Eur J Appl Physiol (1995) 70:115-125
© Springer-Verlag 1995
K.I. Norton . B. Squires • L.H. Norton • N.P. Craig
P. McGrath • T.S. Olds
Exercise stimulus increases ventilation
from maximal to supramaximal intensity
Accepted: 2 August 1994
Abstract This study investigated the influence of an
exercise stimulus on pulmonary ventilation (12E)during
severe levels of exercise in a group of ten athletes. The
altered ventilation was assessed in relation to its effect
on blood gas status, in particular to the incidence and
severity of exercise induced hypoxaemia. Direct
measurements of arterial blood were made at rest and
during the last 15 s of two intense periods of cycling;
once at an intensity found to elicit maximal oxygen
uptake (l?Ozm~x; MAX) and once at an intensity established to require 115% of l;'Ozmax (SMAX). Oxygen
uptake (1)'O2) and ventilatory markers were continually
recorded during the exercise and respiratory flow-volume loops were measured at rest and during the final
30 s of each minute for both exercise intensities. When
compared to MAX exercise, the subjects had higher
ventilation and partial pressure of arterial oxygen
(PaO2) during the SMAX intensity. Regression analysis
for both conditions indicated the levels of PaOz and
oxygen saturation of arterial blood (S~O2) were positively correlated with relative levels of ventilation during exercise. It was apparent that mechanical constraints to ventilate further were not present during the
MAX test since the subjects were able to elevate l;z
during SMAX and attenuate the level of hypoxaemia.
This was also confirmed by analysis of the flow volume
recordings. These data support the conclusions firstly,
K.I. Norton ( I ~ ) " T.S. Olds
Human Bioenergetics Laboratory, School of Sport and Leisure
Studies, The University of New South Wales, P.O. Box 88, Oatley,
NSW 2233, Australia
B. Squires • P. McGrath
Department of Physiology, University of Newcastle,
New South Wales, Newcastle, Australia
L.H. Norton
John Hunter Hospital, Newcastle, New South Wales, Australia
N. P.Craig
The South Australian Sports Institute, Adelaide, South Australia
that overwhelming mechanical constraints on l/E were
not present during the MAX exercise, secondly, the
subjects exhibiting the most severe hypoxaemia had no
consistent relationship with any measure of expiratory
flow limitation, and thirdly, ventilatory patterns during
intense exercise are strong predictors of blood gas
status.
Key words Exercise-induced hypoxaemia
Exercise stimulus - Flow-volume loops
Supramaximal exercise • Ventilatory drives
Introduction
Traditionally, it has been accepted that the pulmonary
system does not represent a limitation to maximal
performance in humans (Cerretelli and di Prampero
1987). This assumption is based upon findings that
arterial blood remains at or near resting oxygen saturation (S~O2) levels (i.e. %S~O2 equal to or greater than
95%) even during intense exercise, in the healthy but
untrained population. Although exceptions to this generalisation had occasionally been reported (Holgram
and Linderholm 1958; Rowell et al. 1964), it was not
until the last decade that several detailed studies have
confirmed the existence of exercise induced hypoxaemia (EIH) in athletes exercising maximally at sea
level (Dempsey et al. 1984; Powers et al. 1989a; TorreBueno et al. 1985).
It has now become evident from other studies that in
a substantial number of high-performance athletes performing aerobic exercise at maximal intensity, the pulmonary system does represent a limitation to performance (Hopkins and McKenzie 1989; Powers et al.
1988, 1989b; Williams et al. 1986). The most obvious
sign of this pulmonary insufficiency has been an exaggerated widening of the alveolar-arterial POz difference
( P A - a O ; ) with increasing exercise intensities (Johnson
et al. 1992; Warren et al. 1991). It has also been
116
reported that arterial oxygen tension ( P a O 2 ) may decrease to below 8 kPa (60 mmHg) and that there is
a concomitant drop in %S~O2 to as low as 85%-86%
(Dempsey et al. 1984; Powers et al. 1989b; Williams
et al. 1986).
Although several mechanisms to explain these findings have been proposed, the precise etiology of EIH
remains to be determined (Johnson et al. 1992; Powers
and Williams 1987). It has been suggested that one
potential cause of EIH may be a relative hypoventilation during maximal exercise (Dempsey et al. 1984;
Miyachi and Tabata 1992). If there is a less than adequate hyperpnoeic response during intense exercise,
this may result in a suboptimal alveolar oxygen partial
pressure (PAO2) and a corresponding reduced driving
force for 02 diffusion according to Fick's Law.
There has been considerable debate, however, as to
whether athletes with the lowest SaO2 levels during
exercise do (Dempsey et al. 1984; Miyachi and Tabata
1992; Powers et al. 1984; 1988; Warren et al 1991), or do
not (Powers et al. 1992; Williams et al. 1986) exhibit
reduced levels of relative ventilation [indicated by
a low ventilatory equivalent for oxygen (pulmonary
ventilation/oxygen uptake, 12E/1202) I. Some authors
have suggested a mechanical limitation exists in trained
subjects exhibiting EIH, preventing further 1/E during
intense exercise (Johnson et al. 1992). On the other
hand, it has been shown that 12E can be increased
during severe exercise in some, but not all, subjects
following chemical stimulation using either dead space
loading (McParland et al. 1991), increased inspired
CO2 breathing (Johnson et al. 1992; Fig. 5) or pharmacological chemoreceptor stimulation (Naeije et al.
1993).
This study was designed to determine the relationship between aspects of 12E during exercise at maximal
and supramaximal levels, and the development of hypoxaemia in a group of high-performance athletes. Specifically, we addressed the question of whether absolute
and relative levels of 12E are altered as a consequence of
an exercise stimulus beyond that intensity found to
elicit maximum oxygen uptake (!)O2m,x). We analysed
the change in 1/E in relation to the development and
severity of EIH. In addition, we investigated possible
reasons for suboptimal ventilatory responses during
exercise.
Methods
EIH (Powers et al. 1988; Williams et al. 1986). The subjects gave
informed consent after being given a detailed description of the test
protocol, special emphasis being placed upon the risks associated
with invasive measurements. All procedures used were performed in
accordance with the guiding ethics for human experimentation outlined by the National Health and Medical Research Council of
Australia.
