Fractures of the humeral diaphysis respond well to nonoperative treatment
with functional braces in the majority of
patients1,2.
However, patients who sustain high-energy injuries often benefit from surgical
stabilization of the fracture. Osteosynthesis with a compression plate is the
so-called gold standard for surgical treatment of these
fractures3-5,
with reported union rates of 95% to
97%4,5.
However, plate osteosynthesis is associated with negative features, including
the need for a direct exposure of the fracture site, increased blood loss, a
risk of radial nerve injury, and disruption of the periosteal blood
supply3-8.
Closed intramedullary stabilization of a humeral shaft fracture can, in
theory, avoid these problems. Unlocked flexible nails have been
used3,9-15,
but they have been associated with problems relating to pain, nonunion, and
lack of rotational control of the fracture. Locked intramedullary nails have
been associated with controversy, as outstanding results have been reported in
some studies and unfavorable outcomes have been documented in
others7,16-25.
Shoulder pain and malfunction have been reported for as many as 16% to 37% of
the patients in recent
studies8,16,18,26.
Evaluation of the literature is complicated by the marked differences in
surgical indications among the
studies25,27.
The primary problems that have been reported with locked intramedullary
nailing are nonunion, shoulder pain, and dysfunction.
The titanium flexible humeral nail (Synthes, Paoli, Pennsylvania) was
developed by the AO as an alternative intramedullary device
(Fig. 1). It is made of
numerous sigmoid-shaped segments that interlock over a central core wire, and
it is available in two diameters, 7.5 and 9 mm. It allows for proximal and
distal interlocking and provides four wires that can be deployed from the tip
of the nail (Figs. 2-A and
2-B,3-A and 3-B).
The device is flexible when it is implanted; however, it becomes rigid when a
screw is placed in the tensioning block, thus driving the block proximally and
resulting in stiffening of the central core wire
(Figs. 4-A and 4-B). Once the
nail is stiffened, proximal and distal holes become available for
interlocking. The nail is designed to be inserted with use of osseous entry
portals that do not violate the rotator cuff or damage the articular surface
of the humeral head during antegrade insertion
(Figs. 2-A and 2-B) and that
are smaller and avoid the olecranon fossa during retrograde insertion
(Figs. 3-A and 3-B).
The purpose of this study was to document the clinical outcome and
complications associated with the use of this flexible humeral nail in a
consecutive series of patients. To our knowledge, no previous studies on the
use of this device have been reported in the English-language literature.
Fifty consecutive patients with fifty-one humeral fractures were enrolled
in this prospective study between March 1997 and December 2000. This flexible
humeral nail was used by the surgeons for all patients who met the entry
criteria during the study period. The study was approved by our institutional
review board, and all patients provided informed consent to participate in the
study. All surgical treatment and clinical follow-up was conducted at a
level-I trauma center and its associated clinic. In order to be included in
this study, patients had to be more than seventeen years old and have
sustained a fracture involving the diaphysis of the humerus. In addition, they
needed to have one or more of the following conditions: an injury involving
multiple trauma resulting in the need for upper extremity weight-bearing, a
bilateral humeral fracture, an open humeral fracture, a severe associated
chest wall injury, or the inability to maintain an acceptable closed reduction
of the fracture. Exclusion criteria were an age of seventeen years or less and
an unwillingness to consent to the use of this device. At the time of
follow-up, all patients were evaluated by one of three authors (J.P.S.,
D.A.V., or J.E.A.). Clinical union was defined as functional motion of the
limb without substantial tenderness at the fracture site and the ability to
perform at least the activities of daily living. Radiographic union was
defined as osseous bridging of three of the four cortices visible on the
anteroposterior and lateral radiographs. Patients were evaluated with use of
the Short Form-36 (SF-36), the shoulder scoring system of Constant and
Murley29, and a
subjective pain-rating scale. The Constant score was obtained as originally
described by Constant and Murley except that dumbbell weights were used to
assess shoulder power in lateral abduction. In addition to the motion
documented for the Constant score, shoulder abduction and elbow flexion and
extension were also documented. All radiographs were reviewed by two authors
(J.P.S. and H.W.H.), who came to a consensus with regard to their
interpretation.
