International Journal of Sports Physiology and Performance, (Ahead of Print)

https://doi.org/10.1123/ijspp.2020-0923
© 2021 Human Kinetics, Inc. ORIGINAL INVESTIGATION

Five-Minute Power-Based Test to Predict Maximal Oxygen

Consumption in Road Cycling
Sebastian Sitko, Rafel Cirer-Sastre, Francisco Corbi, and Isaac López-Laval

Purpose: To examine the ability of a multivariate model to predict maximal oxygen consumption (VO2max) using performance
data from a 5-minute maximal test (5MT). Methods: Forty-six road cyclists (age 38 [9] y, height 177 [9] cm, weight 71.4 [8.6] kg,
VO2max 61.13 [9.05] mL/kg/min) completed a graded exercise test to assess VO2max and power output. After a 72-hour rest,
they performed a test that included a 5-minute maximal bout. Performance variables in each test were modeled in 2 independent
equations, using Bayesian general linear regressions to predict VO2max. Stepwise selection was then used to identify the minimal
subset of parameters with the best predictive power for each model. Results: Five-minute relative power output was the best
explanatory variable to predict VO2max in the model from the graded exercise test (R2 95% credibility interval, .81–.88) and
when using data from the 5MT (R2 95% credibility interval, .61–.77). Accordingly, VO2max could be predicted with a 5MT
using the equation VO2max = 16.6 + (8.87 × 5-min relative power output). Conclusions: Road cycling VO2max can be predicted
in cyclists through a single-variable equation that includes relative power obtained during a 5MT. Coaches, cyclists, and
scientists may benefit from the reduction of laboratory assessments performed on athletes due to this finding.

Keywords: endurance, power output, power meter, training

In recent years, multiple performance factors have been reported.11,12 If VO2max can be attained during a short maximal
identified in road cycling.1 Among the physiological determinants effort, these kinds of intervals could be used to predict this
of performance, maximal oxygen consumption (VO2max) adjusted important cardiorespiratory fitness determining factor: several
to body mass can be highlighted as a key parameter of cardiore- studies have shown that 3- and 5-minute maximal intervals can
spiratory fitness.2 Normally, gas exchange measurements during predict VO2max attained during a GXT in several populations,
graded exercise tests (GXTs) are used to assess VO2max in cyclists. such as trained and untrained adults and adolescents.13–15 Although
Although these determinations allow direct oxygen and carbon these results seem promising, they should not be extrapolated to
dioxide calculations, they also present several setbacks: several specific performance groups such as road cyclists, as previous
different GXT protocols and step durations have been proposed in studies have shown that VO2 behavior during a short interval may
the past. However, previous research highlights that the use of vary depending on the sport discipline, fitness status, and effort
different GXT protocols may yield varying VO2max results.3–5 In duration.11 To date, the usefulness of a short maximal interval for
addition, the VO2max results may also vary for several other the prediction of VO2max in the road cycling population has not
reasons. For instance, GXT length could lead less-motivated been established.
individuals to stop exercising before the VO2max intensity is In recent years, the possibility of assessing the power output in
reached.4 Also, different exercise modalities may render different the field has been provided by the exponential rise in the use of
VO2max results. Finally, reaching the VO2 plateau is necessary for power meters among professional and amateur cyclists. These user-
VO2max determination through GXT, which is not always friendly tools allow the collection of real-time power data with
achieved in these tests.6,7 In these cases, relying on secondary reduced associated costs.16 The possibility of assessing power
criteria or verification tests further complicates the assessment4 and output in the field has generated interest in the indirect measure-
puts the spotlight on alternative VO2max assessment methods. ment of several ventilatory and lactate landmarks.17 In order to
The interest in self-paced short-duration maximal efforts for sustain a specific power output, energy needs to be regenerated at a
the determination of VO2max has risen in recent years. The rate that matches demand, a process that depends on the body’s
estimation of VO2max from these types of efforts has a relevant ability to supply oxygen to the muscle.18 Previous research has
physiological basis: VO2max values obtained during a GXT can be shown linear VO2 responses to increasing power output demands
reached in many cases during the first 2 minutes of a maximal during GXT until the lactate threshold is reached, with nonlinear
exercise bout, with 4- to 5-minute intervals showing VO2 decre- increases afterward.19–21 The VO2 behavior during suprathreshold
ments in the last part of the effort.8–10 This has been confirmed by exercise is related to the cyclists’ performance level and predomi-
studies assessing time to exhaustion at the VO2max intensity in nance of fiber type, with type IIX fibers associated to higher VO2
several endurance sports, such as swimming, cycling, and running, excess at higher power outputs.22 As the relationship between VO2
in which times to exhaustion of 3 to 5 minutes are commonly and power output varies at different exercise intensities, a short
supramaximal exercise bout could represent an opportunity to
Sitko and López-Laval are with the Dept of Physiatry and Nursery, Section of
predict VO2max through the assessment of the power output.
Physical Education and Sports, Faculty of Health and Sport Sciences, University of The direct link between VO2 and power output could represent
Zaragoza, Huesca, Spain. Cirer-Sastre and Corbi are with the National Inst of an opportunity to predict VO2max through the assessment of the
Physical Education of Catalonia (INEFC), University of Lleida (UdL), Lleida, power output measured during a short maximal exercise bout that
Spain. Sitko ([email protected]) is corresponding author. could be easily implemented in a field setting.
1
2 Sitko et al

