The Relationship Between Freely Chosen Cadence

International Journal of Sports Physiology and Performance, 2012, 7, 375-381
© 2012 Human Kinetics, Inc.
www.IJSPP-Journal.com
ORIGINAL INVESTIGATION
The Relationship Between Freely Chosen Cadence
and Optimal Cadence in Cycling
Umberto Emanuele, Tamara Horn, and Jachen Denoth
Purpose: The main aim of this study was to compare the freely chosen cadence (FCC) and the cadence at
which the blood lactate concentration at constant power output is minimized (optimal cadence [Copt]). The
second aim was to examine the effect of a concomitant change of road incline and body position on FCC, the
maximal external power output (Pmax), and the corresponding Copt. Methods: FCC, Copt, and Pmax were analyzed
under 2 conditions: cycling on level ground in a dropped position (LGDP) and cycling uphill in an upright
position (UHUP). Seven experienced cyclists participated in this study. They cycled on a treadmill to test the
2 main hypotheses: Experienced cyclists would choose an adequate cadence close to Copt independent of the
cycling condition, and FCC and Copt would be lower and Pmax higher for UHUP than with LGDP. Results: Most
but not all experienced cyclists chose an adequate cadence close to Copt. Independent of the cycling condition, FCC and Copt were not statistically different. FCC (82.1 ± 11.1 and 89.3 ± 10.6 rpm, respectively) and
Copt (81.5 ± 9.8 and 87.7 ± 10.9 rpm, respectively) were significantly lower and Pmax was significantly higher
(2.0 ± 2.1%) for UHUP than for LGDP. Conclusion: Most experienced cyclists choose a cadence near Copt to
minimize peripheral fatigue at a given power output independent of the cycling condition. Furthermore, it is
advantageous to use a lower cadence and a more upright body position during uphill cycling.
Keywords: FCC, uphill, level ground, body position, power output
On modern multigeared bicycles, the cyclist is free
to choose any available gear ratio to achieve a target
cycling velocity. The choice of the gear ratio at a given
cycling velocity in turn determines the cadence. Thus,
as noted by Vercruyssen and Brisswalter,1 the cadence
is one of the only variables that an athlete can modify
during cycling exercise to optimize performance. Hence,
cadence selection has been of interest since the multigeared chain-driven bicycle was invented in the late 19th
century. For more than 100 years, cyclists, coaches, and
researchers have debated over cadence choice. Even in
the last 10 years, several scientific studies on the freely
chosen cadence (FCC) have been published, further
demonstrating that there is a continuous discussion on the
theory and practice of cycling. Research has shown that
FCC is a highly individual characteristic and is affected
by several “external” factors.2 It has been reported that
external power output (Pext),3 road gradient,4 crank inertial
load,3 drafting,5 fatigue,6 training,7 muscle-fiber-type
composition,8 and circadian rhythm9 influence FCC.
Despite this intense discussion, the underlying factors
leading to the choice of a particular cadence are still a
subject of debate. Theoretically, cyclists should choose
an “optimal cadence” at which relevant physiological
The authors are with the Institute for Biomechanics, ETH
Zurich, Zurich, Switzerland.
and/or biomechanical variables are optimized. However,
several studies document the multitude of definitions
for “optimal cadence” in cycling (for a review see, eg,
Marais and Pelayo10). The term optimal cadence has been
defined as the cadence with the lowest EMG activity for
a given Pext,11 the cadence that minimizes the summed
joint moments,12 the cadence that produces the lowest
neuromuscular fatigue,13 the cadence with the highest
Pext at a specific performance level,14 and the most efficient cadence for energy expenditure.15 The “cadence
paradox,”16 which states that competitive cyclists and
trained subjects choose a higher cadence than the most
efficient cadence (Ceff), has been an especially interesting
topic in the past.17
Of most interest (at least for a competitive road
cyclist) is the cadence that produces the best performance
to win a race, not Ceff. Therefore, the optimal cadence
(Copt) is defined here as the cadence that corresponds to
the apex of the Pext–cadence relationship (Pmax) at a performance level that can be sustained for the given task.14
From a theoretical perspective, cyclists should choose a
cadence near Copt to maximize Pext for the given task or
to minimize peripheral fatigue at a given Pext. Results
from previous studies showed that both FCC3 and Copt18
were significantly lower during uphill cycling than during
cycling on level ground with the same body position.
