JBJS | Relationship of Peak Height Velocity to Other Maturity Indicators in Idiopathic Scoliosis in Girls*

The Journal of Bone & Joint Surgery, Volume 82, Issue 5

Articles | May 01, 2000

Relationship of Peak Height Velocity to Other Maturity Indicators in Idiopathic Scoliosis in Girls*

David G. Little, M.B.B.S., F.R.A.C.S.(Orth)†; Kit M. Song, M.D.‡; Don Katz, C.O.§; John A. Herring, M.D.§

Investigation performed at the Texas Scottish Rite Hospital for Children, Dallas, Texas
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Royal Alexandra Hospital for Children, P.O. Box 3515, Parramatta, New South Wales 2124, Australia. E-mail address for D. G. Little: davidl3@nch.edu.au.
‡Children's Hospital and Medical Center, 4800 Sand Point Way N.E., Seattle, Washington 98105-0371.
Â§Texas Scottish Rite Hospital for Children, 2222 Welborn Street, Dallas, Texas 75219.

The Journal of Bone & Joint Surgery. 2000; 82:685-685

5 Recommendations (Recommend) | 3 Comments | Saved by 3 Users Save Case

text A A A

Abstract

Background: Our aim was to compare height velocity data, obtained from clinical height measurements, for girls who had idiopathic scoliosis with the data for adolescents who did not have scoliosis. We also compared the growth data with chronological age, menarchal age, and Risser sign in terms of their accuracy in the prediction of growth and progression of the scoliosis.

Methods: One hundred and twenty of 371 patients in a database of girls managed with a brace for the treatment of idiopathic scoliosis had sufficient height data for us to quantify their growth peak. Height velocity data was generated from standing-height measurements obtained, in a scoliosis clinic, with a minimum six-month interval between measurements, and the timing of peak height velocity was calculated. The age at menarche was recorded from the patients' records. The Risser sign and Cobb angle were determined by a single observer. Progression of the scoliosis was defined as an increase in the Cobb angle of at least 10 degrees, compared with the curve magnitude at the time of the initial evaluation, after a minimum of six months. Progression to a magnitude requiring surgery was defined as progression of at least 10 degrees to a magnitude of 45 degrees or more.

Results: The height velocity plot grouped by peak height velocity showed a high peak and a sharp decline with values similar to those in normal populations. Extrapolating from percentile charts, 90 percent of our patients ceased growing by 3.6 years after peak height velocity. The growth peak was blunted (averaged over too long a period such that the data for the period of most rapid growth was averaged in with that for a period of slower growth) when chronological age, menarchal age, and Risser sign were used to predict growth; this indicated that these maturity scales grouped the patients poorly in terms of growth.

The primary curve was progressive in eighty-eight of the 120 patients. Sixty of these patients had a curve of more than 30 degrees at peak height velocity, and in fifty (83 percent) of the sixty the curve progressed to 45 degrees or more. The remaining twenty-eight patients had a curve of 30 degrees or less at peak height velocity, with only one curve (4 percent) progressing to 45 degrees or more.

Peak height velocity also grouped patients for maximal progression of the curve more accurately than did the other maturity scales, as most of the curves progressed maximally at peak height velocity. There was a wider spread of timing of maximal progression when chronological age, menarchal age, and Risser sign were used to predict progression.

Conclusions: Height velocities generated from clinical height measurements for patients with idiopathic scoliosis document the growth peak and predict cessation of growth reliably. Knowing the timing of the growth peak provides valuable information on the likelihood of progression to a magnitude requiring spinal arthrodesis.

Figures in this Article

Introduction

Introduction | Materials and Methods | Results | Discussion | References

It is widely accepted that curve progression in idiopathic scoliosis is related to growth^2,5,7,18. Recommendations for bracing and operative intervention rely on the ability to assess the growth potential of these patients. Chronological age, Risser sign, and menarchal age are commonly used to determine growth potential and the potential for future progression of the curve^15,17,21,28.

Buckler⁴ performed a longitudinal study of normal adolescents and reported a method of grouping children by chronological years around the age at which peak height velocity occurred. He found that these methods produced a growth velocity curve most representative of individual growth patterns and that males and females as well as early and late developers behaved similarly with respect to this maturity scale.

The purposes of our investigation were to establish whether height velocity data could be calculated retrospectively from clinical height measurements and to assess whether the growth pattern of girls with adolescent idiopathic scoliosis was similar to the pattern of normal individuals. In addition, we wished to compare chronological age, Risser sign, and menarchal age with the girls' actual growth data. The ability of each maturity scale to predict the timing of maximal progression was also assessed.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 1:: Graph showing the median values for height velocity from the current study of 120 girls with idiopathic scoliosis compared with values for girls with normal posture reported on by Buckler⁴. PHV = peak height velocity.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 2-A:: Figs. 2-A through 2-D: Graphs showing height velocity data for the 120 female patients with idiopathic scoliosis, plotted against each maturity scale. The values in parentheses are the numbers of patients with a height velocity measurement for each period.
Fig. 2-A: Graph showing the percentile plots for age at peak height velocity (PHV).

