Correspondence: Dr H. Ochi, Department of Obstetrics and Gynaecology, Ehime University School of Medicine, Shigenobu, Ehime 79142, Japan.
Objective To investigate the relation between placental embolisation and the diastolic notch in the uterine artery flow velocity waveform of pregnant ewes under general anaesthesia.
Methods Seven pregnant ewes at a gestation 16 to 17 weeks were anaesthesized and micro beads of gelfoam were injected into the uterine artery; changes in the uterine circulation were assessed by Doppler velocimetry.
Results Gelfoam embolisation reduced uterine blood flow in a dose-dependent manner, from a mean (95% CI) of 568 mL/min (495–641) to 159 mL/min (131–187) after the injection of 30 mg of gelfoam, and increased the uterine vascular resistance from 135 mmHg mine L−1 (103–167) to 498 mmHg mino L−1 (422–574). A diastolic notch in uterine artery flow velocity waveform was observed after 20 mg to 25 mg of gelfoam in two ewes and after injection of 30 mg of gelfoam in all seven animals. Injection of 30 mg of gelfoam increased the pulsatility index to 2–4 (1.9–2.9) from 0.6 (0.5–0.7). The mean uterine vascular resistance at the time of the appearance of a diastolic notch was 414 mmHg mine L−1 (377–451).
Conclusion These findings suggest that an elevated pulsatility index and the presence of a diastolic notch in the uterine artery flow velocity waveform are indicators of increased uterine vascular resistance and impaired uterine circulation.
Because only invasive techniques have been available to evaluate the uterine circulation, their use in humans has been limited. Results obtained in experimental animals do not reflect the in vivo changes in physiological and pathological conditions. Technical advances have now made it possible to assess non-invasively the uteroplacental circulation by Doppler velocimetry in normal and complicated human pregnancies. The results of studies using Doppler velocimetry have shown that complicated pregnancies are characterised by the inhibition of the pregnancy-related physiological increase in diastolic flow velocity1–3 and the presence of diastolic notches4,5 in maternal uterine artery flow velocity waveforms. Although it is likely that the decreased diastolic flow velocity and the presence of an early diastolic notch reflect an increase in uterine vascular resistance and a decrease in uterine blood flow and that these findings may precede fetal compromise, there is no direct evidence to support this hypothesis.
We administered gelfoam into the uterine arteries of pregnant ewes to occlude the uterine spiral arteries, as is characteristic of pregnancy-induced hypertension and analysed changes in the uterine circulation by Doppler velocimetry to investigate this hypothesis.
Seven pregnant ewes at a gestation 16 to 17 weeks were anaesthesized. Intramuscular ketamine hydrochloride (300 to 500 mg) was followed by inhalation of 1.0 to 1.5% enflurane in 50% oxygen in air through an indwelling tracheal tube. An electromagnetic flowmeter (Nihon Koden Model 1100, Tokyo, Japan) and a continuous wave Doppler apparatus (Parks Electronics Model 1010-LA, Salem, Oregon, USA) were attached to the main uterine artery of the pregnant horn. A small branch of the ipsilateral uterine artery, distal to the flow transducer and the Doppler apparatus, was cannulated in the direction toward the main uterine artery. The tip of the cannula was placed close to, but not reaching, the main uterine artery. Thus a whirlpool or other forms of artifactual blood flow that can occur if the tip of the cannula reaches the main uterine artery was avoided. After haemodynamic stabilisation (20 to 40 min after the set-up), bolus doses of 5 mg of gelfoam powder (Upjohn, Tokyo, Japan) up to 30 mg were administered into the main uterine artery at 30 to 40 min intervals. The diameters of the gelfoam particles were 50 km. The uterine arterial blood flow and flow velocity waveforms were measured and recorded 10 to 15 min after each injection of gelfoam. Maternal haemodynamic parameters, such as mean arterial pressure, mean venous pressure and pulse rate, were monitored via catheters introduced through the femoral vessels and placed in the iliac vessels. Cardiac output was determined by the dye dilution method. The pulsatility index was calculated as previously described6, and the appearance of a diastolic notch was determined to evaluate experimentally induced changes in uterine arterial flow velocity waveforms. Uterine vascular resistance (UVR) was defined as the quotient of the decrease in blood pressure across the uteroplacental circulation; it was calculated with arterial and venous pressures measured through the iliac catheters, divided by the uterine blood flow7,8. Five consecutive waveforms were analysed, and the results were averaged for the calculation of the pulsatility index. We calculated the mean value of UVR at the time of the appearance of diastolic notches.
After completion of the experiment, the uterus and placenta were excised for histological examination. Each specimen was fixed in a 10% formaldehyde solution, embedded in paraffin, sliced and stained with hematoxylin-eosin stain; values are given as mean and 95% confidence intervals. An analysis of variance test was used for statistical assessment of hemodynamic parameters. The relation between the pulsatility index and UVR was determined by Spearman rank correlation coefficient. The comparison between UVR before the embolisation and UVR at the time of the appearance of diastolic notches was carried out using Wilcoxon signed rank test.
