Eur J Appl Physiol (2010) 108:11331140

DOI 10.1007/s00421-009-1329-6

ORIGINAL ARTICLE

A low carbohydrate diet affects autonomic modulation

during heavy but not moderate exercise
Adriano Eduardo Lima-Silva Romulo C. M. Bertuzzi Flavio O. Pires
Lenise Fronchetti Monique S. Gevaerd Fernando R. De-Oliveira

Accepted: 7 December 2009 / Published online: 20 December 2009

Springer-Verlag 2009

Abstract The aim of this study was to examine the effects

of low carbohydrate (CHO) availability on heart rate
variability (HRV) responses during moderate and severe
exercise intensities until exhaustion. Six healthy males
(age, 26.5 6.7 years; body mass, 78.4 7.7 kg; body
_ 2max ; 39.5 6.6 mL kg-1 min-1)
fat %, 11.3 4.5%; VO
volunteered for this study. All tests were performed in the
morning, after 812 h overnight fasting, at a moderate
intensity corresponding to 50% of the difference between
the first (LT1) and second (LT2) lactate breakpoints and at a
severe intensity corresponding to 25% of the difference
between the maximal power output and LT2. Forty-eight
hours before each experimental session, the subjects performed a 90-min cycling exercise followed by 5-min rest
_ 2max
periods and subsequent 1-min cycling bouts at 125% VO
(with 1-min rest periods) until exhaustion, in order to deplete
muscle glycogen. A diet providing 10% (CHOlow) or 65%
(CHOcontrol) of energy as carbohydrates was consumed for

the following 2 days until the experimental test. The Poicare

plots (standard deviations 1 and 2: SD1 and SD2, respectively) and spectral autoregressive model (low frequency
LF, and high frequency HF) were applied to obtain HRV
parameters. The CHO availability had no effect on the HRV
parameters or ventilation during moderate-intensity exercise. However, the SD1 and SD2 parameters were significantly higher in CHOlow than in CHOcontrol, as taken at
exhaustion during the severe-intensity exercise (P \ 0.05).
The HF and LF frequencies (ms2) were also significantly
higher in CHOlow than in CHOcontrol (P \ 0.05). In addition,
ventilation measured at the 5 and 10-min was higher in
CHOlow (62.5 4.4 and 74.8 6.5 L min-1, respectively,
P \ 0.05) than in CHOcontrol (70.0 3.6 and 79.6 5.1
L min-1, respectively; P \ 0.05) during the severe-intensity exercise. These results suggest that the CHO availability
alters the HRV parameters during severe-, but not moderate-,
intensity exercise, and this was associated with an increase
in ventilation volume.

Communicated by Susan Ward.

Keywords Heart rate variability

Carbohydrate availability
Ventilation

Moderate and severe exercise

A. E. Lima-Silva (&)
Sports Science Research group, Centro de Educacao,
Universidade Federal de Alagoas, Avenida Lorival Melo Mota
S/N, Maceio, Alagoas CEP 57072-970, Brazil
e-mail: adrianosilva@usp.br
R. C. M. Bertuzzi
F. O. Pires
School of Physical Education and Sport,
University of Sao Paulo, Sao Paulo, Brazil
L. Fronchetti
M. S. Gevaerd
Center of Health Science and Sport, University of Santa Catarina
State, Florianopolis, Brazil
F. R. De-Oliveira
Nucleus of Studies of Human Movement,
Federal University of Lavras, Lavras, Brazil

Introduction
The heart rate (HR) during rest and exercise is regulated by
an integrated system that includes the sinus node, as well as
the sympathetic and parasympathetic (vagal) nervous system. The increase in HR during exercise results from vagal
withdrawal and activation of the sympathetic nervous
system. The initial rise in HR during dynamic exercise at
moderate intensities appears to be mediated by the vagal
withdrawal. On the other hand, as the exercise intensity

