Maternal haemodynamics in pre-eclampsia compared with normal pregnancy during caesarean delivery


Dr K Tihtonen, Department of Obstetrics and Gynecology, Tampere University Hospital, PL 2000, 33521 Tampere, Finland. Email


Objective  To determine how pre-eclampsia modifies maternal haemodynamics during caesarean delivery.

Design  Prospective study.

Setting  Tampere University Hospital, Finland.

Population  Ten pre-eclamptic parturients and ten healthy parturients with uncomplicated pregnancies scheduled for elective caesarean section under spinal anaesthesia.

Methods  Haemodynamic parameters were assessed by whole-body impedance cardiography noninvasively.

Main outcome measures  Stroke index (SI), heart rate (HR), cardiac index (CI), systemic vascular resistance index (SVRI) and mean arterial pressure (MAP) were recorded before operation, continuously during caesarean section, during the period of dissipation of anaesthesia and on the second to fifth postpartum day.

Results  Baseline haemodynamics in women with pre-eclampsia differed significantly from healthy women in higher SVRI and MAP and lower SI and CI. In women with pre-eclampsia, preload infusion increased both SI and HR, causing a significant rise in CI, while in healthy parturients, only HR rose. In both the groups, spinal blockade reduced SVRI but CI remained stable. At the moment of delivery, CI increased in both groups. In uncomplicated pregnancies, both SI and HR increased, but in women with pre-eclampsia, SI was not altered and the rise in CI was due to an increase in HR only. After the reversal of anaesthesia, haemodynamics in the control group returned to baseline values, whereas in women with pre-eclampsia, SI and CI fell to levels that were significantly lower than the levels observed before surgery.

Conclusions  In women with pre-eclampsia, inability to increase SI at the moment of delivery may suggest dysfunction of the left ventricle to adapt to volume load caused by delivery and prompts concern for the increased risk of pulmonary oedema.


Pre-eclamptic parturients have an increased risk for caesarean section by reason of the high incidence of intrauterine growth restriction, fetal distress and prematurity.1 Unfortunately, caesarean section increases the risk of cardiopulmonary morbidity associated with pre-eclampsia.2 Pre-eclampsia has also been shown to be a significant risk factor for postoperative cardiac failure.3 Divergent haemodynamic adaptation during the pregnancy may render pre-eclamptic parturients prone to cardiopulmonary complications during caesarean section.

Caesarean delivery involves interventions potentially harmful for haemodynamics. Sympathetic blockade induced by regional anaesthesia disposes to hypotension4–6 and exposes the fetus to the risk of hypoxaemia especially in cases of intrauterine growth restriction. The effects of hypotension may be intensified by the low cardiac output often found in women with pre-eclampsia.7,8 Also, the moment of delivery constitutes a challenge. We and others have shown that in healthy parturients, delivery by caesarean section is associated with a significant rise in cardiac output,9–13 suggesting a rapidly increased volume load in the vasculature. Cardiac output is also known to rise in women with pre-eclampsia at the moment of delivery under general14 and epidural anaesthesia,15 but how pre-eclampsia modifies haemodynamic adaptation to delivery in respect of factors apart from cardiac output is not known.

Since the maternal haemodynamics in pre-eclampsia differ from those in uncomplicated pregnancy,7,8,16,17 we hypothesised that cardiovascular reactions during the course of caesarean section will also differ between pre-eclamptic and healthy parturients. The objective of this study was to define haemodynamic changes during caesarean section in pre-eclamptic parturients. Reference data were obtained from ten healthy pregnant women with uncomplicated pregnancy undergoing elective caesarean delivery, published separately in an earlier paper.13


The study was carried out at Tampere University Hospital. The procedure was explained to each subject and a written consent was obtained. The local Ethics Committee approved the study.

Ten women with pre-eclampsia undergoing elective caesarean delivery were studied. Parturients with pre-existing hypertension, renal disease or heart disease were excluded. Pre-eclampsia was diagnosed if blood pressure exceeded 140/90 mmHg and if proteinuria of more than 300 mg/day persisted.

Of the ten women with pre-eclampsia, six met the criteria of severe disease, with an arterial pressure constantly greater than 160/110 mmHg and proteinuria exceeding 5.0 g/day. Three women received an antihypertensive peroral regimen of labetalol 200–600 mg/day and one was given a labetalol infusion of 0.35 mg/minute (labetalol 500 mg in 500 ml of 0.9% NaCl). No other medication was used.

