Large reductions in serum sodium concentration have been observed in military personnel and athletes during prolonged exercise (3,5,8,9). Because serum sodium concentration reflects the concentration in the extracellular fluid space, reductions in serum sodium concentration may result from decreasing the sodium content or increasing the water volume in the extracellular fluid via excessive fluid consumption (7-9,22).
Although it is clear that reductions in serum sodium concentration can occur via hyperhydration (14,17), they can also occur during prolonged exercise with heavy sweating when the ingested fluid replaces substantially less sodium than that which is lost in sweat (22). Vrijens and Rehrer (22) reported that serum sodium concentration declined in subjects dehydrated by approximately 1% of body weight while ingesting sodium-free fluid during 3 h of cycling at 55% V˙O2max. However, a decline in serum sodium concentration due to sweat sodium losses might only occur in a segment of the population with high sweat sodium losses. Several studies have reported large interindividual variability in sweat sodium losses in small samples of subjects (11,13,21). Thus, subjects that exhibit high sweat sodium losses should be more likely to lower serum sodium concentration via sweating if they do not replace those losses during exercise compared with those who lose small amounts of sodium through sweating. The purpose of this study was to determine the factors associated with changes in serum sodium concentration during an ultradistance triathlon.
Forty-six healthy male (n = 26, age = 44.5 ± 11.4 yr, weight = 77.0 ± 10.3 kg) and female (n = 20, age = 41.8 ± 9.9 yr, weight = 61.8 ± 5.6 kg) athletes were recruited from among the 1700 men and women who participated in the 2003 Hawaii Ironman® Triathlon World Championship in Kailua-Kona, Hawaii (Table 1). All subjects were required to exercise a minimum of three consecutive days in an environment hotter than 30°C before the study in addition to their intense aerobic training performed during the preceding weeks. Subjects signed a consent form approved by the institutional review board at The University of Texas at Austin.
Sweating analysis trial experimental protocol.
During the 3- to 7-d period before the race, all subjects participated in a sweating analysis trial in an outdoor field laboratory in Kailua-Kona, Hawaii, in warm conditions (26.4°C ± 1.7°C wet bulb globe temperature (WBGT) outdoor, 28.3°C ± 1.2°C dry bulb, 65.6% ± 3.8% relative humidity [RH]). Upon arrival, a 5-mL blood sample was taken, and each subject's personal bicycle was mounted on a stationary trainer. After an 8-min warm-up, the subjects cycled for 30 min at 70%-75% of self-reported HRmax. This intensity was chosen to mimic reported race intensity during an Ironman® triathlon (12). If HRmax was unknown, a predictive equation (208 − 0.7 × age) was used (19). Body mass was measured (A & D Medical, Milipitas, CA) immediately after the 8-min warm-up and again after the 30-min exercise bout for calculation of sweating rate.
A waterproof "sweat patch" then, composed of a 7.6 × 7.6-cm gauze sponge (Johnson & Johnson Medical, Arlington, TX) and a Tegaderm® bandage (3M Health Care, St Paul, MN), was applied to the mid-posterior right forearm and right scapula for regional sweat collection. These regional sites were chosen because they are highly correlated with whole body sweat sodium concentration (0.82 vs 0.88, forearm and scapula, respectively) (11). Before application of the patch, the area was cleansed with 70% isopropyl alcohol and rinsed with distilled water. When necessary, the right forearm or back was gently shaved before patch application. The patches were applied after an 8-min warm-up and initiation of sweating. During the patch application, the subjects remained seated on the bicycle, and the patches were removed immediately after the 30-min exercise bout. This duration was chosen so that the patches would not become saturated, which may alter sweat sodium concentration (23).
During the exercise bout, the subjects were fan cooled (4.1 ± 0.8 m·s−1) and wore shorts and a sports bra (women). The subjects were given 400 mL of a 6% carbohydrate-electrolyte beverage (Gatorade, Quaker Oats Company, Barrington, IL) to consume during the 30-min test.
Race day protocol.
