Mean Blood Glucose to A1C: ADAG Formula, GMI, and Why They Disagree (2026)

Mean blood glucose to A1C: the ADAG formula, GMI equation, reference table from 70 to 300 mg/dL, and why your CGM average and lab A1C sometimes don't match.

Your CGM device reported a 90-day average glucose of 148 mg/dL. The ADAG formula suggests an A1C of 6.8%. Your lab result came back at 7.4%. One of those numbers is wrong, and it might not be either one of them.

Mean blood glucose and A1C describe the same underlying variable from two different calculation approaches. The gap between them tells you something real about your physiology, not just measurement error to discard. The A1C Calculator converts any mean glucose to an estimated A1C using both the ADAG formula and GMI, so you can see the difference directly. This guide explains what each formula is, how to read the reference table, and what to do when the numbers disagree.

What Mean Blood Glucose Actually Measures

Mean blood glucose is the arithmetic average of all glucose readings over a period, typically 14, 30, or 90 days. Every reading carries equal weight regardless of when it was taken.

A 90-day mean is the standard input for A1C estimation because red blood cells live approximately 90 days. Glucose binds to hemoglobin throughout that lifespan. A higher average glucose over those 90 days produces more glycated hemoglobin, which is what the A1C lab test directly measures.

Two important things your mean glucose does not capture: time in range and glucose variability. A person with an average of 140 mg/dL could be spending most of their time in the 100-160 range, or they could be oscillating between 60 and 220 with the same mathematical average. The mean is identical in both cases. The clinical risk profile is not.

This is why mean glucose to A1C estimation should be read alongside time in range data from your CGM, not as a standalone number.

The ADAG Formula: How Mean Glucose Converts to A1C

The ADAG (A1c-Derived Average Glucose) formula was published in the American Diabetes Association's journal Diabetes Care in 2008, based on a multicenter study of 507 participants. It established the standard linear relationship between laboratory-measured A1C and mean plasma glucose from CGM and frequent fingerstick measurements combined.

A1C (%) = (Mean Glucose mg/dL + 46.7) / 28.7

Reverse formula (A1C to estimated average glucose):

eAG (mg/dL) = (28.7 × A1C) - 46.7

Worked example at 148 mg/dL mean glucose:

• A1C = (148 + 46.7) / 28.7
• A1C = 194.7 / 28.7
• A1C = 6.78%, rounds to 6.8%

Worked example in reverse at 7.0% A1C:

• eAG = (28.7 × 7.0) - 46.7
• eAG = 200.9 - 46.7
• eAG = 154 mg/dL

The reference table below shows ADAG-estimated A1C at common mean glucose values. Each row also shows the equivalent GMI, which uses a different coefficient set (explained in the next section).

The ADAG formula is built into the A1C Calculator, which accepts inputs in both mg/dL and mmol/L and produces estimated A1C, eAG, and GMI simultaneously.

ADAG reference table: mean blood glucose to estimated A1C

The following values are calculated using the ADAG formula for single-filer glucose averages:

GMI vs ADAG: The Same Input, Two Different Outputs

The Glucose Management Indicator (GMI) was published in 2018 specifically to address a limitation of the ADAG formula: the ADAG study used a mix of CGM readings and fingerstick data. Modern CGMs produce continuous, sensor-derived glucose readings that have different statistical properties from periodic fingerstick measurements.

The GMI formula uses different coefficients calibrated to CGM-derived glucose data:

GMI (%) = 3.31 + (0.02392 × Mean Glucose mg/dL)

Worked example at 148 mg/dL mean glucose:

• GMI = 3.31 + (0.02392 × 148)
• GMI = 3.31 + 3.540
• GMI = 6.85%

Compared to ADAG at the same input: 6.78%. The difference is small at 148 mg/dL. It grows at higher glucose levels.

ADAG vs GMI comparison at key mean glucose values:

At approximately 154 mg/dL, the two formulas produce the same result. Below that level, GMI tends to run slightly higher than ADAG. Above it, ADAG runs higher than GMI, and the gap widens at elevated glucose levels.

Which one should you use? If your data comes from a modern CGM device, GMI is the more appropriate formula because it was derived from CGM data specifically. If you are working from fingerstick averages, ADAG is the relevant calculation. If your device reports both, neither is wrong. They are calibrated to different measurement sources.

