Gestational diabetes needs to be managed

In reply: Taylor welcomes a strong evidence base for clinical practice, which is echoed by d’Emden and colleagues. However, their strong support for the current Australian diagnostic criteria for gestational diabetes mellitus (GDM) is surprising, given that these criteria are the product of an “Ad Hoc Working Party” report.1 This guideline, published in 1991, clearly acknowledged a lack of strong evidence and advocated criteria that were based on best available rounded values for the 95th centile of venous plasma glucose (VPG) levels in the fasting state (5.5 mmol/L) and 2 hours after a 75 g glucose load (8.0 mmol/L) — values that had been published in an unreferenced overview of Australian and European data.1 It advocated future prospective studies, such as those which form the basis of the International Association of the Diabetes and Pregnancy Study Groups (IADPSG) recommendations.2 Contrary to d’Emden et al’s assertions, the 1991 fasting and 2-hour VPG level cut-offs are misaligned. Contemporary data from the 2120 Australian women in the Hyperglycemia and Adverse Pregnancy Outcome study, using the 75 g oral glucose tolerance test, show that 5.5 mmol/L lies at the 98th centile for fasting VPG level, while 8.0 mmol/L corresponds to the 91st centile for 2-hour VPG level (own unpublished data). A recent summary of ambulatory blood glucose monitoring in pregnancy suggested even tighter normative values, with the 97th centiles for fasting blood glucose level (BGL) (4.8 mmol/L) and 2-hour post-meal BGL (6.6 mmol/L) being lower than previous guesstimates.3

Further, the IADPSG diagnostic thresholds are based not on cohort percentiles but on associations with diabetic fetopathy and correspond to equivalent levels of glucose-related risk, after extensive corrections for potential confounders. Considering the randomised controlled trials, we note that although the Australian Carbohydrate Intolerance Study in Pregnant Women permitted recruitment of women with markedly elevated fasting BGLs, there were few such women.4 By contrast, the US study on treatment for mild GDM considered a fasting VPG level of 5.3 mmol/L or higher as indicative of GDM that definitely required treatment and women in this category were excluded from the trial.5 The mean fasting VPG level in both trials was 4.8 mmol/L, thus both included substantial numbers of women with fasting VPG levels in the range specified by IADPSG. Both trials specified fasting BGL targets for women receiving treatment (5.5 mmol/L4 and 5.3 mmol/L5) and treated actively to achieve both these and post-meal glucose targets. Active treatment in both trials reduced important complications related to fetal overgrowth and pre-eclampsia.

Given the congruence of epidemiological and randomised controlled trial data, we believe that it is time for implementation, rather than further procrastination.