Experimental protocol
Prior to the test day, the subjects visited the laboratory on at least
two occasions for familiarization with the equipment and personnel
and so that measurements of 1202max and mechanical efficiency
could be made. The test for 1202max involved a standard cycle
ergometer protocol employing 2-min intervals, beginning at 150 W
and increasing in 25-W increments until the subjects reached exhaustion. Mechanical efficiency while cycling was determined for
each subject using five 4-rain stages at progressively increasing
submaximaI exercise intensities as has been previously described
(Olds et al. 1993). The mean 1202 during the last 2 rain of each stage
was regressed against exercise intensity (in watts). Regression lines
were then used to determine appropriate supramaximal exercise
intensities as outlined below.
On the day of the experiment, the subjects were measured for
stretch standing height to the nearest 0.1 cm with a wall stadiometer
and body mass to the nearest 50 g with a calibrated beam scale.
Following this, the subjects assumed a supine position on a preparation table and electrodes were placed in the standard V5 position for
electrocardiogram recording. A radial artery catheter was inserted
and the subjects then moved to the ergometer and were instructed to
sit comfortably in position on the cycle for at least 15 rain prior to
exercise.
The experimental protocol involved two exercise tests on the cycle
ergometer, each of up to 4-rain duration. The tests were performed
on the same day with 1 h separating each of the following intensities:
(1) once at an intensity found to elicit 1202max (MAX), and (2) once at
an intensity estimated to require 115% of 1202max (SMAX). Each of
the intensities was conducted in an identical fashion and the order
was randomized to preclude any bias arising from the exercise
administration.
Warm-up was not standardised, rather the subjects were instructed to prepare as they would for competition. Warm-up was then
followed by a compulsory 3 min rest period. Baseline, pre-exercise
blood samples were taken at the end of the rest period, prior to each
exercise test. Three maximal voluntary flow-volume loops were then
performed by the subjects. This was done while in position on the
cycle and connected to the respiratory valve apparatus in the same
way as that used during the exercise tests. The subjects remained
attached to the respiratory apparatus following these manoeuvres to
facilitate correct placement of the exercise tidal flow-volume loops
within the voluntary loops.
In addition to pre-exercise measurements, the following variables
were measured during each test: continuous measurements of heart
rate (HR) and respiratory variables, exercise tidal flow-volume loops
over 30 s intervals beginning at 30, 90, 150, and 210 s, and an arterial
blood sample during the last 15 s of exercise as the subject was
nearing exhaustion.
Subjects
Gas analysis
Ten healthy, non-smoking adult men participated in this study. The
subjects were all well trained but varied in performance abilities.
Included in the group were one amateur cyclist, four competitive
triathletes, one mountain bike cyclist eight times national champion), and four Olympic track cycling medalists (including one
world champion). Trained athletes were deliberately targeted for this
study to improve the likelihood of observing significant levels of
The g o 2 m e a s u r e m e n t s were determined using a breath-by-breath
metabolic assessment system (Ametek, model OCM2). For further
analysis, and to overcome some of the random variations in ventilatory measurements, data were averaged and plotted for 15-s
intervals. The gas analysis system incorporated a lowresistance respiratory valve (Rudolph, model 2700) attached to an
117
inertially-compensated unidirectional turbine flowmeter on the inspiratory side. Expired gas samples passed through short, large
diameter tubing (34 mm) to a 4-1 mixing chamber. Gas samples were
continually monitored for 02 and CO2 concentrations by fast response analysers (Ametek S3A/IO2 and CD 3A CO2 analysers,
respectively). A time delay factor was used to align inspiratory gas
volumes with expiratory gas analysis. The gas analysers were calibrated prior to each test with gases whose concentrations had been
verified densitometrically. The flow meter was calibrated throughout the range 0-4001.min- 1 constant flow and 0-1801.min- 1 puisatile flow rate using a Tissot spirometer and calibrated syringe,
respectively. Resistance to gas flow was measured (as a pressure
drop) for both the inspiratory and expiratory sides of the full system
during these calibration checks using a manometer. Resistance to
flow at 225 l'min -1 was less than 5 cmH20 for both sides of the
system and within the range recommended (Thoden 1991).
exercise by adhesive tape. Peak temperature was recorded within
2-rain post-exercise and used to adjust blood gas tensions. A second
blood sample (15 ml) was collected during the exercise test immediately following the first (blood gas) sample. This was subsequently
divided for analysis of lactate, catecholamine (adrenaline and noradrenaline) and potassium concentrations. Lactate concentration was
determined spectrophotometrically (Sigma kit #726-UV) and
serum potassium assessed using flame photometry (Instrumentation
Laboratory, model 943). Plasma catecholamine concentrations were
assessed via gas chromatography-mass spectrometry (Duncan et al.
1988). The sensitivity of the catecholamine assays was 0.05pmol- 1- i with a coefficient of variation of a known standard being
less than 0.4%. All blood analyses were performed in duplicate and
the mean value used in further procedures.
Cycle ergometer
Flow-volume analysis
Flow-volume data were measured using a modified pneumotachograph (P.K. Morgan, model 098) incorporated into the respiratory
valve. The pneumotachograph was attached between the mouthpiece and Rudolph respiratory valve so that a minimal increase in
dead space (approximately 40 ml) resulted. The voltage signal resulting flom the measured pressure differences was sent to a computer
for conversion to real-time flow-volume loops. These loops were
recorded in discrete files, each containing approximately 30 s of
measurements. Average flow-volume loops for each file were deduced by visual inspection when data were superimposed on one
another in a computer printout. The estimated mean loops for each
subject were then scanned into a computer program which allowed
minor smoothing to be performed. Composite group mean flowvolume loops were based on the average of all individual loops
(breaths) produced by analysis of the final 30 s of measurements for
each subject in both exercise tests. The exercise loops were plotted
relative to the group mean forced vital capacity, peak voluntary
inspiratory and expiratory flow rates and expiratory reserve volumes (difference between exercise end tidal volume and end expiration in the voluntary ventilation test). These values were all determined from measurements taken on the mean flow-volume loops of
the individual subject.
All tests were conducted on a sport-specific, air-braked ergometer,
details of which have been published previously (Craig and Walsh
1989). The ergometer was calibrated throughout the physiological
range of exercise intensities using a dynamic calibration torque
meter. During each test the power output in watts and total work
done in kilojoules were continuously measured from flywheel revolution rates and stored on disk using a computer interface. The
subjects were given instructions about the required power output for
each test and were given instantaneous feedback of pedal cadence
via a digital display. The test was concluded if the subject reached
4rain or failed to hold the pedal cadence within 3 rpm of the
required cadence for more than 5 s.
Statistical analyses
Paired Wilcoxon tests were used to determine if there were any
differences between group means in the exercise treatments and
between selected pre-exercise and exercise conditions. Least squares
linear regression analysis was employed to determine the relationship between dependent and independent variables. A probability
level of 0.05 was accepted to indicate significance for all statistical
analyses.
Results
Blood collection and analysis
A catheter was inserted under local anaesthetic (procaine hydrochloride) into the radial artery of the nondominant arm and sutured into
position by an experienced anaesthetist. Catheter patency was maintained using positive pressure infusion (3 ml'h -1) of heparinised
saline (10 U. mi- 1). Withdrawal of blood was performed in an identical fashion prior to and during the exercise tests. Following the
removal of the saline flush, a 2.5-ml sample was collected anaerobically into a heparinized arterial blood gas syringe, capped and
immediately placed in an ice bath. Analysis was performed in duplicate within 20rain of collection using either a hemoximeter
(Radiometer, model OSM2), or blood gas analyser (Ciba-Corning,
model 278) which directly measured blood pH, partial pressures of
oxygen and carbon dioxide (PO2, PCO2) and calculated further
variables based on these measurements. The ideal alveolar gas
equation was used to determine alveolar PO2 (PAO2) and PA-aO2 as
described previously (Hopkins and McKenzie 1989). Both analysers
were calibrated prior to the test sessions using standard solutions.
Blood gas values were adjusted for deviations recorded in body
temperature and pH due to exercise, according to procedures outlined previously (Dempsey et al. 1984; Kelman and Nunn 1966).
Body core temperature was determined using a rectal probe positioned 12-cm past the anal sphincter and held in place during the
P h y s i c a l c h a r a c t e r i s t i c s of the subjects used i n this
s t u d y are s h o w n in T a b l e 1.
T h e l/O2max values m e a s u r e d d u r i n g the p r e l i m i n a r y
test were n o different f r o m those m e a s u r e d d u r i n g
either the M A X or S M A X e x p e r i m e n t s . As expected,
high p o w e r o u t p u t s were a c h i e v e d b y the athletes d u r i n g the exercise tests. All subjects c o m p l e t e d 4 r a i n at
the M A X level a n d the m e a n time for the S M A X
p r o t o c o l was 174 ( S E M 10)s. N o s u b j e c t was able to
m a i n t a i n the S M A X exercise i n t e n s i t y (i.e. 115%
l/O2max) for the full 4 rain; times r a n g e d f r o m 145 to
210 s. P e a k H R d u r i n g b o t h tests a v e r a g e d 189 ( S E M 3)
b e a t s r a i n - 1 w h i c h was n o different f r o m t h a t m e a s u r e d d u r i n g the p r e l i m i n a r y i n c r e m e n t a l test to e x h a u s t i o n [ 1 9 t ( S E M 4) b e a t s . m i n - l l .
T a b l e 2 o u t l i n e s the v e n t i l a t o r y r e s p o n s e s d u r i n g the
two exercise tests. It s h o u l d be p o i n t e d o u t t h a t preexercise d a t a d o n o t reflect r e s t i n g levels since there
were o b v i o u s l y c a r r y - o v e r effects f r o m the w a r m - u p i n
118
Table 1 Physical characteristics and performance attributes of subjects
Variable
Mean
SEM
Age (Year)
Height (cm)
Mass (kg)
Maximal 02 uptake a
(1'm i n - 1)
Power output MAX
(W)
SMAX
22.9
179.0
73.4
4.77
1.9
1.5
2.3
0.19
365
421
15
18
Range
18 35
172.0 187.3
60.6 83.1
3.40 5.47
275-412
306-479
MAX, Intensity found to elicit maximal 0 2 uptake; SMAX, intensity
estimated to require 115% of maximal O2uptake
aMeasured during preliminary laboratory test
addition to expected anticipatory responses. The data
are important, however, since they allow relative changes during the exercise to be seen easily. In addition,
they demonstrated that the subjects were beginning
exercise in the same physiological state. Differences
during exercise (MAX compared to SMAX) were found
for all variables except aerobic power and ventilatory
equivalent for CO2.
Arterial blood gases assessed during the final 15 s of
each exercise test are given in Table 3.
The PaO2 and SaO2 decreased from pre-exercise to
exercise conditions for both exercise intensities
(P < 0.05). There was a difference in P,O2 between
MAX and SMAX exercise intensity. The SaO2 ranged
from 90.8%-95.4% during MAX exercise and from
91.8%-96.5% during the SMAX exercise. The differences between MAX and SMAX were also significant.
When compared to the MAX test results all but two of
the subjects had a rise in S~O2 during the SMAX test
(Fig. 1), coincident with a significant elevation in I) E
(15.71.min -1) from MAX to SMAX exercise. Individual partial pressure of arterial carbon dioxide
(PaCO2) values during the two exercise tests are illustrated in Fig. 1. There were no differences in PaCOa
Table 2 Ventilatory responses pre-exercise
and during the two exercise conditions
between exercise intensities (P = 0.96) although differences were obvious among subjects. Temperature
and pH corrected PaO2 ranged from 11.3 to 14.2kPa
(85.2-106.8 mmHg) during MAX to 11.6 to 14.3 kPa
(87.4-107.5 mmHg) during the SMAX exercise.
Temperature change during exercise averaged
0.5(SEM 0.06)°C (37.8-38.3°C) for the MAX and
0.4(SEM 0.07)°C (37.9-38.3 °C) for the SMAX exercise
intensity.
The PA_aO2 widened from pre-exercise to exercise
under both conditions. The difference between MAX
and SMAX exercise during the last 15 s was significant.
Individual values ranged from 1.9 to 4.0kPa
(14-30mmHg) during MAX and from 0.7-3.8 kPa
(5-29 mmHg) during the SMAX intensity. There was
a significant negative correlation between the level of
PaO2 during exercise and the PA_aO2(Fig. 2).
Table 4 shows the concentrations of various blood
biochemicals pre-exercise and during the exercise tests.
With the exception of blood pH and lactate concentrations there were elevations in all other chemical stimulants of I?E from MAX to SMAX exercise intensity.
During MAX and SMAX exercise intensities respectively, individual concentrations ranged as follows; pH,
7.12-7.25 and 7.12-7.24; lactate, 8.5-20.1 and
9.1 21.0 retool'l- 1; K +, 6.4 7.7 and 7.1-8.1 retool'l- 1;
noradrenaline, 23.2-93.9 and 35.1-114.6 nmol'l- 1 and
adrenaline, 3.8-34.6 and 11.1-60.9 nmol.1-1. On average, plasma noradrenaline concentrations rose to
about 19 and 23 times resting concentrations during
MAX and SMAX exercise whereas adrenaline concentrations increased by 10 and 20 times, respectively.
Flow-volume data were obtained on six subjects
only. The results of the flow-volume analyses collected
during the last 30 s of each test for the six subjects are
summarised in Fig. 3. Tidal volume differed between
the exercise conditions as did measured peak expiratory flow rate (PEFR). The differences in peak inspiratory flow rate (PIFR) measured during the SMAX test
Variable
O2 uptake (1-min 1)
CO2 production (l'min-1)
Pulmonary ventilation
(1.min -1, BTPS)
Pulmonary ventilation/O2
uptake
Pulmonary ventilation/CO2
production
Tidal volume (1)
Respiratory frequency
(breaths.min -1)
MAX
SMAX
Pre-exercise
Exercise
Pre-exercise
Exercise
Mean SEM
0.53 0.11
0.62 0.05
Mean SEM
4.80 0.17
5.53 0.24
Mean SEM
0.62 0.18
0.68 0.61
Mean SEM
4.90 0.19
6.11 0.26 a
16.3
1.5
142.3 7.5
18.8
2.0
158.0 7.9"
30.7
1.3
30.5
1.0
29.6
1.3
32.2
1.3~
26.3
1.0
1.31 0.09
25.7 1.3
2.97 0.12
27.6 0.7
1.49 0.12
25.9 1.5
2.83 0.13 a
12.4
48.2
12.6
55.8 2.6~
1.0
1.9
1.2
BTPS, Body temperature and pressure, saturated. For other definitions see Table 1
~Significant difference MAX versus SMAX exercise (P < 0.05)
119
Table 3 Blood gas values and alveolararterial difference in oxygen tension preexercise and during the two exercise
conditions
Variable
MAX
SMAX
Pre-exercise
Mean
% 02 saturation
97.6
Partial pressure of arterial CO2
(kPa)
5.5
Partial pressure of arterial 02
(kPa)
14.9
Partial pressure of alveolar 02
(kPa)
15.2
Alveolar-arterial difference
in partial pressure of Oz (kPa)
0.3
SEM
0.1
Exercise
Pre-exercise
Exercise
Mean SEM
93.2
0.5
Mean SEM
97.6
0.2
Mean SEM
94.2
0.5 a
0.1
5.6
0.2
5.4
0.1
5.6
0.2
0.4
12.3
0.3
14.9
0.4
12.9
0.3"
0.2
15.4
0.2
15.3
0.3
15.8
0.1"
0.4
3.1
0.3
0.4
0.3
2.9
0.3"
For other definitions see Table 1.
aSignificant difference MAX versus SMAX exercise (P < 0.05)
SaO.
97
(%)
PaCO.z (kPa)
7.3
96
95
94
6.0
93
5.3
92
4.7
91
9 0 -MAX
-
4.0
SMAX
i
r
SMAX
MAX
Fig. 1 Individual arterial oxygen saturation (SaO2) percentages and
arterial carbon dioxide tensions (PaCO2) measured during the final
15 s of each exercise test. For definitions see Table 1
A-aDO2 (kPa)
5.3
4.0
0
I MAX
0 SMAX
r = -.77
0
0 00 1 ~ 0, ~ . ~~
2.7
tion of peak expiratory reserve was used throughout
the range of lung volumes naturally chosen during
MAX exercise for subject A, while subject B did not
appear to be as restricted during the same intensity. In
spite of these differences, subject B regulated to a higher
PaCO2 (and P,O2) during the MAX exercise test. The
flow-volume loops for each minute over the duration of
the exercise tests are also included. The figure demonstrates the expansion of the flow-volume loop as the
exercise progressed, despite the constant exercise intensity.
The 12E/1202 has been regressed against PaO2 (Fig.
5A) and PaCO2 (Fig. 5B) measured during both exercise tests. Figure 6 illustrates the significant relationship found between P,O2 and PAO2 in both exercise
intensities. The relationship between arterial potassium
concentration and corresponding I)E/I)O2 is shown in
Fig. 7.
(p<O.01)
Discussion
•
1,3
i
10.7
11.3
i
12.0
i
i
i
i
12.7
13.3
14.0
14.7
PaO2 (kPa)
Fig. 2 Relationship between partial pressure of alveolar O2-partial
pressure of arterial 02 (PA .02) compared to PaOz during the two
exercise tests
compared to the MAX exercise intensity did not reach
significance (P = 0.06).
