Oral Diseases (2003) 9, 1923 2003 Blackwell Munksgaard

All rights reserved 1354-523X/03

http://www.blackwellmunksgaard.com

Oral and Maxillofacial Pathology

Amelogenesis imperfecta: a classication and catalogue for the 21st century

MJ Aldred1,2,3, R Savarirayan2,3,4, PJM Crawford5
Department of Dentistry, Royal Childrens Hospital, Melbourne, Australia; 2Murdoch Childrens Research Institute, Melbourne, Australia; 3Department of Paediatrics, The University of Melbourne, Melbourne, Australia; 4Genetic Health Services Victoria and Southern Cross Bone Dysplasia Centre, Royal Childrens Hospital, Melbourne, Australia; 5University of Bristol Dental School, Bristol, UK
1

Amelogenesis imperfecta (AI) is a collective term for a number of conditions with abnormal enamel formation. Many cases are inherited, either as an X-linked, autosomal dominant or autosomal recessive trait. Several classications have evolved since 1945, based primarily on phenotype with the mode of inheritance being used in some systems as a secondary factor in allocating a case into a particular category. The benets and shortcomings of these systems are reviewed. As we move into an era of establishing the molecular basis of AI we propose a robust mechanism for classication and cataloguing of the disorder which parallels systems used in medical genetics. This system is applicable to individuals and families irrespective of current or future knowledge of the molecular defect involved. We argue that this system is of more benet to these individuals and families than previous classications. Oral Diseases (2003) 9, 1923 1 Keywords: amelogenesis imperfecta; classication; inherited; enamel; dental; genetic

Introduction
Class A number of individuals, persons or things possessing common attributes and grouped together under a general or class name; a kind, sort, division. (Shorter Oxford English Dictionary, Third Edition, 1965) Classify To arrange or distribute in classes according to a method or system. (ibid) classify verb [T] To divide (things) into groups according to type. (Cambridge International Dictionary of English; http://dictionary.cambridge.org/)

Correspondence: Michael J Aldred, Department of Dentistry, Royal Childrens Hospital, Flemington Road, Parkville, Victoria 3052, Australia. Tel: +613 9439 7415, Fax: +613 9431 6897, E-mail: michael_aldred@bigpond.com Received 21 February 2002; revised 18 October 2002; accepted 30 October 2002

These denitions carry with them the understanding that classication is the process of methodically distributing items according to type or common feature. Where items share several common features, the method of classication may be based on any one of those features best suited to the purpose of the classier. Thus, a gardener might wish to classify plants by season of owering, by ower colour or even by scent. A botanist wishing to consider the relationships between plants would be more interested in a classication that recognized biological relationships. A toxicologist might consider a classication based on the biochemical nature of the toxins produced by plants. Classication in medicine has for long been cliniciancentred and often based on appearance (phenotype); thus many clinicians will have learned classications for Swellings of the neck or White patches of the oral mucosa and so on. These classications are designed to aid diagnosis; however, they do little to further study of the particular condition, or to inform the patient. A second form of classication is increasingly seen in developmental conditions. For example, in OMIM

