Wouldn’t it be wonderful if it were as easy to check your A1C level at home as it is to check your blood glucose! It is the A1C, of course, that best measures the control we have over our diabetes.
Testing A1C 19 Times
It actually can be that easy. You may know that recently several home A1C test kits and meters have become available. You may, however, wonder how accurate they are.
That was my concern. Therefore, I offered up 19 drops of blood for A1C testing during a recent three-month period. These 19 A1C tests compared an A1C meter, an A1C test kit, and one of the country’s largest laboratories.
My meters Web page at “A1C Meters and Kits” describes and links three meters and three test kits. However, only one of the meters, Metrika Inc.’s A1cNow, is currently marketed for home use.
Two of the test kits—the Accu-Base Hemoglobin A1C Sample Collection Kit marketed by Diabetes Technologies Inc. (DTI) and the SimpleChoice A1C, which at that time marketed by SpectRx—are the same. SpectRx licensed the DTI kit.
The two other test kits—BioSafe Collection Kit marketed by BioSafe Laboratories Inc. and the A1C At Home collection kit marketed by FlexSite Diagnostics Inc.—use a filter paper process.
I did not evaluate these two kits, because their laboratories didn't pass the rigorous requirements of the National Glycohemoglobin Standardization Program (NGSP).
”Neither FlexSite nor BioSafe are currently NGSP certified,” writes Randie Little, network coordinator of the National Glycohemoglobin Standardization Program. “I would only recommend a lab that is either using a certified method or is a certified lab. However, most certified labs are clinical trial laboratories. ”
The National Glycohemoglobin Standardization Program, located at the University of Missouri, Columbia, has standardized glycohemoglobin (A1C) test results so that clinical laboratory results are comparable to those reported in the Diabetes Control and Complications Trial (DCCT). The NGSP acts as a reference laboratory for manufacturers of A1C methods to help in standardizing methods and providing comparison data for certification of traceability to the DCCT.
The use of a test method traceable to the DCCT is one part of it. But, “There is one other thing that each laboratory needs to do,” says Dr. David Goldstein, who just stepped down as NGSP Steering Committee chair. “Every laboratory needs to establish its own normal range.” That’s because the different assays that the laboratories can use have “a little bias.”
The reference (normal) range for the High Performance Liquid Chromatography (HPLC) assay used to test SimpleChoice and DTI samples is 4.2 percent to 5.8 percent. For the A1cNow system the expected normal range is 3.9 percent to 6.5 percent. Reports from Unilab define less than 6.0 percent as non-diabetic.
A1cNow does not, of course, use an NGSP certified laboratory, because you do the test at home and don’t send it anywhere. The A1cNow kit uses an immunoassay method, but they did not it compare to an NGSP immunoassay for certification.
The difference between standardized methods and NGSP certified laboratories is one of the most misunderstood aspects of A1C testing.
“A certified laboratory is different than a laboratory that uses a certified method,” writes Dr. Goldstein. “Certified labs are labs that go one step further and these are usually large reference labs.”
The way that Dr. Little puts it is, “As far as saying that the lab or method is ‘standardized’ vs. ‘certified,’ any lab can say that their method is standardized. But the proof of standardization is certification.’
Dr. Little adds that 98 percent of the labs in this country use certified methods. But only nine laboratories in the U.S. are certified at Level I and three more are Level II certified, according to the NGSP list of certified laboratories. These laboratories have documented a certain level of precision and accuracy.
I lived in California at the time of my testing, and only one lab in my state is Level I certified. It is the Quest Diagnostics lab in Van Nuys, but it will accept samples only from people in clinical trials.
The Quest lab in San Jose is the one that analyzes A1C tests that my endocrinologist in Santa Cruz prescribes for testing in a local clinic. Robert Sobel, vice president, Bay Area Technical Operations for Quest, says that, while they use a certified method (Roche Integra), the lab isn’t certified. He also told me that they could not send my blood samples to their lab in Van Nuys.
We had tried to have my samples sent to Quest tested in their Van Nuys facility. We hatched the plan in a conference call among Arthur G. Williams, president and chief scientific officer of Diabetes Technologies Inc. (DTI), Keith D. Ignotz, president and chief operating officer of SpectRx Inc., and myself.