Figure 4 shows examples of two subjects who used
different levels of expiratory and inspiratory flow reserve during both exercise intensities. A high propor-
Previous authors have suggested that EIH may be due
to a veno-arterial shunt (Powers and Williams 1987),
an alteration in the shape of the oxygen dissociation
curve in athletes (Rowell et al. 1964), significant ventilation-perfusion non-homogeneity (Gale et al. 1984)
and/or less than optimal I)E (Miyachi and Tabata
1992). This study was primarily concerned with assessing the incidence and effect of a reduced 12E/1202 during intense exercise. We manipulated these parameters
using an extreme exercise stimulus (supramaximal exercise) to determine whether we could alter 1#E to levels
beyond those found under maximal conditions. We
were also interested in identifying possible reasons for
the reduced 12E/1202 observed in some athletes during
periods of exceptional metabolic disturbance. Our results indicated that a significant level of arterial hypoxaemia (SaOz less than 93%) occurred in half of the
120
Table 4 Blood biochemistrypre-exercise
and during the two exerciseconditions
Variable
MAX
pH
Lactate concentration
(mmol'l t)
K + concentration
(mmol.l -~)
Noradrenaline concentration
(nmol.1-1)
Adrenaline concentration
(nmol'I- 1)
SMAX
Pre-exercise
Exercise
Pre-exercise
Exercise
Mean SEM
7.39 0.01
Mean SEM
7.18 0.01
Mean SEM
7.39 0.01
Mean SEM
7.19 0.02
1.1
0.2
12.9
1.2
1.4
0.3
13.5
1.3
4.4
0.1
6.8
0.1
4.4
0.1
7.7
0.1a
2.6
0.3
49.9
6.1
2.6
0.2
59.2
7.6 a
1.7
0.3
16.4
2.7
1.7
0.2
33.5
5.9a
For definitionssee Table 1
aSignificantdifferenceMAX versus SMAX exercise(P < 0.05)
is 1
12
PEFR
(i.sec-~) 9
PIFR
(I.sec "1)
9
,21 ~
4
~
i
i
VOLUME (L)
Fig. 3 Summary of the group mean (n = 6 subjects) ventilatory
responses under voluntary conditions and during the two exercise
tests. Values were as follows: forced vital capacity 5.38 (SEM 0.5);
tidal volume MAX 2.90 (SEM 0.2), SMAX 2.75 (SEM 0.2); peak
expiratory flow rate (PEFR) (1.s- 1) voluntary loop 12.5 (SEM 1.0);
MAX 8.21 (SEM 0.7), SMAX 9.65 (SEM 0.7); peak inspiratory flow
rate (PIFR) (1.s- ~) voluntary loop 9.68 (SEM 0.5); MAX 7.62 (SEM
0.6), SMAX 8.58 (SEM 0.4). For other definitions See Table 1
subjects exercising at intensities corresponding to
1202max and 30% of the subjects during the supramaximal intensity (115% l)O2max). When the exercise intensity was increased from maximal to supramaximal
levels, all but two of the subjects elevated 12E with
a concomitant attenuation in the degree of hypoxaemia. This occurred in spite of the fact that there
was greater oxygen demand for both the exercise and
respiratory work. Regression analyses indicated that
PaO2 during exercise was positively correlated with the
level of JTE/[')O2 as well as negatively correlated with
PA-aO2. Although cause and effect cannot be estab-
lished, the results would suggest that both 19E/120a and
a widening gradient for 02 across the lung (perhaps as
a result of the lower I/E) contributed to the reduction in
PaO2. These results, together with analysis of respiratory flow-volume data recorded during rest and exercise
indicate that the potential to ventilate further was present during the MAX exercise intensity; however, the
subjects chose not to, despite lower levels of SaO2. This
study has therefore demonstrated that an exercise
stimulus was important for further 1/E when moving
from the maximal to the supramaximal power output.
The widening of P A - a O 2 and the decrease in SaO2
from rest to exercise confirmed recent findings and
would support the view that the pulmonary system
may limit performance in a significant number of elite
level athletes (Dempsey et al. 1989; Hopkins and
McKenzie 1989; Powers et al. 1989b; Williams et al.
1986). Previous reports have indicated SaO 2 levels may
reach as low as 84%-86% during exercise (Powers
et al. 1984, 1988; Williams et al. 1986). The minimal
individual value of 90.8% SaO2 found in the present
study was not as extreme as many of these earlier
reports. Many studies reporting such low values have
used indirect pulse oximetry to assess arterial oxygenation and there have been mixed reports concerning the
appropriateness of this technique during severe exercise
(Norton et al. 1992; Powers et al. 1989a,). Notwithstanding these limitations, differences in the type of
exercise may also account for part of this discrepancy
since directly measured S,O2 values have been as low
as 85% during treadmill running (Rowell et al. 1964).
An important finding in the present study was the
greater 02 transfer during SMAX exercise (compared
to the MAX intensity), achieved by elevating alveolar
pressure through increments in ~;'E (Table 3). A small
[0.4 kPa (2.7 mmHg)] but significant rise in P A O 2 occ u r r e d in association with a mean increase of
15.7 1.rain-1 in I?E (Tables 2, 3). The P a O 2 w a s elevated
by an average of 0.6 kPa (4.7 mmHg) and S~O2 by 1%
during exercise at the higher intensity. The significant
121
SUBJECTA
15
15
IZ
12
PEFR
PEFR
(i.se61) 9
(I.sed 1)
6
PIFR
(I.sec "~ )
9
6
PIFR
6
6
(I.sec "1 )
1in-1
9
9
12
12
4
3
2
1
VOLUME (k)
t
4
2,
2
1
VOLUME (L)
SUBJECT B
PEFR
(i.sec-1)
PIFR
15
15
12
12
PEFR
(i.sec-1)
9
9
6
6
3
3
0
0
3
3
s
PIFR
s
(I.sec "1 )
(I.see 1 )
9
12
9
5
4
3
2
1
0
VOLUME (L)
1_2
5
4
3
2
1
o
VOLUME (L)
Fig. 4 Ventilatory responses for two subjects (A and B) during the two exercise tests. Relevant ventilatory and blood biochemistry data for
each subject are also given for reference (see text for discussion). 12E, Pulmonary ventilation; for other definitions See Table 1 and Figs. 1 and 2
relationship between PAO2 and PaO2 during the exercise tests would suggest alterations in PAO2 may be
important in determining corresponding changes PaO2
during exercise. It should be pointed out that despite an
elevated SaO2 during the SMAX exercise intensity
there was no measured increase in 1202. It is likely the
additional 02 delivered in the circulation (estimated at
about 60 ml with a 1% increase in S~O2) was outside of
the resolution of our measurements for whole-body
1702.
Overall, it has been demonstrated in the present
study that elevations in 12E during the supramaximal
exercise were associated with increased PAO2 and
PaO2, and with a small reduction in Pa-aO2 .
