FLUID LOSS The rate at which sweat is secreted onto the skin depends on the need to lose heat,

which is determined mainly by the rate of heat production and ambient temperature (Maughan & Noakes, 1991). Heat produc-tion is directly proportional to the intensity of exercise where for sports like running and cycling, heat production is therefore a function of speed. For most other sports however, the work intensity is not constant, but consists of rest, low and high intensity exercise periods.

In a marathon race the faster runners lose sweat at a higher rate than the slower runners, but on the other hand they are active for a shorter period of time, so the total sweat loss of these marathon runners is unrelated to finishing time.Among competitors in all sporting events, it is obvious that the rate of sweat secretion varies considerably between individuals. Some people sweat profusely, and a large part of the sweat produces drips from the body surface without evaporating. This may be considered an inefficient sweating mecha-nism: the rate of sweat secretion is not directly linked to the rate ofheat production. The reasons for this excessive sweat secretion in some individuals are not clearly understood. Marathon runners competing under the same conditions and with the same fluid intake at low (10C) ambient temperature may lose from as little as 1% to as much as 6% of the body weight during a race (0.74.2-kg of body mass for a 70-kg man) (Maughan & Miller, 1984). At high ambient temperatures, sweat losses equi-valent to 8% body weight may occur in marathon runners: this amounts to 56L of water for a 70-kg individual (Maughan & Shirreffs 1998).

A loss of body water corresponding to as little as 2% of body weight can result in some impairment of exercise tolerance (Armstrong, Costill & Fink, 1985; Maughan & Shirreffs, 1998; Nielsen, Sjogaard & Boude-Petersen, 1984; Walsh et al., 1994;) and losses in excess of 5% of

body weight can decrease the capacity for work by about 30% (Saltin & Costill, 1988). The fluid losses are distributed in varying proportions between the plasma, the extracellular water and the intracellular water. The reduction of plasma volume may be of particular significance and could have a significant impact on perfor-mance. Blood flow to the muscles must be maintained throughout the exercise period to supply oxygen and substrates to the exercising muscles and to remove carbon dioxide and other waste products. In addition to that, blood flow to the skin must also be adequate to carry heat to the body surface where it can be dissipated. If the blood volume decreases due to fluid loss, there may be difficulty in meeting the requirements of a high flow rate to both muscle and skin. Decreases in plasma volume may also result in significant increases in the viscosity of the blood, forcing the heart to work harder to pump it around the circulatory system (Costill & Fink 1974; Fortneyet al., 1981; Fortney et al., 1984; Greenleaf et al., 1977; Horstman & Horvath 1972; Nadel, 1980; Nadel,Fortney & Wenger, 1980; Nielsen, 1974, Saltin & Costill, 1988; Saltin & Stenberg 1964). Dehydration may therefore be more important than substrate depletion in causing fatigue during prolonged exercise, particularly in hot weather where fluid losses are high and where it is not possible to replace the losses during the exercise period. Performance in events requiring brief bursts of very high intensity exercise may be affected more than events which, involve prolonged relatively low intensity exercise. When the ambient temperature is high and blood volume has been decreased by sweat loss during prolonged exercise, there may be difficulty in meeting the requirements for a high blood flow to both the working muscles and skin. In this situation also, skin blood flow is likely to be compromised, allowing central venous pressure and muscle blood flow to be maintained but reducing heat loss and causing body temperature to rise (Rowell, 1986) and perhaps contributing to the fatigue

EFFECT OF ENVIRONMENT AND TRAINING STATUS ON FLUID BALANCE

The major influence on the physical exchange ofheat between the body and the environment is the ambient temperature. When ambient temperature exceeds skin temperature, heat is gained from the environment by physical transfer, leaving evaporative loss as the only mechanism available to prevent or limit a rise in body temperature. Hence, increased sweating rate in the heat will result in an increased requirement for fluid replacement. A precaution which can be taken to prevent sweat loss, and hence reduce the need for fluid replacement is to limit the extent of warm-up prior to the competition or reduce the clothing worn.

When the humidity is high and particularly whenthere is no wind, evaporative heat loss will be severely affected which, will then limit exercise tolerance by dehydration and hyperthermia rather than by the limited carbohydrate stores in the body. Suzuki (1980) reported that exercise time at 66% VO2maxwas reduced from 91 min when the ambient temperature was 0C to 19 min when the same exercise was performed at 40C.

