Experts debate role of Cobb angle in scoliosis bracing

3D orthosis design must factor in many variables, including spinal alignment, for optimal outcomes

Published in the May 2006 issue of BioMechanics

By Jordana Bieze Foster

The Cobb angle, long considered one of the cornerstones of orthotic scoliosis management, provides essential information about the magnitude of a spinal curve. What may be equally important, however, is what the Cobb angle doesn't tell a practitioner about a patient's spine: details of vertebral rotation, sagittal deformities, or overall spinal alignment. In addition, focusing solely on in-brace reduction of Cobb angle can result in orthoses so uncomfortable that patients will not wear them, thus negating any theoretical advantage that might have been gained.

Concerned that practitioners may be over-reliant on Cobb angle, a group of experts convened in March at the annual meeting of the American Academy of Orthotists and Prosthetists to discuss the merits of a more balanced approach-both literally and figuratively-to scoliosis bracing.

Cobb basics

It is nearly impossible to discuss scoliosis without mentioning the Cobb angle, which has been an industry standard for measuring curve magnitude since 1948. Indeed, a Cobb angle of 15 degrees or greater is what constitutes a diagnosis of scoliosis; smaller curves are characterized as postural asymmetry (see "Curve Concepts," September 2004 supplement, page S2). It is defined as the angle formed at the intersection of a line perpendicular to the endplate of the most angulated inferior vertebra and a line perpendicular to the endplate of the most angulated superior vertebra, as seen on a radiograph in the coronal plane.

The Cobb angle's most important function throughout the course of scoliosis bracing is as an outcomes measure. The decrease in Cobb angle from a baseline x-ray to one taken after a brace has been fit to the patient is a quantitative measure of the amount of in-brace correction being applied to the curve in question. Similarly, Cobb angles calculated from subsequent radiographs can chart a patient's progress during brace treatment.

However, it is critical that the practitioner reestablish the most angulated inferior and superior vertebrae each time a new Cobb angle is measured rather than simply using the same vertebral endplates as were used in the initial Cobb measurement, said Thomas H. Colburn, CO, director of orthotics and prosthetics at Tufts New England Medical Center in Boston, who was one of five speakers at the AAOP symposium.

"By using the same endpoints as the initial measure, you might end up with a smaller Cobb angle," said Colburn, who is also director of education for Boston Brace International. "We need to make sure we have standardization."

Predicting progression

Cobb angle has also been used, with varying degrees of success, to predict which scoliotic curves are most likely to benefit from orthotic management. Researchers have shown that the size of a curve at the time of diagnosis directly impacts the likelihood that the curve will progress, with larger curves most likely to progress and least likely to benefit from bracing-although skeletal immaturity and curve location are also contributing factors (see "Ahead of the Curve," June 2002 supplement, page S2).

Research also suggests the amount of curve correction from baseline that can be achieved with a brace is predictive of how effective orthotic treatment ultimately will be, although opinions differ regarding the amount of in-brace correction needed to achieve a positive outcome. Early reports suggested that curves with at least 50% in-brace correction were least likely to progress once orthotic treatment had been discontinued, and the 50% mark has become an industry standard. In a study of larger curves (Cobb angles between 35 degrees and 45 degrees ) treated with Boston Brace thoracolumbosacral orthoses, Dartmouth University researchers found that patients who were fully compliant were those with the highest magnitude (49%) of in-brace correction. That study was published in the Sept. 15, 2000, issue of Spine.

But Donald E. Katz, CO, director of the orthotics department at Texas Scottish Rite Hospital in Dallas, has found in his own research that successful outcomes can still be attained in patients with lesser amounts of initial in-brace correction. In a 1997 study published in Spine, Katz and colleagues found that 40% in-brace correction was needed for single thoracic curves and just 30% in-brace correction was needed for single thoracolumbar, lumbar, and double major-primary thoracic curve patterns. And in a more recent study of curves greater than 35 degrees, Katz and then-fellow A. Atiq Durrani, MD, found that just 25% in-orthosis correction of the primary curve was predictive of a successful outcome; those findings were published in the Nov. 1, 2001, issue of Spine.

"In-brace correction is important but shouldn't be at the expense of brace tolerance," Katz said at the AAOP symposium. "Trying to 'squeeze out' another five to 10 degrees of correction may render the orthosis intolerable."


As valuable as the Cobb measurement may be, scoliosis experts agree that the magnitude of a curve as seen on a lateral radiograph is but one of many variables that need to be factored into orthotic fabrication decisions. For starters, Cobb angle tells the practitioner nothing about vertebral rotation or the curvature of the spine in the sagittal plane.

"Cobb angle is a two-dimensional descriptor of a three-dimensional curve," Colburn said.

Another variable not encompassed by the Cobb measurement is that of vertebral tilt, which can have clinical implications. Looking at double curve patterns in which the thoracic curve was at least 35 degrees, Katz and Durrani found that bracing was unsuccessful in all cases where the superior vertebral endpoint of the thoracic curve was tilted to a greater degree than the vertebral endpoint shared by the upper and lower curves (the "transitional vertebra"). This end-vertebral angle pattern was seen in only four of 31 patients, however-far less frequently than patterns in which the superior vertebral endpoint of the thoracic curve was less tilted than the shared vertebral endpoint, or in which the two endpoints were tilted symmetrically.

"Varying patterns can have the same thoracic Cobb angle and the same vertebral endpoints," Katz said. "Cobb angle doesn't really describe the relationship of vertebral tilts."

