New haemoglobin A1c: the way it is reported is about to change…
While the world struggles to provide care for increasing numbers of persons with diabetes, there is ongoing discussion behind the doors of an international committee about the name and unit in which the most commonly used measurement of diabetic glycaemic control – haemoglobin A1c (HbA1c) – is to be reported in the future.
Glycated haemoglobins are formed when glucose in the blood spontaneously reacts with the amino groups of the haemoglobin protein. Haemoglobin consists of two α and two β chains. The amino acids that undergo nonenzymatic glycation are the N-terminal valine group and free ɛ-amino groups of lysine residues. Approximately 60% of the glycation happens at the N-terminal valine. Such glycated proteins are quite stable and the amount formed correlates to the exposure to different concentrations of glucose over the previous 2–3 months.
The landmark Diabetes Control and Complications Trial (DCCT) in persons with type 1 diabetes used high-performance liquid chromatography (HPLC) methods to separate the glycated HbA1c from the rest of the haemoglobin A0 molecules and to calculate the percentage of HbA1c.1 The trial proved conclusively that a sustained lowering of the HbA1c concentration led to significantly less microvascular complications in the patients treated more intensively and firmly established HbA1c measurement as a critical measure of diabetes control. The process for standardization of HbA1c measurement within the DCCT was subsequently widened and the USA developed a national glycohaemoglobin standardization programme (NGSP). Since July 1996 the NGSP has been standardizing HbA1c assays in laboratories throughout the USA using the BioRex 70 (Bio-Rad Laboratories, Richmond, CA, USA) HPLC system and the NGSP has developed global reach by being adopted for the standardization of the HbA1c assays and HbA1c analysers of most commercial multinational manufacturers. Independently, the Japanese Diabetes Society and Sweden developed separate standardization schemes. These schemes use more modern HPLC methods than the NGSP standardization and the resulting HbA1c values are between 0.2% (Japan) and 1.1% (Sweden) lower.
Most pathology laboratories offer HbA1c assays and patients often have tests at different laboratories; occasionally one obtains very different HbA1c results for the same patient. Why does this occur? The different types of HbA1c assays (e.g. affinity chromatography, HPLC, cation exchange HPLC and immunologically based HbA1c assays) will produce slightly different HbA1c values, which are further affected by the assay imprecision, the standardization method and the method-specific interferences. There is no argument that HbA1c is a critical measure in diabetes care. Most important international diabetes organizations now recommend HbA1c targets. Many recommend a HbA1c level below which there is only a minor risk of developing long-term microvascular and macrovascular complications of diabetes (generally 6–7%); some organizations also recommend a target above which the treatment regimen should be changed to improve glycaemic control (one such recommended level is 8%). The difference between HbA1c levels indicating adequate control, and levels where the treatment should be changed, is very small. A difference of 1% in HbA1c equates to a difference of approximately 2 mmol/L in mean glucose over the last 3 months. It is therefore vital that all HbA1c laboratory assays produce the same numerical values and provide the same clinical information for both patients and clinicians throughout the world.
In 1994 the International Federation of Clinical Chemistry (IFCC) formed a working group to improve standardization of the HbA1c measurement. The IFCC working group has made significant progress in standardization by developing a clear definition of HbA1c – the β-N-valine glycated haemoglobin (β-N-(1-deoxy)-fructosyl-haemoglobin). This is the major glycation site of the HbA1 molecule. The IFCC also developed two reference methods based on electrospray ionization mass spectrometry or capillary electrophoresis. In addition they have developed pure haemoglobin A0 and HbA1C standards and have linked the new reference methods to results obtained internationally by clinical laboratories.
So far there has been good international communication and the laboratories in Europe, the USA and other countries, including those in Australasia, have accepted the new standards and the new reference methods. The manufacturers’ assays are already calibrated against these new standards. However, the new IFCC reference system will give measurements of HbA1c between 1 and 2% lower than that previously reported and used in the DCCT for persons with type 1 diabetes and the UK Prospective Diabetes Study for persons with type 2 diabetes.2 The reason for this is the lack of specificity of the original BioRex 70 HbA1c assay. The chromatographic ‘HbA1c’ peak measured has been shown to contain various glycated and nonglycated haemoglobin molecules, such as a uraemic adduct, foetal haemoglobin and other interferences. In clinical practice, such artefacts may obscure the desired correlations between blood glucose control and reported HbA1c in patients with haemoglobinopathies, which are not infrequent in Asian and Mediterranean heritage, and in renal disease, a common diabetes complication.
Having spent considerable time and effort to convince clinicians to use HbA1c values of between 6 and 7% as the target level of good control and HbA1c values of greater than 8% as an indication of poor control requiring clinical intervention, there would be a need to re-educate clinicians and patients with new reference values approximately 2 percentage points lower. The alternative would be to adjust the new reference method by using the correlation equation between the DCCT and IFCC methods to report the new IFCC values as a similar, but more accurate value than the current DCCT/NGSP value. However, we now know that our old laboratory reports were wrong because substances chemically different from HbA1c are contributing to the measured HbA1c. The problem with this approach is that we would be setting a precedent by reporting values that we clearly know are incorrect and in the future would be referenced to an incorrect and archaic method and standardization system.
The IFCC HbA1c Working Group has developed a totally new scheme of reference methods, purified standards and a Global Network of certified IFCC Reference Laboratories, which produce a much more robust and reproducible HbA1c assay but HbA1c values that are different from those currently used. Clearly, a change in which the HbA1c normal and treatment ranges were suddenly adjusted would be confusing to patients and clinicians. One suggestion has been to use a new value called ‘mean blood glucose’. This is based on the understanding that HbA1c should reflect the mean blood glucose. The correlation between mean blood glucose and HbA1c was established from subjects who participated in the DCCT and had multiple daily capillary blood glucose measurements over the duration of the study.3
However, the correlation between mean capillary glucose as determined in the DCCT and HbA1C shows a significant scatter for individual patients. This scatter has led to the theory that some persons glycate their proteins faster than others with published reports identifying them as fast and slow glycosylators. The overall amount of HbA1c formed depends not only on the mean glucose concentration, but also on the relative lifespan of the erythrocytes. This adds another level of complexity to this equation. Therefore the proposed compromise has its own problems and might not solve all the problems raised.
To clarify the correlation between mean glucose concentration and HbA1c levels, an extensive study has been funded in the USA. This study is called the EASD/IDF/ADA Study of Relationship between MBG and HbA1c. It involves 60 patients of diverse ethnic groups in each of 11 centres, 30 each of type 1 and type 2 with stable diabetes control (one-third normal or near-normal glycaemic control, one-third with moderate glycaemic control and one-third with poor glycaemic control). All the HbA1c assays will be carried out centrally by both IFCC HbA1c and DCCT HPLC methods. Patients self-monitor blood glucose seven times daily on at least 3 days per week during the 16-week study and undergo continuous glucose monitoring over 48 h with a glucose sensor (which records an average interstitial fluid glucose every 10 min) at baseline, 4, 8, 12 and 16 weeks. Preliminary data were reported at the IDF meeting in December 2006.
In the meantime, practitioners should be aware that the current reporting of HbA1c is under intensive review. Hopefully, the results of the prospective studies will be available in 2007 and a worldwide agreement with regard to the future reporting of HbA1c values will be reached. Expect to hear more information in the near future.
