A new pair of field studies, published in BMC Medicine, offers the most detailed real-world picture to date. A research team led by Michael Siebers and Johannes Fuss of the University of Duisburg-Essen, together with colleagues at the University Hospital Ulm, the German Sport University Cologne, the University Medical Center Mainz, and other institutions, tracked circulating endocannabinoid concentrations in trained runners during actual competitions, from a standard 42.2-kilometer marathon all the way up to 230-kilometer ultramarathons. Their results reveal a strikingly time-dependent chemical choreography, with one key messenger climbing steadily for hours while a second behaves in a surprisingly selective, distance-dependent manner.

The endocannabinoid system is built around lipid messengers that are synthesized on demand from membrane fats and act on two G-protein-coupled receptors, CB1 and CB2. CB1 receptors are densely expressed in brain regions governing mood, anxiety, and pain perception, which is precisely why the system has been a leading suspect in the runner’s high. The two most studied messengers are anandamide (AEA), an arachidonate-derived ethanolamide whose name derives from the Sanskrit word for bliss, and 2-arachidonoylglycerol (2-AG), which belongs to a different chemical family but activates the same receptors. Both are rapidly degraded by dedicated enzymes, fatty acid amide hydrolase for AEA and monoacylglycerol lipase for 2-AG, meaning their blood levels reflect a dynamic balance of production and clearance rather than a passive accumulation.

In the first study, 19 trained runners completed both a full marathon and a duration-matched walking session on separate occasions, an elegant design that held exercise duration constant while varying the intensity and the mode of locomotion. During the marathon, the researchers drew blood at every 14-kilometer checkpoint and again after a 45-minute recovery break, allowing them to chart the trajectory of five lipid signals: AEA, 2-AG, its structural isomer 1-AG, arachidonic acid, and palmitoylethanolamide, an anti-inflammatory lipid that does not bind CB1. Mood was captured with visual analog scales rating euphoria, anxiety, and pain, the three core psychological features of the runner’s high. All measurements were quantified with a standardized liquid chromatography and multiple reaction monitoring assay, the gold-standard technique for targeted lipidomics.

The headline finding concerns anandamide. AEA concentrations rose progressively throughout the marathon and, remarkably, remained elevated even after 45 minutes of rest, indicating that the signal is not a fleeting byproduct of exertion but a sustained shift in circulating lipid tone. By contrast, the duration-matched walking session produced only modest changes in AEA, suggesting that sustained moderate-to-vigorous intensity, not merely hours of movement, is what drives the messenger upward. The ultramarathon study reinforced this pattern: 36 runners competing over 100, 160, or 230 kilometers all finished with AEA levels above their own baselines, regardless of race distance. Whatever mechanism pushes anandamide into the bloodstream, it appears to engage early and keep working across an extraordinary range of exercise durations.

2-AG told a different and more nuanced story. In the marathon, 2-AG increased significantly only during the later stages of running and remained elevated into early recovery, a delayed profile that contrasts sharply with the steady climb of AEA. In the ultramarathon cohort, 2-AG was not elevated during the races themselves in the same way, but post-race concentrations were increased at all three distances, consistent with what the authors interpret as a delayed, recovery-related response rather than an acute exercise signal. This dissociation matters scientifically: it suggests that the two major endocannabinoids are regulated by partly independent mechanisms during prolonged endurance exercise, with AEA tracking the ongoing stress of running and 2-AG lagging behind, perhaps reflecting lipid remodeling or delayed enzymatic activity that continues after the finish line.

The psychological measurements added a crucial affective dimension. Marathon running was associated with significantly higher euphoria and lower anxiety than duration-matched walking, even though both activities lasted the same amount of time. Pain, however, told a more complicated tale: it increased after the 28-kilometer mark of the marathon, showing that the analgesic component of the runner’s high does not simply overwhelm accumulating musculoskeletal damage during very long efforts. In the ultramarathoners, post-race profiles showed increased pain, reduced anxiety, and no significant change in euphoria. Taken together, these patterns hint that different facets of the runner’s high may ride on different chemical rails, with the anxiolytic effect potentially linked to the sustained AEA rise and the euphoric peak perhaps requiring conditions, such as the intensity profile of a marathon, that ultramarathon pacing does not reproduce.

