The aims of this longitudinal analysis of untreated monozygotic twins were to investigate the change of the facial soft tissues during growth, to determine the concordance of soft tissue growth patterns between genetically identical twins, and to assess the genetic component of soft tissue development.
The sample consisted of 33 pairs of untreated monozygotic twins (23 male, 10 female) from the Forsyth Moorrees Twin Study (1959-1975); lateral cephalograms taken from 6 to 18 years of age were analyzed at 3-year intervals. Cephalograms were traced, and longitudinal changes in the soft tissue profile between twins were analyzed with intraclass correlation coefficients and linear regression modelling.
The concordance between monozygotic twins at 18 years of age was moderate to high with intraclass correlation coefficients values between 0.37 and 0.87. Additionally, female twins showed higher concordance at 18 years of age than did male twins for all included variables. However, about 10% to 46% of the twin pairs had large differences in their soft tissue parameters, even after the growth period.
Although monozygotic twins possess the same genetic material, differences in the soft tissues were found. This supports the complex developmental mechanism of the human face and the varying influence of genetic and environmental factors.
The facial profiles of 33 pairs of monozygotic twins from childhood to early adulthood were assessed.
The concordance in facial profiles of genetically identical twins is greater after growth is over.
Sex and skeletal configuration might affect the similarity of twins’ faces.
The facial profile varied considerably during growth, even among genetically identical twins.
Even at adulthood, many monozygotic twins showed considerable differences in facial profiles.
Facial morphology and attractiveness have direct impacts on social interactions, psychological development, and other aspects of personal and professional life. Therefore, the development of the human face during the growth period and the attempt to influence it have attracted considerable interest in orthodontics and dentofacial orthopedics.
Research on the dynamic and complex phenomena of growth and development of the facial soft tissues has traditionally been conducted on photographs, anteroposterior or lateral cephalograms, and 3-dimensional facial scans. Growth of the human face, although initially considered to be mainly driven by the growth of the craniofacial bones, is complex, with various facial components developing proportionally, disproportionally, or independent of changes of the underlying bones.
The craniofacial complex is derived from a complex developmental process, where gene expression and molecular interactions play early embryonic roles, whereas hormonal and biomechanical environmental factors act mainly during the later childhood and pubertal growth periods. Important insights into the relative contributions of genetic and environmental factors might be acquired from studying human twins, since they share all or part of their genome, thus enabling partitioning of genetic and environmental components.
An early study with twins and same-sex singletons indicated that many cephalometric variables are under strong genetic control, especially those pertaining to the vertical dimension, and that heritability is stronger in the anterior than in the posterior craniofacial region. This strong genetic influence on the vertical dimension of the face was confirmed in a subsequent study of Chinese female twins, which indicated that early orthodontic intervention would be better aimed toward the anteroposterior than the vertical dimension. This seems, however, to be refuted by a later study on twins that indicated that, although the facial profile resemblance among twins was high, the shape and sagittal position of the mandible are under stronger genetic control than are its size and vertical relationship to the cranial base. A recent cross-sectional study on female twin patients found that genetic factors account for more than 70% of the phenotypic variations of the size of the face, nose, lip prominence, and interocular distance. Finally, it was reported that although size of the face showed signs of potential dominant genetic influence, facial proportions were influenced more by environmental factors.
Variations in the facial proportions through time using longitudinal growth data have been previously reported, but these can be attributed to a combination of both genetic and environmental factors, since unrelated patients were studied. However, to our knowledge, the soft tissue profile has not yet been assessed using monozygotic twins, who per definition share their genetic material. Therefore, the aims of this retrospective longitudinal cephalometric cohort study were to assess the changes in the facial soft tissues during childhood, adolescent, and early adult growth of untreated monozygotic twins to (1) determine the concordance of soft tissue growth patterns between genetically identical twins and (2) assess possible differences between them as a proxy for the genetic and environmental contributions to facial soft tissue development.
Material and methods
Patients for this retrospective cohort study were recruited from the Forsyth Moorrees Twin Study performed from 1959 to 1975 at the Forsyth Infirmary for Children in Boston after appropriate Institutional Review Board approval (Boston University, number H-31945). The original sample contains records from over 500 families of twins with no previous history of orthodontic treatment. Eligible patients for this study were those with (1) white origin; (2) no history of orthodontic treatment, craniofacial anomalies, or chronic systemic disease; and (3) available lateral cephalograms of good quality with the soft tissue profile clearly discernable. All patients were measured at approximately the same times every 3 years from middle childhood to early adulthood: T1 at 6 years, T2 at 9 years, T3 at 12 years, T4 at 15 years, and T5 at 18 years of age.
Six widely used soft tissue measurements were made on the films of 33 pairs of twins at the 5 time points (T1-T5), including facial convexity angle, nasal prominence, nasolabial angle, upper lip length, upper lip thickness, and soft tissue chin thickness ( Fig 1 ). Since 1 aim of this study was to assess the concordance between twins, the analysis was based on the differences between the 2 twins in a pair instead of overall facial convexity and its change or malocclusion type.
