Prediction of Curve Progression in Idiopathic Scoliosis

        Progression of idiopathic scoliosis correlates with skeletal growth, peaks during the adolescent growth spurt, and usually stabilizes or slows at skeletal maturity. 1 The predictors of skeletal maturity guide the treatment of idiopathic scoliosis by envisaging the risk and timing of curve progression. The maturity indicators that closely correlate with scoliosis progression include chronological age, height, and weight changes; development of secondary sexual characteristics; and menarche; however, skeletal maturity assessment is a more sensitive index of both growth velocity and the cessation of growth. 2 Numerous scoring systems have been developed and used to evaluate skeletal maturity in patients with scoliosis, including the Risser sign, the Oxford stage, and the Greulich and Pyle and Tanner-Whitehouse III skeletal maturity assessments. 3–6 Because of its simplicity, the Greulich and Pyle system 4 is widely used in clinical practice to estimate skeletal age. However, wide spacing of age reduces its sensitivity, particularly during peak height velocity (PHV). 2

        The TannerWhitehouse III system, 5 although a reliable predictor of skeletal maturity, may be too complex and time consuming for routine clinical use. 2,7,8 Cobb angles of 35 ° or greater progressed. Similarly, all patients with initial Cobb angles of 40 ° or greater progressed except those in SS7. Conversely, none of the patients with initial Cobb angles of 15 ° or less or those in SS5, SS6, and SS7 with initial Cobb angles of 30 ° or less progressed. Predictive progression of 67%, 50%, 43%, 27%, and 60% was observed for subgroups SS1/30 °, SS2/20 °, SS3/30 °, SS4/30 °, and SS6/35 ° respectively. Conclusion. This larger cohort shows a strong predictive correlation between SS and initial Cobb angle for probability of curve progression in idiopathic scoliosis. Key words: idiopathic scoliosis , Sanders stage , Risser stage , Cobb angle , curve progression , Lenke classifi cation , skeletal maturity , skeletal age , chronological age , hand fi lms . Level of Evidence: 3 Spine 2015;40:1006–1013 Prediction of Curve Progression in Idiopathic Scoliosis Validation of the Sanders Skeletal Maturity Staging System Prakash Sitoula , MBBS, MS (Ortho) , * Kushagra Verma , MD , † Laurens Holmes Jr , PhD, DrPH , * Peter G. Gabos , MD , * † James O. Sanders , MD , ‡ Petya Yorgova , MS , * Geraldine Neiss , PhD , * Kenneth Rogers , PhD, ATC , * and Suken A. Shah , MD * † Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. SPINE131133_LR 1006 PINE131133_LR 1006 28/05/15 7:47 PM 8/05/15 7:47 PM DEFORMITY Validation of Sanders Skeletal Maturity Staging for Scoliosis Progression • Sitoula et al Spine www.spinejournal.com 1007 Sanders et al9 developed the simplifi ed skeletal maturity staging system from the Tanner-Whitehouse III and Greulich and Pyle systems.

        This system uses hand and wrist radiographs to establish the skeletal maturity stage, which is then correlated with scoliosis progression based on skeletal maturity and curve magnitude in a predictive model. 9 Although this was a novel approach based on prospectively collected data, the study was limited by the small number of patients and exclusion of males. Our present study aimed to validate the fi ndings of Sanders et al9 with a larger cohort, including boys, and to assess the correlation between curve magnitude/skeletal maturity and curve progression in idiopathic scoliosis. MATERIALS AND METHODS Patient Selection After institutional review board approval, we retrospectively identifi ed 1100 consecutive patients with idiopathic scoliosis treated between 2005 and 2011. Girls aged 8 to 14 years ( <2 yr postmenarche) and boys aged 10 to 16 years who had obtained at least 1 hand and spine radiograph on the same day for evaluation of skeletal age and scoliosis curve magnitude were followed to skeletal maturity (Risser stage 5 or fully capped Risser stage 4), curve progression to 50 ° or greater, or spinal fusion.

