This study compared measured serum [Na+] (S[Na+]; brackets denote concentration) with that predicted by the Nguyen-Kurtz equation after manipulating ingested [Na+] and changes in body mass (ΔBM) during prolonged running in the heat. Athletes (4 men, 4 women; 22–36 yr) ran for 2 h, followed by a run to exhaustion and 1-h recovery. During exercise and recovery, subjects drank a 6% carbohydrate solution without Na+ (Na+0), 6% carbohydrate solution with 18 mmol/l Na+ (Na+18), or 6% carbohydrate solution with 30 mmol/l Na+ (Na+30) to maintain BM (0%ΔBM), increase BM by 2%, or decrease BM by 2% or 4% in 12 separate trials. Net fluid, Na+, and K+ balance were measured to calculate the Nguyen-Kurtz predicted S[Na+] for each trial. For all beverages, predicted and measured S[Na+] were not significantly different during the 0%, −2%, and −4%ΔBM trials (−0.2 ± 0.2 mmol/l) but were significantly different during the +2%ΔBM trials (−2.6 ± 0.5 mmol/l). Overall, Na+ consumption attenuated the decline in S[Na+] (−2.0 ± 0.5, −0.9 ± 0.5, −0.5 ± 0.5 mmol/l from pre- to postexperiment of the 0%ΔBM trials for Na+30, Na+18, and Na+0, respectively) but the differences among beverages were not statistically significant. Beverage [Na+] did not affect performance; however, time to exhaustion was significantly shorter during the −4% (8 ± 3 min) and −2% (14 ± 3 min) vs. 0% (22 ± 5 min) and +2% (26 ± 6 min) ΔBM trials. In conclusion, when athletes maintain or lose BM, changes in S[Na+] can be accurately predicted by changes in the mass balance of fluid, Na+, and K+ during prolonged running in the heat.
- Nguyen-Kurtz equation
- carbohydrate-electrolyte solution
- endurance performance
exercise-associated hyponatremia (EAH) is defined as a serum sodium concentration (S[Na+]) less than 135 mmol/l and occurs primarily in endurance events such as marathons, ultramarathons, and triathlons (15). Although the condition is rare, severe EAH (typically S[Na+] < 125 mmol/l) can result in cerebral/pulmonary edema, coma, and death (1, 15, 22). Therefore, research regarding the mechanisms and prevention of EAH has been of utmost importance since it was first described in 1985 (18). Montain et al. (14) and Weschler (27) have completed detailed theoretical quantitative analyses on sodium balance and, by extension, the etiology of EAH, using a prediction equation developed by Kurtz and Nguyen (10) to calculate postexercise S[Na+] on the basis of changes in the mass balance of water, Na+, and K+ (14, 27).
The mathematical model predicts that S[Na+] is most sensitive to changes in total body water and thus the primary cause of EAH is an increase in body mass (BM). That is, the dilution of S[Na+] and the symptoms associated with EAH are mediated by the consumption of fluid at a rate that grossly exceeds sweating rate. The model also predicts that S[Na+] is moderately sensitive to changes in the mass electrolyte balance of Na+ and K+ (ΔE), such that the consumption of a carbohydrate-electrolyte solution (CES) will attenuate the decline in S[Na+] compared with the consumption of water alone (14, 27). This theoretical model provides a practical means for athletes to predict changes in S[Na+] so as to decrease their risk of EAH. However, predictions using this model have not been directly compared with empirical data of postexercise S[Na+] in athletes with known preexercise S[Na+], BM, sweating rates, sweat [Na+] and [K+] (brackets denote concentration), fluid intake, and Na+ and K+ intake in a controlled laboratory setting.
Therefore, the purpose of the present study was to compare measured S[Na+] with that predicted by the Nguyen-Kurtz equation in athletes after prolonged running in the heat to assess the equation's accuracy. A second goal of the study was to measure the effects of beverage [Na+] (0, 18, and 30 mmol/l) and pre- to postexercise ΔBM (+2%, 0%, −2%, and −4%) on measured S[Na+]. Furthermore, we aimed to measure the effects of a BM deficit (−2% and −4%) vs. fluid consumption to match sweating rate (i.e., 0%ΔBM) on prolonged running performance in the heat. We hypothesized that 1) the Nguyen-Kurtz equation would accurately predict the measured postexercise S[Na+], 2) increased [Na+] in a CES would attenuate the decline in S[Na+] compared with a beverage with no Na+, 3) an increase in BM as a result of overdrinking relative to sweat losses (+2%ΔBM) would be associated with a decrease in S[Na+], and 4) a decrease in BM via underdrinking relative to sweat losses (−2% and −4%ΔBM) would impair endurance performance compared with maintenance of preexercise BM (0%ΔBM).