The field-based design is both the study’s greatest strength and its defining constraint. By sampling real competitors mid-race, the researchers captured physiological conditions that no laboratory protocol can faithfully simulate, including accumulated fatigue, dehydration, heat stress, and the psychological texture of competition. The trade-offs are equally real: field studies cannot control diet, pacing, sleep, or terrain with laboratory precision, and blood-based lipid measurements are an indirect window onto brain chemistry, since circulating endocannabinoids do not map one-to-one onto signaling within the central nervous system. The authors also note that the findings document robust correlations between eCB dynamics and affective states during prolonged running; establishing causation, for example by pharmacologically manipulating the system during exercise, remains a challenge for future work.

Even so, the implications are considerable. The endocannabinoid system is a validated target of both cannabis pharmacology and mainstream drug development, and exercise is increasingly recognized as a natural modulator of this network with potential relevance for mood disorders, chronic pain, and stress resilience. Demonstrating that hours of running produce sustained, intensity-dependent elevations of anandamide in real athletes provides an evolutionary and physiological scaffold for the idea that long-distance locomotion is intrinsically rewarding, a hypothesis sometimes framed as the endurance-exercise model of human origins, in which our ancestors’ capacity for persistence hunting and long-distance travel was reinforced by neurochemical rewards. The persistence of elevated AEA into recovery also raises practical questions about whether the mood benefits of a long run outlast the run itself, a possibility the marathon data directly support at the 45-minute mark.

What emerges from these two studies is a revised and richer picture of the runner’s high: not a single chemical switch flipped at some magical mileage, but a layered, time-dependent process in which anandamide builds steadily with sustained effort, 2-AG responds late and lingers into recovery, and the subjective experience, euphoria, calm, and pain, shifts in ways that only partially overlap with the lipid signals. For the millions of runners who chase that elusive feeling each weekend, the new work offers both validation and humility. The chemistry of bliss is real, measurable in a vial of blood taken at kilometer 28, and it follows rules that science is only now beginning to read, one marathon and one ultramarathon at a time.

Subject of Research: Endocannabinoid signaling dynamics during marathon and ultramarathon running and their relationship to acute mood and pain responses

Article Title: Endocannabinoid dynamics across marathon and ultramarathon running: evidence from two field studies

Article References: Siebers, M., Huvermann, D., Siebers, C., Canales-Romero, D., Florea-Ghile, A., Keite, L., John, L., Munk, M., Bizjak, R., Witzel, J., Schulz, S., Kirsten, J., Bindila, L., Hinney, A., Grau, M., Bizjak, D. A., Engler, H., & Fuss, J. (2026). Endocannabinoid dynamics across marathon and ultramarathon running: evidence from two field studies. BMC Medicine, 24(1), Article 469. https://doi.org/10.1186/s12916-026-05186-z

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05186-z

Keywords: endocannabinoids, anandamide, 2-arachidonoylglycerol, runner's high, marathon, ultramarathon, endurance exercise, CB1 receptor, lipidomics, euphoria, anxiety, pain perception

At the heart of the study lies the Krebs cycle, also known as the citric acid cycle, the biochemical engine that operates inside mitochondria and converts nutrients into usable cellular energy. In every aerobic cell, this cyclic series of chemical reactions strips electrons from fuel molecules and hands them to the electron transport chain, ultimately generating ATP, the currency that powers virtually all biological work. The Missouri team discovered that endurance exercise modified this machinery differently depending on the sex of the animal. Female rats responded by producing more of the proteins that constitute the energy cycle, effectively expanding the factory floor. Male rats, by contrast, showed greater chemical modification of their existing proteins, a process known as post-translational remodeling that alters how the same molecular components behave without changing their abundance.

