HbA1c

HbA1c, or glycated hemoglobin, is the percentage of circulating hemoglobin that has glucose irreversibly bound to it, and it serves as the standard laboratory index of average blood glucose over the preceding two to three months. It is one of the most frequently ordered tests in medicine, used both to diagnose diabetes and prediabetes and to monitor the adequacy of glycemic control over time. Its value for longevity-oriented medicine comes from the continuous, graded relationship between glucose exposure and the diseases that shorten healthy lifespan, including coronary disease, kidney failure, retinopathy, and dementia. The headline statistic is that each 1 percentage point of HbA1c is associated with roughly a fifth more diabetes-related mortality in observational diabetes cohorts, and the gradient extends below the diabetic range into values that laboratories still label normal. Unlike a fasting glucose, HbA1c requires no fasting and is far less sensitive to the timing of the last meal, which is what made it practical as both a screening and a monitoring tool. The marker is best understood as a slow-moving running average rather than a real-time signal.

The biochemistry is straightforward and underpins both the strengths and the pitfalls of the test. Glucose diffuses freely into red cells and binds the N-terminal valine of the hemoglobin β-chain without an enzyme, first forming a labile Schiff base and then, through Amadori rearrangement, a stable ketoamine that lasts the lifetime of the cell. Because the reaction is concentration-driven and effectively irreversible, the accumulated fraction of glycated hemoglobin integrates the glucose a person has been exposed to over the lifespan of their red cells. That lifespan, normally 90 to 120 days, is therefore the second determinant of the result, and any condition that shortens or lengthens it moves HbA1c independent of true glycemia. Iron deficiency, hemolysis, transfusion, pregnancy, and inherited hemoglobin variants are the practically important examples. This dual dependence, on glucose and on red cell turnover, is the single most important concept for interpreting a discordant result.

The evidence linking HbA1c to outcomes spans landmark interventional trials and large prospective cohorts. The Diabetes Control and Complications Trial in type 1 diabetes and the United Kingdom Prospective Diabetes Study in type 2 diabetes established that lowering HbA1c reduces microvascular complications, with DCCT cutting retinopathy progression by 76 percent when HbA1c fell from about 9 to 7 percent. For people without diabetes, the ARIC study of 11,092 adults showed HbA1c outperformed fasting glucose for predicting coronary disease and death, and EPIC-Norfolk showed a continuous mortality gradient with the lowest risk below 5.0 percent. The picture is more nuanced for macrovascular disease and aggressive treatment: the ACCORD trial, which targeted an HbA1c below 6.0 percent in high-risk type 2 diabetes, was stopped early after the intensive arm showed 22 percent higher mortality, demonstrating that the observational association is not a simple instruction to drive the number as low as possible by any means. The lesson is that glycemic exposure causes small-vessel disease, but the macrovascular association is partly shared with insulin resistance and other drivers, and the method and population of lowering matter.

Interpretation in practice rests on three ideas. First, the optimal range is tighter than the laboratory range: the lowest long-term risk sits near 5.0 to 5.4 percent, below the 5.7 percent prediabetes threshold, so a normal label is not the same as an optimal value. Second, the result is only as trustworthy as the red cell biology behind it, so a value that disagrees with fingerstick or continuous glucose data should prompt evaluation for iron deficiency, hemoglobin variants, hemolysis, or kidney disease before the number is acted upon. Third, the measurement window dictates cadence: because HbA1c takes about 3 months to fully reflect a change, retesting more often than every 3 months rarely adds information outside of pregnancy or rapidly changing treatment. Complementary markers fill the gaps that HbA1c leaves, including fasting glucose and fasting insulin for early insulin resistance, fructosamine or glycated albumin when red cell turnover is abnormal, and continuous glucose monitoring for variability and time in range that an average cannot capture. Standardization to the NGSP and IFCC scales allows results to be compared across laboratories, with mmol/mol increasingly reported alongside the familiar percentage.

