Andrew J.M. Boulton, MD, FRCP, Honorary Secretary EASD
Christopher D. Saudek, MD, President American Diabetes Association
Introduction
Following exploratory discussions with officers of the American Diabetes Association (ADA) and International Diabetes Federation (IDF), the EASD sub-committee on glycated haemoglobin standardisation arranged and hosted a meeting of interested parties which was recently held at the new EASD headquarters in Düsseldorf, Germany. The principal aim of this meeting was to consider potential solutions to problems associated with glycated haemoglobin measurement and to explore the possibility of achieving international consensus on the introduction of a new "gold standard" for the measurement of haemoglobin A1c (HbA1c). The 12 discussants in attendance represented the European, American and International Diabetes associations, the American and International Federations of Clinical Chemists and CEN (European Committee for Standardisation: www.cenorm.be). The format of the meeting comprised overviews of the problems in the area by representatives of the American and International Associations of Clinical Chemists, and a review of potential pathways to achieving international agreement on standardisation by the representative of CEN. A general discussion on the problems and their potential solutions followed.
Background
The original observation that normal adult haemoglobin could be separated chromatographically into a major and several minor components was made over 40 years ago. It was later recognised that in one of these minor components, HbA1C, glucose is attached non-enzymatically to the N-terminal Valine of the beta chain of HbA0. In subsequent years, a number of different assays were developed to measure glycated haemoglobin that were bases on the small difference in physical behaviour, consequent upon the addition of glucose to haemoglobin. Early assays suffered from several problems, including poor definition of the analyte, the non-specificity of cation exchange chromatography, the fact that non glycated material comprised up to 40% of the HbA1C peak, poor calibration of techniques and so on.
The Problems
The variability in assay methodology, including non-specific measurement procedures, incorrect calibration and inadequate definition of quantities, has traditionally resulted in poor comparability of glycated haemoglobin results.
In 1996, the American Association of Clinical Chemists (AACC), in collaboration with the American Diabetes Association and the Centers for Disease Control addressed these problems by establishing the National Glycohaemoglobin Standardisation Program, with international representation on its Steering Committee (see Little et al 2001). This program adopted the HPLC reference method used in the Diabetes Control and Complications Trial (DCCT) (Bio-Rex 70) as the Designated Comparison Method (DCM). The approach of the National Glycohaemoglobin Standardisation Program enables manufacturers to establish tracability to the DCCT standards through a national reference laboratory, with primary and secondary back-up laboratories in the USA and Europe. These back-up laboratories use varying methodologies including ion-exchange, HPLC, affinity HPLC, capillary electrophoresis and immunoassay. Calibration for standardisation is based upon the "set-point" used in the DCCT. The United Kingdom Prospective Diabetes Study (UKPDS) incorporated the same standardization method, so that its HbA1c results were precisely compatible with those of DCCT.
The National Glycohaemoglobin Standardisation Program has been successful in establishing reproducibility testing of laboratories in the U.S.A., albeit by solely proficiency methods, and to a standard assay that may not be the most accurate measure of true HbA1c. Over 97% of laboratories in the U.S.A., and many in Europe as well as worldwide, now meet acceptable standards through this program.
Current Assay Developments:
In recent years, a new methodology has been developed in The Netherlands that is based on mass spectroscopic assay of the glycosylation of the N-terminus alanine of the B chain of haemoglobin. The IFCC has developed a reference method with definition of the analyte, development of reference methods, a network of reference laboratories in 3 continents and method comparison. To date, IFCC have performed 6 inter-comparison studies with exchange of haemolysates and inter-lab and between lab CV's of 1-2%. Thus the IFCC network of HbA1c reference laboratories is capable of very precise value assignment. A detailed review of the new method and problems and potential solutions for standardisation of measurements of glycated haemoglobin will be published in Diabetologia within a few months of this report, prepared by Miedema et al. on behalf of the IFCC.
A key issue raised by this method of measuring HbA1c is that it yields normal (non-diabetic) values that are significantly lower than those used by the National Glycohaemoglobin Standardisation Program, DCCT and UKPDS. Thus the non-diabetic reference range is about 3-5% rather than 4-6%, and the International Federation of Clinical Chemists (IFCC) has considered "good" glycaemic control to be 5%, and control requiring treatment alteration >6%.
Clinical and Research Needs
The clinical requirements for an HbA1C test include:
There was agreement that the new mass spectroscopy-based method appears to be more accurate and to reflect the "true" HbA1c. There was some discussion about whether this specific method will actually predict complications in the exact way that the HbA1c method used in the DCCT and UKPDS did. The consensus was that there is no reason not to think so.
International standardisation of HbA1c measurement would hold significant advantages, and it would be especially disruptive if major differences in reference ranges develop due to differing methodologies. While national societies have published guidelines for diabetes management based on DCCT-referenced HbA1c readings, and while it is understandable that these guidelines may differ based on differences in philosophy of guideline writing or of diabetes care, grossly different reference ranges would introduce highly undesirable confusion.
There was therefore much discussion as to how the HbA1c should be reported, i.e. whether new values should be reported directly (Option 1 above) or whether the new method should be reported out with currently accepted reference ranges (Option 2). Recognizing the value of the new mass spectroscopy method, there was little support for continuing affinity chromatography as the reference method (Option 3).
The European representatives, on the whole, were most eager to have the new, more precise method utilized as the gold standard. A point was made, however, that this year's reference range for the new method may change again as further methodologic refinements are made.
The ADA representatives expressed the opinion that as 70% of Americans with diabetes are now familiar with the HbA1c levels based on the National Glycohaemoglobin Standardisation Program. While endocrinologists and diabetes specialists could perhaps understand and adapt to a new reference range, generalist physicians and particularly patients would find it much more difficult to do so. To change the normal values so significantly, then, may cause a major setback in the effort to teach physicians and people the use HbA1c as a primary index of control.
The IDF reminded the group that of all those with diabetes worldwide, probably less than 50% had ever had an HbA1c estimation. The question of cost of assays is a key point to consider when advising or adopting a recommended methodology.
The European representatives discussed methods by which harmonisation of standards can be achieved. One method is to organise a CEN workshop that can include European and Non-European experts. This can ultimately result in agreed European or International standards.
Conclusions: