The Role of Cadence on the V˙O2 Slow Component in Cycling and Running in Triathletes

Abstract

The purpose of this study was to compare the effect of two different types of cyclic severe exercise (running and cycling) on the VO2 slow component. Moreover we examined the influence of cadence of exercise (freely chosen [FF] vs. low frequency [LF]) on the hypothesis that: 1) a stride frequency lower than optimal and 2) a pedalling frequency lower than FF one could induce a larger and/or lower VO2 slow component. Eight triathletes ran and cycled to exhaustion at a work-rate corresponding to the lactate threshold + 50% of the difference between the work-rate associated with VO2max and the lactate threshold (delta 50) at a freely chosen (FF) and low frequency (LF: - 10 % of FF). The time to exhaustion was not significantly different for both types of exercises and both cadences (13 min 39 s, 15 min 43 s, 13 min 32 s, 15 min 05 s for running at FF and LF and cycling at FF and LF, respectively). The amplitude of the VO2 slow component (i.e. difference between VO2 at the last and the 3rd min of the exercise) was significantly smaller during running compared with cycling, but there was no effect of cadence. Consequently, there was no relationship between the magnitude of the VO2 slow component and the time to fatigue for a severe exercise (r = 0.20, p = 0.27). However, time to fatigue was inversely correlated with the blood lactate concentration for both modes of exercise and both cadences (r = - 0.42, p = 0.01). In summary, these data demonstrate that: 1) in subjects well trained for both cycling and running, the amplitude of the VO2 slow component at fatigue was larger in cycling and that it was not significantly influenced by cadence; 2) the VO2 slow component was not correlated with the time to fatigue. If the nature of the linkage between the VO2 slow component and the fatigue process remains unclear, the type of contraction regimen depending on exercise biomechanic characteristics seems to be determinant in the VO2 slow component phenomenon for a same level of training.

Full text

Physiology and Biochemistry
Billat VL, Mille-Hamard L, Petit B, Koralsztein JP. The Role of
Cadence on the V
˙O2Slow Component in Cycling and Running
in Triathletes. Int J Sports Med 1999; 20: 429 –437
Accepted after revision: February 24, 1999
■■■■ The purpose of this study was to compare the effect of
two different types of cyclic severe exercise (running and cyc-
ling) on the V
˙O2slow component. Moreover we examined the in-
fluence of cadence of exercise (freely chosen [FF] vs. low fre-
quency [LF]) on the hypothesis that: 1) a stride frequency lower
than optimal and 2) a pedalling frequency lower than FF one
could induce a larger and/or lower V
˙O2slow component. Eight
triathletes ran and cycled to exhaustion at a work-rate cor-
responding to the lactate threshold + 50 % of the difference be-
tween the work-rate associated with V
˙O2max and the lactate
threshold (∆50) at a freely chosen (FF) and low frequency (LF:
– 10% of FF). The time to exhaustion was not significantly differ-
ent for both types of exercises and both cadences (13 min 39s,
15 min 43 s, 13 min 32 s, 15 min 05 s for running at FF and LF and
cycling at FF and LF, respectively). The amplitude of the V
˙O2slow
component (i.e. difference between V
˙O2at the last and the 3rd
min of the exercise) was significantly smaller during running
compared with cycling, but there was no effect of cadence. Con-
sequently, there was no relationship between the magnitude of
the V
˙O2slow component and the time to fatigue for a severe ex-
ercise (r = 0.20, p = 0.27). However, time to fatigue was inversely
correlated with the blood lactate concentration for both modes
of exercise and both cadences (r = – 0.42, p = 0.01). In summary,
these data demonstrate that: 1) in subjects well trained for both
cycling and running, the amplitude of the V
˙O2slow component
at fatigue was larger in cycling and that it was not significantly
influenced by cadence; 2) the V
˙O2slow component was not cor-
related with the time to fatigue. If the nature of the linkage be-
tween the V
˙O2slow component and the fatigue process remains
unclear, the type of contraction regimen depending on exercise
biomechanic characteristics seems to be determinant in the V
˙O2
slow component phenomenon for a same level of training.
■Key words: V
˙O2slow component, fatigue, running, cycling,
cadence, triathletes.
Introduction
At work rates that elicit a lactic acidosis, an additional slow
phase of V
˙O2(V
˙O2slow component) is superimposed upon
the underlying V
˙O2kinetics. Indeed, previous studies of cy-
cling exercise reported that oxygen uptake can attain a steady
state above the lactate threshold, but only in the work-rate
range where lactate remains constant. Moreover, at higher
work rates i.e. above what has been termed “critical power”
[27], oxygen uptake continues to rise until the end of the test
or until exhaustion and will eventually drive V
˙O2to the maxi-
mal oxygen uptake (V
˙O2max) [32,35,42].
Although the mechanism underlying the continuous rise in
V
˙O2during a suprathreshold exercise remains poorly under-
stood, Poole et al., [34] demonstrated that exercising muscle
was the predominant site of the V
˙O2slow component. How-
ever, few studies [8] have investigated the influence of cyclic
exercise modality (running vs. cycling) on the V
˙O2slow com-
ponent. Indeed, the difference between the muscle tension
and contraction characteristics of running and cycling may
provide important insights into the phenomenon of the V
˙O2
slow component whose ultimate causes remain still unsettled.
Therefore, the biomechanical difference between running and
cycling, which imposes different types of muscular contrac-
tion, might imply that the muscle metabolism will have differ-
ent effects considering the modality of exercise.
Given that endurance training decreases the magnitude of the
V
˙O2slow component [19], influence of training was controlled
in a recent study performed with triathletes well trained both
in cycling and running [8]. This study reported a larger V
˙O2
slow component in cycling (ten fold) than in running at the
same relative exerciseintensity (between the lactate threshold
and V
˙O2max of each exercise i.e. ∆50). Whether the type of
cyclic exercise influences the magnitude of the V
˙O2slow com-
ponent has not been well documented or understood.
Indeed, the two types of exercise are quite different from a bio-
mechanical standpoint. Mechanical efficiency of cycling i.e.