123

1134

becomes more severe, the rise in HR is mediated by the

activity of cardiac sympathetic nerves (Tulppo et al. 1999).
Although reduced, vagal modulation remains activated
during moderate-intensity exercise, but it is considerably
diminished during severe exercise (Breuer et al. 1993;
Tulppo et al. 1996). As a result of the fine balance between
vagal and sympathetic activity, the HR increases as exercise intensity increases (Tulppo et al. 1996, 1999; Pichon
et al. 2004; Cottin et al. 2004).
Changes in autonomic regulation of the HR during
exercise have been frequently studied by Poincare plots
and spectral analysis of heart rate variability (HRV)
(Tulppo et al. 1996, 1999; Pichon et al. 2004; Cottin et al.
2004). The first of these methods uses the standard deviation of instantaneous beat-to-beat RR interval variability
and the standard deviation of continuous long-term RR
interval variability to provide the SD1 and SD2 indices,
respectively (Tulppo et al. 1996, 1999). The second method
uses the spectral analysis of short-term HRV records to
provide two frequency bands, the high-frequency band
(HF) and the low-frequency band (LF) (Pichon et al. 2004;
Cottin et al. 2004; Yamamoto et al. 1991). In particular, the
LF provides information on the balance of sympathetic and
parasympathetic nervous system activities, while the SD1
and HF quantify the parasympathetic modulation of HR.
Therefore, as the exercise intensity becomes more severe,
there is an increase in the LF and decreases in the SD1 and
HF because of the balance between parasympathetic and
sympathetic modulation (Pichon et al. 2004; Cottin et al.
2004; Yamamoto et al. 1991).
Factors other than exercise intensity also appear to affect
autonomic regulation during dynamic exercise. Some
studies have shown that manipulations of carbohydrate
(CHO) stores can affect autonomic modulation (Sasaki
et al. 1991; Helge et al. 1996). Sasaki et al. (1991) observed
a rise in plasma norepinephrine levels during exercise at
_ 2max following a
70% of the maximal oxygen uptake VO
7-day low-CHO diet, when compared with the levels following an iso-energetic 7-day high-CHO diet. Similarly,
Helge et al. (1996) demonstrated greater plasma norepinephrine levels and HRs during exercise at *80% of
_ 2max following the ingestion of a high-fat, low-CHO diet
VO
for 7 days. They suggested that the increase in sympathetic
activity was the result of reduced CHO availability.
However, even though these manipulations have been
shown to alter sympathetic activity (Sasaki et al. 1991;
Helge et al. 1996), it is still not known whether they could
modify the HRV parameter during exercise. Additionally,
the interactive effects of CHO manipulation and exercise intensity on the HRV parameters still need to be
established.
Changes in the ventilatory pattern during exercise
may also alter the HRV responses. Studies have shown

123

Eur J Appl Physiol (2010) 108:11331140

that exercise-induced hyperventilation, especially during

severe exercise, appears to indirectly increase the highfrequency component (HF) of HRV via the mechanical
effect of breathing on the sinus node (Pichon et al. 2004;
Cottin et al. 2004, 2005, 2006, 2007). This effect may be
accentuated by the manipulation of CHO stores. Indeed,
higher ventilation volumes at any power output during an
incremental exercise test have been reported when a lowCHO diet is consumed, but these effects are particularly
evident during severe exercise (Heigenhauser et al. 1983;
Busse et al. 1991). Consequently, it could be hypothesized that the hyperventilation induced by low CHO
availability might modify the HRV responses during
exercise. The question that remains is if the interaction
between low CHO availability and hyperventilation is
evident at moderate exercise intensities or only during
severe exercise (Cottin et al. 2004, 2005, 2006). The
investigation into the effects of CHO manipulation on
HRV parameters has practical application, because athletes might make a biased decision in supposing that a
low-CHO diet increases performance due to the increase
in muscle fat oxidative capacity and the reduction in
muscle glycogen oxidation (Simi et al. 1991). Results of
this poor CHO intake could be a greater cardiovascular
stress and a consequent negative effect in performance
(Baldwin et al. 2003).
Therefore, the purpose of this study was to examine the
effects of low CHO availability on HRV and ventilatory
responses during moderate- and severe-intensity exercise.
We hypothesized that there would be an increase in
sympathetic activation due to low CHO availability,
resulting in alterations in the HRV parameters during
exercise. Additionally, we hypothesized that the decrease
in CHO availability would also yield higher ventilation
volumes during exercise, mainly at a severe intensity,
which in turn would induce changes in the HRV
parameters.

Methods
Subjects
Six healthy males between 20 and 30 years of age (age
26.5 6.7 years; body mass 78.4 7.7 kg; height 179.7
_ 2max ; 39.5 6.6 mL
4.2 cm; body fat 11.3 4.5%; VO
kg-1 min-1) were recruited to take part in this study. The test
protocol, benefits and risks were previously explained, and
written consent was obtained. The procedures used were in
agreement with the Ethics Code of the World Medical
Association (Declaration of Helsinki), and the approval of the
Ethics Committee of the University of Santa Catarina State
was obtained.

