Glucose Risk Index Research Articles

Association Between Glycemic Variability and Persistent Acute Kidney Injury After Noncardiac Major Surgery: A Multicenter Retrospective Cohort Study.

While the relationship between glycemic variability (GV) and acute kidney injury (AKI) has been a subject of interest, the specific association of GV with persistent AKI beyond 48 hours postoperative after noncardiac surgery is not well-established. This retrospective cohort study aimed to describe the patterns of different GV metrics in the immediate 48 hours after noncardiac surgery, evaluate the association between GV indices and persistent AKI within the 7-day postoperative window, and compare the risk identification capabilities of various GV for persistent AKI. A total of 10,937 patients who underwent major noncardiac surgery across 3 medical centers in eastern China between January 2015 and September 2023 were enrolled. GV was characterized using the coefficient of variations (CV), mean amplitude of glycemic excursions (MAGE), and the blood glucose risk index (BGRI). Multivariable logistic regression was used to examine the relationship between GV and AKI. Optimal cutoff values for GV metrics were calculated through the risk identification models, and an independent cohort from the INformative Surgical Patient dataset for Innovative Research Environment (INSPIRE) database with 7714 eligible cases served to externally validate the risk identification capability. Overall, 274 (2.5%) of the 10,937 patients undergoing major noncardiac surgery met the criteria of persistent AKI. Higher GV was associated with an increased risk of persistent AKI (CV: odds ratio [OR] = 1.26, 95% confidence interval [CI], 1.08-1.46; MAGE: OR = 1.31, 95% CI, 1.15-1.49; BGRI: OR = 1.18, 95% CI, 1.08-1.29). Compared to models that did not consider glycemic factors, MAGE and BGRI independently contributed to predicting persistent AKI (MAGE: areas under the curve [AUC] = 0.768, P = .011; BGRI: AUC = 0.764, P = .014), with cutoff points of 3.78 for MAGE, and 3.02 for BGRI. The classification of both the internal and external validation cohorts using cutoffs demonstrated good performance, achieving the best AUC values of 0.768 for MAGE in the internal cohort and 0.777 for MAGE in the external cohort. GV measured within 48 hours postoperative period is an independent risk factor for persistent AKI in patients undergoing noncardiac surgery. Specific cutoff points can be used to stratify at-risk patients. These findings indicate that stabilizing GV may potentially mitigate adverse kidney outcomes after noncardiac surgery, highlighting the importance of glycemic control in the perioperative period.

136-LB: The Implication of Glycemic Risk Index Varies with Different Levels of Glycosylation A1c (HbA1c) in Type 1 Diabetes

Objective: Type 1 diabetes (T1D) is characterized by significant fluctuations in blood glucose, necessitating comprehensive assessment of glycemic control. The Glucose Risk Index (GRI) is a new glucose assessment metric derived from continuous glucose monitoring (CGM) that can simultaneously indicate hyperglycemic and hypoglycemic risks. This study explored the implications of GRI for T1D with different HbA1c levels. Methods: We obtained 507 CGM and clinical data points from 281 patients with T1D, some of whom had follow-up data with intervals of over 3 months. The included patients aged 4-72 years with duration greater than 3 months and wore a CGM for at least 14 days. The hypoglycemic index and hyperglycemic index were calculated by time below range (TBR) <54 mg/dL + 0.8 × TBR <70-54 mg/dL and time above range (TAR) >250 mg/dL) + 0.5 × TAR >180-250 mg/dL, respectively. GRI is the sum of 3 times hypoglycemic index and 1.6 times hyperglycemic index. Relationships between GRI and HbA1c, each CGM metrics were explored and compared in patients with HbA1c <7% and ≥7%. Results: The mean GRI and HbA1c were 57.6 ± 20.6 and 7.3 ± 1.1%, respectively, and HbA1c of 7% approximately corresponded to GRI of 55. GRI was correlated with HbA1c (R = 0.536), hypoglycemic index (R = 0.466) and hyperglycemic index (R = 0.654), but the correlation between GRI and HbA1c disappeared in subgroup with HbA1c <7%. Compared with the HbA1c ≥7% group, GRI in the HbA1c <7% group had a stronger correlation with hypoglycemic index and a weaker correlation with hyperglycemic index. Further analysis showed the combination of dichotomous HbA1c (<7% or ≥7%) and GRI (<55 or ≥55) conferred best evaluation of glycemic control. Conclusions: The implication of GRI varies with different levels of HbA1c in T1D, with GRI mainly representing hypoglycemic risk when HbA1c <7%, and hyperglycemic risk when HbA1c ≥7%. A more comprehensive assessment of glycemic control in T1D should be the combination of dichotomous HbA1c and GRI. Disclosure B. He: None. Y. Xie: None. X. Li: None. Z. Zhou: None. Funding National Natural Science Foundation of China (81600649)

Glycemic Control by Treatment Modalities: National Registry-Based Population Data in Children and Adolescents with Type 1 Diabetes

Introduction: The aim of the study was to assess the differences in key parameters of type 1 diabetes (T1D) control associated with treatment and monitoring modalities including newly introduced hybrid closed-loop (HCL) algorithm in children and adolescents with T1D (CwD) using the data from the population-wide pediatric diabetes registry ČENDA. Methods: CwD younger than 19 years with T1D duration >1 year were included and divided according to the treatment modality and type of CGM used: multiple daily injection (MDI), insulin pump without (CSII) and with HCL function, intermittently scanned continuous glucose monitoring (isCGM), real-time CGM (rtCGM), and intermittent or no CGM (noCGM). HbA1c, times in glycemic ranges, and glucose risk index (GRI) were compared between the groups. Results: Data of a total of 3,251 children (mean age 13.4 ± 3.8 years) were analyzed. 2,187 (67.3%) were treated with MDI, 1,064 (32.7%) with insulin pump, 585/1,064 (55%) with HCL. The HCL users achieved the highest median TIR 75.4% (IQR 6.3) and lowest GRI 29.1 (7.8), both p < 0.001 compared to other groups, followed by MDI rtCGM and CSII groups with TIR 68.8% (IQR 9.0) and 69.0% (7.5), GRI 38.8 (12.5) and 40.1 (8.5), respectively (nonsignificant to each other). These three groups did not significantly differ in their HbA1c medians (51.8 [IQR 4.5], 50.7 [4.5], and 52.7 [5.7] mmol/mol, respectively). NoCGM groups had the highest HbA1c and GRI and lowest TIR regardless of the treatment modality. Conclusion: This population-based study shows that the HCL technology is superior to other treatment modalities in CGM-derived parameters and should be considered as a treatment of choice in all CwD fulfilling the indication criteria.

Design of dual hormone blood glucose therapy and comparison with single hormone using MPC algorithm.

The complete automated control and delivery of insulin and glucagon in type 1 diabetes is the developing technology for artificial pancreas. This improves the quality of life of a diabetic patient with the precise infusion. The amount of infusion of these hormones is controlled using a control algorithm, which has the prediction property. The control algorithm model predictive control (MPC) predicts one step ahead and infuses the hormones continuously according to the necessity for the regulation of blood glucose. In this research, the authors propose a MPC control algorithm, which is novel for a dual hormone infusion, for a mathematical model such as Sorenson model, and compare it with the insulin alone or single hormone infusion developed with MPC. Since they aim for complete automatic control and regulation, unmeasured disturbances at a random time are integrated and the performance evaluation is projected through statistical analysis. The blood glucose risk index (BGRI) and control variability grid analysis (CVGA) plot gives the additional evaluation for the comparative results of the two controllers claiming 88% performance by dual hormone evaluated through CVGA plot and 2.05 mg/dl average tracking error, 2.20 BGRI. The MPC developed for dual hormone significantly performs better and the time spent in normal glycaemia is longer while eliminating the risk of hyperglycaemia and hypoglycaemia.

Tuning of an Artificial Pancreas Controller: an in silico methodology based on clinically-relevant criteria.