20– 25 % [43] is much less than that of running at 45 – 70%
The Role of Cadence on the V
˙O2Slow Component in Cycling
and Running in Triathletes
V. L. Billat, L. Mille-Hamard, B. Petit, J. P. Koralsztein
Faculté des Sciences du Sport, Université Lille 2, Centre de Médecine du Sport CCAS, Paris, France
Int J Sports Med 1999; 20: 429– 437
© Georg Thieme Verlag Stuttgart · New York
ISSN 0172-4622
429

[11]. The mechanical work observed in running is due to the
stretch-shortening movements in running which can be at-
tributed to the elastic behaviour of the muscle-tendon com-
plex during contact with the ground [10]. Hence, muscle ten-
sion during running is applied for a shorter duration and at
higher levels than in cycling which has an isometric contrac-
tion phase. In fact, Patterson and Moreno [31] reported that,
as the frequency increased from 60 to 120 rpm, a relatively
large percentage of the average crank force was “isometric-
like” and produced no external work. This difference in muscle
contraction regimen leads to the fact that other than in run-
ning, where the freelychosen stride length allows for the most
economical run [12], this is not the case in cycling [25] and
thus could influence the amplitude of the V
˙O2slow compo-
nent. Indeed, the fact that in running and not in cycling, the
preferred cadence at a particular work rate is the most
economical (i.e. lowest aerobic demand) could have a potential
effect on the V
˙O2slow component.
Therefore, the purpose of this study was to examine the possi-
ble influence of cadence on the difference in the V
˙O2slow
component between two types of cyclic exercises (running vs.
cycling). We postulated that lowering the cadence in cycling
and running could decrease and increase respectively the V
˙O2
slow component. Moreover, we wanted to check the relation-
ship between the magnitude of the V
˙O2slow component and
the time to exhaustion during supra critical-velocity exercise
to allow us to better appreciate the influence of the magnitude
and/or kinetics of the V
˙O2slow component on exercise toler-
ance [32,35].
Methods
Subjects
Eight well-trained triathletes gave their informed consent and
volunteered to participate in this study, which was approved
by the Paris Ethical Committee. The physical characteristics of
the subjects are presented in Table 1. All subjects were highly
motivated and familiar with treadmill running, ergometer cy-
cling and with the sensation and symptoms of fatigue during
heavy exhaustive cycling and running exercise.
Material
Running tests were performed on a motorized treadmill (Gym-
rol 1800, Techmachine, Saint Etienne, France) kept at a 0%
slope for all of the tests, the speed being controlled with a re-
solution of 0.5 m × h–1 (controller provided by the “Centre
d'Enseignement et de Développement pourle Montage en Sur-
face”, Université Joseph Fourier, Grenoble, France). Cycling
tests were carried out on a ERG 600 Bösch electronically-
braked cycle ergometer (Berlin, Germany). Respiratory and
pulmonary gas exchange variables were measured using a
MedGraphics CPMax cart (Medical Graphics, St. Paul, MN,
USA) which was calibrated prior to each test according to the
manufacturer's instructions. Breath-by-breath data were aver-
aged every 15-s (with discrete time-bins). ECG was monitored
from a three lead configuration (Jaeger cardioscope) and the
output signal was fed to the CPX Medical Graphics system for
computation of HR. V
˙O2max was the highest 30-s V
˙O2reached
at the end of an incremental test. The velocity (in running) or
power output (in cycling) when the subject attained V
˙O2max
were termed vV
˙O2max and pV
˙O2max, respectively. Blood sam-
ples were analyzed for blood lactate concentration (Lab) using
a purpose-built analyzer YSI 27 (Yellow Springs Instruments,
Ohio, USA).
Preliminary measurement
Each subject performed tests at the same time of day in a cli-
mate-controlled laboratory (21 to 228C). Subjects were in-
structed not to train hard, or ingest food and beverages con-
taining caffeine three days before testing. Each subject under-
took two preliminary incremental tests, one on the treadmill
and one on the cyclo-ergometer, to determine 1) V
˙O2max, 2)
the work-rate associated with V
˙O2max and 3) the fraction of
V
˙O2max at which the lactate threshold occurred. These two in-
cremental tests (running and cycling) were performed three
days apart and in a randomized order. Subjects performed a
continuous incremental test (3-min stages) to exhaustion.
Duration and workload increments were standardized for run-
ning and cycling as follows. The workload increments were es-
timated to demand a V
˙O2response equal to 2× resting V
˙O2(2
METS i.e. 2 × 3.5 ml × kg–1 × min–1). Each work increment,
Table 1 Physical characteristics, triathlon experience, and training regimen data of the 8 triathletes
Subjects Age Body mass Height Triathlon
Experience
Triathlon
Competitions
Mean Training Distance (km × wk-1)
(yr) (kg) (cm) (yr) (number) Run Bike Swim
1 33 75 175 5 18 60 170 8
2 29 60 166 5 19 40 150 8
3 23 76 170 3 12 70 150 6
4 25 72 175 5 16 40 170 6
5 25 66 177 4 13 50 200 5
6 30 68 170 6 9 60 220 7
7 38 80 185 6 10 60 200 6
8 24 82 175 4 9 50 220 8
Mean 28 72 177 5 13 54 185 7
SD 5 7 5 1 4 11 29 1
Int J Sports Med 1999; 20 V. L. Billat et al.430

ranged between 35 to 50 watts for cycling depending on the
weight of the subject (on basis of 12 ml O2/min/watt according
to recommendations by Astrand and Rodahl [3]). For example,
for a subject of 60 kg the work load increment was 35 watts
([60 (kg) × 7 (ml × kg–1 × min–1)]/12 (ml/min/watt)). For run-
ning, the speed was increased by 2 km × h–1 (33.3 m × min–1)
at each stage, except for the last stage when the increment
was only 1 km × h–1 (16.7 m × min–1). The initial work was set
at 70 to 100 watts for cycling (double that of the workload in-
crements) and at 10 km × h–1 for running.
The power output and velocity associated with V
˙O2max
(pV
˙O2max and vV
˙O2max) were defined for running and cycling
as the minimal work-load at which V
˙O2max occurred [8]. All
subjects were verbally encouraged and gave a maximum ef-
fort.
Blood samples were obtained from the fingertip at the end of
each 3-min stage, immediately after the end of the exercise
test and then 8 minutes into the recovery period.
