Natural history of complex fractures of the proximal humerus
using a three-dimensional classification system
Gordon Edelson, MD,a Husam Safuri, MD,a Joseph Salami, MD,a Fina Vigder, MD,b
and Daniela Militianu, MD,c Tiberias and Haifa, Israel
We studied the nonoperative treatment of proximal
humeral fractures in severe injuries usually treated
surgically. The natural history of 63 patients was followed
prospectively for 2 to 9 years (mean, 42 months) with
a nonrandomized protocol. A 3-dimensional
classification system based on computed tomography
scans was used to categorize the fractures. Assessment
was made for range of motion, function via a validated
testing instrument (Simple Shoulder Test), analog pain
score, avascular necrosis (AVN), and fracture union.
Magnetic resonance imaging for early signs of AVN was
done in 16 cases. After conservative treatment of
complex fractures of the shoulder, motion is considerably
compromised but pain is minimal and functional status is
acceptable to most individuals in this predominantly
older patient population. Status comparable to
a successful surgical shoulder fusion is achieved in most
cases—Nature’s fusion. Contrary to common belief,
AVN, even in severely displaced injuries, is rare. Future
randomized studies based on a 3-dimensional
classification need to be done to compare these natural
history results with various types of surgical interventions.
(J Shoulder Elbow Surg 2008;17:399-409.)
T he classification of complex fractures of the proximal
humerus has long been an area of dispute,5,7,32,33 re-
flecting an inability to agree on the anatomy of these
injuries based on conventional radiographs alone. It
has recently been shown that 3-dimensional (3D) com-
puted tomography (CT) reconstructions, when viewed
in a systematic fashion, can yield superior understand-
ing and an enhanced concurrence among observers
on the nature of these fractures.11 This has led to
From the aOrthopedic Department and bRadiology Department,
Poriya Government Hospital, Tiberias, and cMRI/Radiology
Department, Rambam Medical Center, Haifa.
Reprint requests: Gordon Edelson, MD, Orthopedic Department,
Poriya Government Hospital, Tiberias, Israel (E-mail: Edelson@
bezeqint.il).
Copyright ª 2008 by Journal of Shoulder and Elbow Surgery
Board of Trustees.
1058-2746/2008/$34.00
doi:10.1016/j.jse.2007.08.014
a modification of the Neer classification24 of proximal
humeral fractures to reflect their true 3D anatomy.
A 3D understanding is important, in and of itself,
during any surgical reconstruction, but a classification
system is also useful insofar as it enhances other as-
pects of clinical decision making. For example, in
a particular category of injury, what, if any, surgery
is indicated? In this regard, one must first know the nat-
ural history of the specific fracture type without the
benefit of operative intervention.
To answer to this basic question, we categorized
nonoperative proximal humeral fractures based on
the 3D classification and have followed their clinical
progress. We present the natural history in nonopera-
tive patients with the more complex types of injury,
who historically have been the ones commonly treated
surgically.
MATERIALS AND METHODS
For purposes of this study, complex fractures were de-
fined as (1) displaced 3-part fractures, (2) shield fractures,
or (3) shield fracture variants based on the 3D classification
system (Figure 1, fractures colored in yellow and red). Pa-
tients were placed in these categories by the consensus of
3 orthopaedic surgeons after independently viewing 3D re-
constructions of each fracture from 4 different directions—
from the front, the side, the back, and overhead—a fracture
wheel format (Figure 2). Displacement was defined by the
Neer criterion of angulation greater than 1 cm or 45 be-
tween fracture fragments.24 These measurements were
more easily made on the 3D images than on standard radio-
graphs. Hence, all 3D radiographs were digitized, and the
angulation and displacement of fragments were calculated
by use of standard computer software. Measurements
were made in each of the 4 directions of the fracture wheel
(Figure 3), and the largest amount of displacement or angu-
lation between fragments was the measurement used.
From 1996 to 2005, all patients treated in our hospital
with comminuted fractures of the proximal humerus were ex-
amined with 3D CT reconstructions. Initial attempts to ran-
domize selection for operative versus conservative care
were unsuccessful in our patient population because patients
insisted on proactively choosing rather than being assigned
to a treatment group by lot. Consequently, we carefully ex-
plained the pros and cons of surgery versus nonoperative
treatment (avoiding commonly asked questions such as
‘‘What would you do for a member of your family?’’). The
399
400 Edelson et al J Shoulder Elbow Surg
May/June 2008

