Continuous Glucose Monitoring Data Research Articles

Observed Glycemic and Psychosocial Benefits in the Prospective Bigfoot Unity Real World Study: A Six-Month Analysis.

The Bigfoot Unity Diabetes Management System integrates Abbott FreeStyle Libre 2 continuous glucose monitoring (CGM) data into a smart pen cap and mobile app enabling clinician-directed insulin dose recommendations and real-time alerts. The objective was to analyze real world 6-month glycemic control in a prospective study for individuals using the System for multiple daily insulin injections (MDI). We conducted a 6-month analysis from the BURST study (NCT05088265) of individuals with type 1 or type 2 diabetes (T2D). Participants reported baseline demographics, adverse event and other surveys electronically. Either at-home kit or electronic medical record glycated hemoglobin A1c (HbA1c) data were collected. Of 102 participants in the per-protocol cohort, median age was 59 years, 87% had T2D, 42% used CGM previously, 62% were White non-Hispanic and 59% female. Mean HbA1c decreased from 9.1 ± 1.7% at baseline to 8.0 ± 1.2% at 6 months (mean difference -1.1%, 95% confidence interval -1.4 to -0.8, P < 0.001). At 6 months, time in range 70-180 mg/dL, time <70 mg/dL and <54 mg/dL were 56 ± 23%, 1.0 ± 1.4% and 0.04 ± 0.14%, respectively. Six severe hypoglycemia events occurred in 4 participants (none system related) and no diabetic ketoacidosis events occurred in the per-protocol cohort. In this study primarily of older adults with T2D using MDI, durable glycemic improvement occurred using the System at 6 months, with the frequency of hypoglycemia being substantially below established targets of <4% and <1% for time below 70 and 54 mg/dL, respectively.

The metabolic and circadian signatures of gestational diabetes in the postpartum period characterised using multiple wearable devices.

Gestational diabetes mellitus (GDM) affects 14% of all pregnancies worldwide and is associated with cardiometabolic risk. We aimed to exploit high-resolution wearable device time-series data to create a fine-grained physiological characterisation of the postpartum GDM state in free-living conditions, including clinical variables, daily glucose dynamics, food and drink consumption, physical activity, sleep patterns and heart rate. In a prospective observational study, we employed continuous glucose monitors (CGMs), a smartphone food diary, triaxial accelerometers and heart rate and heart rate variability monitors over a 2 week period to compare women who had GDM in the previous pregnancy (GDM group) and women who had a pregnancy with normal glucose metabolism (non-GDM group) at 1-2 months after delivery (baseline) and 6 months later (follow-up). We integrated CGM data with ingestion events recorded with the smartphone app MyFoodRepo to quantify the rapidity of returning to preprandial glucose levels after meal consumption. We inferred the properties of the underlying 24 h rhythm in the baseline glucose. Aggregating the baseline and follow-up data in a linear mixed model, we quantified the relationships between glycaemic variables and wearable device-derived markers of circadian timing. Compared with the non-GDM group (n=15), the GDM group (n=22, including five with prediabetes defined based on fasting plasma glucose [5.6-6.9 mmol/l (100-125 mg/dl)] and/or HbA1c [39-47 mmol/mol (5.7-6.4%)]) had a higher BMI, HbA1c and mean amplitude of glycaemic excursion at baseline (all p≤0.05). Integrating CGM data and ingestion events showed that the GDM group had a slower postprandial glucose decrease (p=0.01) despite having a lower proportion of carbohydrate intake, similar mean glucose levels and a reduced amplitude of the underlying glucose 24 h rhythm (p=0.005). Differences in CGM-derived variables persisted when the five women with prediabetes were removed from the comparison. Longitudinal analysis from baseline to follow-up showed a significant increase in fasting plasma glucose across both groups. The CGM-derived metrics showed no differences from baseline to follow-up. Late circadian timing (i.e. sleep midpoint, eating midpoint and peak time of heart rate) was correlated with higher fasting plasma glucose and reduced amplitudes of the underlying glucose 24 h rhythm (all p≤0.05). We reveal GDM-related postpartum differences in glucose variability and 24 h rhythms, even among women clinically considered to be normoglycaemic. Our results provide a rationale for future interventions aimed at improving glucose variability and encouraging earlier daily behavioural patterns to mitigate the long-term cardiometabolic risk of GDM. ClinicalTrials.gov no. NCT04642534.