Taken together, these results suggest that both FCC and
Copt are affected in a similar way by road incline, but no
375
376 Emanuele, Horn, and Denoth
study has addressed whether the difference between FCC
and Copt is the same independent of road incline. Moreover, to the best of our knowledge, no study has compared
FCC and Copt at a submaximal performance level.
A study by Emanuele and Denoth18 showed that the
road incline not only decreased Copt but also decreased
Pmax. In addition, that study showed that the change
in body position from a dropped posture (DP) with
the cyclist’s hands on the drops of the handlebars and
arms fully extended to an upright posture (UP) with the
cyclist’s hands on the top portion of the handlebars and
arms fully extended significantly increased Pmax without
affecting the corresponding Copt. These results led to the
conclusion that Copt and FCC should be lower and Pmax
should be higher for uphill cycling with UP than for
level-ground cycling with DP.18 However, the influence
of a simultaneous change in road incline and body position as seen in field observations of cyclists has not been
analyzed in an experimental study.
The main aim of this study was to compare FCC
and Copt at an endurance-performance level that can be
sustained for approximately 1 hour. The hypothesis was
that at a performance level corresponding to the maximal
lactate steady state (MLSS), experienced cyclists would
choose an adequate cadence close to Copt independent of
the cycling condition. The second aim of this study was
to examine the effect of a concomitant change in road
incline and body position on FCC, Copt, and Pmax at an
endurance-performance level corresponding to MLSS.
The hypothesis was that FCC and Copt would be lower
and Pmax higher during uphill cycling with UP than during
level-ground cycling with DP.
Methods
Subjects
Seven well-trained male amateur cyclists (31 ± 6 y,
182.0 ± 4.5 cm, 71.7 ± 6.7 kg) who have competed at
the national level volunteered to participate in this study.
They were all informed of the nature of this study, as well
as the potential risk and discomfort associated with the
experimental procedures, before they gave their written
consent to participate. The ethical committee of ETH
Zurich approved the experimental design.
Determination of FCC. To determine FCC during
level-ground cycling (FCCLG) the subjects cycled with
DP at a road inclination of 0% and a speed of 30.2 km/h.
During this test, the cyclists were encouraged to change
gears and use the gear that felt the most comfortable.
However, because of the limited gear ratios, there were
limitations on the FCCs. Only the following cadences
could be chosen: 60, 65, 70, 75, 80, 85, 90, 95, 100, and
105 rpm. After a 10-minute warm-up at 100 W, Pext was
increased until it reached a level that the cyclist thought he
could maintain for approximately 1 hour. This subjective
estimation of PMLSS was reached within 5 to 7 minutes.
Pext and gear were then held constant for 1 minute, and
the cadence (FCC) and Pext were recorded. The subjects
were then given a 20-minute rest period.
To determine FCC during uphill cycling (FCCUH),
the same protocol was used, except that subjects cycled
with UP on the treadmill with an inclination of 7% and
a speed of 15.1 km/h.
Lactate Minimum Test. The determination of the lactate
minimum power output (PLM) using a lactate minimum
test (LMT) is a reliable and valid method to predict PMLSS
and therefore a good predictor of the specific endurance
performance level that can be sustained for approximately
1 hour.19,20 During the LMT, the participants cycled with
UP and a constant cadence of 80 rpm, which corresponds
to the mean Copt measured by Emanuele and Denoth18
during cycling using an ergometer. The method used to
determine PLM was based on Fontana et al21 (Figure 1).
Briefly, the LMT consisted of 2 incremental tests. The
first incremental test started at 100 W and increased 50
W every minute until exhaustion to induce lactic acidosis,
followed by 1 minute at 100 W and 7 minutes of complete
rest (part 1). After the rest period, a second incremental
test started at 100 W with an increase of 25 W every 90
seconds (part 2). During the LMT, heart rate, oxygen
uptake, carbon dioxide output, minute ventilation, and
breathing frequency were continuously recorded. Blood
lactate concentration (bLa) was measured at rest, at
exhaustion in part 1, during recovery in part 1 (after 2, 4,
Experimental Design
The subjects were asked to attend 3 test sessions within
a 3-week period with at least 2 days between single test
days. To improve the reliability of the measurements, the
participants were asked to control a number of variables;
for the details see Emanuele and Denoth.14,18 The purpose
of the first test session was to determine individual FCC
and estimate the individual Pext at maximal lactate steady
state (PMLSS). The purpose of the second and third test
sessions was to estimate Copt and Pmax at MLSS when
cycling on level ground with DP and when cycling uphill
with UP, respectively.
