Most studies of gestational diabetes mellitus (GDM) have reported a marked reduction in perinatal mortality with appropriate dietary regimens and good medical and obstetrical surveillance. Nevertheless, fetal morbidity, including macrosomia, has remained high and appears to be linked to factors other than plasma glucose control. In a review of six investigations in which insulin therapy was combined with an appropriate diet, the incidence of fetal macrosomia was reduced in five studies as compared with diet-only treatments. Again, the improvement did not always correlate with altered plasma glucose profiles. Other studies suggest that maternal plasma substrate disturbances other than glucose may contribute to the development of fetal macrosomia. To what extent insulin administration reduces morbidity by containing circulating maternal fuels, such as lipids and amino acids, in a more normal range remains to be determined. Moreover, the role of diet, maternal obesity, and weight gain during pregnancy adds to the complexity of factors influencing obstetrical outcome in gestational diabetes. Until the relative importance of all of these variables is adequately assessed, criteria for selection of women with pregnancy-onset diabetes for insulin therapy are most likely to be based on fasting and postprandial plasma glucose concentrations.