Abstract
Background: Nonoperatively treated fractures
of the humeral diaphysis have a high rate of union with good functional
results. However, there are clinical situations in which operative
treatment is more appropriate, and, though interest in plate osteosynthesis
has decreased, intramedullary nailing has gained popularity in recent
years. We report the results of treating fractures of the humeral diaphysis
with a prefabricated brace that permits full motion of all joints
and progressive use of the injured extremity.
Methods: Between 1978 and 1990, 922 patients
who had a fracture of the humeral diaphysis were treated with a
prefabricated brace that permitted motion of adjacent joints. The
injured extremities were initially stabilized in an above-the-elbow
cast or a coaptation splint for an average of nine days (range, zero
to thirty-five days) prior to the application of the prefabricated
brace. Orthopaedic residents, supervised by teaching staff, provided
follow-up care in a special outpatient clinic. Radiographs were
made at each follow-up visit until the fracture healed.
Results: We were able to follow 620 (67 percent)
of the 922 patients. Four hundred and sixty-five (75 percent) of
the fractures were closed, and 155 (25 percent) were open. Nine
patients (6 percent) who had an open fracture and seven (less than
2 percent) who had a closed fracture had a nonunion after bracing.
In 87 percent of the 565 patients for whom anteroposterior radiographs
were available, the fracture healed in less than 16 degrees of varus angulation,
and in 81 percent of the 546 for whom lateral radiographs were available,
it healed in less than 16 degrees of anterior angulation. At the
time of brace removal, 98 percent of the patients had limitation
of shoulder motion of 25 degrees or less.
We were unable to follow most of the patients long-term, as they
did not return to the clinic once the fracture had united and use
of the brace had been discontinued.
Conclusions: Functional bracing for the treatment
of fractures of the humeral diaphysis is associated with a high rate
of union, particularly when used for closed fractures. The residual
angular deformities are usually functionally and aesthetically acceptable. The
present study illustrates the difficulties encountered in carrying
out long-term follow-up of indigent patients treated in charity
hospitals that are affiliated with teaching institutions. These
difficulties are also becoming common with patients insured under
managed-care organizations and are frequent in our peripatetic population.
To our knowledge, the senior one of us (A. S.) and his colleagues11 were the first to report the successful
use of functional bracing for the treatment of fractures of the humeral
diaphysis. The new method permitted freedom of motion of all joints
in the injured extremity4. The
initial report on fifty-one patients showed nonunion in one patient
who had a pathological fracture. Subsequent reports have shown a
high rate of union and few complications3,5,14,16,17.
In recent years, improved operative techniques have made other methods
of treatment available, particularly intramedullary nailing. The
initial intramedullary nailing procedures were followed by a high
rate of complications. Robinson et al.8 reported
a 23 percent rate of delayed union and nonunion (seven of thirty
patients).
Since our study population consisted primarily of indigent patients
seen at teaching institutions, a large number of the patients failed
to return for additional follow-up once the acute symptoms had subsided.
This deprived us of the opportunity to perform an outcomes study.
Between 1978 and 1990, we treated 922 patients who had a fracture
of the humeral diaphysis with a prefabricated brace that permitted
functional motion of the shoulder and elbow. We were able to follow
620 (67 percent) of these patients.
The average age of the patients was thirty-six years (range,
sixteen to eighty-three years) at the time of injury. There were
391 male patients (63 percent) and 229 female patients (37 percent). Three
hundred and three (49 percent) of the fractures were in the right
humerus, and 317 (51 percent) were in the left humerus; 465 (75
percent) were closed, and 155 (25 percent) were open. One hundred
and eighteen (76 percent) of the open fractures were gunshot injuries,
and the remaining thirty-seven (24 percent) resulted from motor-vehicle
accidents. One hundred and ninety-two patients (31 percent) sustained
the fracture in a fall to the ground; 118 (19 percent), from a gunshot;
211 (34 percent), in a motor-vehicle accident; and ninety-nine (16
percent), from various other causes, such as a bicycle accident,
a direct blow to the arm, or a twisting force.