Figure 1 Diagram of the 5 basic types of proximal humeral fracture together with common variations in the 3D clas-
sification system. The major categories are (1) 2-part fracture, (2) 3-part fracture, (3) shield fracture, (4) isolated
greater tuberosity fracture, and (5) fracture-dislocation. Rare additional injuries not pictured are posterior fracture-
dislocations and isolated fractures of the lesser tuberosity. The colored fractures (red and yellow) comprise the com-
plex fractures with which this article deals.
J Shoulder Elbow Surg Edelson et al 401
Volume 17, Number 3

Figure 2 Fracture wheel, illustrating a shield type injury. The fracture wheel concept, viewing the injury from the
front, the side, the back, and overhead, gives a standardized and useful 3D understanding on a 2-dimensional page.

Figure 3 Example of measuring fragment displacement and angulation on each view of a 3D fracture wheel—in this
case using a 3-part injury. The largest displacement between fragments and the largest angle of fragment inclination
were the measurements used. Measurements were made with standard computer software. Fractures adjudged to be
displaced according to Neer’s criteria (see text) were categorized as complex and included in this study.

subsequent decision for conservative treatment or operation
was made at the behest of the patient or because surgery
was contraindicated because of major medical comorbid-
ities (recent heart attack, stroke, chronic dialysis, metastatic
disease, etc.) (9 patients). Exclusions were also made for
multiple trauma (4 patients) or electromyography-docu-
mented axillary nerve damage (3 patients).
During the investigation period, 126 patients with
complex fractures who met the inclusion criteria were seen
at the hospital. Of these patients, 76 were treated
402 Edelson et al J Shoulder Elbow Surg
May/June 2008

Table I Range of motion in different types of complex fractures compared with normal side

Displaced 3-part fracture Shield fracture Variants Shield and 3-part Normal side
(36 patients) (13 patients) (14 patients) (63 patients)

Forward flexion in scapular plain ( ) 117 6 27 108 6 28 106 6 22 157 6 19

External rotation with arm at side ( ) 43 6 23 33 6 18 31 6 21 69 6 18
Internal rotation with arm behind back T10 (63) L3 (64) L3 (63) T7 (62)
External rotation at 90 ( ) 56 6 24 44 6 29 40 6 32 84 6 16
Internal rotation at 90 ( ) 42 6 23 36 6 19 34 6 21 64 6 16

Data are given as mean 6 SD.

T, Thoracic vertebrae; L, lumbar vertebrae.

Table II Comparison of range of motion between complex fractures in

valgus, valgus, and neutral positions

Valgus Varus Neutral

(26 patients) (21 patients) (16 patients)

