Performance of a predictive algorithm in sensor-augmented pump therapy in the prevention of hypoglycaemia

Abstract

The Predictive Low Glucose Management (PLGM) system consists of a Medtronic Veo pump, Enlite sensor, MiniLink REAL-Time transmitter, Bluetooth-RF translator and a predictive algorithm operating from a Blackberry smartphone. The system suspended insulin delivery when the pre-set hypoglycaemic threshold of 4.4mmol/L was predicted to be reached in 30 minutes. The aim of this study was to determine the plasma glucose profile with the PLGM system when hypoglycaemia was induced by (a) moderate-intensity exercise, (b) subcutaneous insulin bolus and (c) increasing the overnight basal infusion rate in individuals with type 1 diabetes. The primary outcome was the plasma glucose nadir following each hypoglycaemic stimulus with and without PLGM. Participants performed 30-60 minutes of moderate-intensity exercise or were administered a subcutaneous insulin bolus following a glucose stabilisation period on basal continuous subcutaneous insulin infusion. In participants studied with increased overnight basal rates, hypoglycaemia was induced by increasing basal rates by 180%. They were randomised and studied on 2 separate days; with PLGM off and with PLGM on. On both days, participants were observed until plasma glucose dropped to 2.8mmol/L or were symptomatic.