Right ventricular function in severe aortic stenosis assessed by echocardiography and MRI

The prevalence of aortic valve stenosis (AS) is increasing due to an ageing population. Despite that right ventricular function has prognostic value for postoperative outcome, the right ventricle (RV) is not extensively studied and often not routinely assessed in AS. Our aim was to explore the relation between severe AS and RV function in a surgical aortic valve replacement (SAVR) cohort, comparing two imaging modalities for RV evaluation.


| BACKGROUND
Aortic stenosis (AS) is a progressive disease known to cause left ventricular (LV) failure.While systolic LV function has been extensively studied, the right ventricle (RV) has received much less attention and is often not routinely assessed in AS.
The number of patients with AS is increasing with an aging population (Danielsen et al., 2014;Iung, 2003).The prevalence of aortic stenosis in Western countries is approximately 12% in those aged ≥75 years (Osnabrugge et al., 2013).Echocardiography is the standard method for diagnosing aortic stenosis, as well as to assess its severity and prognosis.The prevalence of severe AS, defined as a valve area less than 1.0 cm 2 , is 3%-4% (Baumgartner, Falk, et al., 2017;Danielsen et al., 2014;Osnabrugge et al., 2013).Present ESC guidelines recommend aortic valve replacement (AVR) for severe AS when the patient is symptomatic or has LV dysfunction (Baumgartner, Falk, et al., 2017;Baumgartner, Hung, et al., 2017).
RV impairment in AS is associated with increased postoperative mortality and increased treatment-related costs (Bootsma et al., 2018;Duncan et al., 2017;Musa et al., 2016).Recent research showed that patients with severe AS and RV dysfunction had a reduced long-term survival (Galli et al., 2015;Rigolli et al., 2019).Bohbot et al. showed that tricuspid annular plane systolic excursion (TAPSE) <17 mm at the time of AS diagnosis is a marker of poor survival after AVR even in asymptomatic patients with severe AS (Bohbot et al., 2020).Despite these observations, current guidelines for the management of AS do not mention the importance of assessing size and function of the RV.
Our aim was to explore the relation between severe AS and RV function in a Swedish surgical AVR cohort, by both echocardiography and magnetic resonance, to explore the coexistence of RV dysfunction by both modalities and to investigate whether echocardiographic or magnetic resonance imaging methods would be most sensitive in detecting signs of RV dysfunction in patients with severe AS.

| Study population
Patients with severe AS according to the ESC/EACTS guidelines (Baumgartner, Hung, et al., 2017;Vahanian et al., 2012) were screened for inclusion when scheduled for surgical aortic valve replacement (SAVR).Patients were selected based on their voluntary participation and their possibility to undergo cardiovascular magnetic resonance imaging (CMR).Application and approval (No. 2011/105-31) for ethical permit for research involving humans was made via the Regional Ethics Review Board.Written informed consent was obtained from all participating patients.Exclusion criteria were age <18 years, other concomitant heart valve disease, congenital heart disease except bicuspid aortic valve, hemodynamic instability, previous cardiac surgery, atrial fibrillation and ischemic heart disease.
Patients included in the study underwent CMR and transthoracic echocardiography (TTE) on the same day within 1 week preoperatively when possible.

| CMR
CMR images were obtained using a Philips Ingenia 3 T.The present study analysed balanced SSFP cine images (2-, 3-and 4-chamber view and a short-axis stack of the LV/RV from base to apex).All short-axial cine-scans were manually segmented, for RV-and LV-volumes as well as LV mass, using the freely available software Segment version 3.0 R7946 (Heiberg et al., 2010;Schulz-Menger et al., 2020).All manual segmentations were carried out or supervised by the same CMR specialist physician.A semiautomatic recording of right ventricle AV-plane displacement (RV-AVPD) for TAPSE analysis and RV free wall strain (RVFWS) analysis as implemented in Segment was performed.Likewise, a manual marking of the free wall anulus, tracking it between diastole and peak systole for TAPSE, was performed.

| Analysis and statistics
RV function was determined from both TAPSE and RVFWS by TTE and from RV ejection fraction (EF) by CMR.Analysis of TAPSE and RVFWS by CMR was performed for comparison with TTE.In the following, TAPSE and RVFWS refers to TAPSE and RVFWS by TTE unless otherwise stated.RV dysfunction was defined according to the following cut-off values; TAPSE < 17 mm, RVEF < 50% or RVFWS > −20% (Kawel-Boehm et al., 2020;Lang et al., 2015;Morris et al., 2017;Rudski et al., 2010)  between groups.Pearson's test or Spearman's test was applied for correlations.Level of significance was set to p < 0.05 (two-tailed).

