Digestive and Liver Disease 37 (2005) 659–664

Alimentary Tract

Large forehead: A novel sign of undiagnosed coeliac disease

M. Finizioa, G. Quarembab, G. Mazzaccac, C. Ciaccic,∗a Dental School, Chair of Dental and Biomedical Informatics, Second University of Napoli, Naples, Italy

b Department of Public Medicine, Federico II University of Napoli, Naples, Italyc Gastrointestinal Unit, Federico II University of Napoli, via Sergio Pansini 5, 80131 Naples, Italy

Received 24 January 2005; accepted 13 April 2005Available online 24 May 2005

See related commentary, on pages 646–647

Abstract

Background. The development of cranial proportions is the result of genetic, embriogenetic and environmental factors. Coeliac diseaseis a genetically inherited disease that is frequently diagnosed in adulthood in individuals, in whom the disease runs unidentified for years andcan affect child growth from the moment of dietary gluten introduction up to the moment of gluten withdrawal following diagnosis. Data ont

liac diseasea

hy controlsw re.

of the facec

his is thefi ed amongt ia and shorts ental enamelh red for theg patient forc©

K

1

prow

e andpor-

lt of

wth

iveds theus in

1d

he effects of gluten on craniofacial development in coeliac children are not available.Aim. The aim of the present study is to evaluate gluten-related effects on face development in patients with undiagnosed coe

nd their clinical relevance.Methods. The study was a prospective, multivariate analysis. Face photographs of adult patients with coeliac disease and healt

ere marked at six reference points and distances and the ratios among distances were measured by computer program softwaResults. The main finding of the study is that Caucasian Mediterranean adult coeliac individuals tend to have a peculiar aspect

haracterised by a larger forehead when compared to general population controls.Conclusion. The craniofacial morphology of patients with coeliac disease reveals an altered pattern of craniofacial growth. T

rst report of alterations of craniofacial development in coeliac disease. This alteration is a clinical sign that should be includhe extraintestinal manifestation of coeliac disease. It has a frequency comparable to other signs or symptoms such as anaemtature and is a better predictor of coeliac disease than other signs such as recurrent aphthous stomatitis, recurrent abortion and dypoplasia. A large forehead is a sign easily evident visually or with very simple measurements; computer analysis is not requieneral practitioner. This sign along with the presence of other clinical signs and symptoms, should alert a physician to test aoeliac disease.2005 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.

eywords: Anthropometry; Coeliac disease; Coeliac face; Craniofacial development; Forehead

. Introduction

A peculiar aspect of the face of a large number of coeliacatients diagnosed in adulthood was noted during a differentesearch study aimed at evaluating dental enamel defects. Inrder to test this observation, a study on facial anthropometryas planned.

∗ Corresponding author. Tel.: +39 081 7464270; fax: +39 081 7464270.E-mail address: ciacci@unina.it (C. Ciacci).

Craniofacial anthropometry measures the human facskull. Measurements of the craniofacial complex are imtant for studies on human growth[1], population variation[2]and medical problems[3].

The development of cranial proportions is the resugenetic[4], embriogenetic[5] and environmental factors[6].

Craniofacial structures have distinct origins and grorates[7,8].

The neurocranium (i.e. that part of the skull that is derfrom the branchial (or pharyngeal) arches and comprisebones of the face) develops almost completely in the foet

590-8658/$30 © 2005 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.oi:10.1016/j.dld.2005.04.014

660 M. Finizio et al. / Digestive and Liver Disease 37 (2005) 659–664

the womb[9], while the viscerocranium (i.e. that part of theskull that is derived from the branchial arches and comprisesthe bones of the face) undergoes its greater development dur-ing weaning, childhood and dentition[10].

Coeliac disease is a genetically inherited disease[11] thatis frequently diagnosed in adult life[12], in individuals inwhom the disease runs unidentified for many years[13].

It is well recognised that undiagnosed coeliac disease mayinterfere with body development during the early phase of hu-man life, particularly in terms of height and weight[14,15].We hypothesised that these patients may also differ from thegeneral population for their craniofacial development and thata discrimination analysis would enable identification of dis-tinct morphometric features using craniofacial antropometry.

The present study is aimed at investigating the effects ofundiagnosed coeliac disease on facial development and itspossible clinical relevance.

