Glucose Variability: A Review of Clinical Applications and Research Developments

Ageing well with diabetes: the role of technology

Over the past two decades there has been a substantial rise in the adoption of diabetes therapeutic technology among children, adolescents and younger adults with type 1 diabetes, and its use is now also advocated for older individuals. Older people with diabetes are more prone to experience hypoglycaemia because of numerous predisposing factors and are at higher risk of hypoglycaemic events requiring third-party assistance as well as other adverse sequelae. Hypoglycaemia may also have long-term consequences, including cognitive impairment, frailty and disability. Diabetes in older people is often characterised by marked glucose variability related to age-associated changes such as variable appetite and levels of physical activity, comorbidities and polypharmacotherapy. Preventing hypoglycaemia and mitigating glucose excursions may have considerable positive impacts on physical and cognitive function and general well-being and may even prevent or improve frailty. Technology for older people includes continuous glucose monitoring systems, insulin pumps, automated insulin delivery systems and smart insulin pens. Clinical trials and real-world studies have shown that older people with diabetes benefit from technology in terms of glucose management, reductions in hypoglycaemic events, emergency department attendance and hospital admissions, and improvement in quality of life. However, ageing may bring physical impairments and other challenges that hinder the use of technology. Healthcare professionals should identify older adults with diabetes who may benefit from therapeutic technology and then adopt an individualised approach to education and follow-up for individuals and their caregivers. Future research should explore the impact of diabetes technology on outcomes relevant to older people with diabetes.

Optimizing postprandial glucose prediction through integration of diet and exercise: Leveraging transfer learning with imbalanced patient data

BACKGROUND

In recent years, numerous methods have been introduced to predict glucose levels using machine-learning techniques on patients' daily behavioral and continuous glucose data. Nevertheless, a definitive consensus remains elusive regarding modeling the combined effects of diet and exercise for optimal glucose prediction. A notable challenge is the propensity for observational patient datasets from uncontrolled environments to overfit due to skewed feature distributions of target behaviors; for instance, diabetic patients seldom engage in high-intensity exercise post-meal.

METHODS

In this study, we introduce a unique application of Bayesian transfer learning for postprandial glucose prediction using randomized controlled trial (RCT) data. The data comprises a time series of three key variables: continuous glucose levels, exercise expenditure, and carbohydrate intake. For building the optimal model to predict postprandial glucose levels we initially gathered balanced training data from RCTs on healthy participants by randomizing behavioral conditions. Subsequently, we pretrained the model's parameter distribution using RCT data from the healthy cohort. This pretrained distribution was then adjusted, transferred, and utilized to determine the model parameters for each patient.

RESULTS

The efficacy of the proposed method was appraised using data from 68 gestational diabetes mellitus (GDM) patients in uncontrolled settings. The evaluation underscored the enhanced performance attained through our method. Furthermore, when modeling the joint impact of diet and exercise, the synergetic model proved more precise than its additive counterpart.

CONCLUSION

An innovative application of the transfer-learning utilizing randomized controlled trial data can improve the challenging modeling task of postprandial glucose prediction for GDM patients, integrating both dietary and exercise behaviors. For more accurate prediction, future research should focus on incorporating the long-term effects of exercise and other glycemic-related factors such as stress, sleep.

Current landscape and comprehensive management of glycemic variability in diabetic retinopathy

Diabetic retinopathy (DR), a well-known microvascular complication of diabetes mellitus, remains the main cause of vision loss in working-age adults worldwide. Up to now, there is a shortage of information in the study regarding the contributing factors of DR in diabetes. Accumulating evidence has identified glycemic variability (GV), referred to fluctuations of blood glucose levels, as a risk factor for diabetes-related complications. Recent reports demonstrate that GV plays an important role in accounting for the susceptibility to DR development. However, its exact role in the pathogenesis of DR is still not fully understood. In this review, we highlight the current landscape and relevant mechanisms of GV in DR, as well as address the mechanism-based therapeutic strategies, aiming at better improving the quality of DR management in clinical practice.

Orexin neurons track temporal features of blood glucose in behaving mice

Does the brain track how fast our blood glucose is changing? Knowing such a rate of change would enable the prediction of an upcoming state and a timelier response to this new state. Hypothalamic arousal-orchestrating hypocretin/orexin neurons (HONs) have been proposed to be glucose sensors, yet whether they track glucose concentration (proportional tracking) or rate of change (derivative tracking) is unknown. Using simultaneous recordings of HONs and blood glucose in behaving male mice, we found that maximal HON responses occur in considerable temporal anticipation (minutes) of glucose peaks due to derivative tracking. Analysis of >900 individual HONs revealed glucose tracking in most HONs (98%), with derivative and proportional trackers working in parallel, and many (65%) HONs multiplexed glucose and locomotion information. Finally, we found that HON activity is important for glucose-evoked locomotor suppression. These findings reveal a temporal dimension of brain glucose sensing and link neurobiological and algorithmic views of blood glucose perception in the brain's arousal orchestrators.

Continuous Glucose Monitoring of Steroid-Induced Hyperglycemia in Patients With Dermatologic Diseases

BACKGROUND AND OBJECTIVES

Systemic administration of glucocorticoids is a mainstay therapy for various inflammatory diseases and may lead to hyperglycemia, which carries the risk of worsening preexisting diabetes and triggering steroid-induced diabetes. Therefore, we aimed to identify patients at risk and to quantify severity of steroid-induced hyperglycemia (SIH) by continuous glucose monitoring (CGM) in hospitalized patients needing systemic glucocorticoid treatment.

PATIENTS AND METHODS

This prospective study included 51 steroid-naive, dermatological patients requiring systemic high-dose glucocorticoid treatment at the Department of Dermatology of the University Hospital Essen. After careful diabetes-specific assessment at admission, glucose monitoring was performed using a CGM system and glucose profile was analyzed in patients with and without SIH.

RESULTS

SIH occurred in 47.1% of all treated patients, and a relevant part of patients with initial normoglycemia developed SIH (2/10 patients). Doubling of SIH incidence was observed with each severity grade of dysglycemia (4/10 in prediabetes; 9/10 in diabetes). Patients with SIH spend nearly 6 hours daily above targeted glucose range, and severe hyperglycemia was observed for 1.2 hours/day.

CONCLUSIONS

Our study underlines the need for dedicated glucose monitoring in dermatologic patients on systemic glucocorticoid therapy by demonstrating its impact on glucose metabolism.

Association Between Individualized Versus Conventional Blood Glucose Thresholds and Acute Kidney Injury After Cardiac Surgery: A Retrospective Observational Study

OBJECTIVES

This study was designed to compare individualized and conventional hyperglycemic thresholds for the risk of acute kidney injury (AKI) after cardiac surgery.

DESIGN

This was an observational study.

SETTING

The study took place in a single-center tertiary teaching hospital.

PARTICIPANTS

Adult patients who underwent cardiac surgery between January 2012 and November 2021 were enrolled.

MEASUREMENTS AND MAIN RESULTS

Two blood glucose thresholds were used to define intraoperative hyperglycemia. While the conventional hyperglycemic threshold (CHT) was 180 mg/dL in all patients, the individualized hyperglycemic threshold (IHT) was calculated based on the preoperative hemoglobin A1c level. Various metrics of intraoperative hyperglycemia were calculated using both thresholds: any hyperglycemic episode, duration of hyperglycemia, and area above the thresholds. Postoperative AKI associations were compared using receiver operating characteristic curves and logistic regression analysis. Among the 2,427 patients analyzed, 823 (33.9%) developed AKI. The C-statistics of IHT-defined metrics (0.58-0.59) were significantly higher than those of the CHT-defined metrics (all C-statistics, 0.54; all p < 0.001). The duration of hyperglycemia (adjusted odds ratio, 1.09; 95% confidence interval, 1.02-1.16) and area above the IHT (1.003; 1.001-1.004) were significantly associated with the risk of AKI, except for the presence of any hyperglycemic episode. None of the CHT-defined metrics were significantly associated with the risk of AKI.

CONCLUSIONS

Individually defined intraoperative hyperglycemia better predicted postcardiac surgery AKI than universally defined hyperglycemia. Intraoperative hyperglycemia was significantly associated with the risk of AKI only for the IHT. Target blood glucose levels in cardiac surgical patients may need to be individualized based on preoperative glycemic status.

Impact of acute glycemic variability on short-term outcomes in patients with ST-segment elevation myocardial infarction: a multicenter population-based study

BACKGROUND

Given the increasing attention to glycemic variability (GV) and its potential implications for cardiovascular outcomes. This study aimed to explore the impact of acute GV on short-term outcomes in Chinese patients with ST-segment elevation myocardial infarction (STEMI).

METHODS

This study enrolled 7510 consecutive patients diagnosed with acute STEMI from 274 centers in China. GV was assessed using the coefficient of variation of blood glucose levels. Patients were categorized into three groups according to GV tertiles (GV1, GV2, and GV3). The primary outcome was 30-day all-cause death, and the secondary outcome was major adverse cardiovascular events (MACEs). Cox regression analyses were conducted to determine the independent correlation between GV and the outcomes.

