Reliability of knee extension and flexion measurements using the Con-Trex isokinetic dynamometer.

Abstract

The aim of this study was to evaluate the reliability of isokinetic and isometric assessments of the knee extensor and the flexor muscle function using the Con-Trex isokinetic dynamometer. Thirty healthy subjects (15 males, 15 females) were tested and retested 7 days later for maximal strength (isokinetic peak torque, work, power and angle of peak torque as well as isometric maximal voluntary contraction torque and rate of torque development) and fatigue (per cent loss and linear slope of torque and work across a series of 20 contractions). For both the knee extensor and the flexor muscle groups, all strength data - except angle of peak torque - demonstrated moderate-to-high reliability, with intraclass correlation coefficients (ICC) higher than 0.86. The highest reliability was observed for concentric peak torque of the knee extensor muscles (ICC = 0.99). Test-retest reliability of fatigue variables was moderate for the knee extensor (ICC range 0.84-0.89) and insufficient-to-moderate for the knee flexor muscles (ICC range 0.78-0.81). The more reliable index of muscle fatigue was the linear slope of the decline in work output. These findings establish the reliability of isokinetic and isometric measurements using the Con-Trex machine.

Full text

Page 1

Reliability of knee extension and flexion measurements
using the Con-Trex isokinetic dynamometer
Nicola A. Maffiuletti1, Mario Bizzini1, Kevin Desbrosses2, Nicolas Babault3and Urs Munzinger1
1Neuromuscular Research Laboratory, Schulthess Clinic, Zurich, Switzerland,2Occupational Physiology Laboratory, National Institute for Research and Security,
Nancy, France, and3Performance Expertise Centre, Faculty of Sport Sciences, University of Burgundy, Dijon, France
Correspondence
Nicola A. Maffiuletti, Schulthess Clinic, Lengghalde
2, 8008 Zurich, Switzerland
E-mail: nicola.maffiuletti@kws.ch
Accepted for publication
Received 13 March 2007;
accepted 25 May 2007
Key words
fatigue; hamstring; quadriceps; torque; work
Summary
The aim of this study was to evaluate the reliability of isokinetic and isometric
assessments of the knee extensor and the flexor muscle function using the Con-Trex
isokinetic dynamometer. Thirty healthy subjects (15 males, 15 females) were tested
and retested 7 days later for maximal strength (isokinetic peak torque, work,
power and angle of peak torque as well as isometric maximal voluntary contraction
torque and rate of torque development) and fatigue (per cent loss and linear slope of
torque and work across a series of 20 contractions). For both the knee extensor and
the flexor muscle groups, all strength data – except angle of peak torque –
demonstrated moderate-to-high reliability, with intraclass correlation coefficients
(ICC) higher than 0Æ86. The highest reliability was observed for concentric peak
torque of the knee extensor muscles (ICC = 0Æ99). Test–retest reliability of fatigue
variables was moderate for the knee extensor (ICC range 0Æ84–0Æ89) and
insufficient-to-moderate for the knee flexor muscles (ICC range 0Æ78–0Æ81). The
more reliable index of muscle fatigue was the linear slope of the decline in work
output. These findings establish the reliability of isokinetic and isometric
measurements using the Con-Trex machine.
Introduction
In both rehabilitation and sports medicine, accurate measure-
ments of muscle function are required to assess the impact of
therapeutic interventions or the effects of physical training. To
this aim, isokinetic dynamometry has been introduced in the
late 1960s and for more than two decades it has been the
standard research tool to investigate muscle function of single
muscle groups, more particularly the thigh muscles. Isokinetic
muscle strength is typically measured as peak torque, average
work and power (Perrin, 1993; Brown, 2000; Wrigley &
Strauss, 2000). Isometric muscle strength can also be measured
with these machines that includes the assessment of maximal
voluntary contraction (MVC) torque and rate of torque
development (RTD). Muscle fatigue (or endurance) – i.e. the
decline in torque⁄work output across a series of contractions – is
another feature that can be assessed using isokinetic techniques,
usually at a relatively fast concentric velocity over 20–50
consecutive maximal-effort repetitions (Thorstensson & Karls-
son, 1976; Kannus et al., 1992; Kannus, 1994; Emery et al.,
1999; Pincivero et al., 2001).
Trustworthy isokinetic and isometric results require reliable
measurement techniques. Reliability refers to the consistency of a
test and can be expressed asintra and intersession. The measuring
device itself, the procedure for conducting measurements, and
the steadiness of the subject being measured are all helpful to
determine the reliability of the method. Different studies have
demonstrated acceptable intra and intersession reliability for
various kinds of isokinetic machines, such as the Biodex (Brown
et al., 1993; Lund et al., 2005), Cybex (Bandy & McLaughlin,
1993; Li et al., 1996), Kin Com (Tredinnick & Duncan, 1988)
and Merac (Capranica et al., 1998). Other isokinetic dynamom-
eters have been used in scientific studies (Lido, Orthotron,
Technogym, etc.); however, the majority of these dynamometer
manufacturers have left the market in the recent years.
The Con-Trex MJ isokinetic machine has been recently
introduced to the rehabilitation-sports training community.
Several scientific studies have used this device to assess static
(isometric) and dynamic (eccentric and concentric) function of
both the knee extensor and the flexor muscles (Cotte & Ferret,
2003; Bardis et al., 2004; Harrison et al., 2004; Mackey et al.,
2004; Koller et al., 2006; Kilgallon et al., 2007). However, there
is a lack of information on the reliability of the Con-Trex
machine for the assessment of isometric and eccentric variables
– including isokinetic angle of peak torque – as well as for
muscle fatigue assessment.
Clin Physiol Funct Imaging (2007) 27, pp346–353doi: 10.1111/j.1475-097X.2007.00758.x
? 2007 The Authors
Journal compilation ? 2007 Blackwell Publishing Ltd •Clinical Physiology and Functional Imaging 27, 6, 346–353
346

