Acute Cardiac Effects of Severe Pre-Eclampsia

Abstract

BACKGROUND

Pre-eclampsia with severe features (PEC) is a pregnancy-specific syndrome characterized by severe hypertension and end-organ dysfunction, and is associated with short-term adverse cardiovascular events, including heart failure, pulmonary edema, and stroke.

OBJECTIVES

The authors aimed to characterize the short-term echocardiographic, clinical, and laboratory changes in women with PEC, focusing on right ventricular (RV) systolic pressure (RVSP) and echocardiographic-derived diastolic, systolic, and speckle tracking parameters.

METHODS

In this prospective observational study, the authors recruited 63 women with PEC and 36 pregnant control patients.

RESULTS

The PEC cohort had higher RVSP (31.0 ± 7.9 mm Hg vs. 22.5 ± 6.1 mm Hg; p < 0.001) and decreased global RV longitudinal systolic strain (RVLSS) (−19.6 ± 3.2% vs. −23.8 ± 2.9% [p < 0.0001]) when compared with the control cohort. For left-sided cardiac parameters, there were differences (p < 0.001) in mitral septal e′ velocity (9.6 ± 2.4 cm/s vs. 11.6 ± 1.9 cm/s), septal E/e′ ratio (10.8 ± 2.8 vs. 7.4 ± 1.6), left atrial area size (20.1 ± 3.8 cm² vs. 17.3 ± 2.9 cm²), and posterior and septal wall thickness (median [interquartile range]: 1.0 cm [0.9 to 1.1 cm] vs. 0.8 cm [0.7 to 0.9 cm], and 1.0 cm [0.8 to 1.2 cm] vs. 0.8 cm [0.7 to 0.9 cm]). Eight women (12.7%) with PEC had grade II diastolic dysfunction, and 6 women (9.5%) had peripartum pulmonary edema.

CONCLUSIONS

Women with PEC have higher RVSP, higher rates of abnormal diastolic function, decreased global RVLSS, increased left-sided chamber remodeling, and higher rates of peripartum pulmonary edema, when compared with healthy pregnant women.

Pre-eclampsia is a clinical syndrome characterized by hypertension with proteinuria and/or end-organ damage. Although the inciting pathology is not definitively known, pre-eclampsia occurs in 2% to 8% of pregnancies and is associated with early abnormal placentation, endothelial dysfunction, and systemic inflammation in the second half of pregnancy (1–5). Pre-eclampsia can lead to subsequent pulmonary edema (6,7), strokes (8), acute respiratory distress syndrome, placental abruption, HELLP (hemolysis, elevated liver enzymes, and low platelets) syndrome, disseminated intravascular coagulation, acute renal failure, liver rupture, seizures, or death (2,7,9). Maternal perinatal outcomes are more likely to be poor if pre-eclampsia occurs before the 32nd gestational week or is associated with pre-existing comorbidities (2). In particular, black women have higher rates of short- and long-term morbidity and mortality (2,10,11).

Pre-eclampsia is associated with deleterious changes in cardiovascular physiology, including increased systemic vascular resistance (12,13), capillary leak-induced interstitial edema (14), and pulmonary edema (4,15,16). Some studies have reported left ventricular (LV) systolic dysfunction, whereas others have found no changes (4,12,15,17,18). Right ventricular (RV) systolic function has also been reported to decline, although RV ejection fractions (EFs) remain within the normal range (19). By contrast, studies consistently report LV diastolic dysfunction, which can persist well into the postpartum state (20). Diastolic dysfunction likely results from abnormal LV remodeling and hypertrophy (4,17,18,21), with similar processes occurring in the RV (19,21).

Many of the abnormalities in cardiac function associated with pre-eclampsia have been discovered by 2-dimensional (2D) and tissue Doppler echocardiography (21–24). More recently, speckle tracking echocardiography (STE) has also been used in studies of hypertensive disorders of pregnancy (21,25,26). STE is an especially valuable adjuvant echocardiographic modality because STE-derived parameters are less load-dependent than traditional echocardiographic markers and can reveal occult myocardial systolic dysfunction even when conventional echocardiographic measures are normal (27,28).

Although previous studies have focused on the cardiovascular and clinical characteristics associated with pre-eclampsia (18,20,22,23,29), few have focused on women with pre-eclampsia with severe features (PEC) or have correlated echocardiographic parameters with short-term cardiovascular events. Therefore, we prospectively evaluated cardiac function using transthoracic echocardiography in a diverse cohort of women with PEC compared with a cohort of women with normotensive pregnancies. Specifically, we evaluated whether women with PEC would have higher right ventricular systolic pressures (RVSPs) and abnormal RV or LV function. In addition, we assessed the relationship between abnormal echocardiographic features, B-type natriuretic peptide (BNP) levels, and maternal perinatal cardiac complications.

