Abstract
Background:
The Proximal Femur Maturity Index (PFMI) can be used to assess skeletal maturity on existing whole-spine radiographs without additional radiation. However, the relationship between the PFMI at the initiation of bracing for adolescent idiopathic scoliosis (AIS) and subsequent curve progression remains unknown. This study aimed to investigate the relationship between the PFMI and curve progression, and the predictability of risks to adulthood curve progression and surgical thresholds based on the PFMI grade at brace initiation.
Methods:
This was a prospective study of 202 patients with AIS who were prescribed underarm bracing according to the Scoliosis Research Society criteria and had good brace-wear compliance. The patients were followed from brace initiation until complete skeletal maturity. Longitudinal data on the coronal Cobb angle and skeletal maturity assessments using Risser staging, Sanders staging, the distal radius and ulna classification, and the PFMI were collected. Each patient was assessed on whether the major curve progressed to ≥40° (adulthood deterioration) and ≥50° (the surgical threshold). Logistic regressions were used to predict probabilities of curve progression to the 2 thresholds, adjusted for factors that were significant in univariate analyses.
Results:
The PFMI correlated with the other skeletal maturity indices (rs [Spearman rank correlation] = 0.60 to 0.72, p < 0.001 for all). The pre-brace PFMI grade correlated with progression to ≥40° (rrb [rank-biserial correlation] = −0.30, p < 0.001) and to ≥50° (rrb = −0.20, p = 0.005). Based on regression models (p < 0.001) adjusted for the pre-brace major Cobb angle and curve type, brace initiation at PFMI grades 2 and 3 for a curve of ≥30° had predicted risks of 30% (95% confidence interval [CI], 4% to 55%) and 12% (95% CI, 7% to 17%), respectively, for progression to the surgical threshold. Brace initiation at PFMI grade 5 had 0% progression risk.
Conclusions:
The PFMI can be used for predicting curve progression and prognosticating brace outcomes in AIS. Patients with brace initiation at PFMI grade 4 for a curve of <30° or at grade 5 were unlikely to progress to the adulthood deterioration or surgical threshold. In comparison, skeletally immature patients initiating bracing at a PFMI grade of ≤3 for a major curve of ≥30° had a higher risk of progression despite compliant brace wear.
Level of Evidence:
Prognostic Level II. See Instructions for Authors for a complete description of levels of evidence.
Conventionally, Risser staging1 and bone age using a hand-wrist radiograph are used to assist clinical decision-making regarding the timing of brace initiation and timing of weaning or surgery in patients with adolescent idiopathic scoliosis (AIS). However, Risser staging is known for its insensitivity and inaccuracy in growth assessment2,3. In contrast, skeletal bone age determined utilizing Sanders staging4 based on distal phalangeal physeal maturation and the distal radius and ulna (DRU) classification5 based on distal radial and ulnar epiphyseal ossifications were found to be effective for indicating peak growth6,7 and for predicting curve progression8,9. Unfortunately, the need for a hand-wrist radiograph results in additional radiation exposure. It is therefore desirable to assess skeletal maturity with the same spine radiograph used for assessing scoliosis. The Proximal Femur Maturity Index (PFMI) has recently been established and validated for use in predicting peak growth and growth cessation10. It is also advantageous because the proximal femoral epiphyses are readily visible on whole-body or whole-spine images made with biplanar stereoradiography.
Despite the increased interest in the PFMI, its relationship with curve progression, particularly the ability of the PFMI at the time of brace initiation to predict the likelihood of curve progression, remains unknown. This missing knowledge would be beneficial in prognosticating the treatment outcome, specifically the likelihood of bracing success or failure. Determining the risk of curve progression will help clinicians to gauge whether brace treatment is worthwhile and to prognosticate the likelihood of surgery.
The present study aimed to evaluate the clinical relevance of the PFMI in guiding the timing of brace initiation in AIS, specifically the predictability of curve progression at skeletal maturity on the basis of the PFMI grade at the time of brace initiation.
Materials and Methods
Study Design
This prospective study of patients with AIS who were prescribed underarm bracing at our tertiary scoliosis clinic from December 2016 to December 2019 was registered at ClinicalTrials.gov (NCT04746417). Patients were to be followed from brace prescription until complete skeletal maturity at Risser stage 5. Included patients had to have been referred for bracing with a Boston brace in accordance with the Scoliosis Research Society criteria: skeletal immaturity (Risser stage of 0 to 2), a major curve Cobb angle of 20° to 40°, pre-menarchal onset or onset within the first year of menarche if female, and no prior treatment11. Exclusion criteria included loss to follow-up or suboptimal brace compliance of <18 hours/day. Patients without a left hand-wrist radiograph and a posteroanterior spine radiograph made on the same day at brace prescription were also excluded (Fig. 1). Local ethics committee approval and parental and patient informed consent were obtained.
Fig. 1.
Flowchart of patient recruitment at the time of brace prescription. SRS = Scoliosis Research Society.
Data Collection
Patient data collection included chronological age, sex, onset of menarche, and curve type (thoracic or lower [thoracolumbar or lumbar], according to the location of the curve apex) at the time of brace prescription. Longitudinal data on the coronal Cobb angles of the major and minor curves at each consultation as well as skeletal maturity determined using Risser staging, Sanders staging, and the DRU classification were measured by the attending orthopaedic specialist. Supine and first in-brace Cobb angles (routinely assessed during brace fabrication) were recorded, as spinal flexibility and the correction rate in the brace are related to the treatment prognosis for underarm bracing12-16. Brace compliance (hours/day) captured by a thermal sensor embedded in the brace was recorded. The PFMI of each included patient was assessed independently by 1 experienced orthopaedic surgeon blinded to the clinical and radiographic data.
