A decision algorithm to identify patients with high probability of monogenic diabetes due to HNF1A mutations

PURPOSE: To investigate the utility of biomarkers of maturity-onset diabetes of the young (MODY), high-sensitivity C-reactive protein (hsCRP), and 1,5-anhydroglucitol (1,5-AG) in conjunction with other clinical and laboratory features to improve diagnostic accuracy and provide a diagnostic algorithm for HNF1A MODY. METHODS: We examined 77 patients with HNF1A MODY, 88 with GCK MODY mutations, 99 with type 1 diabetes, and 92 with type 2 diabetes. In addition to 1,5-AG and hsCRP, we considered body mass index (BMI), fasting glucose, and fasting serum C-peptide as potential biomarkers. Logistic regression and receiver operating characteristic curves were used in marker evaluation. RESULTS: Concentration of hsCRP was lowest in HNF1A MODY (0.51 mg/l) and highest in type 2 diabetes (1.33 mg/l). The level of 1,5-AG was lowest in type 1 diabetes and HNF1A MODY, 3.8 and 4.7 μg/ml, respectively, and highest (11.2 μg/ml) in GCK MODY. In the diagnostic algorithm, we first excluded patients with type 1 diabetes based on low C-peptide (C-statistic 0.98) before using high BMI and C-peptide to identify type 2 diabetes patients (C-statistic 0.92). Finally, 1,5-AG and hsCRP in conjunction yielded a C-statistic of 0.86 in discriminating HNF1A from GCK MODY. We correctly classified 92.9% of patients with type 1 diabetes, 84.8% with type 2 diabetes, 64.9% HNF1A MODY, and 52.3% GCK MODY patients. CONCLUSIONS: Plasma 1,5-AG and serum hsCRP do not discriminate sufficiently HNF1A MODY from common diabetes types, but could be potentially useful in prioritizing Sanger sequencing of HNF1A gene.