CAM 10120

Continuous or Intermittent Monitoring of Glucose in Interstitial Fluid

Category:Durable Medical Equipment   Last Reviewed:January 2020
Department(s):Medical Affairs   Next Review:February 2020
Original Date:August 2000    

Description:
Tight glucose control in patients with diabetes has been associated with improved health outcomes. Several devices are available to measure glucose levels automatically and frequently (e.g., every 5-10 minutes). The devices measure glucose in the interstitial fluid and are approved as adjuncts to or replacements for traditional self-monitoring of blood glucose levels. Devices can be used on a long-term (continuous) or short-term(often referred to as intermittent) basis.

The following conclusions are based on a review of the evidence, including, but not limited to, published evidence and clinical expert opinion, solicited via BCBSA’s Clinical Input Process.

Type 1 Diabetes
For individuals with type 1 diabetes who are willing and able to use the device, and have adequate medical supervision, who receive long-term (continuous) glucose monitoring (CGM), the evidence includes randomized controlled trials (RCTs) and systematic reviews. Relevant outcomes are symptoms, morbid events, quality of life (QOL), and treatmentrelated morbidity. Systematic reviews have generally found that at least in the short-term, long-term CGM resulted in significantly improved glycemic control for adults and children with type 1 diabetes, particularly highly compliant patients. A 2017 individual patient data analysis, pooling data from 11 RCTs, found that reductions in hemoglobin A1c (HbA1c) levels were significantly greater with real-time CGM than with a control intervention. Two RCTs in patients who used multiple daily insulin injections and were highly compliant with CGM devices during run-in phases found that CGM was associated with a larger reduction in HbA1c levels than previous studies. One of the two RCTs prespecified hypoglycemiarelated outcomes and reported that time spent in hypoglycemia was significantly less in the CGM group. One RCT in pregnant women with type 1 diabetes, which compared real-time CGM with self-monitoring of blood glucose, has also reported a difference in change in HbA1clevels, an increased percentage of time in the recommended glucose control target range, a smaller proportion of infants who were large for gestational age, a smaller proportion of infants who had neonatal intensive care admissions lasting more than 24 hours, a smaller proportion of infants who had neonatal hypoglycemia requiring treatment, and reduced total hospital length of stay all favoring CGM. The evidence is sufficient that the longterm use of CGM provides an improvement in net health outcomes for persons with type 1 diabetes mellitus.

For individuals with type 1 diabetes who have poor control of diabetes despite the use of best practices or when basal insulin levels need to be determined prior to insulin pump initiation who receive short-term glucose monitoring, the evidence includes RCTs and systematic reviews. Relevant outcomes are symptoms, morbid events, QOL, and treatmentrelated morbidity as well as intermediate outcomes related to measures of glucose control such as frequency and time in hypoglycemia and hyperglycemia. The evidence for shortterm monitoring of glycemic control is mixed, and there was no consistency in HbA1c levels. Some trials have reported improvements in glucose control for the intermittent monitoring group but limitations in this body of evidence preclude conclusions. The definitions of control with short-term CGM use, duration of use and the specific monitoring protocols varied. In some studies, short-term monitoring was part of a larger strategy aimed at optimizing glucose control, and the impact of monitoring cannot be separated from the impact of other interventions. Studies have not shown an advantage for intermittent glucose monitoring in reducing severe hypoglycemia events but the number of events reported is generally small and effect estimates imprecise. The limited duration of use may preclude an assessment of any therapeutic effect. Two RCTs of short-term CGM use for monitoring in pregnancy included women with both type 1 and 2 diabetes, with most having type 1 diabetes. One trial reported a difference in HbA1c levels at 36 weeks; the proportion of infants that were large for gestational age (>90th percentile) favored CGM while the second trial did not. The differences in the proportions of infants born via cesarean section, gestational age at delivery, and infants with severe hypoglycemia were not statistically significant in either study. Limitations of the published evidence preclude determining the effects of the technology on net health outcome. Evidence reported through clinical input supports that this use provides a clinically meaningful improvement in net health outcome and is consistent with generally accepted medical practice when used in specific situations such as poor control of diabetes despite the use of best practices or when basal insulin levels need to be determined prior to insulin pump initiation. The evidence is sufficient to determine that the technology results in a meaningful improvement in the net health outcome.

Type 2 Diabetes
For individuals with type 2 diabetes who receive long-term CGM, the evidence includes RCTs. The relevant outcomes are symptoms, morbid events, QOL, and treatment-related morbidity. Most RCTs of CGM in patients with type 2 trials found statistically significant benefits of CGM regarding glycemic control. However, the degree of HbA1c reduction and the difference in HbA1c reduction between groups might not be clinically significant. Moreover, additional evidence would be needed to show what levels of improvements in HbA1c levels over the short-term would be linked to meaningful improvements over the long-term in health outcomes such as diabetes-related morbidity and complications. Also, the variability in entry criteria as well as among interventions makes it difficult to identify an optimal approach to CGM use; the studies used a combination of intermittent and continuous monitoring with a review of data in real-time or at study visits only. Only the DIAMOND RCT (n=158) has used real-time CGM in type 2 diabetes. Selected patients were highly compliant during a run-in phase. The difference in change in HbA1clevels from baseline to 24 weeks was -0.3% favoring CGM. The difference in the proportion of patients with a relative reduction in HbA1c level by 10% or more was 22% favoring CGM. There were no differences in the proportions of patients with an HbA1c level of less than 7% at week 24. There were no events of severe hypoglycemia or diabetic ketoacidosis in either group. The treatment groups did not differ in any of the QOL measures. RCTs using flash glucosesensing technology as a replacement for self-monitoring of blood glucose for the management of insulin-dependent treated type 2diabetes found no difference in HbA1c change at 6 and 12 months between groups. However, time in severe hypoglycemia (<45mg/dL) was reduced for intervention participants. Two trials of CGM have enrolled pregnant women with type 2 diabetes, but the total number of women with type 2 diabetes included in both trials is only 58. One study reported a difference in HbA1c levels at 36 weeks, and the proportion of infants that were large for gestational age (>90th percentile) favored CGM while the second study did not. Neither trial reported analyses stratified by diabetes type. Limitations of the published evidence preclude determining the effects of the technology on net health outcome. Evidence reported through clinical input for long-term (continuous) CGM in patients with type 2 diabetes who do not require insulin did not provide strong support of a safety benefit and clinically meaningful improvement in net health outcome. The evidence is insufficient to determine the effects of the technology on health outcomes.

For individuals with type 2 diabetes who are willing and able to use the device and have adequate medical supervision and who experience significant hypoglycemia on multiple daily doses of insulin or an insulin pump in the setting of insulin deficiency who receive long-term (continuous) glucose monitoring, the evidence includes a systematic review and nonrandomized study with 12-month follow-up. The relevant outcomes are the frequency of and time spend in hypoglyceimia, the incidence of hypoglycemic episodes, complications of hypoglycemia, and QOL. The available studies demonstrate that CGM can significantly reduce time in hypoglycemia and frequency of hypoglycemia events both during the day and at night. At 12-month follow-up, hypoglycemic events were reduced by 40.8% to 61.7% with a greater relative reduction in the most severe thresholds of hypoglycemia. The published evidence supports a meaningful improvement in the net health outcome. Evidence reported through clinical input provides additional clinical context and based on both the published evidence and clinical input the following patient selection criteria are associated with a clinically meaningful improvement in net health outcome and are consistent with generally accepted medical practice: selected patients with type 2 diabetes who are (1) willing and able to use the CGM device and have adequate medical supervision and (2) experiencing significant hypoglycemia on multiple daily doses of insulin or an insulin pump in the setting of insulin deficiency. The evidence is sufficient to determine that the technology results in a meaningful improvement in the net health outcome.

For individuals with type 2 diabetes who require multiple daily doses of insulin and have poor control of diabetes despite the use of best practices or when basal insulin levels need to be determined prior to insulin pump initiation who receive short-term CGM monitoring, the evidence includes RCTs and systematic reviews. The relevant outcomes are symptoms, morbid events, QOL, and treatment-related morbidity. Systematic reviews of three to four RCTs have found statistically significant benefits from CGM regarding glycemic control. However, the degree of HbA1c reduction and the difference in HbA1c reductions between groups may not be clinically significant. Also, the limited number of RCTs and variability among interventions make it difficult to identify an optimal approach to CGM or a subgroup of type 2 diabetes patients who might benefit. Moreover, studies of CGM in patients with type 2 diabetes have generally not addressed the clinically important issues of severe hypoglycemia and diabetic complications. Very few pregnant women with type 2 diabetes have been included in RCTs. Limitations of the published evidence preclude determining the effects of the technology on net health outcome. Evidence reported through clinical input for use of short-term CGM in patients with type 2 diabetes who require multiple daily doses of insulin supports that this use provides a clinically meaningful improvement in net health outcome and is consistent with generally accepted medical practice when used in specific situations such as poor control of diabetes despite use of best practices or when basal insulin levels need to be determined prior to insulin pump initiation. The evidence is sufficient to determine that the technology results in a meaningful improvement in the net health outcome.

Gestational Diabetes
For individuals who are pregnant with gestational diabetes who receive long-term CGM or short-term (intermittent) glucose monitoring, the evidence includes an RCT. The relevant outcomes are symptoms, morbid events, QOL, and treatment-related morbidity. In the RCT, the type of glucose monitoring was unclear. Trial reporting was incomplete; however, there was no difference between the groups for most reported outcomes. The evidence is insufficient to determine the effects of the technology on health outcomes.

Tight glucose control in patients with diabetes has been associated with improved health outcomes. Several devices are available to measure glucose levels automatically and frequently (e.g., every 5-10 minutes). The devices measure glucose in the interstitial fluid and are approved as adjuncts to or replacements for traditional self-monitoring of blood glucose levels. Devices can be used on an intermittent (short-term) basis or a continuous (long-term) basis..

Background
Blood Glucose Control 
The advent of blood glucose monitors for use by patients in the home revolutionized the management of diabetes. Using fingersticks, patients can monitor their blood glucose levels both to determine the adequacy of hyperglycemia control and to evaluate hypoglycemic episodes. Tight glucose control, defined as a strategy involving frequent glucose checks and a target hemoglobin A1c (HbA1c) level in the range of 7%, is now considered standard of care for diabetic patients. Randomized controlled trials assessing tight control have demonstrated benefits for patients with type 1 diabetes in decreasing microvascular complications. The impact of tight control on type 1 diabetes and macrovascular complications such as stroke or myocardial infarction is less certain. The Diabetes Control and Complications Trial (2002) demonstrated that a relative HbA1c level reduction of 10% is clinically meaningful and corresponds to approximately a 40% decrease in risk for progression of diabetic retinopathy and 25% decrease in risk for progression of renal disease.1,

Due to an increase in turnover of red blood cells during pregnancy, HbA1c levels are slightly lower in women with a normal pregnancy compared with nonpregnant women. The target A1c in women with diabetes is also lower in pregnancy. The American Diabetes Association recommends that, if achievable without significant hypoglycemia, the A1c levels should range between 6.0% to 6.5%; an A1c levels less than 6% may be optimal as the pregnancy progresses.2,

Tight glucose control requires multiple daily measurements of blood glucose (ie, before meals and at bedtime), a commitment that some patients may be unwilling or unable to meet. Also, the goal of tight glucose control has to be balanced with an associated risk of hypoglycemia. Hypoglycemia is known to be a risk in patients with type 1 diabetes. While patients with insulin-treated type 2 diabetes may also experience severe hypoglycemic episodes, there is a lower relative likelihood of severe hypoglycemia compared with patients who had type 1 diabetes.3,4, An additional limitation of periodic self-measurements of blood glucose is that glucose levels are seen in isolation, and trends in glucose levels are undetected. For example, while a diabetic patient’s fasting blood glucose level might be within normal values, hyperglycemia might be undetected postprandially, leading to elevated HbA1c levels.

Management 
Recently, measurements of glucose in the interstitial fluid have been developed as a technique to measure glucose values automatically throughout the day, producing data that show the trends in glucose levels. Although devices measure glucose in the interstitial fluid on a periodic rather than a continuous basis, this type of monitoring is referred to as continuous glucose monitoring (CGM). 

Several devices have received approval from the U.S. Food and Drug Administration (FDA). The first approved devices were the Continuous Glucose Monitoring System (MiniMed), which uses an implanted temporary sensor in the subcutaneous tissues, and the GlucoWatch G2 Biographer, an external device worn like a wristwatch that measures glucose in interstitial fluid extracted through the skin by electric current (referred to as reverse iontophoresis). 

Devices subsequently approved include those for pediatric use and those with more advanced software, more frequent measurements of glucose levels, or more sophisticated alarm systems. Devices initially measured interstitial glucose every 5 to 10 minutes and stored data for download and retrospective evaluation by a clinician. With currently available devices, the intervals at which interstitial glucose is measured ranges from every 1 to 2 minutes to 5 minutes, and most provide measurements in real-time directly to patients. While CGM potentially eliminates or decreases the number of required daily fingersticks, it should be noted that, according to the FDA labeling, monitors are not intended as an alternative to traditional self-monitoring of blood glucose levels but rather as adjuncts to monitoring, supplying additional information on glucose trends not available from self-monitoring. Also, devices may be used intermittently (ie, for periods of 72 hours) or continuously (ie, on a long-term basis). 

In addition to stand-alone continuous glucose monitors, several insulin pump systems have built-in CGM. This evidence review addresses CGM devices, not the insulin pump portion of these systems..

Regulatory Status
Multiple CGM systems have been approved by the Food and Drug Administration through the premarket approval process (see Table 1).

CGM devices labeled as “Pro” for specific professional use with customized software and transmission to health care professionals are not enumerated in this list.

Table 1. CGM Systems Approved by the Food and Drug Administration

Device Manufacturer Approval Indications

Continuous Glucose Monitoring System (CGMS®)

MiniMed 1999  3-d use in physician's office
GlucoWatch G2® Biographera   2001  Not available since 2008
Guardian®-RT (Real-Time) CGMS MiniMed (now Medtronic) 2005  
Dexcom® STS CGMS system Dexcom 2006  
Paradigm® REAL-Time System (second generation called Paradigm Revel System) MiniMed (now Medtronic) 2006 Integrates CGM with a Paradigm insulin pump 
FreeStyle Navigator® CGM System Abbott  2008  
Dexcom® G4 Platinum Dexcom  2012 Adults ≥18 y; can be worn for up to 7 d; 
    2014 Expanded to include patients with diabetes 2-17 y 
Dexcom® G5 Mobile CGM Dexcom  2016a Replacement for fingerstick blood glucose testing in patients ≥2 y. System requires at least 2 daily fingerstick tests for calibration purposes, but additional fingersticks are not necessary because treatment decisions can be made based on device readings5, 
Dexcom® G6 Continuous Glucose Monitoring System  Dexcom  2018 

Indicated for the management of diabetes in persons age ≥2 years.

Intended to replace fingerstick blood glucose testing for diabetes treatment decisions.

Intended to autonomously communicate with digitally connected devices, including automated insulin dosing (AID) systems. with 10-day wear 

Freestyle Libre® Pro Flash Glucose Monitoring System  Abbott  2017  Adults ≥18 y. Indicated for the management of diabetes and can be worn up to 10 days It is designed to replace blood glucose testing for diabetes treatment decisions. 
Freestyle Libre® Flash Glucose Monitoring System Abbott  2018 

Adults ≥18 y.

Extended duration of use to 14 days

Guardian Connect  Medtronic MiniMed 2018 

Adolescents and adults (14-75 years)

Continuous or periodic monitoring of interstitial glucose levels.

Provides real-time glucose values, trends, and alerts through a Guardian Connect app installed on a compatible consumer electronic mobile device 

Eversense Continuous Glucose Monitoring System  Senseonics 

2018

2019 

Adults ≥18 y.

Continually measuring glucose levels up to 90 days.

Use as an adjunctive device to complement, not replace, information obtained from standard home blood glucose monitoring devices.

Adults ≥18 y.

Continually measuring glucose levels up to 90 days.

Indicated for use to replace fingerstick blood glucose measurements for diabetes treatment decisions.

Historical data from the system can be interpreted to aid in providing therapy adjustments. 

CGM: continuous glucose monitoring.
a  As a supplement to the G4 premarketing approval. 

Food and Drug product codes: MDS, PQF. QCD.

Related Policies
10103 Blood Glucose Monitors (Glucometers)

10130 Artificial Pancreas Device Systems

Policy
Short-Term Use – (up to three days) 
Short term use of CGMS can be beneficial in patients with diabetes to detect nocturnal hypoglycemia, the dawn phenomenon, and postprandial hyperglycemia and to assist in the management of hypoglycemic unawareness and when significant changes are made to their diabetes regimen (such as instituting new insulin or to pump therapy) as follows: 

Continuous glucose monitoring  may be considered MEDICALLY NECESSARY when used for up to 72 hours as a diagnostic test is covered without prior authorization for Type I and Type II patients on insulin. 

Long-Term Use
Long-term use of CGMS in the treatment of Type I and Type II diabetes who require insulin may be considered MEDICALLY NECESSARY for those who meet all of the following criteria

  • Patient has completed a comprehensive diabetic education program (a primary caregiver may complete this program for pediatric patients); and
  • Patient has a documented frequency of glucose self-testing an average of at least 4 times per day; and
  • Insulin injections are required 3 or more times per day; and
  • Patient self-adjusts insulin dose based on self-testing results and meets one or more of the following:
    • Glycosylated HbA1c values of 7 or greater, or
    • Documented Inadequate glycemic control despite compliance with frequent self-testing and fasting hyperglycemia (greater than 150 mg/dL) or frequent recurring episodes of severe hypoglycemia (less than 70 mg/dL), or
    • Documented hypoglycemia unawareness, episodes of ketoacidosis, or hospitalizations for uncontrolled glucose levels, or
    • Frequent nocturnal hypoglycemia despite appropriate modifications in insulin therapy, or
    • A newly pregnant female with Type I or Type II or one that has developed gestational diabetes that requires insulin therapy

A certificate of medical necessity (CMN) may be substituted for medical record documentation if it addresses all of the criteria noted above and is signed in attestation by the treating physician.

Other uses of continuous monitoring of glucose levels in interstitial fluid as a technique of diabetic monitoring are considered investigational and/or unproven and therefore considered NOT MEDICALLY NECESSARY.

The use of implantable CGM devices is investigational and/or unproven and therefore considered NOT MEDCICALLY NECESSARY.

