A Simple Composite Metric for the Assessment of Glycemic Status from Continuous Glucose Monitoring Data: Implications for Clinical Practice and the Artificial Pancreas

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.

Simulating Realistic Continuous Glucose Monitor Time Series By Data Augmentation

BACKGROUND

Simulated data are a powerful tool for research, enabling benchmarking of blood glucose (BG) forecasting and control algorithms. However, expert created models provide an unrealistic view of real-world performance, as they lack the features that make real data challenging, while black-box approaches such as generative adversarial networks do not enable systematic tests to diagnose model performance.

METHODS

To address this, we propose a method that learns missingness and error properties of continuous glucose monitor (CGM) data collected from people with type 1 diabetes (OpenAPS, OhioT1DM, RCT, and Racial-Disparity), and then augments simulated BG data with these properties. On the task of BG forecasting, we test how well our method brings performance closer to that of real CGM data compared with current simulation practices for missing data (random dropout) and error (Gaussian noise, CGM error model).

RESULTS

Our methods had the smallest performance difference versus real data compared with random dropout and Gaussian noise when individually testing the effects of missing data and error on simulated BG in most cases. When combined, our approach was significantly better than Gaussian noise and random dropout for all data sets except OhioT1DM. Our error model significantly improved results on diverse data sets.

CONCLUSIONS

We find a significant gap between BG forecasting performance on simulated and real data, and our method can be used to close this gap. This will enable researchers to rigorously test algorithms and provide realistic estimates of real-world performance without overfitting to real data or at the expense of data collection.

Continuous glucose monitoring-based composite metrics: a review and assessment of performance in recent-onset and long-duration type 1 diabetes

This study examined correlations between continuous glucose monitoring (CGM)-based composite metrics and standard glucose metrics within CGM data sets from individuals with recent-onset and long-duration type 1 diabetes. First, a literature review and critique of published CGM-based composite metrics was undertaken. Second, composite metric results were calculated for the two CGM data sets and correlations with six standard glucose metrics were examined. Fourteen composite metrics met selection criteria; these metrics focused on overall glycemia (n = 8), glycemic variability (n = 4), and hypoglycemia (n = 2), respectively. Results for the two diabetes cohorts were similar. All eight metrics focusing on overall glycemia strongly correlated with glucose time in range; none strongly correlated with time below range. The eight overall glycemia-focused and two hypoglycemia-focused composite metrics were all sensitive to automated insulin delivery therapeutic intervention. Until a composite metric can adequately capture both achieved target glycemia and hypoglycemia burden, the current two-dimensional CGM assessment approach may offer greatest clinical utility.

Software Packages and Tools for the Analysis of Continuous Glucose Monitoring Data

The advancement of technology in the field of glycemic control has led to the widespread use of continuous glucose monitoring (CGM), which can be nowadays obtained from wearable devices equipped with a minimally invasive sensor, that is, transcutaneous needle type or implantable, and a transmitter that sends information to a receiver or smart device for data storage and display. This work aims to review the currently available software packages and tools for the analysis of CGM data. Based on the purposes of this work, 12 software packages have been identified from the literature, published until December 2021, namely: GlyCulator, EasyGV (Easy Glycemic Variability), CGM-GUIDE^© (Continuous Glucose Monitoring Graphical User Interface for Diabetes Evaluation), GVAP (Glycemic Variability Analyzer Program), Tidepool, CGManalyzer, cgmanalysis, GLU, CGMStatsAnalyser, iglu, rGV, and cgmquantify. Comparison of available software packages and tools has been done in terms of main characteristics (i.e., publication year, presence of a graphical user interface, availability, open-source code, number of citations, programming language, supported devices, supported data format and organization of the data structure, documentation, presence of a toy example, video tutorial, data upload and download, measurement-units conversion), preprocessing procedures, data display options, and computed metrics; also, each of the computed metrics has been analyzed in terms of its adherence to the American Diabetes Association (ADA) 2017 international consensus on CGM data analysis and the ADA 2019 international consensus on time in range. Eventually, the agreement between metrics computed by different software and tools has been investigated. Based on such comparison, usability and complexity of data management, as well as the possibility to perform customized or patients-group analyses, have been discussed by highlighting limitations and strengths, also in relation to possible different user categories (i.e., patients, clinicians, researchers). The information provided could be useful to researchers interested in working in the diabetic research field as to clinicians and endocrinologists who need tools capable of handling CGM data effectively.

