Content uploaded by Malachy P McHugh
Author content
All content in this area was uploaded by Malachy P McHugh on Dec 16, 2014
Content may be subject to copyright.
Original Research
Food Intake and Electrolyte Status of Ultramarathoners
Competing in Extreme Heat
Beth W. Glace, MS, Christine A. Murphy, PAC, and Malachy P. McHugh, PhD
Nicholas Institute of Sports Medicine and Athletic Trauma (B.W.G., M.P.M.), Lenox Hill Hospital (C.A.M.), New York, New York
Key words: ultramarathon, gastrointestinal distress, hyponatremia, fluid, endurance
Objective: To relate changes in laboratory indices to dietary intake during extremely prolonged running and
to determine if dietary intake influences the ability of runners to finish an 160 km trail race.
Methods: We monitored intake and serum chemistries of 26 runners competing in an 160 km foot race in
temperatures which peaked at 38°C. Blood was drawn pre-, mid- and post-race. Dietary intake and incidence of
gastrointestinal distress or changes in mental status were determined by interview with runners approximately
every 13 km. Twenty-three runners completed at least 88 kms and, of these 23 runners, 13 finished 160 km in
a mean time of 26.2 ⫾3.6 hours.
Results: Finishers ingested nearly 30,000 J, 19.4 ⫾8.1 L of fluid and 16.4 ⫾9.5 g of sodium (Na). Sodium
and fluid intake per hour was estimated to be 0.6 g/hour and 0.7 L/hour, respectively. Electrolyte intake during
the first half of the race was similar between those that finished the race and those that did not. Finishers ingested
fluid at a greater rate than non-finishers (p⫽0.01) and tended to meet their caloric needs more closely than did
non-finishers (p⫽0.09). Body weight was unchanged over time (ANOVA, p⫽0.52). Serum Na concentration
tended to fall from 143 to 140 mEq/L during the race (p⫽0.06), and was inversely correlated with weight loss
(p⫽0.009). Serum Na concentration was lower mid-race in runners experiencing changes in mental status than
in runners without changes (p⫽0.04). Fluid intake was inversely correlated with serum Na concentrations (p⫽
0.04). Most of the runners experienced nausea or vomiting; these symptoms were not related to serum sodium
concentration. Hyponatremia (⬍135 mEq/L) was seen in one runner at 88 kms, but resolved by 160 km. Urinary
sodium excretion decreased (p⫽0.002) as serum aldosterone concentration increased pre- to post-race (p⬍
0.001). From start to finish of the race plasma volume increased by 12%.
Conclusions: Food and fluid was ingested at a greater rate than described previously. Runners consumed
adequate fluid to maintain body weight although dietary sodium fell far short of the recommended 1 g/hour. The
rate of fluid intake was greater in finishers than in non-finishers, and finishers tended to more nearly meet their
energy needs. Maintenance of body mass despite large exercise energy expenditures in extreme heat is consistent
with fluid overload during a running event lasting more than 24 hours in hot and humid conditions.
INTRODUCTION
Little work has been done which prospectively and carefully
examines both food and fluid intake and serum chemistries in
ultra-endurance events. The choice of food and beverage be-
comes a crucial factor in performance during such events.
Carbohydrate intake during prolonged exercise has been shown
to improve time to exhaustion by providing exogenous energy
[1], while sufficient fluid intake is necessary to prevent dehy-
dration [2]. In addition, hyponatremia is well-documented during
prolonged exercise [3,4]. Hyponatremic runners seeking medical
assistance present with a variety of symptoms that range from
nausea, weakness, confusion, incoordination to grand mal seizures
and coma [5]. We have previously found that runners exercising
nearly continuously for 24 hours experienced nausea and changes
in mental status [6]. Sodium intake for these runners was much
less than recommended (0.5 g/hour vs. 1–2 g/hour) [7] while fluid
and total energy intake were greater than expected. Several run-
ners had also reported profound post-exercise diuresis. These
symptoms were suggestive of hyponatremia although we had not
assessed serum electrolytes.
The purpose of this study was to document changes in blood
Address reprint requests to: Beth W. Glace, M.S., Nicholas Inst. Sports Medicine and Athletic Trauma, Lenox Hill Hospital, 130 East 77th Street, New York, NY 10021.
E-mail: lb@nismat.org
Presented in part at the American College of Sports Medicine, Indianapolis, IN, June 2000.
