Macrosomia is an important cause of neonatal morbidity. To detect the utility of plasma insulin level to predict neonatal macrosomia, we studied the glycemic and insulin responses to 100 g oral glucose tolerance test (OGTT) in 57 gestational diabetic women on diet therapy. Ten (17.5%) of them had macrosomic newborns. Macrosomia was defined as a birthweight higher than the 90th centile on our population scale. The mothers of the macrosomic newborns exhibited similar parity, body mass index, gestational age at OGTT and at delivery. Fasting capillary plasma glucose value was higher in women with macrosomic newborns (5.3 mmol/l +/- 0.4 vs 4.8 mmol/l +/- 0.4; p less than 0.01) as well as fasting capillary plasma insulin concentration (28.7 mU/L +/- 9.3 vs 17.7 mU/L +/- 6.4; p less than 0.01). The fasting glucose value was found to be more accurate than fasting insulin concentration to predict macrosomia (logistic regression and discriminant analysis). Thus, measuring insulin concentrations during an OGTT seemed to be unhelpful to predict macrosomia in gestational diabetic women.