Glycaemic management in diabetes: old and new approaches

HbA_1c is the most used parameter to assess glycaemic control. However, evidence suggests that the concept of hyperglycaemia has profoundly changed and that different facets of hyperglycaemia must be considered. A modern approach to glycaemic control should focus not only on reaching and maintaining optimal HbA_1c concentrations as early as possible, but to also do so by reducing postprandial hyperglycaemia, glycaemic variability, and to extend as much as possible the time in range in near-normoglycaemia. These goals should be achieved while avoiding hypoglycaemia, which, should it occur, should be reverted to normoglycaemia. Modern technology, such as intermittently scanned glucose monitoring and continuous glucose monitoring, together with new drug therapies (eg, ultra-fast insulins, SGLT2 inhibitors, and GLP-1 receptor agonists), could help to change the landscape of glycaemia management based on HbA_1c in favour of a more holistic approach that considers all the different aspects of this commonly oversimplified pathophysiological feature of diabetes.

Towards a Rational Balanced Pancreatic and Islet Allocation Schema

Allocation of donated organs for transplantation is a complex process that considers numerous factors such as donor, organ and candidate characteristics and practical issues such as geography. Whole pancreas and isolated islet transplantation are lifesaving for certain individuals with diabetes. Herein, we suggest a revised allocation schema that matches donor characteristics with candidate medical condition while allowing for geographic considerations. It is hoped that adoption of this schema will shorten allocation time, decrease organ waste and optimize the parity between organ donor characteristics and candidate state of health.

Quality of Glycemic Control: Assessment Using Relationships Between Metrics for Safety and Efficacy

Numerous methods have been proposed as measures of quality of glycemic control resulting in confusion regarding the best choice of metric to use by clinicians and researchers. Some methods use a single metric such as HbA1c, Mean Glucose, %Time In Range (%TIR), or Coefficient of Variation (%CV). Others use a combination of up to seven metrics, for example, Q-Score, Comprehensive Glucose Pentagon (CGP), and Personal Glycemic State (PGS). A recently proposed Composite continuous Glucose monitoring index utilizes three metrics: %TIR, Time Below Range (%TBR), and standard deviation (SD) of glucose. This review proposes that only two metrics can be sufficient when monitoring an individual patient or when comparing two or more forms of management interventions. These two metrics comprise (1) a measure of efficacy such as Mean Glucose, HbA1c, %TIR, or %Time Above Range (%TAR) and (2) a measure of safety based on risk of hypoglycemia such as %TBR, Low Blood Glucose Index (LBGI), or frequency of specified types of hypoglycemic events per patient year. By analysis of the two-dimensional graphical and statistical relationships between metrics for safety and efficacy and by testing identity versus nonidentity of these relationships, one can improve sensitivity for detection of the effects of medications and of other therapeutic interventions, avoid the need for arbitrary scoring systems for glucose values falling within versus outside the target range, and offer the advantage of conceptual and practical simplicity.

The Ambulatory Glucose Profile: Opportunities for Enhancement

The ambulatory glucose profile (AGP) and the frequency distribution for glucose by ranges are well established as standard methods for display, analysis, and interpretation of glucose data arising from self-monitoring, continuous glucose monitoring, and automated insulin delivery systems. In this review, we consider several refinements that may further improve the utility of the AGP. These include (1) display of the AGP together with information regarding dietary intake, medication administration (e.g., insulin), glucose lowering (pharmacodynamic) activity of medications, and physical activity measured by accelerometers or heart rate; (2) display of average time below range (%TBR), time above range (%TAR), and time in range (%TIR) by time of day to indicate timing of hypoglycemic and hyperglycemic episodes; (3) detailed analysis of postprandial excursions for each of the major meals after synchronizing by onset of meals and adjusting for the premeal glucose levels, enabling comparisons of magnitude, shape, and patterns; (4) methods to characterize distinct patterns on different days of the week; (5) display of glucose on a nonlinear scale to improve the balance between deviations associated with hypoglycemia versus hyperglycemia; (6) use of time scales other than midnight-to-midnight to facilitate analysis of time segments of particular interest (e.g., overnight); (7) options to display individual glucose values to assist in the identification of dates and times of outliers and episodes of hypoglycemia and hyperglycemia; and (8) methods to compare AGPs obtained from different individuals or groups receiving alternative interventions in terms of therapy or technology. These refinements, individually or collectively, can potentially further enhance the effectiveness of the AGP for assessment of glucose levels, patterns, and variability. We discuss several questions regarding implementation and optimization of these methods.

