Radiation therapy for the treatment of bone tumors and soft-tissue
sarcomas may deliver damaging doses of radiation to skeletal bone.
It is known that ionizing radiation has a detrimental effect on
cortical bone10 and that it may
inhibit and delay fracture union12,16.
Stinson et al.15 reported a rate
of long-term fracture of 6 percent (eight of 134 patients) after
a limb-sparing procedure and radiation therapy for the treatment
of a primary soft-tissue sarcoma. Grigsby et al.5 reported
a 4.8 percent prevalence (ten of 207 patients) of femoral neck fracture
at an average of 12.7 years after irradiation of the groin. There
is a paucity of information regarding the treatment and predictability
of union of fractures through previously irradiated bone.
Free vascularized fibular transfer has become an established
procedure for the treatment of major skeletal defects and recalcitrant
nonunions, as well as after tumor resection, in both the upper and
the lower extremities3,8,9,13,17,18.
Han et al.6 reported an 81 percent
rate of union (130 of 160 patients) at an average of forty-two months
after a vascularized bone transfer. Seven patients had had a nonunion
of a pathological fracture after irradiation for the treatment of
a tumor; however, these patients were included in the tumor-resection group
and were not analyzed separately6.
Jupiter et al.9 reported the results
of seven major reconstructions of the femoral shaft that had included
a vascularized free fibular transfer. One of their patients had
had a nonunion of a mid-diaphyseal femoral fracture after irradiation
for the treatment of an osteosarcoma.
To our knowledge, there have been no reported series of patients
in whom vascularized free fibular transfer was used for the treatment
of nonunions in previously irradiated bone. Therefore, the purpose
of the current study was to evaluate the results of this procedure
in patients who had such a nonunion.
We retrospectively reviewed the records of seventeen patients
who had had a nonunion of a fracture associated with irradiation
and had undergone a vascularized free fibular transfer procedure
between April 1984 and October 1995. All patients were followed
for a minimum of two years, and all were followed until the time
of union, failure of treatment, or death. The clinical records and
radiographs were reviewed by one of us (G. P. D.) who was not directly
involved with the care of any of the patients. All free fibular
transfers were carried out, and the patients subsequently were followed, by
the senior one of us (M. B. W.). The data that were recorded for
each patient included age, tumor diagnosis, radiation dosage and
technique, oncological procedures, chemotherapy, date of fracture,
fracture treatment, duration of follow-up, date of vascularized
free fibular transfer, graft length, postoperative complications,
and time to osseous union and functional recovery.
The clinical results were evaluated according to the rating system
described by Mankin et al.11.
A grade of excellent meant that the patient had no evidence of disease,
was pain-free, and had essentially normal function with no limitations
(except with regard to high-performance sports activities); a grade
of good meant that there was some degree of impairment of function
that did not necessitate bracing or use of supports (such as crutches
or a cane) and did not interfere with the patient's occupation or
lifestyle (except with regard to sports activities); and a grade
of fair meant that some form of support was needed for walking or
prehension and there was sufficient pain to impair function. The
graft was considered to have failed if it was removed or if the
limb was amputated.
The radiographs were evaluated for evidence of union as well
as for stress fractures and other complications. Radiographic evidence
of external bridging callus was defined as the end point of union
of the donor and recipient bone.
Clinical and Demographic Data
There were eleven female and six male patients, whose ages ranged
from thirteen years to eighty-two years at the time of the vascularized
free fibular transfer.
The indications for radiation therapy were a bone tumor in eight
patients and a soft-tissue neoplasm in nine patients (Table I). The bone
tumors included Ewing sarcoma in four patients, lymphoma in two,
cavernous hemangioma in one, and metastatic breast carcinoma in
one. The soft-tissue tumors included malignant fibrous histiocytoma
in two patients; rhabdomyosarcoma in two; and reticulum-cell sarcoma,
myxosarcoma, hemangiopericytoma, fibrosarcoma, and a tumor of undetermined
pathology in one patient each. Six patients were diagnosed with
and treated for the tumor at our institution, and eleven were diagnosed
and treated elsewhere. The six patients who were diagnosed at our
institution included the two who had malignant fibrous histiocytoma
and the four who had Ewing sarcoma. As classified according to the
criteria established by the Musculoskeletal Tumor Society4, both malignant fibrous histiocytomas
were stage IIA. All four Ewing sarcomas were stage IIA. The stage
of the tumor was not available from the records for the eleven patients
who were treated at other institutions.
