The accuracy of a continuous blood glucose monitor during surgery.

BACKGROUND: Protocols for tight control of blood glucose can be difficult to achieve in the surgical setting, especially when relying upon intermittent blood glucose testing. A continuous blood glucose monitoring system can facilitate blood glucose management. In the present study, we compared blood glucose measured continuously (STG-22™, Nikkiso, Tokyo, Japan), during surgery with coincident measurements obtained intermittently using a conventional laboratory glucometer (ABL™ 800FLEX (Radiometer Medical Aps, BrJOURNAL/asag/04.02/00000539-200801000-00030/ENTITY_OV0119/v/2017-07-21T020210Z/r/image-pngnshJOURNAL/asag/04.02/00000539-200801000-00030/ENTITY_OV0119/v/2017-07-21T020210Z/r/image-pngj, Denmark). The goal of the study was to determine the reliability and accuracy of the continuous method during surgery. MATERIAL AND METHODS: Twenty-nine patients scheduled for routine surgery with general anesthesia were enrolled in this study. After anesthetic induction, a 20G IV catheter was inserted in a peripheral forearm vein and connected to the continuous blood glucose monitor. A radial arterial catheter was also inserted from which samples for blood glucose estimation were obtained by an anesthesiologist, following an established protocol of discarding 3 mL of blood before the actual blood sampling. Blood glucose was measured by ABL™ 800FLEX immediately after sampling. One hundred points of paired blood glucose values were obtained, which were compared using Bland and Altman analysis. RESULTS: Bias and upper and lower limits of agreement were −2.6, 23, and −28, respectively. The percentage error of the lower/upper limits of agreement was 21% and 18%, respectively. DISCUSSION AND CONCLUSIONS: The blood glucose measurements obtained continuously agreed with the coincident intermittent measurements within 21%. The STG-22™ may still be useful for following changes continuously and reducing the frequency of intermittent measurement, but the need for testing samples with a reliable device is not eliminated.