Subjects

The subjects enrolled in this study were 80 consecutive Caucasian patients (46 males and 34 females) with a Class III malocclusion operated with bilateral sagittal split osteotomy. All patients were selected from the files at the Department of Orthodontics, University of Oslo. Surgery was performed at the Department of Maxillofacial Surgery, Ullevaal University Hospital during the period from 1990 to 1996. No patient received any additional orthognathic surgical procedure and all patients received rigid internal fixation. Practicing orthodontists or postgraduate students carried out pre- and postsurgical orthodontic treatment. No surgery was performed until growth was evaluated to have declined to adult levels. A criterion for inclusion in the study was the availability of standardized lateral cephalograms of adequate quality and resolution exposed according to a strict data collection protocol. The subjects' age at surgery and some presurgical craniofacial and occlusal characteristics are presented in Table 1.

TABLE 1. Patient Characteristics^a

View Fullscreen

Surgical technique

Seven different surgeons performed the operations for the patients studied. After completion of the mandibular split, the teeth were placed in their planned position and stabilized with intermaxillary fixation. The bony segments were fixed using bicortical screws with washer (Salzburg, Howmedica Leibinger; GmbH & Co. KG; Freiburg, Germany) which were placed in the gonial area through a transcutaneous approach. Two screws were placed above the inferior alveolar nerve in the retromolar area and the third screw below the nerve in a more anterior position. After completion of skeletal fixation, the intermaxillary fixation was released and the occlusion and the position of the condyles were checked. Postoperative orthodontic treatment and use of Class III elastics started within 2 to 4 weeks after surgery.

Cephalometric assessment

All lateral cephalograms were taken in the same cephalostat with the teeth in centric occlusion and the lips in relaxed position. Magnification for linear measurements was 5.6%, which was not corrected. The cephalograms were taken at the following occasions: within 1 week before surgery (T1), within 1 week after surgery (T2), 2 months after surgery (T3), 6 months after surgery (T4), 1 year after surgery (T5), and 3 years after surgery (T6). The 2-month radiograph was missing in 3 patients, and the 6-month and 1-year radiographs were missing in 4 patients.

All radiographs were hand-traced on acetate paper by the same examiner. The clearest radiograph in each patient's series was selected and the details of the cranial base structures were drawn. A x-y cranial base coordinate system was constructed on this radiograph through Sella with the x-axis drawn 7° to the Sella-Nasion line and the y-axis passing through Sella perpendicular to the x-axis. The tracings of the different stages were superimposed on the first radiograph and the reference lines were transferred to the tracing of each consecutive cephalogram. During superimposition particular attention was given to fitting the tracing of the cribriform plate and the anterior wall of the sella turcica, areas that undergo minimal remodeling.20 The x and y coordinates for the landmarks were registered by digitization to the Dentofacial Planner computer program (Dentofacial Software Inc, Toronto, Canada). The skeletal and soft tissue landmarks identified and the reference lines used are shown in Figure 1. Cephalometric landmark and some measurement definitions are presented in Table 2.

View Figure

• Download as Powerpoint
• Download Figure

TABLE 2. Cephalometric Landmark and Some Measurement Definitions

View Fullscreen

Error of the method and statistical analysis

Twenty-five radiographs chosen at random were traced and digitized by the same investigator on 2 separate occasions at least 2 weeks apart. Dahlberg's method21 was used to determine the error between the duplicate determinations and the coefficient of reliability was also calculated.22 Systematic error was assessed by a paired t-test at the 10% level as recommended by Houston22 (Table 3).

TABLE 3. Error of the Method Assessed From Duplicate Tracings of 25 Radiographs

View Fullscreen

To test for statistical significance of changes in cephalometric variables between the different stages, Student's t-test for paired data was performed. Pearson's coefficients of correlation were calculated to examine relationships between soft tissue and hard tissue changes. The sample was subdivided according to gender and magnitude of surgical setback. T-tests for independent samples and analysis of variance (ANOVA) were used to compare soft to hard tissue ratios in these subgroups.

