Objective
We sought to determine if increased placental vascular impedance to flow is associated with changes in fetal cardiac function using spatiotemporal image correlation and virtual organ computer-aided analysis.
Study Design
A cross-sectional study was performed in fetuses with umbilical artery pulsatility index >95th percentile (abnormal [ABN]). Ventricular volume (end-systole, end-diastole), stroke volume, cardiac output (CO), adjusted CO, and ejection fraction were compared to those of 184 normal fetuses.
Results
A total of 34 fetuses were evaluated at a median gestational age of 28.3 (range, 20.6–36.9) weeks. Mean ventricular volumes were lower for ABN than normal cases (end-systole, end-diastole) with a proportionally greater decrease for left ventricular volume (vs right). Mean left and right stroke volume, CO, and adjusted CO were lower for ABN (vs normal) cases. Right ventricular volume, stroke volume, CO, and adjusted CO exceeded the left in ABN fetuses. Mean ejection fraction was greater for ABN than normal cases. Median left ejection fraction was greater (vs right) in ABN fetuses.
Conclusion
Increased placental vascular impedance to flow is associated with changes in fetal cardiac function.
Abnormal umbilical artery (UA) Doppler velocimetry reflects increased impedance to blood flow in the placenta. Mathematical modeling of the placental circulation shows that initially, placental resistance and pulsatility index (PI) increase very slowly with fractional terminal vessel obliteration. However, there is a steep increase of the PI after 60-90% of vessels are obliterated.
In human pregnancies, structural heart disease, small for gestational age with normal UA Doppler velocimetry, intrauterine growth restriction (IUGR), twin-to-twin transfusion syndrome, and intraamniotic infection (reported in animal models also ) can result in fetal cardiac dysfunction. The heart is a central organ in the fetal adaptive mechanisms to placental insufficiency and hypoxia. Therefore, it follows that placental insufficiency with increased placental vascular resistance may lead to fetal cardiovascular compromise, and even fetal metabolic acidosis and death. Indeed, severe IUGR due to placental insufficiency contributes to 30% of total perinatal loss and severe morbidity. Monitoring of fetal cardiac function has been proposed as an adjunct to current methods to predict adverse outcome and death in IUGR. Fetuses with abnormal UA Doppler velocimetry have been shown to have similar changes to those observed in adults with atherosclerosis. This may be important in relating placental vascular disease (detected by UA Doppler velocimetry) to the risk for adult cardiovascular disease. Studies report that fetuses with abnormal UA Doppler velocimetry have evidence of higher red blood cell count and hemoglobin concentration, endothelium activation, platelet activation (which promotes thrombosis), platelet consumption, an atherogenic lipoprotein profile, and evidence of intravascular inflammation. Epidemiologic studies and animal models have also established that low-birthweight babies have an increased risk of cardiovascular disease later in life. Thus, the condition that is the focus of our study is the in utero equivalent to fetal atherosclerosis, and this, along with fetal cardiac dysfunction, may have important consequences in fetal programming of cardiac disease and the early onset of disease. Examining fetal cardiovascular parameters is required to gain an understanding of the hemodynamic changes occurring in the setting of increased placental vascular impedance to flow.
However, the repeatability and reproducibility of most fetal echocardiographic measurements determined using 2-dimensional (2D) sonography is poor, particularly for ventricular volume and volume flow estimations. This has been attributed to measurement variation in the atrioventricular (AV) and semilunar valves, which can lead to large differences in the estimated cardiac output (CO). Errors in measuring velocity-time integral or valve area will greatly influence volume flow measurements, particularly because the valve area is related to the square of the radius, thus accentuating any errors. Moreover, the use of 2D measurements to estimate ventricular volume requires assumptions about the 3-dimensional (3D) geometry of the heart that may be invalid, leading to inaccuracy in estimations of CO.