Standardized Hybrid Closed-Loop System Reporting

The hybrid closed-loop (HCL) system has been shown to improve glycemic control and reduce hypoglycemia. Optimization of HCL settings requires interpretation of the glucose, insulin, and factors affecting glucose such as food intake and exercise. To the best of our knowledge, there is no published guidance on the standardized reporting of HCL systems. Standardization of HCL reporting would make interpretation of data easy across different systems. We reviewed the literature on patient and provider perspectives on downloading and reporting glucose metric preferences. We also incorporated international consensus on standardized reporting for glucose metrics. We describe a single-page HCL data reporting, referred to here as "artificial pancreas (AP) Dashboard." We propose seven components in the AP Dashboard that can provide detailed information and visualization of glucose, insulin, and HCL-specific metrics. The seven components include (A) glucose metrics, (B) hypoglycemia, (C) insulin, (D) user experience, (E) hyperglycemia, (F) glucose modal-day profile, and (G) insight. A single-page report similar to an electrocardiogram can help providers and patients interpret HCL data easily and take the necessary steps to improve glycemic outcomes. We also describe the optimal sampling duration for HCL data download and color coding for visualization ease. We believe that this is a first step in creating a standardized HCL reporting, which may result in better uptake of the systems. For increased adoption, standardized reporting will require input from providers, patients, diabetes device manufacturers, and regulators.

Glycemic deviation index: a novel method of integrating glycemic numerical value and variability

BACKGROUND

There are many continuous blood glucose monitoring (CGM) data-based indicators, and most of these focus on a single characteristic of abnormal blood glucose. An ideal index that integrates and evaluates multiple characteristics of blood glucose has not yet been established.

METHODS

In this study, we proposed the glycemic deviation index (GDI) as a novel integrating characteristic, which mainly incorporates the assessment of the glycemic numerical value and variability. To verify its effectiveness, GDI was applied to the simulated 24 h glycemic profiles and the CGM data of type 2 diabetes (T2D) patients (n = 30).

RESULTS

Evaluation of the GDI of the 24 h simulated glycemic profiles showed that the occurrence of hypoglycemia was numerically the same as hyperglycemia in increasing GDI. Meanwhile, glycemic variability was added as an independent factor. One-way ANOVA results showed that the application of GDI showed statistically significant differences in clinical glycemic parameters, average glycemic parameters, and glycemic variability parameters among the T2D groups with different glycemic levels.

CONCLUSIONS

In conclusion, GDI integrates the characteristics of the numerical value and the variability in blood glucose levels and may be beneficial for the glycemic management of diabetic patients undergoing CGM treatment.

Making sense of glucose metrics in diabetes: linkage between postprandial glucose (PPG), time in range (TIR) & hemoglobin A1c (A1C)

While A1C is the standard diagnostic test for evaluating long-term glucose management, additional glucose data, either from fingerstick blood glucose testing, or more recently, continuous glucose monitoring (CGM), is necessary for safe and effective management of diabetes, especially for individuals treated with insulin. CGM technology and retrospective pattern-based management using various CGM reports have the potential to improve glycemic management beyond what is possible with fingerstick blood glucose monitoring. CGM software can provide valuable retrospective data on Time-in-Ranges (above, below, within) metrics, the Ambulatory Glucose Profile (AGP), overlay reports, and daily views for persons with diabetes and their healthcare providers. This data can aid in glycemic pattern identification and evaluation of the impact of lifestyle factors on these patterns. Time-in-Ranges data provide an easy-to-define metric that can facilitate goal setting discussions between clinicians and persons with diabetes to improve glycemic management and can empower persons with diabetes in self-management between clinic consultation visits. Here we discuss multiple real-life scenarios from a primary care clinic for the application of CGM in persons with diabetes. Optimizing the use of the reports generated by CGM software, with attention to time in range, time below range, and postprandial glucose-induced time above range, can improve the safety and efficacy of ongoing glucose management.

