SPINE Volume 27, Number 3, pp E64–E70
©2002, Lippincott Williams & Wilkins, Inc.
An In Vitro Human Cadaveric Study Investigating the
Biomechanical Properties of the Thoracic Spine
Itaru Oda, MD,* Kuniyoshi Abumi, MD,* Bryan W. Cunningham, MSc,†
Kiyoshi Kaneda, MD,* and Paul C. McAfee, MD‡
Study Design. An in vitro human cadaveric study com-
paring the effects of anterior and posterior sequential
destabilization conditions on thoracic functional unit me-
chanics was studied.
Objectives. To investigate the biomechanical proper-
ties of the human thoracic spine.
Summary of Background Data. Few studies have ad-
dressed the mechanical role of the costovertebral joints
under torsion in the stability of the human thoracic spine.
Methods. Sixteen functional spinal units with intact
costovertebral joints were obtained from six human ca-
davers and randomized into two groups based on desta-
bilization procedures: Group 1, anterior to posterior se-
quential resection; and Group 2, posterior to anterior
sequential destabilization. Biomechanical testing was
performed after each destabilization procedure, and the
range of motion under maximum load was calculated.
Results. Group 1: Under flexion– extension, lateral
bending, and axial rotation loading, discectomy increased
the range of motion by 193%, 74%, and 111%, respec-
tively. Moreover, subsequent right rib head resection fur-
ther increased the range of motion by 81%, 84%, and 72%,
respectively. Group 2: Under all loading conditions lami-
nectomy ⫹ medial facetectomy resulted in a 22–30% in-
crease in range of motion. Subsequent total facetectomy
led to an additional 15–28% increase in range of motion.
Conclusion. The rib head joints serve as stabilizing
structures to the human thoracic spine in the sagittal,
coronal, and transverse planes. In anterior scoliosis sur-
gery additional rib head resection after discectomy may
achieve greater curve and rib hump correction. The lateral
portion of the facet joints plays an important role in
providing spinal stability and should be preserved to
minimize postoperative kyphotic deformity and segmen-
tal instability when performing decompressive wide
laminectomy. [Key words: thoracic spine, biomechanics,
costovertebral joint, rib head] Spine 2002;27:E64 –E70
In the thoracic spine the costovertebral joints and their
surrounding ligaments connect adjacent vertebrae and
ribs and provide additional structural stability. There are
some experimental studies dealing with the mechanical
properties of thoracic spine.1,11,12,18 However, mechan-
From the *Department of Orthopaedic Surgery, Hokkaido University
Graduate School of Medicine, Sapporo, Japan, the †Orthopaedic Bio-
mechanics Laboratory, Union Memorial Hospital, Baltimore, and the
‡Scoliosis & Spine Center, St. Joseph Hospital, Towson, Maryland.
Supported by Orthopaedic Associates Research Foundation, Inc.,
Towson, Maryland.
Acknowledgment date: November 2, 2000.
First revision date: February 22, 2001.
Second revision date: May 9, 2001.
Acceptance date: August 1, 2001.
Device status category: 1.
Conflict of interest category: 14.
E64
ical roles of the costovertebral joints and rib cage have
not been well understood.
Recent biomechanical studies using a dog thorax
model investigated the effects of anterior and posterior
sequential destabilization conditions on functional unit
mechanics.8,14 They concluded that the costovertebral
joints play an important role in providing stability, espe-
cially under lateral bending and axial rotation loading,
to the thoracic spine; however, the effects of anatomic
differences between human and dog spines remained un-
determined. Feiertag et al.5 demonstrated the mechanical
significance of rib head joints in a human thorax model;
however, axial rotation tests were not performed.
Kaneda et al.7 reviewed scoliosis cases treated by the
anterior procedure and concluded that rib head resection
combined with discectomy resulted in better correction
of the scoliosis and rib hump versus discectomy alone.
However, no studies to date have evaluated the mechan-
ical effects of rib head resection under axial rotation
loading in the human thoracic spine.
The purpose of this study was to investigate the bio-
mechanical properties of the human thoracic spine and
to indicate the mechanical significance of the anterior
and posterior stabilizers, comparing the effects of se-
quential destabilization conditions on functional unit
mechanics.
