Journal Logo

BASIC SCIENCES: Original Investigations

Effect of intermittent high-intensity exercise on gastric emptying in man


Author Information
Medicine and Science in Sports and Exercise: August 2001 - Volume 33 - Issue 8 - p 1270-1278
  • Free


The ingestion of dilute carbohydrate-electrolyte drinks suitable during exercise has clearly been shown to delay fatigue and improve performance of prolonged strenuous exercise (for reviews see 15,19). The fatigue process is influenced by many factors, among which are depletion of the body’s carbohydrate stores and perturbations in thermoregulatory and cardiovascular control caused by the loss of body water incurred by sweating. Properly formulated solutions ingested during exercise can supply exogenous carbohydrate and water that can help offset the exercise-induced losses. The rates at which ingested drinks leave the stomach and are absorbed by the small intestine have been seen as two factors that can limit the availability of the constituents of the drinks. The rate of gastric emptying is considered to be an important factor limiting fluid replacement (19). Although the volume consumed and the formulation of the beverages are known to have a crucial influence on the rate of gastric emptying and intestinal absorption, it has been assumed that exercise intensity plays a relatively minor role (15,19,27). Although it is known that exercise at intensities greater than about 70% of an individual’s maximum oxygen uptake (V̇O2max) slows gastric emptying, it has been presumed that at high levels of physical activity the intensity and duration of the exercise would preclude drinking and that the exercise duration would be too short for any benefit to be derived from fluid ingested during the exercise (5). The majority of studies in which gastric emptying has been measured during exercise have therefore been carried out at intensities at or below 70% of the subject’s V̇O2max(15,19).

During many sports and other forms of physical activity, the period spent exercising is prolonged but the exercise is intermittent and the intensity is varied. Many of the factors that have been shown to induce fatigue during prolonged constant-intensity exercise have the same effect on intermittent exercise of sufficient duration. Ingestion of carbohydrate-electrolyte drinks is as effective in delaying fatigue during intermittent exercise lasting for at least 40 min as during constant-load exercise (27). It has been assumed that in most forms of intermittent exercise the time spent at relatively low levels of activity is sufficient to allow appropriate amounts of any ingested drinks to be emptied from the stomach and absorbed and that the time spent in sprinting will have little effect. For example, in Danish top-class soccer matches, the mean number of sprints per player was 19 and each sprint lasted approximately 2 s (2).

However, in many instances involving intermittent exercise, the overall intensity of the physical activity might be expected to delay gastric emptying (9,27). The average intensity of effort in top-level soccer has been reported to be above 70% of V̇O2max although the time spent in sprinting is relatively short (2). In spite of the popular appeal of sports with activity patterns that fluctuate between low and high intensity, there appear to be no studies in the literature that have examined the effect of intermittent sprint exercise on gastric emptying of ingested beverages.

The purpose of the present study was to investigate the effect of intermittent high-intensity cycle exercise on gastric emptying and to compare the response with that of exercise at a constant power output equivalent to the average power output of the intermittent exercise trial.


The double-sampling gastric aspiration technique of George (11), as modified by Beckers and colleagues (3), was used in this study. Eight healthy, physically active men with no history of gastrointestinal or metabolic disorders were recruited as subjects. Volunteers gave written informed consent before participation in this investigation, which had received prior approval from the Grampian Research Ethics Committee. The physical characteristics of the subjects were (mean ± SD): age, 25 ± 3 y, height 1.78 ± 0.041 m, body mass 74.4 ± 7.7 kg, and V̇O2max 48.9 ± 4.6 mL·kg−1·min−1. As a preliminary to the study, all potential subjects were screened to identify individuals who could be successfully intubated with the oro-gastric tube. During this process, the successful subjects underwent the major part of the gastric emptying protocol to establish the correct positioning of the tube in the stomach and to familiarize the subjects with this part of the study procedure. In addition, maximum oxygen uptake (V̇O2max) was determined using an incremental exercise test on an electrically braked cycle ergometer (Lode, Groningen, The Netherlands) within 2 wk of beginning the study. Oxygen consumption was measured with an on-line gas analysis system (MedGraphics CPX; Cardiokinetics, Manchester, UK), and heart rate was assessed using a telemetric heart rate monitor (Polar Favor, Bodycare Products, Warwickshire, UK). Attainment of V̇O2max was assumed to have been reached when there was little or no increase in V̇O2 produced by an increase in power output and the respiratory exchange rate was greater than 1.15. Verification of V̇O2max was carried out on a subsequent visit when, after a suitable warm-up exercise, subjects cycled for 3 min at power outputs 25 W below and above the highest power output completed on the first visit. From the results of the oxygen consumption test, power outputs corresponding to 60, 66, 70, and 100% of V̇O2max were calculated for each subject and used in the study trials. The maximum heart rate of subjects was taken as that measured over 60 s of exercise at the highest workload attempted during the V̇O2max test (median (range values) 191 (187–195) b·min−1). Before undertaking the main study trials, subjects participated in a practice trial where they were intubated and then cycled for 60 min after the protocol for the most intense exercise used in the study. Subjects drank 600 mL of the test drink then commenced exercising with the oro-gastric tube in place: at intervals throughout the exercise, the contents of subject’s stomach were aspirated and reinjected. This practice trial was carried out to familiarize the subject with the study procedures to alleviate any stress that unfamiliarity might induce, and to ascertain that the subject could complete the most intense exercise protocol.

