Letters to the Editor
THE PHYSICIAN AND SPORTSMEDICINE - VOL 29 - NO. 7 - JULY 2001
Hyponatremia or Hype?
We have spent months correcting damage among physicians, athletic trainers, coaches, and athletes done by the article published last year by Dr Noakes on hyponatremia (1) and similar comments by him in the October 2000 Runner's World (2). While the article in The Physician and Sportsmedicine contains useful information about the potentially serious medical condition of hyponatremia, it contains this inaccurate statement on page 72: "...the evidence is nonexistent that the modest levels of dehydration in endurance athletes—body mass losses of 2% to 8%—have any health consequences during exercise."
It may be true that endurance athletes and other athletes who participate at very low exercise intensities (eg, most of the athletes competing in ultra events) would be at a low risk (but not no risk) of adverse medical effects associated with the stated degrees of dehydration. However, this statement is not true for endurance athletes and other athletes who participate in shorter events where the intensity is great enough to overwhelm the thermoregulatory system, especially in the presence of extreme environmental conditions. Further, 8% dehydration should never be deemed "a modest level of dehydration." When a 200-lb athlete loses 16 lb of sweat, that is a significant or severe level of dehydration.
Noakes' observations of ultramarathoners may exist because dehydration does not affect heat storage and rectal temperature (Tre) if the exercise intensity is low. A study by Montain et al (3) demonstrated this fact in healthy males who exercised at 25%, 45%, and 65% of VO2max. However, the following physiologic and clinical facts provide strong evidence that the risk and likelihood of serious hyperthermia (above the 39° to 40°C threshold for heatstroke) increases when athletes are dehydrated:
Therefore, we conclude that athletes who exercise at high intensities and wear uniforms in hot environments are at great risk for dangerous hyperthermia, especially when dehydrated to 8%. The average person reading Noakes' article, however, could easily conclude that dehydration has no influence on heat storage during any exercise-related heat stress. In fact, one of us (D.J.C.) answered numerous telephone calls from athletic trainers who, after reading this article, questioned time-honored teachings regarding the need for adequate fluid intake.
We believe that the recent position statements by the American College of Sports Medicine (13) and the National Athletic Trainers' Association (14)—which stress the importance of minimizing fluid losses by encouraging fluid replacement equal to fluid losses—are better approaches to the issue of dehydration. (We were closely involved with both statements.) If the medical staff advocates individualized rehydration protocols, based on sweat rates, that properly maintain a normal hydration status, the problems associated with improper fluid replacement—such as dehydration and hyponatremia—will be eliminated, or at least dramatically diminished.
Douglas J. Casa, PhD, ATC, CSCS
I am incensed by the article on hyponatremia by Flinn and Sherer (1) and the commentary by Dr Noakes (2), both of which fail to address the consensus opinions of hyperthermia as relating to dehydration and vascular shifts as outlined by Dr Carl Gisolfi and Dr David Costill. Their extensive and thorough evaluations refute Noakes' hypothesis. Noakes refers to his articles extensively without one reference to opposing views. For Noakes to ignore this literature implies that people treating hyperthermic athletes should ignore the words of Gisolfi and Costill, which is most inappropriate. Also, to state that hypothermia can exist at 102°F but possibly not at 106°F and above is a misstatement of facts.
I am the medical director of one of the top 25 races in the country, the Crim 10-mile Road Race. The race has been run in Flint, Michigan, in August heat and humidity for 24 years without one death from hyperthermic effect. Each year we treat 60 to 150 cases of mild or moderate-to-severe hyperthermia. We have done this on the basis of hydration, and we will continue to do so despite Noakes' comments. In theory, I am familiar with hyponatremia, but hyperthermia should be largely addressed as fluid-volume depletion.
Jon L. Schriner, DO
The articles on hyponatremia by Flinn and Sherer and Noakes bring up a very good clinical reminder not to treat a collapsed athlete automatically with intravenous (IV) fluids. The mechanism for hyponatremia during endurance events has been suggested by these authors and others to be related to overhydration with hypotonic fluids, increasing the loss of sodium through sweating or inappropriate ingestion of sodium during a long distance event.
In my experience and in some brief studies done on ultramarathoners, I have the suspicion that inappropriate ADH (vasopressin) syndrome may be a significant factor to add to the causes of hyponatremia in some of these athletes. It is my feeling that certain susceptible individuals, when subjected to the stress of long distance events, may develop inappropriate ADH clinical syndrome stimulated by the hemodynamics of long endurance activities. We know this can happen in other clinical "stress" situations in susceptible individuals, and I believe it might be associated with some of the case reports we have seen in the literature.
It is not easy to get levels of antidiuretic hormone done inexpensively by a laboratory, but it may be worthwhile in athletes who experience this syndrome to get more data and information regarding this hormonal possibility.
Barry D. Mink, MD
The recent articles on hyponatremia in The Physician and Sportsmedicine underscore the attention this topic is receiving—for good reasons.