Amelogenesis imperfecta, classication MJ Aldred et al

20

entry #166210 OSTEOGENESIS IMPERFECTA CONGENITA; OIC the phenotype of OI type II can be caused by mutation in either the COL1A1 gene (120150) or the COL1A2 gene (120160) (Online Mendelian Inheritance in Man: http://www.ncbi.nlm.nih.gov/omim/). Such rened classication which allows qualication and a degree of ambiguity recognizes our increasing understanding of the molecular biology of both health and disease and forms a foundation both for study and for the clearer informing of patients. The basis for classication of that group of enamel defects referred to as amelogenesis imperfecta (AI) has not yet been denitively established. The rst contribution on the subject was a paper by Finn in 1938, who dierentiated two groups of tooth anomalies based on clinical characteristics, namely defects of dentine hereditary brown opalescent dentine and enamel brown hypoplasia of enamel. The rst denition of AI as a disease caused by a primary defect in enamel has been attributed to Weinmann et al (1945) who classied AI into two types, namely hypoplastic and hypocalcied. This denition was subsequently adopted by both Darling (1956) and Witkop (1957). Darling (1956) stated that AI implied a generalized fault of enamel structure aecting all the teeth of one or both dentitions, unrelated to any specic time or period of amelogenesis, or to any intermittent dietary abnormality or disease. He also remarked on the inherited pattern of the disease and added that there was a possibility that the condition could occur spontaneously in one or more members of the same family. Subsequently, Witkop and Rao (1971) dened AI as a group of disguring hereditary conditions which aect the clinical appearance of enamel of all or nearly all the teeth, which occur in kindreds such that all the individuals in the kindred show essentially the same defect and which are unassociated with known morphologic or biochemical changes elsewhere in the body. Later work describing variable expressivity in families with AI (Aldred and Crawford, 1988; Backman et al, 1993) suggests that the inclusion of the statement all the individuals in the kindred show essentially the same defect is perhaps inappropriate, a view reinforced by Crawford et al (1988) who described a family aected by autosomal dominant AI associated with taurodontism (ADAIT) in which both hypoplastic and hypomineralized forms of AI were reported. Additionally, in cases of X-linked inheritance (Crawford and Aldred, 1992; Aldred and Crawford, 1993; Backman et al, 1993) the occurrence of the Lyonization (X-inactivation) phenomenon gives rise to dierent appearances in males and females. By Witkop and Raos (1971) strict denition the term AI ought to be applied only to inherited defects unassociated with generalized ndings. Nevertheless, some authors have included enamel defects associated with inherited syndromes and metabolic diseases as examples of AI (Schulze, 1970; Pindborg, 1982). It is important for us to reconsider what we mean by any condition being exclusive of association with any other.

Table 1 Classication of amelogenesis imperfecta proposed by 3 Witkop and Rao (1971) (a) Hypoplastic 1. Autosomal dominant hypoplastichypomaturation with taurodontism (a) Winter type (b) Crawford type 2. Autosomal dominant smooth hypoplastic with eruption defect and resorption of teeth 3. Autosomal dominant rough hypoplastic 4. Autosomal dominant pitted hypoplastic 5. Autosomal dominant local hypoplastic 6. X-linked dominant rough hypoplastic (b) Hypocalcied 1. Autosomal dominant hypocalcied (c) Hypomaturation 1. X-linked recessive hypomaturation 2. Autosomal recessive pigmented hypomaturation 3. Snow-capped teeth, autosomal dominant 4. White hypomature spots?

In a view which argues one gene, one condition then the case for the exclusivity of AI is clear. However, as the complex interactions of both genomic position and biochemical activity are further revealed, hitherto unimaginable relationships may appear. In order to diagnose AI, we suggest a rening of the denition of Witkop and Rao (1971) to dene AI as: A group of conditions, genomic in origin, which aect the structure and clinical appearance of enamel of all or nearly all the teeth, and which may be associated with morphologic or biochemical changes elsewhere in the body. It follows that determining the probable mode of inheritance is integral and essential to the diagnosis of AI, as is the seeking of appropriate genetic advice and counselling. Once this has been done, then attention can be paid to the phenotype, its implications and management. This does not contradict the need for the phenotype to contribute to the establishment of the mode of inheritance. We do, however, believe that this approach is of benet to individuals and families with AI and to the eventual clarication and elaboration of the molecular basis of their disorder. In our attempts to systematize our understanding of AI, we are left with three factors. First, there is sucient variability in the phenotypic expression of AI within individual pedigrees to make appearance invalid as a criterion for sorting. Secondly, with the advances in our knowledge of the human genome it would be logical to use genotypic information to distinguish between forms of AI. Thirdly, patients want information about their condition that relates not only to themselves but also to their families. If he (sic) is able to tell his patients about their past and present symptomsalsowhat is going to happen, as well asthe details they have omitted, people will have no qualms in putting themselves under his care. (Hippocrates of Cos, c. 450 BC)

Oral Diseases

Amelogenesis imperfecta, classication MJ Aldred et al

Table 2 Classication of amelogenesis imperfecta proposed by Winter and Brook (1975) Hypoplasia I Autosomal dominant, thin and smooth with eruption defect and resorption II Autosomal dominant, thin and rough III Autosomal dominant, randomly pitted IV Autosomal dominant, localised V X-linked dominant, rough Hypocalcication I Autosomal dominant Hypomaturation I X-linked recessive II Autosomal recessive pigmented III Snow-capped teeth Hypomaturationhypoplasia with taurodontism I Autosomal dominant with occasional hypoplastic pits II Autosomal dominant with thin hypoplasia

