Heritability of facial soft tissue growth in mono‐ and dizygotic twins at 12 and 17 years of age: A retrospective cohort study

Abstract

Objective

The purpose of this investigation of untreated monozygotic and dizygotic twins was to identify the genetic and environmental components to the facial soft tissue growth.

Settings and Sample Population

The sample consisted of 52 untreated monozygotic twins (36 male and 16 female) and 46 untreated dizygotic twins (23 male and 23 female) from the Forsyth Moorrees Twin Study (1959‐1975).

Materials and Methods

Lateral cephalograms were taken at 12 and 17 years of age and traced to analyse facial convexity, nasolabial angle, upper and lower lip thickness, upper and lower lip profile and nose prominence. The genetic and environmental components of variance were analysed with structural equation modelling for multilevel mixed‐effects model.

Results

At 12 years of age, strong additive genetic influence was seen for facial convexity (70%), upper lip profile (66%) and nose prominence (65%), whereas strong dominant genetic components were found for upper lip thickness (56%). Nevertheless, under unique environment influence were nasolabial angle (58%), lower lip profile (51%) and lower lip thickness (64%). At 17 years of age, only upper lip thickness (55%) and nose prominence (84%) were under strong additive genetic control, while the rest of the variables were under strong dominant genetic control. The only exception was lower lip thickness (61%), which is still influenced by the unique environment.

Conclusion

Although monozygotic/dizygotic twins share at least part of their genome, at both times either additive, dominant or environmental components were found. Nevertheless, at 17 years of age most of the variables are either under additive or dominant genetic influence.

1. INTRODUCTION

The last decade has witnessed an increase of interest in the research on facial attractiveness. ¹ Both evolutionary and socialization theories indicate that attractiveness has a massive influence on human development and interaction. ² As a result, not only the ability to forecast but also to influence facial growth is of tremendous interest. The assessment and prediction of dentofacial growth are perhaps the most important, but still the most subjective aspects of clinical orthodontics. ³

Although growth of the human face was initially considered to be mainly driven by the growth of the craniofacial bones, ⁴ it is highly differentiated. Various facial components are growing proportionally, disproportionally or independent from changes of the underlying bones. ⁵ The craniofacial complex is developed through a complex developmental process, where gene expression and molecular interactions play early embryonic roles. However, hormonal and biomechanical environmental factors influence growth mainly during later childhood and pubertal growth periods. ⁶ , ⁷ , ⁸ , ⁹ The shape and size of craniofacial structures are known to be affected by both genetic and environmental factors, ¹⁰ , ¹¹ , ¹² and their complex interactions present a great challenge in modern biology. Nevertheless, there has been controversy in terms of genetic heritability in the vertical and horizontal aspects from previous studies.

Two early studies claimed that vertical measurements had greater heritability than horizontal measurements. ¹³ , ¹⁴ These findings were confirmed by Manfredi et al, ¹⁵ who concluded that many cephalometric variables are under strong genetic control, especially those pertaining to the vertical dimension. Furthermore, they indicated stronger heritability for the anterior than the posterior craniofacial region. These findings were confirmed in a subsequent twin study of Chinese female twins, which reported that early orthodontic treatment would be better aimed towards the sagittal than the vertical dimension. ¹⁶ On the contrary, Townsend and Richard stated that both anteroposterior and vertical aspects of the mandible presented strong heritability. ¹⁷ It is of note here that even though dental arch morphology (width, length and shape) appears to be under dominant heritability influence, traits concerning the maxillomandibular occlusal relationship seem to be mostly determined by environmental factors. ¹⁸ Recent data indicated that, although the facial profile resemblance among twins was high, genetic control is more attributable to the mandible's sagittal than vertical position and the anterior face height demonstrated higher genetic determination compared with posterior face height. ¹⁹ A three‐dimensional population‐based twin study indicated that genetic factors could explain more than 70% of the phenotypic facial variation in facial size, nose, lips’ prominence and interocular distance. ²⁰

Because twin studies can evaluate the genetic influence on facial morphology by minimizing the environmental contribution, cephalometric studies of twins are a very powerful tool in orthodontic diagnosis, treatment planning, growth prediction and estimation of treatment prognosis. Classical twin study methods have been based on comparisons of the differences within monozygotic (MZ) (identical) or dizygotic (DZ) (fraternal) pairs, whereas the amount of the differences is being taken as an indication of the relative genetic impact. ²¹ Therefore, twin studies can shed some light on the role of genetic and environmental factors. ¹⁴ Nevertheless, evidence of the specific contribution of genes and the environment on the development of each craniofacial characteristic remains limited.

