This is a retrospective analysis based on data from the Bonn registry, a prospective and consecutive data collection at the University Hospital of Bonn. All procedures were conducted in accordance with the Declaration of Helsinki and its amendments. Study participants comprised patients undergoing TTVR for the treatment of symptomatic TR from June 2015 to July 2021. All patients were deemed as being either ineligible or high risk for surgery. A standardized diagnostic workup was performed, including transesophageal echocardiography and left- and right-heart catheterization. In the present analysis, we included patients with sufficient information for the assessment of RV–PA coupling.

Doppler echocardiography was performed according to the current guidelines [5, 6]. TR severity was assessed using standard two-dimensional color Doppler methods integrating quantitative, semi-quantitative, and qualitative parameters according to the most recent guidelines. The TR jet was used to estimate TR peak gradient using the Bernoulli equation. The right atrial (RA) pressure was estimated by its size and collapsibility of the inferior vena cava. ePASP was then obtained by adding these two measures. TAPSE was measured as the peak excursion of the tricuspid annulus during the whole cardiac cycle in the apical four-chamber view. RV fractional area change (RVFAC) was calculated as (RV end-diastolic area − RV end-systolic area)/RV end-diastolic area × 100.

We calculated an RV–PA coupling ratio by dividing TAPSE by PASP at baseline. ePASP and iPASP were applied to the equation as a PA component (i.e., TAPSE/ePASP, TAPSE/iPASP). We also assessed RVFAC/ePASP and RVFAC/iPASP in sensitivity analyses. iPASP was obtained through a right heart catheter at baseline.

The primary outcome was defined as a composite of mortality and rehospitalization due to heart failure within one year after TTVR. Each outcome was also assessed individually. We obtained clinical follow-up data through interviews at scheduled outpatient clinic visits, telephone interviews with their family, or documentation from the general practitioners.

Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Continuous variables are reported as the mean ± standard deviation (SD) or medians with interquartile ranges (IQRs). T-tests are applied for normally-distributed variables and Mann–Whitney U test is conducted for non-normally distributed variables. Categorical variables are presented as numbers and percentages. Chi-square test is applied for the comparison. A test for Pearson’s correlation was conducted to assess the correlation of ePASP and iPASP, whereas a test for Spearman’s correlation was used for echocardiographic and invasive RA pressure measurements.

Given that the primary outcome is time-to-event data, we conducted a receiver operating characteristics (ROC) analysis using a Cox proportional model. The predictive accuracy of the outcome was evaluated using Harrell’s concordance index (Harrell’s C) [7, 8]. A smooth spline curve for hazard ratio (HR) of RV–PA coupling was depicted. The study participants were divided according to the quartiles of RV–PA coupling parameters. The Kaplan–Meier method was used for survival time analyses. A Cox proportional hazard model was conducted to assess the association of RV–PA coupling parameters with outcomes. Age, sex, EuroSCORE II, estimated glomerular filtration rate (eGFR), left ventricular (LV) ejection fraction, and chronic obstructive pulmonary disease were accounted in the model 1, while reduction in TR ≤ moderate at discharge was incorporated in the model 2 to adjust for the association [1, 3, 4, 9]. Furthermore, the association was tested among predefined subgroups (i.e., NYHA I/II or NYHA III/IV, LV ejection fraction ≤ 50% or > 50%, TR severity at baseline ≤ severe or ≥ massive, TR severity at discharge ≤ moderate or ≥ severe). All statistical analyses were performed using EZR version 1.37 (Saitama Medical Center, Jichi Medical University, Saitama, Japan), R version 3.5.2 (R Foundation for Statistical Computing), or IBM SPSS Statistics 26 (IBM Corporation, New York, USA).

During the study period, a total of 206 patients met the inclusion criteria and were entered in the analysis. The mean age was 78.5 ± 7.1 years, and 58.3% were female. The study participants were at a high risk for surgery (EuroSCORE II: 9.0 ± 6.5%), and massive/torrential TR was observed in 100 (48.5%) of these patients. The majority of patients presented with atrial fibrillation (94.2%), preserved left ventricular ejection fraction (55.2 ± 10.1%), and a secondary etiology of TR (96.1%). All patients were treated either with transcatheter edge-to-edge repair (n = 185: 89.8%) or transcatheter annuloplasty (n = 21: 10.2%). No periprocedural death was observed. TTVR reduced the severity of TR (p < 0.001), with 78.0% of patients having TR 2+ or less at discharge.

Variables regarding echocardiography and right heart catheterization are listed in Table 1. The median values of TAPSE, ePASP, and TAPSE/ePASP were 17.0 mm (IQR 15.0, 20.0 mm), 45.0 mmHg (IQR 36.3, 53.0 mmHg), and 0.392 mm/mmHg (IQR 0.287, 0.514 mm/mmHg). The median time from baseline echocardiography to invasive right heart catheterization was one day (IQR − 1, 5 days). The invasively-measured PASP (i.e., iPASP) was 42.0 mmHg (IQR 35.0, 52.0 mmHg), and the median of TAPSE/iPASP was 0.408 mm/mmHg (IQR 0.316, 0.526 mm/mmHg).

