Basketball is a popular team sport characterized by intermittent bouts of high-intensity activity repeated over the duration of a ∼1–2 hr practice or game. Sweat losses vary considerably among basketball players, depending upon body size, exercise intensity, and amount of playing time, among other factors. Although most basketball practices and games are played in indoor temperate environments, sweating rates can be quite high, ranging up to ∼2.5 L/hr (Barnes et al., 2019; Nuccio et al., 2017). High volumes of fluid loss through sweating can increase the risk for significant hypohydration (≥2% body mass loss) and associated detrimental effects on basketball-specific performance (Baker et al., 2007). Thus, individualized sweat testing has been recommended as a tool to help identify players with a higher risk of significant fluid and electrolyte losses (McDermott et al., 2017).
However, standard sweat testing techniques are not easy to implement during live on-court training sessions. For example, the absorbent patch technique requires trained personnel to carefully remove the patch during/after practice, followed by sample storage, and subsequent weighing, and biochemical analyses, to determine local sweating rate (LSR) and electrolyte concentrations (Baker, 2017). Furthermore, determining whole-body sweating rate (WBSR) requires a bit of time and cooperation from the athlete to measure body mass and to assess all fluid/food intake and urine loss during exercise (i.e., mass balance; Cheuvront & Kenefick, 2017). These logistical considerations can be a barrier to conducting standard sweat testing with professional teams, where it is important to avoid disrupting the flow of the practice or distracting the players.
Advancements in wearable technologies have enabled a simpler, less intrusive approach to sweat testing (Choi et al., 2018). For example, our group recently developed a wearable microfluidic device and smartphone image processing platform for in situ measurement of LSR and local sweat chloride concentration ([Cl−]), as well as algorithms to predict WBSR (Gatorade Gx platform). The platform was previously validated with 312 athletes from a wide range of individual and team sports, including competitive youth basketball players (Baker et al., 2020). However, the microfluidic device has not been tested with elite basketball players. Given that some of the highest sweat losses in basketball have been reported among professional players (Osterberg et al., 2009), it is important to determine the validity of the microfluidic device in this cohort. Furthermore, determining the validity of the microfluidic patch when worn on tattooed skin is warranted given the popularity of tattoos among basketball players, and their unknown effects on LSR and sweat electrolyte concentrations (Luetkemeier et al., 2017; Rogers et al., 2019).
Therefore, the primary purpose of this study was to compare measurements of LSR and local sweat [Cl−] using the microfluidic wearable device versus the standard absorbent patch technique in elite basketball players. For ecological validity, the testing took place during live coach-led practices in uncontrolled conditions (e.g., did not control for arm hair, tattoos, or shirt sleeves over patches). A secondary objective was to compare WBSR values predicted from the microfluidic device and algorithms versus measurements of WBSR using the standard mass balance approach. The observational nature of this study also provided the opportunity to measure ad libitum carbohydrate as well as fluid intake during basketball training. Current guidelines suggest that athletes participating in >1 hr of stop-and-go sports, such as basketball, should consume 30–60 g of carbohydrate per hour during exercise to maintain performance (Burke et al., 2011; Thomas et al., 2016). Therefore, since there are few descriptive studies on hydration and nutrition behaviors in elite basketball (Garth & Burke, 2013), another secondary objective was to determine the percentage of players that adhered to current hydration/nutrition guidelines.
Methods
Participants
Subjects in this study were from five different teams of the National Basketball Association’s G League, which is the National Basketball Association’s official minor league for player development. The players met the definition of “elite” per the participant classification framework (McKay et al., 2022). This study (clinical trial identifier: NCT04679532) was approved by the Sterling Institutional Review Board (sterlingirb.com) and has therefore been performed in accordance with the ethical standards in the Declaration of Helsinki. Participants were informed of the experimental procedures and associated risks before providing written informed consent.
Experimental Design
Sweat testing procedures took place during the National Basketball Association G League preseason, approximately 1–2 weeks before the teams’ first regular season games. Players were tested during moderate-intensity practices held in the teams’ training facilities. The training sessions were led by team coaches and consisted of a range of activities including light shooting drills, half-court instruction, and full-court scrimmages. The intensity and duration of the practices, as well as the breaks in activity were not influenced by the investigators. The temperature and relative humidity during practice were measured with a Kestrel 5400 (Nielsen-Kellerman Co.) positioned courtside. Each subject completed one trial and served as their own control. Sweat was collected from the right and left ventral forearm with an absorbent patch (pad size, 11.9 cm2; Tegaderm+ Pad 3582, 3M) and the wearable microfluidic patch (Baker et al., 2020), respectively. This was considered an appropriate comparison between methods, since previous studies have reported no significant bilateral differences in forearm sweating rate and sweat electrolyte concentrations (Baker et al., 2018; Dziedzic et al., 2014; Kenefick et al., 2012). Arm hair was not shaved on either forearm and players could wear long-sleeved shirts or compression arm sleeves as desired.
Measurements
Body mass was measured before and after practice to the nearest 0.1 kg on a digital platform scale (Tanita WB-800S plus), while subjects wore minimal clothing (i.e., compression shorts). Subjects were asked to towel dry before each body mass measurement. During practice, subjects were allowed ad libitum access to water, sports drink (6% carbohydrate, Gatorade Thirst Quencher), and sports nutrition products per their usual options at the teams’ training facilities. One-liter bottles were labeled with each player’s initials and contents of the bottle (sports drink or water) for ease of identification. Drink bottles were kept courtside in ice chests. Players were instructed to drink as much or as little as they wanted but to drink only from their labeled bottles. All drink bottles and nutrition products were weighed before and after consumption to determine fluid and food intake (to the nearest 1 g; Ohaus CS2000). Carbohydrate intake was calculated from the amount of sports drink, or food consumed, and the carbohydrate amount listed on the nutrition facts label of the product. If players had to relieve themselves before the postpractice body mass measurement, participants were asked to collect all urine in a preweighed container provided to them. The urine container was weighed by investigators to determine urine loss (to the nearest 1 g; Ohaus CS2000).
Whole-body sweat loss was calculated from the change in pre- to postexercise body mass, corrected for fluid intake, urine loss, respiratory water loss, and mass loss due to substrate oxidation. Respiratory water loss and mass loss due to substrate oxidation were estimated as 0.2 g/kcal (Cheuvront & Kenefick, 2017), based on energy expenditure derived from a crude estimate of 6.25 metabolic equivalents for moderate basketball activity (average of shooting baskets and playing a game; Ainsworth et al., 2011). WBSR (in liters per hour) was calculated by dividing whole-body sweat loss (in liters) by the duration of exercise. A nominal 1.0 g/ml was used to convert mass to volume.
Absorbent Patch
Before practice, the ventral forearms were wiped clean with alcohol pads and allowed to air dry. Then the absorbent and microfluidic patches were applied to the right and left ventral forearms (about 5 cm below the elbow crease), respectively. An elastic net dressing (Surgilast; Derma Sciences) was put over the absorbent patch on the right forearm to help with adherence to the skin. Absorbent patches were removed upon moderate sweat absorption, but before saturation (as determined by visual inspection), or at the end of practice, whichever came first. Upon removal, the absorbent pad was immediately separated from the Tegaderm using clean forceps and placed in an airtight plastic tube (Sarstedt Salivette). The tubes were sealed with Parafilm™, placed in a plastic bag, and transported to the laboratory for subsequent processing and analyses (which took place within 3 days of sample collection).
The LSR was measured gravimetrically based on the mass of sweat absorbed in the pad (to the nearest 0.001 g using an analytical balance; Mettler Toledo Balance XS204), the pad surface area, and the duration that the patch was on the skin. A standard pad surface area of 11.9 cm2 was used for all patches based on the average ± SD surface area of 11.9 ± 0.1 cm2 measured across 10 randomly selected patches among three lot numbers. Sweat from the absorbent pad was extracted via centrifuge and analyzed for sodium ([Na+]), and potassium ([K+]) concentration, as well as [Cl−] in duplicate by ion chromatography (Dionex ICS-6000, Thermo Scientific). Corrections to local sweat [Na+] and [Cl−] for background electrolytes in the absorbent pads were made according to a previous publication (Baker et al., 2018).
Microfluidic Patch
A detailed description of the microfluidic device and smartphone application-based algorithms (Gatorade Gx Platform) has been published previously (Baker et al., 2020). In brief, the microfluidic patch is a flexible stack of thin-film polymeric materials (3M) consisting of an enclosed multilayered network of microchannels, embedded chloride ion assay, colored dye, and reservoirs that collect microliter amounts of sweat from the skin and route it for colorimetric analysis. A patterned hypoallergenic skin adhesive integrated on the bottom surface of the microfluidic patch defines the sweat collection areas across two distinct microchannels. Microchannel 1 contains an embedded orange dye that mixes with excreted sweat near the inlet and flows down the channel pathway for analysis of local sweat volume. Microchannel 2 contains silver chloranilate used for colorimetric analysis of chloride concentration in excreted sweat. Sweat entering Microchannel 2 mixes with this chloride assay, producing a purple color with an intensity that increases monotonically with [Cl−]. Chloride is the electrolyte of choice for the microfluidic patch because the colorimetric chloride assay is well established, and the chemical reaction is robust at various ambient temperatures. By contrast, sodium assays are less stable, and none exist that precipitate sodium for quantitative measurements via colorimetry.
At the time of absorbent patch removal, photos of microfluidic patches were captured using an image processing software application on a smartphone (iPhone 8, Apple). Supplementary Figure S1 (available online) shows an image of the microfluidic patch on one of the players during a rest period in practice. LSR and sweat [Cl−] were determined using the image processing algorithms previously described (Baker et al., 2020). In addition, WBSR was estimated from the application’s predictive algorithms (based on microfluidic LSR, body mass, sex, air temperature, exercise duration, and exercise mode), as described previously (Baker et al., 2020).
Statistical Analysis
Analyses were carried out using XLSTAT (version 2021.4.1; Addinsoft). In previous work, the correlation between the wearable microfluidic device and absorbent patches for LSR and local sweat [Cl−] were .83 and .92, respectively. A sample size of 48 was required to achieve 80% power in detecting a correlation between patches of .85 using a one-sided hypothesis test versus a correlation of .71 with .05 level of significance. Visual inspection of frequency histograms, and Q–Q plots, as well as Shapiro–Wilk tests were conducted to assess normality of the data. In instances of deviation from normality (WBSR and LSR), data were natural log transformed prior to analyses. These data are presented in their original form in the “Results” section for clarity and ease of practical interpretation. Paired-samples t tests were used to assess the difference between methods for LSR, sweat [Cl−], and WBSR. Scatterplots for LSR and sweat [Cl−] were made and Pearson product–moment correlations were conducted to assess the relation between the absorbent patch and microfluidic patch. Bland–Altman plots for predicted versus measured WBSR were used to illustrate results by individual participants and 95% limits of agreement (LOA) were set around the mean difference (line of bias). For practical context, whole-body sweat loss prediction error (WBSR error multiplied by exercise duration) was also expressed as a percentage of body mass for each subject. Descriptive statistics are presented as mean ± SD. The significance level for all statistical tests was set at alpha = .05.
Results
Sixty-five players agreed to participate in the study. Figure 1 shows a flow diagram of the sample size at each phase from enrollment to analysis and includes reasons for losses and exclusions. One player was excluded because he did not practice on the day of testing. The final sample size analyzed was n = 53 for LSR, n = 41 for sweat [Cl−], and n = 49 for WBSR. The subjects (n = 53) age, body mass, and height were 25 ± 3 years, 92.2 ± 10.4 kg, and 197 ± 8 cm, respectively. Exercise duration was 98 ± 30 min and the conditions in the gymnasium were 21.0 ± 1.2 °C and 42% ± 14% relative humidity. The patches were removed from the forearms after 64 ± 39 min of practice and the amount of sweat collected was 0.73 ± 0.34 g. Local sweat [Na+] and [K+] from the absorbent patch were 38.8 ± 16.7 and 6.0 ± 1.1 mmol/L, respectively.
—Flow diagram of the sample size at each phase from enrollment to analysis, including reasons for losses and exclusions. LSR = local sweating rate; WBSR = whole-body sweat rate.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
—Flow diagram of the sample size at each phase from enrollment to analysis, including reasons for losses and exclusions. LSR = local sweating rate; WBSR = whole-body sweat rate.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
—Flow diagram of the sample size at each phase from enrollment to analysis, including reasons for losses and exclusions. LSR = local sweating rate; WBSR = whole-body sweat rate.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
Figure 2 shows scattergrams of individual data for LSR and local sweat [Cl−] using the absorbent patch and wearable microfluidic device. There were no differences between the absorbent patch and microfluidic patch for LSR (1.25 ± 0.91 mg·cm−2·min−1 vs. 1.14 ± 0.78 mg·cm−2·min−1, p = .34; Figure 2a) or local sweat [Cl−] (30.6 ± 17.3 mmol/L vs. 29.6 ± 19.4 mmol/L, p = .55; Figure 2b). For LSR, the mean bias between methods was −0.11 mg·cm−2·min−1 and 95% LOA was ±1.09 mg·cm−2·min−1. For sweat [Cl−], the mean bias between methods was −1.0 mmol/L and 95% LOA was ±21.1 mmol/L.
—Scattergrams showing individual data for LSR (panel a, n = 53) and local sweat chloride concentration (panel b, n = 41) using the absorbent patch and wearable microfluidic device. Separate analyses were conducted to compare the absorbent patch and wearable microfluidic device for subjects with and without tattooed skin on their ventral forearms. Panel c shows LSR (n = 25 with tattoos; n = 28 without tattoos) and panel d shows sweat chloride concentration (n = 19 with tattoos; n = 22 without tattoos). Horizontal line represents the group median, and the plus sign represents the mean. p > .05 for all comparisons of the absorbent patch versus wearable microfluidic device. LSR = local sweating rate.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
—Scattergrams showing individual data for LSR (panel a, n = 53) and local sweat chloride concentration (panel b, n = 41) using the absorbent patch and wearable microfluidic device. Separate analyses were conducted to compare the absorbent patch and wearable microfluidic device for subjects with and without tattooed skin on their ventral forearms. Panel c shows LSR (n = 25 with tattoos; n = 28 without tattoos) and panel d shows sweat chloride concentration (n = 19 with tattoos; n = 22 without tattoos). Horizontal line represents the group median, and the plus sign represents the mean. p > .05 for all comparisons of the absorbent patch versus wearable microfluidic device. LSR = local sweating rate.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
—Scattergrams showing individual data for LSR (panel a, n = 53) and local sweat chloride concentration (panel b, n = 41) using the absorbent patch and wearable microfluidic device. Separate analyses were conducted to compare the absorbent patch and wearable microfluidic device for subjects with and without tattooed skin on their ventral forearms. Panel c shows LSR (n = 25 with tattoos; n = 28 without tattoos) and panel d shows sweat chloride concentration (n = 19 with tattoos; n = 22 without tattoos). Horizontal line represents the group median, and the plus sign represents the mean. p > .05 for all comparisons of the absorbent patch versus wearable microfluidic device. LSR = local sweating rate.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
Separate analyses were conducted for subjects with (n = 25) and without (n = 28) tattoos to determine if the presence of tattoos impacted the comparison between patch methods for LSR or sweat [Cl−]. Tattoo ink density (i.e., percentage of patch surface area covering tattooed skin) was 55% on average and ranged from 10% to 85%. There were 28 subjects with no tattoos on their ventral forearms and 25 subjects with a tattoo under the patches on one (n = 10 left, n = 2 right) or both (n = 13) of their ventral forearms. LSR for subjects with tattoos was: absorbent 1.22 ± 1.05 mg·cm−2·min−1 versus microfluidic 1.05 ± 0.90 mg·cm−2·min−1 (p = .26; Figure 2c). LSR for subjects without tattoos was: absorbent 1.28 ± 0.78 mg·cm−2·min−1 versus microfluidic 1.22 ± 0.65 mg·cm−2·min−1 (p = .82; Figure 2c). Local sweat [Cl−] for subjects with tattoos was: absorbent 29.7 ± 19.7 mmol/L versus microfluidic 30.8 ± 23.9 mmol/L (p = .72; Figure 2d). Local sweat [Cl−] for subjects without tattoos was: absorbent 31.4 ± 15.3 mmol/L versus microfluidic 28.6 ± 15.0 mmol/L (p = .13; Figure 2d).
Figure 3 shows scatterplots of microfluidic versus absorbent patch results. There were significant correlations between patch methods for LSR (r = .74, p < .0001) and local sweat [Cl−] (r = .83, p < .0001). In addition, there was a significant correlation between absorbent patch LSR and measured WBSR (r = .76, p < .0001), as well as microfluidic patch LSR, and measured WBSR (r = .64, p < .0001).
—Scatterplots of the microfluidic device versus absorbent patch results for LSR (panel a, n = 53) and local sweat chloride concentration (panel b, n = 41). LSR = local sweating rate; CI = confidence interval; Obs = Observed.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
—Scatterplots of the microfluidic device versus absorbent patch results for LSR (panel a, n = 53) and local sweat chloride concentration (panel b, n = 41). LSR = local sweating rate; CI = confidence interval; Obs = Observed.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
—Scatterplots of the microfluidic device versus absorbent patch results for LSR (panel a, n = 53) and local sweat chloride concentration (panel b, n = 41). LSR = local sweating rate; CI = confidence interval; Obs = Observed.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
Figure 4a shows a Bland–Altman plot comparing WBSR values measured via standard mass balance methods during the study versus the WBSR values predicted by the microfluidic patch and algorithms. There was no difference between measured and predicted WBSR (0.97 ± 0.41 L/hr vs. 0.89 ± 0.35 L/hr, p = .22). The mean bias was −0.08 L/hr and 95% LOA between methods was ±0.61 L/hr. Figure 4b shows the whole-body sweat loss prediction error expressed as a percentage of body mass for each subject. Mean whole-body sweat loss error was −0.06% body mass. Out of 49 subjects, n = 35 (71%) predictions were within ±0.5% body mass. Forty-eight (98%) predictions were within ±1.0% body mass, and n = 1 subject’s prediction was +1.5% (2%) of his body mass.