Journal of the American College of Nutrition, Vol. 21, No. 6, 553–559 (2002)
Published by the American College of Nutrition
553
and urine electrolytes and to relate those changes to dietary
intake in runners engaged in a 160 km trail race. We further
sought to determine whether intake had an effect upon the
ability to complete the race and upon gastrointestinal function
or mental status.
METHODS
Subjects
Twenty-six volunteers with a mean age of 48.8 ⫾8.8 years
(SD) were recruited by letter three months prior to the event. To
be eligible, subjects had to have an anticipated finish time of
less than 24 hours by self-report. According to race results from
the previous two years, this gave us a pool of approximately 85
runners to recruit from and limited us to the top 50% of
competitors according to finish times. The 21 men and 5
women who agreed to participate were mailed questionnaires
regarding their medical and running/training history. The re-
search protocol was approved by the Research and Clinical
Investigations Committee of Lenox Hill Hospital, and written
informed consent was obtained prior to subject participation.
Participants were paid $100 at study completion.
Protocol
On the evening prior to the race, before the pre-race meal,
participants provided a urine sample that was placed on ice.
After reviewing the medical history, body mass was determined
in shorts (jog bra for women), with shoes, on a calibrated scale
and body fat was determined with Skindex skinfold calipers
(Caldwell, Justiss & Co., Fayetville, AR) that use a gender- and
age-adjusted formula [8,9]. Blood was drawn from an antecu-
bital vein within two minutes of being seated. An aliquot of
whole blood was analyzed within three hours for hemoglobin
(HemoCue, Lee Diagnostics AB, Medical) and hematocrit. The
remaining sample was centrifuged and separated. The serum
was frozen and later analyzed for aldosterone, sodium, potas-
sium, osmolality and glucose. Plasma volume changes were
calculated according to the methods described by Strauss et al.
using hemoglobin and hematocrit [10]. Runners were provided
with dietary intake records and instructed to record their food
intake for the 12 hours preceding the race, including the pre-
race meal. In addition, they were asked to record the type and
amount of non-steroidal anti-inflammatory drugs (NSAIDs)
pre-race.
The runners were met just prior to the 4:30 a.m. race start,
where the pre-race food records were collected and reviewed.
Temperatures on race day ranged from 21°–38°C. The race
course consisted primarily of dirt roads and trails with 4500
meters of ascents and descents. There were 37 official food
stations spaced throughout the course. Investigators manned 12
food stations, at approximately 13 km intervals. A checklist of
supplied food items was used to record the type and quantity of
food, beverages and NSAIDs taken by the runners at each of
the observed stations. Since runners could eat from food sta-
tions other than the ones monitored by investigators and could
eat foods supplied by support crews, they were also asked to
recall everything they had ingested since we last interviewed
them. Runners were asked to keep any wrappers from foods
eaten on the trail until they next saw us, to help account for
their intake. At each interview the runners were asked if they
had experienced any gastrointestinal symptoms (GiSx) i.e. nau-
sea, vomiting, diarrhea, and were observed for changes in
mental status (MSx) such as confusion, disorientation, or in-
ability to concentrate.
Blood samples were collected again at 88 km and at the
finish line and analyzed or frozen, as described above. Urine
collected pre- and post-race was frozen and later analyzed for
electrolytes and osmolality. Body mass was assessed pre-race,
at 70 km, 110 km, 133 km and immediately after the finish,
after voiding and prior to taking any further food or fluid.
Skinfold measurements were repeated after towel drying. En-
ergy expenditure during the event was estimated using a step-
wise calculation using body mass, mean velocity and time spent
running [11].
Dietary Analysis
The NutriBase Nutrition Management Software, Clinical
Version, (CyberSoft, Inc. Phoenix, AZ) was used to calculate
the nutritional composition of the foods and fluids consumed.
Runners were subsequently interviewed by phone if clarifica-
tion was required, e.g. to obtain brand names of specialized
“sport”supplements. The manufacturers were contacted to
provide nutritional composition of these products, and the data
was added to the nutritional software database. Where possible,
food items were obtained, and the food was weighed in our
laboratory according to typical serving size. For example, we
standardized “handfuls”of a variety of food items by having
several staff members repeatedly take a handful of the specified
food. The food was then weighed, and the mean weight was
used as a typical “handful”. Diets were analyzed for total
calories, macro- and micro-nutrients as well as for total mois-
ture content.