Figure 1 — Schematic of the protocol for the lactate minimum test. The test consists of 2 incremental tests, separated
by 1 minute at 100 W and 7 minutes of complete rest. The first
incremental test started at 100 W with 50-W increments every
minute until exhaustion to induce lactic acidosis. The second
incremental test started at 100 W with 25-W increments every
90 seconds. × = Measurements of blood lactate concentration
(bLa).
FCC and Optimal Cadence in Cycling 377
and 7.5 min), and at the end of each stage in part 2. The
bLa levels during part 2 of the LMT were plotted against
power output, and a third-order polynomial curve was
fitted through these data points.21 The PLM corresponds
to the “lactate minimum” in this polynomial curve. PLM
was used as an estimation of PMLSS at 80 rpm, the exercise
intensity needed for the second and third test sessions.
Determination of Copt. To determine Copt at MLSS
during uphill cycling, the subjects performed 3
incremental exercise tests at the following 3 cadences:
FCCUH – 10 rpm, FCCUH, and FCCUH + 10 rpm in a
randomized order. During these tests, the subjects cycled
with UP on the treadmill at an 7% inclination at 15.1
km/h. At each cadence, a 2-minute warm up at 100 W
was followed by an incremental exercise test comprising
4 increments in Pext (Figure 2): P80MLSS – 20 W (3 min),
P80MLSS – 10 W (3 min), P80MLSS (2 min), and P80MLSS + 10
W (2 min). The incremental test was followed by a 2.5minute cooldown at 100 W. Between the 3 incremental
exercise tests, the subjects were given a 14-minute rest to
prevent fatigue. bLa was measured at the beginning of the
warm-ups and at the end of each stage of the incremental
exercise tests. To determine Copt at MLSS, the increase in
bLa was analyzed. The increase in bLa from rest until the
end of the exercise test was plotted against the cadence
used (Figure 3[a]). A second-order polynomial-regression
line was fitted through these data points to determine the
bLa–cadence relationship. Copt at MLSS corresponds
to the “lactate minimum” in this polynomial curve. A
quadratic curve constrained to pass through the origin
was then fit to Copt at MLSS and PMLSS at 80 rpm to assess
the individual Pmax at MLSS (Figure 3[b]).14
To determine Copt at MLSS during level-ground
cycling, the same protocol was used, except that the
treadmill was set at an inclination of 0%, the speed was
increased to 30.2 km/h, and the participants cycled with
DP. The bLa–cadence relationship for level-ground
cycling was then compared with the bLa–cadence relationship for uphill cycling. Pext for uphill cycling with
UP corresponding to the intersection point of the 2 bLa–
cadence relationships was used to fit a quadratic curve
constrained to pass through the origin to assess the individual Pmax at MLSS for level-ground cycling with DP.
Figure 3 — (a) Blood lactate concentration (bLa) in relation to
cadence. Mean values for bLa at freely chosen cadence (FCC),
FCC – 10 rpm, and FCC + 10 rpm are shown for uphill cycling
(white triangles) and for level-ground cycling (white squares).
Second-order polynomial-regression curves were fitted to the
data on level ground cycling with the dropped posture (LGDP;
dashed line) and uphill cycling with the upright posture (UHUP;
solid line) to assess the optimal cadence (Copt) for UHUP (black
triangle) and Copt for LGDP (black square). The cadences were
normalized relative to the Copt at maximal lactate steady state
(MLSS) for LGDP. (b) Power–cadence relationships at the
performance level corresponding to the MLSS. To assess the
maximal power output (Pmax) at MLSS for UHUP, a quadratic
curve (solid line) constrained to pass through the origin was
fitted to the assessed Copt for UHUP (black triangle) and the
power output at MLSS estimated with the lactate minimum
test at 80 rpm (P80MLSS, white triangle). To assess Pmax at MLSS
for LGDP (black square) the Pext for UHUP corresponding to the
intersection point of the 2 bLa–cadence relationships (white
square) was used to fit a quadratic curve constrained to pass
through the origin (dashed line). These power outputs and
cadences were normalized relative to Pmax at MLSS and the
corresponding optimal cadence (Copt at MLSS) for LGDP.
Equipment
Figure 2 — Schematic of the protocol for test sessions 2
and 3. PMLSS80 = power output at maximal lactate steady state
estimated with a lactate minimum test at 80 rpm. × = Measurements of blood lactate concentration (bLa).
