Characteristics of Adult Celiac Disease in the USA: Results of a National Survey

Transcription

1 THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 96, No. 1, by Am. Coll. of Gastroenterology ISSN /01/$20.00 Published by Elsevier Science Inc. PII S (00) Characteristics of Adult Celiac Disease in the USA: Results of a National Survey Peter H. R. Green, M.D., Stavros N. Stavropoulos, M.D., Sofia G. Panagi, M.D., Susan L. Goldstein, M.A., Donald J. McMahon, Ph.D., Habibul Absan, M.D., M.Sc., and Alfred I. Neugut, M.D., Ph.D. Department of Medicine, Columbia University College of Physicians and Surgeons, and School of Public Health, Columbia University, New York, New York OBJECTIVE: The clinical spectrum of adults with celiac disease in the United States, where the disease is considered rare, is not known. We sought this information by distributing a survey. METHODS: A questionnaire was distributed by way of a celiac newsletter, directly to celiac support groups, and through the Internet. RESULTS: Respondents (1612) were from all United States except one. Seventy-five percent (1138) were biopsy proven. Women predominated (2.9:1). The majority of respondents were diagnosed in their fourth to sixth decades. Symptoms were present a mean of 11 yr before diagnosis. Diarrhea was present in 85%. Diagnosis was considered prompt by only 52% and 31% had consulted two or more gastroenterologists. Improved quality of life after diagnosis was reported by 77%. Those diagnosed at age 60 yr also reported improved quality of life. Five respondents had small intestinal malignancies (carcinoma 2, lymphoma 3) accounting for a relative risk of 300 (60 876) for the development of lymphoma and 67 (7 240) for adenocarcinoma. CONCLUSIONS: Patients with celiac disease in the United States have a long duration of symptoms and consider their diagnosis delayed. Improved quality of life after diagnosis is common. An increased risk of developing small intestine malignancies is present. (Am J Gastroenterol 2001;96: by Am. Coll. of Gastroenterology) INTRODUCTION Adult celiac disease, or gluten-sensitive enteropathy, is considered rare in the United States (1). On the basis of data from Olmsted County, Minnesota, the clinical prevalence of celiac disease in 1991 was considered to be 1 in 4600 of the population (1). In Europe, where the disease is considered common, serological screening studies have demonstrated the prevalence of celiac disease to be between 1 in 250 and 1 in 300 of the general population (2, 3). In Northern Ireland, the prevalence is even greater, 1 in 122 (4). A recent study from Baltimore that screened healthy blood donors for antiendomysial antibodies revealed a prevalence rate of 1 in 250 (5), similar to the European experience. The reason for the lack of recognition and diagnosis of celiac disease in the United States is unclear (6 8). In view of the lack of current data on the clinical spectrum of celiac disease in the United States, we used a questionnaire to obtain clinical information from a large number of adults with celiac disease. We also attempted to determine from this questionnaire the impact of the low rate of recognition of celiac disease in the United States. MATERIALS AND METHODS An outline of the questionnaire is shown in Table 1. The responses were in multiple choice form. Patients were asked to volunteer their name and address and to indicate whether they could be contacted for further information. The study was approved by the Institutional Review Board of Columbia Presbyterian Medical Center. The questionnaire was initially tested on a group of 15 patients with celiac disease and subsequently modified to ensure the clarity of the questions and expand the available selection of possible answers. The questionnaire was distributed by: 1) inclusion in a celiac newsletter with nationwide distribution (Gluten Free Living, Hastings-on-Hudson, NY); 2), to 1500 members of an Internet discussion group; and 3) direct mail to members of celiac support groups in New York, New Jersey, Michigan, and California. Responses were collected between October 1996 and October Data were entered in a computer database (Excel Version 7, Microsoft Corporation, Redmond, WA) and analyzed by two different investigators. Double entry was used to ensure accuracy of the data. Respondents who reported having a lymphoma or small bowel adenocarcinoma were contacted by telephone to obtain more information. Statistical analyses comparing patients diagnosed at age 60 yr to those diagnosed at age 60 yr were carried out using SAS (Cary, NC). All comparisons, except those listed below, were adjusted for gender using Maentel-Haenszel odds ratios or linear regression for discrete or continuous variables, respectively. For dermatitis herpetiformis, Sjögren s syndrome, thyroid disease, aphthae, and irritable bowel syndrome, unconditional logistic regression was used

