Although glycosylated hemoglobin (or HbA1c) has been used for thirty years, now, to allow us to know how well a person's diabetes is controlled, it has never been used to actually diagnose someone as having diabetes in the first place.
To understand what HbA1c is go to this link:
For years the only way to diagnose diabetes, or to screen people to see if they have diabetes, was to measure their blood glucose level after they had had nothing to eat for at least ten hours (called a Fasting Blood Glucose level) or to measure blood glucose at a random time of the day or after a heavy meal or a large glucose drink.
A normal fasting blood glucose level is less than 100mg/dl (5.6 mmol/L). A fasting blood glucose level over 125mg/dl (6.9 mmol/L) means the person has diabetes, and a value in between 100 and 125mg/dl means that the person has impaired fasting glucose (sometimes called Impaired Glucose Tolerance). People in this last category are at higher risk for developing diabetes in the future.
A random blood glucose under 140mg/dl (7.8 mmol/L) is normal. A random blood glucose level over 200mg/dl (11.1 mmol/L) means the person has diabetes, and a value between 140 and 199mg/dl means that the person has impaired glucose tolerance and is at higher risk for developing diabetes in the future.
When using either a fasting blood glucose or a random blood glucose to diagnose diabetes we require that there are TWO abnormal values on different days. This is to make sure that the first measurement wasn't just a laboratory error or mistake. This can happen from time to time. A tube of blood can get labeled for the wrong patient. The person might have just finished eating a donut and had sugar on their skin when the finger prick blood sample was taken. Although rare, it is possible to get a wacky result from a lab test. We would not want to tell someone that they had diabetes unless we were very sure, and so we always like to have two abnormal tests done on different days.
Starting on January 1st 2010 the worldwide medical community (led by the American Diabetes Association, or ADA) has agreed that HbA1c can be used to diagnose diabetes. This is great news. HbA1c can be measured any time of day so you don't have to be fasting for ten hours before getting the test done. HbA1c is a very sensitive and reliable test which gives an average of all the ups and downs of blood glucose in the previous twelve weeks. The other good news is that it is still okay to use the previous tests (based on fasting or random blood glucose measurement). But now finding a HbA1c that is 6.5 percent or higher on two occasions can also be used to diagnose diabetes.
So if you someone is tested in the middle of the day and has a random blood glucose over 200mg/dl AND a HbA1c of 6.5% or higher then they have diabetes. Or if they come in fasting and have a fasting blood glucose over 125mg/dl AND a HbA1c of 6.5% or higher then they have diabetes.
The ADA did two other things, one of which is good and the other of which is a bit confusing. The good thing is that they have stopped using the term "Prediabetes" which I have always thought is a sloppy and inaccurate term (read my answer to "What is Prediabetes?" in The Diabetes Answer Book pages 39-40). Instead the ADA lists "Categories of increased risk for diabetes," which includes impaired fasting glucose (100-124 mg/dl) or impaired random blood glucose levels of 140-199 mg/dl. But when they define the level of HbA1c that gives an increased risk for diabetes they say 5.7-6.4%. Now if the normal non-diabetic range for HbA1c is 4.0-6.0% then it does make sense that anyone with a HbA1c level of 6.1%, 6.2%, 6.3%, or 6.4% is above normal but not yet high enough to be called diabetes. But someone with a HbA1c level of 5.7%, 5.8%, 5.9%, or 6.0% is still in the normal range! Why should we think of them as being at increased risk for diabetes in the future? Well, these various cut points to define what it "normal," what is "slightly abnormal" or "impaired" and what is "really abnormal" or "diabetic" are somewhat arbitrary cut points. Over the past few decades groups of scientific researchers have argued about where to put these cut points and have changed their minds several times!
It certainly makes sense to think that someone who has a HbA1c level in the upper end of the non-diabetic normal range (5.7-6.0%) is probably at a higher risk for developing diabetes in the future than someone who has a HbA1c level in the lower end of that normal range (4.0-4.9%). For me, the fact that we can now use HbA1c to diagnose people with diabetes and can identify people who do not have diabetes yet but are at increased risk is a good thing.
