Table 2.

Short-term intervention studies included in this review.

Author (year)	Primary aim	Metabolomic profiling platform	Biosample	Study population	Exercise/physical activity intervention or test	Sample collection timepoint	Primary findings
Sabatine et al. (2005)	To compare metabolomic pathways activated with exercise in myocardial ischemia cases versus controls.	LC-MS	Plasma	18 ischemia cases (mean (SD) age 64 (10) years) and 18 controls (mean (SD) age 65 (11) years); men and women.	Exercise testing with standard Bruce protocol.	Two samples before and after exercise testing.	Demonstrated significant changes after exercise stress testing in circulating levels of multiple metabolites; identified clusters of metabolites that were altered in all individuals, and metabolites that showed discordant effects after exercise in ischemia cases versus controls.
Enea et al. (2009)	To investigate biochemical changes due to short-term and prolonged PA.	NMR	Urine	22 trained and untrained women; mean (SD) age between 20.7 (0.7) years to 23.0 (1.9) years.	Prolonged exercise test until exhaustion and short-term intensive test.	Two samples before and after each exercise test.	Metabolic changes were observed in urine after a short-term intensive exercise test; the extent of the changes depended on prior training status.
Kirwan et al. (2009)	To demonstrate the potential for 1H NMR analysis in exercise biochemistry.	NMR	Plasma	7 endurance-trained cyclists/ triathletes, men, age not reported.	Intermittent exhaustive cycling.	Eight samples, prior to and immediately following exercise and at 30, 60, 90, 120, 140 and 280 min during the 4-h passive recovery period.	Differences in the response to exhaustive exercise and subsequent recovery varied according to baseline metabolic substrate levels.
Lehmann et al. (2010)	To capture the network of metabolites regulating the beneficial effects of exercise.	LC/TOF-MS	Plasma	21 men in two groups, mean (SD) age 32.6 (6.1) years in first group, 30.9 (5.8) years in second group.	60 or 120-min treadmill run.	Four samples; before and after running and 3 and 24 h into recovery.	An increase in acylcarnitines during exercise supports fat oxidation and may exert beneficial biological functions.
Lewis, et al. (2010)	To explore exercise-induced metabolic responses.	LC-MS	Plasma	Exercise testing protocol (ETT) - 78 men and women, mean (SD) age ranged between 48 (14) and 59 (12) years; Boston Marathon (BM) Cohort- 25 healthy amateur runners men and women mean (SD) age 42 (9) years.	Diagnostic treadmill ETT or bicycle ergometry cardiopulmonary exercise testing.	ETT; three samples at baseline, peak exercise and 60 min after completion; BM - two samples before and after marathon.	Acute and long-term exercise elicited a metabolic response. Metabolic responses differed by BMI by and baseline fitness level, and in those with ischemia.
Pechlivanis et al. (2010)	To investigate changes in urine metabolome elicited by two differing exercise sessions.	NMR	Urine	12 men in two groups; mean (SD) age in exercise group 1; 21 (2) years, mean (SD) age in exercise group 2; 20 (1) years.	Three sets of two 80 m maximal runs, separated by 10 s or 1 min of rest.	Two samples: before and 35 min after exercise session.	Urine metabolomic profiles could distinguish between pre- and post-exercise and between the two different exercise groups; perturbations were more pronounced when the time between exertions was shorter.
Krug et al. (2012)	To explore the response of the metabolome to a PA challenge.	MS/MS and NMR	Plasma, Urine, Exhaled air, Breath condensate.	15 men; mean (SD) age 27.8 (2.9) years.	30-min bicycle ergometer.	Twelve samples: Plasma and exhaled air was collected before the physical activity test (PAT), during the PAT at 15 and 30 min, and after the PAT at 45, 60, 90, and 120 min. EBC was collected at 0, 60- and 120-min. Urine was collected at 0 and 120 min.	Changes in the metabolomic profile were observed immediately follow the initiation of the challenge but returned to almost normal within 2 h of completion. Inter-subject variation increased even more during the challenges.
Nieman et al. (2013)	To investigate changes in serum metabolome elicited by a 3-day period of intensified training.	GC-MS, LC-MS	Serum	15 men (mean (SD) age 35.5 (9.2) years) and women (mean (SD) age 35.1 (8.9) years) runners.	3-day intense running (2.5 h on treadmill/day).	Three samples; pre-exercise, immediately post-exercise and 14 h post-exercise.	Runners experienced a profound systemic shift in blood metabolites related to energy production especially from the lipid super pathway following 3 days of heavy exertion that was not fully restored to pre-exercise levels after 14 h recovery.
Mukherjee et al. (2014)	To understand molecular mechanisms mediating the beneficial adaptations of exercise in older adults.	NMR	Urine	Competitive cyclists (n = 9) and untrained, minimally active controls (n = 8); men aged 50–60 years old.	Submaximal endurance cycle.	Pre-exercise, immediately post exercise and 24 h post exercise.	Post exercise, highly trained competitive athletes have a characteristic metabolic footprint that differs from non-trained, minimally active older individuals.
Nieman et al. (2014)	To identify metabolomic correlates of oxidative stress during endurance exercise.	LC-MS and GC-MS	Plasma	19 cyclists, men aged 27–49 years old.	75-km cycling on an ergometer.	Four samples; before and immediately, 1.5 and 21 h after exercise.	Metabolomics confirms the role of oxidative stress and fatty acid metabolism with endurance exercise.