RESULTS

A total of 7136 patients with STEMI were included. During 30-days follow-up, there was a significant increase in the incidence of all-cause death and MACEs with higher GV tertiles. The 30-days mortality rates were 7.4% for GV1, 8.7% for GV2 and 9.4% for GV3 (p = 0.004), while the MACEs incidence rates was 11.3%, 13.8% and 15.8% for the GV1, GV2 and GV3 groups respectively (p < 0.001). High GV levels during hospitalization were significantly associated with an increased risk of 30-day all-cause mortality and MACEs. When analyzed as a continuous variable, GV was independently associated with a higher risk of all-cause mortality (hazard ratio [HR] 1.679, 95% confidence Interval [CI] 1.005-2.804) and MACEs (HR 2.064, 95% CI 1.386-3.074). Additionally, when analyzed as categorical variables, the GV3 group was found to predict an increased risk of MACEs, irrespective of the presence of diabetes mellitus (DM).

CONCLUSION

Our study findings indicate that a high GV during hospitalization was significantly associated with an increased risk of 30-day all-cause mortality and MACE in Chinese patients with STEMI. Moreover, acute GV emerged as an independent predictor of increased MACEs risk, regardless of DM status.

Comparison of efficacy and safety of insulin degludec/liraglutide and insulin glargine U-100/lixisenatide in individuals with type 2 diabetes mellitus using professional continuous glucose monitoring

AIM/INTRODUCTION

Insulin glargine U100/lixisenatide and insulin degludec/liraglutide are fixed-ratio combinations containing basal insulin and a glucagon-like peptide-1 receptor agonist capable of reducing both fasting and postprandial blood glucose levels with a single formulation. This study aimed to compare the time in range (TIR) and the time below range (TBR) level 1 using professional continuous glucose monitoring and to establish criteria for the differential use of the fixed-ratio combinations.

MATERIALS AND METHODS

Thirty-six outpatients with type 2 diabetes mellitus (24 men and 12 women; average age, 62.1 years) were randomly assigned to the groups. At 0 and 18 weeks, a device was worn to compare the TIR and TBR level 1. The correlation between the C-peptide index at baseline and TIR at 18 weeks was assessed.

RESULTS

The TIR and TBR level 1 showed no significant differences between the two groups. Both groups showed significant positive correlations between the C-peptide index and the TIR (P = 0.002, r = 0.679; P = 0.002, r = 0.681, respectively). The changes in glycemic variability, therapeutic indices, and body mass index were not significantly different among the groups (P > 0.05). The receiver operating curve analysis revealed that the cut-off values of the C-peptide index to achieve TIR of >70% at 18 weeks were 1.258 (sensitivity, 77.8%; specificity, 100%) and 1.099 (sensitivity, 57.1%; specificity, 90.9%) in the insulin glargine U100/lixisenatide and insulin degludec/liraglutide groups, respectively.

CONCLUSIONS

A TIR of >70% was achieved for both fixed-ratio combinations without significant differences.

Glycemic Variability Assessment in Newly Treated Exocrine Pancreatic Insufficiency With Type 1 Diabetes

BACKGROUND

Thirty-nine percent of people with type 1 diabetes may have lowered pancreatic elastase levels, correlated with exocrine pancreatic insufficiency (EPI or PEI). EPI is treated with oral supplementation of pancreatic enzymes. Little is known about the glycemic impact of pancreatic enzyme replacement therapy (PERT) in people with diabetes. This article demonstrates a method of assessing glycemic variability (GV), glycemic outcomes, and other changes in an individual with type 1 diabetes using open-source automated insulin delivery (AID).

METHOD

Macronutrient, PERT intake, and EPI-related symptoms were self-tracked; diabetes data were collected automatically via an open-source AID system. Diabetes data were uploaded via Nightscout to Open Humans and downloaded for analysis alongside self-tracked data (food, PERT). Glycemic outcomes, macronutrients, PERT dosing, and a variety of GV metrics following meals were evaluated for one month before and one month after PERT commencement. Breakfast was assessed independently across both time periods.

RESULTS

In an n = 1 individual using an open-source AID, time in range was already above goal and improved further after PERT commencement. Glucose rate of change and excursions >180 mg/dL were reduced; mean high blood glucose index was reduced overall and more so specifically at breakfast following PERT commencement.

CONCLUSIONS

GV can aid in assessing response to new-onset medications, as was demonstrated in this article for n = 1 individual with type 1 diabetes (using an open-source AID) after commencing PERT for newly identified EPI. GV may be useful for evaluating the efficacy of new-onset medications for people with insulin-requiring diabetes.

Effect of cinnamon spice on continuously monitored glycemic response in adults with prediabetes: a 4-week randomized controlled crossover trial

BACKGROUND

Previous clinical studies showing that cinnamon spice lowers blood glucose concentrations had inconsistent results.

OBJECTIVES

To determine the effect of daily cinnamon spice supplementation in an amount commonly used for seasoning on glucose concentrations in adults with obesity and prediabetes.

METHODS

Following a 2-wk run-in period of maintaining a low polyphenol/fiber diet, 18 participants with obesity and prediabetes underwent a 10-wk randomized, controlled, double-blind, crossover trial (mean age 51.1 y; mean fasting plasma glucose 102.9 mg/dL). The participants were randomly assigned to take cinnamon (4 g/d) or placebo for 4-wk, followed by a 2-wk washout period, and then crossed over to the other intervention for an additional 4-wk. Glucose changes were measured with continuous glucose monitoring. Oral glucose tolerance testing immediately following ingestion of cinnamon or placebo was performed at 4-time points to assess their acute effects both at the baseline and end of each intervention phase. Digestive symptom logs were obtained daily.

RESULTS

There were 694 follow-up days with 66,624 glucose observations. When compared with placebo, 24-h glucose concentrations were significantly lower when cinnamon was administered [mixed-models; effect size (ES) = 0.96; 95 % confidence interval (CI): -2.9, -1.5; P < 0.001]. Similarly, the mean net-area-under-the-curve (netAUC) for glucose was significantly lower than for placebo when cinnamon was given (over 24 h; ES = -0.66; 95 % CI: 2501.7, 5412.1, P = 0.01). Cinnamon supplementation resulted in lower glucose peaks compared with placebo (Δpeak 9.56 ± 9.1 mg/dL compared with 11.73 ± 8.0 mg/dL; ES = -0.57; 95 % CI: 0.8, 3.7, P = 0.027). Glucose-dependent-insulinotropic-polypeptide concentrations increased during oral glucose tolerance testing + cinnamon testing (mixed-models; ES = 0.51; 95 % CI: 1.56, 100.1, P = 0.04), whereas triglyceride concentrations decreased (mixed-models; ES = 0.55; 95 % CI: -16.0, -1.6, P = 0.02). Treatment adherence was excellent in both groups (cinnamon: 97.6 ± 3.4 % compared with placebo: 97.9 ± 3.7 %; ES = -0.15; 95 % CI: -1.8, 0.2, P = 0.5). No differences were found in digestive symptoms (abdominal pain, borborygmi, bloating, excess flatus, and stools/day) between cinnamon and placebo groups.

CONCLUSIONS

Cinnamon, a widely available and low-cost supplement, may contribute to better glucose control when added to the diet in people who have obesity-related prediabetes. This trial was registered at clinicaltrials.gov as NCT04342624.

Short-term glycemic variability in non-diabetic, non-obese dogs assessed by common glycemic variability indices

Glycemic variability (GV) refers to swings in blood glucose levels and is an emerging measure of glycemic control in clinical practice. It is associated with micro- and macrovascular complications and poor clinical outcomes in diabetic humans. Although an integral part of patient assessment in human patients, it is to a large extent neglected in insulin-treated diabetic dogs. This prospective pilot study was performed to describe canine within-day GV in non-diabetic dogs with the aim to provide a basis for the interpretation of daily glucose profiles, and to promote GV as an accessible tool for future studies in veterinary medicine. Interstitial glucose concentrations of ten non-diabetic, non-obese beagles were continuously measured over a 48-h period using a flash glucose monitoring system. GV was assessed using the common indices MAGE (mean amplitude of glycemic excursion), GVP (Glycemic variability percentage) and CV (coefficient of variation). A total of 2260 sensor measurements were obtained, ranging from 3.7 mmol/L (67 mg/dL) to 8.5 mmol/L (153 mg/dL). Glucose profiles suggested a meal-dependent circadian rhythmicity with small but significant surges during the feeding periods. No differences in GV indices were observed between day and night periods (p > 0.05). The MAGE (mmol/L), GVP (%) and CV (%) were 0.86 (± 0.19), 7.37 (± 1.65), 6.72 (± 0.89) on day one, and 0.83 (± 0.18), 6.95 (± 1.52), 6.72 (± 1.53) on day two, respectively. The results of this study suggest that GV is low in non-diabetic dogs and that glucose concentrations are kept within narrow ranges.

Enhancement of Ambulatory Glucose Profile for Decision Assistance and Treatment Adjustments

The ambulatory glucose profile (AGP) lacks sufficient statistical metrics and insightful graphs; indeed, it is missing important information on the temporal patterns of glucose variations. The AGP graph is difficult to interpret due to the overlapping metrics and fluctuations in glucose levels over 14 days. The objective of this proposed work is to overcome these challenges, specifically the lack of insightful information and difficulty in interpreting AGP graphs, to create a platform for decision assistance. The present work proposes 20 findings built from decision rules that were developed from a combination of AGP metrics and additional statistical metrics, which have the potential to identify patterns and insightful information on hyperglycemia and hypoglycemia. The "CGM Trace" webpage was developed, in which insightful metrics and graphical representations can be used to make inferences regarding the glucose data of any user. However, doctors (endocrinologists) can access the "Findings" tab for a summarized presentation of their patients' glycemic control. The findings were implemented for 67 patients' data, in which the data of 15 patients were collected from a clinical study and the data of 52 patients were gathered from a public dataset. The findings were validated by means of MANOVA (multivariate analysis of variance), wherein a p value of < 0.05 was obtained, depicting a strong significant correlation between the findings and the metrics. The proposed work from "CGM Trace" offers a deeper understanding of the CGM data, enhancing AGP reports for doctors to make treatment adjustments based on insightful information and hidden patterns for better diabetic management.

Artificial Intelligence and Machine Learning for Improving Glycemic Control in Diabetes: Best Practices, Pitfalls, and Opportunities

OBJECTIVE

Artificial intelligence and machine learning are transforming many fields including medicine. In diabetes, robust biosensing technologies and automated insulin delivery therapies have created a substantial opportunity to improve health. While the number of manuscripts addressing the topic of applying machine learning to diabetes has grown in recent years, there has been a lack of consistency in the methods, metrics, and data used to train and evaluate these algorithms. This manuscript provides consensus guidelines for machine learning practitioners in the field of diabetes, including best practice recommended approaches and warnings about pitfalls to avoid.

METHODS

Algorithmic approaches are reviewed and benefits of different algorithms are discussed including importance of clinical accuracy, explainability, interpretability, and personalization. We review the most common features used in machine learning applications in diabetes glucose control and provide an open-source library of functions for calculating features, as well as a framework for specifying data sets using data sheets. A review of current data sets available for training algorithms is provided as well as an online repository of data sources.

SIGNIFICANCE

These consensus guidelines are designed to improve performance and translatability of new machine learning algorithms developed in the field of diabetes for engineers and data scientists.

Statistical Packages and Algorithms for the Analysis of Continuous Glucose Monitoring Data: A Systematic Review

BACKGROUND

Continuous glucose monitoring (CGM) measures glucose levels every 1 to 15 minutes and is widely used in clinical and research contexts. Statistical packages and algorithms reduce the time-consuming and error-prone process of manually calculating CGM metrics and contribute to standardizing CGM metrics defined by international consensus. The aim of this systematic review is to summarize existing data on (1) statistical packages for retrospective CGM data analysis and (2) statistical algorithms for retrospective CGM analysis not available in these statistical packages.

METHODS

A systematic literature search in PubMed and EMBASE was conducted on September 19, 2023. We also searched Google Scholar and Google Search until October 12, 2023 as sources of gray literature and performed reference checks of the included literature. Articles in English and Danish were included. This systematic review is registered with PROSPERO (CRD42022378163).

RESULTS

A total of 8731 references were screened and 46 references were included. We identified 23 statistical packages for the analysis of CGM data. The statistical packages could calculate many metrics of the 2022 CGM consensus and non-consensus CGM metrics, and 22/23 (96%) statistical packages were freely available. Also, 23 statistical algorithms were identified. The statistical algorithms could be divided into three groups based on content: (1) CGM data reduction (eg, clustering of CGM data), (2) composite CGM outcomes, and (3) other CGM metrics.

CONCLUSION

This systematic review provides detailed tabular and textual up-to-date descriptions of the contents of statistical packages and statistical algorithms for retrospective analysis of CGM data.

Glycemic Variability As a Prognostic Factor for Mortality in Patients With Critical Illness: A Systematic Review and Meta-Analysis

OBJECTIVES

To perform a systematic review and meta-analysis to evaluate the association of various measures of glycemic variability, including time-domain and complexity-domain, with short-term mortality in patients with critical illness.

DATA SOURCES

We searched Embase Classic +, MEDLINE, and the Cochrane Database of Systematic Reviews from inception to November 3, 2023.

STUDY SELECTION

We included English language studies that assessed metrics of glycemic variation or complexity and short-term mortality in patients admitted to the ICU.

DATA EXTRACTION

Two authors performed independent data abstraction and risk-of-bias assessments. We used a random-effects model to pool binary and continuous data and summarized estimates of effect using odds ratios and mean difference. We used the Quality in Prognosis Studies tool to assess risk of bias and the Grading of Recommendations, Assessment, Development and Evaluations to assess certainty of pooled estimates.

DATA SYNTHESIS

We included 41 studies (n = 162,259). We demonstrate that increased sd, coefficient of variance, glycemic lability index, and decreased time in range are probably associated with increased mortality in critically ill patients (moderate certainty) and that increased mean absolute glucose, mean amplitude of glycemic excursion, and detrended fluctuation analysis may be associated with increased mortality (low certainty).

CONCLUSIONS

We found a consistent association between increased measures of glycemic variability and higher short-term mortality in patient with critical illness. Further research should focus on standardized measurements of glycemic variation and complexity, along with their utility as therapeutic targets and prognostic markers.

Explainable Machine Learning for Real-Time Hypoglycemia and Hyperglycemia Prediction and Personalized Control Recommendations

BACKGROUND

The occurrences of acute complications arising from hypoglycemia and hyperglycemia peak as young adults with type 1 diabetes (T1D) take control of their own care. Continuous glucose monitoring (CGM) devices provide real-time glucose readings enabling users to manage their control proactively. Machine learning algorithms can use CGM data to make ahead-of-time risk predictions and provide insight into an individual's longer term control.

METHODS

We introduce explainable machine learning to make predictions of hypoglycemia (<70 mg/dL) and hyperglycemia (>270 mg/dL) up to 60 minutes ahead of time. We train our models using CGM data from 153 people living with T1D in the CITY (CGM Intervention in Teens and Young Adults With Type 1 Diabetes)survey totaling more than 28 000 days of usage, which we summarize into (short-term, medium-term, and long-term) glucose control features along with demographic information. We use machine learning explanations (SHAP [SHapley Additive exPlanations]) to identify which features have been most important in predicting risk per user.

RESULTS

Machine learning models (XGBoost) show excellent performance at predicting hypoglycemia (area under the receiver operating curve [AUROC]: 0.998, average precision: 0.953) and hyperglycemia (AUROC: 0.989, average precision: 0.931) in comparison with a baseline heuristic and logistic regression model.

CONCLUSIONS

Maximizing model performance for glucose risk prediction and management is crucial to reduce the burden of alarm fatigue on CGM users. Machine learning enables more precise and timely predictions in comparison with baseline models. SHAP helps identify what about a CGM user's glucose control has led to predictions of risk which can be used to reduce their long-term risk of complications.

Glycemic variability: Measurement, target, impact on complications of diabetes and does it really matter?

Over the past two decades, there has been continuous advancement in the accuracy and complexity of continuous glucose monitoring devices. Continuous glucose monitoring provides valuable insights into blood glucose dynamics, and can record glucose fluctuations accurately and completely. Glycemic variability (GV) is a straightforward measure of the extent to which a patient's blood glucose levels fluctuate between high peaks and low nadirs. Many studies have investigated the relationship between GV and complications, primarily in the context of type 2 diabetes. Nevertheless, the exact contribution of GV to the development of diabetes complications remains unclear. In this literature review, we aimed to summarize the existing evidence regarding the measurement, target level, pathophysiological mechanisms relating GV and tissue damage, and population-based studies of GV and diabetes complications. Additionally, we introduce novel methods for measuring GV, and discuss several unresolved issues of GV. In the future, more longitudinal studies and trials are required to confirm the exact role of GV in the development of diabetes complications.

How to Measure Glycemic Variability? A Literature Review

Optimal glycemic control without the presence of diabetes-related complications is the primary goal for adequate diabetes management. Recent studies have shown that hemoglobin A1c level cannot fully evaluate diabetes management as glycemic fluctuations are demonstrated to have a major impact on the occurrence of diabetes-related micro- and macroangiopathic comorbidities. The use of continuous glycemic monitoring systems allowed the quantification of glycemic fluctuations, providing valuable information about the patients' glycemic control through various indicators that evaluate the magnitude of glycemic fluctuations in different time intervals. This review highlights the significance of glycemic variability by describing and providing a better understanding of common and alternative indicators available for use in clinical practice.

Association of Long-Term HbA1c Variability with Anxiety and Depression in Patients with Type 2 Diabetes: A Cross-Sectional Retrospective Study

Purpose

To explore the relationship between long-term glycemic variability and anxiety and depression in patients with type 2 diabetes.

Participants and Methods

A cohort comprising 214 individuals diagnosed with type 2 diabetes participated in this study. Comprehensive demographic and laboratory information was gathered for them. The evaluation of anxiety relied on the 7-item Generalized Anxiety Disorder Scale (GAD-7), while depression was assessed utilizing the 9-item Health Questionnaire (PHQ-9). Based on the presence or absence of anxiety and depression, participants were categorized into either the mood disorder or control groups. Subsequently, univariate and stepwise multiple binary logistic regression analyses were conducted to investigate the potential correlations between factors and the presence of anxiety and depression.

