Journal of Science and Medicine in Sport 22 (2019) 1030–1037

Contents lists available at ScienceDirect

Journal of Science and Medicine in Sport

journal homepage: www.elsevier.com/locate/jsams

Review

Normative Yo-Yo Intermittent Recovery Level 1 and Yo-Yo

Intermittent Endurance Level 1 test values of boys aged 9–16 years
Boris Schmitz a,∗,1 , Carina Pfeifer a,1 , Kiana Kreitz b , Matthias Borowski b , Andreas Faldum b
, Stefan-Martin Brand a
a
Institute of Sports Medicine, University Hospital Muenster, Germany
b
Institute of Biostatistics and Clinical Research, University of Muenster, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: To provide age- and sex-specific reference values of Yo-Yo tests in children and adolescents.
Received 24 January 2019 Design: Systematic review.
Received in revised form 7 April 2019 Methods: A literature search for articles on Yo-Yo Intermittent (YYI) tests was performed in MEDLINE,
Accepted 27 May 2019
SPORTDiscus, Web of Science and Google Scholar. Original reports on healthy children/adolescents
Available online 31 May 2019
6–16 years of age were eligible. For each test, age- and sex-related reference values were calculated
using global means and percentiles.
Keywords:
Results: Ninety-two studies (7398 participants) fulfilled the eligibility criteria. The YYI tests most fre-
Adolescent
Child
quently used were the Yo-Yo Intermittent Recovery Level 1 test (YYIR1, 57.8%), Yo-Yo Intermittent
Physical performance Endurance Level 1 test (YYIE1, 14.7%), Yo-Yo Intermittent Recovery Level 1 Children’s test (YYIR1C, 12.7%),
Physical fitness Yo-Yo Intermittent Endurance Level 2 test (YYIE2, 8.8%) and the Yo-Yo Intermittent Recovery Level 2 test
Physical activity (YYIR2, 5.9%). Of these, 71.6% reported test results of boys, 17.6% reported mixed test results and 10.8%
reported test results of girls. Smoothed centile curves for the YYIR1 and YYIE1 over the entire age range
were generated for boys, revealing constantly increasing performance with increasing age.
Conclusions: YYI tests values differ with respect to age and sex. In boys, development of YYIR1 and YYIE1
test values (6–16 years of age) was different, suggesting better applicability of the YYIR1 test for boys
>13 years of age. The results may be used to rate YYI test performance for continuous screening and to
identify children with low physical fitness. Since limited data was available of females, further research
on YYI tests is needed with respect to sex-specific results.
© 2019 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.

Practical implications 1. Introduction

• We provide the first collection of normative data for Yo-Yo Inter- Physical inactivity is not only a major risk factor for cardiovascu-
mittent (YYI) test results in children and adolescents. lar disease (CVD)1 and all-cause mortality and morbidity2–4 but has
• The results may be used to rate YYI test performance to monitor been described as a global pandemic.5 Recent worldwide estimates
age-related development and identify children with low physical indicate that 23% of men, 32% of women and over 80% of adolescents
fitness levels. (11–17 years of age) do not meet the World Health Organisation
• The provided data will also allow coaches to identify adolescents (WHO) recommendations on physical activity.6,7 Physical activity
with outstanding physical fitness. is an important determinant of physical fitness and physical fitness
itself is one of the most important health markers and a predictor
of morbidity and mortality of CVD and other diseases.8,9 More-
over, physical fitness is a far more objective parameter considering
that most epidemiologic evidence on physical activity and seden-
tary behaviour is based on self-reported data and thus affected by
measurement/ rating error and recall bias.10,11 In addition to the
association between physical fitness and future health in adults,
∗ Corresponding author.
an association between low childhood fitness and increased risk
E-mail address: [email protected] (B. Schmitz).
1
Contributed equally to this work.
for CVD, obesity, musculoskeletal and mental health problems in

https://doi.org/10.1016/j.jsams.2019.05.016
1440-2440/© 2019 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
 B. Schmitz et al. / Journal of Science and Medicine in Sport 22 (2019) 1030–1037 1031

