# DIABETES-EHLB Selected Discussions

## TOPIC #1: Hemoglobin A1c (Hba1C)

I have been looking for a definitive answer concerning my questions(s), that is related to Hemoglobin A1c (Hba1C).

[Q] Is it possible to calculate an individual's Hemoglobin A1c (Hba1C), from their blood glucose readings [trend]? If so, could you point me to the equation?

Thanks much,
Mike

RESPONSE 1: On Thu, 12 Dec 1996, Michael Patterson wrote:

```I have been looking for a definitive answer concerning my questions(s),
that is related to Hemoglobin A1c (Hba1C).

[Q] Is it possible to calculate an individual's Hemoglobin A1c (Hba1C),
from their blood glucose readings [trend]?  If so, could you point
me to the equation?
```

Theoretically, yes. Sampling theory will demonstrate that you would need at a minimum 8 bg tests/day including overnight tests to produce a good estimate of A1c. Any practicing data cruncher would probably want twice that. -Charly Coughran

RESPONSE 2: Having tested frequently enough to get a representative sampling, a rough translation of bg to HbA1c can be made as follows:

```            bg              HbA1c
---------       ------------
60              4
90              5
120              6
150              7
180              8
210              9
240             10
270             11
300             12
330             13
```
John Kinsley

RESPONSE 3: Thanks for the reply. Eight tests a day - that doesn't sound too practical :) I suppose that calculating it is probably something that I shouldn't pursue.

Thanks,
Mike

RESPONSE 4: There is no precise method of converting a group of blood glucose readings to an HbA1c because the two are an "apples and oranges" case (i.e. one is not directly convertible to the other as the two do not measure the same thing.) I remember an extensive discussion in the Compuserve Diabetes Forum about a formula for approximating one given the other(about 2 yrs ago.) You might try there. Dave Groves was the senior sysop when I was there.

Greg

RESPONSE 5: Keep in mind that this is very, very rough. Due to the different lab methodologies and lack of standardization, every lab has a different normal range for A1c. I have seen labs which report the high end of normal at around 5% and at around 7%. Given that kind of variability the minimum error bars here should be + or - 1%. I understand there is a standardization committee at work. Does anyone have any information as to their progress? It would be nice to be able to easily compare A1c values between labs, or ours to the literature.

RESPONSE 6: I got this info. from a government site, so I assume that it's fairly accurate. - [if someone is interested in where I found it I'll post the site]

The hemoglobin Alc test Another test for blood sugar, the hemoglobin A1c test, shows what your average blood sugar was for the past 3 months. It shows how much sugar is sticking to your red blood cells. The doctor does this test to see what level your blood sugar is most of the time. See your doctor for a hemoglobin A1c test every 3 months. To do the test, the doctor or nurse takes a sample of your blood. The blood is tested in a laboratory. The laboratory sends the results to your doctor. If most of the blood sugar tests you do yourself show that your blood sugar is around 150, the hemoglobin A1c test should be almost normal. If most of your tests show high levels of blood sugar, then the hemoglobin Alc test is usually high. Ask your doctor what your hemoglobin A1c test showed.
Please don't take what I say for fact and assume that it's 100% wrong. Since it appears that HbAlc results are correlated to an individual's average blood sugar, it is possible (at least I think) to calculate a 'very rough' HbAlc value from your bg readings. As long as their are plenty of readings within the time range, then I think that you can get a 'very rough' estimate. It looks like John's data would give the equation - f(x) = 2 + (1/30)x; - so maybe it is possible. I might be 100% wrong, so please don't take it for fact. Give it your +1% or -1% and maybe it will come close, as long as their are plenty of readings, etc... A number of variables are there, but I'm thinking 'very rough'???

