ATTD 2025 Oral Abstracts

LEVERAGING T1DEXI STUDY TO DEVELOP DEEP LEARNING MODELS FOR PREDICTING BLOOD GLUCOSE LEVELS

Andrea Calzavara, Francesco Prendin, Giacomo Cappon, Andrea Facchinetti

University of Padova, Department Of Information Engineering, Padova, Italy

Background and Aims: Deep learning (DL) models are rapidly gaining popularity and are increasingly being integrated into advanced decision support systems and artificial pancreas technologies to enhance Type 1 Diabetes (T1D) management. In this context, large datasets play a crucial role for developing accurate blood glucose (BG) forecasting algorithms. Leveraging the extensive T1DEXI dataset, this study aims to develop and evaluate advanced DL algorithms for BG prediction.

Methods: The dataset includes 497 adults from a 4-week real-world trial of at-home exercise sessions. Participants provided data on exercise, food intake, administered insulin, heart rate, and continuous glucose monitoring (CGM) data. We investigated Long Short-Term Memory (LSTM), Convolutional Neural Network-LSTM (CNN-LSTM), and CNN-Transformer models, using an AutoRegressive with eXogenous inputs (ARX) model as baseline comparator. Models employed as input historical CGM measurements, meal logs, and insulin data to forecast BG levels at 30, 45, 60, and 90 minute prediction horizons (PHs). In addition to RMSE, time gain (TG) is used to quantify the time advantage provided by the algorithms, enabling patients to take preemptive actions against impeding adverse events.

Results: presented in Table 1 indicate that the CNN-LSTM and LSTM models achieved the best performance, exhibiting the lowest RMSE and highest TG across all PHs. However, all DL models showed comparable performance, consistently outperforming the ARX model (p-value<0.03).

Conclusions: The T1DEXI dataset enables the development and evaluation DL models that deliver accurate predictions and outperform the ARX model across all PHs.

FLEXIBLE BLOOD GLUCOSE CONTROL: OFFLINE REINFORCEMENT LEARNING FROM HUMAN FEEDBACK

Harry Emerson ¹, Sam Gordon James¹, Ryan Mcconville¹, Matthew Guy²

¹University of Bristol, Bristol, United Kingdom,²University Hospital Southampton, Southampton, United Kingdom

Background and Aims: Reinforcement learning (RL) has shown success in automating insulin dosing in simulated type 1 diabetes (T1D) patients, but is unable to incorporate patient preferences and expertise into T1D decision-making.

Methods: This work proposes a novel method for creating flexible RL blood glucose controllers that adapt to diverse patient instructions. This approach utilises ‘reward sketching’, where patients retrospectively label their glucose and insulin data, identifying desired glucose levels or patterns and the insulin actions they think contribute to positive outcomes. These preferences are then used to fine-tune a pre-trained RL controller to meet patient-specific goals.

Results: The method is evaluated using the UVA/Padova simulator and tested against a diverse set of simulated preferences, including increasing time-in-range (TIR), reducing glycemic variability (CoV) and minimising time-below-range (TBR). Results are compared to a proportional-integral-derivative (PID) controller, tuned to best achieve the specified goal. The RL controller outperforms the PID, achieving greater TIR, CoV and TBR, while obtaining a (17+/-7)% reduction in glycemic risk. The method does not require patients to know how to achieve their goals, only to be able to highlight examples of their goal within their historical data. However, the technique can also use patient knowledge, in the form of simple instructions, to further improve glucose control in these scenarios. For example, instructions on identifying compression lows or pre-empting mealtimes resulted in (1.2+\-0.2)% reduced CoV post-compression and (6+\-3)% increased time-in-range post-meal.

Conclusions: This work presents the first method for effectively incorporating patient preferences into RL-based blood glucose controllers in T1D.

ADDRESSING INDIA’S DIABETES CRISIS WITH SUGARFIT’S IN-HOUSE GPT-DRIVEN CHATBOT INNOVATION

Shivtosh Kumar, Annie R, Mridul Maheshwari, Chhavi Mehra

Ragus Healthcare Pvt Ltd, Bengaluru, India

Background and Aims: India, with over 77 million people living with diabetes, faces the world's highest diabetes burden. High healthcare costs, limited access, and lack of awareness pose significant challenges. To address this, Sugarfit developed a GPT-powered chatbot delivering personalized diabetes care with instant access to expert advice, diagnostic analysis, diet plans, and CGM insights.

Methods: The Sugarfit chatbot focuses on five key areas: Instant Diabetes Q&A: Utilizes users' medical history, reports, and vitals, along with evidence-based health information, to provide personalized responses to diabetes-related queries. Diagnostic Report Analysis: Users can upload reports for personalized assessments and precise unique recommendations for improvements in nutrition, fitness, and mindfulness. Personalized Diet Plans: The chatbot generates tailored diet plans based on individual health profiles, eliminating the need for costly consultations. Personalized Fitness Plans: Custom fitness plans are provided based on users' health conditions, activity levels, and goals. CGM Data Analysis: The chatbot processes CGM data, offering insights into glucose trends, nocturnal patterns, and glycemic variability, along with actionable feedback.

Results: In previous studies with Sugarfit’s virtual coaching, 25,011 participants showed significant improvements over three months. The average reduction in HbA1c was 1.27±1.4% (p <0.001), fasting blood sugar (FBS) levels improved by 30.67±57.8 mg/dL (p <0.001), and weight decreased by 1.0±3.5 kg (p <0.01). These findings demonstrate the effectiveness of personalized coaching. The chatbot aims to scale this success, offering cost-effective, AI-driven diabetes care.

Conclusions: Sugarfit’s GPT-powered chatbot is an innovative solution to India’s diabetes crisis, making care more accessible and affordable, with potential for global impact.

LIVE ON STAGE: THE CREATION OF AN AGP ANALYSIS AI BOT USING ZERO CODE

Leon Tribe

Baker Heart and Diabetes Institute, Melbourne, Australia

Background and Aims: According to the World Health Organization, half the world lacks access to essential health services. Even in advantaged regions, the demand for health care services outstrips supply or is simply unaffordable. The health care industry must get creative if it is to meet these challenges. The aim of this presentation is to show how modern AI technologies can be part of the solution. These services can be created with little to no technical skills and with input from health care professionals to ensure quality is maintained. To show how simple this is to do, the setup of the bot will be shown in real time as part of the presentation.

Methods: Taking the example of the Ambulatory Glucose Profile, this presentation will show how modern AI tools can be used to analyse a person with diabetes' profile and provide near real time analysis and suggestions. Azure AI services will be configured, live on stage.

Results: By the end of the presentation the audience will see how simple the technologies are to set up and how they are applied to medical data and imagery to provide guidance for diagnosis and treatment. The system can be used for training students, providing support for inexperienced practitioners (e.g. GPs unfamiliar with AGP graphs) or as a triage filter to identify high risk individuals for follow-up with a health care practitioner.

Conclusions: Using low-code AI tools like those in this presentation, we can provide automated health care services, addressing service deficits in developed and developing countries, without compromising quality.

SEVERAL SUPERVISED LEARNING ALGORITHMS USED TO ESTIMATE INGESTED GLUCOSE BY TYPE 1 DIABETIS MELLITUS PATIENTS T1DM

Rodolfo Villamizar Mejia, Edward Rodriguez, David Padilla

Universidad industrial de santander, Santander, Bucaramanga, Colombia

Background and Aims: Different blood glucose control approaches for Type 1 Diabetes Patients (T1DM) have been reported as contributing to the development of an artificial pancreas (AP). However, a drawback to obtaining similar performances to those of a health person is that glucose ingested by T1DM patients cannot be directly measured.Thus, this work compares several supervised learning algorithms trained to estimate the ingested glucose, based on predicting future blood glucose profiles once glucose is ingested by the T1DM patient.

Methods: Several supervised learning algorithms (Naïve Bayes, Trees, KNN, SVM, Discriminant Analysis, Long Short-Term Memory Recurrent Neural Networks LSTM-RNN, GAN and GRU) were trained to predict glucose curves within a postprandial horizon and then they are compared with measured or simulated glucose data for T1DM patients, when different glucose quantities are ingested. Then, based on the predicted curve the quantity of glucose ingested by the patient is estimated.

Results: The algorithms were trained by using data from i) the database known as OhioT1DM or ii) in silico adult patients of the UVA/PADOVA Type 1 Diabetes Simulator. For first training the best algorithm was SVM with MSE (379.66) and MAE (0.67), while for the second training the best is GRU with MAE (0,2879), MSE (0,08) and Error (2,2489).

Conclusions: The eight supervised learning algorithms trained can predict postprandial glucose profiles and then ingested glucose by T1DM patients, based on performance indicators. However, this capacity varies significatively by data quality (real or in-silico patients) and when intake is detected, which is the moment in that algorithm start to estimate.

NON-INVASIVE BIOPSY DIAGNOSIS OF DIABETIC KIDNEY DISEASE VIA DEEP LEARNING APPLIED TO RETINAL IMAGES

Shujie Yu ¹, Zhouyu Guan¹, Jie Shen², Carol Y. Cheung³, Gavin Siew Wei Tan⁴, Yih-Chung Tham⁴, Ching-Yu Cheng⁴, Charumathi Sabanayagam⁵, Lee-Ling Lim⁶, Weiping Jia¹, Huating Li¹, Bin Sheng⁷, Tien Yin Wong⁸

¹Department of Endocrinology and Metabolism, Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine., Shanghai, China, ²Medical Records and Statistics Office, Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine., Shanghai, China, ³Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong., Hong Kong, China, ⁴Singapore Eye Research Institute, Singapore National Eye Centre., Singapore, Singapore, ⁵Ophthalmology and Visual Sciences Academic Clinical Program (Eye ACP), Duke-NUS Medical School., Singapore, Singapore, ⁶Department of Medicine, Faculty of Medicine, University of Malaya., Kuala Lumpur, Malaysia, ⁷Shanghai Jiao Tong University, Shanghai, China, ⁸Tsinghua Medicine, Beijing Tsinghua Changgung Hospital, Tsinghua University., Beijing, China

Background and Aims: Improving the accessibility of screening diabetic kidney disease (DKD) and differentiating isolated diabetic nephropathy (DN) from non-diabetic kidney disease (NDKD) are two major challenges in managing individuals with diabetes. We aimed to develop a deep learning system to detect DKD and differentiate isolated DN from NDKD, using retinal images as inputs.

Methods: We developed a retinal image-based AI-DL system——DeepDKD. First, for DKD detection, we used 486,312 retinal images for development and internal validation, and ten multi-ethnic datasets (65,406 participants) for external validation. Second, to differentiate isolated DN from NDKD, we used retinal images from 267 participants who underwent invasive kidney biopsies and three multi-ethnic datasets (244 participants) for development and validation. Finally, we conducted two proof-of-concept studies: a prospective real-world study to evaluate the effectiveness of DeepDKD in screening DKD; and a longitudinal analysis of the effectiveness of DeepDKD in differentiating isolated DN from NDKD on renal function changes with 4.6-year follow-up.

Results: For detecting DKD, DeepDKD achieved AUCs of 0.792-0.826 across the internal and external validation datasets. For differentiating isolated DN from NDKD, DeepDKD achieved AUCs of 0.733-0.906. In the prospective study, compared to the metadata model, DeepDKD could detect DKD with higher sensitivity (P<0.0001). In the longitudinal study, participants of isolated DN and those of NDKD identified by DeepDKD had a significant difference in renal function outcomes over 4.6 years (P=0.001).

Conclusions: Among multi-ethnic populations with diabetes, a retinal image-based AI-DL system showed good performance in detecting DKD and differentiating isolated DN from NDKD, demonstrating its potential for clinical practice.

ADVANCED HYBRID CLOSED LOOP INSULIN DELIVERY SYSTEMS VS. OPEN-LOOP INSULIN DELIVERY DURING RAMADAN IN INDIVIDUALS WITH TYPE 1 DIABETES. THE HYBRIDRAM RANDOMIZED CONTROLLED TRIAL

Mohammad Almehthel ^1,2, Ali Alqahtani², Nouf Aboalsamh², Imad Brema², Ibrahim Alghofaili³, Amal Aljohani³, Mohammed Ahmed², Ahmad Alobedallah², Saad Alzahrani², Nawaf Sharahili², Hala Alsedrah², Razan Alrumaih², Wessal Alshamrani², Mohammad Hassanein⁴

¹University of British Columbia, Division Of Endocrinology, Vancouver, Canada, ²King Fahad Medical City, Obesity, Endocrine And Metabolism Center, Riyadh, Saudi Arabia, ³Prince Sultan Military Medical City, Riyadh, Saudi Arabia, ⁴Dubai Hospital, Dubai, United Arab Emirates

Background and Aims: Advanced Hybrid Closed Loop (AHCL) systems, which adjust insulin delivery based on real-time glucose readings, may offer advantages over Open-Loop Insulin Delivery (OLID) during Ramadan fasting, a period that poses unique challenges for individuals with type 1 diabetes (T1D). This study aimed to compare the efficacy of AHCL and OLID during Ramadan.

Methods: A dual-center, open-label, randomized controlled trial was conducted among individuals with T1D observing Ramadan fasting. Participants were randomized to either AHCL (Medtronic 780G or Tandem Control IQ) or continue OLID. The primary outcome was the difference in Time in Range (TIR) between groups during the 30-day fasting period. Secondary outcomes included Time Below Range (TBR), Time Above Range (TAR), and a composite outcome of TIR ≥ 70%, TBR < 5%, and ≤ 2 Fast Broken Days.

Results: The AHCL group (N=44) had significantly higher median TIR (72.6% vs. 55.8%), lower TBR (1.4% vs. 3.2%), and lower TAR (26% vs. 41%) compared to the OLID group (N=55), all p<0.001. The composite outcome was achieved by 50% of the AHCL group vs. 10.9% of the OLID group (p<0.001).

Conclusions: AHCL systems significantly improved TIR, TAR, and TBR, offering better glycemic control during Ramadan fasting compared to OLID, enhancing safety and quality of life.

EVALUATION OF GLYCAEMIC OUTCOMES SELECTING TWO DIFFERENT PERIODS IN PEOPLE WITH TYPE 1 DIABETES USING AID SYSTEMS

Lucia Babiano-Benito ¹, Francisco Javier Arroyo Díaz², Estela Gil-Poch², Mercedes Fernández-Bueso¹, Marta M. Guijarro-Chacón¹, Jorge González-Prieto¹, Laura M. Martín-Flores¹, Pilar Beato-Víbora¹

¹Hospital Universitario, Badajoz, Spain, ²Badajoz University Hospital, Department Of Paediatrics, Badajoz, Spain

Background and Aims: The time required to achieve representative glucose outcomes in type 1 diabetes people is uncertain. Previous work has shown that 14 days of continuous glucose monitoring data provides a good approximation of glucose metrics. The objective of this study was to evaluate whether there are differences between the glucometrics obtained after 14 and 30 days of sensor glucose data.

Methods: A cross-sectional study was performed. The glycaemic outcomes of the individuals with type 1 diabetes who were treated with the same AID system (MiniMed™ 780G) and with data available in the CareLink System platform in the last year were evaluated. GMI, TIR (70-180mg/dl), time in hypoglycemia < 70 mg/dl and < 54 mg/dl, time in hyperglycemia > 180 mg/dl and > 250 mg/dl were recorded. Additionally, number of hypoglycemia episodes and hyperglycemic episodes per 14 days were computed. Two periods of time, 14 and 30 days were compared.

Results: 452 subjects were included, age: 34 ± 16 years (7-69 years), 53% (n= 239) females. Differences in glycaemic control are summarised in Table 1.

Conclusions: For reliable estimation of hypoglycemia episodes < 70 mg/dl, hyperglycemia episodes > 180 mg/dl, and time > 250 mg/dl (%) 30 days duration, instead of 14 days, may be required. To evaluate TIR, time 180-250 mg/dl, time 54-70 mg/dl, time <54 mg/dl, and GMI (%) either of the periods could be equivalent.

IMPROVEMENT OF GLYCEMIC PARAMETERS AND PARENTAL QUALITY OF LIFE METRICS AFTER THERAPY CHANGE FROM PLGM TO AID IN PRESCHOOLERS

Torben Biester ¹, Kerstin Remus¹, Sarah Biester¹, Heike Saßmann², Britta Klusmeier¹, Olga Kordonouri¹, Felix Reschke¹, Mareike Niemeyer¹, Jantje Weiskorn¹, Thekla Von Dem Berge¹

¹AUF DER BULT Hospital for Children and Adolescents, Diabetes Center, Hannover, Germany, ²Hannover Medical School, Department For Medical Psychology, Hannover, Germany

Background and Aims: In Germany, AID for preschool children is available since May 2022. Children with onset of T1D before that date received predictive low glucose management (PLGM) therapy. The aim of this study was to compare glycemic parameters and parental person-related outcomes (PRO) in a cohort of preschoolers initiating AID therapy.

Methods: All families of preschool-aged children not using AID in the outpatient clinic AUF DER BULT were offered to switch to mylife loop (CamAPS FX with Ypsopump and Dexcom G6 sensor). Families were provided with a pump, compatible phone and all necessary consumables. During the first three months, families could reach out at any time for support and guidance. Over a three months extension, only one regular outpatient visit took place. Anthropometric, glycemic parameters and PRO measures by validated questionnaires were evaluated before a structured AID training and after 3 and 6 months of use. Primary endpoint was 14-day time in range (TIR_{70-180 mg/dl}[%]) after 3 months compared to a 14-day blinded CGM session before switching therapy. Paired t-tests were performed to assess statistical significance.

Results: A total of 27 preschoolers with PLGM therapy participated in this study. TIR increased from 52 to 64%(p<0.001) at primary endpoint and did not change after extension phase. Further results are provided in table. After the study 24/27 participants continued the system.

Conclusions: These data indicate that glycemic control can sustainably improve after switching from PLGM to AID and be accompanied by improvement of parental PRO metrics and therapy confidence.

AUTOMATED INSULIN DELIVERY IS THE GOLD STANDARD IN CHILDREN AND YOUTH WITH TYPE 1 DIABETES AND HIGH-RISK GLYCEMIA: RESULTS FROM THE CO-PILOT RANDOMIZED CONTROLLED TRIAL

Alisa Boucsein ¹, Yongwen Zhou^1,2, Venus Michaels¹, Jillian Haszard³, Craig Jefferies^4,5, Esko Wiltshire^6,7, Ryan Paul^8,9, Martin De Bock^10,11, Benjamin Wheeler^1,12

¹University of Otago, Department Of Women's And Children's Health, Dunedin, New Zealand, ²The Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, China, ³University of Otago, Biostatistics Centre, Dunedin, New Zealand, ⁴Starship Child Health, Auckland, New Zealand, ⁵The University of Auckland, Liggins Institute And Department Of Paediatrics, Auckland, New Zealand, ⁶University of Otago, Paediatrics & Child Health (uow), Wellington, New Zealand, ⁷Te Whatu Ora Capital, Coast And Hutt Valley, Wellington, New Zealand, ⁸Waikato Regional Diabetes Service, Te Whatu Ora Waikato, Hamilton, New Zealand, ⁹University of Waikato, Division Of Health, Hamilton, New Zealand, ¹⁰University of Otago, Department Of Paediatrics, Christchurch, New Zealand, ¹¹Te Whatu Ora, Health New Zealand, Christchurch, New Zealand, ¹²Te Whatu Ora, Health New Zealand, Dunedin, New Zealand

Background and Aims: Questions and safety concerns remain about the use of automated insulin delivery (AID) systems in children and youth with considerably elevated glycemia. While usually excluded from AID use both clinically and in research, these individuals may have the most to gain from modern burden reducing therapies. Therefore, the CO-PILOT trial aimed to evaluate the efficacy and safety of AID in this group.

Methods: In this prospective, multicenter, parallel-group, open-label randomized controlled trial, AID-naïve children and youth (aged 7–25 years) with type 1 diabetes and a current HbA_1c of at least 8.5% were randomly allocated in a 1:1 ratio to 13 weeks of AID (MiniMed™ 780G) or control (standard care of multiple daily injections or continuous subcutaneous insulin infusion).

Results: A total of 80 participants underwent randomization (37 to AID and 43 to control), and all completed the trial. At 13 weeks, the mean (±SD) HbA_1c decreased from 10.5±1.9% to 8.1±1.8% in the AID group and increased marginally from 10.4±1.6% to 10.6±1.8% in the control group (adjusted between-group difference, -2.5 percentage points; 95% CI, -3.1 to -1.8; P<0.001). Participants in the AID group spent on average 8.4 hours more in the target glucose range of 70 to 180 mg/dL per day than those in the control group. One severe hypoglycemia and two diabetic ketoacidosis events occurred in the control group, with no events in the AID group.

Conclusions: In children and youth with type 1 diabetes and high-risk glycemia, AID safely facilitated significant improvements in glycemic outcomes.

REAL-WORLD IMPROVEMENTS IN GLYCEMIC AND PARENT-REPORTED OUTCOMES IN CHILDREN WITH TYPE 1 DIABETES AGED 2-6 YEARS USING THE MEDTRONIC MINIMED 780G: A PROSPECTIVE OBSERVATIONAL STUDY

Jolien De Meulemeester ¹, Laura Valgaerts¹, Sylvie Tenoutasse², Marieke Den Brinker³, Inge Gies⁴, Marie-Christine Lebrethon⁵, Jacques Louis⁶, Sophie Lambert⁷, Nicole Seret⁷, Marlies Van Loocke⁸, Goele Smeets⁹, Annelies Lemay¹⁰, Sofie Ryckx¹¹, Sylvia Depoorter¹², Sara Van Aken¹³, Ellen Maris¹⁴, Pieter Gillard¹, Thierry Mouraux¹⁵, Kristina Casteels¹

¹University Hospitals Leuven – KU Leuven, Leuven, Belgium, ²HUB Hôpital Universitaire de Bruxelles – HUDERF Hôpital Universitaire Des Enfants Reine Fabiola – ULB, Brussels, Belgium, ³University Hospital Antwerp – University of Antwerp, Edegem, Belgium, ⁴Universitair Ziekenhuis Brussel, Brussels, Belgium, ⁵CHU Liège – Hôpital Citadelle, Liège, Belgium, ⁶Grand Hôpital de Charleroi, Charleroi, Belgium, ⁷Clinique CHC MontLégia, Liège, Belgium, ⁸AZ Delta, Roeselare, Belgium, ⁹Jessa Hospital, Hasselt, Belgium, ¹⁰AZ Turnhout, Turnhout, Belgium, ¹¹ZAS Paola, Antwerpen, Belgium, ¹²AZ Sint-Jan, Brugge, Belgium, ¹³University Hospital Ghent, Gent, Belgium, ¹⁴Vitaz, Sint-Niklaas, Belgium, ¹⁵CHU UCL Namur – Godinne, Yvoir, Belgium

Background and Aims: This study evaluated real-world changes in glycemic and parent-reported outcomes 6 months after initiation of the Medtronic MiniMed^TM 780G in Auto Mode in children with type 1 diabetes aged 2-6 years.

Methods: Children aged 2-6 years whose parents agreed to activate Auto Mode on 780G were recruited at 15 Belgian centers between December 2022 and December 2023. Data were collected every 3 months from start to 6 months after start of Auto Mode. Parent-reported outcomes were evaluated through questionnaires (HAPPI-D and Hypoglycemia Fear Survey (HFS) – Parent). Data are reported as median (IQR) or least-squares mean [95% CI].

Results: Of the 150 children included, data were available from 138, 127 and 109 children at start, 3 and 6 months, respectively. Children were 4.0 (3.0-5.0) years old, had T1D for 1.6 (0.8-2.5) years, and 78.5% used an insulin pump before. Time in range (70-180 mg/dL) increased from 57.4% [55.2-59.7] at start to 65.8% [64.0-67.6] at 6 months (p<0.001). HbA1c decreased from 7.6% (7.4-7.8) to 7.2% (7.0-7.4) (p<0.001) and time <70 mg/dL from 5.0% (4.2-5.8) to 4.8% (3.9-5.6) (p=0.397). Parents reported less diabetes-related burden for themselves and their children on the HAPPI-D questionnaire (22.5 points [21.5-23.5] at start vs 21.1 points [20.1-22.2], p=0.003), while scores on the HFS-Parent questionnaire remained stable. The overall performance of the system is shown as composite diabetes nonagon (Figure 1).

Conclusions: Six-month real-world use of the 780G in Auto Mode in children with T1D aged 2-6 years is associated with improved glycemic management and quality of life for their parents.

EARLY INITIATION OF AUTOMATED INSULIN DELIVERY AT TYPE 1 DIABETES DIAGNOSIS IN CHILDREN AND ADOLESCENTS IS ASSOCIATED WITH IMPROVED OUTCOMES AT TWO YEARS POST DIAGNOSIS

Osagie Ebekozien ¹, Elizabeth Mann², Priya Prahalad³, Risa Wolf⁴, Nestoras Mathioudakis⁴, Nicole Rioles¹, Saketh Rompicherla¹, Susan Thapa¹, Robert Vigersky⁵

¹T¹D Exchange, Boston, United States of America, ²University of Wisconsin-Madison, Madison, United States of America, ³Stanford University, Endocrinology, Stanford, United States of America, ⁴John Hopkins University, Baltimore, United States of America, ⁵Medtronic, Diabetes, California, United States of America

Background and Aims: Automated Insulin delivery (AID) is associated with reduction in HbA1c, Severe Hypoglycemia (SH) and Diabetic Ketoacidosis (DKA) for people with type 1 diabetes (PwT1D). There are limited data on whether the timing of AID initiation is associated with T1D outcomes.

Methods: Data were analyzed from the T1D Exchange EHR dataset. Data from children and adolescents with T1D diagnosed between 2020–2022 with at least 2 years outcome follow up data were included in the analysis. The cohort was grouped into four categories based on time of AID initiation: Less than 6 months (Early Users); 6 – 12 months; 13 – 24 months (Late Users) and Non-AID users at 24 months. LOESS plot estimates, and T tests were used to compare HbA1c, SH and DKA outcomes among the groups.

Results: Of 44,665 eligible children, there were 8,875 AID users with a mean age of 13 years, 23% on public health insurance, 57% were Non-Hispanic White. Among AID Users, 7% were Early users, 15% initiated AID within 6-12 months, while 78% were Late users. HbA1c at 2 years post-diagnosis was statistically significantly (p value less than 0.01) lower for the Early user group (HbA1c - 7.1%) as compared to late user group (HbA1c - 7.9%) as shown below. SH and DKA were significantly lower for the early group 0.9 vs 2.7 per 100 person years and 2.3 vs 7.1 per 100 person years respectively as compared to the late group.

Conclusions: Early initiation of AID is associated with lower HbA1c, SH and DKA 2 years post-diagnosis.

GLUCOSE-LOWERING THERAPY BEYOND INSULIN IN TYPE 1 DIABETES ON MINIMED 780G ADVANCED HYBRID CLOSED-LOOP SYSTEM: A RANDOMIZED CONTROLLED STUDY ON SITAGLIPTIN IN EARLY-STAGE DIABETIC NEPHROPATHY

Nancy Elbarbary ¹, Eman Ismail², Manal El-Hamamsy³, Marwa Ibrahim⁴, Amal Elkholy⁵

¹Faculty of Medicine, Ain Shams University, Department Of Pediatrics, Cairo, Egypt, ²Faculty of Medicine, Ain Shams University, Department Of Clinical Pathology, Cairo, Egypt, ³Faculty of Pharmacy, Ain Shams University, Department Of Clinical Pharmacy, Cairo, Egypt, ⁴Ahram Canadian University, Department Of Clinical Pharmacy, Cairo, Egypt, ⁵Faculty of Pharmacy, Ain Shams University, Department Of Clinical Pharmacy, Cairo, Egypt

Background and Aims: Dipeptidyl peptidase-4 (DPP-4) inhibition has beneficial effects on various metabolic indicators in diabetes. Stromal cell-derived factor-1 (SDF-1) is expressed in diverse organs including kidneys which is cleaved and inactivated by DPP-4 enzyme.

Objectives: This trial was performed to assess effect of sitagliptin on diabetic nephropathy when used as an add-on therapy to advanced hybrid closed-loop (AHCL) system in adolescents with T1DM and nephropathy.