Given the easy accessibility to power output measurement by measured height with a SECA 214 stadiometer (Seca, Hamburg,
cyclists, coaches, and scientists and considering the inconve- Germany), which is graduated up to 1 mm. After the body
niences related to assessing VO2max during a GXT, we hypothe- composition assessment, the participants performed a GXT based
sized that a 5-minute interval could be used to predict VO2max on previous protocols.24 The GXT commenced at a workload of
obtained through gas exchange methods in a GXT. Consequently, 150 W, with 25-W increments every 3 minutes until exhaustion. In
the objective of the present study was to examine the ability of a all cases, a 5-minute warm-up at 100 W was allowed before
multivariate model to predict VO2max using performance data commencement of the test. The assessment ended after a plateau
from a 5-minute maximal test (5MT). of the VO2 curve was reached or, when not seen, at voluntary
fatigue when 100% of the estimated maximal heart rate (HRmax), a
rate of perceived exertion of ≥18, and a respiratory exchange ratio
Methods of ≥1.15 were attained.5 We asked the participants to totally refrain
Participants from exercise until the next assessment.
After a 3-day rest, the participants returned to the laboratory
We recruited 46 road cyclists for the study (age 38 [9] y, height 177 for the second assessment, which was based on the testing protocol
[9] cm, body mass 71.4 [8.6] kg, body mass index 22.69 [2.23] kg/m2, suggested by Coggan.25 The warm-up started with 15 minutes
fat mass 11.1% [4.5%], VO2max 61.13 [9.05] mL/kg/min). The at 75% of the HRmax reached during the GXT and was followed by
inclusion criteria were (1) being the current owner of a cycling 3 fast-pedaling intervals (100 rpm and 90% of HRmax) separated by
license: World Tour, Elite/U23, Masters, or recreational; (2) the 1-minute soft-pedaling rests (75% of HRmax). The warm-up ended
absence of surgical procedures and injuries in the 6 months prior with an easy 3-minute rest (75% of HRmax). Finally, the subjects
to the study; and (3) the absence of drug use in the 6 months prior did a 5MT and were instructed to pace it to complete the entire
to the study. The sample was composed of recreationally trained duration. The protocol ended with a 10-minute cooldown at 65%
(n = 14), trained (n = 11), well-trained (n = 10), and professional of HRmax. The participants performed both tests on their own
(n = 11) cyclists according to previous guidelines for the classifi- bikes, set up on the Tacx Neo Smart 2 T bike trainer (Tacx Interna-
cation of endurance athletes.23 After being informed of the benefits tional, Rijksstraatweg, The Netherlands) and had visual access to
and potential risks of the investigation, each participant completed real-time power, heart rate, and cadence data obtained from the
a health-screening questionnaire and provided written informed trainer.26 During both tests, we analyzed the gas exchange (CPX/D;
consent prior to participation in the study. The study followed the Med Graphics, St Paul, MN) and monitored the heart rate with the
ethical guidelines of the 2013 Declaration of Helsinki and received Polar Team Pro tool (Polar Electro, Espoo, Finland) and the rate
approval from the research ethics committee of the autonomous of perceived exertion with the rate of perceived exertion scale
region of Aragon, Spain (PI19/447). category ratio scale 1027

Experimental Design Statistical Analysis

The participants completed the study on 2 separate days, with a 72- We described the performance data as mean (SD) and explored
hour rest in between. On the first day, we assessed the body associations between each variable and VO2max using indepen-
composition of all cyclists through the electrical impedance method dent univariate general linear regressions (Table 1). To predict
(BC-602; Tanita Co, Tokyo, Japan) in the morning hours. We VO2max, we created 2 separate Bayesian general linear models,

Table 1 Univariate Linear Regression Coefficients Between VO2max and Study Variables
Variable Mean (SD) β SE 95% CI R2
Body weight, kg 71.4 (8.6) −0.57 0.14 −0.84 to −0.29 .29
Incremental test
HRmax, beats/min 178 (11) 0.29 0.13 0.04 to 0.54 .12
Power output at VO2max, W 320 (46) 0.13 0.02 0.08 to 0.17 .41
Relative power output at VO2max, W/kg 4.5 (0.8) 10.8 0.68 9.47 to 12.14 .86
HR at VO2max, beats/min 177 (10) 0.28 0.14 0.02 to 0.56 .11
Relative HR at VO2max (%HRmax) 99 (2) −0.43 0.62 −1.63 to 0.79 .03
Cadence at VO2max, rpm 95 (9) 0.45 0.13 0.19 to 0.72 .22
RPE at VO2max, AU 18 (1) 0.69 1.15 −1.47 to 2.91 .03
Coggan test
5-min power output, W 355 (56) 0.04 0.01 0.01 to 0.07 .17
5-min power output, W/kg 5 (0.9) 8.87 0.87 7.21 to 10.55 .71
5-min heart rate, beats/min 178 (10) 0.12 0.14 −0.15 to 0.38 .04
5-min relative heart rate (%HRmax) 100 (5) −0.50 0.27 −1.03 to 0.02 .09
Mean cadence, rpm 95 (9) 0.34 0.14 0.06 to 0.62 .13
5-min RPE, AU 20 (0) −0.51 2.87 −5.87 to 5.29 .02
Abbreviations: AU, arbitrary units; CI, credibility interval; HR, heart rate; HRmax, maximal heart rate; RPE, rating of perceived exertion; β, mean estimated coefficient;
VO2max, maximal oxygen consumption.

(Ahead of Print)
 Five-Minute Test to Predict VO2max in Cycling 3

the first including all performance data obtained during the graded during the 5MT (R2 = .71, 95% credibility interval, .61 to .77). This
test, and the second including data from the 5MT. Then, we carried equation was presented as VO2max = β0 + β1 × 5-minute RPO, and
out a forward stepwise selection analysis to identify which combi- its coefficients are shown in Table 3. As an example, a cyclist
nation of performance variables led to the best predictive equation delivering an RPO of 6 W/kg would have an estimated VO2max
for both scenarios. of 16.61 + 8.87 × 6 = 69.83 mL/min/kg (Table 4). A more detailed
We performed data cleaning, manipulation, and analyses in summary of the results of the complete model and stepwise
Microsoft Excel (version 16.43; Microsoft Corp, Armonk, NY) and selection are presented in Supplementary File 3 (available online).
R (version 4.0.2; R Foundation for Statistical Computing, Vienna, The association between RPO at VO2max and the 5MT is pre-
Austria). We performed Bayesian general linear models using the sented in Figure 1.
Stan computational framework (http://mc-stan.org/) accessed with
the brms package.28 In the Bayesian models, we chose a regular-
ized horseshoe prior to the current study due to the advantages Discussion
discussed by Piironen and Vehtari.29 Stepwise variable selection The aim of this study was to examine the ability of a multivariate
was achieved using the predictive projection technique, accessed
model to predict VO2max using performance data from a 5MT. The
with a projpred package,30 and based on leave-one-out cross-
main findings of the study were as follows: (1) power output
validation.31 We judged the best model size in each scenario by
adjusted to body mass was the best explanatory variable for
the mean log predictive density and the root mean square. After
VO2max and (2) VO2max obtained during a GXT can be predicted
validating the models, we explored the residuals of the final models
through a single-variable equation based on a 5MT, which is
for normality, homogeneity, and independence assumptions. We
presented as VO2max = 16.61 + 8.87 × 5 min RPO.
checked the normality assumption of the residuals by means of a
To the best of our knowledge, this is the first study that has
normal Q–Q plot of residuals. In addition, we assessed the propor-
assessed whether the 5-minute interval proposed in the Coggan
tion of the variance explained using Bayesian R2.32 We present
protocol could be used for the prediction of VO2max obtained
Markov chain Monte Carlo and model diagnoses of all models in
Supplementary Files 2 and 4 (available online).
Table 3 Regression Coefficients for Model 2,
Results Predicting VO2max Based on 5-Minute Test Variables
VO2max Prediction From Performance Data Coefficient Mean SE 95% CI
According to the univariate models, participants with higher (β0) Intercept 16.61 4.42 8.13–25.21
VO2max in the graded test had lower body mass and, in the (β1) 5-min relative power output, W/kg 8.87 0.87 7.21–10.55
same test, achieved higher absolute and relative power, absolute (σ) SD 4.99 0.54 4.07–6.20
heart rate, and cadence (Table 1). The stepwise multiple regression Abbreviations: CI, credibility interval; VO2max, maximal oxygen consumption.
analysis produced an equation for VO2max with a single explana-
tory variable, which was the relative power output (RPO) at
VO2max during the graded test (R2 = .86, 95% credibility inter-
val, .81 to .88). This equation was presented as VO2max = β0 + Table 4 Predicted Values for VO2max According
β1 × VO2max RPO, and its coefficients are shown in Table 2. A to the Results
more detailed summary of the results of the complete model and
stepwise selection are presented in Supplementary File 1 (available RPO at min 5, W/kg Estimated VO2max, mL/min/kg
online). 3 43.2
3.5 47.8
VO2max Prediction From a 5MT 4 51.9
The participants with better VO2max in the graded test achieved 4.5 56.5
higher absolute and RPOs and higher cadence in the 5MT (Table 3). 5 60.9
The VO2max obtained during the 5MT was statistically compara- 5.5 65.5
ble with that attained in the GXT, t(45) = −.73; P = .47, with an 6 69.8
almost perfect linear correlation (adjusted R2 = .988, P < .0001). In
6.5 74.4
line with the final model for the graded test variables, the stepwise
multiple regression analysis produced an equation for VO2max 7 78.7
with a single explanatory variable, which was the average RPO Abbreviations: RPO, relative power output; VO2max, maximal oxygen consumption.

