It is a well-known fact that the human spine grows during the infantile,
juvenile, and adolescent periods of life. Less known and less obvious are the
specifics of vertebral column growth, including the total amount of growth per
year and the amount of growth that occurs at each vertebral segment and in
each of the various planes (axial, coronal, sagittal, and transverse). We do
know that the growth rate of the spine varies according to age. In utero, the
spine grows at an extremely accelerated rate; from birth to two years of age,
it develops at an increased rate; from two to ten years of age, it proceeds to
grow at a steady rate; and finally, during the prepubescent growth phase, it
grows at an increased rate once again.
There is a relationship between growth of the spine and the development of
spinal deformity. Progression of scoliotic spinal deformity occurs during
periods of peak growth
velocity1,2.
The first spinal growth peak occurs after the postnatal period until two years
of age, and the second peak occurs during the prepubescent period as a result
of a strong hormonal
influence2.
Estimates of average total vertebral column growth and vertebral column
growth by region have been proposed by Dimeglio and
Ferran3. Anderson et
al.4 described the
remaining vertebral column growth based on sitting heights. These methods
delineate differences in growth between boys and girls and among different
skeletal ages. According to Dimeglio and
Ferran3, the
longitudinal growth of the thoracic and lumbar vertebrae is 0.8 and 1.1 mm per
year, respectively. According to
Roaf5 and
Taylor6, the
intervertebral thoracic discs grow between 0.2 and 0.6 mm per year and the
intervertebral lumbar discs grow between 0.3 and 0.8 mm per year. According to
our own clinical measurements of patients with adolescent idiopathic
scoliosis, which are less accurate than measurements obtained with use of
scaled radiographs or three-dimensional reconstruction
techniques7, the
growth of the disc and end plate is <1 mm per year.
Normal vertebral morphologic data for pediatric specimens from the
Hamann-Todd Osteological Collection (Cleveland Museum of Natural History,
Cleveland, Ohio) have been described by Zindrick et
al.8. Extensive
morphometric analyses of anatomic scoliotic specimens from two major
osteologic sources, the Hamann-Todd Osteological Collection and the Robert J.
Terry Collection (Smithsonian National Museum of Natural History, Washington,
DC), were made by Parent et
al.9,10
in order to identify typical deformity patterns for thoracic and lumbar
vertebrae in idiopathic scoliosis. Thirty scoliotic specimens (472 vertebrae)
were studied with use of a three-dimensional digitizer and were matched to 512
normal vertebrae. A characteristic three-dimensional deformity pattern was
identified, consisting of progressive vertebral wedging and torsion, decreased
pedicle width on the concave side of the curve, and great variation at the
articular facet surfaces. All findings were increasingly more important toward
the apex of the curve and with increasing severity of the curve.
Winter11
estimated the loss of vertebral growth due to spinal arthrodesis, thereby
creating an indirect assessment of spinal column growth. Winter's data, which
generally consisted of experience and clinical observations, were based on the
Dimeglio data11.
The Dimeglio "formula" states that, for each vertebral segment,
0.7 mm per year of longitudinal growth is lost after posterior
arthrodesis11. The
formula assumes that spinal column growth ceases at fourteen years of age for
girls and at sixteen years of age for boys. This formula will be referred to
frequently in this paper. When the formula is rounded to 1 mm per year, each
vertebral ring apophysis can be assumed to contribute approximately 0.5 mm of
axial column length per year, on the average. Thus, each vertebra (two
vertebral ring apophyses per vertebra) contributes 1 mm per year to vertebral
column height. The average used for this exercise does not take into account
the accelerated spinal growth seen in the first two years after birth and in
the prepubescent period, when vertebral growth is likely to be greater.
Currently available information regarding vertebral column growth is
remarkably limited and poorly defined. Further research is needed to obtain
accurate data with regard to normal growth of the vertebrae and the spinal
column. These data must include accurate coronal, sagittal, axial, and
transverse growth information for normal and abnormal states at each year of
age. Remarkably few specific, accurate, and detailed data regarding spinal
growth are available.