METHODS

STUDY DESIGN AND PARTICIPANTS.

This prospective observational study was approved by the Johns Hopkins University School of Medicine Institutional Review Board and adhered to the Guidelines for Good Clinical Practice. All participants gave written informed consent. Pregnant women with singleton pregnancies >23 weeks of gestational age were recruited from the Johns Hopkins Hospital, the Johns Hopkins Bayview Medical Center, and associated outpatient clinics. Women were excluded if they had multifetal gestations, known valvular or congenital heart disease, cardiomyopathy, pulmonary hypertension, prior cardiac surgery, pulmonary embolism, systemic lupus erythematosus, any connective tissue disease, antiphospholipid syndrome, or interstitial lung disease.

We enrolled patients with pre-eclampsia with severe features (3), which is characterized by the following criteria: systolic blood pressure ≥160 mm Hg or a diastolic blood pressure ≥110 mm Hg measured on 2 occasions at least 4 h apart, and proteinuria >300 mg of protein in a 24-h urine collection or a protein/creatinine ratio ≥0.3 on urinalysis. In the absence of proteinuria, patients were included if they had thrombocytopenia (platelet count <100,000/μl), impaired liver function (liver enzymes twice normal), progressive renal dysfunction (serum creatinine ≥1.1 mg/dl), pulmonary edema, or new-onset visual disturbances. Of note, for the purpose of this paper, we abbreviate pre-eclampsia with severe features as PEC.

Women with PEC were then classified into 2 groups based on the presence or absence of pre-existing chronic hypertension. If they had documented hypertension before the index pregnancy and experienced elevations in blood pressure to meet the criteria for PEC, they were classified as having pre-eclampsia with pre-existing chronic hypertension (PEC-cHTN), referred to as superimposed pre-eclampsia. If they had no pre-existing hypertension and developed pre-eclampsia during the index pregnancy, they were classified as having pre-eclampsia without chronic hypertension (PEC no-cHTN).

A control group of normotensive women with singleton pregnancies >23 weeks of gestation and no exclusion criteria were enrolled at outpatient antenatal care visits. They were matched by gestational age with PEC cohort patients.

The women with PEC underwent echocardiographic and laboratory measures within 24 h of admission and before active labor or delivery but after initial treatment with intravenous magnesium sulfate and oral or intravenous antihypertensive agents (i.e., labetalol, hydralazine, or nifedipine). The control cohort underwent transthoracic echocardiography and venipuncture in the outpatient setting.

TRANSTHORACIC ECHOCARDIOGRAPHY.

The transthoracic echocardiograms were performed by experienced cardiac sonographers using GE or Phillips ultrasound machines (GE Healthcare, Chicago, Illinois, or Philips Healthcare, Andover, Massachusetts) and interpreted by 2 Level III–trained cardiologists certified by the National Board of Echocardiography. In accordance with American Society of Echocardiography (ASE) guidelines (30), 2D methods were used to obtain chamber sizes, and the biplane method of disks was used to calculate LV EF. In addition, LV diastolic parameters were measured and used to estimate LV filling pressures and assign diastolic function grades based on the 2009 and 2016 ASE guidelines criteria (Online Appendix) (31,32). Tricuspid regurgitant velocity was used to estimate RVSP using the modified Bernoulli equation with the addition of estimated right atrial (RA) pressure based on inferior vena cava dimension and collapsibility with the sniff maneuver (30).

After acquisition of standard transthoracic echocardiographic images, apical views were subsequently analyzed using STE software version 3.1.0.3358 (Epsilon, Milwaukee, Wisconsin) (33), which tracks the position and movement of acoustic speckles within a myocardial regional of interest. From the 4-chamber apical view, the RV free wall borders were determined by tracing the RV endocardial border and adjusting the thickness of the myocardium as appropriate. The software divides the RV free wall into apical, mid, and basal segments (34). From the 4-chamber, 2-chamber, and 3-chamber apical views, the LV endocardium was traced and automatically segmented into basal, mid, and apical segments of the 2 opposite walls in each views (33,35,36). All automated border tracings were then reviewed and manually adjusted to ensure adequate tracking. Using the STE software, peak longitudinal systolic strain was then calculated by determining the average maximal extent of longitudinal movement for each LV or RV segment (34,36). By convention, strain is expressed as a negative percentage. Areas with decreased strain (i.e., due to hypokinesis) have less negative percentages, or lower absolute values, than expected for a region of interest.