Methods: Forty-six adolescents on MiniMed 780G system were randomly assigned into two groups; intervention group who received oral sitagliptin 50 mg for three months. The other group used AHCL only and served as a control group. Mean glucose management indicator, percentage of TIR, TBR, TAR and CoV were determined. Urinary albumin creatinine ratio (UACR), SDF-1 levels were assessed.

Results: Both groups were well-matched as regards baseline clinical and laboratory characteristics as well as AHCL system settings (p>0.05). Serum SDF-1 levels were higher in all T1DM patients versus healthy controls (p<0.001). After three months, sitagliptin resulted in a significant decrease of SDF-1 levels from 3.58 ± 0.73 to 1.99 ± 0.76 ng/mL (p<0.001), together with improvement of UACR from 64.30 ± 13.01 to 11.70 ± 5.95 mg/g creatinine (p<0.001). In addition, sitagliptin reduced postprandial glucose, sensor glucose, CoV and total daily dose of insulin while TIR 70–180 mg/dL and insulin-to-carbohydrate ratio were significantly increased.

Conclusions: Sitagliptin as an add on therapy to AHCL had a reno-protective effect and prevented progression of diabetic nephropathy among T1DM patients, in addition to the improvement of TIR while reducing glycemic variability without compromising safety.

IMPACT OF LOWERING TARGET GLUCOSE SETTING WITH THE OMNIPOD® 5 AUTOMATED INSULIN DELIVERY (AID) SYSTEM: EVIDENCE FROM REAL-WORLD USE

Gregory Forlenza ¹, Grazia Aleppo², Eden Miller³, Cari Berget¹, Sean Oser⁴, Daniel Desalvo⁵, Francisco Pasquel⁶, Jennifer Sherr⁷, Anne Peters⁸, Emma Wilmot^9,10, Lauren Huyett¹¹, Irene Hadjiyianni¹¹, José Méndez¹¹, Lindsey Conroy¹¹, Trang Ly¹¹

¹University of Colorado Anschutz Medical Campus, Barbara Davis Center For Diabetes, Aurora, United States of America, ²Northwestern University, Feinberg School Of Medicine, Chicago, United States of America, ³Diabetes and Obesity Care LLC, Bend, United States of America, ⁴University of Colorado School of Medicine, Aurora, United States of America, ⁵Baylor College of Medicine, Houston, United States of America, ⁶Emory University, Atlanta, United States of America, ⁷Yale School of Medicine, New Haven, United States of America, ⁸Keck School of Medicine of the University of Southern California, Los Angeles, United States of America, ⁹University of Nottingham, School Of Medicine, Nottingham, United Kingdom, ¹⁰Royal Derby Hospital, University Hospitals of Derby and Burton NHS Trust, Department Of Diabetes And Endocrinology, Derby, United Kingdom, ¹¹Insulet Corporation, Acton, United States of America

Background and Aims: Target glucose is the setting with the greatest impact on AID with the Omnipod® 5 AID System. Among Omnipod 5 real-world users with <70% time in range (TIR; 70-180mg/dL), half are not primarily using the lowest target and one-third are not using it at all (data on file). While customization is important, those aiming to maximize TIR should use the lowest target whenever possible. This analysis quantified the change in glycemic outcomes when lowering the target setting.

Methods: A retrospective analysis of real-world data from Omnipod 5 users in the US with self-reported T1D aged ≥2y or T2D aged ≥18y and sufficient CGM data (≥90 days of data, ≥75% of days with ≥220 readings) was conducted. Pre- and post-exposure to different target settings on glycemic outcomes within individuals were compared for consecutive periods of ≥14 days where the target setting was 120, 130, 140, or 150mg/dL in the first and 110mg/dL in the second period.

Results: 16,241 users with T1D and 1,858 with T2D met analysis criteria for a change in target setting. Lowering the target to 110 mg/dL significantly improved TIR by median 2.5-11.1% for users with T1D and by 3.0-12.0% for users with T2D (Figure). The increase in time below range (TBR; <70mg/dL) for the largest target decrease was median [IQR] 0.30% [0.01%, 0.83%] and 0.05% [-0.04%, 0.33%], respectively, for T1D and T2D (both p<0.05).

Conclusions: These real-world results demonstrate that lowering the target setting with the Omnipod 5 System was correlated with improved TIR with a minimal impact in TBR.

REAL-LIFE EXPERIENCE WITH THE MEDTRONIC MINIMED™ 780G ADVANCED HYBRID CLOSED-LOOP SYSTEM IN PEDIATRIC PATIENTS AND YOUNG ADULTS WITH TYPE 1 DIABETES

Estela Gil-Poch ¹, Francisco Javier Arroyo Díaz¹, Pilar Beato-Víbora²

¹PEDIATRIC UNIVERSITY HOSPITAL BADAJOZ, Pediatric Endocrinology, Badajoz, Spain, ²Badajoz University Hospital, Badajoz, Spain

Background and Aims: The benefit in metabolic control in patients treated with the Medtronic MiniMed™780G system is well known with extensive results in real life. The objective of the study is to analyze glycemic control in pediatric patients with T1D after treatment switch to the Medtronic™780G system, and this relationship with device settings, insulin requirements and body composition.

Methods: A data set of 106 children with T1D using AHCL between December 2020-September 2024 was analyzed. Baseline data, 6-12-24 and 36 months after the start of treatment were compared.

Results: 106 patients were studied with a mean age of 14.6±3.8 years and a mean duration of T1D of 7.1±4.1 years. The patients had previous treatment with MiniMed™640G system, MiniMed™670G, or multiple daily injections associated CGM with DEXCOMG6™ or flash glucose monitoring with FREESTYLE2. After the start of treatment, a statistically significant reduction in HbA1c was observed, as well as an increase in the time in the target range 70-180 mg/dl. A significant improvement in the time in hypoglycemia appeared with continued use of the system. (Table 1). Initial settings for young children were an active insulin time (AIT) of 3 hours and a glucose target (GT) of 120 mg/dl, and an AIT of 2 hours and a GT of 100 mg/dl for older patients. The insulin requirement decreased but a significant increase in BMI appeared.

Conclusions: The MiniMed™780G system improves metabolic control and this is maintained over time. The change in BMI should be considered due to increasing rates of overweight in pediatric population with T1D.

NEW INSULIN SENSITIVITY MODEL TO IMPROVE GLYCEMIC CONTROL

Pierre Gauthier¹, Thomas Cartier-Michaud¹, Alice Adenis¹, Hector Romeo-Ugalde¹, Sylvain Lachal¹, Thibault Le Roux-Mallouf¹, Chesner Desir², Erik Huneker ¹, Pierre Yves Benhamou³

¹Diabeloop, Research, Grenoble, France, ²Diabeloop, Research, GRENOBLE, France, ³University Hospital Grenoble Alpes, Endocrinology, Grenoble, France

Background and Aims: This study analyzes the relationship between insulin sensitivity factor (ISF) and glycemia in type 1 diabetes (T1D) patients based on real world data in order to improve glycemic control and propose modifications to the Hovorka model to incorporate a more accurate relation between ISF and glycemia.

Methods: We used a dataset of 5201 T1D patients equipped with the Diabeloop DBLG1® system. ISF is considered to be the observed direct effect of insulin injection on glycemia. It was computed as ΔG/ΔIOB where ΔG is the difference between maximum and minimum glucose on 4-hours windows splitted by variations of 20 mg/dl, and ΔIOB is the corresponding change in insulin on board. Peaks were filtered to exclude meal impacts. ISF values were binned by average glycemia levels. A modified Hovorka model incorporating the observed ISF pattern was developed and evaluated in silico.

Results: As shown in Fig. 1, ISF measured on data peaked around 100 mg/dL and showed a biphasic decline at higher levels. Median ISF decreased by 50% from 100 to 160 mg/dL and by 33% from 160 to 240 mg/dL. The modified Hovorka model showed an improved representation of the ISF-glycemia relationship compared to data measurement on Fig. 2.

Conclusions: This study provides a measurement of variation in ISF across glycemic ranges in T1D patients and a modification of the Hovorka model. Incorporating these results in the Hovorka model, could potentially improve accuracy and effectiveness in predicting glycemia dynamics and insulin requirements. These findings could enhance closed-loop insulin delivery systems and personalized diabetes management.

REAL -LIFE FOLLOW -UP OF THE PHYSICAL ACTIVITY IN T1D PATIENTS EQUIPPED WITH AN INSULIN PUMP WITH HYBRID CLOSED LOOP (HCL): THE RAPPID STUDY

Laurent Meyer¹, Michael Joubert ², Bruno Guerci³, Saïd Bekka⁴

¹Hôpitaux universitaires de Strasbourg, Endocrinology-diabetology, Strasbourg, France, ²Caen University Hospital, Diabetes Care Unit, Caen, France, ³Brabois Adults Hospital, CHRU of Nancy and Lorraine University, Endocrinology, Diabetology And Nutrition, Nancy, France, ⁴Institut de Diabétologie et Nutrition du Centre, Mainvilliers, France

Background and Aims: The main objective of this study is to describe system settings and glucose control of T1D patients treated with HCL with regular physical activity in real life.

Methods: Non-interventional, longitudinal, prospective, multicenter study, during 1 month, in 4 diabetes care centers. Seventy-seven patients were included between march and July 2024 (Medtronic 780G, n=37; Tandem Control-IQ, n=25; Ypsopump CamAPS-FX Dexcom G6, n=15), trained for applying the consensual rules of doses adjustments related to physical activity, and informed to report their physical activity sessions on diaries.

Results: Forty three men and 34 women have been included: age = 42.2±14.7years; diabetes duration = 23.4±12.9years; HbA1C = 6.8±0.6%; duration of HCL use = 18.8±11.4 months. A total of 803 physical activity sessions have been recorded (74.2% aerobic, 10.0% anaerobic and 15.8% mixed), with a mean of 13.1±7.2 sessions and a mean total physical activity duration of 19.1±13.6 hours per patient. Hypoglycemia occurred in 168 (20.9%) sessions and led to a temporary stop of physical activity in 25/168 cases (14.9%). Carb intake was necessary in 22.3% sessions with an amount of 33.5±23.1g per physical activiy session. No episode of hyperglycemia led to physical activity cessation. Fifty technical problems were encountered during physical activity sessions (6.2%). Analysis of pump settings and CGM data during physical activity sessions is ongoing.

Conclusions: This is the first large observational study designed to describe precisely routine physical activity sessions, pump settings, clinical behavior and CGM outcomes for T1D patients practicing physical activity with different model of HCL system.

FULLY CLOSED-LOOP IMPROVES GLYCAEMIC CONTROL COMPARED TO PUMP WITH CGM IN ADOLESCENTS WITH TYPE 1 DIABETES AND HBA1C ABOVE TARGET: A TWO-CENTRE, RANDOMISED CROSSOVER STUDY

Nithya Kadiyala ¹, Rama Lakshman¹, Janet Allen¹, Julia Ware^1,2, Charlotte Boughton^1,3, Malgorzata Wilinska^1,2, Ajay Thankamony², Sara Hartnell³, Hood Thabit⁴, Ruben Willemsen⁵, Pratik Shah⁵, Roman Hovorka^1,2

¹University of Cambridge, Wellcome-mrc Institute Of Metabolic Science, Cambridge, United Kingdom, ²University of Cambridge, Department Of Paediatrics, Cambridge, United Kingdom, ³Cambridge University Hospitals NHS Foundation Trust, Wolfson Diabetes And Endocrine Clinic, Cambridge, United Kingdom, ⁴Manchester University NHS Foundation Trust, Diabetes, Endocrinology And Metabolism Centre, Manchester, United Kingdom, ⁵Department of Paediatric Diabetes and Endocrinology, The Royal London Children’s Hospital, Barts Health NHS Trust, London, United Kingdom

Background and Aims: Adolescents with type 1 diabetes (T1D) struggle to reach target HbA1c more than other age groups. Safety and efficacy of fully closed-loop (FCL) insulin delivery (no bolusing) has not been assessed in this cohort. We evaluate the use of FCL with faster insulin aspart in adolescents with T1D and HbA1c above target.

Methods: This two-centre, randomised, crossover study included 24 adolescents (13-19 years) using insulin pump therapy with HbA1c >58mmol/mol (7.5%) (mean±SD, age 16±2 years, HbA1c 9.2±1.6% [78±17mmol/mol]). Participants underwent two 8-week periods of unrestricted living comparing FCL (CamAPS HX) with faster-acting insulin aspart (Fiasp) with standard insulin pump therapy with CGM in random order.

Results: In an intention-to-treat analysis, the proportion of time sensor glucose was in target range (3.9–10.0 mmol/L) was higher during FCL than during pump with CGM use (mean±SD 45.2±7.2% vs. 32.3±12.8%, mean difference 12.9 percentage points [95% CI 8.5, 17.3], p<0.001). Time in hyperglycaemia >13.9mmol/L was lower at 30.0±7.1% with FCL compared to pump with CGM (41.3±18.8%, p<0.001). The proportion of time with sensor glucose <3.9 mmol/L was similar between treatments (median[IQR] FCL 2.78% [2.06, 3.74] vs. pump with CGM 2.97% [1.97, 5.14], p=0.316). HbA1c was lower after FCL than after pump with CGM use (8.8±0.8% [73±9mmol/mol] vs 9.4±1.7% [79±18mmol/mol]). There was no difference in total daily insulin dose (p=0.202). No severe hypoglycaemia or DKA events occurred.

Conclusions: Fully closed-loop insulin delivery with CamAPS HX improved glucose control compared with insulin pump therapy with CGM in adolescents with T1D and HbA1c above target.

IMPROVED GLYCEMIC OUTCOMES WITH THE OMNIPOD® 5 SYSTEM IN PEOPLE WITH TYPE 2 DIABETES USING GLP1-RECEPTOR AGONISTS OR SGLT2 INHIBITORS: SUB-ANALYSIS OF THE SECURE-T2D STUDY

Davida Kruger ¹, Francisco Pasquel², Georgia Davis², David Huffman³, Anne Peters⁴, John Parker⁵, Lori Laffel⁶, Giulio Romeo⁶, Justin Mathew⁷, Kristin Castorino⁸, Kathleen Dungan⁹, Mark Kipnes¹⁰, Edward Jauch¹¹, Tamara Oser¹², Viral Shah¹³, Barry Horowitz¹⁴, Anders Carlson¹⁵, Mark Warren¹⁶, Wasim Deeb¹⁷, John Buse¹⁸, John Reed¹⁹, Jason Berner²⁰, Thomas Blevins²¹, Chris Bajaj²², Lauren Kanapka²³, Dan Raghinaru²³, Trang Ly²⁴, Roy Beck²³

¹Henry Ford Health, Detroit, United States of America, ²Emory University, Atlanta, United States of America, ³University Diabetes and Endocrine Consultants, Chattanooga, United States of America, ⁴Keck School of Medicine of the University of Southern California, Los Angeles, United States of America, ⁵Accellacare, Wilmington Health, Wilmington, United States of America, ⁶Joslin Diabetes Center, Harvard Medical School, Boston, United States of America, ⁷Albert Einstein College of Medicine, Bronx, United States of America, ⁸Sansum Diabetes Research Institute, Santa Barbara, United States of America, ⁹Ohio State University, Columbus, United States of America, ¹⁰Diabetes and Glandular Disease Clinic, San Antonio, United States of America, ¹¹Mountain Area Health Education Center, Asheville, United States of America, ¹²University of Colorado, Aurora, United States of America, ¹³Indiana University School of Medicine, Division Of Endocrinology & Metabolism, Indianapolis, United States of America, ¹⁴Metabolic Research Institute, West Palm Beach, United States of America, ¹⁵HealthPartners, International Diabetes Center, Minneapolis, United States of America, ¹⁶Physicians East, Greenville, United States of America, ¹⁷East Coast Institute for Research at First Choice Endocrinology, Jacksonville, United States of America, ¹⁸University of North Carolina, Chapel Hill, United States of America, ¹⁹Endocrine Research Solutions, Roswell, United States of America, ²⁰East Coast Institute for Research at Georgia Mountain Endocrinology, Canton, United States of America, ²¹Texas Diabetes and Endocrinology, Austin, United States of America, ²²Diabetes and Thyroid Center, Fort Worth, United States of America, ²³Jaeb Center for Health Research, Tampa, United States of America, ²⁴Insulet Corporation, Acton, United States of America

Background and Aims: Despite growing use of non-insulin medications including glucagon-like peptide-1 receptor agonists (GLP1-RA) and sodium-glucose transport protein 2 inhibitors (SGLT2i), many individuals with type 2 diabetes (T2D) are not achieving recommended glycemic targets. The Omnipod® 5 Automated Insulin Delivery (AID) System demonstrated safety and efficacy in adults (≥18y) with T2D and was recently cleared by the US FDA for use in this population (also FDA cleared/CE marked for ages ≥2y with type 1 diabetes). This SECURE-T2D sub-analysis evaluated outcomes in participants using stable doses of GLP1-RA or SGLT2i.

Methods: This multicenter single-arm trial enrolled insulin-treated adults aged 18-75y with T2D with HbA1c <12.0%. After a 14-day period to capture baseline data, participants initiated 13 weeks of AID. Differences in glycemic outcomes according to GLP1-RA/SGLT2i use were evaluated.

Results: A total of 305 participants (mean age 57±11y, 24% Black, 22% Hispanic/Latino, 21% using basal-only) initiated AID. Of these, 55% were using GLP1-RA, 44% were using SGLT2i, and 27% were using both. HbA1c, time in range, and time >250mg/dL (>13.9mmol/L) significantly improved with AID, with similar benefit with or without GLP1-RA/SGLT2i use (Table). No differences were observed for time <70mg/dL (<3.9mmol/L). Total daily insulin was reduced by 16% overall (p<0.001), with similar reductions between users and non-users of GLP1-RA/SGLT2i. Weight increased by 0.8kg in the overall cohort (p<0.001) and did not differ by GLP-1 RA/SGLT2i use.

Conclusions: This sub-analysis provides evidence that insulin-treated adults with T2D using the Omnipod 5 System can achieve similar glycemic benefits with minimal hypoglycemia independently of concomitant treatment with GLP1-RA/SGLT2i therapy.

TECHNOLOGICAL VS CONVENTIONAL TREATMENT IN CHILDREN WITH EARLY ONSET OF TYPE 1 DIABETES MELLITUS: FOLLOW-UP OF 12 YEARS AND EVALUATION OF COMPLICATIONS

Gabriele Olivieri ^1,2, Valeria Favalli^1,2, Federica Sandullo^1,2, Francesca Arrigoni^1,2, Benedetta Dionisi^1,2, Riccardo Foglino^1,2, Camilla Morosini^1,2, Carola Lorentino^1,2, Martina Matarazzo^1,2, Elena Zanardi^1,2, Riccardo Sacconi^3,4, Giuseppe Querques^3,4, Gianpaolo Zerbini^2,4, Pasquale Anthony Della Rosa^4,5, Andrea Rigamonti^1,2, Giulio Frontino^1,2, Riccardo Bonfanti^1,2,4

¹IRCCS San Raffaele Hospital, Department Of Pediatrics, Milan, Italy, ²IRCCS San Raffaele Hospital, Diabetes Research Institute, Milan, Italy, ³IRCCS San Raffaele Hospital, Eye Department, Milan, Italy, ⁴Vita-Salute San Raffaele University, School Of Medicine, Milan, Italy, ⁵IRCCS San Raffaele Hospital, Neuroradiological Department, Milan, Italy

Background and Aims: We studied a cohort of 16 patients, who presented with early onset of the disease (mean 2 yrs) and have been followed for 12 yrs. 8 patients have been treated from the onset with technological therapy (AHCL), while 8 have been treated with conventional MDI therapy (MDI).

Methods: We assessed metabolic control (HbA1c), mean of the area under the curve of HbA1c, number of DKA, severe hypos and complications. CNS was studied with RNM, the presence of diabetic nephropathy with AER and the count of urinary podocytes in the urine, diabetic retinopathy was searched by OCT and angio-OCT.

Results: Hba1c was significantly lower in the CSII group during the follow-up after 12 yr, with a 30% reduction of exposition to glucose during follow-up (AUC HbA1c >6% 1,25+/- 0,7 vs 0,87+/- 0,5 p=0.02). No differences between groups for DKA at onset and during follow-up and for severe hypos. CNS RNM: The CSII group showed a significant minor reduction in gray matter at temporal lobe and limbic system (p<0.001). For both groups no differences in AER (7,5 mg/l vs 7,2 mg/l) and podocituria were found. No clinical signs of retinal involvement were identified, however significant reduction of central retinal thickness (measured by OCT) in MDI vs CSII group (255 +/-12 vs 271 +/-8 micron p=0.002).

Conclusions: Technological therapy has been able to reduce by 30% the exposure to hyperglicaemia, thus reducing the impact of high glucose on tissues target (CNS and retina). Technological treatment should be implemented for all diabetic children from the onset of disease.

SEMAGLUTIDE AS ADJUNCT TO AUTOMATED INSULIN DELIVERY IN ADULTS WITH TYPE 1 DIABETES: A DOUBLE-BLINDED, RANDOMIZED, CROSSOVER TRIAL

Melissa-Rosina Pasqua ¹, Michael Tsoukas¹, Alessandra Kobayati¹, Wedyan Aboznadah¹, Adnan Jafar², Ahmad Haidar¹

¹McGill University Health Centre - Royal Victoria Hospital, Endocrinology & Metabolism, Montreal, Canada, ²McGill University Health Centre - Royal Victoria Hospital, Endocrinology & Metabolism, MONTREAL, Canada

Background and Aims: There are ongoing needs to improve glycemic and metabolic control in type 1 diabetes. The aim of our study was to assess semaglutide as adjunct to automated insulin delivery therapy in adults with type 1 diabetes using insulin pumps.

Methods: We performed a randomized, placebo-controlled, cross-over trial to assess semaglutide vs. placebo with automated insulin delivery therapy. In each intervention, participants were titrated up to 1 mg of semaglutide or placebo over 11 weeks, followed by use of an automated insulin delivery system for 4 weeks. The interventions were separated by 2 weeks washout period. The primary outcome was percentage of time spent in the target glucose range of 3.9 to 10 mmol/L during the last 4 weeks of each intervention. Key secondary outcomes were time spent below 3.9 mmol/L, weight, insulin usage, and HbA1c.

Results: Twenty-eight adult participants were randomized (BMI 32.2 kg/m2 (5.1), HbA1c 7.4% (0.8)); 24 completed the trial. Compared to placebo, semaglutide increased time in target range by mean 4.8 (SD 7.6) percentage points (p = 0.006), without increasing time spent below 3.9 mmol/L (p = 0.190). This coincided with a placebo-adjusted reduction in daily insulin usage by 11.3 units [-23.6, -4.9], reduction in weight by 5.3 (2.9) kg, and reduction in HbA1c by 0.5% [-0.7, -0.2] (p < 0.001 for all comparisons). There were 2 episodes of euglycemic ketosis without acidosis requiring medical assistance during semaglutide use.

Conclusions: Semaglutide improves glycemic control with automated insulin delivery compared to placebo, along with reductions in weight and insulin usage. (NCT05205928).

A HIGHER TIME IN RANGE TARGET CANNOT REPLACE TIME IN TIGHT RANGE AS SHOWN BY REAL-WORLD DATA FROM OVER 140.000 MINIMED 780G SYSTEM USERS

Goran Petrovski, Tim Van Den Heuvel, Arcelia Arrieta, Javier Castaneda, Ohad Cohen

Medtronic International Trading Sarl, Diabetes Operating Unit, Tolochenaz, Switzerland

Background and Aims: Technological advancements like AID systems have greatly improved glycemic control in T1D. With GMI levels of 6.5% now achievable, there is debate on whether Time in Tight Range targets (TITR, 70-140 mg/dL) should be used as a metric for euglycemia or if higher Time in Range targets (TIR, 70-180 mg/dL) suffice. We analyzed real-world data from MiniMed 780G system (MM780G) users to explore if a stricter TIR target can replace TITR.

Methods: Real-world data from 149.079 users were analyzed (until Sep2024). TIR and TITR targets corresponding to GMI<7% and GMI<6.5% were identified using ROC curves. Then, we visualized which percentage of users achieving the TIR target also achieved the TITR target, and vice versa.

Results: Targets corresponding to GMI<7% were TIR>70% and TITR>46%, and to GMI<6,5% were TIR>80 and TITR>58%. For targets corresponding to GMI<7%, the figure shows that 15.0% of users meeting TIR did not meet TITR target, while only 5.9% of users meeting the TITR target did not meet the TIR target. For stricter targets (corresponding to GMI<6.5%), the percentage of users ‘meeting the TIR target but not the TITR target’ was 38.7%, while the ‘percentage of users meeting the TITR target but not the TIR target’ remained low at 6.4%.

Conclusions: This analysis shows that high TIR does not necessarily reflect high TITR, particularly with lower GMI (<6.5%). A higher TIR target can therefore not replace a TITR target as a measure of glycemic control.

IN REAL-LIFE SETTING, HALF OF THE PERSONS WITH TYPE 1 DIABETES REACH COMBINED INTERNATIONAL GLUCOSE TARGETS 6 MONTHS AFTER CLOSED LOOP INITIATION

Sylvie Picard ¹, Blandine Courbebaisse², Joelle Dupont³, Fabienne Amiot-Chapoutot⁴, Emmanuelle Lecornet-Sokol⁵, Estelle Personeni⁶, Clara Bouche⁷, Francois Mougel⁸, Francoise Giroud⁹, Sandrine Lablanche¹⁰, Sophie Borot¹¹

¹CIRDIA, Point Medical, Dijon, France, ²Endocrinology Diabetes, Fontenay-sous Bois, France, ³Loire Private Hospital, Endocrinology Diabetes, Saint Etienne, France, ⁴Endocrinology Diabetes, Troyes, France, ⁵IPE (Paris Endocrinology Institute), Paris, France, ⁶Endocrinology Diabetes, Chatenoy le Royal, France, ⁷Rothschild Foundation Hospital, Endocrinology Diabetes, Paris, France, ⁸Centre Mutualite Francaise, Endocrinology Diabetes, Lons le Saunier, France, ⁹Endocrinology Diabetes, Dijon, France, ¹⁰Grenoble University Hospital, Endocrinology Diabetes, Grenoble, France, ¹¹Besancon University Hospital, Endocrinology Diabetes, Besancon, France

Background and Aims: Hybrid closed loop (HCL) systems were shown to improve glucose outcomes in persons with type 1 diabetes (PWT1D). However, little is known on real-life outcomes in PWT1D with either TIR^70-180<50% or TBR^<70≥4% before HCL.

Methods: Data from consecutive PWT1D who started HCL were retrospectively reported by diabetologists from the CIRDIA (multisite HCL initiation center). All participants had signed an informed consent. 14-day AGP data were reported before HCL and 6±1 months later (M6). Targets were defined as TIR^70-180>70%, TBR^<70<4%, GMI<7% and by the percentage of PWT1D reaching the AGP combined criteria (AGPcc) of TIR, TBR and GMI. Results are means ± SD or percentages.

Results: Between May 2023, and January 2024, a total of 153 HCL were initiated by 6 diabetologists; 6-month data were available in 116. HCL system was Medtronic 780G (78%), Control IQ (13%) or CamAPS (9%). Patients were female (51%), mean age 43.6±13.9 years, T1D for 24.2±12.1 years, BMI 27.3±4.8 kg/m², TIR^70-180 53.7± 16.4%, TBR^<70 2.9±2.9%, GMI 7.7±0.9%. At initiation and M6, TIR target value was reached by respectively 16% and 63%, TBR by 66% and 83%, GMI by 20% and 60%, AGPcc by 8% and 45%. In participants with TIR<50% (n=47) or TBR≥4% (n=40) at baseline, AGPcc was reached by respectively 21% and 53% at M6 versus 58% in those with both TIR≥50% and TBR<4%.

Conclusions: In real-life, 58% of PWT1D with baseline TIR≥50% and TBR<4%, 53% of those with TBR≥4% and 21% of those with TIR<50% reached the AGPcc 6 months after HCL initiation – versus 8% before.

