REVIEW

Thomas W. Martens, MD Gregg D. Simonson, PhD Richard M. Bergenstal, MD

Medical Director, International Diabetes Director of Care Transformation and Training, Executive Director, International Diabetes
Center, HealthPartners Institute, Minneapolis, International Diabetes Center, HealthPartners Center, HealthPartners Institute, Minneapolis,
MN; Consultant, Department of Internal Institute, Minneapolis, MN; Assistant MN; Clinical Professor, Department of Medicine,
Medicine, Park Nicollet Clinic Professor, Department of Family Medicine, University of Minnesota, Minneapolis, MN
Brooklyn Center, MN University of Minnesota Medical School,
Minneapolis, MN

Using continuous glucose monitoring

data in daily clinical practice
ABSTRACT ontinuous glucose monitoring (CGM)
Access to and use of glycemic data are central to optimal
C technology, first developed in the early
2000s, has evolved to include devices with
management of diabetes. Use of continuous glucose longer wear times that do not require calibra-
monitoring (CGM) data to guide the management of tion with fingerstick blood glucose monitoring,
diabetes has increased dramatically thanks to improved and with dramatically improved ease of use and
ease of use, accuracy, and availability. Retrospective availability.1 In parallel with the evolution in
CGM data collected throughout the day and night allow CGM technology, there has been a dramatic
clinicians to visualize glycemic patterns, and single-page increase in clinical use of CGM, both in type
summary views like the Ambulatory Glucose Profile 1 diabetes, where CGM has become standard
(AGP) Report make rapid interpretation both feasible and of care, and in insulin-treated type 2 diabetes.2
intuitive. A systematic approach that integrates retrospec-
tive CGM-generated data at clinic visits and other clinical ■ OVERVIEW OF CGM DEVICES
interactions with personal use of CGM data can optimize Current-generation blood glucose monitor-
glycemic management. ing relies on measurement from whole blood
obtained by fingerstick, while CGM technol-
KEY POINTS ogy derives glucose values from interstitial
CGM is recommended for patients with type 1 diabetes fluid via a tiny electrode inserted beneath the
and patients with type 2 diabetes treated with insulin. skin. Because diffusion of glucose from blood
into the interstitial compartment is slightly
The single-page AGP Report allows for rapid and intuitive delayed, interstitial glucose values are processed
interpretation of CGM data by displaying patterns of mathematically to improve approximation
clinically relevant hypoglycemia, hyperglycemia, and and concordance with capillary glucose levels.
glucose variability. Although device-related delays have been min-
imized in recent CGM devices, typically there
When reviewing the time-in-ranges bar, focus on increas- is a 5- to 10-minute lag between interstitial
ing time in range to more than 70% and decreasing time and blood glucose levels,3 and this should be
below range to less than 4% to improve glycemia. communicated to clinicians and patients.
CGM technology can be broadly divided
Focus also on lifestyle and medication changes that make into 2 categories: devices for personal use by
the AGP curve more flat, narrow, and in-range. patients to monitor glucose on an ongoing
basis and professional devices, or clinic-owned
devices used intermittently to evaluate glucose
metrics and patterns at clinic visits and to
guide counseling and management suggestions.
doi:10.3949/ccjm.91a.23090 Personal use has largely overshadowed professional
CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 91 • NUMBER 10 OCTOBER 2024 611

Downloaded from www.ccjm.org on November 20, 2024. For personal use only. All other uses require permission.
CONTINUOUS GLUCOSE MONITORING DATA

TABLE 1
Currently available continuous glucose monitoring systems
Type of system Description Examples
Real-time Patient-owned Freestyle Libre 3
Measures and displays data continuously (real-time) Dexcom G6 and G7, Stelo (over the counter)
Stores data for retrospective analysis Guardian 3 and 4 and Simplera
Eversense E3

Intermittently scanned Patient-owned Freestyle Libre 2, Rio (over the counter)

Measures glucose continuously but only displays
data when swiped by a reader or smartphone
Also known as “flash” glucose monitoring

Professional Clinic-owned system placed on patient in office Freestyle Libre Pro

Typically worn for 7–14 days Dexcom G6 Pro
Glucose data may be blinded (both systems) or
unblinded (Dexcom G6 Pro) to the patient
Provides data to support medication and lifestyle
changes, guide shared decision-making, and
identify hypoglycemia