         Exclusion criteria were nonidiopathic scoliosis, follow-up of less than 12 months after obtaining the initial hand radiographs, girls older than 2 years postmenarche, skeletal maturity at initial evaluation (Risser stages 4 and 5), skeletal dysplasia, syndromic scoliosis, or previous spinal fusion. The cohort of 161 patients met the inclusion criteria. Data Collection From the patient record, the following demographic data were collected at initial and fi nal follow-up: age, height, weight, family history, duration of menarche (if applicable), bracing, and duration of follow-up. Radiographical data gathered included curve magnitude and location, skeletal maturity as stated by Sanders et al, 9 and Risser stage. Spinal Radiographs All spinal radiographs were standing posteroanterior view (with patient out of the brace, if applicable) on full-length (36-in.) fi lm. The pelvis was included in the fi lm. Curve Pattern Classifi cation A modifi ed Lenke classifi cation system, previously described by Sanders et al, 2 was used to classify the curves at initial and fi nal follow-up visits because the original Lenke classifi cation for scoliosis 10 is intended for surgical curve classifi cation. Six curve types were used: type 1 (main thoracic), type 2 (double thoracic), type 3 (double major; thoracic curve > lumbar curve), type 4 (triple major), type 5 (thoracolumbar/lumbar), and type 6 (double major; thoracolumbar or lumbar curve > thoracic curve).

        Curve Measurement At initial and final follow-up, curve magnitudes were determined according to the Cobb method 11 for proximal thoracic, main thoracic, thoracolumbar, and lumbar curves. To reduce interobserver variability, a single observer (P.S.) performed all radiographical measurements. Levels selected for Cobb angle measurements were similar for initial and fi nal follow-up unless the curve pattern changed. Skeletal Maturity Assessment The simplifi ed skeletal maturity staging system was used on the basis of the posteroanterior radiograph of the left hand and wrist to classify skeletal maturity in 8 stages. Hereafter, this staging system and the stages will be termed as the Sanders Skeletal Maturity Staging System and the Sanders stage (SS), respectively. This system is based on progressive growth and subsequent fusion of epiphyses of small long bones of hand. Few descriptive terms need to be understood for this. Digital epiphysis is termed as “covered” when it is as wide as the metaphysis. For metacarpal heads to be called “covered,” the palmar and dorsal surfaces should be well defi ned. The epiphyseal bend over metaphysis is termed as the “cap.” In SS1 (juvenile slow), digital epiphyses are not covered.

        In SS2 (preadolescent slow), all digital epiphyses are covered. SS3 (adolescent rapid-–early) is characterized by majority of digits being capped. In addition, the epiphyses of second through fi fth metacarpals are wider than their metaphyses. Beginning of closure of distal phalangeal physes heralds SS4 (adolescent rapid-–late). In SS5 (adolescent steady—early), all the distal phalangeal physes are closed while all the remaining physes are open. In SS6 (adolescent steady-–late), the proximal and middle phalangeal physes are closing. SS7 (early mature) is characterized by closure of all phalangeal physes while the distal radius physis remains open. SS8 (mature) is reached when there is complete closure of distal radius physis. A single observer (P.S.) reviewed all hand fi lms. Statistical Analysis Summary statistics were performed for both continuous (mean and SD) and categorical (frequency and percentage) variables. The χ2 test was used to show distribution of curves within the SS as well as curve progression. From the coeffi cient of logistic regression model, we estimated the predictive probability, using the margin model, on the basis of the results of the present cohort to determine whether or not a curve would progress.

        This was done by dichotomizing the curve magnitude at final follow-up into less than 50 ° and 50 ° or greater. Exact logistic regression and the margin probability model were used to predict curve progression given initial curve magnitude and SS. Stability of prediction was tested with 95% confi dence interval and P value. All tests were 2-tailed, and the signifi cance level was less than 0.05. Statistical Package for Social Sciences version 17.0 (Chicago, IL) and STATA version 12.0 (STATACorp, College Station, TX) were used for data analysis. Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. SPINE131133_LR 1007 PINE131133_LR 1007 28/05/15 7:47 PM 8/05/15 7:47 PM DEFORMITY Validation of Sanders Skeletal Maturity Staging for Scoliosis Progression • Sitoula et al 1008 www.spinejournal.com July 2015 RESULTS Patients There were 161 patients: 131 girls (81%) and 30 boys (19%). The mean age for girls was 12.3 ± 1.2 years, and the mean age for boys was 13.9 ± 1.1 years.

