8 The Case Against Bracing
The previous chapter argued in favor of brace treatment; this chapter will make the case against it. To assess the role of bracing in the management of idiopathic scoliosis, it is crucial to understand this deformity and why it is treated.1 We can then begin to understand why the results of bracing for idiopathic scoliosis have been so disappointing.
What Deformity Are We Treating?
Chapter 3 of this book, on the pathogenesis of idiopathic scoliosis, describes in detail the geometry of idiopathic scoliosis and its development. The three-dimensional nature of structural scoliosis has been known for centuries,2 and the essential features of the deformity are first a lordosis, second a rotation/torsion, and third a lateral deformity or scoliosis.3–6 Scoliosis is fundamentally a front-back problem, not a right-left problem. Once the lordosis develops in the mid-lower thoracic region, it progressively gets closer to the axis of spinal-column rotation and tries to get out of the sagittal plane by buckling.6 This is the basis of the Adams forward-bend test, developed in the early 1860s, which compresses the sagittal plane and enhances the rotational prominence in scoliosis (see Fig. 3.5 ).2
Going back to the geometry of the deformity in scoliosis, the posteroanterior (PA) view of the patient shows the lateral curvature with rotation such that the posterior elements turn into the curve concavity and it can be clearly seen that the back of the spine is shorter than the front (see Figs. 3.6 and 3.7 ).7 However, above and below this central area of structural scoliosis are compensatory scolioses that act to bring the spine into straight alignment. The nature of these compensatory scolioses is that of asymmetric kyphoses balancing the central area of lor-doscoliosis. This was precisely Roaf’s concept of curve progression, holding that the central area of lordoscoliosis was compressed by the kyphoses above and below it.4 His classic article should be compulsory reading for anyone trying to understand this three-dimensional deformity in scoliosis.
Unfortunately, the problem isn’t so straightforward, because as soon as the lordosis buckles out of the sagittal plane, in comes the Hueter-Volkmann law to produce asymmetric epiphyseal loading and three-dimensional vertebral wedging that increases progressively toward the curve apex. As the Cobb angle increases progressively from 10 degrees, to 20 degrees, 30 degrees, 40 degrees, and beyond, so does the degree of asymmetric vertebral wedging ( Fig. 8.1 ).1,6,8 Supine, side-bending, or traction films of the patient now become progressively more like the erect film. The components of the jigsaw puzzle of the spinal column fit less well together and the resulting so-called stiffness or lack of flexibility does not come from an added component of soft-tissue stiffness, but simply from the progressive loss of stackability straightness of the spine the greater its curvature becomes. Therefore, if the word “unstackability” existed, it would very nicely describe the problems that occur with increasing curve size.
Even a 20-degree curve causes some loss of stackability, and it would challenge the most fertile mind to devise an orthosis that could not only stop the progression of curvature in scoliosis, but could actually produce some form of correction.
Meanwhile, looking at Scheuermann’s disease, the adolescent idiopathic deformity opposite that of scoliosis, it is a thoracic hyperkyphosis with the sagittal plane of the spine moving progressively further behind the axis of spinal-column rotation (see Fig. 3.11 ).6 Thus, Scheuermann’s disease progresses in the sagittal plane without buckling potential. Because this is a uniplanar deformity, its correction requires extension, and indeed, 1 or 2 years in an extension cast or brace leads to a true physiological correction of the deformity.9 If, therefore, hyperkyphosis needs extension for its correction, lordosis needs flexion, and it is flexion that renders it particularly rotationally unstable.2 It therefore does not appear that the deformity of idiopathic lordoscoliosis is treatable without surgery.
Why Are We Treating This Deformity?
James, in Edinburgh, divided idiopathic scoliosis according to its age of onset into the categories of infantile (birth to 4 years of age), juvenile (age 5 to 9 years), and adolescent (age 10 years to maturity).10 There are, however, only two phases of increased growth velocity: during infancy and again during adolescence.11 At birth, mean body length is 50 cm, and during the first year of life babies grow by half of that (i.e., 25 cm). During the second year of life they grow by half of the latter figure (i.e., 12.5 cm), and in the third year of life they again grow by half of this (i.e., 6 cm).12 The growth rate then remains steady until the adolescent growth spurt, when this trend reverses with increased growth velocity.
