We conducted a prospective clinical trial of the use of BMP in five
consecutive patients with congenital pseudarthrosis of the tibia. The use of
rhBMP-7 (osteogenic protein-1 [OP-1]; Stryker Biotech, Hopkinton,
Massachusetts) and a prospective clinical study protocol were approved by the
institutional review board of the Children's Hospital of New York and Columbia
University. Informed consent for the use of rhBMP-7 was obtained from the
parents of the patients. The use of clinical and radiographic materials was
approved by the parents in accordance with the Health Insurance Portability
and Accountability Act. We chose rhBMP-7 over rhBMP-2 because the former had
been approved for use in the treatment of tibial nonunion, whereas rhBMP-2 had
received only provisional approval for its use, in conjunction with a cage,
during anterior spinal
fusion25-27.
The families of our patients were informed that data concerning the cases
would be submitted for publication.
Patients
Between October 2002 and March 2003, five patients underwent an
osteosynthesis procedure with use of rhBMP-7 and intramedullary nail fixation.
Four of the five patients also had unilateral external fixation, and one
patient (Case 1) had Ilizarov external fixation. The mean age of the patients
at the time of surgery was six years (range, seven months to thirteen years).
Four patients had undergone a mean of three prior surgical procedures (range,
one to four procedures). Table
I summarizes the prior surgical treatment and the type of
pseudarthrosis according to the classification system described by
Crawford28,29,
in which type I indicates anterolateral bowing of the tibia; type II,
anterolateral bowing with increased cortical thickness, a narrow medullary
canal, and a tubular defect; type III, a cystic lesion; and type IV, the
presence of a fracture, a cyst, or a frank pseudarthrosis.
Functional outcome was analyzed before and after the procedure according to
the criteria developed by Johnston, in which grade 1 indicates unequivocal
union; grade 2, an equivocal union (a residual transverse or longitudinal
cortical deficiency) and/or deformity (usually >15° of valgus,
procurvatum, or recurvatum); and grade 3, a persistent nonunion or
refracture7. The
five patients had a mean lower-extremity length discrepancy of 3.3 cm (range,
1 to 5 cm). Four patients were able to walk with a knee-ankle-foot
total-contact orthosis that had an anterior and posterior shell. One patient
was an infant. One patient had pain at the nonunion site preoperatively. Three
of the five patients had neurofibromatosis; the other two had no associated
disorders.
Surgical Treatment
An anterior midline skin incision was made over the nonunion site. After
the removal of local, dense fibrous tissue, rhBMP-7 (3.5 mg) mixed with a
type-I collagen carrier and corticocancellous allograft bone graft was applied
around the nonunion site. Scrupulous attention was paid to removing any
abnormal tissue posteriorly so that healthy soft tissues would be in contact
with the rhBMP-7. The bone ends were resected back to healthy-appearing bone.
A tourniquet was used during the procedure so that the integrity of the
rhBMP-7 allograft composite would be well maintained until the end of the
procedure. Three patients had a second tibial osteotomy after which a
concomitant limb-lengthening procedure was performed over an intramedullary
nail, away from the nonunion site, to treat a limb-length discrepancy. A
unilateral external fixation device was applied in four patients, and an
Ilizarov device was used in one.
We applied static compression to achieve apposition of the bone ends at the
site of the nonunion. Dynamic compression was not applied. One patient (Case
5) who had overlapping of the bone segments at the nonunion site and a 5-cm
lower-extremity length discrepancy had primary tibial distraction at the
nonunion site with a unilateral external fixator. One patient (Case 1) was
treated with intramedullary nailing and an above-the-knee cast and had
placement of rhBMP-7 alone, without supplemental allograft. This patient
required an additional procedure because of failure of union at seven months
postoperatively and underwent the placement of another 7 mg of rhBMP-7 with 6
mL of bone marrow along with application of an Ilizarov frame.
Clinical and Radiographic Evaluation
All patients were evaluated preoperatively and were followed
postoperatively at four-week intervals. Non-weight-bearing anteroposterior and
lateral radiographs of the involved extremity were made at the preoperative
examination, and follow-up radiographs were made at four-week intervals until
one year postoperatively. Four patients underwent a technetium-99m bone scan
six months postoperatively.
