Trueness in the measurement of haemoglobin: consensus or reference method?

Background: For the measurement of haemoglobin a reference method exists: the haemiglobincyanide method. However, a Dutch external quality assessment organization does not use this method in the evaluation of trueness of results. The aim of this work was to assess whether trueness was compromised by the use of a consensus value.

Methods: Five Cell Dyn Sapphires (Abbott) in three independent locations were used to measure haemoglobin concentration. Results were compared to the reference method (haemiglobincyanide). Patient samples with a distribution over clinically relevant concentrations (Hb 2.5-10.2 mmol/L) were used next to samples from external quality assessment rounds. Passing and Bablok regression analysis and Bland-Altman plots were used to evaluate any systematic deviation.

Results: Results measured on the Cell Dyn Sapphires deviated significantly from the results obtained with the reference method. Remarkably, consensus results from external quality control samples also deviated significantly from the reference method.

Conclusions: A significant negative bias exists in the measurement of haemoglobin on Cell Dyn Sapphires. Additionally, the consensus value as reported in external quality control assessment also shows an even greater significant negative bias compared to the reference method. As a reference method is available, external quality assessment would benefit from using this method instead of a consensus value to evaluate trueness.