Ninety-two fractures (15 percent) were in the proximal third
of the humeral diaphysis, 303 (49 percent) were in the middle third,
219 (35 percent) were in the distal third, and six (1 percent) were segmental.
One hundred and one fractures (16 percent) were transverse, 149
(24 percent) were oblique, 364 (59 percent) were comminuted, and
six (1 percent) were segmental. Twelve (2 percent) of the fractures
had an associated inferior glenohumeral subluxation.
Sixty-seven patients (11 percent) had an associated radial nerve
palsy. Fifty-two (78 percent) of the palsies were associated with
a closed fracture and fifteen (22 percent), with an open fracture.
Twenty-eight (42 percent) of the sixty-seven fractures were transverse,
fourteen (21 percent) were oblique, and twenty-five (37 percent)
were comminuted. The mechanism of injury was a fall for sixteen
patients (24 percent), a motor-vehicle accident for twenty-four
(36 percent), a low-velocity gunshot for twenty-three (34 percent),
and unknown for four (6 percent)
In the instances in which a nerve injury was due to a penetrating
injury or a high-velocity gunshot wound, operative exploration was
performed and, if necessary, the nerve was repaired. We were unable
to determine the exact number of patients in this category. The
fractures in these patients were stabilized by other orthopaedic
surgeons within the department, and we were unable to follow them.
Only patients who had an open fracture or an associated injury were
admitted to the hospital. The humeral fractures of an unknown number
of patients who had multiple injuries were treated by operative
or nonoperative methods, and these patients were not included in
our study.
Initially, the injured extremity was stabilized in an above-the-elbow
cast or a coaptation splint that held the elbow in 90 degrees of
flexion, for an average of nine days (range, zero to thirty-five
days). None of the fractures were manipulated. Patients were evaluated
in the outpatient department approximately one week (range, three
days to five weeks) after the initial injury. If the acute symptoms
had subsided and the injured extremity was not swollen, a brace
was applied and the patient was given a collar-and-cuff sling to
wear. The brace consisted of two plastic sleeves that encircled
the arm with two adjustable Velcro straps to hold the sleeves together.
The brace extended from approximately two inches (five centimeters)
distal to the axilla to two inches proximal to the olecranon. Patients
were shown how to adjust the brace and tighten the Velcro straps
several times a day to accommodate the changes in the girth of the
extremity that occurred as the swelling subsided and muscle atrophy
developed (Fig. 1).
The brace was worn at all times, except during bathing.
Patients were instructed in the performance of pendulum exercises
immediately after the application of the initial cast or splint,
and the exercises were continued after the application of the brace. The
collar-and-cuff sling was taken off for a few minutes several times
a day to permit combined active and passive exercises of the elbow
and to regain full extension of this joint as soon as possible. Active
elevation and abduction of the shoulder were not allowed, since
such exercises could lead to angular deformity. The patients also
were instructed not to lean the elbow on the arm of a chair, a table,
or their lap, as leaning on the elbow of a fractured extremity during
the early stages of healing may cause varus angulation. Such angulation is
more likely to occur in association with transverse fractures, particularly
when the bone fragments contact each other; it is less likely to
occur in association with oblique fractures, where elastic pistoning
of the fragments takes place.
Patients were seen one week after the application of the brace,
and radiographs were made to evaluate the position of the fracture.
Once full extension of the elbow had been achieved, use of the collar-and-cuff
sling was discontinued during walking but its use was encouraged
during recumbency. During the next four weeks, patients increased
the frequency and intensity of exercises involving passive flexion
of the shoulder and active flexion and extension of the elbow.
Fracture treatment was the same for the sixty-seven patients
(11 percent) who had an associated radial nerve palsy at the time
of the injury. A cock-up wrist splint was not used in the anticipation that,
once the elbow reached full extension, the wrist would spontaneously
extend to neutral, precluding the development of a permanent flexion contracture
of the wrist and fingers. Patients were instructed to perform active
and passive extension of the wrist and fingers several times a day.
