European Journal of Applied Physiology

https://doi.org/10.1007/s00421-019-04258-8

ORIGINAL ARTICLE

Yo‑Yo intermittent tests are a valid tool for aerobic fitness assessment
in recreational football
Carlo Castagna1,2 · Peter Krustrup3 · Susana Póvoas3,4

Received: 13 June 2019 / Accepted: 1 November 2019

© Springer-Verlag GmbH Germany, part of Springer Nature 2019

Abstract
Purpose To examine the suitability of three versions and two levels of the Yo-Yo intermittent tests for assessing and tracking
aerobic fitness status development in male recreational football players. Sixty-six untrained participants (age 39 ± 6 years,
­VO2max 41.2 ± 6.2 ml kg−1 min−1, body mass 81.9 ± 10.8 kg, height 173.2 ± 6.4 cm) partook in a 12-week recreational football
training program. They were evaluated during the Yo-Yo intermittent endurance level 1 (YYIE1) and 2 (YYIE2) tests and
the Yo-Yo intermittent recovery level 1 test (YYIR1), and during a treadmill test for ­VO2max assessment, at baseline. Thirty-
two out of these 66 participants replicated all these tests at post-intervention. An additional group of 30 male age-matched
recreational football players that afterwards started the 12-week recreational football program (age 39 ± 6 years, V ­ O2max
45.3 ± 5.8 ml kg−1 min−1, body mass 82.5 ± 7.8 kg, height 172.8 ± 5.4 cm) was evaluated at baseline to test cross-validation.
Results The Yo-Yo tests showed very large associations with V ­ O2max at baseline (r = 0.75–0.77; P < 0.0001) and at post-
intervention (r = 0.76–0.82; P < 0.0005). Post-training, very large associations were found between YYIE2 performance
and ­VO2max (r = 0.65, P < 0.0001). Cross-validation revealed small to large differences between the observed and estimated
­VO2max values (1.5–2.96 ml kg−1 min−1) with moderate typical error of estimation (7.9–8.7%) across the tests. Performance
in the YYIE1, YYIE2 and YYIR1 tests of ≥ 1760, 480 and 600 m, respectively, indicated good to excellent ­VO2max values.
Conclusions The Yo-Yo tests considered here showed robust and consistent criterion validity. The YYIE2 could be a more
accurate option to track aerobic fitness development in recreational football players.

Keywords Soccer · Team sports · Maximal oxygen uptake · Field tests · Responsiveness · Endurance

Introduction

Individual level of aerobic fitness has been associated with

health and physical fitness in the general population irre-
spective of sex, age and physical activity/exercise status
(Nes et al. 2014; Pate and Kriska 1984). Given the social
Communicated by Anni Vanhatalo. interest of aerobic fitness development and the reported sup-
posed positive effects on a wide range of non-communicable
* Susana Póvoas diseases (e.g., hypertension, coronary heart disease, stroke,
[email protected]
diabetes, various types of cancer, depression), large-scale
1
Fitness Training and Biomechanics Laboratory, Technical aerobic exercise evaluation and prescription was proposed
Department, Italian Football Federation (FIGC), Coverciano, (Herdy and Caixeta 2016; Nes et al. 2014).
Florence, Italy Nevertheless, direct assessment of aerobic fitness still
2
University of Rome Tor Vergata, Rome, Italy remains an option of limited access to large-scale studies
3
Department of Sports Science and Clinical Biomechanics, due to its given economic and logistical cost, in addition
SDU Sport and Health Sciences Cluster (SHSC), University to the required presence of well-trained personnel. Conse-
of Southern Denmark, Odense, Denmark quently, field tests are a popular option in sport and exercise
4
Research Center in Sports Sciences, Health Sciences for aerobic fitness assessment, being considered easy for
and Human Development, CIDESD, University Institute
of Maia, ISMAI, Maia, Portugal

13
Vol.:(0123456789)
 European Journal of Applied Physiology

detecting metrics such as distance covered at a progressive Schmitz et al. 2018). However, no study has systematically
pace dictated by sound cues (Bangsbo et al. 2008). examined the validity levels of the Yo-Yo intermittent tests
Within field testing, the Yo-Yo intermittent tests, with in recreational football players. Moreover, the published
their different versions and levels, are currently the most studies arbitrarily choose Yo-Yo intermittent versions, since
popular option for aerobic fitness assessment in amateur, no information is available in the literature about the most
professional and recreational football (Schmitz et al. 2018). suitable test for estimating ­VO2max under field conditions in
The reason for their popularity may be their logical validity recreational football players (Krustrup and Bangsbo 2015;
(i.e., comprising intermittent high-intensity protocols and Krustrup and Krustrup 2018; Krustrup et al. 2009, 2018;
shuttle running), practicability and capacity for testing a Milanovic et al. 2015), or for detecting recreational football
large number of participants at the same time with practi- training-induced changes in aerobic fitness. Consequently,
cally no associated costs (Schmitz et al. 2018; Nes et al. evidence on the validity levels of the Yo-Yo intermittent ver-
2014). The feasibility and relevance of the Yo-Yo intermit- sions may be of great interest for training optimisation and
tent protocols for football has promoted a number of stud- assessment (Impellizzeri and Marcora 2009).
ies evaluating their validity and reliability (Bangsbo et al. The aim of this study was therefore, to examine the suita-
2008; Schmitz et al. 2018; Póvoas et al. 2016, 2018). Despite bility of three versions and two levels of the Yo-Yo intermit-
the fact that the Yo-Yo intermittent tests may be considered tent tests for assessing aerobic fitness status (in the untrained
among the most investigated and well-used field tests for and trained conditions) and training-induced changes in
aerobic fitness assessment, some key forms of validity are adult male participants undertaking a recreational football
still to be addressed in untrained participants, where, for intervention. Test validity consistency (pre-to-post-training)
the above-mentioned practical reasons, field testing is more and differences in external responsiveness were assumed as
appealing (Póvoas et al. 2016, 2018; Schmitz et al. 2018). working hypotheses.
Indeed, no published paper addressed so far the external
responsiveness (i.e., sensitivity) of the Yo-Yo intermittent
tests in tracking recreational players’ V ­ O2max development Methods
and the consistence of the criterion validity in the trained
status (i.e., pre-to-post-intervention criterion validity). Fur- Participants
thermore, the estimating power of the equations developed
via criterion validity (i.e., cross-validity) has not yet been Ninety-six male inactive participants volunteered to par-
the topic of published researches. Information about the ticipate in a 12-week recreational football training interven-
reported levels of validity will help to understand which is tion. A first group of sixty-six (age 39 ± 6 years, body mass
the most appropriate Yo-Yo intermittent test to be used both 81.9 ± 10.8 kg, height 173.2 ± 6.4 cm) performed baseline
with initially untrained and successively moderately trained and post-intervention evaluations. Afterwards, a second
individuals, as are those participating in recreational football group of 30 male age-matched recreational football play-
exercise-based programs. ers (age 38.9 ± 6.4 years, body mass 82.5 ± 7.8 kg, height
Recreational football has been shown to be a motivat- 172.8 ± 5.4 cm) that later was engaged in a recreational
ing and effective exercise mode for promoting relevant football training intervention, like the one here considered,
improvement in aerobic fitness (i.e., V ­ O2max) in the short- was selected to test cross-validation, performing the same
term (12–16 weeks) with a sustainable training volume tests and procedures before the beginning of the exercise
(30–60 min twice a week) (Krustrup et al. 2009, 2010, program. Test familiarization took place during the 2 weeks
2018; Krustrup and Krustrup 2018; Milanovic et al. 2015). prior to data collection. The study was conducted according
The agonistic nature of recreational football and its casually to the Declaration of Helsinki guidelines. Written informed
intermittent exercise mode pose questions about the influ- consent was obtained from each of the participants after
ence of team formation (i.e., players’ skill and fitness level) acknowledgement of the risks and benefits of their taking
to warrant game involvement and enjoyment and, conse- part in the study and of the option to withdraw from the
quently, group enhancement in aerobic fitness. The use of study at any time without penalty. The study received ethical
population valid field tests for aerobic fitness may result in clearance from the local Institutional Review Board before
a strategy for monitoring participants’ training improve- commencement.
ments and even possibly grouping players in teams that are
homogeneous with respect to aerobic fitness (Bangsbo et al. Design
2008; Milanovic et al. 2015; Schmitz et al. 2018). The Yo-Yo
intermittent tests have been shown to be sensitive in tracking This study comprised a descriptive longitudinal design aim-
aerobic fitness improvement in recreational football players ing to evaluate three Yo-Yo intermittent protocols, the Yo-Yo
of different skill levels, gender and age (Bangsbo et al. 2008; intermittent endurance level 1 (YYIE1) and 2 (YYIE2) tests,

