The research team, led by Taira Kajisa and Toshiyuki Sakai of Toyo University, equipped 20 male athletes classified as Tier 3 Highly Trained/National with FreeStyle Libre 2 continuous glucose monitoring (CGM) sensors. These devices measure glucose in the interstitial fluid just beneath the skin, providing a reading every 15 minutes without needles during exercise. The cohort contributed 36 race-event datasets spanning track running, road relays, half-marathons, cross country, and race walking, generating 16,360 individual glucose data points collected from three days before to three days after each race.

The study design addressed a critical confound: carbohydrate intake. Athletes, all living in a team dormitory with standardized meals, completed solid carbohydrate consumption at least 120 minutes before race starts, and event regulations prohibited carbohydrate-containing fluids during races. This means the glucose elevations observed mid-race could not be explained by recent fueling. Fifteen of the athletes also wore sensors during high-intensity training sessions prescribed at the identical race pace, on matching course types, within eight days of the competition, enabling a rare within-subject comparison.

The statistical picture was unambiguous. Mean glucose during the pre-race hour climbed to 8.3 ± 1.5 millimolar, well above the cohort’s overall mean of 6.1 millimolar, and then surged to a competition peak of 11.2 ± 2.2 millimolar, nearly double typical resting values. After the finish line, glucose dropped sharply, falling to 6.7 millimolar within an hour. A repeated-measures analysis of variance confirmed a powerful effect of race phase, and paired comparisons adjusted with the Holm-Bonferroni procedure showed the jump from pre-race to during-race glucose carried a large effect size, with Cohen’s dz of 1.45.

The within-subject comparison between competition and training proved the study’s centerpiece. Peak interstitial glucose reached 12.91 ± 1.93 millimolar in races versus 11.02 ± 1.18 millimolar in matched training bouts, a mean difference of 1.89 millimolar with a 95 percent confidence interval of 1.06 to 2.72. The paired t-test yielded p = 0.000131, and the geometric mean ratio of 1.166 translates to approximately 17 percent higher competitive peaks. Exploratory analyses showed the difference was not driven by race distance: short events under 10 kilometers showed +2.17 millimolar, mid-distance events +2.15, and longer events +1.40, a gradient that actually runs opposite to what a pure intensity-difference explanation would predict.

What could drive this competition-specific hyperglycemia? The authors frame their findings within classical stress physiology. Competitive arousal activates the sympathetic nervous system, releasing epinephrine and norepinephrine, which stimulate the liver to break down glycogen and dump glucose into the bloodstream. Causal support exists in the literature: in adrenalectomized humans, adrenaline infusion during exercise directly increases hepatic glucose production. Crucially, muscle glycogen stores, though five times larger than the liver’s roughly 100 grams, cannot release glucose into blood, so the racing hyperglycemia must reflect hepatic mobilization, not muscle fuel leaking into circulation.

A preliminary validation experiment underscored the technique’s reliability. A sub-elite marathoner (personal best under three hours) ran 5 kilometers fasted at 4 minutes per kilometer while wearing both a CGM and periodically sampling finger-prick blood glucose. CGM tracked capillary glucose closely, with the expected lag of about 15 minutes and a modest average underestimate of roughly 1.5 millimolar. During an 18-day monitoring window surrounding a full marathon, this athlete’s glucose spiked above 10 millimolar just before the start, peaked at 11.7 millimolar mid-race, and stayed elevated through the night, while easy practice runs kept glucose hovering near 5.4 millimolar.

The researchers are careful, refreshingly so, about what the data cannot prove. Internal physiological load was not measured; heart rate, power, and blood lactate went unrecorded, and pace was prescribed rather than verified during training. The authors acknowledge they cannot definitively attribute the competition-training difference to psychological stress rather than subtle differences in effort. A single bipolar pre-race questionnaire item on arousal showed no significant link to glucose, but the team itself argues the measure was inadequate, calling instead for validated instruments like the Competitive State Anxiety Inventory-2 and direct measurements of plasma or salivary catecholamines and cortisol.

The practical implications could reshape how endurance athletes prepare. Athletes whose glucose already soars before the gun may not need aggressive pre-race carbohydrate supplementation, suggesting individualized monitoring could personalize fueling strategies. Coaches might deliberately build competition-like stress into training sessions to acclimate athletes’ metabolism to the competitive milieu, since exogenous carbohydrate intake cannot reproduce the endocrine signature of real racing. And the sharp post-race glucose drop within an hour highlights a narrow recovery window when prompt glycogen replenishment matters most.