It is recognised that training will provide some protection against the development of heat illness during exercise in the heat. This adaptation is most marked when training is carried out in a hot environment (Senay, 1979). One of the benefits oftraining in a hot environment is the expansion of plasma volume (Hallberg & Magnusson, 1984), which occurs within a few hours of completion of exercise and may persist for several days (Davidson et al., 1987; Robertson, Maughan & Davidson, 1988). This post-exercise hypervolaemia should be regarded as an acute response rather than an adaptation, although it may appear to be one of the first responses to occur when an individual starts a training programme.

This increased resting plasma volume in the trained athlete allows the endurance-trained individual to maintain a higher total blood volume during exercise (Convertino, Keil &

Greenleaf, 1983) allowing for better maintenance of cardiac output at the cost of a lower circulating haemoglobin concen-tration. In addition, the increased plasma volume is associated with an increased sweating rate which limits the rise in body temperature (Mitchell et al., 1976). The increase in plasma volume increases progressively over the first 6 days, reaching a value of about 23% greater than the control, with littlechange thereafter (Mitchell et al., 1976; Senay, Mitchell & Wyndham,1976). The main adaptation in terms ofan increased sweating rate and an improved thermoregulatory response occurs slightly later than the cardiovascular adaptations (Wyndham et al., 1976). With training, better maintenance of body temperature is achieved at the expense of an increased sweat rate. Although the increased sweat rate allows for a greater evaporative heat loss, the proportion of the sweat which is not evaporated and which therefore drips wastefully from the skin is also increased (American College of Sports Medicine, 1996). Although a high sweat rate may be necessary to ensure adequate evaporative heat loss, unfortunately many individuals have an inefficient sweating mechanism.

EFFECT OF DEHYDRATION ON EXERCISE PERFORMANCE Numerous studies have examined the influence of dehydration on maximal aerobic power and physical work capacity (Sawka, Montain & Ladzka, 1996). The physical work capacity for aerobic exercise of progressive intensity is decreased when a person is dehydrated (Sawka et al., 1996). Physical work capacity has been shown tobe decreased even with marginal (12% body weight loss) water deficits and the reduction is larger with increasing water deficits (Armstrong et al., 1985; Caldwell, Ahonen & Nousiainen, 1984). Dehydration results in much larger decrements of physical work capacity in hot thanin temperate climates (Buskirk, Lampietro & Bass, 1958; Caldwell et al.,1984; Craig & Cummings, 1966; Webster, Rutt & Wettman, 1990). In a temperate environment, less than 3% body weight loss appears not to alter maximal aerobic power. Maximal aerobic power decreased when dehydration equalled or exceeded 3% body

weight. In a hot environment on the other hand, Craig and Cummings (1966) demonstrated that small (2% body weight) to moderate (4% body weight) water deficits resulted in large reductions of maximal aerobic power. Even at low levelsof dehydration (1.8%), high-intensity (90% VO2max) exercise performance time is reduced (Walsh et al., 1994). It is believed that the thermoregulatory system, via increased body temperature plays an important role in the reduced exercise performance mediated by a body water deficit. Table 1 presents a summary of investigations concerning the influence of dehydration on maximal aerobic power and physical work capacity.

On the effects of dehydration on physiologic tolerance to submaximal exercise, Adolph (1947) reported that 16% and 2% of the soldiers suffered exhaustion from heat strain during an endurance (2-23 h) desert walk (at 4-6.5 km/h) (T a ~38C) when they did not drink and did drink water ad libitum, respectively. Ladell (1955) had subjects attempt 140-min walks in a hot (T a ~38C) environment while ingesting different combinations of salt and water.They reported that exhaustion from heat strain occurred in 75% oftheir subjects when not receiving water and in 7% of their subjects when receiving water. Clearly, dehydration increases the incidence of exhaustion from heat strain. Sawka et al.(1985) had subjects attempt lengthy treadmill walks (~25% of VO2maxfor 140 min) in a hot-dry (T a = 49C, rh = 20%) environment when euhydrated and when dehydrated by 3%, 5% and 7% of their body weight. All eight subjects completed the euhydration and 3% dehydation experi-ments, and seven subjects completed the

5% dehydration experiments. In the 7% dehydration experiment, six subjects discontinued after completing only 64 min (mean). Clearly dehydration increases the incidence of exhaustion from heat strain.

Dehydration also impairs competitive middle distance running performance in athletes. Armstrong et al. (1985) had athletes compete in1500-, 5000- and 10000-m races when euhydrated and when dehydrated. Dehydration (2% body weight loss) was achieved by diuretic administration, which decreased plasma volume by 11%. Running performance was impaired at all distances but more in the longer races (~6% for the 5000 and 10000m) than in the shortest race (3% for the 1500m). Burge, Carey & Payne, (1993) who examined simulated 2000-m rowing performance, found that it took the rowers anaverage of 22s longer to complete the task when they were dehydrated than when they were euhydrated. In addition, dehydration reduced average power by 5%.

