Continuous Glucose Monitoring Use in Clinical Trials for On-Market Diabetes Drugs

To the best of our knowledge, there are no published data on the historical and recent use of CGM in clinical trials of pharmacological agents used in the treatment of diabetes. We analyzed 2,032 clinical trials of 40 antihyperglycemic therapies currently on the market with a study start date between 1 January 2000 and 31 December 2019. According to ClinicalTrials.gov, 119 (5.9%) of these trials used CGM. CGM usage in clinical trials has increased over time, rising from <5% before 2005 to 12.5% in 2019. However, it is still low given its inclusion in the American Diabetes Association’s latest guidelines and known limitations of A1C for assessing ongoing diabetes care.

The availability of reliable continuous glucose monitoring (CGM) systems has proven to be a major innovation in diabetes management and research. Most current CGM systems are approved for 7- to 14-day use and use a wire-tipped glucose oxidase sensor inserted in subcutaneous tissue to monitor glucose concentrations in interstitial fluid. One implanted CGM system is approved for longer-term use (90–180 days); it operates with fluorescence-based technology. CGM sensors record a glucose data point every 1–15 minutes (depending on the system), collecting far more granular data and information on glycemic patterns than self-monitoring of blood glucose (SMBG) alone. Real-time CGM or intermittently scanned CGM systems send data continuously or intermittently to dedicated receivers or smartphones, whereas professional CGM systems provide retrospective data, either blinded or unblinded, for analysis and can be used to identify patterns of hypo- and hyperglycemia. Professional CGM can be helpful to evaluate patients when other CGM systems are not available to the patient or the patient prefers a blinded analysis or a shorter experience with unblinded data.In the 20 years since CGM systems first became available to people with diabetes, technological improvements, particularly pertaining to accuracy and form factor, have made CGM increasingly viable for both patient use and clinical investigation (1,2). Average sensor MARD (mean absolute relative difference; a summary accuracy statistic) has decreased from >20 to <10% (3–10), including two systems that do not require fingerstick calibrations and three that are approved to be used for insulin dosing (11). Concurrently, size, weight, and cost of CGM systems have all decreased, while user-friendliness and convenience have increased (12).To encourage use of CGM-derived data, researchers and clinicians have worked to develop a standard set of glycemic metrics beyond A1C. In 2017, two international groups of leading diabetes clinical and research organizations published consensus definitions for key metrics, including clinically relevant glycemic cut points for hypoglycemia (<70 and <54 mg/dL), hyperglycemia (>180 and >250 mg/dL), and time in range (TIR; 70–180 mg/dL) (13,14).CGM-derived metrics provide far greater precision and granularity than is possible with SMBG or A1C data alone (15). Crucially, CGM also allows for the accurate measurement and detection of nocturnal glycemia (16). The use of these metrics enables a more comprehensive understanding of glycemic management that can facilitate individualized treatment for people with diabetes or prediabetes. Although A1C is a useful estimate of mean glucose over the previous 2–3 months, especially when evaluating population health, it is important to include other glycemic outcomes in clinical trials. Furthermore, there is emerging evidence suggesting that TIR predicts the development of microvascular complications at least as well as A1C (17,18).TABLE 1Benefits of CGM Compared With A1C Alone in Assessing Glycemia

CGM	A1C Alone
Facilitates real-time readings of blood glucose levels	Requires SMBG
Provides information on glucose variability, including duration of hypo- and hyperglycemia and nocturnal glycemia	Does not provide information on acute glycemic excursions and time in biochemical hypoglycemia and hyperglycemia
Correlates strongly with 3 months of mean glucose, TIR, and hyperglycemia metrics	Measures average glucose during the past 2–3 months
Provides information on direction of and rate of change in glucose levels	Does not provide information on direction of or rate of change in glucose levels
Provides TIR data (time spent between 70 and 180 mg/dL)	Does not have TIR measurement capability

Open in a separate windowDespite recent standardization of metrics and an emerging consensus around the importance of including CGM-derived outcomes in clinical trials, to our knowledge, there has been no attempt to estimate the historical and current use of CGM in clinical trials of pharmacological agents for diabetes. We sought to analyze the use of CGM in trials of currently available pharmaceutical agents for the treatment of diabetes.