bims-glumda Biomed News
on CGM data in management of diabetes
Issue of 2025–06–22
thirteen papers selected by
Mott Given



  1. Ugeskr Laeger. 2025 Jun 16. pii: V12240902. [Epub ahead of print]187(25):
      This case report highlights the importance of recognizing potential pitfalls in HbA1c measurement errors and being vigilant in terms of discrepancies between HbA1c estimated from continuous glucose monitoring (CGM) and venous HbA1c measurements. The latter is especially relevant for physicians managing diabetes, as by 2025, 4,400 adults with insulin-treated type 2 diabetes are expected to start using CGM devices, many of whom will be monitored in primary care settings.
    DOI:  https://doi.org/10.61409/V12240902
  2. JMIR Diabetes. 2025 Jun 18. 10 e69061
       Background: Continuous glucose monitoring (CGM) is used to assess glycemic trends and guide therapeutic changes for people with diabetes. We aimed to increase patient access to this tool by equipping primary care physicians (PCPs) to accurately interpret and integrate CGM into their practice via a multidisciplinary team approach.
    Objective: The primary objective of this study was to evaluate the feasibility and effectiveness of integrating CGM into primary care clinics using a multidisciplinary approach that included a clinical pharmacist (PharmD) and a certified diabetes care and education specialist (CDCES).
    Methods: Eighteen PCPs received a 1-hour video training module from an endocrinologist teaching a systematic stepwise approach to CGM interpretation. Patient inclusion criteria included type 2 diabetes mellitus, ≥18 years old, hemoglobin A1c (HbA1c) ≥8% or concern for hypoglycemia, and no previous CGM use or an endocrinology visit in the past year. Patients saw physician extenders (CDCES or a PharmD) for professional CGM placement and education on nutrition, medication administration, and physical activity goals based on the PCP's recommendations. The CDCES or PharmD reviewed CGM data with patients and collaborated with PCPs to adjust the care plan, informed by the systematic stepwise approach to CGM interpretation. Patients either converted to personal CGM if desired or had a second professional CGM device placed after ≥1 month from the initial professional CGM placement and obtained a postintervention HbA1c measurement at ≥3 months from the initial HbA1c measurement. The primary outcomes were time in range, HbA1c, and average time from referral to the first CGM device placement. Follow-up continued with the CDCES or PharmD until patients met the study discharge criteria of HbA1c level ≤7%. Paired t tests with 1-sided P values were used to assess changes in glucose metrics from the initial to postintervention measurements. The McNemar test was used to determine the significance of change in patients meeting the goal of ≥70% time in the target range of 70-180 mg/dL.
    Results: The CGM users (n=46) had a mean (SD) age of 62.39 (14.57) years, and 14/46 participants (30%) were female. The mean (SD) time in range increased by 28.06%, from 43.25% (33.41%) at baseline to 71.31% (25.49%) postintervention (P<.001), due to reduced hyperglycemia. The proportion of CGM users meeting the consensus target of the time in range ≥70% increased from 23.81% to 57.14% (P<.001). Postintervention HbA1c decreased by an average of 2.37%, from 9.68% (1.78%) to 7.31% (1.32%; P<.001).
    Conclusions: The integration of CGM into primary care clinics to increase patient access is feasible and effective using a multidisciplinary approach.
    Keywords:  continuous glucose monitoring; diabetes education; multidisciplinary team; primary care physicians; type 2 diabetes
    DOI:  https://doi.org/10.2196/69061
  3. Diabetes Metab Syndr Obes. 2025 ;18 1939-1948
       Background: While lifestyle modification remains fundamental for prediabetes management, the potential added value of real-time continuous glucose monitoring (RT-CGM) in diabetes health education programs warrants investigation. This study evaluated whether an individualized diabetes health education program using RT-CGM could improve glycemic control compared to general dietary guidance in prediabetic individuals.
