Objective Residual aortic regurgitation (AR) after transcatheter aortic valve implantation (TAVI) has been associated with increased mortality. Nevertheless, a considerable proportion of these patients survives and appears to tolerate AR. Identification of patients at higher risk of death may assist in tailoring therapy, but predictors of mortality in this subset of patients is largely unknown.
Methods A total of 1432 patients were included in the German TAVI registry. One-year follow-up data were available for 1318 patients (92%). Of the latter, 201 patients (15.2%) had more-than-mild AR as evaluated by angiography and represent the population of the current analysis. Among these patients, baseline demographic, clinical, echocardiographic and angiographic characteristics were compared among survivors and non-survivors to identify factors associated with mortality at 1 year.
Results Mean age was 81.2±6.6 years and men represented 55%. The mean logistic EuroSCORE was 22±15%. Overall, 92% of patients received the Medtronic CoreValve and 8% received the Edwards Sapien valve. At 1 year, 61 patients (31%) with more-than-mild post-TAVI AR had died. Compared with patients who survived, patients who died had more commonly coronary artery disease, peripheral arterial disease and chronic renal impairment. Patients who died had a lower baseline LVEF (44±18% vs 52±16%, p=0.002), higher prevalence of more-than-mild (≥2+) mitral regurgitation (44% vs 27%, p=0.001), and a higher systolic pulmonary artery pressure (51±18 mm Hg vs 44±19 mm Hg, p=0.002), but the severity of aortic stenosis was similar, and the prevalence and severity of pre-TAVI AR was comparable (any AR in 88% vs 83%, respectively, p=0.29). Using Cox regression analysis, only baseline mitral regurgitation ≥2+ (HR 1.77, 95% CI 1.05 to 2.99, p=0.03) and systolic pulmonary artery pressure (HR 1.15, 95% CI 1.01 to 1.33, p=0.04) were independently associated with 1-year mortality, while female gender was protective (HR 0.54, 95% CI 0.30 to 0.96, p=0.03).
Conclusions We identified preprocedural characteristics associated with 1-year mortality in patients with more-than-mild AR after TAVI. More-than-mild baseline mitral regurgitation, higher systolic pulmonary artery pressure and male gender were independently associated with worse outcome.
Statistics from Altmetric.com
Transcatheter aortic valve implantation (TAVI) has become the treatment of choice in inoperable patients with severe aortic stenosis (AS),1 and is an accepted alternative to surgical aortic valve replacement (SAVR) in high-risk patients.2 However, residual aortic regurgitation (AR), which is mostly of paravalvular origin, is more frequently seen after TAVI than after SAVR.2 Recent reports have suggested that more-than-mild AR negatively impacts early and long-term prognosis following TAVI, and its association with mortality has raised concerns.3–7
The HR for 1-year mortality in patients treated with TAVI with more-than-mild AR has been recently reported to be 2.27 (95% CI 1.84 to 2.81).8 On the other hand, a considerable proportion of these patients survives and appears to tolerate AR. Based on the current literature, a direct causal relationship between AR and mortality (vs AR being a marker of other factors) still needs to be determined. In addition, the role of confounding variables, such as preprocedural AR, LV function and mitral regurgitation (MR), remains unclear at present. Identification of patients with more-than-mild AR at higher risk of death may assist in tailoring therapy, but predictors of mortality in this subset of patients are largely unknown. Therefore, we sought to characterise patients with residual more-than-mild AR after TAVI that do die in the first postprocedural year, and to assess potential predictors of mortality, using data from the multicentre German TAVI registry.
