The management of displaced proximal humeral fractures is challenging and
often reflects the personal experience of the physician treating the injury.
Regardless of the treatment protocol that is employed, these fractures pose
difficulties in terms of restoring humeral alignment, joint surface congruity,
and rotator cuff function while maintaining humeral head vascularity.
Treatment options include nonoperative methods, open reduction and internal
fixation, and hemiarthroplasty. However, each is associated with a high risk
of complications. Nonoperative treatment is preferable for minimally displaced
or impacted fractures, but the results may be poor as the fracture fragments
can become
displaced1,2.
Hemiarthroplasty is most often considered for four-part fractures, fractures
in which the head is split, or
fracture-dislocations3,4.
In most series, the outcomes of hemiarthroplasty have been unpredictable and
at times
disappointing5,6.
Open reduction and internal fixation with standard or modified plates can
provide favorable
results7-9.
However, hardware failure, infection, and osteonecrosis are risks, especially
when the fracture is comminuted or associated with severely osteopenic
bone10. Current
trends in the operative treatment of proximal humeral fractures have been
toward less invasive techniques of reconstruction, with limited soft-tissue
dissection and minimal amounts of hardware to achieve stable internal
fixation11-18.
Lately, low-profile locking plates and intramedullary nails have been
introduced, with promising early
results19-21.
Over the last fourteen years at our institution, we have used transosseous
suture fixation for a large number of displaced proximal humeral fractures.
Predominantly, these injuries have been four-part valgus impacted fractures,
three-part fractures, and two-part fractures of the greater tuberosity with or
without associated dislocation of the humeral head. The purpose of this study
was to present the radiographic and clinical outcomes in a series of 165
consecutive patients who had been treated with this technique at our
institution.
Between 1993 and 2003, 443 patients with an acute, displaced fracture of
the proximal part of the humerus were treated surgically in the Shoulder and
Elbow Unit at Patras University Hospital, Greece. The indication for
transosseous fixation was an isolated, displaced two-part fracture of the
greater tuberosity with or without anterior dislocation of the humeral head, a
three-part fracture, or a four-part valgus impacted fracture.
The contraindications for suture fixation included a two-part fracture of
the surgical neck, a displaced four-part fracture (>7 mm of translation or
an impaction angle of >45°), and a fracture associated with a
dislocation of the humeral head or a head-splitting component.
One hundred and eighty-eight patients met these criteria and were treated
with transosseous fixation with nonabsorbable number-5 Ethibond sutures
(Ethicon, Edinburgh, United Kingdom). Of the remaining 255 patients,
eighty-four had a displaced two-part surgical neck fracture, which was treated
with plate-and-screw osteosynthesis, and 171 had a displaced four-part
fracture, a three or four-part fracture-dislocation, or a head-splitting
fracture. One hundred and thirty-six of those 171 patients underwent primary
shoulder hemiarthroplasty and thirty-five young patients (an age of less than
thirty-nine years) underwent transosseous suture fixation (a so-called
head-preserving subgroup) to avoid the consequences of a primary
hemiarthroplasty. Those thirty-five patients, however, were not included in
this study.
Of the 188 patients who met the criteria for transosseous suture fixation,
twenty were not evaluated because of inadequate follow-up and three died of
unrelated causes before the time of follow-up. Thus, 165 patients were
available for assessment. Ninety-four of the patients were women, and
seventy-one were men. The mean age was fifty-four years (range, eighteen to
seventy-five years). All fractures were classified according to the
four-segment classification system of
Neer1. Forty-five
(27%) of the injuries were four-part valgus impacted fractures; sixty-four
(39%) were three-part fractures; and fifty-six (34%) were isolated fractures
of the greater tuberosity, thirty-six (64%) of which were associated with
anterior dislocation of the shoulder.
Ninety-eight patients (59%) sustained the fracture as a result of a fall
from a standing or low height. The remainder sustained a high-energy injury:
forty-five were injured in a motor-vehicle accident; eight, in a fall from a
height of >2 m; and fourteen, in a sports-related accident. One hundred
(61%) of the fractures occurred in the dominant arm.
Five patients had a clinically detectable neurological deficit
preoperatively. It was attributed to a brachial plexus injury in two of these
patients, to an axillary nerve paresis in two others, and to an isolated
radial nerve deficiency in one. Eighty (48%) of the patients had been
regularly employed prior to the injury: thirty-five had had a sedentary job,
and forty-five had performed manual work.
