Participants in prolonged exercise tasks are encouraged to ingest fluids to prevent the development of excessive dehydration. This behavior is important for preserving performance and attenuating the physiological strain that accompanies dehydration, but it is often neglected that sweat contains electrolytes as well as water and that gross excess of body water relative to sodium content can produce a potentially life-threatening syndrome termed hyponatremia. During the past 15 yr, there have been > 55 published case reports of individuals developing symptomatic hyponatremia, or water intoxication, consequent to participation in prolonged exercise/work in warm weather. Although this medical event remains relatively rare (7,12,13), the frequency appears to be increasing. The potential for serious and deadly outcomes due to hyponatremia makes it important that sports medicine professionals know what the illness is, the risk of developing symptomatic hyponatremia, predisposing conditions and behaviors, and presenting symptoms.
We review the possible mechanisms responsible for the development of hyponatremia associated with exercise and consider the hypothesis that there is a population at increased risk for developing symptomatic hyponatremia during prolonged exercise. Two review papers have been published on the hyponatremia of exercise (3,6). This review provides a critical analysis of published case reports to determine the mediating factors leading to illness.
CAUSE AND SYMPTOMS
The term hyponatremia is strictly defined as a serum sodium level below 135 mEq/L. However, it is a term also used clinically to refer to the syndrome that can occur when there is rapid lowering of blood sodium, usually to levels below 130 mEq/L. In 57 case reports in the literature *, serum sodium concentrations at presentation averaged 121 mEq/L and ranged from 109 to 131 mEq/L.
The symptomatic hyponatremia of exercise arises consequent to prolonged work (typically longer than 6 h) where sweating is the primary means of dissipating heat. Because sweat contains not only water but also small quantities of electrolytes, there is a progressive loss of water, sodium, chloride, and potassium. Sweat electrolyte losses contribute to the development of the syndrome, particularly if sweat sodium losses are high. The condition may also occur when individuals consume low-sodium or sodium-free water in excess of sweat losses during and/or shortly after completing exercise. In either case, the reduction in solute concentration in the extracellular fluid (ECF) promotes movement of water from the ECF into cells. If this fluid shift is of sufficient magnitude and occurs rapidly, it can congest the lungs, swell the brain, and alter central nervous system function. Figure 1 presents the physiological consequences of hyponatremia.
Signs and symptoms of hyponatremia include confusion, disorientation, mental obtundation, headache, nausea, vomiting, aphasia, incoordination, and muscle weakness. Complications of severe and rapidly evolving hyponatremia include seizures, coma, pulmonary edema, and cardiorespiratory arrest. Although the condition is generally treatable without long-term sequelae, death has occurred (8).
Athletes who have developed symptomatic hyponatremia were participating in distance running events of ≥ 42 km and triathlons lasting 9–12 h. In these events, symptomatic hyponatremia occurs in 0.1–4% of the participants (7,12). Hyponatremia, however, is present in a significant number of the athletes seeking medical care; 3–27% of athletes who received medical care after ultramarathon running and/or triathlon participation were diagnosed with hyponatremia (4,7,12).
The hyponatremia associated with exercise is not limited to ultraendurance athletes. The U.S. Army Medical Surveillance Activity reported an average of 19 hyponatremia hospitalizations annually between 1989 and 1999 (13,14). Nearly 85% of these hospitalizations occurred between May and September, with the most cases occurring in July. The proportion of men and women among the cases reflected their overall proportions in the U.S. Army (15% female and 85% male). Whites may be more likely than other racial populations to be afflicted, because 75% of the cases occurred among white soldiers, who comprise only 63% of the U.S. Army population.
Backer et al. (2) reported that symptomatic hyponatremia accounted for 7 of 116 (6%) heat-related illnesses treated by staff at Grand Canyon National Park in the summer of 1993. Interestingly, six of seven cases occurred in females.
In addition to the number of participants developing symptomatic hyponatremia, there is a population who do not present with symptoms but who develop low sodium concentrations during activity and therefore can be considered to be at increased risk for the syndrome. In one prospective study (12), 11 athletes finished a triathlon with serum sodium levels below 130 mEq/L, but only 7 (63%) sought medical attention. In a separate study (9), only five of nine triathletes (56%) who finished the race with serum sodium levels below 130 mEq/L sought medical attention. These two observations suggest that individual susceptibility may play a role in the cause, but more important, document that a fair number of people performing prolonged work are developing low serum sodium levels.