13
European Journal of Applied Physiology

and the Yo-Yo intermittent recovery level 1 test (YYIR1), for at a subjectively set intensity by the participants (2–3 of
validity categories (Impellizzeri and Marcora 2009). These the Börg CR10 scale) and with changes of direction of
tests were chosen due to their popular use in football and, 180° to mimic the evaluation protocols, preceding each
specifically, in recreational football (Bangsbo et al. 2008; Yo-Yo intermittent test. The participants were asked to
Krustrup et al. 2015; Schmitz et al. 2018). progressively increase the intensity towards the end of the
Ninety-six participants agreed to partake in this study. warm-up, nonetheless without reaching maximal speed.
Sixty-six participants performed the considered field tests A 2-min passive recovery preceded each test. On the day
to exhaustion, before a recreational football intervention, before testing, the participants refrained from performing
consisting of 2–3 60-min weekly training sessions in the vigorous physical activity.
form of 45-min small-sided games played on an artificial The proposed Yo-Yo intermittent tests differ in their ini-
pitch (7v7; 43 × 27 m pitch, 83 m2 per player). Allocation tial running speed and progression, and the between-bouts
in the training groups was random, and thus, not guided (40 m) recovery lasts 5–10 s, during which the partici-
by participants’ baseline fitness to satisfy the principle of pants are asked to cover 5–10 m, respectively. The Yo-Yo
spontaneous playing of recreational football. Moreover, a intermittent test protocols were implemented according to
population-tailored continuous incremental treadmill test the procedures suggested by Krustrup et al. (2003, 2006,
(TT) was used to assess the individual V ­ O2max values under 2015).
laboratory conditions. Post-intervention, only 32 out of the The TT (HP Cosmos Quasar, Nussdorf, Germany) con-
original 66 participants performed all the tests (i.e., the three sisted of 3 min of walking at 5 km h−1 and 2 min of run-
Yo-Yo tests and the TT). ning at 8 km h−1 with 0% inclination, and then alternating
An additional group of 30 male age-matched recreational between increase in speed (1 km h−1) and inclination (1%)
football players that afterwards started the 12-week recre- every 30 s until voluntary exhaustion. Expired respira-
ational football program was evaluated at baseline to test tory gas fractions were measured using an open-circuit
cross-validation. breath-by-breath automated gas analysis system (Quark
This design was set to examine, for each of the considered CPET, Cosmed, Rome, Italy). Attainment of ­VO2max was
field tests, the following categories of validity: (1) pre- and assumed when participants achieved a plateau in ­VO 2
post-training criterion validity, (2) test external responsive- despite an increase in exercise intensity and at least one
ness and (3) test cross-validation (Impellizzeri and Marcora of the following criteria: a respiratory exchange ratio
2009). (RER) greater than 1.10 and RPE equal to or higher than
Furthermore, test specificity was assessed using the ROC 7, and HR peak equal or superior to 90% of theoretical
statistic, considering ­VO2max as a dichotomising variable maximal HR (Midgley et al. 2006, 2007). The highest 15-s
with categorisation carried out according to the procedures ­VO2 during the final stages of the test was considered for
provided by Herdy and Caixeta (Herdy and Caixeta 2016). the individual V­ O2max calculations (Midgley et al. 2006,
This with the purpose to find cutoff thresholds to charac- 2007). Data analysis was performed with manual inspec-
terize participants’ fitness level ­(VO2max) from the Yo-Yo tion of each TT data file using an Excel file (Microsoft,
performance (distance covered). Redmont, USA). All the participants were able to meet
Rating of perceived exertion (RPE) was assessed using the above criteria and the attainment of 90% of the age-
the Börg CR10 scale (Börg et al. 2013) after each imple- predicted maximal HR during the TT (Tanaka et al. 2001).
mentation of the considered Yo-Yo tests. To obtain detailed The average TT duration was within the range suggested
information concerning the perceived demands of the test- for ­VO2max assessment (10.7 ± 1.5 min; 95% CI 9.8–10.7)
ing, the participants were asked to rate their perceived effort (Midgley et al. 2006, 2007). During all tests, the HRs were
immediately after exhaustion during the Yo-Yo tests, pro- recorded at 1-s intervals using Polar Team System 2 HR
viding scores for their muscular ­(RPEM), cardiorespiratory monitors (Polar Electro Oy, Kempele, Finland) to assess
­(RPECR) and global ­(RPEG) perception (Weston et al. 2015; exercise cardiovascular demands.
Castagna et al. 2017). Before the beginning of the evaluations, the participants
performed submaximal versions of the selected Yo-Yo
Testing procedures tests and the TT to familiarize themselves with the testing
procedures. The participants were allowed to drink water
The Yo-Yo intermittent and TT tests were performed in ad libitum to ensure proper hydration under all the exer-
random order with at least 4 days’ recovery in between. All cise conditions considered in this study. Immediately after
field tests were carried out on the same artificial turf foot- each test and 15–20 min post-match, RPE was assessed as
ball pitch and at the same time of day to avoid circadian global and differential (cardiorespiratory and muscular)
effects on Yo-Yo test performance. A standardized warm- RPE (Castagna et al. 2017).
up consisting of 10 min of 20 m back and forth running