Physiology and What It Measures

The Glycation Reaction

HbA1c arises from a nonenzymatic reaction between glucose and hemoglobin. Glucose crosses the red cell membrane via GLUT1 and reaches an intracellular concentration that tracks plasma glucose, where it condenses with the free α-amino group of the N-terminal valine on the hemoglobin β-chain. The first product is a labile Schiff base (an aldimine), which can dissociate if glucose falls. Over hours to days, a fraction of these Schiff bases undergo an Amadori rearrangement into a stable ketoamine, and it is this stable, essentially irreversible adduct that the HbA1c assay measures. Because formation is driven by mass action, the steady-state fraction of stable glycated hemoglobin is proportional to the time-averaged glucose concentration the cell has experienced. The reaction also occurs at other hemoglobin residues and on many other proteins, but the β-chain N-terminal adduct is the defined HbA1c species used clinically. This molecular picture explains why HbA1c is unresponsive to a single meal yet faithfully reflects the prevailing average, and why a labile fraction must be removed or accounted for by modern assays to avoid short-term artifact.

Red Cell Lifespan and the Measurement Window

The second determinant of HbA1c is the lifespan of the red cell itself. A normal erythrocyte circulates for 90 to 120 days, accumulating glycation throughout, so the measured fraction is a weighted average across a population of cells of different ages. Younger cells, which dominate after blood loss, hemolysis, transfusion, or erythropoietin therapy, carry less glycation and pull the average down; an older mean cell age, as in iron deficiency or after splenectomy, pulls it up. The weighting is not uniform: because glycation continues to accumulate as cells age, recent glucose is overrepresented, and roughly half the HbA1c value reflects the most recent month. Direct measurements of red cell survival in healthy people (Cohen et al., Blood 2008) show enough variation to shift HbA1c by several tenths of a percentage point at identical mean glucose. This is the physiological basis for the entire category of HbA1c confounders, and it is the reason a result that conflicts with fingerstick or sensor glucose should trigger a look at hematology before a change in management.

From HbA1c to Average Glucose

Clinically, HbA1c is often translated into an estimated average glucose to make it intuitive. The A1c-Derived Average Glucose study (Nathan et al., Diabetes Care 2008) related HbA1c to four weeks of combined continuous and self-monitored glucose in 507 people and produced the linear relationship in which estimated average glucose in mg/dL equals 28.7 times the A1c minus 46.7. By this relationship an HbA1c of 5.4 percent corresponds to an average glucose near 108 mg/dL, and 6.5 percent to about 140 mg/dL. The translation is a population average, and individuals vary in how much glucose it takes to produce a given HbA1c, a phenomenon described as the glycation gap or hemoglobin glycation index. People who glycate more readily run a higher HbA1c at the same mean glucose, which has measurable consequences for both diagnosis and the apparent intensity of their disease. Awareness of this inter-individual variation prevents over-interpretation of small differences in HbA1c between people.

Standardization and Units

HbA1c is one of the most rigorously standardized assays in laboratory medicine, which is what makes its thresholds portable across the world. In the United States, results are reported on the NGSP scale as a percentage, anchored to the assay used in the DCCT, while the IFCC reference method reports the same quantity in mmol/mol of glycated to total hemoglobin. The two scales are linked by the equation IFCC mmol/mol equals (NGSP percent minus 2.15) times 10.929, so 5.4 percent equals about 36 mmol/mol and 6.5 percent equals about 48 mmol/mol. The NGSP harmonization program (Little et al., Clinical Chemistry 2011) reduced what had been chaotic inter-laboratory variation to a tightly controlled system, which is why a numeric HbA1c target is meaningful regardless of where the blood is drawn. Different analytical methods, including ion-exchange high-performance liquid chromatography, boronate affinity chromatography, and immunoassay, handle hemoglobin variants differently, so the platform matters when a variant is present even though the scale is shared.