2. Methods

2.1. Patients

The study was a prospective, multivariate analysis.The study was conducted in an outpatient clinic. Cau-

casian Mediterranean adults from Campania, southern Italy( e re-c ongt o ob-t Thec ongm om-m thes avet 0%).

cialt e unitf

anda edicaa o ac-c ed-i eliacd to ex-c dies.

nd-i ofp adult-h liacd

2

entm ents

were carried out both by callipers and on photographs withspecific software for image analysis in 20 subjects, randomlyselected from the entire sample.

The photographs were taken en face on and in pro-file (right). All underwent teleradiography of the cranium(antero-posterior and latero-lateral). After appropriate com-parisons between the two measurements we decided to con-tinue the study for the rest of the subjects using only pho-tographs in order to reduce time and effort for subjects. It alsoallowed the analyses to be carried out at distance by other peo-ple unaware of diagnosis of the subjects. Photographs weretaken from a distance of 150 cm with a digital camera (3.2megapixel) positioning the subjects’ head in the ‘natural headorientation’ (NHO)[16,17]. The same person took all thephotographs. By applying appropriate software, photographswere rotated to make the interpupillary distance parallel tothe lower edge of the picture. In all cases, the rotation wasless than 3◦.

Six reference points were marked on the photographs ontwo different frames of the face for each of the subjects andthe ratios of the distances were calculated.

Although the adoption of ratios removed the inconsisten-cies in magnification, camera position and subject placement(common to normal hand-held photographs), in order to fur-ther minimise these problems, a specialised photographic rigwas used and specific protocols established in order to pro-v

lesst weree ly enf ssiblem num-b lessi ther sures.

re:

•••• the

rtion

• er

en-t andg hoto-g

naly-s

• m-

• uro-ace

65 males, 155 females), all from Italian parentage werruited. Coeliac patients were consecutively recruited amhose who underwent the annual visit necessary in Italy tain gluten-free food under the National Health System.ontrol group was made up of volunteers recruited amedical and non-medical hospital staff. The Ethical Cittee of the University Federico II of Napoli approved

tudy protocol. All of those who were eligible and asked gheir informed consent and participated in the study (10

Inclusion criteria were: age greater than 18 years.Exclusion criteria were: face asymmetry or previous fa

rauma, plastic surgery and patients persons seen in thor any other gastrointestinal problem.

The control group (group 0) was made up of 100 sex-ge-matched volunteers recruited among students, mnd non-medical staff and friends of coeliac patients whepted to participate in this study. Controls underwent m

cal evaluation to exclude other gastrointestinal and coisease-related symptoms and all underwent blood testslude anaemia and anti-tissue-transglutaminase antibo

Adult coeliac patients were divided in two groups depeng on age of diagnosis. Group 1 (N = 97) was composedatients with coeliac disease diagnosed and treated inood, while group 2 was made up of patients with coeisease diagnosed and treated during childhood (N = 23).

.2. Measurements

In order to check the reliability of the measuremethodology, a pilot study was undertaken. Measurem

l

ide consistent and reproducible photographs[18].The points of measurements in this condition, even if

han those available with calliper and Rx measurements,asily identifiable and associated with those taken direct

ace. For the purpose of data analysis, among all the poeasurements, we chose those that allowed the minorer of errors both inter- and intra-examiners and those

nfluenced by subjects’ mimicry. Moreover, we comparedatios between measurements and not the absolute mea

The reference points used for the measurements we

pupillar: the lower point of left and right pupilstrichion: the midpoint of the hairlinenasion: the midpoint of the nasofrontal suturesubnasal: in the midline, the point of transition fromlower part of anterior nasal spine and the cutaneous poof the upper lipgnathion: in the midline, the lowest point on the lowborder of the chin

In the case of a subject with a receding hairline, we idified the original line passage from the frontal musclealea capitis and it was marked before shooting the praph.

Segments were traced between points and used for ais. The measurement set included:

interpupillar (ip): distance from the lower point of circuference of pupilstrichion-nasion (tn): the forehead length, marker of necranial growth which makes the superior third of the f

M. Finizio et al. / Digestive and Liver Disease 37 (2005) 659–664 661

Fig. 1. The face of a young coeliac woman upon which the reference pointsand segments used in the analysis have been traced.

• nasion-subnasal (ns): the nose length that is the mediumthird of the face

• subnasal-gnathion (sg): the length of the lower third of theface

• trichion-gnathion (tg): total length of the face

Points and segments are shown inFig. 1Six examiners, unaware of the subjects’ diagnosis, exam-

ined the photos of 25 randomly selected subjects by a com-puterised image analyser.

Mean, standard deviations (S.D.), standard error ofmean (S.E.M.) and range (min− max) were calculatedfor all the craniofacial anthropometric measurements. TheKolmogorov–Smirnov test and multivariate analyses weresubsequently performed by means of parametric statisticaltests.

Table 1Box’s Test of Equality of Covariance Matrices that shows that the pop-ulations under study have a multivariate normal distribution and equalvariance–covariance matrices

Box’s M 2.709

F approximately .893df1 3df2 7053832.352Sig. .444

Table 2Stepwise statistics

Step Tolerance F to remove Wilks’ lambda

1 tn/ns 1.000 12.978

2 tn/ns .928 16.681 .991ip/tg .928 5.346 .938

The Wilks’ lambda method was used for choosing the best discriminant vari-ables (shown in the table) among all the craniofacial measurements ratios.Trichion-nasion (tn): the forehead length, marker of neurocranial growth,which makes the superior third of the face; nasion-subnasal (ns): the noselength that is the medium third of the face; interpupillar (ip): distance from thelower point of circumference of pupils; trichion-gnathion (tg): total lengthof the face.

Table 1 shows Box’s Test of Equality of CovarianceMatrices that demonstrates that the populations understudy have a multivariate normal distribution and equalvariance–covariance matrices.

We used a discriminant analysis with stepwise eliminationof variables to build a predictive model of group membershipbased on observed characteristics of each case. The Wilks’lambda method was used for selecting the test set to assessthe success of the discriminant function and also for choosingthe discriminant variables. AsTable 2shows, the variables

Fig. 2. Highlights the findings of the present study. Half photograph of theface of the coeliac woman (panel A) and half photograph of the face ofan age-matched control woman (panel B) were joined together along themidline. The horizontal blue lines are traced to match the reference points.Photographs have been resized in proportion in order to equalise the tgdistance.

662 M. Finizio et al. / Digestive and Liver Disease 37 (2005) 659–664

Table 3Gender, age, anthropometrics and selected craniofacial measurements ratios of the study cohort

Control group(n 100)

Coeliac patients withdiagnosis inadulthood (n 97)

Coeliac patients withdiagnosis inchildhood (n 23)

ANOVA Test for linearity

F p F p

Gender (M/F) 32/68 24/73 9/14 – – – –Age (years) 31.7± 13.7 33.2± 10.7 23.4± 4.3 6.54 0.002 0.771 0.381BMI (kg/height2) 23.6± 4.4 20.2± 3.1 21.6± 3 19.5 0.0001 0.156 0.693tn/ns 125 ± 14.4 132 ± 12.7 128.6 ± 13.1 6.538 0.002 13.066 0.0001ip/tg 34.2± 1.82 34.6± 2.0 34.8± 2.4 1.205 0.302 1.702 0.193ns/tg 29.1± 2.05 28.4± 1.74 29.0± 2.4 3.072 0.048 0.577 0.449sg/tg 34.8± 2.8 34.3± 2.91 33.9± 3.1 1.428 0.242 0.996 0.319tn/ip 105.6± 10.5 108.2± 11.1 106.9± 11.9 1.354 0.260 2.707 0.101

The most significant finding in this study is in bold type. Interpupillar (ip): distance from the lower point of circumference of pupils; trichion-nasion (tn): theforehead length, marker of neurocranial growth which makes the superior third of the face; nasion-subnasal (ns): the nose length that is the medium third ofthe face; subnasal-gnathion (sg): the length of the lower third of the face; trichion-gnathion (tg): total length of the face.

admissible for successful discriminant analysis are the ratiobetween tn/ns and ip/tg.

All the analyses were carried out by means of statisticalsoftware (SPSS 12.1, Chicago, IL, USA). Statistical signifi-cance was accepted atp <0.05.

3. Results

Fig. 2 shows the photographs of a coeliac and a controlwoman, which have been cut into half and placed together toindicate both the measurement points and to visually repre-sent the main finding of the study.

Descriptive statistics for the craniofacial anthropometricmeasurements are shown inTable 3.

Separate analyses by gender showed no differences be-tween women and men for all the variables examined (datanot shown), as previously reported[19]. Thus, all the analyseswere run with both sexes together.

The facial proportions showed no significant correlationwith weight, height and BMI at the moment of examinationfor all groups (ANOVAp = ns, data not shown). In the coeliacgroups, similar non-significant correlations were found formain gastrointestinal findings and laboratory tests at the mo-ment of diagnosis of coeliac disease (ANOVAp = ns, datanot shown). In the coeliac group, none of the segments weresignificantly different statistically by presence or absence ofautoimmune diseases (ANOVAp = ns, data not shown).

