With institutional review board approval, a four-generation pedigree
was obtained (Fig. 1). One
patient presented with tibial lesions and gave a positive family history,
after which medical records and radiographs were reviewed retrospectively to
identify all of that patients' family members who had had lower-extremity
fractures in childhood. All of the affected patients had been treated at our
hospital between 1957 and the time of the study. There were twenty-two
relatives from four generations; three had died before the initiation of the
study.
Radiographs of the tibia and fibula of the six affected patients were
reviewed, and anteroposterior and lateral radiographs of the tibia and fibula
of ten unaffected family members were made at the time of the study. These ten
individuals reported that they had no history of deformity, pain, or fracture
of the lower extremity and had never been treated with bracing, a cast, or
surgery on the lower extremity. The radiographs were inspected for evidence of
previous fractures, lytic lesions, or abnormal mineralization or cortical
defects. Radiographs were not made for non-blood relations (i.e.,
spouses).
Blood samples were collected from fifteen of the nineteen living relatives
as well as from five spouses, for a total of twenty specimens. A karyotype of
Case 4 revealed normal chromosomes. Ten to 20 mL of whole blood was collected,
and DNA isolation and genotyping were performed as previously
described7. To test
the possibility that the osseous lesions were a variant of neurofibromatosis,
linkage analysis of polymorphic loci D17S947, D17S2196, D17S1294, D17S1293,
and D17S1299 surrounding the neurofibromin (NF1) gene on chromosome 17q11.2
was conducted with use of the MLINK version of the LINKAGE program. For this
analysis, we applied a gene frequency of 0.00001 and varying penetrances of 0%
to 99%8.
Case 1. An 11.8-year-old girl presented to our hospital
in 1957 for treatment of a left tibial and fibular fracture that occurred when
a bicycle fell over, striking her leg. She had sustained one previous tibial
fracture. Radiographs revealed a diaphyseal fracture of the tibia and fibula
and a more proximal lytic lesion with a sclerotic border at the junction of
the proximal and middle thirds of the tibia. The fracture was treated with
closed reduction and application of a cast, and it healed uneventfully in
15° of procurvatum.
The patient was followed for 3.5 years after the fracture, and she remained
asymptomatic with a stable deformity. Radiographs made four years following
the fracture showed partial resolution of the lytic lesion. The patient
accompanied a relative to the clinic twenty-seven years after the fracture and
reported that she had had no additional fractures or pain. However, a mild
procurvatum of the tibia was noted clinically. She died at the age of fifty of
unrelated causes.
Case 2. The brother of the patient described above (Case 1)
presented in 1958, at the age of 15.3 years, for treatment of a pseudarthrosis
of the tibia following a fracture of the tibia sustained at the age of twelve
years. The fracture had previously been treated with open reduction and
autologous bone-grafting but had failed to heal. A repeat open reduction,
débridement of the pseudarthrosis, and autologous grafting with bone
from the contralateral tibia were performed. Pathologic examination of the
débrided bone showed cellular fibrous tissue with islands of osteogenic
activity and increased vascularity, resembling "osteitis fibrosa."
The pseudarthrosis healed without additional surgery. Radiographs of the tibia
made when the patient was fifty-seven years of age revealed no evidence of
disease, complete resolution of the lesion, and no limb-length
discrepancy.
Case 3. In 1977, a five-week-old boy presented with enlargement of both
tibiae, which were painful to palpation. His maternal grandmother was the
sister of the two patients described above (Cases 1 and 2). Radiographs
revealed large cystic lytic lesions in the midparts of the tibial shafts with
thinning and expansion of the cortex and an apex lateral bow; the left tibia
showed more involvement than the right
(Figs. 2-A and 2-B). The
patient was treated with a long leg orthosis until the age of two years. He
did well (Fig. 2-C) until the
age of seven years and seven months, at which time he was struck with a
baseball and sustained a fracture of the distal part of the right fibula
(Fig. 2-D). The fracture did
not heal with closed treatment, and he underwent open reduction and internal
fixation (Fig. 2-E). The
fracture united slowly, and the hardware was removed at the age of ten years,
at which time the fracture was thought to be solidly healed. Radiographs made
at the age of twenty-three years showed no evidence of tibial disease on
either side, but a right fibular nonunion was present
(Fig. 2-F).