In this study, the lactate threshold (LT) was defined as the
V
˙O2corresponding to the starting point of an accelerated lac-
tate accumulation (around 4 mmol × l–1) and expressed in %
V
˙O2max [4]. Although using a ramp like increase in work rate
cannot allow precise determination of this blood lactate
threshold, the incremental test gave us a useful index to de-
lineate the domain at which blood lactate starts to dramatical-
ly accumulate in the blood.
To be sure that the triathletes would exercise in the range of
severe-intensity exercise (where neither V
˙O2nor blood lactate
can be stabilized and at which a consequent V
˙O2slow compo-
nent appeared) [20], the work rate for both modes of exercise
was set at 50% of the work rate differencebetween those at the
lactate threshold and at V
˙O2max determined in the incremen-
tal test exercise described above (Wr ∆50). This suprathres-
hold exercise was designated to be a severe intensity all-out
exercise bout that produced fatigue between 10 and 15 min-
utes.
Experimental design and procotols
The subjects ran and cycled to exhaustion at Wr ∆50 at a freely
chosen (FF) and low frequency (LF) cadence. Each of these tests
were separated by one week. After a fifteen minute warm-up
period at 50% of the work rate associated with V
˙O2max
(WrV
˙O2max) (which was below the lactate threshold for all
subjects), the work-rate was increased to the suprathreshold
exercise within 20 s. All subjects were given verbal encourage-
ment throughout each trial. For running, time to fatigue at su-
prathreshold exercise was the time at which the subject's feet
left the treadmill as he placed his hands on the guardrails, and,
for cycling, the time at which the subject was no longer able to
maintain the required power-output. This time was recorded
to the nearest second.
Blood samples were obtained from the fingertip before the
warm-up run, during the last 30 seconds of the warm-up, ev-
ery 3 min of all-out suprathreshold exercise and immediately
after the end of the exercise test and then 8 minutes into the
recovery period.
Data analysis
Statistical analysis was performed using t-tests for paired
comparisons to determine whether V
˙O2max, blood lactate
concentration, respiratory exchange ratio, heart rate, blood
lactate threshold (in % V
˙O2max) and times to exhaustion were
different in cycling and running tests. The amplitude of the
V
˙O2slow component was computed as the difference between
V
˙O2at the last and the 3rd min of the exercise. A two-way AN-
OVA measure tested the overall effect of exercise and frequen-
cy on the V
˙O2slow component. Subsequently, the differences
were located by Scheffes' post hoc test. Results are presented
as means ± standard deviation (SD). Correlations betweenbio-
energetic characteristics were determined using Pearson cor-
relation coefficient. The parameters of the linear distance-time
model were derived using linear regression (Statview 4.5,
Berkeley, CA). The level of statistical significance was set at
p = 0.05.
Results
Incremental tests
Table 2shows the maximal V
˙O2, heart rate and blood lactate,
reached at the end of the incremental tests in cycling and run-
ning. There was no significant difference between triathletes'
V
˙O2max during cycling and running. However, the blood lac-
tate and heart rate were significantly higher in cycling than in
running. The lactate threshold was higher in running than in
cycling relative to the work-rate associated with V
˙O2max
(84.9 ± 0.6 vs. 72.5 ± 4 % WrV
˙O2max). Therefore, when setting
the work-rate for both modes of exercise ∆50 work rate exer-
cise gave 86.3 ± 2 and 92.5 ± 0.7 % WrV
˙O2max for cycling and
running, respectively.
The lactate threshold was determined between the range of
3 –5 mmol × l–1,in accordance with Aunola and Rusko [4]. This
point was closer in definition to the onset of blood lactate ac-
cumulation defined by Sjödin and Jacobs [36], than to the lac-
tate threshold of Farrel et al. [17].
Table 2 Maximal values of metabolic and cardiorespiratory param-
eters (mean ± SD) during the incremental tests for cycling and running
Variables Cycle
Ergometry
Treadmill
Running
p Values
V
˙O2max (l × min-1) 4.33 ± 0.66 4.19 ± 0.68 0.56
V
˙O2max (ml × kg-1 × min-1) 60.2 ± 5.5 59.8 ± 4.7 0.82
RER 1.17 ± 0.04 1.13 ± 0.04 0.01*
HR (beat × min-1) 183 ± 10 189 ± 11 0.65
[La] (mmol × l-1) 8.9± 1.3 7.7 ± 1.0 0.004*
Work rate at V
˙O2max 350 ± 38.5 19.9 ± 0.8 –
(Watt for cycling and
km × h-1 for running)
Lactate Threshold
(%WrV
˙O2max)
72.5 ± 4.0 84.9 ± 0.6 0.001*
* Significant difference, p < 0.05 (RER, Respiratory Exchange Ratio; HR, Heart
Rate)
Effect of Cadence on V
˙O2Slow Component Int J Sports Med 1999; 20 431

The selected warm-up period was below the lactate threshold.
The fifteen minute warm-up period was performed at 53 ± 2 %
and 52 ± 2 % of V
˙O2max for running and cycling, respectively
(50% of WrV
˙O2max).
Constant work rate tests
Table 3shows, the contribution of the slowcomponent to V
˙O2,
the cycling and running maximal V
˙O2heart rate, and blood lac-
tate reached at the end of the all-out severe work tests at free
and low cadence. The mean cadence for exercise freely chosen
by the subjects was higher in running than in cycling (Table 3).
Time to fatigue
The speed and the work rate set in all-out severe exercise were
18.4 ± 1.1 km × h–1 and 302 ± 43 watts, respectively. The time to
fatigue for each exercise (running and cycling) at both caden-
ces (spontaneous and low) were not significantly different
(around 14 minutes, Table 3). Moreover, the time to fatigue be-
tween sports was not significantly different for the free or low
cadence (r = 0.46, p = 0.26 and r = 0.58, p = 0.14 for FF and LF
between cycling and running). However, there was no correla-
tion between cadence (free of low frequency) and time to fa-
tigue in cycling (r = 0.60, p = 0.1).
