J Appl Physiol 108: 245–250, 2010.
First published November 25, 2009; doi:10.1152/japplphysiol.91275.2008.
Carbohydrate exerts a mild influence on fluid retention following
exercise-induced dehydration
Kristin L. Osterberg, Shannon E. Pallardy, Richard J. Johnson, and Craig A. Horswill
Virginia Polytechnic and State University, Blacksburg, Virginia; Gatorade Sports Science Institute, Barrington, Illinois;
and University of Colorado, Denver, Colorado
Submitted 23 September 2008; accepted in final form 19 November 2009
carbohydrate; fluid retention; rehydration; renin; osmolality
rehydration, or restoration of fluid spaces, is
important when acute illness or excessive sweating has compromised hydration status. Most rehydration research to date has
focused on the effects of electrolytes on fluid retention. There is
little argument that sodium, particularly sodium chloride, has the
greatest impact on fluid retention (16, 21, 25, 35, 38). Sodium is
the primary cation in the extracellular fluid and as such serves to
regulate plasma osmolality and helps maintain plasma volume.
Maughan et al. (22) found that potassium chloride was as effective
as sodium chloride in restoring fluid balance; however, this effect
RAPID AND COMPLETE
Address for reprint requests: C. Horswill, Gatorade Sports Science Institute,
Barrington, IL 60010 (e-mail: [email protected]). Address for
other correspondence: K. Osterberg, Virginia Polytechnic and State Univ., 212
War Memorial Hall, Blacksburg, VA 24061 (e-mail: [email protected]).
http://www.jap.org
was likely due to the chloride contribution and its impact on the
extracellular fluid compartment.
Carbohydrate is often included in rehydration solutions to aid in
fluid absorption and add taste to increase sensory acceptance. For
the treatment of dehydration induced by severe diarrhea, the
World Health Organization recommends a standard solution containing 90 mmol/l sodium and 111 mmol/l glucose in oral rehydration solutions (ORS) to assist fluid absorption through the
intestinal sodium-glucose cotransporter (7). For exercise and
sport, rehydration solutions commonly contain a mixture of glucose and fructose in precise ratios based on absorption (8, 12, 32)
and oxidation (1, 19, 20) characteristics. Although multiple carbohydrate substrates facilitate fluid uptake, their role in retention
of fluid once inside the body remains unclear. Maughan et al. (22)
found that a beverage with glucose alone produced significantly
greater urine loss than any of three other beverages that contained
electrolytes. However, no study to date has attempted to determine
whether there is a dose effect of carbohydrate concentration on
fluid retention when electrolyte concentrations are standardized.
Carbohydrate could enhance fluid retention based on several
plausible mechanisms. First, progressively increasing the carbohydrate concentration, energy density, and osmolality has been
shown to decrease the rate of gastric emptying (5, 24, 37) and
absorption (13, 32) of fluid. Slower emptying and absorption may
cause a slower movement of fluid into the bloodstream and sustain
a higher plasma osmolality that would attenuate urine production
by the kidneys. Second, postexercise ingestion of beverages with
higher carbohydrate content could stimulate greater intracellular
fluid retention related to glycogen storage. Glucose transport into
the cell draws water along with the substrate, and glycogen
storage has been shown to increase total body water (27, 36).
Finally, in some people, the insulin response invoked by hyperglycemia may lead to increased renal tubular resorption of sodium
and fluid, which would promote fluid retention (33). Indeed,
physiological hyperinsulinemia causes sodium retention in healthy
individuals (29).
