Shoulder hemiarthroplasty for acute fractures of the proximal humerus: a minimum five-year follow-up.

Abstract

Significant discrepancy in the functional outcome of hemiarthroplasty for proximal humerus fractures has been reported with short or mid-term follow-up. This study reports the long-term results and rate of complications of shoulder arthroplasty in the treatment of proximal humerus fractures. The review comprised 57 patients (44 women, 13 men) who underwent hemiarthroplasty between 1976 and 1996 as treatment of a proximal humerus fracture and who had a minimum 5-year follow-up (mean, 10.3 years). The mean patient age was 66 years at the time of surgery (range, 23-89 years). According to a modified Neer result rating system, results were satisfactory in 27 patients and unsatisfactory in 30. The mean active elevation was 100 degrees (range, 20 degrees -180 degrees ) and external rotation was 30 degrees (range, 0 degrees -90 degrees ). At the most recent follow-up, 9 patients (16%) had moderate or severe pain and 2 required implant revision or removal. The study data suggest that patients undergoing arthroplasty as treatment of an acute fracture of the proximal humerus may achieve satisfactory long-term pain relief; however, the result for overall shoulder motion is less predictable. In view of these results, current indications, surgical technique, and postoperative treatment may need to be revisited.

Full text

Page 1

Shoulder hemiarthroplasty for acute fractures of the proximal
humerus: A minimum five-year follow-up
Samuel A. Antun ˜a, MD,aJohn W. Sperling, MD, MBA,band Robert H. Cofield, MD,bPrincipado de Asturias, Spain,
and Rochester, MN
Significant discrepancy in the functional outcome of
hemiarthroplasty for proximal humerus fractures has
been reported with short or mid-term follow-up. This study
reports the long-term results and rate of complications of
shoulder arthroplasty in the treatment of proximal
humerus fractures. The review comprised 57 patients (44
women, 13 men) who underwent hemiarthroplasty
between 1976 and 1996 as treatment of a proximal
humerus fracture and who had a minimum 5-year follow-
up (mean, 10.3 years). The mean patient age was 66
years at the time of surgery (range, 23-89 years).
According to a modified Neer result rating system, results
were satisfactory in 27 patients and unsatisfactory in 30.
The mean active elevation was 100?(range, 20?-180?)
and external rotation was 30?(range, 0?-90?). At the
most recent follow-up, 9 patients (16%) had moderate or
severe pain and 2 required implant revision or removal.
The study data suggest that patients undergoing
arthroplasty as treatment of an acute fracture of the
proximal humerus may achieve satisfactory long-term
painrelief;however,theresultforoverallshouldermotion
is less predictable. In view of these results, current
indications, surgical technique, and postoperative
treatment may need to be revisited. (J Shoulder
Elbow Surg 2008;17:202-209.)
Since it was popularized by Neer,25shoulder hemi-
arthroplasty has become a standard treatment option
for complex fractures of the proximal humerus when
they are not amenable to fixation or the humeral
head fragment is not viable. The common indications
include selected 3-part fractures, 4-part fractures with
or without dislocation, and head-splitting fractures.
Although the
ported
some,5,8,9,13,17,20,21,25,27-29,31others have ques-
tioned these results.2-4,15,22,33,35Moreover, some au-
thors have reported functional outcome that was
comparable to that obtained with conservative treat-
ment.35However, most series reporting on arthro-
plasty for proximal humerus fractures are either small
or with a short follow-up, and many have an inade-
quate objective assessment. This study evaluated the
long-termoutcome ofpatients that underwent shoulder
arthroplasty as the treatment of a proximal humerus
fracture to determine the results, risk factors for an un-
satisfactory result, and rates of revision surgery.
initial satisfactory outcome re-
Neerhasbeenbyreproducedby
MATERIALS AND METHODS
BetweenMay1976andJune1996,85patientswhosus-
tained a fracture of the proximal part of the humerus were
treated with a primary hemiarthroplasty at the Mayo Clinic.
Of these 85 patients, 57 patients with a complete preopera-
tive evaluation, operative records, and a minimum 5-year
follow-up (average, 10.3 years; range, 5-22 years) or until
the time of the revision surgery were included in the study.
Ten were excluded because they had died of unrelated
causes before 2 years after surgery, and 7 were excluded
because they were lost to follow-up before 2 years. Eleven
had a follow-up of between 2 and 5 years (average, 27
months; range,24-31 months). None ofthese 11had under-
gone another surgical procedure on the shoulder, and all
were satisfied with the result with no or slight pain at the
time of the last clinical follow-up.
Of the 57 patients followed up for a minimum of 5 years,
25 were followed up between 5 and 7 years, 6 between 7
and 10 years, and 26 for more than 10 years. Two patients
required revision surgery6 and 11 yearsfrom the time ofthe
index operation.
There were 44 women and 13 men, their mean age was
66 years (range, 23-89 years), and the dominant extremity
wasinvolvedin41.Themechanismofinitialinjurywasafall
fromstandingheightin50patients,amotor vehicleaccident
in 5, and a seizure in 1 with bilateral fractures. One patient
had a preoperative brachial plexus palsy that resolved by 6
weeks after surgery. The fractures were classified by Neer’s
classification of proximal humerus fractures.24Seven pa-
tients had a 3-part fracture, 32 had a 4-part fracture, 4
had a 3-part fracture dislocation, 9 had a 4-part fracture
and dislocation, and 5 had a head-splitting fracture.
FromtheaDepartmentofOrthopedicSurgery,HospitalValledelNa-
lo ´n-University of Oviedo, and thebDepartment of Orthopedic Sur-
gery, Mayo Clinic.
Reprint requests: John W. Sperling, MD, Mayo Clinic, 200 First St
SW, Rochester, MN 55905 (E-mail: sperling.john@mayo.edu).
Copyright ª 2008 by Journal of Shoulder and Elbow Surgery
Board of Trustees.
1058-2746/2008/$34.00
doi:10.1016/j.jse.2007.06.025
202