Abstract 4138815: Prognostic Implications of Glycemic Variability Assessed by Continuous Glucose Monitoring-derived Coefficient of Variation among Patients Hospitalized for Acute Heart Failure

Background: Recent findings indicate that high glycemic variability (GV) could increase the risk of adverse outcomes in patients with cardiovascular disease. With the advances in technology, the utility of continuous glucose monitoring (CGM) has grown during recent years, which can provide valuable information that is not captured by HbA1c. We used the glucose coefficient of variation (CV, calculated as the ratio of standard deviation divided by the mean glucose level) of the obtained CGM data to analyze GV. Aims: We aimed to analyze CGM-derived CV as an indicator of GV and to evaluate prognostic implications of CGM-derived CV in patients hospitalized for acute heart failure (HF). Methods: Patients hospitalized for HF who had not been diagnosed diabetes were performed 75g-oral glucose tolerance test (OGTT). CV was evaluated using CGM system for at least 72 consecutive hours after being able to have meals stably during hospitalization. The primary outcome was a composite of cardiovascular death or re-hospitalization for HF. Patients were divided into two groups based on the median value of CV. Results: We prospectively analyzed 130 patients (mean age: 76±10 years, male 56%) admitted for acute HF. Following the 75g-OGTT, only 21% of the entire cohort exhibited normal glucose tolerance. During the median follow up period of 905 days, Kaplan–Meier curve analysis revealed that the cumulative incidence of the primary outcome event was significantly higher in patients with CV above median than in those with CV less than median (51.7% versus 24.4%, Log-rank P=0.01)(Figure). Multivariate Cox proportional analysis demonstrated that the risk of CV per 1% increment for the primary outcome was significant adjusted for history of diabetes, CGM-derived mean glucose level, and HbA1c level (hazard ratio: 1.06, 95% confidence interval 1.03–1.09). Conclusions: CGM-derived CV was associated with long term prognosis of patients hospitalized for HF with or without diabetes.

Impact of Continuous Glucose Monitoring Versus Blood Glucose Monitoring to Support a Carbohydrate-Restricted Nutrition Intervention in People with Type 2 Diabetes.

Introduction: Low- and very-low-carbohydrate eating patterns, including ketogenic eating, can reduce glycated hemoglobin (HbA1c) in people with type 2 diabetes (T2D). Continuous glucose monitoring (CGM) has also been shown to improve glycemic outcomes, such as time in range (TIR; % time with glucose 70-180 mg/dL), more than blood glucose monitoring (BGM). CGM-guided nutrition interventions are sparse. The primary objective of this study was to compare differences in change in TIR when people with T2D used either CGM or BGM to guide dietary intake and medication management during a medically supervised ketogenic diet program (MSKDP) delivered via continuous remote care. Methods: IGNITE (Impact of Glucose moNitoring and nutrItion on Time in rangE) study participants were randomized to use CGM (n = 81) or BGM (n = 82) as part of a MSKDP. Participants and their care team used CGM and BGM data to support dietary choices and medication management. Glycemia, medication use, ketones, dietary intake, and weight were assessed at baseline (Base), month 1 (M1), and month 3 (M3); differences between arms and timepoints were evaluated. Results: Adults (n = 163) with a mean (standard deviation) T2D duration of 9.7 (7.7) years and HbA1c of 8.1% (1.2%) participated. TIR improved from Base to M3, 61-89% for CGM and 63%-85% for BGM (P < 0.001), with no difference in change between arms (P = 0.26). Additional CGM metrics also improved by M1, and improvements were sustained through M3. HbA1c decreased by ≥1.5% from Base to M3 for both CGM and BGM arms (P < 0.001). Diabetes medications were de-intensified based on change in medication effect scores from Base to M3 (P < 0.001). Total energy and carbohydrate intake decreased (P < 0.001), and participants in both arms lost clinically significant weight (P < 0.001). Conclusion: Both the CGM and BGM arms saw similar and significant improvements in glycemia and other diabetes-related outcomes during this MSKDP. Additional CGM-guided nutrition intervention research is needed.

Do Metrics of Temporal Glycemic Variability Reveal Abnormal Glucose Rates of Change in Type 1 Diabetes?

