Gastroenterology
Volume 123, Issue 5 , Pages 1726-1729, November 2002

Cancer risk in celiac disease☆

Division of Gastroenterology and Hepatology, Mayo Clinic and Foundation, Rochester, Minnesota

Article Outline

Abstract 

GASTROENTEROLOGY 2002;123:1726-1729

 

See article on page 1428.

Although malignancy occurring in the setting of celiac disease (CD) has been recognized for more than 60 years,1 the occurrence of cancers in association with dermatitis herpetiformis (DH) has been described more recently.2 There has been uncertainty as to the magnitude of cancer risk, particularly with respect to non-Hodgkin lymphoma (NHL). Several questions about cancer risk in CD or DH can be posed. First, what is the risk of malignant lymphoma in patients with CD and DH? Second, are other malignancies seen more commonly in patients with CD and DH? Finally, does a gluten-free diet affect cancer risk and mortality in these conditions?

The study by Askling et al.3 in this issue of GASTROENTEROLOGY provides useful additional information with regard to all 3 questions. The authors used the resources of the Swedish inpatient register, which contains information on all Swedish inpatient care since 1964. For each hospitalization, the registry includes the patient's national registration number, hospitalization dates, and discharge diagnoses. Patients eligible for the study included all individuals discharged at least once with a diagnosis of CD or DH between 1964 and 1994. Because the study included only those patients who required hospitalization, this was not a true population-based study. As the authors have noted, less severe cases of CD and DH that did not require hospitalization would not have been included in the study. Follow-up began on the date of discharge from the initial hospitalization and continued until the patient's death or emigration, or until the end of the follow-up period. Follow-up was conducted through linkage with several national registries, including those of cancer, causes of death, population, and population changes. Cancers diagnosed before entry, occurring within 1 year of initial hospitalization, or incidentally detected during autopsy were excluded. The relative risk of malignancy was estimated by comparing the observed number of cancers in the study patients with the numbers expected, based on national incidence rates of cancer (standardized incidence ratio [SIR]). Although the validity of a diagnostic code for CD or DH in the inpatient registry was not directly confirmed by medical record review, these codes have been found to be 85%–90% accurate for other chronic conditions.

Askling et al.3 identified more than 12,000 patients hospitalized with CD and/or DH (the 2 largest studies before this included a total of approximately 1700 patients4, 5). Among approximately 11,000 patients with CD, a total of 249 cancers were identified, of which 44 were malignant lymphomas (18% of total, 4% expected). The relative risk of cancer compared with that expected was 1.3, or 30% higher, and this was statistically significant (95% confidence interval [CI], 1.2–1.5). Importantly, the risk of lymphoma was increased 6-fold (SIR, 5.9; 95% CI, 4.3–7.9). Patients with CD were at a significantly increased risk of several other types of malignancies, including oropharyngeal carcinoma (SIR, 2.3), esophageal carcinoma (SIR, 4.2), and small intestinal adenocarcinomas (SIR, 10). There were small but significantly increased risks of colorectal carcinoma (SIR, 1.5) and primary liver cancer (SIR, 2.7). A total of 135 cancers were identified among 1354 patients with DH. Patients with DH were found to have a 2-fold increased risk of lymphoma (SIR = 1.9), but this elevation was not statistically significant (95% CI, 0.8–3.9). Furthermore, they were not at increased risk of other types of malignancies.

Patients first hospitalized with CD during childhood (younger than 10 years of age) were not found to have a significantly increased risk of lymphoma (SIR, 1.4; 95% CI, 0.2–5.0), whereas those first hospitalized with CD after 20 years of age had a considerably higher risk (SIR, 7.0; 95% CI, 5.0–9.5). The risk among those diagnosed in adulthood declined during follow-up, reaching unity after 15 years or more. It was also noted that the relative risk of lymphoma decreased over successive calendar years, from a 12-fold increase during the 1970s to only a 3-fold risk at the end of the study period.

