Patient selection

Overall, 739 patients with DRG principal diagnosis of scoliosis were treated from 2016 to 2020. After excluding patients with non-idiopathic and adult scoliosis and incorrect coding, 292 patients with idiopathic scoliosis remained. Finally, 69 patients were included in the research after excluding patients with juvenile scoliosis, patients without brace therapy, and cases with incomplete records (Fig. 1).

Fig. 1Patient characteristics

There were 55 female (79.7%) and 14 male (20.3%) patients. The mean age at first presentation in our hospital was 13.7 ± 1.6 years (boys: 14.8 ± 1.4, girls: 13.5 ± 1.5 years, p = 0.005). The age reported by parents or patients when scoliosis first became apparent was 13.4 ± 1.7 years (boys: 14.3 ± 1.8 years, girls: 13.1 ± 1.6 years, p = 0.027). Age at brace initiation was 14.1 ± 1.5 years (boys: 15.1 ± 1.5 years, girls: 13.9 ± 1.4 years, p = 0.008). Age at menarche of included girls was 12.8 ± 1.2 years, and the period between this age and brace initiation was 1.1 ± 1.7 years (Table 1). The distribution of Risser stages are shown in Table 1 and Fig. 2. Period between first presentation and bracing start was 0.40 ± 0.65 years.

Table 1 Patient and curve characteristics of all included patientsFig. 2

Distribution of Risser stages at brace initiation

Curve characteristics before bracing

In accordance with the Ponseti curve patterns, the distribution of curves across both genders were as follows: combined 30.4%, thoracic 29%, lumbar 24.6%, and thoracolumbar 15.9%. The ‘thoracic right/lumbar left’ curve was the most frequent curve direction (29%), followed by the ‘thoracic right’ with 27.5%, ‘lumbar left’ with 18.8%, and ‘thoracolumbar left’ with 10.1%.

The Cobb angle at the initial presentation of patients was 29.0° ± 9.8° and at bracing start was 30.0° ± 9.2°. In 22 patients (31.9%), the Cobb angle was < 25°, in 36 patients (52.2%) between 25° and 40°, and in 11 patients (15.9%) > 40° (Fig. 3). Curve rotation according to Nash and Moe was Grade 2 in 35 patients (50.7%), Grade 1 in 25 (36.2%), and Grade 3 in nine patients (13.0%) (Table 1).

Fig. 3

Frequency of initial Cobb angles according to SRS criteria

Patients and curve characteristics during bracing

The period of bracing was 2.6 ± 1.5 years and age at brace termination was 16.7 ± 1.3 years (girls: 16.6 ± 1.2 years, boys: 17.4 ± 1.3 years, p = 0.033). Bracing was terminated 3.7 ± 1.2 years after menarche. The RBW did not differ between groups (Fig. 4). The Cobb angle in the first brace after padding was significantly reduced to the initial Cobb angle (20.0° ± 11.8° vs 30.0° ± 9.2°, p < 0.001). Percentual in-brace-correction was 36.4% ± 31.8%. Cobb angle after brace termination was 29.9° ± 11.9°, which means there was no significant change to the Cobb angle before brace therapy (p = 0.903) (Fig. 5, Table 1).

Fig. 4Fig. 5

Cobb angle before, during and after bracing

Outcome and determinants according to subgroupsAll included patients (Supplement Tables 1–3)

After brace weaning related to all 69 included patients, surgery was recommended in 14 cases (20.3%), a curve progression of ≥ 6° was observed in 16 cases (23.2%), a Cobb angle beyond 45° was found in eight cases (11.6%), and a Cobb angle improvement of ≥ 6° was recorded in 14 patients (20.3%) (Fig. 6).

Fig. 6

Outcome after bracing (all included patients)

Patients with later indication for surgery had a significantly higher Cobb angle at initial presentation (38.1° ± 9.7° vs 26.6° ± 8.4°, p < 0.001) and at the beginning of bracing (38.8° ± 9.8° vs 27.8° ± 7.6°, p < 0.001). Furthermore, they had lower Risser stages at the start of bracing (p = 0.010) and higher degrees of rotation according to Nash and Moe (p = 0.030).

