Introduction
Severe aortic stenosis (AS) is a progressive valvular disease of ageing populations, often complicated by pulmonary hypertension (PH). The coexistence of PH in transcatheter aortic valve implantation (TAVI) candidates imposes additional haemodynamic burden, elevating right-ventricular (RV) load and impairing peri-procedural stability [1–3]. While TAVI has transformed AS management across risk strata, the specific mortality burden associated with PH in this population warrants further characterisation [4].
Mechanistically, PH often arises from chronic left-sided pressure overload, pulmonary venous hypertension and eventual pulmonary vascular remodelling [5–7]. These changes can lead to right-ventricular-pulmonary-arterial (RV–PA) uncoupling and diminished cardiac output reserve. Prior research has indicated that PH may be associated with increased frailty and mortality [8–10]. However, with ongoing technological and procedural refinements in TAVI, it is important to synthesise contemporary evidence regarding the observed mortality rates in patients with PH [11].
Previous reviews have been limited by heterogeneous PH definitions, mixed-risk populations and incomplete long-term data [12–15]. Therefore, we conducted a systematic review and descriptive meta-analysis to (i) quantify 1-year all-cause mortality among TAVI patients with PH, (ii) assess the influence of study quality on reported mortality rates and (iii) identify methodological and biological factors contributing to outcome variation. This study aims to provide a comprehensive overview of the current mortality landscape for TAVI patients with PH, highlighting areas for improved standardisation and future research.
Methods
1. Protocol and registration
This systematic review and meta-analysis followed the PRISMA 2020 guidelines and was prospectively registered in PROSPERO (CRD42024582949).
2. Search strategy
A comprehensive electronic search was performed in MEDLINE (Ovid), Embase (Ovid), Cochrane CENTRAL, Scopus and Web of Science from database inception to 18 September 2024, with no language restrictions. The strategy combined text words and MeSH terms related to ‘transcatheter aortic valve implantation’, ‘TAVI’, ‘pulmonary hypertension’ and ‘mortality’. Reference lists of eligible papers and relevant reviews were hand-searched to identify additional studies. Full search syntax is available in Supplementary Appendix 1.
3. Eligibility criteria
Studies were included if they (1) enrolled adults (≥18 years) with severe AS undergoing TAVI, (2) defined PH by echocardiography or right-heart catheterisation, (3) reported ≥1-year all-cause mortality or longer follow-up and (4) were primary observational cohorts with extractable outcome data. Systematic reviews were considered for qualitative synthesis to provide context on prior research but were not included in the quantitative meta-analysis.
Exclusion criteria include duplicate cohorts, non-TAVI populations, small series (<10 patients), abstracts without full text or studies lacking outcome reporting.
4. Study selection
Two independent reviewers screened titles/abstracts and full texts using Covidence. Conflicts were resolved by consensus or by a third reviewer. The database search identified 54 records. After removing duplicates (n = 4), 50 titles/abstracts were screened and 17 studies met the inclusion criteria (11 observational cohorts for quantitative synthesis and 6 systematic reviews for qualitative synthesis). No additional eligible studies were found through citation or grey-literature searches. The study selection process is summarised in the PRISMA 2020 flow diagram (Figure 1).
Figure 1
PRISMA diagram. TAVI, transcatheter aortic valve implantation.
5. Data extraction
Two reviewers independently extracted data using standardised Cochrane templates capturing study design, setting, sample size, mean age, PH definition and measurement method, follow-up duration and primary/secondary outcomes. Extracted data were cross-verified for accuracy before synthesis.
6. Quality and risk-of-bias assessment
Quality of observational cohorts was evaluated with the Newcastle–Ottawa Scale (NOS) (maximum 9 stars: 4 for Selection, 2 for Comparability, 3 for Outcome). Systematic reviews were appraised using the ROBIS tool across four domains (study eligibility, identification/selection, data collection, synthesis) for the qualitative context. Discrepancies were resolved by consensus. The detailed NOS scores for observational studies are presented in Table 1, and ROBIS assessments for included systematic reviews are summarised in Tables 2 and 3.
Table 1
NOS quality scores.
| Author, year | Selection (0–4) | Comparability (0–2) | Exposure/outcome (0–3) | Total score (max 9) |
|---|---|---|---|---|
| Ben-Dor, 2011 [1] | 3 – Clear TAVR cohort, well-defined PH and outcomes | 1 – Adjusted for major confounders (age, EF) | 2 – Outcome data from registry follow-up | 6/9 |
| Boxhammer, 2024 [11] | 4 – Comprehensive systematic selection per PRISMA | 2 – High comparability across included studies | 3 – Detailed exposure definitions and synthesis | 9/9 |
| Miyamoto, 2022 [10] | 3 – Clear inclusion (AS + TAVR + PH), defined outcomes | 2 – Multivariate analysis controlling for confounders | 3 – Prospective data with validated measures | 8/9 |
| Schewel, 2015 [4] | 3 – Consecutive TAVR cohort, PH defined echocardiographically and invasively | 2 – Groups balanced for baseline risk | 3 – Robust outcome assessment (RHC validation) | 8/9 |
| D’Ascenzo, 2015 [5] | 3 – Multi-centre registry, clear inclusion/exclusion | 2 – Adjusted for multiple covariates | 3 – Standardised data collection and outcomes | 8/9 |
| Testa, 2016 [6] | 3 – Defined PH subgroups in TAVR patients | 2 – Adjusted for key comorbidities | 2 – Outcome ascertainment moderate quality | 7/9 |
| Bishu, 2014 [7] | 3 – Consecutive TAVR cases, PH defined by ECHO | 1 – Partial adjustment for confounders | 2 – Retrospective but consistent outcomes | 6/9 |
| Luçon, 2014 [2] | 4 – Very large TAVR cohort (n > 2000), robust selection | 2 – Full multivariate modelling | 3 – Validated outcome measures (mortality) | 9/9 |
| Alushi, 2019 [8] | 3 – Prospective invasive cohort, clear PH definition | 2 – Adjusted for baseline characteristics | 3 – Comprehensive follow-up for outcomes | 8/9 |
| Lindman, 2015 [3] | 3 – PARTNER I registry, well-defined PH and outcomes | 2 – Propensity matching for confounders | 3 – Robust outcome assessment | 8/9 |
| Rodés-Cabau, 2010 [9] | 3 – Canadian multicentre, clear inclusion/exclusion | 2 – Adjusted for baseline variables | 2 – Long-term follow-up for mortality | 7/9 |
Table 2
ROBIS quality assessment (systematic reviews)
| Author, year | Eligibility criteria | Search and selection | Data handling | Synthesis and reporting | Overall judgement |
|---|---|---|---|---|---|
| Khalil, 2024 [12] | Adults with severe AS undergoing TAVR; PH status and mortality outcomes clearly defined | PRISMA-based, multiple databases (PubMed, Embase, Cochrane, WoS), dual screening | Dual extraction, NOS used | Random-effects meta-analysis; sensitivity and subgroup analyses; transparent reporting | Low risk/high quality |
| Tang, 2017 [13] | Clear inclusion: TAVR, PH-assessed, mortality outcomes reported | PRISMA-guided, 2 independent reviewers, comprehensive database search | Data extracted systematically; NOS used | Random-effects model, meta-regression, publication bias (EGGER/ BEGG). | Low risk/high quality |
| Desai, 2023 [14] | Clear inclusion: TAVR + PH studies ≥100 pts. excluded poor-quality studies | PRISMA-based multi-database search, dual screening | Structured data extraction and NOS quality grading | Narrative synthesis (heterogeneity prevented pooling) | Moderate quality/some concerns |
| Meybodi, 2024 [15] | Included TAVR studies with PH data and outcomes; excluded duplicates/overlaps | Comprehensive multi-database search (PubMed, Embase, Scopus, Cochrane, ProQuest), predefined strategy | Extracted ORS, meta-regression to test covariates; publication bias tested (EGGER, Begg) | Pooled 33 studies, random-effects model, heterogeneity and sensitivity analysis described | Low risk/high quality |
| Boxhammer, 2022 | Severe AS + PH, focus on non-invasive diagnostics; English-only inclusion | Systematic PubMed Central® search with explicit search terms and PRISMA flow | Abstract + full-text screening, duplicate removal; descriptive data extraction | Narrative synthesis across 39 studies (echo, CT/MRI, biomarkers) | Moderate quality/some concerns (narrative only) |
| Kokkinidis, 2018 | Severe as undergoing TAVR; studies reporting PH and mortality | Systematic search; inclusion criteria clear | Data pooled appropriately, risk of bias discussed | Random-effects meta-analysis, sensitivity/ subgroup analyses | Moderate-high quality |