Three explanations have been proposed to explain the hyponatremia associated with prolonged exercise: (1) hyponatremia is primarily the result of unreplaced solute losses, (2) hyponatremia is the result of overload of the ECF precipitated by excessive water intake, and (3) salt is lost from the ECF by movement of sodium into a “third space,” perhaps the intestinal lumen (6). What evidence supports these explanations? Which are the most plausible?
Table 1 presents the predicted serum sodium response to prolonged exercise for two individuals of low and average body mass, with three sweat sodium concentrations, who replace their sweat losses with sodium-free fluid. The three sweat concentrations, i.e., 25, 50, and 75 mEq/L, represent the sweat sodium concentrations for a heat-acclimatized, endurance- trained individual; a non-heat-acclimatized, sedentary person; and an individual with a medical condition that limits sweat sodium reabsorption. For the example, it was assumed that total body water was 63% of body mass and that water would distribute within the ECF and ICF until osmotic equilibrium was reached. It was further assumed that the two individuals were running (10 km·h−1), where energy cost would be dependent on body mass. Sweating rates were calculated based on metabolic rate using the generalized equation of Barr and Costill (3). Sweat losses were predicted assuming a 90-km ultramarathon distance.
The data presented in Table 1 illustrate that sweat sodium losses are an important contributor to the reduction of serum sodium when water or sodium-free solutions are used to replace sweat lost during exercise. In the example provided, if a 70-kg man had sweat sodium concentrations of 25, 50, and 75 mEq/L and drank sufficient sodium-free water to fully replace the 8.6 L of sweat loss, serum sodium would be expected to decline ∼ 5, 10, and 15 mEq/L. However, to lower the serum sodium to the average value reported for individuals with symptomatic hyponatremia (121 mEq/L), the 70-kg man would still need to accrue a fluid excess of 5.1, 3.3, and 1.6 L at the low, moderate, and high sweat sodium concentrations, respectively. The fluid excess must result in large part from excessive water or sodium-free fluid intake (relative to sweat and respiratory losses) as the liberation of water associated with muscle glycogen would be expected to add only 0.5–0.7 L of water to the total body water.
Table 1 also illustrates that smaller individuals who have similar sweat sodium concentrations need less fluid excess than larger persons to dilute serum sodium to levels associated with symptoms. Therefore, if a group of people are drinking at the same rate (as may happen if they follow drinking schedules that are not adjusted for body mass), the individuals with smaller body mass will dilute their extracellular sodium levels more quickly and further than will those of larger body mass.
In the mid-1980s, Hiller et al. (4) argued that athletes participating in the Hawaiian Ironman triathlon were developing symptomatic hyponatremia consequent to large unreplaced sodium losses. Their belief was based on the observation that 70% of the hyponatremic athletes treated at the Hawaiian Ironman Triathlon were dehydrated. However, no criteria were given as to how dehydration was determined or what magnitude of water deficit was present, and they appear to have ignored that the metabolic processes during an Ironman-length event would produce ∼ 1 kg body mass loss even without a net body water loss. The dehydration generalization is also in contrast to their data showing that four of seven individuals who finished the Ironman triathlon with serum sodium levels below 130 mEq/L either maintained or gained body mass during the race (9).
For sweat sodium losses to be the primary cause of hyponatremia, a large sodium deficit must accrue during exercise. Sweat sodium levels typically range from 10 to 60 mEq/L. A 70-kg individual with 44 L of total body water who drinks adequate sodium-free fluid to replace 100% of sweat losses would need to develop a 660-mEq sodium deficit to decrease serum sodium from 140 to 125 mEq/L. Assuming a finish time of 9 h for a 90-km ultramarathon footrace and a sweating rate of 0.9 L·h−1 (3), the development of a 660-mEq sodium deficit would require a sweat sodium concentration of 81 mEq/L. At a sweating rate of 1.2 L·h−1, a sweat sodium concentration of 61 mEq/L would be necessary to achieve the required sodium deficit. These are exceptionally high sweat sodium concentrations for physically fit and heat-acclimatized athletes.
