INTRODUCTION

Facial embryology usually follows a symmetrical development pattern. However, the human face does not display perfect symmetry. Asymmetry may be considered subclinical, also named relative symmetry or mild asymmetry, or may be moderate or severe, demanding orthodontic and or orthopedic intervention when it affects facial and smile esthetics.^1–3

The prevalence of facial asymmetries reported in the literature varies between 11% and 37%.^2–6 When evaluation is made through more strict or accurate diagnostic methods, the prevalence of asymmetry can approach or exceed 50%.^7–10 In determining facial asymmetry, it is known that deviations in the mandible are the most striking characteristic of the disharmony, particularly the lateral displacement of the chin in relation to the midsagittal plane.^2,7,11

Mandibular asymmetry (MA) is frequently found in subjects affected by trauma, syndromes, or other disorders of the craniofacial area.¹² However, there is limited knowledge of factors associated with idiopathic asymmetries occurring without accompanying acquired disease or congenital alterations.¹ Epidemiological data on these asymmetries are not abundant in the literature. Among the existing studies, most^4–6 analyzed dentofacial deformities as a whole and cited MA among those. Few^2,3,8,10 investigated MA specifically and evaluated associated factors with large samples. Moreover, those studies did not demonstrate how substantial the associations were.

Therefore, this study aimed at estimating the prevalence of different intensities of MA in orthodontic and orthognathic surgery patients with complete dentitions. Multiple logistic regression was also used to determine demographic and skeletal factors associated with this disharmony.

MATERIALS AND METHODS

For this cross-sectional study, the sample was composed of cone-beam computed tomography (CBCT) images of 1178 individuals, using the database of a service center for dental diagnosis and planning (Compass3D, Belo Horizonte, MG, Brazil), which receives tomographic images from all over the country. The images were obtained from orthodontic and orthognathic surgery patients between 2011 and 2013.

Information from the tomographic images was used exclusively for scientific purposes, fully preserving the anonymity of the patients. Institutional review board approval was obtained before the study was initiated (protocol number 771.293).

Initially, a pilot study with 120 individuals who were not participating in the main study was conducted in order to obtain the proportion of MA in subjects with different sagittal jaw relationships. Epi Info version 7 software (CDC, Atlanta, Ga) was used, evaluating the association between the exposure variables and MA. There was an expected prevalence of 40% of MA in unexposed subjects, using a 95% confidence interval and 80% statistical power. Assuming a proportion of unexposed (skeletal Class I) to exposed (skeletal Classes II and III) of 1:1 and a minimum prevalence ratio of 1.25, a sample size of 1178 individuals would be adequate.

For inclusion in the sample, the following criteria were used: CBCT images requested because of clinical need or because conventional radiography could not meet clinical requirements, thereby conforming to the SEDENTEXCT project and the American Academy of Oral and Maxillofacial Radiology guidelines;^13,14 patients 19 through 60 years of age; no missing teeth other than third molars; and images acquired using the same brand of CBCT machine (i-CAT, Imaging Sciences International, Hatfield, Pa). Exclusion criteria were: prior history of facial fractures or facial surgery; degenerative disease involving the temporomandibular joint; or craniofacial anomalies.

To perform the exams, the CBCT scanner was adjusted to operate under the following specifications: extended field of view (16 × 22 cm or 17 × 23 cm) 120KvP, 3-8mA, and 0.4mm³ voxel size. Patients were asked to occlude at maximum intercuspation and relax their lips.

The DICOM files were imported into SimPlant Ortho Pro 2.0 software (Materialise Dental, Leuven, Belgium), which is capable of providing exact values for the measurements of choice. Landmarks were located using multiplanar reconstruction slices, and a measurement scale of 0.01 mm and 0.01° was used. The landmarks and reference planes used in this study are described in Table 1. The tomographic measurements evaluated are described in Table 2 and illustrated in Figure 1A,B,C.

Table 1 Landmarks and Reference Planes Used in the Study

View Fullscreen

Table 2 Tomographic Measurements Evaluated

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

The outcome was MA, which was later categorized as a qualitative polytomous ordinal variable (relative symmetry, moderate asymmetry, or severe asymmetry). The sample was divided into three groups on the basis of MA intensity by analyzing chin laterality^2,7,11 independently if deviated to the left or right. Asymmetry was defined as the extent to which the patient's gnathion was displaced from the midsagittal plane. Patients with displacements of 2 mm or less were defined as exhibiting relative symmetry.^9,15 Patients whose gnathion was displaced by more than 2 mm and up to 4 mm were defined as exhibiting moderate asymmetry. Patients with gnathion displacement from the midsagittal plane greater than 4 mm were defined as exhibiting severe asymmetry.^7,11,16

The exposure variables and their respective forms of analysis were:

• Demographic: Sex (male or female), Age 19 through 60 years of age (with subjects divided into two groups relative to the median age of 29 years as younger or older).