Subjects undertook four experimental trials, separated by at least 3 d. The gastric emptying measurement period was 60 min on each trial. In the resting trial (R), subjects remained seated on a chair throughout the experimental period: in the continuous steady-state exercise trial subjects cycled at a fixed power output calculated to correspond to 66% of their V̇O2max (C66). During the intermittent high-intensity exercise trials, subjects cycled either at a power output calculated to be equivalent to 60% of their V̇O2max interspersed at fixed intervals with higher intensity sprints on the same cycle ergometer, or they cycled at a power output equivalent to 70% of their V̇O2max interspersed with the sprints. Each bout of high-intensity exercise consisted of three 30-s sprints at 100% of V̇O2max interspersed with 60 s of exercise at the lower intensity for that trial (Fig. 1). The first bout of sprinting commenced 4 min after the start of the exercise period: thereafter, the high-intensity bouts were 10 min apart, which meant that the gastric emptying measurements were always carried out while subjects were exercising at the lower intensity for that trial. On the trial where subjects exercised at 60% of V̇O2max and sprinted at 100% of V̇O2max the average power output was calculated to be equivalent to 66% of their V̇O2max (I66). On the trial where the exercise intensities were 70% of V̇O2max and 100% of V̇O2max, the average power output corresponded to 75% of their V̇O2max (I75). The treatment order was randomized according to a 4-way crossover design, and the same electrically braked cycle ergometer was used for all exercise trials. The ambient temperature of the laboratory was maintained between 18 and 23°C on all trials. The gastric tube remained in place throughout each of the four trials, and measurements of the volume of the gastric contents were carried out every 10 min as described by Beckers et al. (3).

Schematic diagram of the protocol used for the intermittent exercise trials. During the low-intensity exercise bouts, subjects cycled at a power output equivalent either to 60% of their V̇O2max (I66) or 70% of their V̇O2max (I75), and for the high-intensity sprints, they cycled at 100% of their V̇O2max in both trials.

Subjects were asked to refrain from strenuous exercise and alcohol and to eat the same diet for 24 h before each trial day. They reported to the laboratory in the morning and, except for drinking two mugs (∼ 500 mL) of water on waking, they were overnight fasted. After providing a urine sample, body mass was recorded with the subject wearing shorts. The subjects then dressed in shorts, T-shirt, and training shoes on all trials. Subjects then sat at rest and the gastric tube (French Levine, 14 gauge; Vygon Ltd., Ecouen, France) was passed orally and positioned in the stomach (12). The telemetric heart rate monitor was placed on the chest before the subjects sat either on a chair if they were on the resting trial or on the cycle ergometer if they were on an exercise trial.

Five minutes after the subjects had assumed this sitting posture, their resting heart rate was recorded and the residual fasting content of their stomach was aspirated. Subjects then ingested 600 mL of the test drink within 90 s; heart rate was recorded and, if required, subjects then immediately began to exercise. Timing of the trial was started and the initial mixing and sampling of the stomach contents was completed.

The test beverage was a commercially available carbohydrate-electrolyte beverage (Gatorade, Quaker Oats Ltd, Silea, Italy) supplied as a single batch in powder form; the powder was reconstituted in distilled water according to the manufacturer’s instructions. A volume of 600 mL of the test beverage was accurately measured into a glass vessel and mixed with 12 mg of phenol red dye. A 2.5-mL sample of the test solution was retained for analysis of its composition and phenol red concentration. The test drinks were all given at room temperature (18–20°C).

