In this monocentric study, we included 75 patients who underwent either direct transcatheter annuloplasty using Cardioband (Edwards Lifesciences, Irvine, USA), 55 (73%), or transcatheter edge-to-edge repair using TriClip (Abbott Laboratories, Chicago, USA) or Pascal (Edwards Lifesciences, Irvine, USA), 20 (27%), between May 2020 and June 2023. Patients suffered from severe TR and were deemed unsuitable for cardiac surgery by the heart team. The study was approved and overseen by the Local Ethics Committee, and informed consent for retrospective study inclusion was waived (File number: 2022-945).

All patients underwent an electrocardiographically gated, contrast-enhanced CT with full cardiac cycle acquisition using a 320-detector row system (Canon Medical Systems, Japan). Contrast administration was tailored to optimize homogenous right heart enhancement. Axial thin slice (0.5 mm) images were reconstructed in 10% increments throughout the cardiac cycle. All scans exhibited sufficient right-sided contrast opacification for image analysis.

CT data sets were analyzed using heart.ai (LARALAB GmbH, Munich, Germany). heart.ai was used to quantify the right heart morphology for each reconstruction phase throughout the cardiac cycle.

The end-systolic and end-diastolic phases were identified as the phases with minimum and maximum RV volumes. For the CT-based anterior, posterior, septal, and lateral TAPSE measurements, the distance between the systolic and diastolic positions of the tricuspid annulus was measured at the respective locations of the annulus. For TAPSE volume, the annulus of the right ventricle was extracted in each phase. Using the least squares regression method, the best-fit plane was computed for each annulus. The annulus points were projected onto their respective best-fit planes. The annular areas A1 and A2 were calculated using the shoelace formula. The orthogonal distance D between the center of mass of the annulus (An1) and the best-fit plane of the annulus (An2) was evaluated. This distance quantifies the displacement between the two annuli. The TAPSE volume was estimated as D * (A1 + A2)/2.

All patients received transthoracic echocardiograms within 24–48 h prior to the intervention, with assessment following current American Society of Echocardiography (ASE) guidelines (14). Transthoracic echocardiography (TTE)-based TAPSE was measured using M-mode in the apical four-chamber view as the distance between the most apical and basal position of the tricuspid annulus during one cardiac cycle. The severity of TR was graded according to the five-grade scheme proposed by Hahn and Zamorano (15).

Interbeat changes of TAPSE, fractional area change (FAC), RV mid-cavity diameter (RVMID), and left ventricular (LV) ejection fraction (LVEF) were analyzed in transthoracic echocardiography (TTE) using subsequent beats captured in one series. For evaluation of the dependency of TAPSE on body surface area (BSA), TTEs from 30 healthy subjects were analyzed.

Patients received either transcatheter edge-to-edge repair or direct transcatheter annuloplasty. The interventions were performed under general anesthesia with interventional guidance by transesophageal echocardiography (TEE) and fluoroscopy via femoral vein access. Technical success was defined according to TVARC as successful delivery, deployment, and positioning of the device, absence of procedural mortality, and freedom from emergency surgery related to the device (16).

Patients were followed up in our outpatient clinic.

The primary endpoint was a composite of death from any cause or rehospitalization for heart failure (HF). Mortality and HF hospitalizations were assessed by reviewing in-hospital data or by contacting treating general physicians after the index procedure.

Statistical analysis was performed using SPSS Statistics 27 provided by IBM. Continuous and normally distributed variables were expressed as mean ± standard deviation. Receiver operating characteristic (ROC) was performed to determine optimal cutoff values for TAPSE for outcome analysis. Kaplan–Meier analysis was performed for clinical outcomes using time to first event. Differences in time-to-event distributions were evaluated using the log-rank test. The fazard ratio (HR) was calculated using Cox regression. Univariate analyses were initially used, and all parameters with p < 0.1 were then included in the multivariate analysis. A p-value of <0.05 was considered significant.

Of the 75 enrolled patients, 55 patients (73%) received direct transcatheter annuloplasty, and 20 (27%) received edge-to-edge repair using TriClip (Abbott Laboratories, Chicago, USA) or Pascal (Edwards Lifesciences, Irvine, USA).

The mean patient age was 77 ± 8 years, and 61% were female. The mean EuroScore II was 5.4 ± 4.5%, and most patients were severely symptomatic, presenting in New York Heart Association (NYHA) class ≥3 in 88% (66) of cases. A total of 56 patients (75%) had been previously hospitalized for right-sided heart failure and mean N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels at baseline were 2,850 ± 2,393 ng/L. The most common comorbidities included atrial fibrillation (77%) and heart failure with preserved ejection fraction (68%) (Table 1).