Overhydration, both orally and IV, with hypotonic (relative to sweat) solutions is being nailed as the culprit in most of these cases, with some individuals reported to have been overhydrated by as much as 6 to 7 L of free water. I am in full agreement with the opinions of Dr Noakes, and others, outlining the absolute importance of confirming the diagnosis prior to treatment. The injudicious use of IV fluids, especially hypotonic (relative to serum) fluids, can only lead to worsening of hyponatremia, if present, with all of its potential consequences.
However, from an initial prevention standpoint, I believe that overhydration with hypotonic (relative to sweat) fluids is not the primary problem. The primary problem is underreplacement of relevant electrolytes at the rate at which they are lost.
Renal physiology is certainly complex, but the human kidney is more than able to excrete plenty of free water, up to 15 L/day, according to most renal physiology texts. Certainly the rate of free water accumulation during endurance events may overwhelm the kidney's ability to excrete it. However, we must also ask what process in the kidneys might lead to water accumulation. The same renal physiology texts tell us that:
Sodium and fluid are two independently regulated processes. Therefore, if I am losing 1 L/hr of water via sweating, I must replace it at this rate for optimal performance (and survival!). If I am losing sweat sodium at 1,000 mg/L of sweat, I must replace it at this rate.
Increasing the ingestion of relevant electrolytes includes sodium and potassium, and possibly others like calcium, magnesium, and zinc. The problem of "overhydration" and hyponatremia will certainly be reduced, if not eliminated, if we advise athletes about the importance of electrolyte replacement at the rate at which it is lost.
Douglas W. Stoddard, MD
I found both articles on hyponatremia of high standard, thought-provoking, and prideworthy. No expert, I have the following observations:
Heat exhaustion (not heat hyperpyrexia) is prevalent in nonacclimatized persons. Acclimatization is a graduated exposure to increasing loads of exertion over a period (typically 3 weeks) under "hot" conditions expected to be encountered at a scheduled event. I believe this allows sweat glands to conserve sodium, thus precluding hyponatremia from drinking water, the most commonly available source.
Further, convulsions are known to occur in the hyponatremic. Trans-urethral removal of the prostate associated with concurrent fluid overload has in fact resulted in just such incidents.
Shyam M. Pathak, MD
DR NOAKES RESPONDS: It is a pleasure to respond to this series of letters, as it allows an opportunity to clarify some important issues. Perhaps there are three reasons for this unnecessary confusion of the relationships between dehydration, fluid replacement, and exercise-related heat illness.
Confounder #1. The first reason is that popular concepts on this topic have changed incrementally over the past 25 years independent, in my view, of sufficient new scientific information to justify some of those changes. For example, during that time, the American College of Sports Medicine (ACSM) has produced four different position stands pertaining to fluid ingestion and the prevention of heat injury during exercise (1-4). Each is somewhat different, particularly concerning the nature of the claims made for the benefits of such fluid replacement and the volumes of fluid that should be ingested during exercise.
In 1975, the simple advice was that athletes should "be encouraged to frequently ingest fluids during competition" (1). "Frequently" was considered to be every 3 to 4 km in races of 16 km or longer, suggesting that fluid ingestion was of lesser importance in races less than 16 km. This advice was entirely appropriate in 1975 since, for the previous 100 years, athletes had been advised to avoid fluid ingestion during exercise (5). The rationale for fluid ingestion was to "reduce rectal temperature and prevent dehydration." No specific reference was made to the possibility that fluid ingestion alone would influence the risk of, much less prevent, heat illness during exercise.
The subsequent position stand of 1987 (2) introduced two additional statements of relevance to this debate: First, "Fluid consumption before and during the race will reduce the risk of heat injury, particularly in longer runs such as the marathon [emphasis added]" (6-8). The three references cited, however, do not support that conclusion because all of them studied only the effects of fluid ingestion on rectal temperature during exercise. None studied the incidence of heat injury in marathon runners, nor the influence of different rates of fluid ingestion on that incidence. So none could conclude that fluid ingestion reduces the risk of heat injury, however defined, in marathon runners. Hence, this position stand, which introduced a fundamentally novel concept not present in the 1975 position statement, was not based on new evidence.
The second statement was: "Such dehydration will subsequently reduce sweating and predispose the runner to hyperthermia, heat stroke, heat exhaustion, and muscle cramps" (8). Nor is this statement evidence-based since the cited reference did not study the incidence of hyperthermia, heatstroke, heat exhaustion, or muscle cramps in marathon runners or the effects of different rates of fluid ingestion on those incidences. Rather, the study reported a relationship between the postexercise rectal temperature and the extent of weight loss in competitors in 32-km running races. The conclusions that have been applied to that study have been criticized in detail elsewhere, including argument that any apparent relationship between weight loss and the postexercise rectal temperature may have been spurious (5,9). Of interest was the finding that the winners of those races were also the most dehydrated and hyperthermic, a finding at variance with the theory that dehydration impairs athletic performance and increases the risk of heat illness.