I found the information at:

http://www.niddk.nih.gov/LevelBest/chap2.htm - National Institutes of Health

RESPONSE 7: The following is a highlight of a presentation at a recent ADA meeting (see PharmInfoNet: The Effect of Immediate Feedback of HbA1c on Patient Glycemic Control and Their Physicians' Treatment Decisions).
The Effect of Immediate Feedback of HbA1c on Patient Glycemic Control and Their Physicians' Treatment Decisions.
David G. Marrero, Stephanie Kraft, Naomi Fineberg,
Indianapolis, IN
June 8, 1996

Periodic glycated hemoglobin (GHb) testing is recommended for assessing glycemic control and basing treatment decisions in patients with diabetes. Elevated levels of glucose result in the glycosylation of a variety of proteins, including hemoglobin. Glycated hemoglobin is easy to assay in a sample of blood, and because red blood cells have a life span of about 90 days, GHb readings provide a good indication of mean glucose control over the past 90 days. However, current technologies for assaying GHb (the assay is typically done for the GHb known as HbA1c) require laboratory time and therefore, prevent physicians from routinely using the data when it would be most useful: during the patient visit.

Using a new technology, the DCA 2000 System (tm) (Bayer), which enables an HbA1c measurement to be obtained from a capillary sample in 9 minutes, the investigators studied the impact of having an immediately available GHb result on patients' glycemic control and physicians decisions concerning pharmacologic therapy. Faculty and resident physicians were randomly assigned to receive either GHb results immediately during clinic visits with a randomly assigned subset of their adult IDDM and NIDDM patients (n=113) or to have the results placed into their patients' charts (n=175) five days after the clinic encounter. Outcomes included patients' glycemic control as measured by GHb and the appropriateness of physicians' therapeutic decisions regarding glycemic control of their patients over the course of 6 months.

Immediate feedback was associated with improved glycemic control in patients with IDDM and NIDDM who were using insulin (p=.02); and improved therapy decisions for non-insulin requiring NIDDM patients with GHb greater than 9% (p=.004). Furthermore, there was a relationship between the level of physician training and outcomes: faculty, as opposed to resident physicians, receiving immediate feedback were more likely to take approporate action in response to elevated GHb values for both IDDM and NIDDM (p less than .01).

These data suggest that providing GHb results during the clinical encounter may improve the physicians' ability to make appropriate therapeutic decisions that can result in improved glycemic control. In addition, non-insulin requiring NIDDM patients in poor glycemic control may realize particular benefit because their drug therapy can be more effectively adjusted in response to HbA1c levels. Unfortunately, the investigators observed that because the use of GHb as a monitoring tool for diabetes is not used widely, the value of having immediate GHb results may still be limited by physicians' inexperience with interpreting and acting upon GHb results.

RESPONSE 8: John Mack forwarded highlights of a recent presentation on use of the DCA 2000 System for fast HbA1c analysis from a fingerprick sample.

A couple of years ago, the manufacturer was advertising the DCA 2000 in Diabetes Care and similar professional publications. They've stopped; presumably they felt that awareness was sufficient that they need not continue. Someone reading misc.health.diabetes called and found that the list price was about \$2000, with estimated actual cost to an office of about \$1500. Thus it is not an unreasonable investment for a doctor who monitors a number of diabetic patients.

Another way of simplifying HbA1c testing is to do it by mail. I pursued this option because the local SmithKline lab wanted my doctor to charge me \$96 for the test, even though I knew people in other parts of the country were paying \$20-\$30. After some time, I found that Diabetes Support Systems provides a collection kit and instructions for mailing the sample to a lab which cooperates with them. The kit, which is sufficient to collect four samples, costs about \$30, and the lab charges \$15 per test, for a total cost of under \$25 per test. I've done it three times in the past eight months and have been pleased with the response. Without this service; I probably would have had no HbA1c tests in this period otherwise due to the cost and the inconvenience of visiting the lab or doctor's office.

I spoke with someone at DSS about the lab's equipment. It was not a DCA 2000. I do not remember the name of their machine, but they told me that it is exactly the same equipment as SmithKline uses.

DSS advertises in Diabetes Forecast. Their number is 1-800-252-0207.

Edward Reid