Introduction

Calcific aortic stenosis (AS) is the most frequent heart valve disease in developed countries. Current recommendations for aortic valve replacement (AVR), either surgical (SAVR) or transcatheter (TAVR), are driven by AS severity graded by echocardiographic parameters, presence of symptoms (dyspnea, heart failure, angina, syncope) and left ventricular (LV) systolic dysfunction (LV ejection fraction (EF) < 50%) [1, 2]. Additionally, the heart team decides about the therapy modality according to surgical risk stratification, accounting comorbidities, and frailty status [1, 2]. However, clinical outcome is influenced not only by the severity of the AS, but also by the consequences of chronic afterload elevation leading to extent anatomical and functional changes of the myocardium. Parameters reflecting such myocardial remodeling are LV systolic and diastolic dysfunction [3,4,5], left atrial (LA) enlargement, significant mitral (MR) [6] and tricuspid regurgitation (TR) [7], damaged pulmonary vasculature [8] as well as right ventricular (RV) damage [9]. Recent studies have proposed a non-invasive echocardiographic staging classification differentiating the cascade of cardiac damage [10], that has predictive value for the clinical outcome after AVR in either symptomatic [11, 12] and asymptomatic AS [13]. Moreover, a small retrospective analysis showed that invasive staging of cardiac damage is predictive for mortality, especially the progressed stages with pulmonary and RV damage [14]. Although, routine assessment of invasive hemodynamics is not recommended by the current guidelines [1, 2], especially assessment of pulmonary vasculature function, is proved to provide prognostic value in AS patients undergoing AVR [8]. While, on one hand, echocardiography is limited in assessment of pulmonary vasculature, on the other hand, it is necessary for classification of cardiac damage staging. Therefore, the present study aims to proof the invasive assessment of cardiac damage in a large cohort of severe AS, to analyze its prognostic significance for the clinical outcome, and to elaborate its value compared to the echocardiographic classification.

Methods

Patient population and study design

Prospectively collected data of patients with symptomatic severe AS who underwent TAVR at the Asklepios Klinik St Georg between July 2008 and November 2017 (n = 1855) were retrospectively analyzed. Finally, 1400 datasets of full invasive hemodynamic assessments by left- and right-heart catheterizations, and full echocardiographic assessments were included. A detailed specification of the parameters which were assessed and calculated is provided in Online Resources 1. Clinical outcomes after TAVR were assessed according to the Valve Academic Research Consortium (VARC)-2 criteria [15].

Cardiac damage staging classification

Patients were categorized according to the five stages of the cardiac damage staging classification described by Maeder et al.[14] with some modifications (Fig. 1): stage 0 = no signs of cardiac damage; stage 1 = LV damage (LV end-diastolic pressure (LVEDP) > 15 mmHg and/or cardiac output (CO) ≤ 4.0L/min and/or cardiac index (CI) ≤ 2.5L/min/m2); stage 2 = LA and/or mitral valve damage (mean pulmonary artery wedge pressure (PCWP) > 15 mmHg and/or atrial fibrillation); stage 3 = pulmonary and/or tricuspid valve damage (SPAP ≥ 60 mmHg; and/or pulmonary vascular resistance (PVR) > 3 wood units); stage 4 = RV damage (mean right atrial pressure (RAP) > 15 mmHg). The invasive criteria were compared to the following non-invasive cardiac damage staging [10]: stage 0 = no signs of cardiac damage; stage 1 = LVEF < 50% and/or E/e´ > 14 and/or LV mass index (LVMI) > 115 g/m2 (male) or > 95 g/m2 (female); stage 2 = moderate-to-severe MR and/or LA diameter > 47 mm and/or presence of atrial fibrillation; stage 3 = SPAP ≥ 60 mmHg; and/or moderate-to-severe TR; stage 4 = tricuspid annular plane systolic excursion (TAPSE) < 16 mm.

