appl Physiol 113: 763–773, 2013.
40. Zielinski J, Rychlewski T, Kusy K, Domaszewska K, and Laurentowska
M. The effect of endurance training on changes in purine metabolism: A
longitudinal study of competitive long-distance runners. Eur J Appl
Physiol 106: 867–876, 2009.