All lateral cephalograms were taken with the same device (copy of the Broadbent cephalometer) in a standardized position in centric occlusion and the head aligned in natural head position. This position was stabilized with ear rods and a nasal support to prevent variations in the head position. The focus-to-coronal plane distance was 9 cm, and the film-to-coronal plane distance was 150 cm, which resulted in a constant magnification factor of 6%. The subjects were asked to refrain from swallowing during the radiologic examination, with tongue posture subsequently assessed on the cephalograms to ascertain that no children swallowed during the radiographic examination.
After anonymization of all documents with a unique code, the radiographs were traced by 1 person (M.H.Z.) using software (Dolphin Imaging and Management Solutions, Chatsworth, Calif).
Sample size calculation was performed a priori and aimed to find a clinically significant concordance in the primary outcome (facial convexity) between monozygotic twins with the intraclass correlation coefficient (ICC). Based on previous data, we assumed an ICC of 0.50 between twins of each pair at T5 and aimed to find a minimal statistically significant difference of half a standard deviation with a paired t test. Assuming a change of 2° in facial convexity, with a SD of 4° at T5 from a similar study, α of 5%, and power of 80%, we calculated that a sample of 28 twin pairs was needed, to which 5 more pairs were added to account for any missing patient files.
Descriptive statistics (means and standard deviations) were calculated for all variables. For each cephalometric variable, the concordance of monozygotic twins after growth cessation (T5, 18 years) was assessed by calculating ICC values and their 95% confidence intervals (CI).
Additionally, the absolute difference of each cephalometric variable in each twin pair was calculated. We also conducted mixed-effects linear regressions to calculate whether sex and sagittal jaw relationship (ANB angle) were associated with considerable soft tissue differences between twins, while accounting for repeated measurements (T1-T5) per patient. The regression results were expressed as unstandardized regression coefficients (β) and their 95% confidence intervals.
Finally, to identify twins with considerable differences in their soft tissue parameters, the percentages of twin pairs with absolute differences between twins greater than 1 SD for each variable were calculated. Sex, sagittal jaw relationship (ANB angle), and growth from T1 through T5 were again tested for associations with considerable soft tissue differences with mixed-effects binomial regressions. The regression results were expressed as relative risks (RR) and their 95% confidence intervals.
All analyses were run in Stata SE software (version 14.0; StataCorp, College Station, Tex) with an unadjusted α of 5%, since the study’s scope was based on descriptive analyses of concordance and associated factors.
A random sample of 180 cephalograms was chosen and remeasured from the same assessor (M.H.Z.) after 1 month for repeatability. The repeatability and agreement of the repeated measurements were assessed with the concordance correlation coefficient and the Bland-Altman method, whereas the error of the method was calculated with Dahlberg’s formula.
A total of 33 eligible monozygotic twin pairs (66 subjects) were included in the study. Of those 33 pairs, 23 (70%) were male, and 10 (30%) were female. At T1, 2 pairs of twins were skeletal Class I (defined as having an ANB angle >0° but ≤4°), 24 pairs were skeletal Class II, and 7 pairs were discordant, with 6 pairs having 1 twin Class I and 1 twin Class II, and 1 pair having a mixture of Class I and Class III. However, by T5, the distribution had changed, with 2 pairs having skeletal Class III relationships, 3 mixing Class I and Class III in the pair, 5 having Class I, 12 mixing Class I and Class II, and 11 pairs being both skeletal Class II. Some radiographs were missing at some time points or had poor quality; this resulted to a total of 287 lateral cephalograms, with varying sample sizes throughout the study period.
A concordance between monozygotic twins at 18 years (T5) could be seen, but varied considerably among the included variables, with facial convexity being the most concordant (ICC = 0.87), followed by nasal prominence (ICC = 0.68), nasolabial angle (ICC = 0.70), upper lip length (ICC = 0.69), upper lip thickness (ICC = 0.54), and soft tissue chin thickness (ICC = 0.37) ( Table I ). Additionally, a consistent difference between female and male twins was seen: female twins showed higher concordance at T5 than did male twins for all included variables. Overall, the results indicated that the similarity of monozygotic twins in the soft tissue profile in early adulthood varied among the several variables that were studied.