Page 2

The aim of this study was to investigate test–retest reliability
(both within and between sessions) of the knee extensor and the
flexor muscle function (strength and fatigue) assessment using a
new isokinetic device, in a group of healthy individuals from
both sexes.
Methods
Subjects and experimental procedures
According to the recommendations of Walter et al. (1998) on
sample size for reliability studies, 15 healthy men and 15
healthy women of age ranging from 23 to 42 (mean age ± SD:
30 ± 5 years, height: 175 ± 8 cm, mass: 70 ± 13 kg) were
included in this study. All 30 were recreational athletes without
known cardiovascular and orthopaedic problems. They were
recruited from the Clinic staff. The study protocol was approved
by the local ethical committee and written consent forms were
signed prior to participation. The study was conducted
according to the Declaration of Helsinki. Subjects were
instructed to maintain their regular training regimens through-
out the experimental period and not to take part in any vigorous
physical activity for 2 days prior to their test date.
Testing
The subjects were tested during two identical sessions held
7 days apart, at the same time of day. All measurements were
recorded by the same experimenter (NAM) to avoid intertester
variability. The thigh muscles of the dominant limb (the limb
used to kick a ball) were tested by using a commercially
available dynamometer (Con-Trex MJ; CMV AG, Du ¨bendorf,
Switzerland), which allows instantaneous isokinetic and iso-
metric torque recording. Subjects were comfortably seated on
the dynamometer chair, with the hip joint at about 85?
(0? = full extension). The distal shin pad of the dynamometer
was attached 2–3 cm proximal to the lateral malleoulus by using
a strap. To minimize extraneous body movements during thigh
muscle contractions, straps were applied across the chest, pelvis
and mid-thigh. The alignment between the dynamometer
rotational axis and the knee joint rotation axis (lateral femoral
epicondyle) was checked at the beginning of each trial. Gravity
effect torque was recorded on each subject throughout the range
of motion and this was used to correct torque measurements
during all tests. The participants were given standardized
(verbal) encouragement by the investigator (see below), and
were asked to position their arms across the chest with each
hand clasping the opposite shoulder during the testing
procedure. On-line visual feedback of the instantaneous dyna-
mometer torque was provided to the subjects on a computer
screen.
The overview of the experimental protocol is depicted in
Fig. 1. Subjects warmed-up by performing 20 submaximal
(20–80% of the estimated maximum effort) concentric and
eccentric contractions of the thigh muscles (reciprocal for knee
extensors and knee flexors) at slow angular velocities (15
and )15? s)1, respectively). Subjects were also asked to
complete two to three submaximal practice repetitions prior
to each test series. For the isokinetic trials, range of motion was
70?, from 80? to 10? of knee flexion (0? corresponding to knee
Figure 1 Experimental protocol. Rest periods
of 1 min were interspersed between series. KE,
knee extensors; KF, knee flexors; ROM, range
of motion.
? 2007 The Authors
Journal compilation ? 2007 Blackwell Publishing Ltd •Clinical Physiology and Functional Imaging 27, 6, 346–353
Test–retest reliability, N. A. Maffiuletti et al.
347