Case 4. This patient, who was the mother of one patient (Case 5)
and the aunt of another (Case 3), had been seen in 1962 for the treatment of a
nonhealing right tibial fracture that she had sustained at the age of six
years. Radiographs showed a lytic lesion within the cortex of the tibia. The
fracture healed with continued immobilization and did not require surgery.
Radiographs made at the age of forty-four years showed normal findings.
Case 5. A four-year-old boy, the cousin of one patient (Case 3)
and the son of another (Case 4), presented in 1983 for treatment of a right
tibial nonunion seven months following a fracture sustained on a trampoline.
The nonunion was treated with débridement, autologous iliac crest bone
graft, and insertion of a Williams intramedullary rod. The fracture healed
without additional surgery. The patient was followed until the age of
twenty-two years, at which time there had been no additional tibial or fibular
fractures and radiographs showed resolution of the tibial lesion.
Case 6. A twenty-two-month-old girl, the granddaughter of
another patient (Case 2), presented in 1997 for treatment of a pathologic
fracture of the left tibia (Figs. 3-A and
3-B). The medical history was notable for truncus arteriosus, for
which she had had cardiac surgery. A total contact ankle-foot orthosis was
applied for treatment of the fracture and, at the age of four years and one
month, she underwent open débridement, intramedullary nailing with
interlocking screws, and iliac crest bone-grafting for a pseudarthrosis of the
tibia (Figs. 3-C, 3-D, and
3-E). Pathologic specimens from the tibial pseudarthrosis revealed
fibrocartilaginous tissue and dense fibrous scar tissue. The fibrous tissue
was associated with bone trabeculae, some of which were rimmed by a
well-developed layer of osteoblasts (Figs.
3-F and
3-G). Two independent
musculoskeletal pathologists made a diagnosis of osteofibrous dysplasia. Nests
of epithelioid cells were not present, which excluded a diagnosis of
adamantinoma.
At three years and ten months following the surgery, the interlocking
screws had failed and the lesion was still apparent radiographically
(Fig. 3-H). The patient had
been instructed to wear an ankle-foot orthosis full-time for two years
following the surgery and then whenever she left the house. The patient
stopped using the orthosis four years following the surgery. At the age of
eight years and eight months (four years and seven months after the surgery),
the child was walking without pain and also had no pain with a range of
motion, rotation, or stress.
We identified six affected patients, who were followed for 3.5 to
forty-two years after presentation. Their ages at the time of follow-up ranged
from 8.7 to fifty-seven years. The adult heights of the three affected male
patients were 5 ft and 5 in (165.1 cm), 5 ft and 6 in (167.6 cm), and 5 ft and
7.5 in (171.5 cm). They were, on the average, in the 11th percentile (5th,
10th, and 18th percentiles) when compared with age-matched normal standards.
The adult heights of one affected female patient and three unaffected women in
the direct line of transmission, whose children required treatment for tibial
lesions, were, on the average, in the 22.5th percentile (range, 5th to 45th
percentile) and ranged from 5 ft (152.4 cm) to 5 ft and 4 in (162.6 cm)
(Table I).
Three of the six patients who had had documented tibial and/or fibular
involvement as children had no persistent radiographic evidence of disease as
adults. The other three patients had abnormalities on their last available
radiograph, but long-term radiographic follow-up was not possible because of
either death or the young age of the patient.
Radiographs of ten presumably unaffected relatives did not reveal
pathologic lesions in either the tibia or the femur. Seven of the ten
individuals were skeletally mature, and three were children. In seven of these
ten relatives, the transverse diameter of the fibula was reduced and the
cortex appeared striped, while the bone appeared widened on the lateral
radiograph (Figs. 4-A and
4-B). No evidence of a previous fracture was seen. The height of
the seven mature patients ranged from 5 ft and 1 in (154.9 cm) to 5 ft and 7
in (170.2 cm) or the 10th to 45th percentile for gender and age-matched age
groups (Table I).
Analysis of polymorphic markers flanking the NF1 gene on chromosome 17 did
not reveal evidence of linkage, and a common haplotype was not observed in
affected patients. We therefore concluded that it was unlikely that the
osteofibrous dysplasia in this family was caused by mutations in the NF1
gene.