Updated Software for Automated Assessment of Glucose Variability and Quality of Glycemic Control in Diabetes

Background: Glycemic variability is an important factor to consider in diabetes management. It can be assessed with multiple glycemic variability metrics and quality of control indices based on continuous glucose monitoring (CGM) recordings. For this, a robust repeatable calculation is important. A widely used tool for automated assessment is the EasyGV software. The aim of this work is to implement new methods of glycemic variability assessment in EasyGV and to validate implementation of each glucose metric in EasyGV against a reference implementation of the calculations. Methods: Validation data used came from the JDRF CGM study. Validation of the implementation of metrics that are available in EasyGV software v9 was carried out and the following new methods were added and validated: personal glycemic state, index of glycemic control, times in ranges, and glycemic variability percentage. Reference values considered gold standard calculations were derived from MATLAB implementation of each metric. Results: The Pearson correlation coefficient was above 0.98 for all metrics, except for mean amplitude of glycemic excursion (r = 0.87) as EasyGV implements a fuzzy logic approach to assessment of variability. Bland-Altman plots demonstrated validation of the new software. Conclusions: The new freely available EasyGV software v10 (www.phc.ox.ac.uk/research/technology-outputs/easygv) is a validated robust tool for analyzing different glycemic variabilities and control metrics.

Assessment of Glucose Control Metrics by Discriminant Ratio

Objective: Increasing use of continuous glucose monitoring (CGM) data has created an array of glucose metrics for glucose variability, temporal patterns, and times in ranges. However, a gold standard metric has not been defined. We assess the performance of multiple glucose metrics to determine their ability to detect intra- and interperson variability to determine a set of recommended metrics. Methods: The Juvenile Diabetes Research Foundation data set, a randomized controlled study of CGM and self-monitored blood glucose conducted in children and adults with type 1 diabetes (T1D), was used. To determine the ability of the evaluated glycemic metrics to discriminate between different subjects and attenuate the effect of within-subject variation, the discriminant ratio was calculated and compared for each metric. Then, the findings were confirmed using data from two other recent randomized clinical trials. Results: Mean absolute glucose (MAG) has the highest discriminant ratio value (2.98 [95% confidence interval {CI} 1.64-3.67]). In addition, low blood glucose index and index of glycemic control performed well (1.93 [95% CI 1.15-3.44] and 1.92 [95% CI 1.27-2.93], respectively). For percentage times in glucose target ranges, the optimal discriminator was percentage time in glucose target 70-180 mg/dL. Conclusions: MAG is the optimal index to differentiate glucose variability in people with T1D, and may be a complementary therapeutic monitoring tool in addition to glycated hemoglobin and a measure of hypoglycemia. Percentage time in glucose target 70-180 mg/dL is the optimal percentage time in range to report.

A Review of Continuous Glucose Monitoring-Based Composite Metrics for Glycemic Control

We performed a literature review of composite metrics for describing the quality of glycemic control, as measured by continuous glucose monitors (CGMs). Nine composite metrics that describe CGM data were identified. They are described in detail along with their advantages and disadvantages. The primary benefit to using composite metrics in clinical practice is to be able to quickly evaluate a patient's glycemic control in the form of a single number that accounts for multiple dimensions of glycemic control. Very little data exist about (1) how to select the optimal components of composite metrics for CGM; (2) how to best score individual components of composite metrics; and (3) how to correlate composite metric scores with empiric outcomes. Nevertheless, composite metrics are an attractive type of scoring system to present clinicians with a single number that accounts for many dimensions of their patients' glycemia. If a busy health care professional is looking for a single-number summary statistic to describe glucose levels monitored by a CGM, then a composite metric has many attractive features.

Glucose Time In Range, Time Above Range, and Time Below Range Depend on Mean or Median Glucose or HbA1c, Glucose Coefficient of Variation, and Shape of the Glucose Distribution