The purpose of the present study was to determine whether
there is a dose relationship between carbohydrate concentration
of an oral rehydration solution and the percentage of ingested
fluid that is retained to promote rehydration following exerciseinduced dehydration. We compared the effects of three concentrations of carbohydrate: 3, 6, and 12 g/100 ml each
containing standardized electrolyte concentrations (18 mmol/l
Na⫹, 11 mmol/l Cl⫺, 3 mmol/l K⫹). For the control comparison, we used a carbohydrate-free solution with standardized
electrolyte concentrations and a carbohydrate-free, electrolytefree solution (water). We hypothesized that compared with
water, fluid retention, defined as the percentage of the volume
retained relative to volume ingested to restore body mass loss,
would be greater for fluids containing electrolytes and similar
8750-7587/10 $8.00 Copyright © 2010 the American Physiological Society
245
Downloaded from http://jap.physiology.org/ by 10.220.33.4 on June 14, 2017
Osterberg KL, Pallardy SE, Johnson RJ, Horswill CA. Carbohydrate
exerts a mild influence on fluid retention following exercise-induced dehydration. J Appl Physiol 108: 245–250, 2010. First published November 25,
2009; doi:10.1152/japplphysiol.91275.2008.—Rapid and complete rehydration, or restoration of fluid spaces, is important when acute illness
or excessive sweating has compromised hydration status. Many studies have investigated the effects of graded concentrations of sodium
and other electrolytes in rehydration solutions; however, no study to
date has determined the effect of carbohydrate on fluid retention when
electrolyte concentrations are held constant. The purpose of this study
was to determine the effect of graded levels of carbohydrate on fluid
retention following exercise-induced dehydration. Fifteen heat-acclimatized men exercised in the heat for 90 min with no fluid to induce
2–3% dehydration. After a 30-min equilibration period, they received,
over the course of 60 min, one of five test beverages equal to 100%
of the acute change in body mass. The experimental beverages
consisted of a flavored placebo with no electrolytes (P), placebo with
electrolytes (P ⫹ E), 3%, 6%, and 12% carbohydrate solutions with
electrolytes. All beverages contained the same type and concentration
of electrolytes (18 meq/l Na⫹, 3 meq/l K⫹, 11 meq/l Cl⫺). Subjects
voided their bladders at 60, 90, 120, 180, and 240 min, and urine
specific gravity and urine volume were measured. Blood samples were
taken before exercise and 30, 90, 180, and 240 min following exercise
and were analyzed for glucose, sodium, hemoglobin, hematocrit, renin,
aldosterone, and osmolality. Body mass was measured before and after
exercise and a final body mass was taken at 240 min. There were no
differences in percent dehydration, sweat loss, or fluid intake between
trials. Fluid retention was significantly greater for all carbohydrate beverages compared with P (66.3 ⫾ 14.4%). P ⫹ E (71.8 ⫾ 9.9%) was not
different from water, 3% (75.4 ⫾ 7.8%) or 6% (75.4 ⫾ 16.4%) but was
significantly less than 12% (82.4 ⫾ 9.2%) retention of the ingested fluid.
No difference was found between the carbohydrate beverages. Carbohydrate at the levels measured exerts a mild influence on fluid retention in
postexercise recovery.
246
INFLUENCE OF CARBOHYDRATE ON FLUID RETENTION
for beverages containing carbohydrate concentrations of 6% or
less. Further, we hypothesized that a high-carbohydrate (12%)
beverage with electrolytes would result in greater fluid retention compared with beverages with less or no carbohydrate.
METHODS
Table 1. Composition of beverages
Carbohydrate type
Sodium, mmol/l
Potassium, mmol/l
Osmolality, mOsm
Water Placebo*
Placebo* ⫹ Electrolytes
3% ⫹ Electrolytes
6% ⫹ Electrolytes
12% ⫹ Electrolytes
15
18
3
49
Sucrose HFCS†
18
3
187
Sucrose HFCS†
18
3
338
Sucrose HFCS†
18
3
691
*Both placebo beverages contained flavoring (lemon lime) and artificial sweeteners (aspartame). †High fructose corn syrup (HFCS)-42 (58% glucose, 42%
fructose). With complete hydrolysis, the ratio of glucose to fructose for all treatments with carbohydrate is 3.25 to 2.75.
J Appl Physiol • VOL
108 • FEBRUARY 2010 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.4 on June 14, 2017
Participants. Fifteen trained men volunteered to participate in this
study. Participants ranged in age from 19 to 49 (34.4 ⫾ 10.0) yr with an
average body mass of 77.8 ⫾ 6.7 kg and an estimated maximal oxygen
consumption of 56.2 ⫾ 6.3 ml·kg⫺1·min⫺1. The estimate was based in
ECG graded exercise stress test that subjects complete before participation in the study (10). All subjects were recreationally competitive
runners or triathletes who trained year round and maintained their normal
training during the study. The sample size of 15 was chosen to provide
statistical power of 80% based on an expected difference for fluid
retention as reported for rehydration beverages administered in a repeatedmeasures design and with similar sodium and carbohydrate types used
in the current study (15, 34). Data for each subject were collected over
the course of 3 wk, and all subjects were heat acclimated before
beginning the trials (11, 18). The experimental protocol was approved
by a Human Subjects Review Committee, and all athletes gave written
informed consent before participation.
Study design. The study was double-blind, and treatment order was
randomized so that each participant completed all treatments serving
as his own control. Heat-acclimated participants reported to the
facility for each trial at 1445 having abstained from food and fluid for
3 h. Each subject provided a urine sample to assess preexercise
hydration status via urine specific gravity and was weighed without
clothing. A flexible catheter was inserted into the antecubital vein, and
participants sat for 30 min while wearing shorts and t-shirt in a warm
environment (30°C) while allowing for equilibration of fluid spaces
(24). A baseline blood sample was collected at 1530 and subjects
commenced exercise. Each trial consisted of 90 min of moderateintensity exercise at 70 –75% maximal heart rate based on the ECG
maximal exercise stress test, with no fluid consumed to induce 2–3%
loss of body mass. The 90 min exercise session consisted of 30 min
each of stationary cycling, running, and elliptical exercise in the same
order for each subject for each of the trials. The environmental
conditions were 30°C and 50% relative humidity (RH). Following
exercise, participants were weighed without clothing to determine the
amount of sweat loss incurred and then emptied their bladders. After
changing into dry clothing, they were escorted into a thermoneutral
environment (23°C, 30% RH) for the 4-h recovery period. After being
seated for 30 min postexercise, a second blood sample was collected
prior to ingesting any fluids. Beverages were stored at a constant
temperature of 2.8°C. Subjects were then given a volume of beverage
equivalent to 25% of their total change in body mass, which was
determined by subtracting postexercise body mass from preexercise
body mass. Another aliquot (25% of losses) was given at 40 min into
recovery. At 50, 60, 70, and 80 min, participants consumed aliquots of
beverage equivalent to 12.5% of body mass reduction for a total of
100% fluid replacement. At 90 min, a third blood sample was
collected. Participants were asked to void and collect all urine at 60,
90, 120, 180, and 240 min into recovery. Blood sampling was
repeated at 180 and 240 min into recovery before urine collection and
with subjects in a seated position. This rehydration protocol is similar
to that used by Maughan and Leiper (21). Surveys for rating of
gastrointestinal (GI) comfort were also completed at the times of
blood sampling. After the final urine and blood sample, a final nude
body mass was obtained.