Frangenheim first described a tibial diaphyseal lesion that he
termed "osteitis fibroma" in
19219, and
Jaffe10 later
classified that lesion as a variant of fibrous dysplasia. Kempson then
described the unique locally aggressive characteristics of these lesions,
which he called "ossifying
fibroma"11.
Finally, Campanacci proposed the term "osteofibrous dysplasia" in
1976 and described the clinical and pathologic findings associated with that
condition1,2.
While there was variation among the lesions seen in the family in the
present study, the lesions most closely resembled osteofibrous dysplasia.
Clinically, the lesions were noted in childhood, with four of the six patients
presenting before the age of ten years. Osteofibrous dysplasia presents most
often in the first decade of life, with >50% of patients initially seen at
the age of five years or
younger12. One of
the patients in our study was seen with bilateral tibial involvement in the
neonatal period. Neonatal presentation with both unilateral and bilateral
tibial disease has been previously
described13-16.
The lesions were located in the diaphyseal region of the tibia and, in some
patients, in the fibula as well, and they led to bowing and pathologic
fractures2. Of the
seven lesions, six (including the bilateral tibial lesion in Case 3) healed
with relatively minimal treatment compared with that necessary for the vast
majority of congenital pseudarthroses of the tibia due to neurofibromatosis.
The seventh tibial lesion, in Case 6, had not yet healed at the time of
writing, but it was asymptomatic. Roach et al. reported on eleven children
with late-onset pseudarthrosis of a dysplastic tibia and found the prognosis
to be better than that for patients who have fractures prior to their first
birthday17. (It
should be noted that three of the eleven patients in the report by Roach et
al. are members of the family described in the present study.) No patient in
our study had lesions in the femora or pathologic fractures in the upper
extremities, and there were no other known operations for the treatment of
fractures of bones other than the tibia or fibula throughout the pedigree.
This finding is in contradistinction to the usual clinical history of patients
with fibrous dysplasia, in which femoral involvement and polyostotic lesions
are common. In addition, while fibrous dysplasia may be a congenital
condition, it is not heritable like the tibial dysplasia described in the
family in our study.
Radiographically, the lesions seen in members of this pedigree were
variable (Table II). Most were
well-demarcated and lytic, and they had surrounding sclerosis, as has been
described in osteofibrous dysplasia lesions and that is unlike the indistinct
margins and widespread involvement seen in fibrous dysplasia. However, unlike
osteofibrous dysplasia, the lesions were not limited to the cortex of the
bones. The final follow-up radiographs of the adults in this series showed
irregularities in the medullary canal of the fibula, with thickening of the
fibular diaphyseal cortex. This finding also has not been described in
patients with osteofibrous dysplasia, to our knowledge.
Finally, histologic evaluation of tissue from the lesion in the most
recently treated patient revealed a fibrous stroma with spicules of osseous
trabeculae that were characteristically rimmed with
osteoblasts2. While
it is known that the pathologic appearance of tissue from the cystic type of
congenital pseudarthrosis of the tibia may resemble that of osteofibrous
dysplasia5, it is
the combination of clinical, radiographic, and pathologic findings in these
six patients that most closely fits the description of osteofibrous dysplasia
rather than congenital pseudarthrosis of the tibia or fibrous dysplasia.
None of the patients described in the early reports by
Kempson11 or
Campanacci1,2
or in the articles by Komiya and
Inoue18, Nakashima
et al.12, and
Schoenecker et
al.19 had a
positive family history. In 2002, Hunter and Jarvis described two brothers who
had sustained pathologic fractures through small sclerotic diaphyseal tibial
cortical lesions6.
The large kindred described in our study is, to our knowledge, the first to
show clear genetic transmission of lesions characteristic of osteofibrous
dysplasia. The pattern of inheritance appears to be autosomal dominant, as it
involved both sexes and there was vertical transmission of the disease
throughout the pedigree. However, there were members of the family who had an
affected parent and an affected child but did not, themselves, have any
clinical symptoms of the disease as children.
Ten family members who had no clinical evidence of disease were examined
radiographically at the time of this study, and no tibial lesions were found
in those individuals. Similarly, three affected patients who were followed to
skeletal maturity had healed lesions and normal radiographic findings without
evidence of recurrence at the time of this study. One patient (Case 3) had
persistence of a fibular nonunion but a normal-appearing tibia. It is possible
that apparent obligate carriers in this family, such as the woman who was the
mother of Case 6 and the daughter of Case 2, may have had radiographic
evidence of lesions even though they never became clinically symptomatic, and,
if so, the lesions may have disappeared by skeletal maturity without
treatment. Campanacci and others have documented the spontaneous disappearance
of osteofibrous dysplasia lesions in young children who were observed without
surgery2,18-20.