Both 3D and 4-dimensional (4D) sonography have the potential to minimize the limitations inherent in 2D estimations of fetal cardiovascular parameters because: (1) geometric assumptions are not made when assessing ventricular volumes; (2) neither small outflow tract diameters nor angle-dependent Doppler measurements are required for calculation; and (3) from a single cardiac dataset obtained using spatiotemporal image correlation (STIC), all parameters required for calculation (left and right ventricular volumes) are present in the same volume, reducing the risk inherent in measuring 2 chambers at different times when using 2D ultrasound. Indeed, 3D and 4D echocardiography have been used to evaluate cardiovascular parameters in normal fetuses.
Yet, there are insufficient data regarding the fetal cardiovascular response to increased placental vascular impedance to flow determined using 4D sonography. We have previously described a repeatable and reproducible technique to quantify ventricular volume calculations using STIC and virtual organ computer-aided analysis (VOCAL). Subsequently, we quantified fetal cardiovascular parameters (ventricular volume, stroke volume, CO, and ejection fraction) in a group of 184 normal fetuses over a range of gestational ages. Therefore, the objective of this study was to use the same technique to determine if increased placental vascular impedance to flow is associated with changes in fetal cardiac function.
Materials and Methods
Study population
A cross-sectional study was conducted to include pregnancies with increased placental vascular impedance to flow (UA PI >95th percentile ) by searching our database of women enrolled into research protocols that included examination of the fetal heart by 3D and 4D ultrasound. Women were eligible for inclusion if gestational age was determined by either a first- or second-trimester sonographic examination and there was a singleton fetus (>19 weeks of gestation). Women were excluded in the presence of fetal hydrops or chromosomal or congenital abnormalities. A control group, consisting of 184 normal fetuses whose cardiovascular parameters had been previously reported, was used for comparison.
IUGR was defined as an abdominal circumference (AC) <5th percentile for gestational age with UA PI >95th percentile. Estimated fetal weight (EFW) was not used to determine the presence of IUGR. Fetal Doppler recordings were obtained from the UA (free loop of cord), middle cerebral artery, and ductus venosus when possible. Preeclampsia was defined as the presence of systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg, and proteinuria of 300 mg/24 hours or ≥+2 (dipstick) on 2 occasions 6 hours apart. All women provided written informed consent prior to undergoing sonographic examination. Participation was approved by the institutional review board of the Eunice Kennedy Shriver National Institute of Child Health and Human Development and by the Human Investigation Committee of Wayne State University.
Examination technique
Ultrasound examinations were performed by 8 experienced sonographers using systems with STIC capability (Voluson 730 Expert, Voluson E8 Expert; GE Medical Systems, Kretztechnik GmbH, Zipf, Austria) and utilizing a motorized curved-array transabdominal transducer (2-5 or 4-8 MHz). Tissue harmonic imaging was used for each examination, and compound resolution imaging was used at the sonographer’s discretion. A transverse view of the fetal chest at the level of the 4-chamber view was obtained, from which STIC datasets were acquired. The transducer was oriented such that the fetal spine was located posteriorly for each acquisition. Acquisition time was 10 seconds with a sweep angle that was sufficient to encompass the fetal cardiac structures (25-35 degrees). Color Doppler sonography was not utilized during the acquisition process. Adequate cardiac datasets were accepted for postprocessing if acoustic shadowing (signal loss in the sound path secondary to echogenic structures), dropout (signal loss in the sound path without intervening structures), and motion artifact were absent. When multiple STIC datasets were available, the dataset obtained closest to the time of delivery was selected for analysis.
Cardiac datasets were acquired to investigate the following fetal cardiovascular parameters: (1) ventricular volume (mL); (2) stroke volume (mL) (end-diastolic volume – end-systolic volume); (3) CO (mL/min) (stroke volume × fetal heart rate); and (4) ejection fraction (%) (stroke volume/end-diastolic volume × 100%). Fetal biometry of the biparietal diameter, head circumference (HC), AC, and femoral diaphysis length (FL) were obtained using 2D sonography at the time of cardiac dataset acquisition. CO was expressed both as a function of EFW, and as a function of biometric parameters (HC, AC, FL).