Table 3
ROBIS summary (low-to-moderate overall risk)
| Author, year | Type | Inclusion criteria | No. of studies | PH definition | Main findings | Quality tool | Overall judgement |
|---|---|---|---|---|---|---|---|
| Khalil, 2024 [12] | Systematic review + meta-analysis | Severe AS + TAVI, PH & mortality reported | 15 | Echo or RHC | Pooled PH ↑ mortality | ROBIS | Low risk/high quality |
| Tang, 2017 [13] | Systematic review + meta-analysis | TAVI, PH assessed, mortality outcomes | 11 | Per study | PH increased mortality | ROBIS | Low risk/high quality |
| Desai, 2023 [14] | Systematic review | ≥100 pts, TAVI + PH | 10 | Various | Narrative synthesis only | ROBIS | Moderate/some concerns |
| Meybodi, 2024 [15] | Systematic review + meta-analysis | TAVI + PH data | 33 | Various | Pooled analysis confirms risk | ROBIS | Low risk/high quality |
| Boxhammer, 2022 | Systematic review (diagnostic) | Severe AS + PH (non-invasive) | 39 | Echo, CT, MRI | PH diagnosis methods heterogeneous | ROBIS | Moderate/some concerns |
| Kokkinidis, 2018 | Systematic review + meta-analysis | Severe as undergoing TAVI | 20 | Echo-based | PH associated with mortality | ROBIS | Moderate–high quality |
7. Statistical analysis
Study-level event proportions for 1-year all-cause mortality were logit-transformed and pooled using a DerSimonian–Laird random effects model. Results are expressed as pooled proportions with 95% confidence intervals (CIs). Heterogeneity was quantified using Cochran Q and I2 statistics (low < 25%, moderate 25–75%, high > 75%). Subgroup analyses were prespecified by study quality (NOS ≥ 8 vs 6–7) and PH definition (echocardiography vs right-heart catheterisation). Sensitivity analyses used leave-one-out resampling. Publication bias was explored using funnel plots and Egger’s regression test (if ≥10 studies). All analyses were conducted in RevMan 5.3 and R (meta package v6.3-1).
Results
1. Study characteristics
Eleven observational cohorts (N = 10,665 patients) were included in the quantitative meta-analysis. Six systematic reviews were included for qualitative synthesis and contextualisation. Cohort sizes ranged from 277 to 2435 patients, with mean ages between 79 years and 85 years. Most studies defined PH as sPAP ≥ 40 mm Hg by echocardiography, while others used mPAP ≥ 25 mm Hg via right-heart catheterisation. Follow-up durations ranged from 12 months to 36 months. Baseline characteristics and outcome data for all included observational cohorts are shown in Table 4.
Table 4
Characteristics of observational studies.
| Author, year | Population/design | N | PH definition | Mean age (years) | Ejection fraction (%) | 1-year mortality (%) | Key findings |
|---|---|---|---|---|---|---|---|
| Ben-Dor, 2011 [1] | Single-centre TAVI registry | 509 | sPAP ≥ 40 mmHg (echo) | 81.7 | 47.5 | 44.0 | PH independently predicted mortality |
| Boxhammer, 2024 [11] | Single-centre, sex-stratified | 303 | sPAP threshold per study | 82.6 | 55 | 17.2 | Higher risk signal; sex interaction |
| Miyamoto, 2022 [10] | OCEAN-TAVI (multicentre) | 1,872 | sPAP > 36 mmHg (echo) | 84.7 | 58.5 | 15.9 | Residual/new-onset PH ↑ mortality |
| Schewel, 2015 [4] | Single-centre (Hamburg) | 559 | mPAP ≥ 25 mmHg (RHC) | 79.8 | 49.5 | 23.4 | Invasive PH remained an independent predictor |
| D’Ascenzo, 2015 [5] | Multicentre (Italy) | 674 | sPAP > 40 mmHg (echo) | 81.5 ± 5.6 | 51 ± 11.6 | 31.0 | Persistent PH strongest predictor |
| Testa, 2016 [6] | CoreValve multicentre registry | 900 | sPAP > 60 mmHg | 79.5 | 50 | 11.6 | Severe PH predicts 1-year mortality |
| Bishu, 2014 [7] | Single-centre (Mayo Clinic) | 277 | PASP tertiles (≥49 mmHg) | 81 ± 8 | 53 ± 15 | 35.0 | Highest tertile → highest mortality |
| Luçon, 2014 [2] | FRANCE-2 (national) | 2435 | sPAP ≥ 40/ ≥ 60 mmHg | 83 ± 7 | 49 ± 15 | 28.0 | PH independently predicted mortality |
| Alushi, 2019 [8] | Prospective invasive cohort | 617 | Echo/RHC PH | 80 ± 8 | 56 ± 4 | 27.0 | Invasive PH and RV parameters predictive |