In all tests, the subjects exercised on a normal road-racing
bicycle with 24-36-48 teeth chain rings and 12-13-1415-17-18-19-20-21-24 teeth cog set. The bicycle was
adjusted so that the vertical and horizontal positions of
the saddle and the handlebar related to the crank axis
matched each subject’s own bicycle. The racing bicycle
was equipped with a standard crank (length = 170 mm),
clipless pedals, and a professional (8 strain gages) SRM
PowerMeter (Schoberer Rad Messtechnik, Jülich, Germany). Using the SRM PowerMeter, Pext, cadence, and
heart rate were recorded. During all tests except the LMT,
the bicycle was mounted on a treadmill (Woodway, Weil
378 Emanuele, Horn, and Denoth
am Rhein, Germany). Thereby, the bicycle was fixed with
the fork to a sliding carriage, which allowed a horizontal bicycle translation relative to the laboratory (Figure
4). Pext was adjusted by changing the mass of a weight
magazine. This magazine was connected to a wire that
ran over a pulley placed behind the treadmill and was
then tied to the back of the bicycle.3,22 The mass needed
to achieve the target Pext was calculated for each subject
based on his measured Pext without the weight magazine.
During the LMT, the bicycle was mounted on an
indoor trainer (Flow, Tacx, Wassenaar, Netherlands).
The gas-exchange and ventilatory variables were measured breath by breath with an Oxycon Mobile device
(Viasys Healthcare, Höchberg, Germany). The lactate
concentration in the blood samples (20 μL) taken from
the earlobes was analyzed with a Biosen C-Line analyzer
(EKF Industrie-Elektronik, Barleben, Germany).
Statistics
All of the statistical analyses were performed using
SPSS Statistics 17 (SPSS Inc, Chicago, IL). The level of
significance was set at P < .05. Based on the estimated
coefficient of variation for the measured bLa of 5%, the
95% confidence interval (CI) for assessing the individual
Copt at MLSS (Figure 5) was calculated using the modelbased residual bootstrapping method for regression.14,18
Assuming that the cyclists use the next higher (Copt+)
or lower (Copt–) possible cadence to the real Copt (Copt*)
(Figure 5), the theoretical 95% CI for the difference
between FCC and the measured/estimated Copt at MLSS
(ΔC) was calculated using the 95% CI for the individual
Copt at MLSS and the available gear ratios. The theoretical 95% CIs for ΔC were also calculated assuming that
the cyclists chose Copt+ + 5 rpm or Copt– – 5 rpm. The
statistical power for detecting a relevant ΔC over 5 rpm
was calculated for a sample size of 7. All the parameter
values were compared using the Student paired t test.
The variables were summarized with descriptive statistics
(mean ± SD).
Figure 4 — The standard bicycle mounted on the treadmill.
The fork was fixed to the sliding carriage, which allowed horizontal bicycle translation relative to the laboratory (30 cm).
Figure 5 — Schematic of the different cadences. Black squares
mark the available cadences that the cyclist can choose. The
white circle marks the cyclist’s freely chosen cadence (FCC).
The asterisk (*) marks the real optimal cadence (Copt*). The
black triangle marks the measured/estimated optimal cadence
(Copt). Copt– is the next lower available cadence to Copt*. Copt+ is
the next higher available cadence to Copt*. ΔC is the difference
between FCC and Copt. The possible distribution of Copt using
the method to determine Copt described in this study is shown
(thin solid line). The 95% confidence interval (CI) to assess
Copt is marked by the dotted lines.
Results
Based on the coefficient of variation for the measured
increase in bLa, the residual bootstrap yielded a 95%
CI for assessment of the individual Copt at MLSS of 4.8
rpm (Figure 6[a]). Assuming that the cyclists use Copt+
or Copt–, the theoretical 95% CI for ΔC was 0 ± 7.4 rpm
(Figure 6[b]). During level-ground cycling, ΔC of all
cyclists was within this CI, and during uphill cycling
the difference of 1 cyclist was outside this CI. Assuming
that the cyclists use Copt+ + 5 rpm or Copt– – 5 rpm, the
theoretical 95% CIs for ΔC were 7.5 ± 3.7 rpm and –7.5
± 3.7 rpm. During level-ground cycling, ΔC of 6 cyclists
was outside this CI, and during uphill cycling, ΔC of 4
cyclists was outside this CI. Thus, during level-ground
cycling, 6 cyclists used Copt+ or Copt– and 1 cyclist used
Copt+ or Copt+ +5 rpm. During uphill cycling, 4 cyclists
used Copt+ or Copt–, 1 cyclist used Copt– or Copt– – 5 rpm,
and 1 cyclist used Copt+ + 5 rpm.
