Relationship of HLA-DQ8 and Severity of Celiac Disease: Comparison of New Yorkand Parisian Cohorts

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2004;2: Relationship of HLA-DQ8 and Severity of Celiac Disease: Comparison of New Yorkand Parisian Cohorts TIMOTHY C. JOHNSON,* BEVERLY DIAMOND,* LORENZO MEMEO, HORIA NEGULESCU, ZARA HOVHANISSYAN, VIRGINIE VERKARRE, HEIDRUN ROTTERDAM, ALESSIO FASANO, SOPHIE CAILLAT ZUCMAN, # ETIENNE GROSDIDIER,** ROBERT WINCHESTER, CHRISTOPHE CELLIER,** BANA JABRI, and PETER H. R. GREEN* Departments of *Medicine, Surgical Pathology, and Pediatrics, Columbia University College of Physicians and Surgeons, New York, New York; Department of Pathology, University of Chicago, Chicago, Illinois; # Department of Immunology, Necker Faculty-University René Descartes-Paris V, Paris; **Department of Gastroenterology, European Georges Pompidou Hospital, Assistance Publique-Hôpitaux de Paris, Paris; Department of Pathology, Hopital Necker Enfants Malades, Paris, France; and Department of Pediatrics, University of Maryland, College Park, Maryland Background & Aims: Celiac disease is a polygenic disorder associated with HLA-DQ2 or HLA-DQ8, which are present in greater than 90% of patients. The disease is considered milder in the United States compared with Europe. We assessed whether differences in the frequency of HLA type may account for differences in severity of the disease by using cohorts of patients from New Yorkand Paris. Methods: HLA-DQ typing was performed on patients with celiac disease in New Yorkand Paris. Clinical and pathologic data were compared between the New Yorkand Parisian cohorts and also correlated with the different HLA types (HLA-DQ2, HLA-DQ2/-DQ8, HLA-DQ8). Results: Among these patients, the disease was milder in the New Yorkcohort compared with the Parisian cohort. There were fewer patients with a classical presentation (45% and 89%, respectively; P < 0.001) and less severe pathology (total villous atrophy, 64% and 89%, respectively; P < 0.05), and less marked intraepithelial lymphocytosis (intraepithelial leukocytes [IELs]/100 enterocytes, 48.1 and 82.5, respectively; P < ). HLA-DQ2 homozygotes were less prevalent in the New Yorkcohort compared with the Parisian cohort (59% and 79%, respectively; P 0.08). HLA-DQ8 alleles were more prevalent in the New York cohort compared with the Parisian cohort (41% and 21%, respectively; P 0.026). There was, however, no difference in the clinical or pathologic parameters of severity when we compared the groups based on HLA type. Conclusions: HLA-DQ8 alleles were increased in the New Yorkcohort of patients with celiac disease; however, this did not account for less severe manifestations of the disease. Worldwide, celiac disease is one of the most common genetically determined diseases encountered, occurring in an estimated 1 in 266 persons. 1 The disease is an inflammatory intestinal disorder caused by activation of lamina propria HLA-DQ2 or -DQ8 restricted T cells in the small intestine by gluten-containing grains (wheat, barley, and rye) with pathologic sequelae that include crypt hyperplasia, villous atrophy, and an increased number of intraepithelial lymphocytes. 2,3 Celiac disease is a polygenic disorder associated with HLA-DQ2 (DQA1*0501/DQB1*0201) or -DQ8 (DQA1*0301/DQB1*0302). HLA-DQ2 is present in 98% of celiac disease patients in Northern Europe and 92% of celiac disease patients in Southern Europe. In Southern Europe, Israel, Spain, and Finland, HLA- DQ8 is present in roughly 10% of patients with celiac disease. 4,5 Although there appear to be several susceptibility non-hla chromosomal regions, 6 9 the HLA region appears to harbor a predominant celiac disease susceptibility locus. 7,9 Celiac disease was considered rare in the United States. 10 This was thought to be due to a combination of failure of physician recognition 11,12 and possibly because the disease is milder in the United States. Recent studies have, however, shown the disease to be as common in the United States as in Europe. 