This was a prospective cohort study (November 1, 2010-September 1, 2012 inclusive) of healthy volunteers recruited from prenatal obstetric clinics at two tertiary care centres. Eligible subjects were between the ages of 18 and 35 years old at their last normal menstrual period, carrying a healthy singleton pregnancy, and had no previous pregnancy carried beyond 14 weeks gestational age. Exclusion criteria included multiple gestation; any history of cardiac disease or cardiac surgery; any history of thyroid disease, hypertension or diabetes; any significant pregnancy-related complications including gestational diabetes, gestational hypertension, or pre-eclampsia; and/or the presence of pacemakers, surgical clips or other contraindications to CMR. Cardiac imaging by both TTE and CMR was performed in all patients at two time points: initially in the third trimester and then again approximately three months postpartum. Postpartum imaging was performed as a surrogate for the baseline non-pregnant cardiovascular state and used for comparison with third trimester imaging data. The University of Manitoba Human Research Ethics Board approved the study protocol and informed consent was obtained from all patients.

Standard TTE (Vivid 7, GE Healthcare, Milwaukee, WI, US) with a multi-frequency transducer and tissue Doppler imaging (TDI) capability was performed in all patients. All TTE scans were performed with women in the left lateral decubitus position to minimize aorto-caval compression and improve patient comfort in the third trimester. The parasternal long axis view was used to obtain cardiac dimensions for the LV, right ventricle (RV), interventricular septum (IVS), posterior wall thickness (PWT), left atrium (LA) and ascending aorta. The apical four chamber and two chamber views were used to obtain LV end systolic volume (LVESV) and LV end diastolic volume (LVEDV) by manually tracing the endocardium. These measurements allowed for calculation of the LV ejection fraction (LVEF) and stroke volume (SV) using the biplane modified Simpson’s method 40 . Left and right atrial volumes (LA and RA) were calculated in the apical four chamber view at end-systole using the area length method. The calculation of LV mass using LV linear dimensions was performed according to the American Society of Echocardiography recommendations 40 . Diastolic function was evaluated using pulsed wave Doppler obtained at the level of the mitral valve, assessing peak early (E) and late (A) mitral inflow velocities, isovolumetric relaxation time (IVRT) and deceleration time (DT). The E/A ratio was calculated as a marker of diastolic function. Tissue Doppler measurements of systolic (S’), early (E’) and late (A’) tissue velocities were obtained at the medial and lateral mitral annulus. The dimensionless E/E’ ratio was calculated as a non-invasive measure of LV filling pressure 41 . Two observers (RD and DJ), blinded to the clinical data, analyzed the echocardiographic images offline using EchoPAC PC (version 110.0.0, GE Healthcare, Milwaukee, WI, US).

The serial CMR scans were performed using a 1.5 Tesla Siemens Scanner (Sonata, Magnetom, Siemens, Erlangen, Germany). Breath-hold cine imaging was performed using a segmented TrueFISP sequence (temporal resolution 25 ms, slice thickness 8 mm, interslice gap 4 mm, spatial resolution 117 × 192, field of view 360 mm, and GRAPPA parallel imaging) with ECG gating to obtain images representative of the entire cardiac cycle in both long and short axis views. A combination of a 32 channel spine matrix and 6 channel body matrix coil was used. All patients were imaged in a half left lateral decubitus position to minimize aorto-caval compression and improve patient comfort in the third trimester. There were no sedative medications or contrast agents used during this study. Chamber measurements were obtained from the three and four chamber long axis views. Measurement of the aortic sinus, sinotubular junction, and ascending aorta were performed in both the three-chamber and coronal views. Endocardial and epicardial contours were drawn manually for the LV and RV, respectively, at end-systole and end-diastole in each data set with the most basal short axis slice identified as the image which contains at least 50% of circumferential myocardium. Papillary muscles and trabeculations were included in LV and RV mass calculation. Two observers (RD and DJ), blinded to the clinical data, analyzed the CMR images offline using CMR analysis suite (version 3.4.0, Circle Cardiovascular Imaging, Calgary, Alberta, Canada).