Analysis was performed offline (4D View, versions 5.0-7.0; GE Healthcare, Milwaukee, WI) in a standardized manner. In the A plane of the multiplanar display, the fetal heart was reoriented such that the left ventricle was located on the left side of the screen with the apex of the heart directed upward. Next, the ventricular septum was rotated to 90 degrees in both the A and C planes. The AV valves were located by scrolling from front to back in the A plane. The image was optimized by selecting Chroma Color (Sepia) with the addition of speckle reduction imaging (SRI 5). After image brightness and contrast settings were optimized, end-systolic and end-diastolic phases were identified by scrolling through each frame, and locating the image preceding AV valve opening (systole) and following AV valve closure (diastole).
Ventricular volumes were calculated in a semiautomated fashion utilizing VOCAL. VOCAL II was selected and the Contour Finder: Trace option was utilized with 15 degrees of rotation and a sensitivity of 1 (default = 5). The image was enlarged and the reference dot repositioned into the ventricle of interest. Due to the complex geometry of the ventricles, the location of the reference dot within the ventricle was selected to meet the software requirement that the contour only cross the rotation line twice. With these selections, 12 rotational steps were made and a volume was computed. Datasets were accepted for analysis if the ventricular septum, ventricular walls, and AV valves were visible throughout each rotational step.
We previously reported the repeatability and reproducibility of ventricular volume measurements utilizing this technique. Volume measurements were repeatable with good agreement (coefficient of variation [CV] <10%) and excellent reliability (intraclass correlation >0.95) for both intraobserver and interobserver measurements. Additionally, ventricular volumes were reproducible with negligible differences in agreement (CV <1%), good reliability (intraclass correlation >0.9), and minimal bias (mean percent difference −0.4%; 95% limits of agreement, −5.4% to 5.9%) when different STIC datasets for the same patient were compared.
Statistical analysis
Data were first assessed using numerical and graphical techniques, including scatter plots of each response vs gestational age, to determine whether they met assumptions of the statistical tests being used for analysis. All but 2 scatter plots revealed the presence of curvilinear relationships and heteroscedasticity; hence, natural logarithmic transformations (from the Box-Cox family of transformations) of each response and gestational age were performed to linearize the data and correct for heteroscedasticity.
Analyses of covariance based on weighted regression were performed on the transformed data ( Table 1 ), with the weights computed according to the procedure described by Altman. These weights are the best linear unbiased estimates. Ejection fraction for the right and left ventricles were both linear and homoscedastic; therefore, analysis of covariance of 2-factor interaction and main effects multiple regression models were used iteratively to analyze these untransformed data. Residual analysis was performed on all models as a diagnostic measure to assess the aptness of the models fit.
| Cardiovascular parameter a | Normal (95% CI) b | ABN (95% CI) c | Proportion change d |
|---|---|---|---|
| Ventricular volume in end-systole, mL | |||
| Left | 0.43 (0.4–0.5) | 0.12 (0.02–0.2) | −72% |
| Right | 0.65 (0.6–0.7) | 0.39 (0.3–0.5) | −40% |
| Ventricular volume in end-diastole, mL | |||
| Left | 1.28 (1.2–1.4) | 0.64 (0.4–0.9) | −50% |
| Right | 1.57 (1.5–1.7) | 1.09 (0.8–1.3) | −30% |
| Stroke volume, mL | |||
| Left | 0.86 (0.79–0.93) | 0.53 (0.37–0.68) e | −38% |
| Right | 0.92 (0.85–0.99) | 0.71 (0.55–0.86) e | −23% |
| Cardiac output, mL/min | |||
| Left | 119.6 (110–129) | 71.9 (50–94) e | −40% |
| Right | 127.5 (118–137) | 96 (74–118) e | −25% |
| Cardiac output divided by HC, mL/min/cm | |||
| Left | 4.4 (4.1–4.8) | 2.8 (2.0–3.6) e | −36% |
| Right | 4.7 (4.4–5.0) | 3.7 (2.9–4.4) e | −21% |
| Cardiac output divided by AC, mL/min/cm | |||
| Left | 4.8 (4.5–5.2) | 3.2 (2.4–4.0) e | −33% |
| Right | 5.1 (4.8–5.5) | 4.1 (3.3–4.9) e | −20% |
| Cardiac output divided by FL, mL/min/cm | |||
| Left | 21.4 (19.8–23.0) | 14.2 (10.5–18.0) e | −34% |
| Right | 22.6 (21.1–24.2) | 18.4 (14.7–22.1) e | −19% |
| Ejection fraction, % | |||
| Left | 70.4 (69–72) | 82.4 (79–86) e | 17% |
| Right | 60.8 (59–63) | 66 (62–70) e | 9% |
a Mean values adjusted for gestational age;
c Normal vs ABN, P < .0001 (main effects analysis of covariance);
d Proportion change calculated as [(1 – ABN/normal) × 100%];
e ABN right ventricle vs ABN left ventricle for median stroke volume; cardiac output; cardiac output divided by HC, AC, FL; and ejection fraction, P < .05 for all, except P < .001 for ejection fraction (paired test).
