See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/6398171

Cross-validation of the 20- versus 30-s

Wingate anaerobic test

ARTICLE in ARBEITSPHYSIOLOGIE · SEPTEMBER 2007

Impact Factor: 2.19 · DOI: 10.1007/s00421-007-0454-3 · Source: PubMed

CITATIONS READS

20 142

4 AUTHORS, INCLUDING:

Matt Laurent Clay Robinson

Bowling Green State University Affiliation of Crop, Soil, and Environmental …
37 PUBLICATIONS 196 CITATIONS 29 PUBLICATIONS 167 CITATIONS

SEE PROFILE SEE PROFILE

James M Green
University of North Alabama
108 PUBLICATIONS 973 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate, Available from: Matt Laurent
letting you access and read them immediately. Retrieved on: 11 January 2016
Eur J Appl Physiol (2007) 100:645–651
DOI 10.1007/s00421-007-0454-3
O R I G I N A L A R T I CL E
Cross-validation of the 20- versus 30-s Wingate anaerobic test
C. Matthew Laurent Jr. · Michael C. Meyers ·
Clay A. Robinson · J. Matt Green
Accepted: 19 March 2007 / Published online: 12 April 2007
© Springer-Verlag 2007
Abstract The 30-s Wingate anaerobic test (30-WAT) is
the most widely accepted protocol for measuring anaerobic
response, despite documented physical side eVects. Abbre-
viation of the 30-WAT without loss of data could enhance
subject compliance while maintaining test applicability.
The intent of this study was to quantify the validity of the
20-s Wingate anaerobic test (20-WAT) versus the tradi-
tional 30-WAT. Fifty males (mean § SEM; age = 20.5 §
0.3 years; Ht = 1.6 § 0.01 m; Wt = 75.5 § 2.6 kg) were
randomly selected to either a validation (N = 35) or cross-
validation group (N = 15) and completed a 20-WAT and
30-WAT in double blind, random order on separate days to
determine peak power (PP; W kg¡1), mean power (MP;
W kg¡1), and fatigue index (FI; %). Utilizing power
outputs (relative to body mass) recorded during each sec-
ond of both protocols, a non-linear regression equation
(Y20WAT+10 = 31.4697 e¡0.5[ln(Xsecond/1174.3961)/2.63692];
r2 = 0.97; SEE = 0.56 W kg¡1) successfully predicted
(error »10%) the Wnal 10 s of power outputs in the cross-
validation population. There were no signiWcant diVer-
ences between MP and FI between the 20-WAT that
included the predicted 10 s of power outputs (20-WAT+10)
C. Matthew Laurent Jr. (&) · J. Matt Green
Department of Kinesiology, The University of Alabama,
Box 83012, Moore Hall, Tuscaloosa, AL, 35487-0312, USA
e-mail: cmlaurentjr@bama.ua.edu
M. C. Meyers
Human Performance Research Laboratory,
Department of Sports and Exercise Sciences,
West Texas A&M University, Canyon, TX 79016, USA
C. A. Robinson
Department of Agriculture, West Texas A&M University,
Canyon, TX 79016, USA
and the 30-WAT. When derived data were subjected to
Bland–Altman analyses, the majority of plots (93%) fell
within the limits of agreement (§2SD). Therefore, when
compared to the 30-WAT, the 20-WAT may be considered
a valid alternative when used with the predictive non-linear
regression equation to derive the Wnal power output values.
Keywords Leg power · Work capacity · Cycle ergometry ·
Sprint test
Introduction
The most commonly employed protocol for the measure-
ment of anaerobic response is the 30-s Wingate anaerobic
test (30-WAT). Developed almost 30 years ago, the 30-
WAT involves a maximal exertion bout on a cycle ergome-
ter to evaluate peak power (PP), mean power (MP), and
fatigue index (FI; Bar-Or 1987; Bar-Or et al. 1977). When
performing a 30-WAT, a subject typically exhibits a sharp
rise in power output, reaching peak power within the Wrst
few seconds. Typically, subjects are unable to maintain this
output, leading to an exponential decline in power through-
out the remaining duration of the test (Bar-Or et al. 1977;
Marquardt et al. 1993). During this prolonged period of
maximal eVort, the accumulation of [H+] and lactate as
byproducts of anaerobic glycolysis results in a drop in
blood pH (lactic acidosis). The increased acidity impairs
enzyme activity involved in energy metabolism and
reduces maximal muscle Wbre recruitment (Allen et al.
1992; Davis 1985). In addition, the acute increase of blood
glucose as a substrate for glycolysis during maximal exer-
cise can result in hypoglycemia (Vincent et al. 2004).
Together these responses may result in unwelcome physical
side eVects, including fatigue, headache, dizziness, and
123
646
nausea (Jacobs et al. 1982). Subject awareness of these side
eVects during anaerobic testing may result in submaximal
eVort, high attrition rates, or unsuccessful subject comple-
tion of subsequent sport performance testing (Jacobs et al.
1982; Marquardt et al. 1993).
In 1987, Vandewalle et al. reported high correlation
coeYcients between data collected at the 20th and 30th sec-
onds of a WAT. Consequently, they proposed to curtail the
duration of the WAT in order to diminish the stress of the
test and the involvement of the aerobic metabolism. How-
ever, this previous study did not compare the 20-WAT and
30-WAT but the data collected at the 20th and 30th second
of the same test. Hence this previous study failed to prove
that results of a 20-WAT would have been the same as
results collected during the Wrst 20 s of a 30-WAT because
of a possible submaximal eVort strategy during a stressful
30-WAT. More recently, Marquardt et al. (1993) reported
high intraclass correlations (r > 0.95) between results of a
20 and 30 Wingate tests in 14 subjects and concluded “that
a 20-s Wingate test may be a valid and less strenuous alter-
native to the 30-s Wingate test.”
Since its Wrst publication, the 30-WAT has been used in
hundreds of studies published in the international literature
and it would be interesting to be able to compare the results
of a future 20-s anaerobic test with all these experimental
data. Moreover, it is likely that many athletes have per-
formed several traditional 30-WATs and, subsequently,
could be reluctant to perform new tests whose results can-
not be compared with previous scores. The prediction of
30-WAT indices from 20-WAT performances is not indi-
vidual but statistical in the study by Vandewalle et al.
(1987a). In that study, the statistical prediction of the
amount of work performed at the 30th seconds (an equiva-
lent of MP) from 20-s data was probably accurate
(r = 0.989), however, the regression coeYcient between
pedal rate at 20th and 30th was lower (r = 0.882). In the
present study we validate a 20-WAT by comparing the data
collected during 20 and 30 s all-out exercises in 50 sub-
jects. Moreover, we propose a method which enables an
individual and accurate prediction of 30-WAT indices from
the data of a 20-WAT (20-WAT+10).
Methods
Subjects
The participants for this study included 50 male, college-
aged students (mean § SEM; age = 20.5 § 0.3 years;
Ht = 175.8 § 1.2 cm; Wt = 75.5 § 2.6 kg) with no known
cardiovascular/pulmonary disease, metabolic disorders, or
medical contraindications to exercise as determined by self-
response medical history form and interview. Subjects were
123
Eur J Appl Physiol (2007) 100:645–651
randomly selected students who volunteered from univer-
sity level physical education classes. Subjects were encour-
aged not to participate in any strenuous exercise during the
24 h prior to testing, and initially reported to the laboratory
4 h postabsorptive. Subjects were fully informed of the
nature of the study and provided written, informed consent
in accordance with the accepted guidelines of the Institu-
tional Review Board of the university and the American
College of Sports Medicine (1997).
Procedures
All subjects initially reported to the laboratory whereupon
each individual had their height (m) and body mass (kg)
determined using calibrated physician scales. Anaerobic
response was quantiWed utilizing both a 20-WAT and a 30-
WAT, performed in a double blind, randomly determined
order to measure peak power (W kg¡1), mean power
(W kg¡1), and fatigue index (%). The maximum power
achieved during the Wrst 5 s of the test is deWned as PP,
while MP is deWned as the average power achieved
throughout the trial, and the FI reXects the percent power
decline during the trial (Bar-Or 1987; Bar-Or et al. 1977).
Both WAT protocols were conducted while pedaling a
calibrated Monark model 824E cycle ergometer (Monark
AB, Varberg, Sweden) with integrated laser-based sensor
and computer software (Sports Medicine Industries, Inc.,
St. Cloud, MN). Prior to both tests, each subject began a
warm-up phase consisting of alternating three 30-s intervals
of active rest (pedaling against no resistance at 60 rpm)
with three 30-s intervals pedaling against increasing resis-
tance of 25, 50, and 75% of test resistance (Vanderford
et al. 2004). Prior research has indicated that this form of
pre-test loading elicits optimal power production during
supramaximal exercise (Burnley et al. 2005). Test resis-
tance was calculated by multiplying the subject’s body
mass (kg) by 0.075. Following completion of the warm-up,
the subject continued to pedal at 60 rpm with no resistance
for another 2 min until initiation of the WAT. Following a
10-s countdown, the resistance was immediately added and
the subject was verbally encouraged to pedal as fast as pos-
sible for 20 or 30 s. Relative power outputs (W kg¡1) were
measured during each second of the two trials, and PP, MP,
and FI were calculated according to accepted procedures
(Bar-Or 1987; Bar-Or et al. 1977).
Within each subject, the 20-WAT and 30-WAT were
completed at the same time of day, no less than 48 h and no
more than 7 days apart in order to ensure optimal recovery
and minimize any confounding anthropometric changes,
respectively. In accordance with previously published
research regarding test–retest reliability of the WAT (Bar-Or
