Despite extensive research, relatively little is known about the etiology
of adolescent idiopathic scoliosis. Multiple theories have been proposed, and
its etiology is probably multifactorial in
nature1,2.
Recent studies have suggested that genetic factors play a role in some, but
not all, forms of
scoliosis3-5.
A limited number of studies have evaluated gross and microarchitectural
changes in the bone of patients with
scoliosis6-8.
Enneking and Harrington performed histological studies with the light
microscope on inferior articular processes in patients with
scoliosis8. They
evaluated chondrogenesis, osteogenesis, subchondral bone formation, and
cartilage degeneration.
Stilwell7 studied
the mechanisms of bone-remodeling under conditions of chronic asymmetrical
loading using an animal model of scoliosis. That work demonstrated apposition
of cortical bone at the concavity and the resorption of cortical bone at the
convexity in the vertebral bodies. The results of those studies are consistent
with Wolff's law, which states that bone remodels in response to the
mechanical stresses it experiences so as to produce an anatomical structure
best able to resist the applied
stress9.
Although those studies advanced our understanding of the structural changes
within the scoliotic spine, they did not address changes in bone porosity or
bone density with use of modern electron microscopy techniques. The studies
were also conducted when less was known about microadaptations of bone to
eccentric compression and tension environments. Numerous animal studies have
demonstrated differences in cortical bone structure and material organization
within the same bone, in regions that are subject to differences in strain
modes, such as eccentric compression and tension
forces7,10-13.
Those studies demonstrated that cortical thickness increases and porosity
decreases in compression while cortical thickness decreases and porosity
increases in
tension10-13.
The insight for the current study comes from surgical observations in
patients with scoliosis, in whom the facets are noted to have substantial
differences in gross morphology, suggesting bone-remodeling in response to the
eccentric tension and compression forces applied to the convex and concave
sides of the spine, respectively.
The purpose of this study was to analyze, with use of electron microscopy,
the structure of spinal facets in patients with adolescent idiopathic
scoliosis. The facets were evaluated for differences in bone
microarchitecture, including porosity and cortical thickness.
Institutional review board approval for the study was obtained from both
institutions that participated in the study (St. Luke's Children's Hospital,
Boise, Idaho, and the Shriners Hospital for Children, Salt Lake City, Utah).
Informed consent was obtained from all patients.
An Excel database (Microsoft, Redmond, Washington) was used to track
standard demographic information for the patients, including age, sex,
underlying diagnoses, type of spinal deformity, magnitude of scoliosis and
kyphosis, apex of the curve, curve flexibility, number and location of the
facet specimens, and walking status of the patients. Eight patients with
adolescent idiopathic scoliosis, who had a mean age of 16.5 years (range,
eleven to thirty years), underwent standard posterior arthrodesis. The mean
scoliotic deformity for this patient group, determined with use of the Cobb
measurement, was 53° (range, 40° to 68°)
(Table I).
Facet removal was performed in the routine manner, as the standard part of
the surgical treatment of severe scoliosis. The facets were harvested with use
of a 6.2-mm osteotome, providing a right angle in the superomedial corner of
the facet (Fig. 1). The soft
tissue was left intact on the posterior and lateral surfaces, and it was then
removed in the laboratory during processing. Facet pairs at matched anatomic
levels were obtained from three locations: (1) the apex of the curve, (2) one
level cephalad to the apex, and (3) one level caudad to the apex.
During surgery, the facets were placed in individual specimen containers
with 70% alcohol, with labels describing the side (left or right), convexity
or concavity, thoracic level, and name of the patient. The specimens were
delivered to the laboratory and were then assigned a code number. Any
information about the identity of the patient was removed.
Specimen Processing
Each facet pair was identified and placed in a small gauze bag to be
dehydrated in ascending grades of ethyl alcohol and cleared in xylene. The
articular cartilage on the anterior surface and the cut edges of the medial
and cranial surfaces were used to orient the facets and discriminate the right
facet from the left facet of each pair.
Each pair was placed in a specimen cup approximately 5 mm apart, with the
articular cartilage facing the base of the cup and the cranially cut surfaces
facing inward toward each other, and the specimens were embedded in
polymethymlethacrylate according to published
protocols14-17.
The specimens were separated with a band saw into approximately 10 × 10
× 10-mm cubes (each containing a separate facet from the pair). The
cutting was carefully conducted to ensure that only the plastic was cut and
the facets were not damaged. The facets were again oriented according to the
previously mentioned landmark surfaces, and the cubes were marked accordingly.
The specimens were ground manually with increasingly fine grades of sand paper
on a rotary grinding wheel (Buehler, Lake Bluff, Illinois) until the cranial
surface was exposed. This surface was polished to a mirror finish with
0.5-µm levigated alumina (LECO, St. Joseph, Michigan).
Each section was sputter-coated with gold (Hummer VI-A; Anatech,
Alexandria, Virginia) for 1.5 minutes at a current of 10 mA for viewing in a
scanning electron microscope (model 6100; JEOL USA, Peabody, Massachusetts).