Table 4 Classication of amelogenesis imperfecta proposed by Sundell and Koch (1985) 1. Hypoplastic 1.1 Rough 1.1.1 Basic form 1.1.2 Thin enamel 1.1.3 Pitted basic form 1.1.3.1 Pitted thin 1.1.3.2 Pitted with horizontal grooves 1.1.3.3 Pitted with vertical grooves 1.1.4 Horizontal grooves 1.1.5 Unspecied appearance 1.2 Smooth 1.2.1 Thin enamel 2. Hypomineralized 2.1 Hypomaturated 2.1.1 Localized opacities 2.1.2 Generalized opacities 2.2 Hypocalcied 2.2.1 Localized or generalized

21

Table 3 Classication of amelogenesis imperfecta proposed by Witkop and Sauk (1976) Hypoplasia Autosomal dominant pitted hypoplastic Autosomal dominant local hypoplastic Autosomal dominant smooth hypoplastic Autosomal dominant rough hypoplastic Autosomal recessive rough (enamel agenesis) X-linked dominant smooth Hypomaturation Autosomal dominant hypomaturationhypoplasia with taurodontism X-linked recessive hypomaturation Autosomal recessive pigmented Snow-capped teeth Hypocalcied Autosomal dominant hypocalcied Autosomal recessive hypocalcied

Table 5 Classication of amelogenesis imperfecta proposed by Sundell and Valentin (1986) Hypoplastic 1 Rough thin enamel 2 Rough pitted 3 Rough, horizontal grooves 4 Rough, unspecied 5 Smooth thin Hypomineralized 6 Hypomaturation, localized opacities 7 Hypomaturation, generalized opacities 8 Hypocalcication, localized or general

Dominant (autosomal or X-linked) Autosomal recessive Autosomal dominant with incomplete penetrance Dominant (X-linked or autosomal) Dominant (Autosomal or X-linked) Autosomal dominant Autosomal dominant (X-linked dominant or recessive) Autosomal recessive

The incidence of AI has been reported to vary from approximately 1 : 14 000 (Witkop, 1957) to <1 : 800 (Backman and Holmgren, 1988). The condition is not a single entity; there are a number of subtypes, characterized by their varying modes of inheritance and dierent appearances. Before exploring further the nature of any new classication, it is of merit to consider those that have been used before.

The classication of AI
Of the classication systems for AI often cited, the classication presented by Witkop and Rao in 1971 divided the conditions rst into three groups according to the predominant phenotype within a pedigree and then by the particular expression (Table 1). This in turn was rearranged by Winter and Brook in 1975 according to the presumed underlying developmental defect (Table 2). Witkop and Sauk (1976) proposed an elision of these two classications (Table 3). All three of these classications used both clinical features and inheritance patterns to allocate individual diagnoses. Note the confusing and inconsistent use of Arabic and Roman numerals (or lack thereof).

Sundell and Koch (1985) used a classication based solely on clinical ndings in a large study of Swedish children (Table 4). They argued that previous classications were of questionable value as they did not take into account the relative frequencies of the dierent types of AI. This classication was insucient on its own to distinguish between all the cases seen. The incidence of a condition is not particularly relevant to its position in a classication, although parallels might be drawn with ornithology where identication is sometimes aided by reference to a migrants or accidentals list. However, in the present case, the clinical appearance alone would not seem to justify allocation to diagnostic groups where there is known to be considerable variation in appearance, especially within families, of what is an inherited condition. Sundell and Valentin (1986) presented a modied classication of AI (Table 5) formed by the addition of family (genetic) pedigrees into the clinical appearance classication of Sundell and Koch (1985). In their paper, they seem to have confused incidence with existence of a condition. The resulting, and rather confusing, list is very dicult to compare with those previously cited and
Oral Diseases

Amelogenesis imperfecta, classication MJ Aldred et al

22

Table 6 Classication of amelogenesis imperfecta proposed by Witkop (1988) Type I Hypoplastic IA hypoplastic, pitted autosomal dominant lB hypoplastic, local autosomal dominant IC hypoplastic, local autosomal recessive ID hypoplastic, smooth autosomal dominant IE hypoplastic, smooth X-linked dominant IF hypoplastic, rough autosomal dominant IG enamel agenesis, autosomal recessive Type II Hypomaturation IIA hypomaturation, pigmented autosomal recessive IIB hypomaturation, IIC snow-capped teeth, X-linked lID autosomal dominant? Type III Hypocalcied IIA autosomal dominant IIB autosomal recessive Type IV -Hypomaturationhypoplastic with taurodontism IVA Hypomaturationhypoplastic with taurodontism, autosomal dominant IVB Hypoplastichypomaturation with taurodontism, autosomal dominant

Table 7 Classication of amelogenesis imperfecta proposed by Aldred and Crawford (1995) Genetic location Mutation Biochemical outcome Mode of inheritance Phenotype

leaves out at least one well-documented condition, namely ADAIT (Winter et al, 1969; Crawford, 1970; Congleton and Burkes, 1979; Crawford et al, 1988). At one level, this is not surprising as ADAIT has been described in only a few families to date. However, in general terms, the use of a classication based on incidence and generated from one geographical area is likely to restrict the conditions studied to a single genetic pool. Again, the incidence of a condition hardly seems relevant to its position in a classication, neither does the clinical appearance alone seem to justify allocation to diagnostic groups even within families, of what is an inherited condition. Witkop (1988) further developed the predominantly phenotypic classication of Witkop (1957), Witkop and Rao (1971) and Witkop and Sauk (1976) into a further classication (Table 6). Four major types were recognized based on phenotype (hypoplastic, hypomaturation, hypocalcied and hypomaturationhypoplastic) and then subdivided into 15 subtypes based primarily on phenotype and, secondarily, by mode of inheritance. There is clear evidence, from a large number of case studies, of the genetic aetiology of at least 10 forms of AI; further diagnostic confusion arises from phenotypic variability. It would seem that the only denitive and reproducible diagnosis would arise from a study of the genetic composition of aected individuals and families. As already cited, classication relies upon a method or system. Previous eorts to base classes of AI on
Oral Diseases

appearance have been overtaken by two major events; rst, descriptions of phenotypic variation within single pedigrees and, secondly, our increasing knowledge of genomic genetics. In an attempt to use these ndings to bring order to a confused eld, Aldred and Crawford (1995) proposed a new classication based on the molecular defect (when known), biochemical result (when known), mode of inheritance and then phenotype (Table 7). This proposal has not yet been adopted, although some papers in other areas have suggested that this might be useful in other dental genetic disorders (Dean et al, 1997). The lack of present uptake may reect the as yet limited data available for classication based on the molecular defect in an individual or family. In view of this, we present an argument for a modied version of this suggestion which is presently workable, in the best interests of individuals and families with AI and capable of later extension. In other aspects of genetics, a reliance on phenotypic characteristics has proven to be unreliable in the light of molecular genetic investigations. An example of this is in the classication of the various craniosynostosis (premature fusion of one or more of the calvarial sutures) syndromes. There are over 100 of these syndromes that were classied primarily on the basis of their clinical phenotypes (Cohen, 1986). The recent explosion of knowledge regarding the genetic basis of these conditions has rendered this classication less useful as many disorders have been discovered to be allelic conditions. Moreover, many individuals within a particular family who were diagnosed to have dierent clinical conditions (based solely on phenotype) have been shown to have identical gene mutations (Jabs, 1998). It is therefore logical and prudent to consider both the modes of inheritance and molecular genetic data in conjunction with the specic phenotype in any classication where there exists an overlap of clinical features. These systems have proved to be more user-friendly as well as lending themselves to being a better tool for understanding how the genotype leads to the specic phenotype (i.e. socalled functional genomics).

A workable classication for AI

In our previous proposal for classication of AI (Aldred and Crawford, 1995) we had anticipated that the advances in medical molecular genetics would rapidly have been mirrored in dental genetics. Regrettably, our understanding of the molecular genetic basis of dental abnormalities has not progressed as quickly as might have been expected. As a result, we propose a format for the classication of AI (which might also be applied to other inherited dental anomalies) until such time as the molecular basis of these disorders is claried. This proposal might also become the basis of a future catalogue of inherited dental conditions as has been developed in other inherited disorders. In saying this we seek to distinguish between a method of classication and the formation of a catalogue, although it may be that the former eventually contributes to the latter.

