AJSM PreView, published on March 1, 2012 as doi:10.

1177/0363546512436801

Ultrasonographic Assessment of Humeral

Retrotorsion in Baseball Players
A Validation Study
Joseph B. Myers,*yz§ PhD, ATC, Sakiko Oyama,y§ MS, ATC, and John P. Clarke,|| MD
Investigation performed at the University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina

Background: Recently, clinicians and researchers started using diagnostic ultrasound to measure humeral torsion as an alterna-
tive to radiological assessments in overhead-throwing athletes and other clinical populations. Ultrasound appears to be a reliable,
nonradiological alternative, but ultrasound assessment has not been validated against computed tomography (CT), the current
gold standard.
Purpose: This study aimed to establish the validity of an ultrasound assessment of humeral torsion by comparing data obtained
using both ultrasound and CT, the current standard.
Study Design: Cohort study (Diagnosis); Level of evidence, 2.
Methods: Humeral torsion was measured using both ultrasound and CT in 24 collegiate-aged baseball players. Comparisons
between the assessments were made using both regression and Bland-Altman plots. Reliability and precision were also
established.
Results: A strong relationship existed between humeral torsion variables obtained with ultrasound and CT (R = .797, R2 = .635, P
\ .001). The ultrasound assessment yielded reliability coefficients ranging from .991 to .997, with approximately 1° of measure-
ment error. The CT method’s reliability coefficients ranged from .805 to .933, with approximately 3.5° of measurement error.
Conclusion: There was a strong relationship between humeral torsion obtained with ultrasound and CT, the current standard of
assessment. Ultrasound provides a reliable, valid alternative to CT for obtaining an indication of the amount of humeral torsion in
the upper extremity.
Keywords: humeral torsion; ultrasound; computed tomography; baseball

Humeral torsion represents the amount that the distal
humerus is twisted relative to the proximal humerus. The
dominant limb of throwing athletes has repeatedly been
*Address correspondence to Joseph B. Myers, PhD, ATC, University
of North Carolina at Chapel Hill, Department of Exercise and Sport Sci-
ence, CB# 8700, Fetzer Hall, Chapel Hill, NC 27599-8700 e-mail:
joemyers@.unc.edu
y
Sports Medicine Laboratory and Neuromuscular Research Labora-
tory, Department of Exercise and Sport Science, University of North Car-
olina at Chapel Hill, Chapel Hill, North Carolina.
z
Department of Orthopaedics, University of North Carolina at Chapel
Hill, Chapel Hill, North Carolina.
§
Human Movement Science Doctoral Curriculum, Department of
Allied Health Sciences, School of Medicine, University of North Carolina
at Chapel Hill, Chapel Hill, North Carolina.
||
Department of Radiology, University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina.
One or more of the authors has declared the following potential con-
flict of interest or source of funding: This study was funded by the Univer-
sity of North Carolina Translational and Clinical Science Institute and the
National Center for Research Resources (Award Number UL1RR025747).
The American Journal of Sports Medicine, Vol. XX, No. X
DOI: 10.1177/0363546512436801
Ó 2012 The Author(s)
shown to have more humeral torsion than their nondomi-
nant limb as well as the limbs of non–overhead-throwing
athletes.4,10-12,15 This increased humeral torsion has been
linked to upper extremity injuries including throwing-
related shoulder and elbow injuries in overhead-throwing
athletes9,17 and has been demonstrated to largely influence
glenohumeral rotation range of motion.3,8,12
Methods for measuring humeral torsion include the use
of radiographs, magnetic resonance imaging (MRI), com-
puted tomography (CT), and ultrasonography, with the
CT methodology currently considered the gold standard.{
Yet, radiological measures like radiography and CT, as
well as MRI, pose some methodological difficulty that
makes them less useful in both the clinical and research
setting. Those barriers include cost, radiation exposure,
time, and difficulty obtaining accurate measurements.
Diagnostic ultrasound has been shown to be a potential
alterative to radiological/MRI measures, making it a tool
that can be utilized for quick, accurate screenings in the
clinical setting.8,9,13,15-19 Unlike radiological/MRI methods,
{
References 2, 3, 5, 6, 8-12, 15-17, 19.

