Clinical Recommendations for the Use of the Ambulatory Glucose Profile in Diabetes Care

Abstract

Background:

The ambulatory glucose profile (AGP) uses the wealth of data that are generated by continuous glucose monitoring, including flash glucose monitoring technologies, to provide a visual representation of glucose levels over a typical standard day of usually the most recent two weeks for a person with diabetes and helps to identify patterns and trends in glucose control. The AGP allows certain patterns of glucose levels to be identified and analyzed, such that treatment adjustments can be made, and new individual treatment goals can be defined. This helps to ensure increased treatment satisfaction and adherence, quality of life, and an improvement in metabolic management for people with diabetes.

Objective:

To date, a range of approaches exists for interpreting the information contained in an AGP, with different priorities given to identifying and targeting patterns of hypoglycemia and the degree of variability and stability underlying the glucose levels. The objective of the present recommendation is to describe the steps for assessing an AGP in detail and to illustrate these steps using visual examples.

Conclusion:

This paper describes the consensus recommendations from a group of German expert diabetologists on the necessary steps for assessing an AGP in a structured and detailed way and to explain these steps using practical clinical examples.

Keywords

diabetes therapy diabetes self-management ambulatory glucose profile AGP continuous glucose monitoring flash glucose monitoring CGM rtCGM isCGM glycemic control glycemic variability glycemic stability hypoglycemia

Introduction

The ambulatory glucose profile (AGP) is an internationally agreed on and standardized format that allows data obtained by measuring tissue or blood glucose levels to be processed, plotted, and analyzed in terms of treatment-relevant factors.^1,2 It provides a comprehensive overview of the glucose profile for people with diabetes and simplifies clinical decision-making for physicians and the diabetes team. People with diabetes can also gain a better understanding of their own glycemic control, thanks to the highly visual presentation of the data.^3-6 Alongside the AGP, the individual daily or weekly glucose profiles for each patient are also important when analyzing the AGP, because they enable the patterns identified within the AGP to be more precisely characterized. The causes of individual hyperglycemic and hypoglycemic events and trends can therefore be analyzed and discussed together with the person with diabetes.^2,5

The AGP was originally developed by Mazze et al to plot the glucose data measured by patient self-monitoring of blood glucose (SMBG).¹ With the arrival of real-time continuous glucose monitoring (rtCGM) and intermittently scanned CGM (isCGM), also called flash glucose monitoring,⁷ the AGP method was accordingly adopted and modified. The AGP is based on glucose levels measured every one to five minutes, over several days or weeks, being displayed as if all the readings had occurred in a single 24-hour period—the so-called “modal” or standard day. Fourteen days of continuous glucose data are optimal to provide an AGP for a person with diabetes that is sufficient to generate a reliable profile for identification of patterns.⁸ A 14-day AGP has been shown to accurately predict the anticipated glucose control profile for up to 30 days under typical everyday conditions.^3,8,9 The AGP is therefore not only a representation of glucose control at a particular point in time but also reveals daily dynamic changes and enables diabetes healthcare professionals to pinpoint aspects of glucose control for future improvement.

In the AGP, glucose levels over several days or weeks are graphically displayed in the context of a typical 24-hour day.¹⁰ This reflects the changes in glucose levels and displays the actual variability and stability over the selected period. In the analysis of the data, different information can be calculated as objective measures, such as the average glucose level; the time in range (TIR), as well as time above or below range; events with high or low glucose levels and the average duration of these events. Time in range can be shown both as a percentage and also as the number of hours that the measured glucose values are in the target range.³

A number of protocols have been suggested in recent years to optimize analysis of the AGP (see Table 1). A priority has been given to identifying the pattern and risk of hypoglycemia, as well as the degree of variability underlying the glucose levels. The objective of the present recommendation is to describe the steps for assessing an AGP in detail and to illustrate these steps using visual examples. This approach is applicable to type 1 and type 2 diabetes, in children and adults, as well as in pregnancy. The concepts discussed are reflected in analytics provided by all CGM device manufacturers in their reports and are also provided in generic nonproprietary software that can be used with any CGM device. These recommendations were previously documented in a German-language publication.¹¹

Table 1.

