Article Text

Download PDFPDF

Original research article
Outcomes of persistent pulmonary hypertension following transcatheter aortic valve replacement
  1. Ahmad Masri,
  2. Islam Abdelkarim,
  3. Michael S Sharbaugh,
  4. Andrew D Althouse,
  5. Jeffrey Xu,
  6. Wei Han,
  7. Stephen Y Chan,
  8. William E Katz,
  9. Frederick W Crock,
  10. Matthew E Harinstein,
  11. Dustin E Kliner,
  12. Forozan Navid,
  13. Joon S Lee,
  14. Thomas G Gleason,
  15. John T Schindler,
  16. João L Cavalcante
  1. Department of Medicine, Division of Cardiovascular Diseases, University of Pittsburgh, UPMC-Heart and Vascular Institute, Pittsburgh, Pennsylvania, USA
  1. Correspondence to Dr. João L Cavalcante, Heart & Vascular Institute - UPMC, University of Pittsburgh, 200 Lothrop Street, Scaife Hall, S-558 Pittsburgh, PA 15213, USA; cavalcantejl{at}


Objectives To determine the prevalence and factors associated with persistent pulmonary hypertension (PH) following transcatheter aortic valve replacement (TAVR) and its relationship with long-term mortality.

Methods Consecutive patients who underwent TAVR from July 2011 through January 2016 were studied. The prevalence of baseline PH (mean pulmonary artery pressure ≥25 mm Hg on right heart catheterisation) and the prevalence and the predictors of persistent≥moderate PH (pulmonary artery systolic pressure (PASP)>45 mm Hg on 1 month post-TAVR transthoracic Doppler echocardiography) were collected. Cox models quantified the effect of persistent PH on subsequent mortality while adjusting for confounders.

Results Of the 407 TAVR patients, 273 (67%) had PH at baseline. Of these, 102 (25%) had persistent≥moderate PH. Mortality at 2 years in patients with no baseline PH versus those with PH improvement (follow-up PASP≤45 mm Hg) versus those with persistent≥moderate PH was 15.4%, 16.6% and 31.3%, respectively (p=0.049). After adjusting for Society of Thoracic Surgeons Predicted Risk of Mortality and baseline right ventricular function (using tricuspid annular plane systolic excursion), persistent≥moderate PH remained associated with all-cause mortality (HR=1.82, 95% CI 1.06 to 3.12, p=0.03). Baseline characteristics associated with increased likelihood of persistent≥moderate PH were ≥moderate tricuspid regurgitation, ≥moderate mitral regurgitation, atrial fibrillation/flutter, early (E) to late (A) ventricular filling velocities (E/A ratio) and left atrial volume index.

Conclusions Persistency of even moderate or greater PH at 1 month post-TAVR is common and associated with higher all-cause mortality.

  • pulmonary vascular disease
  • transcatheter valve interventions
  • aortic stenosis

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Pulmonary hypertension (PH) is common in patients with severe aortic stenosis (AS) undergoing transcatheter aortic valve replacement (TAVR) and associated with increased mortality post-TAVR.1-7 Despite its prognostic role, it remains a missing component of the traditional Society of Thoracic Surgeons risk score assessment. Although invasive right heart catheterisation (RHC) is the gold standard method for PH evaluation, pre-TAVR PH screening is recommended to be performed non-invasively through transthoracic Doppler echocardiography (TTE) with the estimation of pulmonary artery systolic pressure (PASP) using simplified Bernoulli equation.8 9

Studies have suggested that only severe post-TAVR PH is a predictor of mortality on a relatively short follow-up time.1 4 7 However, PH is a heterogeneous entity in which outcomes can be influenced by the presence of right ventricular dysfunction and other comorbidities. Importantly, most studies have used TTE to diagnose PH, and have not performed comprehensive baseline right-sided assessment with adjustment of these parameters into the measured outcome.4 7 10 Therefore, we sought to understand the prevalence of persistent moderate or greater PH post-TAVR and its effect on mortality in a contemporary cohort of TAVR patients who underwent comprehensive clinical, echocardiographic and invasive haemodynamic evaluation.


Study design

We conducted a retrospective cohort analysis of patients who underwent TAVR at the University of Pittsburgh Medical Center, a large tertiary care centre, between July 2011 and January 2016. Because the purpose was to evaluate the clinical impact of persistent PH post-TAVR, we excluded patients who: (A) had undergone valve-in-valve procedures, (B) lacked baseline RHC data, (C) died prior to post-TAVR 1-month echocardiographic assessment, (D) had uninterpretable tricuspid regurgitation (TR) jet signal to estimate PASP at 1 month post-TAVR, or (E) lacked 1-month follow-up TTE.

