Chapter 3

The Etiology of Spinal Deformities

3.1Introduction

3.2The Development of Idiopathic Scoliosis

3.3Experimental Models

3.4What Pathogenesis Tells Us about Treatment

3 The Etiology of Spinal Deformities

3.1 Introduction

It is probably best to start with the etiology of idiopathic scoliosis and then to go on to discuss the etiology of the other spinal deformities in the classification. For whatever reason right–left asymmetry has attracted an illogical proportion of research effort. Nonstructural scoliosis, for example, secondary to a leg length inequality (see ▶ Fig. 2.4), is indeed a problem of right–left asymmetry and so for that matter is a single congenital hemivertebra (see ▶ Fig. 2.6) that exists only in the coronal plane. However, structural scoliosis (scoliosis with torsion) is a complex deformity involving all three planes.

The deformity of idiopathic scoliosis is three-dimensional resulting from viscoelastic buckling of the spine in both the coronal (producing a lateral bend) and the transverse planes (an axial rotation or torsional buckling) to an approximation of Euler’s laws. ¹ Euler’s laws can be summarized using a rather simplified formula. Euler’s laws dictate how engineers’ beams or columns fail mechanically and this is in two principal ways: simple angular collapse when a stick bends and breaks, for example, and that would be angular kyphosis; or beam buckling whereby the column fails by twisting deformation (buckling) (▶ Fig. 3.1). From a biomechanical standpoint, the critical buckling point is expressed by the formula:

where P is the critical buckling load, N = a constant for end conditions, E = Young’s

modulus, I = moment of inertia, and L = effective length of the column. (For nonlinear cases exhibiting anisotropism, appropriate corrections can be made using the tangent modulus and secant modulus theories). It is recognized that Euler’s theory provides only a simplistic explanation for spinal instability, since the plastic deformations that occur in a curved column like the spine are more complex. We will develop Euler’s laws further as we go through this chapter. While this buckling process is essentially physical, for a progressive deformity to ensue, it is a requirement that the buckling process occurs during the phase of spinal growth, thus the product of spinal buckling can be seen as progressive three-dimensional shape deformation. Observation of this deformation process facilitates an understanding of pathogenesis as well as treatment. Therefore, both biomechanical and biological factors are the critically important issues to consider. ²

It would be sensible to begin by describing normal spinal shape and then we can see how this shape buckles to produce structural scoliosis. The common characteristics of the normal human spine can be summarized as follows. It is a moderately slender tapered column and in cross-section it can be regarded as a prism of variable orientation. ³ The broad side-to-side width of the cervical and lumbar vertebrae dictates that the bases of the prisms are situated anteriorly while the apices are posterior. By contrast the thoracic vertebral bodies are heart-shaped with a greater length from front to back and are thus prismatic with the apices anterior and the bases posterior. Furthermore, the thoracic prisms are asymmetric with the apices slightly to the right of the midline as a result of the lateral force (or lateral restraint) imposed by the thoracic aorta (▶ Fig. 3.2 and ▶ Fig. 3.3). The spinal column or beam does not have to be prismatic in cross-section to buckle, rather prismatic shape does help the buckling process and its appreciation. ³ In addition in the coronal plane a very small lateral curvature exists in the thoracic region (convex to the right usually, secondary to the “aortic force”), and this is an almost invariable finding described by anatomists centuries ago. ⁴ To confirm this tendency for right-sided thoracic direction it is the opposite (i.e., to the left) with situs inversus ^{5,​ 6} (▶ Fig. 3.3b). In this beautiful study utilizing both computed tomography (CT) and magnetic resonance imaging (MRI), the preexisting rotation of the vertebrae from top to bottom in the normal and situs inversus spine is clearly shown. The right-sided preponderance from T6 to T10 to the right in the normal spine and to the left in the situs inversus spine does achieve statistical significance at the 1% level. This nicely demonstrates that just where idiopathic thoracic scoliosis is apical, namely the mid-lower thoracic region; it is predestined to favor the right side. This effect is static dependent upon vertebral shape in the transverse plane. In the lumbar region there is no static predisposition to the left; rather, it is the presence of the abdominal aorta on the left side that produces its effect dynamically, being a pulsatile mass just next to the spine (▶ Fig. 3.3c).

