Lung and cardiac ultrasound for hemodynamic monitoring of patients with severe pre-eclampsia

Authors


ABSTRACT

Objective

To evaluate lung and cardiac ultrasound for the assessment of fluid tolerance and fluid responsiveness before and after delivery in pregnant women with severe pre-eclampsia (PE).

Methods

This was a prospective observational study of singleton pregnant women with severe PE and healthy term controls. Lung ultrasound echo comet score (ECS), which denotes the amount of extravascular lung water, was obtained using the 28-rib interspaces technique. The echocardiographic E/e′ ratio, measured by pulsed-wave and tissue Doppler, was used as a marker of diastolic left ventricular function. Fluid responsiveness was assessed by measuring changes in stroke volume (SV) with passive leg raising (PLR). SV was calculated from the left ventricular flow velocity-time integral measured by pulsed-wave Doppler at baseline and after PLR. Change in SV ≥ 12% was considered to indicate fluid responsiveness. Measurements obtained 1 day before delivery and 1 and 4 days after delivery were compared in the two groups (PE vs controls).

Results

We included 21 women with severe PE and 12 healthy controls. ECS and E/e′ ratio were higher in women with PE than in controls, both before delivery (P = 0.002 and P = 0.02) and 1 day postdelivery (P = 0.02 and P = 0.03); however there was no difference at 4 days postdelivery (P = 0.63 and P = 0.90). The change in SV with PLR before (P = 0.26) and after (P = 0.71) delivery did not differ between groups. An increase in SV ≥ 12% was observed in three (14%) women with PE and four (33%) controls before delivery and in four (19%) women with PE and two (17%) controls 1 day after delivery.

Conclusions

Severe PE is associated with an increase in extravascular lung water, which could in part be caused by disturbed diastolic left ventricular function. Excess lung water can be identified by lung ultrasound assessment in women with severe PE before the appearance of clinical signs. Only a small proportion of these women are fluid responsive. Copyright © 2016 ISUOG. Published by John Wiley & Sons Ltd.

INTRODUCTION

Accurate assessment of maternal hemodynamics is fundamental for appropriate fluid management in patients with severe pre-eclampsia (PE). Insufficient intravascular volume results in decreased oxygen delivery to tissues and exacerbates organ dysfunction[1, 2]. On the other hand, fluid excess can lead to fluid extravasation and pulmonary edema[1, 3]. The risk of fluid over-resuscitation is especially high in women with PE. This condition continues to be the leading cause of pulmonary edema in pregnancy, and maternal deaths associated with poor fluid management in PE cases have been reported[4, 5].

Outside pregnancy, transthoracic cardiac ultrasound (echocardiography) and lung ultrasound have become important diagnostic and monitoring tools in critically ill patients[6-9]. Echocardiography allows a rapid and non-invasive assessment of myocardial contractility and preload, and lung ultrasound can be used to determine the amount of extravascular lung water (EVLW)[10-12]. In pregnancy, there is evidence of a good correlation between non-invasive hemodynamic monitoring by echocardiography and invasive monitoring using a pulmonary artery catheter[13, 14]. Previous studies have shown that invasive hemodynamic monitoring could facilitate fluid management in patients with PE[15, 16]. However, recent studies have examined the utility of echocardiography in combination with lung ultrasound for guiding fluid therapy in patients with severe PE[17-19].

The objective of this study was to evaluate the ability of simple echocardiographic parameters, in combination with lung ultrasound, to assess fluid responsiveness and fluid tolerance in patients with severe PE.

METHODS

This prospective, observational study was performed at a single tertiary perinatal center between April 2015 and August 2016. All women included in the study provided written informed consent for participation, and the National Medical Ethics Committee approved the study.

