Effects of Strength Training on Throwing
Velocity and Shoulder Muscle Performance in
Teenage Baseball Players
Michael I. Wooden, MS, PT, OCS'
Bruce Greenfield, MMSc, PT2
Marie lohanson, MS, PT3
lisa litzelrnan, MA, ATC4
Mary Mundrane, MPT, ATCS
Robert A. Donatelli, MA, PT, 0CS6
F
o r over 20 years, the literature has demonstrated
widespread use of isokinetics, both for nieasurement of strength (1, 5-7,
1.5, 1 7, 19, 22, 24) and for exercise
to improve torque production of various niuscle groups (3, 10, 1 1, 22,
24). lsokinetic (IKN) exercise, in
which the speed is constant, is
thought to be totally accommodating
to the individual because the resistance varies at each point in the
range of motion (ROM) t o match the
force applied (2, 16). T h e major advantage of this fixed-speed, variable
resistance exercise is that maximum
resistance can be applied throughout
the ROIM, provided the effort is
maximum (2, 14). Other advantages
are that I KN exercise accommodates
to pain and fatigue and that speeds
can be increased o r decreased depending on clinical needs (2).
Isokinetic exercise has been
shown t o be an effective means of
increasing "strength" o r torque output. Lesme et al (1 1) studied the effects of short (6-second) and long
(SO-second) burst exercise a t 18 0 ° /
sec on the knee extensors of both
limbs of five healthy males. After 7
JOSPT
\'ohme 15 Sumber .
i\lay 1 902
Exercise protocols designed to improve muscle function and athletic performance are continually developed and revised, often without published research supporting their efficacy. This study
compared the effects of isokinetic (IKN) and accommodative isotonic training in the individualized,
dynamic, variable resistance (IDVR) mode. Twenty-seven teenage baseball players were tested
isokinetically for dominant shoulder rotational peak torque and power and for throwing velocity.
They were then randomly assigned to 5 weeks of IKN training, IDVR training, or a control group of
no training. Following the training period, pretest protocols were repeated. Analysis o i variance of
differences in means and Newman-Keuls post hoc tests showed statistically significant increases in
throwing velocity and external rotator torque in the IDVR group but not the IKN group. External
rotator power improved in both groups. Internal rotator torque and power were not improved in
either group. Results suggest that IDVR may be more effective than IKN training in improving
throwing velocity and external rotator torque production. Clinicians should consider using IDVR
protocols in improving shoulder muscle function and throwing performance.
Key Words: throwing velocity, shoulder, muscle strength training
'Orthopaedic clinical specialist, Physical Therapy Associates of Metro Atlanta; instructor, Division of Physical
Therapy, Department of Rehabilitation Medicine, Emory University, Atlanta, CA
'Physical Therapy Associates of Metro Atlanta; clinical instructor, Division of Physical Therapy, Emory University, Atlanta, GA
'Physical therapist, Physical Therapy Associates of Metro Atlanta, Atlanta, CA
'Athletic trainer, Physical Therapy Associates of Metro Atlanta, Atlanta, CA
'Head athletic trainer, Physical Therapy Associates of Metro Atlanta, Atlanta, GA
'Orthopaedic clinical specialist, Physical Therapy Associates of Metro Atlanta; instructor, Division of Physical
Therapy, Department oi Rehabilitation Medicine, Emory University, Atlanta, CA
weeks of training, both limbs iniproved significantly in peak torque,
work output, and endurance. In a
controlled study of the knee extensors of 16 healthy females, Peterson
et al (22) found increases in concentric and eccentric torque production
after 6 weeks of IKN training. Klingenstierna et al (10) studied the effects of multiple speed IKN training
on the knee flexor and extensor
muscles of eight adult, below-knee
unilateral amputees. Significant increases in torque production were
RESEARCH S T U D Y
noted at most speeds in the amputated and uninvolved linibs. Smith
and Melton (24) compared the effects of low- and high-speed IKN exercise to variable resistance isotonic
(ITN) exercise (using Nautilus e q u i p
nient) on quadriceps i111d hamstrings
strength as well as various functional
tasks: vertical jump, standing broad
jump, and the 40-yard dash. In the
12 adolescent boys studied, all
groups improved in strength measured isokinetically, while the liiglispeed IKN group had the greatest
improvements in all functional tests.
In a study of 32 collegiate tennis
players, Ellenbecker et a1 (3) conipared the effects of concentric IKN
versus eccentric IKN exercise on rotator cuff strength and power and on
tennis serve velocity measured by
video analysis. After 6 weeks of
training, both groups improved significantly in concentric strength.
