The evaluation of measurement uncertainty of HbA1c and its effect on clinical decision levels

INTRODUCTION: The prevalence of diabetes mellitus (DM) is increasing all over the world. Hemoglobin A1c (HbA1c) is one of the diagnostic tests for DM. But is the HbA1c analysis result accurate and absolute? Uncertainty of measurement is a quality parameter of measurement results, and characterizes the dispersion of the values attributed to a measured quantity. The aim of this study was to estimate the measurement uncertainty (MU) of HbA1c and retrospectively re-evaluate patient results with respect to the estimation of uncertainty and to suggest a solution for results that are close to cut-off level.
METHODS: The results of 10212 patients who had an HbA1c analysis performed in our laboratory in 2016 were retrospectively reviewed. The HbA1c levels were measured using a high performance liquid chromatography method. The uncertainty of measurement of the serum HbA1c level was estimated according to the Eurachem/Co-operation on Traceability in Analytical Chemistry Guide CG 4.
RESULTS: The measurement uncertainty (95% confidence interval) of HbA1c was estimated to be ±4.6%. When measurement uncertainty was taken into account, the acceptable range for the 6.5% value typically used to diagnose DM was between 6.2% and 6.8%. It was observed that the results of 1555 patients were affected by uncertainty values.
DISCUSSION AND CONCLUSION: Medical laboratories must produce the necessary data and analytical results in order to achieve the correct interpretation and use of the results. A test result is not sufficiently powerful without an assessment of its reliability. The interpretation of values that are close to cut-off levels may change when evaluated with the uncertainty of measurement. Therefore, reporting HbA1c analysis results with the estimation of the MU is important to illustrate the true limits and the level of confidence.