Healing of the nonunion, incorporation of the rhBMP-7 allograft composite,
resorption of the allograft, radiographic changes in the appearance and size
of the proximal and distal tibial physes, and the lower-extremity length
discrepancy were recorded and analyzed. Lower-extremity lengths were measured
radiographically on scanograms. Functional outcome was graded before and after
the procedure according to the criteria described by
Johnston7.
Complications such as infection, drainage, and wound dehiscence were
recorded.
Radiographic and Functional Outcome
Preoperatively, the lesions were classified as type III (Cases 1, 2, and 3)
and type IV (Cases 4 and 5) according to the radiographic classification
system of
Crawford28,29.
All five patients had an associated pseudarthrosis of the fibula.
Case 1. An eight-year-old girl with a Crawford type-III
pseudarthrosis and no associated disorders demonstrated no evidence of healing
and underwent revision surgery with use of 7 mg of rhBMP-7, corticocancellous
bone allograft, and bone-marrow aspirates at seven months following the index
procedure. Subsequently, there was a persistent transverse cortical deficiency
and most of the allograft was resorbed. The bone scan at six months
postoperatively was negative.
Case 2. A twelve-year-old girl with neurofibromatosis and a
Crawford type-III lesion had undergone four prior surgeries and demonstrated
complete union at six months (Figs. 1-A,
1-B, and 1-C). This patient gained 5 cm in limb length as a result
of a simultaneous lower-extremity lengthening procedure. Complete cortical
restoration and continuity was noted, and alignment was satisfactory.
Incorporation of the rhBMP-7 allograft composite was not observed
radiographically. The bone scan was negative.
Case 3. A one-year-old boy with neurofibromatosis and no history
of surgery had recurrent anterior angulation and an obvious Crawford type-III
pseudarthrosis at the most recent follow-up examination. There was increased
bone density and hypertrophy at the pseudarthrosis site. The bone scan was
negative.
Case 4. A five-year-old girl with neurofibromatosis and a
Crawford type-IV lesion showed deficient healing on radiographs at the time of
the one-year follow-up, although 2.5 cm of the lengthened segment had
consolidated well. The bone scan was negative.
Case 5. A four-year-old boy with a Crawford type-IV congenital
pseudarthrosis of the tibia was treated with intramedullary nailing and
external fixation. The patient had gradual tapering of the tibia without
angular deformity at the pseudarthrosis site. The rhBMP-7 allograft bone
composites were gradually resorbed (Figs.
2-A through 2-D). The bone scan
was negative.
All five patients met the criteria of a Johnston grade-3 outcome
preoperatively. After the rhBMP-7 implantation and other surgical treatment,
the outcome score improved to grade 1 in one patient (Case 2) and remained as
grade-3 congenital pseudarthrosis of the tibia in the other four patients
(Cases 1, 3, 4, and 5).
Radiographically, the allograft bone and the rhBMP-7 composite appeared to
be gradually resorbed without the formation of new bone. Technetium bone
scintigraphy did not reveal radioisotope uptake in the regions in which the
bone graft and rhBMP-7 bone composite were applied around the nonunion
sites.
Complications
Neither abnormal growth nor angular deformity developed in our patients in
any areas other than at the pseudarthrosis site following the use of rhBMP-7.
The proximal and distal tibial physes remained unchanged in radiographic
appearance. Postoperatively, the patients did not experience fever, rash, or
any other systemic illness. One patient had sterile serous drainage from the
wound. A first-generation cephalosporin antibiotic was administered to the
patient for five days. The drainage included fine granules of
corticocancellous allograft bone. Early surgical intervention was not
undertaken because the cultures were persistently negative, there was a lack
of systemic symptoms, the laboratory values were normal, and the wound had a
benign appearance. The cause of the drainage, which eventually stopped after
eight weeks, was uncertain; however, because there was never any definitive
evidence of infection and the bone scan was negative at six months, we believe
that the drainage was not a cause of the persistent nonunion.
Congenital pseudarthrosis of the tibia remains one of the most difficult
pediatric orthopaedic disorders because it is highly resistant to current
treatment
methods1-18.
In a multicenter review of 340 patients who underwent 1287 surgical
procedures, the European Paediatric Orthopaedic Society reported that the rate
of union ranged from 28% to
75%6. These patients
had many different procedures, and few definitive conclusions were drawn.
Anderson et al. demonstrated that intramedullary nailing, with use of the
technique described by Williams, and autogenous bone-grafting resulted in
union of an established pseudarthrosis in nine of ten
patients1. In a
multicenter study of different treatment options, Grill et al. found that
treatment of 108 tibiae with the Ilizarov method combined with other
techniques resulted in eighty-one unions (75%), which was the highest of any
procedure analyzed6.
A recent study by Johnston demonstrated that a grade-1 outcome was obtained in
five of twenty-three patients after only one surgical procedure consisting of
intramedullary nailing and autogenous iliac-crest
bone-grafting7. With
additional surgery, six more patients achieved a grade-1 outcome, or complete
union. One of the critical factors for a favorable outcome was fibular
osteotomy. Dormans et al. reported that eleven of twelve patients had union
after a free vascularized fibular
transfer4. At the
other extreme, one study showed that fourteen of thirty patients required
amputation because the exremities remained dysfunctional after multiple
attempts (a mean of 4.7
procedures)16. We
know of no published reports on the use of BMPs in the treatment of congenital
pseudarthrosis of the tibia.
Since the first description of the osteogenic potential of proteins
extracted from bone by
Urist30, numerous
studies have provided a scientific rationale for the use of BMPs in human
fractures or spinal
fusions25,30-34.
In the two largest reported clinical series,
rhBMP-720 and
rhBMP-221, which
work through nearly identical mechanisms, seemed to be alternative or possibly
superior bone-graft substitutes, perhaps with the ability to replace
autogenous iliac-crest bone graft. Although these early clinical results
demonstrated promise in the treatment of fractures in adults, the results of
BMP treatment were modest compared with the results from animal studies.
Because a single dose of BMP does not govern the entire fracture-healing
process, it may not solve the most challenging problems of congenital
pseudarthrosis of the tibia. Other critical factors may be the length of time
that the BMP remains at the pseudarthrosis site and whether healthy progenitor
cells are present. In our patients, new bone formation around the
pseudarthrosis site was virtually absent in response to rhBMP-7.
Technetium-99m bone-scanning did not reveal any radioisotope uptake in the
area of the rhBMP-7 and corticocancellous allografts. When healing occurred in
the one patient (Case 2), it was achieved through cortical appositional
healing rather than by a bridging callus. One patient (Case 1) underwent a
repeat application of BMP-7 mixed with a bone-marrow aspirate and
corticocancellous allograft at seven months. At the time of the second
surgical procedure, abundant dense fibrous tissue was noted around the
pseudarthrosis site. There was no evidence of new bone formation in the areas
in which the first dose of rhBMP-7 had been placed.
One of the compounding factors that may have caused poor healing in our
series is the high degree of stiffness of the unilateral external fixators.
Four patients had such frames applied. Although we were able to restore the
mechanical axis and maintain alignment with the unilateral external fixators,
we did not create the mechanical compression that is unique to the Ilizarov
device5,6.
Successful clinical results have been reported following the use of the
Ilizarov device, perhaps reflecting the importance of mechanical compression,
which can promote a better biologic stimulus for healing.
With regard to safety, no side effects associated with the use of either
rhBMP-7 or the type-I collagen carrier were noted and there were no
radiographic changes in the proximal or distal tibial physes or abnormal bone
growth related to the use of rhBMPs. One patient had persistent sterile
drainage of fine granules of corticocancellous bone grafts through an
incompletely healed wound; the drainage stopped spontaneously after two
months.
In conclusion, we believe that the use of BMPs alone may not be enough to
overcome the poor healing environment that is associated with congenital
pseudarthrosis of the tibia. A single application of 3.5 mg of rhBMP-7
combined with a type-I collagen and corticocancellous allograft bone graft did
not result in enhanced bone healing in our patients. ?