Global RV longitudinal systolic strain (RVLSS) was calculated as the average of regional strain parameters from the basal, mid, and apical RV free wall segments (34). Global LV longitudinal systolic strain (LVLSS) was calculated as the average of regional strain parameters for the entire length of LV myocardium (basal, mid, and apical segments of 2 opposite walls in each view) (37). To determine the level of intra-observer agreement for RVLSS and LVLSS values, the interpreting cardiologist re-reviewed STE images and calculations from 10 randomly selected patients in a blinded manner. The intraclass correlation coefficient was then calculated and was 0.97 for both global RVLSS and LVLSS. For interobserver reliability, an additional cardiologist, blinded to the findings of the initial interpreting cardiologist, reviewed the STE images and calculations from those same patients. Using Spearman correlation, interobserver agreement was 0.95 for both global RVLSS and LVLSS.

SERUM BNP.

Blood samples were obtained within 24 h of echocardiographic imaging. Samples were collected in serum separation tubes, centrifuged at 4°C, and processed within 4 h of venipuncture. Serum was then aliquoted and stored at −80°C. Serum samples were later thawed and BNP measured in duplicate using the BNP Human ELISA kit (ABCAM, Cambridge, Massachusetts). The intra-assay coefficient of variation is <10%.

OUTCOME MEASURES.

The primary outcome measure was RVSP. Secondary outcome measures were echocardiographic-parameters: left atrial (LA) and RA size, LV and RV size, LV wall thickness, LV EF, RV fractional area of change, transannular planar systolic excursion (TAPSE), S′ velocity, mitral inflow E and A velocities, E/A ratio, septal e′, E/e′ ratio, RVLSS (basal, mid, apical, and global), and LVLSS (basal, mid, apical, and global). Clinical secondary outcomes included eclampsia, stroke, pulmonary edema, and mortality. Pulmonary edema was noted if diagnosed by the treating physician and confirmed by radiographic imaging. Serum BNP levels were the sole biomarker secondary outcome.

STATISTICAL ANALYSIS.

Group demographic, clinical, imaging, and outcome variables are shown in Tables 1 to 3 and Online Tables 1 to 5. Group differences in categorical variables were tested using chi-square tests, or Fisher exact tests when group numbers were small and the large number assumption for chi-square tests did not apply. Analysis of variance was used for continuous, normally distributed variables; mean SD were tabulated. Kruskall-Wallis nonparametric tests were used for variables that were not normally distributed; medians (25th to 75th percentiles) were tabulated. The overall number of significance tests performed was 51, thus, to correct for multiple comparisons, a p value of <0.00098 (≈p < 0.001), that is, 0.05/51, was used as the statistical significance threshold. At the time of study design, the variability of RVSP in this population was unknown. After accruing 47 participants, we derived an intermediate estimate of 7.84 mm Hg for the standard deviation of RVSP. Using a p < 0.00098 and a power of 80%, a power calculation determined that 16 control patients and 32 with pre-eclampsia would be needed to show a statistically significant 10 mm Hg difference in RVSP. More participants than needed were recruited because it was anticipated that there would be an inability to obtain RVSP and secondary echocardiographic parameters in some participants.

TABLE 1.

Participant Demographics and Clinical Data

	Control (n = 36)	Pre-Eclampsia (n = 63)	p Value
Age, yrs	30.7 ± 5.7	29.2 ± 6.2	0.238
Gestational age, weeks	31.8 ± 4.9	33.1 ± 3.6	0.127
Body mass index, kg/m2	29.7 (26.8–36.4)	33.1 (28.0–41.4)	0.042
Race			<0.001*
Asian	3 (8.3)	1 (1.6)
Black	7 (19.4)	40 (63.5)
Other	2 (5.6)	2 (3.2)
White	24 (66.7)	20 (31.7)
Ethnicity			0.336
Hispanic	3 (8.3)	6 (9.5)
Non-Hispanic	33 (91.7)	57 (90.5)
Diabetes			0.843
Nondiabetic	34 (94.4)	57 (90.5)
Gestational	1 (2.8)	2 (3.2)
Pre-gestational	1 (2.8)	4 (6.3)
Highest systolic BP, mm Hg	116 ± 3	185 ± 15	<0.001*
Highest diastolic BP, mm Hg	67 ± 2	103 ± 9	<0.001*

Values are mean ± SD, median (interquartile range), or n (%). The threshold of significance is p < 0.001.

^*Statistically significant difference between groups.

