There has been little quantification of long-term vertical facial changes that might occur after contemporary orthognathic surgery. The aim was to assess ≥10-year follow-up vertical facial changes in patients after Class II and III orthognathic surgery.
Sequential lateral cephalograms of 162 former orthognathic patients had been gathered during routine clinical follow-up before any consideration was given to this current project. For this study, facial patterns were classified according to the Frankfort-mandibular plane and ANB angles and the horizontal distance from the pogonion to the nasion-perpendicular line. Chosen Class II and III groups were divided into longer and shorter-face subgroups on either side of the average FMP angle (longer face >25°). The assessment was made from already-available lateral cephalograms taken before treatment, at debanding, and ≥10-year clinical follow-up.
After obvious improvement with orthodontics and orthognathic surgery, mean FMP angles seemed to move back toward the preexisting vertical dimensions in both Class II and III longer and shorter-face groups over the next decade. However, there was considerable individual variation around the means. Similar posttreatment changes were seen with the ANB angle and the distance from pogonion to the nasion-perpendicular line. Significant correlations were found between the amount of preexisting discrepancy for these variables and their changes during and after treatment.
Significant planned vertical mandibular changes are achieved with Class II and III surgery. However, in the longer term, it would seem that, in general, there is a tendency for postsurgical vertical facial proportions to revert somewhat toward the pretreatment proportions. Wide individual variation in posttreatment musculoskeletal behavior should be expected. Despite these changes, it does not necessarily mean that there should be an expectation of long-term dentofacial collapse after such treatment.
Vertical changes occur over the decade after Class II and III orthognathic surgery.
There is significant variation in the amount of these changes.
The changes occur in both longer and shorter faces.
There is some association between the amount of change and the initial discrepancy.
Much is already known about the range of likely dental, skeletal, and soft-tissue effects of contemporary orthognathic surgical procedures. Those procedures are performed routinely around the world. Apart from newer concepts of surgery-first, with and without the use of skeletal anchorage plates, the techniques used today are essentially the same as those used for the past 20 years or more.
It is well-accepted that changes continue to occur in the aging adult dentofacial skeleton and soft tissues. There does seem, however, to be considerable individual variation in the reported effects—some related to hormonal changes occurring during adult life and others seemingly just a slow continuation of what might have been happening during the teenage growth spurt. ,
Historical assessments of dentofacial relationships were largely based on 2-dimensional images such as lateral cephalograms. , More recently, there has been a move toward using 3-dimensional imaging and software to diagnose and plan orthognathic surgery. Having said that, when looking at the long-term 2-dimensional outcomes of complex treatment, it would seem that much of the historical lateral cephalometric material might still be of use. To discount it would be wasting valuable potential information.
Suppose it were possible to provide further evidence of changes in the overall dentofacial system in the long term after orthodontics and orthognathic surgery. In that case, it might further aid clinicians when setting long-term protocols to maintain the results of their complex treatment. In addition, further study might be valuable to those interested in the various roles that muscular function might play in the development and control of the dentofacial complex, with or without active treatment of any sort.
With all this in mind, this study was undertaken to assess 2-dimensional changes taking place in the longer term after routine orthodontic and orthognathic surgery treatment.
Specifically, the aims were: to assess surgical and long-term vertical facial changes, to assess surgical and long-term horizontal dentoalveolar and skeletal changes, and to look for factors that might have influenced any changes taking place.
Material and methods
The sample consisted of lateral cephalograms from 162 patients who had undergone routine Class II or Class III orthognathic surgery between 1994 and 2008. The full consecutive list during that period included 268 patients—but these 162 patients were those who continued to be followed routinely for clinical examination of the health and relationships of the teeth, jaws, and surrounding tissues—for 10 years or more after the removal of fixed orthodontic appliances. Follow-up supporting lateral cephalometric radiographs had been taken for all patients. Later on, the use of this already-gathered material and the overall study purpose were the subjects of ethics application and review (Latrobe University no. HEC20434). No approval had to be given to take any further radiographs purely for research purposes—only for the use of the already-gathered clinical material. As such, this was a retrospective sample of availability and convenience. The study patients were treated by 1 orthodontist and several experienced oral and maxillofacial surgeons in private practice.
The original treatment had consisted of the placement of contemporary preangulated fixed appliances on all erupted teeth, including the second molars. Maxillary arch expansion, with or without surgical assistance or various premolar extractions, may have been part of the orthodontic management for any patient. All patients had received rigid screw and/or plate fixation at the osteotomy sites. These plates had not been removed during the follow-up period in most patients. High-quality pretreatment (T1), debanding (T2), and ≥10-year (T3) lateral cephalograms exhibiting good soft-tissue definition with lips relaxed and teeth in occlusion were available for all subjects. The overall case sample was initially divided according to their detailed clinical notes on whether they had received some sort of Class II or III orthognathic surgery. All subjects were deidentified by being given study numbers—so that no reference could then be made to the patient’s name during subsequent measuring and tabulation. The radiographs for each subject were also labeled with numbers corresponding to the respective clinical time points (T1 to T3).
After uploading all 486 radiographs to the Dolphin system (Dolphin Imaging and Management Solutions, Chatsworth, Calif), cephalometric landmarks were identified by 1 investigator (M.G.W). Digitization was then undertaken by 3 other trained workers over 18 months. Each worker digitized approximately equal numbers of radiographs. Radiographs were digitized randomly, without reference to the others of the same patient taken at different time points. Cephalometric landmarks used in this study are presented in Table I . Three widely accepted lateral cephalometric measurements were made in this process:
As one measure of vertical mandibular steepness, Frankfort-mandibular plane angle (Go-Mn) (degrees) also reflects anterior vertical facial height.
ANB angle (degrees) as a measure of the anteroposterior dentoalveolar relationship.
The distance from the pogonion to the nasion-perpendicular line (mm) measures the horizontal chin position within the face.
|Porion||The most superior point on the outline of the bony external auditory meatus|
|Orbitale||The most inferior point on the orbital curve|
|Nasion||The most anterior point on the articulation between the frontal and nasal bones|
|Pogonion||The most anterior point on the bony chin|
|Gonion||The most posteroinferior point on the angle of the mandible|
|Menton||The most inferior point on the bony symphysis|
|Point A||The most posterior point on the maxillary labial alveolar curve|
|Point B||The most posterior point on the mandibular labial alveolar curve|
After digitization and measurement, the chosen Class II and III surgery groups were further divided into longer- and shorter-face subgroups, on either side of the average FMP angle: longer (>25°) and shorter (≤25°).
For interest only, it was later noted that the following surgical operations had been performed in each group: (1) Class II longer: total 55, bimaxillary surgery 45, mandible-only 8, and maxilla-only 2; (2) Class II shorter: total 41, mandible-only 38, and bimaxillary surgery 3; (3) Class III longer: total 34, bimaxillary surgery 26, and mandible-only 8; and Class III shorter: total 32, bimaxillary surgery 17, mandible-only 8, and maxilla-only 7.
Various genioplasties had been performed in many patients.
To assess location and measurement error and interworker variability, 10 radiographs were randomly selected from the set digitized by each trained worker. Those radiographs were uploaded again to the Dolphin system, and landmarks relocated (M.G.W) without reference to their original placement. Each worker then redigitized their 10 films and remade the set measurements. Each worker also digitized 10 films previously digitized by one of their colleagues. Differences in measurements of <1 mm or <1° were considered acceptable. At the 95% confidence level, the results of paired t tests indicated that there were no significant differences between the first and second sets of these simple measurements made by each worker on their own set, as well as those from the other sets. ,
After digitization, measurements were imported into an Excel spreadsheet (Microsoft Office 2007; Microsoft Corporation, Redmond, Wash) in which the linear measurement (the distance from pogonion to the nasion-perpendicular line) was multiplied by 0.926 or 0.917, respectively, as necessary to correct for differences in magnification factors of 8% and 9%. The sample characteristics and treatment changes were then statistically analyzed using a commercially available statistical software package, PASW Statistics (version 18.0; SPSS Inc, Chicago, Ill). Independent t tests were used to test for differences in the means for the longer and shorter-face subgroups within the Class II and Class III samples at T1, T2, and T3. Paired t tests determined the statistical significance of the treatment changes between time points for each subgroup. Significance was set at either P <0.05 or P <0.01. Pearson correlation coefficients and associated levels of significance were then calculated to determine the levels of correlation between the T1 to T3 changes occurring in the FMP and ANB angles and the pogonion distance, respectively, and the various T1 measurements.
