Share this article on:

Leonardo da Vinci Meets Celiac Disease

Zanchi, Chiara*; Ventura, Giovanna; Di Leo, Grazia*; Orzes, Nicoletta; Ronfani, Luca*; Not, Tarcisio; Ventura, Alessandro

Journal of Pediatric Gastroenterology & Nutrition: February 2013 - Volume 56 - Issue 2 - p 206–210
doi: 10.1097/MPG.0b013e31827114a4
Original Articles: Gastroenterology

Background and Aim: Leonardo da Vinci's face symmetry derives from 3 equal craniofacial segments: trichion-nasion (tn), which represents the superior third of the face, nasion-subnasal (ns) that is the medium third of the face, and subnasal-gnathion (sg) that is the length of the lower third of the face. It has been reported that adult subjects with celiac disease (CD) can be identified on the basis of a greater extension of the forehead in comparison to the medium third of the face, with a high tn/ns ratio. The aim of the present study was to investigate the correlation between facial asymmetry and CD in childhood and adulthood.

Methods: We studied 126 biopsy-proven patients with CD (76 children and 50 adults) and 102 healthy controls (43 children and 59 adults). Their faces were photographed; the pictures were edited using a software program to calculate the facial segments.

Results: The tn length was significantly different between adult celiac and adult controls (7.43 ± 1.46 cm vs 6.38 ± 1.73 cm, P = 0.001). The cutoff of 6.5 cm tn, derived from receiver operating characteristic curve analysis, identified 43 of 50 patients (sensitivity 86%), but 34 of 59 controls were positive (specificity 54.2%). The positive predictive value was 56%; however, the tn/ns ratio was not significantly different between celiacs and controls. Neither the tn length nor the tn/ns ratio in celiacs correlated to the time of gluten exposure.

Conclusions: Adults, but not children, with celiac disease show a forehead extension significantly greater than controls, but this test's specificity appears too low to be used in the screening of CD.

*Institute for Maternal and Child Health—IRCCS “Burlo Garofolo”

University of Trieste

Department of Gastroenterology and Digestive Endoscopy, Civil Hospital, Gorizia, Italy.

Address correspondence and reprint requests to Chiara Zanchi, Department of Pediatrics, Institute for Maternal and Child Health—IRCCS “Burlo Garofolo”—University of Trieste, Via dell’Istria 65, 34100 Trieste, Italy (e-mail:

Received 5 June, 2012

Accepted 14 August, 2012

The authors report no conflicts of interest.

Celiac disease (CD) is an autoimmune disorder triggered by gluten in genetically predisposed individuals (1). The clinical picture varies from apparently asymptomatic cases to severe malabsorption syndrome. Gluten-free diet is the only effective therapy, whereas unrecognized and untreated cases are exposed to an increased risk of malignancy, osteoporosis, other autoimmune disorders, and an increased mortality (1–3). The detection of some apparently negligible signs (eg, systematic dental enamel defect) may help in recognizing the cases that otherwise may remain unrecognized (4,5).

In his famous “Vitruvian Man,” a mathematic design in which the union between scientific spirit and artistic intuition is expressed, Leonardo da Vinci defines the ideal and perfect proportions of the human body, and divides the face into 3 equal parts (Fig. 1).

A recent study suggests that adult celiac patients have a typical asymmetry of the face, with a forehead that is larger than the other two-thirds of the face. The authors’ suggestion is that this measurement could be useful for the diagnosis and screening of CD because it is feasible at a glance (7). This hypothesis has not yet been confirmed in the literature.

In our work we verified whether the measurement of the length and the proportions of the face's segments according to Leonardo da Vinci may be useful for the clinical diagnosis of CD, both in adulthood and in childhood.

Back to Top | Article Outline



A total of 126 celiac subjects (87 females and 39 males) and 102 healthy controls (62 females and 40 males) were enrolled and 4 subgroups were compared: 50 adults with known CD (from now on referred to as group 1), 76 children with known CD (group 2), 59 healthy adults (group 3), and 43 healthy children (group 4). Celiac adult and children patients (groups 1 and 2) were recruited among the outpatient clinic for periodic follow-up (subjects already on a gluten-free diet) and, in some cases, among the day-hospital patients coming for CD diagnosis formalization. In all of the patients, the diagnosis of CD was based on the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition criteria (positive serology and typical duodenal mucosa aspect). Healthy adult and children controls (groups 3 and 4) were recruited among inpatients accessed to the hospital for acute disease and among outpatients; none of these subjects had a gastroenterological problem. All of the controls had a negative antitransglutaminase antibodies rapid assay.