use. Professional CGM remains useful for individuals Grade B recommendation for intermittently scanned
for whom personal systems are either not needed or not CGM),2 while the American Association of Clinical
available and in specialized research settings. Personal Endocrinology strongly recommends CGM for all
CGM remains the technology of choice for most users. patients with diabetes using basal and bolus insulin
Personal CGM devices can be categorized as real- (ie, treated with both background and mealtime bolus
time devices that measure and display glucose values insulin [Grade A; high strength of evidence]) and for
continuously while worn or intermittently scanned patients with type 2 diabetes treated with less intensive
devices (Table 1). The latter are somewhat simpler insulin regimens (basal insulin only [Grade B; inter-
devices that require the user to scan a sensor worn on the mediate strength of evidence]).10
body to gather glucose data. Both types of CGM devices
can collect 24-hour retrospective data for evaluating ■ THE POWER OF CGM: 2 TYPES OF DATA
patterns and glycemic metrics, and both have utility in
the management of type 1 and type 2 diabetes. Medical nutrition and noninsulin and insulin therapies
directly target physiologic processes to improve glu-
■ EVIDENCE AND GUIDELINES ARE EVOLVING cose management; CGM improves care indirectly by
facilitating changes in lifestyle or diet and improving
Evidence from multiple randomized controlled trials medication adherence without any direct physiologic
supports the value of CGM in the management of dia- impact. The power of CGM is in the 2 types of data
betes, especially for patients who manage their diabetes it provides.
with insulin.4–9 CGM improves both hemoglobin A1c Point-in-time data: A patient with diabetes can
and hypoglycemia relative to fingerstick blood glucose view, on demand, a point-in-time glucose value, a
monitoring in type 1 diabetes.4,5 In patients with type 2 trend arrow indicating whether the glucose is rising
diabetes who use insulin, CGM improves hemoglobin or falling, and a profile of recent glucose levels that
A1c or decreases hypoglycemia to a greater degree than typically represents 8 hours of data. With point-in-
fingerstick blood glucose monitoring.6–9 time data patients can see the impact of diet choices,
Evidence-based guidelines created by specialty and lifestyle choices, and medications at any time, which
advocacy groups have evolved based on this growing allows real-time physiologic feedback to directly guide
body of evidence. The 2024 American Diabetes Asso- management of diabetes day to day.
ciation Standards of Medical Care in Diabetes supports Retrospective data: CGM technology has the capa-
CGM for all individuals with diabetes on insulin ther- bility to collect and display thousands of glucose data
apy (Grade A recommendation for real-time CGM, points retrospectively as composite glucose metrics, and
612 CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 91 • NUMBER 10 OCTOBER 2024

Downloaded from www.ccjm.org on November 20, 2024. For personal use only. All other uses require permission.
 MARTENS AND COLLEAGUES

A.

B.

C.

Figure 1. Example of an Ambulatory Glucose Profile Report. (A) The time-in-ranges graph quickly shows
whether glycemic goals are being met and whether action is needed. Average glucose and glucose man-
agement indicator metrics provide additional information about the need to take action. Glucose variability
reports variations over the course of the report period. Increased variability is a risk factor for hypoglycemia.
(B) The ambulatory glucose profile curve presents a 24-hour picture of all glucose readings collected during
the report period. (C) Ambulatory daily glucose profiles are thumbnails of daily values.
©2024 International Diabetes Center, Minneapolis, MN. Used with permission. Visit AGPreport.org for more information.

CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 91 • NUMBER 10 OCTOBER 2024 613

Downloaded from www.ccjm.org on November 20, 2024. For personal use only. All other uses require permission.
CONTINUOUS GLUCOSE MONITORING DATA