adulthood exists.12–14 Therefore, timely diagnosis of substandard VȮ2max in children aged 6–9 years (r2 = 0.47).31 Overall, YYI tests
fitness levels or physical restraints in children and adolescents is show good validity for game-specific performance and VȮ2max ,
important. even if some studies exist that did not detect a correlation of YYI
Physical fitness depends on different physiological systems test performance and VȮ2max for unknown reasons.28,35,38
including the respiratory, cardiovascular, musculoskeletal, nervous Currently, no normative data for the different YYI tests is avail-
and endocrine-metabolic component and can thus be seen as an able for children and adolescents and individual test results are
integrated measure of body function.15 In addition to the associa- therefore hard to evaluate in terms of cardiorespiratory fitness.
tion of physical fitness and mortality in general, a strong correlation Thus, the primary aim of this study was to develop age- and sex-
between cardiorespiratory fitness (determined by maximal oxy- specific normative values for YYI tests in children and adolescents
gen uptake [VȮ2max ]) and mortality exists.15 Hence, different and to provide a basis for the identification of children with very low
tests have been applied to measure cardiorespiratory fitness also cardiorespiratory fitness by use of the YYI tests. Moreover, using
in children and adolescents15–18 including the 20 m shuttle run this framework, children and adolescents with very high cardiores-
test (20mSRT)12,19,20 with high levels of test-retest reliability and piratory fitness could be identified, which may also be of interest
validity.21,22 The 20mSRT is a maximal multistage filed test,23 for future athletic success.
which has been modified by including an active recovery period
to form the Yo-Yo Intermittent Recovery (YYIR) test.24 The main 2. Methods
objective of the YYIR test is to determine the ability to repeat-
edly perform intense exercise and the potential to rapidly recover 2.1. Study design and participants eligibility criteria
from such exercise.25 It has thus been frequently used in inter-
mittent sports such as soccer or basketball.24,25 During the test, A systematic review (PROSPERO, CRD42018084267) was per-
participants perform 2 × 20-m runs at increasing speed, which are formed to identify original articles reporting on performance
interspersed by 10-s periods of active recovery (2 × 5 m).24 The test testing of <16 years old using one of the five YYI test variants includ-
is performed as maximal performance test (i. e. until total exhaus- ing the YYIR1, YYIR2, YYIE1, YYIE2 and YYIR1C test. Other test
tion is reached). Two different YYIR test levels exist, namely the modifications, such as sub-maximal test versions or the original
Yo-Yo Intermittent Recovery Level 1 (YYIR1) test and the YYIR Level multistage 20mSRT without recovery periods, which has occasion-
2 (YYIR2) test. The YYIR1 test starts at a lower level26 with 4 runs ally been described as ‘Yo-Yo test’ were not considered in this
at 10 to 13 km h−1 (0–160 m) followed by 7 runs at 13.5–14 km h−1 analysis. Only reports on healthy humans with no disability (n ≥ 5)
(160–440 m) and proceeds with 0.5 km h−1 speed increments after were eligible. Data of participants, who were >16 years of age were
every 8 running bouts.25 The YYIR2 test starts with two initial not included as our previous normative analysis of YYI test per-
runs of 13 and 15 km h−1 , respectively, followed by two runs at formance in adults already considered this population.33 Articles
16 km h−1 , three runs at 16.5 km h−1 , 4 runs at 17.0 km h−1 and pro- had to be original reports (not a conference abstract, review or
ceeds with stepwise 0.5 km h−1 speed increments after every 8 book [chapter]) and be published in English. Grey literature, such as
running bouts.24,27 Additional test modifications exist mainly for websites, reference lists, or theses were not included. Articles were
the determination of endurance capacity, the Yo-Yo Intermittent excluded if they (1) reported test results in any other format than
Endurance Level 1 (YYIE1) test (starting at 8 km h−1 )28 and the YYIE maximal distance (m), time (min), speed (m s−1 ) or stages, (2) did
Level 2 (YYIE2) test (starting at 11.5 km h−1 ).29 During both YYIE not report original performance test results [i.e. percent changes,
tests, the recovery period has been shortened to 5 s (2 × 2.5 m) and etc.], (3) were not available as full-text (after an attempt to contact
the increase is reduced to 0.25 km h−1 .30 Additionally, the Yo-Yo the corresponding author), (4) presented performance data only in
Intermittent Recovery Level 1 Children’s test (YYIR1C) was devel- figures, or if (5) the test type was not clearly defined. Furthermore,
oped based on the YYIR1 test, but with a shorter running distance articles were excluded if test performance mean or standard devia-
of only 16 m and active recovery during 2 × 4.0 m.31,32 All Yo-Yo tion (SD)/standard error of the mean (SEM) or the number of tested
Intermittent (YYI) tests have the advantages of a field test method. subjects were missing or not clearly reported in the full-text. These
Testing does not require any special locations (indoor, outdoor) eligibility criteria were chosen also in accordance with the quality
and can be carried out at low costs also in groups. However, some assessment (described below).
essential requirements for good testing practice exist. As discussed
in detail recently33 these include that participants need to under-
stand the test settings and rules including criteria of termination. 2.2. Search strategy and data sources
This might be of pivotal relevance predominantly in the lowest age
groups and can best be addressed by practical familiarization with A systematic literature search was conducted (CP) using
the test. PubMed (MEDLINE database), Web of Science, SPORTDiscus with
The validity of the YYI tests can be assessed by comparison of Full Text and Google Scholar for records published up until August
test results with performance during actual competitive games. 2018. The following key words were used: ‘Yo-Yo intermittent test’
Moreover, YYI test variants have been reported valid to deter- or ‘Yo-Yo intermittent’ or ‘Yo-Yo intermittent recovery test’ or ‘Yo-
mine cardiorespiratory fitness and thus VȮ2max has been used for Yo intermittent recovery’ or ‘Yo-Yo intermittent endurance test’
test validation. For the YYIR1 test, good correlation with high- or ‘Yo-Yo intermittent endurance’ or ‘Yo-Yo Intermittent Recovery
intensity activity and total distance covered during a match has Level 1 Children’s’ or ‘YYIR1’ or ‘YYIR2’ or ‘YYIR’ or ‘YYIE1’ or ‘YYIE2’
been reported for young soccer players (age 1̃4 years, r = 0.77 and or ‘YYIE’ or ‘YYIR1C’ or ‘Yo-Yo IR’ or ‘Yo-Yo IE’ or ‘Yo-Yo IR1 or ‘Yo-
r = 0.65, respectively).34 Also in young soccer players, a good corre- Yo IE1’ or ‘Yo-Yo IR2’ or ‘Yo-Yo IE2’ or ‘Yo-Yo test’ or ‘YoYo test’.
lation with time spent at sprinting speed (>18.0 km h−1 ) during the Manual searches based on references from identified articles were
match has been described (age 1̃5 years, r = 0.63).35 For the YYIR2 also performed. Two reviewers performed full-text screening on
test, performance correlated well with the time above 85% of maxi- potential relevant reports. The individual steps of report identifi-
mal heart rate (HRmax ) during a soccer match in young male soccer cation, screening and processing are documented in the PRISMA
players (age 1̃4 years, r = 0.71).36 In terms of cardiorespiratory fit- flow-chart (Supplementary Fig. 1).39 Search results and fulfilment
ness, YYIE1 test performance showed good correlation with VȮ2max of eligibility criteria were discussed if unclear (BS and CP) until
determined in a laboratory setting (soccer players, age 1̃4 years, consensus was achieved and a third person was consulted upon
r = 0.63)37 and YYIR1C test performance also correlated well with disagreement to determine inclusion.
1032 B. Schmitz et al. / Journal of Science and Medicine in Sport 22 (2019) 1030–1037