Patients underwent comprehensive clinical evaluation by the designated Heart Team and deemed appropriate to undergo TAVR as suggested by the guidelines.9 11 Clinical, laboratorial and procedural data were collected. Baseline PH was defined by clinical RHC pre-TAVR as a mean pulmonary artery pressure (mPAP)≥25 mm Hg according to guidelines.12 Up to 505 patients underwent TAVR during the study period. Persistent moderate PH or greater post-TAVR was defined as PASP>45 mm Hg on Doppler echocardiography using modified Bernoulli equation. Patients were clinically followed up in our programme at 1, 6 and 12 months post-TAVR, and annually thereafter. Primary outcome was all-cause mortality, which was obtained through querying the social security death index (using the updated Social Security Administration Death Master file, for which our healthcare system is exempt from the 3-year delay period by the Social Security Administration).

Study procedures (echocardiogram, RHC and TAVR procedure)

All included patients undergoing TAVR evaluation underwent comprehensive TTE using commercial systems (Philips Medical Systems, Bothell, WA; Siemens Medical Solution, Malvern, PA; or General Electric, Milwaukee, WI). Both baseline and 1-month follow-up TTE studies were systematically independently reviewed as per guidelines13–15 for collection of several imaging parameters. All included patients underwent RHC as part of their clinical care prior to TAVR. Fluid-filled catheters that were connected to pressure transducers were used to measure pressures inside the right atrium, right ventricle (RV), pulmonary artery and pulmonary arterial wedge pressure (PAWP). Lastly, all patients underwent TAVR procedure receiving either balloon-expandable or self-expanding bioprosthesis. Study procedural details are available in online supplementary file 1.

Statistical analysis

Categorical variables are presented as frequency (percentage); continuous variables are presented as mean±SD. For categorical variables, groups are compared using the Χ2 test, except in the case of selected variables with less than five patients in a single cell, where Fisher’s exact test was used. For continuous variables, analysis of variance was used to test for differences in the baseline characteristics across all three groups (no baseline PH vs improved PH vs residual PH) while two-sample t-tests were used to test pairwise differences between the two groups that had PH before TAVR (ie, improved PH vs residual PH). Factors associated with residual moderate PH or greater post-TAVR were assessed using univariable and multivariable logistic regression analyses, adjusting for the baseline mPAP. Survival after TAVR is presented using Kaplan-Meier curves; log-rank tests were used to compare unadjusted survival times between groups. Cox proportional hazards models were used to quantify the risk for patients with residual PH versus those in whom PH resolved while adjusting for Society of Thoracic Surgeons Predicted Risk of Mortality (STS-PROM) (as a proxy for overall patient’s burden of comorbidities, except for PH) and tricuspid annular plane systolic excursion (TAPSE), an established measure of RV systolic function. The proportional hazards assumption was tested by looking at time-dependent covariates (time*covariate interaction variables) to determine whether the main effects varied over time. Statistical analyses were performed using SAS V.9.4 (SAS Institute).


Of the 505 patients identified who underwent TAVR during the study period, 407 met the inclusion criteria. Figure 1 displays the study workflow. Of those, 273 (67%) had baseline PH (mPAP≥25 mm Hg). Patients with baseline PH were further subdivided, based on the 1-month post-TAVR follow-up TTE, into persistent PH group (102 patients or 25%, if PASP>45 mm Hg) and improved PH group (171 patients or 42%, follow-up PASP≤45 mm Hg).

Figure 1

Study population and workflow. mPAP, right heart catheterisation-derived mean pulmonary artery pressure; N, number; PASP, echo-derived pulmonary artery systolic pressure; PH, pulmonary hypertension; RHC, right heart catheterisation; TAVR, transcatheter aortic valve replacement.

Baseline clinical characteristics are shown in table 1. Median STS-PROM score was higher in the residual PH versus improved PH group (9.9% vs 8.2%, p=0.004). Preoperative atrial fibrillation was also more common in patients with residual PH than in those with improved PH (62.7% vs 46.2%, p=0.013).

Table 1

Baseline characteristics

Table 2

Baseline Echocardiographic measurements


Baseline echocardiographic measurements are shown in table 2. Median follow-up echo was obtained 34 days (IQR 29–36 days) post-TAVR procedure. Baseline moderate to severe mitral regurgitation (MR), moderate to severe TR, higher mean PASP, higher early (E) to late (A) ventricular filling velocities (E/A ratio) and larger mean left atrial volume index (LAVI) were seen in patients with residual PH.

Invasive haemodynamics

Baseline RHC data are shown in table 3. Patients with residual PH had higher mPAP (38.9±9.5 vs 34.6±8.4 mm Hg; p<0.001), higher pulmonary vascular resistance (PVR, 3.6±2.3 vs 2.7±2.1 Wood units; p=0.004) and higher transpulmonary gradient (TPG, 15.4±8.6 vs 12.5±8.3 mm Hg; p=0.008). There was no statistically significant difference in diastolic pressure gradient (DPG). In patients with persistent PH, only 13% had a DPG≥7 mm Hg, while 54% had PVR>3 Wood units and 57% had TPG≥12 mm Hg.

Table 3

Baseline right heart catheterization measurements

TAVR procedure

Procedural characteristics are shown in online supplementary table 1. The majority of patients underwent transfemoral approach (74.2%), followed by left subclavian approach (13%). There were no statistically significant differences between all three groups for their respective procedural characteristics.

Supplementary file 1



Over a mean follow-up of 20±13 months, there were 82 deaths. The 2-year all-cause mortality in patients with no PH, improved PH and residual PH was 15.4%, 16.6% and 31.3% (p=0.049), respectively (figure 2). There was no statistically significant difference in mortality between patients with no PH at baseline and those with PH at baseline as defined by RHC (see online supplementary figure S1). When patients with PH at baseline were divided based on post-TAVR echo-derived PASP, there was no statistically significant difference in mortality between patients with no PH at baseline and those with improved PH; p=0.557. On the other hand, residual≥moderate PH as compared with improved PH was associated with an increased risk of all-cause mortality with an unadjusted HR of 1.89 (95% CI 1.12, 3.19, p=0.018). Multivariable analysis showed that persistent PH was independently associated with all-cause mortality (HR=1.82, 95% CI 1.06 to 3.12, p=0.03) even after adjustment for the comprehensive STS-PROM risk score and RV TAPSE.

Figure 2

Survival after transcatheter aortic valve replacement (TAVR) in the study population (starting at 30 days post-TAVR).

Pre-TAVR characteristics associated with persistent PH in unadjusted univariate analysis as well as adjusted for baseline mPAP are shown in table 4. After adjusting for mPAP, baseline≥moderate TR, ≥moderate MR, atrial fibrillation or atrial flutter, E/A ratio and LAVI were associated with persistent PH.

Table 4

Relationship between preoperative factors and residual PH

We performed a secondary analysis where pre-TAVR echo-derived PASP>45 mm Hg was used as the entry criteria to diagnose baseline PH rather than invasive mPAP≥25 mm Hg by RHC. Despite discordance between echo and RHC in the quantification of PH severity,16 survival curves were similar (see online supplementary table). When patient outcomes were analysed based on discordance of pre-TAVR PH diagnosis by echo and RHC, there was no statistically significant survival difference in patients with baseline echo-derived PASP≤45 mm Hg, irrespective of invasively measured mPAP (see online supplementary figure S3). Furthermore, regardless of the PH haemodynamic subtype, all-cause mortality was higher in those with persistent PH versus improved (see online supplementary figures S4 and S5, panels A–C).


In this large single-centre study of 407 consecutive patients undergoing TAVR, persistent moderate or greater PH at 1 month post-TAVR is common (25% of overall TAVR cohort or 37.4% of those with baseline PH) and associated with a high 2-year all-cause mortality (31.3%). The presence of residual PH (PASP by TTE>45 mm Hg) was associated with a 1.8-fold increased risk of mortality, even after adjustment for the comprehensive STS-PROM score as well as an index of RV systolic function (TAPSE), suggesting that persistent PH is an independent contributor to mortality in this population. Importantly, post-TAVR PH improvement at 1 month is associated with similar 2-year all-cause mortality as patients with no baseline PH (16.6% vs 15.4%, p=0.557). Therefore, not all patients with baseline PH are at risk of poor outcomes post-TAVR. The risk appears to be increased to patients when even≥moderate PH remains present at 1 month postprocedure. Thus, identification of both baseline and post-TAVR PH is of paramount importance.

Baseline PH has been shown in many studies to be associated with mortality and worse outcomes in patients undergoing TAVR.1 4–7 17 To our knowledge, only two other studies have evaluated the short-term implications of post-TAVR persistency of PH into outcomes.4 7 Testa et al reported that post-TAVR patients with persistent severe PH (PASP>60 mm Hg at 1 month post-TAVR) had a 2.4-fold increase at 1-year mortality.4 Sinning et al demonstrated that patients with persistent severe PH (PASP>60 mm Hg, 90 days post-TAVR) had higher 2-year mortality than patients with improved PH (50% vs 18.6%, p=0.001).7 The important limitations in both studies included: (1) gold standard RHC was not used to diagnose baseline PH and (2) comprehensive right-sided assessment evaluating RV systolic function was not performed pre-TAVR and/or adjusted into a multivariate model.

