Skeletal maturity and remaining growth represent essential parameters that must be considered in the evaluation of the risk of curve progression in idiopathic scoliosis. Duval-Beaupère et al.1 as well as Lonstein and Carlson2 demonstrated that the main progression of idiopathic scoliosis occurs at the time of the most rapid adolescent skeletal growth, which occurs around eleven to thirteen years of skeletal age in girls and thirteen to fifteen years in boys. This growth phase corresponds to the accelerating phase of the pubertal growth diagram3,4 and is characterized by a gradual increase in the spinal growth rate5. Peak height velocity is a reliable clinical marker for the prediction of remaining growth and the risk of curve progression in idiopathic scoliosis6,7. Yet, during this growth phase, the Risser sign8 remains at grade 0. The Risser grade-1 sign appears only around a skeletal age of 13.5 years in girls and 15.5 years in boys3,4.
Most current clinical and radiographic markers do not help pediatric orthopaedic surgeons to readily distinguish maturity levels prior to Risser grade 1. Triradiate cartilage closure is closely related to peak height velocity7,9, with this period of accelerated spinal growth split into approximately two halves by the closure of the triradiate cartilage, which occurs around a skeletal age of twelve years in girls and fourteen years in boys3,10. Triradiate cartilage closure is mainly used in the preoperative evaluation of the risk for the development of a crankshaft phenomenon after posterior spinal fusion in patients with an immature spine and an open triradiate cartilage11-15.
A more detailed evaluation of skeletal maturity would be suitable during Risser grade 0, which corresponds to the phase of highest spinal growth velocity, in addition to a height measurement every six months and an assessment of the status of the triradiate cartilage. Precise information about maturity levels and remaining growth could help to better evaluate the efficacy of brace treatment, the curve progression risk, or the potential for the development of a crankshaft phenomenon. The method of Sauvegrain et al.16 for the assessment of skeletal age with use of radiographs of the elbow has been used in France over the last forty years, and a detailed analysis of this method has demonstrated its reliability during the time of peak growth velocity17. This method is used during the accelerating phase of the pubertal growth diagram3,4, from a skeletal age of eleven to thirteen years in girls and thirteen to fifteen years in boys, at which time the development of elbow ossification centers is clearly identified. Complete elbow ossification correlates with the end of accelerated growth velocity and marks the beginning of the decelerating growth phase. Derived from the Sauvegrain method, a simplified method described by Diméglio et al.4,17 is characterized by clear and regular morphological development of the olecranon during the accelerating growth phase. Assessment of this apophysis allows skeletal age to be evaluated very easily and accurately at intervals of six months.
The purpose of this study was to determine the accuracy of this simplified method of olecranon staging, to compare it with the time of triradiate cartilage closure, and to investigate its capacity to increase information on skeletal maturation during Risser grade 0 in patients with idiopathic scoliosis.
The medical records and radiographs of patients with juvenile or adolescent idiopathic scoliosis who were followed at our pediatric orthopaedic department between 1988 and 2006 were reviewed. From 561 complete and analyzable records, the charts of 100 boys and 100 girls were randomly selected for this retrospective study. The patients had to have been followed regularly every six months for a minimum of two years during the phase of peak height velocity with a Risser grade 0 (between eleven and thirteen years of skeletal age in girls and between thirteen and fifteen years in boys). The clinical and radiographic evolution of the scoliosis had to be well documented. Anteroposterior and lateral radiographs of the left elbow made to assess the skeletal age during this growth period had to be available. When more than one skeletal age assessment had been performed, only the first elbow radiograph was retained for the review in this study. Clinical growth parameters, such as standing height, sitting height, weight, and menarche, were documented on the checklist and the growth curve that is regularly used in our clinic. Patients with congenital scoliosis as well as scoliosis with an underlying neurological disorder, a syndrome, or an endocrinopathy were excluded from the study protocol. In juvenile idiopathic scoliosis, magnetic resonance imaging was carried out in addition to clinical and neurological examinations to detect any neural axis abnormality.
In the group of 100 boys, fifty-eight patients had juvenile scoliosis, with an onset of between four and ten years, and forty-two had adolescent scoliosis, with an onset after ten years. In the group of 100 girls, forty-nine patients had a juvenile scoliosis and fifty-one had an adolescent scoliosis.
Standing and sitting height were measured in centimeters every six months, including five consecutive measurements during two years, which allowed the calculation of average annual growth velocity for standing and sitting height, respectively, as follows: (last height measurement - first height measurement) / 2 (years). Age at menarche was also documented.
The Sauvegrain method is used shortly before the onset of puberty and during the first two years of the pubertal growth spurt, from a skeletal age of ten to thirteen years in girls and twelve to fifteen years in boys17. This method determines skeletal age from anteroposterior and lateral radiographs of the left elbow and is based on a 27-point scoring system, which assesses four ossification centers. The lateral condyle (1 to 9 points), the trochlea (1 to 5 points), the olecranon apophysis (1 to 7 points), and the proximal radial epiphysis (1 to 6 points) are rated according to their developmental stage. The scores are added together, and the total score allows skeletal age in years to be assessed on a graph for girls and a graph for boys. These graphs contain typical scores of skeletal ages at intervals of six months, which allow a clear and regular differentiation of skeletal age during the phase of peak height velocity.
Derived from the Sauvegrain method, the simplified olecranon method developed by Diméglio4,17 allows skeletal age in intervals of six months to be assessed easily and quickly. Five images of the olecranon are identified, with two ossification nuclei indicating eleven years of age in girls and thirteen years in boys; a half-moon image, 11.5 years in girls and 13.5 years in boys; a rectangular aspect, twelve years in girls and fourteen years in boys; the beginning of fusion, 12.5 years in girls and 14.5 years in boys; and a complete fusion, thirteen years in girls and fifteen years in boys (Fig. 1).