Patients with Cobb angle progression of ≥ 6° at the end of bracing therapy were younger when the curve was initially noticed (12.4 ± 1.5 vs 13.7 ± 1.7 years, p = 0.011) and older when menarche occurred (13.4 ± 1.1 vs 12.6 ± 1.2 years, p = 0.037). The Risser sign was also shifted to smaller degrees (p = 0.015). There was a significantly diminished probability of curve progress ≥ 6° in the Risser 4 stage compared to Risser 0 (p = 0.018; Fig. 7). In the Risser 3 stage, there was no significantly reduced probability of curve progression compared to Risser 0. Nevertheless, 12.3% of all included patients still experienced curve progression ≥ 6° despite Risser 3 or 4 at brace initiation. This corresponds to 21.6% of all Risser 3 and 4 patients.

Fig. 7

Curve progression depending on the Risser stage

Cobb angle progression beyond 45° at the end of brace therapy was noted with higher initial Cobb angle values (45.0° ± 6.2° vs 28.0° ± 7.5°, p < 0.001) and higher initial Nash and Moe grades (p = 0.002).

Cobb angle reduction in brace (%) was the only significant different predictive variable comparing patients with Cobb angle improvement ≥ 6° vs < 6° (54.0% ± 31.2% vs 31.9% ± 30.7%; p = 0.019; Supplement Table 2).

RBW did not have a significant effect on the outcome parameters.

Detailed results of additionally conducted univariate binary logistic regression for significant different predictive variables are shown in Supplement Table 3.

Subgroup 1 (Supplement Tables 4 and 5)

Only nine (13.0%) of the included patients met the SRS research criteria. In three of those nine patients (33.3%), surgery was finally recommended, and curve progression was ≥ 6° in three patients (33.3%). There were no patients with curve progression beyond 45°. Cobb angle improvement ≥ 6° was observed in one patient (11.1%).

In the Mann–Whitney U test, the Cobb angle at brace initiation and initial presentation was significantly higher (36.7° ± 3.1° vs 28.8° ± 3.9°, p = 0.038 and p = 0.020, respectively) in patients for whom surgery later had to be recommended.

Patients with Cobb angle progression ≥ 6° at brace termination showed a significant difference of Nash and Moe grade compared with those with ≤ 5° progression (p = 0.048 in Fisher’s exact test).

Subgroup 2 (Supplement Table 6)

Twenty-two patients of the study population had initial Cobb degrees < 25° at brace initiation. In two patients (9.1%), surgery was finally recommended, curve progression was ≥ 6° in five patients (22.7%), and there was a curve improvement ≥ 6° in five patients (22.7%). There were no patients with curve progression beyond 45°.

The age of patients at brace initiation was significantly younger (11.7 ± 0.6 vs 14.3 ± 1.2 years; p = 0.017) in patients for whom surgery was recommended within the course.

There were no significantly different predictor variables in patients with final Cobb angle progress or improvement ≥ 6°.

Subgroup 3 (Supplement Table 7)

Eleven of our patients had Cobb angles > 40° at brace initiation. In seven patients (63.6%), surgery was finally recommended, and curve progression was ≥ 6° in four patients (36.4%). Of the six patients in this group in whom the Cobb angle was still in the range of 41°–44° at the start of bracing, three showed a Cobb angle progression beyond 45° at brace termination. Two patients (18.2%) had a Cobb angle improvement ≥ 6° over bracing time.

There were no statistical differences in the predictive parameters in patients for whom surgery later had to be recommended.

Patients with Cobb angle progression ≥ 6° at brace termination were significantly older at menarche (13.7 ± 1.0 vs 11.2 ± 1.1 years, p = 0.036) and showed a more frequent curve pattern favouring the thoracic and thoracolumbar curves (p = 0.036).

Subgroup outcome comparison (Table 2)Table 2 Number of patients with the corresponding outcomes after brace weaning

In all subgroups, RBW did not have a significant effect on the outcome parameters.

The comparison of the individual subgroups with the entire group of all patients and the comparison among the subgroups did not reveal any significant differences in the behaviour of curve progression < 6°/ ≥ 6° and curve improvement. Surgery was recommended more frequently to patients in the subgroup with an initial Cobb angle > 40° compared to all patients (p = 0.006) and compared to the subgroup with initial Cobb degree < 25° (p = 0.002).

Compared to all other subgroups and to the all-patients group, Cobb degree progression beyond 45° was significantly more frequent in the subgroup with initial Cobb angles > 40° (41°–44°; p = 0.006–p = 0.044).