OUTCOMES IN THE BIONIC PANCREAS PIVOTAL TRIAL AND IN THE FIRST YEAR OF COMMERCIAL ILET USAGE WERE INDEPENDENT OF USER ENGAGEMENT WITH THE ILET

Steven Russell ^1,2, Rajendranath Selagamsetty¹, Edward Damiano^1,3

¹Beta Bionics, Concord, United States of America, ²Massachusetts General Hospital, Diabetes Unit, Boston, United States of America, ³Boston University, Biomedical Engineering, Boston, United States of America

Background and Aims: Glycemic and continuous glucose monitoring (CGM) outcomes typically improve with increasing user engagement for therapeutic (insulin pumps and pens) and diagnostic (CGMs, flash-glucose monitors, and blood-glucose monitors) devices used to manage type 1 diabetes (T1D). This study aimed to evaluate the effect of user engagement on outcomes during the bionic pancreas pivotal trial (BPPT) and during the first year of commercial iLet usage.

Methods: In this retrospective study, we used frequency of swipes to unlock the iLet as a measurement of user engagement. All actions with the iLet (meal announcements, viewing the glucose graph, changing insulin, viewing alerts, changing volume settings, etc.) begin by swiping to unlock the iLet. We computed correlation coefficients via linear regression between user engagement and clinical outcomes, such as hemoglobin A1c (HbA1c), hypoglycemia, glucose management indicator (GMI), and time in range 70–180 mg/dL (TIR).

Results: In adults (≥18 years old, 216 in the BPPT and 2,755 commercial users) and children (<18 years old, 108 in the BPPT and 540 commercial users), the r² correlation coefficients between frequency of swipes to unlock and each of GMI, time < 54 mg/dL, or TIR was less than 0.01. Furthermore, in the BPPT, the r² correlation coefficient between frequency of swipes to unlock and HbA1c at 13 weeks was less than 0.02.

Conclusions: We conclude that the performance of the iLet in the first year of commercial launch mirrors the performance of the iLet during the BPPT in that outcomes did not depend on how frequently the user interacted with the iLet.

THE CORRELATION BETWEEN FEAR OF HYPOGLYCEMIA AND CONTINUOUS GLUCOSE MONITORING METRICS IN ADULTS WITH TYPE 1 DIABETES USING DIABETES TECHNOLOGY: A ONE-YEAR RETROSPECTIVE SECONDARY-USE STUDY

Laura Valgaerts ¹, Jolien De Meulemeester¹, Margaretha Visser¹, Sara Charleer¹, Bart Keymeulen², Christophe De Block³, Chantal Mathieu¹, Kristina Casteels¹, Pieter Gillard¹

¹University Hospitals Leuven – KU Leuven, Leuven, Belgium, ²University Hospitals Brussels, Brussels, Belgium, ³University Hospitals Antwerp, Antwerp, Belgium

Background and Aims: Although fear of hypoglycemia (FoH) and continuous glucose monitoring (CGM) metrics have been shown to improve with diabetes technology, their underlying correlation remains unclear.

Methods: This retrospective secondary-use study evaluated the correlation between delta (Δ) Hypoglycemia Fear Survey II – Worry subscale (HFS-worry) score and ΔCGM-metrics in 1054 adults with type 1 diabetes (T1D) who used real-time CGM (rtCGM; ALERTT1-study, n=225) or hybrid closed-loop (HCL; INRANGE-study, n=829) for 12 months. Delta refers to the difference in questionnaire scores and CGM-metrics between start and 12 months after start of rtCGM or HCL. Data are reported as mean ± SD.

Results: Mean age was 40 ± 13.7 years, 56.0% were female. At start, HbA1c was 7.5 ± 1.0%, time in range (TIR; 70-180 mg/dL) 58.4 ± 15.8%, time <70 mg/dL 3.5 ± 3.5% and time <54 mg/dL 0.8 ± 1.5%. Delta HFS-worry showed no correlation with Δtime <70 mg/dL (r=0.042, p=0.195) or Δtime <54 mg/dL (r=-0.006, p=0.849), which was confirmed by descriptive analyses of ΔHFS-worry quartiles (Figure 1). Only weak correlations between ΔHFS-worry and ΔTIR (r=-0.074, p=0.024) and Δcoefficient of variation (r=0.113, p=0.008) were found. Also the Δscore of the question about perceived frequency of hypoglycemia on the Diabetes Treatment Satisfaction Questionnaire status showed only a weak correlation with Δtime <70 mg/dL (r=0.168, p<0.001) and Δtime <54 mg/dL (r=0.089, p=0.007).

Conclusions: No clear correlation was found between changes in FoH and CGM-metrics, suggesting that more complex factors beyond glucose management alone likely influence FoH in adults using rtCGM or HCL for one year.

EIGHTEEN-MONTH HYBRID CLOSED-LOOP USE IN VERY YOUNG CHILDREN WITH TYPE 1 DIABETES: A SINGLE-ARM MULTICENTER TRIAL

Julia Ware ^1,2, Janet Allen^1,2, Charlotte Boughton^2,3, Malgorzata Wilinska^1,2, Sara Hartnell³, Ajay Thankamony¹, Carine De Beaufort^4,5,6, Fiona Campbell⁷, Elke Froehlich-Reiterer⁸, Maria Fritsch⁸, Sabine Hofer⁹, Thomas Kapellen¹⁰, Birgit Rami-Merhar¹¹, Martin Tauschmann¹¹, Roman Hovorka^1,2

¹University of Cambridge, Department Of Paediatrics, Cambridge, United Kingdom, ²University of Cambridge, Institute Of Metabolic Science-metabolic Research Laboratories, Cambridge, United Kingdom, ³Cambridge University Hospitals NHS Foundation Trust, Wolfson Diabetes And Endocrine Clinic, Cambridge, United Kingdom, ⁴Centre Hospitalier de Luxembourg, Deccp, Clinique Pédiatrique, Luxembourg, Luxembourg, ⁵University of Cambridge, Faculty Of Science, Technology And Medicine, Luxembourg, Luxembourg, ⁶UZ-VUB, Department Of Paediatric Endocrinology, Brussels, Belgium, ⁷Leeds Children's Hospital, Department Of Paediatric Diabetes, Leeds, United Kingdom, ⁸Medical University of Graz, Department Of Pediatric And Adolescent Medicine, Graz, Austria, ⁹Medical University of Innsbruck, Department Of Paediatrics I, Innsbruck, Austria, ¹⁰University of Leipzig, Hospital For Children And Adolescents, Leipzig, Germany, ¹¹Medical University of Vienna, Department Of Paediatrics And Adolescent Medicine, Vienna, Austria

Background and Aims: We aimed to evaluate the longer-term safety and efficacy of hybrid closed-loop therapy in very young children with type 1 diabetes.

Methods: Following a 16-week multi-center, multi-national, randomized crossover trial comparing hybrid closed-loop (CL) with sensor-augmented pump (SAP) therapy in 74 very young children aged 1-7 years with type 1 diabetes, participants were invited to an extension phase using hybrid closed-loop for a further 18 months. Outcomes were compared to the primary phase SAP period and primary phase CL period.

Results: After the primary study phase, 60 participants were eligible to enroll in the extension. Of these, 49 consented (mean±SD age 6.6±1.5 years) to continue use of hybrid closed-loop for 18 months. The percentage time with glucose in range 3.9-10.0mmol/L was 8.4 percentage points (95% CI 6.7 to 10.1; p<0.001) higher, while HbA1c was 0.4% ([5.0mmol/mol], 95% CI 0.3 to 0.6 [3.7 to 6.2]; p<0.001) lower during the CL extension phase compared to the primary phase SAP period. At 18 months mean HbA1c was 6.7±0.5%, similar to the primary phase CL period (p=0.76). Time in hypoglycemia (<3.9mmol/L) was similar between CL extension phase and both primary phase SAP and CL periods (p=0.31 and p=0.70 respectively). There were two severe hypoglycemia events and one other serious adverse event during the extension phase. One unexpected serious adverse device effect occurred.

Conclusions: Use of the Cambridge hybrid closed-loop system led to sustained improvements in glycemic control lasting more than 18 months in very young children with type 1 diabetes.

MENSTRUAL CYCLE PHASE-SPECIFIC PATTERN DETECTION VIA BLOCK-BASED RECURRENCE QUANTIFICATION ANALYSIS (BLOCKRQA) FOR TARGETED INSULIN CONTROL IN WOMEN WITH TYPE 1 DIABETES (T1D)

Carson Wedding ^1,2, Gregory Forlenza², Taisa Kushner³

¹Colorado School of Mines, Quantitative Biosciences And Engineering, Golden, United States of America, ²University of Colorado Anschutz Medical Campus, Barbara Davis Center For Diabetes, Aurora, United States of America, ³Galois, Inc, Portland, United States of America

Background and Aims: Women with T1D frequently experience variation in insulin needs across the menstrual cycle due to hormonal fluctuations. Most commonly, a decrease in insulin sensitivity during the luteal phase (LP) results in increased hyperglycemia. In this project, we use a novel pattern detection algorithm, BlockRQA, to identify and track complex glycemic patterns which when observed, result in starkly different glycemic outcomes in the luteal vs follicular phases (FP).

Methods: Complex recurrent patterns are identified using data from women with T1D (n=30, 95±12days of data, age = 27±7 years, HbA1c = 6.9±0.8). Patterns are at least 2hr in length, with features determined on the per-hour basis (eg. average blood glucose (BG), total insulin delivery, meals, etc). For each subject, the percent change in total BG risk index of each pattern observed from FP to LP was tracked (Figure 1). The percent change in risk was then averaged across patterns, weighted by number of occurrences.

Results: All subjects demonstrated an overall increase in risk in the LP, with 23 showing >50% mean increase and 9 showing >100% mean increase. For the 30 subjects, the mean percent increase in risk from the FP to LP was 33%±8.4%.

Conclusions: While this work is ongoing, intermediate results show significant promise in identifying and studying patterns that consistently result in different glycemic outcomes depending on menstrual cycle phases. Such analysis is the first step in being able to design novel menstrual-cycle specific interventions and artificial pancreas algorithms that address differing insulin needs across cycle phases, improving glycemic control and outcomes.

IMPACT OF SIMPLERA SYNC™ SENSORS AND EXTENDED INFUSION SETS ON GLYCEMIA AND MINIMED™ 780G SYSTEM PERFORMANCE IN CHILDREN AND YOUNG ADULTS WITH HIGH HBA1C

Venus Michaels¹, Alisa Boucsein¹, Yongwen Zhou^1,2, Shirley Jones¹, Ryan Paul^3,4, Esko Wiltshire^5,6, Craig Jefferies^7,8, Martin De Bock^9,10, Benjamin Wheeler ^1,11

¹University of Otago, Department Of Women's And Children's Health, Dunedin, New Zealand, ²The Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, China, ³Waikato Regional Diabetes Service, Te Whatu Ora Waikato, Hamilton, New Zealand, ⁴University of Waikato, Division Of Health, Hamilton, New Zealand, ⁵University of Otago, Paediatrics & Child Health (uow), Wellington, New Zealand, ⁶Te Whatu Ora Capital, Coast And Hutt Valley, Wellington, New Zealand, ⁷Starship Child Health, Auckland, New Zealand, ⁸The University of Auckland, Liggins Institute And Department Of Paediatrics, Auckland, New Zealand, ⁹University of Otago, Department Of Paediatrics, Christchurch, New Zealand, ¹⁰Te Whatu Ora, Health New Zealand, Christchurch, New Zealand, ¹¹Te Whatu Ora Southern, Dunedin, New Zealand

Background and Aims: The MiniMed™ 780G improves glycemia and reduces burden in type 1 diabetes. We investigated how new all-in-one “Simplera Sync™” sensors and 7-day extended infusion sets (EIS) affect glycemia and system performance in young people with elevated HbA1c levels.

Methods: We conducted a sub-study with two distinct ≥3-month phases (P1, P2) in all 76 participants (aged 7-25 years; pre-enrolment HbA1c ≥ 8.5%) enrolled in a post-RCT extension study. During P1, participants used 780G with Guardian 4™ sensors and 3-day infusion sets. In P2, they transitioned to investigational 780G with Simplera Sync™ and EIS. We compared glycemic and system performance outcomes between pre-automation baseline, P1, and P2.

Results: Baseline HbA1c was 10.4% ± 1.8%, improving to 8.2% ± 1.4% after P1 [p<0.01], and remaining stable at 8.2% ± 1.0% after phase 2 [p=0.36]. Time in range (70-180mg/dL) increased from 22.2% ± 13.3% at baseline, to 58.1% ± 15.1% after P1 [p<0.01] and 60.7% ± 14.0% after P2 [p=0.06]. Time in tight range (70-140 mg/dL) increased from 13.7% ± 9.5% at baseline to 38.0% ± 13.5% after P1 [p<0.01] and 40.9% ± 13.1% after P2 [p=0.02]. No increase in hypoglycemia was seen from P1 to P2. In P2, automation time increased from 76.1% to 86.1% [p<0.01] and sensor wear time from 77.3% to 88.6% [p<0.01] compared to P1.

Conclusions: In this high-risk population, Simplera Sync™ and EIS improved time in automation and sensor wear time with the MiniMed™780G system. This was associated with improved time in tight range with no increase in hypoglycemia.

EFFICACY OF THE OMNIPOD® 5 AUTOMATED INSULIN DELIVERY (AID) SYSTEM COMPARED WITH MULTIPLE DAILY INJECTIONS IN TYPE 1 DIABETES: A MULTINATIONAL RANDOMIZED CONTROLLED TRIAL (RADIANT)

Emma Wilmot ^1,2, Bruno Guerci³, Emeline Renard⁴, Kevin Perge^5,6, Hélène Hanaire⁷, Jacques Beltrand⁸, Pieter Gillard⁹, Aurelie Berot¹⁰, Thomas Crabtree^1,2, Elise Bismuth¹¹, Michael Joubert¹², Marie-Beatrice Saade¹³, Pratik Choudhary¹⁴, Randa Salet¹⁵, Lalantha Leelarathna^16,17, Rachel Reynaud¹⁸, Sandrine Lablanche¹⁹, Cecile Gouillard-Darnaud²⁰, Sze May Ng^21,22, Philippe Lysy²³, Nikolaos Daskas²⁴, Trang Ly²⁵, Marc Nicolino^5,6

¹University of Nottingham, School Of Medicine, Nottingham, United Kingdom, ²Royal Derby Hospital, University Hospitals of Derby and Burton NHS Trust, Department Of Diabetes And Endocrinology, Derby, United Kingdom, ³Brabois Adults Hospital, CHRU of Nancy and Lorraine University, Endocrinology, Diabetology And Nutrition, Nancy, France, ⁴Children’s Hospital, CHRU of Nancy and Lorraine University, Pediatric Medicine Department, Nancy, France, ⁵Hospices Civils de Lyon, Hôpital Femme Mère Enfant, Service d'Endocrinologie et Diabétologie Pédiatrique, Lyon, France, ⁶Université Claude Bernard, Lyon, France, ⁷University Hospital of Toulouse, Department Of Diabetology, Toulouse, France, ⁸Hôpital Necker Enfants malades, APHP Centre, Université de Paris, Paris, France, ⁹University Hospitals Leuven, Department Of Endocrinology, Leuven, Belgium, ¹⁰American Memorial Hospital, CHU of Reims, Departement Of Pediatrics, Reims, France, ¹¹Robert Debré Hospital, Assistance Publique Hopitaux de Paris, Departement Of Pediatric Endocrinology And Diabetology, Paris, France, ¹²Caen University Hospital, UNICAEN, Diabetes Care Unit, Caen, France, ¹³Rennes University Hospital, Pediatric Endocrinology And Diabetology Unit, Department Of Pediatrics, Rennes, France, ¹⁴Diabetes Research Centre, University of Leicester, Leicester, United Kingdom, ¹⁵Nîmes University Hospital, Department Of Pediatrics, Nîmes, France, ¹⁶Diabetes, Endocrine and Metabolism Centre, Manchester University NHS Foundation Trust, Manchester, United Kingdom, ¹⁷Imperial College, Department Of Metabolism, Digestion And Reproduction, London, United Kingdom, ¹⁸Timone Hospital, Aix-Marseille University, Multidisciplinary Pediatrics Department, Marseille, France, ¹⁹Grenoble Alpes University Hospital France, Grenoble Alpes University, Diabetology Department, Grenoble, France, ²⁰Lenval University Pediatric Hospital, Nice, France, ²¹Mersey and West Lancashire Teaching Hospitals, Southport, United Kingdom, ²²Faculty of Health, Social Care and Medicine, Edge Hill University, Ormskirk, United Kingdom, ²³Cliniques Universitaires Saint-Luc, Brussels, Belgium, ²⁴Oxford University Hospitals NHS Foundation Trust, John Radcliffe Hospital, Department Of Paediatric Endocrinology And Diabetes, Oxford, United Kingdom, ²⁵Insulet Corporation, Acton, United States of America

Background and Aims: Despite advances in technology, the prevailing therapy for most people living with type 1 diabetes (T1D) continues to be multiple daily injections (MDI), with metabolic and clinical outcomes remaining at suboptimal levels. This randomized controlled study aimed to demonstrate superior efficacy of the Omnipod® 5 AID System as compared with MDI therapy and sensor use in children and adults with T1D.

Methods: Children and adults aged 4-70 years with T1D ≥1 year and screening HbA1c 7.5-11% (58-97mmol/mol) currently using MDI with a FreeStyle Libre 2 continuous glucose monitor (CGM) for ≥3 months were enrolled across 19 institutions in France, the United Kingdom, and Belgium. Participants completed 14 days of data collection with MDI+CGM and then were randomized 2:1 to use the Omnipod 5 AID System with FreeStyle Libre 2 (intervention arm) or continue with MDI+CGM for 13 weeks (control arm). The primary endpoint is the change in HbA1c at 13 weeks. Secondary endpoints, tested in a pre-specified hierarchical order include CGM metrics and person-reported outcomes.

Results: A total of 188 participants (109 aged <18 years; 79 aged ≥18 years) were randomized (2:1), 125 to the intervention arm and 63 to the control arm (Table), with the distribution of adult and pediatric participants predefined in the study protocol. The final study visit was completed on August 16, 2024. Full results will be available at the time of the presentation.

Conclusions: This large, multinational randomized controlled trial will provide a direct comparison of outcomes with the Omnipod 5 AID System compared with MDI+CGM therapy.

PREDICTORS OF GLYCEMIC IMPROVEMENT IN CHILDREN AND YOUTH WITH SEVERE HYPERGLYCEMIA USING THE MINIMED 780G SYSTEM

Yongwen Zhou ^1,2, Alisa Boucsein¹, Venus Michaels¹, Madeleine Gray¹, Craig Jefferies^3,4, Esko Wiltshire^5,6, Ryan Paul^7,8, Amber Parry-Strong⁵, Maheen Pasha⁹, Goran Petrovski¹⁰, Martin De Bock^11,12, Benjamin Wheeler^1,13

¹University of Otago, Department Of Women's And Children's Health, Dunedin, New Zealand, ²The Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, China, ³Starship Child Health, Auckland, New Zealand, ⁴The University of Auckland, Liggins Institute And Department Of Paediatrics, Auckland, New Zealand, ⁵University of Otago, Paediatrics & Child Health (uow), Wellington, New Zealand, ⁶Te Whatu Ora Capital, Coast And Hutt Valley, Wellington, New Zealand, ⁷Waikato Regional Diabetes Service, Te Whatu Ora Waikato, Hamilton, New Zealand, ⁸University of Waikato, Division Of Health, Hamilton, New Zealand, ⁹Sidra Medicine, Clinical Nutrition And Dietetics, Doha, Qatar, ¹⁰Sidra Medicine, Division Of Endocrinology And Diabetes, Doha, Qatar, ¹¹University of Otago, Department Of Paediatrics, Christchurch, New Zealand, ¹²Te Whatu Ora, Health New Zealand, Christchurch, New Zealand, ¹³Te Whatu Ora, Health New Zealand, Dunedin, New Zealand

Background and Aims: To identify factors with the greatest influence on glycemia in young people with type 1 diabetes (T1D) and very elevated glycemia after 3-months of 780G.

Methods: Data were extracted from two separate and previously published studies with similar inclusion criteria assessing AID efficacy among young people (aged 7-25 years) with glycated hemoglobin A1c (HbA1c) ≥8.5% (≥69 mmol/mol) and naïve to automation. Univariate and multivariate linear models were performed to explore factors leading to the greatest improvements in HbA1c and time in range 3.9-10.0 mmol/L (TIR).

Results: 99 young people (age 17.3± 4.2 years; HbA1c 10.6±1.9% [92±21 mmol/mol]) were included. After 3-months, HbA1c improved to 8.1±1.5% (65±16 mmol/mol) (-2.5±2.1% change), and TIR improved from 24.2±13.5% to 58.4±15.4% (p both <0.001). After adjusting for gender, deprivation index quantities, prior insulin delivery method, time in sensor wear, active insulin time, and glucose target, participants with higher baseline HbA1c >11.0% and higher automation time >80% achieved the greatest HbA1c improvements by 2.4% (p <0.001) and 1.7% (p=0.001). Similar findings were seen for time in range. No other factors, including higher meal announcement frequency >3 times, prior use of multiple daily injection, ethnicity, gender, and age >15 years contributed to statistically significant improvement.

Conclusions: In young people with markedly elevated glycemia, the two most important predictors of glycemic improvement were high baseline HbA1c and high use of automation (>80%). Those with the highest HbA1c significantly benefit from automation. Sociodemographic background and carbohydrate counting adherence/knowledge should not prevent or delay access to AID.

ICAN CGM SYSTEM – FIRST PERFORMANCE DATA IN A PEDIATRIC POPULATION

Jiangfeng Fei ¹, Jiyun Zheng¹, Fei Gao¹, Yuanyuan Yu¹, Ao Gao¹, Qian Zhou², Frank Flacke³, Shuxia Ding⁴, Wenxian Ouyang⁵, Haiying Wu⁶

¹Sinocare Inc., R&d, Changsha, China, ²Sinocare Inc., Department Of Regulatory Affairs, Changsha, China, ³Sinocare Inc., Medical Affairs, Mainz, Germany, ⁴Ningbo University, Department Of Endocrinology, Women And Children's Hospital Of Ningbo University, Ningbo, China, ⁵Hunan Children's Hospital, Department Of Hepatology And Endocrinology, Changsha, China, ⁶Children's Hospital of Soochow University, Department Of Endocrinology, Suzhou, China

Background and Aims: The iCan CGM System, developed by Sinocare Meditech Inc. is equipped with proprietary innovative sensor technology. After receiving CE-mark for a wear-time of 15d in an adult population the next step was to generate pediatric population data. This study was conducted according to international standards in three leading hospitals in China.

Methods: The study has a single arm, prospective multicenter design. This interim analysis reports on 60 patients, age 2-17 yrs., the majority (88.3%) with T1D. At days 1-2, 6-8 and 14-15 the performance of the device was assessed by comparison to reference readings. The data were analyzed according to generally accepted procedures.

Results: We report an overall MARD of 8.60%, a 20/20% performance of 95.36% with a median sensor wear-life of 15d which corresponds with the label wear time of 15d. The MARD results at the 3 accuracy testing visits were estimated to 8.63% (1-2d), 8.07% (6-8d), and 9.11% (14-15d). The Consensus error grid analysis shows 100% (95%CI: 99.87% – 100.00%) in Zones A+B. There were no device related and skin related adverse events reported up to this point.

Conclusions: In this interim analysis of the first performance study in a pediatric population, the iCan CGM System demonstrates high overall accuracy and a single digit MARD throughout the wear-time. Furthermore, it confirmed the outstanding wear-life of the sensor. The overall performance of the iCan CGM System in this population confirms the consistent performances previously seen and reported in adults (NCT06570551).

PERFORMANCE AND ACCURACY OF 15-DAY G7 CONTINUOUS GLUCOSE MONITOR

Satish Garg ¹, Mark Christiansen², Timothy Bailey³, Abdulrahman Jbaily⁴, Sharon Gao⁴, Matthew Johnson⁴, Stayce Beck⁴

¹Barbara Davis Center for Diabetes, University of Colorado, Medicine And Pediatrics, Aurora, United States of America, ²Diablo Clinical Research, Walnut Creek, United States of America, ³Headlands Research AMCR Institute, Escondido, United States of America, ⁴Dexcom Inc., San Diego, United States of America

Background and Aims: Continuous glucose monitoring (CGM) is becoming a standard for glucose monitoring for patients managing their diabetes. We evaluated the accuracy of a new 15-day G7 CGM (Dexcom, Inc., San Diego, CA) with a 12-hour grace period.

Methods: The study included 130 adult participants at six investigational sites in the USA. CGMs were placed on the posterior upper arm. Four separate in-clinic visits (up to 11 hours each) during the 15.5-day wear period. Each visit included hypoglycemic and hyperglycemic challenges and frequent arterialized venous blood glucose measurements made with the YSI 2300 Stat Plus glucose analyzer. The MARD and %20/20 agreement were calculated as previously described (PMID: 35157505). The study was approved by the FDA through the Investigational Device Exemption process (G210190) and registered on ClinicalTrials.gov (NCT05263258).

Results: Participant demographics are summarized in Table 1. A total of 20,310 CGM-YSI matched pairs with values ranging from 40-400 mg/dL (2.2-22.2 mmol/L) were included in the evaluation. Overall, the MARD was 8.0%, the %20/20 agreement rate was 94.2%, and the proportion of CGM readings within 20% of corresponding YSI values was 93.6% (lower bound of the 95% confidence interval, 91.9%). The %20/20 agreement rates on Day 1 and in the first 12 hours of wear were 89.4% and 83.4%, respectively (Table 2). No serious adverse events related to the device were reported during the study.

Conclusions: We conclude that the 15-day G7 CGM provides accurate glucose readings for 15.5 days in adults with diabetes and meets the FDA iCGM requirements.

CGM CHARACTERISTICS IN 383 PREGNANT WOMEN HAVING HAD BARIATRIC SURGERY

Patrick Ritz¹, Hélène Hanaire ², Nicolas Touloupas³, Maram Al Saglab⁴, Camille Ternynck⁵, Damien Subtil⁶, Paul Guerby⁷, Frédérique Rimareix⁸, Anne Vambergue⁴

¹University Hospital Toulouse, Endocrinology And Nutrition, Toulouse, France, ²CHU de Toulouse, Diabetologie, Toulouse, France, ³CHU de Toulouse, Nutrition, Toulouse, France, ⁴CHRU de Lille, Département D’endocrinologie Diabétologie Nutrition, Lille, France, ⁵CHRU de Lille, Ea 2694 – Santé Publique: épidémiologie Et Qualité Des Soins, Lille, France, ⁶CHRU de Lille, Lille, France, ⁷CHRU de Lille, Département D’obstétrique Hôpital Jeanne De Flandres, Lille, France, ⁸University Hospital Toulouse, Diabetologie, Toulouse, France

Background and Aims: We aimed at describing CGM metrics in French pregnant women after bariatric surgery, compared to those of normal pregnancies (1) and studying the relationship with pregnancy outcomes.

Methods: Retrospective study of 383 pregnancies after gastric bypass or sleeve gastrectomy. CGM were recorded 24-28 weeks of gestation. LGA and SGA babies were defined according to AUDIPOG curves. Preterm delivery was considered for birth <37 weeks’ gestation. CGM metrics of 413 non-operated uncomplicated pregnancies were taken from (1).

Results: All CGM parameters differed from those in non-operated pregnancies (Table 1). In 287 bypasses versus 98 sleeves, TIR (63-140mg/dl) was significantly reduced whereas TAR, TBR, CV were increased. Mothers giving birth to LGA had increased mean glucose, and decreased TBR. No differences in CGM were seen in mothers giving birth to preterm infants (vs. mothers without preterm birth), and if the baby was SGA (vs. normal growth). Table 1 comparison between CGM in this cohort and that of non-operated women with uncomplicated pregnancies (ref 1). Mean (SD, or IQR), all P <0.0001.

Conclusions: In conclusion, CGM metrics are different in operated vs non operated women, and between sleeve and bypass. CGM was only different in mothers given birth to LGA babies. 1 Carlson AL,et al. BMJ Open Diab Res Care 2024;12:e003989.