Procedure for Analyzing an Ambulatory Glucose Profile.³

Mazze et al²	Matthaei et al⁵; Siegmund et al⁶
Analyze glucose variability	Verify data quality (setting of target range, TIR)
Verify stability of glucose profile	Identify and reduce hypoglycemic events
Identify and reduce hypoglycemic events	Evaluate and optimize TIR and glucose variability
Summary during therapeutic interview: key messages
New evaluation of patient AGP

Abbreviations: AGP, ambulatory glucose profile; TIR, time in range.

Data Quality

The quality and the quantity of the data obtained are of great importance. For an optimal analysis, the AGP should use continuous data recorded over at least 14 days.⁸ The analysis interval should be between 14 and 28 days at most and reflect a representative period. At least 70% of sensor data should be captured during this period, provided that data gaps are not concentrated within a certain time period.¹² In the case of extensive gaps in the data at certain times of the day (for example, due to longer intervals between scans or transmitter dysfunction), the data can only be evaluated to a limited degree.

Defining a Target Range, Hypoglycemia Thresholds, and TIR

Setting the patient’s individual target range for glucose levels and defining the hypoglycemia thresholds are important, both for evaluating the AGP data itself and for making treatment decisions. The recommendations for these thresholds should always meet each patient’s needs for achieving their individual HbA1c target value and for avoiding severe hypoglycemia. Based on “International Consensus on Use of Continuous Glucose Monitoring” and the consensus report “Standardizing Clinically Meaningful Outcome Measures Beyond HbA1c for Type 1 Diabetes,” the default target glucose range is recommended as 70 to 180 mg/dL (3.9-10 mmol/L) with a secondary target glucose range of 70 to 140 mg/dL (3.9-7.8 mmol/L) in some cases.^12,13

In the AGP format, only one target range can be defined and displayed, with no differentiation between pre- and postprandial glucose levels being possible. Therefore, both indicated ranges for the TIR calculation can be considered to be a compromise—up to 140 mg/dL (7.8 mmol/L) with broad exclusion of the postprandial changes or up to 180 mg/dL (10 mmol/L) to include the postprandial changes. For example, a low percentage TIR can be caused by high glucose variability and/or frequent and long-lasting hyperglycemia or hypoglycemia. This can also be a sign for an increased risk of hypoglycemic and hyperglycemic events in the future. Emerging research indicates that there is a relationship between increased glucose variability (and thus a low TIR) and diabetes-related complications such as retinopathy, nephropathy, and neuropathy.^14-17 The concept of TIR is also easily understood by patients and, since it is expressed as a number, it can be used to evaluate the course of treatment. With the introduction of closed-loop systems, TIR will play an increasingly important role in the future, for example, regarding evaluation of treatment algorithms.

A possible correlation between HbA1c and TIR has been investigated in initial studies of type 1 diabetes.^18,19 Although the correlation between TIR and HbA1c is moderate, suggestions for TIR goals can be derived from HbA1c values. The expert group currently recommends a target for TIR of ≥70% for people with type 1 diabetes of all ages and for people with type 2 diabetes, which is in concordance with the newly published recommendation of the International Consensus on Time in Range.²⁰ This target should be considered as a minimum requirement, since an ideal TIR of 100% can rarely be achieved in people with diabetes. On the other hand, individual TIR targets have to be established together with each person with diabetes, according to their personal needs and priorities.

Identifying and Avoiding Hypoglycemia

A fundamental task of adjusting treatment is to avoid hypoglycemia²¹ and this is one of the biggest challenges in diabetes management.²² Even slight hypoglycemia can be responsible for increased glucose instability (endogenous glucose release/exogenous carbohydrate delivery). Simply eliminating an existing issue with hypoglycemia can lead to a considerable improvement in their glucose profile for many people with diabetes.