Once all FMP angles had been measured, the patients with Class II and III malocclusion could be divided into the main longer and shorter-face study subgroups. Mean ages for these patients with Class II and III malocclusion longer and shorter-face groups are presented in Table II . From the table, it can be seen that there was wide individual variation in ages at the commencement of treatment. Although the mean ages generally reflected patients in their early-to-mid 20s, some were in their late teens, late 20s, and early 30s. The ages at the ≥10-year follow-up were generally in the mid to late 30s, but within one standard deviation, they ranged from early 30s through to late 40s. Mean ages for males and females within the vertical subgroups are also included in Table II , although the sample numbers are unequal and some are quite small. Because of the relatively small numbers in some gender-based subgroups, it was decided to pool the measurement data and focus on the 4 main vertical subgroups.
|Total sample||162||23.70 (3.60)||25.60 (4.64)||37.58 (4.93)|
|Class II longer||55||23.66 (3.18)||25.67 (4.15)||36.64 (4.47)|
|Male||36||25.29 (4.22)||27.10 (4.56)||38.10 (5.21)|
|Female||19||23.03 (3.10)||25.24 (4.17)||36.77 (4.88)|
|Class II shorter||41||21.48 (3.84)||23.54 (4.20)||34.86 (5.05)|
|Male||21||23.15 (4.22)||25.36 (5.10)||36.42 (5.77)|
|Female||20||21.03 (3.77)||23.22 (4.66)||35.35 (4.90)|
|Class III longer||34||25.47 (3.20)||27.10 (5.23)||38.62(5.66)|
|Male||14||28.67 (4.10)||30.52 (5.13)||41.68 (5.49)|
|Female||20||24.20 (3.44)||27.78 (4.82)||36.73 (4.77)|
|Class III shorter||32||24.74 (4.45)||26.55 (5.42)||41.59 (4.81)|
|Male||12||26.07 (4.66)||29.74 (5.23)||40.44 (5.74)|
|Female||20||23.19 (4.21)||25.22 (5.73)||36.19 (4.22)|
Mean values for the Frankfort-mandibular plane angle (gonion to menton) within the various vertical subgroups, T1 to T3, are presented in Table III and Figure 1 . From the table, it can be seen that at T1, there were significant differences between the mean FMP angles within both Class II and III longer and shorter-face subgroups. There was considerable individual variation within each group. These differences between vertical groups were not so apparent at T2. The longer and shorter-face means had all moved toward each other. The various FMP means at T3 then moved back toward those at T1, so differences in the longer and shorter-face means were significant again. Interestingly, the T3 means within all 4 vertical groups were not significantly different from the T1 means within those groups. Changes in the means themselves are also presented in Table III , and the wide ranges of individual variation in both treatment and follow-up FMP angle changes are further highlighted in Table IV .
|Groups||n||Frankfort-MP angle, degrees †||Changes|
|T1||T2||T3||T1 to T2||T2 to T3||Residual change|
|Total sample||162||26.62 (3.57)||26.41 (3.22)||26.46 (4.22)|
|Class II long||55||32.48 (3.63) ∗∗||28.05 (3.09) ††||30.44 (4.98) ∗||4.43 (2.18) ∗∗||2.39 (1.58) ∗||−2.04 (2.43)|
|Class II short||41||20.13 (3.84) ∗∗||23.92 (2.47) ††||22.05 (3.18) ∗||3.79 (1.81) ∗∗||−1.87 (1.52) ∗||1.32 (2.95)|
|Class III long||34||30.94 (3.24) ∗∗||27.82 (3.86) ††||29.52 (4.92) ∗||−3.12 (1.41) ∗∗||1.70 (0.84) ∗||−1.42 (2.08)|
|Class III short||32||20.26 (2.65) ∗∗||24.87 (2.80) ††||22.00 (2.62) ∗||5.01 (1.77) ∗∗||−3.27 (1.32) ∗||1.74 (2.44)|
|Groups||n||T1 to T2||T2 to T3|
|Class II long||55||−2.25° to −6.61° (closing)||0.81° to 3.97° (reopening)|
|Class II short||41||1.98° to 5.60° (opening)||−0.35° to −3.39° (reclosing)|
|Class III long||34||−1.71° to −4.53° (closing)||0.86° to 2.72° (reopening)|
|Class III short||32||3.24° to 6.78° (opening)||−1.98° to −4.59° (reclosing)|
Mean values for the ANB angle within the various vertical subgroups, T1 to T3, are presented in Table V . From the table, it can be seen that at T1, there was no significant difference between the mean ANB angles within the Class II longer and shorter-face groups. However, within the Class III groups, the shorter-face mean ANB angle was significantly different from that of the longer-face subgroup. The T2 means had all moved toward a population average of approximately 2°, and the previously significant differences in the means between the vertical subgroups were now not as apparent. In contrast to the apparent mean changes in the ANB angle during active treatment (to T2), those seen during the follow-up period through to T3 had all moved away somewhat from the population average again. The differences between the ANB means at T1 and T3 were now not significant. As with all the other measurements, there was considerable individual variation in the changes seen within the groups at all time points.