2 AJG January, 2001 Celiac Disease in the USA 127 Table 1. Outline of Questionnaire Demographic, personal and medical data Gender and menopausal status Current age, height and weight Presence of diseases associated with celiac disease Previous bone fractures Presence of a diagnosis of lymphoma or other malignancies and timing of this diagnosis in relation to the diagnosis of celiac disease Family history of celiac disease Prediagnosis data Symptoms Duration of symptoms before diagnosis Diagnosis of celiac disease in childhood Number of physicians consulted before diagnosis Diagnosis data Age, height and weight at diagnosis Method of diagnosis Who made the diagnosis Patient s perspective on diagnosis and treatment Was the diagnosis prompt? Was the diagnosing physician knowledgeable about diagnosis, treatment, and management? Follow-up management Whether the patient had a follow-up biopsy while on a glutenfree diet Whether bone mineral density was measured Whether family members had been screened for celiac disease The gluten-free diet Adherence to the diet Intentional dietary lapses Sources of patient information about celiac diease and the diet Patient s rating of his/her quality of life Before diagnosis bad, fair, good, or excellent After diagnosis same, improved, or worse to adjust for age in addition to gender. Finally, for hip fractures, the comparison was restricted to patients of age 60 yr at the time of the survey (i.e., only the elderly from the early and late diagnosis groups were compared) and unconditional logistic regression was used to adjust for age and gender. Relative risks for small intestinal lymphoma and adenocarcinoma in celiac patients were calculated as ratios of the observed over expected cases. Expected cases were estimated using published age- and sex-specific incidence rates for these neoplasms in the U.S. population (9) and assuming as follow-up period the time between diagnosis of celiac disease and the time of the survey. The 95% confidence limits were calculated assuming a Poisson distribution (10). Table 2. Characteristics of Respondents Characteristic All Patients (n 1508) Biopsy-proven Patients (n 1138) Mean age (range) 53 (18 92) 53 (18 88) Male-to-female ratio 1:2.8 1:2.9 The characteristics of the group as a whole are shown in Table 2. Biopsy confirmation of the diagnosis had been performed in 75% (n 1138). Because biopsy is the gold standard of diagnosis (11), we concentrated on this group to ensure that the participants did in fact have the disease. Their age and gender were similar to those of the entire group (Table 2); the mean age at the time of the survey was 53 yr (range, yr) and women predominated (maleto-female ratio, 1:2.8). The respondents had been diagnosed an average of 7.8 yr before the survey. The age at diagnosis is shown graphically in Figure 1. The majority of patients were diagnosed in their fourth to sixth decade. Women predominated in all age groups except in the very oldest group; however, there were only three patients in this group (80 89 yr). The female predominance decreased with age. Diagnosis Data pertaining to the diagnosis of celiac disease in biopsy proven respondents are presented in Table 3. The mean duration of symptoms before diagnosis was 11 yr (range, 0 70 yr). Diarrhea was present before diagnosis in 85%. Irritable bowel syndrome had been previously diagnosed in 36%. There had been a diagnosis of celiac disease in childhood in only 6% of patients; however, of these patients the majority (62%) had not remained on a gluten-free diet and were rediagnosed later as adults (mean age at rediagnosis, 42 yr). To assess patient satisfaction with their medical care we inquired about the number of physicians consulted and whether the patient considered their physician knowledgeable about the disease. Only 54% considered their diagnosis prompt. Twenty-seven percent had consulted two or more gastroenterologists, and 4% had consulted four or more before diagnosis. The diagnosing physician was considered not knowledgeable about diagnosis by 20%, about treatment by 32%, and about follow-up management by 49% of the respondents. RESULTS General Characteristics of Respondents Responses were received from 1612 individuals (approximate response rate 65%). Adults (age 18 yr) comprised 94% of the respondents and were further analyzed. The majority of respondents were from the northeastern United States (41%). However, all states but one (South Dakota) were represented. Two percent of the respondents were from Canada. Figure 1. Number of biopsy-proven respondents diagnosed in each age group (n 1138).