A standard radiographic trauma series of the shoulder (an anteroposterior
radiograph in the scapular plane as well as lateral and axillary radiographs)
was made for all patients as soon as there was clinical suspicion of a
fracture. In order to reduce discomfort, the axillary radiograph was usually
made with the patient in a supine position, under the supervision of the
attending physician. Additional radiographic views (anteroposterior in
external rotation or in internal rotation with 15° of cephalic tilt) and
computed tomography scans were made for forty of the 165 patients for further
evaluation of the degree of fracture displacemaent.
Surgical Technique
One hundred and fifty-five (94%) of the operations were performed within
five days after the injury; the remaining ten patients, who had been referred
from other hospitals, underwent surgery within three weeks after the injury.
The original technique was developed by the senior author (P.D.), who
performed or supervised 145 of the 165 procedures.
After induction of general anesthesia, the patient was placed in the
beach-chair position with at least 60° of flexion at the waist. Two folded
sheets were placed behind the scapula to bring the shoulder girdle forward,
facilitating access to the glenohumeral joint. A second-generation
cephalosporin was administered preoperatively and for the first postoperative
day. The entire upper extremity was prepared and draped in a manner that
allowed full and unrestricted arm positioning during the procedure.
Fixation of Four-Part Valgus Impacted Fractures
(Figs. 1-A and 1-B)
The fracture area was exposed through a lateral transdeltoid approach,
between the anterior and middle portions of the muscle. The skin incision was
made from the anterolateral tip of the acromion and extended approximately 6
to 7 cm distally. With use of blunt dissection, the deltoid was split for 4 to
5 cm. Rotation and abduction of the proximal part of the humerus in this
surgical window allowed adequate visualization of both tuberosities and of the
metaphyseal area, thus minimizing the risk of iatrogenic injury to the
axillary nerve. In cases with metaphyseal extension, the nerve was identified
and protected by the surgeon's finger.
Soft-tissue attachments to the fracture fragments were carefully preserved
to prevent devascularization of the humeral head. The fracture lines between
the tuberosities were identified and gently separated, facilitating access to
the humeral head. Invariably, the humeral head was facing superiorly with the
tuberosities splayed on either side of it. In every case, both tuberosities
were displaced by >1 cm. Initially, two stay sutures were passed through
each tuberosity fragment (or near its site of tendon insertion in osteoporotic
bone or when extensive comminution was present), and the rotator cuff tendons
were mobilized. Without disimpacting the articular head fragment from its
valgus position, two pairs of sutures were passed transosseously, 1 cm
proximal to the fracture line at both the medial and the lateral border of the
articular head fragment. Finally, another two pairs of sutures were inserted
laterally and medially through 2.7-mm drill holes in the diaphysis. These
sutures were directed into the opposite tuberosity, near the musculotendinous
junction, and onto the neighboring area of the articular segment (i.e.,
through the medial aspect of the diaphysis toward the greater tuberosity and
through the lateral aspect of the diaphysis toward the lesser tuberosity as
well as onto the adjacent articular fragment). Once all sutures were in place,
the tuberosities were approximated to the diaphysis, just below the top of the
head fragment. Each suture was then tied individually, and then to each other,
in a cruciate arrangement that allowed secure fixation of each part to the
others. Further loosening of the sutures, due to fracture compression, was
corrected by tying additional knots between the free sutures, once more in a
cruciate manner. Ten patients (22%) had associated or coexisting tears of the
rotator cuff tendons, which were also repaired with nonabsorbable sutures.
Once the fracture had been repaired, the adequacy of the fixation was
verified by gentle mobilization of the shoulder through 90° of forward
elevation and 30° of both external and internal rotation. Fixation was
considered to be stable if the repaired humeral head and the diaphysis moved
as a single unit. Intraoperative imaging was employed for the initial
procedures. As the experience of the surgeon increased, he checked for
abnormal motion by means of observation and palpation of the fracture site
alone. The deltoid flaps were then reapproximated with use of absorbable
sutures in a figure-of-eight manner. The subcutaneous tissue was closed with
absorbable sutures, and the skin was closed with a subcuticular technique. A
Velpeau dressing secured the arm to the chest wall; it was converted to a
simple sling on the second postoperative day.