13,14 Although there are data on the clinical spectrum of patients with celiac disease in the United States, there are no studies describing the HLA status of people with celiac disease in the United States. We therefore examined the HLA status in a cohort of patients with celiac disease and correlated the findings with severity of disease. In addition, we compared the findings with those of a cohort of French patients with celiac disease. Methods Study Subjects (New YorkCohort) HLA-DQ typing was performed on 44 consecutive unrelated patients who were being assessed for celiac disease at Columbia Presbyterian Medical Center in New York City 2004 by the American Gastroenterological Association /04/$30.00 PII: /S (04)

2 October 2004 HLA DQ8 AND SEVERITY OF CELIAC DISEASE 889 between January 2000 and January HLA-DQ typing also was performed on a cohort of 81 disease controls and 60 non-disease controls in New York. The patients were classified into 1 of the following 3 groups: (1) HLA-DQ2 (homozygous or heterozygous for DQ2 and without DQ8), (2) HLA-DQ2/-DQ8, and (3) HLA-DQ8 (homozygous for DQ8 and without DQ2). Six of the patients were already on a gluten-free diet. The following information was collected prospectively: (1) ethnicity, (2) age at diagnosis, (3) duration of illness, (4) presence of classic symptoms (diarrhea), and (5) positive antibodies, endomysomial antibodies. The biopsy specimens taken of patients when they were consuming a regular diet (n 38) were analyzed. The 6 patients already on a gluten-free diet were not included in the pathologic analysis. The pathologic analysis was determined by 2 pathologists who were unaware of the dietary status of the patients. The degree of villous atrophy was assessed as either severe villous atrophy that included total and subtotal villous atrophy or partial villous atrophy if the crypt-to-villous ratio was less than 1:3. 6 Only in areas where 3 or more villi were observed was the ratio assessed. Intraepithelial lymphocytes were counted per 100 epithelial cells. Epithelial cell changes such as ballooning and cuboidal cells were noted as were the number of mitotic figures in 5 different crypts; the maximum number was recorded. The study was approved by the institutional review board. The HLA typing of the subjects was performed with the DQB SSP Kit from Biotest (Denville, NJ) based on polymerase chain reaction using allele-specific primers. Positive and negative controls were included in the kit. Because of the absolute linkage disequilibrium between the DQA and DQB genes in the Caucasian population, DQA alleles were deduced from the DQB genotype. The same HLA typing kit was used in both the New York and Paris cohort. Study Subjects (Paris Cohort) Similar information was obtained from 66 unrelated adult patients with celiac disease who were assessed at Hùpital Europèen Georges Pompidou in Paris, France, between January 2000 and January HLA-DQ typing also was performed on 193 disease controls and 441 non-disease controls in Paris. Eighteen of the subjects were on a gluten-free diet. Only the biopsy specimens taken from patients when they were consuming a regular diet (n 48) were analyzed. The pathologic analysis was determined by 2 pathologists who were unaware of the dietary status of the patients. Data were extracted from a database in which all the parameters were entered prospectively. The study was approved by the institutional review board. Statistical Analysis Statistical differences within the New York and Paris populations were assessed with t test scores and the Fisher exact test. The assessment of the association of HLA-DQ status and disease severity was performed through analysis of variance testing. All of the criteria for clinical and pathologic presentation listed earlier were analyzed separately among the 3 groups in both the New York and Paris populations, as well as by combining the HLA-DQ2/HLA-DQ8 and HLA-DQ8 groups together. To show a difference between classic presentation, villous atrophy, and HLA-DQ8/non-DQ2 prevalence among the cohorts, each group would need to have 17, 44, and 300 subjects, respectively. The earlier-described criteria were calculated with a power of 0.80 