Poster Presentation Australasian Diabetes in Pregnancy Society Annual Scientific Meeting 2024

Accuracy of blood glucose meters in pregnant women with hyperglycemia (#17)

Jincy Immanuel 1 2 , Tobias Kongbrailatpam 1 , Jonathon Nguyen 1 , Rohit Rajagopal 3 , David Simmons 1
  1. School of Medicine, Western Sydney University, Sydney
  2. Texas Woman's university, Denton, Texas, United States
  3. Macarthur Diabetes Endocrinology & Metabolism Service, Campbelltown Hospital, Sydney

Background: Self-monitoring of blood glucose is an important component of hyperglycaemia in pregnancy management. Previous research has shown varying performance of home blood glucose meters and has highlighted the need to evaluate their accuracy during pregnancy. This study examined the analytical accuracy of four commonly used blood glucose meters in pregnant women and assessed whether changes in hematocrit level during pregnancy affected meter performance.

Methods: Pregnant women with any type of diabetes receiving antenatal care at Campbelltown Hospital (November 2022-December 2023) were enrolled. Capillary and venous blood samples were collected simultaneously. Glucose concentrations from Fingerstick and venous samples were measured using the Ascensia Contour Next, Roche Accu-chek Guide, Abbott Freestyle Neo, and LifeSmart meters and then compared against two reference methods: iSTAT analyzer and the laboratory plasma glucose measured by hexokinase method. Mean bias, the Bland-Altman tool, ISO 15197:2013 criteria, and surveillance error grid (SEG) analysis were conducted for each meter. The influence of hematocrit changes was assessed using an independent t-test by dichotomizing the hematocrit levels at 0.36 L/L.

Results: In total, 824 paired samples from 103 women were analyzed: (mean (SD) age 33.0±4.9 years, European ethnicity 36%, mean (SD) glucose: iSTAT 6.7±2.3 (range 2.5–14.3) mmol/l, hexokinase 6.6±2.4 (range 2.5–14.5) mmol/l, mean hematocrit 0.36 ±0.03 L/L. When using iSTAT as the reference, the percentages of meter measurements above the reference values for meters Ascensia to LifeSmart were 58%, 20%, 4.9%, and 90%, respectively (p<0.001). When using hexokinase as the reference, the percentages were 43%, 45%, 30%, and 88% respectively (p<0.001). Compared to iSTAT, the mean biases and 95% limits of agreement for each meter were: Ascensia 1.3% (-8%­–10.6%), Roche -3.2% (-11.4%–5.0%), Abbott -11.9% (-24.7%–0.8%), and LifeSmart 6.8% (-5.8%–19.4%). When compared to the hexokinase method, the biases and limits of agreement were as follows: Ascensia -0.2% (-8.1%–7.7%), Roche -0.2% (-10.2%–9.8%), Abbott -3.8% (-17.6%–10%), and LifeSmart 6.1% (-5.9%–18.2%). In the SEG analysis, Ascensia and Roche demonstrated compliance with the standard by showing ≥ 97% of pairs within the SEG no-risk green zone when using iSTAT as the reference method. When compared with the hexokinase method, Ascensia, Roche, and Abbott meters demonstrated compliance by achieving ≥ 97% within the no-risk zone. There was no difference in mean bias between the two hematocrit groups.

Conclusion: The performance of meters varied, with Ascensia and Roche demonstrating higher compliance with guidelines.