However, only the concentric training group improved in both eccentric strength and serve velocity.
Ironically, the advantage of fixed
velocity may also be somewhat of a
disadvantage when using IKN training to improve functional activities
o r athletic performance. Despite the
successes reported in previous literature (3, 24). most IKN devices allow
maximum velocities of 500°/sec.
well below those required for most
activities (1 2). For example, the
shoulder joint must accelerate to iln
angular velocity of more than
6,000°/sec when throwing a baseball
(1 8). T h e high angular velocities that
occur with throwing pose a question
as to whether training isokinetically
inhibits limb acceleration and, possibly, optimum improvement in
torque production and performance.
A recent advance in resistive exercise is offered by the Musculoskeletal Evaluation Rehabilitation and
Conditioning (MERAC, Universal
Equipment Co., Cedar Rapids, 1.4)
dynamometer. In addition to isokinetics, the MERAC also provides the
option of acconimodative ITN exer-
cise. This individualixed, dynamic.
variable resistance (I D\'R) mode
works as follows: from a series of
maximnni effort isokinetic contractions, the MERAC computer ilverages the data into a motor perforniance curve (MPC) as shown in Figure 1. T h e system is then programmed for exercise to provide resistance based on a selected percentage of the torque recorded at every
point in the ROM on the curve.
Maximum resistance is provided by
programming 100 percent of the
MPC. This resistance is ITN since it
is no longer speed-constant: however, it is accommodative since the
resistance changes constantly to
match the MPC. With no preset
speed, the limb can accelerate
against the resistance according to
the effort of the individual.
One week aher
completion of training,
the pretest protocols
were repeated.
T h e authors' clinical experience
with IDVR strength training on a variety of patients Iias been positive,
but no known published, controlled
studies demonstrating its efficacy exist. T h e purpose of this study was to
deterniine whether IKN o r lDVR
training would be more effective in
increasing rotator cuff muscle torque
production and power, as well as improving the velocity of throwing ;I
baseball.
METHODS
Subjects were 27 volunteers.
ages 14- 17 (mean 15.5 f .97), who
were junior and senior high sckool
baseball players of various positions.
Average weight was 155.2 Ibs (f
26.0). Twenty-five were right-hand
dominant; two were left handed. In
the previous 6 months, no one reported a history of upper extremity
in-jury o r pain that rendered him unable to compete o r that required examination by a physician. Prior to
participation, testing and training
procedures were explained, and informed consent was obtained from
the subjects' parents. Subjects were
screened with shoulder, elbow, and
wrist ROM tests in all planes and
with shoi~lderlocking and quadrant
tests.
Pretest
Throwing velocity was tested a p
proxiniately 1 hour prior to muscle
torque and power testing in all subjects to reduce the chance of fatigue
influencing the muscle testing. T h e
pretests were not randomized because the gym space and radar
equipment needed for velocity testing were available for a limited time.
In il nlethod modified from Pawlowski et al (1 9), velocitv was tested
indoors to control for weather effects, using a standard pitching distance of 60 ft, 6 in. A pitcher's
mound was not used. Subjects were
instructed to use a wind-up motion
and were allowed as many warm-up
throws as they felt necessary for
maximum effort throwing. Follow*
ing a 2-minute rest, five maximum
effbrt throws were measured in miles
per hour (niph) with a calibrated
Magnum X Ran radar gun (CMI
Corporation, Owensburg, KY). T h e
radar gun was mounted at a height
of 36 in and positioned to the right
of the catcher. T h e mean of the five
throws was recorded. Reliability was
insured by calculating percent of
agreement for all throws in the pretest and posttest. Of the throws, 92.2
percent were within 2 mph of the
means, and 98.9 percent were within
3 rnph.
Dominant shoulder internal and
external rotation torque (ft-lbs) and
power (W) were tested isokinetically
Volume I3 Sumber 3 !day 1992 JOSPT
lsokinetic Graph
40
3530
-
Int.
Rot.
2520
.
Ext.
Rot.
-
Ext
Rot.
Ext.
Rot.
f
15-
I
105
-
\
0
I
I
I
the MERAC table. A plastic overlay,
cut at 30" angles, was placed on the
graduated center line of the template. T h e shoulder joint was positioned in 90" abduction and verified
with a standard goniometer, and the
arm was strapped to the dynamometer attachment. A plumb line was
dropped from the shoulder joint axis
to the center line of the template,
and the apex of the plastic overlay
was located at this point. T h e chair
was then turned t o be aligned with
the 30" line on the overlay. Once
positioned in 90" abduction and 30"
in the POS, the long axis of the upper arm was visualized to be continuous with the dynamometer axis to
ensure that movements tested would
occur around the rotational axis of
Position i n D e g r e e s
IDVR C u r v e Graph
(notor Performvlce Curve)
501
351
8
int.