We aimed to identify the normal range of glucose rates of change (RoC) observed in health and assess whether existing metrics of temporal glycemic variability (GV-timing), such as mean absolute glucose change (MAG) and continuous overlapping net glycemic action (CONGA), are predictive of abnormally rapid RoC in type 1 diabetes (T1D). We identified the normal range of RoC over one-hour intervals from continuous glucose monitoring (CGM) data of healthy individuals. Rapidly rising glucose was defined as RoC values above percentiles 99% (level 1, L1) or 99.9% (level 2, L2), and rapidly falling glucose as below 1% (L1) or 0.1% (L2). The percentage of time these thresholds are exceeded in a given individual is referred to as time in fluctuation (TIF). In a separate CGM dataset of 736 T1D individuals, we calculated TIF-L1 and TIF-L2, and compared them against corresponding values of MAG and CONGA. The extremum percentiles of RoC observed in health are 0.1%: -80 mg/dL/h, 1%: -50 mg/dL, 99%: +56 mg/dL/h, and 99.9%: +89 mg/dL/h. The T1D individuals spend significantly more TIF at rates exceeding these thresholds (TIF-L1: median, 16.7% [interquartile range, 12.7-21.5], TIF-L2: 5.0% [3.1-7.8]) than healthy individuals (TIF-L1: 1.4% [0.6-2.8], TIF-L2: 0.0% [0.0-0.2]). Both MAG and CONGA are highly correlated with TIF-L1 and TIF-L2 (r > .95 in each pairwise comparison). Individuals with T1D spend significant time with glucose RoC exceeding those observed in health. Existing GV-timing metrics are strongly correlated with time with abnormal RoC. Incorporation of a GV-timing metric in clinical practice is recommended.

Identifying the Relationship Between CGM Time in Range and Basal Insulin Adherence in People With Type 2 Diabetes.

The study aimed to determine the relationship between basal insulin adherence and glycemic control evaluated by time in range (TIR) in people with insulin-treated type 2 diabetes (T2D), using data from both continuous glucose monitors (CGM) and connected insulin pens. Furthermore, the study aimed to determine the best basal insulin adherence metric. CGM data and basal insulin data were collected from 106 insulin-treated people (aged ≥18 years) with T2D. Three different adherence metrics were employed (dose deviation, dose deviation ≤20%, and a traditional metric) and a three-step methodology was used to measure insulin adherence level. The coefficient of determination (R2), based on a univariate linear regression analysis, was used to determine the relationship between each adherence metric and TIR. A statistically significant relationship was observed between TIR and adherence quantified as the dose deviation ≤20% metric (R2 = 0.67, P = .006). Neither the relationship between the dose deviation metric and TIR (R2 = 0.43, P = .08) nor the relationship between the traditional metric and TIR (R2 = 0.35, P =.23) was found to be statistically significant. Our study indicates a relationship between basal insulin adherence and TIR in people with insulin-treated T2D. This seems to underscore the role of basal insulin adherence for optimal glycemic outcomes and utilizing TIR as a clinical marker. Furthermore, the results suggest that the magnitude of deviation from the recommended basal insulin dose impacts glycemic control, indicating dose deviation ≤20% as a more accurate metric for quantifying adherence.

Comparison of Computational Statistical Packages for the Analysis of Continuous Glucose Monitoring Data with a Reference Software, "Ambulatory Glucose Profile," in Type 1 Diabetes.

Objective: To compare the accuracy of commonly used continuous glucose monitoring (CGM) analysis programs with ambulatory glucose profile (AGP) and Dexcom Clarity (DC) in analyzing CGM metrics in patients with type 1 diabetes (T1D). Research Methods: CGM data up to 90 days from 152 adults using the same CGM and automated insulin delivery system with T1D were collected. Six of the 19 CGM analysis programs (CDGA, cgmanalysis, Glyculator, iglu, EasyGV, and GLU) were selected to compare with AGP and DC. Metrics were compared etween all tools with two one-sided t-tests equivalence testing. For the equivalence test, the acceptable range of deviation was set as ±2 mg/dL for mean glucose, ±2% for time in range (TIR), ±1% for time above range (TAR), time above range level 1 (TAR1), time above range level 2 (TAR2), and coefficient of variation (CV). Results: All packages were compared with each other for all CGM metrics, and most of them had statistically significant differences for at least some metrics. All tools were equivalent to AGP for mean glucose, TIR, TAR, TAR1, and TAR2 within ±2 mg/dL, ±2%, ±1%, ±1% and 1%, respectively. CDGA, Glyculator, cgmanalysis, and iglu were not equivalent to AGP for CV within ±1%. All tools were equivalent to DC for mean glucose, TIR, and TAR2 within ±2 mg/dL, ±2%, and ±1%, respectively. Glyculator was not equivalent for TAR1, TAR, and CV. CGDA, cgmanalysis, and iglu were not equivalent to DC for TAR1 and TAR. EasyGV and GLU were not equivalent for TAR within ±1%. Conclusions: CGM analysis programs reported CGM metrics statistically differently, but these differences may not be applicable in clinical practice. The equivalence test also confirmed that the differences are negligible for TIR and mean glucose, while they can be important for hyperglycemic ranges and CV. A standardization for CGM data handling and analysis is necessary for clinical studies reporting CGM-generated outcomes.