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 2-B:: Graph showing the percentile plots for chronological age.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 2-C:: Graph showing the percentile plots for the Risser sign.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 2-D:: Graph showing the percentile plots for menarchal age.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 3:: Plot of height velocity curve with median occurrence of menarche, age at menarche plus two years, and Risser grades 1, 4, and 5 in the 120 female patients with idiopathic scoliosis. PHV = peak height velocity.

View Larger

Table I: Method of Assigning Peak Height Velocity

*See the Materials and Methods section of the article for the details of this technique.

Date	Interval (yrs.)	Interval Used for Velocity (yrs.)	Height of Patient (cm)	Height Velocity (cm/yr.)	Age at Peak Height Velocity (yrs.)
May 2, 1988	-	-	148.0	-	-1.8
Jan. 30, 1989	0.75	0.75	152.0	5.3	-1.0
Feb. 12, 1990	1.04	1.04	162.5	10.1	0.0
Apr. 12, 1990	0.16	1.20	164.0	10.0	0.2
Aug. 13, 1990	0.34	0.50	165.5	6.0	0.5
Oct. 15, 1990	0.17	0.51	166.0	3.9	0.7
Feb. 11, 1991	0.33	0.50	168.0	5.0	1.0
Sept. 30, 1991	0.63	0.63	169.0	1.6	1.6
Apr. 27, 1992	0.58	0.58	169.0	0.0	2.2
May 10, 1993	1.04	1.04	169.0	0.0	3.2

Add to My JBJS

Table I: Method of Assigning Peak Height Velocity

*See the Materials and Methods section of the article for the details of this technique.

Date	Interval (yrs.)	Interval Used for Velocity (yrs.)	Height of Patient (cm)	Height Velocity (cm/yr.)	Age at Peak Height Velocity (yrs.)
May 2, 1988	-	-	148.0	-	-1.8
Jan. 30, 1989	0.75	0.75	152.0	5.3	-1.0
Feb. 12, 1990	1.04	1.04	162.5	10.1	0.0
Apr. 12, 1990	0.16	1.20	164.0	10.0	0.2
Aug. 13, 1990	0.34	0.50	165.5	6.0	0.5
Oct. 15, 1990	0.17	0.51	166.0	3.9	0.7
Feb. 11, 1991	0.33	0.50	168.0	5.0	1.0
Sept. 30, 1991	0.63	0.63	169.0	1.6	1.6
Apr. 27, 1992	0.58	0.58	169.0	0.0	2.2
May 10, 1993	1.04	1.04	169.0	0.0	3.2

View Larger

Table II: Percentages of Girls Whose Height Velocity Decreased to Less Than Two Centimeters per Year

	Percent
Peak height velocity
+3 years	81
+3.6 years	90
Age of patient
15 years	45
16 years	79
Risser sign
Grade 4	68
Grade 5	93
Menarche
+2 years	68
+2.7 years	90

Add to My JBJS

Table II: Percentages of Girls Whose Height Velocity Decreased to Less Than Two Centimeters per Year

	Percent
Peak height velocity
+3 years	81
+3.6 years	90
Age of patient
15 years	45
16 years	79
Risser sign
Grade 4	68
Grade 5	93
Menarche
+2 years	68
+2.7 years	90

View Larger

Table III: Timing of Maximal Progression in Eighty-eight Patients with Curve Progression

Maturity Scale	No. of Patients with Maximal Curve Progression
Age at peak height velocity
-3	0
-2	1
-1	11
0	48
1	21
2	5
3	2
Chronological age (yrs.)
9	1
10	2
11	12
12	23
13	27
14	14
15	8
16	1
Risser sign (grade)
0	59
1	10
2	6
3	3
4	9
5	1
Menarche (yrs.)
-3	2
-2	9
-1	28
0	32
1	14
2	3
3	0

Add to My JBJS

Table III: Timing of Maximal Progression in Eighty-eight Patients with Curve Progression

Maturity Scale	No. of Patients with Maximal Curve Progression
Age at peak height velocity
-3	0
-2	1
-1	11
0	48
1	21
2	5
3	2
Chronological age (yrs.)
9	1
10	2
11	12
12	23
13	27
14	14
15	8
16	1
Risser sign (grade)
0	59
1	10
2	6
3	3
4	9
5	1
Menarche (yrs.)
-3	2
-2	9
-1	28
0	32
1	14
2	3
3	0

Materials and Methods

Introduction | Materials and Methods | Results | Discussion | References

We reviewed the charts and radiographs of girls included in a database of 371 female patients who had been managed with a brace and followed from 1977 to 1995. Patients with idiopathic scoliosis who had height measurements and documented follow-up at least from the time that the Risser sign was grade 0 to the time that it was grade 4 were selected. Additional criteria for inclusion in the study were a minimum of six visits to our institution over a minimum of 2.5 years, a curve magnitude of 25 to 45 degrees as measured with the Cobb method, and treatment with either a Boston thoracolumbosacral orthosis or a Charleston bending brace. The patient's age, height, menarchal age, Risser sign, Cobb angle, type of orthosis, and dates and types of surgical procedures were recorded for each visit. The Cobb angle and Risser sign were measured from radiographs for the present study by one of us (D. K.). All radiographs were made with the patient standing and not wearing the brace. In patients who had more than one spinal curve, the primary curve was designated as the largest curve at the time of the most recent follow-up.

