Content uploaded by Teun Van Erp
Author content
All content in this area was uploaded by Teun Van Erp on Jan 12, 2021
Content may be subject to copyright.
Downloaded from https://journals.lww.com/acsm-msse by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3POoLCqWplEemA1mp72O43oKT4GeHhxOY7sddx0MZyNk= on 04/01/2020
Downloadedfromhttps://journals.lww.com/acsm-msse by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3POoLCqWplEemA1mp72O43oKT4GeHhxOY7sddx0MZyNk= on 04/01/2020
Case Report: Load, Intensity, and Performance
Characteristics in Multiple Grand Tours
TEUN VAN ERP
1
, MARCO HOOZEMANS
1
, CARL FOSTER
2
, and JOS J. DE KONING
1,2
1
Department of Human Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam Movement Sciences, Amsterdam,
THE NETHERLANDS; and
2
Department of Exercise and Sport Science, University of Wisconsin –La Crosse, La Crosse, WI
ABSTRACT
VAN ERP, T., M. HOOZEMANS, C. FOSTER, and J. J. DE KONING. Case Report: Load, Intensity, and Performance Characteristics in
Multiple Grand Tours. Med. Sci. Sports Exerc.,Vol.52,No.4,pp.868–875, 2020. Introduction: The aim of this study was to present the
load, intensity, and performance characteristics of a general classification (GC) contender during multiple grand tours (GTs). This study also inves-
tigated which factors influence climbing performance. Methods: Power output (PO) data were collected from a GC contender from the Vuelta a
España 2015, the Giro d’Italia 2017, the Giro d’Italia 2018, and the Tour de France 2018. Load (e.g., Training Stress Score and kJ spent) and intensity
in five PO zones were quantified. One-way ANOVA was used to identify differences between the GTs. Furthermore, performance during the four
GTs was quantified based on maximal mean PO (W·kg
−1
) over different durations and by the relative PO (W·kg
−1
) on the key mountains in the
GTs. Stepwise multiple regression analysis was used to identify which factors influence relative PO on the key mountains. Results: No significant
differences were found between load and intensity characteristics between the four GTs, with the exception that during the Giro d’Italia 2018, a sig-
nificantly lower absolute time was spent in PO zone 5 (P= 0.005) compared with the other three GTs. The average relative PO on the key mountains
(n= 33) was 5.9 ± 0.6 W·kg
−1
and was negatively influenced by the duration of the climb and the total elevation gain before the key mountain,
whereas the gradient of the mountain had a positive effect on relative PO. Conclusions: The physiological load imposed on a GC contender did
not differ between multiple GTs. Climbing performance was influenced by short-term fatigue induced by previous altitude meters in the stage and
the duration and gradient of the mountain. Key Words: GENERAL CLASSIFICATION CONTENDER, PROFESSIONAL, TEAM LEADER,
PERFORMANCE, POWER OUTPUT
Elite road cycling is one of the most physically demanding
sports in the world, especially the 3-wk multistage races:
the Giro d’Italia, the Tour de France, and the Vuelta a
España. A grand tour (GT) typically contains 21 race-days
with only two or three rest days, covering between 3500 and
4000 km in 85–95 h of competition (1). For elite cyclists, to
participate in one of the three GTs is the highlight of the sea-
son and winning a GT is the highest achievement possible in
professional cycling. A GT is a complex race with different
competition elements and therefore different kinds of stages
(e.g., flat, semimountainous, and mountain stages) (2,3).
The introduction of the heart rate and power output (PO)
monitors ensures that more information on the load and
intensity demands of participating in a GT is obtained (2–8).
In 2003, Lucia et al. (4) showed with heart rate data of seven
cyclists the load and intensity demands of two different GTs
(the Tour de France and the Vuelta a España) and did not find
any significant differences between the load and intensity de-
spite the assumed different nature of both races. Vogt et al. (2)
and Sanders and Heijboer (3) discussed, on the basis of heart rate
and PO, the load and intensity characteristics of different terrains
(e.g., flat, semimountain, and mountain stage) in GTs (2,3). In ad-
dition, Rodriguez-Marroyo et al. (5) showed that fatigue sup-
presses the heart rate during a GT and was able to correct the
load and intensity demands of the Vuelta a España by adjusting
heart rate thresholds based on prelaboratory and postlaboratory
tests. Although multiple studies report load and intensity charac-
teristics of GTs, it is still unknown what the load and intensities
are for a General Classification (GC) contender.
Winning a GT is the highest achievement possible for an
elite cyclist and only accomplished by the best and most com-
plete cyclists. With exceptions from (team) time trials (TTs),
differences in the GCs are mostly made by performances up-
hill, which can be defined by the relative PO during that climb.
Therefore, relative PO on the last mountain in a stage is very
important for a GC contender. Pinot and Grappe (9) presented
Address for correspondence: Teun van Erp, Ph.D., Department of Human
Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam Movement
Sciences, Amsterdam, the Netherlands; E-mail: teunvanerp@hotmail.com.
Submitted for publication May 2019.
Accepted for publication October 2019.
0195-9131/20/5204-0868/0
MEDICINE & SCIENCE IN SPORTS & EXERCISE
®
Copyright © 2019 by the American College of Sports Medicine
DOI: 10.1249/MSS.0000000000002210
868
APPLIED SCIENCES
Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
the relative mean maximal PO over different durations
(15–18,000 s), the so-called record PO (9) or maximal power
profile (MPP) (10) of a GC contender achieved in training
and competition. In addition, Sanders and Heijboer (3) pre-
sented the MPP values of nine riders during the Giro d’Italia.