Mixing of the stomach contents was carried out by aspiration and re-injection of 30–50 mL of the liquid in the stomach using a 60 mL catheter tip syringe (Becton Dickinson, Drogheda, Ireland); the time taken to mix the stomach contents was always 1 min. A 2.5 mL sample of the mixed gastric contents was then collected: the difference in the phenol red concentration between the initial test beverage and the mixed contents of the stomach was used to calculate the residual volume of swallowed saliva and secretions from the esophagus and stomach. Samples of the stomach contents were collected at 10 min intervals for 60 min after ingestion of the test beverage. Nine min after ingestion, the gastric contents were mixed as before and a 2.5 mL sample collected. Ten min after ingestion, 5 mL of a known concentration of phenol red was added, and the gastric contents were again mixed before a second 2.5 mL sample was aspirated at 11 min after ingestion. The volumes calculated from these two samples are referred to as those of the 10 min sample point. As phenol red is poorly absorbed by the stomach (4), the differences in the concentration of the dye in gastric samples can be used to calculate the total volume in the stomach and the volume of test beverage remaining in the stomach at these time points (3). The difference between the total gastric volume and test drink volume is the volume of secretions and swallowed saliva in the stomach. Phenol red was measured spectrophotometrically at a wavelength of 560 nm after dilution (1:10) with NaOH-NaHCO3 buffer (250:500 mmol·L−1, pH 9.7). At the end of the 60 min of measurement of gastric emptying, the stomach was washed by instilling 100 mL of distilled water, mixed with the remaining gastric contents and a sample was taken; the total fluid volume of the stomach was then emptied as completely as possible by aspiration.

Comparisons were made of the values for total gastric fluid volume at the end of the experiment as calculated by three different methods: 1) the method of Beckers et al. (3), 2) as estimated from the dilution of the phenol red concentration of the stomach contents by the 100 mL of distilled water wash, and 3) as measured by the actual volume aspirated from the stomach minus the 100 mL of distilled water wash.

Heart rate was recorded at 5 min intervals timed from the end of ingestion of the test drink. In addition, on the intermittent high intensity trials, a heart rate recording was also made within the last 5 s of each of the sprints and the average of the readings for three consecutive sprints was taken as the value for that bout of intense exercise.

At the end of each trial, subjects removed the gastric tube, showered, urinated then dressed in dry shorts before having their body mass measured. The volume of urine passed was recorded and a sample was collected for osmolality and for analysis of sodium, potassium and chloride concentration. Sweat loss was calculated from the change in body mass corrected for fluid ingestion minus the volume urine excreted and the volume of fluid aspirated from the stomach at the end of the measuring period. It was assumed that each gram of the corrected body mass loss was equivalent to one mL of sweat loss.

Sodium and potassium concentration in the test drink, gastric aspirates and urine samples was measured using a flame photometer (Clinical Flame photometer 410C; Corning Ltd., Halstead, Essex, UK); chloride concentration was determined using a coulometric titrator (PCLM 3; Jenway, Dunmow, Essex, UK). Osmolality was measured by freezing point depression (Roebling Automatik Osmometer; Camlab, Cambridge, UK).

Statistical Analysis

Initially the distribution of the data was examined using a normal probability plot and the derived correlation coefficient. Where this analysis suggested that the data were normally distributed, statistically analysis was carried out using a one-way analysis of variance (ANOVA) with factors for subjects and treatment. This was followed by application of the Tukey multiple range test to assess any differences between treatments where appropriate. Data that were not normally distributed were analyzed using the Kruskal-Wallis nonparametric one-way analysis of variance with factors for subjects and treatment in the model. Pairwise differences were assessed using Wilcoxon’s matched-pairs signed ranks test where appropriate. The time to empty half the test meal volume (t1/2) on each trial was calculated by linear regression following logarithmic transformation of the volume of test meal remaining in the stomach at each time point. Parametric data are reported as mean ± SD values and nonparametric data are reported as median and range values in the text, tables and figures. Significance level was taken to be P < 0.05 in all cases.


There were no differences between trials in the mean ± SD concentration of sodium (19 ± 1 mmol·L−1), potassium (3.2 ± 0.1 mmol·L−1) or chloride (11 ± 1 mmol·L−1), or in the osmolality (295 ± 7 mosmol·kg−1) of the test beverage. According to the manufacturers, the composition of the test beverage was (mmol·L−1): carbohydrate, 333; sodium, 18; potassium, 3.0; chloride, 11; osmolality range 270–330 mosmol·kg−1, and the energy content was 1.06 MJ·L−1.

Exercise Intensity

One subject could complete only 50 min of exercise in trial I75, but this subject successfully finished the remaining trials; the other subjects completed all four trials. The power outputs used in the trials corresponded to those that elicited the relevant percentage of V̇O2max determined during that test. The average exercise intensity in each trial was calculated as the time spent at the various power outputs, and was expressed as a percentage of V̇O 2max. The median (range) values were 66.1 (65.2–67.3) % in trial C66, 65.9 (63.3–68.3) % in trial I66 and 74.7 (71.8–75.5) % in trial I75.