The position stand of 1996 (3) proposed that dehydration can "predispose the runner to heat exhaustion or the more dangerous hyperthermia and exertional heatstroke" (10,11). However, this statement is not evidence-based since the two referenced articles are literature that which do not provide any new findings available since the 1987 position stand, which supports the conclusion that dehydrated runners are at increased risk of "dangerous hyperthermia" or heatstroke. Indeed, the 1996 position stand appears to be contradictory since it later states that "excessive hyperthermia may occur in the absence of significant dehydration" especially in short distance races when the rate of heat production is high. That statement is evidence-based.
The 1996 position stand also includes the statement, "Adequate fluid consumption before and during the race can reduce the risk of heat illness, including disorientation and irrational behavior, particularly in longer events such as a marathon [emphasis added]" (6-8). But the statement is not evidence-based since the same three references from the 1987 position stand are again cited, and none measured the effects of different rates of fluid ingestion on the incidence of heat illness, disorientation, or irrational behavior in any running races, including marathon races. The statement also proposes that athletes should be encouraged to "replace their sweat losses or consume 150 to 300 mL every 15 minutes (600 to 1,200 mL per hour)". No scientific rationale for that proposal is provided.
The position stand also includes the statement, "Intravenous (IV) fluid therapy facilitates rapid recovery [in runners with heat exhaustion]" (12,13). But that statement is also not evidence-based since it cites two review articles that do not contain evidence from controlled clinical trials proving that novel claim.
The related 1996 position stand on exercise and fluid replacement (4) extends these new claims by stating that the "most serious effect of dehydration resulting from the failure to replace fluids during exercise is impaired heat dissipation, which can elevate core temperature to dangerously high levels (ie, >40°C)." Later the statement is made that dehydration during exercise "presents the potential for the development of heat-related disorders" including the potentially life-threatening heatstroke (14,15). It is therefore reasonable to surmise that fluid replacement that offsets dehydration and excessive elevation in body heat during exercise may be instrumental in reducing the risk of thermal injury" (10).
Once again, these conclusions are not evidence-based, as all the material cited to support these new ideas are review articles that represent the individual beliefs of the authors. Furthermore none, provides specific data showing either that athletes who are dehydrated are at greater risk of heatstroke during exercise or that fluid ingestion during exercise can reduce that risk. My criticisms of the flaws in the paper of Wyndham (15) and the related paper on which it is based (8), are detailed elsewhere (5,9).
The position stand on exercise and fluid replacement confirms the belief that the rate of fluid ingestion during exercise should equal the sweat rate and that "fluid and carbohydrate requirements can be met simultaneously by ingesting 600 to 1,200 mL/hr of solutions containing 4% to 8% carbohydrate." In addition, runners are encouraged to "consume the maximal amount that can be tolerated." Although there is substantial evidence supporting the proposal that ingesting carbohydrate at rates of about 60 g/hr can enhance performance during prolonged exercise by preventing hypoglycemia, there is no published evidence to prove that fluid should be ingested at rates equal to sweat rates in order to prevent heat injury.
To summarize, over the course of four revisions, the ACSM position stands have become progressively more strident in promoting the belief that high rates of fluid ingestion during exercise are necessary to prevent heatstroke and other heat illnesses. But this advice is not evidence-based, since none of the position stands refers to specific prospective studies from which such definite conclusions can be drawn. Furthermore, not even one published cross-sectional study shows that runners with heat illnesses are more dehydrated than are those who do not develop heat injury during exercise.
Hence, the conclusions of the four ACSM position stands are based on a "reasonable-to-surmise" doctrine. Furthermore, they promote high rates of fluid replacement during exercise, sufficient to replace sweat losses (without ever referring to urine losses), but without providing a scientific validation for that conclusion. Again, the doctrine of "potential for prevention" is invoked. The position stands acknowledge that high exercise intensities are more likely to produce heatstroke regardless of the levels of dehydration that develop.
Confounder #2. The second factor that, in my personal view, has increased the confusion of the role of fluid ingestion during exercise, has been the growing academic influence of certain sports-related industries. For the reality is that nothing will convey a message more powerfully than commercial advertising, which is more likely both to reach the athletes and to have a more lasting impression than are the more cautious and less widely distributed statements of the scientists.
For example, Gatorade's Web site repeats one of the key ACSM guidelines: "The goal of fluid replacement during exercise should be to fully replace sweat losses. The physiological and performance benefits of doing so are well documented" (16). Another publication from the Gatorade Institute states, "Dehydration can imperil health by increasing the risk of heat illnesses such as heat exhaustion and heatstroke" (17). The former document (16) also includes a statement that implies that the ACSM guidelines, which specifically promote the ingestion of sports drinks, are the reason why heat injuries have become less common: "Although athletes and others continue to fall prey to exertional heatstroke, the frequency of such deaths has been dramatically reduced over the years (18), in large part because the necessity of adequate fluid replacement has become well recognized."
This statement can have no factual basis since neither the historical nor the current incidence of heatstroke in "athletes" is known. Nor is this conclusion supported by any specific data in the reference cited (18). This apparently informed opinion can only entrench the belief—among a much broader audience of athletes and coaches without access to the ACSM position stands—that complete replacement of sweat losses during exercise is essential to avoiding heat injury.
