Effect of Gain on Closed-Loop Insulin

The purpose of this study is to test the ability of an advanced external Physiologic Insulin Delivery (ePID) algorithm (a step by step process used to develop a solution to a problem) to get acceptable meal responses over a range of gain. Gain is defined as how much insulin is given in response to a change in a patient's glucose level. This study also examines the effectiveness of the external Physiologic Insulin Delivery (ePID) closed-loop insulin delivery computer software. The investigators would like to assess whether fasting target levels can be achieved as the closed-loop gain increases or decreases, and to evaluate the system's ability to produce an acceptable breakfast meal response.

There have been significant advances in diabetes management technology, including more sophisticated insulin pumps and more accurate real-time continuous glucose monitors. The next technological development is widely thought to be the introduction of an algorithm linking the pump and sensor to form a closed-loop insulin delivery system. The algorithm used for this purpose needs to be robust to changes in an individual's insulin sensitivity, and the sensor's sensitivity to glucose. Insulin sensitivity (how much the patient's glucose level changes in response to a change in insulin delivery) and algorithm gain (how much insulin is delivered in response to a change in glucose) determine the systems overall closed-loop gain. Ideally, the overall gain can be set to achieve the lowest possible peak postprandial glucose response without postprandial hypoglycemia. However, if the algorithm's gain is set to a fixed value and the subject's insulin sensitivity changes, the overall-gain will change. Some degradation in closed-loop performance might be acceptable during periods whenever the subject's insulin sensitivity is low (i.e., the subject is insulin resistant) and the risk of hypoglycemia may actually be reduced. However, if the subject becomes more sensitive the system may become less stable and the risk of postprandial hypoglycemia may increase. In addition to changes in insulin sensitivity, glucose sensors will sometimes over- or under-read blood glucose as sensor sensitivity increases or decreases. This will result in a change in the closed-loop algorithm's effective target. The purpose of this study is to evaluate the ability of an advanced Physiologic Insulin Delivery algorithm to achieve an acceptable breakfast response as the gain and effective target glucose level changes. Specifically: to assess the fasting glucose levels achieved as the overall closed-loop gain and effective target is increased or decreased, and determine the system's ability to produce an acceptable breakfast meal response under these conditions

Study subjects are studied under closed-loop control on three occasions: once with the glucose values used for control equal to blood glucose (NO error), once with values 33% higher than blood glucose (HIGH error), and once with values 20% lower than blood glucose (LOW error). The six different sequences of these three exposures then comprise the six arms of this crossover study.

Subjects were randomized to receive overnight and breakfast closed-loop glucose control glucose on three occasions: first with glucose-value-used-for-control higher than blood glucose (HIGH error), then second with glucose-value-used-for-control lower than blood glucose (LOW error), then third with glucose-value-used-for-control equal blood glucose (NO error).

Subjects were randomized to receive overnight and breakfast closed-loop glucose control glucose on three occasions: first with glucose-value-used-for-control higher than blood glucose (HIGH error), then second with glucose-value-used-for-control equal blood glucose (NO error), then third with glucose-value-used-for-control lower than blood glucose (LOW error).

Subjects were randomized to receive overnight and breakfast closed-loop glucose control glucose on three occasions: first with glucose-value-used-for-control equal blood glucose (NO error), then second with glucose-values-used-for-control higher than blood glucose (HIGH error), then third with glucose-value-used-for-control lower than blood glucose (LOW error).

Subjects were randomized to receive overnight and breakfast closed-loop glucose control glucose on three occasions: first with glucose-value-used-for-control equal blood glucose (NO error), then second with glucose-value-used-for-control lower than blood glucose (LOW error), then third with glucose-value-used-for-control higher than blood glucose (HIGH error).

Subjects were randomized to receive overnight and breakfast closed-loop glucose control glucose on three occasions: first with with glucose-value-used-for-control lower than blood glucose (LOW error), then second with glucose-value-used-for-control equal blood glucose (NO error), then third with glucose-value-used-for-control higher than blood glucose (HIGH error).

Subjects were randomized to receive overnight and breakfast closed-loop glucose control glucose on three occasions: first with glucose-value-used-for-control lower than blood glucose (LOW error), then second with glucose-value-used-for-control equal blood glucose (NO error), then third glucose-value-used-for-control higher than blood glucose (HIGH error),

Overnight and breakfast closed-loop control were performed using a target glucose of 120 mg/dL but with the glucose-value-used-for-control equal to 1.33 times the true glucose value (analogous to higher gain lower target).