The heart rate response followed the expected pattern, with no difference between trials in the preingestion resting heart rate and higher intensity exercise eliciting a higher heart rate (Figure 2). Following ingestion of the test drink, but before starting exercise, heart rate increased compared to the resting value (P = 0.028) to a similar extent in all four trials (P = 0.15). Heart rate was higher during trial I75 (P = 0.001) than during I66 and heart rate was higher (P = 0.02) during I66 than during C66. After the initial 5 min period of exercise, heart rate remained relatively stable during the constant workload trial (P = 0.25) but increased during both sprint trials I66 and I75. Overall, subjects exercised at an average of 78 ± 8% of their maximum heart rate during trial C66, 81 ± 8% during trial I66 and 88 ± 6% during trial I75. The percentage of maximum heart rate was not different in trials C66 and I66 (P = 0.064) and was lower than that on trial I75 (P = 0.001).

Heart rate response (beats min<·;min>−1) for the four experimental trials. Data are reported as mean values. The pooled standard deviation for the four trials at each time point is shown at the top of the graph. A broken line on the graph indicates the median measured maximum heart rate value for the group of subjects.

Gastric Emptying

Total volume in the stomach.

The total volume in the stomach (Table 1) immediately following ingestion of the test drinks was similar in all four trials (P = 0.67). During the remainder of the experimental period, the total volume remaining in the stomach was similar in trials R and C66 (P = 0.87), but the volume was less in these two trials than in trials I66 (P < 0.05) and I75 (P = 0.001). Total gastric volume in trial I66 was less than that in trial I75 (P = 0.001). While differences in the total gastric volume between trial I66 and both trials R and C66 were evident only during the initial 30 min of measurement, the volume in the stomach in trial I75 was greater than in trials R and C66 throughout the 60 min period of measurement. At 60 min, the measurements of the total volume remaining in the stomach were similar whether obtained by aspiration, or by dye dilution using the formula of Beckers and coworkers (3) or from dilution of the dye by the 100 mL of distilled water wash (P = 0.91;Table 2).

Table 1
Table 1:
The median (range) total volume in the stomach at each 10-min sampling point in the rest trial (R), the continuous steady-state moderate exercise trial (C66), the intermittent exercise trial at 60% of V̇O2max with sprints at 100% of V̇O2max (I66) and the intermittent exercise trial at 70% of V̇O2max with sprints at 100% of V̇O2max (I75).
Table 2
Table 2:
Comparison of median (range) volume in the stomach at the end of each study period measured by direct aspiration, and estimated by the dye dilution method of Beckers et al(3) and by the dilution of the dye by the wash volume of distilled water; the aspiration volume is the amount of fluid removed from the stomach minus the 100-mL wash-out volume.

Test beverage volume remaining in the stomach.

The pattern of emptying of the test beverage volume (Figure 3) differed slightly from that of the total gastric volume, indicating differences between trials in the volume of secretions and saliva present. Overall, the speed of gastric emptying of the test solution was not different in trials R and C66 (P = 0.96) which was faster than that during the intermittent sprint trials I66 (P < 0.05) and I75 (P < 0.001). However, at the 10 min sampling point no difference could be detected (P = 0.11) in test drink volumes remaining in the stomach between trials R (457 (196–568) mL), C66 (474 (138–597) mL), I66 (556 (366–598) mL) and I75 (538 (326–577) mL). Over the remaining 50 min of the measurement period, gastric emptying in R and C66 was faster than on I66 (P < 0.040) which was faster than I75 (P = 0.006).

The median volume (mL) of test solution remaining in the stomach at each time point for the four trials. For clarity, the range values are not shown.

The calculated t1/2 for the test beverage was not different (P = 0.71) in trials R (20 (7–30) min) and C66 (21 (7–49) min), and in both of these trials was less (P < 0.005) than in trial I75 (62 (27–100) min). Although t1/2 was less in trial R than in trial I66 (30 (15–74) min;P = 0.002), no differences could be detected in t1/2 values between trials C66 and I66 (P = 0.11) or between trials I66 and I75 (P = 0.12).

In the intermittent exercise trials some of the stomach aspirates, which are usually dyed red with phenol red, were strongly colored yellowish-green indicating the presence of bile. This suggests that some reflux of intestinal contents occurred during these trials. In trial I66 bile staining was present in aspirates collected from two subjects between the 40 and 60 min measurement points: in trial I75 the same two subjects and one other also produced bile stained aspirates. No bile stained aspirates was collected in any of the resting or continuous exercise trials.

Spectrophotometric scanning of the absorption spectra demonstrated that the presence of bile, at the concentrations present in the stomach aspirates, did not interfere with the measurement of the phenol red content of the aspirates read at a wavelength of 560 nm. When known amounts of phenol red, covering a concentration range of 20–100 mg·L−1, were added to the bile stained aspirates, the measured concentrations were not different (P = 0.73) from the values calculated from the amount of dye added plus the dye concentration measured in the original aspirate samples. Therefore the presence of bile in the aspirates did not appear to affect the measurement of the dye concentrations and hence the estimation of gastric emptying. Because of the unknown amount of liquid refluxing into the stomach during the intermittent sprint trials the volume of saliva and gastric secretions present in the stomach could not be estimated.

