Comparison of statural height growth velocity at different cervical vertebral maturation stages


Knowledge of a patient’s stage of growth and development plays a vital role in diagnosis, treatment planning, results, and stability of the outcome. Cervical vertebral maturation (CVM) predicts the stage of growth and development, but its validity has only been investigated restrospectively, using historic samples. Our objective was to assess prospectively whether a correlation exists between CVM stage and statural height growth velocity.


Participants were aged between 8 and 18 years and of both sexes. Standing height was measured every 6 weeks with participants barefoot and in natural head position. CVM stage was assessed from lateral cephalograms taken at the start of treatment. Intraobserver and interobserver reliability of CVM staging and statural height measurements were assessed using the Cohen weighted kappa, percentage of agreement, intraclass correlation coefficient, and Bland-Altman plots, respectively. Analysis of variance was used to test for statistically significant differences between growth velocities at the CVM stages.


We analyzed 108 participants. The peak in statural height growth velocity occurred at CVM stage 3 ( P = 0.001). There was a statistically significant difference in the mean annualized growth velocity between all CVM stages except stages 2 and 4. Girls had their peak pubertal growth spurt an average of 1.2 years earlier than did boys.


This study suggests that there is a significant relationship between CVM stage and statural height velocity.

Graphical abstract


  • This was a prospective study to assess the validity of the CVM staging method.

  • The peak in statural height growth velocity occurred at CVM stage 3 ( P = 0.001).

  • Girls undergo peak pubertal growth spurt on average 1.2 years earlier than boys.

  • CVM staging is valid for identifying the pubertal peak in statural height velocity.

Knowledge regarding the timing and extent of growth for orthodontic patients is essential for managing them optimally and successfully, particularly in those with skeletal discrepancies. Such knowledge plays a vital role in the diagnosis, treatment planning, results, and overall stability of the patients’ outcomes. Depending on the stage of development and growth velocity expected for a patient, different treatment modalities may be considered more appropriate than others. For orthodontists, it therefore is crucial to be able to assess the growth potential of a patient and, when orthognathic surgery or implants are required, to know when growth has ceased ( Fig 1 ).

Fig 1
Changes in height growth velocity from infancy to adulthood
(modified from McNamara ).

Numerous methods have been investigated to identify the stage of growth and development and predict both the timing and potential of this growth. These include chronologic age, dental age, menarche and voice changes, standing height, skeletal maturation of the hand and wrist, and cervical vertebral maturation (CVM).

None of these methods has demonstrated a strong enough correlation to growth with the exception of skeletal age of hand-wrist radiographs and CVM. The principle of using skeletal maturity to determine the most appropriate time for orthodontic treatment has varied in popularity but has always required additional radiation exposure and additional skills for the orthodontist to interpret the hand-wrist radiographs. As a result, alternatives to hand-wrist radiographs were sought using investigations that were more commonplace in orthodontics and more familiar to the orthodontist to facilitate interpretation.

CVM is an alternative method to hand-wrist radiographs that has been shown to be reliable and does not require additional radiation for patients undergoing orthodontic treatment. The CVM index assesses the shape of the cervical vertebrae visible on a standard lateral cephalogram and uses this approach to predict the patient’s stage of growth and development.

Although many studies have looked into the validity of this index, all have used historical samples and have been retrospective. Methodologic flaws, as well as sampling issues, mean that the validity of the CVM method is yet to be shown in a contemporary population sample in a prospective manner.

For the CVM method to be of use clinically today, it must not only be reliable but also be valid with respect to its predictability of growth in a contemporary sample. The aim of this study was therefore to address the concerns raised in previous research and to determine the validity of the CVM stage as a predictor for growth using an appropriately sized, contemporaneous sample of children and adolescents, in a prospective manner. The objective of this study was to assess whether a correlation exists between CVM stage and statural height growth velocity. Phase II of this study will assess the relationship between CVM stage and facial (mandibular) growth.

Material and methods

Routine orthodontic clinical records were collected from all patients according to the departmental protocol of Liverpool University Dental Hospital, United Kingdom, for use in treatment planning. These records included the patient’s initial standing height, a lateral cephalogram radiograph, intraoral and extraoral photographs, and other radiographs as indicated clinically. After this, the patients started orthodontic treatment, as appropriate, to correct their malocclusion. Routine care was provided per the consultant’s treatment plan. Interim and final records were obtained as clinically necessary. In addition to this routinely acquired information, the clinicians undertaking the patients’ treatment took a measurement of standing height at each visit and recorded it in the clinical records.

Participants receiving orthodontic treatment at Liverpool University Dental Hospital in the academic years 2012-2013 and 2013-2014 were eligible to participate in this study. At their first appointment, routine history, examination, and special investigations were undertaken; then the patient and parent or guardian were informed about the study and invited to participate if the inclusion criteria were fulfilled.

Participants were included if they were 18 years or younger, were of either sex, had not received previous orthodontic treatment, and had given informed consent or assent to participate in the study.

Patients diagnosed with any congenital clefts of the lip or palate, or known or suspected craniofacial syndromes or growth-related conditions, were excluded.

Standing height was measured with the patients barefoot, feet together with their heels against the wall and in natural head position (Frankfort horizontal plane parallel to the floor), using a wall-mounted height measure, at every visit. Taking the standing height at each visit (every 6 weeks) was the only intervention that was in addition to routine clinical practice. It was carried out in a designated area by the treating clinician and recorded on a data sheet in the patients’ notes. The height was measured at the same time of the day on each occasion. An annualized growth rate was calculated to allow comparison of growth rates over the same time period. This was also because there were small fluctuations in the growth velocity throughout the year, so an average growth velocity per year was calculated.

Lateral cephalograms (including the cervical vertebrae) were taken at the start of treatment for diagnostic and treatment planning purposes. The lateral cephalograms were staged for maturation using the CVM index described and developed by Bacceti et al.

For analysis, patients were grouped into 4 categories, CVM stages 1 and 2, 3, 4, and 5 and 6. To detect differences between these groups, of the magnitude observed by Franchi et al with 80% power at the 5% significance level, 17 patients per group would be needed, or a total sample size of 68.

Intraobserver and interobserver reliabilities of CVM staging were assessed using the Cohen weighted kappa and percentage of agreement. Intraclass correlation coefficient and Bland-Altman plots were used to assess the reliability of the statural height measurements. Analysis of variance was used to test for statistically significant differences between the statural height velocities at different CVM stages.

The study was conducted in compliance with the principles of the Declaration of Helsinki (1996). Ethical approval was granted from Liverpool East Research Ethics Committee, reference number 13/NW/0408 and protocol number UoL000751, granted on October 30, 2013.

S.H. was calibrated, in the assessment of CVM stage, to J.E.H., who had been calibrated in a previous study. The first CVM assessment was undertaken on a randomized set of full cephalograms by both assessors independently and in duplicate. The second assessment was undertaken 4 weeks later on the same set of images of full cephalograms presented in a different randomized order. To assess the impact of being able to see the dentition on the CVM assessment, the same set of cephalograms with the dentition masked and in a different randomized order was assessed by both assessors a year later.


The intraobserver reliability of the CVM index was characterized as “almost perfect” agreement (weighted kappa, 1; percentage of agreement, 100%). Interexaminer reliability of the CVM index also was characterized as “almost perfect” agreement (full cephalograms: weighted kappa, 0.96; percentage agreement, 99.3%; masked cephalograms: weighted kappa, 0.83; percentage of agreement, 97.4%). Intraobserver and interobserver reliabilities for statural height measurements were characterized as “excellent” (intraclass correlation coefficient, 0.98-0.99). Intraobserver reliability of cephalometric measurements was characterized as “good” (intraclass correlation coefficient, 0.85-0.93).

The total number of patients recruited into the study was 185. However, due to various shortcomings in the data collection, discussed later, the final sample size that could be used in the data analysis was 108. This is demonstrated in Figure 2 , which shows the flow of patients through the study.

Fig 2
Flow of patients through study.

