The objective of this project is to investigate how the glucoregulatory system responds to exercise at different intensities and for different types of exercise including aerobic and resistance for the purpose of developing a better artificial pancreas (AP) control system. The experiments proposed here will specifically reveal how exercise impacts 1) non- insulin mediated glucose uptake, 2) insulin mediated glucose uptake (insulin sensitivity), 3) glucagon responsiveness in euglycemia, and 4) glucagon responsiveness during hypoglycemia. Exercise commonly results in hypoglycemia in persons with type 1 diabetes due to a rapid increase in glucose uptake by working muscles as well as an increase in insulin sensitivity. To date, avoiding exercise-induced hypoglycemia using current closed loop systems has had limited success as most AP algorithms are designed to react to declines in sensed glucose, and this decline may be significantly delayed from the onset of exercise. In this project, we propose to evaluate both insulin and glucagon sensitivity during and after exercise with tracer studies. In the first study, 26 adult subjects with type 1 diabetes will be brought in for three single-tracer, euglycemic experiments during which exercise will be performed. Subjects will receive intravenous insulin, glucose, and octreotide (to suppress endogenous hormone production) and will have blood drawn at regular intervals for measurement of insulin, native glucose, tracer glucose, catecholamine, and fatty acid levels. Glucose control will be managed by an infusion algorithm developed at OHSU. Subjects will repeat this experiment at three insulin infusion rates, designated low, medium, and high. Tracer levels will be used to determine rate of appearance (Ra) and rate of disappearance (Rd) of glucose to assess glucose uptake during exercise. Thirteen subjects will be placed in an arm of moderate aerobic exercise, and 13 will be placed in an arm of intense aerobic exercise. In the second study, subjects will be brought in for two consecutive single-tracer experiments, with two arms as outlined above for moderate and intense aerobic exercise. In the first experiment, subjects will be kept at euglycemia and given three doses of glucagon, one before, one during, and one after exercise. In the second experiment, glucose infusion rate will be clamped at an initial, basal rate during exercise and glucagon will be given before, during, and after exercise. Ra and Rd will be calculated and used to determine changes in glucagon sensitivity during exercise with or without hypoglycemia. Study 1 and 2 will then be repeated for resistance training rather than aerobic as anaerobic exercise is often associated with less dramatic changes in glucose levels. Data from these four studies will be used to develop an integrated model of glucose regulation during exercise that can be integrated in a model predictive control (MPC) algorithm for automated control of insulin and glucagon delivery in type 1 diabetes. In a final crossover study on 20 subjects, this MPC algorithm with exercise detection will be tested alongside a similar algorithm without exercise detection as well as sensor-augmented pumps (SAP) for control of glucose during free- range exercise in a 7 day out-patient study.
Subjects with type 1 diabetes often experience low blood glucose levels during and after exercise, resulting in fear of physical activity. Specific details with regards to how the human glucoregulatory system responds to exercise are currently not known, as we do not know (1) how insulin versus non-insulin mediated glucose uptake change during aerobic and nonaerobic exercise and across different intensities and (2) how glucagon sensitivity changes during euglycemia and hypoglycemia across aerobic and anaerobic exercise at different intensities. The experiments described in this grant will help answer these questions for the purpose of improving an artificial pancreas control system that can perform better during exercise.