Effects of switching from MiniMed™ 640G to 770G on continuous glucose monitoring metrics and DTR-QOL scores: An observational study of Japanese people with type 1 diabetes mellitus.

We evaluated the effect of the MiniMed™ 770G, an insulin pump using hybrid closed-loop technology, on blood glucose management and quality of life in Japanese people with type 1 diabetes. This was a 52-week, prospective, observational study. Fifty Japanese people with type 1 diabetes switched from the MiniMed™ 640G to 770G, and we analyzed the continuous glucose monitoring data of 24 subjects who used auto mode throughout the study. We also analyzed the scores of the Diabetes Therapy-Related Quality of Life questionnaire completed by 26 auto-mode users before and after the treatment change. The baseline time in range 70-180 mg/dL was 67.3 (54.8-78.4)%, with a significant improvement beginning 8 weeks after the switch and lasting until 52 weeks. The baseline time below range <70 mg/dL was 1.9 (0.6-3.6)%, with a significant increase at week 8; however, the mean value was less than 4% throughout the study period. On the other hand, the number of blood glucose measurements significantly increased. While there was no significant difference in the overall change in the total Diabetes Therapy-Related Quality of Life score, there was a significant decrease in the treatment satisfaction score. Use of the MiniMed™ 770G improved continuous glucose monitoring metrics. However, treatment satisfaction decreased, probably due to the increased frequency of blood glucose monitoring necessary to maintain auto mode.

Effect of Interrupting Prolonged Sitting with Frequent Activity Breaks on Postprandial Glycemia and Insulin Sensitivity in Adults with Type 1 Diabetes on Continuous Subcutaneous Insulin Infusion Therapy: A Randomized Crossover Pilot Trial.

Objective: This study examined acute effects of interrupting prolonged sitting with short activity breaks on postprandial glucose/insulin responses and estimations of insulin sensitivity in adults with type 1 diabetes (T1D). Method: In a randomized crossover trial, eight adults (age = 46 ± 14 years [mean ± SD], body mass index [BMI] = 27.2 ± 3.8 kg/m2) receiving continuous subcutaneous insulin infusion (CSII) therapy completed two 6-h conditions as follows: uninterrupted sitting (SIT) and sitting interrupted with 3-min bouts of simple resistance activities (SRAs) every 30 min. Basal and bolus insulin were standardized across conditions except in cases of hypoglycemia. Postprandial responses were assessed using incremental area-under-the-curve (iAUC) and total AUC (tAUC) from half-hourly venous sampling. Meal-based insulin sensitivity determined from glucose sensor and insulin pump (SiSP) was assessed from flash continuous glucose monitor and insulin pump data. Outcomes were analyzed using mixed models adjusted for sex, BMI, treatment order, and preprandial values. Results: Glucose iAUC did not differ by condition (SIT: 19.8 ± 3.0 [estimated marginal means ± standard error] vs. SRA: 14.4 ± 3.0 mmol.6 h.L-1; P = 0.086). Despite CSII being standardized between conditions, insulin iAUC was higher in SRA compared to SIT (137.1 ± 22.7 vs. 170.9 ± 22.7 mU.6 h.L-1; P < 0.001). This resulted in a lower glucose response relative to the change in plasma insulin in SRA (tAUCglu/tAUCins: 0.32 ± 0.02 vs. 0.40 ± 0.02 mmol.mU-1; P = 0.03). SiSP was also higher at dinner following the SRA condition, with no between-condition differences at breakfast or lunch. Conclusion: Regularly interrupting prolonged sitting in T1D may increase plasma insulin and improve insulin sensitivity when meals and CSII are standardized. Future studies should explore underlying mechanistic determinants and the applicability of findings to those on multiple daily injections. Trial Registration: Australian and New Zealand Clinical Trial Registry Identifier-ACTRN12618000126213 (www.anzctr.org.au).

The Effect of Effort During a Resistance Exercise Session on Glycemic Control in Individuals Living With Prediabetes or Type 2 Diabetes: Protocol for a Crossover Randomized Controlled Trial.