Statistical Analysis
Anthropometric measurements, dietary intake and blood
and urine analysis were examined. A one-way, repeated mea-
sure ANOVA was used to examine changes over the duration
of the race. Two-way mixed model ANOVAs were used to
examine changes overtime between runners: (1) with/without
MSx; (2) finishers and non-finishers. Because non-steroidal
anti-inflammatory drugs (NSAIDs) are associated with GI dis-
tress, ANCOVA with NSAID intake as a confounding variable
was used to compare dietary intake between GISx and No GISx
groups. Statistical significance was established for an alpha
level of 0.05. Bonferroni corrections were used for all post-hoc
Food and Fluid Intake during an Ultramarathon
554 VOL. 21, NO. 6
pair-wise comparisons. All results are reported as means with
standard deviations.
RESULTS
Anthropometrics
Thirteen of the 26 subjects completed the 160 km course, in
a mean time of 26.2 ⫾0.4 hours. Of the non-finishers, all
subjects completed a minimum of 45 km. Eleven of the 13
non-finishers had body mass determined at 70 km. However,
most of the non-finishers dropped out of the race while on the
trail; final body mass determination and blood sampling could
not always be performed at the time the runner dropped out,
unless they happened to stop at a monitored aid station. The
mean change in weight from race start till the last time
weighted during the race, or until the finish, was ⫺0.97 ⫾1.8
kg. Finishers and non-finishers did not differ in the magnitude
of weight change, ⫺0.5 ⫾1.5 kg vs. ⫺1.2 ⫾1.8 kg (p⫽0.27),
respectively. Surprisingly, there was no significant change in
body mass over time in the finishers (p⫽0.52). Finishers
decreased mass by 0.5 kg, representing a 0.6% decline from
pre-race values. Although finishers tended to weigh less pre-
race than non-finishers (66.5 kg vs. 73.8 kg, p⫽0.09), the
mean percent body fat did not differ between groups (12.8 ⫾
6.1% vs. 15.5 ⫾5.9%, p⫽0.22, respectively). There was no
significant change in body fat from pre- to post-race (12.8 ⫾
6.1% vs. 11.9 ⫾3.7%, p⫽0.29).
Dietary Intake
An analysis of dietary records collected for the 12 hours
pre-race revealed that 8986 joules were ingested, of which
carbohydrate intake accounted for 59% (Table 1). Three liters
(⫾1.4) of fluid and 3.2 g (⫾1.9) of sodium were consumed
pre-race. Intake is expressed absolutely and per km/kg, in order
to include runners that did not complete the entire course.
Runners who completed 160 km ingested sodium, protein and
fluid at a greater rate than did non-finishers. Carbohydrates
provided 81% of the 29493 J ingested by those completing the
race. Finishers drank a mean of 19.4 ⫾5.6 L, with a range of
11.9 to 28.1 L. Sodium intake was 16.4 ⫾6.8 g, with a range
of 4.9 to 27.5 grams. These were consumed at rates of 0.7
L/hour and 0.6 g/hour, respectively. Notably, 16% of the total
fluid intake was contributed by the moisture contained in solid
foods. The greatest rates of fluid and carbohydrate intake
occurred during the 2nd and 3rd quarters of the race (Fig. 1).
The rate of fluid intake was greater for those runners who went
on to complete the entire course, as compared to those who did
not finish, when adjusted for body weight (1.8 ⫾0.49 vs. 1.3 ⫾
0.4 mL/kg/km, p⫽0.011).
The rates of energy expenditure and intake were calculated
for both finishers and non-finishers. Expenditure was substan-
tially greater than energy intake, 40,061 vs. 21,569 J respec-
tively, resulting in a mean deficit of ⫺161 J per km completed.
The energy deficit per km was greater for non-finishers than for
non-finishers (⫺190.6 vs. ⫺132.1 J/km, p⫽0.01).