No previous investigation has studied the heritability of facial soft tissue growth using a model fitting statistical analysis of lateral cephalometric variables of monozygotic and dizygotic twins. Therefore, the purpose of this retrospective cephalometric cohort study was to determine genetic and environmental impact on facial soft tissue growth using quantitative genetic modelling at 12 and 17 years of age in untreated twins.

2. MATERIAL AND METHODS

2.1. Study sample

Patients for this retrospective investigation were recruited from the Forsyth Moorrees Twin Study sample that was collected from 1959 to 1975 at the Forsyth Infirmary for Children in Boston. The protocol of the study has been approved by the Institutional Review Board (Hersberger et al, 2018). This archive includes nearly 500 twin pairs who registered and came for yearly records. Zygosity determination was determined by serologic testing of 29 factors. All twins were Caucasian and had no previous history of orthodontic treatment. Eligible patients for this investigation were those with (1) no history of orthodontic treatment, craniofacial anomalies or chronic systemic disease and (2) available lateral cephalograms in good quality with the soft tissue profile clearly visible. All patients in this twin sample had been measured annually up until they reached adulthood. For the current investigation, two time points at 12 years and 17 years of age were chosen. The sample consisted of 52 untreated monozygotic twins (36 male and 16 female) and 46 untreated dizygotic twins (23 male and 23 female).

2.2. Cephalometric measurements

All lateral cephalograms were taken in a standardized position in centric occlusion with the same device (copy of the Broadbent cephalometer). The position was stabilized with ear rods and nasal support to prevent variations in the head position. The focus‐coronal plane distance was 150 cm, and the film‐coronal plane distance was 9 cm, ²² which resulted in a constant magnification factor of 6%. The patients were asked to refrain from swallowing during the radiological examination, with tongue posture subsequently assessed on the cephalograms to ascertain that no children swallowed during the radiographic examination. The films were digitally scanned using an Epson Expression 11000XL – Photo Scanner (Epson America, Inc, Long Beach, CA) at a resolution of 300 dpi and 16‐bit grayscale.

After anonymization of all documents with a unique code, the scanned radiographs were traced by two trained and calibrated persons (MHZ and SNP) using OnyxCeph™ (Image Instruments, Chemnitz, Germany). Seven widely used soft tissue measurements were made on the films of 98 siblings at two time points: N‐Sn‐Pog', Cotg‐Sn‐Ls, Pr‐Ls, LbSup_I_Sn‐Pog', LbInf_I_Sn‐Pog', Id‐Li and Sn‐Pr (ǁFH) (Figure 1).

FIGURE 1.

Cephalometric measurements used in this study: A, N‐Sn‐Pog' (facial convexity, in degrees); B, Cotg‐Sn‐Ls (nasolabial angle, in degrees); C, Pr‐Ls (upper lip thickness, in millimetre); D, LbSup_I_Sn‐Pog' (upper lip profile, in millimetre); E, LbInf_I_Sn‐Pog' (lower lip profile, in millimetre); F, Id‐Li (lower lip thickness, in millimetre); and G, Sn‐Pr (nose prominence, in millimetre)

2.3. Statistical analysis

Descriptive statistics included means with standard deviations, while crude differences among MZ and DZ twins with their 95% confidence intervals (CI) were calculated with generalized linear models with standard errors clustered within twin families. Quantitative genetic modelling was used to investigate the heritability of soft tissue morphology. Following the classic twin design, we allowed the variation in the population to be due to three latent components: additive genetic factors (A, estimation of the relative influences of additive genetic factors, the narrow‐sense heritability), non‐additive (dominant) genetic factors (D, dominance and epistatic interactions between loci), environmental factors shared between twins in pairs (C, it is attributed to the non‐genetic sources of variation between individuals that are experienced by multiple individuals in a population; it is typically the largest component of variance in population in natural conditions) and environmental factors unique to an individual (E). A pair of MZ twins was deemed to be genetically identical (ie genetic correlation of 1). ACE and AE models assume that genetic effects are additive (ie there is no epistasis); therefore, by Mendelian inheritance rules, the genetic correlation within a pair of DZ twins is 0.5. ADE models assume interactions between genetic influences inducing non‐additive – dominant – genetic effects, with a genetic correlation within a pair of DZ twins of 0.75 (0.5 additive genetic correlation and 0.25 dominant genetic correlation) (Figure A1). Further assumptions include the lack of assortative mating of the twins’ parents and the equal environments’ assumption for MZ/DZ twins (which was supported by the Forsyth Moorrees Twin Study description), while the AE/ADE models assume no environmental effects shared by both twins of a twin pair (no C). Using these assumptions, we fitted structural equation modelling for multilevel mixed‐effects ACE variance decomposition models in Stata SE 14.2 (StataCorp, College Station, TX). We started by fitting for each variable an ACE, an AE and an ADE model, where the letters indicate that sources were allowed to contribute to the variation. It should be noted that the power of the sample was sufficient to detect additive genetic influence but might too low to detect dominance or shared environment, unless the effect was large. The Akaike information criterion (AIC) ²³ was used to determine the best‐fitting model from the three, where a lower AIC value indicates a better fit of the model to the observed data. Then, the same procedure was done on models adjusted for sex, and again the best‐fitting model was selected from the three. In the end, the best‐fitting crude and adjusted‐for‐sex models were compared to find the most parsimonious model (lowest AIC) that best explained observed variance. Finally, the classical heritability estimate was calculated for all factors as twice the difference of the MZ and DZ correlations for all variables. ²⁴ The analysis was run two times: once at 12 years of age and once at 17 years of age.

