Abstract:Objective To investigate the incidence and risk factors of nocturnal hypoglycemia in type 1 diabetes mellitus (T1DM), and to explore whether nocturnal hypoglycemia can be evaluated by daily capillary blood glucose profiles. Methods A total of 137 T1DM patients hospitalized in Nanjing Drum Tower Hospital between May 2013 and August 2018 were recruited and received continuous glucose monitoring (CGM) for 3 consecutive days during stable therapy period, meanwhile capillary blood glucose measurements (before and after 3 meals as well as at bedtime) were recorded. From data of CGM, the patients were divided into nocturnal hypoglycemia group and non-hypoglycemia group. The demographics, laboratory measurements and dynamic blood glucose parameters were compared between these two groups. Receiver operating characteristic curve (ROC) was used to analyze the optimal cut-off points of fasting blood glucose, postprandial blood glucose, bedtime blood glucose and BGn was used to predict nocturnal hypoglycemia. Multiple regression analysis was used to evaluate the risk factors of nocturnal hypoglycemia. Results (1) Totally, 179 hypoglycemic profiles monitored by CGM were recorded in 137 patients with T1DM, including 50 nocturnal hypoglycemic episodes occurred in 31 patients (27.9%). (2) Logistic regression analysis indicated that glycated hemoglobin A1c (HbA1c), glycated albumin and mean blood glucose were independent negatively correlated with nocturnal hypoglycemia (OR=0.784, 0.021, 0.751, P<0.05), while standard deviation of blood glucose and low blood glucose index were independent positively correlated with nocturnal hypoglycemia (OR=1.641, 3.004, P<0.05). (3) Of the daily capillary blood glucose profiles, levels of fasting blood glucose, blood glucose after dinner and blood glucose at bedtime were independently negative correlated with nocturnal hypoglycemia (OR=0.257, 0.685, 0.708, P<0.05). The cut-off values for prediction of nocturnal hypoglycemia were as follows: levels of fasting glucose was 5.8 mmol/L [sensitivity 80%, specificity 90%, area under curvel (AUC) 0.91], levels of blood glucose after dinner was 8.2 mmol/L (sensitivity 57%, specificity 76%, AUC 0.72), levels of blood glucose at bedtime was 6.7 mmol/L (sensitivity 64%, specificity 90%, AUC 0.80). (4) Based on the capillary blood glucose profiles before meals and at bedtime, a model was established for the prediction of nocturnal hypoglycemia: BGn=bedtime blood glucose × (1-SDSM/MBGSM). When BGn was 5.2 mmol/L, the specificity increased to 93% and AUC was 0.81. Conclusion Nocturnal hypoglycemia is common in patients with T1DM, and individualized blood glucose control targets should be set. Potential increased risk factors of nocturnal hypoglycemia were as follows: the levels of fasting blood glucose in the morning is below 5.8 mmol/L, the levels of blood glucose at bedtime is below 6.7 mmol/L or the model of BGn is below 5.2 mmol/L. Key words: Type 1 diabetes; Nocturnal hypoglycemia; Capillary blood glucose monitoring; Continuous glucose monitoring

Is type 1 diabetes a chaotic phenomenon?

Chaotic time series prediction for glucose dynamics in type 1 diabetes mellitus using regime-switching models

Machine Learning-Based Time in Patterns for Blood Glucose Fluctuation Pattern Recognition in Type 1 Diabetes Management: Development and Validation Study

Dynamic associations between glucose and ecological momentary cognition in Type 1 Diabetes

Multivariate statistical analysis methods to investigate interindividual glucose dynamics for subjects with type 1 diabetes mellitus

Temporal patterns in insulin needs for Type 1 diabetes

The Challenge of Predicting Meal-to-meal Blood Glucose Concentrations for Patients with Type I Diabetes

Novel Detection and Progression Markers for Diabetes Based on Continuous Glucose Monitoring Data Dynamics

Nocturnal Glucose Patterns with and without Hypoglycemia in People with Type 1 Diabetes Managed with Multiple Daily Insulin Injections

The Characteristics of Blood Glucose Fluctuations in Patients with Fulminant Type 1 Diabetes Mellitus in the Stable Stage

Dynamic evaluation of blood glucose fluctuations in type 2 diabetic and normal subjects

Type 1 Diabetes—A Clinical Perspective

Is Psychological Stress a Factor for Incorporation Into Future Closed-Loop Systems?

Identification of Main Factors Explaining Glucose Dynamics During and Immediately After Moderate Exercise in Patients With Type 1 Diabetes

Integrating Bayesian Approaches and Expert Knowledge for Forecasting Continuous Glucose Monitoring Values in Type 2 Diabetes Mellitus

Mapping data to virtual patients in type 1 diabetes

A System of ODEs for Representing Trends of CGM Signals

Clustering of Hypoglycemia Events in Patients With Hyperinsulinism: Extension of the Digital Phenotype Through Retrospective Data Analysis

Value of Daily Capillary Blood Glucose Profiles in Assessing Nocturnal Hypoglycemia of Type 1 Diabetes Mellitus Based on Continuous Blood Glucose Monitoring System

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

Ensemble machine learning reveals key features for diabetes duration from electronic health records