No athlete in the study experienced intra-race hypoglycemia despite the extraordinary glucose turnover, a reassuring finding given the high metabolic demands documented. The cohort’s limitations, all male, all collegiate, all from a single university squad, mean generalization to female athletes, masters competitors, or world-class Tier 4-5 performers requires caution. Yet the study stands as a proof of concept that wearable CGM technology can capture the metabolic fingerprint of competition itself, in the chaotic real world where athletic destinies are decided rather than in sanitized laboratory conditions. Future work pairing glucose data with heart rate variability, psychometric stress scales, and hormone sampling may finally untangle how much of the racing surge comes from the mind rather than the muscles.

Subject of Research: Interstitial glucose responses to sanctioned competition versus high-intensity training in highly trained male collegiate endurance athletes measured by continuous glucose monitoring

Article Title: Interstitial glucose during sanctioned competition and during high‐intensity training in highly trained collegiate male endurance athletes: A real‐world continuous glucose monitoring study

Article References: Kajisa, T., & Sakai, T. (2026). Interstitial glucose during sanctioned competition and during high‐intensity training in highly trained collegiate male endurance athletes: A real‐world continuous glucose monitoring study. Physiological Reports, 14(17), Article e71099. https://doi.org/10.14814/phy2.71099

Image Credits: AI Generated

DOI: 10.14814/phy2.71099

Keywords: continuous glucose monitoring, endurance athletes, competition stress, hepatic glucose output, catecholamines, exercise metabolism, running, race walking, glycogen, sports physiology, interstitial glucose, athletic performance

The research emerged from a larger 12-week randomized controlled trial designed to test whether high-intensity combined training could match traditional training in military-style fitness outcomes. Participants were young adults aged 18 to 27 from the Birmingham, Alabama area, all of whom were sedentary before enrollment. One group, labeled TRAD, performed 30 minutes of cycling at 70 percent of heart rate reserve followed by a full-body resistance routine of squats, presses, rows, and curls at three sets of 13 repetitions. The other group, called HITT, completed ten rounds of a maximal 30-second on/off circuit featuring box jumps, burpees, kettlebell swings, cycling and rowing sprints, battle ropes, and wall balls, then finished the same resistance exercises in superset form at lower volume and shorter rests. Both groups trained in the early morning after an overnight fast, and researchers drew blood before exercise, immediately after, three hours later, and again at 24 hours.

The analytical scale of the study is striking. Using untargeted metabolomics with both reverse-phase and HILIC chromatography in positive and negative ion modes, the team detected 10,793 serum compounds, annotated 5,215 of them, and consolidated these into 3,243 non-redundant features. After statistical filtering with a false discovery rate threshold of 0.10, 2,052 compounds changed significantly in at least one comparison, including 738 annotated endogenous metabolites spanning amino acids, lipids, acyl carnitines, nucleotides, and biogenic amines. Partial least squares-discriminant analysis showed that the serum metabolome clustered distinctly at the immediate and three-hour timepoints in both groups, before settling back toward baseline by 24 hours — a molecular signature of exertion that fades over roughly a day.

Within each group, the response was enormous. In the traditional group, 683 metabolites shifted below the significance threshold at one or more timepoints, compared with 585 in the high-intensity group. Immediately after exercise, both groups showed roughly equal numbers of rising and falling metabolites, but by three hours the picture skewed heavily toward accumulation: 333 upregulated metabolites in TRAD and 284 in HITT, dominated by fatty acids, conjugated fatty amines, and ketones. By 24 hours, the traditional group still had 183 altered metabolites while the high-intensity group had 86, indicating that the longer session left a deeper and more persistent biochemical footprint. The researchers attribute this to the sustained metabolic demand of continuous endurance work and the roughly doubled duration of the traditional prescription.

The timing of specific metabolite classes tells a coherent physiological story. Immediately after exercise, the blood was enriched for the active substrates of ATP generation — glucose, pyruvate, citrate, aconitate, and fumarate — along with products of purine catabolism such as hypoxanthine and uridine monophosphate. Working skeletal muscle, which cannot express xanthine oxidase, releases these purine intermediates into circulation, where the liver salvages them back into the ATP pool or converts them to uric acid. This salvage pathway is more energy-efficient than building nucleotides from scratch, and its appearance in the blood mirrors findings from earlier studies of moderate steady-state cycling and resistance exercise alone.

