Reducing analytical variation between point-of-care and laboratory HbA1c testing

Background: Point-of-care (POC) HbA1c testing allows for timely treatment changes, improved glycemic control, and patient and provider satisfaction. Substantial variation between POC and laboratory HbA1c results has been reported. At our university hospital diabetes clinic, we observed significant negative bias in HbA1c with the DCA Vantage (TM) (Siemens Healthcare Diagnostics, Tarrytown, NY, USA) compared with the Tosoh G8 HPLC laboratory analyzer (Tosoh Bioscience, San Francisco, CA, USA). This led us to systematically analyze the bias with the goal of recalibrating the DCA to minimize bias. Methods: We analyzed 45 patient samples, with HbA1c ranging between 5% and 10.8%, concurrently on two DCA analyzers and on the Tosoh G8 machine. The bias for each sample was the difference between the value on the DCA and the Tosoh 08 analyzer. Based on regression equations derived from the data, a correction factor for each DCA analyzer was calculated. The analyzers were recalibrated and retested for bias. Results: At baseline, the mean bias (range) was -0.5229 (+0.1 to -1.3) for Analyzer 1 and -0.5348 (0.0 to -1.6) for Analyzer 2. After recalibration, the mean bias (range) was 0.000 (+0.6 to -0.6) and 0.0003 (+0.5 to -0.5) for Analyzers 1 and 2, respectively, and the systematic negative bias seen prior to the calibration was almost eliminated. Conclusions: We recommend periodic recalibration of POC analyzers to eliminate systematic unidirectional bias and to harmonize results between the POC and central laboratory analyzers within a healthcare system. Calibration may need to be repeated with any change in the reagent lot.