Study participants

Consecutive patients with a singleton pregnancy complicated by severe PE were included in the study at hospital admission. Severe PE was defined according to the American College of Obstetricians and Gynecologists Task Force on Hypertension in Pregnancy recommendations: new-onset cerebral or visual disturbances; pulmonary edema; thrombocytopenia (platelet count < 100 000/mL); elevated liver enzymes (transaminases) to twice the normal upper limit; severe persistent pain in the right upper or middle upper abdomen that does not respond to medication and is not explained by another condition, or both; renal insufficiency (serum creatinine > 97 µmol/L), or a doubling of serum creatinine concentration in the absence of other renal disease; systolic blood pressure ≥ 160 mmHg or diastolic blood pressure ≥ 110 mmHg on more than one occasion at least 4 h apart while the patient is on bed rest (unless antihypertensive therapy had been initiated before this time)[20]. As per our institution's standard protocol, all patients were managed in a high-dependency setting antepartum and for at least 24 h postdelivery. Blood pressure was monitored continuously. Fluid intake and urine output were assessed hourly and blood tests were repeated at least every 12 h to monitor kidney function, electrolytes, full blood count, transaminases and bilirubin. Magnesium sulfate was used for eclampsia prophylaxis as a 4-g intravenous loading dose, followed by 1 g/h infusion. Antihypertensive treatment with intravenous hydralazine or labetalol was used to maintain systolic blood pressure < 160 mmHg and diastolic blood pressure < 110 mmHg. Intravenous and oral fluid intake was minimized and neither fluids nor diuretics were routinely administered to treat oliguria. Women were delivered following maternal stabilization when the gestational age was ≥ 34 + 0 weeks. For women at < 34 + 0 weeks' gestation, administration of a single course of antenatal corticosteroids was attempted and pregnancy was managed expectantly for up to 48 h in the absence of maternal and/or fetal indications for immediate delivery. Vaginal delivery was considered unless a Cesarean delivery was required for the usual obstetric indications.

Controls were healthy women with a singleton pregnancy at term (≥37 + 0 weeks' gestation) and an estimated fetal weight appropriate for gestational age. They were included in the study at hospital admission for either a planned Cesarean section or induction of labor.

Lung ultrasound and echocardiography

Ultrasound assessment was performed using a Vivid S6 scanner (GE Vingmed Ultrasound, Horten, Norway) with a 3Sc-RC cardiac transducer for both echocardiography and lung ultrasonography. All ultrasound examinations and offline analyses were performed by a single investigator (J.A.). All offline measurements were done using dedicated cardiac software (EchoPac version 201; GE Vingmed Ultrasound).

Inter- and intraobserver reproducibility for lung ultrasound parameters and echocardiographic measurements was assessed by offline analyses in 10 randomly selected subjects (five PE patients and five controls) by two independent operators (an obstetrician (M.L.) and a cardiologist (J.A.)).

Lung ultrasound was performed according to a systematic protocol in supine patients[9, 21]. The echo comet score (ECS) was obtained by the 28-rib interspaces technique, which divides the chest wall into 12 areas on the left (from the second to the fourth intercostal space) and 16 areas on the right (from the second to the fifth intercostal space) anterior and lateral hemithorax[17]. An increased amount of EVLW can be diagnosed by multiple B-lines or ‘comet tails’[22]. B-lines are defined as discrete laser-like vertical hyperechoic reverberation artifacts that arise from the pleural line and extend to the bottom of the screen without fading, and move synchronously with lung sliding[22]. They represent a reverberation artifact through edematous interlobular septa or alveoli[21, 22]. The sum of the B-lines found on each of the 28 chest-wall areas yields the ECS, denoting the amount of EVLW. Lung ultrasound examination has been demonstrated to be an easy-to-learn technique with a steep learning curve (suggested < 10 examinations)[23-25]. No additional training in lung ultrasonography was necessary in the present study for both investigators performing ECS measurements (J.A. and M.L.).

All subjects were studied by standard two-dimensional and Doppler transthoracic echocardiography according to the joint recommendations of the American Society of Echocardiography and the European Association of Cardiovascular Imaging[26]. Patients were studied at rest in the left lateral decubitus position and data were acquired at end expiration from standard parasternal/apical/subcostal views. Maternal heart rate was measured during every echocardiogram, as significant differences in heart rate can further influence other echocardiographic parameters.

The E/e′ ratio (the E-wave represents the early diastolic mitral inflow velocity assessed by pulsed-wave Doppler and the e′-wave represents the early diastolic mitral annulus displacement velocity at the septal side assessed by tissue Doppler) was used as a marker of diastolic function. Standard cardiac settings were used to obtain optimal pulsed-wave and tissue Doppler curves, according to echocardiographic guidelines[26]. The E/e′ ratio is a well-described echocardiographic parameter for assessing diastolic left ventricular function, as the majority of studies have shown a close correlation between this ratio and left ventricular end-diastolic pressure, with few studies reporting conflicting results[27-33].