Rot.
Position in Degrees
FIGURE 1. A) lsokinetic graph, internal and external rotation, tested at SOO0/sec;B) Motor performance
curve from the IDVR graph.
at 500°/sec with a MERAC dynamometer. T h e shoulder was positioned in 90" abduction and 30"
horizontal adduction in the plane of
the scapula (POS), using a procedure
modified from Freedman and
JOSPT Volume 15 Number 5 May 1992
Munro (4) and Greenfield et al (5).
T o minimize body movements, subjects were seated and stabilized in a
MERAC accessories chair. T h e accessory chair was positioned on a
floor template located at the end of
The individualized,
dynamic, variable
resistance training
showed significant
increases in throwing
velocity.
the shoulder joint. T h e chair was
then locked into place (Figure 2).
Dynamometer height, resistance
handle length, and chair position on
the template were recorded for use
in the training and posttest sessions.
Torque was calibrated to baseline prior to each test. Reliability and
validity of the MERAC for torque
measurement, angular velocity, and
shoulder test-retest positioning have
been established previously (5, 13).
Stops were programmed into the
computer to limit the arc of motion
to 90". As a warm-up, subjects performed seven submaximal effort
contractions followed by three maximal effort contractions. After a 90second rest, four maximal effort contractions were recorded. Data for
RESEARCH S T U D Y
of 1 0 repetitions was added each
week until 1 0 sets of 1 0 were performed in week five.
Posttest
O n e week after completion of
training, the pretest protocols were
repeated. Subjects were positioned
and tested by the same investigators
as in the pretest.
DATA ANALYSIS
FIGURE 2. Testing and training position tor shoulder
rotation.
peak torque (PT), peak torque to
body weight ratio (PT:BW), and
power were saved on the MERAC
computer (Figure 1A). Another 90second rest was followed by four
more maximal effort contractions t o
obtain the MPC (Figure 1B). T h e ord e r of torque and MPC testing was
randomized by a coin toss.
In order t o reduce the effect of
size in torque output, peak torque
values were normalized by expressing them as a percentage of body
weight, providing PT:BW ratios (5,
6, 16, 23). To ensure that assignment to training groups had been
well randomized, one-way analysis of
variance (ANOVA) of the pretest
means of velocity, PT:BW, a n d
power was performed. After training, one-way ANOVA of the differences between the pretest and post-
Measure
Training
G~OUP
test means of each measure was
used. Newman-Keuls post hoc multiple comparison tests were then performed t o determine the effects of
the IKN and IDVR training conditions. Minimum level of significance
was established at p < .05.
RESULTS
Analysis of variance of the pretest measurements of throwing velocity, PT:BW, a n d power showed
n o significant differences among the
IKN, IDVR, and control groups. Table 1 lists the effects of the type of
training on the various measurement
parameters.
Throwing Velocity
Analysis of variance of the
change in mean throwing velocity
demonstrated that statistically significant differences between the groups
occurred [F(2,24)= 7.23, p < .05].
Newman-Keuls post hoc tests showed
Change Mean
(SD)
Overall
~ ~ 2 4 )
Overall
P
Velocity (MPH)
IDVR'
IKNn
Training Sessions
Subjects were randomlv assigned
t o IKN training, IDVR training, or a
control group of n o training. Sessions were held between the fall a n d
spring baseball seasons a t regular intervals, three times a week for 5
weeks, for a total of 1 5 sessions. T w o
subjects missed o n e session, all others completed 15. For training, the
subjects were placed in the pretest
position and the MERAC was programmed for either IKN or IDVR
resistance. T h e IKN group exercised
at 500°/sec. T h e IDC'R group exercised at 100 percent of the variable
resistance provided by the pretest
MPC, with n o preset velocity. In
week 1, subjects completed six sets
of 10 maximal effort repetitions with
a 1-minute rest between sets. O n e set
Control
PT:BWm internal rotation
(It-1bs:lbs)
IDVR
IKN
Control
PT:BW external rotation
(It-1bs:lbs)
IDVR
IKN
Control
Power internal rotation (W)
IDVR
IKN
Control
Power external rotation (W)
IDVR
IKN
Control
Individualized, dvnamic, variable resistance mode
" lsokinetic mode
"Peak torque to body weight ratio
TABLE 1. Effect of type of training on throwing velocity, shoulder rotator PT:BW, and power.