Page 2

The average time from the initial injury to arthroplasty
was 3 days (range, 1-30 days). Five patients had addi-
tional fractures at the time of the injury: 1 femoral neck frac-
ture that required a bipolar arthroplasty, 1 ankle fracture
that required open reduction and internal fixation, 1 clavic-
ular fracture that was fixed at the time of the arthroplasty, 1
undisplaced C2 fracture treated conservatively, and 1 small
glenoid rim fracture that did not require fixation. All pa-
tients reported normal shoulder function before the injury.
None had undergone any previous surgical procedure
or reported any significant prior trauma of the injured
shoulder.
The clinical assessment of all patients undergoing shoul-
der surgery at our institution is recorded using a standard
shoulder analysis sheet. Pain is graded on a 1-5 scale: 1
point if there is no pain, 2 for slight pain, 3 for occasional
moderate pain after unusually vigorous activities, 4 for mod-
erate pain, and 5 points for severe pain. Patient satisfaction
was graded as excellent (1 point), good (2 points), satisfac-
tory (3 points), or unsatisfactory (4 points). Active elevation
and external rotation were recorded in degrees. Internal ro-
tation was measured as the posterior spinal segment the
thumb could reach. The results were further evaluated by
age and type of fracture.
A modified Neer result rating system was used.30The re-
sult was considered excellent if the patient had no pain or
slight pain, was satisfied with the procedure, and had at
least140?ofactiveelevationandexternalrotationofatleast
45?. A satisfactory result was no pain, slight pain, or occa-
sional moderate pain only with vigorous activities, the pa-
tient was satisfied with the procedure, and had at least
90?of active elevation and external rotation of at least
20?. If any of the criteria were not met, the result was unsat-
isfactory. Anyone who underwent an additional operation,
including implant revision or removal, was considered to
have an unsatisfactory result.
At the time of follow-up, clinical information on 22 pa-
tients was obtained from the notes recorded by the treating
surgeon on routine clinic visits. The information for 15 pa-
tients was obtained from the local orthopedic surgeon.
The remaining 20 patients returned a standard shoulder
questionnaire.
The preoperative, initial postoperative, and most recent
radiographs were evaluated by 3 observers, and consensus
was reached. The projections for radiographic analysis
were a 40?posterior oblique view with external rotation of
the humerus, a 40?posterior oblique view with internal rota-
tion of the humerus, and an axillary radiograph. Minimum
radiographic follow-up was 5 years for 35 patients, be-
tween 2 and 5 years in 9 patients, and 13 did not have ad-
equateradiographs.Only
radiographic follow-up exceeding 5 years were included.
The radiographs were reviewed to determine the pres-
ence or absence of glenohumeral subluxation, peripros-
thetic radiolucency, shift in the position of the component,
loss of glenoid cartilage, and the presence of bony erosion
of the glenoid:
patients withcomplete
?
Periprosthetic lucency was grade 0 if there was no
radiolucent line, grade 1 if the line was 1 mm or
less and incomplete, grade 2 if the line was 1 mm
wide and complete, grade 3 if the line was 1.5
mm wide and incomplete, grade 4 if the line was
1.5 mm wide and complete, and grade 5 if the
line was 2 mm wide and complete.
Glenohumeral subluxation was evaluated for direc-
tion and the amount of translation of the center of
the prosthetic head relative to the center of the gle-
noid. It was recorded as none, mild if there was
lessthan 25%translation, moderateifthe translation
was 25% to 50%, and severe if the translation was
more than 50%.
Erosion of the glenoid was graded as none, mild
(bony contact between the humeral head and the
glenoid with radiographic sclerosis of the glenoid
subchondral bone), moderate (medialization of the
glenoid subchondral bone with mild hemispheric
conforming deformation of the glenoid), or severe
(complete hemispheric deformation of the glenoid
with superior and inferior osteophyte formation
and central bone loss within 5 mm of the coracoid
base).
Humeral component subsidence was either present
or absent and was measured in millimeters.
?
?
?
Tuberosityreconstructionwasspecificallyassessed radio-
graphically. The vertical distance between the greater tuber-
osity and the proximal edge of the prosthetic humeral head
was measured in millimeters. The position of the greater tu-
berosity was considered adequate if the distance was be-
tween 5 and 20 mm below the humeral head according to
previous anatomic studies.10The presence of malunion,
nonunion, or resorption of the tuberosity was analyzed in
the last radiograph. The distance between the acromion
and the humeral head was also recorded as an index of su-
perior migration of the humeral head.
Operative technique
Allprocedureswereperformedby,orassistedby,1of20
surgeons who were responsible for the trauma patients at
our institution. Only 4 of the 20 surgeons performed more
than 5 hemiarthroplasties.
In 50 shoulders, a long deltopectoral exposure with the
anterior deltoid origin left intact was used. In 7 shoulders,
wider exposure was deemed necessary, and the anterome-
dial approach (deltopectoral exposure plus release of the
clavicular and anterior acromial origins of the deltoid) was
used. After identification of the fracture pattern and confir-
mation that a hemiarthroplasty was required, the proximal
humerus was prepared for implantation of the prosthesis.
The humeral component was press-fit in 8 shoulders and
was inserted with bone cement in 49 shoulders. A total of
33Neerhumeralcomponents(3M,StPaul,MN)and24Co-
field humeral components (Smith and Nephew, Memphis,
TN) were used.
Tuberosity fixation was performed with 20-gauge stain-
less-steel wire in 5 patients or heavy No. 5 nonabsorbable
suture in 52. According to the surgical records, sutures in
38 patients were placed both horizontally (to perform a se-
cure fixation of the tuberosities to the prosthesis) and verti-
cally (to fix the tuberosities to the shaft of the humerus). In
19 patients, only horizontally placed sutures were used.
Additional procedures at the time of shoulder arthro-
plasty included acromioplasty in 4 patients and clavicular
osteosynthesis in 1.
J Shoulder Elbow Surg
Volume 17, Number 2
Antun ˜a, Sperling, and Cofield203