Clinical Evidence

Microvascular Disease and the Causal Core

The strongest and most clearly causal evidence concerns small-vessel disease. The Diabetes Control and Complications Trial randomized 1,441 people with type 1 diabetes and demonstrated that lowering HbA1c from roughly 9 to 7 percent reduced retinopathy progression by 76 percent, nephropathy by 50 percent, and neuropathy by 60 percent. The United Kingdom Prospective Diabetes Study extended this to type 2 diabetes, with UKPDS 33 showing a 25 percent reduction in aggregate microvascular endpoints under intensive control of 3,867 newly diagnosed patients. The observational UKPDS 35 analysis added the continuous dimension, finding a 37 percent lower microvascular complication rate per 1 percentage point lower HbA1c, with benefit extending toward the normal range. Together these results make microvascular protection the clearest reason to lower a high HbA1c, and they establish glycemic exposure as a direct cause rather than a marker of small-vessel injury.

Macrovascular Disease and the Limits of the Association

The relationship between HbA1c and large-vessel disease is real but more complex. Observationally, HbA1c tracks coronary disease and stroke continuously, and in adults without diabetes the ARIC study found risk rising steeply above 6.0 percent, with HbA1c outperforming fasting glucose. Yet interventional trials that lowered HbA1c aggressively in established type 2 diabetes produced limited macrovascular benefit. The ACCORD trial, targeting an HbA1c below 6.0 percent in 10,251 high-risk participants, was halted early after a 22 percent increase in all-cause mortality in the intensive arm, and ADVANCE found no significant macrovascular reduction at a target of 6.5 percent or below. The reconciling interpretation is that chronic hyperglycemia contributes to atherosclerosis, but the macrovascular association is heavily shared with insulin resistance, dyslipidemia, and hypertension, and the harms of intensive pharmacological lowering, especially hypoglycemia, can offset its benefits in advanced disease. The macrovascular lesson is that the number is informative but not a license to drive it to the floor by any means.

Longevity and All-Cause Mortality

For a longevity-oriented reader the general-population mortality data are central. EPIC-Norfolk followed 4,662 men and found all-cause mortality rising continuously with HbA1c, with the lowest risk below 5.0 percent and most of the excess deaths concentrated in people whose HbA1c was still in the non-diabetic range. The ARIC analysis similarly linked higher HbA1c to all-cause mortality in adults without diabetes after adjustment for conventional risk factors. At the very low end the curve often turns upward, so that HbA1c values below about 5.0 percent are associated with higher mortality; this lower limb is widely interpreted as reverse causation, with anemia, liver disease, malnutrition, alcohol use, and frailty lowering HbA1c while independently raising death rates, rather than as evidence that low glucose is harmful in healthy people. Within established diabetes, registry data describe a U-shaped relationship with a nadir near 7.0 to 7.5 percent, reflecting hypoglycemia and comorbidity at the extremes. The practical synthesis is that a value in the low-to-mid 5 percent range, in a person who is otherwise healthy and not anemic, marks the metabolic position associated with the longest survival.

Causality and Mendelian Randomization

Distinguishing cause from correlation is essential because HbA1c sits downstream of both glucose metabolism and red cell biology. Mendelian-randomization studies that use genetic instruments for higher glucose and HbA1c support a causal effect of glycemia on microvascular outcomes such as retinopathy and on type 2 diabetes itself, consistent with the interventional trials. The macrovascular picture from genetic studies is weaker and partly attributable to the overlap of glycemic genetics with adiposity and lipid traits, mirroring the disappointing macrovascular results of intensive-lowering trials. A further subtlety is that some HbA1c-associated genetic variants act through red cell physiology rather than glucose, biasing naive analyses; partitioning HbA1c variants into glycemic and erythrocytic groups has shown that the erythrocytic component can distort the apparent HbA1c-outcome relationship. The honest summary is that glucose exposure is causal for small-vessel disease and diabetes, that its causal contribution to atherosclerotic events is real but smaller and shared with other traits, and that part of the observed HbA1c-mortality association is non-glycemic and reflects the conditions that alter red cell turnover.

