Abstract
Background:
The Risser sign can be assessed with the United States method or the European method. The Tanner-Whitehouse method estimates skeletal age on the basis of hand radiography and digital skeletal age. Digital skeletal age scores between 400 and 425 are associated with the beginning of the curve acceleration phase or peak growth velocity in adolescent idiopathic scoliosis. The first objective of the present study was to evaluate the agreement between the two Risser grading systems. The second objective was to identify which grading system best predicts a digital skeletal age of between 400 and 425. The third objective was to explore a new system that could be used to replace the Risser method.
Methods:
One hundred female patients with adolescent idiopathic scoliosis were recruited in this cross-sectional descriptive study. Each patient was evaluated with posteroanterior spine and hand radiographs. The Risser sign was measured according to both the United States and European grading systems. Digital skeletal age was calculated, and triradiate cartilage ossification was assessed.
Results:
With use of kappa statistics, moderate agreement between the United States and European grading systems was seen. Risser stages alone were not good predictors of the curve acceleration phase. A new system with three groups was tested, and the second group (Risser 0 with closed triradiate cartilage and Risser 1) was the best predictor of a digital skeletal age score of between 400 and 425.
Conclusions:
Two Risser grading systems coexist, and the agreement between them is moderate. No Risser stage was found to be a good clinical landmark for the beginning of the curve acceleration phase of adolescent idiopathic scoliosis. A new group, Risser 0 with closed triradiate cartilage and Risser 1, was the best predictor of the beginning of the curve acceleration phase. This new system is easy to implement and is based on findings that are available on spine radiographs. It should be used at the first visit and for scoliosis follow-up to assess skeletal maturity and correlation with the curve acceleration phase.
Adolescent idiopathic scoliosis progresses during periods of rapid growth1. This curve progression is related to various factors, which include the pattern and magnitude of the curve, the age of the patient at the time of presentation, menarchal status, and Risser sign2.
The assessment of skeletal maturity is of great importance in cases of adolescent idiopathic scoliosis because it is used to determine the interval between follow-up evaluations and the duration of brace treatment as well as to guide surgical planning. Bone age can be assessed on the basis of hand and wrist radiography with use of either the Greulich and Pyle atlas3 or the Tanner-Whitehouse method4, but the most common approach in cases of adolescent idiopathic scoliosis is to assess ossification of the iliac apophysis with use of the Risser sign.
In the early 1900s, it was commonly assumed that the progression of all types of scoliosis stopped completely with the end of vertebral growth5. In 1936, Risser recognized the appearance and completion of ossification of the iliac apophysis as a physiological sign of vertebral growth cessation6. At first, he associated the end of vertebral growth with the culmination of apophysis ossification over the iliac crest. Later, in 1958, he stated that fusion of the iliac apophysis to the iliac crest indicated vertebral growth completion7. In 1965, Stagnara et al.8 suggested that the Risser sign corresponded with fusion of the iliac apophysis to the iliac crest and postulated that this finding corresponded with the termination of deformity progression. This confusion about the sign representing the end of vertebral growth may explain the origin of two different methods of measuring the Risser sign. It was recently noted that the Risser sign has not been clearly described in the literature9.
The first grading system, which is more popular in North America, will be called the United States Risser sign in this paper. This system includes six stages (Fig. 1). Stage 0 represents the absence of an ossification center. Stages 1 to 4 correspond with excursion of the iliac apophysis over quarters defined on the iliac crest. Stage 5 corresponds with the completion of fusion of the iliac apophysis to the iliac crest.
The other grading system, also with six stages, is popular in European countries8 and will be referred to as the European Risser sign in this paper. Stages 0 and 5 share a common definition in the United States and European grading systems, but differences exist between stages 1 to 4. In the European system, the iliac crest is split into thirds (stages 1 to 3) and stage 4 represents the beginning of fusion (Fig. 1).
As already mentioned, the Risser sign is the skeletal maturity indicator that is preferred for follow-up in cases of adolescent idiopathic scoliosis, and thus the coexistence of two different grading systems might have an important clinical impact. Furthermore, the Risser sign is applied in multicentric studies and is included in multicentric databases. Therefore, the fact that the Risser sign is applied in multicenter studies and multicentric databases could become a source of confusion if the grading system is not properly specified first.