BP =blood pressure.

TABLE 3.

Echocardiographic Parameters and BNP in the Control Group and PEC Groups (PEC-cHTN vs. PEC no-cHTN)

	Control (n = 36)	PEC-cHTN (n =17)	PEC no-cHTN (n = 46)	p Value
Right ventricular parameters
RVSP, mm Hg	22.5 ± 6.1	28.6 ± 9.4	31.7 ± 7.3	<0.001*
RA area, cm2	14.6 ± 3.2	14.6 ± 2.8	15.0 ± 3.0	0.750
IVC, cm	1.4 ± 0.4	1.5 ± 0.3	1.5 ± 0.4	0.596
RV FAC, %	49.6 (46.0–54.0)	50.6 (45.0–58.0)	49.5 (45.0–58.0)	0.770
S′, cm/s	14.7 ± 2.3	16.1 ± 3.5	15.4 ± 3.1	0.299
TAPSE, cm	2.4 ± 0.4	2.6 ± 0.4	2.7 ± 0.4	0.060
Basal RV longitudinal strain, %	−24.1 ± 8.2	−20.1 ± 10.0	−19.3 ± 8	0.026
Mid RV longitudinal strain, %	−22.4 ± 6.4	−19.6 ± 4.5	−19.9 ± 6.9	0.07
Apical RV longitudinal strain, %	−25.2 ± 6.4	−17.3 ± 11.5	−20.4 ± 6.9	0.01
Global RV longitudinal strain, %	−23.8 ± 2.9	−19.1 ± 4.0	−19.8 ± 3.0	<0.001*
Left ventricular parameters
LA area, cm2	17.3 ± 2.9	20.5 ± 3.8	20.0 ± 3.9	<0.001*
LV septal wall thickness, cm	0.8 (0.7–0.9)	1.2 (1.0–1.3)	0.9 (0.8–1.1)	<0.001*
LV posterior wall thickness, cm	0.8 (0.7–0.9)	1.2 (1.0–1.3)	0.9 (0.9–1.1)	<0.001*
LVEF, %	61.2 ± 5.1	66.1 ± 4.6	63.3 ± 6.1	0.012
Mitral E, cm/s	84.3 ± 17.9	97.8 ± 18.9	98.8 ± 24.4	0.009
Mitral A, cm/s	57.3 ± 12.7	71.1 ± 27.2	76.1 ± 21.6	<0.001*
Mitral septal e′, cm/s	11.6 ± 1.9	8.9 ± 1.9	9.8 ± 2.5	<0.001*
Mitral E/e′ ratio	7.4 ± 1.6	11.4 ± 2.7	10.6 ± 2.9	<0.001*
Mitral E/A ratio	1.4 (1.2–1.8)	1.4 (1.0–1.7)	1.3 (1.1–1.6)	0.261
Mitral DT, ms	198.3 ± 38.0	195.7 ± 35.0	192.4 ± 53.0	0.917
Basal LV longitudinal strain, %	−16.8 ± 2.3	−13.7 ± 9.8	−16.0 ± 2.4	0.15
Mid LV longitudinal strain, %	−19.0 ± 2.5	−17.1 ± 2.5	−17.5 ± 2.2	0.018
Apical LV longitudinal strain, %	−24.8 ± 3.2	−23.2 ± 2.9	−24.4 ± 3.2	0.17
Global LV longitudinal strain, %	−20.1 ± 1.5	−18.8 ± 1.5	−19.2 ± 1.5	0.02
BNP, pg/ml	44.0 (20.0–81.0)	57.0 (20.0–106.0)	61.0 (26.0–110.0)	0.434

Values are mean ± SD or median (interquartile range). The threshold of significance is p < 0.001. There were no statistically significant differences between PEC-cHTN versus PEC no-cHTN groups (p values not shown).

^*Statistically significant difference between control versus PEC-cHTN and control versus PEC no-cHTN groups.

PEC = pre-eclampsia with severe features; PEC-cHTN = pre-eclampsia with pre-existing chronic hypertension; PEC no-cHTN = pre-eclampsia without pre-existing chronic hypertension; other abbreviations as in Table 2.

Because our preliminary demographic analyses revealed racial differences between groups, all significantly different outcome variables were also stratified by race. We specifically focused on differences between black and white women, and included all stratified variables in the Online Appendix.

RESULTS

STUDY POPULATION.