Backward elimination of variables by stepwise procedure(at each step, removing the variable that contributes least to

Table 4Classification results

Group Predicted group membership Total

Control group Coeliac patients with diagnosis in adulthood

OriginalCount

Control group 63 37 100Coeliac patients with diagnosis in adulthood 35 62 97Coeliac patients with diagnosis in childhood 12 11 23

%

C

D egmen .T ct mea

ross va s othet

Control group 63.0a

Coeliac patients with diagnosis in adulthood 36.1Coeliac patients with diagnosis in childhood 52.2

ross-validatedb

CountControl group 63Coeliac patients with diagnosis in adulthood 35

%Control group 63.0c

Coeliac patients with diagnosis in adulthood 36.1

iscriminant analysis identified two facial proportions (tn/ns and ip/tg she table shows how the groups classify taking in account each subjea Of original grouped cases, 63.5% correctly classified.b Cross validation is done only for cases in predictable groups. In c

hat case.c

Of cross-validated grouped cases, 63.5% correctly classified.

37.0 100.063.9a 100.047.8 100.0

37 10062 97

37.0 100.063.9c 100.0

ts) that correctly classify 63% (about two-third) of the coeliac adults and controlssurements. Subjects percent correctly classified is in bold type.

lidation, each case is classified by the functions derived from all caser than

M. Finizio et al. / Digestive and Liver Disease 37 (2005) 659–664 663

group discrimination) isolated variables shown inTable 3.This procedure is univariate because each variable is consid-ered a single dependent variable in tests of inclusion.

Linear discriminant analysis was successful in distin-guishing between coeliac and control when ratios betweensegment tn and ns (tn–ns) and between ip and tg segments(ip–tg) had been used as variables, and with greater weightof the first ratio.

After stepwise elimination, the classification into thegroups remained significantly different using these variables(p < 0.0001). Weaker relationships were found in the remain-ing measurements (data not shown).

Discriminant analysis indicates that coeliac patients di-agnosed in childhood cannot be identified as a distinctivegroup.Table 4shows that coeliac patients with diagnosis inchildhood is equally attributed to the group coeliac (n 11)and control (n 12). For the 12 coeliac patients classified asbelonging to control, diagnosis of coeliac disease was madein the early months of life (5–14 months) and they reported astrict adherence to gluten-free diet throughout their lives. Forthe other patients classified in the coeliac group, three werediagnosed at age of 7; four were not compliant to diet and wecould not find any particular reason for the remaining five.

In summary, in Caucasian Mediterranean persons, regard-less of gender, the face can be divided into three almost equalparts and the present findings indicate that a large foreheadc

4

iter-r specto paredt thefi t inc or thea

thefi iter-r e dif-f n isp opulat tion.D /nsa d oft ism enceo

andd is, ast s orc mpli-a uliarc liaci to

gender, other anthropometric indices and are also unrelatedto reported symptoms at the moment of diagnosis.

In the present study, we referred to the surface structuresof the face in order to have a comprehensive visual grasp offace shape and its tentative interpretation in terms of differ-ential growth. The method is validated in comparison withother more sophisticated methods[20] and widely used bysurgeons[21,22], orthodontists[23], anatomists[24] and an-thropologists.

Viscerocranium development appears to be more influ-enced by epigenetic factors than neurocranium[25]. Verticalgrowth of the cranium is possibly more affected by gluten in-take in gluten-sensitive individuals in the first 5 years of life.In fact, coeliac patients of our series with very early diseasediagnosis on strict gluten-free diet during childhood do notshow the described craniofacial characteristics.

The possible explanation of the findings of the presentstudy is that gluten causes immunological and/or nutritivealteration during the development of the skull, affecting inparticular vertical growth of the viscerocranium. Malnutri-tion, in particular lack of food-derived proteins, may interferewith skull development as described in experimental stud-ies in animals fed on low-protein diet during early stage oflife [26]. Low calcium and low vitamin D diet cause cranio-facial abnormalities in experimental models[27]. It is wellknown that undiagnosed coeliac disease induces growth fail-u freed

s arep di-m t them ignif-i

pro-g easea thath e af-f ge ofl

ation,ha

ole.I inj l mat-u mea-s

liaca Thisi ain-t n, al andc ner.A ment.T nger( nd

an be a sign of coeliac disease.