3 128 Green et al. AJG Vol. 96, No. 1, 2001 Table 3. Diagnosis of CD in Biopsy-Proven Celiacs Patient Characteristics Percentage Diagnosis in childhood* 6 Diarrhea before diagnosis 85 Relatives with celiac disease 19 Diagnosis considered prompt 54 Consultation of 2 gastroenterologists 27 Consultation of 4 gastroenterologists 4 Prior diagnosis of irritable bowel syndrome 36 Diagnosing MD considered knowledgeable about: Diagnosis 20 Treatment 32 Follow-up care 49 Mean duration of symptoms before diagnosis 11 yr * 62% of these patients went off the diet and were rediagnosed as adults (presumably when symptoms recurred) at a mean age of 43 yr. Sources of Information About Celiac Disease and the Gluten-free Diet Sixty-six percent of the patients were referred to a nutritionist; however, the majority of patients, 88%, claimed they obtained the bulk of their information about celiac disease from celiac patient support groups. Only 10% obtained information from their primary physician and 39% from their gastroenterologist. Follow-up Care Only 15% (167 patients) had their first-degree relatives screened for the disease and 26% (294) had bone mineral density measured. Reduced bone mineral density was reported by 58% (169 patients) of those who had this assessed. Only 40% (419) of those diagnosed at least 1 yr before the survey had a repeat small bowel biopsy on gluten-free diet. Gluten-free Diet and Quality of Life In view of the difficulty in adhering to the gluten-free diet, information concerning dietary compliance was obtained. We asked patients whether they complied with the diet all of the time, most of the time, occasionally, or never. The majority (98%) claimed compliance all or most of the time (68% all and 30% most of the time). We also inquired about intentional dietary lapses, which occurred mainly when eating out of the home, 26% at restaurants and 21% at parties. A lactose-free diet was adhered to by 22%. To assess response to the gluten-free diet, we inquired about the patients perception of their overall quality of life. Quality of life before diagnosis was rated as bad by 30%, fair 33%, good 24%, and excellent 10%. Quality of life after diagnosis was reported to be improved by 77%, unchanged by 15%, and worse by 8%. As a more objective assessment of response to the gluten-free diet we calculated body mass index. Overall, there was a mean increase in the body mass index of 18% between the time of diagnosis and the time of the survey. An increase in body mass index was observed in 76% of the patients. Associated Diseases The percentage of patients reporting conditions associated with celiac disease is shown in Table 4. Thyroid disease was Table 4. Associated Conditions in Biopsy-proven Celiac Patients Associated Condition Respondents (%) Thyroid disease 18 Dermatitis herpetiformis 9.8 Recurrent aphthae 9.0 Type I diabetes mellitus 3.3 Sjögren s syndrome 2.0 reported by 18%, dermatitis herpetiformis 9.8%, recurrent aphthae 9%, type I diabetes mellitus 3.3%, and Sjögren s syndrome 2%. Nonbiopsy-proven Group Twenty-five percent of the respondents claimed to have celiac disease, but had not had biopsy confirmation. Of this group, 62% were diagnosed by a physician, whereas 38% were self-diagnosed. It is unclear whether these patients actually had celiac disease. Although 84% reported improved quality of life after diagnosis, the mean percent increase in body mass index was 9%, significantly less than the 18% increase in body mass index observed in the biopsy-proven group (t test, p 0.001). Diagnosis in the Elderly The usefullness of diagnosing celiac disease in the elderly, who may have minimal symptoms and may find it difficult to change long-standing dietary habits, has not been established. Therefore, we analyzed patients who had been diagnosed at age 60 yr or older and compared them to those diagnosed at age 60 yr. Overall, 214 patients (19%) were diagnosed at age 60 yr. The mean age at diagnosis was 66 yr (range, yr). At the time of the survey, their mean age was 71 yr. In this older group, the female preponderance decreased; male-to-female ratio was 1:2 compared to 1:3 for those 60 yr old at diagnosis (p 0.01). Mean duration of symptoms was 13 yr. Five patients were diagnosed in childhood and rediagnosed in the elderly age range. An improved quality of life after diagnosis was reported by 71% of the elderly. Although this is significantly less than those diagnosed at age 60 yr (79%, p 0.01), it does indicate that the majority of patients believe they benefited from the diagnosis. The mean percentage increase in body mass index was 16% versus 18% in those diagnosed at age 60 yr (p 0.06). To assess whether delay in diagnosis of celiac disease (i.e., diagnosis at age 60 yr) was associated with an increased risk of hip fracture, we analyzed the prevalence of reported hip fractures among patients who were older than 60 yr at the time of the survey. Within this group, we compared patients with a celiac disease diagnosis at age 60 yr to those with a diagnosis at age 60 yr. The prevalence of previous hip fractures in these two groups was 4% and 0.5%, respectively (p 0.1, odds ratio 5.9 [0.7 49] after adjustment for age and sex).