         The mean follow-up was 2.4 ± 1.2 years. During this time, there was a substantial increase in weight (from a mean of 47 ± 11 kg to 57 ± 12 kg) and height (from a mean of 156 ± 10 cm to 163 ± 9 cm). At initial evaluation, 68% of the females were postmenarchal, with a mean duration from initial menarche of 6 ± 5.4 months. Family history was positive in 70 patients (44%), negative in 89 patients (55%), and unknown in 2 patients (1%). Eighty-two patients (51%) were braced. Curve Patterns At fi nal follow-up, distribution of curves according to modifi ed Lenke types was as follows: type 1, 26 patients (16%); type 2, 12 patients (7%); type 3, 63 patients (39%); type 4, 5 patients (3%); type 5, 38 patients (24%); and type 6, 17 patients (11%). Nine patients changed from 1 curve pattern to another: type 1 to type 2, 1 patient; type 1 to type 3, 3 patients; type 3 to type 1, 1 patient; type 3 to type 6, 1 patient; type 6 to type 1, 1 patient; and type 6 to type 3, 2 patients. Radiographs At the initial point of evaluation, the majority of the patients were in Risser stage 0 (n = 84, 52%), followed by stage 3 (n = 27, 17%), stage 2 (n = 26, 16%), and stage 1 (n = 24, 15%).

         According to the SS, most patients were in SS4 (28%), followed by SS3 (25.5%), and the least patients were in SS7 (1.2%) ( Table 1 ). Final Outcome: Predictive Probability for Curve Progression The curve progression at fi nal follow-up was stratifi ed by Cobb angle of less than 50 ° (64% patients) and 50 ° or greater (36% patients) ( Table 1 ). The 50 ° curve magnitude was selected because it often represents a lower limit for surgery. Table 2depicts the univariate analysis for curve progression to 50 ° or greater, which showed statistically signifi cant correlation for sex ( P = 0.03), brace ( P = 0.0001), Risser stage ( P = 0.005), SS ( P < 0.0001), and initial curve magnitude ( P < 0.0001). In contrast, positive family history, menarchal status, and distribution of curves by modifi ed Lenke types were found not to have, or to have a weak association to, curve progression ( P > 0.05). As expected, an increase in curve progression was observed with higher initial curve magnitude ( Table 2 ). Table 3shows the distribution of patients and percentage progression to 50 ° or greater for each curve magnitude/SS subgroup ( Figures 1 and 2 ). The 95% confi dence intervals and P values are listed for each group. In SS2, all patients with initial Cobb angles of 25 ° or greater progressed, and in SS1 and SS3, all patients with initial Cobb angles of 35 ° or greater progressed. Similarly, all patients with initial Cobb angles of 40 ° or greater progressed, except for those in SS7 ( Table 3 ). Conversely, none of the patients in the present series in SS5, SS6, and SS7 with initial Cobb angles of 30 ° or less progressed. In SS1, 50% of patients with an initial Cobb angle of 25 ° progressed ( P = 0.2), and 67% of the patients with an initial Cobb angle of 30 ° progressed ( P = 0.01). In SS2, 50% Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

        TABLE 1. Demographics Variable Frequency Percentage Sex Female 131 81 Male 30 19 Brace No 79 49 Yes 82 51 Family history No 89 55 Yes 70 44 Unknown 2 1 Menarche No 89 68 Yes 42 32 Risser stage 0 84 52 1 24 15 2 26 16 3 27 17 Modifi ed Lenke type 1 26 16 2 12 7 3 63 39 4 53 5 38 24 6 17 11 Sanders Stage 1 7 4.4 2 28 17.4 3 41 25.5 4 45 28 5 7 4.3 6 31 19.2 7 2 1.2 Final outcome No curve progression 103 64 Curve progression to ≥50 ° 58 36 SPINE131133_LR 1008 PINE131133_LR 1008 28/05/15 7:47 PM 8/05/15 7:47 PM DEFORMITY Validation of Sanders Skeletal Maturity Staging for Scoliosis Progression • Sitoula et al Spine www.spinejournal.com 1009 of patients with an initial Cobb angle of 20 ° progressed ( P = 0.005). In SS3 and initial Cobb angles 20 °, 25 °, and 30 °, we observed 8% ( P = 0.3), 40% ( P = 0.06), and 43% ( P = 0.02) progression, respectively. Fifty-one percent of the patients in the present study were braced and half of them progressed ( Table 4 ). Thirty-four percent of the patients who were braced and progressed were in SS1 and SS2. Table 5illustrates the data on bracing and curve progression for various Cobb angles. DISCUSSION Skeletal maturity and curve magnitude are the 2 strongest predictors of curve progression in idiopathic scoliosis. This study sought to develop a predictive model for curve progression using initial curve magnitude and SS with a larger cohort (n = 161). In comparison with the work of Sanders et al, 9 this study offers several new insights: the raw data for rates of curve progression for each Cobb angle/SS subgroup are listed, and the results are likely more generalizable because patients were not excluded by either their sex or curve pattern because no difference in curve progression was found according to the modifi ed Lenke types.