    Methods: In this randomized controlled trial conducted at Guangdong Provincial People's Hospital (initiated September 2022), we enrolled 41 adults (>18 years) with prediabetes, randomly assigning them to either: (1) RT-CGM group (n=20) receiving meal adjustments based on continuous glucose data and energy balance, or (2) control group (n=21) receiving adjustments based solely on energy balance. The study comprised two intensive 14-day education sessions (baseline and 1-year follow-up) with metabolic assessments (HbA1c, fasting blood glucose, BMI, lipid profile, and uric acid) conducted at baseline, 1-year, and 2-year timepoints.
    Results: The RT-CGM group demonstrated significantly greater improvements in HbA1c compared to controls at both 1-year (p=0.007) and 2-year (p=0.033) follow-ups.
    Conclusion: Our findings suggest that incorporating RT-CGM into diabetes health education program can enhance glycemic control in prediabetic individuals compared to general dietary guidance alone. These results support the potential clinical utility of RT-CGM in prediabetes management strategies.
    Keywords:  continuous glucose monitoring; diabetes self-management education; glycosylated hemoglobin; prediabetes
    DOI:  https://doi.org/10.2147/DMSO.S511187
  4. Diabetologia. 2025 Jun 20.
       AIMS/HYPOTHESIS: Glucose variability in people with type 2 diabetes has been associated with increased risk of CVD, and AGEs might be an underlying mechanism. Therefore, this study investigates associations of glucose variability with AGEs in the skin in people with and without impaired fasting glucose, impaired glucose tolerance or diabetes.
    METHODS: We used data from the Maastricht Study, a population-based cohort study. Glucose variability and AGEs in skin were measured by continuous glucose monitoring (CGM) and skin autofluorescence (SAF), respectively. Multiple linear regression was used to test the association of CGM-metrics CV and SD with SAF and adjusted for age, sex, CVD risk factors, nutritional factors and educational level. Interaction analysis was used to test the effect of glucose metabolism status on the association of CV and SD with SAF.
    RESULTS: We included 795 participants (mean ± SD age 59 ± 8.7 years; 49% were female). Glucose metabolism status was stratified into normal glucose metabolism (n = 459), prediabetes (n = 174) and type 2 diabetes (n = 162). Individuals with type 2 diabetes had higher values of SAF (mean ± SD 2.3 ± 0.6 arbitrary units [AU]) than those with prediabetes (2.1 ± 0.4 AU, p = 0.014) and normal glucose metabolism (2.0 ± 0.4 AU, p = 0.007). In the cohort, both SD (0.152 AU [IQR 0.088-0.217]) and CV (0.014 AU [IQR 0.005-0.017]) were significantly associated with SAF in fully adjusted analyses. Glucose metabolism status did not modify the associations of SD and CV with SAF.
    CONCLUSIONS/INTERPRETATION: A higher glucose variability is associated with higher levels of SAF, suggesting that glucose variability plays a role in the formation of AGEs.
    Keywords:  AGE reader; Cardiovascular disease; Coefficient of variation metric; Diabetes; Skin advanced glycation end-products ; Standard deviation metric
    DOI:  https://doi.org/10.1007/s00125-025-06469-5
  5. Comput Methods Programs Biomed. 2025 Jun 09. pii: S0169-2607(25)00295-0. [Epub ahead of print]269 108878
       BACKGROUND AND OBJECTIVE: Diabetes is a global health concern, affecting millions of adults worldwide and exhibiting a growing prevalence. Managing the disease highly relies on continuous glucose monitoring, yet the dense and complex nature of electronic devices data streams poses significant challenges for efficient interpretation. Large Language Models are being widely applied across different domains for their ability to generate human-like text, but still fall short in producing accurate and meaningful text from raw data. To address this limitation, this study proposes a fine-tuning methodology tailored specifically to glucose data, but scalable to other expert-guided domains, enabling the models to generate concise, relevant and safe summaries, bridging the gap between raw data and efficient medical attention.