Heat acclimation and experimental control. To ensure that subjects
were heat acclimated before the experimental trials, each participant
reported to the laboratory for 10 days of exercise in the heat over the
course of 14 –18 days (11). To do so, they exercised at a moderate
intensity for 90 min in 30°C and were permitted to run, cycle, exercise on
the elliptical machine, or any combination of the three. They were given
ad libitum access to fluids during exercise as well as encouraged to drink
fluids during the day to ensure expansion of plasma volume. Experimental trials started within 3 days of finishing the heat acclimation phase.
Diet and exercise controls. For the 24-h period before each experimental treatment, participants were given a control diet containing
3,417 kcal and 2,995 mg of sodium. Subjects were not permitted to eat
food other than what was provided. If they chose to not eat all of the
food that was provided, whatever was eaten before the first trial was
repeated for each subsequent trial, i.e., the 24-h diet was kept constant
across remaining trials for each subject. Because subjects were not
permitted to exercise the day before or the day of the trial, it is
reasonable to assume that subjects would not need more than 3,417
kcal to remain in energy balance. Further, diet carbohydrate content
was not standardized relative to body mass but the absolute amount
was consistent during the preexperiment 24-h period for each subject.
Sodium, caffeine, and alcohol intake was restricted due to the potential retention or diuretic effect.
Participants were given 1 liter of water and were told to drink as
much additional water as they desired but to stop drinking and eating
3 h before the start of the trial. This was done to ensure that they
reported for each trial in a consistent state of euhydration based on
urine specific gravity. Participants who arrived at the laboratory with
a urine specific gravity ⬎1.020 were rescheduled for another day.
Beverages. Subjects replaced 100% of the acute change in body
mass with one of five beverages that included placebo (noncaloric and
flavored water without electrolytes) or four beverages with the same
concentrations of electrolytes but that varied by carbohydrate concentration: 0, 3, 6, and 12% carbohydrate for which proportions of
sucrose, glucose, and fructose were kept consistent. All beverages
were matched for taste and flavor and were given to subjects at a
temperature of 3.3°C (see Table 1 for beverage composition).
Blood analysis. Blood draws were conducted using sterile aseptic
technique. An indwelling 20-gauge venous catheter (BD Insyte Autoguard Shielded IV Catheter, Sandy, UT) and needle-free syringe
system was maintained patent using a sterile saline (Bacteriostatic
0.9% sodium chloride injection, USP, Hospira, Lake Forest, IL
60045) lock in between blood draws. After each blood draw, ⬃3 ml
of sterile saline was used to flush and lock the catheter until the next
blood draw. Before each blood draw, the saline was removed and
discarded, and ⬃1.5 ml of blood was withdrawn and discarded before
obtaining the sample for analysis.
Two aliquots of blood were taken from each subject at each draw.
Blood from a heparinized syringe (⬃2 ml) was immediately used to
measure blood glucose, hematocrit, and hemoglobin using a blood gas
analyzer (GEM Premier 3000, Instrumentation Laboratories, Lexington,
247
INFLUENCE OF CARBOHYDRATE ON FLUID RETENTION
RESULTS
Fluid balance. Preexercise urine specific gravity, dehydration,
body mass loss, and total fluid intake did not differ between
treatments (see Table 2). A one-way ANOVA for the order of the
trials unrelated to treatments showed that there were no differences in the mean sweat loss. This supports the stability of the
physiological state of the subjects and the consistency of dehydration responses during the experimental period. The mean of the
coefficient of variation for each subject’s preexercise body weight
was 0.65%, further suggesting the consistency of the state of the
subjects at the start of each experimental trial.
The 12% carbohydrate beverage (P ⬍ 0.001) resulted in
significantly less absolute urine loss than all other treatments
(0.35 ⫾ 0.17 liter). P ⫹ E (0.58 ⫾ 0.16 liter) was not different
from P (0.64 ⫾ 0.21 liter), 3% (0.49 ⫾ 0.13 liter), or 6% (0.48 ⫾
0.28 liter) (see Figs. 1 and 2, respectively, for urine specific
gravity and urine loss at each time point).