The subtle radiographic finding of a change in fibular shape may indicate
involvement in asymptomatic individuals. The variable expressivity and
decreased penetrance that we observed in this family, although poorly
understood, are typical of many inherited single-gene disorders. For example,
neurofibromatosis and osteogenesis imperfecta are two classic autosomal
dominant diseases with considerable interfamilial and intrafamilial clinical
variation.
Because of the autosomal dominant transmission with variable clinical
penetrance in this pedigree and the clinical resemblance to congenital
pseudarthrosis of the tibia, the differential diagnosis in this family
included neurofibromatosis. However, no patient had café au
lait spots or other clinical findings consistent with neurofibromatosis.
The DNA analysis excluded neurofibromin as a candidate gene and supported the
clinical evidence that the disease in this family was distinct from
neurofibromatosis21.
In 1976, Beals and Fraser reported on a similar family who had congenital
anterolateral bowing of the tibia with pseudarthrosis and fibular hypoplasia
associated with pectus
excavatum22. Six
individuals were thought to have mild bowing of the tibia bilaterally with
mild attenuation of the distal parts of the fibular shafts, and three had
frank pseudarthrosis of the tibia that required surgical treatment. Autosomal
dominant inheritance was proposed, and there were no unaffected carriers of
the disease. Clinical examination of all family members revealed no
café au lait spots or other evidence of neurofibromatosis.
Biopsies showed "an abundance of osteocytes, osteoclasts, and fibrous
tissue." Pectus excavatum was present in all nine affected patients but
not in any of the seven unaffected relatives. The resemblance of the kindred
in this paper to that in ours is striking, but the members of the family that
we described reported no history of pectus deformities. Physical examination
did not specifically address the shape of the sternum, however, so the
presence or absence of a pectus deformity cannot be definitively
established.
Another familial bone dysplasia that can produce lytic lesions in the long
bones is familial expansile osteolysis, which was described by Crone and
Wallace in 199023.
They observed large lytic lesions that resulted in expansion and thinning of
the cortex in multiple long bones. However, the age at presentation of this
condition is much older than that of osteofibrous dysplasia, with familial
expansile osteolysis presenting between the ages of fifteen and forty-five
years. Familial expansile osteolysis may become so aggressive that the
dysplasia resembles Gorham disease, and spontaneous regression does not occur.
Also, familial expansile osteolysis is associated with hearing loss and poor
dentition. We believe that the condition in the kindred that we studied is
distinct from familial expansile osteolysis.
A finding of note in our study was that none of the five affected patients
who had reached maturity had evidence of adamantinoma on the most recent
radiographs. In fact, the tibiae were for the most part normal as seen
radiographically. Although some authors have proposed that osteofibrous
dysplasia is a precursor lesion for
adamantinoma24,25,
the lesions found in this family did not follow that pattern. Recent
cytogenetic studies have presented supporting evidence that osteofibrous
dysplasia and adamantinoma have a common
histogenesis26-28.
The biopsy specimens that we studied were not subjected to immunohistochemical
staining, so we cannot comment further on the similarities or differences
among classic osteofibrous dysplasia, adamantinoma, and the tibial and fibular
dysplasia seen in this family.
In conclusion, previously undescribed osteofibrous dysplasia-like lesions
occurred in six individuals in three generations of a family of twenty-two
members. These lesions were radiographically variable, but they resembled
osteofibrous dysplasia in that they were found in the tibia and sometimes also
in the fibula and they predisposed the bones to pathologic fracture. The
fractures all healed (with the exception of the one in the most recently
treated patient, who was still being followed at the time of writing),
although some required surgical treatment with bone graft and, occasionally,
intramedullary fixation. None of the lesions progressed to adamantinoma.
The occurrence of osteofibrous dysplasia in three generations of this
family suggests an autosomal dominant inheritance and that a single gene is
causative. We are currently applying a positional cloning approach to identify
this gene, which could reveal important insights into the pathogenesis of the
disease. ?