Explaining Variation in Sweat Sodium Concentration: Effect of Individual Characteristics and Exercise, Environmental, and Dietary Factors
Design
Stepwise multiple regression on three models: model 1 — 1,944 sweat tests from 1,304 subjects (forearm patches); model 2 — subset with energy expenditure, 1,003 tests from 607 subjects; model 3 — subset with energy expenditure, dietary sodium and VO2max, n=48, whole-body washdown.
Protocol
Retrospective analysis of data compiled from 1944 sweat tests; stepwise multiple regression (P < 0.05 threshold for inclusion), with T values expressing each factor's relative importance (from the abstract).
Three models by available variables: model 1 (1,944 tests from 1,304 subjects), model 2 (energy expenditure subset: 1,003 tests from 607 subjects), model 3 (energy expenditure, dietary sodium and VO2max subset: n = 48).
Whole-body sweat [Na+] predicted from forearm sweat patches in models 1 and 2 and measured directly by whole-body washdown in model 3.
Candidate factors: age group, race/ethnicity, sex, body mass, VO2max, exercise mode, duration, energy expenditure, whole-body sweating rate, fluid balance, pre-exercise urine specific gravity, air temperature, relative humidity, season, dietary and exercise sodium intake.
Key Findings
The models explain only 17-23% of the variation in whole-body sweat [Na+] (adjusted r² 0.17, 0.19, 0.23) — the majority is unexplained
Most important factors: energy expenditure (directly associated with sweat [Na+], likely via sweating rate) and season — warmer months (proxy for heat acclimatization) were associated with LOWER sweat [Na+]
Other significant predictors (model 1): exercise mode (cycling higher), male sex, whole-body sweating rate, body mass (negative); model 2 adds air temperature
NOT significant in any model: age group, race/ethnicity, relative humidity, exercise duration, pre-exercise urine specific gravity, exercise fluid balance, dietary or exercise sodium intake, relative VO2max
Practical Application
Across 1,944 sweat tests, individual characteristics plus exercise, environmental and dietary factors explained only 17-23% of sweat sodium concentration — sodium intake, humidity and duration explained none of it. Population averages and rules of thumb therefore predict a given athlete's sweat sodium poorly; a personal measurement is the only way to know.
Heat acclimatization matters: sweat [Na+] was lower in warmer months, so a test done in early spring can overstate summer losses.
Not from this paper; the base's own practice: test in race-relevant conditions (temperature, intensity, duration) and retest after major changes in fitness, body composition, or season.
Limitations
From the design as the abstract reports it: a retrospective, observational regression analysis of compiled sweat tests — associations, not tested causes.
In models 1 and 2 (the large samples) whole-body sweat [Na+] was predicted from forearm patches; only model 3 (n = 48) measured it directly by whole-body washdown.
Season is used as a proxy for heat acclimatization, as the authors state; acclimatization itself was not measured.
The models accounted for only 17%-23% of the variation, so most of what sets an individual's sweat [Na+] remains unexplained.
How this entry was checked and changed
Content correction 2026-09-27
"Two cyclists with identical mass, sweat rate, environment can differ by 40+ mmol/L" -> not in the abstract; removed.
Journal line said "Journal of Sports Sciences" -> the DOI is Journal of Applied Physiology.
Added the abstract's main finding (energy expenditure; warmer months = lower sweat [Na+]) and the three-model design with sample sizes; "retesting" and "test in race conditions" labelled as house practice. Sport line "cycling" -> "general".
2026-09-28: added Protocol and Limitations from the abstract (Europe PMC, DOI 10.1152/japplphysiol.00391.2022, PMID 36227164).