1987; Granier et al. 1995; Kaczkowski et al. 1982;
Vandewalle et al. 1987a), a PP increase ¸8% was considered
Eur J Appl Physiol (2007) 100:645–651
to be a distorted response, presumably due to a learning
eVect. In these instances, the initial protocol was repeated
48 h later and used for comparison. In 13 instances, the
diVerence between the two protocols still exceeded the
stated criteria and, consequently, these subjects were omit-
ted from the study. To assess the presence of negative side
eVects following testing, a basic approach was taken. Sub-
jects were asked to provide a yes/no answer with respect to
presence of nausea, light-headedness, leg fatigue, and/or
any other physical side eVects. This approach was advanta-
geous in that it was not diYcult to administer and permitted
the evaluation of side eVects immediately proximal to the
test. This was important as side eVects may often be acute
and transient and in such case a written survey completed
by subjects after recovering from testing would provide less
valid information.
Statistical analyses
Data were subjected to multivariate analyses of variance
(MANOVAs), where appropriate, utilizing the Statistical
Package for Social Sciences (SPSS, Inc., Chicago, IL) to
determine a signiWcant main eVect between protocols.
When univariate post hoc procedures (AVOVA) indicated
signiWcant diVerences between relative MP outputs and FI,
the sample of 50 subjects were randomly assigned to either
a validation group (N = 35) or cross-validation group
(N = 15). Both linear and non-linear regression analyses
were employed to develop an equation in order to deter-
mine the best model to predict the Wnal 10 s of relative
power outputs. These values were derived from power out-
puts obtained during seconds 11–20 of the 20-WAT rela-
tive to those obtained during seconds 21–30 of the 30-WAT
from the validation group. The non-linear formula
employed to Wt the data was a three parameter, peak, log-
normal equation of the form Y = a e¡0.5[(ln (X/X0)/b)2],
where a, b, and X0 are the parameters Wtted, e¡0.5 is an
exponential function, ln is the command to perform a nor-
mal logarithmic procedure, and X is the second for which
the relative power output is being determined. Following
regression analyses, the predicted Wnal 10 s were combined
with the observed 20-WAT data (20-WAT+10) recorded in
the cross-validation group to demonstrate the validity and
applicability of the derived equation to a target population.
ANOVAs were performed to identify signiWcant diVer-
Table 1 Mean anaerobic power
Variable
and capacity for 20-WAT and
30-WAT (N = 50)
Peak power (W kg¡1)
Mean power (W kg¡1)
Fatigue index (%)
* SigniWcant at P < 0.05 level
647
ences among the PP, MP, and FI of the 20-WAT+10 and
the 30-WAT relative power output. To further assess valid-
ity and reliability, the Bland–Altman method of comparison
was employed to determine the limits of agreement
(mean § 2SD) between the two protocols (Altman and
Bland 1983; Bland and Altman 1986) within the cross-vali-
dation group. Statistical signiWcance was determined
a priori at the P = 0.05 level. All data are presented as
mean § SEM unless otherwise noted.
Results
Peak power, mean power, and fatigue index
The mean relative power outputs recorded during the
20-WAT as well as the 30-WAT along with their respective
P values and percent diVerences are presented in Table 1.
As expected, a signiWcant MANOVA was found for the
Wilks’ Lambda rank variable (F3,80 = 0.723; P < 0.001;
n ¡
= 0.998) between protocols. Subsequent ANOVAs
indicated no signiWcant diVerences between relative PP out-
puts recorded during the 30-WAT and the 20-WAT; how-
ever, signiWcant diVerences were found with regard to
relative MP and FI. Mean relative power outputs were sim-
ilar throughout the Wrst 20 s of both trials, however, a sub-
stantial power decline was observed during the Wnal 10 s of
the 30-WAT (Fig. 1).
Linear versus non-linear regression prediction model
Results from linear and non-linear regression analyses
revealed that the non-linear model yield similar coeYcients
of determination between models. There was an additional
two percent of variation observed was explained when
using a non-linear versus a linear prediction model
(r2 = 0.97 vs. r2 = 0.95, respectively). Additionally, a
noticeable discrepancy was revealed in the standard error of
the estimate between models, with the non-linear model
reporting an error of »10% (SEE = 0.56 W kg¡1) when
compared to an error of »15% (SEE = 0.75 W kg¡1) of the
linear model observed within the measured population.
Consequently, the non-linear regression model was
employed to validate and cross-validate the 20-WAT ver-
sus 30-WAT.
30-WAT 20-WAT P DiVerence
Mean § SEM Mean § SEM (%)
6.8 § 0.2 6.7 § 0.2 0.54 1.5
4.5 § 0.1 5.0 § 0.2 0.01* 10.0
58.3 § 1.7 46.1 § 1.7 <0.01* 20.9
123
648 Eur J Appl Physiol (2007) 100:645–651

