ADAG is not a standalone blood test you can order at a lab. It stands for “A1C-Derived Average Glucose,” the name of a landmark international study that created the formula doctors now use to translate your HbA1c percentage into an estimated average glucose (eAG) value in the same everyday units you see on a home glucose meter. The idea is straightforward: instead of hearing “your A1C is 7 percent” and having to guess what that means for daily blood sugar, you hear “your estimated average glucose is about 154 mg/dL,” a number that connects directly to the readings you already track. But the conversion rests on assumptions about how glucose sticks to hemoglobin, and those assumptions do not hold equally for everyone.
What the ADAG Study Actually Did
The ADAG study, published in 2008 in Diabetes Care, enrolled 507 people across ten international centers: 268 with type 1 diabetes, 159 with type 2, and 80 without diabetes. Over three months, each participant collected roughly 2,700 glucose measurements through a combination of continuous glucose monitoring and traditional fingerstick checks. Researchers then compared those real-world averages to the A1C drawn at the end of the period and ran the numbers through regression analysis. The result was a tight linear relationship, with about 84 percent of the variation in average glucose explained by A1C alone. The equation that emerged, eAG (mg/dL) = 28.7 × A1C − 46.7, is the one labs and diabetes apps still use today.1PubMed Central. Translating the A1C assay into estimated average glucose values
One of the study’s strengths was that the relationship held across subgroups. Age, sex, diabetes type, race and ethnicity, and smoking status did not significantly alter the regression line.2PubMed Central. Translating the A1C assay into estimated average glucose values That broad consistency is why the formula was adopted widely. But “did not differ significantly” in a study of 507 people is not the same as “works perfectly for every individual.” The 16 percent of variation not captured by the equation has to come from somewhere, and for some groups of people, that gap matters clinically.
Why Hemoglobin Captures a Three-Month Window
The reason A1C reflects long-term glucose at all is a slow chemical reaction. Glucose in your bloodstream attaches nonenzymatically to the amino-terminal end of hemoglobin’s beta chain, forming a compound called hemoglobin A1c. This attachment happens continuously and irreversibly over the roughly 120-day lifespan of a red blood cell.3PubMed. The glycosylation of hemoglobin: relevance to diabetes mellitus Higher average blood sugar means more glucose molecules bump into hemoglobin, so a higher fraction ends up glycated. When a lab measures A1C, it is essentially reading a chemical diary of your glucose levels over the previous two to three months, weighted toward the most recent weeks because younger red cells make up more of the circulating pool.
This passive tagging mechanism is what makes A1C so useful: you cannot game it the way you might game a single fasting glucose draw by eating carefully the night before. But it also means anything that changes how long red blood cells live, how readily hemoglobin gets glycated, or how much hemoglobin is available will throw off the reading and, by extension, the eAG derived from it.
What Your eAG Number Means in Practice
When a lab report shows an eAG alongside your A1C, the intent is to hand you a number you can compare against your meter or continuous glucose monitor readings. An A1C of 6 percent corresponds to an eAG of about 126 mg/dL. At 7 percent, it is roughly 154 mg/dL. At 8 percent, around 183 mg/dL. The conversion is linear, so each one-percentage-point rise in A1C adds about 29 mg/dL to the estimated average.
Worldwide, there is consensus that HbA1c should be reported in both NGSP (percent) and IFCC (mmol/mol) units, along with eAG in either mg/dL or mmol/L.4NGSP. IFCC Standardization of HbA1c In practice, many labs in the United States still report only the percent, and eAG appears as an optional add-on line. If your report does not include it, you can calculate it yourself from the equation above, or look it up on conversion charts from the American Diabetes Association.
Does Seeing eAG Help People Manage Diabetes Better?
The original hope was that expressing results in familiar glucose units would improve understanding and motivation. A randomized trial compared patients counseled with the term “estimated average glucose” against patients counseled with the standard A1C percentage. Both groups showed meaningful improvements in knowledge and understanding at follow-up, with composite survey scores improving by about a third in each group. But the eAG group did not improve significantly more than the A1C group.5PubMed. A randomized comparison of the terms estimated average glucose versus hemoglobin A1C In other words, education mattered, but the specific format of the number did not seem to be the decisive factor.
A separate analysis found that when labs began routinely reporting eAG alongside A1C, there were modest population-level shifts in both glycemic and cholesterol control, suggesting the additional context might nudge clinical decision-making even if individual patients do not perceive a dramatic difference.6PubMed Central. Translating the HbA1c assay into estimated average glucose values in children and adolescents with type 1 diabetes mellitus The practical upshot is that eAG is a communication tool, not a diagnostic test in its own right. It does not provide new biological information; it repackages the A1C in more intuitive language.
Racial and Ethnic Differences in Glycation
This is where the neat conversion starts to creak. Several studies have found that for the same measured blood glucose concentration, A1C levels tend to be higher in Black individuals than in white individuals. One analysis concluded that HbA1c overestimates mean glucose in Black persons compared with white persons, possibly because of racial differences in the rate at which hemoglobin becomes glycated.7PubMed. Racial Differences in the Relationship of Glucose Concentrations and Hemoglobin A1c Levels If A1C runs higher than the actual glucose average, then the eAG derived from it will also run higher, potentially leading to over-treatment or unnecessary diagnostic labels.
The reasons remain under investigation. Candidate explanations include differences in red blood cell lifespan, in the balance of glucose between the inside and outside of cells, and in nonglycemic genetic factors that influence how readily hemoglobin picks up glucose molecules.8PubMed Central. Racial and ethnic differences in the relationship between HbA1c and blood glucose: implications for the diagnosis of diabetes None of these has been pinpointed as the single driver, and the size of the discrepancy varies between studies. What is consistent is the direction: the ADAG equation, derived from a mixed population, may systematically overstate glucose exposure for some racial groups and understate it for others.
This has real consequences. If a clinician uses A1C alone to diagnose prediabetes or diabetes, and the patient’s biology produces a higher A1C at the same glucose level, the threshold will be crossed sooner. The reverse is also true: conditions that lower A1C can mask genuine hyperglycemia. For anyone in a population where these discrepancies have been documented, it is worth confirming A1C-based assessments with direct glucose measurements when something does not add up.
Sickle Cell Trait and False Reassurance
Sickle cell trait, carried by roughly 8 percent of Black Americans, provides one of the starkest examples of how the ADAG formula can mislead. A large study found that for a given fasting glucose, people with sickle cell trait had HbA1c values about 0.3 percentage points lower than those without the trait. That gap widened at higher glucose concentrations. The downstream effect was dramatic: the prevalence of prediabetes diagnosed by A1C was about 29 percent in participants with sickle cell trait compared with about 49 percent in those without it, even when actual glucose levels were similar.9JAMA. Association of Sickle Cell Trait With Hemoglobin A1c in African Americans
Because sickle cell trait alters the structure and turnover of hemoglobin, the chemical tagging process that A1C measures does not proceed normally. The result is an A1C that looks reassuringly low while blood sugar may be running dangerously high. Any eAG calculated from that A1C will inherit the same false reassurance. Diabetes guidelines increasingly recommend using fructosamine, glycated albumin, or direct glucose testing when sickle cell trait or other hemoglobin variants are present.
Pregnancy Throws Off the Numbers
Pregnant women present a different challenge. Blood volume expands, red cell turnover speeds up, and hemoglobin concentration drops, all of which change how much glycation accumulates. Research on pregnant women with diabetes found that the standard ADAG equation substantially overestimated average glucose. At an A1C of 8.0 percent, the standard formula predicts an average glucose around 10.2 mmol/L, but the pregnancy-specific estimate landed between 7.4 and 7.7 mmol/L depending on gestational week. Even at an A1C of 6.0 percent, the pregnancy-adjusted average glucose was only 6.4 to 6.7 mmol/L instead of the roughly 7.0 mmol/L the standard formula would give.10PubMed Central. Translating HbA(1c) measurements into estimated average glucose values in pregnant women with diabetes
A separate study found that anemia during pregnancy, which is common, was associated with significantly higher HbA1c and eAG values even after accounting for actual glucose levels.11International Journal of Pathology. Determining how the gestational changes, including low hemoglobin influence the HbA1c and estimated average glucose relationship That means pregnant women with low hemoglobin might look worse on paper than their glucose control actually warrants, while those with faster red cell turnover might look better. Neither direction is harmless when tight glucose targets are needed to protect both mother and baby. For this reason, many obstetric guidelines rely more heavily on fingerstick logs or continuous glucose monitoring than on A1C during pregnancy.
Chronic Kidney Disease and ESA Use
Kidney disease introduces its own distortions. As kidney function declines, red blood cell lifespan shortens, which should theoretically lower A1C by giving hemoglobin less time to accumulate glucose. On top of that, many patients with chronic kidney disease receive erythropoiesis-stimulating agents (ESAs) to treat anemia. These drugs flood the circulation with young red cells that have had minimal time in a high-glucose environment, pulling measured A1C down even further. A study of patients with type 2 diabetes and chronic kidney disease found that ESA use was the primary driver of the weakened relationship between A1C and actual average glucose, causing a systematic underestimation of glucose exposure from A1C.12PubMed. Defining the relationship between average glucose and HbA1c in patients with type 2 diabetes and chronic kidney disease
If you are on dialysis or receiving ESAs, your eAG may paint an optimistic picture that does not match the glucose swings your body is actually experiencing. Clinicians managing diabetes in kidney disease often supplement A1C with glycated albumin or fructosamine, both of which reflect shorter windows of glycemic control and are not affected by red cell turnover.
Aging and a Slowly Drifting Baseline
Even in people without diabetes, A1C rises with age. Data from the Framingham Offspring Study and the National Health and Nutrition Examination Survey showed that each additional year of age was associated with roughly a 0.01-percentage-point increase in A1C, independent of glucose tolerance status.13PubMed Central. Effect of Aging on A1C Levels in Individuals Without Diabetes That sounds trivial in a single year, but over two or three decades it adds up. A healthy 70-year-old may have an A1C that sits a third of a percentage point higher than a healthy 40-year-old at identical glucose levels.
