Phenome‐wide association study in adult coeliac disease: role of HLA subtype

Abstract

SummaryBackgroundCoeliac disease arises in genetically susceptible individuals, in particular in carriers of HLA‐DQ2/DQ8 risk alleles and is associated with various comorbidities. Coeliac disease may confer an increased mortality, but the data are conflicting.AimsWe aimed to characterize mortality and morbidity in patients with coeliac disease with a special focus on the role of the number of HLA risk alleles.MethodsWe studied coeliac disease‐associated morbidity and mortality in ~500 000 participants of the UK Biobank including 2482 individuals with the diagnosis of coeliac disease. We used an unbiased, multivariable Phenome‐Wide Association Study (PheWAS) method to uncover the coeliac disease‐associated disorders. The tag SNP approach was used to divide the coeliac disease subjects into HLA‐DQ2/DQ8‐based risk categories.ResultsWe found 225 ICD‐10 codes significantly associated with coeliac disease. During the median follow‐up of 10.7 years, coeliac disease individuals (n = 2482) had higher overall mortality (HR 1.6 [95% CI, 1.4‐1.8]) than controls and both an increased occurrence of and an increased mortality from cancer, respiratory and cardiovascular diseases (HR 1.4‐1.6). Coeliac disease individuals with 2 HLA‐DQ2/8 risk alleles had a similar overall mortality as coeliac disease participants with 0‐1 HLA‐DQ2/8 alleles, but were more likely to die from lymphoproliferative diseases (HR 7.6 [95% CI, 1.01‐57.25]).ConclusionsOur data suggest that the increased mortality from lymphoproliferative diseases is restricted to those coeliac patients with 2 HLADQ2/8 alleles and that a combination of coeliac disease and HLADQ2/8 alleles is needed to increase the susceptibility. Once confirmed, closer monitoring may be warranted in this high‐risk subpopulation.

Systematic review with meta‐analysis: Cause‐specific and all‐cause mortality trends across different coeliac disease phenotypes

“…As shown in Figure 1, a literature search identified 5898 records, from which 25 eligible studies were identified 10–16,18,19,23–38 . Of these, seven were older reports 11–13,23–26 for which updated or more comprehensive reports on the same population of patients were available 10,16,27,30,31 .…”

“…A post‐hoc analysis of cause‐specific mortality in dermatitis herpetiformis showed reduced mortality from cardiovascular diseases (pooled HR 0.70, 95% CI 0.53–0.92) and no increase in mortality from malignancies overall (pooled HR 0.87, 95% CI 0.62–1.23). Four studies reported outcomes on cause‐specific mortality in undiagnosed coeliac disease 19,28,32,33 . However, as the sample size of studies on undiagnosed coeliac disease was generally small, no further analysis on cause‐specific mortality in these patients was possible.…”

“…Data on mortality in coeliac disease compared to the general population was available for only a few countries. Specifically, multiple studies were conducted in the UK 15,28,30–32,34–38 and Scandinavian countries (Finland, Sweden, Denmark), 14,16,18,27,29,33 two were conducted in Italy, one of which was a multicentre study 10 also including the patients described in the other paper 12 and one was conducted in the USA on undiagnosed coeliac disease 19 . Overall, studies conducted in Scandinavian countries showed no increase in mortality (pooled HR 1.01, 95% CI 0.91–1.13; I 2 = 81%) while in the UK there was overall a slight increase in mortality (pooled HR 1.23, 95% CI 1.03–1.47; I 2 = 91%).…”

“…Finally, one study investigated the relationship between the number of HLA‐DQ2/DQ8 alleles (0–1 vs 2) and mortality, both in celiac patients and controls, and found no significant relationship with all‐cause mortality in either group. However, an increased risk of death from lymphoproliferative disorders was found only in celiac patients with 2 HLA‐DQ2/DQ8 alleles (HR 7.6, 95% CI, 1.01–57.25) but not in coeliac patients with 0–1 alleles, or controls irrespective of number of alleles 28 . Among the included studies, none evaluated the role of seronegative coeliac disease or complicated coeliac disease on mortality.…”

“…Possibly explaining these discrepancies, Tio et al. could only include data from 3 studies in their analysis of cardiovascular death, 17 while we included data from 13 studies in our analysis of cardiovascular death, many of them published after 2012 10,14,16,18,27–33,35,38 …”