Type 2 diabetes (T2D) is preceded by prediabetes, and these conditions place a great burden on patients and society. These conditions are significantly associated with poor glycemic control, which is improved by resistance exercise. It has been suggested that resistance exercise should be performed with a high degree of effort to improve glucose metabolism, but this is associated with negative psychological responses that might lead to lower long-term adherence. This study aimed to investigate the effect of the degree of effort during a resistance exercise session on glycemic control and psychological responses in individuals living with prediabetes or T2D. This study will be a crossover, 3-arm, randomized controlled trial. A total of 15 participants living with prediabetes or T2D will be thoroughly familiarized with 7 resistance exercises; afterward, they will perform 3 randomized experimental sessions, each lasting approximately 48 hours each, separated by at least 4 washout days. In 2 of these sessions, supervised resistance exercise will be performed, but the sessions will differ in the degree of effort in each set (high vs low) and will be equalized in terms of total weight lifted and session duration. For this, proximity to failure will be manipulated by changing the number of sets per exercise, the number of repetitions per set, and the resting interval between sets and exercises. Participants will also complete a sedentary (control) session, where they will not perform any exercise. In response to each session, psychological responses will be assessed (exertion, affect, enjoyment, self-efficacy, and discomfort). Glycemic control will be assessed by a continuous glucose monitoring device every 5 minutes, throughout the approximately 48 hours of each experimental session. Food and drink will be individually prescribed by a registered dietitian nutritionist and provided to participants, in order to control for the confounding effect of energy intake and diet composition. Physical activity levels will be assessed by accelerometry. Randomization will be done using the opaque, sequentially numbered envelopes technique. Participants and researchers will be blinded for continuous glucose monitoring and accelerometry data, and data will be analyzed by a blinded statistician. This study has been funded, and data collection is expected to take place between June 2024 and May 2025. Final manuscript submission should happen by August 2025. The results of this project might encourage individuals living with prediabetes and T2D to engage in resistance exercise while better informing exercise specialists on how to best incorporate resistance exercise in their client's or patient's routine. ClinicalTrials.gov NCT06208189; https://clinicaltrials.gov/study/NCT06208189. PRR1-10.2196/63598.

Diabetic Peripheral Neuropathy and Glycemia Risk Index in Type 2 Diabetes: A Cross-Sectional Study.

Diabetic peripheral neuropathy (DPN) is a prevalent chronic complication of diabetes which is linked to chronic hyperglycemia and glycemic variability. This study aimed to investigate the association between the glycemia risk index (GRI) and DPN in patients with type 2 diabetes mellitus (T2DM) using continuous glucose monitoring (CGM) data. From 2019 to 2023, 862 adults diagnosed with T2DM were enrolled at a tertiary care diabetes center in Ningbo, China. The medical history and laboratory parameters were recorded. Neurophysiological examinations were performed to evaluate DPN. The CGM data were recorded for 14 days, and the GRI was calculated based on these data. Multivariate logistic regression analyses were conducted to assess the odds ratio (OR) for DPN with an increased GRI. The prevalence of DPN in the ascending GRI quartiles was 41.6%, 47.9%, 49.1%, and 59.5%, respectively (P for trend < 0.001). In the multivariable logistic analysis, the highest GRI quartile exhibited a 63% greater risk of DPN (OR 1.631, 95% CI: 1.071 to 2.484, P = 0.023) than the lowest quartile after adjusted for age, sex, body mass index, diabetes duration, blood pressure, creatinine, urinary albumin-to-creatinine ratio, lipid profile and glycated hemoglobin. High GRI levels, as measured by CGM, were associated with a greater likelihood of DPN in T2DM patients.

Increased bolus overrides and lower time in range: Insights into disordered eating revealed by insulin pump metrics and continuous glucose monitor data in Australian adolescents with type 1 diabetes

AimsTo determine the prevalence of disordered eating behaviors (DEB) in a population of Australian adolescents with T1D and to investigate clinical parameters, insulin pump therapy (IPT) and continuous glucose monitor (CGM) data trends, and psychological attributes associated with DEB. Methods50 participants (27 female, 23 male, average age 15.2 years, average duration of diabetes 6.2 years) were recruited. Diabetes Eating Problem Survey-Revised (DEPS-R) and Strengths and Difficulties Questionnaires were completed. Prevalence of disordered eating was reported, and associations with clinical parameters, insulin pump therapy (IPT) and continuous glucose monitor (CGM) metrics were assessed. ResultsTwenty-four participants (48 %) had an elevated DEPS-R score. Participants with elevated DEPS-R were more likely to be female (75 % vs 31.6 %, p = 0.004), have a higher HbA1c (8.2 %/67 mmol/mol vs. 6.9 %/51 mmol/mol, p < 0.002) and BMI Z-score (+1.28 SD vs +0.76 SD, p = 0.040). They had lower time in range, 3.9–10 mmol/L (50.3 % vs. 63.8 %, p = 0.01) and higher mean glucose (10.0 mmol/L vs. 8.3 mmol/L, p = 0.005). Of the 60 % using IPT, participants with elevated DEPS-R had increased meal bolus overrides (7.9 % vs 3.8 %, p = 0.047). Reported difficulties on SDQ were higher in the elevated DEPS-R group (18.3 vs 10.5, p < 0.002). ConclusionsDEB are common in Australian adolescents with T1D and associated with increased dysglycemia. Diabetes technology cannot be solely relied upon for detection of DEB and there remains a need for routine screening.