• Skeletal: Side of mandibular displacement determined by displacement of gnathion from the midsagittal plane (right or left), Sagittal jaw relationship determined by the ANB angle and classified as Class I (from 0° to 4.5°), Class II (>4.5°) or Class III (<0°), as suggested by Tweed.¹⁷ Vertical skeletal relationship determined by the N-Ba PtLeft-Gn angle and classified as horizontal (>93°), normal (from 87° to 93°), or vertical (<87°) as suggested by Ricketts.¹⁸ Cranial base angle determined by the N-S-Ba angle as acute (<127°), normal (from 127° to 136°) or obtuse (>136°) as suggested by Björk.¹⁹ Maxillary asymmetry (absent or present) determined by the distance from the anterior nasal spine to the midsagittal plane (when this distance exceeded 2 mm, the patients were defined as having maxillary asymmetry as suggested by Haraguchi et al.)⁷

Since tomographic measurements were made by three examiners, the method error was determined by the intraobserver and interobserver intraclass correlation coefficients (ICC). The three examiners analyzed 10% of the sample at two different times, with a two-week interval. The intraobserver and interobserver ICC value was >0.90 for the evaluated measurements (considered highly reliable), except for the ANB angle which displayed an interobserver ICC of 0.89 (considered reliable). The results demonstrated good repeatability and reproducibility of the method.

Statistical analyses were conducted using STATA version 13.0 (StataCorp, College Station, Tex). To verify the association between exposure variables and mandibular asymmetries, a bivariate analysis was initially conducted using the chi-square test (X²). Simple and adjusted ordinal logistic regression (partial proportional odds model) was then performed. This model was chosen because the outcome was polytomous and categorized ordinally. Only exposure variables with P < .20 in the bivariate analysis were entered into the multiple logistic regression analysis. The final model was used to estimate OR for those variables selected after adjustment and differences were classified as statistically significant for variables with P < .05 and a 95% confidence interval (95% CI).

RESULTS

The mean age of the subjects was 32.28 (SD = 11.33). Displacement of gnathion in relation to the midsagittal plane was 2.51 mm (SD = 2.69) in absolute terms, varying from 19.94 mm for the right side to 21.49 mm for the left side. Other characteristics of the exposure variables are shown in Table 3. Table 4 shows that the prevalence of relative mandibular symmetry, moderate asymmetry, and severe asymmetry on the total sample was 55.2% (95% CI 52.3–58.0), 27.2% (95% CI: 24.8–29.9), and 17.6% (95% CI: 15.5–19.8), respectively.

Table 3 Characteristics of Sampled Individuals

View Fullscreen

Table 4 Bivariate Analysis of the Association Between Mandibular Asymmetries and the Exposure Variables, Using the X² Test

View Fullscreen

From the bivariate analysis (Table 4), the variables age (P = .151), side of mandibular displacement (P = .008), sagittal jaw relationship (P = .074), and maxillary asymmetry (P < .001) were included in the regression model. In the adjusted regression model (Table 5), it was observed that mandibular asymmetries were independently associated only with the side of mandibular displacement (P = .006) and maxillary asymmetry (P < .001). For the ordinal logistic regression analysis, both groups of mandibular asymmetries were compared to the relative symmetry group.

Table 5 Unadjusted and Adjusted Odds Ratio (OR) and 95% Confidence Intervals for Moderate and Severe Mandibular Asymmetries Among Adults According to Demographic and Biological Characteristics

View Fullscreen

When compared to subjects with symmetry, the odds of moderate MA were 1.5 times higher when there was a mandibular deviation to the left (OR = 1.50; 95% CI: 1.01–2.24) and 2.04 times higher when maxillary asymmetry was present (OR = 2.04; 95% CI: 1.11–3.76). The odds of severe MA were 2.09 times higher if there was mandibular deviation to the left (OR = 2.09; 95% CI: 1.27–3.44) and 4.93 times higher when maxillary asymmetry was present (OR = 4.93; 95% CI: 2.64–9.20) compared to subjects with mandibular symmetry.

The model showed an adjusted coefficient of determination (R²) of 0.076 (Table 5), which means that only 7.6% of MA could be explained by the model of variables analyzed in this study.

DISCUSSION

Epidemiological studies are often highly clinically relevant since they enable professionals to know the frequency of disharmonies and also to identify specific groups of people who are prone to the alteration. This epidemiological study on non-syndromic and non-pathological mandibular asymmetries had two main objectives: to estimate the prevalence of different intensities of MA in orthodontic and orthognathic surgery patients and to investigate demographic and skeletal factors associated with these disharmonies. Combined with the use of regression analysis, some uncertainties surrounding MA may be better understood. Prevalence values of 55.2%, 27.2%, and 17.6% were observed for relative mandibular symmetry, moderate MA, and severe MA, respectively. Mandibular asymmetries were independently associated with the variables “side of mandibular deviation” and “maxillary asymmetry.”