The control group comprised ten healthy parturients with normal pregnancy undergoing elective caesarean section. The description of these women is given in an earlier publication.13 The main characteristics of the women are shown in Table 1.

Table 1.  Subject details
Subject detailsPre-eclamptic parturients, n= 10Healthy parturients, n= 10P
  1. NS, not significant. Values are expressed as mean (range).

Age (years)34 (24–43)30 (24–38)NS
Percentage of primiparous8050NS
Gestational age at delivery (weeks)34 (28–38)40 (39–41)<0.0001
BMI (kg/m2)26 (22–30)30 (25–41)0.049
Systolic blood pressure (mmHg)166 (137–204)127 (101–171)0.001
Diastolic blood pressure (mmHg)105 (90–122)66 (61–72)<0.0001
Daily proteinuria (g/day)6.3 (1.52–11.35)<0.3<0.0001
Haematocrit (%)35 (31–39)36 (33–40)NS

Since gestational age at delivery differed between healthy controls and pre-eclamptic parturients, the baseline haemodynamics of an additional ten healthy primiparous women were recorded at gestational week 34 (the mean gestational age for pre-eclamptic parturients). These ten women had an uncomplicated pregnancy and were matched by age and body mass index (BMI) with those of healthy parturients. The baseline haemodynamic measurements were not different between the healthy parturients at 34 weeks of gestation and the control group of the study. Thus, the difference in gestational age between the study groups was not considered to influence the results.

CircMon B202 (JR Medical Ltd, Tallinn, Estonia) was used for the noninvasive estimation of heart rate (HR), stroke volume, cardiac output and systemic vascular resistance. This cardiovascular monitoring device applies whole-body impedance cardiography to measure cardiac output. Impedance cardiography is based on the measurement of changes in electrical impedance in the vasculature during the cardiac cycle. Current and voltage electrodes were applied to the wrists and to the ankles. Impedance changes during systole and diastole were used to calculate stroke volume. A more detailed description of this method is available in other publications.18,19

Stroke volume (quantity of blood ejected from the left ventricle with each beat), cardiac output (stroke volume × HR) and systemic vascular resistance (impediment to blood flow in peripheral vasculature) were indexed to body surface area to minimise the effect of the body size on these parameters (i.e. cardiac index [CI], stroke index [SI] and systemic vascular resistance index [SVRI]).

Blood pressure was measured by an automated noninvasive arterial pressure device (Hewlett-Packard M1960A, Boeblingen, Germany) for every 5 minutes and after the induction of spinal anaesthesia every 2 minutes. Mean arterial pressure (MAP) was calculated as diastolic blood pressure plus one-third of pulse pressure.

The baseline haemodynamic values were recorded prior to preload infusion in left lateral position. An intravenous catheter was then inserted into a peripheral vein. Women in both groups received hydroxyethyl starch (6%) 10 ml/kg over 15–30 minutes as preloading infusion. Thereafter, hydroxyethyl starch (6%) 10 ml/kg/hour was infused during the operation.

Once preloading was completed, whole-body impedance cardiography recording was continued throughout the operation. The recording was completed after the operation, when anaesthetic effect of spinal blockade had disappeared and the woman was able to bend and extend her ankles.

Spinal anaesthesia was induced after fluid infusion. It was performed using a 25- or 27-gauge spinal needle at the L2–L3 or L3–L4 intervertebral space, with the woman in the right lateral position. The women received a mean 2.5 ml (range 2.4–2.7 ml) of 0.5% hyperbaric bupivacaine. Thereafter, they were repositioned supine with slight left lateral position. The surgical procedure was initiated when the level of sensory block was deemed adequate.

Maternal hypotension was defined as a decrease in systolic arterial pressure to 80% of baseline or a systolic pressure of less than 100 mmHg. Hypotension was treated by intravenous ephedrine infusion: ephedrine 50 mg in 100 ml of 0.9% NaCl was given until normotension was restored; typically 40 mg of ephedrine was needed. After the operation, bleeding was estimated and volume depletion replaced by Ringer infusion or erythrocytes, if considered necessary.