Table 3 Delay before exhaustion (tlim), maximal values of metabolic and cardiorespiratory parameters (mean ± SD) during the constant work-
load tests for cycling and running in the two cadences: Free frequency (FF)and Low Frequency (LF)
Variables Treadmill Running Cycle Ergometry p values
A sports effect
B cadence effect
A × B interaction
FF LF FF LF
Cadence (rotation or stride per min) 92.1 ± 2.5 83.0 ± 2.3 82.5 ± 10.0 74.3 ± 8.9 0.0008*
0.001*
0.86
tlim (min, sec) 13 min 39 s
± 3 min 07 s
15 min 43 s
± 4 min 11 s
13 min 32 s
± 4 min 52 s
15 min 05 s
± 4 min 43 s
0.80
0.24
0.87
work rate or speed (Watt or km× h-1) 18.4 ± 1.1 18.4 ± 1.1 254.5 ± 37 254.5 ± 37 –
relative work rate (%WrV
˙O2max) 92.4 ± 0.3 92.4 ± 0.3 86.3 ± 1.9 86.3 ± 1.9 0.0001*
–
–
V
˙O2(l × min-1) 3.95 ± 0.92 4.03 ± 0.58 4.09 ± 0.58 4.13 ± 0.74 0.63
0.82
0.93
V
˙O2(ml × kg-1 × min-1) 54.3 ± 9.0 56.0 ± 7.8 56.5 ± 3.8 57.3 ± 8.5 0.53
0.64
0.86
relative V
˙O2(%V
˙O2max) 90.7 ± 11.1 94.0 ± 13.1 94.5 ± 9.8 95.8 ± 15.8 0.54
0.61
0.82
HR (beats × min-1) 185 ± 12 184 ± 12 182 ± 14 182 ± 16 0.54
0.90
0.88
[La] (mmol × l-1) 6.1 ± 2.5 6.4 ± 2.3 9.2 ± 0.7 7.5 ± 2.3 0.01*
0.38
0.24
RER 1.11 ± 0.07 1.06 ± 0.06 1.15 ± 0.04 1.12 ± 0.04 0.01*
0.03
0.58
V
˙E(l × min-1) 139.9 ± 28.8 135.0 ± 20.6 159.1 ± 28.1 159.5 ± 38.5 0.04*
0.83
0.80
∆[La] (6 – 3 min) (mmol × l–1) 1.5 ± 0.9 1.2 ± 0.7 1.9 ± 0.7 1.5 ± 0.7 0.25
0.16
0.87
∆V
˙O2(6 –3 min) ml × min-1) 81.4 ± 27.6 77.6 ± 39.5 212.8 ± 72.5 191.4 ± 108.5 0.0006*
0.694
0.782
RER, Respiratory Exchange Ratio; HR, Heart Rate; ∆V
˙O2(ml × min-1)) and ∆[La] difference of V
˙O2and lactatemia between the sixth and the third minute of exer-
cise. * Significantly different
Int J Sports Med 1999; 20 V. L. Billat et al.432

V
˙
O2slow component and maximal V
˙
O2attained in the severe
all-out exercise
As shown in Table 3, 1) the V
˙O2slow component was signifi-
cantly higher in cycling than in running, 2) it was not influ-
enced by the cadence (212.8± 73 and 191.4 ± 108ml of O2/min
in cycling at FF and LF vs. 81.4 ± 17.6 and 77.6 ± 39 ml of O2/min
in running, p = 0.0006). Thus there was no effect of cadence on
the V
˙O2slow component.
Moreover, in cycling there was no significant difference be-
tween the maximal value of V
˙O2at the end of the all-out severe
exercise and V
˙O2max obtained in the incremental test (Table
3). On the contrary, for running at free cadence, maximal V
˙O2
was significantly lower than V
˙O2max obtained in the incre-
mental test. However, by running at a reduced cadence as for
cycling, triathletes attained their V
˙O2max at the end of the
all-out severe exercise.
By examining the individual V
˙O2response in the all-out severe
exercise in running and cycling for free and low frequency, we
can observe that in each exercise (2 sports × 2 cadences) there
was no difference between the maximal value of V
˙O2attained
in the four ∆50 exercises and the V
˙O2max attained in the ramp
protocole (chi 2 = 7.5, p = 0.27). The average tendency, there-
fore, is to reach V
˙O2max in cycling, and that the reduction in
cadence in either sport (running or cycling) has no effect on
the V
˙O2achieved (chi 2 = 1.8, p = 0.41).
Blood lactate and the V
˙
O2slow component
Blood lactate level at the end of the exhaustive exercise was
not significantly different from that at the end of the incre-
mental test in all constant load tests but for cycling at low ca-
dence where blood lactate was significantly lower. In all-out
severe exercise, maximal blood lactate was higher in cycling.
However, for each sport there was no influence of cadence on
blood lactate (Table 3). Considering both cadences (free and
low), the amplitude of the V
˙O2slow component was signifi-
cantly correlated with the blood lactate accumulated in the
same time (between the 3rd and the 6th min) for cycling
(r = 0.50, p < 0.05, n = 16) but not for running (r = – 0.08,
p = 0.78, n = 16).
The V
˙
O2slow component and time to fatigue
No significant correlation was found between the amplitude of
the V
˙O2slow component and the duration tolerated in this su-
pra-threshold exercise (r = –0.20 for both running and cycling
and r = – 0.02 and – 0.37 for running and cycling considered
separately).
Blood lactate and time to fatigue
A significant inverse relationship was found between the mag-
nitude of blood lactate accumulated and time to fatigue in se-
vere exercise considering both modes of exercise and both ca-
dences (r = – 0.42, p = 0.01, n = 32). Examining each mode or
frequency only, there was a close significant relationship for
running at low frequency between lactate accumulation and
time to fatigue (r = – 0.46, p = 0.07, r = – 0.43, p = 0.07, respec-
tively n = 16).
The V
˙
O2slow component and cardio-ventilatory variables
Heart frequency was similar between modes of exercise and
cadence. However, end RER, and V
˙E values were significantly
higher in cycling than in running at cessation of all-out severe
exercise, cadence having no influence in each mode of exercise
(Table 3).
End V
˙E/V
˙O2was similar in running and cycling with no effect
of cadence in each mode of exercise (Table 3). Nevertheless,
the “drift” in V
˙E between the 3rd and 6th min was significantly
larger during cycling than during running (∆V
˙E = 43.6 ± 16.0 l/
min and 49.9 ± 18.1l/min in FF and LF in running vs.
18.4 ± 11.7 l/min and 16.0 ± 5.6 l/min in FF and LF for running,
p < 0.001).