Subjects with facial asymmetry, or who had had facial trauma and/or facial plastic surgery, and bold subjects in whom a precise extracranic cephalometric analysis was difficult, were excluded. Moreover, subjects with chronic diseases (rheumatologic, endocrinologic, and metabolic diseases) were excluded.

The present study took place at the Pediatric Department of the IRCCS Burlo Garofolo of the University of Trieste, Italy, and at the Gastroenterology and Digestive Endoscopy Division of the Civil Hospital of Gorizia, Italy.

Back to Top | Article Outline


Pictures were taken from a distance of 150 cm with a digital camera (3.2 megapixels) and with the subject's head in a “natural head orientation”; natural head orientation is the position that the physician considers the natural head position, that is, the position studied to make the cephalometric analysis using extracranic reference points instead of intracranic ones (8–11).

Basically, natural head position is the natural position assumed by the head when a subject is asked to stand up and to look straight forward to a distant point at an eye level (9–12). The same person, who did not know the diagnosis, took all the pictures with a constant zoom.

The pictures were thereafter digitally edited using Adobe Photoshop CS (Adobe, San Jose, CA). Each picture was rotated until the interpupil distance resulted parallel to the lower side of the picture frame; in most cases, the rotation needed was <3 degrees. Then, 6 reference points were traced, and the distances between these points and the ratio between these distances were measured (Fig. 2).

We decided to consider also the ratio between the measurements, besides the absolute measurements, to reduce the mistakes because of the camera and the subject position as much as possible, and to obtain the more coherent and reproducible pictures as possible.

Reference points were as follows:

1. trichion: the median point of the hair line;

2. pupillar: the lowest point of the pupils;

3. nasion: the median point of the nasofrontal suture;

4. subnasal: the point on the median line, between the lowest part of the anterior nasal spine and the cutaneous portion of the upper lip;

5. gnathion: the lowest point on the lower chin margin, on the median line.

Segments used for the analysis were as follows:

1. interpupillar (ip): distance between the lowest points of the pupils circumference;

2. trichion-nasion (tn): the superior third of the face (it represents the forehead width);

3. nasion-subnasal (ns): the median third of the face (it represents the nose length);

4. subnasal-gnathion (sg): the inferior third of the face;

5. trichion-gnathion (tg): the total face height.

Then the same anthropometric ratios used by Finizio et al (7) (tn/ns, ip/tg, ns/tg, sg/tg, tn/ip) were calculated for each subject.

In the present study, we focused on the forehead. A wide forehead was assessed on the absolute tn segment length and on the ratio between the tn and the ns (tn/ns) segments.

All of the control subjects were tested by the rapid immunoglobulin A antitissue transglutaminase test; rapid assay was performed following the manufacturer's instructions (Eu-tTG Quick, Eurospital, Italy) by the same person (the physician).

Written informed consent was obtained from all of the individuals and from the parents of the children enrolled. The study was approved by the independent ethical committee of the Institute of Child Health IRCCS “Burlo Garofolo” (C.I.B.27/09).

Back to Top | Article Outline

Statisical Analysis

Continuous variables are summarized as means and standard deviations (SD), categorical variables as proportions and percentages. Differences in the study outcomes between cases and controls (celiac vs nonceliac) were explored using a nonparametric test (Mann-Whitney test). To explore the influence of possible explanatory variables (age/time of gluten exposure and malabsorption symptoms) on differences in main study outcomes (tn and tn/ns) between cases and controls stratification, Spearman correlation and multivariate linear regression analysis were carried out.

To identify the best tn cutoff, a receiver operating characteristic (ROC) curve analysis was carried out. Sensitivity, specificity, and positive and negative predictive values were calculated using this cutoff. All data management and analysis were done by the SPSS package (version 11.5, SPSS Inc, Chicago, IL) and STATA software (StataCorp, College Station, TX).