Do case studies support the argument that the hyponatremia of exercise can be produced from excessive salt losses without excessive water intake? Vrijens and Rehrer (15) recently reported one case of symptomatic hyponatremia consequent to cycling for 2.5 h and consuming 2.5 L of water during exercise. Sweat loss was ∼ 2.8 L. Serum sodium declined from 144 to 128 mEq/L during exercise. Assuming a total body water estimate of 50.7 L for the 84.5-kg individual, the calculated sweat sodium concentration would be in excess of 100 mEq/L to lower the serum sodium level so dramatically. Some individuals do, in fact, achieve such levels. Smith et al. (10) reported that a 24-yr-old infantryman who experienced hyponatremic heat exhaustion had sweat sodium concentrations of 81 and 103 mEq/L in samples obtained through pilocarpine iontophoresis.
It has been argued by others that fluid intake in excess of sweat losses is the major reason for the development of hyponatremia (3,6). Based on the predictions presented in Table 1, it would be expected that few individuals would develop symptomatic hyponatremia during prolonged exercise by sodium losses, alone. Indeed, fluid intake in excess of sweat losses would be required for each prediction to lower sodium levels from 140 to 121 mEq/L. Published case studies also support the argument that excess fluid intake contributes to the development of the syndrome. From 57 individual case reports in the literature (see Acknowledgments), sufficient data were presented to estimate sweat water losses in 24 cases. In this subset, 20 of 24 individuals reported fluid intake in excess of predicted sweat losses. In the prospective case presented by Armstrong et al. (1), the volunteer drank 1.8–2.1 L·h−1 despite sweating only 0.5–0.8 L·h−1. Similarly, four soldiers afflicted during U.S. Army training reported fluid consumption rates that were approximately two times higher than predicted for their metabolic rate and the weather conditions (8).
An alternative explanation for the lowering of serum sodium is that sodium leaves the ECF and enters “an as yet unidentified third space”—possibly the gastrointestinal tract (6). The notion originates from observations that there are upper limits of gastric emptying, that sodium is secreted into the intestinal lumen when carbohydrate is present in the intestine, and that gastric emptying and intestinal absorption are reduced during exercise. It appears that the third space is intended to explain how ingested hypotonic fluid could lower the concentration of sodium in the ECF while the ingested fluid is still in the gastrointestinal tract. Thus, this explanation provides a mechanism for rapid lowering of the serum sodium concentration after drinking independent of water entering the vascular space and expanding the extracellular and intracellular fluid spaces.
The weakness of the third space proposal is that intestinal perfusion studies report only modest movement of sodium into the intestinal lumen after ingestion of water. There is sodium secretion following the ingestion of carbohydrate solutions, but the magnitude appears proportional to the carbohydrate content. Furthermore, if large quantities of water were retained in the gastrointestinal tract, diarrhea would be expected. Yet, diarrhea is not a prevalent symptom of the hyponatremia associated with exercise. Lastly, the prospective data in the literature do not require a third space to account for serum sodium dilution. The calculations used to formulate secreted into the intestinal lumen when carbohydrate is present in the intestine, and that gastric emptying and intestinal absorption are reduced during exercise. Table 1 accurately predict the sodium dilution from the prospective studies of both Armstrong et al. (1) and Speedy et al. (11). Thus, in these cases the sodium dilution could be accounted for by the relative expansion of the total body water and sodium deficit.
The observation that symptomatic hyponatremia associated with exercise arises after prolonged and excessive fluid intake has led to the suggestion that impaired renal responses and/or function plays a role in the development of hyponatremia, in that it would be expected that normal fluid regulatory mechanisms would prevent gross overhydration. In support of such a hypothesis, hospitalized patients who develop symptomatic hyponatremia have impaired renal water excretion, often associated with an inappropriate (relative to the osmotic and volume status) secretion of arginine vasopressin (AVP) during the fluid overload. Whether such a mechanism plays a role in the retention of excess fluid during exercise remains unclear. In prospective studies, there was a dilution of serum sodium consequent to fluid overload with no change in AVP concentration (1,12). Armstrong et al. (1) observed no change in AVP concentration as a subject drank sufficient fluid to accrue a 4-L body water excess during 4 h of exercise. It was not until after the subject complained of nausea, a stimulus for AVP secretion, and discontinued drinking that AVP levels rose. Similarly, Speedy et al. (12) reported low AVP levels in athletes participating in an ultraendurance triathlon despite serum sodium levels below 130 mEq/L. These observations suggest that inappropriate AVP secretion is not necessary for the hyponatremia associated with exercise.
