Automated blood glucose monitor is insensitive to hematocrit variation and common saccharide interferents

Introduction

Luminous Medical is developing an automated, patient-attached system (Argus Glucose Monitoring System) that uses a flow-through glucose oxidase sensor to measure plasma glucose levels in whole blood. The Argus GMS will aid caregivers in achieving glycemic control in critical care patients. The FDA has recently indicated that ICU glucose monitors need to satisfy tighter accuracy requirements than those specified by ISO 15197 (± 15 mg/dl or ± 20%). Luminous conducted a study to demonstrate Argus GMS accuracy under conditions of variable hematocrit, and in the presence of common saccharide interferents.

Methods

Hematocrit Study Group: one unit of whole blood was collected under an IRB-approved protocol from two healthy volunteers. Blood samples were prepared at three hematocrit levels (~24, 36, and 48%). Three glucose levels were tested at each hematocrit: native glucose, native plus 100, and native plus 200 mg/dl. Repeats were measured for each hematocrit/glucose combination (n = 99 total samples in this study group). Interferent Study Group: blood samples were prepared at native hematocrit (~37%) and native glucose level (~100 mg/dl). Testing followed established interference testing guidelines [1]. Blood samples were spiked with concentrated interferent solution or with equal volume of normal saline. Testing alternated between interferent-containing and interferent-free samples (n = 5 test and five control samples for each interferent). Saccharide interferents included maltose (55 mg/dl), galactose (15 mg/dl), fructose (18 mg/dl), and xylose (40 mg/dl). In both study groups, blood samples were measured in vitro, in real time, using the Argus GMS in multisample mode. Plasma measurements from two YSI 2700 instruments were averaged to provide reference glucose values.

Results

Hematocrit Study Group: the Argus GMS measured plasma glucose across a range of 50 to 350 mg/dl with RMS error of 5.5 mg/dl (mean error 2.6, 95% CI = -7.0 to 12.2, R = 0.998). The Argus GMS showed no sensitivity to hematocrit variation. Interferent Study Group: the Argus GMS showed minimal glucose measurement bias for each of the four saccharide interferents (maltose 2.21 mg/dl, galactose 4.09 mg/dl, fructose 2.63 mg/dl, and xylose 2.03 mg/dl).

Conclusions

The Argus GMS provides highly accurate glucose readings in whole blood samples with no observable sensitivity to hematocrit variation and with minimal influence from saccharide interferences. Argus demonstrated measurement accuracy within ± 10 mg/dl or ± 10% at 95% confidence limits.