Oral Presentation Australasian Diabetes in Pregnancy Society and Society of Obstetric Medicine Australia and New Zealand Joint Scientific Meeting 2021

Accuracy of continuous glucose monitoring during periods of predicted acute glycaemic variability in pregnancy in women with Type 1 Diabetes Mellitus in the inpatient setting: A pilot study (#10)

Emma Croker 1 , Naeel Mohammad 1 , Christopher W Rowe 1 2 , Katie Wynne 1 2
  1. Department of Endocrinology, John Hunter Hospital, New Lambton, NSW, Australia
  2. School of Medicine and Public Health, University of Newcastle, Newcastle, NSW, Australia

Background: Subcutaneous continuous glucose monitors (CGM) use during pregnancy in ambulant women with Type 1 diabetes mellitus (T1DM) improves glycaemic control and neonatal outcomes. Women may require intensive inpatient glycaemic management during glucocorticoids, intercurrent illness or intrapartum to reduce maternal-foetal risk. CGM has the potential to enhance glycaemic monitoring during these episodes and improve maternal experience by minimising capillary glucose ‘finger-prick’ testing. The accuracy of CGM during episodes of acute glycaemic variability has not been evaluated in pregnancy.

Methods: Observational, retrospective study of pregnant women with T1DM using CGM whist on IV insulin infusion with 30-60minutely capillary glucose (CapG) monitoring. CapG was paired with nearest time-point CGM glucose.

Results: Data available for eleven episodes (intercurrent illness=3, intrapartum=8). Median gestation was 37 weeks; CGM Dexcom 5 (n=9) and Carelink (n=2). Pearson’s correlation for CGM and CapG was 0.80 (p<0.005) for n=256 paired data points (figure1). Overall accuracy of CGM compared to CapG was reasonable, with mean absolute relative difference (MARD) 11.4% (SD 11.6) for n=256 data points. MARD calculated for each episode ranged from 1.1-32.8% (median 11.9%,IQR 6.59) (figure2). Significant positive correlation between increasing CV% of CapG and MARD (Spearman rho 0.68,p=0.02) and positive correlation between increased rate of change of CapG and increased CapG to CGM glucose (linear regression coefficient 0.69, p<0.005). No apparent correlation between MARD and number of data points or number of calibrations. In 5/10 capillary-detected hypoglycaemic events (glucose<3.8mmol/L), CGM levels lagged behind, with up to 70% difference between CGM and CapG in this range (red box, figure 1); six hypoglycaemic events occurred during one admission.

Conclusion: Capillary glucose and CGM were reasonably correlated during IVI in pregnancy. Rate of change of glucose may predict discrepancy between capillary and CGM glucose. Further study of accuracy of CGM under glycaemic extremes in pregnancy is planned in prospective studies.

61313a4ef1591-Abstract+publication+picutre.png