Indeed, as I have argued in the disputed article (19), the hyponatremia of exercise is always either a self-inflicted, or, less commonly, an iatrogenic condition, that occurs when affected athletes and the physicians who treat them follow this informed opinion to its literal conclusion. I have been unable to understand why the guidelines for fluid replacement during exercise (4) fail to include the warning that drinking too much is dangerous. This is particularly surprising since the most recent position stand (3) includes the statement that: "Excessive consumption of pure water or dilute fluid during exercise may lead to a harmful dilutional hyponatremia," but the message is distorted in the position stand on exercise and fluid replacement (4), which concludes only that "One major rationale for inclusion of sodium in rehydration drinks is to avoid hyponatremia."
A probable result of this subtle distortion is to promote the unproven concept that the ingestion of sports drinks during exercise is more beneficial than water because, unlike water, any sodium-containing fluid ingested during exercise will prevent the development of hyponatremia. In contrast, these position stands should really warn that it is the gross overconsumption of any fluid—even those containing sodium in the relatively low concentrations found in popular sports drinks—that causes hyponatremia (19-21).
Confounder #3. The final possible reason for confusion is that the diagnostic features of the different exercise-related heat illnesses include just about anything anyone ever wants to include (22,23). In particular, those with special expertise in laboratory research but who may not treat athletes with heat illnesses describe diagnostic criteria that, in my clinical experience, are not particularly helpful for managing collapsed athletes. For example, in their letter, Drs Casa and Armstrong list what they consider the "obvious signs and symptoms of heat exhaustion." But their "obvious" signs and symptoms are so nonspecific that they are without any clinical value whatsoever. If one were to diagnose heat exhaustion on the basis of their extensive list of nonspecific signs and symptoms, the condition would be overdiagnosed, and conclusions of the effects of fluid ingestion on the incidence of heat exhaustion, so defined, would be quite meaningless. Later in their letter, Casa and Armstrong extend their definition by suggesting that any degree of "heat storage" during exercise should be considered a heat illness.
In contrast, the view of a clinician treating collapsed runners might be that a rise in body temperature to above 40°C during exercise is, by itself, not abnormal since such values are frequently measured in the top finishers (8,24-26). Furthermore, rectal temperatures in excess of 40°C can be induced quite readily by vigorous running in hot conditions for short durations (26) so that dehydration cannot be a factor.
Thus, unless the core temperature is raised above 40.5°C and is associated with specific symptoms, most particularly a change in cerebral function that persists when the patient is lying supine with the legs elevated, I would caution that a diagnosis of heat illness, specifically heatstroke, is not appropriate (22,23). Nor would many who measure rectal temperatures frequently in healthy athletes completing short distance races agree with Casa and Armstrong's suggestion that the threshold for heatstroke is 39° to 40°C. My personal impression is that rectal temperatures become a concern only when they are above 41°C in athletes who exhibit the other diagnostic features of heatstroke.
Because the classically defined symptoms of the so-called heat illnesses are, with the clear exception of heatstroke, nonspecific and are not associated with rectal temperatures even close to those measured in patients with heatstroke (28), I have proposed that patients with these nonspecific symptoms should not be diagnosed as suffering from "heat" illnesses (22,23). Rather, any definition should consider that postural hypotension may be the important pathophysiologic abnormality common to all these so-called "heat" illnesses (23,29,30).
The definition of what exactly constitutes the heat illnesses has been further confused because some physicians who provide medical care at endurance events have not been exposed to the more subtle complexities of the research findings in this field. Rather, they assume that the opinions of the laboratory-based experts are directly applicable to the clinical conditions that they treat, as does Dr Schriner in his letter. As a result, the diagnostic classification of the heat illnesses and the related treatment protocols are seldom challenged by those clinicians who provide medical care at endurance events.
I have been fortunate to experience the three sides of this debate, first as an endurance athlete, next as a clinician, and finally as a laboratory-trained scientist. I began marathon running in 1972 in the era when runners were still being advised not to drink during exercise (5). At that time, watering stations were provided at one or occasionally two points during marathon races, and there were usually no medical facilities to treat collapsed athletes. My first published articles (31-33) emphasized the need to provide more fluid during such races. By 1976, I was involved in the medical care of athletes after marathons and ultramarathons in South Africa. This led, over the next two decades, to a series of studies (29,30,34-36) of the etiology of postexercise collapse in endurance athletes. These studies have, at least in my biased opinion, helped to develop a rational, more scientifically based approach for the management of collapsed endurance athletes (23,30,37).
These management principles are followed at the 226-km South African Ironman Triathlon, of which I am currently the medical director. In the inaugural (2000) race, the weights of all competitors were recorded before and after the race, as were their blood sodium concentrations and their postrace rectal temperatures. In addition, all competitors were examined medically after the race to evaluate the efficacy of the preventive measures introduced to safeguard the health of the athletes.
