Running · 6 min read
Carbohydrate taken during a marathon supplies fuel while the body’s stored carbohydrate is being used. For many runners, 60–90 g per hour is a useful range to develop towards in training, with the final amount determined by race intensity, duration and gastrointestinal tolerance. Understanding the physiology helps turn that range into a plan that can be followed for the whole race.
By Chris Gkoufas ·

Carbohydrate is stored as glycogen in skeletal muscle and the liver. Muscle glycogen supplies fuel within the working muscle, while liver glycogen helps support blood glucose. Running also uses fat, with the relative contribution of carbohydrate generally increasing as exercise intensity rises. These fuels support the continual regeneration of ATP required for muscular contraction. Hargreaves and Spriet, 2020
During prolonged exercise, the availability of carbohydrate becomes an important constraint on maintaining intensity. Taking carbohydrate during the race helps maintain blood glucose and carbohydrate oxidation, meaning its use as fuel, and can improve endurance performance. Jeukendrup, 2014
This is the physiological reason to begin fuelling before a pronounced loss of energy develops. A gel taken after pace has collapsed cannot retrospectively supply the preceding hour. We recommend distributing intake from early in the race, using a schedule already practised during longer runs.
The hourly intake is also different from total energy expenditure. Runners do not need to replace every calorie used while running. Carbohydrate consumed during the event contributes alongside fuel already stored in the body; the practical aim is to support the required effort with an amount that can be absorbed and tolerated.
Established guidance recommends 30–60 g/hour for exercise lasting around 1–2.5 hours and up to approximately 90 g/hour for longer endurance events. Higher intakes should use multiple transportable carbohydrates, usually a glucose source combined with fructose. Thomas et al., 2016
For a marathon, duration alone does not settle the amount. A runner completing five hours at a relatively low absolute intensity may use carbohydrate at a different rate from a faster runner racing hard for three hours. Tolerance also varies. We use the 60–90 g/hour range as a practical starting point for planning many marathon efforts, then develop the intake through training rather than imposing the upper end on every runner.
Someone comfortable with 40 g/hour should work progressively towards the intended race intake. Moving directly to 90 g/hour on race morning adds an untested demand at the point when there is least room to resolve it. A lower amount that has been tolerated reliably is preferable to repeatedly forcing an intake that causes substantial symptoms.
These during-exercise recommendations are usually expressed in grams per hour, rather than grams per kilogram of body mass. Intestinal delivery is an important limit on the use of ingested carbohydrate, so a larger runner does not automatically need a proportionately larger hourly dose. Jeukendrup, 2014
Before ingested carbohydrate can be used by working muscle, it has to move through the digestive system and into the circulation. Glucose and fructose use different intestinal transport pathways. Combining them allows a higher rate of carbohydrate delivery and oxidation than relying on glucose alone at high intakes. Maltodextrin, a common gel ingredient, is broken down into glucose and uses that glucose pathway. Jeukendrup, 2014
The practical consequence is to look at the carbohydrate blend when planning higher intakes. Packet count alone says little: two products described as gels can contain different carbohydrate amounts and different mixtures. The nutrition label and instructions for use are more informative than the size of the packet.
Fluid matters too. Some concentrated gels are designed to be taken with water, while other products have different instructions. Rehearse the intended combination of gel and drink, because that is the combination the stomach will encounter during the race. A product tolerated on its own may feel different when combined with a carbohydrate drink and a pre-race breakfast.
Start with the carbohydrate target, subtract the contribution from any drink, and supply the remainder through gels or another familiar source. The examples below show the arithmetic for one hour. They assume that the stated drink quantity is already appropriate for the runner’s fluid needs.
| Hourly total | Example intake |
|---|---|
| 60 g | Three gels containing 20 g each |
| 60 g | Two gels containing 30 g each |
| 60 g | 30 g from a drink plus one 30 g gel |
| 90 g | Three gels containing 30 g each |
At 60 g/hour, a four-hour race requires 240 g of carbohydrate during the event. If each gel contains 30 g and the drinks contain no carbohydrate, that total is eight gels. One gel every 30 minutes, beginning at roughly 15 minutes, spreads the intake across the race: 15, 45, 75, 105, 135, 165, 195 and 225 minutes.
The same calculation can be adapted for a different product. If a drink provides 30 g each hour, one 30 g gel each hour brings the total to 60 g. The drink’s contribution should be based on the amount consumed, rather than the carbohydrate in a full bottle that may not be finished.
Race logistics determine whether the calculation is usable. Check where water is available, how products will be carried and whether aid-station drinks match those used in training. A spare gel is useful if one is dropped or the race takes longer than expected. Avoid trying to recover a missed dose by taking several packets together; resume a manageable schedule and assess how the stomach is responding.
The gastrointestinal system can adapt to repeated feeding during exercise. Research on gut training supports practising carbohydrate intake as part of preparation, although improvements in comfort and absorption are not identical in every athlete. Jeukendrup, 2017
Use relevant long runs to develop the routine, beginning at an amount that is comfortable. Keep the breakfast, product and drink combination reasonably consistent while changing the hourly intake. This makes the result easier to interpret than changing the gel, concentration, meal and pace together.
Record the amount consumed and when symptoms appeared. Discomfort beginning after several hours requires a different review from a product that causes nausea immediately. Persistent pain, vomiting or diarrhoea warrants advice from a sports dietitian or clinician. Repeatedly forcing a poorly tolerated strategy is not useful gut training.
Intakes above 90 g/hour are receiving growing attention. A 2026 review discusses potential applications around 120 g/hour in trained athletes, while emphasising that broad recommendations for 120–200 g/hour exceed the established performance evidence. Higher carbohydrate oxidation does not itself establish a faster marathon. Morton et al., 2026 For recreational runners, developing a consistent, well-tolerated plan within conventional guidance is the more appropriate priority.
Sweat losses depend on the runner and the conditions. Sweat rate and sweat sodium concentration can both vary, which limits the value of a single fixed drinking recommendation for every marathoner. Baker, 2017 Our articles on sweat rate and sweat sodium explain how these measurements inform planning.
A carbohydrate drink supplies both fluid and fuel. If obtaining the desired carbohydrate amount would require excessive drinking, use a more concentrated carbohydrate source and manage fluid separately. Drinking to thirst helps guard against overconsumption; excess fluid can produce exercise-associated hyponatraemia, and sodium intake does not make overdrinking safe. Hew-Butler et al., 2015
The finished fuelling plan should be straightforward: an hourly carbohydrate amount, familiar products, usable timings and a fluid strategy suited to the conditions. The purpose of the science is to make those choices well founded, so that on race day the runner can follow a routine already established in training.
Endurance Performance Profiling