The treatment of the neoplasm consisted of excision combined
with chemotherapy and radiation in four patients, excision and radiation
in six patients, radiation and chemotherapy without excision in two
patients, and radiation alone in five patients. The excision consisted
of tumor resection with margins that were clear of tumor. There
were no tumor recurrences during the study period.
The dose of radiation was well documented for fourteen patients;
it was unknown for the other three, as the radiation therapy had
been given at another institution and documentation was not available.
The average dose given was 5564 centigray (range, 3000 to 9000 centigray).
One patient received 2000 centigray of intraoperative radiation.
There was one soft-tissue complication, consisting of wound
necrosis and infection, which was directly attributable to the radiation
therapy. This complication occurred in a forty-one-year-old patient who
had a malignant fibrous histiocytoma of the anterior compartment
of the leg and had received 5000 centigray preoperatively and 2000
centigray postoperatively. He underwent multiple d衲idements and
eventually required a rectus abdominis free microvascular muscle
flap to cover the defect.
The average interval between the completion of the radiation
therapy and the occurrence of the fracture was 111 months (range,
zero to 228 months) (Table II). There were no recurrent tumors
at the primary site, and none of the fractures were associated with a
recurrent tumor. All but one fracture were considered pathological
(related either to osteoradionecrosis or to bone resorption due
to tumor) because they were spontaneous or were associated with minimal
trauma. The remaining fracture, which was subtrochanteric, occurred
during placement of an intramedullary rod for the treatment of an
impending pathological fracture.
There were fourteen femoral fractures in thirteen patients, a
humeral fracture in three patients, and a tibial fracture in one
patient. Four of the femoral fractures were subtrochanteric, six
were mid-diaphyseal, and four were distal diaphyseal or metaphyseal.
Of the four subtrochanteric fractures, three were treated initially
with internal fixation with an intramedullary rod and one was treated with
a dynamic compression screw and a seven-hole plate. Of the six mid-diaphyseal
femoral fractures, five were treated initially with internal fixation
with an intramedullary rod and one was treated with a ten-hole plate.
Of the three distal femoral fractures, one was treated initially
with internal fixation with an intramedullary rod; one, with dual
plates; and one, with closed reduction and external stabilization
with a cast. The supracondylar femoral fracture was treated with
internal fixation with a supracondylar compression screw and a side-plate.
The tibial fracture was treated initially with two lag-screws and
external stabilization with an above-the-knee cast. Of the three humeral
fractures, two were treated with a coaptation cast or a splint and
one, with a cast followed by internal fixation with an eight-hole
plate with six screws.
Two patients had secondary autogenous iliac-crest onlay bone-grafting
to the fracture site after the initial internal fixation. Four patients
had adjuvant treatment with pulsed electromagnetic-field stimulation
to the fracture site after the initial internal fixation procedure.
One patient had bone-marrow aspirate injected percutaneously into
the nonunion site. None of these interventions led to union of the fracture.
None of the patients had wound-healing problems associated with
the initial treatment. The average interval between the pathological
fracture and the vascularized free fibular transfer was nineteen
months (range, one to fifty-two months).
A total of eighteen fibular transfer procedures were performed
in the seventeen patients (Fig. 1-A,Fig. 1-B,Fig. 1-C,Fig. 1-D,Fig. 1-E,Fig. 1-F,Fig. 1-G,Fig. 1-H, and Table III). One patient required a second
free fibular transfer one year after the initial fibular transfer
for the treatment of a second fracture distal to the original fracture
and the original fibular graft in the same femur. All fibular transfers
were applied as onlay grafts because none of the nonunions were
associated with a large segmental osseous defect. The fibular graft
was fixed with screws proximal and distal to the nonunion site in
all patients. Cancellous autogenous bone from the iliac crest was
used as an additional graft at the proximal and distal junctions
and at the fracture site in all patients. The average length of
the transferred fibular segment was 13.5 centimeters (range, nine
to twenty centimeters). The average operating time was 9.5 hours (range,
7.3 to 12.1 hours).
