How to make full use of "continuous glucose monitoring data" to guide clinical diagnosis and treatment?

This article is contributed by Dr.
Liu Wei, and Yimaitong is authorized to publish it
.

"This study shows that'Detrend Fluctuation (DFF)' may provide a reference for guiding clinical diabetes classification, and we look forward to further verification of it in multi-center and large-scale studies in the future
.

" Professor Ji Linong
.

Diagnosis classification is an important prerequisite for precise treatment of diabetes
.

Although there are clinical markers such as C-peptide and diabetes-related autoantibodies to assist clinical classification, the current classification of diabetes still mainly depends on the personal experience of clinicians
.

Dynamic glucose monitoring can provide at least 7 days of continuous glucose change data, which provides a convenient means to outline the pathophysiological changes of diabetes from a digital perspective.
How to make full use of these fluctuation data for clinical services is the focus of current research
.

Professor Ji Linong’s team used de-trend fluctuations to analyze the dynamic glucose monitoring data, and tried to differentiate between type 1 diabetes and type 2 diabetes through data fluctuations
.

The research was officially published in the international journal Mediators of Inflammation (IF: 6.
577) in March 2021.
Dr.
Liu Wei from the Department of Endocrinology, Peking University People’s Hospital, Chen Jing and He Luxi, graduate students of Beijing Institute of Technology, are the co-first authors of the article.
Peking University Professor Ji Linong from the Department of Endocrinology, People's Hospital, and Professor Shi Dawei from Beijing Institute of Technology are the co-corresponding authors of the article
.

Research background and design The prevalence rate of diabetes in my country is as high as 11.
2%, which is the country with the largest number of diabetic patients in the world.
The prevalence and awareness data of epidemiological surveys over the past 40 years show that the task of diabetes prevention and treatment is extremely difficult.

.

Diabetes is the general term for a group of diseases caused by the dysfunction of the glucose homeostasis regulation system.
Hyperglycemia is a sign of diabetes and the main manifestation of steady-state imbalance.
The fluctuation of glucose level itself contains the characteristics of specific metabolic disorders
.

With the application of continuous glucose monitoring (CGM) technology, it has become a reality to continuously depict the fluctuations of glucose levels in the body within 7 days or longer intervals, laying the foundation for precise control of blood glucose and individualized treatment, and blood glucose levels are also affected.
It is considered to be the fifth vital sign after heart rate, blood pressure, respiration, and body temperature
.

Diabetes hyperglycemia diagnostic criteria and blood glucose control goals are based on numbers.
How to better establish the relationship between blood glucose fluctuation data and disease classification and control is the focus of research and development in the field of diabetes diagnosis and treatment in the era of big data
.

Diagnosis and treatment of diabetes are inseparable from "numbers"
.

The diagnostic criteria of diabetes are blood glucose and glycosylated hemoglobin levels, and the goal of diabetes control is glycosylated hemoglobin levels.
Therefore, numbers play a very important role in the clinical diagnosis and treatment of diabetes
.

In recent years, progress in the field of glucose monitoring has allowed us to obtain a large number of values ​​related to blood glucose fluctuations, and more efficient, standardized, and reasonable application of these data is a clinical problem that needs to be solved urgently
.

The team of Professor Ji Linong and Professor Dawei Shi of Beijing Institute of Technology used the analysis method of Detrended Fluctuation Function (DFF) to analyze the glucose data obtained by FGM, and explore the relationship between this indicator and insulin β cell function, and its clinical guidance The potential of diabetes typing
.

DFF is a method of analyzing the volatility of data.
It can filter out the trend component of its own evolution, and the remaining spread series are mainly volatility components
.

Its main advantage is that it can effectively eliminate the false long-range correlation caused by the non-stationarity of the time series, thereby revealing the long-range correlation of the dynamic behavior of the complex system
.

DFF has been often used in climate, hydrology, geology and other fields in the past.
In this study, researchers used DFF to analyze the volatility of glucose data to understand its correlation with serum C peptide, which reflects the function of pancreatic β-cells, and To further verify its role in guiding diabetes classification
.

Distinguish between type 1 diabetes and type 2 diabetes: DFF has potential.
The study enrolled 78 subjects with type 1 diabetes and 59 subjects with type 2 diabetes.
All subjects wore FGM for 14 days and collected FGM data for analysis.

.

In all subjects, the blood glucose fluctuations described by DFF were negatively correlated with fasting C-peptide (r = -0.
667; P <0.
001), and the correlation coefficient was higher than other indicators describing blood glucose fluctuations, namely MAGE, SD , Mean BG and TIR
.