MEMBRANES FOR CGM AND MULTIANALYTE SENSING

Andre Martinez

DSM Biomedical, Exton, United States of America

Background and Aims: Multianalyte sensors operate on principles similar to CGM. One component required for these sensors is the outer membrane layer intended to limit the rate of analyte permeation. For oxygen dependent sensors, this layer must also correctly balance the permeability ratio of oxygen to analyte to enable reliable detection over physiological variation of oxygen and analyte concentrations. Sparsa TPUs were investigated to determine applicability for use as analyte limiting membranes for glucose, lactate, and ketone sensors.

Methods: To determine analyte permeability, free standing films were cast and loaded into Franz cells to perform permeability experiments. The reciever side of cells was sampled periodically and subjected to an enzymatic assay that generated colorimetric differnces based on analyte concentration that were quantified spectroscopically. These values were used to determine permeability. A quartz crystal microbalance was used to quantify non-specific protein binding by BSA on Sparsa surfaces. Sparsa microstructure was investigated via AFM and SAXS. A potentiostat and glucose sensors perchased from Pinnacle Technology were coated with Sparsa and subjected changes in glucose concentrations to establish response curves.

Results: Sparsa permeability was determined appropriate for acting as an analyte limiting membrane for glucose, lactate, and ketone sensing. Casting environment had little impact on the structure or performance of Sparsa membranes. Varying formulation hydrophilicity changed sensitivity to glucose for a sensor coated with Sparsa. Non-fouling behaviour was exhibited, similar to that of a PEG brush.

Conclusions: Sparsa is a versatile membrane material suitable for use in CGM-like devices to monitor various analytes including glucose, lactate, and ketones.

THE DIURNAL VARIATION OF SENSOR-DETECTED, AND PATIENT-REPORTED HYPOGLYCAEMIA IN PEOPLE LIVING WITH TYPE 1 AND TYPE 2 DIABETES

Jonah Thomas ¹, Pratik Choudhary^2,3, Vaios Koutroukas^3,4

¹University of Leicester, Leicester Diabetes Centre, Leicester, United Kingdom, ²University Hospitals of Leicester NHS Foundation Trust, Leicester Diabetes Centre, Leicester, United Kingdom, ³University of Leicester, Diabetes Research Centre, Leicester, United Kingdom, ⁴Leicester Diabetes Centre, Leicester, United Kingdom

Background and Aims: It is understood that rates of hypoglycaemia vary over a 24-hour period. We therefore aimed to examine how the rates of sensor-detected and patient-reported hypoglycaemia changes over the course of a day.

Methods: Data from the HypoMETRICS study were analysed. All episodes of level 1 (<=3.9mmol/mol), and level 2 (<3.0mmol/mol) sensor-detected hypoglycaemia (SDH) and patient-reported hypoglycaemia (PRH; measured via the HypoMETRICS mobile application) were identified. The hour-block each event occurred in was recorded, and the probability of experiencing an event within that hour-block, over the study duration was calculated.

Results: A total of 594 participants data were included (type 1; 277, mean age; 54±16 years, male; 327, HbA1c; 7.5±1.1mmol/mol, mean valid days; 68.4 days). In type 1 diabetes, the risk of level 1 SDH and PRH is low over night but higher during the day (SDH day=6.3% vs SDH night=5.3%; PRH day=4.3% vs PRH night=3.7%, p<0.001). Individuals with type 2 sees a higher risk of level 1 and PRH over the night than the day (SDH level 1 day=3.5% vs SDH level 1 night=4.2%, p<0.001; PRH day=2.7% vs PRH night=4.3%, p=0.034). Risk of clinically relevant level 2 hypoglycaemia is low and flat over the day and night in both groups (type 1=3.1%; type 2=2.2%).

Conclusions: The risk of symptomatic hypoglycaemia is higher during the day in those with type 1 whilst higher during the night in those with type 2. The rates of clinically relevant level 2 hypoglycaemia are low and constant in both groups across the 24 hours.

INCREASED FREQUENCY OF HYPOGLYCEMIC EVENTS MAY CONTRIBUTE TO ELEVATED FRACTURE INCIDENCE IN YOUNG INDIVIDUALS WITH TYPE 1 DIABETES

Ellen Åkesson ¹, Auste Pundziute Lyckå², Piotr Michno^3,4, Alexander Thorén⁵, Stefan Särnblad⁶, Ragnar Hanås^7,8, Karin Åkesson^4,9

¹Sahlgrenska University Hospital, Gothenburg, Sweden, ²Department of Paediatrics, Institute of Clinical Sciences, Sahlgrenska Academy, University Of Gothenburg, Gothenburg, Sweden, ³Department of Ortopedics, Ryhov County Hospital, Jönköping, Sweden, ⁴Department of Biomedical and Clinical Sciences (BKV) Linköping University, Linköping, Sweden, ⁵Centre of Registers Västra Götaland, Gothenburg, Sweden, ⁶Örebro University, School Of Medical Sciences, Örebro, Sweden, ⁷NU Hospital Group, Department Of Pediatrics, Uddevalla, Sweden, ⁸Sahlgrenska Academy, Institute of Clinical Science, Gothenburg University, Gothenburg, Sweden, ⁹Department of pediatrics, Ryhov County Hospital, Jönköping, Sweden

Background and Aims: This study aims to compare the incidence of fractures in young T1D individuals vs controls and to explore the association between the frequency of hypoglycemic events and fracture incidence. Previous research has indicated an elevated fracture incidence among individuals with T1D. The predominant hypothesis attributes this increased risk to a higher prevalence of osteoporosis.

Methods: Data on 22,077 (55% males) individuals diagnosed with T1D before the age of 18 years were retrieved from the Swedish National Diabetes Register (NDR) and compared with 110,390 age- and sex-matched controls.

Results: There is a significantly increased risk of fractures in T1D individuals compared to controls (Figure1). This elevated risk isńt seen before T1D diagnosis but is apparent already within the first few years after diagnosis. It is observed in both sexes, with females’ relative risk (RR) 1.4 (95%CI 1.35-1.47) and males’ RR 1.27 (95%CI 1.23-1.30) during the follow up period. In the National Patient Register 0.56 hypoglycemic events/T1D individual with at least one fracture vs 0.3/T1D individual without any fracture were reported. Mean age in the fracture group was 27 years, compared to 22 years in the non-fracture group. A logistic regression model, adjusted for age and duration of T1D, still showed a significantly higher frequency of hypoglycemic events throughout the follow-up period (p < 0.03).

Conclusions: The increased incidence of fractures in individuals with T1D may, to some extent, be attributed to a higher risk of accidents resulting from hypoglycemic events. Continuous glucose monitoring systems with alarms could be crucial tools to reduce this risk.

COMBINING TIME BELOW RANGE WITH AWARENESS STATUS IDENTIFIES THOSE WITH TYPE 1 DIABETES AT THE HIGHEST RISK OF SEVERE HYPOGLYCEMIA: RESULTS FROM THE SFDT1 STUDY

Dulce Canha ¹, Pratik Choudhary², Isabela Banu³, Brigitte Delemer⁴, René Valéro⁵, Nathalie Ronci⁶, Blandine Delenne⁷, Lise Dufaitre⁸, Tiphaine Trecan⁹, Pauline Schaepelynck¹⁰, Caroline Sanz¹¹, Sopio Tatulashvili¹², Emmanuel Cosson¹², Gloria Aguayo¹, Jean-Pierre Riveline¹³, Guy Fagherazzi¹

¹Luxembourg Institute of Health, Department Of Precision Health, STRASSEN, Luxembourg, ²University Hospitals of Leicester NHS Foundation Trust, Leicester Diabetes Centre, Leicester, United Kingdom, ³Hôpital Paris Saint-Joseph, Service De Diabétologie Endocrinologie, Paris, France, ⁴CHU de Reims - Hôpital Robert Debré, Service D'endocrinologie, Diabétologie Et Nutrition, REIMS CEDEX, France, ⁵AP-HM Hôpital de la Conception, Marseille, France, ⁶Ronci Nathalie, Blanquefort, France, ⁷Centre Hospitalier du Pays d'Aix, AIX EN PROVENCE, France, ⁸Hôpital Saint-Joseph Marseille, Marseille, France, ⁹Hôpital Lariboisière - AP-HP, Paris, France, ¹⁰AP-HM Hôpital Sainte Marguerite, Marseille, France, ¹¹Clinique pasteur, Toulouse Cedex, France, ¹²CHU Avicenne, Service D’endocrinologie-diabétologie-nutrition, Bobigny, Paris, France, ¹³INSERM UMR-S1151, CNRS UMR-S8253, Université Paris Cité, Institut Necker Enfants Malades, Paris, France

Background and Aims: Severe hypoglycemia events (SHE), defined by critically low blood glucose levels which require the need of external assistance, remain frequent in people with type 1 diabetes (T1D) despite advanced diabetes technologies. We tested whether time below range (<70 or <54 mg/dl (TBR70, TBR54)) was a relevant marker of SHE and how impaired awareness of hypoglycemia (IHA) modifies this relationship.

Methods: SFDT1 participants using continuous glucose monitoring (CGM) with valid TBR data were included. IAH was assessed using the Gold Score (≤2=no-IAH, 3=undetermined, ≥4=IAH). SHE frequency was self-reported 12 months post-inclusion. Age and sex-adjusted logistic regression models analyzed associations between TBR and SHE, including TBR-IAH interactions.

Results: Among 567 participants (mean age: 41.9 years; 51.3% female), 12.5% reported at least one SHE. SHE prevalence was 13.5%, 8.9%, 11.1%, and 15.3% for TBR70 categories <2%, 2-3.9%, 4-5.9%, and ≥6%, respectively. Only patients with TBR70≥6% and IAH had a significant increase in SHE risk (2.47 [1.03, 5.91]). For TBR54, SHE prevalence was 11.6%, 11.0%, and 18.1% for categories 0%, 0.1-1.9%, and ≥2%, respectively. Only opatients with TBR54≥2% and IAH had a significantly higher SHE risk (3.47 [1.27, 9.43]).

Conclusions: People with T1D remain at risk of SHE despite the use of technology. TBR70 alone is not discriminative for high-risk SHE, but combining TBR70 with awareness status identifies those at the highest risk, with even better discriminative power when using TBR54.

ASSOCIATIONS OF CGM METRICS WITH STIMULATED C-PEPTIDE MEASURES IN YOUTH WITH RECENT-ONSET TYPE 1 DIABETES

Linda Dimeglio ¹, Colleen Bauza², Bruce Buckingham³, Mark Clements⁴, Gregory Forlenza⁵, Jennifer Sherr⁶, Roy Beck⁷, Lauren Kanapka², Craig Kollman⁸, Shannon Beasley⁹, Erin Cobry¹⁰, Lisa Norland³, Ryan Mcdonough⁴, Michelle Van Name⁶, Antoinette Moran¹¹

¹Indiana University School of Medicine, Indianapolis, United States of America, ²Jaeb Center for Health Research, Tampa, United States of America, ³Stanford University, Stanford, United States of America, ⁴Children's Mercy Hospital, Kansas City, United States of America, ⁵University of Colorado Anschutz Medical Campus, Barbara Davis Center For Diabetes, Aurora, United States of America, ⁶Yale School of Medicine, New Haven, United States of America, ⁷JAEB Center for Health Research, Research, Tampa, United States of America, ⁸Jaeb Center for Health Research Foundation, Inc., Tampa, United States of America, ⁹University of Minnesota, Minneapolis, United States of America, ¹⁰Barbara Davis Center for Diabetes, University of Colorado, Aurora, United States of America, ¹¹University of Minnesote, Minneapolis, United States of America

Background and Aims: The primary endpoint of studies investigating disease-modifying interventions in persons with new-onset type 1 diabetes (T1D) typically are c-peptide measures from mixed-meal tolerance tests (MMTTs) as a surrogate for beta-cell function. Continuous glucose monitors (CGMs) are widely used with increased wearability and accuracy. Given that MMTTs are cumbersome and perceived as burdensome, we investigated if CGM glucose measures could be employed as surrogates for stimulated c-peptide values in persons with new-onset T1D.

Methods: 103 children with recent-onset T1D were enrolled in a trial investigating verapamil and intensive glycemic management (NCT04233034) with AID. Participants used CGM continuously and had five MMTTs performed at 3-month intervals, from baseline to 52-weeks. We compared 14 CGM-measured metrics to MMTT C-peptide measures.

Results: CGM metrics at 52-weeks moderately correlated with c-peptide area under the curve (Table 1). Time-in-range (TIR) and coefficient of variation (CV) were the best individual predictors of peak -c=peptide ≥0.2pmol/mL with AUC-ROCs of 0.71 and 0.72. A multivariate model of TIR, time <70mg/dL, and CV had a sensitivity of 68.4% and specificity of 75.0%.

Conclusions: CGM measures correlate with stimulated c-peptide measures, however, like previous findings in pre-clinical T1D, the strength of the correlations and sensitivity and specificity of CGM-derived measures are not high enough to replace c-peptide measures in new onset T1D clinical trials.

IMPAIRED AWARENESS OF HYPOGLYCAEMIA DOES NOT IMPACT THE DETECTION OF HYPOGLYCAEMIA ON CGM: THE HYPO-METRICS STUDY

Patrick Divilly ^1,2, Natalie Zaremba^2,3, Pratik Choudhary⁴

¹St Vincent's University Hospital, Endocrinology, Dublin, Ireland, ²King's College London, London, United Kingdom, ³Radboud University Medical Centre, Nijmegen, Netherlands, ⁴University Hospitals of Leicester NHS Foundation Trust, Leicester Diabetes Centre, Leicester, United Kingdom

Background and Aims: Over half of sensor-detected hypoglycaemia (SDH) is not recognised by people with type 1 diabetes (T1D). We explored the impact of intact (NAH) vs impaired (IAH) awareness of hypoglycaemia on rates of person-reported hypoglycaemia (PRH) and SDH, and the proportion of SDH that were detected in the Hypo-METRICS study.

Methods: We analysed data from 276 people with T1D (IAH=58; Gold score >4). All wore blinded continuous glucose monitor (CGM) and recorded PRH on the Hypo-METRICS app for 10-weeks. PRH were classified as symptomatic or asymptomatic (detected by technology or routine monitoring). SDH with associated PRH ± 1 hour was considered detected.

Results: Participants with IAH were older (52 vs 44 yrs;p<0.001). There were no significant differences in gender (women 52 vs 54%;p=0.6), CGM use (69 vs 78%;p=0.3), median[IQR] time in range (62[54-73] vs 60[51-71]%;p=0.4) or time <3.9mmol/l (5.4[2.9-7.9] vs 4.3 [2.4-7.9]%;p=0.3) between IAH and NAH. There were no significant differences in the rates of SDH<3.9mmol/l (7.0[4.2-10.6] vs 6.4[3.7-10.2] episodes/week;p=0.4), SDH<3.0mmol/l (1.4[0.5-3.0] vs 1.0[0.4-2.4] episodes/week;p=0.1), PRH (3.3[2.5-5.8] vs 4.0[2.3-5.9] episodes/week;p=0.8), the detection of SDH <3.9mmol (38[29-47] vs 42[28-53]%;p=0.3) or the detection of SDH <3.0mmol/l (46[36-67] vs 47[25-66]%;p=0.6). Participants with IAH reported fewer symptomatic PRH (2.3[1.4-3.7] vs 3.0[1.7-4.9] episodes/week;p=0.007), but higher rates of technology PRH (1.0[0.2-2.3] vs 0.5[0.1-1.2] episode/week;p=0.002) than those with NAH.

Conclusions: We could not differentiate between IAH or NAH with CGM based metrics. In the era of CGM, technological awareness may be a suitable substitute for symptomatic awareness.

GENETIC VARIATION AND HYPOGLYCEMIA IN TYPE 1 DIABETES MELLITUS; A REAL-WORLD STUDY USING CONTINUOUS GLUCOSE MONITORING

Daphne Katsarou ¹, Prodromos Sakaloglou², Panagiota Christou³, Maria Christou³, Eleni Georga⁴, Ioannis Georgiou², Costas Papaloukas⁵, Dimitrios Fotiadis⁶, Stelios Tigas³

¹University of Ioannina, Dept. Of Materials Science And Engineering & Dept. Of Biological Applications And Technology, Ioannina, Greece, ²University of Ioannina, Laboratory Of Medical Genetics In Clinical Practice, Ioannina, Greece, ³University of Ioannina, Endocrinology And Diabetes Department, Ioannina, Greece, ⁴University of Ioannina, Dept. Of Materials Science And Engineering, Ioannina, Greece, ⁵University of Ioannina, Dept. Of Biological Applications And Technology & Biomedical Research Institute, Foundation For Research And Technology-hellas, Ioannina, Greece, ⁶University of Ioannina, Dept. Of Materials Science And Engineering, University Of Ioannina & Biomedical Research Institute, Foundation For Research And Technology, Ioannina, Greece

Background and Aims: Evidence suggests that genetic polymorphism may influence the risk of hypoglycemia in people with type 1 diabetes mellitus (PwT1D). The aim of the study was to examine the frequency and severity of hypoglycemia in relation to certain candidate gene polymorphisms in PwT1D.

Methods: 74 adult PwT1D using continuous glucose monitoring (CGM) took part in the study. TBR₇₀ (time below range, 54–69 mg/dL) and TBR₅₄ (<54 mg/dL) were recorded. Three different gene polymorphisms and their association with hypoglycemia were examined: (i) the ADRB2 Arg16Gly polymorphism, (ii) the ACE I/D polymorphism (insertion/deletion polymorphism of the angiotensin-converting enzyme gene) and (iii) the AGTR2 1675G/A polymorphism. Sanger sequencing was performed for DNA SNP analysis and PCR amplification/gel electrophoresis for the I/D polymorphism detection.

Results: The 1675G/A polymorphism of AGTR2 was associated with higher TBR 54 (2.24±2.77 vs 0.96±1.62%, p=0.04) and numerically higher TBR 70 (5.47±4.47 vs 3.44±3.36, p=0.13) in heterozygous carriers of the A allele compared to those carrying the GG genotype. TBR₅₄ was 2.2-fold higher among patients homozygous for the ACE II polymorphism compared to carriers of the D allele. Homozygosity of the A allele of the ADRB2 Arg16Gly polymorphism was associated with a 1.5-fold higher TBR₇₀ compared with that of carriers of the G allele (p=0.07).

Conclusions: Differences in the rate of hypoglycemia were observed in relation to genetic variants known to affect the risk of hypoglycemia in this population of PwT1D. Larger studies in different populations are needed to further explore these findings.

RISKS OF GLYCAEMIC EXCURSIONS AROUND EXERCISE SESSIONS AND THE DAYS WITHOUT EXERCISE WHEN IMPLEMENTING A 2-MONTH COMBINED TRAINING PROGRAM IN ADULTS WITH TYPE 1 DIABETES

Elodie Lespagnol ¹, Léo Duriez¹, Serge Berthoin¹, Cassandra Parent¹, Julie Dereumetz¹, Sémah Tagougui¹, Angéline Melin¹, Madleen Lemaitre², Pierre Fontaine³, Anne Vambergue⁴, Remi Rabasa-Lhoret⁵, Elsa Heyman¹

¹Univ. Lille, Univ. Artois, Univ. Littoral Côte d’Opale, ULR 7369 - URePSSS - Unité de Recherche Pluridisciplinaire Sport Santé Société, Lille, France, ²Department of Diabetology, Endocrinology, Metabolism and, Nutrition, Lille University Hospital, Lille, France, Lille, France, ³Santelys Loos, Université de Lille, Lille, France, ⁴CHRU de Lille, Département D’endocrinologie Diabétologie Nutrition, Lille, France, ⁵Institut de recherches cliniques de Montréal (IRCM), Montréal, Canada

Background and Aims: Implementation of exercise programs in individuals with type 1 diabetes (T1D) may precipitate glycaemic fluctuations, which should be better characterized for developing appropriate glucose management recommendations. We aimed to assess glycaemic excursions and their progression throughout a 2-month training program, distinguishing those observed around exercise sessions from those of non-exercising days.

Methods: Nineteen (13 female) adults with T1D participated in 2-3 supervised exercise (aerobic/strength) 90-min weekly sessions over 2 months. Glycaemic excursions (continuous glucose monitoring) were calculated for specific periods (24-hour, nocturnal; periods before, during, after exercise sessions) and compared between day with and without exercise session (linear mixed models, logistic regressions).

Results: During the 2-month training, nights following exercise showed reduced hyperglycaemic (>10.0mmol⋅L⁻¹) risk vs. non-exercise nights. This difference attenuated with weeks of training, in parallel to a progressive increased risk of time >16.7mmol⋅L⁻¹ during early and late recovery of exercise sessions. Overall, independently of exercise session occurrence, level 1 and 2 nocturnal hypoglycaemic risk increased as the training weeks passed.

Conclusions: Initially, acute exercise sessions reduced nocturnal hyperglycaemia without increasing hypoglycaemia. However, nocturnal hypoglycaemia risk progressively increased after several weeks of training, highlighting the necessity for vigilant glycaemic supervision, particularly at night, even on non-exercise days.

GLYCEMIC VARIABLITY AND TIME IN RANGE AMONG TODDLERS AND PRESCHOOLERS WITH TYPE 1 DIABETES, SINGLE CENTRE EXPERIENCE

Safinaz Elhabashy¹, Nouran Salah ², Eman Sakr²

¹ain shams university, cairo, Egypt, ²Ain Shams University, Pediatrics, Cairo, Egypt

Background and Aims: Attaining tight glycemic control without risking hypoglycemia is crucial in toddlers and preschoolers with type 1 diabetes (T1D) to avoid cognitive impairment later in life. Hence, this study aims to assess the glycemic parameters among toddlers and preschoolers with T1D in relation to different basal insulins.

Methods: Sixty toddlers and preschoolers with T1D with mean age 3.53 ± 1.17 years (range, 2-6) and mean diabetes duration 9.37 ± 1.85 months were assessed for insulin type and dose, auxological parameters, frequency of hyouglycemia and DKA and continuous glucose-monitoring (CGM) for glycemic parameters.

Results: The mean time in range (TIR) of the studied cohort was 55.07 ± 24.05%, and their mean coefficient-of variation (CV) was 42.82 ± 11.69%. The TIR was significantly higher in the degludec group (69.36 ± 18.54) and the glargine group (55.43 ± 26.51) than the NPH group (32.56 ± 9.11), p<0.001. Meanwhile, the CV was significantly lower in the degludec group (35.12 ± 6.47) than the gargine (44.1 ± 13.13) and the NPH (53.8 ± 7.54) groups, p<0.001. The insulin degludec and glargine groups had significantly lower HbA1c (p=0.002), hypoglycemia (p=0.006), severe hypoglycemia (p=0.029) and IDD (p=0.015) than the NPH group.

Conclusions: Conclusion: Toddlers with T1D have poor glycemic control in the form of low time in range and high glycemic variability. Insulin degludec and glargine resulted in better HbA1c and TIR with reduced hypoglycemia, IDD and CV than NPH among toddlers and preschoolers with T1D.

GLYCEMIC CONTROL WITH CONTINUOUS GLUCOSE MONITORING IN PEOPLE WITH TYPE 2 DIABETES REGARDLESS OF THERAPY: A PROPENSITY-SCORE MATCHED DATA ANALYSIS FROM A LARGE DATA SET

Satish Garg ¹, Irl Hirsch², Janet Snell-Bergeon³, Enrico Repetto⁴, Brian Ulmer⁴, Christopher Perkins⁴, Richard M. Bergenstal⁵

¹Barbara Davis Center for Diabetes, University of Colorado, Medicine And Pediatrics, Aurora, United States of America, ²University of Washington, seattle, United States of America, ³Barbara Davis Center for Diabetes Adult Clinic, Aurora, United States of America, ⁴Roche, Indiannapolis, United States of America, ⁵International Diabetes Center, HealthPartners Institute, St Louis Park, United States of America

Background and Aims: Use of continuous glucose monitoring (CGM) improves glucose control in individuals with both type 1 diabetes (T1D) and type 2 diabetes (T2D) treated with insulin and non-insulin therapies. However, the impact of the frequency of CGM use on HbA1c in patients with T2D has not been evaluated.

Methods: This retrospective 12-month propensity-score analysis, matched for age, gender and ethnicity (controls), used Optum's de-identified Market Clarity Data of >79 million people. We evaluated HbA1c values after initiating CGM in a cohort of insulin-treated and non-insulin-treated adults with T2D. Treatment groups were non-insulin therapies (NIT), basal insulin therapies (BIT), and prandial insulin therapies (PIT). Frequency was categorized by days of CGM sensor use: >0 to £60 sensor days (LOW), >60 to £300 sensor days (MEDIUM), and >300 sensor days (HIGH). The primary outcome was change in HbA1c among CGM users compared with control patients not using CGM.

Results: A total of 9,612 adults (³18 years) with suboptimally controlled T2D (Control n=4,806, LOW n=928, MEDIUM, n=1,679, HIGH, n=2,199) were included in the analysis. Individuals with >80% (>300 sensor days) CGM use achieved a 0.6% HbA1c reduction compared with no CGM use (Figure). HbA1c values decreased as CGM frequency increased.

Conclusions: Greater and sustained improvements in glycemic control with CGM use were observed in adults with T2D suboptimally controlled glycemia treated with intensive and non-intensive therapies compared to those without CGM use. Higher CGM frequency was associated with greater HbA1c reductions.

USE OF CONTINUOUS GLUCOSE MONITORING IN PEOPLE WITH TYPE 2 DIABETES AND ITS EFFECT ON METABOLIC CONTROL

Jorge González Prieto ¹, Marta M. Guijarro Chacón², Lucia Babiano-Benito¹, Laura Martín Flores¹, Francisco Miguel Morales Perez², Claudia García Lobato²

¹Hospital Universitario Badajoz, Badajoz, Spain, ²Hospital Universitario Badajoz, Endocrinología Y Nutrición, Badajoz, Spain

Background and Aims: There is evidence about the benefit of continuous glucose monitoring (CGM) systems in people with type 1 diabetes (T1D). However, less clear is in patients with type 2 diabetes (T2D). The objective was to assess whether our patients with T2D on a basal-bolus regimen show an improvement in their metabolic control.

Methods: Individuals with T2D who had used FreeStyle Libre 2 with download available in Libreview, were evaluated. Parameters related to monitoring use at 3 months and 1 years after CGM initiation were retrospectively compared with the initial download.

Results: A total of 80 individuals with T2D were included: age: 70±10 years (50 - 90); 57% (n = 46) men; use of CGM 1,14±0,52 years; HbA1c prior to CGM initiation: 7.85 ±1,16%). The results of the metabolic analysis are shown in Table 1. * GMI: Glucose Management Indicator. CV: Coefficient of variation. Data expressed as mean ± SD HbA1c after to CGM initiation: 7,57 ± 1,43 (p=0,067). Last HbA1c 7,61 ± 1.03 (p=0.430).

Conclusions: The use of continuous glucose monitoring in people with type 2 diabetes has shown benefits in improving metabolic control and decreasing hypoglycemia in some studies. However, no metabolic benefits were observed at 3 months and 1 year of use compared with the initial download in our cohort. Furthermore, no improvement in glycosylated hemoglobin was found. Despite this, other benefits such as improved quality of life could justify its use in in these people.

