The impact of maternal plasma volume expansion and antihypertensive treatment with intravenous dihydralazine on fetal and maternal hemodynamics during pre-eclampsia: a clinical, echo-Doppler and viscometric study




To establish the effects of plasma volume expansion (PVE) followed by intravenous dihydralazine (DH) administration on maternal whole blood viscosity (WBV) and hematocrit, uteroplacental and fetoplacental downstream impedance and umbilical venous (UV) volume flow in pre-eclampsia.


In 13 pre-eclamptic women maternal and fetal hemodynamics were established by means of combined measurement of maternal arterial blood pressure (BP), WBV, hematocrit and uterine artery (UtA) resistance index (RI) in addition to umbilical artery (UA) pulsatility index (PI) and UV volume flow obtained from UV vessel area and UV time-averaged flow velocity. In each woman all parameters were measured four times at baseline, after PVE, after DH and 24 h after the start of treatment.


Maternal diastolic BP, hematocrit and WBV display a significant reduction after PVE. In the fetus UA PI decreases significantly whereas a significant increase in UV cross-sectional area was detected. After maternal DH administration, arterial systolic and diastolic BP and UA PI show a significant decrease compared with the measurements following PVE. At 24 h, only maternal systolic and diastolic BP display a significant further decrease. No significant changes were established for the UtA RI, UV time-averaged velocity and UV volume flow during the entire study period.


During pre-eclampsia, maternal PVE followed by DH administration results in a significant reduction in maternal diastolic BP, maternal hematocrit and WBV. Maternal PVE is associated with a significant increase in UV cross-sectional area and a non-significant rise of 11% in UV volume flow. Maternal DH administration does not result in any change in UV cross-sectional area. However, UA PI decreases significantly after both PVE and DH treatment. Copyright © 2004 ISUOG. Published by John Wiley & Sons, Ltd.


In pre-eclampsia, maternal hemodynamics are characterized by relative hypovolemia, raised total peripheral resistance and impaired uteroplacental perfusion1, 2. Hematocrit is raised. The same applies to red cell aggregation, which especially affects low-shear flow such as in the intervillous space of the placenta and therefore may reduce oxygen supply to the fetus. An increase in red cell aggregation causes a steep rise in blood viscosity. Hemodilution has been proposed as an effective treatment in pre-eclampsia to lower the increased blood viscosity3, 4. This may not only be expected to improve uteroplacental perfusion, but also to lower maternal systemic blood pressure (BP), one of the symptoms of pre-eclampsia5. Maternal administration of antihypertensive drugs in addition to plasma expanders has been advocated to further reduce systemic vascular resistance and BP6. Dihydralazine (DH) is a drug frequently used in the treatment of pre-eclampsia. It has vasodilating properties, but crosses the placenta.

The aim of the present study was to establish the effects of maternal administration of plasma expanders followed by intravenous DH on maternal whole blood viscosity (WBV) and hematocrit, uteroplacental and fetoplacental downstream impedance and umbilical venous (UV) volume flow in 13 women with pre-eclampsia.



UV volume flow, umbilical artery (UA) and uterine artery (UtA) flow velocity waveforms were recorded in a consecutive series of 13 pre-eclamptic women with a singleton pregnancy. The women were enrolled from the obstetric high-care unit during the second half of 2002. Eight women were nulliparous and five were multiparous. Maternal age was in the range 25–39 (median, 30) years. The women were not known to have pre-existing hypertension, renal disease or heart disease. Pre-eclampsia was defined as the occurrence of a diastolic BP > 90 mmHg measured in a sitting position and proteinuria ≥ 0.3 g/L. A total of 5/13 patients displayed HELLP syndrome, which was defined as the simultaneous occurrence of a platelet count < 100 × 109/L, serum aspartate aminotransferase (ASAT) and serum alanine aminotransferase (ALAT) concentrations > 30 U/L and a haptoglobin value < 0.3 g/L7. Moreover, 7/13 women used the antihypertensive drug, methyldopa, before admission. In each woman a blood test was performed including a complete blood count, urine analysis, serum creatinine, uric acid, liver enzymes, WBV and platelet count before fluid infusion. Pregnancy duration was determined from the last menstrual period and confirmed by measurement of fetal crown–rump length or by biparietal diameter before 20 weeks' gestation. Fetal growth restriction, as defined by a fetal upper abdominal circumference measurement below the 5th centile, existed in 6/13 fetuses8. Gestational age varied between 27 and 33 (median, 30) weeks. On admission, UA pulsatility index (PI) was above the 95th centile of the normal reference chart9 in 4/13 fetuses. Individual clinical data and laboratory findings of the 13 pre-eclamptic women are presented in Table 1.

Table 1. Individual clinical data and laboratory findings of pre-eclamptic patients (n = 13) on admission
PatientAge (years)ParityGA (weeks + days)BP (mmHg)Associated HELLP syndromeUric acid (mmol/L)Proteinuria (g/L)Liver function testsPlatelet count (109/L)Haptoglobin (g/L)
  1. BP, blood pressure; GA, gestational age; M, multiparous; N, nulliparous.

 129N29 + 3180/1100.273.4Normal2250.72
 232N30 + 5170/1100.4322.0Increased182
 328M29 + 3165/1050.273.1Normal179< 0.05
 439M30 + 2202/1150.3011.2Normal1170.51
 533N31 + 5222/104+0.334.6Increased 66< 0.05
 727N27 + 3180/120+0.438.8Increased 890.05
 831N29 + 6150/95+0.300.6Increased 61< 0.05
 936M31155/107+0.383.0Increased 820.25
1025N30 + 4161/1140.300.4Increased1340.13
1130M30 + 5160/1000.291.0Increased 800.97
1230N27 + 5183/101+0.220.8Increased 84< 0.05

Maternal hematocrit, WBV in addition to UV volume flow, UA and UtA PI were determined on four different occasions. The first measurement (Measurement 1) served as baseline before any intravenous fluid treatment; the second measurement (Measurement 2) was carried out 1 h after starting maternal administration of pasteurized plasma as a volume expander. A total of 500 mL was administered during the first hour of treatment. The third measurement (Measurement 3) was taken 1 h after initiating maternal antihypertensive treatment by means of DH, which was administered intravenously 1 h after starting plasma infusion, as a bolus of 10 mg during the first 60 min followed by 1 mg/h or more depending on the patients' condition. The fourth and final measurement was performed between 20 and 25 h following the baseline measurement. During the first 24 h of treatment, fluid intake and diuresis were recorded and BP was either monitored continuously by means of an intra-arterial catheter in the maternal radial artery using a Hewlett-Packard monitor system (HP64S, McMinnville, OR, USA) (n = 7) or by means of an automatic BP device (Dinamap 1846 SX, Critikon Inc., Tampa, FL, USA) (n = 6). Arterial BP measurements were recorded and the mean of at least three values (between three and nine values) was taken over a period of 15 min for further analysis. Sonographic measurements were performed in the same time period as the maternal blood samples. In 10/13 women, after the third measurement had been performed, plasma administration was continued for another 24 h with a total amount ranging from 250 to 1000 mL. Antihypertensive treatment was aimed at achieving a diastolic BP ≤ 90 mmHg.

Whole blood viscosity determination

Blood samples (6 mL) were collected during the first 24 h, simultaneously with the four Doppler ultrasound measurements, for hematocrit determination and for WBV. For practical reasons the authors did not have the opportunity to perform viscosity measurements in all patients. Data were available in a subset of 10/13 women for hematocrit determination and in a subset of 8/13 women for WBV determination. The latter was determined using the Contraves LS 30 cone-plate viscometer (Contraves, Zurich, Switzerland) allowing measurements at different shear rates (0.19, 0.87, 4.04, 18.74 and 87 s−1). The temperature was fixed at 37°C. For statistical analysis the WBV at the lowest (0.19 s−1), medium (4.04 s−1) and highest (87 s−1) shear rates was selected.

Umbilical venous volume flow

A Toshiba SHH 140A (Toshiba Corporation, Medical Systems Division, Tokyo, Japan) unit was used for all Doppler and B-mode ultrasound recordings.