Fixation of Three-Part Fractures
(Figs. 2-A and 2-B)
The same principles of fixation were used for three-part fractures. Again,
the sutures were placed through the displaced greater or lesser tuberosity,
the humeral head with the intact lesser or greater tuberosity, and the
diaphysis. In this type of fracture, the humeral head typically is rotated
either internally or externally, and attention should be paid to achieving an
adequate reduction in both the frontal and the sagittal plane. Eighteen
patients (28%) with a three-part fracture had associated rotator cuff tears,
which were repaired with nonabsorbable sutures.
Fixation of Two-Part Fractures of the Greater Tuberosity
(Figs. 3-A and 3-B)
When the fracture of the greater tuberosity was associated with an anterior
dislocation of the shoulder, the patient was lightly sedated in the emergency
department to facilitate reduction. One or a maximum of two efforts were made
to reduce the dislocation by closed means. If closed reduction failed, the
patient was transferred to the operating theater for reduction under general
anesthesia. Twenty-nine (52%) of the fifty-six patients with a two-part
fracture of the greater tuberosity had a characteristic longitudinal tear in
the rotator cuff, through the rotator interval. Five of these patients had a
complete avulsion of the supraspinatus tendon, three had a complete avulsion
of the infraspinatus tendon, and two had a combined avulsion of both tendons.
The cortical edge of the tuberosity fragment was reduced to align with the
edge of the fracture bed on the proximal part of the humerus. Two or three
pairs of sutures were then passed through the bone, in the upper and lower
portions of the tuberosity fragment, and then through corresponding drill
holes in the humerus. The sutures were tied in a cruciate fashion with care
taken to prevent overreduction or further comminution. The longitudinal tears
in the rotator cuff were also repaired with nonabsorbable sutures.
Rehabilitation Protocol
All patients participated in a closely monitored three-phase rehabilitation
program. Pendulum exercises were started on the second postoperative day and
continued until the third or fourth postoperative week. The second phase
included active-assisted range-of-motion exercises for five to ten weeks. In
the final phase, active dynamic shoulder motion and strengthening exercises
were prescribed until the sixth postoperative month.
Outcome Assessment
The functional outcome was assessed independently by two senior
physiotherapists using the parameters of the Constant score. Pain, performance
of daily activities, range of motion, and strength were scored on a scale of 1
to 100 points, with 100 points being an excellent score. The isometric power
of the shoulder was assessed according to a scale that awarded a maximum of 25
points when a patient could hold a weight of 12 kg at 90° of shoulder
abduction or if the maximum weight that he or she could hold in this position
was <12 kg but was equal to the greatest weight that could be held on the
contralateral, uninjured side.
Standardized anteroposterior and axillary radiographs of the shoulder were
made postoperatively; at one, three, six, and twelve months; and at the last
visit. These radiographs were assessed by an independent specialist registrar
(G.K.) for the quality of the reduction, progression of healing, malunion,
nonunion, heterotopic ossification, and signs of subacromial impingement,
humeral head osteonecrosis, or posttraumatic osteoarthritis.
On the basis of two measures—varus or valgus deformity of the head
with respect to the humeral shaft (in degrees) and residual displacement of
the greater tuberosity (in millimeters)—the quality of the reduction on
the first postoperative radiograph was classified as excellent/very good (mild
deformity [a varus or valgus angle of <20° and displacement of the
greater tuberosity of <3 mm]), good (moderate deformity [a varus or valgus
angle of between 20° and 40° and displacement of the greater
tuberosity of between 3 and 6 mm]), or poor (severe deformity [a varus or
valgus angle of >40° and displacement of the greater tuberosity of
>6 mm]). At the final radiographic assessment, persistent malunion was
classified as mild, moderate, or severe with use of the same criteria as was
used for the rating of the postoperative quality of the reduction. Fracture
union was considered to have occurred when the patient reported no shoulder
pain or activity-related discomfort and the fracture lines were no longer
visible on the plain radiographs. Inherent to this definition was the absence
of evidence of loss of reduction when compared with the reduction seen on
previous postoperative images. Heterotopic bone formation was graded with a
modification of the system of Brooker et
al.22. Class I
indicated islands of bone in the soft tissues; class II, at least 1 cm of
space between apposing bone surfaces; class III, <1 cm of space between
apposing bone surfaces; and class IV, complete bridging. Subacromial
impingement was demonstrated radiographically by migration or flattening of
the greater tuberosity, subacromial erosion, calcification, or malunion of the
humeral head. However, the diagnosis of this condition depended largely on
clinical parameters, such as a positive Neer test, a painful arc of abduction,
or a drop arm sign. Posttraumatic osteoarthritis was graded according to the
Kellgren-Lawrence
scale23 as 0 (no
features of osteoarthritis), 1 (questionable osteophyte development), 2
(definite osteophyte observed with a minimal loss of joint space), 3 (moderate
loss of joint space width), or 4 (severe loss of joint space width with
subchondral bone sclerosis). Humeral head osteonecrosis was defined by
destruction of the trabecular architecture with loss of osseous substance in
an articular segment. It was considered to be complete if the entire humeral
head segment had been reabsorbed and partial if only a portion of the humeral
head had collapsed.