No attempt was made to document a history of smoking, drinking,
or recreational drug use, since the patients, with few exceptions,
belonged to a population in which habitual use of these substances
is common. Questions related to these habits probably would not
have been answered accurately. We believed that, under those circumstances,
it would have been very difficult to obtain reliable data. We have
no information on the amount of time away from work, since many
of our patients either were unemployed or did not hold a permanent
working position.
We were able to follow 620 (67 percent) of the 922 patients who
had a fracture of the humeral diaphysis to the point of complete
healing. A large number of patients discontinued their visits to
the outpatient clinics as soon as the injured extremity became painless
and functional. Despite numerous attempts, we could not obtain long-term
follow-up for these patients.
The functional brace was removed upon confirmation of clinical
and radiographic union of the fracture, which occurred at an average
of 11.5 weeks (range, five to twenty-two weeks) (Fig. 2-A,Fig. 2-B,Fig. 3-A,Fig. 3-B,andFig. 3-C
). The 465 closed fractures healed at a median of 9.5 weeks (range,
five to nineteen weeks) and the 155 open fractures, at a median
of fourteen weeks (range, eight to twenty-two weeks). The median healing
time was twelve weeks (range, eight to twenty-two weeks) for the
101 transverse fractures, ten weeks (range, five to seventeen weeks) for
the 149 oblique fractures, eleven weeks (range, five to eighteen
weeks) for the 364 comminuted fractures, and twelve weeks (range,
eight to twenty-one weeks) for the six segmental fractures. The median
healing time was ten weeks (range, five to fourteen weeks) for the
ninety-two fractures located in the proximal third of the humeral
diaphysis, ten weeks (range, six to twenty-two weeks) for the 303
fractures located in the middle third, nine weeks (range, six to
twenty-two weeks) for the 219 fractures located in the distal third,
and twelve weeks (range, eight to twenty-one weeks) for the six
segmental fractures.
Sixteen patients (3 percent) required operative intervention
because of a nonunion. The mechanism of injury was a fall to the
ground for four of these patients, a motor-vehicle accident for
nine, and a low-velocity gunshot for three. Nine open fractures
and seven closed fractures did not unite. Of the nine nonunions
of open fractures, four (two transverse and two comminuted) were
in the middle third of the humerus and five (two transverse and
three comminuted) were in the distal third of the humerus. Of the
seven nonunions of closed fractures, four (three transverse and
one comminuted) were in the middle third of the humerus and three
(two transverse and one comminuted) were in the distal third. Distraction
between the fragments in the nonunion was observed in seven patients
on radiographs made with the patient standing.
Four (less than 1 percent) of the patients had a refracture between
the second and the eighth week after removal of the brace. Two of
the refractures occurred after a fall; one, during sports activities; and
one, from an unknown mechanism. The four refractures healed following
reapplication of the brace.
The most recent anteroposterior radiographs were available for
measurement of the humeral angulation in 565 patients (91 percent)
(Fig. 4),
and the most recent lateral radiographs were available for measurement
of angulation in 546 patients (88 percent) (Fig. 5). The 101 transverse
fractures healed in an average of 9 degrees of varus angulation;
the 149 oblique fractures, in an average of 4 degrees; and the 364
comminuted fractures, in an average of 8 degrees.
Nerve function did not return in one of the sixty-seven patients
who had a radial nerve palsy. Because of the relatively short duration
of follow-up of many of the patients, we are not in a position to state
the ultimate degree of recovery that might have taken place. As
we stated, an unknown number of patients who had a repair of the
lacerated nerve were followed at other services and were not seen
in our clinic.
The twelve patients who had an inferior subluxation of the shoulder
demonstrated spontaneous correction of the subluxation, but no accurate
data was kept concerning the speed of recovery.