Table 2 Regression Coefficients for Model 1, VO2max Prediction Based

on Incremental Test Variables
Coefficient Mean SE 95% CI
(β0) Intercept 12.16 3.11 6.10–18.25
(β1) Relative power output at VO2max, W/kg 10.80 0.68 9.47–12.14
(σ) SD 3.51 0.40 2.84–4.39
Abbreviations: CI, credibility interval; VO2max, maximal oxygen consumption.

(Ahead of Print)
4 Sitko et al

Figure 1 — Left panel, association between power outputs in the graded exercise test and 5MT (r = .61); right panel, association between observed
and predicted VO2max. 5MT indicates 5-minute test; VO2max, maximal oxygen consumption.

through laboratory gas exchange methods. The testing proposed by these methods may be deemed acceptable, the protocol compliance
Coggan has been widely researched in the last years, though most might be affected by the practical difficulties associated with the
studies have focused on the 20-minute interval for the assessment adjustments to the power output increments that these tests estab-
of threshold power.17 Previous studies have shown strong correla- lish. With regard to this, Astorino et al38 showed that a self-paced
tions between RPO during short intervals, maximal aerobic power, cycle protocol elicited higher VO2max compared with a ramp
and VO2max. Novak et al2 reported strong correlations between protocol. Furthermore, Foster et al39 compared the physiological
RPO during 15-second to 10-minute intervals, maximal aerobic responses between a self-paced laboratory 5-km cycle time trial and
power (r = .70–.94, P < .001), and VO2max (r = .82–.88, P < .001). a GXT using a cycle ergometer. They found that VO2max, heart
Sanders et al33 reported that an 8-minute interval could be used as rate, ventilation, and blood lactate levels were significantly greater
a predictor of endurance performance, given its relationship with in the 5-km time trial compared with the GXT.39 Contrary to the
several laboratory-set thresholds, although its relationship with GXT, the 5MT performed in our study was self-paced.25 The
VO2max in particular was not tested in this study. Finally, Eston cyclists were required to produce the highest possible power output
et al34 showed strong correlations between 2- and 4-minute sub- and could use their own pacing strategy. This brings us to another
maximal tests and VO2max obtained from graded exercise test- question: Could the pacing strategy impact peak oxygen uptake
ing (r = .79–.92). values obtained during the 5MT? Two previous studies have
In line with these previous findings, the RPO during the 5MT assessed VO2 behavior during 5-minute all-out efforts performed
in our study was strongly correlated to the RPO at the VO2max by cyclists who followed different pacing strategies.9,10 In both
intensity (r = .61) and VO2max itself (R2 = .71, 95% credibility cases, although the mean VO2 values observed during the 5-minute
interval, .61 to .77). Accordingly, we proposed a predictive formula intervals were similar, faster paces during the start of the interval
for estimating VO2max from RPO during a 5MT. One of the testing yielded higher peak oxygen uptake values. Our study was not
protocols that is most widely used in the coaching field was designed to assess the differences in pacing, and this could repre-
designed by Coggan and includes a 5-minute effort among the sent an interesting opportunity for future research.40
recommended intervals.25 Furthermore, Coggan proposed a power In contrast with previous research, which normally focused on
profile chart in which the mean power output during a 5-minute specific fitness levels (either recreational, trained, or elite), the
effort could be used to categorize riders from novice to world class. current study included participants with a wide range of fitness
When applying our predictive formula to the rider categorization levels (11 World Tour, 12 U23/Elite, 12 Master, and 11 recrea-
suggested by Coggan, world class and exceptional riders would be tional cyclists).17 In addition, compared with preceding investiga-
represented by VO2max values above 77.5 and 71.9 mL/kg/min, tions, this study’s sample size was also larger. Consequently, the
respectively.25 These values are in line with those previously equation obtained in this study is generic and not specific to fitness
reported for elite cyclists and meet the criteria used to categorize levels. On the one hand, this strengthens the practical application of
endurance athletes according to their fitness level.19,23,35 Given all our findings, as it could be used at many performance levels. In the
the above, and that field testing can be done more easily as opposed authors’ opinion, an equation that can be used among broad cycling
to laboratory testing, we consider that our predictive formula may populations could be very interesting for coaches who typically
represent an interesting option for cyclists, coaches, and scientists work with athletes of different fitness levels. On the other hand,
who do not have gas exchange measurement instruments available. however, slight differences in the coefficients could exist if we had
Several authors have proposed methods for predicting VO2max estimated independent equations for each fitness level. In this
through incremental tests based on power output.36,37 Although regard, the diagnostic plot provided in the right panel of Figure 1
(Ahead of Print)
 Five-Minute Test to Predict VO2max in Cycling 5

