J Dent Res 82(7): 523-527, 2003
© 2003 International and American Associations for Dental Research
Clinical
Segregation Analysis of Mandibular Prognathism in Libya
A.A. El-Gheriani1,2, B.S. Maher2, A.S. El-Gheriani3, J.J. Sciote4, F.A. Abu-shahba5, R. Al-Azemi4,7, and M.L. Marazita1,2,6,*
1 Department of Human Genetics, Graduate School of Public Health, 500 Cellomics Bldg., University of Pittsburgh, 100 Technology Dr., Pittsburgh, PA 15219;
2 Center for Craniofacial and Dental Genetics, Division of Oral Biology, School of Dental Medicine, University of Pittsburgh, PA;
3 Department of Prosthodontics, Faculty of Dentistry, Benghazi, Libya (currently Department of Restorative Dentistry, Ajman University of Science and Technology, Ajman, UAE);
4 Department of Orthodontics, School of Dental Medicine, University of Pittsburgh, PA
7 (currently Department of Orthodontics, Ministry of Health, Kuwait);
5 Department of Orthodontics, Faculty of Dentistry, Benghazi, Libya; and
6 Department of Oral and Maxillofacial Surgery, School of Dental Medicine, University of Pittsburgh, PA;
*corresponding author,
/ ABSTRACTThe etiology of mandibular prognathism has been attributed tovarious genetic inheritance patterns and some environmentalfactors. The variation in inheritance patterns can be partlydue to the use of different statistical approaches in the respectivestudies. The objective of this study was to investigate therole of genetic influences in the etiology of this trait. Weperformed segregation analysis on 37 families of patients currentlybeing treated for mandibular prognathism. Mandibular prognathismwas treated as a qualitative trait, with cephalometric radiographs,dental models, and photographs used to verify diagnosis. Segregationanalysis of a prognathic mandible in the entire dataset supporteda transmissible Mendelian major effect, with a dominant modeof inheritance determined to be the most parsimonious.
KEY WORDS: mandible • mandibular prognathism • Class III malocclusion • genetics
/ INTRODUCTIONMandibular prognathism is a common finding, with prevalencevarying by race (Singh, 1999)—in particular, with higherprevalence in East Asians (Allwright and Bundred, 1964), Africans(Garner and Butt, 1985), and Caucasians (Emrich et al., 1965),respectively—and also varying by age, ranging from anapproximate prevalence of 0.5% in children 6-14 yrs old (Newman, 1956)to a range of 2-4% in adults (Jorgenson, 1990).
Despite many years of investigation, the relative contributionsof genetic and environmental components in the etiology of non-syndromicmandibular prognathism are unclear. Mossey (1999) states thatthis is due to a lack of research dedicated to this problem,relatively imprecise measuring tools and limited knowledge aboutthe genetic mechanisms involved, and the precise nature andeffects of environmental influences. It should also be addedthat little is known about the interaction between genetic andenvironmental factors in the causation of mandibular prognathism.
Mandibular prognathism cases have been associated with variousenvironmental etiologies, such as: enlarged tonsils (Angle, 1907);endocrine imbalances (Downs, 1928); posture, trauma,and disease (Gold, 1949); hormonal disturbance (Pascoe et al.,1960); congenital anatomic defects (Monteleone and Duvigneaud, 1963);and instrument deliveries (Schoenwetter, 1974). Familialaggregation of mandibular prognathism has also been describedand ascribed to a variety of genetic models, including autosomal-recessive(Downs, 1928; Iwagaki, 1938), autosomal-dominant (Stiles and Luke, 1953;Wolff et al., 1993), and a polygenic model of transmission(Litton et al., 1970). However, Kraus et al.(1959) maintainedthat "the role of heredity could not be discerned". Althoughthe genetic models differ, there is a consensus that there isa role for genetics in determining the occurrence of mandibularprognathism. Understanding the specific genetic factors contributingto variation in the risk for mandibular prognathism would bea major advance in dentofacial orthopedics and oral and maxillofacialsurgery.
The goal of the current study was to apply modern methods ofsegregation analysis to examine specific genetic models of thefamilial transmission of mandibular prognathism in a seriesof large Libyan families.
/ MATERIALS & METHODSSubjects
The authors identified 37 probands with mandibular prognathismfrom the patient base of several dental clinics in Benghazi,Libya. Complete family histories for each proband and the affectionstatus of other individuals in each family were confirmed bycephalometric, photographic, and/or dental models. The studysample of 37 families, comprised of 1013 individuals (285 first-degreerelatives, 658 extended relatives, and 33 individuals relatedto the proband through marriage), is summarized in Table 1.Consenting probands (or their parent/guardian) were asked todiscuss this study with other family members prior to participationin the study. The study protocol was approved by the Universityof Pittsburgh Institutional Review Board, and informed consentwas obtained from all subjects.
