The Use of an Automated, Portable Glucose Control System for Overnight Glucose Control in Adolescents and Young Adults With Type 1 Diabetes

Abstract

OBJECTIVEA key milestone in progress towards providing an efficacious and safe closed-loop artificial pancreas system for outpatient use is the development of fully automated, portable devices with fault detection capabilities to ensure patient safety. The ability to remotely monitor the operation of the closed-loop system would facilitate future physician-supervised home studies.RESEARCH DESIGN AND METHODSThis study was designed to investigate the efficacy and safety of a fully automated, portable, closed-loop system. The Medtronic Portable Glucose Control System (PGCS) consists of two subcutaneous glucose sensors, a control algorithm based on proportional-integral-derivative with insulin feedback operating from a BlackBerry Storm smartphone platform, Bluetooth radiofrequency translator, and an off-the-shelf Medtronic Paradigm Veo insulin pump. Participants with type 1 diabetes using insulin pump therapy underwent two consecutive nights of in-clinic, overnight, closed-loop control after a baseline open-loop assessment.RESULTSEight participants attended for 16 overnight studies. The PGCS maintained mean overnight plasma glucose levels of 6.4 ± 1.7 mmol/L (115 ± 31 mg/dL). The proportion of time with venous plasma glucose <3.9, between 3.9 and 8 (70 and 144 mg/dL), and >8 mmol/L was 7, 78, and 15%, respectively. The proportion of time the sensor glucose values were maintained between 3.9 and 8 mmol/L was greater for closed-loop than open-loop (84.5 vs. 46.7%; P < 0.0001), and time spent <3.3 mmol/L was also reduced (0.9 vs. 3%; P < 0.0001).CONCLUSIONSThese results suggest that the PGCS, an automated closed-loop device, is safe and effective in achieving overnight glucose control in patients with type 1 diabetes.

Day and Night Closed-Loop Control Using the Integrated Medtronic Hybrid Closed-Loop System in Type 1 Diabetes at Diabetes Camp

“…Further advances in both sensor performance and algorithms incrementally demonstrated improved protection against hypoglycemia, reduced variability, and decreased mean glucose levels in controlled inpatient settings (11)(12)(13) as well as in outpatient studies (2,4,6). The strategies for different control algorithms are well described in several recent reviews (14,15).…”

Artificial Pancreas Using a Personalized Rule-Based Controller Achieves Overnight Normoglycemia in Patients with Type 1 Diabetes

“…We hypothesized that our algorithm proposing a microbolus every 5 min, mimicking the physiological pulsatile insulin secretion, 27,28 could contribute to insulin action optimization and favor stability. This pulsatile administration is different than the strategy used in other CL studies, [7][8][9]13,15,29 in which a change in the insulin delivery rate is made so insulin is distributed throughout the actuation period.…”

“…5 The most widely used strategies for CL controllers have been focused on proportional integral derivative control 6,7 and model predictive control. 8,9 Other approaches use an inverse controller 10 or imitate the lines of reasoning of diabetes caregivers, using methods such as fuzzy logic control 11 or rule-based algorithms.…”

“…8,9 Other approaches use an inverse controller 10 or imitate the lines of reasoning of diabetes caregivers, using methods such as fuzzy logic control 11 or rule-based algorithms. 12 In the last few years, some clinical experience with CL control algorithms has yielded favorable results in regulating BG levels in subjects with T1DM, both in-clinic [6][7][8][9]13 or very closely supervised conditions. 11 The percentage of time in BG target range was generally defined as the main goal.…”

Processing Diabetes Mellitus Composite Events in MAGPIE

“…Within this context several algorithms have been proposed to predict blood glucose levels in a given time horizon, in order to calculate the right dose of insulin to be injected by the insulin pump. These prediction algorithms are based on different techniques like rules [12], artificial neural networks [32], proportional integral derivative controllers [28], and neuro-fuzzy techniques [45]. There are also algorithms focusing on denoising CGM readings [16] that can help on improving the prediction.…”