Skeletal Maturity Parameters
Skeletal maturity at brace prescription was assessed using Risser staging17, Sanders staging4, the DRU classification5, and the PFMI (Fig. 2)10. Risser stages ranged from 0 to 2, with stage 0− indicating open triradiate cartilage (TRC) and stage 0+ indicating closed TRC. Sanders stages ranged from SS1 to SS8. SS3A and 3B indicated epiphyseal capping of digits and either absence or presence of epiphyseal capping at the distal radius, respectively. SS7A and SS7B indicated that all phalangeal physes are completely closed, with the medial growth plate of the distal ulna having ≤50% or >50% fusion, respectively18. Radius grades in the DRU system ranged from R1 to R11, and ulna grades ranged from U1 to U9. PFMI grades ranged from 0 to 6.
Fig. 2.
Schematic diagrams and definitions of the Proximal Femur Maturity Index grades.
Study Outcomes
The primary study outcomes were whether the major curve progressed to the adulthood deterioration threshold19 of 40° and to the surgical threshold20 of 50°. These thresholds are considered clinically relevant, as curves reaching 40° have been reported to have deterioration of approximately 1°/year during adulthood21 and those with deterioration to 50° usually require surgical intervention. When a patient had a major curve Cobb angle of 35° to 40° at the initiation of bracing, only a subsequent increase of >5°22 was considered to be curve progression to the adulthood deterioration threshold. The change in Cobb angle was calculated by comparing the major curve Cobb angle at either complete skeletal maturity or surgical intervention with that at brace initiation. The clinical relevance of the PFMI in guiding brace initiation was evaluated in terms of whether this skeletal maturity measure had the ability to predict curve progression to the 40° or 50° threshold.
Statistical Analysis
Each patient was evaluated for whether the major curve progressed to ≥40° and to ≥50°. We investigated the relationships of these 2 primary outcomes of brace treatment with each of the following parameters at brace initiation: skeletal maturity status, major Cobb angle (in degrees), sex, menarchal status (pre- or post-menarche), curve type, supine flexibility (in percent), and correction rate in the brace (in percent). The PFMI was tested for any correlations with other skeletal maturity indices using Spearman rank (rs) correlation tests. Since the bracing criteria included the Risser staging, the PFMI and Risser staging were cross-tabulated in relation to the primary outcomes. Logistic regressions were used to model each of the 2 binary bracing outcomes (i.e., curve progression to ≥40° [yes or no] and curve progression to ≥50° [yes or no]) on the basis of the PFMI grade and other factors identified as significantly correlated with that outcome in the univariate analyses (using a chi-square test of independence and point-biserial and rank-biserial correlation tests, according to the type of variable), and to generate predicted probabilities of curve progression. The risk of curve progression for each PFMI grade at brace initiation was estimated using the predicted probabilities (ranging from 0 to 1, based on which progression risks were represented in percentages) with true mean value and 95% confidence interval (CI), with delineation of pre-brace major Cobb angle being <30° or ≥30° because there are reports of a natural history of rapid progression of curves of ≥30°21,23. Evaluation of regression models was performed using the Hosmer-Lemeshow goodness-of-fit test, Nagelkerke R2 value (indicating the proportion for variation that is explained by the model24), accuracy (percentage of predictions that are correct), and confusion matrix. Regression models were also established using Risser staging, Sanders staging, and the DRU classification. A p value of <0.05 was considered significant.
Results
A total of 202 patients (84.7% female) who had compliance with bracing and follow-up until complete skeletal maturity were analyzed. The mean follow-up duration (and standard deviation) was 3.1 ± 0.8 years, and the mean major Cobb angle at brace initiation was 29.3° ± 3.9°. Patient demographics are presented in Table I. Curve progression (>5° increase in the major Cobb angle) was demonstrated in 18.8% (38) of the patients, and 2.5% required surgical intervention.
TABLE I.
Patient Demographics at Brace Initiation*
| Sex | |
| Female | 171 (84.7%) |
| Male | 31 (15.3%) |
| Age, female/male (yr) | 12.2 ± 1.1/13.5 ± 1.0 |
| Menarchal status | |
| Pre-menarchal | 90 (52.6%) |
| Post-menarchal/time since menarche | 81 (47.4%)/7.8 ± 5.8 mo |
| Risser stage | |
| 0− | 52 (25.7%) |
| 0+ | 63 (31.2%) |
| 1 | 41 (20.3%) |
| 2 | 46 (22.8%) |
| PFMI grade | |
| 1 | 1 (0.5%) |
| 2 | 19 (9.4%) |
| 3 | 76 (37.6%) |
| 4 | 88 (43.6%) |
| 5 | 18 (8.9%) |
| Sanders stage | |
| 1 | 1 (0.5%) |
| 2 | 20 (9.9%) |
| 3A | 28 (13.9%) |
| 3B | 40 (19.8%) |
| 4 | 44 (21.8%) |
| 5 | 13 (6.4%) |
| 6 | 40 (19.8%) |
| 7A | 13 (6.4%) |
| 7B | 3 (1.5%) |
| DRU radius grade | |
| 5 | 3 (1.5%) |
| 6 | 45 (22.3%) |
| 7 | 73 (36.1%) |
| 8 | 65 (32.2%) |
| 9 | 16 (7.9%) |
| DRU ulna grade | |
| 4 | 11 (5.4%) |
| 5 | 56 (27.7%) |
| 6 | 84 (41.6%) |
| 7 | 46 (22.8%) |
| 8 | 5 (2.5%) |
| Curve type | |
| Thoracic | 101 (50.0%) |
| Thoracolumbar or lumbar | 101 (50.0%) |
| Curve magnitude (°) | |
| Major curve | 29.3 ± 3.9 |
| Minor curve | 22.5 ± 5.0 |
| Correction rate by bracing† (%) | 46.3 ± 22.5 |
| Supine flexibility‡ (%) | 32.9 ± 18.7 |
| Progression: yes/no | |
| Pre-brace major Cobb angle | |
| All angles | 38 (18.8%)/164 (91.2%) |
| 20°-29.9° | 18 (15.0%)/102 (85.0%) |
| ≥30° | 20 (24.4%)/62 (75.6%) |
The values are given as the mean ± standard deviation or as the number with the percentage in parentheses; the percentage is the row percentage for progression and the column percentage for all other variables. PFMI = Proximal Femur Maturity Index, DRU = Distal Radius and Ulna classification.
Correction rate = (pre-brace Cobb angle – first in-brace Cobb angle)/pre-brace Cobb angle × 100%.
Supine flexibility = (pre-brace Cobb angle – supine Cobb angle)/pre-brace Cobb angle × 100%.
Table II compares the distributions of patient parameters between the groups with and without curve progression to each threshold. Progression to ≥40° occurred in 11.4% of the patients, and progression to ≥50° occurred in 4.5%. PFMI grade 2, SS1 and SS2 (the only patient with SS1 had curve progression), R5, U4, and Risser stage 0− were the maturity gradings in each system at brace initiation associated with the highest percentages of patients having major curve progression to ≥40°. Similarly, PFMI grade 2, SS2, R5, U4, and Risser stage 0− at brace initiation were associated with the highest percentages of progression to ≥50°. The PFMI grades correlated strongly with the Sanders stages (rs = 0.72, p < 0.001), radius (rs = 0.66, p < 0.001) and ulna (rs = 0.60, p < 0.001) grades, and Risser stages (rs = 0.65, p < 0.001). There was no progression to ≥50° when curves of <30° were braced at PFMI grade 3, 4, or 5. Cross-tabulation revealed that PFMI grade 3 corresponded to 55.2% (16) of the 29 patients with Risser stage 0−, 35.9% (14) of the 39 with stage 0+, 25% (6) of the 24 with stage 1, and 14.3% (4) of the 28 with stage 2 when they were braced for a <30° curve (Table III). Patients with PFMI grade 4 or 5 were at Risser stage 0+, 1, or 2.
TABLE II.
Curve Progression to Adulthood Deterioration (40°) and Surgical (50°) Thresholds*
| Parameter | Curve Progression to ≥40° | Curve Progression to ≥50° | ||
|---|---|---|---|---|
| Yes | No | Yes | No | |
| Sex | ||||
| Female | 18 (10.5%) | 153 (89.5%) | 8 (4.7%) | 163 (95.3%) |
| Male | 5 (16.1%) | 26 (83.9%) | 1 (3.2%) | 30 (96.8%) |
| Menarchal status | ||||
| Pre-menarchal | 17 (18.9%) | 73 (81.1%) | 7 (7.8%) | 83 (92.2%) |
| Post-menarchal | 1 (1.2%) | 80 (98.8%) | 1 (1.2%) | 80 (98.8%) |
| Curve type | ||||
| Thoracic | 13 (12.9%) | 88 (87.1%) | 8 (7.9%) | 93 (92.1%) |
| Thoracolumbar or lumbar | 10 (9.9%) | 91 (90.1%) | 1 (1.0%) | 100 (99.0%) |
| Risser stage | ||||
| 0− | 15 (28.8%) | 37 (71.2%) | 7 (13.5%) | 45 (86.5%) |
| 0+ | 5 (7.9%) | 58 (92.1%) | 2 (3.2%) | 61 (96.8%) |
| 1 | 2 (4.9%) | 39 (95.1%) | 0 | 41 (100%) |
| 2 | 1 (2.2%) | 45 (97.8%) | 0 | 46 (100%) |
| PFMI grade | ||||
| 1 | 0 | 1 (100%) | 0 | 1 (100%) |
| 2 | 8 (42.1%) | 11 (57.9%) | 3 (15.8%) | 16 (84.2%) |
| 3 | 11 (14.5%) | 65 (85.5%) | 5 (6.6%) | 71 (93.4%) |
| 4 | 4 (4.5%) | 84 (95.5%) | 1 (1.1%) | 87 (98.9%) |
| 5 | 0 | 18 (100%) | 0 | 18 (100%) |
| Sanders stage | ||||
| 1 | 1 (100%) | 0 | 0 | 1 (100%) |
| 2 | 7 (35.0%) | 13 (65.0%) | 4 (20.0%) | 16 (80.0%) |
| 3A | 8 (28.6%) | 20 (71.4%) | 3 (10.7%) | 25 (89.3%) |
| 3B | 5 (12.5%) | 35 (87.5%) | 1 (2.5%) | 39 (97.5%) |
| 4 | 1 (2.3%) | 43 (97.7%) | 1 (2.3%) | 43 (97.7%) |
| 5 | 1 (7.7%) | 12 (92.3%) | 0 | 13 (100%) |
| 6 | 0 | 40 (100%) | 0 | 40 (100%) |
| 7A | 0 | 13 (100%) | 0 | 13 (100%) |
| 7B | 0 | 3 (100%) | 0 | 3 (100%) |
| DRU radius grade | ||||
| 5 | 1 (33.3%) | 2 (66.7%) | 1 (33.3%) | 2 (66.7%) |
| 6 | 12 (26.7%) | 33 (73.3%) | 4 (8.9%) | 41 (91.1%) |
| 7 | 5 (6.8%) | 68 (93.2%) | 3 (4.1%) | 70 (95.9%) |
| 8 | 4 (6.2%) | 61 (93.8%) | 1 (1.5%) | 64 (98.5%) |
| 9 | 1 (6.3%) | 15 (93.8%) | 0 | 16 (100%) |
| DRU ulna grade | ||||
| 4 | 6 (54.5%) | 5 (45.5%) | 3 (27.3%) | 8 (72.7%) |
| 5 | 7 (12.5%) | 49 (87.5%) | 2 (3.6%) | 54 (96.4%) |
| 6 | 7 (8.3%) | 77 (91.7%) | 3 (3.6%) | 81 (96.4%) |
| 7 | 3 (6.5%) | 43 (93.5%) | 1 (2.2%) | 45 (97.8%) |
| 8 | 0 | 5 (100%) | 0 | 5 (100%) |
| Pre-brace major Cobb angle | ||||
| 20°-29.9° | 7 (5.8%) | 113 (94.2%) | 1 (0.8%) | 119 (99.2) |
| ≥30°† | 16 (19.5%) | 66 (80.5%) | 8 (9.8%) | 74 (90.2) |
The values are given as the number with the row percentage in parentheses.
