bims-glumda Biomed News
on CGM data in management of diabetes
Issue of 2025–12–07
23 papers selected by
Mott Given
  1. Identifying Hypothyroidism as Complication of Type 1 Diabetes from Continuous Glucose Monitoring Data.
  2. Enhanced Detection of Adverse Glycemic Events in Hospital with Unmasked Continuous Glucose Monitoring Versus Point-of-Care Testing in People with Type 1 Diabetes.
  3. Continuous Glucose Monitoring in Older Adults With Diabetes and Alzheimer Disease and Related Dementias-Promise and Perspective.
  4. AttenGluco: Multimodal Transformer-Based Blood Glucose Forecasting on AI-READI Dataset.
  5. "What is it like to live with a CGM?": A qualitative study on type 1 diabetes patients' experiences.
  6. Kalman Filter-based Microwave Blood Glucose Monitoring using A Priori Physiological Model.
  7. Clustering type 1 diabetes patients based on blood glucose level measurements.
  8. Glycemic variability before and during Ramadan fasting among adults with type 2 diabetes and hepatic cirrhosis: a prospective paired cohort using real‑time CGM.
  9. Evaluating the Cost-Effectiveness of Real-Time Continuous Glucose Monitoring Versus Self-Monitoring of Blood Glucose in the Treatment of Patients with Insulin-Treated Type 2 Diabetes in Australia.
  10. Evaluating the Effect of Input Features on Deep Learning Models for Blood Glucose Forecasting.
  11. A deep learning-derived digital biomarker of dysglycemia and its association with genetic risk of type 2 diabetes.
  12. Continuous Glucose Monitoring in Insulin-Treated Older Adults With Diabetes and Alzheimer Disease and Related Dementias.
  13. Predicting Treatment Outcome of Patients with Type 2 Diabetes on Once-Daily Basal Insulin Injections using Machine Learning.
  14. Continuous glucose monitoring patterns among pregnant individuals with varying levels of glucose intolerance.
  15. Understanding Glycemia in the Post Discharge Period Through Blinded Continuous Glucose Monitoring.
  16. Hypoglycemic coma owing to sensor device dysfunction after an automatic smartphone operating system update.
  17. Luseogliflozin Suppresses Late-Night Basal Glycemic Variability in Type 2 Diabetes: A Retrospective Observational Study.
  18. Managing Diabetes One Step at a Time in Low- and Middle-Income Countries: The Promise of Wearable Devices.
  19. GlucoNet: A Hybrid Learning Methodology for Non-Invasive Blood Glucose Estimation from PPG.
  20. Multi-Metric Subgroup Analysis for Glucose Forecasting in Type 1 Diabetes.
  21. Commonly used C-peptide assays show differing associations with CGM metrics.
  22. Analyzing Basal Insulin Dosing as Perturbations in Compositional Glycemic Profiles for Time in Range in Type 1 Diabetes.
  23. From laboratory to pre-market prototyping: lessons learned from a noninvasive blood glucose level measurement device.
  1. Annu Int Conf IEEE Eng Med Biol Soc. 2025 Jul;2025 1-6
    Identifying Hypothyroidism as Complication of Type 1 Diabetes from Continuous Glucose Monitoring Data.
    Agnese Piersanti, Agnese Callegari, Libera L Del Giudice, Christian Gobl, Laura Burattini, Andrea Tura, Micaela Morettini.
      Recent studies have shown that type 1 diabetes mellitus (T1DM) is an important risk factor for the development of hypothyroidism. In this regard, a timely intervention is fundamental to limit adverse effects. Providing real-time measurements of interstitial glucose, Continuous Glucose Monitoring (CGM) devices may represent a powerful source of data to feed machine-learning based algorithms for the discovery of hidden patterns related to the development of diabetes complications such as hypothyroidism. Aim of this study was to setup a machine-learning-based approach capable to identify subjects with hypothyroidism among those with T1DM, starting from CGM tracings. CGM data acquired during a period of 26 weeks and relating to 79 subjects with T1DM taken from the REPLACE-BG campaign database, of which 51 had hypothyroidism and 28 had T1DM with no other complication, were used. The CGM traces were pre-processed to handle the presence of missing data and 41 features were extracted with the use of AGATA software. The feature set was then reduced through Two-Step Decision Tree-Embedded Feature Selection (DT-EFS), leading to the inclusion of 8 final features. The best performing model was the decision tree, showing the following testing performances: area under receiver operating characteristics of 72.3%, accuracy of 71.4%, precision of 74.6%, F1 score of 70.1%, sensitivity of 71.4% and specificity of 69.5%. The 8 features identified herein describe the long-term variability of the subjects' glycemic trace which may suggests a possible connection with the presence of hypothyroidism in T1DM.Clinical Relevance-This establishes the possibility to automatically detect hypothyroidism in T1DM from clinically meaningful CGM glycemic patterns.
    DOI:  https://doi.org/10.1109/EMBC58623.2025.11252628
  2. J Diabetes Sci Technol. 2025 Dec 02. 19322968251395088
    Enhanced Detection of Adverse Glycemic Events in Hospital with Unmasked Continuous Glucose Monitoring Versus Point-of-Care Testing in People with Type 1 Diabetes.
    Ray Wang, Mervyn Kyi, Brintha Krishnamoorthi, Ailie Connell, Cherie Chiang, Debra Renouf, Rahul Barmanray, Spiros Fourlanos.
       BACKGROUND: In people with type 1 diabetes (T1D) admitted to hospital, adverse glycemic events (AGE), both hypoglycemia and hyperglycemia, bestow risk for adverse outcomes. Continuous glucose monitoring (CGM) use is increasingly common amongst people with T1D. We investigated AGE frequency in hospital, based on CGM versus point-of-care (POC) blood glucose measures.
    METHODS: In this multi-center retrospective analysis of non-critically ill hospitalized adults with T1D who continued wearing their unmasked CGM (FreeStyle Libre 1/2, Dexcom G5/G6, Medtronic Guardian 3) during admission and received standard ward-based POC testing, we compared CGM- and POC-based AGE detection of hypoglycemia (<70 mg/dL) and hyperglycemia (>180 mg/dL).
    RESULTS: In 253 admissions, 127 837 CGM and 5508 POC glucose measures were analyzed, yielding 1391 CGM-detected hyperglycemia AGE and 317 CGM-detected hypoglycemia AGE. For CGM-detected AGE with a concurrent POC AGE evident, CGM detected hyperglycemia a median [interquartile range, IQR] of 70 minutes [22, 166] before POC and at lower glucose concentrations (187 vs 223 mg/dL, P < .0001) and detected hypoglycemia a median [IQR] of 38 minutes [14, 65] before POC and at higher glucose concentrations (67 vs 56 mg/dL, P < .0001). A quarter of CGM-detected AGE were not detected by POC. Only 3% of POC-detected AGE were not detected by CGM.
    CONCLUSIONS: Almost all AGE in hospital were detected by CGM, with few detected by POC alone. Compared to POC, CGM detected AGE earlier, with a lesser glycemic extreme, although unmasked CGM use may have influenced these results. Detecting AGE in hospital appears superior with CGM compared to POC glucose alone in people with T1D.
    Keywords:  CGM; T1D; T1DM; adverse glycemia; hospital; inpatient
    DOI:  https://doi.org/10.1177/19322968251395088
  3. JAMA Netw Open. 2025 Dec 01. 8(12): e2541947
    Continuous Glucose Monitoring in Older Adults With Diabetes and Alzheimer Disease and Related Dementias-Promise and Perspective.
    Thaer Idrees, Elena Toschi.
      
    DOI:  https://doi.org/10.1001/jamanetworkopen.2025.41947
  4. Annu Int Conf IEEE Eng Med Biol Soc. 2025 Jul;2025 1-7
    AttenGluco: Multimodal Transformer-Based Blood Glucose Forecasting on AI-READI Dataset.
    Ebrahim Farahmand, Reza Rahimi Azghan, Nooshin Taheri Chatrudi, Eric Kim, Gautham Krishna Gudur, Edison Thomaz, Giulia Pedrielli, Pavan Turaga, Hassan Ghasemzadeh.
      Diabetes is a chronic metabolic disorder characterized by persistently high blood glucose levels (BGLs), leading to severe complications such as cardiovascular disease, neuropathy, and retinopathy. Predicting BGLs enables patients to maintain glucose levels within a safe range and allows caregivers to take proactive measures through lifestyle modifications. Continuous Glucose Monitoring (CGM) systems provide real-time tracking, offering a valuable tool for monitoring BGLs. However, accurately forecasting BGLs remains challenging due to fluctuations due to physical activity, diet, and other factors. Recent deep learning models show promise in improving BGL prediction. Nonetheless, forecasting BGLs accurately from multimodal, irregularly sampled data over long prediction horizons remains a challenging research problem. In this paper, we propose AttenGluco1, a multimodal Transformer-based framework for long-term blood glucose prediction. AttenGluco employs cross-attention to effectively integrate CGM and activity data, addressing challenges in fusing data with different sampling rates. Moreover, it employs multi-scale attention to capture long-term dependencies in temporal data, enhancing forecasting accuracy. To evaluate the performance of AttenGluco, we conduct forecasting experiments on the recently released AIREADI dataset, analyzing its predictive accuracy across different subject cohorts including healthy individuals, people with prediabetes, and those with type 2 diabetes. Furthermore, we investigate its performance improvements and forgetting behavior as new cohorts are introduced. Our evaluations show that AttenGluco improves all error metrics, such as root mean square error (RMSE), mean absolute error (MAE), and correlation, compared to the multimodal LSTM model, which is widely used in state-of-the-art blood glucose prediction. AttenGluco outperforms this baseline model by about 10% and 15% in terms of RMSE and MAE, respectively.
    DOI:  https://doi.org/10.1109/EMBC58623.2025.11251776
  5. BMC Health Serv Res. 2025 Dec 05. 25(1): 1575
    "What is it like to live with a CGM?": A qualitative study on type 1 diabetes patients' experiences.
    Jeong Hee Kang, Youngran Yang, Seonmi Yeom.
       BACKGROUND: The use of continuous glucose monitoring (CGM) is increasing in people with type 1 diabetes mellitus (T1DM) due to its convenience and usefulness for glucose management.
    PURPOSE: This qualitative study explores the experiences and challenges of patients with T1DM regarding their CGM use.
    METHODS: Twenty-nine participants were recruited from a hospital and an online community between May 26 and October 5, 2022. Individual in-depth interviews were conducted, audio-recorded, and subsequently transcribed verbatim. Data were analyzed using Braun and Clarke's six-phase framework for thematic analysis.