It is during the first year or two of life that a progressive thoracic scoliosis can impair the proper development of the heart and lungs,13,14 and for this reason progressive infantile idiopathic scoliosis threatens health in adulthood. By contrast, there is no such risk if a deformity occurs after the age of 5 years.15
James had some difficulty in identifying true cases of idiopathic scoliosis of juvenile onset because if a deformity existed at, say, the age of 6 years, it was more likely to be a carryover from infancy than a disorder of true juvenile onset.10 Consequently, there is much in favor of having two types of scoliosis according to age of onset: (1) early-onset idiopathic scoliosis beginning before the age of 5 years; and (2) late-onset idiopathic scoliosis beginning after the age of 5 years.16
Early-onset idiopathic scoliosis is therefore treated to prevent future cardiopulmonary problems, whereas treatment for late-onset idiopathic scoliosis is done to ease deformity and improve appearance, although not without the potential for producing significant psychosocial distress. Eating disorders are 10 times more common in girls with significant scoliosis than in their counterparts with straight backs, and at the worst end of the spectrum is frank osteoporosis just when peak bone mass occurs.17 Once the deformity is corrected, such patients very quickly catch up with their straight-backed peers and become “normal” girls.
When bracing was first introduced for idiopathic scoliosis it was erroneously thought that preventing progression of the condition (particularly to beyond 60 degrees) would reduce the likelihood of the spine and chest providing a hostile environment for the internal organs.18–20 At the same time, surgery was a major undertaking, with the control of progression rather than correction being the main goal, and was accompanied by the possibility of serious neurological complications and pseudarthrosis requiring reoperation.21 Indeed, James not only performed massive spinal fusions, but reinspected them at intervals of about 3 months to make sure that union was occurring, and if not, applied supplementary bone grafts.10 Perhaps not unreasonably, children and families faced with the possibility of a hideous deformity with a shortened lifespan or a very major operation opted for any treatment offered, even to the point of spending 23 hours per day in a brace.
However, the deformity of late-onset idiopathic scoliosis is not and never has been a long-term organic health problem, and if the deformity does become unacceptable to the patient and family, a single modern surgical operation with only a few days spent in the hospital can easily restore body symmetry with minimal risks.
Is the Deformity in Scoliosis Treated Successfully Without Surgery?
Late-onset Idiopathic Scoliosis
All sorts of contraptions have been prescribed for scoliosis over the centuries. Most were in the form of racks or turn-buckles, from the design of which it would seem that pain was the chief objective.22 Indeed, having a severe scoliosis was a serious stigma for which these sorts of horrific devices were deemed entirely appropriate. Unfortunately, patients with idiopathic scoliosis are, regrettably, still stigmatized, and endure continued bullying in schools. It is incomprehensible that idiopathic scoliosis can be regarded as a simple cosmetic condition when it can cause significant psychosocial distress at a very vulnerable age, weight loss, osteoporosis at the extreme, and social outcasting as the norm.17 This is what drives scoliosis surgeons to ever-better goals for their patients.
The Milwaukee brace was introduced to support poliomyelitic scoliosis after surgical intervention, and was never designed to be a nonoperative treatment of any type of spinal deformity.23 As the polio epidemics ended with successful vaccinations, the Milwaukee brace came to be prescribed for idiopathic scoliosis instead. Looking at the explanations and cases that Blount and Moe23 reported in their textbook on treatment with the Milwaukee brace, it is clear that what was being treated was a straightforward right/left spinal asymmetry in the frontal plane, and indeed, that still seems to be the prevailing view. No adequate biomechanical explanation for the possible efficacy of brace treatment was put forward other than, to begin with, simple stretching of the spine with the upper part of the orthosis under the mandible and the lower part fitted to the pelvis. Serious problems with dentition then led to the orthosis having a cervical choker rather than an occipitomandibular piece.24
It was then conjectured that the Milwaukee brace might work by way of three-point fixation: at the top and bottom, and with a pad just below the middle on the convex side. X-ray films of patients in the brace were encouraging with, for example, a 30-degree standing curve when out of the brace and a 20-degree curve in the brace. The curve was not allowed to move from this semi-improved position for 23 hours a day, whereas a nonbraced child would move its spine through as full a range as possible thousands of times a day during normal activities of daily living.16