1
2 Myers et al
which directly measure torsion from angles derived from
both the proximal and distal humerus, the ultrasound tech-
niques described in the literature indirectly measure
humeral torsion by calculating the forearm inclination
angle relative to a standardized humeral position.8,9,13,15-19
Ultrasound measurements of humeral torsion have reliably
produced results that mimic the findings obtained using
radiographic measures in the literature.8,9,13,15-17 But to
date, a direct comparison of the ultrasound measurement
with the CT gold standard has not been conducted. This
level of validity is necessary before it can be concluded
that the ultrasound assessment is a viable alternative.
The purpose of the current study was to compare the
amount of humeral torsion obtained through the ultrasound
measurement with the CT measurement in a cohort of
college-aged baseball players.
MATERIALS AND METHODS
Participants
Twelve male collegiate-aged baseball players (age, 21.8 6
1.5 years; height, 185.2 6 9.5 cm; weight, 91 6 13.2 kg)
participated in the current study. A combination of both
pitchers and position players with varying injury histories
were enrolled to ensure that sufficient variability existed
across the 24 upper extremities assessed (bilateral assess-
ment in the 12 players) for the analyses performed. All par-
ticipants were healthy at the time of testing.
Procedures
All participants attended one testing session, during which
humeral torsion was assessed using both ultrasound and
CT measurements. Before participation, all participants
read and signed the informed consent forms approved by
the University of North Carolina at Chapel Hill Institu-
tional Review Board.
Humeral torsion was assessed using the indirect ultra-
sound technique described by Myers et al,8 Whiteley
et al,15 and Yamamoto et al.19 Before the current study,
the investigators established construct validity, reliability,
and precision of the torsion assessment, yielding intrases-
sion, intersession, and intertester reliability coefficients
ranging from .96 to .98, with an average of 2.3° of measure-
ment error.8 Participants lay supine on a treatment table
with 90° of shoulder abduction and elbow flexion. A tester
positioned a 4-cm linear array ultrasound transducer (LOG-
IQe, General Electric, Milwaukee, Wisconsin) on the partic-
ipant’s anterior shoulder with the ultrasound transducer
level with the plane of the treatment table (verified with
a bubble level) and aligned perpendicular to the long axis
of the humerus in the frontal plane. The second tester
then rotated the humerus so that the bicipital groove
appeared in the center of the ultrasound image, with the
line connecting the apexes of the greater and lesser tubercles
parallel to the horizontal plane. A grid was applied to the
display of the ultrasound unit to aid examiners with posi-
tioning of the humeral tubercles. The second tester then
The American Journal of Sports Medicine
Figure 1. (A) Ultrasonographic assessment of humeral tor-
sion. (B) Ultrasonographic image of the proximal humerus
where the apexes of the humeral tubercles are aligned
horizontally.
placed a digital inclinometer on the ulnar side of the fore-
arm, pressing firmly against the ulna, and recorded the fore-
arm inclination angle with respect to the horizontal (Figure
1). Because the ulna extends perpendicular to the elbow epi-
condylar axis (line connecting the medial and lateral epicon-
dyles), this angle reflects the angular difference between
the epicondylar axis (distal humerus) and the line perpendic-
ular to the line connecting the apexes of the greater and
lesser tubercles (proximal humerus), thus representing an
ultrasound-assisted approximation of humeral torsion.
Three trials were performed bilaterally by 2 separate asses-
sors, both of whom have extensive experience in the ultra-
sonography measurement technique. Each assessor was
blinded to the results obtained by the other assessor, allow-
ing for intertester reliability (intraclass correlation coeffi-
cients) and precision (standard error of measurement) to
be established for the data collected in the current study.
Previous work by the investigators has demonstrated that
ultrasonography assessment has high intrarater, interrater,
and intersession reliability. Therefore, the mean of all
6 trials (3 from each assessor) was calculated to represent
the humeral torsion angle for each limb.
Either before or after the ultrasound measurement
(depending on the availability of the CT scanner), each par-
ticipant was escorted from a sports medicine research lab-
oratory to the radiology department of a university medical
center. In all cases, the CT and ultrasound measurements
were performed within 1.5 hours of each other. Computed
tomography was conducted on a Siemens Somatom Defini-
tion AS (Siemens Medical Solutions USA Inc, Malvern,
Pennsylvania) CT scanner. In most cases, the CT image
was obtained with the participants’ arms placed at their
sides. In 4 cases, the larger size of the participants dictated
that the image be obtained with the arm in full elevation.
In all cases, the limb was oriented parallel to the scanning
table, confirmed with the laser sighting provided by the CT
unit. Two-millimeter transverse plane slices were obtained
for the entire proximal humerus (from the acromion, distal
to the surgical neck) and distal humerus (from the distal
shaft, proximal to the supracondylar ridges, distal to the
elbow where the radius and ulna could be appreciated).
Scanning these regions ensured that we obtained the nec-
essary imaging to calculate the humeral torsion angle and
Vol. XX, No. X, XXXX
Figure 2. Calculation of humeral torsion (u) from computed
tomography scans obtained of both the proximal (A) and dis-
tal (B) humerus. (A) Dashed line: proximal humeral axis. Solid
line: line perpendicular to the proximal humeral axis. (B) Solid
line: distal humeral axis. Dashed line: transpose of the line
perpendicular to the proximal humeral axis from A.
the bicipital forearm angle from the CT images, while min-
imizing participant’s radiation exposure.
To calculate humeral torsion from the series of CT
images, 2 assessors independently selected the proximal
and distal humeral slices that each thought best portrayed
the necessary landmarks to calculate the humeral torsion
angle. To ensure that the most appropriate slice was
included for analysis, the chosen slice and the slice imme-
diately above and below were used to calculate the
humeral torsion angle for each limb. The humeral torsion
angle was calculated as the angle between the proximal
and the distal humeral axes (Figure 2). The proximal
humeral axis was established by identifying a line extend-
ing from one edge of the humeral articular surface to the
other and then drawing a line perpendicular to the line
(Figure 2A). The distal humeral axis was defined as
a line connecting the most prominent points on the
humeral epicondyles (Figure 2B). The mean of the humeral
torsion angle from the 3 series of slices represented the
humeral torsion angle obtained by each assessor for each
limb. Each assessor independently calculated the humeral
torsion angles from the obtained CT scans, allowing for the
calculation of intertester reliability and precision.
As a secondary analysis, to compare the humeral torsion
obtained with the ultrasound measurement and CT scans
using the equivalent anatomic landmarks, the bicipital
forearm angle was calculated from the CT images. The
bicipital forearm angle was defined as the angle between
a line that connected the apexes of the greater and lesser
tubercles (on the slice when the tubercles were most prom-
inent) and a line that was perpendicular to the distal
humeral axis (to represent the ulna) (Figure 3). All CT
image processing was conducted using ImageJ Imaging
Processing and Analysis software (National Institutes of
Health, Bethesda, Maryland).
Data Analysis
Agreement between the ultrasound and CT humeral tor-
sion variables was determined through both linear regres-
sions (with accompanying Pearson correlation coefficients)
and interpretation of the Bland-Altman plots. All limbs
Ultrasound of Humeral Retrotorsion in Baseball Players 3
Figure 3. Calculation of bicipital forearm angle (O) from com-
puted tomography scans obtained of both the proximal (A)
and distal (B) humerus. (A) Solid line: line connecting the
apexes of the lesser and greater tubercles. (B) Dashed line:
distal humeral axis. Solid line: line perpendicular to the distal
humeral axis. Dotted line: transpose of the line connecting
the apexes of the lesser and greater tubercles from A.
were treated as independent; thus, 24 limbs (12 partici-
pants assessed bilaterally) went into all analyses.
While the methods used in the current study have been
previously reported to be reliable and precise, it was impor-
tant for us to demonstrate that these specific data used to
establish validity were collected with reliable methods. As
such, both reliability and precision were established for all
variables in the current study. For the ultrasound humeral
torsion measurements, 3 trials obtained from each assessor
were used to calculate intratester (intraclass correlation coef-
ficient [ICC] = 3,1) and intertester reliability (ICC = 3,k) and
precision (standard error of measurement [SEM]). Similarly,
reliability and precision (ICC = 3,1 and 3,k; SEM) of the CT
humeral torsion data were calculated within and between
assessors. The mean of the CT humeral torsion angles
obtained across assessors represented the humeral torsion
variable that was compared with the ultrasound variable.
For the bicipital forearm angle (obtained from the CT
images), only one assessor obtained this variable given that
it was a variable of secondary importance. The intrarater reli-
ability and precision (ICC = 3,1 and SEM) were obtained from
the 3 trials completed for each limb. All statistical analyses
were conducted using SPSS 17 (SPSS Inc, Chicago, Illinois).
RESULTS
The descriptive statistics for all variables appear in Table 1.
The regression analysis demonstrated that the humeral tor-
sion data obtained with the ultrasound measurement had
a strong relationship with and were significant predictors
of both CT humeral torsion data (R = .797, R2 = .635, P \
.001) and CT bicipital forearm angle (R = .924, R2 = .853,
P \ .001). All regression statistics appear in Table 2, with
the corresponding scatter plots presented in Figures 4
and 5. The Bland-Altman plots demonstrate the measure-
ment bias and 95% limits of agreement between the ultra-
sound and CT variables in Figures 6 and 7.
For all variables assessed in the current study, the
intratester and intertester reliability yielded ICCs of .80
or higher. The SEM associated with the CT variables
4 Myers et al The American Journal of Sports Medicine