—(a) Bland–Altman plot comparing WBSR values measured using the standard mass balance approach versus WBSR values predicted by algorithms in the smartphone application (n = 49). Mean bias was −0.08 L/hr (p = .22). The lower and upper bounds of the 95% LOA were −0.69 and 0.53 L/hr, respectively. (b) Scattergram showing the whole-body sweat loss prediction error expressed as a percentage of body mass for each subject (n = 49). Horizontal line represents the group median, and the plus sign represents the mean. WBSR = whole-body sweat rate; LOA = limits of agreement.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
—(a) Bland–Altman plot comparing WBSR values measured using the standard mass balance approach versus WBSR values predicted by algorithms in the smartphone application (n = 49). Mean bias was −0.08 L/hr (p = .22). The lower and upper bounds of the 95% LOA were −0.69 and 0.53 L/hr, respectively. (b) Scattergram showing the whole-body sweat loss prediction error expressed as a percentage of body mass for each subject (n = 49). Horizontal line represents the group median, and the plus sign represents the mean. WBSR = whole-body sweat rate; LOA = limits of agreement.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
—(a) Bland–Altman plot comparing WBSR values measured using the standard mass balance approach versus WBSR values predicted by algorithms in the smartphone application (n = 49). Mean bias was −0.08 L/hr (p = .22). The lower and upper bounds of the 95% LOA were −0.69 and 0.53 L/hr, respectively. (b) Scattergram showing the whole-body sweat loss prediction error expressed as a percentage of body mass for each subject (n = 49). Horizontal line represents the group median, and the plus sign represents the mean. WBSR = whole-body sweat rate; LOA = limits of agreement.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
The players’ total sweat loss was 1.51 ± 0.66 L and they consumed 723 ± 438 g of fluid/food (of which 712 ± 419 g was fluid). The players drank 429 ± 406 g sports drink and 283 ± 337 g water. The net change in body mass from before to after practice was −1.1% ± 0.7%. Six players finished practice with >2% body mass loss (Figure 5a). The players consumed 27 ± 25 g of carbohydrate during practice for a rate of 17 ± 15 g/hr. Thirteen players met the carbohydrate recommendation of 30–60 g/hr, while the other 40 players did not. Ten players ingested no carbohydrate during practice (Figure 5b).
—Frequency histograms showing the players’ fluid balance (panel a, n = 49) and carbohydrate intake (panel b, n = 53) during practices.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
—Frequency histograms showing the players’ fluid balance (panel a, n = 49) and carbohydrate intake (panel b, n = 53) during practices.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
—Frequency histograms showing the players’ fluid balance (panel a, n = 49) and carbohydrate intake (panel b, n = 53) during practices.
Citation: International Journal of Sport Nutrition and Exercise Metabolism 32, 5; 10.1123/ijsnem.2022-0017
Discussion
Sweat testing was conducted in elite basketball players to compare measurements of LSR and local sweat [Cl−] obtained with the microfluidic wearable device versus the standard absorbent patch technique. WBSR values predicted from the microfluidic device and smartphone application-based algorithms were also compared with measurements of WBSR using the standard mass balance approach. The main finding of this study was that there were no differences between the absorbent patch and microfluidic patch for LSR or local sweat [Cl−]. Furthermore, there was no significant difference between measured and predicted WBSR, with the microfluidic patch and algorithms predicting measured WBSR within 95% LOA of ±0.61 L/hr compared with the reference method. These results are consistent with previous research using the microfluidic device in competitive youth basketball players and extend the validation of the device to elite basketball players. It is also important to note the ecological validity of these results given that the testing took place in uncontrolled conditions. The microfluidic device produced results similar to reference methods during live coach-led practices, despite the presence of arm hair, tattoos, and/or the wearing of shirt sleeves over the patches.
To put these results in practical context, the microfluidic patch WBSR prediction errors were expressed as a percentage of each player’s body mass to show how the over- or underestimation in WBSR could impact fluid balance (if the athlete were to consume the same amount of fluid that the microfluidic patch/app predicted was lost via sweating). As shown in Figure 4b, 98% of the players’ predicted fluid balance was within ±1.0% body mass, thus avoiding significant under- and overhydration (McDermott et al., 2017). For one player the WBSL prediction error was equivalent to +1.5% of body mass. Still, significant overhydration should be avoided in practice if fluid intake plans aim for up to 1% body mass loss. Moreover, it is noteworthy that some of the differences between measured and predicted WBSR could be due to the reference methods used to measure WBSR in this study. This includes imprecision associated with body mass measurements in the field (e.g., scale sensitivity and trapped sweat in compression clothing) and the crude estimation of nonsweat sources of mass loss.
