Feasibility Evaluation of an Artificial Pancreas With Glucose Prediction Trust Index

NCT ID: NCT03092310

Last Updated: 2018-07-12

Basic Information

• Recruitment Status: COMPLETED
• Clinical Phase: NA
• Total Enrollment: 15 participants
• Study Classification: INTERVENTIONAL
• Study Start Date: 2017-03-20
• Study Completion Date: 2017-12-20

Brief Summary

The purpose of this pilot study is to establish that closed-loop insulin delivery with a target enchanted model predictive control (eMPC)/Health Monitoring System (HMS) algorithms with a trust index of the predicted glucose value is safe and effective, to analyze and learn to improve upon the accuracy of the predicted glucose values, and to collect efficacy data to inform a future larger study.

Detailed Description

This is a feasibility study of an artificial pancreas (AP) system with our previously validated target eMPC and HMS algorithms, with the addition of a trust index of the predicted glucose value integrated into the portable Artificial Pancreas System (pAPS). The system will be evaluated on up to 20 subjects, with the goal of 10 subjects completing a 48-hour closed-loop session at one clinical site (William Sansum Diabetes Center). During the session subjects will bolus for all meals and snacks and perform a 45-minute walking session.

The purpose of this pilot study is to establish that closed-loop insulin delivery with a target enchanted model predictive control (eMPC)/Health Monitoring System (HMS) algorithms with a trust index of the predicted glucose value is safe and effective, to analyze and learn to improve upon the accuracy of the predicted glucose values, and to collect efficacy data to inform a future larger study.

Conditions

Type 1 Diabetes Mellitus

Study Design

• Allocation Method: NA
• Intervention Model: SINGLE_GROUP
• Primary Study Purpose: TREATMENT
• Blinding Strategy: NONE

Study Groups

Artificial Pancreas

The artificial pancreas device will employ its enhanced Model Predictive Control (MPC) algorithm with a target glucose level of 110 mg/dL with a trust index for MPC-predicted glucose values, weighing future glucose predictions and only acting on predictions with higher weight in the trust index. The Health Monitoring System algorithm uses the same CGM data as the MPC control algorithm but utilizes a separate algorithm for trending and predictions of future glucose values. Using a redundant and independent algorithm is an important safety feature of the overall AP device.

Group Type: EXPERIMENTAL

Artificial Pancreas Device

Intervention Type: DEVICE

Artificial Pancreas Device

Interventions

Artificial Pancreas Device

Artificial Pancreas Device

Intervention Type: DEVICE

Eligibility Criteria

Inclusion Criteria

• Age ≥ 18 and ≤ 75 years at the time of screening.
• Clinical diagnosis of type 1 diabetes for at least one year.
• Has been using an insulin pump for at least 6 months at the time of screening.
• HbA1c < 10%, as performed by point of care or central laboratory testing. A1c will be assessed at the screening visit, or if already completed within 2 months of the screening visit, the prior lab value may be used in lieu of repeating this assessment.
• Serum creatinine of < 2.0 mg/dL. Serum creatinine will be assessed at the screening visit, or if already completed within 2 months of the screening visit, the prior lab value may be used in lieu of repeating this assessment.
• Bolus for all meals and snacks that contain ≥ 5 grams of carbohydrate.
• Willing to perform at least 7 fingerstick blood glucose tests a day.
• Willing to refrain from taking acetaminophen products for the duration of the clinical trial. If acetaminophen is taken, subject is to avoid making any insulin dosing decisions based on CGM for at least 12 hours.
• Willing to abide by the study protocol and use study-provided devices, including the Omnipod, Dexcom CGM, glucometer, ketone meter and pAPS tablet.

Exclusion Criteria

• Pregnancy
• One or more episodes of hypoglycemia requiring an emergency room visit or hospitalization in the past 6 months.
• One or more episodes of hyperglycemia requiring an emergency room visit or hospitalization in the past 6 months.
• Known unstable cardiac disease or untreated cardiac disease, as revealed by history or physical examination.
• Dermatological conditions that would preclude wearing a CGM sensor or Pod.
• One or more seizures in the past year.
• Screening A1c ≥ 10% or serum creatinine ≥ 2.0 mg/dL.
• Any condition that could interfere with participating in the trial, based on investigator judgment.
• Concurrent use of any non-insulin glucose-lowering agent (including GLP-1 agonists, Symlin, DPP-4 inhibitors, SGLT-2 inhibitors, biguanides, sulfonylureas and naturaceuticals), defined as using any of these medications within 30 days of the screening visit or during the study.
• Participation in another pharmaceutical or device trial at the time of enrollment or during the study.

• Minimum Eligible Age: 18 Years
• Maximum Eligible Age: 75 Years
• Eligible Sex: ALL
• Accepts Healthy Volunteers: No

Sponsors

Harvard University

OTHER

Sponsor Role: collaborator

Sansum Diabetes Research Institute

OTHER

Sponsor Role: lead

Responsible Party

Jordan Pinsker

Research Physician

Responsibility Role: PRINCIPAL_INVESTIGATOR

Principal Investigators

Jordan E Pinsker, MD

Role: PRINCIPAL_INVESTIGATOR

Sansum Diabetes Research Institute

Eyal Dassau, PhD

Role: PRINCIPAL_INVESTIGATOR

Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University

Francis J Doyle III, PhD

Role: PRINCIPAL_INVESTIGATOR

Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University

Alejandro Leguna, PhD

Role: PRINCIPAL_INVESTIGATOR

Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University

Locations

William Sansum Diabetes Center

Santa Barbara, California, United States

Countries

United States

References

Pinsker JE, Laguna Sanz AJ, Lee JB, Church MM, Andre C, Lindsey LE, Doyle FJ 3rd, Dassau E. Evaluation of an Artificial Pancreas with Enhanced Model Predictive Control and a Glucose Prediction Trust Index with Unannounced Exercise. Diabetes Technol Ther. 2018 Jul;20(7):455-464. doi: 10.1089/dia.2018.0031.

Reference Type: RESULT

PMID: 29958023 (View on PubMed)

Other Identifiers

Pro00020556

Identifier Type: -

Identifier Source: org_study_id