A1C from Average Glucose

What the A1C-glucose relationship actually is

Hemoglobin A1c (HbA1c) measures the percentage of hemoglobin molecules in your red blood cells that have glucose chemically bound to them — a process called glycation. Because red blood cells live ~120 days and glycation is irreversible, A1C reflects average glucose exposure weighted across that lifespan. Higher average glucose = more glycation = higher A1C. The Nathan 2008 ADAG study (Nathan et al., Diabetes Care) enrolled 643 subjects and derived the formula: estimated average glucose (mg/dL) = 28.7 × A1C − 46.7. Solving for A1C gives the reverse calculation this tool performs.

CGM era — GMI vs lab A1C

Continuous glucose monitors (Dexcom G7, Libre 3, Medtronic Guardian) compute "GMI" — Glucose Management Indicator — using essentially the same formula but applied to 14+ days of CGM data rather than ~90 days of red blood cell history. The 2018 Bergenstal et al. consensus paper renamed the metric from "estimated A1C" to GMI specifically because individual A1C-GMI gaps were clinically meaningful: ~50% of T1D and T2D adults had gaps over 0.3%, and ~10% had gaps over 0.7%. Both numbers can be "right" — they reflect different averaging windows and individual glycation kinetics.

Why the formula isn\'t accurate for everyone

Several conditions shift the A1C-glucose relationship: (1) Iron-deficiency anemia and recent blood loss — increases A1C above what glucose predicts. (2) Hemolytic anemia, sickle cell disease, thalassemia — usually decreases A1C below predicted. (3) Recent transfusion — shifts depending on donor RBC age. (4) Pregnancy — A1C runs ~0.5% lower than glucose predicts due to increased RBC turnover. (5) Chronic kidney disease and uremia — generally shifts A1C upward. (6) Some ethnicities — Black, Hispanic, and South Asian populations average slightly higher A1C than non-Hispanic Whites at the same fasting glucose (Selvin 2011 Lancet meta-analysis).

The 14-day CGM standard

The 2019 International Consensus on Time in Range (Battelino et al., Diabetes Care) established 14 days of CGM data with ≥70% wear time as the minimum for reliable metrics. Below 14 days, single-day variations (illness, unusual meals, stress) skew the average meaningfully. Above 30 days, the additional data adds diminishing returns. For most clinical purposes, 14–30 days of consistent CGM data is the sweet spot for estimating what your next lab A1C will show.

A1C target ranges

• <5.7% (108 mg/dL avg): non-diabetic range
• 5.7–6.4% (115–137 mg/dL avg): prediabetes (ADA)
• ≥6.5% (140+ mg/dL avg): diabetes diagnostic threshold
• <7.0% (154 mg/dL avg): ADA standard treatment target for most adults with diabetes
• <6.5% (140 mg/dL avg): tighter target for younger, healthier adults without hypoglycemia risk
• <8.0% (183 mg/dL avg): less stringent target for older adults or those with comorbidities

Related tools

• A1C ↔ eAG Converter (the forward direction)
• Glucose Unit Converter (mg/dL ↔ mmol/L)
• HOMA-IR Insulin Resistance Calculator
• Glycemic Load Calculator

Frequently asked questions

Why does my CGM "GMI" differ from my lab A1C?

GMI (Glucose Management Indicator) is essentially the same calculation as estimated A1C from eAG. The difference between GMI and lab A1C is real and reflects two things: (1) individual variation in red blood cell glycation kinetics, and (2) the lab A1C reflects ~120 days of glucose history weighted toward the most recent 30 days, while GMI reflects only the period covered by your CGM data. The 2018 Bergenstal et al. consensus paper (Diabetes Care) defined GMI specifically because the lab A1C-CGM gap was clinically meaningful in many people — neither is "wrong," they're measuring slightly different things. A consistent gap of 0.3–0.5% between GMI and A1C is normal.

Is the Nathan formula accurate for everyone?

No. The 2008 Nathan A1C-Derived Average Glucose (ADAG) study enrolled 643 subjects with various forms of diabetes and pooled their data — but individual variation was substantial. About 15% of subjects had eAG values that differed by more than 15% from their measured A1C. Anemia, recent blood loss, hemoglobinopathies, pregnancy, and chronic kidney disease all shift the relationship. For most healthy and stable T1D/T2D adults, the formula is accurate within ±0.3%, which is clinically usable. For complex cases, lab A1C remains the standard.

How long does my CGM data need to cover for an accurate estimate?

Minimum 14 days of CGM wear time with at least 70% data capture, per the 2019 International Consensus on Time in Range guidelines. The 14-day window correlates ~0.84 with subsequent 90-day A1C in stable subjects. 30 days is better; 90 days approaches the full A1C measurement period. If you've recently changed treatment (new meds, major diet change, surgery), the CGM-based estimate will be ahead of the lagging lab A1C until ~6 weeks after the change stabilises.

Can I use this to skip lab A1C tests?

Not really, even with CGM. Lab A1C is the standardised diabetes diagnostic and treatment-monitoring measure that endocrinologists, insurance, and clinical trials use. CGM-derived GMI is increasingly accepted as supplementary (and ADA guidelines explicitly endorse it for in-between-visit tracking) but it doesn't replace the official measure. Use this calculator to predict what your next lab A1C will likely show — useful for catching changes early — but plan on getting the lab measurement on schedule with your care team.

My average glucose looks normal but my A1C estimate is borderline. Why?

Several possibilities. The relationship between average glucose and A1C is non-linear at the extremes — glycation kinetics differ between low-normal and elevated ranges. Time spent above target ranges contributes more to A1C than time-in-range glucose does. Spike behaviour matters: someone averaging 110 mg/dL with brief 250 mg/dL spikes vs steady 110 mg/dL throughout will have different A1Cs despite identical averages. Check your "Time in Range" and standard deviation — those metrics tell a fuller story than average alone.