120 60
y = 1.2153x + 32.71
R = .797; R² = 0.6349 1.96 + SD

100 50

40
T (Degrees))

80

T Scan HT
Bias

30
US HT

60

US HT - CT
1.96 - SD
20
40

10
20

0
0 0 10 20 30 40 50 60 70 80
0 10 20 30 40 50 60
CT Scan HT (Degrees) (US HT + CT Scan HT)/2

Figure 4. Scatter plot and regression demonstrating the Figure 6. Limits of agreement plot between ultrasound and
relationship between ultrasound and computed tomography computed tomography humeral torsion (24 limbs).
humeral torsion (24 limbs).
20
120
y = 0.9332x + 11.316 1.96 + SD
R = .924; R² = 0.8533
15
100
earm Angle (Degrees)

10
80
US HT ((Degrees)

Bias
5
60
cipital Fore

0
40
US HT - Bic

1.96 - SD

–5
20
U

–10
0
0 10 20 30 40 50 60 70 80 90 100
0 10 20 30 40 50 60 70 80 90
Bicipital Forearm Angle (Degrees) (US HT + Bicipital Forearm Angle)/2

Figure 5. Scatter plot and regression demonstrating the Figure 7. Limits of agreement plot between ultrasound
relationship between ultrasound humeral torsion and com- humeral torsion and computed tomography humeral bicipital
puted tomography bicipital forearm angle (24 limbs). forearm angle (24 limbs).

