Artificial Pancreas Systems for Type 1 Diabetes

Recruiting at 2 trial locations

Overseen By Deborah Branigan

Age: 18+

Sex: Any

Trial Phase: Academic

Sponsor: Oregon Health and Science University

Must be taking: Insulin

Must not be taking: Beta blockers, Corticosteroids

Disqualifiers: Cardiovascular disease, Renal insufficiency, Liver disease, others

No Placebo Group All trial participants will receive the active study treatment (no placebo)

Approved in 3 Jurisdictions This treatment is already approved in other countries

Compensation: Varies

Number of Visits: 2

29 Participants Needed

Trial Summary

Do I have to stop taking my current medications for the trial?

The trial requires you to stop taking any medication intended to lower glucose other than insulin, such as metformin or liraglutide. You also cannot use beta blockers, non-dihydropyridine calcium channel blockers, or corticosteroids. If you're on any of these, you would need to stop them to participate.

What data supports the idea that Artificial Pancreas Systems for Type 1 Diabetes is an effective treatment?

The available research shows that Artificial Pancreas Systems, which use a Model Predictive Control (MPC) algorithm, are effective in managing blood sugar levels for people with Type 1 Diabetes. Studies have demonstrated that these systems improve safety by reducing the risk of low blood sugar events and effectively correcting high blood sugar levels. For example, one study found that the system increased the time patients spent in a safe blood sugar range from 68.8% to 76.9%. Additionally, these systems have been tested in clinical trials with positive results, showing improved performance over previous methods without increasing the risk of low blood sugar. This suggests that Artificial Pancreas Systems are a promising treatment option for better managing Type 1 Diabetes compared to traditional methods. ^{1 2 3 4 5}

What safety data exists for artificial pancreas systems in treating type 1 diabetes?

Several studies have evaluated the safety of artificial pancreas systems using Model Predictive Control (MPC) algorithms. These studies include clinical trials approved by the US Food and Drug Administration, demonstrating the systems' ability to reduce hypoglycemia and improve hyperglycemia correction. Enhancements like velocity-weighting and velocity-penalty MPC have been shown to improve safety by reducing controller-induced hypoglycemia and promoting effective hyperglycemia correction. Additionally, adaptive and enhanced MPC strategies have been tested in silico and in clinical trials, showing improved performance and safety by personalizing insulin delivery and minimizing hypoglycemia risk. Overall, these systems have been shown to increase time in safe glycemic ranges without significantly increasing hypoglycemia instances. ^{1 2 3 4 5}

Is the Model Predictive Control closed-loop system a promising treatment for Type 1 Diabetes?

Yes, the Model Predictive Control closed-loop system is a promising treatment for Type 1 Diabetes. It can safely and effectively manage blood sugar levels by automatically adjusting insulin delivery based on real-time needs. This system can handle different situations like meals and exercise, and it has been shown to reduce the risk of low blood sugar while improving high blood sugar correction. Clinical trials have demonstrated its benefits, making it a strong option for managing Type 1 Diabetes. ^{2 4 6 7 8}

What is the purpose of this trial?

An artificial pancreas (AP) is a control system for automatic insulin delivery. The investigators have implemented a missed meal bolus detection algorithm for use within an AP control system. If a meal is detected that was not reported by the user, the system shall calculate the amount of meal insulin that will be dosed and deliver that insulin. The investigators will test how well the new algorithm manages glucose compared to the participant's usual care including the tslim X2 pump with Control IQ enabled. This type of algorithm may improve glucose control for high risk patient populations.

Research Team

Leah Wilson, MD

Principal Investigator

Oregon Health and Science University

Eligibility Criteria

Adults over 18 with Type 1 Diabetes, living within 40 miles of the study site, who have been using a t:slim X2 insulin pump and Dexcom G6 CGM with Control IQ for at least 12 weeks. They must not be pregnant or planning pregnancy without proper contraception, have an HbA1c between ≥7.5% and ≤10%, no severe liver disease, infections, seizure disorders, recent drug trials participation, bleeding issues or allergies to Fiasp insulin.

Inclusion Criteria

I have been using a t:slim X2 insulin pump with Dexcom G6 and Control IQ for at least 12 weeks.

Lives with another person age 18 or older who will sleep in the house at night and that can attend the training on using the system

Lives within 40 miles of enrollment site

Exclusion Criteria

I have not had major surgery in the last 30 days.

