In-Hospital Outcomes After Tricuspid Valve Surgery in the United States

Abstract

Background

The impact of institutional experience on postoperative outcomes among patients undergoing isolated or concomitant tricuspid valve operations is not well studied.

Objectives

The authors characterized the presence of a potential volume-outcome relationship between center-level cardiac valve procedural volume and risks of in-hospital mortality as well as resource utilization in a national cohort.

Methods

The 2016 to 2022 Nationwide Readmissions Database was reviewed to identify all adults undergoing isolated or concomitant tricuspid valve replacement or repair. A multivariable regression model was developed to evaluate mortality and hospitalization costs. Hospitals were categorized as high (HVH) or low volume based on explanatory analysis. The association of HVH status with postoperative complications, length of stay, and 30-day nonelective readmission were subsequently examined.

Results

Of an estimated 35,567 patients, 21,723 (61.0%) were treated at HVH. Compared to low-volume hospital, HVH patients were older, less commonly females, and more commonly of the highest income quartile. Patients at HVH more frequently received an isolated operation (21.0 vs 20.2%; P = 0.01). Following risk-adjustment, treatment at an HVH was associated with lower odds of mortality (adjusted OR: 0.56; 95% CI: 0.47-0.68) and cardiac complications (adjusted OR: 0.83; 95% CI: 0.73-0.94). However, HVH status was associated with increased costs by $6,070, but lower odds of nonhome discharge and 30-day nonelective readmission.

Conclusions

HVH status was associated with lower odds of in-hospital mortality and reduced risk of 30-day nonelective readmission, but greater hospitalization costs. Increased outreach and resource availability at low-volume centers may help alleviate this disparity and enhance the overall value of care in tricuspid valve surgery.

Central Illustration

Despite affecting over 1.6 million individuals in the United States and the risk of progressive heart failure, tricuspid valve disease remains undertreated with only 24% of eligible patients receiving operative repair or replacement.1, 2, 3 This may be in part due to the significant risk of morbidity associated with tricuspid valve operations, particularly in the immediate postoperative period. Moreover, a study of the Society for Thoracic Surgeons database found 86% of tricuspid valve procedures to be concomitant.³ Patient characteristics associated with the development of postoperative adverse events have included the degree of right ventricular dysfunction and liver disease, among others.⁴ Furthermore, prior work has demonstrated substantial center-level variation in both clinical outcomes as well as resource utilization among those undergoing tricuspid valve repair or replacement (TVR).^5,6

Over the past 4 decades, a mounting body of evidence has suggested the presence of a positive volume-outcome relationship for a myriad of complex procedures.^7,8 Institutional and surgical expertise along with the availability of adjunct complex resources at “expert” centers are thought to, at least in part, underlie this observation. In fact, advocacy organizations such as the Leapfrog Group have established minimum volume requirements to optimize outcomes at the population level.⁹ Although this phenomenon has been demonstrated in many aspects of cardiac surgery, the volume-outcome relationship in isolated tricuspid valve operations has not been elucidated thus far.

In the present work, we characterized acute clinical outcomes and hospitalization costs associated with isolated/concomitant tricuspid valve operations in a national cohort. We hypothesized increasing center tricuspid operative volume to be independently associated with reduced mortality and complications during index hospitalization.

Methods

Nationwide Readmissions Database

This was a retrospective cohort study of the 2016 to 2022 Nationwide Readmissions Database (NRD). Maintained as part of the Healthcare Cost and Utilization Project, the NRD is the largest all-payer readmissions database in the United States and provides accurate for ∼60% of all U.S. hospitalizations using the survey-weighting methodology.¹⁰ Within each calendar year, unique patient identifiers assigned by the NRD are used to track in-state readmissions. Due to the deidentified nature of the NRD, this study was deemed exempt from full review by the Institutional Review Board at the University of California-Los Angeles.

Study cohort and definitions

All elective adult (≥18 years) hospitalizations entailing isolated or concomitant TVR were identified using the relevant International Classification of Diseases-10th Revision (ICD-10) procedure codes.¹¹ As shown in Supplemental Figure 1, records missing key data, including age, sex, mortality, or transfer status, were excluded from the analysis (4.7%) (Figure 1). To enhance cohort homogeneity, entries involving heart transplantation, ventricular assist device placement, or endocarditis were not considered (1.3%).