2.4. Method error

Intra‐observer method error was assessed using the coefficient of reliability and the method suggested by Bland and Altman. ²⁵ The reliability of the method was tested by tracing and measuring 50 selected lateral cephalograms twice from the same assessor (MHZ and SNP) with a one‐month time interval and has been previously reported. ²⁶

2.5. Sample size calculation

Sample size calculation was performed a priori not for the present study, but a previous study of this project, ²⁶ which aimed to find a clinically significant concordance in facial convexity between solely MZ twins and calculated that a sample of 28 twin pairs would be needed to provide adequate power. The present study included a convenience sample of all eligible patients in the Forsyth Moorrees Twin Study, and its sample was doubled by adding the DZ twins and, therefore, was deemed to be adequately powered.

3. RESULTS

At both times (12 years and 17 years), a total of 98 siblings were included in the examination (Table 1). These pertained to 52 MZ twins and 46 DZ twins with average facial convexity at age 12 of 159.6° (160.0° at age 17), nasolabial angle of 108.2° (109.4° at age 17), upper lip thickness of 3.5 mm (3.9 mm at age 17), upper lip profile of 1.3 mm (1.0 mm at age 17), lower lip profile of 1.0 mm (0.7 mm at age 17), lower lip thickness of 3.6 mm (3.8 mm at age 17) and nose prominence of 3.3 mm (4.3 mm at age 17). Insignificant differences were found between MZ and DZ twins for all variables at 12 and 17 years of age (<1° and <1m in all case), and only the variable of facial convexity showed a statistically significant difference of 3.2° (as the 95% CI excluded zero) at both 12 and 17 years of age. Crude or adjusted‐for‐sex models were initially run (Table A1) and compared in Table 2. The models with the lowest AIC values were chosen, and only the results of the best‐fitting model have been taken into consideration. In most of the cases at 12 years of age, an ADE (additive‐dominant‐unique environmental) model better explained the observed variance, which indicated either non‐additive (dominant), additive or unique environment genetic influence. The sole exception was lower lip profile and facial convexity, for which an ACE model or AE model was more appropriate. Models adjusted for sex were chosen for all variables apart from the measurement of nasolabial angle. Under strong additive genetic influence were facial convexity (70%), upper lip profile (66%) and nose prominence (65%). In contrast, under strong dominant genetic influence was upper lip thickness (56%). Interestingly, environmental factors common for the twins accounted for 49% of the variance for lower lip profile. In contrast, under unique environment influence were lower lip thickness (64%), nasolabial angle (58%) and lower lip profile (51%). In general, high classical heritability estimates were seen for all variables, except lower lip thickness and lower lip profile, while upper lip thickness had the highest estimate.

TABLE 1.