Fig. 1 Average relative power (N =50)

outputs of all subjects (N = 50) 7.0
between 20-WAT and 30-WAT
6.0

5.0

4.0

W·kg-1
3.0

2.0

1.0

20-WAT 30-WAT
0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Second

Predicted versus observed MP and FI
The relative power outputs of the 20-WAT+10 compared to
the observed relative power outputs of the 30-WAT for all
35 subjects in the validation group (Fig. 2) was modeled
and curve Wtted for all individuals in a three parameter,
peak, log-normal equation (Y20WAT+10 = 31.4697 e¡0.5 [ln
(Xsecond/1174.3961)/2.63692]). The model was signiWcant
and the slope and intercept in the equation were signiW-
cantly diVerent from zero (P < 0.001).
The mean relative power outputs recorded during the 20-
WAT+10 as well as the 30-WAT, their respective P values,
and percent diVerences are presented in Table 2. MANOVA
indicated no signiWcant main eVect for the Wilks’ Lambda
rank variable (F3,80 = 0.182; P = 0.909; n ¡
= 0.082)
between a 20-WAT+10 and 30-WAT.

Fig. 2 Average relative power (N = 35)

outputs of the validation group 7.0
(N = 35) between 20-WAT+10
and 30-WAT 6.5

6.0

5.5

5.0
W·kg-1

4.5

4.0

3.5

3.0

2.5
30-WAT 20-WAT+10
2.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Seconds

Table 2 Mean anaerobic power and capacity for the 20-WAT+10 and the 30-WAT in the validation group (N = 35)
Variable 30-WAT Mean § SEM 20-WAT+10 Mean § SEM P DiVerence (%)

Peak power (W kg¡1) 6.6 § 0.2 6.6 § 0.2 0.75 0.0

Mean power (W kg¡1) 4.3 § 0.1 4.4 § 0.1 0.71 2.3
Fatigue index (%) 59.4 § 1.7 57.5 § 1.2 0.20 3.3
No values were signiWcantly diVerent at the P < 0.05 level

123
Eur J Appl Physiol (2007) 100:645–651 649

Table 3 Mean anaerobic power and capacity for the 20-WAT+10 and the 30-WAT in the cross-validation group (N = 15)
Variable 30-WAT Mean § SEM 20-WAT+10 Mean § SEM P DiVerence (%)

Peak power (W kg¡1) 6.9 § 0.2 7.1 § 0.3 0.06 2.9

¡1
Mean power (W kg ) 4.6 § 0.2 4.6 § 0.2 0.58 0.0
Fatigue index (%) 55.4 § 2.4 58.2 § 1.7 0.08 4.9
No values were signiWcantly diVerent at the P < 0.05 level

