NSJ-Spine 10/99.

live 9/7/99 5:57 AM Page 163

J Neurosurg (Spine 2) 91:163–169, 1999

Transpedicular instrumentation as an adjunct in the

treatment of thoracolumbar and lumbar spine
tuberculosis with early stage bone destruction

TAO-CHEN LEE, M.D., KANG LU, M.D., LIN-CHENG YANG, M.D.,

HSUAN-YING HUANG, M.D., AND CHENG-LOONG LIANG, M.D.
Departments of Neurosurgery, Anesthesiology and Pathology, Chang Gung Medical College,
and Chang Gung Memorial Hospital, Kaohsiung, Taiwan; and Department of Medical Research,
Ta Jen College of Technology and Medicine, Ping Tung, Taiwan

Object. Because modern imaging techniques now allow for early diagnosis of spinal tuberculosis, more conservative
management options are possible. The authors evaluated the effectiveness of transpedicular instrumentation for treatment
of thoracolumbar and lumbar spinal tuberculosis in patients with mild bone destruction and the main symptom of “insta-
bility catch” (a sudden painful “snap” that occurs when one extends from a forward bent to an upright position).
Methods. Eighteen patients (nine men and nine women, age range 49–71 years) with spinal tuberculosis were treated
with transpedicular instrumentation that was supplemented with posterolateral fusion and chemotherapy. All patients
were wheelchair dependent or bed-ridden due to severe instability catch, with a mean symptom duration of 2.5 months
(range 1–6 months). Two contiguous vertebrae were involved in 17 patients, and a single vertebrae was involved in one.
In five patients mild neurological deficits (Frankel Grade D) were present. During surgery, the screws were implanted
into the two nonaffected pedicles nearest the lesion to stabilize the involved segments. No attempt at radical debridement
or neural decompression was undertaken. The follow-up period ranged from 21 to 40 months. Postoperatively the insta-
bility catch was relieved within 10 days (excellent outcome) and within 1 month (good outcome) in seven and eight
patients, respectively, and within 3 months (fair outcome) in two; in the remaining patient, the symptom did not resolve
(poor outcome). A short duration of symptoms (generally  3 months) and bone destruction of less than 50% in the
involved vertebral bodies were observed in patients who made a good or excellent outcome. During the follow-up peri-
od, good maintenance of spinal alignment, stabilization of the involved segment, and resolution of the inflammatory
process were shown; however, there was no strong evidence that fusion had occurred at the bony defect. Patients in whom
a fair outcome was achieved experienced a longer duration of symptoms, and in each, one vertebral body with greater
than 50% bone destruction was demonstrated. However, good maintenance of spinal alignment was also shown during
the follow-up period. The patient whose outcome was poor had the longest history (6 months) of symptoms and the most
extensive involvement of the spine ( 50% destruction of two adjacent lumbar vertebral bodies). Postoperatively, implant
failure occurred and the patient developed a wound infection.
Conclusions. Transpedicular instrumentation provides rapid relief of instability catch and prevents late angular
deformity in patients with thoracolumbar and lumbar spinal tuberculosis in whom limited ( 50%) bone destruction
of the involved vertebral bodies has been shown and whose main symptom is instability catch.

KEY WORDS • transpedicular instrumentation • spinal tuberculosis • instability catch

tuberculosis is not an uncommon disease enti-
S
PINAL
ty in underdeveloped and developing countries.3,6 In
the United States, the incidence of spinal tuberculo-
sis has been increasing in the past 10 years.22,26
The “Hong Kong operation,” which involves radical re-
section of the tuberculous-infected focus and bone grafting,
was developed prior to the advent of modern imaging
modalities.9,10 At that time, early diagnosis of spinal tuber-
culosis was difficult, and many patients with this disease
had already developed severe neurological deficits and
advanced bone destruction by the time of diagnosis. Conse-
quently, radical debridement, neural decompression, and re-
construction of the vertebral column were often warranted.
With the advent of computerized tomography (CT) scan-
ning and magnetic resonance (MR) imaging, early diagno-
sis of spinal tuberculosis has become possible.4,7,12,28 In
addition, more effective regimens of antituberculous che-
motherapy have become available.20,24 Therefore, the treat-
ment strategy for this disease entity has been revised and
has become more conservative in recent years.1,6,8,11,20,22–24,26
Modern transpedicular instrumentation can provide rigid
stabilization of the thoracolumbar and lumbar spine with
only short segments required. It has been proven effective
for the treatment of many thoracolumbar and lumbar spinal
disorders that result in segmental instability.2,5,13–19,27,29 This
operation, which is performed via a posterior approach, is
especially feasible for treatment of spinal tuberculosis,
because the tuberculous infection usually involves the ante-
rior column.4,7,12,26,28,30 Its main shortcomings are that it does
not allow for radical resection of the infectious focus and

