Are continuous glucose monitors overrated and overused?

The popularity of continuous glucose monitoring (CGM) is surging in the U.S., but whether patients are being helped is unclear. Not only is CGM currently being used by 2 million patients with type 2 diabetes, but start-ups promote them as biofeedback "wellness tools" for people without diabetes even though there is no evidence that CGM improves dietary choices, weight loss, or other outcomes in persons with obesity or prediabetes. Arguing in an American Family Physician editorial a year ago that "Continuous Glucose Monitoring in Type 2 Diabetes Is Not Ready for Widespread Adoption," Dr. Sandy Robertson and colleagues pointed out that no long-term studies had demonstrated that CGM improved patient-oriented outcomes in this population compared to finger-stick or no self-glucose monitoring, and that "unnecessary monitoring not only wastes money but can negatively impact quality of life." I've also written on this blog before that patients with type 2 diabetes who are not using insulin do not benefit from self-monitoring.

In this context, two articles in the June 1 issue of AFP emphasized the relatively narrow evidence-based indications for considering CGM. A POEM (Patient-Oriented Evidence That Matters) reported on a systematic review and meta-analysis that found little benefit except in "patients using intensive insulin therapy" who are insensitive to extreme hyperglycemia or hypoglycemia; and a Diagnostic Tests review of the FreeStyle Libre 14-Day "flash" CGM system reported "convenience, possible cost savings, and improvement in treatment satisfaction" for selected patients, but noted that "in adults with insulin-treated type 2 diabetes, there is conflicting evidence whether it reduces [hemoglobin] A1c levels and hypoglycemic time and events."

Two recent studies in JAMA shed additional light on the effects of CGM on glycemic control in patients with insulin-treated type 2 diabetes. In the first study, a multicenter randomized, controlled trial of 175 adults using basal-only insulin with a mean hemoglobin A1c level of 9.1%, participants assigned to the CGM group spent more time in the normal glucose range (70 to 180 mg/dL) and had a lower average hemoglobin A1c at 8 months of follow-up (8.0% vs. 8.4%) than participants in the traditional glucose monitoring group. Severe hypoglycemic events were rare in both groups. In the second study, a retrospective cohort of more than 40,000 patients with type 1 and insulin-treated type 2 diabetes at Kaiser Permanente California, those who initiated CGM "had significant improvements in hemoglobin A1c and reductions in emergency department visits and hospitalizations for hypoglycemia" compared to those who did not initiate CGM. Though the study authors attempted to control for the many baseline differences between the two groups (for example, CGM initiators were "more likely White and English-language speakers and less likely living in a deprived neighborhood"), unmeasured residual confounding may still have affected the results.

Financial conflicts of interest complicate interpretation of the evidence regarding the effectiveness of CGM. A "summary review of recent real-world evidence" on flash CGM published last year in the American Diabetes Association's Clinical Diabetes journal not only was funded by Abbott Diabetes Care (which makes the FreeStyle Libre system) and acknowledged the assistance of a medical communications company in writing the manuscript, but both authors also served on Abbott's advisory board. At AFP, such close ties to a manufacturer of CGM devices would disqualify authors from writing a review article on diabetes or related topics, as we recognize that disclosure is not enough to mitigate biases introduced by these conflicts.