For a pre-brace Cobb angle of ≥35°, there must have been curve progression of >5° to be classified as curve progression to the ≥40° threshold.
TABLE III.
Cross-Tabulation of the PFMI and Risser Staging with Curve Progression to Adulthood Deterioration (40°) and Surgical (50°) Thresholds
| Risser Stage* | ||||
|---|---|---|---|---|
| 0− | 0+ | 1 | 2 | |
| Curve progression to ≥40° | ||||
| Pre-brace major Cobb angle | ||||
| <30° | ||||
| PFMI grade | ||||
| 1† | 0/1 | |||
| 2 | 4/12 (33.3%) | |||
| 3 | 1/16 (6.25%) | 0/14 | 2/6 (33.3%) | 0/4 |
| 4† | 0/23 | 0/14 | 0/19 | |
| 5† | 0/2 | 0/4 | 0/5 | |
| ≥30° | ||||
| PFMI grade | ||||
| 1 | ||||
| 2 | 4/7 (57.1%) | |||
| 3 | 6/16 (37.5%) | 2/13 (15.4%) | 0/4 | 0/3 |
| 4 | 3/10 (30.0%) | 0/11 | 1/11 (9.1%) | |
| 5† | 0/1 | 0/2 | 0/4 | |
| Curve progression to ≥50° | ||||
| Pre-brace major Cobb angle | ||||
| <30° | ||||
| PFMI grade | ||||
| 1† | 0/1 | |||
| 2 | 1/12 (8.3%) | |||
| 3† | 0/16 | 0/14 | 0/6 | 0/4 |
| 4† | 0/23 | 0/14 | 0/19 | |
| 5† | 0/2 | 0/4 | 0/5 | |
| ≥30° | ||||
| PFMI grade | ||||
| 1 | ||||
| 2 | 2/7 (28.6%) | |||
| 3 | 4/16 (25.0%) | 1/13 (7.7%) | 0/4 | 0/3 |
| 4 | 1/10 (10%) | 0/11 | 0/11 | |
| 5† | 0/1 | 0/2 | 0/4 | |
The denominators are the numbers of patients with the particular Risser stage who corresponded to the PFMI grade in the first column. The numerators are the numbers of patients with curve progression to the threshold. Combinations in which some patients progressed to the threshold are indicated in bold.
This PFMI grade had a 0% curve progression rate.
The PFMI grades correlated with progression of the major curve to ≥40° (rrb [rank-biserial correlation] = −0.30, p < 0.001) and to ≥50° (rrb = −0.20, p = 0.005) (Table IV), and the strengths of the correlations were comparable with those of the Sanders staging (progression to ≥40°: rrb = −0.37, p < 0.001; progression to ≥40°: rrb = −0.24, p = 0.001). The pre-brace major Cobb angle correlated with progression to both thresholds, (p < 0.01), menarchal status at brace initiation (p < 0.001) and the correction rate in the brace (p = 0.009) correlated with curve progression to ≥40°, and curve type correlated with progression to ≥50° (p = 0.035).
TABLE IV.
Correlation of Factors with Curve Progression to Adulthood Deterioration (40°) and Surgical (50°) Thresholds*
| Categorical Variables | Curve Progression to ≥40°: Yes vs. No | Curve Progression to ≥50°: Yes vs. No | ||
|---|---|---|---|---|
| Chi-Square† | P Value | Chi-Square† | P Value | |
| Sex | 0.82 | 0.366 | 0.13 | >0.99 |
| Pre- vs. post-menarchal status | 14.11 | <0.001 | 4.09 | 0.067 |
| Thoracic vs. thoracolumbar or lumbar curve | 0.44 | 0.506 | 5.70 | 0.035 |
| Continuous Variables | rpb | P Value | rpb | P Value |
|---|---|---|---|---|
| Pre-brace major Cobb angle in ° | 0.20 | 0.005 | 0.21 | 0.002 |
| Correction rate by bracing in % | -0.19 | 0.009 | -0.10 | 0.143 |
| Supine flexibility in % | -0.10 | 0.176 | -0.12 | 0.099 |
| Ordinal Variables | rrb | P Value | rrb | P Value |
|---|---|---|---|---|
| Risser stage | −0.26 | <0.001 | −0.21 | 0.003 |
| PFMI grade | −0.30 | <0.001 | −0.20 | 0.005 |
| Sanders stage | −0.37 | <0.001 | −0.24 | 0.001 |
| DRU radius grade | −0.22 | 0.002 | −0.17 | 0.014 |
| DRU ulna grade | −0.21 | 0.003 | −0.14 | 0.053 |
Significant p values (<0.05) are indicated in bold. rpb = point-biserial correlation coefficient, rrb = rank-biserial correlation coefficient.
The Fisher exact test was used instead of the chi-square test if the expected cell count was <5.
Table V shows the probabilities of progression to each threshold as predicted by the logistic regression models based on the PFMI. The model for curve progression to ≥40°, which adjusted for menarchal status (for girls), the pre-brace major Cobb angle, and the correction rate (in percent) in the brace, had a chi-square value (with 7 degrees of freedom) of 47.238, p value of <0.001, and accuracy of 91.6%. The model for curve progression to ≥50°, which adjusted for the pre-brace major Cobb angle and curve type, had a chi-square value (with 6 degrees of freedom) of 27.758, p value of <0.001, and accuracy of 96.0%. The Hosmer and Lemeshow test and Nagelkerke R2 (see Appendices I and II) indicated a good fit for each regression model, comparable with those of models using other skeletal maturity indices (see Appendix III). For progression to ≥40°, the PFMI had high predictive ability that was comparable with that of Sanders staging and better than that of Risser staging. For progression to ≥50°, the true-positive rates of all prediction models appeared suboptimal because of the small number of patients with progression (9) and the predicted probability cutoff value of 0.5 for classifying a case as being positive. Independent of these metrics based on positive versus negative case classification, the amount of explained variance of curve progression and goodness of fit were adequate for generating predicted probabilities with 95% CIs. While there was only 1 patient who started bracing at PFMI grade 1, mean predicted risks for progression to ≥40° when bracing started at PFMI grade 2 were calculated as 27% (95% CI, 12% to 42%) for curves of <30° and 57% (95% CI, 35% to 79%) for curves of ≥30°. For progression to ≥50°, brace initiation at PFMI grade 2 for curves of <30° had an 8% risk (95% CI, 3% to 13%), and the risk was 30% (95% CI, 4% to 55%) for curves of ≥30°. If bracing curves of ≥30° commenced at PFMI grade 4, the predicted risk of progression to ≥50° was 3% (95% CI, 1% to 4%). No patient at PFMI grade 5 at brace initiation had progression, resulting in a calculated 0% risk of curve progression to the 2 thresholds regardless of pre-brace major-curve magnitude (Fig. 3). These 18 patients without progression who had started bracing at PFMI grade 5 were at Risser stages 0+ to 2 (Fig. 4), SS4 to SS7B, R7 to R9, and U6 to U8.
Fig. 3.
Predicted risk of curve progression to the adulthood deterioration (Fig. 3-A) and surgical thresholds (Fig. 3-B) according to PFMI grading. The values are shown as the mean predicted probability and 95% CI.
Fig. 4.
Number of patients with a pre-brace Cobb angle of <30° (Fig. 4-A) or ≥30° (Fig. 4-B) who had curve progression to ≥50°, shown according to Risser stage and PFMI grade.
TABLE V.
Observed Curve Progression Rate and Predicted Probability of Curve Progression to Each Threshold by PFMI Grade at Brace Initiation*
| Curve Progression Rate; Predicted Probability of Progression | |||||
|---|---|---|---|---|---|
| PFMI Grade 1 | PFMI Grade 2 | PFMI Grade 3 | PFMI Grade 4 | PFMI Grade 5 | |
| Curve progression to ≥40° | |||||
| Pre-brace major Cobb angle | |||||
| <30° | 0% (0/1) | 33.3% (4/12); 0.27 (0.12, 0.42); | 7.5% (3/40); 0.05 (0.03, 0.07) | 0% (0/56); 0.017 (0.009, 0.024) | 0% (0/11); 0.0 (0.0, 0.0) |
| ≥30° | NA; NA | 57.1% (4/7); 0.57 (0.35, 0.79) | 22.2% (8/36); 0.26 (0.19, 0.33) | 12.5% (4/32); 0.09 (0.05, 0.13) | 0% (0/7); 0.0 (0.0, 0.0) |
| Curve progression to ≥50° | |||||
| Pre-brace major Cobb angle | |||||
| <30° | 0% (0/1); — | 8.3% (1/12); 0.08 (0.03, 0.13) | 0% (0/40); 0.016 (0.010, 0.022) | 0% (0/56); 0.0018 (0.0012, 0.0024) | 0% (0/11); 0.0 (0.0, 0.0) |
| ≥30° | NA; NA | 28.6% (2/7); 0.30 (0.04, 0.55) | 13.9% (5/36); 0.12 (0.07, 0.17) | 3.1% (1/32); 0.03 (0.01, 0.04) | 0% (0/7); 0.0 (0.0, 0.0) |
The curve progression rate is expressed as the percentage in the relevant patients, with the fraction with curve progression in parentheses, and the predicted probability of progression (range, 0 to 1) is given with the 95% confidence interval generated from the regression model in parentheses. All models for which results are shown were significant (p < 0.05). NA = not applicable, because no patients fell in the category. A dash indicates that the probability was not calculated because only 1 patient fell in the category.
Discussion
Curve magnitude and growth potential at brace initiation are risk factors for curve progression after bracing25. In this study, we demonstrated that the PFMI can be used for estimating the curve progression risk and prognosticating the bracing outcome. Brace initiation at a PFMI grade of 3 or lower for a curve of ≥30° was found to carry a much higher risk of curve progression to the adulthood deterioration and surgical thresholds compared with brace initiation at PFMI grade 4, as well as compared with a curve of <30°. The risk of brace initiation at PFMI grade 5 was found to be minimal. These guidelines can aid clinicians in identifying patients with a high risk of progression, and provide clinicians with the evidence that facilitates their discussion with patients and parents about the expectations of the brace treatment outcome.
