REF-915 · Almond et al., 2005 · New England Journal of Medicine, Vol. · Paper · PubMed

Hyponatremia among Runners in the Boston Marathon

B+moderate evidenceMeasured on recreational to highly trained athletesWritten before October 2026, being re-checked

Runners in the 2002 Boston Marathon were recruited one or two days before the race and gave demographic and training data; after the race they gave a blood sample and reported fluid intake and urine output, and were weighed before and after. Of 766 enrolled, 488 gave a usable finish-line blood sample.

Thirteen percent had hyponatraemia (serum sodium 135 mmol per liter or less) and 0.6 percent had critical hyponatraemia (120 mmol per liter or less). In multivariate analysis, hyponatraemia was associated with weight gain during the race, a racing time over 4:00 hours and body-mass-index extremes.

Female sex, the composition of fluids drunk, and use of non-steroidal anti-inflammatory drugs were not associated with it.

Participants

n=766 runners enrolled from the 2002 Boston Marathon; 488 (64 percent) provided a usable blood sample at the finish line.

Protocol

Recruitment one or two days before the race; survey of demographics and training history.
After the race: blood sample; questionnaire on fluid consumption and urine output during the race; pre- and post-race weights.
Univariate and multivariate regression for risk factors.

Key Results

Incidence: 13 percent hyponatraemia (serum sodium ≤ 135 mmol per liter); 0.6 percent critical hyponatraemia (≤ 120 mmol per liter).
Univariate associations: substantial weight gain, more than 3 liters of fluid during the race, drinking fluids every mile, racing time > 4:00 hours, female sex, low body-mass index.
Multivariate associations:
weight gain — odds ratio 4.2 (95 percent CI 2.2 to 8.2)
racing time > 4:00 hours vs < 3:30 hours — odds ratio 7.4 (95 percent CI 2.9 to 23.1)
body-mass-index extremes
Not associated (authors' conclusion): female sex, composition of fluids ingested, use of nonsteroidal anti-inflammatory drugs.

Practical Application

"How much should I drink in a marathon?" Not so much that you finish heavier: among 488 Boston Marathon finishers, 13 percent were hyponatraemic and 0.6 percent critically so, and weight gain during the race (odds ratio 4.2) and finishing slower than 4:00 hours (odds ratio 7.4 vs under 3:30) were the independent risk factors. Drinking more than 3 liters or at every mile carried the risk in the univariate analysis.
What was drunk mattered less than how much: the composition of fluids ingested was not associated with hyponatraemia. Inference, not tested here: switching water for a sports drink is not a substitute for drinking less.
Slower runners have more time and more aid stations to overdrink — they are the ones to warn.
Not from this paper: the consensus built on this and later studies recommends drinking to thirst (REF-776, A-); the base's drinking and sodium framework is SYNTH-68, and the marathon framework is SYNTH-64.

Limitations

One race, one year; observational — associations, not causes.
488 of 766 enrolled gave a usable blood sample; fluid intake and urine output were self-reported.
Pre-race sodium is not reported in the abstract, so starting values are unknown.
Written from the abstract and record metadata only (full text not read).
How this entry was checked and changed
Why this level

observational cohort with 488 analysed runners