The mean observed and predicted relative power outputs FI observed during the 30-WAT trials, with 93% of all
for the 20-WAT+10 as well as the 30-WAT in the cross- observations falling within the §2SD limits of agreement.
validation group, their respective P values, and percent
diVerences are presented in Table 3. Furthermore, ANO-
VAs indicated no signiWcant diVerences between relative Discussion
PP outputs, relative MP outputs, or FI between the two pro-
tocols in the cross-validation group (Fig. 3). The 30-WAT has been shown to be highly reliable and
Figures 4, 5 illustrate the limits of agreement (§2SD) as applicable in a wide variety of settings, having test–retest
determined by the Bland–Altman method of comparison. reliability of 0.90–0.97 (Bar-Or 1987; Bar-Or et al. 1977;
Predicted values of MP and FI demonstrated acceptable Kaczkowski et al. 1982; Vandewalle et al. 1987b). The
agreement when compared to the actual values of MP and 30-WAT has been used to evaluate anaerobic performance

Fig. 3 Average relative power (N = 15)

outputs of the cross-validation 7.0
group (N = 15) between
20-WAT+10 and 30-WAT 6.5

6.0

5.5

5.0
W·kg-1

4.5

4.0

3.5

3.0

2.5
30-WAT 20-WAT+10
2.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Seconds

Fig. 4 Bland–Altman plots of (N = 15)

relative mean power outputs 2.0
Observed Difference Between Protocols

illustrating the upper and lower 1.8

1.6
levels of agreement (dashed 1.4
lines) between 30-WAT and 1.2
20-WAT+10 1.0
0.8
0.6
0.4
0.2
0.0
-0.2
-0.4
-0.6
-0.8
-1.0
-1.2
-1.4
-1.6
-1.8
-2.0
3.0 3.5 4.0 4.5 5.0 5.5 6.0
Mean Power (W·kg-1)

123
650 Eur J Appl Physiol (2007) 100:645–651

Fig. 5 Bland–Altman plots of (N = 15)

fatigue index illustrating the 24.0
upper and lower levels of 22.0

Observed Difference Between Protocols

agreement (dashed lines) 20.0
between 30-WAT and 18.0
20-WAT+10 16.0
14.0
12.0
10.0
8.0
6.0
4.0
2.0
0.0
-2.0
-4.0
-6.0
-8.0
-10.0
-12.0
45.0 50.0 55.0 60.0 65.0 70.0
Fatigue Index (%)

