At the 24-hour World Championships, the athletes who drank the most fluid ran the fewest kilometers. The correlation was strong: ρ = −0.776, p = 0.005. More fluid, worse race. That single finding — from 12 elite ultra-runners tracked by Lavoue et al. in 2020 — tells you everything you need to know about why the drink-on-a-schedule vs. drink-to-thirst debate has no clean winner.
A 2025 narrative review from the University of Arkansas H₂O Laboratory synthesized 6 peer-reviewed studies across ultra-endurance running. The conclusion isn’t that one strategy beats the other. It’s that both break down when applied blindly.
What the 2025 Narrative Review on Programmed Hydration Actually Found
Researchers Wierick, Perez, Zhao, and McDermott reviewed 5 field-based ultra-endurance studies and 1 laboratory protocol. They compared two approaches. Programmed fluid intake (PFI) means drinking on a fixed schedule. Thirst-driven fluid intake (TDFI) means drinking only when you feel thirsty. Their core finding: neither is universally superior.
Thirst worked well in field conditions. Schedules showed lab advantages that dissolved in real races.
The review recommends a hybrid model: a minimum fluid floor derived from your sweat rate, with thirst calibrating intake above that floor. That’s not a compromise between two bad options. It’s the approach that matches what the research actually shows.
Before getting into the evidence, one mechanical fact matters: sweat rate in ultra-endurance running spans 0.4 to over 2.5 liters per hour. That’s a sixfold range driven by temperature, pace, and individual physiology. A schedule written for one end of that range can injure an athlete at the other end.
Translation: if your coach built your hydration plan on a 20°C training day, and you race in 33°C heat, that plan may need to double. Thirst alone may not catch up fast enough.
Why Thirst Works Better in the Field Than in the Lab
The Western States 100 is 161 km. In 2014, Hoffman and Stuempfle tracked 383 starters at a peak temperature of 39°C. Two-thirds of athletes used thirst-driven drinking in at least one segment. The result: finish time showed no correlation with body mass change (r = 0.06). Thirst-guided athletes and schedule-guided athletes crossed the line at the same speed.
Thirst isn’t just a sensation. It encodes both fluid volume and osmolality, the salt concentration of your blood. When both signals are working, thirst adjusts not just how much you drink but what you crave. Electrolytes and plain water balance out naturally. A schedule can’t do that. It delivers fluid on a clock regardless of your internal chemistry.
But thirst has a documented lag. In heat above 25°C or at faster-than-expected paces, the signal arrives 15–30 minutes after the deficit starts. Think of thirst as a fuel gauge connected to a sensor halfway down the tank. By the time the needle moves, you’ve already been running on fumes for a few miles.
That lag is where thirst breaks down and a floor rate becomes necessary.
The Lab Study: PFI Wins on Power, Loses on Time
The one laboratory study in the 2025 review — Jeker et al., 2022 — put 8 male cyclists through 5 hours at 30°C, then a 20 km time trial. PFI produced 278 ± 41 watts versus 263 ± 39 watts for thirst-driven drinking. A 5.7% power advantage. Meaningful.
Then the researchers added bathroom stops to the clock.
PFI athletes produced 1,552 mL of urine during the effort. Thirst-driven athletes produced 690 mL. Total race time including urination: PFI 334.5 minutes, TDFI 333.8 minutes. The statistical difference: p = 0.46. No difference at all.
The lab advantage evaporated under race-day logistics.
The Counterintuitive EAH Finding That Changes the Whole Picture
Exercise-associated hyponatremia (EAH) means blood sodium drops below 135 mmol/L. It causes nausea, confusion, and in severe cases, seizures. The standard story is simple: athletes drink too much plain water and dilute their sodium. Don’t overdrink, problem solved.
The Arnaoutis study (n=62, 44 km Olympus Marathon, 2020) breaks that story. Athletes who developed EAH drank 21% less total fluid and consumed 37% less sodium than the normal-sodium finishers. They didn’t overdrink. They retained fluid through inappropriate AVP secretion, a hormonal pathway that causes the kidneys to hold water regardless of intake.
This matters because it reframes the risk calculus. You can’t prevent EAH by watching your total fluid volume alone. At the 246 km Petite Trotte à Léon ultramarathon, 65% of male finishers met biochemical EAH criteria post-race. The hyponatremic runners actually lost less body mass (−2.06 kg) than the normal-sodium finishers (−3.36 kg). They retained fluid despite, not because of, their drinking behavior.
EAH at long distances isn’t just a hydration problem. It’s partly a hormonal one.
Programmed vs. Thirst-Driven Hydration: PFI vs. TDFI Side by Side
| Dimension | Programmed (PFI) | Thirst-Driven (TDFI) |
|---|---|---|
| EAH risk | Higher if schedule exceeds sweat rate in cool/slow conditions | Lower in experienced athletes; AVP-driven EAH possible regardless |
| Dehydration risk | Lower if schedule is individualized; higher if it’s generic | Higher in heat/high-intensity when thirst lags 15–30 min |
| Lab TT power (5h, 30°C) | 278 ± 41 W (5.7% advantage) | 263 ± 39 W |
| Total time with bathroom | 334.5 min (no difference) | 333.8 min |
| 161 km field performance | No finish-time advantage; BM loss 2–3% | No penalty; BM loss 2–3% (r = 0.06) |
| Sodium coupling | Requires separate sodium protocol | Thirst encodes osmolality naturally |
| GI distress | Excess volume slows gastric emptying | Underdinking cuts splanchnic blood flow |
No row gives PFI a clean field-race win. The table doesn’t say TDFI is better either — it says the decision depends on race length, temperature, and pace.
The 13-Fold Sodium Problem
Sweat sodium concentration in marathoners ranges from 7.0 to 95.5 mmol/L. That’s a 13-fold spread. Mean is 42.9 ± 18.7 mmol/L, but the mean is nearly useless. About 20% of male athletes exceed 60 mmol/L. These are the salty sweaters who can develop hypernatremia if they drink plain water on a standard schedule.
Sweat sodium doesn’t track with sweat rate, age, body size, or training history. Each athlete needs individual testing. A generic 300 mg/hour sodium recommendation fits the statistical average athlete. Real athletes aren’t average.
This is why the sodium piece of hydration planning can’t be solved by a drinking schedule alone.
A Real-World Example: Two Approaches, One Race
Take David, a 42-year-old trail runner with 8 years of ultra experience preparing for his first 100-mile race in July. His pre-race sweat testing puts him at 1.2 L/h in temps around 28°C. He’s a moderate sodium sweater at 52 mmol/L — slightly above average.
His training coach had him on a rigid 750 mL/hour schedule built from winter sweat tests at 14°C. By mile 60, he was urinating every 40 minutes, felt bloated, and dropped to a 14-minute mile pace. He finished but felt terrible. Blood work afterward showed sodium at 133 mmol/L — mild EAH despite never feeling thirsty.
The next year, he used a hybrid approach: a minimum floor of 600 mL/hour (from his hot-weather sweat test) and thirst as his ceiling. He carried a sodium-matched electrolyte drink built around his test result. At mile 60 he was urinating twice since the start. He finished 47 minutes faster.
The difference wasn’t willpower. It was matching the protocol to his actual physiology in the actual conditions.
Building the Hybrid Model: Floor, Ceiling, Sodium
The 2025 Wierick review’s recommendation maps to three practical steps.
Step 1 — Establish a sweat-rate floor. Test in conditions close to your race: same temperature range, similar pace, 60–90 minutes. Calculate fluid consumed minus urine produced, divided by hours. That number sets your minimum intake rate for race conditions. Don’t use a winter test for a summer race.
Step 2 — Let thirst calibrate above the floor. In cool or temperate conditions, thirst is accurate enough to be your primary signal after the floor is satisfied. In heat above 25°C, drink at floor rate by default and let thirst push you slightly higher when you feel the cue arriving.
Step 3 — Match sodium to your sweat category. If you haven’t done a sweat sodium test, use behavioral cues: salt rings on dark kit, white residue on skin, strong craving for salty food mid-run. High-signal sweaters need a sodium-forward electrolyte product. Low-signal sweaters may not need supplementation for efforts under 4 hours.
AthleteOS builds this protocol automatically. The race-day hydration brief it generates combines your recorded sweat rate, expected race temperature, and pace estimate to produce a minimum floor rate. It then layers in your sweat sodium category (low, medium, or high) to produce sodium targets that match individual variability, not population averages. Slower athletes or those racing in cool conditions get a ceiling reminder so the floor rate doesn’t push them into EAH territory. Get your race-day hydration brief at myathleteos.com/signup.
Where the Classic 2% Rule Breaks Down
That threshold comes from largely unblinded lab studies where athletes knew their hydration status. When researchers removed that knowledge — giving IV rehydration or placebos without telling athletes which was which — the performance decrement at 2–3% body mass loss disappeared. The 2% rule is real in psychology; its physiology is shakier than most guidelines admit.
At the Western States 100, the top-10 finishers regularly showed more than 2% body mass loss at the 90 km checkpoint. Finish time didn’t correlate with body mass change at r = 0.06. The number barely moves.
That doesn’t mean dehydration doesn’t matter. It means the specific 2% cutoff was set with less scientific rigor than its authority implies.
For more on how training load and readiness interact with race-day fueling decisions, see how Zone 2 training builds aerobic base and managing your fitness score through a build block. If you’re fueling specifically for your Ironman run leg, the triathlon fueling and race-pacing guide covers the additional dimension of heat plus glycogen depletion together.
Thirst is smarter than a clock. But it needs a floor underneath it.