Policy Guidelines
This policy only evaluates continuous (real time or intermittent) intersitital glucose monitors and does not evaluate insulin pumps. Insulin pump systems with a built-in continuous glucose monitor and a low-glucose suspend feature are addressed in evidence review 10130 (artificial pancreas device systems). 

Short-term intermittent monitoring is generally conducted over 72-hour periods. It may be repeated subsequently depending on the patient’s level of diabetes control. 

Best practices in diabetes control include compliance with a self-monitoring blood glucose regimen of 4 or more fingersticks each day and use of an insulin pump or multiple daily injections of insulin. During pregnancy, 3 or more insulin injections daily could also be considered best practice for patients not on an insulin pump prior to the pregnancy. Prior shortterm (72-hour) use of an intermittent glucose monitor would be considered a part of best practices for those considering long-term use of a continuous glucose monitor. 

Significant hypoglycemia may include recurrent, unexplained, severe (generally blood glucose levels <50 mg/dL) hypoglycemia or impaired awareness of hypoglycemia that puts the patient or others at risk. 

Women with type 1 diabetes taking insulin who are pregnant or about to become pregnant with poorly controlled diabetes are another subset of patients to whom the policy statement on intermittent monitoring may apply. 

The strongest evidence exists for use of continuous glucose monitoring devices in patients age 25 and older. However, age may be a proxy for motivation and good control of disease, so it is also reasonable to select patients based on their ability to self-manage their disease, rather than their age. Multiple CGM devices have FDA labeling related to age.

Providers board-certified in endocrinology and/or providers with a focus on the practice of diabetes care may be considered qualified to evaluate and oversee individuals for continuous (i.e., long-term) monitoring. 

See the Codes table for details

Rationale  
This evidence review was created in August 2000 and has been updated regularly with searches of the MEDLINE database. The most recent literature update was performed through November 20, 2019. This review was informed by a TEC Assessment (2003).[6]

The following conclusions are based on a review of the evidence, including but not limited to, published evidence and clinical expert opinion, solicited via BCBSA’s Clinical Input Process.

Evidence reviews assess the clinical evidence to determine whether the use of technology improves the net health outcome. Broadly defined, health outcomes are the length of life, quality of life (QOL), and ability to function -- including benefits and harms. Every clinical condition has specific outcomes that are important to patients and managing the course of that condition. Validated outcome measures are necessary to ascertain whether a condition improves or worsens; and whether the magnitude of that change is clinically significant. The net health outcome is a balance of benefits and harms.

To assess whether the evidence is sufficient to draw conclusions about the net health outcome of technology, two domains are examined: the relevance, and quality and credibility. To be relevant, studies must represent one or more intended clinical uses of the technology in the intended population and compare an effective and appropriate alternative at a comparable intensity. For some conditions, the alternative will be supportive care or surveillance. The quality and credibility of the evidence depend on study design and conduct, minimizing bias and confounding that can generate incorrect findings. The randomized controlled trial (RCT) is preferred to assess efficacy; however, in some circumstances, nonrandomized studies may be adequate. RCTs are rarely large enough or long enough to capture less common adverse events and long-term effects. Other types of studies can be used for these purposes and to assess generalizability to broader clinical populations and settings of clinical practice.

Technical reliability is outside the scope of these reviews, and credible information on technical reliability is available from other sources.

The evidence review focuses on the clinical utility of continuous glucose monitoring (CGM) systems. That is, their ability to provide additional information on glucose levels leads to improved glucose control, or to reduce the morbidity and mortality associated with clinically significant severe and acute hypoglycemic or hyperglycemic events. Because diabetic control encompasses numerous variables, including the diabetic regimen and patient self-management, RCTs are important to isolate the contribution of interstitial glucose measurements to overall diabetes management.

For the evaluation of the clinical utility of CGM, studies would need to use the test as either an adjunct or a replacement to current disease status measures to manage treatment decisions in patients with diabetes. Outcomes would include measures of glucose control, QOL and measures of disease progression.

CGM Devices for Long-Term Use in Type 1 Diabetes
In some parts of the analysis of type 1 diabetes, BCBSA combines discussion of indications 1 (long-term) and 2 (short-term) glucose monitoring because several systematic reviews and RCTs provided information relevant to both indications.

Clinical Context and Therapy Purpose
The purpose of long-term CGM glucose monitoring devices is to provide a testing option that is an alternative to or an improvement on existing testing used in the management of individuals with type 1 diabetes.

The question addressed in this evidence review is: Does long-term use of a CGM device improve the net health outcome for individuals with type 1 diabetes?

The following PICOs were used to select literature to inform this review.

Patients
The relevant population of interest is individuals with type 1 diabetes. All individuals with type 1 diabetes require engagement in a comprehensive self-management and clinical assessment program that includes assessment of blood glucose control.

Interventions
The testing being considered is the use of a CGM device to assess blood glucose levels as part of optimal diabetes management.

Comparators
The following practice is currently being used to measure glucose levels: capillary blood sampling (finger stick) for self-monitoring of blood glucose (SMBG). Standard treatment for patients with type 1 diabetes includes injection of long-acting basal insulin plus multiple daily injections (MDI) of rapid-acting insulin boluses as required for meal intake. Activity level may require patients need to modify the timing and dose of insulin administration. Individuals with type 1 diabetes may also use an insulin pump either for initial treatment or convert to pump use after a period of MDI. Individuals are required to check their blood glucose before making preprandial insulin calculations, in response to symptoms of hypoglycemia or related to activity-related insulin adjustments.

Outcomes
The general outcomes of interest are change in hemoglobin A1c (HbA1c) levels, time spent in hypoglycemia and hyperglycemia, time in range (generally glucose of 70-180 mg/dl), the incidence of hypoglycemic events, complications of hypoglycemia, and QOL. To assess short-term outcomes such as HbA1c levels, a minimum follow-up of 8 to 12 weeks is appropriate.

Study Selection
Methodologically credible studies were selected using the following principles:

  1. To assess efficacy outcomes, comparative controlled prospective trials were sought, with a preference for RCTs;
  2. In the absence of such trials, comparative observational studies were sought, with a preference for prospective studies.
  3. To assess long-term outcomes and adverse effects, single-arm studies that capture longer periods of follow-up and/or larger populations were sought.
  4. Studies with duplicative or overlapping populations were excluded.

Systematic Reviews
A number of systematic reviews and meta-analyses have assessed RCTs evaluating CGM for long-term, daily use in treating type 1 diabetes.7,8,9,10,11,12, These systematic reviews have focused on slightly different populations, and some did not separate long-term CGM from intermittent glucose monitoring.[10] The most recent meta-analysis and the only analysis to use individual patient data were published by Benkhadra et al (2017).[13] The meta-analysis evaluated data from 11 RCTs that enrolled patients with type 1 diabetes and compared real-time CGM with a control intervention. Studies in which patients used insulin pumps or received multiple daily insulin injections were included. Reviewers contacted corresponding study authors requesting individual patient data; data were not obtained for one trial. Mean baseline HbA1c levels were 8.2% in adults and 8.3% in children and adolescents. The overall risk of bias in the studies was judged to be moderate. In pooled analyses, there was a statistically significantly greater decrease in HbA1c levels with realtime CGM vs control conditions. Overall, the degree of difference between groups was 0.26%. In subgroup analyses by age, there was a significantly greater change in HbA1c levels among individuals 15 years and older, but not among the younger age groups. There were no significant differences between groups in the time spent in hypoglycemia or the incidence of hypoglycemic events. Key findings are shown in Table 2.

Table 2. Individual Patient Data Meta-Analytic Outcomes for Real-Time CGM in Type 1 Diabetes    

No. of Trials

N

Group

Point Estimate

95% Confidence Intervals

p

Change in HbA1c levels, %

8

1371

Overall

-0.258

0.464 to -0.052

0.014

7

902

Age >15 y

-0.356

0.551 to -0.160

<0.001

7

178

Age 13-15 y

-0.039

-0.320 to 0.242

0.787

7

291

Age ≤12 y

-0.047

0.217 to 0.124

0.592

Time spent in hypoglycemia <60 mg/dL, min

4

706

Overall

-8.549

-31.083 to 13 985

0.457

4

467

Age >15 y

-8.095

-32.615 to 16.425

0.518

3

109

Age 13-15 y

-13.966

31.782 to 3.852

0.124

3

130

Age ≤12 y

-9.366

19.898 to 1.167

0.081

Incidence of hypoglycemic events <70 mg/dL, mean no. events

3

351

Overall

0.051

-0.314 to 0.416

0.785

277 

Age >15 y

-0.074 

-0.517 to 0.368 

0.742 

47 

Age 13-15 y 

0.536 

0.243 to 1.316   

0.177 

27 

Age ≤12 y 

0.392 

0.070 to 0.854 

0.097 

Adapted from Benkhadra et al (2017).13,
CGM: continuous glucose monitoring: HbA1c: hemoglobin A1c.

Earlier meta-analyses of glucose monitoring devices for type 1 diabetes tended to combine studies of intermittent glucose monitoring with studies of long-term CGM. Several reported separate subgroup analyses for long-term CGM. A Cochrane review by Langendam et al (2012) assessed CGM in type 1 diabetes in adults and children included RCTs; it compared CGM with conventional SMBG.[9] In pooled analysis (6 studies; n=963 patients) of studies of long-term CGM, the average decline in HbA1c levels 6 months after baseline was statistically significantly larger for CGM users than for SMBG users (mean difference [MD], -0.2%; 95% confidence interval [CI], -0.4% to -0.1%), but there was no difference in the decline in HbA1c levels at 12 months (1 study, n=154 patients; MD, 0.1%; 95% CI, -0.5% to 0.7%). In a meta-analysis of 4 RCTs (n=689 patients), there was no significant difference in the risk of severe hypoglycemia between CGM and SMBG users and the CI for the relative risk was wide (relative risk, 1.05; 95% CI, 0.63 to 1.77), indicating lack of precision in estimating the effect of CGM on hypoglycemia risk. Reviewers were unable to compare the longer-term changes in HbA1c levels or hypoglycemia outcomes for real-time CGM. Trials reporting results by compliance subgroups found larger treatment effects in highly compliant patients.

A systematic review by Wojciechowski et al (2011) evaluating CGM included RCTs conducted in adults and children with type 1 diabetes.[11] Reviewers selected studies having a minimum of 12 weeks of follow-up and requiring patients to be on intensive insulin regimens. Studies compared CGM with SMBG; there was no restriction on the type of CGM device but CGM readings had to be used to adjust insulin dose or modify diet. Fourteen RCTs met the eligibility criteria. Study durations ranged from three to six months. Baseline mean HbA1c levels ranged from 6.4% to 10%. Five included studies found a statistically significant decrease in HbA1clevels favoring CGM, while nine did not. In a pooled analysis, there was a statistically significant reduction in HbA1clevels with CGM compared with SMBG (weighted mean difference [WMD], -0.26%; 95% CI, -0.34% to -0.19%). For the subgroup of 7 studies that reported on long-term CGM, this difference was statistically significant (WMD = -0.26; 95% CI, -0.34 to -0.18). In a subgroup analysis by age, there were significant reductions in HbA1c levels with CGM in 5 studies of adults (WMD= -0.33; 95% CI, -0.46 to -0.20) and in 8 studies with children and/or adolescents (WMD= -0.25; 95% CI, -0.43 to -0.08). Four of the studies provided data on the frequency of hypoglycemic episodes. Pooled results showed a significant reduction in hypoglycemic events for CGM vs SMBG (standardized mean difference, -0.32; 95% CI, -0.52 to -0.13). In five studies reporting the percentage of patients with severe hypoglycemic episodes, there were no differences in the percentages of patients with severe hypoglycemic episodes using CGM and SMBG. 

Randomized Controlled Trials 
Recent RCTs not included in the meta-analyses above are described next. For example, van Beers et al (2016) published a crossover RCT comparing CGM with SMBG and focused on patients with impaired hypoglycemia awareness.[14] Eligible patients were 18 to 75 years old, were treated with insulin infusion pumps or multiple daily insulin injections, undertook at least 3 SMBG measurements per day, and had impaired awareness of hypoglycemia (ie, Gold score ≥415). The trial used an artificial pancreas device system without using the low-glucose suspend feature. After a 6-week run-in phase (during which patients received education about diabetes management), 52 patients received both 16 weeks of CGM and 16 weeks of SMBG, in random order. There was a 12-week washout period between interventions. All patients were included in the primary intention-to-treat analysis. Six patients withdrew early from the study. 

The primary outcome (time spent in normoglycemia [4-10 mmol/L]) was significantly higher in the CGM phase than in the SMBG phase. The percentage of time spent in normoglycemia was 65.0% in the CGM phase and 55.4% in the SMBG group (MD=9.6%; p<0.001). The sequence allocation did not affect the primary endpoint. Most other CGMderived outcomes (eg, number and duration of nocturnal hypoglycemia events) also significantly favored the CGM group. The total number of severe hypoglycemic events (ie, those needing third-party assistance) was 14 in the CGM phase and 34 in the SMBG phase, which differed significantly between groups (p=0.033). The number of patients with 1 or more severe hypoglycemic events during the intervention period, however, did not differ significantly between phases 10 in the CGM phase and 18 in the SMBG phase (p=0.062). HbA1coutcomes did not differ significantly (eg, change in HbA1c levels from baseline was -0.1% in both phases; p=0.449). Regarding hypoglycemia awareness (1 of 4 variables), the Gold score at the study endpoint differed significantly (mean, 4.6 for the CGM phase vs 5.0 for the SMBG phase, p=0.035); 3 other variables related to hypoglycemia awareness did not differ between groups. 

Two, 2017 RCTs evaluated long-term CGM in patients with type 1 diabetes treated with multiple daily insulin injections. Both trials used the Dexcom G4 CGM device. Lind et al (2017) reported on a crossover study with 142 adults ages 18 and older who had baseline HbA1c levels of 7.5% or higher (mean baseline HbA1c level, »8.5%).[16] There was a 6- week run-in period using a CGM device with masked data and patients were excluded from further participation if they did not believe they would use the device more than 80% of the time or did not perform an adequate number of calibrations during the run-in period. Enrolled patients underwent 26-week treatment periods with a CGM device and conventional therapy using SMBG, in random order. There was a 17-week washout period between intervention phases. The primary endpoint was the difference in HbA1c levels at the end of each treatment period. Mean HbA1c levels were 7.9% during CGM use and 8.4% during conventional therapy (MD = -0.4%; p<0.01). There were a large number of secondary endpoints. A portion of them was prespecified, and analyses took into consideration the statistical impact of multiple comparisons; the remaining secondary outcomes were considered descriptive, and p-values were not reported. Among the prespecified secondary outcomes, treatment satisfaction (measured by the Diabetes Treatment Satisfaction Questionnaire) was significantly higher in the CGM phase than in the conventional treatment phase (p<0.001). Hypoglycemia outcomes were secondary descriptive outcomes. There was 1 (0.7%) severe hypoglycemic event during the CGM phase and 5 (3.5%) events during conventional therapy. The percentage of time with hypoglycemia (<70 mmol/L) was 2.8% during CGM treatment and 4.8% during conventional therapy. 

In the second study, Beck et al (2017) randomized 158 patients on a 2:1 basis to 24 weeks of CGM (n=105) or usual care (n=53).[17] The trial included patients with type 1 diabetes who were aged 25 or older and had baseline HbA1c levels between 7.5% and 10%. Before randomization, patients underwent a two-week period using a CGM system (without seeing data from the CGM) to ensure compliance. To be eligible, patients had to wear the CGM on at least 85% of days, calibrate the device at least twice daily, and perform SMBG at least 3 times daily. The primary outcome (change in HbA1c levels at 24 weeks) was 1.0% in the CGM group and 0.4% in the usual care group (p<0.001), with a between-group difference of 0.6%. Prespecified secondary outcomes on the proportion of patients below a glycemic threshold at 24 weeks also favored the CGM group. The proportion of patients with HbA1c levels less than 7.0% was 18 (18%) in the CGM group and 2 (4%) in the control group (p=0.01). The proportion of patients with HbA1c levels less than 7.5% was 39 (38%) in the CGM group and 6 (11%) in the control group (p<0.001). Moreover, prespecified secondary outcomes related to hypoglycemia also differed significantly between groups, favoring the CGM group. The time spent in hypoglycemia less than 70 mg/dL was 43 minutes per day in the CGM group and 80 minutes per day in the usual care group (p=0.002). Comparable numbers for time spent at less than 50 mg/dL were 6 minutes per day in the CGM group and 20 minutes per day in the usual care group (p=0.001). The median change in the rate per 24 hours of hypoglycemia events lasting at least 20 minutes at less than 3.0 mmol/L (54 mg/dL) fell by 30% from 0.23 at baseline to 0.16 during follow-up in the CGM group but was practically unchanged (0.31 at baseline and 0.30 at follow-up) in the usual care group (p=0.03).[18]QOL measures assessing overall well-being (World Health Organization Well-Being Index), health status (EQ-5D-5L), diabetes distress (Diabetes Distress Scale), hypoglycemic fear (worry subscale of the Hypoglycemia Fear Survey), and hypoglycemic confidence (Hypoglycemic Confidence Scale) have also been reported.[19] There were no significant differences between CGM and usual care in changes in wellbeing, health status, or hypoglycemic fear. The CGM group demonstrated a greater increase in hypoglycemic confidence (p=0.01) and a greater decrease in diabetes distress (p=0.01) than the usual care group. 

Pregnant Women 
One trial of real-time CGM in pregnant women with type 1 diabetes has been reported. Study characteristics, results, and gaps are summarized here and in Tables 3 to 6. Feig et al (2017) reported results of 2 multicenter RCTs in women ages 18 to 40 with type 1 diabetes who were receiving intensive insulin therapy and who were either pregnant (≤13 weeks and 6 days of gestation) or planning a pregnancy.[20] The trial enrolling pregnant women is reviewed here. Women were eligible if they had a singleton pregnancy and HbA1c levels between 6.5% and 10.0%. The trial was conducted at 31 hospitals in North America and Europe. Women were randomized to CGM (Guardian REAL-Time or MiniMed Minilink system) plus capillary glucose monitoring or capillary glucose monitoring alone. Women in the CGM group were instructed to use the devices daily. Women in the control group continued their usual method of capillary glucose monitoring. The target glucose range was 3.5 to 7.8 mmol/L and target HbA1c levels were 6.5% or less in both groups. The primary outcome was the difference in change in HbA1c levels from randomization to 34 weeks of gestation. The proportion of completed scheduled study visits was high in both groups; however, participants using CGM had more unscheduled contacts, which were attributed both to sensor issues and to sensor-related diabetes management issues. The median frequency of CGM use was 6.1 days per week (interquartile range, 4.0-6.8 d/wk) and 70% of pregnant participants used CGM for more than 75% of the time. The between-group difference in the change in HbA1c levels from baseline to 34 weeks of gestation was statistically significant favoring CGM (MD = -0.19%; 95% CI, -0.34 to -0.03; p=0.02). Women in the CGM group spent an increased percentage of time in the recommended glucose control target range at 34 weeks of gestation (68% vs 61%, p=0.003). There were no betweengroup differences in maternal hypoglycemia, gestational weight gain, or total daily insulin dose. A smaller proportion of infants of mothers in the CGM group were large-forgestational-age (odds ratio [OR], 0.51; 95% CI, 0.28 to 0.90; p=0.02). In addition, for infants of mothers in the CGM group, there were fewer neonatal intensive care admissions lasting more than 24 hours (OR=0.48; 95% CI, 0.26 to 0.86; p=0.02), fewer incidences of neonatal hypoglycemia requiring treatment with intravenous dextrose (OR=0.45, 0.22 to 0.89; p=0.025), and reduced total hospital length stay (3.1 days vs 4.0 days; p=0.0091). Skin reactions occurred in 49 (48%) of 103 CGM participants and 8 (8%) of 104 control participants. 

Table 3. RCT Characteristics for Real-Time CGM in Pregnant Women With Type 1 Diabetes 

Study; Registration

Countries

Sites

Dates

Participants

Interventions

 

 

 

 

 

Active

Comparator

Feig et al (2017)20,; NCT01788527

Canada, England, Scotland, Spain, Italy, Ireland, U.S.

31

 

2013-

2016

Pregnant women (<14 wk gestation) with type 1 diabetes receiving intensive insulin therapy with HbA1c levels between 6.5% and 10.0% (mean, 6.9%); mean age, 31 y

CGM (real-time) (n=108)

SMBG (n=107)

CGM: continuous glucose monitoring: HbA1c: hemoglobin A1c; RCT: randomized controlled trial; SMBG: self- monitored blood glucose.

Table 4. RCT Outcomes for Real-Time CGM in Pregnant Women With Type 1 Diabetes 

 

Infant

 

Maternal

Study

Large-for- Gestational Age

Gestational Age at Delivery, wk

Severe Hypoglycemia

Caesarean Section

HbA1c Levels: Change From Baseline to 34 Wk of Gestation

Severe Hypoglycemia

Feig et al (2017)20,

 

 

 

 

 

 

n

211

201

200

202

173

214

CGM

53 (53%)

Median, 37.4

15 (15%)

63 (63%)

-0.54

11 (11%)

Control

69 (69%)

Median, 37.3

28 (28%)

74 (73%)

-0.35

12 (12%)

TE (95% CI)

OR=0.51 (0.28 to 0.90)

NR

OR=0.45 (0.22 to 0.89)

NR

-0.19% (-0.34% to -0.03%)

NR

p

0.02

0.50

0.025

0.18

0.02

1.0

Values are n or n (%) or as otherwise indicated.
CI: confidence interval; CGM: continuous glucose monitoring; HbA1c: hemoglobin A1c; NR: not reported; OR: odds ratio; RCT: randomized controlled trial; TE: treatment effect. 

The purpose of the gaps tables (see Tables 5 and 6) is to display notable gaps identified in each study. This information is synthesized as a summary of the body of evidence following each table and provides the conclusions on the sufficiency of evidence supporting the position statement.

Table 5. Relevance Gaps of RCTs for Real-Time CGM in Pregnant Women With Type 1 Diabetes 

Study

Populationa

Interventionb

Comparatorc

Outcomesd

Follow-Upe

Feig et al (2017)20,

4. Run-in period requirement may have biased selection to highly compliant participants

3. More unscheduled contacts in CGM group

3. More unscheduled contacts in CGM group

 

 

The study limitations stated in this table are those notable in the current review; this is not a comprehensive limitations assessment. CGM: continuous glucose monitoring; RCT: randomized controlled trial. 
a Population key: 1. Intended use population unclear; 2. Clinical context is unclear; 3. Study population is unclear; 4. Study population not representative of intended use. 
b Intervention key: 1. Not clearly defined; 2. Version used unclear; 3. Delivery not similar intensity as comparator; 4.Not the intervention of interest. 
c Comparator key: 1. Not clearly defined; 2. Not standard or optimal; 3. Delivery not similar intensity as intervention; 4. Not delivered effectively. 
d Outcomes key: 1. Key health outcomes not addressed; 2. Physiologic measures, not validated surrogates; 3. No CONSORT reporting of harms; 4. Not establish and validated measurements; 5. Clinical significant difference not prespecified; 6. Clinical significant difference not supported. 
e Follow-Up key: 1. Not sufficient duration for benefit; 2. Not sufficient duration for harms.

Table 6. Study Design and Conduct Gaps of RCTs for Real-Time CGM in Pregnant Women With Type 1 Diabetes

Study

Allocationa

Blindingb

Selective Reportingd

Data Completenesse

Powerd

Statisticalf

Feig et al (2017)20,

 

1. Not blinded; chance of bias in clinical management

 

 

 

3, 4. Treatment effects and confidence intervals not calculated for some outcomes

The study limitations stated in this table are those notable in the current review; this is not a comprehensive limitations assessment. CGM: continuous glucose monitoring; RCT: randomized controlled tria. 
a Allocation key: 1. Participants not randomly allocated; 2. Allocation not concealed; 3. Allocation concealment unclear; 4. Inadequate control for selection bias. 
b Blinding key: 1. Not blinded to treatment assignment; 2. Not blinded outcome assessment; 3. Outcome assessed by treating physician.
c Selective Reporting key: 1. Not registered; 2. Evidence of selective reporting; 3. Evidence of selective publication.
d Data Completeness key: 1. High loss to follow-up or missing data; 2. Inadequate handling of missing data; 3. High number of crossovers; 4. Inadequate handling of crossovers; 5. Inappropriate exclusions; 6. Not intent to treat analysis (per protocol for noninferiority trials). 
e Power key: 1. Power calculations not reported; 2. Power not calculated for primary outcome; 3. Power not based on clinically important difference. 
f Statistical key: 1. Analysis is not appropriate for outcome type: (a) continuous; (b) binary; (c) time to event; 2. Analysis is not appropriate for multiple observations per patient; 3. Confidence intervals and/or p values not reported; 4. Comparative treatment effects not calculated. 

CGM Implanted Device for Long-Term Use
The Eversense Continuous Glucose Monitoring System is implanted in the subcutaneous skin layer and provides continuous glucose measurements over a 40-400 mg/dL range. The system provides real-time glucose values, glucose trends, and alerts for hypoglycemia and hyperglycemia and low glucose through a mobile application installed on a compatible mobile device platform. The Eversense CGM System is a prescription device indicated or use in adults (age 18 and older) with diabetes for up to 90 days. The device was initially approved as an adjunctive glucose monitoring device to complement information obtained from standard home blood glucose monitoring devices. Prescribing providers are required to participate in insertion and removal training certification.

Three nonrandomized prospective studies and three postmarketing registry studies that were intended to evaluate the accuracy and safety of the device in adults are summarized in Tables 76 and 87. In the three pivotal trials, accuracy measures included the mean absolute relative difference (MARD) between paired samples from the implanted device and a reference standard blood glucose measurement (Yellow Springs Instrument). Device development led to a demonstration of the increasing accuracy of the sensors. However, the accuracy tends to be lower in hypoglycemic ranges. Outcomes could not be differentiated for T1D vs T2D. Serious adverse events were primarily limited to skin reactions. Trends in secondary glycemic measures were variably reported but the studies were not designed to acquire clinical outcome data.

Data from the PRECISE, PRECISE II, AND PRECISION pivotal trials were provided to the Food and Drug Administration for the initial approval of Eversense as an adjunctive device. Expanded approval was granted in June 2019 and Eversense is now approved as a device to replace fingerstick blood glucose measurements for diabetes treatment decisions.[50]Historical data from the system can be interpreted to aid in providing therapy adjustments. No new clinical studies were conducted to support the change in the indications for the device. The sponsor previously performed clinical studies to establish the clinical measurement performance characteristics of the device, including accuracy across the claimed measuring range (40 to 400 mg/dL glucose), precision, claimed calibration frequency (every 12 hours), the wear period for the sensor (90 days), and performance of the alerts and notifications. This same clinical study information was used to support what the Food and Drug Administration considered a reasonable assurance of safety and effectiveness of the device for the replacement of fingerstick blood glucose monitoring for diabetes treatment decisions. As a condition of approval, the sponsor is required to conduct a post-approval-study. The study design is a nonblinded, prospective, multi-center, single-arm longitudinal cohort study intended to evaluate the safety and effectiveness of diabetes management with the Eversense CGM System non-adjunctively compared to self-monitoring of blood glucose using a blood glucose meter in participants with either Type 1 or Type 2 diabetes.[50] Subjects will serve as their own control, with baseline SMBG use to manage their diabetes for the first six months of the study followed by the use of the CGM nonadjunctively for the next six months. Total follow-up duration is 12 months. Approximately 925 subjects will be screened to achieve an enrollment such that approximately 740 subjects will be available for analysis at the end of the study. The investigation will include both clinic visits and home use of the device.

Three post-marketing registry studies of the Eversense device have been recently published. Glucometric results and adverse events are summarized in Table 8. Sanchez et al (2019) reported glucometric and safety data on the first 205 patients in the U.S. to use the Eversense device for at least 90 days (Table 8).[49] Of the 205 patients, 62.9% reported having T1D, 8.8% T2D, and 28.3% were unreported; results were not reported separately by diabetes type. Diess et al (2019) reported safety outcomes for 3,023 patients from 534 sites in Europe and South Africa who had used the device for 6 months or longer.[48] There were no serious adverse events, and the most commonly reported adverse events were sensor site infection and skin irritation. Tweden reported accuracy and safety data from 945 patients in Europe and South Africa who used either the 90-day or 180 day Eversense system for 4 insertion-removal cycles.[53] The percentage of patients using the 180-day system increased from cycle 1 to 4 as the device became more widely available (9%, 39%, 68% and 88% in cycles 1-4). There was no evidence of degradation of performance of the device over repeated insertion/removal cycles. Adverse events were not otherwise reported

Table 7. Summary of Key Nonrandomized Trials: Implanted CGM Study Characteristics 

Study Study Type Country Dates Participants Test/Treatment FollowUp

Kropff (2017)42

PRECISE

Prospective Single-arm Blinded Germany, Netherlands, U.K 2014- 2015 Adults (≥ 18 years) with T1 or T2 diabetes using insulin (N=71) Implanted CGM 180 days

Christiansen (2018)41

PRECISE II

Prospective Single-arm Blinded United States 2016

Adults (≥ 18 years) with T1D (67.8%) T2D (32.2%)

Insulin use: Total: (75.6%)

T2D: (6.6%)

(N=90)

Implanted CGM

Single sensor=75

Bilateral sensor =15

90 days

Christiansen (2019)44

PRECISION

Prospective Single-arm Blinded United States 2017- 2018

Adults (≥ 18 years) with T1D (71.4%) T2D (28.6%)

Insulin use: Total: (85.7%)

T2D: NR

(N=35)

Implanted CGM

Single sensor=8

Bilateral sensor=27

90 days
Deiss et al (2019)48

Prospective Single-arm Blinded

Postmarketing registry

Europe and South Africa 2016- 2018

Adults (≥ 18 years) with T1D or T2D (% not reported)

Consecutive patients who reached 4 sensor insertion/removal cycles

Total N=3023; 6 months of use (N=969), 1 year of use (N=173

Implanted CGM

Single sensor (90-day or 180 days)

Up to 1 year
Sanchez et al (2019)49

Prospective Single-arm Blinded

Postmarketing registry

United States 2018- 2019

Consecutive patients who reached a 90-day wear period of the device (62.9% T1D, 8.8% T2D, 28.3% unreported)

(N=205)

Implanted CGM 90 days
Tweden et al (2019)53

Prospective Single-arm Blinded

Postmarketing registry

Europe and South Africa 2016- 2019

Adult patients with T1D or T2D (% not known) for whom the Eversense CGM System was prescribed and inserted by their health care provider across approximately 1000 centers in Europe and South Africa

(N=945)

Implanted CGM

90 day system or 180 day system

4 insertionremoval cycles

CGM: continuous glucose monitoring; NR: not reported; T1D: type 1 diabetes; T2D: type 2 diabetes.

Table 8. Summary of Key Nonrandomized Trials: Implanted CGM Study Results 

Study Efficacy Outcomes Efficacy Results Adverse Events

Kropff (2017)42

PRECISE

N=71 N=71
MARD (glucose range 40- 400mg/dl 11.1% (glucose >75mg/dl) 14 device/ procedure-related nonsevere adverse events 11 participants (total number of 147 sensors implanted, used, and removed)

Christianen (2018)41

PRECISE II

N=90 N=90
MARD (glucose range 40- 400mg/dl)

8.8%

95% CI: 8.1%-9.3%

14 device/ procedure-related nonsevere adverse events in 7 participants (total number of 106 sensors implanted, used, and removed)

Christiansen (2019)44

PRECISION

N=35 N=35
MARD (glucose range 40- 400mg/dl)

9.6%

95% CI: 8.9%-10.4%

8 device/ procedure-related nonsevere adverse events in 5 participants (total number of 62 sensors implanted, used, and removed)
Deiss et al (2019)48   N=3023
  NR (safety only)

133 adverse events (85 procedure-related, 22 device-related, 6 drug-related, 4 device/procedure related; 16 not related)

No related serious adverse events through 4 insertion/removal cycles.

infection (n=29 patients); adhesive patch irritation (n=20 patients); unsuccessful first removal attempt (n=23 patients)

Sanchez et al (2019)49 N=205 N=205
MARD (glucose range 40- 400mg/dl) 11.2% (SD 11.3%, median 8.2%). 10 (5%) transient skin irritation, redness, and/or swelling. 4 (2%) mild infection, 3 (1.5%) hypoglycemia that was self-treated, 4 (2%) failure to remove the sensor on the first attempt, and 5 (2.5%) skin irritation due to the adhesive

 

Mean SG (mg/dL)

161.8

Median 157.2 (IQR 138.4 to 178.9)

% SG values in hypoglycemia (<54mg/dL), 24-hour period 1.2% (18.0 minutes)
% SG values in hypoglycemia (<54mg/dL), nighttime 1.7%
TIR, 24-hour period 62.3% (~15 hours)
TIR, nighttime 61.8%
Tweden et al (2019)53  

No evidence of degradation of performance from the repeated insertion and removal procedures occurring in approximately the same subcutaneous tissue of the body.

Adverse events otherwise not reported.

MARD (glucose range 40- 400mg/dl) Mean 11.5% to 11.9% during each sensor cycle
Mean SG (mg/dL) 156.5 to 158.2 mg/dL across four sensor cycles
% SG values in significant hypoglycemia (<54mg/dL), 24- hour period 1.1% to 1.3% (16 to 19 minutes)
% SG values in significant hypoglycemia (<70mg/dL), 24- hour period 4.6% to 5.0% (66 to 72 minutes)
TIR, 24-hour period 63.2% to 64.5% (910 to 929 minutes)
Time in hyperglycemia (>180- 250mg.dL), 24-hour perio 22.8% to 23.2% (328 to 334 minutes)
Time in signficant hyperglycemia (>250 mg/dL), 24-hour period 8.1% to 8.8% (117 to 127 minutes)

CGM: continuous glucose monitoring; CI: confidence interval; MARD: mean absolute relative difference; SD: standard deviation; SG: sensor glucose; TIR: time in range 

Section Summary: CGM Devices for Long-Term Use in Type 1 Diabetes 
Numerous RCTs and several systematic reviews of RCTs have evaluated CGM in patients with type 1 diabetes. A 2017 individual patient data analysis, using data from 11 RCTs, found that reductions in HbA1c levels were significantly greater with real-time CGM compared with a control intervention. In addition, a 2012 meta-analysis of 6 RCTs found a significantly larger decline in HbA1c levels at 6 months in the CGM group than the SMBG group. There are few studies beyond 6 months. Two recent RCTs in patients who used multiple daily insulin injections and were highly compliant with CGM devices during run-in phases found that CGM was associated with a larger reduction in HbA1c levels than previous studies. Reductions were 0.4% and 0.6%, respectively, compared with approximately 0.2% to 0.3% in previous analyses. One of the 2 RCTs prespecified hypoglycemia-related outcomes and time spent in hypoglycemia was significantly lower in the CGM group. 

One RCT in pregnant women with type 1 diabetes (n=215) has compared CGM with SMBG. Adherence was high in the CGM group. The difference in the change in HbA1c levels from baseline to 34 weeks of gestation was statistically significant favoring CGM, and women in the CGM group spent an increased percentage of time in the recommended glucose control target range at 34 weeks of gestation. There were no between-group differences in maternal hypoglycemia, gestational weight gain, or total daily insulin dose. A smaller proportion of infants of mothers in the CGM group were large for gestational age, had neonatal intensive care admissions lasting more than 24 hours, and had neonatal hypoglycemia requiring treatment. The total hospital length of stay was shorter by almost 1 day in the CGM group. 

Three nonrandomized prospective studies and three postmarketing registry studies assessed the accuracy and safety of an implanted glucose monitoring system that provides continuous glucose monitoring for up to 4 insertion/removal cycles as an adjunct to home glucose monitoring devices. Accuracy measures included the mean absolute relative difference between paired samples from the implanted device and a reference standard blood glucose measurement. The accuracy tended to be lower in hypoglycemic ranges. Limitations on the evidence include lack of differentiation in outcomes type 1 diabetes vs type 2 diabetes and variability in reporting of trends in secondary glycemic measures. The initial approval of the device has been expanded to allow the device to be used for glucose management decision making. The same clinical study information was used to support what the Food and Drug Administration considered a reasonable assurance of safety and effectiveness of the device for the replacement of fingerstick blood glucose monitoring for diabetes treatment decisions. As a condition of approval, the sponsor is required to conduct an additional post-approval-study.

CGM Devices for Short-Term Use in Type 1 Diabetes
Meta-analyses of glucose monitoring devices for type 1 diabetes tend to combine studies of intermittent glucose monitoring with studies of long-term CGM. For this body of evidence, there is variability in the definitions of intermittent monitoring and the specific monitoring protocols used. Also, many of the trials of intermittent monitoring have included additional interventions to optimize glucose control (eg, education, lifestyle modifications). 

Clinical Context and Therapy Purpose
The purpose of the short-term use of CGM devices is to provide a testing option that is an alternative to or an improvement on existing testing used in the management of individuals with type 1 diabetes.

The question addressed in this evidence review is: Does the short-term use of a CGM device improve the net health outcome for individuals with type 1 diabetes?

The following PICOs were used to select literature to inform this review. 

Patients
The relevant population of interest are individuals with type 1 diabetes. All individuals with type 1 diabetes require engagement in a comprehensive self-management and clinical assessment program that includes assessment of blood glucose control. Individuals with type 1 diabetes may have poorly controlled diabetes, despite current use of best practices, including situations such as unexplained hypoglycemic episodes, hypoglycemic unawareness, suspected postprandial hyperglycemia, and recurrent diabetic ketoacidosis. In addition, individuals with type 1 diabetes may need to determine basal insulin levels prior to insulin pump initiation.

Interventions
The testing being considered is the short-term use of a CGM device to assess blood glucose levels as part of optimal diabetes management. Short-term use is generally for 72 hours. However, reports of use range from 3-30 days.

Comparators
The following practice is currently being used to measure glucose levels: capillary blood sampling (finger stick) for SMBG. Standard treatment for patients with type 1 diabetes includes injection of long-acting basal insulin plus MDI of rapid-acting insulin boluses as required for meal intake. Activity level may require patients need to modify the timing and dose of insulin administration. Individuals with type 1 diabetes may also use an insulin pump either for initial treatment or convert to pump use after a period of MDI. Individuals with type 1 diabetes may also use an insulin pump either for initial treatment or convert to pump use after a period of MDI. Individuals are required to check their blood glucose before making preprandial insulin calculations, in response to symptoms of hypoglycemia or related to activity-related insulin adjustments

Outcomes
For short-term use of CGM, the general outcomes of interest includetime in range (generally glucoses of 70-180 mg/dl), frequency and time spent in hypoglycemia and, frequency and time spent in hyperglycemia for the duration of the monitoring. Repeat CGM may be necessary to assess the impact of changes in management.

CGM devices and self-glucose monitor devices may be used in the home, outpatient, or inpatient setting and patients are monitored by endocrinologists, diabetologists, internists and primary care physicians and clinicians.

Study Selection
Methodologically credible studies were selected using the following principles:

  1. To assess efficacy outcomes, comparative controlled prospective trials were sought, with a preference for RCTs;
  2. In the absence of such trials, comparative observational studies were sought, with a preference for prospective studies.
  3. To assess long-term outcomes and adverse effects, single-arm studies that capture longer periods of follow-up and/or larger populations were sought.
  4. Studies with duplicative or overlapping populations were excluded.

Systematic Reviews
Two meta-analyses were identified that reported separate subgroup analyses for intermittent monitoring. In a Cochrane review by Langendam et al (2012), 4 studies (total n=216 patients) compared real-time intermittent glucose monitoring systems with SMBG, and the pooled effect estimate for change in HbA1clevels at 3 months was not statistically significant (MD change, -0.18; 95% CI, -0.42 to 0.05).[9]The meta-analysis by Wojciechowski et al (2011), which assessed RCTs on CGM (described previously), also included a separate analysis of 8 RCTs of intermittent monitoring.[11] On pooled analysis, there was a statistically significant reduction in HbA1c levels with intermittent glucose monitoring compared with SMBG (WMD= -0.26; 95% CI, -0.45 to -0.06).

Randomized Controlled Trials
The largest RCT was the Management of Insulin-Treated Diabetes Mellitus (MITRE) trial, published by Newman et al (2009); it evaluated whether the use of the additional information provided by minimally invasive glucose monitors improved glucose control in patients with poorly controlled insulin-requiring diabetes.[21] This 4-arm RCT was conducted at secondary care diabetes clinics in four hospitals in England. This trial enrolled 404 people over the age of 18 years, with insulin-treated diabetes (types 1 or 2) for at least 6 months, who were receiving 2 or more injections of insulin daily. Most (57%) participants had type 1 diabetes (41% had type 2 diabetes, 2% were classified as “other”). Participants had to have 2 HbA1c values of at least 7.5% in the 15 months before trial entry and were randomized to 1 of 4 groups. Two groups received minimally invasive glucose monitoring devices (GlucoWatch Biographer or MiniMed Continuous Glucose Monitoring System [CGMS]). Intermittent glucose monitoring was used (ie, monitoring was performed over several days at various points in the trial). These groups were compared with an attention control group (standard treatment with nurse feedback sessions at the same frequency as those in the device groups) and a standard control group (reflecting common practice in the clinical management of diabetes). Changes in HbA1c levels from baseline to 3, 6, 12, and 18 months were the primary indicator of short- to long-term efficacy. At 18 months, all groups demonstrated a decline in HbA1c levels from baseline. Mean percentage changes in HbA1c levels were -1.4% for the GlucoWatch group, -4.2% for the CGMS group, -5.1% for the attention control group, and -4.9% for the standard care control group. In the intentionto-treat analysis, no significant differences were found between any groups at any assessment times. There was no evidence that the additional information provided by the devices changed the number or nature of treatment recommendations offered by the nurses. Use and acceptability indicated a decline for both devices, which was most marked in the GlucoWatch group by 18 months (20% still using GlucoWatch vs 57% still using the CGMS). In this trial of unselected patients, glucose monitoring (CGMS on an intermittent basis) did not lead to improved clinical outcomes.

Pregnant Women
Systematic Reviews
Voormolen et al (2013) published a systematic review of the literature on CGM during pregnancy.[22] They identified 11 relevant studies (total n=534 women). Two were RCTs, one of which was the largest of the studies (n=154). Seven studies used CGMs that did not have data available in real-time; the remaining four studies used real-time CGM. Reviewers did not pool study findings; they concluded that the evidence was limited to the efficacy of CGM during pregnancy. The published RCTs are described next.

Randomized Controlled Trials
Two RCTs of intermittent glucose monitoring in pregnant women with type 1 or type 2 diabetes are summarized in Tables 9 to 12 and the following paragraphs. While both trials included a mix of women with type 1 and type 2 diabetes, most women had type 1 diabetes in both trials, so the trials are reviewed in this section.

Secher et al (2013) randomized 154 women with type 1 (n=123) and type 2 (n=31) diabetes to real-time CGM in addition to routine pregnancy care (n=79) or routine pregnancy care alone (n=75).[24] Patients in the CGM group were instructed to use the CGM device for 6 days before each of 5 study visits and were encouraged to use the devices continuously; 64% of participants used the devices per-protocol. Participants in both groups were instructed to perform eight daily self-monitored plasma glucose measurements for six days before each visit. Baseline mean HbA1clevels were 6.6% in the CGM group and 6.8% in the routine care group. The 154 pregnancies resulted in 149 live births and 5 miscarriages. The prevalence of large-for-gestational-age infants (at least 90th percentile), the primary study outcome, was 45% in the CGM group and 34% in the routine care group. The difference between groups was not statistically significant (p=0.19). Also, no statistically significant differences were found between groups for secondary outcomes, including the prevalence of preterm delivery and the prevalence of severe neonatal hypoglycemia. Women in this trial had low baseline HbA1c levels, which might explain the lack of impact of CGM on outcomes. Other factors potentially contributing to the negative findings included the intensive SMBG routine in both groups and the relatively low compliance rate in the CGM group.

Murphy et al (2008) in the U.K. randomized 71 pregnant women with type 1 (n=46) and type 2 (n=25) diabetes to CGM or usual care.[23]The intervention consisted of up to 7 days of CGM at intervals of 4 to 6 weeks between 8 weeks and 32 weeks of gestation. Neither participants nor physicians had access to the measurements during sensor use; data were reviewed at study visits. In addition to CGM, the women were advised to measure blood glucose levels at least seven times a day. Baseline HbA1c levels were 7.2% in the CGM group and 7.4% in the usual care group. The primary study outcome was maternal glycemic control during the second and third trimesters. Eighty percent of women in the CGM group wore the monitor at least once per trimester. Mean HbA1c levels were consistently lower in the intervention arm, but differences between groups were statistically significant only at week 36. For example, between 28 weeks and 32 weeks of gestation, mean HbA1c levels were 6.1% in the CGM group and 6.4% in the usual care group (p=0.10). The prevalence of large-for-gestational-age infants (at least 90th percentile) was a secondary outcome. Thirteen (35%) of 37 infants in the CGM group were large-for-gestational age compared with 18 (60%) of 30 in the usual care group. The odds for reduced risk of a large-for-gestational-age infant with CGM was 0.36 (95% CI, 0.13 to 0.98; p=0.05).

Table 9. RCT Characteristics for Intermittent CGM in Pregnant Women With Type 1 Diabetes 

Study; Registration Countries Sites Dates Participants Interventions
          Active Comparator
Secher et al (2013)24; NCT00994357 Denmark 1 2009-2011 Pregnant women with type 1 (80%) or type 2 (20%) diabetes; mean gestational age, <14 wk); median HbA1c level, 6.7%; median age, 32 y CGM (for 6 d before each study visits; encouraged to used continuously) plus SOC (n=79) SOC (n=75)
Murphy et al (2008)23; ISRCTN84461581 U.K 2 2003-2006 Pregnant women with type 1 (65%) or type 2 (35%) diabetes; mean gestational age, 9.2 wk; mean HbA1c level, 7.3%; mean age, 31 y CGM (up to 7 d of CGM at intervals of 4-6 wk) plus SOC (n=38) SOC (n=33)

CGM: continuous glucose monitoring; HbA1c: hemoglobin A1c; RCT: randomized controlled trial; SOC: standard of care. Table 109. RCT Results for Intermittent CGM in Pregnant Women With Type 1 Diabetes 

Table 10. RCT Results for Intermittent CGM in Pregnant Women With Type 1 Diabetes 

Study

Infant 

 

Maternal  

 

Large-for- Gestational Age 

Gestational Age at Delivery 

Severe Hypoglycemia 

Caesarean Section 

 

HbA1c Levels at 36 Weeks of Gestationa 

Severe Hypoglycemia 

 

 

Days 

 

 

 

 

Secher et al (2013)24 

 

 

 

 

 

n

154

154

145

154

 

154

CGM

34 (45%)

Median, 263

9 (13%)

28 (37%)

Median, 6.0%

16%

Control

25 (34%)

Median, 264

10 (14%)

33 (45%)

Median, 6.1%

16%

TE (95% CI)

NR

NR

NR

NR

NR

NR

p

0.19

0.14

0.88

0.30

0.63

0.91

 

 

Weeks

 

 

 

 

Murphy et al (2008)23,

 

 

 

 

 

n

71

71

68

69

71

NR

CGM

13 (35%)

Mean, 37.6

3 (8%)

27 (71%)

Mean, 5.8%

 

Control

18 (60%)

Mean, 37.5

5 (17%)

21 (61%)

Mean, 6.4%

 

TE (95% CI)

OR=0.36 (0.13 to 0.98)

NR

NR

NR

0.6% (CI NR)

 

p

0.05

0.80

0.50

0.40

0.007

 

Values are n or n (%) or as otherwise indicated.
CGM: continuous glucose monitoring; CI: confidence interval; HbA1c: hemoglobin A1c; NR: not reported; OR: odds ratio; RCT: randomized controlled trial; TE: treatment effect. 
a N inconsistently reported for HbA1c outcome.

In summary, 2 trials of intermittent glucose monitoring conducted in Europe included pregnant women with type 1 or 2 diabetes, with most having type 1 diabetes. Secher et al (2013) included intermittent, real-time monitoring23,; Murphy et al (2008) included intermittent, retrospective monitoring with CGM.24, The intervention started in early pregnancy in these studies; mean age was in the early thirties and mean baseline HbA1c level was greater than 6.5%. There was no statistically significant difference between CGM and routine care for maternal HbA1c levels at 36 weeks in Secher; the difference in HbA1c levels at 36 weeks was about 0.6% (p=0.007) in Murphy. Secher also reported no difference in severe maternal hypoglycemia. The proportion of infants that were large for gestational age (>90th percentile) was higher in the CGM group in Secher, although not statistically significantly higher; the difference in large for gestational age was statistically significantly lower for CGM in Murphy. The differences in the proportions of infants born via caesarean section, gestational age at delivery, and infants with severe hypoglycemia were not statistically significant in either trial.

Tables 11 and 12 display notable gaps identified in each study.

Table 11. Relevance Limitations of RCTs of Intermittent CGM in Pregnant Women With Type 1 Diabetes 

Study

Populationa

Interventionb

Comparatorc

Outcomesd

Follow-Upe

Secher et al (2013)24 

4. Study population had relatively low HbA1c levels

4. Only 64% of the participants used devices per protocol 

 

 

 

Murphy et al (2008)23,

 

 

 

 

 

The study limitations stated in this table are those notable in the current review; this is not a comprehensive limitations assessment. CGM: continuous glucose monitoring; HbA1c: hemoglobin A1c; RCT: randomized controlled trial. 
a Population key: 1. Intended use population unclear; 2. Clinical context is unclear; 3. Study population is unclear; 4. Study population not representative of intended use. 
b Intervention key: 1. Not clearly defined; 2. Version used unclear; 3. Delivery not similar intensity as comparator; 4.Not the intervention of interest. 
c Comparator key: 1. Not clearly defined; 2. Not standard or optimal; 3. Delivery not similar intensity as intervention; 4. Not delivered effectively. 
d Outcomes key: 1. Key health outcomes not addressed; 2. Physiologic measures, not validated surrogates; 3. No CONSORT reporting of harms; 4. Not establish and validated measurements; 5. Clinical significant difference not prespecified; 6. Clinical significant difference not supported. 
e Follow-Up key: 1. Not sufficient duration for benefit; 2. Not sufficient duration for harms.

Table 12. Study Design and Conduct Limitations of RCTs of Short-Term CGM in Pregnant Women With Type 1 Diabetes

Study

Allocationa

Blindingb

Selective Reportingd

Data Completenesse

Powerd

Statisticalf

Secher et al (2013)24,

 

1. Not blinded; chance of bias in clinical management

 

 

 

3, 4. Treatment effects and confidence intervals not calculated

Murphy et al (2008)23, 

 

1. Not blinded; chance of bias in clinical management 

 

 

 

3, 4. Treatment effects and confidence intervals not calculated for some outcomes 

The study limitations stated in this table are those notable in the current review; this is not a comprehensive limitations assessment. CGM: continuous glucose monitoring; RCT: randomized controlled trial. 
a Allocation key: 1. Participants not randomly allocated; 2. Allocation not concealed; 3. Allocation concealment unclear; 4. Inadequate control for selection bias. 
b Blinding key: 1. Not blinded to treatment assignment; 2. Not blinded outcome assessment; 3. Outcome assessed by treating physician.
c Selective Reporting key: 1. Not registered; 2. Evidence of selective reporting; 3. Evidence of selective publication.
d Data Completeness key: 1. High loss to follow-up or missing data; 2. Inadequate handling of missing data; 3. High number of crossovers; 4. Inadequate handling of crossovers; 5. Inappropriate exclusions; 6. Not intent to treat analysis (per protocol for noninferiority trials). 
e Power key: 1. Power calculations not reported; 2. Power not calculated for primary outcome; 3. Power not based on clinically important difference. 
f Statistical key: 1. Analysis is not appropriate for outcome type: (a) continuous; (b) binary; (c) time to event; 2. Analysis is not appropriate for multiple observations per patient; 3. Confidence intervals and/or p values not reported; 4. Comparative treatment effects not calculated. 

Section Summary: Glucose Monitoring Devices for Short-Term Use in Type 1 Diabetes 
For short-term monitoring of type 1 diabetes, there are few RCTs and systematic reviews. The evidence for short-term monitoring on glycemic control is mixed, and there was no consistency in HbA1c levels. Some trials have reported improvements in glucose control for the intermittent monitoring group but limitations in this body of evidence preclude conclusions. The definitions of control with short-term CGM use, duration of use and the specific monitoring protocols varied. In some studies, short-term monitoring was part of a larger strategy aimed at optimizing glucose control, and the impact of monitoring cannot be separated from the impact of other interventions. Studies have not shown an advantage for intermittent glucose monitoring in reducing severe hypoglycemia events but the number of events reported is generally small and effect estimates imprecise. The limited duration of use may preclude an assessment of any therapeutic effect. Two RCTs of short-term CGM use for monitoring in pregnancy included women with both type 1 and 2 diabetes, with most having type 1 diabetes. One trial reported a difference in HbA1c levels at 36 weeks; the proportion of infants that were large for gestational age (>90th percentile) favored CGM while the second trial did not. The differences in the proportions of infants born via cesarean section, gestational age at delivery, and infants with severe hypoglycemia were not statistically significant in either study. Limitations of the published evidence preclude determining the effects of the technology on net health outcome. Evidence reported through clinical input supports that this use provides a clinically meaningful improvement in net health outcome and is consistent with generally accepted medical practice when used in specific situations such as poor control of diabetes despite the use of best practices and to help determine basal insulin levels prior to insulin pump initiation. Further details from clinical input are included in the Clinical Input section and the Appendix. 

CGM Devices for Use in Type 2 Diabetes
There is limited ability to distinguish between indications 3 and 4 (long-term) and 5 (short-term) glucose monitoring in the analysis of the data for type 2 diabetes, consistent with the literature.

Clinical Context and Therapy Purpose
The purpose of long-term CGM and short-term glucose monitoring devices is to provide a treatment option that is an alternative to or an improvement on existing therapies in individuals with type 2 diabetes.

The question addressed in this evidence review is: Does the use of long-term or short-term CGM glucose monitoring devices improve the net health outcome for individuals with type 2 diabetes?

The following PICOs were used to select literature to inform this review.

Patients
The relevant population of interest are individuals with type 2 diabetes. All individuals with type 2 diabetes require engagement in a comprehensive self-management and clinical assessment program that includes assessment of blood glucose control. Some individuals with type 2 diabetes may have poorly controlled diabetes, despite current use of best practices, including situations such as unexplained hypoglycemic episodes, hypoglycemic unawareness, and persistent hyperglycemia and A1C levels above target. In addition, some individuals with type 2 diabetes may need to determine basal insulin levels prior to insulin pump initiation.

Interventions
The testing being considered is the use of long-term or short-term CGM devices to assess blood glucose levels as part of optimal diabetes management.

Comparators
The following practice is currently being used to measure glucose levels: SMBG (capillary blood sampling (finger stick) using blood glucose meters) and periodic measurement of HbA1c.

Outcomes
The general outcomes of interest are a change in HbA1clevels, frequency of and time spent in hypoglycemia, frequency and time spent in hyperglycemia, complications of hypoglycemia and hyperglycemia, and QOL. To assess short-term outcomes such as HbA1clevels, a minimum follow-up of 8 to 12 weeks is appropriate. To assess long-term outcomes such as time spent in hypoglycemia, the incidence of hypoglycemic events, complications of hypoglycemia, and QOL, follow-up of six months to one year would be appropriate. CGM devices and self-glucose monitor devices may be used in the home, outpatient, or inpatient setting and patients are monitored by endocrinologists, diabetologists, internists and primary care physicians and clinicians.

Study Selection
Methodologically credible studies were selected using the following principles:

  1. To assess efficacy outcomes, comparative controlled prospective trials were sought, with a preference for RCTs
  2. In the absence of such trials, comparative observational studies were sought, with a preference for prospective studies.
  3. To assess long-term outcomes and adverse effects, single-arm studies that capture longer periods of follow-up and/or larger populations were sought.
  4. Studies with duplicative or overlapping populations were excluded.

Systematic Reviews
The systematic reviews by Poolsup et al (2013)[10] and Gandhi et al (2011),[8] previously described, also reported on the efficacy of CGM in patients with type 2 diabetes. A comparison of the trials of type 2 diabetes included in the systematic reviews and meta-analyses in these reviews is shown in Table 13.

Table 13. Comparison of CGM Trials for Type 2 Diabetes Included in Systematic Reviews 

Primary Study 

Ida et al (2019)38 

Poolsup et al (2013)10, 

Gandhi et al (2011)8, 

Ehrhardt et al (2011)28ª

 

Cosson et al (2009)27b,

 

Allen et al (2008)25b

 

Yoo et al (2008)43b,

Beck et al (2017)17a 

 

 

 

Ajjan et al (2016) 43b 

 

 

 

Haak et al ((2017)46b 

 

 

 

CGM: continuous glucose monitoring.
a These studies used real-time CGM (RT-CGM) devices compared to SMBG
b These studies used retrospective CGM (r-CGM) devices compared to SMBG

A summary of the characteristics of systematic reviews is shown in Table 14. Results are briefly described in Table 13 and the following. Gandhi et al (2011) identified 3 RCTs studying patients with type 2 diabetes (1 study included both types of diabetes).[8] There was a mix of patients with type 2 diabetes who did and did not require insulin. Two of the three trials evaluated retrospective CGM of different lengths and durations, and the third evaluated real-time intermittent glucose monitoring. Patients in the trials had baseline HbA1clevels greater than 8%. In a meta-analysis of the 3 trials, there was a statistically significant reduction in HbA1clevels for CGM compared with SMBG in adults with type 2 diabetes (WMD = -0.70; 95% CI, -1.14 to -0.27). Poolsup et al (2013) conducted a meta-analysis of 4 trials evaluating adults with type 2 diabetes.[10] Three trials in Poolsup et al (2013) overlapped with those of Gandhi et al (2011); the remaining trial also evaluated real-time CGM but with a longer period of use (2 weeks on and 1 week off for 3 months). In a pooled analysis, CGM had greater efficacy regarding HbA1clevels than SMBG. The pooled mean difference in HbA1clevel was -0.31% (95% CI, -0.6% to 0.02%; p=0.04). Because of a lack of statistical heterogeneity among studies, subgroup analyses (eg, by type of CGM device) were not performed. Table 14. Systematic Review Characteristics for CGM in Type 2 Diabetes 

Table 12. Systematic Review Characteristics for CGM in T2D 

Study

Dates

Trials

Participants

N (Range)

Design

Duration

Ida et al (2019)38

1960-2018

7

Adults with T2D

669 (25-224)

RCT

At least 8 wk

Poolsup et al (2013)10 

To 2013 

Adults with T2D 

228 (25-100) 

RCT 

At least 8 wk (median, 3 mo) 

Gandhi et al (2011)8,

1996-2010

3

 

Adults with T2D

128 (25-57)

RCT

At least 8 wk (median, 3 mo)

CGM: continuous glucose monitoring; HbA1c: hemoglobin A1c; RCT: randomized controlled trial; T2D: type 2 diabetes.

Table 15. Meta-Analytic Results for CGM in Type 2 Diabetes 

Study 

Reduction in HbA1c Levels (Mean Difference) 

Hypoglycemic Events (Mean Difference 

Diabetes Complications (retinopathy, nephropathy, neuropathy, diabetic foot) 

Health-Related Quality of Life 

Ida et al (2019)38 

 

 

 

 

Total N

660

285

 

 

PE (95% CI)

-0.42 (-0.70 to -0.13)

-0.35 (-0.59 to -0.10)ª 

NR 

 Multiple diabetes-specific scales used in each study, therefore, results could not be combined for metaanalyse

p

0.004

0.0006 

 

 

I2

64%

0% 

 

 

Poolsup et al  (2013)10,

 

 

 

 

Total N

228

NR

NR

NR

PE (95% CI)

-0.31 (-0.60 to -0.02)

 

 

 

p

0.04

 

 

 

I2

0%

 

 

 

Gandhi et al (2011)8,

 

 

 

 

Total N

128

NR

NR

NR

PE (95% CI)

-0.70 (-1.14 to -0.27)

 

 

 

p

NR

 

 

 

I2

0%

 

 

 

CGM: continuous glucose monitoring; CI: confidence interval; HbA1c: hemoglobin A1c; NR: not reported; PE: pooled effect. 

Randomized Controlled Trials 
Several RCTs of CGM in adults with type 2 diabetes are summarized in Tables 16 to 19. The largest and most recent studies are also briefly summarized in the following paragraphs. The trials were conducted in North America, Europe, and Asia. Baseline HbA1clevels were between 8.5% and 9.0% in the RCTs, with participants having a mean baseline age range in the mid-50s and early-60s. The RCTs used a mixed of intermittent and continuous, real-time monitoring.

A large RCT, Multiple Daily Injections and Continuous Glucose Monitoring in Diabetes (DIAMOND), was reported by Beck et al (2017).[31] DIAMOND was performed at 25 endocrinology practices in North America (22 in the U.S., 3 in Canada) and enrolled adults with type 2 diabetes receiving multiple daily injections of insulin. One-hundred fifty-eight patients were randomized in 2 groups: CGM and usual care (n=79 in each group). Patients compliant during a run-in period were eligible for randomization. Patients in both groups were given a blood glucose meter. Participants in the CGM group were given a Dexcom G4 Platinum CGM System (Dexcom) and instructions on use. Change in HbA1c level from baseline to 24 weeks was the primary outcome. Analyses were adjusted for baseline HbA1c levels and the clinic was performed using intention-to-treat analysis with missing data handling by multiple imputations. At baseline, the mean total daily insulin dose was 1.1 U/kg/d. Week 24 follow-up was completed by 97% of the CGM group and 95% of the control group. Mean CGM use was greater than six d/wk at one month, three months, and six months. The adjusted difference in mean change in HbA1c level from baseline to 24 weeks was -0.3% (95% CI, -0.5% to 0.0%; p=0.022) favoring CGM. The adjusted difference in the proportion of patients with a relative reduction in HbA1c level of 10% or more was 22% (95% CI, 0% to 42%; p=0.028) favoring CGM. There were no events of severe hypoglycemia or diabetic ketoacidosis in either group. The treatment groups did not differ in any of the QOL measures.

The RCT by Sato et al (2016) included 34 patients with type 2 diabetes who were at least 20 years old and on insulin injection therapy, had HbA1clevels between 6.9% and 11.0% during the previous 3 months, with HbA1c fluctuations within 0.5%.[30] All patients conducted SMBG and used CGM devices that do not have data available in real-time (ie, data were viewed retrospectively by physicians). Devices were used for four to five days before each of three clinic visits, two months apart. At clinic visits, patients were evaluated and suggestions made to improve glucose control by lifestyle changes and by changing medication doses. In the intervention group, but not the control group, patients and physicians had access to CGM data at the clinic visits. The primary endpoint was a change in HbA1c levels from baseline, which did not differ significantly between groups at the end of the trial, between the first and second visits, or between the second and third visits. HbA1c levels changed little in either group. In the intervention group, the mean baseline HbA1c level was 8.2%, and the mean final HbA1c level was also 8.2%. Comparable percentages in the control group were 8.2% and 7.9%. In this trial, conducted in Japan, decisions on medication doses were made only by the physician at clinic visits, and practices may differ in other countries.

Ehrhardt and colleagues published 2 reports (2011, 2012) from an RCT evaluating the largest sample (n=100) in the Poolsup et al (2013) systematic review (accounting for 45% of the weight in the pooled analysis of HbA1c levels).[29][28] The trial evaluated the intermittent use of a CGM device in adults with type 2 diabetes treated with diet/exercise and/or glycemia-lowering medications but not prandial insulin who had an initial HbA1clevel of at least 7% but not more than 12%. The trial compared real-time CGM with the Dexcom device used for four, two-week cycles (two weeks on and one week off) with SMBG. The primary efficacy outcome was a mean change in HbA1c levels. Mean HbA1c levels in the CGM group were 8.4% at baseline, 7.4% at 12 weeks, 7.3% at 24 weeks, and 7.7% at 52 weeks. In the SMBG group, these values were 8.2% at baseline, 7.7% at 12 weeks, 7.6% at 24 weeks, and 7.9% at 52 weeks. During the trial, the reduction in HbA1c levels was significantly greater in the CGM group than in the SMBG group (p=0.04). After adjusting for potential confounders (eg, age, sex, baseline therapy, whether the individual started taking insulin during the study), the difference between groups over time remained statistically significant (p<0.001). The investigators also evaluated SMBG results for both groups. The mean proportions of SMBG tests less than 70 mg/dL were 3.6% in the CGM group and 2.5% in the SMBG group (p=0.06).

Table 16. RCT Characteristics for Glucose Monitoring in Type 2 Diabetes 

Study; Registration 

Countries 

Sites 

Dates 

Participants 

Interventions

 

 

 

 

 

Active 

Comparator 

Haak et al (2017)45,46 

France, Germany, UK 

26 

2013-2014 

Adults (≥18y) with T2D on intensive insulin therapy (MDI or CSII),HbA1c levels (7.5-12.0%), SMBG >10/week 

Flash sensor-based glucose monitoring (n=149) 

SMBG (n=75) 

Beck et al (2017) (DIAMOND)31,; NCT02282397  

U.S., Canada 

25 

2014-2016 

Adults with T2D using multiple daily injections of insulin with HbA1c levels 7.5%-10.0% (baseline mean, 8.5%); mean age, 60 y 

Real-time CGM (n=79) 

SMBG (n=79) 

Sato et al (2016)30,; UMIN: 000012034a  

Japan 

2012-2014 

Adults with T2D using insulin with HbA1c levels 6.9%-11.0% (baseline mean, 8.2%); mean age, 62 y   

CGM for 4-5 d every 4 mo; reviewed at study visits (n=17)  

"Blinded" CGM (n=17)

 

Ehrhardt et al  (2011)28,

U.S.

 

NR

 

Adults with T2D using oral  antidiabetic agents without prandial insulin with HbA1c levels 7.0%-12.0% (baseline mean, 8.3%), mean age, 58 y

Real-time CGM for  4 cycles of 3 wk (n=50)

SMBG (n=50) 

Cosson et al (2009)27,

France

5

NR

Adults with T1D or T2D treated with oral antidiabetic agents with or without insulin with HbA1c levels 8.0%-10.5% (baseline mean, 9.1% in T2D); mean age, 57 y in T2D

CGM for 48 h at baseline and 3 mo; CGM data shared with physician and patient (n=11 with T2D)

"Blinded" CGM (n=14 in T2D)

Allen et al (2008)25,

U.S.

2

NR

Adults with T2D not receiving insulin with HbA1c levels >7.5% (baseline mean, 8.6%), not participating in physical activity; mean age, 57 y

Diabetes education plus CGM for 3 d (n=27)

Diabetes education (n=25)

Yoo et al (2008)26,

Korea

4

2007

Adults with T2D using oral antidiabetic agents or insulin with HbA1c levels 8.0%-10.0% (baseline mean, 9%); mean age, 56 y

CGM (3 d at a time for 3 mo) (n=32)

SMBG (n=33)

CGM: continuous glucose monitoring; CSII: continuous subcutaneous insulin infusion; HbA1c: hemoglobin A1c; NR: not reported; MDI: multiple daily injections; RCT: randomized controlled trial; SMBG: self-monitored blood glucose; T1D: type 1 diabetes; T2D: type 2 diabetes.

a Registered with the University Hospital Medical Information Network in Japan.

Most RCTs used a type of intermittent monitoring; some reported data for patients in real-time while others provided data reviewed only at study visits. Four of the six RCTs of CGM in type 2 diabetes reported a statistically significant larger decrease in HbA1c levels with CGM than with control. Beck et al (2017) reported more patients in CGM with a relative reduction in HbA1c levels of greater than 10% at 24 weeks but no difference in the QOL measures.[29] In Cosson et al (2009), the comparative treatment effect was not reported, but the CGM group had a statistically significant reduction in HbA1c levels from baseline to 3 months.[26] Few other outcomes were reported. No trials reported on follow-up beyond six months. Thus the effect of CGM on outcomes related to diabetic complications is unknown. Only two RCTs used blinded CGM; in one, there was no difference in reduction in HbA1c levels between CGM and control. Haak et al (2017)[32][46]reported the use of flash glucose-sensing technology as a replacement for SMBG for the management of insulindependent treated T2D and found no difference in HbA1c change at six months between groups. Intervention group participants did experience reductions in time in hypoglycemia compared with control for mild (<70mg/DL), moderate (<55mg/DL) and severe hypoglycemia (<45mg/dL) of 43%, 53% and, 64% respectively.

Table 17. RCT Outcomes for Glucose Monitoring in Type 2 Diabetes.

Study

Reduction in HbA1c Levels (Mean Range), %

HbA1c Level <7.0%, n (%)

Relative Reduction in HbA1c Level ≥10%, n (%)

Hypoglycemic or Ketoacidosis Events

Diabetes Complications(retinopathy, nephropathy, neuropathy, diabetic foot)

Health-Related Quality of Life

 

Baseline to 24 Wk

At 24 Wk

At 24 Wk

 

 

DTSQ Overall Mean Score at 24 Wk

kHaak et al (2017)32,45,46,46

 

 

 

 

 

 

N

224

NR

NR

 

NR

224

Flash monitor

8.6 to 3.7

 

 

3 serious hypoglycemic eventsa

 

(mean±SE)

13.1 (0.50)

Control SMBG

8.75 to 8.34

 

 

1 serious hypoglycemic eventsa

 

9.0(0.72)

TE (95% CI)

NR

 

 

 

 

NR

p

0.8222

   

 

 

<0.0001

 

Baseline to 24 WK

At 24 Wk 

At 24 Wk 

 

 

DDS Overall Mean Score at 24 Wk 

Beck et al (2017)31,

 

 

 

 

 

 

N

158

158

158

158

NR

150

CGM

8.6 to 7.7

11 (14%)

40 (52%)

0

 

Baseline: 1.78

24 weeks:  1.61

Control

8.6 to 8.2

9 (12%)

24 (32%)

0

 

Baseline: 1.69

24 weeks: 1.78

TE (95% CI)

-0.3 (-0.5 to 0.0)

3% (-9% to 14%)

22% (0% to 42%)

 

 

0.22 (0.08 to 0.36)

p

0.022

0.88

0.028

 

 

0.009

 

Baseline to 8 Mo

 

 

 

 

 

Sato et al (2016)30,    

 

 

 

 

 

 

N

34

NR

NR

NR

NR

NR

CGM

8.2 to 8.2

 

 

 

 

 

Control

8.2 to 7.9

 

 

 

 

 

TE (95% CI)             

NR

 

 

 

 

 

p

>0.05

 

 

 

 

 

 

Baseline to 12 Wk

 

 

 

 

 

Ehrhardt et al (2011)28,

 

 

 

 

 

 

N

100

NR

NR

NR

NR

NR

CGM

8.4 to 7.4

 

 

 

 

 

Control

8.2 to 7.7

 

 

 

 

 

TE (95% CI)

NR

 

 

 

 

 

p

0.006

 

 

 

 

 

 

Baseline to 3 Mo

 

 

Time Spent With Hypoglycemia, min

 

 

Cosson et al (2009)27,

 

 

 

 

 

 

N

25

NR

NR

19

NR

NR

CGM

9.2 to 8.6

 

 

18

 

 

Control

9.0 to 8.8

 

 

11

 

 

TE (95% CI)

NR

 

 

NR

 

 

 

Baseline to 8 Wk

 

 

 

 

 

Allen et al (2008)27,

 

 

 

 

 

 

N

46

NR

NR

NR

NR

NR

 

 

 

 

 

 

 

CGM

8.9 to 7.7

 

 

 

 

 

Control

8.4 to 8.1

 

 

 

 

 

TE (95% CI)

NR

 

 

 

 

 

p

<0.05

 

 

 

 

 

 

Baseline to 3 Mo

 

 

 

 

 

Yoo et al (2008)28,

 

 

 

 

 

 

N

57

NR

NR

NR

NR

NR

CGM

9.1 to 8.0

 

 

 

 

 

Control

8.7 to 8.3

 

 

 

 

 

TE (95% CI)

NR

 

 

 

 

 

p

0.004

 

 

 

 

 

CGM: continuous glucose monitoring; CI: confidence interval; DDS: Diabetes Distress Scale; HbA1c: hemoglobin A1c; NR: not reported; RCT: randomized controlled trial; TE: treatment effect. 

a serious hypoglycemic event defined as requiring third-party assistance.

Tables 18 and 19 display notable gaps identified in each study.

Table 18. Relevance Limitations of RCTs for Glucose Monitoring in Type 2 Diabetes 

Study; Trial 

Populationa 

Interventionb 

Comparatorc 

Outcomesd 

Follow-Upe 

Haak et al (2017)45,46 

 

 

 

1. Did not include outcomes on diabetic complications 

1. Follow-up not sufficient to determine effects on diabetic complications 

Beck et al (2017)31,; DIAMOND  

 

 

 

1. Did not include outcomes on diabetic complications 

1. Follow-up not sufficient to determine effects on diabetic complications 

Sato et al (2016)30, 

 

 

 

1. Focused on HbA1c; did not include outcomes on adverse events, QOL, or diabetic complications

 1. Follow-up not sufficient to determine effects on diabetic complications

Ehrhardt et al (2011)28, 

 

 

 

1. Focused on HbA1c; did not include outcomes on adverse events, QOL, or diabetic complications

6. No justification for clinically significant difference 

1. Follow-up not sufficient to determine effects on diabetic complications; patients reportedly followed for 52 wk but data not reported.

