RESUMO
INTRODUCTION: HbA1c is an important part of assessing the diabetic control and since the use of point-of-care devices for monitoring HbA1c is increasing, it is important to determine how these devices compare to the central laboratory. METHODS: One hundred and twenty patient samples were analyzed on the Bio-Rad Variant™II and one POC analyzer (Sakae A1c Gear). Three patient sample pools containing ~5%, ~7%, and ~10% HbA1c levels were run over 20 days. Three reagent lots and three instruments were evaluated for the A1c Gear. RESULTS: The 120 patient samples showed strong correlation (R(2)>0.989) when compared to the Variant™II with means=8.06% and 7.81%, for Variant IIand A1c Gear, respectively. Changing reagent lots or instruments had no impact for the A1c Gear. The ~5%, ~7%, and ~10% pools within-run and between-run imprecision was between 0.87-1.33% and 1.03-1.32%, and 1.41-2.35% and 1.24-1.89% with total imprecision of 1.67-2.35% and 1.61-2.31% for the A1c Gear and Variant II, respectively. The A1c Gear showed a small negative bias (0.25% HbA1c) across HbA1c measurement ranges of <11.5%. This bias was, however, acceptable and not considered to be clinically significant. CONCLUSIONS: The A1c Gear meets the criteria of total CV <3% leading us to the conclusion that the A1c Gear can give results as precise as the laboratory at the POC.