This paper presents a methodology to tune an artificial pancreas controller by minimizing the time spent in endangering glycaemic ranges (hypo- and hyperglycaemia). The risk associated to the patient's glycaemia is evaluated with an objective metric (the blood glucose risk index), which has an established clinical relevance. The tuned controller is validated in the UVA/Padova environment where the resulting artificial pancreas achieves minimal glucose risk index in realistic 24-hour long scenarios with unannounced glucose intake.

Comparison of Glycemic Variability Indices: Blood Glucose, Risk Index, and Coefficient of Variation in Predicting Adverse Outcomes for Patients Undergoing Cardiac Surgery

Fluctuations in blood glucose (glycemic variability) increase the risk of adverse outcomes. No universally accepted tool for glycemic variability exists during the perioperative period. The authors compared 2 measures of glycemic variability-(1) coefficient of variation (CV) and (2) the Blood Glucose Risk Index (BGRI)-in predicting adverse outcomes after cardiac surgery. Prospective, observational study. Single-center, teaching hospital. A total of 1,963 adult patients undergoing cardiac surgery. None. Postoperative blood glucose levels were measured hourly for the first 24 hours and averaged every 4 hours (4, 8, 12, 16, 20, and 24 hours). Glycemic variability was measured by CV and the BGRI. The primary outcome, major adverse events (MAEs), was a predefined composite of postoperative complications (death, reoperation, deep sternal infection, stroke, pneumonia, renal failure, tamponade, and myocardial infarction). Logistic regression models were constructed to evaluate the association. Predictive ability was measured using C-statistics. Major adverse events were seen in 170 (8.7%) patients. Only the fourth quartile of CV showed association (odds ratio [OR] 1.91; 95% confidence interval [CI] [1.19-3.14]; p = 0.01), whereas BGRI was related significantly to MAE (OR 1.20; 95% CI [1.10-1.32]; p < 0.0001). The predictive ability of CV and BGRI increased on adding the standard Society of Thoracic Surgeons (STS) risk index. The C-statistic for STS was 0.68, whereas STS + CV was 0.70 (p = 0.012) and STS + BGRI was 0.70 (p = 0.012). Both CV and the BGRI had good predictive ability. The BGRI being a continuous variable could be a preferred measure of glycemic variability in predicting adverse outcomes (cutoff value 2.24) after cardiac surgery.

A personalized diet and exercise recommender system for type 1 diabetes self-management: An in silico study

Management of diet and exercise levels needs to be personalized for patients with Type 1 Diabetes (T1D) to reduce the number of hypoglycemia events and to achieve a good glycemic control. This study developed a model-based Recommender system that could provide an (optimal) personalized intervention on diet and exercise for T1D patients, which could be potentially implemented as a mobile application (app) for self-management of T1D in the future work. At each intervention time, the Recommender makes prediction of blood glucose based on a patient-specific model of glucose dynamics, and then provides optimal interventions, which could be a meal/snack size or a target heart rate during exercise, by minimizing a risk function with constraints under a future time horizon. Simulations were conducted to evaluate the Recommender through 30 virtual subjects generated from a modified UVa/Padova simulator with an added exercise-glucose subsystem. The performance of the Recommender was compared to two self-management schemes: the Starter scheme and the Skilled scheme, where the Skilled represents an off-line optimal scheme providing a lower bound on the risk index. Compared to the Starter, the Recommender reduced the mean Low Blood Glucose Index by 84% and reduced the Blood Glucose Risk Index by 49% (P < 0.05), and it had comparable performance as the Skilled. The Recommender also reduced the number of hypoglycemia events during and post-exercise compared to the Starter and the Skilled.