The average duration of the surgery was seventy minutes (range, fifty-two
to 105 minutes), and the average duration of the hospital stay was 4.7 days
(range, 3.5 to seven days). The average follow-up period was 5.4 years (range,
three to eleven years). All neurological deficits accompanying the initial
injury resolved within four to six months after the surgery, except for one
brachial plexus injury that persisted after fracture consolidation. No
iatrogenic injuries of the axillary nerve supply to the anterior deltoid were
noted at the time of final follow-up. Three postoperative superficial
infections were managed with oral antibiotics, and a deep infection developed
five weeks after the operation in one patient. It was treated effectively with
surgical débridement and drainage, and the fracture united
uneventfully.
Functional Outcome (Table
I)
At the time of final follow-up, the mean Constant score was 91 points
(range, 54 to 100 points) and the mean Constant score as a percentage of the
score for the unaffected shoulder, unadjusted for age or gender, was 94%
(range, 60% to 100%). The average amount of active elevation with the patient
standing was 167° (range, 140° to 180°), and the average amount of
external rotation was 75° (range, 50° to 90°). The mean amount of
internal rotation, measured as the highest posterior spinal segment reached by
the thumb with the arm in adduction, was T10 (range, L3 to C7).
Three patients reported stiffness and compromised shoulder motion, which
was consistent with a diagnosis of adhesive capsulitis. Each of these patients
had cooperated poorly with the rehabilitation protocol. Despite an intensive
course of physiotherapy, the symptoms decreased in only one of these patients;
the remaining two patients declined additional treatment.
One hundred and forty-four patients (87%) were satisfied with the result;
they had no pain with vigorous activities and were able to resume previous
levels of daily and recreational activities. Thirty of the thirty-five
patients who had been regularly employed in a sedentary job prior to the
injury returned to their full work duties within seven months. Of the
forty-five patients who had been regularly employed in a manual job prior to
the injury, thirty-nine returned to their previous work duties.
Radiographic Outcome
(Table
IIandFig.
4)
The quality of the reduction as seen on the first postoperative radiograph
was assessed as excellent/very good in 155 (94%) of the 165 patients, good in
seven (4%), and poor in three (2%). In the group with an initial good
reduction, partial osteonecrosis developed in one patient and nonunion
developed in another. In the group with a poor reduction, partial
osteonecrosis developed in one patient and one patient had complete collapse
of the head. The remaining six patients with a good or poor reduction
demonstrated a persistent malunion, which was moderate in five and severe in
one, at the last radiographic assessment. During the follow-up period, a
malunion also developed in three patients who had had an excellent/very good
reduction; all three of these malunions were mild. Overall, nine (5%) of the
165 patients had a malunion; six of these patients had sustained a three-part
fracture of the greater tuberosity.
According to the definition of fracture union described previously, all
fractures had united within two to three months, with the exception of two
three-part fractures of the greater tuberosity that showed no signs of healing
at four and six months. One of those fractures had had a good postoperative
reduction. Revision osteosynthesis with use of a plate and screws,
supplemented with bone graft, led to union of both of these fractures within
four months.
Heterotopic ossification was detected in fifteen (9%) of the 165 patients
and was classified as class I in twelve patients and class II in three. Ten of
these patients had been involved in road traffic accidents and had sustained
other skeletal injuries.
Complications
Overall, osteonecrosis of the humeral head developed in eleven patients
(7%) and total collapse was detected in four (2%). Three of these cases were
attributed to the nature of the fracture; the other patient had a poor
postoperative reduction. Each of these four patients with total collapse was
treated with conversion to a hemiarthroplasty, at a mean of sixteen months
postoperatively. Partial osteonecrosis, present in seven patients (4%), was
most often located at the anterolateral portion of the head. None of these
seven patients had severe pain or clinically relevant restriction of shoulder
motion.