Amelogenesis imperfecta, classication MJ Aldred et al

Table 8 New scheme proposed for the classication and cataloguing of amelogenesis imperfecta Mode of inheritance (autosomal dominant autosomal recessive X-linked isolated case) Molecular basis [chromosomal localization locus mutation (when known)] Biochemical outcome (putative result of mutation when known) Phenotype (description of clinical and or radiographic features and or other relevant ndings)

Whereas all previous authors have used a general division by phenotype as a primary discriminator, we propose that, until such time as the molecular basis of all forms of AI is established, the primary structure for the classication of AI be based on the mode of inheritance, with the clinical and radiographic appearances (and any other features) being the secondary discriminators (Table 8). As information on the molecular basis of the disorder in particular individuals or families becomes known, this can be added to the description of the cases. Such a concept is not particularly novel, apart from in the dental literature. A revision of inherited dentine defects along these lines has been suggested (Dean et al, 1997) to avoid overdue emphasis on phenotypic characteristics. Furthermore, this system forms the basis of Online Mendelian Inheritance in Man. References
Aldred MJ, Crawford PJM (1988). Variable expression in amelogenesis imperfecta with taurodontism. J Oral Pathol 2 Med 17: 327323. Aldred MJ, Crawford PJM (1993). Clinical features of a family with X-linked amelogenesis imperfecta mapping to a new locus (AIH3) on the long arm of the X chromosome. Oral Surg Oral Med Oral Pathol 76: 187191. Aldred MJ, Crawford PJM (1995). Amelogenesis imperfecta towards a new classication. Oral Dis 1: 25. Backman B, Holmgren G (1988). Amelogenesis imperfecta: a genetic study. Hum Hered 38: 189206. Backman B, Lundgren T, Engstrom EU et al (1993). The absence of correlations between a clinical classication and ultrastructural ndings in amelogenesis imperfecta. Acta Odontol Scand 51: 7989. Cohen MM (1986). Craniosynostosis: diagnosis, evaluation and management. Raven Press: New York. Congleton J, Burkes EJ (1979). Amelogenesis imperfecta with taurodontism. Oral Surg Oral Med Oral Path 48: 540544.

Crawford JL (1970). Concomitant taurodontism and amelogenesis imperfecta in the American Caucasian. J Dent Child 37: 8387. Crawford PJM, Aldred MJ (1992). X-linked amelogenesis imperfecta. Presentation of two kindreds and a review of the literature. Oral Surg Oral Med Oral Pathol 73: 449455. Crawford PJM, Evans RD, Aldred MJ (1988). Amelogenesis imperfecta: autosomal dominant hypomaturation-hypoplasia type with taurodontism. Br Dent J 164: 7173. Darling AI (1956). Some observations on amelogenesis imperfecta and calcication of the dental enamel. Proc R Soc Med 49: 759765. Dean JA, Hartseld JK, Wright JT, Hart TC (1997). Dentin dysplasia type II linkage to chromosome 4q. J Craniofac Genet Dev Biol 17: 172177. Finn SB (1938). Hereditary opalescent dentin. 1. An analysis of the literature on hereditary anomalies of tooth colour. J Am Dent Assoc Dental Cosmos 25: 12401249. Jabs EW (1998). Toward understanding the pathogenesis of craniosynostosis through clinical and molecular correlates. Clin Genet 53: 7986. Pindborg JJ (1982). Aetiology of developmental enamel defects not related to uorosis. Int Dent J 32: 123134. Schulze C (1970). Developmental abnormalities of the teeth and jaws. In: Gorlin RJ, Goldman HM, eds. Thomas oral pathology, 6th edn. St Louis: Mosby, pp. 96178. Sundell S, Koch G (1985). Hereditary amelogenesis imperfecta. I. Epidemiology and clinical classication in a Swedish child population. Swed Dent J 9: 157169. Sundell S, Valentin J (1986). Hereditary aspects and classication of hereditary amelogenesis imperfecta. Community Dent Oral Epidemiol 14: 211216. Weinmann JP, Svoboda JF, Woods RW (1945). Hereditary disturbances of enamel formation and calcication. J Am Dent Assoc 32: 397418. Winter GB, Brook AH (1975). Enamel hypoplasia and anomalies of the enamel. Dent Clin North Am 19: 324. Winter GB, Lee KW, Johnson NW (1969). Hereditary amelogenesis imperfecta. A rare autosomal dominant type. Br Dent J 127: 157164. Witkop CJ (1957). Hereditary defects in enamel and dentin. Acta Genet Stat Med 7: 236239. Witkop CJ (1988). Amelogenesis imperfecta, dentinogenesis imperfecta and dentin dysplasia revisited: problems in classication. J Oral Pathol 17: 547553. Witkop CJ, Rao SR (1971). Inherited defects in tooth structure. Birth Defects Orig Artic Ser 7: 153184. Witkop CJ, Sauk JJ (1976). Heritable defects of enamel. In: Stewart RE, Prescott GH, eds. Oral facial genetics, Chapter 7, Mosby: St Louis.

23

Oral Diseases