MATERIALS AND METHODS
Eight very active subjects (4 men, 4 women; 22–36 yr) volunteered to participate in this study. Subjects were informed of the experimental procedures and associated risks before providing written, informed consent. This study was approved by the Institutional Review Board for the Protection of Human Subjects at the Pennsylvania State University. Preliminary screening included a physical exam and a graded-exercise test on a treadmill to determine maximal oxygen uptake (V̇o2max). Criteria for the subjects' inclusion were V̇o2max ≥50 ml·kg−1·min−1 for men and ≥45 ml·kg−1·min−1 for women, running ≥32 km per wk, and not currently taking medications or oral supplements that could interfere with study results. All women were eumenorrheic with regular cycles (natural, n = 2; oral contraceptive users, n = 2) and were tested within 7 days after the onset of menstruation. No subjects reported having experienced EAH previously. Sample characteristics are presented in Table 1.
All subjects had been engaged in ≥12 wk of regular running before participation in the study and maintained a consistent training schedule until completion of all experimental trials. Subjects were instructed to eat their typical prerace diet the evening before and to abstain from heavy exercise, alcohol, and caffeine at least 24 h before each trial. Diet logs were kept by the subjects to facilitate consistent food and fluid consumption for 24 h before each trial. Subjects reported to the laboratory at 0700 on the morning of test days after an overnight fast. Immediately upon arrival, a blood sample was obtained to confirm normal baseline hydration and Na+ status. Subjects were considered euhydrated when serum osmolality was <290 mosmol/l (22) and eunatremic when S[Na+] was 135–145 mmol/l.
There were 12 experimental trials (3 beverages at 4 different %ΔBM). In separate trials, subjects drank fluid or no fluid to 1) maintain BM (0%), 2) increase BM by 2%, 3) decrease BM by 2%, or 4) decrease BM by 4%. The beverages were 1) a 6% carbohydrate solution without Na+ (Na+0), 2) a 6% carbohydrate solution with 18 mmol/l Na+ (Na+18), and 3) 6% carbohydrate solution with 30 mmol/l Na+ (Na+30). Experimental trials were scheduled at least 1 wk apart and were assigned in random order. Both the subject and investigator were blinded to the beverage consumed during the trials.
The night before each experiment, subjects swallowed an ingestible temperature sensor for the measurement of body core temperature (Tc). On test days, subjects had an 18-gauge Teflon catheter placed in an antecubital vein, voided their bladder, and then entered an environmental chamber set at 30°C and 40% relative humidity. Next, the subject sat quietly for 30 min before the baseline heart rate (HR), blood pressure (BP), Tc, and blood sample were obtained. Next, the subject's initial BM was measured to the nearest 0.05 kg. All BM measurements during the experiment were taken with the subject wearing lightweight running shorts, sport bra (women), thin socks, and running shoes. Next, the subject ran for seven 15-min bouts (70% V̇o2max) each separated by 2 min of rest (2 h of interval running total). Twelve minutes into each running bout, HR, BP, Tc, rating of perceived exertion (RPE), and a blood sample were obtained. Urine samples were collected during rest periods as needed so that ΔBM measurements did not include bladder volume. The criteria for terminating a trial before the planned 2 h were S[Na+] ≤ 132 mmol/l or Tc > 39°C.
During each rest period, the subjects were toweled off and then had their BM measured. During the 0%ΔBM trials, subjects drank fluid (either Na+0, Na+18, or Na+30) volumes during the rest periods to maintain their initial BM. During the −2 and −4%ΔBM trials, fluid was restricted until the subjects reached their target BM. If the subjects' BM fell below their target BM, they ingested enough fluid (either Na+0, Na+18, or Na+30) to maintain the desired %ΔBM. During the +2%ΔBM trials, subjects drank the necessary fluid (either Na+0, Na+18, or Na+30) volumes to gain 2% of their BM by the end of the 2 h of interval running. The 2% gain in BM was titrated over the 2-h interval running protocol so that BM gain was achieved gradually (to maximize fluid retention and minimize gastrointestinal discomfort).
At the end of the 2-h interval running protocol, subjects voided and then had their BM measured. Next, subjects drank the appropriate volume of Na+0, Na+18, or Na+30 or drank no fluid to maintain the desired %ΔBM. The performance run commenced 15 min after the end of the final interval running bout. During the performance test subjects ran at a speed corresponding to 85% V̇o2max until exhaustion. Subjects were instructed to run until volitional fatigue and received a monetary incentive. No fluid was consumed and no verbal encouragement or feedback on run time or distance was given to subjects during the performance run. Time to exhaustion was recorded to the nearest second. HR, BP, Tc, and RPE were recorded at the end of the performance run. Seven subjects repeated one trial to determine the repeatability of the performance test. The beverage and %ΔBM of repeat trials were selected at random and the subjects were blinded as to which trial they were repeating.
During the first 10 min of recovery, subjects walked at 2.5 mph for a gradual cooldown. Next, subjects sat quietly in the environmental chamber (30°C and 40% relative humidity) for a 50-min recovery period to allow fluid compartments to stabilize and capillary filtration pressure to return to resting values. This was important because the Nguyen-Kurtz equation was derived from measurements made at rest. The subjects' BM was measured at 10 min, 30 min, and end of the recovery period. Subjects drank fluid (either Na+0, Na+18, or Na+30) or no fluid during the 50-min recovery period to maintain the desired %ΔBM. Urine samples were collected at the beginning (as needed) and at the end of the 50-min recovery period. During exercise and recovery, fans were placed around the subject to promote evaporation of sweat and minimize the amount of sweat trapped in their clothing and shoes.
Sterile sweat patches (PharmChem) were placed on the forehead, forearm, scapula, upper chest, and anterior thigh of subjects during the second rest period. The subjects' skin was cleaned with an alcohol swab and dried before the sweat patches were applied. The patches were removed when an adequate sample was obtained. The patches were then placed in air-tight plastic tubes (Sarstedt Salivette) and then centrifuged at 4°C for 15 min. The sweat samples were aliquoted into cryovials and refrigerated until analysis.
Blood, urine, and sweat analysis.
Venous blood samples (9 ml each) were drawn without stasis. A 2-ml aliquot was transferred into an EDTA-treated test tube and immediately analyzed for hematocrit (microhematocrit centrifugation) and hemoglobin (Hemacue Hb 201+). The remaining 7-ml aliquot was transferred into a serum separator tube, allowed 30–60 min to clot, and then centrifuged at 4°C for 15 min. All of the following measurements were made in triplicate. Serum was analyzed for [Na+], [K+] (S[K+]), and osmolality (Sosm; freezing point depression, Advanced DigiMatic Osmometer model 3D2). Urine samples were analyzed for volume (Uvol), [Na+] (U[Na+]), and [K+] (U[K+]). Sweat was analyzed for [Na+] (Sw[Na+]) and [K+] (Sw[K+]). Serum [Na+] and [K+] were measured by the ion-specific electrode method (Diamond Diagnostics, Ciba Corning 614), and urine and sweat [Na+] and [K+] were measured via flame photometry (Instrumentation Laboratory model IL943). The two methods for [Na+] and [K+] analysis were compared in a subset of serum samples, and the coefficient of variation (CV) between methods was 0.8 and 2.5% for [Na+] and [K+], respectively.
HR was measured by use of a Polar monitor, and BP was measured by brachial auscultation (sphygmomanometry). RPE was assessed by the Borg scale (3). BM was measured to the nearest 0.05 kg via a Seca 770 scale. A CorTemp Disposable Temperature Sensor (COR-100) and CorTemp Recorder (CT-2000) were used to measure Tc.
Mean arterial pressure (MAP) was calculated as MAP = [1/3] pulse pressure + diastolic BP. The percent change in plasma volume (PV) from baseline (ΔPV) was calculated from hematocrit and hemoglobin (5). Volume of sweat loss (Swvol) was calculated from ΔBM corrected for fluid consumed and urine excreted. Total sweat Na+ and K+ loss were calculated from Sw[Na+] and Sw[K+] and Swvol. Total urine Na+ and K+ loss were calculated from U[Na+] and U[K+] and Uvol. The net change in total Na+ and K+ (ΔE) was calculated as [Na+ + K+ intake] − [(Na+ and K+ sweat loss) + (Na+ and K+ urine loss)].
Predicted postexperiment S[Na+] was calculated according to Kurtz and Nguyen (10): where S[Na+]i was initial serum [Na+], TBWi was initial total body water (TBW) (assume = 0.73 of fat-free mass, Ref. 20), and ΔTBW was the pre- to postexperiment change in TBW (assume = ΔBM).
Sweat [Na+] and [K+] were measured in samples collected from all five sites. However, the [Na+] and [K+] reported and used for all calculations in the present study were from the forearm and chest, respectively, since these regional sites have been reported to be the most highly correlated with whole body sweat [Na+] and [K+] (21).
The survey was administered immediately after the 2-h interval running period and at the end of the 50-min recovery period. The survey consisted of 100-point visual analog rating scales that assessed lightheadedness, windedness, stomach bloating, sloshing of stomach contents, stomach upset, side stitch/ache, total body fatigue, muscle cramping, and the saltiness of the beverages. Subjective ratings were assessed by having subjects place a pen mark on the 100-point scales between the extreme answers at the opposite ends of the line [ranging from “none” (0) to “very” or “severe” (100)] to represent their perceived intensity of the attribute (see Ref. 13 for a more detailed description of linear rating scales).