Expert consensus has suggested classifying hypoglycemia into three distinct levels: level 1 <70 mg/dL (3.9 mmol/L); level 2 <54 mg/dL (3.0 mmol/L); and level 3, severe hypoglycemia in which the patient is reliant on external assistance.^12,13,23

The AGP format allows a detailed analysis of hypoglycemia in terms of its frequency (frequency of the events), duration (in minutes), depth (glucose level), and periodicity (number of events per time interval). Treatment recommendations can be adjusted based on this information (Figure 1). Whenever an increased frequency of asymptomatic or severe hypoglycemia occurs, avoiding further hypoglycemia should be the focus of treatment adjustment.

Figure 1.

Examples are shown of hypoglycemic events from individual ambulatory glucose profiles that have been evaluated by the expert group in terms of frequency, duration, depth, and periodic occurrence, with corresponding recommendations for action.

The assessment of different levels of hypoglycemia data from an AGP obtained over 14 days are described in greater detail in Figure 1, along with the possible therapeutic consequences in each case: if the glucose level is between 70 and 54 mg/dL (3.9-3.0 mmol/L) for less than 60 minutes, no direct therapeutic interventions are required (Figure 1, row 1). If glucose levels remain between 70 and 54 mg/dL (3.9-3.0 mmol/L) for longer than 60 minutes, the subsequent action must be determined on an individual basis (Figure 1, row 2). This is also advisable if the frequency of these events increases from 5 to 10 over 14 days (Figure 1, row 3). If the glucose level falls below 54 mg/dL (3.0 mmol/L), the episodes become longer and/or the frequency increases, therapeutic interventions are required (Figure 1, rows 4 and 5). The occurrence of periodic events with glucose levels below 54 mg/dL (3.0 mmol/L) requires immediate therapeutic intervention regardless of the duration of these events (Figure 1, rows 6-8). With a measurement period of 14 days, a frequency of 10 means that the patient suffers a hypoglycemia episode about every two days. A frequency of 15 would mean a hypoglycemic event every day on average.

It must be pointed out that even with one severe hypoglycemic event (defined as >60 minutes and <54 mg/dL [3.0 mmol/L] in this expert recommendation), a therapeutic intervention may be necessary and avoiding further hypoglycemic episodes is the primary goal.

Analyzing Glucose Variability

Emerging data shows that, regardless of HbA1c values, fluctuations in glucose levels are associated with an increased risk of diabetes complications.^24-29 Therefore, an essential component of any AGP evaluation is to assess the variability in glucose values.^2,5,6 Glucose fluctuations within a day (intraday) and from one day to another (interday) must both be taken into account here.

Glucose variability is a measure of the scatter of the glucose levels at any time point and is differentiated into the interquartile range (IQR, 25th to 75th percentile) and the interdecile range (IDR, 10th to 90th percentile) both of which are arranged around the median line (Figure 2). The IQR covers the 50% of all the sensor glucose values that are closest to the median. The IQR is more likely to be affected by the specifics of treatment, such as carbohydrate factors for matching insulin doses with carbohydrate consumption, correction factors/insulin sensitivity, and basal insulin rates or basal insulin dose (Figure 3). The IDR reflects 80% of all the sensor glucose values, including those readings outside of the IQR that reveal the consequences of individual behavior rather than the aspects of day-to-day treatment. For example, the IDR can reflect skipped meals, inappropriate injection-meal interval (IMI), irregular daily routines with varying times for meals, exercise or alcohol, as well as incorrect carbohydrate factors for some injected insulin doses (Figure 3).

Figure 2.

The median value, the interquartile range, and the interdecile range for the glucose levels measured in a person with diabetes are displayed.

Figure 3.

Example of the interquartile range and interdecile range representations of glucose variability, along with possible causes for these fluctuations in glucose levels.

