Available online at www.sciencedirect.

com

Journal of Science and Medicine in Sport 14 (2011) 541546

Original research

Isokinetic dynamometry applied to shoulder rotators Velocity

limitations in eccentric evaluations
Gisele G. Zanca, Ana B. Oliveira, Michele F. Saccol, Stela M. Mattiello-Rosa
Department of Physical Therapy, Federal University of So Carlos, So Carlos, SP, Brazil
Received 11 November 2010; received in revised form 20 April 2011; accepted 5 May 2011

Abstract
The objectives of this study were to evaluate if collegiate overhead athletes, with and without shoulder pain, and non-athletes could reach
a preset velocity in internal and external shoulder rotation isokinetic evaluations; and to evaluate the correlation between torque and velocity.
Controlled laboratory study, cross-sectional. Evaluations were performed using the isokinetic dynamometer Biodex System 3. Participants
were assessed seated, with the arm at 90 of shoulder abduction and 90 of elbow flexion, from neutral rotation to 90 of external rotation.
Five maximal contractions of isokinetic concentric and reactive eccentric internal and external rotation were performed at the velocities
90 /s, 180 /s and 240 /s. Data were processed with using MatLab. Most participants did not reach the isokinetic phase during eccentric tests
at 180 /s and 240 /s, particularly in the external rotators evaluation. High correlations between torque and velocity of eccentric tests were
found. The groups presented no differences in maximal velocity attained in trials which preset velocity was not reached. These results call
into question the use of reactive eccentric tests at velocities higher than 180 /s for the isokinetic evaluation of shoulder external rotators in
collegiate overhead athletes and non-athletes in this specific position. In such cases, careful evaluation of the velocity is recommended to
determine if the isokinetic phase was reached.
2011 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.

Keywords: Athlete; Muscle strength; Shoulder joint; Torque

1. Introduction Even for athletes, the use of high velocities to evaluate

Isokinetic evaluation is a widely used tool for studying
the shoulder muscle performance of healthy and injured over-
head athletes.13 Considering that high velocities are reached
during the throw, the greatest velocities available in isoki-
netic dynamometers are used to evaluate internal and external
rotation of the shoulder.46
The mechanical reliability and validity of the isokinetic
dynamometer Biodex System 3, in concentric mode, were
evaluated by Drouin et al.,7 who observed that, during tests
with angular velocity of 300 /s or more, the preset velocity
was not reached. The authors questioned whether the torque
applied during a relatively short range of motion (ROM) was
sufficient for attaining the higher selected velocities.7
Corresponding author.
E-mail address: stela@ufscar.br (S.M. Mattiello-Rosa).
muscle performance seems to be questionable. Alfredson
et al.8 did not use velocities greater than 180 /s in the concen-
tric mode to evaluate female volleyball players and sedentary
women because participants from both groups were not able
to perform the test. Mullaney and McHugh9 have used 120 /s
to evaluate non-athletes because they were unable to reach
peak torque during the isokinetic phase at 180 /s. Both studies
have used the seated position, 90 of shoulder abduction and
90 of elbow flexion (90/90 position),8,9 which has been rec-
ommended because it approximates the throwing position.10
It seems coherent that the torque used for isokinetic evalua-
tion of muscle performance be restrict to the isokinetic phase
of the movement evaluated. According to Handel et al.,11
torque generated beyond this phase is subject to acceler-
ation artifacts and thus may not reliably represent muscle
performance.
Our group has been evaluating athletes with and without
shoulder pain using isokinetic dynamometry, emphasizing

