This study examines the time of occurrence of peak angular velocity in a group of girls with AIS and looks for correlations between height velocity and angle velocity.

Before attempting any further discussion of these findings, some limitations of the present study should be noted. Many shortcomings appear whenever scoliosis and/or growth are studied [14, 16–20] and they are probably the reason for the shortage of prospective and conclusive studies on this subject. Prospective works are difficult to perform because of the long follow-up period and this is why most papers in medical literature are transversal studies or retrospective cohort studies. The work by Nachemson *et al* [14, 20] deserves especial mention as the only prospective work on the effectiveness of bracing in the treatment of AIS.

It is very difficult to form homogeneous samples as growth in girls with the same chronological age is very variable. Little [7] demonstrated that the height pattern is similar among girls with AIS when they are grouped according to PHV. Considering that in many cases, this PHV is unknown, for the purpose of this study girls were grouped using the date of menarche as the reference for comparing patients. Although mean age of diagnosis for this sample was 11.6 years (approximately 1 year before occurrence of menarche (mean 12.8, DE 1.2)), note that diagnosis was made in first evaluation 1 or 2 years after menarche in 50% of girls. Our results, as well as the data presented by Little [7] show that menarche occurs one year after PHV. Therefore, considering that the mean age of diagnosis was 11.6 years, this means that at the beginning of follow-up, the PHV had already taken place in 50% of cases.

Contrary to what the first observations [21] seemed to indicate, curve progression in idiopathic scoliosis is not linear. It is frequent to observe curves that remain stable for a time and from a point, and then start to progress. On the other hand, it is not unusual to observe short periods of progression in a growth spurt followed by stabilization. When studying angular velocities, a repeated angle increase of 0 was frequently observed, even in girls with progressive curves. This represented a handicap for the use of both parametric and non-parametric statistical tests and specially, those for univariate analysis. This conditioned the fact that, when searching for a correlation, calculations of r and R^{2} were not feasible and the authors had to make an approximation by using absolute values of height and angle instead of the height and angle velocities [10].

The authors would like to mention the inference of the brace in this sample. Although the period after menarche is a phase of growth deceleration, the effect of the brace is, in our opinion, partially responsible for the deceleration velocity after menarche in braced curves. As observed in Figures 1 and 3, there is an important decrease in angular velocities in the perimenarchal period followed by a second growth spurt which occurs during the 6-12 months after menarche. Even though a period called *âge heureux* (happy period) consisting of a period of a relative stabilization of the curve, followed by a second growth spurt has been described [22], we think that in the sample, this decreasing velocity was due to the brace effect. It is not unusual to observe important initial reductions with the use of the brace, and in posterior controls, to find that the curve has returned to its initial values. If the second peak of PAV observed in graphs correlates to the cessation of bracing in these patient is another question that deserves further research.

In this study, graphic assessment suggests a relationship between height growth velocity and curve progression velocity, though frequent progression in spurts did not allow a statistical relationship to be found between angle and height velocities. Nevertheless, this problem was partially solved by correlating absolute values of height and angle over a time interval. In our previous work [10], a statistically significant correlation between growth rates was noted up until 2.5 years after menarche, specially in AIS managed by observation (r = 0.632, p < 0.001).