J. Neurosurg: Spine / Volume 91 / October, 1999 163

NSJ-Spine 10/99.live 9/7/99 5:57 AM Page 164
does not enable satisfactory interbody fusion. However, in a
recent study by the Medical Research Council Working
Party, the authors showed that medical treatment alone can
achieve complete or partial fusion at the bone defect in more
than 80% of the patients with spinal tuberculosis.20 This
result encouraged us to introduce transpedicular instrumen-
tation for the treatment of spinal tuberculosis in patients
whose bone defect is too mild to warrant more radical inter-
ventions but in whom chemotherapy might be inadequate.
Clinical Material and Methods
Between January 1993 and February 1996, 24 patients
with severe instability catch, whose imaging findings sug-
gested thoracolumbar or lumbar spinal tuberculosis under-
went surgery in which transpedicular instrumentation was
placed and a tissue biopsy sample was obtained. Of these,
18 patients in whom there was histological proof of tuber-
culous granulomatous inflammation were included in this
retrospective study. The other six patients (including one
in whom a final diagnosis of Salmonella infection was
made) with pathological findings of only acute or chronic
inflammation were excluded.
The following imaging features were considered patho-
gnomonic of spinal tuberculosis: 1) plain x-ray film evi-
dence of focal erosion in the anterior corner of the vertebral
body adjacent to the end plate; 2) CT evidence of anterior
vertebral body destruction with paraspinal soft-tissue in-
volvement; and 3) MR imaging findings indicative of ver-
tebral body and paraspinal soft-tissue mass with decreased
signal intensity on T1-weighted sequence, increased signal
intensity on T2-weighted sequence, and rims of enhance-
ment on gadolinium-DTPA–enhanced sequences.4,7,12,28
Preoperatively, all patients had undergone plain radiogra-
phy and MR studies, and two patients had undergone an
additional CT study.
The indications for placement of transpedicular instru-
mentation were radiologically visible bone destruction in the
thoracolumbar or lumbar spine and the clinical symptom
of instability catch severe enough to render the patients ful-
ly wheelchair dependent or bedridden. “Instability catch,”
which implies an unstable spine, was defined by Nachem-
son21 as “a sudden painful snap when someone extends from
a forward bent position into the upright position.”
There were nine men and nine women, with a mean age
of 58.8 years (range 49–71 years). The spinal lesions were
located in the thoracolumbar (five cases) or lumbar (13
cases) spine, from T-9 to L-5. Two contiguous vertebrae
were involved in 17 patients and a single vertebra in one.
Prominent ( 15˚) angular deformity was found only in one
patient. The extent of bone involvement was determined
using MR imaging, which is sensitive in detecting early
stage inflammatory process.4,7 All patients experienced an
intractable instability catch, fulfilling the criteria by which
we determined the need for surgical stabilization. In addi-
tion to the back pain, five patients had sustained a mild neu-
rological deficit (Frankel Grade D) that resulted from gran-
ulation tissue or abscesses that caused neural compression.
The duration of clinical symptoms ranged from 1 to 6
months, with a mean of 2.5 months. In all patients a moder-
ate elevation of erythrocyte sedimentation rate was dem-
onstrated (range 35–70 mm/hour; mean 50 mm/hour).
Because tuberculosis remains endemic and bacillus Calmette–
164
T. C. Lee, et al.
Guérin vaccine is still administered in this country, the tuber-
culin skin test is not diagnostic and was therefore not used in
this series for preoperative diagnosis of tuberculosis.
During the operations, the screws were inserted into the
two nonaffected cranial and caudal pedicles nearest to the
bone destruction. That is, the screws could be inserted into
the infected vertebra if the infectious focus did not involve
the pedicles. The instrumentation was routinely supple-
mented with posterolateral arthrodesis in which autogenous
bone harvested from posterior iliac crest was used. A biop-
sy sample of the granulation tissue was obtained via a
transpedicular route or by a limited laminectomy, depend-
ing on the location of the infectious focus. When accessi-
ble, the abscess was drained during the operation. How-
ever, extensive laminectomy or radical debridement was
not attempted because neurological function worse than
Frankel Grade D was not demonstrated in any of the pa-
tients in this series. The biopsy specimens were sent for
pathological examination, and the pus obtained from the
abscesses was sent for culture for mycobacterium tubercu-
losis. As the polymerase chain reaction technique did not
become available in our hospital until 1997, none of the
specimens obtained in this series was examined using this
new technique. The patients, who wore Taylor braces, were
encouraged to move around on the 2nd postoperative day.
Antituberculous chemotherapy (rifampin 450 mg daily,
isoniazid 300 mg daily, and ethambutol 1200 mg daily) was
instituted the day on which radiological diagnosis of spinal
tuberculosis was made and it was continued for 12 months.
Results
The treatment results were shown in Table 1. The duration
of surgery ranged from 3 to 4.5 hours (mean of 3.5 hours).
The blood loss ranged from 100 to 300 ml (mean of 220 ml).
No patient required intraoperative blood replacement.
During the operation, the granulation tissue was found in
the vertebral bodies in 10 patients and in the epidural space
in eight. Histological examination of the granulation tissues
invariably showed features typical of tuberculous granulo-
matous inflammation: infiltration of lymphocytes and epi-
theloid cells with areas of caseating necrosis and sporadic
Langhans’-type giant cells (Fig. 1). Acid-fast bacilli were
not found in any of the biopsy specimens. Epidural or intra-
osseous abscesses were found and drained during surgery
in six patients. However, mycobacterium tuberculosis grew
in only one of the cultures from the abscess specimens. The
erythrocyte sedimentation rate returned to normal in all
patients after completion of chemotherapy.
The assessment of operative results was based mainly on
the relief of instability catch because patients in this series
experienced no, or only mild, neurological deficits. The
results were categorized as follows: excellent, complete re-
lief of instability catch within 10 days postoperatively; good,
complete relief of instability between 10 days and 1 month
postoperatively; fair, complete relief of instability between 1
and 3 months postoperatively; and poor, persistent instabili-
ty catch throughout the follow-up period. The follow-up
periods ranged from 21 to 40 months (mean of 28 months).
Excellent/Good Treatment Results
According to this categorization, excellent and good
treatment results were achieved in seven and eight patients,
J. Neurosurg: Spine / Volume 91 / October, 1999
NSJ-Spine 10/99.live 9/7/99 5:57 AM Page 165