    RESULTS: The participants' ages ranged from 19 to 64 years, and the duration of CGM use varied from 2 to 84 months. The qualitative results revealed three themes: typical diabetes care, tied to CGM for life, and conditions for better use. The immediate monitoring capability of CGM was highlighted as a key advantage, liberating individuals from the fear of hypoglycemia and enhancing their overall quality of life. However, the participants also encountered inevitable inconveniences and shared common concerns regarding the accuracy of CGM and the financial burden associated with CGM costs.
    CONCLUSIONS: Healthcare providers and policymakers must address the concerns of patients with T1DM and implement educational activities to bridge the gap between self-monitoring of blood glucose (SMBG) and CGM.
    Keywords:  Blood glucose self-monitoring; Continuous glucose monitoring; Diabetes mellitus
    DOI:  https://doi.org/10.1186/s12913-025-13705-6
  6. Annu Int Conf IEEE Eng Med Biol Soc. 2025 Jul;2025 1-5
    Kalman Filter-based Microwave Blood Glucose Monitoring using A Priori Physiological Model.
    Boyu Ren, Shanaz X Chen, Zheng Gong, Yifan Chen.
      Diabetes mellitus (DM) is a highly prevalent endocrine metabolic disorder, and traditional finger-prick blood glucose monitoring methods pose infection risks. Microwave medical sensing and imaging (MMSI) leverages the dielectric properties of human tissues to facilitate non-invasive, continuous glucose monitoring, demonstrating significant potential for clinical application. This paper proposes a Kalman filter-based algorithm using a physiological glucose concentration model to optimize blood glucose estimation in microwave sensing systems utilizing microgel sensors. The algorithm integrates the physiological glucose model as the prediction step of the Kalman filter and employs a first-order Taylor series expansion to linearize the nonlinear variations in glucose concentration. This approach ensures that the filtering process incorporates both the current actual state and the model's linear approximation. The results indicate that, compared to traditional linear Kalman filtering, the improved algorithm enhances filtering effectiveness by 73.8%, significantly boosting system performance and enabling continuous glucose monitoring.
    DOI:  https://doi.org/10.1109/EMBC58623.2025.11253995
  7. Annu Int Conf IEEE Eng Med Biol Soc. 2025 Jul;2025 1-7
    Clustering type 1 diabetes patients based on blood glucose level measurements.
    Ciro Rodriguez-Leon, Maria Dolores Aviles, Oresti Banos, Pablo Lopez-Ibarra, Manuel Munoz-Torres, Miguel Quesada-Charneco, Claudia Villalonga.
      Improving diabetes mellitus management requires accurate blood glucose predictions to prevent dangerous glycemic extreme events; however, traditional models often struggle with the inherent fluctuations in blood glucose levels. In this work, the identification of distinct patient groups exhibiting common characteristics based on longitudinal continuous glucose monitoring (CGM) data is explored. This clustering approach aims to offer valuable insights for both enhancing predictive modeling accuracy and informing clinical relevance. Raw CGM data was transformed into structured feature sets using statistical descriptors calculated across clinically meaningful time intervals. Following outlier removal, multiple clustering algorithms were evaluated, with the optimal solution selected based on internal metrics and validation by clinical experts. The analysis revealed six distinct patient groups, each characterized by unique glycemic behaviors, including well-controlled, prone to hyperglycemia, and prone to hypoglycemia profiles. It is proposed that these identified clusters can improve glucose prediction by strategically balancing personalized and generalized approaches. Furthermore, valuable insights into individual metabolic variability can be obtained, potentially supporting the development of tailored treatment strategies. While acknowledging limitations inherent in data transformation and expert-driven evaluation, this clustering methodology represents a significant step towards more precise and data-driven diabetes mellitus management.
    DOI:  https://doi.org/10.1109/EMBC58623.2025.11253861
  8. Diabetes Res Clin Pract. 2025 Nov 27. pii: S0168-8227(25)00969-6. [Epub ahead of print]231 112955
    Glycemic variability before and during Ramadan fasting among adults with type 2 diabetes and hepatic cirrhosis: a prospective paired cohort using real‑time CGM.
    Sarah Firdausa, Irsan Hasan, Dicky L Tahapary, Ignatius Bima Prasetya, Suharko Soebardi, Cleopas Martin Rumende, Hamzah Shatri, Cosphiadi Irawan, Wismandari Wisnu.
       AIMS: To compare glycemic variability and continuous glucose monitoring (CGM)-derived metrics before and during Ramadan fasting among adults with type 2 diabetes and cirrhosis.
    METHODS: In this single-center prospective paired-cohort study (Jakarta, January-March 2025), adults with diabetes and cirrhosis wore a real-time CGM system for 14 days pre-and during Ramadan fasting. The primary outcome was coefficient of variation (CV). Secondary outcomes included time in range (TIR), time above range (TAR), time below range (TBR), and ambulatory glucose profiles (AGP).
    RESULTS: Thirty-two adults completed paired CGM assessments. Median CV was 26.95 % pre-Ramadan and 25.45 % during Ramadan (p = 0.84). TIR showed a modest, non-significant increase (+4.0 pp), TAR a mild rise (+5.2 pp), and TBR remained low (<1% in both periods, p = 0.75). High-risk patients (IDF-DAR risk score > 6) showed greater increases in TAR versus low-moderate risk (p = 0.048). AGP demonstrated reduced daytime variability with distinct post-meal excursions at suhoor and iftar.