Height measurements were recorded at each clinic visit by the nursing staff. The protocol dictated that the patient be measured standing, without shoes, and that the height be recorded in centimeters with a wall-mounted ruler with a perpendicular slide. Different nurses measured the patients, who attended clinics as dictated by their clinical follow-up needs.

Height measurements for each patient were used to calculate height velocities in centimeters per year. For this purpose, a year is expressed as a decimal value (six months equals 0.5 year). One year is assumed to be 365 days. Height velocity is thus calculated as growth (in centimeters) during the time-interval divided by the time-interval (in years).

An example of the calculation for an individual patient is highlighted in Table I. A minimum interval of six months (183 days) is used. The patient was seen in February, April, and August of 1990. When the height velocity for August was calculated, the height measurement for April was skipped over, as it was made after an interval of less than six months. The use of a minimum six-month period for the calculation of height velocity reduces the effects of measurement error in the values for height and reflects the methods used by Buckler⁴ and Anderson et al.¹. The use of a larger time-interval can blunt the growth peak by inclusion of data over too long a period such that the period of most rapid growth is averaged in with a period of slower growth. Use of the height measurement from February allows calculation of a height velocity in centimeters per year over a six-month interval. There was 3.0 centimeters of growth over the six months, yielding a height velocity of 6.0 centimeters per year. Peak height velocity for the patient in Table I occurred in February 1990, when the height velocity was 10.1 centimeters per year. We then assigned a negative value to the years preceding and a positive value to the years following peak height velocity (PHV -1, PHV +1, and so on). If the first height velocity value available for a patient was the largest and was also more than nine centimeters per year, this was designated as peak height velocity and the patient was included in the study. This assumption was based on the fact that nine centimeters per year is two standard deviations from median values six months before and after the growth peak⁴. Any patients in whom the growth peak, as defined above, was not recorded were specifically excluded.

Height velocity was plotted against four maturity scales: chronological age, age at peak height velocity, Risser sign, and menarchal age. Percentile plots were generated by calculating the tenth, twenty-fifth, fiftieth (median), seventy-fifth, and ninetieth percentiles for a given time-period. For example, the median height velocity at the menarchal age plus one year was calculated from data at the menarchal age plus 0.5 year to the menarchal age plus 1.5 years. For menarche and peak height velocity, six monthly values were calculated immediately around those milestones.

We defined cessation of growth as a height velocity of less than two centimeters per year. A height velocity of two centimeters per year was shown by Buckler⁴ to be within one centimeter of complete cessation of growth. Also, a measurement error of one centimeter in height will produce a velocity over six months of two centimeters per year. We believed that this was the closest that we could come to accurately determining a reasonable end point of growth in our study.

We used the regression model of Bjure et al.³ to correct for the patients' decrease in height due to the curvature: log y = 0.011x - 0.177, where y is the loss in height (centimeters) caused by a curve magnitude of x (degrees as determined with the Cobb method). This method addresses the loss in height from the increase in the Cobb angle in the coronal plane as seen on posteroanterior radiographs. This allowed estimation of the amount that progression of the curve may have affected the height of the individuals and thus their height velocity.

We defined progressive curves as those increasing by at least 10 degrees during the follow-up interval⁶. The timing of maximal progression of the progressive curves was identified and was compared, with use of an F test for variance, among the maturity scales to ascertain which scale grouped this most tightly.

We defined progression of at least 10 degrees, to 45 degrees or more, as progression to a magnitude requiring surgery (a surgical magnitude). This definition is based on our clinical experience in recommending surgery at curve magnitudes of 45 degrees or more. We recognize that progression to 45 degrees or more is not an absolute indication for surgery and that recommendations vary among surgeons and patients. Our definition serves only to define a common end point.

We evaluated the extent of progression to a surgical magnitude relative to the size of the primary curve at peak height velocity. On the basis of the work of Duval-Beaupç±¥⁹, we hypothesized that determination of the curve magnitude at peak height velocity may aid in the identification of curves at high risk of progression and of leading to a spinal arthrodesis.

Results

Introduction | Materials and Methods | Results | Discussion | References

One hundred and twenty patients met all of the entry criteria. Of these patients, thirty-eight were included because their first recorded height velocity was more than nine centimeters per year. Overall, 1227 data points on height velocity were generated, with a mean of 10.2 data points per patient. The median duration of follow-up was 5.0 years (range, 2.5 to 12.5 years). The plot for median peak height velocity retained the characteristics of an individual plot, with a high peak and sharp decline, and was comparable with the plot of data generated for normal girls by Buckler⁴ (Fig. 1). The only area of disagreement between the two sets of data was that the median peak height velocity in the current study was higher.

According to the regression model of Bjure et al.³, the mean loss of height (and standard deviation) of the patients in our study was 1.4 ± 0.5 centimeters. Accordingly, a 30-degree curve results in a loss of trunk height of approximately 1.4 centimeters. A 40-degree curve results in a loss of 1.8 centimeters. If a curve of 30 degrees were to progress to 40 degrees in a six-month interval, the height increase would be reduced by 0.4 centimeter in six months (0.8 centimeter per year). Thus, even rapid progression of moderate curves blunts the growth peak only slightly. Progression did not noticeably blunt the growth peak in our study; the peak was detectable for all patients, with a median peak height velocity of 9.5 centimeters per year.