The German TAVI registry is a multicentre prospective national registry, which has been designed to monitor the use and outcome of TAVI in daily clinical practice, and to evaluate safety, effectiveness and health-economical data. The registry is completely independent from industry, driven by the scientific interest of the participating hospitals, and financed by the Institut für Herzinfarktforschung (Institute of Myocardial Infarction Research) in Ludwigshafen, Germany. Details of the registry have been described elsewhere.9 ,10
Since January 2009, all participating hospitals committed to include all consecutive patients treated with TAVI, independent of the device or access site used. The registry stopped inclusion in June 2010; after that all patients were included in the prospective German Aortic Valve Replacement Registry.11
Proposed inclusion criteria for treatment were severe symptomatic AS and (A) age ≥80 years and a logistic EuroScore12 ≥20% or (B) logistic EuroScore <20% and at least one of the following criteria: liver cirrhosis, chronic pulmonary disease or porcelain aorta. Technical feasibility according to the available prostheses sizes was also required for inclusion. Ultimately, risk assessment and patient inclusion were confirmed by an independent senior cardiologist and senior cardiac surgeon (local heart team). The study complied with the Declaration of Helsinki and the locally appointed ethics committee has approved the study protocol. All patients provided written informed consent before the procedure and also gave written informed consent for processing of their anonymous data.
The present analysis includes all patients with more-than-mild AR after TAVI and a known survival status at 1 year. Among these patients, baseline demographic, clinical, echocardiographic and angiographic characteristics were compared among survivors and non-survivors to identify factors associated with 1-year mortality.
The German TAVI registry was open to all commercially available prostheses, which were the Medtronic CoreValve (Medtronic Inc., Minneapolis, MN, USA) and the Edwards Sapien (Edwards Lifesciences, Irvine, CA, USA) prosthesis at that time. Details of the valves used have been described elsewhere.13 ,14 Briefly, the Medtronic CoreValve consists of a trileaflet porcine pericardial tissue valve, which is mounted and sutured in a self-expanding nitinol stent frame. The profile size is 18 French. With the current generation, four different device sizes are available for different annulus dimensions, but only the 26 mm and 29 mm prostheses were available during the study period. Vascular access options were the transfemoral, transsubclavian and transaortic routes. The Edwards Sapien and more recently the Edwards Sapien XT device are balloon-expandable prostheses with a bovine pericardial tissue valve. Vascular access options included the transfemoral, transapical and transaortic routes. Initial transfemoral procedures were performed with the RetroFlex delivery catheter followed by the RetroFlex II and more recently the NovaFlex catheters incorporating a flexible nose cone. The profile size of the transfemoral vascular sheath was 22–24 French with the first generation Edwards Sapien device and decreased to 18–19 French with the Edwards Sapien XT valve. Two device sizes were available at the time of enrolment, the 23 mm and the 26 mm devices. Technical details of the implant procedure have been described elsewhere.13 ,14
Analysis of the severity of aortic and mitral regurgitation
Preinterventional AR and MR were evaluated by echocardiography and were graded as absent (0), mild (1/4), moderate (2/4), moderate-to-severe (3/4) and severe (4/4). The evaluation was done by the treating physician.
The degree of postprocedural AR was angiographically evaluated at the end of the TAVI procedure after final device deployment and removal of the catheter and guide wire as previously described.3 Qualitative angiographic assessment of the severity of AR was performed by visual estimation of the concentration of contrast medium in the LV, using the method of Sellers et al.15 AR was classified into four grades: absent (0), mild (1/4), moderate (2/4), moderate-to-severe (3/4) and severe (4/4). The evaluation was performed by the treating physician.
Data collection: Data were collected via the internet by the Institut für Herzinfarktforschung at the Heart Centre Ludwigshafen.
