Key:
NZ    Nahida Zaman
JL    Dr Joyce Lee

NZ    You are listening to Nahida Zaman with Diabetes Radio.  My guest today is Dr Joyce Lee.  Dr Lee, hello and a warm welcome to Diabetes Radio.

JL    Thank you for having me.

NZ    Dr Lee, you specialise in Paediatric Endocrinology and are Assistant Professor within the Department of Paediatrics and Communicable Diseases at the University of Michigan’s CS Mott Children’s Hospital.  Can you please give the Diabetes Radio listener some background into your research interests?

JL    I am a Paediatric Endocrinologist and both clinically and for my research interests I study children with obesity and diabetes.

NZ    And Dr Lee, what is the focus of your clinical work?

JL    I am very interested in trying to identify whether there are certain tests or predictors that can help determine which kids are at greatest risk for developing type II diabetes.

NZ    Doctor, you recently published a paper in the November 2011 issue of the ADA journal Diabetes Care.  The paper was titled Evaluation of Non-Fasting Tests to Screen for Childhood and Adolescent Dysglycemia.  The paper received international recognition and was picked up by both the medical and mainstream news as it showed that the American Diabetes Association’s preferred testing method failed at identifying children with diabetes.  Can you please walk our listeners through the history, why and how did haemoglobin A1C become a testing method recommended by the ADA?

JL    Haemoglobin A1C is a marker that is actually traditionally used to measure control of glucose in patients with diabetes so the higher your A1C the higher your blood sugars but more recently it has been used to actually try and screen or identify individuals with diabetes so they may not know that they have diabetes yet but they could have this test to potentially find out if they are at risk for having it.  A1C is simpler to measure because you do not have to fast or not drink or eat anything for eight to 12 hours before you did the test.  The other measures that have traditionally been used are either a fasting glucose, again where you have to fast for eight to 12 hours, or a glucose tolerance test which also requires fasting and so because of convenience issues, this idea that you have to have a patient come back to get the test, A1C was more of a more preferred option for that reason.

NZ    Can you please give us some background on your study?

JL    We looked at a sample of 254 kids who were overweight or obese so that means the body mass index greater than the 85th percentile, we are interested in that group because they are a group that is actually recommended for diabetes screening by the American Diabetes Association as well as the American Academy of Paediatrics.

NZ    Doctor, what were the objectives of your study?

JL    The objective of our study was to look at non-fasting tests, tests for which you do not have to fast to try and identify which kids are at greatest risk for having what we call dysglycemia so that is either having diabetes, diabetic range glucose levels or pre-diabetes which is potentially having higher glucose levels that put you at risk for diabetes which is in contrast to having absolutely normal glucose so in terms of our hypotheses we wanted to look at a variety of tests including haemoglobin A1Cs but as well a one hour glucose challenge test, so that is a test where an individual can have eaten their breakfast or lunch and they just drink 50 grams of glucola and they have their glucose measured 60 minutes later.  We also did a random glucose where they may have had a meal but we just check their blood sugar at that time and we also looked at other markers including a urine dipstick so they essentially provided a urine sample.  We wanted to look at all these different tests to see whether they would be predictive of pre-diabetes or diabetes based on the Gold Standard test.  The Gold Standard test is a two hour glucose tolerance test so again they have to come in fasting for at least eight to 12 hours, they drink 75 grams of glucola and then two hours later they have a glucose drawn so what we did was we had these kids come in for two different visits to our clinical research unit and in the first visit they came in and had the Gold Standard test and that test identified them as having dysglycemia which is the pre-diabetes or diabetes or having normal glucose tolerance and then we had them come in within one to three weeks and have some of those non-fasting tests and we essentially compared the test performance.

NZ    And what were the results of the study?

