Laura E. Mitchell
Twin studies of oral clefts (OCs) have in general been undertaken to address one of two questions: (1) Are OCs in twins related to or a consequence of the twinning process? and (2) What can twins tell us about the genetic contribution to OCs in the general population of, predominantly singleton (i.e., nontwin), affected individuals? Clearly, the answers to these two questions are not independent. If OCs in twins are a consequence of the twinning process, information from twins will not be useful in establishing the genetic contribution to OCs in the general population of affected individuals.
Prior to examining the literature, it is helpful to have an appreciation for some of the challenges presented by twin studies of OCs. Although OCs are relatively common malformations, the co-occurrence of twinning and OCs is rare. Population-based data from Denmark (Christensen and Fogh-Andersen, 1993a,b), which has a high ascertainment of OC cases (Christensen, 1999), suggests that only approximately 1/21,000 newborns will be a twin with an OC. Hence, it is extremely difficult to ascertain an unbiased sample of OC twins that is sufficiently large to address the questions of interest with any degree of precision. This problem is compounded by the fact that data from twins with cleft lip with or without cleft palate (CL/P) and those with cleft palate only (CP) must be evaluated separately since, with rare exceptions (e.g., Van der Woude's syndrome), the etiologies of these disorders differ (Fraser, 1970). In addition, twins with syndromic forms of CL/P and CP must be excluded from any analysis that focuses on the causes of nonsyndromic OCs, and monozygotic twins (i.e., identical twins derived from a single sperm and egg) and dizygotic twins (i.e., fraternal twins derived from the fertilization of two eggs) must be evaluated separately.
Accurate assessment of zygosity is an extremely important aspect of any twin study. Fortunately, recent advances in molecular genetics have made it possible to assign zygosity with a very high degree of accuracy in a relatively inexpensive and easy manner. DNA fingerprinting (Eufinger et al., 1993) or the evaluation of six to eight polymerase chain reaction (PCR) systems (Eufinger et al., 1995; Martin et. al., 1997) can be used to obtain an accurate zygosity diagnosis. The latter approach is generally preferred since it can be performed using DNA from buccal swabs, blood spots (e.g., Guthrie card), or venous blood, whereas DNA fingerprinting requires 0.5 to 1.0 ml of ethylenediaminetetraacetic acid (EDTA) blood (Eufinger et al., 1995).
Unfortunately, the majority of published twin studies of CL/P and CP predate the use of DNA-based assignment of zygosity. In fact, many of these studies did not assess zygosity directly but, rather, used Weinberg's difference method to estimate the number of mono- and dizygotic twin pairs. This method estimates the number of monozygotic pairs by subtracting twice the number of opposite-sex pairs (which must be dizygotic) from the total number of pairs. The assumptions underlying this method are that the sexes of dizygotic twins are independent and that same-sex dizygotic pairs are representative of opposite-sex dizygotic pairs. The validity of the first of these assumptions has, however, been challenged (James, 1979; Allen, 1981; Boklage, 1985; James, 1992), and the validity of the second is questionable when dealing with conditions, such as CL/P, that exhibit distorted sex ratios. [Although some studies have detected a predominance of females among individuals with CP, there is evidence that this may be an artifact of ascertainment, which typically includes only cases requiring surgical intervention (Christensen et al., 1992).]
Hence, estimates of twin concordance for CL/P and CP based on Weinberg's difference method may be biased and must be interpreted with caution.
Oral Clefts and the Twinning Process
Previous studies have suggested that twins are more likely than singletons to be affected with congenital malformations (Hay and Wehrung, 1970; Myrianthopoulos, 1976; Layde et al., 1980; Little and Nevin, 1989; Mastroiacovo et al., 1999), although such differences have not been identified in all studies (Windham and Bjerkedal, 1984; Ramos-Arroyo, 1991). When detected, the excess risk for congenital malformations in twins has generally been attributed to an increased risk only among monozygotic twins (Myrianthopoulos, 1976; Layde et al., 1980). It has been suggested that this may reflect either developmental disturbances associated with the monozygotic twinning process (Myrianthopoulos, 1978; Nance, 1981) or a common etiology for monozygotic twinning and specific malformations (Myrianthopoulos, 1978; Schinzel et al., 1979; Nance, 1981).
