Findings Between March 31, 2009, and June 2, 2014, we assessed 8820 women for eligibility and recruited 1555, with
a mean BMI of 36·3 kg/m² (SD 4·8). 772 were randomly assigned to standard antenatal care and 783 were allocated
the behavioural intervention, of which 651 and 629 women, respectively, completed an oral glucose tolerance test.
Gestational diabetes was reported in 172 (26%) women in the standard care group compared with 160 (25%) in the
intervention group (risk ratio 0·96, 95% CI 0·79–1·16; p=0·68). 61 (8%) of 751 babies in the standard care group were
large for gestational age compared with 71 (9%) of 761 in the intervention group (1·15, 0·83–1·59; p=0·40). Thus, the
primary outcomes did not diff er between groups, despite improvements in some maternal secondary outcomes in the
intervention group, including reduced dietary glycaemic load, gestational weight gain, and maternal sum-of-skinfold
thicknesses, and increased physical activity. Adverse events included neonatal death (two in the standard care group
and three in the intervention group) and fetal death in utero (ten in the standard care group and six in the intervention
group). No maternal deaths were reported. Incidence of miscarriage (2% in the standard care group vs 2% in the
intervention group), major obstetric haemorrhage (1% vs 3%), and small-for-gestational-age infants (≤5th customised
birthweight centile; 6% vs 5%) did not diff er between groups.