1440-2440/$ see front matter 2011 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.jsams.2011.05.002
542 G.G. Zanca et al. / Journal of Science and Medicine in Sport 14 (2011) 541546
shoulder internal and external rotation. In order to reproduce
throwing movement, it has been recommended the use of high
velocities, and we have noticed that participants have diffi-
culties reaching these speeds. In such cases, the analysis of
torque data without simultaneous examination of the veloc-
ity could lead to evaluation of a torque produced beyond the
isokinetic phase, which would compromise the principles of
the test.
Due to the importance of considering the isokinetic phase
in the evaluation of muscle performance and to the difficulty
of individuals in reaching high velocities, this study proposes
to analyze isokinetic evaluation data with respect to velocity.
Therefore, the objectives were (1) to determine if collegiate
overhead athletes, with and without shoulder pain, and non-
athletes could reach a preset velocity in isokinetic evaluations
of shoulder internal and external rotation in 90/90 position;
(2) to evaluate the correlation between torque and velocity in
eccentric tests.
2. Methods
Fifty-one male collegiate overhead athletes (hand-
ball, baseball, volleyball and water polo players), were
evaluated and divided into two groups: athletes with
shoulder pain (GAP; n = 21; 21.6 2.1 years; 176 8 cm;
77.5 11.7 kg; 7.33 4.23 years practicing sports), and
asymptomatic athletes (GAA; n = 30; 21.7 2.7 years;
178 9 cm; 76.2 11.9 kg; 7.27 4.21 years practicing
sports). The inclusion criteria for the GAP were complaints of
pain during the throw, and at least two positive tests of the fol-
lowing ones, for pain criterion: apprehension,12 relocation,12
Neer,13 Hawkins and Kennedy14 and Jobe and Moynes.15
We also evaluated 18 non-athletes (GNA; 22.4 2.5 years;
175 6 cm; 77.9 12.2 kg) who had no previous history of
injury to the neck or shoulder. The study excluded individ-
uals with generalized joint laxity, previous neck or shoulder
surgery or systemic diseases. This study was approved by
the Ethics Committee of the University, all participants gave
their written and informed consent agreement to participate,
and the rights of volunteers were protected.
The evaluations were carried out using the isokinetic
dynamometer Biodex Multi-Joint System 3 (Biodex Med-
ical System Inc., NY, USA) and recorded at 100 Hz. The
dynamometer calibration was checked before every evalua-
tion session. For isokinetic concentric mode was used cushion
at 1, and for eccentric reactive mode, at 4.
The evaluations were preceded by a warm-up based on
active-free shoulder movements. The participants were evalu-
ated in seated position while secured with pelvic and diagonal
straps for trunk stabilization. The dominant arm (throw-
ing arm) was positioned at 90 of shoulder abduction and
90 of elbow flexion. The olecranon was aligned with the
dynamometers mechanical axis of rotation. Gravity correc-
tion was performed with the arm relaxed at 90 of shoulder
abduction, in neutral rotation.
Shoulder internal and external rotation torque was eval-
uated through a ROM of 90 , from neutral rotation to 90
of external rotation, at three angular velocities: 90 /s, 180 /s
and 240 /s. For familiarization, the participants performed
three submaximal repetitions at each velocity and mode.
After 1 min of rest, five maximal reciprocal repetitions were
performed, and the participants received standardized verbal
encouragement to develop maximal strength in all con-
tractions. The concentric tests were carried out before the
eccentric tests4 in ascending order of velocity.16 Two minutes
of rest were allowed between each velocity. The torque limit
for the eccentric reactive mode was adjusted for each vol-
unteer. The concentric peak torque was used as reference by
adding, in general, from 40 to 60 Nm to this value, according
to their response to the familiarization trials.
Data from the isokinetic tests were exported and processed
using MatLab (version 7.0.1, MathWorks Inc., Natick, MA,
USA). Velocity, torque, and temporal axis data were utilized.
An analysis was carried out to verify if the preset velocity had
been reached in each trial, considering a 2% margin of error.
Although the accuracy for velocity is given as 1 /s, a detailed
analysis of the velocity recordings showed that there is an
error of 2% identified when comparing periods of plateau
and the preset velocity.
When the preset velocity was reached (successful trials),
the period in which it was maintained (isokinetic phase) was
identified. The trial with the longest isokinetic phase was con-
sidered as the best trial. In the cases that the preset velocity
was not reached in any of the five repetitions, the trial that
attained the highest angular velocity was considered the best
trial. Finally, a cross-correlation analysis was applied to ver-
ify the coherence and phasic relationships between torque
and velocity in eccentric evaluations.
Statistical analysis was carried out with SigmaPlot for
Windows (version 11.0, Systat Software Inc., Chicago, IL,
USA). The ShapiroWilks test showed that data were non-
normal distributed. The KruskalWallis test was used for
group comparisons of maximum velocity in repetitions where
the preset velocity was not reached. When its prerequisites
were met, the Chi-square test was used to compare the num-
ber of participants who did or did not reach the test velocity
among the groups. Significance was set at 5% for all analyses.
3. Results
All participants reached the preset velocities in the five
repetitions during the concentric test as well as in the eccen-
tric test at 90 /s (Fig. 1). However, in the eccentric tests at
180 /s and 240 /s, a large number of participants from the
three groups did not reach the isokinetic phase, particularly
during the evaluation of shoulder external rotators (Fig. 1C).
In the eccentric test of internal rotators at 180 /s, 4.8%
of the participants from the GAP did not reach the preset
velocity in any trial, and at 240 /s, 23.8% of the GAP, 13.3%
of the GAA, and 5.6% of the GNA did not reach it. In the
 G.G. Zanca et al. / Journal of Science and Medicine in Sport 14 (2011) 541546 543