Eur J Appl Physiol (2010) 108:11331140

Preliminary test
_ 2max and the first (LT1) and second (LT2) lactate
The VO
thresholds were determined during an incremental test
on an electrically braked cycle ergometer (Ergo Fit 167,
Pirmansens, Germany). The test was carried out between
7:00 and 9:00 h after 8- to 12-h overnight fasting. The test
started with a 5-min warm-up at 50 W followed by
increases of 20 W every 3 min until volitional exhaustion.
At the end of each stage, 25 ll of blood was drawn from
the ear lobe and immediately analyzed to determine the
blood lactate concentrations (YSI 1500 Sport, Yellow
Springs Instruments, Yellow Springs, OH, USA). The LT1
and LT2 were determined by linear regression analysis
(Ribeiro et al. 1986). Briefly, the lactate-time function was
fitted with two initially unknown intercepts calculated from
every possible intersection between any points before the
LT1 and after the LT2. The intercepts that best shared the
curve in three roughly linear segments were determined for
the case in which the highest R2 and the lowest standard
error of the estimation were attained (Pires et al. 2008).
The lactate breakpoints were used to determine the
workload of the CHO depletion protocol and experimental
tests. The workload of these trials was determined by using
the lactate breakpoints method, because it has been considered as an accurate and reliable method for indentifying
metabolic transition points and determining trial intensities
with different metabolic stresses (Weltman et al. 1989).
Additionally, the protocol with 20 W/3-min increases has
been described as being most appropriate for the identification of lactate thresholds due to the blood lactate kinetics
_ 2max was assumed when at
(Stockhausen et al. 1997). VO
least one of the following criteria was attained: HR greater
than 90% of the maximal age-predicted heart rate; RER
_ 2 rate of increase smaller than
greater than 1.1; a VO
-1
150 mL min (Howley et al. 1995). The maximal workload (WMAX) was determined as the highest workload
reached with a pedal frequency between 70 and 80 rpm.
Manipulation of CHO availability
Two days (48 h) prior to each experimental test, the subjects performed a CHO depletion protocol. The CHO
depletion was obtained by subjects cycling for 90 min at
the power output required for a 50% difference between
_ 2max ). This
LT1 and LT2, named as DLT50% (*70% VO
intensity was such that the subjects could complete 90 min
of cycling. At the conclusion of this, the subjects rested for
5 min, and then a series of 1-min bouts at 125% VO2max
separated by 1-min rest periods was carried out until
volitional fatigue. This combined protocol (prolonged and
intermittent exercises) was previously validated for the
reduction of muscle glycogen content in both fast and slow-

1135

twitch fibers (Gollnick et al. 1973, 1974; Heigenhauser

et al. 1983). After the CHO depletion exercise protocol,
subjects consumed either a low- (10% CHO, 55% protein
and 35% fat) or a moderate-CHO (65% CHO, 15% protein
and 20% fat), iso-energetic diet for 48 h. It has been
observed that the consumption of a diet with a moderate or
low CHO content over a 2-day period restores or reduces
the endogenous CHO stores, respectively (Heigenhauser
et al. 1983; Grisdale et al. 1990; Glass et al. 1997).
A dietician created all the diets by using food plans
and considering the body mass and food preferences of
each subject. The subjects were given a list of food
choices that described the allowed quantity of each food
group and provided the recommended daily energy
intake. Subjects recorded all food-intake during the 48 h
after the CHO depletion exercise protocol until the
experimental session. Subjects were instructed by the
dietician in correctly recording the food intakes, and a
2-day simulation was applied for familiarization with
food intake records. Diet records were subsequently
analyzed for caloric contributions of fats, proteins,
and carbohydrates (DietWin professional software). The
subjects were asked to abstain from alcohol, caffeine,
and vigorous exercise during the 48 h diet prior to the
experimental sessions.
Experimental tests
The experimental protocol was performed in the morning after
812 h overnight fasting. The subjects completed four
experimental tests in a random order, separated by 7-day
washout periods. The four tests included exercise to exhaustion after the low and moderate CHO content diets at both
moderate (MOD-CHOcontrol and MOD-CHOlow, respectively) and severe exercise intensities (SEV-CHOcontrol
and SEV-CHOlow, respectively). The moderate-intensity
exercise was defined as a power output corresponding to
DLT50% intensity (i.e., below LT2). The severe-intensity
exercise was defined as a power output corresponding to
25% of the difference between WMAX and LT2, named as
DWL25% (above LT2). Subjects were only informed about
the workload used in each experimental test after completion of the entire study.
Measurement and analysis of gas exchange
_ 2 ; carbon dioxide
During all tests, the oxygen uptake VO
_ 2 and ventilation V_E were measured
production VCO
using a portable metabolic measurement system (Aerosport
Kb1C, Aerosport Inc., An Arbor, MI, USA). The subjects
breathed through a mask connected to a pneumotach sys_ 2 by a
tem. Samples of expired air were analyzed for VO
_
galvanic fuel cell and for VCO2 by a non-dispersive

