More Diabetes Answers

Can HbA1c be used to diagnose diabetes?

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:

http://gomeworld.typepad.com/morediabetesanswers/2010/01/what-is-glycosylated-hemoglobin-or-hemoglobin-a1c-hba1c.html

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.

Can a bleeding ulcer cure diabetes!?

Just the other day one of my patients thought that her diabetes had been "cured" after being in hospital with a bleeding stomach ulcer.  When she came out of the hospital her HbA1c, which had previously been in the 8.0 -9.0 percent range, came back at 4.9 percent!  She was delighted until I explained the reason.  She had lost several pints of blood from her bleeding ulcer and so was quite unwell with severe anemia when she was admitted to hospital.  In order to treat her she had been treated with a transfusion of several pints of blood.  That blood had been donated by one or more non-diabetic individuals.  When this was mixed with her own blood the average amount of HbA1c dropped right into the normal range because the transfused blood had very little glucose attached to it.  Over the next 10 to 12 weeks I'm afraid that my patient's HbA1c will rise back up to the 8-0 percent range as glucose gets attached to her own hemoglobin and to the hemoglobin from the transfused blood.

To understand more about HbA1c look at page 83 in The Diabetes Answer Book or check out this post:

What is glycosylated hemoglobin (or hemoglobin A1c, HbA1c)?

http://gomeworld.typepad.com/morediabetesanswers/2010/01/what-is-glycosylated-hemoglobin-or-hemoglobin-a1c-hba1c.html

What is glycosylated hemoglobin (or hemoglobin A1c, HbA1c)?

As I said in The Diabetes Answer Book (page 83):

Hemoglobin is a protein inside all of your red blood cells.  It helps carry oxygen to your tissues to keep you healthy.  Every day your body makes a few million new red blood cells in your bone marrow.  These get pumped round and round your body for the next two or three months.  When the red blood cells get old, your body destroys them in your spleen.  Since the hemoglobin is floating around amongst the glucose in your blood, a small amount of glucose gets attached to the hemoglobin.  Hemoglobin with glucose attached to it is called glycosylated hemoglobin or hemoglobin A1c (HbA1c).  Brand new red cells have no glucose attached to them, but the longer they stay in your bloodstream, the more glucose gets attached.  The higher the glucose level in your blood, the more HbA1c you form.

We can measure what percentage of your hemoglobin has glucose attached to it.  This tells us what the average glucose level was in your blood for the previous ten to twelve weeks.  To give you an idea of how to use this information, a non-diabetic person's blood glucose will vary between about 20 and 160 mg/dl (3.9-8.9 mmol/L) throughout the day.  Over time about 4 to 6 percent of the person's hemoglobin will get glycosylated.  So the normal range for HbA1c is about 4 to 6 percent.  Now, imagine the worst possible blood glucose control.  If someone let his or her blood glucose average remain at 400 to 600 mg/dl (22-33 mmol/L) for months on end, the HbA1c could rise above 20 percent.  I have never seen HbA1c higher than 25 percent, but it can almost get to that level if someone leaves his or her diabetes completely untreated for long enough.

Doctors pay so much attention to HbA1c that it is important for you to understand what it is and how to interpret it.  I will refer back to this in other blog posts. 

Does insulin glargine (Lantus) cause cancer?

What a great, if scary, question!  I'm afraid the answer that I give you is not as straightforward as you would like.  The answer is, "Probably not, but we don't know for sure."  You might wonder why someone even asked such a question.  I mean if you have diabetes then insulin is just the hormone that your body doesn't make enough of.  Taking insulin should be good for you, not bad for you, right?  Absolutely.  Insulin has been a miraculous advance in the treatment of diabetes and the millions of people who take insulin are much healthier because of it.  So why wonder if insulin glargine (Lantus) might cause cancer?  Let me tell you what is and is not known.

By looking at large populations of people around the world researchers have known for well over a decade that people who are very overweight and people who have type 2 diabetes have an increased risk for several types of cancer (cancer of the colon, cancer of the pancreas and cancer of the breast).  It is not known why.  One possible reason is because people who are overweight and people with type 2 diabetes have a lot more insulin in their body and insulin stimulates the growth of cells, so maybe it stimulates the growth of abnormal cancer cells.