Forward flexion in 123 6 28 126 6 21 118 6 35

scapular plain ( )
External rotation with 38 6 22 43 6 28
arm at side ( )
Internal rotation with L1 (63) L1 (64)
arm behind back
External rotation 64 6 25 52 6 28
at 90 ( )
Internal rotation 39 6 24 34 6 21
at 90 ( )
Data are given as mean 6 SD.
L, Lumbar vertebrae; T, thoracic vertebrae.
conservatively. Of these conservatively treated patients, 13
died or were otherwise lost to sufficient follow-up, leaving
a final figure of 63 patients in this study.
In the final study group, 74% were women. The mean pa-
tient age was 66 years (range, 37-87 years). The mean fol-
low-up period was 32 months (range, 24-84 months). The
right shoulder was involved in 54%. The dominant arm
was involved in 57%. Shield or shield variant injuries com-
prised 46%. Complex 3-part fractures constituted 54%.
The operative group consisted of 50 patients. The mean
age (64 years; range, 42-83 years), sex (70% female),
and severity of injury (48% shield or shield variants) in the
operative group were similar to those of the nonoperative
group. The surgical group underwent a total of 6 different
procedures (percutaneous pinning, open screw-and-washer
fixation, standard plating, locked plating, Rush rod with fig-
ure-of-8 wiring, and hemiarthroplasty). These procedures
were performed by a number of different surgeons of vary-
ing experience. The subsequent diversity and small size of
the operative groups did not permit significant statistical
analysis for comparison with the nonoperative group. The
surgical patients, therefore, are not considered further in
this report.
Conservative care consisted of sling immobilization for 3
to 4 weeks, followed by active-assisted range-of-motion exer-
cises progressing to resisted strengthening at 3 months. Exer-
cises were done under the supervision of physical therapists.
Patients were seen by one or both of the senior authors for
39 6 21
T9 (63)
53 6 33
39 6 22
Figure 4 Graphic representation of visual analog pain scores at
final follow-up in complex humeral head fractures.
follow-up at 3 months, 6 months, and at 6-month intervals
thereafter during the study period. Standard anteroposterior
radiographs in neutral, internal rotation, and axillary or
modified Velpeau axillary views were obtained at each re-
turn visit. History and physical examination were
performed and a standardized questionnaire completed.
A validated testing instrument21 (Simple Shoulder Test
[SST]) and a visual analog pain scale ranging from 1 to
10 were administered by a physician to all patients.
Magnetic resonance imaging (MRI) studies looking for
avascular necrosis (AVN) were performed in the first 16
cases, 4 months after initial injury. This aspect of the study
was curtailed thereafter because of funding constraints,
but evidence of AVN continued to be carefully pursued
with standard radiographic studies.
Statistical methods
Continuous variables are presented as mean 6SD and
categoric variables as frequency and percentage. Paired t
tests were conducted to compare injured with noninjured
sides. Differences in range-of-motion modalities between
subgroups were analyzed by the independent t test for 2
groups and by the nonparametric Kruskal-Wallis ANOVA
test for more than 2 groups. Correlation between age and
range-of-motion performance was measured by Spearman
correlation. A 2-tailed P value <.05 was considered to be
significant. Comparison between nonoperative and opera-
tive groups regarding sex, handedness, and severity of in-
jury was analyzed by use of the c2 test (P < .05).
J Shoulder Elbow Surg
Volume 17, Number 3
Figure 5 Graphic presentation of results of SST in patients with complex fractures. A normal population of healthy
65-year-old patients usually answers yes to all 12 questions.13
Comparison of the 2 groups in terms of age was done with
a standard t test.
RESULTS
At final follow-up, mean motion was as follows. For-
ward flexion in the scapular plane on the injured side
was 112 6 37 . This represented 72% of the compa-
rable motion on the uninjured side. External rotation
with the elbow at the side in a mid-axillary plane
was 37 6 18 , which represented 53% of the normal
side. Internal rotation with the elbow at the side was to
the L1 (63) vertebral level compared with T7 (62) on
the normal side. External rotation and internal rotation
with the shoulder abducted to 90 in the scapular
plane were 56 6 28 and 36 6 21 , respectively.
This represented 60% and 55%, respectively, of the
normal side (Table I).
More extensive and comminuted fractures did mar-
ginally less well with regard to motion. As a result,
shield fractures and fracture variants tended to
achieve less motion than did 3-part injuries (Table I).
This bimodal tendency, however, rose to statistical sig-
nificance only for internal rotation (P >.001). Regard-
ing varus, valgus, or neutral position of the head
fragment, there was no significant difference in terms
of range of motion (Table II), again with the exception
of internal rotation, which favored the neutral position
(P ¼ .008). The mean time to maximum motion recov-
ery was 6.8 6 3 months.
The final pain score estimate for all fracture types on
the visual analog pain scale (which ranged from 1-10)
was 0.4 6 1.2 at rest and 2.8 6 2.3 with activity (Fig-
ure 4). This level was reached on average 5.2 6 2
months after injury.
Edelson et al 403
Results of the SST combined for all categories
are shown in Figure 5. Normal patients, aged 65
years, without shoulder pathology can perform all
12 functions.21 In our study, post-fracture patients
performed, on average, 9.79 6 2.45 functions at
final follow-up.
Of the 16 complex injuries examined by MRI, 14
showed no evidence of AVN (Figure 6, A and B).
Two of the sixteen did show early MRI evidence of lo-
calized avascularity (Figure 6, C), but neither of these
developed radiographic or clinical signs of AVN with
follow-up of greater than 3 years. Clinical and radio-
graphic AVN went on to develop in only 1 patient in
the overall study group of complex injuries—in this
case, a head-splitting shield type (Figure 7). This pa-
tient had not been examined with an MRI scan. First ev-
idence of AVN was noted on standard radiographic
views at 1 year after injury, progressing to complete
collapse by 3 years. At 4 years’ follow-up, she had
minimal discomfort, a functional range of motion,
and 9 of 12 on the SST examination. She did not
want further treatment.
Although malunions were universal among the
complex fractures, only 1 nonunion occurred, in a 3-
part injury. The patient was a 57-year-old woman
who had removed her sling after 1 week. The early re-
moval of the immobilization device may have contrib-
uted to her nonunion. Though subsequently painful,
she declined any surgical intervention.
Women did not have significantly different overall
motion compared with men. Younger patients, aged
under 55 years (9 patients), tended to have improved
motion, which rose to statistical significance in for-
ward flexion (P ¼ .003) and external rotation with
the arm at the side (P ¼ .037).
404 Edelson et al J Shoulder Elbow Surg
May/June 2008

