Surgical Treatment of Tricuspid Valve Regurgitation in Patients Undergoing Left Ventricular Assist Device Implantation: Interim Analysis of the TVVAD Trial

Structured Abstract

Background:

Right heart failure (RHF) remains a serious complication of left ventricular assist device (LVAD) therapy. Many patients presenting for LVAD implantation have significant tricuspid valve regurgitation (TR). It remains unknown whether concurrent tricuspid valve surgery (TVS) reduces post-operative RHF.

Objectives:

The primary aim was to identify whether concurrent TVS reduced the incidence of moderate or severe RHF within the first 6-months after LVAD implantation.

Methods:

Patients with moderate or severe TR on pre-operative echocardiography were randomized to either LVAD implantation alone (No TVS) or LVAD implantation with concurrent TVS. Randomization was stratified by pre-operative right ventricular dysfunction. The primary endpoint was the frequency of moderate or severe RHF within 6-months after surgery.

Results:

This report describes a planned interim analysis of the first 60 randomized patients. The TVS group (n=32) had mild or no TR more frequently on follow-up echo studies, compared to the No TVS group (n=28). However, at 6-months, the incidence of moderate and severe RHF was similar in each group (TVS: 46.9% versus No TVS: 50%, p=0.81). There was no significant difference in post-operative mortality or requirement for right ventricular assist device between the groups. There were also no significant differences in secondary endpoints of functional status and adverse events.

Conclusions:

The presence of significant TR prior to LVAD is associated with a high incidence of RHF within the first 6 months after surgery. TVS was successful in reducing post-implant TR compared with No TVS but was not associated with lower incidence of RHF.

Video abstract:

Impact of Tricuspid Valve Treatment on Post-operative Right Heart Failure. Patients with moderate or severe TR undergoing LVAD implantation were randomized to undergo concurrent TVS versus No TVS. At 6-months, there was no significant difference in the incidence rate or severity of RHF between TVS and No TVS.

Download video file ^{(30.2MB, mp4)}

Introduction

Right heart failure (RHF) remains an important complication of left ventricular assist device (LVAD) therapy. Yet, improvements in durable LVAD design have not reduced the frequency of this adverse event^1–5. The etiology of post-LVAD RHF is multifactorial including pre-existing right ventricular (RV) dysfunction, pulmonary hypertension, and mechanical left ventricular (LV) unloading causing septal shift impairing right ventricular performance⁵. Notably, many patients being considered for LVAD implantation have moderate or severe tricuspid valve regurgitation (TR) which has been associated with post-LVAD RHF^6–8. Data from the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) indicates that concurrent tricuspid valve (TV) procedures are common at a rate of 16.5%^{9, 10}. However, it is unclear if concurrent tricuspid valve surgery (TVS) leads to reduced post-LVAD RHF. Most data on the subject has come from retrospective studies reaching conflicting conclusions^{4, 6, 7, 9–12}. The Surgical Treatment of Tricuspid Valve Regurgitation in Patients Undergoing LVAD Implantation (TVVAD) is the first randomized clinical trial (RCT) designed to answer this important question, whether concurrent surgical correction of moderate or severe TR in patients undergoing LVAD implantation would reduce the rate of RHF during the first 6 months after surgery.

Methods

Trial Design and Oversight

Patients scheduled for LVAD implantation who had moderate or severe TR were randomized to LVAD implantation alone (No TVS) versus LVAD with concurrent TVS (TVS) (Figure 1). Participants were randomized in a 1:1 ratio. The trial protocol was submitted to clinicaltrials.gov (NCT03775759, IRB Approval 6/5/2018) and began enrollment in December 2018. Patients required a qualifying echo within one-week of their scheduled surgery to complete screening. This trial was approved by the Institutional Review Board at Duke University Medical Center (Pro00072372). A data safety monitoring board was convened which independently reviewed outcomes relative to patient safety.

