Abstract
Background: Classification of fractures of the greater tuberosity
has shown poor reliability, in part as a result of an inability to assess
fracture displacement accurately. We used fluoroscopic images of prepositioned
osteotomized greater tuberosity fragments in cadavers to determine the
accuracy of radiographic interpretation, the interobserver reliability, and
the effect that radiographs might have on surgical decision-making.
Methods: Twelve osteotomies of the greater tuberosity (three each
with 2, 5, 10, and 15 mm of displacement) were created in whole-body cadavers.
Six fluoroscopic images (anteroposterior views in external and internal
rotation, anteroposterior views in neutral rotation with 15° of cephalic
and 15° of caudal tilt, a lateral outlet view, and an axillary view) were
made after each osteotomy. Four experienced orthopaedic surgeons measured
displacement in millimeters on seventy-two randomized images. Four views in
sequence (the anteroposterior view in internal rotation and the outlet view
together, then the axillary view, and then the anteroposterior view in
external rotation) of each osteotomy pattern were then viewed, and each
surgeon was asked whether surgery would be indicated on the basis of each set
of images.
Results: No one fluoroscopic view was significantly more accurate
than another. There was a trend toward increased accuracy of imaging of
minimally displaced (=5 mm) tuberosity fragments with the anteroposterior
view in external rotation. When viewed sequentially, the anteroposterior view
in external rotation, evaluated last, altered treatment in nine of forty-eight
situations. There was substantial agreement (? = 0.71) among the
surgeons with respect to their recommendations for treatment of the displaced
greater tuberosities after they had inspected the four images.
Conclusions and Clinical Relevance: To our knowledge, we are the
first to examine the accuracy and reliability of interpreting images of known
displacements of the greater tuberosity. Multiple radiographic views are
needed to evaluate displacement of the greater tuberosity appropriately. The
anteroposterior view in external rotation can profile the greater tuberosity
and help demonstrate small displacements. Treatment decisions should be
consistent between surgeons when multiple views are used.
Fractures of the proximal part of the humerus are not uncommon, and
approximately 15% to 40% of such fractures involve the greater
tuberosity1-5.
Most authors have agreed that nondisplaced fractures of the greater tuberosity
can be treated nonoperatively, but there is debate regarding the degree of
displacement that requires surgery. Recent cadaveric and clinical studies have
indicated that =5 mm of displacement can cause impingement and a poor
functional
outcome3,6,7.
Accurate radiographic assessment of fractures of the greater tuberosity can
be difficult because of small fragment sizes and superimposition of the
fragment on the humeral
head8-11.
Numerous authors have investigated the reliability and reproducibility of both
the Neer and the AO/ASIF system for classification of proximal humeral
fractures12,13.
Many of those studies have highlighted the difficulty of reliably assessing
the number of fracture fragments or their displacement on either standard
radiographs or computerized tomography
scans12,14-17.
To our knowledge, no one has evaluated the accuracy of assessing known
displacements on radiographs.
The purpose of the current study was to evaluate the ability of surgeons
with experience with shoulder fractures to assess displacement of fractures of
the greater tuberosity; we did this by comparing their measurements of
displacements on radiographs with actual displacements in cadaveric specimens.
We also investigated the interobserver reliability of the surgeons' decisions
whether to operate on greater tuberosity fractures on the basis of their
radiographic appearance and how various radiographic projections might
influence operative decision-making.
Twelve isolated greater tuberosity fragments were created in whole-body
cadavers after the proximal part of the humerus was dissected through a
standard deltopectoral approach. The deltoid insertion was partially
transected to improve exposure of the greater tuberosity, but the insertions
of the supraspinatus, infraspinatus, and teres minor were left intact.
The tuberosity osteotomies were performed with an oscillating saw. The
osteotomy plane was made directly at the base of the tuberosity and did not
extend into the articular surface of the humeral head. Also, the osteotomy
plane did not involve the intertubercular groove of the proximal part of the
humerus.
