To evaluate the performance of the ACCU-CHEK Inform II glucose meter system for use in critically ill patients, and to assess the effects of blood oxygen tension, hematocrit, sodium levels on glucose measurements. A multicenter study involving the collection and testing of remnant deidentified whole blood samples from the blood gas laboratory at 3 institutions. Glucose meter measurements were compared to the Hitachi/Cobas c501 Isotope Dilution Gas Chromatograph aligned glucose hexokinase reference method, and the hospital’s laboratory comparison method. Blood gas, whole blood electrolyte, glucose, and hematocrit test results were documented for each sample. Acceptance criteria described in ISO15197:2013 and Clinical Laboratory Standards Institute (CLSI) POCT12-A3 were used to assess clinical performance. Glucose meter results correlated with both the Hitachi/Cobas c501 hexokinase, y = 0.988x − 0.4, R2 = 0.982, N = 579, hospital reference methods, y = 0.974x + 5.5, R2 = 0.972, N = 590. When compared to the glucose hexokinase method, glucose meter results met the International Organization for Standardization and the CLSI acceptance criteria. Glucose measurements were negatively biased at elevated sodium levels, with an approximate bias of −7 (mg/dL, %) at sodium level of 160 mmol/L. The glucose meter system correlated with both glucose hexokinase and hospital laboratory comparison methods, with performance meeting International Organization for Standardization CLSI acceptance criteria. However, meter results were influenced by high blood sodium concentration.
From the *Point-of-Care Testing Center for Teaching and Research, Department of Pathology and Laboratory Medicine, School of Medicine, University of California, Davis, CA; †Department of Pathology and Clinical Laboratories, Duke Medical Center, Durham, NC; ‡Department of Clinical Chemistry, Isala, Zwolle, the Netherlands; and §Roche Diagnostics, Indianapolis, IN.
Reprints: Richard F. Louie, PhD, FACB, Pathology and Laboratory Medicine, POCT•CTR, School of Medicine, 3455 Tupper Hall, University of California, Davis, CA 95616. E-mail: firstname.lastname@example.org.
The research was funded by Roche Diagnostics.
The authors declare no conflicts of interest.