At the time of the free fibular transfer for the fourteen femoral
fractures, the initial fixation was maintained in seven patients,
an intramedullary-rod exchange was performed in four, and an exchange
of the condylar plate was performed in two (one of whom had insertion
of a compression plate and the other, an intramedullary rod). One
patient who had had no prior stabilization had stabilization with
an external fixator at the time of the fibular transfer.
The patient who had a nonunion of a tibial fracture also had
a malunion of the distal aspect of the tibia with 30 degrees of
varus angulation due to an old previous fracture. This patient underwent
a simultaneous corrective osteotomy for the treatment of the distal
malunion, a free fibular transfer to the site of the proximal nonunion,
and intramedullary fixation across both sites.
Of the three patients who had a nonunion of a humeral fracture,
two were treated initially with a cast alone. This was unsuccessful
and led to a nonunion in both patients. The first operation in these two
patients was the vascularized free fibular transfer. One patient
had plate-and-screw fixation, which was exchanged for an external
fixator at the time of the free fibular transfer. The average duration
of follow-up after the fibular transfer was fifty-seven months (range,
twenty-eight to 112 months).
Rate of Union
Fifteen of the eighteen fractures united primarily following
the vascularized fibular transfer. Six fractures united by six months;
seven, by one year; and two, by two years. Two patients (Cases 4
and 16) had a residual nonunion: one had a subtrochanteric nonunion
and the other, a nonunion of the distal aspect of the femur. An
additional patient (Case 2) had union after a secondary cancellous
bone-grafting procedure to one end of the transferred fibula. The
time to union averaged 9.4 months (range, three to twenty-four months).
Functional Result
Thirteen patients had an excellent result; one, a good result;
two, a fair result; and one, a failure of treatment.
All thirteen patients who had an excellent result returned to
their prefracture activity level. All patients who had had a fracture
of the lower extremity were pain-free and walked without a limp.
The patient who had a good result (Case 3) was satisfied but
had a limp and was not able to participate in sports.
The two patients who had a fair result were older in age. One
(Case 4) had a persistent nonunion but was able to function with
use of a brace and to walk about the house. The other patient (Case
15), who was eighty-eight years old at the time of the most recent
follow-up, had a healed fracture but sustained a second pathological
fracture proximal to the vascularized free fibula. The patient was treated
with a brace. At the time of this writing, the fracture had not
healed but the patient was able to walk about the house.
The patient who was considered to have had failure of the treatment
(Case 16) eventually had an above-the-knee amputation.
Donor-Site Morbidity
There were no donor-site wound complications, and no patient
perceived functional impairment of the donor limb. The clinical
records did not suggest any objective findings of an abnormal nature related
to the donor limb.
Secondary Operations Following the
Vascularized Free Fibular Transfer
Three patients (Cases 2, 3, and 4) underwent removal of the intramedullary
rod and d衲idement for the treatment of a deep infection at the
site of the rod. One of these patients (Case 2) required secondary
autogenous iliac-crest bone-grafting at the proximal end of the
free fibular transfer for the treatment of a delayed union.
Failure of Treatment
The patient (Case 16) who had failure of the treatment had a
distal femoral nonunion that had been treated with external fixation
at the time of the vascularized free fibular transfer. The fracture
did not heal, and after one year the ununited distal femoral fragment
was resected and a custom total knee endoprosthesis was inserted.
The proximal stem of the prosthesis subsequently became loose, and
the patient had an above-the-knee amputation two years after the
vascularized free fibular transfer.