Average cephalometric measurements for the 98 MZ/DZ twins at 12 years and 17 years of age

		12 years of age			17 years of age
Variable		n	Mean (SD)	Difference (95% CI)*	n	Mean (SD)	Difference (95% CI)*
Facial convexity (N’‐Sn‐Pog') (°)	MZ	52	158.1 (5.1)	3.2 (0.8, 5.6)	52	159.9 (6.1)	3.2 (0.1, 6.2)
	DZ	46	161.3 (4.7)		46	163.1 (6.3)
Nasolabial angle (Cotg‐Sn‐Ls) (°)	MZ	52	108.1 (9.2)	0.6 (−3.8, 5.0)	52	109.1 (9.8)	0.7 (−3.7, 5.0)
	DZ	46	108.5 (9.9)		46	109.8 (9.2)
Upper lip thickness (Pr‐Ls) (mm)	MZ	52	3.5 (0.5)	0 (−0.2, 0.3)	52	3.9 (0.6)	−0.1 (−0.3, 0.2)
	DZ	46	3.5 (0.5)		46	3.8 (0.6)
Upper lip profile (LbSup_I_Sn‐Pog') (mm)	MZ	52	1.3 (0.6)	−0.1 (−0.4, 0.2)	52	1.0 (0.6)	−0.2 (−0.4, 0.1)
	DZ	46	1.2 (0.5)		46	0.9 (0.5)
Lower lip profile (LbInf_I_Sn‐Pog') (mm)	MZ	52	1.1 (0.6)	−0.3 (−0.5, 0)	52	0.8 (0.7)	−0.2 (−0.5, 0)
	DZ	46	0.9 (0.6)		46	0.6 (0.5)
Lower lip thickness (Id‐Li) (mm)	MZ	52	3.6 (0.5)	−0.1 (−0.3, 0.2)	52	3.8 (0.6)	−0.1 (−0.4, 0.1)
	DZ	46	3.6 (0.5)		46	3.7 (0.7)
Nose prominence (Sn‐Pr) (mm)	MZ	52	3.3 (0.7)	0 (−0.3, 0.4)	52	4.3 (0.7)	0 (−0.3, 0.4)
	DZ	46	3.3 (0.6)		46	4.3 (0.7)

Abbreviations: CI, confidence interval; DZ, dizygotic twins; MZ, monozygotic twins; SD, standard deviation.

*From generalized linear model accounting for within‐twins clustering with robust standard errors

TABLE 2.

Parameter estimates of genetic and environmental effects on cephalometric measurements at 12 years of age

			A		C [D]		E
Variable	Adjusted	Model	b (SE)	%	b (SE)	%	b (SE)	%	AIC	rMZ	rDZ	h2
Facial convexity (°)	–	AE	18.8 (4.7)	72%	–	–	7.3 (2.5)	28%	581.53	0.72	0.37	0.70
	–	ADE	18.8 (<0.1)	72%	[<0.1] (<0.1)	[<1%]	7.3 (2.3)	28%	581.53	0.78	0.37	0.70
	Sex	AE*	17.7 (4.6)	70%	–	–	7.4 (2.6)	30%	581.28	0.71	0.35	0.71
	Sex	ADE*	17.7 (<0.1)	70%	[<0.1] (<0.1)	[<1%]	7.4 (2.3)	30%	581.28	0.71	0.35	0.71
Nasolabial angle (°)	–	ADE*	37.4 (<0.1)	42%	[<0.1] (<0.1)	[<1%]	52.7 (14.0)	58%	717.21	0.38	0.31	0.16
	Sex	ADE	36.7 (<0.1)	41%	[<0.1] (<0.1)	[<1%]	52.9 (13.8)	59%	718.83	0.38	0.30	0.16
Upper lip thickness (mm)	–	ADE	<0.1 (<0.1)	<1%	[0.2] (<0.1)	[59%]	0.1 (<0.1)	41%	144.03	0.59	0.15	0.88
	Sex	ADE*	<0.1 (<0.1)	<1%	[0.1] (<0.1)	[56%]	0.1 (<0.1)	44%	142.43	0.55	0.15	0.81
Upper lip profile (mm)	–	ADE	0.2 (<0.1)	70%	[<0.1] (<0.1)	[<1%]	0.1 (<0.1)	30%	143.98	0.71	0.41	0.60
	Sex	ADE*	0.2 (<0.1)	66%	[<0.1] (<0.1)	[<1%]	0.1 (<0.1)	34%	135.33	0.67	0.38	0.57
Lower lip profile (mm)	–	ADE	0.2 (<0.1)	50%	[<0.1] (<0.1)	[<1%]	0.2 (<0.1)	50%	177.81	0.47	0.46	0.02
	Sex	ACE*	<0.1 (<0.1)	<1%	0.2 (0.1)	49%	0.2 (<0.1)	51%	175.18	0.47	0.48	‐0.01
Lower lip thickness (mm)	–	ACE	<0.1 (<0.1)	<1%	0.1 (<0.1)	33%	0.2 (<0.1)	67%	149.48	0.30	0.37	‐0.13
	Sex	ADE*	0.1 (<0.1)	36%	0.1 (<0.1)	[<1%]	0.2 (<0.1)	64%	149.41	0.32	0.34	‐0.03
Nose prominence (mm)	–	AE	0.3 (0.1)	70%	–	–	0.1 (<0.1)	30%	182.65	0.73	0.41	0.64
	–	ADE	0.3 (<0.1)	70%	[<0.1] (<0.1)	[<1%]	0.1 (<0.1)	30%	182.65	0.73	0.41	0.64
	Sex	ADE*	0.2 (<0.1)	65%	[<0.1] (<0.1)	[<1%]	0.1 (<0.1)	35%	172.10	0.63	0.42	0.42

