Suppression of tumorigenicity (ST2) belongs to the Toll-like/IL-1 receptor family and exists in two different forms, one as a transmembrane form (ST2L) and the other as a soluble form (sST2) [7]. ST2L interacts with IL-33 as a ligand-receptor complex and acts in a complex signaling cascade against cardiac remodeling and fibrotic remodeling processes. However, high mechanical stress responses in the heart and lungs result in increased secretion of sST2 from alveolar epithelial cells and cardiac myocytes. These bind with higher affinity to Interleukin (IL)-33 and thus prevent cardioprotective signaling. High plasma sST2 levels are therefore associated with an increased risk of adverse outcomes in patients with severe AS, heart failure and PH [8, 9].

GDF-15 is a member of the transforming growth factor-beta (TGF-$\beta$) superfamily and is secreted by numerous cells such as macrophages, cardiomyocytes, pulmonary endothelial cells and vascular smooth muscle cells [10, 11]. The growth factor plays a crucial role in inflammatory processes, tissue injury and apoptosis and has been found to be eleveated in numerous pathological conditions such as cardiovascular and pulmonary disease [12, 13].

H-FABP is a cytoplasmic protein secreted by cardiomyocytes in the context of acute ischemic heart disease. At the molecular level, H-FABP is involved in lipid metabolism, transporting fatty acids from the cell membrane to mitochondria for eventual oxidation [14]. This biomarker has already found its way into clinical practice, as it is already available as a rapid test to diagnose myocardial infarction at an earlier stage [15].

IGF-BP2 is an important member of the insulin-like growth factor family regulating the activity of the insulin-like growth factor (IGF) in most tissues and organs including liver, heart, CNS and reproductive organs [16]. IGF-BP2 exerts an inhibitory effect on the growth hormone IGF-1, which has a cardioprotective function by downregulating the renin-angiotensin-aldosterone system [17]. Elevated IGF-BP2 levels, through IGF-1 inhibition, thus lead to a consecutive unopposed renin-angiotensin-aldosterone effect, resulting in cardiac remodeling and left ventricular dysfunction [18].

suPAR is the soluble form of the cell membrane protein urokinase-type Plasminogen Activator Receptor (uPAR), is released into the blood during inflammation of any kind and therefore provides information about inflammatory activity in the human body. Numerous studies have described increased plasma concentrations in patients with coronary heart disease, myocardial infarction and chronic heart failure [19].

BNP is a cardiac hormone, which is released by cardiomyocytes in the course of stretching processes of the left ventricle. Especially in patients with pressure and volume overload, the plasma concentration of BNP is significantly increased. Evidence based, it is already handled as a relevant heart failure biomarker in clinical practice. Patients with moderate to severe AS showed significantly increased mortality at higher BNP plasma concentrations compared to patients with baseline BNP at follow-up [20].

Troponin is a relevant protein complex consisting primarily of three subunits. Two of them, troponin T and troponin I are specifically formed in the myocardium and are relevantly involved in the interaction of actin and myosin filaments. Any form of damage to cardiac myocytes will result in increased release of troponin. This is exploited clinically for early detection of myocardial infarction [21].

Between 2016 and 2018, 179 patients with severe, primary degenerative AS planning for TAVR procedure were enrolled in current study. Corresponding data analyses were performed at Paracelsus Medical University Hospital Salzburg and Kepler University Hospital Linz in accordance to to principles of the Declaration of Helsinki and Good Clinical Practice.

