Eur J Appl Physiol (2007) 101:359368
DOI 10.1007/s00421-007-0506-8
ORIGINAL ARTICLE
Effects of intermittent hypoxic training on cycling performance
in well-trained athletes
Belle Roels David J. Bentley Olivier Coste
Jacques Mercier Gregoire P. Millet
Accepted: 9 June 2007 / Published online: 17 July 2007

Springer-Verlag 2007
Abstract This study aimed to investigate the effects of a
short-term period of intermittent hypoxic training (IHT) on
cycling performance in athletes. Nineteen participants were
randomly assigned to two groups: normoxic (NT, n = 9)
and intermittent hypoxic training group (IHT, n = 10). A
3-week training program (5 1 h1 h 30 min per week)
was completed. Training sessions were performed in
normoxia (~30 m) or hypoxia (simulated altitude of
3,000 m) for NT and IHT group, respectively. Each subject
performed before (W0) and after (W4) the training
B. Roels
O. Coste
UPRES EA 3759Multidisciplinary Approach of Doping,
700 avenue Pic St Loup, 34090 Montpellier, France
B. Roels
Centre of Sport Medicine and Human Performance,
Brunel University, School of Sport and Education,
London UB8 3PH, UK
D. J. Bentley
Department of Human and Health Science,
University of Westminster, London, UK
D. J. Bentley
Health and Exercise Science, School of Medical Sciences,
University of New South Wales, Sydney, Australia
O. Coste
Direction Regionale et Departementale de la Jeunesse et des
Sports, 190 avenue du Pe`re Soulas, 34094 Montpellier, France
J. Mercier
Laboratoire de physiologie des Interactions EA 701,
Institut de Biologie, Bvd Henri IV, 34060 Montpellier, France
G. P. Millet (&)
ASPIRE, Academy for Sports Excellence,
PO Box 2287, Doha, Qatar
e-mail: gregoire.millet@aspire.qa
program, three cycling tests including an incremental test
to exhaustion in normoxia and hypoxia for determination
of maximal aerobic power VO _ 2 max and peak power out-
put (PPO) as well as a 10-min cycle time trial in normoxia
(TT) to measure the average power output (Paver). No
_ 2 max was observed between the
significant difference in VO
two training groups before or after the training period.
When measured in normoxia, the PPO significantly in-
creased (P < 0.05) by 7.2 and 6.6% in NT and IHT groups,
respectively. However, only the IHT group significantly
improved (11.3%; P < 0.05) PPO when measured in hy-
poxia. The NT group improved (P < 0.05) Paver in TT by
8.1%, whereas IHT group did not show any significant
difference. Intermittent training performed in hypoxia was
less efficient for improving endurance performance at sea
level than similar training performed in normoxia. How-
ever, IHT has the potential to assist athletes in preparation
for competition at altitude.
Keywords Aerobic
Altitude
Endurance
Power

