Table 3.

Long-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
Yan et al. (2009)	To investigate alterations in metabolic phenotype of professional athletes induced by long-term training.	GC/TOF-MS	Venous blood	28 men: 16 professional rowers (mean age (SD) 23 (2.7) years) and twelve controls (mean age (SD) 23.8.(1.5) years)	2 weeks of technical and aerobic exercise.	Three samples: before training and after one- and two-weeks training.	Significant metabolomic differences observed between professional athletes and controls. Long-term strength and endurance training induced distinct separation between athletes by seniority, training stage and training
Huffman et al. (2011)	To profile the metabolome at baseline, after 6 months of exercise training, and 2 weeks after exercise training cessation.	Targeted MS	Plasma	53 middle-aged, overweight and moderately obese, inactive participants, men and women (age not reported)	6 months supervised aerobic exercise training.	Three samples; baseline, after 6 months of exercise training and two weeks after exercise cessation.	Improvements in insulin sensitivity with the intervention was associated with reductions in circulating free fatty acids and increased levels of glycine and proline. These changes persisted two weeks post intervention cessation.
Neal et al. (2012)	To investigate physiological adaptation with two endurance-training periods differing in intensity distribution.	NMR	Urine	12 cyclists, men, mean age (SD) 37 (6) years	Two 6-week training periods following two models in a crossover design. 1. polarized training model (—75–80% low intensity, 5–10% moderate intensity, 15–20% at high intensity); and 2. threshold training model (40–50% moderate intensity, little high intensity and the remainder at low intensity).	Two samples; before and after exercise.	Some markers of cellular energy stress were modified with low intensity but not with high intensity training.
Pechlivanis et al. (2013)	To investigate changes in the human serum metabolome elicited by two differing exercise sessions.	NMR	Serum	14 men in two groups; group A mean (SD) age 21 (2) years; group B mean (SD) age 20 (2) years.	8-week training program involving sets of two 80 m maximal runs, separated by (group A) 10 s or (group B) 1 min of rest.	Two samples: during a training session in week 1 and during a training session in week 8.	Serum metabolomic profiles could distinguish between pre- and post-exercise and between the two different exercise groups; there was greater dispersion among the post exercise samples, indicating greater variation in the response to exercise.
Huffman et al. (2014)	To evaluate the effect of exercise training on muscle metabolic signatures.	MS/MS GC-MS	Muscle	112 men and women at risk of metabolic disease.	Six months following one of the six interventions: (i) low-amount moderate-intensity exercise; (ii) low-amount vigorous-intensity exercise; (iii) high-amount vigorous-intensity exercise; (iv) resistance training; (v) linear combination of low-amount vigorous-intensity exercise and resistance training; (vi) continued inactivity, age 18–70 years.	Two samples; before and 16–25 h after the last exercise bout.	Dose response relationship between exercise regime energy expenditure and increase in muscle concentrations of acyl carnitines. The smallest changes were seen in the resistance training group.
Felder et al. (2017)	To identify endogenous metabolites that distinguish the trained from the untrained state.	LC-MS	Serum	37 men, 45–65 years.	Electronically braked cycle ergometer before and after completion of a ten-week exercise program.	Two samples: before and after exercise program.	Detected metabolites that differed between the trained and untrained state.
Meucci et al. (2017)	To explore metabolomic changes associated with PA.	GC/TOF-MS	Urine	22 recreationally active overweight preadolescents, boys and girls 8–12 years old.	4 week or 8 week supervised, play-based PA for 6 h per day, 5 h per week, and one group with unsupervised play.	Two samples pre and post intervention.	After an eight-week PA intervention there were significant metabolomic differences between a control group who did not undergo the intervention, and between the pre- and post-exercise samples of the children who underwent the intervention. No change was evident after four weeks.
Brennan et al. (2018a)	To explore metabolomic changes associated with varying exercise interventions.	LC-MS	Plasma	216 middle-aged abdominally obese men and women; age 52.5 (8.0) years.	Four intervention groups varying in exercise amount and intensity for 6 months.	Three samples; baseline and at 48 h after the last session at 16 and 24 weeks.	There were no differences between specific intervention groups, or in the exercise group compared to the control group.
Brennan et al. (2018b)	To examine changes in the metabolome following chronic aerobic exercise.	LC-MS/MS	Plasma	Three hundred abdominally obese men and women, mean (SD) age ranged between groups from 51.8 (8.3) to 55.1 (6.6) years.	Four intervention groups varying in exercise amount and intensity for 6 months.	Three samples; baseline and at 48 h after the last session at 24 weeks.	Few metabolites changed with exercise compared to controls, therefore associations between adipose tissue deposits and metabolites where not sue to exercise induced adipose tissues reduction.
Grapov et al. (2020)	To investigate how the effects of insulin resistance on the metabolome are altered by exercise and fitness.	GC-MS	Plasma	12 insulin resistant obese women aged 30–50 years.	Cycle Ergometer before and after a 14-week training and weight loss intervention.	11 samples at 5 min increments during 30 min cycling and postcycling cool down. The same 11 sample procedure was performed before and after the intervention.	The long-term intervention had little impact on the global metabolome, however the excise itself strongly impacted the metabolome.
Sakaguchi et al. (2020)	To investigate the chronic effect of inspiratory muscle training (IMT) on the metabolome.	NMR	Serum	28 cyclists randomized to 3 levels of IMT intensity for 11 weeks or a control group, men aged 20–40 years old.	Inspiratory Muscle Training for 11 weeks.	Two samples; one week before and 11 weeks after IMT.	Metabolites shifts did not differ by IMT intensity level or in comparison to controls, indicating physical training has negligible effects on the serum metabolome.

All Individuals were ‘healthy’ at blood draw unless otherwise stated.

GC–MS – Gas Chromatography–Mass Spectrometry.

GC/TOF-MS – Gas Chromatography/Time of Flight-Mass Spectrometry.

IMT – Inspiratory Muscle Training.

LC-MS - Liquid Chromatography-Mass Spectrometry.

MS – Mass Spectrometry.

MS/MS – Tandem Mass Spectrometry.

NMR - Nuclear Magnetic Resonance Spectroscopy.

PA – Physical Activity.

SD – Standard Deviation.