Page 3

fully extended). Concentric measurements involved three
continuous, reciprocal (maximal) knee extensions and flexions,
which were performed at three preset constant angular
velocities, in the following order: 60, 120, 180? s)1(slow to
fast) (Wilhite et al., 1992). Eccentric measurements consisted of
three maximal contractions at a single velocity of )60? s)1.
Knee flexor and extensor trials were performed as discrete
movements in a single direction (i.e. non-reciprocal). For both
concentric and eccentric repetitions, subjects were exhorted to
push⁄pull as hard and as fast as possible and to complete the full
range of motion. For the isometric trials, the knee joint was
fixed at an angle of 60? of flexion, which has been demonstrated
to be the angle of maximal isometric force generation
(Thorstensson et al., 1976). Two isometric knee extensions
and flexions were performed, and subjects were consistently
asked to produce their maximal force rapidly (as fast and
forceful as possible) and then to maintain the contraction for
4–5 s. After the maximal isokinetic and isometric contractions,
subjects completed a fatigue test consisting of 20 continuous,
reciprocal (maximal) knee extensions and flexions, which were
performed at 180? s)1(total duration ?20 s). Throughout the
fatigue test, subjects were exhorted to push⁄pull as hard and as
fast as possible and to complete the full range of motion. It was
consistently verified that the average peak torque of the first
three to five contractions was similar to the values recorded
during the three-repetition series performed at 180? s)1in the
first part of the session, to ensure the maximal effort. Whatever
the action mode and the velocity, subjects recovered passively
for 60 s between series of measurements (isokinetic trials) or
between repetitions (isometric trials). The Con-Trex software
consistently indicated the duration of both contraction and rest
phases.
Criterion measures
Torque, position and angular velocity data were recorded from
the isokinetic dynamometer with a sampling rate of 100 Hz. For
concentric and eccentric strength trials, the software calculated a
large number of parameters, but we retained only those
commonly used in isokinetic studies, namely the peak torque,
the average work, the average power and the angle of peak
torque (Brown et al., 1992; Perrin, 1993; Brown, 2000; Wrigley
& Strauss, 2000). These variables were calculated for the knee
extensor and the flexor muscles. For each angular velocity, only
the two trials giving the highest peak torque were considered.
The velocity throughout each repetition was analysed and it was
also verified that, at the faster angular velocity, peak torque was
developed during the constant velocity period. Isometric torque
data were exported and analysed with Windows-based software
(Acqknowledge; Biopac Systems Inc., Santa Barbara, CA, USA).
Isometric MVC torque, i.e. the highest torque value, and
maximal RTD, i.e. the highest positive value from the first
derivative of the torque signal (greatest slope of torque-time
curve), were calculated for all knee extensor and flexor trials.
For the fatigue test, only peak torque and average work
associated with the 20 consecutive repetitions were retained.
Data associated with the first contraction of each series were
consistently removed from the analyses (see Fig. 2). For
respective muscles, torque (Fig. 2a) and work (Fig. 2b) values
associated with the first four repetitions (contraction 2–5) and
last four repetitions (contraction 17–20) of the series were
averaged and considered as prefatigue and postfatigue data,
respectively. Peak torque and work losses were then calculated as
the per cent difference between postfatigue and prefatigue data
(Thorstensson & Karlsson, 1976). Moreover, the decline in
torque and work via the negative slope was determined by linear
regression analysis from the second to the last (20th) repetition
(Pincivero et al., 2000, 2001), for respective muscle groups (see
Fig. 2).
Statistical analyses
For muscle strength parameters (peak torque, work, power,
angle of peak torque, isometric MVC torque and RTD), within-
session (trial 1 versus trial 2) and between-session (average of
trials 1 and 2 of session 1 versus session 2) reliability were
calculated. For muscle fatigue variables (prefatigue, postfatigue,
per cent loss and slope of both peak torque and work data), only
test–retest reliability was calculated.
Relative reliability concerns the degree to which individuals
maintain their position in a sample with repeated measurements
(Atkinson & Nevill, 1998). We assessed this type of reliability
with the intraclass correlation coefficient (ICC) (2,1), a two-
way random effects model with single-measure reliability in
which variance over the repeated session is considered (Shrout &
Fleiss, 1979). The ICC indicates the error in measurements as a
proportion of the total variance in scores. As a general rule, we
considered an ICC over 0Æ90 as high, between 0Æ80 and 0Æ90 as
moderate and below 0Æ80 as insufficient (Vincent, 1999).
Absolute reliability is the degree to which repeated measure-
ments vary for individuals (Atkinson & Nevill, 1998); this type
of reliability is usually expressed as a proportion of the
measured values, i.e. coefficient of variation (CV). CV refers to
intrasubject variation between two measurements. For each
subject, CV was calculated as: (SD of two measurements⁄mean of
two measurements) · 100. To interpret the CV values, we used
the arbitrary suggestions made by Stokes (1985) with an
analytical goal of 15% or below.
Following a repeated measures ANOVA on trials within the
same session, paired sample t-tests (bilateral) were used to
determine whether the average muscle strength and fatigue data
obtained in the session 1 were significantly different from
session 2. Statistical significance was set at P<0Æ05. All statistical
procedures were performed with SPSS 12.0.1 statistical software
(SPSS Inc., Chicago, IL, USA).
Results
For all the criterion measures, as (i) CV were not significantly
different (unpaired Student?s t-test) and (ii) average ICC were
Test–retest reliability, N. A. Maffiuletti et al.
? 2007 The Authors
Journal compilation ? 2007 Blackwell Publishing Ltd •Clinical Physiology and Functional Imaging 27, 6, 346–353
348