Another landmark of skeletal maturity is triradiate cartilage closure, which occurs generally before the beginning of iliac crest apophysis ossification during Risser stage 010.
Bone age in children can be assessed with use of the Greulich and Pyle atlas, which is a reference atlas that contains radiographs of the hand and wrist associated with different skeletal ages and sex. Another method that involves the use of radiographs of the hand and wrist is the Tanner-Whitehouse method4. That method is a more accurate research tool11, although it is not used as widely in clinical settings.
With the Tanner-Whitehouse method, the distal radial and ulnar epiphyses as well as the metacarpal and phalangeal epiphyses of the first, third, and fifth digits are considered when assessing skeletal age. A maturity score is assigned to each bone, and the sum of the scores gives the RUS (radius, ulna, small bones) score. The digital skeletal age (DSA) score is obtained when individual radial and ulnar scores are removed from the RUS score.
Digital skeletal age is highly correlated with the curve acceleration phase of curve progression in adolescent idiopathic scoliosis12. Sanders et al. found that a DSA score of between 400 and 425 corresponded with the beginning of the curve acceleration progression period of adolescent idiopathic scoliosis12.
Measurement of the digital skeletal age in the clinical context is time-consuming, requires comparison with reference books, and necessitates an extra radiograph of the left hand. The Risser sign is easy to use and can be measured on standard serial spinal radiographs.
Objectives
The first objective of the present study was to evaluate the agreement between the European and United States Risser grading systems. The second objective was to find which Risser stage, whether United States or European, best predicted the beginning of the curve acceleration phase corresponding to a DSA score of between 400 and 425, as defined by Sanders et al.12. The third objective was to explore new landmarks that could modify the Risser grading systems to better predict the same DSA score of between 400 and 425.
The independent variable in this cross-sectional descriptive study was the Risser stage, and the dependent variable was the DSA score.
All female patients who were seen for the first time in the scoliosis clinic of Sainte-Justine Hospital between September 2006 and January 2008 were included if they fulfilled the inclusion and exclusion criteria, using a sample of convenience. The inclusion criteria were (1) an age of between ten and seventeen years, and (2) the availability of a posteroanterior radiograph of the spine (including both full iliac crests) and a radiograph of the left hand and wrist, made on the same day. The exclusion criteria were (1) metabolic bone disease, (2) growth disturbance, (3) chromosomal anomalies, and (4) previous spine or pelvic surgery.
Digital standing posteroanterior spine radiographs were reviewed for all patients. The United States Risser sign was assessed during a first session. During a second session, two days later, the European Risser sign was assessed by the same observer (M.-L.N.), who was blinded to the personal information of the patient. Both iliac crests were evaluated, but only the most advanced apophyseal ossification center was used. The status of the triradiate cartilage (i.e., open or closed) was also assessed.
Radiographs of the left hand and wrist were reviewed to evaluate bone age and the DSA score according to the Tanner-Whitehouse method. A detailed description of the method is available as supplemental material in a recent study by Sanders et al.13. The single observer who performed this evaluation (M.-L.N.) was blinded to the chronological age and Risser sign of the patients.
Statistical Analysis
To fulfill the first objective, kappa statistics were tested to investigate agreement between the European and United States Risser signs. Two analyses were performed: one included all patients at Risser stages 0 to 5, and the other included patients at Risser stages 1, 2, 3, and 4. The second analysis excluded stages 0 and 5 because the definition of those stages is the same in both systems. A Spearman correlation coefficient was also calculated.
For the second objective, two one-way analyses of variance were conducted, one with the United States Risser grading system as the independent variable and the DSA score as the dependent variable, and the other with the European grading system as the independent variable and the DSA score as the dependent variable.
