ESPE Abstracts

Objective: To provide etiology-based diagnostics to pediatric patients with diabetes in Belgium using routine clinical phenotyping and thorough genotyping.

Methods: A Belgian GENEPEDIAB study consortium was created to screen, using routine diagnostic tools, for monogenic forms of diabetes in pediatric patients followed in convention centers for type 1 or type 2 diabetes, while presenting atypical biological and clinical features of the disease (e.g. lack of anti-islet antibodies, persistence of C-peptide secretion and low glycemic variability; features not considered by the MODY calculator). We compiled the most representative clinical features of monogenic diabetes into a new diagnostic tool, the DIAMODIA score. Patients enrolled in GENEPEDIAB were phenotyped (e.g. glycemic variability, glucose tolerance, multiplex serum protein assays) and genotyped for “type 1 diabetes risk” (HLA typing and SNP-based diabetes risk score determination as established by the team of Prof A Hattersley (Exeter, UK). Patients fulfilling sufficient criteria were genotyped (restricted gene panel [University of Antwerp] then whole-exome sequencing using NGS). Gene-phenotype correlations were performed using bioinformatics (Prof. N. Limaye, de Duve Institute). Phenotyping analyses were compared to a control cohort of 295 patients with classical type 1 diabetes.

Results: A cohort of 750 patients diagnosed with type 1 and type 2 diabetes was evaluated. Our DIAMODIA score identified a subgroup of 107 patients (the “ADia cohort”) likely to present atypical diabetes. Data derived from the ADia cohort confirmed the following characteristics: absence of anti-islet antibodies (44%), presence of residual C-peptide secretion (median random C-peptide levels of 0.5±0.7 nM), diabetes control within recommended targets (median HbA1C: 6.6%). In ADia patients, routine MODY gene panel analysis identified 23 class 5 variants and 5 class 3 variants (10 HNF1A, 11 GCK, 4 HNF4A, 2 HNF1B, 2 ABCC8, 2 KCNJ11). In-depth WES analysis retrieved class 3 variants in 19 patients, providing our DIAMODIA score a yield of 21.9% for class 5 and 18.1% for class 3 gene variant positivity. Familial clustering analysis of variants is ongoing. Also, low levels of glycemic variability were key in identifying ADia patients with monogenic gene variants: a delta GTAA1C and delta IDAA1C score of >2 helped discriminate monogenic vs type 1 diabetes patient cohorts.

Conclusion: In our GENEPEDIAB cohort, phenotyping and genotyping helped us decipher new molecular elements in patients with atypical diabetes. Our consortium will now evaluate how etiology-based diagnosis of atypical forms of diabetes in children and adolescents might impact patient education and treatment individualization.