    CONCLUSIONS: In adults with diabetes and compensated cirrhosis (Child-Pugh A), Ramadan fasting did not worsen glycemic variability, and hypoglycemia risk remained minimal. Fasting appeared safe in well-selected patients, particularly those classified as low-to-moderate IDF-DAR risk, whereas high-risk or decompensated cirrhosis patients should be counseled against fasting.
    Keywords:  Continuous glucose monitoring; Glycemic variability; Liver cirrhosis; Ramadan; Time in range; Type 2 diabetes mellitus
    DOI:  https://doi.org/10.1016/j.diabres.2025.112955
  9. Adv Ther. 2025 Dec 05.
    Evaluating the Cost-Effectiveness of Real-Time Continuous Glucose Monitoring Versus Self-Monitoring of Blood Glucose in the Treatment of Patients with Insulin-Treated Type 2 Diabetes in Australia.
    Hamza Alshannaq, David Simmons, Jessica Y Matuoka, Richard F Pollock, Matthew Tucker, Moin U Ahmed, Greg J Norman.
       INTRODUCTION: Type 2 diabetes (T2D) is a major public health concern in Australia, associated with substantial clinical, humanistic, and economic burden. The condition is linked to high rates of cardiovascular and microvascular complications, premature mortality, and reduced quality of life. Effective glycemic management is central to reducing these adverse outcomes. Real-time continuous glucose monitoring (RT-CGM) has been shown to improve glycemic control in insulin-treated T2D compared with self-monitoring of blood glucose (SMBG). However, evidence of its cost-effectiveness in the Australian setting is limited. This study aimed to evaluate the cost-effectiveness of Dexcom ONE+ RT-CGM versus SMBG in adults with insulin-treated T2D in Australia.
    METHODS: A lifetime economic evaluation was conducted using version 10 of the IQVIA CORE Diabetes Model. The analysis simulated clinical and economic outcomes for two subgroups: those on intensive insulin therapy (IIT) and non-intensive insulin therapy (NIIT). Treatment effects were sourced from clinical trials and real-world evidence. Outcomes included life years, quality-adjusted life years (QALYs), and direct healthcare costs. Incremental cost-effectiveness ratios (ICERs) were calculated as cost per QALY gained. Scenario and sensitivity analyses tested robustness.
    RESULTS: RT-CGM was dominant compared to SMBG in both IIT and NIIT subgroups. In IIT, RT-CGM yielded 0.567 additional QALYs and cost savings of AUD 9869. In NIIT, it yielded 0.319 additional QALYs and savings of AUD 5253. Results were robust across sensitivity analyses. Health equity considerations were also identified, particularly for Indigenous populations and those with youth-onset T2D.
    CONCLUSIONS: RT-CGM was dominant in both insulin-treated subgroups, improving patient outcomes while reducing healthcare costs. These findings highlight the potential value of RT-CGM for broad reimbursement in Australia and the importance of addressing inequities in glycemic management, particularly among Indigenous Australians and younger individuals with T2D.
    Keywords:  Cost-effectiveness; Diabetes complications; Glycemic control; Real-time continuous glucose monitoring; Self-monitoring of blood glucose; Type 2 diabetes
    DOI:  https://doi.org/10.1007/s12325-025-03430-1
  10. Annu Int Conf IEEE Eng Med Biol Soc. 2025 Jul;2025 1-7
    Evaluating the Effect of Input Features on Deep Learning Models for Blood Glucose Forecasting.
    Andrea Calzavara, Francesco Prendin, Giacomo Cappon, Simone Del Favero, Andrea Facchinetti.
      In type 1 diabetes (T1D), predicting future blood glucose (BG) concentration tens of minutes in advance is a key element in decision support systems and artificial pancreas devices. Recently, given the availability of large datasets, deep learning (DL) models for BG forecasting have been investigated, showing promising results. However, the impact of typical input features-such as meal, insulin, and physical activity (PA)- on their performance remains still underexplored. The aim of this study is to assess how these inputs may affect DL models performance.We trained and evaluated five DL models on four weeks of daily-life data from 497 individuals with T1D. Seven input configurations, ranging from univariate continuous glucose monitoring (CGM) models to comprehensive approaches incorporating CGM, insulin, carbohydrate (CHO) intake, heart rate (HR), and exercise data, were evaluated. Results indicate that incorporating additional features progressively enhances performance at the 30-minute prediction horizon (PH), with all models showing similar Root Mean Squared Error (RMSE) and Time Gain (TG). For the CNN-Transformer, the model showing the greatest improvement, the univariate approach achieved an RMSE of 21.02 ± 3.5 mg/dL and a TG of 10.38 ± 1.33 minutes. Incorporating all factors reduced RMSE to 18.63 ± 4.18 mg/dL and increased TG to 12.12 ± 2.64 minutes. Notably, prediction accuracy during exercise improved only when PA data were included, reducing RMSE from 28.72 ± 9.18 mg/dL to 24.7 ± 7.8 mg/dL. While these improvements are statistically significant, their potential clinical benefit remains limited due to the modest magnitude of change.
    DOI:  https://doi.org/10.1109/EMBC58623.2025.11252941
  11. NPJ Metab Health Dis. 2025 Dec 02. 3(1): 46
    A deep learning-derived digital biomarker of dysglycemia and its association with genetic risk of type 2 diabetes.