| Lindman, 2015 [3] | PARTNER I registry | 2180 | mPAP ≥ 25 (mod/sev ≥ 35) | 83 ± 8 | 55 | 25.0 | PH predicted mortality – - esp. in women |
| Rodés-Cabau, 2010 [9] | Canadian multicentre | 339 | sPAP > 60 mmHg | 81 ± 8 | 55 ± 14 | 22.1 | PH associated with late mortality |
2. Quantitative data synthesis (meta-analysis)
Across the 11 observational cohorts, the pooled 1-year allcause mortality was 24.5% (95% CI 20.0%-29.6%; I2 = 96.7%). High-quality studies (NOS ≥ 8; n = 7, N = 8640) reported a pooled 1-year mortality of 23.7% (95% CI 19.7%-28.1%), while moderatequality studies (NOS 6–7; n = 4, N = 2025) showed 26.2% (95% CI 13.2%–45.3%). The pooled 1-year mortality across all cohorts is illustrated in the forest plot in Figure 2. Subgroup analyses stratified by study quality are shown in Figure 3 (high-quality studies) and Figure 4 (moderate-quality studies). Due to significant heterogeneity in reporting and definitions across the included studies, a robust quantitative subgroup analysis based on specific PH definitions (e.g., echocardiographic vs right-heart catheterisation) or severity was not feasible within the scope of this meta-analysis.
Figure 2
Overall 1-year mortality forest plot. CI, confidence interval.
Figure 3
High-quality studies (NOS ≥ 8). CI, confidence interval; NOS, Newcastle–Ottawa Scale.
Figure 4
Moderate-quality studies (NOS 6–7). CI, confidence interval; NOS, Newcastle–Ottawa Scale.
3. Qualitative synthesis of systematic reviews
Six systematic reviews were included to provide a broader context of the existing literature on PH and TAVI outcomes. These reviews consistently reported an association between PH and adverse outcomes after TAVI, highlighting the complexity and variability in PH definitions and prognostic implications. They also underscored the need for standardised assessment and reporting. Details of these systematic reviews and their quality assessments are provided in Tables 2 and 3.
Discussion
This systematic review and descriptive meta-analysis provide a comprehensive synthesis of 1-year all-cause mortality rates among patients with PH undergoing TAVI. Drawing on data from 11 observational cohorts comprising over 10,000 patients, our findings indicate a substantial pooled 1-year mortality rate of 24.5% in this specific patient population. This underscores that patients with PH continue to experience high mortality rates following TAVI, even amidst advancements in procedural technology and patient selection. The consistency of these rates across high- and moderate-quality studies (23.7% vs 26.2%) suggests a persistent clinical challenge (see Figures 2–4).
1. Mechanistic and pathophysiological insights from literature
The observed high mortality rates in patients with PH undergoing TAVI are likely influenced by a complex interplay of factors, as highlighted by existing literature. Chronic left-sided pressure overload in AS can lead to post-capillary PH, which may progress to combined pre- and post-capillary disease involving increased pulmonary vascular resistance (PVR) and pulmonary vascular remodelling [5–7]. These changes can impair RV–pulmonary arterial (PA) coupling, limiting the RV’s ability to adapt to haemodynamic shifts during and after TAVI. Individual studies, such as those by Alushi et al. [8] and Schewel et al. [4], have demonstrated that invasively measured PH, particularly when associated with elevated PVR or reduced RV function, is strongly associated with mortality. Similarly, Lindman et al. [3] and Ben-Dor et al. [1] observed that PH severity correlates with 1-year mortality, even after adjusting for comorbidities and baseline left ventricular function in their respective cohorts. Furthermore, the persistence or progression of PH post-procedure, rather than its absolute baseline value, has been suggested by some studies (e.g., D’Ascenzo et al. [5], Testa et al. [6] and Luçon et al. [2]) to be a powerful determinant of long-term outcomes. These insights, derived from individual study analyses, provide a pathophysiological context for the high mortality rates observed in our descriptive meta-analysis.