[in this window]
[in a new window]
/ Table 1. Numbers of Individuals by Phenotype and Sex in 37 Families Ascertained through Probands with Mandibular Prognathism
Procedure
Assessment
We determined mandibular prognathism by assessing one or moreof the following orthodontic records: a lateral cephalogram,orthodontic study models, and/or lateral profile facial photographs(Fig.). This assessment was done by Drs. A.S. El-Gheriani andF.A. Abu-shahba. The highest level of evidence was consideredto be the radiographic records; if radiographic records werenot available, then dental models were used. If no dental modelswere available, photographs were utilized. Cases for which affectionstatus consensus could not be reached were classified as unknown.
View larger version (114K):
[in this window]
[in a new window]
/ Figure. Shown in three different subjects are the diagnostic criteria for inclusion, either (a) a cephalogram, (b) a study model, or (c) a lateral photograph.
All 37 probands had a lateral cephalometric radiograph as partof their treatment record, and a confirmed negative ANB anglewas a prerequisite for enrollment in the study; however, wewere able to obtain lateral cephalometric radiographs on 11subjects for further evaluation at the University of PittsburghSchool of Dental Medicine. These cephalograms were measuredand compared with selected linear and angular values found inthe Bolton Growth Study (Broadbent et al., 1975) (Table 2).Since the data are stratified by age and sex, we chose pooledsex measurements at age 12 as the mean value for each measurementfor comparative purposes. Although inter- and intra-rater reliabilitymeasurements were not assessed, we used the 11 cephalographsas confirmation of the clinical diagnosis made on all subjects,since the cephalometric values confirmed a skeletal class IIIdiagnosis in all cases.
View this table:[in this window]
[in a new window]
/ Table 2. Cephalometric Analysis Results
This criterion establishes the forward relationship of the mandibleto the maxilla. Affection status of relatives (on whom radiographicrecords were not available) was determined by the presence ofan edge-to-edge incisor relationship or an anterior crossbitein dental casts articulated in centric jaw relation. The finaldiagnostic tool was the utilization of lateral photographs onindividuals for whom radiographic and dental cast records wereunavailable. The subject was determined to have a prognathicappearance (concave facial profile) when viewed in a lateralphotograph. Individuals for whom these records were not availablewere classified as unknown.
Statistical methods
We investigated the role of genetics and other influences inthe skeletal malocclusion family patterns by fitting class Aregressive models (Bonney, 1984) as implemented by the REGDroutine (for dichotomous traits) in S.A.G.E. release 3.1 [SAGE, 1994].These models assume that variation in a trait among individualsresults from a major gene effect and residual variation whichmay reflect both familial correlations and individual variation.The class A models assumed that the similarity between siblingsis due only to common parentage. We tested hypotheses by fittinga more complete model and comparing the resulting likelihoodwith those of reduced nested models.
The analysis model allowed for statistical detection of majortransmissible effects in a dichotomous trait, but did not allowfor explicit tests of multigenic/multifactorial contributionsto the trait. Such tests are not readily available for analysisof a dichotomous trait in extended kindreds. The hypothesisof most interest was whether there was a major effect that wouldthen allow the application of the powerful gene-mapping methodsto identify specific gene(s) involved in mandibular prognathism.Such gene-mapping methods will also allow for assessment ofmultigene contributions to the trait, and also gene x environmentalinteractions.
The overall model that was used allows for two alleles (A andB) at a single locus, resulting in three genotypes or "types"(AA, AB, BB). For our analyses, we assumed the distributionof the three types in the population to be in Hardy-Weinbergequilibrium. In the "no transmission" test, there is only onetype.
The parameters of the models include the regression coefficientsfor each "type" (ßAA, ßAB, ßBB),and the frequency of allele A (denoted qA). Individuals of typesAA, AB, and BB were assumed to transmit the A allele to theiroffspring with probabilities AA, AB, and BB. These transmissionprobabilities were used to calculate the probabilities of allthree types for individuals whose parents are in the pedigree.
The segregation analyses consisted of fitting a series of modelsranging from simple no-transmission models to the most generaltransmission model. The general model was fitted and comparedwith various more restricted submodels. To guard against thepresence of multiple local maxima, we used several initial estimatesof the parameters. Models that were tested for skeletal malocclusionwill include no-transmission models, A-dominant, A-recessive(B-dominant), additive and co-dominant major gene models, anda general model. The no-transmission model assumed one typewith all individuals being independent of one another. The A-dominantmodel restricted ßAA = ßAB, while a B-dominantmodel restricted ßAB = ßBB. The additivemodel assumed ßAB to be the average of ßAAand ßBB. The co-dominant major gene model put no restrictionson the types. Under these genetic models, the transmission parameterswere restricted to the Mendelian expected values of AA = 1,AB = 0.5 and BB = 0. For us to utilize the general model inthese analyses, there were no restrictions on the regressioncoefficients of the types or on tAB, allowing for tests of Mendeliantransmission.