suggests that the prediction error at the 2 extremes of the observed Acknowledgments
VO2max range was small and comparable between the participants
with the highest and lowest fitness levels. However, as inherent in The authors would like to thank all the World Tour riders for their
linear modeling, prediction uncertainty is still lower at the center willingness to share their data and participate in this study.
and higher at the extremes. For this reason, it might be interesting to
complement the present results in future research by addressing the
possible differences in the coefficients and/or prediction accuracy References
when fitting separate, level-specific equations instead of the generic
1. Phillips KE, Hopkins WG. Determinants of cycling performance: a
one developed in this study.
review of the dimensions and features regulating performance in elite
Several aspects should be taken into account when attempting
cycling competitions. Sport Med Open. 2020;6(1):23. doi:10.1186/
to extrapolate our results to the field: modern power meters are
s40798-020-00252-z
generally precise but may present variable trueness.41 Therefore,
2. Novak AR, Dascombe BJ. Physiological and performance character-
the extrapolation of our predictive formulas to tests performed with
istics of road, mountain bike and BMX cyclists. J Sci Cycling. 2014;
other measuring systems must be exercised with caution. Further-
3(3):9–16.
more, power variability, pedaling characteristics, cadence, and
3. Jamnick NA, Botella J, Pyne DB, Bishop DJ. Manipulating graded
biomechanics are different in the field compared with the labora-
exercise test variables affects the validity of the lactate threshold and
tory setting.42,43 The protocol used in the current study was
VO2peak. PLoS One. 2018;13(7):e0199794. PubMed ID: 30059543
characterized by several limitations. First, we obtained VO2max
doi:10.1371/journal.pone.0199794
values from a GXT based on previous research and characterized
4. Poole DC, Jones AM. Measurement of the maximum oxygen uptake
by 3-minute steps. It is well known that the GXT protocol design
VO2max: VO2peak is no longer acceptable. J Appl Physiol. 2017;
may influence VO2max testing results, and the evidence regarding
122(4):997–1002. PubMed ID: 28153947 doi:10.1152/japplphysiol.
the suitability of this duration for the assessment of VO2max is
01063.2016
contradictory: while some studies show equal VO2max values
5. Beltz NM, Gibson AL, Janot JM, Kravitz L, Mermier CM, Dalleck
obtained with 1-, 2-, and 3-minute steps, others have reported
LC. Graded exercise testing protocols for the determination of
that shorter tests provide higher values, and vice versa.3–5,44–48
VO2max: historical perspectives, progress, and future considerations.
Second, several authors have suggested that the inclusion of veri-
J Sports Med. 2016;2016(1):1–12. doi:10.1155/2016/3968393
fication bouts after a GXT could help confirm whether true
6. Lucía A, Rabadán M, Hoyos J, et al. Frequency of the VO2max plateau
VO2max was attained during the test.5,49 This procedure has,
phenomenon in world-class cyclists. Int J Sports Med. 2006;27(12):
however, several limitations, such as a lack of standardization,
984–992. PubMed ID: 16739087 doi:10.1055/s-2006-923833
homogeneity, and scarcity of previous research with study samples
7. Caputo F, Denadai BS. Exercise mode affects the time to achieve
such as ours.50,51 In this study, VO2max obtained during the
