Discussion

In this exploratory study we demonstrate an echocardiographic evidence that suggests a subclinical increase in estimated mean PAP and estimated PVR, and a modest decrement in RV function in patients with an established diagnosis of early and late PEC compared with age- and gestationally- matched pregnant controls. Pulmonary artery acceleration time at cut off value of 100 ms differentiated patients with early PEC compared with controls, substantiating the likelihood of increased PA pressure. Likewise, validated equations that estimate mean PAP based on PAAT revealed an increase in estimated MPAP in PEC compared with controls. Estimated PVR was significantly elevated in both early and late PEC compared to controls. We detected significantly increased RV MPI and reduced mid-point TTAD in early PEC, but the change in RVMPI did not reach the threshold that is considered to be clinically significant (RVMPI>5), and other parameters of RV function, such as RV FAC, indexed RV SV, indexed RV CO, TAPSE, and TV S’ were unchanged. In the absence of clinical or echocardiographic evidence of valvular, or systolic, or diastolic LV dysfunction, we speculate that the increased estimated MPAP, decreased PAAT and increased estimated PVR are due to a pulmonary vascular phenomenon rather than a left heart disease etiology. Our study highlights the value of echocardiography as a safe and non-invasive tool in the assessment of cardiac structure and function in PEC patients.
Non-invasive assessment of the pulmonary circulation and RV function and structure in PEC is an evolving area of investigation. We are aware of only two studies assessing PA flow dynamics non-invasively in PEC, and both studies corroborate our findings in general, although by using slightly different methodology, without distinguishing between early and late onset PEC and without excluding patients with severe PEC[17, 18]. Importantly, although the gold standard for diagnosing PAH is right heart catherization, it is invasive and increases the risk of maternal and fetal radiation exposure. Echocardiographic estimation of PAP provides an invaluable non-invasive and safe screening tool for increased PA pressure that warrants further follow up and investigation. Similarly, echocardiography is a more feasible, bedside option of measuring RV function and structure compared to the gold standard, cardiac magnetic resonance imaging[19].
A recent meta-analysis of 21 published studies indicate that estimation of mean PAP using PAAT shows very good correlation with values obtained using catheterization, that a cutoff value of PAAT<100 ms has 84% sensitivity and 90% specificity in the diagnosis of pulmonary hypertension and that PAAT can be reliably measured in >90% of patients[14]. In our study we used both the method of using the PAAT<100ms cutoff value and also the estimation of MPAP to test whether there is evidence for increased PAP in PEC.
In our early PEC cohort, we found estimated mean PAP to be above the maximum of normal range. Both Caglar et al.[17] and Vaught et al.[18] found significantly increased estimated PA pressures in preeclamptic patients, but their reported values were below or at the limit of maximum normal pressure (20mmHg). The difference could be methodological, since they used TV regurgitant jet velocity for their estimate and we based our estimate on PAAT. Alternatively, the difference may be in the patient cohort analyzed as our findings suggest a greater effect in early PEC, and Caglar et al[17]. and Vaught et al. [18] did not analyze early PEC separately.
Our study is the first to assess TTAD in preeclamptic patients. TTAD detects the systolic displacement of the TV annulus, relative to the apex, at three points. As such, TTAD serves as a surrogate of the longitudinal deformation of the tricuspid valve annulus. The lateral point of TTAD (TTAD L) shows close correlation with TAPSE and mid-point TTAD (TTAD MP) has been shown to be more predictive of RV dysfunction in patients with PAH compared with TAPSE[20] and to better correlate with cardiac magnetic resonance imaging-measured RV EF[21]. As such, TTAD L did not show a difference between controls and PEC, whereas TTAD MP was significantly decreased in early PEC vs early controls.