Materials and Methods
Specimen Preparation and Biomechanical Testing. A to-
tal of six fresh-frozen human cadaveric specimens were used
in this investigation. The specimens were harvested from
four female and two male cadavers, and the average age ⫾
standard deviation was 68.8 ⫾ 9.8 years. To document ab-
normalities and degenerative changes, anteroposterior and
lateral radiographs were taken. No fractures or other abnor-
malities were found beyond the normal degenerative
changes. In preparation for biomechanical testing, the spec-
imens were thawed at room temperature and cleaned of all
residual musculature, with care taken to preserve all liga-
mentous structures. Sixteen thoracic functional spinal units
(FSUs) with bilateral ribs were then prepared. The integrity
of the ligaments, joint articulations, transverse processes,
and the posterior 5 cm of the ribs, including their vertebral
articulations, were preserved. The spinal levels tested were
as follows: T3–T4, 1 FSU; T4 –T5, 3 FSUs; T5–T6, 4 FSUs;
T6 –T7, 4 FSUs; T7–T8, 2 FSUs; and T8 –T9, 1 FSU. The
ranges of motion (ROMs) of these thoracic segments are
reported to be comparable each other.17
All biomechanical testing was performed using an MTS 858
Bionix Test System (MTS System Inc., Minneapolis, MN). The
 Biomechanical Properties of the Thoracic Spine • Oda et al E65
Figure 1. A typical load– displacement curve is presented. The
fourth cycle was used for data analysis, and the range of motion
(ROM) was defined as the angular displacement under the maxi-
mum load.
upper and lower vertebral bodies were secured in rectangular
steel tubing containers using eight-point compression screws.
For all testing modes the axis of rotation was centered at the
midcolumn of the FSUs, as determined by the center of the
intervertebral disc. For flexion– extension and lateral bending
testing the specimens were placed horizontally, and a rota-
tional torque generated by the MTS machine resulted in pure
flexion– extension or bilateral lateral bending moment around
the axis previously defined at setup.9,10 This testing procedure
was a constrained type of loading and permitted 1 degree of
freedom. Nondestructive biomechanical testing was first per-
formed on the intact specimens to evaluate the intact stability
of the motion segments and included flexion– extension (⫾ 2
Nm), bilateral lateral bending (⫾ 2 Nm), and bilateral axial
Figure 2. The anterior to poste-
rior destabilization procedures
(Group 1) are shown: (1) discec-
tomy and transection of the an-
terior–posterior longitudinal liga-
ments preserving the bilateral rib
head joints (disc); (2) resection of
the right rib head joint (RRH); (3)
removal of the right costotrans-
verse joint (RCTJ); and (4) resec-
tion of the left rib head joint
(LRH).
rotation (⫾ 2 Nm with 100 N compressive preload). These
maximum loading levels chosen were within the range of pre-
vious in vitro studies.3,8 –11,12,14 Each testing mode was re-
peated over a period of five ramp cycles at a rate of ⫾ 0.4 Nm/s.
The first three served as conditioning cycles with the data from
the fourth cycle used for computational analysis. Angular dis-
placement of the FSUs was measured using an Angular Dis-
placement Transducer (Trans Tek, Inc., Ellington, CT) built in
the MTS machine. Load– displacement data acquisition was
performed using transducers built in the MTS machine through
an analog-to-digital DAS16G Metrabyte board (Metrabyte,
Taunton, MN) interfaced with an IBM 486 PS/2 (IBM, Boca
Raton, FL). Therefore, cyclic load– displacement curves were
obtained immediately after testing (Figure 1).
Destabilization Procedures. After intact specimen analysis
the FSUs were randomized into two groups based on destabi-
lization procedures. These destabilization procedures were de-
termined by referring two previous reports.8,14
Group 1 (Anterior to Posterior: T4 –T5, 2 FSUs; T5–T6, 2
FSUs; T6 –T7, 2 FSUs; T7–T8, 1 FSU; T8 –T9, 1 FSU). The
following was performed in Group 1: 1) total discectomy and
transection of the anterior–posterior longitudinal ligaments
preserving the bilateral rib head joints (Disc), 2) resection of the
right rib head joint including radiate and intra-articular liga-
ments (RRH), 3) removal of the right costotransverse joint
including costotransverse and superior costotransverse liga-
ments (RCTJ), and 4) resection of the left rib head joint (LRH)
(Figure 2).
Group 2 (Posterior to Anterior: T3–T4, 1 FSU; T4 –T5, 1
FSU; T5–T6, 2 FSUs; T6 –T7, 2 FSUs; T7–T8, 1 FSU; T8 –T9,
1 FSU). The following was performed in Group 2: 1) laminec-
tomy combined with bilateral medial facetectomy at the medial
margin of the pedicles (LMF), 2) bilateral total facetectomy
(TF), 3) resection of the right costovertebral joint (RCVJ), and
4) removal of the left costovertebral joint (LCVJ) (Figure 3).