Data analysis: Statistical analysis was performed using the SAS statistical package, V.9.1 (Cary, North Carolina, USA). Absolute numbers and percentages as well as means (with SD) are computed to describe the patient population. Categorical values between survivors and non-survivors were compared by χ2 test and continuous variables were compared by two-tailed Wilcoxon rank sum test. Cox regression analysis was used to analyse baseline factors associated with mortality at 1 year. Factors included in the analysis were age, gender, coronary artery disease, peripheral arterial disease, reduced LV EF ≤30%, decompensated heart failure, chronic renal impairment, preprocedural AR, MR ≥2/4 and systolic pulmonary artery pressure. The variables entered into the model had a p value <0.1 in the univariate analysis and/or were clinically relevant to the outcome. The logistic EuroScore was not included into the model, because most of the individual factors were already included. One-year survival rates were analysed using Kaplan-Meier curves and were compared using the log-rank test. Patients with a known survival status at hospital discharge and at 30 days but not at 1 year were included in the survival analysis. p Values <0.05 were considered statistically significant. All p values are results of two-tailed tests. The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
A total of 1432 patients were enrolled between January 2009 and June 2010 at 27 German hospitals. One-year follow-up data were available for 1318 patients (92%). Of the latter, 201 patients (15.2%) had more-than-mild AR after TAVI as evaluated by postprocedural angiography and represent the population of the current analysis (figure 1).
The study population's mean age was 81.2±6.6 years and men represented 55%. The mean logistic EuroScore was 22±15%. Overall, 92% of patients received the Medtronic CoreValve prosthesis and 8% received the Edwards Sapien valve. At 1 year, 61 patients (31%) with more-than-mild post-TAVI AR had died. Baseline clinical and echocardiographic characteristics of the study population stratified according to the survival status at 1 year are shown in table 1. Compared with patients who survived, patients who died more commonly had coronary artery disease (73% vs 55%, p=0.01), peripheral arterial disease (34% vs 17%, p=0.007) and chronic renal impairment (69% vs 50%, p=0.01). Patients who died had a lower baseline LVEF (44±18% vs 52±16%, p=0.002), higher prevalence of more-than-mild (≥2/4) MR (44% vs 27%, p=0.001) (figure 2), and a higher systolic pulmonary artery pressure (51±18 mm Hg vs 44±19 mm Hg, p=0.002), but the severity of aortic stenosis was similar, and the prevalence and severity of pre-TAVI AR was comparable (any AR in 88% vs 83%, respectively, p=0.29) (figure 3). The baseline logistic EuroScore was higher in non-survivors (29±17 % vs 20±13%, p<0.001).
Details of the TAVI procedure and the degree of residual post-TAVI AR are shown in online table 1. The transfemoral route was the most commonly used access in both groups, but was more frequently used in survivors compared with non-survivors (97.1% vs 83.6%, respectively, p<0.001). The CoreValve prosthesis was used in 93.6% and 88.5% of patients, respectively (p=0.22). Implantation of the TAVI prostheses resulted in immediate and marked improvement of the haemodynamic valve status, reducing the transvalvular mean pressure gradient from 50.4±19.2 mm Hg (survivors) and 45.7±21.0 mm Hg (non-survivors) to 5.1±8.3 mm Hg and 4.0±6.5 mm Hg in both groups, respectively. Postprocedurally, there were no significant differences in the grades of residual AR between the study groups, with most patients having moderate degrees of AR (91.4% vs 88.5% in survivors and non-survivors, respectively, p=0.52).
Inhospital events of survivors compared with non-survivors are shown in table 2. Almost 50% of mortality events occurred during initial hospitalisation (figure 4). Haemodynamic complications such as low cardiac output (35.0% vs 7.2%), respiratory failure (40.0% vs 6.4%) and renal failure requiring dialysis (18.3% vs 2.1%) were significantly higher in non-survivors, while rates of myocardial infarction, stroke, vascular complications and advanced atrioventricular block were similar. Seven patients had a repeat interventional procedure for valve-related dysfunction, six were among survivors and one was among non-survivors (p=0.44).
For patients who survived the inhospital period, the New York Heart Association (NYHA) functional class at 30 days was not significantly different between both groups, but there was a strong trend towards a worse functional class in the 31 patients who died in the following year compared with patients who survived (20.0% vs 49.5% with NYHA I and 20.0% vs 5.3% with NYHA IV for non-survivors and survivors, respectively, both p=0.08).