For bivariate analysis, the Shapiro-Wilk and Kolmogorov-Smirnov tests were used to test for normal distribution. Student t test was used to determine the differences of the mean among groups, and Pearson correlation coefficient was utilized to assess correlations. For nonparametric data, the Mann-Whitney U test and Wilcoxon signed rank test were used to examine the difference between parameter medians, and Spearman rank correlation coefficient (r s ) was utilized to assess correlations. A P value < .05 was considered statistically significant for all comparisons. Statistical analyses were performed using SPSS, version 14 (IBM Corp, Armonk, NY) and SAS System for Windows, version 9.2 (SAS Institute Inc, Cary, NC).
Results
Patient population
Thirty-four cases met the inclusion criteria; clinical and sonographic data are shown in Table 2 . Ventricular volume (end-systole, end-diastole), stroke volume, CO, adjusted CO, and ejection fraction were determined and compared to that of 184 normal cases previously reported by our group (each with normal UA PI and AC measurements).
| Parameter | Value |
|---|---|
| Clinical characteristics | |
| Gestational age at delivery, wk | 31.6 (23.0–41.1) |
| Interval between acquisition of STIC datasets and delivery, d | 10 (1–133) |
| Birthweight, g | 1000 (282–3750) |
| Term deliveries (≥37 wk of gestation) | 20% (7/34) |
| Preeclampsia | 53% (18/34) |
| Perinatal death | 18% (6/34) |
| Sonographic characteristics | |
| Gestational age at evaluation, wk | 28.3 (20.6–36.9) |
| IUGR present | 71% (24/34) |
| Symmetrical IUGR a | 54% (13/24) |
| Asymmetrical IUGR b | 46% (11/24) |
| HC <10th percentile | 50% (17/34) |
| Amniotic fluid index, mm | 67 (17–178) |
| UA end-diastolic velocity present | 74% (25/34) |
| UA AEDV | 20% (7/34) |
| UA REDV | 6% (2/34) |
| MCA PI | 1.7 (0.74–2.59) |
| DV PIV c | 0.65 (0.24–3.06) |
| DV (reversed or absent a-wave) c | 7% (2/29) |
a HC/abdominal circumference <95th percentile for gestational age ;
b HC/abdominal circumference >95th percentile for gestational age ;
Sonographic evaluation was performed at a median gestational age of 28.3 (range, 20.6–36.9) weeks. UA waveform analysis demonstrated the presence of end-diastolic velocity in 74% (n = 25), absence in 20% (n = 7), and reversed in 6% (n = 2). Since cardiovascular parameters did not statistically differ among these 3 groups, they were considered as a single group with UA PI >95th percentile (ABN). Doppler velocimetry of the middle cerebral artery and ductus venosus was available for 100% (n = 34) and 85% (n = 29) of women, respectively.