13
 European Journal of Applied Physiology

Statistical analyses 21 participants (P = 0.001; 0.90 power for r = 0.80) for the
criterion validity assessment at pre- and post-intervention
All the reported data are expressed as means ± standard devi- as well as for the cross-validation calculations and 25 par-
ations and 95% confidence intervals (95% CI), unless stated ticipants (P = 0.001; 0.90 power for r = 0.80), for external
otherwise. The Shapiro–Wilk W test was used to verify nor- responsiveness, respectively.
mality assumption. Comparisons between data sets were
performed with one-way repeated-measurements analysis of
variance (ANOVA) with a post hoc Bonferroni test. Cohen’s
d was used to evaluate the effect size, with values above 0.8, Results
between 0.8 and 0.5, between 0.5 and 0.2, and lower than 0.2
considered as large, moderate, small and trivial, respectively Criterion validity
(Cohen 1988). Pearson correlation (r) and regression equa-
tions were used to assess the associations between variables. The participants’ pre-intervention mean V ­ O 2max was
The magnitude of the reported effects was described using 41.2 ± 6.2 ml kg −1 min −1 (39.7−42.8, n = 66) with a
the Hopkins et al.’s (Hopkins et al. 2009) criteria as follows: median of 41.4 ml kg−1 min−1 (38.8–43.0). In the YYIE1,
trivial, r < 0.1; small, 0.1 < r < 0.3; moderate, 0.3 < r < 0.5; YYIE2 and YYIR1 tests performed during the pre-
large, 0.5 < r < 0.7; very large, 0.7 < r < 0.9; nearly perfect, training assessment, the participants covered 1602 ± 678
r > 0.9; and perfect, r = 1.0. Within test conditions, variabil- (1436−1769), 447 ± 174 (405–491) and 650 ± 286 m
ity was expressed as coefficient of variation (%CV). Estima- (581−721), respectively. Very large associations were
tion error was reported as typical error of estimation (TEE) reported between the YYIE1 and the YYIE2 (r = 0.80;
according to the procedures proposed by Hopkins et al. 0.69−0.87, P < 0.0001) and YYIR1 (r = 0.87; 0.80−0.92,
(Hopkins et al. 2009). Relative consistency was assessed P < 0.0001), respectively. Similarly, YYIR1 performance
using the intraclass correlation coefficient (­ ICC3.1) with 95% was very largely associated with YYIE2 distance (r = 0.87;
CI (Weir 2005; Hopkins 2000). According to Fleiss (Fleiss 0.79−0.92, P < 0.0001). Table 1 shows the regression sta-
2011), ICC values of 0.75–1.00 were considered excellent, tistics of the equations depicting the associations between
0.41–0.74 good and 0.00–0.40 poor. ­VO 2max and the field tests. The Yo-Yo tests were very
The smallest worthwhile change (SWC) in measure- largely and significantly (P < 0.0001) associated with
ments was considered to test the practical difference ­VO2max values. The magnitude of the typical error of esti-
between variables and calculated as 0.2 times the standard mation among the regression equations ranged from small
deviation of the variable mean (Hopkins et al. 2009). The to moderate (0.98−1.45 METs). The participants achieved
ROC curve statistic was used to explore the possibility of 187 ± 10 (184–189), 183 ± 10 (180–185), 180 ± 10
detecting a cutoff measure for Yo-Yo intermittent tests to (177–182) beats·min −1 during the YYIE1, YYE2 and
characterize the individual level of aerobic fitness (i.e., YYIR1, respectively. This corresponds to 99 ± 2, 97 ± 3
­VO2max). For this purpose, ­VO2max dichotomisation was and 95 ± 3% of the participants’ maximal HR during the
performed using the procedures suggested by Herdy and YYIE1, YYIE2 and YYIR1 tests, respectively.
Caixeta (Herdy and Caixeta 2016), considering the group At post-intervention, the associations between Yo-Yo
median as reference. Values above and below the group intermittent tests and ­VO2max were very large and signifi-
­VO2max median were considered as good to excellent and cant (Table 2), with absolute values higher than at base-
fair to very poor, respectively. Significance was set at 5% line. The corresponding TEE of the regression equations
(p < 0.05) for all calculations. Power calculation suggested ranged from small to moderate (0.74−1.44 METs).

Table 1  Regression statistics YYIE1 YYIE2 YYIR1

(n = 66) and association
between field tests performance Pearson r 0.77*** (0.65−0.85) 0.75*** (0.62 − 0.84) 0.75*** (0.62−0.84)
and players’ ­VO2max at
Intercept 29.9 (27.4−32.5) 28.1 (25.1−31.2) 30.7 (28.1−33.2)
pre-intervention (baseline)
conditions Slope 0.007 (0.006−0.009) 0.029 (0.023−0.036) 0.016 (0.013−0.020)
TEE (ml kg−1 min−1) 4.0 (3.4−4.9) 4.1 (3.5−5.0) 4.2 (3.6−5.1)
TEE (standardized) 0.7 (0.6−0. 8) 0.7 (0.6−0.8) 0.7 (0.6−0.8)
TEE as CV% 10.3 (8.7 −12.6) 10.5 (8.9−12.8) 10.5 (8.9−12.8)

TEE typical error of estimate, CV coefficient of variation, YYIE1 Yo-Yo intermittent endurance test level 1,
YYIE2 Yo-Yo intermittent endurance test level 2, YYIR1 Yo-Yo intermittent recovery test level 1
***P < 0.0001; (95% CI)

13
European Journal of Applied Physiology

Table 2  Regression statistics YYIE1 YYIE2 YYIR1

and association between field
tests performance and players’ Pearson r 0.82*** (0.66−0.91) 0.81*** (0.63 −0.90) 0.76*** (0.55 −0.88)
­VO2max at post-intervention
Intercept 32.1 (28.4−35.7) 31.3 (27.2−35.3) 33.8 (29.9−37.7)
(n = 32)
Slope 0.006 (0.005−0.008) 0.020 (0.015−0.026) 0.012 (0.008−0.016)
TEE (ml kg−1 min−1) 3.3 (2.6−4.4) 3.4 (2.7−4.6) 3.8 (3.0−5.1)
TEE (standardized) 0.6 (0.5−0.8) 0.6 (0.5−0.8) 0.7 (0.5−0.9)
TEE as CV% 7.9 (6.2−10.8) 7.8 (6.2−10.7) 8.8 (6.9−12.0)

TEE typical error of estimate, CV coefficient of variation, YYIE1 Yo-Yo intermittent endurance test level 1,
YYIE2 Yo-Yo intermittent endurance test level 2, YYIR1 Yo-Yo intermittent recovery test level 1
***P < 0.0001; (95%CI)