Operating conditions of the scanning electron microscope included a 20-kV
accelerating voltage, 0.8-mA probe current, 50-µm aperture size, and 15-mm
working distance. Backscattered electrons were collected with a solid-state
annular backscatter electron detector (Tetra; Oxford Instruments,
Buckingshire, United Kingdom). Images were captured as 512 × 512-pixel
digital images with a gray-level depth of eight bits (256 distinct gray
levels) with use of a computer-controlled image capture and retrieval system
(ISIS 300 series; Oxford Instruments). Each specimen was aligned so that the
cortical surface was parallel to the top of the image and the articular
surface was parallel to the bottom of the image. Adjacent, nonoverlapping
images at 150 times magnification, which spanned the width of the cortex from
the medial surface to the lateral surface and captured >80% of the cortex
of the facet, were then obtained. One to two images at a magnification of
twenty times, which captured a picture of the entire cranial surface of the
facet, were made of each specimen.
The backscatter electron images captured in this study were analyzed to
determine both cortical porosity (void spaces) and cortical thickness. Image
analysis was performed with use of a Scion Image program (Image 1.61; National
Institutes of Health, Bethesda, Maryland). The images at 150 times
magnification were used for the cortical porosity analysis. Included in the
porosity count were all void spaces that were approximately twice the height,
or more, of the cortical bone above it. Large cracks, surface contaminants,
and other preparation artifacts were manually excluded from the analysis with
use of digital image-processing methods. Void spaces were considered to be any
contiguous region, thirty pixels or more in size, with gray levels between
zero and twenty; therefore, lacunae, which have a mean size of twelve to
twenty-five pixels, were not included in the porosity count.
The images at twenty times magnification were used for the cortical
thickness analysis. Each image was centered under a grid, which divided the
full facet into twenty-five even sections that were numbered sequentially from
one to twenty-five. Five numbers between one and twenty-five were selected
with use of a random number generator. These numbers were used to designate
exact, unbiased points from where the cortical thickness was measured.
Measurements (in micrometers) were made by extending a line from the base of
the cortex to the spot at the top of the cortex, which allowed the line to be
perpendicular to the dorsal surface of the facet.
All analyses for cortical porosity and cortical thickness were conducted at
three levels of depth on the facet. After each full round of analysis was
complete, the cranial surface of the facet was manually ground to reveal a
deeper layer of the facet, polished to a mirror finish, and sputter-coated
with gold. Images at a magnification of 150 and twenty times were made of the
cortex for each of the three levels according to the previously described
methods.
When the cortical thickness and percent porosity had been measured for all
three levels for each facet, these numbers were averaged together to create a
mean cortical thickness and mean percent porosity for each facet. The means of
these two groups (concave and convex) were compared with use of the paired
sample t test at an a = 0.05 significance level.
Atotal of twenty-four facet pairs from matched anatomic levels, which had
been obtained from eight patients, were analyzed. The mean porosity for the
concave facets was 16.5% ± 5.8%
(Fig. 2). For the convex side,
the mean porosity was 24.1% ± 6.2%
(Fig. 3). The concave side was
significantly less porous than the convex side, with a mean difference of 8%
(p = 0.03).
The mean cortical thickness for the concave facets was 798 ± 266
µm. For the convex facets, the average cortical thickness was 377 ±
124 µm. The facets on the concave side had a significantly thicker cortex
than did the facets on the convex side; the mean difference was 421 µm (p
< 0.01) (Fig. 4).
Alimited number of experimental studies have been performed on bone from
patients with adolescent idiopathic scoliosis. Anatomic studies in scoliosis
have focused upon pedicle
morphology18-20,
vertebral morphology, and vertebral
rotation21,22.
Limited information is available on the histological architecture of bone in
these
patients6-8.
Histological studies by
Stilwell7 were
performed on specimens obtained from animal models of scoliosis. Those studies
showed decreased chondrogenesis and osteogenesis, disorganized columnation,
and premature growth cessation in the cartilaginous end plates of the
vertebral body on the concave side of the curve. McCarroll and
Costen6 performed a
biopsy of the convex anterior aspect of the vertebral bodies, which showed
abnormalities of cartilaginous growth. The findings in both of those studies
are consistent with the principles described by Hueter-Volkmann and Delpech,
which suggest that cartilaginous growth is affected by the mechanical pressure
on the
tissue23,24.
Enneking and Harrington performed histological studies with use of light
microscopy on inferior articular processes in patients with
scoliosis8. They
evaluated chondrogenesis, osteogenesis, subchondral bone formation, and
cartilage degeneration. On the basis of their findings, Enneking and
Harrington concluded that scoliosis was probably not a disorder of
endochondral growth or a result of an underlying pathological bone condition,
and they suggested that further research should focus upon extraosseous
tissues.
The work of Wolff
9 was based upon the
observation that bone alters its configuration in response to
stress25-27.