All parametric data was reported as mean ± standard deviation (SD). Categorical data was reported as “n” (percentage). Comparison of non-pregnant and third trimester means for each imaging modality was performed using a paired Student’s t-test. Linear regression analyses were performed to assess the correlation between TTE and CMR for CO, stroke volume (SV) and LV mass. Bland-Altman plots of difference versus mean were performed for CO, SV and LV mass with 95% agreement limits of ± 1.96 SD. The Mann–Whitney U test was used to measure the intra and inter-observer variability for LVEDV and LV mass for both imaging modalities. Statistical significance was defined as p < 0.05. SAS version 8.01 (SAS Institute Inc., Cary, NC, US) was used to perform the analysis.

The baseline clinical characteristics of the study population (n = 34) are summarized in Table 1. The mean maternal age at the last normal menstrual period was 29 ± 3 years with a mean pre-pregnancy BMI of 24 ± 4 kg/m². The mean gestational age at third trimester imaging was 237 ± 16 days (34 weeks ± 16 days) and mean number of days for postpartum imaging was 107 ± 25 days (16 weeks ± 25 days).

Data are expressed as mean ± SD or number (percentage). BSA, body surface area; BMI, body mass index; BP, blood pressure; GA, gestational age; LNMP, last normal menstrual period.

The structural and systolic functional parameters for both imaging modalities in the third trimester and postpartum (baseline) settings are summarized in Table 2. There was a statistically significant increase in LV end diastolic diameter (LVEDD) and LVEDV in the third trimester using both TTE and CMR, with narrower confidence intervals using CMR. Both TTE and CMR demonstrated a statistically significant increase in LV mass in the third trimester. By TTE, there was an increase in LV mass from 111 ± 10 g to 163 ± 11 g, representing a 47% increase above baseline values. Similarly, CMR demonstrated a 48% increase in LV mass over the same period, from 121 ± 5 g to 179 ± 5 g, with tighter confidence intervals. The increase in heart rate (HR) and SV between baseline and the third trimester observed with both imaging modalities, resulted in an appropriate increase in CO during pregnancy. Using TTE and CMR, there was an increase in CO from 3.5 ± 0.8 L/min to 6.3 ± 0.7 L/min and from 4.0 ± 0.3 L/min to 7.4 ± 0.4 L/min, respectively (p < 0.05) (Table 2).

Data are expressed as mean ± SD. CO, cardiac output; HR, heart rate; IVS, interventricular septum; LV, left ventricle; LVEDD, left ventricular end-diastolic diameter; LVEF, LV ejection fraction; LVEDV, LV end-diastolic volume; LVESD, LV end-systolic diameter; LVESV, LV end-systolic volume; PASP, pulmonary artery systolic pressure; PWT, posterior wall thickness; RV, right ventricle; RVEDD, right ventricular end-diastolic diameter; RVEF, RV ejection fraction; RV FAC, fractional area change; SV, stroke volume; TAPSE, tricuspid annular plane systolic excursion; *P < 0.05, third trimester vs. baseline within each imaging modality.

Linear regression analysis for SV, CO and LV mass demonstrated good correlation between both TTE and CMR at baseline and in the third trimester as shown in Figure 1. The baseline R-values for SV, CO and LV mass were 0.85, 0.82 and 0.59, respectively. In the third trimester, the R-values for SV, CO and LV mass were 0.83, 0.81 and 0.67, respectively. The Bland-Altman plots for the same parameters are shown in Figure 2. At both time points, TTE showed a slightly negative bias on the Bland-Altman plots in the assessment of SV, CO and LV mass as compared to CMR.

