Predictive Factors of Insulin Discontinuation after Adding a GLP-1 Receptor Agonist in Patients with Type 2 Diabetes Treated with Insulin in Combination with Oral Antihyperglycemic Agents

The aim of this 6-month (mo) retrospective study was to evaluate the effectiveness of GLP1 receptor agonists (GLP1RA) in patients with T2D treated with insulin plus oral antihyperglycemic drugs and to determine predictors of successful insulin discontinuation (ID), defined as the achievement of A1C <7% 6 mo after the initiation of the GLP1RA in those patients who stopped insulin over the follow-up. The best predictive model of successful ID was estimated by logistic regression.

159 patients were included, mean age 61.8 y, A1C 8.42%, BMI 36.7 kg/m2 and insulin dose 55.9 U/d; 61.6% received liraglutide, 25.2% exenatide BID and 13.2% exenatide QW. A1C (-0.99%) and weight (-4.2kg) significantly decreased at 6 mo. 19.5% of patients discontinued insulin (table), but only 7.5% of patients stopped insulin and maintained A1C<7% at 6 mo. Fasting plasma glucose, A1C, duration of T2D, duration of insulin therapy, insulin dose and diabetic renal disease were all negatively associated to successful ID in the univariate analysis, but only insulin dose remained significant in the multivariate analysis (OR 0.88, 95CI% 0.80-0.97).

In summary, in a real-world setting, few patients treated with GLP1RA achieved good glycemic control after stopping insulin. Baseline insulin dose was the main baseline predictor of successful ID.