Adaptation
Introduction
Over the last few decades many athletes and coaches have
used altitude training in various forms to help improve
performance at altitude and/or at sea level. The traditional
approach to altitude training was for athletes to live and
train at moderate altitude. The effects of this form of
stimulus on endurance performance have been researched
extensively (Fulco et al. 2000; Levine 2002; Wilber 2001).
A recent approach has been for athletes to live and sleep at
altitude and train near sea level, the so-called live high-
train low (LHTL) method or the opposite live low-train
123
360
high (LLTH) method, to live and sleep at sea level and
train at altitude (Wilber 2001). Because the geography of
many countries does not allow LHTL or LLTH, other
strategies have been developed for athletes, such as being
briefly exposed to hypoxia. Intermittent hypoxic exposure
with (IHE) or without (IHT) exercise training is based on
the assumption that brief exposure to hypoxia (minutes to
hours) is sufficient to stimulate EPO release, and ultimately
increase red blood cell (RBC) concentration (Rusko 1996;
Wilber 2001) and to induce peripheral modifications in
skeletal muscle that in turn might increase performance
(Dufour et al. 2006; Roels et al. 2007). Rodriguez et al.
(1999) found that hypoxic exposure (4,0005,000 m) for
only 90 min, three times a week over 3 weeks, signifi-
cantly increased hematocrit (Hct), RBC, reticulocytes
(RET) and haemoglobin (Hb) concentration. However, no
control group (training in normoxia) was included. Dufour
et al. (2006) found that a 6-week IHT with two weekly
high-intensity (second ventilatory threshold; VT2) and
moderate-duration (2440 min) training sessions, enhanced
endurance performance of competitive distance runners.
The IHT group significantly improved VO _ 2 max (+5%) both
in normoxic and hypoxic conditions, without any changes
in blood O2-carrying capacity. Also, the time to exhaustion
increased for the IHT group only (+35%), without signifi-
cant changes in the VO _ 2 kinetics. This study suggested that
IHT induces peripheral muscle adaptations, which result in
an enhanced endurance performance. Aside from this
investigation data regarding the effects of IHE or IHT on
haematological and muscular variables and endurance
performance are inconclusive (Dufour et al. 2006; Geiser at
al. 2001; Rodriguez et al. 1999, 2000, 2003; Roels et al.
2005, 2007; Terrados et al. 1988; Truijens et al. 2003).
Katayama et al. (2003) investigated the effect of a 3-
week IHE program of 90 min a day, three times a week at
simulated hypobaric hypoxia of 4,500 m in trained runners
and found that the sea-level 3,000 m running time and the
time to exhaustion during a maximal exercise test was
significantly enhanced, while there was no change in the
resting haematological parameters. Therefore, IHE could
improve endurance performance at sea level in trained
runners. In addition, Hendriksen and Meeuwsen (2003)
observed an improvement in maximal power output and
mean and peak anaerobic power at sea level 9 days after
the end of a 10-day intermittent hypoxic training period at
a simulated altitude of 2,500 m (105 min per day). No
significant differences were observed after the end of a
similar training performed at sea level. In contrast, 5 weeks
of intermittent hypoxic training (simulated altitude of
2,500 m) in a swimming flume improved sea level per-
formance (100- and 400-m freestyle time trials) and
_ 2 max : However, no additional effect of a hypoxia
VO
stimulus was reported (Truijens et al. 2003). Roels et al.
123
Eur J Appl Physiol (2007) 101:359368
(2005) also showed that 4 weeks of interval training under
hypoxia conditions, i.e. simulated altitude of 3,000 m for
approximately 114 min per week, did not induce a greater
increase in performance compared to similar normoxic
training; moreover, no changes in haematological variables
were observed.
Zoll et al. (2006) found that specific muscular transcript
level adaptations in the vastus lateralis participated in
enhancing endurance performance after a 6-week IHT with
two sessions performed at VT2 per week, in long-distance
runners. Indeed, combining a hypoxic stimulus with
endurance training resulted in transcriptional adaptations of
the vastus lateralis muscle in athletes and in turn a sig-
nificant increase in time to exhaustion and VO _ 2 max : The
same research group (Ponsot et al. 2006) also found that
this 6-week IHT of 2 sessions at VT2 in well-trained ath-
letes, qualitatively improved the mitochondrial function by
increasing the respiratory control by creatine, providing a
tighter integration between ATP demand and supply.
Therefore, they suggested that adaptations of mitochondrial
function accompany the increase in endurance performance
after IHT. Geiser et al. (2001) investigated if IHT induces
greater improvements than the same intensity training in
normoxia and also if there was a difference in the effect of
high- versus moderate-intensity training on aerobic per-
formance variables. Previously untrained subjects were
assigned to four groups, Norm-high, Norm-low, Hyp-high
and Hyp-low. The training consisted of 30 min a day of
cycling, 5 days per week for 6 weeks at an intensity of
80% and 67% of VO _ 2 max for the high and low intensity
groups, respectively, and this under normoxic (600 m) or
hypoxic (3850 m) conditions for the normoxic or hypoxic
groups, respectively. They found no significant differences
in performance enhancement between IHT and the similar
normoxic protocol. However, IHT increased the endurance
performance to a greater extent when measured under
hypoxic conditions. In addition, the training intensity had
no effect on the gain of VO _ 2 max : However, the Hyp-high
group was the most effective group to increase the muscle
oxidative capacity.
The effects of high intensity interval training performed
in hypoxia on endurance performance in hypoxia in trained
athletes has not yet been clearly established and could be a
factor in the adaptive response to simulated altitude train-
ing (Geiser et al. 2001).
Collectively data regarding the effects of IHT on
endurance performance measured in hypoxia and normoxia
are minimal and inconclusive in trained athletes. Therefore,
the present study reports on the effects of a 3-week inter-
mittent hypoxic training, including two high intensity
interval sessions and three continuous training sessions per
week, on sea level and hypoxic cycling performance. We
hypothesized that IHT would increase cycling performance
Eur J Appl Physiol (2007) 101:359368
both at sea level and in hypoxia more than a similar
normoxic training.
Methods
Subjects
Nineteen healthy male endurance-trained athletes (cyclists
and triathletes) participated in the study, which was ap-
proved by the institutional ethics committee (Nmes,
France). All subjects gave written informed consent and
were sea-level residents. Each subject was familiarized
with the testing protocols and equipment used in the
experiment. None of the subjects had a history of recent
travel to altitude, i.e. 3 weeks prior to training. Subjects
were initially randomized into two groups: an intermittent
hypoxic training group (IHT; n = 10), and a normoxic
training group (NT, n = 9). Groups were of comparable
training status in terms of average power output and
average oxygen consumption measured during a laboratory
time-trial (TT) prior the experimental period. Subject
characteristics for each group are presented in Table 1. One
athlete of the NT group did not complete the study due to
illness.
Experimental design
The endurance-training program consisted of two interval
and three continuous training sessions per week for
3 weeks. The continuous sessions consisted of 60 min at
60% of VO _ 2 max : The interval-training required subjects to
perform two sets of three repetitions of 2 min duration at
an intensity of 100% of PPO (measured in either hypoxia
or normoxia). Two minute of rest was allowed between
each repetition with 6 min rest between each set. Each
training session began and ended with 15 min warm-up and