The median gestational age at delivery was 31.6 (range, 23.0–41.1) weeks, with 20% (n = 7) of cases delivering at term (≥37 weeks of gestation). The median interval between acquisition of STIC datasets and delivery was 10 (range, 1–133) days. The median birthweight was 1000 (range, 282–3750) g. Of the 34 ABN cases, IUGR occurred in 71% (n = 24), preeclampsia in 53% (n = 18), and perinatal death in 18% (n = 6). Since no significant differences were found between cardiovascular parameters in those with and without IUGR, preeclampsia, or perinatal death, they were analyzed as a single group.
Ventricular volumes are lower in the presence of increased placental vascular impedance to flow
After adjusting for gestational age, mean volumes for the left ventricle (mL) were lower in end-systole (ABN: 0.12 vs normal: 0.43; P < .0001) and end-diastole (ABN: 0.64 vs normal: 1.28; P < .0001) ( Figure 1 and Table 1 ). Similarly, after adjusting for gestational age, mean volumes for the right ventricle (mL) were also lower in end-systole (ABN: 0.39 vs normal: 0.65; P < .0001) and end-diastole (ABN: 1.09 vs normal: 1.57; P < .0001) ( Figure 1 and Table 1 ). Moreover, there was a proportionately greater decrease [(1 – ABN/normal) × 100%] in left ventricular volume as compared to the right, in both end-systole (left: −72% vs right: −40%) and end-diastole (left: −50% vs right: −30%) ( Table 1 ).
Right ventricular volumes are greater than the left in the presence of increased placental vascular impedance to flow
Median right ventricular volumes (mL) were significantly greater than the left in end-systole (ABN right: 0.28 vs ABN left: 0.10; P < .001) and end-diastole (ABN right: 0.75 vs ABN left: 0.53; P < .001) ( Table 3 ). When a ratio of right to left ventricular volume was calculated, right ventricular volumes were greater in end-systole (median right/left: 3.2; interquartile range [IQR], 1.8–5.4) and end-diastole (median right/left: 1.5; IQR, 1.1–2.6), an effect that was independent of gestational age (systole: r s = 0.15; P = nonsignificant; diastole: r s = 0.15; P = nonsignificant). For the same median ratio of right to left ventricular volume, this was significantly different between the ABN and normal groups in both end-diastole (ABN: 1.5; IQR, 1.1–2.6 vs normal: 1.2; IQR, 0.9–1.7; P < .05) and end-systole (ABN: 3.2; IQR, 1.8–5.4 vs normal: 1.6; IQR, 1.1–2.4; P < .05).
| Cardiovascular parameter | Left ventricle | Right ventricle | P value |
|---|---|---|---|
| Volume in end-systole, mL | 0.10 (0.03–0.17) | 0.28 (0.16–0.55) | < .001 |
| Volume in end-diastole, mL | 0.53 (0.29–0.91) | 0.75 (0.5–1.52) | < .001 |
| Stroke volume, mL | 0.43 (0.26–0.78) | 0.5 (0.31–0.92) | < .05 |
| Cardiac output, mL/min | 61.1 (34.9–104.9) | 65.9 (45.3–127) | < .05 |
| Cardiac output adjusted by EFW, mL/min/kg | 67.8 (44.7–84.1) | 77.4 (66.1–123.8) | < .05 |
| Cardiac output adjusted by AC, mL/min/cm | 3.06 (1.82–4.3) | 3.11 (2.36–5.39) | < .05 |
| Cardiac output adjusted by HC, mL/min/cm | 2.69 (1.48–3.91) | 2.63 (1.94–5.06) | < .05 |
| Cardiac output adjusted by FL, mL/min/cm | 13.3 (7.5–18.7) | 14.0 (10.1–24.0) | < .05 |
| Ejection fraction, % | 83.7 (77.4–87.9) | 64.7 (58.1–76) | < .001 |
Stroke volume and CO are lower in the presence of increased placental vascular impedance to flow
Mean stroke volume (adjusted for gestational age) (mL) was lower for the left (ABN: 0.53 vs normal: 0.86; P < .0001) and right (ABN: 0.71 vs normal: 0.92; P < .0001) ventricle ( Figure 2 and Table 1 ). Similarly, mean CO (adjusted for gestational age) (mL/min) was lower for the left (ABN: 71.9 vs normal: 119.6; P < .0001) and right (ABN: 96 vs normal: 127.5; P < .0001) ventricle ( Figure 2 and Table 1 ). Median fetal heart rate (beats/min) was not significantly different between the ABN and normal groups (ABN: 141; IQR, 131–145 vs normal: 140; IQR, 134–147; P = .44).