Within 2 h of the race starting time, a subset of the subjects (n = 20) reported to the field laboratory adjacent to the race starting line. The prerace body mass and a 5-mL blood sample via venipuncture were taken. Body mass and serum sodium concentration for this subset were similar to that measured before the sweating analysis trial (71.6 ± 14.9 vs 71.6 ± 15.1 kg, 146.5 ± 3.2 vs 145.7 ± 1.5 mEq·L−1, P > 0.05, respectively). Therefore, baseline body mass and serum sodium for all subjects (N = 46) are defined either as the prerace measures (subset; n = 20) or as the presweating analysis trial measures (n = 26). Environmental conditions for the race were similar to those during the sweat test: 27.6°C ± 0.8°C DB and 69.5% ± 2.8% RH. Upon completing the race, all subjects (N = 46) were escorted from the finish area to the field laboratory. The body weight and a 5-mL blood sample were taken within 5-15 min of finishing the race. Fluid and food intake during the race were self-reported by the participants to the researchers at the 56-mile point of the bike segment, after the bike segment, and/or at the finish area (subset; n = 20). Subjects were free to consume any race-provided or self-provided food and fluid during the race. Food, fluid, or sodium supplementation was not provided or was limited by the research staff. Total sodium ingestion was the sum of sodium intake from pills plus sodium ingested from foods and fluid. Fluid intake was the sum of all fluid ingested during the event.
Blood sampling and analysis.
Blood samples were collected in glass test tubes that contained no additives and were placed on ice. After clotting occurred, serum was separated by centrifugation for 20 min. Serum was extracted into 5-mL plastic test tubes, and serum sodium concentration was determined by chemical analysis (Nova 8, Waltham, MA).
Sweating rate and sweat sodium concentration analysis.
Upon removal of the sweat patches, the gauze sponge was immediately separated from each Tegaderm® bandage and placed into plastic syringes. The sweat content of the sponges was squeezed by the plunger into two 5-mL plastic test tubes and then capped. Sweat electrolyte concentration was determined with a Nova 5 analyzer with a coefficient of variation of 2% for sodium concentration. Sweating rate was calculated as the change in body mass during the 30-min exercise bout. Mass change due to respiratory and metabolic losses was not included in this calculation because they are minimal during 30 min of exercise. Corrections were made for fluid consumption and sweat contained in clothing worn during the test.
Mean, SD, and coefficient of variation (SD/mean) were calculated for descriptive purposes. Multiple regression analysis was used to determine the contributions of variables to changes in race day serum sodium concentrations. A two-tailed Student's t-test was used to compare prerace-postrace and sex differences. Pearson product moment correlations were used for comparisons. Statistical significance was defined as P < 0.05.
Sweating analysis trial.
Serum sodium concentration at rest before the sweating analysis trial averaged 145.5 ± 2.2 mEq·L−1 (N = 46) (Table 1). Men had a significantly higher absolute sweating rate than women (1.53 ± 0.36 vs 1.11 ± 0.38 L·h−1, respectively, P = 0.001), whereas sweating rate per kilogram of body mass was not significantly different (20.15 ± 5.15 vs 17.98 ± 6.18 mL·kg−1·h−1, men vs women, respectively, P = 0.198). Sweat sodium concentration was also not significantly different between sexes (44.97 ± 16.4 vs 39.95 ± 12.6 mEq·L−1, men vs women, respectively, P = 0.262). Although men had a 27% higher rate of sweat sodium loss than women, the difference was not statistically significant (0.94 ± 0.49 vs 0.74 ± 0.41 mEq·kg−1·h−1, respectively, P = 0.155) because of the large interindividual variability that exists in sweat sodium losses (11,13,21). Sweating analysis data for the group (N = 46) and sexes are presented in Table 2.
Race day results.
On average, subjects lost 1.45 ± 1.6 kg (2.08% ± 2.1%) of body mass during the race. Mean baseline serum sodium concentration for all subjects (N = 46) was 145.4 ± 2.1 mEq·L−1, with no difference between women and men (145.3 ± 2.7 vs 145.4 ± 1.6 mEq·L−1, respectively, P = 0.852). Postrace serum sodium concentration (142.8 ± 4.4 mEq·L−1) was significantly lower than baseline, with an average decline of 2.5 ± 4.8 mEq·L−1 (t(45) = 3.575, P = 0.001). Postrace serum sodium concentration was not significantly different between women and men (141.9 ± 4.1 vs 143.6 ± 4.6 mEq·L−1, respectively, P = 0.208). The mean decline in serum sodium concentration was also not significantly different between sexes (−3.4 ± 4.6 vs −1.9 ± 4.9 mEq·L−1, women vs men, respectively, P = 0.284).