Energy Delivery to the Duodenum

Energy delivery to the duodenum was similar in trials R and C66 (P = 0.93) which was greater than that during I66 (P = 0.02) and I75 (P = 0.001)(Table 3). In line with the change in test drink volume emptied from the stomach, the amount of energy delivered over each 10 min period following ingestion varied in trials R (P = 0.013), C66 (P = 0.01) and I66 (P = 0.001) but was similar in trial I75 (P = 0.86).

Table 3
Table 3:
The median (range) cumulative amount of energy delivered to the duodenum during the rest trial (R), the continuous steady-state moderate exercise trial (C66), the intermittent exercise trial at 60% of V̇O2max with sprints at 100% of V̇O2max (I66), and the intermittent exercise trial at 70% of V̇O2max with sprints at 100% of V̇O2max (I75).

The initial mixing of the ingested drink with the small volume of residual gastric contents remaining after placement of the oro-gastric tube produced a slight change in the electrolyte concentration and osmolality of the test drink (Table 4). Thereafter, the sodium, potassium (not shown) and chloride concentration of the gastric contents increased and the osmolality decreased over the 60 min period of gastric measurement in all trials (Table 4). The changes in the composition of the gastric contents over time in all trials were consistent with the suggestion of a relatively greater proportion of gastrointestinal secretions being incorporated into a decreasing volume of the test drink present in the stomach. There was no difference found between trials in the concentration of the measured electrolytes or the osmolality of the gastric contents at the end of the gastric measurement period.

Table 4
Table 4:
The mean (SD) concentration of electrolytes present and the osmolality of the gastric contents during the rest trial (R), the continuous steady-state moderate exercise trial (C66), the intermittent exercise trial at 60% of V̇O2max with sprints at 100% of V̇O2max (I66) and the intermittent exercise trial at 70% of V̇O2max with sprints at 100% of V̇O2max (I75).

Fluid Balance

The mean body mass of the subjects was similar (P = 1.00) at the start of each of the trials (74.4 ± 8.6 kg). Median (range) sweat loss, calculated from the difference in body mass at the beginning and end of each trial and adjusted for fluid intake, urine excreted and the volume aspirated from the stomach, was less in the rest trial (0.06 (−0.03 to 0.48) kg than in the exercise trials (P < 0.002). Median water losses were similar (P = 0.76) in trials C66 (0.63 (0.20 to 1.02) kg) and I66 (0.69 (0.52 to 1.10) kg), and were less (P < 0.05) than in trial I75 (1.15 (0.60 to 1.70) kg).

Although there was considerable variability between subjects in the volume of urine produced before each trial, median (range) urine volumes were similar (P = 0.66) before the start of each trial (Table 5). Urine osmolality was similar at the start of the trials (P = 0.70); the median values suggested that the subjects were adequately hydrated at the beginning of all trials, although one individual consistently produced urine with a high osmolality (Table 5). No differences in electrolyte concentration or content of the urine collected before each trial were detected. Median urine volumes excreted at the end of trials R, C66 and I66 were similar (P = 0.46), but the volume produced at the end of trial I75 was less than that produced following trial R (P = 0.013;Table 5). The osmolality of the urine collected after each trial was similar (P = 0.24). The sodium content of the urine collected at the end of trials R, C66 and I66 were similar (P = 0.53) and greater than that of the urine collected after trial I75 (P = 0.015). There were no other differences in the electrolyte content of urine collected at the end of the trials (Table 5).

Table 5
Table 5:
The median (range) volume, electrolyte content, and osmolality of the urine samples collected before and after the rest trial (R), the continuous steady-state moderate exercise trial (C66), the intermittent exercise trial at 60% of V̇O2max with sprints at 100% of V̇O2max (I66), and the intermittent exercise trial at 70% of V̇O2max with sprints at 100% of V̇O2max (I75).


The present study demonstrates that intermittent sprint exercise can cause a slowing of gastric emptying of an ingested beverage relative to that measured at rest or during continuous steady-state exercise. While cycle exercise above 70% of an individual’s V̇O2max has been shown to slow gastric emptying (7), physical activity at lower intensities has been considered to have little effect on the rate of emptying from the stomach (10,19,24,25). Using an isotopic water tracer method, we have previously shown that cycle exercise at 40% of V̇O2max and above can reduce the availability of ingested fluids, and that the effect is proportional to the exercise intensity (16). In that study, we could not distinguish between effects on gastric emptying or on intestinal fluid absorption. However, the majority of studies investigating the effect of cycle exercise on gastric emptying have found no significant slowing in the rate at exercise intensities at or below 70% of V̇O2max(15,19).