Results

The prevalence of anxiety disorders is 23%, and depression is 32%. The prevalence of smoking, diabetic autonomic neuropathy, stroke, and osteoporosis in the mood disorder group was significantly higher than that in the control group (P < 0.05), the glycated hemoglobin A1c variability score (HVS), mean hemoglobin A1c value, total cholesterol, urinary albumin/creatinine and systemic immune-inflammatory index (SII) were significantly higher in the control group (P < 0.05). The level of high-density lipoprotein in the mood disorder group was significantly lower than the control group (P < 0.05). In stepwise multiple binary logistic regression analyses, the main factors associated with anxiety were depression (P < 0.001, OR=117.581) and gender (P < 0.001, OR=9.466), and the main factors related to depression included anxiety (P < 0.001, OR=49.424), smoking (P=0.042, OR=2.728), HVS (P=0.004, OR=8.664), and SII (P=0.014, OR=1.002).

Conclusion

Persistent fluctuations in blood glucose levels have been linked to anxiety and depression. Consequently, maintaining an optimal level of glycemic control and minimizing fluctuations becomes imperative in the comprehensive management of diabetes.

Are postprandial glucose responses sufficiently person-specific to use in personalized dietary advice? Design of the RepEAT study: a fully controlled dietary intervention to determine the variation in glucose responses

Introduction

An elevated postprandial glucose response is associated with an increased risk of cardiometabolic diseases. Existing research suggests large heterogeneity in the postprandial glucose responses to identical meals and food products between individuals, but the effect of other consumed meals during the day and the order of meals during the day on the heterogeneity in postprandial glucose responses still needs to be investigated. In addition, the robustness of the glucose responses to meals or foods is still unknown.

Objectives

The overall aim of the project is to assess whether the glucose response to a meal is sufficiently person-specific to use in personalized dietary advice. We aim to answer the question: "How replicable are glucose responses to meals within individuals and how consistent is the variation in glucose responses between individuals?"

Methods

The question will be assessed under standardized conditions of a 9-week fully controlled dietary intervention in which all meals are the same between individuals and consumed in a fixed order at a fixed time. 63 apparently healthy men and women with a BMI of 25-40 kg/m2 and aged 45-75 years were enrolled in the RepEAT study (NCT05456815), of whom 53 participants completed the study. The RepEAT study comprised a fully controlled dietary intervention of nine weeks, consisting of three repetitive periods of three weeks. Within each three-week period, a variety of meals and food products were offered during breakfast, lunch, dinner and in between meal snacks. Throughout the dietary intervention, glucose was continuously monitored using Freestyle Libre Pro IQ monitors. Physical activity was monitored using the ActiGraph and ActivPAL. To measure the association between glucose responses and an individual's phenotype, various measurements were performed before the start of the dietary intervention including an oral glucose tolerance test, a high-fat mixed meal challenge, assessment of body fat distribution including liver fat (MRI/MRS), and cardiometabolic markers.

Discussion

The repetitive and fully controlled nature of the dietary study allows detailed assessment of the replicability of the glucose responses to meals and food products within individuals. Furthermore, the consistency of the variation between individuals independent of insulin resistance will be determined.

A Consensus Statement for Continuous Glucose Monitoring Metrics for Inpatient Clinical Trials

Diabetes Technology Society organized an expert consensus panel to develop metrics for research in the use of continuous glucose monitors (CGMs) in a hospital setting. The experts met virtually in small groups both before and after an April 13, 2023 virtual meeting of the entire panel. The goal of the panel was to develop consensus definitions in anticipation of greater use of CGMs in hospital settings in the future. Establishment of consensus definitions of inpatient analytical metrics will be easier to compare outcomes between studies. Panelists defined terms related to 10 dimensions of measurements related to the use of CGMs including (1) hospital hypoglycemia, (2) hospital hyperglycemia, (3) hospital time in range, (4) hospital glycemic variability, (5) hospital glycemia risk index, (6) accuracy of CGM devices and reference methods for CGMs in the hospital, (7) meaningful time blocks for hospital glycemic goals, (8) hospital CGM data sufficiency, (9) using CGM data for insulin dosing, and (10) miscellaneous factors. The panelists voted on 51 proposed recommendations. Based on the panel vote, 51 recommendations were classified as either strong (43) or mild (8). Additional research is needed on CGM performance in the hospital. This consensus report is intended to support that type of research intended to improve outcomes for hospitalized people with diabetes.

Key indices of glycaemic variability for application in diabetes clinical practice

Near normal glycaemic control in diabetes consists to target daily glucose fluctuations and quarterly HbA1c oscillations in addition to overall glucose exposure. Consequently, the prerequisite is to define simple, and mathematically undisputable key metrics for the short- and long-term variability in glucose homeostasis. As the standard deviations (SD) of either glucose or HbA1c are dependent on their means, the coefficient of variation (CV = SD/mean) should be applied instead as it that avoids the correlation between the SD and mean values. A CV glucose of 36% is the most appropriate threshold between those with stable versus labile glucose homeostasis. However, when near normal mean glucose concentrations are achieved a lower CV threshold of <27 % is necessary for reducing the risk for hypoglycaemia to a minimal rate. For the long-term variability in glucose homeostasis, a CVHbA1c < 5 % seems to be a relevant recommendation for preventing adverse clinical outcomes.

A 2-dimensional model framework for blood glucose prediction based on iterative learning control architecture

The accurate, timely, and personalized prediction for future blood glucose (BG) levels is undoubtedly needed for further advancement of diabetes management technologies. Human inherent circadian rhythm and regular lifestyle resulting in similarity of daily glycemic dynamics play a positive role in the prediction of blood glucose. Inspired by the iterative learning control (ILC) method in the field of automatic control, a 2-dimensional (2-D) model framework is constructed to predict the future blood glucose levels by taking both the short-range information within a day (intra-day) and long-range information between days (inter-day) into account. In this framework, the radial basis function neural network was applied to capture nonlinear relationships in glycemic metabolism, that is, short-range temporal dependence and long-range contemporaneous dependence on previous days. We build models for each patient, and the models were tested on the in silico datasets at various prediction horizons (PHs). The learning model developed in the 2-D framework successfully increases the accuracy and reduces the delay of predictions. This modeling framework provides a new point of view for BG level prediction and contributes to the development of personalized glucose management, such as hypoglycemia warning and glycemic control.

Performance and patients' satisfaction with the A7+TouchCare insulin patch pump system: A randomized controlled non-inferiority study

BACKGROUND

We assessed the performance and patient satisfaction of a new insulin patch pump, the A7+TouchCare (Medtrum), compared with the Omnipod system.

METHODS

This multicenter, randomized, open-label, controlled study enrolled 100 adult patients with type 1 or type 2 diabetes mellitus (A1C ≥ 6.5% and ≤ 9.5%, i.e., 48 to 80 mmol/mol) who were assigned with the Omnipod or with the A7+TouchCare pump for 3 months. The primary study outcome was the glucose management indicator (GMI) calculated with continuous glucose monitoring (CGM).

RESULTS

Premature withdrawals occurs respectively in 2 and 9 participants in the Omnipod and TouchCare groups. In the Per Protocol analysis, the difference in GMI between groups was 0.002% (95% confidence interval -0.251; 0.255). The non-inferiority was demonstrated since the difference between treatments did not overlap the pre-defined non-inferiority margin (0.4%). There was no significant difference in CGM parameters between groups. On average, patients in both groups were satisfied/very satisfied with the insulin pump system. Patients preferred Omnipod as an insulin management system and especially the patch delivery system but preferred the A7+TouchCare personal diabetes manager to control the system.

CONCLUSIONS

This study showed that the A7+TouchCare insulin pump was as efficient as the Omnipod pump in terms of performance and satisfaction.

CLINICAL TRAIL REGISTRATION

The study was registered in the ClinicalTrials.gov protocol register (NCT04223973).

A study of factors influencing long-term glycemic variability in patients with type 2 diabetes: a structural equation modeling approach

Aim

The present study aims to utilize structural equation modeling (SEM) to investigate the factors impacting long-term glycemic variability among patients afflicted with type 2 diabetes.

Method

The present investigation is a retrospective cohort study that involved the collection of data on patients with type 2 diabetes mellitus who received care at a hospital located in Chengdu, Sichuan Province, over a period spanning from January 1, 2013, to October 30, 2022. Inclusion criteria required patients to have had at least three laboratory test results available. Pertinent patient-related information encompassing general demographic characteristics and biochemical indicators was gathered. Variability in the dataset was defined by standard deviation (SD) and coefficient of variation (CV), with glycosylated hemoglobin variation also considering variability score (HVS). Linear regression analysis was employed to establish the structural equation models for statistically significant influences on long-term glycemic variability. Structural equation modeling was employed to analyze effects and pathways.

Results

Diabetes outpatient special disease management, uric acid variability, mean triglyceride levels, mean total cholesterol levels, total cholesterol variability, LDL variability, baseline glycated hemoglobin, and recent glycated hemoglobin were identified as significant factors influencing long-term glycemic variability. The overall fit of the structural equation model was found to be satisfactory and it was able to capture the relationship between outpatient special disease management, biochemical indicators, and glycated hemoglobin variability. According to the total effect statistics, baseline glycated hemoglobin and total cholesterol levels exhibited the strongest impact on glycated hemoglobin variability.

Conclusion

The factors that have a significant impact on the variation of glycosylated hemoglobin include glycosylated hemoglobin itself, lipids, uric acid, and outpatient special disease management for diabetes. The identification and management of these associated factors can potentially mitigate long-term glycemic variability, thereby delaying the onset of complications and enhancing patients' quality of life.