2.3. Study selection and data extraction for each category providing the individual mean and 95% confi-
dence interval (CI) of each study as well as the calculated global
Data were extracted using Excel (Microsoft Office 2016, mean and global SD. The I2 statistic was calculated for each cat-
Microsoft Corporation, Redmond, USA) into a standard extraction egory to quantify heterogeneity within the respective individual
table by two reviewers (BS and CP) including information on first study results.41 All forest plots including I2 statistics are available
author, year of publication, description of participants (total num- in the online repository.
ber, anthropometric data, activity level), performance data and test
type. In cases where articles reported on test results over different 3. Results
age groups, separate data on participants <16 years of age were
extracted if possible. In case of redundant data reported in sepa- 3.1. Study selection and characteristics
rate publications, only data from the earliest report were extracted.
In studies where both girls and boys were tested in combina- The procedure of study identification, selection and final inclu-
tion (mixed data) and data could not be extracted separately, test sion is shown in Supplementary Fig. 1. During the screening stage,
results were not included in the quantitative analysis (Supplemen- 591 records did not meet the inclusion criteria (see methods) and
tary Fig. 1). In cases where intra- and inter-seasonal performance were removed. Of the remaining 537 reports, 445 studies were
changes were reported, the best test data were extracted. In stud- excluded (details in Supplementary Fig. 1) and 92 studies with a
ies where performance data were reported involving any type of total of 7398 participants were included in the qualitative synthe-
experimental condition with hypothesized effect on test perfor- sis. Each individual included study (or sub-study) with information
mance, only pre-intervention data were extracted if available or on author, mean age, sex, activity status and test results is presented
data of (untreated) control groups were used. Participants were in Supplementary Table 1, sorted by test and test results. Seventy-
classified by sex as well as active or inactive in accordance with three studies were included in the quantitative synthesis since 19
author’s description in the original article. studies reported mixed results of boys and girls.

