A comparison of three methods of measuring plasma glycosylated hemoglobin

ABSTRACT

Background: The current clinical dilemma of diabetes mellitus is prevention of micro- and macrovascular complications. These complications are directly linked to the degree of hyperglycaemia in diabetics. For long-term control of the glycaemic state, measurement of glycohaemoglobin in blood is essential, and glycohaemoglobin harmonization is encouraged for significant inter-laboratory variability. Objectives: To evaluate the performance of plasma glycosylated hemoglobin (HbA1c) assay on Primus PDQ and Bio-Rad D10 analyzers with respect to imprecision against the HbA1c goal, recovery and correlation to the HbA1c test on Cobas Integra 800 automated analyzer. Methods: PDQ employs high performance liquid chromotography (HPLC) boronate affinity chromatography. D10 uses HPLC cation-exchange chromatography. Cobas Integra 800 uses an immunoturbidimetric method. Imprecision studies were performed using patient samples and commercial control sera. Each sample was assayed 20 times and run every morning and afternoon for 5 consecutive days to evaluate between-run data. Recovery was assessed using a patient’s sample with low HbA1c level, admixed with another sample containing high HbA1c level. Comparison was made using 60 patient samples on PDQ and D10 using Cobas Integra 800 as the reference system. Ten venous samples from each group of patients with thalassaemia and uraemia were assayed for HbA1c on Cobas Integra 800, PDQ and D10. Results: The within-run coefficients of variation (CVs) on patient samples were 0.5% and 0.6% for PDQ and 1.1% and 1.3% for D10. The between-run CVs were 1.0% and 0.8% for PDQ and 0.3% and 1.2% for D10. PDQ yielded CVs of 1.1% and 0.7% for level 1 (normal) and level 2(abnormal) controls respectively while D10 demonstrated CVs of 0.5% for each level of controls. The between-run CVs for normal and abnormal controls were 1.7% and 0.8% respectively for PDQ, and 1.0% and 0.7% respectively for D10. Mean recovery for PDQ and D10 were 99.7% and 101.7% respectively. Both PDQ and D10 correlated well with Cobas Integra 800 [Cobas800]. HbA1c method by PDQ correlated well with that by D10. Conclusion: Both the HPLC systems demonstrated acceptable data, good recovery and are comparable with our current operating system (Cobas Integra 800). Both these systems appear to be satisfactory analytical alternatives to the Cobas Integra 800.