Research & Education

Hemoglobin A1c: Less Reliable Than We Thought?


Hemoglobin A1c or glycated hemoglobin is typically taken to represent a three to four month average blood glucose level. This is because red blood cells are assumed to live approximately 120 days. But is this true for everyone? Or do different pathological conditions or genetic polymorphisms affect the lifespan and behavior of red blood cells? Research indicates that hemoglobin A1c (HbA1c) is more variable than is generally acknowledged. This has important implications for physicians and patients as HbA1c may not be as reliable a marker as is usually believed.

For starters in a somewhat paradoxical situation individuals with type 2 diabetes (T2D) may have an artificially low HbA1c. Hyperglycemia reduces the lifespan of erythrocytes giving these cells less time to become glycated even in the presence of a high glucose concentration. In such a situation it’s unlikely that a type 2 diabetic would have an HbA1c in the non-diabetic range but it might be substantially lower than would be expected based on their typical post-prandial blood glucose or their glucose level throughout the day in general. This could lead the patient and their physician to believe the diabetes is better controlled than it actually is.

A study that looked at the relationship between erythrocyte survival time and HbA1c quantified these measurements in 23 men with T2D (non-smokers age 40-77). Researchers found that “there is a hyperglycemia-related decrease in erythrocyte survival that results in a progressively greater underestimation of the severity of the hyperglycemia the higher the GHb [glycated hemoglobin] percentage.”

This study also determined that erythrocyte lifespan decreased by about 7 days for every 1.0% increase in HbA1c. In four subjects with HbA1c >12% mean erythrocyte survival time was just 81 days—almost 35% less than the mean of 123 applied in the study.

Contrast this with the opposite situation in exceptionally healthy people: in these individuals erythrocyte survival time is longer than normal. This gives the cells more time to accumulate sugar and become glycated which could result in an artificially high HbA1c. Healthcare professionals may encounter such patients: their fasting glucose is normal and if they check postprandial glucose intermittently that is normal as well. All other metabolic markers look fine too which makes an elevated HbA1c a mystery. The mystery is explained by the longer lifespan of the erythrocytes and is not cause for alarm. (Patients in this context will likely not have an HbA1c in the diabetic or pre-diabetic range but the HbA1c will be higher than otherwise expected—for example 5.4 rather than 5.0. And it is not unheard of for the HbA1c to indeed be in the low end of the pre-diabetic range yet all other parameters look good.)

In these cases it may be helpful to look at fructosamine which usually reflects glycated albumin. (In some cases it reflects total glycated serum proteins but since albumin is the most abundant protein in blood it is usually taken to be glycated albumin.) Albumin has a half-life of approximately 20 days so the fructosamine measurement reflects more recent blood glucose levels (2-3 weeks). Fructosamine is not believed to be affected by red blood cell lifespan.


Another factor affecting erythrocyte lifespan and HbA1c is anemia. There are many different types of anemia but some of them result in a shortened lifespan for red blood cells. This may result in an artificially low HbA1c. On the other hand iron deficiency anemia (IDA) is associated with higher HbA1c. In a study that compared HbA1c in 120 subjects with IDA and well-controlled T2D to 120 iron-sufficient type 2 diabetic subjects those with IDA had higher HbA1c. In subjects younger than 50 those with IDA had a mean HbA1c of 6.71 (± 1.64) compared to 5.53 (± 0.59) in those who were iron sufficient. Not surprisingly the numbers were slightly higher for subjects older than 50: mean HbA1c was 6.94 (± 1.27) in those with IDA and 5.77 (± 0.72) in those who were not anemic. 

In a separate study involving non-diabetic subjects with IDA mean HbA1c at baseline was 7.4 (± 0.8). After a 3 month intervention with supplemental iron (100mg/d) this decreased to 6.2% (± 0.6). The study authors noted “Iron deficiency must be corrected before any diagnostic or therapeutic decision is made based on HbA1c.” (On the other hand one study found that individuals with IDA had lower HbA1c than healthy subjects and that it increased with iron supplementation. So this should be evaluated on a case-by-case basis.)

It is also essential to remember that many people with HbA1c in the healthy range are at risk for insulin resistance. They may also have fasting glucose at a healthy level but very often these two measurements remain normal only because pathologically and chronically elevated insulin is keeping them in check. This is a dramatically underappreciated issue as hyperinsulinemia—even in those with normal fasting glucose HbA1c and response to an OGTT—is a driving factor behind a staggering number of health issues that have exploded in incidence in recent years.

Genetic Polymorphisms

Finally ethnic and genetic factors influence erythrocyte physiology and HbA1c. It is well known that certain traits either persist or are eliminated through the generations. (For example lactase gene persistence among populations with a history of dairying or a specific polymorphism that indicates adaptation to a fat- and protein-rich diet among Arctic peoples.) So it shouldn’t surprise us that there are polymorphisms that affect red blood cells and even hemoglobin itself. 

People of African Mediterranean or Southeast Asian descent may have hemoglobin variants that influence the accuracy of HbA1c and there are several other variants found in individuals from other parts of the world. Some laboratories take hemoglobin variant into account in their assays but some do not. (Detailed information on the 20 most popular testing methods and whether they account for Hb variant can be found here.) 

Bottom line: HbA1c is not as clear-cut a measurement as we might think. Like any other marker in the blood it should not be evaluated in isolation. It is one tile in a broad mosaic that is best taken as a whole and not the sum of its individual parts.