The conclusions that can be drawn from studies of the prevalence of congenital malformations in twins and singletons are limited by a number of factors. Several of these studies did not assess zygosity directly but, rather, compared the prevalence of malformations in same-sex twins and opposite-sex twins (Hay and Wehrung, 1970; Layde et al., 1980; Little and Nevin, 1989; Ramos-Arroyo, 1991). The rationale for this comparison is that all opposite-sex twins are dizygotic, whereas same-sex twins represent a mixture of monozygotic and dizygotic pairs. Hence, assuming that opposite-sex dizygotic twins are representative of same-sex dizygotic twins, any differences between same- and opposite-sex pairs must be attributed to the monozygotic twins. However, as previously stated, this assumption is questionable when dealing with conditions that exhibit distorted sex ratios. An additional limitation of these studies is that the risk to twins was assessed using information from both members of each twin pair. This will lead to an overestimate of the risk to twins relative to singletons since the members of a twin pair are not independent. Finally, most of these studies were based on relatively small samples and thus had limited statistical power to evaluate the relationship between twinning and specific malformations (Mastroiacovo et al., 1999).
A small number of studies have compared the prevalence of CL/P and CP in twin and singleton births and in same- and opposite-sex twin pairs. In general, these studies are limited by concerns regarding the completeness of case ascertainment and the potential for biased ascertainment of twin pairs (Christensen and Fogh-Andersen, 1993b). Based on these studies, there is some evidence that twins have a higher prevalence of CL/P, but not CP, relative to singletons (Table 17.1). However, in three of these studies (Layde et al., 1980; Little and Nevin, 1989; Ramos-Arroyo, 1991), the number of twins with CL/P was very small (n < 10), and the two largest studies (Hay and Wehrung, 1970; Mastroiacovo et al., 1999) provided conflicting results. Interpretation of the data from same- and opposite-sex twin pairs is also hampered by the small number of affected twins observed in the majority of studies. However, there appears to be no clear tendency for the prevalence of CL/P to be higher in same-sex than in opposite-sex twin pairs. Conclusions regarding differences in the prevalence of CP in same-sex and opposite-sex pairs are precluded by the small number of affected individuals in all but one study (Mastroiacovo et al., 1999).
Studies in Denmark have provided the most compelling evidence that the prevalence rates of CL/P and CP are not significantly different in twins and singletons (Christensen and Fogh-Andersen, 1993a,1993b). These studies had the advantage of near complete ascertainment since, in Denmark, individuals with CL/P or CP have been carefully registered for over five decades and treatment is both free and highly centralized (Christensen, 1999). In addition, zygosity was directly assessed (using 8-17 blood, serum, and enzyme types) in all twin pairs that included at least one affected individual, and the analyses were restricted to individuals who had nonsyndromic forms of CL/P and CP. Although a slight excess of twins with CL/P was detected in this population (33 expected vs. 47 observed), this difference was not statistically significant and there was no evidence that the risk of CL/P differed in monozygotic and dizygotic twins (Christensen and Fogh-Andersen, 1993a). Further, there was no significant difference in the observed and expected numbers of twins with CP (16 expected vs. 9 observed) and no evidence that the risk of this malformation differed by zygosity (Christensen and Fogh-Andersen, 1993b).