 

Cosson et al (2009)27,

 

 

 

1. Focused on HbA1c; did not include outcomes on adverse events, QOL, or diabetic complications

1. Follow-up not sufficient to determine effects on diabetic complications

Allen et al (2008)25,

 

 

 

1. Focused on HbA1c; did not include outcomes on adverse events, QOL, or diabetic complications

1. Follow-up not sufficient to determine effects on diabetic complications

Yoo et al (2008)26,

 

 

 

1. Focused on HbA1c; did not include outcomes on adverse events, QOL, or diabetic complications

1. Follow-up not sufficient to determine effects on diabetic complications

The study limitations stated in this table are those notable in the current review; this is not a comprehensive limitations assessment. HbA1c: hemoglobin A1c; QOL: quality of life; RCT: randomized controlled trial.

a Population key: 1. Intended use population unclear; 2. Clinical context is unclear; 3. Study population is unclear; 4. Study population not representative of intended use. 
b Intervention key: 1. Not clearly defined; 2. Version used unclear; 3. Delivery not similar intensity as comparator; 4.Not the intervention of interest. 
c Comparator key: 1. Not clearly defined; 2. Not standard or optimal; 3. Delivery not similar intensity as intervention; 4. Not delivered effectively. 
d Outcomes key: 1. Key health outcomes not addressed; 2. Physiologic measures, not validated surrogates; 3. No CONSORT reporting of harms; 4. Not establish and validated measurements; 5. Clinical significant difference not prespecified; 6. Clinical significant difference not supported. 
e Follow-Up key: 1. Not sufficient duration for benefit; 2. Not sufficient duration for harms.

Table 19. Study Design and Conduct Limitations of RCTs for Glucose Monitoring in Type 2 Diabetes

Study; Trial

Allocationa

Blindingb

Selective Reportingc

Data Completenessd

Powere

Statisticalf

Haak et al (2017)32,45,46 

 

1.Pre-randomization blinded run-in phase for both groups. Control group only blinded for last 2 weeks of study 

 

 

 

3, 4. Treatment effects and CIs not calculated 

Beck et al (2017)31; DIAMOND

 

1. Not blinded; chance of bias in clinical management

 

 

 

 

Sato et al (2016)30,

 

 

 

 

 

3, 4. Treatment effects and CIs not calculated

Ehrhardt et al (2011)28,

 

1. Not blinded; chance of bias in clinical management

1. Registration not reported

 

3. No justification for difference used for power calculation

3, 4. Treatment effects and CIs not calculated

Cosson et al (2009)27,

 

 

1. Registration not reported

2. Unclear how missing data were handled in analyses

1.-3. No power calculations

3, 4. Treatment effects and CIs not calculated

Allen et al (2008)25,

 

1. Not blinded; chance of bias in clinical management

1. Registration not reported

 

2, 3. Power not calculated a priori; convenience sample size

3, 4. Treatment effects and CIs not calculated

Yoo et al (2008)26,

 

1. Not blinded; chance of bias in clinical management

1. Registration not reported

 

 

3, 4. Treatment effects and CIs not calculated

The study limitations stated in this table are those notable in the current review; this is not a comprehensive limitations assessment. CI: confidence interval; RCT: randomized controlled trial. 
a Allocation key: 1. Participants not randomly allocated; 2. Allocation not concealed; 3. Allocation concealment unclear; 4. Inadequate control for selection bias. 
b Blinding key: 1. Not blinded to treatment assignment; 2. Not blinded outcome assessment; 3. Outcome assessed by treating physician.
c Selective Reporting key: 1. Not registered; 2. Evidence of selective reporting; 3. Evidence of selective publication.
d Data Completeness key: 1. High loss to follow-up or missing data; 2. Inadequate handling of missing data; 3. High number of crossovers; 4. Inadequate handling of crossovers; 5. Inappropriate exclusions; 6. Not intent to treat analysis (per protocol for noninferiority trials). 
e Power key: 1. Power calculations not reported; 2. Power not calculated for primary outcome; 3. Power not based on clinically important difference. 
f Statistical key: 1. Analysis is not appropriate for outcome type: (a) continuous; (b) binary; (c) time to event; 2. Analysis is not appropriate for multiple observations per patient; 3. Confidence intervals and/or p values not reported; 4. Comparative treatment effects not calculated. 

Haak et al (2017a)[45] reported the results of a 12-month open-access extension of the REPLACE RCT comparing flash glucose-sensing technology in individuals with type 2 diabetes treated with intensive insulin therapy. Summaries of the study characteristics and results are provided in Table 20 and 21 respectively.

Generally, the impact on outcomes of reduction in time in hypoglycemia and reduction in nocturnal hypoglycemia was maintained at 12 months. There was no change in time in range (70-180mg/dl). Sensor utilization was maintained at a rate of 83.6% (SD: 13.8%) of daily intended use. Adverse events were reported in five participants in the open-access phase, which lead to withdrawal from the study. Two participants died but death was not judged to be a device or study-related. Three participants experienced sensor site complications severe enough to warrant study discontinuation. Table 2019. Summary of Key Nonrandomized Trial Characteristics. 

Table 20. Key RCT Characteristics for CGM in Pregnant Women With Gestational Diabetes 

Study

Study Type 

Country

Dates

Participants 

Treatment

Follow-up 

Haak (2017a45

Prospective Open Access Extension France, Germany, UK 2013-2015

Adults (≥18 years) REPLACE trial intervention group participants who completed 6-month treatment phase

(N=139)

Flash glucose sensor use for selfmanagement of T2D using insulin

(Insulin pen device = 94%)

12 months

T2D: type 2 diabetes.

Table 21. Summary of Key Nonrandomized Trial Results

Study

Reduction in Time in Hypoglycemia Hours/day (12 months)

Frequency of Hypoglycemic Events/day (12 months) Reduction in Time in Nocturnal Hypoglycemia Hours per 7 hours (12 months

Frequency of Nocturnal Hypoglycemia Events per 7 hours (12 months 

Change From Baseline Hypoglycemic Events/day (12 months)

Haak (2017a)45,

N (108) 

N (108) 

N (108) 

N (108) 

N (108) 

  Mean (SD) Mean (SD) Mean (SD)

Mean (SD)

 (%)

Glucose <70 mg/dl

-0.70 (1.85)

p=0.0002

-.0.27 (0.67)

p<0.0001

-0.31 (0.84)

p=0.0002

-0.1 (0.33)

p=0.0021

-40.8

p<0.0001

Glucose <55 mg/dl

-0.40 (1.09)

p=0.0002

0.20 (0.49)

p<0.0001

-0.19 (0.57)

p=0.0008

-0.9 (0.21)

p<0.0001

-56.5

p<0.0001

Glucose <45 mg/dl

-0.23 (0.73)

p= 0.0013

-0.13 (0.35)

p<0.0002

-0.12 (0.42)

p=0.0032

-0.05 (0.15)

p=0.0008

-61.7

p=0.0001

SD: standard deviation.

Pregnant Women
As discussed in the section on CGM in pregnant women, two RCTs have evaluated short-term glucose monitoring in pregnant women with type 1 and type 2 diabetes. Most women had type 1 diabetes in both trials. There were 25 (35%) women with type 2 diabetes in Murphy et al (2008)[23] and 31 (20%) with type 2 diabetes in Secher et al (2013).[24] Results for women with type 2 diabetes were not reported in Murphy et al (2008). Secher et al (2013) reported that 5 (17%) women with type 2 diabetes experienced 15 severe hypoglycemic events, with no difference between groups; other analyses were not stratified by diabetes type.

Section Summary: CGM for Use in Type 2 Diabetes
Most RCTs of CGM in patients with type 2 trials found statistically significant benefits of CGM regarding glycemic control. However, the degree of HbA1c reduction and the difference in HbA1c reduction between groups might not be clinically significant. Moreover, additional evidence would be needed to show what levels of improvements in HbA1c levels over the short-term would be linked to meaningful improvements over the long-term in health outcomes such as diabetes-related morbidity and complications. Also, the variability in entry criteria as well as among interventions makes it difficult to identify an optimal approach to CGM use; the studies used a combination of intermittent and continuous monitoring with a review of data in real-time or at study visits only. Only the DIAMOND trial (n=158) used real-time CGM in type 2 diabetes. Selected patients were highly compliant during a run-in phase. The difference in change in HbA1clevels from baseline to 24 weeks was -0.3% favoring CGM. The difference in the proportion of patients with a relative reduction in HbA1clevel by 10% or more was 22% favoring CGM. There were no differences in the proportions of patients with an HbA1c level of less than 7% at week 24. There were no events of severe hypoglycemia or diabetic ketoacidosis in either group. The treatment groups did not differ in any of the QOL measures. RCTs using flash glucose-sensing technology as a replacement for SMBG for the management of insulin-dependent treated type 2 diabetes found no difference in HbA1c change at 6 and 12 months between groups. However, time in severe hypoglycemia (<45mg/dL) was reduced for intervention participants. Two trials of CGM have enrolled pregnant women with type 2 diabetesbut the total number of women with type 2 diabetes included in both trials is only 58. One study reported a difference in HbA1c levels at 36 weeks, and the proportion of infants that were large for gestational age (>90th percentile) favored CGM while the second study did not. Neither trial reported analyses stratified by diabetes type. Limitations of the published evidence preclude determining the effects of the technology on net health outcome. Evidence reported through clinical input for long-term (continuous) CGM in patients with type 2 diabetes who do not require insulin did not provide strong support of a safety benefit and clinically meaningful improvement in net health outcome. Evidence reported through clinical input for use of short-term CGM in patients with type 2 diabetes who require multiple daily doses of insulin supports that this use provides a clinically meaningful improvement in net health outcome and is consistent with generally accepted medical practice when used in specific situations such as poor control of diabetes despite use of best practices and to help determine basal insulin levels prior to insulin pump initiation. Further details from clinical input are included in the Clinical Input section and the Appendix.

Use of Long-Term (Continuous) CGM in Individuals with Type 2 Diabetes on Multiple Daily Doses of Insulin with Significant Hypoglycemia in the Setting of Insulin Deficiency

Clinical Context and Therapy Purpose
The purpose of long-term CGM glucose monitoring devices is to provide a treatment option that is an alternative to or an improvement on existing therapies in individuals with type 2 diabetes(T2DM).

The question addressed in this evidence review is: Does the use of long-term CGM glucose monitoring devices improve the net health outcome for individuals with type 2 diabetes who are on multiple daily doses of insulin with significant hypoglycemia in the setting of insulin deficiency?

The following PICOs were used to select literature to inform this review.

Patients
The relevant population of interest is a subgroup of individuals with type 2 diabetes who are willing and able to use the device and have adequate medical supervision and who experience significant hypoglycemia on multiple daily doses of insulin or an insulin pump in the setting of insulin deficiency who receive long-term (continuous) glucose monitoring.

Interventions
The testing being considered is the use of long-term CGM devices to assess blood glucose levels and detect hypoglycemia as part of optimal diabetes management.

Comparators
The following practice is currently being used to measure glucose levels: SMBG (capillary blood sampling (finger stick) using blood glucose meters) and periodic measurement of HbA1c.

Outcomes
The general outcomes of interest are the frequency of and time spent in hypoglycemia, the incidence of hypoglycemic episodes, complications of hypoglycemia, and QOL. To assess short-term outcomes a minimum follow-up of 8 to 12 weeks is appropriate. To assess long-term outcomes follow-up of six months to one year would be appropriate. CGM devices and self-glucose monitor devices may be used in the home, outpatient, or inpatient setting and patients are monitored by endocrinologists, diabetologists, internists and primary care physicians and clinicians.

Study Selection
Methodologically credible studies were selected as described above in the section CGM Devices for Use in Type 2 Diabetes

Systematic Reviews
Meta-analytic results for long-term CGM in type 2 diabetes are summarized in Table 15. The largest and most recently published systematic review of RCTs (Ida et al [2019][38]) reported a statistically significant reduction in hypoglycemic events in 285 subjects for CGM with a mean reduction of -0.35 (mean difference -0.59 to -0.10, p=0.0006).

Key Non-Randomized Trials
Twelve-month open-access, follow-up results for long-term CGM in 108 individuals with type 2 diabetes treated with intensive insulin therapy are summarized in Table 20 (Haak [(2017)][45]). Hypoglycemia was analyzed using 3 different glucose level thresholds (<70 mg/dl, <55 mg/dl, and <45 mg/dl). At all three glucose level thresholds, there were statistically significant reductions in time in hypoglycemia, frequency of hypoglycemic events, time in nocturnal hypoglycemia, and frequency of nocturnal hypoglycemia. Change for hypoglycemic events per day at 12 months compared to baseline was also significant: -40.8% (glucose <70 mg/dl, p<0.0001); -56.5% (glucose <55 mg/dl, p<0.0001); -61.7% (glucose <45 mg/dl, p=0.0001).

Section Summary: Use of Long-Term (Continuous) CGM in Individuals with Type 2 Diabetes on Multiple Daily Doses of Insulin with Significant Hypoglycemia in the Setting of Insulin Deficiency
A recently published systematic review and 12-month follow-up study using long-term CGM in patients with type 2 diabetes demonstrate that CGM can significantly reduce time in hypoglycemia and frequency of hypoglycemia events both during the day and at night. At 12-month follow-up, hypoglycemic events were reduced by 40.8% to 61.7% with a greater relative reduction in the most severe thresholds of hypoglycemia. The published evidence supports a meaningful improvement in the net health outcome. Evidence reported through clinical input provides additional clinical context and based on both the published evidence and clinical input the following patient selection criteria are associated with a clinically meaningful improvement in net health outcome and are consistent with generally accepted medical practice: selected patients with type 2 diabetes who are (1) willing and able to use the CGM device and have adequate medical supervision and (2) experiencing significant hypoglycemia on multiple daily doses of insulin or an insulin pump in the setting of insulin deficiency. Further details from clinical input are included in the Clinical Input section and the Appendix.

CGM Use in Pregnant Women With Gestational Diabetes
Clinical Context and Therapy Purpose
The purpose of long-term (continuous) CGM and short-term (intermittent) glucose monitoring devices is to provide a treatment option that is an alternative to or an improvement on existing therapies in women with gestational diabetes.

The question addressed in this evidence review is: Do the use of long-term (continuous) CGM and short-term (intermittent) glucose monitoring devices improve the net health outcome for women with gestational diabetes?

The following PICOs were used to select literature to inform this review.

Patients
The relevant population of interest are women with gestational diabetes.

Interventions
The therapies being considered are devices that provide continuous, long-term glucose levels to the patient to direct insulin regimens and intermittent (ie, 72 hours), the results of short-term monitoring of glucose levels are used by the provider to optimize management.

Comparators
The following practice is currently being used to measure glucose levels: capillary blood sampling (finger stick) for blood glucose meters for self-monitoring.

Outcomes
The general outcomes of interest are a change in HbA1clevels, time spent in hypoglycemia, the incidence of hypoglycemic events, complications of hypoglycemia and QOL.

To assess short-term outcomes such as HbA1clevels, time spent in hypoglycemia, the incidence of hypoglycemic events and, complications of hypoglycemia, a minimum follow-up of 8 to 12 weeks is appropriate. To assess long-term outcomes such as QOL and maternal and infant outcomes, follow-up of 24 to 36 weeks would be appropriate.

Study Selection
Methodologically credible studies were selected using the following principles:

  1. To assess efficacy outcomes, comparative controlled prospective trials were sought, with a preference for RCTs;
  2. In the absence of such trials, comparative observational studies were sought, with a preference for prospective studies.
  3. To assess long-term outcomes and adverse effects, single-arm studies that capture longer periods of follow-up and/or larger populations were sought.
  4. Studies with duplicative or overlapping populations were excluded.

Randomized Controlled Trials
One trial of glucose monitoring in women with gestational diabetes has been published. Trial characteristics, results, and limitationsare shown in Tables 22 to 25. In the RCT, Wei et al (2016) evaluated the use of CGM in 120 women with gestational diabetes at 24 to 28 weeks.[33] Patients were randomized to prenatal care plus CGM (n=58) or SMBG (n=62). The CGM sensors were reportedly inserted for 48 to 72 hours on weekdays; it is not clear whether the readings were available in real-time. The investigators assessed a number of endpoints and did not specify primary outcomes; a significance level of p less than 0.05 was used for all outcomes. The groups did not differ significantly in a change in most outcomes, including a change in maternal HbA1c levels, rates of preterm delivery before the 35th gestational week, cesarean delivery rates, proportions of large-for-gestational-age infants, or rates of neonatal hypoglycemia. Women in the CGM group gained significantly less weight than those in the SMBG group.

Table 22. Key RCT Characteristics for CGM in Pregnant Women With Gestational Diabetes 

Study 

Countries 

Sites 

Dates 

Participants

                    Interventions 

 

 

 

 

 

Acitve 

Comparator 

Wei et al (2016)32,

China

1

2011-2012 

Pregnant women with gestational diabetes diagnosed between 24 and 28 wk of gestation; mean HbA1clevel, 5.8%; mean age, 30 y 

CGM (48- 721 on weekdays) (n=51)                              

SMBG (n=55) 

CGM: continuous glucose monitoring; HbA1c: hemoglobin A1c; RCT: randomized controlled trial; SMBG: self-monitored blood glucose.

Table 23. RCT Outcomes for CGM in Pregnant Women With Gestational Diabete 

Study

Infant 

 

 

 

Maternal

 

Large-forGestational Age, n (%) 

Gestational Age at Delivery, wk 

Severe Hypoglycemia, n (%) 

Caesarean Section, n (%) 

HbA1c Levels at 36 Wk of Gestationa   Severe Hypoglycemia
Wei et al (2016)32   

 

 

     

106 

106 

106 

106   

NR 

CGM

18 (35) 

Mean, 37.4 

4 (8) 

31 (60)  Mean, 5.5% 

 

Control

29 (53) 

Mean, 37.5 

7 (13) 

38 (69) 

Mean, 5.6% 

 

TE (95% CI) 

NR 

NR 

NR 

NR  NR 

 

0.07 

0.92 

0.41  0.37  0.09   

Values are n (%) or as otherwise indicated.
CGM: continuous glucose monitoring; CI: confidence interval; HbA1c: hemoglobin A1c; NR: not reported; OR: odds ratio; RCT: randomized controlled trial; TE: treatment effect.
a N inconsistently reported for HbA1c outcome.