Continuous Glucose Data Construction and Risk Assessment Application of Diabetic Retinopathy Complications for Patients with Type 2 Diabetes Mellitus

Managing diabetes mellitus (DM) includes achieving acceptable blood glucose levels and minimizing the risk of complications from DM. The appropriate glucose sensing method is continuous glucose monitoring (CGM). Effective evaluation metrics that reflect glucose fluctuations can be realized. However, compared with self-monitoring of blood glucose (SMBG), CGM data are not easy to obtain. Therefore, this article studies a fusion model to achieve this objective, including Gaussian process regression (GPR) and long short-term memory (LSTM). Compared with the three commonly used LSTM, GPR, and support vector machine, the proposed model can construct accurate results. By using the constructed CGM data, the conventional metrics, such as the mean amplitude of glycemic excursion (MAGE), mean blood glucose (MBG), standard deviation (SD), and time in range (TIR), are calculated. These metrics and other variables are input into statistical methods to realize diabetic retinopathy risk assessment. In this way, the relationship between the glycemic variability of the constructed CGM data by the mathematical model and DR could be achieved. The utilized statistical methods include single-factor analysis and binary multivariate logistic regression analysis. Results show that fasting blood glucose, disease course, history of hypertension, MAGE and TIR are independent risk factors for DR.

Validating the Use of Continuous Glucose Monitors With Nondiabetic Recreational Runners.

Continuous glucose monitors (CGMs) are becoming increasingly popular among endurance athletes despite unconfirmed accuracy. We assessed the concurrent validity of the FreeStyle Libre 2 worn on 2 different sites at rest, during steady-state running, and postprandial. Thirteen nondiabetic, well-trained recreational runners (age = 40 [8]y, maximal aerobic oxygen consumption = 46.1 [6.4]mL·kg-1·min-1) wore a CGM on the upper arm and chest while treadmill running for 30, 60, and 90minutes at intensities corresponding to 50%, 60%, and 70% of maximal aerobic oxygen consumption, respectively. Glucose was measured by manually scanning CGMs and obtaining a finger-prick capillary blood glucose sample. Mean absolute relative difference, time in range, and continuous glucose Clarke error grid analysis were used to compare paired CGM and blood glucose readings. Across all intensities of steady-state running, we found a mean absolute relative difference of 13.8 (10.9) for the arm and 11.4 (9.0) for the chest. The coefficient of variation exceeded 70%. Approximately 47% of arm and 50% of chest paired glucose measurements had an absolute difference ≤10%. Continuous glucose Clarke error grid analysis indicated 99.8% (arm) and 99.6% (chest) CGM data fell in clinically acceptable zones A and B. Time-in-range analysis showed reduced accuracy at lower glucose levels. However, CGMs accurately detected trends in mean glucose readings over time. CGMs are not valid for point glucose monitoring but appear to be valid for monitoring glucose trends during steady-state exercise. Accuracy is similar for arm and chest. Further research is needed to determine whether CGMs can detect important events such as hypoglycemia during exercise.

The emotional burden of type 1 diabetes: A cross-sectional study to understand associations between diabetes distress and glucose metrics in adulthood.

Advancements in type 1 diabetes (T1D) management, such as continuous glucose monitoring (CGM), have helped people achieve narrower glucose ranges, but associations between CGM and diabetes distress are unclear. Although higher HbA1c is associated with higher distress, associations with other glucose metrics are unknown. To better understand this relationship, we characterized diabetes distress in a sample of CGM users and compared differences in glucose metrics (measured via CGM) between those with higher versus lower distress. CGM users with T1D from the T1D Exchange Registry completed an online survey including diabetes distress (DDS-2) and shared CGM data (N = 199). CGM metrics were computed from all available data within 3 months prior to survey completion. Participants were grouped by distress level: lower (DDS-2 < 3, n = 120) or higher (DDS-2 ≥ 3, n = 79). Welch's t-tests were used to compare mean differences in CGM metrics between groups and MANCOVA was used to further probe mean differences. Approximately 39.7% participants reported higher diabetes distress. Welch's t-tests revealed participants with higher distress spent significantly more time in higher glucose ranges (above 180 mg/dL and above 250 mg/dL), less time in target glucose ranges (between 70 and 180 mg/dL and between 70 and 140 mg/dL) and had higher glucose management index values compared to those with lower distress (p < 0.01). MANCOVA models showed similar results. CGM users continue to experience diabetes distress. Moreover, higher distress appears to be associated with hyperglycaemia. These findings provide support for broader screening efforts for diabetes distress.