There are also curvatures in the sagittal plane in the ambulant human spine and there are four primary curves: lordoses in the cervical and lumbar regions and kyphoses in the thoracic spine and the sacrum. Therefore, it can be seen that the “normal” human spine is not symmetrical in any of its three planes and that the nature of this asymmetry varies along the length of the column. Flexion in the thoracic spine is limited by the relatively reduced disk height and with the rib cage imparting marked rigidity. The presence of a kyphosis indicates that the axis of longitudinal rotation in the thoracic spine is located in front of the spine and thus protects the otherwise rotationally unstable prismatic transverse plane shape from rotation on flexion or compressive loading (▶ Fig. 3.4 and ▶ Fig. 3.5). The thoracic spine is thus normally under tension and not compression.

In idiopathic scoliosis, however, the back of the spinal column rotates toward the curve concavity (concordant rotation) as inspection of every radiograph will confirm (▶ Fig. 3.5), thus indicating that the spinal column is shorter at the back than the front (the back of the spine is running the 200 m on the inside lane) and thus confirms the presence of lordosis in every case (▶ Fig. 3.6). ⁷ Meanwhile, with a kyphotic thoracic spine compressed to failure, the spinous processes buckle toward the curve convexity and this can be seen in virtually every case of Scheuermann’s disease of any degree when there is quite often a mild coronal plane curvature with either no rotation or rotation toward the convexity ⁸ (discordant rotation) (see ▶ 5). Cervical and lumbar vertebrae with their axes of rotation lying behind the vertebral bodies because of their lordotic sagittal shape are inherently rotationally unstable on flexion but are protected by three factors: (1) low flexural stiffness as a result of the anteriorly based prismatic shape and increased disk height permitting considerable flexion; (2) powerful posterior musculoligamentous structures; and (3) lumbar facet joints with virtually sagittal plane orientation although the lower two lumbar facet joints permit a small amount of rotation. As a result of the natural three-dimensional spinal asymmetry we can regard any axial load as having eccentricity and the spine as a whole as being prebuckled, theoretically reducing the critical load requirements as compared to a straight or single curved column.

Along with restraints imposed by the intervertebral disks, ligaments, and facet joints, the sagittal curvatures of the spine determine the centroidal axes of rotation of any given motion segment in any plane (anterior in the thoracic spine and posterior in the cervical and lumbar spines). The spine exhibits variable stiffness (flexural, torsional, and shear rigidity) throughout its length and at different magnitudes of loading and direction of loading—that is, it is anisotropic, nonlinear, and inhomogeneous. ²

Motion segments display coupling of rotations and translations such as the thoracic spine axial torque producing a primary axial rotation motion and a secondary frontal rotation (lateral bend) about a helical axis of rotation; the direction to which the posterior elements point is constant, toward the convexity of the curve for the kyphotic thoracic spine. This contrasts with the invariable concave side posterior rotation seen in idiopathic scoliosis. As an aside, the apparent coupling of lateral flexion that accompanies forward flexion in scoliotic spines is no more than true forward flexion of a segment rotated about its longitudinal axis, the local plane of flexion/extension now pointing laterally. It can be seen straightaway that if lateral flexion occurs in the kyphotic thoracic spine it does not produce the natural scoliotic spinous process rotation toward the curve concavity and therefore efforts at trying to produce experimentally right–left asymmetry, where most attention has been directed, will naturally be fruitless—lateral flexion doesn’t occur first.

In terms of the end restraints of the spinal column, the spine can be thought of as permitting translations and rotations only in the sagittal plane, since it is a requirement that the eyes are level and face forward and the pelvis is similarly constrained; the presence of the thoracic aorta may be imagined as perhaps a lateral restraint or a laterally deflecting force.