2.4. Quality assessment 3.2. Results by sex, test type and age

The methodological quality of the studies was assessed using the 3.2.1. Results of girls
critical appraisal tool by Brink and Louw.40 The tool consists of 13 Of girls, only limited data (10.8% of all included studies) was
items to assess the quality of a study. Individual items can be scored available with respect to the different age categories and test types.
as ‘yes’, ‘no’ or ‘not applicable’. For our analysis, we determined For available data, test global means, SDs and quantiles for the
(in accordance with the above mentioned inclusion criteria) that YYIR1, YYIE1/2 and YYIR1C test are shown in Fig. 1, computed from
the following items (8 out of 13) had to be scored ‘yes’. A detailed 17 studies/subgroups with a total of 889 female participants.42–49
description of the subject sample was available, the reference stan-
dard was explained, the competence or qualification of rater(s) 3.2.2. Results of boys
performing the test was clarified, the stability of the variable being The majority of included studies reported test results of boys
measured was taken into account, the execution of the test was (71.6% of the studies) and results by test type and age were as
described in (sufficient) detail to permit replication, the execution follows. For the computation of YYIR1 test global means, SDs
of the reference standard was described in (sufficient) detail, the and quantiles (Fig. 2A), 92 studies or subgroups with a total of
statistical methods were appropriate and withdrawals from the 2832 participants were available.34,35,42,50–60–91 Computation of
study were explained. Studies were rated by two reviewers (BS YYIR2 test global means, SDs and quantiles (Fig. 2B) involved
and CP). Disagreements were resolved by discussion if necessary. 8 studies/subgroups reporting on 147 participants.36,50,55,92–94
Researchers were not blinded to study authors, results or publica- Twenty-four studies/subgroups reporting on 848 participants were
tion journal. available for generation of YYIE1 test global means, SDs and quan-
tiles (Fig. 2C).37,95–104 Computation of YYIE2 test global means,
2.5. Data treatment and statistical data analyses SDs and quantiles (Fig. 2D) included 13 studies/subgroups and 498
participants.35,98,105–109 For the generation of YYIR1C test global
In most cases age was reported as age at last birthday given as means, SDs and quantiles (Fig. 2), 8 studies/subgroups and 665
mean and SD. Subjects were then assigned to respective age cate- participants were available.47–49,98,110
gories. The combined age category of 6–9 years included subjects Since the YYIR1 and the YYIE1 test were the most frequently
older than 6 years of age who had not yet completed their 9th year used test types, smoothed centile curves over the complete age
of life. Each following age category then represents the next span range could be generated for these two tests and are presented
of one year, respectively. In rare cases, age was already given as a in Fig. 3A/B. Boys’ mean performance regarding YYIE1 and YYIR1
span of a year (eg. 10–11 years) and was extracted without modi- test improved steadily from age group 6–9 years to 15–16 years
fications. For overall analysis of tested participants by test type, a (Fig. 3C). For both test types, the average increase in mean
weighted mean age was generated and a corresponding SD of 0.25 test results seemed to decline when boys became older than
was assumed if age was reported over a span of a year. Data for 12–13 years. However, age-related changes, i.e. the boys’ perfor-
YYI test performance was analysed as maximal distance (m). Max- mance improvements were not identical between the YYIR1 and
imal test duration (min), speed (m s−1 ) or stages were recalculated YYIE1 (Fig. 3).
for comparison. SD was calculated from SEM using the equation
√
SD = n· SEM where n is the number of subjects. For each category 3.3. Risk of bias
(i. e. age, sex and test type) a global mean and global SD was calcu-
lated based on the reported means and SDs of the individual studies As detailed above, the quality of the included studies was
as weighted mean of the individual reported means, with weights assessed using the critical appraisal tool by Brink et al.40 The tool
built by the number of subjects per study as described.33 Each global does not incorporate a quality score and the impact of each item
SD was calculated using the formula described in Supplementary is considered individually. Since the current analysis includes test
File 1. Finally, global mean and global SD were translated into nor- results obtained from single observations (i. e. does not include data
mal quantiles for each category. Individual forest plots were created from two consecutive tests), the following 5 items were found not
 B. Schmitz et al. / Journal of Science and Medicine in Sport 22 (2019) 1030–1037 1033

Fig. 1. Yo-Yo Intermittent test results of girls by test type and age. Qualitative ratings: values above the 80th centile, ‘very high’; between the 60th and 80th centiles, ‘high’;
between the 40th and 60th centiles, ‘moderate’; between the 20th and 40th centiles, ‘low’; values below the 20th centile, ‘very low’.

applicable: blinding of raters to their own prior findings, evaluation standard and index test. All other items were scored with ‘yes’ by
of time period between reference standard and index test, testing of definition of the inclusion criteria. Other criteria for the assessment
interrater reliability and blinding of raters to the test results, vari- of bias, such us blinding of test participants or raters to test results
ation of the order of examination, and independence of reference were unfeasible for this type of test.
1034 B. Schmitz et al. / Journal of Science and Medicine in Sport 22 (2019) 1030–1037

Fig. 2. Yo-Yo Intermittent test results of boys by test type and age. Qualitative ratings: values above the 80th centile, ‘very high’; between the 60th and 80th centiles, ‘high’;
between the 40th and 60th centiles, ‘moderate’; between the 20th and 40th centiles, ‘low’; values below the 20th centile, ‘very low’.

4. Discussion to generate normative age- and sex-specific reference values. Over-

all, the YYIR1 was the most frequently performed test variant
This study systematically analysed test data of the most com- in children and adolescents up to 16 years of age with 57.84%,
mon YYI tests from 7398 children and adolescents aged 6–16 years followed by the YYIE1, which was performed in 14.71% of the
 B. Schmitz et al. / Journal of Science and Medicine in Sport 22 (2019) 1030–1037 1035