In summary, comparisons of the prevalence of CL/P and CP in singletons and twins, same-sex and opposite-sex twin pairs, and monozygotic and dizygotic pairs do not provide compelling evidence that these malformations are associated with twinning in general or with monozygotic twinning in particular. Additional evidence for or against such associations could be obtained by comparing other characteristics of CL/P and CP in twins and singletons. Specifically, demonstration of differences between affected singletons and twins with respect to well established characteristics of these conditions (e.g., sex ratio, pattern of familial recurrence, severity of defect) would provide some support for the hypothesis that these malformations are associated with the twinning process. Such differences have not been extensively investigated. However, the Danish data provide no evidence that the sex ratios among individuals with CL/P or CP differ in twins and singletons (Christensen and Fogh-Andersen, 1993a,1993b) or that either the distribution of CL/P types (i.e., unilateral vs. bilateral, cleft lip vs. and palate) (Christensen and Fogh-Andersen, 1993a) or familial recurrence patterns for CL/P (Mitchell and Christensen, 1997) differ in twins and singletons.
|
TABLE 17.1. Prevalence per 10,000 of Cleft Lip with or without Cleft Palate (CL/P) and Cleft Palate (CP) in Singleton and Twin Births |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
In conclusion, additional, larger studies of the familial, clinical, and epidemiological characteristics of CL/P and CP in twins are required to definitively rule out the existence of an association between twinning and these conditions. Such studies should include accurate assessment of zygosity and samples of sufficient size to evaluate the characteristics of these conditions separately in monozygotic and dizygotic twins. Given that CL/P (and possibly CP) exhibits a distorted sex ratio, samples adequate for the analysis of zygosity and sex-specific subgroups would be particularly desirable. However, since the available data provide little evidence that the causes of CL/P or CP differ in twins and singletons, it appears likely that cautious use of twin data may provide important insights regarding the genetic contribution to these conditions in the general population of affected individuals.
What Twins Tell us about the Genetic Contribution to Oral Clefts
In the classic twin design, concordance for a trait is measured in monozygotic twins (who are genetically identical) and dizygotic twins (who, on average, share only one-half of their genes) and the relative contribution of genetic and nongenetic factors to disease risk is estimated. A basic assumption of the classic twin approach is that the environment (both pre- and postnatal) of monozygotic and dizygotic twins is similar and, therefore, that any differences in their concordance rates must be attributable to differences in their degree of genetic similarity. Although this assumption has been challenged (Phillips, 1993), it appears to be valid in many circumstances (Duffy, 1993; Christensen et al., 1995). For additional information regarding the strengths and limitations of the classic twin approach, the reader is referred to the many reviews of this topic (Bundey, 1991; Bryan, 1992; LaBuda et al., 1992; Hall, 1996; Martin et al., 1997).
In many twin studies, only the subset of twin pairs that includes at least one affected member is sampled. Each independently ascertained twin is called a proband. Two methods are commonly employed to estimate twin concordance rates from such data. The pairwise concordance rate, which is the probability that both members of a twin pair are affected given that at least one is affected, and the probandwise concordance rate, which is the probability that a twin is a affected given that the cotwin is a proband (Allen et al., 1967). The probandwise concordance rate is generally preferred over the pairwise rate because it provides estimates of risk for individuals rather than pairs. In addition, estimates of probandwise concordance are directly comparable to estimates of risk for other types of relatives and can be used to infer the relative importance of genetic factors for the trait ofinterest (McGue, 1992). In contrast to pairwise concordance rates, probandwise concordance rates can also be compared across studies with different ascertainment schemes (McGue, 1992).
Comparison of probandwise concordance rates in mono- and dizygotic twins can provide a number of clues regarding the genetic contribution to a trait. If a trait is determined by a single gene, the probandwise concordance rate for monozygotic cotwins should be no greater than four-fold higher than the corresponding rate for dizygotic cotwins. For example, concordance rates for a fully penetrant, autosomal dominant condition would be 100% and 50% in monozygotic and dizygotic twins, respectively, giving rise to a twofold difference in risk. The corresponding values for a fully penetrant, autosomal recessive condition would be 100% and 25%, corresponding to a fourfold difference. Hence, relative risks that exceed 4 are incompatible with simple, single-gene inheritance and indicate that more than one gene and/or environmental factor must impact on the disease risk.