Both studies give a rare glimpse of the performances necessary
to compete in a GT. However, PO values on the key moun-
tains for a GC contender in a GT are still unknown.
Several factors can influence the performance of GC con-
tender on the key mountains during a GT. Because of the very
demanding nature of GTs, elite cyclists accumulate fatigue over
the 3 wk of racing, and this may reflect in several physiological
and psychological changes that affect performance (4,11–13).
Recently, Rodriguez-Marroyo et al. (5) showed that long-term fa-
tigue accumulated by professional cyclists over 21 d of racing re-
sulted in a performance decline of ~10% and affected maximal
and submaximal endurance performance. Furthermore, it is ex-
pected that fatigue sustained during the stage (short-term fatigue)
has an influence on the performance at the end of a stage (14). Fur-
thermore, mountain characteristics such as the length and the gra-
dient of the mountain will have an influence on the performance
and thus on relative PO (7,15). Lastly, environmental conditions
such as heat (16) and altitude (17) could impair performances.
Although there is growing interest in the load, intensity, and
performance characteristics of elite cyclists competing in GTs
(2–5), the reported values in the literature are based on the av-
erage of multiple cyclists. In GTs, teams select different types
of riders, which have different roles throughout the race (e.g.,
sprinter, domestic, and TT specialist) and thus do not have to
perform all 21 stages at their top level (18,19). This contrasts
with a GC contender who must perform every day to prevent
losses in time. Therefore, load and intensity demands for a
GC contender could be higher than the reported values for the av-
erage team. However, there is evidence of “pacing”across the
duration of GTs, whereby GC riders only ride at maximal effort
during a few stages (20). In addition, to the best of the authors’
knowledge, no study reported performance characteristics of a
GC contender or factors that could influence the performance
of a GC contender during a GT. Therefore, our research question
was threefold: first, to present the load and intensity characteris-
tics of a GC contender for multiple GTs; second, to present the
relative PO, which is needed on the key mountains to compete
for the victory in multiple GTs; and third, to examine which fac-
tors are associated with the relative PO on the key mountains.
METHODS
Participant. The athlete is a professional cyclist (24 yr,
69.9 kg, 185 cm) competing in the Union Cycliste Internationale
World Tour Series and provided written informed consent for a
detailed analysis of his PO data collected during the Vuelta a
España 2015, the Giro d’Italia 2017, the Giro d’Italia 2018, and
the Tour de France 2018, where the athlete competed for the
GC victory. During the 7 yr that the athlete was a professional cy-
clist, he was highly successful and in 2017 and 2018 ranked as a
top 10 cyclist in the world according to the World Tour ranking.
The main specialty of the athlete are TTs, where he became
World Champion in 2017 (individual and team TT) and won a
silver medal at the 2016 Olympics. Furthermore, the athlete
won 2 GCs in the World Tour, a 7-d race: the BinckBank Tour
in 2017 and the Giro d’Italia in 2017. Until 2019, the athlete
won 19 races including TTs and road races in all the GTs.
Research design. During four GTs, PO data were col-
lected from as many as possible stages and uploaded to a cen-
tral database. Because of sponsor changes, the collection of the
PO data was done with different brands of power meters and
with different power meters on different bikes (i.e., one road
bicycle, two reserve road bicycles, and two TT bicycles).
The PO data from the Vuelta a España 2015 and the Giro
d’Italia 2017 were collected by Pioneer power meters (SGY-
PM910H2; Pioneer, Kawasaki, Japan) and from the Giro
d’Italia 2018 and the Tour de France 2018 by Shimano power
meter (FC-RC9100-P; Shimano, Sakia, Japan). The athlete
and mechanics were informed about the importance of the zero
calibration and were instructed to do the zero calibration be-
fore every ride. Because of malfunctions, crashes and bicycle
changes PO data were not recorded for two stages at the Vuelta
a España 2015 (i.e., stages 9 and 19), two stages at the Giro
d’Italia 2017 (i.e., stages 3 and 4), two stages at the Giro d’Italia
2018 (i.e., stages 4 and 21), and two stages at the Tour de France
2018 (i.e., stages 2 and 13). All collected PO data were manually
checked, and spikes were manually corrected when necessary.
Dataweresampledat1HzduringtheVueltaaEspaña2015,
the Giro d’Italia 2017, and the Tour de France 2018. Using an-
other bicycle computer during the Giro d’Italia 2018, the data
were collected at 0.2 Hz. All bicycle computers measured race
characteristics (i.e., distance, duration, and elevation (gain))
and environmental characteristics (i.e., temperature).
Load and intensity characteristics. Load characteris-
tics are based on PO and measured as the total mechanical en-
ergy spent (in kilojoules spent) and Training Stress Score
(TSS). (10) TSS was calculated according to equation 1.
TSS ¼tNP IF
FTP 3600
100 ½1
where tis the duration of the race in seconds and IF the intensity
factor (see equation 3). NP is the normalized power as calculated
with equation 2, where p
i
is the floating mean power for 30-s time
segments and Nis the total number of time segments. The func-
tional threshold power (FTP) was determined as 95% of the
highest 20-min mean maximal PO obtained in races from that par-
ticular season (10). FTP was established at 408 W (5.8 W·kg
−1
)
for the Vuelta a España 2015, 409 W (5.8 W·kg
−1
)fortheGiro
d’Italia 2017, 417 W (5.9 W·kg
−1
) for the Giro d’Italia 2018,
and 417 W (6.0 W·kg
−1
) for the Tour de France 2018.
NP ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
NX
N
i¼1
p4
i
4
v
u
u
t½2
IF ¼NP
FTP
½3
ELITE CYCLISTS COMPETING IN CYCLING GRAND TOURS Medicine & Science in Sports & Exercise
®
869
APPLIED SCIENCES
Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
In addition, similar to previous research (21,22), load metrics
(TSS and kJ spent) are expressed relatively per kilometer (kJ
spent·km
−1
and TSS·km
−1
). Intensity distribution was quanti-
fied based on the time spent in five different PO zones. The
five PO zones were based on a percentage of FTP based on
guidelines provided by Allen and Coggan (10): zone 1,
≤55% of FTP; zone 2, 56%–75% FTP; zone 3, 76%–90%
FTP; zone 4, 91%–105% FTP; and zone 5, ≥106% FTP.
Performance characteristics. To assess the perfor-
mance during the four GTs, the MPP and the relative PO on
the key mountains were analyzed. The MPP corresponded
to the highest mean maximal power developed in each GT
by the athlete for the durations of 5, 10, and 30 s and 1, 5,
10, 20, 60, 120, and 180 min. The MPP was expressed in relation
to body mass of the cyclist (W·kg
−1
) to compensate for changes
in body mass. Body mass was measured by the team for the four
GTs, and the mean body masses were 69.9, 70.3, 70.5, and
69.0 kg for the Vuelta a España 2015, the Giro d’Italia 2017,
the Giro d’Italia 2018, and the Tour de France 2018, respectively.
To analyze the performance on the key mountains in the
GTs, the last mountain in a stage with an uphill finish and
mountains with a significant importance (e.g., finish directly
after descend) were selected by the use of www.touretappe.
nl (23) and www.procyclingstats.com. (24) Based on visual
inspection of the PO, speed, and altitude data, the key moun-
tains were manually selected and PO data were collected and
expressed in relation to the athlete’sbodymass(W·kg
−1
).
Factors that influence performance. To investigate
the effect of long-term fatigue on the climbing performance
at the key mountains, we reported on which race-day the key
mountains were in the GT (stage number). To determine the
effect of short-term fatigue during the stage, the duration, dis-
tance, total elevation gain (TEG), kJ spent, TSS, kJ spent·km
−1
,
and TSS·km
−1
were measured from the start of the stage till the
start of the key mountain. Furthermore, to investigate the influ-
ence of mountain characteristics on relative PO, we measured
the duration and the gradient of the key mountains. Finally, the
temperature and average altitude measured by the bike computer
on the key mountain were analyzed to investigate the effect of
heat stress and altitude on climbing performance.
Statistical analysis. Descriptive data were reported as
mean (±SD). All parameters except stage number were col-
lected by analyzing the collected PO data from the four GTs.
Differences between the four GTs were determined using
one-way ANOVA. Bonferroni’spost hoc test was applied to
identify differences when the ANOVA indicated a significant
main effect. Stepwise multiple linear regression (SMLR) analysis
was used to identify the best predictors of the relative PO on the
key mountains. Before SMLR, the collinearity between variables
was determined by Pearson correlations. For variables with a cor-
relation of r> 0.7, the variable with the highest correlation with
performance (relative PO on the key mountain) was used for
SMLR. Based on a visual inspection of the relation between du-
ration of the climb and relative PO, duration was logarithmically
transformed for better fitting the data. Regression coefficients
(intercept and slopes) are presented, and uncertainties in the
coefficients are presented as 95% confidence intervals (CI).
PO analysis was performed using Golden Cheetah (Golden
Cheetah, Version 3.4), and statistical analysis was performed
using SPSS (IBM SPSS Statistics version 23; IBM Corpora-
tion, Armonk, NY). The level of statistical significance was
set at P< 0.05. The following criteria were adopted to interpret
the magnitude of the correlation (r) between the measures: <0.1,
trivial; 0.1–0.3, small; 0.3–0.5, moderate; and >0.5, large (25).
RESULTS
In total, 76 stages were analyzed, which were collected dur-
ing the Vuelta a España 2015 (n= 19), the Giro d’Italia 2017
(n= 19), the Giro d’Italia 2018 (n= 19), and the Tour de
France 2018 (n= 19). In the GCs, the athlete finished sixth,
first, second, and second, respectively. In total, the athlete fin-
ished in 14 stages in the top 3, from which he won 6 stages
(two road races and four TTs) and was GC leader for 5, 9,
and 1 d in the Vuelta a España 2015, the Giro d’Italia 2017,
and the Giro d’Italia 2018, respectively.
Table 1 presents basic descriptive and load characteristics
for the four GTs. No significant differences were observed be-
tween the four GTs.The GT with the lowest load recorded was
the Tour de France 2018 and the highest load recorded was the
Giro d’Italia 2018; the differences between both of them were
6% and 11% for kJ spent and TSS, respectively.