15 min cool-down. Training sessions were performed on
subjects own bicycle, fixed on an electromagnetic resistant
home-trainer (Elite Travel, Milan, Italy). Training sessions
Table 1 Subjects characteristics
W0 W4
NT IHT NT IHT
Age (years) 24.2 0.4 24.4 0.3
Height (cm) 181.6 0.7 180.1 0.5
Weight (kg) 71.3 0.9 73.2 0.8 72.0 0.8 72.6 0.8
Mean SE; W0: pre; W4: post; NT: normoxic training group
(n = 8); IHT: intermittent hypoxic training group (n = 10)
361
were performed in normoxia (PIO2 of 160 mmHg) or in a
hypoxic (PIO2 of 100 mmHg, simulated altitude of
~3000 m) environment for NT and IHT group, respec-
tively. Average duration of the hypoxic stimulus per week
during the training period was 382 min. No supplementary
interval training outside the two supervised interval-train-
ing sessions was performed. Each training session was
meticulously controlled by the same researcher. Table 2
presents the workloads of each group averaged over
3 weeks for the interval training and continuous workload
sessions. No significant differences in (relative) training
workloads during the training sessions were observed be-
tween the groups.
Before (W0) and after (W4) the training program, a
medical examination and determination of physical char-
acteristics was completed (Table 1). Subjects also per-
formed an incremental exercise test to exhaustion, and a
laboratory TT. The incremental test was performed in both
normoxic and hypoxic conditions. In contrast, TT was
performed only in normoxia. Blood samples at rest were
obtained at W0 and W4, and repeated at the end of each
training week (W1W3).
Iron supplementation (Oligo-essentiels, Laboratoires
Richelet, Paris, France) of 1 ml per day (i.e. 5 mg per day)
was initiated in both groups 2 weeks before the beginning
of the experiment and continued during the training to
prevent low iron state which could limit erythropoiesis. A
physician was in attendance at all times and was respon-
sible for the safety of the subjects during the study.
The weekly training regime, outside the training pro-
gram was controlled and documented.
Training outside experimental design
The training completed outside the experimental design
was recorded daily using a computerized training diary
during the 3 weeks of training period. Briefly, the type of
activity, i.e. cycling, swimming, running, etc., and the
intensity was recorded with this training diary. The training
intensity was divided into five intensity levels (Mujika
Table 2 The averaged percentage of target power output (target PO),
i.e. 100% PPO for interval sessions and 60% of VO2max and the
averaged absolute power output (Wabsolute; W) for continuous training
sessions, of normoxic training (NT, n = 8) and intermittent hypoxic
training (IHT; n = 10) groups
Continuous Interval
% of Wabsolute % of Wabsolute
target PO target PO
NT 91.7 0.6 185.0 7.5 91.0 0.7 310.9 4.5
IHT 91.8 1.6 130.0 0.3 91.8 1.0 259.4 0.3
Mean SE
123
362
et al. 1996). All training sessions outside the protocol were
individually timed and each exercise categorized according
to the five intensity levels. The performed training duration
was multiplied by its corresponding multiplying factor, i.e.
2, 4, 6, 10 and 16, respectively, and the sum was then
divided by the overall duration of the session to calculate
the average intensity of each training session. The recorded
parameters were the number and average intensity of the
sessions, calculated according to Mujika et al. (1996) as it
was not sensible to compare hourly volume as the subjects
performed exercise training in different modes (swimming,
running). Table 3 presents the average intensity of addi-
tional training outside the experimental design of each
training week for both groups. No significant differences in
(relative) training workloads outside the training sessions
were observed between the groups.
Environmental stimulus
A simulated altitude of 3,000 m was used because this
level of hypoxia was sufficient for erythropoietic stimula-
tion, and still well tolerated by subjects (Villa et al. 2005).
Hypoxic gas mixture was delivered continuously by a
system, which enriches the inspired air by modifying
nitrogen content (Altitrainer 200, SMTEC, Geneva,
Switzerland). Briefly, O2 content of the mixture is contin-
uously displayed and can be expressed either by equivalent
altitude (~3,000 m) or oxygen partial pressure
(PIO2~100 mmHg), taking into account barometric pres-
sure. Users inhale the mixture contained in the tank
through a Hans Rudolph two-way respiratory valve.
Respiratory mask could at all times be removed and so
subjects found themselves immediately in normoxic con-
ditions. A breath-by-breath analyzer can be easily attached
to the system in order to measure respiratory exchange.
This device has been used previously by our laboratory and
further details can be found in Roels et al. (2005).
Performance tests
An incremental test to exhaustion was performed in
normoxic and hypoxic conditions to determine maximal
Table 3 The average intensity of additional training outside the
experimental design of each training week for normoxic training (NT;
n = 8) and intermittent hypoxic training (IHT; n = 10) groups
W1 W2 W3
NT 3.1 0.05 3.1 0.06 3.2 0.1
IHT 2.9 0.09 3.0 0.05 3.0 0.07
Mean SE; W1: first week of training protocol; W2: second week of
training protocol; W3: third week of training protocol
123
Eur J Appl Physiol (2007) 101:359368
oxygen uptake VO _ 2 max ; ml kg1 min1 ; peak power
output (PPO; W), maximal ventilation VE _ max ; 1 min1 ;
and maximal heart rate (HRmax; bpm). Test began at an
initial power output of 60 W for 3 min and then the
workload was increased by 30 W every minute until
exhaustion. Exhaustion coincided with heart rate (HR)
approaching an age predicted maximum value (220age); a
plateau in VO _ 2 despite an increase in mechanical power
output; and an RER > 1.1. Breath-by-breath data were re-
duced to 30-s averages and VO _ 2 max was determined as the
highest 30-s VO _ 2 average. PPO was defined as the highest
mechanical power maintained for 1 min.
A cycle TT was performed to determine the highest
power output over a period of 10 min in normoxia. The TT
was performed to assess the influence of training on
endurance performance in normoxia by determining the
mechanical power output (averaged power over the 10 min
TT: Paver; W), percentage of maximal oxygen consumption
%VO _ 2 max ; i.e. the highest 30 s averaged VO _ 2 measured
during the TT expressed as a percentage of the VO _ 2 max
measured during the incremental test to exhaustion in
normoxia, peak ventilation VE _ peak ; 1 min1 ; and peak
heart rate (HRpeak, bpm). Subjects were informed of the
elapsed time but received no feedback on power output or
on performance.
In this 10-min cycle, TT is considered to be long enough
to obtain an aerobic contribution for energy provision, can
be easily administered and has been used in other studies
(Roels et al. 2005).
All the tests were performed on a bicycle equipped with
a SRM road professional powermeter (Schoberer Rad
Messtechnik, Julich, Welldorf, Germany). Saddle height on
cycle ergometer measured during the first test was kept
identical for the remaining tests. Power output and pedaling
cadence were recorded with an acquisition frequency of
1 s. Averaged values of 30-s were stored. Calibration of the
powermeter occurred as for the manufacturers recommen-
dations.
Physiological measurements
Pre (W0) and post training (W4), respiratory exchange
measurements were made using a K4b2 (Cosmed, Rome,
Italy). K4b2 setup was changed in order to calculate ven-
tilatory variables with the accurate FIO2 rather than with
the default version. Aforementioned physiological vari-
ables were measured, breath-by-breath, and averaged every
30 s. Before each test, the system was calibrated using
ambient air, whose partial O2 composition was assumed to
be 20.93% and a gas of known CO2 (5%) and O2 (16%)
concentration. Calibration of turbine flowmeter of the K4b2
was performed with a 3-l syringe (Quinton Instruments,
Seattle, WA). During the different tests and training ses-
Eur J Appl Physiol (2007) 101:359368 363