Background: Examine the expected relationships between time in range (%TIR), time above range (%TAR), and time below range (%TBR) with median glucose (or %HbA1c) and %coefficient of variation (%CV) of glucose for various shapes of the glucose distribution. Methods: We considered several thresholds defining hypoglycemia and hyperglycemia and examined wide ranges of median glucose and %CV using three models for the glucose distribution: gaussian, log-gaussian, and a modified log-gaussian distribution. Results: There is a linear relationship between %TIR and median glucose for any specified %CV when median glucose is well removed from the threshold for hypoglycemia. %TIR reaches a peak when median glucose is close to 120 mg/dL and declines both at higher and lower median glucose values. There is a nearly linear relationship for %TAR and median glucose for a wider range of glucose (80-220 mg/dL). Risk of hypoglycemia is minimal when %CV is below 20%, but rises exponentially as %CV increases or as median glucose decreases. Similar results were obtained for a wide range of possible shapes of glucose distribution. These simulations are consistent with results from clinical studies. Conclusion: Both %TIR and %TAR are approximately linearly related to mean and median glucose (or %HbA1c). %TAR provides linearity over a wider range than %TIR. Risk of hypoglycemia (%TBR) is critically dependent on both glycemic variability (%CV) and mean or median glucose. These relationships support the use of %TIR, %TAR, and %TBR as metrics of quality of glycemic control for clinical, research, and regulatory purposes.

Diabetes Healthcare Professionals Use Multiple Continuous Glucose Monitoring Data Indicators to Assess Glucose Management

BACKGROUND

Continuous glucose monitoring (CGM) offers multiple data features that can be leveraged to assess glucose management. However, how diabetes healthcare professionals (HCPs) actually assess CGM data and the extent to which they agree in assessing glycemic management are not well understood.

METHODS

We asked HCPs to assess ten de-identified CGM datasets (each spanning seven days) and rank order each day by relative glycemic management (from "best" to "worst"). We also asked HCPs to endorse features of CGM data that were important in making such assessments.

RESULTS

In the study, 57 HCPs (29 endocrinologists; 28 diabetes educators) participated. Hypoglycemia and glycemic variance were endorsed by nearly all HCPs to be important (91% and 88%, respectively). Time in range and daily lows and highs were endorsed more frequently by educators (all Ps < .05). On average, HCPs endorsed 3.7 of eight data features. Overall, HCPs demonstrated agreement in ranking days by relative glycemic control (Kendall's W = .52, P < .001). Rankings were similar between endocrinologists and educators (R² = .90, Cohen's kappa = .95, mean absolute error = .4 [all Ps < .05]; Mann-Whitney U = 41, P = .53).

CONCLUSIONS

Consensus in the endorsement of certain data features and agreement in assessing glycemic management were observed. While some practice-specific differences in feature endorsement were found, no differences between educators and endocrinologists were observed in assessing glycemic management. Overall, HCPs tended to consider CGM data holistically, in alignment with published recommendations, and made converging assessments regardless of practice.

Role of Composite Glycemic Indices: A Comparison of the Comprehensive Glucose Pentagon Across Diabetes Types and HbA1c Levels

Background: Complex changes of glycemia that occur in diabetes are not fully captured by any single measure. The Comprehensive Glucose Pentagon (CGP) measures multiple aspects of glycemia to generate the prognostic glycemic risk (PGR), which constitutes the relative risk of hypoglycemia combined with long-term complications. We compare the components of CGP and PGR across type 1 and type 2 diabetes. Methods: Participants: n = 60 type 1 and n = 100 type 2 who underwent continuous glucose monitoring (CGM). Mean glucose, coefficient of variation (%CV), intensity of hypoglycemia (INT_hypo), intensity of hyperglycemia (INT_hyper), time out-of-range (TOR <3.9 and >10 mmol/L), and PGR were calculated. PGR (median, interquartile ranges [IQR]) for diabetes types, and HbA1c classes were compared. Results: While HbA1c was lower in type 1 (type 1 vs. type 2: 8.0 ± 1.6 vs. 8.6 ± 1.7, P = 0.02), CGM-derived mean glucoses were similar across both groups (P > 0.05). TOR, %CV, INT_hypo, and INT_hyper were all higher in type 1 [type 1 vs. type 2: 665 (500, 863) vs. 535 (284, 823) min/day; 39% (33, 46) vs. 29% (24, 34); 905 (205, 2951) vs. 18 (0, 349) mg/dL × min²; 42,906 (23,482, 82,120) vs. 30,166 (10,276, 57,183) mg/dL × min², respectively, all P < 0.05]. Across each HbA1c class, the PGR remained consistently and significantly higher in type 1. While mean glucose remained the same across HbA1c classes, %CV, TOR, INT_hyper, and INT_hypo were significantly higher for type 1. Even within the same HbA1c class, the variation (IQR) of each parameter in type 1 was wider. The PGR increased across diabetes groups; type 2 on orals versus type 2 on insulin versus type 1 (PGR: 1.6 vs. 2.2 vs. 2.9, respectively, P < 0.05). Conclusion: Composite indices such as the CGP capture significant differences in glycemia independent of HbA1c and mean glucose. The use of such indices must be explored in both the clinical and research settings.