The evidence to date suggests that hyponatremia occurs in persons with normal renal function. Irving et al. (5) were unable to identify renal compromise in eight athletes who developed symptomatic hyponatremia after participation in an 88-km ultramarathon footrace. Similarly, many case studies report that the individuals urinated before the development of symptoms and produced relatively dilute urine at presentation (1,2).
An alternative explanation for the retention of water during the hyponatremia of exercise is that changes in the control factors determining water and sodium reabsorption limit urine volume. Heavy exercise lowers both renal blood flow and glomerular filtration rate. Exercise can decrease urine flow rate 20–60%, and this effect can persist even if an overly hydrated person drinks during exercise. Exercise will also increase sympathetic nerve activity to the kidney as well as various hormones, such as renin and angiotensin, that increase tubular sodium reabsorption and iso-osmotic reabsorption of water. The reduction in glomerular filtration alone or in combination with iso-osmotic reabsorption of water at the proximal tubule would decrease the amount of free water delivered to the collecting duct and thus lower urine production without invoking the inappropriate AVP hypothesis.
In summary, both large salt losses via sweat and excessive water intake, singly or in combination, contribute to lowering serum sodium during prolonged exercise. To a degree, the kidneys are able to maintain serum sodium within normal limits by excreting free water. During exercise, however, increased renal sympathetic nerve activity and activation of the renin-angiotensin system reduce urine flow substantially, thus limiting the ability of the kidneys to compensate for imbalances in the intake and loss of salt and water and increasing the risk of developing hyponatremia. Figure 2 illustrates the various factors that contribute to the development of hyponatremia during exercise. Preventive actions that can be taken are to not drink in excess of predicted sweat losses and to ingest sodium when the work task will produce persistent and profuse sweating.
A factor that has received little consideration is whether the persons afflicted with symptomatic hyponatremia may be more susceptible to fluid-electrolyte imbalance than their nonafflicted peers. It is clear from the epidemiological data that hyponatremia associated with prolonged exercise is quite rare. Furthermore, it has been reported in prospective studies that only 44–64% of ultramarathon participants who finish the race with serum sodium levels < 130 mEq/L require medical attention (9,12). Given these two observations, it seems important to consider whether there is a population at risk for developing symptomatic hyponatremia during prolonged exercise.
In 1995, Smith et al. (10) reported the case of a British infantryman who developed symptomatic hyponatremia twice. The attending physicians performed both a sweat test and genetic studies to determine whether the soldier had cystic fibrosis. The sweat test was chosen because high concentrations of sodium chloride in sweat are so characteristic of cystic fibrosis that they provide the basis of a widely used diagnostic test. The sweat test produced elevated sweat sodium and chloride levels (81–103 and 102–143 mEq/L, respectively) consistent with a diagnosis cystic fibrosis. Subsequent genetic testing confirmed this diagnosis. Clinical evaluation showed that the soldier had normal respiratory and pancreatic function, in contrast to the usual clinical picture of cystic fibrosis. Thus, this study identified a soldier with cystic fibrosis who had no overt disease other than an impaired ability of his sweat glands to reabsorb sodium and chloride.
Cystic fibrosis is an autosomal recessive disorder caused by certain mutations on the cystic fibrosis transmembrane regulatory (CFTR) gene. The classic form of cystic fibrosis is characterized by repeated bouts of pneumonia and usually digestive abnormalities. A relatively small number of CFTR mutations are associated with most cases of the classic form, and 4–5% of northern European adults are heterozygous carriers of the classic form. However, > 800 variants of the CFTR gene have been identified, many of which are not known to be associated with any clinical abnormalities. The prevalence of most of these variants in the population is not known, but because many of these variants are not associated with known disease, it is possible that they are common enough in the population to contribute to a significant number of cases of hyponatremia associated with exercise.
Sweat testing was recently performed on two Marine recruits who developed symptomatic hyponatremia consequent to Marine basic training at Parris Island, SC (Dr. Scott Flinn, Personal Communication). Both had sweat chloride levels high enough to be consistent with a diagnosis of cystic fibrosis. One of these patients, a Mexican American, tested negative on a genetic panel for cystic fibrosis. However, the panel consisted of the 14 most common mutations found in northern Europeans and Ashkenazi Jews with cystic fibrosis. We cannot rule out the possibility that this individual has other variants of the gene. Future research in this area will be helpful in determining why some individuals develop low serum sodium levels during prolonged work.
