Modeling and modulating insulin delivery in automated insulin delivery systems to accommodate for meal compositions

点击翻译按钮获取中文摘要

PROJECT SUMMARY The candidate, Laya Ekhlaspour, MD, is dedicated to advancing diabetes management by decreasing the burden of diabetes care. This proposal will provide a structured clinical research training experience with formal mentorship that will enable Dr. Ekhlaspour to become an independent clinical researcher with expertise in closed loop systems. Currently, carbohydrates are considered the primary and major determinant affecting postprandial glucose control and the insulin bolus is based solely on the carbohydrate content of the meal along with the blood glucose at the time of the bolus. Given the known benefits of normalizing postprandial glucose excursions, the goal of this research is to specifically address problems related to providing adequate insulin coverage for meals that have variable fat and protein content that might result in both early hypoglycemia and prolonged hyperglycemia, when automated insulin systems are used. This proposal will take advantage of existing data collected during an observational study at the Barbara Davis Center, and the feasibility closed-loop trials at Stanford. In order to characterize the effect of macronutrient food content on postprandial glucose levels, a retrospective review of meals with known glucose, protein and fat content combined with CGM and insulin values in both open loop and closed loop situations will be conducted. This analysis will provide data to allow modeling for insulin requirements with meals of variable patterns of postprandial glycemic levels. It will also determine whether this modeling will allow for setting the percent of insulin required upfront and how long the insulin delivery needs to be extended in pumps which allow for extended meal boluses. The result of this analysis can be used in closed-loop control real-time modeling of meal boluses when there is a real-time adaptation to food absorption patterns, which will contribute to eventual fully closed-loop glucose control. The developed model will be based on assessing the hourly glucose and insulin requirements which could be integrated into prandial dosing algorithms in closed loop systems in order to optimize postprandial glycemic control. The long-term goal is to have a fully-closed loop algorithm, which will recognize the need for additional insulin with a high fat or protein meal based on the CGM postprandial tracing and insulin requirements during closed-loop control. This will reduce the burden of diabetes management significantly because the patient does not have to announce a meal. The proposed studies will provide the preliminary meal data for modeling of meal-responses that will account for both the carbohydrate, protein and fat content of the meal without user input. The following step will be to validate this model through in silico experiments and then conduct a randomized, controlled trial of the implementation of a potential refined meal algorithm that could handle postprandial glucose levels without an extended bolus as part of a competitive R01 grant application.