Machine Learning Algorithms Based on Time Series Pre-Clustering for Nocturnal Glucose Prediction in People with Type 1 Diabetes.

Background: Machine learning offers new options for glucose prediction and real-time glucose management. The aim of this study was to develop a machine learning-based algorithm that takes into account glucose dynamics patterns for predicting nocturnal glucose in individuals with type 1 diabetes. Methods: To identify glucose patterns, we applied a hierarchical clustering algorithm to real-time continuous glucose monitoring data obtained from 570 adult patients. Machine learning algorithms with or without pre-clustering were used for modeling. Results: Eight clusters without nocturnal hypoglycemia and six clusters with at least one low-glucose episode were identified by the cluster analysis. When forecasting time series without hypoglycemia with a prediction horizon (PH) of 15 or 30 min, gradient boosting trees (GBTs) with pre-clustering and random forest (RF) with pre-clustering outperformed algorithms based on medoids of time series clusters, the Holt model, and GBTs without pre-clustering. When forecasting time series with low-glucose episodes, a model based on the pre-clustering and GBTs provided the highest predictive accuracy at PH = 15 min, and a model based on RF with pre-clustering was the best at PH = 30 min. Conclusions: The results indicate that the clustering of glucose dynamics can enhance the efficacy of machine learning algorithms used for glucose prediction.

Prediction of Incident Diabetic Retinopathy in Adults With Type 1 Diabetes Using Machine Learning Approach: An Exploratory Study.

Early detection and intervention are crucial for preventing vision-threatening diabetic retinopathy (DR) in adults with type 1 diabetes (T1D). This exploratory study uses machine learning on continuous glucose monitoring (CGM) data to identify factors influencing DR and predict high-risk individuals for timely intervention. Between June 2018 and March 2022, adults with T1D with incident DR or no retinopathy (control) were identified. The CGM data were collected retrospectively for up to seven years before the date of defining incident DR or no retinopathy. A mixture of three machine learning algorithms was trained and evaluated in two different scenarios, using different glycemic features extracted from CGM traces (scenario 1), and the two principal components (two PCs; exposure to hyperglycemia and hypoglycemia risk) of those features (scenario 2). Classifiers were evaluated through 10-fold cross-validation using the receiver operating characteristic area under the curve (AUC-ROC) to select the best classification model. The CGM data of 30 adults with incident DR (mean±SD age of 21.2±9.4 years, glycated hemoglobin [HbA1c] of 8.6%±1.0%, and body mass index [BMI] of 24.5±4.8 kg/m2) and 30 adults without DR (age of 41.8±14.7 years, HbA1c of 7.0%±0.9%, and BMI of 26.2±3.6 kg/m2) were included in this analysis. In scenario 2, classifiers outperformed scenario 1, resulting in an average AUC-ROC increase to 0.92 for two of three models, indicating that the two PCs captured vital classification data, representing the most discriminative aspects and enhancing model performance. Machine learning approaches using CGM data may have potential to aid in identifying adults with T1D at risk of DR.

Impact of Hypoglycemia on Glucose Variability over Time for Individuals with Open-Source Automated Insulin Delivery Systems

This study investigates glucose conditions preceding and following various hypoglycemia levels in individuals with type 1 diabetes using open-source automated insulin delivery (AID) systems. It also seeks to evaluate relationships between hypoglycemia and subsequent glycemic variability. Methods: Analysis of continuous glucose monitor (CGM) data from 122 adults with type 1 diabetes using open-source AID from the OpenAPS Data Commons was conducted. This study comprehensively analyzed the effects of hypoglycemia on glycemic variability, covering time periods before and after hypoglycemia. Results: Glucose variability normalization post-hypoglycemia can take up to 48 h, with severe hypoglycemia (41–50 mg/dL) linked to prolonged normalization. A cyclical pattern was observed where hypoglycemia predisposes individuals to further hypoglycemia, even with AID system use. A rise in glucose levels often precedes hypoglycemia, followed by an elevated mean time above range (TAR) post-hypoglycemia, indicating a ‘rebound’ effect. The experimental results are further validated on T1DEXI data (n = 222), originating from commercial AID systems. Different hypoglycemia categorization approaches did not show significant differences in glycemic variability outcomes. The level of hypoglycemia does influence the pattern of subsequent glucose fluctuations. Conclusion: Hypoglycemia, especially at lower levels, significantly impacts subsequent glycemic variability, even with the use of open-source AID systems. This should be studied further with a broader set of commercial AID systems to understand if these patterns are true of all types of AID systems. If these patterns occur in all types of AID systems, it underscores potential opportunities for enhancements in AID algorithms and highlights the importance of educating healthcare providers and people with diabetes about post-hypoglycemia glucose variability.