Figure 1.

The Relationship Between Hospital Tricuspid Volume and Outcomes

(A) Functional relationship demonstrating lower risk of mortality as annual tricuspid operation volume increases. (B) Functional relationship demonstrating increased mean hospitalization expenditures as annual tricuspid operation volume increases. Hospitals that performed 12 or more annual tricuspid operations comprised the HVH cohort. HVH = high-volume hospital; LVH = low-volume hospital; TVR = tricuspid valve repair/replacement.

Patient and hospital characteristics, including age, sex, primary payer, income quartile, hospital teaching status, and urban or rural location, were defined in accordance with the NRD Data Dictionary.¹² Comorbidities, such as heart failure, pulmonary hypertension, chronic kidney disease, liver disease, and atrial fibrillation, were identified using appropriate diagnosis codes reported elsewhere.¹³ The burden of chronic conditions was quantified using the van Walraven modification of the Elixhauser Index, a validated score of 30 comorbidities. In-hospital complications obtained using ICD-10 coding relevant to procedure-related admission were categorized as, cardiac (cardiac tamponade, myocardial infarction, ventricular fibrillation, and ventricular tachycardia), infectious (wound infection, postoperative seroma, cellulitis, and superficial/deep site infection), respiratory (acute respiratory distress syndrome, pneumonia, pneumothorax, prolonged ventilation, and respiratory failure), renal (acute renal failure requiring dialysis), and receipt of blood transfusion.¹³ Major adverse events (MAEs) were defined as a composite of in-hospital mortality and the development of any complication listed above. Separately, the annual institutional cardiac valve volume was calculated as the sum of mitral, aortic, pulmonary, and tricuspid valve interventions performed at each hospital.

Study outcomes

The primary outcome of interest was in-hospital mortality and the development of postoperative complications. Hospitalization costs, length of stay, nonhome discharge, and 30-day readmission were also examined. Hospitalization costs were derived by applying hospital-specific cost-to-charge ratios provided by the Healthcare Cost and Utilization Project to total charges recorded during the index hospitalization. Costs were inflation-adjusted to 2022 U S. dollars using the Personal Health Care Price Index. Cost analyses were further adjusted for length of stay and discharge disposition to account for differences in postoperative resource utilization.

Statistical analysis

Categorical and continuous variables are summarized as proportions (%) and median (IQR), respectively. Between-group comparisons were conducted using the Pearson chi-square test for categorical and the Mann-Whitney U test for continuous variables. These analyses represent unadjusted comparisons and do not account for patient-, procedural-, or hospital-level covariates. Multilevel mixed regression models were developed to characterize the association of high-volume hospital (HVH) with outcomes of interest and marginal estimates were calculated. The first level of each model includes patient characteristics whereas unique hospital identifiers constituted the second level. The Least Absolute Shrinkage and Selection Operator was used to guide variable selection. Briefly, this automated algorithm reduces model overfitting and enhances out-of-sample performance.¹⁴ Furthermore, length of stay and discharge disposition were adjusted in cost analysis models. Subsequently, the annual tricuspid operation volume was incorporated into the multivariable regression using restricted cubic splines.¹⁵ We aimed to identify the volume at which an inverse relationship between annual hospital caseload and mortality occurs. To facilitate a bivariate comparison of clinical outcomes and resource use across institutions with varying levels of expertise, this volume threshold was used to categorize hospitals as low-volume hospital (LVH) or HVH volume, as shown in Figure 1A and Supplemental Figure 2. Hospitals performing 12 or more tricuspid valve operations annually were classified as HVHs, whereas those performing fewer than 12 cases per year were categorized as LVHs. Regression estimates are shown as adjusted ORs (AORs) for logistic and beta coefficients (β) for linear models, both with 95% CIs. All statistical tests were 2-sided with a significance threshold of α = 0.05. All analyses were conducted using Stata (version 16.1; StataCorp).