and a P value of To assess the variability in the pathologic interpretation of villous atrophy between both centers, 40 pathologic slides of study subjects were coded, exchanged, and blindly read. The coefficient of variation was determined. Results Characteristics of the Study Population The clinical, pathologic, and DQ status of the patients is shown in Table 1. There were differences between the New York and Parisian cohorts. Women predominated in both groups although proportionally men were more represented in the New York cohort. In addition, the New York cohort was significantly older at diagnosis (46.2 vs yr; P ), but had a similar duration of symptoms (3.4 vs. 3.2 yr; P 0.30). However, fewer patients in the New York cohort had the more severe classic presentation with diarrhea (45.5% vs. 89.7%; P 0.001). Fewer patients in the New York cohort had more severe degrees of villous atrophy (total villous atrophy, 64% vs. 89.1%; P 0.05). The number of intraepithelial lymphocytes per 100 epithelial cells in the biopsy specimens of the Parisian cohort was greater compared with the New York cohort (82.5 vs. 48.1; P ). Fewer patients in the New York cohort were homozygous for DQ2 compared with the Paris cohort (59% vs. 79%, P 0.08), with more having at least one allele of the DQ8 (41% vs. 21%, P 0.026). We assessed the HLA-DQ type in a larger number of patients in each center (disease and non-disease controls). The HLA alleles in the New York non-disease cohort (n 60) consisted of 46% DQ2, 17% DQ8, and 38% DQ2/DQ8 individuals, while the HLA alleles in the Paris cohort (n 441) consisted of 24% DQ2, 6.5% DQ8, and 69% DQ2/DQ8 negative individuals. The HLA alleles in the New York disease control cohort (n 81) consisted of 68% DQ2, 14% DQ2/DQ8, and 19% DQ8, with 33% having at least one allele of the DQ8. The HLA alleles in the Paris disease control cohort (n 193) consisted of 95% DQ2, 3.6% DQ2/DQ8, 1.6% DQ8, with 5.2% having at least one allele of the DQ8. Analysis of the ethnicity pattern of the subjects from New York revealed the majority were of Eastern European Jewish origin (59.1% paternal, 54.5% maternal),

3 890 JOHNSON ET AL. CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 2, No. 10 Table 1. Characteristics of New York and Paris Populations Variables New York Paris P a Total number of patients Men 18 (41%) 14 (21%) 0.05 Women 26 (59%) 52 (79%) 0.05 HLA type DQ2 26 (59%) 52 (79%) 0.08 DQ2/DQ8 12 (27%) 9 (14%) 0.08 DQ8/non-DQ2 6 (14%) 5 (7%) 0.08 DQ2/DQ8 DQ8/non-DQ2 18 (41%) 14 (21%) Age at diagnosis, yr Duration of illness, yr Presence of diarrhea (%) Gluten-free diet (%) TVA (%) PVA (%) IELs (per 100 enterocytes) TVA, total villous atrophy and subtotal villous atrophy; PVA, partial villous atrophy; IELs, intraepithelial lymphocytes. a P values for the comparisons of clinical severity variables among all 3 groups were calculated using Fisher exact test, t test, and analysis of variance. followed by Irish (22.7% paternal, 18.2% maternal), and Italian (6.8% paternal, 11.4% maternal), with a minority of French, English, German, and Scottish. The ethnic origins of the Parisian cohort were French and Italian (92% maternal and paternal), and Maghred, North African (8% maternal and paternal). Comparison of Clinical Severity Among the Human Leukocyte Antigen DQ Groups When we compared the characteristics of the patients from New York according to HLA-DQ status (Table 2), there was no significant difference in the ethnic origins of the patients, age at onset of disease, presence of diarrhea, presence of endomysial antibodies, or duration of symptoms among the different HLA groups. Comparison of Pathologic Severity We examined the biopsy specimens of 38 of the patients in the New York cohort (Table 3). Total villous atrophy was present in 64% and partial villous atrophy was present in 36%. The mean intraepithelial lymphocyte (IEL) count was per 100 epithelial cells (normal, Table 2. Clinical Severity and HLA Status of New York Cohort Variables HLA-DQ2 HLA-DQ2/8 HLA-DQ8/ HLA-DQ2/DQ8 non-dq2 P a and HLA-DQ8 Number Sex Men 10 (38%) 7 (54%) 1 (20%) (44%) 0.7 Women 16 (62%) 