Relationship of PFMI with Brace Outcomes
The PFMI grades at brace initiation were found to be clinically relevant, as they significantly correlated with whether curves progressed to the adulthood deterioration and surgical thresholds, with correlation strength comparable to those demonstrated by Sanders staging, Risser staging, and the DRU classification. Initiating brace use at PFMI grade 1 or 2 resulted in the highest percentage of curves that progressed to both thresholds compared with other PFMI grades. This places the previous finding of peak height velocity (PHV) occurrence at PFMI grade 3 in perspective10,26. Curves prompting the need for bracing before PHV (i.e., at SS2, R5, or U4) demonstrated the highest rate of progression to the surgical threshold, as in previous studies involving these other indices4,8.
Predicted Curve Progression Risks Based on Pre-Brace PFMI Grade
Significant regression models for the likelihood of curve progression based on PFMI grading and curve magnitude at brace prescription were established in patients with good compliance, who could be identified because the prospective study design including recording of brace-wear compliance using a thermal sensor. Poor brace compliance was thus eliminated as a confounder for curve progression, in contrast to previous studies27-31. Bracing curves of ≥30° at PFMI grade 2 yielded a high predicted risk of 57% for progression to the adulthood deterioration threshold. The progression risk decreased to 26% for PFMI grade 3 and 9% for grade 4. Bracing curves of ≥30° at PFMI grade 2 carried a 30% risk of progression to the surgical threshold, and the risk was 12% if bracing started at PFMI grade 3. Thus, a trend of decreasing curve progression risk with increasing skeletal maturity status at the time of brace initiation was demonstrated. This is supported by the fact that skeletal maturity is one of the key factors in treatment prognosis8 and the relationship of remaining growth potential to curve progression32-34. In the present study, in which all patients had good brace-wear compliance, bracing resulted in smaller predicted probabilities of curve progression than those previously estimated. The highest risks of progression to the surgical threshold reported by Cheung et al. in 2018 were 40% for curves of 30° to 34.9° and 63% for curves of ≥35° when patients presented at stage U4 of the DRU classification8, whereas the highest predicted risk estimated in the present study was 30%, when curves of ≥30° were braced at PFMI grade 2, or 32% if initiated at U4. However, as previously noted8, the higher predicted risks in previous studies are believed to be underestimates because those studies lacked an objective measure of brace compliance, resulting in a heterogeneous group of patients that included some with poor compliance.
Clinical Application of the PFMI
The PFMI has the capability of estimating the curve progression risk of patients treated with bracing. Based on the predicted risks of progressions to the 40° and 50° thresholds, we identified patients for whom bracing of curves of ≥30° began at PFMI grade 2 as high-risk patients. Curves braced at PFMI grade 3 also had a considerable progression risk. We recommend that clinicians be vigilant when bracing immature patients at PFMI grade 3 or less and to better prepare these patients regarding the potential of brace treatment failure despite good brace compliance. This vigilance may involve closer follow-up intervals and earlier consideration of surgical intervention if deterioration occurs, to avoid unnecessarily prolonged bracing and to achieve optimal surgical outcomes such as better curve correction35. In comparison, curves of <30° braced at PFMI grade 4 (1.7% risk of progression to ≥40°, 0.2% risk of progression to ≥50°) and grade 5 are likely to have favorable bracing outcomes (see Table V and Appendix IV).
In addition, the PFMI can be utilized in identifying patients with minimal progression risk among those considered skeletally immature based on Risser staging. Approximately 60% to 70% of patients have been reported to be at Risser stage 025 at brace prescription, and the surgery rate was higher for those with open TRC (63.0% compared with 32.4% for those with closed TRC)36. Our results (see Appendix IIIa) agree with the finding of Karol et al. that curves of ≥30° in patients with open TRC had the highest risk of surgery, while curves of <30° had a 16.2% progression rate despite adequate bracing36. Among our patients with curves of <30° and Risser stage 0− at brace initiation, none of those at PFMI grade 3 had curve progression to the surgical threshold (Fig. 4-A, Table III). PFMI grade 3 represents open or closed TRC, with medial and lateral physeal beaking of the femoral head and a greater trochanteric notch. PFMI grade 3 can indicate an additional maturity status between Risser stages 0− and 0+ in which the TRC is nearly closed or not completely closed. This can aid in differentiating patients at Risser stage 0− who have less progression risk. Among patients starting bracing at Risser stage 0+, 1, or 2, identifying those at PFMI grade 5 can identify patients with minimal risk of curve progression to ≥50° (Fig. 4-B).
Limitations
Very immature PFMI grades of 0 and 1 could have been better represented if the study population had included patients with juvenile idiopathic scoliosis (JIS). However, PFMI grades 0 and 1 appear less relevant to our study population as the PFMI provides coverage of the peripubertal growth, and the timing of AIS presentation (10 to 18 years of age) does not coincide with the onset of the pubertal growth spurt37. Despite laying the foundation for the clinical utility of the PFMI in braced patients, validation in other ethnic groups and other regions should be performed38.
Conclusion
To our knowledge, this is the first prospective study on the utility of the PFMI in guiding brace initiation in patients with AIS. With use of the PFMI, the bracing outcome can be predictable for selected patients and the follow-up regimen can be customized for patients with a high progression risk. In particular, a patient with a major Cobb angle of ≥30° and a PFMI grade of 3 or less at the time of brace initiation was found to have a high risk of curve progression to the adulthood deterioration and surgical thresholds, whereas brace initiation for patients at PFMI grade 5 or those with a pre-brace major Cobb angle of <30° and PFMI grade 4 had a minimal risk of curve progression.