    METHODS: This study introduces a novel continuous glucose monitoring framework that involves fine-tuned GPT models using structured datasets generated through an expert-guided data modeling based on Fuzzy Logic and prompt engineering for task contextualization. A new evaluation methodology is defined to assess the performance of the Large Language Models across different critical domains where expert knowledge is fundamental to characterize temporally dependent data and ensure valuable insights.
    RESULTS: Fine-tuned GPT-4o achieved the highest performance, with an average score of 96% across all metrics. GPT-4o-mini followed with 76% score, while GPT-3.5 scored 72%. The use of fuzzy knowledge-based prompts proved more effective in scenarios with full data availability, or in scenarios with a simplified data availability when the models are not fine-tuned; domain-guided prompts improved output relevance and stability in fine-tuned models with less data availability.
    CONCLUSIONS: These results indicate the capability of our methods to align Large Language Models with the task of generating human-like text from raw data, highlighting their potential to manage diabetes by complex glucose patterns interpretation, alleviating the burden on healthcare systems.
    Keywords:  Continuous glucose monitoring; Evaluation methodology; Fine-tuning process; IoT-data fuzzy summarization; Large language models; Prompt engineering
    DOI:  https://doi.org/10.1016/j.cmpb.2025.108878
  6. Fed Pract. 2025 Feb;42(2): 1-6
       Background: Idiopathic postprandial syndrome (IPP) presents with hypoglycemic-like symptoms in the absence of biochemical hypoglycemia and remains a diagnosis of exclusion. Its pathophysiology is poorly understood. The diagnosis requires thorough evaluation and the Whipple triad criteria. Treatment typically involves dietary modifications, including reduced carbohydrate intake, increased protein and fiber, and frequent small meals. Continuous glucose monitoring (CGM) may be a useful adjunct in correlating symptoms with glucose trends, but its role is still evolving.
    Case Presentation: A 41-year-old male veteran presented with chronic postprandial episodes characterized by lightheadedness, nausea, tremulousness, anxiety, and other adrenergic symptoms occurring after carbohydrate-heavy meals. An extensive workup was unremarkable. CGM confirmed normoglycemia during episodes, ruling out true hypoglycemia and supporting a diagnosis of idiopathic postprandial syndrome. He was referred to a nutritionist for guidance on a high-protein, high-fiber, low-carbohydrate diet and subsequently reported symptomatic improvement.
    Conclusions: This case highlights the importance of recognizing IPP as a distinct clinical entity, especially due to its nonspecific clinical presentation. Early identification allows for a more accurate diagnosis and targeted treatment through tailored dietary and behavioral strategies, helping to alleviate symptoms.
    DOI:  https://doi.org/10.12788/fp.0541
  7. Diabetes Obes Metab. 2025 Jun 16.
    Hyogo Diabetes Hypoglycemia Cognition Complications (HDHCC) Study Group
       AIMS: Type 2 diabetes mellitus (T2DM) is known to be a risk factor for cognitive dysfunction and dementia. Time in range (TIR), which is derived from continuous glucose monitoring (CGM), has been widely used as an indicator of the quality of glycemic control. While cross-sectional studies have reported an association between CGM-derived TIR and cognitive function scores, few studies have longitudinally investigated the relationship between the two. This study aimed to prospectively investigate the association between CGM-derived TIR and changes in multiple cognitive function scores.
    MATERIALS AND METHODS: The present study used baseline and 2-year data from an ongoing multicenter cohort study. This study included 197 T2DM patients aged ≥60 years with undiagnosed dementia. Participants were examined with the mini-mental state examination (MMSE), the Japanese version of the Montreal cognitive assessment (MoCA-J) and the digit symbol substitution test (DSST) at both baseline and 2 years. Multiple regression analyses were performed to investigate the association between TIR and changes in cognitive function test scores over 2 years.