Lead author Scott Rector, a Curators’ Distinguished Professor of Medicine and of Nutrition and Exercise Physiology at Mizzou, framed the distinction as a matter of strategy rather than superiority. Female liver cells, in essence, generated more proteins, while male liver cells changed more of the proteins they already had. He emphasized that these are subtle adjustments in the cellular energy process and that neither sex necessarily ends up with a better modification. The observation is significant because it demonstrates that even a deeply conserved pathway like the Krebs cycle can be tuned by exercise in ways that diverge between the sexes, hinting at why metabolic diseases sometimes present and progress differently in men and women.

Despite the divergent molecular routes, the study found that the destination was beneficial for both sexes. Exercise training improved liver health across the board, with measurable reductions in liver fat, improved cholesterol use and disposal, and fewer markers of fibrosis, the accumulation of scar tissue that typically accompanies liver injury or chronic disease. Fibrosis is a critical clinical indicator because progressive scarring can push a fatty liver toward cirrhosis or cancer, outcomes that can be fatal. The fact that both males and females showed these improvements, even while achieving them through different molecular means, underscores the robustness of exercise as a therapeutic intervention for the liver.

One sex-specific difference did emerge at the whole-body level. Male rats appeared to expel more cholesterol from the body than females did, suggesting that the male liver prioritizes export of this lipid, while females may handle cholesterol through other pathways. Cholesterol disposal is a central concern in metabolic health because excess cholesterol contributes to cardiovascular disease and is implicated in the progression of fatty liver conditions. The researchers note that why these different adaptations occur remains unclear, and Rector’s team plans to investigate how the observed changes in the cellular energy cycle might ripple outward to shape whole-body metabolism, the integrated process by which the body creates and uses energy.

The clinical stakes of this work are considerable. The findings point toward metabolic dysfunction-associated steatotic liver disease, or MASLD, the modern name for what was previously called nonalcoholic fatty liver disease. MASLD is the most common chronic liver condition, and its prevalence is climbing in lockstep with rising rates of obesity worldwide. In its most severe forms, MASLD can progress to cirrhosis or liver cancer, both of which can be fatal. Because current treatments are limited and largely rely on lifestyle change, understanding precisely how exercise remodels liver metabolism could enable more precise and personalized interventions, matched to a patient’s sex and metabolic profile rather than applied as a one-size-fits-all prescription.

Methodologically, the study exemplifies the multi-omics approach that defines the MoTrPAC initiative, one of the largest concerted efforts in biology to map the molecular effects of exercise across tissues. Rather than measuring a single molecule or pathway, multi-omics studies integrate data on proteins, their chemical modifications, metabolites, and other molecular layers over time, capturing what the authors describe as temporal and sexually dimorphic remodeling. This means the researchers tracked not only which molecules changed but when they changed during the training period, revealing a dynamic choreography of adaptation rather than a static before-and-after snapshot. Such temporal resolution is essential for distinguishing early signaling events from the stable structural changes that ultimately define a trained liver.

The choice of an experimental animal model is also central to the study’s design. By controlling diet, exercise intensity, and environment in rats, the researchers could isolate the effects of endurance training itself, something that is extraordinarily difficult in human studies where adherence, diet, and genetics introduce confounding variables. The trade-off is that findings in rats must ultimately be validated in people, but the conservation of core metabolic pathways between rodents and humans makes the liver an especially promising tissue for translation. The Krebs cycle, mitochondrial protein remodeling, and cholesterol transport all operate on principles shared across mammals, giving the results a plausible path toward clinical relevance.

The research was a collaborative effort spanning institutions. In addition to Rector, Mizzou study authors include Taylor Kelty, a research assistant professor at NextGen Precision Health, and John Thyfault, professor of cell biology and physiology at the University of Kansas Medical Center, who served as co-corresponding author. Rector’s laboratory is a member of MoTrPAC, which is supported by the NIH Common Fund through cooperative agreements managed by the National Institute of Diabetes and Digestive and Kidney Diseases, the National Institute of Arthritis and Musculoskeletal Diseases, and the National Institute on Aging. The authors declared no competing interests, and the paper appeared in Cell Reports under the title describing endurance exercise as eliciting temporal and sexually dimorphic multi-omics remodeling of liver metabolism.

