But the A1c is still an imperfect tool for diagnosis, because, while it works well in large studies where you want a fast way to summarize the blood sugar status of thousands of people, it often fails in individuals.
Anemia, abnormally long-lived red blood cells, and some genetic conditions may give an individual an A1c test result that has no relationship to that individual's actual blood sugar level.
The only test that accurately determines whether a person has abnormal blood sugar is the Oral Glucose Tolerance Test (OGTT). Unfortunately, this test is expensive and time consuming so it can't be used for routine screening.
The current formula used to equate an A1c to a blood sugar is the ADAG formula. I've created a calculator that will let you convert any A1c result into the average blood sugar that is supposed to correlate to it. You can find it here:
A1c/Average Glucose Calculator
Remember that the average glucose is an abstract concept that bears only a slight relationship to actual blood sugar. You can have an average blood sugar of 150 mg/dl by veering from 40 mg/dl (a serious hypo) to 260 mg/dl--both dangerous blood sugar levels, or you can get it by having a blood sugar that varies between 140 and 160, a pre-diabetic blood sugar that, while it is high enough to cause problems, causes only slow developing problems that take years to manifest.
If you are diagnosed using a Oral Glucose Tolerance Test (OGTT) you should be called diabetic if your blood sugar goes over 200 mg/dl but some tests only look at the 2 hour number and diagnose you if that is over 200 mg/dl. It might have come down after going as high as 300 mg/dl the first hour, so the reliance on one reading may miss the accurate diabetes diagnosis.
If your blood sugar on the (GTT is over 140 mg/dl (7.7 mmol/L) at 2 hours but under 200 mg/dl (11 mmol/L) you'll be told you have "prediabetes" which many doctors treat as unimportant. It isn't unimportant. Blood sugars in that range are high enough to cause heart disease, "diabetic" nerve damage, and "diabetic" retinal damage. (Details HERE)
What A1c Is Normal?
A study just published in Diabetes Care gives more insight into what really is a "normal" A1c. It analyzed the data from 2,494 Australian clinic patients to come up with the range that seemed to correlate to diabetes and then applied it to another 6,015 people to see if it could be used as an effective screening tool.
A1C for Screening and Diagnosis of Type 2 Diabetes in Routine Clinical Practice.
Zhong X. Lu, et al. Diabetes Care April 2010 vol. 33 no. 4 817-819. doi: 10.2337/dc09-1763
This study found:
1. An A1c under 5.5% was a good indicator of normal blood sugar (probably determined with a OGTT, though this isn't stated in the abstract.)
2. An A1c over 7% was almost always an indicator of diabetes.
3. In those with A1cs of 5.6% to 6.9% between 61.9–69.3% had "abnormal glucose status."
What you should take from this is that if you are given an A1c screening test, you need to ask your doctor what the actual result was. If it is over 5.5%, you should test your blood sugar with a meter at home. You can learn how HERE)
If you see blood sugars routinely over 140 mg/dl (7.7 mmol/L) two hours after meals, consider yourself as having "prediabetes" and take steps to lower your blood sugar. You'll find instructions HERE.
If your A1c is much lower than predicted by your average blood sugars, it is possible you are one of the people for whom this test isn't accurate. You can read more about that HERE.
A1c Doesn't Correlate to Complications anywhere NEAR as Well as Post-Meal Blood Sugar
Whatever your A1c might be, keep in mind that the reason doctors rely on it is because is is cheap, and they don't have the time to look at your blood sugar logs to see what is really happening. That's why after your diagnosis most doctors use the A1c, exclusively, to measure your progress and likelihood of developing complications. Unfortunately, that is where it fails the most.
A lot of research makes it crystal clear that post-meal blood sugar highs predict the development of complications much better than A1c. This is because as mentioned above, the same A1c may reflect widely different patters of blood sugar behavior.
The weight of the evidence suggests that blood sugars that rise over 150 mg/dl(8.3 mmol/L) cause complications, especially if they stay elevated over that level for more than a brief time. A two hour reading of 150 mg/dl should set alarm bells ringing no matter what your A1c might be. Conversely, if you are keeping your two hour readings at 100 mg/dl(5.6 mmol/L) or lower, and still have a high A1c, you are almost certainly less likely to develop the classic diabetic complications.
Two different studies of neuropathy found that the only measure of blood sugar control that correlated with the presence of neuropathy was the 2 hour OGTT result. Neither A1c or fasting glucose had any predictive value.
Most doctors consider any A1c under 8% as "good control." My lab still prints that on their lab result sheet. If you think blindness and amputation are good, that might be an okay level. For those of us who want normal health, a normal A1c is required. The Australian study suggests that level should be 5.5% or under.
NOTE: Dr. Richard K. Bernstein, for whom I have immense respect insists that only a 4.3% A1c is normal. But I have seen people with 5.4% A1cs whose numbers after eating any amount of carbohydrate was never over 115 mg/dl (6.4 mmol/L).
That makes me think that Bernstein's number is defining "normal" in a way that eliminates all but a tiny percentage of the population, which means that though the 4.3% A1c might be perfect, it probably isn't at all normal. If you can get lower than 5.5%, great--especially if the low reading isn't due to anemia which gives very low readings no matter what your blood sugar might be.
My own A1c is always considerably higher than my post-meal testing would predict, as are those of many people I hear from. I don't have any classic complications 12 years after diagnosis, so I no longer stress about this. Watch actual highs by testing routinely 1 and 2 hours after meals, and you'll know exactly how you're doing and can ignore concepts like "average" which aren't as helpful as "actual."