However, the V
˙O2slow component was not correlated with the
increase of V
˙E (r = 0.34, p = 0.20 for running and r = 0.05,
p = 0.85 in cycling). This decrease of running economy (defined
as the V
˙O2at the same speed) due to an increase in V
˙O2of
+ 2.27 ml/min/kg at the third minute during LF, induced a
higher value of V
˙O2at the end of the constant severe exercise
with the triathletes reaching their V
˙O2max at the low stride-
frequency, unlike FF. This difference, however, was not signifi-
cant (t = – 1.6, p = 0.32). For cycling the low cadence did not
modify the economy of exercise since V
˙O2at the third minute
was the same at both exercise frequencies (t = – 0.62, p = 0.55).
Discussion
The purpose of this study was to examine 1) the influence of
cadence of exercise (spontaneous vs. low in running and cy-
cling) on the V
˙O2slow component during an exhaustive exer-
cise in triathletes equally trained in cycling and running, 2)
whether the magnitude of the V
˙O2slow component influenced
the duration of severe exercise (i.e. time limit: tlim).
This study showed that, 1) triathletes well trained in both run-
ning and cycling (V
˙O2max = 60 ml/min/kg for both exercises)
had a larger V
˙O2slow component in cycling and this difference
was not reduced by a lower cadence (– 10 % of the spontaneous
one), 2) the V
˙O2slow component (in cycling) was correlated
with blood lactate accumulation but was not correlated with
time to fatigue. Time to fatigue at this severe intensity was
only correlated with the blood lactate concentration.
The V
˙
O2slow component is larger in cycling than
in running and is not influenced by the cadence
The V
˙O2slow component is larger in cycling than in running
Most of the studies which have reported a V
˙O2slow compo-
nent were performed during cycling [20]. However, this pres-
ent study confirmed that for the same subject (triathlete), the
amplitude V
˙O2slow component at fatigue was significantly
greater in cycling than running. These results are in accord-
ance with those of a previous study also carried out with ten
other triathletes at a work rate corresponding to 90 % of work
rate (Wr) at V
˙O2max [8].
Considering the absence of the V
˙O2slow component in run-
ning despite continuous blood lactate accumulation, this pres-
ent study is in accordance with that of Nagle et al. [28] who
Effect of Cadence on V
˙O2Slow Component Int J Sports Med 1999; 20 433

underlined that during running “lactic acid is produced in
work requiring 65– 90% of V
˙O2max even when a reasonably
steady state of V
˙O2is attained” (duration 30 min at
86.4 ± 2.6 % of V
˙O2max). Costill [15] reported that three highly
trained distance runners (V
˙O2max > 70 ml/kg/min) were able
to sustain a V
˙O2steady-state at 90% of V
˙O2max for more than
25 minutes (10km performed on treadmill) with end blood
lactate of 5 mM. In the same manner, a recent study [7]
showed that thirteen high level runners (V
˙O2max =
74.9 ± 3.0 ml/min/kg) were able to maintain a V
˙O2steady-state
at 91% of V
˙O2max for 17 minutes with an end blood lactate of
6.5 ± 2.1 mM. They ran at 90% of the velocity associated with
V
˙O2max (vV
˙O2max ) a velocity well above their critical velocity
(86 ± 1.5% vV
˙O2max). This present study, performed with
triathletes, confirmed that the V
˙O2slow component was very
slight in running (25 ml/min) and was ten fold higher in cy-
cling (250 ml/min). In fact, 5 and 4 out of the 8 subjects
reached their V
˙O2max in cycling for FF and LF, respectively,
and only 1 and 3 out of the 8 subjects reached their V
˙O2max
in running for FF and LF, respectively.
Given that Poole et al. [33] demonstrated that the majority of
the V
˙O2slow component is attributable to factors within the
working limbs [20], and considering the absence of V
˙O2slow
component for severe running by highly trained distance run-
ners and triathletes, we hypothesized that not only the train-
ing and the relative intensity, but also the type of contraction
regimen could influence the V
˙O2slow component. The main
difference between these two types of exercise is from the bio-
mechanical standpoint: the mechanical efficiency of cycling is
much less than in running due to the fact that the contraction
regimen is quite different. A greater efficiency is seen in run-
ning due to the stored elastic energy contributing to the posi-
tive mechanical work [10,13]. On the contrary, cycling is char-
acterized by longer phases of isometric contraction [31]. This
could explain our results reporting a larger lactate production
above the lactate threshold at the same relative fraction of the
difference between the work-rate at V
˙O2max and the work rate
associated with the lactate threshold (∆50). This decrease of
muscle pH which displaces the equilibrium of the Lohman
reaction [21] accelerates the dephosphorylation of phospho-
creatine (PC) to creatine. The decrease of the intramuscular
pH due to lactate accumultion might be the signal “turn on”
of a slow and continuous increase of V
˙O2[24] and decrease in
maximal explosive power, due to a decreased (PC) as demon-
strated by Capelli et al. [9].
Moreover, it has been suggested that the V
˙O2slow component
may primarily be related to motor unit recruitment patterns
during exercise depending upon the contribution of less effi-
cient, fast twitch motor units [5,6]. The hypothesis that the
V
˙O2slow component arises from the recruitment of a fast
twitch fiber population with slow kinetics is consistent with
the notion that the V
˙O2kinetics are limited by fiber mitochon-
drial content [34]. Moreover, it has been shown that isolated
mitochondria from type II fibers exhibit a 18% lower P/O ratio
[45]. Similarly, Whipp [42] considers that a (if not the) major
contributor to the V
˙O2excess is likely to be the high energy
cost of contraction in the type II fibers recruited proportionally
more at high work rates and requiring a large high-energy
phosphate cost of force production.
To our knowledge,no data are available in the literature to pro-
vide an answer to the question if the difference in V
˙O2slow
component between running and cycling can be accounted
for by a difference in percentage of fast twitch fibres recruited
during these two types of exercises. However, comparing max-
imal treadmill and bicycle exercises, Okita et al. [30] showed,
that in cycling exercise, there was severe metabolic stress on
the quadriceps, leading to near complete depletion of PC and
to a severe decrease in pH. Based on the findings of depleted
PC, they concluded that all the available muscle fibers, includ-
ing slow and fast twitch fibers, had been recruited and ex-
hausted. In treadmill exercise this was not the case since a sig-
nificant amount of PC (about one third) still remained and pH
decrease was less severe both in the quadriceps and in the calf.