Page 4

almost identical between men and women, all data were
collapsed across gender.
Muscle strength
Mean isokinetic and isometric strength data for sessions 1 and 2,
together with the associated CV and ICC for within and
between-session reliability are presented in Table 1 (knee
extensors) and Table 2 (knee flexors). For both muscle groups,
isokinetic peak torque, work and power as well as isometric
MVC torque and RTD demonstrated moderate-to-high reliabil-
ity, with CV lower than 9Æ7% and ICC higher than 0Æ86.
Concentric peak torque at the fastest velocity (180? s)1)
significantly increased from session 1 to session 2 for both the
knee extensor and the flexor muscles (P<0Æ05). In the same
way, eccentric peak torque (P<0Æ05), work (P<0Æ01) and power
? 2007 The Authors
Journal compilation ? 2007 Blackwell Publishing Ltd •Clinical Physiology and Functional Imaging 27, 6, 346–353
(P<0Æ01) of the knee extensor muscles as well as angle of peak
torque at 60? s)1(P<0Æ05) significantly increased from session
1 to session 2. For the knee extensor muscles, angle of peak
torque data showed a high reliability for the concentric but not
for the eccentric mode (low ICC). Both within and between-
session reliability for the knee flexors angle of peak torque was
insufficient (low ICC), except for 180? s)1.
To gain insight into (between-session) reliability results, CV
and ICC data were grouped as a function of muscle group (knee
extensors versus flexors), isokinetic variable (peak torque versus
work versus power versus angle of peak torque) and angular
velocity()60versus0versus60versus120versus180? s)1).We
observedthehighestICCandthelowestCV(i.e.betterreliability),
respectively, for knee extensors (Fig. 3a), peak torque (Fig. 3b)
and 180? s)1data (Fig. 3c), while knee flexors, angle of peak
torque and eccentric data showed the lowest reliability.
(a)
(b)
Figure 2 Decline in knee extensor and knee
flexor peak torque (a) and average work (b)
across the 20 repetitions (mean values of 30
subjects, session 1). The error bars represent
SD. The average values of contraction 2–5 and
17–20 (shaded areas) were considered as
prefatigue and postfatigue values, respectively.
The difference between prefatigue and postfa-
tigue data (peak torque and work loss) was
calculated in per cent values. Moreover, torque
and work losses were quantified as the slope of
the linear fit for respective muscle groups.
KE, knee extensors; KF, knee flexors.
Test–retest reliability, N. A. Maffiuletti et al.
349