Figure 1 and Graphical Abstract:

Impact of Tricuspid Valve Treatment on Post-operative Right Heart Failure. Patients with moderate or severe TR undergoing LVAD implantation were randomized to undergo concurrent TVS versus No TVS. At 6-months, there was no significant difference in the incidence rate or severity of RHF between TVS and No TVS.

Participants

All patients scheduled to undergo LVAD implantation at Duke University Medical Center were screened for inclusion. Inclusion criteria included age of 18 years or older, planned LVAD implantation, and evidence of moderate or severe TR on pre-operative echocardiogram performed within one-week prior to surgery. Transthoracic (TTE) or transesophageal echocardiography (TEE) was allowed for screening. Patients with mild TR during surgical planning were reassessed during pre-procedure, intra-operative TEE for evidence of moderate or severe grade TR and randomized if they had upgraded TR severity. Both destination therapy and bridge-to-transplant indications for LVAD were included. Exclusion criteria included previous TVS, previous LVAD, planned concurrent RV assist device (RVAD) at the time of LVAD implantation, pre-implant RVAD or extracorporeal membrane oxygenation, pregnancy, or planned thoracotomy approach. Hospital records were reviewed to determine pre-implant characteristics, INTERMACS clinical profile, etiology of cardiomyopathy, history of prior cardiac surgery, type of pre-operative support, hemodynamic measurements, and blood laboratory values.

Interventions

Anesthesia and cardiopulmonary bypass were conducted according to institutional protocol and all patients were studied with an intra-operative TEE examination. A 1:1 block randomization scheme, stratified by the degree of RV dysfunction (none/mild versus moderate versus severe) based on intra-operative TEE assessment, was used to assign patients to each study arm. RV dysfunction severity and TR severity were evaluated by board-certified cardiologists and cardiac anesthesiologists, based on the guidelines and recommendations of the American Society of Echocardiography^{13, 14}.

Patients randomized to TVS underwent either TV ring annuloplasty or valve replacement at the time of LVAD implantation. The decision to perform annuloplasty versus replacement and the type and size of ring or valve to be used was made at the discretion of the four implanting surgeons for the study (CAM, JNS, BSB, MAD). Patients randomized to the control arm underwent LVAD implantation without TVS. All patients were administered an inhaled pulmonary vasodilator during weaning from cardiopulmonary bypass. Medical therapy for RHF post-implantation was consistent between the two groups and involved use of inotropic medications, inhaled pulmonary dilators, and diuretic medications. All but two patients were implanted with the HeartMate 3 LVAD (Abbott, Chicago, IL).

A post-operative inotrope weaning protocol was applied to all patients after sternal closure, when hemodynamically stable with no evidence of bleeding. Epinephrine and norepinephrine were weaned by 0.005 ug/kg/min every hour if the following criteria were met: (1) central venous pressure was <15mmHg, mean arterial pressure > 70 mmHg, SvO₂ >60%, cardiac index > 2; (2) urine output was more than 0.5 mL/kg/hr; and (3) absence of low flow or suction alarms during the previous 12-hours. The pulmonary artery catheter was removed after epinephrine, norepinephrine, and pulmonary vasodilator had been weaned off. Dopamine and dobutamine were also weaned by 1 µg/kg/min per 12-hours and milrinone was weaned by 0.125 µg/kg/min per 12-hours per protocol, if the following criteria were met: (1) mean arterial pressure was >70mmHg; (2) urine output was ≥ 0.5 mL/kg/hr; (3) absence of worsening signs or symptoms of RHF (ie. elevated jugular venous pressure, increased ascites, edema, weight gain); (4) absence of no low flow or suction alarms during the previous 12-hours. The decision to implant an RVAD was made by the surgeon. RVAD application was reserved for cases in which LVAD filling remained inadequate despite maximal intravenous inotropic support and inhaled pulmonary vasodilators.