Fluid responsiveness was assessed 1 day before delivery (within 24 h before delivery) and 1 day postdelivery (within 24 h after delivery) by measuring changes in stroke volume (ΔSV) while performing passive leg raising (PLR) to a 45° angle for 2 min. We calculated SV as the product of the area of the left ventricular outflow tract (cm2) and left ventricular flow velocity-time integral (VTI) (cm). Left ventricular outflow tract area was calculated using the formula: (left ventricular outflow tract diameter)2 × π/4, measured on the parasternal long-axis view[8, 26]. Left ventricular outflow tract diameter was only measured once, as it remains constant in a given patient. VTI was measured with pulsed-wave Doppler on the apical long-axis view. A second measurement of VTI was obtained approximately 30 s after PLR and was compared with the baseline value. Fluid responsiveness was defined as ΔSV ≥ 12%[8, 18].

Statistical analysis

For continuous variables, data were expressed as median and interquartile range. Categorical data were summarized as frequencies and percentages. For comparisons between the two study groups (women with severe PE vs controls), the Mann–Whitney U-test was used for continuous variables and the chi-square test or Fisher's exact test for categorical variables, as appropriate. Changes in parameters over time (before delivery vs 1 day postdelivery vs 4 days postdelivery) within each group were analyzed by repeated-measures ANOVA. Measurements before delivery vs 1 day postdelivery, and 1 day postdelivery vs 4 days postdelivery within each group were compared by the Wilcoxon signed-rank test. Intraclass correlation coefficient (ICC) was used to test inter- and intraobserver reproducibility of the analyzed parameters. Statistical analysis was done using IBM SPSS Statistics for Windows Version 21.0 (IBM Corp., Armonk, NY, USA), and for all tests two-tailed P ≤ 0.05 was considered to be statistically significant.

RESULTS

Study population characteristics

Twenty-one women with severe PE and 12 healthy controls were included in the study; characteristics of the participants are shown in Table 1. Maternal ethnicity was Caucasian except for one woman in the PE group, who was Asian. No woman admitted to tobacco smoking or alcohol or drug abuse during pregnancy. No participant had chronic hypertension, or pre-existing or gestational diabetes mellitus. There were more nulliparous women in the severe PE group than in the control group (P = 0.03). None of the parous women had had PE or delivered preterm in their previous pregnancy. As expected from the study design, gestational age at admission in the control group was higher (P < 0.001) and there were more small-for-gestational-age neonates in the severe PE group (P = 0.03). The high rate of Cesarean section among controls was a reflection of the criteria used for inclusion in this group.

Table 1. Clinical characteristics of 21 women with singleton pregnancy complicated by severe pre-eclampsia (PE) and 12 healthy controls included in the study
CharacteristicSevere PE (n = 21)Controls (n = 12)P
  1. Data are given as median (range) or n (%). Comparison of groups by Mann–Whitney U-test or Fisher's exact test.
  2. BMI, prepregnancy body mass index; GA, gestational age at admission; MA, maternal age; SGA, small-for-gestational age (defined as a neonate with birth weight < 5th centile for gestational age).
MA (years)28 (21–44)33 (24–42)0.54 
BMI (kg/m2)23 (19–32)33 (17–29)0.93 
Nulliparous16 (76)4 (33)0.03 
GA (weeks)33 + 0 (24 + 1 to 39 + 4)39 + 3 (37 + 0 to 42 + 1)< 0.001
Cesarean delivery16 (76)11 (92)0.37 
SGA7 (33) 0.03 

Severe features of the inclusion criteria were present in the PE group: hypertension in all 21 (100%) cases; neurological symptoms in seven (33%); elevated liver enzymes in six (29%); thrombocytopenia in two (10%); and right upper abdominal pain in two (10%).

In the control group, planned Cesarean section was performed in eight women (for previous Cesarean delivery in six (75%) cases and for breech presentation in two (25%) cases). Labor was induced in four women in the control group (post-term pregnancy was the indication for induction in all cases). After induction, one (25%) woman delivered vaginally and three (75%) delivered by emergency Cesarean section.

No woman in the control group received intravenous medication or intravenous fluid at the time of the sonographic examinations. In the PE group, all 21 patients were treated with magnesium sulfate as per our institution's protocol before delivery as well as at 24 h postdelivery. Magnesium sulfate was, therefore, administered as a continuous intravenous infusion of 1 g/h (as a 50-mL/h infusion) during two sonographic examinations: before delivery and 1 day postdelivery. In 12 (57%) women with PE, magnesium sulfate was the only intravenous fluid intake. In nine (43%) patients with PE, an additional 30 mL/hour infusion of crystalloids was administered before delivery and during the first ultrasound examination, while always limiting administration of intravenous fluid to 80 mL/h. Intravenous hydralazine was necessary for blood-pressure control in five (24%) women with PE before delivery. On day 4 postpartum, no woman with PE was receiving any intravenous fluid or medication.