studies. The YYIR1C was performed in 12.75% of the studies

and the YYIE2 and YYIR2 were performed in 8.82% and 5.88%
of the studies, respectively. This cross-sectional approach sug-
gested that YYI tests performance increased with age and boys
tended to perform better than girls, even though data on girls was
scarce.
For the two most common tests, the YYIR1 and the YYIE1 test, a
centile framework for boys aged 9–16 years was generated which
allows qualitative interpretation of test results and the classifica-
tion of age-specific physical fitness as follows. Values above the
80th centile are rated ‘very high’, values between the 60th and 80th
centiles are rated ‘high’, values between the 40th and 60th centiles
are rated ‘moderate’, values between the 20th and 40th centiles
are rated ‘low’ and values below the 20th centile are rated ‘very
low’.12 Moreover, normal maturation-related longitudinal changes
can be monitored using these references and the effects of inter-
ventions (exercise training, general increase of regular physical
activity) may be controlled. Of note, the majority of included stud-
ies described their participants as “physically active” and reported
involvement in ball team sports most frequently represented by
soccer, rugby/football and basketball. In detail, of all tested par-
ticipants who performed the YYIR1, 98.76% were described as
physically active (male: 98.51%, female: 100%). Among participants
performing the YYIR2, 100% were physically active. For the YYIE1,
78.96% (male: 96.36%, female: 50%), for the YYIE2, 92.44% (male:
95.98%, female: 45.65%) and for the YYIR1C, 42.12% (male: 42.9%,
female: 40.15%) were physically active. Therefore, individuals per-
forming at values below the 20th centile of the YYIR1 or the YYIE1
may not necessarily be at a critical fitness level. However, physi-
cal activity in children and adolescents should be conceived as the
‘norm’ rather than physical inactivity.6,7 This is also of relevance
with respect to the reported association between physical fitness in
childhood and future health in adults including the observation that
low childhood fitness may lead to an increased risk for CVD, obesity,
musculoskeletal and mental health problems in adulthood.12–14
Of note, specific associations between 20mSRT performance and
health indicators have also been described.111 Even though the gen-
eral description of “physical activity” of the included studies was
not investigated in detail in the current analysis, an overall com-
parison of included study participants described as “inactive” to
participants described as “active” suggested that regular physical
activity may lead to at least 20% increased test results over all age
categories.
With respect to the yearly increase in test performance, it is
of interest to compare the YYIR1 test data with normative data on
20mSRT performance recently presented by Tomkinson et al.12 The
group also provided centile curves of boys (age range 9–17 years)
showing comparable characteristics to the YYIR1 test with the
largest rate of increase at the age of 12 years. Since the YYIR1
was developed based on the 20mSRT, both tests are described to
mainly estimate the anaerobic performance capacity of the partici-
pant. By contrast, the YYIE1 has been introduced to address aerobic
endurance capacity and thus cardiorespiratory fitness. Interest-
ingly, the development of age-related YYIE1 performance over
all age groups showed different characteristics compared to the
YYIR1 and 20mSRT in that a strong increase per age category in
YYIE1 performance was observed already at younger age. Of note,
the included studies for both tests, the YYIR1 and the YYIE1 test,
involved a high number of “physically active” participants (both
>95% for boys). Based on these observations and the different physi-
ological requirements of the two tests, age-specific YYIR1 and YYIE1
Fig. 3. Quantile curves for (A) Yo-Yo Intermittent Recovery Level 1 test and (B) Yo-
test results might indicate different physiological stages during
Yo Intermittent Endurance Level 1 test results by age of boys. Qualitative ratings:
values above the 80th centile, ‘very high’; between the 60th and 80th centiles, ‘high’; maturation. To this respect, it has been suggested that, irrespec-
between the 40th and 60th centiles, ‘moderate’; between the 20th and 40th centiles, tive of exercise training, aerobic capacity increases with age in
‘low’; values below the 20th centile, ‘very low’. (C) Standardised differences in mean childhood. In girls, a plateau for aerobic capacity and thus cardiores-
Yo-Yo Intermittent Recovery Level 1 (YYIR1) test and Yo-Yo Intermittent Endurance
piratory fitness is reached at about 15 years of age, while in boys
Level 1 (YYIE1) test performance of boys by age.
1036 B. Schmitz et al. / Journal of Science and Medicine in Sport 22 (2019) 1030–1037

