It is the author’s intent to use principles of mechanics, diagrams, models and examples to establish a background of organized understanding that will enable us to more accurately evaluate scoliosis deformities and render more complete and effective orthotic treatment. The difference between good and bad orthotic treatment of scoliosis lies not so much in the Cobb angle result; it lies in considering the total status of the deformity, treating all aspects, minimizing side effects and complications, and maximizing patient comfort, cosmesis, and sense of normalcy during treatment.
The spinal deformity we call scoliosis has several components and must be appreciated on two levels. We need to stand close to observe how the deformity manifests itself in certain areas of the spine and in individual curves. Focusing on localized features of the x-ray, we are able to note such things as vertebral tilt, rotation, and Cobb angle.
It is equally, if not more, important to step back a bit from the patient and x-ray to note the alignment of the upper thorax with respect to the pelvis and how well the “family” of curves are “balanced” about the centerline. It is also from a distance that we must observe and evaluate the patient’s spinal curvature in the sagittal plane.
One of the most important and basic principles in the treatment of idiopathic scoliosis (IS) is often overlooked: that any effort to treat any one curve in either the frontal or the sagittal plane will affect all curves in both planes. Secondary effects may be large or small, beneficial or detrimental. The author occasionally relates scoliosis geometry and treatment mechanics to Milwaukee orthosis design. That helps connect theory to practice. Patient examples are presented for the same reason. An understanding of the fundamental mechanical principles of IS treatment should be helpful in orthotic design and fitting, regardless of the type of orthosis.
GEOMETRY OF SCOLIOSIS
We should note and evaluate four deformity components in each scoliosis patient: 1) lateral vertebral displacement from the midsagittal plane; 2) vertebral rotation about the longitudinal axis; 3) vertebral lateral tilt; and 4) abnormal curvature in the sagittal plane.
To evaluate vertebral lateral displacement, we first must establish a useful reference midline (Figure 1). In the frontal plane projection, the healthy spinal column is normally vertically stacked on the sacrum and supported by it, so it seems most logical and proper to view the mediolateral position of spine elements relative to a vertical reference line passing through the center of the sacrum. This quickly tells us where each and every part of the spine is located relative to where it should be in the frontal projection. It helps us to evaluate the overall balance of the spine before, during, and after treatment. Lateral deviation from the midline is important for two reasons: 1) it is a major factor in cosmetic appearance, and 2) the size of gravitational loads tending to increase the magnitude of a curve are directly related to the spine’s deviation from the midline. If we do not evaluate the lateral deviation from sacral midline, the Cobb angle may fool us into an erroneous conclusion about the severity of lumbar and thoracolumbar scoliosis.
The patient in Figure 2 presented with a 29° left thoracolumbar curve (left). Months later, leftward deviation from midline increased significantly, even at the L3, L4 level (Figure 2, right). However, the Cobb angle has decreased. Misalignment of superior elements in the direction of a curve convexity has the effect of reducing the Cobb angle even when/as the deformity progresses.
Rotation of spine elements about the longitudinal axis in the vicinity of curve apices is familiar to all of us. None of us needs to be reminded to look for and evaluate the asymmetrical features that are manifestations of this aspect of scoliosis. If anything, too much is made of rotation. People have postulated deep and complex reasons for rotation based either in the anatomy or in the etiology of the disease. The truth is that there is nothing at all mysterious about this rotation. Rotation is a common and predictable aspect of beam buckling when the impressed bending loads are not in the direction of the beam’s least bending resistance (Figure 3).
Lateral vertebral tilt traditionally has been evaluated in an interesting way. We universally measure the total change in lateral tilt from the bottom of a given curve to the top of the curve and call that total change the Cobb angle. We virtually never measure its absolute value. Lateral tilt, in the absolute sense, is greatest at the ends of a curve. It is perhaps unfortunate that we do not take note of the magnitude and location of maximum lateral tilt. There are empirical indications that it is significant. The pioneers in spine surgery learned by trial and error that it is unwise to end a fusion right at the end of a curve. To do this is to terminate stability exactly where the lateral tilt is greatest and thus where transverse shear stress is large.