        Previous authors have reported good sensitivity, specifi city, and accuracy when estimating PHV from serial height measurements to predict scoliosis curve progression. 1,12 However, this method requires serial offi ce visits to measure PHV, has limited prognostic value at a single offi ce visit, and is often discovered after curve progression. Dimeglio et al13 described the modifi ed Sauvegrain method to predict skeletal maturity in children. Its narrow window of use is limiting as it can be used only during the pubertal growth spurt while elbow physes remain open. The Risser sign is a popular and simple method for everyday clinical use, but it has a number of limitations. 8,14 It does not accurately predict the growth spurt as the apophysis fi rst appears (Risser stage 1), about 6 months after PHV. 13 Therefore, many children who are in their PHV are included in Risser 0 along with many others who are not yet in PHV. Furthermore, the iliac Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. TABLE 2. Categorical Variables Stratifi ed by Curve Progression Variable No Curve Progression Curve Progression ≥50 ° n (%) n (%) P Sex 0.029* Female 89 (68%) 42 (32%) Male 14 (47%) 16 (53%) Brace 0.0001* No 64 (81%) 15 (19%) Yes 39 (48%) 43 (52%) Family history 0.917 No 57 (64%) 32 (36%) Yes 45 (64%) 25 (36%) Unknown 1 (50%) 1 (50%) Menarche 0.073 No 56 (63%) 33 (37%) Yes 33 (79%) 9 (21%) Risser stage 0.005* 0 43 (51%) 41 (49%) 1 20 (83%) 4 (17%) 2 19 (73%) 7 (27%) 3 21 (78%) 6 (22%) Modifi ed Lenke type 0.226 1 19 (73%) 7 (27%) 2 8 (67%) 4 (33%) 3 34 (54%) 29 (46%) 4 2 (40%) 3 (60%) 5 27 (71%) 11 (29%) 6 13 (72%) 4 (28%) Sanders stage 0.0001* 1 2 (29%) 5 (71%) 2 8 (29%) 20 (71%) 3 27 (66%) 14 (34%) 4 36 (80%) 9 (20%) 5 4 (57%) 3 (43%) 6 24 (77%) 7 (23%) 7 2 (100%) 0 (0%) Initial curve magnitude 0.0001* 10 7 (100%) 0 (0%) (continued) TABLE 2. (Continued ) Variable No Curve Progression Curve Progression ≥50 ° n (%) n (%) P 15 20 (100%) 0(0%) 20 28 (80%) 7 (20%) 25 17 (71%) 7 (29%) 30 20 (61%) 13 (39%) 35 10 (45%) 12 (55%) 40 1 (7%) 14(93%) 45 0 (0%) 5 (100%) *n, number of patients; %, percentage progression; P, probability: value set at 5%. SPINE131133_LR 1009 PINE131133_LR 1009 28/05/15 7:47 PM 8/05/15 7:47 PM DEFORMITY Validation of Sanders Skeletal Maturity Staging for Scoliosis Progression • Sitoula et al 1010 www.spinejournal.com July 2015 0 with closed triradiate cartilage, and/or Risser 1, marked the curve acceleration phase.

        Currently, this system has not been validated to describe the risk of progression of curves to the surgical range. All children in the present series with initial Cobb angles of 40 ° or greater and SS of less than 7 had curve progression. Conversely, none of the patients with initial Cobb angles of 15 ° or less progressed. This latter fi nding is in agreement with previous authors. 9,17 All of the patients in SS2 with initial Cobb angles of 25 ° or greater progressed, which correlated apophysis exhibits a high incidence of abnormal development, including incomplete excursion and fragmentation. 15 Finally, the iliac apophysis is better visualized in an anteroposterior view of the pelvis than in a posteroanterior view, and radiographs for scoliosis are now routinely obtained posteroanteriorly to reduce radiation to breast tissue, making it diffi cult to interpret the full extent of ossifi cation of the iliac apophysis. 16 Nault et al8 modifi ed the Risser staging system by subdividing Risser 0 into 2 groups: open or closed triradiate cartilage.