Relationship Between Time in Range and Dusk Phenomenon in Outpatients with Type 2 Diabetes Mellitus

Purpose

The dusk phenomenon refers to a spontaneous and transient pre-dinner hyperglycemia that affects glucose fluctuation and glycemic control, and the increasing use of continuous glucose monitoring (CGM) has facilitated its diagnosis. We investigated the frequency of the dusk phenomenon and its relationship with the time in range (TIR) in patients with type 2 diabetes mellitus (T2DM).

Patients and Methods

This study involved 102 patients with T2DM who underwent CGM for 14 days. CGM-derived metrics and clinical characteristics were evaluated. A consecutive dusk blood glucose difference (pre-dinner glucose minus 2-hour post-lunch glucose) of ≥ 0 or once-only dusk blood glucose difference of < 0 was diagnosed as the clinical dusk phenomenon (CLDP).

Results

We found that the percentage of CLDP was 11.76% (10.34% in men, 13.64% in women). Compared with the non-CLDP group, the CLDP group tended to be younger and have a lower percentage of TIR (%TIR^3.9-10) and higher percentage of time above range (%TAR^>10 and %TAR^>13.9) (P ≤ 0.05). Adjusted for confounding factors, the binary logistic regression analysis showed a negative association of CLDP with %TIR (odds ratio < 1, P < 0.05). We repeated the correlation analysis based on 70%TIR and found significant differences in hemoglobin A1c, fasting blood glucose, mean blood glucose, standard deviation of the sensor glucose values, glucose coefficient of variation, largest amplitude of glycemic excursions, mean amplitude of glycemic excursions, glucose management indicator, and percentage of CLDP between the two subgroups of TIR ≤ 70% and TIR > 70% (P < 0.05). The negative association between TIR and CLDP still remained after adjustment by the binary logistic regression analysis.

Conclusion

The CLDP was frequently present in patients with T2DM. The TIR was significantly correlated with the CLDP and could serve as an independent negative predictor.

The significance of coefficient of variation as a measure of hypoglycaemia risk and glycaemic control in real world users of the automated insulin delivery MiniMed 780G system

AIM

Use of the MiniMed 780G system (MM780G) can result in a reduction in mean and standard deviation (SD) of sensor glucose (SG) values. We assessed the significance of the coefficient of variation (CV) as a measure of hypoglycaemia risk and glycaemic control.

MATERIALS AND METHODS

Data from 10 404 MM780G users were analysed using multivariable logistic regression to assess the contribution of CV to (a) hypoglycaemia risk, measured as not reaching target <1% for time below range (TBR), and (b) achieving targets of time-in-range (TIR) >70% and glucose management indicator <7%. CV was compared with SD and low blood glucose index. To assess the relevance of CV <36% as a therapeutic threshold, we identified the CV cut-off point that optimally discriminated users at risk of hypoglycaemia.

RESULTS

The contribution of CV was the smallest in terms of risk of hypoglycaemia (vs. low blood glucose index and SD) and TIR and glucose management indicator targets (vs. SD). In all cases the models with SD showed the best fit. A CV <43.4% (95% CI: 42.9-43.9) was the optimal cut-off point with a correct classification rate of 87.2% (vs. 72.9% for CV <36%).

CONCLUSION

For MM780G users, CV is a poor marker for hypoglycaemia risk and glycaemic control. We recommend using, for the former, TBR and whether the TBR target is met (and not using CV <36% as a therapeutic threshold for hypoglycaemia); for the latter, TIR, time above range, whether targets are met and a discrete description of mean SG and SD of SG values.

Cardiovascular autonomic neuropathy is associated with increased glycemic variability driven by hyperglycemia rather than hypoglycemia in patients with diabetes

AIM

Cardiac autonomic neuropathy (CAN) has been suggested to be associated with hypoglycemia and impaired hypoglycemia unawareness. We have assessed the relationship between CAN and extensive measures of glucose variability (GV) in patients with type 1 and type 2 diabetes.

METHODS

Participants with diabetes underwent continuous glucose monitoring (CGM) to obtain measures of GV and the extent of hyperglycemia and hypoglycemia and cardiovascular autonomic reflex testing.

RESULTS

Of the 40 participants (20 T1DM and 20 T2DM) (aged 40.70±13.73 years, diabetes duration 14.43±7.35 years, HbA1c 8.85±1.70%), 23 (57.5%) had CAN. Despite a lower coefficient of variation (CV) (31.26±11.87 vs. 40.33±11.03, P=0.018), they had a higher CONGA (8.42±2.58 vs. 6.68±1.88, P=0.024) with a lower median LBGI (1.60 (range: 0.20-3.50) vs. 4.90 (range: 3.20-7.40), P=0.010) and percentage median time spent in hypoglycemia (4 (range:4-13) vs. 1 (range:0-5), P=0.008), compared to those without CAN. The percentage GRADE_Euglycemia (3.30±2.78 vs. 5.69±3.09, P=0.017) and GRADE_Hypoglycemia (0.3 (range: 0 - 3.80) vs. 1.8 (range: 0.9-6.5), P=0.036) were significantly lower, while the percentage median GRADE_{Hyperglycemia} (95.45 (range:93-98) vs. 91.6 (82.8-95.1), P=0.013) was significantly higher in participants with CAN compared to those without CAN.

CONCLUSION

CAN was associated with increased glycemic variability with less time in euglycemia attributed to a greater time in hyperglycemia but not hypoglycemia.

Efficacy of polyethylene glycol loxenatide versus insulin glargine on glycemic control in patients with type 2 diabetes: a randomized, open-label, parallel-group trial

Objective: This trial aimed to evaluate the glycemic control of polyethylene glycol loxenatide measured with continuous glucose monitoring (CGM) in patients with type 2 diabetes mellitus (T2DM), with the hypothesis that participants given PEG-Loxe would spend more time in time-in-range (TIR) than participants were given insulin glargine after 24 weeks of treatment. Methods: This 24-week, randomized, open-label, parallel-group study was conducted in the Department of Endocrine and Metabolic Diseases, Longhu Hospital, Shantou, China. Participants with T2DM, who were ≥45 years of age, HbA1c of 7.0%-11.0%, and treated at least 3 months with metformin were randomized (1:1) to receive PEG-Loxe or insulin glargine. The primary endpoint was TIR (blood glucose range: 3.9-10.0 mmol/L) during the last 2 weeks of treatment (weeks 22-24). Results: From March 2020 to April 2022, a total of 107 participants with T2DM were screened, of whom 78 were enrolled into the trial (n = 39 per group). At the end of treatment (weeks 22-24), participants given PEG-Loxe had a greater proportion of time in TIR compared with participants given insulin glargine [estimated treatment difference (ETD) of 13.4% (95% CI, 6.8 to 20.0, p < 0.001)]. The tight TIR (3.9-7.8 mmol/L) was greater with PEG-Loxe versus insulin glargine, with an ETD of 15.6% (95% CI, 8.9 to 22.4, p < 0.001). The time above range (TAR) was significantly lower with PEG-Loxe versus insulin glargine [ETD for level 1: -10.5% (95% CI: -14.9 to -6.0), p < 0.001; ETD for level 2: -4.7% (95% CI: -7.9 to -1.5), p = 0.004]. The time below range (TBR) was similar between the two groups. The mean glucose was lower with PEG-Loxe versus insulin glargine, with an ETD of -1.2 mmol/L (95% CI, -1.9 to -0.5, p = 0.001). The SD of CGM glucose levels was 1.88 mmol/L for PEG-Loxe and 2.22 mmol/L for insulin glargine [ETD -0.34 mmol/L (95% CI: -0.55 to -0.12), p = 0.002], with a similar CV between the two groups. Conclusion: The addition of once-weekly GLP-1RA PEG-Loxe to metformin was superior to insulin glargine in improving glycemic control and glycemic variability evaluated by CGM in middle-aged and elderly patients with T2DM.

Continuous Glucose Monitoring in Patients Following Simultaneous Pancreas-Kidney Transplantation: Time in Range and Glucose Variability

Simultaneous pancreas-kidney transplantation (SPKT) can improve long-term patient survival and restore endogenous insulin secretion in recipients with type 1 diabetes (T1D). There are currently few data on glucose fluctuations assessed by continuous glucose monitoring (CGM) after SPKT. Aim: to evaluate CGM-derived time in range (TIR) and glucose variability (GV) in patients with T1D and functioning pancreatic grafts after SPKT. Fifty-four CGM recordings from 43 patients, 15 men and 28 women, aged 34 (31; 39) years were analyzed. Time since SKPT was up to 1 year (group 1, n = 13), from 1 to 5 years (group 2, n = 15), and from 5 to 12 years (group 3, n = 26). TIR (3.9-10 mmol/L), Time Above Range (TAR), Time Below Range (TBR), and GV parameters were estimated. There were no differences in mean glucose (5.5 [5.1; 6.2], 5.9 [5.4; 6.2], and 5.9 [5.6; 6.7] mmol/L), TIR (97.6 [92.8-99.1], 97.2 [93.2; 99.1], and 97.5 [93.4; 99]%); TAR (0, 1.8 [1.3; 3.7], and 2.5 [2; 5]%), TBR (5 [3.3; 12.7], 4.1 [2.2; 10.1], and 3.5 [1.3; 6.5]%) and GV parameters between three groups (all p > 0.05). Thus, recipients with functioning pancreatic grafts demonstrate remarkably high TIR and low GV after SPKT.