Once the fracture was clinically stable and there were radiographic
signs of healing, the patients were asked by the examining resident
to flex, abduct, and rotate the shoulders. The range of motion was
recorded on specially designed forms, and the final recording reflected
the motion measured at the time of the last contact with the patient
(Fig. 6).
Elbow motion was recorded for 301 (49 percent) of the patients (Fig. 7).
We have no information on the number of deep infections that
may have occurred in the patients who were admitted to the hospital
and never transferred to our clinic, since they were treated by
a different service. None of the patients who had an open fracture
that was treated by us had an infection at the time of transfer
to our clinic.
Nonoperative management still remains the treatment of choice
for most fractures of the humeral diaphysis5,12.
A high rate of union and satisfactory functional results have given
credence to this method3,5,8,11,14,16,17.
Ostermann et al.5 reported nonunion
of four (2 percent) of 191 fractures. Zagorski et al.17 reported that three (2 percent)
of 170 patients had a nonunion, and the average varus-valgus angulation
in their series was 5 degrees. Sharma et al.16 found
a residual varus angulation of less than 5 degrees in more than
50 percent of forty patients who were treated with a functional
brace. In another, comparable group of twenty-five patients treated
with a coaptation splint, only four (16 percent) had a final varus
angulation of less than 5 degrees.
Rüedi et al.10 reported nonunion,
infection, and other complications after the treatment of fractures
of the humeral diaphysis, particularly after operative treatment. Habernek
and Orthner1 found no instance
of nonunion, infection, or secondary radial nerve palsy in a group
of nineteen patients treated operatively. In a study of thirty-nine
patients treated with the use of a retrograde nail, Rommens et al.9 reported that 95 percent (thirty-seven)
had union and 92 percent (thirty-six) had excellent shoulder function.
However, Robinson et al.8 reported
a postoperative complication rate of 87 percent (twenty-six patients),
a reoperation rate of 70 percent (twenty-one patients), and a rate
of delayed union or nonunion of 23 percent (seven patients) in a
group of thirty patients. Postoperative shoulder function was poor
in seven of the eighteen patients for whom functional assessments
had been performed. Riemer et al.7 found
complications in seven of twelve patients in whom a fracture had
been treated with nine-millimeter-diameter nails. Hems and Bhullar2 reported a 33 percent rate of nonunion
(seven of twenty-one patients) and a 16 percent rate of shoulder
pain (four of twenty-five patients) related to use of an interlocking
nail through an antegrade approach.
In our study of functional bracing for the treatment of fractures
of the humeral diaphysis, we found a rate of nonunion of less than
2 percent in the 465 patients who had a closed fracture and of 6
percent in the 155 patients who had an open fracture. The final
angular deformities were cosmetically acceptable in most instances:
87 percent of the 565 fractures for which anteroposterior radiographs were
available healed in less than 16 degrees of varus angulation, and
only 2 percent had a permanent varus angular deformity of more than
25 degrees (Fig. 4). These degrees of angulation are considered
by most to be aesthetically acceptable3,5,8,14,16,17.
We believe that the low prevalence of refracture of less than 1
percent (four patients) is related to the fact that fractures treated
without rigid immobilization heal with abundant and stronger periosteal
callus13.
The brace was removed at an average of 11.5 weeks (range, four
to twenty-two weeks) after the initial injury. Since the functional
brace does not immobilize any of the joints in the injured extremity
for more than one week, permanent limitation of function is unlikely
and many patients are able to perform most functional activities
of daily living before the fracture is healed and the brace is removed.
Sixty percent of the patients had a full range of shoulder motion
upon discontinuing use of the functional brace. At the time of brace
removal, 98 percent of the patients had limitation of shoulder motion
of 25 degrees or less (Fig. 6). The patients did not receive supervised
physical or occupational therapy throughout the treatment period.