Blood and Urine Analysis
Twelve of the 13 finishers and nine of the 13 non-finishers
had blood drawn mid-race. Glucose concentration were
90(5.0), 97(5.4), and 95 mg/dL (5.3 mmol/L) pre-, mid- and
post-race, i.e. there was no evidence of hypoglycemia. Serum
potassium concentration was 4.15 mEq/L, 5.85 and 4.18 mEq/L
pre-race, mid-race and post-race, respectively (effect of time
p⬍0.001). Serum sodium concentrations tended to fall from
143.9 mEq/L pre-race, to 140.8 mEq/L mid-race, and to 140.2
mEq/L post-race (p⫽0.06). Hyponatremia, as defined by a
serum Na concentration of less than 135 mEq/L, was observed
in only one runner, at 90 km. The runner complained of both
vomiting, which began at 12 km, and confusion, beginning at
88 km. The runner’s serum Na concentrations were as follows
pre-, mid-, and post-race: 144 mmol/L, 133 mmol/L and 136
mmol/L, respectively. Changes in body mass inversely paral-
leled the changes in serum Na; his mass was 71.2 kg pre-, 73.9
kg mid-, and 73.0 kg post-race. His mean Na and fluid intakes
were 17.1 g and 24.2 L, respectively. Anecdotally, the runner
reported that he had not urinated prior to his episode of con-
fusion, but that after drinking a beer during the race he began
voiding and “felt much better.”
Table 1. Nutrient Intake
Joules Fluid (L) Carb (g) Fat (g) Protein (g) Fiber (g) Na (g) K (g)
Pre-Race (n ⫽26) 8986 ⫾3359 3.2 ⫾1.4 318 ⫾144 70 ⫾39 67 ⫾66 31.7 ⫾8.7 3.2 ⫾1.9 1.9 ⫾1.0
Race All (n ⫽26) 21589 ⫾12803 14.1 ⫾6.5 1084 ⫾609 95 ⫾73 84 ⫾62 23 ⫾16 11.4 ⫾7.4 5.1 ⫾3.3
Race Finishers
(n ⫽13) 29493 ⫾3481 19.4 ⫾5.6 1419 ⫾622 137 ⫾75 128 ⫾56 32 ⫾15 16.4 ⫾6.8 7.2 ⫾3.1
Rate of Intake (mL) (g) (mg) (mg) (mg) (mg) (mg)
All /kg/km (n ⫽26) 2.15 ⫾1.07 1.57 ⫾0.5 0.13 ⫾0.07 0.01 ⫾0.007 0.009 ⫾0.006 0.003 ⫾0.002 1.3 ⫾0.7 0.06 ⫾0.03
Non-Finishers /kg/km
(n ⫽13) 2.15 ⫾0.99 1.33* ⫾0.41 0.12 ⫾0.07 0.008 ⫾0.007 0.006* ⫾0.004 0.002 ⫾0.002 0.94* ⫾0.58 0.06 ⫾0.04
Finishers Per /kg/km
(n ⫽13) 2.15 ⫾1.18 1.8* ⫾0.49 0.14 ⫾0.07 0.01 ⫾0.007 0.01* ⫾0.006 0.003 ⫾0.001 1.60* ⫾0.69 0.06 ⫾0.03
* Finishers and non-finishers, pⱕ0.01. Values are ⫾SD.
Food and Fluid Intake during an Ultramarathon
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION 555
There was no relationship between total sodium intake and
post-race serum (r ⫽0.07, p⫽0.83) or urinary sodium (r ⫽
0.41, p⫽0.17). Post-race serum sodium was affected, how-
ever, by fluid intake during the second half of the race (r ⫽
⫺0.58, p⫽0.04) (Fig. 2a); high fluid intakes were correlated
with lower post race serum sodium. Post-race urinary sodium
also tended to be inversely correlated with 2nd half fluid
intakes (r ⫽⫺0.53, p⫽0.06), (Fig. 2b). No relationship was
seen between dietary potassium intake and serum or urinary
potassium levels.
Loss of body weight was inversely correlated with changes
in serum sodium (r ⫽0.58, p⫽0.009). There was no change
in serum osmolality from pre- to post-race (290 ⫾11 vs. 283 ⫾
12 mOsm/kg H
2
O, p⫽0.15). Serum aldosterone increased
significantly during the race, from 8.3 ⫾4.9 to 49.2 ⫾3.1 (p⫽
0.001), and mean sodium excretion fell from 74.1 mEq/L
pre-race to 19.4 mEq/L post-race (p⫽0.002). A significant
difference in plasma volume changes was observed between
finishers and non-finishers mid-race. By 90 km non-finishers
experienced a 5% decrease in plasma volume while those who
went on to complete the course had expanded plasma volume
by 3% (Fig. 3). Overall, finishers increased plasma volume by
11.8% from pre- to post-race.
