REF-1354 · Cheuvront & Kenefick, 2014 · Comprehensive Physiology · Paper · PubMed

Does losing 2% of body weight in sweat slow me down?

For endurance this review sets the threshold at ≥2%, via lost volume; strength and power suffer marginally but maybe importantly, thinking little. Heat makes it worse (REF-062).

Main caveat: Read from the abstract only. The studies behind the ≥2% threshold, their environments and populations, the size of the endurance loss, and the authors' figures…
moderate evidenceMeasured on non-athletes to highly trained athletes

Cheuvront and Kenefick, from the U.S. Army Research Institute of Environmental Medicine, review the physiology and the assessment of dehydration and evaluate the literature on dehydration and performance.

Their organising idea is that osmolality and volume both matter, and that losing hypo-osmotic body water (intracellular dehydration) is physiologically distinct from losing iso-osmotic body water (extracellular dehydration).

Sweat loss is of the first kind (the base's gloss, from the same group's 2015 review, which contrasts the hypertonic hypovolaemia of sweating with the iso-osmotic hypovolaemia of cold and altitude, REF-062). They argue that this distinction, which they call underappreciated, is the single most essential aspect of dehydration assessment.

They review how diagnostic accuracy and biological variation decide how useful any assessment method can be, why proper baselines are needed, and how large a deficit must be before the body's osmotic or volume-mediated compensatory responses become reliable and detectable.

That physiology, they say, informs their reading of the performance literature: it suggests a dehydration threshold of 2% or more for impaired endurance performance, mediated by volume loss; strength and power show a marginal but potentially important impairment with no clear threshold and no plausible mechanism; and any effect on cognition appears small and related mainly to distraction or discomfort.

The cost of dehydration for a given sport therefore depends on what the task is made of — endurance, strength, cognition or motor skill.

Our reading: this is the physiological backbone of the base's "about 2%" line — the deficit costs endurance mainly by shrinking circulating volume — and the same group's 2015 review adds that the cost depends on the environment (REF-062); this abstract does not say how much a 2% deficit costs, nor anything about how much to drink.

Protocol

Design: narrative review of the physiology of dehydration, of dehydration-assessment methods (with diagnostic and biological-variation analyses) and of the dehydration-performance literature.
Studies and participants: n = not given in the abstract (no search method, study count or populations stated).
Outcomes covered: body-water deficit by type (hypo-osmotic vs iso-osmotic), assessment methods and baselines, compensatory physiological responses, endurance performance, strength and power, cognition.
Funding and interests: both authors are at the Thermal and Mountain Medicine Division of the U.S. Army Research Institute of Environmental Medicine, Natick (a military research institute); funding and conflicts of interest are not available from the abstract or the PubMed record, which lists no grant.

Key Results

Question (abstract)What the review concludes
What matters most in assessing dehydration?Telling hypo-osmotic loss (intracellular dehydration) from iso-osmotic loss (extracellular dehydration) — "the single most essential aspect"
What makes an assessment method useful?Diagnostic accuracy and biological variation analyses; a proper baseline; a deficit large enough to provoke reliable, detectable osmotic or volume responses (sizes not in the abstract)
Endurance performanceA ≥2% dehydration threshold for impairment, mediated by volume loss
Strength and powerMarginal but potentially important impairment; no clear threshold or plausible mechanism
CognitionPotential impairment appears small, related primarily to distraction or discomfort
A given sportDepends on the makeup of the task (endurance, strength, cognitive, motor skill)
Size of the loss per % of deficit; environment; drinking advicenot in the abstract

Mechanisms

Authors' synthesis (their analysis of the literature; no pooled measurement is described): compensatory osmotic and volume responses become reliable and detectable only once the deficit is large enough, and their reading of the performance studies suggests a ≥2% threshold for impaired endurance performance.
Authors' interpretation: the endurance loss is mediated by volume loss; for strength and power no plausible mechanism is known; cognitive effects come mainly from distraction or discomfort.
Inference: sweat is hypotonic, so a sweating athlete loses more water than salt: body fluids become more concentrated while plasma volume falls (REF-062 calls this the hypertonic hypovolaemia of sweat loss). On this review's reading, it is the volume part that costs endurance; the abstract gives no numbers.
Inference: a "volume-mediated" cost is a cardiovascular one, which fits the same group's view that cardiovascular strain is the lynchpin of the cost in the warm (REF-062) and the review of heat, dehydration and blood flow to brain and muscle (REF-228). How much of the heart-rate drift in a long ride is due to fluid loss rather than heat is not in this abstract.

Practical Application

Does losing 2% of body weight in sweat slow me down? For endurance this review sets the threshold at ≥2%, via lost volume; strength and power suffer marginally but maybe importantly, thinking little. Heat makes it worse (REF-062).
Read the 2% as a threshold, not a rule to drink to: it marks where the endurance cost begins, in the authors' synthesis (B+, review, from the abstract). Not from this paper; the base's own practice (SYNTH-68): drink to thirst, with a weighed sweat rate as the ceiling, never a target (REF-776).
A hot long ride is where it bites: the same group's later review found a >2% deficit usually costs aerobic performance in warm-hot conditions, sometimes in temperate ones, not in the cold (REF-062); a short strength session is not where dehydration matters most (the abstract's "no clear threshold" for strength and power).
One reading is not a diagnosis: the authors stress proper baselines and biological variation for any assessment method (their figures are not in the abstract), so a single hydration reading taken without the athlete's own baseline is weak evidence (inference). Not from this paper; the method the base uses for sweat loss (SYNTH-68, REF-916): weigh before and after representative sessions, correcting for what was drunk and passed.
What it is good for as a source: that a fluid deficit costs endurance mainly through lost circulating volume, and that the effect is task-specific. What it does not show: a dose of performance lost per percent, the role of the environment, or how to drink.

Limitations

Read from the abstract only. The studies behind the ≥2% threshold, their environments and populations, the size of the endurance loss, and the authors' figures for the assessment methods could not be checked.
The threshold is the authors' synthesis. No search method or pooled effect is described in the abstract, so "≥2%" cannot be traced to specific trials here.
Environment not stated in the abstract. The same group's 2015 review makes the cost depend on heat (REF-062); read alone, this abstract can be misquoted as "2% always costs performance".
Laboratory evidence. Most dehydration-performance studies impose a deficit before or during a laboratory test (inference); self-paced racing with drinking on demand is a different situation.
Funding and interests: both authors work for the U.S. Army Research Institute of Environmental Medicine; funding and conflicts are not available from the abstract or the PubMed record. No device maker, federation or data platform is named.
How this entry was checked and changed
Why this level

an authoritative narrative review from the leading group on hydration physiology (the same group as REF-062); the abstract calls its look at the performance literature "a unique evaluation" but names no search, no study count and no pooled effect, so the ≥2% threshold is the authors' synthesis and cannot be traced to studies from the abstract. Read from the abstract (closed access; Europe PMC holds no full text and OpenAlex lists no open copy)