For the public, the takeaway is both simple and nuanced. Consistent endurance exercise demonstrably improves the metabolic machinery of the liver in both sexes, reducing fat accumulation, improving cholesterol handling, and limiting scarring. But the cellular language in which those improvements are written differs between males and females, with females building more of the energy-producing apparatus and males chemically retuning what they already possess. As researchers continue to decode these sex-specific molecular transducers of physical activity, the long-term promise is a future in which exercise prescriptions and therapies for liver disease are tailored not just to how much a person moves, but to the distinct biology their body uses to respond.

Subject of Research: Sex-specific molecular remodeling of liver metabolism by endurance exercise in a rat model

Article Title: Exercise rewires liver metabolism differently based on sex

Article References: Exercise rewires liver metabolism differently based on sex. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: endurance exercise, liver metabolism, Krebs cycle, mitochondria, sex differences, MASLD, fatty liver disease, cholesterol, fibrosis, MoTrPAC, multi-omics, Cell Reports

Completing a marathon with type 1 diabetes is a metabolic balancing act of unusual complexity. Every kilometre of running consumes muscle glycogen and blood glucose, while carbohydrate intake, insulin sensitivity, adrenaline, dehydration and core temperature all push glucose levels in different directions at different times. Too much circulating insulin, or too little carbohydrate on board, risks hypoglycaemia, a dangerous drop in blood sugar that can cause confusion, collapse or worse. Too little insulin risks hyperglycaemia and ketoacidosis. Despite the well-documented health benefits of regular exercise, fear of these low-glucose episodes remains the single biggest barrier to physical activity for people with type 1 diabetes, affecting up to 45 per cent of those living with the condition.

To examine how amateur runners actually manage this challenge, the researchers designed an observational study built around the 2025 Poznań Marathon, a standard 42-kilometre road race. They recruited 20 amateur runners: 10 with type 1 diabetes of at least one year’s duration and 10 controls without diabetes. The two groups were well matched, showing no significant difference in age (a mean of 35.4 years in the diabetes group versus 39.7 years in the controls), and each group contained eight men and two women. Among the runners with type 1 diabetes, the median duration of the condition was 16.5 years and median glycated haemoglobin, a measure of long-term glucose control, stood at 6.4 per cent, indicating generally well-managed diabetes.

Before the race, each runner with type 1 diabetes followed an individualised insulin strategy agreed in advance. The target pre-race glucose range was set at 140 to 200 mg/dL, deliberately above the normal fasting range to create a safety buffer for the exercise-induced drops to come. Runners using multiple daily injections reduced their basal insulin dose by 25 per cent, while those using non-hybrid insulin pumps cut basal delivery by 50 per cent. Participants on hybrid closed-loop systems, which automatically adjust insulin delivery, instead set a target glucose of 150 mg/dL. Five runners used multiple daily injections, three used continuous subcutaneous insulin infusion pumps, and two used automated insulin delivery systems, reflecting the full spectrum of modern insulin therapy.

Glucose was assessed at five checkpoints along the course: the start line, 10 km, 19 km, 30 km and the finish. At each point, capillary glucose was measured with a standard fingerstick glucometer and compared against readings from two different CGM systems, one intermittently scanned and one transmitting in real time. Carbohydrates or insulin were administered as required throughout the race. The researchers also evaluated the accuracy of the CGM devices using mean absolute relative difference, or MARD, a standard metric that expresses the average absolute percentage difference between sensor readings and reference glucose values. The lower the MARD, the more faithfully the sensor tracks true blood glucose.

The headline performance result was striking in its ordinariness. Marathon completion times did not differ significantly between the groups, with a median finishing time of 228 minutes for the runners with type 1 diabetes and 248 minutes for the controls. In other words, with careful preparation, the runners with diabetes were not merely surviving the distance; they were racing it on equal terms. During the race, they consumed a median of 53.5 grams of carbohydrate per hour, equivalent to 2.61 grams per kilogram of body weight across the entire marathon, a fueling rate consistent with general endurance-sport guidance.

