Retrospective Analysis of Point-of-Care and Laboratory-Based Hemoglobin A1c Testing

Abstract

Abstract Background Glycemic control is essential to diabetic management, and hemoglobin A1c (Hb A1c) has long been used for this purpose. Though laboratory-based testing is standard, point-of-care (POC) systems provide rapid results in clinic, allowing more timely patient management. A negative bias with POC testing has been observed, and our aim is to further characterize these discrepancies at our institution. Methods A medical record search identified patients who underwent laboratory-based and/or POC Hb A1c testing (DCA Vantage™) at our medical center from July 2015 to April 2016. Patients who underwent both tests within 30 days were grouped by age, sex, and test interval (same day, <1 day, ≤15 days, or ≤30 days). Mean laboratory-based and POC values were compared using the paired t-test. Correlation statistics were determined using the Deming regression. Results In total, 40503 data points were gathered from the database, comprising 28555 laboratory-based Hb A1c tests and 11948 POC-based Hb A1c tests. A total of 28292 unique patients were identified, of which 493 underwent both tests within 30 days. While DCA and laboratory-based testing was highly correlated, there was a mean negative bias of 0.18% with POC testing. Bias was greater for women [0.17% higher (95% CI, 0.063%–0.284%), P = 0.002] and children aged 0–13 years [0.52% higher (95% CI, 0.141%–0.891%), P = 0.007]. Conclusions There is a consistent negative bias with POC testing, most pronounced in the female and pediatric populations. Further studies will determine what variables contribute to this discrepancy and how clinical management is modified. POC testing using the DCA Vantage should be interpreted cautiously.