2. Determinants of heterogeneity
A prominent feature of the current synthesis is the substantial heterogeneity observed across studies (I2 ≈ 96.7%). This high degree of heterogeneity is not purely statistical but reflects genuine clinical and methodological diversity among the included observational cohorts. Key factors contributing to this variability include the definition and quantification of PH, which varied markedly – ranging from echocardiographic systolic pulmonary artery pressure (sPAP ≥ 40 mmHg) to invasive mean pulmonary artery pressure (mPAP ≥ 25 mmHg). Echocardiographic estimations are known to have variability and can sometimes overstate prevalence, whereas invasive right-heart catheterisation, while offering greater precision, was used in only a subset of cohorts. The inconsistent application of PH severity classifications (mild, moderate or severe) and the distinction between pre- and post-capillary phenotypes further contribute to this heterogeneity. Additionally, differences in device generation, operator experience and evolving procedural strategies (e.g., transfemoral vs transapical approaches) across the study periods may also play a role. Despite this variability, the consistent finding of high 1-year mortality rates across studies, even in contemporary series such as Boxhammer [11] and Miyamoto [10], highlights the enduring clinical significance of PH in this population.
3. Comparison with previous reviews
Earlier systematic reviews, including those by Tang [13], Kokkinidis (2018) and Desai [14], have consistently reported an association between PH and mortality after TAVI [16]. More recent meta-analyses by Khalil [12] and Meybodi [15] have further confirmed this association. Our descriptive meta-analysis contributes to this body of evidence by specifically quantifying the pooled 1-year all-cause mortality rate in TAVI patients with PH from primary observational cohorts and stratifying these results by study quality. While previous reviews often mixed different endpoints or included systematic reviews in their quantitative synthesis, our approach provides a focused descriptive estimate of mortality from primary data. The stratification by study quality (NOS ≥ 8 vs 6–7) yielded remarkably consistent mortality estimates (23.7% vs 26.2%), suggesting that the observed high mortality is robust across studies of varying methodological rigour. This synthesis, by focusing on primary observational data for quantitative pooling and using systematic reviews for qualitative context, offers a refined perspective on the mortality burden.
4. Clinical implications
The consistent finding of high 1-year mortality rates in TAVI patients with PH has direct implications for clinical practice. While PH should not necessarily contraindicate TAVI, its presence signals the need for enhanced peri-procedural vigilance and post-procedural surveillance. Comprehensive pre-TAVI right-heart evaluation, including assessment of RV systolic function, tricuspid regurgitation severity and PVR, should be routinely incorporated for candidates with suspected PH. Optimisation of volume status and afterload reduction prior to TAVI may improve haemodynamic stability, particularly in patients with combined post-capillary or mixed PH. Post-procedural reassessment of pulmonary pressures and RV function can help identify patients with persistent PH who may benefit from closer follow-up or consideration of adjunctive pharmacotherapy targeting pulmonary vasculature, as suggested by individual studies [17,18].
Future research should aim to standardise PH assessment and classification in TAVI populations. Integrating echocardiographic, invasive and biomarker-based measures into a unified framework could refine prognostic modelling. Prospective, multicentre studies with standardised right-heart catheterisation and longitudinal follow-up are essential to delineate reversible vs fixed components of PH. Moreover, clinical trials exploring pulmonary vasodilator or RV-targeted therapies – particularly in patients with mixed or residual PH – may open new avenues for improving outcomes beyond mechanical valve replacement.
Conclusion
Patients with PH undergoing TAVI exhibit a high 1-year all cause mortality rate, approaching 25%. This descriptive meta-analysis highlights that despite technological progress and procedural refinements, PH continues to represent a significant clinical challenge in this population. The substantial heterogeneity observed across studies underscores the need for standardised PH phenotyping and the integration of comprehensive right-heart functional assessment into pre-TAVI evaluation. Such measures are essential to improve risk stratification, optimise patient selection and guide post-procedural management, ultimately aiming to improve outcomes for these vulnerable patients.
Notes
[5] Conflicts of interest Conflicts of interest
None declared.
[6] Registration
PROSPERO CRD42024582949.
[7] Data availability
All data derived from published sources.
[8] Ethics
Not applicable (systematic review of published studies).
[9] Declaration of interest, funding and ethical statement
The authors declare that they have no conflicts of interest related to this work. No funding sources were received for the preparation, analysis or publication of this systematic review. This study is a systematic review and meta-analysis of data extracted exclusively from previously published studies. Therefore, ethical committee approval and patient consent were not required, as no new human participants or identifiable data were involved. All authors have reviewed and approved the final version of the manuscript and agree to be accountable for all aspects of the work.
[10] Central message
This is the largest systematic review evaluating cardiac CT for the detection of pulmonary hypertension, including 43 studies and ~40,000 patients.
[11] Perspective message
The review highlights critical methodological gaps and provides clear recommendations for future research, including standardisation of CT thresholds, prospective validation and uniform reporting.