Prevalence of Autoimmune Thyroiditis in Children with Celiac Disease andEffect of Gluten Withdrawal

Antonella Meloni, MD, Chiara Mandas, MD, Rita Desiree Jores, MD, PhD, and Mauro Congia, MD

Objective To study the prevalence of autoimmune thyroiditis (AT) in Sardinian children with celiac disease (CD)and the effects of a gluten-free diet (GFD) on thyroid function.Study design Children with biopsy-proven CD (n = 324; female:male 2:1; mean age, 6.6 years) followed from 1 to15 years, were retrospectively evaluated for AT at onset of CD and during GFD. Serum thyroid peroxidase and thy-roglobulin antibodies (AbTPO, AbTG), thyroid-stimulating hormone (TSH), free thyroxine (FT4), free triiodothyronine(FT3), and thyroid ultrasonography were considered. Age-matched Sardinian schoolchildren (n = 8040), previouslyevaluated for antithyroid antibodies and thyroid function, were used as controls.Results Thirty-four patients with CD (10.5%) developed AT (female:male 4,5:1; mean age, 10.5 years), 11 at onsetof CD and 23 during GFD, with a higher prevalence than controls (P = 2.9�13). Twenty-eight patients were euthyroidand 6 hypothyroid. AbTPO and/or AbTG persisted elevated for 2 to 9 years despite the GFD in 9 of 11 patients withAT at onset of CD.Conclusions AT is strongly associated with CD in Sardinian children, has an age of onset of 10.5 years, and ap-pears to be gluten-independent. In children with CD with AT, the female:male bias reported in adult AT is presentbefore puberty. (J Pediatr 2009;155:51-5).

Celiac disease (CD) is an autoimmune-mediated enteropathy triggered and maintained by the ingestion of gluten-con-taining cereals (wheat, rye, and barley) in genetically predisposed individuals.1 CD is considered to be a worldwide dis-order, more common in Western countries. In Europe and in the United States, the disease has been found to affect

about 1 in 100 people2; among Sardinians with CD, it has been reported with a similar incidence.3 Like type 1 diabetes, rheu-matoid arthritis, and multiple sclerosis, CD has a strong genetic association with particular HLA class II alleles, namely the HLAgenes encoding the class II DQ (a1*0501, b1*02) molecule.4

Sardinia is an ancient genetic isolate with a peculiar distribution of HLA class II alleles and haplotypes.5 The CD predisposingHLA class II DQ (a1*0501, b1*02) molecule is very frequent among Sardinians (43%), and this HLA molecule is almost alwaysencoded in cis by the DRB1*0301, DQA1*0501, DQB1*0201 haplotype.6

An increased prevalence of autoimmune disorders, including type 1 diabetes and autoimmune thyroid diseases, has beenrepeatedly reported in patients with CD.7-9

Two hypotheses have been suggested to explain this association: (1) sharing of 1 or more genes responsible for CD and for thecoexpressed autoimmune disease; or (2) continued gluten exposure in untreated CD might lead to loss of the intestinal barrierfunction and to alterations of the systemic immune response, ultimately helping to induce other autoimmune disorders.10

The first concept implies that the comorbidity associated with CD is part of the individual genetic background.11-13 On thecontrary, the second concept may have important clinical implications because when the gluten induced immune reactivity andthe interplay between autoimmune predisposing genes and trigger(s) is blocked, this comorbidity could be eliminated.14,15

In this retrospective study, a large cohort of 324 children with CD from Sardinia, followed for 1 to 15 years, was evaluated forthe occurrence of autoimmune thyroiditis (AT). The aims of the study were to establish the prevalence of AT in CD and toverify if gluten exposure before the diagnosis of CD is correlated with development of thyroid autoimmunity, or vice versaif early gluten withdrawal is able to prevent the future development of AT.

From the Pediatric Clinic II, ‘‘Microcitemico’’ HospitalASL 8, Department of Biological Sciences andBiotechnology, University of Cagliari, Sardinia, Italy

Supported by a grant from Fritz-Thyssen-Stiftung, Ger-many (R.-D.J.). The authors declare no potential conflictsof interest, real or perceived.

0022-3476/$ - see front matter. Copyright � 2009 Mosby Inc.

