Abstract Background The increased prevalence of diabetes in socially disadvantaged populations requires solutions that are amenable to these communities. Unmanaged diabetes can cause cardiovascular disease, kidney damage, blindness, and limb amputation. Diabetes is traditionally diagnosed using fasting/oral glucose tolerance or hemoglobin A1c (HbA1c) tests; both of which require access to medical and laboratory facilities. Self-collected capillary dried blood spot (DBS) testing for HbA1c could alleviate health disparities by offering access to low cost and minimally invasive remote collection. The objective of this study was to validate self-collected capillary DBS specimens for clinical HbA1c testing. Methods DBS HbA1c was measured using the hemolysate application of the Roche cobas c513 Tina-quant Hemoglobin A1cDx Gen.3 assay. Calibration was performed according to the manufacturer’s package insert. Patient samples were prepared by collecting two 3mm DBS punches. Non-biological standards were prepared by placing two blank 3mm DBS punches into an empty tube and adding 10uL of standard. After 24 hours, samples were resuspended in 750uL of hemolyzing reagent and incubated for 75 minutes at room temperature (RT) while shaking at 800 RPM. Preliminary data showed that non-biological standards spotted onto DBS cards did not behave like real patient samples. A correction factor (CF) was derived (n=134) and challenged (n=20) to account for the systematic bias observed between matched patient DBS and venous specimens. Imprecision was established by running two patient DBS specimens as controls. Accuracy was evaluated using residual College of American Pathologists (CAP) proficiency testing (PT) material (n=32). Linearity was confirmed by running 5 levels of Maine Standards Linearity HbA1c Kit in triplicate. Stability of the DBS specimen and the hemolysate supernatant extracted from the DBS specimen were observed at RT. Results Specimens tested between venous and DBS HbA1c to derive the CF (n=134) had a bias of 0.26% (4.32%) and a linear equation of (y=0.938x +0.631; R=0.996). The resulting CF is: reported HbA1c% = (DBS HbA1c% - 0.631)/0.938. Matched venous and DBS specimens that challenged the CF (n=20) had a bias of -0.10% (-1.83%) and a linear equation of (y=1.06x -0.41; R=0.988). Imprecision was 2.3% and 1.3% for patient DBS control specimens with mean concentrations of 5.3% and 5.9%, respectively. Residual CAP PT samples (n=32) had an average bias of -0.05% (-0.40%). DBS HbA1c concentrations from 4.2-15.5% were linear (y=0.903x +1.15; R=0.999). The DBS specimen is stable for 60 days at RT and the hemolysate supernatant is stable for at least 10 hours at RT. Conclusions Self-collected capillary DBS specimens are acceptable for the clinical measurement of HbA1c, and effectively increase accessibility to diabetes screening and monitoring. Applying a CF to DBS patient samples removes the systematic bias observed between matched venous and DBS samples. Non-biological standards and capillary whole blood do not behave the same on DBS cards. Alternative PT for DBS specimens should be investigated as at-home collection advances.
Read full abstract