Each tuberosity was displaced an equal measured amount superiorly and
posteriorly along a 45° angle between the vertical and horizontal (to
mimic the in vivo pull of the supraspinatus, infraspinatus, and teres minor on
a fractured greater tuberosity, which has commonly been found to displace
posterosuperiorly)8.
Along this angle, the osteotomized tuberosities were displaced 2, 5, 10, or 15
mm posterosuperiorly and were sutured in place with multiple opposing number-5
nonabsorbable sutures through bone tunnels in both the proximal part of the
humerus and the tuberosity to hold the tuberosity securely in the desired
position. The stability of the suture fixation was then tested through a range
of motion of the shoulder and was double-checked after each image was made.
Twelve tuberosity displacements (three 2-mm, three 5-mm, three 10-mm, and
three 15-mm displacements) were created. The soft tissue was then replaced
over the proximal part of the humerus.
Six different fluoroscopic images were made of each displaced tuberosity.
The c-arm of the fluoroscopy unit was positioned flush with the shoulder for
all views, in an effort to standardize magnification across all images. A
pilot study was performed to determine fluoroscopic magnification and to
compare it with standard radiographic magnification. The fluoroscopic
magnification was found to be 31% compared with 20% for the standard
radiographs18. A
standard anteroposterior view (perpendicular to the shoulder) with the arm in
15° of internal rotation and a lateral outlet view were obtained. In
addition, we obtained two views that require repositioning of the arm, an
axillary view and a "true" anteroposterior view (perpendicular to
the glenohumeral joint) with the humerus in 30° of external rotation (the
"gunslinger" view). To evaluate the effect of cephalad or caudal
rotation, two standard anteroposterior views with the fluoroscopy unit placed
in 15° of cephalad or caudal tilt were obtained as well. For both of those
views, the arm was held in neutral rotation. A total of seventy-two images
(six projections for each of twelve displaced tuberosities) were made (Figs.
1-A,
1-B,
1-C,
1-D,
1-E,
1-F).
Each image was given a coded moniker and randomized into a book of images.
Four attending orthopaedic surgeons of the Mount Sinai Department of
Orthopaedics, two of whom were trained in shoulder surgery and two of whom had
general trauma experience, were asked to determine the amount of displacement
(in millimeters) of the tuberosity on each image by measuring from the edge of
the defect in the humeral head to the closest edge of the greater tuberosity.
The observers were not asked to indicate the direction of the displacement.
Each observer interpreted the images independently.
The surgeons were given a standard ruler, not calibrated for magnification,
and a goniometer to make their measurements, and all knew that each image
showed an isolated osteotomy of the greater tuberosity. The type of projection
was not specified. In addition to determining the amount of displacement of
each fragment, the surgeons indicated whether they thought that the displaced
tuberosity would require operative fixation in a real-life situation.
After viewing the seventy-two images in random order and answering the two
questions (regarding the amount of displacement and whether surgery was
required), the surgeons were presented with a series of four images (the
anteroposterior view in internal rotation, the outlet view, the axillary view,
and the anteroposterior view in external rotation) of each osteotomized
tuberosity, with the twelve osteotomized tuberosities presented in random
order. The four fluoroscopic views were chosen because they represent
orthogonal radiographs that are typically made for a patient with a fracture
of the greater tuberosity. The anteroposterior views with 15° of caudal or
cephalad tilt were not included because they are not typically made in the
initial evaluation of a patient with a greater tuberosity fracture. The four
images of each osteotomized tuberosity were presented in a sequential fashion
in the same order to each surgeon: each reviewer first evaluated the
anteroposterior view in internal rotation and the outlet view together, then
assessed the axillary view, and finally evaluated the anteroposterior view in
external rotation. After each sequential image was presented, the surgeon was
asked to indicate whether he would recommend surgery solely on the basis of
the radiographic appearance. The surgeon was not given guidance regarding how
much displacement requires surgery. There is controversy about the precise
indication for surgery in the literature, but for this project we considered 5
mm of displacement to be the cutoff for "appropriate" operative
treatment.