V̇O2max without influencing maximal exercise time at the intensity
5MT was statistically comparable with that attained in the GXT,
associated with V̇O2max in triathletes. Int J Sports Med. 2006;27(10):
t(45) = −.73; P = .47, with an almost perfect linear correlation
798–803. PubMed ID: 16586327 doi:10.1055/s-2005-872962
(adjusted R2 = .988, P < .0001). Finally, the implementation of
8. Hanon C, Leveque JM, Thomas C, Vivier L. Pacing strategy and V̇O2
our 5MT requires taking into consideration that time to exhaustion kinetics during a 1500-m race. Int J Sports Med. 2008;29(3):206–211.
at VO2max intensity varies among cyclists of the same apparent PubMed ID: 17990206 doi:10.1055/s-2007-965109
level and that VO2 kinetics vary among individuals.12,52 Also, as it 9. Aisbett B, Le Rossignol P, McConell GK, Abbiss CR, Snow R.
is commonly reported in the literature, a VO2 plateau was not seen Effects of starting strategy on 5-min cycling time-trial performance.
in all cases during the GXT.6 Secondary measures, such as rating J Sports Sci. 2009;27(11):1201–1209. PubMed ID: 19724963 doi:
of perceived exertion, respiratory exchange ratio, heart rate, and 10.1080/02640410903114372
voluntary fatigue, were used as criteria for test termination in these 10. Aisbett B, Lerossignol P, McConell GK, Abbiss CR, Snow R.
cases. Although our GXT was performed until voluntary fatigue Influence of all-out and fast start on 5-min cycling time trial perfor-
when a VO2max plateau was not observed, which minimizes the mance. Med Sci Sports Exerc. 2009;41(10):1965–1971. PubMed ID:
risk of early termination and VO2max underestimation, some 19727014 doi:10.1249/MSS.0b013e3181a2aa78
authors have suggested that secondary termination criteria may 11. Caputo F, Mello MT, Denadai BS. Oxygen uptake kinetics and time
affect VO2max calculations during GXT.4 to exhaustion in cycling and running: a comparison between trained
and untrained subjects. Arch Physiol Biochem. 2003;111(5):461–466.
Practical Applications PubMed ID: 16026035 doi:10.3109/13813450312331342337
12. Billat V, Faina M, Sardella F, et al. A comparison of time to
By using a simple 5MT implemented into a field training session, exhaustion at VO2max in elite cyclists, kayak paddlers, swimmers and
coaches and cyclists with a wide range of performance levels can runners. Ergonomics. 1996;39(2):267–277. PubMed ID: 8851531
assess VO2max through a single variable predictive equation. This doi:10.1080/00140139608964457
assessment may contribute to more accurate and individualized 13. Macnaughton L, Croft R, Pennicott J, Long T. The 5 and 15 minute
training programs and may also reduce the need for access to a runs as predictors of aerobic capacity in high school students. J Sports
laboratory setting to evaluate this cardiorespiratory parameter. Med Phys Fitness. 1990;30(1):24–28. PubMed ID: 2366531
14. Cooper K, Shafer A. Validity and reliability of the polar A300’s
Conclusions fitness test feature to predict VO2max. Int J Exerc Sci. 2019;12(4):
393–401.
In this study, in a sample of 46 cyclists with a wide range of fitness 15. Dexheimer JD, Brinson SJ, Pettitt RW, Schroeder ET, Sawyer BJ,
levels, VO2max obtained through gas exchange testing during a Jo E. Predicting maximal oxygen uptake using the 3-minute all-out
GXT could be predicted by a single variable formula based on the test in high-intensity functional training athletes. Sports. 2020;8(12):
RPO obtained during a 5MT. 155. doi:10.3390/sports8120155

