Key takeaways:

• Time in range and coefficient of variation are the two most important CGM metrics for assessing safety and glycemic outcomes.
• Health care professionals should not compare metrics between different CGM systems.

PHILADELPHIA — Health care professionals should use the full spectrum of continuous glucose monitoring metrics to assess complication risk for people with diabetes, according to a speaker at the Heart in Diabetes CME Conference.

George Grunberger, MD, FACP, MACE, chairman at Grunberger Diabetes & Endocrinology in Bloomfield Hills, Michigan, discussed four key CGM metrics: mean glucose, coefficient of variation, glucose management indicator and time in range. Grunberger said each of the metrics not only provides different information on glycemic control, but may also provide information on a person’s risk for different cardio-renal-metabolic conditions. Grunberger presented a framework for assessing CGM metrics.

“We have to always prioritize risk reduction over intensification,” Grunberger said during a presentation.

Metrics for safety and glycemic control

Grunberger said coefficient of variation, which assesses how variable glucose levels are within patients, is one of the most important metrics due to its link to hypoglycemia risk. In an international consensus paper published in Diabetes Care in 2019, the authors stated that patients should have a coefficient of variation of 36% or less.

“Clinically, coefficient of variation is important, especially for worry about hyperglycemia and especially when people have a high risk, [such as] older people, pregnant women and people with hyperglycemia unawareness,” Grunberger said.

Additionally, Grunberger said, high coefficient of variation has been more strongly linked to early risk for atherosclerosis than both fasting glucose and HbA1c.

Time in range, which measures the percentage of time a person has a glucose level between 70 mg/dL and 180 mg/dL, is the best metric for determining glycemic control, Grunberger said. Target time in range should be 70% or higher for most people with type 1 or type 2 diabetes, though Grunberger noted target ranges and percent time in range differ for older adults, pregnant women with type 1 diabetes and those with gestational diabetes.

Study data have linked time in range to several diabetes complications. A 2025 study published in Diabetologia found incidence of diabetic retinopathy increased by 18% for each 5 percentage point decrease in time in range for people with type 2 diabetes. A reanalysis of the Diabetes Control and Complications Trial published in Diabetes Care in 2018 found that each 10 percentage point decline in time in range increased microalbuminuria adverse outcome risk by 40%. Multiple studies have also linked time in range and adverse cardiovascular outcomes, with decreased time in range associated with higher risk for all-cause and CVD mortality.

Grunberger said some people in the diabetes field have advocated for the FDA to allow time in range to be used as a primary efficacy endpoint in drug trials.

“Half of the randomized controlled trials in type 2 diabetes lately have used time in range as primary endpoint,” Grunberger said. “On average it takes about 13 days, so less than 2 weeks, to see an effective treatment. Clearly, it’s a lot faster than you would do with HbA1c.”

Grunberger said time below range must also be reported with time in range due to the risk for complications associated with hypoglycemia, and close attention must be paid to any patient with a time below range higher than 4%.

When examining both metrics, time in range is the best for predicting outcomes, Grunberger said, while coefficient of variation can identify hypoglycemia danger that time in range may miss.

Additional CGM metrics

Grunberger also highlighted the utility of mean glucose and glucose management indicator. He said health care professionals should target a mean glucose of 154 mg/dL for people with diabetes. The metric calculates the average glucose for the entire CGM wear period for a patient.

Grunberger said health care professionals should never compare mean glucose between different CGM systems, as there can be up to a 14.2% difference between brands.

“Using a different device in the same patient might be dangerous,” Grunberger said.

Glucose management indicator is an estimated HbA1c figure calculated by CGM. Grunberger said most adults should have a target glucose management indicator of less than 7%.

In 2025, a paper published in Diabetologia outlined an updated formula for calculating glucose management indicator. Grunberger said the updated formula was needed to correct for discordance between the metric and HbA1c as well as to improved adjustments for race and ethnicity.

The updated glucose management indicator formula is linked to a diabetes complication. Data published in Diabetologia in 2025 found the updated glucose managing indicator was more closely correlated with incident retinopathy in people with type 1 diabetes than HbA1c.

“The latest [ADA] Standards of Care actually list glucose management indicator alongside with HbA1c and time in range as primary glycemic assessment tools,” Grunberger said. “It’s coming of age slowly, but surely.”

When glucose management indicator and HbA1c differ by 0.5 percentage points or more, it may be due to nonglycemic reasons and may indicate chronic kidney disease, iron deficiency or hemoglobinopathy, Grunberger said.