Endocrine Abstracts

Various criteria have been proposed to diagnose and treat gestational diabetes due to uncertainties about the influence of mild degrees of maternal glucose intolerance on pregnancy outcomes. We used receiver operating characteristics (ROC) curve analysis to correlate macrosomia (more than or equal to 4000g) with varying thresholds for the 2 hour 75 gm oral glucose tolerance test (OGTT) results in pregnant women screened at our hospital from 1998-2001.

Women with a 2 hour OGTT value >6.8mmol per litre were compared with those >7.8mmol per litre (WHO criterion). All received similar education, monitoring and clinic follow up with initiation of insulin for capillary glucose results persistently >7.0mmol per litre.

47 of 168 women (28%) with 2-hour value >6.8 had macrosomic babies as opposed to 15 of 61 (25%) with >7.8. A 6.8 mmol per litre cut-off involved follow-up of an additional 107 women but identified 32 additional macrosomic infants. 2 hour values of 4.8, 5.8, 6.8 and 7.8 gave sensitivities of 90, 82, 44 and 14%, and specificities of 10, 22, 51 and 82% respectively on ROC curve analysis. The proportion of macrosomics identified were 31%, 28% and 25% using thresholds of 5.8, 6.8 and 7.8. Defining macrosomia as 4500 grammes resulted in no significant change in sensitivity or specificity. The macrosomia rate in patients with a 2 hour value >7.8mmol per litre (25%) is comparable to 20-30% found by others. Of these, 22% treated with insulin had macrosomic babies compared with 25 % in the untreated group.

Present guidelines are based on postpartum maternal glycaemic outcomes rather than fetal outcomes. Adopting a progressively lower 2-hour glycaemic threshold identified macrosomics at a similar rate although with lower specificity, justifying a lower threshold to prompt monitoring during pregnancy. We suggest that this would improve fetal morbidity and help achieve the goals of the St Vincent's declaration.