The percentile plot for peak height velocity showed a sharp peak and decline (Fig. 2-A); however, in the graphs for chronological age (Fig. 2-B), Risser sign (Fig. 2-B), and menarchal age (Fig. 2-D) the peak was blunted, indicating that girls growing quickly were grouped with those growing slowly. The interquartile ranges (from the twenty-fifth to the seventy-fifth percentiles) at the maximum height velocity on the graphs were 2.6 for peak height velocity, 4.1 for chronological age, 3.7 for Risser sign, and 3.7 for menarchal age.

The percentile line crossing at two centimeters per year, which we defined as growth cessation in the present study, was calculated for the commonly used milestones of Risser signs of grades 4 and 5 and menarchal age plus two years. The ninetieth percentile for height was followed until it crossed below two centimeters per year to predict with 90 percent certainty that growth had ceased. It is at least 90 percent certain that a patient has ceased growing when he or she has reached the age at peak height velocity plus 3.6 years, the age at menarche plus 2.7 years, or a Risser sign of grade 5 (Table II). In our study, chronological age was not helpful in determining cessation of growth, as 55 percent of the girls who were fifteen years old and 21 percent of those who were sixteen years old had a growth velocity of more than two centimeters per year.

The mean time for the Risser sign to proceed from grade 1 to 4 was 1.4 ± 0.8 years, and the mean time for the Risser sign to proceed from grade 4 to 5 was 1.7 0.9 years. The first sign of ossification of the iliac apophysis was noted at a median of nine months after the growth peak (peak height velocity plus 0.75 year). Eighty-five percent (102) of the 120 patients were at or past their growth peak by the time that the iliac apophysis became visible.

Menarche occurred at a median of seven months after the growth spurt (peak height velocity plus 0.6 year). It occurred between six months before and six months after peak height velocity in 41 percent of the girls. Thirteen of the girls had menarche two or more years after peak height velocity. The growth pattern of these sixteen patients clearly followed peak height velocity rather than menarchal age. The mean height velocity for twelve of these patients at menarche was 2.4 centimeters per year, indicating that they were near the end of growth at menarche. The other four patients had surgery before menarche.

Eighty-eight (73 percent) of the 120 curves were progressive. Comparison with use of the F test showed that the timing of maximal progression was significantly more closely grouped around peak height velocity than around a specific chronological age (p = 0.0001) or menarchal age (p = 0.02) (Table III). Fifty-nine (67 percent) of the eighty-eight curves had maximal progression while the Risser sign was still grade 0, but ten (11 percent) of the eighty-eight had maximal progression when the Risser sign was grade 4 or grade 5. As the Risser data was skewed and the units were ordinal, an F test was not applied.

Fifty-one (58 percent) of the eighty-eight progressive curves increased by at least 10 degrees to a magnitude of 45 degrees or more (progression to a surgical magnitude). The mean magnitude at peak height velocity for curves progressing to 45 degrees or more was 34.6 ± 7.1 degrees; the mean for curves not progressing to at least 45 degrees was 25.7 ± 6.9 degrees. This difference was highly significant according to the t test (p < 0.001). Sixty (68 percent) of the eighty-eight progressive curves had a magnitude of more than 30 degrees at peak height velocity, whereas the remaining twenty-eight were 30 degrees or less. We found that fifty (83 percent) of the sixty progressive curves that were more than 30 degrees at peak height velocity increased to a surgical magnitude, whereas only one (4 percent) of the twenty-eight that were 30 degrees or less at peak height velocity did so. This was a significant difference according to Fisher's exact test (p < 0.001). This data generates a sensitivity of 98 percent (fifty of fifty-one), a specificity of 73 percent (twenty-seven of thirty-seven), and an accuracy of 88 percent (seventy-seven of eighty-eight) for the statement that progressive curves of more than 30 degrees at peak height velocity will progress to 45 degrees or more, whereas curves of 30 degrees or less at peak height velocity, although progressive, will not increase to a surgical magnitude.

All fifty-one curves that progressed 10 degrees to 45 degrees or more were more than 30 degrees by menarche (sensitivity, 100 percent). However, this indicator is nonspecific, as seventeen (46 percent) of thirty-seven curves that did not progress to a surgical magnitude were also 30 degrees or more by menarche (specificity, 54 percent; accuracy, 81 percent).

