Gut training for runners: better tolerance, unproven absorption

Can you actually train your gut to take more carbohydrate?

Partly, and probably not in the way the phrase implies. Repeated rehearsal of race feeding plausibly changes what a runner can swallow at race intensity without symptoms. Whether it raises the maximum rate at which carbohydrate crosses the intestinal wall is a separate claim, and that one currently rests on physiology rather than on outcome trials in runners. The distinction matters because a stomach that has stopped complaining and a gut that is absorbing more are two different things, and only one of them is what the phrase gut training is usually sold on.

The practical consequence comes first: if a runner's problem is cramping, nausea or an urgent stop during long races, rehearsing the exact fuel at the exact intensity is a reasonable thing to do and the reasoning behind it is coherent. If the expectation is a measured increase in absorptive capacity that converts into a faster marathon, the chain of evidence has a gap in the middle of it, and the gap is named below.

What has to happen between the gel and the muscle?

Gut training is a claim about one link in a long chain. Written out as discrete steps, with the evidence tier for each:

1. Gastric emptying. Carbohydrate has to leave the stomach. Rate depends on the volume in the stomach and on how concentrated the solution is. Established physiology.
2. Digestion to single sugars. Maltodextrin and sucrose are broken down at the intestinal brush border before anything can be absorbed. Established physiology.
3. Transport across the intestinal wall. Glucose crosses on a sodium-coupled transporter; fructose crosses on a different one. This is the reason mixed glucose–fructose formulations exist at all: two routes rather than one. Established physiology.
4. Delivery and oxidation. Absorbed sugar reaches the liver, a portion is retained there, and the remainder is available to working muscle. A paper in European Journal of Applied Physiology (2021) on exogenous carbohydrate notes that carbohydrate is one of the fundamental fuels during prolonged exercise and is widely consumed with the aim of improving performance, while the regulation of that carbohydrate from ingestion through to its use in the working muscle is still not fully appreciated.

One more physiological fact sits across the whole chain: during hard running, blood flow is redistributed towards working muscle and away from the splanchnic circulation. That is established physiology, and it is why a gel that behaves impeccably on an easy run can behave differently at race pace.

Which part of the chain is plausibly trainable, and which is only asserted?

Gastric emptying. The stomach adapts to habitual volume and habitual concentration. This is a mechanistically plausible adaptation and it is the most commonly offered explanation for why regular practice with race drinks feels easier over weeks. Neither of the two papers cited here reports a controlled trial in runners measuring the size of that adaptation across a training block.

Transporter capacity. Intestinal sugar transporter expression responding to habitual substrate load is a general property of intestinal physiology. That is a mechanism, stated as a mechanism. What it is not is a demonstrated finding in trained runners. Here the two papers have to be read against each other. The 2014 Nutrients review, quoted above for the risks of high intake, also lists gut training among the nutritional manipulations it proposes, and states that training the gut with a high intake of carbohydrate may increase absorption capacity and probably prevent gastrointestinal distress. It states that as a proposal, not as a measured result, and neither paper cited here reports an absorption rate measured in runners before and after such a programme. The 2021 review arrives at the same place from the other side: it names the relocation of glucose transporters in epithelial cells as a key step in regulating carbohydrate, and its conclusion asks for future work on the influence of longitudinal training on those regulation processes, the gut among them.

Symptom tolerance. Part of what changes with practice is not absorptive at all. It is knowing the product, the concentration, the timing and the sensation, and no longer interpreting normal fullness at mile 18 as a crisis. This is the least glamorous link, it is largely behavioural and perceptual, and it is also the one a runner can most obviously act on.

What are the limits, and who is this not going to help?

A paper in Nutrients (2014) on carbohydrate-dependent, exercise-induced gastrointestinal distress sets out three categories of cause: mechanical, ischaemic and nutritional, and states that among the nutritional factors a high intake of carbohydrate and the use of hyperosmolar solutions increase the risk of symptoms. Two things follow from that.

First, feeding practice addresses one category out of three. A runner whose symptoms are mechanical – jostling, posture, the simple fact that running shakes the abdomen in a way cycling does not – or ischaemic, tied to high intensity, heat or fluid loss, is unlikely to resolve them by rehearsing gels. Second, and less comfortably, a gut-training programme that consists mainly of pushing intake higher is deliberately increasing the variable that paper identifies as a nutritional risk factor. Concentration and total dose are the things being manipulated, in the direction of more symptoms rather than fewer, in the hope that adaptation outruns the provocation. The same review does give a concentration it regards as workable: glucose at about 6 per cent, or glucose plus fructose at 8 to 10 per cent, recommended as a way of raising carbohydrate intake without delaying gastric emptying. A rehearsal programme that keeps pushing past that is no longer rehearsing the race drink; it is testing a different one.