UV cross-sectional area (mm2) was expressed as the mean of three tracings of the inner edge of the vessel10. The software program used to obtain these tracings was developed in the authors' department using Labview Imaq Vision software (National Instruments, Austin, TX, USA). UV maximum flow velocity (mm/s) was measured using a combined two-dimensional (2D) real-time color-coded Doppler system with a transducer carrier frequency of 3.75 MHz. The system operates at output intensities of < 100 mW/cm2 spatial peak temporal average in both imaging and Doppler mode. Sample volume length for all flow velocity waveforms was in the range 0.2–0.3 cm, and the high-pass wall filter was set at 100 Hz. All ultrasound studies were carried out with the woman in the semi-recumbent position. All ultrasound/Doppler recordings were performed by the same investigator (S.B.).

A more detailed description of both vessel area and UV flow velocity measuring techniques for calculation of volume flow has been presented in detail elsewhere10. In brief, UV volume flow was calculated according to the formula: UV volume flow (mL/min) = 0.06 × time-averaged velocity (mm/s) × cross-sectional vessel area (mm2).

Arterial Doppler data acquisition

UA flow velocity waveforms were recorded in all patients from the ascending vasculature of the UA bilaterally11. Flow velocity waveforms were defined as abnormal if there was persistent notching, or if the resistance index (RI) was > 95th centile whether or not a notch was present12. UA flow velocity waveforms from the free loop of the umbilical cord were recorded for calculation of the PI. All Doppler recordings and 2D real-time images were stored on SVHS videotape in PAL format using a Panasonic AG 7350 (Matsushita Electric Industrial Co., Takutsuki, Osaka, Japan) for off-line analysis.

Statistical analysis

All calculations were performed with the SPSS 10 software package (SPSS Inc., Chicago, IL, USA). Data were reported as median values with ranges. Non-parametric, two-way analysis of variance (Friedman) and the Wilcoxon matched-pair signed rank test were used to assess differences between paired variables such as measurements for maternal systolic and diastolic BP, hematocrit and WBV in addition to fetal and maternal Doppler ultrasound data. A value of P < 0.05 was considered to be statistically significant.


Table 2 demonstrates median values with ranges for maternal arterial systolic and diastolic BP, maternal hematocrit and WBV in addition to UtA, UA and UV Doppler ultrasound data at each of the four measurements. Median maternal arterial systolic BP only becomes significantly reduced after DH administration and 24 h following the baseline measurement. Median maternal arterial diastolic BP already displays a significant reduction after administration of the plasma expander. The same pattern exists for median maternal hematocrit, median maternal WBV (apart from the low-shear rate subset at 24 h) and median UA PI. Median UV cross-sectional vessel area increases significantly (apart from the measurement at 24 h). No significant changes were established for the median UtA RI, median time-averaged velocity and median UV volume flow during the entire study period.

Table 2. Median and range values for maternal arterial systolic and diastolic blood pressure (n = 13), hematocrit (n = 10), maternal blood viscosity measurements (n = 8) and fetal/maternal Doppler flow parameters (n = 13)
1 (baseline)2 (after PVE)3 (after dihydralazine)4 (after 24 h)
  • *, *†

    P ≤ 0.05 Measurements 2, 3 and 4 compared with Measurement 1.

  • P ≤ 0.05 Measurement 3 compared with Measurement 2.

  • P ≤ 0.05 Measurement 4 compared with Measurement 3. BP, blood pressure; PVE, plasma volume expansion; UV, umbilical venous; WBV, whole blood viscosity.