visually as composite and daily views for retrospective CGM metrics will be. This can be especially important
analysis. when counseling people using intermittently scanned
Point-in-time and retrospective data support CGM technology. More frequent scanning leads to more
diabetes management in complementary ways. Ret- complete data collection, with better insights into day
rospective data allow for shared decision-making and night patterns, frequency of hypoglycemia, and
and optimized evaluation of the safety and efficacy variability in glucose levels throughout the day.
of glycemic management during clinical interactions. Central to optimal and efficient use of CGM data
The power of retrospective CGM data lies not in the is a structured approach to its evaluation. To guide
thousands of individual data points, but in composite decision-making, we employ a 3-step evaluation process:
summary reports. Just as electrocardiographic reports Determine Where to Act.
have evolved toward a standardized layout, presenta-
tion of CGM data has evolved toward the Ambulatory Step 1: Determine whether action is needed
Glucose Profile (AGP), a standardized single-page sum- Time in ranges. The upper third of the AGP Report
mary report (Figure 1). Major CGM manufacturers use (Figure 1A) provides a summary of glycemic metrics.
slight variations of the AGP Report to display data in The time-in-ranges bar graph allows rapid determi-
a format that is familiar and accessible. While reports nation of whether glycemic goals are being met and
vary by manufacturer and device, AGP reports typically whether action is needed to improve glucose manage-
include the data elements described in this article. ment. The time-in-ranges graph displays:
There are several mechanisms for obtaining ret- • Percentage of time spent in prespecified glycemic
rospective CGM and AGP data. CGM data from ranges for the number of days included in the AGP
the sensor are sent to a reader or smartphone device Report—arguably the single most important measure
either in real time or when the device is intermittently in determining the need for action regarding the
scanned. For intermittent scanning, the sensor should adequacy and safety of the patient’s glycemic regimen
be scanned at least every 8 hours to capture all retro- • Time above range, defined as the high range of 181
spective CGM data. Once transferred to a receiver or to 250 mg/dL and very high range greater than 250
smartphone, the data can be uploaded from the device mg/dL
to an industry-based cloud data repository from which • Time-in-range target of 70 to 180 mg/dL
they can be easily viewed by the patient or, with per- • Time below range, in the low range of 69 to
mission (typically by an email invitation), remotely by 54 mg/dL and clinically significant very low range
the diabetes care team. All major CGM manufacturers below 54 mg/dL.
have proprietary cloud-based repositories. If a clinician Comparison of time in range to consensus goals
does not have access to a patient’s cloud-based data, it on the time-in-ranges graph permits the clinician or
is feasible in clinical settings to view retrospective data patient to decide quickly whether to act.
on a smartphone or reader directly. Glycemic metrics The patient represented in Figure 1 has not met any
and the AGP are typically available on these devices, of the 5 time-in-ranges goals. Action is needed because
although the format is slightly less accessible. the patient has too much time below range at 9% (goal
< 4%) and too much time above range at 25% (goal
■ THE AMBULATORY GLUCOSE PROFILE: 3 STEPS < 25%). Optimized glycemic management should focus
on increasing time in range (70–180 mg/dL) while
Because CGM technology can capture glycemic data minimizing time below range (< 70 mg/dL). Another
of a 24-hour day-night cycle over several weeks, CGM- approach is to focus on “more green” (more time in the
derived glycemic metrics and patterns displayed in an target range of 70 to 180 mg/dL) and “less red” (less
AGP Report provide a robust picture of glycemia on time with a glucose level below 70 mg/dL). This is also
both a daily and time-averaged basis. Consensus panel a patient-friendly way to communicate what the goal for
guidance recommends at least 14 days of CGM data CGM “time in ranges” is. Time in range and time below
with a minimum of 70% sensor wear to generate an range can be thought of together as a composite measure
AGP Report that enables optimal analysis and decision- reflecting the adequacy of glycemic management.14
making.11 This recommendation is based on data sug- The goals for time above range, time in range, and
gesting a strong correlation between 14-day CGM met- time below range were chosen by the International
rics that measure time within recommended ranges and Consensus on Time in Range (Table 2).15 Time in
CGM metrics collected over longer periods of time.12,13 range greater than 70% has been shown in multiple
The more complete the data, the more reliable the analyses to correlate loosely with a hemoglobin A1c of
614 CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 91 • NUMBER 10 OCTOBER 2024

Downloaded from www.ccjm.org on November 20, 2024. For personal use only. All other uses require permission.
 MARTENS AND COLLEAGUES