The glucose traces themselves told a reassuring story. Median capillary glucose measured by glucometer stood at 183.5 mg/dL at the start, within the planned pre-race target, then fell to 119.5 mg/dL at 10 km, rose to 142.5 mg/dL at 19 km, dipped to 121.5 mg/dL at 30 km and finished at 108.5 mg/dL. These values remained within or close to a safe range throughout, showing that the pre-race insulin reductions and steady carbohydrate intake kept the runners’ blood sugar from collapsing under the metabolic demands of the distance. Only two hypoglycaemic measurements occurred, and both were in the same participant, who nevertheless completed the race. Notably, that runner had started with a glucose level below 140 mg/dL, beneath the study’s recommended pre-race floor, and consumed 49.5 grams of carbohydrate per hour, slightly less than the group median.

But the study also delivered a caution about the very technology that made it possible. CGM accuracy deteriorated substantially during the marathon. The intermittently scanned system differed from glucometer measurements by an average of approximately 43 per cent, and the real-time system by approximately 37 per cent. Both sensors overestimated capillary glucose, by +32.2 mg/dL and +50.4 mg/dL respectively. This matters because a runner who trusts an inflated sensor reading may believe their glucose is safe when it is in fact falling toward hypoglycaemia. Sensor error during prolonged exercise is thought to arise from a combination of factors, including reduced subcutaneous blood flow as the body shunts blood to working muscle, sweat interfering with sensor adhesion, compression of the sensor site, and the lag between interstitial fluid glucose, which CGM devices measure, and blood glucose, which changes fastest during rapid metabolic swings.

The authors drew a practical conclusion from this discrepancy. In a statement, they said: In this small observational study, all runners with type 1 diabetes completed the marathon, with performance comparable to controls. The runner who experienced low blood sugar had started the race with a glucose level below 140 mg/dL. During the marathon, CGM readings differed from glucometer measurements. For longer endurance events, runners should therefore consider checking their glucose with a glucometer, especially when the sensor reading does not match how they feel. That advice effectively reframes CGM as a trend-monitoring tool rather than a standalone decision-making instrument during ultra-endurance efforts, with fingerstick confirmation reserved for moments when symptoms and sensor numbers diverge.

The research team emphasised that the findings should not be read as a green light for unsupervised endurance racing. They noted that fear of hypoglycaemia is the main barrier to physical activity and affects up to 45 per cent of people with type 1 diabetes despite the major health benefits of regular exercise, and that managing glucose is challenging when levels change rapidly and responses vary between individuals. Their message was nonetheless an optimistic one: with appropriate education and careful blood sugar management, people with type 1 diabetes can successfully take part in even very demanding endurance exercise. They advised anyone with type 1 diabetes preparing for a marathon to discuss an individual glucose, carbohydrate and hydration plan with their doctors before the event, stressing that the most important element is an appropriate insulin management strategy, including reductions in basal and prandial insulin, and that baseline glucose control, exercise experience, diabetes duration and complications all shape individual risk. Some people, they added, should consult a cardiologist before starting endurance training. The team, which supports many athletes with type 1 diabetes, including competitors at the Olympic Games, Ironman triathlon finishers and a runner who completed ten marathons in ten consecutive days, presents the study as further evidence that the condition need not disqualify anyone from the marathon start line, provided the science of glucose management is respected as rigorously as the training plan itself.

Subject of Research: Glucose management, carbohydrate intake and CGM accuracy in amateur marathon runners with type 1 diabetes during a real-world 42 km race

Article Title: Running marathons with type 1 diabetes can be safe, shows small study using continuous glucose monitoring devices

Article References: Running marathons with type 1 diabetes can be safe, shows small study using continuous glucose monitoring devices. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: type 1 diabetes, marathon, continuous glucose monitoring, hypoglycaemia, endurance exercise, insulin adjustment, carbohydrate intake, CGM accuracy, MARD, EASD, sports medicine, glucose management