Transthoracic echocardiography was performed using common ultrasound devices (iE33 and Epiq 5; Philips Healthcare, Hamburg, Germany). Severe AS was classified according to current valid guidelines of European Society for Cardiology measuring. An AV Vmax (maximal velocity over aortic valve) of 4.0 m/s, an AV dpmean (mean pressure gradient over aortic valve) $\ge$40 mmHg and an aortic valve area $\le$1.0 cm${}^2$ formed the definition of severe AS. Patients with low-flow, low-gradient AS and a stroke volume 35 mL/m${}^2$ were excluded from the study, so the patient population presented here includes solitary individuals with high pressure gradients without a low-flow situation. Simpson’s method was applied to receive left ventricular ejection fraction (LVEF). To graduate mitral, aortic, and tricuspid valve regurgitation in minimal, mild (I), moderate (II) and severe (III) spectral and color-Doppler images were used. The maximum tricuspid regurgitant jet velocity (TRV) is obtained by continuous wave Doppler over the tricuspid valve and is used to calculate the pulmonary artery pressure (PAP) using the formula 4 $\times$ TRV${}^2$. To finally obtain the sPAP, which is decisive for PH, the right atrial pressure (RAP) had to be estimated. This corresponds to the central venous pressure and is determined by the diameter of the inferior vena cava (IVC). The following RAP determination was performed in the respective cohorts from Salzburg and Linz: With an IVC diameter $\ge$21 mm and a respiratory caliber fluctuation 50%, a RAP of 15 mmHg was assumed. For an IVC diameter 21 mm as well as a respiratory caliber fluctuation $\ge$50%, a RAP of 3 mmHg was calculated. Other scenarios not corresponding to the above constellations were ascribed an intermediate value of 8 mmHg. Finally, the simplified Bernoulli equation (4 $\times$ TRV${}^2$) + RAP leads to an sPAP result. An sPAP $\ge$40 mmHg was used as the cut-off value to determine PH in accordance with the current literature [22, 23, 24]. In particular, Schewel et al. [25] compared echocardiographically obtained sPAP with invasively RHC obtained sPAP data. The correlation coefficient of r = 0.820 was in a very satisfactory range. This study also demonstrated that a cut-off value $\ge$40 mmHg had better overall statistical goodness criteria than a cut-off value $\ge$45 mmHg or $\ge$50 mmHg.

Blood samples were obtained on the day of hospitalization and thus one day before the actual TAVR procedure under fasting conditions using a vacuum-containing system. The collection tubes were centrifuged, the plasma obtained was separated from the blood components and then frozen at –80 °C to analyze the total of 179 samples at similar time points under same conditions.

Plasma levels of sST2, GDF-15, H-FABP, IGF-BP2 and suPAR were measured by using enzyme-linked immunosorbent assay (ELISA) kits (sST2: Duoset DY523, GDF-15: DY957, H-FABP: DY1678, IGF-BP2: DY674, suPAR: DY807, R&D Systems, Minneapolis, MN, USA). Instructions of the manufactures were performed for adequate preparation of reagents. In summary, serum samples and standard protein were loaded onto the wells of ELISA plates (Nunc MaxiSorp flat-bottom 96 well plates, VWR International GmbH, Vienna, Austria) and incubated for two hours. The plates were treated with Tween 20/PBS solution (Sigma Aldrich, St. Louis, MO, USA) and subsequently a biotin-labeled antibody was added. The subsequent incubation time was another two hours. A washing process again was performed and streptavidin-horesradish-peroxidase solution was added to the wells. A color reaction was generated after adding tetramethylbenzidine (TMB; Sigma Aldrich, St. Louis, MO, USA). Optical density was determined at 450 nm on an ELISA plate-reader (iMark Microplate Absorbance Reader, Bio-Rad Laboratories, Vienna, Austria).

Statistical analysis was performed using SPSS (Version 25.0, SPSS Inc., Armonk, NY, USA).

First of all, the Kolmogorov-Smirnov test was applied to test variables for normal distribution. Normally distributed metric data was expressed as mean $\pm$ standard deviation (SD) and analyzed using an unpaired student’s t-test. Not-normally distributed metric data was expressed as median and interquartile range (IQR) and the Mann-Whitney-U-test was applied for statistical analysis. Frequencies/percentages were used for categorial clinical data and compared using the chi-square test.

To determine an optimal cut-off value of examined cardiovascular biomarkers according to a prediction of an sPAP $\ge$40 mmHg, Area Under the Receiver Operator Characteristics (AUROC)-curves with Area Under the Curve (AUC) and separate analysis of Youden Index (YI) was performed.

Correlation analysis was performed using Pearson’s rank-correlation coefficient to draw conclusions about a relationship between echocardiographic sPAP and cardiovascular biomarkers.

Kaplan-Meier curves were carried out to detect overall 1-year survival of patients in dependence of sPAP, whereby the currently accepted classification into three severity levels (I: sPAP 40 mmHg, II: sPAP 40–59 mmHg; III: sPAP $\ge$60 mmHg) was used.

At last, a univariate Cox proportional hazard regression model was used to calculate hazard ratio (HR) and 95% confidence interval (CI) for several influencing factors associated with 1-year-mortality in patients undergoing TAVR procedure. For better comparability, a z-transformation was absolved for metric data. Afterwards, multivariate Cox regression was performed to assess independent predictors of mortality. Therefore, again covariates associated with mortality in the univariate analysis (p 0.100) were entered and a backward variable elimination was done.