Statistical Analysis
The accuracy of the observers' measurements of the displacements on the
seventy-two images was determined by calculating the average error—i.e.,
the amount that a measured value differed from the known value, averaged over
the data set. The average error of the measurement on each fluoroscopic view
was also determined for each osteotomy displacement (2, 5, 10, and 15 mm) and
overall. The difference between the accuracy of the shoulder surgeons and that
of the general orthopaedists was evaluated for significance with use of a
Wilcoxon signed-rank test, which allows paired analysis, with a significance
level of p < 0.05. An analysis-of-variance procedure called the Friedman
test, ranked for nonparametric distributions, with a significance level of p
< 0.05, was used to evaluate the significance of the accuracy of each
surgeon when measuring displacement. The Friedman test was also used to
evaluate significance in a comparison of the accuracies of the six
fluoroscopic views.
The amount of agreement among the surgeons' decisions about whether an
operation was indicated after they had viewed the four sequential images of
each osteotomy displacement (2, 5, 10, and 15 mm) was also determined. The
number of times that the axillary view or the anteroposterior view in external
rotation (reviewed third and last, respectively) changed the decision
regarding operative treatment was determined as well.
The interobserver reliability for determining the displacement and whether
an operation was indicated was described with use of weighted kappa
coefficients19,20.
This method relates the observed proportion of agreements with the proportion
expected by chance. According to guidelines described by Landis and Koch, a
value of >0.80 indicates "excellent" agreement; 0.61 to 0.80,
"substantial" agreement; 0.41 to 0.60, "moderate"
agreement; 0.21 to 0.40, "fair" agreement; and =0.20,
"poor" or "slight"
agreement21. A
biostatistician at the Mount Sinai School of Medicine performed all
statistical analyses.
Accuracy of Fluoroscopic Imaging for Measurement of Displacement
The anteroposterior view in external rotation and the antero-posterior view
with 15° of caudal tilt were found to be the most accurate for the
assessment of all patterns of displacement, with an average error of 5.1 mm
overall (range of error for all views, 5.1 to 6.5 mm); however, this
difference was not significant (p = 1.0). The 2-mm displacements were measured
most accurately on the anteroposterior view in external rotation (average
error, 0.7 mm; p = 0.22), whereas the 5 and 10-mm displacements were measured
most accurately on the anteroposterior view with 15° of caudal tilt
(average error, 1.8 and 6.3 mm, respectively). Finally, the 15-mm
displacements were measured most accurately on the axillary view (average
error, 8.5 mm). The average error did not differ significantly among the
images for any amount of displacement
(Table I). The data revealed a
trend toward increasing inaccuracy as displacement increased. The average
error increased with increasing displacement for all of the images except for
the caudal and cephalad views showing 2 and 5-mm displacements.
Accuracy and Agreement of Surgeons When Measuring Displacement
The shoulder surgeons were more accurate in assessing the amount of
displacement on the individual images of the displaced tuberosities, with an
average error of 4.4 mm (4.1 and 4.7 mm) compared with an average error of 6.9
mm (6.5 and 7.2 mm) for the general orthopaedists. However, this difference
did not reach significance in the grouped comparison (p = 0.076).
Operative Recommendations Based on Individual Images of Displaced
Tuberosities
The individual images of the tuberosities with 2 mm of displacement were
used as an indication for nonoperative treatment by the four surgeons in
sixty-four (89%) of seventy-two instances
(Table II). Nonoperative
treatment was recommended in all thirty-six instances in which an
anteroposterior view in internal rotation, an axillary view, or an outlet view
of a tuberosity with 2 mm of displacement was evaluated. Conversely, surgery
was thought to be indicated in thirty-two of the seventy-two instances in
which an image of a tuberosity with 5 mm of displacement was examined. With
regard to the images of the tuberosities with 5 mm of displacement, only the
outlet view led to complete agreement among the surgeons that operative
treatment was not appropriate
(Table II). Tuberosities with
10 or 15 mm of displacement were considered to be an indication for an
operation in 110 of 144 instances. The recommendation to operate on a
tuberosity with 10 or 15 mm of displacement was based more frequently on the
anteroposterior view in external rotation (twenty-three of twenty-four images)
than on the axillary view (fourteen of twenty-four) or the outlet view
(thirteen of twenty-four). The interobserver reliability among all surgeons
with regard to the choice of treatment based on the individual seventy-two
images was moderate, with a kappa value of 0.45. The general orthopaedists
agreed on the treatment choice more reliably (kappa value, 0.61) than did the
shoulder surgeons (kappa value, 0.38).