Tables 24 and 25 display notable limitations identified in each study.

Table 24. Relevance Limitations of RCTs for CGM in Pregnant Women With Gestational Diabetes 

Study; Trial

Allocationa

Blindingb

Selective Reportingc

Data Completenessd

Powere

Statisticalf

Wes et al (2016)32

3. Not report 

1. Not blinded; chance of bias in clinical management

1. Registration not reported 

5. Exclusions not well justified 

1. No power calculations reported; primary outcome not specified 

3, 4. Treatment effects and CIs not calculated 

The study limitations stated in this table are those notable in the current review; this is not a comprehensive limitations assessment. CGM: continuous glucose monitoring; CI: confidence interval; RCT: randomized controlled trial.
a Allocation key: 1. Participants not randomly allocated; 2. Allocation not concealed; 3. Allocation concealment unclear; 4. Inadequate control for selection bias.
b Blinding key: 1. Not blinded to treatment assignment; 2. Not blinded outcome assessment; 3. Outcome assessed by treating physician.
c Selective Reporting key: 1. Not registered; 2. Evidence of selective reporting; 3. Evidence of selective publication.
d Data Completeness key: 1. High loss to follow-up or missing data; 2. Inadequate handling of missing data; 3. High number of crossovers; 4. Inadequate handling of crossovers; 5. Inappropriate exclusions; 6. Not intent to treat analysis (per protocol for noninferiority trials).
e Power key: 1. Power calculations not reported; 2. Power not calculated for primary outcome; 3. Power not based on clinically important difference.
f Statistical key: 1. Analysis is not appropriate for outcome type: (a) continuous; (b) binary; (c) time to event; 2. Analysis is not appropriate for multiple observations per patient; 3. Confidence intervals and/or p values not reported; 4.Comparative treatment effects not calculated.

Section Summary: CGM Use in Pregnant Women With Gestational Diabetes
The RCT in women with gestational diabetes was conducted in China with the intervention starting in the 2nd or 3rd trimester and mean baseline HbA1c level less than 6.0%. The type of CGM monitoring was unclear. Trial reporting was incomplete; however, there were no differences between groups for most reported outcomes.

Summary of Evidence

The following conclusions are based on a review of the evidence, including but not limited to, published evidence and clinical expert opinion, solicited via BCBSA’s Clinical Input Process.

Type 1 Diabetes
For individuals with type 1 diabetes who are willing and able to use the device, and have adequate medical supervision, who receive long-term CGM, the evidence includes RCTs and systematic reviews. The relevant outcomes are symptoms, morbid events, QOL, and treatment-related morbidity. Systematic reviews have generally found that at least in the short-term, long-term CGM resulted in significantly improved glycemic control for adults and children with type 1 diabetes, particularly highly compliant patients. A 2017 individual patient data analysis, pooling data from 11 RCTs, found that reductions inHbA1c levels were significantly greater with real-time CGM than with a control intervention. Two RCTs in patients who used multiple daily insulin injections and were highly compliant with CGM devices during run-in phases found that CGM was associated with a larger reduction in HbA1c levels than previous studies. One of the two RCTs prespecified hypoglycemia-related outcomes and reported that time spent in hypoglycemia was significantly less in the CGM group. One RCT in pregnant women with type 1 diabetes, which compared real-time CGM with self-monitoring of blood glucose, has also reported a difference in change in HbA1clevels, an increased percentage of time in the recommended glucose control target range, a smaller proportion of infants who were large for gestational age, a smaller proportion of infants who had neonatal intensive care admissions lasting more than 24 hours, a smaller proportion of infants who had neonatal hypoglycemia requiring treatment, and reduced total hospital length of stay all favoring CGM. The evidence is sufficient that the long-term use of CGM provides an improvement in net health outcomes for persons with type 1 diabetes mellitus.

For individuals with type 1 diabetes who receive short-term glucose monitoring, the evidence includes RCTs and systematic reviews. The relevant outcomes are symptoms, morbid events, QOL, and treatment-related morbidity as well as intermediate outcomes related to measures of glucose control such as frequency and time in hypoglycemia and hyperglycemia. The evidence for short-term monitoring of glycemic control is mixed, and there was no consistency in HbA1c levels. Some trials have reported improvements in glucose control for the intermittent monitoring group but limitations in this body of evidence preclude conclusions. The definitions of control with short-term CGM use, duration of use and the specific monitoring protocols varied. In some studies, short-term monitoring was part of a larger strategy aimed at optimizing glucose control, and the impact of monitoring cannot be separated from the impact of other interventions. Studies have not shown an advantage for intermittent glucose monitoring in reducing severe hypoglycemia events but the number of events reported is generally small and effect estimates imprecise. The limited duration of use may preclude an assessment of any therapeutic effect. Two RCTs of short-term CGM use for monitoring in pregnancy included women with both type 1 and 2 diabetes, with most having type 1 diabetes. One trial reported a difference in HbA1c levels at 36 weeks; the proportion of infants that were large for gestational age (>90th percentile) favored CGM while the second trial did not. The differences in the proportions of infants born via cesarean section, gestational age at delivery, and infants with severe hypoglycemia were not statistically significant in either study. Limitations of the published evidence preclude determining the effects of the technology on net health outcome. Evidence reported through clinical input supports that this use provides a clinically meaningful improvement in net health outcome and is consistent with generally accepted medical practice when used in specific situations such as poor control of diabetes despite the use of best practices and to help determine basal insulin levels prior to insulin pump initiation. The evidence is sufficient to determine that the technology results in a meaningful improvement in the net health outcome.

Type 2 Diabetes
For individuals with type 2 diabetes who receive long-term CGM, the evidence includes RCTs. The relevant outcomes are symptoms, morbid events, QOL, and treatment-related morbidity. Most RCTs of CGM in patients with type 2 trials found statistically significant benefits of CGM regarding glycemic control. However, the degree of HbA1c reduction and the difference in HbA1c reduction between groups might not be clinically significant. Moreover, additional evidence would be needed to show what levels of improvements in HbA1c levels over the short-term would be linked to meaningful improvements over the long-term in health outcomes such as diabetes-related morbidity and complications. Also, the variability in entry criteria as well as among interventions makes it difficult to identify an optimal approach to CGM use; the studies used a combination of intermittent and continuous monitoring with a review of data in real-time or at study visits only. Only the DIAMOND RCT (n=158) has used real-time CGM in type 2 diabetes. Selected patients were highly compliant during a run-in phase. The difference in change in HbA1c levels from baseline to 24 weeks was -0.3% favoring CGM. The difference in the proportion of patients with a relative reduction in HbA1c level by 10% or more was 22% favoring CGM. There were no differences in the proportions of patients with an HbA1c level of less than 7% at week 24. There were no events of severe hypoglycemia or diabetic ketoacidosis in either group. The treatment groups did not differ in any of the QOL measures. RCTs using flash glucosesensing technology as a replacement for SMBG for the management of insulin-dependent treated type 2 diabetes found no difference in HbA1c change at 6 and 12 months between groups. However, time in severe hypoglycemia (<45mg/dL) was reduced for intervention participants. Two trials of CGM have enrolled pregnant women with type 2 diabetes, but the total number of women with type 2 diabetes included in both trials is only 58. One study reported a difference in HbA1c levels at 36 weeks, and the proportion of infants that were large for gestational age (>90th percentile) favored CGM while the second study did not. Neither trial reported analyses stratified by diabetes type. Limitations of the published evidence preclude determining the effects of the technology on net health outcome. Evidence reported through clinical input for long-term (continuous) CGM in patients with type 2 diabetes who do not require insulin did not provide strong support of a safety benefit and clinically meaningful improvement in net health outcome. The evidence is insufficient to determine the effects of the technology on health outcomes.

For individuals with type 2 diabetes who are willing and able to use the device and have adequate medical supervision and who experience significant hypoglycemia on multiple daily doses of insulin or an insulin pump in the setting of insulin deficiency who receive long-term (continuous) glucose monitoring, the evidence includes a systematic review and nonrandomized study with 12-month follow-up. The relevant outcomes are the frequency of and time spent in hypoglycemia, the incidence of hypoglycemic episodes, complications of hypoglycemia, and QOL. The available studies demonstrate that CGM can significantly reduce time in hypoglycemia and frequency of hypoglycemia events both during the day and at night. At 12-month follow-up, hypoglycemic events were reduced by 40.8% to 61.7% with a greater relative reduction in the most severe thresholds of hypoglycemia. The published evidence supports a meaningful improvement in the net health outcome. Evidence reported through clinical input provides additional clinical context and based on both the published evidence and clinical input the following patient selection criteria are associated with a clinically meaningful improvement in net health outcome and are consistent with generally accepted medical practice: selected patients with type 2 diabetes who are (1) willing and able to use the CGM device and have adequate medical supervision and (2) experiencing significant hypoglycemia on multiple daily doses of insulin or an insulin pump in the setting of insulin deficiency. The evidence is sufficient to determine that the technology results in a meaningful improvement in the net health outcome.

For individuals with type 2 diabetes who receive short-term CGM monitoring, the evidence includes RCTs and systematic reviews. The relevant outcomes are symptoms, morbid events, QOL, and treatment-related morbidity. Systematic reviews of three to four RCTs have found statistically significant benefits from CGM regarding glycemic control. However, the degree of HbA1c reduction and the difference in HbA1c reductions between groups may not be clinically significant. Also, thelimited number of RCTs and variability among interventions make it difficult to identify an optimal approach to CGM or a subgroup of type 2 diabetes patients who might benefit. Moreover, studies of CGM in patients with type 2 diabetes have generally not addressed the clinically important issues of severe hypoglycemia and diabetic complications. Very few pregnant women with type 2 diabetes have been included in RCTs.Limitations of the published evidence preclude determining the effects of the technology on net health outcome. Evidence reported through clinical input for use of short-term CGM in patients with type 2 diabetes who require multiple daily doses of insulin supports that this use provides a clinically meaningful improvement in net health outcome and is consistent with generally accepted medical practice when used in specific situations such as poor control of diabetes despite use of best practices and to help determine basal insulin levels prior to insulin pump initiation. The evidence is sufficient to determine that the technology results in a meaningful improvement in the net health outcome.

Gestational Diabetes
For individuals who are pregnant with gestational diabetes who receive long-term CGM or short-term (intermittent) glucose monitoring, the evidence includes an RCT. The relevant outcomes are symptoms, morbid events, QOL, and treatment-related morbidity. In the RCT, the type of glucose monitoring was unclear. Trial reporting was incomplete; however, there was no difference between the groups for most reported outcomes. The evidence is insufficient to determine the effects of the technology on health outcomes. 

Clinical Input From Physician Specialty Societies and Academic Medical Centers
While the various physician specialty societies and academic medical centers may collaborate with and make recommendations during this process, through the provision of appropriate reviewers, input received does not represent an endorsement or position statement by the physician specialty societies or academic medical centers, unless otherwise noted.

2019
In response to requests, while this policy was under review in 2019, clinical input on the use of continuous or intermittent monitoring of glucose in the interstitial fluid was received from 3 respondents, including 3 physician-level responses identified through one specialty society including 2 physicians with academic medical center affiliations. Evidence from clinical input is integrated within the Rationale section summaries and the Summary of Evidence.

2008
In response to requests, input was received from 1 physician specialty society and 4 academic medical centers while this policy was under review in 2008. Input concurred that continuous glucose monitoring, particularly intermittent glucose monitoring, was helpful in a subset of patients with diabetes. Reviewers commented that this monitoring can improve diabetes care by reducing glucose levels (and improving hemoglobin A1clevels) and/or by reducing episodes of hypoglycemia. Reviewers argued that there is persuasive data from case reports to demonstrate the positive impact of intermittent glucose monitoring.

Practice Guidelines and Position Statements
American Association of Clinical Endocrinologists and the American College of Endocrinology
The AACE and the ACE published a consensus statement on outpatient glucose monitoring.[36] Their recommendations on continuous glucose monitoring (CGM) included:

Type 1 diabetes, adults: “CGM recommended, especially for patients with history of severe hypoglycemia, hypoglycemia unawareness and to assist in the correction of hyperglycemia in patients not at goal. CGM users must know basics of sensor insertion, calibration and real-time data interpretation.”

Type 1 diabetes, children: Same as adults, except that more training and follow-up are needed.

Type 2 diabetes receiving insulin, sulfonylureas, or glinides: “Data on CGM in T2DM [type 2 diabetes mellitus] are limited at this time. Trials assessing the use of CGM in T2DM are ongoing."

The AACE and the ACE (2018) published a consensus statement on a T2D management algorithm. It is recommended that therapy be evaluated regularly including the results of A1C, SMBG records (fasting and postprandial) or continuous glucose monitoring tracings.[35]

In 2019, the AACE and the ACE 2015 Clinical Practice Guidelines for Developing a Diabetes Mellitus Comprehensive Care Plan further supplemented by an AACE/ACE Consensus Statement on Comprehensive Type 2 Diabetes Management. The statement supports consideration of the use of personal CGM devices for those patients who are on intensive insulin therapy (three to four injections/day or on an insulin pump), for those with a history of hypoglycemia unawareness, or those with recurrent hypoglycemia. Regarding the duration of use the statement reads; “While these devices could be used intermittently in those who appear stable on their therapy, most patients will need to use this technology on a continual basis.”[39]

National Institute for Health and Care Excellence
The National Institute for Health and Care Excellence (2016) updated its guidance on the diagnosis and management of type 1 diabetes in adults.[34] The guidance stated that realtime CGM should not be offered “routinely to adults with type 1 diabetes” but that it can be considered in the following:

"...adults with type 1 diabetes who are willing to commit to using it at least 70% of the time and to calibrate it as needed, and who have any of the following despite optimised use of insulin therapy and conventional blood glucose monitoring:

  • More than 1 episode a year of severe hypoglycaemia with no obviously preventable precipitating cause.
  • Complete loss of awareness of hypoglycaemia.
  • Frequent (more than 2 episodes a week) asymptomatic hypoglycaemia that is causing problems with daily activities.
  • Extreme fear of hypoglycaemia.
  • Hyperglycaemia (HbA1c [hemoglobin A1c] level of 75 mmol/mol [9%] or higher) that persists despite testing at least 10 times a day. Continue real‑time continuous glucose monitoring only if HbA1c can be sustained at or below 53 mmol/mol (7%) and/or there has been a fall in HbA1c of 27 mmol/mol (2.5%) or more."

American Diabetes Association
The American Diabetes Association (2019) “Standards of Medical Care in Diabetes: Diabetes Technology" included the following statement:[40]

" SMBG or CGM is especially important for insulin-treated patients to monitor for and prevent hypoglycemia and hyperglycemia. Most patients using intensive insulin regimens (MDI or insulin pump therapy) should assess glucose levels using SMBG or a CGM prior to meals and snacks, at bedtime, occasionally postprandially, prior to exercise, when they suspect low blood glucose, after treating low blood glucose until they are normoglycemic, and prior to critical tasks such as driving"[40]

The Standards (2019) also state that data provided by CGM "offers opportunities to analyze data more granularly than was previously possible, providing additional information to aid in achieving glycemic targets," and note that metrics have been proposed that may include the following:[52]

  1. average glucose
  2. percentage of time in hypoglycemic ranges (i.e. <54 mg/dL [level 2], 54-70 mg/dL [level 1])
  3. percentage of time in target range (i.e. 70-180 mg/dL)
  4. percentage of time in hyperglycemic range (>180 mg/dL)

Endocrine Society
The Endocrine Society (2016) published clinical practice guidelines that included the following recommendations on CGM[36]:

6. "Real-time continuous glucose monitors in adult outpatients

6.1 We recommend real-time continuous glucose monitoring (RT-CGM) devices for adult patients with T1DM [type 1 diabetes mellitus] who have A1C levels above target and who are willing and able to use these devices on a nearly daily basis.

6.2 We recommend RT-CGM devices for adult patients with well-controlled T1DM who are willing and able to use these devices on a nearly daily basis.

Use of continuous glucose monitoring in adults with type 2 diabetes mellitus [T2DM]

6.3 We suggest short-term, intermittent RT-CGM use in adult patients with T2DM (not on prandial insulin) who have A1C levels ≥7% and are willing and able to use the device."

International Consensus on Time in Range
In 2019, consensus recommendations on clinical targets for CGM data interpretation were published and endorsed by the American Diabetes Association, American Association of Diabetes Educators, European Association for the Study of Diabetes, Foundation of European Nurses in Diabetes, International Society for Pediatric and Adolescent Diabetes, JDRF, and Pediatric Endocrine Society.[51]

U.S. Preventive Services Task Force Recommendations
Not applicable. 

Ongoing and Unpublished Clinical Trials
Some currently unpublished trials that might influence this review are listed in Table 26.

Table 26. Summary of Key Trials 

NCT No.

Trial Name

Enrollment

Date

Ongoing

 

 

 

NCT03908125a 

A Post- Approval Study to Evaluate the Long-term Safety and Effectiveness of the Eversense® Continuous Glucose Monitoring (CGM) System 

400 

Mar 2023 

NCT03808376a

PROMISE Study: A Prospective, Multicenter Evaluation of Accuracy and Safety of an Implantable Continuous Glucose Sensor Lasting up to 180 Days

180

Mar 2020

NCT03445065a

Benefits of a Long Term Implantable Continuous Glucose Monitoring System for Adults With Diabetes - France Randomized Clinical Trial

324

Jun 2020

Unpublished

 

 

 

NCT03263494 CGM Intervention in Teens and Young Adults With T1D (CITY): A Randomized Clinical Trial to Assess the Efficacy and Safety of Continuous Glucose Monitoring in Young Adults 14-<25 With Type 1 Diabetes

200

Jul  2019

NCT02838147

Effect of a Continuous Glucose Monitoring on Maternal and Neonatal Outcomes in Gestational Diabetes Mellitus: A Randomized Controlled Tria

200

Jul 2019

NCT: national clinical trial.
a Denotes industry-sponsored or cosponsored trial.  

