Recommended pre-analytical plasma glucose sampling methodology may distort gestational diabetes mellitus prevalence: implications for diagnostic thresholds.

ABSTRACT

AIM: Current International Association of the Diabetes and Pregnancy Study Groups/World Health Organization gestational diabetes mellitus (GDM) diagnostic thresholds are based on a landmark study in which the pre-analytical plasma glucose sampling methodology is unclear. Worldwide, plasma glucose pre-analytical sampling methodology practices are divergent. We considered the effects of pre-analytical plasma glucose sampling methodology on GDM prevalence and gestational outcomes. METHODS: This is a retrospective observational cohort study of 1178 pregnant women undergoing an oral glucose tolerance test (OGTT). Of the 1178 pregnant women, a subset of 892 non-GDM women with singleton births undergoing OGTT between 24 and 28 weeks' gestation were investigated for large for gestation age (LGA) outcomes. OGTT were determined using traditional methods (sodium fluoride tubes batched at roomed temperature). We modelled the potential effects of using a recommended pre-analytical plasma glucose methodology (lyophilized citrate tubes) on GDM prevalence. RESULTS: The GDM prevalence in our cohort was 13.5%. The incidence of LGA showed a linear association with maternal plasma glucose that was similar to the association observed in the Hyperglycemia and Adverse Pregnancy Outcome study. Frequency of LGA exceeded 10% at HAPO glucose category 4 (fasting, 4.8 to 4.9 mmol/l; 1-h, 8.7 to 9.5 mmol/l) for fasting and 1-h plasma glucose. The use of a recommended pre-analytical method is projected to increase the prevalence of GDM to 39.2%. CONCLUSION: We challenge the consensus that recommended pre-analytical plasma glucose methodologies are optimal for the accurate diagnosis of GDM. Recommended pre-analytical plasma glucose methods may profoundly over-diagnose GDM. Centres using recommended pre-analytical plasma glucose methodologies may need to reappraise their diagnostic thresholds.