Closed Loop Glucose Control in Patients With Type 1 and Type 2 Diabetes

Tight glucose control in the intensive care unit (ICU) setting is difficult to achieve. The investigators hypothesize that a closed loop glucose control system based on artificial intelligence (AI) will improve upon the glucose control currently achieved by the current open loop manual methods, and that this improved glucose control may improve the outcomes of critically ill patients, including those with COVID-19. This Earl Feasibility Study will test the ability of a prototype artificial intelligence based closed loop glucose control system named FUSION, to provide safe and effective glucose control in subjects with type 1 and type 2 diabetes in a clinical research center (CRC) setting. Subjects with type 1 diabetes have been chosen as safe and effective glucose control is difficult to achieve in these subjects during meal challenges. Subjects with type 2 diabetes have been chosen as they are insulin resistant, which makes their insulin resistance profile similar to that of ICU patients. As this is the first in human study of a new medical device, the controlled environment of the CRC is preferable to the less controlled environment of an ICU setting. The prototype FUSION system to be used in this study will consist to two Dexcom G6 continuous glucose monitors (CGM), the AI-based glucose control software run on an all-in-one medical computer, and two syringe pumps. The prototype system is housed on a medical cart. Based on the average glucose value of the two Dexcom G6 CGM's, and the rules of the FUSION systems AI-based glucose control software, the FUSION system will make rate adjustments every 5-10 minutes to the intravenous infusion rates of short acting insulin (NovoLog) and dextrose (D10NS) under its control, in an attempt to keep the subjects glucose in the range of 100-140 mg/dL. The FUSION system only requires entry of the subjects study identification number and weight in kilograms to initiate the system. For safety reasons, the subjects will have their blood glucose independently measured every 10-60 minutes on the point of care Nova StatStrip system, throughout the 24 hour study period. The study has halting criteria to avoid recurrent instances of severe hypoglycemia (\< 54 mg/dL). The average of the two CGM's, that is used by the FUSION system for glucose control, will be used for statistical analysis.