How can we integrate the results of this study into our understanding? First, we address the question, “what is the risk of lymphoma in patients with CD and DH?” The importance of this question has been highlighted by recent studies suggesting that the true prevalence of CD may be as high as 3–4 cases per 1000 persons,6 even in the United States.7 Studies of prognosis must use a representative and well-defined sample of patients observed starting at a similar point in the course of their disease.8 Population-based studies of incidence cohorts are especially valuable in this respect because they encompass a cohort of patients manifesting the full spectrum of disease severity followed up from time of diagnosis. The only true population-based study of mortality in patients with CD is from Edinburgh and the Lothian region of Scotland.5 Among 653 patients, Logan et al.5 reported an overall mortality almost twice as high as that of the general population, including 17 deaths from lymphoproliferative diseases (30-fold increase in risk). Studies of several referral-based cohorts have shown an increased risk of malignant lymphoma, but with wide variation in relative risk.4, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18 A referral-based cohort study of more than 1000 patients from Italy confirmed that NHL is the leading cause of mortality in CD, and the risk of death caused by NHL was almost 70-times higher than expected.4 A study from Birmingham, England, reported a significant excess of NHL, with a relative risk of 42.7.14 In a large, multicenter, case-control study of Italian patients, the odds ratio for NHL associated with CD was 3.1 (95% CI, 1.3–7.6).19 The current study by Askling et al.3 reports a 6-fold increased risk of lymphoma.

How can we explain the marked variation in reported lymphoma risk? Ascertainment bias may explain some of these differences. Such bias occurs when CD cases that have long been clinically silent are found because of the diagnosis of lymphoma, whereas uncomplicated CD cases remain undiagnosed. If lymphoma risk is then calculated by using cancers from both the visible and the submerged parts of the “celiac iceberg,” but excluding clinically silent cases of CD from the denominator, the true lymphoma risk may be overestimated. To reduce the potential for biased results, the authors of the present study excluded cancers diagnosed before, in association with, or up to 1 year after the diagnosis of CD from the calculation of lymphoma risk.3 Secondly, the lower relative risks seen in recent studies may be caused by changing secular trends in CD-associated lymphoma risk. Earlier diagnosis of CD through more widespread use of sensitive serologic tests, greater diagnostic awareness among physicians, and greater adherence to gluten-free diets14 may explain the decreasing risk of lymphoma seen over time in the present study.

The study by Askling et al.,3 although not truly population-based, is important because it is considerably larger than previous studies of malignancy in CD. It adds weight to the evolving evidence that suggests that CD-associated lymphoma risk may not be as high as previously noted. Although the magnitude of risk remains debatable, even the lower estimates of risk point to an unequivocal association between celiac disease and malignant lymphoma. The 2-fold risk of lymphoma among patients with DH, reported by Askling et al.,3 is also considerably lower than the previously reported risk.20, 21

We return to the second question, “are other malignancies seen more commonly in patients with CD and DH?” Four case series9, 11, 13, 14 and the single population-based study5 have all confirmed an increased risk of oropharyngeal squamous carcinoma, esophageal squamous carcinoma, and adenocarcinoma of the small intestine in patients with CD. In contrast, a large cohort study from Italy4 reported no excess of nonhematologic malignancies. CD may predispose a patient to malignancy of the upper gastrointestinal tract as a result of nutritional deficiency. The pharyngeal and esophageal tumors that evolve in the setting of CD are frequently squamous in nature. The association between iron-deficiency anemia, esophageal webs, and esophageal squamous carcinoma of the esophagus has been described.22 Four case series reporting malignancy and mortality in patients with DH found no increase in nonhematologic cancers.2, 21, 23, 24 The present study by Askling et al.3 also found significantly increased risks of the aforementioned malignancies in patients with CD but not in those with DH.3 In addition, a previously unrecognized increased risk of colorectal cancer and primary liver cancers, and a reduced risk of breast cancer, were noted in patients with CD.3 Although the weight of evidence confirms the increased risk of oropharyngeal, esophageal, and small intestinal carcinomas in CD but not in DH, further studies will be required to substantiate the newly reported findings with regard to colorectal, liver, and breast cancer.