Back to Top | Article Outline



There was an overlap between the mean age of celiac patients and of healthy controls (21.57 ± 18.3 years vs 21.70 ± 13.90 years; range respectively of 2–88 and 1–68). CD patients’ mean age at diagnosis was 18.40 ± 18.20 years (range 0.6–83). Study group features are shown in Table 1.

The diagnosis of CD was based on the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition's criteria: all of the patients were tested positive for IgA antitransglutaminase antibodies (mean value 53 ± 31 IU/mL; normal values <7 IU/mL), and all of them presented an atrophic intestinal mucosa that was classified according to Oberhiber criteria: 85 of 126 were classified with 3C degree, 35 with 3b degree, and 6 of 126 with 3A degree.

At the diagnosis time, 27 adults of 50 and 63 pediatric patients of 76 suffered from typical symptoms (diarrhea, weight loss, failure to thrive, iron-deficiency anemia), whereas 29 and 12 children experienced atypical symptoms (autoimmune disease, alopecia, vitiligo, dermatitis herpetiformis). Only 7 adults and 1 child were asymptomatic and had been detected by CD family screening.

Back to Top | Article Outline

Morphostructural Differences Among Celiac and Nonceliac Subjects

Mean tn, ns, sg, and tg were greater in adult celiac subjects compared with healthy age-matched controls (respectively tn 7.43 ± 1.46 cm vs 6.38 ± 1.72 cm, P = 0.001; ns 5.89 ± 1.04 cm vs 5.25 ± 1.45 cm, P = 0.012; sg 7.57 ± 1.48 cm vs 6.91 ± 1.77 cm, P = 0.041; tg 20.89 ± 3.73 cm vs 18.55 ± 4.8 cm, P = 0.006) (Table 2). Mean tn, ns, sg, and tg lengths in celiac children, instead, were not statistically different from the healthy matched controls (respectively, tn 6.26 ± 2.22 cm vs 7.05 ± 2.86 cm, P = 0.09; ns 4.73 ± 1.8 cm vs 5.3 ± 2.0 cm P = 0.13; sg 5.90 ± 2.12 cm vs 6.72 ± 2.7 cm, P = 0.06; tg 16.89 ± 6 cm vs 19.06 ± 7.43 cm, P = 0.08). There was no difference between the tn/ns ratio, ns/tg ratio, and sg/tg ratio in celiac and nonceliac subjects, both in adults (1.26 ± 0.12 vs 1.22 ± 0.12, P = 0.083) and in children (1.35 ± 0.19 vs 1.34 ± 0.19, P = 0.92) (Table 2).

Fifty-four percent adult celiacs (n = 27) had typical malabsorption symptoms at onset. The presence of these symptoms did not influence the tn values (7.4 SD 1.1 vs 7.5 SD 1.8; P = 0.9) nor the tn/ns ratio (1.24 SD 0.13 vs 1.28 SD 0.1; P = 0.3).

There was no correlation between the tn value and the time of gluten exposure, and between the tn/ns ratio and the time of gluten exposure in adult celiac subjects (Spearman correlation coefficient = 0.1 and 0.02, respectively). There was no correlation between the craniofacial parameters and both the antitransglutaminase serum concentrations and the intestinal atrophy degree at the diagnosis time.

Multivariate analysis evaluating the influence of age and symptoms at onset on the face proportions confirmed that tn is different in celiac subjects and in healthy controls, and it only depends on being affected by CD or not.

Using a cutoff tn value of 6.5 cm, identified by the ROC curve analysis, for defining celiac and nonceliac subjects (if tn ≥6.5 then celiac subject, if tn <6.5 then nonceliac subject), this test has a sensibility of 86% and a specificity of 54% (Table 3). The positive predictive value (the chance that tn ≥6.5 correctly identifies a celiac subject) is 56%, whereas the negative predictive value (the chance that tn <6.5 correctly rules out a nonceliac subject) is of 78% (Table 3).

The cutoff values for ns, sg, and tg segments in adults (respectively, 5.3, 6.4, and 17.7), obtained with the ROC curve analysis, allow us to create a sensible test (sensibility of 86% for ns, 90% for sg, and 90% for tg) but with a low specificity and a poor positive predictive value (respectively, 49% and 59% for ns; 35% and 36% for sg; 44% and 58% for tg).

