Robert C Hawkins1. 1. Department of Pathology and Laboratory Medicine, Tan Tock Seng Hospital, Singapore 308433, Singapore. Robert_Hawkins@ttsh.com.sg
Abstract
BACKGROUND: We examined the analytical performance of two Roche Accu-Chek Go meters (reporting whole blood glucose) and two Medisense Optium meters (reporting plasma glucose) against laboratory plasma glucose measurement. METHODS: One hundred and twenty heparinized whole blood specimens were analyzed in duplicate on each of the four meters and compared to plasma glucose measurement on the Roche DP Modular system (glucose oxidase methodology). RESULTS: There was a significant difference in imprecision between the two Medisense meters (CV 3.7% and 4.9%), and between the Roche (CV 2.5% and 2.9%) and Medisense meters. The two Roche meters showed a consistent positive bias compared to the Modular, while the two Medisense meters over-read at low glucose concentrations and under-read at high glucose concentrations. Total error for the Roche meters averaged 10.8% and 8.9%, and for the Medisense meters averaged 12.5% and 15.6%, respectively. All results for all meters fell in error grid zones A and B. CONCLUSION: Neither meter met the ADA, CLIA or biological variation goals, although they came close to the NCCLS guideline. Neither meter reported plasma glucose equivalents. The Accu-Chek Go performed slightly better than the Optium meter, but the positive bias was puzzling and requires further explanation.
BACKGROUND: We examined the analytical performance of two Roche Accu-Chek Go meters (reporting whole blood glucose) and two Medisense Optium meters (reporting plasma glucose) against laboratory plasma glucose measurement. METHODS: One hundred and twenty heparinized whole blood specimens were analyzed in duplicate on each of the four meters and compared to plasma glucose measurement on the Roche DP Modular system (glucose oxidase methodology). RESULTS: There was a significant difference in imprecision between the two Medisense meters (CV 3.7% and 4.9%), and between the Roche (CV 2.5% and 2.9%) and Medisense meters. The two Roche meters showed a consistent positive bias compared to the Modular, while the two Medisense meters over-read at low glucose concentrations and under-read at high glucose concentrations. Total error for the Roche meters averaged 10.8% and 8.9%, and for the Medisense meters averaged 12.5% and 15.6%, respectively. All results for all meters fell in error grid zones A and B. CONCLUSION: Neither meter met the ADA, CLIA or biological variation goals, although they came close to the NCCLS guideline. Neither meter reported plasma glucose equivalents. The Accu-Chek Go performed slightly better than the Optium meter, but the positive bias was puzzling and requires further explanation.