Flash Continuous Glucose Monitoring: A Summary Review of Recent Real-World Evidence

The International Diabetes Federation estimated the global prevalence of diabetes in 2019 to be 9.3% (463 million people), and this proportion is expected to rise to 10.2% (578 million people) by 2030 (7), with associated annual costs of ∼$2.25 trillion (7). This figure includes $1.5 trillion in direct costs and $730 billion in indirect costs (e.g., absenteeism and societal costs) resulting from uncontrolled diabetes (7).

Therefore, preventing or reducing the severity of acute and long-term diabetes complications through patient-centered care remains a primary goal of diabetes management, and maintaining optimal glycemic control is central to achieving this goal (8,9). Although improvements in glycemic control should always be a priority, it is also important to strike a balance among the clinical benefits of new diabetes technologies, the initial and ongoing costs associated with their use, and the long-term gains for health and well-being.

Many studies have demonstrated that lower A1C levels result in lower health care resource utilization and associated costs (10–14). In a recent U.K. analysis using the IMS Core Diabetes Model, the per-person cost reductions that could be achieved over time were calculated based on a 0.4% (4.4 mmol/mol) reduction from baseline A1C (12). As shown in Table 1, the cost savings are most notable for individuals with the highest baseline A1C levels.

The global HAT (Hypoglycemia Assessment Tool) study, a 6-month retrospective and 4-week prospective investigation of 27,585 insulin-treated patients (type 1 diabetes, n = 8,022; type 2 diabetes, n = 19,563) in 24 countries noted the costs of inadequate glycemic control (15). During the prospective period, 83% of patients with type 1 diabetes and 46.5% of those with type 2 diabetes reported hypoglycemia, resulting in increased blood glucose monitoring, a marked increase in contact with health care providers, and increased hospitalizations. Significant indirect costs were incurred during the prospective period, with lost work time averaging 2.0 days for patients with type 1 diabetes and 1.8 days for those with type 2 diabetes. Other studies have had similar findings (15). Importantly, any level of hypoglycemia confers substantial indirect costs on employers as well as on individuals with diabetes because of increased work days lost (16–18).

Although well-designed RCTs provide high levels of evidence, there is growing recognition for the complementary relationship between RCTs and real-world prospective and retrospective observational studies. An increasing number of payers and regulators now request that pharmaceutical and medical device companies provide real-world evidence alongside RCT findings when evaluating the safety and effectiveness of new drugs and medical devices (19–22).

Large RCTs have clearly established that use of flash CGM improves glycemic control, reduces hypoglycemia, and achieves higher treatment satisfaction scores among individuals with type 1 diabetes (1,2) or type 2 diabetes (4,5) who are treated with intensive insulin therapy. Now real-world studies are investigating the use of flash CGM within different clinical settings and diverse diabetes populations.

As presented in Table 2, results from recently published prospective, observational studies closely align with glycemic benefits reported in earlier RCTs and also demonstrate the value of flash CGM use on cost outcomes and quality of life (QoL) measures (23–31). While these studies confirm significant reductions in A1C (23–25,32) and hypoglycemia (23,24) within large populations with type 1 or type 2 diabetes, they also provide strong evidence linking metabolic outcomes of flash CGM use to reductions in health care resource utilization. For example, one prospective, observational study assessed the impact of flash CGM in an unselected real-world cohort of 1,913 adults with type 1 diabetes (23). Over the 12-month study period, admissions for severe hypoglycemia and/or diabetic ketoacidosis (DKA) decreased from 3.3 to 2.2% (P = 0.031), and fewer individuals reported severe hypoglycemic events (7.8 vs. 14.6%, P <0.0001) or experienced hypoglycemic coma (1.1 vs. 2.7%, P = 0.001). Although measures of general and diabetes-specific QoL were relatively high at baseline and remained stable throughout the study, treatment satisfaction was increased by study end. Moreover, fewer subjects were absent from work (2.9 vs. 5.8%, P <0.0001), a metric that is often not reported in RCTs and provides an informative indicator of economic benefit.

Similarly, an analysis of a Dutch registry assessed the impact of flash CGM use among 1,365 individuals with diabetes (77% with type 1 diabetes, 16% with type 2 diabetes, and 7% with other types of diabetes) (24). After 12 months of flash CGM use, A1C decreased from 8.0 to 7.4%, with the greatest reductions occurring among participants with a baseline A1C >8.6%. The percentage of patients experiencing any hypoglycemic event decreased from 93.5 to 91.0% (P <0.05), and the number of diabetes-related hospital admissions decreased from 144 before baseline to 22 at 12 months (P <0.001). Additionally, flash CGM users reported reduced diabetes burden with the SF-12 (12-item Short Form, v. 2) survey, EQ-5D-3 L (EuroQol 5-Dimension, three-level version) instrument, and DVN-PROM (Diabetes Vereniging Nederland Patient-Reported Outcomes Measure) questionnaire. The majority of study participants reported fewer hypoglycemic events (77%), less severe hypoglycemia (78%), more frequent insulin dose adjustments (80%), better understanding of their glucose fluctuations (95%), and less worry about their diabetes among housemates and family members (62%). Moreover, 81.7% felt no inhibitions about measuring their glucose in the presence of strangers, which was consistent with an increased frequency of sensor scanning. Also, in addition to the cost savings associated with reduced hospitalizations, there were fewer absences from work (7.8 vs. 18.55%, P <0.001).