Growth-Related Pathogenesis Hypotheses Applicable to Infantile,
Juvenile, and Adolescent Idiopathic Scoliosis
It is well known that idiopathic scoliosis is a complex three-dimensional
deformation of the vertebral column that causes subsequent deviations in the
frontal plane, the sagittal profile, and the transverse plane (as a result of
torsion). After the onset of the initial deformity, progressive scoliosis
evolves with growth of the spine. The progressive scoliosis enters a
biomechanical cycle that involves subsequent asymmetrical loading of the
vertebral ring apophyses, altered vertebral growth and shape, and the
development of the global picture of the scoliotic
deformity12-19.
The growing vertebrae and their ring apophyses respond to loading. Axial
loading of the ring apophyses affects the longitudinal growth in accordance
with the Hueter-Volkmann
principle20. This
principle states that increased pressure retards growth and, conversely,
reduced pressure accelerates growth. Thus, in idiopathic scoliosis, the
initially parallel vertebrae and discs in the frontal plane subsequently
become wedged in all three dimensions, causing the vertebrae and spinal column
to rotate (vertebral torsion).
Stokes et al.16
have done substantial work in this area. In particular, they made use of two
animal models (rat tail and bovine tail) to investigate the hypothesis that
the vertebral wedging that occurs with growth in patients with progressive
spinal deformity results from asymmetric loading in a "vicious
cycle" (Fig. 1). They
also thoroughly characterized the growth-modulation process, and they
developed computer models to test various aspects of the "vicious
cycle" (the loading of the spine with scoliotic curves to see if the
"vicious cycle" of stress-modulated growth explains scoliosis
progression, for example). They found that plausible magnitudes of skeletal
loading and growth sensitivity to load can predict that a substantial
component of scoliosis progression during growth is biomechanically
mediated.
Inspired by the work of Stokes et
al.13,14,19,
Villemure et al. demonstrated the feasibility of biomechanical modeling of
vertebral growth and growth
modulation12
(Fig. 2-A). The model includes
the osseoligamentous structures of the spine as well as the vertebral growth
and the growth modulation process that results from loading of the growth
plates. On the basis of these biomechanical simulations and the observations
of idiopathic scoliosis by Villemure et
al.21, five
different hypotheses regarding pathogenesis have been
tested22 with use
of initial geometrical eccentricity (gravity-line imbalance of 3 mm or 2°
of rotation) at the thoracic apex to trigger the self-sustaining deformation
process over a period of twenty-four months
(Fig. 2-B). Their results
support the conclusion that there is not a unique simulated pathogenesis in
adolescent idiopathic scoliosis that results in the development of scoliotic
deformities. The eccentric loading in the frontal plane, whether combined or
not with sagittal imbalance, generated the closest representation of scoliotic
deformities. Their observations support the hypothesis that the essential
scoliotic lesion is one of precarious coronal
balance23 and
reduced thoracic kyphosis associated with coronal plane
asymmetry24,25.
The biomechanical simulations suggested that the thoracic segment is more
sensitive to imbalances in the frontal plane and that the sagittal equilibrium
is much more stable than that in the coronal plane.
Further, it has been proposed that it is not likely that scoliosis has a
typical pattern of evolution; rather, the patterns of evolution are more
likely to be variable and patient
specific21. On
evaluation of idiopathic scoliotic curves of twenty-eight adolescent patients,
no consistent pattern of the plane of maximal deformity or the evolution of
the thoracic kyphosis was observed. Patient-specific evolution of scoliosis is
a new concept and is convincingly demonstrated in the accurate and thorough
biomechanical clinical analysis of these patients. It is reasonable to state
that the Hueter-Volkmann principle still applies. The patient-specific
evolution of deformity is likely due to many other associated variables, such
as age of onset, location of initial deformity, loads on the spine,
maturational status, mechanobiologic factors (such as the sensitivity of
growth cells to loads), and factors associated with the body habitus.
The neurocentral junction also has been identified as a potential cause of
adolescent idiopathic scoliosis. Asymmetrical (left-sided or right-sided)
growth at this site is believed to lead to pedicle-length asymmetry, which
then causes vertebral rotation and, ultimately, the development of scoliotic
curves. A computer model was used to investigate whether pedicle asymmetry,
either alone or in combination with other deformations, could be involved in
the pathomechanisms leading to
scoliosis26.