aerobic capacity continues to increase until the age of about 18 5. Limitations

years.112 Of note, the ratio of aerobic capacity to body mass in chil-
dren is comparable to that of adults, while anaerobic capacity is Some limitations for the presented analysis may exist. First, we
significantly lower than that of adults both in comparison to aerobic have grouped tested subjects into age categories and inter-annual
capacity and in comparison to body mass.113 So while children are differences have thus not been taken into account, which leads to
very well adapted to aerobic exercise, anaerobic exercise quickly reduced accuracy especially in the youngest age group of 6–9 years.
leads to fatigue.112 This is also documented by the oxygen (O2 ) Available data on YYI test results of girls was scarce and more data
uptake kinetics during exercise in children, which is much steeper will be needed to generate test-specific normative data of girls over
early under load compared to adults.114 Thus, the YYIR1C test has all age groups. The generated centile curves for the YYIE1 did not
been developed with a shorter running distance of 16 m and shorter include original data on age categories 9–11 years and may thus be
recovery distance of 2 × 4 m but similar speed increments as the affected by the conformation process. The necessary pooling and
YYIR1 test.31 It is therefore suggested especially for the youngest transformation of data may also have affected the presented results
and/ or less fit children.31,46 However, our analysis did not iden- to some extent. Second, our analysis was based on the reported
tify a sufficient number of studies reporting on the YYIR1C test to sex and the effect of gender has not been addressed. Third, besides
generate centile curves over the entire age range of 9–16 years for methodological quality assessment of the included studies, selec-
comparison to the development of test results with age to the YYIR1 tion bias within individual studies may have occurred in terms of
and YYIE1 test. test termination (i.e. the test is not stopped at the earliest time
Since it has been suggested that response to different train- point violating test requirements) or data partitioning and report-
ing strategies may be monitored by YYI tests27 and YYI tests may ing of data subsets (i.e. reporting on best test results). Reporting and
be used to monitor age-specific development of physical fitness, publication bias may have affected the present analysis since some
effectors of test reproducibility are important. For the YYIR1 test, data/studies may have remained unreported or were not published
reported reproducibility was higher in competitive U17 soccer because of unexpected/contradictory, negative or not significant
players (coefficient of variation [CV] = 7.9%) compared to U13 soc- test results. In addition, a number of studies reported mixed test
cer players (CV = 17.3).63 High reproducibility was also observed for results of boys and girls and were not included in our quantitative
different groups of elite youth soccer players performing the YYIE2 analysis. Furthermore, the record search was limited to studies pub-
test (CV = 3.9%)115 or the YYIE1 test (CV = 5.7%).37 For the YYIR1C lished in English and inclusion of data reported in other languages
test, mainly performed in younger children, reported reproducibil- may have altered the results of age groups with smaller sample
ity was better in children aged 8–9 years (CV = 13.0%) compared to sizes.
children between 6–7 years of age (CV = 26.0%).31 Combined, these
observations suggest that test reproducibility increases with age 6. Conclusions
but may potentially be higher in more experienced participants
already at younger age. This might be explained by the fact that, per Of the available YYI test variants, the YYIR/2, the YYIE1/2 and
definition, the YYI tests are maximal performance tests. It is thus an the YYIR1C have been used in children and adolescents. Of these,
essential requirement that children perform YYI tests at a maximal the YYIR1 and the YYIE1 have been used most frequently and over
effort, which may be evaluated through ratings of perceived exer- the entire age range from 9 to 16 years in boys, while data on girls is
tion or the use of heart rate monitors. However, if the child is not underrepresented. Our analysis provides evidence that YYIR1 and
used to all-out exercise efforts, perceived exertion might be prob- YYIE1 test reference values differ depending on age and sex. The
lematic when performing the test for the first time. It could thus presented results may be used by practitioners and researchers to
be necessary that full test familiarization is performed before the rate YYI test performance and monitor age-related development
actual test (with sufficient offset, 1̃ week). In general, it is manda- of physical fitness. With regard to varying reproducibility values,
tory that the test is supervised and documented by experienced caution is warranted when using YYI tests to rate physical fitness
personnel, at best by at least two raters. This will help to prevent especially in younger children.
documentation and procedure errors and will also limit the effect
of observer errors including failure of early or late test termina- Funding
tion. Moreover, YYI tests with shorter running distances, such as
the YYIR1C test or lower initial running speeds, such as the YYIE1 None.
should be used for younger or less fit children. Even if individual
tests are performed considering these requirements, test perfor- Acknowledgments
mance values at the lowest level (at or below the 10% level) should
be interpreted with great care. While it is possible that comparably We thank Franziska Breulmann for reviewing the extracted data
low test performance reveals low physical fitness or might even after entry into the databases.
indicate a certain delay in development or unknown health prob-
lems, individual circumstances leading to the test result should be Appendix A. Supplementary data
identified and this should include detailed questioning of the par-
ticipant. Any adverse events should be documented and physical Supplementary material related to this article can be found, in
examinations should be performed if indicated. When test results the online version, at doi:https://doi.org/10.1016/j.jsams.2019.05.
remain at a low level also after retest within a certain time period 016.
and/or show no improvement with age, it might be adequate to
involve the child into regular or individually optimized physical
activity programs. This could lead to increased test results and References 2
thus improved physical fitness to prevent future health risks. The
here presented framework might also be used by coaches to iden- 1. Hamer M, Chida Y. Walking and primary prevention: a meta-analysis of prospec-
tify children or adolescents with outstanding physical fitness (at or tive cohort studies. Br J Sports Med 2008; 42(4):238–243.

above the 90% level) and test results may be compared to the YYI
test normative data of adults at different athletic levels presented
in our previous study.33 2
Further references can be found in Multimedia Component 3.
 B. Schmitz et al. / Journal of Science and Medicine in Sport 22 (2019) 1030–1037 1037