6 (46%) 4 (80%) (56%) 0.7 Ethnicity (%) M P M P M P c M P c Eastern European Jewish Irish Italian French English German Scottish Age at diagnosis, yr Duration of illness, yr EMA positive (%) Presence of diarrhea (%) M, maternal; P, paternal; EMA, endomysial antibody. a P values for the comparisons of clinical severity variables among all 3 groups were calculated using Fisher exact test, t test, and analysis of variance. b P value between HLA-DQ2/DQ8 and HLA-DQ8 was calculated using 2, t test, and analysis of variance testing. c P values for ethnicity were calculated in combination for all ethnic groups. There was no significant difference between the HLA-DQ groups in terms of ethnicity. P b

4 October 2004 HLA DQ8 AND SEVERITY OF CELIAC DISEASE 891 Table 3. Pathologic Severity and HLA Status of the New York Cohort Variables HLA-DQ2 HLA-DQ2/DQ8 HLA-DQ8 P a HLA-DQ8 HLA-DQ2/8 and P b Number TVA (%) PVA (%) Epithelium altered (%) Mitosis (greatest per 5 crypts) c IELs (per 100 enterocytes) d TVA, total villous atrophy and subtotal villous atrophy; PVA, partial villous atrophy; IELs, intraepithelial lymphocytes. a P values for the comparisons of pathologic severity among all groups were calculated using Fisher exact test, t test, and analysis of variance testing. b P value between HLA-DQ2/8 and HLA-DQ8 were calculated using 2, t test, and analysis of variance testing. c HLA-DQ2/DQ8 group had significantly increased number of crypt mitosis compared with the HLA-DQ2 group. d HLA-DQ8 group had a trend increase in the number of IELs per 100 enterocytes compared with the HLA-DQ2 group. 30), mean number of mitoses (greatest per 5 crypts) was , and 82% had alteration in the epithelial cells. There was no significant difference in the severity of villous atrophy between the 3 HLA groups or the combined DQ8 group. There was, however, a suggestion that HLA-DQ8 had more severe pathologic changes. The HLA-DQ2/-DQ8 group had a significantly greater number of crypt mitoses (P 0.047). This difference was lost when the DQ8 groups were combined (P 0.16). In addition, there was a trend in severity favoring HLA-DQ8 in respect to the number of IELs (HLA-DQ2 vs. HLA-DQ8, 44.8 and 59.8 IELs per 100 epithelial cells, respectively; P 0.085). The trend was maintained when we compared the DQ2 group with the combined DQ8 group (HLA-DQ2 vs. HLA-DQ2/DQ8 and HLA-DQ8; 44.8 and 53.3 IELs per 100 epithelial cells, respectively; P 0.07). Comparison of Human Leukocyte Antigen DQ Status and Severity of Disease in Paris Cohort We examined the different HLA-DQ subgroups among the cohort from Paris. In respect to the clinical parameters (Table 4), the only significant difference was the absence of men in the DQ8-containing groups (DQ2, 27%; DQ8, 0%; P 0.03). In addition, there was no difference in pathologic severity between the different HLA-DQ groups (Table 5). When we compared the pathologic interpretation of the pathologists in both centers, we found a good agreement between the assessment of the degree of villous atrophy (, 0.79; 95% confidence interval, ). Table 4. Clinical Severity and HLA Status of Paris Cohort Variables HLA-DQ2 HLA-DQ2/ DQ8 HLA-DQ8 P a HLA-DQ2/ DQ8 and HLA-DQ8 Number Sex Men 14 (27%) 0 (0%) 0 (0%) (0%) 0.03 Women 38 (73%) 9 (100%) 5 (100%) (100%) 0.03 Ethnicity (%) c M P M P M P c M P c French/Italian Maghreb Age at diagnosis, yr Duration of illness, yr Presence of diarrhea (%) EMA positive (%) M, maternal; P, paternal. a P values for the comparisons of pathologic severity among all groups were calculated using the Fisher exact test, t test, and analysis of variance testing. b P value between HLA-DQ2/8 and HLA-DQ8 were calculated using the Fisher exact test, t test, and analysis of variance testing. c P values for ethnicity were calculated in combination for all ethnic groups. There was no significant difference between the DQ groups in terms of ethnicity. P b