4 AJG January, 2001 Celiac Disease in the USA 129 Table 5. Small Bowel Lymphoma and Adenocarcinoma Relative Risks in Biopsy-proven Celiac Patients Neoplasm Observed Expected RR Primary small bowel (60 876) lymphoma Small bowel adenocarcinoma (7 240) Malignancy We inquired whether respondents had malignancies known to be associated with celiac disease. Those subjects who reported non-hodgkin s lymphoma, including enteropathy associated T-cell lymphoma (EATCL), or small bowel adenocarcinoma were contacted by phone and mail to obtain further clinical information and confirm the diagnosis. Review of the pathology records confirmed the diagnosis of small bowel adenocarcinoma in two patients (0.2%) and that of non-hodgkin s lymphoma in five (0.4%). The lymphomas were EATCL (n 3), B-cell diffuse large cell involving neck nodes (n 2), and mycosis fungoides (n 1). The five patients with small intestine neoplasms (two with adenocarcinoma and three with EATCL) were analyzed further. Their mean age was 54 yr (range, yr). All patients reported symptoms in childhood that were consistent with celiac disease. Only one, however, was actually diagnosed with celiac disease as a child; this patient subsequently went off the gluten-free diet when he was thought to have grown out of the disease. Two patients had symptoms in childhood requiring diets that excluded glutencontaining foods, and the remaining two patients had abdominal pain, diarrhea, and anemia as children. In adult life, all had anemia before diagnosis, three had diarrhea, and three had a previous diagnosis of irritable bowel syndrome. Symptoms were present for a mean of 9 yr before the diagnosis of celiac disease. Diagnosis of the neoplasm was prompted by the development of bowel obstruction in three patients, refractory sprue after 16 yr on a gluten-free diet in one patient, and anemia in one patient. The malignancy was diagnosed simultaneously with celiac disease in the two patients with adenocarcinoma and subsequently to the diagnosis of celiac disease in those with EATCL (at 1, 16, and 26 yr after the diagnosis of celiac disease). The neoplasm was located in the jejunum in all patients, whereas one had lymphoma in the duodenum as well. One patient was alive at 13 yr after the diagnosis, and the others had the diagnosis 2 yr earlier. For our survey population of celiac patients we estimated the relative risk for the development of small bowel lymphoma to be 300 (60 876) and for small intestinal adenocarcinoma to be 67 (7 240) compared to the U.S. population (Table 5). DISCUSSION There is little clinical information available about celiac disease in the United States where the disease is considered rare. Given the low rate of diagnosis of adult celiac disease in the United States, the lack of a registry of celiac patients, and the relatively small populations of patients even in referral centers, we elected to proceed with a self-administered survey distributed to geographically diverse celiac groups, despite the selection bias inherent in this methodology. We attempted to reach as many patients with celiac disease as possible by contacting patients through support groups, the Internet, and a celiac disease newsletter. By these methods, we obtained information on 1138 biopsyproven adult celiac patients in the United States. The majority (41%) of respondents was from the northeastern United States, but all states but one were represented. The large number of participants in this survey and their wide geographic distribution suggests that the group may be fairly representative of patients with celiac disease in the United States. If we assume a clinical prevalence of one person with celiac disease per 4600 of the U.S. population (1), our survey may represent as many as 2% of the patients with celiac disease in the United States. The demographic characteristics of adult celiac patients in the United States mirror those observed in Europe with a peak in diagnosis in the fourth to sixth decade and a female predominance, which decreases with age (12 14). Six percent of biopsy-proven patients in our study had been diagnosed in childhood. The majority of these had not maintained a gluten-free diet and had been rediagnosed as adults in their fourth decade. It is common for compliance to the diet to be poor among adolescents because of a remission in symptoms during that age period (15). In a study from Switzerland, approximately 20% of children diagnosed with celiac disease over a period of 23 yr had a parent or child initiate reintroduction of gluten to the diet and were rediagnosed (relapse on follow-up biopsy) up to 15 yr after reintroduction of gluten. The majority had relapsed within 2 yr (16). More than 80% of patients in this survey presented with diarrhea. This is higher than series of patients from Europe (12, 17) or Canada (18). Diarrhea with or without the malabsorption syndrome is regarded as the classical presentation of celiac disease (19). In countries where the disease is considered to be common, the percent of patients presenting with diarrhea has been declining. In a study from Edinburgh only 50% presented with diarrhea (17). Other presentations have been termed atypical, silent, or subclinical (19) and include anemia, bone disease, vague GI symptoms, or abnormal blood films. The higher rate of patients with the typical or classical presentation in our study suggests decreased awareness of celiac disease and its various atypical presentations among physicians in the United States. Nearly 40% of celiac patients in our study claimed to have a previous diagnosis of irritable bowel syndrome. It is not usual practice to screen these patients for celiac disease. However, this diagnostic group represents a large percentage of patients seen in a gastroenterology practice (20) and may in fact harbor a large number of undiagnosed celiac patients in the United States.

5 130 Green et al. AJG Vol. 96, No. 1, 2001 Other investigators (21 23) have also noted a long duration of symptoms before diagnosis (mean of 11 yr in our study). In one study of 408 patients from Germany (23), the interval from onset of symptoms to the first visit to a physician was greatly surpassed by the interval from the first visit to a physician to the time of diagnosis. Thus, the long duration of symptoms was mostly due to a physician delay in reaching the diagnosis rather than a patient delay in seeking medical attention. The findings of consultations with multiple physicians and a previous diagnosis of irritable bowel syndrome in a substantial number of patients suggest that a similar situation exists in the United States. This is also reflected in the opinion of the patients that their diagnosis was not prompt and their physicians were not knowledgeable about the disease. A plausible explanation would be that physicians regard adult celiac disease as rare and fail to consider it in clinical situations other than the classical chronic diarrhea and malabsorption. An association between celiac disease and autoimmune disorders, such as type I diabetes, autoimmune thyroid disease, and Sjögren s syndrome, has been well documented in the literature (24). In our study, the percentage of celiac patients who had a diagnosis of type I diabetes and Sjögren s were 3.3% and 2%, respectively, compared to estimates of % and 3.3% reported in the literature (24, 25). The strength of the association with thyroid disease varies widely between studies (1.8 14%) (25, 26). Our high estimate of 18% is probably attributable in part to the difficulty of discriminating between autoimmune and nonautoimmune thyroid disease in a self-administered patient survey. Dermatitis herpetiformis and recurrent oral aphthae are relatively common extraintestinal manifestations of celiac disease, and were present in 9.8% and 9% of patients, percentages similar to those reported by other investigators (1, 27). Overall, the close agreement between our prevalence estimates of associated disorders and extraintestinal manifestations of celiac disease, and those reported in the literature supports the validity of our population as a representative sample of adult celiac