Page 3

Postoperatively, the arm was placed in a shoulder immo-
bilizer. The physical therapy program started with passive
motion exercises within the first 48 hours and during a vari-
ableperiodbetween3and6weeks,followedbyastandard
active-assisted motion program progressing to isometric
strengthening.
Statistical analysis
All calculations and data analysis was done with SPSS
12.0 statistical software (SPSS, Chicago, IL). The Wilcoxon
rank sum test (2 groups) or analysis of variance (3 or more
groups) were performed to detect differences between
groups of patients. For all analyses, the level of significance
was set at P <.05.
RESULTS
Complications and reoperations
Three of the 57 shoulders required another opera-
tion, with 2 having component revision or removal.
One patient sustained an acute posterior dislocation
immediately after surgery that was attributed to defi-
ciency in tuberosity fixation and required another sur-
gical procedure to reattach the tuberosities. Five years
after the second operation, the shoulder remained sta-
ble, had limited motion, and was moderately painful
with unusual activities.
A 23-year-old patient with severe limitation of
range of motion and moderate shoulder pain under-
went a greater tuberosity osteotomy for repositioning,
scar tissue removal, and repair of a rotator cuff tear.
Nine months after the second procedure, shoulder
function was not improved, and radiographic evi-
dence showed tuberosity resorption. A third operation
was performed to transfer the pectoralis major superi-
orly to gain shoulder flexion, with poor results. Six
years after the index procedure, the arthroplasty was
removed and a shoulder fusion was performed.
One patient with an uncemented hemiarthroplasty
underwent revision to a cemented total shoulder
arthroplasty 11 years after the first operation owing
to symptomatic humeral component loosening and
glenoid arthritis (Figure 1).
Three patients, all with limited motion and moder-
ate pain or pain after unusual activities, had sub-
acromial injections due to impingement or rotator
cuff pathology, with only temporary relief.
Pain
At the most recent follow-up, 48 patients (84%) had
no pain, mild pain, or pain only after unusual activi-
ties; 7 had moderate pain; and 2 had severe pain.
One of the 2 with severe pain was revised 11 years
postoperatively due to humeral component loosening,
and the other refused further surgery because of se-
vere medical problems. One of the 7 patients with
moderate pain required 2 additional surgical proce-
dures and finally underwent a shoulder fusion,
4 had dementia, and the other 2 refused another
operation.
With the numbers available, we could not detect
a significant difference in pain relief with regard to
the patient’s age, type of fracture, or tuberosity
fixation technique (Table I).
Range of motion
At the most recent follow-up, the average active
elevation was 100?(range, 20?-180?), and external
rotation was 30?(range, 0?-90?). Average internal
rotation was the ability of the thumb to reach L5
(range,abdomen-T11).Anactivepostoperativeeleva-
tion of at least 90?was achieved by 33 of the 57
patients.
With the numbers available, we could not detect
a significant difference in postoperative range of
motion with regard to type of fracture or tuberosity
fixation technique. Patients younger than 70 years
had significantly better shoulder elevation than older
patients: 110?vs 89?(P ¼.038), with no significant
difference in external rotation.
Patient satisfaction
According to patients’ self-assessments, 28 shoul-
ders were graded as excellent, 16 as good, 2 as