The costovertebral joint consists of the costotransverse joint
and rib head joint.
E66 Spine • Volume 27 • Number 3 • 2002

Figure 3. The posterior to ante-

rior destabilization procedures
(Group 2) are shown: (1) laminec-
tomy with bilateral medial facet-
ectomy (LMF); (2) bilateral total
facetectomy (TF); (3) resection of
the right costovertebral joint
(RCVJ); and (4) removal of the left
costovertebral joint (LCVJ).

Data Analysis. Biomechanical testing was first performed

with the intact specimens and then repeated after each de-
stabilization procedure. ROM was defined as the angular
displacement under maximum load (Figure 1). The ROMs
were normalized to the intact condition for each specimen.
Using a Wilcoxon signed rank test, statistical significance of
each destabilization procedure was determined by compar-
ing the ROMs before and after the resection. A one-way
analysis of variance combined with a Student-Newman-
Keuls test was used to compare anterior and posterior desta-
bilization procedures. Statistical significance was indicated
at P ⬍ 0.05.

Results
Flexion–Extension
Mean values and standard deviations of the ROMs un-
der flexion– extension loading are shown in Figure 4.
Group 1. Discectomy significantly increased ROM by
193%, and subsequent right rib head resection further
increased the ROM by 81% (P ⬍ 0.05). Resection of
the right costotransverse joint did not significantly in-
crease the ROM, and subsequent left rib head resec-
tion resulted in an additional 114% increment in
ROM (P ⬍ 0.05).
Group 2. Laminectomy ⫹ medial facetectomy led to a
22% increase in ROM (P ⬍ 0.05), and subsequent total
facetectomy resulted in an additional 28% increase (P ⬍
0.05). Removal of the right costovertebral joint demon-
strated a 15% increase (P ⬍ 0.05); however, resection of
the left costovertebral joint did not significantly increase
the ROM (P ⬎0.05). Figure 4. Range of motion of the thoracic functional units under
flexion– extension loading (percentage to control): (A) Group 1; (B)
Group 2. Error bars indicate standard deviations. Group 1: Disc ⫽
Lateral Bending discectomy with ALL/PLL transection; RRH ⫽ right rib head resec-
Mean values and standard deviations of the ROMs un- tion; RCTJ ⫽ removal of the right costotransverse joint; LRH ⫽ left
der lateral bending loading are illustrated in Figure 5. rib head resection. Group 2: LMF ⫽ laminectomy with medial
facetectomy; TF ⫽ total facetectomy; RCVJ ⫽ removal of the right
Group 1. Discectomy increased the ROM by 79%, and costovertebral joint; LCVJ ⫽ resection of the left costovertebral
subsequent right rib head resection resulted in a 103% joint. Not significant, P ⬎ 0.05.
 Biomechanical Properties of the Thoracic Spine • Oda et al E67
Figure 5. Range of motion of the thoracic functional units under
bilateral lateral bending loading (percentage to control): (A)
Group 1; (B) Group 2. Error bars indicate standard deviations.
Group 1: Disc ⫽ discectomy with ALL/PLL transection; RRH ⫽
right rib head resection; RCTJ ⫽ removal of the right costo-
transverse joint; LRH ⫽ left rib head resection. Group 2: LMF ⫽
laminectomy with medial facetectomy; TF ⫽ total facetectomy;
RCVJ ⫽ removal of the right costovertebral joint; LCVJ ⫽
resection of the left costovertebral joint. Not significant, P ⬎
0.05.
increase (P ⬍ 0.05). Removal of the right costotransverse
joint exhibited a 60% increase (P ⬍ 0.05), and addi-
tional left rib head resection demonstrated a 104% fur-
ther increase in ROM (P ⬍ 0.05).
Group 2. Laminectomy ⫹ medial facetectomy led to a
22% change in ROM (P ⬍ 0.05), and subsequent total
facetectomy further increased ROM by 15% (P ⬍
0.05). Removal of the right costovertebral joint dem-
onstrated a 17% increase (P ⬍ 0.05), and subsequent
left costovertebral joint resection led to a 49% in-
crease (P ⬍ 0.05).
Axial Rotation
Mean values and standard deviations of the ROMs un-
der axial rotation loading are given in Figure 6.