RV dysfunction and eventually RV failure are associated with decreased survival in PH.17 18 However, in neither of the prior published studies was RV function and/or TR integrated into the multivariate model4 and/or measured.7 Our findings demonstrated that these worse outcomes persisted even with adjusting for comprehensive STS-PROM score and RV function measured by TAPSE.

RHC prior to TAVR allows accurate PH diagnosis, prognostication and haemodynamic PH classification into isolated precapillary, isolated postcapillary or combined pre-post capillary PH, according to the guidelines.12 19 In our study, patients with residual PH had higher baseline mPAP, TPG and PVR, with similar PAWP, signalling a higher prevalence of the combined pre-post capillary PH phenotype. While our cohort is underpowered to detect statistically significant differences based on PH subtype, trends for worse outcomes in patients with persistent PH post-TAVR were consistent in all PH subtypes.

In our study, baseline characteristics associated with increased likelihood of persistent≥moderate PH were ≥moderate TR, ≥moderate MR, atrial fibrillation/flutter, E/A ratio and LAVI. These findings suggest worse haemodynamic profile with higher burden of concomitant valvular heart disease and worse diastolic function. In a prior study by Testa et al, left ventricular ejection fraction <30%, atrial fibrillation and an organic cause of MR were associated with persistent PH.4 However, Testa et al did not report the prevalence of TR, hence an important limitation of their analysis. Schwartz et al have recently shown in a large single-centre study that patients with baseline≥moderate TR had a higher 6-month PASP (49.4±19 vs 39.8±13 mm Hg) supporting our observations.18

Overall, our study expands the findings by Testa et al and Sinning et al and provides crucial proof of the importance of continued PH assessment by showing that even moderate or greater PH at 1 month post-TAVR is strongly associated with worse survival, independent of baseline RV function. Thus, these considerations should be incorporated into management decisions made jointly with the patient and clinical team. With recent advances in the medical therapeutic options for patients with PH, further studies are needed to determine whether baseline and persistent PH in patient with severe AS can be modified by further therapeutic strategies, beyond TAVR.


First, although this represents a large contemporary TAVR cohort from a single centre, this remains an observational study with the limitations that accompany such design, including the exclusion of 75 patients due to missing data (52 patients without baseline RHC data, and 23 patients without follow-up echo at 1 month). However, all included patients underwent a uniform protocolled comprehensive evaluation by the Heart Team prior to TAVR, and all echocardiograms were systematically and independently reviewed. Second, we used RHC to define pre-TAVR PH, but used PASP by TTE to define persistent PH post-TAVR. While RHC remains the most accurate method to diagnose PH, it is not common practice to be performed in the post-TAVR setting. The currently accepted practice of using echocardiography for clinical decisions pre and post-TAVR, and the role of RHC in the care of patients undergoing TAVR need to be further defined particularly for those patients with persistent PH. Lastly, the clinical implications of persistent PH into other important post-TAVR metrics such as quality of life, functional capacity and hospital readmissions need to be equally evaluated in future studies.


Longitudinal evaluation of PH severity is necessary in TAVR patients. Persistent moderate or greater PH at 1 month post-TAVR is common and strongly associated with all-cause mortality. PH improvement post-TAVR is associated with an improved survival similar to patients without baseline PH. Whether persistent PH should be a modifiable target for future therapies beyond TAVR requires further investigation.

Key messages

What is already known on this subject?

Severe pulmonary hypertension (PH) at baseline is associated with worse outcomes after transcatheter aortic valve replacement (TAVR).

What might this study add?

Persistent moderate to severe PH at 1 month is common and is associated with increased mortality, independent of baseline right ventricular function.

How might this impact on clinical practice?

  • Persistent moderate to severe PH at 1 month is common and is associated with increased mortality, independent of baseline right ventricular function.

  • Evaluation of PH at post-TAVR setting is necessary as it might be a target to further improve TAVR outcomes.



  • Contributors AM, IA, JSL, TGG, JTS and JLC: designed the study, data extraction, wrote and revised the manuscript. JX and WH: data extraction. MSS and ADA: designed the study, performed the analysis, wrote and revised the manuscript. SYC, WEK, FWC, MEH, DEK and FN: designed the study, reviewed the manuscript.

  • Funding This work was supported by the National Institutes of Health (HL096834, HL124021 to SYC) and the American Heart Association (14GRNT19600012 to SYC).

  • Competing interests JLC and TGG received investigator-initiated grant support from Medtronic. None of the other coauthors have any conflict of interest relevant to the content of this manuscript.

  • Ethics approval IRB-University of Pittsburgh.

  • Provenance and peer review Not commissioned; externally peer reviewed.