Discussion

Introduction | Materials and Methods | Results | Discussion | References

The aim of all maturity scales is to group children in such a way that they behave similarly with regard to growth and development. Peak height velocity is a maturity scale directly related to growth and the adolescent growth spurt. It uses the time of the growth peak as the fixed point by which the patients are grouped. Buckler⁴ showed that this method groups males and females as well as early and late developers so that their growth patterns are very similar and the composite graph produced is representative of a typical individual growth pattern. Buckler also showed that more than 90 percent of both male and female subjects are within 2.5 centimeters of their final adult height three years after the growth peak (peak height velocity plus three years). Our investigation confirmed similar timing for patients with idiopathic scoliosis. We found a 90 percent probability of cessation of growth at peak height velocity plus 3.6 years. Once peak height velocity is identified, the timing of cessation of growth can be reliably predicted, so that patients can be told well in advance how long they will need to wear an orthosis. The graph generated with clinical height measurements showed a higher peak height velocity in these patients than in individuals with normal posture. In a longitudinal study, Lonç ²-Duek et al.¹⁴ also found significantly increased peak height velocity in children in whom scoliosis developed compared with children in whom it did not. The interquartile range in our charts (2.6) was higher than that reported by Buckler⁴ (1.8), indicating greater variability. This difference is likely explained by increased measurement error, although the possibility remains that this is a real effect in patients with idiopathic scoliosis. The height measurements in our study were made by multiple observers in a clinical setting, whereas the measurements in Buckler's study were all made by a single observer in a research setting. The interquartile range of height velocity at peak height velocity was far smaller than the range with the other scales, again indicating a better grouping of patients by this scale, regardless of error in height measurement.

A possible criticism of the use of height measurements of patients with idiopathic scoliosis is that progression of the curve reduces their height and blunts the growth peak. The regression model of Bjure et al.³ attempts to address the loss of height caused by increased deformity in the coronal plane. According to this formula, peak height velocity is diminished by only 0.8 centimeter per year in a patient with a curve progressing from 30 to 40 degrees in six months. Eighty-eight (73 percent) of the 120 curves in this series progressed at least 10 degrees during the period in which the height measurements were made, yet this did not blunt the growth peak. Our results confirm that it is possible to construct height velocity plots accurately even with progression of the curve.

The Risser sign is widely used to assess spinal maturity, but the shortcomings of this approach have been recently reported^13,22. The time taken to proceed from Risser grade 1 to Risser grade 4 or 5 is highly variable, as shown in the present study and others^19,21,26,29. This variation makes predictions of the cessation of growth unreliable. Our data does show that the Risser sign has some value in determining when growth has ceased. Two-thirds of female patients have ceased growing by the time that the Risser sign is grade 4, but to be 90 percent certain of cessation of spinal growth the clinician must wait an average of another nineteen months for the Risser sign to reach grade 5. These values are in agreement with the work of Anderson et al.¹, who noted a mean of 1.2 centimeters of trunk growth in girls and 2.5 centimeters in boys after the appearance of grade 4.

Most of the eighty-eight patients exhibited maximal curve progression during the period when the Risser sign was grade 0. As most patients pass in and out of the growth peak during this stage, this finding is not surprising. However, 11 percent of the patients had their maximum point of curve progression after the Risser sign had reached at least grade 4. Apparently, late progression did not occur with the other maturity scales; thus, we believe that the Risser sign is an inferior maturity indicator for defining the risk of progression of scoliosis. Only one patient had maximal curve progression after the Risser sign had reached grade 5, and this patient's curve progression was not predicted by any of the scales. Waiting for the appearance of grade 5 is safe, but it occurs at an indefinite time (average, 1.7 years after grade 4)^{13,19,21,26,29}.

Menarche is often used by clinicians as a reliable point in time on which to base predictions of spinal growth in girls. However, it occurs at varying stages of puberty and may be delayed in girls who are active in sports. The present study confirmed a relationship between menarche and the growth spurt in patients with idiopathic scoliosis. Forty-one percent of the girls had the onset of menses within one year of peak height velocity, but 59 percent had menarche outside this time-period. Though the median peak height velocity in our study was 9.5 centimeters per year, the median peak according to menarchal age was 6.8 centimeters per year, occurring one year prior to menarche. This indicates that when menarchal age was used to predict growth, girls growing rapidly were grouped with those growing slowly, severely blunting the growth peak. The results of previous studies have also suggested that menarche occurs later in patients who have scoliosis than in those who do not^8,16,27, although the authors of one study concluded that it occurs earlier¹⁰. Sixteen (13 percent) of the 120 girls in our investigation had menarche more than two years after peak height velocity was reached. In these patients, growth patterns followed those predicted by peak height velocity and not those predicted by menarche. The mean height velocity at menarche for these girls (excluding four who had surgery before menarche) was 2.4 centimeters per year, indicating that growth had nearly ceased. The use of menarche as a maturity indicator in these girls could have led to the initiation of bracing when growth had in fact ceased. It is our belief that peak height velocity is a better predictor of spinal growth in patients with delayed menarche and in fact serves to identify this group.

Menarche proved to be a reasonable landmark for predicting progression in the present study. A curve magnitude of more than 30 degrees at menarche was a sensitive predictor of progression to a surgical magnitude. However, as menarche occurs at a median of seven months after peak height velocity, the landmark was much less specific than peak height velocity. According to our data, curves are highly unlikely to progress maximally more than two years after menarche.

Tanner stages^24,25 have been used by orthopaedists to estimate maturity of patients with scoliosis, but they were not available for use in the present study. Buckler⁴ showed that Tanner stages reflect growth in an indirect fashion, with a spread of stages over the growth peak, and that the stages have a different relationship to growth in boys and girls. This was also shown by Lonç ²-Duek et al.¹⁴.