Symptoms of impingement were detected in four patients (2%) between six
months and one year after the surgery. Three patients were treated
successfully with one, two, or three subacromial injections of corticosteroid
and local anesthetic. The other patient, a seventy-two-year-old woman,
remained symptomatic eighteen months postoperatively. She did not wish to
undergo subacromial decompression as she was satisfied with the outcome of the
fracture fixation and experienced only moderate pain with forward elevation of
>150°. At the time of the last follow-up, two patients (1%) had
symptomatic posttraumatic osteoarthritis that was seen radiographically and
was classified as grade 2 and 3.
The overall reoperation rate was 4% during the average 5.4-year follow-up
period. One deep infection required surgical drainage, there were two
conversions to plate osteosynthesis as described, and there were four
conversions to a hemiarthroplasty.
The ideal treatment of a displaced proximal humeral fracture remains
controversial and is still a matter of debate in the literature. Misra et
al.24 conducted a
meta-analysis in 2001 and concluded that, although the quality of reports
seems to be improving, the current published literature is inadequate to allow
evidence-based clinical decision-making.
With the exceptions of minimally displaced fractures, which can be
successfully managed by nonoperative means, and of four-part fractures,
head-splitting fractures, and fracture-dislocations, which are usually treated
with hemiarthroplasty, open reduction and internal fixation is usually
indicated for displaced two, three, and four-part valgus impacted
fractures3,25-27.
In contrast to conventional plate methods, there is a growing trend toward
less invasive surgical exposures and osteosynthesis techniques involving
screws, wires, sutures, and tubular or locking plates, in an endeavor to
minimize soft-tissue detachments, periarticular scarring, and vascular insult
to the articular humeral head segment. Percutaneous pin, cannulated screw, and
intramedullary wire and nail fixation techniques are reasonable alternatives,
but there is always the risk of breakage, migration, joint or neurovascular
penetration, and superficial infection with the use of these
implants11,28,29.
McLaughlin30
first suggested the use of sutures for fixation of avulsed tuberosity
fractures. Neer31
advocated the use of wires and silk or nylon sutures for the fixation of
three-part fractures and reported excellent or satisfactory results in 86% of
patients so treated. Cuomo et al. 32 reported good to excellent results in
eighteen of twenty-two patients in whom a two-part or three-part fracture had
been treated with wire or suture fixation. Flatow et
al.33 reported that
twelve patients with a displaced fracture of the greater tuberosity had a good
or excellent result at a mean of five years after suture fixation.
To date, there have been few reports of large series of patients treated
with suture fixation. One of the most recent studies is that by Park et
al.15, who reported
on twenty-seven patients in whom a total of twenty-eight fractures of the
proximal part of the humerus had been treated with nonabsorbable sutures.
Thirteen fractures were of the greater tuberosity, nine were two-part surgical
neck fractures, and six were three-part fractures. Overall, there were
twenty-five excellent or satisfactory results and few complications. Hockings
and Haines13
reported only one case of osteonecrosis and a mean Constant score of 86% in a
study of eleven patients in whom a four-part valgus impacted fracture had been
treated exclusively with transosseous sutures. Finally, Banco et
al.34 and Iannotti
et al.25 described
the parachute and figure-of-eight techniques of transosseous fixation with use
of Dacron tapes and number-2 FiberWire sutures, respectively.
To our knowledge, the current study involved one of the largest reported
series of displaced proximal humeral fractures managed solely with
nonabsorbable sutures. However, our treatment regimen has several critical
differences when compared with the suture-fixation techniques described in
similar reports. First, we prefer the lateral transdeltoid approach as it
provides adequate exposure of the fracture while reducing soft-tissue trauma.
The axillary nerve, which is the major anatomical structure in danger, is
located approximately 5 to 6 cm below the tip of the acromion. With the
exposure that we used, the split in the deltoid ends far more proximally, as
only 1 to 2 cm of metaphyseal exposure is required for the drill holes. In a
cadaveric study, Gardner et
al.35 investigated
an extended anterolateral acromial approach to the proximal part of the
humerus. The axillary nerve typically was found deep to the anterior deltoid
raphe, approximately 6.3 cm from the undersurface of the acromion.
Subsequently, they used this approach in sixteen patients with a proximal
humeral fracture, and none had complications related to the surgical approach.
In our series, we found no cases of axillary neurapraxia that could be
attributed to the surgical approach. We believe that the main advantage of the
lateral approach when compared with the more standard deltopectoral approach
is the preservation of the remaining blood supply of the humeral head,
especially in patients with a four-part valgus impacted fracture.