Peake et al. (2014)	To assess effects of high-intensity interval training (HUT) vs. work-matched moderate-intensity continuous exercise (MOD) on metabolism.	GC-MS	Plasma.	10 well-trained cyclists and triathletes; men mean (SD) age 33.2 (6.7) years.	Electromagnetically braked cycle ergometer until exhaustion.	Four samples: before, immediately-post and 1- and 2- hours post exercise.	Although many metabolic changes were the same, there were also some distinct differences in specific metabolites following HUT vs. MOD.
Ra et al. (2014)	To identify salivary fatigue markers.	CE-TOFMS	Saliva.	37 fatigued soccer players, mean (SD) age 20.6 (0.04) years.	3 consecutive days of soccer.	Two samples; before and after 3 days of soccer.	Salivary metabolites increased after three days of soccer may represent novel biomarkers of fatigue.
Zheng et al. (2014)	To investigate the effects of PA on the adolescent metabolome.	NMR	Urine, plasma.	192 overweight adolescents; 12–15 years old.	Pedometer for 7 consecutive days.	Two samples: at baseline and twelve weeks later.	No strong correlation could be identified neither between the plasma nor the urine metabolome and daily PA.
Wang et al. (2015)	To determine the effects of different levels of training exercises on the urine metabolome.	NMR	Urine.	12 professional half-pipe snowboarders, men and women age 25–25 years.	Strength, endurance, and trampoline exercises at three different intensity levels.	Three samples; at three timepoints throughout.	A PCA plot was able to distinguish between the different intensity levels. Results show that organisms reach a relatively stable physical state to adapt to the training load after long-term training.
Danaher et al. (2016)	To compare the metabolomic effects of workload matched high intensity trials.	GC-MS	Plasma.	7 untrained men, mean (SD) age 22.9 (5) years.	30 min cycling at two different intensities.	Four samples; at rest (preexercise), 10 min into exercise, immediately after exercise and after 60 min recovery.	The high intensity protocol elicited greater metabolic changes relating to lipid metabolism and glycolysis than the moderate intensity. The changes were more pronounced throughout the recovery period than during the exercise.
Muhsen et al. (2016)	To explore the metabolomic effects of submaximal exercise.	LC-MS	Urine.	10 men and women, 23–48 years.	Braked cycle ergometer at light and moderate intensities for 45 min.	Three samples: Day before, day of and day after exercise.	PCA could separate between pre- and post-exercise samples.
Zafeiridis et al. (2016)	To compare the metabolic profile of three aerobic exercises matched for effort/strain.	NMR	Plasma.	9 active men; mean (SE) age 20.5 (0.7) years.	Continuous, long-interval (3 min), and short-interval (30 s) bouts of exercise.	Two samples for each protocol; pre and post exercise.	OPLS demonstrated a distinct separation in metabolomic profiles between pre and post exercise samples, but no differences between the three different regimens.
Berton et al. (2017)	To explore the metabolomic response to resistance exercise.	NMR	Serum.	10 men, mean (SD) age 24 (2) years.	Resistance exercise: leg press and knee extension exercises.	Six samples: 1-hour pre and immediately post exercise and 5, 15, 30- and 60-minutes post exercise.	Resistance exercise induced changes in the serum metabolome.
Couto et al. (2017)	To investigate the effect of a swimming training session on oxidative stress markers of asthmatic compared to non-asthmatic elite swimmers using exhaled breath metabolomics.	GC-MS	Exhaled breath.	20 elite swimmers (including 9 with asthma); men and women aged 13–24 years old.	1-hour Swimming training session.	Two samples; before and after exercise.	In well-trained athletes, swimming is associated with a decrease in oxidative stress markers independently of the presence of asthma, although a more pronounced decrease was seen in controls.
Karl et al. (2017)	To explore metabolomic changes associated with strenuous exercise.	LC-MS	Plasma.	25 army Soldiers, men mean (SD) age 19 (1) years.	4-day, 51-km cross-country ski march.	Two samples before and after march.	Observed increases in energy metabolism, lipolysis, fatty acid oxidation, ketogenesis, and branched-chain amino acid catabolism.
Messier et al. (2017)	To determine whether the metabolomic pathways used during endurance exercise differ according to whether the effort is performed at sea level or at moderate altitude.	NMR	Plasma.	20 men, mean (SD) age 39 (4.3) years.	60 min cycle ergometer.	Two samples before exercise and after 60 min (when participants were still peddling).	At similar exercise intensity, substrate use during endurance exercise differed by altitude level.
Nieman et al. (2017)	To determine the relationship between exercise-induced increases in IL-6 and lipid-related metabolites.	LC-MS and GC-MS	Plasma.	24 runners, men aged 22–55 years old.	Treadmill to exhaustion.	Two samples before and after exercise testing.	There was a clear separation in metabolomic profiles pre and post exercise.
Prado et al. (2017)	To understand exercise-induced changes during a soccer match.	MS	Urine and blood.	30 professional soccer players, men aged 18–20 years old.	Soccer/Football match.	Two samples pre-match and postmatch.	Hypoxanthine and related metabolites were upregulated in urine after a soccer match, suggesting adenosine monophosphate deamination was increased.