Patterns of Glycemic Variability During a Diabetes Self-Management Educational Program

Background: Variations in blood glucose levels over a given time interval is termed as glycemic variability (GV). Higher GV is associated with higher diabetes-related complications. The current study was done with the aim of detecting the sensitivity of various GV indices among individuals with type 2 diabetes mellitus of different glycemic control status. Methods: We performed a longitudinal study among individuals with type 2 diabetes mellitus (T2DM) who were participating in a two-week diabetes self-management education (DSME) program. Participants were categorized by their HbA1c as poor (≥8%), acceptable (7%–8%), and optimal control (<7%). Continuous glucose monitoring (CGM) sensors recorded interstitial glucose every 15 min from day 1. The evaluated GV measures include standard deviation (SD), coefficient of variation (CV), mean amplitude of glycemic excursion (MAGE), continuous overlapping net glycemic action (CONGA), mean of daily difference for inter-day variation (MODD), high blood glucose index (HBGI), and low blood glucose index (LBGI). Results: A total of 41 study participants with 46347 CGM values were available for analysis. Of 41 participants, 20 (48.7%) were in the poor, 10 (24.3%) in the acceptable, and 11 (26.8%) in the optimal control group. The GV indices (SD; CV; MODD; MAGE; CONGA; HBGI) of poorly controlled (77.43; 38.02; 45.82; 216.63; 14.10; 16.62) were higher than acceptable (50.02; 39.32; 30.79; 138.01; 8.87; 5.56) and optimal (34.15; 29.46; 24.56; 126.15; 8.67; 3.13) control group. Glycemic variability was reduced in the poorly and acceptably controlled groups by the end of the 2-week period. There was a rise in LBGI in the optimally controlled group, indicating pitfalls of tight glycemic control. Conclusion: Indices of glycemic variability are useful complements, and changes in it can be demonstrated within short periods.

Glucose Variability: A Review of Clinical Applications and Research Developments

Glycemic variability (GV) is a major consideration when evaluating quality of glycemic control. GV increases progressively from prediabetes through advanced T2D and is still higher in T1D. GV is correlated with risk of hypoglycemia. The most popular metrics for GV are the %Coefficient of Variation (%CV) and standard deviation (SD). The %CV is correlated with risk of hypoglycemia. Graphical display of glucose by date, time of day, and day of the week, and display of simplified glucose distributions showing % of time in several ranges, provide clinically useful indicators of GV. SD is highly correlated with most other measures of GV, including interquartile range, mean amplitude of glycemic excursion, mean of daily differences, and average daily risk range. Some metrics are sensitive to the frequency, periodicity, and complexity of glycemic fluctuations, including Fourier analysis, periodograms, frequency spectrum, multiscale entropy (MSE), and Glucose Variability Percentage (GVP). Fourier analysis indicates progressive changes from normal subjects to children and adults with T1D, and from prediabetes to T2D. The GVP identifies novel characteristics for children, adolescents, and adults with type 1 diabetes and for adults with type 2. GVP also demonstrated small rapid glycemic fluctuations in people with T1D when using a dual-hormone closed-loop control. MSE demonstrated systematic changes from normal subjects to people with T2D at various stages of duration, intensity of therapy, and quality of glycemic control. We describe new metrics to characterize postprandial excursions, day-to-day stability of glucose patterns, and systematic changes of patterns by day of the week. Metrics for GV should be interpreted in terms of percentiles and z-scores relative to identified reference populations. There is a need for large accessible databases for reference populations to provide a basis for automated interpretation of GV and other features of continuous glucose monitoring records.

Metrics to Evaluate Quality of Glycemic Control: Comparison of Time in Target, Hypoglycemic, and Hyperglycemic Ranges with "Risk Indices"

OBJECTIVE

We sought to cross validate several metrics for quality of glycemic control, hypoglycemia, and hyperglycemia.

RESEARCH DESIGN AND METHODS

We analyzed the mathematical properties of several metrics for overall glycemic control, and for hypo- and hyperglycemia, to evaluate their similarities, differences, and interrelationships. We used linear regression to describe interrelationships and examined correlations between metrics within three conceptual groups.