ranged from 2.6° to 5.1°. The error associated with the torsion data obtained with CT and ultrasound, with 63%
ultrasound variables was lower, ranging from 0.8° to of the variance in CT humeral torsion being explained by
1.5°. All ICCs and corresponding SEMs appear in Table 1. the ultrasound data (Figure 4). These results suggest that
the assessment of humeral torsion with ultrasound provides
a viable alternative to the CT methodology, especially in the
DISCUSSION clinical setting when time, cost, and radiation are concerns.
While the relationship between the ultrasound and CT
The primary purpose of the current study was to establish methods is strong, there is still 37% of the variance that
the validity of using diagnostic ultrasound for measuring is unaccounted for in the regression model. We suspect
humeral torsion by comparing it to humeral torsion data that this variability stems from several methodological
collected with CT, the current gold standard. To meet this issues. With CT, the humeral head axis was defined as
purpose, humeral torsion was assessed in 24 limbs using a line perpendicular to the line bisecting the articular mar-
both the ultrasound and CT methods. Agreement between gins of the humeral head. The articular surface was iden-
the 2 methods was assessed through both regression analy- tified visually on each CT scan slice by the 2 assessors,
sis and visual interpretation of Bland-Altman plots. potentially resulting in measurement error in calculating
The results of the current study suggest that there is the humeral head axis. When we examined the reliability
a strong positive relationship (R = .797) between humeral of the CT data (the current gold standard), it exhibited
Vol. XX, No. X, XXXX
TABLE 1
Humeral Torsion Descriptive and Reliability Statisticsa
Dominant Nondominant
Limb,b deg Limb,b deg
US humeral torsion 74.2 6 14.5 61.2 6 14.4
CT humeral torsion 32.4 6 11.4 25.2 6 7.7
CT bicipital forearm angle 68.3 6 14.2 52.5 6 12.6
a
US, ultrasound; CT, computed topography; NC, not calculated.
b
Values presented as mean 6 standard deviation.
c
Values presented as intraclass correlation coefficient (standard error of measurement).
TABLE 2
Predicting Computed Tomography (CT) Variables From Ultrasound Humeral Torsion (Regression)
R R2
CT humeral torsion .797 .635
CT bicipital forearm angle .924 .853
lower reliability and greater error of measurement than
the ultrasound measurements. Thus, measurement error
associated with the CT method may account for some of
this variability, suggesting that, while commonly used in
the literature, CT may not best represent a true gold stan-
dard. The analysis of 95% limit of agreement demonstrates
that the difference between the torsion measured using
ultrasound and CT can be different as much as 18° higher
or lower (Figure 6). However, this analysis assumes that
both measures are reliable and therefore may be affected
by the lower reliability of the CT methodology.
An additional source of variability may be that the ultra-
sound data are affected by the amount of elbow carrying
angle present given that the ulnar inclination is used to indi-
rectly measure the distal humeral angle. It has been previ-
ously demonstrated that the typical carrying angle for male
athletes as in the current study is approximately 10° when
the arm is in extension.1,14,20-22 The carrying angle decreases
as the arm moves into flexion.14,21 In the current study, the
arm was placed at 90° of elbow flexion during the ultrasound
assessment, thus minimizing this source of variability.
A final source of variability may be the inherent differ-
ences between the 2 methods given that different anatomic
landmarks and ultimately a different angle are used
between the direct method (CT) and the indirect method
(ultrasound) compared in the current study. With the
ultrasound method, the primary landmarks of importance
are the humeral tubercles and bicipital groove, while these
landmarks are not used for the CT measurements. The
result is a consistent bias of approximately 38° that is evi-
dent in Figure 6. Potentially, variability with location of
the tubercles and bicipital groove relative to the axis of
the humerus (from CT scans) may account for some of
the variability between the ultrasound and CT humeral
torsion measurements.7 Because of the bias present, it
must be emphasized that each method is measuring a dif-
ferent anatomic angle; thus, the data obtained from each
method cannot be directly compared. But clinically, both