The results suggest that, perhaps because the triathletes were advised to drink no more than between 500 to 800 mL per hour during exercise, the race was one of the safest Ironman triathlons on record. The percentage of finishers requiring medical care, all for trivial conditions, was 7%, compared with a more common frequency of 20% to 30% in the Hawaiian (38) and New Zealand (39) Ironman triathlons. Furthermore, only 3 triathletes (1%) finished the race with serum sodium concentrations below 135 mmol/L, diagnosed as asymptomatic hyponatremia. This is the lowest reported incidence in any Ironman triathlon of which we are aware and compares with values of 29% and 18% in the Hawaiian (40) and New Zealand (39) Ironman races, respectively.
I emphasize these points to indicate that I have at least some practical experience about the scientific issues about which I write. I return now to the exact issues raised in the different letters.
What is particularly remarkable about the letter from Casa and Armstrong is that, from my entire article, they select just one sentence that they consider inaccurate and have apparently spent months trying to correct. Yet they actually agree with what I have written. To quote: "It may be true that endurance athletes who participate at very low exercise intensities (most of the athletes competing in ultra events) would be at low risk (but not no risk) of adverse medical effects associated with the stated degrees of dehydration".
Their acceptance of the truth of my statement is especially important because Armstrong is listed as one of the principal authors of two ACSM position stands (3,4). My understanding is that the ACSM position stands focus principally on endurance events, especially races of 10 to 42 km in which the exercise intensity will be lower than in short distance races that last less than 30 minutes, in which significant dehydration is also unlikely so that dehydration cannot be a factor influencing the risk of heat illness. If my arguments are accepted, then future ACSM position statements should acknowledge that dehydration of up to 8% (at least) may have no medical consequences during more prolonged exercise when the exercise intensity is appropriate for the duration of the exercise and the environmental conditions are within those prescribed as of either low or moderate risk according to the ACSM guidelines (3).
Indeed, I offer data (table 1) that strongly support the correctness of this new interpretation of Drs Casa and Armstrong. These data are from all athletes whose rectal temperatures and body weight changes were measured in the 2000 South African Ironman Triathlon. Table 1 includes data from individual athletes who completed either the running or the cycling legs as part of relay teams, who competed at the same time as individual triathlon competitors.
A total of 201 athletes, to our knowledge the largest such group yet reported, in the different disciplines (triathletes, cyclists, runners) are grouped according to the percentage of weight lost during the race, a surrogate measure of dehydration. Thus, athletes in group 1 lost between 12% and 8% of their starting body weight during the race (mean, -8.4%); group 2 lost from 8% to 4% - (means, -5.0% to -5.6%); and group 3 lost from 4% to 1% (means, -2.8% to -2.6%). Seven triathletes but no cyclists or runners (group 1) reached levels of dehydration in excess of 8%. Yet regardless of their sporting discipline or the level of dehydration during the race, there were no differences in the rectal temperatures measured within 2 to 3 minutes of the athletes' crossing the finish line between group 1, 2, and 3. Indeed, mean rectal temperatures were low in all groups of athletes; the highest value of 40.5°C was measured in the athlete who finished second in the individual triathlon after he had led the race until the last kilometer. That athlete developed profound postural hypotension (85/50 mm Hg) at the finish line but was discharged fully recovered, within 20 minutes of lying supine in the head-down position with appropriate cooling. His data are not included in table 1 as he chose not to participate in the weighing section of the trial.
Hence, whatever the findings in the laboratory, in cross-sectional studies in actual competition, there is no relationship between the level of dehydration reached during the race and postrace rectal temperature. This confirms our previous conclusion from earlier field studies (25,26) involving many fewer athletes. In fact, the rectal temperatures reached by even the most dehydrated triathletes, runners, and cyclists were no higher than those of the least dehydrated athletes despite the fact that, in each sporting discipline, the most dehydrated group finished the race faster than those who were less dehydrated (data not shown). The finding that the more dehydrated athletes are also the faster finishers is frequently observed. Nor were postexercise rectal temperatures anywhere near the range that might suggest risk of heatstroke as also found, but not recognized, in the study of Wyndham and Strydom (8,15).
Thus, these data do not support the conclusion, now also included in the recent position statement of the National Athletic Trainers' Association, that all athletes involved in prolonged exercise should, or need to, "minimize fluid losses by encouraging fluid replacement equal to fluid losses." This may indeed by a "time-honored" dogma, but it has no scientific basis—at least if it is advocated on the grounds that this practice is necessary to prevent hyperthermia, heatstroke, or exercise-associated collapse during prolonged exercise, including marathons and Ironman triathlons, run at usual intensities and in the environmental conditions considered safe according to the ACSM guidelines. I submit that, if we wish to ensure that ACSM position stands do no harm, then the authors of future position stands should, at the outset, ensure that their conclusions are first evidence-based. Then, if they become time-honored, they are at least more likely also to be correct.
The gist of Casa and Armstrong's criticism is that my conclusions, here confirmed (see table 1), are not relevant to higher exercise intensities or when the exercise is undertaken in severe environmental conditions, "where the intensity is great enough to overwhelm the thermoregulatory system," especially if the efficiency of that system is impaired by what they consider are the detrimental effects of dehydration. These specified conditions—high-intensity exercise of short duration in extremely hot environmental conditions—are, of course, precisely opposite to those favoring the development of hyponatremia, which was the focus of the original article (19).