Secondary Fractures
Two patients had a second fracture in the same femur, one and
four years after the initial fracture united following the fibular
transfer. One of these patients (Case 15) was an eighty-two-year-old woman
who initially had a nonunion of a distal femoral fracture that was
treated with a condylar blade-plate at the time of the vascularized
free fibular transfer. The fracture healed uneventfully after seven
months. Four years later, the patient sustained a mid-diaphyseal
femoral fracture proximal to the plate and fibula. The plate was
removed, and an intramedullary rod was placed without additional
bone graft. The rod was noted to be broken three years later, with
no evidence of union at the fracture site. However, because the patient
was only slightly symptomatic, no additional operation was undertaken.
At the time of the latest follow-up, the patient was eighty-eight
years old and pain-free; she was able to walk with a walker and
a brace and to live independently.
The other patient (Case 5) had a second fracture in the femur
distal to the site of the vascularized free fibular transfer. The
fracture was treated successfully with a second free fibular transfer
and revision of the internal fixation. (This patient is discussed
further in the section on infection.)
Complications of Fixation
One patient (Case 13), who had fixation with an intramedullary
reconstruction rod at the time of the vascularized free fibular
transfer for the treatment of a subtrochanteric nonunion, was noted
to have proximal migration of the femoral screw through the femoral
head and into the acetabulum four months postoperatively. The screw
was exchanged for a shorter screw. The patient did well, with no
symptoms related to the hip. Healing of the nonunion was confirmed
at seven months.
Infection
Four patients had a deep infection.
One patient (Case 3) had cultures that were positive for Pseudomonas
aeruginosa after corrective osteotomy of the distal aspect
of the tibia and intramedullary-rod fixation combined with the vascularized
free fibular transfer. The patient was managed with a six-week course
of intravenous gentamicin therapy (240 milligrams per day) and also
with intravenous ceftazidime therapy (two grams every eight hours
for two months). The patient did well for seven months but then
required emergency removal of the rod because of septicemia. The
patient was managed again with gentamicin (400 milligrams intravenously
every twenty-four hours) and with imipenem (750 milligrams intravenously
every eight hours) for a total of six weeks. At eighteen months
the nonunion had healed, and at three years the patient had a good
functional recovery.
Another patient (Case 2) had a superficial perioperative wound
infection, which healed with local wound care. One year later, it
was noted that the intramedullary rod had broken and the proximal end
of the transferred fibula had not united. The rod was removed, and
intraoperative cultures were found to be positive for Propionibacterium
acnes. The femur was debrided with extensive intramedullary
reaming. The patient received cefazolin (one gram every eight hours)
intravenously for six weeks. A second intramedullary rod was subsequently
inserted, and the ununited proximal aspect of the fibula was augmented
with an iliac-crest autograft. At the time of the most recent follow-up two
years later (forty-two months after the vascularized free fibular
transfer), the fracture had healed and there were no signs of persistent
deep infection.
Another patient (Case 4), who had intramedullary-rod fixation
at the time of the vascularized free fibular transfer for the treatment
of a subtrochanteric femoral nonunion, had no evidence of union
and had radiolucent areas adjacent to the rod at fifteen months
postoperatively. Deep infection was suspected, and an indium bone
scan was suggestive of infection. The intramedullary rod was removed, and
the femur was reamed extensively. The cultures of deep-wound specimens
were positive for coagulase-negative Staphylococcus organisms. The
patient was treated with a six-week course of intravenous vancomycin
therapy (750 milligrams twice a day). A functional cast-brace was
applied, but no additional operations were performed. Four years
later, there was no evidence of union, but the patient was pain-free
and able to walk with the brace.