among children, adolescents, and adults with activity levels
ranging from sedentary to athletic (Bar-Or 1987; Groussard
et al. 2003; Mastrangelo et al. 2004; Van Someren and
Palmer 2003; Vincent et al. 2004). However, the limitations
and applicability of the WAT to a variety of populations,
including high-level athletes, has been well documented in
other studies (Beneke et al. 2002; Medbø and Tabata,
1993). Additionally, detrimental physical side eVects have
been documented, and may constitute a threat to optimal
subject compliance (Maud and Schultz 1989; Ulmer 1996).
Abbreviation of the 30-WAT without loss of data could
potentially ensure subject compliance while maintaining
test applicability (Marquardt et al. 1993; Smith and Hill
1991). Therefore, the purpose of this study was to quantify
the validity of and, subsequently, cross-validate the 20-
WAT versus 30-WAT.
As expected, results from this study indicate an abbrevi-
ated version of the 30-WAT produced signiWcantly diVer-
ent relative MP outputs and FI values. In the study by
Marquardt et al. (1993), average power output (PO) from
seconds 15–20 during 20-WAT was apparently correlated
with PO from seconds 15–20 during 30 WAT instead of PO
from seconds 25–30, which did not enable the computation
of FI during 30-WAT. In a study by Vandewalle et al.
(1987a), the correlation coeYcient between velocity at 20
and 30 s was lower than the correlation coeYcient between
MP (0.882 vs. 0.989) although these velocities were mea-
sured during the same exercise. In the present study, the
fatigue indices (FI) corresponding to 20-WAT+10 and 30-
WAT were calculated for exercises performed on separate
days. Nevertheless, the diVerences between FI were only
3.3 and 4.9% (Tables 2, 3 for the validation and cross vali-
dation studies, respectively) although the test–retest corre-
lation coeYcients in the literature are low for the fatigue
index. Consequently, it is possible to predict the Wnal 10 s
of relative power outputs and attain similar relative MP out-
puts and FI values observed during a traditional 30-WAT
with the use of the derived non-linear regression equation.
Comparison of the 20-WAT versus the 30-WAT
As seen in Figs. 1, 2, 3, each subject achieved maximal
power output within the initial 5 s of the anaerobic bouts.
This was followed by a gradual decline in relative power, as
expected, throughout the remainder of the test, resulting in a
minimal power output during the Wnal 5 s, as observed in
earlier studies (Ansley et al. 2004; Bar-Or 1987; Calbet et al.
2003; Gastin 2001; Granier et al. 1995; Murphy et al. 1986;
Smith and Hill 1991; Vandewalle et al. 1987a). The lack of
signiWcant diVerence in PP reported in this study indicates
that similar eVort was demonstrated over both trials.
Despite similar relative power outputs observed during
the Wrst 20 s of both trials, signiWcant diVerences were
observed between the overall relative MP outputs and FI
values between the two protocols, as would be expected.
While lactate and acidity were not directly assessed in the
current study, it is plausible that these by-products poten-
tially associated with fatigue may have diVered between tri-
als and consequently aVected performance. However, it is
strongly emphasized that without measures, any conclu-
sions as such would be speculative at best.
Validity of the 20-WAT+10 versus the 30-WAT
Following the derivation of the non-linear regression equa-
tion, predicted power output and FI values resulted in simi-
lar relative MP outputs and FI values between the 20-
WAT+10 and the 30-WAT in the validation and cross vali-
dation groups (Figs. 1, 2). Both PP and MP values from the
20-WAT+10 and 30-WAT met or exceeded the test–retest

