Monitoring of mitral- and tricuspid valve interventions with CardioMEMS: Insights beyond imaging

Background: Mitral- and tricuspid regurgitation are associated with significant


| INTRODUCTION
Mitral-and tricuspid regurgitation (TR) are common and associated with significant morbidity and mortality. [1][2][3][4][5] Following recent guideline recommendations, 6,7 catheterbased interventions on both the mitral-and the tricuspid valve are increasingly performed, including edge-to-edge repair (mitral: TMVR, tricuspid: TTVR) or bicaval valve implantation for TR (bi-CAVI). All of these approaches have been shown to improve outcome. [8][9][10][11][12] However, patient selection and assessment of procedural success remain a major challenge, particularly in tricuspid interventions. Echocardiographic quantification of the extent of residual TR after TTVR is challenging but remains the only established parameter to assess procedural success. However, residual TR severity strongly depends on loading conditions and afterload, thus, accurate assessment of procedural success by imaging at certain time points may be difficult, if not impossible. 13 Conventional invasive hemodynamic assessment requires repeat hospitalization and, following TTVR, is associated with the same risk, as the edge-to-edge repaired tricuspid valve has to be passed with the Swan-Ganz catheter.
The CardioMEMS sensor (Abbott), which is implanted in the pulmonary artery, allows continuous monitoring of cardiac output (CO) and pulmonary artery pressures (PAP).
Based on these measurements, interventional success beyond imaging and conventional biomarkers is feasible.
The present patient series aims to elaborate on the hemodynamic changes in patients after TTVR, bi-CAVI or TMVR, using CardioMEMS. This study aimed to elaborate (1) the additional value of CardioMEMS on top of echocardiography, (2) investigate the hemodynamic changes post intervention, and (3) determine the difference in length of hospital stay and events between patients with CardioMEMS and a control group. This study should be regarded as a proof-of-concept study.  Figure 1, Panel D) were recruited and compared to a control group without CardioMEMS sensor. All patients provided written informed consent, and the local ethics committee approved for a deeper understanding of the prerequisites for optimal patient selection and management for catheter-based interventions.

K E Y W O R D S
cardiac output, mitral-and tricuspid regurgitation, pulmonary hypertension, valvular diseases the study. Echocardiography was performed approximately 1 month prior to the procedure. Outpatient follow-up visits, including a 6-minutes walk test (6MWT) and echocardiography assessment, were performed 3-12 months after the intervention. CardioMEMS data were obtained and analysed 1-3 months before and 3-12 months after the intervention until the patients came to the outpatient visit.

| Cardiac monitoring
The Abbott CardioMEMS pressure sensor has been developed for remote cardiac monitoring of heart failure patients. 14 Percutaneous femoral venous access was used to catheterize the left-sided pulmonary artery. A dorsal lower-lobe branch with a diameter of approximately 8 mm was chosen to place the sensor. The procedure is safe with a device-and procedure-related complication rate of 1% each as shown in a randomized controlled trial. 14 The device continuously measured and remotely transmitted PAP and CO as patients were asked to transmit data daily using their electronic bedside unit. According to the follow-up program of the heart failure unit, patients were contacted by phone when PAP increased significantly.

| Interventional valve procedure
According to established guidelines, our interdisciplinary heart team indicated TTVR, bi-CAVI, or TMVR based on each patient's specific clinical and anatomic conditions. 6,7 All procedures were performed under general anaesthesia with transesophageal echocardiographic guiding and fluoroscopy.

| Statistical analysis and outcome
We show mean and standard deviation for continuous parameters and total numbers for categorical data. Continuous parameters were tested for normal distribution using the Shapiro-Wilk-Test. In the case of normal distribution, preand post-interventional measurements were compared with Student's t-test. In the case of non-normal distribution, the Wilcoxon rank-sum test was used. In addition, we applied boxplots and line charts to show differences between preand post-interventional values. In all patients, periprocedural events (vascular complication, bleeding) and major adverse events (hospitalization for heart failure, death, reoperation) 1 year after valve intervention were assessed. We used SPSS (IBM SPSS Statistics, Version 27.0) for all analyses and set the significance level at an alpha of .05.

