JOurn1al or in.ln err.3du.3u 1:e -e-arch
2 - No. en iler..'l:ece eni r -'ï¿½.,
Does Warming Up Influence Shoulder Stiffness or External Rotation in
Healthy, Non-Elite Tennis Players?
Daniel S. Peterson, Jeff T. Wight, Guy B. Grover, and Mark D. Tillman
During overhead motion sports, massive loads are placed on the shoulder. Research has shown that these loads
are closely associated with stiffness and flexibility at the shoulder. Understanding the way tension and
flexibility change is crucial for assessment as well as the importance of warming up. The purpose of this study was
to investigate changes in stiffness and range of motion (ROM) about the shoulder before and after a sport
Amateur tennis athletes (11 males and 11 females) were tested for both external motion and torque about
the shoulder. Both dominant and non-dominant arms were taken to a maximum external position and were
measured for torque at the shoulder and maximum ROM. Subjects then completed a tennis specific warm-
up consisting of light stretching, dry tennis serves, and finally several live serves. Subjects were again tested for
ROM and torque.
A two-way ANOVA was carried out (pre/post warm-up, dominant/non-dominant arms) showing significantly
larger post warm-up ROM, as well as larger torque values on the dominant arm. T-tests were also performed
to determine differences between pre- and post-warming up trials of each arm. Dominant arm external rotation
was shown to significantly increase while the non-dominant arm ROM did not. Tests of the onset of stiffness
(when the shoulder reached 2 Nm of torque) and slope of the angle/torque graph were calculated as measures
of stiffness, and neither showed any significant changes between arms or across warm-up trials. Within each
trial, subjects completed three measured external ROM repetitions. Significant differences were shown between
the first and third repetition within trials.
These findings show the importance of warming up as a tool to immediately increase ROM. Studies have shown
that overhead sport athletes with larger ROM about the shoulder are less susceptible to injury. Also, these results
are important to clinicians or therapists who are assessing ROM. Due to changes in the internal workings of
the shoulder as well as increased comfort of the subject, the arm's ROM is dynamic when "cold" and may
change significantly even within repetitions of the same trial.
The balance between mobility and stability in the shoulder is one of great complexity. Even in normal
movements people can experience pain and dislocations due to the relative instability of the joint. Athletes
who participate in activities involving overhead arm motions, such as throwing in baseball or serving in tennis,
put themselves at an especially high risk for shoulder-related injuries. Overhead motions are particularly
damaging due both to the extreme loads they place on the shoulder, as well as the severe arm positions which
are associated with these loads. These repetitive loads and exaggerated ranges of motion can cause significant
injury to athletes, and it is important to understand the circumstances around their occurrence.
Dynamic motions can be described by two measures: range of motion (ROM) and stiffness (torque required
to produce that ROM). Differences in ROM have been well documented and are highly variable both between
and within individuals. The variability in passive ROM is attributed to several factors, one of which being
the experience level of the athlete. Kilber, Chandler, Livingston, & Roetert (1996) showed that subjects who
have been participating in overhead motions for extended periods often show a decrease in internal
rotation. Moderate negative correlations between dominant internal rotation and both age and years of play
were reported. This is probably due to posterior capsule tightness, resulting from chronic microtraumatization of
the surrounding structures (Schmidt-Wiethoff, Rapp, Mauch, Schneider, & Appell 1996). It is unclear if
these adaptations are normal, but some studies have shown negative effects associated with lack of internal
rotation, including increased injury (Chandler, Kibler, Uhl, Wooten, Kiser, & Stone, 1990).
Differences in ROM have also been observed between athletes of different sports. Chandler et al. (1990) noted
that tennis athletes, when compared to other athletes, showed significantly different internal and external ranges
of motion. These tennis athletes were shown to be tighter in internal rotation and looser in external rotation for
both the dominant and non-dominant arms (Williford, East, Smith, & Burry, 1986).
Differences have also been shown within overhead-motion athletes (Ellenbecker & Mattalino, 1999). Research
has shown that while baseball pitchers had no significant difference between extremities in total ROM, tennis
athletes showed significantly less dominant arm total ROM (Ellenbecker, Roetert, Bailie, Davies, & Brown,
2002). These changes in ROM can have significant implications resulting in chronic injuries. Specifically, a
reduced internal ROM has been shown to be associated with increased injury and pain (Chandler et al.,
1990; Ellenbecker et al., 2002).
There have been some discrepancies in measured changes of external ROM. After testing professional tennis
athletes, Schmidt- Wiethoff et al. (1996) determined that there was indeed an increase in external rotation,
while Ellenbecker & Roetert (2003) noted that while internal rotation of non-dominant shoulders was less,
external rotation between dominant and non-dominant arms were not different.
As there are some differences between dominant and non-dominant arms, changes in ROM can also be
expected between athletes and non-athletes. In addition to shoulder stretching completed by most athletes,
the violent motions experienced during play induce changes in ROM. Both dominant and non-dominant shoulders
of experienced tennis athletes have been shown to have lesser internal ROM when compared to non-athletes
(Borsa, Dover, Wilk, & Reinold 2006).
Stiffness of the shoulder is another important factor in the assessment of dynamic motion. For overhead-
motion sports, similar external ranges of motion are necessary. Due to individual differences in shoulder stiffness,
the process in which this ROM is reached is extremely varied (Chow, Wight, Grover, Tillman, 2006). For this reason
it is important to note how to best measure these differences as well as the way in which they change.
Stretching has been shown to have an impact on stiffness and flexibility. Although benefits have been variable
with regards to certain activities, specifically maximal contractions, stretching has been recognized as a potential
way to decrease injury and increase performance in overhead motions such as pitching and tennis serving
(Fowles, Sale, & MacDougall, 2000; Young & Elliott 2001; and Nelson & Kokkonen 2001). As integral elements
of overhead-motion athletics, the shoulder and elbow joint have been scrutinized to determine how
stretching influences overhead sports mechanics and affects ROM and strength. However, most previous studies
have not taken into account the immediate changes in stiffness due to stretching. Most studies consider passive
ROM but neglect the change in ROM associated with warming up. Some studies have shown benefits but
were focused only on the lower body (Weijer, Gomiak, & Shamus 2003; Williford et al., 1986). Also, these
studies included a different type of "warm-up" than is often associated with overhead motion. A warm-up of the
lower body usually consists of superficial and deep-heat modalities, or repetitive exercises at low intensity, such
as jogging, bicycling (Taylor, Waring, & Brashear, 1995) and stair climbing (Hubley, Kozey, & Stanish,
1984; Wessling, Davane, & Hylton, 1987; and Williford et al., 1986).With overhead sports, a warm-up is geared
more toward in-game activity. For example, warming up the shoulder for a tennis match more often
includes stretching followed by some practice swings and light serves, followed by full speed serves. This is the
type of warm-up that was integrated into the present study.
The purpose of the study described here was to determine if any differences in stiffness or flexibility (ROM)
occurred within tennis athletes before and after warming up. The immediate effect of warming up on changes in
ROM on the shoulder may have implications in both clinical and research settings. If there are in fact differences
in pre and post collections, it would provide insight as to how data should be collected in the future. In addition,
when analyzing progress and recovery of shoulder injuries, it is important to know how best to assess ROM.
Further, changes in stiffness and flexibility may give insight into how important stretching really is in
A total of 22 participants, with an average height and weight of 174.5 ï¿½ 10.7 cm and 69.2 ï¿½ 12.9 kg,
respectively, were measured for pre- and post-external ROM changes. Subjects included 11 males and 11
females, and all were in good health. None of the subjects had sustained significant shoulder injuries in
recent months. All subjects had some experience with tennis, though the experience level ranged from two months
to several years. Most of the subjects were chosen from sport and recreation tennis classes on campus.
Other subjects were recruited from non-sport classes, provided they had "moderate tennis experience."
Apparatus and Instrumentation
The device used in this study (shown in Figures 1 and 2) was previously fabricated in conjunction with
ongoing projects at the University of Florida. Its specific purpose is to provide an accurate measure of both ROM
and stiffness. This equipment was used by seating the subject in a chair with the arm abducted to 90 degrees.
The weight of the arm was supported by a piece of PVC pipe (5" diameter) held in place by rollers on a post.
The subject's arm was inserted through the PVC, and secured by a bicycle tire tube that was inflated to secure
the humerus. The axis of the humerus was aligned with the center of a bicycle wheel and was horizontally
adducted 150 to align with the scapular plane. The forearm was flexed and the wrist strapped to a bar extending
from the perimeter of the bicycle wheel. A potentiometer (Model 536, Spectrol Electronics, City of Industry, CA)
was used to measure the rotation of the wheel. The wheel was manually rotated into maximum internal or
external rotation by a line connected to the perimeter of the wheel down through a pulley. A load cell (Model
SBO-3000, Transducer Technologies, Temecula, CA) was used to measure the force required to attain the ROM.
A strap also surrounded the shoulder in order to reduce rotation about the scapula. The data were amplified
before being processed by a Labview program running on a laptop computer.
Figure 1. Instrumentation
Figure 2. Subject in External Rotation
The duration of testing was approximately 45 minutes per subject. Subjects were given informed consent
paperwork, then measured for height and weight. Each subject was then seated and familiarized with the
procedure.The height of the arm support and bike wheel was adjusted to insure that the arm was abducted to
90 degrees and the axis of the humerus was in line with the center of the wheel. A horizontal bar from the
outer diameter of the wheel was then strapped to the wrist of the subject. For smaller subjects, the seat was
raised with supports for alignment purposes.Each subject was instructed to execute one practice rotation to
become comfortable with the position of maximal external ROM. Maximal ROM was determined by the subject as
the point of an "intense stretch without significant pain." The order of testing (external, internal dominant and
non-dominant) was randomized. The subject was then instructed to execute three recorded maximum
rotation repetitions. A repetition began with the forearm perpendicular to the ground. The arm was taken
to maximum rotation by way of a line tied to the outer rim of the bicycle wheel. The subject's arm was then
brought back to the starting position. Each subject then underwent external and internal maximum rotations for
both arms. During internal rotation, the shoulder was secured to the chair to reduce scapular motion. Then
the subject was instructed to warm-up as he or she normally would before a tennis match. Subjects
stretched slightly, executed several "dry" serves and finally hit several live serves until they reached "game
status." Subjects were then retested for all measures (internal and external rotation) for both arms.
The data collected corresponded to a series of three repetitions, all with synchronized force and ROM data. The
force data were multiplied by the distance from the bicycle wheel axis to the bar which was placed on the wrist.
This calculation yields torque, which is necessary to standardize forearm lengths between subjects. To
facilitate analysis, it was necessary to split the data into three separate repetitions. Once this was done, ROM
and torque were extrapolated for each repetition.
Due to difficulties with the amplifier and load cell, the data from 8 of the 22 subjects had to be dropped from
the study. In the process of comparing these remaining 14 subjects, several significant changes in maximum
external rotation emerged. For each measure (Max External ROM, Max Torque, Slope, and Onset) a two way
ANOVA was performed comparing dominant and non-dominant arms with pre and post warm-up repetitions.
To further analyze specific differences between arms, T-tests were used to compare dominant pre- versus
post-warm-up, non-dominant pre- versus post-warm-up, Dominant pre- versus non-dominant pre- and
dominant post- versus non-dominant post-. All data were taken from the third repetition from each trial.
Maximum External Rotation
As seen in Figure 3, the repeated-measures two way ANOVA comparing dominant and non-dominant arms with
pre and post warm-up maximum ranges of motion showed a significant difference between pre- and post- warm-
up of both arms, with post-warm-up having larger external ROM. No significant difference was shown
between dominant and non-dominant arms (Table 1).
When T-tests were run, dominant arm post-warm-up maximum external ROM were significantly larger than
pre-warm-up external ROM. Also, dominant arm pre-warm-up ROM were shown not to be significantly larger
than non-dominant ROM.Neither non-dominant pre- vs. post- ranges nor pre-warm-up dominant vs. non-
dominant ranges were shown to be significantly different.
The repeated-measures two-way ANOVA showed significantly larger dominant arm torques when compared to
the non-dominant arm.
T-tests revealed dominant arms to have undergone significantly more torque during the post-warm-up reps than
non-dominant arms. There were no significant differences shown between pre- versus post- repetitions of
either dominant or non-dominant arms.
Stiffness may be defined as the slope of the curve relating torque and rotation angle (Novotny, Wooley, Nichols,
& Beynnon, 2000). The two points designated to determine this slope are the angles corresponding to a torque of
1 and 2 Nm. Only limited amounts of data were available within this range, and no significant difference in
stiffness could be identified for any of the test conditions. Onset of stiffness was also calculated. This measure
was described as the angle at which 2 Nm of torque was reached. Onset also showed no significant
statistical differences in any of the testing areas.
Finally, to determine the possible changes associated with loosening within the three trial-repetitions, a t-test
was run to determine differences between the first and third repetitions within each trial. This test
showed significantly larger third repetition ROMs compared to first repetitions of that same trial.
Table la: Data Analysis
p Condition p
Maximum Angle (ï¿½) Onset (ï¿½)
(All) Dom vs. non-dom
p = 0.283 (All) Dom vs. non-dom p = 0.332
(All) Pre vs. post p = 0.038a (All) Pre vs. Post p = 0.095
Associations p = 0.259 Associations p = 0.215
Maximum Torque (Nm) Slope (ï¿½/Nm)
(All) Dom vs. non-dom p =0.04 a (All) Dom vs. non-dom p = 0.410
(All) Pre vs. post p = 0.387 (All) Pre vs. Post p = 0.320
Associations p = 0.637 Associations p = 0.289
Table 1b: Data Analysis
T-Tests; Mean (SD)
Type Pre Post p=
Maximum Angle (ï¿½)
Dom 107.08 (14.32) 114.65 (10.30) 0.041a
Non-Dom 105.53 (8.24) 109.38 (10.79) 0.145
p= 0.685 0.113
Dom 88.7 (17.1) 97.6 (12.407) 0.076
Non-Dom 88.2 (12.8) 91.97 (15.9) 0.280
p= 0.88 0.188
Maximum Torque (Nm)
Dom 7.4415 (4.66) 7.74 (4.04) 0.753
Non-Dom 4.95 (2.45) 5.67 (2.66) 0.093
p= 0.069 0.038a
Dom 6.53 (2.24) 13.59 (14.40) 0.304
Non-Dom 6.15 (2.61) 6.35 (1.35) 0.795
p= 0.814 0.354
Table 1c: Data Analysis
First to Third Rep Variability (ï¿½)
Average Number p
1st Rep Average 107.34 (13.2) p <.001
3rd Rep Average 110.43 (12.1)
*A forearm rotation of 0ï¿½ was designated as 90ï¿½ of shoulder abduction with the ulna/radius parallel to the
(a, Significant difference)
Change in Maximum ROM (pre and post warm-up)
. . _----
104 * Dominant Arm
102 * Non-dominant Arm
Figure 3. Changes in maximal external rotation from pre- to post-warm-up trials.
Though there were some difficulties encountered in the collection of data, several important points can be taken
from this study. With regards to maximal external rotation, the dominant arm showed a significant increase from
pre- to post-warm-up while the non-dominant arm did not. Though not significant, there was a change between
non-dominant arm averaging 105.50 before warming up and 109.00 after. Extensive "whole body" warm ups can
lead to increases in core muscle temperature, which increases and prolongs connective tissue and
musculotendinous extensibility (Lehmann, Masock, Warren, & Koblanski, 1970). In this study, the focus was not
the "whole body" warm-up, but rather stressed the changes due to the more specific tennis warm-up. This
tennis warm-up consisted of several light serves, followed by full speed serves until the subject felt to be at "in-
game status." Since the change associated with the non-dominant arm was not significant, the difference
in dominant arm ROM is most likely due to the stretching involved with the practice serves as opposed to any
whole-body warm-up that did occur.
There was no significant difference between dominant and non-dominant arms in pre-warm-up external
rotation. Research on this topic has been mixed, some showing that highly trained overhead athletes show
larger ranges of motion of the dominant arm (Young & Elliott 2001; Lehmann et al., 1970), and others showing
no difference (Nelson & Kokkonen 2001). Subjects in the studies noted, however, were professional, while in
the current study the subjects tested were of a much lower caliber.
