Abstract
BACKGROUND: Hyperglycemia is an independent risk factor for perioperative complications and mortality after cardiac surgery. Freestyle Libre Continuous glucose monitor (CGM) is an interstitial glucose monitoring system, which has been proven to be stable and accurate in various disease. MATERIALS AND METHODS: 86 patients with hyperglycemia after cardiac surgery between October 2021 and October 2022 in the cardiovascular ICU, were enrolled in this study. We evaluated systematic errors and numerical accuracy in blood glucose (BG) between two groups using Bland-Altman analysis and mean absolute relative difference (MARD). The clinical accuracy was assessed by consensus and Clarke error grid. RESULTS: 435 paired sensor-reference values were enrolled between the Freestyle and venous BG. The MARD was 18.2% with the Bland-Altman standard bias of 31.87 mg/dL. 1851 paired values were enrolled between the Freestyle and arterial BG. The MARD was 21.5%, and the Bland-Altman standard bias was 41.45 mg/dL. 432 paired values were enrolled between the arterial and venous BG. The MARD was 8.4%, and the Bland-Altman standard bias was - 9.95 mg/dL. CONCLUSIONS: Our study demonstrated the relatively low accuracy of Freestyle Libre CGM for critical patients after cardiac surgery. The arterial blood gas glucose analysis acts as an effective tool to improve BG management.