DERIVED TIME-IN-RANGE (DTIR) IN PEOPLE WITH TYPE 2 DIABETES (PWT2D) SWITCHING TO IGLARLIXI FROM PREMIXED INSULIN: A POST-HOC ANALYSIS OF THE SOLI-SWITCH STUDY

Martin Haluzík ¹, Malha Azzouz², Mohammed Al Mehthel^3,4, Jothydev Kesavadev⁵, Soon Jib Yoo⁶, Ramazan Sari⁷, Khier Djaballah⁸, Lydie Melas-Melt⁹, Felipe Lauand⁸

¹Diabetes Centre, Institute for Clinical and Experimental Medicine (IKEM), Prague, Czech Republic, ²Department of Endocrinology, Faculty of Medicine, University of Algiers, Algiers ¹6000, Algiers, Algeria, ³Department of Endocrinology, The University of British Columbia, Vancouver, British Columbia, Canada, ⁴Obesity, Endocrine and Endocrine Centre, King Fahad Medical City, Riyadh, Saudi Arabia, ⁵Jothydev's Diabetes Research Centre, Trivandrum, Kerala, India, ⁶Division of Endocrinology and Metabolism, Department of Internal Medicine, Bucheon St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul, Korea, Republic of, ⁷Division of Endocrinology and Metabolism, Department of Internal Medicine, Faculty of Medicine, Akdeniz University, Antalya, Turkey, ⁸Sanofi, paris, France, ⁹Ividata Life Sciences on behalf of Sanofi, Levallois-Perret, France

Background and Aims: Time-in-range (TIR) is linked to HbA1c, 5% increase in TIR leads to significant improvements in glycemic control and fewer diabetes complications. Derived TIR (dTIR) represents proportion of self-measured plasma glucose (SMPG) measurements within target glucose range. This Soli-SWITCH post-hoc analysis assessed dTIR (using capillary SMPG) in PWT2D switching to iGlarLixi from premixed-insulin.

Methods: Soli-SWITCH (EudraCT: 2021-003711-25), a 24-week, single-arm study, included PWT2D (≥18 years) who received premixed-insulin (≥3 months and <10 years) ± oral antidiabetics, switched to iGlarLixi at baseline. Primary endpoint of this post hoc analysis was change in dTIR (baseline to Week 24). Secondary endpoints were: changes in derived time-above-range (dTAR) and time-below-range (dTBR) (baseline to Week 24), proportion of participants with dTIR increase ≥5%, and those with dTIR >70%, dTAR <25%, and dTBR <4% (at Week 24). dTIR (70 − 180 mg/dL), dTAR (>180 mg/dL), dTBR (<70 mg/dL) were calculated from 7-point SMPG values.

Results: From baseline to Week 24, least squares mean change (95% CI) in dTIR and dTAR was 30.4% (26.4, 34.5) and −30.4% (−34.5, −26.4), respectively, while mean dTBR remained unchanged (0.0%; 95% CI: −0.6, 0.7) (Table). At Week 24, 65.6% (57.6, 73.0) and 55.4% (47.3, 63.3) participants achieved dTIR ≥70% and dTIR increase ≥5%, respectively. Additionally, 54.8% (46.6, 62.7) participants achieved triple target of dTIR >70%, dTAR <25%, and dTBR <4%.

Conclusions: In PWT2D switching from premixed-insulin, iGlarLixi improved dTIR and reduced dTAR, without increasing dTBR, making it an effective once-daily option to simplify insulin therapy without increasing hypoglycemia risk.

COST-EFFECTIVENESS OF REAL-TIME CONTINUOUS GLUCOSE MONITORING SYSTEM VERSUS SELF-MONITORING OF BLOOD GLUCOSE IN PEOPLE WITH TYPE 2 DIABETES ON INSULIN THERAPY

Johan Jendle ¹, Sabrina Ilham², Hamza Alshannaq², Gregory Norman², Peter Lynch²

¹Örebro University, School Of Medicine, Department Of Medical Sciences, Örebro, Sweden, ²Dexcom, Inc., Global Access, Value And Evidence, San Diego, United States of America

Background and Aims: Real-time continuous glucose monitoring (RT-CGM) improves glycemic control and results in cost savings for individuals with Type 2 Diabetes on insulin therapy (T2D IT). We examined the cost-effectiveness of RT-CGM for people with T2D IT in Sweden.

Methods: The analysis used IQVIA CORE Diabetes Model v10 over a 50-year time horizon from a Swedish payor perspective. Cohort characteristics and clinical effectiveness data were sourced from the Steno2Tech trial, demonstrating a 0.9% reduction in HbA1c and a 1.1 kg/m² reduction in body mass index (BMI) in RT-CGM vs. SMBG users after 12-month of follow-up. HbA1c and BMI trajectories were based on the Swedish National Diabetes Registry. A utility benefit for avoidance of fingerstick (0.03) was applied to RT-CGM users. RT-CGM (Dexcom ONE+) and SMBG costs were based on local prices. Costs and outcomes were discounted at 3% rate. Sensitivity analyses were conducted.

Results: RT-CGM was associated with an incremental quality-adjusted life year (QALY) of 0.458, incremental lifetime costs of 130,717 Swedish Krona (SEK), and an incremental cost-effectiveness ratio (ICER) of SEK 285,221 per QALY compared with SMBG, significantly below the willingness to pay threshold of SEK 500,000 in Sweden. Sensitivity analyses modeling HbA1c and BMI trajectories using UKPDS equations reduced the ICER by 45%, and RT-CGM became cost-saving under an upward HbA1c progression scenario. RT-CGM users experienced a relative reduction in cumulative incidence of ocular, renal, and neurological outcomes by 3.19%, 3.16%, and 2.76% respectively.

Conclusions: Dexcom ONE+ is a cost-effective intervention for people with T2D IT in Sweden compared to SMBG.

QUANTIFYING INSULIN RESISTANCE VIA CONTINUOUS GLUCOSE MONITORING: FROM CONCEPT TO CLINICAL REALITY - A PRELIMINARY STUDY AMONG DIVERSE DIABETES PHENOTYPES

Shmuel Levit ¹, Elvira Ahmetgalieva², Chen Ryder³, Ildar Musin², Vyacheslav Levit⁴

¹Institute of Endocrinology, Assuta Medical Center, Diabetes & Metabolism, Tel Aviv, Israel, ²Kazan National Research Technological University, Department Of Medical Innovations, Kazan, Russian Federation, ³Western Galilee Academic College, Brain & Behavior Research Institue, Acre, Israel, ⁴Clinical City Hospital N8, Cheliabinsk, Russian Federation

Background and Aims: Precise quantification of insulin resistance (IR) remains a clinical challenge. This study aimed to develop a novel IR calculation method using continuous glucose monitoring data. Our objectives were to enhance diabetes diagnosis accuracy and facilitate personalized treatment strategies. We sought to provide clinicians with a practical tool for precise IR assessment.

Methods: This retrospective study evaluated 42 subjects (5 controls, 37 diabetics), categorizing them into No, Mild, and High IR groups based on age, BMI, diabetes onset, glucose toxicity, inflammation, insulin needs, nocturnal hyperglycemia, medications, and endocrinopathies. A proprietary algorithm derived an IR parameter from CGM data (Medtronic Guardian 4). One-way ANOVA with post-hoc Tukey HSD tests was used for comparisons between groups.

Results: ANOVA revealed significant differences (F(3,38) = 56.112, p<0.001, η² = 0.816). The High IR group showed higher levels compared to all other groups (p<0.001). No significant differences were observed between Control, No IR, and Mild IR groups. Table 1: Insulin Resistance Levels Across Study Groups

Insulin resistance: ^aHigh, ^bLow (^b<^a significantly). ****p<0.001. Significant differences between lettered groups.

Conclusions: This study demonstrates significant differences in CGM-derived IR parameters between clinically assessed groups, particularly in distinguishing high IR. The non-invasive nature of this method could revolutionize IR assessment. Further research with larger sample sizes is needed to validate these findings across the full spectrum of insulin resistance.

DERIVED TIME-IN-RANGES IN INSULIN-NAÏVE OLDER ADULTS WITH TYPE 2 DIABETES TREATED WITH INSULIN GLARGINE 300 U/ML VS DEGLUDEC 100 U/ML: A BRIGHT TRIAL POST-HOC ANALYSIS

Thomas Martens ¹, Andreas Liebl², Yingying Luo³, Maria Aileen Mabunay⁴, Felipe Lauand⁵, Lydie Melas-Melt⁶, Javier Escalada⁷

¹International Diabetes Center, HealthPartners Institute, Minnesota, United States of America, ²Center for Diabetes and Metabolism, m&i-Fachklinik, Bad Heilbrunn, Germany, ³Peking University People’s Hospital, Department Of Endocrinology And Metabolism, Beijing, China, ⁴Sanofi, Singapore, Singapore, ⁵Sanofi, Paris, France, ⁶Ividata Life Sciences, Levallois-Perret, France, ⁷Clínica Universidad de Navarra, Pamplona, Spain

Background and Aims: Optimal glycaemic control with minimal hypoglycaemia is vital for people with type 2 diabetes (PWT2D) with advancing age. BRIGHT study compared the efficacy and safety of insulin glargine 300 U/mL (Gla-300) vs insulin degludec 100 U/mL (IDeg-100) in insulin-naïve PWT2D. This post-hoc analysis assessed derived time-in-range (dTIR) (using capillary 8-point self-monitored blood glucose profiles) in PWT2D comparing Gla-300 vs IDeg-100 in participants aged <70 and ≥70 years.

Methods: BRIGHT (NCT02738151) was multicentre, randomised, parallel-group, 24-week trial. This analysis assessed changes in dTIR (70–180 mg/dL), derived time-below-range (dTBR <70 mg/dL), derived time-above-range (dTAR >180 mg/dL) from baseline to Week 24 and number of participants (%) achieving dTIR ≥70%, ≥5% improvement in dTIR and triple target (dTIR >70%, dTBR <4%, dTAR <25%) at Week 24.

Results: From 768 (<70 years) and 161 participants (≥70 years), 392 (<70 years) and 74 (≥70 years) received Gla-300 and 376 (<70 years) and 87 (≥70 years) received IDeg-100 at baseline. Mean±SD dTIR change in participants <70 years was +29.41±32.45% (Gla-300) and +25.56±33.10% (IDeg-100) while +33.95±32.01% (Gla-300) and +26.27±27.58% (IDeg-100) for participants ≥70 years at Week 24. Proportion of participants achieving triple target were 42.2% (<70 years) and 38.4% (≥70 years) with Gla-300 while 39.7% (<70 years) and 33.3% (≥70 years) with IDeg-100 at Week 24 (Table 1). At Week 24, proportion of glucose values in dTIR for participants receiving Gla-300 were 77.2% (<70 years) and 79.8% (≥70 years) and for IDeg-100 were 77.4% (<70 years) and 74.9% (≥70 years) (Figure 1). dTBR was low and at target for all groups.

Conclusions: In PWT2D, Gla-300 and IDeg-100 improved dTIR, reduced dTAR, without increasing dTBR in older participants (≥70 years).

IMPACT OF PRIMARY CARE-INITIATED CONTINUOUS GLUCOSE MONITORING ON CLINICAL OUTCOMES IN PATIENTS WITH TYPE 2 DIABETES

Jovan Milosavljevic ¹, Clyde Schechter², Shivani Agarwal¹

¹Fleischer Institute for Diabetes and Metabolism, Division of Endocrinology, Albert Einstein College of Medicine and Montefiore Medical Center, Bronx, United States of America, ²Montefiore Medical Center, Albert Einstein College of Medicine, Department Of Family And Social Medicine, Bronx, United States of America

Background and Aims: This study aimed to evaluate the effects of CGM prescribed by primary care on glycemic control and hospitalization outcomes in a large safety-net hospital network.

Methods: We conducted a retrospective cohort study of adults with T2D on diabetes medications seen in primary care from July 2020 to July 2023. Primary outcomes were changes in HbA1c (linear mixed-effect model) and the incidence of hospitalizations and/or emergency department (ED) visits (recurrent event frailty Cox regression model), using time-fixed and time-varying covariates.

Results: Among 32,020 patients with T2D (mean age 62.0 ± 13.6 years; 59% female; 41% Hispanic; 38% non-Hispanic Black; 14% Medicaid; 34% Medicare), 3,457 (10.8%) were prescribed CGM by primary care providers. CGM initiators experienced worsening HbA1c levels prior to starting CGM and demonstrated significant reversal and improvement of HbA1c after initiation (-0.42 within the first year [95% CI -0.51 to -0.32]) (Figure 1). CGM users achieved a 0.5%-point HbA1c reduction significantly faster than non-users (median time: users 308 days [95% CI 288–330]; non-users 840 days [95% CI 822–859]; HR 1.87 [95% CI 1.78–1.96]). CGM initiation was associated with a reduced risk of recurrent all-cause hospitalizations and ED visits (HR 0.86 [95% CI 0.82–0.90]).

Conclusions: CGM prescribed in primary care settings for patients with T2D resulted in accelerated glycemic control and reduced hospitalizations. These findings support the need for broader CGM implementation in primary care to improve outcomes, particularly for underserved populations.

FRONTIER – FREESTYLE LIBRE USE IN ONTARIO AMONG PATIENTS WITH T2D – EVIDENCE FROM REAL-WORLD PRACTICE: PATIENTS AGED >65 YEARS ON NON-INSULIN THERAPIES

Alexandria Ratzki-Leewing ¹, Stewart Harris¹, Remi Rabasa-Lhoret², Yeesha Poon³

¹Western University, Epidemiology And Biostatistics, London, Canada, ²Montreal Clinical Research Institute, Montréal, Canada, ³Abbott Diabetes Care, Alameda, United States of America

Background and Aims: This study investigated healthcare resource utilization (HCRU) and HbA1c at a population level before and after FreeStyle Libre (FSL) adoption in people with T2D aged >65 years on non-insulin therapies GLP-1 RA and/or oral therapy).

Methods: The large, real-world population in the Institute for Clinical Evaluative Sciences database (Ontario, Canada) was used to identify patients aged >65 years with T2D using GLP-1 RA and/or oral therapy who initiated FSL and remained active on FSL for 24 months (Figure 1). ED visits and hospitalization (overall, with DKA, and with hypoglycemia) were quantified during the 12 months before the index date and the last 12 months of the 24-month follow-up period. HbA1c data were from the last tests in each period.

Results: 1,774 patients were included (GLP-1 RA ± lifestyle/oral glucose lowering therapy, 273; oral therapy only, 1,501). After starting FSL, HbA1c was statistically significantly reduced in both groups, with a larger mean reduction in the GLP-1 RA group than the oral therapy group (−0.56% vs. −0.27%; Figure 2). Overall, ED visits and hospitalization fell in both groups, with statistically significant reductions in the oral therapy group (Table). Lower trends were seen for ED visits and hospitalization with DKA and hypoglycemia.

Conclusions: For people with T2D aged >65 years on GLP-1 RA and/or oral therapy, initiation of FSL was associated with a significant mean reduction in HbA1c. HCRU improvements were significant only for patients using oral therapy, which may reflect the smaller number of patients using GLP-1 RA.

SHORT-TERM CONTINUOUS GLUCOSE MONITORING REVEALS BLOOD GLUCOSE INSIGHTS AND PROMOTES BEHAVIORAL AWARENESS IN PEOPLE WITH NON-INSULIN-TREATED TYPE 2 DIABETES, ESPECIALLY WITH SHORTER DIABETES DURATION

Tanja Thybo, Emma Nielsen, Nuri Cayuelas I Mateu, Anne-Marie Karin Wegener

Danish Diabetes Association, Glostrup, Denmark

Background and Aims: CGM has shown potential as an educational tool for persons with Type 2 Diabetes (PwT2D) not receiving insulin therapy. This study investigates which insights PwT2D gain regarding the influence of their lifestyle on blood glucose levels during a two-week period of CGM usage with minimal external guidance.

Methods: As a walk-in setup in 8 cities in Denmark, 455 PwT2D not receiving insulin therapy were recruited to wear a CGM device for up to two weeks. Minimal instructions were given, and participants completed a baseline questionnaire on-site and an emailed follow-up survey after two weeks to assess their CGM usage and their insights gained regarding lifestyle impacts on glucose levels.

Results: 340 completed the final survey. Among these, 75% used the CGM for two weeks and 80% found the CGM highly useful and informative. A larger share of the participants with a diabetes duration of less than five years compared to all other participants reported to have gained insights into how specific food items (88% vs. 78%), portion sizes (80% vs. 60%), and physical activity (65% vs. 55%) affected their glucose levels. Notably, the knowledge gained was not influenced by the participants’ educational length.

Conclusions: Short-term CGM usage with minimal guidance provides benefits for PwT2D not treated with insulin. It offers personalized insights into the effects of lifestyle choices on glucose levels, which can support individualized therapy decisions and promote behavioral changes that may enhance glycemic control. The duration of diabetes influences the insights gained, while education level has minimal impact on knowledge acquisition.

THE IMPACT OF SHORT-TERM CONTINUOUS GLUCOSE MONITORING COMBINED WITH PATIENT EDUCATION ON HBA1C LEVELS IN PATIENTS WITH TYPE 2 DIABETES: AN ONGOING CLINICAL TRIAL

Nuphar Vinegrad ^1,2,3, Sara Menashe Auman^3,4,5 , Liron Sabag^3,4,5, Peleg Naftali^3,4,5, Shulamit Witkov^3,4, Doron Pasternak⁶, Idit F Liberty^2,3

¹Soroka Univerity medical Center, Endocrinology, Beer Sheva, Israel, ²Ben Gurion University, The Faculty Of Health Sciences At Ben-gurion University Of The Negev, Beer Sheva, Israel, ³Soroka University Medical Center, Diabetes Clinic, Beer Sheva, Israel, ⁴Soroka University Medical Center, Diet And Nutrition Unit, Beer Sheva, Israel, ⁵Ben Gurion University, the Department Of Epidemiology, Biostatistics And Community Health Sciences, Beer Sheva, Israel, ⁶Soroka University Medical Center, Clinical Reaserch Center, Beer Sheva, Israel

Background and Aims: Diabetes is a leading cause of morbidity and mortality, with better effective glycemic control linked to fewer complications. Continuous glucose monitoring (CGM) may help improve glycemic management. This study evaluates the impact of a 7- to 10-day intensive educational intervention using CGM on patients with type 2 diabetes receiving optimal medical therapy (GLP-1 agonists, Metformin, and SGLT2 inhibitors) and HbA1c levels ≥8%.

Methods: We conducted an open-label, randomized, clinical trial including 34 patients. The intervention group used CGM, while the control group conducted 4 daily finger-prick glucose measurements. Both groups documented their meals. Following the monitoring period, all participants received tailored medical and lifestyle recommendations from a diabetes specialist and a registered dietician diabetes educator.

Results: Baseline characteristics were comparable between groups. Three months post-intervention, both groups showed a significant reduction in HbA1c with an average decrease of 0.42% (p=0.035). However, no significant difference was observed in ΔHbA1c between the CGM and the control group. Notably, patients in the CGM group required fewer medical interventions and received more lifestyle modifications compared to the control group, where all patients received additional medications, including basal insulin. Patients using CGM reported greater convenience.

Conclusions: Short-term CGM combined with personalized medical and nutritional guidance may enhance glycemic control in patients with type 2 diabetes, as assessed 3 months post-intervention. Although both the CGM and the daily finger-pricks group improved glycemic control, CGM users reported greater satisfaction and required fewer additional treatments. Further studies are needed to evaluate the long-term effects of this approach.

DIAGAME DIGITAL TWIN AND SERIOUS GAMING FOR PATIENTS WITH DIABETES

Ryan De Vries ¹, Edouard Reinders², Elma Cruts – Van De Meulengraaf², Pleun Wouters - Van Poppel², Pieter Van Gorp³, Peter Hilbers¹, Uzay Kaymak⁴, Harm Haak⁵, Natal Van Riel¹

¹Eindhoven University of Technology, Biomedical Engineering, Eindhoven, Netherlands, ²Máxima MC, Eindhoven, Netherlands, ³Eindhoven University of Technology, Industrial Engineering & Innovation Sciences, Eindhoven, Netherlands, ⁴Jheronimus Academy of Data Science, ‘s-Hertogenbosch, Netherlands, ⁵Eindhoven University of Technology, Eindhoven, Netherlands

Background and Aims: The DiaGame project develops and tests a new concept to empower patients with diabetes in their self-management. We combine real world data with advanced computational algorithms and models and serious gaming.

Methods: Data has been collected in daily-life settings in sixty participants (52 type 2 diabetes, 8 type 1 diabetes) using a continuous glucose monitor, smartwatch with in-house software, and public smartphone application acquiring data on blood glucose levels, heart rate, step count, acceleration, participant-reported dietary intake, physical activity, insulin dosing, and mood. We develop and validate a physiology-based mathematical model of the glucose-insulin system to describe and predict glycemic dynamics. This model simulates glucose and insulin concentrations following a mixed meal or OGTT and has previously been validated.

Results: The model has been successfully extended to describe the subcutaneous absorption of various insulin analogues, which is important for patients using insulin (type 1 diabetes and type 2 diabetes with intensive insulin therapy). The model accurately captures patient-specific glycemic dynamics of people with T1D. To realize personalized models for T2D patients it is important to consider residual insulin production capacity. By analyzing and comparing personalized models the independent and coherent effects of various lifestyle factors on glycemic dynamics are elucidated.

Conclusions: Patient digital twins are developed by updating personalized prediction models with patient-gathered, real world data. Ultimately, these digital twins will be employed to aid self-management education through a personalized interactive serious game. A prototype of this game is currently being evaluated by patients and caregivers.

UNDERSTANDING CONTINUOUS GLUCOSE MONITOR (CGM) AND INSULIN PUMP USE IN FEDERALLY QUALIFIED HEALTH CENTERS IN THE U.S

Ashby F. Walker ¹, Benjamin Churba¹, Gabby Malden¹, Timothy Long², Warria Esmond³, Young-Rock Hong¹, Crystal Walker², Matthew Gurka⁴, Kyra Vandoren³, Nicole Marlow¹, Stephanie Tejo Martinez², Angel Rivas², Sarah Westen⁵, Lucia Siman³, Andrew Hamilton³, Zihan Zhu³, Arjit Jain³, Michael Haller⁶

¹University of Florida, Health Services Research, Management, And Policy, Gainesville, United States of America, ²Health Choice Network, Miami, United States of America, ³AllianceChicago, Chicago, United States of America, ⁴University of Virginia, Public Health Sciences, Charlottesville, United States of America, ⁵University of Florida, Clinical And Health Psychology, Gainesville, United States of America, ⁶University of Florida, Pediatrics, Gainesville, United States of America

Background and Aims: Federally Qualified Health Centers (FQHCs) are medical homes for 30 million underserved individuals in the U.S. However, research regarding technology use in diabetes has not included data from FQHCs. This study aimed to describe technology use in a national cohort of FQHC patients with Type 1 Diabetes (T1D) and Type 2 Diabetes (T2D) treated with either multiple daily injections (MDI) of insulin or continuous subcutaneous insulin infusion (CSII).

Methods: A large data platform from 40 FQHCs in 17 states was leveraged to identify and describe documented use of CGM and insulin pumps for adults with T1D/T2D. Data was summarized according to sex, race and ethnicity, Federal Poverty Level (FPL), language preference, and Healthcare Effectiveness Data and Information Set (HEDIS) diabetes-related metrics.

Results: A cohort of n=16,267 was identified: 2,249 with T1D and 14,018 with T2D (MDI); 53.8% were below the poverty line (FPL<100); 54.3% female; 23.4% Non-Hispanic White(NHW) 28.0% Non-Hispanic Black(NHB), 0.4% American Indian/Alaska Native(AIAN), 2.0%, Native Hawaiian/Pacific Islander(NHPI), 1.2% Non-Hispanic Asian, and 39.0% Hispanic, and 5.9% Other/Multiracial; 47.6% publicly insured, 22.1% uninsured, and 30.3% privately insured; 34% non-English speaking; 25.0% rural and 75.0% urban. For T1D, 59.2% had HbA1c >9% with documented use <5% for CGM and <8% for insulin pumps; for T2D 50.5% had HbA1c >9% with documented use <2% for CGM and <3% for insulin pumps.

Conclusions: There is incredibly low CGM and pump use documented for people with T1D and T2D seen for care at FQHCs across all demographic groups. Interventions to increase technology equity must include FQHC populations.

DOES TECHNOLOGY USE AFFECT THE TRANSFER OF YOUTH WITH TYPE 1 DIABETES FROM PEDIATRIC TO ADULT CARE?: RESULTS FROM THE INTERNATIONAL SWEET DATABASE

Barbara Piccini ¹, Reinhard Holl², Carine De Beaufort³, Christina Kanaka Gantenbein⁴, Faisal Malik⁵, Hwa Young Kim⁶, Itxaso Rica⁷, Jawad Mirza⁸, Joseph Leung⁹, Lukana Preechasuk¹⁰, Nancy Elbarbary¹¹, Violeta Iotova¹², Sonia Toni¹

¹Meyer University Children's Hospital IRCCS, Diabetology And Endocrinology Unit, Florence, Italy, ²University of Ulm, Institute Of Epidemiology And Medical Biometry- Zibmt, Ulm, Germany, ³Faculty of Science - Technology and Medicine- University of Luxemburg, Pediatric Clinic/centre Hospitalier De Luxemburg, Luxembourg, Luxembourg, ⁴Aghia Sophia Children's Hospital, Division Of Endocrinology, Diabetes And Metabolism Of The First Dept Of Pediatrics Of The National And Kapodistrian University Of Athens Medical School, Athens, Greece, ⁵University of Washington School of Medicine, Department Of Pediatrics, Seattle, United States of America, ⁶Seoul National University Bundang Hospital, Division Of Pediatric Endocrinology And Metabolism, Department Of Pediatrics, Seongnam, Korea, Republic of, ⁷Hospital Universitario Cruces, Pediatric Endocrinology Department, Barakaldo, Spain. Ciberdem. Biocrucesbizkaia, Barakaldo, Spain, ⁸Mubarak alkabeer hospital, Ministry of health, Pediatric Endocrine And Diabetic Unit, Department Of Pediatrics, Kuwait City, Kuwait, ⁹University of British Columbia, Endocrinology Department, Vancouver, Canada, ¹⁰Faculty of Medicine Siriraj Hospital- Mahidol University, Siriraj Diabetes Center Of Excellence, Bangkok, Thailand, ¹¹Ain Shams University, Diabetes Unit, Department Of Pediatrics, Cairo, Egypt, ¹²Medical University of Varna, Department Of Pediatrics, Varna, Bulgaria

Background and Aims: Transfer of youth with type 1 diabetes mellitus (T1DM) from pediatric to adult care is extremely heterogeneous. This study aims to examine the influence of technology use on age of transfer worldwide.

Methods: The SWEET (Better control in Pediatric and Adolescent diabeteS: Working to crEate CEnTers of Reference) database was analyzed (2016-2023, centers >20 patients, > 5 years data). Transfer was assumed in T1DM patients 14-24 years without visits for≥2 years. Presumed age at transfer and technology use (pump [CSII], continuous glucose monitoring [CGM], closed loop [CL], telemedicine [TM]) were evaluated across geographical regions.

Results: Among 33428 patients (51% male, 105 centers), estimated age at transfer of T1DM patients (90%) was 18.6 years in Asia/Middle East/Africa, 17.2 in Australia/New Zealand, 18.3 in North America, 17.0 in South America and 18.0 in Europe (p <0.0001). CSII/CGM use was associated with earlier transfer in Europe and conversely with later transfer in North America (>1 year gap, [Table 1]).

Conclusions: This study demonstrates significant heterogeneity in age at transfer of T1DM youth to adult care worldwide. Technology use may influence the transfer differently across regions, requiring additional research.

THE IMPACT OF DIABETES TECHNOLOGIES ON PSCYHOSOCIAL ASPECTS OF LIFE WITH TYPE 1 DIABETES AMONG ADULT MALES AND FEMALES: BODY SHAPE DISSATISFACTION AND SOCIAL STIGMATIZATION

Basak Sasmazer ¹, Mevhibe Irem Yildiz², Sukru Keles³, Selcuk Dagdelen⁴

¹Hacettepe University Faculty of Medicine, Department Of Internal Medicine, Ankara, Turkey, ²Hacettepe University Faculty of Medicine, Department Of Psychiatry, Ankara, Turkey, ³Karadeniz Technical University Faculty of Medicine, Department Of Medical History And Ethics, Trabzon, Turkey, ⁴Hacettepe University Faculty of Medicine, Department Of Endocrinology And Metabolic Diseases, Ankara, Turkey

Background and Aims: Insulin pumps and sensors might trigger body dissatisfaction and social stigmatization mostly among adolescent females. Here we aimed to investigate impact of technology-use on body dissatisfaction and social stigmatization in an adult type 1 diabetic population with both sexes.