The question then arises as to why the majority of
subjects tolerated the decreased %SaO2 during exercise
and chose not to elevate I~E? This was particularly
pertinent under conditions of the MAX test. Several
potential causes of the reduced l)'E/12Oz have been
proposed. These have included a mechanical limitation
to ventilate further at the lung volumes naturally
chosen during exercise (Johnson et al. 1992), respiratory muscle fatigue (Aaron et al. 1992; Gallagher and
Younes 1989; Johnson et al. 1992; Levine and Henson
122
Fig. 5 Relationship between an
indicator of relative ventilation
(ventilatory equivalent for
oxygen; pulmonary
ventilation/oxygen uptake,
1)E/1202) and arterial oxygen
tension (P,O2; A) and between
12E/1202 and arterial carbon
dioxide tension (PaCO2; B)
during the two exercise tests.
For other definitions See Table 1
~/E/%/02
45
40
~/E/~702
45
A
r = .71
(p<O.01)
O
30
25
•
1 2.0
PaO 2 (kPa)
14.7
2/
../
o
1 2.0
(p<O.O1)
• MAX
o SMAX
10.7
i
13.3
14.7
16.0
17.3
PAOz(kPa)
Fig. 6 Relationship between arterial oxygen tensions (PaO2) and
alveolar (PAO2) during the two exercise tests. For other definitions
See Table 1
VE/V02
45.
O
40
35
30
• .;.o oo; o r=53
25
• MAX
o SMAX
20
6.0
615
" ' , , . ~ e
_~
_
•• •00
0
• MAX
o SMAX
o
O ~
25
~ 2.7
1 3m3
PaOz (kPa)
•
(p~O.01)
, 20
l I .~
10.7
/
r = -.79
30
20
13.3
40
35
B
0
71s 810 81s 91o
[K +] (mmol.Iq)
Fig. ? Relationship between ventilatory equivalent for oxygen
([2E/1/O2) and arterial potassium concentration [K +] recorded during the final 15 s of both M A X and SMAX exercise conditions. For
definitions See Table 1 and Fig 5
1988) and differences among subjects in their ventilatory response to a given level of metabolic stimuli
(Cerretelli and di Prampero 1987; Johnson et al. 1992).
Recently, it has been suggested that a mechanical limitation to l)'~ is the primary reason for a constraint of
1410
1417
4.0
4.7
5.3
6.0
6.7
73
PaCO2 (kPa)
the hyperpnoeic response during heavy and maximal
treadmill exercise (Johnson et al. 1992). These authors
have shown that inducing either hypercapnia or greater
hypoxaemia during maximal exercise failed to increase
I/E, inspiratory or expiratory pressure in the majority
of their subjects, although several individuals did show
a moderate elevation in 1/~. In contrast, McParland et
al. (1991) have demonstrated an elevation in I?E during
maximal exercise with increased dead space loading.
The results of the present study indicated that the
subjects operated to increase I?E from the MAX to the
SMAX exercise intensity by augmenting respiratory
frequency (fR) and slightly decreasing tidal volume (Vr)
(Table 2). The potential to increase 17E further, therefore, existed during all exercise intensities, yet this option was not adopted, at the expense of arterial blood
gas homeostasis. Although a mechanical limitation to
VE was shown to be absent among the subjects in this
study, the results do not explain why a greater ventilatory response was not achieved during the MAX exercise intensity. Such an increase in I) E (for example, to
the same level as that found during the SMAX test)
would presumably have attenuated the degree of hypoxaemia.
The respiratory flow-volume analyses in the present
study failed to show any clear association between the
levels of SaO 2 o r PaO2, and the fraction used of the
potential flow-volume area, forced vital capacity (FVC)
or PIFR and PEFR. Although there was a range in the
fraction of FVC chosen as VT during exercise (MAX,
42%-70% and SMAX, 41%-60%), typically this was
substantially less than 70% of FVC and consistent with
previous reports (Whipp and Pardy 1986). Similarly,
analysis of the final 30 s of exercise flow-volume loops
revealed a range in flow rates (as a fraction of the
maximum possible generated voluntarily at a given
lung volume) for both the inspiratory (MAX,
63% 86% and SMAX, 71%-96%) and expiratory manoeuvres (MAX, 54%-81% and SMAX, 55%-89%).
This simplistic approach may be somewhat misleading
since most athletes did approach their maximal ability
to generate flow throughout at least part of their
123
expiratory effort (Fig. 3). It should be pointed out
however, that these high flow rates were only seen
during the final minute or so of the exercise tests (for
example see Fig. 4). The subjects tested by Johnson et
al. (1992) were often found to exhibit a mechanical
limitation of lung and respiratory muscle for producing
alveolar ventilation. However, we did not find that
those of our subjects who suffered the greatest hypoxaemia also demonstrated the greatest expiratory
flow limitation. Two examples of individuals who displayed variable levels of expiratory and inspiratory
flow limitation are presented in Figure 4, together with
indices of blood gas status during exercise. Despite the
potential for loss of resolution by visually estimating
the mean flow-volume loop over each collection period,
it is clear that during the MAX test at least, only subject
A was encroaching upon his voluntary capacity for
flow generation at about mid-tidal volume. Each minute of exercise also resulted in a reduced 'reserve' for
flow generation in both subjects. Thus, differences in
blood gas status during exercise would not appear to be
explained on the basis of possible mechanical constraints only since subject A had a lower PaCO2.
Research concerning the possibility of respiratory
muscle fatigue during exercise has suggested that inspiratory muscle fatigue either does not occur or has no
functional significance during intense exercise of short
duration (Bye et al. 1983; Coast et al. 1990; Gallagher
and Younes 1989; Johnson et al. 1992; Levine and
Henson 1988). This study provided support for this
opinion during events of 4 rain or less.
There have been differing thoughts concerning the
potential for training-induced adaptations in peripheral chemosensitivity (Cerretelli and di Prampero
1987; Johnson et al. 1992; Saunders et al. 1976; Scoggin
et al. 1978). Of particular significance in this study was
the range in P~CO2 among the subjects during exercise
[MAX, 4.8-6.9kPa (35.9-51.7mmHg) and SMAX,
4.2-6.5 kPa (31.7-48.6 mmHg)]. The consistency within subjects to regulate to a particular individual PaCOz
in the two exercise intensities (while PaO2 and SaOz
varied) was apparent (Fig. 1). The reason for some
subjects tolerating high P,CO2 during exercise (particularly MAX intensity) is not known). When the data
for both conditions were combined (n -- 20), there were
significant correlations between P~CO2 (r = 0.79) and
PaO2 (r = 0.71) levels and I)z/l)'02 during exercise (Fig.
