What are the potential causes for significant discrepancies between predicted A1c by Dexcom CGM and standard serum A1c by lab draw?

When interpreting A1C and CGM results, it is important to remember that neither measure is perfect. As in all clinical conditions, it is important to evaluate A1C and CGM results in the context of the clinical situation. If there is a disparity between the two measures, one should review the limitations of each and determine whether the patient has a condition that may be interfering. There are numerous limitations/considerations for both measures! A few are summarized here:

We have known for a long time that A1C potentially can be falsely high or low in a very long list of conditions. Common conditions include anemia, hemoglobinopathies, opioid use, and chronic alcoholism. Conditions with low red cell turnover (e.g., vitamin B12 or folate deficiency) may lead to a disproportionate number of older red cells and a falsely high A1C. Rapid red cell turnover (e.g., hemolysis or treatment with erythropoietin) leads to a greater proportion of younger red cells and falsely low A1C values. Interestingly, several studies have shown that some ethnic groups (e.g., African American, Hispanic American, Asian American) have slightly (~0.3-0.4 points) higher A1C values than Non-Hispanic White patients with similar glucose concentrations.

CGM device measures interstitial fluid glucose. Due to the delayed equilibration between the different physiologic compartments, CGM glucose values may be lower than venous plasma glucose when blood glucose concentrations are rapidly rising. CGM is also less accurate in the extreme low or high. Medications including acetaminophen, hydroxyurea, and high-dose vitamin C may cause false elevations in CGM glucose values with some, but not all, CGM devices. Another source for error with CGM occurs if the sensor is not worn for most of the time. That results in missing data. For sensors that need to be scanned, if the sensor is not scanned at least every 8 hours, data from the previous scans may be lost. So it is important to review reports and see what percentage of the time the sensors are being used and for what period of time is reflected in the report. A1C measurement reflects the blood glucose values for the past few months and CGM reports typically only report the last couple of weeks.