Methods: Patients with T1DM on MDI (controls, n: 72, F/M: 41/31) and insulin pump (cases, n:45, F/M: 33/12) were included in a prospective case-control study. A hybrid model involving both quantitative and qualitative research methods was used. Body Cathexis Scale (BCS), Rosenberg Self-Esteem Scale (RSES), Multidimensional Scale of Perceived Social Support (MSPSS), and Type 1 Diabetes Stigma Assessment Scale (DSAS-1) were utilized for quantitative evaluation. In-depth standardized interviews were conducted with 20 cases, as part of a qualitative research method. Pregnancy and Munchausen Syndrome were excluded.

Results: Cases had lower body satisfaction without statistical significance [BCS-scores: 140 (132-154) vs 148 (138-163), p=0.077]. RSES-scores were similarly comparable between study groups [1 (0-2) vs 1 (0-2), p=0.209]. Perceived social support were similarly comparable [MSPPS: 73 (58-80) vs 72 (60-84), p=0.828]. Perceived and/or experienced stigmatization was similarly high in both groups [DSAS-1: 42.78±15.34 vs 44.82±14.18, p=0.464]. In-depth interviews showed personal experiences of body dissatisfaction and social stigma due to the visibility of technologies, and self-esteem problems.

Conclusions: Not only visible technologies, MDI also might cause psychosocial complications in adults with T1DM. As body-dissatisfaction and social stigma are interindividually varying issues, the pscyhosocial aspects cannot be captured by only-quantitative measures, should also be supported by in-depth interviews as shown here in this study.

MARKED BENEFIT OF AUTOMATED INSULIN DELIVERY (AID) SYSTEMS IN PRESCHOOL CHILDREN WITH TYPE 1 DIABTES

Jantje Weiskorn ¹, Marie Auzanneau^2,3, Marie-Anne Burckhardt⁴, Nicole Blauensteiner⁵, Ulrike Schierloh⁶, Stefanie Lanzinger^2,3

¹Childreńs Hospital AUF DER BULT, Diabetology, Endocrinology, Clinic Research, Hannover, Germany, ²University Ulm, Epidemiology And Biometry, Ulm, Germany, ³German Centre for Diabetes Research (DZD), Diabetes Research, Neuherberg, Germany, ⁴University Childreńs Hospital Basel UKBB, Pediatric Endocrinology And Diabetology, Basel, Switzerland, ⁵Medical University Vienna, Department Of Pediatrics And Adolescent Medicine, Vienna, Austria, ⁶Centre Hospitalier de Luxembourg, Deccp, Clinique Pédiatrique, Luxembourg, Luxembourg

Background and Aims: Objective: To analyse continuous glucose monitoring (CGM) profiles and BMI after one year of AID use in different age groups.

Methods: Using data from the German/Austrian/Luxembourgian/Swiss Diabetes Prospective Follow-Up (DPV) registry of 2019–2023, we included children and adolescents with Type-1 diabetes (T1D) > 6 months, age <18 years, at least one year of AID use, and available CGM profiles (time in range (TIR), time in tight range (TiTR), time below range (TBR)). Adjusted longitudinal regression models were performed to assess the effect of one year under AID in different age groups.

Results: 1,183 individuals (median age: 12.2 years, 53.9% male and 32.6% with migration background) were included. In preschool children (age <6 years, n=47), TIR increased from 67.3% [95%-CI: 63.6-71.2%] to 71.2% [67.7-75.0%], p=0.042, and TiTR from 45.2% [41.6-49.0%] to 49.1% [45.6-53.0%], p=0.037. In school children (age: 6- <12 years, n=515), TIR improved from 66.1% [65.2-67.0%] to 67.2% [66.4-68.1%], p=0.006. In adolescents (age: 12-<18 years, n=621), TIR and TiTR changed not significantly, but HbA1c increased from 7.6% [7.5-7.7%] to 7.7% [7.6-7.8%], p=0.017. In school children and adolescents, BMI-SDS increased from +0.45 [+0.37-+0.53] to +0.51 [+0.43-+0.59], p<0.003, and from +0.62 [+0.54-+0.70] to +0.67 [+0.59-+0.75], p<0.002 respectively, and TBR decreased from 2.8% [2.6-3.0%] to 2.5% [2.3-2.7%], p<0.004 and from 2.3% [2.2-2.5%] to 2.0% [1.9-2.2%], p<0.001, respectively.

Conclusions: Based on real-world data, preschool children particularly benefit from AID in terms of TIR and TiTR. In school children and adolescents, AID has a favorable effect on TBR, but is not beneficial for BMI-SDS.

ASSOCIATION OF SOCIAL DISADVANTAGE WITH HEALTH CARE USE INDICATORS AND DIABETES OUTCOMES AMONG INSULIN PUMP USERS WITH TYPE 1 DIABETES (T1D) IN ONTARIO, CANADA

Youstina Soliman ¹, Karl Everett², Rayzel Shulman¹, Peter Austin², Lorraine Lipscombe¹, Gillian Booth¹, Alanna Weisman¹

¹University of Toronto, Toronto, Canada, ²ICES, Toronto, Canada

Background and Aims: In Ontario, individuals with T1D can access insulin pumps through a publicly-funded program, which requires regular follow-up to maintain eligibility. Although social disadvantage is a predictor of worse diabetes outcomes, we hypothesized this risk would be mitigated among pump users due to program requirements. We evaluated associations between social disadvantage and health care use indicators and diabetes outcomes among insulin pump users.

Methods: We conducted a population-based retrospective cohort study using administrative health data in Ontario, Canada. New insulin pump users on/after April 11, 2012 were followed until March 31, 2021 for the following outcomes: annual rate of A1C tests, ambulatory diabetes appointments, and diabetes-related hospitalizations/emergency department visits, and A1C levels. Associations between social disadvantage (defined by Ontario marginalization index material resources quintile) and outcomes were estimated by Poisson and linear regressions. Generalized estimating equations were used to account for clustering by residential area. All results are reported for most vs. least disadvantaged quintile with 95% confidence intervals.

Results: Among 15,755 pump users (median age 34, mean A1C 8.2%), social disadvantage was not associated with the numbers of A1C tests/year or ambulatory diabetes appointments/year [rate ratios 1.00 (0.97 - 1.03) and 0.97 (0.93 – 1.02)]. However, social disadvantage was associated with minimally higher A1C [0.12% (0.06 - 0.17)] and higher rates of diabetes-related hospitalizations [rate ratios 1.76 (1.41-2.19) for hypoglycemia and 2.07 (1.64 – 2.62) for hyperglycemia].

Conclusions: An insulin pump program with regular follow-up requirements may mitigate disparities in clinical care and glycemic levels, but not diabetes-related hospitalizations.

TRANSDERMAL INSULIN DELIVERY USING MTC-Y CARRIER: A NOVEL ALTERNATIVE TO SUBCUTANEOUS INJECTIONS FOR SUSTAINED RELEASE AND IMPROVED PATIENT COMFORT IN DIABETES TREATMENT

Pawel Biernat

BIOTTS S.A., Bielany Wroclawskie, Poland

Background and Aims: Transdermal drug delivery is a widely accepted method of administration, offering improved patient compliance. The BIOTTS transdermal therapeutic system (TTS) uses patches to release drugs at controlled rates. MTC-Y, a proprietary multifunctional carrier, enhances bioavailability and facilitates macromolecular drug transport through the skin. This study focused on developing a transdermal formulation for NPH insulin, used by over 75 million diabetes patients globally, aiming to increase patient comfort and reduce side effects.

Methods: To assess skin permeability, tests on porcine skin (structurally similar to human skin) revealed that the MTC carrier transported insulin effectively, reaching concentrations of 80 ng/mL after 72 hours. Subsequent in vivo studies on male Sprague–Dawley rats further validated these findings. Plasma insulin concentrations were measured over 108 hours.

Results: Plasma insulin concentrations were measured over 108 hours, demonstrating sustained release without skin irritation (Fig.1). The transdermal system showed 55.1% bioavailability compared to subcutaneous (SC) insulin (Tab.1), with a more consistent and prolonged release profile. In subsequent studies using Göttingen mini pigs, transdermal insulin exhibited around 70% efficacy compared to SC administration, particularly at the six-hour peak (Fig.2).

Fig.1 Insulin concentration in rat blood plasma

Fig.2 Blood glucose concentration (Göttingen mini pigs)

Conclusions: These results highlight the potential of transdermal insulin as a viable alternative to daily injections, offering enhanced patient comfort and compliance. The sustained insulin levels observed in transdermal delivery fall within the therapeutic window, supporting the formulation’s promise for future clinical applications.

DETECTING AND DESCRIBING HYPOGLYCAEMIC EPISODES USING CONNECTED INSULIN PENS AND CONTINUOUS GLUCOSE MONITORING (CGM) DATA

Klemen Dovč ¹, Andreas Trier Poulsen², Anne Kaas², Charlotte Schmitt², Sudha Mj³, Tadej Battelino¹

¹University Children's Hospital, University Medical Centre Ljubljana, and Faculty of Medicine, University of Ljubljana, Department Of Endocrinology, Diabetes And Metabolism, Ljubljana, Slovenia, ²Novo Nordisk A/S, Søborg, Denmark, ³Novo Nordisk Service Centre India Private Ltd, Bangalore, India

Background and Aims: Hypoglycaemia remains a concern for insulin-treated people living with diabetes (PLWD). This real-world study assessed the frequency of continuous glucose monitoring (CGM)-based hypoglycaemia (level 1: <3.9 mmol/L for ≥15 minutes; level 2: <3.0 mmol/L for ≥15 minutes) over 24 hours by age and connected pen insulin regimen.

Methods: Data were collected from PLWD (aged 18–100 years) using CGM and a connected pen to deliver their bolus insulin (aspart or fast-acting aspart). Data included days with ≥15% CGM coverage. Mean frequency of hypoglycaemia onset was calculated per individual per hour.

Results: Data from 65,634 individuals (mean age: 43.09 years), including 8,206,388 CGM days with bolus injections, were analysed. Observed frequency of combined (level 1 + 2) hypoglycaemia varied over 24 hours but was highest between 13:00–16:00; level 1 episodes peaked during the afternoon, whilst the frequency of level 2 episodes increased during the evening/overnight (Figure). Adults aged 18–39 years experienced higher frequencies of hypoglycaemia versus other subgroups. Lower frequencies of hypoglycaemia were observed in individuals who used a connected pen for basal and bolus insulin compared with those only using a connected pen to deliver bolus insulin.

Conclusions: Observed frequency of hypoglycaemic episodes in PLWD varies by time, age and usage of connected pens for bolus and basal insulin.

ADOSHELL®, A NON-FIBROTIC IMMUNOPROTECTIVE HYDROGEL SHOWS EFFICIENT AND SUSTAINED IN VIVO FUNCTIONALITY FOR CLINICAL USE

Xavier Gaume ¹, Romain Besnard¹, Julie Brun¹, Camille Gautier¹, Clément Cocita¹, Jonna Saarimäki-Vire², Timo Otonkoski², Diego Balboa³, Anne-Lise Gaffuri¹, Ouardane Jouannot¹, Nicolas Laurent¹, Rosy Eloy¹, Martin Gaudier⁴, Karim Bouzakri⁵, François Pattou⁶

¹ADOCIA, Lyon, France, ²Helsinki University, Helsinki, Finland, ³Helsinki University, Helsinki, France, ⁴ADOCIA, LYON, France, ⁵Centre Européen d’Etude du Diabète, Strasbourg, France, ⁶CHU Lille - INSERM, Lille, France

Background and Aims: AdoShell® is an easily implantable and retrievable scaffold for islet transplantation to cure diabetes without immunosuppression. It is based on a permselective hydrogel in a supporting frame.

Methods: Primary islets or stem-cell derived islets (SCDI) were encapsulated in AdoShell®. Functionality was evaluated in vitro by GSIS and in vivo by measuring glycemia and C-peptide secretion after implantation intraperitoneally in rodents. For in vitro evaluation of immunoprotection, naked and encapsulated islets were incubated for 4 days with activated PBMC. Human-scale AdoShell® was implanted by laparoscopy in minipig.

Results: Allogeneic transplant of encapsulated rat islets in immunocompetent diabetic rats showed good tolerance. Islet survival, insulin secretion and glycemic regulation were demonstrated for 74 days (n=5 rats). Using human islets, AdoShell® prevented activated human PBMC mediated cell death, confirming immunoprotection against allogenic rejection. In vivo, human C-peptide secretion of AdoShell® Islets implanted in immunodeficient mice reached a plateau of 633.0±168.1 pM (n=10 mice) and was maintained for 3 months (Figure). In vitro secretion indexes of explants were similar to pre-implantation levels. A human-scale AdoShell® able to deliver a therapeutic dose was successfully manufactured and implanted by laparoscopy in minipig. This proof of concept opens the pathway towards clinical application. The compatibility of AdoShell® with SCDI was tested in immunodeficient mice demonstrating cluster maturation and functionality for 2 months.

Conclusions: These results show the tolerance, immunoprotection, and compatibility of AdoShell® with human islets. Scalability and mini-invasive surgery pave the way for translation in humans, while compatibility with SCDI promises to make islet transplantation widely available.

POST-PRANDIAL GLUCOSE EXCURSION WITH INHALED INSULIN IN YOUTH COMPARED WITH ADULTS WITH TYPE 1 DIABETES

Michael Haller ¹, Lauren Kanapka²

¹University of Florida, Gainesville, United States of America, ²Jaeb Center for Health Research, Tampa, United States of America

Background and Aims: Inhaled Technosphere insulin (TI) has been shown to reduce post-prandial glucose excursions in adults with type 1 diabetes (T1D). The aim of this study was to compare glucose excursions in youth using TI to bolus for a standardized meal with glucose excursions in adults using either TI or rapid acting analogue (RAA) insulin to bolus.

Methods: Individuals included in the analyses were multiple daily injection insulin users with T1D who were participants in the INHALE-1 (pediatrics) or INHALE-3 (adults) randomized trials comparing TI versus RAA as bolus insulin. For the first dose of TI, an in-clinic meal challenge was performed with a standardized meal in 113 youth 6-<18 years old and in 51 adults 22-77 years old. Additionally, 26 adults had a meal challenge using RAA insulin. The TI dose was determined by rounding the usual RAA dose to the nearest whole number, multiplying by 2, and then rounding down to the nearest 4-unit dose of TI.

Results: Mean (±SD) glucose excursion was less with TI (69±56 mg/dL in youth and 73±50 mg/dL in adults) compared with RAA in adults (101±45 mg/dL); and time to peak glucose was shorter with TI (69±34 in youth and 71±36 minutes in adults) than with RAA (78±31 minutes in adults) as seen in the Figure.

Conclusions: The post-prandial glucose excursion with TI appears similar in youth to that which has been reported in adults and is substantially lower than the excursion following a bolus with RAA insulin.

COMPARISON OF A REGIMEN OF INHALED TECHNOSPHERE INSULIN PLUS INSULIN DEGLUDEC VERSUS USUAL CARE IN ADULTS WITH TYPE 1 DIABETES (T1D)

Anastasios Manessis ¹, Roy Beck², Ryan Bailey³, Peter Calhoun³, Katrina Ruedy³, Irl Hirsch⁴

¹NYC Research, Inc, New York, United States of America, ²JAEB Center for Health Research, Research, Tampa, United States of America, ³Jaeb Center for Health Research, Tampa, United States of America, ⁴University of Washington, seattle, United States of America

Background and Aims: Despite advances in insulin delivery methods, most people with T1D experience post-meal hyperglycemia and do not achieve glycemic goals. Inhaled Technosphere Insulin (TI), with a more rapid onset of action than subcutaneously-injected rapid-acting analogue (RAA) insulin, has the potential to improve glycemic control. The aim of this study was to assess change in glycemic outcomes when switching from an insulin regimen using RAA for bolusing to TI.

Methods: In 42 adults with T1D, glycemic outcomes using a regimen of TI and degludec for 13 weeks were compared to a prior period of usual care consisting of automated insulin delivery (AID, N=15), multiple daily injections (MDI, N=23), or a non-AID pump (N=4). The primary outcome was non-inferiority for daytime time-in-range 70-180 mg/dL (TIR).

Results: Mean daytime TIR increased from 50% at baseline to 55% after 13 weeks using TI plus degludec (mean change 5.1%, 95% CI 0.3% to 9.8%, P for non-inferiority <0.001). Mean HbA1c decreased from 7.6% to 7.4% (mean change -0.23%, 95% CI -0.42 to -0.04, P=0.02 for superiority). HbA1c was <7.0% in 14% at baseline and 31% after 13 weeks (P=0.02 for superiority); 29% improved >0.5% and 12% worsened >0.5%. Mean glucose by hour is shown in the Figure. Treatment group difference in time <54 mg/dL was 0.2% (95% CI -0.1% to 0.5%).

Conclusions: Switching from insulin delivery with predominately AID or MDI to a regimen of TI and degludec was non-inferior for TIR and showed an improvement in HbA1c.

EFSITORA VS. DEGLUDEC IN PATIENTS WITH TYPE 1 DIABETES (QWINT-5)

Richard M. Bergenstal ¹, Ruth S. Weinstock², Chantal Mathieu³, Yukiko Onishi⁴, Vishali Vijayanagaram⁵, Michelle L. Katz⁶, Molly C. Carr⁷, Annette M. Chang⁶

¹International Diabetes Center, HealthPartners Institute, St Louis Park, United States of America, ²State University of New York Upstate Medical University, Syracuse, United States of America, ³University Hospitals Leuven – KU Leuven, Leuven, Belgium, ⁴The Institute of Medical Medicine, Asahi Life Foundation, Tokyo, Japan, ⁵Eli Lilly and Company, Gurgaon, India, ⁶Eli Lilly and Company, Insulins And Devices, Indianapolis, United States of America, ⁷Eli Lilly and Company, Indianapolis, United States of America

Background and Aims: The study aimed to evaluate the efficacy and safety of a novel once-weekly basal insulin efsitora alfa (efsitora) compared to daily insulin degludec in patients with type 1 diabetes (T1D). The primary objective was to demonstrate non-inferiority in glycaemic control, with key secondary objectives assessing safety and other efficacy endpoints.

Methods: This was a phase 3, randomised, open-label, treat-to-target trial involving adults with T1D and HbA1c 7.0% to 10.0%. Participants were randomly assigned to either once-weekly efsitora or daily insulin degludec, both with insulin lispro, over a 52-week treatment period.

Results: Efsitora demonstrated noninferior glycaemic control to degludec, with HbA1c change from baseline to week 26 of 7.88% to 7.41% with efsitora and 7.94% to 7.36% with degludec (ETD: 0.052% [95% CI, -0.077 to 0.181]; p=0.43. The rate of level 2 (<54 mg/dL [3·0 mmol/L]) or level 3 hypoglycaemia was higher with efsitora versus degludec (14.0 vs 11.6 events per patient-year of exposure; estimated rate ratio 1.21 [95% CI, 1.04 to 1.41], p=0.016) from week 0-52, with the highest rates occurring from week 0-12. Nocturnal level 2 or 3 hypoglycaemia rates were similar between groups. The incidence of severe hypoglycaemia was higher with efsitora (35/343 [10%]) versus degludec (11/349 [3%]).

Conclusions: Once-weekly efsitora showed similar HbA1c reduction compared to degludec. The imbalance in hypoglycaemia for efsitora versus degludec suggests the need for further evaluation of efsitora dose initiation and optimization of basal-bolus insulin in T1D.

EVOLUTION OF A NOVEL CLINICAL DECISION SUPPORT (CDS) ALGORITHM WITH INSULIN EFSITORA ALFA IN ADULTS WITH TYPE 2 DIABETES (T2D)

Michelle L. Katz ¹, Christof Kazda², Jie Xue³, Juan Frias⁴, Jit Mitra⁵, Stephanie Reeve⁶, Stephanie Gleissner⁶, Eyal Dassau⁶

¹Eli Lilly and Company, Insulins And Devices, Indianapolis, United States of America, ²Eli Lilly and Company, Clinical Research - Incretins & Incretin Combinations, Fegersheim, France, ³Eli Lilly and Company, Advanced Analytics, Indianapolis, United States of America, ⁴Biomea Fusion, Downtown Redwood City, United States of America, ⁵Eli Lilly and Company, Gurgaon, India, ⁶Eli Lilly and Company, Indianapolis, United States of America

Background and Aims: Efsitora is a weekly basal insulin. We previously evaluated a basal titration algorithm that incorporated FBG trends, prior insulin doses, population pharmacokinetic modeling, and hypoglycaemia for efsitora dosing and demonstrated safety and feasibility (cohort 1).

Methods: Algorithm refinements in the starting dose, safety, and titration modules were simulated, incorporated and tested in cohort 2 of a multicenter, open-label early feasibility study in adults (≥18 and ≤75 years) with T2D with or without current basal insulin use. During the 16-week treatment period, investigators requested a dose recommendation from the CDS algorithm and either overrode or administered the recommended dose. BGs, blinded CGM, hypoglycaemia, and dose overrides were captured.

Results: Participants (N=32) had a mean (SD) age of 60.1 (8.7) years, duration of diabetes of 17.1 (6.8) and 16.5 (10.1) years and HbA1c of 8.5 (1.1) % and 8.3 (1.1) % for basal switch and insulin naïve participants. CDS dosing recommendation overrides were recorded for 3 of 491 injections (0.6%). A significant reduction from baseline to Week 16 in HbA1c was observed for basal switch (mean [SD]= -0.6 [0.7]; p=0.003) and insulin naïve participants (-0.9 [1.2]; p=0.009). This was achieved while improving CGM-detected level 2 hypoglycaemia events from the earlier algorithm (switch 18.6 to 10.5 and naïve 22.4 to 8.8 events/pt/yr.). Switch and naïve participants improved by 16.2% (p=0.003) and 25.7% (p=<0.001) to 47.3% and 61.7% in time spent between 3.9-10.0 mmol/L (70-180 mg/dL), respectively, from baseline to endpoint.

Conclusions: The novel CDS algorithm-maintained investigator trust and glycaemic efficacy with improved hypoglycaemia after algorithm refinement.

CHANGE IN CARDIOVASCULAR BIOMARKERS IN OVERWEIGHT AND OBESE ADULTS WITH TYPE 1 DIABETES (T1D) ON TIRZEPATIDE

Satish Garg ¹, Gurleen Kaur², Emma Mason², Kelly Ciesco², Drew Renner², Janet Snell-Bergeon²

¹Barbara Davis Center for Diabetes, University of Colorado, Medicine And Pediatrics, Aurora, United States of America, ²Barbara Davis Center for Diabetes Adult Clinic, Aurora, United States of America

Background and Aims: Two-thirds of adults with T1D are overweight/obese, contributing to sub-optimal glucose control, cardiovascular, and renal disease. Tirzepatide is a dual incretin approved for use in obesity with or without T2D. We examined cardiovascular and renal biomarkers with longer-term over 21 months off-label use of tirzepatide among overweight/obese adults with T1D.

Methods: 84 people with T1D who were prescribed tirzepatide (cases) and 38 controls not on any anti-obesity medication were analyzed. Controls were frequency matched for age, diabetes duration, sex, HbA1c, and BMI. Data was collected from the Electronic Medical Records before initiating tirzepatide, and over 21 months of treatment. Mixed-effects models were used to examine changes in lipids, blood pressure, and eGFR over time in cases compared to controls.

Results: Baseline characteristics were similar, except tirzepatide users had a higher baseline BMI than controls (Table). HbA1c decreased more in patients using tirzepatide vs controls (-0.50±0.07% and -0.24±0.09% respectively, p=0.017) and cases lost a least square mean of 37±2.1lbs (-15.5% body weight) compared to a 3.3±2.9lbs (+2% body weight) gain in controls. Tirzepatide users had significant improvement in total and LDL cholesterol, triglycerides, systolic blood pressure, and eGFR (Table); which remained significant even after adjusting for weight and HbA1c. The eGFR declined significantly in controls but not in tirzepatide users.

Conclusions: In conclusion, use of tirzepatide in patients with T1D was associated with improvement in cardiovascular biomarkers and preservation of kidney function independent of weight changes and/or HbA1c levels. We recommend a RCT with tirzepatide in overweight/obese patients with T1D to assess cardiovascular markers.

LONG-TERM EFFECTIVENESS OF TIRZEPATIDE IN OVERWEIGHT AND OBESE ADULTS WITH TYPE 1 DIABETES (T1D)

Janet Snell-Bergeon ¹, Amanda Rewers¹, Christie Beatson¹, Monica Lanning¹, Satish Garg²

¹Barbara Davis Center for Diabetes Adult Clinic, Aurora, United States of America, ²Barbara Davis Center for Diabetes, University of Colorado, Medicine And Pediatrics, Aurora, United States of America

Background and Aims: Tirzepatide is a dual incretin approved for use in patients with T2D and overweight/obesity (OW/OB). Short-term observational data suggests that it is effective in reducing weight and HbA1c in OW/OB adults with T1D. We analyzed long-term off-label use of tirzepatide among OW/OB adults with T1D.

Methods: 84 people with T1D who were prescribed tirzepatide (cases) and 38 controls not on any anti-obesity medication were analyzed. Controls were frequency matched for age, diabetes duration, sex, HbA1c, and BMI. Data was collected before initiating tirzepatide, and over 21 months of treatment. Mixed-effects models were used to examine change in weight, BMI, and HbA1c over time in cases compared to controls.

Results: Baseline characteristics were similar in cases vs. controls (mean age 41±11 and 41±7 years, p=0.98, diabetes duration 27±13 and 28±11 years, p=0.54, HbA1c 7.0±0.9 and 7.1±1.0%, p=0.42, 63% and 68% female, p=0.57), respectively. Cases had a slightly higher baseline BMI than controls (35.2±4.8 vs. 33.3±4.2 kg/m2, p=0.03). Patients using tirzepatide lost significantly more weight than controls (Figure 1A), and weight loss continued to increase over 21 months (-23.4% vs. +2% respectively; p<0.0001). HbA1c improved in cases (-0.50±0.07%) more than in controls (-0.24±0.09%, p=0.017, Figure 1B). Time-in-range increased significantly in cases (9.8±1.3%) and in controls (6.3±2.0%) but did not differ by group (p=0.14). No hospitalization from severe hypoglycemia or diabetic ketoacidosis were reported.

Conclusions: In conclusion, longer-term use of tirzepatide was associated with continued weight loss of >23% over 21 months and rapid and sustained improvement in HbA1c in OW/OB patients with T1D on tirzepatide.

CGM MARKERS ACCURATELY PREDICT PROGRESSION TO STAGE 3 TYPE 1 DIABETES

Andrea Steck, Fran Dong, Cristy Geno Rasmussen, Kathleen Waugh, Kimber Simmons, Brigitte Frohnert, Marian Rewers

Barbara Davis Center for Diabetes, University Of Colorado Anschutz Medical Campus, Aurora, United States of America

Background and Aims: With the increase in general population screening for risk of type 1 diabetes, there is an urgent need to define biomarkers for disease progression that are both accurate and practical for the clinical setting. Our aim was to evaluate the accuracy of continuous glucose monitoring (CGM) markers for progression to Stage 3 type 1 diabetes (T1D) in our combined ASK (Autoimmunity Screening for Kids) and DAISY (Diabetes Autoimmunity Study in the Young) populations.

Methods: A total of 165 persistently autoantibody positive (Ab+) subjects (64% multiple Ab+) were followed for a median of 23 (9-42) months after their baseline CGM. Of these, 42 (25%) progressed to clinical diabetes at a median of 14 (range:11-18) years. The full Cox PH model included single vs multiple Ab+, sex, ZnT8A positivity, IA2A positivity, HbA1c>5.5% and TA140>10%, with 3 factors staying significant after stepwise regression.

Results: Multiple CGM metrics had receiver operating characteristic area under the curve (ROC AUC) for Stage 3 T1D prediction of 0.79 or higher compared to 0.73 for HbA1c. In survival analyses, the risk of progression to diabetes was 47% in 2 years and 64% in 3 years in those with TA140>10%, compared to only 8% and 17% respectively in the other participants (p<0.0001). The predictive ability was further improved in a combined model including TA140>10%, HbA1c>5.5% and ZnT8A positivity with a 65% 2-year risk.