|Groups||n||ANB, degrees||Pogonion to nasion-perpendicular, mm|
|Total sample||162||2.09 (2.90)||1.19 (2.13)||0.96 (2.27)||−1.92 (3.24)||0.47 (2.29)||−1.72 (3.02)|
|Class II long||55||7.22 (2.92) †||3.21 (2.22) †||4.55 (2.89) †||T1 to T3 †||−8.14 (3.18) †||−2.11 (2.47) ∗∗||−3.85 (3.19) †||T1 to T3 ∗|
|Class II short||41||6.88 (2.88) †||2.50 (1.94) †||3.64 (2.02) †||T1 to T3 †||−6.46 (2.88) †||2.05 (1.94) ∗∗||−1.47 (2.02) †||T1 to T3 ∗|
|Class III long||34||−3.81 (2.64) ∗||1.03 (1.66) †||−0.64 (1.84) †||T1 to T3 †||3.47 (2.64) ∗||1.38 (1.66) †||2.99 (1.84) †||T1 to T3 †|
|Class III short||32||−6.58 (3.17) ∗||1.57 (1.85) †||2.69 (1.97) †||T1 to T3 †||8.84 (4.45) ∗||1.88 (3.10) †||5.27 (4.02) †||T1 to T3 †|
Mean values for the horizontal distance from pogonion to the nasion-perpendicular line, within the various vertical subgroups, T1 to T3, are also presented in Table V . From the table, it can be seen that, at commencement, T1, there was no significant difference between the mean distance for the longer and shorter-face Class II subgroups. However, there was a significant difference between the Class III longer and shorter-face means. The T2 means for the Class II vertical subgroups were significantly different. Those for the Class III groups were not. The differences in means for this pogonion distance at T3, within the Class II and Class III groups, respectively, were not significant. However, overall changes in the means from T1 to T3 for both Class II groups were significant. The T1 to T3 changes for the Class III groups were not.
The calculation of Pearson correlation coefficients showed significant correlations between T1 and T2 and T2 and T3 measurements. Within the total sample, for instance, the more severe the initial T1 cephalometric measurement, the greater the change seen through active treatment (to T2), for all 3 chosen cephalometric measurements: ANB ( r = 0.554; P <0.01), FMP ( r = 0.796; P <0.01), and pogonion to the nasion-perpendicular line ( r = 0.532; P <0.01). This reflects the presenting morphology and that treatment was planned to deal with the existing individual dentofacial discrepancy.
In contrast to the T2 calculations, the correlation coefficients involving T3 were not as strong. For those involving the T1 to T3 changes, significant correlations were still found within the total sample for ANB ( r = 0.532; P <0.0.05) and FMP ( r = 0.544; P <0.05), but not for pogonion to the nasion-perpendicular line. These ANB and FMP correlations would also seem to reflect the fact that there does seem to be a relationship between expected posttreatment change and the originally-assessed necessary amount of correction. The fact that there was no significant T1 to T3 correlation for pogonion to the nasion-perpendicular line in this sample may well have been affected by the fact that many anteroposterior advancement or reduction sliding genioplasties were performed, as deemed necessary.