2. Biswas A, Oh PI, Faulkner GE et al. Sedentary time and its association with risk for 32. Bendiksen M, Ahler T, Clausen H et al. The use of Yo-Yo intermittent recovery
disease incidence, mortality, and hospitalization in adults: a systematic review level 1 and Andersen testing for fitness and maximal heart rate assessments of
and meta-analysis. Ann Intern Med 2015; 162(2):123–132. 6- to 10-year-old school children. J Strength Cond Res 2013; 27(6):1583–1590.
3. O’Donovan G, Hamer M, Stamatakis E. Relationships between exercise, smok- 33. Schmitz B, Pfeifer C, Kreitz K et al. The Yo-Yo intermittent tests: a systematic
ing habit and mortality in more than 100,000 adults. Int J Cancer 2017; review and structured compendium of test results. Front Physiol 2018; 9:870.
140(8):1819–1827. 34. Castagna C, Impellizzeri F, Cecchini E et al. Effects of intermittent-endurance
4. Lee IM, Shiroma EJ, Lobelo F et al. Effect of physical inactivity on major non- fitness on match performance in young male soccer players. J Strength Cond Res
communicable diseases worldwide: an analysis of burden of disease and life 2009; 23(7):1954–1959.
expectancy. Lancet 2012; 380(9838):219–229. 35. Rebelo A, Brito J, Seabra A et al. Physical match performance of youth foot-
5. Kohl 3rd HW, Craig CL, Lambert EV et al. The pandemic of physical inactivity: ball players in relation to physical capacity. Eur J Sport Sci 2014; 14(Suppl
global action for public health. Lancet 2012; 380(9838):294–305. 1):S148–S156.
6. Global Recommendations on Physical Activity for Health, Geneva, World Health 36. Da Silva CD, Natali AJ, de Lima J Roberto Perrout et al. Yo-Yo IR2 test and margaria
Organization, 2010. test: validity, reliability and maximum heart rate in young soccer players. Revista
7. Global Status Report on Noncommunicable Diseases, Geneva, World Health Orga- Brasileira de Medicina do Esporte 2011; 17(5):344–349.
nization, 2014. 37. Wong P, Chaouachi A, Castagna C et al. Validity of the Yo-Yo intermittent
8. Blair SN, Kohl 3rd HW, Paffenbarger Jr RS et al. Physical fitness and all- endurance test in young soccer players. Eur J Sport Sci 2011; 11(5):309–315.
cause mortality. A prospective study of healthy men and women. JAMA 1989; 38. Martinez-Lagunas V, Hartmann U. Validity of the Yo-Yo Intermittent Recovery
262(17):2395–2401. Test Level 1 for direct measurement or indirect estimation of maximal oxygen
9. Kriemler S, Manser-Wenger S, Zahner L et al. Reduced cardiorespiratory fit- uptake in female soccer players. Int J Sports Physiol Perform 2014; 9(5):825–831.
ness, low physical activity and an urban environment are independently 39. Moher D, Liberati A, Tetzlaff J et al. Preferred reporting items for systematic
associated with increased cardiovascular risk in children. Diabetologia 2008; reviews and meta-analyses: the PRISMA statement. BMJ 2009; 339:b2535.
51(8):1408–1415. 40. Brink Y, Louw QA. Clinical instruments: reliability and validity critical appraisal.
10. Jefferis BJ, Parsons TJ, Sartini C et al. Objectively measured physical activity, J Eval Clin Pract 2012; 18(6):1126–1132.
sedentary behaviour and all-cause mortality in older men: does volume of activ- 41. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med
ity matter more than pattern of accumulation? Br J Sports Med 2018. http://dx. 2002; 21(11):1539–1558.
doi.org/10.1136/bjsports-2017-098733 [Epub ahead of print]. 42. Myburgh GK, Cumming SP, Silva MC et al. Maturity-associated variation in
11. Arem H, Moore SC, Patel A et al. Leisure time physical activity and mortality: functional characteristics of elite youth tennis players. Pediatr Exerc Sci 2016;
a detailed pooled analysis of the dose-response relationship. JAMA Intern Med 28(4):542–552.
2015; 175(6):959–967. 43. Moss SL, McWhannell N, Michalsik LB et al. Anthropometric and physical per-
12. Tomkinson GR, Lang JJ, Tremblay MS et al. International normative 20 m shuttle formance characteristics of top-elite, elite and non-elite youth female team
run values from 1 142 026 children and youth representing 50 countries. Br J handball players. J Sports Sci 2015; 33(17):1780–1789.
Sports Med 2017; 51(21):1545–1554. 44. Wright MD, Hurst C, Taylor JM. Contrasting effects of a mixed-methods high-
13. NCD Risk Factor Collaboration (NCD-RisC). Trends in adult body-mass index intensity interval training intervention in girl football players. J Sports Sci 2016;
in 200 countries from 1975 to 2014: a pooled analysis of 1698 population- 34(19):1808–1815.
based measurement studies with 19.2 million participants. Lancet 2016; 45. Attene G, Pizzolato F, Calcagno G et al. Sprint vs. intermittent training in young
387(10026):1377–1396. female basketball players. J Sports Med Phys Fitness 2014; 54(2):154–161.
14. Ruiz JR, Sui X, Lobelo F et al. Association between muscular strength and mor- 46. Povoas SC, Castagna C, da Costa Soares JM et al. Reliability and construct validity
tality in men: prospective cohort study. BMJ 2008; 337:a439. of Yo-Yo tests in untrained and soccer-trained schoolgirls aged 9-16. Pediatr
15. Ortega FB, Ruiz JR, Castillo MJ et al. Physical fitness in childhood and adolescence: Exerc Sci 2016; 28(2):321–330.
a powerful marker of health. Int J Obes (Lond) 2008; 32(1):1–11. 47. Skoradal MB, Purkhus E, Steinholm H et al. “FIFA 11 for Health” for Europe in the