5 892 JOHNSON ET AL. CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 2, No. 10 Table 5. Pathologic Severity and HLA Status of Paris Cohort Variables HLA-DQ2 HLA-DQ2/DQ8 HLA-DQ8 P a HLA-DQ2/DQ8 and HLA-DQ8 P b Number TVA (%) PVA (%) IELs (per 100 enterocytes) c N/A d c 1.0 TVA, total villous atrophy and subtotal villous atrophy; PVA, partial villous atrophy; N/A, not available. a P values for the comparisons of pathologic severity among all groups were calculated using Fisher exact test, t test, and analysis of variance testing. b P value between HLA-DQ2/8 and HLA-DQ8 was calculated using the Fisher exact test, t test, and analysis of variance testing. c Only one HLA-DQ2/DQ8 subject was analyzed. d There were no IELs analyzed in HLA-DQ8 subjects in the Paris cohort. Discussion Celiac disease is a genetically determined disease as shown by studies that show a sibling recurrence risk for celiac disease of 10% 15 and a concordance of 70% for identical twins. 16 The disease is considered to be polygenic. However, the major association consistently has been with the HLA status. 7 Greater than 98% of patients have either HLA-DQ2 or -DQ8, with HLA-DQ2 occurring in greater than 90%. 17 When we examined the HLA-DQ2/-DQ8 status of a series of consecutive patients seen in a celiac disease center in New York, we found an increased number of patients with at least 1 allele for HLA-DQ8 compared with previous reports in the literature, 17 as well as a cohort of French patients from Paris. Overall, 41% of the New York cohort had at least 1 HLA-DQ8 allele compared with 21% of the Paris cohort. The New York cohort had a decreased percent of patients who were homozygous for HLA-DQ2, the classic celiac disease pattern, compared with the cohort from Paris (59% compared with 79%). The increased prevalence of HLA-DQ8 in the New York celiac patient cohort compared with the Parisian cohort was confirmed by the results in the disease and non-disease control patients. We compared the clinical and pathologic severity of the disease between patients diagnosed with celiac disease in New York and Paris. The New York cohort differed from the Parisian cohort in that they were older, fewer had the classic presentation with diarrhea, and there were fewer patients with more severe degrees of villous atrophy. At least for these limited numbers of patients from New York and Paris, the disease appears to be milder in the United States compared with France. Similar findings were reported in abstract form when a cohort of patients from Italy were compared with a cohort from the United States. 18 We cannot assess if these differences are due to quantitative differences in the amount of gluten in the diet of the United States and European countries. The HLA status of patients with celiac disease determines whether isolated intestinal lymphocytes and intestinal biopsy specimens in culture react with specific immunogenic gliadin peptides, 2,19,20 indicating that the genetic make up of patients with celiac disease influences their reaction to gliadin. Therefore, we were interested in determining if the HLA status of the patients influenced the clinical or pathologic presentation of the disease. In view of the milder phenotype of the patients with celiac disease in the New York group and the greater percentage of patients with HLA-DQ8 we considered whether the presence of DQ8 was associated with a milder variety of celiac disease. We could not, however, find an influence of HLA-DQ status on the clinical severity of the disease for either the New York or Paris cohort. There were differences in the pathologic findings in the cohorts from New York and Paris. The Paris group had more patients with total villous atrophy (90%) compared with the New York group (64%). However, we could find no significant difference in the degree of villous atrophy or the number of IELs per 100 epithelial cells among the different HLA-DQ groups from each cohort. When we looked at the histologic findings of the New York cohort in greater detail there was a suggestion that there were signs of greater severity in the DQ8 group. This was manifested by increased IELs and crypt mitoses. These findings, however, were not significant. HLA-DQ2 predominates in most patients with celiac disease. 17 However, in a study of patients with celiac disease from Israel only 70% were homozygous for DQ2, 21 confirming an ethnic influence on HLA-DQ status. There have been only a few investigators who have examined the effect of HLA-DQ status on severity of disease. Clerget-Darpoux et al. 22 estimated the risk for developing celiac disease to be 6.8 times greater for those having a double dose of DQB1*0201 (DQ2 homozygotes) than for other dimer carriers. Bouguerra et al. 23 established that the presence of a double dose of the

