Type 1 Diabetes (T1D)
Conditions
Keywords
Automated Insulin Delivery Systems
Brief summary
Automated insulin delivery (AID) systems are the current gold standard for the management of type 1 diabetes (T1D), improving glycemic control, reducing hypoglycemia, and decreasing treatment burden. Although both commercial AID (C-AID) and open-source AID (OS-AID) systems have demonstrated significant clinical benefits, direct randomized comparisons between these systems remain scarce. Moreover, an important limitation of currently available AID systems is their reliance on user-initiated meal announcements and carbohydrate counting, which remain major contributors to postprandial dysglycemia and treatment burden. While emerging evidence suggests that AID systems may partially compensate for missed meal announcements, their comparative performance under these challenging real-world conditions is unknown. The NOMAD study is an international, multicenter, open-label, randomized, non-inferiority crossover trial designed to compare an open-source AID system with commercially available AID systems (including Tandem Control-IQ and Omnipod 5) in adults with type 1 diabetes currently using an open-source AID system. The study specifically evaluates glycemic outcomes following standardized unannounced meal challenges performed under free-living conditions. A total of 33 participants will be enrolled and complete two 4-week intervention periods, each consisting of a 2-week run-in phase followed by a 2-week data collection phase, with crossover between treatment arms. Participants will continue using their own Dexcom G6 or Dexcom G7 continuous glucose monitoring system throughout the study. The primary outcome is the percentage of time spent in the target glucose range (3.9-10.0 mmol/L) during the 4 hours following standardized unannounced meal challenges. Secondary outcomes include additional CGM metrics (time above and below range, glucose variability, mean glucose, and GMI), insulin delivery characteristics, and patient-reported outcomes evaluating treatment satisfaction, usability, and diabetes-related burden.
Detailed description
AID systems have become the preferred treatment for T1D, improving glycemic control, reducing hypoglycemia, and decreasing treatment burden compared with conventional insulin therapy. Both C-AID and OS-AID systems have demonstrated substantial clinical benefits, including increased time in range and improved patient-reported outcomes. However, despite their growing adoption, few randomized controlled trials have directly compared these systems under standardized conditions. One of the major remaining challenges for current AID systems is the management of meals. Most commercially available systems are hybrid closed-loop systems that require users to announce meals and estimate carbohydrate intake before eating. Inaccurate carbohydrate estimation, delayed boluses, or missed meal announcements are common in daily life and contribute substantially to postprandial hyperglycemia and overall treatment burden. Although AID algorithms can partially compensate for these situations through automated insulin adjustments, their effectiveness varies between systems and has not been directly compared under real-world conditions. This study aims to compare the performance of open-source and commercial AID systems during standardized unannounced meal challenges performed under free-living conditions. Adults with T1D who are experienced users of an OS-AID system will participate in a randomized, open-label, crossover trial in which they will use both their usual OS-AID system and a commercially available AID system. Throughout the study, participants will use the same Dexcom continuous glucose monitoring system to ensure consistent glucose assessment across study periods. The primary objective is to determine whether OS-AID is non-inferior to commercial AID for maintaining postprandial glucose control following unannounced meals, as measured by the percentage of time spent in the target glucose range during the 4 hours after meal initiation. Secondary objectives include comparing overall glycemic outcomes, glucose variability, insulin delivery characteristics, and patient-reported outcomes related to treatment satisfaction, diabetes burden, and usability. The study will also explore differences in insulin dosing behavior, including user-initiated boluses, between the two AID approaches. By evaluating glycemic performance in a common real-world scenario where meal announcements are omitted, this study will provide evidence regarding the comparative effectiveness, safety, and user experience of open-source and commercial AID systems. The findings are expected to inform clinical decision-making, support individualized selection of AID systems, and contribute to the development of future automated insulin delivery technologies that further reduce the burden of diabetes self-management.
Interventions
DEVICEOpen-Source Automated Insulin Delivery (OS-AID)
Participants will use their own OS-AID system, including Loop, AndroidAPS, or OpenAPS, in combination with a compatible insulin pump and Dexcom G6 or Dexcom G7 continuous glucose monitoring (CGM). These systems use open-source algorithms to automatically adjust insulin delivery based on real-time CGM data through adaptive basal insulin modulation and, depending on the platform, automated correction boluses. Participants will continue using their established OS-AID settings throughout the intervention period without software updates and will be asked to perform two unannounced meal-challenges.
DEVICECommercial Automated Insulin Delivery (C-AID)
Participants will use a commercially available hybrid closed-loop AID system compatible with Dexcom G6 or Dexcom G7 CGM. Depending on study allocation and participants' usual device, the commercial system will include Tandem t X2 with Control-IQ technology or Omnipod 5. Participants not already using a compatible commercial AID system will receive study equipment and standardized training before initiating the intervention. Commercial AID systems automatically adjust insulin delivery based on continuous glucose measurements using manufacturer-specific algorithms while continuing to recommend user-initiated meal announcements and boluses for optimal performance.
Sponsors
Institut de Recherches Cliniques de Montreal
Stanford University
University of Alberta
Study design
Allocation
RANDOMIZED
Intervention model
CROSSOVER
Primary purpose
TREATMENT
Masking
NONE
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Males and females ≥ 18 years old currently residing in Canada or the United States of America (USA) * Having a clinical diagnosis of T1D for at least one year (based on the investigator's judgment; C peptide level and antibody determinations are not needed.) * Having been on OS-AID therapy for at least 2 weeks and using a Dexcom G6 or G7 CGM. * If using an ultra-rapid acting insulin, be willing to switch to a rapid-acting insulin, as the former are not recommended in the C-AID systems used in the study (due to the risk of cannula or pod blockage). * Accepting that personal pump setting parameters will be collected by the research team. This access will be limited to the study period.
Exclusion criteria
* Using regular insulin (Novolin ge Toronto or Humulin R) or U200 or U500 concentrated insulin (e.g. Humalog U200, Entuzity U500) * Participant-reported clinically significant nephropathy (eGFR \< 25 ml/min/1.73m2, planned or on dialysis), neuropathy (e.g., known uncontrolled gastroparesis) or retinopathy (e.g., proliferative retinopathy with ongoing active treatment such as laser photocoagulation or planned surgery) as judged by the investigator * Recent (\<3 months) acute macrovascular event (e.g., acute coronary syndrome or cardiac surgery) * Anticipated therapeutic change (including change of CGM sensor \[other than Dexcom sensors\] or AID system type, new antidiabetic drug initiation) between admission and end of the study * % of time spent out of automated mode exceeding 10% in last month's trial start * Anticipated need to use hydroxyurea during the study period (as it can interfere with CGM readings) * Pregnancy (ongoing or current attempt to become pregnant) * Breastfeeding * Plan to go abroad in a foreign country during the study period * Severe hypoglycemic episode within one month of screening * Severe hyperglycemic episode (including DKA) requiring hospitalization in the last 3 months * Current use of glucocorticoid medication (except low stable dose and inhaled steroids and stable adrenal insufficiency treatment e.g., Cortef®) * Current use of SGLT-2 inhibitors or GLP1 agonists or other antidiabetic agents (Metformin, DPP-4 inhibitors) unless at a stable dose for at least 3 months, without anticipated change during the study and appropriate ketone testing is performed (for iSGLT2 treatment). * Known or suspected allergies to study products (e.g., Dexcom adhesive) * Other serious medical illnesses likely to interfere with study participation or with the ability to complete the study by the judgment of the investigator * Anticipation of a significant change in exercise or diet regimen between admission and end of the study (i.e., starting or stopping an organized sport; planned significant diet change) * Anticipated radiologic examination at the time of study assessment incompatible with CGM wear (e.g., MRI) * In the opinion of the investigator, a participant who is unable or unwilling to observe the contraindications of the study device.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Percentage of time in target glucose range during the 4 hours following unannounced meal challenges | During the final 2 weeks of each 4-week intervention period (following the run-in phase). | Percentage of time with sensor glucose between 3.9 and 10.0 mmol/L (70-180 mg/dL) during the 4 hours following standardized unannounced meal challenges, measured using continuous glucose monitoring (CGM). Data will be analyzed at the meal level using all meal challenges completed during each intervention period. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Time in tight range | Final 2 weeks of each intervention period. | Percentage of time with sensor glucose between 3.9 and 7.8 mmol/L (70-140 mg/dL), measured by CGM. |
| Time Above Range (>10 mmol/L) | Final 2 weeks of each intervention period. | Percentage of time with sensor glucose \>10.0 mmol/L (\>180 mg/dL), measured by CGM. |
| Time Above Range (>13.9 mmol/L) | Final 2 weeks of each intervention period. | Percentage of time with sensor glucose \>13.9 mmol/L (\>250 mg/dL), measured by CGM. |
| Mean Glucose | Final 2 weeks of each intervention period | Mean sensor glucose concentration (mmol/L) measured using CGM. |
Countries
Canada
Contacts
CONTACTJoséphine Molveau, Ph.D
CONTACTValérie Boudreau, DtP, PhD
PRINCIPAL_INVESTIGATORRémi Rabasa-Lhoret, M.D. Ph.D
IRCM
Outcome results
None listed