Glycemic Variability and Prognosis of Patients with Intracerebral Hemorrhage: A Meta-Analysis

Glycemic disorder may affect the outcomes of patients with intracerebral hemorrhage (ICH). However, the association between glycemic variability (GV) and prognosis in these patients remains to be determined. We performed a meta-analysis to compressive the influence of GV on functional outcome and mortality in patients with ICH. Observational studies comparing the risks of poor functional outcome (defined as modified Rankin Scale>2) and all-cause mortality between ICH patients with higher versus lower acute GV were retrieved by systematic search of Medline, Web of Science, Embase, CNKI, and Wanfang databases. A random-effect model was used to pool the data after incorporating the between-study heterogeneity. Sensitivity analyses were performed to evaluate the stability of the findings. Eight cohort studies involving 3400 patients with ICH were included in the meta-analysis. The follow-up duration was within 3 months after admission. All of the included studies used standard deviation of blood glucose (SDBG) as the indicator of acute GV. Pooled results showed that ICH patients with higher SDBG were associated with a higher risk of poor functional outcome as compared to those with lower SDBG [risk ratio (RR): 1.84, 95% confidence interval (CI): 1.41 to 2.42, p<0.001, I2=0%]. In addition, patients with higher category of SDBG were also associated with a higher mortality risk (RR: 2.39, 95% CI: 1.79 to 3.19, p<0.001, I2=0%). In conclusion, high acute GV may be a predictor of poor functional outcome and mortality of patients with ICH.

Long-Term Effects and Potential Impact of Early Nutrition with Breast Milk or Infant Formula on Glucose Homeostasis Control in Healthy Children at 6 Years Old: A Follow-Up from the COGNIS Study

There is scarce evidence about early nutrition programming of dynamic aspects of glucose homeostasis. We analyzed the long-term effects of early nutrition on glycemic variability in healthy children. A total of 92 children participating in the COGNIS study were considered for this analysis, who were fed with: a standard infant formula (SF, n = 32), an experimental formula (EF, n = 32), supplemented with milk fat globule membrane (MFGM) components, long-chain polyunsaturated fatty acids (LC-PUFAs), and synbiotics, or were breastfed (BF, n = 28). At 6 years old, BF children had lower mean glucose levels and higher multiscale sample entropy (MSE) compared to those fed with SF. No differences in MSE were found between EF and BF groups. Normal and slow weight gain velocity during the first 6 months of life were associated with higher MSE at 6 years, suggesting an early programming effect against later metabolic disorders, thus similarly to what we observed in breastfed children. Conclusion: According to our results, BF and normal/slow weight gain velocity during early life seem to protect against glucose homeostasis dysregulation at 6 years old. EF shows functional similarities to BF regarding children's glucose variability. The detection of glucose dysregulation in healthy children would help to develop strategies to prevent the onset of metabolic disorders in adulthood.

Association Between the Variability of Glycated Hemoglobin and Retinopathy in Patients with Type 2 Diabetes Mellitus: A Meta-Analysis

Visit-to-visit variability of glycated hemoglobin (HbA1c) is a marker of long-term glycemic fluctuation, which has been related to increased risk of macrovascular complications in patients with type 2 diabetes mellitus (T2DM). The association between HbA1c variability and retinopathy in patients with T2DM, however, has been inconsistent in previous studies. In order to fully evaluate the above association, we conducted a meta-analysis. Observational studies related to the aim of the meta-analysis were identified by search of PubMed, Web of Science, and Embase databases. Studies with HbA1c variability evaluated as the standard deviation (SD) and/or the coefficients of variation (CV) of HbA1c were included. The results were analyzed using a random-effects model that incorporated potential heterogeneity between studies. Twelve observational studies involving 44 662 T2DM patients contributed to the meta-analysis. Overall, 5150 (11.5%) patients developed retinopathy. Pooled results showed that compared to patients with lower HbA1c variability, T2DM patients with higher HbA1c-SD (relative risk [RR]: 1.48, 95% confidence interval [CI]: 1.24 to 1.78, p<0.001, I2=34%) and higher HbA1c-CV (RR: 1.29, 95% CI: 1.05 to 1.59, p=0.02, I2=0%) were both associated with higher risk of DR. For studies with HbA1c-SD, the association was not significantly affected by study characteristics such as country, study design, mean age, disease duration, adjustment of mean HbA1c, or quality scores (p for subgroup difference all>0.05). In conclusion, higher HbA1c variability may be associated with an increased risk of retinopathy in patients with T2DM.

Association between Intensity Levels of Physical Activity and Glucose Variability among Children and Adolescents with Type 1 Diabetes

Studies would indicate a reduction in hemoglobin A1c levels following moderate and/or vigorous physical activity (PA) for people managing diabetes. However, prior investigations rarely looked at glucose variability in an adolescent population.

PURPOSE

The purpose of this investigation was to test the relationship between physical activity intensity levels and glucose variability in a sample of adolescents with type 1 diabetes mellitus, and if the amount of time accumulated for each intensity level is predictive of changes in glucose variability.

METHODS

Glucose variability was determined using continuous glucose monitor data and physical activity intensity time was retrieved through Fitabase®. Both glucose and physical activity data were collected over a two-week timeframe. Data analysis was completed using Pearson's correlation and a simple linear regression with a p-value of 0.05 to determine significance.

RESULTS

A significant inverse relationship was observed (p = 0.04) between glucose variability and average minutes of daily moderate-intensity activity (r = -0.59), as well as moderate and vigorous physical activity (MVPA) combined (r = -0.86; p = 0.03). A simple linear regression indicated that only MVPA was a significant predictor of glucose variability (β = -0.12; 95% CI: -0.23--0.01, p = 0.03).

CONCLUSION

These data demonstrated that the total amount of daily physical activity is important when properly managing type 1 diabetes mellitus, but time spent in MVPA over two weeks may have an inverse relationship with glucose variability in children and adolescents over a span of two weeks.

Decreasing complexity of glucose time series derived from continuous glucose monitoring is correlated with deteriorating glucose regulation

Most information used to evaluate diabetic statuses is collected at a special time-point, such as taking fasting plasma glucose test and providing a limited view of individual's health and disease risk. As a new parameter for continuously evaluating personal clinical statuses, the newly developed technique "continuous glucose monitoring" (CGM) can characterize glucose dynamics. By calculating the complexity of glucose time series index (CGI) with refined composite multi-scale entropy analysis of the CGM data, the study showed for the first time that the complexity of glucose time series in subjects decreased gradually from normal glucose tolerance to impaired glucose regulation and then to type 2 diabetes (P for trend < 0.01). Furthermore, CGI was significantly associated with various parameters such as insulin sensitivity/secretion (all P < 0.01), and multiple linear stepwise regression showed that the disposition index, which reflects β-cell function after adjusting for insulin sensitivity, was the only independent factor correlated with CGI (P < 0.01). Our findings indicate that the CGI derived from the CGM data may serve as a novel marker to evaluate glucose homeostasis.

Making sense of glucose sensors in end-stage kidney disease: A review

Diabetes mellitus remains the leading cause of end-stage kidney disease worldwide. Inadequate glucose monitoring has been identified as one of the gaps in care for hemodialysis patients with diabetes, and lack of reliable methods to assess glycemia has contributed to uncertainty regarding the benefit of glycemic control in these individuals. Hemoglobin A1c, the standard metric to evaluate glycemic control, is inaccurate in patients with kidney failure, and does not capture the full range of glucose values for patients with diabetes. Recent advances in continuous glucose monitoring have established this technology as the new gold standard for glucose management in diabetes. Glucose fluctuations are uniquely challenging in patients dependent on intermittent hemodialysis, and lead to clinically significant glycemic variability. This review evaluates continuous glucose monitoring technology, its validity in the setting of kidney failure, and interpretation of glucose monitoring results for the nephrologist. Continuous glucose monitoring targets for patients on dialysis have yet to be established. While continuous glucose monitoring provides a more complete picture of the glycemic profile than hemoglobin A1c and can mitigate high-risk hypoglycemia and hyperglycemia in the context of the hemodialysis procedure itself, whether the technology can improve clinical outcomes merits further investigation.

Glucose variability: A physiological correlate of eating disorder behaviors among individuals with binge-spectrum eating disorders

OBJECTIVES

Elevated glucose variability may be one mechanism that increases risk for significant psychological and physiological health conditions among individuals with binge-spectrum eating disorders (B-EDs), given the impact of eating disorder (ED) behaviors on blood glucose levels. This study aimed to characterize glucose variability among individuals with B-EDs compared with age-matched, sex-matched, and body mass index-matched controls, and investigate the association between frequency of ED behaviors and glucose variability.

METHODS

Participants were 52 individuals with B-EDs and 22 controls who wore continuous glucose monitors to measure blood glucose levels and completed ecological momentary assessment surveys to measure ED behaviors for 1 week. Independent samples t-tests compared individuals with B-EDs and controls and multiple linear regression models examined the association between ED behaviors and glucose variability.