Fluid and sodium balance data, physiological variables, time to exhaustion, and subjective ratings were analyzed by two-way ANOVA (beverage vs. %ΔBM) with repeated measures. The Tukey honestly significant difference test was performed when main effects were found. PROC MIXED in SAS 9.1 was used to perform the ANOVA tests. An intraclass correlation coefficient was used to test the reliability of performance between repeated trials and the reliability between predicted and measured postexperiment S[Na+] (23). The slope and intercept of the regression line and line of identity of the scatterplot for predicted vs. measured S[Na+] were compared to determine the accuracy of the Nguyen-Kurtz equation in predicting postexperiment S[Na+]. A paired t-test was also used to determine whether there was a significant difference between predicted and measured postexperiment S[Na+]. The significance level for all statistical tests was set to alpha = 0.05. All data are presented as means ± SE.
Fluid and sodium balance.
The measured net %ΔBM and net ΔE for each trial are presented in Table 2. The measured %ΔBM was similar to the target %ΔBM in all except the −4%ΔBM trials. Five of the eight subjects did not have sweating rates high enough to reach −4%ΔBM. Their actual ΔBM after total fluid restriction was −2.9 ± 0.1%. Because these five subjects (4 women, 1 man) did not consume any fluid during this trial, a beverage Na+ comparison at −4%ΔBM for these subjects was not possible. Thus these five subjects were only asked to complete one −4%ΔBM trial. A total of 86 trials were completed (24 trials at +2%ΔBM, 24 trials at 0%ΔBM, 24 trials at −2%ΔBM, and 14 trials at −4%ΔBM), all of which were included in the statistical analyses. The target −4%ΔBM trials that were not completed (10 total) were treated as missing data in the analyses (when computing ANOVAs, PROC MIXED still analyzes data from all subjects, even if some have missing data). The n for each trial is provided in all tables and figures. Also, it should be noted that, for simplicity, all trials will hereafter be referred to as the target %ΔBM (as indicated in all tables and figures), not the measured %ΔBM.
Fluid, Na+, K+, and carbohydrate intake during each trial are presented in Table 3. As predicted, subjects consumed significantly more fluid and carbohydrate during +2% vs. 0% vs. −2% vs. −4%ΔBM, respectively. Na+ intake was significantly higher during Na+30 vs. Na+18 during the +2% and 0%ΔBM trials.
Fluid, Na+, and K+ loss during each trial are presented in Table 4. There were no significant differences among beverages within each %ΔBM. Total Uvol was significantly higher during +2% vs. 0% (P < 0.0001), −2% (P < 0.0001), and −4% (P < 0.0001) ΔBM trials. There were no statistical differences in urine Na+ loss, sweat volume, or sweat Na+ loss among trials. No subject became hyponatremic during the trials (the lowest measured S[Na+] was 136 mmol/l).
Predicted vs. measured S[Na+].
The relation between predicted and measured postexperiment S[Na+] are presented in Fig. 1. When +2%, 0%, −2%, and −4%ΔBM were all included in the analysis (left), the slope (0.71) of the regression line was significantly different from one and the intercept (42) was significantly different from zero. Additionally, the paired t-test results showed a significant difference between predicted and measured S[Na+] (P = 0.00) when all trials were included in the analysis. However, when the +2%ΔBM trials were excluded from the analysis (right), the slope (0.84) and intercept (23) of the regression line were not significantly different from one and zero, respectively (additionally, paired t-test P = 0.42). Furthermore, the intraclass correlation coefficient between predicted and measured S[Na+] was 0.90. The mean difference between predicted and measured postexperiment S[Na+] was −0.8 ± 0.2 for all trials and −0.2 ± 0.2 when the +2%ΔBM trials were excluded.
There were no differences in baseline S[Na+] and Sosm among trials. The mean S[Na+] and Sosm at baseline were 142.4 ± 0.2 and 285.5 ± 0.3, respectively. The ΔS[Na+], ΔSosm, and %ΔPV from pre- to postexperiment are presented in Fig. 2. During the +2%ΔBM trials, the decrease in Sosm was significantly smaller with Na+18 and Na+30 vs. Na+0.
Table 5 presents Tc, HR, MAP, and RPE results at the end of 2-h interval running and the 50-min recovery period for each trial. There were no significant differences among trials at baseline. Preexperiment resting Tc, HR, and MAP were 37.09 ± 0.03, 58 ± 1, and 84 ± 1, respectively. RPE was 8.9 ± 0.2 at the end of the first running bout. Significant main effects of %ΔBM at the end of exercise and the 50-min recovery period are shown in Table 5. There were no significant differences in Tc, HR, MAP, or RPE among beverages at the end of exercise or the 50-min recovery period.
