OBJECTIVE To test whether the human fetus accommodates to the increasing glucose requirements of late pregnancy with an increased maternal-fetal glucose concentration gradient and whether there are differences in pregnancies with fetal growth restriction (FGR) according to clinical severity. METHODS Umbilical venous glucose concentration was measured in 77 normal pregnancies (appropriate for gestational age [AGA]) and 42 pregnancies complicated by FGR at the time of fetal blood sampling. In 40 AGA and in all FGR cases, a maternal "arterialized" blood sample was collected simultaneously. Growth-restricted fetuses were subdivided into three groups according to fetal heart rate (FHR) recordings and Doppler measurements of the umbilical artery pulsatility index (PI): group 1 (normal FHR and PI; 12 cases), group 2 (normal FHR, abnormal PI; 17 cases) and group 3 (abnormal FHR and PI; 13 cases). RESULTS In normal pregnancies with increasing gestational age, there was a significant decrease (P < .001) of umbilical venous glucose concentration and a significant increase of the maternal-fetal glucose concentration difference (P < .001). In addition, there was a significant relation between fetal and maternal glucose concentrations (P < .001). In FGR pregnancies, the maternal-fetal glucose concentration difference was significantly higher in fetuses of groups 2 and 3 compared with normal pregnancies and FGR pregnancies of group 1. CONCLUSIONS In human pregnancy, the fetal glucose concentration is a function of both gestational age and the maternal glucose concentration. In FGR pregnancies, as an accommodation of the fetus to a restricted placental size and placental glucose transport capacity, the maternal-fetal glucose concentration difference is increased, and this increase is a function of the clinical severity.