The GMI Calculator covers the GMI formula in depth, including the difference in clinical interpretation between GMI and standard A1C thresholds for diabetes diagnosis and management.

Why Your CGM Average and Lab A1C Sometimes Don't Match

This is where the real variation occurs, and why the formulas are estimates rather than exact conversions.

Hemoglobin variants. The A1C lab test measures glycated hemoglobin as a percentage of total hemoglobin. In patients with hemoglobin variants (HbS, HbC, HbE), some assay methods produce falsely low or falsely high A1C results. Mean glucose from CGM is unaffected by hemoglobin variants. A patient with sickle cell trait may have a lower measured A1C than their average glucose would predict, not because their glucose control is better but because the assay is measuring differently.

Red blood cell lifespan. The standard A1C assumes red blood cells live approximately 90 days. Conditions that shorten red blood cell lifespan (hemolytic anemia, blood loss, recent transfusion, end-stage kidney disease with erythropoietin treatment) reduce A1C below what average glucose would predict. The blood cells are replaced before they accumulate as much glycated hemoglobin. Conversely, conditions that extend red blood cell lifespan (iron deficiency anemia, splenectomy) elevate A1C above the glucose-based prediction.

The 90-day vs recent glucose weighting. A1C is not a true 90-day average. Glucose from the most recent 30 days contributes approximately 50% of the A1C result, while glucose from days 60-90 back contributes proportionally less. If your glucose control improved significantly in the last month, your A1C will reflect that more heavily than a straight 90-day mean would suggest.

CGM sensor calibration. Most modern CGMs measure interstitial glucose, not plasma glucose. Interstitial glucose lags behind plasma glucose by 5 to 15 minutes. During rapid glucose swings, this lag introduces a systematic gap. Calibration accuracy also varies by sensor generation and placement site. Lab glucose from a venous plasma draw is the reference measurement; CGM-derived values are always estimates of that reference.

Which Number to Trust When They Don't Match

If your estimated A1C from mean glucose and your lab A1C disagree by more than 0.5%, the gap is worth understanding rather than averaging away.

Trust the lab A1C for: diabetes diagnosis (lab A1C is the clinical diagnostic criterion), medication dosing decisions where threshold values matter (like starting insulin at 9%), and any insurance or regulatory documentation.

Trust the CGM-derived GMI for: understanding how your glucose is tracking between lab draws, identifying improvement or deterioration that happened too recently to fully show in A1C, and in populations where hemoglobin variants make A1C interpretation unreliable.

When your lab A1C is consistently higher than your CGM estimate: Consider whether you have iron deficiency anemia, which is the most common cause of A1C elevation beyond what average glucose explains. A ferritin level will either confirm or rule this out quickly.

When your lab A1C is consistently lower than your CGM estimate: Hemolytic conditions and hemoglobin variants are the main causes. If you are of African, Mediterranean, or Southeast Asian descent, asking your provider about hemoglobin variant testing is reasonable.

For most people without the conditions above, the ADAG formula and measured A1C will agree within 0.5%. A gap larger than that is worth one conversation with your provider before attributing it to normal variation.

The Blood Glucose to A1C Calculator walks through the full calculation with a reference table and conversion chart for both mg/dL and mmol/L inputs.

If you are monitoring multiple metabolic lab values together, the Corrected Calcium Calculator handles a related calibration challenge that appears on the same standard metabolic panel: adjusting a total lab value to reflect the biologically active fraction when protein levels are abnormal.

The Number That Gets Closest to the Truth

No single number captures everything about glucose control. Mean glucose tells you where the average landed. A1C tells you what the biology recorded over the previous three months. GMI gives you a CGM-calibrated estimate of where A1C is likely to sit. Time in range tells you about variability that the mean obscures.

For clinical decisions, lab A1C remains the standard. For tracking progress between lab draws and understanding the direction your glucose is moving, mean glucose from CGM and the GMI calculation are the most responsive measures. Run your mean glucose through the A1C Calculator to see the ADAG estimate, the GMI, and the eAG in one calculation. Then compare against your next lab draw to see how closely your specific physiology tracks the population average the formulas were built on.