(Ahead of Print)
6 Sitko et al

16. Passfield L, Hopker J, Jobson S, Friel D, Zabala M. Knowledge 2006;97(5):535–541. PubMed ID: 16779551 doi:10.1007/s00421-
is power: issues of measuring training and performance in cycling. 006-0213-x
J Sports Sci. 2017;35(14):1426–1434. PubMed ID: 27686573 doi: 35. Sallet P, Mathieu R, Fenech G, Baverel G. Physiological differences
10.1080/02640414.2016.1215504 of elite and professional road cyclists related to competition level and
17. Sitko S, Cirer-Sastre R, Corbi F, López-Laval I. Power assessment in rider specialization. J Sports Med Phys Fitness. 2006;46(3):361–365.
road cycling: a narrative review. Sustainability. 2020;12(12):5216. PubMed ID: 16998438
doi:10.3390/su12125216 36. González-Haro C, Galilea PA, Drobnic F, Escanero JF. Validation of
18. Bassett DR, Howley ET. Limiting factors for maximum oxygen a field test to determine the maximal aerobic power in triathletes and
uptake and determinants of endurance performance. Med Sci Sports endurance cyclists. Br J Sports Med. 2007;41(3):174–179. PubMed
Exerc. 2000;32(1):70–84. PubMed ID: 10647532 doi:10.1097/ ID: 17178775 doi:10.1136/bjsm.2006.031310
00005768-200001000-00012 37. Santtila M, Häkkinen K, Pihlainen K, Kyröläinen H. Comparison
19. Lucía A, Hoyos J, Santalla A, Pérez M, Chicharro JL. Curvilinear between direct and predicted maximal oxygen uptake measure-
VO2: power output relationship in a ramp test in professional cyclists: ment during cycling. Mil Med. 2013;178(2):234–238. PubMed
possible association with blood hemoglobin concentration. Jpn J ID: 23495472 doi:10.7205/MILMED-D-12-00276
Physiol. 2002;52(1):95–103. PubMed ID: 12047807 doi:10.2170/ 38. Astorino TA, McMillan DW, Edmunds RM, Sanchez E. Increased
jjphysiol.52.95 cardiac output elicits higher VO2max in response to self-paced exer-
20. Muniz-Pumares D, Pedlar C, Godfrey R, Glaister M. The effect of cise. Appl Physiol Nutr Metab. 2015;40(3):223–229. PubMed ID:
the oxygen uptake-power output relationship on the prediction of 25682980 doi:10.1139/apnm-2014-0305
supramaximal oxygen demands. J Sports Med Phys Fitness. 2017; 39. Foster C, Green MA, Snyder AC, Thompson NN. Physiological
57(1–2):1–7. PubMed ID: 26564270 responses during simulated competition. Med Sci Sports Exerc. 1993;
21. Lamberts RP, Lambert MI, Swart J, Noakes TD. Allometric scaling 25(7):877–882. PubMed ID: 8350712 doi:10.1249/00005768-
of peak power output accurately predicts time trial performance and 199307000-00018
maximal oxygen consumption in trained cyclists. Br J Sports Med. 40. Chidnok W, Dimenna FJ, Bailey SJ. VO2max is not altered by self-
2012;46(1):36–41. PubMed ID: 21821613 doi:10.1136/bjsm.2010. pacing during incremental exercise. Eur J Appl Physiol. 2013;113(2):
083071 529–539. PubMed ID: 22941093 doi:10.1007/s00421-012-2478-6
22. Lucía A, Rivero JLL, Pérez M, et al. Determinants of V̇O2 kinetics at 41. Maier T, Schmid L, Müller B, Steiner T, Wehrlin J. Accuracy of
high power outputs during a ramp exercise protocol. Med Sci Sports cycling power meters against a mathematical model of treadmill
Exerc. 2002;34(2):326–331. PubMed ID: 11828244 doi: 10.1097/ cycling. Int J Sports Med. 2017;38(6):456–461. PubMed ID:
00005768-200202000-00022 28482367 doi:10.1055/s-0043-102945
23. De Pauw K, Roelands B, Cheung SS, De Geus B, Rietjens G, 42. Bertucci W, Grappe F, Groslambert A. Laboratory versus outdoor
Meeusen R. Guidelines to classify subject groups in sport-science cycling conditions: differences in pedaling biomechanics. J Appl
research. Int J Sports Physiol Perform. 2013;8(2):111–122. PubMed Biomech. 2007;23(2):87–92. PubMed ID: 17603128 doi:10.1123/
ID: 23428482 doi:10.1123/ijspp.8.2.111 jab.23.2.87
24. Mcnaughton LR, Roberts S, Bentley DJ. The relationship among 43. Jeffries O, Waldron M, Patterson SD, Galna B. An analysis of
