The retrospective multi-centre study compared percutaneous annuloplasty and T-TEER of the tricuspid valve in a real-world cohort. Patients with an age ≥18 years, high surgical risk and symptomatic severe to torrential TR of different pathologies and despite optimal medical therapy underwent either Cardioband® implantation (Edwards Lifesciences, Irvine, CA, USA) or T-TEER (TriClip®, Abbott, Chicago, Illinois, USA, or PASCAL®, Edwards Lifesciences, Irvine, California, USA) from 2019 to 2022. Patients with combined procedures of the tricuspid and mitral valve were excluded from the analysis. In case of two sequential TR procedures, only the first intervention was evaluated. The study was approved by the institutional ethics committee of the Charité – Universitätsmedizin Berlin, Germany (EA1/005/23).

Interventional therapy of TR was recommended by local heart team consisting of interventional cardiologists, cardiovascular surgeons, anesthetists, heart failure specialists, and imaging experts. The appropriate device was selected based on individual tricuspid anatomy and function according to current recommendations (5, 7). Patients with a systolic pulmonary artery pressure ≥75 mmHg measured by right heart catheterization did not undergo an interventional therapy. Evaluation of tricuspid regurgitation and its interventional therapy comprised an echocardiographic assessment, right and left heart catheterization as well as computed tomography in line with the guidelines of the ESC and EACTS (3). Echocardiography was performed according to the standards of the American Society of Echocardiography (ASE) and the European Association of Cardiovascular Imaging (EACVI) using a GE healthcare Vivid E9 or E95 with a M5S(c) 1.5–4.5 MHz or 4Vc-D 3D/4D Phased Array 1.4–5.2 MHz transducer (GE Vingmed, Horton, Norway) (11, 12). TR was quantified by assessment of the effective regurgitant orifice area (EROA) and regurgitant volume (RegVol) using the proximal isovelocity surface area (PISA) method, vena contracta measurements, as well as hepatic vein reflux. TR severity was graded into five categories as proposed by Hahn et al. (13). TOE or intracardiac ultrasound was used for guidance of the interventional therapy. All procedures were performed as described previously under general anesthesia (8, 10, 14).

Outcome parameters were the following procedural characteristics: technical and device success and complications based on the Mitral Valve Academic Research Consortium (MVARC) criteria (15) as well as TR reduction and learning curves, i.e., changes of outcome and procedural characteristics over time. Technical and device success were combined and defined as successful implantation of the device and retrieval of the delivery system without conversion to open heart surgery or other interventional therapy as well as TR reduction of at least one grade at the time of leaving the operating room (adapted from MVARC criteria). TR grade was quantified at the end of the procedure and at discharge.

SPSS Statistics version 28 for Windows (IBM Corporation, New York, NY, USA) was used for data analysis. Categorical and ordinal variables were listed in percentages and were analysed by chi-squared test or Mann-Whitney U-test, respectively. Continuous variables were presented as median with 25th and 75th percentile or mean with standard deviation depending on their skewness and uniform per variable for a better inter- and intragroup comparison. Continuous variables were assessed by t-test (normally distributed parameters) and Wilcoxon-test or Mann-Whitney U-test (not normally distributed parameters). The impact of TR reduction on right ventricular function and learning curves were evaluated by linear regression analysis. TR improvement (difference between TR grade after intervention and at baseline) was assumed to be a continuous variable. A p-value of <0.05 was considered statistically significant.

Technical and device success was achieved in 97% of patients in the T-TEER group and 98% of patients in the annuloplasty group (p = 0.502). Reasons for device failure were two single leaflet device attachments (SLDA) after T-TEER. Two additional SLDA occurred during the intra-hospital stay. One patient with planned Cardioband® implantation developed cardiac arrest due to iatrogenic coronary dissection after engagement of the right coronary artery, necessitating cardiopulmonary resuscitation. Therefore, the Cardioband® implantation was aborted. We did not include this case in the further analysis of procedural characteristic and durable repair. None of the patients died during the procedure. One patient of the annuloplasty group suffered from a major vascular access complication, a dissection of the left iliac artery due to pronounced kinking, treated by stent implantation. The implantation of the Cardioband® was performed during a second procedure without further complications.

