Using insulin pumps and sensors to monitor blood glucose has become standard of care for children with diabetes. The continous infusion of insulin in the subcutaneous tissue leads to increased risk of lipohypertrophy that might influence insulin absorption. We are in this study using ultrasound to detect subcutaneous hyperechogenicity to guide rotation of insulin infusion and investigate if avoidance of areas with hyperechogenicity improve insulin absorption seen as decrease in insulin needs.

An increasing proportion of patients with diabetes mellitus type 1 (T1D) is using diabetes devices in order to achieve better glycemic control, improved quality of life and more flexible lifestyle (1-4). Continuos subcutaneous insulin infusion (CSII) or insulin pumps are infusing small doses of insulin by a catheter fastened to the skin by an infusion set. Modern systems with automated insulin delivery (AID) as the precursor to the completely automated artificial pancreas, give algorithm-based corrections on basis of the interstitial fluid glucose measurement from the continuos glucose monitor (CGM). The first studies on these new more automated closed loop systems show improvements in Time in Range (TIR) but still with room for improvement. (5). The absorption of insulin is a very important measure that potentially could vary a lot in relation to both age, sex, timing, exercise, blood flow and a whole range of other aspects, but especially also according to the tissue that insulin is infused into. The skin site used for the infusion set is therefore increasingly important, and this could therefore account for at least some of the glycemic variability seen even in patients using these highly automated systems.

Ultrasound have shown promising results to investigate the subcutaneous areas by means of both density and vascularity. Regarding the density of tissue, hyper- and hypo- echogenicity are important measures. An- other newly study have also shed light on the subcutis and the findings of lipohypertrophy or pre-clinical lipohypertrophy by ultrasound (6). A study from our team similarly found proportions of hyperechogenicity among 60%-75% of all insulin pump users, with as well neovascularization in about 10% of the users (7). The hyper echogenicity is an ultrasonic sign of increased density of the tissue, and whether this is inflammation, lipohypertrophy, or fibrosis is unknown, but since this is seen in most areas used for insulin pumps, the con- sequences of the unspecific increased tissue density on the absorption of insulin is highly relevant.

The risk factors for developing hyperechogenicity besides infusion of insulin is unknown, but the occurrence of eczema and subtle changes in dermis are suspected to be associated with an increased risk of subcutaneous reactions. The subtle nonvisible changes in the skin could be level of natural moisturizing factor (NMF), lipids, cytokines and microbiome. Simultaneously investigation of the skin barrier and subcutis have not been investigated previously, why investigating the interaction between the two is important.

To maximize the effect of insulin pumps, there is therefore a lack of knowledge on the skin sites, especially in means of the consequences of the subcutaneous changes as hyper-echogenicity and vascularization among insulin pump users.

4. Aim The primary aim of present study is therefore to investigate how insulin infusion in hyperechogenic areas of subcutis affects the effect of insulin on blood glucose.

The secondary aim is to investigate the concordance of eczema and wounds and hyperechogenicity.

5. Study design Study type: Prospective longitudinal observational and controlled intervention study

Study design:

The study consists of two phases of two weeks. In the first two weeks they are choosing infusion sites as they used to, but with clear mapping of the different available skin infusion sites of 10 cm2 including the ultrasound measurement, where they report by electronical diary which skin sites are used when. In the next two weeks they are told which skin sites are seen with echogenicity and are instructed to avoid skin sites with hyper- echogenicity in subcutis. The devices will be used in accordance with the CE-label during the full study. In the second phase of two weeks, they as well do report the skin sites by electronical diary.

6. Study population Hybrid closed loop pump users between the age of 6-18 years with at least 6 months of pump history.

7. Study Intervention Study intervention: to avoid infusion sites with hyper echogenicity in subcutis used in controlled manner.

Justification for why randomization is not used: The primary outcome is not to test the intervention but to test the influence on insulin absorption depending on where the infusion set is inserted. Therefore, it is more important to have enough treatment periods in both hyperechogenic and normo-echogenic subcutaneous areas which is how the sample size is tested than to test the intervention which is yet a clinical highly relevant second part of the study.

10. Outcomes and statistical measures Primary analysis compares in paired design with paired t-test the second day of using infusion set in respec- tively areas with and without hyper echogenicity in subcutis. Primary outcome is Insulin/kg/day. Secondary outcomes are: Insulin-dose-adjusted A1C (IDAA1C), Insulin/kg/day/carbohydrate, Coefficients of variants (CV), Time in Range (TIR), Time below range (TBR), Time above range (TAR), Time in Target Range (TTR), glycemic risk index (GRI) and insulin sensitivity estimation based on Schiavon et al (10). Secondary analysis compares the primary and secondary outcomes in a paired t-test for effect of intervention. Further, examine the concordance of cutan and sub-cutan changes and the skin barrier regarding natural moisturizing factors (NMF), microbiome, cytokines, and lipids from the Tape strips.

Other analyses: Descriptive analysis of duration and size of hyperechogenicity and primary analysis of areas with or without vascularization.

For all analysis the second day of using the infusion set will be used to avoid the first 24 hours where the algorithm adjusts.

Adjusting variables: age, gender, diabetes duration, BMI, physical activity and insulin pump use duration

9.1 Sample size calculation: Assumptions: Alpha = 0.05, Beta = 0.2 and an expected increase of 25% in insulin/kg/day at hyperechogenic areas compared to control areas results in necessary data from at least 12 hyperechogenic areas and 12 areas without hyper echogenicity. Hyperechogenic areas are expected to be found in 50% of the participants. A total of 30 participants are planned to be recruited.

11. Ethical considerations The participants will receive information before the study to give both oral and written informed consent to participation in this study. After consent the researchers on the project, sponsor, representatives for the sponsors as well as authorities have access to e-journal information to give proper quality surveillance of the study. The study will be approved by both the Danish Data Agency and The National Committee on Health Research Ethics prior to initiation through a merged application in the Capital Region of Denmark for both Committees. No payment for patients participating in the study will be given. Processing of personal data will agree with the General Data Processing Regulation and the Danish Data Protection Act.

12. Reporting and dissemination Anonymous raw data would be made public after the study and a detailed communication plan would be made to ensure that as many as possible receive the results from this study. Publications in international peer-reviewed journals as well as international and national conference presentations will be carried out irrespective of positive, negative, or inconclusive results. The protocol will be made public at www.clinical- trials.gov prior to study initiation.

Inclusion Criteria:

• type 1 diabetes,
• using CE marked automatic insulin delivery system,
• using insulin pump for at least 6 months

Exclusion Criteria:

• Do not use bolusguide (entering carbs),
• not trained in carbohydrate counting,
• unable to understand and read Danish

Depending on Danish law regarding GDPR and identification of the individual