about 7.0%.16,17 A hemoglobin A1c target of 7.0% or TABLE 2

less is supported by multiple landmark diabetes studies,
Glucose targets in healthy
including the UK Prospective Diabetes Study and the
Diabetes Control and Complications Trial data.18,19 and at-risk adults
Additionally, evidence continues to build supporting Target levels
time in range itself as a key indicator of long-term com-
Healthy
plication risk.20–25 Interest is also building for using time nonpregnant Older and
in range as a surrogate for hemoglobin A1c, or even as Glucose values adults high-risk adults
a direct glycemic measure in place of hemoglobin A1c,
Time above range
for purposes of quality measurement.
Time in range also provides glucose data over a > 250 mg/dL < 5% < 10%
much shorter timeframe than hemoglobin A1c. This > 180 mg/dL < 25% < 50%a
frees clinicians from the traditional hemoglobin A1c– Time in range
based 3-month cycle for visits, allows for more fre- 70–180 mg/dL > 70% > 50%
quent changes to the diabetes regimen, and potentially
reduces clinical inertia. The same international CGM Time below range
consensus committee has created modified (less strin- < 70 mg/dL < 4%b < 1%
gent) time-in-range goals for individuals with reduced < 54 mg/dL < 1% NA
life expectancy or significant comorbidities.15 a
Includes values > 250 mg/dL.
Average glucose and glucose management indi- b
Includes values < 54 mg/dL.
cator, 2 glycemic metrics on the AGP Report (Figure
Based on information from reference 15.
1A), may help determine whether action is needed.
The average glucose reflects values over the data col-
lection period. The directly related glucose manage- or other changes, improving the reliability of the glu-
ment indicator, expressed as a percentage, can be used cose management indicator measure.29
clinically to estimate the hemoglobin A1c, a measure Glucose variability is a measure of variation in glu-
familiar to clinicians and patients. cose readings at a given time of day over the course of
The glucose management indicator is a calculation the AGP Report period (Figure 1A). Increased glucose
based on CGM-derived average glucose, and often does variability is an important risk factor for hypoglycemia
not align exactly with laboratory-measured hemoglobin and likely correlates with longer-term vascular risk.30
A1c for a variety of reasons.26 It is based purely on gly- Glucose variability is expressed on the AGP Report in
cemia over the period reflected on the AGP Report and terms of percent coefficient of variation. An important
can vary from the 3-month time-averaged hemoglobin clinical correlate is that if the percent coefficient of
A1c due to short-term clinical impacts (eg, change in variation is elevated (> 36%), the likelihood of hypo-
diet, use of steroids, or short-term stress). Calibration glycemia is high; by consensus, the target for glucose
accuracy of individual sensors can impact the accuracy variability is 36% or less.15,31 Often, high glycemic
of glucose management indicator estimates. Addition- variability is associated with changes in diet, physical
ally, the glucose management indicator is a derived activity, or lack of adherence to medication, such as
value based on a linear regression equation and may not skipping insulin doses or taking rapid-acting insulin
accurately correlate with laboratory hemoglobin A1c at after the meal rather than before.
the extremes of hemoglobin A1c values (ie, people with
hemoglobin A1c in the normal range or above 10%). Step 2: Identify where action is needed
Conversely, laboratory-measured hemoglobin A1c can Evaluation of time in range allows rapid determination
vary significantly from measures of true glycemia based of whether a change in therapy is needed. Further data
on many factors impacting the life span of red blood are needed to determine where the changes are needed.
cells.27,28 For that, it is necessary to review the AGP curve and
Variance between the glucose management indi- the daily glucose profiles.
cator and hemoglobin A1c is common, expected, and The AGP curve is a “modal day” view, representing
often related to known factors impacting hemoglobin all the glucose readings from the entire AGP Report
A1c measures. More recent data suggest that extending period combined and presented over a single 24-hour
the data collection period for CGM metrics beyond 14 period (Figure 1B). The AGP curve has a thick median
days may decrease the impact of short-term behavioral line, 25% to 75% interquartile range lines (indicating
CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 91 • NUMBER 10 OCTOBER 2024 615

Downloaded from www.ccjm.org on November 20, 2024. For personal use only. All other uses require permission.
CONTINUOUS GLUCOSE MONITORING DATA

TABLE 3 a profile of excessive variability to a profile that is as

close to normoglycemia as can be done safely. Focus-
Case presentation: clinical data
ing first on hypoglycemia is important in improving
Demographics short-term safety; decreasing excessive variability can
Male, age 65 dramatically improve hypoglycemia risk. The focus
Body mass index: 42.7 should be on making the AGP curve flat, narrow, and
Blood pressure: 127/75 mm Hg in range by keeping the median line as flat as possible,
Social history: single, no children, retired
the spread between the 95% and 5% lines as narrow as
Medical history possible, and the whole curve in range between 70 and
22-year history of type 2 diabetes 180 mg/dL to the extent possible.
Hypertension Daily glucose profiles provide thumbnails of daily
Hypercholesterolemia values (Figure 1C). When a glycemic concern is iden-
Diabetic neuropathy
Diabetic retinopathy tified on the AGP curve, daily glucose profiles facilitate
Peripheral vascular disease further evaluation:
• Is the issue caused by a glycemic pattern observed
Health habits on multiple days or a single day?
Nonsmoker, occasional alcohol • Is there a difference on specific days of the week
No regular physical activity
Eats 3 meals per day, often with evening snacks
(eg, weekday vs weekend)?
• Is an outlying value causing the pattern in the AGP
Current diabetes medications curve (a target for discussion of lifestyle and dietary
Metformin extended release 1,000 mg 2 times per day issues that can impact glycemia) or an artifact?
Dulaglutide 4.5 mg (maximum dose) once weekly Possible causes of artifacts in CGM data can include
Insulin glargine 60 units/day at bedtime
Insulin lispro 10 units with meals
compression of the CGM sensor during sleep, displaced
or malfunctioning sensor electrodes, or connectivity
Laboratory results problems. Typically, artifactual CGM data, or data for
Hemoglobin A1c 8.5% (reference range 4–5.6) which there is no rational clinical explanation, may be
Fasting blood glucose 165 mg/dL (70–99) dismissed in evaluating the AGP Report.
Estimated glomerular filtration rate > 60 mL/minute/1.73 m2
Step 3: Act on the glycemic data
The AGP Report augments shared decision-making
with the patient, enhancing the ability to work together
to develop a plan focused on lifestyle and medication
where 50% of the values fall at that specific time), and
changes that address glycemic patterns identified in
5% to 95% lines as outer boundaries. The curve allows
Step 2. Abnormal patterns can be a target for intensi-
evaluation of crucial questions:
fication or reduction in therapy; they can also suggest
• Is there a pattern of dangerous hypoglycemia at a
potential changes such as reducing carbohydrate intake
certain time of day?
or increasing physical activity to improve troublesome
• Is there a postprandial pattern or hyperglycemia
patterns. For safety, we address patterns of hypoglycemia
throughout the entire day or night or both?
first and then consider hyperglycemia patterns, either at
• Is there excessive variability suggesting a role for
the current or at subsequent visits, to further optimize
modifying diet, physical activity, or medication
glycemic patterns. We recommend focusing on no more
adherence?
than 1 or 2 glycemic patterns of concern at a time.
The AGP curve shows patterns of hypoglycemia (time
below range) and hyperglycemia (time above range)
■ CASE PRESENTATION
that indicate quickly where action is needed.
The goal of glycemic therapy is to optimize normo- Michael, a 65-year-old man on a regimen of basal and
glycemia while minimizing hypoglycemia. Therefore, bolus insulin (background and mealtime bolus insu-
the AGP curve can help focus therapies on interven- lin) along with noninsulin therapies, is not meeting
tions that reduce variability (ie, “flattening” the median glycemic goals. His demographic and clinical data are
line and reducing the spread of the 95% and 5% lines) outlined in Table 3 and his glucose data in Figure 2.
while decreasing hyperglycemia and hypoglycemia Based on the “Determine Where to Act” guide,
by improving time in range. The goal in evaluating the first step is to review time-in-ranges metrics, with
therapies based on the AGP curve is to move from special attention to time below range. Michael’s time
616 CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 91 • NUMBER 10 OCTOBER 2024