Figure 1 presents the relative and absolute time spent in the five
PO zones. Post hoc test analyses revealed a lower absolute time
spent in PO zone 5 during the Giro d’Italia 2018 (334 ± 8 min)
compared with the Vuelta a España 2015 (531 ± 9 min;
P= 0.005), the Giro d’Italia 2017 (533 ± 11 min; P= 0.005),
and the Tour de France 2018 (493 ± 9 min; P= 0.039). No sig-
nificant differences were observed between the other PO
zones in the four GTs.
TABLE 1. Average load and intensity characteristics for the GTs where the athlete competed for the GC victory (mean ± SD).
Vuelta a España 2015 (nStages = 19) Giro d’Italia 2017 (nStages =19) Giro d’Italia 2018 (nStages =19) Tour de France 2018 (nStages =19)
Distance, km 168 ± 53.5 181 ± 55.1 176 ± 58.2 167 ± 55.4
Duration, min 260 ± 85 272 ± 95 266 ± 91 255 ± 86
PO,W 227±36.2 203±43.2 238±47.2 228±36.5
PO, W·kg
−1
3.25 ± 0.52 2.89 ± 0.61 3.38 ± 0.67 3.30 ± 0.53
TEG, m 2413 ± 1397 2583 ± 1438 2245 ± 14784 2076 ± 1436
Total kJ spent, kJ 3623 ± 1196 3643 ± 1185 3770 ± 1359 3530 ± 1192
Total TSS™233±80.5 237±84.9 234±91.6 209±74.6
kJ spent·km
−1
22.4 ± 4.8 21.4 ± 5.9 24.0 ± 11.7 22.7 ± 7.6
TSS·km
−1
1.48 ± 0.46 1.48 ± 0.75 1.39 ± 0.38 1.38 ± 0.60
No significant difference between GTs is found (P>0.05).
Norm, normalized.
http://www.acsm-msse.org870 Official Journal of the American College of Sports Medicine
APPLIED SCIENCES
Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
Figures 2 and 3 and Table 2 give a detailed overview of the
athlete’s performance during the four GTs. Figure 2 presents
the relative MPP (5, 10, and 30 s and 1, 5, 10, 20, 60, 120,
and 180 min), and Figure 3 presents the relative PO of all
key mountains. Table 2 provides a summary of the key moun-
tains and themountain characteristics.The average relative PO
on all the key mountains for all four GTs combined was
5.9 ± 0.6 W·kg
−1
.
Table 3 presents the correlation matrix between all the fac-
tors with a moderate correlation with relative PO on the key
mountain or factors identified by SMLR. SMLR analysis
showed that in the final regression model climbing perfor-
mance was significantly associated with a combination of the du-
ration (log
10
transformed; −1.52 W·kg
−1
(95% CI, −1.81 to
−1.22), r
2
= 0.61, P= 0.0001), the TEG before the mountain
(·10
−3
); −0.23 W·kg
−1
(95% CI, −0.32 to −0.14), r
2
= 0.13,
P= 0.0001), and gradient of the mountain (0.12 W·kg
−1
(95%
CI,0.07to0.17),r
2
= 0.11, P= 0.0001), whereas the intercept
was 7.64 W·kg
−1
(95% CI, 7.11 to 8.17). A total of 86% of the
variance of the relative PO on the key mountain can be explained
by those three factors, described in equation 4.
PO Wkg−1
¼7:64 þlog10 duration mountain minðÞðÞ
ð−1:52 þgradient mountain %ðÞÞ
ð0:12 −TEG before mountain mðÞ10−3
Þ0:23 ½4
DISCUSSION
This is, to the authors’knowledge, the first study describing
the individual load, intensity, and performance characteristics
of a GC contender during multiple GTs. The presented data
cover one GT in which the athlete finished first and two GTs
where the athlete finished second. Therefore, the present study
gives a unique insight into the load, intensity, and performance
characteristics of the fight for a GT victory. Furthermore,
SMLR analysis showed that the climbing duration and the gra-
dient of the key mountain combined with indicators of short-
term fatigue determined 86% of the variance of the relative
PO on the key mountain.
Load and intensity. An impressive energy expenditure
during racing between 74,123 and 79,166 kJ or a TSS between
4400 and 4983 au is necessary to finish a GT in the top 10. Sim-
ilar to Lucia et al. (4), this study did not find any significant dif-
ferences between the load of the four GTs despite the fact that
all GTs have different combinations of flat, semimountain,
and mountain stages and (team) TTs. The largest load differ-
ences occurred between the Giro d’Italia 2018 and Tour de
France 2018, which was 6% and 11% for kJ spent and TSS, re-
spectively. The differences between kJ spent and TSS are prob-
ably caused by the quadratic relation with exercise intensity,
which is integrated within the calculation of TSS (26). Our data
showed that despite competing for the GC, ~80% of the time in
FIGURE 2—MPP from a GC contender during different GTs. The values indicate the maximal PO for different durations (5, 10, and 30 s and 1, 5, 10, 20,
60, 120, and 180 min) achieved in the GTs participated by the athlete and are expressed in relation to the body mass of the athlete.
FIGURE 1—Intensity distribution expressed as relative (A) and absolute(B) time spent in different PO zones in a GC contender during different GTs. Ab-
solute time spent in PO zone 5 during the Giro d’Italia 2018 is significantly (P< 0.05) different from the other GTs.
ELITE CYCLISTS COMPETING IN CYCLING GRAND TOURS Medicine & Science in Sports & Exercise
®
871
APPLIED SCIENCES
Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
a GT is spent in the low-intensity zones (PO zones 1, 2, and 3).