Segregation analysis is sensitive to the method of identifyingfamilies, i.e., the ascertainment scheme used. In these studies,families were ascertained through probands. The analysis wastherefore adjusted for the ascertaining scheme by conditioningon the probands when the likelihood calculations were performed.
Hypotheses were tested according to the Likelihood Ratio Criterion,which is the ratio of the maximum value of the likelihood underthe most general model to the likelihood under a restrictedmodel. Each null hypothesis corresponds to one or more restrictionsbeing placed on the most general model to the likelihood undera restricted model. The models were evaluated with a test statisticdefined as minus the log of the likelihood ratio criterion.The distribution of this statistic was approximated by a 2 distributionwith the number of degrees of freedom equal to the differencein the numbers of parameters between the model under the nullhypothesis and the more general model. The null hypothesis wasrejected if this statistic is greater than the 2 value correspondingto the desired significance level ( = 0.05). Additionally, equallylikely models were assessed based on Akaike’s informationcriteria (AIC; Akaike, 1974). For any given model, the AIC =-2ln(L) + 2 number of parameters estimated. The model with thelowest AIC was considered to be the most parsimonious amongequally likely models.
/ RESULTSThe segregation analysis results are summarized in Table 3.Compared with the general model, the no-transmission model (p= 0.01) and co-dominant major gene model (p < 0.001) wereboth rejected. Autosomal-dominant, -recessive, and additivemajor locus models could not be rejected by the likelihood ratiotest (all p-values 0.14). Of the models that were not rejected,the autosomal-dominant major locus model was determined to bethe most parsimonious (AIC = 531.65).
View this table:[in this window]
[in a new window]
/ Table 3. Results of Segregation Analysis
/ DISCUSSION
Our results support the previous findings that there is a hereditarycomponent to the expression of this phenotype. We were ableto rule out the "no transmission" and co-dominant models. Byutilizing the AIC, we were able to conclude that, among theautosomal-dominant, -recessive, and additive models, the autosomal-dominantmodel was the most parsimonious. Our conclusion of autosomal-dominantinheritance was in agreement with Wolff et al.(1993), who utilizedpictures or authentic descriptions to determine affection status,but disagreed with the polygenic conclusion of Litton et al.(1970).Note that Litton et al.(1970) relied on articulated dental casts,while we used cephalograms as a means of determining affectionstatus.
Now that we have statistical genetic evidence for major locusinvolvement in mandibular prognathism, gene-mapping studieswill be feasible. Gene-mapping studies would have the potentialto identify the gene or genes involved in the trait, providingpowerful confirmation for a genetic basis for mandibular prognathism.Such studies will also allow for tests of multiple genes todetect any etiological heterogeneity in the trait.
Careful clinical characterization of mandibular prognathismwill be key to additional progress in our understanding of thegenetics of this common disorder. The Steiner analysis is readilyknown and easy to perform, and, as Saunders et al.(1980) reported,there is a positive correlation between the ANB angle amongfamily members. Note, however, that the important points arethe relative position of the mandible to the cranial base/maxilla,and also which aspects of the skeletal morphology seem to beassociated with the relative mandibular prognathism.
The drawback to basing this study on cephalometric records wasthe ethical implication of unnecessary radiation exposure. Therefore,the probands who were required to have radiographs taken aspart of their routine treatment were confirmed radiographically;however, the affection status of other relatives could not beverified unless they were also under orthodontic treatment.The use of dental models, as proposed by Ngan et al.(1997),to diagnose Class III malocclusion overcomes the radiation exposureissue; however, the contributions of each growth site and anatomicarea cannot be determined in the overall phenotypic expressionwhen diagnosis is based on dental models.
The cephalometric analysis (Table 2) highlights the difficultyin the study of this trait. The ages ranged from 10 to 38 yrs,and the ANB angles ranged from -0.9° to -9.1°. Eventhough all 11 subjects have a negative ANB, there were noticeabledifferences in their cephalometric measures that warrant thenotion that there may be multiple factors or multiple typesof mandibular prognathism (Jacobson et al., 1974). The Boltonstandards were compared with our measured values because thereare no definitive cephalometric norms for the Libyan population,and Libyan schoolchildren have been shown to have a cephalicindex similar to that of European schoolchildren (Gardiner, 1982).