Skeletal age and spinal growth were studied by Anderson et al.¹. They noted little difference in the plots of skeletal age and chronological age, and they noted marked blunting of the skeletal age curve. They emphasized that individual peaks in growth velocity were much higher than the mean for a given skeletal age. Anderson et al. also noted that the peak in spinal growth was seen over a range of four skeletal age years. Other authors^4,7,11 showed that spinal growth continues after closure of the limb physes, including those of the hand and wrist. It has also been noted that skeletal age is delayed in patients with idiopathic scoliosis⁸. This finding indicates that knowing a patient's skeletal age tells the clinician little about present and future spinal growth. We did not have skeletal ages available for our study, but we believe that the results of Anderson et al. and those of others show it to have a weak relationship with spinal growth.

The timing of maximal curve progression has been shown by this and other investigations to be associated with the adolescent growth peak, with a marked reduction in progression after growth has ceased^2,5,7,18,28. The present study showed that peak height velocity groups girls with progressive idiopathic scoliosis significantly better for maximal progression than do the other maturity scales. There was a larger spread in the timing of maximal progression when chronological age, Risser sign, and menarche were used to define the time scale. For a maturity scale to be useful, curves should behave in a similar fashion at a similar time. The majority of the curves in our study progressed maximally at peak height velocity. Most of the curves progressed when the Risser sign was grade 0 but there was a large spread, out to grade 5. As Risser grade 0 does not represent one point in time, the scale does not aid the clinician by alerting him or her that progression may be imminent.

An important finding in the present study is that a curve magnitude of more than 30 degrees at peak height velocity is highly prognostic of progression to a surgical magnitude. This information is complementary to that derived from the investigations by Lonstein and Carlson¹⁵ and Peterson et al.¹⁷, which allows a prognosis, on the patient's first visit, about the risk of progression of the curve and thus about the need for a brace. Our data is not applicable to a patient on the first visit but will be useful in advising patients being managed with a brace for the treatment of a progressive curve about the likelihood of progression to a surgical magnitude. The sensitivity, specificity, and accuracy of peak height velocity and curve magnitude in predicting progression to a surgical magnitude while the patient is wearing a brace are similar to those of the complex logistic regression model of Peterson et al.¹⁷ in predicting progression of 6 degrees or more in patients under observation.

Further study will be required to assess the applicability of peak height velocity to male patients with idiopathic scoliosis, as the present study included only girls. Male patients often have late progression, and the Risser sign is even less reliable than it is for female patients^12,23. As males have no menarche, there is no good way of predicting remaining growth. Timing of peak height velocity applied equally well to normal males and females in Buckler's study⁴. Validation of a similar relationship in individuals with idiopathic scoliosis would be extremely valuable.

Our retrospective study has certain limitations. The height measurements that we used were not specifically collected for the study, and the patients' heights were measured at varying intervals by multiple observers. Despite this, the assigned peak height velocity generated a reliable growth curve and predicted cessation of growth and timing of maximal progression better than the other scales. Height measurements made in a clinical setting are clearly subject to error; however, the rapidity of onset and the magnitude of the adolescent growth spurt vastly diminish the importance of this error.

We used height measurements made with the patient standing to determine peak height velocity. This peak is really a summation of two other peaks, that of limb growth and that of spinal growth, which occur nearly six months apart^4,7. However, the peak in limb growth is smaller than that in spinal growth, which occurs later. Spinal growth continues after limb growth has ceased, so any prediction of the cessation of growth is equivalent to a prediction of the end of spinal growth¹¹. We do not believe that measurements made with the patient sitting would add to the usefulness of the system, as these measurements are more prone to error than those made with the patient standing^4,18.

It may not be as easy to assign peak height velocity prospectively for patients as it was to assign this parameter in this retrospective review. It is necessary to see a patient three times to obtain two velocities and observe a trend. Height measurements may be obtained from the family or primary-care physician. If only two height measurements and one velocity are available and the velocity is more than nine centimeters per year, there is approximately a 90 percent probability that the patient is in the growth peak, as these values are above the ninetieth percentile for velocity outside the peak (Fig. 2-A). According to our data, menarche and the first appearance of the iliac apophysis occur at a median of seven and nine months, respectively, after peak height velocity. If the iliac apophysis is visible, there is an 85 percent probability that the patient is at or beyond peak height velocity. The median growth velocity curve is schematically related to other median values for maturity indicators (Fig. 3).

The present study has established that it is possible to construct height velocity curves for patients with idiopathic scoliosis that are comparable with the curves for normal patients, even if the scoliosis is progressive. Use of peak height velocity and curve magnitude to predict curves at risk of progression to a surgical magnitude will be of major research value. We believe that future studies purporting to measure brace effectiveness should also include an analysis of the actual growth of the patients in terms of the timing of peak height velocity rather than in terms of the indirect information provided by menarchal age and the Risser sign. Height velocity data has also been shown to predict the occurrence of the crankshaft phenomenon better than twenty other available methods²⁰.

Knowing that a patient's height velocity is increasing tells the clinician that the curve is at high risk of progression, and, when the timing of peak height velocity is known, accurate prediction of the cessation of growth and likely cessation of progression can be made. This retrospective evaluation provides the basis for the prospective research required to validate whether measurement of height velocity will be of value in the clinical care of patients with idiopathic scoliosis.

