Analytic Bias Among Certified Methods for the Measurement of Hemoglobin A_1c

Abstract

We studied the magnitude, significance, and origin of an analytic bias that emerged between our point-of-care (POC) and our central laboratory (CL) methods for the measurement of hemoglobin A1c (HbA1c) and evaluated the analytic accuracy of 7 commonly used HbA1c methods relative to the National Glycohemoglobin Standardization Program (NGSP) reference method. The POC and CL methods were compared by split-sample analysis of clinical specimens and time series analyses of the HbA1c results reported for a 33-month period. The relative accuracies of 7 HbA1c methods were evaluated using College of American Pathologists proficiency survey results. Long-term drifts in the CL- and POC-analyzed test results caused the median intermethod bias [(POC result)-(CL result)] to increase from -0.4% to -0.9% HbA1c. Systematic biases, drifts in analytic performance over time, and intermethod variability were frequently observed among the 7 NGSP-certified HbA1c methods. Intermethod variability is a potential source of inaccuracy whenever HbA1c results are interpreted relative to universal, fixed, clinical decision thresholds.

Six of Eight Hemoglobin A1c Point-of-Care Instruments Do Not Meet the General Accepted Analytical Performance Criteria

“…There is demonstrated benefit in using POC instruments for the measurement of Hb A 1c in certain clinical situations (2-4), but recently concerns have been raised about the performance of NGSP-certified POC instruments compared with laboratory-based methods (17 ). The overall imprecision as determined by means of an EP-5 protocol is very important for interpretation of Hb A 1c results (variability in the patient vs analytical variability).…”

Performance of point-of-care HbA1ctest devices: implications for use in clinical practice – a systematic review and meta-analysis

“…Some laboratory instruments have been reported to have systematic biases against reference laboratories [46] which may have led to under-or over-estimation of biases in some studies in our review. Only nine studies reported to use an NGSP or IFCC reference laboratory for the comparator method which generally resulted in lower variability.…”

Nurse-led diabetic retinopathy screening: a pilot study to evaluate a new approach to vision care for Canadian Aboriginal peoples

“…The average level of blood glucose in the 30 days immediately preceding the blood sample contributes ~50% of the HbA1c results whereas ~10% is from the past 90–120 days [18,19]. The accuracy of the A1C Now+ was tested on participants who were diabetic and non-diabetic (n=189) in a multiple-site study conducted in the USA, with a 99% accuracy rate found when compared to the National Glycohemoglobin Standardization Program reference results [20]. The present study identified 2 categories of HbA1c: normal (7% or lower) and elevated (7.1% or higher).…”