The structural complexity of the spine in terms of its myriad components is not only complicated by the inconveniences of creep, relaxation, anisotropism, and nonlinearity, but by additional, self-generating deforming forces consequent upon growth. Bone obeys the Hueter(1)-Volkmann(2) law as well as Wolff’s(3) law and the vertebral bodies exhibit dynamic deformation as a result of tensile and compressive forces that result in stimulating and inhibiting bone growth, respectively, and further three-dimensional asymmetry—only, of course, during spinal growth (not just at the age of 15 when the pelvis matures but up to possibly the age of 25 ⁹), and then afterward not until osteoporotic or osteoarthritic old age. The more that growth makes the vertebrae asymmetric, the more the spinal column becomes less easy to straighten and requires alternative treatment strategies (see Chapter 4).

3.2 The Development of Idiopathic Scoliosis

Now let’s get away from biomechanics for a while to more clinical matters. From the foregoing it would not seem to require much more effort to figure out how idiopathic scoliosis develops especially as the information has been in the literature for decades, if not centuries. The concept of the essential lesion in idiopathic scoliosis as being a relatively fixed lordotic area that, under the influence of transverse or coronal plane asymmetry, rotates to the side and gives rise to a lateral curvature has consistently been put forward as the prime etiology. ^{10– 12} Adams’ work in 1865 describes his dissections of bodies with scoliotic spines 40 years before X-rays were invented. He showed quite clearly in his dissections that there was always a lordosis at the apex of the curve in structural scoliosis. ¹⁰ Unfortunately, Adams seems to have been relegated to the “forward bend test” but his work really should be read fully by all interested scoliosis surgeons. Another intellectual giant in scoliosis research was Robert Roaf in Liverpool, an individual of the utmost brilliance while at the same time being incredibly pleasant and having the time to help young surgeons, for example, R.A.D., with an interest in scoliosis surgery. As we shall see he also made a considerable contribution to the treatment of spinal deformities. ¹² However he told us all about the cause of structural scoliosis although it seems that his papers have gone largely unread by the non-UK scoliosis world. His classic “The Basic Anatomy of Scoliosis” ¹¹ clearly explained how the deformity of idiopathic scoliosis develops and all the Leeds Scoliosis Study Group has done subsequently is to add a little more detail. He pointed out that under the restraining influence of the anterior chest and abdominal wall increased length of the anterior elements can be accommodated only by lateral deviation and rotation of the vertebral bodies and that the degree of torsion varies with relative lengthening of the anterior elements. ▶ Fig. 3.7, an interpretation of one of his figures, represents how this is accomplished with reference in particular to the line of the anterior longitudinal ligament and the line of the interspinous ligaments. He stated that “simple three-dimensional geometry shows that if the vertebral bodies are rotated farther from the midline than the spinous processes there must be lordosis – that is relative lengthening of the anterior longitudinal ligament” (note the use of the word “simple”).

He then went on to measure the anterior and posterior spinal lengths from T1 to L5 in both normal and scoliotic spines confirming the relative shortness of the back of the spine with scoliosis (▶ Table 3.1). He then described how a scoliotic spine can be divided into five regions, as shown in ▶ Fig. 3.8. Then for those still struggling with this concept of lordosis as the essential lesion, ▶ Fig. 3.9 shows how lordosis behaves when flexed and the only methods by which it can be corrected. This is of fundamental importance because it indicates precisely how and why, having understood how the deformity develops, you can correct it logically. ¹¹ This is why in Leeds, particularly for larger more stiff curves, we introduced the concept of shortening the front of the spine (anterior multiple diskectomy), the Leeds Procedure. ¹³ Perhaps we should have called it Roaf’s procedure but we wanted to draw attention to all the work on the pathogenesis of idiopathic scoliosis that we have done in Leeds over the years. We felt that to understand how this three-dimensional deformity develops would be an essential prerequisite for understanding how it can be corrected.

In Leeds we borrowed 11 articulated skeletons with idiopathic scoliosis from the Royal College of Surgeons of Edinburgh Museum (courtesy of Professor D. C. Mekie, the curator) and studied them every which way including carrying out the same anterior and posterior vertebral body height measurements as Roaf had done. We found precisely the same results, namely over the apical region, the anterior vertebral body heights were significantly bigger than the posterior (▶ Table 3.2). ¹⁴

Another great was Edgar Somerville from Oxford who probably carried out the first sensible experiments to produce structural scoliosis (in rabbits) ¹⁵ (▶ Fig. 3.10). Indeed, the title of his seminal paper was “Rotational Lordosis: the development of the single curve.” He felt the concept of lordosis as being the essential lesion was so straightforward that he spent the rest of his professional life looking after children with developmental hip dysplasia! It was really he that inspired the Leeds Group to go on and do all the work they did and he was a frequent and most welcome visitor and advisor to the Leeds Unit.