Glucose metabolism modeling of diabetes patients with different intensities of aerobic exercise: an in silico study

Exercise is vital for diabetics to improve their blood glucose level. However, the quantitative relationship between exercise modes (including types, intensity, time, etc.) and the blood glucose is still not clear. In order to answer these questions, this paper established a blood glucose metabolic model based on ordinary differential equation method. Furthermore, a silico method was adopted to study the effects of different aerobic exercise intensities (light, moderate and vigorous) on blood glucose and optimal strategies of insulin infusion for type 1 diabetes mellitus (T1DM) and type 2 diabetes mellitus (T2DM). Additionally, the universality of proposed model and insulin infusion strategies was verified based on 1 000 virtual diabetes patients' simulation. The experimental results showed that: (1) Vigorous-intensity aerobic exercise may result in hypoglycemia (< 3.89 mmol/L), which was so harmful to health that diabetics should avoid. Compared with moderate-intensity exercise, the light-intensity aerobic exercise intuitively lowered blood glucose slowly and caused a relative long high-blood-glucose (> 6.11 mmol/L) period, however, its overall blood glucose risk index (BGRI) was lower. (2) Insulin dosage of the optimized strategies decreased by 50% and 84% for T1DM and T2DM when they did moderate intensity exercise. As for light intensity exercise, the dosage of insulin was almost the same as they didn't do exercise, but BGRI decreased significantly. (3) The simulations of 1 000 virtual diabetic patients manifested that the proposed model and the insulin infusion strategies had good universality. The results of this study can not only help to improve the quantitative understanding about the effects of aerobic exercise on blood glucose of diabetic patients, but also contribute to the regulation and management of blood glucose in exercise mode.

Optimal Insulin Bolus Dosing in Type 1 Diabetes Management: Neural Network Approach Exploiting CGM Sensor Information.

Type 1 diabetes (TID) therapy is based on multiple daily injections of exogenous insulin. The so-called insulin bolus calculators facilitate insulin dose calculation to the patients by implementing a standard formula SF which, besides some patient-related parameters, also considers the current value of blood glucose concentration (BG), normally measured by the patient through a fingerprick device. The recent approval by the U.S. Food and Drug Administration to use the measurements collected by wearable continuous glucose monitoring (CGM) sensors for insulin dosing of fers new perspectives. Indeed, CGM sensors provide real-time information on both glucose concentration and rate of change, currently not considered in the SF. The purpose of this work is to preliminary investigate the possibility of using neural networks (NN)s for the calculation of meal insulin bolus dose exploiting CGM-based information. Using the UVa/Padova TID Simulator, we generated data of 100 subjects in 9-h, single-meal, noise-free scenarios. In particular, for each subject we analyzed different meal conditions in terms of carbohydrate intakes, preprandial BG and glucose rate-of -change. Then, a fully-connected feedforward NN was trained, with the aim of estimating the insulin bolus needed to obtain the best glycemic outcomes according to the blood glucose risk index (BGRI). Preliminary results show that by using the NN to calculate insulin doses lower BGRI values are obtained, on average, compared to the SF. These results encourage further development of the approach and its assessment in more challenging scenarios.

In Silico Assessment of Literature Insulin Bolus Calculation Methods Accounting for Glucose Rate of Change.

The standard formula (SF) used in bolus calculators (BCs) determines meal insulin bolus using "static" measurement of blood glucose concentration (BG) obtained by self-monitoring of blood glucose (SMBG) fingerprick device. Some methods have been proposed to improve efficacy of SF using "dynamic" information provided by continuous glucose monitoring (CGM), and, in particular, glucose rate of change (ROC). This article compares, in silico and in an ideal framework limiting the exposition to possibly confounding factors (such as CGM noise), the performance of three popular techniques devised for such a scope, that is, the methods of Buckingham et al (BU), Scheiner (SC), and Pettus and Edelman (PE). Using the UVa/Padova Type 1 diabetes simulator we generated data of 100 virtual subjects in noise-free, single-meal scenarios having different preprandial BG and ROC values. Meal insulin bolus was computed using SF, BU, SC, and PE. Performance was assessed with the blood glucose risk index (BGRI) on the 9 hours after meal. On average, BU, SC, and PE improve BGRI compared to SF. When BG is rapidly decreasing, PE obtains the best performance. In the other ROC scenarios, none of the considered methods prevails in all the preprandial BG conditions tested. Our study showed that, at least in the considered ideal framework, none of the methods to correct SF according to ROC is globally better than the others. Critical analysis of the results also suggests that further investigations are needed to develop more effective formulas to account for ROC information in BCs.

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

We sought to cross validate several metrics for quality of glycemic control, hypoglycemia, and hyperglycemia. 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. 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 MR, one can more closely approximate the properties of BGRI, LBGI, or HBGI, and of GRADE, GRADE%Hypoglycemia, or GRADE%Hyperglycemia. 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.

A Neural-Network-Based Approach to Personalize Insulin Bolus Calculation Using Continuous Glucose Monitoring.

In type 1 diabetes (T1D) therapy, the calculation of the meal insulin bolus is performed according to a standard formula (SF) exploiting carbohydrate intake, carbohydrate-to-insulin ratio, correction factor, insulin on board, and target glucose. Recently, some approaches were proposed to account for preprandial glucose rate of change (ROC) in the SF, including those by Scheiner and by Pettus and Edelman. Here, the aim is to develop a new approach, based on neural networks (NN), to optimize and personalize the bolus calculation using continuous glucose monitoring information and some easily accessible patient parameters. The UVa/Padova T1D Simulator was used to simulate data of 100 virtual adults in a single-meal noise-free scenario with different conditions in terms of meal amount and preprandial blood glucose and ROC values. An NN was trained to learn the optimal insulin dose using the SF parameters, ROC, body weight, insulin pump basal infusion rate and insulin sensitivity as features. The performance of the NN for meal bolus calculation was assessed by blood glucose risk index (BGRI) and compared to the methods by Scheiner and by Pettus and Edelman. The NN approach brings to a small but statistically significant ( P < .001) reduction of BGRI value, equal to 0.37, 0.23, and 0.20 versus SF, Scheiner, and Pettus and Edelman, respectively. This preliminary study showed the potentiality of using NNs for the personalization and optimization of the meal insulin bolus calculation. Future work will deal with more realistic scenarios including technological and physiological/behavioral sources of variability.

In silico testing of optimized Fuzzy P+D controller for artificial pancreas

Background and objectivesDespite therapeutic advances, a complete cure has not been found yet for patients with type 1 diabetes (T1D). Artificial pancreas (AP) is a promising approach to cope with this disease. The controller part of the AP can compute the insulin infusion rate that keeps blood glucose concentration (BGC) in normoglycemic ranges. Most controllers rely on model-based controllers and use manual meal announcements or meal detection algorithms. For a fully automated AP, a controller only using the patient's BGC data is needed. MethodsAn optimized Mamdani-type hybrid Fuzzy P+D controller was proposed. Using the University of Virginia/Padova Simulator, a 36h scenario was tested in nine virtual adult patients. To take into account the effect of continuous glucose monitor noise, the scenario was repeated 25 times for each adult. The main outcomes were the percentage of time BGC levels in the euglycemic range, low blood glucose index (LBGI), and blood glucose risk index (BGRI), respectively. ResultsThe obtained BGC values were found to be in the euglycemic range for 82.6% of the time. Moreover, the BGC values were below 50mg/dl, below 70mg/dl and above 250mg/dl for 0%, 0.35% and 0.74% of the time, respectively. The BGRI, LBGI, and high blood glucose index (HBGI) were also found as 3.75, 0.34 and 3.41, respectively. The proposed controller both increases the time the BGC levels in the euglycemic range and causes less hypoglycemia and hyperglycemia relative to the published techniques studied in a similar scenario and population.