The IQR and IDR can give the diabetes healthcare professional an indication of whether the glucose fluctuations are caused more by treatment factors or by behavior and therefore allow more targeted treatment decisions to be made (Figure 3). Factors that lead to significant fluctuations in the IDR can be considered as uncommon changes in daily routine, whereas the IQR reflects more-regular “errors” in treatment. For example, if the AGP shows patterns in which the IQR reaches the hypoglycemic range, then adjusting the insulin dose before this time point can be recommended.⁴ What must be noted is that irregular routines and unplanned meals can also lead to a change in the IQR. If a wide IQR develops despite a regular daily routine, additional education for the patient should be considered—for example, in order to reinforce the calculation and application of carbohydrate factors for matching insulin doses with carbohydrate consumption. If the discrepancy between the IQR and the IDR is too large, the individual daily profiles can provide information about which patterns underlie the glucose variability. For example, to establish whether the discrepancies develop on certain days of the week or are associated with other recurring events.

As a treatment goal, Monnier et al have suggested that glucose variability be measured and targeted using a score of the mean amplitude of glycemic excursion (MAGE) of ≤40 mg/dL (≤2.2 mmol/L).³⁰ The coefficient of variation of the 24-hour mean glucose values (%CV) has more-recently been proposed as a standard parameter to represent glucose variability because this value is independent of the average value (mean),³¹ and this has been supported by the 2017 international consensus on CGM.¹² A threshold for CV of <36% is suggested to differentiate between stable and unstable glucose variability.³² The basis for this recommendation was the observation that the frequency of hypoglycemic events rose significantly if this value was exceeded.³²

Glucose levels in metabolically healthy individuals are also subject to fluctuations under physiological conditions. For example, raised glucose levels after meals occur, as do values below 70 mg/dL (3.9 mmol/L), although values below 54 mg/dL (3.0 mmol/L) rarely occur (Figure 4).^32,33 A comparison of the glucose profiles for a metabolically healthy person (Figure 4) and those of a person with diabetes (Figure 2) illustrate and emphasize the high glucose variability seen in people with diabetes.

Figure 4.

Example of an ambulatory glucose profile from a metabolically healthy individual.

Checking the Stability of the Glucose Profile

A next step in an effective AGP analysis is the assessment of the stability of the glucose profile. The peaks and the time course of the rises and falls in glucose levels must be taken into account, as well as the gradient of the change in glucose. The latter is defined as the absolute change in the median glucose per hour and is evaluated using the change in the median line ((mg/dL)/h; Figure 5²). The time of any postprandial rises and the postprandial glucose levels should also be analyzed over three to four hours following meals.

Figure 5.

Illustration of glucose stability is using the positive and negative gradients of the median curve.

An improvement in glucose stability can only be achieved after examining and improving the glucose variability. In one study of the initial clinical experience with flash glucose monitoring and AGP, the stability of the glucose in the participant group (type 1 and type 2 diabetes) was 14.04 (mg/dL)/h.³⁴ An accepted “normal” value has not yet been defined for stability in this context and a metric for glucose stability cannot be provided using the established software for AGP analysis.

Evaluate the Glucose Exposure

Glucose exposure may be defined as the area under the curve of the median line (AUC_Median), which is estimated using the trapezoidal rule.^2,35 This value is dependent on the peak in the median values and on the glucose stability. Glucose variability and glucose stability both indirectly influence glucose exposure, and both should therefore be optimized before attempting to lower glucose exposure. Unlike glucose stability, the glucose exposure enables a consideration of the measured values over a period of 24 hours (mg/dL)/h compared to mg/dL × 24 hours). Similar to the glucose stability, a quantitative analysis and an evaluation of the glucose exposure based on this are not yet possible with the software tools currently available.