Other than the fact that the phenotype is difficult to definecorrectly, craniofacial growth, and particularly the growthof the mandible, is highly variable and is reported to continueinto the late teens and well beyond the third decade of life(Behrents, 1985), although Tollaro et al.(1994) reported thata distinctive Class III pattern could be detected in childrenwith complete deciduous dentitions (age 4-6 yrs). It is notreasonable for clinicians to delay treatment for those individualsexhibiting mandibular prognathism in an attempt to see the finalphenotype, and conversely, could one compare the untreated casesof mandibular prognathism (individuals in their late 20s andbeyond) with those of pre-pubescent children who are still growing?Litton et al.(1970) used a cut-off point of 15 years of age,such that all unaffected children below the age of 15 were classifiedas unknown, since there is potential for them to develop thiscondition at a later time. We did not incorporate such a conditioninto our study and will be looking at the incidence of mandibularprognathism with increasing age for future studies.
The emphasis now should be placed on devising a safe and acceptablemethod of not only diagnosing mandibular prognathism, but alsoinvestigating the inheritance patterns of each skeletal morphologiccharacteristic that may contribute to it. Once a definitivemethod of phenotype classification is developed, this studywill be expanded to other population groups in the US, Europe,Africa, and the Middle East. Finally, given this statisticalevidence of a major transmissible effect in mandibular prognathism,we will begin linkage studies to identify the specific gene(s)involved in the trait.
/ ACKNOWLEDGMENTSThis study was supported by NIH grant T35-DE07336. The resultsof this paper were obtained by use of the program package S.A.G.E.,which is supported by a US Public Health Service Resource Grant(1 P41 RR03655) from the NationalCenter for Research Resources.We also thank Drs. M.K. Nair, F. Hamad, A. Sultan, R.E. Ferrell,M. Rahouma, H. Ben Khayal, Mr. F Reybod, and the study subjectsfor their contributions to this research.
Received September 5, 2002; Last revision January 13, 2003; Accepted April 15, 2003
/ REFERENCESAkaike H (1974). A new look at the statistical model identification. IEEE Trans Automatic Control AC 19:616–623.
Allwright WC, Bundred WH (1964). A survey of handicapping dentofacial anomalies among Chinese in Hong kong. Int Dent J 14:505–519.
Angle EH (1907). Treatment of malocclusion of the teeth. 7th ed. Philadelphia: S.S. White Manufacturing Company.
Behrents RG (1985). The biological basis for understanding craniofacial growth during adulthood. Prog Clin Biol Res 187:307–319.[Medline]
Bonney GE (1984). On the statistical determination of major gene mechanisms in continuous human traits: regressive models. Am J Med Genet 18:731–749.[ISI][Medline]
Broadbent BH Sr, Broadbent BH Jr, Golden WH (1975). Bolton standards of dentofacial growth. St. Louis: The C.V. Mosby Company.
Downs WG (1928). Studies in the causes of dental anomalies. J Dent Res 28:267–379,
Emrich RE, Brodie AG, Blayney JR (1965). Prevalence of class I, class II, and class III malocclusions (Angle) in an urban population; an epidemiological study. J Dent Res 44:947–953.[Abstract/FreeFullText]
Gardiner JH (1982). An orthodontic survey of Libyan schoolchildren. Br J Orthod 9:59–61.[Medline]
Garner LD, Butt MH (1985). Malocclusion in black American and Nyeri Kenyans. An epidemiologic study. Angle Orthod 55:139–146.[ISI][Medline]
Gold JK (1949). A new approach to the treatment of mandibular prognathism. Am J Orthod 35:893–912.
Iwagaki H (1938). Hereditary influence of malocclusion. Am J Orthod Oral Surg 24:328–336.
Jacobson A, Evans WG, Preston CB, Sadowsky PL (1974). Mandibular prognathism. Am J Orthod 66:140–171.[ISI][Medline]
Jorgenson RJ (1990). Mandibular prognathism. In: Birth defects encyclopedia. Buyse ML, editor. Cambridge: Blackwell Scientific Publications.
Kraus BS, Wise WJ, Frie RH (1959). Heredity and the craniofacial complex. Am J Orthod 45:172–217.
Litton SF, Ackermann LV, Isaacson RJ, Shapiro BL (1970). A genetic study of class III malocclusion. Am J Orthod 58:565–577.[ISI][Medline]
Monteleone L, Duvigneaud JD (1963). Prognathism. J Oral Surg 21:190–195.
Mossey PA (1999). The heritability of malocclusion: Part 2. The influence of genetics in malocclusion. Br J Orthod 26:195–203.[Abstract/FreeFullText]
Newman GV (1956). Prevalence of malocclusion in children six to fourteen years of age and treatment in preventable cases. J Am Dent Assoc 52:566–575.