With the median follow-up duration of 201 days (IQR 98–424 days), 29 patients had died, and 40 patients had experienced rehospitalization due to heart failure, resulting in 57 cases of the primary outcome. The associations of right heart parameters with the primary outcome are listed in Supplemental Table 1. Compared to TAPSE/ePASP, TAPSE/iPASP showed better predictability for the primary outcome (Fig. 1A): the c-statistics was 0.565 (95% CI 0.488–0.643) for TAPSE/ePASP and increased to 0.695 (95% CI 0.631–0.759) when iPASP was applied to the formula (i.e., TAPSE/iPASP) (p < 0.001). The trend was consistent for RV–PA coupling using RVFAC (Fig. 1B). RVFAC itself was not associated with the primary outcome (Supplemental Table 1). Similarly, RVFAC/ePASP was not predictive for the outcome, while RVFAC/iPASP predicted the primary endpoint. As shown in Fig. 1, integrating the iPASP measurement led to a better prediction for the outcome (c-statistics: 0.495 for RVFAC/ePASP [95% CI 0.411–0.578]; c-statistics for RVFAC/iPASP: 0.643 [95% CI 0.566–0.719]).

In addition, we separately conducted ROC analyses in patients with severe and those with massive/torrential TR. TAPSE/iPASP showed a better discrimination irrespective of the TR severity, whereas the difference in Harrell’s C was numerically higher in patients with more advanced TR (delta c-statistics: 0.102 ± 0.035 in severe TR, p = 0.003; delta c-statistics: 0.179 ± 0.056 in massive/torrential TR, p = 0.001).

There was a significant correlation between ePASP and iPASP in the total cohort (correlation coefficient 0.50, p < 0.001), whereas the correlation was attenuated in patients with massive to torrential TR (severe TR: correlation coefficient 0.588, p < 0.001; massive to torrential TR: correlation coefficient 0.337, p = 0.001; interaction p = 0.01) (Fig. 2). A similar trend was observed in the correlation between TR pressure gradient and iPASP (severe TR: correlation coefficient 0.546, p < 0.001; massive to torrential TR: correlation coefficient 0.331, p = 0.001; interaction p = 0.035).

In addition, the echocardiographic RA pressure (median 15 mmHg [IQR 8–15 mmHg]) was modestly correlated with the invasive measurement (median 13 mmHg [IQR 9–18 mmHg]) (correlation coefficient 0.165, p value = 0.023) (Fig. 3). The RA pressure was likely to be underestimated in patients with the echocardiographic measurement of 3 and 8 mmHg. The miscalculation was also observed in patients with the echocardiographic RA pressure of 15 mmHg: a quarter of patients showed less than 10 mmHg of the invasive measurement, whereas another quarter showed more than 20 mmHg of the invasively-measured RA pressure.

We divided the study participants according to TAPSE/iPASP (quartile 1: ≤ 0.316; quartile 2: 0.317–0.407, quartile 3: 0.408–0.526, quartile 4: ≥ 0.527). The baseline characteristics are listed in Table 1. Patients with lower TAPSE/iPASP were more likely to be male, having prior coronary artery bypass graft surgery, higher PA resistance and RA pressure, compared to patients with higher TAPSE/iPASP mm/mmHg. TR reduction to ≤ 2+ at discharge was consistently observed regardless of TAPSE/iPASP (Supplemental Table 2).

The spline curve depicts the linear association of TAPSE/iPASP with the primary endpoint (Fig. 4). TAPSE/iPASP was inversely associated with the primary outcome (per 0.1-point increase: adjusted-HR 0.67, 95% CI 0.56–0.82, p < 0.001) in the multivariable Cox proportional model (Table 2). Other parameters associated with the outcome were sex (male: adjusted-HR 2.26, 95% CI 1.21–4.20, p = 0.01) and eGFR (per 1 ml/min/1.73 m² increase: adjusted-HR 0.98, 95% CI 0.97–0.99, p = 0.008).

The association of TAPSE/iPASP with the outcome remained significant after adjusting for the TR reduction (Table 2). Also, the observed association was consistent among the predefined subgroups, including NYHA functional class, LV ejection fraction, the severity of TR at baseline, and the reduction in TR at discharge (Supplemental Fig. 1).

Event-free survival curves were estimated and depicted using a Kaplan–Meier method (Fig. 5). According to the TAPSE/iPASP quarters (i.e., ≤ 0.316; 0.317–0.407; 0.408–0.526; ≥ 0.527), the event-free survivals were 43.4%, 48.3%, 77.9%, and 85.4% at one year after TTVR. Similar findings were observed for each outcome. The event-free survival from heart failure rehospitalization were 52.0%, 61.8%, 86.4%, and 87.3%, whereas survival from death were 65.0%, 82.1%, 85.9%, and 90.6% at one-year follow-up after TTVR.

Several limitations to our study should be acknowledged. First, the single-center, retrospective nature of this study, and the small sample size may limit the generalizability of the results. Nevertheless, the present study cohort represents data from one of the high-volume centers for TTVR. Second, no follow-up data on RV–PA coupling using iPASP was available. An early change in RV–PA coupling was reported in the most recent study, which needs to be validated in future studies using invasive measurements of right heart pressures. Also, factors associated with a change in RV–PA coupling following TTVR should be elucidated in further investigations. Third, the study cohort was regarded as one homogenous group, but we did not assess more detailed etiology of TR (i.e., atrial or ventricular secondary TR, cardiac implantable electronic device lead-induced TR). Fourth, since two different PASP values (echocardiographic and invasive) were measured at different times, patients might have been subjected to different hemodynamic influences. Finally, TAPSE reflects the longitudinal RV function but does not account for global RV function or might overestimate RV function in the presence of TR [10]. Nonetheless, the current analysis showed comparable c-statistics between the RV–PA coupling using TAPSE and RVFAC, which implies their feasibility in clinical practice for risk stratification in patients with TR. Magnetic resonance tomography might be more accurate for assessing function [10]. Nevertheless, impaired renal function or implanted pacemaker lead may limit its clinical use or feasibility of measurements. In contrast, TAPSE is the most commonly used and easily reproducible marker of RV function in clinical practice. Therefore, TAPSE/iPASP can be generalizable in each individual and each hospital.