Page 5

Table 1 Reliability of isokinetic and isometric muscle strength data for the knee extensors.
Mean ± SD
Within-session
reliability
Between-session
reliability
Session 1Session 2 CV ICCCV ICC
Peak torque (N m)
60? s)1
120? s)1
180? s)1
)60? s)1
Average work (J)
60? s)1
120? s)1
180? s)1
)60? s)1
Average power (W)
60? s)1
120? s)1
180? s)1
)60? s)1
Angle of peak torque (?)
60? s)1
120? s)1
180? s)1
)60? s)1
Isometric MVC torque (N m)
Isometric RTD (N m s)
177Æ7 ± 46Æ1
152Æ4 ± 40Æ2
132Æ5 ± 37Æ9
202Æ4 ± 68Æ0
177Æ6 ± 45Æ9
154Æ9 ± 40Æ4
135Æ6 ± 38Æ9*
211Æ4 ± 61Æ2*
2Æ8
1Æ9
1Æ9
3Æ4
0Æ996
0Æ998
0Æ999
0Æ995
3Æ2
2Æ8
3Æ3
7Æ0
0Æ978
0Æ991
0Æ993
0Æ965
169Æ9 ± 42Æ1
152Æ9 ± 40Æ8
136Æ7 ± 39Æ7
172Æ0 ± 56Æ4
168Æ1 ± 41Æ2
152Æ9 ± 38Æ4
137Æ5 ± 38Æ4
182Æ6 ± 53Æ3**
2Æ7
2Æ1
2Æ4
3Æ9
0Æ996
0Æ998
0Æ999
0Æ994
4Æ1
3Æ7
4Æ0
7Æ2
0Æ966
0Æ979
0Æ986
0Æ971
119Æ0 ± 28Æ1
187Æ1 ± 48Æ5
218Æ8 ± 64Æ1
125Æ3 ± 41Æ0
120Æ4 ± 30Æ1
188Æ9 ± 46Æ2
223Æ0 ± 64Æ3
133Æ5 ± 39Æ9**
2Æ8
2Æ6
2Æ9
4Æ3
0Æ996
0Æ996
0Æ998
0Æ993
4Æ7
3Æ4
4Æ6
7Æ3
0Æ956
0Æ983
0Æ981
0Æ971
53Æ4 ± 6Æ9
47Æ6 ± 7Æ1
47Æ1 ± 6Æ6
60Æ4 ± 7Æ1
214Æ1 ± 61Æ9
521Æ2 ± 161Æ3
55Æ2 ± 7Æ4*
48Æ1 ± 6Æ7
47Æ2 ± 6Æ8
62Æ3 ± 7Æ9
221Æ4 ± 57Æ6
513Æ2 ± 143Æ9
4Æ0
3Æ5
4Æ1
6Æ9
4Æ4
8Æ3
0Æ927
0Æ968
0Æ943
0Æ766
0Æ983
0Æ923
4Æ3
4Æ3
4Æ0
7Æ1
5Æ5
9Æ1
0Æ914
0Æ920
0Æ911
0Æ483
0Æ972
0Æ874
CV, coefficient of variation; ICC, intraclass correlation coefficients; MVC, maximal voluntary contraction; RTD, rate of torque development. Signifi-
cantly higher than session 1 at *P<0Æ05 and **P<0Æ01, respectively.
Table 2 Reliability of isokinetic and isometric muscle strength data for the knee flexors.