123
Eur J Appl Physiol (2007) 100:645–651
reliability values previously reported with respect to power
outputs observed between two separate 30-WATs (Bar-Or
1987; Kaczkowski et al. 1982). These Wndings suggest that
administering a 20-WAT is still advantageous if (a) the
derived regression equation is employed to determine rela-
tive power outputs of the Wnal 10 s of a WAT to maintain
the validity typically associated with the original protocol
or (b) by acknowledging the percent diVerences that exist
between the protocols reported in this study.
Physical response
Detrimental physical responses and subsequent subject
apprehension have been reported to occur both during and
after the 30-WAT, including nausea, dizziness, headaches,
and vomiting resulting in less than optimal compliance with
the 30-WAT (Jacobs et al. 1982; Maud and Schultz 1989).
Following the 30-WAT in this study physical discomfort was
observed as three subjects vomited, and approximately 25%
subjects reported nausea, light-headedness, headaches, and/
or leg fatigue, despite a 4-min cool-down and provision of a
high-carbohydrate drink upon request. While these eVects
can only be reported subjectively, the only detrimental side
eVect reported following the 20-WAT was leg fatigue. The
minimal side eVects (aside from leg exhaustion) seem to
indicate a reduced level of discomfort among subjects
following the 20-WAT, which could increase subject compli-
ance as well as the test–retest reliability of the protocol.
Conclusion
The major Wnding from this study is that a 20-WAT can be
considered a valid alternative to 30-WAT when used con-
comitantly with the prediction of the Wnal 10 s utilizing the
regression equation derived in this study. The use of this
abbreviated 20-WAT+10 protocol may reduce subject dis-
comfort both during and following the test, thereby maximiz-
ing subject compliance and enhancing the applicability and
repeatability of the WAT across a variety of populations.
References
Allen DG, Westerblad JA, Lee JA, Lannergren J (1992) Role of exci-
tation–contraction coupling in muscle fatigue. Sports Med
13:116–126
Altman DG, Bland JM (1983) Measurement in medicine: the analysis
of method comparison studies. Statistician 32:307–317
American College of Sports Medicine (1997) Policy statement regard-
ing the use of human subjects and informed consent. Med Sci
Sports Exerc 29:5
Ansley L, Robson PJ, Gibson A, Noakes TD (2004) Anticipatory pac-
ing strategies during supramaximal exercise lasting longer than
30 s. Med Sci Sports Exerc 36:309–314
651
Bar-Or O (1987) The Wingate anaerobic test: an update on methodol-
ogy, reliability and validity. Sports Med 4:381–394
Bar-Or O, Dotan R, Inbar O (1977) A 30 s all-out ergometric test: its
reliability and validity for anaerobic capacity. Isr J Med Sci
13:126
Beneke R, Pollmann C, Bleif I, Leithauser RM, Hutler M (2002) How
anaerobic is the Wingate anaerobic test for humans? Eur J Appl
Physiol 87:388–392
Bland JM, Altman DG (1986) Statistical methods for assessing agree-
ment between two methods of clinical measurements. Lancet
1:307–310
Burnley M, Doust JH, Jones AM (2005) EVects of prior warm-up re-
gime on severe-intensity cycling performance. Med Sci Sports
Exerc 37:838–845
Calbet JA, De Paz JA, Garatachea N, Cabeza de Vaca S, Chavarren J
(2003) Anaerobic energy provision does not limit Wingate exer-
cise performance in endurance-trained cyclists. J Appl Physiol
94:668–676
Davis JA (1985) Anaerobic threshold: review of the concept and direc-
tions for future research. Med Sci Sports Exerc 17:6–21
Gastin PB (2001) Energy system interaction and relative contribution
during maximal exercise. Sports Med 31:725–741
Granier P, Mercier B, Mercier J, Anselme F, Prefaut C (1995) Aerobic
and anaerobic contribution to Wingate test performance in sprint
and middle-distance runners. Eur J Appl Physiol 70:58–65
Groussard C, Machefer G, Rannou F (2003) Physical Wtness and plas-
ma non-enzymatic antioxidant status at rest and after a Wingate
test. Can J Appl Physiol 28:79–92
Jacobs I, Bar-Or O, Karlsson J, Dotan R, Tesch P, Kaiser P, Inbar O
(1982) Changes in muscle metabolites in females with 30-s
exhaustive exercise. Med Sci Sports Exerc 14:457–460
Kaczkowksi W, Montgomery DL, Taylor AW, Klissouras V (1982) The
relationship between muscle Wber composition and maximal anaer-
obic power and capacity. J Sports Med Phys Fitness 22:407–413
Marquardt JA, Bacharach DA, Kelly JM (1993) Comparison of power
outputs generated during 20 and 30 s Wingate tests. Res Q Exerc
Sport 64:A33–A34
Mastrangelo MA, Chaloupka EC, Kang J, Lacke CJ, Angelucci JA,
Martz WP, Biren GB (2004) Predicting anaerobic capacity in 11–
13 year-old boys. J Strength Cond Res 18:72–76
Maud PJ, Schultz BB (1989) Norms for the Wingate anaerobic test with
comparison to another similar test. Res Q Exerc Sport 60:144–151
Medbø JI, Tabata I (1993) Anaerobic energy release in working muscle
during 30 s to 3 min of exhausting bicycling. J Appl Physiol
75:1654–1660
Murphy MM, Patton JF, Frederick FA (1986) Comparative anaerobic
power of men and women. Aviat Space Environ Med 57:636–641
Smith JC, Hill DW (1991) Contribution of energy systems during a
Wingate power test. Br J Sports Med 25:196–199
Ulmer HV (1996) Concept of an extracellular regulation of muscular
metabolic rate during heavy exercise in humans by psychophysi-
ological feedback. Experientia 52:416–420
Van Someren KA, Palmer GS (2003) Prediction of 200-m sprint kaya-
king performance. Can J Appl Physiol 28:505–517
Vanderford ML, Meyers MC, Skelly WA, Stewart CC, Hamilton KL
(2004) Physiological and sport-speciWc skill response of Olympic
youth soccer athletes. J Strength Cond Res 18:334–342
Vandewalle H, Heller J, Pérès G, Raveneau S, Monod H (1987a) Etude
comparative entre le Wingate test et un test force-vitesse sur ego-
cycle. Sci Sports 2:279–284
Vandewalle H, Pérès G, Monod H (1987b) Standard anaerobic exer-
cise tests. Sports Med 4:268–289
Vincent S, Berthon P, Zouhal H, Moussa E, Catheline M, Betue-Ferrer
D, Gratas-Delamarche A (2004) Plasma glucose, insulin and cat-
echolamine responses to a Wingate test in physically active wom-
en and men. Eur J Appl Physiol 91:15–21
123