It is well known that growth disturbances may occur after injury to
growth plates in children. However, in the treatment of infections or benign
tumors adjacent to the physis, it is unclear what extent of curettage would
cause early closure of the growth plate. We report a case of tuberculosis of
the proximal part of the tibia in a child, in whom the growth plate of the
proximal part of the tibia was curetted surgically over a large area and yet
continued to grow.
Skeletal tuberculosis without spinal involvement in children is
rare1, and it does
not have any pathognomonic
radiographic2 or
clinical characteristics. Skeletal tuberculosis is often difficult to diagnose
initially; several weeks or months may be required to diagnose it
correctly3. In the
case reported here, it took eighty days to confirm the diagnosis. Despite this
delay, the involved growth plate regenerated, and the tibia was still growing
nine years later.
The patient and family were informed that data concerning the case would be
submitted for publication.
Athirty-month-old child complained of pain in the right thigh and
knee. Because coughing and sneezing were also observed, a pediatrician
diagnosed a common cold and prescribed a nonsteroidal anti-inflammatory drug.
On the next day, the child was examined by an orthopaedic surgeon because of
the right thigh and knee pain. The right knee was not swollen or warm, and a
radiograph showed normal findings. Two weeks later, the patient was examined
by us. By then, the right knee was swollen and tender, although an
anteroposterior radiograph of the knee did not show any abnormality. A
T2-weighted magnetic resonance image of the knee showed low signal intensity
only in the center of ossification in the proximal epiphysis of the right
tibia (Figs. 1-A and 1-B).
The patient had been vaccinated with BCG (bacille Calmette and Guerin) at
eight months of age. The tuberculin skin test was positive, which is normal
after the vaccination. There was no family history of tuberculosis.
Serum studies included a normal leukocyte count, a slightly elevated
C-reactive protein level (0.73 mg/dL [7.3 mg/L]), and an erythrocyte
sedimentation rate of 47 mm/h. As a result, a diagnosis of bacterial infection
of the proximal part of the tibia was made.
An antibiotic agent (cefroxadine, 420 mg/day) was administered for six
weeks with a subsequent decrease in pain, and all blood-test results improved
to within the normal range. Three weeks after discontinuation of the
antibiotic, the knee became swollen again, and there was local heat to
palpation. Curettage and débridement of the proximal metaphysis of the
tibia was done to drain the pus and obtain tissue for biopsy. Histologic
examination of granulation tissue from the cavity of the proximal part of the
tibia revealed Langerhans giant cells, lymphocytes, and epithelioid cell
proliferation, which were considered to be diagnostic of tuberculosis of the
proximal part of the tibia. Chemotherapy with isoniazid, 200 mg/day, and
rifampicin, 140 mg/day, was begun. However, despite this treatment the
surgical wound remained open, and a radiograph showed that the radiolucent
lesion in the proximal part of the tibia had increased in size (Figs.
2-A and 2-B).
Additional cultures of specimens from the wound and polymerase chain
reaction tests did not reveal any evidence of Mycobacterium
tuberculosis. Forty days after the initial surgical procedure, a biopsy
of the granulation tissue in the fistula demonstrated acid-fast bacilli. A
second surgical procedure, performed through a posterior approach,
demonstrated an osteolytic lesion of the tibial metaphysis and epiphysis
extending through the growth plate. This lesion was curetted aggressively. As
a result of the curettage, the center of the growth plate contained a defect
that measured 1.5 × 2.0 cm (Figs. 3-A
and 3-B). Chemotherapy was switched to 200 mg/day of isoniazid,
140 mg/day of rifampicin, and 28 mg/72 hr of streptomycin for one month and
then to 200 mg/day of isoniazid and 140 mg/day of rifampicin for nine more
months.
After the second surgical procedure, the tibia was immobilized in a long
leg cast and no weight-bearing was permitted for three months. Range-of-motion
knee exercises and walking with an ischial weight-bearing brace were then
begun. Postoperative radiographs showed reduction of the area of the bone
defect in the proximal tibial epiphysis, and weight-bearing was permitted
(Figs. 4-A and 4-B).
A follow-up radiograph had been made annually for the nine years prior to
the time of writing (Fig. 4-C),
and the growth plate of the proximal part of the tibia was still open when the
patient was twelve years of age (Fig.
4-D). The leg lengths remained equal, and there was no angular
deformity of the involved knee.
The incidence of bone tuberculosis in children is low. Pulmonary
lesions are rarely associated with tuberculous
osteomyelitis4, and
the BCG vaccination itself may cause
osteomyelitis5,
although the origin of the infection is sometimes unclear. The route of
infection in our patient was also unclear. Bacteriological and histological
examination suggested the tuberculosis after surgical débridement and
curettage. However, polymerase chain reaction tests did not reveal the DNA of
Mycobacterium tuberculosis in the material draining from the infected
area of the tibia. It took eighty days after the patient's first visit to the
hospital to establish the correct diagnosis. Chen et al. reported an average
6.6-month delay in the diagnosis of tuberculous
osteomyelitis3,
indicating the difficulty in making the diagnosis quickly.
There are few reported cases in which the origin of osteomyelitis was in
the epiphysis of the proximal part of the
tibia6-8.
Radiographs of our patient showed an osteolytic lesion in both the metaphysis
and the epiphysis, although the main osteomyelitic lesion was thought to be in
the metaphysis. During the first surgical débridement, an abscess with
posterior cortical disruption was found in the metaphysis. The growth plate
did not appear to be damaged even though the initial magnetic resonance image
had shown a change in signal intensity in the proximal tibial epiphysis
alone.
It is well known that growth disturbances occur after trauma to growth
plates. Shear stress, extreme compression force, and traction on the physis
can injure the physeal cartilage. Drill holes through the physis as well as
curettage of the growth plate are also thought to cause growth disturbances in
some instances. However, the extent of curettage of the growth plate that
would cause its early closure is unknown. Makela et al. reported that injury
to 7% of the cross-sectional area of a rabbit femoral growth plate caused
growth disturbance9.
Osterman reported, in an experimental study, that resection of 65% of the
physis caused early closure but the closure could be prevented by inserting an
interpositional fat
graft10. One
possible mechanism of early closure of the growth plate was described by
Barash and
Siffert11, who
suggested that progression of ossification of the epiphysis facilitated the
formation of bridging callus through the growth plate but not the metaphysis.
Kameyama reported that the low growth-plate activity due to aging would be
expected to retard the regeneration of the growth
plate12.
In our patient, a large area of the metaphysis and epiphysis was curetted
through the growth plate. The growth plate was damaged in an area 1.5 cm wide
and 2 cm deep, which was half of the whole diameter in the sagittal plane and
one-third of the whole diameter in the coronal plane, or one-sixth of the
total cross-sectional area of the physis.
This case should remind orthopaedic surgeons to consider tuberculous
osteomyelitis in the differential diagnosis of bone infection in a child.
Thorough surgical débridement is needed to eradicate the infection even
if the physis is involved, since there may be full recovery following
substantial physeal damage, as there was in our patient, despite the extensive
curettage. ?