6 October 2004 HLA DQ8 AND SEVERITY OF CELIAC DISEASE 893 DQB1*0201 allele, compared with a single dose in the presence of HLA-DQ8, increased the risk for penetrance of the disease. Greco et al. 24 examined 145 celiac disease patients and examined their HLA status and respective clinical severity of disease. The results showed no significant difference between the HLA-DQ2 homozygotes and HLA-DQ2/-DQ8 groups in terms of severity of disease. In a study by Drago et al., 18 celiac patients with HLA-DQ8 had a nonsignificant increase in total villous atrophy compared with individuals homozygous for DQ2. Similarly, we could not show a significant influence of HLA-DQ type on severity of disease. We have shown an increased rate of HLA-DQ8 carriage in a cohort of patients with celiac disease from New York. We realize that major shortcomings of this series were the small number of patients included and referral bias in both cities. However, there is no evidence that HLA-DQ8 influences the phenotype of patients with celiac disease. It is apparent that celiac disease is a complex entity with multiple genetic factors that interact with environmental factors to produce the disease. The environmental factors include gluten ingestion, 25 as well as the timing of and quantity of gluten ingestion, 26,27 breast feeding, 26 and possibly smoking. 28,29 Genes outside the HLA region may be more important in determining phenotype. The reason that celiac disease may be more silent in the United States remains obscure. References 1. Fasano A, Catassi C. Current approaches to diagnosis and treatment of celiac disease: an evolving spectrum. Gastroenterology 2001;120: van de Wal Y, Kooy YM, van Veelen PA, Pena SA, Mearin LM, Molberg O, Lundin KE, Sollid LM, Mutis T, Benckhuijsen WE, Drijfhout JW, Koning F. Small intestinal T cells of celiac disease patients recognize a natural pepsin fragment of gliadin. Proc Natl Acad Sci U S A 1998;95: Sollid LM. Coeliac disease: dissecting a complex inflammatory disorder. Nat Rev Immunol 2002;2: Polvi A, Arranz E, Fernandez-Arquero M, Collin P, Maki M, Sanz A, Calvo C, Maluenda C, Westman P, de la Concha EG, Partanen J. HLA-DQ2-negative celiac disease in Finland and Spain. Hum Immunol 1998;59: Ciclitira PJ, King AL, Fraser JS. AGA technical review on celiac sprue. American Gastroenterological Association. Gastroenterology 2001;120: Greco L, Corazza G, Babron MC, Clot F, Fulchignoni-Lataud MC, Percopo S, Zavattari P, Bouguerra F, Dib C, Tosi R, Troncone R, Ventura A, Mantavoni W, Magazzu G, Gatti R, Lazzari R, Giunta A, Perri F, Iacono G, Cardi E, de Virgiliis S, Cataldo F, De Angelis G, Musumeci S, Ferrari R, Balli F, Bardella M-T, Volya V, Catassi C, Torre G, Eliaou J-F, Serre J-C, Clerget-Darpoux F. Genome search in celiac disease. Am J Hum Genet 1998;62: Liu J, Juo SH, Holopainen P, Terwilliger J, Tong X, Grunn A, Brito M, Green P, Mustalahti K, Maki M, Gilliam TC, Partanen J. Genome wide linkage analysis of celiac disease in Finnish families. Am J Hum Genet 2002;70: Naluai AT, Nilsson S, Gudjonsdottir AH, Louka AS, Ascher H, Ek J, Hallberg B, Samuelsson L, Kristiansson B, Martinsson T, Nerman O, Sollid LM, Wahlstrom J. Genome-wide linkage analysis of Scandinavian affected sib-pairs supports presence of susceptibility loci for celiac disease on chromosomes 5 and 11. Eur J Hum Genet 2001;9: King AL, Ciclitira PJ. Celiac disease: strongly heritable, oligogenic, but genetically complex. Mol Genet Metab 2000;71: Talley NJ, Valdovinos M, Petterson TM, Carpenter HA, Melton LJ 3rd. Epidemiology of celiac sprue: a community-based study. Am J Gastroenterol 1994;89: Fasano A. Where have all the American