The methodology used to define the midsagittal plane in this study was previously validated by Damstra el al.²⁰ Once this plane was determined, mandibular asymmetry was classified into three groups: relative mandibular symmetry, moderate MA, and severe MA. According to some scholars,^7,11,16 the clinical expression of asymmetry only occurs when there is a skeletal deviation of at least 4 mm. Asymmetries below that can be verified but may not require comprehensive treatment. In other words, human sensitivity for perceiving facial imbalance increases when the skeletal deformation is close to or larger than 4 mm. However, the expression of asymmetry or its masking depends on individual characteristics such as the thickness of the soft tissue in the area. Because of that, some other authors consider there to be facial asymmetry in cases with skeletal deviations larger than 2 mm.^7,9 Thus, this study evaluated asymmetry and divided it into three distinct categories.

For statistical analysis of the data in this study, ordinal regression models were used. These models enable researchers to estimate the odds of occurrence of an adverse event, such as MA. In ordinal logistic regression, the OR can be interpreted as the number of times the odds of the outcome occurring increases when the exposure variable is modified in a unit/category. Thus, the OR is used as a measure of association intensity.²¹ When compared to subjects with symmetry, the odds of moderate MA were 1.5 times higher when there was a lateral deviation to the left and approximately two times higher when maxillary asymmetry was present. The odds of severe MA was about two times higher when there was a lateral deviation to the left and approximately five times higher when maxillary asymmetry was present, compared to subjects with mandibular symmetry. Other studies also pointed out a relationship between the presence of MA and lateral deviation to the left,^2,8 as well as the concomitant presence of maxillary asymmetry.^7,15 However, those studies did not demonstrate how intense the association was.

The higher frequency of deviation of gnathion to the left may be explained by the dominant growth potential of the right side of the face given the dominance of the right side when evaluating the dimensions of the skull and brain.^7,12 Another possible mechanism causing this facial laterality could be imbalanced development of cells during embryogenesis since asymmetry exists in cell polarization and division.¹²

As for maxillary asymmetry, which was measured by the lateral deviation of anterior nasal spine (ANS), it has been reported that it is usually of a smaller magnitude than that of MA,^2,7,11 suggesting that most maxillary asymmetries are secondary to asymmetric mandibular growth.¹ It is important to mention that, in this study, maxillary asymmetry was evaluated by the lateral displacement of ANS to the midsagittal plane, as was also done by Haraguchi et al.⁷ However, since it is known that mandibular asymmetry is also related to vertical differences in maxillary heights¹⁶ (also described as maxillary cant), future studies should be designed to analyze this variable.

The findings from this study demonstrated that there was no independent association between MA and sex, age, sagittal jaw relationship, vertical skeletal relationship, or cranial base angle in the evaluated subjects. Such results are consistent with most reports in the literature as to sex and age.^8–10 However, there were some differences related to the other variables that were analyzed.

Some authors found that facial asymmetry was equally distributed in skeletal Class I, II, and III,⁸ while others demonstrated asymmetry was more frequently associated with Class III,^6,10,22 or less common in Class II.² The current findings showed that severe MA was more frequent in skeletal Class III individuals, but moderate MA was more frequent in Class II individuals. However, these differences were not statistically significant.

Regarding differences related to vertical skeletal pattern, Good et al.²² evaluated 66 orthodontic patients and suggested that MA was related to an increased lower anterior face height. Celik et al.²³ evaluated condylar and ramal vertical asymmetry in 101 adult patients with different vertical growth patterns and concluded that the asymmetry index values were not statistically different between the groups, similar to what was observed in the present study.

Kim et al.¹⁵ observed a larger volume of the cranial base on the side contralateral to the mandibular deviation. However, studies performed by Kwon et al.²⁴ and Baek et al.¹⁶ did not verify morphological differences in the cranial base between symmetric and asymmetric subjects. In evaluating the angle of the cranial base, the current study did not observe an association between this variable and mandibular asymmetry.

In this study, the final regression model explained only 7.6% of the variance. Other variables that may have predictive value should be investigated in additional studies. A limitation of the study was that it was cross-sectional in design, and the temporal relationship between the outcome and predictors could not be defined. The cross-sectional design makes it difficult or impossible to discern the order of occurrence of events in time, which may lead to a so-called reverse causality bias.²¹ Additionally, mandibular functional shifts that could contribute a postural component to mandibular asymmetries were not analyzed.

A large sample size and robust statistical testing were strengths of this study. The purpose was to give professionals a better understanding about the factors associated with mandibular asymmetries and the intensity of these associations.

The results indicated that clinical investigation of skeletal mandibular asymmetries are necessary for orthodontic and orthognathic surgery patients, since their prevalence was found to be quite high. If misdiagnosed, asymmetries may lead to extended treatment time or compromised outcomes. These asymmetries happen independently of the individual's sex, age, sagittal jaw relationship, vertical skeletal relationship, or cranial base angle. However, since the odds of mandibular asymmetries were higher when chin deviation was to the left, as well as when maxillary asymmetry was present, whenever patients present with those characteristics, the existence of mandibular asymmetry should be carefully considered.

CONCLUSIONS

• The prevalence of mandibular asymmetry in adults was 44.8%, of which 27.2% was for moderate MA and 17.6% for severe MA.

• The odds of presenting with MA were significantly higher when the chin was deviated to the left or when maxillary asymmetry was present.