Furthermore, ~10% is spent around FTP (PO zone 4) and ~10%
is spent in the highest PO zone or above FTP (PO zone 5). The
only significant difference found between the four analyzed
GTs was a lower absolute time spent in PO zone 5 during the
Giro d’Italia 2018 (334 ± 8.3 min) compared with the other
three GTs. The lack of significant different load and intensity
between the four analyzed GTs could suggest that GC con-
tenders unconsciously pace their efforts during a GT (4,20).
For example, the Giro d’Italia 2017 had the highest amount of dis-
tance, duration, and TEG, whereas the PO, intensity factor, and kJ
spent·km
−1
were the lowest of all the four studied GTs. The lower
time spent in PO zone 5 during the Giro d’Italia 2018 could be
caused by the higher overall pace during this GT. This is shown
by the higher PO and intensity factor. This is strengthened by
the results of the regression analysis that TEG before the key
mountain had a decremental effect on relative PO. A high amount
of elevation gains in a GT results in a lower PO on the key moun-
tains and thus a lower load and intensity on the key mountains.
This will result in an overall lower load and intensity demand
for the specific GT. This is in agreement with Lucia et al. (4),
who stated that a shorter or less mountainous GT is compensated
by a higher relative intensity compared with a GT that covers a lot
of kilometers or altitude meters.
Performances. Reported MPP values in this study are
higher compared with the values reported by Sanders and
Heijboer (3), which is probably caused by averaging the
values of a whole team (Fig. 2). Cyclists have different roles
during a GT (e.g., domestique, sprinter, GC contender, and
TT specialist), and different roles mean different priorities.
For example, a GC contender must perform at his maximal
on some moments during all the 21 stages, whereas sprinters
and their domestiques pace mountain stages to be within the
time limit and therefore do not have to compete maximal every
stage. Thus, averaging MPP values of GC contender and
domestiques will result in lower MPP values compared with
MPP values of a GC contender alone. In agreement with Pinot
and Grappe (9), the term “record PO”is preferred to describe
performance instead of “mean maximal PO,”because the
highest PO obtained during competition is not the maximal
that can be achieved by the athlete. This could also be the rea-
son why the values reported in this study are slightly lower
compared with Pinot and Grappe (9). The MPP values pre-
sented in this study were affected by fatigue, tactical racing,
and the ability to achieve a maximal effort exactly for each du-
ration, whereas MPP values achieved during training are not
influenced by the factors mentioned previously (Fig. 2). For
the first time, climbing performances of a GC contender in a
GT are described. On average, an impressive 5.9 ± 0.6 W·kg
−1
for 27 ± 13 min is necessary on the key mountains to compete
for the GC victory. The presented data are from a GC contender
specialized in TTs, and thus, he gained time on his direct compet-
itor (i.e., first or second place in the GT) in the specific GTs dur-
ing those TTs (i.e., 113, 257, 50, and 14 s, in the respective GTs).
Therefore, this GC contender can have a more defensive race
strategy throughout the mountain stages because his strategy is
not to lose time in these stages and gain time during the TTs.
Thus, it could be that the described climbing values are even
FIGURE 3—Relative PO at different key mountains in a GCcontender during multiple GTs(Alto de la Mesa
1
;AltodeCazorla
2
,AltodeCapileira
3
,AltoEls
Cortals d’Encamp
4
,AltoCampoo
5
, Alto de Sotres
6
, Alto Ermita de Alba
7
, Puerto de Cotos
8
, Blockhaus
9
,Oropa
10
,Selvino
11
, Umbrailpass
12
,Pontives
13
,
Piancavallo
14
, Asiogo
15
,Etna
16
, Montevergine di Mercogliano
17
, Gran Sasso d’Italia
18
,Osimo
19
,Zoncolan
20
,Sappada
21
, Pratonevoso
22
, Bardonecchia
23
,Cervinia
24
,
Mur d Bretagne
25
, Col de la Colombiere
26
, La Rosiere
27
,Alped’Huez
28
,Mende
29
,PicduNore
30
, Col du Portillon
31
, Saint-lary Soulan
32
,Cold’Aubisque
33
).
TABLE 2. Mountain characteristics and PO from the key mountains for the GTs where the
athlete competed for the GC victory (mean ± SD).
VueltaaEspaña
2015 (n=8)
Giro d’Italia
2017 (n=7)
Giro d’Italia
2018 (n=9)
Tour de France
2018 (n=9)
Distance, km 10.0 ± 6.0 11.0 ± 3.4 9.4 ± 4.3 8.4 ± 4.5
Duration, min 27.5 ± 13.8 31.5 ± 11.8 25.4 ± 12.4 23.7 ± 14.5
PO,W 407±43.8 398±25 427±51 405±50
PO, W·kg
−1
5.8±0.6 5.7±0.4 6.0±0.7 5.9±0.7
Gradient,% 7.8±1.9 7.2±1.1 7.0±2.8 7.3±1.6
Temperature, °C 23.7 ± 7.8 24.6 ± 2.6 17.1 ± 4.5 24.7 ± 3.4
Stage number 11.4 ± 6.0 15.9 ± 3.7 13.3 ± 4.6 13.3 ± 4.0
Altitude, m 1041 ± 464 982 ± 437 1192 ± 541 1118 ± 428
http://www.acsm-msse.org872 Official Journal of the American College of Sports Medicine
APPLIED SCIENCES
Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
higher in GC contenders specialized in climbing. Furthermore,
this GC contender was GC leader for multiple days in the
Vuelta a España 2015 (5 d) and the Giro d’Italia 2017
(9 d). Defending the GC lead could result in a different race
strategy, and thus a different MPP or intensity distribution,
compared with a rider challenging the race leader.