The Cobb measurement by itself also doesn't describe the alignment of the spine, which is typically assessed in terms of decompensation (malalignment between the C7 midpoint and the central sacral line that bisects the sacrum at S1 on a lateral radiograph) and apical vertebral translation (distance that the most laterally deviated vertebra in a curve is displaced from a plumb line connecting C7 with the sacrum on a lateral radiograph). And spinal alignment, particularly in-brace alignment, may be even more important to patient outcome than curve correction.

For example, a patient might have an upper thoracic curve and a lower lumbar curve, both on the same side of the central sacral line. Fabricating a brace for that patient based solely on the two Cobb angles without considering spinal alignment, would result in a straighter but laterally angled spine that might prove more problematic for the patient than the original scoliosis.

"With Cobb angle, you know where you're at, but you're not going to know exactly where you're going," said Keith M. Smith, CO, a St. Louis orthotist affiliated with Washington University.

Spinal malalignment can also render Cobb measurements less meaningful. A lateral shift of the upper spine in the direction of a lower curve, for example, will reduce the Cobb angle for that curve even though it is in fact progressing, said J. Martin Carlson, CPO, president of Tamarack Habilitation Technologies in Blaine, MN.

Finding a balance

To quantify spinal alignment, Carlson and others have proposed an "overall balance summation" measure in which lateral deviation in one direction would be assigned a positive value and lateral deviation in the other would be assigned a negative value; an optimally aligned spine would have a score of zero. Notably, spinal curvatures would not preclude such optimal alignment, as long as the spine overall is balanced.

Reliability and sensitivity of this type of summation system have not yet been established, Carlson noted. Nor are its advocates dismissing the importance of the Cobb angle.

"We do not propose that OBS should replace Cobb angle," Carlson said. "They should be dual measurements."

The therapeutic question that arises, then, is to what extent bracing should emphasize curve correction, and to what extent it should emphasize spinal alignment. The final answer remains unclear, but experts agree that both factors need to be considered.

"We need to strive to reduce Cobb angle as much as possible, but we do need to make sure we have a well-balanced spine inside the brace," Colburn said.

To be sure, the medical literature is lacking in research on the relative effectiveness of spinal orthoses for addressing spinal alignment. In a 1983 study published in Spine, Connecticut researchers found no significant improvement in decompensation in 22 patients treated with a Milwaukee brace but did see mild improvement in half the patients. Investigators from the duPont Institute in Wilmington, DE, reported that decompensation after an average of two years was improved in 70% of 71 patients treated with the Wilmington brace in whom baseline decompensation exceeded 1 cm; that study was published in the October 1987 issue of Clinical Orthopaedics and Related Research. In the September 1990 issue of Spine, researchers from Yale University found improved decompensation in two of 11 patients treated with the nocturnal Charleston bending orthosis and further decompensation in three patients. In the May-June 1995 issue of the German publication Zeitschrift fur Orthopadie und ihre Grenzgebiete, investigators from the Universitatsklinik in Bonn reported that decompensation improved in 61.9% of 26 patients treated with the active-derotation Cheneau orthosis to 35% of baseline, whereas in 38.5% of 26 patients treated with a Milwaukee brace decompensation improved to 46.6%.

Poor alignment may not lead to poor outcomes, however. Katz and Durrani analyzed the lumbar pelvic relationship (LPR) in terms of the angle between the iliac crest line and a line tangent to the most inferior vertebra in the lumbar curve, and found that this measure of asymmetry had implications only for double-curve patterns in which the thoracic curve was greater than 35 degrees . In those 24 patients, successful outcomes were seen in 75% of cases in which LPR was less than 12 degrees but only 12.5% of cases in which the LPR was greater than 12 degrees . However, in patients with single or double curves in which the primary curve was lumbar or thoracolumbar, LPR was not predictive of therapeutic outcome despite being greater on average than in the double-curve patterns with primary thoracic curves (18.2 degrees versus 11.8 degrees).

"It is as appropriate to treat the less balanced spine as it is to treat the more balanced spine," Katz said.

Clinical implications

Management strategies to optimize both curve correction and alignment may also differ depending on the type of brace. Andrew J. Mills, CO, managing director of the Spine Corporation in the U.K., suggested at the AAOP symposium that SpineCor's dynamic bracing concept incorporates both elements.

"With dynamic bracing it's not a question of compromising between the two," Mills said. "Initially balance may be worse in the brace, but its dynamic nature allows the body to improve posture as the Cobb angle is reduced."

An additional clinical question yet to be addressed is the issue of compensatory curves created as a result of primary curve correction. Several symposium speakers noted that these compensatory curves may actually improve spinal alignment, which theoretically will benefit the patient. However, as Katz pointed out, studies of the natural history of primary scoliosis curves have shown that double curves are more likely to progress than single curves, thus raising the question of whether-or in which cases-a bracing strategy that results in a double curve will, in fact, be beneficial.

And ultimately, experts agree, bracing biomechanics alone will go only so far toward achieving a positive outcome.

"The most important factor is making an orthosis the patient will actually wear," Katz said.

Jordana Bieze Foster is a freelance writer in Massachusetts and the former editor of BioMechanics magazine.

Copyright: 2006 CMP Media, LLC Copyright 2008 Jordana Foster – 24 Kirkland Dr, Stow, MA – Email: – Fax: (815) 346-5239