celiacs gone? Acta Paediatr Suppl 1996;412: Ferguson A. Celiac disease, an eminently treatable condition, may be underdiagnosed in the United States. Am J Gastroenterol 1997;92: Not T, Horvath K, Hill ID, Partanen J, Hammed A, Magazzu G, Fasano A. Celiac disease risk in the USA: high prevalence of antiendomysium antibodies in healthy blood donors. Scand J Gastroenterol 1998;33: Fasano A, Berti I, Gerarduzzi T, Not T, Colletti RB, Drago S, Elitsur Y, Green PH, Guandalini S, Hill ID, Pietzak M, Ventura A, Thorpe M, Kryszak D, Fornaroli F, Wasserman SS, Murray JA, Horvath K. Prevalence of celiac disease in at-risk and not-at-risk groups in the United States: a large multicenter study. Arch Intern Med 2003;163: Bevan S, Popat S, Braegger CP, Busch A, O Donoghue D, Falth- Magnusson K, Ferguson A, Godkin A, Hogberg L, Holmes G, Hosie KB, Howdle PD, Jenkins H, Jewell D, Johnston S, Kennedy NP, Kerr G, Kumar P, Logan RF, Love AH, Marsh M, Mulder CJ, Sjoberg K, Stenhammer L, Walker-Smith J, Marossy A-M, Houlston RS. Contribution of the MHC region to the familial risk of coeliac disease. J Med Genet 1999;36: Greco L, Romino R, Coto I, Di Cosmo N, Percopo S, Maglio M, Paparo F, Gasperi V, Limongelli MG, Cotichini R, D Agate C, Tinto N, Sacchetti L, Tosi R, Stazi MA. The first large population based twin study of coeliac disease. Gut 2002;50: Sollid LM. Molecular basis of celiac disease. Annu Rev Immunol 2000;18: Drago S, Fornaroli F, Sdepanian V, Di Pierro M, Kryzsak D, Sewell P, Madeiros L, Horvath K, Piezak M, Catassi C, De Angelis G, Guandalini S, Hakall A, Fasano A. Comparison of intestinal tissue damage and HLA haplotype between American and Italian celiac disease patients. Gastroenterology 2002;122:A van de Wal Y, Kooy Y, van Veelen P, Pena S, Mearin L, Papadopoulos G, Koning F. Selective deamidation by tissue transglutaminase strongly enhances gliadin-specific T cell reactivity. J Immunol 1998;161: Mazzarella G, Maglio M, Paparo F, Nardone G, Stefanile R, Greco L, van de Wal Y, Kooy Y, Koning F, Auricchio S, Troncone R. An immunodominant DQ8 restricted gliadin peptide activates small intestinal immune response in in vitro cultured mucosa from HLA-DQ8 positive but not HLA-DQ8 negative coeliac patients. Gut 2003;52: Tighe MR, Hall MA, Ashkenazi A, Siegler E, Lanchbury JS, Ciclitira PJ. Celiac disease among Ashkenazi Jews from Israel. A study of the HLA class II alleles and their associations with disease susceptibility. Hum Immunol 1993;38: Clerget-Darpoux F, Bouguerra F, Kastally R, Semana G, Babron MC, Debbabi A, Bennaceur B, Eliaou JF. High risk genotypes for celiac disease. C R Acad Sci III 1994;317: Bouguerra F, Babron MC, Eliaou JF, Debbabi A, Clot J, Khaldi F, Greco L, Clerget-Darpoux F. Synergistic effect of two HLA heterodimers in the susceptibility to celiac disease in Tunisia. Genet Epidemiol 1997;14: Greco L, Percopo S, Clot F, Bouguerra F, Babron MC, Eliaou JF, Franzese C, Troncone R, Clerget-Darpoux F. Lack of correlation

7 894 JOHNSON ET AL. CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 2, No. 10 between genotype and phenotype in celiac disease. J Pediatr Gastroenterol Nutr 1998;26: Marsh MN. Gluten, major histocompatibility complex, and the small intestine. A molecular and immunobiologic approach to the spectrum of gluten sensitivity ( celiac sprue ). Gastroenterology 1992;102: Ivarsson A, Hernell O, Stenlund H, Persson LA. Breast-feeding protects against celiac disease. Am J Clin Nutr 2002;75: Hernell O, Ivarsson A, Persson LA. Coeliac disease: effect of early feeding on the incidence of the disease. Early Hum Dev 2001;65(Suppl):S153 S Snook JA, Dwyer L, Lee-Elliott C, Khan S, Wheeler DW, Nicholas DS. Adult coeliac disease and cigarette smoking. Gut 1996;39: Vazquez H, Smecuol E, Flores D, Mazure R, Pedreira S, Niveloni S, Maurino E, Bai JC. Relation between cigarette smoking and celiac disease: evidence from a case-control study. Am J Gastroenterol 2001;96: Address requests for reprints to: Peter Green, Columbia University, Department of Gastroenterology, 161 Fort Washington Avenue, Room 645, New York, New York pg11@columbia.edu; fax: (212)