Procedure time and duration of radiation exposure were significantly shorter in the T-TEER group compared to the annuloplasty group (T-TEER: 102 [73–122] min and 10 [7–15] min vs. annuloplasty: 165 [143–236] min and 61 [44–74] min, p < 0.001; Figures 2, 3). On average, 1.9 ± 0.6 devices per patient were implanted in the T-TEER group, mostly into the anteroseptal, followed by the posteroseptal commissure. Annuloplasty devices comprised the sizes C to F. On average, 121 ± 62 ml of contrast were used for Cardioband® implantation compared to none for T-TEER. Annuloplasty patients presented significantly more often a new or progressive but hemodynamically not relevant pericardial effusion compared to patients after T-TEER. Moreover, four cases of transient bradycardia and three incidents of right coronary artery injuries requiring intervention occurred during percutaneous annuloplasty. Detailed procedural complications are listed in Table 4.

TR was reduced at least one grade in 100% of patients, with 91% of T-TEER patients and 80% of annuloplasty patients achieving a TR grade ≤2 post-intervention. The mean reduction was 2.4 ± 0.8 grades after T-TEER and 2.5 ± 0.8 grades after annuloplasty (Figure 4). Moreover, we observed a significant decrease in annulus diameter in both groups after the procedure, which was more pronounced after annuloplasty (Figure 5). Grading of TR severity by transthoracic echocardiography at discharge was consistent with the intraprocedural assessment at the end of the intervention in 69% of patients in the T-TEER group and in 64% of patients in the annuloplasty group (p = 0.650 for comparison of both groups).

In both groups, we observed a significant decrease of right ventricular function as assessed by right ventricular fractional area change (RVFAC) and a significant reduction of right atrial area, while tricuspid annular plane systolic excursion (TAPSE) was only reduced after annuloplasty (Table 3). Moreover, we detected a significant impact of TR reduction on RVFAC after T-TEER, but not after annuloplasty (linear regression analysis for T-TEER: F-value 7.7, p = 0.008, regression coefficient 46.1, p < 0.001).

The procedure time of annuloplasty showed a significant reduction throughout the study period. There was neither a relevant effect of a learning curve on TR reduction in both groups nor on procedure time in the T-TEER group.

TR reduction of at least one grade in 100% of patients in our study cohort is comparable to results of previous T-TEER and annuloplasty trials (8, 14, 16, 17). Device selection was based on tricuspid anatomy and function assessed by echocardiography, computed tomography, and right heart catheterization (5, 7, 8, 16, 17). Figure 6 illustrates our algorithm for device selection, which aligns with similar recommendations published in the literature (5). Although percutaneous annuloplasty is primarily applied in annulus dilatation in atrial TR, TR etiologies did not differ significantly between both groups and annuloplasty as well as T-TEER resulted in a significantly reduced tricuspid annular diameter. Interestingly, the algorithm is not well reflected in the baseline measurements which in fact showed a greater tenting height in the annuloplasty group and a wider coaptation gap in the T-TEER group. The reason for this apparent discrepancy is that a substantial number of cases (n = 10) could not be treated with the preferred device—either due to RCA proximity (for annuloplasty) or due to gap size (for T-TEER). This resulted in patients with large gaps (often associated with larger tenting height) being treated by annuloplasty especially in the early phase of the study period. Similarly, patients with suitable valve anatomy for annuloplasty had to undergo T-TEER when RCA proximity made annuloplasty not feasible. With the availability of larger edge-to-edge devices, e.g., TriClip G4 and PASCAL Ace, we observed a reduction of tenting height in the annuloplasty group and an increase of baseline gap size in T-TEER patients. In both groups, patients had right ventricular leads. Lurz et al. already reported comparable efficacy of T-TEER in patients with and without right ventricular leads (7, 18). However, TR improvement was more pronounced in patients with a commissural or central lead compared to a lead in contact to the leaflet body (18). In line with the results by Lurz et al., we also observe no significant difference in TR reduction in patients with or without right ventricular leads in both groups. We observed a decrease in right ventricular function, as measured by TAPSE, after interventional therapy, possibly driven by reduction of regurgitation volume in both groups and the annuloplasty device itself. Therefore, we recommend further follow-up visits including assessment of functional status, medical therapy, and echocardiography. This is necessary to prevent the potential oversight of clinical deterioration associated with an afterload mismatch and the decrease in right ventricular function. At baseline, the majority of our patients were either receiving diuretic treatment or undergoing dialysis in addition to heart failure therapy (Table 1). It is advisable to adjust diuretic therapy depending on venous congestion during the follow-up visits.