Overnight and breakfast closed-loop control were performed using a target glucose of 120 mg/dL and glucose-value-used-for-control equal to the true glucose value.

Overnight and breakfast closed-loop control were performed using a target glucose of 120 mg/dL but with the glucose-value-used-for-control equal to 0.8 times the true glucose value (analogous to lower gain higher target).

Glucose Area Under the Curve (AUC) Breakfast defines the total exposure to glucose during breakfast. Breakfast is typically considered the most difficult meal to control; low AUC is desirable.This outcome measure was analyzed for each of the three calibration error values (high error, no error and low error).

On day #1, day #2 and day #3 (each day could be 24 hours to 7 days apart from prior one, and completed within 6 week period) 8:00 AM to 2:00 PM on day following admission, with samples obtained every 10-15 minutes, for each sequence of calibration errors

On day #1, day #2 and day #3 (each day could be 24 hours to 7 days apart from prior one, and completed within 6 week period) 8:00 AM to 12:00 PM on day following admission, with samples obtained every 10-15 minutes, for each sequence of calibration errors

Night-time in target range 5.0-8.33, following the 3 hour controller initialization period blood glucose remained at or near target.

On day #1, day #2 and day #3 (each day could be 24 hours to 7 days apart from prior one, and completed within 6 week period) 12:00 AM to 6:00 AM on day following admission, with samples obtained every 10-15 minutes, for each sequence of calibration errors

Inclusion Criteria: Type 1 diabetes for > 3 years Manage diabetes using a continuous glucose monitor and continuous subcutaneous insulin infusion pump Non obese (BMI < 30) Aged 18 - 75 years old HbA1c < 8 % Exclusion Criteria: renal or hepatic failure cancer or lymphoma Malabsorption or malnourishment Hypercortisolism Alcoholism or drug abuse Anemia (hematocrit < 36 in females and <40 in males) Eating disorder Dietary restrictions Acetaminophen allergy Chronic acetaminophen use Glucocorticoid therapy History of gastroparesis Use of Beta blockers

Steil GM, Panteleon AE, Rebrin K. Closed-loop insulin delivery-the path to physiological glucose control. Adv Drug Deliv Rev. 2004 Feb 10;56(2):125-44. doi: 10.1016/j.addr.2003.08.011.

Steil GM, Rebrin K, Janowski R, Darwin C, Saad MF. Modeling beta-cell insulin secretion--implications for closed-loop glucose homeostasis. Diabetes Technol Ther. 2003;5(6):953-64. doi: 10.1089/152091503322640999.

Steil GM, Rebrin K, Darwin C, Hariri F, Saad MF. Feasibility of automating insulin delivery for the treatment of type 1 diabetes. Diabetes. 2006 Dec;55(12):3344-50. doi: 10.2337/db06-0419.

Weinzimer SA, Steil GM, Swan KL, Dziura J, Kurtz N, Tamborlane WV. Fully automated closed-loop insulin delivery versus semiautomated hybrid control in pediatric patients with type 1 diabetes using an artificial pancreas. Diabetes Care. 2008 May;31(5):934-9. doi: 10.2337/dc07-1967. Epub 2008 Feb 5.

Steil GM, Palerm CC, Kurtz N, Voskanyan G, Roy A, Paz S, Kandeel FR. The effect of insulin feedback on closed loop glucose control. J Clin Endocrinol Metab. 2011 May;96(5):1402-8. doi: 10.1210/jc.2010-2578. Epub 2011 Mar 2.

Loutseiko M, Voskanyan G, Keenan DB, Steil GM. Closed-loop insulin delivery utilizing pole placement to compensate for delays in subcutaneous insulin delivery. J Diabetes Sci Technol. 2011 Nov 1;5(6):1342-51. doi: 10.1177/193229681100500605.

Buchanan TA, Xiang AH, Peters RK, Kjos SL, Berkowitz K, Marroquin A, Goico J, Ochoa C, Azen SP. Response of pancreatic beta-cells to improved insulin sensitivity in women at high risk for type 2 diabetes. Diabetes. 2000 May;49(5):782-8. doi: 10.2337/diabetes.49.5.782.

Panteleon AE, Loutseiko M, Steil GM, Rebrin K. Evaluation of the effect of gain on the meal response of an automated closed-loop insulin delivery system. Diabetes. 2006 Jul;55(7):1995-2000. doi: 10.2337/db05-1346.