Downloaded from www.ccjm.org on November 20, 2024. For personal use only. All other uses require permission.
 MARTENS AND COLLEAGUES

Figure 2. Patient’s Ambulatory Glucose Profile Report.

©2024 International Diabetes Center, Minneapolis, MN. Used with permission. Visit AGPreport.org for more information.

CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 91 • NUMBER 10 OCTOBER 2024 617

Downloaded from www.ccjm.org on November 20, 2024. For personal use only. All other uses require permission.
CONTINUOUS GLUCOSE MONITORING DATA

TABLE 4
Coding for continuous glucose monitoring
CPT code Description Comments
95249 Personal (patient-owned) CGM: sensor placement, hook-up, One-time code for initial start-up and education
calibration, patient training, and printout

95250 Professional CGM (office-owned CGM), sensor placement, Billing code covers the cost of sensors and placement
hook-up, calibration, patient training, removal of sensor, by clinician/staff
and printout

95251 CGM data analysis and interpretation with report by clinician Can be billed no more frequently than every 30 days
Coding guidelines
72 hours of data are required for billing any of these codes.
-25 modifier for CGM codes can be used if billing for CGM interpretation on the same day as a Problem Visit code (99212-99215).
If a significant and separately identifiable service took place:
• 99212-99215: Pre-CGM evaluation (+) -25 95249: CGM start-up and instruction
• 99212-99215: E and M code for problem visit (+) -25 95251: CGM analysis, interpretation, and report.
CGM = continuous glucose monitoring; CPT = current procedural terminology