In most competitive sports there is considerable variation in exercise intensity, and in situations where the activity is prolonged, participants are advised to ingest fluid and nutrients in order to delay the fatigue process (27). It has been presumed that if the average exercise intensity is less than 70% of an individual’s V̇O2max gastric emptying of ingested fluids would be essentially the same as that during rest (6,27). The present study has shown that while exercise at a constant work load equivalent to 66% of V̇O2max did not affect gastric emptying, cycling exercise incorporating short high intensity intermittent sprints giving an overall intensity of 66% V̇O2max slowed gastric emptying relative to rest and to the continuous exercise trial. There was no difference in the overall heart rate during trials C66 and I66 suggesting that generally exercise intensity was similar, however, heart rate was elevated during the sprints and the physiological stress may have been greater during trial I66. Increasing the intensity of the moderate level of exercise while keeping the same intensity, duration and number of the sprints caused a further slowing of gastric emptying.

In the present study, subjects had to exercise at the intensity and duration that was imposed by the study protocol. Individuals competing in team sports have more control of the levels of activity at which they exercise and therefore may be expected to regulate the intensity to conserve energy allowing them to participate throughout the duration of play. Longer periods of low-intensity exercise could therefore be predicted to occur in team sports than in the present study. However, in a recent study, we have shown that during 30 min of a competitive soccer match the median (range) volume of a sports drink emptied from the stomach (206 (34–523) mL) was reduced compared with the volume that was emptied over the same time period when the subjects undertook low-intensity walking exercise (489 (235–734) mL) (14). The overall intensity of the exercise was calculated to be 59 (22) % V̇O2max during the soccer match, but the pattern of activity was intermittent with periods of low-level activity punctuated by short bursts of high-intensity sprinting. Others have suggested that the mean exercise intensity is about 75% of V̇O2max during a professional soccer match (for reviews see 2,9). Intermittent high-intensity cycling or running exercise, at an apparently overall moderate activity level, can inhibit the gastric emptying of carbohydrate-electrolyte drinks. However, there is great variation in the levels of activity and the length of rest periods both between different sports and between individuals participating in the sport.

Several mechanisms have been proposed whereby exercise intensity may affect the rate of gastric emptying, but there is little evidence to suggest which of these factors plays the major role in influencing emptying from the stomach (19). Normally, the rate of gastric emptying of ingested liquids is regulated to limit the rate at which the constituents enter the small intestine (18,19,25). Although increasing the volume of the stomach contents increases the rate of emptying, increasing the nutrient content, osmolality, or acidity of liquids in the stomach slows gastric emptying (17,19). Gastric emptying of liquids into the small intestine is brought about by a relatively higher intragastric than duodenal pressure (18). Contractile activity in the stomach and gastroduodenal junction is controlled by an integrated system of neural, mainly of vagal origin, and hormonal factors that modulate the intrinsic muscular tone of this region of gut (17).

Any exercise-induced release of vasoactive hormones, many of which inhibit gastric contractility, could slow gastric emptying (19). A greater reduction in splanchnic blood flow produced by the intermittent high-intensity exercise may also play a role in slowing the rate of emptying (26). This may cause a reduction in the rate of intestinal absorption and therefore a build up of nutrients in the intestinal lumen, or an increase in the concentration of absorbed nutrients in the capillary bed of the villi if intestinal blood flow was relatively stagnant. In both situations, inhibitory feedback mechanisms would be triggered and gastric emptying would be slowed (18,25). The exercise-induced reduction in splanchnic blood flow may be sufficient to cause a level of hypoxia that perturbs the normal function of the gastroduodenal junction (22). Others have suggested that there may be a mechanical effect caused by upper body movements that influence gastric motility (20). Although upper body movement is relatively slight during cycling exercise (20,25), intra-abdominal pressure may be increased due to the greater recruitment of abdominal muscles to aid respiration during bouts of intense exercise (1). In the present study, the observation that, at least in some individuals, there was reflux of intestinal contents during the later stages of the high-intensity exercise may suggest that this type of exercise can cause greater pressure in the duodenum than in the stomach.

Both hyperthermia and dehydration have been shown to slow gastric emptying by mechanisms that are not clearly understood (21,25). In the present study, subjects on the higher-intensity intermittent exercise trial (I75) were hypohydrated to a greater extent because of the higher sweat loss and demonstrated slower gastric emptying than on either of the other two exercise trials or the trial at rest. However, although gastric emptying was slower in the intermittent sprint trial (I66) than in the steady-state exercise trial with the same average exercise intensity (C66), there was no difference in the level of hydration exhibited by subjects in these two trials. It is possible that the differences in gastric emptying between the exercise trials could have been due to differences in body temperature, which was not measured in this study. Nevertheless, as the main effect on gastric emptying occurred over the 10-min period immediately after ingestion, and there was no difference in emptying between the rest trial and steady-state exercise trial, it is unlikely that hyperthemia or hydration status were factors in the observed results.