The fourth patient (Case 5) had a nonunion of a fracture of the
proximal aspect of the femur that had been treated initially with
a compression hip-screw and plate. The fracture united six months
after the vascularized free fibular transfer. One year after the
original fracture had healed, the patient sustained a second fracture
distal to the distal end of the fibular graft. The original plate
was removed, and a second free fibular transfer was done adjacent
to the distal fracture with intramedullary-rod fixation. Intraoperative
cultures at the time of this procedure were positive for Pseudomonas
aeruginosa; therefore, the patient was given a six-week
course of gentamicin therapy (400 milligrams intravenously every
twenty-four hours) and ceftazidime (two grams intravenously every
eight hours). The fracture appeared to progress to radiographic union;
however, due to a persistently draining sinus fourteen months later,
the intramedullary rod was removed, the sinus tract was excised,
and extensive reaming of the intramedullary canal was performed.
The patient was treated with six more weeks of gentamicin therapy
(400 milligrams intravenously every day) and a two-month course
of imipenem therapy (750 milligrams intravenously every eight hours).
At the time of the most recent follow-up 7.8 years later, the patient
was asymptomatic with no evidence of infection, and she was able
to bear weight fully without support. Radiographs suggested union
of both femoral fractures.
Radiation in sufficiently high doses is well known to have detrimental
effects on normal bone. Maeda et al.10 studied
the effects of a single 3500-centigray dose of external beam radiation
to the shafts of rat femora. At two weeks, there was a loss of cells
in the diaphyseal cortex of the irradiated femora and decreased
bone-turnover activity as reflected by the significant surface-area
reduction (p < 0.01) of both osteoid seams and resorption spaces
in the intracortical region of the irradiated bone. At fourteen
to eighteen weeks, a point-counting method1,10 revealed
increased porosity of bone compared with that in the control rats.
Fracture-healing has also been shown to be impeded by both prefracture
and postfracture irradiation1,7,12,14,16.
Cutright and Brady2 studied
the long-term effects of radiation on the vascularity of rat bone.
They observed a marked decrease (average, 30 percent) in bone mass
with exposures of more than 4000 centigray compared with that of
the nonirradiated bone. Moreover, vascularity (measured with use
of the isotope chromium-51) decreased in the irradiated bone to 28
percent of that in the nonirradiated bone with exposures of more
than 4000 centigray. Eight months following irradiation, there was
no improvement in bone vascularity in association with any of the
doses studied (2000, 4000, 6000, 8000, and 10,000 centigray). The
average dose of radiation received by the patients in our series
was 5564 centigray. The treatment of the tumors varied in terms
of surgery and chemotherapy, but all patients received a minimum
of 3000 centigray and some received as much as 9000 centigray. All fractures
occurred within the exposure field of the radiation therapy. All
but one fracture was spontaneous or associated with minimal trauma,
suggesting that the fractures were indeed of a pathological nature.
It is noteworthy that none of the fractures were due to recurrence
of the tumor. The delay between the radiation therapy and the fracture
averaged 111 months, suggesting that the deleterious effects of
radiation on bone are long-term and perhaps gradually progressive
over time. The decision to proceed to vascularized free fibular
transfer was made after other avenues of treatment had been considered.
The treatment of radiation-induced pathological fractures is
challenging, and the surgery is complicated. Attempts to achieve
union of these fractures are associated with many pitfalls and complications.
The rate of deep infection in the current series was appreciable
(four of eighteen fractures). This high rate was likely influenced
by the multiple previous operations, the long duration of the operative
procedure (average, 9.5 hours), and possibly the irradiation.
The overall rate of union was high; sixteen of the eighteen fractures
healed at an average of 9.4 months following the fibular transfer.
There were fourteen excellent or good results in this series. The
two patients who had a fair result included one who had a second
fracture in the field of the radiation therapy and one who had an
infection and a persistent nonunion. Both patients required a brace for
walking. The only patient who had a failure of treatment eventually
required an above-the-knee amputation.
On the basis of this review, we recommend that a fracture occurring
within the field of therapeutic radiation should initially be treated
with open reduction and stable internal fixation with a nonvascularized
autograft. If there is no evidence of union at six months and the
internal fixation is still sound, a vascularized free fibular transfer
with additional corticocancellous autografting should be performed.
In conclusion, we suggest that microvascular free fibular transfer
combined with autografting in association with stable fracture fixation
is an appropriate treatment option for difficult nonunions in previously
irradiated bone.