Fig. 1. Maximum velocity reached by the subjects from the groups of athletes with shoulder pain (GAP), asymptomatic athletes (GAA) and non-athletes
(GNA), in their best trial, for each of the test velocities (90 /s, 180 /s and 240 /s), for concentric (A) and eccentric (C) shoulder external rotation, and concentric
(B) and eccentric (D) shoulder internal rotation.

eccentric test of external rotators at 180 /s, 19% of the GAP
and 20% of the GAA did not reach the preset velocity in any
of the trials, while at 240 /s, 66.7% of the GAP, 73.3% of
the GAA, and 44.4% of the GNA did not reach the isokinetic
phase. For this last condition, the numbers of participants who
reached or did not reach the preset velocity were compared
among groups. The Chi-square test revealed no intergroup
differences (p > 0.05).
The above data refers to the participants who did not reach
the preset velocity in any of the five trials. When the perfor-
mance of the participants who reached the preset velocity was
evaluated, we observe that a large percentage did not reach
the isokinetic phase in all trials (Fig. 2). Once again, difficulty
was observed mainly in the eccentric test of external rotators
at 240 /s.
There was no intergroup difference in the maximum veloc-
ity reached during the trials where the isokinetic phase was
not reached (Table 1).
High correlations17 were found between torque and
velocity in eccentric tests, and phase offset was consistent.
The mean standard deviation of maximal correlations
(mean standard deviation of phase offset) for internal
and external rotations were, respectively, 0.97 0.02
(0.07 0.05 s) and 0.95 0.03 (0.07 0.06 s) at
90 /s; 0.93 0.03 (0.10 0.04 s) and 0.88 0.10
(0.09 0.10 s) at 180 /s; and 0.91 0.04 (0.10 0.04 s)
and 0.89 0.04 (0.08 0.04 s) at 240 /s.
4. Discussion
The results demonstrated that higher velocities were not
always reached during eccentric tests, in all evaluated groups.
This behavior was not observed in the concentric tests, at
any evaluated velocity, or in the eccentric test at 90 /s. The
velocity 240 /s was not reached in most eccentric tests of
external rotators. If we consider only the number of successful
trials, the performance was even worse. Only six participants
out of a total of 69 reached 240 /s in the five repetitions of
the eccentric test of external rotators. The high correlations
between torque and velocity were found with a mean phase
offset of 0.070.10 s, demonstrating the delay between torque
544 G.G. Zanca et al. / Journal of Science and Medicine in Sport 14 (2011) 541546

Fig. 2. Number of successful trials performed by participants from the groups of athletes with shoulder pain (GAP), asymptomatic athletes (GAA) and
non-athletes (GNA) in the eccentric tests of shoulder external (A) and internal (B) rotation, at 180 /s and 240 /s.