Many children with significant scoliosis also have abnormal spine curvature in the sagittal plane. These abnormal curvatures occur in a spectrum of patterns. The ends of the spectrum are two very different types of curvature. What the author calls Type I (Figure 4, left) is characterized by a relatively vertical sacrum and all sagittal curves at a reduced magnitude. Figure 4 (right) is Type II or what Dr. Blount (WP Blount, Medical College of Wisconsin, personal communications, 1980) referred to as the “horizontal sacrum” cases. These have an abnormally long lumbar lordosis extending well into the thoracic spine. The thoracic kyphosis typically is reduced in magnitude and displaced cephalad, often extending well into the cervical spine. Cervical lordosis is also reduced, or it may be nonexistent. Thoracic hypokyphosis is worth noting and studying because it can be nearly as cosmetically disturbing as the asymmetrical manifestations of scoliosis. In addition, it is one of the important factors in thoracic distortion and volume reduction that, in extreme cases, lead to cardiopulmonary compromise. Some of our traditional and current ways of treating scoliosis can create or exacerbate abnormal sagittal curvature.
EFFECT AND OPTIMIZATION OF ORTHOTIC FORCES ON THE THORACIC SPINE
In some cases, corrective or holding forces can be as damaging as they are helpful.
The force vector of the thoracic pad is directed anteromedially. It can be resolved and represented by two orthogonal component vectors (Figure 5). The anteriorly directed component of that force, pushing against the rib prominence, tends to reduce the rotational aspect of the deformity (Figure 6). Inasmuch as rotation is coupled to vertebral tilt and lateral misalignment, it also reduces these aspects of the deformity. Finally, the anterior component also reduces thoracic kyphosis (Figure 7). This may be detrimental to the child, depending on the pre-existing amount of thoracic kyphosis.
The child whose x-ray appears in Figure 8A, had a thoracic kyphosis of 25° before beginning Milwaukee orthosis treatment. The x-ray in Figure 8B, taken several months after the completion of treatment, shows her thoracic kyphosis had been reduced to a lordosis of 4°. This undesirable modification of sagittal curvature, as a side effect of scoliosis treatment, is not uncommon. Fortunately, it is avoidable by the measures explained below.
The medially directed component of the thoracic pad force pushes on the ribs, which in turn push the errant vertebrae back toward the centerline. This is generally beneficial, but we must be careful that in our zeal to reduce the Cobb angle of an individual curve we do not increase the lateral misalignment above and below the apex of that curve, decreasing the balance of the family of scoliosis curves. We must never forget that we are treating children, not Cobb angles.
The medially directed component of a thoracic pad force has another interesting effect. That effect is on vertebral tilt. This is best illustrated by a demonstration.
Figure 9 shows a simple frame for holding the spine models and applying forces in any desired direction. The columns set in this frame do not fully represent the human spine. The restraint condition, at the cephalad end in particular, would be more valid if it allowed some realignment. However, the model is sufficiently valid to illustrate qualitatively the effect of applied holding forces. The model illustrates quickly and clearly the importance of applying a thoracic pad at the correct level.
Figure 10 shows that a force applied directly to the center of the undeformed model results, as expected, in a curve with apex at the point of force application. If we attach a rib to the column and apply the medially directed force to the rib, we see that something different happens (Figure 11). The rib transmits not only a force but a tilting moment at the spinal column where it is attached. The tilt is toward the side of the force application. The apex of the resulting curve occurs below the level of the rib attachment. For maximum effect in reducing an existing scoliosis curve, pushing against ribs leading to the spine at and above the curve apex is indicated. Tests on subsequent models substantiate this finding.
The spine model in Figure 12A was fabricated to include a lateral curve. When a holding force of one kilogram is applied to a rib just superior to the apex, correction is excellent (Figure 12B). Below the curve apex, the column is tilted rightward to begin with, and applying the same force to one of those ribs results in a much less satisfactory realignment (Figure 12C).
The reader may have noted the slight upward direction of the forces in this demonstration. This is intentional and meant to simulate the slightly upward pull common for strap-mounted thoracic pads.
The caudad slope of the rib in this model is meant to approximately represent such slope in the lower thorax. More cephalad ribs generally have less of such slope.
Particular care should be taken not to place right thoracic pads too low in cases of the common right thoracic/left lumbar double major curves. Low placement will exacerbate the tilt component of the deformity exactly where it is maximum. This is a common mistake.
The patient with the anteroposterior (AP) x-ray in Figure 13A was treated with a Milwaukee orthosis. In addition to a left lumbar pad, a left low oval thoracic pad was installed to push the lower thoracic vertebrae to the right and reduce the rightward tilt. The right thoracic pad was positioned very high and adjusted for moderate pressure only so it would not prevent the left pads from being effective in pushing the spine to the right.