Mean ± SD
Within-session
reliability
Between-session
reliability
Session 1Session 2 CVICC CV ICC
Peak torque (N m)
60? s)1
120? s)1
180? s)1
)60?s)1
Average work (J)
60? s)1
120? s)1
180? s)1
)60? s)1
Average power (W)
60? s)1
120? s)1
180? s)1
)60? s)1
Angle of peak torque (?)
60? s)1
120? s)1
180? s)1
)60? s)1
Isometric MVC torque (N m)
Isometric RTD (N m s)
109Æ2 ± 30Æ9
96Æ7 ± 28Æ2
88Æ3 ± 27Æ1
129Æ5 ± 38Æ0
109Æ6 ± 30Æ5
98Æ1 ± 28Æ2
90Æ9 ± 25Æ9*
129Æ7 ± 40Æ5
3Æ6
2Æ9
2Æ7
3Æ4
0Æ995
0Æ998
0Æ997
0Æ996
3Æ1
4Æ2
4Æ1
6Æ4
0Æ988
0Æ987
0Æ988
0Æ971
111Æ0 ± 32Æ0
101Æ7 ± 29Æ7
93Æ4 ± 28Æ4
123Æ8 ± 37Æ7
110Æ6 ± 29Æ6
101Æ5 ± 28Æ8
94Æ8 ± 27Æ2
124Æ5 ± 37Æ8
3Æ4
3Æ6
3Æ2
4Æ0
0Æ995
0Æ996
0Æ996
0Æ993
3Æ8
5Æ5
4Æ1
5Æ7
0Æ977
0Æ972
0Æ985
0Æ976
71Æ7 ± 21Æ9
119Æ7 ± 36Æ4
144Æ7 ± 44Æ2
90Æ6 ± 27Æ6
76Æ4 ± 20Æ0
122Æ5 ± 37Æ5
148Æ7 ± 42Æ7
91Æ8 ± 27Æ9
8Æ8
6Æ2
4Æ9
4Æ0
0Æ859
0Æ978
0Æ992
0Æ993
8Æ4
6Æ9
6Æ1
5Æ9
0Æ912
0Æ962
0Æ960
0Æ975
29Æ7 ± 6Æ8
35Æ7 ± 7Æ7
43Æ8 ± 16Æ5
24Æ1 ± 9Æ5
106Æ5 ± 32Æ5
247Æ2 ± 77Æ8
27Æ8 ± 5Æ6
35Æ2 ± 10Æ2
40Æ0 ± 15Æ1
25Æ3 ± 8Æ1
107Æ7 ± 33Æ2
252Æ3 ± 82Æ6
10Æ2
10Æ2
8Æ4
13Æ5
3Æ4
9Æ7
0Æ769
0Æ706
0Æ907
0Æ931
0Æ991
0Æ906
11Æ1
10Æ2
11Æ6
18Æ1
4Æ7
9Æ3
0Æ520
0Æ703
0Æ836
0Æ731
0Æ975
0Æ902
CV, coefficient of variation; ICC, intraclass correlation coefficients; MVC, maximal voluntary contraction; RTD, rate of torque development. *Sig-
nificantly higher than session 1 at P<0Æ05.
Test–retest reliability, N. A. Maffiuletti et al.
? 2007 The Authors
Journal compilation ? 2007 Blackwell Publishing Ltd •Clinical Physiology and Functional Imaging 27, 6, 346–353
350