Discussion and Recommendations

Continuous monitoring of glucose levels has brought glucose monitoring to a higher level and helps us to evaluate glucose control much better and enables treatment adjustments to be more precise and efficient. By its nature, it also results in a large body of data that can be systematically analyzed in a structured way using the AGP. When reviewing the data, it should be ensured that the analysis interval is between 14 and 28 days at most and reflects a representative period. It is recommended that no more than one or two therapeutic adjustments are made at each AGP evaluation appointment. When considering the AGP and the analysis of the data, it is also recommended to proceed in the following order:

Reviewing the Quality of the Data, Definition of the Target Range (TIR)

A prerequisite for analyzing the AGP is that it reflects continuous collection of glucose data over 14 days and has captured at least 70% of the sensor data. Furthermore, effective use of the AGP assumes that the target glucose range has been defined. It is a limitation that it is not currently possible to differentiate in the AGP between preprandial and postprandial glucose target ranges. A target range of 70 to 180 mg/dL (3.9-10 mmol/L) or 70 to 140 mg/dL (3.9-7.8 mmol/L) must be chosen against which the TIR can be calculated. The general recommendation is that the target glucose range should be adapted to meet the individual patient’s needs.

The expert group currently recommends, in accordance with recommendations of the International Consensus on Time in Range¹⁸ that the target for TIR in people with type 1 diabetes, independent of age, or with type 2 diabetes should be ≥70%. The consensus group also recommends targets for time below 70 mg/dL (3.9 mmol/L) of <4% and below 54 mg/dL (3.0 mmol/L) of <1% and to minimize time in hyperglycemia >180 mg/dL (10 mmol/L) to <25%, and <5% for readings >250 mg/dL (13.9 mmol/L).

Because an ideal TIR of 100% in people with diabetes is not a realistic goal, the TIR targets should be considered as minimum requirements for good diabetes care. There is evidence of an association between the TIR and clinical trial outcomes.^{13,16,17,36,37} However, the evidence associating TIR with diabetes outcomes is still limited, meaning that additional reports on clinical experience with this measure of glycemic control will be important for verifying these recommendations.

Identifying and Reducing Hypoglycemia Patterns

The AGP allows differentiated consideration of hypoglycemia by frequency, duration, depth, and periodicity, and thus, enables adaptation of the treatment recommendations. With the occurrence of one event of pronounced hypoglycemia (defined as lasting >60 minutes and <54 mg/dL [3.0 mmol/L] in this expert recommendation), a therapeutic intervention may be necessary and avoiding further hypoglycemic episodes should become the focus of therapy.

Reducing Glucose Variability

The degree of glucose variability is reflected in the IQR and the IDR and can be affected by both treatment factors and patient behavior. In this context, therapeutic factors are considered to be the main reason behind changes in the IQR, whereas the IDR is particularly affected by behavioral factors. A %CV of <36% is currently identified as a treatment goal.³⁰

Checking the Glucose Stability

An improvement in the stability (the peak and time course of the glucose rises and falls, as well as their gradient) can only be achieved after examining and improving the degree of glucose variability. It is recommended to keep the gradient of the median line as low as possible and to approximate the profile of a metabolically healthy person. The glucose stability can be visually assessed using the AGP, although no software tools for quantitative analysis are yet available.

Checking the Glucose Exposure

As soon as appropriate software tools become available, the glucose exposure should be used as an additional parameter for evaluation as part of the AGP review.

Summary

The AGP uses the wealth of data that are generated by CGM technologies and provides a rapid visual representation of the glucose profile for a person with diabetes that helps to identify patterns and trends in glucose control. The AGP should be reviewed by diabetes healthcare professionals together with the person with diabetes. This enables any abnormal features to be identified and investigated, such that treatment adjustments can be made and new individual treatment goals can be defined. This helps to ensure increased treatment satisfaction, quality of life, and an improvement in metabolic management for people with diabetes.

Footnotes

Acknowledgements

The authors thank Dr Katharina Fritzen and Professor Oliver Schnell, Sciarc GmbH, Baierbrunn, and Dr Rob Brines from Bite Medical Consulting for their support in the preparation of this manuscript.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: The authors are members of the national advisory board of Abbott Diabetes Care.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The preparation of this manuscript was supported by Abbott Diabetes Care.