Figure 13B is a photo of S.M. in her Milwaukee orthosis. Figure 13C shows her AP x-ray in the orthosis after 6 months of treatment.
The author can find no basis in mechanics or experience for always keeping the pads below the apex of scoliosis curves. The people who advocate that in their “recipe” for scoliosis treatment are obtaining Cobb angle reductions primarily by means of abdominal and circumferential compression, not from precise pad placement.
Blount and Moe 1 discovered early that a snug pelvic section that reduced lumbar lordosis is an important aspect of the orthotic treatment of scoliosis. The abdominal compression stabilizes the lumbar spine in reduced lordosis so that a lumbar pad can be effective. The snug pelvic section also provides a stable foundation for the superstructure.
Finally, the snug circumference (especially if it is extended to include the lower thorax), and abdominal compression have the effect of increasing intra-abdominal pressure. The increased pressure, which is further elevated with each pulmonary inspiration, pushes downward on the pelvic floor and upward on the thorax (Figure 14). This has the effect of superimposing a longitudinal stretch on the lumbar and thoracic spine. These are powerful tools capable of significant Cobb angle reductions and are features in many scoliosis orthosis designs. However, a note of caution is warranted. Concern has been expressed regarding the potential effect on kidney function, 2,3 respiratory mechanics, 4,5 and esophagitis. 6,7
It has long been recognized that passive reduction of lumbar lordosis will induce an active reduction of thoracic kyphosis in the standing posture of a neurologically normal person. Body casts employing this principle more than 40 years ago were referred to as “antigravity” casts. A little more recently we have the excellent experimental research work report by Lindh. 8 It should be noted that in the Lindh experiments, strong abdominal compression was part of the technique for passive reduction of lumbar lordosis. Such strong reduction of lumbar lordosis is contraindicated on many patients because it induces excessive active reductions in thoracic kyphosis.
When a thoracolumbosacral orthosis (TLSO) is applied for treatment of scoliosis, reduction of thoracic kyphosis may be as great as 25° (Figure 15). This tendency also occurs with a Milwaukee orthosis, but the Milwaukee superstructure gives us an opportunity to more precisely manage the sagittal curvature of the mid and upper spine.
Exactly how one may design a Milwaukee-type cervico-thoracolumbosacral orthosis to avoid creating or aggravating the thoracic hypokyphosis is described in the following paragraphs. Understanding how those design interventions work will help us to understand how other designs may be beneficially altered.
First, when appropriate, we need to reduce the anteriorly directed component of any thoracic pad force. 9,10 This is done by extending the thoracic pad laterally at least to the midaxillary line, and using a short anterior outrigger (Figure 16) or no outrigger at all. This creates a more medially directed pull on the pad. In addition, the right posterior upright should be bent to a position where it is close to, but not pushing anteriorly on, the thoracic pad. There must be room in the left posterior midthorax for motion during back-arching movements and postures. The neck ring must be positioned a bit more anterior than otherwise. Figure 17 illustrates thoracic pad force vector management.
Another design feature that will greatly help to maintain thoracic kyphosis is to exert a well-dispersed pressure against the lower anterior thorax. This is accomplished by means of a cloth panel, such as that shown in Figure 18.
When sacral inclination is abnormally vertical (Type I, see Figure 4), passive forces to further reduce lumbar lordosis should be minimized as much as is practical. However, when the patient has an abnormally horizontal sacrum (Type II, Figure 5), a pelvic section with strong lumbar flexion will help to normalize sagittal curvature. Other measures, outlined in the previous two paragraphs, are appropriate for both Type I and Type II hypokyphosis.
The patient with thoracic scoliosis whose AP x-ray appears in Figure 1 presented with a hypokyphotic thoracic spine. A fabric panel was used to apply pressure to the lower anterior thorax. The thoracic pad force and superstructure alignment were managed as described. Figure 19 shows his lateral x-rays before and during treatment. As shown, instead of exacerbating his thoracic hypokyphosis, the orthosis helped to bring his sagittal spine curvature to a more normal configuration.
BIOMECHANIC ANALYSIS OF ORTHOTIC TREATMENT OF THORACOLUMBAR AND LUMBAR SCOLIOSIS CURVES
The orthotic treatment of thoracolumbar and lumbar curves usually can be accomplished with a low-profile orthosis with minimal coverage and constraint of the patient’s thorax. This analysis relates to a specific TLSO design to better illustrate the interaction among forces, design, and desired effect. For this purpose, we have chosen the low-profile Gillette-style TLSO. To simplify our discussion, we refer exclusively to left lumbar or thoracolumbar curves because IS exhibits such a strong pattern of producing left curves in this region of the spine.