The whole-body impedance cardiography derived SI, HR, CI and SVRI were analysed as average values for 30 seconds at the following time points: baseline, after preloading, at the point of lowest measured blood pressure after spinal blockade, immediately after delivery of newborn and placenta, at skin closure, after the anaesthesia had disappeared and on the second to fifth postpartum day.

Statistical analysis was performed using SPSS for Windows (version 11.0; SPSS Inc., Chicago, IL, USA) on a standard PC. Outcome parameters were analysed using repeated measurements analysis of variance (RANOVA) to detect significant changes in the entire group and coherence between the groups over time. Differences between the groups at each time point were examined post hoc using independent samples t test. A paired samples t test was used to detect changes within the groups. Differences in haemodynamic parameters between successive measurement points were expressed as percent changes to describe reactions to interventions. Differences were considered statistically significant if P≤ 0.05. All data are presented as mean and SD, SE of the mean or range, when appropriate.


The main results concerning SI, HR, CI, SVRI and MAP at baseline, during caesarean section, after the disappearance of anaesthesia and in the early postpartum period are shown in Figure 1.

Figure 1.

Trends of haemodynamic variables during caesarean section in healthy (▪) and pre-eclamptic (•) parturients. Values are expressed as mean and SE of the mean. 1 = baseline in left lateral position, 2 = after preloading with 6% hydroxyethyl starch, 3 = after spinal blockade (at the point of the lowest measured arterial blood pressure), 4 = before surgery, 5 = immediately after delivery of the newborn and the placenta, 6 = skin closure and 7 = disappearance of anaesthesia.

The women in the control group and those with pre-eclampsia differed significantly from each other in baseline measurements. The mean MAP and SVRI were significantly higher and, in contrast, SI and CI significantly lower in women with pre-eclampsia than in normal pregnancies. Based on RANOVA, SVRI (P= 0.002) and MAP (P= 0.012) remained at a significantly higher level, and SI (P= 0.069) and CI (P= 0.063) showed a tendency to remain lower in pre-eclamptic parturients compared with healthy parturients during the whole surgical procedure. The coherence between the groups during caesarean section was significantly different in respect of SI (P= 0.01) and SVRI (P= 0.002).

The haemodynamic reactions to fluid preloading, spinal anaesthesia, delivery and disappearance of anaesthesia, expressed as percent changes of haemodynamic parameters between the successive points of analysis are compared between the groups as shown in Table 2.

Table 2.  Percentual changes in haemodynamic parameters between successive periods of analysis
 Healthy parturientsPre-eclamptic parturientsP
  1. NS, not significant.

  2. Values are expressed as mean percent and SD. P values refer to the difference between the groups and values less than 0.1 are given.

Fluid preloading 
SI (ml/m2)−3.017.3+10.812.30.056
HR (beats/minute)+21.212.3+11.713.2NS
CI (l/minute/m2)+18.536.3+23.418.4NS
SVRI (dyne × sec/cm5× m2)−1.926.9−11.121.4NS
MAP (mmHg)+9.99.9+6.211.4NS
Spinal blockade 
SI (ml/m2)+2.75.2+0.77.9NS
HR (beats/minute)−2.918.5−2.610.8NS
CI (l/minute/m2)+3.825.7+4.319.9NS
SVRI (dyne × sec/cm5× m2)−19.340.0−8.032.9NS
MAP (mmHg)−21.422.9−9.27.8NS
Delivery of the fetus and the placenta 
SI (ml/m2)+16.114.1−
HR (beats/minute)+13.925.7+19.931.4NS
CI (l/minute/m2)+34.236.1+21.638.2NS
SVRI (dyne × sec/cm5× m2)−5.738.9−16.527.8NS
MAP (mmHg)+15.422.0−7.416.80.03
Disappearance of anaesthesia 
SI (ml/m2)−2.513.1−
HR (beats/minute)−11.712.7−18.68.2NS
CI (l/minute/m2)−14.714.4−33.317.30.03
SVRI (dyne × sec/cm5× m2)+49.052.0+83.864.1NS
MAP (mmHg)+20.424.3+14.126.1NS

Preload administration was associated with an elevation in SI, HR and CI in women with pre-eclampsia, whereas in normal pregnancies, only HR increased. The difference in SI reaction after preload infusion almost reached statistical significance (P= 0.056), while changes in other parameters were similar in the respective groups.

