Evaluation of Bio-Rad D100 for HbA1c testing of dried blood spot samples

Abstract

Abstract Hemoglobin A1c (HbA1c) testing is critical for diabetes diagnosis and monitoring. Although many studies have demonstrated the feasibility of measuring HbA1c from dried blood spots (DBS) using immunoassays, enzymatic assays, and certain high-performance liquid chromatography (HPLC) assays, there is limited evidence supporting the use of the Bio-Rad D100 HPLC method. We evaluated the feasibility of testing HbA1c from DBS using Bio-Rad D100 to support the assessment of glycemic control for the growing field of telehealth. Deidentified residual EDTA whole blood samples were processed, and HbA1c was measured on the Bio-Rad D100 on day zero. Concurrently, DBS were prepared for each sample by spotting 50 µL of blood onto a Whatman® 903 filter paper. The samples were dried overnight and stored at room temperature for two days. On day 2, a DBS disc with a diameter of 3.2 mm was eluted with 4.5 mL of Bio-Rad diluent for 20 minutes. The supernatants of the DBS samples were then processed on the platform. HbA1c values from the DBS samples were compared to the EDTA whole blood results. Compared to the EDTA samples, the HbA1c concentrations of the corresponding DBS samples were elevated, especially for samples with original results less than 7.0%. Furthermore, the chromatograms of the DBS samples were significantly altered; we observed baseline elevation and the appearance of unknown minor peaks. This issue was investigated. First, blank DBS samples were created by directly spotting filter paper with diluent or adding diluent into an empty EDTA tube and then spotting the filter paper with the mixture of diluent and EDTA. The blank DBS samples were tested, and the lack of peaks suggested that the interference was not from the diluent, EDTA, or the filter paper. DBS samples were also prepared with sodium heparin whole blood, finger-stick capillary blood, and multiple brands of filter papers and tested using a different HPLC method. However, the interference was still present in those samples, indicating the interference is most likely due to the time-dependent degradation of the samples themselves. Of note, the alternate HPLC method showed less chromatogram alteration. To prevent possible adduct formation, DBS samples were eluted in the presence of various concentrations of cysteine (Jeppsson et al. 1986). However, the pretreatment did not eliminate the chromatogram changes. Finally, a multiple regression model was created using the results of day two A1c, A0, and F peaks, and a correction factor was generated. The corrected HbA1c concentrations were well within +/-6% of the initial HbA1c concentrations. In conclusion, the Bio-Rad D100 HPLC method appears particularly susceptible to stability changes in DBS samples, and further analysis is needed before employing a correction factor.