Reduced RV function in PEC was shown by previous studies, yet with the use of different indices of assessment[17, 22]. The study by Melchiorre et al. [22] demonstrated biventricular increased wall thickness, and diastolic dysfunction, as well as reduced LV EF and cardiac output. Comparatively, we did not find a change in cardiac output or a biventricular diastolic dysfunction. These differences may be related to the study population. Our study did not include women with the diagnosis of severe PEC, and we excluded patients with low EF, whereas the study by Melchiorre et al [22] did not use these exclusions. Additionally, the latter study used TDI to assess strain and strain rate, a technique that suffers from some limitations[23]. The study by Vaught et al .[18] was consistent with our finding of increased RV MPI in PEC patients compared with controls, yet it demonstrated biventricular hypertrophy. This difference may be related to the study population. Vaught et al[18] studied parturients with early PEC and severe findings and we excluded patients with the diagnosis of severe preeclampsia. Collectively, our study aligns with the literature in terms of the presence of subclinical decrement in RV function and increased PA pressure in patients with early PEC. Notably, the mild changes in RV function can be considered to be subclinical. Whether the mildly reduced RV function is a consequence of increased afterload or whether it develops independently due to circulating mediators is yet to be determined. Intriguingly, a recent study found persistent abnormalities in RV function in women who formerly had early onset PEC for up to three years post-delivery[24].
Preexisting pulmonary arterial hypertension (PAH) portends an increased mortality in pregnancy and current recommendation for patients with preexisting pulmonary hypertension is to avoid becoming pregnant, due to a 15-40% incidence of mortality among women who become pregnant with preexisting PAH[25]. There is evidence that PEC and PAH share common mechanisms in their etiology [7, 26-29]. Therefore, we reason that changes in the pulmonary circulation and right ventricle function in PEC warrant further investigation.
Our study is limited by its small sample size and needs to be validated using a larger cohort that will also allow correlating echocardiographic estimates of MPAP and PVR with indices of disease severity. The small sample size precludes appropriate adjustment for comorbidities, which should be addressed in subsequent larger studies. Seven of the nineteen PEC patients in this study received anti-hypertensive therapy during the echo exam, which may have altered the echocardiographic findings compared with controls. However, none of the three drugs, labetalol, hydralazine or nifedipine are likely to cause an increase in PA pressure as we observed in this study in the PEC group. Instead, all three drugs have been proposed in the past to be used to treat PAH, although they were not shown to be effective. Given the cross-sectional design of our study, our echo exams were conducted very close to delivery. Hence the findings may represent a mixed pattern of pregnancy and labor rather than a single pattern. This can be remedied by studying serial exams in future studies. While PEC and control patients were matched by gestational age, there may be likely other systematic differences between the groups that could confound the results. Furthermore, due to the lack of 3-D capability of our echocardiography instrumentation at the time of the study, we were unable to measure 3-D volumetric LV EF, known to be more accurate echocardiographic modality of measurement of LV EF[30]. Furthermore, due to the lack of suitable apical LV 2 chamber views, we were unable to measure 2-D LVEF using the biplane modified Simpson methodology. The area-length method of LVEF estimation suffers from the limitations of geometric assumption of LV shape as well as sensitivity to foreshortening and wall distortions[9]. Yet, none of our patients had diseases that would lead to LV wall motion abnormalities known to affect the accuracy of this methodology. Rather, we used the area-length methodology to estimate LV EF as recommended by the American Society for Echocardiography and European Society of Cardiovascular Imaging. A normal LV FAC and the absence of clinical history or symptomatology of heart failure make us deduct that the elevated estimated MPAP is not due to a left heart pathology.
It is to be emphasized that echocardiographic indices of PAH are only estimations and they are used to raise the probability of PAH rather than to establish a conclusive PAH diagnosis.
In conclusion, non-invasive assessment of PA hemodynamics and RV function may have utility in assessing the cardiovascular status of PEC patients. Our findings suggest that there is an increase in estimated MPAP and estimated PVR in PEC, particularly in case of early onset. There is some evidence of a parallel mild deterioration of RV function, the significance of which is yet to be determined in future longitudinal studies. An adequately powered longitudinal study would also assess whether there is a correlation between echocardiographic findings of pulmonary hemodynamics and maternal and fetal outcomes and to assess responsiveness to antihypertensive therapy.

Acknowledgements/Sources of funding

Supported by the CounterACT Program, National Institutes of Health Office of the Director (NIH OD), the National Institute of Neurological Disorders and Stroke (NINDS), and the National Institute of Environmental Health Sciences (NIEHS), Grant Numbers= (UO1 ES027697) to