External responsiveness (0.76−0.94), 0.89 (0.78−0.95) and 0.89 (0.79−0.95) for

YYIE1, YYIE2 and YYIR1, respectively.
The participants’ pre-to-post-intervention changes are shown
in Table 3. Large and significant (P < 0.001) differences Yo‑Yo intermittent test rating of perceived exertion
were reported at post-intervention for all the considered
variables. After 12 weeks of recreational football practice, The ­RPEG was 7.5 ± 1.5 (7.1−7.9), 7.6 ± 1.4 (7.2−8.0) and
8, 25, 44 and 40% increases were shown in ­VO2max, YYIE1, 7.8 ± 1.3 (7.5−8.2) for YYIE1, YYIE2 and YYIR1, respec-
YYIE2 and YYIR1, respectively. The absolute and relative tively. The differences across the Yo-Yo tests in ­RPEG
changes in YYIE2 performance were largely associated with were trivial (d = 0.084−0.20, P = 0.07−0.49). ­RPECR for
those shown in ­VO2max (r = 0.65; 0.38−0.82, P < 0.0001 YYIE1, YYIE2 and YYIR1 was 8.1 ± 1 (7.6−8.5), 8.2 ± 1.3
and r = 0.64; 0.37−0.82, P < 0.0001, respectively). Small (7.9−8.5) and 8.2 ± 1.4 (7.8−8.5), respectively. Differ-
associations were reported between pre-to-post changes in ences between tests in R ­ PECR were trivial (d = 0.0−0.1,
YYIE1 performance and V ­ O2max training variations when P = 0.55−0.95). The ­R PE M corresponded to 6.5 ± 2.1
expressed as either absolute or relative values (r = 0.34; (6.0−7.1), 6.7 ± 2.0 (6.2−7.2) and 6.7 ± 2.0 (6.2−7.2) after
− 0.02−0.62, P = 0.07 and r = 0.35; − 0.01−0.63, P = 0.06, YYIE1, YYIE1 and YYIR1, respectively. Between-test dif-
respectively). Associations between training-induced vari- ferences in ­RPEM were trivial (d = 0.0−0.1, P = 0.41−0.88).
ations in YYIR1 performance and V ­ O2max were moderate Differences across the differential RPE within each of the
(r = 0.47; 0.14−0.71, P = 0.008 and r = 0.50; 0.16−0.73, selected Yo-Yo tests were significant and large and indicated
P = 0.005 for absolute and relative changes, respectively). ­RPECR > RPEG > RPEM (d = 1.0−2.1, P = 0.01−0.0001).

Yo‑Yo test cross‑validation

Discussion
The regression equations (Yo-Yo intermittent test perfor-
mance vs ­VO2max), developed with the 66 participants’ base- This is the first study to address several levels of criterion
line data supplemented with data from the extra 30 partici- validity and external responsiveness of three Yo-Yo intermit-
pants who performed the three Yo-Yo intermittent tests and tent tests (YYIE1, YYIE2 and YYIR1) in the same research
the treadmill test for V
­ O2max assessment, revealed trivial to design and with adult male recreational football players. The
small, not significant differences between the observed and results suggested the validity of the considered Yo-Yo inter-
estimated ­VO2max values (Table 4). TEE ranged from 1.03 mittent tests with information helpful for supporting their
to 1.83 METs (moderate). contextual and aim-guided choice. Specifically, this study
revealed Yo-Yo intermittent tests’ interchangeability, with
Yo‑Yo intermittent test specificity between-performance associations at the upper end of the
very large range of the magnitude-based inference category
The ROC analyses reported cutoff distances of 1760, 480 (0.80–0.87, P < 0.0001). Moreover, at pre-intervention (i.e.,
and 600 m for YYIE1, YYIE2 and YYIR1, respectively. under untrained conditions), criterion validity, assumed as
The correspondent measure of test sensitivity and speci- the individual association between Yo-Yo maximal perfor-
ficity for YYIE1, YYIE2 and YYIR1 was 63.64/93.94%, mance and V ­ O2max, was large to very large, indicating the
69.70/96.97% and 72.73/87.88%, respectively. The area suitability of these field tests for providing valid information
under the curve (AUC) for the Yo-Yo intermittent tests about participants’ aerobic fitness. The pre-intervention level
was significant (P = 0.0001) and corresponded to 0.87 of criterion validity was confirmed with test cross-validation

13
 European Journal of Applied Physiology

resulting in no significant differences between observed and

ICC intraclass correlation coefficient, TEM%CV typical error of measurement as a percentage of the CV, YYIE1 Yo-Yo intermittent endurance test level 1, YYIE2 Yo-Yo intermittent endurance
estimated ­VO2max values. Criterion validity magnitude was
maintained post-intervention (“longitudinal” criterion valid-
ity), with large to nearly perfect associations between Yo-Yo
performance and ­VO2max values. Test external responsive-
SWC

114

121
1.2

40

59
ness reported differences in the magnitude of the relation-
2

ships between pre-to-post-intervention Yo-Yo intermittent

14.0 (11.1; 19.3)

17.8 (13.9; 24.6)

21.4 (16.7; 29.8)

test performance and ­VO2max changes (Fanchini et al. 2014,
1.8 (1.5; 2.4)

5.6 (4.5; 7.6)

5.6 (4.5; 7.6)

TEM%CV

2015; Impellizzeri and Marcora 2009). Indeed, only the

changes in YYIE2 performance showed a large associa-
tion with variations in the criterion values, with YYIR1 and
YYIE1 showing moderate and small associations, respec-
tively. Interestingly, exertion perception in the considered
0.84*** (0.70;

0.87*** (0.75;

0.89*** (0.78;

0.78*** (0.58;

0.67*** (0.42;
(0.95–0.99)

categories (global, cardiorespiratory and muscular) did not

0.98***

differ across the field testing conditions (i.e., Yo-Yo inter-

0.92)

0.93)

0.95)

0.89)

0.83)
ICC

mittent tests).
The Yo-Yo intermittent tests are considered to be the
most thoroughly investigated field tests in football, with
­ O2max and Yo-Yo intermittent test performances values (n = 32)

data reported across recreational and competitive levels,

age and gender (Schmitz et al. 2018). With reference to
the published literature, this study’s YYIE2 performance
0.64

0.96

0.81

1.19

1.28

0.90
d

results were remarkably lower than those reported in top-

elite football players (2438 ± 531 m) and lower than those
reported in younger (18–26 years) male inactive participants
% Difference

(670 ± 267 m) (Schmitz et al. 2018; Krustrup et al. 2015).

1.3 ± 2.5

8.0 ± 8.5

8.0 ± 8.5

25 ± 24

44 ± 33

40 ± 40

In the YYIR1, the participants covered a distance similar

to that reported in inactive younger (18–25 years) male
participants (706 ± 504 m) (Schmitz et al. 2018). In adult
(− 501; − 254)

(− 234; − 126)

(− 345; − 140)

test level 2, YYIR1 Yo-Yo intermittent recovery test level 1, SWC smallest worthwhile change
(− 4.4; − 1.9)
− 1.21 (− 1.99;

240 ± 52 (133;

competitive football players, a distance of 1890 ± 457 m was

− 377 ± 330

− 180 ± 145

− 243 ± 274
− 3.1 ± 3.3
Post-intervention Difference

reported, corresponding to ~ 190% more coverage than the

− 0.44)

recreational players considered in this study (Schmitz et al.