References

Introduction | Materials and Methods | Results | Discussion | References

Anderson, M.; Hwang, S.-C.; and Green, W. T.: Growth of the normal trunk in boys and girls during the second decade of life. Related to age, maturity, and ossification of the iliac epiphyses. J Bone Joint Surg,47-A: 1554-1564, Dec 1965.47-A1554 1965

Ascani, E.; Bartolozzi, P.; Logroscino, C. A.; Marchetti, P. G.; Ponte, A.; Savini, R.; Travaglini, F.; Binazzi, R.; and Di Silvestre, M.: Natural history of untreated idiopathic scoliosis after skeletal maturity. Spine,11: 784-789, 1986.11784 1986 [PubMed]

Bjure, J.; Grimby, G.; and Nachemson, A.: Correction of body height in predicting spirometric values in scoliotic patients. Scandinavian J. Clin. and Lab. Invest.,21: 191-192, 1968.21191 1968

Buckler, J. M. H.: A Longitudinal Study of Adolescent Growth. London, Springer, 1990.

Bunnell, W. P.: The natural history of idiopathic scoliosis before skeletal maturity. Spine,11: 773-776, 1986.11773 1986 [PubMed]

Carman, D. L.; Browne, R. H.; and Birch, J. G.: Measurement of scoliosis and kyphosis radiographs. Intraobserver and interobserver variation. J Bone Joint Surg,72-A: 328-333, March 1990.72-A328 1990

DiMeglio, A., and Bonnel, F.: Le Rachis en Croissance. Paris, Springer, 1990.

Drummond, D. S., and Rogala, E. J.: Growth and maturation of adolescents with idiopathic scoliosis. Spine,5: 507-511, 1980.5507 1980 [PubMed]

Duval-Beaupç±¥, G.: La croissance des scoliotiques: hypothç²¥ et è³µde prè«©minaire. Acta Orthop. Belgica,38: 365-376, 1972.38365 1972

Goldberg, C. J.; Dowling, F. E.; and Fogarty, E. E.: Adolescent idiopathic scoliosis - early menarche, normal growth. Spine,18: 529-535, 1993.18529 1993 [PubMed]

Howell, F. R.; Mahood, J. K.; and Dickson, R. A.: Growth beyond skeletal maturity. Spine,17: 437-440, 1992.17437 1992 [PubMed]

Karol, L. A.; Johnston, C. E., II; Browne, R. H.; and Madison, M.: Progression of the curve in boys who have idiopathic scoliosis. J Bone Joint Surg,75-A: 1804-1810, Dec 1993.75-A1804 1993

Little, D. G., and Sussman, M. D.: The Risser sign: a critical analysis. J. Pediat. Orthop.,14: 569-575, 1994.14569 1994

Loncar-Duek, M.; Pecina, M.; and Prebeg, Z.: A longitudinal study of growth velocity and development of secondary gender characteristics versus onset of idiopathic scoliosis. Clin. Orthop.,270: 278-282, 1991.270278 1991 [PubMed]

Lonstein, J. E., and Carlson, J. M.: The prediction of curve progression in untreated idiopathic scoliosis during growth. J Bone Joint Surg,66-A: 1061-1071, Sept 1984.66-A1061 1984

Normelli, H.; Sevastik, J.; Ljung, G.; Aaro, S.; and Jonsson-Soderstrom, A. M.: Anthropometric data relating to normal and scoliotic Scandinavian girls. Spine,10: 123-126, 1985.10123 1985 [PubMed]

Peterson, L.-E.; Nachemson, A. L.; and Members of the Brace Study Group of the Scoliosis Research Society: Prediction of progression of the curve in girls who have adolescent idiopathic scoliosis of moderate severity. Logistic regression analysis based on data from the brace study of the Scoliosis Research Society. J Bone Joint Surg,77-A: 823-827, June 1995.77-A823 1995

Risser, J. C., and Ferguson, A. B.: Scoliosis: its prognosis. J Bone Joint Surg,18: 667-670, July 1936.18667 1936

Risser, J. C.: The iliac apophysis: an invaluable sign in the management of scoliosis. Clin. Orthop.,11: 111-119, 1958.11111 1958 [PubMed]

Sanders, J. O.; Little, D. G.; and Richards, B. S.: Prediction of the crankshaft phenomenon by peak height velocity. Spine,22: 1352-1357, 1997.221352 1997 [PubMed]

Scoles, P. V.; Salvagno, R.; Villalba, K.; and Riew, D.: Relationship of iliac crest maturation to skeletal and chronologic age. J. Pediat. Orthop.,8: 639-644, 1988.8639 1988

Shuren, N.; Kasser, J. R.; Emans, J. B.; and Rand, F.: Reevaluation of the use of the Risser sign in idiopathic scoliosis. Spine,17: 359-361, 1992.17359 1992 [PubMed]

Suh, P. B., and MacEwen, G. D.: Idiopathic scoliosis in males. A natural history study. Spine,13: 1091-1095, 1988.131091 1988 [PubMed]

Tanner, J. M.; Whitehouse, R. H.; and Takaishi, M.: Standards from birth to maturity for height, weight, height velocity, and weight velocity: British children, 1965. I. Arch. Dis. Child.,41: 454-471, 1966.41454 1966 [PubMed]

Tanner, J. M.; Whitehouse, R. H.; and Takaishi, M.: Standards from birth to maturity for height, weight, height velocity, and weight velocity: British children, 1965. II. Arch. Dis. Child.,41: 613-635, 1966.41613 1966 [PubMed]