Transpedicular instrumentation for spinal tuberculosis

TABLE 1
Clinical characteristics in 18 patients with spinal tuberculosis treated with transpedicular instrumentation placement and chemotherapy*
Preop Characteristics

Dur- Site No. of Postop Characteristics

ation of of In- Verte-
Age Symp- volved brae W/ Fusion Spinal Follow
Case (yrs) Frankel tom Verte-  50% Relief Frankel at Align- Compli- Up
No. Sex IC Grade (mos) brae VBD of IC Grade Defect ment cations (mos) Outcome

1 49, F + E 2 L2–3 0 10 days E – yes – 30 excellent

2 55, M + E 2 L4–5 0 10 days E – yes – 24 excellent
3 58, M + E 1 L-4 0 10 days E – yes – 25 excellent
4 51, F + E 3 L3–4 0 10 days E – yes – 36 excellent
5 54, F + E 2 T9–10 0 10 days E – yes – 25 excellent
6 59, M + E 3 L3–4 0 10 days E – yes – 30 excellent
7 58, F + E 1 L2–3 0 10 days E – yes – 23 excellent
8 59, F + E 1 T10–11 0 10 days to 1 mo E – yes – 21 good
9 71, F + E 3 L3–4 0 10 days to 1 mo E – yes – 24 good
10 56, M + D 2 L4–5 0 10 days to 1 mo E – yes – 29 good
11 54, F + E 3 L4–5 0 10 days to 1 mo E – yes – 40 good
12 62, M + E 3 L3–4 0 10 days to 1 mo E – yes – 30 good
13 53, M + E 1 T10–11 0 10 days to 1 mo E – yes – 30 good
14 64, M + D 2 L3–4 0 10 days to 1 mo E – yes – 25 good
15 60, M + E 2 L1–2 0 10 days to 1 mo E – yes – 23 good
16 62, F + D 3 T12– L1 1 1 mo to 3 mos E – yes infection 28 fair
17 65, F + D 5 L3–4 1 1 mo to 3 mos E – yes – 30 fair
18 68, M + D 6 L4–5† 2 ‡ D – no|| infection 36 poor
& implant
failure
*Abbreviations: IC = instability catch; VBD = vertebral body destruction; + = present; – = absent.
† There was a skip lesion at cervical spine.
‡ Instability catch was not relieved during the follow-up period.
|| The implant was removed due to complication.