Group 1. Discectomy resulted in 111% increment in
ROM (P ⬍ 0.05), and right rib head resection presented
Figure 6. Range of motion of the thoracic functional units under
bilateral axial rotation loading (percentage to control): (A)
Group 1; (B) Group 2. Error bars indicate standard deviations.
Group 1: Disc ⫽ discectomy with ALL/PLL transection; RRH ⫽
right rib head resection; RCTJ ⫽ removal of the right costo-
transverse joint; LRH ⫽ left rib head resection. Group 2: LMF ⫽
laminectomy with medial facetectomy; TF ⫽ total facetectomy;
RCVJ ⫽ removal of the right costovertebral joint; LCVJ ⫽
resection of the left costovertebral joint. Not significant, P ⬎
0.05.
an additional 72% increase (P ⬍ 0.05). Resection of the
right costotransverse joint resulted in a 66% increase (P
⬍ 0.05), and additional left rib head resection exhibited
a 189% increase (P ⬍ 0.05).
Group 2. Laminectomy ⫹ medial facetectomy and sub-
sequent total facetectomy significantly increased the
ROM by 30% and 15%, respectively (P ⬍ 0.05). Right
costovertebral joint resection indicated a 43% increase
(P ⬍ 0.05), and subsequent left costovertebral joint re-
section resulted in a 39% increment (P ⬍ 0.05).
Anterior Versus Posterior Destabilization
The destabilizing effects of total posterior element re-
moval (laminectomy ⫹ total facetectomy), total discec-
tomy, and total discectomy combined with unilateral rib
head resection are compared because these three proce-
dures are common in surgical correction of the spinal
deformity (Figure 7). A one-way analysis of variance
E68 Spine • Volume 27 • Number 3 • 2002
Figure 7. Ranges of motion of the thoracic functional units after
posterior and anterior destabilizations are compared (percentage
to control). Error bars indicate standard deviations. Posterior ele-
ments ⫽ removal of whole posterior ligamentous complex; Disc ⫽
total discectomy combined with ALL/PLL transection; Disc ⫹ Rt.
Rib Head ⫽ total discectomy with ALL/PLL transection and right
rib head resection. #P ⬍ 0.05 versus posterior elements (one-
way analysis of variance with Student-Newman-Keuls post hoc
test). *P ⬍ 0.05 versus both posterior elements and disc
(one-way analysis of variance with Student-Newman-Keuls
post hoc test).
demonstrated statistical significance among the three
procedures under all loading conditions (flexion–
extension, F ⫽ 55.54, P ⫽ 0.000; lateral bending, F ⫽
15.85, P ⫽ 0.000; axial rotation, F ⫽ 19.96, P ⫽ 0.000).
Total discectomy exhibited significantly greater ROM
versus posterior element removal (P ⬍ 0.05) except in
lateral bending. Discectomy combined with unilateral
rib head resection demonstrated a significant increase in
ROM versus other two resections.
Discussion
In this study the basic biomechanical properties of the
human thoracic functional spinal units were investigated
using human cadaveric specimens. From the anatomic
point of view, human spines are ideal for performing in
vitro biomechanical testing. However, the stabilizing ef-
fects of extrinsic stabilizers, such as the neuromuscular
system, were not considered in the current study. As well,
functional spinal units including bilateral costovertebral
joints were used; however, structural effects of the rib
cage were not investigated. Thus, there are some limita-
tions when applying the acquired data to clinical
situations.
In this study for flexion– extension and lateral bend-
ing testing, the specimens were placed horizontally,
and a rotational torque, generated by the MTS ma-
chine, resulted in pure flexion– extension or bilateral
lateral bending moment. A pure bending moment re-
sults in the most uniform loading of the construct.13
Although the in vivo axis of rotation of the thoracic
spine combined with rib cage after sequential destabi-
lization remains unknown, the fixation method was
constrained using a consistent loading axis as previ-
ously described. Therefore, the authors understand
that there are some study limitations. The results could
differ for some extent if the unconstrained type of
loading methods were performed; however, the results
of this study closely resemble the previous uncon-
strained loading tests.8,14
Feiertag et al. reported that rib head resection after
total discectomy indicated significant increases in tho-
racic spinal motion in the sagittal and coronal planes;
however, axial rotation tests were not performed.5
Therefore, the effects of rib head resection on thoracic
spine stability in the transverse plane were uncertain.