sions, HR was constantly recorded by means of a HR Results

monitor (S810, Polar, Kempele, Finland).
Haematological variables
Blood samples at rest were obtained by venapuncture by a
physician at W0 and W4. Samples were analyzed in the
2 hours following collection using a blood analyzer (Pentra
120 Retic, abx, Montpellier, France) for RBC (106 mm3),
Hb (g dl1), Hct (%), mean cell volume (MCV; lm3),
averaged globular content of hemoglobin (AGCH; pg),
averaged corpuscular concentration of haemoglobin
(ACCH; g dl1), red blood disc (RBD; 103 mm3), reticu-
locytes (RET; % and 106 mm3) and averaged reticulocytar
volume (ARV, lm3).
Statistical analysis
All values are reported as mean standard error (SE). The
effects of the two training methods (IHT vs. NT) on
measured variables was, after analysis of normality and
homogeneity of variance of the tested samples, analyzed
using a two-way (training group time) variance (ANO-
VA) with repeated measures on the second factor. Signif-
icant effects were subsequently analyzed using Student
Newman-Keuls post-hoc test. The ratio of true variance
over error variance (F) was calculated. Pearson correlations
were also completed to indicate significant correlations
between the performance parameters. All analyzes were
completed using SigmaStat 2.3 (Jandel Corporation, San
Rafael, CA) and statistical significance was accepted at
P < 0.05.
Incremental test to exhaustion
There were no significant differences in the initial PPO and
_ 2 max measured in normoxia or hypoxia between the two
VO
experimental groups. Physiological variables obtained
from the incremental test performed in normoxic and
hypoxic conditions are presented in Table 4. PPO signifi-
cantly increased after training (P < 0.001; F = 15.0) for
both groups when measured in normoxic conditions but no
differences were found between groups. PPO values were
341.7 3.5 and 339.0 4.9 W at W0 and 366.3 3.2 and
361.5 4.4 W at W4 for NT and IHT group respectively.
In hypoxic conditions, only the IHT group increased
(P < 0.05; F = 8.09) PPO (11.3%) from 282.5 3.4 to
314.5 3.7 W, respectively, at W0 and W4. No significant
differences were found in the NT group: 301.1 2.5 to
313.8 1.9 W at W0 and W4, respectively.
Cycle time-trial
At the start of the training period there were no significant
differences between groups in terms of average power
output (Paver) during TT. Only the NT group improved
(P < 0.05; F = 7.05) Paver by 8.1%. Paver values were
259.1 5.0 W at W0, and 280.1 5.8 W at W4, whereas
IHT group did not show any significant difference
(255.3 4.1 to 257.5 4.0 W at W0 and W4, respec-
tively). There were no significant differences for the other
physiological variables measured during the cycle TT
between each training group (Table 5).

Table 4 Incremental test to exhaustion under normoxic and hypoxic condition

Normoxia Hypoxia
NT (n = 8) IHT (n = 10) NT (n = 8) IHT (n = 10)
W0 W4 W0 W4 W0 W4 W0 W4

PPO 341.7 3.48 366.3 3.2** 339.0 0.48 361.5 4.41** 301.1 2.45 313.8 1.89 282.5 3.41 314.5 3.67 *
(W)
_ 2 max
VO 58.1 0.83 61.0 1.15 58.5 0.7 58.3 0.59 47.6 0.81 48.8 1.08 48.3 0.94 48.6 0.64
(ml kg1 min1)
HRmax 190.1 1.1 188.3 1.54 189.7 1.1 189.4 0.97 187.6 1.44 187.4 1.15 188.2 0.96 188.9 1.06
(bpm)
_ max
VE 159.9 2.6 174.3 2.5 154.2 1.4 155.0 1.74 151.2 3.26 167.5 3.38* 140.6 2.27 142.5 1.91

(l min1)

RPE 17.9 0.16 17.9 0.14 16.4 0.15 16.9 0.16 17.7 0.26 18.0 0.21 17.5 0.16 17.8 0.15
Mean SE; W0: pre; W4: post; NT: normoxic training group; IHT: intermittent hypoxic training group; PPO: peak power output; VO _ 2 max :
highest value of the oxygen consumption averaged over 30 s; HRmax: highest value of heart rate averaged over 30 s; VE _ max : highest value of
ventilation averaged over 30 s; RPE: rate of perceived exertion; *: P < 0.05, **: P < 0.001 for the differences within a group versus W0; :
P < 0.05 for the differences between groups at a matched time point