The mean age at the start of treatment in the final sample (n = 108) was 13.9 years (SD, 1.7 years; range, 10.16-18.56 years). For the sample size in each group to be sufficient to provide adequate power to detect a statistically significant difference if it existed, data from patients at CVM stages 1 and 2 and CVM stages 5 and 6 were combined. Those in CVM stages 3 and 4 were kept separate because these stages were of particular interest in this research.

The mean annualized growth velocity (MAGV) in statural height was calculated for each patient; the peak in statural height growth velocity occurred at CVM stage 3 ( P = 0.001).

Table I shows the MAGV in centimeters per year in each CVM group. There was an increase in the MAGV from CVM stage 2 to stage 3 by almost 5 centimeters per year. The MAGV then dropped by 4.3 centimeters per year at CVM stage 4 relative to CVM stage 3. At CVM stages 5 and 6, the MAGV dropped by a further 3.5 centimeters per year to 1.5 centimeters per year in the MAGV. This pattern would support the idea that standing height velocity is greatest at CVM stage 3 and has dropped significantly by stages 5 and 6.

Table I
Mean annualized growth rate by CVM stage
CVM stage Patients (n) Mean annualized growth rate (cmy −1 ) SD (cmy −1 )
1 and 2 14 4.51 2.71
3 22 9.39 4.44
4 33 5.00 2.33
5 and 6 39 1.56 2.34
Total 108 4.59 4.06
cmy 1 , Centimeters per year.

The trend in the MAGV through the CVM stages is shown in Figure 3 . This demonstrates the increase in the MAGV that occurs at CVM stage 3. The growth velocity then decreased from CVM stage 3 to stage 4, and further reduced going into CVM stages 5 and 6. There were 2 outliers in the CVM stage 3 group who showed particularly high growth velocities. There were also 3 outliers at CVM group 5 and 6, suggesting that a few subjects had some late growth.

Fig 3
Box plot of MAGV at different CVM stages.

When assessing the MAGV split between sexes ( Table II ), the same pattern was seen as described above, with the MAGV highest at CVM stage 3 and decelerating toward stage 5. However, growth velocities in boys generally were higher than in girls, and their peak growth velocity was 10.4 centimeters per year (95% confidence interval [CI], 7.68-13.30) of MAGV compared with only 7.5 centimeters per year (95% CI, 4.83-10.26) of MAGV for girls, but this difference was not statistically significant in our sample ( P = 0.36).

Table II
MAGV in each CVM group
CVM stage Patients (n) MAGV (cmy 1 ) 95% CI
Lower bound Upper bound
Girls Boys Girls Boys Girls Boys Girls Boys
1 and 2 9 5 3.87 2.16 5.65 3.47 2.21 1.34 5.53 9.96
3 8 14 7.54 3.25 10.44 4.79 4.83 7.68 10.26 13.2
4 17 16 4.43 2.20 5.60 2.38 3.31 4.33 5.57 6.87
5 and 6 26 13 1.77 2.70 1.13 1.39 0.68 0.29 2.86 1.97
cmy 1 , Centimeters per year.

Table III displays the mean ages of subjects at CVM stage 3. The mean ages at CVM stage 3 were 12.4 (SD, 1.4) years for girls and 13.6 (SD 0.9) years for boys. This confirms previous research reporting that girls reach puberty before boys, and the difference was significant (−1.20 years; 95% CI, −2.12 to −0.28).

Table III
Average ages of subjects at CVM stage 3
Sex Patients (n) Mean age (y) SD Minimum Maximum 95% CI
Lower bound Upper bound
Girls 8 12.4 1.14 10.55 13.61 11.46 13.36
Boys 14 13.6 0.90 11.80 14.51 13.06 14.09
Total 22 13.2 1.12 10.55 14.51 12.65 13.65

Analysis of variance was used to test for statistically significant differences in the MAGV between the different CVM stages. The null hypothesis, stating that there was no difference between the 4 groups, was rejected, since the differences were statistically significant ( P = 0.0001). Table IV demonstrates the statistically significant difference between 2 CVM stages.

Dec 8, 2018 | Posted by in Orthodontics | Comments Off on Comparison of statural height growth velocity at different cervical vertebral maturation stages
Premium Wordpress Themes by UFO Themes