Reproducibility of a prediabetes classification in a contemporary population.

Aims: To assess whether meeting both fasting plasma glucose (FPG) and HbA1c criteria for prediabetes in people at high risk indicates with near certainty the presence of dysglycemia on repeat testing. Methods: Observational study using data from Vitamin D and Type 2 Diabetes (D2d) study. HbA1c, FPG were measured at screening visit 1; FPG, HbA1c and 2 h plasma glucose (2hPG) measured at screening visit 2 (a median of 21 days later); participants classified as having normal glucose regulation (all 3 tests in normal range), prediabetes or diabetes (at least 1 of 3 tests in diabetes range). A predictive model was developed to estimate the probability of confirming dysglycemia and for detecting diabetes at screening visit 2 based on values of FPG and HbA1c at screening visit 1. Results: Of 1271 participants who met both FPG and HbA1c criteria for prediabetes at screening visit 1, 98.6% exhibited dysglycemia (defined as prediabetes or diabetes) on repeat testing (84.5% were classified as having prediabetes, 14.1% were reclassified as having diabetes). Of those with diabetes, 62.6% were identified by 2hPG alone. Conclusions: Combined measurement of FPG and HbA1c is a reliable and reproducible measure to identify presence of dysglycemia among people at high risk. A prediction model is provided to help clinicians decide whether an oral glucose tolerance test will provide value in detecting diabetes based on the 2hPG criterion