Three hours later, the emphasis shifted to recovery and fuel replenishment. Both groups showed elevated levels of 12,13-diHOME, an adipose-derived exerkine known to stimulate fatty acid uptake in skeletal muscle, alongside palmitoylcarnitine, which ferries fatty acid chains across the mitochondrial inner membrane for beta-oxidation, and a broad suite of long-chain fatty acids. Ninety-seven metabolites rose in both groups at this timepoint, most of them lipid-related, representing the largest shared signature in the study. The authors interpret this coordinated lipid mobilization as systemic replenishment of energy stores drained by the bout — the metabolic equivalent of refueling after a long drive.

Only a handful of metabolites distinguished the two training modes directly. Four met the false discovery threshold: capryloylglycine was lower in the traditional group immediately after exercise; hydroxynorleucine and acetylcholine were higher in that group at three hours; and D-mannose was lower in the traditional group at three hours. The acetylcholine finding is particularly intriguing because high-intensity exercise has previously been linked to reduced circulating choline and acetylcholine, yet choline itself fell in both groups here, suggesting that another mechanism — perhaps reduced acetyl-CoA availability or diminished choline acetyltransferase activity — drove the drop in the high-intensity group. The mannose difference may reflect the liver responding to glycolytic stress, since epinephrine-driven glycogenolysis releases mannose that cells can phosphorylate and channel into glycolysis or glycogen synthesis.

Some of the distinguishing metabolites remain biochemical mysteries. Capryloylglycine, a conjugate of the medium-chain fatty acid caprylic acid and glycine, has been described as a principal pancreatic metabolite in pigs and is elevated in the muscle of older adults, and genetic defects in mitochondrial beta-oxidation raise its levels — but why a traditional combined session would lower it acutely is unknown, since glycine availability was similar between groups. Hydroxynorleucine does not even appear as an identified entry in the Human Metabolome Database. The authors are candid that these findings need replication and mechanistic follow-up before any physiological meaning can be assigned.

Unbiased clustering across time revealed that most response patterns were shared between the two prescriptions, with key metabolites such as glucose, lactate, cortisol, palmitoylcarnitine, and glutathione rising or falling in parallel. Yet each mode also produced unique signatures: the traditional group showed a distinct cluster containing succinate that fell equally at both early timepoints, while the high-intensity group had a cluster with ornithine, glycine, and citrulline that dropped immediately and then climbed steadily through 24 hours. These idiosyncratic patterns, layered on top of the shared core response, are the clearest evidence yet that exercise dose and structure write distinguishable chemical signatures into the blood.

The study has caveats worth noting. Participants consumed a standardized protein drink after the immediate post-exercise blood draw to mimic real-world training, which could have contributed to some of the three-hour changes, though the lipid and amino acid patterns closely match those of fasted studies. All participants were untrained, so unaccustomed exertion itself likely drove much of the response, and the design covaried for sex rather than testing sex-specific effects, which the team has reported separately. Still, as the first timecourse study of the acute circulating metabolomic response to combined endurance and resistance exercise, the work establishes that a 45-minute high-intensity session and a 90-minute traditional session converge on the same fundamental biology — fuel burning, purine salvage, lipid mobilization — while leaving behind subtly different molecular fingerprints. For time-pressed exercisers and military planners alike, that suggests the shorter workout may deliver much of the same metabolic conversation, just in a more compressed dialect.

Subject of Research: Acute serum metabolomic responses to traditional versus high-intensity combined endurance and resistance exercise in young adults

Article Title: Serum metabolomics signatures after an acute bout of combined traditional or high‐intensity tactical training in young adults

Article References: Graham, Z. A., Pathak, K. V., Garcia‐Mansfield, K., Lavin, K. M., Torres, A. R., McAdam, J. S., Broderick, T., Pirrotte, P., & Bamman, M. M. (2026). Serum metabolomics signatures after an acute bout of combined traditional or high‐intensity tactical training in young adults. Physiological Reports, 14(18), Article e71094. https://doi.org/10.14814/phy2.71094

Image Credits: AI Generated

DOI: 10.14814/phy2.71094

Keywords: metabolomics, exercise physiology, high-intensity interval training, resistance training, serum metabolites, combined exercise, fatty acid oxidation, purine salvage, recovery, military fitness, young adults, Physiological Reports