satisfactory, and 11 as unsatisfactory; of these, 2
underwent additional surgery, and 9 had moderate
to severe shoulder pain. With the numbers available,
we could not detect a significant difference in patient
satisfaction with regard to patient age, type of
fracture, or tuberosity fixation technique.
Radiographic analysis
The immediate postoperative radiograph for 46 of
the 57 shoulders showed a well-positioned greater
tuberosity, defined as a greater tuberosity-humeral
head distance of 5 to 20 mm. In 7 shoulders, the
greatertuberositywas
(greater tuberosity-humeral head distance more than
20 mm). In 4 shoulders, the greater tuberosity was
superiorly malposition above the humeral head. The
average distance between the acromion and the hu-
meral head was 10.8 mm on the postoperative radio-
graph (range, 2-32 mm). Fourteen components were
slightly malpositioned; 5 had been implanted in varus
and 9 in valgus. Five cemented humeral components
showed periprosthetic lucency after surgery; it was
less than 1 mm and incomplete in 4 shoulders and
1.5 mm wide and complete in 1 shoulder.
At the time of most recent radiographs, some
degree of glenohumeral subluxation was present in
26 of the 35 shoulders (85%) that could be evaluated
inferiorlymalpositioned
204 Antun ˜a, Sperling, and Cofield
J Shoulder Elbow Surg
March/April 2008

Page 4

radiographically. Eighteen shoulders had isolated
superior subluxation that was mild in 6, moderate in
8, and severe in 4. Six shoulders had isolated anterior
subluxation, noted as mild in 4 and moderate in 2.
Two shoulders had mild posterior subluxation.
Humeral periprosthetic lucency was present at the
most recent follow-up in 13 of 31 shoulders with
a cemented humeral component. It was grade 1 in 5
shoulders, grade 3 in 7, and grade 4 in 1. One of 4
shoulders that received uncemented components
presented with grade 4 periprosthetic humeral
lucency andrequiredrevisionsurgeryforsymptomatic
loosening.
Narrowing of the joint space occurred in 15 of 35
shoulders, of which 2 had evidence of cartilage wear
without bony erosion, 8 had mild bony glenoid
erosion, 4 had moderate glenoid erosion (1 was re-
vised to a total shoulder arthroplasty), and 1 shoulder
had severe superior glenoid erosion.
Bony union of the greater tuberosity in a good posi-
tion was achieved in 22 of 35 shoulders (Figure 2).
The tuberosity progressed to nonunion in 1 shoulder,
the tuberosity malunited in 3, and the tuberosity re-
sorbed in 9 (Figure 3). All shoulders with nonunion
or resorption had an unsatisfactory result. The aver-
age distance between the acromion and the humeral
head at the most recent follow-up was 5.6 mm (range,
0-16 mm), demonstrating a clear tendency for supe-
rior humeral head displacement due to rotator cuff in-
sufficiency with time.
Rating results
Results were excellent in 6, satisfactory in 21, and
unsatisfactory in 30. Two unsatisfactory results were
due to additional surgery, including implant revision
or removal. In the other 28 shoulders, reasons for an
unsatisfactory result were moderate or severe pain
Figure 1 A and B, A 64-year-old patient with a 4-part fracture of the proximal humerus. C and D, In these radio-
graphs taken 11 years after humeral head replacement, severe glenoid bony erosion and humeral uncemented
stemlooseningcanbeobserved.Thepatienthadsevereshoulder pain,activeelevationof80?, andexternalrotation
to 0?. E and F, The patient underwent revision of the humeral stem and implantation of a glenoid component.
J Shoulder Elbow Surg
Volume 17, Number 2
Antun ˜a, Sperling, and Cofield205