While findings from these real-world studies further support the metabolic benefits associated with flash CGM use reported in RCTs, the reductions in hospitalizations (23–25,33) and in absenteeism (23,24) also demonstrate the immediate and substantial economic benefits of flash CGM use within populations with type 1 or type 2 diabetes. The improvements observed in treatment satisfaction (23,29), levels of hypoglycemia fear (26,29), sense of well-being (29), and other health-related measures (23–25,28,29) additionally support patient-reported outcomes described in RCTs (1,5).

As use of flash CGM technology continues to grow, large national and commercial database studies are being investigated to discern the impact of flash CGM on both A1C and acute diabetes-related events (Table 3). Two recent analyses showed significant reductions in all-cause hospitalizations and diabetes-related events among adults with either type 1 or type 2 diabetes who acquired flash CGM.

Analysis of a French national reimbursement claims database identified 74,158 adults with diabetes (type 1: n = 33,203, type 2: n = 40,955) who initiated flash CGM during the last 6 months of 2017 (34). Over the next 12 months, yearly hospitalization rates for DKA and acute hyperglycemia were reduced by 52% among patients with type 1 diabetes and by 47% for those with type 2 diabetes. The reduced rates were most evident for people with very low or very high adherence to self-monitoring of blood glucose (SMBG).

Significant reductions in acute diabetes-related adverse events (ADEs) and all-cause hospitalizations (ACH) were noted among 1,244 adults with type 2 diabetes treated with rapid- or short-acting insulin who acquired flash CGM (35). At 6 months post-acquisition, ADE rates decreased from 0.158 to 0.077 events/patient-year (hazard ratio [HR] 0.49 [95% CI 0.34–0.69], P <0.001). Hospitalizations also decreased from 0.345 to 0.247 events/patient-year (HR 0.72 [95% CI 0.58–0.88], P = 0.002). These findings equate to numbers needed to treat of 12 and 10 for 1 year to avoid one ADE or 1 ACH, respectively.

Strong evidence for the clinical and economic benefits associated with use of flash CGM has also emerged from studies of individuals treated less intensively with insulin or noninsulin therapy (30,31,33). In addition to significant reductions in ADEs and hospitalization rates (33), there were substantial and sustained reductions in A1C among adults with type 2 diabetes treated with long-acting insulin or noninsulin therapies (30). For those not treated with insulin, the reductions in A1C were similar to what would be expected from adding insulin glargine (36). Further studies (34,37) have established that there is no correlation between previous frequency of daily blood glucose monitoring and ADEs.

These findings challenge the view that CGM should be made available only to patients who are treated with intensive insulin therapy and who have a documented history of frequent blood glucose monitoring. This perception may have reduced the coverage of CGM offered by some commercial and public insurers. For example, the Centers for Medicare & Medicaid Services currently limit coverage to patients who administer three or more insulin injections per day (or use an insulin pump) and perform four or more glucose tests per day.

Many individuals with diabetes experience poor glycemic control (38,39), which puts them at increased risk for acute adverse glycemic events (40,41) and the long-term development of microvascular and macrovascular disease (42–44). In addition to its clinical consequences, uncontrolled diabetes is driving an inordinate economic burden on private payers and national health care systems (7).

Numerous studies have shown that optimizing A1C levels and reducing the incidence of severe hypoglycemia and DKA can significantly lower health care costs (10–14,45–47). However, achieving optimal diabetes control necessitates expanded adoption of diabetes medications and technologies that are effective, safe, and feasible in real-world clinical settings.

Flash CGM provides immediate information about an individual’s current and projected glucose level using rate-of-change arrows, which allows users to respond promptly to mitigate or prevent pending hypoglycemia or hyperglycemia. Findings from large RCTs and prospective, observational studies have shown that use of flash CGM is associated with improved overall glycemic control (23–26,28), reductions in hypoglycemia (23–25), fewer diabetes-related hospitalizations (23–25,33), decreased absenteeism (23,24), and improvements in treatment satisfaction (23,29) and measures of well-being (23,25,28,29). These outcomes indicate both clinical and economic benefits, and use of flash CGM can enable these outcomes regardless of therapy and previous blood glucose monitoring frequency (30,31,33,34,37).

Given the growing global prevalence of diabetes and its associated costs, there is an opportunity to take advantage of flash CGM to facilitate improvements in metabolic control and patient QoL while reducing the projected costs of diabetes care.

The authors thank Christopher G. Parkin, CGParkin Communications, Inc., for his thoughtful assistance in developing this manuscript.

Funding for the development of this manuscript was provided by Abbott Diabetes Care.

C.J.B. has served on advisory boards for Abbott Diabetes Care, Boehringer Ingelheim, Lexicon, Novo Nordisk, and Sanofi. J.R.G. has served on advisory boards and/or speaker bureaus for Abbott Diabetes Care.

C.J.B. and J.R.G. wrote, reviewed, and approved the manuscript for submission. C.J.B. is the guarantor of this work and takes responsibility for the integrity of the data and the accuracy of the content.