RESULTS

There were consistently high correlations between %Time in range (%TIR) and previously described risk indices (M₁₀₀, Blood Glucose Risk Index [BGRI], Glycemic Risk Assessment Diabetes Equation [GRADE], Index of Glycemic Control [IGC]), and with J-Index (J). There were also high correlations among %Hypoglycemia, Low Blood Glucose Index (LBGI), percentage of GRADE attributable to hypoglycemia (GRADE_{%Hypoglycemia}), and Hypoglycemia Index, but negligible correlation with J. There were high correlations of percentage of time in hyperglycemic range (%Hyperglycemia) with High Blood Glucose Index (HBGI), percentage of GRADE attributable to hyperglycemia (GRADE_{%Hyperglycemia}), Hyperglycemia Index, and J. %TIR is highly negatively correlated with %Hyperglycemia but very weakly correlated with %Hypoglycemia. By adjusting the parameters used in IGC, Hypoglycemia Index, Hyperglycemia Index, or in M_R, one can more closely approximate the properties of BGRI, LBGI, or HBGI, and of GRADE, GRADE_{%Hypoglycemia}, or GRADE_{%Hyperglycemia}.

CONCLUSIONS

Simple readily understandable criteria such as %TIR, %Hypoglycemia, and %Hyperglycemia are highly correlated with and appear to be as informative as "risk indices." The J-Index is sensitive to hyperglycemia but insensitive to hypoglycemia.

Beyond HbA1c: Comparing Glycemic Variability and Glycemic Indices in Predicting Hypoglycemia in Type 1 and Type 2 Diabetes

BACKGROUND

Hypoglycemia is the major impediment to therapy intensification in diabetes. Although higher individualized HbA1c targets are perceived to reduce the risk of hypoglycemia in those at risk of hypoglycemia, HbA1c itself is a poor predictor of hypoglycemia. We assessed the use of glycemic variability (GV) and glycemic indices as independent predictors of hypoglycemia.

METHODS

A retrospective observational study of 60 type 1 and 100 type 2 diabetes subjects. All underwent professional continuous glucose monitoring (CGM) for 3-6 days and recorded self-monitored blood glucose (SMBG). Indices were calculated from both CGM and SMBG. Statistical analyses included regression and area under receiver operator curve (AUC) analyses.

RESULTS

Hypoglycemia frequency (53.3% vs. 24%, P < 0.05) and %CV (40.1% ± 10% vs. 29.4% ± 7.8%, P < 0.001) were significantly higher in type 1 diabetes compared with type 2 diabetes. HbA1c was, at best, a weak predictor of hypoglycemia. %CV_CGM, Low Blood Glucose Index (LBGI)_CGM, Glycemic Risk Assessment Diabetes Equation (GRADE)Hypoglycemia_CGM, and Hypoglycemia Index_CGM predicted hypoglycemia well. %CV_CGM and %CV_SMBG consistently remained a robust discriminator of hypoglycemia in type 1 diabetes (AUC 0.88). In type 2 diabetes, a combination of HbA1c and %CV_SMBG or LBGI_SMBG could help discriminate hypoglycemia.

CONCLUSION

Assessment of glycemia should go beyond HbA1c and incorporate measures of GV and glycemic indices. %CV_SMBG in type 1 diabetes and LBGI_SMBG or a combination of HbA1c and %CV_SMBG in type 2 diabetes discriminated hypoglycemia well. In defining hypoglycemia risk using GV and glycemic indices, diabetes subtypes and data source (CGM vs. SMBG) must be considered.

Glycaemic variability-based classification of impaired glucose tolerance vs. type 2 diabetes using continuous glucose monitoring data

Many glycaemic variability (GV) indices extracted from continuous glucose monitoring systems data have been proposed for the characterisation of various aspects of glucose concentration profile dynamics in both healthy and non-healthy individuals. However, the inter-index correlations have made it difficult to reach a consensus regarding the best applications or a subset of indices for clinical scenarios, such as distinguishing subjects according to diabetes progression stage. Recently, a logistic regression-based method was used to address the basic problem of differentiating between healthy subjects and those affected by impaired glucose tolerance (IGT) or type 2 diabetes (T2D) in a pool of 25 GV-based indices. Whereas healthy subjects were classified accurately, the distinction between patients with IGT and T2D remained critical. In the present work, by using a dataset of CGM time-series collected in 62 subjects, we developed a polynomial-kernel support vector machine-based approach and demonstrated the ability to distinguish between subjects affected by IGT and T2D based on a pool of 37 GV indices complemented by four basic parameters-age, sex, BMI, and waist circumference-with an accuracy of 87.1%.