Urbaniak, J. R.; Schaefer, W. W.; and Stelling, F. H., III: Iliac apophyses: prognostic value in idiopathic scoliosis. Clin. Orthop.,116: 80-85, 1976.11680 1976 [PubMed]

Warren, M. P.; Brooks-Gunn, J.; Hamilton, L. H.; Warren, L. F.; and Hamilton, W. G.: Scoliosis and fractures in young ballet dancers. Relation to delayed menarche and secondary amenorrhea. New England J. Med.,314: 1348-1353, 1986.3141348 1986

Weinstein, S. L., and Ponseti, I. V.: Curve progression in idiopathic scoliosis. J Bone Joint Surg,65-A: 447-455, April 1983.65-A447 1983

Zaoussis, A. L., and James, J. I. P.: The iliac apophysis and the evolution of curves in scoliosis. J Bone Joint Surg,40-B(3): 442-453, 1958.40-B(3)442 1958

Topics

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 2-B:: Graph showing the percentile plots for chronological age.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 2-C:: Graph showing the percentile plots for the Risser sign.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 2-D:: Graph showing the percentile plots for menarchal age.

View Large | Download Slide(.ppt)
Add to My JBJS

View Larger

Table I: Method of Assigning Peak Height Velocity

*See the Materials and Methods section of the article for the details of this technique.

Date	Interval (yrs.)	Interval Used for Velocity (yrs.)	Height of Patient (cm)	Height Velocity (cm/yr.)	Age at Peak Height Velocity (yrs.)
May 2, 1988	-	-	148.0	-	-1.8
Jan. 30, 1989	0.75	0.75	152.0	5.3	-1.0
Feb. 12, 1990	1.04	1.04	162.5	10.1	0.0
Apr. 12, 1990	0.16	1.20	164.0	10.0	0.2
Aug. 13, 1990	0.34	0.50	165.5	6.0	0.5
Oct. 15, 1990	0.17	0.51	166.0	3.9	0.7
Feb. 11, 1991	0.33	0.50	168.0	5.0	1.0
Sept. 30, 1991	0.63	0.63	169.0	1.6	1.6
Apr. 27, 1992	0.58	0.58	169.0	0.0	2.2
May 10, 1993	1.04	1.04	169.0	0.0	3.2

Add to My JBJS

Table I: Method of Assigning Peak Height Velocity

*See the Materials and Methods section of the article for the details of this technique.

Date	Interval (yrs.)	Interval Used for Velocity (yrs.)	Height of Patient (cm)	Height Velocity (cm/yr.)	Age at Peak Height Velocity (yrs.)
May 2, 1988	-	-	148.0	-	-1.8
Jan. 30, 1989	0.75	0.75	152.0	5.3	-1.0
Feb. 12, 1990	1.04	1.04	162.5	10.1	0.0
Apr. 12, 1990	0.16	1.20	164.0	10.0	0.2
Aug. 13, 1990	0.34	0.50	165.5	6.0	0.5
Oct. 15, 1990	0.17	0.51	166.0	3.9	0.7
Feb. 11, 1991	0.33	0.50	168.0	5.0	1.0
Sept. 30, 1991	0.63	0.63	169.0	1.6	1.6
Apr. 27, 1992	0.58	0.58	169.0	0.0	2.2
May 10, 1993	1.04	1.04	169.0	0.0	3.2

View Larger

Table II: Percentages of Girls Whose Height Velocity Decreased to Less Than Two Centimeters per Year

	Percent
Peak height velocity
+3 years	81
+3.6 years	90
Age of patient
15 years	45
16 years	79
Risser sign
Grade 4	68
Grade 5	93
Menarche
+2 years	68
+2.7 years	90

Add to My JBJS

Table II: Percentages of Girls Whose Height Velocity Decreased to Less Than Two Centimeters per Year

	Percent
Peak height velocity
+3 years	81
+3.6 years	90
Age of patient
15 years	45
16 years	79
Risser sign
Grade 4	68
Grade 5	93
Menarche
+2 years	68
+2.7 years	90

View Larger

Table III: Timing of Maximal Progression in Eighty-eight Patients with Curve Progression

Maturity Scale	No. of Patients with Maximal Curve Progression
Age at peak height velocity
-3	0
-2	1
-1	11
0	48
1	21
2	5
3	2
Chronological age (yrs.)
9	1
10	2
11	12
12	23
13	27
14	14
15	8
16	1
Risser sign (grade)
0	59
1	10
2	6
3	3
4	9
5	1
Menarche (yrs.)
-3	2
-2	9
-1	28
0	32
1	14
2	3
3	0

Add to My JBJS

Table III: Timing of Maximal Progression in Eighty-eight Patients with Curve Progression

Maturity Scale	No. of Patients with Maximal Curve Progression
Age at peak height velocity
-3	0
-2	1
-1	11
0	48
1	21
2	5
3	2
Chronological age (yrs.)
9	1
10	2
11	12
12	23
13	27
14	14
15	8
16	1
Risser sign (grade)
0	59
1	10
2	6
3	3
4	9
5	1
Menarche (yrs.)
-3	2
-2	9
-1	28
0	32
1	14
2	3
3	0