Abbreviations: A, additive genetic variance; AIC, Akaike information criterion; b, regression coefficient; C, shared environment variance; D, dominant genetic variance; E, unique environment variance; h², classic heritability; rDZ, correlation of dizygotic twins; rMZ, correlation of monozygotic twins; SE, standard error.

*In italics is given the most parsimonious model that best describes each cephalometric variable.

In the secondary analysis at 17 years of age, the same number of high‐quality radiographs was evaluated. Small differences were found between MZ and DZ siblings for all variables, and as before, only the facial convexity variable showed a statistically significant difference of 3.2° (the 95% CI excluded zero). Crude or adjusted‐for‐sex models were initially selected (Table A2) and compared in Table 3. Models adjusted for sex were chosen for facial convexity, upper and lower lip thickness and upper lip profile. Half of the variables (nasolabial angle, upper and lower lip profile and lower lip thickness) were better explained through an ADE model, whereas for the remaining variables (upper lip thickness, facial convexity and nose prominence) an AE model was more appropriate to explain the observed variance, which indicated either additive or unique environmental genetic influence. Under strong dominant genetic influence at 17 years of age were nasolabial angle (71%) and upper (60%) and lower (51%) lip profile. On the contrary, facial convexity (83%), upper lip thickness (55%) and nose prominence (84%) were under strong additive genetic influence. In contrast, unique environmental factors accounted for 15%‐61% of the observed variance, with lower lip thickness being the most affected (61%). Overall, high classical heritability estimates were observed for all variables except for lower lip thickness, with highest value seen for nasolabial angle.

TABLE 3.

Parameter estimates of genetic and environmental effects on cephalometric measurements at 17 years of age

			A		C [D]			E
Variable	Adjusted	Model	b (SE)	%	b (SE)	%	b (SE)		%	AIC		rMZ		rDZ		h2
Facial convexity (°)	–	AE	34.9 (5.7)	85%	–	–	6.3 (1.6)		15%	611.13		0.84		0.25		1.17
	Sex	AE*	30.8 (5.7)	83%	–	–	6.2 (1.6)		17%	605.21		0.83		0.33		1.01
Nasolabial angle (°)	–	ADE*	<0.1 (<0.1)	<1%	[63.1] (<0.1)	[71%]	25.2 (5.3)		29%	703.89		0.75		0.04		1.42
	Sex	ADE	<0.1 (<0.1)	<1%	[63.0] (<0.1)	[71%]	25.2 (5.3)		29%	705.77		0.75		0.04		1.41
Upper lip thickness (mm)	–	ADE	<0.1 (<0.1)	<1%	[0.3] (<0.1)	[69%]	0.1 (<0.1)		31%	164.50		0.71		0.15		1.12
	Sex	AE*	0.1 (<0.1)	55%	–	–	0.1 (<0.1)		45%	133.33		0.52		0.39		0.27
Upper lip profile (mm)	–	ADE	<0.1 (<0.1)	<1%	[0.2] (<0.1)	[66%]	0.1 (<0.1)		34%	140.74		0.72		0.14		1.16
	Sex	ADE*	<0.1 (<0.1)	<1%	[0.1] (<0.1)	[60%]	0.1 (<0.1)		40%	131.88		0.67		0.16		1.02
Lower lip profile (mm)	–	ADE*	<0.1 (<0.1)	<1%	[0.2] (<0.1)	[51%]	0.2 (0.1)		49%	176.03		0.55		0.07		0.95
	Sex	ADE	0.1 (0.1)	24%	[0.1] (<0.1)	[25%]	0.2 (0.1)		51%	176.13		0.54		0.13		0.82
Lower lip thickness (mm)	–	ADE	<0.1 (<0.1)	<1%	[0.2] (<0.1)	[50%]	0.2 (0.1)		50%	184.10		0.47		–0.12		0.30
	Sex	ADE*	<0.1 (<0.1)	<1%	[0.1] (<0.1)	[39%]	0.2 (<0.1)		61%	163.43		0.40		–0.11		0.26
Nose prominence (mm)	–	AE*	0.4 (0.1)	84%	–	–	0.1 (<0.1)		16%	186.16		0.86		0.33		1.06
	Sex	ADE	0.3 (0.5)	53%	[0.2] (<0.4)	[31%]	0.1 (<0.1)		16%	189.15		0.85		0.32		1.07

Abbreviations: A, additive genetic variance; AIC, Akaike information criterion; b, regression coefficient; C, shared environment variance; D, dominant genetic variance; E, unique environment variance; h², classic heritability; rDZ, correlation of dizygotic twins; rMZ, correlation of monozygotic twins; SE, standard error.

*In italics is given the most parsimonious model that best describes each cephalometric variable.

Lastly, the analysis of the repeated measurements showed high reliability and small limits of agreement in all instances, which supported the robustness of the method.

4. DISCUSSION

In this investigation, we analysed seven cephalometric measurements in 52 participants at 12 years of age, and 46 participants were studied at 17 years of age. Apart from facial convexity, all variables showed small differences in the average cephalometric measurements. This finding is in accordance with a previous study, which indicates a substantial concordance for most of these variables. ²⁶ This is somewhat understandable since twins brought up in the same family are subject to essentially identical environmental influences and have an identical or at least similar genome. ²⁷ However, classical heritability estimates showed that not all the examined measurements presented similar heritability, while lower lip thickness presented the least heritability at both time points (−0.03 at 12 years and 0.26 at 17 years).

According to the results of the present investigation, the so‐called ADE model was considered to better explain the observed variance for most of the variables at 12 years of age. This implies that either deviations from the mean phenotype are due to the inheritance of a particular allele and this allele's relative effect on the phenotype, or that dominant genetic variance involves deviation due to interactions between alternative alleles at a specific locus. At 17 years of age for half of the variables, an ADE model, and for the rest an AE model, was considered to better explain the observed variance. However, a recent published study ¹⁹ regarding the heritability of mandibular cephalometric variables with completed craniofacial growth found that the best‐fitting models for almost all cephalometric variables were either an ACE or an AE model. This might be because their measurements were all skeletal and not soft tissue measurements. Furthermore, it is useful to do an adjustment for sex as well. Since heritability of craniofacial components cannot only be influenced by age but also by sex, ²⁸ models adjusted for the latter were chosen for most of the variables, as they fitted the data better than the conventional models. At 12 years of age, models modified for sex were chosen for all the variables apart from nasolabial angle. At 17 years of age, models adjusted for sex were chosen for facial convexity, upper lip thickness, upper lip profile and lower lip thickness.

However, in our investigation, the ACE model was deemed to be appropriate for lower lip profile at 12 years only. This indicates that either additive genetic factors or shared respective specific environmental factors play an important role in the expression of this specific variable, whereas non‐additive genetic factors are of smaller importance. Accordingly, it changed from E to D, which indicated, at 12 years of age, a mainly unique environment (51%) and, at 17 years of age, a mostly dominant (51%) genetic influence. It was particularly noteworthy to see that the main contributions of the factors A, C, D and E changed for most variables between 12 and 17 years of age. The contribution of upper lip thickness changed from D to A, which indicated, at 12 years of age, a mainly dominant (56%) and, at 17 years of age, a mostly additive (55%) genetic influence. A possible explanation for this could be that up to 12 years of age the upper lip is mostly determined by the sagittal growth of the maxilla, which has mostly ceased at 17 years of age and sagittal mandibular growth or vertical changes mostly influence the position/thickness of the upper lip. However, the contribution of upper lip profile changed from A to D, which showed, at 12 years, a mainly additive (66%) and, at 17 years of age, a mostly dominant (54%) genetic impact. On the contrary, lower lip thickness was mainly under environmental control at both time points; consequently, the classical heritability estimates were lowest for this variable at 12 and 17 years of age. Nevertheless, the contribution of lower lip profile changed from E to D, which showed, at 12 years, a mainly environmental (51%) influence, and at 17 years of age, a mostly dominant (51%) genetic impact. This is in accordance with a previous study whether anterior vertical parameters are more heritable than posterior ones ¹⁵ or whether the lower third of the face is under strong genetic control. ²⁹

In addition, facial convexity and nasal prominence were mainly under additive genetic control at both time points. In both cases, they were mainly influenced by additive genetic influence with 70% and 65%, respectively, at 12 years of age and 83% and 84%, respectively, at 17 years of age.

