Bracing (and screening)--yes or no?

Late-onset (adolescent) idiopathic scoliosis is the commonest spinal deformity seen in our clinics. Like many orthopaedic deformities the degree of the condition determines its importance. When schoolchildren are examined many small inconsequential curves may be seen since it is difficult to keep a straight spine in the face of threedimensional growth for 20 years. The condition becomes progressively less prevalent when assessed by increasing magnitude of the curve. In order to separate potentially troublesome scoliosis from unimportant ‘schooliosis’, the Scoliosis Research Society suggested that in the former the Cobb angle should be more than 10°. To that definition must be added ‘concordant’ rotation in which the posterior elements are directed towards the concavity of the curve; otherwise there is no separation of structural from nonstructural curves. Most patients with late-onset idiopathic scoliosis require no treatment and suffer no functional limitations. In assessment of their management the advice given must be evidence-based. Personal experience is very valuable but it can also be inflexible. Use of the Cobb angle in analysing scoliosis both epidemiologically and clinically has presented problems. The deformity of structural scoliosis is three-dimensional and as the lordotic apex twists so the frontal plane depicts the deformity less well. For example, the problems presented by a curve of 40° are much more than twice as great as in one of 20° and therefore, since the Cobb angle is not linearly proportional to the severity of the deformity, it cannot be handled by simple means and changes in percentage. Surgeons treating scoliosis can intuitively picture what a Cobb angle of, say, 45° represents and still tend to think in terms of this measurement. It can be argued that we should carry on using it since there is not a better means of assessment. However, having an intuitive feeling about the Cobb angle is not sufficient reason for using it as the critical variable in assessing the size of the curve in studies on the natural history or of response to treatment. The Cobb angle is a continuous variable but we tend to categorise it by grouping together, for example, curves of 11° to 15° in 5° batches. Now we have been invited, on the basis of the apparent ability of the Milwaukee brace to prevent progression of a Cobb angle of 6°, to endorse treatment with a brace and to resurrect screening in school. What is ‘magic’ about 6° of progression? The answer is that this is probably the smallest difference in the Cobb angle that can be measured accurately, although reliable detection of this magnitude may not be possible. However, we are surely not saying that an increase in the Cobb angle from 40° to 46° is the same as from 10° to 16° or 100° to 106°! We cannot treat a continuous variable which is nonlinear in a categorical fashion, particularly if this is of the same order of magnitude as the intrinsic error of the measurement. We cannot treat children with potentially serious spinal deformities on the basis of rough guides. Apart from the problems of measuring the severity of the curve, is there evidence to support the positivists in their enthusiasm for treatment with a brace? Is the evidence valid or should the negativists, as we have been gratuitously termed, fight back?