We wanted the most accurate possible comparison between the results that I could get from my lab and those from the SimpleChoice A1C test kit, which that company licensed at that time from DTI. Subsequently, SpectRx discontinued the Simple Choice. A Level I certified laboratory, Premier Laboratories in Kansas City, Missouri, tests blood samples sent to DTI.
SpectRx paid for me to have 10 A1C tests on the same day. I went to Unilab for five of them and sent in five SimpleChoice kits. The Unilab test results—all of which were taken from the same large vial of my blood but analyzed separately—were 5.8 percent, 5.9 percent, 5.9 percent, 5.9 percent, and 5.9 percent. The spread of the SimpleChoice tests was a bit wider: 6.4 percent, 6.6 percent, 6.7 percent, 6.7 percent, and 6.7 percent, which averaged about 6.6 percent.
The difference was remarkable. That difference—variation—between the two laboratories is unacceptable, according to my local endocrinologist, Dr. Jeffrey F. Williams. He adds that the precision or interassay results are good. But which is more accurate?
As much as I would like to believe that my A1C was at the 5.9 percent level of the Unilab tests, I have reluctantly come to the conclusion that my A1C was the mean of that reported by SimpleChoice, 6.6 percent. There is no arguing with the fact that the SimpleChoice and DTI tests are the only A1C results I could get that are tested in Level I certified labs.
While the spread of the SimpleChoice results was wider than those from Unilab, I believe that is because all the tests came from the same drop of blood. My guess is that five separate blood draws would not have been as close. Since Unilab is not an NGSP certified lab, it seems to me that their results were the inaccurate ones.
The tests do show that you can confidently track your A1C trend. You can do that, however, only if you use the same lab every time.
I did not make a head-to-head comparison between the Metrika A1cNow meter and my usual lab or with SimpleChoice. I did run five A1cNow tests in a 15-minute period. The results were 6.4 percent, 6.4 percent, 6.5 percent, 6.5 percent, and 6.7 percent. The range was exactly the same as that of the SimpleChoice test.
Then, you might think, that the A1cNow meter is just as accurate as the SimpleChoice/DTI kit. But SimpleChoice/DTI use a Level I certified laboratory, while the A1cNow meter uses the immunoassay method, which does not have as high a standard of accuracy as the HPLC method used in the laboratory where SimpleChoice/DTI samples are tested.
At about the same time I got one A1C test each from DTI, A1cNow, SimpleChoice, and Unilab. The results varied from 5.7 percent to 7.0 percent. Obviously, this business is not as accurate as it needs to be.
In addition, ease of use is a factor that impinges on accuracy. Every time I used an A1cNow meter I felt it was easily subject to operator error. A sample that is either too large or too small will cause an error. It is also highly sensitive to temperature and light. The meter must be at room temperature (64° F to 82° F) for at least one hour before opening it. You also have to test within 15 minutes of opening the pouch containing the meter. Then, you can’t disturb the meter while waiting for the test results.
On the other hand, the DTI/Simple Choice kit seems to me to be a model of thoughtful, intelligent design. The instructions are straightforward. The kit is simple to use and is subject to only the most gross operator error.
Accuracy is the most important factor, but it isn’t everything. As with blood glucose testing, time, sample size, and cost are three more important factors.
How fast you can get your results is of great importance if you have to make changes in your diet and treatment. Here, the A1cNow test is the clear leader. You get your result in eight minutes. The DTI/SimpleChoice test results should come back to you in a week. A local lab test will take much longer, since you normally have to have an appointment with your physician to get the prescription and then wait for the results to get back to you.
DTI/SimpleChoice requires the smallest sample size, 5 microliters of blood or 5/1000th of a milliliter. The A1cNow requires twice as much. Unilab requires one tube of blood, about 4 milliliter.
Your cost will depend largely on whether you have health insurance or not. Insurance will normally cover the cost of lab testing. Few, if any, insurance plans will reimburse the costs of the alternatives. Diabetes Technologies Inc. offers the Accu-Base Hemoglobin A1C Sample Collection Kit for $21.95 plus $3.85 shipping. Retailers set the price of the SimpleChoice A1C test kit, but it is generally about $25. Metrika sells an A1cNow meter for $22.95.
