Smart Titration of Long-Acting Insulin in Insulin-Naïve Type 2 Diabetic Subjects

Insulin-naïve subjects with type 2 diabetes (T2D) begin their therapy with a low initial insulin dose (IID) which is gradually increased to reach the optimal insulin dose (OID) and optimal blood glucose target (OBG). However, for those individuals with a high insulin need (HIN), the STANDARD IID recommended by American Diabetes Association (ADA) guidelines may not be sufficient for achieving OBG in a timely manner. To address this limitation, we propose a SMART IID strategy and assess its possible benefits and risks on blood glucose control during the insulin titration period. A 52-week basal insulin titration trial was performed on 300 in silico insulin-naïve subjects of the T2D Padova simulator, to determine their subject-specific OID. Based on the final OID, subjects were classified as having HIN or low insulin need (LIN). A logistic regression model was developed to classify subjects as HIN or LIN based on certain characteristics available at the start of the trial. The subjects classified as HIN by the model were then assigned to a higher IID (SMART). A second 52-week titration trial was conducted using the SMART IID on the subjects classified as HIN. To assess effectiveness and safety of SMART IID, we compared time above 180 mg/dL (Ta), time in target 70-180 mg/dL (Tt), time below 70 mg/dL (Tb), and cumulative number of hypoglycemic events (NH) in the two experiments at 90, 124, 180, 270 and 365 days, with 124 being the time to reach the OBG. Results found that Ta was always significantly lower and Tt was always significantly higher with SMART IID than STANDARD titration (p