More and more children and adolescents are using diabetes devices attached to the skin. The attachment of infusions sets to the skin provoke allergenic or toxic eczema, the continous infusion of insulin provoke subcutaneous changes and prolonged wear time seems to increase the risk of these complication. On the other site fewer skin stripping episodes with longer wear time and filtration of the insulin may be beneficial for the skin, therefore comparing the occurrence of subcutaneous hyper echogenicity, eczema and the skin barrier in users of 3 and 7 days infusions set is highly relevant. The primary aim of present study is to investigate if the use of three days insulin infusion set is superior to seven days in preventing the occurrence of hyperechogenicity in areas recently used for insulin infusion and if the type of insulin matter comparing insulin aspart - Novo Rapid, or insulin lispro - Humalog

Background 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. Continuous 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 as the precursor to the completely automated artificial pancreas, give algorithm-based corrections on basis of the interstitial fluid glucose measurement from the continuous 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. Several studies have observed better outcome in those treated with Medtronic 780g using the most aggressive targets (blood glucose target of 5.5 mmol/l and insulin action time of 2 hours). The ability of change in settings to predict future outcomes is though not explored and the importance of settings versus behavior is unknown. One type of behavior is timing of shift of insulin infu-sions set, prolonged use is associated with higher blood glucose, probably caused by decrease in absorption. Other studies indicate that the depth of subcutaneous tissue may also be important, since buttocks seems to be better than abdomen in small children 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 im-portant, and this could therefore account for at least some of the glycemic variability seen even in pa-tients using these highly automated systems. New infusions set with extended wear time have been developed and tested in adults showing good results although tendency to increase in sensor glucose after 6 days. The extended wear time have not been tested in children and none have examined the extend of lipohypertrophy with extended wear time or the importance of skin depth.

Ultrasound have shown promising results to investigate the skin depth as well as subcutaneous areas by means of both density and vascularity. Regarding the density of tissue, hyper- and hypo- echogenicity are important measures that might reflect different grades of lipohypertrophy. Lipodystrophy (lipo-hypertrophy and lipoatrophy) are a well-known complications to insulin infusion. The importance of type of insulin has only been investigated in cross-sectional designs finding no differences related to insulin type. Just one old study showed a positive effect of changing from fast-acting to rapid-acting insulin on lipohypertrophy. Similar there is low evidence for the importance of insulin type on lipoatrophy which is also a rare condition. Current guideline recommend change in type of insulin with lipoatrophy whereas if experiencing lipohypertrophy patients are asked to find new areas for infusion. Another newly study have also shed light on the subcutis and the findings of lipohypertrophy or pre-clinical lipohypertrophy by ultrasound. A study similarly found proportions of hyperechogenicity among 60%-75% of all insulin pump users, with as well neovascularization in about 10% of the users. These findings that may reflect fibrosis, inflammation, or lipohypertrophy seems im-portant to investigate and test how to prevent them from arising.

The attachment of infusions sets to the skin provoke allergenic or toxic eczema and prolonged wear time seems to increase the risk of this complication. On the other site fewer skin stripping episodes with longer wear time may be beneficial for the skin, therefore comparing the occurrence of eczema and the skin barrier in users of 3 and 7 days infusions set is highly relevant. The subtle nonvisible changes in the skin could be level of natural moisturizing factor (NMF) which can be investigated by tape strips. The skin barrier can also be studied by electric impedance spectroscopy (EIS) a new method that have been shown to detect skin barrier dysfunction in normal looking non-lesional skin of children and predict who is more likely to develop atopic dermatitis. It is able to discriminate between skin without visible changes and healthy skin as well as show a significant increase in EIS that correlated with healing. EIS has also been used to study irritancy from detergents which may be a good surrogate for skin irritation at diabetes treatment sites. Recent studies have been in animal models but historically irritation ( e.g. from sodium laural sulfate has been studied quite extensively with EIS in humans. The microcirculation may be an important marker of skin recovery and skin inflammation seen as hyperemia. Tissue Viability Imaging (TiVI) uses polarization light spectroscopy to study micro vascularization by esti-mating the red blood cell (RBC) concentration in the skin. The method is very robust and offers instantaneous capturing of data and low variability, though normal values for children and different areas are scarce. The non-invasive properties make EIS and TiVi promising new tools for detecting early signs of skin impairment and changes that could reveal if skin have been impaired by the infusions set and guide when to reuse a site.

4. Aim The primary aim of present study is to investigate if the use of three days insulin infusion set is superior to seven days in preventing the occurrence of hyperechogenicity in areas recently used for insulin infusion and if the type of insulin matter comparing insulin aspart - Novo Rapid, or insulin lispro - Humalog. Second-ly, to estimate the glycemic variables/day of using the infusions set within the recommended period or outside recommendations (> 72 hours for three-days set and > 168 hours for seven-days set). Thirdly, to investigate occurrence of eczema and skin barrier changes depending on the use of three- or seven-days infusions set.