The TLSO exerts lumbar pressure by means of a lumbar pad illustrated in Figure 20. For thoracolumbar scoliosis (apex at T12, L1), a low left thoracic pad is also used (Figure 21). The thoracic pad is fabricated of one-millimeter thick polyethylene, faced with six-millimeter polyethylene padding. It is flexible enough to follow the exact contour of the thorax. It is slung in place in the fashion of a Milwaukee thoracic pad (Figure 22) and is easily adjusted for position and pressure. Lead foil strips are built into all pads so that their placement can be precisely and easily checked on an x-ray.
The trim lines of the orthosis anteriorly are equivalent to those of a Milwaukee orthosis pelvic section. Posteriorly and laterally, they differ considerably. Two posterior thoracic extensions are created paraspinally and terminate two centimeters inferior to the scapulae. These extensions are about seven centimeters wide, with lateral borders sweeping almost straight downward to blend with the usual lateral trim line above the waist. On the same side as the lumbar pad, a trochanteric extension is created by continuing the inferior posterior border straight around laterally past midline and then sharply upward to blend with the rectus femoris relief. On the opposite side, the trim line sweeps upward to blend with rectus remoras relief along a line passing superior to the greater trochanter. Reinforcement is provided by two metal reinforcement bars paraspinally and by a semicircumferential bar in the waist groove on the same side as the lumbar pad (Figure 23).
This orthosis leaves a major portion of the trunk uncovered and unconstrained, allowing a great deal of openness, torso motion, and exercise in the orthosis. Passive control of the upper thorax and neck position is absent, by choice, with this low profile design. We are relying significantly on the patients’ desire and ability to line up their shoulders and head in a proper relationship to their lower torso and in proper relationship to the horizon (Figure 24). Thus, this orthosis usually is not appropriate for patients with neuromuscular diseases.
With that brief description of a TLSO design for treating lumbar and thoracolumbar adolescent IS we discuss the most important forces exerted between the body and the orthosis in the coronal and sagittal planes. We demonstrate the relationship of these forces to orthosis stabilization and design.
The three arrows in Figure 25 represent the most important forces exerted by the orthosis in the coronal plane. They are the lumbar force, FL, the gluteus medius force, FGM, and the trochanteric force, FT. Of course, the body is exerting equal and opposite forces on the orthosis (Figure 26). FL is the only directly therapeutic force of the three. It is generated by a lumbar pad pressing anteriorly and medially. This serves to derotate and translate the misaligned lumbar vertebrae back to their proper position (see cross-section view in Figure 20). As the orthosis is cinched tight, the two other forces, FGM and FT, also are generated. FGM is a force on the right hip in the area between the iliac crest and the trochanter. The third force, FT, occurs in the area of the left greater trochanter. The top two forces tend to push the orthosis into a leftward list. The trochanteric force, FT, serves to resist that action and maintain proper alignment of the orthosis.
If the trochanteric extension is absent, FT must be generated along the lateral, inferior border of the orthosis just above the trochanter (Figure 27). With a smaller moment arm, this force is not as effective in maintaining the alignment of the orthosis. The orthosis tilts, and the left lateral inferior edge sinks into the soft tissues of the hip. The misalignment and discomfort limit the effectiveness of the orthosis.
As mentioned, superior to the lumbar pad, the patient provides active control to maintain the head and shoulders in an acceptable attitude and alignment. This active control obviates the need for another passive holding force on the right side above the lumbar curve. In fact, such a force would usually increase the leftward misalignment of the upper thorax (Figure 28). Such leftward misalignment would decrease the Cobb angle of the curve but is not in the patient’s greater best interest.
To examine the important forces in the sagittal plane, we use a lateral view of the orthosis (Figure 29). Posterior pressure near the top of the thoracic extensions (represented by F1) and near the bottom of the buttocks (represented by F3) combined with firm abdominal pressure (represented by F2) passively hold the lumbar spine in a relatively flexed position. This lumbar flexion brings the lumbar vertebrae to a more posterior location, where it is possible to effectively push against them with the lumbar pad. Proper maintenance of these forces, including the lumbar pad force, requires a stable posterior contour and a stable anterior-posterior dimension.