Operative Recommendations Based on Four Sequential Fluoroscopic
Images of a Displaced Tuberosity
Nonoperative treatment was recommended in eleven of the twelve instances in
which the four images (anteroposterior in internal rotation, outlet, axillary,
and anteroposterior in external rotation) of the three tuberosities with 2 mm
of displacement were viewed sequentially by the four surgeons. In nine of the
twelve instances in which the four images of the tuberosities with 5 mm of
displacement were evaluated, an operation was recommended. In all of the
instances in which all four images of the tuberosity fragments with 10 or 15
mm of displacement were viewed, an operation was recommended.
The "appropriate" treatment (nonoperative management for a
tuberosity displaced <2 mm and operative management for a tuberosity
displaced =5 mm) was recommended after the first two views (the
anteroposterior view with internal rotation and the outlet view) were examined
in thirty-seven of fortyeight instances. The axillary view (the next in the
sequence) changed the treatment in only one situation; the treatment
recommendation for one tuberosity with 5 mm of displacement was altered from
nonoperative to operative. Thus, after examination of three views, the
"appropriate" treatment was recommended in thirty-eight of
forty-eight instances.
The "appropriate" treatment was recommended after examination
of all four views in forty-two of forty-eight instances. Interestingly, the
anteroposterior view in external rotation, which was the last of the four
images to be examined, altered treatment in nine of the forty-eight cases. In
five of those nine cases, the recommendation for nonoperative treatment was
changed to operative treatment for a tuberosity displaced 5 mm. In the
remaining four cases, the recommendation was changed to operative treatment
for a tuberosity displaced 2 mm.
The interobserver reliability between the observers with regard to their
recommendations for treatment of the displaced greater tuberosity fragments on
the basis of four sequential views ranged from 0.41 to 0.84 ("moderate
to excellent"). The interobserver reliability between the shoulder
surgeons (kappa value, 0.84) and the general orthopaedists (kappa value, 0.82)
were both "excellent." Overall, the kappa value for the agreement
among all four surgeons after they had evaluated the four views was 0.71,
which is "substantial" agreement.
The ability to accurately assess displacement of greater tuberosity
fractures is important because the amount of displacement has direct
implications with regard to treatment and long-term
outcome2,3,7,8,12,22.
Historically, it was thought that all fractures with >10 mm of displacement
required open reduction and internal fixation, but recent reports have shown
that nonoperative treatment of fractures with <10 mm of displacement may
have an unsatisfactory
result3,6.
Studies of nonoperative treatment of fractures of the greater tuberosity have
documented that displaced fractures have poorer results than nondisplaced
fractures23,24.
The displaced fragment has been implicated in mechanical blockage of abduction
(due to impingement on the acromion) and of external rotation (due to
impingement on the posterior aspect of the
glenoid)25-28.
Also, the malunited fragment can affect the function of the rotator cuff.
Recent studies of operative repair of displaced fractures have shown better
results. One of us (E.F.) and
colleagues8
previously found that twelve patients treated with suture fixation of a
displaced fracture of the greater tuberosity had either a good or an excellent
result, with an average of 170° of active elevation. Kim and
Ha29 reported that
twenty-three patients in whom a displaced fracture of the greater tuberosity
had been initially treated nonoperatively for at least six months had
improvement two years following arthroscopic reduction of the tuberosity and
subacromial decompression. These studies highlight the need to diagnose
displaced greater tuberosity fractures accurately and to determine the amount
of displacement with appropriate radiographic projections.
Authors have commented on the difficulty of measuring the amount of
displacement of the tuberosity accurately on standard
radiographs11,30-32.