Factors associated with mortality at 1 year
Most of the mortality events at 1 year were reported as being of cardiovascular (39.3%) or unknown (42.8%) aetiology, while only 17.9% were reported as non-cardiovascular deaths. Factors independently associated with 1-year mortality are shown in figure 5. Using Cox regression analysis, only baseline MR ≥grade 2/4 (HR 1.77, 95% CI 1.05 to 2.99, p=0.03) and systolic pulmonary artery pressure (HR 1.15, 95% CI 1.01 to 1.33, p=0.04) were independently associated with 1-year mortality, while female gender was protective (HR 0.54, 95% CI 0.30 to 0.96, p=0.03). Kaplan-Meier survival curves in patients with and without MR ≥grade 2/4 are shown in figure 6.
Despite the difficulties in accurately characterising post-TAVI AR, especially paravalvular leaks, its impact on short-term and long-term mortalities has been consistently reported across different studies.3–7 16–18 The risks and impact of AR can be decreased by careful preprocedural imaging and device selection, standardisation of procedural technique, prompt recognition and adequate management. However, with the current generation TAVI devices, residual AR cannot be entirely eliminated. In addition, corrective measures (such as postdilatation, snaring, valve-in-valve implantation and conversion to SAVR) have all been associated with increased risk of adverse events.19–22 Therefore, TAVI operators are often uncertain about the impact of residual AR on outcomes in a particular patient, and about the necessity and aggressiveness of corrective measures used to reduce post-TAVI AR.
In the current study, we evaluated baseline clinical, echocardiographic and angiographic characteristics of patients with residual more-than-mild AR after TAVI that do die in the first postprocedural year, and sought to assess predictors of mortality. Our aim was to characterise an ‘extreme-risk’ subgroup of patients, in which more-than-mild AR should probably not be accepted. In a multivariate Cox regression analysis, we found that more-than-mild preprocedural MR, an elevated pulmonary artery pressure and male gender were independently associated with 1-year mortality, while pre-existing AR did not appear to be protective. From a physiological point of view, it seems plausible that postprocedural AR would be better tolerated if there is some degree of AR before TAVI, because of the chronic adjustment to volume overload already made by the LV. In a recent study by Hayashida et al,23 patients with residual moderate AR (n=89) and no preprocedural AR had worse outcome compared with those who had preprocedural AR, a finding that could not be confirmed in our analysis. Whether the nature and localisation of high-velocity paravalvular regurgitant jets after TAVI impose a different workload on the LV compared with pre-existing transvalvular regurgitation is currently unknown and remains to be determined, but our findings challenge the concept of a protective preconditioning effect provided by pre-existing AR.
The independent association of pre-TAVI more-than-mild MR with 1-year mortality in our patient population is an important finding. MR is present in most patients with severe aortic stenosis, and large series have reported moderate or severe MR in 22–48% of patients undergoing TAVI,1 ,24–26 where MR is typically left untreated. In a recent study by Toggweiler et al,27 moderate or severe MR in patients undergoing TAVI was associated with a higher early, but not late, mortality rate. At 1-year follow-up, MR was improved in 55% of patients with moderate or severe MR at baseline, and improvement was more likely in patients with high transaortic gradients, with functional MR, without pulmonary hypertension and without atrial fibrillation. Notably, the impact of post-TAVI AR on MR improvement was not evaluated in this study. But in the study by Hayashida et al,23 postprocedural MR increased in parallel with the degree of postprocedural AR despite the absence of significant differences in preprocedural MR, and improvement in MR was only observed in patients with no or mild AR (there was no improvement in patients with moderate AR and deterioration in patients with severe AR). Although the mechanism underlying this finding is unclear, impaired haemodynamics and worsening heart failure resulting from post-TAVI AR may aggravate MR and lead to higher mortality. A malpositioned deeply implanted bioprosthesis can also have an impact on AR and possibly on MR. But even if MR remains unchanged in these patients, the coexistence of two types of left-sided valvular insufficiencies or a ‘double reflux’ may exacerbate the volume overload imposed on the LV, potentially worsening heart failure and adversely affecting short-term and long-term outcomes. A similar pathophysiological explanation can be hypothesised in patients with an elevated pulmonary artery pressure, but these explanations should be further explored.