Outcomes

The primary outcome was moderate or severe RHF during the first 6-months post-surgery. A committee comprised of cardiologists and cardiac surgeons, who were members of the study not involved in care decisions and management, adjudicated the primary outcome for each subject utilizing the INTERMACS definition for RHF. There were two criteria based on whether RHF occurred during the index hospitalization versus during a subsequent readmission. In the setting of the index hospitalization, moderate grade RHF required 7-14-days of either intravenous inotropic agents or pulmonary vasodilator support, and severe grade required more than 14-days of either intravenous inotropic agents or pulmonary vasodilator support. Severe acute grade required that the patient have RVAD implantation or experience death due to RHF. In the setting of readmissions, moderate grade RHF required documentation of elevated right atrial pressure and clinical manifestation of RHF on two separate readmissions within a 3-month period, whereas severe RHF required intravenous inotropic support, RVAD during readmission, or death due to RHF.

Secondary outcomes included rates of serious adverse events, degree of post-operative TR, 6-minute walk test (6MWT) at 3- and 6-months post-implantation, Kansas City Cardiomyopathy Questionnaire (KCCQ) short version scores at 6-months and index hospitalization length of stay (LOS) and intensive care unit (ICU) LOS in days. Serious adverse events included all-cause mortality, atrioventricular block resulting in pacemaker placement, major bleeding as defined by INTERMACS, tricuspid prosthetic dysfunction, endocarditis, stroke, pump thrombosis, and mechanical ventilation for longer than 72-hours.

Sample size

The sample size was determined using a Lan-DeMets spending function with O’Brien-Fleming stopping boundaries to achieve 80.2% power to detect a 25% difference between the groups. Thus, a sample size of 100 patients was determined, assuming 60% and 35% RHF incidence in the No TVS and TVS groups, respectively. The interim analysis was to be performed when the primary outcome was determined for first 60 subjects (60% of the initial target sample size). This study had a planned interim analysis to utilize an adaptive group sequential trial design with an efficacy boundary based on conditional power (CP) that would allow for a pre-specified, one-time increase in sample size (max = 280) to determine the sample size needed for adequate statistical power to detect a clinically meaningful benefit (15% difference). Pre-specified efficacy (Z-statistic < −2.28) and futility (CP < 10%) boundaries were defined for the primary endpoint and allowed for the trial to be terminated early if overwhelming evidence for efficacy or futility was demonstrated by the interim analysis.

Statistical analysis

Continuous data are presented using the mean ± SD or median with 25^th and 75^th percentiles (Q1, Q3), while categorical data are shown as counts with percentages based on non-missing data. For the primary endpoint analysis, a two-sided Z-test was performed to determine if TVS was associated with a difference in RHF frequency. For the secondary analyses, continuous variables were compared between the two study arms using the Wilcoxon rank-sum test or Student’s t-test, based on distribution, while differences between categorical variables were compared using a chi-square or Fisher’s exact test (expected cell counts <5). Adjustment for multiple testing for secondary outcomes was done using the Benjamini-Hochberg ad-hoc procedure for controlling the False Discovery Rate (FDR). A p-value < 0.05 was considered statistically significant, and all analyses were done using SAS 9.4 (SAS Institute, Inc., Cary, NC).

Results

Baseline characteristics of the two groups

Of the 60 patients included in this analysis, 32 were randomized to TVS and 28 to No TVS (Supplemental Figure 1). No one was lost to follow-up. There were no statistically significant differences in pre-operative baseline characteristics. The average age was similar in each arm (TVS: 59.3 ± 12.0 vs No TVS: 58.4 ± 12.2 years). Patients in both arms were predominantly African American (63.8%) and male (80.0%). The patients were predominantly INTERMACS profile 2 (TVS: 53.1% vs No TVS: 57.1%). Hemodynamic parameters were similar between the two groups and hemodynamic predictors of RHF, such as pulmonary artery pulsatility index, suggested high risk for RHF in both groups^{15, 16}. The baseline distributions of RV dysfunction and TR severity were similar between the two groups. All baseline characteristics are presented in Table 1.