The microfluidic patch predicted local sweat [Cl−] within 10 mmol/L of standard methods for 70% of players and within 20 mmol/L for 95% of players, which is likely sufficient to identify players with high versus lower risk for large sweat electrolyte losses. As discussed previously, the microfluidic patch measures chloride because there is a well-established colorimetric assay for chloride, while no reliable assay is available for sodium. Since most electrolyte replacement guidelines are centered around sodium, the smartphone application associated with the microfluidic patch converts local sweat [Cl−] to whole-body sweat [Na+]. This conversion uses regression equations developed from previously published data demonstrating significant relations between local and whole-body sweat [Cl−], and whole-body sweat [Cl−] and whole-body sweat [Na+] (Baker et al., 2018, 2019, 2020).
It was of interest to determine the potential impact of tattoos on the accuracy of the microfluidic device since some previous studies (Luetkemeier et al., 2017, 2020), but not others (Beliveau et al., 2020; Rogers et al., 2019), have suggested that tattoos may alter LSR and sweat electrolyte concentrations. In the present study, 25 of the 53 players (47%) had a permanent tattoo on at least one of their ventral forearms under the absorbent and/or microfluidic patches. However, there were still no significant bilateral differences in LSR or sweat [Cl−] when analyses were conducted separately for the 25 tattooed subjects. These results are consistent with others reporting no effect of tattoos on LSR and sweat electrolyte concentrations during exercise-induced sweating (Beliveau et al., 2020; Rogers et al., 2019) and corroborates the validity of local sweat sampling and analyses from tattooed skin.
While the microfluidic wearable is a valid and practical method for sweat testing in the field, it is not without limitations. As noted in Figure 1, we were unable to obtain viable data from n = 11 subjects for LSR and n = 23 subjects for local sweat [Cl−]. One of the most common limitations was the patch becoming unadhered from the arm during high impact aspects of practice (scrimmages). This was a limitation in 14% of players (n = 9) with the microfluidic patch versus 3% (n = 2) with the absorbent patch and affected the ability to obtain both LSR and local sweat [Cl−] measurements. This percentage of patch failure was higher than our previous validation study (Baker et al., 2020), but much of the prior testing was conducted in a controlled laboratory setting (cycling), or in the field with individual sports involving no contact (e.g., running, track and field, tennis), or team sports (e.g., soccer) with typically less physical contact than occurs with basketball. Future research is needed to determine if a stronger patch adhesive or application to a different anatomical location may reduce the prevalence of this failure mode for high contact sports such as basketball. The other limitation with the microfluidic patch occurred when sweat did not advance far enough in Microchannel 2 (chloride channel) to initiate the colorimetric reaction. Insufficient sweat volume was a limitation in 19% of players (n = 12) with the microfluidic patch versus 1.5% (n = 1) with the absorbent patch and tended to occur with players who had less practice time on court and lower sweating rates (∼0.5 mg·cm−2·min−1). This limitation only affected the ability to obtain local sweat [Cl−] measurements because sufficient sweat was still collected in Microchannel 1 (volume channel). Future research is needed to address this limitation, perhaps by applying the patch to a location with a higher LSR (e.g., back) during shorter duration exercise or players with lower sweating rates.
There are few descriptive studies on hydration and nutrition behaviors during exercise in elite basketball players (Garth & Burke, 2013). Current guidelines suggest that athletes participating in >1 hr of stop-and-go sports should consume 30–60 g of carbohydrate per hour during exercise (Burke et al., 2011; Thomas et al., 2016). In the present study, 75% of players consumed less than 30–60 g carbohydrate per hour. Nearly a fifth of players (10 out of 53) ingested no carbohydrate at all. For most players in the present study, who drank enough fluid to avoid significant hypohydration, they could consider solid/semisolid forms (gels, chews, and bars) of carbohydrate to increase intake to 30–60 g/hr. By contrast, the 12% of players who accrued significant hypohydration (>2% body mass loss) should be advised to increase their carbohydrate intake through sports drink and/or water plus solid/semisolid carbohydrate sources (Pfeiffer et al., 2010a, 2010b). However, there are potential limitations in interpreting the carbohydrate intake data since the present study was not a game setting and no instructions were given to players. Alternative reasons for not needing/wanting to consume 30–60 g/hr during practice (e.g., significant carbohydrate intake before exercise or players attempting to reduce caloric intake to lose weight) cannot be ruled out.
In conclusion, the wearable microfluidic device and smartphone application algorithms provide similar LSR, local sweat [Cl−], and WBSR results compared with the standard field-based methods in elite male basketball players, with or without tattooed skin, during typical moderate-intensity practice sessions that include a mix of noncontact drills and live high contact scrimmages. With respect to hydration and nutrition behaviors, only 12% of players accrued >2% body mass loss, while 75% of players did not meet the carbohydrate intake recommendations (30–60 g/hr) during practice. More research regarding carbohydrate intake of elite athletes during basketball training is needed.