A surgical splint was present in the immediate postsurgical radiographs of 3 patients. Due to the increase in anterior facial height caused by the splints and the resulting influence on soft tissue landmark position, all data based on the immediate postoperative cephalogram (T2) for these 3 patients were excluded.

Amount, direction, and time course of soft tissue changes

Upper lip

A posterior repositioning of the upper lip landmarks Sn, Sls, and Ls as an immediate result of mandibular setback was observed. This immediate backward movement was most pronounced at Ls (0.7 ± 1.6 mm; P < .001) and gradually decreased in the more superior lip points. Posterior movement of Ls and Sls continued during the postoperative period. At the end of the observation period, the net posterior movement of Ls was 1.6 ± 1.6 mm (P < .001). Although the decrease in the distance from Ls to the esthetic line as a result of surgery was 2.0 ± 1.5 mm, only 0.6 ± 1.5 mm (P < .001) of this change was maintained at the end of follow-up. Along with the posterior movement of the upper lip landmarks, an accompanying inferior movement could be observed. As an immediate response to surgery, point Stm^s moved down (1.0 mm ± 1.7 mm) and continued to move in an inferior direction throughout the postoperative period to reach a net inferior position of 1.9 ± 1.6 mm (P < .001) (Table 5). A straightening of the upper lip as evidenced by the 6.3 ± 5.3° (P < .001) decrease in upper lip inclination and the 6.1 ± 6.1° (P < .001) increase in the nasolabial angle was also observed. The upper lip thickness decreased, the decrease being most pronounced in the measurement Ls to U1 (2.2 ± 1.9 mm; P < .001), but also obvious in measurements Sls to A and Sn to A.

TABLE 5. Changes in Soft Tissue Variables Among 80 Mandibular Setback Patients During the Various Time Intervals^a

View Fullscreen

Lower lip and chin

As an immediate effect of mandibular setback, all soft tissue landmarks were relocated in a posterior direction. The greatest change was observed at point Mlf (6.5 ± 3.4 mm; P < .001), followed by Pg’ (6.0 ± 3.9 mm; P < .001) and finally by Li (5.4 ± 3.1 mm; P < .001) (Figure 2). During the first 2 months postoperatively, Li moved more posteriorly than Mlf, whereas Pg’ did not follow this movement. In the subsequent postoperative intervals the lower lip remained more or less stationary, whereas a reversal in direction of movement at Mlf and Pg’ could be seen. Statistically significant anterior movement of Mlf and Pg’ then continued throughout the rest of the observation period. Thus, at the end of the follow-up period, point Li demonstrated the greatest amount of posterior relocation (6.5 ± 2.8 mm; P < .001) compared to Mlf and Pg’ which were relocated 5.9 ± 3.0 mm and 4.5 ± 3.7 mm (P < .001), respectively. The distance from Li to the esthetic line showed a net increase of 2.2 ± 2.4 mm (P < .001). In the vertical plane, as a direct effect of surgery, points Li and Stmⁱ showed an inferior movement of 2.3 ± 3.6 mm and 1.8 ± 2.6 mm (P < .001), respectively. Postoperatively, Li moved in a superior direction (1.5 ± 2.4 mm P < .001) approaching its original preoperative vertical position, whereas Stmⁱ maintained a net inferior position of 1.5 ± 2.3 mm (P < .001). Mean surgical changes in the vertical position of landmarks Mlf, Pg’ and Me’ were minimal and statistically insignificant. At the end of the observation period, only Me’ demonstrated a net superior movement (0.8 ± 1.9 mm P < .001).

A change in lower lip inclination (7.7 ± 9.5° P < .001) was noted at the time of surgery with the lower lip becoming more procumbent. This initial change then gradually diminished so that, by the end of the observation period, the lower lip had more or less regained its presurgical inclination. An increase in mentolabial fold depth of 0.9 ± 1.3 mm (P < .001) was also observed as a result of surgery. At the end of follow-up only 0.4 ± 1.1 mm (P < .001) of this deepening was maintained. A slight net decrease in soft tissue thickness at the mentolabial fold region (Mlf-B) (0.3 ± 0.8 mm; P < .01) was evident (initial increase followed by a greater decrease). An initial increase in lower lip thickness (Li-L1), which gradually decreased again to reach its preoperative dimension, was also observed. Soft tissue thickness at the chin (Pg’-Pg) similarly remained mostly unchanged.

