Table 2 Summary of the main publications including data on the effect on pain of different treatments presented in the order of citation in the manuscript

 S Negrini/Cochrane Database Syst Rev 2015 [39]

Review including 7 studies (5 initially planned as RCTs and 2 as prospective controlled trails) with a total of 662 adolescents, comparing braces with other treatment.

PedsQL (only 1 item in one of the 4 dimensions focus specifically on pain).

Back pain did not change in the long term (very low quality evidence)

The authors highlight that it was not possible to perform a meta-analysis due to important clinical differences among studies.

 T Maruyama/Physiother Theory Pract 2011 [40]

Systematic review including 20 studies: 2 controlled clinical trials and 18 case-control studies. No RCTs found.

Child Health Questionnaire (1 study) and VAS (1 study)

Compared with observation bracing does not influence back pain or HRQoL. Conflicting evidence reported for studies comparing bracing with other forms of treatment

 ME Alves de Araujo/J Bodyw Mov Ther 2012 [49]

RCT comparing Pilates-based therapy (N = 20) to weekly meetings with no therapeutic intervention (N = 11). Age ranged from 18 to 25 years

Borg CR 10

Pain decreased from 5.3(1.5) to 1.8 (1.9) (P = 0.0001) in the experimental group and from 4.4(2.3) to 3.8(2.7) in the control group (NS)

All patients were female students with minor non-structural dorso-lumbar scoliosis (Cobb angles 7.6(3.5) (Experimental) and 7.1(2.8) (Control) respectively.

 M Monticone/Eur Spine J 2014 [51]

RCT comparing self-correction, task-oriented spinal exercises and education (N = 55) with traditional spinal exercises (N = 55). Evaluations pre-, post-treatment, and at 1-year follow-up. Mean age at baseline 12.5 and 12.4 years respectively (NS)

SRS-22

Pain scores at the 3 evaluations were: 3.8 (0.4), 4.6 (0.3), and 4.7 (0.2) in the experimental program and 3.9 (0.5), 4.3 (0.3), and 4.2 (0.4) in the control group. P < 0.001 for time, group, and interaction effects.

 DK Noh/J Back Musculoskelet Rehabil 2014 [52]

Comparison of a corrective spinal technique (CST) approach with a conventional exercise (CE) program. Two groups of 16 AIS patients each. Mean age 13.8 years Versus 14.9 (N.S.)

SRS-22

Pain scores improved from baseline to follow-up in both groups. Results were from 4.5 to 4.9 in CST group and from 3.8 to 4.6 in CE group.

The 2 groups were already small at baseline and the attrition rate was substantial with N = 8 (CST) and 4 (CE) at follow-up.

 KA Zapata/Pediatr Phys Ther 2015 [53]

Randomized trial comparing 8 weeks of weekly supervised spinal stabilization exercises (N = 17) with 1-time treatment (N = 17). Patients with AIS aged 10–17 years

Numeric Pain Rating Scale (NPRS)

Oswestry

Patient-specific functional scale (PSFS)

NPRS scores improved from 5.4(1.5) to 1.5(1.8) in the supervised group and from 5.4(1.3) to 3.4(1.7) in controls (P ≤ 0.05 & > MCID)

The reduction in OSW was similar in both groups and the improvement in PSFS was higher in the supervised group but the difference between the 2 groups < MCID in adults.

All the patients included in the trial had low back pain at baseline.

 M Tavernaro/Scoliosis 2012 [54]

Cross-sectional study followed by retrospective case-control study to verify the impact of a complete rehabilitation team in adolescent patients with bracing (N = 28 AIS and 10 hyperkyphosis). Mean age 15.8 years

SRS-22

Pain scores were 3.93 ± 0.55 among those treated by the team and 3.54 ± 0.83 in those who were not (N.S.)

Other variables were significantly improved by the team management.

 EM Bunge/Eur Spine J 2007 [63]

Cross-sectional evaluation of HRQoL of AIS patients after completing treatment. Patients had been braced (N = 36), treated by brace and surgery (31), or only by surgery (30). Mean age 16.3 years.

SRS-22

Mean scores for Pain domain in the 3 groups were:

Brace: 4.5 (0.57)

Brace & surg: 4.1 (0.90)

Only surgery: 4.1 (0.71)

Pain was not significantly correlated with the time span between completing treatment and filling out the questionnaire.

 PR Rushton/Spine 2013 [68]

Review of the literature and statistical analysis evaluating the effect of surgery on HRQoL of adolescents with AIS and 2 years follow-up.

SRS-24

81 % of the included cohorts reported a statistically significant improvement of pain. The reduction was above MCID only in 1/12 cohorts

 MC Hawes/Disabil Rehabil 2008 [56]

Systematic review of studies on surgery for AIS with ≥ 10 patients and followed for ≥2 years after surgery. 82 articles (5780 patients) included.

Trials’ results presented as “pain-free”: Yes, No, or Not tested/reported. Yes meant that most or all patients reported minimal or no pain and none reported severe, chronic, or increased pain post-op

The authors conclude that there is no evidence to support that the result in terms of reduced magnitude of the spinal curve is correlated with reduced pain.

65 % of articles did not include pain in their outcome. Of those who did, definitions were quite different

 Z Landman/Spine 2011 [57]

Multicenter, prospective, consecutive clinical series.

N = 1433 patients.

Changes in pain assessed in 295 patients with complete data and 2 years follow-up

SRS-22 (Z-test for proportions used to analyze preop and postop differences)

Mild to severe pain within the last month in 73.2 % preop, 53.6 % at 1 year and 53.2 % at 2 years. Pain at rest in 70.5 % preop, 56.9 % at 1 year and 60 % at 2 years.

Compared with preoperative data, at 2 years 40 % of patients reported a decrease in pain, 38.6 % no change, and 20.3 % an increase in pain.

A significant disagreement between preoperative pain reported by physicians (44 %) and patients (77.9 %) was found.

Patients who were more satisfied with their appearance reported less pain.

 DW Roberts/Spine 2011 [23]

Longitudinal multicenter cohort study to compare functional outcomes between genders before and after surgery. N = 744. Mean age was 15.2 years for males and 14 years for females.

SRS-30

Postop pain improved significantly in both genders from 4.1 to 4.3 in girls and from 4.3 to 4.5 in boys. The difference between genders is NS.

The pain reduction at 2 years does not seem clinically meaningful.

 T Akazawa/Spine 2012 [58]

Case-control study to compare healthy subjects (N = 80) with idiopathic (N = 80) and non-idiopathic (N = 56) scoliosis patients and ≥ 21 years after surgery.

SRS-22

Roland-Morris (RDQ)

Pain dimension scores were 4.3(0.6) in AIS patients and 4.2(0.5) in controls (NS)

RDQ scores were 2.4(4.1) and 1.4(3.1) respectively.

At long-term postoperative follow-up AIS patients had similar pain scores as healthy controls

 TP Bastrom/Spine 2013 [60]

Review of a prospective multicenter database registry. N = 584 AIS patients treated surgically with ≥ 2 years follow-up. Age at surgery 14.7 ± 2 years

SRS-22

Self-reported pain Vs pain free (1^st 6 months post-op or 6–24 months post-op). The focus is on the patients with unexplained pain >6 months postoperative

The 2-year pain scores were 4.1 ± 0.7 in patients with postoperative pain and 4.5 ± 0.6 in the postop pain free group (P < 0.001). These 2 groups were also significantly different in terms of preoperative pain (3.8 ± 0.8 Vs 4.1 ± 0.7).

 CB Sieberg/J Pain 2013 [61]

Prospective multicenter registry examining postoperative outcomes of AIS patients.

N = 169 at baseline, 1 and 2 years postop and N = 69 for 5 years follow-up.

SRS-30

Moderate to severe pain in the past month was reported preop by 35 % of patients. The figures were 11 % at 1 year, 15 % at 2 years, and 15 % at 5-years post-surgery.

Pain often to very often at rest was reported by 43 % at baseline, 5 % at 1- and 2-years follow-up and 8 % at final follow-up.

Examining the evolution of pain from preoperatively to 5 years follow-up the authors describe 5 trajectories distinguishable on preop age, mental health, and self-image.

 LY Carreon/Spine 2011 [62]

Longitudinal cohort of AIS patients treated surgically and evaluated preoperatively and 2 years postop. N = 745. Mean age 14.2 years

SRS-22 and SRS satisfaction

Mean pain domain scores improved from 4.16 ± 0.71 preop to 4.31 ± 0.72 (P < 0.0001)

Pain and satisfaction domains were significantly correlated (Spearman 0.260)

The pain reduction does not seem clinically relevant

 EM Bunge/Eur Spine J 2007 ^a [63]

Cross-sectional evaluation of HRQoL of AIS patients after completing treatment. Patients had been braced (N = 36), treated by brace and surgery (31), or only by surgery (30). Mean age 16.3 years.

SRS-22

Mean scores for Pain domain in the 3 groups were:

Brace: 4.5 (0.57)

Brace & surg: 4.1 (0.90)

Only surgery: 4.1 (0.71)

A subset of patients’ satisfaction scores was compared with that of their surgeons. No significant differences were reported

 MO Andersen/Spine 2006 [64]

Longitudinal study of consecutive AIS patients (N = 181) treated by brace (BT = 82) or surgery (ST = 99). Follow-up 9.7 years

VAS (6 items)

SF-36

Among the 6 VAS scores only “Do you feel leg pain right now” was significantly different with 0.5 (0.2–0.9) in the BT group and 0.2 (0.0–0.3) in the ST group (P = 0.034).

The results for the mean Bodily pain dimension of the SF-36 were 74.6 (BT) and 71.4 (ST) which are significantly lower than the mean of 79.8 of a 408 Danish control cohort.

The difference in pain does not seem clinically meaningful

 B Poitras/Spine 1994 [65]

Retrospective cohort study of patients referred to a single center for AIS compared with a sample (N = 1755) of the general population.

Among the 723 patients treated surgically, 555 completed the questionnaire. Follow-up 10–30 years.

Pain assessed by Postal questionnaire (taken from McGill Pain Questionnaire, Oswestry and Roland Morris)

Controls evaluated by telephone survey.

Back pain in the past year was reported by almost ¾ of the responders. For controls the same figure was 56 %.

The authors found no correlations with several surgical technical variables evaluated (curve correction, number of vertebrae fused, distal level of fusion).

 PR Rushton/Spine 2013 [55]

Review of the literature and statistical analysis to compare untreated adolescents with AIS with normal controls.

SRS-22r

Of the included cohorts 81 % reported statistically significant worse pain among untreated scoliotic patients but the difference was clinically important only in 5 % of cohorts.

 VM Pham/Ann Readapt Med Phys 2008 [69]

Comparison of 3 groups of patients with AIS 32 without brace, 41 treated full-time with a corset and 35 wearing the brace during the night only. The 3 groups were significantly different in age, Risser, Cobb angle, etc.

VAS

Quality of Life Profile for Spine Deformities

No significant differences in the intensity of pain (VAS) were found

 OF Ugwonali/Spine J 2004 [70]

Cross-sectional questionnaire-based study. Patients with AIS were braced (N = 78) or observed (N = 136)

Mean age was respectively 13.6 Versus 13.8 years (N.S.)

Child Health Questionnaire (CHQ Parental form-28) and PODCI

No differences in pain domains were found between the 2 groups despite the fact that Cobb angles were significantly bigger in the braced group. Regression analysis showed a significant association of the CHQ bodily pain dimension and age but not with gender, cobb angle or treatment.

No information gathered from the adolescents themselves is included in the study.

The authors conclude that bracing does not decrease QoL