Two hundred seventy-two twin pregnancies were studied by analysis of the fetal umbilical artery waveforms recorded using continuous-wave Doppler ultrasound. After the first 100 cases, a management strategy was adopted whereby all twin gestations would be studied between 28-32 weeks gestation and the results made available to the referring obstetrician, thereby influencing clinical management. Perinatal mortality and morbidity, in pregnancies with both fetuses alive at 28 weeks gestation, were compared between the patients studied in the first group of 100 and those studied after the Doppler examination was introduced as a clinical service. There was a decrease in perinatal mortality, both uncorrected (57.9 per 1000 to 17.9 per 1000; P <.05) and corrected (42.1 per 1000 to 8.9 per 1000). Fetal deaths were reduced from six to one (P <.05). This decrease in perinatal mortality was achieved without any appreciable change in the gestational age at delivery or mode of delivery between the two groups. There was a reduction in the number of infants requiring neonatal intensive care (from 38% to 24%; P <.01).