We have observed that large constant stresses on polypropylene at body temperatures will cause polypropylene to “creep” to a more relaxed configuration. This relaxation seriously degrades the corrective force exerted by the orthosis. This is the reason for the metal reinforcement bars. The left paraspinal bar maintains the contour supporting the lumbar pad. The right paraspinal bar maintains the relative void in the right lumbar area. The semicircumferential metal waist bar prevents the orthosis shell from creeping to a larger anterior-posterior dimension, thereby losing the ability to exert an anterior force component. It is interesting to point out that the Monel (Special Metals Corp., Huntington, WV) metal band served this purpose in the original leather Milwaukee girdle design. Finally, we know that increasing intra-abdominal pressure superimposes a traction force (represented by FIP in Figure 29) and thus partially unweights the spine. This obviously aids in curve correction.
We must not conclude this discussion without explicitly addressing a potential TLSO complication. We refer to how passive reduction of lumbar lordosis, common with most, if not all, lumbar scoliosis treatment, induces a corresponding active reduction of thoracic kyphosis. That induced kyphosis reduction may be of no consequence or may even be beneficial, as it was for patient A.W. in Figure 30. A.W. had a pre-existing thoracic kyphosis of 53°. Her lateral x-ray in the low-profile orthosis prescribed for left lumbar scoliosis shows a kyphosis reduction to 37°. However, M.M. in Figure 31 had a pre-existing thoracic kyphosis of only 15°. Her in-orthosis lateral x-ray shows a virtually complete loss of thoracic kyphosis. It is our experience that alteration of sagittal spine curvature during scoliosis treatment is a secondary effect that is not fully reversed when scoliosis treatment is discontinued.
For the practitioner, it seems right to follow a biomechanics analysis with a clinical example. One example proves nothing but does give some “life” to the analysis and helps one remember the principles taught.
With this in mind, examine a set of x-rays. Figure 32 is the AP x-ray of an 11-year, 9-month-old child. The x-ray reveals a 25° left lumbar curve with the apex or horizontal vertebra at the L1–L2 disc. A compensating right thoracic curve of lesser magnitude occurs above. Next, observe the alignment. The entire vertebral column between T10 and L4 lies to the left of the centerline. The vertebral column above T8 is well centered over the sacrum. The position of the pedicle rings on the x-ray indicated a grade 1 rotation in the vicinity of the apex of the curve. This corresponds with a left lumbar prominence of 1.5 centimeters measured on the patient. Finally, this AP film shows that the pelvis is level. The AP x-ray is surprisingly reliable as an indicator of leg length discrepancy (discrepancies greater than one centimeter should be compensated by a shoe heel elevation for the duration of orthotic treatment). The left side bending film (Figure 33) indicates that the left lumbar curve can unbend to −5°, giving an impression of flexibility. However, it is even more important to view the right side bending film (Figure 34) because, to correct this curve, the lumbar vertebrae must be pushed to the right. The right side bending film reveals that the centers of the bodies of L3, L4, and L5 do not reach the reference centerline on right side bending. The entire mediolateral mobility range of motion of the lower lumbar vertebrae is to the left of center (Figure 35). It will be difficult to push these vertebrae back into alignment. The right side bending film also indicates that the compensating right thoracic curve is flexible, as we would expect.
The lateral standing x-ray (Figure 36) reveals a lumbar lordosis of 60°, which is probably near the upper limit of what we would consider normal. The thoracic kyphosis, at 45°, is a bit larger than average but not abnormal.
This x-ray information is important to the orthotist. The order form should contain data on the location, size, and flexibility of each curve. It also should indicate size of thoracic kyphosis and amount of pelvic obliquity. It is also helpful to have a rough facsimile of the mediolateral spine alignment on the order form (Figure 37).
Now that we have completed the interpretation of the patient’s x-rays, we must define what must be accomplished for the patient. The lumbar vertebrae must be pushed to the right toward their proper position on the centerline. The good alignment of the upper thorax over the pelvis should be preserved. The rotational component of the deformity and the Cobb angle should be reduced.
A Gillette-style TLSO was provided for this patient based on the biomechanics principles presented earlier. Figures 38 and 39 show her x-rays in the orthosis. The radio-opaque markers in the lumbar pad (Figure 38) indicate that it is well located between the 12th rib and the iliac crest. The two markers shown here bracket the crest of the pad.