As a result of surgery, facial convexity (G-Sn-Pg’) was increased (5.9 ± 3.4°; P < .001). About two-thirds of this improvement was maintained at the end of the follow-up period. Composite tracings demonstrating the average profile of the 80 patients before and 3 years after surgery are shown in Figure 3. Values for some of the measurements describing the average soft tissue profile for males and females at the end of follow-up are given in Table 6.

View Figure

• Download as Powerpoint
• Download Figure

TABLE 6. Mean Values and One Standard Deviation Describing the Soft Tissue Profiles for Male and Female Setback Patients 3 Years After Surgery^a

View Fullscreen

Relation between soft tissue and hard tissue movements

Correlations

Changes in all soft tissue landmarks were found to correlate with movement of their corresponding hard tissue structures. Strongest correlations were generally observed between Pg’-Pg and Mlf-B, followed by Me’-Me and Li-L1. Correlations coefficients between the magnitude of surgical (T1-T2) and net (T1-T6) skeletal change on one hand, and the net (T1-T6) effects on the soft tissue, on the other hand, were calculated and are presented in Table 7. Correlations between the net soft tissue change and the net skeletal change were higher than with the immediate skeletal change. A weak, but significant correlation (r = 0.33; P < .01) was also found between magnitude of skeletal setback at Pg and horizontal change in upper lip position. In the vertical plane, correlations between soft tissue and hard tissue movements for the chin and mentolabial fold (in the range of 0.4 to 0.5; P < .001) were lower than those in the sagittal plane. Correlation coefficients of 0.55 and 0.50 (P < .001) were observed between setback at Pg and the vertical change at Stm^s and Stmⁱ, respectively. Weaker correlations were found between the change in anterior facial height (vertical movement of Me) and these measurements.

TABLE 7. Pearson Correlation Coefficients Between Hard Tissue Movement and Soft Tissue Changes^a,b

View Fullscreen

No associations were found between skeletal repositioning in the horizontal or vertical plane and the change in soft tissue thickness at the chin, mentolabial fold, the lower lip, or the change in mentolabial fold depth. A weak, but significant correlation was, however, found between the magnitude of setback and the net decrease in upper lip thickness (Ls to U1) (r = 0.47; P < .001).

Preoperative thickness of both the upper and lower lip was significantly correlated with the net change in their thickness (r = 0.54 and r = 0.36; P < .001, respectively). No similar associations were found for preoperative soft tissue thickness of the mentolabial fold or chin. A significant correlation was, however, found between the preoperative mentolabial fold depth and the net change in its depth (r = 0.57; P < .001).

Influence of skeletal relapse on postoperative soft tissue changes

Percentages of the net change (T1-T6) of some soft tissue landmarks and both the net (T1-T6) and surgical movement (T1-T2) of the underlying hard tissue landmarks were calculated (Table 8). Due to the large standard deviations encountered, the 5% trimmed mean is also presented. Percentages of the net soft tissue change to the net skeletal change were generally higher compared to those of the net soft tissue change to the surgical skeletal change.

TABLE 8. Percentages of Soft Tissue Changes to Hard Tissue Movement^a

View Fullscreen

A paired t-test was used to analyze whether the magnitude of postoperative soft tissue changes differed significantly from the amount of relapse of the underlying skeletal landmarks. No statistically significant differences existed between relapse at Pg and Pg’ or Me and Me’. Statistically significant differences were found between changes at Mlf and point B, the skeletal landmark moving more anteriorly (0.7 ± 1.1 mm; P < .001) than its soft tissue counterpart. Whereas Li showed a postoperative posterior movement, L1 demonstrated an anterior movement resulting in a difference of 2.3 ± 1.7 mm (P < .001).

