Materials and Methods

Study design: A prospective cross-sectional study carried out at the University of Alabama At Birmingham from June 2016 to December 2016 and approved by the Institutional Review Board of the University of Alabama At Birmingham (RB-170119002) and written informed consent was obtained from all subjects.
Study population: Patients were enrolled sequentially as they met study inclusion criteria identified by the Obstetric Anesthesiology fellow at the time of daily patient board check out. A total of 46 patients with singleton pregnancies were enrolled into four groups: early and late PEC with early and late gestational age-matched normotensive pregnant controls. The incidence of PEC was defined and classified according to the American College of Obstetricians and Gynecologists (ACOG) (2013) revised criteria[2]. Early PEC was defined as PEC diagnosed at or before 34 weeks of gestation and late PEC was defined as PEC ≥ 34 weeks of gestation. Blood pressure at the time of enrollment was measured manually according to the guidelines of the National High Blood Pressure Education Program Working Group on High Blood Pressure in pregnancy. All of the measurements were performed in the left arm, in the sitting position, with the arm at the level of the heart. Patients underwent physical exams to rule out signs of severe PEC in the form of visual disturbances, headaches or right upper quadrant abdominal pain. Patients with severe preeclampsia, with history of heart failure or ejection fraction (EF) <40%, valvular heart disease, chronic hypertension, PA stenosis, congenital heart disease, and those who were in eclampsia, were excluded from the study. None of the patients had chronic kidney disease.
Echocardiography: TTE was performed on all participants upon the diagnosis of PEC and admission to the ward before delivery. All TTE exams were performed by a board-certified single provider (AZ) with a 10-year experience in echocardiography and ultrasonography, using the same device (SPARQ (Philips Vingmed Ultrasound, Horten, Norway) and transducer (S5-1, MHz phased-array). A second board certified observer (BJM) who was blinded to the patients’ group designation and who had not attended the initial examination analyzed the echocardiographic exams post hoc. The values entered by the initial examiner were concealed prior to the examinations entered by the second observer. Electrocardiography was recorded continuously during echocardiographic studies. Two-dimensional, M-mode, and tissue Doppler TTE imaging were performed according to American Society of Echocardiography guidelines[9].
A conventional, focused, and modified apical four-chamber view (mA4CW), parasternal long axis (PLAX), parasternal short axis (PSAX), left parasternal RV outflow, and the left parasternal RV inflow (RV) views were obtained. RV basal diameter (RV-Bd) was measured from mA4CW, focusing on RV at end-diastole. The right atrium (RA) long and short axis diameters (RAd-lax/RAd-sax) and right atrial area (RAA) were measured from mA4CW at end-systole. The RVOT outflow tract (RVOT) diameter was measured from left PSAX at end-diastole. RV free wall thickness (RV-FWT) was measured from left PLAX at end-diastole using M-mode. Pulsed wave Doppler was performed at the mitral, tricuspid, pulmonary, and aortic valves. RV systolic function was evaluated using the fractional area of change (FAC), tricuspid annular plane systolic excursion (TAPSE) and tissue Doppler imaging of the peak systolic velocity of the lateral tricuspid annulus. TAPSE was measured using M mode from the apical 4 chamber view. RV diastolic function was estimated using early and late diastolic waves across the tricuspid valve and across the tricuspid annulus. RV myocardial performance index (RV MPI), a representation of the global systolic and diastolic function of the RV was calculated as described before using pulsed wave Doppler velocity and the formula RV MPI = (RV isovolumic contraction time (IVCT)) + RV isovolumic relaxation time (IVRT)/ RV ejection time (ET) [10]. RV ejection time was measured from the pulsed wave Doppler across the right ventricular outflow tract in the left parasternal RV outflow view. RV MPI reference range RV MPI < 0.4. RV cardiac output and stroke volume were calculated in accordance of guidelines[9]..
Apical 4 chamber views were specifically optimized to visualize the RV to obtain echocardiographic cine loops by recording 3 consecutive heart cycles with a frame rate > 60 frames/sec. Offline analyses were performed using a dedicated software (Qlab 10.3, Philips Healthcare). To assess tissue-tracking tricuspid annular displacement (TTAD) user defined 3 points were selected in RV focused view as follows: at the point of insertion of the lateral (TTAD L), and septal (TTAD S) leaflets and at midpoint (TTAD MP) of the tricuspid valve to the tricuspid annulus and the RV apex. The software automatically tracked tissue-tracking tricuspid annular displacement (TTAD) and calculated the TTAD at the RV free wall, the TTAD at the interventricular septal wall and the TTAD at the midpoint of TV annulus, as well as the percent displacement of midpoint (TTAD MP%). In some patients (n=1 in Early C, n=2 in early PEC, n=3 in late C and n=2 in late PEC) the image quality was inadequate to reliably assess TTAD and in the case of one patient there was no agreement between the two observers, resulting in exclusion from the analysis. The resulting number for TTAD measurements are listed in Table 3. The apical 4 chamber and the parasternal SAX views were used to measure left ventricular (LV) EF by the area-length technique[9]. LV endocardial fractional shortening measured by M mode, fractional area of change, and lateral mitral annular peak tissue velocity measured by tissue Doppler imaging, were all measured from the parasternal long axis, parasternal SAX, and apical 4 chamber views, respectively, according to guidelines[9].