Results

Unadjusted findings

Of an estimated 35,567 patients, 21,723 (61.0%) were treated at HVH and 13,844 (38.9%) at LVH with detailed characteristics shown in Table 1. Of 2,769 centers, 531 (19.2%) were HVH. Compared to LVH, HVH patients were older (69 years [59-76] vs 68 [56-75]; P < 0.001), less commonly females (55.8% vs 58.4%; P = 0.01), and less frequently of the lowest income quartile (19.7% vs 27.8%; P < 0.001). Notably, patients at HVH more commonly underwent an isolated operation (21.0% vs 20.2%; P = 0.01), followed by a greater frequency of concomitant tricuspid and aortic procedures (13.5 vs 11.5; P = 0.01). Patients at HVH were more likely to receive a bioprosthetic valve replacement (14.4% vs 11.9%; P = 0.01) compared to others. Furthermore, patients at HVH less frequently presented with chronic lung disease (16.1% vs 18.3%; P < 0.001), diabetes (16.4% vs 18.4%; P < 0.001), and liver disease (4.9% vs 6.0%; P < 0.001). Those at HVH were more commonly treated at metropolitan teaching centers (93.3% vs 69.1%; P < 0.001).

Table 1.

Patient and Hospital Characteristics of Patients Undergoing Tricuspid Valve Replacement or Repair Stratified by Hospital Volume Status

	HVH (n = 21,723)	LVH (n = 13,844)	P Value
Age, y, median [IQR]	69 [59-76]	68 [57-75]	<0.001
Female (%)	55.8	58.4	0.01
Elixhauser Comorbidity Index, median [IQR]	6 [5-8]	6 [5-8]	0.18
Valve replacement or repair (%)			0.004
TV repair	82.6	84.8
TV-bioprosthetic	14.4	11.9
TV-mechanical	3.0	3.3
Operation type (%)			0.01
Isolated	21.0	20.2
TV-mitral	50.9	51.4
TV-aortic	13.5	11.5
TV-CABG	14.6	16.8
Income quartile (%)			<0.001
>75%	26.4	18.6
51%-75%	27.8	25.2
26%-50%	26.0	28.4
0%-25%	19.7	27.8
Insurance coverage (%)			<0.001
Private	26.8	23.0
Medicare	63.1	61.7
Medicaid	7.6	11.4
Other payer	2.5	3.9
Comorbidities (%)
Chronic lung disease	16.1	18.3	<0.001
Diabetes	16.4	18.4	<0.001
Late-stage kidney disease	3.5	4.0	0.11
Liver disease	4.9	6.0	<0.001
Pulmonary circulation disorders	30.9	34.3	<0.001
Hospital teaching status (%)			<0.001
Nonmetropolitan	7.7	3.5
Metropolitan Nonteaching	5.9	27.4
Metropolitan teaching	93.3	69.1

CABG = coronary artery bypass grafting; HVH = high-volume hospital; LVH = low-volume hospital; TV = tricuspid valve.

On unadjusted analysis, those at HVH had lower in-hospital mortality (3.7% vs 5.7%; P < 0.001). Patients at HVH less commonly experienced cardiac (14.6% vs 16.6%; P < 0.001), infectious (14.8% vs 17.6%; P < 0.001), and respiratory (18.8% vs 24.5%; P < 0.001) complications. Compared to LVH, individuals at HVH had greater costs ($67,621 [48,500-95,213] vs $62,714 [46,982-87,651]; P < 0.001), and reduced rates of 30-day nonelective readmission (13.4% vs 14.5%; P = 0.05) (Table 2).

Table 2.

Unadjusted and Risk-Adjusted Outcomes of Patients Undergoing Tricuspid Valve Replacement or Repair Stratified by Hospital Volume Status