346)

2018). Greater distances were reported in top-elite football

players (2302 ± 509 m) (Schmitz et al. 2018). When per-
forming the YYIE1, male football players reported distances
2025 ± 742***

664 ± 224***

937 ± 368***
44.8 ± 5.7***
81.7 ± 8.9***

in the range of 1890–3044 m, revealing the feasibility and

3692 ± 543
Table 3  Players’ pre-to-post-intervention changes in body mass, V

performance variability of this test protocol in football play-

ers across the competitive levels and ages (Schmitz et al.
2018). This study’s participants reported a mean YYIE1
covered distance that was lower than that covered by elite
and amateur football players, but remarkably higher than
3452 ± 571

1648 ± 604

484 ± 199

695 ± 294
82.9 ± 9.8

41.7 ± 5.9

inactive female participants (535 ± 332 m, n = 543) (Schmitz

et al. 2018). These data, together with the baseline V ­ O2max
results, confirm the untrained status of this study’s partici-
***P < 0.001; (95% CI); d = Cohen’s d

pants, promoting the internal validity of this study (Herdy

Pre-intervention

and Caixeta 2016; Schmitz et al. 2018).

Test interchangeability should be a preliminary feature
when looking at test inclusion in an evaluation test battery.
VO2max (ml kg−1 min−1)

Indeed, collinearity (i.e., very large between-tests associa-

VO2max (mL min−1)

tion) promotes test redundancy and collection of data that

Body mass (­ kg−)

may result of limited further interest (Impellizzeri and Mar-

YYIR1 (m)
YYIE1 (m)

YYIE2 (m)

cora 2009). Thus, the evaluation of test interchangeability

Variable

is a useful procedure to guide the selection of field tests to

be included in assessment batteries. The considered Yo-Yo

13
European Journal of Applied Physiology

Table 4  Cross-validation values YYIE1 YYIE2 YYIR1

performed using the calibration
equation of Table 2 in an Pearson r 0.65*** (0.39−0.81) 0.64*** (0.37−0.81) 0.59*** (0.30−0.78)
independent group of 30 players
Estimated ­VO2max 43.8 ± 5.3 47.2 ± 6.8 42.4 ± 1.6
that performed the three Yo-Yo
intermittent tests considered in Observed ­VO2max 45.3 ± 5.8 45.3 ± 5.8 45.3 ± 5.8
this study Difference 1.5 ± 4.7 (− 0.2 to 3.2) 1.9 ± 5.2 (− 0.1 to 3.9) 3.0 ± 5.5 (1.1 −4.8)
d = 0.32; P = 0.08 d = 0.36; P = 0.06 d = 0.85; P = 0.002
TEE 3.3 (2.7–4.4) 3.9 (3.9–5.1) 3.6 (2.9–4.8)
TEE (standardized) 0.7 (0.6−1.0) 0.8 (0.6−1.0) 1.5 (1.2−2.0)
TEE as CV% 7.9 (6.3−10.6) 8.7 (6.9−11.7) 8.3 (6.6−11.2)

TEE typical error of estimate, CV coefficient of variation, YYIE1 Yo-Yo intermittent endurance test level 1,
YYIE2 Yo-Yo intermittent endurance test level 2, YYIR1 Yo-Yo intermittent recovery test level 1
***P < 0.0001; (95% CI); d = Cohen’s d

intermittent tests were very largely mutually associated, intermittent tests may help in evaluating the endurance level
suggesting a high degree of common shared variance in of the participants, grouping them in terms of initial aerobic
their performance. These results seem to question the spe- performance to optimize training outcomes, and in evalu-
cific use of one of these examined Yo-Yo intermittent tests, ating intermittent high-intensity endurance performance
suggesting a general intermittent high-intensity ability in development during the training process (Milanovic et al.
untrained individuals. However, test duration may inform 2015). The criterion validity of Yo-Yo tests has been the
test inclusion for practical issues and physiological demands topic of several investigations providing valuable informa-
(Krustrup et al. 2003, 2006, 2015). In this study, the average tion regarding the suitability of their use in specific popula-
duration of the considered tests was 12.0 ± 4.7 min (95% CI tions (Bangsbo et al. 2008; Schmitz et al. 2018). With the
10.9–13.2), 2.9 ± 1.1 min (95% CI 2.7–3.2) and 5.4 ± 2.4 min aim of testing different levels of criterion validity, the con-
(95% CI 4.4–5.7) for YYIE1, YYIE2 and YYIR1, respec- structs of baseline and longitudinal criterion validity were
tively. Although no physiological measurements were per- considered in this study. Baseline criterion validity, con-
formed during the tests besides HR assessment, test duration sidered as the association between Yo-Yo performance and
may suggest YYIE1 as a mainly aerobic test and YYIE2 and individual ­VO2max in the preliminary phases of the imple-
YYR1 as possessing a strong anaerobic nature (Krustrup mentation of a recreational football intervention, revealed
et al. 2003, 2006, 2015). Biopsy studies provided mecha- global large to very large associations. Furthermore, the
nistic evidence of the association between-test duration resulting regression equations provided TEEs that are prac-
and aerobic involvement, with YYIE2 performance being tically acceptable considering the sustainability offered by
heavily supported by the anaerobic metabolism pathway in the Yo-Yo intermittent tests (Hopkins et al. 2009). However,
untrained individuals (Krustrup et al. 2015). In light of this the estimation error was higher than the SWC predicted from
study, tests’ durations and available physiological literature baseline ­VO2max values (0.35 vs 0.98 − 1.45 METs, respec-
on Yo-Yo tests, YYIE2 and YYIE1 may be positioned at the tively), suggesting caution when using regression equations
lower and upper ends of the anaerobic–aerobic continuum, to estimate aerobic fitness (Bangsbo et al. 2008). Baseline
respectively (Krustrup et al. 2003, 2015). Despite its conven- criterion validity was confirmed post-intervention, with the
ience and physiological nature, YYIE1 has been addressed associations ranging from large to nearly perfect, revealing
by only 3% of the published studies that used or examined Yo-Yo test longitudinal criterion validity.
Yo-Yo tests (Schmitz et al. 2018). From the collected data, These study results confirm the Yo-Yo intermittent tests
YYIE1 may be considered, in terms of duration and distance baseline criterion validity shown in football populations
covered, similar to YYIR1 in male competitive football play- and in untrained participants reported in previous validation
ers (Castagna et al. 2006a; Krustrup et al. 2003, 2006, 2015; studies. Indeed, Krustrup et al. (2003) reported very large
Thomas et al. 2006). associations (r = 0.71, n = 15) between YYIR1 performance
The Yo-Yo intermittent tests were originally proposed and ­VO2max in elite football players. These results were simi-
as sustainable, low-cost field tests for assessing football- lar (r = 0.74) to those reported by Rampinini et al. (2010) in
specific endurance (Bangsbo et al. 2008; Schmitz et al. male competitive football players. An effect of training status
2018). Their interest applies not only to competitive, but on the magnitude of the association was reported by Thomas
also recreational football, and particularly when dealing with et al. (2006), who found very large values (r = 0.87, n = 19)
large populations (Krustrup et al. 2009, 2010, 2018; Krus- in recreational active participants. A moderate association
trup and Krustrup 2018). In recreational football, the Yo-Yo (r = 0.46, n = 24) was reported by Castagna et al. (2006b) in