If you don’t have health insurance, a lab test will be much more expensive than the alternatives. Quest charges individuals $99.25 for an A1C blood draw and test.
Whether to continue to have your A1C checked at a local lab or to use an A1C meter or test kit depends a lot on your particular circumstances. You need to weigh the pros and cons. However, don’t think so much about it that you forget to get tested one way or another.
What does the A1C Measure?
When you get an A1C test, you are, of course, checking your blood. That is the simplistic answer. It is, however, a lot more than that.
The A1C is the new standard name for blood level tests that have variously been called glycated hemoglobin, glycohemoglobin, glycosylated hemoglobin, HbA1c. It describes a series of stable minor hemoglobin components formed slowly and nonenzymatically from hemoglobin and glucose.
The rate of formation of these hemoglobin components is directly proportional to the glucose level. Your A1C level provides you with your glycemic history of the previous 120 days, since that is the average life span of your red blood cells.
The A1C “most accurately reflects the previous 2-3 months of glycemic control,” according to the American Diabetes Association's Position Statement on Tests of Glycemia in Diabetes.
That’s the establishment position. But even this careful statement is subject to some caveats.
The A1C reflects the last four months, says Dr. David Goldstein, a pediatric endocrinologist practicing in Columbia, Missouri, and the previous chair of the National Glycohemoglobin Standardization Program. “But it reflects the last month much stronger than four months ago. It is a weighted average. About half of what goes into the glycohemoglobin is the past month or so.”
Ron Sebol, a retired electronics engineer living near Columbia, Maryland, says, however, that the A1C measures an even shorter period. A standard text, Ellenberg & Rifkin’s Diabetes Mellitus, bases the chapter on the kinetics of glycation on differential equations that were computer modeled and presented a graph of a square wave response. Ron emphasizes that the equations and the computer model were validated in a test conducted in a hospital setting and are therefore beyond mere conjecture.
“That is how A1C reflects a stepwise abrupt change in average blood glucose,” he writes me “To an electronics engineer, but not the doctors who wrote the paper, the graph was instantly recognizable as an exponential of the same sort as describes charging a capacitor via a series resistor.”
This means, he says, that in only one week the A1C has changed more than 50 percent. By the 47th day it has changed 99 percent.
There is less dispute over how often we should get our A1C checked. If you are meeting your treatment goals and have stable control you should be tested twice a year, the ADA says. When you are not meeting your goals or you change your therapy, you should test four times a year.
Your goal should be an A1C of less than 7.0 percent, the ADA says. Many people believe that this level is too high. The Diabetes Control and Complications Trial (DCCT) showed that near normalization of glycemic levels prevent complications. That trial set an A1C level of 4.0 percent to 6.0 percent as normal. The United Kingdom Prospective Diabetes Study (UKPDS), the other major diabetes study, set the normal level as less than 6.2 percent.
Dr. Richard K. Bernstein, an endocrinologist in practice in Mamaroneck, New York, and author of Dr. Bernstein’s Diabetes Solution, is perhaps the sharpest critic of the ADA’s treatment goals.
“I feel diabetics are entitled to the same blood sugars as non-diabetics,” he wrote me recently. “This means that an appropriate A1C would be in the vicinity of 4.5%…This numbers happens to be the value for blood sugar that I’ve seen over and over on the non-diabetic meter salesmen that visit my office. It also happens to be the value of my own A1C. Since an A1C of 6% corresponds to an average blood sugar that is more than 50% above normal, it certainly does not meet the guidelines for good health and longevity.”
Many people question Dr. Bernstein’s goals. That includes two contributors to this article, Ron Sebol and Dr. Little.