RESULTS

Individuals with B-EDs demonstrated numerically higher glucose variability than controls (t = 1.42, p = .08, d = 0.43), although this difference was not statistically significant. When controlling for covariates, frequency of ED behaviors was significantly, positively associated with glucose variability (t = 3.17, p = .003) with medium effect size (f2  = 0.25). Post hoc analyses indicated that binge eating frequency was significantly associated with glucose variability, while episodes of 5+ hours without eating were not.

DISCUSSION

Glucose variability among individuals with B-EDs appears to be positively associated with engagement in ED behaviors, particularly binge eating. Glucose variability may be an important mechanism by which adverse health outcomes occur at elevated rates in B-EDs and warrants future study.

PUBLIC SIGNIFICANCE

This study suggests that some individuals with binge ED and bulimia nervosa may experience elevated glucose variability, a physiological symptom that is linked to a number of adverse health consequences. The degree of elevation in glucose variability is positive associated with frequency of eating disorder behaviors, especially binge eating.

Glycemic Variability in Type 1 Diabetes Mellitus Pregnancies—Novel Parameters in Predicting Large-for-Gestational-Age Neonates: A Prospective Cohort Study

Pregnancies with type 1 diabetes mellitus (T1DM) have a high incidence of large-for-gestational-age neonates (LGA) despite optimal glycemic control. In recent years, glycemic variability (GV) has emerged as a possible risk factor for LGA, but the results of the conducted studies are unclear. This study analyzed the association between GV and LGA development in pregnancies with T1DM. This was a prospective cohort study of patients with T1DM who used continuous glucose monitoring (CGM) during pregnancy. Patients were followed from the first trimester to birth. GV parameters were calculated for every trimester using the EasyGV calculator. The main outcomes were LGA or no-LGA. Logistic regression analysis was used to assess the association between GV parameters and LGA. In total, 66 patients were included. The incidence of LGA was 36%. The analysis extracted several GV parameters that were significantly associated with the risk of LGA. The J-index was the only significant parameter in every trimester of pregnancy (odds ratios with confidence intervals were 1.33 (1.02, 1.73), 3.18 (1.12, 9.07), and 1.37 (1.03, 1.82), respectively. Increased GV is a risk factor for development of LGA. The J-index is a possible novel GV parameter that may be assessed in all three trimesters of pregnancy together with glycated hemoglobin and time-in-range.

Patient-Tailored Decision Support System Improves Short- and Long-Term Glycemic Control in Type 2 Diabetes

BACKGROUND

The increasing prevalence of type 2 diabetes mellitus (T2D) and specialist shortage has caused a healthcare gap that can be bridged by a decision support system (DSS). We investigated whether a diabetes DSS can improve long- and/or short-term glycemic control.

METHODS

This is a retrospective observational cohort study of the Diabetiva program, which offered a patient-tailored DSS using Karlsburger Diabetes-Management System (KADIS) once a year. Glycemic control was analyzed at baseline and after 12 months in 452 individuals with T2D. Time in range (TIR; glucose 3.9-10 mmol/L) and Q-Score, a composite metric developed for analysis of continuous glucose profiles, were short-term and HbA1c long-term measures of glycemic control. Glucose variability (GV) was also measured.

RESULTS

At baseline, one-third of patients had good short- and long-term glycemic control. Q-Score identified insufficient short-term glycemic control in 17.9% of patients with HbA1c <6.5%, mainly due to hypoglycemia. GV and hyperglycemia were responsible in patients with HbA1c >7.5% and >8%, respectively. Application of DSS at baseline improved short- and long-term glycemic control, as shown by the reduced Q-Score, GV, and HbA1c after 12 months. Multiple regression demonstrated that the total effect on GV resulted from the single effects of all influential parameters.

CONCLUSIONS

DSS can improve short- and long-term glycemic control in individuals with T2D without increasing hypoglycemia. The Q-Score allows identification of individuals with insufficient glycemic control. An effective strategy for therapy optimization could be the selection of individuals with T2D most at need using the Q-Score, followed by offering patient-tailored DSS.

Compensatory thalamocortical functional hyperconnectivity in type 2 Diabetes Mellitus

Type 2 diabetes mellitus (T2DM) is associated with brain damage and cognitive decline. Despite the fact that the thalamus involves aspects of cognition and is typically affected in T2DM, existing knowledge of subregion-level thalamic damage and its associations with cognitive performance in T2DM patients is limited. The thalamus was subdivided into 8 subregions in each hemisphere. Resting-state functional and structural MRI data were collected to calculate resting-state functional connectivity (rsFC) and gray matter volume (GMV) of each thalamic subregion in 62 T2DM patients and 50 healthy controls. Compared with controls, T2DM patients showed increased rsFC of the medial pre-frontal thalamus, posterior parietal thalamus, and occipital thalamus with multiple cortical regions. Moreover, these thalamic functional hyperconnectivity were associated with better cognitive performance and lower glucose variability in T2DM patients. However, there were no group differences in GMV for any thalamic subregions. These findings suggest a possible neural compensation mechanism whereby selective thalamocortical functional hyperconnectivity facilitated by better glycemic control help to preserve cognitive ability in T2DM patients, which may ultimately inform intervention and prevention of T2DM-related cognitive decline in real-world clinical settings.

Machine Learning Models for Nocturnal Hypoglycemia Prediction in Hospitalized Patients with Type 1 Diabetes

Nocturnal hypoglycemia (NH) is a dangerous complication of insulin therapy that often goes undetected. In this study, we aimed to generate machine learning (ML)-based models for short-term NH prediction in hospitalized patients with type 1 diabetes (T1D). The models were trained on continuous glucose monitoring (CGM) data obtained from 406 adult patients admitted to a tertiary referral hospital. Eight CGM-derived metrics of glycemic control and glucose variability were included in the models. Combinations of CGM and clinical data (23 parameters) were also assessed. Random Forest (RF), Logistic Linear Regression with Lasso regularization, and Artificial Neuron Networks algorithms were applied. In our models, RF provided the best prediction accuracy with 15 min and 30 min prediction horizons. The addition of clinical parameters slightly improved the prediction accuracy of most models, whereas oversampling and undersampling procedures did not have significant effects. The areas under the curve of the best models based on CGM and clinical data with 15 min and 30 min prediction horizons were 0.97 and 0.942, respectively. Basal insulin dose, diabetes duration, proteinuria, and HbA1c were the most important clinical predictors of NH assessed by RF. In conclusion, ML is a promising approach to personalized prediction of NH in hospitalized patients with T1D.

Preferences for Risks and Benefits of Islet Cell Transplantation for Persons With Type 1 Diabetes With History of Episodes of Severe Hypoglycemia: A Discrete-Choice Experiment to Inform Regulatory Decisions

BACKGROUND

The advisory panel for US Food and Drug Administration (FDA) recently endorsed pancreatic islet cell transplantation (ICT) therapy for suboptimally controlled type 1 diabetes (T1D), and FDA approval is under consideration. An important part of regulatory approval includes the patient perspective, through discrete choice. We developed a discrete-choice instrument and used it to determine how 90 people with T1D weigh the risks and benefits of ICT to inform regulatory decisions.

METHODS

Sawtooth software created a random, full-profile, balanced-overlap experimental design for a measure with 8 attributes of ICT risks/benefits, each with 3 to 5 levels. We asked 18 random task pairs, sociodemographics, diabetes management, and hypoglycemia questions. Analysis was performed using random parameters logistic regression technique.

RESULTS

The strongest preference was for avoiding the highest chance (15%) of serious procedure-related complications (β = -2.03, P < 0.001). The strongest positive preference was for gaining 5-y insulin independence (β = 1.75, P < 0.001). The desire for 5-y HbA1C-defined clinical treatment success was also strong (β = 1.39, P < 0.001). Subgroup analysis suggested strong gender differences with women showing much higher preferences for all benefits (68% higher for 5-y insulin independence), and men were generally more risk averse than women. Those with high versus low diabetes distress showed 3 times stronger preference for 5-y insulin independence but also twice preference to avoid risks of serious complications.

CONCLUSION

Despite showing the most preference for avoiding serious ICT complications, people with T1D had a strong preference for achieving ICT benefits, especially insulin independence. We identified important attributes of ICT and demonstrated that patients are willing to make these trade-offs, showing support for the introduction of ICT.

Kernel machine learning methods to handle missing responses with complex predictors. Application in modelling five-year glucose changes using distributional representations

BACKGROUND AND OBJECTIVES

Missing data is a ubiquitous problem in longitudinal studies due to the number of patients lost to follow-up. Kernel methods have enriched the machine learning field by successfully managing non-vectorial predictors, such as graphs, strings, and probability distributions, and have emerged as a promising tool for the analysis of complex data stemming from modern healthcare. This paper proposes a new set of kernel methods to handle missing data in the response variables. These methods will be applied to predict long-term changes in glycated haemoglobin (A1c), the primary biomarker used to diagnose and monitor the progression of diabetes mellitus, making emphasis on exploring the predictive potential of continuous glucose monitoring (CGM).

METHODS

We propose a new framework of non-linear kernel methods for testing statistical independence, selecting relevant predictors, and quantifying the uncertainty of the resultant predictive models. As a novelty in the clinical analysis, we used a distributional representation of CGM as a predictor and compared its performance with that of traditional diabetes biomarkers.

RESULTS

The results show that, after the incorporation of CGM information, predictive ability increases from R²=0.61 to R²=0.71. In addition, uncertainty analysis is useful for characterising some subpopulations where predictivity is worsened, and a more personalised clinical follow-up is advisable according to expected patient uncertainty in glucose values.