peak power output, lactate threshold, and short-distance cycling variability in power output during indoor and outdoor cycling time
performance. J Strength Cond Res. 2006;20(1):157–161. PubMed trials. Int J Sports Physiol Perform. 2019;14(9):1273–1279. doi:10.
ID: 16506862 doi:10.1519/00124278-200602000-00025 1123/ijspp.2018-0539
25. Allen H, Coggan A. Workout guide. In: Jardine R, ed. Training and 44. Pierce SJ, Hahn AG, Davie A, Lawton EW. Prolonged incremental
Racing With a Power Meter. Boulder, CO: VeloPress; 2019:341. tests do not necessarily compromise V̇O2max in well-trained athletes.
26. Tacx International. 2021. https://tacx.com/es/producto/neo-2t-smart/. J Sci Med Sport. 1999;2(4):356–363. PubMed ID: 10710013 doi:
Accessed February 4, 2021. 10.1016/S1440-2440(99)80008-5
27. Williams N. The Borg rating of perceived exertion (RPE) scale. 45. Buchfuhrer MJ, Hansen JE, Robinson TE, Sue DY, Wasserman K,
Occup Med. 2017;67(5):404–405. doi:10.1093/occmed/kqx063 Whipp BJ. Optimizing the exercise protocol for cardiopulmonary
28. Bürkner PC. Brms: an R package for Bayesian multilevel models assessment. J Appl Physiol Respir Environ Exerc Physiol. 1983;
using Stan. J Stat Softw. 2017;80(1):1–28. doi:10.18637/jss.v080.i01 55(5):1558–1564. PubMed ID: 6643191 doi:10.1152/jappl.1983.55.
29. Piironen J, Vehtari A. Comparison of Bayesian predictive methods 5.1558
for model selection. Stat Comput. 2017;27(3):711–735. doi:10.1007/ 46. Roffey DM, Byrne NM, Hills AP. Effect of stage duration on
s11222-016-9649-y physiological variables commonly used to determine maximum
30. CRAN. CRAN—Package projpred. 2020. https://cran.r-project.org/ aerobic performance during cycle ergometry. J Sports Sci. 2007;
web/packages/projpred/index.html. Accessed November 19, 2020. 25(12):1325–1335. PubMed ID: 17786685 doi:10.1080/02640410
31. Loo. 2020. Efficient leave-one-out cross-validation and WAIC for 601175428
Bayesian models. https://mc-stan.org/loo/. Accessed November 19, 2020. 47. Pollock ML, Bohannon RL, Cooper KH, et al. A comparative
32. Gelman A, Goodrich B, Gabry J, Vehtari A. R-squared for Bayesian analysis of four protocols for maximal treadmill stress testing. Am
regression models. Am Stat. 2019;73(3):307–309. doi:10.1080/ Heart J. 1976;92(1):39–46. PubMed ID: 961576 doi:10.1016/S0002-
00031305.2018.1549100 8703(76)80401-2
33. Sanders D, Taylor RJ, Myers T, Akubat I. A field-based cycling test to 48. Zhang YY, Johnson MC, Chow N, Wasserman K. Effect of exercise
assess predictors of endurance performance and establishing training testing protocol on parameters of aerobic function. Med Sci Sports
zones. J Strength Cond Res. 2020;34(12):3482–3488. PubMed ID: Exerc. 1991;23(5):625–630. PubMed ID: 2072842 doi:10.1249/
28368958 doi:10.1519/JSC.0000000000001910 00005768-199105000-00016
34. Eston RG, Faulkner JA, Mason EA, Parfitt G. The validity of 49. Schaun GZ. The maximal oxygen uptake verification phase: a light at
predicting maximal oxygen uptake from perceptually regulated the end of the tunnel? Sport Med—Open. 2017;3(1):44. doi:10.1186/
graded exercise tests of different durations. Eur J Appl Physiol. s40798-017-0112-1

(Ahead of Print)
 Five-Minute Test to Predict VO2max in Cycling 7

50. Nolan PB, Beaven ML, Dalleck L. Comparison of intensities and rest J Appl Physiol. 2017;123(3):696. PubMed ID: 28947630 doi:10.
periods for VO2max verification testing procedures. Int J Sports Med. 1152/japplphysiol.00338.2017
2014;35(12):1024–1029. PubMed ID: 24886925 doi:10.1055/s- 52. Lucía A, Hoyos J, Santalla A, Pérez M, Chicharro JL. Kinetics of VO2
0034-1367065 in professional cyclists. Med Sci Sports Exerc. 2002;34(2):320–325.
51. Pettitt RW, Jamnick NA. Commentary on “measurement of the PubMed ID: 11828243
maximum oxygen uptake V̇O2max : V̇O2peak is no longer acceptable.”

(Ahead of Print)