5A, B). Thus, greater ventilatory responses were associated with lower P,CO2 and higher PaO2. Although
these significant relationships (and those of Fig.6) are
clearly shown in the present study, there are inconsistencies in the literature. For example, Powers et al.
(1992) have not found any significant relationship between PAO2 and P~O2 in their subjects exercising to
maximal levels on a cycle ergometer. Furthermore, they
have reported no relationship between 12E/1202 and
PaO2 or PaCO2 at maximal exercise intensity, suggesting that reduced 12E/1202 played an insignificant role
in the development of EIH. In support of the present
findings, however, Miyachi and Tabata (1992) have
obtained similar results to those of the present study
such as a significant correlation between l)'z/l)'02 and
SaO2 of 0.74 (0.87 in the present study). They have
concluded that the relatively low ventilatory response
may account for about 50% of the arterial 02 desaturation in exercise. It is not known why these differences
have arisen. One possibility may be the relative homogeneity of the subjects with respect to their training in
the study of the Powers et al. (1992) (compared to the
other two groups) which decreased the spread for the
variables assessed.
One possibility for differences in the ventilatory response among individuals in this study could be related
to differences in responses to the same level of biochemical stimulation and/or variations in the concentrations of chemoreceptor stimulants during exercise
(Cerretelli and di Prampero 1987; Johnson et al. 1992;
Miyachi and Tabata 1992). For example, in the study
by Miyachi and Tabata (1992) a significant positive
relationship has been found between S~O2 during
maximal exercise and the subjects' ventilatory response
to hypercapnia at rest. The moderate association
(r = 0.53) found between [K +] and (/z/l/02 during the
two exercise conditions (Fig. 7) shows at least one
possible explanation for the greater ventilatory drive in
some subjects. It is not known why [K ÷] would vary
among subjects during exercise to the extent found in
this study [and independently of absolute exercise
intensity and estimated muscle mass (estimated as
a constant fraction of body mass; kg)], however, previous reports have shown Na+-K ÷ pump activity to be
influenced by training (McKenna et al. 1993).
When compared to the MAX intensity, the elevated
12~ during the SMAX intensity was likely to have been
in response to the influence of more extreme levels of
stimulants such as K + and catecholamines (Table 4)
during the supramaximal exercise test which would
undoubtedly have exacerbated ventilatory drive (Wasserman et al. 1986). Greater afferent stimuli from the
exercising muscles generating larger forces could also
have contributed to the enhanced ventilatory drive
(McClosky and Mitchell 1972). Since l/Oz did not
change from MAX to SMAX, the COz production
must have also been similar. The increased elimination
of CO2 during the SMAX exercise, with the same
measured PaCO2, may have been due to a reduced time
for CO2 to diffuse completely into the fluid compartments of the body.
The significant lowering of PaO2 and SaO2 during
the experiments would seem to suggest the exercise
protocol may also be important, in particular the intensity and duration of the test. Early phases of the exercise period have been shown to be associated with
accumulation of body CO2 stores which are far greater
than for 02 (Wasserman et al. 1986). Thus, ventilatory
drive has been found to be relatively sluggish at this
time (Fig. 4), in spite of rapidly lowered PaO2 and PAO2
(Hopkins and McKenzie 1989). Since it has previously
been shown that 12z lags behind rather than leads
124
I/CO2 in response to exercise (Wasserman et al. 1986),
it would be expected that CO2 stores would stimulate
1)E earlier (faster VE kinetics) during SMAX exercise
(Ward et al. 1983). In this way it may have been too late
to catch up with the required ventilatory response
during the MAX exercise intensity (which was stopped
at 4 min).
In summary, this study demonstrated that high!y
trained and motivated athletes were able to increase VE
when power output was increased from maximal to
supramaximal levels of intensity. That is, there does not
appear to have been a mechanical constraint to ventilation at maximal exercise levels in the subjects participating in this study. As a consequence of this elevation in
VE arterial blood gas status was affected during the
more intense exercise test. Specifically, less extreme
levels of arterial hypoxaemia were found during the
more severe exercise intensity since the majority of
subjects increased 1/E. The precise mechanism operating to constrain the ventilatory effort during the lower
exercise intensity is still unknown. However, the ventilatory response during severe exercise is the net result
of a combination of many stimuli, including both neural and humoral factors. It has been suggested that
these must be integrated to produce the most effective
possible mechanical response (Johnson et al. 1992;
Wasserman et al. 1986), presumably with the lowest
energy cost, in the face of dramatic blood biochemistry
shifts.
Acknowledgements The authors wish to thank Dr G. McLeay, S.
Woolford, P. Bourdon and T. Stanef for their technical assistance.
This study was supported by grants from the AustraIian Sports
Commission and Australian Research Council.
References
Aaron EA, Johnson BD, Pegelow D, Dempsey JA (1992) The oxygen
cost of exercise hyperpnea: implications for performance. J Appl
Physiol 72:1818 1825
Bye PTP, Farkas GA, Roussos C (1983) Respiratory factors limiting
exercise. Annu Rev Physiol 45:439-451
Cerretelli P, Di Prampero PE (1987) Gas exchange in exercise. In:
Farhi LE, Tenney SM (eds, vol IV) Handbook of physiology. The
respiratory system vol IV, chapt. 16. Washington DC: American
Physiological Society, pp 297 339
Coast JR, Clifford PS, Henrich TW, Stray-Gunderson J, Johnson
RL, (1990) Maximal inspiratory pressure following maximal exercise in trained and untrained subjects. Med Sci Sports Exerc
22:811-815
Craig NP, Walsh CBV (1989) Cycling. In: Draper J, Telford R
(eds) Sports specific guidelines for the physiological assessment of
the elite athlete. Australian Sports Commission Canberra, pp
21-57
Dempsey JA, Hanson PG, Henderson KS (1984) Exercise induced
arterial hypoxemia in healthy human subjects at sea level.
