Performance of an automated blood glucose monitor in the OR/ICU

Introduction

DexCom and Edwards Lifesciences collaborated to develop an intravenous blood glucose monitor for use in a critical care environment. The system provides automatic, real-time monitoring and trending of blood glucose (BG) values for critical care patients. The system is designed to address the needs of clinicians to monitor BG in critical care units. The sensor was also designed to accurately measure BG even in the presence of many drugs commonly used in the hospital.

Methods

This was a blinded, multicenter study evaluating 17 adult patients presenting to the OR and ICU. They were monitored for up to 72 hours with reference samples taken at 4-hour intervals and measured with a YSI-2300 laboratory analyzer. Data are from initial roll-in patients used for training and assessing protocol complications. Reference blood samples were taken from either a peripheral vein, a central venous catheter or an arterial line. Glucose values were calculated prospectively but not displayed during the study.

Results

Of 211 paired BG measurements, 94.3% were within Clarke Error Grid A, 5.7% in B and 0% in C, D and E (Figure 1a). The mean absolute relative difference was 6.5%. No significant errors due to drugs were measured. These results compare favorably with a previous study of 50 subjects who were monitored continuously for up to 72 hours in a diabetic in-clinic study (Figure 1b) where large excursions of BG could safely be evaluated.

Figure 1

Clarke Error Grid for (a) OR/ICU and (b) in-clinic.

Conclusions

The results of this study help illustrate the capability of the system to provide safe, accurate, automated glucose measurements in critical care patients.