Background: Sensor-augmented pump therapy (SAPT) with algorithms to predict impending low blood glucose and suspend insulin delivery has the potential to reduce hypoglycemia exposure. The aim of this study was to determine whether predictive low glucose management (PLGM) system is effective in preventing insulin-induced hypoglycemia in controlled experiments. Methods: Two protocols were used to induce hypoglycemia in an in-clinic environment. (A) Insulin bolus: Insulin was administered as a manual bolus through the pump. (B) Increased basal insulin: Hypoglycemia was induced by increasing basal rates overnight to 180%. For both protocols, participants were randomized and studied on 2 separate days; a control day with SAPT alone and an intervention day with SAPT and PLGM activated. The predictive algorithm was programmed to suspend basal insulin infusion when sensor glucose was predicted to be <80 mg/dL in 30 min. The primary outcome was the requirement for hypoglycemia treatment (symptomatic hypoglycemia or plasma glucose <50 mg/dL) and was compared in both control and intervention arms. Results: With insulin bolus, 24/28 participants required hypoglycemia treatment with SAPT alone compared to 5/28 participants when PLGM was activated (P ≤ 0.001). With increased basal rates, all the eight SAPT-alone participants required treatment for hypoglycemia compared to only one with SAPT and PLGM. There was no post pump-suspend hyperglycemia with insulin bolus (P = 0.4) or increased basal rates (P = 0.69) in participants with 2-h pump suspension on intervention days. Conclusions: SAPT with PLGM reduced the requirement for hypoglycemia treatment following insulin-induced hypoglycemia in an in-clinic setting. © 2016, Mary Ann Liebert, Inc. 2016.