However, simulations with asymmetrical pedicle geometry did not produce
structural scoliosis, vertebral rotation, or wedging. Also, simulations with
asymmetry of the pedicle growth rate did not cause scoliosis independently and
did not amplify the scoliotic deformity caused by other deformations that were
tested in the previous model. Thus, the results of the computer simulation do
not support the hypothesis that asymmetrical growth at the neurocentral
junction is a cause of adolescent idiopathic scoliosis. This finding concurs
with the findings from recent experiments on animals in which growth at the
neurocentral junction was unilaterally restricted and no scoliosis, vertebral
wedging, or rotation was
noted26,27.
Correction by Modifying Growth—What Are the Possibilities?
The idea of growth-correction surgery is an intriguing one. Even though our
current knowledge is likely sufficient for the correction of scoliosis by
growth modulation in the human clinical scenario, we have not yet defined the
surgical indications, the upper and lower age limits, and the upper and lower
curve magnitudes in the laboratory or in the clinic. The concept of correcting
spinal deformity by means of harnessing growth has great appeal for a number
of reasons. First, the potential exists to correct the scoliosis naturally.
Second, the potential exists to correct the scoliosis without fusion. Third,
because the patients who are most likely to be eligible for a surgical
procedure will be children with moderate, progressive scoliosis, the potential
exists to eliminate the need for spinal bracing during the teen years.
A possible surgical treatment for adolescent idiopathic scoliosis in spines
that still have the potential to grow would consist of modifying the force
equilibrium transmitted to the spine to reverse the deformation process. Other
ways to trigger vertebral growth modulation have been tested in the past. The
results of animal studies have shown that rib resection, shortening, or
lengthening can
induce28-31
as well as
correct32-34
scoliosis. Xiong and
Sevastik35 reported
on the case of a young patient with scoliosis who was treated by shortening of
three ribs on the curve concavity and who subsequently showed a continued
reduction of spinal curvature. In a biomechanical study in which they used a
finite element model of the
trunk36,37,
Grealou et al.36
and Carrier et
al.37,38
demonstrated that, although rib surgeries produce only slight immediate
geometrical changes, concave-side rib-shortening and convex-side
rib-lengthening can induce load patterns on the end plates (in the range of
1.5 to 2.5 Nm) that could act against the progression of scoliosis
(Fig. 2-C). This theory was
recently tested37
with use of a computer model that integrated the whole rib cage, ligament
relaxation, spine growth, and growth modulation. The results, which were
expected on the basis of the load patterns induced on the vertebral end
plates, confirmed the potential of rib-shortening to act against progression
of scoliotic deformity. Decreased vertebral wedging in the frontal plane was
observed for all vertebrae in the apical region, a finding that was reflected
in the correction of the spinal curvature (slight correction of 10% over the
course of twenty-four months). However, because of the limited correction,
this approach should be combined with other means of controlling
deformity.
Computer simulation was further
used38 to find the
optimal parameters for rib surgery for the correction of scoliotic
deformities. The optimization results showed that no "typical"
optimal operation exists and that the parameters maximizing the correction
varied greatly according to the geometric configuration of the patient's trunk
and the correction objective.
Growth modulation of the spine has been demonstrated in the
laboratory13,39-41,
most notably through the use of a technique of surgical stapling of an
intervertebral disc interspace. Betz et
al.42 recently
reported on the clinical use of growth correction surgery with use of
intervertebral stapling for moderate idiopathic and syndromic scoliosis. This
early report demonstrated varied results. In select cases, there appeared to
be a stabilization or a reversing of the scoliotic deformity.
Using an external fixator to impose an angular deformity on rat-tail
vertebrae, Mente et
al.43 tested the
following three hypotheses: (a) a vertebral wedge deformity created by chronic
static asymmetrical loading will be corrected by reversal of the load
asymmetry; (b) a vertebral wedge deformity created by chronic static
asymmetrical loading will remain if the load is simply removed; and (c)
vertebral longitudinal growth rates, altered by chronic static loading, will
return to normal after removal of the load. Their results showed that a
vertebral wedge deformity can be corrected by reversing the load used to
create it and that vertebral growth is not permanently affected by applied
loading.