Patients completed the pain scale data forms during clinic visits. They
were asked to rate the pain numerically on a scale that ranged from 0 to 10,
with 0 indicating no pain, 1 to 3, minimal pain; 4 to 6, moderate pain; or 7
to 10, severe pain. Patients were not to score the pain as 0 unless they had
absolutely no pain in the shoulder. A minimal pain score (1 to 3) was defined
as rare or occasional pain that neither limits activities nor requires any
medication stronger than occasional nonsteroidal anti-inflammatory medications
or acetaminophen. Moderate pain was defined as more frequent pain that may
occasionally impede functional activities or sleep but does not require any
medication stronger than those listed above. Severe pain was defined as
frequent pain that impairs function or sleep and requires narcotic
analgesia.
Forty-one patients with forty-two fractures had been followed for more than
six months. The mean duration of follow-up for these patients was twenty-two
months (range, seven to forty-two months). Nine patients, including two who
had died, were lost to follow-up. The mean age of the forty-one patients was
thirty-eight years (range, eighteen to eighty-six years). There were twelve
female and twenty-nine male patients. The mechanism of injury was a
motor-vehicle accident for twenty-seven patients, a fall for four, a gunshot
wound for four, a motorcycle accident for two, a pedestrian-motor-vehicle
accident for three, and an accident in which the individual was struck by a
train. The mean Injury Severity Score was 18 points (range, 4 to 50 points).
Twelve limbs sustained ipsilateral injuries in addition to the humeral
fracture. There were fifteen open fractures, two vascular injuries, and one
nerve injury. Nine fractures involved the proximal third of the humerus,
twenty-five involved the midshaft, two were in the distal third, and six were
segmental. The fracture pattern was transverse in seven, oblique in nine,
spiral in one, comminuted in seventeen, and segmental in eight. Twenty nails
were placed in nineteen patients in an antegrade manner, and twenty-two were
inserted in a retrograde manner. Twenty-four 7.5-mm nails and eighteen 9-mm
nails were implanted.
Surgical Technique
Our treatment protocol for closed humeral fractures involved intramedullary
stabilization in the first few days after the injury, depending on the
patient's condition and the surgeon's schedule. Open fractures were irrigated
and débrided immediately following admission, and the nail was placed
at the time of that surgery. The retrograde technique was used for fractures
in the proximal third of the humerus, the antegrade technique was used for
fractures involving the distal third, and the technique used for midshaft
fractures was chosen at the discretion of the attending surgeon. The
postoperative rehabilitation regimen included pendulum exercises beginning on
the first day following surgery, with advancement to passive range-of-motion
exercises of the elbow and shoulder within a few days as the pain decreased.
Active range-of-motion exercises were discouraged until approximately six
weeks following surgery or until callus was visualized on radiographs.
Antegrade insertion is accomplished with the patient in the beach-chair
position or lying supine with a 30° rotation of the torso and with
fluoroscopy available during the procedure. A 3 to 4-cm incision is made
either straight laterally or anterolaterally from the acromion. The deltoid
muscle is split along its fibers. Care is taken to limit the length of the
split to <5 cm to prevent damage to the axillary nerve. The greater
tuberosity is located, and the fibers of the rotator cuff insertion are
identified. A sharp curved awl is used to make an entry portal approximately 1
to 1.5 cm distal to the insertion of the rotator cuff. The entry point should
be made distal enough to ensure that there will be no damage to the fibers of
the broad insertion of the rotator cuff. It is important to make certain that
the awl is directed toward the medial cortex of the humerus to prevent
disruption of the lateral cortex distal to the opening with subsequent
reaming. A ball-tipped guide-wire is passed across the fracture site following
reduction. It is important to remember the maneuver that is necessary to
obtain reduction, since the guide-wire will have to be removed to implant the
noncannulated nail. The proximal aspect of the humerus must be reamed to 10.5
mm to accept the proximal end of the nail. The ends of the deployable wires
should be cut off prior to antegrade insertion, since they are deployed only
with retrograde insertion. The guide-wire is removed and the nail is implanted
with use of the insertion jig. If passage is difficult, the surgeon should
flex the nail more to free the tip from the sides of the medullary canal. Once
the nail is inserted, the stiffening screw is placed through the insertion jig
and a stab hole is made through the skin. A measuring guide is provided to
determine the length of the stiffening screw. The fracture is impacted, and
then the nail is stiffened by placing the stiffening screw
(Figs. 4-A and 4-B). Proximal
interlocking is then accomplished with a screw placed through a second stab
incision. When proximal interlocking has been accomplished, the reduction is
checked. If a gap persists at the fracture site, the fracture should be
impacted again. Rotation is assessed and adjusted as needed. The limb is then
placed on a Mayo tray, and distal interlocking is performed with use of the
free-hand technique with fluoroscopy.
Retrograde insertion follows similar steps except that the starting point
is approximately 2 to 3 cm proximal to the olecranon fossa. The entry point is
directly posterior on the humerus. The deployable wires may be used, if
needed, for a fracture with proximal extension. Positioning of the patient for
retrograde nailing is either supine or lateral. There is a guide to assist in
opening the canal at the appropriate angle to minimize anterior notching.
After opening the canal, it is very important to size the opening with the
dummy nail provided with the set, to ensure that the entry portal is of
sufficient size to accommodate the nail. Following insertion of the nail, the
peripheral wires may be deployed from the tip of the nail if needed. The four
wires are only used during retrograde implantation with fractures that extend
proximally near the surgical neck of the humerus. The wires are deployed with
use of a device provided with the set that assists in maintaining rotational
control of the humeral head. The remainder of the procedure is very similar to
the antegrade nailing technique.
Statistical Analysis
Means and proportions were used to describe the study population with
respect to clinical and outcome characteristics. For statistical comparisons,
the t test and Fisher exact test were used when comparing group means and
proportions, respectively. For continuous variables, when more than two group
means were compared, analysis of variance was used. Finally, linear regression
was used to determine the clinical characteristics that were independently
associated with the outcome measures of interest (the Constant score, rating
on the pain scale, and SF-36 physical and mental component scores). A p value
of <0.05 (two-sided) was considered significant.
At the time of follow-up, thirty-nine fractures in thirty-eight patients
had achieved clinical and radiographic union without the need for secondary
procedures. Two fractures (5%) in our series had a nonunion. One additional
patient had an amputation six days following the fracture because of
complications of a vascular bypass procedure. The vascular injury was due to
the primary injury and was not a result of the stabilization of the humerus.
The mean time to clinical union was twelve weeks (range, four to fifty weeks).
Radiographic union was achieved at a mean of ten weeks (range, six to sixteen
weeks). Thirty-one patients (thirty-two fractures) had a full range of motion
at both the shoulder and the elbow. Ten patients (24%) had some loss of
motion: six of them had loss of motion at the shoulder and four had loss of
motion at the elbow. The details of the range of motion and associated
injuries in these ten patients are provided in Tables
I and
II. All patients with loss of
shoulder motion and three of the four patients with loss of elbow motion had
undergone retrograde nailing. The association between loss of motion and
retrograde nailing was significant (p = 0.01).
Twenty-one shoulders had no pain at the most recent follow-up evaluation.
Seventeen shoulders were associated with minimal discomfort; three, moderate
discomfort; and one, severe pain. Two patients (two shoulders) with moderate
pain and the one patient (one shoulder) with severe pain had undergone
retrograde nailing. Forty fractures (95%) had anatomic alignment (=5°
of angulation in either plane), and two had mild varus alignment of 8° and
10°. During seventeen of the twenty-two retrograde procedures, wires were
deployed from the tip of the nail. During the twenty procedures that had
antegrade nailing, wires cut at the proximal end of the nail were left
prominent above the cortex of the humerus in four patients and the proximal
end of the nail was left prominent (mean, 3.5 mm; range, 1 to 10 mm) in six
patients. One of the patients who had an antegrade procedure was noted to have
extensive reaming of the lateral cortex on the postoperative radiograph. Of
the twenty-two patients who had retrograde nailing, three had notching of the
anterior cortex. None of the four entry-point errors led to clinical
complications.
Four patients (10%) in our series had five clinical complications related
to the procedure. Two patients who had antegrade nailing had nonunion. Union
was achieved following exchange nailing in one patient and following treatment
with a compression plate and bone-grafting in the other. Two patients who had
antegrade nailing sustained a broken nail. In one patient, the nail broke
after the patient fell down the stairs fifteen weeks following the index
procedure. The patient was successfully treated with an exchange nailing with
a flexible humeral nail. The other patient was lifting weights (performing
curls under the supervision of a physical therapist), at six weeks following
surgery, when the nail broke. This fracture was treated with a compression
plate. Both fractures healed. A deep infection developed in one patient
following humeral nailing of a grade-IIIA open
fracture28. The
infection resolved with surgical débridement and placement of
antibiotic beads, and the nail was retained. Union was achieved, and the
patient had a good result. All five complications of nailing occurred in
patients managed with the use of a 7.5-mm nail. In contrast, no fracture site
complications occurred in any of the eighteen patients who had fixation with a
9-mm nail. The association between nail size and complications was nearly
significant (p = 0.06).
Thirty-six patients with at least one year of follow-up (mean, twenty-three
months; range, twelve to forty-two months) had a mean score of 90 points
(range, 50 to 100 points), according to the shoulder scoring system of
Constant and
Murley29. The mean
score was 75 points for patients who were more than fifty years old compared
with 94 points for patients who were fifty years of age or younger; the
difference was significant (p < 0.01). The mean Constant scores were also
significantly different with respect to fracture pattern and fracture
location. The patients who had a spiral, transverse, oblique, or comminuted
fracture had a mean Constant score of 92 points (range, 88 to 100 points),
whereas those who had a segmental fracture had a significantly lower mean
score of 80 points (p = 0.04). In addition, the patients who had a fracture of
the proximal, middle, or distal third of the humerus had a mean Constant score
of 92 points (range, 89 to 98 points), whereas the patients with a segmental
fracture had a significantly lower mean score of 76 points (p = 0.02). With
the numbers available, no significant relationship was detected between the
Constant score and the nail insertion technique, nail size, or the occurrence
of a complication.
Thirty-eight patients were evaluated with use of the validated SF-36
questionnaire. The mean mental component score was 46 points (range, 21 to 68
points) and the mean physical component score was 40 points (range, 23 to 62
points) for the group as a whole at the time of the most recent follow-up. The
mean physical component score was significantly decreased (p < 0.01) for
those who had a complication (28 points) compared with those who had no
complication (43 points). The physical function, pain index, and vitality
subscores of the physical component scores were also significantly decreased
for patients who had a complication compared with those without a complication
(p < 0.05). No other variables with a significant effect on outcome could
be identified, with the small numbers available. In the multivariate analysis,
certain variables, including the insertion technique, nail size, fracture
pattern, anatomic location of the fracture, open fracture, ipsilateral
extremity injury, associated skeletal injury, complications, and an age of
more than fifty years, were compared with the outcome measures (the Constant
score, rating on the pain scale, and the physical and mental component scores
of the SF-36). Older age (p = 0.003) was found to be significantly associated
with a lower Constant score. No significant associations with the mental
component score were detected. A significant association was found between the
occurrence of a complication and a lower physical component score (p = 0.02).
Given the low statistical power associated with this small number of patients,
no other significant associations were found.
The results of the use of intramedullary nailing in the treatment of
humeral fractures have been mixed, with good outcomes reported in some
studies7,9-11,13,17,19,24,25,30
and poor outcomes reported in
others8,15,16,18,20,31.
The primary problems have been nonunion and shoulder dysfunction associated
with antegrade nailing that violates the rotator
cuff3,8,15,16,18,23,30,31.
A recent meta-analysis pooled the data of the randomized prospective studies
that compared plate and nail
fixation31. The
pooled data comprised 195 patients and demonstrated that plates were
associated with significantly better outcomes than were intramedullary nails
with respect to the rate of reoperation and the prevalence of shoulder
problems (p < 0.05). Methods designed to prevent shoulder pain have
included the use of small unlocked flexible nails, the use of retrograde
nailing, or an anterior acromial approach for antegrade nailing. All of these
solutions have drawbacks, including nonunion, malunion, supracondylar
fracture, elbow pain or motion limitation, and decreased stability at the
fracture site.
A major goal of this study was to evaluate shoulder function, with use of
outcome measures, after intramedullary nailing of humeral shaft fractures with
a nail that avoids surgical insult to the rotator cuff. The results in our
patients were very encouraging. Shoulder function for all patients was
excellent (a mean Constant score of 90 points), especially for those who were
fifty years of age or younger as they had a mean score of 94 points.
Similarly, the patients had a good result with respect to both the range of
motion and the pain score. Thirty-six shoulders (86%) had a full range of
motion, and thirty-eight (90%) had minimal or no pain. Shoulder function was
found to be impaired in patients with a segmental fracture pattern as well as
in older patients. We believe that the less favorable results with segmental
fractures, which generally covered more than one-third of the length of the
diaphysis, occurred because they may have required somewhat more conservative
physical therapy in the early postoperative period. We found that the SF-36 is
not as helpful in determining shoulder function as are the clinical
examination and the Constant score. On the basis of the data that we obtained,
it appears that shoulder pain is minimized and shoulder function is improved
with use of the flexible humeral nail compared with antegrade stabilization of
the fracture with other intramedullary
devices15,16,18,20,31.
Retrograde intramedullary nailing is another procedure that has been
advocated by some in an attempt to prevent shoulder pain and
dysfunction9,11,12,14,24,25.
Both flexible unlocked
nails9,11,12,14
and solid locking intramedullary
nails24,25
have been used. The flexible nails used in the studies did not allow for
static locking, which decreased the rotational control of the
fracture11,17,24,25.
The solid nails require a large entry portal, and there is a risk of
supracondylar fracture as well as elbow pain if the portal goes through the
olecranon
fossa24,25.
The titanium flexible humeral nail allows interlocking nailing with use of a
relatively small entry portal that does not penetrate the olecranon fossa. The
surgical technique for retrograde implantation is somewhat more demanding than
the antegrade technique. Great care must be taken to avoid anterior cortical
notching during reaming. This nail thus allows implantation from either
direction with the same device, giving the surgeon maximum versatility. With
the numbers available, no significant difference was detected between the
antegrade and retrograde approaches in terms of union, Constant score, or
shoulder pain. However, our impression was that the retrograde technique is
more demanding and should be reserved for fracture patterns with proximal
extension that require that technique.
The rate of complications associated with the smaller (7.5-mm) nail was
significantly higher than that associated with the 9-mm nail. Riemer et
al.23 previously
noted an increased prevalence of complications in patients with medullary
canals of <10 mm. All five of the complications in our study occurred in
patients who had a smaller medullary canal that would accommodate only a
7.5-mm nail.
We concluded that the titanium flexible humeral nail provides an option for
locked intramedullary fixation of humeral fractures while avoiding entry
portals that may interfere with shoulder or elbow function. We believe that
avoiding the rotator cuff and the humeral articular surface may improve the
outcome and decrease shoulder pain when intramedullary nail fixation is
selected. Because small diameter nails were associated with an increased rate
of complications, patients with a medullary canal that is =8 mm in diameter
should be evaluated carefully and options other than intramedullary fixation
should be considered. This device had a relatively low complication rate and
yielded excellent results in terms of union, alignment, and shoulder function.
Additional clinical studies including prospective, randomized comparisons are
needed to fully elucidate the role of this nail in the treatment of patients
with a humeral shaft fracture.