Prediabetes and Future Diabetes Risk

HbA1c in the 5.7 to 6.4 percent band identifies a population at high risk of developing diabetes, and the systematic review by Zhang et al. (Diabetes Care 2010) quantified this gradient, with 5-year incidence rising from roughly 9 to 25 percent at 5.5 to 6.0 percent to 25 to 50 percent at 6.0 to 6.5 percent. Because the test needs no fasting and reflects months of glucose, it is convenient for screening, but it does not always agree with fasting glucose or the oral glucose tolerance test, and a meaningful minority of people are classified differently by each. The disagreement is partly biological, since the three tests probe different aspects of glucose handling, and partly the result of HbA1c confounders. In practice a borderline HbA1c is best confirmed and supplemented with a fasting glucose and, where insulin resistance is suspected, a fasting insulin, since compensatory hyperinsulinemia can precede any rise in average glucose by years.

Optimal versus Conventional Ranges

The conventional laboratory framework labels HbA1c below 5.7 percent as normal, 5.7 to 6.4 percent as prediabetes, and 6.5 percent or above as diabetes, but these are diagnostic boundaries derived from the risk of microvascular disease and the distribution of the population, not optimal targets for a healthy person. The cohort data place the lowest cardiovascular and all-cause mortality risk near 5.0 to 5.4 percent, comfortably below the 5.7 percent threshold, so a result of 5.6 percent is normal by label yet above the band associated with the best long-term outcomes. The gap between normal and optimal is the central teaching point of the marker: a person whose HbA1c drifts from 5.2 to 5.6 percent has moved meaningfully along the risk gradient while remaining within the normal range throughout. The optimal framing must always be qualified for people with established diabetes, in whom a target that low is neither expected nor safe, and for older or frail individuals, in whom avoiding hypoglycemia takes precedence over a tight number.

Interpretation in Practice

When HbA1c Misleads

The most important interpretive skill is recognizing when HbA1c does not reflect true glycemia. Any condition that shortens red cell survival, including G6PD deficiency, sickle cell and other hemoglobinopathies, hereditary spherocytosis, hemolytic anemia, recent blood loss, and recent transfusion, lowers HbA1c and can mask hyperglycemia. Conditions that lengthen red cell survival or age the red cell population, particularly iron deficiency and vitamin B12 deficiency, raise HbA1c independent of glucose, so a mildly elevated HbA1c in an iron-deficient person should be rechecked after repletion. Pregnancy lowers HbA1c through accelerated turnover, kidney disease can move it in either direction through carbamylation and shortened survival, and hemoglobin variants can interfere with specific assays in a method-dependent way. The unifying rule is that a result discordant with fingerstick or continuous glucose data is a hematology question first and a glucose question second, and that fructosamine or glycated albumin, which reflect glycation of serum proteins over two to three weeks and are independent of red cell lifespan, are the appropriate alternatives when red cell turnover is abnormal.

Retest Cadence and Complementary Markers

Because HbA1c takes about 3 months to fully express a change in average glucose, the natural retest interval is roughly quarterly, and testing much more often outside of pregnancy or a rapid treatment change yields noise rather than signal. HbA1c is an average and therefore blind to variability, so it should be paired with other markers that capture what an average cannot: fasting glucose and fasting insulin or HOMA-IR detect insulin resistance earlier in its course, since the body defends a normal glucose with rising insulin long before HbA1c moves; continuous glucose monitoring reveals time in range, post-meal excursions, and hypoglycemia that a single number conceals; and fructosamine provides a shorter two-to-three-week window when a faster read is needed or when red cell biology is disturbed. Used this way, HbA1c is the stable backbone of glycemic assessment, accurate for trend and prognosis, while the complementary markers supply the resolution and the early warning that its long, slow memory cannot.