in range, 34%, is well below the clinical target of 70% daily dose of insulin equally between basal insulin and
or greater, and his time below range is 0%. We quickly bolus insulin. This would “rebalance” the basal and
determine that action is needed to improve his glyce- bolus insulin by redistributing the total daily dose of
mic profile. insulin 50:50 between basal and bolus.
Step 2, “Identify where to act,” requires review With a current total daily insulin dose of 90 units
of the AGP curve and daily glucose profiles. Several (60 units of basal and 30 units of bolus insulin),
patterns are apparent. Michael has a “stairstep” rise in we would add 10% (roughly a total daily dose of
glycemia during the day, corresponding with breakfast, 100 units), split that between basal (50 units) and bolus
dinner, and an evening snack. Overnight, median glu- (50 units) dosing, and then divide the bolus insulin
cose drops from 250 mg/dL at midnight to 170 mg/dL at between the 3 meals for a new insulin regimen of
6 am. The pattern of an exaggerated overnight drop in 50 units of glargine at bedtime with 16 units of lispro
glucose and a stairstep rise during the day suggests too with meals. CGM-based management allows a more
much basal (background) insulin and too little bolus rapid cycle time. We could revisit titration in 2 weeks
(mealtime bolus) insulin. Michael’s average glucose of with a new AGP profile and continue titration until
203 mg/dL without hypoglycemia also demonstrates the regimen is optimized to match individual basal and
that the total daily dose of insulin is inadequate. bolus insulin needs.
Step 3, “Act on the glycemic data,” involves
adjusting Michael’s therapies. We address any pattern ■ CGM CLINICAL PEARLS
of hypoglycemia first, as that is the biggest short-term
risk to patients with diabetes. Michael has no signif- Modern CGM technology is typically straightforward
icant hypoglycemia, so our next move is to optimize and easy to use. Online videos and web-based instruc-
insulin therapy to address hyperglycemia. Michael’s tion can be helpful at start-up. Additionally, care
insulin regimen contains an excessive amount of team–based resources like trained and designated staff
basal insulin relative to mealtime insulin. As a rule of can help ensure that data are available to clinicians at
thumb, the balance between basal and bolus insulin the time of clinical interactions. Building the team is a
is typically 50:50 (with some individual variation in worthwhile effort to ensure success. Coding for CGM
this balance).32 This imbalance is reflected in the AGP is shown in Table 4.
curve, which shows a drop in glucose overnight (due Difficulties encountered by users of CGM technol-
to too much basal insulin), then a rise, with meals, ogy often revolve around problems with sensor adhe-
throughout the day (due to too little mealtime insulin). sion or with skin irritation and dermatitis. Trimming
A reasonable intervention would be to increase the of body hair in the area of sensor placement can be
total daily dose of insulin by 10%, then divide the total helpful, and various available skin protectants and
618 CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 91 • NUMBER 10 OCTOBER 2024

Downloaded from www.ccjm.org on November 20, 2024. For personal use only. All other uses require permission.
 MARTENS AND COLLEAGUES

barriers can help both with adhesion and irritation and variations between readings and between devices
issues. Adhesive overlays are widely available and can are to some degree expected. All CGM sensors are
address adhesion issues. For patients experiencing sig- known to be less accurate in the hypoglycemia range.
nificant challenges, local diabetes educators often have Concerning symptoms or discordant data may warrant
significant expertise in overcoming these challenges confirmation with an alternate technology. Unexpected
and can be an ideal resource. or outlying CGM data should optimally be confirmed
Some commercially available CGM sensors have with blood glucose monitoring if there are questions
not been approved for use with magnetic resonance regarding the validity of data. ■
imaging, computed tomography, or radiographic tech-
nologies, and consideration should be given to removal ■ DISCLOSURES
before such testing. We recommend checking with Dr. Martens has disclosed consulting for Sanofi; serving as an advisor or
the manufacturer’s recommendation for use of CGM review panel participant for Eli Lilly; teaching and speaking for Abbott
Diabetes Care, Dexcom, and Eli Lilly; serving as a research principal
sensors with these technologies. investigator for Abbott Diabetes Care, Dexcom, and Novo Nordisk; serving
Therapeutic substances can variably interfere with as a co-principal investigator for Capillary Biomedical, Eli Lilly, Insulet Cor-
poration, Sanofi, and Tandem Diabetes Care. Dr. Simonson has disclosed
glucose sensing by CGM sensors. Interference by ther- consulting for Abbott Diabetes Care and teaching and speaking for Abbott
apeutic quantities of acetaminophen has largely been Diabetes Care and Sanofi. Dr. Bergenstal has disclosed intellectual property
rights (royalties or patent sales) for Medtronic; consulting for Abbott Dia-
overcome, but high-dose aspirin and vitamin C can betes Care, Dexcom, Lilly, Medtronic, Novo Nordisk, and Tandem Diabetes;
affect glucose readings, as can hydroxyurea and, for some serving as an advisor or review panel participant for Abbott Diabetes Care,
Dexcom, Lilly, Medtronic, and Novo Nordisk; serving as a research principal
sensors, alcohol.33 Review of interfering substances based investigator for Abbott Diabetes Care, Insulet Corporation, Lilly, Medtronic,
on CGM manufacturer recommendations is advisable. Novo Nordisk, and Tandem Diabetes; and serving as a research co-principal
investigator for Abbott Diabetes Care, Dexcom, Insulet Corporation, Lilly,
Finally, no technology is immune from variance and Medtronic, Novo Nordisk, and Tandem Diabetes.
errors. Neither blood glucose monitoring nor CGM The authors’ employer, nonprofit HealthPartners Institute, contracts for
technology is a “gold standard” in evaluating glucose, their services and no author receives personal income for these services.

