Hubert W Vesper1, Enada Archibold, Gary L Myers. 1. Division of Laboratory Sciences, National Center for Environmental Health, Centers for Disease Control and Prevention, 4770 Buford Hwy NE MS F25, Atlanta, GA 30341, United States. HVesper@cdc.gov
Abstract
BACKGROUND: The trueness of glucose monitors is commonly assessed using whole-blood samples and a clinical analyzer as a comparison method. In this study, the effect of specimen matrix on trueness of one clinical analyzer and one glucose monitor was compared with a gas chromatography-mass spectrometry (GC/MS) reference method, using split-sample comparison with capillary whole blood (CWB), venous whole blood (VWB), and plasma (PL). METHODS: CWB was analyzed by the glucose monitor and the GC/MS reference method. VWB was analyzed by the glucose monitor, clinical analyzer, and the GC/MS method. PL was analyzed by the clinical analyzer and the GC/MS reference method. RESULTS: For the glucose monitor, the bias was 0.4% and -18.2% for CWB and VWB, respectively. The clinical analyzer had a bias of -25.4% for VWB and -12.0% for PL and a proportional bias was detected in both specimens. Using the clinical analyzer as a comparison method, the glucose monitor had a proportional bias of -9.8%. CONCLUSION: The trueness of clinical analyzers can be affected by the specimen matrix that needs to be assessed before they are used as comparison method to assess trueness of glucose monitors.
BACKGROUND: The trueness of glucose monitors is commonly assessed using whole-blood samples and a clinical analyzer as a comparison method. In this study, the effect of specimen matrix on trueness of one clinical analyzer and one glucose monitor was compared with a gas chromatography-mass spectrometry (GC/MS) reference method, using split-sample comparison with capillary whole blood (CWB), venous whole blood (VWB), and plasma (PL). METHODS: CWB was analyzed by the glucose monitor and the GC/MS reference method. VWB was analyzed by the glucose monitor, clinical analyzer, and the GC/MS method. PL was analyzed by the clinical analyzer and the GC/MS reference method. RESULTS: For the glucose monitor, the bias was 0.4% and -18.2% for CWB and VWB, respectively. The clinical analyzer had a bias of -25.4% for VWB and -12.0% for PL and a proportional bias was detected in both specimens. Using the clinical analyzer as a comparison method, the glucose monitor had a proportional bias of -9.8%. CONCLUSION: The trueness of clinical analyzers can be affected by the specimen matrix that needs to be assessed before they are used as comparison method to assess trueness of glucose monitors.