In past experience, the loosening of staples that have been applied across
the intervertebral disc has been problematic. Other prototypes that include
better bone fixation may address this issue. Betz et
al.42, who made use
of a Nitinol staple with a shape-memory-alloy composition that clamps down on
the vertebral body after insertion, reported no adverse effects and no staple
dislodgment or migration during the short follow-up period (mean, eleven
months; range, three to thirty-six months) of their study.
Principle of Growth Modulation and Asymmetric Growth in Reversing
Spinal Deformity
As was stated earlier, the growing vertebrae and their ring apophyses will
respond to applied loads. The mechanism for doing so is loading of the ring
apophyses in an asymmetric manner to modulate the longitudinal growth
according to the Hueter-Volkmann principle. Increased pressure retards growth
and, conversely, reduced pressure accelerates growth. Thus, in a patient with
idiopathic scoliosis who has an appropriate amount of growth remaining, the
condition of wedged vertebrae and column rotation (vertebral torsion) is
predicted to reverse
itself43.
Angular Change at the Single Vertebra or Full-Spine Level
Assuming that the physeal width of the ring apophysis is 55 mm and the
axial growth per vertebra per year is 1 mm (rounded from 0.7 mm per segment
per year, as mentioned above), it would be possible to obtain a 1° to
2° correction per year per
level13. If the
required forces were applied over a spinal region of five segments, a 5°
to 10° change per year could be expected. Thus, if there were three years
of spinal column growth remaining, growth modulation and fusionless methods
could induce a possible full correction in a 30° five-segment apex curve
prior to the cessation of growth and the onset of skeletal maturity.
Stokes and
Gardner-Morse17
simulated curve progression in an analytical model that included measured
velocity of spinal growth in adolescence, analytical estimates of load
asymmetry, data on growth sensitivity or mechanical loading from animal
studies, and comparison of predictions with known rates of scoliosis
progression. They found that if the initial Cobb angle was 26° at the age
of eleven years, then the average final lumbar spinal curve magnitude would be
34° at the age of sixteen years when spinal loading was equivalent to 50%
of muscle maximum effort and 42° when spinal loading was equivalent to 75%
of muscle maximum effort.
The biomechanics of physeal compression by applied loads of a long bone
versus a ring apophysis may be different and requires further study. A
bone-physis-bone staple compression model may not be equivalent to a
bone-physis-disc-physis-bone compression model. In other words, the stiffness
of compression of the spinal disc model may be less than that of the long bone
model and potentially less predictable. Additional research will be needed,
including biomechanical testing after the application of staples across a
spinal disc.
If the spine is mechanically tethered at a corrected position, will
subsequent asymmetric spinal growth maintain it? This is an important question
that needs to be answered. According to the Hueter-Volkmann theory, corrective
asymmetric growth after tethering should, in principle, allow for definitive
correction and maintenance of improved alignment of the spine. Remodeling of
vertebral wedging to a corrected vertebral morphology should allow for
maintenance of correction after the removal of asymmetric loading when the
spine deformity has been corrected.
Before growth modulation approaches can be used in the clinical setting,
further simulation modeling is needed to identify the most effective
techniques for achieving successful growth correction. Our current knowledge
with regard to vertebral growth and spine-segment growth is lacking in detail
and accuracy. Additional morphologic data will need to be accurate and
authoritative and must include all ages, from birth to skeletal maturity. The
data must also include coronal, sagittal, axial, and transverse plane
morphology in both normal and abnormal states.
Simulation modeling cannot be validated until accurate data have been
accrued. Despite the current limitations, modeling can yield valuable insights
into the biomechanics of scoliosis and the development of better treatments.
Modeling studies will be most useful, however, when the results are considered
in combination with the results of other types of studies, such as clinical
studies and in vitro and in vivo laboratory experiments.
Our knowledge of vertebral column growth is remarkably limited. Further
detailed morphologic research is needed in relation to growth, especially as
orthopaedic surgeons approach the era of growth modulation in the correction
of scoliosis. Nonetheless, through application of the knowledge that is
currently available to us, it should be possible to correct a scoliosis by
5° to 10° per year over a five-segment apex, assuming that the spine
continues to have the potential for growth during that time. ?