Using a dog thorax model, Oda et al.8 and Takeuchi et
al.14 demonstrated biomechanical significance of the
rib head joint in the thoracic spine stability. The data
from these dog studies closely resemble the results of
the current study. A recent biomechanical study per-
formed by Thometz et al. demonstrated that distrac-
tion with rib resection preserving costovertebral joints
on the convexity (thoracoplasty) did not significantly
influence rotation of the apical vertebral body in the
coronal and transverse planes,16 suggesting that cos-
tovertebral junction plays an important role in stabi-
lizing the spine. In this in vitro investigation removal
of unilateral rib head joint presented significant in-
creases in ROM under flexion– extension, lateral
bending, and axial rotation loading. Hence, the rib
head joints serve as stabilizing structures to the human
thoracic spine in the sagittal, coronal, and transverse
planes. However, Broc et al. reported that transtho-
racic microdiscectomy, including unilateral resection
of the pedicle and rib head, did not overtly destabilize
the thoracic spine.3 Their microdiscectomy technique
preserved anterior–posterior longitudinal ligaments
and 75% of the intervertebral disc. Thus, with an in-
tact intervertebral disc, unilateral rib head removal
would affect thoracic spinal stability to a lesser extent
versus the rib head resection after total discectomy.
In this study the discectomy with anterior–posterior
longitudinal ligament transection demonstrated a sig-
nificantly greater increase in ROM versus total resec-
tion of the posterior ligamentous complex in flexion–
extension and bilateral axial rotation testing modes
(P ⬍ 0.05). The intervertebral disc can be regarded as
the most important stabilizer in the thoracic functional
unit mechanics. For spinal deformity correction, anterior
disc release is biomechanically more effective in pro-
ducing segmental rotation versus posterior destabiliza-
tion alone. Moreover, Kaneda et al. reported that rib
head resection after partial discectomy significantly
increases the correction rate of the axial rotational
deformities.6 They stated that the correction rate of
the rotational deformity in the transverse plane im-
proves from 15% to 58% after resection of the rib
head joint. In this study unilateral rib head resection
after discectomy resulted in a significant increase in
ROM in all testing modes (P ⬍ 0.05). Thus, rib head
 Biomechanical Properties of the Thoracic Spine • Oda et al E69
resection should be combined with discectomy to
achieve greater curve and rib hump correction.
In this study bilateral total facetectomy after laminec-
tomy combined with bilateral medial facetectomy re-
sulted in a significant increase in ROM in all testing
modes (P ⬍ 0.05). In the clinical setting the lateral por-
tion of the facet joints should be preserved to minimize
postoperative kyphotic deformity and segmental insta-
bility when performing decompressive wide laminec-
tomy. Moreover, removal of the total posterior ligamen-
tous complex increased the ROM by 38%, 37%, and
45% under flexion– extension, lateral bending, and axial
rotation, respectively. If a resection of total posterior
elements is performed, posterior reconstruction may be
required to avoid postoperative deformity and instabil-
ity. However, these effects of posterior element removal
should be less in the in vivo situation because of addi-
tional stability provided by the rib cage and neuromus-
cular system. Using a human torso model, Feiertag et
al. demonstrated that unilateral total facetectomy did
not significantly increase the ROM in the thoracic
spine.5 Therefore, posterior reconstruction may not be
required after unilateral total facetectomy in the tho-
racic spine.
Yoganandan et al. demonstrated that a two-level lam-
inectomy decreases the strength and stability of the tho-
racic spine under axial compression loading18; however,
other destabilizing procedures were not performed. Ta-
neichi et al. performed a statistical analysis of cases with
metastatic thoracic spinal tumors and reported that de-
struction of the costovertebral joint more strongly in-
duced vertebral collapse than the size of metastasis in the
vertebral body,15 suggesting that the thoracic spine com-
pressive failure load decreases after costovertebral joint
injuries. In this study neither nondestructive nor destruc-
tive axial compression tests were performed. Currently,
the structural effects of the costovertebral articulations
against axial compression loading remain unknown;
hence, further biomechanical studies are required for
certainty.
The three-column theory of thoracolumbar spinal sta-
bility is espoused by Denis.4 However, this theory should
not be applied to the thoracic spine between T1–T2 and
T9 –T10 because of the costovertebral joints and rib cage
as stabilizing structures. As Berg described the sternal–
rib complex as a possible fourth column in thoracic spine
fractures,2 our results indicated that the costovertebral
joints serve as significant stabilizing structures to the hu-
man thoracic spine.