In order to compare this simplified method of skeletal age assessment from the olecranon with the entire Sauvegrain method, the elbow radiographs of the 100 girls and 100 boys with idiopathic scoliosis were reviewed once by three independent and experienced pediatric orthopaedic surgeons who had all routinely used both methods for a minimum of two years in their practice. No time limit was imposed for skeletal age determination with use of the Sauvegrain method or the olecranon method, which were used respectively in two separate review sessions. Prior to the review, the name and age of the patient were omitted from all radiographs, which were marked with consecutive numbers from one to 200 by an individual who did not participate in the study. Apart from gender, information about the patients was not accessible to any of the observers during the analysis of the radiographs. The skeletal age as well as the respective chronological age of the patient were then stored in a data file.
After the review of the elbow radiographs was completed, the three observers analyzed a series of five consecutive posteroanterior radiographs of the spine made when the patient was at Risser grade 0 during the growing period of peak height velocity. The time of triradiate cartilage closure was compared with the chronological age and the skeletal age from the olecranon. Triradiate cartilage closure was considered to correspond to stage 2 of the Oxford method18 for assessing skeletal age from the pelvis. The radiographic status of the triradiate cartilage was discussed by the three observers in order to minimize intraobserver and interobserver errors.
Statistical Analysis
The concordance correlation coefficient of Lin19 was used to analyze the concordance between skeletal age data assessed with the olecranon method and skeletal age data assessed with the Sauvegrain method for each observer separately. Furthermore, the concordance between the three observers was analyzed for each method of skeletal age assessment. According to the Landis and Koch scale for intraclass correlation coefficients20, the concordance was considered to be excellent if r > 0.80, good if r = 0.61 to 0.80, fair if r = 0.41 to 0.60, and poor if r = 0.40.
Standing Height Velocity, Sitting Height Velocity, and Menarche
Growth velocities of standing and sitting height were analyzed during the first two years of the pubertal growth spurt. Clinically, the beginning of this phase was detected by a sharp increase in growth velocity when the skeletal age was approximately eleven years in girls and thirteen years in boys.
In the boys, the average annual standing height velocity (and standard deviation) was 8.11 ± 0.88 cm per year (range, 6.0 to 10.0 cm per year). The annual average sitting height velocity was 4.18 ± 0.67 cm per year (range, 2.5 to 5.5 cm per year).
In the girls, the average annual standing height velocity was 7.22 ± 0.75 cm per year (range, 6.0 to 8.5 cm per year). The annual average sitting height velocity was 3.22 ± 0.62 cm per year (range, 2.0 to 4.5 cm per year). Menarche occurred at an average chronological age of 13.11 ± 1.03 years (range, 11.0 to 16.0 years).
Concordance Between the Olecranon and Sauvegrain Methods
An excellent concordance was noted between skeletal ages assessed by means of the olecranon method and skeletal ages assessed with use of the Sauvegrain method (average concordance, r = 0.977 for boys and r = 0.938 for girls). Correlation coefficients and respective confidence intervals in Table I demonstrate that this finding applies for each of the three observers equally. There was no difference when boys were compared with girls.
Interobserver Concordance of the Olecranon and Sauvegrain Methods
The correlation coefficients in Table II show that the concordance among the three observers was equally excellent for skeletal ages assessed by means of the olecranon method or by the Sauvegrain method. A comparison of the genders demonstrated no discrepancy between the skeletal ages of boys and girls.
Comparison of Chronological and Skeletal Age Data
In the boys, the average chronological age was 13.89 ± 1.01 years (range, 12.2 to 15.5 years), while the average skeletal age determined with the olecranon method was 13.82 ± 0.69 years (range, 13.0 to 15.0 years).
In the girls, the average chronological age was 12.01 ± 1.16 years (range, 10.0 to 14.3 years), while the average skeletal age was 12.17 ± 0.68 years (range, 11.0 to 13.0 years).
Chronological age and skeletal age assessed with the olecranon method were considered to be the same if the difference was less than six months. Skeletal age was defined as advanced if it exceeded chronological age by more than six months and as delayed if it was less than the chronological age by at least six months. Table III shows that chronological age and skeletal age corresponded to one another within a six-month range in 49% of the boys and 51% of the girls.
Triradiate Cartilage Closure
The time of triradiate cartilage closure was compared with the chronological age and with the skeletal age assessed with the olecranon method. The determination of triradiate cartilage closure and skeletal age was based on the consensus of the three observers who worked together in this part of the study. However, this review of the spine radiographs and all available elbow radiographs for each patient was completed only after the initial review of independent skeletal age-grading had been completed.
In the boys, the average chronological age at the time of triradiate cartilage closure was 14.55 ± 1.01 years (range, 12.4 to 16.2 years), and the average skeletal age was 14.02 ± 0.44 years (range, 13.5 to 15.0 years).
In the girls, the average chronological age at the time of triradiate cartilage closure was 12.08 ± 0.87 years (range, 10.3 to 13.8 years), while the average skeletal age was 12.11 ± 0.40 years (range, 11.5 to 13.0 years).
Triradiate cartilage closure occurred at the time of the appearance of the rectangular shape of the olecranon apophysis in 65% of the boys and 61% of the girls and corresponded to a skeletal age of fourteen years in the boys and twelve years in the girls (Fig. 1). In 91% of the boys and in 88% of the girls, triradiate cartilage closure occurred within six months before to six months after the appearance of the rectangular shape of the olecranon.