Comparison of concordance rates in dizygotic twins with risks to full sibs also provides information regarding the influence of nongenetic factors. Specifically, an increase in risk to dizygotic cotwins relative to full sibs (who are genetically equivalent to cotwins) suggests that shared environmental factors must contribute to disease risk. Finally, monozygotic twin concordance rates less than 100% indicate that genetic alterations occurring after cleavage of the embryo and/or other noninherited factors (e.g., intrauterine environmental differences in the allocation of cells and in the placental vascular supply to each twin, stochastic events) influence disease risk.
Several studies have reported twin concordance rates for CL/P and CP (Metrakos et al., 1958; Hay and Wehrung, 1970; Shields et al., 1979; Christensen and Fogh-Andersen, 1993a,b;
Nordstrom et al., 1996; Natsume et al., 2000). However, estimates from the majority of these studies are largely inadequate for several reasons: concordance was assessed in same- and opposite-sex pairs and used to estimate the expected concordance rates in mono- and dizygotic twins (Hay and Wehrung, 1970), the number of available twin pairs was small (Table 17.2), and pairwise rather than probandwise concordance rates were reported (Metrakos et al., 1958; Hay and Wehrung, 1970; Shields et al., 1979; Nordstrom et al., 1996).
Probandwise concordance rates for CL/P suggest that the risk to monozygotic cotwins of affected individuals is less than 100% but six- to 19-fold higher than the risk to dizygotic cotwins (Table 17.2). In addition, the observed risk to dizygotic cotwins does not appear to be markedly different from the 3% to 5% risk generally quoted for the full sibs of individuals with CL/P (Mitchell and Risch, 1992). These observations suggest that CL/P is influenced by genetic factors but that it is unlikely to be determined by a single gene. Moreover, non-genetic factors (or postcleavage genetic alternations) are also implicated in the development of CL/P since the probandwise concordance rate for monozygotic twins is less than 100%.
Estimates derived from the most recent Danish study (Christensen and Fogh-Andersen, 1993a) suggest that as much as 73% of the variation in liability to CL/P (i.e., the heritability of CL/P) may be determined by genetic effects. This estimate is likely the most accurate assessment of the genetic contribution to CL/P since it is based on twins for whom zygosity was established using extensive typing of blood, serum, and enzymes, as well as an accurate estimate of the population prevalence of CL/P in Denmark during the relevant time period (Christensen et al., 1992).
Twin studies suggest that CP is also influenced by both genetic and nongenetic factors and unlikely to be inherited in a simple Mendelian fashion (Table 17.2). However, the relatively small number of dizygotic twin pairs in the individual studies makes it difficult to determine if the risk to dizygotic cotwins is similar to or greater than the risk to full sibs, which has been estimated to be approximately 2% to 3% (Fitzpatrick and Farrall, 1993; Christensen and Mitchell, 1996). Twin-based estimates of heritability for CP are also difficult to derive for this reason.
Additional, larger studies of twins with CL/P and CP will be required to obtain more precise estimates of twin concordance and heritability for these traits. Although heritability estimates provide useful information regarding the relative genetic contribution to a trait, such estimates can and have been obtained from other types of relative and it is generally recognized that the risk for both CL/P and CP is determined by genetic factors. Moreover, as the quest to identify the specific genes involved in these conditions has been initiated (Wyszynski et al., 1996a; Schutte and Murray, 1999), the availability of precise, twin-based estimates of heritability for CL/P and CP is unlikely to significantly influence the current direction of research regarding the etiology of these conditions. Hence, the ability to obtain heritability estimates does not by itself provide strong justification for additional twin studies of CL/P and CP. Twin studies of CL/P and CP can, however, provide information that would be useful for genetic counseling purposes [e.g., what is the risk that the un-affected member of a monozygotic twin pair discordant for CL/P will have an affected child? (Wyszynski et al., 1996b)]. In addition, information on probandwise twin concordance rates can be used to establish the mode of inheritance, which is helpful in determining the most appropriate methods for the identification of CL/P and CP susceptibility loci.