For the primary outcome variable, the mean volume for
percent fluid retained was directionally greatest for the 12%
beverage but was not statistically different from that of 3% or
6%. Percent retention for P ⫹ E was not different from that of
P while the retention for both of these trials was less than that
for the 12% beverage (P ⬍ 0.05). Percent fluid retained was less
for P than all carbohydrate beverages (P ⬍ 0.001, see Fig. 3).
Using the mean and SD, we calculated a statistical power of
84% for this variable and an effect size of 0.43 for the ANOVA
results (3). The regression analysis showed that none of the
Fig. 1. Urine specific gravity (mean ⫾ SD) before exercise and throughout
recovery period. P ⫹ E, placebo with electrolyte. P, placebo without electrolyte. 3%, 6%, and 12% are carbohydrate solutions with electrolytes. P ⬍ 0.05
at following times: at 60 min, P ⫹ E, 3% ⬍ 6%; at 90 min, P, P ⫹ E ⬍ 6, 12%,
3% ⬍ 12%; at 120 min, P, P ⫹ E ⬍ 6, 12%, 3% ⬍ 12%; at 180 min, P, P ⫹
E ⬍ 6, 12%, 3, 6% ⬍ 12%; 240 min, P, P ⫹ E ⬍ 6, 12%.
variables considered accounted for significant variance in percent fluid retained. Final body mass was not significantly
different between trials.
Blood measures. Blood glucose was not different at baseline
or postexercise but was significantly lower at several time
points during postexercise recovery (see Fig. 4). Blood sodium
was not different pre- or immediately postexercise. At 90 min,
blood sodium (P ⬍ 0.001) was significantly lower for P and
P ⫹ E compared with all other treatments. P remained lower
than 12% at all time points.
Serum insulin was not different pre- or postexercise. Insulin
(P ⬍ 0.001) was significantly lower for P and P ⫹ E than all other
treatments. Insulin (P ⬍ 0.001) was significantly higher for 12%
at 180 and 240 min compared with all other treatments (see Fig. 5).
For all trials the relative plasma volume decreased on average 7.1 ⫾ 4.8% following exercise and dehydration and was
only significantly different at 90 min for 3% (P ⫽ 0.034)
during the recovery period (see Fig. 6).
Serum osmolality was not different pre- or postexercise for
any of the treatments. Osmolality (P ⬍ 0.001) was significantly
higher for the 12% beverage than all other beverages at 90 min
but was not significantly different at any other time point.
Serum aldosterone was not different pre- or postexercise or at
90 min postexercise. Aldosterone was significantly lower at 180
min (P ⫽ 0.001) and 240 min (P ⫽ 0.006) for P and 12%
compared with P ⫹ E and 3%. The 6% beverage was not different
from the others at any time point. However, the pattern of change
in mean values for serum aldosterone was similar for all trials.
Table 2. Preexercise urine specific gravity, body mass loss, and postexercise fluid intake
Preexercise USG
Preexercise body mass, kg
Postexercise body mass, kg
Change in mass, kg
Volume ingested, liters
Placebo
Placebo ⫹ Electrolytes
3%
6%
12%
1.008 ⫾ 0.005
77.76 ⫾ 6.92
75.77 ⫾ 6.85
1.99 ⫾ 0.40
1.99 ⫾ 0.39
1.007 ⫾ 0.003
77.86 ⫾ 6.85
75.76 ⫾ 6.61
2.10 ⫾ 0.38
2.10 ⫾ 0.38
1.007 ⫾ 0.004
77.76 ⫾ 6.75
75.71 ⫾ 6.61
2.05 ⫾ 0.43
2.05 ⫾ 0.43
1.008 ⫾ 0.005
77.84 ⫾ 6.77
75.76 ⫾ 6.60
2.07 ⫾ 0.41
2.07 ⫾ 0.41
1.006 ⫾ 0.003
77.97 ⫾ 7.01
75.93 ⫾ 6.91
2.04 ⫾ 0.40
2.04 ⫾ 0.41
Values are means ⫾ SD. USG, urine specific gravity.
J Appl Physiol • VOL
108 • FEBRUARY 2010 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.4 on June 14, 2017
MA). Hematocrit and hemoglobin values were used to calculate the
change in relative plasma volume using the method of Dill and Costill
(9). A second sample (⬃7 ml) was drawn into a nonheparinized vacutainer, allowed to stand for 20 min at room temperature, and centrifuged
for 15 min at 3,500 rpm. The serum was aliquoted and stored for
measurement of concentrations of sodium and potassium and osmolality
(Fiske 2400, Norwood, MA). Aliquots of serum were frozen at ⫺80°C
and later assayed for concentrations of insulin (ELISA assay LINCO
Research, St. Charles, MO), renin (SensoLyteTM 520, AnaSpec, San Jose,
CA), and aldosterone (ELISA, Alpha Diagnostics Intl, San Antonio TX).
Urine analysis. All urine samples were collected in disposable
specimen containers. Within 5 min of the collection, urine volume
was determined by mass and urine specific gravity was assessed (A
300 Clinical Refractometer).
Statistical analysis. SPSS version 14.0 was used to analyze the data.
ANOVA adjusted for two-way repeated measures (beverage treatment ⫻
time) was used to determine differences in mean values over time while
one-way ANOVA was used for variables measured one time. A backwards regression strategy was used to identify independent variables that
were associated with fluid retention including sweat losses, fluid intake,
and total carbohydrate. Effect size was calculated using the difference
between means for a comparison, divided by the pooled SD for those two
means (12). Data are reported as means ⫾ SD. A probability level of 0.05
was selected for statistical significance.
248
INFLUENCE OF CARBOHYDRATE ON FLUID RETENTION
Fig. 4. Blood glucose concentration (mean ⫾ SD) before exercise and during
postexercise recovery. P ⬍ 0.05 at the following times: at 90 min, 12% ⬎ P, P ⫹ E,
3%; at 180 min, 12% ⬎ P, P ⫹ E, 3%, 6%; at 240 min, 6, 12% ⬍ P, P ⫹ E.
Serum renin (P ⫽ 0.02) was significantly higher for all
carbohydrate beverages at 90 min compared with P and P ⫹ E.
The 12% beverage (10.17 ⫾ 9.69 ng/100 ml) approached
significance (P ⫽ 0.06) at the 240 min time point compared
with P (4.05 ⫾ 2.99 ng/100 ml).
GI and sensory ratings. No significant treatment ⫻ time
interactions were detected in any of the sensory or GI ratings.
others (30) who have shown that electrolytes influence fluid
retention, but our data also show that the greatest fluid retention tended to occur when the electrolytes are combined with a
12% carbohydrate-containing solution.
We proposed several mechanisms by which carbohydrate
dose could impact fluid retention. Increasing the carbohydrate
content of a beverage raises both the caloric content and
osmolality of the beverage. Both characteristics can reduce the
rate of gastric emptying (4, 5, 24) and rate of fluid absorption
(13, 32). A slow rate of fluid entering the vascular space would
sustain the body’s response to the state of dehydration and
modulate the bolus of fluid initially presented to the kidney
during rehydration. Beverages containing carbohydrate concentrations between 0 and 6% empty from the stomach at
similar rates (24, 28, 31) so we did not expect differences
between these beverage treatments. Somewhat in contrast to
our hypothesis was the difference (P ⫽ 0.034) in relative
plasma volume for the 3% beverage at 90 min (greater relative
plasma volume compared with that of other beverage treatments). Because of the lower osmolality of the 3% treatment
compared with the other carbohydrate-containing beverages
DISCUSSION
In this study, beverages containing carbohydrate and electrolytes contributed to greater fluid retention than did water
alone in recreational adult male athletes who rehydrated following exercise-induced dehydration. We did not find significant differences in fluid retention between beverages containing 3% and 6% carbohydrate and the placebo with electrolytes.
The solution containing 12% carbohydrate and electrolytes was
associated with the greatest fluid retention compared with the
placebo and placebo with electrolytes. There was no statistical
difference between placebo and placebo with electrolytes,
although a trend existed for greater fluid retention for the
electrolyte beverage (66.3 ⫾ 14.4% vs. 71.8 ⫾ 9.9%). Our data
are similar to Nielson et al. (25), Maughan and Leiper (21), and
Fig. 3. Percent fluid retained (mean ⫾ SD) during the recovery period. *P ⬍
0.05 for P ⬍ 3, 6, 12%. ^P ⬍ 0.05 for P and P ⫹ E ⬍ 12%.
J Appl Physiol • VOL
Fig. 5. Serum insulin concentration (mean ⫾ SD) before exercise and throughout recovery period. P ⬍ 0.05 the following times: at 90 min, P and P ⫹ E ⬍
3% ⬍ 6% ⬍12%; at 180 min and 240 min, P, P ⫹ E, 3%, 6%, ⬍ 12%.
108 • FEBRUARY 2010 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.4 on June 14, 2017
Fig. 2. Urine volume (mean ⫾ SD) at each time point during recovery period.
P ⬍ 0.05 at following times: at 90 min, 12% ⬍ P, P ⫹ E; at 180 min, 12% ⬍
P, P ⫹ E, 3, 6%, P ⬎ P ⫹ E, 3%, 6%; at 240 min, 12% ⬍ P, P ⫹ E, 6%.