Table 1:

Pre-operative Baseline Characteristics for TVS versus No TVS

Characteristics	No TVS	TVS
Baseline clinical characteristics
N	28	32
Age, mean (SD a ) – years	58.4 (12.2)	59.3 (12.0)
Sex, n (%)
Male	23 (82.1%)	25 (78.1%)
Female	5 (17.9%)	7 (21.9%)
Race, n (%)
Black or African American	18 (64.3%)	19 (63.3%)
White or Caucasian	9 (32.1%)	11 (36.7%)
BSA b , mean (SD) – m2	2.1 (0.3)	2.1 (0.3)
Pre-op chronic conditions, n (%)
Insulin dependent diabetes	3 (10.7%)	6 (18.8%)
Stroke or TIAc	4 (14.3%)	5 (15.6%)
CKDd, ≥ Stage 3	18 (64.3%)	17 (53.1%)
DVTe or PEf	5 (17.9%)	5 (15.6%)
Previous Cardiac Surgery, n (%)	9 (32.1%)	8 (25.0%)
Pre-operative heart failure characteristics
INTERMACS Profile, n (%)
2 – Progressive decline	16 (57.1%)	17 (53.1%)
3 – Stable but inotrope dependent	4 (14.3%)	10 (31.3%)
4 – Resting symptoms	8 (28.6%)	5 (15.6%)
Etiology of Cardiomyopathy, n (%)
Ischemic	10 (35.7%)	12 (37.5%)
Non-ischemic	18 (64.3%)	16 (50.0%)
Mixed	0 (0.0%)	4 (12.5%)
Type of pre-operative support, n (%)
IVg inotropes	19 (67.9%)	27 (84.4%)
IABPh	14 (50.0%)	15 (46.9%)
Temporary LVAD	0 (0.0%)	2 (6.3%)
Pre-operative hemodynamic measurements and indices
PVR i , median (Q1, Q3 j ) – mmHg/L/min	3.0 (2.1, 8.0)	3.7 (2.3, 5.2)
CO k , mean (SD) – L/min	3.7 (1.0)	4.2 (1.5)
CI l , mean (SD) – L/min/m2	1.8 (0.5)	2.0 (0.6)
PAm Systolic pressure, mean (SD) – mmHg	53.6 (13.3)	61.3 (11.2)
PA Diastolic pressure, mean (SD) – mmHg	26.9 (8.0)	31.7 (7.4)
PCWP n , mean (SD) -- mmHg	24.9 (8.0)	26.6 (7.9)
CVP o , mean (SD) -- mmHg	16.2 (6.4)	14.8 (5.7)
RVSWi p , mean (SD) – mmHg mL/m2	438.3 (219.7)	606.1 (350.7)
PAPi q , median (Q1, Q3)	1.76 (0.96, 2.82)	2.29 (1.45, 2.92)
CVP/PCWP, mean (SD)	0.7 (0.2)	0.6 (0.5)
Pre-operative echocardiogram measurements
TR severity, n (%)
Moderate	13 (46.4%)	18 (56.3%)
Severe	15 (53.6%)	14 (43.8%)
RV dysfunction severity, n (%)
None/mild	4 (14.3%)	5 (15.6%)
Moderate	14 (50.0%)	15 (46.9%)
Severe	10 (35.7%)	12 (37.5%)
Pre-operative labs, mean (SD)
Creatinine, mg/dL	1.4 (0.5)	1.8 (1.1)
Bilirubin, total, mg/dL	1.6 (0.9)	1.9 (2.4)

Abbreviations:

^aSD - standard deviation,

^bBSA – body surface area,

^cTIA – transient ischemic attack,

^dCKD – chronic kidney disease,

^eDVT – deep vein thrombosis,

^fPE – pulmonary embolism,

^gIV – intravenous,

^hIABP – intra-aortic balloon pump,

ⁱPVR – pulmonary vascular resistance,

^jQ – quartile,

^kCO – cardiac output,

^lCI – cardiac index,

^mPA – pulmonary artery,

ⁿPCWP – pulmonary capillary wedge pressure,

^oCVP – central venous pressure,

^pRVSWi – right ventricular stroke work index,

^qPAPi – pulmonary artery pulsatility index

Tricuspid valve surgical interventions

Within the TVS arm, 84.4% of patients received annuloplasty repair versus 15.6% valve replacement. (Supplemental Table 1). The most common ring was the 28 mm Medtronic TRI-AD 2.0 Semi-Rigid ring (Medtronic, Inc., Minneapolis, MN). The most common valve was the Medtronic Mosaic Mitral Valve. Patients who underwent TVS demonstrated immediate improvement in TR severity as demonstrated on the intra-operative post-LVAD TEE compared to those who randomized to the LVAD alone arm (p < 0.001). The distribution of TR severity for each group, on the follow-up TTEs at both one-week post-surgery and at 3-6-months post-surgery, remained stable when compared to the distribution of TR severity on the intra-operative TEE post-LVAD (Figure 2). At 6-months, no patients in the TVS group had moderate or greater TR (TVS: 52.6% none, 47.4% mild). In contrast, more than half of the patients in the No TVS group had persistent moderate or severe TR (No TVS: 12.5% none, 25.0% mild, 43.8% moderate, 18.8% severe).

Figure 2:

TR severity by timepoint. This graph demonstrates the percentage of all patients in each group experiencing none, mild, moderate, or severe TR at the pre-operative, intra-operative post-LVAD, 1-week post-LVAD, and 3–6-month post-LVAD timepoints. There was immediate improvement in TR to none or mild grade in the TVS group that lasts up to 6-months post-surgery, whereas most subjects in the No TVS group continued to experience moderate or severe TR.

Primary endpoint: Right heart failure within 6-months

At 6-months, there was no significant difference in the incidence of RHF between the two groups (TVS: 46.9% versus No TVS: 50%, p = 0.81). Of the cases which met the RHF primary endpoint, only two occurred after the index hospitalization. The distribution of RHF severity was also similar between the two groups (Figure 3). Interim analysis of the primary endpoint demonstrated futility in determining a statistically significant difference between the two groups (CP = 1.6%) based on the initial target sample size (n = 100). As a result, the sample size was not increased, and further trial enrollment was terminated.

Figure 3 and Central Illustration:

The primary endpoint: incidence of right heart failure within 6 months of surgery. This graph compares the incidence of right heart failure between each group. It also compares the fraction of subjects with moderate, severe, or severe acute RHF in each group. There was no significant difference in RHF incidence between the TVS (46.9%) and No TVS (50%) group (p = 0.81) and the distribution of RHF severity was similar as well.

Length of Stay and Readmissions

The median (Q1, Q3) hospital length of stay was 23 (17, 31) days in the TVS group and 15.5 (13, 23) days in the No TVS group (p = 0.10). The median (Q1, Q3) ICU length of stay was 5.5 (4, 11) days in the TVS group versus 6.5 (4, 9) days in the No TVS group (p = 0.92). There was no statistically significant difference in the number of readmissions observed in each group (p = 0.65). Most patients in both groups did not experience a readmission during the follow-up period (TVS: 53.1% vs No TVS: 64.3%). Among the readmissions, 12.5% were due to RHF in the TVS group, and none in the No TVS group (Table 2). Of the reasons for readmissions, gastrointestinal bleeding and driveline infection were the most common.