Several approaches (e.g., linkage, association studies) can be used to identify disease-causing or diseasepredisposing genes (Lander and Schork, 1994). In general, linkage strategies are more appropriate for conditions determined by a small number of genes, each of which has a relatively major impact on risk. In contrast, association studies are better suited for conditions determined by a relatively large number of genes which have a modest impact on risk. Hence, it is helpful to have some understanding of the mode inheritance of a trait (i.e., the number of genes involved and the magnitude of their effect) prior to undertaking studies aimed at disease-gene identification.
Analysis of the familial recurrence patterns exhibited by a trait can provide insight regarding the number of genes involved in determining a trait and the magnitude of their effect. Specifically, the pattern of decline in λR - 1, where λR is the ratio of risk to a relative of type R (R = M, 1, 2, 3 for monozygotic twin, parent/offspring, second-degree, and third-degree relatives, respectively) compared to the population prevalence is determined by the underlying genetic model, and analysis of the observed decline in λR - 1 from monozygotic cotwins to first-degree relatives, as well as from first- to second-, and second- to third-degree relatives, can be used to assess the mode of inheritance (Risch, 1990). (Additional detail regarding this approach is provided in Chapter 19.) The lack of good estimates of λR, for monozygotic twins has, however, limited the usefulness of this approach for establishing the mode of inheritance of CL/P and CP (Mitchell and Risch, 1992). Hence, an important goal of future twin studies of CL/P and CP should be the precise estimation of monozygotic probandwise concordance rates.
|
TABLE 17.2. Monozygotic (MZ) and Dizygotic (DZ) Twin Concordance Rates for Cleft Lip with or without Cleft Palate (CL/P) and Cleft Palate (CP) |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Twin studies of CL/P and CP that incorporate the collection of DNA samples from both members of the twin pair and their parents may also help to determine the contribution of a specific genetic locus to disease risk. Information from monozygotic and dizygotic twins can be used to estimate the coefficient of genetic contribution, which measures the genetic contribution of a putative susceptibility locus to a multilocus disease (Rotter and Landaw, 1984). In addition, information from dizygotic twins can be used to estimate the relative increase in risk to sibs compared to the general population (As) that is attributable to a putative susceptibility locus (Risch, 1987), and concordant dizygotic pairs can be used in affected sib pair or affected relative pair linkage analyses. In conclusion, twin studies offer the opportunity to obtain additional insights regarding the genetic contribution to CL/P and CP. Such studies provide information that cannot be obtained by other methods and, hence, provide a useful complement to more traditional family-based and epidemiological investigations of these conditions.
Summary
Twin studies have played an important role in attempts to unravel the genetic contribution to many diseases (Martin et al., 1997) but have not been extensively utilized to examine the genetic contribution to common congenital anomalies. The paucity of twin research on congenital anomalies is partly attributable to the commonly held belief that the causes of these conditions differ in twins and singletons. However, this belief is not well substantiated for either CL/P or CP.
The rarity of twins with specific malformations is also a limiting factor in twin studies of congenital anomalies. Ascertainment of twins with a specific malformation, from a single source, is unlikely to provide sufficient numbers to address the questions of interest. Hence, twin studies of CL/P and CP are likely to require multicenter, multinational collaborations. Such studies are subject to concerns regarding heterogeneity resulting from differences in case definition and ascertainment, as well as the potential for etiological heterogeneity across centers. Careful consideration of study design and the development of common protocols and diagnostic criteria will be essential factors in future twin studies of CL/P and CP. Although metaanalysis of observational data remains controversial, a well-designed prospective meta-analysis of OCs in twins may provide advantages over both retrospective meta-analyses of data from individual studies and studies coordinated through a single center (Margitic et al., 1995). One multicenter, international study of twins with CL/P is currently nearing completion (L.E. Mitchell, unpublished data). The results of this study should provide additional clues regarding the etiology of this condition and important insights regarding the design and implementation of future multicenter twin studies of OCs.