patients. An intestinal biopsy is considered the gold standard for the diagnosis of celiac disease (11, 28). Surprisingly, 25% of respondents to our survey claiming to have celiac disease had not undergone a biopsy, although a physician had diagnosed two thirds of them. In view of the nearly 100% specificity of the antiendomysial antibody for celiac disease, it has been suggested a biopsy is not necessary for diagnosis (29). However, a baseline biopsy is important because patients may have an atypical course and require follow-up biopsies to evaluate response to gluten exclusion or persistent symptoms. We speculate that the lack of biopsy in some patients (especially relatives of patients with biopsy-diagnosed celiac disease) may reflect an empirical trial of a gluten-free diet, with or without the acquiescence of their physician. Other patients may start a gluten-free diet as a natural dietary cure for symptoms related to a functional or organic GI disorder (including undiagnosed celiac disease). These patients may comprise a significant proportion of the one third of patients in the nonbiopsy-proven group who were self-diagnosed. It is not clear whether these patients have celiac disease. Some may have the disease, as it is common in our experience for relatives of celiac patients to empirically start a gluten-free diet if symptoms develop. The nonbiopsy-proven group did not have as great an increase in body mass index on a gluten-free diet as the biopsy-proven group suggesting that many did not have the disease. Our survey confirms the increased incidence of small bowel malignancies, adenocarcinoma, and EATCL, in celiac disease (30 35). These malignancies are rare in the general population (9). Our survey obviously did not include patients who may have died from these often rapidly fatal malignancies. Therefore, our relative risks may underestimate the true risk of developing small bowel lymphoma or adenocarcinoma. The malignancies occurred in the jejunum. The patients presented with bowel obstruction, refractory sprue, and anemia. All had a long history of symptoms suggestive of celiac disease since childhood, although only one had been diagnosed as a child. The patients with EATCL had been diagnosed with celiac disease before the diagnosis of lymphoma, whereas those with adenocarcinoma had the diagnosis of celiac disease simultaneously established with the diagnosis of carcinoma, although one had an initial diagnosis in childhood. This particular patient, however, had been told he had grown out of the disease and had not remained on a gluten-free diet. All these patients probably had celiac disease since childhood. Adherence to a gluten-free diet reduces the risk of developing malignancies (35). Therefore, earlier diagnosis may have prevented the development of malignancy. We have no objective measure of how well the patients with EATCL had adhered to the diet from the time of celiac disease diagnosis to the time of EATCL diagnosis. About 20% of patients in our study were diagnosed with celiac disease at age 60 yr. Men were more prominent in this group. These patients also had a long duration of symptoms. This elderly group appeared to have benefited from being diagnosed, as the majority was able to adhere to the gluten-free diet and reported improved quality of life. There was a trend toward more hip fractures in elderly patients diagnosed at age 60 yr compared to those diagnosed at age 60 yr. Earlier diagnosis may prevent the development of osteoporosis and subsequent fractures (36). The large group of adult celiac patients in our study provides data as to the clinical spectrum of celiac disease in the United States. The majority of patients presented with diarrhea had a long duration of symptoms before diagnosis and considered the diagnosis delayed. This is probably due in part to the perception among physicians that the disease is rare. Celiac disease results in a markedly increased risk for the development of small bowel malignancies. Because a gluten-free diet may prevent this complication, early diagnosis is important.