Appendix
Supporting material provided by the authors is posted with the online version of this article as a data supplement at jbjs.org (http://links.lww.com/JBJS/H825).
Footnotes
Investigation performed at the Department of Orthopaedics and Traumatology, The University of Hong Kong, Pokfulam, Hong Kong SAR, People’s Republic of China
A commentary by Joshua M. Pahys, MD, is linked to the online version of this article.
Disclosure: This study was funded in part by the Scoliosis Research Society, which did not play a role in the investigation. The Disclosure of Potential Conflicts of Interest forms are provided with the online version of the article (http://links.lww.com/JBJS/H824).
References
- 1.Risser JC. The Iliac apophysis; an invaluable sign in the management of scoliosis. Clin Orthop. 1958;11(11):111-9. [PubMed] [Google Scholar]
- 2.Minkara A, Bainton N, Tanaka M, Kung J, DeAllie C, Khaleel A, Matsumoto H, Vitale M, Roye B. High Risk of Mismatch Between Sanders and Risser Staging in Adolescent Idiopathic Scoliosis: Are We Guiding Treatment Using the Wrong Classification? J Pediatr Orthop. 2020. Feb;40(2):60-4. [DOI] [PubMed] [Google Scholar]
- 3.Cheung PWH, Cheung JPY. Does the Use of Sanders Staging and Distal Radius and Ulna Classification Avoid Mismatches in Growth Assessment with Risser Staging Alone? Clin Orthop Relat Res. 2021. Nov 1;479(11):2516-30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Sanders JO, Khoury JG, Kishan S, Browne RH, Mooney JF, 3rd, Arnold KD, McConnell SJ, Bauman JA, Finegold DN. Predicting scoliosis progression from skeletal maturity: a simplified classification during adolescence. J Bone Joint Surg Am. 2008. Mar;90(3):540-53. [DOI] [PubMed] [Google Scholar]
- 5.Cheung JP, Samartzis D, Cheung PW, Cheung KM, Luk KD. Reliability Analysis of the Distal Radius and Ulna Classification for Assessing Skeletal Maturity for Patients with Adolescent Idiopathic Scoliosis. Global Spine J. 2016. Mar;6(2):164-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sanders JO, Browne RH, McConnell SJ, Margraf SA, Cooney TE, Finegold DN. Maturity assessment and curve progression in girls with idiopathic scoliosis. J Bone Joint Surg Am. 2007. Jan;89(1):64-73. [DOI] [PubMed] [Google Scholar]
- 7.Cheung JP, Cheung PW, Samartzis D, Cheung KM, Luk KD. The use of the Distal Radius and Ulna Classification for the prediction of growth: peak growth spurt and growth cessation. Bone Joint J. 2016. Dec;98-B(12):1689-96. [DOI] [PubMed] [Google Scholar]
- 8.Cheung JPY, Cheung PWH, Samartzis D, Luk KD. APSS-ASJ Best Clinical Research Award: Predictability of Curve Progression in Adolescent Idiopathic Scoliosis Using the Distal Radius and Ulna Classification. Asian Spine J. 2018. Apr;12(2):202-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sitoula P, Verma K, Holmes L, Jr, Gabos PG, Sanders JO, Yorgova P, Neiss G, Rogers K, Shah SA. Prediction of Curve Progression in Idiopathic Scoliosis: Validation of the Sanders Skeletal Maturity Staging System. Spine (Phila Pa 1976). 2015. Jul 1;40(13):1006-13. [DOI] [PubMed] [Google Scholar]
- 10.Cheung PWH, Canavese F, Chan CYW, Wong JSH, Shigematsu H, Luk KDK, Cheung JPY. The Utility of a Novel Proximal Femur Maturity Index for Staging Skeletal Growth in Patients with Idiopathic Scoliosis. J Bone Joint Surg Am. 2022. Jan 6;104(7):630-40. [DOI] [PubMed] [Google Scholar]
- 11.Rowe DE. The Scoliosis Research Society Brace Manual.Accessed Accessed 2023 Nov 16. http://www.srs.org/UserFiles/file/bracing-manual/section1.pdf [Google Scholar]
- 12.Katz DE, Durrani AA. Factors that influence outcome in bracing large curves in patients with adolescent idiopathic scoliosis. Spine (Phila Pa 1976). 2001. Nov 1;26(21):2354-61. [DOI] [PubMed] [Google Scholar]
- 13.Cheung JPY, Cheung PWH. Supine flexibility predicts curve progression for patients with adolescent idiopathic scoliosis undergoing underarm bracing. Bone Joint J. 2020. Feb;102-B(2):254-60. [DOI] [PubMed] [Google Scholar]
- 14.Wong LPK, Cheung PWH, Cheung JPY. Supine correction index as a predictor for brace outcome in adolescent idiopathic scoliosis. Bone Joint J. 2022. Apr;104-B(4):495-503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Cheung JPY, Cheung PWH, Yeng WC, Chan LCK. Does Curve Regression Occur During Underarm Bracing in Patients with Adolescent Idiopathic Scoliosis? Clin Orthop Relat Res. 2020. Feb;478(2):334-45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Cheung JPY, Yiu KKL, Vidyadhara S, Chan PPY, Cheung PWH, Mak KC. Predictability of Supine Radiographs for Determining In-Brace Correction for Adolescent Idiopathic Scoliosis. Spine (Phila Pa 1976). 2018. Jul 15;43(14):971-6. [DOI] [PubMed] [Google Scholar]