Leftward misalignment of the spine has been reduced significantly at all levels between T7 and L5. Alignment of the vertebral elements above T8 remains excellent. The Cobb angle of the curve has been reduced from 25° to 12°. Vertebral rotation has been reduced.
A check of the lateral x-ray (Figure 30) reveals that the thoracic kyphosis has been reduced from 45° to 30°. This reduction of thoracic kyphosis is an “induced” active postural compensation in reaction to the passive orthotic reduction of the patient’s lumbar lordosis. A reduction of 10° to 20° is customary and is the reason we do not think a low-profile orthosis is appropriate for patients with scoliosis with a pretreatment kyphosis of less than 15°.
This analysis of the biomechanics of one example (Figure 40) of a low-profile TLSO was prepared with the hope that it will contribute to an understanding of general principles involved in treating lumbar and thoracolumbar scoliosis.
Virtually all of the analyses, diagrams, biomechanic models, patient radiographs, and orthotic examples presented originated during the author’s tenure as Director of Habilitation Technology at Gillette Children’s Hospital. Spinal orthotic services were delivered under the orthopedic leadership of Dr. Robert B. Winter, Chief of Gillette’s spine service. Dr. John Lonstein also provided direction and review to spinal orthotic progress at Gillette. The importance of Drs. Winter and Lonstein to the author’s work notwithstanding, the reader should not assume total endorsement of this article by either of those men.
The Gillette Medical Education and Research Fund underwrote a multitude of costs associated with photographic documentation and artwork. The author acknowledges the important contributions of the very capable spinal orthotists and technicians (Fred Sutterfield, Douglas Grimm, Frank Ransom, James Isenor, Charles Schemitsch, Paul Swanlund, Catherine Voss, and others) he had the privilege to supervise. Finally, and most importantly, the author acknowledges and thanks the patients and parents who trusted us to do our best and suffered the consequences of our ever-present limitations.
This document includes some excerpts, photos and drawings which were published in 1980 and 1981 by Gillette Children’s Hospital within a set of booklets, A Thoraco-Lumbo-Sacral Orthosis for Idiopathic Scoliosis, part I and part II, which were authored by J. Martin Carlson. The biomechanical principles illustrated in the photos and drawings are as accurate today as they were 20 years ago. The Gillette staff have continued to evolve and change material and design details to further improve cosmesis and orthotic options. Mr. Carlson and the JPO thank Gillette Children’s Specialty Health Care for their permission to republish.
There are some excerpts, photos and drawings in this document which were published in Orthopädie-Technik in articles authored by J. Martin Carlson (August 1986, October 1991, April 1995 and June 1995). Mr. Carlson and the JPO thank Verlag Orthopädie-Technik for their permission to republish.
1. Blount WP, Moe JH. The Milwaukee Brace. Baltimore/London: Williams & Wilkins; 1980.
2. Aaro S, Berg U. The immediate effect of Boston brace on renal function in patients with idiopathic scoliosis. Clin Orthop 1982; 170: 243–247.
3. Berg U, Aaro S. Long-term effect of Boston brace treatment on renal function in patients with idiopathic scoliosis. Clin Orthop 1983; 180: 169–172
4. Kennedy JD, Robertson CF, Hudson I, et al. Effect of bracing on respiratory mechanics in mild idiopathic scoliosis. Thorax 1989; 44: 548–553.
5. Kennedy JD, Robertson CF, Olinsky A, et al. Pulmonary restrictive effect of bracing in mild idiopathic scoliosis. Thorax 1987; 42: 959–961.
6. Kling TF Jr, Drennan JC, Gryboski JD. Esophagitis complicating scolioses management with the Boston thoracolumbosacral orthosis. Clin Orthop 1981; 159: 208–210.
7. Gryboski JD, Kocoshis SA, Seashore JH, et al. Body-brace oesophagitis, a complication of kyphoscoliosis therapy. Lancet 1978; 2( 8087): 449–451.
8. Lindh M. The effect of sagittal curve changes on brace correction for idiopathic scoliosis. Spine 1980; 5: 26–36.
9. Winter RB, Carlson JM. Modern orthotics for spinal deformities. Clin Orthop 1977; 126: 74–86.
© 2003 American Academy of Orthotists & Prosthetists
10. Carlson JM, Winter RB, Lonstein J. Biomechanische aspekte in der rumpforthetik des kindes. Orthopädie-Technik 1991; 42: 713–721.