A Closed-Loop Artificial Pancreas Based on Risk Management

Control algorithms that regulate blood glucose (BG) levels in individuals with type 1 diabetes mellitus face several fundamental challenges. Two of these are the asymmetric risk of clinical complications associated with low and high glucose levels and the irreversibility of insulin action when using only insulin. Both of these nonlinearities force a controller to be more conservative when uncertainties are high. We developed a novel extended model predictive controller (EMPC) that explicitly addresses these two challenges. Our extensions to model predictive control (MPC) operate in three ways. First, they explicitly minimize the combined risk of hypoglycemia and hyperglycemia. Second, they integrate the effect of prediction uncertainties into the risk. Third, they understand that future control actions will vary if measurements fall above or below predictions. Using the University of Virginia/Padova Simulator, we compared our novel controller (EMPC) against optimized versions of a proportional-integral-derivative (PID) controller, a traditional MPC, and a basal/bolus (BB) controller, as well as against published results of an independent MPC (IMPC). The BB controller was optimized retrospectively to serve as a bound on the possible performance. We tuned each controller, where possible, to minimize a published blood glucose risk index (BGRI). The simulated controllers (PID/MPC/EMPC/BB) provided BGRI values of 2.99/3.05/2.51/1.27 as compared to the published IMPC BGRI value of 4.10. These correspond to 73/79/84/92% of BG values lying in the euglycemic range (70-180 mg/dl), respectively, with mean BG levels of 151/156/147/140 mg/dl. The EMPC strategy extends MPC to explicitly address the issues of asymmetric glycemic risk and irreversible insulin action using estimated prediction uncertainties and an explicit risk function. This controller reduces the avoidable BGRI by 56% (p < .05) relative to a published MPC algorithm studied on a similar population.

Anticipatory Behavior in Blood Glucose Control: Using Meal Prior Probabilities to Prepare for Future Meal Disturbances

AbstractAutomatic regulation of blood glucose in patients with Type 1 Diabetes is challenged by unknown or unannounced food consumption. Yet we know most food is ingested in discrete meals, spaced by brief periods of fasting. Treating meals as discrete events, this paper proposes a set of prior probabilities relating distinct meals independently of daily patterns. Using these prior probabilities, closed-loop blood glucose control can attain anticipatory behavior, implicitly lowering glucose levels when patients should be hungry and meals are most likely to occur. This improves overall performance and individual meal responses. The benefits occur because blood glucose prediction, which anticipates future food intake, can achieve near zero mean errors even over a long prediction horizon. As a side effect, such predictors can easily be tuned or tested against unrestricted out-patient data which, by definition, contains unknown meals. We validate our approach in both prediction and control of blood glucose levels. We see improved prediction accuracy over 1-4 hour horizons and a significant reduction in the blood glucose risk index for simulated closed-loop control.

Real-Time Continuous Glucose Monitoring Together with Telemedical Assistance Improves Glycemic Control and Glucose Stability in Pump-Treated Patients

Real-time continuous glucose monitoring (CGM) has recently been incorporated into routine diabetes management because of the potential advantages it offers for glycemic control. The aim of our study was to evaluate the impact of the use of real-time CGM together with a telemedicine system in hemoglobin A1c and glucose variability in patients with type 1 diabetes treated with insulin pumps. Ten patients (five women, 41.2 [range, 21-62] years old, duration of diabetes 14.9 [range, 3-52] years) were included in this randomized crossover study. Patients used the DIABTel telemedicine system throughout the study, and real-time CGM was used for 3 days every week during the intervention phase. At the end of the control phase, a blind 3-day CGM was performed. Glucose variability was evaluated using the Glucose Risk Index (GRI), a comparative analysis of continuous glucose values over two consecutive hours. Hemoglobin A1c decreased significantly (8.1 +/- 1.1% vs. 7.3 +/- 0.8%; P = 0.007) after the intervention phase, while no changes were observed during the control phase. The mean number of daily capillary glucose readings was higher during the intervention phase (4.7 +/- 1.1 vs. 3.8 +/- 1.0; P < 0.01), because of an increase in random analyses (1.22 +/- 0.3 vs. 0.58 +/- 0.1; P < 0.01), and there was also a significant increase in the mean number of bolus doses per day (5.23 +/- 1.1 vs. 4.4 +/- 0.8; P < 0.05). The GRI was higher during the control phase than during the experimental phase (9.6 vs. 6.25; P < 0.05). Real-time CGM in conjunction with the DIABTel system improves glycemic control and glucose stability in pump-treated patients with type 1 diabetes.