    Jian Shao, Ying Pan, Jingnan Xue, Haonan Pan, Jing Wang, Shaoyun Li, Zedong Nie, Yuefei Li, Zijian Tian, Yu Zhao, Huyi Feng, Kaixin Zhou.
      Type 2 diabetes is a global health burden driven by genetic and environmental factors. Continuous glucose monitoring (CGM) can effectively guide lifestyle interventions in non-diabetic. However, predefined CGM metrics fail to fully capture the dysglycemic information contained in the high-dimensional time-series CGM data. This study employed deep learning to learn dysglycemia features from CGM data associated with diabetes and derived a digital biomarker of dysglycemia, validated against traditional dysglycemic biomarkers and diabetes polygenic risk score (PRS). Output of the deep learning model, called the deep learning-score, was significantly associated with multiple existing dysglycemic biomarkers and PRS of diabetes (P = 0.007). Moreover, existing CGM metrics were not associated with prevalent diabetes after adjusting for the deep learning-score, while the deep learning-score remained significantly associated with prevalent diabetes (P < 0.001) in a regression analysis. This digital biomarker demonstrated potential for providing dynamic feedback on dysglycemia and improving long-term intervention adherence.
    DOI:  https://doi.org/10.1038/s44324-025-00089-8
  12. JAMA Netw Open. 2025 Dec 01. 8(12): e2541939
    Continuous Glucose Monitoring in Insulin-Treated Older Adults With Diabetes and Alzheimer Disease and Related Dementias.
    Pareeta Kotecha, Steven M Smith, William T Donahoo, Steven T DeKosky, Jiang Bian, Jingchuan Guo.
       Importance: Cognitive impairment and glycemic control have a bidirectional association. Understanding the impact of continuous glucose monitoring (CGM) vs self-monitoring of blood glucose (SMBG) is important for treating older adults with Alzheimer disease and related dementias (ADRD) and diabetes that is treated with insulin.
    Objective: To compare the risks of glycemic outcomes and related adverse events between CGM users and prevalent SMBG users in insulin-treated older adults with ADRD and diabetes.
    Design, Setting, and Participants: This retrospective, prevalent-new user cohort study utilized a nationwide, 15% random sample of Medicare claims data (Parts A, B, and D) from January 2016 to December 2020 to ensure balanced cohorts. Insulin-treated older adults (≥66 years) with ADRD and diabetes were included. Individuals in hospice care or with a professional CGM use were excluded. Analysis was carried out from August 2023 to December 2024.
    Exposure: Therapeutic CGM use vs prevalent SMBG use.
    Main Outcomes and Measures: Primary outcomes included hypoglycemia hospitalizations, hyperglycemia crisis, and all-cause mortality; falls and all-cause hospitalizations were secondary outcomes. Upper respiratory tract infections served as a negative control outcome. A 1:1 propensity score matching with a 0.25 caliper was carried out for confounding control. Cox proportional hazards models were used for outcome analysis.
    Results: In this cohort of 2022 insulin-treated older adults with diabetes and ADRD (1011 CGM users and 1011 SMBG users; mean [SD] age, 76.4 [6.7] years; 1133 female [56.0%]), CGM use was significantly associated with lower risk of all-cause hospitalization (hazard ratio [HR], 0.86; 95% CI, 0.76-0.96) and mortality (HR, 0.57; 95% CI, 0.48-0.67) compared with SMBG use. CGM use had lower point estimates for hypoglycemia hospitalization (HR, 0.66; 95% CI, 0.40-1.08) and falls (HR, 0.86; 95% CI, 0.68-1.08) and higher point estimates for hyperglycemia crisis (HR, 1.38; 95% CI, 0.99-1.94) vs SMBG use, but these findings were not significant. Exploratory subgroup analyses showed varied benefits. The negative control outcome showed no significant differences across analyses.
    Conclusions and Relevance: In this cohort study of insulin-treated older adults with ADRD and diabetes, CGM use was associated with improved long-term clinical outcomes. Pragmatic (ie, evaluating the effectiveness of healthcare interventions in everyday settings) trials are needed to validate these findings and to assess the feasibility of CGM use in this population.
    DOI:  https://doi.org/10.1001/jamanetworkopen.2025.41939
  13. Annu Int Conf IEEE Eng Med Biol Soc. 2025 Jul;2025 1-7
    Predicting Treatment Outcome of Patients with Type 2 Diabetes on Once-Daily Basal Insulin Injections using Machine Learning.
    Ali Mohebbi, Niels-Kristian Kjoller, Alexander R Johansen, Henrik Bengtsson, Morten L Jensen, Bente M Stallknecht, Morten Morup.
      Basal insulin has been and remains a common and cost-effective intensification step from insufficient oral antidiabetic drug (OAD) treatment for people with type 2 diabetes (T2D), but individualized intensification alternatives are rapidly increasing. Recently, decision support tools to assist healthcare professionals based on machine learning (ML) algorithms are becoming more popular. By means of ML, the aim of this pilot study is to explore to what extent patient characteristics and continuous glucose monitoring (CGM) data enhance the ability to predict a successful basal insulin treatment outcome beyond what can be predicted based on hemoglobin A1C (HbA1c) alone at treatment initiation. Clinical data were acquired from four different trials with a total of 222 poorly regulated (HbA1c ≥ 7% ) patients with T2D on OAD initiating basal insulin treatment. HbA1c, patient characteristics, and consensus CGM metrics (based on three days) were available and systematically added as input to three classification models, respectively, based on logistic regression and Gaussian process (GP) classification with linear and both linear and nonlinear kernels. Classification models predicted a binarized HbA1c value after six months as either acceptable (HbA1c < 7%) or suboptimal (HbA1c≥7%) using a repeated stratified cross-validation setup. The consensus metrics based on only three days of CGM show a trend towards slightly improved performance when added on top of HbA1c. However, it appears difficult to accurately predict a binarized HbA1c outcome based on the considered patient information to a satisfactory level for clinical use. Future research should consider the outlined limitations associated with this study and suggested considerations for improvement. However, this pilot study can be considered an initial attempt towards leveraging the potential of ML and CGM data for personalised and cost-effective treatment decision-support for basal insulin initiation.