13
 European Journal of Applied Physiology

amateur football players. Large associations (r = 0.63–0.65), more aerobic tests did not follow the same trend in tracking
were reported in young football players for baseline criterion ­VO2max changes in this population of recreational football
validity in YYIE1 (Castagna et al. 2006a; Wong et al. 2011). players. Given the interest of this issue and the reduced num-
In untrained participants with ­VO2max similar to this study’s ber of participants involved in the responsiveness assessment
participants (40.0 ± 6.2 ml·kg−1·min−1), Krustrup et al. (32 out of 66), further studies are warranted.
(2015) reported very large associations for baseline criterion A practical feature of field tests is their ability to catego-
validity (r = 0.77, P < 0.01) for YYIE2. Interestingly, small rize the results to classify participants’ fitness (Herdy and
non-significant associations were found in the same study Caixeta 2016; Nes et al. 2014). In this study, the ROC statis-
when assessing baseline criterion validity in football players tics were used, assuming group ­VO2max to dichotomise the
(60.9 ± 5.5 ml kg−1 min−1). In a population of professional participants’ fitness level (Impellizzeri and Marcora 2009).
football players, Bradley et al. (2011) reported a large and This procedure enabled the determination of performance
significant association (r = 0.52, P < 0.01) between YYIE2 thresholds, with participants achieving a distance equal to
distance and ­VO2max. Despite the significantly lower YYIE2 or greater than 1760, 480 and 600 m in YYIE1, YYIE2 and
distance in the untrained participants, test performance was YYIR1 tests, considered to possess good to excellent V ­ O2max
related to aerobic metabolism enzymes, as confirmed by the values, respectively (Herdy and Caixeta 2016). These per-
large association (r = 0.57, P < 0.01) between YYIE2 perfor- formance references have a great practical value for motiva-
mance and muscle citrate synthase concentration (Krustrup tion and for the fitness assessment of recreational football
et al. 2015). The results of this study are globally in line with players. The popularity of the Yo-Yo tests and the practical
the above-reported baseline criterion validity studies, addi- interest of fitness categorisation in research and daily prac-
tionally providing longitudinal criterion and cross-validation tice warrant further studies with larger cohorts.
data to further support the validity (i.e., criterion) of the As previously discussed, test choice must or should be
considered Yo-Yo intermittent tests (Bangsbo et al. 2008; performed taking into consideration a number of features
Krustrup et al. 2015; Schmitz et al. 2018). supporting test aim, validity, reliability and sustainability
Test external responsiveness is considered as the ability (Impellizzeri and Marcora 2009; Póvoas et al. 2016, 2018).
of a test to measure changes in the reference variable (Impel- The population characteristics (i.e., physical and physiologi-
lizzeri and Marcora 2009). As far as the authors of this study cal), the expected demands of the test, and the impact of test-
are aware, only a few papers have addressed the external ing on the targeted participants must be in the checklist of
responsiveness of Yo-Yo intermittent tests (Krustrup and sports scientists and fitness instructors when implementing
Bangsbo 2001; Fanchini et al. 2014, 2015). Furthermore, a test battery. This particularly when the tests are maximal
none of them have examined this level of validity taking into in nature and the supposed participants are evaluated in their
consideration training-induced changes in ­VO2max. This level untrained status, as in most recreational football studies. The
of validity supports the usefulness of a field test across time, participants’ perceived exertion may guide this preliminary
defining the ability of a test to describe variation in the crite- evaluation of the “cost–benefit” of implementing a maxi-
rion variable (gold standard, V ­ O2max) (Impellizzeri and Mar- mal test in untrained participants. With the aim of obtaining
cora 2009). In this study, a large association was observed information on the perceived exertion in the Yo-Yo intermit-
between YYIE2 and V ­ O2max changes, while the correspond- tent tests, different forms of RPE have been used (Castagna
ing magnitudes were small and moderate for YYIE1 and et al. 2017; Weston et al. 2015). Interestingly, participants’
YYIR1, respectively. The corresponding magnitude of the reports have shown that differential RPE scores across the
association expressed considering their 95% CI ranged from different Yo-Yo intermittent tests were not different when
moderate to very large, trivial to large and trivial to very considering either the practical (Cohen’s d, trivial) and
large for the YYIE2, YYIE1 and YYIR1, respectively. This the null hypothesis (not significant, P > 0.05) approach. At
data depict the functional interest of the considered Yo-Yo baseline assessment, the participants perceived the Yo-Yo
intermittent tests in recreational football players’ aerobic intermittent tests as a mainly cardiorespiratory-demanding
fitness assessment, suggesting YYIE2 as the potentially ­(RPECR) effort. Indeed, ­RPECR was higher than ­RPEG and
more responsive test of the three field tests here proposed. ­RPEM (large, P < 0.001). This is line with the physiological
With this study design, it is difficult to provide a mechanis- aim of the proposed Yo-Yo intermittent tests, which origi-
tic explanation for the different responsiveness of the field nally were designed to evaluate participants’ endurance
tests here considered. Interestingly, the most demanding and, (both aerobic and anaerobic) and thus, heavily stressed the
supposedly, more anaerobic test, the YYIE2 test, reported cardiovascular system (Bangsbo et al. 2008; Castagna et al.
greater responsiveness. We may speculate that probably 2006a, b; Krustrup et al. 2003, 2006, 2015). Despite the
the training intervention caused physiological changes that differences in testing time to exhaustion due to test protocol
improved the aerobic contribution during the YYIE2. How- speeds, recovery times and accumulated number of 20-m
ever, it is challenging to explain why the other potentially shuttle runs across the Yo-Yo tests, R ­ PEM did not differ