    RESULTS: Multivariate regression analysis showed that there was a significant association between TIR and changes in MMSE (ΔMMSE) over 2 years (standard partial regression coefficient [β] = 0.187, p = 0.005). Similarly, multivariate regression models showed a significant association between TIR and ΔMoCA-J (β = 0.218, p = 0.001) and ΔDSST (β = 0.164, p = 0.036).
    CONCLUSIONS: In patients with T2DM with undiagnosed dementia, CGM-derived TIR might be associated with overall cognitive decline and reduced processing speed.
    Keywords:  Montreal cognitive assessment; cognitive decline; continuous glucose monitoring; digit symbol substitution test; mini‐mental state examination; time in range; type 2 diabetes
    DOI:  https://doi.org/10.1111/dom.16511
  8. Placenta. 2025 Jun 09. pii: S0143-4004(25)00213-9. [Epub ahead of print]168 46-55
       INTRODUCTION: Gestational diabetes mellitus (GDM) increases the risk of pathological fetal growth, including rates of large-for-gestational age (LGA) infants, which in turn increases the risk of offspring later developing cardiometabolic complications. Recent continuous glucose monitoring (CGM) studies have revealed that temporal periods of mild hyperglycaemia are linked to LGA, and too tight glycaemic control can increase periods of maternal hypoglycaemia and increase the risk of delivering small-for-gestational age (SGA) infants. The underlying mechanisms are unclear but likely involve the placenta.
    METHODS: Ex vivo human placental explants from term uncomplicated pregnancies were cultured in varying glucose concentrations for 48 h to recapitulate in vivo maternal glucose profiles. Glucose, osmolality, human chorionic gonadotrophin (hCG) and lactate dehydrogenase (LDH) were measured in conditioned medium, and RNA sequencing performed, followed by functional enrichment analysis (FEA).
    RESULTS: Medium changes every 6-18 h in variable (5/5.5 mM), or constant 5 mM or 7 mM glucose were appropriate to model maternal normoglycaemia, periods of mild hypoglycaemia and periods of mild hyperglycaemia, respectively. There were 61 differentially expressed genes (DEGs) in explants cultured in mild hyperglycaemic conditions and 54 DEGs in mild hypoglycaemic conditions. FEA revealed that transcripts altered by mild hyperglycaemia were associated with vascular development and lipid metabolism/homeostasis, whilst those altered by mild hypoglycaemia were associated with cell turnover.
    CONCLUSIONS: Together this data demonstrates that subtle changes in maternal glucose impact the placenta and may contribute to altered fetal growth. This highlights the importance of employing CGM in pregnancies complicated by GDM and utilising physiological glucose levels in ex vivo/in vitro placental studies.
    Keywords:  Continuous glucose monitoring; Diabetes; Glucose; Placenta; Pregnancy; Transcriptome
    DOI:  https://doi.org/10.1016/j.placenta.2025.06.006
  9. Ther Adv Endocrinol Metab. 2025 ;16 20420188251346260
      Iatrogenic hypoglycaemia remains a major barrier to optimal glycaemic control required to prevent long-term complications in people with type 1 diabetes (pwT1D). Hypoglycaemia is the consequence of the interaction between absolute or relative insulin excess from treatment and compromised physiological defences against falling plasma glucose. With a longer duration of diabetes and repeated exposure to hypoglycaemia, pwT1D can develop impaired awareness of hypoglycaemia (IAH). IAH increases the risk of severe hypoglycaemia six-fold, causing significant morbidity, and, if left untreated, death. Over the last few decades, a stepwise change in diabetes management has been the introduction and widespread uptake of novel technologies, including continuous glucose monitoring (CGM) and automated insulin delivery (AID) systems. These technologies aim to improve glycaemic control whilst minimising hypoglycaemia. Alarms and safety functions, such as suspension of insulin delivery, can help to reduce the hypoglycaemia burden. This review examines the role of continuous glucose monitors and AID systems in managing IAH, exploring evidence for their impact on symptomatic awareness and identifying areas for future research. In conclusion, there is strong evidence that CGM and AID systems improve glycaemic control and reduce the hypoglycaemia burden. However, despite the use of these technologies, severe hypoglycaemic episodes are not entirely eliminated, and it remains unclear whether their implementation restores the physiological symptoms and counter-regulatory response to hypoglycaemia.