Time to exhaustion and repeatability of the performance run are presented in Fig. 3. There were no significant differences among beverages (i.e., no effect of beverage [Na+]) so the average performance times of Na+0, Na+18, and Na+30 were calculated and presented. Time to exhaustion was significantly shorter during the −2% vs. 0% (P = 0.01) and +2% (P = 0.0005) ΔBM trials and during the −4% vs. 0% (P = 0.002) and +2% (P = 0.0001) ΔBM trials. There was no statistical performance difference between −2 and −4%ΔBM (P = 0.56) or between 0 and +2%ΔBM (P = 0.49) trials. The intraclass correlation coefficient between repeated trials was 0.96 and the CV between repeated trials was 10 ± 2%.
Subjects' responses on the 100-point visual analog rating scales are presented in Table 6. At the end of the 2-h interval running period, subjects felt significantly more lightheadedness, windedness, and total body fatigue during the −2% vs. the +2% (P = 0.03, P = 0.04, and P = 0.01, respectively) and 0% (P = 0.02, P = 0.01, and P = 0.02, respectively) ΔBM trials and during the −4% vs. the +2% (P < 0.0001, P = 0.0002, and P = 0.001, respectively) and 0% (P < 0.0001, P = 0.0001, and P = 0.003, respectively) ΔBM trials. There was also a significant difference between the −2% and −4%ΔBM trials for ratings of lightheadedness (P = 0.03) and windedness (P = 0.01). Additionally, subjects rated their muscle cramping higher during the −4% vs. the +2% (P = 0.04) and 0% (P = 0.04) ΔBM trials. At the end of recovery, subjects felt significantly more lightheadedness (P = 0.006 and P = 0.006, respectively) and total body fatigue (P = 0.02 and P = 0.04, respectively) during the −4% vs. the 0% and +2%ΔBM trials.
At the end of the 2-h interval running, subjects rated their level of stomach bloating (P < 0.0001, P < 0.0001, and P < 0.0001, respectively) and sloshing of stomach contents (P = 0.003, P = 0.0009, and P = 0.0009, respectively) higher during the +2%ΔBM vs. the 0%, −2%, and −4%, ΔBM trials. At the end of the recovery period, subjects felt significantly more bloated during the +2%ΔBM vs. the 0% (P < 0.0001), −2% (P < 0.0001), and −4% (P = 0.0002) ΔBM trials.
There were no statistically significant differences among beverages (i.e., no effect of beverage [Na+]) within %ΔBM levels for any of the subjective ratings at the end of 2-h interval running or recovery. Likewise, there were no statistical differences among beverages in the subjects' ratings of beverage saltiness at the end of the 2-h interval running (31 ± 5, 33 ± 5, 32 ± 4 during the Na+0, Na+18, and Na+30 trials, respectively) or at the end of recovery (28 ± 5, 33 ± 5, 34 ± 5 during the Na+0, Na+18, and Na+30 trials, respectively).
The main findings from this study were 1) the Nguyen-Kurtz equation accurately predicted the measured postexperiment S[Na+] during the 0%, −2%, and −4%ΔBM trials, but not the +2%ΔBM trials, 2) Na+ consumption attenuated the decline in S[Na+] from pre- to postexperiment during the 0% and +2%ΔBM trials, but the differences among beverages Na+0, Na+18, and Na+30 were not statistically significant, and 3) prolonged running performance was impaired when subjects incurred a 2% and 4% BM deficit due to fluid restriction.
Predicted vs. measured S[Na+].
The results confirm the predictions of the Nguyen-Kurtz equation (10) when subjects drink to match sweating rate (0%ΔBM) or restrict fluid consumption and lose BM (−2% or −4%) during endurance exercise. These results support the notion that changes in S[Na+] can be predicted by changes in the net mass balance of fluid, Na+, and K+ from pre- to postexercise. As indicated in Fig. 2, the pre- to postexercise ΔS[Na+] in the present study was most sensitive to the ΔBM (i.e., fluid balance). As predicted, drinking any of the fluids in the present study at a rate greater than sweating rate (+2%ΔBM trials) leads to dilution of S[Na+] and restricting fluid intake (a decrease in BM) leads to an increase in S[Na+] even in the presence of a substantial sweat Na+ loss. Moreover, Na+ consumption influences the relation between ΔBM and the ΔS[Na+]. As indicated in Fig. 2, Na+ consumption (i.e., 0% and +2%ΔBM trials) attenuated the decline in S[Na+], and the higher the [Na+] in the beverage, the greater the attenuation.