        They reported that a Risser stage Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. TABLE 3. Number, Percentage, and Probability of Progression of Idiopathic Scoliosis to Surgery Curve SS1 SS2 SS3 SS4 SS5 SS6 SS7 p (n), % (95% CI), P p (n),% (95% CI), P p (n),% (95% CI), P p (n),% (95% CI), P p (n),% (95% CI), P p (n),% (95% CI), P p (n),% (95% CI), P 10 ° … 0 (2), 0% 0 (3), 0% 0 (2), 0% … … … 15 ° … 0 (2), 0% 0 (6), 0% 0 (7), 0% 0 (1), 0% 0 (4), 0% … 20 ° … 4 (8), 50% (15–85%), 0.005* 1 (12), 8% ( −7% to 24%), 0.30 2 (9), 22% ( −5% to 49%), 0.11 … 0 (6), 0% … 25 ° 1 (2), 50% ( −19% to 92%), 0.16 4 (4), 100% 2 (5), 40% ( −2% to 83%), 0.06 0 (4), 0% 0 (2), 0% 0 (6), 0% 0 (1), 0% 30 ° 2 (3), 67% (13%–80%), 0.01* 5 (5), 100% 3 (7), 43% (7%–80%), 0.02* 3 (11), 27% (9%–54%), 0.04* 0 (1), 0% 0 (6), 0% … 35 ° 1 (1), 100% 2 (2), 100% 4 (4), 100% 1 (9), 11% ( −9% to 32%), 0.30 1 (1, 100%) 3 (5), 60% (17%–98%), 0.01* … 40 ° 1 (1), 100% 4 (4), 100% 2 (2), 100% 3 (3), 100% 1 (1), 100% 3 (3), 100% 0 (1), 0% 45 ° … 1 (1), 100% 2 (2), 100% … 1 (1), 100% 1 (1), 100% … Unshaded cells correspond with combinations of Sanders Skeletal Maturity Staging System and curve magnitude for which surgery would be a possible treatment if curve of ≥50 ° was accepted as the threshold for surgery. * SS indicates Sanders stage; p, patients who had curve progression to ≥50 °; n, number of patients; %, percentage progression; CI, confi dence interval; P , probability: value set at 5%. Figure 1. Depiction of curve progression by Sanders stage. Curve progression to 50 ° or greater (taken as the threshold for surgery).

        Figure 2. Illustration of curve progression by initial curve magnitude. Curve progression to 50 ° or greater (taken as the threshold for surgery). SPINE131133_LR 1010 PINE131133_LR 1010 28/05/15 7:47 PM 8/05/15 7:47 PM DEFORMITY Validation of Sanders Skeletal Maturity Staging for Scoliosis Progression • Sitoula et al Spine www.spinejournal.com 1011 well with the study of Sanders et al. 9 We observed a difference in curve progression with initial Cobb angles of 25 ° and 30 ° and SS1 and SS2. All children with SS2 with initial Cobb angles of 25 ° and 30 ° progressed as opposed to 50% and 67% progression, respectively, for SS1. This difference in curve progression in contrast to maturity stage may be related to an intrinsic difference in curve progression between juvenile and adolescent curves or it may be due to a small sample size (sparse data bias). This can be verifi ed by future studies with larger sample size in these 2 stages. We urge caution: some curves may not behave predictably, which could skew results in a small sample. All patients with initial Cobb angles of 35 ° or greater and SS1, SS2, and SS3 progressed as observed in the predictive model of Sanders et al. 9 Beyond SS4, the threshold for curve progression diminished for curves 35 ° or less.

        This was in agreement with the curve acceleration phase as this stage occurred approximately 12 months after the curve acceleration phase. 9 In the authors’ view, patients with SS1, SS2, and SS3 with curves greater than 20 ° or 25 ° are at greatest risk for curve progression into the surgical range. It is these patients who are ideally suited for early intervention, such as bracing. Although this study lacks details regarding bracing (duration of bracing, patient compliance, and age of initiation) and is underpowered to detect the potential infl uence of bracing, bracing is an inextricable confounding factor. Bracing is likely to reduce the number of patients with signifi cant progression as predicted by the SS. This may be a direction for future prospective studies. Thirty-six percent of curves in the present series progressed to 50 ° or greater. Curve progression was observed in 53% of the male children and in 32% of the female children.