But to "prove" that my conclusions are "damaging" if applied to events of short distance and high intensity, they are forced to introduce two variables that have no relevance in the real sporting world and that destroy their argument, thereby nullifying their criticisms.
First, they argue that laboratory-based studies prove that high levels of dehydration are dangerous during exercise in extreme environmental conditions. The specific environmental temperatures needed to prove this effect in two separate studies were 35°C and 42°C. But the ACSM guidelines do not condone competitive exercise under such conditions. Rather, the guidelines indicate that the risk of heat injury would be "very high" (3) at a dry-bulb temperature of 42°C, even when the relative humidity is only 15%. Similarly at any humidity greater than 50%, a dry-bulb temperature of 35°C would also constitute "very high" risk.
Hence, Casa and Armstrong are guilty of trailing a red herring by quoting those laboratory studies as their "proof" that dehydration increases the risk of heatstroke during exercise—for the simple reason that athletes should never be allowed to compete under those extreme environmental conditions without the potential risk of legal consequences. I assume that Casa and Armstrong do not wish to promote participation in exercise under conditions that are unacceptable to the guidelines established by a committee of which Armstrong was a member. What their diversionary argument effectively confirms is that it is the environmental conditions, not the level of dehydration, that determine the risk of heat injury during exercise.
The second reason why the criticism of Casa and Armstrong has no practical validity is because, although they now acknowledge that levels of dehydration of up to 8% are not associated with documented medical risks during prolonged exercise, they argue that those levels of dehydration would be of dire consequence during exercise when the exercise intensity is "great." But the practical relevance of that claim completely escapes me. For example, a 70-kg athlete could only become 8% dehydrated if he were to exercise in usual environmental conditions at a sufficiently high intensity to sustain a sweat rate of 1.6 L/hr while ingesting no fluid for at least 3.5 hours. None of the runners or cyclists in the South African Ironman Triathlon was able to reach levels of dehydration of 8%, for example, despite exercising for up to 6 hours, although some of the faster triathletes achieved that level (see table 1).
I doubt that there is any athlete in the world able to sustain a "great" exercise intensity for 3.5 hours (and without drinking) in order to reach the point where his or her level of dehydration would be sufficient to place the athlete at theoretical risk of heatstroke. Surely after the 3.5 hours of exercise necessary to introduce a "significant" or "severe" level of dehydration, his or her previously "great" exercise intensity would have dropped to approximate that of the athletes reported in table 1 and in whom no "dangerous" effects of severe levels of dehydration were apparent, since postrace rectal temperatures were independent of the level of dehydration.
Their additional statement that "athletes who exercise at high intensities and wear uniforms in hot environments are at great risk for dangerous hyperthermia" is obvious. Whether being dehydrated to "-8%" alters that risk is not established but is irrelevant for the very reasons detailed above. The danger in those circumstances is posed by the environmental conditions and the clothing that is worn, not by the extent of any dehydration. Prevention of risk requires that the exercise and the clothing, not the dehydration, be avoided until more usual environmental conditions again prevail.
Casa and Armstrong are also guilty of incorrect and selective quotation from the classic studies of Adolph, cited in Brown (41). They use that reference to support their argument that "someone who runs or cycles in the heat" at "a larger dehydration level" will develop the nonspecific symptoms of heat illness that they list. However, Adolph did not study cyclists and runners but army personnel who walked for 8 or more hours in desert heat while they were denied access to fluid (9). Furthermore, there were two important features of that study that are never quoted because they are so detrimental to the argument that dehydration is especially dangerous because it increases the risk specifically of heatstroke during prolonged exercise.
First, the common physiologic abnormality that the subjects developed was not hyperthermia, but postural hypotension. Rectal temperatures at exhaustion were seldom higher than those we measured in our healthy triathletes at the completion of the Ironman (see table 1). The other signs and symptoms listed by Casa and Armstrong are nonspecific and by themselves would not explain the physical incapacitation of the military personnel when they became dehydrated by more than 8%.
Second, the really telling conclusion from that study and one that can never be quoted because it makes a mockery of the entire argument that "significant" dehydration carries grave health risks, was that "there were no obvious after effects of dehydration.... We do know that man can suffer a water deficit so incapacitating that he can neither walk nor stand; yet he recovers quickly within a few minutes of water ingestion, and his feelings of well-being within half an hour or less after he begins drinking. With a meal or two intervening, his recovery is practically complete in 6 to 12 hours" (42).
A final comment regarding that letter is appropriate. I have considered why carefully conducted laboratory studies by expert scientists consistently show that the rectal temperature rises in proportion to the degree of dehydration (7,43-45) during exercise in the heat, yet as shown in our other studies (25,26) and confirmed in the data presented in table 1, no such relation is found in field studies.
One key difference is that rates of body cooling by convection may be different in laboratory studies and field trials performed out of doors. In addition, most of those laboratory studies were performed in environmental conditions that would be rated "high risk" by the ACSM guidelines.
For example, the facing wind speed, the main determinant of convective heat loss, in those four laboratory studies was either 0.3 km/hr (7), 3.6 km/hr (45), 5.7 km/hr (43), or 9 km/hr (44), and it is not clear whether such cooling affected the whole body. All of these are well below the wind speeds generated by athletes, especially cyclists, exercising at the "great" intensities necessary to cause heat injury according to the theory of Casa and Armstrong. Furthermore it is known that convective heat loss increases with increasing wind speed (46). Thus, the convective cooling provided in those laboratory studies was less than that found in usual exercise conditions and would have substantially affected the athlete's ability to lose heat normally.
Furthermore, the environmental conditions present in the two studies that most clearly establish a proportional effect of dehydration on the elevation of rectal temperature were performed in environmental conditions (33°C, humidity 50% (44) and 33°C, humidity 64% (45)) that constitute "borderline very high" and "very high" risk, respectively, of heat injury, according to the ACSM guidelines (3). Exercise physiologists are welcome to study physiologic phenomena under environmental conditions in the laboratory that have little practical application to competitive sport. But they should exhibit sufficient clinical judgment not to extrapolate their data to the quite different environmental conditions under which competitive sport can be safely conducted (3).
Hence, it makes no sense to quote these studies as proof of the "dangers" of dehydration when they were performed under the very environmental conditions that the ACSM guidelines describe as unsafe for competitive sport. Surely there should be a consistency of logic in the pronouncements of the ACSM?
Perhaps these criticisms explain why, when exercise is performed under more usual, competitively safe environmental conditions in the laboratory (25°C, relative humidity 55%, and facing wind speed 13 to 15 km/hr in a properly constructed wind tunnel so that the athlete's entire body was subject to rapidly moving air), drinking fluid according to the ACSM guidelines did not offer any performance advantages over ad libitum ingestion (47). Rectal temperatures were not measured because the emphasis in that trial was on athletic performance. Indeed, two other studies (48,49) have failed to show that high rates of fluid ingestion during exercise are better than "ad libitum" intakes or those that either replace sweat rates or are "as much as can be tolerated" (4).
Dr Schriner is quite at liberty to be incensed by my article, but he should be wary lest his hostility affect his scientific and clinical judgment. Points that might assuage his anger include the following:
First, I suspect that Drs Costill and Gisolfi seldom ventured near medical tents at the finish of sporting events and most certainly were not qualified to treat patients with heatstroke in those medical tents. I am unaware of any of their "consensus opinions of hyperthermia as relating to dehydration and vascular shifts," nor of any of their articles referring to the management of heatstroke. It is for this reason that I was unable to quote their (nonexistent) studies.
As indicated earlier in this article, the scientific contributions of Costill and Gisolfi that are quoted in the ACSM position stands were laboratory-based and showed only that ingesting fluid during prolonged exercise in hot conditions in which convective cooling was inadequate caused the rectal temperature to be reduced by about 0.7°C during 2 hours of exercise (7,43). Mean rectal temperatures at the termination of exercise were below 40°C in both studies; no athletes developed any medical complaints during exercise regardless of whether or not they ingested fluids. I was unable to find reference to the words "heat injury, heat cramps, or heatstroke" in either article. Therefore it is not clear to me exactly how these studies relate to the prevention or management of heatstroke in competitive athletes racing out-of-doors. Nor were the studies ever designed to measure the effects of fluid ingestion on the incidence of heat injury.
Had I been able to find any studies that established that connection, I am sure I would have quoted those studies and any others that contest my previous conclusion (9). However, as detailed above, there are no such studies, so the ACSM guidelines—at least as they relate to fluid ingestion and the prevention of heat injury in runners competing in usual environmental conditions—are not evidence-based.
Second, Schriner is at liberty to treat heatstroke as a "fluid volume depletion" condition, but he must do so at both his own risk and that of his patients. The work of Costill and Gisolfi—however certain he may be of their findings—will provide him no legal comfort should he require it because their work has no bearing whatsoever on the treatment of heatstroke. Perhaps he would be better advised to follow the ACSM guidelines (3) and others (23) that advocate rapid whole-body cooling as the treatment of choice in heatstoke. Interestingly, the ACSM guidelines are silent on the role of IV fluids in the management of heatstroke, which is particularly surprising if dehydration is considered to be such an important cause of heatstroke.
Dr Mink is correct that the physiologic abnormality causing the hyponatremia of exercise is an inadequate renal response to high rates of fluid ingestion during very prolonged exercise, leading to fluid retention (19,20,50). Indeed, the remarkable finding is the inability of affected athletes to excrete what in some is a massive fluid overload of up to 6 L (8% of body weight) (50). Their inability to excrete this fluid excess continues for some hours after exercise. The logical conclusion is that there must be an inappropriate secretion of hormones that prevents the normal renal response to fluid overload.