Because the accuracy of the Nguyen-Kurtz equation was confirmed in the present study, it is logical to conclude that the assumption made in calculating predicted S[Na+], i.e., that ΔTBW = ΔBM, is a valid one. However, Noakes et al. (19) contend that, because of 0.7-kg loss from fuel oxidation, 0.4-kg gain from metabolic water production, and 1.5-kg gain from water released with glycogen utilization, 70-kg endurance athletes can lose ≥3% of their BM over the course of a marathon without experiencing a change in TBW. If this assessment was correct, then using ΔBM as a surrogate for ΔTBW in the Nguyen-Kurtz equation would cause the measured decrease in S[Na+] to be larger (more negative) than that predicted for the 0%, −2%, and −4%ΔBM trials. Noakes et al.'s theory was not supported in the present study, since there was no significant difference between predicted and measured S[Na+] during the 0%, −2%, and −4%ΔBM trials. Although it is possible that pre- to postexercise ΔBM overestimates sweat losses to some extent (because of endogenous water production and/or weight loss from the oxidation of glycogen and fatty acids), it is apparently not enough to affect S[Na+]. Similar to Noakes et al.'s view, Hew (8) suggested that runners who finished a marathon in a 3-kg BM deficit were in a state of euhydration, not dehydration (deficit in body water). Hew came to this conclusion after conducting a retrospective analysis of pre- and postrace measurements of BM and S[Na+] in runners who participated in the Houston Marathon. In Hew's analysis, a scatterplot of ΔBM vs. ΔS[Na+] illustrates that a 3-kg loss in BM corresponded to a zero ΔS[Na+] from pre- to postrace and that a zero kg ΔBM corresponded to a 6 mmol/l decrease in S[Na+]. Hew interpreted these data as evidence that 3 kg of endogenous water production caused the 6 mmol/l decrease in S[Na+] despite a zero ΔBM. However, Hew did not consider the impact of Na+ intake and loss on postrace S[Na+]. In the present study, a 0%ΔBM also corresponded to a decrease in S[Na+] (−2 mmol/l); however, the decrease in S[Na+] is associated with a Na+ deficit (ΔE = −140 mmol/l, Table 2) since runners consumed a Na+-free beverage to replace sweat losses. It is possible that the runners in Hew's analysis also incurred a Na+ deficit during the marathon, which would account, at least in part, for the 6 mmol/l decrease in S[Na+]. However, it is difficult to draw any conclusions from Hew's analysis because the runners' sweating rate, sweat [Na+], Na+ intake, and actual fluid intake were not measured.
When subjects overdrank relative to their sweat losses (+2%ΔBM trials), the Nguyen-Kurtz equation was not accurate. During the +2%ΔBM trials, the S[Na+] predicted by the Nguyen-Kurtz equation was significantly less than the measured S[Na+]. The mean difference between predicted and measured S[Na+] was −2.6 ± 0.5 mmol/l. The subjects' renal systems were effective in excreting excess fluid, as indicated by the significantly larger urine volumes collected during the +2%ΔBM trials (Table 4). Because of the high rate of urine excretion during the +2%ΔBM trials, subjects drank a substantial volume of fluid (∼900 ml) during the 50-min recovery period to compensate for urine losses and maintain +2%ΔBM. It is possible that fluid absorption was not complete and a portion of the ingested fluid volume remained in the subjects' stomachs at the time of postexperiment BM measurements. Accordingly, at the end of the 50-min recovery period, subjects rated their level of stomach bloating and sloshing of stomach contents significantly higher during the +2%ΔBM vs. the 0%, −2%, and −4%, ΔBM trials. Consequently, measured S[Na+] was not as diluted as would be predicted by the ΔTBW in the Nguyen-Kurtz equation.
It is interesting to compare the measured S[Na+] results of the present study with that predicted by Montain et al. (14), who used a mathematical model and theoretical conditions to predict S[Na+]. In Montain et al.'s model, the environmental conditions, running speeding, body composition, and sweat S[Na+] were systematically varied. The scenario that is most comparable with the experimental conditions and subjects' physical characteristics of the present study is illustrated in Fig. 1 of Montain et al. (14). In this figure, the theoretical ambient temperature was 28°C and the athlete weighed 70 kg (63% of which was water), had a sweat [Na+] of 50 mmol/l, and was running at 10 km/h. By comparison, in the present study the ambient temperature was 30°C and, on average, the subjects weighed 66 kg, had a sweat [Na+] of 56 mmol/l, and ran at 10.3 km/h. In Montain et al.'s theoretical example (14), the athlete consumed 800 ml of water per hour to maintain BM. After 2 h of running, the athlete's S[Na+] decreased from 140 to ∼139 mmol/l. In the present study, when subjects drank 870 ml of the Na+-free beverage to maintain BM, S[Na+] decreased from 141.6 to 139.6 mmol/l. Thus the ΔS[Na+] predicted by the mathematical model was similar to the ΔS[Na+] measured in the present study. In Fig. 4 of Ref. 14, Montain et al. illustrated the theoretical effect of CES (17 mmol/l Na+ and 5 mmol/l K+) consumption on S[Na+] under the same conditions described previously. Montain et al.'s model demonstrated that fluid replacement with CES attenuated the dilution of S[Na+]. However, the difference in S[Na+] between CES and water after 2 h of exercise was only ∼1 mmol/l. Similarly, the difference between Na+18 and Na+0 during the 0% BM trials was 1.2 mmol/l in the present study.