    DOI:  https://doi.org/10.1109/EMBC58623.2025.11253142
  14. Prim Care Diabetes. 2025 Dec 02. pii: S1751-9918(25)00225-6. [Epub ahead of print]
    Continuous glucose monitoring patterns among pregnant individuals with varying levels of glucose intolerance.
    Rachel D Gordon, Jiqiang Wu, Jennifer Grasch, Neeharika Nandam, Patrick Schneider, Steven Gabbe, Mark B Landon, Maged M Costantine, Kartik K Venkatesh, Elizabeth O Buschur.
      We conducted a prospective observational cohort of pregnant individuals to compare continuous glucose monitoring trends between pregnant individuals with varying levels of glucose intolerance. We found that individuals who had an abnormal value on an oral glucose tolerance test had similar glucose trends to those with gestational diabetes mellitus.
    Keywords:  Continuous glucose monitor; Functional data analysis; Gestational diabetes; Glucose intolerance; Pregnancy
    DOI:  https://doi.org/10.1016/j.pcd.2025.11.006
  15. Endocr Pract. 2025 Nov 27. pii: S1530-891X(25)01294-7. [Epub ahead of print]
    Understanding Glycemia in the Post Discharge Period Through Blinded Continuous Glucose Monitoring.
    Rahul Patel, Mollie Y O'Connor, Amy Sabean, Annabelle Ashley, Hui Zheng, Joyce Yan, Barbara A Steiner, Nillani Anandakugan, Melissa Calverley, Rachel Bartholomew, Evelyn Greaux, Mary Larkin, Steven J Russell, Melissa S Putman, Kristen L Flint.
       OBJECTIVES: Transitions of care at hospital discharge are critical time periods for people with diabetes, but little is known about glycemia in the immediate post-discharge period.
    METHODS: In this observational study, participants wore blinded Dexcom G6 Pro CGM during hospital admission on non-intensive care floors and then continued wearing blinded CGM for up to 10 days post-hospital discharge. Clinical data were extracted from the electronic medical record. Percent time in range 70-180 mg/dl (TIR), above range (TAR), and below range (TBR) from the inpatient and post-discharge periods were calculated. The percentage of participants achieving TIR >50% and >70%, incidence of hypoglycemia after discharge, change in inpatient and post-discharge CGM metrics, and predictors of post-discharge glycemia were determined.
    RESULTS: A total of 24 adults (mean age 65.7 ± 13.6 years, 37.5% female) wore CGM after discharge with mean TIR 43.9 ± 33.2%, mean TAR 55.9 ± 33.3%, and median TBR 0% (0, 0.04). Of these participants, 41.7% had TIR >50%, and 29.2% had TIR >70%. Glycemia pre- and post-discharge was similar (post-discharge vs inpatient TIR -3.7 ± 22.5%, p=0.4). Of the clinical factors assessed, only inpatient glycemia was associated with achieving post-discharge glycemic targets. CGM detected 13 episodes of hypoglycemia occurring in six participants (25%) post-discharge.
    CONCLUSIONS: Glycemia during the post-discharge period is suboptimal, and glucose levels in the hospital may be an important predictor of glycemia after hospitalization. CGM may be useful in identifying hypoglycemia after discharge. Further studies are needed to understand the utility of CGM after hospital discharge.
    Keywords:  continuous glucose monitoring; hospital discharge; hypoglycemia
    DOI:  https://doi.org/10.1016/j.eprac.2025.11.011
  16. Ther Adv Endocrinol Metab. 2025 ;16 20420188251400550
    Hypoglycemic coma owing to sensor device dysfunction after an automatic smartphone operating system update.
    Tasuku Sato, Satomi Hashiguchi, Hiroya Kitsunai, Yumi Takiyama, Hiroshi Nomoto.
      
    Keywords:  continuous glucose monitoring; hypoglycemia; smartphone; type 1 diabetes
    DOI:  https://doi.org/10.1177/20420188251400550
  17. Cureus. 2025 Nov;17(11): e95930
    Luseogliflozin Suppresses Late-Night Basal Glycemic Variability in Type 2 Diabetes: A Retrospective Observational Study.
    Keishi Yamauchi.
       INTRODUCTION:  Glucose fluctuations have been implicated in the development of diabetic macroangiopathy. Sodium-glucose cotransporter 2 (SGLT2) inhibitors are known not only for their glucose-lowering effects but also for their protective effects on the kidneys and heart. Basal glycemic variability (GV) during late-night periods, when external factors such as meals and physical activity are minimized, may represent a clinically important marker of intrinsic glucose regulation. However, little is known about the impact of SGLT2 inhibitors on basal GV. This study investigated the acute (within 24 hours) effect of the first administration of luseogliflozin, an SGLT2 inhibitor, on nocturnal basal GV in patients with type 2 diabetes mellitus.