For uphill cycling with UP and level-ground cycling
with DP, the determined FCC and Copt at MLSS were not
significantly different. The statistical power to detect a
relevant difference of 5 rpm between FCC and Copt at
MLSS was .84 for a sample size of 7.
FCC at the subjective self-selected PMLSS (278 ±
30 W) was significantly lower (P < .05) when cycling
uphill with UP (82.1 ± 11.1 rpm) than when cycling on
level ground with DP (89.3 ± 10.6 rpm). PMLSS at 80 rpm
estimated using the LMT was 261 ± 24 W. The assessed
individual Copt at MLSS was significantly higher (P < .01)
for level-ground cycling with DP (87.7 ± 10.9 rpm) than
for uphill cycling with UP (81.5 ± 9.8 rpm). Copt at MLSS
for level-ground cycling was 7.5% ± 2.3% higher than for
uphill cycling. Finally, Pmax at MLSS was significantly
higher (2.0% ± 2.1%; P < .05) for uphill cycling with
UP than for level-ground cycling with DP (Figure 3[b]).
FCC and Optimal Cadence in Cycling 379
Figure 6 — (a) The relationship between optimal cadence (Copt) and freely chosen cadence (FCC). The data for uphill cycling with
the upright posture (white triangles) and level-ground cycling with the dropped posture (white squares) are shown for all subjects
(n = 7). The solid line represents the identity line (FCC = Copt). The 95% confidence interval (CI) for assessing Copt is shown by the
2 dashed lines. The dotted lines mark the available cadences that the cyclist could choose. (b) The individual differences between
FCC and Copt (ΔC). The ΔCs for uphill cycling with the upright posture and level-ground cycling with the dropped posture are shown
for all subjects (n = 7). The solid lines mark the theoretical 95% CI for ΔC assuming that the cyclists choose the next higher (Copt+)
or lower (Copt–) possible cadence to the real Copt. The dashed lines mark the theoretical 95% CIs for ΔC assuming that the cyclists
choose the next higher Copt+ + 5 rpm or Copt– – 5 rpm. The gray zone marks the intersection of 2 CIs. For the data in this gray zone
it is not possible to determine if the cyclists used Copt+ or Copt+ + 5 rpm or Copt– or Copt– – 5 rpm.
Discussion
FCC and Copt
The main purpose of this study was to compare FCC
and Copt at an endurance-performance level that corresponded to MLSS. The hypothesis was that experienced
cyclists would choose an adequate cadence close to
Copt independent of the cycling condition. The results
revealed that most but not all experienced cyclists choose
the next higher (Copt+) or the next lower (Copt–) possible
cadence to Copt, independent of the cycling condition.
However, the results also showed that, independent of
road incline and body position, FCC and Copt at MLSS
were not significantly different. Here it must be stated
that the statistical power to detect a difference between
FCC and Copt at MLSS was greater than .8 only for an
assumed difference of 5 rpm between FCC and Copt at
MLSS. On the other hand, a difference between FCC and
Copt at MLSS of less than 5 rpm is not relevant from a
performance-related point of view, because such a difference reduces Pext less than 0.5%.14 Furthermore, most of
the individual differences between FCC and Copt at MLSS
can be attributed to the discrete available gear ratios and
the precision in determination of Copt at MLSS.
According to Emanuele and Denoth 14 C opt is
defined as the cadence that corresponds to the apex of
the Pext–cadence relationship at a specific performance
level, for example, at a fixed bLa. From a theoretical
perspective, Copt can be indirectly assessed by measuring
different variables at a constant Pext while manipulating
the cadence. One of these indicative variables is bLa.