- 17.Nault ML, Parent S, Phan P, Roy-Beaudry M, Labelle H, Rivard M. A modified Risser grading system predicts the curve acceleration phase of female adolescent idiopathic scoliosis. J Bone Joint Surg Am. 2010. May;92(5):1073-81. [DOI] [PubMed] [Google Scholar]
- 18.Cheung PWH, Cheung JPY. Sanders stage 7b: Using the appearance of the ulnar physis improves decision-making for brace weaning in patients with adolescent idiopathic scoliosis. Bone Joint J. 2021. Jan;103-B(1):141-7. [DOI] [PubMed] [Google Scholar]
- 19.Weinstein SL, Dolan LA, Spratt KF, Peterson KK, Spoonamore MJ, Ponseti IV. Health and function of patients with untreated idiopathic scoliosis: a 50-year natural history study. JAMA. 2003. Feb 5;289(5):559-67. [DOI] [PubMed] [Google Scholar]
- 20.Weinstein SL, Dolan LA, Wright JG, Dobbs MB. Effects of bracing in adolescents with idiopathic scoliosis. N Engl J Med. 2013. Oct 17;369(16):1512-21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Alcala C, Mehbod AA, Ramos O, Dawson JM, Denis F, Garvey TA, Perra JH, Beauchamp E, Transfeldt EE. Moderate scoliosis continues to progress at 30-year follow-up: a call for concern? Spine Deform. 2023. Sep 27. Epub ahead of print. [DOI] [PubMed] [Google Scholar]
- 22.Pruijs JE, Hageman MA, Keessen W, van der Meer R, van Wieringen JC. Variation in Cobb angle measurements in scoliosis. Skeletal Radiol. 1994. Oct;23(7):517-20. [DOI] [PubMed] [Google Scholar]
- 23.Weinstein SL. The Natural History of Adolescent Idiopathic Scoliosis. J Pediatr Orthop. 2019. Jul;39(6,)(Supplement 1 Suppl 1):S44-6. [DOI] [PubMed] [Google Scholar]
- 24.Nagelkerke NJD. A note on a general definition of the coefficient of determination. Biometrika. 1991;78(3):691-2. [Google Scholar]
- 25.Hawary RE, Zaaroor-Regev D, Floman Y, Lonner BS, Alkhalife YI, Betz RR. Brace treatment in adolescent idiopathic scoliosis: risk factors for failure-a literature review. Spine J. 2019. Dec;19(12):1917-25. [DOI] [PubMed] [Google Scholar]
- 26.Escalada F, Marco E, Duarte E, Ma Muniesa J, Boza R, Tejero M, Cáceres E. Assessment of angle velocity in girls with adolescent idiopathic scoliosis. Scoliosis. 2009. Sep 16;4:20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Aulisa AG, Giordano M, Falciglia F, Marzetti E, Poscia A, Guzzanti V. Correlation between compliance and brace treatment in juvenile and adolescent idiopathic scoliosis: SOSORT 2014 award winner. Scoliosis. 2014. Jun 13;9:6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Brox JI, Lange JE, Gunderson RB, Steen H. Good brace compliance reduced curve progression and surgical rates in patients with idiopathic scoliosis. Eur Spine J. 2012. Oct;21(10):1957-63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Lou EH, Hill DL, Raso JV, Moreau M, Hedden D. How quantity and quality of brace wear affect the brace treatment outcomes for AIS. Eur Spine J. 2016. Feb;25(2):495-9. [DOI] [PubMed] [Google Scholar]
- 30.Katz DE, Herring JA, Browne RH, Kelly DM, Birch JG. Brace wear control of curve progression in adolescent idiopathic scoliosis. J Bone Joint Surg Am. 2010. Jun;92(6):1343-52. [DOI] [PubMed] [Google Scholar]
- 31.Rahman T, Bowen JR, Takemitsu M, Scott C. The association between brace compliance and outcome for patients with idiopathic scoliosis. J Pediatr Orthop. 2005. Jul-Aug;25(4):420-2. [DOI] [PubMed] [Google Scholar]
- 32.Ghanem I, Rizkallah M. The impact of residual growth on deformity progression. Ann Transl Med. 2020. Jan;8(2):23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Dimeglio A, Canavese F. Progression or not progression? How to deal with adolescent idiopathic scoliosis during puberty. J Child Orthop. 2013. Feb;7(1):43-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Zapata KA, Sucato DJ, Lee MC, Jo CH. Skeletally Immature Patients With Adolescent Idiopathic Scoliosis Curves 15°-24° Are at High Risk for Progression. Spine Deform. 2019. Nov;7(6):870-4. [DOI] [PubMed] [Google Scholar]
- 35.Charles YP, Daures JP, de Rosa V, Diméglio A. Progression risk of idiopathic juvenile scoliosis during pubertal growth. Spine (Phila Pa 1976). 2006. Aug 1;31(17):1933-42. [DOI] [PubMed] [Google Scholar]
- 36.Karol LA, Virostek D, Felton K, Jo C, Butler L. The Effect of the Risser Stage on Bracing Outcome in Adolescent Idiopathic Scoliosis. J Bone Joint Surg Am. 2016. Aug 3;98(15):1253-9. [DOI] [PubMed] [Google Scholar]
- 37.Limony Y, Kozieł S, Friger M. Age of onset of a normally timed pubertal growth spurt affects the final height of children. Pediatr Res. 2015. Sep;78(3):351-5. [DOI] [PubMed] [Google Scholar]
- 38.Donzelli S, Fregna G, Zaina F, Livetti G, Reitano MC, Negrini S. Predictors of Clinically Meaningful Results of Bracing in a Large Cohort of Adolescents with Idiopathic Scoliosis Reaching the End of Conservative Treatment. Children (Basel). 2023. Apr 13;10(4):719. [DOI] [PMC free article] [PubMed] [Google Scholar]