Acknowledgements
This work was supported in part by a grant (DEI 1388) from the National Institutes of Health.
References
Allen, G (1981). Errors of Weinberg's difference method. Prog Clin Biol Res 69A: 71–74.
Allen, G, Harvarld, B, Sheilds, J (1967). Measures of twin concordance. Acta Genet 17: 475–481.
Boklage, CE (1985). Interaction between opposite-sex dizygotic fetuses and the assumptions of Weinberg's difference method in epidemiology. Am J Hum Genet 37: 591–605.
Bryan, EM (1992). The role in twins in epidemiologic research. Paediatr Perinat Epidemiol 6: 460–464.
Bundey, S (1991). Uses and limitations of twin studies. J Neurol 238: 360–364.
Christensen, K (1999). The 20th century Danish facial cleft population—epidemiological and genetic-epidemiological studies. Cleft Palate Craniofac J 36: 96–104.
Christensen, K, Fogh-Andersen, P (1993a). Cleft lip (+/- cleft palate) in Danish twins, 1970–1990. Am J Med Genet 47: 910–916.
Christensen, K, Fogh-Andersen, P (1993b). Isolated cleft palate in Danish multiple births. Cleft Palate Craniofac J 30: 469–474.
Christensen, K, Holm, NV, Olsen, J, et al. (1992). Selection bias in genetic-epidemiologic studies of cleft lip and palate. Am J Hum Genet 51: 654–659.
Christensen, K, Mitchell, LE (1996). Familial recurrence-pattern analysis of nonsyndromic isolated cleft palate—a Danish registry study. Am J Hum Genet 58: 182–190.
Christensen, K, Vaupel, JW, Holm, NV, Yashin, Al (1995). Mortality among twins after age 6: fetal origins hypothesis versus twin methods. BMJ 310: 432–436.
Duffy, DL (1993). Twin studies in the medical research. Lancet 341: 1418–1419.
Eufinger, H, Rand, S, Scholz, W, Machtens, E (1993). Clefts of the lip and palate in twins: use of DNA fingerprinting for zygosity determination. Cleft Palate Craniofac J 30: 564–568.
Eufinger, H, Schwartz, CE, Schutte, U (1995). Use of single- and multi-locus and polymerase chain reaction systems for zygosity determination—clinical applications in twins with clefts of the lip and palate. Acta Genet Med Gemellol (Roma) 44: 25–30.
Fitzpatrick, D, Farrall, M (1993). An estimation of the number of susceptibility loci for isolated cleft palate. J Craniofac Genet Dev 13: 230–235.
Fogh-Andersen, P (1942). Inheritance of Harelip and Cleft Palate. Copenhagen: Arnold Busck.
Fraser, FC (1970). The genetics of cleft lip and palate. Am J Hum Genet 22: 336–352.
Hall, JG (1996). Twinning: mechanisms and genetic implications. Curr Opin Genet Dev 6: 343–347.
Hay, S, Wehrung, DA (1970). Congenital malformations in twins. Am J Hum Genet 22: 662–678.
James, WH (1979). Is Weinberg's differential rule valid? Acta Genet Med Gemellol (Roma) 28: 69–71.
James, WH (1992). The current status of Weinberg's differential rule. Acta Genet Med Gemellol (Roma) 41: 33–42.
LaBuda, MC, Gottesman, II, Pauls, DL (1992). Usefulness of twin studies for exploring the etiology of childhood and adolescent psychiatric disorders. Am J Med Genet 48: 47–59.
Lander, ES, Schork, NJ (1994). Genetic dissection of complex traits. Science 265: 2037–2048.
Layde, PM, Erickson, JD, Falek, A, McCarthey, BJ (1980). Congenital malformations in twins. Am J Hum Genet 32: 69–78.