Volume 15 Number 5 May 1992 JOSPT
RESEARCH STUDY
IDVR velocity increases t o be significantly greater than the IKN ( p <
.05) and control ( p < .0 1 ) groups.
T h e IKN group was not significantly
greater than the control.
Internal Rotation of PRBW
Analysis of variance of the
change in mean PT:BW showed that
a statistically significant difference
between the groups existed [F(2,24)
= 5.00, p < .05]. Post hoc tests demonstrated IDVR mean increases t o
be significantly greater than the IKN
group ( p < .05), but that neither the
IDVR nor IKN groups were significantly greater than the control
group.
Internal Rotation Power
Analysis of variance of the
change in mean power showed no
significant differences between any
of the groups [F(2,24)= 1.39, p >
.05].
External Rotation PRBW
Analysis of variance of the
change in mean PT:BW demonstrated that a statistically significant
difference between groups existed
[F(2,24)= 6.92, p < .01]. Post hoc
tests showed changes in the IDVR
group to be significantly greater
than the IKN ( p < .05) and control
( p < . O l ) groups. T h e IKN group
was not significantly greater than the
control.
External Rotation Power
Analysis of variance of the
change in mean power demonstrated
that a statistically significant difference between groups existed
[F(2,24)= 9.8 1 , p < .001]. Post hoc
tests showed that mean power increases for both the IDVR and IKN
groups were significantly greater
than the control group.
JOSPT * Volume 15 * Number 5 * May 1992
DISCUSSION
T h e results suggest that the
IDVR training mode was more effective than IKN exercise for improving throwing velocity and external
rotator PT:BW. These findings differ from those of Smith and Melton
(24).who found IKN exercise to be
more advantageous than I T N training on Nautilus equipment, which is
at least partially accommodating.
Both IDVR and IKN exercises were
effective in increasing external rotator power. A surprising finding of
this study was that internal rotator
PT:BW and power did not improve
significantly in either condition. This
could have been due to the testing
and training position of 90" abduction and 30" in the POS. Although
the 90" abduction component simulates throwing, previous studies have
shown decreases in internal rotator
torque output ( 6 )and subscapularis
electromyographic activity ( 2 1 ) with
progressive degrees of shoulder abduction. Conversely, the POS component may have enhanced the mechanical advantage and length-tension relationship of the external
rotators ( 4 , 6 , 21). Abduction of the
shoulder in the POS does not fully
simulate a throwing motion. However, the POS was used in the present study because of previous work
by Greenfield et al ( 1 5) indicating
that significantly greater external rotator torque was produced in the
POS compared to the frontal plane.
Internal rotator torque output was
the same in both planes.
T h e IDVR training group
showed significant increases in
throwing velocity, which may have
been due, in part, to improvements
in external rotator PT:BW and
power. Pedegana et al ( 2 0 ) reported
a relationship between throwing velocity and strength of the external
rotators, which are active concentrically in cocking the shoulder and eccentrically in controlling arm movement during the follow-through
phase ( 8 , 9 ) . Increases in external rotator concentric torque production
would seem also to have improved
eccentric function, a concept s u p
ported by previous studies ( 3 , 22).
Similar to the findings in the present
study, Pedegana et al found no significant relationship between internal rotator strength and throwing
velocity (20).A more important factor in improvements seen in the
IDVR groups seemed to be the ability t o accelerate the arm freely, simulating the activity of throwing. T h e
IKN group could not accelerate beyond .iOOO/sec, far below the angular velocity required for throwing
(18). Using IDVR, with no fixed
speed, subjects were able to accelerate according to their abilities.
For future study, the authors
recommend also testing isotonically
and recording speeds of contraction
to evaluate the speed factor. Narrowing the age group studied should
be considered, since 13- and 17-yearold boys may respond differently to
training effects.
CONCLUSIONS
Results suggest that the IDVR
mode may be more effective than
the IKN mode in improving throwing velocity and shoulder external
rotator torque production. This may
be the result of IDVR exercise allowing the arm to accelerate more freely
during a maximal effort contraction.
T h e lack of improvement of internal
rotator torque production may have
been due to the testing and training
position used, which put the external
rotators at more of a mechanical advantage. Clinicians should consider
using IDVR protocols, but they
should use different positions of
shoulder abduction to attempt to involve all muscle groups.
lOSP-r'
ACKNOWLEDGMENTS
T h e authors gratefully acknowledge Becky Eisman, PT, for assist2?7
RESEARCH STUDY
ance w i t h t h e t r a i n i n g sessions and
Steve Cole, P h D , f o r t h e statistical
analysis.