respectively. These 15 patients (83%, Cases 1–15; Table 1)
had satisfactory treatment results with similar clinical pic-
tures and imaging findings and will be discussed as one
group hereafter. These patients had a short history of clini-
cal symptoms prior to admission—less than 3 months in all
of them. Most (13 of 15) patients had not experienced neu-
rological deficits and in only two was a mild (Frankel Grade
D) neurological deficit demonstrated. The bone involve-
ment in this group was shown to have the following charac-
teristics: 1) only short segments (one or two vertebrae) were
involved; 2) none of the involved vertebrae showed more
than 50% destruction of the body (Fig. 2 upper left and cen-
ter); 3) absence of a prominent ( 15˚) angular deformity
was demonstrated; and 4) there was absence of a skip lesion
elsewhere. All of these 15 patients experienced great im-
provement of the preexisting instability catch, despite the
superficial wound pain, and they were able to move around
on foot or in a wheelchair by the 3rd postoperative day.
Of the patients in this group, the two patients in whom a
Frankel Grade D neurological deficit was demonstrated im-
proved slowly, but progressively, and became neurologically
normal between the 1- and 2-month follow-up examination.
In all 15 patients, examination of follow-up plain radi-
ographs showed posterolateral fusion masses (Fig. 2 upper
right). Although a small radiolucent zone ( 2 mm) was
shown around the screws in the patients, follow-up radiogra-
phy demonstrated obvious remineralization at the bone
defect, maintenance of the spinal alignment, and stabiliza-
tion of the involved segment in all 15 cases (Fig. 2 lower left
and center). In eight patients titanium implants were used,
and they underwent MR imaging during the follow-up ex-
aminations. These MR studies invariably demonstrated reso-
lution of the inflammatory process with marked shrinkage of
the preexisting paraspinal soft-tissue mass, normalization of
the signal intensities of the involved vertebrae, and filling of
the bone defect with a solid mass (Fig. 2 lower right). How-
ever, absence of trabeculae within the solid masses implied
that there was no strong evidence of fusion at the defects.
Fair Treatment Results
Two patients (Case 16 and 17) in this series were cate-
gorized as having fair outcomes. In one patient there was a
5-month history of symptoms. In both patients two verte-
brae were involved and mild neurological compromise was
demonstrated (Frankel Grade D). In both cases, the pre-
dominantly involved vertebra showed more than 50% de-
struction of the vertebral body on preoperative imaging
studies (Fig. 3 upper left and right). Postoperatively, both
patients experienced some improvement of the instability
catch and were able to use a wheelchair within 1 week and
to begin ambulating with a walker within 1 month. Return
to normal neurological function was demonstrated between
the 2- and 3-month follow-up visits. One presented with
postoperative wound infection. We partially opened the
wound to drain the whitish pus. Culture of the pus did not
grow bacteria or tuberculous bacilli. Because the patient’s
general condition was good, we treated her with frequent
dressing changes and continued antituberculous medica-
tion. Thereafter the wound healed completely within 1
month. Despite the presence of an obvious ( 2 mm) area