Lung ultrasound and echocardiography measures

ECS was significantly higher in women with severe PE than in controls, both before delivery (P = 0.002) and 1 day postdelivery (P = 0.02); however this difference was not present 4 days postdelivery (P = 0.63) (Figure 1). Significant changes in ECS over time were observed in the severe PE group (ANOVA, P = 0.01); ECS values before delivery vs 1 day postdelivery did not differ significantly (P = 0.28), but there was a significant decrease in ECS between days 1 and 4 postdelivery (P < 0.001). In the control group, no significant change in ECS values before and after delivery was observed (ANOVA, P = 0.36). Figure 2 presents lung ultrasound images before and after delivery in a study patient with severe PE, showing the lung ultrasound pattern corresponding to EVLW. A decrease in EVLW following delivery can be clearly seen.

Figure 1.

Box-and-whisker plot of echo comet score (ECS), denoting amount of extravascular lung water, in 21 pregnant women with severe pre-eclampsia (PE) and in 12 healthy controls. ECS was higher in PE patients than in controls, before delivery (image) (P = 0.002) and 1 day postdelivery (image) (P = 0.02), but not 4 days postdelivery (image) (P = 0.63). *P < 0.05. Boxes with internal lines represent median and interquartile range and whiskers are range.

Figure 2.

Lung ultrasound images in woman with severe pre-eclampsia before delivery (a) and 1 day (b) and 4 days (c) after delivery, showing B-lines (image). Note decrease in number of B-lines, illustrating a progressive decrease in extravascular lung water, following delivery.

Heart rate was comparable between both groups at all echocardiographic assessments (before delivery, P = 0.95; 1 day postdelivery, P = 0.57; 4 days postdelivery, P = 0.37). The E/e′ ratio was significantly higher in women with severe PE than in controls, both before delivery (P = 0.02) and 1 day postdelivery (P = 0.03); however this difference was not present 4 days postdelivery (P = 0.90) (Figure 3). In the severe PE group, there was a significant change in E/e′ ratio following delivery (ANOVA, P = 0.03); the E/e′ ratio before delivery vs 1 day postdelivery did not differ significantly (P = 0.15), however there was a significant decrease in E/e′ ratio between days 1 and 4 postdelivery (P = 0.01). No such change was observed in the control group (ANOVA, P = 0.90).

Figure 3.

Box-and-whisker plot of E/e′ ratio, a marker of diastolic function, in 21 women with severe pre-eclampsia (PE) and in 12 healthy controls. E/e′ ratio was higher in PE patients than in controls, before delivery (image) (P = 0.02) and 1 day postdelivery (image) (P = 0.03), but not 4 days postdelivery (image) (P = 0.90). *P < 0.05. Boxes with internal lines represent median and interquartile range and whiskers are range.

Change in SV with PLR did not differ between groups before delivery (mean ± SD, 3 ± 8% vs 7 ± 9%; P = 0.26) or 1 day postdelivery (mean ± SD, 6 ± 9% vs 7 ± 9%; P = 0.71). We observed a ΔSV of ≥ 12% (indicating fluid responsiveness) in three (14%) women with severe PE and four (33%) controls before delivery and in four (19%) women with severe PE and two (17%) controls 1 day postpartum. No significant ΔSV with PLR before vs after delivery was observed in either group (P = 0.16 for the severe PE group and P = 0.58 for the control group, respectively (ANOVA)).

Reproducibility of lung ultrasound and echocardiographic measures

Intra- and interobserver agreement were excellent for ESC score (ICC, 0.989 (95% CI, 0.955–0.997) and 0.986 (95% CI, 0.943–0.996), respectively), and were strong for both E/e′ ratio (ICC, 0.839 (95% CI, 0.630–0.972) and 0.697 (95% CI, 0.165–0.915), respectively) and ΔSV (ICC, 0.810 (95% CI, 0.405–0.949) and 0.754 (95% CI, 0.279–0.933), respectively).

DISCUSSION

In this study we have shown that severe PE is associated with an increase in EVLW and disturbed diastolic left ventricular function compared with healthy pregnancies. We found only a small proportion of women with severe PE in our study cohort to be fluid responsive.