Figure 6 Examples of MRI studies done in complex fractures 4 months after injury. Shield in varus (A) and head-split-
ting shield in valgus (B). Neither of these severe injuries showed MRI signs of AVN (middle images). Radiographic
follow-up at 3 years also showed no evidence of AVN or head collapse (right images). C, Complex 3-part injury in
neutral position. The middle image shows initial MRI evidence of localized AVN. However, there was no radio-
graphic or clinical evidence of AVN at 3.5-year follow-up (right image).

DISCUSSION ages are not usually obtained. However, the efficacy

To understand the natural history of a fracture, the of the 3D system has been favorably compared to
injury must be clearly defined. The 3D classification the Neer and AO classification systems.11 We have
used in this study is not widely used, because 3D im- found 3D images extremely valuable in arriving at
J Shoulder Elbow Surg Edelson et al 405
Volume 17, Number 3

a common language among practitioners as to the

Figure 7 Singular case of AVN in our study group. A, Initial head-
splitting shield injury. B, Radiograph at 6-year follow-up. C, Clinical
forward flexion at 6 years.
true nature of the fracture anatomy.
The Shield injury, seen well only in 3 dimensions, is
the Rosetta Stone for understanding complex fractures
of the proximal humerus. It is composed of the greater
and lesser tuberosities united by the dense bone of the
biceps groove. Although Nature’s variability is end-
less, there is an underlying continuity to the progres-
sion of severity in most fractures of the proximal
humerus,1,11 leading to complex injuries. The hard
bone of the glenoid is the anvil upon which the head
breaks. In the simpler 2-part injury, the major fracture
line is confined to the weak metaphyseal bone of the
surgical neck. In 3-part injuries, in addition to the sur-
gical neck, the greater tuberosity fractures. With in-
creased severity, the major fracture line proceeds
forward from the greater tuberosity to the lesser tuber-
osity, creating a full-blown Shield configuration. Fig-
ure 8 shows the evolution of a 3-part injury to
a shield injury. The Shield itself is then commonly split
internally into several fragments. Usually, comminu-
tion in the greater tuberosity predominates, but in se-
vere injuries, additional fracture lines open up to
produce the Shield variants we have noted (Figure 1).
By the outcome measures used in this study, we
have found a more benign natural history for these se-
vere injuries than is sometimes believed.8,25,34 Our
work confirms the natural history studies done by sev-
eral previous authors.27,37,38
Although malunion is universal, most patients even-
tually become relatively pain-free with sufficient mo-
tion and strength to perform basic activities of daily
living. Even in the most severe of these injuries, the
usual result of conservative treatment is similar to that
of a successful shoulder fusion15,16,30 (Nature’s fu-
sion) (Figure 9)—that is, minimal pain, forward flexion
to or slightly above shoulder level, and rotation depen-
dent in large part on compensatory scapular22 and el-
bow movement.35
Just as in shoulder fusions, limited motion after
malunion may be of major concern to some patients,
especially the young and active. There is no question
that judicious surgery in expert hands can yield
outstanding results superior to those of benign
neglect.12,26,28,31 However, the results of open
reduction–internal fixation or hemiarthroplasty are
unpredictable14,20,26 and may be no more satisfac-
tory than the natural history outcomes we have
demonstrated.
Regarding AVN, our results were also of interest,
especially in the shield fracture and its more severe
variants. In the majority of the existing literature, it is
proposed that the principal blood supply of the
humeral head is the anterior circumflex, which enters
the head at the upper lateral aspect of the bicipital
groove.10,13 The shield injury separates this area of
vascular ingress from the head. The shield is, in fact,
406 Edelson et al J Shoulder Elbow Surg
May/June 2008