Then it was observed that better improvement was obtained when the pelvic component of upright posture flattened the lumbar lordosis of the scoliotic spine. No adequate biomechanical explanation was put forward for this, but clearly, with a flattened lumbar lordosis the upper torso was pitched forward, leading to spontaneous thoracic hyperex-tension, which in turn pushed the lordosis back towards the sagittal plane,16 the opposite of the forward-bend test (see Fig. 4.1 ). Although this might reduce the magnitude of the scoliotic component of the deformity, it unquestionably increased the thoracic lordosis, with a reportedly detrimental effect on pulmonary function.25
“Evidence-based medicine” has now been an “in” phrase for several years, and means “the integration of individual clinical expertise with the best available external evidence from systematic research, particularly concerning the scientific principles governing treatment.”26 It is difficult to identify any criteria by which the nonoperative treatment of late-onset idiopathic scoliosis adheres to the principles of evidence-based medicine. Not only was the orthosis used for treating it devised for a completely different type of scoliosis under entirely different circumstances, but the early experience with this orthosis was reported at a time when it was not conventional to apply statistical methods or any other rigorous analytical approach to the problem. Yet this is what we have done, this is what has happened, and because we are the senior scoliologists of the day, we tell others that they had better do the same. None of us could possibly dissent. In the 1970s, the designer of the Milwaukee brace, Dr. Walter Blount, and his colleagues, reported retrospectively on only half of a total number of 94 patients treated with the brace.27 Notwithstanding, they went on to state “there have been no published data with regard to long-term end results comparable to the available follow-up studies of untreated patients.”
The Campbell Clinic reported on 52 patients, of an original cohort of 125, for whom treatment with the brace began at the age of 14 years and lasted until nearly the age of 17 years (despite spinal growth being effectively complete in girls by the age of 15 years).27 Mean Cobb angles were initially >40 degrees, and improvement was between 0% and 20%. Then the Minneapolis Group reported their results with patients aged 8 to 16 years.28 Of their original 133 patients, 30 were lost to follow-up and 29 were treated surgically “because of a poor response to the brace.” Thus, 59 patients (44%) have to be regarded as having experienced treatment failure. The mean final curve in the 74 patients (56%) who were followed was only a few degrees smaller than the original curve. The authors stated quite rightly that “the role of the Milwaukee brace and the treatment of idiopathic scoliosis is still unclear.” They then asked the important question: “What then is the proper role of the Milwaukee brace in scoliosis treatment?”, and emphasized that “further follow-up must be obtained on these patients.” Despite this important question, no scientific data were being gathered about the efficacy of the brace and no prospective studies were undertaken. Rather, it was stated by more senior scoliosis surgeons that the Milwaukee brace worked, and junior surgeons were obliged to agree.
Then, in 1984, Miller and his colleagues in Gothenburg pointed out that there were no controlled studies of brace treatment and reported retrospectively about 144 braced patients versus 111 untreated patients with both groups having a mean deformity of <30 degrees.29 No evidence was provided in favor of bracing. Nonetheless, the Gothenburg investigators felt that a controlled, randomized prospective study was warranted. That was somewhat surprising, because this type of prospective investigation should be based on clear retrospective evidence of benefit, so as to determine, for example, how long treatment is required, for which group (boys or girls), and for what age range. Because the Swedish retrospective study demonstrated no benefit from brace treatment, the need for a prospective study is questionable.
In the next decade, Caroline Goldberg went from Dublin to Boston to determine the efficacy of the Boston underarm brace in late-onset idiopathic scoliosis. She and her coworkers compared 32 braced patients in Boston, for whom relevant data were adequate, with 32 nonbraced controls in Dublin, and their results were published in 1993.30 There was again no difference between the braced and nonbraced patients, from which the authors concluded that “this raises very seriously the question of whether bracing can be considered an effective way of altering the natural history of late-onset idiopathic scoliosis.”
Bracing has persisted despite the lack of new data in support of its efficacy. Its proponents state that “as high quality clinical research studies have been available in the 1980s and early 1990s the proper place of brace treatment for adolescent idiopathic scoliosis has become apparent,”31 an extraordinary assertion in that precisely the opposite conclusion is contained in the literature.