Surprisingly, during the GT won by this athlete, the Giro
d’Italia of 2017, the athlete had a slightly lower performance
with 5.7 ± 0.4 W·kg
−1
on the key mountains compared with
the other GTs (Table 2). This could be explained by the results
of the SMLR analyses. During this GT, the average duration
on the key mountains was longer. As shown by SMLR analy-
sis, duration has a large negative influence on the relative PO
on the key mountains. It was expected that the duration of the
mountain has a large influence on climbing performance as it
is well established that PO is influenced by duration, described
by the power–duration relationship (27,28). This power–
duration relationship is also clearly visible in the MPP
(Fig. 2). Surprisingly, from all the parameters to assess short-
term fatigue, TEG before the key mountain has the largest in-
fluence on the performance on the key mountains, whereas
load measurements before the key mountain (i.e., TSS and
kJ spent) did not have any significant influence. However,
load values expressed per kilometer (i.e., TSS·km
−1
and kJ
spent·km
−1
) before the key mountain did show a moderate re-
lationship with performance on the key mountains (Table 3).
In addition, collinearity analysis showed that they could be in-
terchangeable used for TEG before the key mountain in the
SMLR analysis. Thus TSS·km
−1
and kJ spent·km
−1
may be
good indicators for short-term fatigue. One of the reasons
that relative load values (TSS·km
−1
and kJ spent·km
−1
)are
better indictors for short-term fatigue compared with abso-
lute load values may be that professional cyclists are highly
endurance trained and a long low-intensity race will not
cause the same amount of fatigue as a shorter high-intensity
race. The load measurements relative to distance (TSS·km
−1
and kJ spent·km
−1
) are in line with previously reported mea-
surements (3,22).
Furthermore, gradient had a large effect on the relative PO
on the key mountain; 14.5% of variance of the PO was
determined by the gradient of the mountain. A 1% increase
in steepness means 0.12 W·kg
−1
higher relative PO on the
mountain. This is in agreement with Padilla et al. (7), which di-
vided mountains during the three GTs into three categories
based on the length and steepness and found higher PO on
the steeper category mountains compared with the less steeper
category mountains. One of the reasons could be that a steeper
mountain means less drafting and thus less help from
domestiques. Therefore, gradient of the mountain also influ-
ences race tactics, which indirectly influence PO. Further-
more, a mountain with a lower gradient could mean some
flatter easier parts, and thus, a more stochastic PO as drafting
behind domestiques has a bigger effect with high speeds.
One of the hypotheses of this study was that long-term fa-
tigue (stage number) would influence the performances on
the key mountain in a GT. However, stage number did not sig-
nificantly (P= 0.12) influence relative PO on the key moun-
tain. This is not in line with previous research that found a
decrement of performances after a GT of ~10% comparing a
laboratory exercise test before and after a GT in non-GC con-
tenders (5). Although not significant, the SMLR analyses indi-
cate a PO decrement of −0.015 W·kg
−1
·d
−1
in a GT, which
means a difference of ~5% between a performance on the first
day and the last day of a GT. One of the specific determinants
of a successful GC contender is the ability to sustain accumu-
lating load while holding a high-performance level. In the
study of Rodriguez-Marroyo et al. (5), cyclists with different
specialties (none GC contenders) were studied, which could be
a reason for the lower decrement in performance described in this
study. Furthermore, a GC contender is protected by domestiques
throughout the whole GT to save energy. In addition, we assume
that with a higher number of climbs, long-term fatigue will be of
significant influence on the performance on the key mountain.
Another reason that stage number did not significantly influence
performance is that most GTs have their mountain stages in week
2 or week 3, and therefore, almost all performances analyzed in
this study were already affected to some extent by long-term fa-
tigue. Lastly, it could be that fatigue influences the race intensity
before the key mountain and thus had a smaller effect on the per-
formance on the last mountain.
TABLE 3. Correlation matrix for parameters with a moderate(r> 0.3) association with performance on the key mountain or with a significant influence on performance based on stepwise mul-
tiple regression analysis.
Performance Mountain Characteristics Long-Term Fatigue Short-Term Fatigue
PO, W·kg
−1
Duration, Min Gradient, % Altitude, m Stage Number TEG kJ Spent·km
−1
TSS·km
−1
Performance
PO, W·kg
−1
—
Mountain characteristics
Duration, min −0.78 —
Gradient, % 0.27 0.11 —
Altitude, m −0.62 0.44 −0.02 —
Long-term fatigue
Stage number −0.44 0.17 −0.06 0.30 —
Short-term fatigue
TEG −0.45 0.03 0.09 0.63 0.23 —
kJ spent·km
−1
−0.40 0.01 −0.01 0.55 0.26 0.72
a
—
TSS·km
−1
−0.40 −0.01 −0.08 0.49 0.29 0.72
a
0.97
a
—
Performance, the relative PO on the last mountain; Mountain characteristics, duration and gradient of the last mountain; Stage number, amount of day sin the GT; Short-term fatigue,
TEG, kJ spent·km
−1
, and TSS·km
−1
measured before the last mountain.
a
Amoderateorhighercollinearity.