Spinal blockade caused similar changes in both groups: SVRI and MAP were reduced, while CI and SI were preserved stable. Eighty percent of women in the normal pregnancies and 30% of women with pre-eclampsia became hypotensive. In both groups, ephedrine increased MAP (14% in healthy parturients and 9% in pre-eclamptic parturients) and SVRI (14 and 16%, respectively). Furthermore, haemodynamic changes after delivery were alike between parturients treated and not treated with ephedrine in both study groups.

Immediately after delivery, the mean CI was 22% higher compared with the value after spinal blockade (52% higher compared with the baseline values) in women with pre-eclampsia and 34% higher (47% higher compared with baseline) in women in the control group. However, in pre-eclamptic pregnancies, no increase in SI occurred and the rise in CI was due entirely to an increase in HR, while in normal pregnancies, a significant increase in both SI and HR was observed. The difference in the percent change in SI between the groups was significant (P= 0.02). Among pre-eclamptic parturients, SI was increased in only one woman. This woman had severe pre-eclampsia complicated by HELLP (haemolysis, elevated liver enzymes and low platelet count) syndrome. Her SI increased to 12% at the moment of delivery, however, remaining lower than the mean SI in the women in control group.

The simultaneous decrease in SVRI was 17% compared with the value after spinal blockade, and the mean MAP decreased, reaching the lowest value during the surgical procedure in pre-eclamptic parturients. In the control group, SVRI decreased 6%, while the mean MAP remained stable (P= 0.03 for the percent change difference in MAP between the groups). Bleeding was moderate (550 ml or less) and similar in both groups.

With reversal of anaesthesia, the haemodynamic values were restored to baseline level in normal parturients. In women with pre-eclampsia, SVRI and MAP returned to baseline, but profound decreases in CI, SI and HR were recorded. During this period, the values reached the nadir of the whole study period. The changes in SI and CI differed between the groups (P= 0.04 and P= 0.03, respectively).

During the postpartum days, there were no significant differences in SI, HR or CI between the groups, but SVRI and MAP still remained significantly higher in women with pre-eclampsia than in normal parturients.

Three of four pre-eclamptic parturients receiving labetalol medication had severe disease. There were no differences in age, gestational age or BMI between women with and without antihypertensive medication. Furthermore, the values of the haemodynamic parameters at baseline and during the surgical procedure between the medicated and the nonmedicated parturients were comparable. In the baseline measurements, only MAP was higher in parturients with severe pre-eclampsia compared with the rest of the pre-eclamptic parturients. Otherwise, the haemodynamic parameters were similar during caesarean section.


The haemodynamics of pre-eclampsia have remained a controversial issue. Cardiac output has been found to be high with high systemic vascular resistance,16 high with low to normal vascular resistance17 and low with high systemic vascular resistance.7,8 Besides the varying baseline haemodynamics of pre-eclampsia, they may be further modified by treatment7,8 or pre-existent disease such as diabetes and renal or cardiovascular disorders. The progress of pre-eclampsia also influences the haemodynamics.20 In our study, there were profound differences between the study groups in the baseline measurements, suggesting a state of low cardiac output with high vascular resistance in pre-eclampsia. CIs were uniformly low, the range in CI in women with pre-eclampsia was narrow and the highest CI did not reach the mean value of CI in normal pregnancies. As the indication for caesarean delivery was in all cases worsening pre-eclampsia, this might suggest that at least, the end stage of pre-eclampsia would be associated with low CI.

In our data, the low CI seen in pre-eclamptic parturients was due to low SI. Women with severe pre-eclampsia had massive proteinuria (mean 8.1 g/day), which presumably caused depletion in intravascular volume, reflected in low SI. Besides assumed hypovolemia, the low SI could be further explained by possible depressed left ventricular performance attributable to the high blood pressure in pre-eclampsia. Borghi et al.21 have shown adaptive changes as a reaction to high afterload in the left ventricular structure and function in pre-eclampsia—higher left ventricular mass, higher left ventricular end-systolic and end-diastolic volumes and lower ejection fraction compared with normal pregnancies.