Importantly, based on the current analysis, a direct causal relationship between more-than-mild MR or an elevated pulmonary artery pressure and mortality in patients with more-than-mild AR after TAVI cannot be determined. Careful analyses of the influence of all degrees of MR and AR on LV geometry and remodelling28 and the determination of the precise cause of death (cardiovascular vs non-cardiovascular) would be needed to confirm the strength and nature of this relationship, which is not possible from the available registry data. However, several factors suggest that a significant proportion of deaths in our cohort could be related to a worse haemodynamic status, including the significantly higher haemodynamic complications during initial hospitalisation, while other TAVI-related complications (such as stroke or vascular complications) were comparable; the worse NYHA functional class at 30 days; and the minority of patients that died due to a clear non-cardiovascular cause during the follow-up period.
Finally, an interaction between gender and outcome after TAVI has been recently observed, and female sex has been associated with improved outcomes and identified as a predictor of mid-term survival.29 The protective effect of female sex in our cohort with significant post-TAVI AR could be a mere extension to this observation, but could also be related to sex-specific pathoanatomical differences in LV geometry, function and recovery. In a recent study by Stangl et al,30 women undergoing TAVI had more severe aortic stenosis, higher LVEF, smaller end-diastolic and end-systolic diameters and more concentric hypertrophy at baseline compared with men. Regression of hypertrophy occurred in men and women, but improvement of the EF was statistically significant only in women. Theoretically, these factors may ameliorate the negative haemodynamic consequences of residual AR in women, but further studies with a detailed and serial imaging assessment are needed to confirm or dispute this finding.
The study has all the shortcomings of a registry, yet its value lies in the relatively large number of patients recruited. Echocardiographic and angiographic data and all adverse events were not centrally adjudicated, which may have contributed to the differences based on differences in self-reports, and the angiographic technique for AR assessment was not standardised. In addition, echocardiographic follow-up data were not captured in this registry, which prevents assessment of the evolution of MR and AR over time. As expected, there was some overlap in mortality predictors of patients with more-than-mild AR (study population) and those with no/mild AR (see online table S2). Further studies with detailed haemodynamic assessment at baseline and at long-term follow-up are needed, in order to define whether the observed mortality predictors have a different haemodynamic impact in both patient subgroups.
We identified preprocedural characteristics associated with 1-year mortality in patients with more-than-mild AR after TAVI. More-than-mild baseline MR, higher systolic pulmonary artery pressure and male gender were independently associated with worse outcome. These findings could be useful to guide clinical decision making in this challenging subgroup of patients.
What is known on this subject?
Residual aortic regurgitation (AR) after transcatheter aortic valve implantation (TAVI) is associated with increased mortality. However, a considerable proportion of these patients survives and appears to tolerate AR. Identification of patients at higher risk of death may assist in tailoring therapy, but predictors of mortality in this subset of patients is largely unknown.
What might this study add?
We identified preprocedural characteristics associated with 1-year mortality in patients with more-than-mild AR after TAVI. More-than-mild baseline mitral regurgitation and higher systolic pulmonary artery pressure were independently associated with worse outcome, while female gender was protective.
How might this impact on clinical practice?
This analysis provides insight into a vulnerable group of patients that are most likely to be adversely affected by post-TAVI AR, and could be useful to guide clinical decision making for prevention and treatment of this complication.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Contributors All authors of this manuscript had: (1) substantial contributions to the conception or design of the work; or the acquisition, analysis or interpretation of data for the work; and (2) participated in drafting the work or revised it critically for important intellectual content; and (3) finally approved the version to be published; and (4) agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Competing interests MA-W and GR report grants from Medtronic. MA-W reports personal fees from Edwards Lifesciences. RZ reports personal fees from Medtronic and Edwards Lifesciences. All other authors do not report any conflict of interest in relation to the contents of this manuscript.
Ethics approval Ethics committees of participating centres.
Provenance and peer review Not commissioned; externally peer reviewed.
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.