JL    So in terms of our results, what we found was that there were tests that performed quite poorly in children and two of the tests that performed quite poorly included the haemoglobin A1C which is now the preferred method for identifying kids with diabetes by the ADA so it did not do a good job discriminating which kids might have dysglycemia versus which do not and another test that did not perform so well as well was the urinalysis but what we did find was that there are a couple of tests that are more promising so the random test for glucose and the one hour glucose challenge test, those tests had better test performance and could conceivably be used in a clinical setting to try and triage or identify which kids need to get more definitive testing for dysglycemia.

NZ    Doctor, were there any limitations to your study?

JL    Yes there are some limitations so we drew our sample from a convenience sample of children from South East Michigan so this may not be generalisable to the rest of the US or UK population.  As well we did have reasonable participation of black adolescents but for example we did not have a large proportion of Hispanic or Asian adolescents so we could not assess how well those tests were in those sub-populations.

NZ    What conclusions should be drawn from your results?

JL    So I think one of the most important, I think, policy conclusions is that the American Diabetes Association came out recommending use of haemoglobin A1C to diagnose diabetes for adults as well as in children.  When these recommendations came out they principally did this using data from adult studies but they did not have much information about test performance on children at the time but based on this study that we published and another study that we published in the past using national data from the US, what we found is that haemoglobin A1C really has poor test performance and does not identify or discriminate well which kids have dysglycemia so we would caution the use of this test for screening and for diagnostic purposes in children being screened or diagnosed with type II diabetes.

NZ    Does this highlight the need for improved non-fasting testing strategies?

JL    I think it does and as well I think there also needs to be better risk scores, clinical risk scores for children so what I mean by that is that in adults there have been a lot of studies that have just looked at demographics like your age, your body mass index and other characteristics so for example if you have a family history of diabetes, if you have signs of having high insulin or insulin resistance so they use those clinical scores which can just be based on potentially a history and a physical exam to try and predict which individuals might be at risk of having diabetes so what I think we need in the paediatric population is a similar type of study where we look at clinical characteristics of children, we maybe combine this with the non-fasting tests and the more promising ones such as the one hour glucose challenge test or the random glucose and we see whether those tests in combination with the clinical risk score can really improve detection of dysglycemia in adolescents.

NZ    Other paediatric organisations like the American Academy of Paediatrics have not endorsed the use of HbA1C yet.  Why has this been the case and how do you see the results of your paper impacting the recommendations of groups such as the AAP?

JL    I have enquired with colleagues across the country as to why they have not endorsed it yet.  I am not clear on the extent at which Paediatricians are aware of this new guideline because it was published in the American Diabetes Association Diabetes Care Journal but that is not a journal that is probably widely read by the general Paediatricians or standard practitioners who care for children in clinical practice so I think there is probably some time for those guidelines to disseminate down and then for those organisations or those individuals to come together and make a determination about whether or not this test is appropriate.  I think based on our studies, we initially had thought that A1C might be a reasonably predictable variable for identifying adolescents with dysglycemia but we did not find that this was the case so I think what I am hoping is that the impaired data that we have used or impaired data that we generated to identify A1C as a core identifier, a marker for dysglycemia in adolescents really needs to be taken into consideration when the, for example, AAP comes together to make a determination on this recommendation.

NZ    Doctor, do you see your results as a potential game changer?

JL    I do in that it could really lead to a different approach to screening in children.  Traditionally what the ADA and the AAP have recommended is essentially a fasting plasma glucose or a two hour glucose tolerance test.  Now the ADA have come out and said no, we refer A1C to those other tests but I think based on the previous data in children, my hope is that these non-fasting tests, the random glucose or the one hour glucose challenge test and potentially clinical risk score which we are developing now would be a first line for Paediatricians and General Practitioners to evaluate those at risk for diabetes.

NZ    How do you see your data helping to drive future recommendations about screening for diabetes in children?

JL    Our studies are some of the first studies to provide this empiric data and test performance in children so again I think our hope is that the individuals who write guidelines for screening steer clinicians away from the tests that do not respond so well and steer them towards the tests that do seem to have better test performance based on the empiric data.

NZ    Dr Joyce Lee, thank you so much for taking the time to speak with us today.