The V
˙O2slow component is not influenced by the cadence
for each cyclic exercise (running and cycling)
In cycling, the muscle mass activated is mainly the vastus late-
ralis and medialis [39]. However, for cyclists having a certain
pedalling skill, there is a positive utilization for knee flexors
(biceps femoris) to decrease peak pedal force and alleviate
muscle activity for knee extensors [39]. Takaishi et al. [39]
speculated that pedalling skills that decrease muscle stressin-
fluence the preferred cadence selection, contributing to re-
cruitment of ST muscle fibers with fatigue resistance and high
mechanical efficiency despite increased oxygen consumption
caused by increase repetitions of leg movement. Indeed, Pat-
terson and Moreno [31] demonstrated that when the pedalling
rate was increased from 60 to 120rpm, the resultant force on
the pedals averaged over a crank cycle was isometric-like and
produced no external work. Their results suggested that pedal-
ling at 90 rpm might minimize peripheralforces and therefore
peripheral muscle fatigue even though this rate might result in
higher oxygen uptake. Consequently, a high frequency of ped-
alling is less economical than a low frequency. For instance,
Gaesser [18] reported that the V
˙O2slow component was signif-
icantly higher when cycling at 100 rpm than at 50rpm. More-
over, Takaishi et al. [38] demonstrated that the optimal pedal-
ling rate estimated from neuromuscular fatigue in working
muscles was not coincident with the pedalling rate at which
the smallest V
˙O2was obtained, but with the preferred pedal-
ling rate of the subjects. They suggested that the reason that
cyclists preferred a higher pedalling rate was closely related
to the development of neuromuscular fatigue in the working
muscles. In the previous [8] and present study performed with
triathletes, the subjects were free to choose their most com-
fortable cadence and usually adopted a frequency around
85 rpm, which may have increased the magnitude of the slow
component. Therefore, we hypothesized that in cycling, lower-
ing this cadence to 70 rpm could have reduced the V
˙O2slow
component.
In running, the minimum of metabolic energy expenditure
during running with the freely chosen step frequency may de-
rive from a compromise which minimizes the step-average
force exerted by the muscles to sustain the impact against the
ground and the stiffening of the limbs [13]. Indeed, Cavanagh
and Williams [12] reported that the freelychosen stride length
in running allows for the most economical run. However, Mor-
gan et al. [26] reported that this was not systematically the
case for recreational distance runners, and half of the triath-
Int J Sports Med 1999; 20 V. L. Billat et al.434

letes examined in this study were first of all cyclists or swim-
mers and not runners.
Therefore, we hypothesized that the difference between spon-
taneous and most economical cadence which exists in cycling,
but not in running, could be one of the factors contributing to
the larger V
˙O2slow component in cycling for the same subject.
In order to test the possible influence of correspondence be-
tween freely chosen cadence and the economical one [12], we
imposed lower frequency in running and cycling, to respec-
tively induce or reduce a V
˙O2slow component. Some authors
[16,22] have demonstrated that while cycling at pedalling
rates of 80–90 rpm (as employed for free cadence in the pres-
ent study), the vastus lateralis shortens at a rate between 1 and
1.5 fibre lengths/sec. They underlined that this contraction
velocity is closer to the velocity of peak efficiency of a type I fi-
bre than a type II fibre. However, the lower cadence did not
change the V
˙O2slow component probably due to the occur-
rence of two contrary factors:
1. Firstly, an increase in the V
˙O2slow component by lowering
the cadence which is probably due to the increased muscle
force required to meet the higher resistance per cycle at a
lower pedal frequency is known to determine (as opposed
to velocity of contraction) the degree of type II fiber
recruitment when the metabolic cost of exercise is held
constant [1]. In our study, work rate at V
˙O2max was 302 ±
43 W and spontaneous cycling cadence was 82.5 ± 10 rpm.
Therefore since power output is equal to the force divided
by the velocity, the average force is equal to 302/
82.5 = 3.66 kgF. A decrease of 10% of the cadence induces
an increase of an average force of 366 gF. Cerretelli et al.
[14] demonstrated that in isometric contractions (5– 40 %
of maximal tension), V
˙O2was linearly related to the force
exerted.
2. Secondly, the factor which could decrease the V
˙O2slow
component when cadence is reduced is the fact that inter-
nal work and total mechanical work are decreased due to
less frequent accelerations of the limbs in regard to the
body center of mass.
In running, the slower stride frequency did not induce a V
˙O2
slow component, since the V
˙O2values at the third minute and
last minute of severe exercise were both higher than for spon-
taneous cadence, and the triathletes reached their V
˙O2max at
the end of the all-out severe run. Cavanagh and Williams [12]
have clearly demonstrated that for ten runners (mean V
˙O2max
64.7 ml/min/kg) at 14 km/h, an increase or decrease of 20% of
the stride frequency induced a mean increase of V
˙O2of 2.6 and
3.4 ml/min/kg, respectively. The simple process of shortening
or lengthening the stride has an important effect on all of the
active musculature: each muscle is forced to work on a slightly
different region of its force-velocity curve and consequently
changes in efficiency are seen [44]. However, in this study,
even if the lower cadence (– 10 %) decreased the running econ-
omy (i.e. + 2.27 ml/min/kg of V
˙O2at the same speed), this de-
crease was not significant and did not influence the time to fa-
tigue.
Therefore, we can suppose that a 10% modification of cadence
was insufficient to induce an important change of running
economy or a significant decrease of internal work in cycling,
but to compare the V
˙O2slow components between the sponta-
neous and lower cadence we needed to be sure that triathletes
would be able to sustain the severe exercise at least 6 minutes.