Severe PE is associated with an increase in EVLW before delivery and immediately postpartum. The amount of EVLW then decreases rapidly in the first days following delivery, such that at day 4 postpartum we observed no difference in EVLW between women with PE and healthy controls. These findings are in accordance with the well-established association of PE with a higher incidence of pulmonary edema[4]. Furthermore, we found that lung ultrasound can identify increased levels of EVLW in women with PE before clinical signs of pulmonary edema appear (none of the patients included had clinical signs of pulmonary edema despite high ECS values). Therefore, lung ultrasound could help reduce complications associated with fluid over-resuscitation and identify those that need diuretic therapy among patients with severe PE. Lung ultrasound is relatively quick to learn, with a steep learning curve[23-25]. Following appropriate training, anesthesiologists and obstetricians should be able to utilize effectively lung ultrasound as a point-of-care investigation. Abdominal probes, available in almost all maternity wards, can also be used for this purpose. Our study suggests excellent interobserver reproducibility of lung ultrasound measurements between cardiologists, who use lung ultrasound regularly to assess pulmonary congestion, and obstetricians who are familiar with the technique.

Echocardiography and PLR could be used to identify women with PE who will respond to fluids by increasing their cardiac output. Such patients will benefit from further fluid administration when their cardiac output is inadequate to meet their metabolic demands. Based on our results, only a small proportion of women with severe PE are fluid responsive (14% before delivery and 19% in the first day following delivery). Our findings are in accordance with the impedance cardiography data of Vartun et al.[34, 35], who found 4–15% of pregnant women to be fluid responsive. Brun et al.[18] found a higher (52%) proportion of fluid responsiveness in severe PE complicated by oliguria, which highlights the importance of basing fluid therapy in this disease on an accurate assessment of the maternal hemodynamic profile.

In our study, women with PE had a higher E/e′ ratio, both before delivery and on the first day postdelivery, suggesting higher end-diastolic left ventricular pressure in women with severe PE than in healthy controls. It should be noted that the E/e′ ratios in women with PE were still within the normal or intermediate reference range for a normal population (septal E/e′ ratio ≤ 15)[29]. Nevertheless, our findings suggest that higher EVLW in PE not only is caused by increased vascular permeability due to endothelial dysfunction, but could also be due to an increase in pulmonary intravascular hydrostatic pressure as a consequence of volume overload and/or disturbed left ventricular diastolic function. This observation is further corroborated by the echocardiographic findings of Melchiorre et al.[36, 37], who observed parameters of diastolic dysfunction more frequently in women with PE than in control pregnancies, and Zieleskiewicz et al.[17], who found an association between diastolic left ventricular parameters and EVLW in severe PE. We have analyzed and demonstrated dynamic changes in EVLW and left ventricular diastolic function in women with PE, before and after delivery. Of note, we found no worsening of left ventricular diastolic function or further increase in EVLW after blood-flow redistribution associated with delivery.

This study has limitations inherent to the small study population. We focused only on patients with the most severe forms of PE, as these are critically ill pregnant women who need an accurate assessment of their hemodynamics for good clinical management. Unfortunately, there can be no optimal controls in these types of study owing to the specific characteristics of PE patients. Gestational age is known to be an important determinant of hemodynamic status in pregnancy. However, to match controls by gestational age, we would have to have included either patients with spontaneous preterm labor or those with medically indicated preterm delivery. In the first case, maternal hemodynamics would be influenced by labor itself and its possible infectious etiology. In the second case, medically indicated preterm delivery is most often performed, if not for hypertensive disorders of pregnancy, for fetal intrauterine growth restriction[38]. This is most often due to placental malfunction, which again makes such patients non-optimal controls[39]. Outside of pregnancy, different echocardiographic parameters are used to assess diastolic function. We used only the E/e′ ratio, which was found to be a simple echocardiographic parameter for evaluating diastolic function with high feasibility. Further studies using additional echocardiographic indices are needed to draw firmer conclusions on left ventricular end-diastolic pressure and diastolic function in PE. Moreover, assessing ECS with the 28-rib interspaces technique is time consuming. Further studies are needed to evaluate simplified techniques, such as the eight anterior regions method, which could be more suitable for everyday clinical practice.

In conclusion, simple and easy-to-use echocardiographic parameters, in combination with lung ultrasound, can be used for assessing fluid responsiveness and fluid tolerance in women with severe PE. Lung ultrasound seems a promising method for identifying women with severe PE in whom positive fluid balance should be avoided, as it may lead to pulmonary edema, whereas echocardiography could help clinicians to identify patients with severe PE who are still fluid responsive and in whom further fluid administration could be beneficial.

Ancillary