Fig. 1
figure 1

Invasive and echocardiographic criteria for staging of cardiac damage of severe aortic stenosis

Full size image

Statistics

Continuous variables are described as means and standard deviations or medians and interquartile range (IQR), as appropriate; they were compared by one-way analysis of variance (ANOVA) with post hoc Bonferroni correction for multiple comparisons and t tests. Categorical data are described with absolute and relative frequencies and compared by the Fisher’s exact test. A two-tailed p value < 0.05 was considered statistically significant. Hazard ratios (HR) were calculated in a multivariable analysis, including all variables with a p value < 0.05 in the univariate analysis, was used to determine independent predictors of 1-year mortality after TAVR. A p value < 0.05 was considered statistically significant. Kaplan–Meier analysis was used to estimate the incidence of clinical outcomes at 1- and 4-year follow-up. Statistical analyses were performed with IBM Statistical Package for Social Sciences, version 26.0.0 (SPSS, Inc., Chicago, Illinois).

Results

Distribution of cardiac damage stages

Regardless the classification modality, most of the patients were stratified in stage 2 (echocardiographic criteria 30%, invasive criteria 33.5%; Fig. 2, graphic abstract). Stage 4 was more frequent if stratified by echocardiography (21.6%) than invasively (10%). A deeper analysis of matching both staging systems revealed that only 21.6%, 33.7%, 42.9%, 38.7%, 23.4% of patients classified as stage 0, 1, 2, 3, 4 by echocardiography are classified as same stage by invasive classification (Fig. 3). The other patients would be re-classified to another stage after invasive assessment. The correlation was only moderate (r = 0.391; 0 < 0.001).

Fig. 2
figure 2

Distribution of stages of cardiac damage in severe aortic stenosis depending on different assessment modalities: echocardiographic and invasive criteria

Full size image
Fig. 3
figure 3

Overlap of cardiac damage stages using different assessment modalities: echocardiographic and invasive criteria

Full size image

Baseline characteristics

Mean age of the population was 81.5 ± 6.8 years and 46.3% were males (Table 1). With increasing stage, prevalence of impaired renal function, atrial fibrillation and COPD increased. Moreover, these patients presented with increased surgical risk scores and impaired functional capacity (NYHA ≥ III).

Table 1 Demographics
Full size table

Echocardiographic characteristics and biomarkers.

Echocardiographic characteristics are summarized in Table 2. The higher the staging class the less aortic valve area, mean transvalvular gradients, LVEF, and TAPSE was documented. LA diameter, E/e’ and LV diameters were decreased in patients with a lower damage stage. Furthermore, progressed cardiac damage was associated with elevated prevalence of moderate-to-severe MR and TR as well as increased SPAP (Table 3). Elevated CRP-, Creatinine- and NT-proBNP-levels were noticed in patients with advanced cardiac damage.

Table 2 Echocardiographic data and biomarkers
Full size table
Table 3 Invasive hemodynamic data
Full size table

Invasive characteristics

Progressed cardiac damage was associated with significantly lower mean aortic transvalvular pressure gradients, systolic arterial pressures and stroke volume indexes. Furthermore, with increasing stage elevated LVEDP, PAPs, PCWP, and RAP were documented. The prevalence of normal-flow, low-gradient pattern decreased, while paradoxical and classical low-flow, low-gradient pattern increased with stage severity (Table 3).

Clinical outcomes

Median follow-up time of the whole population was 1.8 years and mean follow-up time was 2.2 years with an interquartile range of 0.8 (Q1)–3.4 (Q3) years. Acute device success was achieved in 94.9%. There were no differences regarding TAVR-associated complications documented (Online resources 2). From stage 0 to 4, all-cause in-hospital (stage 0/1/2/3/4: 0/3.9/4.3/6.2/8.6%; p = 0.008) and 1-year mortality (10.1/16.1/17.3/22/27.1%; p = 0.001) significantly increase. This effect becomes clearer in the analysis of the 4-year overall (29.5/40.4/39.4/54.6/62.2%; p < 0.001) and cardiovascular (4/6.4/8.7/15.9/25.9%; p < 0.001) mortality (Fig. 4a, b). Figure 4c, d shows similar overall and cardiovascular survivals after graduation by echocardiographic staging classification. Pairwise comparison of the stages graduated by echocardiographic vs. invasive criteria showed similar results regarding survival (Online resources 3a–j).