From August 1, 2014, to December 31, 2015, we recruited 102 pregnant women: 64 with pre-eclampsia and 38 without. One woman with PEC was withdrawn because she ultimately did not meet the criteria for PEC. Two in the control group withdrew before echocardiographic imaging. Thus, the final cohort included 63 patients with PEC and 36 control patients. Demographic and clinical characteristics are shown in Table 1. Mean gestational ages were comparable (33.1 ± 3.6 weeks [PEC] vs. 31.8 ± 4.9 weeks [control]). All other group characteristics were also similar except for race, with a higher percentage of black women in the PEC group. Among the patients with PEC, 17 had PEC-cHTN and 46 had PEC no-cHTN. There were no demographic or clinical differences between these subgroups (Online Table 1).

OUTCOME MEASURES.

Tables 2 and 3 compare echocardiographic values in our groups. The control group’s parameters were within the normal range for women in their age group, and none had abnormal strain or systolic or diastolic dysfunction.

TABLE 2.

Echocardiographic Parameters and BNP in the Control Group Versus Patients With Pre-Eclampsia

	Control (n = 36)	Pre-Eclampsia (n = 63)	p Value
Right ventricular parameters
RVSP, mm Hg	22.5 ± 6.1	31.0 ± 7.9	<0.001*
RA area, cm2	14.6 ± 3.2	14.9 ± 2.9	0.601
IVC, cm	1.4 ± 0.4	1.5 ± 0.4	0.404
RV FAC, %	49.6 (46–54)	50.1 (46–55)	0.770
S′, cm/s	14.7 ± 2.3	15.6 ± 3.2	0.179
TAPSE, cm	2.4 ± 0.4	2.6 ± 0.4	0.030
Basal RV longitudinal strain, %	−24.1 ± 8.2	−19.5 ± 8.5	0.0326
Mid RV longitudinal strain, %	−22.4 ± 6.4	−19.8 ± 6.4	0.110
Apical RV longitudinal strain, %	−25.2 ± 6.4	−19.7 ± 8.1	0.0055
Global RV longitudinal strain, %	−23.8 ± 2.9	−19.6 ± 3.2	<0.001*
Left ventricular parameters
LA area, cm2	17.3 ± 2.9	20.1 ± 3.8	<0.001*
LV septal wall thickness, cm	0.8 (0.7–0.9)	1.0 (0.8–1.2)	<0.001*
LV posterior wall thickness, cm	0.8 (0.7–0.9)	1.0 (0.9–1.1)	<0.001*
LVEF, %	61.2 ± 5.1	64.0 ± 6.1	0.017
Mitral E, cm/s	84.28 ± 17.9	98.5 ± 22.9	0.002
Mitral A, cm/s	57.3 ± 12.7	74.7 ± 23.1	<0.001*
Mitral septal e′, cm/s	11.6 ± 1.9	9.6 ± 2.4	< 0.001*
Mitral E/e′ ratio	7.4 ± 1.6	10.84 ± 2.8	<0.001*
Mitral E/A ratio	1.4 (1.2–1.8)	1.3 (1.0–1.6)	0.109
Mitral DT, ms	198.3 ± 38.0	193.2 ± 53.0	0.606
Basal LV longitudinal strain, %	−16.8 ± 2.3	−15.5 ± 5.1	0.25
Mid LV longitudinal strain, %	−19.0 ± 2.5	−17.4 ± 2.2	0.0077
Apical LV longitudinal strain, %	−24.8 ± 3.2	−24.0 ± 3.2	0.33
Global LV longitudinal strain, %	−20.1 ± 1.5	−19.1 ± 1.5	0.0106
BNP, pg/ml	44.0 (20.0–81.0)	49.9 (20.0–100.0)	0.22

Values are mean ± SD or median (interquartile range). The threshold of significance is p < 0.001.

^*Statistically significant difference between groups.

A = Mitral inflow velocity of late diastolic filling; BNP = B-type natriuretic peptide level; DT = Deceleration time; E = mitral inflow velocity of early diastolic filling; e′ = tissue Doppler mitral annular velocity; IVC = inferior vena cava; LA area = left atrial area; LV = left ventricular; LVEF = left ventricular ejection fraction; RA area = right atrial area; RV FAC = right ventricular fractional area of change; RVSP = right ventricular systolic pressure; S′ = peak systolic velocity at the tricuspid annulus; TAPSE = transannular planar systolic excursion.

RIGHT-SIDED CARDIAC PARAMETERS.