Symptomatology
Gastrointestinal symptoms were experienced by 17 of the
26 participants. Upper Gi symptoms were far more prevalent
than lower, with 15/26 describing nausea or vomiting but only
3/26 runners describing diarrhea or intestinal cramping. Run-
ners with GiSx had greater energy intake, and specifically
greater carbohydrate intake, than did runners not experiencing
symptoms (Table 2). The rate of intake of other nutrients was
similar between GiSx and No GiSx groups. Most runners used
NSAIDs (18/26); however, their use was not associated with
GiSx (p⫽0.50). There were no differences in body mass
changes or in serum sodium changes over time between those
who did and those who did not use NSAIDs. Runners without
GiSx had completed longer runs during training or racing than
had runners with symptoms, 112 vs. 68 km (p⫽0.02).
Nine of 26 runners complained of MSx, which ranged in
severity from mild confusion to hallucinations. Dietary intakes
were not different between runners with and without MSx. Of
the finishers, mid-race serum Na was lower in those reporting
MSx (137.5 mg/dL) compared to those with no MSx (141.9
Fig. 1. Intake per quarter. Fluid: effect of time, p⫽0.04; Carbohydrate:
effect of time, p⫽0.004. Values are ⫾SD.
Fig. 2. Relationship of fluid intake to post-race serum and urinary
sodium concentrations.
Fig. 3. Percent change in plasma volume from pre-race to mid-race for
finishers and non-finishers *p⫽0.047, and from pre- to post-race.
Values are ⫾SD.
Food and Fluid Intake during an Ultramarathon
556 VOL. 21, NO. 6
mg/dL, p⫽0.04). The rate of serum sodium change did not
differ between MSx (⫺0.037 mg/dL/km) and no MSx (⫺0.038
mg/dL/km). The rate of dietary potassium intake was lower in
seven runners who complained of muscle cramping (27.2 mg/
km) compared to those without cramping (46.8 mg/km, p⫽
0.046). There was no association between cramping and so-
dium or fluid intake prior to, or during, the race.
DISCUSSION
While other studies have examined dietary intake during
prolonged exercise, many have used small sample populations
[12,13], were limited by using diet recall methods immediately
following the race [12] or used dietary data collected during
subsequent telephone surveys [14]. While conveniently ap-
plied, these methods are associated with inaccuracies in dietary
data and might be expected to be even less reliable after
physically exhausting exercise. Furthermore, the ultra-endur-
ance events described were of far shorter duration than the
event being reported upon here [13,14]. This study is unique in
that we repeatedly interviewed athletes regarding, and ob-
served, food intake throughout more than 24 hours of contin-
uous exercise.
Our sample population was fairly representative of the field
of entrants: the mean finish times were 26.2 vs. 25.1 hours,
respectively. In previous years the finish rate for this race has
been approximately 75%; in 1999 the rate of completion was
only 44% for the overall field of entrants and 50% for those
runners for whom we collected data. The attrition rate was
likely due to the extreme environmental conditions of high
humidity and temperature.
Our most important finding was that body mass was very
well-maintained throughout the race, despite extreme energy
expenditure and thermal stress. A 3% to 4% decrease in body
mass during ultra-marathons has been reported previously [15].
It has been suggested that failure to lose approximately 2 kg
might actually represent fluid overload since the utilization of
fat and glycogen stores should result in net loss of mass. The
estimated energy deficit during this race ought to have resulted
in a loss of approximately 0.64 kg of body mass, if all of the
unmet energy needs came at the expense of body fat [17]. The
estimated expense of body fat would account for all of the
measured change in body mass and would suggest that runners
experienced a slight relative gain in body fluids. Recommen-
dations for fluid consumption during ultra-marathons have var-
ied from 0.5 to 1 L/hour [18] to 0.6 to 2.4 L/hour [2]. Our data
demonstrates that an intake of 0.74 liters of moisture per hour
was more than sufficient to maintain body weight during such
prolonged but low-intensity exercise and supports recommen-
dations of fluid consumption of no more than 1 L/hour in
ultra-marathons.
Fluid overload has been suggested to be the primary factor
contributing to low serum sodium levels [19]. It has been
estimated that ultra-distance athletes would need to lose about
4% of their body weight in order to maintain serum sodium [19,
20]. We observed a strong indirect relationship between weight
change and serum sodium. Weight maintenance may represent
an inappropriate retention of fluid in the vascular space, as
suggested by the expansion of blood volume. Despite the cooler
night temperatures and decreased pace during the latter stages
of the race, fluid intake was similar between the first and
second halves of the event. It is feasible that 19 liters of fluid
ingested in one day exceeded sweat rate and the kidney’s
capacity to excrete excess fluid in some individuals.