Fig. 4
figure 4

Survival analyses according to stage of cardiac damage after TAVR using echocardiographic or invasive criteria. Kaplan–Meier plots comparing overall and cardiovascular 4-year survival in cardiac damage stages stratified by invasive (a, b) and by echocardiographic (c, d) criteria

Full size image

While NYHA functional class at baseline differed significantly between the stages, with the greatest prevalence of NYHA IV in stage 4 (Fig. 5a; p = 0.007), improvement in functional capacity after TAVR remained similar in all patients after 1 year regardless the staging class (Fig. 5b; p = 0.074).

Fig. 5
figure 5

Changes of functional capacity after transcatheter aortic valve replacement depending on stage of cardiac damage using invasive criteria. New York Heart Association (NYHA) functional class at baseline (a) and after 1 year (b)

Full size image

Predictors of 1-year mortality

Online resources 2 shows the univariate analysis for the whole population. Eleven significant parameters from the univariate analysis were included in the multivariable Cox regression and identified the following independent risk factors for 1-year mortality (Table 4): age (HR 1.04; 95% CI 1.02–1.06; p < 0.001), eGFR (HR 0.99; 95% CI 0.98–0.99; p < 0.001), COPD (HR 1.28; 95% CI 1.02–1.59; p = 0.031), atrial fibrillation (HR 1.25; 95% CI 1.04–1.52; p = 0.020), peripheral vascular disease (HR 1.37; 95% CI 1.13–1.66; p = 0.001), classical low-flow, low-gradient aortic stenosis (HR 1.48; 95% CI 1.17–1.87; p = 0.001), moderate-to-severe paravalvular leakage (HR 2.10; 95% CI 1.34–2.00; p = 0.001), and staging cardiac damage by invasive criteria with a 16% increase in mortality per each increment of stage (HR 1.16; 95% CI 1.03–1.32; p = 0.018). While staging cardiac damage by echocardiographic criteria was significantly relevant in the univariate Cox regression, it was not significant anymore in the multivariable analysis.

Table 4 Multivariable Cox regression for 1-year mortality
Full size table

Discussion

This study represents the largest single-center analysis of invasive hemodynamics in patients with severe AS undergoing TAVR. The main findings can be summarized as follows: (1) Based on invasive hemodynamic assessment, extra-aortic valvular cardiac damage is highly prevalent (graphic abstract); (2) Invasive staging classification of cardiac damage is feasible and predicts independently clinical outcome after TAVR.

Current guidelines

AS is the most common heart valve disease in high-income countries with a prevalence of approximately 2% in elderly patients, that will massively increase in the future with the aging population [3]. According to the practice guidelines, diagnostic of AS is based on echocardiographic criteria [1, 2]. Furthermore, the LVEF is the only echocardiographic parameter mentioned in the guidelines playing a role in decision-making regarding treatment. AVR is today the only therapy option to improve symptoms and survival. Whether patients undergo SAVR or TAVR is a decision of the interdisciplinary heart team considering surgical risk scores as well as patient’s frailty status [1, 2]. Unfortunately, except of LVEF and SPAP, there are no other parameters mentioned embodying the cardiac damage according to AS.