Page 5

associatedwithlimited motion in8patients,lackofac-
tive elevation beyond 90?in 9, lack of active external
rotation exceeding 20?in 4, and limitation of both ac-
tive elevation and external rotation in 7.
We compared the clinical information available 2
years after the surgical procedure with that obtained
at the last follow-up to identify changes in the clinical
outcome over time. All patients with excellent or satis-
factory results at the latest follow-up had the same re-
sult at 2 years. Of the 30 patients with an
unsatisfactory result at the latest clinical evaluation,
25 were categorized as unsatisfactory 2 years after
surgery, 3 were considered satisfactory but shoulder
pain developed subsequently, and 3 had excellent re-
sults at 2 years but severe pain developed, and 1 re-
quired revision for aseptic loosening.
DISCUSSION
Treatment of complex fractures of the proximal hu-
merus is a challenging problem. Hemiarthroplasty
has become a standard procedure in most orthopedic
units when the humeral head is nonviable or
Table I Results by age, method of tuberosity reconstruction, type of fracture, and greater tuberosity healinga
Variable
Follow-up
(months)
Pain
(points)
Active elevation
(deg)
Active external
rotation (deg)
Satisfaction
(points)
Age, years
>70
<70
P
Tuberosity reconstruction
Horizontal
Vertical + horizontal
P
Type of fracture
3-part
4-part
3-part, dislocation
4-part, dislocation
Head split
P
Greater tuberosity
Adequate healing
Nonunion-resorption
P
102 6 55.5
141 6 62.1
.022
2.2 6 1.2
2 6 1.1
NS
89 6 36.2
110 6 43.9
.038
29 6 19.2
31 6 24.7
NS
2.1 6 1.2
1.7 6 1
NS
109 6 52.1
128 6 63.8
NS
2.1 6 1.3
2.1 6 1.1
NS
108 6 43.9
97 6 40.9
NS
33 6 24.9
30 6 20.7
NS
2.1 6 1
1.7 6 1.1
NS
90 6 35
124 6 62.3
114 6 68.9
140 6 71.9
133 6 65.7
NS
1.5 6 1.6
2.1 6 1.1
1.5 6 0.5
2.9 6 1.2
1.6 6 0.5
NS
92 6 28.7
98 6 40.8
102 6 45.7
90 6 45.9
146 6 34.3
NS
34 6 5.3
29 6 23.7
36 6 17.5
27 6 28.9
37 6 20.4
NS
1.4 6 1.1
1.9 6 1
1.5 6 1
2.6 6 1.2
1.2 6 0.4
NS
141 6 58.4
105 6 68.8
.036
2 6 1.1
2.4 6 1.5
NS
107 6 36.2
69 6 25
.002
29 6 21.6
22 6 16.7
NS
1.8 6 1.1
2.1 6 1.2
NS
NS, Not significant.
aThe values are given as the mean 6 the standard deviation.
Figure 2 A, A 59-year-old patient with a 4-part fracture of the proximal humerus. B, A postoperative radiograph
showsadequatereconstructionofthegreatertuberosity.C,At11yearsaftersurgery,thepatienthasmildpain,active
elevation of 90?, and external rotation of 15?. The radiograph shows healing of the tuberosity in good position and
mild glenoid cartilage erosion.
206Antun ˜a, Sperling, and Cofield
J Shoulder Elbow Surg
March/April 2008