Conclusions: CGM TA140>10% is highly predictive of progression to clinical diabetes within the next 2-3 years. Prediction can be further improved in a model incorporating HbA1c and antibody data.

INTEGRATION OF CONTINUOUS GLUCOSE MONITORING AND INSULIN DOSING DATA INTO ELECTRONIC HEALTH RECORDS AT DIABETES CENTER (REAL WORLD DATA)

Robert Bem, David Vávra, Tomas Neskudla, Martin Haluzík

IKEM, Diabetes Center, Prague, Czech Republic

Background and Aims: Continuous glucose monitoring (CGM) and insulin data integration into the electronic health record (EHR) plays a crucial role in caring for people with type 1 diabetes (T1D). The study assessed the feasibility of integrating semi-automated CGM and insulin data into the EHR and their association with diabetes control for adults with T1D at our diabetes center.

Methods: Between 2012 and 2023, data from adults with T1D were used for analysis (from 1765 in 2012 to 2609 subjects in 2023). We assessed the integration of CGM and insulin data into the EHR, including data from sensors, smart pens, and insulin pumps. Data were integrated semi-automatically by processing PDF records sent by patients via email or downloaded during visits, and we evaluated their association with diabetes control.

Results: Since 2012, there has been a significant increase in the percentage of individuals with integrated CGM and insulin data in EHR (2012: 5.3%; 2016: 41.2%; 2020: 59.6%; 2023: 78%; all p<0.05). The number of integrated data points per person increased significantly during the period. A decrease in HbA1c was recorded (2012: 69; 2016: 65.5; 2020: 64.6; 2023: 56.7 mmol/mol; all p<0.05), with significantly lower HbA1c levels in individuals with integrated data compared to those without (56.3 vs. 59.9 mmol/mol in 2023). Fig: Effect of data integration into EHR on diabetes control in 2012-2023.

Conclusions: In our real-world study, integrating CGM and insulin dosing data into the EHR was associated with improved diabetes control in adults with T1D. EXCELES, No. LX22NPO5104.

GLUCOSE VARIABILITY ASSESSED BY CGM CORRELATES WITH INSULIN SECRETION IN ISLET TRANSPLANTATION RECIPIENTS

Willemijn De Vos, Mandala Ajie, Dirk Jan Cornelissen, Eelco De Koning

Leiden University Medical Center, Internal Medicine, Leiden, Netherlands

Background and Aims: Allogeneic and autologous islet transplantation (ITx) is performed to stabilise glycaemic control by reestablishing endogenous insulin secretion. Continuous glucose monitoring (CGM) can improve glycaemic management post-ITx. We hypothesised that CGM metrics of glucose variability are an indicator of islet graft function during follow-up.

Methods: We performed twenty-three allogeneic and autologous islet transplantations. We determined peak C-peptide levels during a mixed meal tolerance test (MMTT) and measured HbA1c three months post-ITx. CGM metrics included time in range (3.9-10.0 mmol/L), time in clinically relevant hypoglycaemia (<3.0 mmol/L) and glucose variability (%CV and SD). Using Spearman-correlation analysis we determined the relation between measures of glucose variability and graft function.

Results: Patients (14M/9F), with BMI 24.04±2.79 were 47.6±11.9 years old. HbA1c decreased from 53.8±15.5 mmol/mol at baseline to 46.1±10.3 mmol/mol, and TIR was 81.5±22.8% at three months. Time in clinically relevant hypoglycaemia decreased 9.4 fold from 0.75 [0.18,1.96] to 0.08 [0.00,0.50]%. There was low glucose variability (%CV 25.7±8.3% and SD 2.0±1.0) and during the MMTT the peak C-peptide was 1.4±0.88 nmol/L at three months. Glucose variability (%CV and SD) was negatively correlated with peak C-peptide (R=-0.71, P=0.00032 and R=-0.7, P=0.00039, respectively) while TIR was positively correlated with peak C-peptide (R=0.63, P=0.0021).

Conclusions: Glucose variability may be used as an indicator of insulin secretory capacity in islet transplant recipients.

IN-HOSPITAL DIABETES MANAGEMENT BY A DIABETES TEAM AND INSULIN ALGORITHMS BASED ON CONTINUOUS GLUCOSE MONITORING OR POINT-OF-CARE GLUCOSE TESTING IN PATIENTS WITH TYPE 2 DIABETES

Mikkel Thor Olsen ¹, Carina Klarskov¹, Signe Jensen¹, Louise Rasmussen¹, Birgitte Lindegaard^2,3, Jonas Andersen^3,4, Hans Gottlieb⁵, Suzanne Lunding⁶, Ulrik Pedersen-Bjergaard^1,3, Katrine Hansen⁷, Peter Kristensen^1,3

¹North Zealand Hospital, Endocrinology And Nephrology, Hilleroed, Denmark, ²North Zealand Hospital, Pulmonary And Infectious Diseases, Hilleroed, Denmark, ³Faculty of Health and Medical Sciences, University of Copenhagen, Clinical Medicine, Copenhagen, Denmark, ⁴North Zealand Hospital, Orthopedic Surgery, Hilleroed, Denmark, ⁵Herlev-Gentofte Hospital, Orthopedic Surgery, Herlev, Denmark, ⁶Herlev-Gentofte Hospital, Infectious Diseases, Herlev, Denmark, ⁷Steno Diabetes Center Copenhagen, Herlev, Denmark

Background and Aims: Around 20% of hospitalised patients have diabetes, highlighting the need for good management strategies. The DIATEC trial investigates the glycaemic and clinical effects of inpatient continuous glucose monitoring (CGM)-guided insulin titration and diabetes teams.

Methods: This two-center trial randomised 166 non-intensive care unit patients with type 2 diabetes. Diabetes management was performed by regular staff, guided by diabetes teams using insulin titration algorithms based on either point-of-care glucose testing or CGM. The primary outcome was the difference in time in range (TIR) (3.9–10.0 mmol/l) between the two arms.

Results: The CGM-arm achieved a higher TIR (median, IQR) of 77·6% (24·4) vs 62·7% (31·5) in the POC-arm (p <·001). Time above range (TAR) >10·0 mmol/l was lower in the CGM-arm of (median, IQR) 21·1% (24·8) vs 36·5% (30·3) in the POC-arm (p=·.001). Time below range (TBR) <3·9 mmol/l was reduced by CGM, with a relative difference to POC of 0·57 (95% CI 0·34–0·97) (p =·042). Prolonged hypoglycaemic events decreased (IRR 0·13, 95% CI 0·04–0·46) (p =·001), and the coefficient of variation (mean, SD) was lower in the CGM-arm of 25·4% (6·3) vs 28·0% (8·2) in the POC-arm (p =·024). Total insulin doses (mean, SD) were reduced in the CGM-arm with 24·1 IU/day (13·9) vs 29·3 IU/day (13·9) in the POC-arm (p =·049). A composite of complications was lower in the CGM-arm (IRR 0·76, 95% CI 0·59–0·98) (p=·032).

Conclusions: In-hospital CGM increased TIR by 15%. CGM also reduced TAR, TBR, glycaemic variability, insulin use, and a composite of in-hospital complications.

ENDOSCOPIC DUODENAL-JEJUNAL BYPASS LINER (DJBL) TREATMENT FOR TYPE 2 DIABETES AND OBESITY: COMPARING 6 VS 12 MONTH IMPLANTATION IN A WORLDWIDE REGISTRY

Robert Ryder ¹, Mahender Yadagiri¹, Lynne Munro², Harry Frydenberg², Sigal Fishman³, Charlotte De Jonge⁴, Jan Willem Greve⁵, James Byrne⁶, Jessica Mcmaster⁷, Jacob Chisholm⁸, Lilian Kow⁸, John Mason⁹, Julian Teare¹⁰, Ricardo Victor Cohen¹¹, Justin Bessell¹², Harald Sourij¹³, Piya Sen Gupta¹,¹⁴

¹City Hospital, Diabetes, Birmingham, United Kingdom, ²Epworth Hospital, Bariatric Surgery, Richmond, Australia, ³Tel Aviv Sourasky Medical Center, Gastroenterology, Tel Aviv, Israel, ⁴Catharina Ziekenhuis, Eindhoven, Netherlands, ⁵Zuyderland Medical Center, Bariatric Surgery, Heerlen, Netherlands, ⁶University Hospital Southampton, Surgery, Southampton, United Kingdom, ⁷Princess Alexandra Hospital, Gastroenterology, Woolloongabba, Australia, ⁸Flinders Medical Centre, Adelaide Bariatric Centre, Bedford Park, Australia, ⁹Trafford General Hospital, Gastroenterology, Manchester, United Kingdom, ¹⁰St Mary's Hospital, Gastroenterology, London, United Kingdom, ¹¹Oswaldo Cruz German Hospital, Center For Obesity And Diabetes, Sao Paulo, Brazil, ¹²Calvary North Adelaide Hospital, Surgery, Adelaide, Australia, ¹³Medical University of Graz, Internal Medicine, Graz, Austria, ¹⁴Guy's and St Thomas' NHS Foundation Trust, Diabetes, London, United Kingdom

Background and Aims: DJBL (AKA EndoBarrier® or RESET®) is a 60cm duodenal-jejunal bypass liner which is endoscopically implanted for up to one-year. It mimics the by-pass part of roux-en-y bariatric surgery. Uncertainty exists re risk:benefit of DJBL which has been used as treatment for obesity, both with and without diabetes.

Methods: In view of this, during 2017, an independent, secure, online registry was established under the auspices of the Association of British Clinical Diabetologists, for the collection of safety and efficacy data worldwide.

Results: As of September 2024, data had been entered on 1101 patients, of whom 261 (age 52.3±10.1 years, 53% male, 82% white ethnicity, BMI 39.9±7.2kg/m2) had both 6- and 12-month data entered. DJBL had considerable impact on weight and HbA1c (Table). Though there was significant further reduction in HbA1c and weight between 6- and 12-months, it is clear (Table), that most of the improvement has already occurred by 6-months. The higher the HbA1c the greater the fall (Table). In the full registry, 46/1101(4.2%) experienced serious adverse events (SAE). 20/46 (43.5%) SAE would have been avoided by removal at 6-months (11 liver abscess, 6 GI bleed, one cholecystitis, one pancreatitis, one abdominal abscess). It was particularly noteworthy that 11/13(85%) liver abscess SAE would have been avoided by removal at 6-months.

Conclusions: This international data from the DJBL registry suggests that most of the benefits of DJBL occur by 6-months and there may therefore be situations where removal at 6-months might be considered to reduce the risk of SAE, especially the liver abscess SAE.

LONG-TERM EFFECTIVENESS, SAFETY AND PATIENT EXPERIENCES WITH ORAL SEMAGLUTIDE IN T2DM: THE LARGEST, REAL-WORLD, INDIAN EXPERIENCE THUS FAR

Purvi Chawla ¹, Manoj Chawla², Banshi Saboo³, Rajesh Deshmane⁴, Rutul Gokalani⁵, Nanditha Arun⁶, Gaurav Beswal⁷, Awais Pathan⁸, Kiran Shah⁹

¹Bharatiya Arogya Nidhi Hospital, Mumbai, India, ²Lina Diabetes Care Mumbai Diabetes Research Centre, Mumbai, India, ³Diacare, Diabetes, Ahmadabad, India, ⁴Shree Mahalaxmi Diabetic Care Centre, Kolhapur, India, ⁵AHC Diabetes Clinic, Ahmedabad, India, ⁶Dr A Ramchandran Diabetes Hospital, Chennai, India, ⁷Good Hope Diabetes & Thyroid Clinic, Mumbai, India, ⁸Dr. Pathan's Diabetes Specialties Clinic, Mumbai, India, ⁹Grant Govt. Medical College & Sir J. J. Group of Hospitals, Mumbai, India

Background and Aims: Oral semaglutide, the first, oral GLP1RA approved for T2DM in India, was launched in January 2022. Growing prescriptions of GLP1RAs are based on the glycemic, weight and pleitropic cardiovascular and renal benefits. This retrospective, real-world study assessed the long-term effectiveness, safety and patient experiences in Indian patientswith T2DM, making it the largest study in India, thus far.

Methods: It was conducted as a multi-centric, single-arm study including 373 patients from 16 centres across India after a central Ethics Committee approval. Adult patients with T2DM, on oral semaglutide therapy for nine to twelve months with/without other anti-diabetic agents with data available at the respective clinics, were included. The primary endpoints included change in HbA1c (%), body weight (kg), and waist circumference (cms) from baseline to end of study. Secondary endpoints werechanges in glycemic, anthropometric and laboratory parameters and composite outcomes like proportion of participants achieving Hba1c and body weight decrease of ≥1% and ≥5%, respectively.Exploratory endpoints assessed change in food preference and other such parameters.

Results: Significant reductions (p<0.0001) in the primary endpoints (HbA1c, body weight, waist circumference) with mean absolute differences being 1.13%, 6 kgs and 6.2 cms respectively were seen. Secondary outcomes showed significant improvements with no major safety concerns and a positive trend in the exploratory outcomes.

Conclusions: This studyunderscored the effectiveness and safety of oral semaglutide in people with T2DM as seen over nine to twelve months of use. These results are aligned with findings in the global, real-world studies conducted.

IMPACT OF DIABETES DEVICES ON SKIN IN YOUTH AND YOUNG ADULTS WITH TYPE 1 DIABETES

Valeria Fadhel Hernandez, Kasserine Taylor, Angela J. Karami, Emily Fivekiller, Cari Berget, Samantha Lange, Erin Cobry

Barbara Davis Center for Diabetes, University of Colorado, Aurora, United States of America

Background and Aims: Diabetes technologies are continuously being developed, leading to increased device use and attainment of glycemic targets. Continuous glucose monitors (CGMs), insulin pumps, and automated insulin delivery systems are now the standard of care. However, a significant barrier to device use is the effect of devices and adhesives on the skin, which has not been well evaluated in youth and young adults (YYA) with type 1 diabetes (T1D).

Methods: This study consisted of a one-time in-house skin survey completed by YYA 2-25 years old with a diagnosis of T1D utilizing a device (CGM and/or pump). The survey assessed device-related skin reactions, locations and types of reactions, and the impact of skin-related issues on daily life.

Results: 77 participants enrolled (Mean Age: 14.2±6.1 yrs, Mean T1D duration 5.9±4.9, Mean HbA1c 7.0±1.0, 46 Female, 88% NHW, 100% CGM users, 90.1% pump users). Devices were worn on upper arm (73%), abdomen (56%), upper buttocks (47%), and thighs (43%). 53% reported CGM reactions, most occurring on the upper arm. 49% reported pump infusion site reactions, with most occurring on the abdomen. Redness and itching were most common. Of those who endorsed difficulty maintaining device adhesion (N=60), over-tapes were commonly used (93%).

Conclusions: Skin reaction in YYA with T1D are frequently associated with diabetes devices. It is important to address skin care during routine T1D visits to identify barriers to device use, including skin reactions and/or difficulty with device adhesion, in order to improve success with device wear. Additional research is needed to address solutions to these barriers.

EVALUATING USER EXPERIENCE AND PREFERENCES OF DECISION SUPPORT VS. PERSONALIZED FEEDBACK SYSTEMS IN DIABETES MANAGEMENT

Annanda Fernandes Moura B Batista ¹, Linda Gonder-Frederick², Ralf Nass², Chiara Fabris², Chaitanya Koravi², Giulio Santini², Charlotte Loving Barnett², Emma Emory², Laura Kollar², Jacob Hellmann², Marc Breton²

¹University of Virginia - Center for Diabetes Technology, School Of Medicine, Charlottesville, United States of America, ²University of Virginia, Center For Diabetes Technology, Charlottesville, United States of America

Background and Aims: Psychological factors such as perceived trust, usefulness, and ease of use impact technology acceptance. Understanding these factors can help optimize Decision Support Systems (DSS) for diabetes management. This study examines perceptions of usefulness, ease of use, and trust in two diabetes management technologies: DSS and Personalized Feedback (PF).

Methods: 53 participants used both support system interventions for 8 weeks in sequential fashion. The PF program provided personalized information on estimated HbA1c, hypoglycemia risk, insulin sensitivity, and insulin on board. The DSS offered insulin and carbohydrate recommendations for exercise and bedtime, a smart bolus calculator, and treatment optimization based on CGM and insulin data. Data were analyzed using the Fisher-Freeman-Halton Exact Test and repeated measures MANOVA. Participants rated ease of use, usefulness, trust, and their preference for each system after use.

Results: There were no significant differences in ease of use or usefulness (p > 0.05). PF scored slightly higher for trustworthiness than DSS (p < 0.05), though with minimal effect size. Preferences indicated DSS was the least preferred 15(28.3%) and PF the most preferred 22(41.5%), with 16(30.2%) showing no preference. Repeated measures MANOVA showed no significant differences in perceived technology burden between programs (p = 0.349).

Conclusions: No significant differences emerged between DSS and PF in ease of use, usefulness, or acceptance. However, PF was marginally more trusted. Findings indicate that preferences for support systems may be shaped more by individual characteristics than by the systems themselves.

MODIFICATION OF THE EXERCISE’S ACUTE EFFECT BY BODY MASS INDEX IN ADULTS WITH T1D IN THE T1DEXI STUDY

Maria Carolina Fragozo-Ramos ¹, Vihangkumar Naik¹, Jaime Gallo-Villegas², Jose Garcia-Tirado¹

¹University of Bern, University Clinic For Diabetes, Endocrinology, Nutritional Medicine, And Metabolism (udem), Bern, Switzerland, ²University of Antioquia, Sports Medicine Postgraduate Program, And Grinmade Research Group, Sicor Center, Faculty Of Medicine, Medellín, Colombia

Background and Aims: Despite the well-known effects of exercise in individuals with T1D, the effect modification by BMI is still not fully understood.

Methods: We performed a post-hoc analysis of T1DEXI data. A linear mixed-effects model was conducted to estimate whether BMI modifies the acute effect of exercise on exercise days compared to sedentary days as per CGM-based metrics. Only participants with sedentary days and post-exercise data were included in the analysis. The model included fixed effects for independent variables and a random intercept to account for intra-patient correlation, adjusting for sex, exercise modality, and HbA1c. Models were stratified by therapy (AID vs. non-AID), with significance set at α=0.05. We report the coefficients with their 95%CI. Analyses were conducted using STATA14.0.

Results: 316 participants(≥18y) were included in the analysis (Table 1). BMI was found to significantly modify the acute effect of exercise on %TITR, %TAR, and %CV in AID users, with (ß=-0.55;p=0.037), (ß=0.32p=0.048), and (ß=0.33;p=0.012), respectively. However, no significant modification by BMI was observed on the acute effect of exercise on %TIR (ß=-0.34;p=0.137) or time spent in hypoglycemia for AID users, nor for any CGM metrics in non-AID users after adjusting for covariates (Table 2).

Conclusions: BMI seems to modify the exercise’s acute effect on %TITR, %TAR, and %CV in AID users. As BMI increases, %TAR and %CV also increase, while %TITR decreases. Given the growing prevalence of overweight and obesity in individuals with T1D, personalized exercise approaches and adjunct therapies beyond insulin may be necessary for individuals with higher BMI.

PROGRESSIVE DECLINE IN MEDICATION DEPENDENCY OVER 270 DAYS WITH PRECISION LIFESTYLE INTERVENTIONS IN T2D

Chhavi Mehra, Annie R, Ankitha Sequeira, Shivtosh Kumar

Ragus Healthcare Pvt Ltd, Research And Development, Bengaluru, India

Background and Aims: Urbanisation, ageing populations, and sedentary lifestyles drive the rise in Type 2 Diabetes (T2D), leading to significant health and economic challenges. The SugarFit Diabetes Reversal and Management Program (SDRMP) integrates dietary changes, physical activity, and continuous support, evaluating its effectiveness in improving health outcomes.

Methods: This retrospective study evaluated 1,036 participants (788 male, 248 female) with T2D who enrolled in SDRMP over 9-months. The program included personalized dietary counselling, customised exercise plans, and remote health coaching with continuous support via virtual consultations and regular feedback. Clinical parameters (glycated haemoglobin (HbA1c), fasting blood sugar (FBS), body mass index (BMI), and weight), were assessed at the baseline and 3 quarterly follow ups. System calculated medication scores led to medication titration up or down, be it OHA or insulin. Paired t-tests evaluated the program’s efficacy.

Results: Participants experienced a remarkable decline in HbA1c levels, from 8.60±1.48% at Q0 to 6.68±0.84% at Q1, 6.56±0.90% at Q2, and 6.57±0.81% at Q3. Weight dropped from 77.69±14.11 kg at Q0 to 72.63±11.94 kg at Q3. Further progress was noted in FBS, decreasing from 154.10±60.14 mg/dL at Q0 to 112.67±23.20 mg/dL at Q3. Both HbA1c and weight reductions were highly significant (p<0.001). Medication scores also dropped notably from 1.24±1.47 to 0.18±0.22, indicating reduced dependency on diabetes medications.

Conclusions: The SDRMP proved effective in improving glycemic control (HbA1c by 2.03%), reducing weight (5.06 kg), and decreasing medication dependency. These findings highlight its potential as a valuable tool for managing type 2 diabetes.

CLUSTERS OF USER EXPERIENCES RELATED TO A BI-HORMONAL FULLY CLOSED LOOP SYSTEM IN REAL-WORLD TYPE 1 DIABETES CARE

Giesje Nefs ^1,2,3,4, Theodorus Jansen⁵, Michel Klaassen⁵, Lyan Vlaskamp⁵, Maureen Witkop⁵, J Hans Devries⁶, Arianne Van Bon⁷

¹Radboud university medical center, Nijmegen, Netherlands, ²Diabeter, Rotterdam, Netherlands, ³Tilburg University, Tilburg, Netherlands, ⁴Diabeter Centrum Amsterdam, Amsterdam, Netherlands, ⁵Inreda Diabetic B.V., Goor, Netherlands, ⁶Amsterdam UMC, Amsterdam, Netherlands, ⁷Rijnstate Hospital, Arnhem, Netherlands

Background and Aims: A Dutch bi-hormonal fully closed loop system (FCL) improved glycemic time in range (TIR), well-being (WHO-5), and diabetes-specific distress (PAID) among adults with type 1 diabetes. This mixed-method study examined (a) how glycemic and person-reported outcomes (PROs) cluster together; and (b) qualitative user experiences.

Methods: In the FREE 1 multi-center single-arm intervention study, participants completed questionnaires at FCL start and follow-up (three, six, nine, twelve months). K-means cluster analysis identified subgroups with similar outcomes. Eight continuing and four discontinued users were interviewed. Clustering variables were change scores (last-first assessment) for PAID-20, sleep quality (short PSQI), system trust (short TAS), usability appraisals (purpose-designed questions), TIR, time above range (TAR), and time below range (TBR).

Results: Four clusters contained people with (a) large TBR improvement, moderate TIR and distress improvement (n=7), (b) the largest TIR and TAR improvement, but smallest distress improvement, negative sleep changes, lowest usability appraisal and trust (n=11), (c) moderate distress and TIR improvement, and high usability appraisal (n=29), and (d) large TIR, TAR, distress and sleep improvements, and high usability and trust, with small TBR improvement (n=17). Benefits mostly outweighed burdens (glucagon replacement, multiple components, alarms). FCL use took continued daily effort, but also represented a positive life changer. Discontinuation reasons were often present early, stacking over time. Participants stressed expectations and manufacturer support.

Conclusions: On the subgroup and individual level, there is variety in FCL experience and outcomes. To facilitate FCL adoption, expectation management seems key. Study funder: Inreda Diabetic. Dutch Trial Register-NL9578.

A NOVEL APPROACH FOR HEALTHCARE QUALITY MEASUREMENT, IN PREGESTATIONAL TYPE 1 DIABETES – A CASE STUDY WITH THE MINIMED™ 780G SYSTEM

Tali Cukierman-Yaffe ¹, Keren Levi², Rakefet Yoeli²

¹Sheba Medical Center, Tel-Aviv university, Division Of Endocrinology Diabetes And Metabolism, Ramat-Gan, Israel, ²Sheba Medical Center, Department Of Obstetrics And Gynecology, Ramat, Israel

Background and Aims: Treatment of pregestational women with Type 1 Diabetes (T1D) requires tight glucose control and a multidisciplinary approach to reduce risks like LGA macrosomia and neonatal hypoglycemia. Continuous quality assessment is crucial for achieving optimal pregnancy outcomes. This analysis evaluates the use of the Galaxy platform for real-time quality control in monitoring maternal glucose indices and their impact on pregnancy outcomes in women using the MiniMed™ 780G system. The goal is to implement the platform for continuous quality assurance and benchmarking to enhance outcomes, particularly in relation to neonates born LGA (large for gestational age).

Methods: This analysis uses data from the Galaxy platform at Sheba Medical Center (2021–2024). Glycemic and therapy data of women using the MiniMed™ 780G in automode during their third trimester were combined with EMR data on pregnancy outcomes. These outcomes were compared to a historical cohort of women not using the Hybrid Closed Loop (HCL) system.

Results: This analysis pertains to 21 pregnancies. 23.8% of neonates were LGA. TIR was lower than reported in the historical cohort of 38 women who did not use the HCL system; however, pregnancy outcomes were similar.

Conclusions: The Galaxy platform provides ongoing data to assess the MiniMed™ 780G system's therapy. It showed effective glucose management with LGA rates comparable to the historical cohort, despite lower TIR. The platform offers real-time assessment of glucose control during pregnancy. As technology advances, the platform has the potential to assess the relationship between glucose metrics and outcomes across larger, multi-center international studies.

PREDICTING GESTATIONAL DIABETES MELLITUS (GDM) USING CONTINUOUS GLUCOSE MONITORING (CGM)-MEASURED DATA FROM COMBINED US AND UK COHORTS

Celeste Durnwald ¹, Zoey Li², Anders Carlson³, Richard M. Bergenstal³, Peter Calhoun², Judy Sibayan², Elizabeth Norton¹, Tanja Kleinhappel⁴, Graham Law⁴, Roy Beck⁵, Helen Murphy⁶, Eleanor Scott⁷

¹Maternal Fetal Medicine Research Program, University of Pennsylvania, Department Of Obstetrics And Gynecology, Philadelphia, United States of America, ²Jaeb Center for Health Research, Tampa, United States of America, ³International Diabetes Center, HealthPartners Institute, St Louis Park, United States of America, ⁴School of Health & Care Sciences, Lincoln, United Kingdom, ⁵Jaeb Center for Health Research Foundation, Inc., Tampa, United States of America, ⁶University of East Anglia, Norwich Medical School, Norwich, United Kingdom, ⁷Leeds Institute of Cardiovascular and Metabolic Medicine, University of Leeds, Leeds, United Kingdom

Background and Aims: Previous studies have described higher glucose levels measured by CGM with the presence of GDM. Using data from two prospective observational studies (US-based (GLAM) and UK-based (MAGIC), we aimed to predict subsequent diagnosis of GDM using CGM-derived glucose data in early pregnancy.

Methods: A total of 1130 pregnant participants were included. Presence of GDM was determined through an oral glucose tolerance test (OGTT) using locally applicable established diagnostic criteria. CGM metrics during gestational weeks 13-15, 21-22, 27-28 prior to the OGTT were calculated. Area under the receiver operating characteristic curve (AUROC) was used to assess performance of CGM metrics in predicting GDM.

Results: Pregnant participants with GDM showed higher glucose levels throughout gestation when compared to participants without GDM (Figure 1). Second trimester % time >120 mg/dL and >140 mg/dL (TA140) had AUROCs of 0.76 and 0.77, respectively, in predicting GDM. AUROC of using gestational week 13-15 TA140 ranged from 0.66 to 0.71 (Figure 1). For participants with week 13 TA140 >5%, 19% would subsequently be diagnosed with GDM using OGTT criteria. For those with ≤5% TA140, 95% would also test negative on the OGTT.

Conclusions: In a combined cohort spanning both the US and UK regions, CGM-measured glucose metrics were predictive of GDM, from as early as 13-15 weeks gestation. Use of CGM to screen for hyperglycemia risk during early pregnancy warrants further study.