123

1136

infrared sensor. V_E was calculated by the difference in

pressure across the pneumotach orifice plate. Prior to each
exercise test, the gas analyzers were calibrated using room
air and known gas concentrations (16% O2 and 5% CO2).
The air volume was calibrated with a 3 L syringe. The
expired respiratory gases were recorded every 20-s and
smoothed to 1-min averages. The last 1-min average for
each 5-min interval was used for the analysis.
Measurement and analysis of heart rate variability
The RR intervals were recorded by a heart rate monitor
(Polar Vantage NV, Kempele, Finland) and transferred to a
computer for further analysis (Polar Precision Performance
software). The heart rate monitor was previously validated
in a pilot study conducted in our laboratory. In order to
validate the device, measurements obtained by the Polar
Vantage NV and the AT/CODAS electrocardiogram were
carried out in real-time. Comparisons between the equipment showed no significant difference between the AT/
CODAS (526 162 ms) and the Polar Vantage NV
(527 162 ms) for a HR range from 68 to 168 bpm. A
high coefficient of determination (r2 = 0.99) and a low
standard error of estimation (1.6 ms) were found. Validity
evidence has been obtained by others (Goodie et al. 2000;
Carrol et al. 1991).
The following 60-s intervals were selected: 01, 45,
910, 1415, 1920, 2425 min and the last min before
exhaustion for the moderate exercise intensity, and 01,
45, 910 min and the last min before exhaustion for the
severe exercise intensity. For each segment, all RR
intervals were plotted against previous RR intervals, and
the Poicare plot analysis was employed. The standard
deviations of the variabilities of instantaneous RR intervals (SD1) and continuous RR intervals (SD2) were
obtained (Tulppo et al. 1996, 1999). The average HR for
each segment was also recorded.
An autoregressive model (order 20) was also applied to
estimate the power spectrum densities of the HRV
(Boardman et al. 2002; Tulppo et al. 2003). It has been
suggested that the spectral analysis method should only be
applied in stationary signals (Cottin et al. 2004, 2005). To
ensure that the signals were stationary, the data collected
during the last 5-min before exhaustion were accepted as
leveling off when the difference between the average and
the variance for two consecutive 2.5-min intervals was
lower than 5% (Casadei et al. 1996). Around 400 RR
intervals were utilized in each window. The LF and HF
were quantified by measuring the area under two frequency
bands of the spectrum: LF, from 0.04 to 0.15 Hz; HF, from
0.15 to 0.4 Hz. The spectral values were expressed as
absolute values (ms2) as well as normalized units (nu). The
ratio of LF/HF was also calculated.

123

Eur J Appl Physiol (2010) 108:11331140

Statistical analysis
The data distribution was verified by ShapiroWilks test.
As the data did not show a Gaussian normal distribution,
the Friedman test was applied to evaluate the changes over
time for the HF and LF components and V_E : When a significant difference was detected, the Wilcoxon signedranked test was carried out to identify the differences
between each 60-s segment of these variables. The
Spearman-Rank test (q) was used to analyze the relationship between the HRV and V_E : To make any future comparisons easier, all data are presented as mean SD, even
if they have not shown normal distributions. The statistical
significance was accepted as P \ 0.05.

Results
No differences were found between the prescribed and
recorded diets in terms of total energy intake and macronutrient composition. Subjects reported that they had not consumed alcohol or caffeine or performed vigorous exercise
during the diet periods. The average power outputs for the
experimental tests were 161.0 16.2 W (70.5 15.7%
_ 2max ) and 207.1 20.0 W (78.3 12.0% VO
_ 2max ) for
VO
the moderate- and severe-intensity exercise tests, respectively. The time to exhaustion was not different between the
low- and moderate-CHO experimental conditions during
moderate-intensity exercise (MOD-CHOcontrol: 57.2
24.2, MOD-CHOlow: 56.7 21.7 min; P [ 0.05), but it
was significantly reduced in the low-CHO trial during
severe-intensity exercise (SEV-CHOcontrol: 22.5 8.0,
SEV-CHOlow: 18.3 5.2 min; P \ 0.05).
During moderate-intensity exercise, the HR increases
significantly until 20-min (P \ 0.05) but no change was
detected thereafter (Table 1). The SD1 parameter
decreased significantly from 1 to 5 min and then remained
unchanged until exhaustion, for both experimental conditions (Table 1). A similar response was observed for the
SD2 parameter (Table 1). There were no significant differences between MOD-CHOcontrol and MOD-CHOlow in
the SD1, SD2 or HR values at any time point (Table 1).
Moreover, there were no significant differences between
the experimental conditions for the HF and LF when
expressed in absolute units (ms2) (Fig. 1 upper), nu or the
LF/HF ratio. There was no change in the V_E pattern via
manipulation of CHO availability (Table 1). No significant
correlation was found between V_E and SD1 or between V_E
and SD2 at any time point, except for V_E and SD1 at 5 min
(r = -0.78; P \ 0.05) and V_E and SD2 at 20 min (r =
-0.82; P \ 0.05). There was no significant correlation
between the HRV parameters expressed in the frequency
domain and V_E at any time point.