As shown in Figure 1, serum sodium concentration changes ranged from a decline of 13 mEq·L−1 to an increase of 8 mEq·L−1. The change in serum sodium concentration was significantly and negatively correlated with body mass change (r2 = 0.30, r = −0.55, P = 0.001) (Fig. 1A) and percent change in body mass (r2 = 0.30, r = −0.55, P = 0.001) (Fig. 1B). Within women, the change in serum sodium concentration was negatively correlated with change in body mass (r = −0.56, P = 0.010). In men, change in serum sodium concentration was significantly and negatively correlated with change in body mass (r = −0.54, P = 0.005), sweating rate (r = −0.49, P = 0.012), and rate of sweat sodium loss (r = −0.44, P = 0.023), as presented in Table 3.
Multiple regression analysis in men indicated that the rate of sweat sodium loss (mEq·kg−1·h−1) (b = −4.060, β = −0.401, P = 0.017) and change in body mass (b = −1.412, β = −0.504, P = 0.004) accounted for approximately 46% of the change in serum sodium concentration (R2 = 0.46, adjusted R2 = 0.41, F(2, 23) = 9.743, P = 0.001). In women, the change in body mass alone acounted for 31% of the change in serum sodium concentration (r2 = 0.31, P = 0.010).
Intake during the race of macronutrients, sodium, potassium, and fluid was recorded for 20 subjects (9 men and 11 women) (Table 4). In women (n = 11), the rate of sodium intake (mEq·kg−1·h−1) was correlated with the rate of sweat sodium loss (mEq·kg−1·h−1; r = 0.64, P = 0.035). This relationship was not significant in men (r = 0.27, P = 0.486). As a percentage of total sodium intake, the source of sodium from food, fluid, and pills varied widely and was not significantly different between sexes (P > 0.05). The sodium source for men, as a percentage of mean total intake, was as follows: 13.1% ± 15.5% (n = 7) from food, 82.8% ± 16.4% (n = 9) from fluid, and 21.4% ± 11.6% (n = 3) from pills. Although twice as many women ingested pills, the mean values were not significantly different: 9.7% ± 7.9% (n = 9) from food, 70.1% ± 23.8% (n = 11) from fluid, and 38.9% ± 22.4% (n = 6) from pills. Women displayed a significantly higher rate of fluid intake per kilogram of body mass during the race than men (17.2 ± 4.7 vs 10.9 ± 3.7 mL·kg−1·h−1, respectively, t(18) = −3.258, P = 0.004).
The main finding of this study is that changes in serum sodium concentration during an ultraendurance triathlon can largely be attributed to changes in body mass and rate of sweat sodium loss in men and the changes in body mass alone in women. In our study, approximately one-third of the change in serum sodium concentration for the group (N = 46) was associated with the change in body mass (r2 = 0.30). As shown in Figure 1, subjects that decreased body mass by 2.94 kg or approximately 4% had no change in serum sodium concentration. A 2-3 kg or 4% loss of body weight is similar to losses reported by other researchers during an ultradistance triathlon (4,5,10,16). Because changes in body mass during exercise are largely due to changes in water balance, as body mass decreased less or increased, serum sodium concentration would be maintained or decline. Conversely, a large fluid loss and decline in body mass may result in an increase or maintenance of serum sodium concentration, potentially offsetting large sodium losses from sweating.
Although the mixed sex group and the female subjects showed negative correlations between change in serum sodium concentration and body mass change (r = −0.55, P = 0.001; r = −0.56, P = 0.010, respectively), similar to a previous study (16), male subjects in our study demonstrated an additional relationship. In men, the change in serum sodium concentration during the race was also significantly negatively correlated with the rate of sweat sodium loss (mEq·kg−1·h−1; r = −0.44, P = 0.023). This relationship for the male subjects implies that those individuals with higher losses of sodium through sweating were those that displayed a larger decline in serum sodium concentration. In men, approximately 30% of the change in serum sodium concentration was accounted for by changes in body mass, and an additional 16% was accounted for by the rate of sweat sodium loss. Thus, men tended to not match their sweat sodium losses with sufficient sodium intake, which contributed to the decline in serum sodium concentration. Estimated sodium losses on the basis of the rate of sweat sodium losses measured during the sweating analysis trial are substantial, amounting to an estimated 734 mEq (16.9 g) of sodium during the race (12.7 h × 1.35 L·h−1 × 42.8 mEq·L−1; mean race time × mean sweating rate × mean sweat sodium concentration, respectively).
Approximately 30% of the decline in serum sodium concentration in women was accounted for by changes in body mass. However, women did not exhibit as clear of a picture as men regarding the contribution of sweat sodium loss to the decline in serum sodium concentration. Two factors may explain this observation. Although men and women had a nonsignificant difference in the rate of sweat sodium loss (0.94 ± 0.49 vs 0.74 ± 0.41 mEq·kg−1·h−1, P = 0.155), women displayed a statistically significant relationship between the rate of sweat sodium loss and the rate of sodium intake (mEq·kg−1·h−1; P = 0.035). This relationship did not exist for men. This implies that the female subjects more closely matched their sweat sodium losses with sodium intake, which was more easily accomplished as women had lower rates of sweat sodium losses than men. As has been previously reported, higher rates of sodium ingestion in female endurance athletes during 4 h of running can attenuate the decline in serum sodium concentration (20). The women in the current study most likely maintained their serum sodium concentration by replacing losses via sodium ingestion that offset their sweat sodium losses. This occurred although the women had a larger rate of fluid intake (mL·kg−1·h−1) than men (P = 0.004), which would potentially result in a greater dilution of the serum sodium concentration. Because of this better matching of sodium intake to loss in women, the relationship between sweat sodium losses and changes in serum sodium concentration was not apparent. Possible reasons for the relationship between increased sodium intake and sweat sodium losses may be that athletes in general, and women in particular, are responding to a sodium appetite mechanism (1,2,6) or that these highly trained athletes are simply aware of their high losses and are attempting to maintain a balance. However, it is not entirely clear as to why a relationship between rate of sweat sodium losses and sodium intake existed in women but not men. It may simply be due to women having lower sweat sodium losses, thus requiring less sodium intake to match those losses.
Baseline serum sodium concentration for women and men was similar (145.3 ± 2.7 vs 145.4 ± 1.6 mEq·L−1, P = 0.852, respectively). In a previous report of an ultraendurance triathlon, it was observed that women had a lower postrace serum sodium level than men (18). However, in the present study, postrace serum sodium concentration was not significantly different between women and men (141.9 ± 4.1 vs 143.6 ± 4.6 mEq·L−1, P = 0.208, respectively). A possible reason that the women in the present study did not display even more of a decline in serum sodium concentration despite larger fluid intake may be due to more closely matching their sodium losses with sodium ingestion.
Although field studies provide valuable information on physiological responses of athletes during actual performance, additional sources of error may be present than that in a laboratory setting. In the present study, the sweat patch technique was used to determine sweat sodium concentration and the resulting rate of sweat sodium loss. An overestimation of whole body sweat sodium losses has been reported with some patch measurements (15), possibly because of leaching when patches become saturated (23). To minimize possible overestimation, large absorbent patches were used in the present study, and they were placed on the subjects for only 30 min so they would not become saturated. Of note is that the mean sweat sodium concentration in our study (43 mEq·L−1) was lower than that reported during studies using the whole body wash-down technique (51.6 mEq·L−1) (15). It should be noted that the calculation of sweat sodium losses during the race was made on the basis of the preliminary sweat sodium analysis trial so that we could compare race intake and sweat sodium losses. Another potential source of error in our study is the reporting of nutrients, as we collected self-reported nutritional intake during the race. Athletes were free to consume their own or course-supplied nutrients during the race. Food, fluid, and sodium were not provided to the participants by the research staff. Upon finishing, participants were walked by escorts from the finish area to the field laboratory where blood samples, body mass measurements, and final nutrient intake were taken. To minimize experimental error due to potential fluid shifts postexercise, the majority of the blood samples were taken 5-8 min after completing the race, with only a few samples drawn up to 15 min postrace.
In conclusion, this study demonstrated that the change in serum sodium concentration during ultraendurance exercise such as the Ironman® triathlon is related to both change in body mass and rate of sweat sodium loss. Although the relationship of the rate of sweat sodium loss to changes in serum sodium was clear in men (r = −0.44, P = 0.023), it was not as clear in women (r = 0.10, P = 0.678), most likely because of women more closely offsetting rates of sodium loss with sufficient sodium intake (r = 0.64, P = 0.035). Although women more closely matched their sodium losses via sodium ingestion, which was more easily accomplished as they had lower sweat sodium losses than men, women consumed more fluid compared with men. This greater fluid intake in women may account for part of their slightly larger decrease in race day serum sodium than men. In summary, changes in serum sodium concentration during ultraendurance exercise are largely due to an interaction of sweat sodium loss, sodium ingestion, and fluid balance.