Finally, we come back to the question, “does a gluten-free diet affect cancer risk and mortality in CD and DH?” The main cause of mortality in CD, particularly within 5 years of diagnosis, is malignancy.4 Three case series4, 14, 15 and the Scottish population-based study5 provide direct and indirect evidence that prompt diagnosis and institution of a gluten-free diet lead to decreased cancer incidence and increased survival in CD. In the Scottish study, patients with CD who were diagnosed at younger than the age of 15 years had a mortality similar to that expected, and the excess mortality was restricted to those diagnosed after age 15 (standard mortality ratio, 1.9).5 Among the 210 patients with CD in the Birmingham study, the relative risks of cancer overall, oropharyngeal cancer, and esophageal cancer were not significantly elevated among those on a strict gluten-free diet, in contrast with those who were on a reduced-gluten or normal diet.14 Although the risk of NHL remained significantly elevated among those on a gluten-free diet (SIR, 16.7), the magnitude of risk was lower than that seen in noncompliant patients (SIR, 77.8).14 A Finnish study reported cancer incidence and mortality rates similar to the general population in a cohort of patients with CD who were 83% compliant with a gluten-free diet.15 Finally, the multicenter Italian cohort study showed that overall mortality among those adherent to a gluten-free diet was similar to, if not lower than, expected, whereas it was 6-times higher than expected among noncompliant patients with CD.4 In contrast, a single published case-control series,25 as well as a case series recently presented in abstract form,26 did not show a difference in mortality, despite adherence to a gluten-free diet. However, in the latter study, follow-up after initiation of a gluten-free diet was limited (mean, 5.9 years).26 The present study from Sweden provides more indirect evidence of the benefit of a gluten-free diet.3 Individuals first hospitalized with CD younger than 10 years of age were not found to have a significantly increased risk of malignancy, whereas those hospitalized with CD who were older than 20 years of age had a considerably higher risk. In addition, the authors observed a decreasing relative risk of malignancy over successive calendar years. The most likely explanation for this finding is the increasing recognition of silent and mild cases of CD caused by aforementioned reasons. Overall, the evidence suggests that early diagnosis is associated with reduced cancer incidence and enhanced survival in CD. Presumably, much of this benefit results from prompt institution of a gluten-free diet. Neither prior studies nor the study by Askling et al.3 provide sufficient evidence to draw conclusions regarding survival benefit for patients with DH diagnosed at an early age.

In summary, it is clear that patients with CD and DH are at increased risk of lymphoma, but the risk may be lower than previously estimated. Patients with CD, but not those with DH, are at increased risk of oropharyngeal, esophageal, and small bowel carcinoma. CD may also confer an increased risk of colorectal cancer and primary liver cancer and a decreased risk of breast cancer, but these findings need further study. Survival in CD is improved by early diagnosis and strict adherence to a gluten-free diet.