123
364
Table 5 Cycle time trial
NT (n = 8)
W0
Paver (W) 259.1 5.0
_ 2 max
%VO 99.1 0.9
HRpeak (bpm) 182.9 1.5
_ peak (l min1)
VE 145.2 2.5
RPE 17.7 0.1
Mean SE; W0: pre; W4: post; NT: normoxic training group; IHT: intermittent hypoxic training group; Paver: Average 10-min power output;
%VO_ 2 max : the maximal oxygen consumption averaged over 30 s measured during the time trial expressed as a percentage of the maximal
oxygen consumption averaged over 30 s measured during the incremental test to exhaustion under normoxia; HRpeak: highest value of heart rate
_ peak : highest value of ventilation averaged over 30 s; RPE: rate of perceived exertion; *: P < 0.05 for the differences
averaged over 30 s; VE
within a group versus W0
Haematological variables
Haematological variables measured during the training
period are represented in Table 6. No significant differ-
ences were found between the two training groups, before,
during or after the experimental period.
Discussion
The present study showed that two interval and three
continuous training sessions per week performed over a
period of 3 weeks significantly improved PPO measured in
normoxia and cycle time-trial performance. However,
these improvements were not further enhanced by per-
forming the same training in hypoxic conditions. Specifi-
cally the findings of the present study demonstrate that
training in hypoxia and normoxia increased PPO at sea
level (normoxia) to the same extent. Of practical signifi-
cance, exercise training in hypoxia increased PPO in
hypoxic conditions. In contrast to these findings, the
improvement in endurance performance following the
different exercise training interventions was not associated
with any changes in VO _ 2 max or in haematological vari-
ables.
Effects of training in hypoxia or normoxia
on endurance performance
In the present study PPO and cycle time-trial performance
improved after 3 weeks of training in both hypoxic and
normoxic conditions. Therefore, the present study is in
accordance with Terrados et al. (1988) who investigated
the effect of intermittent hypobaric hypoxic training
(2,300 m; 45 sessions per week for 34 weeks) in elite
cyclists. These authors found an increase in physiological
work capacity and maximal power output; however, no
123
Eur J Appl Physiol (2007) 101:359368
IHT (n = 10)
W4 W0 W4
280.1 5.8 * 255.3 4.1 257.5 4.0
96.9 1.4 94.6 1.1 95.6 0.8
179.6 1.8 186.2 1.2 183.7 0.8
136.9 2.4 144.8 2.5 142.4 1.8
17.0 0.2 16.9 0.1 16.5 0.1
difference was observed between the experimental hypoxic
group and a control normoxic group, i.e. no additional
effect of the hypoxic stimulus. In addition, Roels et al.
(2005) concluded that IHT of two high-intensity cycling
sessions per week (10090% relative peak power output,
identical to the present high-intensity sessions) for 7 weeks
at a simulated altitude of 3,000 m for well-trained cyclists
and triathletes did not improve performance at sea level,
i.e. PPO and 10-min cycle time trial nor VO _ 2 max to a
greater extent than a similar sea level training. Similarly,
Truijens et al. (2003) found that 5 weeks of intermittent
hypoxic training at a simulated altitude of 2,500 m, in a
swimming flume improved sea-level performance in well-
trained swimmers, but no significant additive effect of
hypoxic training was observed. Geiser et al. (2001) have
also shown, in previously untrained subjects, that IHT of
6 weeks, with five 30 min sessions per day at 3,850 m,
improved normoxic performance to the same extent as a
similar normoxic training while improved to a larger extent
the performance at altitude. However, this study indicates
that the training intensity performed may reflect the sub-
sequent adaptations observed. The peripheral mechanisms
behind this are largely not understood and could be a
function of differential adaptations in skeletal muscle
oxidative capacity. That aside the data from this study and
others indicated that no additional effect on endurance
performance occurs when a hypoxic stimulus is superim-
posed on an intensified training stimulus.
The present study did not observe any changes in Paver
for the IHT group. The reasons of differences between NT
and IHT on 10-min time-trial are unclear. In normoxia, the
present improvement after six sessions of interval-training
at PPO is in agreement with previous studies (Stepto et al.
1999; Weston et al. 1997). Stepto et al. (1999) found that
interval training at 85% of peak power output in well-
trained endurance cyclists increased significantly the
simulated 40 km TT performance. Moreover, these authors
Eur J Appl Physiol (2007) 101:359368 365

5.06 0.005

32.0 0.007
1.14 0.005
0.051 0.002

Mean SE; W0: pre; W1: first week of experimental training; W2: second week of experimental training; W3: third week of experimental training W4: post; NT: normoxic training group
(n = 8); IHT: intermittent hypoxic training group (n = 10); RBC: red blood cell; Hb: haemoglobin; Hct: hematocrit; RBD: red blood disc; MCV: mean cell volume; AGCH: averaged globular
found a reduced plasma lactate concentration at the same

14.80 0.1
46.27 0.4
92.0 0.4
246.7 1.4
29.39 0.1

105.6 0.3
absolute work rate after intense interval training. In addi-
IHT tion, Weston et al. (1997) observed a significant increase in
muscle buffering capacity in well-trained cyclists after six
high-intensity training sessions. These results indicated that

0.046 0.001
skeletal muscle buffering capacity may be an important
5.09 0.07

0.88 0.03
14.7 0.2
44.61 0.5
88.14 0.5
241.7 5.4
28.94 0.2
32.90 0.1

99.33 0.9
determinant of relatively short-duration (< 60 min)
endurance cycling activity (Weston et al. 1997). Why the
W4
NT

performance did not improve to the same extent after IHT

might be related to several mechanisms including a lower

0.052 0.002
5.0 0.05

1.07 0.03
absolute power output developed during the hypoxic
14.65 0.1
45.60 0.3
91.83 0.4
257.5 2.4
29.48 0.2
32.17 0.1

106.3 0.6
interval-session (85.7 8.1%; mean % of normoxic PPO).
Indeed, a study in untrained subjects has demonstrated that
IHT

content of haemoglobin; ACCH: averaged corpuscular concentration of haemoglobin; RET: reticulocytes; ARV: averaged reticulocytar volume
the absolute intensity of exercise is more important than the
relative exercise intensity in terms of acute skeletal muscle
0.046 0.003
5.01 0.06

0.90 0.04
14.63 0.2
44.24 0.5
88.50 0.4
241.8 4.7
29.25 0.2
33.01 0.1

99.13 0.4

adaptations to training in hypoxia (Wadley et al. 2005). In

the present study, both groups trained at the same relative
intensity; however, it might be the absolute training
W3
NT

intensity that induces larger performance changes in the

NT group. Furthermore, it has been suggested that hypoxic
0.049 0.001
4.79 0.05

32.07 0.04
1.01 0.02
14.07 0.1
43.89 0.3
91.71 0.4
247.4 3.2
29.43 0.1

107.6 0.7

training induces a decrease in muscular power due to the

reduced training intensity at altitude (detraining) (Boning
IHT

1997). It has also been shown that post-training period

following a return to sea level after an altitude training
0.045 0.003

camp is detrimental in term of specific performance. This

5.06 0.05

32.71 0.09
0.89 0.04
14.75 0.2
45.04 0.5
89.25 0.5
225.9 4.5
29.15 0.2

101.1 1.0

detrimental effect is probably due to the re-adaptations of

performance and physiological parameters to normal O2
W2
NT

content (Dick 1992). Although this loss of adaptation has

not been described after IHT, one may suggest that the
0.053 0.002
4.89 0.04

32.11 0.03
1.09 0.04

post-performance that occurred exactly at this time might

14.37 0.1
44.80 0.3
91.71 0.4
262.1 4.2
29.43 0.1

106.3 0.3

have been affected.

