Incidence of Malignancies in Diagnosed Celiac Patients

A Population-Based Estimate

Tuire Ilus MD; Katri Kaukinen MD; PhD; Lauri J Virta MD; PhD; Eero Pukkala PhD; Pekka Collin MD; PhD


Am J Gastroenterol. 2014;109(9):1471-1477. 

In This Article


This is the largest study published to date on the risk of malignancies in clinically diagnosed celiac disease. The prevalence of biopsy-proven celiac disease in Finland is 0.6%, which is the highest reported so far.[12,14,15] This documented prevalence cannot be compared with the serological prevalence of the disease, which extends to this level and beyond it in many countries. This high prevalence of clinically diagnosed celiac disease implies that we had a highly representative cohort of diagnosed celiac disease and were not selecting only the most symptomatic or only screen-detected cases. This large, unselected cohort should finally give us a realistic projection of the cancer risk in clinically diagnosed celiac disease.

The overall risk of malignancies was decreased in female celiac disease patients, largely owing to the significantly reduced number of breast cancers, which is the most common cancer in Finnish women. Also contributing to the decrease in the overall cancer risk were the reduced numbers of lung, pancreatic, renal, and bladder cancers.

This study confirmed an increased risk of NHL in celiac patients, but the SIR of 1.94 was much lower than that reported in many previous studies (Table 4). The SIR for small-intestinal cancer in celiac disease patients was also increased, presumably owing to a long-standing inflammation of the small-intestinal mucosa.[19–23] Nevertheless, NHL was detected in only 0.4% and small-intestinal cancer in 0.08% of celiac patients within the observation period. As none of the cases of NHL or small-intestinal cancer was detected before the age of 30 years, the absolute risk of these malignant conditions seems to be negligible in younger celiac patients. For comparison, a study by Catassi et al.[24] investigated the occurrence of celiac disease in NHL. The odds ratio (adjusted for age and sex) for NHL of any primary site associated with celiac disease was 3.1 (95% CI 1.3–7.6). The authors concluded that the risk was not great enough to justify mass screening for celiac disease in NHL.

The risk of female breast cancer has been found to be decreased in all but one study (Table 4). Whether this is at least partly owing to the reduced reproductive period with late menarche and early menopause known to affect females with celiac disease is highly speculative.[25,26] However, it is unlikely that this alone would explain such a low SIR.

This study revealed a modestly increased risk of colon cancer in celiac patients, which has been similarly demonstrated in one study, covering only hospitalized patients.[3] The increased risk was observed only after a longer follow-up. A gluten-free diet contains less fiber than the normal diet,[27] and it alters the intestinal microbiota,[28] which may in turn influence the cancer risk in celiac patients. The risk of rectal cancer was not increased. This is not necessarily a contradictory observation, but it may also support the role of the diet, as dietary fiber is argued not to protect from rectal cancer owing to its short retention time.[29]

The slightly increased risk of skin basal cell carcinoma in celiac patients may well be coincidental, but on the other hand the regular visits of patients to the health-care facilities may make diagnosis of this cancer more feasible.

The main strength of this study is that it covers the whole spectrum of celiac disease, enabling accurate risk estimates with relatively narrow CIs also in rare malignancies such as small-intestinal cancer. The series is also particularly up-to-date, as the study period began in 2002, and it well describes the current celiac disease population. In Finland, celiac patients mostly visit the same primary health-care units as the population in general, and thus cancer detection among them is in no way facilitated by the follow-up of a specialist.

One limitation of the study is that celiac disease and dermatitis herpetiformis patients could not be analyzed separately, as differentiation between these two conditions was not reliably documented in the Social Insurance Register. Our previous study showed that 17% of celiac patients had skin involvement,[17] but the disease entity is largely the same, as is also the risk of malignancies.[7] In the case of patients diagnosed before 2003, the time from diagnosis was unknown. On the other hand, celiac disease is currently understood as a continuum from latent disease to classical small-intestinal lesions, and the diagnostic delay is predominantly long, which would imply that the precise time of the diagnosis is in any case more or less artificial.[4] A gluten-free diet is thought to prevent from malignancies including lymphomas,[30–32] but there is also contradictory evidence that compliance with the diet does not alter the lymphoma risk.[33] Although general compliance with the diet is high in Finland,[16] the low incidence of malignancies in this study does not unquestionably demonstrate a protective effect of a gluten-free diet on the cancer risk in celiac disease.

The risk of NHL was increased within 2 years after the diagnosis of celiac disease (Table 3). As to the cancer risk in general, the SIR remained at the population-based level immediately after the detection of celiac disease, but it was higher after 5 years from the diagnosis. The increased risk of colon cancer may cause concern (Table 3). However, in two large colonoscopy studies the numbers of detected colorectal adenomas and carcinomas including celiac patients were equal to those in controls.[34,35] This makes the rationale of surveillance colonoscopies in celiac patients questionable unless further studies can confirm this association.

We were not able to investigate the effect of many acknowledged confounding factors such as obesity and smoking on the cancer risk. Nevertheless, it is possible to pursue some speculations. A representative sample of Finnish adult celiac disease patients showed most to be of normal weight, but with a significantly lower body mass index than the population in general, both at diagnosis and after 1 year on a gluten-free diet.[36] Patients with celiac disease and dermatitis herpetiformis have been proved to smoke less than individuals in general.[37–39] Supporting this, the risk of lung cancer was significantly decreased in this study. Previous studies have also pointed in this direction, although the risk reduction has not been statistically significant in all (Table 4). Low smoking prevalence would also contribute to the decreased risk of renal and bladder cancers.

We conclude that the overall incidence of malignancy among clinically diagnosed celiac patients was not increased, although an increased SIR for malignancy was seen after 5 years from the diagnosis. The risks of NHL and small-intestinal cancer are in contrast increased, but to a lesser extent than previously described. This study confirms a good prognosis of clinically diagnosed celiac patients.