Eur J Appl Physiol (2010) 108:11331140

1137

Table 1 SD1, SD2, heart rate (HR) and ventilation V_E during moderate intensity exercise after moderate (MOD-CHOcontrol) and low (MODCHOlow) CHO content diet
Time (min)
1

10

15

20

25

exh

SD1 (ms)
MOD-CHOcontrol
MOD-CHOlow
SD2 (ms)
MOD-CHOcontrol
MOD-CHOlow

47.7 9.1a

2.5 0.7

2.3 0.7

2.3 0.6

2.2 0.6

2.1 0.7

2.1 1.0

49.1 9.6a

2.4 0.8

1.9 0.5

2.2 0.6

2.2 0.5

2.2 0.6

2.1 0.5

7.1 4.5

9.1 4.4

8.3 7.4

6.6 5.3

5.6 3.1

5.8 2.4

6.1 2.0

7.9 3.2

5.4 2.0

5.3 3.3

154 17b

105.2 23.4a
106.5 36.7

8.6 1.8
10.5 2.3

HR (bpm)
MOD-CHOcontrol
MOD-CHOlow

115 12a

140 17a

146 18a

150 18b

112 12

141 15

156 16

163 18

149 15

151 15

155 15

160 15

165 16

VE (L min-1)
MOD-CHOcontrol
MOD-CHOlow

9.5 0.7a

48.8 2.9

50.3 2.9

53.7 3.3

59.5 6.4

53.8 5.3

63.9 6.8

10.9 0.9a

50.4 1.0

53.2 2.2

52.3 2.2

49.8 1.4

52.1 3.4

68.7 13.3

Significantly different from subsequent time

Significantly different from values at the exhaustion (P \ 0.05)

During severe-intensity exercise, the SD1 and SD2

parameters were significantly reduced over time in both
experimental conditions (Table 2). However, at exhaustion in SEV-CHOlow, the SD1 parameter was significantly higher than that in SEV-CHOcontrol (3.8 3.4 vs.
2.1 0.7 ms, Table 2). The SD2 parameters at 5 min
and at exhaustion were also higher in SEV-CHOlow
than those in SEV-CHOcontrol (SEV-CHOlow: 8.9 2.6
and 5.2 3.0 ms; SEV-CHOcontrol: 5.9 1.8 and
2.6 0.3 ms, Table 2). These alterations slightly lowered the HR values at exhaustion for SEV-CHOlow than
for SEV-CHOcontrol (177 9, 182 10 bpm, respectively, Table 2). In addition, the HF and LF were significantly higher in SEV-CHOlow than in SEV-CHOcontrol
when expressed in ms2 (P \ 0.05) (Fig. 1 lower), and
the total energy spectrum was also significantly higher in
SEV-CHOlow. However, the HF and LF expressed in nu
and as the LF/HF ratio were not different between SEVCHOcontrol and SEV-CHOlow. V_E was significantly higher
in SEV-CHOlow than in SEV-CHOcontrol at 5 and 10 min
(SEV-CHOlow: 70.0 3.6 and 79.6 5.1 L min-1,
respectively; SEV-CHOcontrol: 62.5 4.4 and 74.8
6.5 L min-1, respectively, Table 2). Significant correlations were found between V_E and SD1 at 5 min
(r = 0.75; P \ 0.05), 10 min (r = 0.68; P \ 0.05) and
exhaustion (r = 0.63; P \ 0.05). The only significant
correlation between V_E and SD2 was found at 5 min
(r = 0.64; P \ 0.05). There was no any significant correlation between the HRV parameters expressed in the
frequency domain and V_E at any time point.

Fig. 1 The high-frequency (HF) and low-frequency (LF) bands at

exhaustion for moderate (upper) and severe (lower) exercise
performed after following moderate- (Control) or low- (Low)
carbohydrate diets. Values are mean and SDs

Discussion
The main finding of this study was that low CHO availability altered the HRV parameters during severe-, but not
moderate-, intensity exercise. These alterations in the HRV
parameters were accompanied by a significant increase in
V_E that was associated with the SD1 parameter.

123

1138
Table 2 SD1, SD2, heart rate
(HR) and ventilation V_E
during severe intensity exercise
after moderate (SEVCHOcontrol) and low (SEVCHOlow) CHO content diet

Eur J Appl Physiol (2010) 108:11331140

Time (min)
1

10

exh

SD1 (ms)
SEV-CHOcontrol

48.0 4.1a

1.9 0.5

1.9 0.6

2.1 0.7

SEV-CHOlow

48.5 2.9a

2.5 0.9

2.4 0.8

3.8 3.4c

SEV-CHOcontrol

103.6 22.5a

5.9 1.8

4.6 2.9

2.6 0.3

SEV-CHOlow

104.7 22.5a

8.9 2.6a,c

5.4 2.3

5.2 3.0c

118 8a

157 8a

169 6a

182 10

SD2 (ms)

Significantly different from

subsequent time

Significantly different from

values at the exhaustion

Significantly different from

values in the SEV-CHOcontrol
(P \ 0.05)