■ REFERENCES 9. Aronson R, Brown RE, Chu L, et al. IMpact of flash glucose Moni-
toring in pEople with type 2 Diabetes Inadequately controlled with
1. Role of continuous glucose monitoring in diabetes treatment. non-insulin Antihyperglycaemic ThErapy (IMMEDIATE): a random-
Arlington, VA: American Diabetes Association; 2018. ized controlled trial. Diabetes Obes Metab 2023; 25(4):1024–1031.
2. American Diabetes Association Professional Practice Committee. doi:10.1111/dom.14949
7. Diabetes technology: standards of care in diabetes—2024. 10. Grunberger G, Sherr J, Allende M, et al. American Association
Diabetes Care 2024; 47(suppl 1):S126–S144. of Clinical Endocrinology clinical practice guideline: the use of
doi:10.2337/dc24-S007 advanced technology in the management of persons with diabetes
3. Basu A, Dube S, Slama M, et al. Time lag of glucose from intravascular mellitus. Endocr Pract 2021; 27(6):505–537.
to interstitial compartment in humans. Diabetes 2013; 62(12): doi:10.1016/j.eprac.2021.04.008
4083–4087. doi:10.2337/db13-1132 11. Danne T, Nimri R, Battelino T, et al. International consensus on use
4. Beck RW, Riddlesworth TD, Ruedy K, et al. Continuous glucose of continuous glucose monitoring. Diabetes Care 2017; 40(12):
monitoring versus usual care in patients with type 2 diabetes receiv- 1631–1640. doi:10.2337/dc17-1600
ing multiple daily insulin injections: a randomized trial. Ann Intern 12. Xing D, Kollman C, Beck RW, et al. Optimal sampling intervals to
Med 2017; 167(6):365–374. doi:10.7326/M16-2855 assess long-term glycemic control using continuous glucose moni-
5. Martens T, Beck RW, Bailey R, et al. Effect of continuous glucose toring. Diabetes Technol Ther 2011; 13(3):351–358.
monitoring on glycemic control in patients with type 2 diabetes doi:10.1089/dia.2010.0156
treated with basal insulin: a randomized clinical trial. JAMA 2021; 13. Riddlesworth TD, Beck RW, Gal RL, et al. Optimal sampling dura-
325(22):2262–2272. doi:10.1001/jama.2021.7444 tion for continuous glucose monitoring to determine long-term
6. Haak T, Hanaire H, Ajjan R, Hermanns N, Riveline JP, Rayman G. glycemic control. Diabetes Technol Ther 2018; 20(4):314–316.
Flash glucose-sensing technology as a replacement for blood doi:10.1089/dia.2017.0455
glucose monitoring for the management of insulin-treated type 14. Ceriello A, Prattichizzo F, Phillip M, Hirsch IB, Mathieu C,
2 diabetes: a multicenter, open-label randomized controlled trial. Battelino T. Glycaemic management in diabetes: old and new ap-
Diabetes Ther 2017; 8(1):55–73. proaches. Lancet Diabetes Endocrinol 2022; 10(1):75–84.
doi:10.1007/s13300-016-0223-6 doi:10.1016/S2213-8587(21)00245-X
7. Beck RW, Riddlesworth T, Ruedy K, et al. Effect of continuous 15. Battelino T, Danne T, Bergenstal RM, et al. Clinical targets for con-
glucose monitoring on glycemic control in adults with type 1 dia- tinuous glucose monitoring data interpretation: recommendations
betes using insulin injections: the DIAMOND randomized clinical from the International consensus on time in range. Diabetes Care
trial. JAMA 2017; 317(4):371–378. 2019; 4 2(8):1593–1603. doi:10.2337/dci19-0028
doi:10.1001/jama.2016.19975 16. Beck RW, Bergenstal RM, Cheng P, et al. The relationships be-
8. Lind M, Polonsky W, Hirsch IB, et al. Continuous glucose monitoring tween time in range, hyperglycemia metrics, and HbA1c. J Diabe-
vs conventional therapy for glycemic control in adults with type 1 tes Sci Technol 2019; 13(4):614–626.
diabetes treated with multiple daily insulin injections: the GOLD doi:10.1177/1932296818822496
randomized clinical trial [published correction appears in JAMA 17. Vigersky RA, McMahon C. The relationship of hemoglobin A1C
2017; 317(18):1912]. JAMA 2017; 317(4):379–387. to time-in-range in patients with diabetes. Diabetes Technol Ther
doi:10.1001/jama.2016.19976 2019; 21(2):81–85. doi:10.1089/dia.2018.0310

CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 91 • NUMBER 10 OCTOBER 2024 619

Downloaded from www.ccjm.org on November 20, 2024. For personal use only. All other uses require permission.
CONTINUOUS GLUCOSE MONITORING DATA