The RV end diastolic diameter (RVEDD) increased significantly from the non-pregnant state to the third trimester, as determined by both TTE (32 ± 5 mm to 39 ± 3 mm) and by CMR (33 ± 4 mm to 39 ± 3 mm) (Table 2). RV mass as analyzed by CMR increased from a mean of 51 ± 5 g at baseline to 71 ± 6 g in the third trimester (p < 0.05), while RV volume increased from 93 ± 4 mL to 115 ± 4 mL (p < 0.05). There was no change in the RV systolic function as measured by TTE or CMR.

Both LA and RA enlargement was observed from the non-pregnant state to the third trimester, using TTE and CMR (Table 3). The LA end systolic volume measured by CMR increased by approximately 50% from 32 ± 3 mL to 48 ± 3 mL in the third trimester. Similarly, the RA end systolic volume measured by CMR increased by 55% from 29 ± 2 mL to 45 ± 4 mL. A smaller degree of biatrial enlargement was observed using TTE.

Data are expressed as mean ± SD. LA, left atrial; RA, right atrial. *P < 0.05, third trimester vs. baseline within each imaging modality.

There was no statistically significant change in aortic root or ascending aortic dimensions demonstrated from non-pregnant to third trimester imaging using either TTE or CMR (Table 3).

Diastolic parameters obtained by TTE are outlined in Table 4. There was an increase in trans-mitral A wave velocity from 0.7 ± 0.1 cm/s to 1.3 ± 0.2 cm/s from the non-pregnant state to the third trimester (p < 0.05). The A wave increase resulted in a decreased E/A ratio in the third trimester from 1.3 ± 0.1 cm/s to 0.7 ± 0.2 cm/s (p < 0.05), consistent with mild diastolic dysfunction. The mitral DT, IVRT, and tissue Doppler parameters (including E/E’) did not change significantly from non-pregnant to third trimester imaging.

Data are expressed as mean ± SD. A, late diastolic transmitral velocity; A’, late diastolic myocardial velocity; E, early diastolic transmitral velocity; E’, early diastolic myocardial velocity; DT, deceleration time; IVRT, isovolumic relaxation time; S’, systolic myocardial velocity; TDI, tissue Doppler imaging. *P < 0.05, third trimester vs. baseline.

The intra- and inter-observer variability of LVEDV and LV mass using both imaging techniques are shown in Table 5. CMR yielded higher reproducibility of cardiac dimensions compared to TTE.

Absolute values are defined as the population mean ± SD of absolute differences between repeated measurements. % values are defined as the population mean ± SD of absolute differences of repeated measurements normalized by the average of the two repeated measurements. CMR, cardiovasculart MR; LVEDV, left ventricular end-diastolic volume; LV mass, left ventricular mass; TTE, transthoracic echocardiography.

The maternal cardiovascular system undergoes significant physiologic changes in order to support the developing fetus. Previous studies, using TTE in normal pregnancy, have reported that the HR typically increases by 15-20%, SV increases by 20-25%, and CO increases by 30- 50% 4 9 10 11 13 15 . In our study, we demonstrated that the HR increased by approximately 20%, SV increased by 40%, and CO increased by 80-85% using TTE and CMR. Although the increase in HR seen in our study was similar to previous studies, we observed a greater increase in SV and CO during pregnancy. Recent improvements in endocardial visualization by both TTE and CMR, as well as differences in study design and imaging time points may account for this finding. To the best of our knowledge, this is the first study to longitudinally assess CO and SV using CMR in pregnancy, which is known to have improved accuracy and reproducibility for assessment of LV volumes 24 . Such findings may have implications in pre-conception counseling and managing women with pre-existing heart disease or myocardial dysfunction in pregnancy.