generation and the velocity response. This result suggests that
the inefficiency in reaching the preset velocity may be related
to the participants difficulty in generating torque.
Our results are different from others that used the angular
velocity of 300 /s.46 A possible explanation for these con-
flicting findings could be the population studied, since we
evaluated collegiate overhead athletes from several sports,
while others have evaluated elite handball players6 and col-
legiate pitchers.5 Considering the sport level and the more
specific training of the pitchers to throw at high velocities, the
population from these studies could be a factor that has influ-
enced the different findings. However, Noffal4 has evaluated
collegiate baseball players (no pitchers) and college students
that did not participate in throwing sports. Therefore, the dif-
ferences found between our study and the previous ones seem
to be related mainly to the position of the participants during
the test (seated versus supine), and/or to the ROM (90 versus
120/150 ). These hypotheses are reinforced considering that
both studies8,9 which reported that their participants could
not reach velocities of 180 /s or higher have used the same
position and ROM of our study.
Scapular stabilization depends more on muscle activity in
the seated position than in the supine position, when the body
weight contributes to the stabilization of the scapula, result-
ing in a steady base for the activity of the shoulder rotator
muscles. This steady base could benefit the lengthtension
relation of these muscles, particularly the external rotators,
considering that all of them are attached to the scapula. This
could partially explain the evaluated participants difficulty
in generating torque sufficient to reach high velocities during
eccentric tests of the external rotators in seated position. The
lesser difficulty observed in the eccentric tests of the internal
rotators could be related to the activity of the muscles that
are not attached to the scapula, such as latissimus dorsi and
pectoralis major. While external rotators depend on scapu-
lar stabilization for better performance, internal rotators do
not totally depend on this factor. Therefore, the good per-
formance of external rotators in a seated position would be
associated with efficient performance of the serratus anterior
for scapular stabilization,18 which could have happened to
the participants who reached the preset velocity.
By contrasting our results with other studies in the liter-
ature, it may seem advantageous to use the supine position
in the isokinetic evaluation of shoulder internal and external
rotators. However, it can be considered less functional, par-
ticularly when evaluating the throwing movement. The effect
of individual positioning on the isokinetic evaluation of the
shoulder rotator muscles needs further investigation.
Another factor that could have contributed to reduced
eccentric performance of the external rotators is arm position.