Researchers have done experiments in laboratories showing that insulin can stimulate the growth of breast cancer cells.  Insulin glargine (made by the Pharmaceutical Company Sanofi-Aventis) is an analog of human insulin.  This means that the scientists have modified the natural human insulin molecule by removing certain parts of the molecule (amino acids) and adding extra ones.  This is done all the time to create new drugs to treat all kinds of illness.  People who take steroid inhalers for asthma or steroid pills to treat their arthritis are taking analogs of the natural steroid hormone, cortisol, that we all make in our bodies.  So insulin glargine (Lantus) is a modified kind of insulin that is a bit different from naturally occurring human insulin.  Well, it turns out that in laboratory studies insulin glargine stimulates cell growth even more strongly than natural human insulin, so it is certainly possible that people taking Lantus to treat their diabetes could have a higher risk of cancer.

What's the best way to know for sure?  By far the best way would be to find a large number of people (several thousand at least) with type 2 diabetes who need insulin treatment and ask them if they would let you randomly assign them to use insulin glargine (Lantus) or another insulin (such as human NPH insulin, perhaps).  It is likely that they could get similar improvements in their blood glucose levels and could have their diabetes controlled with either insulin for many years.  Then, we would keep track of how they were doing for 5, 10, 15 years or more.  During that time some of those thousands of people would get cancer and some of them would die from cancer or from other causes.  Because the people were randomly assigned to either Lantus or NPH at the beginning, if we found a different rate of cancer or death between the two groups of people that would be very powerful evidence.  We would know with confidence whether one type of insulin gave less risk of cancer than the other.  Unfortunately that study has not been done.

There has been a small study where 1017 patients were randomized in this way and during four years of follow-up there were 20 new cases of cancer in the people taking glargine insulin (Lantus) and 31 cases in patients taking human insulin.  In other words the patients taking Lantus insulin seemed to be LESS likely to develop cancer than those taking human insulin.  The trouble was that there were too few people in the study and they weren't followed for long enough for us to know for sure.

Well, if no one has done a larger and longer study like this (a randomized controlled trial) how else might you answer this question?  You could try looking to see what happened to the many thousands of people who have been started on different types of insulin over the past decade or so.  This is called a cohort study.  For those of you who want to learn more details about the advantages or disadvantages of cohort studies compared to randomized controlled trials (RCTs) the entry in Wikipedia is very helpful:

http://en.wikipedia.org/wiki/Cohort_study

 One of the problems with a cohort study is that you can never be sure why particular patients were started on particular types of insulin.  It certainly wasn't random!  And so you never know whether what you find is because of the type of insulin the person is taking or because of some other factor that you don't know about.

Here's a great example of why you should be very cautious about drawing conclusions from cohort studies.  After the end of World War 2 the number of households that had a radio in them went up dramatically over the next decade.  During that same decade the birthrate also went up dramatically.  If you were naive you could conclude that having a radio in your house increases the risk that you will get pregnant!  But if you had done a randomized controlled study after World War 2 where half the households had been given radios and the other half were not given radios I  seriously doubt that you would have seen a difference in the birth rates in those two groups!

When large groups of people (over 100,000) with type 2 diabetes who were taking insulin in Germany and in the United Kingdom were studied the number of people who developed cancer while taking insulin glargine were slightly lower than those taking other types of insulin.  But when the researchers divided up the patients into those who were taking a small, medium or large dose of insulin, and when they looked at those who took only one type of insulin compared with those who took more than one type of insulin they concluded that patients who took only insulin glargine, and took it in high doses, were at higher risk for developing cancer.  This has set off a flurry of angry name-calling in the medical journals with various experts criticizing each other for the methods they have used and the conclusions that they have come to.  Truthfully the only way to know for sure would be for some important regulatory agency like the Federal Drug Administration (the FDA) to demand that a large randomized controlled trial be done.

So where does this leave you, if you are a someone taking insulin glargine (Lantus)?  I think that with what we know today for sure the benefits of taking it far outweigh the risks.  If some group of researchers do decide to do a proper randomized controlled study to see whether taking insulin glargine (or any other kind of insulin) causes a lower or higher risk of cancer that would be great.  I would love to know the answer.  But for now I suggest that you continue to take the insulin you are taking and focus on things that you CAN do to lower your risk of cancer in the future.  Don't smoke.  Don't drink alcohol to excess.  Eat a healthy diet. Get regular exercise.  Do something fun every day.  Try to keep your stress level down.