Background: The optimal treatment of diabetes in pregnancy requires accurate measurement of blood glucose levels, in order to minimize adverse outcomes for both mother and neonate. Self-monitoring of blood glucose is routinely used to measure glycemic control and to assess whether treatment targets are being met; however, the accuracy of blood glucose meters in pregnancy is unclear.
Materials and methods: Pregnant women with gestational, type 1, or type 2 diabetes mellitus were eligible to participate. Nonfasting capillary blood glucose levels were measured in duplicate using the BGStar(®) (Sanofi, Sydney, Australia) and FreeStyle Lite(®) (Abbott, Sydney) blood glucose meters. Venous blood samples were collected and analyzed for plasma glucose, hematocrit, and glycated hemoglobin. Capillary blood glucose was compared with plasma glucose and further assessed according to International Organization for Standardization (ISO) 15197:2013 standards.
Results: One hundred ten women were recruited, providing 96 samples suitable for analysis. The mean ± SD laboratory plasma glucose level was 4.6 ± 1.4 mmol/L; the BGStar and FreeStyle Lite capillary blood glucose values were 5.3 ± 1.4 mmol/L and 5.0 ± 1.3 mmol/L, respectively. Both meters showed a positive bias (0.42 mmol/L for the FreeStyle Lite and 0.65 mmol/L for the BGStar). Furthermore, neither meter fulfilled the ISO 15197:2013 standards, and there was a nonsignificant improvement in meter performance at blood glucose levels of ≤4.2 mmol/L. Hematocrit did not affect the results of either blood glucose meter. Clarke Error Grid analysis demonstrated that approximately 70% of the results of both meters would lead to appropriate clinical action.
Conclusions: The BGStar and FreeStyle Lite blood glucose meters did not meet ISO 15197:2013 recommendations for blood glucose monitoring systems when assessed in a population of women with diabetes in pregnancy. Clinicians should consider this difference in blood glucose readings when making diabetes-related treatment decisions.