Cardiac damage staging by echocardiographic parameters

Echocardiographic assessment of AS grade is based on evaluation of the hemodynamics by continuous and pulsed-wave Doppler, quantifying the LV outflow tract diameter, LVEF and stroke volume index [1, 2]. The assessment of cardiac damage associated with AS is not state-of-the-art today. Staging classification was first introduced by Genereux et al. analyzing the population form the PARTNER 2 trial with 1661 patients [10]. Noninvasive cardiac damage staging was strongly associated with 1-year overall survival after AVR (stage 0/1/2/3/4: 95.6/90.8/85.6/78.7/75.5%; p < 0.001) as well as cardiovascular survival (97.7/98.7/90.8/87.6/81.6%; p < 0.001) with a 1.46-fold increase in mortality per each increment in stage [10]. The concept of cardiac damage staging in AS is based on elevated LV afterload due to the stenotic aortic valve, leading to formation of myocardial fibrosis with LV systolic and diastolic dysfunction [3,4,5, 16]. Obviously, the chronic pressure overload leads to a sequential progression of damage from LV (stage 1) to LA enlargement, and/or significant MR (stage 2) [6], and/or the presence of atrial fibrillation [17] , following to pulmonary vasculature impairment, and/or significant TR (stage 3) [7, 8], resulting in RV damage (stage 4) [9]. These parameters affect negatively the prognosis of severe AS, even after AVR. This hypothesis was corroborated by Fukui et al. [12] who analyzed 689 AS patients 24-month after TAVR. The authors demonstrated a 1.4–2.8-fold increase in mortality per each increment of stage (stage 1/2/3/4: 82.8/73.5/58.5/46.4%) [1, 2]. Furthermore, Vollema et al. in an analysis of 1.189 patients with symptomatic severe AS, showed a 1.28-fold increase in mortality per each increment of stage with the lowest 5-year survival rates in stage 4 (73/68/61/45/41%; p < 0.001) [11]. In a recently published manuscript by Vollema et al., analyzing LV global longitudinal strain (GLS) in severe AS classified by the proposed staging classification, the authors showed that an incorporation of the LVGLS could improve the prognostic value of the classification by identifying patients in a more progressed stage [18]. In a further investigation, Tastet et al. suggested a modified staging classification in patients with moderate-to-severe (AVA < 1.5 cm2) asymptomatic AS with preserved LVEF [13]. Due to underestimation of LV dysfunction by LVEF in concentric remodeling and LV hypertrophy [5, 19], the LVEF cutoff was increased to 60% and LVGLS < 15% was added. Moreover, slow-flow state (SVI < 30 ml/m2) was added to stage 4 as a heart failure parameter. The results showed that > 50% of all asymptomatic patients were in a progressed cardiac damage stage (≥ 2) with a 2.6-fold increased mortality risk.

Instead of LAVi, we used the LA diameter as proposed by the echocardiographic guidelines for LA assessment [20] and could show comparable results for 1- and 4-year overall and cardiovascular survival (Fig. 4). Noticeably, patients in stage 1 and 2 showed very similar survival curves. Probably, the use of LAVi could result in a better stratification as shown in the previous publications [10, 11, 13]. The distribution of cardiac damage stages was comparable to the recently published data with the majority of patients (30%) classified in stage 2 (Fukui et al. 62% [12]; Genereux et al. 51% [10]; Vollema et al. 49% [11]; Tastet et al. 46% [13]). Interestingly, in our population RV damage was more often prevalent (> 20%) compared to the other groups (8.7% [10], 12% [11], 4% [12]). In the PARTNER 2 trial [10] patients with very severe impaired LVEF < 20% and very advanced renal insufficiency (creatinine > 3 mg/dl) were excluded, what could be a reason for lower prevalence of RV damage of the population. Compared to Vollema’s population [11], our patients were approximately 10 years older with noticeably higher prevalence of moderate-to-severe MR (6 vs. 41.9%) and TR (6 vs. 29.3%), that may lead different distribution of stages. Conclusively, this non-invasive staging scheme is immensely helpful in understanding the pathomorphological consequences of the AS disease more deeply and could be beneficial in therapy management [13, 21].

Cardiac damage staging by invasive parameters

Current guidelines are not recommending the routine assessment of invasive hemodynamics in patients with severe AS [1, 2]. Only in special situations, like uncertainty of AS severity in the non-invasive diagnostics, or assessment of SPAP in asymptomatic patients to drive decision-making regarding AVR. Nevertheless, invasive hemodynamics in AS could be of importance to understand extent of cardiac damage. Maeder et al. first analyzed 421 patients with severe AS undergoing AVR (SAVR 70%; TAVR 30%) stratified by invasive criteria [14]. Stage 0: no cardiac damage; LV damage (stage 1) was defined as LVEDP > 15 mmHg as proposed by the guidelines for heart failure [22]. As Maeder et al. showed similar survival curves for stages 0/1/2, we modified their criteria by adding impaired CO and/or CI, which are also proposed by the guidelines [22]. LA and/or mitral valve damage (stage 2) was defined as elevated PCWP [14], that is proposed to diagnose heart failure with preserved EF [22], and is the criterion distinguishing between post- and precapillary pulmonary hypertension [23]. Elevated PVR was used for defining pulmonary vasculature damage (stage 3), being the parameter to distinguish between isolated post-capillary and combined post- and precapillary pulmonary hypertension [23]. Furthermore, an elevated SPAP (> 60 mmHg), a proved predictor for mortality, was also used for stage 3 [1, 2, 8]. An elevated RAP (> 15 mmHg) was suggested as the criterion for RV damage (stage 4) [14].