The Contribution of Postprandial Glucose Levels to Hyperglycemia in Type 2 Diabetes Calculated from Continuous Glucose Monitoring Data: Real World Evidence from the DIALECT-2 Cohort.

Traditional glycemic monitoring in type 2 diabetes is limited, whereas continuous glucose monitoring (CGM) offers better insights into glucose fluctuations. This study aimed to determine the correlations and relative contributions of fasting plasma glucose (FPG) and postprandial plasma glucose (PPG) levels to hyperglycemia. We utilized CGM and recorded carbohydrate intake data from lifestyle diaries of 59 patients enrolled in the Diabetes and Lifestyle Cohort Twente (DIALECT-2). Correlations between FPG and the glucose management indicator (GMI), FPG and Time Above Range (TAR), PPG and GMI, and PPG and TAR were conducted. Daily and mealtime relative contributions of PPG and FPG to glycated hemoglobin (HbA1c) and GMI were determined, considering two ranges: on target (<7.0%, 53 mmol/mol) and not on target (≥7.0%, 53 mmol/mol). Correlations between mealtime PPG and carbohydrate consumption were examined. FPG and PPG correlated with GMI (r = 0.82 and 0.41, respectively, p < 0.05). The relative contribution of PPG in patients with HbA1c, GMI, and TAR values not on target was lower than in patients with HbA1c, GMI, and TAR values on target. When analyzing different mealtimes, patients with target GMI values had a higher PPG (73 ± 21%) than FPG after breakfast (27 ± 21%, p < 0.001). Individuals with elevated GMI levels had lower PPG after lunch (30 ± 20%), dinner (36 ± 23%), and snacks (34 ± 23%) than FPG. PPG after breakfast positively correlated (r = 0.41, p < 0.01) with breakfast carbohydrate intake. Both PPG and FPG contribute to hyperglycemia, with PPG playing a larger role in patients with better glycemic control, especially after breakfast. Targeting PPG may be crucial for optimizing glucose management.

Predicting the Risk of Adverse Neonatal Outcomes in Women With Insulin-Treated Diabetes: Is It Time for a Pregnancy-Specific Glycemic Risk Index (GRI)?

The Glycemia Risk Index (GRI) describes the quality of glycemic control, emphasizing extreme hypoglycemia and hyperglycemia more than less extreme values. However, a pregnancy-specific GRI (pGRI), tailored to the tighter target glucose range required during pregnancy, has not been established. We retrospectively evaluated clinical, metabolic, and Continuous Glucose Monitoring (CGM) data across pregnancy in women with insulin-treated diabetes, managed between September 2021 and March 2024 at the University Hospital of Pisa. First and second levels of hyperglycemia (TAR1: 140-180 mg/dL, TAR2: >180 mg/dL) and hypoglycemia (TBR1: 63-54 mg/dL, TBR2: <54 mg/dL) were used to calculate the pGRI at each trimester. Logistic regression analysis investigated the association between pGRI and risk of at least one adverse neonatal outcome (among preterm delivery, macrosomia, large for gestational age, small for gestational age, neonatal hypoglycemia, neonatal jaundice, and neonatal intensive care unit admission). Of 45 pregnant women, 25 (56%) experienced at least one adverse neonatal outcome. In the third trimester, women with adverse outcomes had significantly higher total TAR (26 [12-32]% vs 10 [4-23]%, P = .018) and lower TIR (71 [64-83]% vs 88 [75-92]%, P = .007). Specifically, the difference was notable in TAR2 (6 [2-15]% vs 1 [0-4]%, P = .004), whereas TAR1 was comparable between the 2 groups. Accordingly, third trimester pGRI was higher in women with adverse neonatal outcomes (38 [18-49]% vs 18 [10-31]%, P = .013) and, at logistic regression, slightly but significantly increased the risk of adverse neonatal outcomes (1.044 [1.004-1.086], P = .024). Pregnant women with insulin-treated diabetes reporting adverse neonatal outcomes spent more time in hyperglycemia, particularly in extreme hyperglycemia. Therefore, the level of hyperglycemia should always be assessed during pregnancy. The pGRI, emphasizing extreme hyperglycemia, may be a novel comprehensive tool for assessing the risk of adverse neonatal outcomes.