Little, J, Nevin, N (1989). Congenital anomalies in twins Northern Ireland. I. Anomalies in general and specific anomalies other than neural tube defects and defects of the cardiovascular sustem, 1974–1978. Acta Genet Med Gemellol (Roma) 38: 1–16.
Margitic, SE, Morgan, TM, Sager, MA, Furberg, CD (1995). Lessons learned from a prospective meta-analysis. J Am Geriatr Soc 43: 435–439.
Martin, N, Boomsma, D, Mackey, JF (1997). A twin-pronged attach on complex traits. Nat Genet 17: 387–392.
Mastroiacovo, P, Castilla, EE, Arpino, C, et al. (1999). Congenital malformations in twins: an international study. Am J Med Genet 83: 117–124.
McGue, M (1992). When assessing twin concordance, use the probandwise not the pairwise rate. Schizophr Bull 18: 171–176.
Metrakos, JD, Metrakos, K, Baxter, H (1958). Clefts of the lip and palate in twins. Plast Reconstr Surg 22: 109–122.
Mitchell, LE, Christensen, K (1997). Evaluation of family history data for Danish twins with non-syndromidic cleft lip with or without cleft palate. Am J Med Genet 72: 120–121.
Mitchell, LE, Risch, N (1992). Mode of inheritance nonsyndromic cleft lip with or without cleft palate: a reanalysis. Am J Hum Genet 51: 323–332.
Myrianthopoulos, NC (1976). Congenital malformations in twins. Acta Genet Med Gemellol (Roma) 25: 331–335.
Myrianthopoulos, NC (1978). Congenital malformations: the contribution of twins studies. Birth Defects 14: 151–165.
Nance, WE (1981). Malformations unique to the twinning process. Prog Clin Biol Res 69A: 123–133.
Natsume, N, Sato, F, Hara, K, et al. (2000). Description of Japanese twins with cleft lip, cleft palate or both. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 89: 6–8.
Nordstrom, REA, Laatikainen, T, Juvonen, TO, Ranta, RE (1996). Cleft-twin sets in Finland 1948–1987. Cleft Palate Craniofac J 33: 340–347.
Phillips, DIW (1993). Twin studies in medical research: can they tell us whether diseases are genetically determined? Lancet 341: 1008–1009.
Ramos-Arroyo, MA (1991). Birth defects in twins: study in a Spanish population. Acta Genet Med Gemellol (Roma) 40: 337–340.
Risch, N (1987). Assessing the role of HLA-linked and unlinked determinants of disease. Am J Hum Genet 40: 1–14.
Risch, N (1990). Linkage strategies for genetically complex traits. II. The power of affected relative pairs. Am J Hum Genet 46: 229–241.
Rotter, JI, Landaw, EM (1984). Measuring the genetic contribution of a single locus to a multilocus disease. Clin Genet 26: 529–542.
Schinzel, A, Smith, DW, Miller, JR (1979). Monozygotic twinning and structural defects. J Pediatr 95: 921–930.
Schutte, BC, Murray, JC (1999). The many faces and factors of orofacial clefts. Hum Mol Genet 8: 1853–1859.
Shields, ED, Bixler, D, Fogh-Andersen, P (1979). Facial clefts in Danish twins. Cleft Palate Craniofac J 16: 1–16.
Windham, GC, Bjerkedal, T (1984). Malformations in twins and their siblings, Norway, 1967–1979. Acta Genet Med Gemellol (Roma) 33: 87–95.
Wyszynski, DF, Beaty, TH, Maestri, NE (1996a). Genetics of non-syndromic oral clefts revisited. Cleft Palate Craniofac J 33: 406–417.
Wyszynski, DF, Lewanda, AF, Beaty, TH (1996b). Phenotypic discordance in a family with monozygotic twins and non-syndromic cleft lip and palate. Am J Med Genet 66: 468–470.