References

Mazze

Lucido

Langer

, et al. Ambulatory glucose profile: representation of verified self-monitored blood glucose data. Diabetes Care. 1987;10(1):111-117.

Mazze

Strock

Wesley

, et al. Characterizing glucose exposure for individuals with normal glucose tolerance using continuous glucose monitoring and ambulatory glucose profile analysis. Diabetes Technol Ther. 2008;10(3):149-159.

Bergenstal

Ahmann

Bailey

, et al. Recommendations for standardizing glucose reporting and analysis to optimize clinical decision making in diabetes: the Ambulatory Glucose Profile. J Diabetes Sci Technol. 2013;7(2):562-578.

Matthaei

Assessing the value of the Ambulatory Glucose Profile in clinical practice. Br J Diab Vasc Dis. 2014;14(4):148-152.

Matthaei

Dealaiz

Bosi

, et al. Consensus recommendations for the use of Ambulatory Glucose Profile in clinical practice. Br J Diab Vasc Dis. 2014;14(4):153-157.

Siegmund

Matthaei

Reuter

, et al. Ambulatory glucose profile. Recommendations for use in clinical practice. Diabetes Stoffwechsel und Herz. 2015;24:115-120.

Petrie

Peters

Bergenstal

, et al. Improving the clinical value and utility of CGM systems: issues and recommendations: a joint statement of the European Association for the Study of Diabetes and the American Diabetes Association Diabetes Technology Working Group. Diabetes Care. 2017;40(12):1614-1621.

Riddlesworth

Beck

Gal

, et al. Optimal sampling duration for continuous glucose monitoring to determine long-term glycemic control. Diabetes Technol Ther. 2018;20(4):314-316.

Mazze

Strock

Bergenstal

, et al. Characterization of glucose metabolism. Staged Diabetes Management. Chichester, UK: Wiley-Blackwell; 2011.

10.

Evans

Cranston

Bailey

CJ.

Ambulatory glucose profile (AGP): utility in UK clinical practice. Br J Diabetes. 2017;17(1):28-33.

11.

Kroeger

Reichel

Siegmund

Ziegler

Praxisbezogene Empfehlungen zum Ambulanten Glukoseprofil. Diabetologie und Stoffwechsel. 2018;13(2):174-183.

12.

Danne

Nimri

Battelino

, et al. International consensus on use of continuous glucose monitoring. Diabetes Care. 2017;40(12):1631-1640.

13.

Agiostratidou

Anhalt

Ball

, et al. Standardizing clinically meaningful outcome measures beyond HbA1c for type 1 diabetes: a consensus report of the American Association of Clinical Endocrinologists, the American Association of Diabetes Educators, the American Diabetes Association, the Endocrine Society, JDRF International, The Leona M. and Harry B. Helmsley Charitable Trust, the Pediatric Endocrine Society, and the T1D Exchange. Diabetes Care. 2017;40(12):1622-1630.

14.

Gandhi

Kovalaske

Kudva

, et al. Efficacy of continuous glucose monitoring in improving glycemic control and reducing hypoglycemia: a systematic review and meta-analysis of randomized trials. J Diabetes Sci Technol. 2011;5(4):952-965.

15.

Sartore

Chilelli

Burlina

, et al. The importance of HbA1c and glucose variability in patients with type 1 and type 2 diabetes: outcome of continuous glucose monitoring (CGM). Acta diabetologica. 2012;49(suppl 1):s153-s160.

16.

Beck

Bergenstal

Cheng

, et al. The relationships between time in range, hyperglycemia metrics, and HbA1c. J Diabetes Sci Technol. 2019;13(4):614-623.

17.

Zhou

, et al. Association of time in range, as assessed by continuous glucose monitoring, with diabetic retinopathy in type 2 diabetes. Diabetes Care. 2018;41(11):2370-2376.

18.

Vigersky

McMahon

The relationship of hemoglobin A1C to time-in-range in patients with diabetes. Diabetes Technol Ther. 2018;21(2):81-85.

19.

