Race fuelling advice has drifted upwards for a decade, from 60g of carbohydrate an hour to 90g and then, among elite marathoners, to figures that would once have been considered reckless. A study published in the Journal of Applied Physiology now provides the mechanistic evidence behind that drift, and it also documents the cost. Eight elite male marathoners with a mean personal best of 2:22:54 completed three 120-minute runs while ingesting labelled glucose-fructose drinks at 60, 90 and 120g an hour.
The metabolic results were unambiguous. Whole-body carbohydrate oxidation averaged 3.07 g per minute at the highest dose, against 2.46 at 90g an hour and 2.09 at 60g an hour. Exogenous oxidation during the second hour — the measure of how much of the ingested fuel was actually being burned rather than sitting in the gut — reached 1.68 g per minute at 120g an hour, against 1.31 and 0.89 at the lower rates. The use of carbon-13 labelling is what makes those numbers credible: it distinguishes ingested carbohydrate from the body's own stores rather than inferring the difference.
The performance-relevant finding is the one about running economy. Oxygen cost at the highest intake was 8.1 mL per kilogram per kilometre lower than at 60g an hour. That is not a trivial difference. For an athlete running at marathon intensity, a reduction of that size in the oxygen required to cover each kilometre is the kind of margin that separates a controlled second half from a survival exercise, and it arrives without any change to fitness, shoes or course.
The complication is gastrointestinal. Moderate or severe symptoms were common in all three trials, not only at the top dose, and the peak severity of nausea, stomach fullness and abdominal cramping was greatest at 120g an hour. That pattern matters for interpretation. It suggests that the gut cost is not a cliff that appears at some threshold but a gradient that steepens, and that anyone moving up the scale is trading measurable metabolic benefit against symptoms they will have to train themselves to tolerate.
The sample is the obvious limitation: eight athletes, all male, all elite, all of whom had spent 24 hours on a high-carbohydrate diet before each trial. None of that generalises automatically to female marathoners, to runners with different gut tolerance, or to anyone racing at four hours rather than two and a quarter. What the study does establish is that the ceiling on useful carbohydrate intake sits higher than the 60 to 90g band that dominated advice until recently, and that reaching it is a training problem rather than a purchasing one.