18. Diabetes Control and Complications Trial Research Group, 26. Bergenstal RM, Beck RW, Close KL, et al. Glucose management
Nathan DM, Genuth S, et al. The effect of intensive treatment indicator (GMI): a new term for estimating A1C from continu-
of diabetes on the development and progression of long-term ous glucose monitoring. Diabetes Care 2018; 41(11):2275–2280.
complications in insulin-dependent diabetes mellitus. N Engl J doi:10.2337/dc18-1581
Med 1993; 329(14):977–986. 27. Holt RIG, DeVries JH, Hess-Fischl A, et al. The management of
doi:10.1056/NEJM199309303291401 type 1 diabetes in adults. a consensus report by the American
19. Intensive blood-glucose control with sulphonylureas or insulin
Diabetes Association (ADA) and the European Association for the
compared with conventional treatment and risk of complications in
Study of Diabetes (EASD). Diabetes Care 2021; 44(11):2589–2625.
patients with type 2 diabetes (UKPDS 33). UK Prospective Diabetes
doi:10.2337/dci21-0043
Study (UKPDS) Group [published correction appears in Lancet 1999;
354(9178):602]. Lancet 1998; 352(9131):837–853. 28. Krhač M, Lovrenčić MV. Update on biomarkers of glycemic control.
20. Beck RW, Bergenstal RM, Riddlesworth TD, et al. Validation of time World J Diabetes 2019; 10(1):1–15. doi:10.4239/wjd.v10.i1.1
in range as an outcome measure for diabetes clinical trials. Diabetes 29. Tozzo V, Genco M, Omololu SO, et al. Estimating glycemia from
Care 2019; 42(3):400–405. doi:10.2337/dc18-1444 HbA1c and CGM: analysis of accuracy and sources of discrepancy.
21. Lu J, Wang C, Shen Y, et al. Time in range in relation to all-cause Diabetes Care 2024; 47(3):460–466. doi:10.2337/dc23-1177
and cardiovascular mortality in patients with type 2 diabetes: 30. Ceriello A, Monnier L, Owens D. Glycaemic variability in diabetes:
a prospective cohort study. Diabetes Care 2021; 44(2):549–555. clinical and therapeutic implications. Lancet Diabetes Endocrinol
doi:10.2337/dc20-1862 2019; 7(3):221–230. doi:10.1016/S2213-8587(18)30136-0
22. Bergenstal RM, Hachmann-Nielsen E, Kvist K, Peters AL, Tarp JM, 31. Rama Chandran S, Tay WL, Lye WK, et al. Beyond HbA1c: compar-
Buse JB. Increased derived time in range is associated with reduced ing glycemic variability and glycemic indices in predicting hypogly-
risk of major adverse cardiovascular events, severe hypoglycemia, cemia in type 1 and type 2 diabetes [published correction appears
and microvascular events in type 2 diabetes: a post hoc analysis of
in Diabetes Technol Ther 2019; 21(5):303]. Diabetes Technol Ther
DEVOTE. Diabetes Technol Ther 2023; 25(6):378–383.
2018; 20(5):353–362. doi:10.1089/dia.2017.0388
doi:10.1089/dia.2022.0447
32. Samson SL, Vellanki P, Blonde L, et al. American Association of
23. Lu J, Ma X, Zhou J, et al. Association of time in range, as assessed by
continuous glucose monitoring, with diabetic retinopathy in type 2 Clinical Endocrinology consensus statement: comprehensive type 2
diabetes. Diabetes Care 2018; 41(11):2370–2376. diabetes management algorithm—2023 update [published correc-
doi:10.2337/dc18-1131 tion appears in Endocr Pract 2023; 29(9):746] [published correction
24. Mayeda L, Katz R, Ahmad I, et al. Glucose time in range and pe- appears in Endocr Pract 2023; 29(12):1025]. Endocr Pract 2023;
ripheral neuropathy in type 2 diabetes mellitus and chronic kidney 29(5):305–340. doi:10.1016/j.eprac.2023.02.001
disease. BMJ Open Diabetes Res Care 2020; 8(1):e000991. 33. Heinemann L. Interferences with CGM systems: practi-
doi:10.1136/bmjdrc-2019-000991 cal relevance? J Diabetes Sci Technol 2022; 16(2):271–274.
25. Raj R, Mishra R, Jha N, Joshi V, Correa R, Kern PA. Time in range, as doi:10.1177/19322968211065065
measured by continuous glucose monitor, as a predictor of micro-
vascular complications in type 2 diabetes: a systematic review. BMJ Address: Thomas W. Martens, MD, International Diabetes Center, Health-
Open Diabetes Res Care 2022; 10(1):e002573. Partners Institute, Suite 600, 3800 Park Nicollet Blvd., Saint Louis Park,
doi:10.1136/bmjdrc-2021-002573 MN 55416; [email protected]

620 CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 91 • NUMBER 10 OCTOBER 2024

Downloaded from www.ccjm.org on November 20, 2024. For personal use only. All other uses require permission.