# Comparative Performance of a Routine Immunoturbidimetric Glycated Hemoglobin Assay and High-Pressure Liquid Chromatography Method

## TL;DR

This study compares two methods for measuring HbA1c, a blood test for diabetes, and finds they agree well but one tends to underestimate high values.

## Contribution

The study evaluates the clinical suitability of an immunoturbidimetric method for HbA1c testing in a local setting.

## Key findings

- Both methods showed satisfactory precision with coefficients of variation under 3%.
- The immunoturbidimetric method systematically underestimated HbA1c levels, especially at higher values.
- A strong correlation (r = 0.88) was found between the immunoturbidimetric and HPLC methods.

## Abstract

Introduction

Hemoglobin A1c (HbA1c) testing is used for the diagnosis of diabetes and glycemic monitoring. This study compares an immunoturbidimetric assay with high-performance liquid chromatography (HPLC) to assess analytical agreement and determine the clinical suitability of the immunoturbidimetric method for diagnosis versus long-term monitoring in a local setting, in alignment with international standards.

Methods

A comparative cross-sectional study was conducted in a routine laboratory in Madagascar. Samples were tested on the same day using the Bio-Rad D-10 (Bio-Rad Laboratories, Hercules, CA) and Mindray BS-300 (Mindray, Shenzhen, China) automated analyzers. Analytical accuracy was assessed using control materials. The two methods were compared by evaluating the correlation using Pearson's coefficient (r) and linear regression, and the concordance using the Bland-Altman plot and Passing-Bablok regression. The HbA1c results were divided into a total group and subgroups (HbA1c < 6.5%, 6.5-8%, and > 8%).

Results

Intra-assay, inter-assay, and overall precision are satisfactory (coefficients of variation (CV) <3%) for routine use of both techniques. A good overall correlation between the immunoturbidimetric technique and HPLC was found with r = 0.88. The Bland-Altman plot shows a proportional increase in mean bias with HbA1c values and the regression equation \begin{document}y = 0.8144\,x + 0.3065\end{document}.

Conclusion

This study showed good correlation between the two techniques, but a systematic underestimation of HbA1c levels by the immunoturbidimetry assay was observed, especially for high values.