J Physiol (Lond) 355:161 175
Dempsey JA, Henke KG, Aaron EA (1989) Pulmonary physiology:
feedback and feedforward mechanisms. In: Skinner JS (ed) Future directions in exercise and sport science research. Human
Kinetics, Champaign, Ill., pp 313 327
Duncan MW, Compton PJ, Lazarus L, Smythe GA (1988) Measurement of norepinephrine and 3,4-dihydroxyphenyglycol in urine
and plasma for the diagnosis of pheochromocytoma. N Engl
J Med 319:136-142
Gale GE, Torre-Bueno JR, Moon RE, Saltzman HA, Wagner PD
(1985) Ventilation-perfusion inequality in normal humans during
exercise at sea level and simulated altitude. J Appl Physiol
58:978-988
Gallagher CG, Younes M (1989) Effect of pressure assist on ventilation and respiratory mechanics in heavy exercise. J Appl Physiol
66:1824-1837
Holgram A, Linderholm H (1958) Oxygen and carbon dioxide
tensions of arterial blood during heavy and exhaustive exercise.
Acta Physiol Scand 44:203-215
Hopkins SR, McKenzie DC (1989) Hypoxic ventilatory response
and arterial desaturation during heavy work. J Appl Physiol
67:1119-1124
Johnson BD, Saupe KW, Dempsey JA (1992) Mechanical constraints on exercise hyperpnea in endurance athletes. J Appl
Physiol 73:874-886
Kelman GR, Nunn JF (1966) Nomograms for correction of blood
PO2, PCO2, pH and base excess for time and temperature.
J Appl Physiol 21:1484 1490
Levine S, Henson D (1988) Low-frequency diaphragmatic fatigue in
spontaneously breathing humans. J Appl Physiol 64:672-680
McClosky DI, Mitchell JH (1972) Reflex cardiovascular and respiratory responses originating in exercising muscle. J Physiol
(Lond) 224:173 186
McKenna M J, Schmidt TA, Hargreaves M, Cameron L, Skinner SL,
Kjeldsen K (1993) Sprint Training increases human skeletal
muscle Na+/K + ATPase concentration and improves K + regulation. J Appl Physiol 75:173 180
McParland C, Mink J, Gallagher CG (1991) Respiratory adptations
to dead space loading during maximal incremental exercise.
J Appl Physiol 70:55-62
Miyachi M, Tabata I (1992) Relationship between arterial oxygen
desaturation and ventilation during maximal exercise. J Appl
Physiol 73:2588 2591
Naeije R, Melot C, Niset G, Delcroix M, Wagner PD (1993) Mechanisms of improved arterial oxygenation after peripheral
chemoreceptor stimulation during hypoxic exercise. J Appl
Physiol 74:1666-1671
Norton LH, Squires B, Craig NP, McLeay G, McGrath P, Norton
KI (1992) Accuracy of pulse oximetry during exercise stress
testing. Int J Sports Med 13:523-527
Olds TS, Norton KI, Craig NP (1993) Mathematical model of
cycling performance. J Appl Physiol 75:730 737
Powers SK, Dodd S, Freeman J, Ayers GD, Samson H, McKnight
T (1989a) Accuracy of pulse oximetry to estimate HbO2 fraction
of total Hb during exercise. J Appl Physiol 67:300-304
Powers SK, Dodd S, Lawler J, Landry G, Kirtley M, McKnight T,
Grinton S (1988) Incidence of exercise-induced hypoxemia in
elite endurance athletes at sea level. Eur J Appl Physiol
58:298-302
Powers SK, Dodd S, Woodyard J, Beadle RE, Church G (1984)
Hemoglobin saturation during incremental arm and leg exercise.
Br J Sports Med 18:212-216
Powers SK, Lawler J, Dempsey JA, Dodd S, Landry G (1989b)
Effects of incomplete pulmonary gas exchange on ~'O2 . . . .
J Appl Physiol 66:2491-2495
Powers SK, Martin D, Cicale M, Collop N, Huang D, Criswell
D (1992) Exercise-induced hypoxemia in athletes: role of inadequate hyperventilation. Eur J Appl Physiol 65:37-42
Powers SK, Williams J (1987) Exercise-induced hypoxemia in highly
trained athletes. Sports Med 4:46 53
Rowell LB, Taylor HL, Wang Y, Carlson WS (1964) Saturation of
arterial blood with oxygen during maximal exercise. J Appl
Physiol 19:284-286
Saunders NA, Leeder SR, Rebuck AS (1976) Ventilatory response to
carbon dioxide in young athletes. Am Rev Respir Dis
113:497-502
125
Schaffartzik W, Poole DC, Derion T, Tsukimoto K, Hogan MS,
Arcas JP, Bebout DE, Wagner PD (1992) VA/Q distribution
during heavy exercise and recovery in humans: implications for
pulmonary edema. J Appl Physiol 72:1657-1667
Scoggin CH, Doekel RD, Kryger MH, Zwillich CW, Weil JV (1978)
Familial aspects of decreased hypoxic drive in endurance athletes. J Appl Physiol 44:464-468
Thoden JS (1991) Testing aerobic power. In: McDougall JD, Wenger HA, Green HJ (eds) Physiological testing of the high-performance athlete, 2nd ed. Human Kinetics, Champaign, Ill, p 156
Torre-Bueno JR, Wagner PD, Saltman HA, Gale GE (1985) Diffusion limitation in normal humans exercising at sea level and
simulated altitude. J Appl Physiol 58:989-995
Ward SA, Whipp BJ, Koyal S, Wasserman K (1983) Influence of
body CO2 stores on ventilatory dynamics during exercise. J Appl
Physiol 55:742-749
Warren GL, Kureton KJ, Middendorf WF, Ray CA, Warren
JA (1991) Red blood cell pulmonary capillary transit time
during exercise in athletes. Med Sci Sports Exerc 23:
1353-1361
Wasserman K, Whipp BJ, Casaburi R (1986) Respiratory control
during exercise. In: Geiger SR (ed, vol. II) Handbook of physiology. The respiratory system, vol II, chapt 17. American Physiological Society, Washington, DC, pp 595-619
Whipp BJ, Pardy RL (1986) Breathing during exercise. In: Macklem
PT, Mead J (eds, vol. III) In: Handbook of physiology. The
respiratory system, vol III, chapt 34. American Physiological
Society, Washington, DC, pp 605-629
Williams JH, Powers SK, Stuart MK (1986) Hemoglobin desaturation in highly trained athletes during heavy exercise. Med Sci
Sports Exerc 18:168-173