5. Study design Study type and design: Randomized controlled intervention study with crossover between 3 and 7-days infusions set and randomised to Novorapid or humalog

6. Study population: Users of Medtronic 780G age 7 and 18 years of age

7. Study Intervention The rationale behind the intervention is to prevent the occurrence of hyperechogenicity as well as eczema by comparing type of infusions set and the occurrence of inflammation and subcutaneous changes.

The investigators uses a non-blinded randomization since it is impossible to blind type of infusion set. Two types of insulin are included since the manufacturer uses different preservatives and normally Lispro is initiated when there is lipoatrofi when on Novorapid (currently the type of insulin with subsidies), indicating they may have different in-flammatory properties.

10. Outcomes and statistical measures Primary analysis: Hyper echogenicity at the infusion sites identified by ultrasound as a 4 level variable (None, hyperechogenicity, hypoechogenicity and new vascularization all registered as yes/no) is the primary outcome. Primary analyses will be a paired analysis of the frequency of hyper echogenicity at the infusion sites after 4 months between the cross over design of three-days vs. seven-days infusion sites comparing the same type of insulin. Secondary analyses will be unpaired analysis of the frequency of hyper echogenicity at the infusion sites after 4 and 8 months between the different insulin types adjusted for type of infusions set used during the observed period.

Confounders or adjusting variables: age, sex, diabetes duration, BMI, type of equipment, Filaggrin

9.1 Sample size calculation Assumptions: Alpha = 0.05, Beta = 0.2 and a clinical meaningful difference in the occurrence of hyperechogenicity is 20% the study aim at 39 in each group to show a significant difference. If the outcome is mean glucose per day assuming a mean of 8.3 mmol/l, a clinical meaningful difference is 0.5 mmol/l and SD is 0.8 mmol/l the study aim at 42 in each group to show this difference. Because of the repeated shift and many observations per patient the power is probably better than shown here why the study aim of including 80 patients with 20 in each of the four groups.

12 Risks and adverse events Participation in this study is not assumed to have any significant risks of adverse events caused by the study. The tests to be done do not have any risks either, Hba1c and Filaggrin will be collected simultane-ously with other blood-tests planned and participation will thereby not cause extra blood tests or any pain. The primary cost for the participants is the demand for extra time from the visits (+ 30 minutes). The use of ultrasound is non-invasive with no risks related, but only skin touching. The EIS measurements will be taken using a Nevisense® device (SciBase) is painless and leave no scar on the skin. Use of the TiVi (Tissue Viability Imaging) system for investigating microvascular event demands no injections, contrast agents, or skin contact. It is a photography showing blood flow. This implies that there is no risk or discom-fort to the test subject.

Both infusion sets and insulin types are approved and a part of the standard of care. The insulin Lispro though is second choice only when there are problems like lipoatrofi when using Novorapid, therefore no additional risks are expected according to this intervention design. Proper reporting of adverse events in means of both insulin type and infusion set type are included in the design.

13. Ethical considerations The participants will receive information before the study to give both oral and written informed consent to participation in this 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. Pro-cessing of personal data will agree with the General Data Processing Regulation and the Danish Data Pro-tection Act.

This project includes children and adolescents from age 7 to 18 years at recruitment. It is important to include children since the skin barrier and healing properties differ according to age, why results from adolescents and adults cannot be extended to younger children. The skin undergoes changes during pu-berty, why differential results regarding skin barrier are expected and must be brought out between pre- and postpubertal children. No risks and possible adverse events will be caused by this study in means of the development of the child or adolescent. Information both oral and written will be given both parents prior to consent and written consent from both parents or other parent authority for patients under 18 years, as well as from the patient itself when between 15-17 years and only from the patient when over 18 years. An independent consent will be obtained from the participant itself during the study, when the adolescent gets 15 or 18 years old. Besides an oral expression from the patient up to 15 will be collected when appropriate and when child do understand the purpose of study. The informing health care pro-fessional and study coordinator do have relevant experience with children and adolescents to give the information properly for different age groups and is well-experienced in doing so from previous research projects in pediatric patients.

Participants in the study are all patients at SDCC and are therefore covered by the Patient Insurance.

Inclusion Criteria:

• Diagnosed with T1D
• Diabetes duration of more than 6 months prior to inclusion.
• Currently using the CE-marked hybrid closed-loop system - Medtronic 780G pump and the corresponding sensor with automated insulin delivery.
• Planning to using the CE-marked hybrid closed-loop system - Medtronic 780G pump and the corresponding sensor with automated insulin delivery.
• Being between 7 and 18 years of age prior to inclusion
• Insulin needs per day above 8 units

Exclusion Criteria:

• Those who are unable to read and understand Danish
• Those with impaired cognitive development that may interfere with the ability to answer questionnaires in Danish and/or be reached by phone or videocall

Not allowed according to Danish law