The hyponatremia associated with prolonged exercise arises primarily from fluid overload, underreplacement of sodium losses, or both. The reduction in extracellular solute leads to movement of water into the intracellular space. If the resulting cellular swelling is of sufficient magnitude, symptoms of central nervous system dysfunction will occur. Competitive and recreational athletes as well as occupational workers performing prolonged work should be taught that persistent excessive fluid intake can be harmful and that fluid intake should not exceed sweat losses. During prolonged exercise lasting in excess of 3–4 h, snacks or fluids containing sodium chloride should be ingested to offset the loss of salt in sweat. The latter recommendation is especially prudent for individuals who know that they lose excessive amounts of salt in their sweat. Several possible contributing mechanisms were discussed, with the most intriguing being that some persons may have variants of the cystic fibrosis gene.
* Due to limitations on the number of references, not all case reports are cited. References are available on request to the authors.
1. Armstrong, L.E., Curtis, W.C., Hubbard, R.W., Francesconi, R.P., Moore, R., Askew E.W.. Symptomatic hyponatremia during prolonged exercise
in heat. Med. Sci. Sports Exerc. 25: 543–549, 1993.
2. Backer, H.D., Shopes, E., Collins S.L.. Exertional heat illness and hyponatremia in hikers. Am. J. Emerg. Med. 17: 532–539, 1999.
3. Barr, S.I., Costill D.L.. Water: can the endurance athlete get too much of a good thing? J. Am. Diet. Assoc. 89: 1629–1635, 1989.
4. Hiller, W.D.B., O’Toole, M.L., Fortess, E.E., Laird R.H.. Medical and physiological consideration in triathlons. Am. J. Sports Med. 15: 164–167, 1987.
5. Irving, R.A., Noakes, T.D., Buck, R., Van Zyl Smit, R., Raine, E., Godlonton, J., Norman R.J.. Evaluation of renal function and fluid homeostasis during recovery from exercise
-induced hyponatremia. J. Appl. Physiol. 70: 342–348, 1991.
6. Noakes, T.D. The hyponatremia of exercise
. Int. J. Sports Nutr. 2: 205–228, 1992.
7. Noakes, T.D., Norman, R.J., Buck, R.H., Godlonton, J., Stevenson, K., Pitts G.C.. The incidence of hyponatremia during prolonged ultraendurance exercise
. Med. Sci. Sports Exerc. 22: 165–170, 1990.
8. O’Brien, K., Montain, S.J., Corr, W.P., Sawka, M.N., Knapik, J., Craig S.C.. Overhydration hyponatremia in Army trainees. Mit. Med. 2000.
9. O’Toole, M.L., Douglas, P.S., Laird, R.H., Hiller W.D.B.. Fluid and electrolyte status in athletes
receiving medical care at an ultradistance triathlon. Clin. J. Sport Med. 5: 116–122, 1995.
10. Smith, H.R., Dhatt, G.S., Melia, W.M.A., Dickinson J.G.. Lesson of the Week: Cystic fibrosis presenting as hyponatraemic heat exhaustion. Br. Med. J. 10: 579–580, 1995.
11. Speedy, D.B., Noakes, T.D., Rogers, I.R., Hellemans, I., Kimber, N.E., Boswell, D.R., Campbell, R., Kuttner J.A.. A prospective study of exercise
-associated hyponatremia in two ultradistance triathletes. Clin. J. Sport Med. 10: 136–141, 2000.
12. Speedy, D.B., Noakes, T.D., Rogers, I.R., Thompson, J.M.D., Campbell, R.G.D., Kuttner, J.A., Boswell, D.R., Wright, S., Hamlin M.. Hyponatremia in ultradistance triathletes. Med. Sci. Sports Exerc. 31: 809–815, 1999.
13. USACHPPM. Hyponatremia hospitalizations, U.S. Army 1989–1996. Med. Surveill. U. S. Army Center for Health Promotion and Preventive Medicine. Monthly Rep. 9–10, 1997.
14. USACHPPM. Overhydration and hyponatremia among active duty soldiers, 1997–1999. Med. Surveillance Monthly Rep. 6: 9–11, 2000.
15. Vrijens, D.M.J., Rehrer N.J.. Sodium
-free fluid ingestion decreases plasma sodium
in the heat. J. Appl. Physiol. 86: 1847–1851, 1999.