Figure 8 Schematic and clinical illustrations from an overhead perspective showing progression of injury severity
from 3-part pattern to shield fracture. As the head is forced down and back against the anvil of the glenoid, the major
fracture line proceeds forward across the bicipital groove to include the lesser tuberosity. The shield thus formed (ar-
rowheads) commonly undergoes additional internal comminution as shown here. The surgical neck fracture, which is
present in both of these injuries, is not seen from the overhead view.

an anatomic neck fracture. It is usually not recognized
as such because of the foreshortened anatomy of the
neck in the proximal humerus. Shield injuries, and
particularly their comminuted variants, might thus be
expected to result in full or partial avascularity of the
head and an unacceptable clinical outcome.3,18 It is
based on this avascular hypothesis that a decision
for operation, particularly hemiarthroplasty, is usually
made.
It appears from our study that AVN, even subtle
forms demonstrable on MRI, may be more uncom-
mon than previously believed. We were unable to
find information in the literature regarding the timing
of MRI/AVN changes, specifically in the humeral
head after injury. Thus, on the advice of our experi-
enced MRI consultant, we proceeded on the assump-
tion that some MRI evidence of avascularity would be
discernible by 4 months after the initial devasculariz-
ing trauma. This contention is supported in the litera-
ture regarding MRI findings in femoral head
injuries.2,4,19 However, because the timing of MRI
evidence of AVN in the humeral head after fracture
has not been specifically reported in the literature,
this portion of our work should be approached with
due caution.
It may be that the blood supply to the humeral head
or the process of revascularization after injury may be
more variable and complex6,17 than presumed by the
limited number of cadaveric injection studies upon
which current concepts are based. It may be that a ma-
jority of AVN cases occur after surgical intervention
rather than from a lack of it. The relatively common
occurrence of AVN after open reduction–internal fixa-
tion has been noted by many authors.9,12,36 It was
unfortunate that, because of financial constraints,
MRI studies could not be done in all of our cases.
J Shoulder Elbow Surg Edelson et al 407
Volume 17, Number 3