Numerous authors have paraphrased the work of
Wolff28, and most
have suggested that "bone has the ability to remodel, by altering its
size, shape, and structure, to meet the mechanical demands placed on
it."26 The
findings of
Stilwell7, McCarroll
and Costen6,
Enneking and
Harrington8, as well
as those in our study are all consistent with Wolff's law, i.e., "that
bone formation will increase in areas of compression and decrease in areas of
tension."29
The insightful work of
Stilwell7, McCarroll
and Costen6, and
Enneking and
Harrington8 provides
a foundation for further study of bone-remodeling in scoliosis. Those studies
were conducted at a time when less was known about bone-remodeling under
eccentric tension-compression environments, and when the electron microscope
was not readily available for such studies. Recent animal studies of cortical
bone under compressive stress have demonstrated that cortical bone exhibits
decreased cortical porosity and increased cortical thickness compared with
bone under tensional
stress10-13,30-32.
The results of our study suggest that scoliotic deformities apply eccentric
tension and compression stresses to the convex and concave portions of the
curve, respectively, producing bone microarchitectural changes in the spinal
facets similar to those seen in animal studies of eccentrically loaded
diaphyseal
bone10-13,30-32.
These changes are consistent with Wolff's
law9.
A recent study by Pazzaglia et
al.33 evaluated the
effects of eccentric tension and compression forces on the growth and
development of bone in a rat model. In that model, the bent tail of rats was
used to investigate the effects of mechanical forces on bones and joints.
Although the study was not on the thoracic spine, the eccentric forces were
applied to a multisegmented structure with intrinsic flexibility through disc
spaces. Their study showed increased density and trabecular width on the
compression side. Their results were similar to ours, and they support
bone-remodeling in accordance with Wolff's
law9.
Human studies of bone-remodeling are limited because of the lack of
availability of tissue, especially for younger subjects. Future studies of
human facets in scoliosis offer the opportunity to further define the
microarchitectural response of human bone and to validate animal models of
bone response to eccentric loading conditions. Our study evaluated specimens
near the apex of the curve, which is typically in the region of maximum
deformity. Presumably, these regions are subject to the greatest differences
in tension and compression forces, as applied to the two facets at the same
anatomic level.
Our results suggest that these microarchitectural bone changes are in
response to eccentric tension-compression forces because the patterns of
remodeling are similar to those seen in other studies of eccentrically loaded
bone10-13,30-32.
Although the results of our study demonstrated bone-remodeling in an eccentric
loading environment, it remains unclear whether the remodeling changes were
solely in response to eccentric loads, or whether an underlying pathological
bone condition also contributed to these changes.
Although numerous studies have suggested that changes in osseous tissue are
a secondary response to scoliosis, recent studies have suggested that an
underlying pathological process in bone, i.e., osteoporosis, may be related to
the structural changes in
scoliosis34-41.
Cheng et al. demonstrated decreased bone-mineral density and lower osteocyte
counts in iliac-crest biopsy specimens from patients with idiopathic
scoliosis34. Cheng
et al. and Thomas et al. also demonstrated that adolescent idiopathic
scoliosis may be associated with generalized low bone-mineral
status34,35,39.
Studies by Cheng et al. and by others have also suggested that the osteopenia
in idiopathic scoliosis may be related to the primary etiology of the disease
and is not a result of the asymmetrical mechanical forces associated with the
scoliosis35-38.
Velis et al. suggested that patients with lumbar scoliosis and osteoporosis
are more likely to have lateral spondylolisthesis and segmental instability
develop than are patients with lumbar scoliosis and normal bone
density41.
Additional work on the underlying structure of bone in patients with
scoliosis is essential for a more complete understanding of osseous tissue in
such patients and to determine whether osteoporosis or another underlying bone
disorder contributes to the development or the progression of deformity.
Future studies should evaluate facets in regions outside the areas of
scoliotic deformity, to assess for differences in bone microarchitecture.
Recent studies have attempted to produce animal models of scoliosis,
although numerous researchers have attempted to produce these deformities for
at least seventy years. The work of Braun et
al.42 showed great
promise in developing a model of scoliosis that will be valuable in the
assessment of interventions that attempt to modulate spinal growth to correct
scoliotic deformities. Other animal models may also be developed. Evaluation
of spinal facets may be useful in validating these animal models of
scoliosis.
In conclusion, facet microarchitecture in patients with scoliosis
demonstrates remodeling similar to that in animal models of bone subject to
eccentric tension-compression forces. These changes suggest that bone in
patients with scoliosis remodels appropriately as a secondary response to
spinal deformity and is consistent with Wolff's
law9. It remains
unclear whether an underlying bone disorder also contributes to the
microarchitectural changes, and recent studies have suggested that
osteoporosis may contribute to the development and/or progression of spinal
deformity35-41.
Additional study of facet architecture should provide an improved
understanding of the strain modes and pathophysiology of scoliosis, and it may
also validate animal models of scoliosis.