The change of inheritance for the nasolabial angle was interesting as well. It shifted from E to D, which indicated, at 12 years, a mainly environmental (58%) and, at 17 years, a mostly dominant (71%) genetic influence. Therefore, it is no surprise that classical heritability estimates at 17 years of age were highest for this variable. These findings were confirmed in a recent published study, ²⁶ which pointed out that the concordance between facial convexity, nasolabial angle and nasal prominence was high at the end of craniofacial growth. This might be related to the strong genetic component of the nose, ³⁰ something that seems to be confirmed by genome‐wide association assessments. ³¹ Since identical twins are considered to have almost identical genetic material, it might be assumed that high concordance rates for specific soft tissue variables indicate that these traits are determined to a higher degree genetically and less by environmental influence. Pertinent literature of craniofacial growth demonstrated increasing heritability estimates of cephalometric variables with the age. ³² Apart from upper lip thickness, where the heritability declined from 0.81 to 0.27, our results basically correspond to these findings. Consequently, comparison of hereditary characteristics is more valid when the growth is completed. Although there is research reporting lifelong craniofacial growth, the remaining skeletal growth two years after completion of the pre‐adolescent growth spurt is insignificant, not only from a clinical but also theoretical point of view. ³³ Results of previous studies indicated high heritability, with vertical skeletal variables showing more heritability than horizontal ones. ¹³ , ¹⁴ , ³¹ , ³⁴ This is in contrast to the results of a previous investigation reporting that horizontal linear variables are more determined by genetic factors, than vertical variables. ¹⁹ Although our study did not analyse classical horizontal or vertical measurements, facial convexity is strongly related to the sagittal dimension. Therefore, our result for this specific variable is on par with the cited study. Additionally, some investigations indicated that the lower third of the face seems to be under strong genetic control. ²⁹ , ³⁵ , ³⁶ However, our data indicated that the contribution of lower lip profile changed from mainly environmental influence to mostly dominant genetic influence at 17 years of age. Nevertheless, for lower lip thickness, environmental influence was pre‐dominant at both time points. In summary, other than lip thickness, all other variables were either under additive or under non‐additive (dominant) genetic influence at 17 years of age. It might be assumed that certain characteristics were determined more by genetic and less environmental influences. This might indicate that room for individual development exists in MZ twins. It is known from pertinent research that environmental factors like the function and dysfunction of the tongue, cheeks and lips, oral muscles and impaired oral breathing or certain disorders of mastication and body posture play an important role in the development of occlusion. ¹² , ³⁷ , ³⁸ , ³⁹ Our results support these clinical observations related to lower lip thickness and lower lip profile. However, the remaining variables describing the lip position (upper lip profile and upper lip thickness and nasolabial angle) showed higher heritability. This indicates an existence of an integrated balance between morphologic units in the dentofacial complex, which are under strong genetic control, and units that may accommodate more to environmental factors for final establishment of the variety of occlusion ³⁵ and soft tissue profile.

The strengths of the present study included the evaluation of both MZ and DZ twins that enabled a more accurate partitioning of the genetic and environmental components of facial soft tissue. However, it should be noted that results from twin studies are difficult to compare. Because of sample size, differences in zygosity maturation and statistical methods, inconsistency or similarity should be interpreted with caution. Only few studies used model fitting analysis for cephalometric twin studies, ¹⁹ , ²⁸ , ³⁴ and a direct comparison of the results is often challenging. Even though the included twins had no craniofacial anomalies and systemic diseases, other factors such as habits, allergies or airway disorders could have been present and influenced their craniofacial growth. Because there was no access to medical records so many years after the sample was gathered, the presence or absence of the mentioned conditions could not be investigated. Additionally, all included twin pairs were Caucasians, which might limit generalization of the findings to other populations.

Additionally, this was a purely clinical/radiographic study, and genome‐wide association studies are needed to identify the exact involvement of any genes in facial soft tissue growth. Until such studies are accessible, the present investigation shows that most of the components of facial soft tissue are under considerable genetic control. Therapeutic approaches can only influence the basic growth pattern within the individual biological limits, and the environmental contribution on soft tissue variability should not be ignored.

5. CONCLUSION

The present study on untreated monozygotic/dizygotic twins in adolescence and early adulthood indicates that:

1. At 12 years of age, strong genetic components (either additive or dominant) were identified for more than half of facial soft tissue variables. However, at 17 years of age all variables apart from one (lower lip thickness) were mainly under strong additive or dominant genetic control.