INFLUENCE OF CARBOHYDRATE ON FLUID RETENTION
and by having transportable substrate not found in the carbohydrate-free treatments, the 3% treatment may have provided a
stimulus for absorption and an initial restoration of the relative
plasma volume measured at the 90-min point. While we are not
familiar with other data showing statistically significant differences between 3 and 6%, prior research shows trends for faster
fluid absorption and directionally greater relative plasma volume when beverages of lower carbohydrate (2 vs. 6%) (14) or
lower osmolality for the same absolute carbohydrate content
are ingested (13, 31), respectively. Beverages with 8% carbohydrate and higher appear to empty more slowly and retard
intestinal fluid uptake (8, 14, 32). While we did not measure
gastric emptying or absorption, the end point of percent of fluid
volume retained that we observed is consistent with the hypothesis: we found no difference in fluid retained for the 3%
and 6% carbohydrate beverage while the beverage with 12%
carbohydrate tended to retain more fluid [effect sizes of 0.8
compared with 3% and 0.4 compared with 6% (6)].
In parallel with the slower appearance of fluid in the vascular
space, the appearance of a fluid that provides a great amount of
carbohydrate would potentially contribute to a high blood osmolality, which could sustain the response to the state of dehydration
and support fluid retention. We did not see differences between
the beverages for overall restoration of relative plasma volume at
the end of the observation period. This may not be unreasonable
because the relative plasma volume can be restored even in the
absence of fluid ingestion (17). However, with the increase in
plasma glucose during the 12% carbohydrate treatment, we observed an elevation in plasma osmolality that would be expected
to stimulate vasopressin (2, 26). The rise in glucose would also
stimulate insulin, which has been shown to also increase urinary
sodium reabsorption, even in healthy individuals (29). We did
observe a higher insulin concentration for the 12% carbohydrate
treatment compared with the noncarbohydrate trials. Renin concentrations were also elevated for all carbohydrate trials with a
tendency for the elevation to be sustained for the highest carbohydrate treatment. These responses may have contributed to the
greater fluid retention for the 12% trial compared with effects of
the treatments devoid of carbohydrate.
We also hypothesized that fluid retention could be greater for
12% carbohydrate due to the water associated with the storage of
J Appl Physiol • VOL
carbohydrate as glycogen. Olsson and Saltin (27) found that total
body water increased an average of 2.2 liters with high vs. low
glycogen storage. The average amount of carbohydrate consumed
was 244 g for 12% compared with 124 g for 6% and 62 g for 3%.
It is possible that the 12% carbohydrate beverage resulted in
greater glycogen storage compared with beverages containing
little or no carbohydrate. Assuming a 3:1 ratio of water to
carbohydrate, the mean difference in fluid retention between the
12% and 6% beverage would theoretically be 360 g of water. We
found a mean difference of ⬃100 g of fluid retention between the
6 and 12% beverages but 178 g difference between P and P ⫹ E,
suggesting that glycogen storage was not a primary mechanism of
fluid retention in this study. There were no differences across
treatment in final body mass. However, mean body mass for the
12% beverage was 281 g more than the placebo ⫹ electrolytes
and 254 g greater than 6%. The differences between the other
beverages were ⬍30 g, indicating the 12% beverage caused the
body to retain more fluid. Nonetheless, without analyzing water
content of muscle biopsies or the volume of intracellular fluid
space using whole body tracer techniques, we cannot know for
sure the compartment in which the fluid was retained.
All efforts were made to minimize intrasubject variability by
requiring subjects to become heat acclimatized, report to each
study in a well-hydrated state based on urine specific gravity,
and control exercise, calories, and sodium intake during the
24-h period preceding each experimental trial. The urine specific gravity tests and consistency of preexercise body weights
for each subject (0.65% for the mean of the coefficient of
variation) provide confidence in our data. We did not test
subjects to confirm heat acclimation; however, by following
standardized published protocols (11, 18), standardizing the
exercise protocol to induce dehydration, and demonstrating no
statistical difference for exercise-induced loss in body mass for
the order of the trials, we are confident that variability in
factors that could affect the results, e.g., plasma volume,
muscle glycogen content, were minimized. Because the total
dose of carbohydrate for each individual varied dependent on
total fluid volume administered for rehydration (100% of body
mass change) and that volume was influenced by slight differences in intraindividual sweat rate, we used multiple regression
to account for subtle variations in the treatments to examine
more precisely the impact of carbohydrate dose on the percentage of fluid retained. Regardless, we were unable to detect a
clear association of fluid retention and carbohydrate dose. At
the end of each session of exercise after 30 min of rest, it was
assumed that subjects had reestablished some degree of homeostasis while remaining in a state of dehydration. We did
not measure core temperature or skin blood flow, so that lack
of confirmation remains a limitation of the study. However, by
design and with the consistency of the responses, we expected
this to have little or no impact on the internal validity of
beverage comparisons within the study.
The accuracy of using the change in body mass as a surrogate
for hydration status is debatable (23). A certain amount of the
change in body mass can be attributed to nonsweat losses
(respiratory water) and nonfluid loss (substrate that is oxidized). Some fluid would also be gained as substrate is
oxidized and lost, but water is formed in the process. Because
each subject performed the same workout and reported to the
study in the same nutrition and hydration state, the environment
was consistent across trials, and the error in using body mass for
108 • FEBRUARY 2010 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.4 on June 14, 2017
Fig. 6. Plasma volume (mean ⫾ SD) before exercise and throughout the postexercise recovery period. P ⬍ 0.05 at the following time: at 90 min, 3% ⬎ P, P ⫹
E, 6, 12%.