Table 2:

Table of secondary endpoint results

	No TVS	TVS	p-value
6-minute walk test, feet
2-4 months
Median (Q1, Q3)	929 (720, 1132)	890 (600, 1120)	n.s.
4-6 months
Median (Q1, Q3)	999 (702, 1211)	1005.0 (784, 1174)	0.917
KCCQ (6 month)
Physical Limitation
Median (Q1, Q3)	66.7 (50.0, 83.3)	85.4 (56.3, 91.7)	n.s.
Symptom Frequency
Median (Q1, Q3)	87.5 (66.7, 100.0)	87.5 (70.8, 95.8)	n.s.
Quality of Life
Median (Q1, Q3)	62.5 (43.8, 81.3)	75.0 (56.3, 87.5)	n.s.
Social Limitation
Median (Q1, Q3)	75.0 (58.3, 100.0)	83.3 (58.3, 100.0)	n.s.
Summary
Median (Q1, Q3)	71.6 (47.7, 82.0)	78.4 (62.8, 93.2)	0.654
Hospital length of stay days, median (Q1, Q3)	15.5 (13.0, 23.0)	23.0 (17.0, 31.0)	0.103
ICU length of stay days, median (Q1, Q3)	6.5 (4.0, 9.0)	5.5 (4.0, 11.0)	0.917
Total days of inotropic support, median (Q1, Q3)	7 (6, 14)	7 (5, 11)	n.s.
Total days of pulmonary vasodilator support, median (Q1, Q3)	3 (2, 5.3)	3 (1, 4)	n.s.
Total days of RVAD support in those requiring RVAD, median (Q1, Q3)	6 (5, 8)	3 (2, 5)	n.s.
Number of readmissions, N(%)			0.654
1	6 (21.4%)	10 (31.3%)
2	2 (7.1%)	0 (0.0%)
3	2 (7.1%)	4 (12.5%)
4	0 (0.0%)	1 (3.1%)
Number of readmissions for right heart failure, N (%)
1	0 (0.0%)	4 (12.5%)	n.s.
NYHA Class, N (%)
1 Month			n.s.
Class 1	1 (4.5)	0 (0.00)
Class 2	10 (45.5)	11 (47.8)
Class 3	10 (45.5)	9 (39.1)
Class 4	1 (4.5)	3 (13.0)
3 Month			n.s.
Class 1	5 (22.7)	3 (10.7)
Class 2	11 (50.0)	19 (67.9)
Class 3	5 (22.7)	5 (17.9)
Class 4	1 (4.5)	1 (3.6)
6 Month			0.913
Class 1	7 (35.0)	6 (25.0)
Class 2	11 (55.0)	14 (58.3)
Class 3	2 (10.1)	4 (16.7)
Post-op labs (6-months)
Creatinine
1-month
Median (Q1, Q3)	1.00 (0.90, 1.50)	1.20 (1.00, 1.40)	n.s.
3-months
Median (Q1, Q3)	1.20 (1.00, 1.60)	1.20 (1.10, 1.60)	n.s.
6-months
Median (Q1, Q3)	1.30 (1.00, 1.70)	1.25 (1.10, 1.50)	n.s.
Bilirubin, total
1-month
Median (Q1, Q3)	0.90 (0.70, 1.40)	1.05 (0.80, 1.70)	n.s.
3-months
Median (Q1, Q3)	1.00 (0.70, 1.10)	0.90 (0.60, 1.20)	n.s.
6-months
Median (Q1, Q3)	0.80 (0.60, 1.10)	1.00 (0.70, 1.20)	n.s.

^*n.s. = not significant

NYHA Class, KCCQ, and 6-minute walk test

NYHA class was similar between the two groups at 1-, 3-, and 6-months. At 6-months, there was no statistically significant difference in the distribution of NYHA classes between the two groups (p = 0.91). At 6-months, most patients were classified as NYHA Class II (TVS: 58.3% vs No TVS: 55.0%, p = 0.91). Both groups demonstrated an improvement in 6MWT distances from the early timepoint to the later timepoint, and the distances walked at these timepoints were similar between the two groups. At the later timepoint, the median (Q1, Q3) in the TVS group was 1005 (784, 1174) feet and 999 (702, 1211) feet in the No TVS group (p-value = 0.92). At 6-months, KCCQ results assessing physical limitation, symptom frequency, quality of life, social limitation, and summary were also similar between the two groups (p-value = 0.65) (Table 2).