J. Neurosurg: Spine / Volume 91 / October, 1999 165

NSJ-Spine 10/99.live 9/7/99 5:57 AM Page 166
FIG. 1. Photomicrograph of the granulation tissue obtained in a
patient, demonstrating caseating necrotic debris in the right upper
corner surrounded by an admixture of lymphocytes and epitheloid
histiocytes. A Langhans’-type giant cell can also be seen in the left
lower corner (arrow). H & E, original magnification  160.
of radiolucency around the screws and minimal remineral-
ization at the bone defect, maintenance of the spinal align-
ment on follow-up radiographs was demonstrated in both
patients (Fig. 3 lower left). The long-term (30-month) fol-
low-up MR images revealed shrinkage of the preexisting
paraspinal soft-tissue mass, normalization of the signal
intensity of the involved vertebrae, and filling of the bone
defect with a solid mass in both patients, but no evidence of
fusion at the bone defect (Fig. 3 lower right).
Poor Treatment Result
Only one patient (Case 18) was categorized as having a
poor result. This patient had the longest history (6 months)
of clinical symptoms as well as extensive ( 50%) destruc-
tion of the vertebral body of the two affected lumbar verte-
brae. He also harbored a tuberculous lesion in the cervical
spine region that was observed at a previous operation.
During the instrumentation placement procedure, infectious
material was found in the vertebral body while we probed a
presumed healthy pedicle and into the rostral portion of one
involved vertebral body. In the study by Oga and col-
leagues,23 they reported that mycobacterium tuberculosis
does not adhere well to inert surfaces and that spinal instru-
mentation does not pose a risk of spreading the infection;
therefore, we chose to place the screw in the infectious focus
rather than the adjacent noninvolved vertebra. Unfortun-
ately, this resulted in loosening of the screw and spread of
infectious material from the vertebral body into the subcu-
taneous space through the iatrogenic pedicular tunnel.
Consequently, the patient developed a wound infection,
with draining sinuses, and pullout of the implant. A subse-
quent operation was performed to remove the implant and
debride the subcutaneous granulation tissue. The wound
healed approximately 1 month after this subsequent surgery.
The follow-up imaging studies demonstrated progression of
166
T. C. Lee, et al.
the bone destruction. This patient experienced persistent
instability catch and neurological deficit and was bedridden
throughout the 36-month follow-up period. Although we
had planned further surgical correction with anterior de-
bridement and fusion, the patient’s general condition re-
mained poor, and he was eventually lost to follow-up study.
Discussion
Despite the general acceptance of the Hong Kong opera-
tion for the treatment of spinal tuberculosis, a recent study
in Korea by the Medical Research Council Working Party
found that chemotherapy alone is efficacious for spinal
tuberculosis, even when myelopathy or obvious bone
destruction is present.20 They reported that good clinical
recovery was attained in 92% of patients and that complete
or partial fusion with a mean increase of kyphosis less than
18˚ by 36 months was obtained in 84% of patients. Their
finding that a tuberculous lesion may heal with spontaneous
fusion indicates that the Hong Kong operation is no longer
the sole treatment option for patients with spinal tuberculo-
sis. In recently published papers, authors have tended to
emphasize the importance of tailoring the management
options to the individual needs of the patients with spinal
tuberculosis. Nussbaum, et al.,22 have treated patients with
spinal tuberculosis according to the degree of bone destruc-
tion. In their series, aggressive debridement and fusion were
performed only in patients in whom advanced vertebral
body involvement resulted in kyphosis. Rezai and associ-
ates,26 have reported that radical operative management was
required only when the extent of vertebral body destruction
exceeds 50%. Boachie-Adjei and Squillante1 have suggest-
ed that neurological function as well as bone involvement
should be carefully evaluated before making surgery-relat-
ed decisions for the treatment of patients with spinal tuber-
culosis. They recommended that patients with Frankel
Grade A or B lesions undergo rapid decompressive surgery,
while those with Grade C or D lesions may be treated
expectantly with chemotherapy alone.
Because most of the patients in our series were neurolog-
ically normal and had no severe bone destruction, theoreti-
cally they could be expected to have had good treatment
outcomes by undergoing antituberculous chemotherapy
alone. However, this conservative method cannot prevent
the possible progression of kyphotic deformity, and it usual-
ly requires long-term rest to achieve relief of severe back
pain.24 Based on analysis of the literature, a rigid stabiliza-
tion system provides the best solution, not only to prevent
kyphosis but also to achieve relief of pain due to spinal
instability.2,5,13–18,27,29 This concept was introduced by Garst,6
who first used the technique of posterior fusion without
instrumentation to replace drastic anterior interbody fusion
for spinal tuberculosis. He reported that this safer procedure
can enable the spine to become stable and pain free and that
it can be performed in developing countries where facilities
for transthoracic procedures are not available. Guven and
coworkers8 have used posterior instrumentation (with a
screw and/or hook system) for the treatment of thoracolum-
bar spinal tuberculosis in 10 neurologically intact patients,
and clinical and radiological evidence of stable fusion was
obtained in all their patients.
In the present study, we used transpedicular instrumen-
J. Neurosurg: Spine / Volume 91 / October, 1999
NSJ-Spine 10/99.live 9/7/99 5:57 AM Page 167