| RESULTS
Between January 2020 and September 2020, 36 patients were enrolled. Thirty-two patients did not receive a CardioMEMS sensor and served as the control group. In the control group, 10 patients received isolated TMVR, 10 patients isolated TTVR, 10 patients TMVR and TTVR, and two patients bi-CAVI. Four patients received a CardioMEMS sensor at an average of 20 days prior to intervention ( Table 1). One of these patients underwent TMVR, one TTVR, one TMVR and TTVR, and one bi-CAVI. Baseline characteristics are presented in Table 1.
Patient 1 underwent successful TMVR (reduction of MR from Grade 3 to Grade 1, transvalvular gradient 4 mmHg). CardioMEMS data revealed significantly decreased systolic, diastolic, and mean PAP (all p < .001) and T A B L E 1 Baseline characteristics for all patients (n = 36).  Figure 2), while echocardiography remained inconclusive (Table 2). At the 6month follow-up, 6MWT improved (Table 4). Thirteen months later, successful TTVR was performed (reduction of TR from grade 4 to grade 2, transvalvular gradient 2 mmHg). The 6MWT was worse, but the patient suffered from knee pain. On CardioMEMS, systolic PAP and CI increased (p = .001), while mean and diastolic PAP did not change significantly. The evolution of pulmonary pressures, particularly CO after TMVR and TTVR was not predictable from echocardiographic data nor NT-pro BNP serum levels (Tables 2-4). The patient was followed for a total of 16 months. Patient 2 underwent highly successful isolated TTVR, as determined by echocardiography (reduction of TR from grade 4 to grade .5, transvalvular gradient 1 mmHg).
However, diastolic and mean PAP, as well as CI, remained unchanged. Only systolic PAP increased significantly (p = .006; Table 2, Figure 2). At the 6-month follow-up, NT-pro BNP serum levels changed moderately by 10%. The patient refused 6MWT. Patient 3 underwent successful TMVR and TTVR in one procedure (reduction of MR from grade 4 to grade 1, transvalvular gradient 3 mmHg; reduction of TR from grade 5 to grade 3, transvalvular gradient 1 mmHg). TR deteriorated after 4 months to grade 5, while the TMVR result remained excellent (mild residual MR). The patients received a CardioMEMS sensor and underwent subsequent treatment of TR with bi-CAVI. Over the following 6 months, systolic, diastolic, and mean PAP continuously decreased and CI increased (all p < .001; Table 2, Figure 2), and the patient improved in terms of symptoms (NYHA III-II, Table 4). NT-pro BNP decreased by 38%.
Patient 4 presented with a history of isolated surgical TV repair due to severe TR. She had severe MR and underwent successful TMVR (reduction of MR from grade 4 to grade 1, transvalvular gradient 5 mmHg). The tricuspid valve was not suitable for TTVR for anatomical reasons after previous TR surgery. At follow-up, systolic, diastolic, and mean PAP decreased, CO and CI increased, although TR was still severe (Table 2, Figure 2). At the 6-month follow-up, NT-pro BNP significantly decreased by 71%, and 6MWT and leg edema improved.
In the control group, 10 patients were recruited with isolated TMVR, 10 patients with isolated TTVR, 10 patients with TMVR and TTVR, and two patients with bi-CAVI. Baseline characteristics are shown in Table 1, periprocedural echocardiography data in Table 3, and follow-up and hospital stay data in Table 4.

| Outcome
None of the patients implanted with CardioMEMS suffered a major adverse event or a periprocedural event.
In the TMVR group without CardioMEMS, one patient underwent reoperation.
In the TTVR group, two patients were hospitalized for heart failure, of whom one died shortly thereafter. Another patient died outside the hospital.
In the group with TMVR and TTVR, two patients were hospitalized for heart failure, and two patients died.
Among patients with bi-CAVI, one patient was hospitalized for bradycardia and received a pacemaker.
Periprocedural complications also occurred only in the control group. One patient needed a blood transfusion, and one received an arterial stentgraft at the access site.