Hypoxia per se affects the structural and biochemical
IHT

properties of skeletal muscle by inducing a change in the

profile of type I (oxidative) to type II (glycolytic) fibres
0.046 0.001
4.95 0.05

0.73 0.01

(Ishihara et al. 1994; Itoh et al. 1995). This adaptation

14.43 0.2
43.70 0.4
88.6 0.4
236.6 4.0
29.15 0.2
33.0 0.1

101.9 0.6

will affect the changes in substrate utilization (Roberts

et al. 1996a, b), by shifting to a more efficient and eco-
W1
NT

nomic mode of oxygen utilization due to the lack of

oxygen in the tissue under hypoxic conditions. However,
0.071 0.004
5.15 0.05
14.96 0.09

31.80 0.07
1.34 0.06
47.06 0.3
91.57 0.4
245.6 1.3
29.14 0.2

104.7 0.6

it was observed that 3 weeks of IHT seemed to alter the

intrinsic properties of mitochondrial function in vastus
IHT

lateralis muscle in well-trained athletes (Roels et al.

2007). They observed a change in the muscle substrate
Table 6 Haematological variables

0.046 0.003

preference with a shift towards less fat oxidation and an

0.92 0.04
5.13 0.1
15.14 0.2
45.71 0.6
89.0 0.5
253.1 5.2
29.58 0.2
33.14 0.1

104.2 0.9

increase in carbohydrate oxidation, whereas the normoxic

control group did not show any significant differences
W0
NT

(Roels et al. 2007). This shift in substrate utilization

might explain the lack of improvement in Paver under
RBD (103 mm3)
RBC (106 mm3)

RET (10 mm )

normoxic conditions and is a detrimental effect of hy-

3

poxia on the peripheral level and upon the immediate

MCV (lm )

ARV (lm3)
ACCH (pg)
AGCH (%)
3
Hb (g dl1)

performance under normoxic conditions. Unfortunately,

RET (%)
Hct (%)

these potential explanations are hypothetical and further

investigation is necessary.