During pregnancy, total blood volume expansion leads to an increase in ventricular preload and compensatory structural changes in the LV. The increases in LV volume and myocardial mass are physiologic adaptations enabling the maternal heart to support the demands of the developing fetus. Previous studies have reported an increase in LVEDV and LV mass in the third trimester, relative to the non-pregnant state 11 12 14 . In the current study, using both TTE and CMR, we demonstrated an increase in LVEDV of 20-30% with an increase in LV mass of 45-50%, similar to previous studies. Our finding of slightly higher values for LV volume and mass both at baseline and in the third trimester for CMR, as compared to TTE, is a reflection of differing image analysis techniques. Whereas CMR routinely includes papillary muscles in volume and mass analyses, they are excluded in echocardiographic measurements 24 26 . Nonetheless, although TTE consistently underestimated LV volumes and mass during the peripartum period as compared to CMR, there was good correlation between both imaging modalities.

Although RV dilation is an accepted physiologic change in pregnancy, poor visualization of this chamber on TTE and lack of standardized parameters have accounted for the paucity of data examining RV remodeling during pregnancy. Whereas quantitative assessment of the RV using TTE is difficult due to its complex geometry, CMR serves as the gold standard for non-invasive assessment of this chamber due to its higher spatial resolution and low intra- and inter-observer variability 42 . A previous study using TTE demonstrated an increase in RVEDD by 18% during the third trimester 17 , which is very similar to the values obtained in our study. Using CMR in our study, we also provide new reference values for an increase in RV volume and mass with preserved RV systolic function during pregnancy. In women with pre-existing congenital cardiac conditions that involve the RV, the peripartum use of CMR may be beneficial in the optimal management of this patient population.

During pregnancy, there is physiological dilatation of both atria due to an increase in the effective circulating blood volume. Similar to previous studies that demonstrated an increase in LA volumes by approximately 30% in the third trimester 10 14 43 , our study confirmed LA volume increases of 39% and 50% by TTE and CMR, respectively. Although few studies have examined RA volumes during pregnancy, Campos et al. demonstrated a 19% increase in RA dimensions in the third trimester by TTE 17 . Our study demonstrated an increase in RA volume of 34% and 55% by TTE and CMR, respectively. With the physiological increase in both atria during pregnancy, compensatory changes in diastolic function occur as well.

Traditional Doppler techniques have been used to characterize diastolic function in pregnancy. Previous echocardiographic studies have demonstrated a decline in the ratio of passive filling to atrial contraction-related ventricular filling during diastole in pregnant women at term 44 45 46 47 48 . This decrease in E/A ratio that has been postulated to reflect mild diastolic dysfunction during the pregnant state was supported in our current study. As compared to these conventional measures of diastolic dysfunction, TDI parameters are independent of preload and afterload, which are both altered by pregnancy. Our results reflect no significant change in TDI parameters during the peripartum state.

Although TTE is limited in its ability to assess aortic root and ascending aorta dimensions, CMR can provide multi-planar reformations of the great vessels allowing for greater visibility and more accurate quantification 49 . A recent comparative study of TTE and CMR in pregnant women with pre-existing aortopathies reported that CMR altered clinical decision-making in up to 50% of the study population 35 . Although the clinical utility of CMR was well reported in this study, no CMR reference parameters for aortic dimensions during a normal healthy pregnancy have been previously been established. An important new finding of our study was to provide longitudinal aortic root measurements during healthy pregnancies using CMR. In our study, we demonstrated that aortic root and ascending aorta measurements remain relatively stable during the peripartum state. This finding may have implications in pre-conception counseling and managing women with pre-existing aortopathies in pregnancy.

There are several limitations to the current study. First, we used the postpartum period as a surrogate for the pre-pregnancy state. This was done for logistical reasons as predicting when and whom to image pre-conception to allow creation of a study cohort is neither cost-effective nor practical. Second, since we did not image pregnant women during the first and second trimesters, the degree of structural and functional change in earlier pregnancy was not determined. We performed imaging in the third trimester as several studies have shown that many hemodynamic changes will be fully evolved by 28 weeks gestation 4 9 . Finally, due to the prolonged length of the CMR examinations for the pregnant women, CMR-based blood flow velocity measurements were not obtained and thus could not be compared to the analogous values measured with TTE.