Foster

Beck

Miller

, et al. State of Type 1 diabetes management and outcomes from the T1D exchange in 2016-2018. Diabetes Technol Ther. 2019;21(4):66-72.

20.

Battellino

Danne

Amiel

, et al. Clinical targets for continuous glucose monitoring data interpretation: recommendations from the International Consensus on Time in Range. Diabetes Care. 2019;42(8):1593-1603.

21.

American Diabetes Association A. Glycemic targets: standards of medical care in diabetes—2019. Diabetes Care. 2019;42(Supplement 1):S61-S70.

22.

Shafiee

Mohajeri-Tehrani

Pajouhi

, et al. The importance of hypoglycemia in diabetic patients. J Diabetes Metab Disord. 2012;11(1):1-7.

23.

International Hypoglycaemia Study Group. Glucose concentrations of less than 3.0 mmol/L (54 mg/dL) should be reported in clinical trials: a joint position statement of the American Diabetes Association and the European Association for the Study of diabetes. Diabetes Care. 2017;40(1):155-157.

24.

Lin

Yang

, et al. Variation of fasting plasma glucose: a predictor of mortality in patients with type 2 diabetes. Am J Med. 2012;125(4):416.e9-e18.

25.

Gorst

Kwok

Aslam

, et al. Long-term glycemic variability and risk of adverse outcomes: a systematic review and meta-analysis. Diabetes Care. 2015;38(12):2354-2369.

26.

Bonke

Donnachie

Schneider

, et al. Association of the average rate of change in HbA1c with severe adverse events: a longitudinal evaluation of audit data from the Bavarian Disease Management Program for patients with type 2 diabetes mellitus. Diabetologia. 2016;59(2):286-293.

27.

Wan

Fung

Fong

, et al. Association of variability in hemoglobin A1c with cardiovascular diseases and mortality in Chinese patients with type 2 diabetes mellitus — A retrospective population-based cohort study. J Diabetes Complications. 2016;30(7):1240-1247.

28.

Lee

C-L

Sheu

Lee

I-T

, et al. Trajectories of fasting plasma glucose variability and mortality in type 2 diabetes. Diabetes Metab. 2018;44(2):121-128.

29.

Wightman

Sainsbury

Jones

GC.

Visit-to-visit HbA1c variability and systolic blood pressure (SBP) variability are significantly and additively associated with mortality in individuals with type 1 diabetes: an observational study. Diabetes Obes Metab. 2018;20(4):1014-1017.

30.

Monnier

Colette

Glycemic variability: should we and can we prevent it?

Diabetes Care. 2008;31(suppl 2):S150-S154.

31.

DeVries

JH.

Glucose variability: where it is important and how to measure it. Diabetes. 2013;62(5):1405-1408.

32.

Monnier

Colette

Wojtusciszyn

, et al. Toward defining the threshold between low and high glucose variability in diabetes. Diabetes Care. 2017;40(7):832-838.

33.

Marks

. Glycaemic stability in healthy subjects: fluctuations in blood glucose concentration during the day. In: Andreani

Marks

Lefèbvre

Hrsg , eds. Hypoglycemia. New York, NY: Raven Press; 1987:19-24.

34.

Distiller

Cranston

Mazze

First clinical experience with retrospective flash glucose Monitoring (FGM) analysis in South Africa: characterizing glycemic control with ambulatory glucose profile. J Diabetes Sci Technol. 2016;10(6):1294-1302.

35.

Matthews

Altman

Campbell

, et al. Analysis of serial measurements in medical research. BMJ (Clinical research ed). 1990;300(6719):230-235.

36.

Omar

Salama

Allam

, et al. Association of time in blood glucose range with outcomes following cardiac surgery. BMC Anesthesiology. 2015;15(1):14.

37.

Rickels

Peleckis

Markmann

, et al. Long-term improvement in glucose control and counterregulation by islet transplantation for Type 1 diabetes. J Clinl Endocrinol Metab. 2016;101(11):4421-4430.