For the third objective, a modified Risser grading system was tested. These modifications involved the division of Risser stage 0 into two groups, depending on the status of the triradiate cartilage (open or closed). The first, group -1, included all subjects at Risser 0 with an open triradiate cartilage. The second, group 0, included all subjects at Risser 0 with a closed triradiate cartilage. A one-way analysis of variance was performed with Bonferroni multiple comparisons with use of the six groups (-1, 0, Risser 1, Risser 2, Risser 3, Risser 4). According to multiple comparisons, three groups were then defined: group A (Risser 0 with open triradiate cartilage), group B (Risser 0 with closed triradiate cartilage and Risser 1), and group C (Risser 2, 3, and 4).
Intraobserver and interobserver reliability were assessed for the new system. Thirty-five posteroanterior spine radiographs were randomly selected from among those for the 100 patients in the present study. Those radiographs were saved on three disks, and three observers (M.-L.N., P.P., S.P.) reviewed the thirty-five radiographs independently. One observer was an experienced spine surgeon, and the two others were senior orthopaedic residents undergoing graduate studies in scoliosis. All thirty-five radiographs were classified according to the new system as A, B, or C. Two observers repeated the same classification two weeks later. Kappa statistics were performed to determine the interobserver and intraobserver reliability.
Source of Funding
Federal funds were received in support of this work. The funding source did not play a role in the investigation.
One hundred female patients with a mean chronological age (and standard deviation) of 13.3 ± 1.4 years (range, ten to seventeen years) were recruited. As no patient had reached Risser stage 5, the analyses excluded Risser stage 5.
Concordance Between the United States and European Risser Grading Systems
Moderate agreement was found between the United States and European grading systems when all stages were included (? = 0.532). Almost perfect agreement (96.6%) was observed for Risser 0. Agreement was 95% (nineteen of twenty) for United States Risser 1 and 65.5% (nineteen of twenty-nine) for European Risser 1 (Table I). The second kappa analysis of twenty patients with Risser stages 1 to 4 showed low agreement (? = 0.336). The Spearman correlation coefficient between both grading systems was 0.97. The United States grading system always gave higher correlation values than the European grading system did.
Predictors of a DSA Score of Between 400 and 425
Digital skeletal age variations for each Risser stage and for both grading systems are presented in Figures 2-A and 2-B.
For Risser 0, the mean DSA scores for the United States and European grading systems were 376 and 378, respectively. For each stage of both grading systems, the mean DSA scores are presented in Figure 3 and in Table II (with 95% confidence intervals). For all stages, the European system was associated with a higher DSA score than the United States system was. Given the confidence intervals, there were no significant differences between the mean DSA scores associated with the two grading systems at any Risser stage.
New System
As no significant difference was apparent between the mean predicted DSA score for each Risser stage, it was not possible to determine if one grading system was superior to the other for predicting a DSA score of between 400 and 425.
For Risser 0 in both grading systems, many patients had DSA values of >400, as seen in Figures 2-A and 2-B, meaning that even at Risser 0, some patients had already reached the target interval of the DSA score.
Analysis of the six new groups (-1, 0, Risser 1, Risser 2, Risser 3, Risser 4) resulted in a new distribution in which there were no more patients with a DSA score of >400 in group -1 (Risser 0 with an open triradiate cartilage) (Figs. 4-A and 4-B). The mean DSA score for each new group is shown in Figure 5. From the one-way analysis of variance and the multiple comparisons (Table III), three groups can be organized: group A (Risser 0 with an open triradiate cartilage), group B (Risser 0 with a closed triradiate cartilage and Risser 1), and group C (Risser 2, 3, and 4).
The mean DSA scores for these three new groups are presented in Table IV (along with the 95% confidence intervals). Again, there was no significant difference between the United States and European systems, but the three groups were significantly different from each other. Group B corresponded best to the target DSA interval of 400 to 425, with a mean DSA score of 432 for the United States grading system and 444 for European grading system.
The intraobserver reliability study showed almost perfect agreement, with kappa values of 0.914 and 0.829. The interobserver reliability study also had strong agreement, with kappa values of 0.786, 0.871, and 0.828.
As the definitions of the two Risser grading systems are different, the moderate agreement (? = 0.532) noted was expected. Stages 0 and 5 have the same description in the United States and European Risser systems, so when Risser 0 was excluded from the analysis, the kappa value dropped to 0.336. These results are concordant with those in the study by Bitan et al.14, who reported a kappa value of 0.422 for females and 0.356 for males in the evaluation of 301 radiographs by two different investigators, one of whom used the United States system and the other of whom used the European system.
A limitation of the Risser sign is the high frequency of anomalous iliac apophysis development. Risser reported that 10% of all patients have short excursions of ossification15. Zaoussis and James16 and Shuren et al.17 noted 40% and 41% rates of anomalous development (fragmented ossification, short excursion, posterior ossification, ossification progression from the posteromedial to the anteromedial position).
Another limitation of the Risser sign is that radiographs that are used for the evaluation of scoliosis are usually made posteroanteriorly rather than anteroposteriorly. Ossification of the iliac apophysis is best viewed on anteroposterior radiographs. Izumi18 reported that there was agreement between posteroanterior and anteroposterior Risser staging in only 58% (fifty-two) of eighty-nine cases studied.
In previous reports in the literature, assessment of the Risser stage showed fair to excellent interobserver and intraobserver reproducibility. A study by Dhar et al.19 showed interobserver agreement of 89.2% and intraobserver agreement of 93.4%. Another study by Goldberg et al.20 showed excellent interobserver agreement, with a kappa value of 0.80. A study by Shuren et al.17 showed kappa values for interobserver agreement of between only 0.31 and 0.53.
As noted by Bitan et al.14 and demonstrated in the present study, the coexistence of two grading systems for the same sign to assess skeletal maturity is a substantial potential source of confusion, especially in the era of multicentric trials and databases. One easy solution would be always to specify which grading system is being used. The coexistence of two systems could also have an impact on other aspects of follow-up and treatment as variations may occur from one surgeon to another, depending on the grading system used. A surgeon using the United States Risser system could stop brace treatment at a younger skeletal age in comparison with another surgeon using the European system.
Many orthopaedic surgeons are not aware of the coexistence of these two systems, and that was why we used kappa statistics to evaluate the concordance between them.
Even though the Risser sign has many limitations, it can be measured on almost all routine radiographs and it is a well-established variable for the follow-up of patients who have adolescent idiopathic scoliosis. Thus, finding a Risser sign equivalent to the DSA score, corresponding to the beginning of the curve acceleration phase, is clinically relevant.
Prediction of Curve Acceleration Phase with the Risser Sign
Sanders et al.12 studied the relationship of different maturity indicators and the curve acceleration phase. They observed that the RUS score of the Tanner-Whitehouse method was the indicator that best correlated (Pearson r = 0.93) with the beginning of this curve acceleration phase. When only the DSA score was considered, the correlation was also 0.93, but with a tighter fit of the curve at the beginning of the curve acceleration phase. The authors reported a relationship between a DSA score of 400 to 425 and the beginning of the most rapid progression of curve magnitude. According to our analysis of variance, there was no significant difference between the United States and European grading systems with respect to the mean DSA score associated with each Risser stage.
As expected, the European Risser method was associated with a higher DSA score, likely because the Risser stage requires longer iliac crest ossification excursion in the European system.
The important point here is that the beginning of ossification of the iliac apophysis seems to be a good indicator of the start of the curve acceleration phase, independent of the distance traveled by the ossification center. However, there was a problem with the Risser 0 group, which included patients with a wide distribution of DSA scores. This problem was considerable enough to make the beginning of iliac ossification a non-useful parameter in the clinical context because some patients at Risser stage 0 could have a DSA score of >400, in the critical range of the beginning of the curve acceleration phase.
In our study, patients in group B (Risser 0 with closed triradiate cartilage and Risser 1) best corresponded with a DSA score of between 400 and 425, representing the beginning of the curve acceleration phase.
A limitation of this new system is that it is less specific than the DSA score. However, this limitation is acceptable because the DSA score is not as suitable in the clinical context and requires an additional radiograph of the left hand. Other limitations include errors introduced by anomalous iliac apophysis development and variations in the Risser sign between anteroposterior and posteroanterior radiographs.