Estimation of PA pressure by echocardiography: We first attempted to estimate PA systolic pressure using the Bernoulli equation (PAP = 4 x TRvel2) from the tricuspid regurgitation jet velocity, in case a tricuspid regurgitant jet was discerned[11]. Since we only had three patients with reliably detectable tricuspid regurgitation jet, these data are not reported.
Since in most cases the tricuspid regurgitant jet was not discerned, the mean pulmonary artery pressure (MPAP) was estimated from the PAAT measured across the pulmonary valve using pulsed Doppler waveform. We were unable to use the most commonly used formulas that rely on linear equations to estimate MPAP (if PAAT is ≥ 120ms, estimated MPAP = 79 - (0.45x PAAT), if PAAT is < 120 ms, estimated MPAP = 90 - (0.62xPAAT), because several (5/11) control pregnant patients had PPAT beyond the cut-off limit of these equations (170 ms). Instead we used a logarithmic equation according to the formula validated by Kitabatake et al[12] and Yared et al [13], which is not limited by the extremes:
estimated MPAP=10-0.0068*(PAAT)+2.1
PAAT-based analysis was also analyzed by categorical assessment using cutoff value of PAAT =100 ms. It was shown that in non-pregnant individuals, PAAT<100ms predicts PAH and increased pulmonary vascular resistance with a sensitivity of 84% and specificity of 90% [14].
Pulmonary vascular resistance was estimated from estimated MPAP and RV cardiac output (RV CO) using the following formula[15]:
‘Estimated PVR=estimated MPAP/RV CO’
Estimation of left atrial pressure by echocardiography: We utilized a noninvasive method to assess the filling pressure for the left ventricle, which is based in measuring the E and A waves of mitral valve flow velocity and by measuring the velocity of the mitral annulus (e’). Mitral E/A >2 and E/e’ >14 show good correlation with elevated LV filling pressures [16].
Statistical analyses: Statistical analyses were performed using Prism (GraphPad Software, La Jolla, CA), unless noted otherwise. The primary variables for assessing differences was estimated mean pulmonary artery pressure and PAAT. Since estimated pulmonary artery pressure is directly calculated from PAAT, we considered these as a single primary variable. All other variables reported are secondary variables. Continuous variables were presented as mean ± standard deviation. Normality test was calculated by four different methods automatically by Prism: Anderson-Darling (A2*) (AD), D’Agostino-Pearson omnibus (K2) (DP), Shapiro-Wilk (W) (SW) and Kolmogorov-Smirnov (distance) (KS). PAAT, RVMPI, maternal age, GA at admission, diastolic BP, Heart rate, RV FAC, RV COi and TAPSE passed normality tests by all four methods, estimated pulmonary artery pressure, RV SVi, Systolic BP passed three, BMI, estimated PVR, TTAD S and TTAD MP % passed two, MV E/e’, TTAD L and TTAD MP only passed DP, MV E/A only passed KS, mean BP and GA at delivery did not pass any. For variables that failed at least two normality tests we used Kruskal-Wallis test with Dunn’s post-hoc analysis. Variables that tested at least three normality tests were considered to be continuous variables and ANOVA was used with Bonferroni’s post-hoc analysis. Categorical variables were analyzed using first by Fisher’s exact test for 2x4 tables (VassarStats free online calculator developed by Richard Lowry, PhD;http://vassarstats.net/fisher2x4.html) followed by pairwise Fisher’s exact test with Bonferroni’s correction. A P value <0.05, after applying correction for multiple comparisons, was considered to be statistically significant. Inter-observer reliability was determined using Bland– Altman methodology and expressed as bias (mean difference) and 95% limits of agreement (2 X SD mean difference). The correlation between estimated MPAP and mean AP was analyzed using linear regression analysis. Since this is an exploratory study with the main purpose to determine whether a larger study is warranted, the sample sizes were not based on power calculation.