123
366
Effects of training in hypoxia on endurance
performance in hypoxia
There is a scarcity of studies examining the effect of
endurance training performed in hypoxia on performance
in hypoxia in experienced endurance athletes. In the pres-
ent study, we found that PPO measured in an incremental
exercise test to exhaustion in hypoxia significantly im-
proved following IHT. In a recent investigation maximal
running velocity, measured during an incremental exercise
test performed in hypoxia, improved following IHT in
trained athletes (Dufour et al. 2006). This is similar to that
observed in the present investigation. Therefore, the data
from his study and others confirms that altitude training or
simulated altitude training could be useful preparation for
performance and/or competition at altitude (Dufour et al.
2006; Wilber 2001).
A possible mechanism could be in the hypoxic venti-
latory response (HVR) after hypoxic stimulus and
endurance training. HVR is enhanced by intermittent
hypoxic exposure (Katayama et al. 1998) and is a posi-
tive adaptation because of its increased VE _ which im-
proves alveolar O2 pressure and increases arterial
oxygenation under hypoxic conditions (Huang et al.
1984). In contrast, HVR decreases after endurance
training (Katayama et al. 1999). In addition, not only
HVR but also cardiovascular responses to hypoxia might
be an underlying mechanism. HVR and systolic blood
pressure responses were significantly increased after a 2-
week intermittent hypoxic training program (5 sessions
per week; 30 min per day; 70% VO _ 2 max ; 4,500 m), but
decreased significantly in a normoxic control group
(Katayama et al. 2004). However, it could be that other
peripheral muscle factors are associated with the changes
observed following the training in hypoxia. For example,
Hoppeler and Fluck (2003) observed a significant in-
crease in vastus lateralis sub-sarcolemmal mitochondria
to compensate decreased O2 availability after an inter-
mittent hypoxic training protocol. Other studies in trained
runners have shown that a 6-week period of low to
moderate intensity training in hypoxia (as compared to
the same training in normoxia) resulted in improvements
in endurance performance which were associated with
increases in mitochondrial function as well as increases in
the expression of genes related to lactate shuttling and
glucose transport, all of which could be important in
terms of endurance performance (Ponsot et al. 2006; Zoll
et al. 2006). These studies suggest that hypoxic training
induce muscular adaptations, which are less expressed or
even absent after normoxic training and which may ex-
plain the observed increase in PPO at altitude. However,
further studies are required to confirm at a protein level,
123
Eur J Appl Physiol (2007) 101:359368
skeletal muscle adaptations to training in hypoxia as
compared to training in normoxia in athletes.
Effects of training on haematological parameters
No changes in RBC, Hb, and Hct nor in the other measured
haematological variables were observed in this study.
Thus, data from this study seems to indicate that hypoxic
exposure needs to exceed 382 min per week to obtain an
effect on haematological variables. However, Rodriguez
et al. (2000) found that hypoxic exposure (4,0005,000 m)
of only 90 min, three times a week for 3 weeks, i.e.
270 min per week, significantly increased Hct, RBC, RET,
and Hb concentration. A continuous exposure of 90 min
seems to represent the minimal stimulus to trigger an acute
EPO-secretion. However, no control group (training in
normoxia) was included in this investigation to establish
whether the training stress in hypoxia was associated with
any of these changes. Moreover, it has been shown that
haematological variables are already significantly in-
creased after 114 min at 3,000 m (Eckardt et al. 1989).
However, although the short duration of the hypoxic
stimulus used in this study might be responsible for the
lack of changes in the haematological variables, other
studies using longer duration of hypoxic stimulus (Terrados
et al. 1988; Vallier et al. 1996) have observed no changes
in selected haematological variables. The interval training
completed by our subjects was at a power output of 100%
of PPO. This form of training induces and elevated blood
lactate and associated metabolic acidosis (Billat 2001).
Metabolic acidosis inhibits EPO secretion, which might be
an explanation for the lack of haematological changes
found in this study (Eckardt et al. 1990). Beside the
duration of the hypoxic stimulus, the level of the hypoxic
stimulus might also be a factor involved in the EPO
secretion and in turn the increased RBC. The present study
used a simulated altitude of 3,000 m, which was lower than
the altitudes used in previous studies (Garcia et al. 2000;
Rodriguez et al. 2000).
Inadequate iron stores are a contributing factor in the
lack of changes in haematological variables. Extra iron is
necessary for the increase in production of red cells stim-
ulated by exposure to altitude (Rodriguez et al. 2000).
However, in this investigation all athletes were given iron
supplements to offset any iron deficiency. Therefore, it is
unlikely this factor was influential in the present study.
Finally, using hypoxic training for enhancement of sea-
level performance based on increased EPO, leading to in-
creased RBC (Levine 2002) is highly debated as Calbet
et al. (2005) suggested that an increased RBC and in-
creased Hb was only responsible for 16% of the variance in
increased VO_ 2 max and in turn in endurance performance.
Eur J Appl Physiol (2007) 101:359368
High intensity interval training and endurance
performance
Of minor note, the data from this investigation shows
3 weeks with two sessions of interval training (100% PPO)
and three steady state, submaximal work-training sessions
are sufficient to obtain significant improvements in sea-
level endurance performance in athletes with already a
substantial endurance training background. These adapta-
tions were associated with no improvements in VO _ 2 max
under hypoxic and normoxic conditions or cycle TT
%VO _ 2 max : This is not an uncommon finding with other
studies demonstrating that endurance performance and
PPO can be elevated by as much as 10% by performing 34
weeks of interval training (23 times per week; 80% PPO)
and without an increase in VO _ 2 max (Stepto et al. 1999;
Weston et al. 1997). Therefore, these studies and the data
from the present investigation show that for well-trained
athletes improvement in endurance performance may be
independent of change in VO _ 2 max and that other maximal
(i.e. PPO) sub-maximal variables can be improved with
short term (34 weeks) intensified training and may influ-
ence performance to a greater extent.
In summary, 3 weeks of intermittent hypoxic training
improved PPO measured in hypoxia in well-trained ath-
letes. However, intermittent hypoxic training was less
efficient than similar normoxic training in term of norm-
oxic endurance performance. The data from this study
suggest that the IHT model can be used for athletes pre-
paring for competition at altitude.
Acknowledgments This study was supported by the International
Olympic Committee and by the French Ministry of Sport. Additional
funding for the study was provided by Faculty Research Grants
Westminster University, London, UK and University of New South