Intermittent bradycardia occurred only in the annuloplasty group, most likely due to mechanical compression on the conduction system in the region of Koch's triangle (19). While the observed bradycardia was usually self-limiting and of short duration, operators should be aware of this possible complication occurring during placement of the last anchors. Compression or injury of the right coronary artery (RCA) requiring percutaneous coronary intervention after anchor placement (5% of our patients) was mostly caused by severe proximity of the RCA to the tricuspid annulus. In addition, we found new or progressive pericardial effusions in 27% of patients after percutaneous annuloplasty, which were most likely due to pericardial irritation resulting from anchor insertion. The pericardial effusions were usually constant until discharge and had no clinical relevance.

In the T-TEER cohort, four SLDA occurred during the procedure (n = 2) and up to discharge (n = 2) between the posterior and septal (n = 2) as well as the anterior and septal leaflets (n = 2). All T-TEER patients with SLDA presented a massive TR with degenerative and short leaflets as well as a larger tricuspid annulus diameter at baseline (45 [44–49] mm) compared to those without SLDA (40 [38–44] mm). However, the tenting height and coaptation gaps were comparable between the two groups, measuring approximately 4 ± 3 mm (n = 4) and ranging from 2 to 5 mm (n = 3) in the SLDA group, respectively. All patients underwent a re-intervention (T-TEER) resulting in a moderate (n = 1), severe (n = 2) or massive TR (n = 1) at discharge. Based on these findings, we propose that the combination of shortened and degenerative leaflets, along with excessive annular dilation, may represent potential risk factors for the development of SLDAs in T-TEER. SLDA is a well-known complication in tricuspid and mitral edge-to-edge repair (20–22). Known risk factors comprise a big coaptation gap between both target leaflets, a severe tethering resulting in a small leaflet insertion and a high mechanical stress on the target leaflet caused by the devices as well as the leaflet tissue itself (20, 22, 23). Current strategies to treat residual TR include stabilization of the device by implanting another device or, in case of no available landing zone, the implantation of a cava valve (24, 25).

We observed a significant learning curve only in the annuloplasty group in terms of procedural duration. Specifically, the first studies of Cardioband® implantation showed a mean procedure time of 4.2 h (8), which was significantly longer than the mean procedure time of 2.8 h in our cohort. This suggests that further development of annuloplasty devices, the use of intracardiac ultrasound and increasing experience of interventional cardiologists may enable further reduction in procedure times. However, it is likely that percutaneous annuloplasty will continue to take longer than T-TEER, as it involves significantly more procedural steps. In the T-TEER group, the duration of procedures also tended to get shorter during the study period, but this did not reach statistical significance. There was no relevant change in the number of implanted devices over time. No further improvements in terms of a learning curve were observed during the study period, in particular no improvement in TR reduction in either group.

In summary, the following aspects might be considered when selecting patients for TR procedure beyond anatomical characteristics: The longer procedure time of percutaneous annuloplasty results in longer duration of general anesthesia and mechanical ventilation, which might increase complications such as pneumonia due to microaspiration and prolonged ventilation as well as delirium in older patients. The implantation requires the use of contrast, which should be taken into account in patients with renal failure including patients with cardiorenal syndrome. However, more severe TRs were treated successfully with annuloplasty compared to T-TEER. T-TEER might be challenging in patients with a big coaptation gap leading to an increased risk of SLDA. Nevertheless, percutaneous annuloplasty and T-TEER resulted in a comparable TR reduction in our study cohort. Long-term data for both devices revealed a durable TR improvement, which was associated with symptoms relief and reduction of mortality and heart failure hospitalizations (19, 26–28). However, prospective, randomized trials are needed to better understand pros and cons of both devices and its use in different TR etiologies.

The present study utilized a retrospective design, and patients were not randomized to specific treatment options. Device selection was based on the TR anatomy. As a result, baseline characteristics showed differences in three aspects: presence of CIEDs, bilirubin levels, and peripheral artery disease. These differences reflect real-world data. Therefore, the present study serves as a hypothesis-generating investigation for further randomized-controlled trials.