RVSP could not be calculated in 19 of 63 (30%) of the PEC participants and 13 of 36 (36%) of the control patients because they had inadequate Doppler envelopes due to limited tricuspid regurgitation. There were no differences in demographic or clinical characteristics between the women with or without RVSP measurements except for body mass index (BMI), which was higher in those without RVSP measurements (35.6 ± 7.2 vs. 33 ± 9; p = 0.012). Among those in whom RVSP could be calculated, PEC was associated with higher values (31.0 ± 7.9 mm Hg) than control patients (22.5 ± 6.1 mm Hg) (p < 0.001) (Table 2). Also, PEC-cHTN and PEC no-cHTN subgroups each had higher RVSP when compared with control patients (28.6 ± 9.4 mm Hg [PEC-cHTN] and 31.7 ± 7.3 mm Hg [PEC no-cHTN]; p < 0.001) (Table 3). When compared with the control patients, there were no differences in RA area, or RV function parameters, including fractional area of change, S′ velocity, and TAPSE (Table 2). In addition, there were no differences in right-sided parameters between the 2 PEC subgroups (Table 3).

For RV strain analysis, 1 participant in the control group (2%) and 6 in the PEC group (9.5%) were excluded due to poor image quality. Women with PEC had decreased global RVLSS compared with the control group, at –19.6 ± 3.2% vs. –23.8 ± 2.9% (p < 0.001) (Table 2), with 39% meeting the diagnostic echocardiographic criteria for reduced global RVLSS (34). Global RVLSS was reduced in both PEC groups (Table 3) when individually compared with the control group; however, there was no difference in RV strain, either global or regional, between PEC groups.

LEFT-SIDED CARDIAC PARAMETERS.

There were no differences in systolic function parameters between the PEC and control groups. Mitral septal e′ velocities, which are lower in patients with abnormal LV relaxation, were lower in the PEC group (9.6 ± 2.4 cm/s vs. 11.6 ± 1.9 cm/s; p < 0.001) (Table 2). Mitral septal E/e′ ratios, which are elevated in patients with increased LV filling pressures, were higher in the PEC group (10.8 ± 2.8 vs. 7.4 ± 1.6; p < 0.001) (Table 2). There were no differences in E/A ratios. LA area, which is typically greater with left-sided volume or pressure overload, was greater in the PEC group (20.1 ± 3.8 cm² vs. 17.3 ± 2.9 cm²; p < 0.001) (Table 2). On the basis of the 2016 ASE guidelines, 8 of 63 women (12.7%) with pre-eclampsia had diagnostic echocardiographic criteria for grade II diastolic dysfunction. Using the older 2009 ASE guidelines, 18 (28.6%) had diastolic dysfunction, with 9 having grade I dysfunction and the other 9 having grade II diastolic dysfunction. Also, LV posterior and septal wall thickness were significantly greater in the PEC group compared with control patients (median [interquartile range]: 1.0 cm [0.9 to 1.1 cm] vs. 0.8 cm [0.7 to 0.9 cm], p < 0.001; and 1.0 cm [0.8 to 1.2 cm] vs. 0.8 cm [0.7 to 0.9 cm], p < 0.001) (Table 2).

For LV strain analysis, STE data could not be obtained in 15 women (23.8%) in the PEC group because of poor image quality. There were no differences in demographic or clinical characteristics between the women with or without STE data except for BMI, which was higher in those with unobtainable STE data (40 ± 9.1 vs. 34 ± 7.6; p = 0.012). Among those with interpretable STE data, women with pre-eclampsia had a trend toward lower global LVLSS compared with the control group, at −19.1 ± 1.5% vs. 20.1 ± 1.5% (p = 0.0106), which was not statistically significant (Table 2).

When comparing PEC subgroups (PEC-cHTN and PEC no-cHTN), there were no differences in diastolic, systolic, or strain parameters (Table 3). The PEC-cHTN and PEC no-cHTN subgroups each had lower mitral septal e′ values, higher mitral E/e′ ratios, thicker septal and posterior walls, and larger LA areas when compared with the control group (Table 3).

LABORATORY PARAMETERS.

BNP levels were not significantly different between the PEC and control groups (Tables 2 and 3).

MATERNAL PERINATAL CLINICAL OUTCOME MEASURES.

There were no adverse clinical events in the control group. In the PEC group, 1 patient died due to delayed postpartum hemorrhage unrelated to pre-eclampsia. No patients with PEC developed eclamptic seizures or cerebrovascular accidents; however, 6 (9.5%) developed pulmonary edema, 2 of whom had PEC-cHTN (Online Table 2). There were no statistically significant differences in demographics or clinical characteristics between the PEC women with and without pulmonary edema (Online Table 3). BNP levels were higher in the PEC group with pulmonary edema; however, this was not statistically significant (106.0 pg/ml [72 to 449 pg/ml] vs. 37.4 pg/ml [20 to 89 pg/ml]; p < 0.016) (Online Table 3).