6 AJG January, 2001 Celiac Disease in the USA 131 Reprint requests and correspondence: Peter H.R. Green, M.D., 161 Fort Washington Avenue, New York, NY Received Feb. 2, 2000; accepted July 27, 2000 REFERENCES 1. Talley NJ, Valdovinos M, Petterson TM, et al. Epidemiology of celiac sprue: A community-based study. Am J Gastroenterol 1994;89: Catassi C, Ratsch IM, Fabiani E, et al. Coeliac disease in the year 2000: Exploring the iceberg. Lancet 1994;343: Ascher H, Kristiansson B. Childhood celiac disease in Sweden. Lancet 1994;344: Johnston SD, Watson RGP, McMillan SA, et al. Prevalence of coeliac disease in Northern Ireland. Lancet 1997;350: Not T, Horvath K, Hill ID, et al. Celiac disease risk in the USA: High prevalence of antiendomysial antibodies in healthy blood donors. Scand J Gastroenterol 1998;33: Fasano A. Where have all the American celiacs gone? Acta Paediatr 1996;4: Ferguson A. Celiac disease, an eminently treatable condition, may be under diagnosed in the United States. Am J Gastroenterol 1997;92: Podolsky DK, LaMont JT. So, where are all the celiacs? Gastroenterology 1999;116: Chow JS, Chen CC, Absan H, et al. A population-based study of the incidence of malignant small bowel tumors: SEER, J Epidemiol 1996;25: Breslow NE, Day NE. Statistical methods of cancer research, vol 2. The design and analysis of cohort studies. IARC Sci Publ No. 82. Lyon, France: IARC 1987: Walker-Smith JA, Guandalini S, Schmitz J, et al. Revised critera for the diagnosis of celiac disease. Arch Dis Child 1990;65: Swinson CM, Levi AJ. Is coeliac disease underdiagnosed? Brit Med J 1980;281: Hankey GL, Holmes GKT. Coeliac disease in the elderly. Gut 1994;35: Ciacci C, Cirillo M, Sollazzo R, et al. Gender and clinical presentation in adult celiac disease. Scand J Gastroenterol 1995;30: Mayer M, Greco L, Troncone R, et al. Compliance of adolescents with coeliac disease with a gluten free diet. Gut 1991; 32: Shmerling DH, Frankx J. Childhood celiac disease: A longterm analysis of relapses in 91 patients. J Pediatr Gastroenterol Nutr 1986;5: Logan R, Tucker G, Rifkind E, et al. Changes in clinical features of coeliac disease in adults in Edinburgh and the Lothians Brit Med J 1993;286: Pare P, Douville P, Caron D, et al. Adult celiac sprue: Changes in the pattern of clinical recognition. J Clin Gastroenterol 1988;10: Ferguson A, Arranz E, O Mahony S. Clinical and pathological spectrum of coeliac disease Active, silent, latent, potential. Gut 1993;34: Everhart JE, Renault PF. Irritable bowel syndrome in officebased practice in the United States. Gastroenterology 1991; 100: Dickey W, McConnell JB. How many hospital visits does it take before celiac sprue is diagnosed? J Clin Gastroenterol 1996;23: Cooper BT. The delayed diagnosis of celiac disease. NZ Med J 1986;99: Lanisch PG, Martinez Schramm A, Petersen F, et al. Diagnostic intervals for recognizing celiac disease. Z Gastroenterol 1996;34: Collin P, Maki M. Associated disorders in celiac disease: Clinical aspects. Scand J Gastroenterol 1994;29: Collin P, Reunala T, Pukkala E, et al. Celiac disease-associated disorders and survival. Gut 1994;35: Counsell CE, Taha A, Ruddell WS. Celiac disease and autoimmune thyroid disease. Gut 1994;35: Lahteenoja H, Toivanen A, Viander M, et al. Oral mucosal changes in celiac patients on a gluten free diet. Eur J Oral Sci 1998;106: Trier JS. Diagnosis of celiac sprue. Gastroenterology 1998; 115: Valdimarsson T, Franzen L, Grodzinsky E, et al. Is small bowel biopsy necessary in adults with suspected celiac disease and IgA anti-endomysium antibodies? Dig Dis Sci 1996;41: Swinson CM, Slavin G, Coles EC, et al. Coeliac disease and malignancy. Lancet 1983;i: Selby WS, Gallagher ND. Malignancy in a 19-year experience of adult celiac disease. Dig Dis Sci 1979;24: Holmes GKT, Dunn GI, Cockel R, et al. Adenocarcinoma of the upper small bowel complicating coeliac disease. Gut 1980; 21: Wright DH. Enteropathy associated T cell lymphoma. Cancer Surv 1997;30: Cooper BT, Read AE. Celiac disease and lymphoma. Q J Med 1987;63: Holmes GKT, Prior P, Lane MR, et al. Malignancy in coeliac disease-effect of a gluten free diet. Gut 1989;30: Corazza GR, Di Sario A, Cecchetti L, et al. Bone mass and metabolism in patients with celiac disease. Gastroenterology 1995;109:122 8.