References

  1. Effect of intensive therapy on the microvascular complications of type 1 diabetes mellitus. Jama. May 15 2002;287(19):2563-2569. PMID 12020338.
  2. Management of Diabetes in Pregnancy: Standards of Medical Care in Diabetes-2018. Diabetes Care. Jan 2018;41(Suppl 1):S137-s143. PMID 29222384.
  3. Pazos-Couselo M, Garcia-Lopez JM, Gonzalez-Rodriguez M, et al. High incidence of hypoglycemia in stable insulin-treated type 2 diabetes mellitus: continuous glucose monitoring vs. self-monitored blood glucose. Observational prospective study. Can J Diabetes. Oct 2015;39(5):428-433. PMID 26254702.
  4. Gehlaut RR, Dogbey GY, Schwartz FL, et al. Hypoglycemia in type 2 diabetes--more common than you think: a continuous glucose monitoring study. J Diabetes Sci Technol. Sep 2015;9(5):999-1005. PMID 25917335.
  5. Food and Drug Administration (FDA). Summary of Safety and Effectiveness (SSED): Dexcom G5 Mobile Continuous Glucose Monitoring System. 2016; https://www.accessdata.fda.gov/cdrh_docs/pdf12/P120005S041b.pdf. Accessed October 18, 2019..
  6. Blue Cross and Blue Shield Technology Evaluation Center (TEC). Use of intermittent or continuous interstitial fluid glucose monitoring in patients with diabetes mellitus. TEC Assessments. 2003;Volume 18:Tab 16..
  7. Floyd B, Chandra P, Hall S, et al. Comparative analysis of the efficacy of continuous glucose monitoring and self- monitoring of blood glucose in type 1 diabetes mellitus. J Diabetes Sci Technol. Sep 2012;6(5):1094-1102. PMID 23063035.
  8. Gandhi GY, Kovalaske M, Kudva Y, et al. Efficacy of continuous glucose monitoring in improving glycemic control and reducing hypoglycemia: a systematic review and metaanalysis of randomized trials. J Diabetes Sci Technol. Jul 2011;5(4):952-965. PMID 21880239.
  9. Langendam M, Luijf YM, Hooft L, et al. Continuous glucose monitoring systems for type 1 diabetes mellitus. Cochrane Database Syst Rev. Jan 18 2012;1:CD008101. PMID 22258980.
  10. Poolsup N, Suksomboon N, Kyaw AM. Systematic review and meta-analysis of the effectiveness of continuous glucose monitoring (CGM) on glucose control in diabetes.Diabetol Metab Syndr. Jul 23 2013;5(1):39. PMID 23876067.
  11. Wojciechowski P, Rys P, Lipowska A, et al. Efficacy and safety comparison of continuous glucose monitoring and self-monitoring of blood glucose in type 1 diabetes: systematic review and meta-analysis. Pol Arch Med Wewn. Oct 2011;121(10):333-343. PMID 22045094.
  12. Yeoh E, Choudhary P, Nwokolo M, et al. Interventions that restore awareness of hypoglycemia in adults with type 1 diabetes: a systematic review and meta-analysis. Diabetes Care. Aug 2015;38(8):1592-1609. PMID 26207053.
  13. Benkhadra K, Alahdab F, Tamhane S, et al. Real-time continuous glucose monitoring in type 1 diabetes: a systematic review and individual patient data meta-analysis. Clin Endocrinol (Oxf). Mar 2017;86(3):354-360. PMID 27978595.
  14. van Beers CA, DeVries JH, Kleijer SJ, et al. Continuous glucose monitoring for patients with type 1 diabetes and impaired awareness of hypoglycaemia (IN CONTROL): a randomised, open-label, crossover trial. Lancet Diabetes Endocrinol. Nov 2016;4(11):893-902. PMID 27641781.
  15. Gold AE, MacLeod KM, Frier BM. Frequency of severe hypoglycemia in patients with type I diabetes with impaired awareness of hypoglycemia. Diabetes Care. Jul 1994;17(7):697-703. PMID 7924780.
  16. Lind M, Polonsky W, Hirsch IB, et al. Continuous glucose monitoring vs conventional therapy for glycemic control in adults with type 1 diabetes treated with multiple daily insulin injections: The GOLD randomized clinical trial. Jama. Jan 24 2017;317(4):379-387. PMID 28118454.
  17. Beck RW, Riddlesworth T, Ruedy K, et al. Effect of continuous glucose monitoring on glycemic control in adults with type 1 diabetes using insulin injections: The DIAMOND randomized clinical trial. Jama. Jan 24 2017;317(4):371-378. PMID 28118453.
  18. Riddlesworth T, Price D, Cohen N, et al. Hypoglycemic event frequency and the effect of continuous glucose monitoring in adults with type 1 diabetes using multiple daily insulin injections. Diabetes Ther. Aug 2017;8(4):947-951. PMID 28616804.
  19. Polonsky WH, Hessler D, Ruedy KJ, et al. The impact of continuous glucose monitoring on markers of quality of life in adults with type 1 diabetes: further findings from the DIAMOND randomized clinical trial. Diabetes Care. Jun 2017;40(6):736-741. PMID 28389582.
  20. Feig DS, Donovan LE, Corcoy R, et al. Continuous glucose monitoring in pregnant women with type 1 diabetes (CONCEPTT): a multicentre international randomised controlled trial. Lancet. Nov 25 2017;390(10110):2347- 2359. PMID 28923465.
  21. Newman SP, Cooke D, Casbard A, et al. A randomised controlled trial to compare minimally invasive glucose monitoring devices with conventional monitoring in the management of insulin-treated diabetes mellitus (MITRE). Health Technol Assess. May 2009;13(28):iii-iv, ix-xi, 1-194. PMID 19476724.
  22. Voormolen DN, Devries JH, Evers IM, et al. The efficacy and effectiveness of continuous glucose monitoring during pregnancy: a systematic review. Obstet Gynecol Surv. Nov 2013;68(11):753-763. PMID 24193194.
  23. Murphy HR, Rayman G, Lewis K, et al. Effectiveness of continuous glucose monitoring in pregnant women with diabetes: randomised clinical trial. BMJ. Sep 25 2008;337:a1680. PMID 18818254.
  24. Secher AL, Ringholm L, Andersen HU, et al. The effect of real-time continuous glucose monitoring in pregnant women with diabetes: a randomized controlled trial. Diabetes Care. Jul 2013;36(7):1877-1883. PMID 23349548.
  25. Allen NA, Fain JA, Braun B, et al. Continuous glucose monitoring counseling improves physical activity behaviors of individuals with type 2 diabetes: A randomized clinical trial. Diabetes Res Clin Pract. Jun 2008;80(3):371-379. PMID 18304674.
  26. Yoo HJ, An HG, Park SY, et al. Use of a real time continuous glucose monitoring system as a motivational device for poorly controlled type 2 diabetes. Diabetes Res Clin Pract. Oct 2008;82(1):73-79. PMID 18701183.
  27. Cosson E, Hamo-Tchatchouang E, Dufaitre-Patouraux L, et al. Multicentre, randomised, controlled study of the impact of continuous sub-cutaneous glucose monitoring (GlucoDay) on glycaemic control in type 1 and type 2 diabetes patients. Diabetes Metab. Sep 2009;35(4):312-318. PMID 19560388.
  28. Ehrhardt NM, Chellappa M, Walker MS, et al. The effect of real-time continuous glucose monitoring on glycemic control in patients with type 2 diabetes mellitus. J Diabetes Sci Technol. May 2011;5(3):668-675. PMID 21722581.
  29. Vigersky RA, Fonda SJ, Chellappa M, et al. Short- and long-term effects of real-time continuous glucose monitoring in patients with type 2 diabetes. Diabetes Care. Jan 2012;35(1):32-38. PMID 22100963.
  30. Sato J, Kanazawa A, Ikeda F, et al. Effect of treatment guidance using a retrospective continuous glucose monitoring system on glycaemic control in outpatients with type 2 diabetes mellitus: A randomized controlled trial. J Int Med Res. Feb 2016;44(1):109-121. PMID 26647072.
  31. Beck RW, Riddlesworth TD, Ruedy K, et al. Continuous glucose monitoring versus usual care in patients with type 2 diabetes receiving multiple daily insulin injections: a randomized trial. Ann Intern Med. Sep 19 2017;167(6):365-374. PMID 28828487.
  32. Wei Q, Sun Z, Yang Y, et al. Effect of a CGMS and SMBG on maternal and neonatal outcomes in gestational diabetes mellitus: a randomized controlled trial. Sci Rep. Jan 27 2016;6:19920. PMID 26814139.
  33. Bailey TS, Grunberger G, Bode BW, et al. American Association of Clinical Endocrinologists and American College of Endocrinology 2016 outpatient glucose monitoring consensus statement. Endocr Pract. Feb 2016;22(2):231-261. PMID 26848630.
  34. National Institute for Health and Care Excellence (NICE). Type 1 diabetes in adults: diagnosis and management [NG17]. 2016; https://www.nice.org.uk/guidance/ng17? unlid=382286372016220232952. Accessed October 18, 2019.
  35. American Diabetes Association (ADA). 6. Glycemic Targets. Diabetes Care. Jan 2017;40(Suppl 1):S48-S56. PMID 27979893.
  36. Peters AL, Ahmann AJ, Battelino T, et al. Diabetes technology-continuous subcutaneous insulin infusion therapy and continuous glucose monitoring in adults: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. Nov 2016;101(11):3922-3937. PMID 27588440.
  37. Centers for Medicare & Medicare Services. Durable Medical Equipment, Prosthetics/Orthotics & Supplies Fee Schedule 2017; https://www.cms.gov/medicare/medicare-fee-forservice-payment/dmeposfeesched/index.html. Accessed October 18, 2019.
  38. Ida, SS, Kaneko, RR, Murata, KK. Utility of Real-Time and Retrospective Continuous Glucose Monitoring in Patients with Type 2 Diabetes Mellitus: A Meta-Analysis of Randomized Controlled Trials.. J Diabetes Res, 2019 Feb 19;2019:4684815. PMID 30775385.
  39. Garber, AA, Abrahamson, MM, Barzilay, JJ, Blonde, LL, Bloomgarden, ZZ, Bush, MM, Dagogo-Jack, SS, DeFronzo, RR, Einhorn, DD, Fonseca, VV, Garber, JJ, Garvey, WW, Grunberger, GG, Handelsman, YY, Hirsch, II, Jellinger, PP, McGill, JJ, Mechanick, JJ, Rosenblit, PP, Umpierrez, GG. CONSENSUS STATEMENT BY THE AMERICAN ASSOCIATION OF CLINICAL ENDOCRINOLOGISTS AND AMERICAN COLLEGE OF ENDOCRINOLOGY ON THE COMPREHENSIVE TYPE 2 DIABETES MANAGEMENT ALGORITHM - 2019 EXECUTIVE SUMMARY.. Endocr Pract, 2019 Feb 12;25(1). PMID 30742570.
  40. American Diabetes Association. 7. Diabetes Technology: Standards of Medical Care in Diabetes-2019. Diabetes Care, 2018 Dec 19;42(Suppl 1). PMID 30559233.
  41. Christiansen, MM, Klaff, LL, Brazg, RR, Chang, AA, Levy, CC, Lam, DD, Denham, DD, Atiee, GG, Bode, BB, Walters, SS, Kelley, LL, Bailey, TT. A Prospective Multicenter Evaluation of the Accuracy of a Novel Implanted Continuous Glucose Sensor: PRECISE II.. Diabetes Technol. Ther., 2018 Jan 31;20(3). PMID 29381090.
  42. Kropff, JJ, Choudhary, PP, Neupane, SS, Barnard, KK, Bain, SS, Kapitza, CC, Forst, TT, Link, MM, Dehennis, AA, DeVries, JJ. Accuracy and Longevity of an Implantable Continuous Glucose Sensor in the PRECISE Study: A 180-Day, Prospective, Multicenter, Pivotal Trial.. Diabetes Care, 2016 Nov 7;40(1). PMID 27815290.
  43. Ajjan, RR, Abougila, KK, Bellary, SS, Collier, AA, Franke, BB, Jude, EE, Rayman, GG, Robinson, AA, Singh, BB. Sensor and software use for the glycaemic management of insulin-treated type 1 and type 2 diabetes patients.. Diab Vasc Dis Res, 2016 Mar 24;13(3). PMID 27000105.
  44. Christiansen, MM, Klaff, LL, Bailey, TT, Brazg, RR, Carlson, GG, Tweden, KK. A Prospective Multicenter Evaluation of the Accuracy and Safety of an Implanted Continuous Glucose Sensor: The PRECISION Study.. Diabetes Technol. Ther., 2019 Mar 30;21(5). PMID 30925083.
  45. Haak, TT, Hanaire, HH, Ajjan, RR, Hermanns, NN, Riveline, JJ, Rayman, GG. Use of Flash Glucose-Sensing Technology for 12 months as a Replacement for Blood Glucose Monitoring in Insulin-treated Type 2 Diabetes.. Diabetes Ther, 2017 Apr 13;8(3). PMID 28401454.
  46. Haak, TT, Hanaire, HH, Ajjan, RR, Hermanns, NN, Riveline, JJ, Rayman, GG. Flash Glucose-Sensing Technology as a Replacement for Blood Glucose Monitoring for the Management of Insulin-Treated Type 2 Diabetes: a Multicenter, Open-Label Randomized Controlled Trial.. Diabetes Ther, 2016 Dec 22;8(1). PMID 28000140.
  47. Centers for Medicare & Medicare Services. Durable Medical Equipment (DME) Center; https://www.cms.gov/Center/Provider-Type/Durable-Medical-Equipment-DMECenter.html. Accessed October 28, 2019.
  48. Deiss D, Irace C, Carlson G et al. Real-World Safety of an Implantable Continuous Glucose Sensor over Multiple Cycles of Use: A Post-Market Registry Study.. Diabetes Technol. Ther., 2019 Aug 17. PMID 31418587.
  49. Sanchez P, Ghosh-Dastidar S, Tweden KS et al. Real-World Data from the First U.S. Commercial Users of an Implantable Continuous Glucose Sensor.. Diabetes Technol. Ther., 2019 Aug 7. PMID 31385732.
  50. Food and Drug Administration. Summary of Safety and Effectiveness Data: Eversense Continuous Glucose Monitoring System (2019). https://www.accessdata.fda.gov/cdrh_docs/pdf16/P160048S006b.pdf. Accessed October 28, 2019..
  51. Battelino T, Danne T, Bergenstal RM et al. Clinical Targets for Continuous Glucose Monitoring Data Interpretation: Recommendations From the International Consensus on Time in Range. Diabetes Care, 2019 Jun 10;42(8). PMID 31177185.
  52. American Diabetes Association. 6. Glycemic Targets: Standards of Medical Care in Diabetes-2019. Diabetes Care. 2019 Jan;42(Suppl 1):S61-S70. PMID: 30559232..
  53. Tweden KS, Deiss D, Rastogi R et al. Longitudinal Analysis of Real-World Performance of an Implantable Continuous Glucose Sensor Over Multiple Sensor Insertion and Removal Cycles.. Diabetes Technol. Ther., 2019 Nov 8. PMID 31697182.  

Coding Section 

Codes

Number

Description

CPT

95250

Ambulatory continuous glucose monitoring of interstitial tissue fluid via a subcutaneous sensor for a minimum of 72 hours; sensor placement, hook-up, calibration of monitor, patient training, removal of sensor, and printout of recording

 

95249

; same as 95250 patient owned equipment (effective 01/01/18)

 

95251

; physician interpretation and report

 

0446T

Creation of subcutaneous pocket with insertion of implantable interstitial glucose sensor, including system activation and patient training

 

0447T

Removal of implantable interstitial glucose sensor from subcutaneous pocket via incision

 

0448T

Removal of implantable interstitial glucose sensor with creation of subcutaneous pocket at different anatomic site and insertion of new implantable sensor, including system activation

HCPCS

A9276

Sensor; invasive (e.g., subcutaneous), disposable, for use with interstitial continuous glucose monitoring system, one unit=1 day supply

 

A9277

Transmitter; external, for use with interstitial continuous glucose monitoring system

 

A9278

Receiver (monitor); external, for use with interstitial continuous glucose monitoring system

 

K0553

Supply allowance for therapeutic continuous glucose monitor (CGM) system, includes all supplies and accessories, 1 month supply = 1 unit of service.

 

K0554

Receiver (monitor), dedicated, for use with therapeutic continuous glucose monitor system.

 

S1030

Continuous non-invasive glucose monitoring device, purchase (for physician interpretation of data, use CPT code)

 

S1031 

Continuous non-invasive glucose monitoring device, rental, including sensor, sensor replacement, and download to monitor (for physician interpretation of data, use CPT code) 

ICD-10-CM 

E10.10-E13.9 

Diabetes mellitus code range 

ICD-10-PCS 

 

ICD-10-PCS codes are only used for inpatient services. There is no specific ICD-10- PCS code for this monitoring. 

Type of Service 

Medicine 

 

Place of Service 

Outpatient 

 

Procedure and diagnosis codes on Medical Policy documents are included only as a general reference tool for each policy. They may not be all-inclusive.

This medical policy was developed through consideration of peer-reviewed medical literature generally recognized by the relevant medical community, U.S. FDA approval status, nationally accepted standards of medical practice and accepted standards of medical practice in this community, Blue Cross and Blue Shield Association technology assessment program (TEC) and other non-affiliated technology evaluation centers, reference to federal regulations, other plan medical policies, and accredited national guidelines.

"Current Procedural Terminology © American Medical Association.  All Rights Reserved" 

History From 2013 Forward     

01/06/2020 

Interim review to remove requirement that this device be ordered by an endocrinologist. No other changes made. 

12/18/2019 

Interim review to add language regarding implantable CGM devices. Also updating description, guidelines, rationale and references. 

02/12/2019 

Annual review, no change to policy intent. Updating description, background, regulatory status, rationale, references and coding. 

03/13/2018 

Annual review, no change to policy intent. Updating background, description, regulatory status, HCPCS coding in guidelines, rationale and references. 

12/19/2017 

Interim review, correcting typo and doing minor editing in description and rationale related to type 2 diabetes. No change in policy intent. 

12/12/2017 

Interim review, updating policy verbiage significantly to include reformatted medical necessity criteria and criteria to allow for coverage for Type II diabetes. 

05/15/2017 

Interim Review. Updated Policy statement and Policy guidelines.  

02/01/2017

Annual Review, increasing the glucose level for hypoglycemia from 50 to 70 in policy statement. Also updating background, description, rationale and references. 

02/10/2016 

Annual review, no change to policy intent. 

2/05/2015

Interim update. Removing verbiage related to artificial pancreas as a new policy, CAM 10130 has been created to address that issue. Updated background, description, related policy, regulatory status, rationale, references and coding. 

11/10/2014

Annual review, no change to policy intent. Added coding section. Updated regulatory status, policy guidelines, rationale and references.

11/11/2013

Updated Description and Rationale.


Go Back