Gestational diabetes mellitus affects 2–9% of pregnancies1,2 and is associated with increased perinatal complications, but many expert committees, including the UK National Institute for Clinical Excellence (NICE),3 advise against routine screening on the grounds that it is of unproven effectiveness.4,5 This must now change with publication of the Australasian Carbohydrate Intolerance Study (ACHOIS).6
In this well-conducted trial pregnant women were screened between 24 and 34 weeks of gestation by a 50-g oral glucose challenge test (OGCT) with 1 hour glucose above 7.8 mmol/L followed by a 75-g formal glucose tolerance test (GTT). One thousand women with a fasting glucose less than 7.8 mmol/L and a 2-hour glucose between 7.8 and 11 mmol/L on the GTT were randomised to intervention or control groups. When the trial was designed these levels fulfilled the WHO criteria for impaired glucose tolerance.7 The intervention group had dietary advice, regular assessment of blood glucose and insulin as required to keep pre-prandial and 2-hour post-prandial glucose levels below 5.5 and 7 mmol/L, respectively. Women in the control group were told they did not have gestational diabetes, as indeed they did not under the old nomenclature. Their caregivers were told to treat them as if screening had not been done (i.e. further testing for gestational diabetes was only done for clinical indications). All women were followed up and analysis was by intention to treat.
The primary outcome measure, a composite of serious perinatal complications (death, shoulder dystocia, bone fracture and nerve palsy), was dramatically lower in the intervention arm (1% vs 4%; relative risk 0.33; 95% confidence interval 0.14–0.75). All eight ‘avoidable’ adverse outcomes (four deaths, three nerve palsies and one fracture) were in the control arm. The intervention arm had more induced labours (39% vs 29%) and neonatal admissions (71% vs 61%), a similar rate of caesarean section (31% vs 32%) and a marked reduction in birthweight over 4 kg (10% vs 21%). Overall, 34 women needed treating to prevent one serious perinatal complication. These are dramatic benefits among women with relatively mild disease; after all women with the old definition of gestational diabetes, 2-hour glucose over 11 mmol/L, had been excluded.
For the first time there is demonstrable clinical validity to screening for gestational diabetes on the basis of improved obstetric outcomes. The challenge is how to provide such screening and treatment. There is debate on how to screen for gestational diabetes and whether entire populations should be screened.8 There remain some differences in the values used for a diagnosis of gestational diabetes after an oral glucose tolerance test and these can influence population incidences.9 Strategies other than that described in the ACHOIS study have been proposed to select and treat women with glucose intolerance.10,11 There are ongoing studies that may be able to better define perinatal risk.12
ACHOIS supports formal screening between 28 and 34 weeks of gestation and demonstrates that conventional management strategies improve meaningful outcomes. Obstetricians must now respond to the challenge of providing such a service that might result, in a unit where 4000 women give birth, in 200 pathological pregnancies requiring complex multidisciplinary care.