CONCLUSIONS

The proposed methods have proven to deal effectively with missing data. They also have the potential to improve the results of predictive tasks by including new complex objects as explanatory variables and modelling arbitrary dependence relations. The application of these methods to a longitudinal study of diabetes showed that the inclusion of a distributional representation of CGM data provides greater sensitivity in predicting five-year A1c changes than classical diabetes biomarkers and traditional CGM metrics.

Impact of anxiety, depression and disease-related distress on long-term glycaemic variability among subjects with Type 1 diabetes mellitus

BACKGROUND

Anxiety, depression, and disease-related distress are linked to worse overall glycaemic control, in terms of HbA1c. This study was aimed to evaluate whether traits of these emotional disorders are associated with long-term glycaemic variability in subjects with Type 1 diabetes.

METHODS

Longitudinal retrospective study. Six-year HbA1c data (2014-2019) from 411 subjects with Type 1 diabetes who had participated in a previous study to design a diabetes-specific quality of life questionnaire in the year 2014 were included. Scores for Spanish versions of the Hospital Anxiety and Depression Scale (HADS) and Problem Areas in Diabetes (PAID) scale were obtained at baseline, along with sociodemographic and clinical data. Long-term glycaemic variability was measured as the coefficient of variation of HbA1c (HbA1c-CV). The association between HADS and PAID scores and HbA1c-CV was analysed with Spearman correlations and multiple regression models, both linear and additive, including other covariates (age, sex, diabetes duration time, type of treatment, baseline HbA1c, use of anxiolytic or antidepressant drugs, education level and employment status).

RESULTS

Scores of depression, anxiety and distress were positively and significantly correlated to HbA1c-CV in univariate analyses. Multiple regression study demonstrated an independent association only for diabetes distress score (p < 0.001). Age, diabetes duration time, baseline HbA1c, education level and employment status were also significantly associated with HbA1c-CV. However, when subjects were analyzed separately in two age groups, distress scores were associated with HbA1c-CV only among those aged 25 years or older, while anxiety scores, but not distress, were associated with HbA1c-CV among those younger than 25 years.

CONCLUSIONS

Psychological factors, particularly disease-related distress and anxiety, are associated with long-term glycaemic variability in subjects with Type 1 diabetes.

Large-Scale Data Analysis for Glucose Variability Outcomes with Open-Source Automated Insulin Delivery Systems

Open-source automated insulin delivery (AID) technologies use the latest continuous glucose monitors (CGM), insulin pumps, and algorithms to automate insulin delivery for effective diabetes management. Early community-wide adoption of open-source AID, such as OpenAPS, has motivated clinical and research communities to understand and evaluate glucose-related outcomes of such user-driven innovation. Initial OpenAPS studies include retrospective studies assessing high-level outcomes of average glucose levels and HbA1c, without in-depth analysis of glucose variability (GV). The OpenAPS Data Commons dataset, donated to by open-source AID users with insulin-requiring diabetes, is the largest freely available diabetes-related dataset with over 46,070 days’ worth of data and over 10 million CGM data points, alongside insulin dosing and algorithmic decision data. This paper first reviews the development toward the latest open-source AID and the performance of clinically approved GV metrics. We evaluate the GV outcomes using large-scale data analytics for the n = 122 version of the OpenAPS Data Commons. We describe the data cleaning processes, methods for measuring GV, and the results of data analysis based on individual self-reported demographics. Furthermore, we highlight the lessons learned from the GV outcomes and the analysis of a rich and complex diabetes dataset and additional research questions that emerged from this work to guide future research. This paper affirms previous studies’ findings of the efficacy of open-source AID.

Glycaemia dynamics in gestational diabetes mellitus

Pregnant women may develop gestational diabetes mellitus (GDM), a disease of pregnancy characterised by maternal and fetal hyperglycaemia with hazardous consequences to the mother, the fetus, and the newborn. Maternal hyperglycaemia in GDM results in fetoplacental endothelial dysfunction. GDM-harmful effects result from chronic and short periods of hyperglycaemia. Thus, it is determinant to keep glycaemia within physiological ranges avoiding short but repetitive periods of hyper or hypoglycaemia. The variation of glycaemia over time is defined as 'glycaemia dynamics'. The latter concept regards with a variety of mechanisms and environmental conditions leading to blood glucose handling. In this review we summarized the different metrics for glycaemia dynamics derived from quantitative, plane distribution, amplitude, score values, variability estimation, and time series analysis. The potential application of the derived metrics from self-monitoring of blood glucose (SMBG) and continuous glucose monitoring (CGM) in the potential alterations of pregnancy outcome in GDM are discussed.

Carbohydrate-Energy Replacement Following High-Intensity Interval Exercise Blunts Next-Day Glycemic Control in Untrained Women

Background

Improved glycemic control has been reported for ∼24 h following low-volume high-intensity interval exercise (HIIE), but it is unclear if this is a direct effect of exercise or an indirect effect of the exercise-induced energy deficit. The purpose of this study was to investigate the effect of carbohydrate-energy replacement after low-volume HIIE on 24 h glycemic control in women.

Methods

Seven untrained women (age: 22 ± 2 yr; BMI: 22 ± 3 kg/m2; VO2peak: 33 ± 7 ml/kg/min) completed three 2-day trials in the mid-follicular phase of the menstrual cycle. Continuous glucose monitoring was used to measure blood glucose concentrations during, and for 24 h following three conditions: (1) HIIE followed by a high-carbohydrate energy replacement drink (EX-HC); (2) HIIE followed by a non-caloric taste-matched placebo drink (EX-NC); and (3) seated control with no drink (CTL). HIIE involved an evening session (1,700 h) of 10 × 1-min cycling efforts at ∼90% maximal heart rate with 1 min recovery. Diet was standardized and identical across all three 2-day trials, apart from the post-exercise carbohydrate drink in EX-HC, which was designed to replenish the exercise-induced energy expenditure. Postprandial glycemic responses to the following days breakfast, snack, lunch, and dinner, as well as 24 h indices of glycemic control, were analyzed.

Results

The day after HIIE, postprandial glycemia following breakfast and snack were reduced in EX-NC compared to EX-HC, as reflected by lower 3 h glucose mean (breakfast: 5.5 ± 0.5 vs. 6.7 ± 1, p = 0.01, Cohen’s d = 1.4; snack: 4.9 ± 0.3 vs. 5.7 ± 0.8 mmol/L, p = 0.02, d = 1.4) and/or area under the curve (AUC) (breakfast: 994 ± 86 vs. 1,208 ± 190 mmol/L x 3 h, p = 0.01, d = 1.5). Postprandial glycemic responses following lunch and dinner were not different across conditions (p &gt; 0.05). The 24 h glucose mean (EX-NC: 5.2 ± 0.3 vs. EX-HC: 5.7 ± 0.7 mmol/L; p = 0.02, d = 1.1) and AUC (EX-NC: 7,448 ± 425 vs. EX-HC: 8,246 ± 957 mmol/L × 24 h; p = 0.02, d = 1.1) were reduced in EX-NC compared to EX-HC.

Conclusion

Post-exercise carbohydrate-energy replacement attenuates glycemic control the day following a single session of low-volume HIIE in women.

Glycemic variability assessed by continuous glucose monitoring in hospitalized patients with community-acquired pneumonia

Background

Glycemic variability (GV) has only been sparsely studied in patients with community-acquired pneumonia (CAP). This study aimed to quantify in-hospital GV in CAP patients, including determining the impact of type 2 diabetes mellitus (T2DM) and glucocorticoid (GC) treatment on GV.

Methods

This is a prospective cohort study of CAP patients (N = 40) with or without T2DM and treated or not with GCs. The primary endpoint was GV measured as glucose standard deviation (SD), coefficient of variation (CV), and postprandial glucose excursions (PPGE) based on continuous glucose monitoring (CGM). Analysis of glucose data was split into daytime and nighttime when possible.

Results

Patients included had a mean age of 74 (range 55 to 91) years. SD (95%CI) increased by a factor of 1.93 (1.40 to 2.66) and 2.29 (1.38 to 3.81) in patients with T2DM and not treated with GCs during the daytime and the nighttime, respectively (both P < 0.01), and by a factor of 1.42 (1.04 to 1.97) in patients treated with GCs but without T2DM during the daytime (P = 0.031) compared to patients without T2DM and not treated with GCs. CV (95%CI) increased by 5.1 (0.0 to 10.1) and 8.1 (1.0 to 15.2) percentage points during the daytime and the nighttime, respectively, in patients with T2DM and not treated with GCs compared to patients without T2DM and not treated with GCs (P = 0.046 and P = 0.026, respectively). PPGE (95% CI) increased during lunch by 2.5 (0.7 to 4.3) mmol/L (45 (13 to 77) mg/dL) in patients with T2DM and treated with GCs compared to patients without T2DM and not treated with GCs (P = 0.018).

Conclusions

CAP patients receiving GCs, especially those with T2DM, are at great risk of developing high GV and therefore require clinical attention to mitigate GV. This applies particularly during the daytime. Results support the 1 to 2-h post-lunch screening procedure for glucocorticoid-induced hyperglycemia in patients without diabetes. SD was positively correlated with hospital length of stay.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12890-022-01874-7.