|Facial convexity||0.87 (0.75-0.95)||0.69 (0.55-0.81)||0.77 (0.62-0.87)||0.63 (0.50-0.75)||0.74 (0.61-0.84)||0.78 (0.61-0.90)|
|Nasal prominence||0.68 (0.44-0.86)||0.40 (0.23-0.59)||0.58 (0.41-0.75)||0.45 (0.31-0.59)||0.48 (0.33-0.64)||0.49 (0.22-0.76)|
|Nasolabial angle||0.70 (0.50-0.87)||0.46 (0.30-0.62)||0.58 (0.39-0.75)||0.47 (0.33-0.61)||0.44 (0.29-0.60)||0.61 (0.37-0.82)|
|Upper lip length||0.69 (0.44-0.87)||0.49 (0.33-0.64)||0.76 (0.61-0.86)||0.56 (0.42-0.69)||0.54 (0.38-0.69)||0.64 (0.41-0.83)|
|Upper lip thickness||0.54 (0.26-0.80)||0.26 (0.13-0.42)||0.51 (0.31-0.70)||0.28 (0.17-0.43)||0.27 (0.14-0.45)||0.50 (0.23-0.76)|
|Soft tissue chin thickness||0.37 (0.10-0.73)||0.39 (0.24-0.56)||0.53 (0.35-0.71)||0.46 (0.32-0.60)||0.40 (0.25-0.56)||0.56 (0.31-0.80)|
The descriptive statistics about the average values and the mean differences between twins for each variable at each time point are given in Table II . Overall, no clear pattern could be seen for differences between twins of each pair from 6 to 18 years (T1-T5). The regression modeling of the differences between twins and their variations during growth showed no clear pattern for most outcomes ( Table III ; Figs 2-4 ). The only exception was the nasolabial angle, where the difference between twins decreased at 15 years compared with 6 years (unstandardized regression coefficient [b] = −2.85°; 95% CI = −5.69-0; P = 0.05). Additionally, differences in the nasolabial angle between twins were associated with sagittal jaw relationship, with larger ANB values leading to greater differences in nasolabial angle between twins (b = 0.54°; 95% CI = 0.14-0.93; P <0.05). Finally, sex was significantly associated with differences in upper lip length between twins; male twins had larger differences in upper lip length than did female twins (b = 0.92 mm; 95% CI = 0.11-1.74; P <0.05).
|Values at each time||Absolute differences between twins in a pair|
|Time||Patients (n)||Mean||SD||Twin pairs (n)||Mean||SD|
|Facial convexity (°)||T1||48||158.10||4.51||24||3.21||3.40|
|Nasal prominence (mm)||T1||48||9.81||3.31||24||1.83||1.32|
|Nasolabial angle (°)||T1||48||117.25||11.12||24||9.11||7.65|
|Upper lip length (mm)||T1||48||21.41||2.24||24||1.51||1.22|
|Upper lip thickness (mm)||T1||48||13.23||1.57||24||1.47||1.30|
|Soft tissue chin thickness (mm)||T1||48||11.48||2.3||24||1.63||1.74|
|Facial convexity (°)||Time||T1||Reference|
|Nasal prominence (mm)||Time||T1||Reference|
|Nasolabial angle (°)||Time||T1||Reference|
|T4||−2.85||−5.69, 0.00||0.05 ∗|
|ANB||0.54||0.14, 0.93||0.007 ∗|
|Upper lip length (mm)||Time||T1||Reference|
|Male||0.92||0.11, 1.74||0.03 ∗|
|Upper lip thickness (mm)||Time||T1||Reference|
|Soft tissue chin thickness (mm)||Time||T1||Reference|
The numbers of twin pairs with considerable differences (>1 SD) in soft tissue variables are shown in Table IV . The percentages of twin pairs ranged from 10% to 46% for the various outcomes with no differences across time points. The only exception was the nasolabial angle, where the differences between twins tended to decrease with time ( P = 0.022). This was also confirmed by the results of the binominal regression ( Table V ); twins at T4 were 77% less likely to have considerably different nasolabial angles (RR = 0.23; 95% CI = 0.07-0.72; P = 0.012) compared with T1 ( Table V ). Additionally, male twin pairs were much more likely to have considerably different upper lip lengths between the 2 twins than were female twin pairs (RR = 3.36; 95% CI = 1.21-9.34; P = 0.020).
|T1 (24 twin pairs)||T2 (33 twin pairs)||T3 (33 twin pairs)||T4 (31 twin pairs)||T5 (22 twin pairs)||P value|
|Variable||n (%)||n (%)||n (%)||n (%)||n (%)|
|Facial convexity||4 (17%)||6 (18%)||5 (15%)||6 (19%)||3 (14%)||0.98|
|Nasal prominence||4 (17%)||7 (21%)||4 (12%)||6 (19%)||5 (23%)||0.85|
|Nasolabial angle||11 (46%)||12 (36%)||11 (33%)||3 (10%)||4 (18%)||0.02 ∗|
|Upper lip length||4 (17%)||8 (24%)||8 (24%)||9 (29%)||8 (36%)||0.63|
|Upper lip thickness||6 (25%)||11 (33%)||10 (30%)||10 (32%)||7 (32%)||0.97|
|Soft tissue chin thickness||7 (29%)||10 (30%)||8 (24%)||4 (13%)||5 (23%)||0.53|