AIM: To determine the contribution of HLA-DQA1* and HLA-DQB1* genes to the risk of coeliac disease (CD) in a cohort of children with type 1 diabetes mellitus (T1DM) from northern Italy. METHODS: Three hundred and fifty-seven children with T1DM, attending the Childhood Diabetes Unit of the University of Verona, have been regularly tested for serum IgA endomysial antibodies (EMA). All patients with positive EMA underwent small bowel biopsy to confirm the diagnosis of CD. HLA typing was performed in subjects with T1DM and CD, and in a control group of 79 EMA-negative patients with T1DM. RESULTS: Of the 357 patients tested, 25 (7%) had CD. The frequency of HLA-DQA1*0501-DQB1*0201 (T1DM + CD 68% vs T1DM 62%) and of DQA1*0301-DQB1*0302 (T1DM + CD 40% vs T1DM 35%) haplotypes, between T1DM patients with and without CD, was statistically comparable. A trend towards a reduction of the risk of CD (p = 0.055, OR: 0.22, CI 0.05: 1.04) was observed in patients with T1DM (28% vs T1DM + CD 2%) who did not carry either the HLA-DQA1*0501-DQB1*0201 or the DQA1*0301-DQB1*0302 haplotype. CONCLUSION: A high prevalence of HLA-DQA1* and -DQB1* susceptibility haplotypes for CD was observed both in EMA-negative diabetics and in those with associated CD. The implementation of screening programmes of CD in a T1DM population, based on the identification of HLA susceptibility haplotypes, seems to be of limited usefulness. Serial serologic screening of diabetic patients remains the advisable strategy.
AIM: To determine the contribution of HLA-DQA1* and HLA-DQB1* genes to the risk of coeliac disease (CD) in a cohort of children with type 1 diabetes mellitus (T1DM) from northern Italy. METHODS: Three hundred and fifty-seven children with T1DM, attending the Childhood Diabetes Unit of the University of Verona, have been regularly tested for serum IgA endomysial antibodies (EMA). All patients with positive EMA underwent small bowel biopsy to confirm the diagnosis of CD. HLA typing was performed in subjects with T1DM and CD, and in a control group of 79 EMA-negative patients with T1DM. RESULTS: Of the 357 patients tested, 25 (7%) had CD. The frequency of HLA-DQA1*0501-DQB1*0201 (T1DM + CD 68% vs T1DM 62%) and of DQA1*0301-DQB1*0302 (T1DM + CD 40% vs T1DM 35%) haplotypes, between T1DM patients with and without CD, was statistically comparable. A trend towards a reduction of the risk of CD (p = 0.055, OR: 0.22, CI 0.05: 1.04) was observed in patients with T1DM (28% vs T1DM + CD 2%) who did not carry either the HLA-DQA1*0501-DQB1*0201 or the DQA1*0301-DQB1*0302 haplotype. CONCLUSION: A high prevalence of HLA-DQA1* and -DQB1* susceptibility haplotypes for CD was observed both in EMA-negative diabetics and in those with associated CD. The implementation of screening programmes of CD in a T1DM population, based on the identification of HLA susceptibility haplotypes, seems to be of limited usefulness. Serial serologic screening of diabeticpatients remains the advisable strategy.