Standardization of glycated haemoglobin assays throughout the Northern region of England: a pilot study

We describe a pilot study designed to assess the comparability of measurements of glycated haemoglobin among 15 laboratories in the North of England. We also evaluated a means of improving agreement and aligning results by referencing locally measured values to those obtained by the central biochemistry laboratory of the Diabetes Control and Complications Trial research group. Blood samples from 50 diabetic and non-diabetic subjects were analysed for glycated haemoglobin content in the participating and reference laboratories using a variety of routinely available methods. The mean CV for these results was 15.3% (95% confidence interval 14.0% to 16.5%). Using a regression equation relating a subset of seven of these results to their assigned reference values, a glycated haemoglobin index was calculated for all the other samples distributed. The mean inter-laboratory CV improved to 4.6% (95% confidence interval 4.0% to 5.1%), p<0.0001. The percentage bias of results from the reference method also improved from 15.1% (95% confidence interval 9.4 to 20.1) to 4.67% (95% confidence interval 4.05 to 5.25) after alignment, p<0.001. This study demonstrated that substantial method related differences between reported glycated haemoglobin results exist. These can be reduced using a simple calibration strategy in which data are correlated to an established method with associated extensive clinical interpretive value as established by the DCCT.