The subjects tested consisted of a wide variety of individuals. Due to different pain tolerances and differences
in anatomy, the torque required to reach maximum ROM varied greatly. For this reason, the "onset" of stiffness
was also calculated. This onset was determined to be the angle at which two Nm were reached. This measurement
is important because it removes the factor of excess force being applied to the arm to reach large ROM. Measure
of onset closest to significance (p = 0.076) was the dominant arm's onset pre- to post-warm-up. This
difference, though not significant, is notable. Since the maximum torque, which correlates to the maximum
external ROM, was not significantly different pre- to post-warm-up, it should not have had a large effect on
the difference in maximum ROM. Therefore, barring differences in torque properties of the arm through the
ROM, differences in onset should mirror differences in maximal ROM. Wight et al. showed that the arm's
stiffness values do in fact stay relatively constant through the ROM (Chow et al., 2006). Large standard
deviations associated with the maximum torque values, however, may have complicated these results, as
some subjects did in fact experience larger maximal torques pre- to post-warm-up. The significant
difference between post-dominant and post-non-dominant maximum torque may have resulted from the
subjects' level of comfort in compromised arm positions. Subjects' dominant arms may be more accustomed
to extreme ROM as they are exposed to them when serving.
The slope data collected varied greatly within and between subjects. This measure was calculated by choosing
two points along the x-axis of the angle/torque graph, corresponding to one and 2 Nm. In short, this method
detects changes in distance traveled (in degrees) between two set torque measurements. Novotny et al. (2000),
also using torque as a measurement, designated one Nm as the "onset of stiffness" and measured from that point
to 4 Nm as the stiffness range. When comparing across subjects, standardization of forces by converting to torque
is necessary. For many of the subjects, inexperience, low pain tolerance, and hesitance to move through a full
ROM, caused extremely low maximum torque values. These problems limited the data to the smaller range of one
to 2 Nm. For follow-up studies, extensive familiarization would be necessary to ensure the subject is very
comfortable with the setup and arm motions.
In addition, to ensure comfort of the subject, a non-threatening apparatus is helpful. Subjects with their arm
in compromised situations under control of a machine may experience apprehension and tightening of the shoulder.
If available, EMG measurements on the rotator cuff muscles could show whether muscle stiffness would in fact
affect these measurements. As reported by Novotny et al. (2000), repeated testing may improve familiarization
with the apparatus.
The large changes in first to third repetitions within trials underscore the variability of ROM testing.
Though differences were quite varied, as a whole the third repetition was larger than the first by approximately
30. Some of the variance may be attributed to the subjects becoming more familiar with the motion and limits
of their shoulder ROM, though a trial repetition was incorporated into the procedure to eliminate this problem.
In future studies, more familiarization measures should be incorporated into the procedure. When studying
less experienced athletes who often are accustomed to extreme ROM, it is recommended that the subjects be
made as comfortable as possible with the apparatus, the motion, and the amount of stiffness in the joint that elicits
a thorough stretch. Also, additional adjustments should be made to the seating device to allow better measures
of small individuals.
A tennis-specific warm-up was shown to produce significant changes in external rotation about the shoulder.
No significant differences between pre- and post-non-dominant arms were shown, implying that the change was
due to the stretch/serving, as opposed to any whole-body warm-up that occurred. No significant differences in
the stiffness were shown as measured by slope of the torque vs. angle graph. A change in onset of stiffness (2
Nm) between pre- and post-warm-up on the dominant arm was shown, though not to a significant level (p =
0.076). Significant differences in the maximum torque used to bring subjects to their maximum angle were
also shown, but only between dominant and non-dominant arms. The difference in maximum torque within pre/
post repetitions of either arm was shown not to be significant.
The fact that there were indeed changes in ROM not due to whole body warm-up shows the importance
of standardizing warm-up procedure if ROM testing or analysis is to be completed. When testing subjects it is
crucial to know whether any warm-up has been completed, as results will be strongly influenced. Significant
changes can be seen within multiple repetitions of one measurement trial. Specifically when trainers or
other clinicians are assessing ROM for clearance to play, close attention must be paid to the number of
maximum stretch repetitions applied. Also, as shown by previous studies, increased ROM in the shoulder may
reduce injury when competing in overhead motion sports. This study shows the possible increase in ROM, and
further decrease in injury that could stem from thorough, sport specific stretching.
1. Borsa PA, Dover GC, Wilk KE, Reinold MM.Med Sci Sports Exerc. 2006 Jan;38(1):21-6. "Glenohumeral range
of motion and stiffness in professional baseball pitchers". Med Sci Sports Exerc. 2006 Jan;38(1):21-6.
2. Chandler TJ, Kibler WB, Uhl TL, Wooten B, Kiser A, Stone E. "Flexibility comparisons of junior elite tennis players
to other athletes". Am J Sports Med. 1990 Mar-Apr;18(2):134-6.
3. Chow JW, Wight JT, Grover GB, Tillman MD. "Exploring the Relationships Between Shoulder Flexibility and
Mechanics of the Tennis Serve". (Report Submitted to United States Tennis Association). Gainesville, FL.
University of Florida, Biomechanics Department, 2006.
4. Ellenbecker TS, Mattalino AJ. "Glenohumeral joint range of motion and rotator cuff strength following
arthroscopic anterior stabilization with thermal capsulorraphy". J Orthop Sports Phys Ther. 1999 Mar;29(3): 160-7.
5. Ellenbecker T, Roetert EP. "Age specific isokinetic glenohumeral internal and external rotation strength in elite
junior tennis players." J Sci Med Sport. 2003 Mar;6(1):63-70.
6. Ellenbecker TS, Roetert EP, Bailie DS, Davies GJ, Brown SW. "Glenohumeral joint total rotation range of motion
in elite tennis players and baseball pitchers". Med Sci Sports Exerc. 2002 Dec;34(12):2052-6.
7. Fowles JR, Sale DG, MacDougall JD. "Reduced strength after passive stretch of the human plantarflexors", Journal
of Applied Physiology, Vol. 89, Issue 3, 1179-1188, September 2000.
8. Hubley C, Kozey J, Stanish W. "The effect of static stretching exercises and stationary cycling on range of motion
at the hip joint". J Orthop Sports Phys Ther. 1984;6:104-109
9. Kibler WB, Chandler TJ, Livingston BP, Roetert EP. "Shoulder range of motion in elite tennis players. Effect of age
and years of tournament play". Am J Sports Med. 1996 May-Jun;24(3):279-85.
10. Lehmann JF, Masock AJ, Warren CG, Koblanski JN. "Effect of therapeutic temperatures on tendon extensibility".
Arch phys med Rehabil. 1970; 51 :481-87.
11. Nelson AG, Kokkonen J. "Acute ballistic muscle stretching inhibits maximal strength performance" Research
Quarterly for Exercise and Sport, December 01, 2001.
12. Novotny JE, Woolley CT, Nichols CE 3rd, Beynnon BD. "In vivo technique to quantify the internal-external
rotation kinematics of the human glenohumeral joint." J Orthop Res. 2000 Mar;18(2):190-4.
13. Schmidt-Wiethoff R, Rapp W, Mauch F, Schneider T, Appell HJ. "Shoulder rotation Characteristics in
professional tennis players." Int J Sports Med. 2004 Feb;25(2):154-8.3)
14. Taylor BF, Waring CA, Brashear TA. "The effects of therapeutic application of heat or cold followed by static stretch
on hamstring muscle length". J Orthop Sports Phys Ther. 1995;21:283-286.
15. Weijer VC, Gorniak GC, Shamus E. "The effect of static stretch and warm-up exercise on hamstring length over
the course of 24 hours". J Orthop Sports Phys Ther. 2003 Dec;33(12):727-33
16. Wessling KC, Devane DA, Hylton CR. "Effects of static stretch versus static stretch and ultrasound combined
on triceps surae muscle extensibility in healthy woman". Phys Ther. 1987; 67:674-679
17. illiford HN, East JB, Smith FH, Burry LA. "Evaluation of warm-up for improvement in flexibility". Am J Sports
Med. 1997; 14:316-319.
18. Young W, Elliott S. "Acute effects of static stretching, proprioceptive neuromuscular Facilitation stretching,
and maximum voluntary contractions on expolsive force production and jumping performance". Research Quarterly
for Exercise and Sport, September 01, 2001.
Back to the Journal of Undergraduate Research
College of Liberal Arts and Sciences I University Scholars Program I University of Florida I| UNIVERSITY of
SUniversity of FLORIDAidaGainesvilleFL32611;(352)846-2032.
@ University of Florida, Gainesville, FL 32611; (352) 846-2032.