Transpedicular instrumentation for spinal tuberculosis

FIG. 2. Imaging studies obtained in a 62-year-old man whose treatment outcome was good. Upper Left:
Preoperative plain radiograph revealing slight posterior displacement of the L-3 vertebra onto the L-4 vertebra and bone
erosion (arrow) at the anteroinferior aspect of the L-3 body. Upper Center: Preoperative noncontrast T1-weighted MR
image (TR 516 msec, TE 17 msec) of the affected L-3 and L-4 vertebrae and the paravertebral soft-tissue mass (arrow),
demonstrating low signal intensity. Upper Right: Anteroposterior radiograph obtained at 30-month follow-up exami-
nation revealing the intertransverse fusion mass (arrow) and limited laminectomy. Lower Left and Center: Lateral flex-
ion (left) and extension (right) radiographs obtained at 30-month follow-up examination revealing maintenance of the
spinal alignment, presence of minimal radiolucency around the upper screws, and remineralization at the bone defect.
Lower Right: Postoperative T1-weighted MR image (TR 516 msec, TE 17 msec) without contrast enhancement obtained
at 30-month follow up, demonstrating normalization of the signal intensity of the involved vertebrae, shrinkage of the
paravertebral soft-tissue mass, and filling of the bony defect with a solid mass. However, absence of trabeculae inside the
solid mass implies that there is no strong evidence of fusion at the bone defect.

tation because it has been proven to be a good method for
stabilizing the thoracolumbar and lumbar spine and is not
as drastic an intervention as the anterior radical operation.
This posterior approach has a special advantage in that it
avoids contact with the infectious focus, because the
tuberculous infection usually involves the anterior col-
umn.26,28,30 The fact that tuberculous lesions may heal in
spontaneous fusion makes this posterior approach feasi-
ble.20 During the placement of instrumentation, we insert-
ed the screws into the healthy pedicles adjacent to the
bone destruction. This method has also been used by Rath
and colleagues25 who suggested that screws can be placed
into affected vertebrae if the upper part of the vertebral
body is not destroyed by infection.
Analysis of our results shows that transpedicular instru-
mentation is highly effective for quick relief of instability in
patients with early stage spinal tuberculosis. Good or excel-
lent operative results were obtained in 15 patients (83%); in
none of these patients was a vertebra with greater than 50%
vertebral body destruction or a history of symptoms lasting
longer than 3 months demonstrated, and most (13 of 15)
patients were neurologically intact. In addition to providing
rapid relief from instability catch, the instrumentation also
preserved spinal alignment and stabilized the involved seg-
ment in these patients. Although the mass in the quiescent
tuberculous lesions in these patients cannot be considered to
be as good as a fusion mass, we believe, it may function as
a “spacer” to support the anterior column. In experienced
hands, this operation is not time consuming nor does it
require blood replacement. It can allow early ambulation
and shorten hospital stay.
In the two patients in whom fair operative results were
obtained, relatively extensive involvement of the spine was
demonstrated (one vertebra with  50% destruction of the
vertebral body). Compared with the 15 patients in whom
satisfactory operative results were achieved, these two pa-