CO adjusted for fetal size remains lower in the presence of increased placental vascular impedance to flow
Fetal CO was adjusted for EFW, which was calculated using biometric parameters (biparietal diameter, HC, AC, FL) obtained at the time of cardiac volume acquisition. Adjustment for EFW (CO/EFW) demonstrated that neither the left nor the right CO (mL/min/kg) changed significantly as gestation advanced (left CO/EFW ABN: r s = 0.13, P = nonsignificant; right CO/EFW ABN: r s = 0.22, P = nonsignificant).
For the left ventricle, the median CO adjusted for EFW (mL/min/kg) was significantly lower in the presence of increased placental vascular impedance to flow (ABN: 67.8; IQR, 44.7–84.1 vs normal: 90; IQR, 56–127.5; P < .05) ( Figure 3 ). However, for the right ventricle, the median CO adjusted for EFW (mL/min/kg) was not significantly different between ABN and normal groups (ABN: 77.4; IQR, 66.1–123.8 vs normal: 99.9; IQR, 72.2–126; P = nonsignificant) ( Figure 3 ).
Fetal CO was also adjusted for fetal size, by dividing the CO (mL/min) by the following biometric parameters (in cm): AC, HC, and FL. In the presence of increased placental vascular impedance to flow, values adjusted for gestational age were significantly lower for the mean left CO(AC) (mL/min/cm) (ABN: 3.2 vs normal: 4.8; P = .0001), CO(HC) (mL/min/cm) (ABN: 2.8 vs normal: 4.4; P < .0001), and CO(FL) (mL/min/cm) (ABN: 14.2 vs normal: 21.4; P = .0001) ( Figure 4 and Table 1 ). Similarly, values adjusted for gestational age were significantly lower for the mean right CO(AC) (mL/min/cm) (ABN: 4.1 vs normal: 5.1; P < .0001), CO(HC) (mL/min/cm) (ABN: 3.7 vs normal: 4.7; P < .0001), and CO(FL) (mL/min/cm) (ABN: 18.4 vs normal: 22.6; P < .0001) ( Figure 4 and Table 1 ).
Right ventricular stroke volume, CO, and adjusted CO are greater than the left in the presence of increased placental vascular impedance to flow
Median right ventricular stroke volume (mL) was significantly greater than the left side (ABN right: 0.5 vs ABN left: 0.43; P < .05). Median right CO (mL/min) was also significantly greater than the left side (ABN right: 65.9 vs ABN left: 61.1; P < .05) ( Table 3 ).
After CO was adjusted for EFW and each of the 3 biometric parameters, the median right ventricular CO remained significantly greater than the left side: EFW (mL/min/kg) (ABN right: 77.4 vs ABN left: 67.8; P < .05); AC (mL/min/cm) (ABN right: 3.11 vs ABN left: 3.06; P < .05); HC (mL/min/cm) (ABN right: 2.63 vs ABN left: 2.69; P < .05); and FL (mL/min/cm) (ABN right: 14.0 vs ABN left: 13.3; P < .05) ( Table 3 ).
Ejection fraction is higher in the presence of increased placental vascular impedance to flow, and is greater on the left side
Mean ejection fraction (adjusted for gestational age) (%) was significantly higher for the left (ABN: 82.4 vs normal: 70.4; P < .0001) and right (ABN: 66.0 vs normal: 60.8; P < .0001) ventricle ( Figure 5 and Table 1 ). Moreover, the median left ejection fraction (%) was significantly greater than the right side (ABN left: 83.7 vs ABN right: 64.7; P < .001) ( Table 3 ).