Maternal arterial systolic BP (mmHg)175150–222165140–218150*140–205140*117–170
Maternal arterial diastolic BP (mmHg)10690–120101*77–12590*75–10580*59–87
Maternal hematocrit (%)35.031.0–44.532.5*28.0–42.533.5*29.0–41.532.5*27.0–40.0
Maternal WBV (mPa*s) at low shear rate (0.187 s−1)17.88.42–30.016.0*6.5–31.416.4*8.3–28.616.44.4–26.2
Maternal WBV (mPa*s) at medium shear rate (4.04 s−1)6.95.3–11.05.2*4.2–10.75.6*3.7–9.95.5*3.5–8.1
Maternal WBV (mPa*s) at high shear rate (87 s−1)4.53.42–5.183.6*3.1–5.33.8*3.8–4.03.7*2.9–4.2
Uterine artery resistance index0.630.5–0.90.640.5–0.80.640.4––0.8
Umbilical artery pulsatility index1.510.9–2.21.32*0.9–1.961.21*†0.8–1.71.25*0.8–1.7
UV cross-sectional area (mm2)33.622.2–43.037.7*27.6–52.338.7*25.4–47.834.427.6–53.5
UV time-averaged velocity (mm/s)55.144.2–81.458.143.2–70.256.535.7–69.556.335.6–77.9
UV volume flow (mL/min)109.258.8–164.6121.477.9–215.6127.354.3–189.5116.164.5–202.4

After maternal DH administration, median arterial systolic and diastolic BP and UA PI show a statistically significant decrease compared with the measurement following plasma volume expansion (PVE). At 24 h, only median maternal arterial systolic and diastolic BP display a significant further decrease.

Figure 1 shows the relationship between maternal WBV and shear rate at each of the four measurements. Note that median values for maternal WBV after PVE and after DH administration coincide.

Figure 1.

Relationship between maternal whole blood viscosity (mPa*s) and shear rate (s−1) at each of the four measurements. Baseline measurement (○), measurement after plasma volume expander (✶), dihydralazine treatment (●) and after 24 h (▵).

Abnormal UtA flow velocity waveforms were present in 8/13 women and persisted throughout the four measurements.

Median gestational age at delivery was 32 (range, 28–34) weeks and the median lag time between ultrasound examinations and delivery was 7 (range, 2–30) days. Median umbilical cord pH at delivery was 7.23 (range, 7.01–7.30) and median fetal birth weight was 1350 (range, 635–2305) g. A total of 3/13 newborns had a birth weight below the 5th centile according to the Kloosterman tables adjusted for maternal parity and fetal sex13.


The objective of this study was to assess circulatory effects of hemodilution and antihypertensive treatment with the vasodilator DH in pre-eclampsia on both mother and fetus. In the pre-eclamptic syndrome an increased vascular resistance and a diminished circulating blood volume with impaired perfusion of various organs, including the uteroplacental unit, are important pathophysiological features14.

Generally, antihypertensive drugs are administered in the course of pre-eclampsia to avoid maternal vascular complications induced by elevation of maternal arterial BP. Their administration may result in a marked reduction in BP affecting both the maternal14 and fetal circulations15. Therefore, it has been suggested that antihypertensive treatment, especially with a vasodilator, should be preceded by administration of a plasma expander so as to normalize maternal circulatory volume14, 16.

Effects of maternal plasma volume expansion

Maternal site

In the present study a small but significant fall in maternal arterial diastolic BP following PVE was established. This is in line with some studies17–19 but not with others14, 20. It is known that PVE will give rise to an increase in maternal cardiac output in addition to a reduction in peripheral vascular resistance. The latter is documented by the drop in maternal arterial diastolic BP. In pre-eclampsia there is a considerable change in the rheological properties of maternal blood, including an increase of red cell aggregation, which causes a steep rise in WBV. PVE will lead to hemodilution as expressed by a lower hematocrit. This will result in a drop in maternal WBV. In the present study, baseline measurements of maternal WBV were above the normal range for normotensive pregnant women according to Buchan21. A significant reduction in maternal WBV was observed after PVE at virtually all shear rates. However, the viscosity drop is more pronounced at the lowest shear rate as shown in Figure 1. Since the viscous forces dominate in the capillary system where shear rates are low, a flow resistance reduction can be expected and can play an important role in regulating the flow-state. This suggests a relationship between reduction in viscosity and drop in BP after PVE. A further explanation for the drop in maternal arterial diastolic BP could be that PVE induces an activation or release of a substance at endothelial level such as nitric oxide and prostacyclin, with vasodilatation as a result22. Moreover, another pathway that seems to be involved in the possible beneficial effects of PVE is the suppression of the renin–angiotensin system at the uteroplacental level23, 24.

