Drinking and sodium
Drink to thirst by default, plan drinking only in the heat when the effort is hard or over ~60 min, and never beyond your measured sweat loss, treat finishing heavier as the warning, and fit sodium to the fluid you actually replace; salt prevents neither hyponatraemia in someone who overdrinks nor most cramps
The question it answers
How much should this athlete drink in training and in a race, does salt matter for them, how do they find their own sweat rate, and how do they avoid both a performance-costing deficit and dangerous overdrinking?
Default (what to prescribe first)
Drink to thirst. Using thirst to guide drinking should limit overdrinking while still preventing excessive dehydration, and the cues persist in hot and cold conditions (REF-776, A-, consensus, Grade 1C). The ultramarathon nutrition review says euhydration can generally be maintained by drinking to thirst (REF-663, B+, narrative review).
Never drink more than you lose. EAH, and all known deaths from it, come from drinking hypotonic fluid (water or sports drink) beyond what is lost, likely with non-osmotic antidiuretic hormone secretion (REF-776, Grade 1A). Almost all symptomatic cases gained or held body weight when some loss was expected (REF-776).
In the heat, keep the net deficit under about 2% of body mass. The UCI statement sets <2%, especially in heat, with strategies from ad libitum to a plan built on measured sweat rates (REF-748, A, cycling, written for professional riders). Losses >2% usually impair aerobic performance in warm-hot conditions, sometimes in temperate ones and not in the cold (REF-062, B+, narrative review).
Sodium follows fluid. There is no evidence athletes need more dietary sodium day to day, and sodium around exercise should be planned against how much is drunk and lost, not sweat sodium alone (REF-220, B+, narrative review).
Weigh yourself before and after representative sessions to learn your sweat rate: body-mass change, corrected for what was drunk, passed and trapped in clothing, is the simplest and most accurate whole-body method (REF-916, B+, methods review). The session details below are the base's own practice.
Decision — who → what
| Situation | Drink and sodium (volumes and cut-offs are practice unless a refId carries them) | Source and how sure |
|---|---|---|
| Short session, under ~60-90 min, cool or temperate | Thirst; water is fine; no sodium target; start euhydrated | An editorial says thirst may suffice for low-to-moderate exercise <60 min in cool climates (REF-225, C). >2% loss did not impair submaximal aerobic performance in cold-cool conditions (REF-062, B+). The 60-90 min line is practice |
| 1-3 h training | Thirst-led with a bottle always available; a hypotonic carbohydrate-electrolyte drink during the ride; sodium to taste | Hypotonic drinks held plasma volume best over 0-180 min of exercise (−6.3% vs isotonic −8.7%, water −7.5%) (REF-224, A, 28 studies; a surrogate, no performance outcome). A model found sodium replacement unnecessary for a soccer match or an elite marathon in every realistic scenario (REF-223, C, no participants). SYNTH-02's 500-800 mL/h is its own practice; inference: never above the athlete's weighed loss |
| Hot race or key session over ~60 min | Plan bottles per hour from a weighed sweat rate so the deficit stays under ~2%, and never above 100% of the loss; heat-acclimate first; pre-cool | Planned drinking favoured in heat for high intensity and >60 min, because thirst may appear only after ~1.5 L is lost (REF-225, C, editorial). ~2.4% hypohydration cut 15-min cycling work 165 → 152 kJ at 34 °C (REF-226, B, 7 active men, blinded). Consensus order: acclimatization, then euhydration, then cooling (REF-061, A); protocol SYNTH-04 |
| Hot race: pre-start sodium | Optional; rehearse first | ~757 mL of a 164 mmol/L sodium drink (≈2.9 g) finished 45 min before a run at 32 °C raised plasma volume 4.5% and time to exhaustion 75 → 96 min in those who ran to exhaustion (REF-007, B, 8 trained male runners; running and an exhaustion test, transfer to a cycling race is inference) |
| Marathon, Ironman, ultra | Thirst, not a schedule (in the heat the hot-race row above applies: a plan from a weighed sweat rate, never above the loss, with thirst overriding it downward); weigh before and after; finishing heavier, or not lighter, is the warning. Sodium in drinks is fine, but it is not a licence to drink more | Sports drinks (about 10-38 mmol/L sodium) are far below blood (~140 mmol/L), so they cannot prevent EAH when intake stays excessive (REF-776, Grade 1C). Body-weight loss under 0.75 kg after a standard marathon went with asymptomatic EAH (REF-776). Among 488 Boston Marathon finishers, 13 percent were hyponatraemic and 0.6 percent critically; weight gain (odds ratio 4.2) and a finish over 4:00 hours vs under 3:30 (odds ratio 7.4) were independent risk factors, and the composition of fluids was not (REF-915, B+, one race, observational) |
