The purpose of this study was to establish normative data for regional sweat sodium concentration ([Na+]) and whole-body sweating rate in athletes. Data from 506 athletes (367 adults, 139 youth; 404 male, 102 female) were compiled from observational athlete testing for a retrospective analysis. The participants were skill/team-sport (including American football, baseball, basketball, soccer and tennis) and endurance (including cycling, running and triathlon) athletes exercising in cool to hot environmental conditions (15-50 °C) during training or competition in the laboratory or field. A standardised regional absorbent patch technique was used to determine sweat [Na+] on the dorsal mid-forearm. Whole-body sweat [Na+] was predicted using a published regression equation (y = 0.57x+11.05). Whole-body sweating rate was calculated from pre- to post-exercise change in body mass, corrected for fluid/food intake (ad libitum) and urine output. Data are expressed as mean ± SD (range). Forearm sweat [Na+] and predicted whole-body sweat [Na+] were 43.6 ± 18.2 (12.6-104.8) mmol · L(-1) and 35.9 ± 10.4 (18.2-70.8) mmol · L(-1), respectively. Absolute and relative whole-body sweating rates were 1.21 ± 0.68 (0.26-5.73) L · h(-1) and 15.3 ± 6.8 (3.3-69.7) ml · kg(-1) · h(-1), respectively. This retrospective analysis provides normative data for athletes' forearm and predicted whole-body sweat [Na+] as well as absolute and relative whole-body sweating rate across a range of sports and environmental conditions.

Advanced capabilities in noninvasive, in situ monitoring of sweating rate and sweat electrolyte losses could enable real-time personalized fluid-electrolyte intake recommendations. Established sweat analysis techniques using absorbent patches require post-collection harvesting and benchtop analysis of sweat and are thus impractical for ambulatory use. Here, we introduce a skin-interfaced wearable microfluidic device and smartphone image processing platform that enable analysis of regional sweating rate and sweat chloride concentration ([Cl−]). Systematic studies (n = 312 athletes) establish significant correlations for regional sweating rate and sweat [Cl−] in a controlled environment and during competitive sports under varying environmental conditions. The regional sweating rate and sweat [Cl−] results serve as inputs to algorithms implemented on a smartphone software application that predicts whole-body sweating rate and sweat [Cl−]. This low-cost wearable sensing approach could improve the accessibility of physiological insights available to sports scientists, practitioners, and athletes to inform hydration strategies in real-world ambulatory settings.

The purpose of this study was to expand our previously published sweat normative data/analysis (n = 506) to establish sport-specific normative data for whole-body sweating rate (WBSR), sweat [Na + ], and rate of sweat Na + loss (RSSL). Data from 1303 athletes were compiled from observational testing (2000-2017) using a standardized absorbent sweat patch technique to determine local sweat [Na + ] and normalized to whole-body sweat [Na + ]. WBSR was determined from change in exercise body mass, corrected for food/fluid intake and urine/stool loss. RSSL was the product of sweat [Na + ] and WBSR. There were significant differences between sports for WBSR, with highest losses in American football (1.51 ± 0.70 L/h), then endurance (1.28 ± 0.57 L/h), followed by basketball (0.95 ± 0.42 L/h), soccer (0.94 ± 0.38 L/h) and baseball (0.83 ± 0.34 L/h). For RSSL, American football (55.9 ± 36.8 mmol/h) and endurance (51.7 ± 27.8 mmol/h) were greater than soccer (34.6 ± 19.2 mmol/h), basketball (34.5 ± 21.2 mmol/h), and baseball (27.2 ± 14.7 mmol/h). After ANCOVA, significant between-sport differences in adjusted means for WBSR and RSSL remained. In summary, due to the significant sport-specific variation in WBSR and RSSL, American football and endurance have the greatest need for deliberate hydration strategies.

We sought to assess the accuracy of current or developing new prediction equations for resting metabolic rate (RMR) in adolescent athletes. RMR was assessed via indirect calorimetry, alongside known predictors (body composition via dual-energy X-ray absorptiometry, height, age, and sex) and hypothesized predictors (race and maturation status assessed via years to peak height velocity), in a diverse cohort of adolescent athletes (n = 126, 77% male, body mass = 72.8 ± 16.6 kg, height = 176.2 ± 10.5 cm, age = 16.5 ± 1.4 years). Predictive equations were produced and cross-validated using repeated k-fold cross-validation by stepwise multiple linear regression (10 folds, 100 repeats). Performance of the developed equations was compared with several published equations. Seven of the eight published equations examined performed poorly, underestimating RMR in >75% to >90% of cases. Root mean square error of the six equations ranged from 176 to 373, mean absolute error ranged from 115 to 373 kcal, and mean absolute error SD ranged from 103 to 185 kcal. Only the Schofield equation performed reasonably well, underestimating RMR in 51% of cases. A one- and two-compartment model were developed, both r2 of .83, root mean square error of 147, and mean absolute error of 114 ± 26 and 117 ± 25 kcal for the one- and two-compartment model, respectively. Based on the models’ performance, as well as visual inspection of residual plots, the following model predicts RMR in adolescent athletes with better precision than previous models; RMR = 11.1 × body mass (kg) + 8.4 × height (cm) − (340 male or 537 female).