Funding for this study was provided by the Gatorade Sports Science Institute (Barrington, IL).
The authors thank the subjects for their enthusiastic participation. Special thanks to Simon S. Schenk, Jacob J. Baty, and Laura Dierenfield for their assistance.
Jeffrey J. Zachwieja and John R. Stofan are employed at the Gatorade Sports Science Institute. Edward F. Coyle was a member of the Gatorade Sports Science Institute Sports Medicine Review Board until 2007.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Beauchamp GK, Bertino M, Burke D, Engelman K. Experimental sodium depletion and salt taste in normal human volunteers. Am J Clin Nutr
2. Fitzsimmons JT. Angiotensin, thirst, and sodium appetite. Physiol Rev
3. Garigan TP, Ristedt DE. Death from hyponatremia
as a result of acute water intoxication in an Army basic trainee. Mil Med
4. Hiller W. Dehydration and hyponatremia
during triathlons. Med Sci Sports Exerc
. 1989;21(5 suppl):S219-21.
5. Hiller W, O'Toole M, Fortess E, Laird R, Imbert P, Sisk T. Medical and physiological considerations in triathlons. Am J Sports Med
6. McCance RA. Experimental sodium chloride deficiency in man. Proc R Soc Lond B Biol Sci
7. Montain S, Sawka M, Wenger C. Hyponatremia
associated with exercise
: risk factors and pathogenesis. Exerc Sport Sci Rev
8. Noakes TD, Goodwin N, Rayner BL, Branken T, Taylor RK. Water intoxication: a possible complication during endurance exercise
. Med Sci Sports Exerc
9. O'Brien K, Montain S, Corr W, Sawka M, Knapik J, Craig S. Hyponatremia
associated with overhydration in U.S. Army trainees. Mil Med
10. O'Toole M, Douglas P, Laird R, Hiller D. Fluid and electrolyte status in athletes receiving medical care at an ultradistance triathlon
. Clin J Sport Med
11. Patterson MJ, Galloway SD, Nimmo MA. Variations in regional sweat composition in normal human males. Exp Physiol
12. Roalstad M, Crosby L, O'Toole M, Zigler A, Grotke G, Hiller D. Heart rate monitoring during an ultraendurance event. Med Sci Sports Exerc
. 1987;19(2 suppl):S71.
13. Robinson S, Robinson AH. Chemical composition of sweat. Physiol Rev
14. Rogers G, Goodman C, Rosen C. Water budget during ultra-endurance exercise
. Med Sci Sports Exerc
15. Shirreffs SM, Maughan RJ. Whole body sweat collection in humans: an improved method with preliminary data on electrolyte content. J Appl Physiol
16. Speedy DB, Noakes TD, Kimber NE, et al. Fluid balance during and after an ironman triathlon
. Clin J Sport Med
17. Speedy DB, Noakes TD, Rogers IR, et al. A prospective study of exercise
in two ultradistance triathletes. Clin J Sport Med
18. Speedy DB, Noakes TD, Rogers IR, et al. Hyponatremia
in ultradistance triathletes. Med Sci Sports Exerc
19. Tanaka H, Monahan KD, Seals DR. Age-predicted maximal heart rate revisited. J Am Coll Cardiol
20. Twerenbold R, Knechtle B, Kakebeeke TH, et al. Effects of different sodium concentrations in replacement fluids during prolonged exercise
in women. Br J Sports Med
21. Verde T, Shephard RJ, Corey P, Moore R. Sweat composition in exercise
and in heat. J Appl Physiol
22. Vrijens DM, Rehrer NJ. Sodium-free fluid ingestion decreases plasma sodium during exercise
in the heat. J Appl Physiol
23. Weschler LB. Sweat electrolyte concentrations obtained from within occlusive coverings are falsely high because sweat itself leaches skin electrolytes. J Appl Physiol
Keywords:©2010The American College of Sports Medicine
HYPONATREMIA; EXERCISE; SODIUM SUPPLEMENTATION; TRIATHLON