PREDICTING ADVERSE PERINATAL OUTCOMES (APOS) USING CONTINUOUS GLUCOSE MONITORING (CGM)-MEASURED DATA FROM COMBINED US AND UK COHORTS

Eleanor Scott ¹, Zoey Li², Peter Calhoun², Tanja Kleinhappel³, Graham Law³, Judy Sibayan², Elizabeth Norton⁴, Anders Carlson⁵, Richard M. Bergenstal⁵, Roy Beck⁶, Helen Murphy⁷, Celeste Durnwald⁴

¹Leeds Institute of Cardiovascular and Metabolic Medicine, University of Leeds, Leeds, United Kingdom, ²Jaeb Center for Health Research, Tampa, United States of America, ³School of Health & Care Sciences, Lincoln, United Kingdom, ⁴Maternal Fetal Medicine Research Program, University of Pennsylvania, Department Of Obstetrics And Gynecology, Philadelphia, United States of America, ⁵International Diabetes Center, HealthPartners Institute, St Louis Park, United States of America, ⁶JAEB Center for Health Research, Research, Tampa, United States of America, ⁷University of East Anglia, Norwich Medical School, Norwich, United Kingdom

Background and Aims: Previous studies have associated higher glucose levels with the development of APOs such as large for gestational age(LGA) infants or hypertensive disorders of pregnancy(HDP). In this analysis, we used CGM-derived glucose data from two prospective US-based (GLAM) and UK-based (MAGIC) studies to predict APOs.

Methods: A total of 1130 pregnant participants were included. Outcomes of interest included LGA, HDP, and neonatal intensive care unit(NICU) admissions. CGM metrics during gestational weeks 13-15, 21-22, and 27-28 prior to the oral glucose tolerance test(OGTT) were calculated. Area under the receiver operating characteristic curve(AUROC) was used to assess performance of CGM metrics in predicting APOs.

Results: In both the US and UK studies, maternal glucose levels were consistently higher throughout gestation for those with vs. without HDP and for those with infants admitted vs. not admitted to the NICU. Maternal glucose levels were higher in later gestation for those with vs. without LGA. AUROCs using % time >140 mg/dL(TA140) for predicting LGA, HDP, and NICU admissions were 0.53, 0.55, and 0.56, respectively using gestational week 13 CGM data and increased to 0.57, 0.60, and 0.58 using gestational week 22 CGM data. First trimester TA140 achieved similar sensitivity to second trimester OGTT in predicting APOs (Figure 1).

Conclusions: In a combined analysis using US and UK data, CGM-measured metrics had comparable performance to OGTT in predicting APOs, even when using first trimester CGM metrics. Use of early CGM as a screening tool for gestational hyperglycemia and earlier clinical management for potential prevention of APOs warrants further study.

BLUNTED CARDIAC AUTONOMIC NERVOUS SYSTEM FLEXIBILITY IN TYPE1 DIABETES INDIVIDUALS COMPARED TO HEALTHY PEERS

Ivana Potočnik¹, Eva Gabrovšek², Tadej Battelino³, Klemen Dovč⁴, Nejka Potocnik ¹

¹University of Ljubljana, Faculty Of Medicine, Institute Of Physiology, Ljubljana, Slovenia, ²Faculty of Health Sciences, University of Ljubljana, Department Of Physiotherapy, Ljubljana, Slovenia, ³Medical University of Graz, Division Of Endocrinology And Diabetology, Department Of Internal Medicine, Graz, Austria, ⁴University Children's Hospital, University Medical Centre Ljubljana, and Faculty of Medicine, University of Ljubljana, Department Of Endocrinology, Diabetes And Metabolism, Ljubljana, Slovenia

Background and Aims: Autonomic nervous system dysfunction (ANSD) is a common complication of type 1 diabetes (T1D). The flexibility of cardiac autonomic nervous activity (CANF) upon increased physiological demand, such as endurance exercise (EE), could serve as an early marker of ongoing ANSD. We aimed to determine CANF in participants with T1D compared to healthy controls.

Methods: Recreational athletes with T1D, achieving recommended glycemic outcomes, and age, gender and maximal aerobic power (VO2max) matched controls (12/12) performed submaximal graded cycling (GC) at three different occasions: before, half an hour and 24 hours after EE in the form of 120min cycling at 65%VO2max. During each GC, heart rate (HR) was measured, and HR variability indices were determined at rest and in the recovery. The root mean square of successive differences (RMSSD), power of the high (HF) and low-frequency (LF) bands, LF/HF ratio, HR recovery in 30 and 60 seconds (HRR30, HRR60) were calculated. A two way ANOVA for repeated measures was performed using SPSS.

Results: An interaction (time X T1D/controls) was found in HR (p = 0.006), RMSSD (p=0,002) and HF (p=0.018) at rest and in HRR30(60) (p=0.007(0.003)). Biphasic time response was detected in all above-mentioned indices in controls (p<0.001 for all) revealing cardiac autonomic dysfunction half an hour and supercompensation 24 hours after EE, while in T1D subjects only HRR30 showed a main time effect (p=0.048).

Conclusions: CANF is blunted in T1D participants compared to controls. Indices of HR variability in response to EE could potentially serve as surrogates for early detection of ANSD in T1D.

IMPROVEMENT IN GLYCEMIC CONTROL AND ELIMINATION OF EXOGENOUS INSULIN USE IN PATIENTS WITH TYPE 1 DIABETES INFUSED WITH FULLY DIFFERENTIATED ISLET CELLS: TRIAL VX-880-101 (FORWARD)

Michael Rickels ¹, Trevor Reichman², James Markmann¹, Jon Odorico³, Piotr Witkowski⁴, John Fung⁴, Martin Wijkstrom⁵, Fouad Kandeel⁶, Anne Peters⁷, Chantal Mathieu⁸, Leslie Kean⁹, Bote Bruinsma¹⁰, Chenkun Wang¹⁰, Janet Hong¹⁰, Bastiano Sanna¹⁰, Gautham Marigowda¹⁰, Felicia Pagliuca¹⁰, Doug Melton¹⁰, Camillo Ricordi¹¹

¹University of Pennsylvania Perelman School of Medicine, Philadelphia, United States of America, ²Toronto General Hosp., University of Toronto, Toronto, Canada, ³University of Wisconsin, Madison, United States of America, ⁴University of Chicago, Chicago, United States of America, ⁵University of Pittsburgh Medical Center, Pittsburgh, United States of America, ⁶City of Hope National Medical Center, Duarte, United States of America, ⁷University of Southern California, Los Angeles, United States of America, ⁸Katholieke Universiteit Leuven, Leuven, Belgium, ⁹Dana Farber/Boston Children’s Hospital, Boston, United States of America, ¹⁰Vertex Pharmaceuticals Incorporated, Boston, United States of America, ¹¹University of Miami, Miami, United States of America

Background and Aims: VX-880, is an allogeneic stem cell-derived, fully differentiated islet cell therapy, evaluated in a Phase 1/2 trial.

Methods: This single-arm, open-label trial evaluates VX-880 with concomitant steroid-free immunosuppression in adults with T1D, impaired hypoglycemia awareness, and at least 2 severe hypoglycemic episodes (SHEs) the year before enrollment.

Results: Fourteen participants received VX-880: 2 initially received half the dose and 12 the full dose (single infusion) (May2024). At baseline, participants had multiple SHEs, HbA1c greater than 7%, and no detectable C-peptide. After infusion, all participants demonstrated engraftment and VX-880 function, evidenced by clinically meaningful levels of glucose responsive C-peptide production on Day (D) 90 MMTT. Twelve participants receiving the full dose (single infusion) had marked glycemic control improvement (mean HbA1c:baseline=7.8%;D180=6.2%), glucose time in range (TIR) of greater than 70% (mean TIR:baseline=49.5%;D180=88.3%), and were free of SHEs at greater than 90 days post-infusion. Marked reduction in exogenous insulin use was seen in 11/12 participants (last visit), with 9/12 stopping exogenous insulin use at/or after D180. There were no VX-880-related serious adverse events and two deaths (unrelated to VX-880): one due to cryptococcal meningitis infection after complications from elective sinus surgery, prolonged high-dose steroid use (protocol-prohibited), and immunosuppressive medication; one due to progression of pre-existing neurocognitive impairment from traumatic brain injury (motor vehicle accident from a SHE before enrollment). VX-880 safety was consistent with immunosuppressive regimen, islet infusion procedure, and long-standing T1D.

Conclusions: VX-880 has curative potential in T1D with elimination of severe hypoglycemia, significantly improved glycemic control, and freedom from exogenous insulin use.

GLUCINDA: A TOOL FOR IMPORTING AND HARMONIZING CGM DATA FOR RESEARCH

Petra Baumann ¹, Alexander Grebien¹, Christoph Buhlheller¹, Christian Taucher¹, Florian Posch², Lia Bally³, Julia Mader¹

¹Medical University of Graz, Division Of Endocrinology & Diabetology, Graz, Austria, ²Medical University of Graz, Graz, Austria, ³Inselspital, Universitätsspital Bern, Bern, Switzerland

Background and Aims: In addition to personalized diabetes management, CGM data are essential for clinical research. Many studies rely on real-world data from participants’ personal CGM devices that are delivered in multiple formats, mostly without meta information on data structure. Importing, harmonising, and ensuring the quality of the data is therefore a time-consuming and challenging task.

Methods: CGM exports are provided in a multitude of file (CSV, XLSX, TXT) and table formats (number of unusable header rows, differing column names etc.). Sometimes, even the file structure changes mid-document. To deal with this multitude of characteristics, we developed a software tool in Python: GluCINDa (Glucose: Coherent Import. Neat Data.) takes on multiple exchange formats, detects glucose measurements with date/time information and produces a harmonised dataset.

Results: A GluCINDa file (CSV) includes: participant ID, date, time, glucose value, glucose unit, glucose source column name, source file name and further notes if required. Moreover, a documentation file allows tracing for each parsed value. GluCINDa also generates a report file with a basic summary about the extracted data and a separate CGM file. Various settings can also be set manually in GluCINDa’s advanced mode (e.g. additional Feather export, skip documentation to save memory).

Conclusions: Not only is GluCINDa faster than regular importing (which sometimes requires thousands of lines of code), it also minimizes human error and data loss due to unruly raw data. Additional functionalities, such as parsing of glucose data from other sources than CGM or independence from a glucose-related keyword in the header are work in progress.

THE IMPACT OF SGLT2I AND GLP1RA ON ETHNICITY-SPECIFIC ALL-CAUSE MORTALITY IN PATIENTS WITH TYPE 2 DIABETES AND CARDIOVASCULAR/RENAL DISEASE

Lynne Chepulis ¹, Mark Rodrigues¹, Han Gan², Rinki Murphy³, Leanne Te Karu³, Ryan Paul^1,4

¹University of Waikato, Division Of Health, Hamilton, New Zealand, ²University of Waikato, Department Of Mathematics, Hamilton, New Zealand, ³University of Auckland, Faculty Of Medical And Health Sciences, Auckland, New Zealand, ⁴Waikato Regional Diabetes Service, Te Whatu Ora Waikato, Hamilton, New Zealand

Background and Aims: SGLT2i and GLP1RA use associates with improved outcomes in patients with type 2 diabetes (T2D) but there is limited data regarding their effect on mortality for Indigenous Māori vs other groups. This study reports 3-year mortality rates in a New Zealand (NZ) T2D population following the funded availability of SGLT2i and GLP1RA (empagliflozin/dulaglutide/liraglutide; EDL) with prioritised access for Māori and Pacific people.

Methods: Data were collected from primary care records for those aged 18-75y with T2D (Feb 2021 to Dec 2023; n=4648) and were linked to national medication dispensing and mortality records via health identifier numbers. Mortality rates were calculated for those with cardiovascular and/or renal disease/risk (CVRD), including by ethnicity and who did/did not initiate EDL (≥ 2 dispensing events).

Results: Overall, 11,881 had CVRD (of which 5463 received EDL) and 932 deaths were recorded (7.8%). The relative 3-Year mortality risk (RR) in CVRD patients was significantly lower after initiating EDL compared to no EDL (3.6% vs 11.5%, RR 0.31 [95% CI 0.27,0.36]; P < 0.001). Reductions in age-adjusted 3-year mortality following EDL use in CVRD patients was greatest for NZ Māori (-12.0%, RR 0.23; P < 0.001) compared to Pacific (-3.7%; RR 0.38; P = 0.04) and non-Māori, non-Pacific people (-5.6%; RR 0.35; P < 0.001). Higher reductions in mortality with EDL use associated with increased patient age and HbA1c (P < 0.001).

Conclusions: SGLT2i/GLP1RA use demonstrates ethnicity-specific improvements in mortality in T2D, with the greatest improvements in Indigenous Māori. We will also report on non-CVRD patients and other ethnic groups.

CARDIOPULMONARY AND AEROBIC EXERCISE TESTS IN TYPE 1 DIABETES UNDER DIFFERENT CIRCUMSTANCES: THE ACT-ONE STUDY

Francesca De Ridder ¹, Kristien Ledeganck², Klemen Dovč³, Nejka Potocnik⁴, Michael Oberosler⁵, Silvia Kheir⁵, Jana Windhaber⁶, Julia Mader⁷, Tadej Battelino⁵, Christophe De Block⁸

¹Antwerp University Hospital, Edegem, Belgium, ²University of Antwerp, Faculty Of Medicine & Health Science, Laboratory of Experimental Medicine And Pediatrics, Antwerp, Belgium, ³University Children's Hospital, University Medical Centre Ljubljana, and Faculty of Medicine, University of Ljubljana, Department Of Endocrinology, Diabetes And Metabolism, Ljubljana, Slovenia, ⁴University of Ljubljana, Faculty Of Medicine, Institute Of Physiology, Ljubljana, Slovenia, ⁵Medical University of Graz, Division Of Endocrinology And Diabetology, Department Of Internal Medicine, Graz, Austria, ⁶Medical University of Graz, Department Of Pediatric And Adolescent Surgery, Graz, Austria, ⁷Medical University of Graz, Division Of Endocrinology & Diabetology, Graz, Austria, ⁸Antwerp University Hospital, Endocrinology-diabetology, Brasschaat, Belgium

Background and Aims: We investigated optimal insulin adjustments for prolonged (90minutes) aerobic exercise, aiming for maximal euglycemia (70-180mg/dL, time in range, TIR).

Methods: Adults with type 1 diabetes using multiple daily injections (MDI, degludec/glargine as basal insulin (n=19)) or automated insulin delivery (AID, Medtronic 780G (n=27)) were included. Participants performed first a baseline maximal exercise test (CPET) to determine the VO2peak, followed by two 90-minute aerobic exercise tests (AEX) at 45% VO2 peak. Thirty minutes prior to AEX1 and AEX2, a standardized breakfast was given with a 50% insulin bolus. For MDI, basal insulin was administered at 50% vs. 80% dose. For AID, the temporary target of 150 mg/dL was set 60 vs. 30minutes before exercise until 60minutes after exercise. Venous glucose levels were analyzed with EKF.

Results: The participants (MDI vs. AID) had a mean age of 40.1±9.7y vs. 28.0±9.8y, a mean diabetes duration of 19.3±9.9y vs. 14.8±8.1y and a mean HbA1c of 6.7±0.6% vs. 7.0±0.5%. Within the MDI group, 50% vs. 80% of basal insulin dose were compared respectively with TIR being 71.8±31.9% vs. 77.8±34.0% during exercise. After exercise, TIR was 67.7±25.8% vs. 78.0±29.5%. In the AID group, temporary target 60 vs. 30minutes prior to aerobic exercise were compared respectively: TIR during exercise was 79.3±28.3% vs. 80.2±27.6%. After exercise, TIR was 82.6±21.0% vs. 80.3±22.7%.

Conclusions: MDI protocols resulted in ≥70% TIR for 56% vs. 67% of participants using 50% vs. 80% basal insulin. AID protocols succeeded in ≥70% TIR for 73% vs. 76% of participants using temporary target 60 vs. 30minutes prior to exercise.

DISORDERED EATING BEHAVIORS AND PREDICTORS IN YOUTH WITH TYPE 1 DIABETES AND CELIAC DISEASE

Alda Troncone¹, Marco Marigliano², Enza Mozzillo³, Alberto Casertano⁴, Elisa Morotti⁵, Dario Iafusco⁶, Stefano Passanisi⁷, Fortunato Lombardo⁷, Delvecchio Maurizio⁸, Riccardo Pertile⁹, Roberto Franceschi ¹⁰

¹University of Campania, Naples, Italy, ²University and Azienda Ospedaliera Universitaria Integrata of Verona, Department Of Surgery, Dentistry, Pediatrics And Gynecology, Section Of Pediatric Diabetes And Metabolism, Verona, Italy, ³Università degli Studi di Napoli Federico II, Naples, Italy, Department Of Translational Medical Science, Section Of Pediatrics, Naples, Italy, ⁴University of Naples Federico II, Department Of Translational Medical Science, Section Of Pediatrics, Naples, Italy, ⁵University and Azienda Ospedaliera Universitaria Integrata of Verona, Verona, Italy, ⁶University of Campania L. Vanvitelli, Pediatric Diabetology “g.stoppoloni”, Napoli, Italy, ⁷University of Messina, Department Of Human Pathology Of Adulthood And Childhood G. Barresi, Messina, Italy, ⁸Università degli Studi dell'Aquila, L’Aquila, Italy, ⁹Azienda Provinciale Per i Servizi Sanitari del Trentino, Clinical And Evaluative Epidemiology Unit, Health Management, Trento, Italy, ¹⁰Diabetology Unit, Pediatric Clinic - S.chiara Hospital Of Trento, Trento, Italy

Background and Aims: Data on Disordered Eating Behaviors (DEBs) occurrence in cohorts with Type 1 Diabetes (T1D) and Celiac Disease (CD) are very few and somehow conflicting. This study aimed to analyze differences in the prevalence of DEBs between T1D+CD vs. T1D and to identify the main predictors of DEBs.

Methods: An observational case-control study, comparing subjects aged 8-18 years with T1D and CD, with people with T1D only, was performed. Eating behavior was assessed using Children’s Eating Attitude Test (ChEAT) in subjects with T1D and the ones with double disease.

Results: 69 youths with T1D and CD, 84 controls with T1D only, were included in the study. The first group reported a prevalence of DEBs of 42.4% compared to 9.5% of subjects with T1D only (<0.001). Items regarding self-control around food and calorie checks had the highest score in the T1D+CD group. In both group multiple logistic regression analysis showed that to be male (OR= 1.450) and with higher BMI z-score (OR= 1.310) gave a higher probability to present a ChEAT total score > clinical cutoff.

Conclusions: Subjects with T1D and CD present a higher prevalence of DEBs. Personalized care plans and interventions for subjects with T1+CD, designed to prevent their evolution to eating disorders (EDs) and that consider relevant DEB predictors, are needed.

RELATIONSHIP OF TIME IN TIGHT RANGE-TIME IN RANGE WITH LIPID PROFILE IN TYPE 1 DIABETES PATIENTS WITH DIFFERENT GLUCOSE MEASUREMENT AND INSULIN TREATMENT MODALITIES

Efstratios Kardalas, Lina Mandrapylia, Dimitra Mpaikousi, Theodora Stratigou, Marinella Tzanella

Evangelismos Hospital, Endocrinology And Diabetes Department, Athens, Greece

Background and Aims: Type 1 diabetes (T1D) is often accompanied with abnormal lipid metabolism. The aim of this study was to evaluate the association of the latter with time in tight range (TITR) and time in range (TIR) in patients on multiple daily insulin regimen (MDIR) or on continuous glucose measurement system (CGMS) or on advanced closed hybrid loop (ACHL).

Methods: 210 patients with T1D (60 on ACHL, 85 with CGMS and 65 on MDIR) were retrospectively studied. The association between TITR/TIR and lipid metabolism parameters, non adjusted and adjusted for HbA1c and other confounding factors, was studied.

Results: Patients exhibited moderate glycemic control (HbA1c: 7.2± 1.3%; TIR 57.5 ± 12.0% and TITR 46.2± 13.4%). ACHL patients had statistically significant higher TITR and TIR levels while mean HbA1c was markedly elevated in the MDIR group compared to the other groups. Additionally, triglycerides and LDL values were strongly and negatively associated with TITR and TIR while HDL was positively associated with TITR and TIR in all 3 groups. Interestingly, the correlation of TITR to triglycerides and LDL persisted even after adjustment for HbA1 and other confounding factors except T1D duration and age in all groups. This observation did not apply to TIR. Finally, all previous correlations were stronger among ACHL patients compared to the other ones.

Conclusions: TITR and TIR are tightly correlated to lipid levels in people with T1D. Further studies are needed to ‘explore’ the significance of these glycemic control parameters regarding optimal lipid metabolism control in these patients.

PRE- AND POST-TRANSPLANT CHARACTERISTICS OF INDIVIDUALS WITH TYPE 1 DIABETES WHO RECEIVED A SOLID ORGAN TRANSPLANT IN THE UNITED STATES

Rebecca Molinsky¹, Vincent Tan¹, Catherine Vick², Andrew Joyce², Erica Briola², Jack Varon¹, Martin Wijkstrom ³

¹Vertex Pharmaceuticals Incorporated, Boston, United States of America, ²Venebio Group, Glen Allen, United States of America, ³University of Pittsburgh Medical Center, Pittsburgh, United States of America

Background and Aims: Individuals with type 1 diabetes (T1D) may have a solid organ transplant due to consequences of their T1D, as a treatment for severe T1D, or due to another underlying condition. The epidemiology of transplants in the T1D population is not well described.

Methods: We conducted a descriptive analysis using data from the Scientific Registry of Transplants in adults with T1D who received solid organ transplants from 2000 to 2022 in the US to describe trends over time and pre- and post- transplant characteristics.

Results: Among 37,699 distinct patients with a history of T1D, 42,654 transplants occurred. Our results focus on the first and most common transplant types (n=37,531) including kidney n=15,581 (42%), simultaneous kidney/pancreas n=15,372 (41%), pancreas n=2,871 (8%), liver n=1,928 (5%), lung n=968 (3%), and heart n=811 (2%). Among kidney, kidney/pancreas, and pancreas transplants, 93% had T1D listed as the primary reason for transplant. At the time of transplant, mean age was 45 years; 42% female; 82% white; 46% were privately insured and 47% on Medicare. The most common induction immunosuppression regimens used included anti-thymocyte globulin (49%) and basiliximab (18%). The most common maintenance immunosuppression regimens used were mycophenolic acid (89%), tacrolimus (88%), and corticosteroids (75%).

Conclusions: A considerable number of kidney and equal numbers of simultaneous pancreas-kidney transplants have been performed among individuals with T1D. This continued requirement for transplants highlights the severity and consequences of T1D, and the need for treatments that address the underlying cause of disease.

INSULIN THERAPY DE-INTENSIFICATION WITH IGLARLIXI (IDEAL) RANDOMISED CONTROLLED TRIAL: PATIENT-REPORTED OUTCOMES

Ondrej Zizka ^1,2, Peter Novodvorsky², Lenka Thieme², Ivana Lankova², Alica Vesela³, Emil Zahumensky⁴, Tomas Edelsberger⁵, Marie Loblova⁶, Filip Hruby⁷, Milos Mraz^1,2, Martin Haluzík^1,2

¹Charles University, First Faculty Of Medicine, Prague, Czech Republic, ²Institute for Clinical and Experimental Medicine, Diabetes Centre, Prague, Czech Republic, ³EDUMED s.r.o, Nachod, Czech Republic, ⁴Internal and Diabetology Clinic, Valasske Klobouky, Czech Republic, ⁵Private Diabetology Clinic, Krnov, Czech Republic, ⁶Diaclinic s.r.o., Ceske Budejovice, Czech Republic, ⁷Institute for Clinical and Experimental Medicine, Department Of Data Science, Prague, Czech Republic

Background and Aims: IDEAL examined the efficacy and safety MDI insulin regimen de-intensification into once-daily iGlarLixi in people with type 2 diabetes (PwT2D) and showed comparable glycaemic control, improvement in several CGM outcomes together with the reduction of body weight, total daily insulin dose and number of insulin injections. Here we report patient-reported outcomes.

Methods: IDEAL was a 5-centre, open-label, parallel-group, phase 4 RCT with a 24-week active treatment period. Participants were randomised in a 1:1 fashion into iGlarLixi or continuation with MDI and received the following questionnaires: Diabetes Treatment Satisfaction Questionnaire static (DTSQs) at baseline (BL) and at week 24 (W24), Diabetes Treatment Satisfaction Questionnaire change (DTSQc) at W24, Hypoglycaemia Attitudes and Behaviour Scale (HABS) at BL and W24, and the Audit of Diabetes Dependent Quality of life (ADDQoL) at BL and W24.

Results: At W24, the (median [IQR]) DTSQs satisfaction score was higher in the iGlarLixi vs MDI group: (35.0 [31.0, 36.0] vs 29.0 [24.0, 32.0], p < 0.001). The DTQSc (W24) satisfaction score was higher in the iGlarLixi vs MDI group: (16.0 [11.5, 18.0] vs 8.5 [5.0, 15.0], p = 0.006). Significant differences in favour of iGlarLixi were observed in the following DTSQs and/or DTSQc items: convenience, continuation, current treatment satisfaction, treatment recommendation, flexibility, perceived hypoglycaemia and treatment understanding. No changes were observed with the ADDQoL and HABS questionnaires.

Conclusions: Insulin therapy de-intensification from MDI into iGlarLixi in PwT2D led to significant improvement in patient satisfaction with diabetes treatment as evidenced by the data from DTSQs/DTSQc questionnaires.

ORAL PRESENTATIONS SESSION 04: NEW MEDICATIONS FOR DIABETESINSULIN THERAPY DE-INTENSIFICATION WITH IGLARLIXI (IDEAL) RANDOMISED CONTROLLED TRIAL – CGM RELATED OUTCOMES

Peter Novodvorsky ¹, Lenka Thieme¹, Ivana Lankova¹, Alica Vesela², Emil Zahumensky³, Tomas Edelsberger⁴, Marie Loblova⁵, Filip Hruby¹, Milos Mraz¹, Martin Haluzík¹

¹Institute for Clinical and Experimental Medicine, Prague, Czech Republic, ²EDUMED s.r.o, Nachod, Czech Republic, ³Internal and Diabetology Clinic, Valasske Klobouky, Czech Republic, ⁴Private Diabetology Clinic, Krnov, Czech Republic, ⁵Diaclinic s.r.o., Ceske Budejovice, Czech Republic

Background and Aims: The IDEAL RCT examined the efficacy and safety of MDI de-intensification into once-daily administered iGlarLixi in people with type 2 diabetes (PwT2D) and showed comparable glycaemic control, reduction of body weight, total daily insulin dose and number of insulin injections. Here we report CGM-related outcomes.

Methods: IDEAL was a 5-centre, open-label, parallel-group, phase 4 RCT with a 24-week active treatment period. Participants were randomised (1:1) into iGlarLixi or continuation with MDI and were applied Freestyle Libre Pro iQ before the onset (BL) and at the end of the treatment period (W24).

Results: Altogether 73/91 (80.2%) of IDEAL participants were analysed. Median (IQR) %TIR increased in iGlarLixi group (70 [57 – 88] at BL vs 80 [61 – 92] at W24, p=0.039), but remained unchanged in MDI group (78 [67 – 89] vs 82 [64 – 89], p=0.728). In iGlarLixi group reductions in %TAR (26 [11 – 44] vs 13 [6 – 35], p=0.04), %of time in level 2 hyperglycaemia (2 [1 – 9] vs 1 [0 – 4], p=0.019), mean glucose (8.6 [7.4 – 10.1] vs 7.7 [6.9 – 9.2] mmol/L, p=0.012) and glycaemia risk index (GRI) (26 [10 – 46] vs 19 [8 – 35], p=0.036) were detected. Median (IQR) %TBR (< 3.9 mmol/L) remained unchanged in iGlarLixi group (0 [0 – 1] vs 1 [0 – 2], p=0.337) and MDI group (0.5 [0 – 2] vs 1 [0 – 3.75], p=0.147).

Conclusions: Insulin therapy de-intensification from MDI into iGlarLixi in PwT2D showed improvement in several CGM metrics without increase in hypoglycaemia.