We could not observe any change in the UtA resistance index after hemodilution. This is in agreement with another report25 and could be explained by the assumption that the uteroplacental circulation is a low-impedance vascular system and therefore the degree of pulsatility does not reflect downstream impedance26, 27.

Fetal site

In pre-eclampsia, an association between increased impedance in the fetal circulation and abnormal fetal blood rheology with or without fetal growth restriction has been described by several authors 28, 29. For example, Buchan30 reported a higher median WBV, hematocrit and fibrinogen concentration in cord blood of pregnancies complicated by pre-eclampsia as compared with controls. In the present study, UA PI decreased following maternal PVE, reflecting reduced downstream impedance at the fetoplacental level. This is in agreement with a previous report31 in which a reappearance of end-diastolic flow velocity in the UA was established after hemodilution.

Hemodilution was associated with a significant increase in UV cross-sectional vessel area, resulting in a non-significant increase of 11% in UV volume flow. Regulation of blood volume between the fetus and the placenta is controlled by the different pressure gradients between placental and the umbilical blood flow32. In an animal experimental study33, it was reported that hypervolemic hemodilution with albumin leads to slight fetal hemodilution, which can be identified by a further decrease in fetal colloidosmotic pressure. The elevated hydrostatic pressure in the intervillous space, induced by PVE, may be the mechanism underlying the net flow of water into the fetal circulation. Our data are at variance with a previous study34 in which a significant increase in UV volume flow was observed after maternal hemodilution with 500 mL dextran.

Effects of dihydralazine treatment

Maternal site

Both maternal arterial systolic and diastolic BP are significantly reduced following administration of DH. However, no significant changes were observed in UtA RI, which may be due to its state of maximum dilatation35–37 nor in maternal hematocrit and maternal WBV.

Fetal site

DH crosses the placental barrier into the fetal circulation38, which may explain the further reduction in fetoplacental downstream impedance expressed by the UA PI. This does not exclude, however, a possible prolonged plasma expander effect. DH administration is not associated with a further increase in UV cross-sectional vessel area. Although Jouppila et al.39 reported no change in the intervillous blood flow, they found an elevated UV volume flow following DH administration. This is at variance with our data, which demonstrated an unaltered UV perfusion.

In the human fetus there is a lack of information concerning the interaction between maternal and fetal control mechanisms, nonetheless fetal well-being may deteriorate after the mother is necessarily treated with antihypertensive drugs. Indeed, DH as a vasodilating agent increases vascular space and induces a further decline in the already reduced circulatory filling. Depending on the severity of the relative hypovolemic state, compensatory mechanisms may fall short resulting in an increasingly insufficient perfusion of maternal organs, including the uteroplacental unit, which may also lead to fetal distress. These adverse effects of antihypertensive therapy can be avoided by correcting circulating volume by PVE14. Besides, Heyl et al. showed that in the event of fetal growth restriction, 14 successive days of treatment with hydroxyethylstarch in combination with Ringer appears to be a promising approach to counteract restricted fetal growth40. This combined treatment seems to be useful for the fetus according to the improvement of the UA Doppler and UV cross-sectional area results within therapy. However, PVE carries a serious risk of volume overload, which may lead to pulmonary and cerebral edema in pre-eclamptic patients with low colloid osmotic pressure and capillary leakage. For that reason, PVE should not be applied without monitoring pulmonary capillary wedge pressure7.

It can be concluded that during pre-eclampsia, maternal administration of PVE, followed by antihypertensive treatment by means of DH, led to a significant reduction in maternal diastolic BP, maternal hematocrit and WBV. Maternal PVE is associated with a significant increase in UV cross-sectional vessel area and a non-significant rise of 11% in UV volume flow. Maternal DH administration does not result in any change in UV cross-sectional vessel area. However, UA PI, as a measure of fetoplacental downstream impedance, decreases significantly after both plasma and DH treatment. We suggest that fetal hemodynamic Doppler ultrasound monitoring is of significant importance in achieving a balance between benefit to the mother and risk to the fetus during antihypertensive treatment in the course of pre-eclampsia.