Second, with our technique, the displaced tuberosities in four-part valgus
impacted fractures always are pulled down below the level of the head in an
attempt to avoid subacromial impingement, and they are sutured not only to
each other but also to the articular fragment and to the medial and lateral
aspects of the diaphysis in a manner that we believe neutralizes the deforming
muscular forces. Similarly, in the treatment of three-part fractures of the
greater tuberosity, the sutures are also passed through the intact lesser
tuberosity. This provides a stable construct and restores the normal
functional balance of the involved tendons, thus allowing early motion of the
shoulder joint. In the treatment of isolated two-part fractures of the greater
tuberosity, the displaced tuberosity also is reduced to its anatomical
position, thus avoiding mechanical blockage to abduction of the shoulder, or
obstruction of external rotation due to posterior displacement of the greater
tuberosity.
Third, we avoid reducing the impacted articular component of the four-part
valgus impacted fracture to its anatomical position, thus minimizing the risk
of further disruption of the posteromedial hinge
(Figs. 1-A and 1-B). The
overall prevalence of total osteonecrosis of the head in this subgroup of
patients was only three (7%) of forty-five, one of the lowest rates that has
been
reported13,17,36,37.
Despite incomplete fracture reduction with our technique, it seems that the
disruption of normal anatomy does not affect shoulder joint biomechanics. The
moment arm of the rotator cuff muscles remains normal as we suture the
tuberosities below the level of the impacted head. Use of this technique is
supported further by the very low rate of early degenerative arthritis seen in
this series.
Partial head osteonecrosis and malunion were well tolerated by most of our
patients, especially the younger ones. Severe complications developed in only
four of the ten patients with good or poor postoperative reduction (one had
total necrosis, two had partial osteonecrosis, and one had nonunion). Two of
these patients required a reoperation. In addition, three patients who had had
an excellent/very good reduction had a malunion. However, all of these
patients with a malunion had a good Constant score at the last clinical
assessment.
McLaughlin30
emphasized that normal anatomy is not crucial for normal function of the
shoulder joint, and some authors have suggested that some residual deformity
may be acceptable in the setting of percutaneous
fixation11.
Careful patient selection and adherence to the defined indications are
important for the success of our therapeutic regimen. The final decision
regarding which intervention is undertaken should be made at the time of the
surgery, when the true nature of each fracture can be fully recognized. Should
the humeral head be found to be completely dislocated or to have undergone
substantial translation (>1 cm) or comminution, a primary hemiarthroplasty
should be performed. Furthermore, we believe that most two-part surgical neck
fractures should be treated with plate-and-screw osteosynthesis. We do not
recommend transosseous suture fixation for this type of fracture as rotational
instability and unstable fixation between the large proximal fragment and the
narrow diaphysis often can cause problems.
Finally, completion of the full rehabilitation program is integral to
obtaining the optimum outcome. A major variation from previous protocols that
was employed in our patient cohort was the institution of pendulum exercises
on the second postoperative day and the use of such exercises for the first
three to four weeks. A full range of motion could be established in this
fashion without exerting undue stress on the fixation.
Neer31 emphasized
the importance of early rehabilitation in the management of proximal humeral
fractures, and its value was demonstrated in the studies by Koval et
al.38 and
Kristiansen et
al.39. We believe
that a good functional outcome can be achieved only with individualized and
supervised monitoring of the patient's progress.
The strength of this study lies in the large number of patients as well as
the strict criteria for assessment of long-term clinical and radiographic
outcomes, but there were also limitations. While we attribute our good
outcomes to precise surgical technique and a rigorous rehabilitation protocol,
refined assessment of these parameters is difficult. Furthermore,
reproducibility of the radiographic measurements is difficult to obtain. Often
the humeral head is either internally or externally oriented, depending on the
position of the radiographic plate and the orientation of the beam. Efforts
were made in all cases to obtain a good-quality radiographic trauma series and
to define strict criteria for inclusion of patients in the treatment protocol.
The lack of a control group is also an important weakness of our study. In the
future, a multicenter study should be conducted to compare our technique with
other stabilization techniques, such as fixation with the new locking plates.
Despite these limitations, we recommend open reduction and suture fixation for
acute four-part valgus impacted, three-part, and two-part greater tuberosity
fractures of the proximal part of the humerus with the intention of achieving
stable osteosynthesis, an adequate rotator cuff repair, and early mobilization
of the shoulder joint. ?