That the prime lesion in structural scoliosis is an apical lordosis is best appreciated in the thoracic spine which is normally kyphotic and therefore the presence of lordosis would indicate a fairly radical departure from normal sagittal shape. The presence of the essential thoracic lordosis in idiopathic thoracic scoliosis can always be verified; but, with increasing degrees of apical rotation, the more increasingly oblique the lateral spinal radiograph is, which spuriously gives the impression of kyphosis when of course it is just another oblique view of the same lordo-scoliosis. ¹⁶ This is precisely why publications and presentations continue to look at and measure lateral radiographs of the patient. It was, for instance, thought that more than 60 degrees of kyphosis on a lateral view, of say a neurofibromatosis patient, was the critical level beyond which anterior surgery was mandated. ¹⁷ All this means is that the scoliosis has got a lot bigger and so has the degree of rotation. It is nothing whatever to do with kyphosis. In recent classifications of idiopathic scoliosis, the concept of kyphosis still appears and is thought to be critical in the classification. ¹⁸ Of course it is quite right that the more the rotation the more the need for anterior surgery. Because the bigger the Cobb angle and the more rotation the stiffer the curve is and requires an anterior space-making procedure (anterior multiple diskectomy) to allow metalwork to adequately correct the deformity. ¹³

It was Stagnara who showed that true views of the scoliosis can be taken when the patient or X-ray beam is rotated to the amount of apical rotation and then true plan d’election views were obtained (see ▶ Fig. 2.10). ⁷ On the anteroposterior (AP) view when the deformity is seen en face, the Cobb angle is much greater; but, more importantly, when a true lateral is taken of the curve apex, that is, 90 degrees to the true plan d’election AP, the lordosis is unmasked (▶ Fig. 3.6). The lordosis is still present on true lateral views of the curve apex even in very big curves. However, of course, the degree of apical lordosis does not increase pari passu with the lateral curvature because the farther the lordosis moves from the sagittal plane then the less it is compressed. While a number of possible etiological factors other than an abnormality in the sagittal plane, particularly neuromuscular factors, have their proponents as causative in idiopathic scoliosis, ^{19– 29} it has to be said that the presence of such factors (e.g., proprioception and equilibrial dysfunction) have in the main been demonstrated by comparing scoliotic patients with straight-backed controls. The finding of a marginally and sometimes significantly increased prevalence, but not a consistent presence, in the scoliotic population has perversely demonstrated that such factors cannot be etiological as they are not present in all cases(4) and indeed are not uncommon in straight backed controls. The only consistently present abnormality in all children with idiopathic scoliosis is an abnormality of spinal shape in the sagittal plane. Crucially, as we shall see in the next chapter, a painstaking 6-year prospective epidemiological survey involving 16,000 schoolchildren demonstrated that the lordosis developed before the scoliosis, confirming that it was the primary abnormality. ³⁰

Quite why the search for the cause of idiopathic scoliosis should continue is barely explicable unless the work of Adams, Roaf, and Somerville, let alone the Leeds Scoliosis Study Group, is either unfamiliar or for whatever reason disregarded. Of course publications from England do not seem to travel well!