The clinical implication is subtle but real. Using fixed A1C thresholds to screen older adults for diabetes may overcapture people whose hemoglobin glycation has simply drifted upward with age. Whether this drift represents a true physiological change in glycation rate, a slight shortening of red cell lifespan, or something else entirely is still debated. For practical purposes, the takeaway is that an eAG of, say, 135 mg/dL in a 75-year-old does not carry the same significance as the same number in a 35-year-old.
Children and Adolescents
The ADAG study enrolled adults. Whether the same formula applies to children was an open question for years. A pediatric study that compared A1C with continuous glucose monitoring averages in children and adolescents with type 1 diabetes found a tight correlation, with about 90 percent of the variation in average glucose explained by A1C.14PubMed Central. Translating the HbA1c assay into estimated average glucose values in children and adolescents with type 1 diabetes mellitus A separate analysis concluded that the adult-derived ADAG formula could be used in children, though with some caution because the distribution of A1C values in pediatric populations tends to be skewed differently than in adults.15PubMed. Paediatric estimated average glucose in children with Type 1 diabetes In practice, most pediatric endocrinologists use the same conversion, sometimes adjusting their interpretation when a child’s A1C sits at the extremes.
How Continuous Glucose Monitors Fit In
The rise of continuous glucose monitors has given patients a direct measure of average glucose over weeks, bypassing hemoglobin entirely. The CGM-derived metric, often called glucose management indicator (GMI), is calculated from the sensor’s glucose readings over 14 days or more. It mirrors the concept behind eAG but uses real glucose data instead of an inference from hemoglobin glycation.
This has created a new kind of confusion. Patients often notice that their GMI and their lab-reported eAG do not match. A 2024 analysis in Diabetes Care explicitly used the ADAG regression equation to compute eAG from A1C and compared it against CGM-derived averages, documenting the sources of discrepancy between the two.16PubMed Central. Estimating Glycemia From HbA1c and CGM: Analysis of Accuracy and Sources of Discrepancy Part of the gap comes from biological variation in glycation rate, the same issue underlying racial and age-related differences. Part comes from the fact that CGM sensors measure interstitial glucose, not blood glucose, and have their own calibration drift. Neither number is wrong in an absolute sense; they are measuring related but not identical things. When the two diverge substantially, it is worth discussing with your doctor which better reflects your real-world glucose control and whether one of the interfering factors described above could be at play.
When A1C and eAG Are Not Enough
For patients where A1C is unreliable, several alternative markers exist. Fructosamine measures glycation of serum proteins, primarily albumin, and reflects average glucose over the previous two to three weeks rather than two to three months. Glycated albumin is similar but more specific, and it captures postprandial glucose spikes more sensitively than A1C does. A third marker, 1,5-anhydroglucitol, drops when glucose rises above the kidney’s reabsorption threshold, providing a real-time snapshot of recent hyperglycemic episodes.17PubMed Central. Alternative biomarkers for assessing glycemic control in diabetes: fructosamine, glycated albumin, and 1,5-anhydroglucitol
None of these alternatives has the decades of outcome data backing A1C. The big trials linking A1C to the risk of complications, the ones that gave us treatment targets of 7 percent for most adults, used A1C as their metric. Translating a glycated albumin value into the same kind of complication-risk estimate remains an active area of research. Still, for clinical scenarios where A1C is known to be misleading, such as hemoglobin variants, dialysis, recent blood transfusion, or pregnancy, these shorter-window markers provide a genuinely independent check on glucose control that the ADAG formula simply cannot offer.
Common Misconceptions Worth Clearing Up
Perhaps the most persistent misunderstanding is that eAG tells you what your blood sugar “should” read on any given day. It does not. It is an average across thousands of readings over months, and averages hide enormous variation. Two people with the same eAG of 154 mg/dL might have very different daily profiles: one might hover steadily between 130 and 180, while the other swings wildly from 60 to 300. The average is the same, but the second person’s variability carries higher risk for complications, something eAG by itself cannot reveal.
Another misconception is that the ADAG formula “corrects” A1C for individual differences. It does not. It simply converts the percentage to a different unit using a population-level regression line. All the biological variability that can make A1C inaccurate for a particular person travels straight through the equation and into the eAG number. If your A1C is misleadingly low because of sickle cell trait, your eAG will be misleadingly low by exactly the same proportion. The formula translates; it does not validate.
Finally, some patients assume that if their eAG matches their meter average reasonably well, their A1C must be accurate for them. This is closer to true, but not guaranteed: home meters measure capillary glucose at selected times of day, often missing overnight lows and post-meal peaks. A CGM average is a better comparator, and even then, the biological sources of discrepancy described above mean that a close match could be coincidental rather than confirmatory. The safest approach is to treat eAG as one piece of a larger picture that includes direct glucose measurements, time-in-range data from a CGM if available, and clinical context about any conditions that could skew hemoglobin-based metrics.