In pre-eclamptic parturients, preload infusion augmented SI and CI to the level of healthy parturients, while in uncomplicated pregnancies, only HR increased. It might be assumed that in normal pregnancies, intravascular volume and the preload of the heart were already sufficient and only higher HR was required to circulate the extra volume infused. In contrast, the more pronounced rise in SI in pre-eclamptic parturients might reflect an existing hypovolemia and inadequate preload, which was corrected by fluid administration. This finding was similar to that previously reported. Circulatory volume has been demonstrated to increase during crystalloid fluid preloading measured indirectly by an increase in central venous pressure22,23 and an increase in cardiac output estimated by Doppler ultrasound24 in women with pre-eclampsia.

After preload infusion, SVRI showed a tendency to decrease in pre-eclamptic parturients. Therefore, despite the vasoconstriction in pre-eclampsia, the vasculature was able to dilatate as a reaction to volume load. MAP remained stable, suggesting that the reduction in SVRI was sufficient to compensate the increase in CI, preventing blood pressure from rising.

At the moment of delivery, a significant increase in SI and CI in normal pregnancies conveys the view of momentarily increased preload of the heart, as the blood from the contracting uterus and pooling in the lower extremities is expelled into the vasculature.13 The unchanged SI observed in women with pre-eclampsia during delivery was an unexpected reaction significantly different from the women in the control group. It may indicate poorer preload of the heart compared with normal parturients. By reason of the hypovolemia and poorly perfused placenta during pregnancy, there is less pooling blood volume in the uterus and lower extremities and the volume load caused by delivery might be less than in uncomplicated pregnancies. Heavy bleeding during the operation might also reduce the volume load, but there was no difference in the amount of bleeding between the groups. A third explanation for the lacking increase in SI in pre-eclamptic parturients could be restricted cardiac adaptation to sudden volume load. Adaptive changes in the heart may lead to poorer left ventricular performance in pre-eclampsia.21,25 A state of diastolic dysfunction could explain the unchanged SI measurements after delivery in our women with pre-eclampsia. Presumably, the heart might be able to compensate sudden volume load only by increasing HR and not by adjusting SI, despite decreased MAP after delivery.

At the moment of delivery, there was also a marked decrease in SVRI in both groups, but in pre-eclamptic parturients, the simultaneous increase in CI was not enough to compensate the reduction in SVRI and, consequently, MAP decreased, while in healthy parturients, it remained stable. This sudden opening of the peripheral circulation concomitant with inadequate cardiac output may lead to underfilling of the vasculature and compromise the oxygenation of tissues. However, the low blood pressure was later corrected by increasing SVRI.

In our study, although there was no clinical case of pulmonary oedema, the observed haemodynamic pattern of failed increase in SI after delivery accompanied by volume load, high HR and lowering CI underlines the risk of relative heart failure and pulmonary oedema after delivery in pre-eclampsia. Low SI and CI, and so the risk of pulmonary oedema, persisted and even worsened during the study period, until the anaesthesia had disappeared. Our haemodynamic findings are in keeping with the clinical fact that the risk of postpartum pulmonary oedema is high during the first postpartum day in pre-eclampsia.

During the later postpartum period, the CI as well as SI in the women with pre-eclampsia had been re-established at the same level as in normal pregnancies, indicating rapid recovery of the heart, but vasospastic features of the disease persisted during the second to fifth postpartum day. However, the mean MAP and SVRI in pre-eclamptic parturients were below the baseline during the postpartum period. The lowered afterload might be one factor allowing the heart to recover from the adaptive state of pre-eclampsia and it could also induce CI to increase. Withdrawal of the excessive fluid compartment from the extravascular space during early puerperium26 may also enhance SI and CI in pre-eclamptic parturients.

In conclusion, we found pre-eclampsia to be a state of low cardiac output and high systemic vascular resistance. Volume preload administration attenuated the vasospastic and hypovolemic features of the condition, re-establishing CI at the level of normal parturients and reducing SVRI. At the moment of delivery, pre-eclamptic parturients were not able to increase SI, thus differing from healthy parturients. The unchanged and even diminishing SI could be due to lower volume preload after delivery or due to diastolic dysfunction of the left ventricle in pre-eclampsia to adapt to sudden volume load at the moment of delivery, exposing the parturient to the risk of pulmonary oedema. During the early puerperium, CI in women with pre-eclampsia recovered to the level of healthy parturients, but vasospastic features still persisted.


We are grateful to Pirjo Järventausta, RN, and Satu Ruusuvuori, RN, for their valuable technical assistance. This study was supported by the Medical Research Fund of Tampere University Hospital.