Although Na+ ingestion attenuated the decrease in S[Na+] during the 0% and +2%ΔBM trials (Fig. 2), the difference in the decrease in S[Na+] among beverages did not reach statistical significance. Two studies have been able to demonstrate a significant effect of beverage Na+ on the ΔS[Na+] during exercise. In a study by Vrijens and Rehrer (26), the measured rate of ΔS[Na+] was significantly greater (more negative) with water than with a CES containing 18 mmol/l Na+ (−2.5 vs. −0.9 mmol·l−1·h−1) during 3 h of continuous cycling. Twerenbold et al. (25) measured the ΔS[Na+] after 4 h of running in women who consumed equal volumes of either water, CES with 410 mg/l Na+ (∼17 mmol/l Na+), or CES with 680 mg/l Na+ (∼28 mmol/l Na+), in separate trials. The subjects finished the trials in positive fluid balance (+2%ΔBM), but the decrease in S[Na+] was significantly greater when runners drank water compared with the CES with 680 mg/l (−6.2 vs. −2.5 mmol/l). The difference in the results of these two studies and the present study may, in part, be explained by the longer exercise times in the Vrijens and Rehrer and Twerenbold et al. studies (3–4 h vs. ∼2 h in the present study). Their findings support the notion that Na+ ingestion becomes even more critical as the duration of exercise increases.
During the 0%ΔBM trials, the measured S[Na+] was similar to that predicted by the Nguyen-Kurtz equation when subjects consumed Na+0 (−2.0 vs. −2.7 mmol/l), Na+18 (−0.8 vs. −1.6 mmol/l), and Na+30 (−0.5 vs. −0.6 mmol/l). In the present study, to maintain preexperiment S[Na+] while also consuming enough fluid to replace sweat losses and maintain BM, a higher [Na+] in the CES would have been required. According to the mass balance calculations, the subjects' average ΔE due to sweat Na+ and K+ losses was −70 mmol·l−1·h−1 and sweating rate was 1.2 l/h. Thus a CES with a combined [Na+] and [K+] of 58 mmol/l (i.e., 70 mmol/l divided by 1.2 l/h) would be required to maintain preexperiment S[Na+], when consumed at a rate equal to sweating rate. That is, to maintain Na+ balance, the CES should be similar in composition to that of the athlete's sweat (the mean sweat [Na+] in the present study was 56 mmol/l). This point illustrates the influence of sweat Na+ losses on the ΔS[Na+] and the importance of individualized fluid and electrolyte replacement programs for endurance athletes.
Because only four men and four women were tested in the present study, it would be difficult to make any firm conclusions regarding sex differences. However, it is important to note that there were no indications of sex-related differences in the baseline S[Na+], pre- to postexperiment ΔS[Na+], or fluid retention, nor in the relation between predicted and measured S[Na+].
It has been shown in clinical settings that significant hyperglycemia can induce dilutional hyponatremia by promoting the osmotic shift of water from the intracellular to the extracellular space (10). Although each beverage used in the present study was composed of the same 6% carbohydrate solution, the volume of fluid ingested and therefore the quantity of carbohydrate consumed varied among trials and subjects. Blood glucose concentration was not measured; however, given that the subjects were healthy individuals, it is highly unlikely that significant hyperglycemia developed. Therefore, it is also unlikely that differences in carbohydrate ingestion among trials affected the ΔS[Na+].
Although ΔBM clearly made a difference, there was no effect of beverage [Na+] on the time to exhaustion during the performance run in the present study (data not shown). There have been mixed results in the literature regarding the effect Na+ intake and S[Na+] on endurance performance. For example, performance was not affected by plasma [Na+] or the rate of change in plasma [Na+] in female endurance athletes running for a period of 4 h in various environmental conditions (25). Conversely, Vrijens and Rehrer (26) found that a high rate of change (decrease) in plasma [Na+] was correlated with a decreased time to exhaustion during 3 h of cycling. Furthermore, preexercise Na+ loading (with 164 mmol Na+/l beverage) is associated with improved cycling time trial performance compared with preexercise consumption of an equal volume of a no or low-Na+ (10 mmol/l) beverage (4, 24). One factor that probably contributed to the improved performance with Na+ ingestion in the preexercise loading studies was the concomitant improved maintenance of PV (4, 24). In these studies, increased performance is likely mediated by improved cardiovascular and/or thermoregulatory function conferred by the maintenance of PV (6, 7, 16, 17). In the present study, there were no significant differences in the %ΔPV among beverages (Fig. 2), which may explain the lack of a beverage effect on endurance running performance.