        Bunnell 18 Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. TABLE 4. Demographic Data on Bracing Brace n % Curve Progression Sanders Stage No (n [%]) Yes (n [%]) 1 2 3 4 5 6 7 No 79 49 64 (81%) 15 (19%) 0 6 (8%) 22 (28%) 27 (34%) 4 (5%) 19 (24%) 1 (1%) Yes 82 51 39 (48%) 43 (52%) 7 (8%) 21 (26%) 20 (24%) 18 (22%) 3 (4%) 12 (15%) 1 (1%) n indicates number of patients; %, percentage progression TABLE 5. Bracing and Curve Progression for Various Cobb Angles Cobb Angle Bracing Curve Progression No (n {%}) Yes (n {%}) P 10 ° No (n [%]) 7 (100%) {100%} Yes (n [%]) 15 ° No (n [%]) 18 (100%) {90%} Yes (n [%]) 2 (100%) {10%} 20 ° No (n [%]) 21 (95%) {75%} 1 (5%) {14%} 0.003 Yes (n [%]) 7 (54%) {25%} 6 (46%) {86%} 25 ° No (n [%]) 8 (89%) {47%} 1 (11%) {14%} 0.1 Yes (n [%]) 9 (60%) {53%} 6 (40%) {86%} 30 ° No (n [%]) 8 (80%) {40%} 2 (20%) {15%} 0.1 Yes (n [%]) 12 (52%) {60%} 11 (48%) {85%} 35 ° No (n [%]) 2 (33%) {20%} 4 (67%) {33%} 0.5 Yes (n [%]) 8 (50%) {80%} 8 (50%) {67%} 40 ° No (n [%]) 3 (100%) {21%} 0.6 Yes (n [%]) 1 (8%) {100%} 11 (92%) {79%} 45 ° No (n [%]) 4 (100%) {80%} Yes (n [%]) 1 (100%) {20%} [%] indicates row percentage; {%}, column percentage. SPINE131133_LR 1011 PINE131133_LR 1011 28/05/15 7:47 PM 8/05/15 7:47 PM DEFORMITY Validation of Sanders Skeletal Maturity Staging for Scoliosis Progression • Sitoula et al 1012 www.spinejournal.com July 2015 ➢ Key Points ‰ This study sought to validate the Sanders Skeletal Maturity Staging System to predict curve progression in idiopathic scoliosis. ‰

        Curve progression in idiopathic scoliosis correlated strongly with initial curve magnitude and skeletal maturity in female children and male children. ‰ All patients in Sanders stage (SS) 2 with initial Cobb angles of 25 ° or greater progressed, and patients in SS1 and SS3 with initial Cobb angles of 35 ° or greater progressed. Similarly, all patients with initial Cobb angles of 40 ° or greater progressed except those in SS7. Conversely, none of the patients with initial Cobb angles of 15 ° or less or those in SS5, SS6, and SS7 with initial Cobb angles of 30 ° or less progressed. Risk of progression was similar to that described by Sanders et al . 9 reported similar sex difference in curve progression: male children, 53%, and female children, 35%. This difference may be due to fewer male children in our cohort; Lonstein and Carlson 17 reported no signifi cant sex difference in scoliosis curve progression in their cohort of 727 patients. Sanders et al9 used only female children in their cohort. Patients were not evenly distributed in all stages. With our inclusion criteria, there were only 2 patients (1.2%) in SS7 in this series and neither progressed. All patients in SS8 were excluded because they were skeletally mature. Although children in SS7 (early mature stage) have some growth potential, 2 no curve progression was observed in the 2 patients in this stage. In the recently published Bracing for Adolescent Idiopathic Scoliosis Trial study, 19 bracing reduced the progression of curves in idiopathic scoliosis in 72% of patients as compared with 48% in the observation group.