Adverse events

There was a trend toward lower all-cause mortality within the 6-month follow-up period in the TVS group (TVS: 12.5% vs No TVS: 25.0%, p = 0.65). There were 4 deaths in the TVS group: one due to RHF, two due to respiratory arrest, and one due to sepsis from Aspergillus pneumonia. There were 7 deaths in the No TVS group: three due to RHF, two due to refractory ventricular tachycardia, one due to abdominal sepsis, and one due to mesenteric ischemia. There was no difference in RVAD implantation rate between the groups (TVS: 9.4% vs No TVS: 14.3%, p = 0.91). No subjects experienced TV prosthetic dysfunction, endocarditis, or pump thrombosis. Both the TVS and No TVS groups had similar rates of major bleeding, stroke, prolonged ventilation, and new requirement for dialysis (Table 3).

Table 3:

Table of adverse events

	No TVS	TVS	p-value
All-cause mortality, N(%)	7 (25.0%)	4 (12.5%)	0.654
RVAD, N(%)	4 (14.3%)	3 (9.4%)	0.913
AV Block, N(%)	1 (3.6%)	0 (0.0%)	n.s.
Tricuspid prosthetic dysfunction, N(%)	0 (0%)	0 (0%)	n.s.
Endocarditis, N(%)	0 (0%)	0 (0%)	n.s.
Major bleeding, N(%)			0.799
No major bleeding	16 (57.1%)	16 (50.0%)
Early Post-op Bleeding	4 (14.3%)	9 (28.1%)
Late Post-op Bleeding	8 (28.6%)	7 (21.9%)
Stroke, N(%)	2 (7.1%)	3 (9.4%)	n.s.
Pump thrombosis, N(%)	0 (0%)	0 (0%)	n.s.
Prolonged mechanical ventilation (>72hrs), N(%)	4 (14.3%)	6 (18.8%)	n.s.
New dialysis requirement, N(%)	1 (3.6%)	1 (3.1%)	n.s.

^*n.s. = not significant

Discussion

RHF after durable LVAD remains an important clinical challenge which is most prevalent during early recovery and substantially limits the overall effectiveness of the therapy^{2, 4, 17, 18}. While the advent of the fully magnetically levitated LVAD design has reduced rates of adverse events related to hemocompatibility, rates of RHF have not declined with this newer design¹. A substantial number of patients who present for LVAD therapy have pre-operative TR^{8, 19–21}. Usually, the TR is classified as functional or secondary to RV dysfunction, volume overload and pulmonary hypertension²². Significant numbers of patients have undergone LVAD implantation with concurrent TVS, but the impact of TVS on RHF post-implantation is incompletely understood.

There are many retrospective, non-randomized studies that have evaluated the association between concurrent TVS at the time of LVAD implantation and post-LVAD outcomes. These previous studies are either single center case series or reviews of larger databases^{9, 11}. In both instances, significant numbers of patients appear to be undergoing concurrent TVS, and outcomes of these groups have been compared to cohorts who do not receive TVS. Unfortunately, in these studies, the groups are not necessarily similar with regard to baseline risk for RHF, and meaningful conclusions are limited due to the lack of randomization.