All rights reserved. 10.1016/j.jpeds.2009.01.013

AbTG Thyroglobulin antibody

AbTPO Thyroid peroxidase antibody

AT Autoimmune thyroiditis

CD Celiac disease

FT3 Free triiodothyronine

FT4 Free thyroxine

GFD Gluten-free diet

TSH Thyroid-stimulating hormone

US Thyroid ultrasonography

51

THE JOURNAL OF PEDIATRICS � www.jpeds.com Vol. 155, No. 1

Methods

PatientsChildren with biopsy-proven CD (n = 324) from Sardinia(223 girls; mean age at diagnosis, 6.6 years; range, 10 monthsto 18 years), followed between 1992 and 2007 (mean follow-up period, 8 years) at the Pediatric Gastroenterological Unitin Cagliari, Italy, were included in the study. Clinical datawere collected retrospectively; affected family members ofprobands, patients with preexisting autoimmune conditions(type 1 diabetes, autoimmune thyroid disorders, Addison’sdisease), and patients with Down syndrome or Turner syn-drome were excluded from the study.

Screening for Celiac DiseaseThe diagnosis of CD was based on the revised criteria of theEuropean Society for Pediatric Gastroenterology, Hepatol-ogy, and Nutrition,16 for example, clinical history, positivityof IgA antiendomysial antibody, typical histological featureson small intestinal biopsy, and clinical response to the gluten-free diet (GFD) (Table I).17 All patients were typed foranti-gliadin antibody (AGA, both IgA and IgG) and antien-domysial antibody (EMA); most patients were also typedfor anti-tissue transglutaminase antibody IgA (tTG-IgA)and anti-actin antibody IgA (AAA-IgA). The latter has beenpreviously associated with the severity of intestinal damage.18

Mucosal atrophy has been graded according to the Marshclassification,19 modified by Oberhuber in type 3a, 3b, and3c.20 After diagnosis of CD, all patients followed a GFD,and the dietary compliance was evaluated by assessment ofAGA, EMA, and tTG every 12 months.

Thyroid Function AssessmentScreening for thyroid peroxidase antibody (AbTPO) and thy-roid globulin antibody (AbTG) was performed in all 324 pa-tients at the onset of CD and once per year after theinstitution of the GFD. When abnormal values were detected,a complete thyroid evaluation, including serum-free triiodo-thyronine (FT3), free thyroxine (FT4), thyroid-stimulatinghormone (TSH), and thyroid ultrasonography, was per-formed.

Thyroid autoantibody serum titers were detected usingcompetitive radioimmunoassay (RIA Medical System;Genoa, Italy) with coated-tube technique thyroperoxidaseantibodies (AbTPO, normal values <50 IU/mL), RIA-immu-noradiometric assay (IRMA, Medical System; Genoa, Italy)with solid-phase technique antithyroglobulin (Ab TG, nor-mal values <100 IU/mL). Serum FT3 and FT4 were detectedby RIA (Medical System; Genoa, Italy) using commercial kits(normal values, 0.8 to 2 ng/mL for FT4; 1.40 to 4.2 pg/mL forFT3), whereas serum TSH was measured by a chemilumines-cent method (Medical System; Genoa, Italy), with normalvalues ranging between 0.3 and 5 mU/mL.

Ultrasonography of the thyroid gland was performed witha 7.5-MHz linear probe. Thyroid echographic images wereclassified and graded in 4 patterns according to Sostre

52

et al21 and Marcocci et al.22 We restricted the diagnosis ofAT to patients with high titers of AbTPO and/or AbTG asso-ciated with an abnormal thyroid echographic pattern. There-fore, patients with mild elevation of AbTPO and/or AbTGwere eliminated from the statistical analysis to avoid possibleoverestimates. Thyroid function was classified according tothe American Thyroid Association Guidelines.23

Control SubjectsWe compared the frequency of AT among patients with CDwith the prevalence of antithyroid antibodies (ATA) in anage-matched Sardinian background population also evalu-ated for thyroid function.24

In this cohort of 8040 schoolchildren (4194 boys, 3846girls, ages 6 to 15 years) ATA were detected in 235 children(2.92%).

Statistical AnalysisData were evaluated by c2 test with Yates corrections, or bythe Wilcoxon test, setting the level of significance at .05.

Results

Clinical data of patients with CD with AT are summarized inTable II (available at www.jpeds.com).

Overall AT was found in 34 (28 girls, 6 boys) of 324patients with CD; 23 were on GFD when AT was diagnosed(Table III). Considered as a whole, a high prevalence of ATamong children with CD (10.5%), compared with the Sar-dinian pediatric background population (2.92%), was found(P = 2.9�13). This prevalence is similar to or even lower thanthat reported in other studies conducted in adults and chil-dren.12,25 Euthyroidism was observed in 28 of 34 patients(82.4%), and 6 were hypothyroid. In 3, hypothyroidism pre-ceded diagnosis of CD, and the other 3 patients developedhypothyroidism during GFD (Table II).

The thyroid ultrasonography showed an abnormal patternwith varying severity, suggestive of an autoimmune infiltrate

Table I. Clinical manifestations and serological markersof CD in 324 Sardinian children

Patients (%)

Clinical formClassic symptoms 39Nonclassic symptoms 56Iron deficiency anemia 22Failure to thrive 18Abdominal pain 5Pubertal delay 4Vomiting 3Diarrhea 2Chronic constipation 2Silent 5

Serological markersAGA IgA 73AGA IgG 88tTG-IgA 95EMA 99AAA-IgA 64

Meloni et al

July 2009 ORIGINAL ARTICLES

Table III. Prevalence of AT among patients with CD according to sex, age (in years), and treatment status

CD CD + AT AT in untreated patients (GCD) AT in treated patients (GFD)

No. of patients 324 34/324 11/34 23/34Female/male (ratio) 223/101 (2.2:1) 28/6 (4.5:1) 10/1 (10:1) 18/5 (3.6:1)Mean age �SD 6.6 � 4.5* 10.5 � 2.9* 9.5 � 2.9 11.0 � 2.8Age range 1-18 5-16 5-15 7-16

*Statistically significant difference (P< 1.4�6) between the age of onset of CD alone and CD with AT.

(Table II). Two patients with isolated high levels of AbTPOand AbTG at CD diagnosis developed an overall AT witha grade 2 of Sostre after 1 year of GFD. A significant differ-ence in the age of onset between patients with CD and pa-tients with CD and AT was observed: 6.6 versus 10.5 years,respectively (P < 1.4�6; Table III). We were aware that a de-tection bias might have influenced the different ages of dis-eases onset because, by definition, AT was looked for onlyafter the diagnosis of CD. On the other hand, no significantdifference in the age of AT onset among patients at CD diag-nosis (9.5� 2.9) and patients on GFD (11.0� 2.8) was found(Table III).

The female/male ratio was higher in patients with CD andAT (F/M 4.5:1), compared with patients with CD alone (F/M2.2:1; Table III). This finding was unexpected in our cohortof pediatric patients because a similar high F/M ratio in AT iscommonly described in late adolescence and adulthood.26,27

Thus, to determine a possible role of sex hormones as an ex-planation for the observed F/M ratio, patients were subdi-vided into 2 groups: prepubertal (age #12 years) andpubertal (age >12 years) (Table IV). Unexpectedly, the F/M ratio in prepubertal (12:1) was even higher than in puber-tal patients (1:1), and the difference, when compared with theF/M ratio of prepubertal patients with CD alone (2:1),reached statistical significance (P < .0068; Table IV).

The Figure shows the relationship of AT with the durationof gluten exposure.

At onset of CD, 11 of 324 patients receiving a gluten-con-taining diet were affected by AT. After diagnosis of CD, anadditional 23 patients placed on GFD developed AT after 1to 11 years, and 3 of these shifted toward hypothyroidism.No difference was found in the serum titers of AbTPO,AbTG, and thyroid echographic pattern between patientswith AT coexisting at the onset of CD and patients develop-ing AT during GFD (Table II). In 9 of 11 patients with ATcoexisting at the diagnosis of CD, very high titers of AbTGand/or AbTPO, suggestive of active AT, persisted despite 1to 9 years of GFD (data not shown).

Prevalence of Autoimmune Thyroiditis in Children with Celiac D

Discussion

Our study documents the association between AT and CD inSardinian children. Because it includes a large number of sub-jects, it provides the actual prevalence of AT in our pediatricpopulation with CD. Previous studies on the prevalence of ATin Sardinian patients with CD, or vice versa on the occurrenceof CD in AT, have been conducted only in adults.9,28

We found that the prevalence of AT in CD is about 4 timesgreater than that found in the age matched Sardinian school-children background population, indicating that CD is a riskfactor predisposing to AT.

There are several possible explanations for the increasedrisk of developing AT in patients with CD: (1) the CD condi-tion carries 1 or more genes (MHC or other autoimmunitygenes) favoring AT; and (2) the timing of CD diagnosisand duration of gluten exposure might increase toxic and im-munological effects of gluten, increasing the risk of AT.14,15,29

A lack of correlation between CD status and occurrence ofAT, favoring the first hypothesis, is suggested by our findings.Indeed, AT developed in patients on a gluten-containing dietor in patients on GFD; in these latter patients, after manyyears (Table II; Figure) GFD did not improve the outcomeof the disease, as demonstrated by very high titers of AbTGand/or AbTPO after 1 to 9 years of follow-up. There wasno significant difference in the mean age between patientswith AT coexisting at onset of CD and patients developingAT during the GFD: around 10.5 years in both groups, indi-cating that AT is independent of gluten exposure (Table III).The age of patients with CD with AT is the same observed inSardinian children with isolated AT.30 Finally, the significantdifference between the age of onset of CD (6.6 years) and CDwith AT (10.5 years) suggests that CD and AT are two differ-ent autoimmune diseases, which association is likelyexplained by a shared genetic background.

Our data, showing that the duration of gluten exposure isnot of crucial importance in favoring the development of

Table IV. Comparison of female/male ratio according to age (prepubertal versus pubertal) in patients with CD andpatients with CD and AT

Patients CD F M F/M CD + AT F M F/M P value

Total (n) 290 195 95 2:1 34 28 6 4.6:1 NSPrepubertal (#12 y) 254 169 85 2:1 25 24 2 12:1 <.007Pubertal (>12 y) 36 26 10 2.6:1 8 4 4 1:1 NS

P value determined by c2 test.F, female; M, male; NS, P value not significant.

isease and Effect of Gluten Withdrawal 53

THE JOURNAL OF PEDIATRICS � www.jpeds.com Vol. 155, No. 1

Figure. Vertical bars represent the number of patients with CD diagnosed with AT. The first bar represents patients with AT at CDdiagnosis, the following represent patients developing AT during GFD.

AT, are in agreement with some authors11-13 and in contrastwith others.14,15,29 Ventura et al14 found that the prevalenceof AT correlated with the duration of gluten exposure andaugmented with the age at CD diagnosis. In addition, thyroidautoantibodies had a tendency to disappear following theGFD. Possible explanations for these variable results could de-pend on different follow-up periods (15 years in our study and6 months in the work of Ventura et al) and on age stratifica-tions chosen by Ventura that did not take into account thatthe prevalence of certain autoimmune diseases raises withaging.

A peculiar and previously unreported finding of our studyconcerns the F/M ratio of patients with CD and AT and therelationship between AT and puberty.

The prevalence of girls in AT is well known, and the sex ra-tio favoring females is reported to be age-related, startinggenerally after puberty.26,27 This observation has supportedthe belief in a fundamental role of sexual hormones in favor-ing the development of AT. Accordingly, in the Sardinianschoolchildren population,24 a significant increase of thyroidautoimmunity was more evident in females than in males, inparticular over 11 years (P < .01). Interestingly, in patientswith CD with AT, the high F/M ratio was observed before pu-berty (Table IV), suggesting either a more complicated roleof the sex hormones in conferring predisposition to AT, or,alternatively, an influence of the genetic background of CDin anticipating the high F/M ratio of AT.

The high frequency of AT among patients with CD, evenon GFD, suggests clinical usefulness for a longitudinal fol-low-up of thyroid function. n

54

The authors are grateful to Prof. A. Cao for critical reading of the man-uscript.

Submitted for publication Aug 29, 2008; last revision received Nov 24, 2008;

accepted Jan 8, 2009.

Reprint requests: Dr Antonella Meloni, Pediatric Clinic II, ‘‘Microcitemico’’

Hospital ASL 8, Department of Biological Sciences and Biotechnology,

University of Cagliari, Sardinia, Italy. E-mail: ameloni@mcweb.unica.it.

References

1. Green PH, Cellier C. Celiac disease. N Engl J Med 2007;357:1731-43.

2. Fasano A, Catassi C. Current approaches to diagnosis and treatment of

celiac disease: an evolving spectrum. Gastroenterology 2001;120:636-51.

3. Meloni G, Dore A, Fanciulli G, Tanda F, Bottazzo GF. Subclinical coeliac

disease in schoolchildren from northern Sardinia. Lancet 1999;353:37.

4. Sollid LM. Coeliac disease: dissecting a complex inflammatory disorder.

Nat Rev Immunol 2002;2:647-55.

5. Lampis R, Morelli L, De Virgiliis S, Congia M, Cucca F. The distribution

of HLA class II haplotypes reveals that the Sardinian population is genet-

ically differentiated from the other Caucasian populations. Tissue Anti-

gens 2000;56:515-21.

6. Congia M, Frau F, Lampis R, Frau R, Mele R, Cucca F, et al. A high fre-

quency of the A30, B18, DR3, DRw52, DQw2 extended haplotype in Sar-

dinian celiac disease patients: further evidence that disease susceptibility

is conferred by DQ A1*0501, B1*0201. Tissue Antigens 1992;39:78-83.

7. Collin P, Reunala T, Pukkala E, Laippala P, Keyrilainen O, Pasternack A.

Coeliac disease–associated disorders and survival. Gut 1994;35:1215-8.

8. Counsell CE, Taha A, Ruddell WS. Coeliac disease and autoimmune thy-

roid disease. Gut 1994;35:844-6.

9. Velluzzi F, Caradonna A, Boy MF, Pinna MA, Cabula R, Lai MA, et al.

Thyroid and celiac disease: clinical, serological, and echographic study.

Am J Gastroenterol 1998;93:976-9.

10. Fasano A. Systemic autoimmune disorders in celiac disease. Curr Opin

Gastroenterol 2006;22:674-9.

Meloni et al

July 2009 ORIGINAL ARTICLES

11. Sategna Guidetti C, Solerio E, Scaglione N, Aimo G, Mengozzi G. Dura-

tion of gluten exposure in adult coeliac disease does not correlate with

the risk for autoimmune disorders. Gut 2001;49:502-5.

12. Ansaldi N, Palmas T, Corrias A, Barbato M, D’Altiglia MR,

Campanozzi A, et al. Autoimmune thyroid disease and celiac disease

in children. J Pediatr Gastroenterol Nutr 2003;37:63-6.

13. Viljamaa M, Kaukinen K, Huhtala H, Kyronpalo S, Rasmussen M,

Collin P. Coeliac disease, autoimmune diseases and gluten exposure.

Scand J Gastroenterol 2005;40:437-43.

14. Ventura A, Magazzu G, Greco L. Duration of exposure to gluten and risk

for autoimmune disorders in patients with celiac disease: SIGEP Study

Group for Autoimmune Disorders in Celiac Disease. Gastroenterology

1999;117:297-303.

15. Ventura A, Neri E, Ughi C, Leopaldi A, Citta A, Not T. Gluten-depen-

dent diabetes-related and thyroid-related autoantibodies in patients

with celiac disease. J Pediatr 2000;137:263-5.

16. Walker-Smith J, Guandalini S, Schmitz J, Shmerling DH, Visakorpi J.

Revised criteria for diagnosis of coeliac disease: Report of Working

Group of European Society of Paediatric Gastroenterology and Nutri-

tion. Arch Dis Child 1990;65:909-11.

17. Fasano A. Clinical presentation of celiac disease in the pediatric popula-

tion. Gastroenterology 2005;128:S68-73.

18. Clemente MG, Musu MP, Troncone R, Volta U, Congia M, Ciacci C,

et al. Enterocyte actin autoantibody detection: a new diagnostic tool in

celiac disease diagnosis: results of a multicenter study. Am J Gastroen-

terol 2004;99:1551-6.

19. Marsh MN, Crowe PT. Morphology of the mucosal lesion in gluten sen-

sitivity. Baillieres Clin Gastroenterol 1995;9:273-93.

20. Oberhuber G. Histopathology of celiac disease. Biomed Pharmacother

2000;54:368-72.

21. Sostre S, Reyes MM. Sonographic diagnosis and grading of Hashimoto’s

thyroiditis. J Endocrinol Invest 1991;14:115-21.

Prevalence of Autoimmune Thyroiditis in Children with Celiac D

22. Marcocci C, Vitti P, Cetani F, Catalano F, Concetti R, Pinchera A. Thy-

roid ultrasonography helps to identify patients with diffuse lymphocytic

thyroiditis who are prone to develop hypothyroidism. J Clin Endocrinol

Metab 1991;72:209-13.