16. Ruiz JR, Castro-Pinero J, Artero EG et al. Predictive validity of health-related Faroe Islands: effects on health markers and physical fitness in 10- to 12-year-
fitness in youth: a systematic review. Br J Sports Med 2009; 43(12):909–923. old schoolchildren. Scand J Med Sci Sports 2018; 28(Suppl 1):8–17.
17. Catley MJ, Tomkinson GR. Normative health-related fitness values for children: 48. Larsen MN, Nielsen CM, Orntoft C et al. Fitness effects of 10-month frequent low-
analysis of 85347 test results on 9-17-year-old Australians since 1985. Br J Sports volume ball game training or interval running for 8–10-year-old school children.
Med 2013; 47(2):98–108. Biomed Res Int 2017; 2017, 2719752.
18. Artero EG, Lee DC, Ruiz JR et al. A prospective study of muscular strength 49. Geertsen SS, Thomas R, Larsen MN et al. Motor skills and exercise capacity are
and all-cause mortality in men with hypertension. J Am Coll Cardiol 2011; associated with objective measures of cognitive functions and academic perfor-
57(18):1831–1837. mance in preadolescent children. PLoS One 2016; 11(8):e0161960.
19. Leger LA, Mercier D, Gadoury C et al. The multistage 20 metre shuttle run test 50. Karakoc B, Akalan C, Alemdaroglu U et al. The relationship between the yo-yo
for aerobic fitness. J Sports Sci 1988; 6(2):93–101. tests, anaerobic performance and aerobic performance in young soccer players.
20. Olds T, Tomkinson G, Leger L et al. Worldwide variation in the performance of J Hum Kinet 2012; 35:81–88.
children and adolescents: an analysis of 109 studies of the 20-m shuttle run test 51. Benounis O, Benabderrahman A, Chamari K et al. Association of short-passing
in 37 countries. J Sports Sci 2006; 24(10):1025–1038. ability with athletic performances in youth soccer players. Asian J Sports Med
21. Domone S, Mann S, Sandercock G et al. A method by which to assess the scalabil- 2013; 4(1):41–48.
ity of field-based fitness tests of cardiorespiratory fitness among schoolchildren. 52. Spencer M, Pyne D, Santisteban J et al. Fitness determinants of repeated-sprint
Sports Med 2016; 46(12):1819–1831. ability in highly trained youth football players. Int J Sports Physiol Perform 2011;
22. Castro-Pinero J, Artero EG, Espana-Romero V et al. Criterion-related validity of 6(4):497–508.
field-based fitness tests in youth: a systematic review. Br J Sports Med 2010; 53. Carvalho HM, Bidaurrazaga-Letona I, Lekue JA et al. Physical growth and changes
44(13):934–943. in intermittent endurance run performance in young male Basque soccer play-
23. Leger LA, Lambert J. A maximal multistage 20-m shuttle run test to predict VO2 ers. Res Sports Med 2014; 22(4):408–424.
max. Eur J Appl Physiol Occup Physiol 1982; 49(1):1–12. 54. Deprez D, Vaeyens R, Coutts AJ et al. Relative age effect and Yo-Yo IR1 in youth
24. Bangsbo J, Iaia FM, Krustrup P. The Yo-Yo intermittent recovery test: a useful soccer. Int J Sports Med 2012; 33(12):987–993.
tool for evaluation of physical performance in intermittent sports. Sports Med 55. Chuman K, Hoshikawa Y, Iida T et al. Relationships between Yo-Yo intermittent
2008; 38(1):37–51. recovery tests and development of aerobic and anaerobic fitness in U-13 and
25. Krustrup P, Mohr M, Amstrup T et al. The yo-yo intermittent recovery test: U-17 soccer players. Int J Sport Health Sci 2011; 9:91–97.
physiological response, reliability, and validity. Med Sci Sports Exerc 2003; 56. Abian-Vicen J, Puente C, Salinero JJ et al. A caffeinated energy drink improves
35(4):697–705. jump performance in adolescent basketball players. Amino Acids 2014;
26. Castagna C, Impellizzeri FM, Belardinelli R et al. Cardiorespiratory responses to 46(5):1333–1341.
Yo-yo Intermittent Endurance Test in nonelite youth soccer players. J Strength 57. Deprez D, Fransen J, Lenoir M et al. The Yo-Yo intermittent recovery test level 1
Cond Res 2006; 20(2):326–330. is reliable in young high-level soccer players. Biol Sport 2015; 32(1):65–70.
27. Krustrup P, Mohr M, Nybo L et al. The Yo-Yo IR2 test: physiological 58. Souhail H, Castagna C, Mohamed HY et al. Direct validity of the yo-yo intermit-
response, reliability, and application to elite soccer. Med Sci Sports Exerc 2006; tent recovery test in young team handball players. J Strength Cond Res 2010;
38(9):1666–1673. 24(2):465–470.
28. Castagna C, Impellizzeri FM, Chamari K et al. Aerobic fitness and yo-yo continu- 59. Casamichana D, Castellano J. Time-motion, heart rate, perceptual and motor
ous and intermittent tests performances in soccer players: a correlation study. behaviour demands in small-sides soccer games: effects of pitch size. J Sports Sci
J Strength Cond Res 2006; 20(2):320–325. 2010; 28(14):1615–1623.
29. Bradley PS, Bendiksen M, Dellal A et al. The application of the Yo-Yo intermittent 60. Doncaster G, Iga J, Unnithan V. Assessing differences in cardiorespiratory fitness
endurance level 2 test to elite female soccer populations. Scand J Med Sci Sports with respect to maturity status in highly trained youth soccer players. Pediatr
2014; 24(1):43–54. Exerc Sci 2018; 30(2):216–228.
30. Reilly T, Bangsbo J, Franks A. Anthropometric and physiological predispositions
for elite soccer. J Sports Sci 2000; 18(9):669–683.
31. Ahler T, Bendiksen M, Krustrup P et al. Aerobic fitness testing in 6- to 9-year-old
children: reliability and validity of a modified Yo-Yo IR1 test and the Andersen
test. Eur J Appl Physiol 2012; 112(3):871–876.