A p-value 0.050 was considered statistically significant.

A total of 179 patients with severe AS from the University Hospitals of Salzburg and Linz were included in the study. Echocardiographically, 82 patients (45.8%) showed an sPAP 40 mmHg equivalent to the absence of PH and 97 patients (54.2%) showed an sPAP $\ge$40 mmHg consistent with the echocardiographic criterion of PH.

Table 1 shows the collected baseline characteristics of the overall cohort as well as those of the classification into patients with sPAP 40 mmHg and sPAP $\ge$40 mmHg. The sPAP groups were compared with each other for significance and the corresponding p-values were documented.

The overall cohort had a mean age of 82.7 $\pm$ 4.8, with a male:female ratio of 50.8% vs. 48.2%. Regarding concomitant disease, arterial hypertension was documented in 78.8%, general cardiovascular disease in 72.1% and diabetes mellitus in 23.5% of patients. Echocardiographic left ventricular ejection fraction (LVEF) averaged 55.0 $\pm$ 10.9% and sPAP 44.6 $\pm$ 15.2 mmHg.

Patients with an sPAP $\ge$40 mmHg, in contrast to those with an sPAP 40 mmHg, showed significantly higher STSScores (3.3 $\pm$ 1.5 vs. 2.6 $\pm$ 1.3; p = 0.025), significantly higher percentages of atrial fibrillation (46.4% vs. 28.0%; p = 0. 012), significantly lower tricuspid annular plane systolic excursion (TAPSE) (20.8 $\pm$ 3.9 mm vs. 23.1 $\pm$ 3.2 mm; p = 0.008) and ultimately significantly higher percent distributions with respect to mitral valve (33.0% vs. 15.9%; p = 0.009) and tricuspid valve insufficiency (28.9% vs. 6.1%; p 0.001).

Fig. 1 provides an overview of the corresponding plasma concentrations of the determined cardiovascular biomarkers depending on the sPAP obtained ($\ge$40 mmHg vs. 40 mmHg).

Fig. 1.

Biomarker analysis. Serum concentrations of sST2 (A), GDF-15 (B), H-FABP (C), IGF-BP2 (D), suPAR (E) and BNP (F) in patients with an sPAP 40 mmHg and with an sPAP $\ge$40 mmHg.

sST2 (Fig. 1A) showed significantly higher plasma concentrations in patients with an sPAP $\ge$40 mmHg than with an sPAP 40 mmHg (16467.1 $\pm$ 10606.6 pg/mL vs. 11563.7 $\pm$ 6708.6 pg/mL; p = 0.010). Similar constellations were found for GDF-15 (Fig. 1B) with plasma concentrations of 785.9 $\pm$ 1034.8 pg/mL vs. 357.2 $\pm$ 683.2 pg/mL (p = 0.005), IGF-BP2 (Fig. 1D) with plasma concentrations of 203352.4 $\pm$ 169893.4 pg/mL vs. 94235.2 $\pm$ 137450.8 pg/mL (p = 0.029) and suPAR (Fig. 1E) with plasma concentrations of 3951.6 $\pm$ 1468.3 pg/mL vs. 3000.5 $\pm$ 1127.4 pg/mL (p = 0.018). BNP (Fig. 1F) with plasma concentrations of 3369.0 $\pm$ 4978.0 pg/mL vs. 1195.0 $\pm$ 1024.2 pg/mL (p 0.001) showed the best result among all biomarkers studied. cTnI presented with plasma concentrations of 27.0 $\pm$ 18.5 pg/mL vs. 16.0 $\pm$ 13.5 pg/mL also statistically significant (p = 0.039).

Only H-FABP (Fig. 1C) did not show significant differences between the sPAP groups (0.5 $\pm$ 2.4 ng/mL vs. 0.4 $\pm$ 1.3 ng/mL; p = 0.069).

To analyze sST2, GDF-15, H-FABP, IGF-BP2 and suPAR as potential biomarkers for prediction of an sPAP $\ge$40 mmHg in patients with severe AS before TAVR, AUROC-curves regarding plasma level concentration of examined biomarkers were figured out. Therefore AUC, cut-off values with YI as well as sensitivity and specificity were extracted in addition to ROC curves (Fig. 2).

Fig. 2.

AUROC analysis of biomarkers. AUROC analyses of sST2 (A), GDF-15 (B), H-FABP (C), IGF-BP2 (D), suPAR (E) and BNP (F) for prediction of sPAP $\ge$40 mmHg with concerning cut-off values, Youden Index, sensitivity and specificity.