Ingestion of fluids during intense intermittent exercise has been shown to delay fatigue and to improve exercise performance and endurance capacity (27). Therefore, it must be assumed that in many situations the gastric emptying rate of ingested solutions is sufficient to give some degree of benefit to an individual undertaking variable-intensity exercise that include high-intensity sprints. In the present study, a greater volume of fluid and greater amount of energy were delivered to the small intestine during the constant load trial than during either of the intermittent trials. However, expressed as a percentage of initial body mass, the levels of hypohydration at the end of trials C66 (range 0.3–1.1%) and I66 (range 0.1–1.2%) were not different (P = 0.25) and less than the 2% value normally quoted as the minimal level at which impairment of work capacity and heat tolerance can be detected (15). Levels of hypohydration at the end of trial I75 (range 0.9–2.2%) were probably sufficient to reduce physical performance in at least some of the subjects.

The amount of carbohydrate delivered by the test solution to the small intestine was greater during trial C66 than I66 or I75. During the first 30 min of exercise, the calculated median (range) amount of carbohydrate delivered to the duodenum was 23.3 (9.9–35.0) g during trial C66, 19.9 (11.5–29.6) g during trial I66, and 14.9 (4.7–20.1) g during trial I75. Coyle and Montain (8) have proposed that a carbohydrate intake of between 0.5–1.0 g·min−1, preferably absorbed during the early part of exercise, is sufficient to delay fatigue during prolonged exercise by maintaining blood glucose levels and carbohydrate oxidation. This would suggest that for most of the subjects in trials C66 and I66 and some of the subjects in trial I75, the rates of gastric emptying supplied adequate amounts of the ingested test solution to have a significant effect on maintaining blood glucose levels (8) and glycogen sparing (28), assuming that intestinal absorption rates were not limiting (8).

The present study has shown that intermittent high-intensity exercise slows gastric emptying, however, sufficient water and carbohydrate can be delivered to the small intestine to benefit exercise capacity in sports with repetitive, high-intensity, sprint activities. The differences in the volumes emptied between the intermittent intensity exercise trials and the rest trial may appear small but, in the early period following ingestion, they represent a significant proportion of the total volume ingested. Ingestion of dilute carbohydrate-electrolyte drinks is helpful to individuals undertaking intermittent high-intensity exercise, but the slowing of gastric emptying caused by the exercise may reduce the benefit especially if larger volumes are ingested over a longer duration than were used in the present study. The rate of gastric emptying of liquids is greatly affected by the volume in the stomach (23). As the gastric volume decreases, there is less distention in the distal region of the stomach, which reduces the activity of the stretch receptors in the gastric mucosa, resulting in a reduction in gastric motility that slows emptying (13,18). Repeated drinking to maintain a large fluid volume in the stomach can sustain high rates of gastric emptying of carbohydrate solutions (5,15). Whether maintaining a relatively large volume of test beverage in the stomach is sufficient stimulus to override the inhibition caused by bouts of sprinting or whether the ingested fluid would accumulate in the stomach and cause gastrointestinal distress remains to be investigated.

N.B. was supported by a student grant from the Gatorade Sports Science Institute. This study was funded by Aberdeen University, UK. The sports drink was kindly donated by Gatorade, Quaker Oats, Silea, Italy.

The results of the present study do not constitute endorsement of the product by the authors or ACSM.

The authors thank the reviewers for their time and effort in contributing to the manuscript.

Address for correspondence: John B. Leiper, Ph.D., Department of Biomedical Sciences, Human Physiology Building, University Medical School, Foresterhill, Aberdeen, AB25 2ZD, United Kingdom. E-mail:


1. Aliverti, A., S. J. Cala, R. Duranti, et al Human respiratory muscle actions and control during exercise. J. Appl. Physiol. 83: 1256–1269, 1997.
2. Bangsbo, J. Fitness Training in Football: A Scientific Approach. Bagsvaerd, Denmark: HO+Storm, 1994,pp. 57–77.
3. Beckers, E. J., N. J. Rehrer, F. Brouns, F. ten Hoor, and W. H. W. Saris. Determination of total gastric volume, gastric secretion and residual meal using the doubling sampling technique of George. Gut 29: 1725–1729, 1988.
4. Bloom, D. S., E. D. Jacobson, and M. I. Grossman. Validation of dilution indicators in the stomach. Gastroenterology 52: 205–211, 1967.
5. Brouns, F. Gastric emptying as a regulatory factor in fluid uptake. Int. J. Sports Med. 19: S125–S127, 1998.
6. Brouns, F., N. J. Rehrer, E. J. Beckers, and W. H. M. Saris. Effects of exercise on gastrointestinal function: implications for water and substrate provision during exercise.In:Intermittent High Intensity Exercise, D. A. D. Macleod, R. J. Maughan, C. Williams, C. R. Madeley, J. C. M. Sharp, and R. W. Nutton (Eds.). London: Spon, 1993,pp. 117–137
7. Costill, D. L., and B. Saltin. Factors limiting gastric emptying during rest and exercise. J. Appl. Physiol. 37: 679–683, 1974.
8. Coyle, E. F., and S. J. Montain. Carbohydrate and fluid ingestion during exercise: are there trade-offs? Med. Sci. Sports Exerc. 24: 671–678, 1992.
9. Ekblom, B. Applied physiology of soccer. Sports Med. 3: 50–60, 1986.
10. Fordtran, J. S., and B. Saltin. Gastric emptying and intestinal absorption during prolonged severe exercise. J. Appl. Physiol. 23: 331–335, 1967.
11. George, J. D. New clinical method for measuring the rate of gastric emptying: the double sampling test meal. Gut 9: 237–242, 1968.
12. Hassan, M. A., and M. Hobsley. Positioning of subject and of nasogastric tube during a gastric secretion study. Br. Med. J. I: 458–460, 1970.
13. Hunt, J. N., and M. T. Knox. Regulation of gastric emptying.In:Handbook of Physiology. Section 6: Alimentary Canal, Vol. IV, Motility. Washington, DC: American Physiological Society, 1968,pp. 1917–1935.
14. Leiper, J. B., A. S. Prentice, C. Wrightson, and R. J. Maughan. Gastric emptying of a carbohydrate-electrolyte drink during a soccer match (Abstract). In:Proceedings of the 18th IBC Conference; January 27–28, 2000, Padua, Italy, Berichte der Osterreichische Gesellschaft für Klinische Chemie (Suppl), 14, 2000.
15. Maughan, R. J. Carbohydrate-electrolyte solutions during prolonged exercise. In:Perspectives in Exercise Sciences and Sports Medicine, Vol. 4, Ergogenics-Enhancement of Performance in Exercise and Sport, D. R. Lamb and M. H. Williams (Eds.). Carmel, IN: Benchmark Press, 1991, pp. 35–85.
16. Maughan, R. J., J. B. Leiper, and Mcgraw, B. A. Effects of exercise intensity on absorption of ingested fluids in man. Exp. Physiol. 75: 419–421, 1990.
17. Meyer, J. H. Motility of the stomach and gastroduodenal junction.In:Physiology of the Gastrointestinal Tract, 2nd Ed., Vol 1. L. R. Johnson, J. Christensen, M. J. Jackson, E. D. Jacobson, and J. H. Walsh (Eds.). New York: Raven Press, 1987,pp. 613–629.
18. Minami, H., and R. W. Mccallum. The physiology and pathophysiology of gastric emptying in humans. Gastroenterology 86: 1592–1610, 1984.
19. Murray, R. The effects of consuming carbohydrate-electrolyte beverages on gastric emptying and fluid absorption during and following exercise. Sports Med. 4: 322–351, 1987.
20. Neufer, P. D., A. J. Young, and M. N. Sawka. Gastric emptying during walking and running. Eur. J. Appl. Physiol. 58: 400–445, 1989.
21. Neufer, P. D., A. J. Young, and M. N. Sawka. Gastric emptying during exercise: effects of heat stress and hypohydration. Eur. J. Appl. Physiol. 58: 433–439, 1989.
22. Nielsen, H. B., L. B. Svendsen, T. H. Jensen, and N. H. Secher. Exercise-induced gastric mucosal acidosis. Med. Sci. Sports Exerc. 27: 1003–1006, 1995.
23. Noakes, T. D., N. J. Rehrer, and R. J. Maughan. The importance of volume in regulating gastric emptying. Med. Sci. Sports Exerc. 23: 307–313, 1991.
24. Rehrer, N. J., E. Beckers, F. Brouns, F. ten Hoor, and W. H. M. Saris. Exercise and training effects on gastric emptying of carbohydrate beverages. Med. Sci. Sports Exerc. 21: 540–549, 1989.
25. Rehrer, N. J., F. Brouns, E. J. Beckers, and W. H. M. Saris. The influence of beverage composition and gastrointestinal function on fluid and nutrient availability during exercise. Scand. J. Med. Sci. Sports 4: 159–172, 1994.
26. Rowell, L. B. Human cardiovascular adjustments to exercise and thermal stress. Physiol. Rev. 54: 75–159, 1974.
27. Shi, X., and C. V. Gisolfi. Fluid and carbohydrate replacement during intermittent exercise. Sports Med. 3: 157–172, 1998.
28. Yaspelkis, B. B., J. G. Patterson, P. A. Anderla, Z. Ding, and J. L. Ivy. Carbohydrate supplementation spares muscle glycogen during variable-intensity exercise. J. Appl. Physiol. 75: 1477–1485, 1993.


© 2001 Lippincott Williams & Wilkins, Inc.