Emanuele and Denoth14 showed that to assess Copt at
a fixed bLa, the Pext–cadence relationship can be fitted
using a quadratic-regression line that is constrained to
pass through the origin. This result implies that Copt can
also be assessed using a quadratic bLa–cadence relationship at a constant Pext. Thus, the results from this study
confirm the assumption that most competitive cyclists
choose a cadence near Copt to minimize peripheral fatigue
at a given Pext or to maximize Pext for the given task. This
assumption has been stated by different authors but has
never been experimentally demonstrated. The following additional indicative variables for Copt at a constant
Pext have been analyzed previously: neuromuscular
fatigue,13 EMG activity,11,23 and time to exhaustion.24 All
of the single data sets of these studies were well fitted
with a second-order polynomial-regression curve (R2 =
.88–.99) to assess Copt. However, none of these studies
assessed Copt and compared it with FCC. Thus, this is
the first study to show that most competitive cyclists
choose an adequate cadence close to the individual Copt
at a specific endurance-performance level. Furthermore,
the comparison between level-ground cycling with DP
and uphill cycling with UP revealed that not only interindividual differences in FCC but also intraindividual
differences in FCC are related to Copt. These are experimental indications of a causative relationship between
FCC and Copt. This relationship can also be deduced
from the observation that both FCC6 and Copt25 decrease
380 Emanuele, Horn, and Denoth
with increasing fatigue. Thus, the cadence paradox16
is resolved because most competitive cyclists choose
a cadence near Copt, which is clearly a higher cadence
than the most efficient cadence. Kohler and Boutellier16
have already explained that FCC exceeds Ceff. In their
theoretical study, they concluded that FCC is not fixed but
depends on the duration of the race and ranges between
Copt and Ceff. However, in their conclusion they neglected
the fact that not only FCC but also Copt and Ceff depend on
performance level or, rather, race duration.14,26 Even their
own calculation showed that Copt, Ceff, and the difference
between these cadences depend on performance level or,
rather, on type II fiber recruitment. Thus, a change in the
difference between Ceff and FCC can be explained by
the change in the difference between Copt and Ceff. Other
studies that have tried to explain the cadence paradox
have shown a reduced difference between FCC and Ceff
with fatigue.6,17,27 Argentin et al,27 for example, showed
a significant shift in FCC (87–68 rpm) toward Ceff (65
rpm) after 2 hours of cycling at 65% of maximal aerobic
power. As Copt decreases with fatigue, which is primarily from an increase in the internal power,14 the reduced
difference between FCC and Ceff with fatigue can also
be explained by the reduced difference between Copt and
Ceff with fatigue. Regrettably, no study has simultaneously
analyzed the influence of fatigue on FCC, Copt, and Ceff
to confirm this theoretical assumption.
The hypothesis that most competitive cyclists choose
a cadence near Copt should be confirmed in future studies by simultaneously analyzing the influence of other
“external” factors on FCC and Copt. The factors of interest
could include Pext, fatigue, training, muscle-fiber-type
composition, saddle height, and crank length. Furthermore, it is not assumed that all competitive cyclists
chose an optimal cadence. Therefore, the percentage of
cyclists that chose too high or too low a cadence must be
estimated in a future study using a large number of subjects. Studies should also be performed with recreational
cyclists and noncyclists to show if the subjects choose an
optimal cadence independent of cycling experience. It is
recommended that future studies be performed with more
cadences tested to enhance precision in determining Copt.
Level Ground Versus Uphill
By comparing level-ground cycling with DP and uphill
cycling with UP, this study had a second aim: to determine
the influence of a concomitant change in road gradient
and body position on FCC, Pmax, and Copt. Concerning FCC, experimental studies have revealed that it is
influenced by road gradient3 but not by body position.28
Thus, a concomitant change in the road gradient and body
position should theoretically have the same influence on
FCC as only a change of road gradient. Our results on
the treadmill support the results from Hansen et al,3 who
also observed a lower FCC when subjects cycled uphill
on a treadmill (69 rpm at 150 W and 73 rpm at 250 W)
than while cycling on level ground (75 rpm at 150 W and
82 rpm at 250 W). These results also agree with observations of cyclists during their normal training29 or during
competitions.30 A previous study showed that a change
in the road gradient (0 vs 7%) alone influences Pmax and
Copt and that a change in body position (DP vs UP) alone
influences Pmax but not Copt.18 These results led to the
assumption that Copt is lower and Pmax is higher for uphill
cycling with UP than for level-ground cycling with DP.
This assumption has now been experimentally confirmed.
Practical Applications
The results from this study suggest that most competitive
cyclists freely choose a cadence near Copt to minimize
fatigue at a given Pext. Thus, athletes should not attempt
to copy the cadence used by successful cyclists and
coaches should avoid forcing athletes to use a desired
cadence. On the other hand, some cyclists freely choose
too high or too low a cadence. Thus, the comparison of
the individual FCC and the individual Copt at a specific
endurance-performance level is an adequate method to
optimize performance.