249
250
INFLUENCE OF CARBOHYDRATE ON FLUID RETENTION
DISCLOSURES
The study was conducted at the Gatorade Sports Science Institute, which is a part
of the Gatorade Company (funding source). C. A. Horswill is currently employed by
Gatorade.
REFERENCES
1. Adopo E, Peronnet F, Massicotte D, Brisson GR, Hillaire-Marcel C.
Respective oxidation of exogenous glucose and fructose given in the same
drink during exercise. J Appl Physiol 76: 1014 –1019, 1994.
2. Beretta-Piccoli C, Weidmann P, Bianchetti MG, Brown JJ, Fraser R,
Lever AF, Link L, Robertson JI. Effect of oral glucose loading on
plasma insulin, potassium, renin and aldosterone in normal subjects and
patients with primary hyperaldosteronism. Clin Exp Hypertens A 4:
1541–1558, 1982.
3. Borenstein M, Rothstein H, Cohen J, Shoenfeld D, Berlin J, Lakatos
E. Power and Precision. Englewood, NJ: Biostat, 2001.
4. Brouns F, Senden J, Beckers EJ, Saris WHM. Osmolarity does not
affect the gastric emptying rate of oral rehydration solutions. JPEN J
Parenter Enteral Nutr 19: 403–406, 1995.
5. Calbet JA, MacLean DA. Role of caloric content on gastric emptying in
humans. J Physiol 498: 553–559, 1997.
6. Cohen J. Statistical Power Analysis for the Behavioral Sciences. Hillsdale, New Jersey: Lawrence Erlbaum, 1988.
7. Committee on Nutrition AAoP. Oral fluid therapy and posttreatment
feeding after enteritis. edited by Barness LA. Elk Grove, IL: American
Academy of Pediatrics, 1993, p. 220 –226.
8. Davis JM, Burgess WA, Slentz CA, Bartoli WP. Fluid availability of
sports drinks differing in carbohydrate type and concentration. Am J Clin
Nutr 51: 1054 –1057, 1990.
9. Dill DB, Costill DL. Calculation of percentage changes in volumes of
blood, plasma, and red cells in dehydration. J Appl Physiol 17: 247–248,
1974.
10. Foster C, Jackson AS, Pollock ML, Taylor MM, Hare J, Sennett SM, Rod
JL, Sarwar M, Schmidt DH. Generalized equations for predicting functional capacity from treadmill performance. Am Heart J 107: 1229 –1234,
1984.
11. Fox RH, Goldsmith R, Hampton IF, Hunt TJ. Heat acclimatization by
controlled hyperthermia in hot-dry and hot-wet climates. J Appl Physiol
22: 39 –46, 1967.
12. Gardner MJ, Altman DG. Confidence intervals rather than P values:
estimation rather than hypothesis testing. Br Med J 292: 746 –750, 1986.
13. Gisolfi CV, Summers RW, Lambert GP, Xia T. Effect of beverage
osmolality on intestinal fluid absorption during exercise. J Appl Physiol
85: 1941–1948, 1998.
J Appl Physiol • VOL
14. Gisolfi CV, Summers RW, Shedl HP, Bleiler TL. Intestinal water
absorption from select carbohydrate solutions in humans. J Appl Physiol
73: 2142–2150, 1992.
15. Gonzalez-Alonso J, Heaps CL, Coyle EF. Rehydration after exercise
with common beverages and water. Int J Sports Med 13: 399 –406, 1992.
16. Greenleaf JE, Jackson GR, Geelen G, Keil LC, Hinghofer-Szalkay H,
Whittman JH. Plasma volume expansion with oral fluids in hypohydrated
men at rest and during exercise. Aviat Space Environ Med 69: 837–844,
1998.
17. Hayes PM, Lucas JC, Shi X. Importance of post-exercise hypotension in
plasma volume restoration. Acta Physiol Scand 169: 115–124, 2000.
18. Houmard JA, Costill DL, Davis JA, Mitchell JB, Pascoe DD, Robergs
RA. The influence of exercise intensity on heat acclimation in trained
subjects. Med Sci Sports Exerc 22: 615–620, 1990.
19. Jentjens RL, Shaw C, Birtles T, Waring RH, Harding LK, Jeukendrup AE. Oxidation of combined ingestion of glucose and sucrose during
exercise. Metabolism 54: 610 –618, 2005.
20. Jentjens RL, Underwood K, Achten J, Currell K, Mann CH, Jeukendrup AE. Exogenous carbohydrate oxidation rates are elevated after
combined ingestion of glucose and fructose during exercise in the heat. J
Appl Physiol 100: 807–816, 2006.