    PATIENTS AND METHODS:  Ten patients with type 2 diabetes who met the criteria were retrospectively selected. Luseogliflozin (2.5 mg/day) was initiated after baseline assessment. Continuous glucose monitoring (CGM; FreeStyle Libre Pro, Abbott Diabetes Care Inc., Alameda, California) was performed, and the coefficient of variation (CV) of glucose was calculated as an index of glycemic variability (GV). Nocturnal basal GV was evaluated between 11:00 P.M. and 3:00 A.M. on the night before and the night after the first administration. This study was retrospective because CGM data before and after luseogliflozin initiation were obtained from existing hospital records, without prior intent to evaluate this specific effect.
    RESULTS:  CGM revealed heterogeneous nocturnal glucose patterns with small but clinically relevant fluctuations. Luseogliflozin significantly reduced nocturnal glucose CV from 11.9% ± 3.8% at baseline to 7.2% ± 4.1% after treatment (p = 0.0048). These findings indicate that even after the first dose, luseogliflozin may acutely stabilize basal glucose fluctuations during late-night periods.
    CONCLUSION:  This is the first study to demonstrate that an SGLT2 inhibitor can suppress nocturnal basal GV immediately after initiation. Given that GV has been associated with sympathetic activation and vascular injury, these results suggest that reducing basal GV with SGLT2 inhibitors may contribute to vascular protection. Further studies with larger sample sizes and longer follow-up are warranted to confirm these observations.
    Keywords:  coefficient of variation; diabetes; glycemic variability; late-night; sglt2 inhibitor
    DOI:  https://doi.org/10.7759/cureus.95930
  18. Chronic Dis Transl Med. 2025 Dec;11(4): 279-283
    Managing Diabetes One Step at a Time in Low- and Middle-Income Countries: The Promise of Wearable Devices.
    Safayet Jamil, Masoud Mohammadnezhad, Abdulrakib Abdulrahim, Faisal Muhammad, Hafiz T A Khan.
      The global burden of diabetes mellitus disproportionately affects low- and middle-income countries (LMICs), where limited healthcare infrastructure hampers timely and effective disease management. Wearable technologies, such as continuous glucose monitors (CGMs), insulin pumps, and fitness trackers, offer a transformative opportunity to bridge care gaps by enabling real-time monitoring, personalized feedback, and improved glycemic control. Evidence shows how wearables enhance patient engagement, support clinical decision-making, and reduce complications. However, significant barriers such as cost, digital illiteracy, poor system integration, and data privacy concerns impede widespread adoption in LMICs. Case studies from Ghana, China, and Ethiopia illustrate these devices' potential and challenges in resource-limited settings. Policy interventions, such as public-private partnerships, subsidies, simplified interfaces, and digital literacy programs, are essential to overcome these obstacles. Furthermore, integrating wearable data into national health systems and leveraging artificial intelligence can improve individualized care and long-term outcomes. As mobile phone use increases in LMICs, coupling wearables with mHealth platforms could further empower self-management. With targeted investments and regulatory support, wearable technologies can be pivotal in advancing equitable, proactive, and data-driven diabetes care across underserved populations.
    Keywords:  continuous glucose monitoring (CGM); diabetes management; low‐ and middle‐income countries (LMICs); mobile Health (mHealth); wearable technology
    DOI:  https://doi.org/10.1002/cdt3.70018
  19. Annu Int Conf IEEE Eng Med Biol Soc. 2025 Jul;2025 1-6
    GlucoNet: A Hybrid Learning Methodology for Non-Invasive Blood Glucose Estimation from PPG.
    Rahul Verma, Prateek Agrawal, Rashmi Kumari, Surita Sarkar, Pabitra Das, Amit Acharyya.
      The growing prevalence of the diabetic population across the world makes it absolutely necessary to monitor the blood glucose level (BGL) on a regular basis. Although, conventional BGL monitoring methods are widely used for clinical intervention as well as individual monitoring of BGL, those are invasive, painful, and unsuitable for continuous monitoring. These limitations lead to the development of noninvasive alternative methodologies, such as photoplethysmography(PPG) based glucose level estimation. PPG has emerged as a promising method for monitoring BGL, offering several advantages over other optical techniques. In this work, a deep learning architecture, GlucoNet has been introduced that combines Convolutional Neural Network (CNN) and Bidirectional Long Short-Term Memory (BiLSTM) modules for learning both spatial and temporal features from PPG to estimate blood glucose levels from PPG signals. Our model was trained and validated on the publicly available online dataset and real-world dataset to estimate its robustness. Our proposed model outperforms the state of the art methods by achieving a mean absolute error (MAE) of 2.15 mg/dL, root mean squared error (RMSE) of 3.28 mg/dL, and an R2 of 0.99. Furthermore, our model demonstrates exceptional clinical performance, with 100% of predictions falling within the clinically acceptable zones A and B of the Clarke Error Grid analysis.Clinical relevance-This study introduces a non-invasive, real-time methodology for blood glucose level estimation using only PPG signal. This model ensures painless, and continuous monitoring of BGL, and can be used as a reliable solution for early detection and proper management of diabetes and related complications.
