Estimating A1C Levels from CGM Data

No need to increase A1C goals to avoid daily hypoglycemic events

with David Price MD

Young woman wearing glucose monitor smiling

Since the summer of 1976 when it was first shown to be related to blood glucose, A1C levels have been a dominant indicator of glycemic control in both type 1 and type 2 diabetes. According to the Mayo Clinic, A1C levels should be measured quarterly if the person with diabetes is on insulin. Though A1C reflects blood glucose levels, it is an indicator of the average levels over 3 months; therefore, it cannot inform a patient about the acute glycemic excursions that may happen daily.  

The use of continuous glucose monitors (CGM) can, in some cases, provide minute-to-minute information on interstitial glucose levels, which helps individuals plan and adjust within their daily lives. However, since CGMs provide data on interstitial glucose levels, there is not a point-by-point relationship with A1C levels measured from blood. How to bring CGM data and A1C levels into a mutually beneficial relationship is the question brought up by John Welsh MD, PhD (Dexcom, Inc, San Diego, CA) in a presentation titled: “Should Target Glucose Values Be Increased to Avoid Severe Hypoglycemia? Real-World Data Say ‘No.’ “

Dr. Welsh’s presentation took advantage of earlier work done by Bergenstal et al. who developed a regression equation to estimate A1C from CGM data. They called this value a glucose management indicator (GMI).  Two data sets were used, and a pooled regression equation was formed:

GMI (%) = 3.31 + 0.02392 (14 day mean CGM mg/dL)

While GMI may not match A1C values exactly, more than 80% of study participants had an absolute difference of less than 0.5% between GMI and A1C levels. Thus, GMI provides a readily available indicator of glycemic control and was recently included in the standardized glucose management metrics for clinical care.   

The ENDO2021 presentation by Dr. Welsh used GMI data to look at time below range in patients using CGM to help address the fear of hypoglycemia. This fear has caused some to suggest that target A1C levels should be increased to avoid hypoglycemia events. 

Using recommendations from the February 2019 Advanced Technologies & Treatments for Diabetes (ATTD) Congress which state that patients should strive for less than an hour a day (4% of their day) with moderate hypoglycemia (<70mg/dL) and no more than 15 minutes (1%) in a severe hypoglycemic state (54 mg/dL), the group looked at the percent of time spent in at these hypoglycemic levels as a function of the GMI readings. 

Their results suggest that only at the very lowest levels of GMI, <6.5% (which is to say, an estimated A1C level of <6.5%) were there increases in hypoglycemic events. Patients with higher GMIs (>8.5%) did not have any fewer hypoglycemic events than those within a more acceptable range.

chart mapping GMI and percentage of hypoglycemic events


Vice President of Medical Affairs at Dexcom, David Price MD, offered more about GMI, its practical use, and thoughts on raising the target A1C levels. Dr. Price said, “Loosening control by raising A1C or GMI may not diminish hypoglycemia or risk of severe hypoglycemic events and may result in excessive hyperglycemia.”

Regarding how GMI might be used in elderly patients he said, “For frail, older, high risk patients, the goal is to minimize hypoglycemia to <1% (15 minutes per day) and avoid hyperglycemia, with <50% of time at >180 mg/dl and less than 10% at >250 mg/dL. So, decisions should not be based on A1C or on GMI alone.” 

When asked about children, Dr. Price suggested the following, “GMI and time in range can be used as a motivation. GMI should not be used alone but should be used with other CGM metrics. If GMI is improved and it does not result from excess hypoglycemia, glycemic control is improved.”

The idea of using all the available CGM metrics seems like a good way for every individual to make specific and daily decisions about their glucose management. However, data from the ENDO2021 presentation by Welsh et al. does not support the notion that increased A1C or GMI levels track with fewer hypoglycemic events. 

Continue Reading:
Clinical Takeaways: Microvascular Risk, Metabolic Disease, Continuous Glucose Monitoring
close X