The V
˙
O2slow component is not related with time
to exhaustion in severe exercise
The absence of a correlation between the V
˙O2slow component
and time to exhaustion questions the mechanism of fatigue
during this severe intensity exercise. Indeed, regardless of the
fact whether triathletes reached their V
˙O2max (in cycling) or
not (in running), they had the same time to fatigue in both ex-
ercises. Whipp [42] suggested that the more rapidly the slow
component projects toward V
˙O2max, the shorter the tolerable
duration of the exercise test. In Nagle et al.'s study [28], sub-
jects did not reach their V
˙O2max and were not completely ex-
hausted. In the present study, both cases existed since during
running subjects did not reach their V
˙O2max, but were com-
pletely exhausted and in cycling, they reached their V
˙O2max
and were exhausted, too (after the same duration). This time
to fatigue is well above the duration of all-out exercise recom-
mended by Astrand and Saltin [2] which clearly pointed out
that it was possible to measure V
˙O2max with work-rate ex-
hausting subjects in 2–8 min. Jones [23] suggested that the
sensation of effort presumably reflects the magnitude of the
voluntary motor command generated. The second source of
sensory information, described as a sense of force or tension,
is derived from peripheral receptors in muscles, tendons and
the skin and is assumed to reflect the actual force exerted by
the muscle.
There might be different reasons for stopping cycling and run-
ning, even if the time to fatigue was not significantly different.
However, Vollestad et al. [40] demonstrated that during an ex-
haustive cycling exercise at 91% of V
˙O2max (20 minutes) about
95% of the fibres are activated in the vastus lateralis (VL) and
90% in the 10 first minutes. During severe exercise at 91% of
V
˙O2max, fibres of type I, IIA, IIAB, and IIB were recruited (with
a 1: 1 relationship between the fraction of active fibers in the
VL and the intensity determined as a percentage of V
˙O2max, if
exercise was performed until exhaustion). The authors did not
focus on the V
˙O2slow component. However, from their data
we could calculate an average V
˙O2slow component of 224 ml/
min between the 5th and the 20th minute (6.4% of V
˙O2max).
They suggested that at this exercise at 91% of V
˙O2max, the ac-
cumulation of lactate in muscle and blood in the first part of
exercise indicated a large anaerobic breakdown of glycogen;
this depletion suggested that even for this severe exercise of
10 minutes the lack of glycogen might be an important factor
in development of fatigue (as for longer and slower exercise).
For a 5000 m running race which elicits about 94% of V
˙O2max
during 14 minutes, Newholme et al. [29] suggested that the
possible major cause of fatigue (defined as the inability to
maintain power output), could be the depletion of glycogen.
However, the nature of the linkage between the V
˙O2slow com-
ponent and the fatigue process remains unclear [34] but we
can postulate that glycogen and PC depletion was responsible
for fatigue since the V
˙O2slow component is associated with
the increase of PC dephosphorylation and the Bohr effect
[39,41].
Effect of Cadence on V
˙O2Slow Component Int J Sports Med 1999; 20 435

Conclusion
The fact that cycling exercise induces a more pronounced V
˙O2
slow component could indicate that the type of exercise con-
traction regimen (isometric vs. eccentric) is significant in the
process of PC dephosphorylation and consequently in the V
˙O2
slow component phenomenon. However, it is also a determi-
nant to consider in further research the role of the handgrip in-
creasing with fatigue during cycling should be a part (signifi-
cant) of the difference of V
˙O2slow component in running and
cycling in triathletes well trained for both exercises. To further
investigate the hypothesis of the importance of the type of
muscular contraction regimen on V
˙O2slow component, it is,
for instance, possible to examine the influence of slope on the
V
˙O2slow component in running, since the slope decreases the
stretch-shortening cycle of running [11].
Acknowledgements
This study was supported by grants from Caisse Centrale des
Activités Sociales d'Electricité et Gaz de France.
References
1. Ahlquist LE, Bassett DR, Sufit R, Nagle FJ, Thomas DP. The effect
of pedalling frequency on glycogen depletion rates in type I and
II quadriceps muscle fibers during submaximal cycling exercise.
Eur J Appl Physiol 1992; 65: 360– 4
2. Astrand PO, Saltin B. Oxygen uptake during the first minutes of
heavy muscular exercise. J Appl Physiol 1961; 16: 971– 6
3. Astrand PO, Rodahl K. Textbook of work physiology. Physiol-
ogical bases of exercise. New York: McGraw-Hill publisher,
1986: 336
4. Aunola S, Rusko H. Reproducibility of aerobic and anaerobic
thresholds in 20–50 year old men. Eur J Appl Physiol 1984; 53:
260–6
5. Barstow TJ, Mole PA. Linear and nonlinear characteristics of
oxygen uptake kinetics during heavy exercise. J Appl Physiol
1991; 71: 2099– 106
6. Barstow TJ, Jones AM, Nguyen PH, Casaburi R. Influence of mus-
cle fiber type and pedal frequency on oxygen uptake kinetics of
heavy exercise. J Appl Physiol 1996; 81: 1642– 50
7. Billat VL, Binsse V, Haouzi P, Koralsztein JP. High level runners
are able to maintain a V
˙O2steady-state below V
˙O2max in an
all-out run over their critical velocity. Arch Physiol Bioch 1998;
