The sub-two-hour marathon performances of Sabastian Sawe and Yomif Kejelcha in London last month put marathon carbohydrate intake back on the front page, with Maurten and Adidas confirming that Sawe averaged close to 115 grams of carbohydrate per hour through the race. A new systematic review published in the journal Sports Medicine pulls together the underlying physiology and asks a more practical question: how much can a recreational marathoner change their own fuelling tolerance, and over what timeline? The answer, drawn from eight controlled trials and a Monash University randomised study, is that a two-week protocol of in-run feeding can cut gastrointestinal symptoms by an average of forty-seven per cent and improve carbohydrate absorption by forty-five to fifty-four per cent.
The mechanism the review describes is straightforward. The intestinal lining produces more SGLT1 and GLUT5 transporters when it is asked to absorb carbohydrate repeatedly during exercise, and the stomach's resting volume expands gradually when it is fed and drunk through long sessions. The two effects compound: the stomach can hold more, and the gut can move more of what reaches it into the bloodstream. Within two weeks of a structured protocol, the same person who reported nausea and bloating at sixty grams per hour can typically tolerate ninety grams per hour without flagging a symptom on the validated GI questionnaire that the review's authors used as a common measurement instrument.
The Monash University randomised trial is the strongest single piece of evidence in the review. Twelve well-trained marathoners followed a two-week protocol of in-run feeding three times a week, with intake building from sixty grams per hour in week one to ninety grams per hour by the end of week two. A matched control group continued normal training without targeted fuelling. The intervention group reduced GI complaint scores by sixty to sixty-three per cent and improved a fixed-time-trial performance by around five per cent at the end of the protocol. The control group recorded no change in either measure across the same fortnight.
The review is careful about the practical limits. Modelled exogenous carbohydrate bioavailability for male and female runners came out at ninety-three and one hundred and eight grams per hour respectively, which the authors describe as the realistic upper bound for trained recreational marathoners in a single race. Pushing higher than that on race day, without the absorption capacity to back it up, is the consistent trigger for symptomatic distress in the eight trials. The review also notes that gut training adapts in both directions: stop the protocol for three weeks and the absorption gain measurably decays, which has implications for tapered race-week training.
The takeaways for amateur marathon coaches are concrete. Adding a structured fuelling block in the eight-to-ten-week range before a goal marathon, with three weekly long-run sessions that progress from sixty to a target intake of ninety grams per hour, is supported by the strongest single piece of evidence in the literature. The review recommends a one-to-zero-point-eight ratio of glucose-derived to fructose-derived carbohydrate to use both transporter systems in parallel, which matches the formulation used in Maurten's race-day mix and the Precision Fuel & Hydration PF 90 gel. The headline of the review is less about elite sub-two-hour science and more about a recreational marathon population still routinely fuelling at thirty to forty grams per hour where the absorption ceiling is sitting twice as high.