REFERENCES
I . Cook, EE, Gray VL, Savinar-Noque E,
Medeiros I: Shoulder antagonistic
strength ratios: A comparison between
collegiate level baseball pitchers and
nonpitchers. I Orthop Sports Phys Ther
8:451-461, 1987
2. Davies C: A Compendium of Isokinetics in Clinical Usage: Workshop and
Clinical Notes, Lacrosse, WI: S and S
Publishers, 1 984
3. Ellenbecker T, Davies C, Rowinski Ml:
Concentric versus eccentric isokinetic
strengthening of the rotator cuff. Objective versus functional test. Am I
Sports Med 16:64-69, 1988
4. Freedman L, Munro RR: Abduction of
the arm in the scapular plane: Scapular
and glenohumeral movements. I Bone
joint Surg 48A: 1503- 15 10, 1966
5. Greenfield B, Donatelli R, Wooden MI,
Wilkes I: lsokinetic evaluation of
shoulder rotational strength between
the plane of the scapula and the frontal
plane. Am I Sports Med 18:124- 128,
1990
6. Hinton RY: lsokinetic evaluation of
shoulder rotational strength in high
school baseball pitchers. Am I Sports
Med 16:274-279, 1988
7. lvey FM, Calhoun lH, Rusche K, Bierschenk I: lsokinetic testing of shoulder
8.
9.
10.
1 I.
12.
13.
14.
15.
16.
strength: Normal values. Arch Phys
Med Rehabil66:384-386, 1985
lobe FW, Moynes DR, Tibone IE, Perry
I: An EMC analysis of the shoulder in
throwing and pitching: A second report. Am ) Sports Med 12:218-220,
1984.
lobe FW, Tibone IE, Perry I, Moynes
D: An EMC analysis of the shoulder in
throwing and pitching. A preliminary
report. Am I Sports Med 1 1:3-5, 1983
Klingenstierna U, Renstrom P, Crimby
C, Morelli B: lsokinetic strength training in below knee amputees. Scand I
Rehabil22:39-43, 199 1
Lesme C, Costil E, Coyle E, Fink W:
Muscle strength and power changes
during maximal isokinetic training.
Med Sci Sports 10:266-269, 1978
Malone TR (ed): Evaluation of Isokinetic Equipment, pp 2-4. Baltimore:
Williams and Wilkins, 1988
Martin PE: Validation of torque measurement systems. An engineering
study conducted for Universal Equipment Co. Unpublished. Department of
Health and Physical Education, Arizona State University, Tempe, Az,
1987
Moffroid M, Whipple R, Hofkosh I,
Lowman E, Thistle H: A study of isokinetic exercise. Phys Ther 49:735747, 1969
Murray MP, Core DR, Cardner CM,
Mollinger LA: Shoulder motion and
muscle strength of normal men and
women in two age groups. Clin Orthor
192:269-273, 1985
Nicholas I, Robinson LR, Logan A,
17.
18.
19.
20.
2 1.
22.
23.
24.
Robertson R: lsokinetic testing in
young nonathletic able-bodied subjects. Arch Phys Med Rehabil 70:210213, 1985
Otis IC, Warren RF, Backus 51, Santer
TI, Mabrey ID: Torque production in
the shoulder of the young adult male.
Am I Sports Med 18: 1 19- 123, 1990
Pappas AM, Zawacki RM, Sullivan TI:
Biomechanics of baseball pitching. A
preliminary report. Am I Sports Med
l3:2 16-22,?, 1985.
Pawloski D, Perrin D: Relationship between shoulder and elbow isokinetic
peak torque, torque acceleration energy, average power, total work and
throwing velocity in intercollegiate
pitchers. Athl Train 24:129-13 1, 1989
Pedegana LR, Elsner RC, Roberts D,
Lang I, Farewell V: The relationship of
upper extremity strength to throwing
speed. Am I Sports Med 10:352-354,
1982
Perry I: Biomechanics of the Shoulder.
In: Rowe CR (ed), The Shoulder, New
York: Churchill Livingstone, 1988
Peterson S, Wessell I, Ragnall K, Wilkins H, Quinney A, Wegner H: Influence of concentric resistance training
on concentric and eccentric strength.
Arch Phys Med Rehabil 71:101-105,
1990
Schlinkman B: Norms for high school
football players derived from Cybex
data reduction computer. I Orthop
Sports Phys Ther 5:243-245, 1984
Smith MI, Melton P: lsokinetic versus
isotonic variable resistance training.
Am I Sports Med 9:275-278, 198 1
Volume 15 Number 5 May 1992 JOSPT