Objective: the aim of our study was to evaluate the efficacy of Doppler velocimetry and cardiotocography in surveillance of the fetus affected by severe fetal growth retardation (FGR) and hence their capacity in predicting adverse perinatal outcome. This could therefore permit the timing of delivery thus reducing perinatal morbidity and mortality. Methods: 53 women with high risk pregnancies, all suffering from various pathologies such as gestational hypertension, pre­eclampsia, HELLP syndrome, Antithrombin III deficit, and in whom FGR (fetal abdominal circumference <2.5th percentile for gestational age) diagnosed between the 24th and 35th week of gestational age, were retrospectively enrolled in the study. Doppler velocimetry was carried out on the main vascular districts - both arterial [umbilical artery (UA)，middle cerebral artery (MCA), aorta (Ao) and uterine artery] and venous [umbilical vein (UV), ductus venosus (DY), inferior vena cava (IYC)]. Also evaluated was the amniotic fluid index (AFI). Daily non stress tests (NST) were conducted at least thrice a day for an overall period of not shorter than 60 min. Also considered were the methods of delivery and the perinatal outcome, e.g. gestational age at birth, perinatal mor­tality, incidence resuscitation, etc. Results: The 53 patients studied were subdivided into three groups on the basis of the initial velocimetric exam of the umbilical and uterine arteries. Group A was comprised of those with altered waveforms of the UA, Group B those with altered waveforms on the uterine artery while Group C contained those with altered waveforms both of the UA and uterine artery. The period of study for group C was on average six days (p < 0.05) while for groups A and B they were on average 10 and 22 days, respectively. Group C also presented higher incidences of altered waveforms in the venous compartment, i.e. absent or reversed end diastole (ARDEF) was observed in 89% of the cases versus 0% and 7.6% observed in Group A and B, respectively. Altered waveforms in the DY and pulsations in the UV were both observed in Group C while the same was not observed in either group A or B. The NST did not show any substantial difference between the groups - only as the appearance of decelerations present in over half the cases in all groups at the end of the study. Group C also presented higher incidences in adverse perinatal outcome as compared to groups A and B such as as low birth weight (868 g vs 1,324 g & 1,397 g, p < 0.001), neonatal resuscitation (52.6% vs 0% & 7.6% p = 0.001), longer periods of admission to neonatal intensive care unit (67 days vs 32 & 33 p < 0.001) and perinatal mortality (36% vs 0 & 0, p < 0.05). The velocimetric indices which appear to better predict perinatal mortality are those related to the venous compartment g1v10g a diagnostic accuracy of 92.8% (Kl > 0.75) in case of pulsations in the UV, 86.6% (Kl > 0.75) with alterations on the DY, and of 78.5% (KI > 0.40) for those on the IVC. Cardiotocography revealed to be less capable in predicting perinatal mortality giving a diagnostic accuracy of 66.6% (Kl > 0.40) with a non reactive, non variable NST. Conclusion: The data presented show that velocimetric modifications in the fetal venous compartment constitute a relevant pro­gnostic sign in the prediction of perinatal mortality and neonatal resuscitation. The further the vessels compromised are from the heart, the higher is the relative risk for perinatal mortality, equal to 5.0 (95% C.I. = 0.61-40.9) with alterations on the IVC, of 8.2 (95% CI= 1.04-61.5) when they involve the DY, and of 18.0 (95% = 2.44-133) when pulsations are obtained on the UV.