To evaluate the relationship between abnormal Doppler velocimetry and adverse perinatal outcomes in pregnancies with borderline amniotic fluid index (AFI). Medical records of 3740 pregnancies with known AFI and Doppler velocimetry measurements within 2 weeks of delivery, and delivered between January 1996 and December 2003, were retrospectively analyzed. Borderline AFI was defined as 5 cm < AFI < or = 8 cm. For the umbilical artery, Doppler velocimetry was considered abnormal when the S/D ratio was greater than 3.0 or when end-diastolic flow was absent; whereas, for the uterine artery, the S/D ratio was greater than 2.6 or presence of end-systolic notch was defined as abnormal. Adverse perinatal outcomes (small for gestational age, cesarean section for fetal distress, 5 min Apgar score of less than 7, respiratory distress syndrome, NICU admission, and perinatal death) according to Doppler velocimetry in borderline AFI were evaluated. Compared to the normal AFI group (n = 3523), the borderline AFI group (n = 217) had significantly higher incidence of adverse perinatal outcome (28.8% vs 54.8%). There was 3-fold increase in the incidence of adverse perinatal outcome among women with the borderline AFI in comparison to normal AFI (OR, 3.00; CI, 2.27-4.00). When abnormal Doppler velocimetry was associated with borderline AFI, a 5-fold increase in the incidence of adverse perinatal outcome was noted (OR, 5.26; CI, 3.00-9.21). In the borderline AFI group, the presence of abnormal Doppler velocimetry measurement was related to increased risk of adverse perinatal outcome. Thus, borderline AFI of 5-8 cm, especially when associated with abnormal Doppler velocimetry, mandates closer antenatal surveillance.