We suspect that this is indeed due to a circulating factor, the blood concentrations of which must decay quite slowly after exercise. Our clinical experience is that patients with this condition pass urine relatively slowly for the first few hours after exercise, and even their peak rates of diuresis during recovery are likely to be less than 600 mL/hr. In contrast, when tested under resting conditions, athletes with a history of exercise-related hyponatremia can reach peak rates of diuresis of up to 900 mL/hr in response to lesser degrees of fluid overload (51). But plasma ADH concentrations have not been shown to be elevated in athletes with the hyponatremia of exercise (40,52). Hence, we propose that the most likely explanation is that another hormone or hormones—perhaps as yet unidentified or when acting in concert—cause the abnormal renal response that leads to the abnormal fluid retention that causes the hyponatremia of exercise.
Perhaps the important clinical point is that, whereas at rest athletes can maintain normal fluid balance provided that their fluid intake is less than 900 mL/hr (51), during exercise the peak rates of urine production fall in some individuals. If they then ingest fluid at rates that exceed their combined rates of sweat and urine production, they will develop symptomatic hyponatremia if their total fluid excess exceeds about 4% of body weight.
Dr Stoddard argues the popular belief that it is the failure to replace sodium losses, rather than fluid overload, that causes the hyponatremia of exercise. The ambiguity in the ACSM position stand (4), described above, contributes to that incorrect conclusion.
In fact, all the scientific evidence suggests that sodium losses are not increased in athletes who develop symptomatic hyponatremia during exercise (50,52). With regard to sodium replacement during exercise, there appears to be little benefit if only half of the fluid loss is replaced (53), as is the usual practice. If there is full fluid replacement during exercise, partial sodium replacement reduces the extent, but does not completely prevent a fall, in the serum sodium concentration during exercise (54). Furthermore in that study, the extent of the diuresis produced by high rates of fluid ingestion during exercise was the best (inverse) predictor of the extent to which the serum sodium concentration fell during exercise, and not the relatively small amounts of sodium ingested from the sports drink. Unfortunately, this important distinction was not reflected in the title of that article (54), which continues to reflect the bias of the ACSM position stand (4). Hence, it is the ability to mount an appropriate diuretic response in the face of a high rate of fluid intake that protects against hyponatremia.
But if there is full replacement of both sodium and water losses during exercise, the serum sodium concentration increases and the plasma volume expands (55). As yet there are no studies of full sodium replacement with excessive fluid replacement; therefore, the possible role of full sodium replacement in the prevention of hyponatremia due to fluid overload is not established, despite the misleading inference in the ACSM position stand on exercise and fluid replacement, described earlier.
However, subjects in the South African Ironman Triathlon who ingested sodium in quantities approximating their losses generally reported positive effects. Thus, the ingestion of higher doses of sodium, especially during very prolonged exercise, may yet prove to be of some benefit even if not through purely physiologic effects on fluid and sodium balance. We are actively studying this possibility.
Stoddard repeats the popular belief that dehydration is still the "No. 1" concern during exercise in the heat and that dehydration alters the body's ability to maintain heat balance. The contrary evidence has been presented at the beginning of this response, including the data reported in table 1. Rather, the clear evidence is that fluid overload is the real potential killer in modern endurance sports (20, 21). For example, Speedy et al (21) have reviewed 57 cases of exercise-associated hyponatremia reported in the medical literature since 1985. These do not include at least 3 more recent fatal cases (56-60). I have been unable to locate, anywhere in the published scientific literature, a similar body of evidence showing a greater number of equally serious cases resulting solely from "dehydration" during the same period of observation.
On behalf of all the authors, I thank Dr Pathak for his complimentary comments. Certainly, lack of preacclimatization to heat is recognized as an important risk factor for heatstroke and is so identified in the ACSM position statements (2,3). Such preacclimatization does indeed reduce sodium losses in sweat and would therefore theoretically influence the risk of developing hyponatremia during exercise. However, the evidence is that the hyponatremia of exercise is due to fluid overload rather than sodium losses, so that any effect of acclimatization on sodium losses during exercise would not likely prevent hyponatremia in athletes who chose to follow inappropriate guidelines and to drink to excess during prolonged exercise.
Pathak's point that hyponatremic convulsions are well recognized in medical conditions associated with fluid overload is well taken. Two points are relevant. First, that high rates of IV fluid therapy are sometimes given to patients undergoing certain surgical conditions in the belief that such fluid assists in the maintenance of cardiac function and systolic blood pressure. The same logic explains why IV fluids were introduced early as the treatment of choice for the management of postexercise collapse (61). Yet it is the reduced peripheral vascular resistance, rather than an impaired cardiac filling secondary to fluid loss, that more probably explains the hypotension in both conditions. Hence, the more logical medical intervention in both conditions should be to offset the cardiovascular effects of reduced peripheral vascular resistance.
Second, perhaps if urologists had been involved in the early management of cases of postexercise collapse, including those due to symptomatic hyponatremia, they would have alerted sports physicians earlier to the dangers (19) of giving IV fluids to collapsed or unconscious patients whose serum sodium concentrations are not known.