Back to Article Outline

References 

  1. Fairly N, Mackie F. The clinical and biochemical syndrome in lymphoma and allied diseases involving the mesenteric lymph glands. BMJ. 1937;1:375–380
  2. Leonard JN, Tucker WF, Fry JS, Coulter CA, Boylston AW, McMinn RM, et al. Increased incidence of malignancy in dermatitis herpetiformis. Br Med J Clin Res Ed. 1983;286:16–18
  3. Askling J, Linet M, Gridley G, Halstensen T, Ekstrom K, Ekbom A. Cancer incidence in a population-based cohort of individuals hospitalized with celiac disease or dermatitis herpetiformis in Sweden. Gastroenterology. 2002;123:1428–1435
  4. Corrao G, Corazza GR, Bagnardi V, Brusco G, Ciacci C, Cottone M, et al. Mortality in patients with coeliac disease and their relatives: a cohort study. Lancet. 2001;358:356–361
  5. Logan RF, Rifkind EA, Turner ID, Ferguson A. Mortality in celiac disease. Gastroenterology. 1989;97:265–271
  6. Catassi C, Ratsch IM, Fabiani E, Rossini M, Bordicchia F, Candela F, et al. Coeliac disease in the year 2000: exploring the iceberg. Lancet. 1994;343:200–203
  7. Not T, Horvath K, Hill ID, Partanen J, Hammed A, Magazzu G, et al. Celiac disease risk in the USA: high prevalence of antiendomysium antibodies in healthy blood donors. Scand J Gastroenterol. 1998;33:494–498
  8. Laupacis A, Wells G, Richardson WS, Tugwell P. Users' guides to the medical literature. V. How to use an article about prognosis. Evidence-Based Medicine Working Group. JAMA. 1994;272:223–234
  9. Harris OD, Cooke WT, Thompson H, Waterhouse JA. Malignancy in adult coeliac disease and idiopathic steatorrhoea. Am J Med. 1967;42:899–912
  10. Brandt L, Hagander B, Norden A, Stenstam M. Lymphoma of the small intestine in adult coeliac disease. Acta Med Scand. 1978;204:467–470
  11. Selby WS, Gallagher ND. Malignancy in a 19-year experience of adult celiac disease. Dig Dis Sci. 1979;24:684–688
  12. Cooper BT, Holmes GK, Cooke WT. Lymphoma risk in coeliac disease of later life. Digestion. 1982;23:89–92
  13. Swinson CM, Slavin G, Coles EC, Booth CC. Coeliac disease and malignancy. Lancet. 1983;1:111–115
  14. Holmes GK, Prior P, Lane MR, Pope D, Allan RN. Malignancy in coeliac disease—effect of a gluten free diet. Gut. 1989;30:333–338
  15. Collin P, Reunala T, Pukkala E, Laippala P, Keyrilainen O, Pasternack A. Coeliac disease—associated disorders and survival. Gut. 1994;35:1215–1218
  16. Ferguson A, Kingstone K. Coeliac disease and malignancies. Acta Paediatr. 1996;412(suppl):78–81
  17. Cottone M, Termini A, Oliva L, Magliocco A, Marrone C, Orlando A, et al. Mortality and causes of death in celiac disease in a Mediterranean area. Dig Dis Sci. 1999;44:2538–2541
  18. Schweizer JJ, Oren A, Mearin ML, Working Group for Celiac Disease and Malignancy of the European Society of Paediatric Gastroenterology, Hepatology and Nutrition . Cancer in children with celiac disease: a survey of the European Society of Paediatric Gastroenterology, Hepatology and Nutrition. J Pediatr Gastroenterol Nutr. 2001;33:97–100
  19. Catassi C, Fabiani E, Corrao G, Barbato M, De Renzo A, Carella AM, et al. Risk of non-Hodgkin lymphoma in celiac disease. JAMA. 2002;287:1413–1419
  20. Sigurgeirsson B, Agnarsson BA, Lindelof B. Risk of lymphoma in patients with dermatitis herpetiformis. BMJ. 1994;308:13–15
  21. Collin P, Pukkala E, Reunala T. Malignancy and survival in dermatitis herpetiformis: a comparison with coeliac disease. Gut. 1996;38:528–530
  22. Larsson LG, Sandstrom A, Westling P. Relationship of Plummer-Vinson disease to cancer of the upper alimentary tract in Sweden. Cancer Res. 1975;35:3308–3316
  23. Reunala TL, Leonard JN. Malignant disease in dermatitis herpetiformis. Clin Dermatol. 1991;9:369–373
  24. Swerdlow AJ, Whittaker S, Carpenter LM, English JS. Mortality and cancer incidence in patients with dermatitis herpetiformis: a cohort study. Br J Dermatol. 1993;129:140–144
  25. Nielsen OH, Jacobsen O, Pedersen ER, Rasmussen SN, Petri M, Laulund S, et al. Non-tropical sprue. Malignant diseases and mortality rate. Scand J Gastroenterol. 1985;20:13–18
  26. Green PHR, Fleischauer AT, Bhagat G, Leffler D, Neugut AI. Increased risk for non-Hodgkin's lymphoma persists despite a gluten-free diet (abstr). Gastroenterology. 2002;122:A16

☆ Address requests for reprints to: Edward V. Loftus, Jr., M.D., Division of Gastroenterology and Hepatology, Mayo Clinic and Foundation, 200 First Street SW, Rochester, Minnesota 55905. e-mail: loftus.edward@mayo.edu; fax: (507) 284-5486.

PII: S0016-5085(02)00305-0

Refers to article:

  • Cancer incidence in a population-based cohort of individuals hospitalized with celiac disease or dermatitis herpetiformis

    Johan Askling, Martha Linet, Gloria Gridley, Trond S. Halstensen, Karin Ekström, Anders Ekbom
    Gastroenterology November 2002 (Vol. 123, Issue 5, Pages 1428-1435)

Gastroenterology
Volume 123, Issue 5 , Pages 1726-1729, November 2002

Article Tools