“The DCCT data stand as a contradiction,” Ron writes. “I posted some calculations based on the DCCT and in them found that a person with an A1C of 6 had a 98.5% chance of being complication free over a 15 year term. The dramatic shift in eating life style that Bernstein requires has virtually no payoff in risk reduction since going from 1.8% risk to zero is not worth the price in life style. Add to that with strategic use of supplements known to be able to reverse glycation damage, and you have essentially a zero risk of complications with a diet much more mainstream than Bernstein insists is needed. Alpha lipoic acid, acetyl-L-carnitine, benfotiamine, and pycnogenol can, I am convinced, make an A1c of six or even 6.5, entirely safe. What cures also prevents. It is the difference between having to limit to Bernstein’s 12 grams of carb per meal vs. limiting to 50 that is at issue.”
There is too much risk for hypoglycemia with the level that Dr. Bernstein recommends, Dr. Little believes. “Certainly a normal range goal would be appropriate for some people with diabetes but not all. One has to weigh the risks and benefits and there are certainly risks of a person with diabetes having too low an A1C. A person without diabetes and one with diabetes may have the same A1C but their blood glucose swings would be different; the person with diabetes would be a much higher risk of hypoglycemia.”
Robert Ray, Ph.D., the CEO of FlexSite Diagnostics, just sent me this thoughtful message that I need to include here to update and clarify this article.
“We have recently had some unfavorable response to your article,” he begins. “I would like to point out some inaccuracies and I am sure you will want to correct them.
“On the second page of the article you say, ‘I did not evaluate these two kits [BioSafe Collection Kit and FlexSite A1c At-Home(R)], because their laboratories didn’t pass the rigorous requirments of the National Glycohemoglobin Standardization Program (NGSP).’
“I can’t speak for BioSafe, of course, but as for FlexSite this is not true. It is true that we have not recently applied for NGSP certification of our lab. It is not true that we ‘didn’t pass.’ For several years in the past we successfully went through the NGSP certification process and certified our laboratory. However, since the technology we use (Roche Unimate immunoassay) is routinely certified by NGSP by Roche, we felt that certifiying our lab as well was redundant so we elected to rely on the Roche certification of the method/instrument. As Randie Little points out in your reference to her comments in your article, ‘98 per cent of the labs in this country use certified methods;’ i.e. their labs are not certified. The fact that of the thousands of labs in the country only nine are level I certified tells us that it is not practical for all labs to be certified.
“We supplement Roche’s certification with participation in routine proficiency testing as do all licensed labs in the U.S. We use the CAP (College of American Pathologists) proficiency survey samples which the NGSP lab analyzes for A1C. The survey is conducted twice a year and over 2800 labs in the country participate using one or more of the 20+ methods/instruments currently in use. In the latest survey of three unknown samples the results from our lab are:
Sample NGSP FlexSite FlexSite
“These results for FlexSite, both on the liquid survey sample and the dried sample, are well within the accepted limits of accuracy practiced by labs around the U.S. and speak to the accuracy of our lab.
“In addition, we occasionally test ourselves against NGSP samples. Not long ago we purchased about 40 samples from the NGSP lab with their results and ran them through our lab both as wet and dried samples. The results correlated to better than 99 percent accuracy. As a result, we are confident that our lab is accurate within the expected limits for this test. We do it this way because it is much more economical and it gives us the same results. However, we don’t show up on the NGSP website because of this approach.
“On the overall question of accuracy of the various methods I would advise you to use considerable caution when trying to conclude which methods are accurate and which might not be. A1C analysis is a very challenging analytical chemical problem and the chemists who have worked out the 20+ commercial methods reported in the CAP survey have all worked hard to achieve the best accuracy possible for their method. Every method has its strengths and weaknesses, its accuracies and inaccuracies. Anyone who says a particular method is the most accurate method is likely to be selecting facts to suit his/her agenda. The comment from your endo that the differences between the methods you tested are unacceptable is understandable, but do not reflect the realities of A1C testing. On the other hand the NGSP has done a commendable job of trying to bring the various methods into the best possible agreement. They have made considerable progress. Their ongoing collaboration with the IFCC (International Federation of Clinical Chemists) to develop a ‘reference’ method is making good progress and will help bring the various methods that much closer together.
“In the mean time you make the points that it is the trend from a given method that is valuable and that A1C testing is the key test. We couldn’t agree more.”
This article originally appeared on mendosa.com, October 28, 2003.
Go back to Home Page
Go back to Diabetes Directory