J. Neurosurg: Spine / Volume 91 / October, 1999 167

NSJ-Spine 10/99.live 9/7/99 5:57 AM Page 168
FIG. 3. Imaging studies obtained in a 65-year-old woman whose
treatment outcome was considered fair. Upper Left: Preoperative
plain radiograph demonstrating severe bone destruction (arrow) at
the inferior portion of the L-3 and superior portion of L-4 vertebral
bodies. Upper Right: Preoperative T1-weighted, gadolinium-DTPA
contrast–enhanced MR image (TR 716 msec, TE 11 msec) reveal-
ing extensive bone destruction of the L-4 vertebra ( 50% destruc-
tion of the vertebral body). Areas of high signal intensity can be
seen at the involved bodies and the paravertebral soft-tissue mass
(arrow). The site of bone destruction shows low signal intensity.
Lower Left: Postoperative plain radiograph obtained at 30-month
follow up demonstrating an area of radiolucency around the upper
screws as well as minimal remineralization at the bone defect;
however, spinal alignment has been maintained. Lower Right:
Postoperative gadolinium-DTPA contrast–enhanced MR image
(TR 400 msec, TE 14.6 msec) at 30-month follow-up examination
revealing almost complete disappearance of the paravertebral soft-
tissue mass and normalization of the signal intensity of the L-3 and
L-4 vertebral bodies. Although the bone defect has been filled with
a solid mass, there is no evidence of bony fusion.
tients had a longer history of symptoms ( 3 months).
Moreover, both had sustained a neurological deficit, al-
though it was mild. We believe that the spinal alignment in
these two patients was maintained because of the spacer
effect that was created by the formed mass in the quiescent
tuberculous lesion. However, the persistent obvious radiolu-
cent zone around the screws and the rather late relief ( 1
month) of instability catch indicated poor stabilization of the
involved spinal segment in these two patients. Because of
their relatively extensive bone destruction of the vertebrae,
a more aggressive treatment strategy—with extensive de-
bridement and fusion—may have been preferable. In cases
of recurrence of back pain due to obvious implant failure or
168
T. C. Lee, et al.
recurrence of disease, the use of the conventional radical
surgery (debridement and fusion) is recommended.
In our series, only one patient was categorized as hav-
ing a poor result. The poor outcome in this case was part-
ly related to the relatively late diagnosis, multiple infec-
tious foci, and advanced bone destruction, and it was partly
caused by surgery-related complications. The direct inser-
tion of a screw into the infectious focus resulted in implant
failure, as well as spreading of infectious material to the
subcutaneous space. Our experience in this case further
supports that patients with relatively extensive ( 50%)
vertebral body destruction should be initially treated with
anterior debridement and fusion. In addition, we must
emphasize that direct insertion of a screw into an infected
focus should be avoided during transpedicular instrumenta-
tion placement in patients with spinal tuberculosis.
Conclusions
In the era of modern imaging modalities and effective
antituberculous medication, patients with early diagnosed
spinal tuberculosis may be treated with more conservative
surgical methods or even medical treatment alone. In
patients with a severe instability catch caused by bone
destruction, surgical stabilization may facilitate the reso-
lution of this pain. Transpedicular instrumentation, a pos-
teriorly placed system, and posterolateral fusion are espe-
cially advantageous for thoracolumbar and lumbar spinal
tuberculosis, which frequently involves the anterior por-
tions of the vertebral bodies. This less drastic modality can
achieve rapid relief of instability catch in patients with
limited ( 50% destruction) vertebral body involvement.
Although this procedure did not result in obvious bone
fusion at the involved site, spinal alignment was main-
tained during the follow-up period that ranged from 21 to
40 months. However, further follow-up review is neces-
sary to monitor for possible disease recurrence or implant
failure. In our experience, the screws can be inserted into
the healthy pedicles immediately adjacent to the area of
bony destruction. However, direct insertion of the screws
into the infectious focus should be avoided.
References
1. Boachie-Adjei O, Squillante RG: Tuberculosis of the spine.
Orthop Clin North Am 27:95–103, 1996
2. Bradford DS: Instrumentation of the lumbar spine: an overview.
Clin Orthop 203:209–218, 1986