anatomic angles represent the amount of twisting present
at the humerus. Despite the potential sources of variability
described above, the data obtained with the CT
Ultrasound of Humeral Retrotorsion in Baseball Players 5
Intrarater Reliability Interrater Reliability
and Precisionc and Precisionc
.997 (0.8) .991 (1.5)
.933 (2.6) .805 (5.1)
.948 (2.9) NC
Standard Error of Estimate P Value
6.31 \.001
6.04 \.001
methodology mimic data for overhead-throwing athletes
presented in the literature.4,11,12
Given that ultrasound assessment of humeral torsion is
an indirect measure of humeral torsion, a subsequent anal-
ysis was conducted where the ultrasound measurement was
directly compared with an equivalent angle (bicipital fore-
arm angle) calculated from the CT images. As would be
expected, there was strong agreement between the 2 angles
(Figure 5), with a bias of approximately 7° between the 2
measures (Figure 7). This small bias as well as remaining
14% variability that was unaccounted for in the regression
model was most likely a result of the minimal role that car-
rying angle plays when the arm is flexed to 90° as well as
the influence of the soft tissue that is present between the
inclinometer and the ulna when performing the ultrasono-
graphic method. Given the strong agreement, we are confi-
dent that the ultrasound method is accurate in calculating
the intended angle. The higher reliability associated with
identifying the landmarks needed for the bicipital forearm
angle (on the CT images), compared with the primary CT
measurement used in the current study, suggests that
more research should focus on the potential usefulness of
the bicipital forearm angle as an indicator of humeral tor-
sion when CT is used. It was the authors’ experience that
it was much easier to identify the glenoid tubercles used
for the bicipital forearm angle compared with finding the
articular margins for the standard CT assessment. Overall,
more research should focus on best establishing the gold
standard for humeral torsion assessment.
Clinically, results of this study demonstrated that the
ultrasound assessment can be used as a viable screening
tool for the amount of humeral torsion present in
overhead-throwing athletes as well as other patient popu-
lations. For example, Myers et al9 recently demonstrated
that collegiate pitchers with a history of throwing-related
elbow injury exhibited increased humeral torsion com-
pared to pitchers with no injury history. Whiteley et al17
prospectively determined that the amount of humeral tor-
sion present in adolescent baseball players was predictive
of injury. The same ultrasound methodology (as in the cur-
rent study) was utilized by both Myers et al9 and Whiteley
6 Myers et al
et al,17 suggesting that clinicians can use the methods for
injury risk screening.
In addition to screening for potential injury risk, the ultra-
sound methodology can also be used as an adjunct screening
tool for the assessment of humeral rotation range of motion.
Assessing humeral rotation range of motion either as part of
preparticipation screenings or the injury evaluation process
in the overhead-throwing athlete is a very common clinical
practice. Yet, it has been previously shown that the amount
of humeral torsion plays a significant role in the amount of
humeral rotation range of motion observed clinically.3,8,12 A
clinician may use ultrasound to assess humeral torsion as
part of his or her humeral rotation range of motion screening
procedures to determine how much of the amount of humeral
rotation range of motion observed results from humeral tor-
sion versus soft tissue tightness.8
The current study has 2 limitations that warrant
acknowledgment. In the current study, only 24 limbs from
12 participants were assessed. Typically, regression analy-
ses warrant a larger sample size. But the costs and radia-
tion exposure associated with CT hindered our sample
size. Despite the lower sample size, a strong relationship
between the CT and ultrasound assessments was demon-
strated. Additionally, the current study only examined colle-
giate baseball players. Baseball players consistently exhibit
more humeral torsion characteristics in their dominant
limbs than the normal population. To minimize the effects
of this limitation, we not only included their dominant
limb but also included their nondominant limbs in the anal-
yses. Previous work has shown that the nondominant limbs
in baseball players mimic the amount of torsion present in
patients with no history of overhead sport participation.8,15