107: 1 – 8
8. Billat VL, Richard R, Binsse VM, Koralsztein JP, Haouzi P. V
˙O2
slow component for a severe exercise depends on type of exer-
cise and is not correlated with time to fatigue. J Appl Physiol
1998; 85: 2118–2124
9. Capelli C, Antonutto G, Zamparo P, Girardis M, di Prampero PE.
Effects of prolonged cycle ergometerexercise on maximal mus-
cle power and oxygen uptake in humans. Eur J Appl Physiol
1993; 66: 189 – 95
10. Cavagna GA, Saibene FP, Margaria R. Mechanical work in run-
ning. J Appl Physiol 1964; 19: 249– 56
11. Cavagna GA, Kaneko M. Mechanical work and efficiency in level
walking and running. J Physiol 1977; 268: 467– 81
12. Cavanagh PR, Williams KR. The effect of stride length variation
on oxygen uptake during distance running. Med Sci Sports Ex-
erc 1982; 14: 30– 5
13. Cavagna GA, Willems PA, Franzetti P, Detrembleur C. The two
power limits conditioning step frequency in human running. J
Physiol 1991; 437: 95–108
14. Cerretelli P, Veicsteinas A, Fumagalli M, Dell'orto L. Energetics of
isometric exercise in man. J Appl Physiol 1976; 41: 136– 41
15. Costill DL. Metabolic responses during distance running. J Appl
Physiol 1970; 28: 251 –5
16. Coyle EF, Labros S, Sidossis S, Horowitz JF, Beltz JD. Cycling effi-
ciency is related to the percentage of type I muscle fibers. Med
Sci Sports Exerc 1993; 25: 1269– 74
17. Farrel PE, Wilmore JH, Coyle EF, Billing JE, Costill DL. Plasma lac-
tate accumulation and distance running performance. Med Sci
Sports Exerc 1979; 11: 338 –44
18. Gaesser GA. O2uptake during high intensity cycling at slow and
fast cadences (abstract). Physiologist 1992; 35: 210
19. Gaesser GA. Influence of training and catecholamines on exer-
cise V
˙O2response. Med Sci Sports Exerc 1994; 26: 782–8
20. Gaesser GA, Poole D. The slow component of oxygen uptake ki-
netics in humans. Exerc Sport Sci Rev 1996; 24: 35 –70
21. Harris RC, Sahlin K, Hultman E. Phosphagen and lactate con-
tents of m. quadriceps of man after exercise. J Appl Physiol
1977; 43: 852– 7
22. Horowitz JF, Sidossis LS, Coyle EF. High efficiency of type I mus-
cle fibers improves performance. Int J Sports Med 1991; 15:
152 – 7
23. Jones LA. The senses of effort and force during fatiguing con-
tractions. In: Gandevia SC et al. (eds). Fatigue. Neural and Mus-
cular Mechanisms. New York: Plenum Press, 1995: 305 – 13
24. Mahler M. Kinetics and control of oxygen consumption in skele-
tal muscle. In: Cerretelli P, Whipp BJ (eds). Exercise Bioenerget-
ics and Gas Exchange. Elsevier/North-Holland: Biomedical
Press publishers, 1980: 53 – 66
25. Marsh AP, Martin PE. The association between cycling experi-
ence and preferred and most economical cadences. Med Sci
Sports Exerc 1993; 24: 782– 8
26. Morgan D, Martin P, Craib M, Caruso C, Clifton R, Hopewell R. Ef-
fect of step length optimization on the aerobic demand of run-
ning. J Appl Physiol 1994; 77: 245– 51
27. Moritani T, Nagata A, De Vries HA, Muro M. Critical power as a
measure of physical working capacity and anaerobic threshold.
Ergonomics 1981; 24: 339 –50
28. Nagle FJ, Robinhold D, Howley E, Daniels J, Baptista G, Stoede-
falke K. Lactic acid accumulation during running at submaximal
aerobic demands. Med Sci Sports Exerc 1970; 2: 182 –6
29. Newsholme EA, Blomstrand E, Ekblom B. Physical and mental
fatigue: metabolic mechanisms and importance of plasma ami-
no acids. Br Med Bull 1992; 48: 477– 95
30. Okita K, Nishijima H, Yonezawa K, Ohtsubo M, Hanada A, Kohya
T, Murakami T, Kitabatake A. Skeletal muscle metabolism in
maximal bicycle and treadmill exercise distinguished by using
in vivo metabolic freeze method and phosphorus-31 magnetic
resonance spectroscopy in normal men. Am J Cardiol 1998; 81:
106 – 9
31. Patterson RP, Moreno MI. Bicycle pedalling forces as a function
of pedalling rate and power output. Med Sci Sports Exerc 1990;
22: 512 – 6
32. Poole DC, Ward SA, Gardner GW, Whipp BJ. Metabolic and re-
spiratory profile of the upper limit for prolonged exercise in
man. Ergonomics 1988; 31: 1265–79
33. Poole DC, Schaffartzik W, Knight DR, Derion T, Kennedy B, Guy
HJ, Prediletto R, Wagner PD. Contribution of exercising legs to
the slow component of oxygen uptake kinetics in humans. J
Appl Physiol 1991; 71: 1245– 53
34. Poole DC, Barstow TJ, Gaesser GA, Willis WT, Whipp BJ. V
˙O2
slow component: physiological and functional significance.
Med Sci Sports Exerc 1994; 26: 1354– 8
35. Roston WL, Whipp BJ, Davis JA, Cunningham DA, Effros RM,
Wasserman K. Oxygen uptake kinetics and lactate concentra-
tion during exercise in humans. Am Rev Respir Dis 1987; 135:
1080– 4
Int J Sports Med 1999; 20 V. L. Billat et al.436

36. Sjödin B, Jacobs I. Onset of blood lactate accumulation and
marathon running performance. Int J Sports Med 1981; 2: 23 –6
37. Stringer WS, Wasserman K, Casaburi R, Porszasz J, Maehara K,
French W. Lactic acidosis as facilitator of oxyhemoglobin disso-
ciation during exercise. J Appl Physiol 1994; 76: 1462– 7
38. Takaishi T, Yasuda Y, Ono T, Moritani T. Optimal pedalling rate
estimated from neuromuscular fatigue for cyclists. Med Sci
Sports Exerc 1996; 28: 1492– 7
39. Takaishi T, Yamamoto T, OnoT, Ito T, Moritani T. Neuromuscular,
metabolic, and kinetic adaptations for skilled pedalling per-
formance in cyclists. Med Sci Sports Exerc 1998; 30: 442 –9
40. Vollestad NK, Bolm PCS. Effect of varying exercise intensity on
glycogen depletion in human muscle fibers. Acta Physiol Scand
1985; 125: 395– 405
41. Wasserman K, Hansen JE, Sue DY. Facilitation of oxygen con-
sumption by lactic acidosis during exercise. News Physiol Sci
199 1; 6 : 2 9 – 34
42. Whipp BJ. The slow component of O2uptake kinetics during
heavy exercise. Med Sci Sports Exerc 1994; 26: 1319 –26
43. Whipp BJ, Wasserman K. The efficiency of muscular work. J Appl
Physiol 1969; 26: 644– 8
44. Wilkie DR. The relation between force and velocity in human
muscle. J Physiol 1950; 110: 249–80
45. Willis WT, Jackman MR. Mitochondrial function during heavy
exercise. Med Sci Sports Exerc 1994; 26: 1347–54
Corresponding Author:
Véronique Billat, Ph.D.
Centre de Médecine du Sport CCAS
2, avenue Richerand
F-75010 Paris
France
Phone: + 33 (1) 42020818
Fax: + 33 (1) 42392083
E-mail: Veronique.Billat@Wanadoo.fr
Effect of Cadence on V
˙O2Slow Component Int J Sports Med 1999; 20 437