Goldberger has recommended that athletes engaged in very
prolonged exercise ingest 1 g/hour of sodium [7]. The athletes
we observed ate approximately half that amount. Serum sodium
concentrations overall tended to fall throughout the race, and
plasma volume was expanded under environmental conditions
which would be expected to result in a contraction of plasma
volume and in hemoconcentration [21]. While serum sodium
would typically be expected to drive plasma expansion, Gast-
mann et al. have suggested that plasma expansion during 24
hours of exercise may result from a carbohydrate-induced os-
motic gradient [22]. Serum Na concentrations declined inde-
pendently of serum osmolality; the high rate of carbohydrate
intake may have contributed to serum osmolality and to the
retention of fluid in the vascular space [19, 23, 24]. In a
previous study at the same race we found that runners ingested
nearly identical amounts of moisture and sodium while com-
peting under far more mild conditions. We suspect that had
conditions been less severe plasma volume may have ex-
panded, and serum Na concentration declined, to an even
greater degree.
Another important finding of this study is that runners
seemed willing and able to ingest far more food than has been
described. We believe that the extreme duration of exercise,
Table 2. Nutrient Intake and Gastrointestinal Distress
Joules
(J/kg/km) CHO
(g/kg/km) Fat (g/kg/km) Protein
(g/kg/km) Fluid
(mL/kg
⫺1
䡠km
⫺1
)Na
(mg 䡠kg
⫺1
䡠km
⫺1
)
GI 0.61 ⫾0.06# 0.15 ⫾0.02* 0.01 ⫾0.002 0.009 ⫾0.002 1.6 ⫾0.1 1.4 ⫾0.2
No GI 0.33 ⫾0.08# 0.09 ⫾0.02* 0.01 ⫾0.002 0.009 ⫾0.002 1.5 ⫾0.1 1.1 ⫾0.2
# Groups significantly different, p⫽0.005.
* Groups significantly different, p⫽0.02. Values are ⫾SD.
Food and Fluid Intake during an Ultramarathon
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION 557
and therefore the reduced intensity of exercise, greatly influ-
enced the amount and type of foods ingested. Running pace
averaged 6.5 km/hour overall and only 5.7 km/hour in the
second half, as compared to that of participants in a 100 km run
who averaged nearly 10 km/hour [13]. While runners engaged
in higher intensity exercise often rely heavily upon the carbo-
hydrate contained in liquids or fruits to provide energy [25],
runners in this study ate substantial quantities of solid food,
including items such as fried chicken and peanut butter sand-
wiches, obtaining 17% of total energy from fats. Finishers
consumed approximately 30,000 Joules, with a range of 11,832
to 47,359 Joules, at a mean rate of 176 J/km. This is similar to
the data collected during a previous 160 km race [6], during the
Alaskan Iditasport [12] and the Tour de France [26]. However,
the observed rate we observed was fourfold greater than de-
scribed by others during running races [13].
Dietary intake may have affected the ability of runners to
finish the race. The deficit between estimated energy expended
and energy consumed tended to be larger in non-finishers
compared to finishers: finishers tended to more nearly meet
their energy needs. Rehrer et al. estimated fluid intake to be
56.8 mL/km during a 67 km race, or half the rate we observed
[14]. Our volunteers consumed a large amount of solid food
which contributed a significant amount of moisture and may, in
part, explain the high fluid intakes.
The incidence of mental status changes was quite high
during the race. We found that mental status changes were
associated with lower serum Na concentration and that serum
Na concentrations were associated with high fluid intakes.
Serum sodium concentration was lower mid-race in those ath-
letes experiencing MSx, but there was no difference in dietary
intake of sodium between those that did and did not report
MSx. While high fluid intake increased the likelihood of com-
plaints of mental status change, finishers did ingest fluid at a
greater rate than those who dropped out of the race. The mental
status changes, then, were not typically severe enough to result
in withdrawal from the competition. In a previous study of
ultra-marathoners competing on the identical course in less
severe conditions we similarly found that those with mental
status changes ingested more fluid than those without symp-
toms, but that overall finish time was not affected by the
occurrence of mental status change [6].