13
European Journal of Applied Physiology

across the selected versions of the Yo-Yo intermittent tests in the preliminary assessment of individual aerobic fitness in
and was of lower magnitude than R ­ PECR. The magnitude previously untrained recreational football participants. The
(trivial) of the reported differential RPE across the pro- initial stage of the test requiring lower running speed may be
posed Yo-Yo intermittent tests suggests their equivalence in a practical feature of YYIE1 for promoting its use in baseline
terms of subjective internal load. According to the reported assessments. Moreover, results obtained by this study’s authors
differential RPE scores within each of the selected Yo-Yo have shown that the YYIE1 may be the best choice for assess-
intermittent tests, exhaustion in these tests was probably ing individual H ­ Rmax in recreational football players (Póvoas
induced by central, mainly cardiovascular, variables, sup- et al. 2019). In light of this study’s results, YYIR1 should not
porting the internal validity of this study and the subjective be the first choice when dealing with male untrained partici-
validity of these tests (i.e., face validity) (Impellizzeri and pants involved in a recreational football intervention. This is
Marcora 2009). Future studies on the consequential validity due to the lower consistency of the improvement rank order
of the Yo-Yo intermittent tests in recreational football play- in aerobic fitness (ICC = 0.67, good) promoted by the recrea-
ers should address delayed post-test onset of muscle soreness tional football training.
to confirm this study data (Messick 1995). According to this study’s findings, the use of intermittent
Yo-Yo tests in recreational football is supported with respect to
their validity and reproducibility, suggesting a differential use
Conclusion in the training process to test participants’ progress. The use
of YYIE1 may be advisable at baseline assessment of aerobic
The results of this study suggest the validity, usefulness fitness, introducing YYIE2 in the later stages of the training
and sustainability of the intermittent Yo-Yo tests consid- control and regulation process. This proposed staged and pro-
ered in this study when dealing with participants enrolled gressive use of Yo-Yo intermittent tests, although reducing
in recreational football interventions. Although the Yo-Yo the information collected in the follow-up process for research
performances showed very large inter-test associations, the purposes, may have some practical interest in daily practice.
various degrees of validity, addressed for the first time in this Practitioners may use the provided fitness category thresh-
study in combination, may result of some use in guiding test olds to gain information on the participants’ aerobic fitness
choice in recreational football players. The baseline and lon- at baseline and in the re-assessment when adopting the pro-
gitudinal validity of YYIE1, YYIE2 and YYIR1 were equiv- posed differential use of the Yo-Yo intermittent tests. Given
alent, with very large magnitudes reported. Furthermore, the observed training effect on the associations between Yo-Yo
cross-validation using an independent group of recreational performance and V ­ O2max, we may speculate that training inter-
football players not included in this study’s pre-to-post- ventions considering higher training volume or intensity, pos-
intervention calculations, and with similar characteristics, sibly providing different physiological adaptations, may guide
provided encouraging results confirming the large associa- test choice as per professional players (Krustrup et al. 2015;
tion between participants’ Yo-Yo intermittent test perfor- Milanovic et al. 2015). The practical interest of the procedures
mance and V ­ O2max. However, given the magnitude of the addressed and proposed in this study warrants further focused
error in estimation, caution should be exercised when using research.
the developed equations at baseline. The encouraging mag-
nitude of association (r ≥ 0.80) between test performance
and ­VO2max at post-training warrants further confirmatory Author contributions Susana Póvoas and Carlo Castagna conceived
and designed the research. Testing and data collection was performed
studies with larger sample sizes (i.e., n > 32) to address the by Susana Póvoas. Carlo Castagna and Susana Póvoas analyzed the data
general interest of predictive equations in training prescrip- and wrote the first draft of the manuscript. Peter Krustrup contributed
tion (Bangsbo et al. 2008). for the study design, interpreted the results, and edited the manuscript.
Test external responsiveness assessment promoted the All authors read and approved the final version of the manuscript.
interest of YYIE2 with its moderate to very large r values
(95%CI 0.38–0.82). It might be speculated that this result Compliance with ethical standards
may be due to the mainly anaerobic nature of YYIE2 in
Conflict of interest The authors do not have any conflict of interests
untrained participants and its consequently higher sensitiv- and state that the results of the present study do not constitute endorse-
ity to training-induced changes in aerobic fitness (Krustrup ment by EJAP. The authors alone are responsible for the content and
et al. 2015). The reported 44% improvement in YYIE2 perfor- writing of the manuscript.
mance (large change) may support the above speculations and
the reported excellent ICC warrant consistency in individual
improvements in the considered participants. In this regard,
the reproducibility of YYIE1 was shown to be nearly perfect
(ICC = 0.89, very large to nearly perfect), suggesting its use

13
 European Journal of Applied Physiology

References Krustrup P, Mohr M, Nybo L, Jensen JM, Nielsen JJ, Bangsbo J (2006)
The Yo–Yo IR2 test: physiological response, reliability, and appli-
cation to elite soccer. Med Sci Sports Exerc 38(9):1666–1673
Bangsbo J, Iaia FM, Krustrup P (2008) The Yo–Yo intermittent
Krustrup P, Nielsen JJ, Krustrup BR, Christensen JF, Pedersen H,
recovery test : a useful tool for evaluation of physical perfor-
Randers MB, Aagaard P, Petersen AM, Nybo L, Bangsbo J (2009)
mance in intermittent sports. Sports Med 38(1):37–51
Recreational soccer is an effective health-promoting activity for
Börg G, Hassmen P, Lagerstrom M (2013) De novo transcript
untrained men. Br J Sports Med 43(11):825–831. https​://doi.
sequence reconstruction from RNA-seq using the Trinity plat-
org/10.1136/bjsm.2008.05312​4
form for reference generation and analysis. Nature Protocols
Krustrup P, Christensen JF, Randers MB, Pedersen H, Sundstrup
56(6):679–685
E, Jakobsen MD, Krustrup BR, Nielsen JJ, Suetta C, Nybo L,
Bradley PS, Mohr M, Bendiksen M, Randers MB, Flindt M, Barnes
Bangsbo J (2010) Muscle adaptations and performance enhance-
C, Hood P, Gomez A, Andersen JL, Di Mascio M, Bangsbo J,
ments of soccer training for untrained men. Eur J Appl Physiol
Krustrup P (2011) Sub-maximal and maximal Yo–Yo intermit-
108(6):1247–1258. https​://doi.org/10.1007/s0042​1-009-1319-8
tent endurance test level 2: heart rate response, reproducibility
Krustrup P, Bradley PS, Christensen JF, Castagna C, Jackman S, Con-
and application to elite soccer. Eur J Appl Physiol 111(6):969–
nolly L, Randers MB, Mohr M, Bangsbo J (2015) The Yo–Yo
978. https​://doi.org/10.1007/s0042​1-010-1721-2
IE2 test: physiological response for untrained men versus trained
Castagna C, Impellizzeri FM, Belardinelli R, Abt G, Coutts A,
soccer players. Med Sci Sports Exerc 47(1):100–108. https​://doi.
Chamari K, D’Ottavio S (2006a) Cardiorespiratory responses
org/10.1249/MSS.00000​00000​00037​7
to Yo–yo intermittent endurance test in nonelite youth soccer
Krustrup P, Williams CA, Mohr M, Hansen PR, Helge EW, Elbe AM,
players. J Strength Cond Res 20(2):326–330
de Sousa M, Dvorak J, Junge A, Hammami A, Holtermann A,
Castagna C, Impellizzeri FM, Chamari K, Carlomagno D, Rampinini
Larsen MN, Kirkendall D, Schmidt JF, Andersen TR, Buono P,
E (2006b) Aerobic fitness and yo–yo continuous and intermit-
Rorth M, Parnell D, Ottesen L, Bennike S, Nielsen JJ, Mendham
tent tests performances in soccer players: a correlation study. J
AE, Zar A, Uth J, Hornstrup T, Brasso K, Nybo L, Krustrup BR,
Strength Cond Res 20(2):320–325
Meyer T, Aagaard P, Andersen JL, Hubball H, Reddy PA, Ryom
Castagna C, Francini L, Povoas SC, D’Ottavio S (2017) Long sprint
K, Lobelo F, Barene S, Helge JW, Fatouros IG, Nassis GP, Xu JC,
abilities in soccer: ball vs running drills. Int J Sports Physiol Per-
Pettersen SA, Calbet JA, Seabra A, Rebelo AN, Figueiredo P, Pov-
form 12(9):1256–1263. https​://doi.org/10.1123/ijspp​.2016-0565
oas S, Castagna C, Milanovic Z, Bangsbo J, Randers MB, Brito
Cohen J (1988) Statistical power analysis for the behavioral sciences,
J (2018) The “football is medicine” platform-scientific evidence,
2nd edn. Edn, Lawrence Erlbaum Associates, Hillsdale
large-scale implementation of evidence-based concepts and future
Fanchini M, Castagna C, Coutts AJ, Schena F, McCall A, Impel-
perspectives. Scand J Med Sci Sports 28(Suppl 1):3–7. https:​ //doi.
lizzeri FM (2014) Are the Yo–Yo intermittent recovery test
org/10.1111/sms.13220​
levels 1 and 2 both useful? Reliability, responsiveness and inter-
Messick S (1995) Standards of validity and the validity of standards in
changeability in young soccer players. J Sports Sci 32(20):1950–
performance assessment. Educ Meas 14(4):5–8
1957. https​://doi.org/10.1080/02640​414.2014.96929​5
Midgley AW, Mc Naughton LR, Wilkinson M (2006) Criteria and
Fanchini M, Schena F, Castagna C, Petruolo A, Combi F, McCall A,
other methodological considerations in the evaluation of time at
Impellizzeri M (2015) External responsiveness of the Yo–Yo
VO2max. J Sports Med Phys Fitness 46(2):183–188
IR test level 1 in high-level male soccer players. Int J Sports
Midgley AW, McNaughton LR, Polman R, Marchant D (2007) Cri-
Med 36(9):735–741. https​://doi.org/10.1055/s-0035-15472​23
teria for determination of maximal oxygen uptake: a brief cri-
Fleiss J (2011) Reliability of measurements. The design and analysis
tique and recommendations for future research. Sports Med
of clinical experiments. Wiley, New York, pp 1–31
37(12):1019–1028
Herdy AH, Caixeta A (2016) Brazilian cardiorespiratory fitness clas-
Milanovic Z, Pantelic S, Covic N, Sporis G, Krustrup P (2015) Is Rec-
sification based on maximum oxygen consumption. Arq Bras
reational soccer effective for improving VO2max a systematic
Cardiol 106(5):389–395. https​://doi.org/10.5935/abc.20160​070
review and meta-analysis. Sports Med 45(9):1339–1353. https​://
Hopkins WG (2000) Measures of reliability in sports medicine and
doi.org/10.1007/s4027​9-015-0361-4
science. Sports Med 30(1):1–15
Nes BM, Vatten LJ, Nauman J, Janszky I, Wisloff U (2014) A simple
Hopkins WG, Marshall SW, Batterham AM, Hanin J (2009) Progres-
nonexercise model of cardiorespiratory fitness predicts long-term
sive statistics for studies in sports medicine and exercise sci-
mortality. Med Sci Sports Exerc 46(6):1159–1165. https​://doi.
ence. Med Sci Sports Exerc 41(1):3–13. https​://doi.org/10.1249/
org/10.1249/MSS.00000​00000​00021​9
MSS.0b013​e3181​8cb27​8
Pate RR, Kriska A (1984) Physiological basis of the sex difference in
Impellizzeri FM, Marcora SM (2009) Test validation in sport physi-
cardiorespiratory endurance. Sports Med 1(2):87–98
ology: lessons learned from clinimetrics. Int J Sports Physiol
Póvoas SC, Castagna C, da Costa Soares JM, Silva P, Coelho ESMJ,
Perform 4(2):269–277
Matos F, Krustrup P (2016) Reliability and construct validity
Krustrup P, Bangsbo J (2001) Physiological demands of top-class
of Yo–Yo tests in untrained and soccer-trained schoolgirls aged
soccer refereeing in relation to physical capacity: effect of
9–16. Pediatr Exerc Sci 28(2):321–330. https​://doi.org/10.1123/
intense intermittent exercise training. J Sports Sci 19:881–891
pes.2015-0212
Krustrup P, Bangsbo J (2015) Recreational football is effective in
Póvoas SCA, Krustrup P, Castagna C, da Silva PMR, Coelho ESMJ,
the treatment of non-communicable diseases. Br J Sports Med
Pereira RLM, Larsen MN (2018) Reliability of submaximal Yo–
49(22):1426–1427. https:​ //doi.org/10.1136/bjspor​ ts-2015-094955​
Yo tests in 9- to 16-year-old untrained schoolchildren. Pediatr
Krustrup P, Krustrup BR (2018) Football is medicine: it is time for
Exerc Sci 30(4):537–545. https​://doi.org/10.1123/pes.2017-0139
patients to play! Br J Sports Med 52(22):1412–1414. https​://doi.
Póvoas SCA, Krustrup P, Pereira R, Vieira S, Carneiro I, Magalhaes J,
org/10.1136/bjspo​r ts-2018-09937​7
Castagna C (2019) Maximal heart rate assessment in recreational
Krustrup P, Mohr M, Amstrup T, Rysgaard T, Johansen J, Steens-
football players a study involving a multiple testing approach.
berg A, Pedersen PK, Bangsbo J (2003) The Yo–Yo intermittent
Scand J Med Sci Sports. 10:11. https:​ //doi.org/10.1111/sms.13472​
recovery test: physiological response, reliability, and validity.
Rampinini E, Sassi A, Azzalin A, Castagna C, Menaspa P, Carlo-
Med Sci Sports Exer 35(4):697–705. https​://doi.org/10.1093/
magno D, Impellizzeri FM (2010) Physiological determinants
bioin​forma​tics/btp61​6
of Yo–Yo intermittent recovery tests in male soccer players. Eur

13
European Journal of Applied Physiology

J Appl Physiol 108(2):401–409. https​://doi.org/10.1007/s0042​ Weston M, Siegler J, Bahnert A, McBrien J, Lovell R (2015) The appli-
1-009-1221-4 cation of differential ratings of perceived exertion to Australian
Schmitz B, Pfeifer C, Kreitz K, Borowski M, Faldum A, Brand SM football league matches. J Sci Med Sport 18(6):704–708. https​://
(2018) The Yo–Yo Intermittent tests: a systematic review and doi.org/10.1016/j.jsams​.2014.09.001
structured compendium of test results. Front Physiol 9:870. https​ Wong P, Chaouachi A, Castagna C, Lau PWC, Chamari K, Wisløff
://doi.org/10.3389/fphys​.2018.00870​ U (2011) Validity of the Yo–Yo intermittent endurance test in
Tanaka H, Monahan KD, Seals DR (2001) Age-predicted maximal young soccer players. Eur J Sport Sci 11:309–315. https​://doi.
heart rate revisited. J Am Coll Cardiol 37(1):153–156 org/10.1080/17461​391.2010.52157​9
Thomas A, Dawson B, Goodman C (2006) The Yo–Yo test: reliability
and association with a 20-m shuttle run and VO2max. Int J Sports Publisher’s Note Springer Nature remains neutral with regard to
Physiol Perf 1:137–149 jurisdictional claims in published maps and institutional affiliations.
Weir JP (2005) Quantifying test-retest reliability using the intra-
class correlation coefficient and the SEM. J Strength Cond Res
19(1):231–240

13