I have had diabetic ketoacidosis in the last 6 months.

I am not on long-term immunosuppressive medication.

Timeline

Screening

Participants are screened for eligibility to participate in the trial

1-2 weeks

Intervention Period

Participants will use the MPC closed-loop system for 7 days using Fiasp insulin. The first 6 hours will be spent in clinic being trained on the system, then eating a meal. Then the participant will take the system home to continue using for 7 days.

7 days

1 visit (in-person), remainder at home

Control Period

Participants will continue their normal diabetes regimen which includes the t:slim X2 pump with Dexcom G6 CGM and Control IQ for 7 days. Participants will share their pump download and Dexcom Clarity data with study staff after the 7 days is complete.

7 days

1 visit (in-person), remainder at home

Follow-up

Participants are monitored for safety and effectiveness after treatment

1-2 weeks

Treatment Details

Interventions

• MPC closed-loop system
• t:slim X2 pump with Dexcom G6 CGM and Control IQ

Trial Overview The trial is testing a new missed meal bolus detection algorithm in an artificial pancreas system against usual care involving the t:slim X2 pump with Control IQ technology. The goal is to see if this new feature can better manage glucose levels in high-risk patients by automatically delivering insulin when meals are detected but not reported by users.

Participant Groups

Group I: MPC Arm Experimental Treatment 1 Intervention

Participants will use the MPC closed-loop system for 7 days using Fiasp insulin. The first 6 hours will be spent in clinic being trained on the system, then eating a meal. Then the participant will take the system home to continue using for 7 days.

Group II: Control IQ arm Active Control 1 Intervention

Participants will continue their normal diabetes regimen which includes the t:slim X2 pump with Dexcom G6 CGM and Control IQ for 7 days. Participants will share their pump download and Dexcom Clarity data with study staff after the 7 days is complete.

MPC closed-loop system is already approved in United States, European Union, Canada for the following indications:

Approved in United States as t:slim X2 insulin pump with Control-IQ technology for:

• Type 1 diabetes mellitus

Approved in European Union as t:slim X2 insulin pump with Control-IQ technology for:

• Type 1 diabetes mellitus

Approved in Canada as t:slim X2 insulin pump with Control-IQ technology for:

• Type 1 diabetes mellitus

Who Is Running the Clinical Trial?

Oregon Health and Science University

Lead Sponsor

Trials

1,024

Recruited

7,420,000+

University of Washington

Collaborator

Trials

1,858

Recruited

2,023,000+

MultiCare Rockwood Northpointe Specialty Center

Collaborator

Trials

1

Recruited

30+

Findings from Research

The adaptive model predictive control (MPC) strategy using a run-to-run (R2R) approach significantly improved glucose control in type 1 diabetes, increasing time in range by 11.39% and reducing time spent above 180 mg/dl by 48.74% over two months in silico simulations.

This approach effectively managed the large glucose variability typical in type 1 diabetes without increasing the risk of hypoglycemia, highlighting its potential for real-life outpatient use in future artificial pancreas studies.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Toward a Run-to-Run Adaptive Artificial Pancreas: In Silico Results. Toffanin, C., Visentin, R., Messori, M., et al.[2019]

The novel Model Predictive Control (MPC) law for the Artificial Pancreas significantly enhances safety by reducing the risk of hypoglycemia while also improving the correction of hyperglycemia, addressing two critical challenges in diabetes management.

This MPC law was tested in four FDA-approved clinical trials, with the largest involving 29 participants over three months, demonstrating its effectiveness in real-world settings compared to previous strategies.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Velocity-weighting & velocity-penalty MPC of an artificial pancreas: Improved safety & performance. Gondhalekar, R., Dassau, E., Doyle, FJ.[2020]

The proposed enhancements to the artificial pancreas controller significantly improved blood glucose management, increasing the time spent in the safe glycemic range (70-180 mg/dL) from 68.8% to 76.9% without increasing hypoglycemia risk.

These advancements include a personalized insulin-blood glucose model, an asymmetric cost function to prioritize preventing hypoglycemia, and a dynamic insulin-on-board algorithm, all of which were validated through both in silico tests and clinical trials.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Enhanced Model Predictive Control (eMPC) Strategy for Automated Glucose Control. Lee, JB., Dassau, E., Gondhalekar, R., et al.[2019]

The novel Model Predictive Control (MPC) law for an Artificial Pancreas effectively manages blood glucose levels in people with type 1 diabetes by using asymmetric input costs to enhance safety during both hyperglycemia and hypoglycemia.