Although dehydration has been postulated to contribute to
GI distress [14], we found no relationship between markers of
dehydration (i.e. weight loss, osmolality) and GI symptoms.
The lack of a relationship could be due to the minimal dehy-
dration experienced by runners in this study. Hypertonic car-
bohydrate beverages have been shown to increase risk of GI
problems in triathletes [25]. We did find an increased risk of
developing GI symptoms with increasing caloric intake, spe-
cifically of carbohydrate intake. The rate of carbohydrate intake
was 54 g/hour, plus 5 g/hour of fat and protein, an intake which
exceeds the limit suggested by Kreider et al. during ultra-
distance running, of 15–50 g/hour [2]. As with the mental status
change, however, most of the GI symptoms were not severe
enough to warrant withdrawal from the race.
Rehrer et al. and Keefe et al. reported that lower GI symp-
toms were far more prevalent than upper GI symptoms in
shorter ultra-marathon events [14,27]. However, we found a
much greater incidence of moderate to severe upper GI distress,
i.e. nausea or vomiting, than of lower GI symptoms. The high
incidence of upper GiSx may be a function of race duration [6].
Better trained runners were less likely to experience GI distress.
A protective effect of training on the incidence of GI symptoms
has been reported previously [6,28]. Brocke et al. found that
ultra-distance athletes often experience endotoxemia during
very prolonged exercise with symptoms of nausea and vomit-
ing, secondary to increased intestinal permeability [29]. Inter-
estingly, better trained runners were found to have better im-
munity to endotoxins and fewer symptoms. The authors
speculated that training may confer a protective effect by
increasing antibodies to endotoxins.
CONCLUSION
We studied food and moisture intake, markers of serum
electrolyte status and disturbances in gastrointestinal and men-
tal status in runners engaged in 21 to 30 hours of near contin-
uous exercise. The slow pace in this long duration race may
have resulted in low sweat rates, while simultaneously allowing
for greater food and fluid consumption than previously re-
ported. Runners ate more than has been previously described in
running races and ingested sufficient fluid to maintain body
weight. Although high fluid intakes were associated with de-
creased serum sodium and increased risk of mental status
change, finishers ingested fluid at a greater rate than those who
did not complete the race and more closely met their energy
requirements. Collectively, we observed a minimum of weight
loss, a high incidence of upper gastrointestinal distress with
high carbohydrate intake and an expansion in plasma volume
despite extreme environmental conditions and exercise stress.
ACKNOWLEDGMENT
The authors wish to thank Maria Pagano for her enthusiastic
assistance despite severe sleep deprivation, and Dr. Maria
Devita for lending her expertise in renal handling of fluids and
electrolytes.
REFERENCES
1. Coyle E, Coggan A, Hemmert M, Ivy J: Muscle glycogen utiliza-
tion during prolonged strenuous exercise when fed carbohydrate.
J Appl Physiol 61:165–172, 1986.
Food and Fluid Intake during an Ultramarathon
558 VOL. 21, NO. 6
2. Kreider R: Physiologic considerations of ultraendurance perfor-
mance. Int J Sports Nutr 1:3–27, 1991.
3. Backer H, Shopes E, Collins S, Barkan H: Exertional heat illness
and hyponatremia in hikers. Am J Emerg Med 17:532–539, 1999.
4. Frizzel R, Lang G, Lowance D, Laham S: Hyponatremia and
ultramarathon running. JAMA 255:772–774, 1986.
5. Surgenor S, Uphold R: Acute hyponatremia in ultra-endurance
athletes. Am J Emerg Med 12:441–444, 1994.
6. Glace B, Murphy C, McHugh M, Nicholas S: Food and fluid
intake: Relationship to gastrointestinal and mental function During
an ultramarathon. Int J Sports Nutr, in press, 2002.
7. Goldberger E: “A Primer of Water, Electrolyte and Acid-Base
Syndromes.”Philadelphia: Lea and Febinger, 1986.
8. Jackson AS, Pollock ML, Ward A: Generalized equations for
predicting body density of women. Med Sci Sports Exerc 12:175–
181, 1980.
9. Jackson AS, Pollock ML: Generalized equations for predicting
body density of men. Br J Nutr 40:497–504, 1978.
10. Strauss M, Davis J, Rosenbaum J, Eoaamwial E: “Water diuresis”
produced during recumbency by the intravenous infusion of iso-
tonic saline solution. J Clin Invest 30:862–868, 1951.