A 2% and 4% BM deficit was associated with a significantly decreased time to exhaustion in the present study. The impaired performance could be attributed to the significantly higher Tc and greater cardiovascular strain (as indicated by higher HR) after the 2-h interval running period when subjects lost BM vs. when they consumed enough fluid to either maintain or increase BM by 2%. Likewise, after the 2-h interval running period, subjects rated their perceived exertion, lightheadedness, windedness, and total body fatigue significantly higher during the −2% and −4%ΔBM trials vs. the +2% and 0%ΔBM trials. It is also possible that the performance differences among levels of %ΔBM could be caused by differences in the amount of carbohydrate consumed. Although each beverage consisted of the same carbohydrate concentration (6%), subjects consumed more total carbohydrate as total fluid volume intake increased (i.e., +2% > 0% > −2% > −4%ΔBM trials; Table 3). Thus the subjects' run time to exhaustion was likely influenced by carbohydrate availability. Below et al. (2) suggest that fluid and carbohydrate ingestion have independent and additive beneficial effects on endurance performance, which may have been the case in the present study.
It is also important to note that there was no difference in time to exhaustion, sweating rate, or Tc between 0% and +2%ΔBM trials in the present study. These results are consistent with those of Latzka et al. (12), who demonstrated that hyperhydration via water or glycerol (∼1.5-liter increase in TBW) provided no performance or thermoregulatory advantage compared with the maintenance of euhydration during 2 h of compensable exercise-heat stress.
EAH tends to occur more commonly in running than cycling endurance events. Therefore, since the primary aim was to determine the effects of Na+ intake and ΔBM on the development of EAH, treadmill running was the mode of endurance exercise used in the present study. To test the effects of experimental manipulations on endurance performance, one could administer either a time-to-exhaustion or time-trial exercise test. Time-trial tests are thought to be more meaningful performance tests owing, in part, to the lower variability associated with time-trial vs. time-to-exhaustion protocols (9, 11). A time trial is practical for a cycling test in which the subject can simply control the pace by adjusting the revolutions per minute. However, a time-to-exhaustion test was used in the present study because it would have been difficult (and distracting) for subjects to control their own treadmill speed for a time trial. Although the CV for time to exhaustion in the present study (10%) was higher than the typical CV for time trials (2–3%, Ref. 9 and 11), the intraclass correlation coefficient (0.96) indicated excellent reliability between repeated time-to-exhaustion tests (23), supporting its use as a meaningful performance test. Moreover, the 37 and 63% decrease in time to exhaustion (compared to 0%ΔBM trials) during the −2%ΔBM and −4%ΔBM trials, respectively, far outweighed the 10% variability between repeated tests.
Limitations and future directions.
During the −2%ΔBM trials fluid was restricted until the subjects' target BM was reached. Perhaps a more realistic simulation would have been to provide fluid intake at more even allotments throughout the 2-h interval running protocol (i.e., smaller volumes given during each rest period rather than a larger volume toward the end of the 2-h running). A practice trial in the heat to determine each subject's sweating rate would have facilitated a titrated fluid intake procedure. However, because of the time commitment already involved with experiment participation (2 screening days and up to 13 experimental trials), a practice trial was not conducted. Regardless, it was more important to ensure that subjects would achieve −2%ΔBM during the trial than attempt to titrate fluid intake.
The results of this study showed that the ΔS[Na+] from pre- to postexercise can be predicted simply from the mass balance of fluid, Na+, and K+. However, it is important to note that these results are only applicable to individuals who have no history of EAH (because subjects in the present study had no history of EAH). In future studies, it would be important to test the accuracy of the Nguyen-Kurtz equation in athletes who have a history of EAH to determine whether the mass balance model still holds true.
Summary and practical recommendations.
In summary, the main findings from this study were 1) the Nguyen-Kurtz equation accurately predicts postexperiment S[Na+] when subjects drink to match sweating rate (0%ΔBM) or restrict fluid consumption and lose BM (−2% or −4%) during endurance exercise, 2) the Nguyen-Kurtz equation does not accurately predict S[Na+] when athletes overdrink relative to their sweat loss and increase their BM (+2%), 3) compared with Na+-free beverages, consumption of beverages with Na+ attenuates the decline in S[Na+] from pre- to postexercise, and 4) prolonged running performance is impaired when subjects incur a 2% and 4% BM deficit due to fluid restriction. It is clear that both the volume and [Na+] of fluid consumed during exercise has implications for sodium balance and that a ≥2% BM deficit impairs endurance performance. Therefore, the present study results suggest that the optimal hydration practice for endurance athletes is to consume fluids at a similar rate (to avoid BM gain and ≥2% BM loss) and similar composition to that of their sweat losses.
Support for this study was provided by the Gatorade Sports Science Institute and the General Clinical Research Center Grant MO1 RR-010732.
The authors are grateful to the subjects for participation in this study. Additionally, we thank Mosuk Chow for statistical consultation; Jane Pierzga, John Jennings, Matt Kenney, Jose Flores, Ben Miller, Doug Johnson, and Randy McCullough for technical assistance; and the General Clinical Research Center nursing staff for medical support.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society