Injection of gelfoam powder embolised the small uterine arteries or spiral arteries located in the decid-ual segment, but there were no histological changes in the larger arteries or the placental vessels (Fig. 1). There were no significant changes in the maternal general circulation throughout the experiment, indicating that the pre-uteroplacental circulation was not affected by gelfoam embolisation. Gelfoam embolisation significantly reduced uterine blood flow and increased UVR in a dose-dependent manner (Table 1). Before embolisation, the flow velocity waveform in a representative pregnant ewe was characterised by a systolic peak tailing off towards diastole with a fairly constant and elevated flow velocity (Fig. 2a). Gelfoam caused a dose-dependent decrease in the diastolic flow velocity but did not affect systolic flow velocity (Fig. 2 b-d). A diastolic notch was observed after administration of 20 to 30 mg of gelfoam (Fig. 2 c, d). These flow velocimet-ric changes resulted in a dose-dependent increase in the pulsatility index (Table 1).
Table 1. Haemodynamic parameters in gelfoam-embolised pregnant ewes. Values are given as mean (95% confidence intervals). IAP = iliac arterial pressure; IVP = iliac venous pressure; UBF = uterine blood flow; UVR = uterine vascular resistance; PI = pulsatility index.
Gelfoam dose (mg)
Cardiac output (L/min)
79.4 (74. 1–84.7)
Pulse rate (bpm)
UVR (mmHg. mino L−1)
1.6 (1–2 2–0)
There was a significant correlation between the pulsatility index and UVR in each pregnant ewe (Fig. 3). A diastolic notch in the uterine artery flow velocity waveform was observed after 20 mg to 25 mg of gelfoam in two ewes and after injection of 30 mg of gelfoam in all seven animals. The mean (95% CI) UVR at the time of the appearance of a diastolic notch was 414 mmHg• min• L−1 (377–451). This value was significantly higher than the value before the embolisation (135 mmHg• min• L−1; P = 0.018 (103–167)).
In previous experimental studies the major branches of the uterine arteries were ligated to reduce the uteroplacental circulation, reflecting the human pathological uteroplacental circulation associated with pregnancy-induced hypertension (PIH)9. It is, however, unlikely that this type of impairment in the uterine circulation occurs in vivo, since even in severe PIH the major branches of the uterine artery are not occluded. In the present study we used gelfoam to embolise uterine arterioles or spiral arteries instead of ligating the major branches of the uterine artery in attempt to induce the circulation consistent with those observed in human PIH9. Gelfoam embolised uterine arterioles or spiral arteries within the decidual segment and did not affect the major branches of the uterine artery. Clapp et al.12 failed to embolise uterine spiral arteries using microspheres. This discrepancy is attributable to the difference in particle size. The gelfoam particles have a diameter of 50 pm whereas the microspheres used by Clapp et al. have a diameter of 15 pm. The changes in uterine flow velocity waveforms induced by gelfoam embolisation in pregnant ewes in the present study appeared to be consistent with the changes seen in human PIH.
Abnormalities in uterine artery flow velocity waveforms, such as an elevated pulsatility index and diastolic notches in women with PIH are correlated with an abnormal perinatal outcome1,3, but the relation of these waveform abnormalities to uterine blood flow and UVR has not been clarified.
Gelfoam embolisation inhibited in a dose-dependent manner the pregnancy-related increase in diastolic flow velocity. A diastolic notch was observed after the adminstration of 20 to 30 mg of gelfoam, but systolic velocity was unchanged. These findings are consistent with features of PIH, in which the uterine arterial flow velocity waveform is characterised by a prominent systolic peak and a deep diastolic basin1,3. Thus, gelfoam-induced embolisation of the utero-placental vessels successfully induced the haemodynamic changes typical of PIH.
The pulsatility index and UVR showed a significant correlation in the present study, suggesting that the uterine arterial pulsatility index is inversely correlated with uterine blood flow. A decreased uterine blood flow or an increased uterine arterial pulsatility index can lead to fetal compromise. Previous studies13–15 have shown an increased incidence of fetal compromise in the presence of an elevated uterine arterial pulsatility index.
The mean value of UVR upon the appearance of the diastolic notch reflected a markedly elevated UVR and a severely impaired uterine circulation. Recent studies4,5 have investigated the significance of a diastolic notch in PIH. Park et al.4 reported that the presence or absence of a diastolic notch is a more useful predictor of perinatal outcome than is an elevated systolic:diastolic ratio alone.
The fact that the diastolic notch appeared only in the presence of a markedly elevated uterine vascular resistance in the gelfoam-embolised ewes suggests that they indicate a severe impairment of the utero-placental circulation. Although the pulsatility index was strongly correlated with uterine vascular resistance, other maternal circulatory factors such as the heart rate, haematocrit and blood pressure, can influence the uterine artery flow velocity waveforms and pulsatility index16. Thus, the use of the pulsatility index alone to assess the uteroplacental circulation may lead to some false-negative results. The present findings suggest that an elevated pulsatility index and the presence of a diastolic notch, which indicate a severely impaired uteroplacental circulation, can predict with greater accuracy a poor perinatal outcome in women with pregnancy-induced hypertension than the use of the pulsatility index alone.