Echocardiographic data showed higher E/e′ ratios in the 6 patients with PEC with pulmonary edema compared with the 57 without (15.3 vs. 9.9) (Online Table 3), and both had higher values than control patients (7.4) (p < 0.001). Otherwise, there were no statistically significant differences in other echocardiographic parameters. Of note, 2 of the 6 women (Participants #2 and #6) (Online Table 2) were diagnosed with pulmonary edema 9 h before undergoing echocardiography. Of the remaining 4 patients, 3 developed pulmonary edema 6 to 9 h after echocardiography, and 1 patient developed pulmonary edema 144 h later.

There were no differences in intravenous fluid administration, urine output or total measurable fluid balance between the pre-eclamptic patients who developed pulmonary edema and those who did not.

RACE-STRATIFIED POST HOC ANALYSIS OF ECHOCARDIOGRAPHIC AND LABORATORY VARIABLES.

Post hoc analyses stratifying outcome variables by race revealed no statistically significant differences in RVSP or other parameters in black patients, most likely due to the smaller number of black women in the control group (7 control patients compared with 40 with PEC) (Online Table 4). Nonetheless, there was a trend for lower global RVLSS in black participants. In the 24 control patients and 20 PEC white women, statistically significant differences persisted in mitral E, mitral A, and septal e′ values, as well as in E/e′ ratios (Online Table 5).

DISCUSSION

In a racially diverse cohort, women with PEC had higher RVSP levels, diminished RVLSS, increased LA size, abnormal LV cardiac relaxation, increased LV wall thickness, and increased LV filling pressures (Central Illustration). Another important finding is that 13% of PEC participants had grade II diastolic dysfunction. Our results provide further evidence that diastolic dysfunction and LV remodeling occur during pre-eclampsia (18,21,23,38–40). In addition, we confirm a reduction in RVLSS, likely due to a combination of intrinsic subclinical RV dysfunction and increased RV afterload (21), and increased pulmonary artery pressures (i.e., elevated RVSP) (19).

CENTRAL ILLUSTRATION. Pre-Eclampsia With Severe Features: Effects on the Heart.

The effects of pre-eclampsia with severe features (PEC) on the heart. LV = left ventricle; RVLSS = right ventricular longitudinal systolic strain; RVSP = right ventricularsystolic pressure.

Notably, our study includes a high proportion of U.S. black women and women who presented early (<34 weeks gestation) with pre-eclampsia, and is significant for only enrolling patients with PEC. Thus, our cohort represents a unique study population with more severe manifestations of pre-eclampsia and a greater risk of developing acute cardiac complications. Most importantly, 10% of women with pre-eclampsia had pulmonary edema, which is 2 to 3 times higher than previously reported rates (41). Although reduced colloid oncotic pressure, increased hydrostatic pressure, intrinsic lung and pulmonary vasculature dysfunction, and the routine administration of magnesium sulfate are all associated with PEC and can contribute to pulmonary edema (42,43), our study suggests that cardiac dysfunction played a contributing role. All the women who developed pulmonary edema had abnormally elevated septal E/e′ ratios, which suggests high LV filling pressures and diastolic dysfunction. Our findings suggest that greater use of echocardiography in patients with severe pre-eclampsia may help identify particularly high-risk women and help improve clinical outcomes. Imaging findings could also affect fluid management strategies and antihypertensive therapies, especially in patients with elevated RVSP levels, RV abnormalities, or E/e′ ratios.

Also, BNP, a biomarker of increased myocardial stress (44–46), was higher in the women who developed pulmonary edema. In our study, this difference did not reach statistical significance, likely due to our study’s limited power and conservative p value after applying Bonferroni correction. However, other studies have found increased BNP and N-terminal pro-BNP levels in women with pre-eclampsia, particularly in those with early-onset (<34 weeks gestation) pre-eclampsia (44–46). We suggest that elevated BNP levels may help identify women at higher risk of developing pulmonary edema, although larger cohort studies measuring BNP levels at multiple time points are needed to definitively determine the risk of developing short-term cardiovascular events.

In addition, our study highlights the importance of STE. We found that women with PEC have lower RVLSS parameters despite having normal TAPSE, fractional area of change, and S′ values. Because RVLSS is a less load-dependent parameter, STE-derived RV parameters may be more sensitive markers that can identify early RV dysfunction in the hypervolemic state associated with pregnancy. In contrast to other studies (25,26), we did not find a significant difference in global LVLSS, perhaps because we continued treatment with antihypertensive medications at the time of echocardiography, or because more women in our cohort with unobtainable STE data had higher BMI levels. Either reason may have attenuated any differences in LVLSS.