HR (bpm)
SEV-CHOcontrol
SEV-CHOlow

115 8

156 7

169 7

177 9c

7.1 0.8a
9.7 0.8a

62.5 4.4b
70.0 3.6b,c

74.8 6.5
79.6 5.1c

84.8 8.0
85.4 8.1

VE (L min-1)
SEV-CHOcontrol
SEV-CHOlow

Low CHO availability did not change the autonomic

modulation during moderate-intensity exercise. In contrast,
during severe-intensity exercise, the low CHO availability
altered the autonomic modulation, yielding higher values
of SD1 and HF during exercise. These findings suggest that
the parasympathetic control of HR is increased during
severe-intensity exercise in response to a low-CHO diet.
Contrary to our findings, parasympathetic control has been
shown not to be activated during severe-intensity exercise,
and vagal withdrawal is typically reported (Breuer et al.
1993; Tulppo et al. 1996, 1999). However, other counterregulatory alterations caused by low-CHO availability may
have affected the beat-to-beat oscillation during the severeintensity exercise. For example, several authors have
suggested that the greater V_E induced by high-intensity
exercise may lead to alterations in sinus node depolarization, thereby modifying the cardiac rhythm (Cottin et al.
2004, 2006; Blain et al. 2005). Interestingly, the increased
SD1 and HF values during severe-intensity exercise after a
low-CHO diet were accompanied by a subtle and significant reduction in HR at exhaustion (Table 2). The lowCHO diet had no effect on the HR during the preceding
time points. However, there was a significant increase in
V_E during the 5 and 10-min of the exercise. These results
suggest that the respiratory mechanics may affect the sinus
node activity, modifying the beat-to-beat oscillations over
time but without modification of the mean HR.
Several studies have demonstrated that the increase in
the HF during exercise performed above the first ventilatory threshold could be explained by the mechanical effect
of the breathing muscles on the sinus node (Cottin et al.
2004, 2006, 2007). Cottin et al. (2005), studying elite
trotting horses, demonstrated a higher HF energy (expressed in absolute units) during severe-intensity exercise, and
it was influenced by greater ventilation values. In addition,
Pichon et al. (2004) found a positive relationship between

123

the HF parameter and ventilation volume at intensities

_ 2max in humans. Even though
between 60 and 80% of VO
there were no associations between the HF parameters and
V_E in the present study, strong correlations were found
between V_E and the SD1 index, suggesting that ventilation
can affect the instantaneous RR oscillations. Indeed, we
showed that the low-CHO diet induced an increase in the
V_E ; HF and SD1 parameters only during severe-intensity
exercise, in which the ventilation volume has a greater
influence on the HRV (Cottin et al. 2004, 2006, 2007).
Therefore, the increase of the SD1 and HF in the SEVCHOlow condition might have been the result of a higher
ventilation volume induced by low CHO stores, as these
parameters are associated with instantaneous RR oscillations. Based on this supposition, it seems reasonable to
assume that the mechanical effect of breathing on the sinus
node, instead of vagal reactivation, might be causing the
alteration in the HF during severe-intensity exercise.
Similar to the findings for HF, the absolute LF values
also increased during severe-intensity exercise in response
to low CHO availability when compared to the control
condition. The amplitude of the LF, expressed in absolute
units, reflects the magnitude of sympathetic activity
(Yamamoto et al. 1991; Nakamura et al. 1993; Pagani et al.
1997; Tulppo et al. 2005). The higher LF observed in our
results could suggest a slightly altered sympathetic outflow
following the low-CHO diet. These results corroborate the
findings of Helge et al. (1996), who showed higher plasma
norepinephrine concentrations and HRs during severe_ 2max ) after
intensity exercise (*81% of pre-training VO
7 weeks of endurance training while using a low-CHO diet.
Similarly, Havemann et al. (2006) demonstrated higher
HRs for a given power output and compromised performances during high-intensity 1-km cycling sprints, but not
moderate-intensity 4-km sprints, in response to a lowCHO, high-fat dietary intervention. The low-CHO, high-fat