    Keywords:  automated insulin delivery; continuous glucose monitoring; hybrid closed loop; impaired awareness of hypoglycaemia (IAH); severe hypoglycaemia
    DOI:  https://doi.org/10.1177/20420188251346260
  10. Cureus. 2025 May;17(5): e84177
      Digital diabetes management technologies (DDMTs) have emerged as promising tools for improving glycemic control in patients with type 2 diabetes mellitus (T2DM) receiving home-based care. This systematic review evaluates the effectiveness of various DDMTs, including mobile health applications, continuous glucose monitoring (CGM), telemedicine, smart insulin pens, and artificial intelligence-driven decision support systems, in optimizing blood glucose levels. A comprehensive literature search across PubMed, Embase, Scopus, Web of Science, and the Cochrane Library identified nine high-quality systematic reviews published between 2020 and 2024. These reviews synthesized evidence from randomized controlled trials (RCTs) and observational studies, with sample sizes ranging from small pilot studies to large-scale trials. The findings indicate that DDMTs significantly improve HbA1c levels, fasting blood glucose, and postprandial glucose compared to standard self-care practices. Mobile applications and CGM systems demonstrated notable reductions in HbA1c, while telemedicine interventions enhanced patient adherence and engagement. Personalized coaching and real-time feedback were key factors in intervention success. However, challenges such as digital health literacy, cost barriers, and long-term adherence remain concerns. Some studies highlighted the need for sustained engagement to maintain long-term benefits. While DDMTs offer a viable alternative to traditional diabetes management, future research should focus on standardizing interventions, addressing accessibility issues, and evaluating their cost-effectiveness. This review contributes to the growing evidence supporting DDMTs in T2DM management and underscores the potential of digital health innovations in improving glycemic outcomes and patient self-care in home settings.
    Keywords:  diabetes; digital technology; mhealth; self-management; telemedicine; type 2 diabetes
    DOI:  https://doi.org/10.7759/cureus.84177
  11. Biosens Bioelectron. 2025 Jun 07. pii: S0956-5663(25)00551-2. [Epub ahead of print]287 117677
      Non-invasive, real-time, and continuous monitoring of trace amounts of glucose in near-neutral biofluids is significant for the daily care and treatment of diabetic patients or people with suboptimal health status. Despite improved sensing performance with novel low-dimensional materials or porous structures in various enzymatic and non-enzymatic electrochemical glucose sensors, they still suffer from high cost, poor long-term stability, and performance fluctuations in varied temperature and pH. This work synergistically combines an Au-modified porous laser-induced graphene (LIG) gate electrode with an organic electrochemical transistor (OECT) to create a flexible non-enzymatic glucose sensor. The resulting OECT-based non-enzymatic glucose sensor exhibits significantly enhanced sensitivity in near-neutral biofluids, the limit of detection (LOD) (0.08 μM in pH = 7.4), excellent stability over time (degradation of ∼10 % in 180 days) and against temperature changes (30 °C-40 °C), self-pH calibration capabilities, and uncompromised sensing performance with shrinking sizes. The highly consistent laser patterning technique and in situ galvanic reduction process for electrode modifications not only provide a simple yet versatile approach to creating low-cost, compact sensing platforms for precise and real-time sweat glucose measurements but also support scalable production, allowing the correlation study of key biomarkers in sweat and blood.
    Keywords:  Laser-induced graphene nanocomposite; Long-term stability; Non-enzymatic glucose sensor; Organic electrochemical transistor; Self-pH calibration
    DOI:  https://doi.org/10.1016/j.bios.2025.117677