Distribution of stages were comparable between our population (stage 0/1/2/3/4: 9.9/23.6/33.5/23.1/10%) and Maeder et al. (16/27/36/17/4%) [14]. These results are consistent with the echocardiographic findings of previous publications with the majority of patients classified as stage 2 [10,11,12, 14]. Nevertheless, our population showed to be significantly older (75 vs. 82 years), with a higher prevalence on comorbidities (renal dysfunction, COPD, stroke, diabetes mellitus, coronary artery disease, significant MR), and NYHA ≥ III. That contributes to represent a real-world TAVR population compared to Maeder et al., where 70% have been treated by SAVR [14]. Obviously and consistent with previous studies, higher rates of low-flow, low-gradient pattern were seen in progressed stages [7, 11].

Maeder et al. could show that stages 3 and 4 were associated with a 3.5- and 5.4-fold increased mortality risk after AVR, whereas stages 0–2 showed no significant difference in mortality after AVR [14]. Furthermore, we found in our population a 1.16-fold increase in mortality per each increment in stage. In the multivariate analysis, invasive classification was superior to echocardiographic. With the addition of invasive CO/CI measurement picturing the hemodynamic measurement of LVEF, we could show a significant difference between stage 0 and 1/2 regarding mortality. Nevertheless, stages 1 and 2 showed very similar survival curves. Although, correlation between echocardiographic and invasive staging classification was only moderate (r = 0.391; p < 0.001; Fig. 3), direct comparison of each stage showed no significant difference regarding survival rates (Supplemental Data 1 A-J). It should be considered to add other invasive parameters to the classification to distinguish between stage 1 and 2 more precisely. Moreover, 14.6% classified by echocardiographic criteria in stage 0–2 would be re-classified after RHC in stage 3 or 4 according to the invasive staging. Re-classification of stages could have an immense value in decision-making and therapy management, especially in SAVR versus TAVR discussion.

Although, echocardiography can diagnose PH with a sufficient sensitivity and specificity in AS [24], invasive assessment remains the gold standard understanding pulmonary pressures, especially in calculating the PVR. In previous publications, the predictive value of PH and the importance of the different PH subtypes, as well as the effect of PAP recovery after TAVR, have been demonstrated [8]. Furthermore, RV dysfunction has been recognized as a predictor for mortality after AVR. Especially worsening RV function was fourfold more frequent after SAVR than after TAVR with a significant increase of mortality [9, 25]. Hence, Tastet et al. suggest symptomatic patients to recommend TAVR preferably to SAVR if they are at least stage 3 [10]. Moreover, asymptomatic patients should undergo complementary echocardiographic staging with GLS [13, 18] and invasive assessment to identify patients with severe AS who showed rapid disease progression to determine best timing for AVR. Obviously, all the assessed parameters, invasive and non-invasive, are not only related to the AS but could also be due to other concomitant conditions as arterial hypertension or coronary artery disease. Nevertheless, these patients are at higher mortality risk and will benefit from the optimal therapy modality. Cardiac damage staging could be useful, if integrated into risk stratification assessment of patients with AS adding to surgical risk scores and patient’s frailty.

Limitations

This is a single-center study, thereby reflecting only restricted experience. It represents a retrospective analysis, although the data were collected prospectively. Only patients with complete left- and right-heart catheterization data were included in the analysis; individuals with incomplete data were excluded, thus it is not a consecutive series of patients. Since patients underwent left and right heart catheterization directly before TAVR, there could be discordant findings in measurements depending on sedation at that time. LAVi, that is crucial for echocardiographic cardiac damage stage 2, was not documented for all patients. We decided to take the LA diameter instead, as part of the current guidelines for quantification of the LA [20]. Unfortunately, this is a retrospective study with incomplete follow-up status at 4-year follow-up.

Conclusions

An invasive cardiac damage staging classification of patients with severe symptomatic AS could predict mortality after TAVR and be useful in therapeutic management and decision-making. Further studies are needed to adapt and complement non-invasive and invasive staging modalities to all valvular diseases, especially regarding the clinical status of the patients to simplify decision-making and optimization of therapy management.