They were instructed by the residents or attending surgeons in the
exercises during the weeks preceding the development of intrinsic
fracture stability, and this was followed by active and passive
exercises consisting of flexion, extension, abduction, and adduction
of the shoulder. We found that, on examination of the patients who
were seen several months after completion of healing, there was
an increase in the range of motion with continued use of the extremity.
Patients were instructed not to actively abduct or elevate the extremity
against gravity until early bridging callus was seen on radiographs;
otherwise, an angular deformity could develop.
Of the 301 patients for whom the range of elbow motion was recorded,
230 (76 percent) had a full range of motion upon removal of the
brace and the remaining seventy-one (24 percent) had limitation of
flexion and extension at that time (16 percent had less than 10
degrees of limitation; 5.6 percent, 11 to 25 degrees; and 2 percent,
more than 25 degrees) (Fig. 7). We assumed that, as in the case
of the shoulder, continued use of the extremity would result in
an improved range of motion.
Functional bracing does not restore anatomical alignment of a
fracture of the humeral diaphysis, but the final angular deviations
were cosmetically and physiologically acceptable3,5,8,14,16,17 and
should not be considered complications but simply minor deviations
from the normal. Permanent loss of the carrying angle of the elbow
does not result in a functional loss and is cosmetically well tolerated.
This is particularly true for patients with large muscular or flabby
arms3,5,8,14,16,17. Malrotation
of the fragments can be kept within functional and aesthetically
acceptable parameters with early active contraction of the flexors
and extensors of the elbow joint. We believe that the triceps, brachialis,
and biceps muscles experience a coiling of their fibers as the bone
fragments rotate after the injury but recoil as the muscles contract during
activity. This recoiling appears to align the fragments in a parallel
direction, correcting the malrotation15.
Manipulation of fractures of the humeral diaphysis is not necessary
with functional bracing since, in most instances, the gravity-dependent
position of the extremity and compression of the soft tissues about
the humerus restore adequate alignment to the fragments. This concept
is substantiated by the fact that only 2 percent of the fractures
healed with angular deformities of more than 25 degrees (Figs. 4
and 5).
Associated radial nerve palsy is a common complication of fractures
of the humeral diaphysis. In our series, the prevalence was 11 percent
(sixty-seven patients). Spontaneous recovery is likely to occur
in virtually all instances if the fracture is closed and the palsy
develops at the time of the injury3,5,8,14,16,17.
Only one of our patients who had a radial nerve palsy failed to
demonstrate improvement of nerve function. However, since the duration
of follow-up was not long, we do not know the maximum extent of
recovery. We do not consider routine exploration of the radial nerve
to be necessary in patients who have a closed fracture or a fracture
produced by a low-velocity gunshot. Operative exploration of the
radial nerve is indicated when a patient has a sharp, penetrating
injury; a high-velocity gunshot wound; or an open fracture associated
with severe soft-tissue injury.
Associated inferior subluxation of the glenohumeral joint is
rarely seen, but it appears to be more common in association with
fractures located in the proximal third of the diaphysis. We identified
inferior subluxation in twelve patients (2 percent). Early voluntary
contractions of the biceps and triceps muscles rapidly restore congruity of
the glenohumeral joint. The active contraction of the two muscles,
which attach proximally on the scapula and distally on the humeral
diaphysis, restores the normal relationship between the humeral
head and the glenoid fossa (Figs. 2-A and 2-B).
The high prevalence of union (97 percent) in this combined group
of closed and open fractures of the humeral diaphysis gives credence
to our long-held hypothesis that motion at the fracture site is an
important factor in osteogenesis13,15.
We do not know the exact mechanism through which motion encourages
fracture-healing. Nonetheless, we suspect that pain and the subsequent
irritation created by motion between the fragments results in a
cascade of favorable events, such as increased vascularity, piezoelectric
potentials, and local chemical and thermal changes13.
The cost of care of fractures of the humeral diaphysis with nonoperative
functional bracing is lower than that of operative treatment, since
hospitalization is not required, the rate of nonunion is low, and
recovery occurs within a short period. These factors need to be
considered when treating such fractures, particularly in view of
the rising cost of health care.