The development of single structural curves at other sites and other curve patterns such as double or multiple curves is perhaps less easy to visualize. However, again, the essential lesion lies in the sagittal plane. Looking first at a single curve, usually only the central three or four vertebrae or so are truly lordotic. Inspect a posteroanterior (PA) radiograph of a left thoracolumbar curve (see ▶ Fig. 2.20). Lordotic vertebrae are defined as those that are laterally displaced to the convex side with the spinous processes rotated back toward the curve concavity (concordant rotation). The apical vertebrae will be confirmed as being so orientated but above and below these vertebrae will be neutral vertebrae with the spinous process in the center and beyond that, convex to the left, will be vertebrae whose spinous processes have remained rotated to the left, that is, lateral spinal curvatures with rotation to the convex side (discordant rotation). ³¹ Therefore, the compensatory curves above and below the structural curve are now kyphoscolioses. That of course must be the case because the central lordoscoliosis must be balanced above and below by asymmetric kyphoses to maintain overall three-dimensional balance in the spine. Moreover, the apex of a single thoracic curve is invariably located at T8 or T9. Should this lordotic area extend slightly lower or slightly higher than usual then there is insufficient “room” to fit in an asymmetric kyphosis before the spine meets the midline. A contralateral lordoscoliosis is therefore produced for balance purposes with the development of a typical double structural curve (see ▶ Fig. 2.16). This can be seen on any PA radiograph of a double curve. If the upper end of a structural thoracic curve has the same problem, then a contralateral lordoscoliosis is produced above with the production of the typical thoracic double major as described by Moe ³² (see ▶ Fig. 2.17). For single curves the lordotic apex never extends below L2 and so apical at T12/L1 is the thoracolumbar curve and apical at L2 is the lumbar curve. A thoracolumbar or lumbar single curve must therefore be either more lordotic or have greater flexural stiffness and there is evidence of both.

It is recognized that Euler’s theory provides only a simplistic explanation for spinal instability since the plastic deformations that occur in a curved column like the spine do not occur at the elastic buckling load (except in the case of a straight beam-column), and are not strictly analogous with a sudden deflection and collapse that occurs at critical load in a true elastic beam column. However, in all age groups, girls with scoliosis, even with curvatures uncoiled, are significantly taller than straight-backed counterparts. ^{33,​ 34} Furthermore, the finding that girls have more slender spines than boys at any given age would suggest that they have greater effective column length, L, in the Euler equation, a significant mechanism by which the spines of girls have a greater propensity to buckle under load, apart from any other difference there may be in flexural stiffness, EI, in such individuals.

Now we have to add in the effect of biological growth and then the scene is set for the development of the progressive deformity. Careful studies of sagittal profile during growth in normal children have demonstrated that the thoracic kyphosis gradually diminishes becoming minimal in the preadolescent growth period in both boys and girls (▶ Fig. 3.11). ^{35,​ 36} At this time girls are growing and developing quickly thus magnifying any trend toward flattening of the normal thoracic kyphosis. ^{33,​ 37} Boys, meanwhile, do not regain their normal thoracic kyphosis until 2 or 3 years later when they are going through their peak adolescent growth velocity. This accounts for the greater preponderance of the opposite deformity (Scheuermann’s disease) in boys. ^{38,​ 39}

Interestingly, the pattern of growth and development during adolescence does have a distinctly familial trend ^{40– 42} and of course several studies have shown that idiopathic scoliosis is a strongly familial condition. ^{41,​ 43– 48} Thus many of the factors that might disadvantage the stability of the spinal column are biological ones that would be expected to have similar genetic expression in family members. Therefore, there are a host of genes active in the spine at this time that would very much go against the rather simplistic notion that there can be a single gene for idiopathic scoliosis. Importantly, the suggestion that normal spinal shape might be a familial matter ⁴⁹ has been proven in our three-dimensional shape study of twins/siblings during growth. Using surface topography measurement in schools but also at the Annual British Twin Conference it was shown that correspondence in lateral profile rose from being insignificant in unrelated children through mixed-sex siblings, same sex siblings, mixed sex dizygotic twins, same sex dizygotic twins, through to monozygotic twins the correlation coefficients steadily increased in magnitude indicating ever closer correspondence so that with identical twins the lateral spinal profiles were virtually the same (▶ Fig. 3.12). ⁵⁰ Lateral spinal profile is therefore a very important genetic element in the pathogenesis of idiopathic scoliosis.

Related posts:

The Beginnings of Surgery for Spinal Deformities Deformities Associated with Neurofibromatosis Scheuermann’s Disease Miscellaneous Conditions Associated with Spine Deformities Idiopathic Scoliosis Basic Principles

Stay updated, free articles. Join our Telegram channel

Join