Long-Term Prognosis of Celiac Disease: A special focus on phenotype and other contributory factors

Celiac disease is an autoimmune enteropathy provoked by dietary gluten. Traditionally, the disease has been considered to cause diarrhea, malabsorption, and failure to thrive in children. With increasing knowledge and better diagnostic facilities, the clinical presentation of celiac disease has changed in recent decades. Now it is widely recognized as a systemic disease manifesting with numerous symptoms of varying severity and occurring in different organs. Celiac disease has been associated with both increased risk of malignancies and mortality, which have been attributed especially to the increased incidence of non-Hodgkin’s lymphoma (NHL). Yet it is largely obscure whether the risks have remained the same despite changes in celiac disease presentation and also in the long term. Possible factors predicting the prognosis are also poorly understood.

This dissertation consists of three studies investigating the prognosis of celiac disease in different disease phenotypes and in the long term. In Study I, the risk of malignancies was studied in a large, prospectively collected cohort of celiac disease patients diagnosed between 1960 and 2000 in Tampere University Hospital catchment area. The cohort involved 1,125 small bowel biopsy-verified celiac disease patients and 335 patients presenting with the skin manifestation of celiac disease, i.e. dermatitis herpetiformis (DH). Patient- and disease-specific clinical characteristics at the time of diagnosis and the findings of the duodenal biopsies were collected retrospectively from the medical records. The risk of all cancers and the most common individual cancers was assessed by comparing the number of cancers detected in the patient cohort with the expected number of cancers based on the cancer incidence of the population in the same region and further calculating the standardized incidence ratios (SIR). Special attention was paid to evaluating cancer risk in different disease phenotypes and determining prognostic factors for the risk.

In Study II, the same patient cohort was used as in Study I to assess overall and cause-specific mortality. Due to the selection process of the reference cohort, 68 celiac patients were excluded, leaving 1,093 celiac and 299 DH patients in the final cohort. The hazard ratios (HR) for mortality were estimated by comparing data from the study cohort with data from a cohort of matched reference individuals. Mortality in different disease phenotypes and prognostic factors for mortality were estimated. The large population based celiac cohort in Study III included 12,803 biopsy- verified celiac disease and DH patients diagnosed during the period 2005–2014, who were entitled to reimbursement for the costs of gluten-free diet. These adult patients were extracted from the dietary grant registry maintained by the Social Insurance Institution of Finland. Overall and cause-specific mortality in the cohort were compared with the mortality of a cohort of matched reference individuals.

The results showed that celiac disease, in its full spectrum, was not associated with an increased overall risk of cancer or mortality, not even over a long period of time. However, there were differences in risk between different cancers, causes of death, and celiac disease phenotypes. The increased risk of NHL (SIR 3.0, 95% confidence interval (CI) 1.9–4.5, p<0.001) (I) was at a level similar to that reported in earlier studies, but the increase in NHL-related mortality (HR 4.6, 95% CI 2.0– 10.7, p<0.001 (II) and HR 4.5, 95% CI 2.8–7.2, p<0.001 (III)) was lower than previously reported. The risk of gastrointestinal cancers was also increased (SIR 1.4, 95% CI 1.1–1.9, p<0.05) (I), mainly due to a higher incidence of liver cancers. In contrast, the relative risk of breast cancer was decreased (SIR 0.7, 95% CI 0.5–0.99, p<0.05) (I). Mortality from cancers overall, gastrointestinal cancers or cardiovascular diseases did not differ from those of the reference individuals. Apart from mortality from lymphoproliferative diseases, only mortality from non-malignant digestive diseases (HR 2.2, 95% CI 1.4–3.4, p<0.001) (III) and diseases of the central nervous system (HR 2.1, 95% CI 1.0–4.4, p<0.05) (II) was increased, while mortality from alcohol-related diseases was decreased (HR 0.3, 95% CI 0.1–1.0, p=0.050) (II).

In childhood celiac disease and in patients presenting with malabsorption or diagnosed by screening, the relative risks of NHL and gastrointestinal cancers were not increased. The risk of developing NHL was higher among men (SIR 3.6, 95% CI 1.8–6.5, p<0.001) than women (SIR 2.6, 95% CI 1.3–4.6, p<0.01). The risk of death from NHL increased with increasing age at diagnosis and was significant among patients diagnosed with celiac disease at an age over 40 (II) up to 50 (III) years.

This study demonstrated that the prognosis of celiac disease from the perspective of cancer risk and mortality is good in the majority of treated patients. Patients diagnosed with celiac disease at an older age and male patients have an increased risk of developing NHL.