ELITE CYCLISTS COMPETING IN CYCLING GRAND TOURS Medicine & Science in Sports & Exercise
®
873
APPLIED SCIENCES
Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
Our hypothesis was that a high environmental temperature
would negatively influence performance on the key mountain
because it is well known that heat stress impairs performance
(16). This study did not find any significant influence of tem-
perature on performance on the key mountain. The reason for
this could be that only four mountains were recorded with a
temperature above 30°C. From those four mountains, two
mountains were shorter than 10 min, and thus, heat stress
would probably not influence the performance on those
2 mountains. Furthermore, we hypothesized that altitude will
negatively influence the PO on the key mountains, as it is well
known that hypoxia impairs performance (17). However, in
the present study, we did not find any significant influence
of altitude on the PO at the key mountains. From the 33 ana-
lyzed mountains, the average altitude was ~1000 m, from
which only 7 mountains were on average above 1500 m. This
could result in a decline in PO of approximately 5.3% and
8.7% at 1000 and 1500 m, respectively (29). However, the
standard preparation of the athlete for a GT was a high-
altitude training camp for a minimum of 14 d, which reduces
the influence of altitude significantly to approximately 3.0%
at 1000 m and 5.4% at 1500 m (29). Therefore, it is likely that
the influence of altitude is somewhat limited because of accli-
matization. Furthermore, race tactics and averaging the PO of
the whole mountain could also blunt the effect of altitude.
Limitations. The presented data were primarily collected
for monitoring of training load of the athlete and are therefore
not without limitations. The measurements of PO were done
with two different brands of power meters and the athlete
had multiple bikes (one road bicycle, two reserve road bicycles,
and two TT bicycles) during the GTs, and therefore, the pre-
sented data of the GTs were collected with different power me-
ters. The mechanics of the team had the task to zero calibrate the
power meter every day before the race, although the authors did
not control this. Furthermore, within the team, no in-house cal-
ibration was performed after receiving the power meter from the
manufacturers. Because of malfunctions, crashes, and bicycle
changes, PO was not measured during eight stages divided
equally over the four GTs. Furthermore, four different brands
of bicycle computers were used to collect the altitude data
which can influence the measurements of altitude (30). In addi-
tion, FTP was obtained by taking 95% of the highest 20-min
mean maximal PO obtained during the particular season.
Nimmerichter et al. (31) showed that variations within a season
can be up to 0.4 W·kg
−1
. Furthermore, the correction factor of
95% is a somewhat arbitrary choice, and it is shown that a low
or high anaerobic capacity has an influence on this correction
factor (32). Both limitations of FTP could have an influence
on the presented TSS values and intensity zones. Furthermore,
ideally, PO intensity zones would be anchored around physio-
logical thresholds, such as the first and second lactate or ventila-
tory thresholds (33,34). However, during the time of the data
collection, no regular and controlled laboratory exercise testing
was implemented within the team. Therefore, in this study, the
boundaries of the intensity zones are based on Coggan and Allen
(10), which are somewhat arbitrary chosen, and the boundaries
are not equally spaced. This could influence the reported inten-
sity distribution. Lastly, the key mountains were manually se-
lected based on visual inspection of PO, speed, and altitude
profile and could therefore be slightly different in length com-
pared with the official length of the mountains.
CONCLUSIONS
To conclude, overall load and intensity characteristics in four
different GTs did not differ when competing for the GC. An im-
pressive 5.7 to 6.0 W·kg
−1
on the key mountains is necessary to
compete for the victory in multiple GTs. Short-term fatigue
in combination with the gradient and duration of the climb de-
termines 86% of the variance of the relative PO on the key
mountains.
The authors would like to thank the cyclist for his participation in this
investigation. Theauthors declare that they have no conflict of interest.
No funding is used for this research. The results of the study are pre-
sented clearly, honestly, and without fabrication, falsification, or inap-
propriate data manipulation. The results of the present study do not
constitute endorsement by the American College of Sports Medicine.
REFERENCES
1. Lucia A, Hoyos J, Chicharro JL. Physiology of professional road cy-
cling. Sports Med. 2001;31(5):325–37.
2. Vogt S, Schumacher YO, Roecker K, et al. Power output during the
Tour de France. Int J Sports Med. 2007;28(9):756–61.
3. Sanders D, Heijboer M. Physical demands and power profile of
different stage types within a cycling grand tour. Eur J Sport
Sci. 2018;1–9.
4. Lucia A, Hoyos J, Santalla A, Earnest C, Chicharro JL. Tour de
France versus Vuelta a Espana: which is harder? Med Sci Sports
Exerc. 2003;35(5):872–8.
5. Rodriguez-Marroyo JA, Villa JG, Pernia R, Foster C. Decrement in
professional cyclists’performance after a grand tour. Int J Sports
Physiol Perform. 2017;12(10):1348–55.
6. Padilla S, Mujika I, Orbananos J, Angulo F. Exercise intensity during
competition time trials in professional road cycling. Med Sci Sports
Exerc. 2000;32(4):850–6.
7. Padilla S, Mujika I, Santisteban J, Impellizzeri FM, Goiriena JJ. Ex-
ercise intensity and load during uphill cycling in professional 3-week
races. Eur J Appl Physiol. 2008;102(4):431–8.