21. Maughan RJ, Leiper JB. Sodium intake and post-exercise rehydration in
man. Eur J Appl Physiol 71: 311–319, 1995.
22. Maughan RJ, Owen JH, Shirreffs SM, Leiper JB. Post-exercise rehydration in man: effects of electrolyte addition to ingested fluids. Eur J Appl
Physiol 69: 209 –215, 1994.
23. Maughan RJ, Shirreffs SM, Leiper JB. Errors in the estimation of hydration
status from changes in body mass. J Sports Sci 25: 797–804, 2007.
24. Murray R, Bartoli W, Stofan J, Horn M, Eddy D. A comparison of the
gastric emptying characteristics of selected sports drinks. Int J Sport Nutr
9: 263–274, 1999.
25. Nielsen B, Sjogaard G, Ugelvig J, Knudesn B, Dohlmann B. Fluid
balance in exercise dehydration and rehydration with different glucoseelectrolyte drinks. Eur J Appl Physiol 55: 318 –325, 1986.
26. Nutzi D, Beretta-Piccoli C, Ferrier CP, Link L, Gerber A, Weidmann
P. Studies on acute glucose-induced aldosterone suppression: role of
renin-angiotensin system. Klin Wochenschr 62: 213–217, 1984.
27. Olsson KE, Saltin B. Variation in total body water with muscle glycogen
changes in man. Acta Physiol Scand 80: 11–18, 1970.
28. Ploutz-Snyder L, Foley J, Ploutz-Snyder R, Kanaley J, Sagendorf K,
Meyer R. Gastric gas and fluid emptying assessed by magnetic resonance
imaging. Eur J Appl Physiol Occup Physiol 79: 212–220, 1999.
29. Quinones GA, Natali A, Baldi S, Frascerra S, Sanna G, Ciociaro D,
Ferrannini E. Effect of insulin on uric acid excretion in humans. Am J
Physiol Endocrinol Metab 268: E1–E5, 1995.
30. Ray ML, Bryan MW, Ruden TM, Baier SM, Sharp RL, King DS.
Effect of sodium in a rehydration beverage when consumed as a fluid or
meal. J Appl Physiol 85: 1329 –1336, 1998.
31. Rogers J, Summers RW, Lambert GP. Gastric emptying and intestinal
absorption of a low-carbohydrate sport drink during exercise. Int J Sport
Nutr Exerc Metab 15: 220 –235, 2005.
32. Ryan AJ, Lambert GP, Shi X, Chang RT, Summers RW, Gisolfi CV.
Effect of hypohydration on gastric emptying and intestinal absorption
during exercise. J Appl Physiol 84: 1581–1588, 1998.
33. Sechi LA, Bartoli E. Molecular mechanisms of insulin resistance in
arterial hypertension. Blood Press Suppl 1: 47–54, 1996.
34. Shirreffs SM, Aragon-Vargas LF, Keil M, Love TD, Phillips S.
Rehydration after exercise in the heat: a comparison of 4 commonly used
drinks. Int J Sport Nutr Exerc Metab 17: 244 –258, 2007.
35. Shirreffs SM, Taylor AJ, Leiper JB, Maughan RJ. Post-exercise
rehydration in man: effects of volume consumed and drink sodium
content. Med Sci Sports Exerc 28: 1260 –1271, 1996.
36. Tarnopolsky MA, Bosman M, MacDonald JR, Vandeputte D, Martin
J, Roy BD. Postexercise protein-carbohydrate and carbohydrate supplements increase muscle glycogen in men and women. J Appl Physiol 83:
1877–1883, 1997.
37. Vist GE, Maughan RJ. The effect of osmolality and carbohydrate content
on the rate of gastric emptying of liquids in man. J Physiol 486: 523–531,
1995.
38. Wemple RD, Morocco TS, Mack GS. Influence of sodium replacement
on fluid ingestion following exercise-induced dehydration. Int J Sport Nutr
7: 104 –116, 1997.
108 • FEBRUARY 2010 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.4 on June 14, 2017
hydration would be systematic and not invalidate our interpretation of the results. However, a source of error that could influence
our results is that the actual volume of water ingested for each
treatment would be systematically less as the carbohydrate concentration increased with the treatment. This could potentially
favor the impact of carbohydrate on fluid retention because the
vascular space and kidneys would receive a smaller volume of
fluid and presumably retain a greater amount. This also could have
contributed to the difference observed between the 12% carbohydrate treatment and the treatments without carbohydrate. Future
work might account for this by first matching the water needed to
replace fluid loss, and then add the carbohydrate in graded
amounts to the treatment.
In summary, our data suggest the presence of carbohydrate in
combination with the electrolytes used in this study promoted
fluid retention more so than ingesting water alone for rehydration
following exercise-induced dehydration. While we did not observe a clear dose effect for carbohydrate, a trend for the greatest
fluid retention occurred when electrolytes were combined with a
12% carbohydrate-containing solution. Lower concentrations of
carbohydrates appeared to exert only a mild influence on fluid
retention when electrolytes are standardized.