STUDY LIMITATIONS AND STRENGTHS.

Obtaining accurate echocardiographic measures is inherently dependent on adequate visualization of the heart, which is more technically challenging to acquire in advanced pregnancy. Also, RVSP cannot be estimated in the 30% of patients who have minimal tricuspid regurgitation. Furthermore, noninvasive measurement of RVSP is not as accurate as direct measurement of pressures during right heart catheterization, which remains the gold standard for ascertaining intracardiac pressures (47). However, it is not practical or ethical to perform right heart catheterization in pregnant women for research purposes alone. STE parameters also require adequate 2D image quality and could not be measured in 24% of our patients, especially in those with higher BMI levels. Additionally, STE results are derived by application-specific algorithms, which may limit comparison of our results to other studies using alternative STE applications but should not influence the differences we found between PEC patients and control patients within our study. We also note that magnesium sulfate, administered to patients with PEC, but not to control patients, affects cardiac conduction (48) and can potentially affect echocardiographic parameters. We also did not have echocardiographic data on participants before pregnancy or their diagnosis of preeclampsia, and did not serially assess cardiac function postpartum. Therefore, we were not able to assess whether there was pre-existing or sustained cardiac dysfunction. Our study was not adequately powered to evaluate the association between echocardiographic parameters and adverse maternal perinatal cardiac outcomes so analysis of these outcomes was exploratory in nature. Finally, we did not match control and PEC patients by race, ethnicity, or other characteristics, although we did match by gestational age. Future studies should incorporate a greater number of black women in both control and study cohorts.

Despite these limitations, our study has many strengths. We studied a relatively large number of women with severe pre-eclampsia and control patients using state-of-the-art echocardiographic assessments of systolic and diastolic function, and observed higher RVSP levels, LV diastolic dysfunction, decreased RV strain, and abnormal cardiac remodeling in women with PEC. Cohort studies with repeated echocardiographic measures are needed to confirm a predictive association between abnormal echocardiographic parameters and subsequent adverse cardiovascular events. We speculate that women diagnosed with PEC and with diastolic dysfunction, abnormal strain parameters, LV remodeling, and/or elevated RVSP are more likely to develop short-term adverse cardiovascular outcomes, such as pulmonary edema. In addition, it is possible that these same abnormal echocardiographic parameters predict long-term cardiovascular events, because it is well-known that a significant subset of women who have had preeclampsia are more prone to develop coronary artery disease, cardiomyopathy, and other cardiovascular diseases (20,38,49–53). The women with echocardiographic abnormalities in particular may also benefit from serial clinical and imaging follow-up, and aggressive postpartum surveillance and treatment of their cardiovascular risk factors.

CONCLUSIONS

Women with severe pre-eclampsia have elevated RVSP values, diastolic dysfunction, diminished RV strain, and abnormal cardiac remodeling when compared with normotensive pregnant control patients. Although mean RVSP measurements are in the mildly elevated range for nonpregnant patients, 13% of patients with PEC had echocardiographic evidence of diastolic dysfunction, 39% had abnormal RV strain, and 10% developed pulmonary edema. Future studies should determine the timeline for development of the abnormal echocardiographic findings associated with pre-eclampsia and whether they are associated with short- and long-term cardiovascular sequelae.

PERSPECTIVES.

COMPETENCY IN MEDICAL KNOWLEDGE:

When severe, pre-eclampsia can be associated with left atrial enlargement, increased left ventricular mass, diastolic dysfunction, and reduced right ventricular strain.

TRANSLATIONAL OUTLOOK:

Further studies are needed to delineate the time course over which abnormalities of cardiac structure and function develop in patients with pre-eclampsia and the long-term impact of these changes on cardiovascular outcomes.

ABBREVIATIONS AND ACRONYMS

2D

2-dimensional

ASE

American Society of Echocardiography

BMI

body mass index

BNP

B-type natriuretic peptide

EF

ejection fraction

LA

left atrial

LV

left ventricular

LVLSS

left ventricular longitudinal systolic strain

PEC

pre-eclampsia with severe features

PEC-cHTN

pre-eclampsia with pre-existing chronic hypertension

PEC no-cHTN

pre-eclampsia without pre-existing chronic hypertension

RA

right atrial

RV

right ventricular

RVLSS

right ventricular longitudinal systolic strain

RVSP

right ventricular systolic pressure

STE

speckle tracking echocardiography

TAPSE

transannular planar systolic excursion