Case report: 39° curve down to 5.8° after about 29 months of bracing

Scoliosis may be a single (C) curve or a double (S) curve. This case has one lumbar curve with its apex at L1 and a Cobb angle of 38.65°. The range 25°–40° is where bracing has the best evidence in a child who is still growing: the BrAIST trial (NEJM 2013) of 242 children with a Cobb angle of 20°–40° recorded 72% success in the braced group versus 48% with observation alone, and the 2016 SOSORT guidelines recommend bracing for growing children with a curve above about 20° and a risk of progression.

According to the clinic, the patient was 15 and at Risser 3 at the first visit, meaning the skeleton was partly ossified with a limited amount of growth remaining. Risser 3 is within SOSORT’s indication for bracing (Risser 0–3) but outside the range of the BrAIST trial (Risser 0–2), so BrAIST’s figures do not apply directly to this patient. The higher the Risser grade, the less growth remains and the lower the risk of progression usually is compared with Risser 0.

Two X-ray views and the back photo were taken at the same time. The left film is flipped to look as if seen from behind, the right film is seen from the front. The orange lines are the Cobb-angle measurement lines and the red boxes give each measurement. Result: 38.65°.

The back photo shows a shifted trunk and uneven waistlines, matching the lumbar curve on the film. Surface photos only illustrate; they do not replace X-rays.

This is an important figure for predicting how well the brace works, because it was measured at the start of wear, at the same time point and in the same unit as the studies of in-brace correction. The clinic took two films when the patient first wore the brace: one in the day brace and one in the night brace, to be compared with the initial out-of-brace film (38.65°). On each film the software records two angles: day brace 7.9° and 9.8°; night brace 2.8° and 0.8°. Following the clinic’s guidance, the lumbar curve is taken as 9.8° (day brace) and 0.8° (night brace).

How to read: in-brace correction is the percentage by which the curve is reduced while wearing the brace compared with not wearing it. The night brace corrected about 9° more than the day brace (9.8° versus 0.8°).

On how the reference film is obtained: in practice there is no other way. Nobody orders two X-ray examinations days apart (one on arrival and one when the brace is delivered), because of unnecessary radiation exposure; the professional rule is a minimum of 3 months between films. The pre-treatment film (2 August 2022) is therefore the “out of brace” film, and the film at first fitting (18 December 2022, about 4.5 months later) is the “in brace” film. This is also the usual meaning of “initial in-brace correction” (the first in-brace film compared with the pre-treatment film). In the study abstracts we read (Emans, Lou, Brown, Bonsignore-Opp, Davis, D’Amato), none states the interval between the two films, so we cannot say how far apart they were taken. This case’s comparison uses that same usual approach, bearing in mind that the curve may have changed slightly between the two films. Bracing lasts several years, so the interval does not remove the value of the comparison.

Compared with the research: in the Boston-brace study of Emans et al. (Spine 1986; 295 patients) the best in-brace correction averaged about 50%, measured when the brace was started; Brown et al. (J Pediatr Orthop 2026; 50 children, Rigo Chêneau braces) recorded 55 ± 20%. These are the same kind of figure at the same time point, so they can be compared: this case is above the average. Details in section 6.

More than two years later (30 December 2024, the patient aged 17) the day brace had been weaned and only the night brace remained. The final films are therefore only two: out of brace and in the night brace. Out of brace the measured angle is 5.8° (other angle: 2.4°); in the night brace −4.5° (other angle: 2.1°), meaning the curve is pushed past straight to the opposite side.

Compared with the first film: out of brace the Cobb angle fell from 38.65° to 5.8°, about 85% (32.9°). By convention an angle below 10° is not regarded as scoliosis, and the measurement error of a Cobb angle (about ±5°) is larger than the value left. That the night brace still over-corrects (−4.5°) is regarded by the clinic as a very good prognostic sign.

The clinic’s protocol: wear the brace until skeletal maturity is complete and the curve no longer responds to corrective force, then about a year more to be safe; after that the day brace is weaned first, then the night brace is reduced gradually. At the final film the patient had passed the day-brace weaning step and was at the night-brace stage. Complete night-brace weaning and further follow-up are needed for a final conclusion on the criterion of avoiding surgery; at 5.8°, far below the 50° threshold, this case is currently very far from the surgical threshold.

We searched PubMed for the average in-brace correction. We found no formal meta-analysis giving one overall average: the systematic review of Peeters et al. (Spine 2022; 28 studies) only synthesised predictive factors (curve flexibility is strongly linked to in-brace correction; CAD/CAM and plaster-cast braces do not differ clearly in correction), and the meta-analysis of Zheng et al. (Children 2023) reports only differences in degrees. We therefore compiled individual studies.

Arithmetic: for the five full-time-brace groups with a reported mean (Emans, Lou, Brown and the two groups of Bonsignore-Opp; 474 patients in all), the sample-size-weighted mean is about 47%. Individual study means range from 22% to 55% for full-time braces, up to 73% for thoracolumbar curves and 96–98% for the Providence night brace measured on supine films (supine films usually give a higher correction than standing films). This is our own arithmetic, not a formal meta-analysis.

Interpretation: the sample case’s day-brace correction (75%) is about 28 percentage points above the weighted mean, but close to the mean for thoracolumbar curves in Davis’s study (73%), because lumbar curves respond better than thoracic ones. The night-brace correction (98%) matches what is reported for the Providence night brace (96–98%, although on supine films).

The final result, in the same unit (degrees of improvement): in the Boston study of Emans et al., at least a year after brace wear stopped, 49% of patients had a curve within ±5° of the start, 39% improved by 5°–15° and only 4% improved by more than 15°; the average out-of-brace correction was only about 23% at the start of weaning and 11% at the last follow-up. In Brown et al. the mean final major curve was 32 ± 12° (starting curves 20°–40°). This case improved by 32.9° (about 85%) out of the brace, which falls in the rare group in those studies. Note that the Boston study is old, covers many curve patterns, and the sample case is still wearing a night brace.

On accompanying outcomes: in D’Amato et al., the share with no progression beyond 5° was 93% for thoracolumbar and 94% for lumbar curves (all 102 children: 74%); in Brown et al., the group with correction of 45% or more had an 88% success rate. These rates use different criteria and cannot be compared directly with BrAIST’s rate of avoiding surgery.

Surface photos help parents see the change in trunk posture over time. The photos are kept as in the clinic’s original, including the notes written on the body.

Before comparing, three different kinds of number must be kept apart, because they measure different things and cannot be compared directly: (1) a trial success rate is the share of patients with a good outcome, e.g. in BrAIST a Cobb angle still below 50° at skeletal maturity (i.e. below the surgical threshold); (2) in-brace correction is the percentage by which the curve is reduced while wearing the brace; (3) the 85% of this case is the percentage reduction of one patient’s Cobb angle (38.65° → 5.8°).

These studies show associations at group level; they do not prove that those factors produced the result in this one patient. No study allows an exact success rate to be calculated for this case; an individual estimate can be made with the BrAIST-Calc tool (Dolan et al., Spine 2024), which needs age, sex, BMI, Risser stage, Cobb angle and average daily wear hours. For this case age (15), sex (female), Risser (3), Cobb angle (38.65°) and wear hours (about 22 hours/day) are known; only BMI is missing.