The results from the second part of this study confirmed that under real cycling conditions it is advantageous from a performance-related point of view to use a
lower cadence and a more upright body position during
uphill cycling.
Conclusion
Despite the intense discussion on the FCC in cycling, the
underlying factors that lead cyclists to choose a particular
cadence remained under debate. This study demonstrated
that interindividual, as well as intraindividual, differences
in FCC are related to differences in Copt. These results
indicate a causative relationship between FCC and Copt.
As expected from a theoretical perspective, most cyclists
freely choose a cadence near Copt to minimize peripheral
fatigue at a given Pext or to maximize Pext for the given
task. Thus, the cadence paradox, which states that competitive cyclists choose a higher cadence than Ceff, has
been resolved.
Furthermore, by comparing level-ground cycling
with DP and uphill cycling with UP, this study confirmed
that under real cycling conditions it is advantageous to
use a lower cadence and a more upright body position
during uphill cycling.
Acknowledgments
These studies were supported by the Swiss Federal Council of
Sports (ESK/BASPO). The authors are grateful to the subjects
for having participated in these studies. They also thank Marco
Hitz and Peter Schwilch for their technical support.
References
1. Vercruyssen F, Brisswalter J. Which factors determine the
freely chosen cadence during submaximal cycling? J Sci
Med Sport. 2010;13(2):225–231. PubMed doi:10.1016/j.
jsams.2008.12.631
FCC and Optimal Cadence in Cycling 381
2.Hansen EA, Smith G. Factors affecting cadence choice
during submaximal cycling and cadence influence on performance. Int J Sports Physiol Perform. 2009;4(1):3–17.
PubMed
3. Hansen EA, Jorgensen LV, Jensen K, Fregly BJ, Sjogaard
G. Crank inertial load affects freely chosen pedal rate
during cycling. J Biomech. 2002;35(2):277–285 Erratum
in J Biomech. 2002;35(2):1521. PubMed doi:10.1016/
S0021-9290(01)00182-8
4.Lucia A, Hoyos J, Chicharro JL. Preferred pedalling
cadence in professional cycling. Med Sci Sports Exerc.
2001;33(8):1361–1366. PubMed
5. Hausswirth C, Lehenaff D, Dreano P, Savonen K. Effects of
cycling alone or in a sheltered position on subsequent running performance during a triathlon. Med Sci Sports Exerc.
1999;31(4):599–604. PubMed doi:10.1097/00005768199904000-00018
6.Vercruyssen F, Hausswirth C, Smith D, Brisswalter J.
Effect of exercise duration on optimal pedaling rate choice
in triathletes. Can J Appl Physiol. 2001;26(1):44–54.
PubMed
7. Hausswirth C, Argentin S, Bieuzen F, Le Meur Y, Couturier
A, Brisswalter J. Endurance and strength training effects on
physiological and muscular parameters during prolonged
cycling. J Electromyogr Kinesiol. 2010;20(2):330–339.
doi:10.1016/j.jelekin.2009.04.008. PubMed
8. Hansen EA, Andersen JL, Nielsen JS, Sjogaard G. Muscle
fibre type, efficiency, and mechanical optima affect freely
chosen pedal rate during cycling. Acta Physiol Scand.
2002;176(3):185–194. PubMed doi:10.1046/j.1365201X.2002.01032.x
9. Moussay S, Dosseville F, Gauthier A, Larue J, Sesboue
B, Davenne D. Circadian rhythms during cycling exercise
and finger-tapping task. Chronobiol Int. 2002;19(6):1137–
1149. PubMed doi:10.1081/CBI-120015966
10. Marais G, Pelayo P. Cadence and exercise: physiological
and biomechanical determinants of optimal cadences—
practical applications. Sports Biomech. 2003;2(1):103–
132. PubMed doi:10.1080/14763140308522811
11.MacIntosh BR, Neptune RR, Horton JF. Cadence,
power, and muscle activation in cycle ergometry. Med
Sci Sports Exerc. 2000;32(7):1281–1287. PubMed
doi:10.1097/00005768-200007000-00015
12. Marsh AP, Martin PE, Sanderson DJ. Is a joint momentbased cost function associated with preferred cycling
cadence? J Biomech. 2000;33(2):173–180. PubMed
doi:10.1016/S0021-9290(99)00155-4
13. Takaishi T, Yasuda Y, Ono T, Moritani T. Optimal pedaling
rate estimated from neuromuscular fatigue for cyclists.