Acknowledgments
The authors thank the G-League and players for participating in this study. The authors also thank Bridget Sopena and Kevin Luhrs for assistance with study set-up; Eric Freese, Anthony Wolfe, Jon Davis, Kris Osterberg, Sarah Stinman, Peter Youngman, and Shanice Johnson for assistance with data collection; and Shankang Xu, Julian Archer, Peter John D. De Chavez, Jessica Wallace, and Jeffrey P. Model for assistance with data analysis. This study was funded by the Gatorade Sports Science Institute, a division of PepsiCo, Inc. This study is registered in clinicaltrials.gov under the clinical trial identifier NCT04679532. Author contributions: This study was designed by Baker and Ghaffari; data were collected by Baker, Keyes, Engel, and King; biochemical analyses of sweat samples was conducted by Brown; data interpretation was completed by Baker, Seib, and Aranyosi; and manuscript write-up was completed by Baker, Seib, Aranyosi, and Ghaffari. All authors approved the final version of the paper. Baker, Keyes, Brown, Engel, and King are employed by the Gatorade Sports Science Institute, a division of PepsiCo, Inc. Seib, Aranyosi, and Ghaffari are employed by Epicore Biosystems, Inc. The views expressed in this manuscript are those of the authors and do not necessarily reflect the position or policy of PepsiCo, Inc. or Epicore Biosystems, Inc.
References
Ainsworth, B.E., Haskell, W.L., Herrmann, S.D., Meckes, N., Bassett, D.R., Jr., Tudor-Locke, C., Greer, J.L., Vezina, J., Whitt-Glover, M.C., & Leon, A.S. (2011). 2011 Compendium of physical activities: A second update of codes and MET values. Medicine & Science in Sports & Exercise, 43(8), 1575–1581. https://doi.org/10.1249/MSS.0b013e31821ece12
Baker, L.B. (2017). Sweating rate and sweat sodium concentration in athletes: A review of methodology and intra/interindividual variability. Sports Medicine, 47(Suppl. 1), 111–128. https://doi.org/10.1007/s40279-017-0691-5
Baker, L.B., De Chavez, P.J.D., Ungaro, C.T., Sopena, B.C., Nuccio, R.P., Reimel, A.J., & Barnes, K.A. (2019). Exercise intensity effects on total sweat electrolyte losses and regional vs. whole-body sweat [Na(+)], [Cl(-)], and [K(+)]. European Journal of Applied Physiology, 119(2), 361–375. https://doi.org/10.1007/s00421-018-4048-z
Baker, L.B., Dougherty, K.A., Chow, M., & Kenney, W.L. (2007). Progressive dehydration causes a progressive decline in basketball skill performance. Medicine & Science in Sports & Exercise, 39(7), 1114–1123. https://doi.org/10.1249/mss.0b013e3180574b02
Baker, L.B., Model, J.B., Barnes, K.A., Anderson, M.L., Lee, S.P., Lee, K.A., Brown, S.D., Reimel, A.J., Roberts, T.J., Nuccio, R.P., Bonsignore, J.L., Ungaro, C.T., Carter, J.M., Li, W., Seib, M.S., Reeder, J.T., Aranyosi, A.J., Rogers, J.A., & Ghaffari, R. (2020). Skin-interfaced microfluidic system with personalized sweating rate and sweat chloride analytics for sports science applications. Science Advances, 6(50), Article e3929. https://doi.org/10.1126/sciadv.abe3929
Baker, L.B., Ungaro, C.T., Sopena, B.C., Nuccio, R.P., Reimel, A.J., Carter, J.M., Stofan, J.R., & Barnes, K.A. (2018). Body map of regional versus whole body sweating rate and sweat electrolyte concentrations in men and women during moderate exercise-heat stress. Journal of Applied Physiology, 124(5), 1304–1318. https://doi.org/10.1152/japplphysiol.00867.2017
Barnes, K.A., Anderson, M.L., Stofan, J.R., Dalrymple, K.J., Reimel, A.J., Roberts, T.J., Randell, R.K., Ungaro, C.T., & Baker, L.B. (2019). Normative data for sweating rate, sweat sodium concentration, and sweat sodium loss in athletes: An update and analysis by sport. Journal of Sports Sciences, 37(20), 2356–2366. https://doi.org/10.1080/02640414.2019.1633159
Beliveau, J., Perreault-Briere, M., Jeker, D., Deshayes, T.A., Duran-Suarez, A., Baker, L.B., & Goulet, E.D.B. (2020). Permanent tattooing has no impact on local sweat rate, sweat sodium concentration and skin temperature or prediction of whole-body sweat sodium concentration during moderate-intensity cycling in a warm environment. European Journal of Applied Physiology, 120(5), 1111–1122. https://doi.org/10.1007/s00421-020-04350-4