Distance changes whether any of this matters. In a race short enough to be covered largely by muscle glycogen, absorption rate is not the limiting factor and the whole question is close to irrelevant. In events long enough that ingested carbohydrate becomes a meaningful share of the energy supply, the same physiology becomes the central problem. The mechanism is the same; the practical weight it carries is not.

Individual spread is large and belongs next to any average. Capacity to absorb fructose in particular varies considerably between people, which is established physiology rather than a trial result. Some runners will report no symptoms at intakes that leave others unable to continue, and a programme built on a group mean will be wrong for a substantial minority in both directions. Runners with a diagnosed gastrointestinal condition are outside the reasoning here altogether; that is the point where the question becomes a clinician's.

What do the gut-training measurements actually measure?

Almost every number attached to this topic is a surrogate, and they are not interchangeable.

• Exogenous carbohydrate oxidation rate. Derived from stable isotope tracing and calculation, not read directly off the intestine. It is the closest available proxy for absorptive throughput – and it is a proxy, not a finishing time.
• Gastric emptying. A measured quantity, and a measurement of one step only. Faster emptying does not establish that more sugar crossed the intestinal wall.
• Symptom scores. Self-reported, unblindable, and sensitive to expectation. A runner who has spent six weeks practising fuelling knows it, and knows what the protocol is supposed to do.
• Intestinal permeability and barrier markers. A different construct entirely. Permeability describes how leaky the barrier is; it does not describe how much carbohydrate moves across it per hour. A change in one does not license a claim about the other.

Definition before number. Tolerance has no single value: it is a symptom score, at a stated intake, at a stated intensity, over a stated duration. A tolerance established at easy pace does not transfer to race pace, because the splanchnic blood flow conditions are not the same. Any figure quoted without its protocol is not a measurement.

Modelled versus measured. Watch and app estimates of carbohydrate use are model outputs fitted from pace and heart rate. They are not measurements of absorption, and they cannot serve as independent confirmation of an adaptation they were never able to observe.

Where does the chain break, and what would close it?

The sequence being sold is: weeks of repeated high-carbohydrate feeding, then a higher absorption rate during running, then fewer symptoms in a race, then a faster finish. The first arrow is the one where measurement in runners stops. The last arrow is rarely tested at all, because the studies that could test it are long, expensive and impossible to blind convincingly.

What would settle it: a randomised trial in endurance-trained runners comparing a defined weeks-long feeding protocol against a matched control, with exogenous carbohydrate oxidation and symptom scores measured during a run at race intensity, and a performance endpoint recorded rather than inferred. Until something like that exists, the 2021 European Journal of Applied Physiology review is the honest summary of where the biology stands. Its conclusion names the potential barriers it found — gastric emptying, intestinal absorption, splanchnic and muscle blood flow, muscle uptake and oxidation — and calls for future work on the influence of longitudinal training on those regulation processes, the gut among them. A training protocol cannot be more certain than the physiology it is built on.

Verdict: what the mechanism licenses

Ranked by the strength of evidence behind each, not by how prominently each is marketed:

• Rehearsing the specific fuel, concentration and timing intended for race day – mechanistically plausible, and consistent with human descriptive work identifying carbohydrate dose and solution concentration as nutritional causes of distress. This is the part of gut training that survives scrutiny.
• Raising absorptive capacity itself through repeated exposure – proposed by the 2014 review as something that may increase absorption capacity, and named by the 2021 review as a question future work should settle; not measured in runners in either paper cited here. Interesting, not bankable.
• Using permeability or barrier markers as proof that gut training worked – biomarker-level at best, and a biomarker of a different construct than the one at issue.
• Expecting feeding practice to resolve mechanical or ischaemic symptoms – not addressed by either paper cited here, and mechanistically unlikely given that those causes sit outside the nutritional category.

The criterion a runner can apply to the next gut-training claim they encounter is simply this: ask which of the four links in the chain the claim is actually about, and whether the number offered as proof measured that link or a neighbouring one. A gut that has stopped complaining is a real and useful result. It is not evidence that anything was absorbed faster.

Sources

• Exogenous carbohydrate and regulation of muscle carbohydrate utilisation during exercise (Eur J Appl Physiol, 2021) – used for the statement that exogenous carbohydrate is a fundamental fuel during prolonged exercise and that its regulation from ingestion to use in working muscle remains incompletely characterised.
• Carbohydrate-dependent, exercise-induced gastrointestinal distress (Nutrients, 2014) – used for the three categories of cause (mechanical, ischaemic, nutritional) and for the point that high carbohydrate intake and hyperosmolar solutions increase the risk of gastrointestinal symptoms.