RESUMO
BACKGROUND: Hemoglobin A1c is currently used both for monitoring and diagnosing patients with diabetes mellitus. Point-of-care (POC) instruments have been proposed to improve glycemic control of patients with diabetes by providing a rapid result if the analytical performance is acceptable. However, the performance of most current POC devices is unsatisfactory. In this study, we evaluated 3 POC devices: cobas b101 (Roche), Afinion (Alere) and B-Analyst (Menarini). METHODS: The CLSI protocols EP-5 and EP-9 were applied to investigate precision, accuracy, and bias. Bias was compared with an ion-exchange chromatography method (Adams Arkray HA-8160, Menarini) of which the accuracy was assessed by three secondary reference methods (SRMs) and a capillary zone electrophoresis-based laboratory method (Capillarys Flex Piercing 2, Sebia). RESULTS: Using the IFCC units (mmol/mol), total CV's for cobas b101, Afinion and B-Analyst ranged from 2.0 to 3.9%, from 2.3 to 3.7% and from 2.3 to 2.9%, respectively, compared to a total CV ranging from 0.9 to 1.2% for the Adams Arkray HA-8160. A high correlation was found between the HA-8160 and the three POC devices (r = 0.992, r = 0.968, r = 0.998 for cobas b101, Afinion and B-Analyst, respectively). However, all three showed a significant bias of -2.2 mmol/mol, -2.2 mmol/mol and 0.5 mmol/mol, respectively. CONCLUSIONS: Only the B-Analyst met the current HbA1c quality specifications for precision, whereas cobas b101 and Afinion failed on the low level. Significant bias was observed for all three POC instruments. As low precision and bias are required for both monitoring and diagnosis of diabetes, we believe that POC devices should be carefully selected.