In more prolonged exercise, sweat rates of 2-3 L h-1 are possible. The combined effects of progressive dehydration and a rising body temperature pose a considerable threat to the runner. During a marathon race even at high ambient temperatures, runners may lose as much as 8% of body weight, corresponding to about 13% of total body water (Costill, 1972). Even in an event as short as 10km, losses corresponding to more than 2% of body weight are possible. Many of the highest values of rectal temperatures in distance runners have been observed after races at distances less than the marathon. Among competitors in a 14-km road race, Sutton (1990) reported more than 30 cases over a period of years where rectal temperatures exceeded 42C. From these cases, hyperthermia may be more common when the rate of heat production is very high as in races of 10km (England et al., 1982). At such high exercise intensities, skin blood

flow is likely to be reduced with a larger fraction of the cardiac output being directed to the working muscles, and so heat loss will be reduced. For an extensive review on the factors influencing fluid loss during exercise, and the effects of dehydration on exercise performance, see Lamb and Brodowicz (1986) and Maughan and Shirreffs (1998).

FLUID REPLACEMENT DURING EXERCISE Physical working capacity is impaired when fluid losses cause even small reductions in plasma volume (Saltin, 1964). In prolonged exercise, the maintenance of plasma volume may be of major importance. A progressive rise in heart rate normally occurs during exercise at a constant work load: this reflects the increased cardiac output necessary to meet the requirements for an increased skin blood flow to promote heat loss while maintaining the supply of oxygen and substrates to the working muscles (Rowell, 1974). The redistribution of regional blood flow includes a decreased flow from the intestines tothe liver, which may reduce the rate of gastric emptying and intestinal absorption. The maximum rate at which water can be absorbed from the gut during strenuous exercise may therefore becomeinadequate to match the high rates of sweat loss which can occur (Costill, Kammer & Fisher,1970).

The ability to perform prolonged exercise in the heat requires replacement of water losses to prevent dehydration. Numerous studies have shown that water intake during prolonged exercise is effective in improving performance and indelaying the onset of fatigue (Armstrong et al., 1985, Below et al., 1995; Costill et al., 1970; Montain & Coyle, 1992; Sawka & Pandolf, 1990; Sawka et al., 1984; Sawka, 1992). To sustain a high rate of work output or exercise performance in heat requires replacement of water losses toprevent dehydration, but exercise performance may also be limited by the availability of carbohydrate as fuel for the working muscles.

Carbohydrate-electrolyte fluid ingestion during exercise therefore, has dual functions of supplying water to replace the losses incurred by sweating and of providing a source of carbohydrate fuel to supplement the bodys limited stores (Lamb & Brodwicz, 1986; Co yle & Coggan, 1984; Maughan & Shireffs, 1998; Murray, 1987; Singh et al., 1997; Singh et al., 1995; Tsintzas et al., 1995). Below et al. (1995) have shown that ingestion of water and carbohydrate have independent and additive effects on exerciseperformance. However, in one study, water replacement at a rate of 100ml every 10 min did not improve endurance during cycle exercise at 70%VO2max, whereas ingestion of 4% glucose-electrolyte solution did significantly extend further the exercise time (Maughan, Fenn & Leiper, 1989). The rates at which substrate and water can be supplied during exercise are limited by the rates of gastric emptying and intestinal absorption. It is not clear which of these processes is limiting. It is commonly assumed that rate of gastric emptying is the limiting factor and thatit determines the maximum rates of fluid and substrate availability (Lamb & Brodwicz 1986; Murray, 1987). Increasing the carbohydrate content of drinks will slow the rate of gastric emptying (Costill & Saltin, 1974), thereby decreasing the rate at which fluid loss can be replaced. The presence of glucose and sodium in the lumen of the small intestine stimulates water absorption, provided that the osmolality of the solution is not high (Leiper & Maughan, 1988). On the other hand, increasing the carbohydrate content of drinks will increase the amount of fuel which can be supplied, but will also tend to decrease the rate atwhich water can be made available. Where provision of water is the first priority, the carbohydrate content of drinks will be low, thus restricting the rate at which substrate is provided. The composition of drinks to be taken will thus be influenced by the relative importance of the need to supply fuel and water; this in turn depends on the intensity and duration of the exercise task, the ambient temperature and humidity and the physiological and biochemical characteristics of the individual athlete. When water replacement is the first priority, an isotonic or moderately hypotonic solution containing glucose

and sodium will be most effective (Farthing, 1988). Most commercial sports drinks contain 6 8% carbohydrate, about 2025 mmol.L-1 sodium and a low (45 mmol.L -1 ) concentration of potassium. These formulations represent a compromise between that which will give the highest rates of fluid replenishment and that which canprovide most carbohydrate. From a summary of all these studies, it has been reported that similar rates of gastric emptying are seen for beverages in which the carbohydrate content ranges from 0 to 10% (Maughan, 1997). Using a wide variety of experimental models, the effects of feeding different types and amounts of beverages have been extensively investigated. Most have shown a positive effect of fluid ingestion on performance. Coyle et al., (1986) demons-trated that a fixed workload can be sustained 30% longer (from 3 to 4h) when carbo-hydrate solutions are given during exercise. Similarly, five other well controlled trials where continuous cycling exercise for 2h or more was performed, showed improvements in exercise with administration of carbohydrate-containing drinks (Bjrkman et al., 1984; Coggan & Coyle 1988; Coggan & Coyle, 1989; Coyle et al., 1983; Ivy et al., 1979). However, three studies showed no significant effects of administration of carbohydrate-containing drinks (Brookes, Davies & Green, 1975; Felig et al., 1982; Flynn et al., 1987) and in none of these studies was performance adversely affected. In all the studies where positive effects were seen, the total volume of fluid consumed varied from 0.4 to 2.7L and the carbohydrate intake ranged from 120 to 410g. In studies employing prolonged intermittent cycle exercise followed by brief high intensity sprint, performance of sprint lasting 12 to 14minwas enhanced if solutions containing 50, 60 or 75g/L of various sugars were given at a rate of 8.5 ml/kg/hr (Mitchell et al., 1988) or when a 6% sucrose solution (42g in 692ml) was given (Murray et al., 1989b). Similarly, 47g glucose-sucrose mixture with added electrolyte or 55g glucose polymer-fructose mixture with added

electrolytes also improved performances of a sprint ride lasting about 6 minutes (Murray, 1987). However, when low glucose polymer, fructose (76g in 1.27L) or when 810% carbohydrate polymer solutions were given, no improvements in performances were seen (Kingwell et al., 1989; Murray, 1987; Murray et al., 1989a). As with cycling, running performance too improved when carbohydrate-containing beverages were given (90g sucrose in 500ml (Sasaki et al., 1987); 7% carbohydrate- electrolyte solution (Macaraeg, 1983). In a study involving prolonged walking, walking time increased when 120g glucose polymer was given in 1.5L compared with a placebo (Ivy et al., 1983). In another experimental model where subjects are able to adjust the treadmill speed while running either to cover the maximum distance possible in a fixed time or complete a fixed distance in the fastest time possible, Williams and coworkers, showed an increased running speed in the closing stages of the trails when carbohydrate-containing drinks were given (Williams, 1989; Williams et al., 1990). In field studies, despite the many practical difficulties associated with the conduct of such studies and the lack of well-controlled studies, studies have shown the ergogenic effects of administration of fluids or glucose-electrolytes solutions. A 7mile (11.2km) walk/run course was completed when given glucose-electrolyte solution as compared to no ingestion of fluids (mean 7.5 km) (Cade et al., 1972). In a study where matched groups of marathon runners consumed 1.4L of either water or a glucose-electrolyte solution, 60% of the competitors were able to run faster than expected when drinking carbohydrate-electrolyte solution compared with 40% when water was given (Maughan & Whiting, 1985). Similarly, soccer players utilised 31% less glycogen when 7% glucose polymer was given during a practice match compared to the placebo group (Leatt, 1986). This reduced utilisation translates to beneficial effect in the latter stages of the game by players taking the glucose polymer, where it has been shown that soccer players beginning the game with low muscle glycogen concentration covered less distance

during the game, particularly during the second half, and spent more time walking and running at low speed (Saltin & Karlsson, 1977). Despite the clear evidence demons-trating the negative effects of dehydration on exercise performance and the positive effects of fluid ingestion (Herbert, 1983), most athletes still do not drink enough to match fluid losses during exercise (Nokes, 1992). This may in part be due to the relative insensitivity of the thirst mechanism in humans, and some degree of dehydration, as much as 2% of body weight is normally incurred before the stimulus to drink is initiated (Adolph, 1947). However, this level of dehydration is sufficient to impair exercise performance and thermoregulatory capacity. Drinking can bestimulated to some extent by improving beverage palatability (Hubbard, Szylk & Armstrong, 1990) or by educating the athletes through increasing the awareness of the need for a consciouseffort to increase fluid intake in situations where dehydration is likely to occur.