Familial correlations and heritability of maxillary midline diastema☆1☆2☆3☆4☆5☆6☆7☆8

Material and methods 

Data collection 

Fifty-two sample subjects from the Department of Orthodontics and the School of Dentistry at Case Western Reserve University, Cleveland, Ohio, were selected as probands for this study. Selection was based on a clinically visible MMD of 0.5 mm or more with the maxillary canines erupted past the height of contour of the clinical crown. Subjects under the age of 12 years or missing maxillary anterior teeth were excluded. Probands were recruited by clinical examinations, examinations of plaster study models, and photographs of the face and teeth. In all, 30 families participated in the study—15 black, 14 white, and 1 mixed race.

Proband subjects were mailed several copies of a 7-question survey for their biological family members to complete and return. Self-addressed, stamped envelopes accompanied the questionnaires, as did letters of consent and instructions for completing the survey. Biological family members over the age of 12 were asked to complete the survey, answering biographical data including name, sex, ethnicity, and specific relationship to the proband. The questionnaires sought information about MMD, now or in the past, but after the age of 12 years. We attempted to obtain an estimate of diastema size on an ordinal, pictorial scale. The questionnaire also asked for information on restorative or orthodontic treatment to close the MMD and about missing permanent teeth. A brief medical history of each subject was obtained. Subjects were asked to complete the survey for biologic family members who were unavailable because geographic location or death. Participants were encouraged to contact other biologic family members. Completed, returned surveys were coded by assigning a family identification number, and each family member was assigned a personal identification number, with the proband number as the first digit. Data on some family members were obtained by telephone interviews or in person. Questions were identical to the information gathered via the questionnaire. Pedigrees for the 30 families were constructed on the relationships reported in the returned survey. Conventional signs were used, with males designated by squares and females by circles. Coloring in the square or circle indicated the affected status of each person in each family. Leaving the representative symbol unfilled designated unaffected subjects, and deceased persons were indicated by a slash mark. A representative pedigree is shown in the Figure.

View Large Image Download to PowerPoint

Fig. Family pedigree data; proband indicated by arrow.

Results 

Summary data are shown in Table I for both races, including the numbers of pedigrees, individuals, females, males, nuclear families, founders, and nonfounders.

Table I. Descriptive statistics for samples

		Black		White
	Category	Total	Total affected with MMD	Total	Total affected with MMD
	Pedigrees	15		14
	Individuals	212	115	150	85
	Females	103	68	80	51
	Males	107	47	70	34
	Nuclear families	49		32
	Founders	64		46
	Non-founders	148		104
	Minimum pedigree size	8		7
	Maximum pedigree size	26		16

Table II. Familial correlations and standard errors

	Group	Relationship	Number of pairs	Correlation	Standard error
	Black	Mother-father	30	−0.2765	0.1648
		Parent-offspring	240	0.0213	0.0781
		Sibling	303	0.1899	0.0985
	White	Mother-father	28	−0.4496	0.1404
		Parent-offspring	171	0.1577	0.0715
		Sibling	172	0.0874	0.0917

Discussion 

Heritability is defined as the ratio of total genotypic variance to the total phenotypic variance; it must be between 0 and 1.28 The lower limit occurs when all the observable phenotypic variance is attributable to nongenetic factors, and the upper limit results when all the observable phenotypic variation is attributable to underlying genetic makeup.28 Therefore, the magnitude of this ratio is a function of the genetic and nongenetic factors. In principle then, a phenotypic trait under Mendelian control could have a heritability of 0, if all people in the population examined had the same genotype and thereby no genetic variation. On the other hand, heritability could approach 1.0 if there were no differences in the environmental factors affecting the trait. Clearly, one must consider the impact of differences in these sources of variation that vary across the human population. Greater diversity in the environmental factors would lower heritability even though the underlying biologic mechanism is identical. Similarly, populations with relatively homogeneous genetic makeup would produce a lower value of heritability. Although no human population is truly genetically homogeneous, it is reasonable to assume that small closed groups, as in the current study, might be more genetically homogeneous than larger mixed populations. Perhaps this explains, in part, the findings of this study. Our results suggest that MMD is more heritable in the white population (0.32 ± 0.14) than in the black sample (0.04 ± 0.16). If the white sample had more uniform environmental influence or greater genetic variance, it would be reflected in a higher heritability for the trait. For example, if tooth size discrepancy was largely responsible for the reported MMD in the white sample, and tooth size is highly heritable, as stated by Chung and Niswander,29 this could account for the result. It is possible that the white population had fewer causes for the trait, with those causes having a higher heritability. The discrepancy in heritability between the black and white races might also be due to environmental factors. Greater environmental influences in the black group, such as periodontal drift, unreported habits, inaccurately reported missing teeth, or excessive incisor proclination, would cause heritability to be correspondingly lower. Also, with the small sample size, inaccuracies in reporting tend to be amplified in the intrafamily correlation calculations. Regardless of the lower calculated heritability in the black sample, the pedigree data strongly suggested a genetic influence in the expression of MMD. The results suggest that either dominance or the dental phenotype in the black population might, to a greater extent, be due to environmental factors.

However, heritability must be used with caution to determine the genetic basis for a phenotype. Unlike laboratory animal studies in which environmental factors can be standardized, human study populations are influenced by many environmental factors. The familial correlations from our study show some interesting patterns. The negative spousal correlations indicate negative assortative mating. Because the sibling correlations are higher than the parent-offspring correlations, it would appear that environmental factors play a significant role in the phenotypic variance of MMD. The parent-offspring and the sibling correlations might not be considered high in the present study. However, these low familial correlation values are accompanied by high standard errors as well, indicating that a larger sample size than we used is desirable.

Although the results indicate a possible genetic basis for MMD, certain shortcomings in our study must be considered before conclusions are drawn. The sample was chosen from an orthodontic graduate clinic. Unlike a random sample, in which one would expect every person in the population to have an equal probability to participate in the study, there was a bias in our sample selection. Nevertheless, because we corrected the likelihood for this method of ascertainment, our estimates are expected to be unbiased in large samples. Second, data collected via questionnaires are prone to errors—in particular, recall bias—as family members were requested to recall information about themselves and biological family members. However, self-reporting of the MMD trait was considered to be reliable because it is so easily identifiable. The self-reporting of MMD size on an ordinal scale proved to be unreliable. Many subjects could not understand the pictorial scale as drawn. The resulting data were therefore based on the dichotomous scale of affected versus unaffected. Clearly, a quantitative study measuring MMD space would have allowed for more powerful statistical analyses, but this was beyond the scope of the current project. Although the results warrant a more detailed search for a genetic mechanism, future efforts must overcome these limitations.

Conclusions 

Thirty families, singly ascertained through probands, consisting of 420 white and black people, were screened for MMD. The intrafamilial correlations were estimated and used to estimate heritability. Analysis of the data led to the following conclusions: