Investigator has moved to another institution

We aim to test the efficacy of a new method for determining individual insulin sensitivity (IS) based on sensor-augmented-insulin pump (SAP) data in order to customize the insulin to carbohydrate ratio (CR) in adolescents with type 1 diabetes (T1D).

To date, the individual insulin sensitivity (IS) could only be investigated by intensive and invasive research techniques that are not feasible to perform in an outpatient setting for pediatric patients with diabetes.

Recently published studies have demonstrated the efficacy of an algorithm to calculate the patient specific insulin sensitivity to customize the CR for adult patients with T1D. The algorithm has been validated in adult patients, however not yet investigated in the pediatric population with T1D.

The aims of our study are:

1. to customize the CR of pediatric subjects with T1D using the individualized insulin sensitivity index (ISind) to improve post-prandial blood glucose control after a standard meal.
2. to test, under free living condition (at home), the efficacy of the customized CR in improving post-prandial glycemic control for pediatric subjects with T1D.

This approach would have at least two potential benefits for pediatric patients with T1D:

1. To provide a non-invasive tool for individualizing their home insulin therapy;
2. To offer a reliable instrument for adjusting the meal bolus of the current hybrid closed loop (HCL) systems to account for the inter-subject variability in insulin action.

A new method to assess insulin sensitivity (IS) has been proposed and investigated by the PI and his group at the University of Padova. The new insulin sensitivity index, named "SISP" is calculated from data derived from insulin pump and continuous glucose monitoring (CGM) uploads. The efficiency of the SISP has been tested in-silico using the University of Virginia/Padova T1D simulator by mimicking a single-meal scenario with patient-specific optimized carbohydrate ratio (CR) (increased or decreased by 20%) and optimal CR. In all the simulations the use of the optimal CR, calculated with the proposed method, has improved the overall glycemic control. The simulator (S2013) used for this purpose has been valdated and is approved by the FDA as a substitute for preclinical trials for insulin treatments, including closed-loop algorithms. It is comprised of data from 100 in-silico patients that represent the biological variability of a generic real diabetic population. Thus, an algorithm that is tuned on the basis of in-silico analysis can be safely implemented in real-life setting.

The method to estimate SISP and to optimize the CR from SAP data, could be easily applied to the daily management of patients with T1D and in a closed-loop context since several closed-loop algorithms, currently used in clinical trials, are based on the pre-programmed open-loop insulin therapy.

Once the individualized SISP is calculated, it can be used to customize the CR using the in-silico tested algorithm to determine an individualized CR (CRIND).

Consequently, the CRIND can be tested in outpatient setting safely, and adjusted in a run-to-run framework, using a well described approach of self-learning, the latter allowing titration of the insulin therapy based on CGM data using a self-learning algorithm as previously described.

• Phase 1: "Control period". This phase represents the control frame-time, during which patients, once enrolled, will use their SAP without any adjustment of the CR and correction factors, according to the parameters recorded at the screening visit. This phase is aimed to record the CGM data and will represent the control period of the study. It lasts 3 weeks.
• Phase 2: "Build-up period" This phase is aimed to obtain an adequate amount of data for customizing the insulin pump parameters to the specific features of each subject according to the proposed algorithms. It consists of three standardized meals (Figure 2) and a run-in phase.

Subjects will be randomly assigned to two different pre-meal insulin CR groups in a 1:1 ratio to determine parameters that will be used to adjust the IS algorithm for pediatric patients with T1D. The post-prandial blood glucose pattern after a pre-meal bolus of CR, CR 20% increased, CR 20% decreased are validated parameters necessary to customize the algorithm for a specific patient population with T1D, therefore subjects will be challenged with two different CRs, depending on the randomization, to collect sufficient data to fine tune the algorithm.

Each subject will go through three meal studies;

Group 1. Meal 1: CR with 20% increase; Meal 2: CR home; Meal 3: CR individualized Group 2. Meal 1: CR with 20% decrease; Meal 2: CR home; Meal 3:CR individualized

• Before the 1st meal subjects will be randomly assigned to receive a modified premeal bolus (increased or decreased of 20%, CR+/-20%) in a 1:1 ratio. This change of the home CR (CRHOME) will allow us to estimate the accuracy and error of the CRHOME and to calculate, according to the described algorithm ((21) and below), the optimized CR (CRIND). This latter will be tested in both groups as third meal at home.
• The three meals are followed by a 7 day- run-in period necessary to collect a minimum amount of CGM data to run the algorithm calculations used in the intervention phase. During this period subjects will adopt the CRIND obtained after the standard meals.
• Phase 3: "Intervention period". This phase is aimed to test the efficacy of the CRIND on mitigating post-prandial hyperglycemia by a home-based study, using periodical adjustment of CR and basal rate according to the "run-to-run" approach. It consists of a self-learning algorithm method to adjust insulin regimen (the basal rate and the CRIND) based on CGM post-meal blood glucose patterns and pre-meal insulin bolus.

Subjects will use the SAP tuned according to the individualized ISR2R and CRR2R obtained from the run-in period, along with an individualized insulin basal rate (BasalR2R). During the run-to-run period subjects will receive weekly revised parameters based on run-to-run algorithm according to data analysis of the past seven days. It will last 3 weeks.

Inclusion Criteria:

• Age 12-21 year old.
• Clinical diagnosis of T1D ≥1 year
• Treatment with sensor augmented pump therapy (insulin pump + glucose sensor) for at least 1 month
• HbA1c <10%
• Be able to comprehend written and spoken English

Exclusion Criteria

• Pregnancy, breast feeding, or plans to get pregnant for the next 12 months
• On medications that affect insulin sensitivity; has other medical conditions known to affect insulin sensitivity;
• Cognitive impairment