| DISCUSSION
The present study demonstrates that hemodynamic monitoring using CardioMEMS adds important information regarding procedural success of interventional treatment of MR and TR on top of imaging and laboratory data. Successful transcatheter treatment of MR and TR resulted in the following effects as determined on CardioMEMS measurements: (1) PA pressures significantly decreased and CO increased after TMVR, (2) PA pressures decreased, and CO increased after bi-CAVI, while (3) divergent changes of PAP and CO were observed after TTVR, (4) periprocedural hospital stay was not reduced in the CardioMEMS group, and (5) events occurred more often in the control group.
The benefits of both TMVR and TTVR are still discussed controversially.
For TMVR, the randomized controlled trial COAPT showed positive effects of TMVR on outcome, while MITRA-FR did not. 12,15 Therefore, the current recommendation to perform mitral edge-to-edge repair is limited to a recommendation grade 2a if the inclusion criteria of COAPT are met. 6,7 Evidence of positive hemodynamic effects, such as a reduction in pulmonary pressure and an increase in RV performance after TMVR, could help overcome such controversies and aid in defining suitable patients. In both of our patients with TMVR, positive changes were observed. Herrmann et al. and Veenis et al. published similar results as case reports. 16,17 Procedural success of TTVR has previously been defined as a reduction of regurgitation severity of only ≥1 degree on echocardiography. 18 On follow-up echocardiography, such modest improvements in valve function may be lost due to variability caused by loading conditions because echocardiography is performed at a specific time point. The presence of multiple regurgitant jets after edge-to-edge repair further complicates the assessment. Therefore, some authors already advocate using 3D echocardiography, which is technically more difficult in the edge-to-edge repaired valve and resource intensive. 19,20 Furthermore, it has been known for some time that although echocardiography is consistent with invasive hemodynamic parameters, for example, in terms of sPAP, 21 it is still only a snapshot with large observer variability.
Stocker et al. recently demonstrated that invasively measured pulmonary pressures are predictive of outcome in patients after TTVR. 22 However, the evolution of pulmonary pressures after TTVR, which should lead to an improvement of RV function, is completely unknown. Here, we observed no changes of mean PAP in our tricuspid edge-to-edge repaired patients, but a significant increase in systolic PAP and CO. We know, that preserved/improved RV function is a robust indicator of good prognosis. 23 However, passing the tricuspid valve for right heart catheterization may also be difficult or even damage the edge-to-edge repaired valve and is, therefore, rarely performed. The use of CardioMEMS avoids classic right heart catheterization for hemodynamic monitoring and is, therefore, a perfect alternative to monitor procedural success of tricuspid edge-to-edge repair.
In addition, several studies showed beneficial effects on the right heart after TTVR based on echocardiography. 11,24 In our patients, these changes could not be consistently observed, which could be due to several reasons: echocardiography is inaccurate due to observer variance, deterioration of TR even after successful repair, lack of patient adherence to therapy (e.g., after improvement of symptoms), small number of patients with short follow-up time.
Little is known about the hemodynamic effects of biCAVI. Currently, outcome data relate to functional improvements, mainly NYHA score. 8 Some data also indicate improvement of RV function and CO in addition to a reduction of TR. [25][26][27] In our patient, CO increased, and PA pressures decreased, which was not measurable on echo or clinical indices (Table 2).
Considering all four patients together, CardioMEMS data reflect functional parameters and diuretic dose and can, in complex cases, be more useful than echocardiography or biomarker data.
Regarding outcome, events occurred less frequently in the CardioMEMS group than in the control group. In addition, events appeared to be more frequent in patients with tricuspid interventions. However, because of the small number of patients, these results should be interpreted with caution.
CardioMEMS has been shown to have the potential to reduce patient hospitalizations, which is tremendously important for improving patient health but also contributes significantly reducing costs in healthcare systems. 14 This can certainly put implantation costs into perspective.
Our small cohort may be useful for establishing future larger studies in a very understudied area. Invasive hemodynamic monitoring with CardioMEMS may help to standardize the evaluation of procedural success and relate these data with outcome after valve interventions.

| LIMITATIONS
This is a small case series from a single center. Therefore, results should be interpreted as hypothesis-generating.
CardioMEMS measurements were not compared with conventional right heart catheterization beyond baseline measurements, particularly because right-heart catheter is associated with risk after TTVR.
The patients provided a varying amount of CardioMEMS data. While one patient sometimes even provided data twice a day, others provided data only twice a week. Therefore, the amount of data collection time points varied between patients, which may bias the statistical analysis. Comparison of data from observational periods may be statistically significant but may not always lead to clinical consequences.
In addition, it must be mentioned that the CardioMEMS Cardiac Output algorithm has not yet been validated. The corresponding study (Validation of CardioMEMS HF System Cardiac Output Algorithm IDE [VICTOR], NCT05428384) is currently in progress.

| CONCLUSION
Invasive monitoring using CardioMEMS provides important information beyond conventional imaging and changes in biomarker serum levels after mitral-and tricuspid valve interventions. Such data pave the way for a deeper understanding of the prerequisites for optimal patient selection and management for catheter-based interventions.

ACKNO WLE DGE MENTS
Many thanks to Sophie Lindenthal for her assistance with the graphic design.

FUNDING INFORMATION
This research received no external funding.

CONFLICT OF INTEREST STATEMENT
V. Dannenberg reports industrial relationships with Abbott and Edwards Lifesciences; J. Mascherbauer with Abbott, Edwards Lifesciences, Boston Scientific, and Shockwave Medical, which are not related to this work and C. Hengstenberg with Edwards Lifesciences, Abbott, Boston Scientific, and Medtronic, which are not related to this work. The remaining authors have nothing to disclose.