Accuracy of intermittently scanned continuous glucose monitoring during caesarean delivery in pregnant women with insulin-treated diabetes

Abstract

AimContinuous Glucose Monitoring (CGM) systems are not currently recommended to guide intrapartum glucose and insulin infusion, due to insufficient data. In this study, intrapartum accuracy of intermittently scanned CGM (isCGM), compared to simultaneously measured capillary glucose (CG), was evaluated. MethodsPaired isCGM (Freestyle Libre 2) - CG data during caesarean delivery in pregnant women with insulin-treated diabetes were prospectively collected. The isCGM accuracy was assessed by MARD and Clarke Error Grid analysis. Moreover, the impact on intrapartum management was evaluated. ResultsSixty-eight paired isCGM-CG data of 19 women were evaluated. The overallMARD was 9.28 %. All values were in A and B zones of Clarke Error Grid. Forty-six (68 %) isCGM-CG pairs were in the same glycemic range, meaning the same intrapartum management. All discordant data were identified by checking CG in case of isCGM above 110 mg/dL or less than 70 mg/dL [chi-square 21.76, p < 0.001]. At ROC curve, isCGM above 110 mg/dL was associated with 100 % sensitivity to discordant result at CG (AUC 0.859, p < 0.001). ConclusionThe accuracy of isCGM during caesarean delivery was good, particularly for glucose values between 70 and 110 mg/dL, when CG confirmation could be safely avoided.