Back to Top | Article Outline


Our study confirms the presence of asymmetry of the face proportions in celiac adults; in particular celiac adults have a wider forehead than healthy controls. Contrarily to previously published data (7), we did not find a statistically significant difference in the segments ratios between celiac subjects and healthy controls, neither in adults nor in children, but we found a significant difference between all of the face segments (tn, ns, sg, and tg) in celiac and healthy adults.

As a test for CD, the measurement of the face segments (tn, ns, sg e tg) in adults is not enough specific to be proposed as a screening test in clinical celiac diagnosis; however, a wider forehead characterized 86% adult celiacs and it could be used to raise clinical suspect of CD in adult patients.

Our data suggest that face proportion alterations are irreversible, but at the same time preventable by early diagnosis in pediatric age because there are no differences between celiac and healthy children. This is possibly because of the brief time of gluten exposure and a longer exposure to gluten may be necessary to induce morphostructural alterations (13–16).

Moreover, our results differ from the previous ones for the absence of any correlation between facial asymmetry and clinical history and the time of gluten exposure. This could be because of a shorter time of gluten exposure in all of the pediatric patients (younger than 9 years) and a too long time of gluten exposure in adult patients (none of the adults had received diagnosis in pediatric age).

We could speculate that the time of gluten exposure was too long in adults to prevent the morphostructural face asymmetry and too brief in children (early diagnosis in all children) to permit the bone face alterations.

Early starting a gluten-free diet could rapidly reestablish normal bone metabolism, avoiding alterations in the face proportions. All of the children enrolled in our study received a strict gluten-free diet and the mean age at diagnosis was very precocious (5.54 ± 2.92 years). As formerly demonstrated in other studies (17–20), bone metabolism alterations, basically caused by calcium and vitamin D malabsorption and by inflammatory cytokines produced by the celiac subject receiving a gluten-containing diet, are completely and rapidly reversed if a gluten-free diet is started in pediatric age and rigorously followed. Because the relation between bone malformation and inflammatory bowel diseases has a reasonable basis (21–26), the hypothesis that bone metabolism alterations in the early phases of the splanchnocranium development can determine an alteration of the facial morphology in the celiac subject receiving a gluten-containing diet is interesting and deserves being tested in other prospective studies.

The mechanisms that regulate craniofacial growth and development are complex and include interactions between genes, hormones, nutrients, and epigenetic factors, during the first years of life, that will give the craniofacial bones their final morphology (27,28). Disturbances in any of these mechanisms may result in an aberrant growth pattern. Recently an interesting correlation has been demonstrated between the disturbances in dental development and vertical growth of the face (29). In conclusion, we can hypothesize that particular craniofacial alterations of celiacs derive from immunological and nutritive alteration induced by gluten exposure during the growth, not only through direct damage to skull bone, but also through damage to alveolar bone, root, and enamel dental development, affecting in particular vertical growth of the viscerocranium.

In conclusion, our study confirms the tendency of adult celiac subjects to having a wider forehead than healthy controls, but this alteration is not strong enough to be used as a “new clinical sign of CD,” as previously suggested (7). The craniofacial alterations in celiacs seem to depend on the gluten exposition through pediatric age because celiac patients diagnosed in pediatric age have normal face proportion. A large forehead is a sign at the glance evident, no cost, so if associated with the presence of other clinical sign and symptoms, should alert a physician to test a patient for CD. Other prospective studies, such as case-finding studies, should be conducted to understand the role of the exposure to gluten in the splanchnocranium development of celiac subjects.