Effect of magnitude of surgical setback on soft tissue changes

The 80 mandibular setback patients were divided into 4 subgroups according to the amount of surgical repositioning at Pg. Patients with setbacks smaller than 3.0 mm formed the small setback group (S), setbacks between 3.0 and 6.0 mm formed the moderately small setback group (MS), setbacks between 6.0 and 9.0 mm the moderately large setback group (ML) and setbacks greater than 9.0 mm the large setback group (L). Ratios of net soft tissue (T1-T6) to net hard tissue (T1-T6) movement for the 4 subgroups were calculated and are presented in Table 9. Group (S) showed the largest variance in soft tissue percentages. Since such ratios with large standard deviations are not clinically useful, this group was excluded from further statistical analysis. Standard deviations for the upper lip, mentolabial fold, and chin in group (L) were smaller compared to groups (ML) and (MS). Standard deviations of the percentages for the upper lip decreased with increasing magnitude of setback. A trend for the percentages of the lower lip and soft tissue chin to decrease from group (L) to group (MS) could also be observed. Statistical comparison (ANOVA) between the 3 subgroups revealed statistically significant differences in the percentages for the lower lip and soft tissue chin only (P < .05).

TABLE 9. Percentages of Net (T1–T6) Soft to Hard Tissue Movement Among Setback Patients^a

View Fullscreen

Effect of gender on soft tissue response

Following exclusion of group (S) due to its considerable variability, the remaining 62 patients were divided into male and female subgroups. No statistically significant differences were found between the 2 subgroups in magnitude of setback at Pg, vertical surgical changes at Me, or skeletal relapse. Ratios of net soft tissue (T1-T6) to net hard tissue (T1-T6) movement for the 2 subgroups were calculated and are presented in Table 10. Females generally demonstrated higher percentages of soft tissue movement than males, but the differences were statistically significant for the ratios of Ls: Pg and Pg’: Pg (P < .05) only.

TABLE 10. Percentages of Net (T1–T6) to Hard Tissue Movement Among Male and Female Setback Patients Following Exclusion of Patients with the Smallest Setbacks

View Fullscreen

Ratios of soft tissue to hard tissue movements

An accurate prediction of the postoperative facial profile is an essential step of the treatment planning process for combined surgical orthodontic therapy. As computers become more powerful and software more intuitive, computerized prediction tracings and videoimaging procedures for planning orthognathic surgery are increasing in popularity. The database upon which commercially available surgical prediction software packages are based has been derived from studies that have either reported mean ratios of soft tissue to hard tissue movements or provided linear regression equations. Ratios generated for the present sample of 80 patients are generally comparable to those described by other authors (Table 11). Due to the high individual variability encountered, it was considered appropriate to report the 5% trimmed mean rather than the absolute mean value for the total sample to discount for the effect of some of the outliers and to provide clinically useful data. Percentages for the upper lip (25%), lower lip (100%), and mentolabial fold (106%) are somewhat higher than figures previously reported. The percentage for the chin (94%) is intermediate between values reported in other studies.

TABLE 11. Literature Review of Soft Tissue to Hard Tissue Ratios Following Mandibular Setback Surgery^a

View Fullscreen

While experience and research25,26 have shown the use of software based on such data to be clinically helpful, it is important to realize that the computerized prediction of soft tissue profile changes following orthognathic surgery is not a precise science and that treatment forecasts are likely to be only as accurate as the database utilized. Arguments against the accuracy of such predictions include: (1) large individual variability in soft tissue response (eg, soft-tissue thickness, tonicity, posture, muscle pull, etc); (2) differences in skeletal stability among patients operated with the same procedure; (3) difficulty in executing the surgical procedure exactly as planned; and (4) inaccuracies inherent in the cephalometric method itself used to generate the predictions. Realizing these shortcomings while acknowledging the clinical usefulness of treatment simulations, there clearly is room for further improvement and refinement of the currently available database. The large number of patients followed in the present study permitted dividing the sample according to gender and magnitude of surgical repositioning and investigating the impact of skeletal relapse. Analyzing these variables revealed influences in postsurgical soft tissue response with possible clinical implications.