The mean LVEF was 54 ± 9%. The mean TAPSE was 19 ± 6 mm and was reduced (defined as TAPSE as <17 mm) in 27 cases (36%). The mean vena contracta was 13.4 ± 5.2 mm. Tricuspid regurgitation was graded severe in 32 (43%), massive in 19 (25%), and torrential in 24 (32%) patients (Table 2).

TTE data were analyzed for interbeat changes of TAPSE, FAC, RVMID, and LVEF which is displayed in Supplementary Table S1. Compared to the mean change of TTE-based TAPSE or LVEF, the mean changes of FAC and RVMID were significantly larger. There was no significant difference in relative change of LVEF compared to TAPSE.

To analyze whether TAPSE is dependent on BSA, we analyzed 30 TTEs from healthy subjects without clinical or echocardiographic RV dysfunction. We found a moderate correlation between TAPSE and BSA (Pearson’s correlation R = 0.535, P = 0.002).

For further analysis of tricuspid annular excursion, septal, anterior, posterior, and lateral TAPSE was analyzed in CT (Figure 1). The mean septal CT-TAPSE was 15 ± 10 mm, and the lateral CT-TAPSE was 15 ± 11 mm. The anterior TAPSE was 13 ± 6 mm, and the posterior TAPSE was 12 ± 6 mm (Table 3, Supplementary Table S2). There were no significant differences between the TAPSE subgroups in the overall study population. The mean volume of TAPSE between systole and diastole was 13.7 ± 5.6 ml.

We found a moderate correlation between TTE-measured TAPSE and CT-TAPSE which was best for lateral TAPSE (R for Pearson correlation for lateral CT-TAPSE 0.363, p < 0.001). Since we found a significant correlation between TTE-TAPSE with BSA, the index values for CT-TAPSE (iTAPSE) were used for further evaluation. Linear regression analysis showed that iTAPSE volume and right ventricular ejection fraction (RVEF) were more closely associated with anterior and posterior iTAPSE than with septal and lateral iTAPSE (Supplementary Table S3).

Procedure time and fluoroscopy time were not different in patients with posterior iTAPSE >4.5 mm/m² and <4.5 mm/m² (all P > 0.05; Table 4). Compared to patients receiving TEER, patients receiving direct annuloplasty had significantly longer procedure times (187 ± 62 vs. 114 ± 64 min; P < 0.001) and fluoroscopy times (52 ± 17 vs. 19 ± 10 min; P < 0.01).

The mean follow-up was 193 ± 146 days. During the follow-up time, 18 patients reached the primary endpoint, of whom 11 patients were hospitalized for heart failure and 7 patients died (Table 4).

The receiver operator curve showed the largest area under the curve for the primary outcome for posterior iTAPSE (AUC 0.72) and iTAPSE volume (AUC 0.72, Supplementary Figure 2). For outcome prediction, a posterior iTAPSE of 4.5 mm/m² and an iTAPSE volume of 9 ml/m² performed best.

At 1 year, the estimated combined endpoint was observed significantly less frequently in patients with posterior iTAPSE >4.5 mm/m² (17.2% vs. 63.6%; HR 0.225, CI 0.087–0.581; P < 0.001; Figure 2). Furthermore, an iTAPSE volume of >9 ml/m² was also associated with better outcome (16.4% vs. 57.1%; HR: 0.269 CI 0.105–0.686; P = 0.003).

Then we divided the cohort into three subgroups. In the first group, patients had a preserved RVEF and preserved posterior iTAPSE (defined as >4.5 mm/m²). In the second group, patients had either a preserved posterior iTAPSE or a preserved RVEF, and in the third group, there were patients with reduced TAPSE and reduced RVEF. We found that patients with impaired RVEF and iTAPSE and patients who had either preserved RVEF or iTAPSE had comparable outcomes (Supplementary Figure S2). In this subgroup, eight patients had reduced RVEF but preserved iTAPSE, and six patients had reduced iTAPSE but preserved RVEF.

Univariate analysis showed that echocardiographic-measured LVEF and artificial intelligence-measured CT RVEF, posterior iTAPSE, and TAPSE volume were associated with the primary outcome at 1 year. On multivariate Cox regression, posterior iTAPSE remained a significant predictor for outcomes (p = 0.022, Table 5). Both TTE-measured TAPSE (HR 1.013, CI 0.918–1.117; p = 0.803) and CT-measured lateral TAPSE (HR 0.919, CI 0.837–1,009; p = 0.075) failed in predicting outcomes at 1 year.

The prevalence of atrial fibrillation was 77% in our cohort. When analyzing only patients with atrial fibrillation, we found that posterior iTAPSE (HR 0.647 CI 0.475–0.881; p = 0.006) but not RVEF (HR 0.955 CI 0.9–1.012; p = 0.12), both evaluated as continuous variable, were significantly associated with the combined endpoint.