Figure 9 Nature’s fusion. Typical clinical outcome in the types of severe shield and shield variant fractures are illus-
trated. Left to right, 3D views followed by range of motion in forward flexion, external rotation, and internal rotation.
408 Edelson et al
However, all cases were followed carefully for 2 years
or more (mean, 3.5 years) with standard 3-view serial
radiographs (anterior-posterior views in neutral and
internal rotation and an axillary view). Some evidence
of AVN will usually be detectable by 2 years after in-
jury9,19,23 with these type of radiographs, although
complete collapse may take longer to manifest.2
Parenthetically, there is no apparent difference in
outcome between complex valgus or varus injuries.
Current thinking holds that valgus injuries are in a cat-
egory by themselves with superior clinical outcomes
and less tendency toward AVN.29,31 However, this
supposition is as yet unproven by controlled or natural
history studies. Treated by conservative means, our
study suggests that the results of the 2 types of injury
are essentially the same (Table II).
One obvious criticism of this study is the potential
for selection bias between operative and nonopera-
tive cases. Surgical patients might have had worse
fractures that, if treated conservatively, would have
yielded more unfavorable outcomes than those we
demonstrate. This does not appear to be the case inso-
far as the overall number of severe fractures (shield
and shield fracture variants) did not differ significantly
between the operative and nonoperative groups. The
age and sex of the 2 groups were also comparable.
Nonetheless, selection bias could have been reduced
by a randomized protocol. We regret our inability to
achieve this. However, we still believe that the
information contributed by this study is worthy of
cautious consideration.
We hope that fully randomized natural history
outcome studies will be done in the future for all cate-
gories of proximal humeral fractures in the 3D classifi-
cation system. Importantly, answers also need to
follow regarding the outcome of different surgical
interventions in each 3D category so that comparisons
with the natural histories can be made.
We thank K. Williamson for graphics and Y. Schwartz for
help in preparing the manuscript.
REFERENCES
1. Aschauer E, Resch H. Four-part proximal humeral fractures. In:
Warner JJP, Iannotti JP, Flatow EL, editors. Complex and revision
problems in shoulder surgery. 2nd ed. Philadelphia: Lippincott;
2005. p. 289-311.
2. Berquist TH. Pelvis, hips, and thigh. In: Berquist TH, editor. MRI of
the musculoskeletal system. Philadelphia: Lippincott; 2001.
p. 195-291.
3. Bigliani LU. Fractures of the shoulder: part I. Fractures of the
proximal humerus. In: Rockwood CA Jr, Green DP, Bucholz RW,
editors. Fractures in adults. 3rd ed. Philadelphia: Lippincott;
1991. p. 871-927.
4. Bluemke DA, Zerhouni EA. MRI of avascular necrosis of bone. Top
Magn Reson Imaging 1996;8:231-46.
5. Brien H, Noftall F, MacMaster S, Cummings T, Landells C,
Rockwood P. Neer’s classification system: a critical appraisal.
J Trauma 1995;38:257-60.
J Shoulder Elbow Surg
May/June 2008
6. Brooks CH, Revell WJ, Heatley FW. Vascularity of the humeral
head after proximal humeral fractures. J Bone Joint Surg Br
1993;75:132-6.
7. Burstein AH. Fracture classification systems: do they work and are
they useful? J Bone Joint Surg Am 1994;76:792-3.
8. Compito CA, Self EB, Bigliani LU. Arthroplasty and acute shoulder
trauma. Clin Orthop Relat Res 1994:27-36.
9. Connor PM, Flatow EL. Complications of internal fixation of
proximal humeral fractures. Instr Course Lect 1997;46:25-37.
10. Duparc F, Muller JM, Freger P. Arterial blood supply of the
proximal humeral epiphysis. Surg Radiol Anat 2001;23:185-90.
11. Edelson G, Kelly I, Vigder F, Reis ND. A three-dimensional
classification for fractures of the proximal humerus. J. Bone Joint
Surg Br 2004;86:413-25.
12. Gerber C, Werner CML, Vienne P. Internal fixation of complex
fractures of the proximal humerus. J Bone Joint Surg Br 2004;86:
848-55.