        In our cohort, 51% of patients underwent bracing. Although there is likely an inhibitory effect with bracing, patients who were braced had a 52% rate of curve progression. In contrast, patients who were not braced had a 19% rate of curve progression. This contradictory effect of bracing can be explained by the retrospective nature of this study, the unmatched braced versus not braced groups, and the higher preponderance (34%) of patients within SS1 and SS2 in the bracing group. Despite benefi t of bracing shown by the BrAIST study, 19 for high-risk curves, there may still be a higher rate of curve progression. With refi nement and multi-institutional data, SS may be useful in guiding the usefulness of brace treatment. There are a few limitations with this study. This is a retrospective study with its inherent biases. Also, in some Cobb angle/SS subgroups, the numbers of patients were not adequate to make statistically stable predictions ( Table 3 ). The absence of patients in a particular Cobb angle/SS subgroup or a wide confi dence interval indicates potential areas for further analysis with a larger sample. For some patients, we defi ned fully capped Risser stage 4 as the endpoint for skeletal maturity although a small percentage of patients may have late progression. Finally, although we did record whether or not patients used a brace as part of the treatment and we performed a secondary analysis of data on bracing, it was not our intent to evaluate the effi cacy of bracing in this cohort. The factors such as in-brace correction, compliance, duration of bracing, and age at which bracing was initiated and stopped were not evaluated. In addition, our study is overall underpowered to detect differences in braced versus not braced patients. CONCLUSION In summary, this larger cohort validates the predictive model presented by Sanders et al9 in determining curve progression in idiopathic scoliosis. It further expands the model to include male and female patients of all curve types. This information may be helpful when counseling patients and families.

        Prospective studies with larger cohorts are needed to further validate these findings. References 1. Little DG , Song KM , Katz D , et al. Relationship of peak height velocity to other maturity indicators in idiopathic scoliosis in girls . J Bone Joint Surg Am 2000 ; 82 : 685 – 93 . 2. Sanders JO , Browne RH , McConnell SJ , et al. Maturity assessment and curve progression in girls with idiopathic scoliosis . J Bone Joint Surg Am 2007 ; 89 : 64 – 73 . 3. Acheson RM . The Oxford method of assessing skeletal maturity . Clin Orthop 1957 ; 10 : 19 – 39 . 4. Gruelich WW , Pyle SI . Radiographic Atlas of Skeletal Development of the Hand and Wrist. 2nd ed. Stanford, CA : Stanford University Press ; 1959 . 5. Tanner JM , Healy MJR , Goldstein H , et al. Assessment of Skeletal Maturity and Prediction of Adult Height (TW3 Method). 3rd ed. London : Saunders Ltd. ; 2001 . 6. Risser JC . The Iliac apophysis; an invaluable sign in the management of scoliosis . Clin Orthop 1958 ; 11 : 111 – 9 . 7. Hsieh CW , Liu TC , Wang JK , et al. Simplifi ed radius, ulna, and short bone-age assessment procedure using grouped-Tanner-Whitehouse method . Pediatr Int 2011 ; 53 : 567 – 75 . 8. Nault ML , Parent S , Phan P , et al. A modifi ed Risser grading system predicts the curve acceleration phase of female adolescent idiopathic scoliosis . J Bone Joint Surg Am 2010 ; 92 : 1073 – 81 . 9. Sanders JO , Khoury JG , Kishan S , et al. Predicting scoliosis progression from skeletal maturity: a simplifi ed classifi cation during adolescence . J Bone Joint Surg Am 2008 ; 90 : 540 – 53 . 10. Lenke LG , Betz RR , Harms J , et al. Adolescent idiopathic scoliosis: a new classifi cation to determine extent of spinal arthrodesis . J Bone Joint Surg Am 2001 ; 83-A : 1169 – 81 . 11. Cobb JR . Outline for the study of scoliosis . In: Blount WP , Banks SW , eds. The American Academy of Orthopaedic Surgeons Instructional Course Lectures. 5th ed. Ann Arbor, MI : JW Edwards ; 1948 : 261 – 75 . 12. Calvo IJ . Observations on the growth of the female adolescent spine and its relation to scoliosis . Clin Orthop 1957 ; 10 : 40 – 7 . 13. Dimeglio A , Charles YP , Daures JP , et al. Accuracy of the Sauvegrain method in determining skeletal age during puberty . J Bone Joint Surg Am 2005 ; 87 : 1689 – 96 . 14. Little DG , Sussman MD . The Risser sign: a critical analysis . J Pediatr Orthop 1994 ; 14 : 569 – 75 . 15. Shuren N , Kasser JR , Emans JB , et al. Reevaluation of the use of the Risser sign in idiopathic scoliosis . Spine 1992 ; 17 : 359 – 61