At the same time, the blood glucose fluctuation index described by DFF showed an obvious bimodal distribution in the overall subjects (Figure 1), so the researchers further explored the ability of this index to classify diabetes
.

Figure 1 DFF is normally distributed among the subjects.
In order to better discover and verify the potential of DFF as a basis for guiding diabetes classification, the researchers randomly divided the subjects into a discovery set (Discovery Cohort) and a verification set ( Validation Cohort), and repeat the sampling process 10 times to increase the stability of the verification results
.

Analysis of receiver operating characteristics (ROC) of 10 discovery sets shows that the confidence interval (CI) of the area under the curve (AUC) is [0.
846, 0.
868], and the confidence interval of the 10 validation sets is [0.
799, 0.
862]
.

In addition, the sensitivity and specificity CIs of DFF for typing diagnosis are both 70% and 80%, indicating the potential ability of DFF to distinguish between type 1 and type 2 diabetes.

.

Experts comment that Professor Ji Linong’s recent advances in the medical field are inseparable from the cooperation between multiple disciplines.
This research uses mathematical algorithms to analyze glucose data, extracts the fluctuation information behind the numbers, and analyzes the information from a clinical perspective.
, The discovery of the relationship between fluctuation information and clinical classification and evaluation of islet function is an important exploration in this field
.

The study shows that DFF may provide a reference for guiding clinical diabetes classification, and it is expected that it will be further verified in multi-center and large-scale studies in the future
.

At the same time, it also proposes possible directions for clinically more effective use of the large amount of data and information provided by FGM to assist clinical decision-making
.

With the advent of the era of big data and the penetration and development of artificial intelligence in various fields, it is believed that this type of research will open up new ideas and directions for the development of similar fields, and we look forward to more similar studies in the future to provide more information for the diagnosis and treatment of diabetes.
Clinical optimization plan
.

The research author introduces Dr.
Liu Wei, co-first author, deputy chief physician, associate professor, and postgraduate tutor of the Endocrinology Department of Peking University People’s Hospital
.

Presided over a project of the National Natural Science Foundation of China and participated in a number of national 973 and 863 projects
.

Chen Jing is the co-first author of the Beijing Institute of Technology, School of Automation, School of Control Science and Engineering, and Master's degree candidate of the Institute of Medical-Industry Convergence
.

He Luxi, co-first author, graduated from Beijing Institute of Technology with a bachelor's degree in Automation, currently studying in Berkeley, California, and Professor Shi Dawei Co-corresponding author, professor, doctoral supervisor, assistant to the dean, research on intelligent perception and motion control, School of Automation, Beijing Institute of Technology Director of the Institute of Medical-Industry Integration, PI Cooperative Research Team Leader of "Precision Medicine System Intelligent Decision and Control"
.

Selected into the fourteenth batch of the National Overseas High-level Talent Introduction Program Youth Project, and the third batch of the China Association for Science and Technology Youth Talent Support Project
.

Professor Ji Linong, co-corresponding author Ji Linong, director of the Department of Endocrinology, Peking University People's Hospital, director of Peking University Diabetes Center, and doctoral supervisor
.

As the main investigator, lead the development of 50 new drug clinical trials, including 3 in phase I, 6 in phase II, 31 in phase III, and 9 in phase IV, of which 10 are China’s self-developed Class 1 new drug registration clinical studies.
So far, the experts who have led the most clinical trials of new drugs in the field of endocrine and metabolic diseases in China have provided evidence-based medical evidence that has contributed to major changes in international diabetes guidelines and clinical practice
.

As the “Key Clinical Specialty of the Ministry of Health” and the “Clinical Research Center in the Diabetes Field of Beijing”, the team he led has been supported by a number of National Natural Science Foundations, and has served as a major national scientific research project including 863, key research and development projects, and Beijing Municipal Science and Technology Commission Chief scientist of major issues
.

Published more than 400 papers in leading professional journals at home and abroad (including New England Journal of Medicine, Lancet, British Medical Journal, Lancet Diabetes & Metabolism, Diabetes Care, Nat Rev Endocrinol, Genetic Medicine, Diabetes, Cardiovascular Diabetology, etc.
), and were selected Elsevier Highly Cited Scholar in 2020
.

He won the second prize of National Science and Technology Progress Award (the second and third winners respectively) and the Chinese Medical Award (the second complete winner) twice
.

At the same time, he is the only expert in China who has held important leadership positions in several authoritative international diabetes academic organizations
.

Professor Ji Linong is committed to introducing advanced disease diagnosis and treatment technology and management concepts, extensively conducting clinical research on the evaluation of new diabetes drugs, and applying emerging medical technologies to the individualized diagnosis and treatment of diseases on the basis of standardized disease diagnosis and treatment to promote the development of precision medicine
.