8. Padilla S, Mujika I, Orbananos J, Santisteban J, Angulo F, Jose
Goiriena J. Exercise intensity and load during mass-start stage races
in professional road cycling. Med Sci Sports Exerc.2001;33(5):
796–802.
9. Pinot J, Grappe F. A six-year monitoring case study of a top-10
cycling Grand Tour finisher. J Sports Sci. 2015;33(9):907–14.
10. Allen H, Coggan A. Training and racing with a power meter. 2nd ed:
Velopress; 2010.
11. Sanders D, Heijboer M, Hesselink MKC, Myers T, Akubat I.
Analysing a cycling grand tour: can we monitor fatigue with intensity
or load ratios? JSportsSci. 2018;36(12):1385–91.
12. Lastella M, Roach GD, Halson SL, Martin DT, West NP, Sargent C. The
impact of a simulated grand tour on sleep, mood, and well-being of com-
petitive cyclists. J Sports Med Phys Fitness. 2015;55(12):1555–64.
13. Rodriguez-Marroyo JA, Pernia R, Villa JG, Foster C. Reliability and
seasonal changes of submaximal variables to evaluate professional
cyclists. Int J Sports Physiol Perform. 2017;10(12):1356–62.
14. Abbiss CR, Laursen PB. Models to explain fatigue during prolonged
endurance cycling. Sports Med. 2005;35(10):865–98.
http://www.acsm-msse.org874 Official Journal of the American College of Sports Medicine
APPLIED SCIENCES
Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
15. Nimmerichter A, Eston R, Bachl N, Williams C. Effects of low and
high cadence interval training on power output in flat and uphill cy-
cling time-trials. Eur J Appl Physiol. 2012;112(1):69–78.
16. Nybo L. Cycling in the heat: performance perspectives and cerebral
challenges. Scand J Med Sci Sports. 2010;20(3):71–9.
17. Schuler B, Thomsen JJ, Gassmann M, Lundby C. Timing the arrival
at 2340 m altitude for aerobic performance. Scand J Med Sci Sports.
2007;17(5):588–94.
18. Padilla S, Mujika I, Cuesta G, Goiriena JJ. Level ground and uphill
cycling ability in professional road cycling. Med Sci Sports Exerc.
1999;31(6):878–85.
19. Mujika I, Padilla S. Physiological and performance characteris-
tics of male professional road cyclists. Sports Med. 2001;31(7):
479–87.
20. Foster C, Hoyos J, Earnest C, Lucia A. Regulation of energy expen-
diture during prolonged athletic competition. Med Sci Sports Exerc.
2005;37(4):670–5.
21. Rodriguez-Marroyo JA, Villa G, Garcia-Lopez J, Foster C. Compar-
ison of heart rate and session rating of perceived exertion methods of
defining exercise load in cyclists. JStrengthCondRes.2012;26(8):
2249–57.
22. Sanders D, van Erp T, de Koning JJ. Intensity and load characteristics
of professional road cycling: differences between men’sandwomen’s
races. Int J Sports Physiol Perform. 2018;14(3):296–302.
23. www.Touretappe.nl. Tour Etappe 2018 [cited 20 Dec 2018].
24. www.procyclingstats.com 2018 [cited 20 Dec 2018].
25. Cohen J. Statistical Power Analysis for the Behavioral Sciences.
2nd ed. Hillsdale (NJ): Lawrence Erlbaum Associates; 1988. p. 567.
26. van Erp T, Hoozemans M, Foster C, de Koning JJ. The influence
of exercise intensity on the association between kilojoules spent
and various training loads in professional cycling. Int J Sports
Physiol Perform. 2019;1–6. In press. https://doi.org/10.1123/
ijspp.2018-0877.
27. Morton RH. A 3-parameter critical power model. Ergonomics. 1996;
39(4):611–9.
28. Morton RH, Hodgson DJ. The relationship between power output
and endurance: a brief review. Eur J Appl Physiol Occup Physiol.
1996;73(6):491–502.
29. Bassett DR Jr, Kyle CR, Passfield L, Broker JP, Burke ER. Compar-
ing cycling world hour records, 1967–1996: modeling with empirical
data. Med Sci Sports Exerc. 1999;31(11):1665–76.
30. Menaspa P, Impellizzeri FM, Haakonssen EC, Martin DT, Abbiss
CR. Consistency of commercial devices for measuring elevation
gain. Int J Sports Physiol Perform. 2014;9(5):884–6.
31. Nimmerichter A, Williams C, Bachl N, Eston R. Evaluation of a field
test to assess performance in elite cyclists. Int J Sports Med. 2010;
31(3):160–6.
32. Klitzke Borszcz F, Ferreira Tramontin A, Pereira Costa V. Is the func-
tional threshold power interchangeable with the maximal lactate steady
state in trained cyclists? Int J Sports Physiol Perform. 2019;1–21.
33. Seiler S. What is best practice for training intensity and duration dis-
tribution in endurance athletes? Int J Sports Physiol Perform. 2010;
5(3):276–91.
34. Seiler STE. Intervals, thresholds, and long slow distance: the role of
intensity and duration in endurance training. Sportscience. 2009;
13(11):32–53.
ELITE CYCLISTS COMPETING IN CYCLING GRAND TOURS Medicine & Science in Sports & Exercise
®
875
APPLIED SCIENCES
Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