Wales, Sydney, Australia.
References
Billat V (2001) Interval training for performance: a scientific and
empirical practice. Part 1: aerobic interval training. Sports Med
31(1):1331
Boning D (1997) Altitude and hypoxia traininga short review. Int J
Sports Med 18:565570
Calbet JA, Boushel R, Lundby C (2005) Comments on point:coun-
terpoint Positive effects of intermittent hypoxia (live high:train
low) on exercise performance are/are not mediated primarely by
augmented red cell volume. J Appl Physiol 100:749
Dick FW (1992) Training at altitude in practice. Int J Sports Med
13(suppl 1):S203S206
Dufour SP, Ponsot E, Zoll J, Doutreleau S, Lonsdorfer-Wolf E, Geny
B, Lampert E, Fluck M, Hoppeler H, Billat V, Mettauer B,
Richard R, Lonsdorfer J (2006) Exercise training in normobaric
hypoxia in endurance runners. I. Improvement in aerobic
performance capacity. J Appl Physiol 100:12381248
367
Eckardt KU, Boutellier U, Kurtz A, Schopen M, Koller EA, Bauer C
(1989) Rate of erythropoietin formation in humans in response to
acute hypobaric hypoxia. J Appl Physiol 66:17851788
Eckardt KU, Kurtz A, Bauer C (1990) Triggering of erythropoietin
production by hypoxia is inhibited by respiratory and metabolic
acidosis. Am J Physiol 258:R678R683
Fulco CS, Rock PB, Cymerman A (2000) Improving athletic
performance: is altitude residence or altitude training helpful?
Aviat Space Environ Med 71:162171
Garcia N, Hopkins SR, Powell FL (2000) Effects of intermittent
hypoxia on the isocapnic hypoxic ventilatory response and
erythropoiesis in humans. Respir Physiol 123:3949
Geiser J, Vogt M, Billeter R, Zuleger C, Belforti F, Hoppeler H
(2001) Training high-living low: changes of aerobic performance
and muscle structure with training at simulated altitude. Int J
Sports Med 22:579585
Hendriksen IJ, Meeuwsen T (2003) The effect of intermittent training
in hypobaric hypoxia on sea-level exercise: a cross-over study in
humans. Eur J Appl Physiol 88:396403
Hoppeler H, Fluck M (2003) Plasticity of skeletal muscle mitochon-
dria: structure and function. Med Sci Sports Exerc 35:95104
Huang SY, Alexander JK, Grover RF, Maher JT, McCullough RE,
McCullough RG, Moore LG, Sampson JB, Weil JV, Reeves JT
(1984) Hypocapnia and sustained hypoxia blunt ventilation on
arrival at high altitude. J Appl Physiol 56:602606
Ishihara A, Itoh K, Itoh M, Hirofuji C, Hayashi H (1994) Hypobaric-
hypoxic exposure and histochemical responses of soleus muscle
fibers in the rat. Acta Histochem 96:7480
Itoh K, Itoh M, Ishihara A, Hirofuji C, Hayashi H (1995) Influence of
12 weeks of hypobaric hypoxia on fibre type composition of the
rat soleus muscle. Acta Physiol Scand 154:417418
Katayama K, Sato Y, Ishida K, Mori S, Miyamura M (1998) The
effects of intermittent exposure to hypoxia during endurance
exercise training on the ventilatory responses to hypoxia and
hypercapnia in humans. Eur J Appl Physiol Occup Physiol
78:189194
Katayama K, Sato Y, Morotome Y, Shima N, Ishida K, Mori S,
Miyamura M (1999) Ventilatory chemosensitive adaptations to
intermittent hypoxic exposure with endurance training and
detraining. J Appl Physiol 86:18051811
Katayama K, Matsuo H, Ishida K, Mori S, Miyamura M (2003)
Intermittent hypoxia improves endurance performance and
submaximal exercise efficiency. High Alt Med Biol 4:291304
Katayama K, Sato Y, Matsuo H, Ishida K, Iwasaki K, Miyamura M
(2004) Effect of intermittent hypoxia on oxygen uptake during
submaximal exercise in endurance athletes. Eur J Appl Physiol
92:7583
Levine BD (2002) Intermittent hypoxic training: fact and fancy. High
Alt Med Biol 3:177193
Mujika I, Busso T, Lacoste L, Barale F, Geyssant A, Chatard JC
(1996) Modeled responses to training and taper in competitive
swimmers. Med Sci Sports Exerc 28:251258
Ponsot E, Dufour SP, Zoll J, Doutrelau S, NGuessan B, Geny B,
Hoppeler H, Lampert E, Mettauer B, Ventura-Clapier R, Richard
R (2006) Exercise training in normobaric hypoxia in endurance
runners. II. Improvement of mitochondrial properties in skeletal
muscle. J Appl Physiol 100:12491257
Roberts AC, Butterfield GE, Cymerman A, Reeves JT, Wolfel EE,
Brooks GA (1996a) Acclimatization to 4,300-m altitude
decreases reliance on fat as a substrate. J Appl Physiol
81:17621771
Roberts AC, Reeves JT, Butterfield GE, Mazzeo RS, Sutton JR,
Wolfel EE, Brooks GA (1996b) Altitude and beta-blockade
augment glucose utilization during submaximal exercise. J Appl
Physiol 80:605615
123
368
Rodriguez FA, Casas H, Casas M, Pages T, Rama R, Ricart A,
Ventura JL, Ibanez J, Viscor G (1999) Intermittent hypobaric
hypoxia stimulates erythropoiesis and improves aerobic capac-
ity. Med Sci Sports Exerc 31:264268
Rodriguez FA, Ventura JL, Casas M, Casas H, Pages T, Rama R,
Ricart A, Palacios L, Viscor G (2000) Erythropoietin acute
reaction and haematological adaptations to short, intermittent
hypobaric hypoxia. Eur J Appl Physiol 82:170177
Rodriguez FA, Murio J, Ventura JL (2003) Effects of intermittent
hypobaric hypoxia and altitude training on physiological and
performance parameters in swimmers [abstract]. Med Sci Sports
35:S115
Roels B, Millet GP, Marcoux CJ, Coste O, Bentley DJ, Candau RB
(2005) Effects of hypoxic interval training on cycling perfor-
mance. Med Sci Sports Exerc 37:138146
Roels B, Thomas C, Bentley DJ, Mercier J, Hayot M, Millet GP
(2007) Effects of intermittent hypoxic training on amino and
fatty acid oxidative combustion in human permeabilized muscle
fibers. J Appl Physiol 102:7986
Rusko HR (1996) New aspects of altitude training. Am J Sports Med
24:S48S52
Stepto NK, Hawley JA, Dennis SC, Hopkins WG (1999) Effects of
different interval-training programs on cycling time-trial perfor-
mance. Med Sci Sports Exerc 31:736741
Terrados N, Melichna J, Sylven C, Jansson E, Kaijser L (1988)
Effects of training at simulated altitude on performance and
muscle metabolic capacity in competitive road cyclists. Eur J
Appl Physiol Occup Physiol 57:203209
123
The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.
Eur J Appl Physiol (2007) 101:359368
Truijens MJ, Toussaint HM, Dow J, Levine BD (2003) Effect of high-
intensity hypoxic training on sea-level swimming performances.
J Appl Physiol 94:733743
Vallier JM, Chateau P, Guezennec CY (1996) Effects of physical
training in a hypobaric chamber on the physical performance of
competitive triathletes. Eur J Appl Physiol Occup Physiol
73:471478
Villa JG, Lucia A, Marroyo JA, Avila C, Jimenez F, Garcia-Lopez J,
Earnest GP, Cordova A (2005) Does intermittent hypoxia
increase erythropoiesis in professional cyclists during a 3-week
race ? Can J Appl Physiol 30:6173
Wadley GD, Lee-Young RS, Canny BJ, Wasuntarawat C, Chen ZP,
Hargreaves M, Kemp BE, McConell GK (2005) Effect of
exercise intensity and hypoxia on skeletal muscle AMPK
signaling and substrate metabolism in humans. Am J Physiol
290:E694-E702
Weston AR, Myburgh KH, Lindsay FH, Dennis SC, Noakes TD,
Hawley JA (1997) Skeletal muscle buffering capacity and
endurance performance after high-intensity interval training by
well-trained cyclists. Eur J Appl Physiol Occup Physiol 75:713
Wilber RL (2001) Current trends in altitude training. Sports Med
31:249265
Zoll J, Ponsot E, Dufour SP, Doutrelau S, Ventura-Clapier R, Vogt M,
Hoppeler H, Richard R, Fluck M (2006) Exercise training in
normobaric hypoxia in endurance runners. III. Muscular adap-
tations of selected gene transcripts. J Appl Physiol 100:1258
1266