Drink on schedule or when thirsty? 24 small trials of long rides and runs were pooled. In races against the clock, the water made no clear difference to finishing times

How much to drink during a long run or ride is one of the oldest arguments in endurance sport. One camp says plan it: weigh yourself, work out your sweat rate and drink to a target. The other says thirst is a good enough guide. Drinking too little lets the body's water run down, and drinking far too much can dilute the blood's sodium, which is its own danger. Researchers in Shijiazhuang and Jilin, China, set out to see what the randomised trials actually show, and published a systematic review with meta-analysis in Frontiers in Nutrition on 27 October.

They searched six databases up to 1 June 2026 for trials in healthy adults doing at least 60 minutes of continuous endurance exercise, in which the drinking strategy itself was assigned at random. Twenty-four studies made the cut, all crossover trials where the same people tried each approach, with 6 to 20 participants each and 225 in total. Most were cyclists or runners in a laboratory, often in heat of 30 °C or more. Of the 190 people whose sex was reported, only 20 were women.

Drinking against little or no fluid gave a split picture. In self-paced tests, where people rode or ran a set distance or time as hard as they could, performance was no different on average: minus 0.1%, with a range from 3.6% worse to 3.5% better (four studies, 30 people). In open-ended tests, where people kept going until they couldn't, they lasted about 34% longer with fluid, but that came from only three studies and 21 people, and the range of results a new study might expect still included no benefit at all.

The body's readings leaned towards drinking. Core temperature ended about 0.4 °C lower on average (nine studies) and heart rate about 7 beats a minute lower (12 studies). Both results varied a lot from trial to trial. Planned drinking against drinking to thirst made no clear difference to self-paced performance (0.7%, three studies, 27 people). It did mean people finished having lost about 1.5 percentage points less of their body weight, which the authors point out is simply what happens when you drink more, not a benefit in itself. There was too little evidence to say anything about how much sodium helps.

The authors rate every main result as very low certainty. The trials were small and mostly men, and the authors are open about how the analysis changed after it was registered and that parts of their search records weren't kept. What the review does show is narrower than either camp might like: drinking during long exercise eases some of the strain on the body, but these trials can't name an ideal amount, and they don't show that drinking to a plan beats drinking to thirst when the clock is running.