Closing the Loop: The Promise of Automated Insulin Delivery

Type 1 diabetes mellitus (T1DM) is managed by attempting to mimic physiologic insulin production with exogenous insulin to minimize both hypo- and hyperglycemia. Historically this has been accomplished through injections of basal insulin combined with frequent use of rapid acting insulin to compensate for food and/or correct hyperglycemia. Insulin doses are calculated and administered independently by people with diabetes. However, this task is burdensome and despite advancements in insulin action/delivery, hypo- and hyperglycemia remain important contributors to morbidity and mortality from T1DM. Insulin delivery via continuous subcutaneous insulin infusion pumps provides greater flexibility and precision in delivering insulin. Additionally, the use of continuous glucose monitors (CGM) has allowed people with diabetes to more closely monitor their blood glucose and intervene sooner, to prevent severe excursions. However, insulin pumps and CGM do not alleviate the need of the person with diabetes to continuously calculate and direct the device to administer insulin appropriately. Mathematical algorithms, which combine data from CGM and insulin delivery by insulin pump, often referred to as an artificial pancreas, represent a step forward in managing T1DM. These novel systems can autonomously calculate insulin requirements and deliver insulin in response to changing blood glucose levels with significant improvement in glycemic control as well as improvement in quality of life for the user. This presentation will explore the past, present, and future of automated insulin delivery.