The pregnancy outcome of 33 women with gestational diabetes who were treated with glibenclamide and changed to insulin if glibenclamide failed, were compared with the pregnancy outcome of 21 women with gestational diabetes treated conventionally with insulin. The pregnancy outcome, with regard to the overall glycaemic control, rates of preterm labour, neonatal hypoglycaemia, fetal macrosomia, perinatal morbidity and mortality, were not statistically different between the two treatment groups. The limited number of women studied, and the non-random allocation of these women to each treatment group however, could have influenced these results. There were a few observed differences in the pregnancy outcome between the two treatment groups, which although were not statistically significant, caused some concern. In particular we noted an increased rate of fetal macrosomia in the glibenclamide treated group, which in theory could have been drug mediated.