13. Gerber C, Schneeberger AG, Vinh TS. The arterial vascularization
of the humeral head. An anatomical study. J Bone Joint Surg Am
1990;72:1486-94.
14. Goldman RT, Kaval KJ, Cuomo F, Gallagher MA, Zuckerman JD.
Functional outcome after humeral head replacement for acute
three- and four-part proximal humeral fractures. J Shoulder Elbow
Surg 1995;4:81-6.
15. Harryman DT II, Walker ED, Scott BS, Harris MS, Sidles JA,
Jackins SE, et al. Residual motion and function after glenohumeral
or scapulothoracic arthrodesis. J Shoulder Elbow Surg 1993;2:
275-85.
16. Hawkins RJ, Neer CS II. A functional analysis of shoulder fusions.
Clin Orthop Relat Res 1987:65-76.
17. Hertel R, Hempfing A, Stiehler M, Leunig M. Predictors of humeral
head ischemia after intracapsular fracture of the proximal humerus.
J Shoulder Elbow Surg 2004;13:427-33.
18. Hoffmeyer P. The operative management of displaced fractures of
the proximal humerus. J Bone Joint Surg Br 2002;84:469-80.
19. Lafforgue P. Pathophysiology and natural history of avascular
necrosis of bone. Joint Bone Spine 2006;73:500-7.
20. Mansat P, Guity MR, Bellumore Y, Mansat M. Shoulder
arthroplasty for late sequelae of proximal humeral fractures.
J Shoulder Elbow Surg 2004;13:305-12.
21. Matsen FA III, Lippitt SB, Sidles JA, Harryman DT II. Practical
evaluation and management of the shoulder. Philadelphia:
Saunders; 1994. p. 363-420.
22. McCully SP, Kumar N, Lazarus MD, Karduna AR. Internal and
external rotation of the shoulder: effects of plane, end-range
determination, and scapular motion. J Shoulder Elbow Surg
2005;4:602-9.
23. Naranja RJ, Iannotti JP. Displaced three- and four-part proximal
humerus fractures: evaluation and management. J Am Acad
Orthop Surg 2000;8:373-82.
24. Neer CS II. Displaced proximal humeral fractures. I. Classification
and evaluation. J Bone Joint Surg Am 1970;52:1077-89.
25. Neer CS II. Displaced proximal humeral fractures. II. Treatment of
three-part and four-part displacement. J. Bone Joint Surg Am 1970;
52:1090-103.
26. Plausinis D, Kwon YW, Zuckerman JD. Complications of humeral
head replacement for proximal humeral fractures. J Bone Joint
Surg Am 2005;87:204-13.
27. Rasmussen S, Hvass I, Dalsgaard J, Christensen BS, Holstad E.
Displaced proximal humeral fractures: results of conservative
treatment. Injury 1992;23:41-3.
28. Resch H, Povacz P, Frohlich R, Wambacher M. Percutaneous
fixation of three- and four-part fractures of the proximal humerus.
J Bone Joint Surg Br 1997;79:295-300.
29. Resch H, Beck E, Bayley I. Reconstruction of the valgus-impacted
humeral head fracture. J Shoulder Elbow Surg 1995;4:73-80.
30. Richards RR, Beaton D, Hudson AR. Shoulder arthrodesis with plate
fixation: functional outcome analysis. J Shoulder Elbow Surg
1993;2:225-39.
J Shoulder Elbow Surg
Volume 17, Number 3
31. Robinson CM, Page RS. Severely impacted valgus proximal
humeral fractures. J Bone Joint Surg Am 2004;86(Suppl 1, Pt 2):
143-55.
32. Sidor M, Zuckerman J, Lyon T, Koval K, Cuom F, Shoenberg N.
The Neer classification system for proximal humeral fractures: an
assessment of interobserver reliability and intraobserver reproduc-
ibility. J Bone Joint Surg Am 1993;75:1745-50.
33. Siebenrock K, Gerber C. The reproducibility of classification of
fractures of the proximal end of the humerus. J Bone Joint Surg
Am 1993;75:1751-5.
34. Stableforth PG. Four-part fractures of the neck of the humerus.
J Bone Joint Surg Br 1984;66:104-8.
Edelson et al 409
35. Wakabayashi I, Itoi E, Minagawa H, et al. Does reaching the
back reflect the actual internal rotation of the shoulder? J Shoulder
Elbow Surg 2006;15:306-10.
36. Wijgman AJ, Roolker W, Patt TW, Raaymakers MD, Marti RK.
Open reduction and internal fixation of three and four-part fractures
of the proximal part of the humerus. J Bone Joint Surg Am 2002;11:
1919-25.
37. Young TB, Wallace WA. Conservative treatment of fractures and
fracture-dislocations of the upper end of the humerus. J Bone Joint
Surg Br 1985;67:373-7.
38. Zyto K. Non-operative treatment of comminuted fractures of the
proximal humerus in elderly patients. J Injury 1998;29:349-52.