This is the first randomized study to examine the impact of concurrent TVS during LVAD implantation. Concurrent TVS achieved significant reduction in post-implant TR on follow up, protocolized echocardiograms. Conversely, the cohort that did not receive TVS displayed a significant rate of moderate or severe TR at the 6-month follow-up. However, the primary endpoint, moderate or severe RHF at 6 months, was not different between the groups (Video abstract). Furthermore, a pre-determined futility analysis concluded that further enrollment was unlikely to demonstrate a clinically meaningful difference between the groups with regards to the primary endpoint. Consistent with the presence of significant preoperative RV dysfunction and TR, the overall rate of post-operative RHF and death during the 6-month follow up was high relative to these overall outcomes in the INTERMACS registry or MOMENTUM trial reports^{1, 2, 23}. Therefore, this study appears to have captured a subset of LVAD cases which were high risk for RHF. Secondary endpoints were also not significantly different between groups. The incidence of new atrial fibrillation and new ventricular tachycardia was similar between the two groups. While mortality was not statistically significantly lower in the TVS group, there may be a trend toward lower mortality with TVS, and it is our intention to analyze the entire cohort over a longer duration. However, it seems unlikely that concurrent TVS would impact overall mortality if it does not significantly reduce RHF.

The conclusions described in this manuscript agree with what has been observed in retrospective reviews of large databases. Robertson et al. utilized the STS National Database and found that concurrent TVS for significant TR did not reduce early mortality or RVAD requirement¹¹. Similarly, Song et al. utilized the INTERMACS database and concluded that while TVS improved post-operative TR severity, it did not reduce late mortality or right ventricular failure⁹. Dunlay et al. conducted a systematic review and found no differences between patients who underwent TVS versus No TVS with regards to post-operative outcomes, including the use of RVAD, renal dysfunction, and early mortality ²⁴.

Limitations

Several limitations of this study should be acknowledged. This interim analysis, while powered to detect a significant difference in the primary endpoint, is not sufficiently powered to detect differences in all of the secondary endpoints, such as survival and need for RVAD. We will be following up for late survival in the full study cohort. Of note, most deaths were not related to RHF. As a single-center study, the results may not be generalizable to other providers. The results of this trial suggest that post-LVAD TR severity is not a major driver of RHF and its associated adverse events. TR severity may be a secondary manifestation of other issues related to advanced heart failure and its correction does not improve one’s predisposition to developing RHF. This suggests that the aggressive use of concurrent TVS does not lead to reduced rates of RHF when considering allcomers with moderate or severe TR. Since half of the patients in this study had moderate TR, the results should not be extended to patients with severe TR. A larger study focused on patients with pre-operative severe TR could potentially reveal a significant benefit. Also, this study was designed to assess effects of TVS on early RHF outcomes, given its higher prevalence during the early post-operative period. There may be a later benefit from concurrent TVS noted beyond 6-months. Future directions for this study will be to complete follow-up of all enrolled patients beyond 1-year post-surgery.

Conclusion

In summary, patients with significant preoperative TR are at high risk for RHF during the early post implant period. Concurrent TVS for patients undergoing LVAD implantation who have significant preoperative TR, did not appear to reduce early RHF adverse event rates, despite achieving significant reduction in TR based on follow up echocardiograms.

Central Message:

This clinical trial demonstrated that treatment of tricuspid regurgitation at the time of left ventricular assist device (LVAD) implantation does not reduce right heart failure incidence at 6 months.

Perspective Statement:

LVAD therapy is the preferred treatment for patients with end-stage heart failure who are not candidates for heart transplantation and can be complicated by post-op incidence of right heart failure (RHF). Patients with moderate or severe tricuspid valve regurgitation (TR) before LVAD implantation should be made aware that surgical correction of TR does not reduce this risk.

Glossary of abbreviations:

RV

right ventricle

LV

left ventricle

LVAD

left ventricular assist device

RVAD

right ventricular assist device

TR

tricuspid valve regurgitation

RHF

right heart failure

RCT

randomized clinical trial

TEE

transesophageal echocardiography

TTE

transthoracic echocardiography

TV

tricuspid valve

TVS

tricuspid valve surgery

6MWT

6-minute walk test

LOS

length of stay

ICU

intensive care unit

KCCQ

Kansas City Cardiomyopathy Questionnaire

NYHA

New York Heart Association

INTERMACS

Interagency Registry for Mechanically Assisted Circulatory Support

CP

conditional power

FDR

false discovery rate