Eur J Appl Physiol (2010) 108:11331140

diet in this study was associated with changes in sympathetic

activation, namely an increase in the LF parameter. Together,
these results provide evidence for an alteration in the sympathetic outflow in response to dietary manipulation, supporting the notion that low CHO availability can influence
sympathetic activity during severe-intensity exercise.
Typically, a reduction in CHO intake is accompanied by
an increase in fat or protein intake (Helge et al. 1996;
Havemann et al. 2006; Foo et al. 2009). In the present study,
we used a low-CHO diet with high protein content in order to
replace the energy lost by the lower CHO content. Previous
studies have shown that a chronic high-protein diet can affect
the cardiovascular system by inducing more considerable
aortic atherosclerosis, impairing the ability to generate new
vessels in response to tissue ischemia, and reducing the
vascular regenerative capacity (Foo et al. 2009). Even
though the diets here were manipulated for a short term, it is
possible that the high-protein diet could have caused some
acute impact on heart autonomic modulation in the present
study. Future studies should be designed to establish the
isolated effects of low-CHO and high-protein diets on
autonomic modulation during exercise.
Some methodological aspects need to be addressed here.
HRV is an indirect measure of parasympathetic and
sympathetic activities in the heart, and therefore, neither
cardiac sympathetic nor cardiac parasympathetic nerve
activity was directly measured in the present study. As
such, the present data provided only an indirect assessment
of cardiac autonomic activity. Another important methodological consideration is the frequency domain analysis of
the HRV signal during the high exercise intensity. Frequency domain analysis may be inadequate for the characterization of HRV dynamics during non-stationary data
sets, as may be found during high-intensity exercise
(Kanters et al. 1996). Although the time effect on the RR
intervals cannot be discarded, we measured the HRV in
data sets in which the difference between the average and
the variance of two consecutive 2.5-min RR intervals was
lower than 5%. Therefore, a stationary characteristic may
be suggested in the temporal series data.
In summary, a low CHO availability influences the HRV
parameters during severe-intensity exercise, but it does not
have any effect during moderate-intensity exercise. In
addition, the changes in V_E during severe exercise that
were caused by low CHO availability seem to be associated
with autonomic modulation, as observed by alterations in
the SD1, SD2 and HF parameters of the HRV. Because a
poor nutritional choice with a low CHO content could be
used by athletes hoping to improve their performance by
increasing their muscle fat oxidative capacity and reducing
their muscle glycogen oxidation, the results presented here
have practical implications as the HRV, and consequently
the HR, can be affected by a low CHO availability.

1139
Acknowledgments The authors are grateful to Professor Julia
Goedecke for guidelines and suggestions, and to Michele Souza,
Fernando Adami and Tony Fernandes for technical assistance. The
Flavio O Pires is grateful for CAPES for his doctoral scholarship.

References
Baldwin J, Snow RJ, Gibala MJ, Garnham A, Howarth K, Febbraio
MA (2003) Glycogen availability does not affect the TCA cycle
or TAN pools during prolonged, fatiguing exercise. J Appl
Physiol 94:21812187
Blain G, Meste O, Bermon S (2005) Influences of breathing patterns
on respiratory sinus arrhythmia in humans during exercise. Am J
Physiol Heart Circ Physiol 288:H887H895
Boardman A, Schlindwein FS, Rocha AP, Leite A (2002) A study on
the optimum order of autoregressive models for heart rate
variability. Physiol Meas 23:325336
Breuer HW, Skyschally A, Schulz R, Martin C, Wehr M, Heusch G
(1993) Heart rate variability and circulating catecholamine
concentrations during steady state exercise in healthy volunteers.
Br Heart J 70:144149
Busse MW, Maassen N, Konrad H (1991) Relation between plasma
K? and ventilation during incremental exercise after glycogen
depletion and repletion in man. J Physiol 443:469476
Carrol T, Godsen R, Tangeman C (1991) The polar vantage XL heart
rate monitor: an analysis of its internal consistency and computer
interface. Med Sci Sports Exerc suppl 23(4):14
Casadei B, Moon J, Johnston J, Caiazza A, Sleight P (1996) Is
respiratory sinus arrhythmia a good index of cardiac vagal tone
in exercise? J Appl Physiol 81:556564
Cottin F, Medigue C, Lepretre PM, Papelier Y, Koralsztein JP, Billat V
(2004) Heart rate variability during exercise performed below and
above ventilatory threshold. Med Sci Sports Exerc 36:594600
Cottin F, Medigue C, Lopes P, Petit E, Papelier Y, Billat VL (2005) Effect
of exercise intensity and repetition on heart rate variability during
training in elite trotting horse. Int J Sports Med 26:859867
Cottin F, Lepretre PM, Lopes P, Papelier Y, Medigue C, Billat V
(2006) Assessment of ventilatory thresholds from heart rate
variability in well-trained subjects during cycling. Int J Sports
Med 27:959967
Cottin F, Medigue C, Lopes P, Lepretre PM, Heubert R, Billat V
(2007) Ventilatory thresholds assessment from heart rate variability during an incremental exhaustive running test. Int J Sports
Med 28:287294
Foo SY, Heller ER, Wykrzykowska J, Sullivan CJ, Manning-Tobin
JJ, Moore KJ et al (2009) Vascular effects of a low-carbohydrate
high-protein diet. Proc Natl Acad Sci USA 106(36):15418
15423
Glass C, Knowlton RG, Sanjabi PB, Sullivan JJ (1997) The effect of
exercise induced glycogen depletion on the lactate, ventilatory
and electromyographic thresholds. J Sports Med Phys Fitness
37:3240
Gollnick PD, Armstrong RB, Sembrowich WL, Shepherd RE, Saltin
B (1973) Glycogen depletion pattern in human skeletal muscle
fibers after heavy exercise. J Appl Physiol 34:615618
Gollnick PD, Piehl K, Saltin B (1974) Selective glycogen depletion
pattern in human muscle fibres after exercise of varying intensity
and at varying pedalling rates. J Physiol 241:4557
Goodie JL, Larkin KT, Schauss S (2000) Validation of Polar heart
rate monitor for assessing heart rate during physical and mental
stress. J Psychophysiol 14:159164
Grisdale RK, Jacobs I, Cafarelli E (1990) Relative effects of glycogen
depletion and previous exercise on muscle force and endurance
capacity. J Appl Physiol 69:12761282