Calculating the Mean Amplitude of Glycemic Excursions from Continuous Glucose Data Using an Open-Code Programmable Algorithm Based on the Integer Nonlinear Method

The mean amplitude of glycemic excursions (MAGE) is an essential index for glycemic variability assessment, which is treated as a key reference for blood glucose controlling at clinic. However, the traditional “ruler and pencil” manual method for the calculation of MAGE is time-consuming and prone to error due to the huge data size, making the development of robust computer-aided program an urgent requirement. Although several software products are available instead of manual calculation, poor agreement among them is reported. Therefore, more studies are required in this field. In this paper, we developed a mathematical algorithm based on integer nonlinear programming. Following the proposed mathematical method, an open-code computer program named MAGECAA v1.0 was developed and validated. The results of the statistical analysis indicated that the developed program was robust compared to the manual method. The agreement among the developed program and currently available popular software is satisfied, indicating that the worry about the disagreement among different software products is not necessary. The open-code programmable algorithm is an extra resource for those peers who are interested in the related study on methodology in the future.

Glycemic Variability Percentage: A Novel Method for Assessing Glycemic Variability from Continuous Glucose Monitor Data

BACKGROUND

High levels of glycemic variability are still observed in most patients with diabetes with severe insulin deficiency. Glycemic variability may be an important risk factor for acute and chronic complications. Despite its clinical importance, there is no consensus on the optimum method for characterizing glycemic variability.

METHOD

We developed a simple new metric, the glycemic variability percentage (GVP), to assess glycemic variability by analyzing the length of the continuous glucose monitoring (CGM) temporal trace normalized to the duration under evaluation. The GVP is similar to other recently proposed glycemic variability metrics, the distance traveled, and the mean absolute glucose (MAG) change. We compared results from distance traveled, MAG, GVP, standard deviation (SD), and coefficient of variation (CV) applied to simulated CGM traces accentuating the difference between amplitude and frequency of oscillations. The GVP metric was also applied to data from clinical studies for the Dexcom G4 Platinum CGM in subjects without diabetes, with type 2 diabetes, and with type 1 diabetes (adults, adolescents, and children).

RESULTS

In contrast to other metrics, such as CV and SD, the distance traveled, MAG, and GVP all captured both the amplitude and frequency of glucose oscillations. The GVP metric was also able to differentiate between diabetic and nondiabetic subjects and between subjects with diabetes with low, moderate, and high glycemic variability based on interquartile analysis.

CONCLUSION

A new metric for the assessment of glycemic variability has been shown to capture glycemic variability due to fluctuations in both the amplitude and frequency of glucose given by CGM data.

Application of Glycemic Variability Percentage: Implications for Continuous Glucose Monitor Utilization and Analysis of Artificial Pancreas Data

BACKGROUND

The problem of glycemic variability has been widely acknowledged in patients with diabetes with severe insulin deficiency. In a companion article, we proposed a novel metric, the glycemic variability percentage (GVP), for assessing glycemic variability that accounts for both the amplitude and frequency of glycemic fluctuations.

METHOD

We applied the new metric, the GVP, to a previously reported case of a subject using an earlier generation continuous glucose monitoring (CGM) device, in which successive periods of use were associated with an apparent decrease in glycemic variability. Results were compared with histogram distributions for the rate of change of glucose as well. The GVP was also applied to data from a published study of a bihormonal artificial pancreas system comparing results from open loop and closed loop in adolescents and in adults.

RESULTS

The GVP was able to quantify the changes in glycemic variability during successive periods of CGM use. Application of the GVP to a published study of a bihormonal artificial pancreas found an increase in glycemic variability compared with other accepted metrics which suggested a decrease in glycemic variability.

CONCLUSION

The GVP may be a clinically useful tool in characterizing the change in glycemic variability in subjects using CGM devices. Compared with metrics, such as the standard deviation, that focus solely on the amplitude of oscillations, the GVP, which measures both frequency and amplitude, may also be a more useful tool in assessing the true level of glycemic variability in artificial pancreas studies.