. Conclusion

The main finding of the study is that Caucasian Medanean adult coeliac individuals tend to have a peculiar af the face characterised by a large forehead when com

o general population controls. To our knowledge, this isrst evidence of alterations of craniofacial developmenoeliac disease. We propose the name of coeliac face fbove-described sign.

Due to the homogeneity of the cohort under studynding must be limited to the Caucasians from the Medanean region at the moment. Other ethnic groups haverent average facial proportions and, if this disproportioresent, it must be measured and compared to general p

ion proportions of the same ethnic group under examinaiscriminant analysis identified two facial proportions (tnnd ip/tg segments) that correctly classify about two-thir

he coeliac adults and controls (Table 4). The hypothesisade that this alteration of growth pattern is the consequf gluten intake in gluten-sensitive individuals.

Data from the coeliac patients with early diagnosisiet therapy for coeliac disease support the hypothes

hey can be classified in the group of coeliac patientontrols depending on the age at diagnosis and the conce to gluten-free diet. Our results indicate that the pecraniofacial development (i.e. a large forehead) of coendividuals with diagnosis in adulthood is unrelated

-

re[14,15]and affects bone metabolism and that gluten-iet normalises bone mass in coeliac children[28].

In coeliac disease, however, craniofacial dimensionoorly related to nutritional indices. In fact, craniofacialensions are not related to weight, height and BMI aoment of measurements, although these factors are s

cantly related directly to diagnosis of coeliac disease.The possibility exists that body weight and height that

ressively vary along time can be affected by coeliac dislways during the growth, but craniofacial dimensions,ave fixed particular developmental time periods, can b

ected only if the disease develops in the very early staife.

It has been demonstrated also that the opposite situypernutrition, can affect craniofacial dimensions inobesedolescents [29].

We cannot exclude that immunity may also play a rn fact, El-Bialy et al. analysed craniofacial morphologyuvenile diabetes patients and found decreased skeletaration and decreased cephalometric linear and angularurements when compared with the control group[30].

In conclusion, the craniofacial morphology of coedults reveals an altered pattern of craniofacial growth.

s a clinical sign that should be included among the extrestinal manifestation of coeliac disease. In our opinioarge forehead is a sign easily evident to visual graspomputer analysis, is not required by a general practitiolarge forehead is easily detectable by manual measurehe empirical method we propose is to place the forefiindex finger) on the midpoint of the hairline (trichion) a

664 M. Finizio et al. / Digestive and Liver Disease 37 (2005) 659–664

the thumb on the midpoint of the nasofrontal suture (nasion).Then, maintaining the obtained distance of fingers, one placesthe forefinger on the nasion and if a significant space existsfrom the thumb and the base of the nose one can considerthat the forehead is larger than average. In our experience,at least 80% of the population with a southern Italy ancestrydo not show this sign. The physician should, in its presencetogether with other clinical signs and symptoms, be alertedto test the patient for coeliac disease.

Predictive value of the large forehead cannot be measuredby cross sectional analysis in a case–control study. In fact,longitudinal data on the incidence of new coeliac diseaseamong originally healthy persons with measured length offorehead are needed.

Conflict of interest statement

None declared.

Acknowledgements

The Authors are grateful to Prof. Stefania Montagnani forher stimulating suggestions. They are grateful also to Ms.Wendy Parsons for revising the manuscript. They thank Ms.A ingt rs.F inc

R

iatric

talians::328–

New

y ofppon

cepts

ic Sci

aun-

89.

[9] Hahn von Dorsche S, Fanghanel J, Kubein-Meesenburg D, Nagerl H,Hanschke M. Interpretation of the vertical and longitudinal growthof the human skull. Anat Anz 1999;181:99–103.

[10] Ochoa BK, Nanda RS. Comparison of maxillary and mandibulargrowth. Am J Orthod Dentofacial Orthop 2004;125:148–59.

[11] Marsh MN. Coeliac disease. Oxford Blackwell Scientific Publica-tions; 1992.

[12] Ciacci C, Cirillo M, Sollazzo R, Savino G, Sabbatini F, MazzaccaG. Gender and clinical presentation of adult coeliac disease. ScandJ Gastroenterol 1995;11:1077–81.

[13] Catassi C, Ratsch IM, Fabiani E, Bordicchia F, Candela F, CoppaGV, et al. Coeliac disease in the year 2000: exploring the iceberg.Lancet 1994;343:200–3.

[14] Lavy U, Bauer CH. Pathophysiology of failure to thrive in gastroin-testinal disorders. Pediatr Ann 1978;7:743–9.