| Who is at risk of EAH | Anyone whose intake stays above sweat, respiratory and urine losses; long events; the severe cases mostly recreational. Warn slower finishers and those at body-mass-index extremes | REF-776. In one marathon cohort, after adjustment: weight gain, a finish over 4:00 hours and body-mass-index extremes; female sex, fluid composition and NSAID use were not associated (REF-915, B+). Inference: a slower athlete spends more hours near fluid, so a fixed mL/h plan adds up to more water |
| Salty sweater (white residue, stinging sweat — practice cues, not validated) | Sodium to taste in training; a sweat-sodium test earns its cost only for very long events where the athlete plans to drink back most losses | Sweat sodium: only 17-23% of its variation explained by the measured factors; sodium intake, humidity and duration explained none (REF-222, B+, 1,944 tests). Targeted sodium needed only in a 160 km ultra with sweat sodium ≥40 mmol/L and >80% of losses replaced (REF-223, C, model, running). A sweat-sodium result depends on collection site, timing, skin cleaning, storage and analysis — ask how it was obtained (REF-916, B+) |
| Cramp-prone athlete | Do not prescribe salt or extra fluid as the fix; stop, stretch gently until it releases, ride or run on easier; then (practice) work on pacing, conditioning for race demands, and heat | Dehydration and electrolyte explanations rest on anecdote, 18 case reports and one case-control study of 10, and four prospective cohorts do not support them (REF-232, B+). Risk factors: cramp history, fatigue, heat and humidity, novel or harder exercise; sweat losses a minor contributor for most; magnesium no meaningful benefit (REF-233, B+) |
| Masters athlete | In heat, plan drinking rather than relying on thirst alone, still capped by weighed loss | UCI: thirst and sweat capacity decline, so hydration in heat needs a plan (REF-748, A, a position statement line; no primary masters hydration study in the base). This departs from REF-776's thirst rule; the cap at measured loss is the base's reconciliation (inference) |
| Woman | No separate volume rule | Women sweated less than men only at the highest heat-loss requirement (REF-712, B, 8 + 8, cycling); sweat rate did not differ by cycle phase (REF-709, B+); in Boston, female sex went with EAH only before adjustment (REF-915). Inference: lower sweat loss means a fixed plan overshoots sooner |
The ultra conflict, stated. The ISSN ultra stand gives 450-750 mL/h (~150-250 mL every 20 min), rising in heat, and says sodium may be needed above what most commercial products provide (>575 mg/L) (REF-528, A, running).
The hyponatraemia consensus says drink to thirst and that sodium does not protect someone who overdrinks (REF-776, A-), and the ultra review agrees on thirst (REF-663, B+).
The base's resolution (inference, not tested): use the stand's range only as a planning ceiling for athletes whose weighed sweat rate is at or above it, never as a floor to drink to; thirst overrides the schedule downward.
One world-class mountain runner drank 0.87 L/h on average (1 L/h with about 1.5 g of sodium per litre after the first 4 h) in a 100-mile race at up to 40 °C and still finished 4.3% lighter (REF-1194, C+, case study), above the ~2% heat limit in this framework's default; the authors wrote that higher intakes "would be required" but may not be practical, and part of the mass lost over 14 h is fuel oxidised, not water (inference).
An observation, not a test of either rule.
The sweat-rate test (house protocol)
The method — body-mass change corrected for non-sweat mass changes and sweat trapped in clothing — is the simplest and most accurate for whole-body sweat rate (REF-916, B+).
The session details are the base's own practice: after voiding, weigh nude and towel-dried; do a representative session of ≥60 min at race intensity and in race-like conditions; weigh again nude and dried. Sweat loss (L) ≈ mass before − mass after (kg) + fluid drunk (L) − any urine passed; divide by hours. Mass change ÷ start mass × 100 is the % deficit.
Before and after
How to progress / periodise
Not from a study; the base's own practice: do one sweat test in base training and one in the hot months; rehearse the race drinking and fuelling in long sessions, as the ultra stand's individualised framing asks (REF-528); retest after heat acclimation (SYNTH-04) because sweat sodium tracked season (REF-222). Settle the plan before race day: riders told they were 2% dehydrated when actually ~1.5% down did 5.6 ± 6.1% less work (REF-227, B, 9 active men, heat).
When NOT to use it / red flags
What the evidence does and does not show
Evidence gaps
How this entry was checked and changed
(the safety rule rests on an A- consensus whose recommendations are mostly graded 1C and a B+ Boston Marathon cohort; the weigh-in method on a B+ methods review; the heat-performance threshold on a B+ review, an A cycling position statement and two small B crossovers in active men; sodium on a B+ review and a C model with no participants; drinking volumes, the sweat-test session details and the hour cut-offs are house practice)