123

1140
Havemann L, West SJ, Goedecke JH, Macdonald IA, St Clair Gibson
A, Noakes TD et al (2006) Fat adaptation followed by
carbohydrate loading compromises high-intensity sprint performance. J Appl Physiol 100:194202
Heigenhauser GJF, Sutton JR, Jones NL (1983) Effect of glycogen
depletion on the ventilatory response to exercise. J Appl Physiol
54:470474
Helge JW, Richter EA, Kiens B (1996) Interaction of training and diet
on metabolism and endurance during exercise in man. J Physiol
492(Pt 1):293306
Howley ET, Bassett DR Jr, Welch HG (1995) Criteria for maximal
oxygen uptake: review and commentary. Med Sci Sports Exerc
27:12921301
Kanters JK, Hjgaard MV, Agner E, Holstein-Rathlou NH (1996)
Short- and long-term variations in non-linear dynamics of heart
rate variability. Cardiovasc Res 31:400409
Nakamura Y, Yamamoto Y, Muraoka I (1993) Autonomic control of
heart rate during physical exercise and fractal dimension of heart
rate variability. J Appl Physiol 74:875881
Pagani M, Montano N, Porta A, Malliani A, Abboud FM, Birkett C
et al (1997) Relationship between spectral components of
cardiovascular variabilities and direct measures of muscle
sympathetic nerve activity in humans. Circulation 95:14411448
Pichon AP, de Bisschop C, Roulaud M, Denjean A, Papelier Y (2004)
Spectral analysis of heart rate variability during exercise in
trained subjects. Med Sci Sports Exerc 36:17021708
Pires FO, Lima-Silva AE, Oliveira EM, Rumenig-Souza E, Kiss MA
(2008) Ventilation behavior in trained and untrained men during
incremental test: evidence of one metabolic transition point.
J Sports Sci Med 7:335343
Ribeiro JP, Hughes V, Fielding RA, Holden W, Evans W, Knuttgen
HG (1986) Metabolic and ventilatory responses to steady state

123

Eur J Appl Physiol (2010) 108:11331140

exercise relative to lactate thresholds. Eur J Appl Physiol Occup
Physiol 55:215221
Sasaki H, Hotta N, Ishiko T (1991) Comparison of sympatho-adrenal
activity during endurance exercise performed under high- and
low-carbohydrate diet conditions. J Sports Med Phys Fitness
31:407412
Simi B, Sempore B, Mayet M-H, Favier RJ (1991) Additive effects of
training and high-fat diet on energy metabolism during exercise.
J Appl Physiol 71:197203
Stockhausen W, Grathwohl D, Burklin C, Spranz P, Keul J (1997) Stage
duration and increase of work load in incremental testing on a cycle
ergometer. Eur J Appl Physiol Occup Physiol 76:295301
Tulppo MP, Makikallio TH, Takala TE, Seppanen T, Huikuri HV
(1996) Quantitative beat-to-beat analysis of heart rate dynamics
during exercise. Am J Physiol 271(1 Pt 2):H244H252
Tulppo MP, Makikallio TH, Laukkanen RT, Huikuri HV (1999)
Differences in autonomic modulation of heart rate during arm
and leg exercise. Clin Physiol 19:294299
Tulppo MP, Hautala AJ, Makikallio TH, Laukkanen RT, Nissila S,
Hughson RL et al (2003) Effects of aerobic training on heart rate
dynamics in sedentary subjects. J Appl Physiol 95:364372
Tulppo MP, Huikuri HV, Tutungi E, Kimmerly DS, Gelb AW,
Hughson RL et al (2005) Feedback effects of circulating
norepinephrine on sympathetic outflow in healthy subjects. Am
J Physiol Heart Circ Physiol 288:H710H715
Weltman A, Weltman J, Rutt R, Seip R, Levine S, Snead D et al
(1989) Percentages of maximal heart rate, heart rate reserve, and
VO2peak for determining endurance training intensity in sedentary women. Int J Sports Med 10:212216
Yamamoto Y, Hughson RL, Peterson JC (1991) Autonomic control of
heart rate during exercise studied by heart rate variability
spectral analysis. J Appl Physiol 71:11361142