The nondominant limb data that appear in Table 1 are con-
sistent with torsion present in non–overhead-throwing ath-
lete populations presented in the literature.8,15
CONCLUSION
There was a strong relationship between humeral torsion
data obtained with ultrasound and CT, the current gold
standard of assessment. Ultrasonography provides a valid,
reliable alternative to CT for measuring humeral torsion in
the upper extremity.
ACKNOWLEDGMENT
The authors thank Elizabeth Hibberd, Casey Shutt, Joan
Bowen, and Yolanda Dunsmore for their assistance with
participant recruitment and data collection.
REFERENCES
1. An KN, Morrey BF, Chao EY. Carrying angle of the human elbow joint.
J Orthop Res. 1984;1(4):369-378.
For reprints and permission queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav
The American Journal of Sports Medicine
2. Boileau P, Bicknell RT, Mazzoleni N, Walch G, Urien JP. CT scan
method accurately assesses humeral head retroversion. Clin Orthop
Relat Res. 2008;466(3):661-669.
3. Chant CB, Litchfield R, Griffin S, Thain LM. Humeral head retrover-
sion in competitive baseball players and its relationship to glenohum-
eral rotation range of motion. J Orthop Sports Phys Ther.
2007;37(9):514-520.
4. Crockett HC, Gross LB, Wilk KE, et al. Osseous adaptation and range
of motion at the glenohumeral joint in professional baseball pitchers.
Am J Sports Med. 2002;30(1):20-26.
5. Hernigou P, Duparc F, Hernigou A. Determining humeral retroversion
with computed tomography. J Bone Joint Surg Am.
2002;84(10):1753-1762.
6. Ito N, Eto M, Maeda K, Rabbi ME, Iwasaki K. Ultrasonographic mea-
surement of humeral torsion. J Shoulder Elbow Surg. 1995;4(3):157-
161.
7. Kummer FJ, Perkins R, Zuckerman JD. The use of the bicipital groove
for alignment of the humeral stem in shoulder arthroplasty. J Shoul-
der Elbow Surg. 1998;7(2):144-146.
8. Myers JB, Oyama S, Goerger BM, Rucinski TJ, Blackburn JT,
Creighton RA. Influence of humeral torsion on interpretation of poste-
rior shoulder tightness measures in overhead athletes. Clin J Sport
Med. 2009;19(5):366-371.
9. Myers JB, Oyama S, Rucinski TJ, Creighton RA. Humeral retrotorsion
in collegiate baseball pitchers with throwing-related upper extremity
injury history. Sports Health. 2011;3(4):383-389.
10. Osbahr DC, Cannon DL, Speer KP. Retroversion of the humerus in
the throwing shoulder of college baseball pitchers. Am J Sports
Med. 2002;30(3):347-353.
11. Pieper HG. Humeral torsion in the throwing arm of handball players.
Am J Sports Med. 1998;26(2):247-253.
12. Reagan KM, Meister K, Horodyski MB, Werner DW, Carruthers C,
Wilk K. Humeral retroversion and its relationship to glenohumeral
rotation in the shoulder of college baseball players. Am J Sports
Med. 2002;30(3):354-360.
13. Schwab LM, Blanch P. Humeral torsion and passive shoulder range
in elite volleyball players. Phys Ther Sport. 2009;10(2):51-56.
14. Van Roy P, Baeyens JP, Fauvart D, Lanssiers R, Clarijs JP. Arthro-
kinematics of the elbow: study of the carrying angle. Ergonomics.
2005;48(11-14):1645-1656.
15. Whiteley R, Ginn K, Nicholson L, Adams R. Indirect ultrasound mea-
surement of humeral torsion in adolescent baseball players and non-
athletic adults: reliability and significance. J Sci Med Sport.
2006;9(4):310-318.
16. Whiteley RJ, Adams RD, Nicholson LL, Ginn KA. Shoulder proprio-
ception is associated with humeral torsion in adolescent baseball
players. Phys Ther Sport. 2008;9(4):177-184.
17. Whiteley RJ, Adams RD, Nicholson LL, Ginn KA. Reduced humeral
torsion predicts throwing-related injury in adolescent baseballers. J
Sci Med Sport. 2010;13(4):392-396.
18. Whiteley RJ, Ginn KA, Nicholson LL, Adams RD. Sports participation
and humeral torsion. J Orthop Sports Phys Ther. 2009;39(4):256-263.
19. Yamamoto N, Itoi E, Minagawa H, et al. Why is the humeral retrover-
sion of throwing athletes greater in dominant shoulders than in non-
dominant shoulders? J Shoulder Elbow Surg. 2006;15(5):571-575.
20. Zampagni ML, Casino D, Martelli S, Visani A, Marcacci M. A protocol
for clinical evaluation of the carrying angle of the elbow by anatomic
landmarks. J Shoulder Elbow Surg. 2008;17(1):106-112.
21. Zampagni ML, Casino D, Zaffagnini S, Visani A, Marcacci M. Trend of
the carrying angle during flexion-extension of the elbow joint: a pilot
study. Orthopedics. 2008;31(1):76.
22. Zampagni ML, Casino D, Zaffagnini S, Visani AA, Marcacci M. Esti-
mating the elbow carrying angle with an electrogoniometer: acquisi-
tion of data and reliability of measurements. Orthopedics.
2008;31(4):370.