11. “ACSM: Guidelines for Exercise Testing and Prescription,”3rd ed.
Philadelphia, PA: Williams and Wilkins, p 277, 1995.
12. Case S, Evans D, Tibbets G, Case S, Miller D: Dietary intakes of
participants in the IditaSport Human Powered Ultra-Marathon.
Alaska Med. 37:20–24, 1995.
13. Fallon K, Broad E, Thompson M, Reull P: Nutritional and fluid in
a 100-km ultramarathon. Int J Sports Nutr 8:24–35, 1998.
14. Rehrer N, Brouns F, Beckers E, Frey W, Villiger B, Riddoch C,
Menheere P, Saris W: Physiologic changes and gastrointestinal
symptoms as a result of ultra-endurance running. Eur J Appl
Physiol 64:1–8, 1992.
15. Speedy DB, Noakes TD, Kimber NE, Rogers IR, Thompson JMD,
Boswell DR, Ross JJ, Campbell RGD, Gallagher PG, Kuttner JA:
Fluid balance in the Ironman triathlon. Clin J Sport Med 11:44–50,
2001.
16. Speedy D, Campbell R, Mulligan G, Robinson D, Walker C,
Gallagher P, Arts J: Weight changes and serum sodium concen-
trations after an ultradistance multisport triathlon. Clin J Sports
Med 7:100–103, 1997.
17. Katch FI, McArdle WD: “Nutrition, Weight Control, and Exer-
cise.”Philadelphia, PA: Lea & Febiger, pp 155–178, 1988.
18. Burke LM: Fluid balance during team sports. J Sports Sci 15:287–
295, 1997.
19. Speedy D, Noakes T, Rogers I, Thompson J, Campbell R, Kuttner
J, Boswell D, Wright S, Hamlin M: Hyponatremia in ultradistance
triathletes. Med Sci Sports Exerc 31:809–815, 1999.
20. Speedy DB, Rogers IR, Noakes TD, Wright S, Thompson JM,
Campbell R, Hellemans I, Kimber NE, Boswell DR, Kuttner JA,
Safih S: Exercise-induced hyponatremia in ultradistance triathletes
is caused by inappropriate fluid retention. Clin J Sport Med 10:
272–278, 2000.
21. Hiller W: Dehydration and hyponatremia during triathlons. Med
Sci Sports Exerc 21(5S):S219–S221, 1989.
22. Montain S, Laird J, Latzka W, Sawka M: Aldosterone and vaso-
pressin responses in the heat: hydration level and exercise intensity
effects. Med Sci Sports Exerc 29:661–668, 1997.
23. Gastmann U, Dimeo F, Huonker M, Bocker J, Steinacker J, Pe-
tersen K, Wieland H, Keul J, Lehmann M: Ultratriathlon-related
blood-chemical and endocrinological responses in nine athletes.
J Sports Med Phys Fitness 38:18–23, 1998.
24. Noakes T, Norman R, Buck R, Godlonton J, Stevenson K, Pitt-
away D: The incidence of hyponatremia during prolonged ultra-
endurance exercise. Med Sci Sports Exerc 22:165–170, 1990.
25. Rehrer N, van Kemenade M, Meester W, Brounds F, Saris W:
Gastrointestinal complaints in relation to dietary intake in triath-
letes. Int J Sports Nutr 2:48–59, 1992.
26. Saris W: The Tour de France: food intake and energy expenditure
during extreme sustained exercise. Cycling Sci 12:17–28, 1990.
27. Keefe E, Lowe D, Goss J, Wayne R: Gastrointestinal symptoms of
marathon runners. West J Med 141:481–482, 1984.
28. Brouns F, Beckers E. Is the gut an athletic organ? Digestion
absorption and exercise. Sports Med 15:242–257, 1993.
29. Brocke-Utne J, Gaffin L, Wells M, Gathiram P, Sohar E, James M,
Morell D, Norman R: Endotoxaemia in exhausted runners after a
long-distance race. S Afr Med J 73:533–536, 1988.
30. Knochel J: Clinical complications of body fluid and electrolyte
imbalance. In Buskirk E, Puhl S (eds): “Body Fluid Balance
Exercise and Sport.”Boca Raton, FL: CRC Press, pp 297–317,
1996.
Received December 24, 2001; revision accepted July 3, 2002.
Food and Fluid Intake during an Ultramarathon
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION 559
