Evaluation of the analytical and clinical accuracy of four blood glucose meters in pregnant women with hyperglycaemia.

Abstract

Aims: Physiological changes during pregnancy can infuence the performance of blood glucose meters. This study aimed to evaluate the analytical and clinical accuracy of glucose meters in pregnant women with hyperglycaemia.

Materials and methods: Glucose was measured by four commonly used meters among consecutive women with diabetes in pregnancy. Capillary and venous samples were collected concurrently and compared with i-STAT (amperometry) and laboratory (hexokinase) glucose as reference methods. Bland-Altman plot, International Organization for Standardization criteria, surveillance error grid (SEG) and haematocrit influence were assessed.

Results: In total, 824 paired samples from 103 women were analysed (GDM 57%, mean i-STAT capillary glucose 6.7 ± 2.3 mmol/L [121 ± 41 mg/dL], laboratory glucose 6.6 ± 2.4 mmol/L [119 ± 43 mg/dL], median haematocrit 0.36 L/L). Mean capillary glucose measured on all meters was significantly different from that measured on i-STAT (all p < 0.001), whereas venous glucose measured on Contour Next, Accu-Chek Guide and the laboratory (plasma) was similar. Contour Next had the lowest bias when using both reference methods (mean bias [95% limits of agreement] meter vs. i-STAT: Contour Next 1.3% [-8% to 10.6%], Accu-Chek Guide -3.2% [-11.4% to 5%], FreeStyle Optium Neo -11.9% [-24.7% to 0.8%] and LifeSmart 6.8% [-5.8% to 19.4%]; meter versus laboratory: -0.2% [-8.1% to 7.7%], -0.2% [-10.2% to 9.8%], -3.8% [-17.6% to 10%] and 6.1% [-5.9% to 18.2%]), respectively. Only Contour Next and Accu-Chek Guide had ≥97% of pairs within the SEG no-risk zone during both comparisons. Meters did not show haematocrit-related bias.

Conclusions: Accuracy of meters was higher when using venous samples than when using capillary samples. Contour Next and Accu-Chek Guide meters met accuracy standards in all analyses.