23. Surks MI, Chopra IJ, Mariash CN, Nicoloff JT, Solomon DH. American

Thyroid Association guidelines for use of laboratory tests in thyroid dis-

orders. JAMA 1990;263:1529-32.

24. Loviselli A, Velluzzi F, Mossa P, Cambosu MA, Secci G, Atzeni F, et al.

The Sardinian Autoimmunity Study: 3, studies on circulating antithy-

roid antibodies in Sardinian schoolchildren: relationship to goiter prev-

alence and thyroid function. Thyroid 2001;11:849-57.

25. Sategna-Guidetti C, Volta U, Ciacci C, Usai P, Carlino A, De

Franceschi L, et al. Prevalence of thyroid disorders in untreated adult ce-

liac disease patients and effect of gluten withdrawal: an Italian multicen-

ter study. Am J Gastroenterol 2001;96:751-7.

26. Chiovato L, Lapi P, Fiore E, Tonacchera M, Pinchera A. Thyroid auto-

immunity and female gender. J Endocrinol Invest 1993;16:384-91.

27. Demirbilek H, Kandemir N, Gonc EN, Ozon A, Alikasifoglu A,

Yordam N. Hashimoto’s thyroiditis in children and adolescents: a retro-

spective study on clinical, epidemiological and laboratory properties of

the disease. J Pediatr Endocrinol Metab 2007;20:1199-205.

28. Meloni GF, Tomasi PA, Bertoncelli A, Fanciulli G, Delitala G, Meloni T.

Prevalence of silent celiac disease in patients with autoimmune thyroid-

itis from Northern Sardinia. J Endocrinol Invest 2001;24:298-302.

29. Toscano V, Conti FG, Anastasi E, Mariani P, Tiberti C, Poggi M, et al.

Importance of gluten in the induction of endocrine autoantibodies

and organ dysfunction in adolescent celiac patients. Am J Gastroenterol

2000;95:1742-8.

30. Prinzis AGM, Cambuli VM, Concas D, Foschini ML, Marras V,

Carta D, Loche S, Mariotti S. The female sex predominance in Hashi-

moto’s thyroiditis is a pubertal phenomenon. J Endocrinol Invest

2007;30:97.

isease and Effect of Gluten Withdrawal 55

THE JOURNAL OF PEDIATRICS � www.jpeds.com Vol. 155, No. 1

Table II. Clinical data at disease onset of 34 pediatricpatients with CD with AT

Patient SexCD

onsetAT

onsetAbTPOU/mL

AbTgIU/mL Ultrasonography (Sostre grade)

1* F 5 5 1000 327 32† F 7 7 440 250 23 F 7 7 970 840 34 F 7 7 555 1500 25 F 8 8 1000 800 2

6† F 10 10 250 600 17 F 11 11 900 600 2

8† F 11 11 300 280 29 F 12 12 140 160 1

10 F 12 12 4494 314 311 M 15 15 615 215 312 F 2 7 1000 370 313 F 2 7 100 288 114 F 6 7 1000 100 215 F 5 8 100 1500 116 M 4 9 550 150 217 F 5 9 55 360 318 F 7 9 4000 6400 419 F 8 9 260 380 220 F 3 10 1900 350 121 F 4 10 1926 340 222 F 6 10 420 600 2

23* F 7 10 5900 225 324* F 1 11 2600 230 325 F 1 12 1000 730 226 M 6 12 255 67 127 F 6 12 1000 50 328 M 12 13 3400 1200 229 F 10 14 110 480 230 F 6 14 500 314 231 M 13 14 170 22 232 M 5 15 235 222 233 F 7 16 250 100 234 F 15 16 430 250 2

The first column shows the patients with CD with AT numbered from 1 to 34; the second showsthe sex; the third and fourth show the age at CD and at AT onset in years; the fifth and sixthshow the titers of circulating thyroid autoantibodies; the seventh shows the thyroid echo-graphic pattern. Normal ranges: AbTPO <50 IU/mL, AbTG <100 IU/mL.*Patients 1, 23, 24: autoimmune hypothyroidism after 1, 7, and 11 years of GFD, respectively.†Patients 2, 6, and 8: autoimmune hypothyroidism at onset of CD.

55.e1 Meloni et al