Are soft tissue changes accompanying a large mandibular setback as predictable as those with a small setback? Contrary to what may be expected, the results of the present study indicate that soft tissue changes following small setbacks are less predictable compared to large setbacks. This is clearly reflected by the large standard deviations of the percentages for the small setback group (Table 9). The reason for the poor predictability with small setbacks may be related to the fact that there usually is a relatively greater rotation of the distal segment with a greater component of vertical repositioning in such cases. Results of the present study indicate that changes in the vertical dimension are less predictable than in the horizontal plane. The influence of the correction of the anterior crossbite and normalization of incisal relationship may also be more influential than the setback per se. Such findings imply extra skepticism when using computerized predictions in cases to receive small mandibular setbacks. Statistical analysis also indicates that soft-tissue to hard-tissue ratios for the lower lip and chin gradually decreased with decreasing magnitude of setback (P < .05).

The other question is related to the effect of gender on soft tissue changes. The use of ratios rather than absolute measurements in the present study has eliminated the effect of size differences between males and females. Our findings indicate that soft tissue movement in response to skeletal repositioning is somewhat greater in females than in males (statistically significant for the upper lip and chin, P < .05). The issue of gender difference has previously been addressed by a study on a Chinese sample.18 Interestingly, the authors found statistically significant differences (P < .05) between the sexes in soft-to-hard tissue change ratios for the lower lip and the chin, with females demonstrating higher ratios than males (12% and 11% difference for the soft tissue chin and lower lip, respectively). The authors attributed these differences to the greater soft tissue thickness in males compared to females. What is the clinical relevance of these findings? Higher ratios for females compared to males observed in the present study (20% and 14% difference for the upper lip and soft tissue chin, respectively) indicate that utilization of separate prediction data based on gender would probably be a step forward towards a more accurate prediction.

Another factor investigated was soft tissue thickness and preoperative morphology. Results of the present study reveal associations between the preoperative thickness of both the upper and lower lip and the net change in thickness (r = 0.54 and 0.36 P < .001, respectively) in the sense that the greater the preoperative soft tissue thickness, the greater the expected change. A significant correlation was also found between the preoperative mentolabial fold depth and the net change in its depth (r = 0.57 P < .001). Even for the areas where a relationship between preoperative morphology and soft tissue thickness apparently exists, correlation coefficients are probably too weak to provide clinically useful predictions. This is in agreement with the conclusions of Gjørup and Athanasiou15 and Chunmaneechote and Friede17 who found low regression coefficients when testing if the soft tissue thickness at lip and chin regions could act as predictors of the ratios of soft to hard tissue changes.

An additional factor to consider when predicting long-term profile changes following orthognathic surgery is the effect of skeletal relapse. Previous studies providing prediction data for mandibular setback surgery have calculated the ratios between the amount of change in hard and soft tissues over the same interval (net soft tissue change to net skeletal setback) and stated that the soft tissue chin and mentolabial fold generally follow their corresponding bony structures in an almost 1:1 relation.6–19 When calculated in the same way, our data also demonstrate a ratio close to 1:1. In the present study we attempted to provide, besides the soft to hard tissue ratios calculated over the same interval, alternative ratios that incorporate the effect of mean long-term skeletal relapse (Table 8). When relapse was accounted for, ratios for all mandibular structures dropped, the drop gradually increasing from the lower lip (12%) to the mentolabial fold (20%) to the soft tissue chin (27%). By expressing the long-term soft tissue changes as a percentage of surgical skeletal setback, rather than as a percentage of the long-term skeletal change, prediction ratios are obtained that are likely to generate a more realistic estimate and one that is more appropriate to demonstrate to the patient.

Despite several sources of inaccuracy, a profile prediction for mandibular setback surgery continues to be a valuable tool as it still can provide a useful estimate that facilitates communication and treatment planning. Taking the various factors investigated in the present study in consideration, different soft to hard tissue algorithms may be created for various situations. Thus different values may for example be applied for a female with a large setback as compared to a male where a small setback is planned. If in addition, average relapse is also accounted for, it is likely that an even more realistic long-term prediction is obtained. Since the newer versions of computerized prediction software allow the user to modify the soft to hard tissue ratios, it is recommended that these data should be adjusted, based on gender, magnitude of skeletal repositioning and each team's own long-term stability data for mandibular setback surgery. A future study designed to compare computerized predictions based on these modifications with the actual postoperative patient profiles should be undertaken to confirm the usefulness of the present findings.