[15] Booth IW. The nutritional consequences of gastrointestinal diseasein adolescence. Acta Paediatr Scand Suppl 1991;373:91–102.

[16] Lundstrom F, Lundstrom A. Natural head position as a basis forcephalometric analysis. Am J Dentofacial Orthop 1992;101:244–7.

[17] Bass NM. Measurement of the profile angle and the aesthetic analysisof the facial profile. J Orthod 2003;30:3–9.

[18] Dobrostanski T. Craniofacial photography in the living. In: ArnoldE, Clement JG, Ranson DL, editors. Craniofacial identification inforensic medicine. 1998, 137–49.

[19] Kharbanda OP, Sidhu SS, Sundrum KR. Vertical proportions of face:a cephalometric study. Int J Orthod 1991;29:6–8.

[20] Nute SJ, Moss JP. Three-dimensional facial growth studied by opticalsurface scanning. J Orthod 2000;27:31–8.

[21] Kolar JC, Salter EM. Craniofacial anthropometry. In: Practical mea-surement of the head and face for clinical, surgical and research use.Springfield: Charles C. Thomas; 1997, 334 pp.

[ ree-atingery. J

[ dimen-thod-

[ n-l

[ rin-

[ jec-1–9.

[ andcal-

314.[ nges

eliacNutr

[ llofontol

[ iale I).

ssunta Di Gilio and Mrs. Floriana Curcio for authorishe publication of their photographs. They thank also Miorella De Marco and Mr. Ugo Nicodemo for their helpomputer analysis of photographs.

eferences

[1] Cappa M, Loche S. Evaluation of growth disorders in the paedclinic. J Endocrinol Invest 2003;26(Suppl.):54–63.

[2] Farkas LG, Hreczko TA, Kolar JC, Munro IR. Vertical and horizonproportions of the face in young adult North American caucasrevision of neoclassical norms. Plast Reconstr Surg 1985;7537.

[3] Farkas LG. Anthropometry of the head and face in medicine.York: Elsevier; 1981.

[4] Suzuki A, Takahama Y. A cephalometric study on the similaritcraniofacial morphology between children and their parents. NiKyosei Gakkai Zasshi 1988;47:7–719.

[5] Cobourne MT. Construction for the modern head: current conin craniofacial development. J Orthod 2000;27:307–14.

[6] Jantz RL. Cranial change in Americans: 1850–1975. J Forens2001;46:784–7.

[7] Moore KL. The developing human. 4th ed. Philadelphia: W.B. Sders; 1988.

[8] Sperber GH. Craniofacial embryology. 4th ed. Oxford: Wright; 19

22] Landes CA, Zachar R, Diehl T, Kovacs AF. Introduction of a thdimensional anthropometry of the viscerocranium. Part II. Evaluosseous and soft tissue changes following orthognathic surgCraniomaxillofac Surg 2002;30:25–34.

23] Bishara SE, Jorgensen GJ, Jakobsen JR. Changes in facialsions assessed from lateral and frontal photographs. Part I. Meology. Am J Orthod Dentofacial Orthop 1995;108:389–93.

24] Zollikofer CPE, Ponce De Leon MS. Visualizing patterns of craiofacial shape variation inHomo sapiens. Proc R Soc Lond B BioSci 2002;269:801–7.

25] Mossey PA. The heritability of malocclusion. Part 1. Genetics, pciples and terminology. Br J Orthod 1999;26:103–13.

26] Miller JP, German RZ. Protein malnutrition affects the growth tratories of the craniofacial skeleton in rats. J Nutr 1999;129:206

27] Engstrom C, Linde A, Thilander B. Craniofacial morphologygrowth in the rat. Cephalometric analysis of the effects of a lowcium and vitamin D-deficient diet. J Anat 1982;134(Pt 2):299–

28] Barera G, Beccio S, Proverbio MC, Mora S. Longitudinal chain bone metabolism and bone mineral content in children with cdisease during consumption of a gluten-free diet. Am J Clin2004;79:148–54.

29] Ohrn K, Al-Kahlili B, Huggare J, Forsberg CM, Marcus C, DahG. Craniofacial morphology in obese adolescents. Acta OdScand 2002;60:193–7.

30] El-Bialy T, Aboul-Azm SF, El-Sakhawy M. Study of craniofacmorphology and skeletal maturation in juvenile diabetics (typAm J Orthod Dentofacial Orthop 2000;118:189–95.