Table 1
Maximum velocity reached during repetitions which the isokinetic phase was not reached in eccentric tests at 180 /s and 240 /s by the groups of athletes with
shoulder pain (GAP), asymptomatic athletes (GAA) and non-athletes (GNA).
GAP GAA GNA p Value
IR 180 /s 136.8 (16.7173.5) 149.0 (85.1175.5) 166.7 (166.3171.0) 0.23
ER 180 /s 155.7 (42.6169.7) 142.7 (98.4172.5) 168.7 (100.6174.2) 0.17
IR 240 /s 197.3 (114.7232.9) 196.7 (109.5233.6) 214.1 (163.4231.0) 0.67
ER 240 /s 181.2 (101.4218.6) 184.7 (95.5234.8) 199.0 (141.6232.9) 0.13
ER, external rotation; IR, internal rotation.
Data are median (minimummaximum).
 G.G. Zanca et al. / Journal of Science and Medicine in Sport 14 (2011) 541546
Because the starting point for the humerus is the end ROM of
external rotation, there can be a deficit in the external rotator
muscles activity, because they are shorted, which could com-
promise the length/tension relation. Therefore, these muscles
would be less capable of generating torque and, consequently,
of accelerating the dynamometer shaft up to greater veloci-
ties.
The ROM used to test the shoulder rotators also
could interfere in the maximum velocity reached. Previous
studies46 used ROM of 120 and 150 , while the ROM in
the present study was limited to 90 , the same of other studies
that reported similar difficulties.8,9 We selected this ROM in
order to avoid the symptoms in athletes with shoulder pain
during an excessive ROM of internal rotation. It could be
suggested that if there had been a greater ROM to allow
for torque development, the participants could have reached
240 /s with less difficulty. However, considering that all par-
ticipants reached the preset velocity in the concentric tests
and in most repetitions of the eccentric evaluations of internal
rotators, the ROM does not seem to have been a determin-
ing factor for these results. Therefore, the hypothesis that the
position of the individual interferes with their performance is
reinforced.
Some dynamometer settings could also have potentially
affected the participants performance. The end stop cush-
ion regulates the point at the deceleration starts. The higher
the cushion the shorter the amount of time spent at preset
velocity. We used cushion set at 1 for isokinetic concentric
mode, which is recommended by the equipment manual for
testing protocols. In the eccentric tests, a softer cushion, at
4, was selected in order to avoid a hard end of motion, con-
sidered uncomfortable by participants with symptoms during
the preliminary testing, mainly around 90 of external rota-
tion. It is possible that with a harder cushion, the participants
could have had more time to reach the preset velocity. On
the other hand, if the participants had experienced pain dur-
ing the isokinetic evaluation, the strength production could
be affected and the ability to reach the preset velocity would
be even more compromised. Moreover, cushioning was the
same for internal and external rotations, but the difficulty was
found mainly for external rotation.
Another adjustable setting in reactive eccentric mode is the
torque limit, which determines a value that, if is exceeded,
the dynamometer shaft stops rotating. To initiate the move-
ment, subject needs to meet a torque threshold of 10% torque
limit. To our knowledge, there is no standardized procedure
to determine this parameter; therefore, we have used the peak
torque in concentric tests as reference. In general, it was
added from 40 to 60 Nm in peak torque of reference to deter-
mine the torque limit. This value was also adjusted according
difficulties on familiarization trials.
Previous studies that evaluated eccentric shoulder rota-
tions at high velocities did not state which mode was used.46
It is possible that the passive mode had been used, which guar-
antees that the preset velocity is reached, since it is under
complete control of the machine. Nevertheless, when using
545
high velocities we have considered safer to use the reactive
mode instead of passive one. At the passive mode, the sub-
ject could not produce sufficient torque, and consequently
shoulder structures could be damaged.
It was expected that athletes would present less difficult
during the tests at high velocities than non-athletes as an
adaptation to throwing performance, since the external rotator
muscles act eccentrically to decelerate the humerus at the end
of the throw, which happens at high velocities.19 However,
despite the overhead sports experience of GAP and GAA,
the three groups had the same difficulty in reaching greater
velocities in the eccentric test. The literature shows that over-
head athletes present, as an adaptation to training, an increase
in the strength of internal rotators without a concomitant
gain of strength of external rotators.20 It has been suggested
that this adaptation could contribute to a better throwing
performance.4 Moreover, the high velocities reached during
a throw are a summation of numerous joint movements and
muscular forces21 that are very difficult to reproduce in the
evaluation of an isolated movement. Concerning the athletes
with shoulder pain, all of them have maintained a regular
routine of sports training, despite of shoulder pain. It is pos-
sible that this level of shoulder pain, which allows continuing
training, did not affect the ability to reach high velocities dur-
ing isokinetic evaluations, or that the regular sports practice
have conserved this skill.
The difficulty of athletes as much as non-athletes during
the eccentric tests at 180 /s or higher, particularly for exter-
nal rotation, highlights the need for careful use of this type
of evaluation. Considering that torque used to analyze mus-
