Clinicians use glucose and hemoglobin A1c (HbA1c) levels to determine whether or not their patients are in control of their diabetes. Just how precise are these values and the established guidelines? That was the question this informative discussion tried to answer.

Mitchell Scott, PhD Washington University, gave an overview of the measures of laboratory test precision and the types of error that contribute to the inaccuracy of results. The coefficient of variation (CV) is a measure of between run precision. It is calculated by dividing the mean by the standard deviation and is the most useful value for clinicians to consider. At his lab (Barnes Jewish Hospital), a glucose value of 88mg/dl has a 1.4% CV and a confidence interval (CI) of 85–91. A glucose of 283 mg/dl has a 1.1% CV and a CI of 276–290. In his experience, the glucose meters used in the hospital are less accurate than the main lab, with an 8.4% CV, a range 153–210, and CI of 139–195.

There are several types of errors that contribute to the overall error of the result. Biological variability includes within patient and between patient variability. The biological variability alone of fasting plasma glucose is quite high (6.9% CV), which means that a glucose value of 126 mg/dl has a CI of 103–149 (Sacks et al, Clinical Chemistry 2002; 48:436). Additional routine sources of error include sample misidentification, inappropriate sample storage, drawing at the incorrect time, phlebotomy technique, drawing in incorrect tubes, and IV contamination. Processing and storage delays can cause major problems: if samples are not centrifuged promptly, glucose in the sample decreases 7% per hour at room temperature. Dr. Mitchell strongly suggests that clinicians centrifuge their samples promptly, repeat any borderline glucose values, and call the lab if lab results do not correlate with their clinical observations.

David B. Sacks, MB, ChB, FRCPath of Harvard Medical School presented changes to the HbA1c test and the global implications for clinical practice. The concentration of glycated hemoglobin (GHb) represents integrated glucose in blood over a 6–8 week period. It is therefore a very useful criterion for glucose control and in most major clinical trials it is used as a predictor of the risk of complications. The ADA recommends regular monitoring of GHb, but there are 30 different tests available worldwide.

The objective of the National Glycohemoglobin Standardization Project (NGSP) is to standardize the HbA1c result (a measure of GHb) so clinicians can easily compare results and accurately assess risk. The NGSP works with manufacturers to calibrate instruments, certify methods, and certify individual labs via proficiency testing. This effort has led to improved reproducibility and consistency: in 1993, 50% of labs were reporting HbA1c as a measure of GHb; in 2006, 99% are reporting HbA1c. NGSP methods are used in the European Union, India, China, Australia, Latin America, North America, South Africa, and Russia.

Yet there is a potential source of confusion for clinicians. The IFCC (International Federation of Clinical Chemistry) has also set out to develop an HbA1c reference method and primary reference materials. As it turns out, the IFCC values are lower than NGSP values. For example, a 7.0 NGSP = 5.3 IFCC. There is a formula that can be used to convert between the two, but this makes it more difficult to accurately determine a patient's glycemic control. Although the guidelines may be changing in the future, a consensus has not been reached at this time. It is therefore important that clinicians continue to follow their current guidelines for HbA1c when monitoring their patients.