In a clinical trial involving 32 subjects, the MPC law demonstrated its ability to predictively suspend and resume insulin delivery, showcasing its potential for safe outpatient use in managing diabetes.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Periodic zone-MPC with asymmetric costs for outpatient-ready safety of an artificial pancreas to treat type 1 diabetes. Gondhalekar, R., Dassau, E., Doyle, FJ.[2020]

The model predictive control (MPC) algorithm for an artificial pancreas significantly improved glucose control in outpatients with type 1 diabetes, increasing the percentage of time spent in the target glucose range by 10.03% during overnight use.

The MPC-based system also demonstrated safety by reducing the time spent in the hypoglycemic range by 1.34% over long-term use (more than 1 month), indicating it effectively manages blood sugar levels without increasing the risk of low blood sugar.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Effectiveness and safety of a model predictive control (MPC) algorithm for an artificial pancreas system in outpatients with type 1 diabetes (T1D): systematic review and meta-analysis. Kang, SL., Hwang, YN., Kwon, JY., et al.[2022]

The model predictive control (MPC) algorithm effectively managed blood glucose levels in six Type 1 diabetes patients, successfully bringing hyperglycemia to normoglycemia within 2 to 3 hours during fasting conditions.

Both intravenous and simulated subcutaneous glucose monitoring routes achieved similar glucose control outcomes, maintaining levels within the physiological range, indicating the MPC's adaptability and efficacy in different monitoring scenarios.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

On-line adaptive algorithm with glucose prediction capacity for subcutaneous closed loop control of glucose: evaluation under fasting conditions in patients with Type 1 diabetes. Schaller, HC., Schaupp, L., Bodenlenz, M., et al.[2011]

The proposed adaptive model predictive control (MPC) algorithm for artificial pancreas systems dynamically adjusts insulin delivery based on real-time estimates of plasma insulin concentration, enhancing the system's responsiveness to changing glucose levels.

In simulation studies, the adaptive MPC effectively optimized insulin delivery while incorporating safety constraints to prevent overdosing, demonstrating improved efficacy in managing blood sugar levels without needing manual meal input.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Plasma-Insulin-Cognizant Adaptive Model Predictive Control for Artificial Pancreas Systems. Hajizadeh, I., Rashid, M., Cinar, A.[2020]

Model predictive control (MPC) offers significant advantages for developing a closed-loop artificial pancreas, such as the ability to incorporate constraints on insulin delivery and adapt to various factors like meals and exercise.

The success of an artificial pancreas relies not only on the choice of control algorithm, whether MPC or PID, but also on the engineers' experience in implementing these algorithms in real human studies, highlighting the importance of practical application over theoretical simulation.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Algorithms for a closed-loop artificial pancreas: the case for model predictive control. Bequette, BW.[2021]

References

1. United States pubmed.ncbi.nlm.nih.gov

Toward a Run-to-Run Adaptive Artificial Pancreas: In Silico Results. [2019]

2. England pubmed.ncbi.nlm.nih.gov

Velocity-weighting & velocity-penalty MPC of an artificial pancreas: Improved safety & performance. [2020]

3. United States pubmed.ncbi.nlm.nih.gov

Enhanced Model Predictive Control (eMPC) Strategy for Automated Glucose Control. [2019]

4. England pubmed.ncbi.nlm.nih.gov

Periodic zone-MPC with asymmetric costs for outpatient-ready safety of an artificial pancreas to treat type 1 diabetes. [2020]

5. England pubmed.ncbi.nlm.nih.gov

Effectiveness and safety of a model predictive control (MPC) algorithm for an artificial pancreas system in outpatients with type 1 diabetes (T1D): systematic review and meta-analysis. [2022]

6. England pubmed.ncbi.nlm.nih.gov

On-line adaptive algorithm with glucose prediction capacity for subcutaneous closed loop control of glucose: evaluation under fasting conditions in patients with Type 1 diabetes. [2011]

7. England pubmed.ncbi.nlm.nih.gov

Plasma-Insulin-Cognizant Adaptive Model Predictive Control for Artificial Pancreas Systems. [2020]

8. United States pubmed.ncbi.nlm.nih.gov

Algorithms for a closed-loop artificial pancreas: the case for model predictive control. [2021]