As early as 1912, Phemister commented that rotational anteroposterior
radiographs were needed to visualize the displaced tuberosity that is
superimposed on the humeral
head10. The
previous study by one of us (E.F.) and
colleagues8
indicated that the outlet and anteroposterior views were reliable to assess
superior displacement of the tuberosity and that an adequate axillary view
could be used to evaluate posterior displacement.
We thought that our observers had difficulty assessing tuberosity
displacement accurately. The anteroposterior view in external rotation (the
gunslinger view) tended to be the most accurate image for the assessment of
the tuberosities with 2 mm of displacement. Rotational views of individual
tuberosities were also found to be most helpful when the surgeons were
deciding on the best treatment modality, especially for tuberosity fragments
with moderate displacement (=10 mm). We found that most of the views were
adequate for determining "appropriate" treatment of minimally
displaced tuberosities (2 mm). However, the difficulty in assessing
tuberosities with 5 mm of displacement was again demonstrated by the
infrequency with which an operation was recommended (thirty-two of seventy-two
situations) when single images were viewed. The agreement among our observers
with respect to operative decision-making after they evaluated single images
of displaced tuberosities was "moderate" (kappa value, 0.45),
which is similar to previous
data15,16.
Although our study of the accuracy and reliability of seventy-two
individual images of osteotomized tuberosities was useful for determining
which projection profiled the greater tuberosity best, those data may not be
very applicable clinically. Rarely are surgeons required to measure
displacement or recommend treatment on the basis of one radiographic
projection. The data that we obtained from the surgeons' interpretations of
the series of images of a given fracture may be more clinically relevant. We
found that the four surgeons recommended operative treatment for the
tuberosities with 5 mm of displacement in nine of twelve instances and they
always recommended operative treatment for the tuberosities with 10 or 15 mm
of displacement after they had reviewed all four images (anteroposterior with
internal rotation, outlet, axillary, and anteroposterior with external
rotation). Evaluation of the anteroposterior view in external rotation (viewed
last) resulted in a change in treatment in nine of the forty-eight cases,
although only five of the nine changes were "appropriate."
Conversely, the axillary view (the third to be evaluated) altered treatment in
only one situation. Unlike previous
investigators14-17,
we found substantial agreement (kappa value, 0.71) among surgeons with regard
to their recommendations for treatment of fractures on the basis of four
sequential images.
This study, although novel, has limitations. Of some concern is the high
average error values for the measurements of the displacement. The surgeons'
measurements on the individual views differed from the actual values by an
average of 4.4 mm (shoulder surgeons) and 6.9 mm (general orthopaedists). We
also found that as the displacement of the tuberosities increased, so did the
average error of the surgeons' measurements (regardless of the radiographic
projection). Although this may not be clinically important in the evaluation
of fractures that are displaced a large amount (>15 mm), it can be very
important in the evaluation of those displaced 5 mm. One reason for high error
values may be a lack of familiarity with greater tuberosity fractures, as may
be evidenced by the higher average error value for the general orthopaedists
than for the shoulder specialists. However, the high error values may also
have been a result of the quality of the imaging. A fluoroscopic mini-c-arm
imager was used to generate paper images of each displaced tuberosity. A
standard radiograph of each tuberosity would have improved the image
quality.
We used standard rulers that did not account for any magnification error
that may have occurred during the imaging. We believe that this replicates
what actually occurs in the clinical setting when physicians use a ruler to
measure fracture displacement. Our pilot study suggested that the
magnification with our technique was slightly greater than that found on
standard radiographs. This would be expected to produce an overestimation of
fracture displacement due to fluoroscopic magnification. However, we observed
both overestimation and underestimation error in the measurements of the
displacement, a finding that cannot be explained solely by the fluoroscopic
magnification.