Burke, L.M., Hawley, J.A., Wong, S.H., & Jeukendrup, A.E. (2011). Carbohydrates for training and competition [review]. Journal of Sports Sciences, 29(Suppl. 1), S17–S27. https://doi.org/10.1080/02640414.2011.585473
Cheuvront, S.N., & Kenefick, R.W. (2017). CORP: Improving the status quo for measuring whole body sweat losses (WBSL). Journal of Applied Physiology, 123(3), 632–636. https://doi.org/10.1152/japplphysiol.00433.2017
Choi, J., Ghaffari, R., Baker, L.B., & Rogers, J.A. (2018). Skin-interfaced systems for sweat collection and analytics. Science Advances, 4(2), Article eaar3921. https://doi.org/10.1126/sciadv.aar3921
Dziedzic, C.E., Ross, M.L., Slater, G.J., & Burke, L.M. (2014). Variability of measurements of sweat sodium using the regional absorbent patch method. International Journal of Sports Physiology and Performance. 9(5), 832–838. https://doi.org/10.1123/ijspp.2013-0480
Garth, A.K., & Burke, L.M. (2013). What do athletes drink during competitive sporting activities?. Sports Medicine, 43(7), 539–564. https://doi.org/10.1007/s40279-013-0028-y
Kenefick, R.W., Cheuvront, S.N., Elliott, L.D., Ely, B.R., & Sawka, M.N. (2012). Biological and analytical variation of the human sweating response: Implications for study design and analysis. American Journal of Physiology—Regulatory, Integrative and Comparative Physiology, 302(2), R252–R258. https://doi.org/10.1152/ajpregu.00456.2011
Luetkemeier, M.J., Allen, D.R., Huang, M., Pizzey, F.K., Parupia, I.M., Wilson, T.E., & Davis, S.L. (2020). Skin tattooing impairs sweating during passive whole body heating. Journal of Applied Physiology, 129(5), 1033–1038. https://doi.org/10.1152/japplphysiol.00427.2019
Luetkemeier, M.J., Hanisko, J.M., & Aho, K.M. (2017). Skin tattoos alter sweat rate and Na+ concentration. Medicine & Science in Sports & Exercise, 49(7), 1432–1436. https://doi.org/10.1249/MSS.0000000000001244
McDermott, B.P., Anderson, S.A., Armstrong, L.E., Casa, D.J., Cheuvront, S.N., Cooper, L., Kenney, W.L., O’Connor, F.G., & Roberts, W.O. (2017). National Athletic Trainers’ Association position statement: Fluid replacement for the physically active. Journal of Athletic Training, 52(9), 877–895. https://doi.org/10.4085/1062-6050-52.9.02
McKay, A.K.A., Stellingwerff, T., Smith, E.S., Martin, D.T., Mujika, I., Goosey-Tolfrey, V.L., Sheppard, J., & Burke, L.M. (2022). Defining training and performance caliber: A participant classification framework. International Journal of Sports Physiology and Performance, 17(2), 317–331. https://doi.org/10.1123/ijspp.2021-0451
Nuccio, R.P., Barnes, K.A., Carter, J.M., & Baker, L.B. (2017). Fluid balance in team sport athletes and the effect of hypohydration on cognitive, technical, and physical performance. Sports Medicine, 47(10), 1951–1982. https://doi.org/10.1007/s40279-017-0738-7
Osterberg, K.L., Horswill, C.A., & Baker, L.B. (2009). Pregame urine specific gravity and fluid intake by National Basketball Association players during competition. Journal of Athletic Training, 44(1), 53–57. https://doi.org/10.4085/1062-6050-44.1.53
Pfeiffer, B., Stellingwerff, T., Zaltas, E., & Jeukendrup, A.E. (2010a). CHO oxidation from a CHO gel compared with a drink during exercise. Medicine & Science in Sports & Exercise, 42(11), 2038–2045. https://doi.org/10.1249/MSS.0b013e3181e0efe6
Pfeiffer, B., Stellingwerff, T., Zaltas, E., & Jeukendrup, A.E. (2010b). Oxidation of solid versus liquid CHO sources during exercise. Medicine & Science in Sports & Exercise, 42(11), 2030–2037. https://doi.org/10.1249/MSS.0b013e3181e0efc9
Rogers, E., Irwin, C., McCartney, D., Cox, G.R., & Desbrow, B. (2019). Tattoos do not affect exercise-induced localised sweat rate or sodium concentration. Journal of Science and Medicine in Sport, 22(11), 1249–1253. https://doi.org/10.1016/j.jsams.2019.06.004
Thomas, D.T., Erdman, K.A., & Burke, L.M. (2016). American college of sports medicine joint position statement. Nutrition and athletic performance. Medicine & Science in Sports & Exercise, 48(3), 543–568. https://doi.org/10.1249/MSS.0000000000000852