This analysis identified an sST2 plasma level (Fig. 2A) of 16045.84 pg/mL as an optimal cut-off value concerning an sPAP $\ge$40 mmHg (AUC 0.613; 95% CI 0.529–0.698; p = 0.010; YI 0.26; sensitivity 0.51; specificity 0.76) (Fig. 2A). GDF-15 (Fig. 2B) provided a cut-off value of 1117.54 pg/mL (AUC 0.624; 95% CI 0.541–0.708; p = 0.005; YI 0.23; sensitivity 0.48; specificity 0.75), IGF-BP2 (Fig. 2D) a cut-off value of 107028.43 pg/mL (AUC 0.643; 95% CI 0.520–0.766; p = 0.029; YI 0.27; sensitivity 0.85; specificity 0.42), suPAR (Fig. 2E) a cut-off value of 3782.84 pg/mL (AUC 0.605; 95% CI 0.520–0.690; p = 0.018; YI 0.24; sensitivity 0.57; specificity 0.67), BNP (Fig. 2F) a cut-off value of 2248.00 pg/mL (AUC 0.692; 95% CI 0.611–0.773; p 0.001; YI 0.36; sensitivity 0.55; specificity 0.82) and cTnI (not graphically shown) a cut-off value of 20.50 pg/mL (AUC 0.704; 95% CI 0.584–0.825; p = 0.002; YI 0.36; sensitivity 0.72; specificity 0.65). Merely the determined cut-off value of H-FABP (Fig. 2C) with 1.44 did not provide a significant result with a p = 0.070. All in all, the investigated biomarkers showed a rather moderate sensitivity and specificity, with BNP and cTnI performing best in a single biomarker determination.

Pearson’s correlation analysis between sPAP and the corresponding biomarkers sST2, GDF-15, H-FABP, IGF-BP2, suPAR and BNP is shown in Fig. 3. Pearson’s correlation coefficient (r) was used to describe potential associations.

Fig. 3.

Correlation of sPAP and biomarkers. Correlation analyses between sPAP and cardiovascular biomarkers of sST2 (A), GDF-15 (B), H-FABP (C), IGF-BP2 (D), suPAR (E) and BNP (F).

Correlation analysis revealed a significant, but moderate linear relationship between sPAP and BNP (r: 0.367; p 0.001) or between sPAP and IGF-BP2 (r: 0.324; p = 0.003) and a significant, but weak linear relationship between sPAP and GDF-15 (r: 0.204; p = 0.014). The other biomarkers such as sST2 (r: 0.117; p = 0.157), H-FABP (r: 0.070; p = 0.406), suPAR (r: 0.126; p = 0.129) and cTnI (r: 0.174; p = 0.115; not graphically demonstrated) showed no significant correlations.

Kaplan-Maier curves were performed with regard to 1-year survival in dependence of severity of sPAP (Fig. 4).

Fig. 4.

Kaplan-Meier curves for detection of 1-year survival in dependence of several risk groups. I: sPAP 40 mmHg; II: sPAP 40–59 mmHg; III: sPAP $\ge$60 mmHg; Log-rank test I vs. II: p = 0.123; I vs. III: p = 0.007; II vs. III: p = 0.154.

Patients with an sPAP 40 mmHg were classified in group I and, according to current studies, show a generally low risk for the presence of PH. 17.6% of these patients died within one year. In contrast, 28.4% deaths occurred in patients with an sPAP of 40–59 mmHg and thus an intermediate risk for PH (group II). Regarding group I and group II, the log-rank test showed no significant difference with p = 0.123. Patients with an sPAP $\ge$60 mmHg and a pronounced risk for the presence of PH (group III) showed a mortality rate of 41.4% and therefore a significant difference (p = 0.007) to group I. In contrast, no relevant significance (p = 0.154) could be detected between group II and III.

To investigate several influencing variables concerning 1-year mortality after TAVR, a univariate and multivariate Cox proportional hazard regression was presented (Table 2).

The result of univariate analyses showed agreement (p 0.100) with echocardiographic data (diameter of interventricular septum, sPAP and mitral insufficiency $\ge$II°), with concomitant disease (cardiovascular disease with 3 vessels) and with laboratory data (cTnI and IGF-BP2). After inclusion of these data in a multivariate analysis, cTnI (p = 0.003) and IGF-BP2 (0.008) remained as independent factors for increased mortality.