To our knowledge, this is the first study in which skeletal maturity indicators on spine radiographs were associated with a parameter corresponding with the curve acceleration phase of adolescent idiopathic scoliosis. While no Risser stage was found to be a good clinical landmark for the beginning of the curve acceleration phase, the new group (Risser 0 with a closed triradiate cartilage and Risser 1) was found to be the best predictor of the beginning of the curve acceleration phase. The clinical relevance of these findings leads us to recommend that patients with adolescent idiopathic scoliosis with a closed triradiate cartilage and iliac apophysis ossification that covers a maximum of one-third of the iliac crest be carefully followed until at least one-third of the iliac apophysis is ossified. If brace treatment is initiated, follow-up is generally indicated until more complete iliac apophysis ossification is present.
Ylikoski
M. Growth and progression of adolescent idiopathic scoliosis in girls. J Pediatr Orthop B.
2005;14:320-4.[PubMed][CrossRef]
Lonstein
JE;
Carlson
JM. The prediction of curve progression in untreated idiopathic scoliosis during growth. J Bone Joint Surg Am.
1984;66:1061-71.[PubMed]
Greulich
WW;
Pyle
SI. Radiographic atlas of skeletal development of the hand and wrist. Stanford, California: Stanford University Press; 1959.
Tanner
JM;
Healy
MJR;
Goldstein
H;
Cameron
N. Assessment of skeletal maturity and prediction of adult height: TW3 method. 3rd ed.London: WB Saunders; 2001.
Risser
JC;
Ferguson
AB. Scoliosis: its prognosis. J Bone Joint Surg.
1936;18:667-70.
Risser
JC. Important practical facts in the treatment of scoliosis. Instr Course Lect.
1948;5:248-60.
Risser
JC. The iliac apophysis; an invaluable sign in the management of scoliosis. Clin Orthop.
1958;11:111-9.[PubMed]
Stagnara
P;
Desbrosses
J;
Michel
CR;
Dupeloux
J;
Fauchet
R;
Berthou
JD;
Perdriolle
R. [Structural scoliosis. Terminal results of orthopedic treatments during the growth period]. Rev Chir Orthop Reparatrice Appar Mot.
1965;51:33-52. .[PubMed]
Brand
RA. 50 years ago in CORR: the iliac apophysis: an invaluable sign in the management of scoliosis Joseph C. Risser MD CORR
1958;11:111-119. .
Sanders
JO. Maturity indicators in spinal deformity. J Bone Joint Surg Am.
2007;89
Suppl 1:14-20.[PubMed][CrossRef]
Fry
EI. Assessing skeletal maturity: comparison of the atlas and individual bone technics. Nature.
1968;220:496-7.[PubMed][CrossRef]
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;89:64-73.[PubMed][CrossRef]
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;90:540-53.[PubMed][CrossRef]
Bitan
FD;
Veliskakis
KP;
Campbell
BC. Differences in the Risser grading systems in the United States and France. Clin Orthop Relat Res.
2005;436:190-5.[PubMed][CrossRef]
Risser
JC. Scoliosis: past and present. J Bone Joint Surg Am.
1964;46:167-99.[PubMed]
Zaoussis
AL;
James
JI. The iliac apophysis and the evolution of curves in scoliosis. J Bone Joint Surg Br.
1958;40:442-53.[PubMed]
Shuren
N;
Kasser
JR;
Emans
JB;
Rand
F. Reevaluation of the use of the Risser sign in idiopathic scoliosis. Spine (Phila Pa 1976).
1992;17:359-61.[PubMed][CrossRef]
Izumi
Y. The accuracy of Risser staging. Spine (Phila Pa 1976).
1995;20:1868-71.[PubMed][CrossRef]
Dhar
S;
Dangerfield
PH;
Dorgan
JC;
Klenerman
L. Correlation between bone age and Risser's sign in adolescent idiopathic scoliosis. Spine (Phila Pa 1976).
1993;18:14-9.[PubMed][CrossRef]
Goldberg
MS;
Poitras
B;
Mayo
NE;
Labelle
H;
Bourassa
R;
Cloutier
R. Observer variation in assessing spinal curvature and skeletal development in adolescent idiopathic scoliosis. Spine (Phila Pa 1976).
1988;13:1371-7.[PubMed] [CrossRef]