Back to Top | Article Outline


1. Di Sabatino A, Corazza GR. Coeliac disease. Lancet 2009; 373:1480–1493.
2. Mustalahti K, Collin P, Sievänen H, et al. Osteopenia in patients with clinically silent coeliac disease warrants screening. Lancet 1999; 354:744–745.
3. Ventura A, Not T, Tommasini A, et al. Autoantibodies as predictors of disease. Lancet 2004; 364:1403–1404.
4. Mäki M, Aine L, Lipsanen V, et al. Dental enamel defects in first-degree relatives of coeliac disease patients. Lancet 1991; 337:763–764.
5. Martelossi S, Zanatta E, Del Santo E, et al. Dental enamel defects and screening for coeliac disease. Acta Paediatr Suppl 1996; 412:47–48.
6. Deleted in proof.
7. Finizio M, Quaremba G, Mazzacca G, et al. Large forehead: a novel sign of undiagnosed coeliac disease. Dig Liver Dis 2005; 37:659–664.
8. Ochoa BK, Nanda RS. Comparison of maxillary and mandibular growth. Am J Orthod Dentofacial Orthop 2004; 125:148–159.
9. Bass NM. Measurement of the profile angle and the aesthetic analysis of the facial profile. J Orthod 2003; 30:353.
10. Lundstrom F, Lundstrom A. Natural head position as a basis for cephalometric analysis. Am J Orthod Dentofacial Orthop 1992; 101:244–247.
11. Lundstrom A, Lundstrom F, Lebret LM, et al. Natural head position and natural head orientation: basic consideration in cephalometric analysis and research. Eur J Orthod 1995; 17:111–120.
12. Halazonetis DJ. Estimated natural head position and facial morphology. Am J Orthod Dentofacial Orthop 2002; 121:364–368.
13. Suzuki A, Takahama Y. A cephalometric study on the similarity of craniofacial morphology between children and their parents. Nippon Kyosei Gakkai Zasshi 1988; 47:7–719.
14. Cobourne MT. Construction for the modern head: current concepts in craniofacial development. J Orthod 2000; 27:307–314.
15. Jantz RL. Cranial change in Americans: 1850–1975. J Forensic Sci 2001; 46:784–787.
16. Mossey PA. The heritability of malocclusion: part 1—Genetics, principles and terminology. Br J Orthod 1999; 26:103–113.
17. Zanchi C, Di Leo G, Ronfani L, et al. Bone metabolism in celiac disease. J Pediatr 2008; 153:262–265.
18. Mora S. Celiac disease: a bone perspective. J Pediatr Gastroenterol Nutr 2003; 37:409–411.
19. Barera G, Beccio S, Proverbio MC, et al. Longitudinal changes in bone metabolism and bone mineral content in children with celiac disease during consumption of a gluten free diet. Am J Clin Nutr 2004; 79:148–154.
20. Mora S, Barera G, Beccio S, et al. A prospective, longitudinal study of the long term effect of treatment on bone density in children with celiac disease. J Pediatr 2001; 139:516–521.
21. Georgios K, Helias A, Athanassios K, et al. Craniofacial surgical management of a patient with systematic juvenile idiopathic arthritis and Crohn's disease. J Craniofac Surg 2009; 20:948–950.
22. Lobe SL, Bernstein MC, German RZ. Life-long protein malnutrition in the rat (Rattus norvegicus) results in altered patterns of craniofacial growth and smaller individuals. J Anat 2006; 208:795–812.
23. Miller JP, German RZ. Protein malnutrition affects the growth trajectories of the craniofacial skeleton in rats. J Nutr 1999; 129:2061–2069.
24. Engström C, Linde A, Thilander B. Craniofacial morphology and growth in the rat. Cephalometric analysis of the effects of a low calcium and vitamin D-deficient diet. J Anat 1982; 134 (Pt 2):299–314.
25. Miheller P, Lorinczy K, Lakatos PL. Clinical relevance of changes in bone metabolism in inflammatory bowel disease. World J Gastroenterol 2010; 16:5536–5542.
26. Ghishan FK, Kiela PR. Advances in the understanding of mineral and bone metabolism in inflammatory bowel diseases. Am J Physiol Gastrointest Liver Physiol 2011; 300:G191–G201.
27. Liu YP, Behrents RG, Buschang PH. Mandibular growth, remodeling and maturation during infancy and early childhood. Angle Orthod 2010; 80:97–105.
28. Sadeghianrizi A, Forsberg CM, Marcus C, et al. Craniofacial development in obese adolescent. Eur J Orthod 2005; 27:550–555.
29. Vesterbacka M, Ringden O, Remberger M, et al. Disturbances in dental development and craniofacial growth in children treated with hematopoietic stem cell transplantation. Orthod Craniofac Res 2012; 15:21–29.

bone metabolism; celiac disease; facial asymmetry; forehead extension; gluten

Copyright 2013 by ESPGHAN and NASPGHAN