Conclusions
The basic biomechanical properties of the human tho-
racic functional spinal units were investigated using
human cadaveric specimens. The intervertebral disc
can be regarded as the most important stabilizer in the
thoracic functional unit mechanics. The rib head joints
serve as stabilizing structures to the human thoracic
spine in the sagittal, coronal, and transverse planes. In
anterior scoliosis surgery additional rib head resection
after discectomy may achieve greater curve and rib
hump correction. The lateral portion of the facet joints
plays an important role in providing stability to the
thoracic spine and should be preserved to minimize
postoperative kyphotic deformity and segmental in-
stability when performing decompressive wide
laminectomy.
Key Points
● The intervertebral disc can be regarded as the
most important stabilizer in the thoracic functional
unit mechanics.
● The rib head joints serve as stabilizing structures
to the human thoracic spine under flexion–
extension, lateral bending, and axial rotation load-
ing, and its resection after discectomy increases
range of motion by approximately 80% under all
loading modes.
● In the thoracic spine total resection of the poste-
rior ligamentous complex leads to an approxi-
mately 40% increase in range of motion under flex-
ion– extension, lateral bending, and axial rotation
loading.
References
1. Andriacchi T, Schultz A, Belytschko T, et al. A model for studies of mechan-
ical interactions between the human spine and rib cage. J Biomech 1974;7:
497–507.
2. Berg EE. The sternal-rib complex: a possible fourth column in thoracic spine
fractures. Spine 1993;18:1916 –9.
3. Broc GG, Crawford NR, Sonntag VKH, et al. Biomechanical effects of trans-
thoracic microdiscectomy. Spine 1997;22:605–12.
4. Denis F. The three column spine and its significance in the classification of
acute thoracolumbar spinal injuries. Spine 1983;8:817–31.
5. Feiertag MA, Horton WC, Norman JT, et al. The effect of different surgical
release on thoracic spinal motion. Spine 1995;20:1604 –11.
6. Jiang H, Raso JV, Moreau MJ, et al. Quantitative morphology of the lateral
ligaments of the spine: assessment of their importance in maintaining lateral
stability. Spine 1994;19:2676 – 82.
7. Kaneda K, Shono Y, Satoh S, et al. Anterior correction of thoracic scoliosis
with Kaneda anterior spinal system: a preliminary report. Spine 1997;22:
1358 – 68.
8. Oda I, Abumi K, Lu DS, et al. Biomechanical role of the posterior elements,
costovertebral joints, and rib cage in the stability of the thoracic spine. Spine
1996;21:1423–9.
9. Oda I, Cunningham BW, Abumi K, et al. The stability of reconstruction
methods after thoracolumbar total spondylectomy: an in vitro investigation.
Spine 1999;24:1634 – 8.
10. Oda I, Cunningham BW, Haggerty CJ, et al. Biomechanical properties of
anterior thoracolumbar multisegmental fixation: an analysis of construct
stiffness and screw/rod strain. Spine 2000;25:2303–11.
11. Panjabi MM, Brand RA, White AA. Mechanical properties of the human
thoracic spine. J Bone Joint Surg Am 1976;5:642–51.
12. Panjabi MM, Hausfeld JN, White AA. A biomechanical study of the liga-
mentous stability of the thoracic spine in man. Acta Orthop Scand 1981;52:
315–26.
13. Panjabi MM. Biomechanical evaluation of spinal fixation devices: I. A con-
ceptual framework. Spine 1988;13:1129 –33.
14. Takeuchi T, Abumi K, Shono Y, et al. Biomechanical role of the interverte-
bral disc and costovertebral joint in stability of the thoracic spine: a canine
model. Spine 1999;24:1414 –20.
E70 Spine • Volume 27 • Number 3 • 2002

15. Taneichi H, Kaneda K, Takeda N, et al. Risk factors and probability of

vertebral body collapse in metastasis of the thoracic and lumbar spine. Spine Address correspondence to
1997;22:239 – 45.
16. Thometz JG, Liu XC, Lyon R. Three-dimensional rotations of the thoracic Itaru Oda, MD
spine after distraction with and without rib resection: a kinematic evaluation Department of Orthopaedic Surgery
of the apical vertebra in rabbits with induced scoliosis. J Spinal Disord Hokkaido University Graduate School of Medicine
2000;13:108 –12.
North-15, West-7, Kita-ku
17. White AA III, Panjabi MM. Clinical Biomechanics of the Spine, 2nd ed.
Sapporo, Hokkaido 060-8638, Japan
Philadelphia: Lippincott, 1990:102–3.
18. Yoganandan N, Maiman DJ, Pintar FA, et al. Biomechanical effects of lam-
E-mail: odaitr@aol.com
inectomy on thoracic spine stability. Neurosurgery 1993;32:604 –10.