2. Two variables (facial convexity and nasal prominence) were mainly under additive genetic control at both time points.

3. Lower lip thickness was mainly under environmental control at both time points.

4. For nasolabial angle and lower lip profile, mainly environmental factors were found at 12 years of age but changed to mainly dominant genetic components at 17 years of age.

5. Apart from lower lip thickness, high heritability was found for most variables.

CONFLICT OF INTEREST

None. All authors declare that there are no conflicts of interest.

AUTHOR CONTRIBUTIONS

Hersberger‐Zurfluh MA collected the data, involved in interpretation of data and drafted the manuscript. Papageorgiou SN collected the data, analysed the data, involved in interpretation of data, critically revised the article and approved the submitted and final versions. Motro M designed the article, involved in interpretation of data, critically revised the article and approved the submitted and final versions. Kantarci A designed the article, involved in interpretation of data, critically revised the article and approved the submitted and final versions. Will LA designed the article, involved in interpretation of data, critically revised the article and approved the submitted and final versions. Eliades T designed the article, involved in interpretation of data, critically revised the article and approved the submitted and final versions.

1.

FIGURE A1.

Additional details of the statistical model. A pair of MZ twins is genetically identical (ie genetic correlation of 1). ACE and AE models assume that genetic effects are additive (ie there is no epistasis); therefore, by Mendelian inheritance rules the genetic correlation within a pair of DZ twins is 0.5. ADE models assume interactions between genetic influences inducing non‐additive – dominant – genetic effects, with a genetic correlation within a pair of DZ twins of 0.75 (0.5 additive genetic correlation and 0.25 dominant genetic correlation). Further assumptions include the lack of assortative mating of the twins parents and the equal environments’ assumption for MZ/DZ twins (which was supported by the Forsyth Moorrees Twin Study description), while the AE/ADE models assume no environmental effects shared by both twins of a twin pair (no C)

TABLE A1.

Model selection for cephalometric measurements at 12 years of age for the unadjusted model based on the Akaike information criterion

	Crude				Adjusting for sex
Variable	ACE	AE	ADE	Choice	ACE	AE	ADE	Choice
Facial convexity (°)	583.40	581.53	581.53	AE/ADE	583.28	581.28	581.28	AE/ADE
Nasolabial angle (°)	721.02	719.21	717.21	ADE	722.64	720.83	718.83	ADE
Upper lip thickness (mm)	146.59	146.59	144.03	ADE	146.97	144.97	142.43	ADE
Upper lip profile (mm)	147.96	145.98	143.89	ADE	139.26	137.33	135.33	ADE
Lower lip profile (mm)	178.12	179.81	177.81	ADE	175.18	177.54	175.54	ACE
Lower lip thickness (mm)	149.48	152.48	150.48	ACE	150.73	151.41	149.41	ADE
Nose prominence (mm)	184.50	182.65	182.65	AE/ADE	175.83	174.10	172.10	ADE

A, additive genetic variance; C, shared environment variance; D, dominant genetic variance; E, unique environment variance.

TABLE A2.

Model selection for cephalometric measurements at 17 years of age for the adjusted for sex model based on the Akaike information criterion

	Crude				Adjusting for sex
Variable	ACE	AE	ADE	Choice	ACE	AE	ADE	Choice
Facial convexity (°)	613.13	611.13	612.49	AE	607.21	605.21	606.89	AE
Nasolabial angle (°)	708.40	706.40	703.89	ADE	710.10	708.10	705.77	ADE
Upper lip thickness (mm)	170.89	168.89	164.50	ADE	135.15	133.33	135.37	AE
Upper lip profile (mm)	143.87	143.87	140.74	ADE	134.19	132.19	131.88	ADE
Lower lip profile (mm)	180.39	178.39	176.03	ADE	176.20	176.20	176.13	ADE
Lower lip thickness (mm)	190.26	188.26	184.10	ADE	168.55	166.55	163.43	ADE
Nose prominence (mm)	188.16	186.16	187.75	AE	189.34	187.34	189.15	ADE

A, additive genetic variance; C, shared environment variance; D, dominant genetic variance; E, unique environment variance.

Hersberger‐Zurfluh MA, Papageorgiou SN, Motro M, Kantarci A, Will LA, Eliades T. Heritability of facial soft tissue growth in mono‐ and dizygotic twins at 12 and 17 years of age: A retrospective cohort study. Orthod Craniofac Res. 2022;25:530–540. doi: 10.1111/ocr.12565

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.