In 1994 the Minneapolis Group published the results for more than 1000 patients who had been braced between 1954 and 1979.32 The braced patients were compared with 727 patients who were not braced but were followed for evidence of curve progression. Overall, there was no curve progression in 77%, another 22% needed operative intervention, and the remaining 791 were managed with a brace only. Failure was defined as an increase in Cobb angle of at least 5 degrees or surgical intervention. The failure rate of curves of <30 degrees was 40% at the cessation of treatment and more than 50% at the latest follow-up. As compared with the previous study29 it was suggested that failure rates were lower in the braced group. Noonan and Weinstein expressed concern about the large number of patients being excluded from the study and that only 28% of the study questionnaires had been completed.33
Then, in 1995, the Puerto Rico Group compared bracing in 54 patients with 47 controls.34 There was a significantly greater number of patients in the control group who required surgery, but more than twice as many in the latter group (77%) had thoracic curves as did those in the treated group (46%). Thoracic curves have always been shown to be associated with a much greater progression potential than other spinal curves in scoliosis.35 Furthermore, the Puerto Rico Group stated categorically that patients “fully complied with treatment,” whereas the excellent study in Oxford by Houghton et al, using hidden compliance meters, showed that upper-middle-class Oxford schoolchildren didn’t wear their braces much more than 10% of the prescribed time despite what they or their parents said.36
With so many stakeholders it was clear, despite the absence of any evidence-base from retrospective studies about bracing favorably altering the natural history of scoliosis, that a prospective controlled trial of bracing would be undertaken. The results of this trial were published in 1995,37 and the authors quite rightly stated that as regards previous reports, “none of these studies met the stringent criteria for scientific evidence that must be used to prove the effectiveness of treatment.”
They went on to state that “a well-designed study must include a large cohort of similar patients with similar patterns and sizes of deformity and that they should be randomized to different treatment methods and followed until at least skeletal maturity.” Unfortunately, randomization was impossible because centers that used bracing would not stop using it and those that did not brace would not use a treatment that they did not believe worked. Furthermore, as will be seen, the patterns of deformity were significantly different.
They decided to use three arms in a study of 286 patients, consisting of: (1) 129 observed patients; (2) 111 braced patients; and (3) 46 electrically stimulated patients, although this last treatment had been discarded years earlier ( Table 8.1 ). Perhaps they thought that the electrically stimulated patients would form another control group.
Another problem was their selection of treatment failure as an increase of at least 6 degrees in the Cobb angle. That is close to the measurement error of the angle itself, but more importantly, a range of 20 degrees to 26 degrees is different from a range of 40 degrees to 46 degrees. As the lordosis in scoliosis buckles out of the sagittal plane, curve size diminishes with rotation. A curve of 40 degrees is therefore much more than twice the size of a curve of 20 degrees (see Fig. 2.9 ). Accordingly, mean values and percentage changes are difficult to interpret.
In any event, the braced patients did best, with a 36% failure rate, the observed patients had a 52% failure rate, and the electrical-stimulation group had a 63% failure rate. These differences in proportions were statistically significant, and from them it could be interpreted that bracing eases idiopathic scoliosis, observation does nothing, and electrical stimulation worsens the curves in scoliosisi!
On the face of it, therefore, the braced group in the study looked as if it had done better, but the following article, in the same edition of the Journal of Bone and Joint Surgery in which the study findings were published, reported on factors in the study that would be indicative of curve progression ( Table 8.2 ).38 The most dominant such factor was curve site, with thoracic curves being significantly more progressive than thoracolumbar curves. Meanwhile, when the trial investigators examined the proportions of the more progressive thoracic curves, they found that almost 90% in the stimulated group, 81% in the observed group, and a mere 68% in the braced group were thoracic curves. Similar proportions existed in the Puerto Rico Study. It would be difficult to better stack the odds in favor of bracing.
Then, in 1996, the Iowa Group reported on 102 of 111 patients treated with a Milwaukee brace, with a mean time from the cessation of bracing to follow-up of more than 6 years.39 Although there were no controls, the authors did not favor bracing, and concluded that “it is currently impossible to state that bracing effectively alters the natural history in immature patients who are at high risk for curve progression.”
Brace | Observed | Electrical | |
Thoracic | 68 | 81 | 89 |
Thoracolumbar | 32 | 19 | 11 |
When the available English-language literature was recently comprehensively reviewed, it showed no evidence in favor of bracing as altering the natural history of scoliosis.40
Related posts:
7 The Case for Bracing
6 The Importance of the Sagittal Plane: Spinopelvic Considerations
22 Complications in Surgery for Spinal Deformity
28 The Impact of Genetics Research on Adolescent Idiopathic Scoliosis
17 Surgical Treatment of the Right Thoracic Curve Pattern
10 Biomechanics and Reduction of Scoliosis
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