The authors, affiliated to Linkoping University and Region
Ostergotland, confirm that the study was conducted in accordance with good research practice and performed in accordance with the Swedish law and the Declaration of Helsinki.

| RESULTS
During the recruitment period (April 2014-Feb 2020) 233 patients were screened for participation.Forty-nine patients met the inclusion criteria and none of the exclusion criteria and consented to participate in the study.Nineteen patients were assessed with TTE and CMR more than 1 week before surgery (8-70 days), due to prioritization of other patients.All patients had symptoms related to AS and none to RV dysfunction.Two-thirds of the studied cohort were men.No significant differences in baseline characteristics were observed between the groups except for smoking habits (p = 0.02) (Table 1).Survival rate was 100% 1 year after surgery.
Sixteen (33%) patients were found to have RV dysfunction defined as one or a combination of the following: TAPSE less than 17 mm, RVEF < 50% or RVFWS > −20%.RVFWS was reduced in 12 patients, of which 6 also had reduced RVEF or TAPSE.One patient had decreased RV function measured with all three parameters (Figure 1).RVD1 was smaller in the RV dysfunction group (p = 0.02).
One patient in the cohort had RVD1 > 41 mm.When dividing the patients in groups according to each RV impairment measurement, both low TAPSE and RVFWS were related to smaller RVD1, thus only the association with TAPSE was significant when indexed to BSA (Table 2).Three patients had FAC < 35% of which one also had  2 and Table S1).
Four out of six patients with LVEF < 50% had RV dysfunction.
When comparing TAPSE by TTE with TAPSE by CMR we found a significant correlation both for manual and semiautomatic analysis of CMR images (both r = 0.42 p = 0.003) (Table S2).

| DISCUSSION
RV dysfunction is a known risk factor for adverse outcome, both early and late after surgery in AS patients (Bohbot et al., 2020;Bootsma et al., 2018;Galli et al., 2015;Rigolli et al., 2019).Our objective was to describe RV function, studied by both CMR and echocardiography, since global RV function, obtainable by CMR, is difficult to assess by echocardiography, however is more accessible and routinely included in the diagnostics and follow-up of patients with aortic stenosis.
Patients were classified as having RV dysfunction if any of the three measurements (RVFWS, TAPSE, RVEF) were below the cut-off values described in the method.All three measurements were chosen since previous studies showed that RVEF by CMR, and TAPSE and RVFWS by echocardiography were individual predictors of prognosis in cardiac disease (Balderas-Muñoz et al., 2017;Bohbot et al., 2020;Galli et al., 2015;Morris et al., 2017;Park et al., 2014;Rigolli et al., 2019).
We described their covariation and found that all patients with RVEF < 50% also had changes in RVFWS and some also in TAPSE.
The main finding of this study was that in addition to these patients, 12 showed a reduction of either TAPSE, RVFWS or both, and that thus the echocardiographic measurements of longitudinal RV function seemed to be an earlier sign of beginning RV dysfunction.

| Methods for RV evaluation
In our study cohort the prevalence of RVEF < 50% was 8.3%, in agreement with previous findings (Rigolli et al., 2019).
RVEF from CMR detects another aspect of RV function than the longitudinal RV function by echocardiography.In the literature the effect of AS on LVEF appears later than LV global strain reduction (Miyazaki et al., 2011).We speculate that the RVEF may follow a similar pattern and therefore a smaller number of patients had reduced RVEF compared to the number of patients with changes in TAPSE or RVFWS from TTE.
There are different views on whether to include septal segments in the evaluation of RV strain as the septum is also part of the LV (Lang et al., 2015;Morris et al., 2017).Averages including septal segments could thereby dilute changes detected in the RV-free wall.
EF by CMR reflects both radial and longitudinal RV function.In our material we used three measurements for the evaluation of RV systolic function.As seen in Figure 1 some of the patients had RV impairment according to two or three measurements while others only had impairment in one of the parameters.Our material as well as the literature indicates that the use of RVFWS correlates better than TAPSE to RVEF (Figure 3 and Table S2) (Carluccio et al., 2018;Focardi et al., 2015;Guendouz et al., 2012).We chose to include all three parameters since we wanted to detect even subtle changes of known prognostic value and some of our patients only had reduced TAPSE.
When investigating FAC and RVD1 as alternative measures of RV function we could only find three patients with FAC < 35% and one of these patients had affected RVFWS.No association was found between FAC and RVEF or TAPSE in our material.The RVD1 was significantly smaller in the RV dysfunction group but when adjusting for body surface area there was no difference between F I G U R E 2 Mean indexed aortic valve area respective median left ventricular ejection fraction was found to be significantly different between the right ventricular function groups.
the groups except for patients with reduced TAPSE.The RVFWS analysis by CMR was at times unreliable due to suboptimal tracking.
Nowadays, AS patients planned for surgical AVR rarely have signs of RV failure due to the possibility to do transcatheter aortic valve implantation (TAVI).Therefore, it is important to demonstrate how some of the most common measures for RV evaluation vary within a cohort of SAVR patients.Further, it was observed that all patients with RVEF < 50% also had reduced RVFWS.
Even though our material is small it reflects the complexity of RV evaluation.CMR has been shown to be an accurate method for determining RVEF though it may be physically and psychologically challenging for the patients and less accessible, especially in smaller hospitals.FAC by echocardiography is less robust and changes probably appear later as more pronounced RV impairment develops, therefore FAC in our patient material was less frequently affected than the longitudinal measures TAPSE and RVFWS.Measurement of FAC by tracing of the entire RV might also be difficult in this patient group due to LV hypertrophy (Lang et al., 2015).