Another shortcoming of this study is that the estimates of displacement of
the tuberosity fragments were based solely on single images. Unfortunately,
when each surgeon was presented with four sequential images of a displaced
tuberosity, he was asked to determine only whether surgery was indicated. As
we stated, it is likely that a surgeon's error in estimating displacement on
the basis of all four views is more clinically relevant. All four surgeons
indicated, after they had completed the tests, that their criterion for
recommending open reduction and internal fixation was =5 mm of displacement
but they often modified that criterion. Some of this indecision may have
contributed to their difficulty in deciding on surgery for the tuberosities
with 5 mm of displacement. One surgeon reported that he accepted up to 5 mm of
superior displacement but up to 10 mm of posterior displacement. This study
might have been more meaningful if a precise surgical indication had been
recommended or if the surgeons had been asked to estimate displacement after
evaluating the full series of views.
Of secondary concern is how easily a limb with a proximal humeral fracture
can actually be manipulated to obtain these views. It may be that patients
will not tolerate the positioning required to obtain the anteroposterior view
in external rotation. Anecdotally, it is understood how difficult it can be to
obtain an axillary view of these patients, and often a Velpeau axillary view,
as recommended by Neer, may be
necessary27.
This study revealed some of the difficulties associated with accurately
assessing fracture displacement. However, we believe that using multiplane
radiographic imaging that includes an anteroposterior view in external
rotation (if it can be tolerated by the patient) in association with the
standard trauma series of anteroposterior, outlet, and axillary views may
improve the accuracy and reliability of diagnostic imaging. We recommend using
these four radiographic views to assess displacement of greater tuberosity
fractures, as we think that these multiple views can help surgeons to make
predictable, appropriate choices for the treatment of these fractures.
?
Horak J, Nilsson BE. Epidemiology of
fracture of the upper end of the humerus. Clin Orthop.1975;112:
250-3.112250
1975
[PubMed]
Lind T, Kroner K, Jensen J. The
epidemiology of fractures of the proximal humerus. Arch Orthop Trauma
Surg.1989;108:
285-7.108285
1989
[PubMed][CrossRef]
McLaughlin HL. Dislocation of the
shoulder with tuberosity fracture. Surg Clin North Am.1963;43:
1615-20.431615
1963
[PubMed]
Meyerding HW. Fracture-dislocation of
the shoulder. Minn Med.1937;20:
717-26.20717
1937
Rose SH, Melton LJ 3rd, Morrey BF,
Ilstrup DM, Riggs BL. Epidemiologic features of humeral fractures. Clin
Orthop.1982;168:
24-30.16824
1982
Bigliani LU, Flatow EL, Pollock RG.
Fractures of the proximal humerus. In: Harryman DT, Matsen FA, Rockwood CA,
Wirth MA, editors. 2nd ed., vol. 1. The
shoulder. Philadelphia: WB Saunders; 1998. p
337-90.1337
1998
Bono CM, Renard R, Levine RG, Levy AS.
Effect of displacement of fractures of the greater tuberosity on the mechanics
of the shoulder. J Bone Joint Surg Br.2001;83:
1056-62.831056
2001
[PubMed][CrossRef]
Flatow EL, Cuomo F, Maday MG, Miller SR,
McIlveen SJ, Bigliani LU. Open reduction and internal fixation of two-part
displaced fractures of the greater tuberosity of the proximal part of the
humerus. J Bone Joint Surg Am.1991;73:
1213-8.731213
1991
[PubMed]
Morris ME, Kilcoyne RF, Shuman W, Matsen
F 3rd. Humeral tuberosity fractures: evaluation by CT scan and management of
malunion. Orthop Trans.1987;11:
242.11242
1987
Phemister DB. Fractures of the greater
tuberosity of the humerus. With an operative procedure for fixation.