3. Chen WJ, Chen CH, Shih CH: Surgical treatment of tubercu-
lous spondylitis. 50 patients followed for 2–8 years. Acta
Orthop Scand 66:137–142, 1995
4. Desai SS: Early diagnosis of spinal tuberculosis by MRI. J
Bone Joint Surg (Br) 76:863–869, 1994
5. Dick W: The “ fixateur interne” as a versatile implant for spine
surgery. Spine 12:882–900, 1987
6. Garst RJ: Tuberculosis of the spine: a review of 236 operated
cases in an underdeveloped region from 1954 to 1964. J Spinal
Disord 5:286–300, 1992
7. Gupta RK, Gupta S, Kumar S, et al: MRI in intraspinal tuber-
culosis. Neuroradiology 36:39–43, 1994
8. Guven O, Kumano K, Yalcin S, et al: A single stage posterior
approach and rigid fixation for preventing kyphosis in the treat-
ment of spinal tuberculosis. Spine 19:1039–1043, 1994
9. Hodgson AR, Stock FE: Anterior spine fusion for the treatment
J. Neurosurg: Spine / Volume 91 / October, 1999
NSJ-Spine 10/99.live 9/7/99 5:57 AM Page 169
Transpedicular instrumentation for spinal tuberculosis
of tuberculosis of the spine. J Bone Joint Surg (Am) 42:
295–310, 1960
10. Hodgson AR, Yau ACMC, Kwon JS: A clinical study of 100
consecutive cases of Pott’s paraplegia. Clin Orthop 36:
128–150, 1964
11. Janssens JP, de Haller R: Spinal tuberculosis in a developed coun-
try. A review of 26 cases with special emphasis on abscesses and
neurologic complications. Clin Orthop 257:67–75, 1990
12. Kim NH, Lee HM, Suh JS: Magnetic resonance imaging for the
diagnosis of tuberculous spondylitis. Spine 19:2451–2455, 1994
13. Lee TC: Complications of transpedicular reduction and stabi-
lization of the thoracolumbar spine. J Formos Med Assoc
94:738–741, 1995
14. Lee TC: Pedicle fixation: an adjuvant for the treatment of thora-
columbar metastases. Ann Acad Med Singapore 22:
418–421, 1993
15. Lee TC: Reduction and stabilization without laminectomy for
unstable degenerative spondylolisthesis: a preliminary report.
Neurosurgery 35:1072–1076, 1994
16. Lee TC: A simple method for transpedicular instrumentation of
the lower lumbar spine. How we do it: technical note. Acta
Neurol Taiwan 6:12–19, 1997
17. Lee TC: Transpedicular reduction and stabilization for post-
laminectomy lumbar instability. Acta Neurochir 138:139–145,
1996
18. Lee TC: Treatment of unstable thoracolumbar spine with “AO
internal fixator.” Acta Neurol Sinica 2:316–319, 1993
19. Lee TC, Yang LC, Chen HJ: Effect of patient position and
hypotensive anesthesia on inferior vena caval pressure. Spine
23:941–948, 1998
20. Medical Research Council Working Party on Tuberculosis of the
Spine: controlled trial of short-course regimens of chemotherapy
in the ambulatory treatment of spinal tuberculosis. Results at
three years of a study in Korea. J Bone Joint Surg (Br) 75:
240–248, 1993
J. Neurosurg: Spine / Volume 91 / October, 1999
21. Nachemson A: The role of spine fusion. Question 8. How do
you define instability? How is this diagnosed, and what surgi-
cal treatment policy do you follow? Spine 6:306–307, 1981
22. Nussbaum ES, Rockswold GL, Bergman TA, et al: Spinal
tuberculosis: a diagnostic and management challenge. J Neuro-
surg 83:243–247, 1995
23. Oga M, Arizono T, Takasita M, et al: Evaluation of the risk of
instrumentation as a foreign body in spinal tuberculosis. Clin-
ical and biologic study. Spine 18:1890–1894, 1993
24. Pun WK, Chow SP, Luk KD, et al: Tuberculosis of the lumbo-
sacral junction. Long-term follow-up of 26 cases. J Bone Joint
Surg (Br) 72:675–678, 1990
25. Rath SA, Neff U, Schneider O, et al: Neurosurgical manage-
ment of thoracic and lumbar vertebral osteomyelitis and disci-
tis in adults: a review of 43 consecutive surgically treated pa-
tients. Neurosurgery 38:926–933, 1996
26. Rezai AR, Lee M, Cooper PR, et al: Modern management of
spinal tuberculosis. Neurosurgery 36:87–98, 1995
27. Roy-Camille R, Saillant G, Mazel C: Internal fixation of the
lumbar spine with pedicle screw plating. Clin Orthop 203:
7–17, 1986
28. Shanley DJ: Tuberculosis of the spine: imaging features. AJR
164:659–664, 1995
29. Steffee AD, Biscup RS, Sitkowski DJ: Segmental spine plates
with pedicle screw fixation. A new internal fixation device for
disorders of the lumbar and thoracolumbar spine. Clin Orthop
203:45–53, 1986
30. Tibau R, Fuster S, Auleda J, et al: Tuberculosis of the neural
arch. A report of four cases. Int Orthop 18:119–121, 1994
Manuscript received January 7, 1999.
Accepted in final form July 23, 1999.
Address reprint requests to: Tao-Chen Lee, M.D., 123, Ta-Pei
Road, Niao-Sung Hsiang, Kaohsiung Hsien, Taiwan.
169