Habernek, H., and Orthner, E.:
A locking nail for fractures of the humerus. J Bone Joint Surg, 73-B(4):
651-653, 1991.
Hems, T. E., and Bhullar, T. P.: Interlocking nailing
of humeral shaft fractures: the Oxford experience 1991 to 1994.
Injury, 27: 485-499, 1996.
Klestil, T.; Rangger, C.; Kathrein,
B.; Huber, B.; and Waldegger, M.: Sarmiento bracing of humeral shaft
fractures - a comparative study. SOT, 20, 1997.
Latta, L. L.; Sarmiento, A.; and Tarr,
R. R.: The rationale of functional bracing of fractures. Clin. Orthop.,
146: 28-36, 1980.
Ostermann, P. A. W.; Ekkernkamp, A.;
and Muhr, G.: Functional bracing of shaft fractures of the humerus
- an analysis of 195 cases. Orthop. Trans., 17: 937, 1993-1994.
Ricciardi-Pollini, P. T., and Falez,
F.: The treatment of diaphyseal fractures by functional bracing.
Results in 36 cases. Italian J. Orthop. and Traumatol., 11: 199-205,
1985.
Riemer, B. L.; Foglesong, M. E.; Burke,
C. J., III; and Butterfield, S. L.: Complications of Seidel intramedullary
nailing of narrow diameter humeral diaphyseal fractures. Orthopedics,
17: 19-29, 1994.
Robinson, C. M.; Bell, K. M.; Court-Brown,
C. M.; and McQueen, M. M.: Locked nailing of humeral shaft fractures.
Experience in Edinburgh over a two-year period. J. Bone and Joint
Surg., 74-B(4): 558-562, 1992.
Rommens, P. M.; Verbruggen, J.; and
Broos, P. L.: Retrograde locked nailing of humeral shaft fractures.
A review of 39 patients. J Bone Joint Surg, 77-B(1): 84-89,
1995.
Rüedi, T.; Moshfegh, A.; Pfeiffer,
K. M.; and Allgöwer, M.: Fresh fractures of the shaft of the humerus:
conservative or operative treatment? Reconstr. Surg. and Traumatol.,
14: 65-74, 1974.
Sarmiento, A.; Kinman, P. B.; Galvin,
E. G.; Schmitt, R. H.; and Phillips, J. G.: Functional bracing of
fractures of the shaft of the humerus. J Bone Joint Surg,
59-A: 596-601, July 1977.
Sarmiento, A., and Latta, L. L.: Closed
Functional Treatment of Fractures, pp. 497-545. New York, Springer,
1981.
Sarmiento, A.; Latta, L. L.; and
Tarr, R. R.: Principles of fracture healing. Part II. The effects of
function on fracture healing and stability. In Instructional Course
Lectures, the American Academy of Orthopaedic Surgeons. Vol. 33,
pp. 83-106. St. Louis, C. V. Mosby, 1984.
Sarmiento, A.; Horowitch, A.; Aboulafia,
A.; and Vangsness, C. T., Jr.: Functional bracing for comminuted
extra-articular fractures of the distal third of the humerus. J.
Bone and Joint Surg., 72-B(2): 283-287, 1990.
Sarmiento, A., and Latta, L. L.: Functional Fracture
Bracing: Tibia, Humerus, and Ulna. New York, Springer, 1995.
Sharma, V. K.; Jain, A. K.; Gupta,
R. K.; Tyagi, A. K.; and Sethi, P. K.: Non-operative treatment of fractures
of the humeral shaft: a comparative study. J. Indian Med. Assn.,
89: 157-160, 1991.
Zagorski, J. B.; Latta, L. L.; Zych,
G. A.; and Finnieston, A. R.: Diaphyseal fractures of the humerus.
Treatment with prefabricated braces. J Bone Joint Surg, 70-A:
607-610, April 1988.