| RV volume vs compensation mechanisms
AVAi, LVEF, LVSVi, RVSVi and systolic blood pressure were significantly lower in the RV dysfunction group.In the natural course of the aortic stenosis disease, the LV remodelling induces increased LV filling pressure which transmits backwards and stimulates compensatory RV remodelling (Cavalcante et al., 2017).But only a small number of patients showed signs of elevated left ventricular end diastolic pressure at rest, with no difference between groups.In the low TAPSE group there was a difference in E/A ratio though no other signs of elevated filling pressure were present.MAPSE, although, was significantly reduced in patients with lower RVFWS indicating discrete LV impairment (Table S1).The lack of significant differences in LV global strain and MAPSE between the total RV impairment group and normal RV group may be a question of power since there was a tendency towards lower values in the RV dysfunction group.
However, other mechanisms such as biventricular dysfunction caused by ventricular interdependence as well as primary RV impairment may also be present.Another possible explanation for the small RV volumes could be attributed to a low LVSV caused by a narrower valve, resulting in a reduced venous return volume and consequently smaller RV volumes.This explanation gains plausibility considering the observed correlations between TAPSE and AVAi, as well as the correlation with LVSVi (Figure 3).

| Limitations
Our cohort is small and contains only patients with severe AS without concomitant heart disease.This highlights the effects of a pure AS, but findings may not be applicable to patients with additional heart conditions.We have not performed any measurements concerning loading conditions and venous return and whether these explains the results, the myocardial factor itself or a combination of both is not possible to determine from our study.There are only a few patients in each subgroup, when divided by the different measures for RV impairment, making these results more difficult to interpret.We regard this as an explorative study to generate hypotheses for future studies on a larger cohort of AS patients.Further, studies with followup data post-surgery would give more information about the predictive value of preoperative RV assessment.

| Summary
In patients with AS without concomitant cardiac disease, RV dysfunction, defined by RVEF by CMR, TAPSE or RVFW strain by TTE, was associated with a smaller aortic valve area and reduced stroke volumes.It is likely that the more pronounced the AS the higher the likelihood of RV dysfunction.We have demonstrated by using three parameters for RV evaluation that longitudinal RV function (TAPSE, RVFWS) by echocardiography frequently was impaired despite a normal RVEF.We found that all patients with reduced RVEF also had reduced RVFWS thereby the CMR examination did not identify additional patients with RV impairment.
RV dysfunction in AS is an important prognostic factor (Bootsma et al., 2018;Duncan et al., 2017;Galli et al., 2015;Musa et al., 2016;Rigolli et al., 2019).The medical community needs an increased awareness about the importance of RV assessment in patients with AS.Hence a detailed analysis of right ventricular function, primarily with TTE, should be included in the diagnostic imaging of severe aortic stenosis for risk evaluation.
RVFWS > −20%.Patients with RV dysfunction showed significantly smaller indexed AVA (AVAi) (p = 0.004) and lower left ventricular ejection fraction (LVEF) (p = 0.03) compared to patients with maintained RV function (Figure 2).In patients with reduced TAPSE the AVAi was significantly smaller.This was not found in the groups with reduced RVFWS or RVEF.RV and LV indexed stroke volumes (RVSVi/LVSVi) were significantly reduced in the RV dysfunction group, p = 0.01 respectively p < 0.001.Both low TAPSE and RVFWS was significantly associated with smaller RVSVi (p = 0.009, p = 0.03) and LVSVi volumes (p = 0.03, p = 0.002) (Tables2 and Table S1).Indexed LV and RV end-diastolic volumes (EDVi) were equal between the normal RV and RV impairment groups as well as the percentage of patients with signs of elevated LV filling pressures.Patients with reduced TAPSE had a smaller indexed RV end-diastolic volume (RVEDVi).For patients with reduced RVFWS we found a lower mitral annular plane systolic excursion (MAPSE) compared to patients with normal RVFWS (p = 0.02).(Table

F
I G U R E 3 Significant correlations between RV impairment measurements, aortic valve area and left ventricular and right ventricular volumes.AVAi, indexed aortic area; LVEF, left ventricular ejection fraction, LVSVI, left ventricular stroke volume index; RVEDVi, right ventricular end-diastolic volume index; RVEF, right ventricular ejection fraction; RVFWS, right ventricular free wall strain; RVSVI, right ventricular stroke volume index; TAPSE, tricuspid annular plane systolic excursion.
Note: No significant differences between characteristics were observed except for smoking habits.
Baseline measurements stratified by RV function.