cle performance should refer to the isokinetic phase of the
movement, the simultaneous analysis of velocity and torque
data is recommended.
The results point to the need for new studies investigating
the methodological aspects of isokinetic tests at high veloc-
ities, in order to increase our understanding of the effects of
position, ROM and number of successful trials performed on
a test. Such understanding would contribute to the develop-
ment of procedures that ensure a more accurate evaluation of
isokinetic peak torque. Furthermore, dynamometer settings,
like cushioning and torque limit, and the mode used should
be reported in the studies in order to guarantee the repro-
ducibility of the methods and to allow adequate comparisons
between studies.
5. Conclusion
The evaluation of eccentric torque of shoulder rotation
at high velocities in the position 90/90, in collegiate over-
head athletes with and without pain and in non-athletes,
should be used carefully, particularly for the external rota-
tors. The high correlations found between torque and velocity
demonstrated that this problem is related to the difficulty in
generating torque with the external rotators in this specific
position.
546 G.G. Zanca et al. / Journal of Science and Medicine in Sport 14 (2011) 541546
6. Practical implications
The difficulty during the eccentric evaluations of shoulder
external rotators at high velocities was found in collegiate
overhead athletes as much as in non-athletes. Therefore, it
is questionable the validity of this kind of evaluation for
this population.
The difficulty observed is probably due to the relative exter-
nal rotators weakness with the shoulder abducted at 90 ,
in seated position.
For evaluations at high velocities, it is recommended
simultaneous analysis of torque and velocity, in order to
guarantee that the peak torque analyzed is a reliable mea-
sure of muscle strength.
Acknowledgments
The authors are deeply grateful to the volunteers who
participated in this study. Gisele Garcia Zanca and Michele
Forgiarini Saccol were the recipients of research from
Fundaco de Amparo Pesquisa do Estado de So Paulo.
References
1. Stuelcken MC, Ginn KA, Sinclair PJ. Shoulder strength and range of
motion in elite female cricket fast bowlers with and without a history
of shoulder pain. J Sci Med Sport 2008;11(6):575580.
2. Stickley CD, Hetzler RK, Freemyer BG, et al. Isokinetic peak torque
ratios and shoulder injury history in adolescent female volleyball ath-
letes. J Athl Train 2008;43(6):571577.
3. Saccol MF, Gracitelli GC, Silva RT, et al. Shoulder functional ratio in
elite junior tennis players. Phys Ther Sport 2010;11(1):811.
4. Noffal GJ. Isokinetic eccentric-to-concentric strength ratios of the
shoulder rotator muscles in throwers and nonthrowers. Am J Sports
Med 2003;31(4):537541.
5. Carter AB, Kaminski TW, Douex Jr AT, et al. Effects of high volume
upper extremity plyometric training on throwing velocity and functional
strength ratios of the shoulder rotators in collegiate baseball players. J
Strength Cond Res 2007;21(1):208215.
6. Andrade MS, Fleury AM, Lira CAB, et al. Profile of isokinetic
eccentric-to-concentric strength ratios of shoulder rotator muscles in
elite female team handball players. J Sports Sci 2010;28(7):743749.
7. Drouin JM, Valovich-McLeod TC, Shultz SJ, et al. Reliability and valid-
ity of the Biodex system 3 pro isokinetic dynamometer velocity, torque
and position measurements. Eur J Appl Physiol 2004;91(1):2229.
8. Alfredson H, Pietil T, Lorentzon R. Concentric and eccentric shoulder
and elbow muscle strength in female volleyball players and non-active
females. Scand J Med Sci Sports 1998;8(5):265270.
9. Mullaney MJ, McHugh MP. Concentric and eccentric muscle fatigue
of the shoulder rotators. Int J Sports Med 2006;27(9):725729.
10. Scoville CR, Arciero RA, Taylor DC, et al. End range eccentric antag-
onist/concentric agonist strength ratios: a new perspective in shoulder
strength assessment. J Orthop Sports Phys Ther 1997;25(3):203207.
11. Handel M, Dickhuth HH, Mayer F, et al. Prerequisites and limitations
to isokinetic measurements in humans: Investigations on a servomotor-
controlled dynamometer. Eur J Appl Physiol 1996;73(34):225
230.
12. Cools AM, Cambier DC, Witvrouw EE. Screening the athletes shoul-
der for impingement symptoms: a clinical reasoning algorithm for early
detection of shoulder pathology. Br J Sports Med 2008;42(8):628635.
13. Neer CS. Anterior acromioplasty for the chronic impingement syn-
drome in the shoulder: a preliminary report. J Bone Joint Surg
1972;54(1):4150.
14. Hawkins RJ, Kennedy JC. Impingement syndrome in athletes. Am J
Sports Med 1980;8(3):151158.
15. Jobe FW, Moynes DR. Delineation of diagnostic criteria and a
rehabilitation program for rotator cuff injuries. Am J Sports Med
1982;10(6):336339.
16. Dvir Z. Isokinetics: testing, interpretation, and clinical applications.
New York: Churchill Livingston; 1995.
17. Domholdt E. Physical therapy research principles and applications.
second ed. Philadelphia: W.B. Saunders Company; 2000.
18. Phadke V, Camargo PR, Ludewig PM. Scapular and rotator cuff muscle
function during arm elevation: a review of normal function and alter-
ations with shoulder impingement. Rev Bras Fisioter 2009;13(1):19.
19. Escamilla RF, Andrews JR. Shoulder muscle recruitment patterns
and related biomechanics during upper extremity sports. Sports Med
2009;39(7):569590.
20. Ellenbecker T, Roetert EP. Age specific isokinetic glenohumeral inter-
nal and external rotation strength in elite junior tennis players. J Sci
Med Sport 2003;6(1):6370.
21. Ellenbecker TS, Davies GJ. The application of isokinetics in test-
ing and rehabilitation of the shoulder complex. J Athl Train
2000;35(3):338350.