Med Sci Sports Exerc. 1996;28(12):1492–1497. PubMed
14.Emanuele U, Denoth J. Power–cadence relationship in
endurance cycling. Eur J Appl Physiol. 2012;112(1):365–
375. PubMed doi:10.1007/s00421-011-1987-z
15. Coast JR, Welch HG. Linear increase in optimal pedal rate
with increased power output in cycle ergometry. Eur J Appl
Physiol Occup Physiol. 1985;53(4):339–342. PubMed
doi:10.1007/BF00422850
16.Kohler G, Boutellier U. The generalized force–velocity
relationship explains why the preferred pedaling rate of
cyclists exceeds the most efficient one. Eur J Appl Physiol.
2005;94(1-2):188–195. PubMed doi:10.1007/s00421-0041283-2
17.Brisswalter J, Hausswirth C, Smith D, Vercruyssen F,
Vallier JM. Energetically optimal cadence vs. freelychosen cadence during cycling: effect of exercise duration. Int J Sports Med. 2000;21(1):60–64. PubMed
doi:10.1055/s-2000-8857
18. Emanuele U, Denoth J. Influence of road incline and body
position on power–cadence relationship in endurance
cycling. Eur J Appl Physiol. 2012;112(7):2433–2441.
doi:10.1007/s00421-011-2213-8. PubMed
19. Baron B, Noakes TD, Dekerle J, et al. Why does exercise
terminate at the maximal lactate steady state intensity? Br J
Sports Med. 2008;42(10):828–833. PubMed doi:10.1136/
bjsm.2007.040444
20.Tegtbur U, Busse MW, Braumann KM. Estimation of
an individual equilibrium between lactate production
and catabolism during exercise. Med Sci Sports Exerc.
1993;25(5):620–627. PubMed
21. Fontana P, Boutellier U, Knopfli-Lenzin C. Time to exhaustion at maximal lactate steady state is similar for cycling
and running in moderately trained subjects. Eur J Appl
Physiol. 2009;107(2):187–192. PubMed doi:10.1007/
s00421-009-1111-9
22.Coleman DA, Wiles JD, Davison RC, Smith MF,
Swaine IL. Power output measurement during treadmill
cycling. Int J Sports Med. 2007;28(6):525–530. PubMed
doi:10.1055/s-2006-955888
23. Marsh AP, Martin PE. The relationship between cadence
and lower extremity EMG in cyclists and noncyclists. Med
Sci Sports Exerc. 1995;27(2):217–225. PubMed
24. Nielsen JS, Hansen EA, Sjogaard G. Pedalling rate affects
endurance performance during high-intensity cycling.
Eur J Appl Physiol. 2004;92(1-2):114–120. PubMed
doi:10.1007/s00421-004-1048-y
25. MacIntosh BR, Svedahl K, Kim M. Fatigue and optimal
conditions for short-term work capacity. Eur J Appl
Physiol. 2004;92(4-5):369–375. PubMed doi:10.1007/
s00421-004-1177-3
26. Foss O, Hallen J. The most economical cadence increases
with increasing workload. Eur J Appl Physiol. 2004;92(45):443–451. PubMed doi:10.1007/s00421-004-1175-5
27. Argentin S, Hausswirth C, Bernard T, et al. Relation between
preferred and optimal cadences during two hours of cycling
in triathletes. Br J Sports Med. 2006;40(4):293–298, discussion 298. PubMed doi:10.1136/bjsm.2005.020487
28.Jobson SA, Nevill AM, George SR, Jeukendrup AE,
Passfield L. Influence of body position when considering
the ecological validity of laboratory time-trial cycling performance. J Sports Sci. 2008;26(12):1269–1278. PubMed
doi:10.1080/02640410802183585
29. Sassi A, Rampinini E, Martin DT, Morelli A. Effects of
gradient and speed on freely chosen cadence: the key role
of crank inertial load. J Biomech. 2009;42(2):171–177.
PubMed doi:10.1016/j.jbiomech.2008.10.008
30. Lucia A, Hoyos J, Chicharro JL. Physiology of professional
road cycling. Sports Med. 2001;31(5):325–337. PubMed
doi:10.2165/00007256-200131050-00004