    DOI:  https://doi.org/10.1109/EMBC58623.2025.11253359
  20. Annu Int Conf IEEE Eng Med Biol Soc. 2025 Jul;2025 1-4
    Multi-Metric Subgroup Analysis for Glucose Forecasting in Type 1 Diabetes.
    Daphne N Katsarou, Eleni I Georga, Stelios Tigas, Costas Papaloukas, Dimitrios I Fotiadis.
      Type 1 Diabetes (T1D) management remains challenging due to the complexity of glucose regulation, demanding accurate and reliable short-term glucose prediction models. This study investigates the use of population-based and personalized predictive models to enhance glucose prediction in T1D patients. We develop an XGBoost-based model, optimized using Bayesian methods, to predict glucose concentration at 15, 30, and 60-minute prediction horizons. The model was tested across various patient subgroups categorized by glucose patterns, glucose variability (GV), and other clinical factors. Results show that the population-based model consistently outperformed personalized models, achieving RMSE values of 17.39 mg/dL, 27.28 mg/dL, and 40.69 mg/dL at 15, 30, and 60 minutes, respectively. Subgroups with higher GV exhibited poorer prediction accuracy. This study highlights the potential of combining population-based models with subgroup-specific optimizations to improve glycemic control in T1D patients. Accurate glucose prediction is crucial for improving glycemic control and reducing risks, ultimately enhancing patient outcomes.Clinical relevance- This study shows that population-based models can enhance glucose prediction in Τ1D, helping clinicians optimize insulin therapy and improve glycemic control.
    DOI:  https://doi.org/10.1109/EMBC58623.2025.11254917
  21. Diabetes Obes Metab. 2025 Dec 04.
    Commonly used C-peptide assays show differing associations with CGM metrics.
    Catriona Clarke, Mark W J Strachan, Fraser W Gibb.
      
    DOI:  https://doi.org/10.1111/dom.70351
  22. Annu Int Conf IEEE Eng Med Biol Soc. 2025 Jul;2025 1-7
    Analyzing Basal Insulin Dosing as Perturbations in Compositional Glycemic Profiles for Time in Range in Type 1 Diabetes.
    Lyvia Biagi, Alvis Cabrera, Arthur Bertachi, Ernesto Estremera, Josep Antoni Martin-Fernandez, Josep Vehi.
      Basal insulin dosing is essential in multiple daily injection therapy for Type 1 Diabetes, yet adjustments remain challenging. This study models basal insulin changes as perturbations in Compositional Data Analysis (CoDa) to assess their impact on glycemic control. In silico simulations were conducted with a virtual cohort of 49 adults over 180 days, considering controlled meal intake and insulin therapy. Basal insulin adjustments (±5%, ±10%, ±15%) were analyzed using CoDa to assess the effects on time in range (TIR: 70-180 mg/dL) and the glycemic distribution. A multivariate linear regression model predicted outcomes based on basal adjustments. The CoDa framework quantified the impact of basal insulin perturbations on TIR, as shown by multivariate models with high R2 values (up to 0.85). Furthermore, we observed a strong correlation between a CoDa-derived log-ratio coordinate and an established glycemic risk index (r = -0.9141). A visualization panel integrates: (i) arithmetic mean TIR (clinically common), (ii) geometric mean (aligned with CoDa) and (iii) glycemic profile distributions (to describe daily variability). This improvement aids physicians in adjusting therapy.Clinical relevance- This study introduces a decision support system leveraging CoDa analysis to predict glycemic outcomes based on basal insulin adjustments. By modeling basal perturbations, the approach enables disease profiling and personalized treatment strategies, reducing the trial-and-error process in basal insulin titration and optimizing T1D management.
    DOI:  https://doi.org/10.1109/EMBC58623.2025.11253387
  23. Annu Int Conf IEEE Eng Med Biol Soc. 2025 Jul;2025 1-4
    From laboratory to pre-market prototyping: lessons learned from a noninvasive blood glucose level measurement device.
    David Naranjo-Hernandez, Javier Reina-Tosina, Laura M Roa, Gerardo Barbarov-Rostan, Alejandro Talaminos Barroso, Arturo Moreno, Dario Postigo, Jose Manuel Fornes.
      The transition from laboratory prototype to marketable medical device presents significant challenges in biomedical engineering. This paper describes the methodological approach and practical experiences in evolving the Glucube® system, a noninvasive blood glucose monitoring device based on near-infrared spectroscopy (NIRS), from laboratory prototype to pre-industrial product. The process encompassed multiple aspects including user-centered design methodology, risk analysis and management, hardware and firmware adaptations, housing redesign, calibration procedures, and development of supporting software infrastructure. Key challenges addressed include optimization of measurement repeatability, signal quality evaluation, and compliance with medical device regulations. The resulting Class IIb medical device features a modular architecture comprising the measuring device, smartphone application interface, and web services for glucose estimation and data management. The methodology and lessons learned provide valuable insights for accelerating the development cycle of medical devices while maintaining compliance with regulatory requirements.Clinical Relevance- This work addresses the critical need for noninvasive blood glucose monitoring solutions, offering practical insights into developing medical devices that can potentially improve diabetes management adherence by eliminating the pain and inconvenience of traditional finger-prick methods. The described methodology can serve as a reference for developing other medical devices intended for clinical practice.
    DOI:  https://doi.org/10.1109/EMBC58623.2025.11254386
This page is being maintained by Thomas Krichel. It was last updated on 2025‒12‒07 at 0:15.