Ann Surg.1912;56:
440-9.56440
1912
[PubMed][CrossRef]
Wallace WA, Hellier M. Improving
radiographs of the injured shoulder. Radiography.1983;49:
229-33.49229
1983
[PubMed]
Neer CS 2nd. Displaced proximal humeral
fractures. I. Classification and evaluation. J Bone Joint Surg
Am.1970;52:
1077-89.521077
1970
Jakob RP, Kristiansen T, Mayo K, Ganz R,
Müller ME. Classification and aspects of treatment of fractures of the
proximal humerus. In: Bateman JE, Welsh PR, editors. Surgery of the
shoulder. Philadelphia: BC Decker; 1984. p
330-43.330
1984
Bernstein J, Adler LM, Blank JE, Dalsey
RM, Williams GR, Iannotti JP. Evaluation of the Neer system of classification
of proximal humeral fractures with computerized tomographic scans and plain
radiographs. J Bone Joint Surg Am.1996;78:
1371-5.781371
1996
[PubMed]
Sidor ML, Zuckerman JD, Lyon T, Koval K,
Cuomo F, Schoenberg N. The Neer classification system for proximal humeral
fractures. An assessment of interobserver reliability and intraobserver
reproducibility. J Bone Joint Surg Am.1993;75:
1745-50.751745
1993
[PubMed]
Sjoden GO, Movin T, Guntner P, Aspelin
P, Ahrengart L, Ersmark H, Sperber A. Poor reproducibility of classification
of proximal humeral fractures. Additional CT of minor value. Acta
Orthop Scand.1997;68:
239-42.68239
1997
[CrossRef]
Siebenrock KA, Gerber C. The
reproducibility of classification of fractures of the proximal end of the
humerus. J Bone Joint Surg Am.1993;75:
1751-5.751751
1993
[PubMed]
Clarke IC, Gruen T, Matos M, Amstutz HC.
Improved methods for quantitative radiographic evaluation with particular
reference to total-hip arthroplasty. Clin Orthop.1976;121:
83-91.12183
1976
[PubMed]
Fleiss JL. Statistical methods
for rates and proportions. 2nd ed. New York: Wiley;
1981.
1981
Dunn G. Design and analysis of
reliability studies: the statistical evaluation of measurement errors.
New York: Oxford University Press; 1989.
1989
Landis JR, Koch GG. The measurement of
observer agreement for categorical data. Biometrics.1977;33:
159-74.33159
1977
[PubMed][CrossRef]
Stevens JH. The action of the short
rotators on the normal abduction of the arm, with a consideration of their
action in some cases of subacromial bursitis and allied conditions. Am
J Med Sci.1909;138:
870-84.138870
1909
[CrossRef]
Olivier H, Duparc J, Romain F. Fractures
of the greater tuberosity of the humerus. Orthop Trans.1986;10:
223.10223
1986
Paavolainen P, Bjorkenheim JM, Slatis P,
Paukku P. Operative treatment of severe proximal humeral fractures.
Acta Orthop Scand.1983;54:
374-9.54374
1983
[PubMed][CrossRef]
Codman EA. The shoulder.
Rupture of the supraspinatus tendon and other lesions in or about the
subacromial bursa. Boston: Thomas Todd;
1934.
1934
Depalma AF, Cautilli RA. Fractures of
the upper end of the humerus. Clin Orthop.1961;20:
73-93.2073
1961
[PubMed]
Neer CS 2nd. Shoulder
reconstruction. Philadelphia: Saunders; 1990. p
377-80.377
1990
Neviaser JS. Complicated fractures and
dislocations about the shoulder joint. J Bone Joint Surg Am.1962;44:
984-98.44984
1962
Kim SH, Ha KI. Arthroscopic treatment of
symptomatic shoulders with minimally displaced greater tuberosity fracture.
Arthroscopy.2000;16:
695-700.16695
2000
[PubMed][CrossRef]
Ahovuo J, Paavolainen P, Bjorkenheim JM.
Fractures of the proximal humerus involving the intertubercular groove.
Acta Radiol.1989;30:
373-4.30373
1989
[PubMed][CrossRef]
Mills HJ, Horne G. Fractures of the
proximal humerus in adults. J Trauma.1985;25:
801-5.25801
1985
[PubMed][CrossRef]
Neer CS 2nd. Fractures about the
shoulder. In: Rockwood CA Jr, Green DP, editors. Fractures in
adults. 2nd ed. Philadelphia: Lippincott; 1984. p
675-721.675
1984