	Unadjusted			Adjusted (AOR/β)
	HVH	LVH	P Value	HVH	95% CI	P Value
Clinical outcomes
In-hospital mortality	3.7	5.7	<0.001	0.56	0.47-0.68	<0.001
Complications (%)a
Blood transfusion	24.8	23.6	<0.001	1.26	1.05-1.51	0.01
Cardiac	14.6	16.6	<0.001	0.83	0.73-0.94	0.01
Infectious	14.8	17.6	<0.001	0.81	0.70-0.90	<0.001
Respiratory	18.8	24.5	<0.001	0.84	0.74-0.94	<0.001
Renal	23.9	23.3	0.55	0.96	0.87-1.07	0.45
Resource utilization
Nonhome discharge	18.7	22.3	<0.001	0.85	0.74-0.98	0.03
30-day readmission	13.4	14.5	0.05	0.89	0.81-1.00	0.04
Duration of stay (days) [IQR]	8 [6-11]	8 [6-11]	0.59	−0.13	−0.46 to 0.20	0.44
Costs (USD $1,000s) [IQR]	67.2 [48.5-95.2]	62.7 [46.9-87.0]	<0.001	6.07	3.32-8.83	<0.001

Estimates from dichotomous outcomes are reported as % or AORs and continuous outcomes as median with IQR or β-coefficients.

AOR = adjusted OR; USD = U.S. dollar; other abbreviations as in Table 1.

^aDefinition of complications: In-hospital complications were categorized as, cardiac (cardiac tamponade, myocardial infarction, ventricular fibrillation, ventricular tachycardia), infectious (wound infection, postoperative seroma, cellulitis, superficial/deep site infection), respiratory (acute respiratory distress syndrome, pneumonia, pneumothorax, prolonged ventilation, respiratory failure), renal (acute renal failure requiring dialysis), and receipt of blood transfusion.

Risk-adjusted outcomes

Following comprehensive risk-adjustment, HVH was associated with significantly reduced odds of mortality (AOR: 0.56; 95% CI: 0.47-0.68) (Table 2). Furthermore, as tricuspid operation volume increased, the risk of mortality decreased (Figure 1A). In addition, HVH was linked with lower odds of cardiac (AOR: 0.83; 95% CI: 0.73-0.94), infectious (AOR: 0.81; 95% CI: 0.72-0.90), and respiratory (AOR: 0.84; 95% CI: 0.74-0.94) complications. Furthermore, HVH status was associated with increased costs by $6,070 (95% CI: 3,323-8,831), lower nonhome discharge (AOR: 0.85; 95% CI: 0.74-0.98), and decreased odds of 30-day nonelective readmission (AOR: 0.89; 95% CI: 0.81-1.00). Moreover, as the volume of tricuspid valve operations increases, the costs of tricuspid valve replacement or repair increase (Figure 1B). Notably, as TVR volume increases, the risk of MAE decreases (Figure 2).

Figure 2.

Major Adverse Events by Tricuspid Valve Procedural Volume

Contour plot demonstrating lower risk of major adverse events (MAEs) as tricuspid volume increases. Abbreviation as in Figure 1.

In a sensitivity analysis of those who underwent an isolated tricuspid operation, HVH remained associated with increased mortality (AOR: 0.77; 95% CI: 0.62-0.91). In addition, HVH status was linked with greater cardiac (AOR: 0.85; 95% CI: 0.72-0.97), respiratory (AOR: 0.81; 95% CI: 0.76-0.96), and infectious (AOR: 0.78; 95% CI: 0.61-0.89) complications. In addition, HVH status was associated with greater hospitalization costs by $9,320 (95% CI: 6.23-12.97) and higher odds of 30-day nonelective readmission (AOR: 0.81; 95% CI: 0.76-0.91) (Supplemental Table 1).

Patient factors associated with care at high-volume hospitals

Several patient factors were found to be significantly associated with the receipt of care at HVH. Younger age (AOR: 0.99; 95% CI: 0.98-0.99) (Figure 3), Medicaid coverage (AOR: 0.47; 95% CI: 0.42-0.53), and the lowest quartile of income (AOR: 0.77; 95% CI: 0.70-0.85) were associated with decreased odds of care at HVH. However, greater Elixhauser Comorbidity Index (AOR: 1.02; 95% CI: 1.01-1.03) and elective admission (AOR: 1.45; 95% CI: 1.36-1.57) were associated with increased odds of care at HVH.

Figure 3.

Patient Factors Associated With Care at High-Volume Hospitals

Several demographic factors were found to be associated with a tricuspid valve replacement or repair at a high-volume hospital. Reference: Low-volume hospital. ∗P < 0.05.

Discussion

The substantial risk associated with tricuspid operations has been cited as a major reason for delaying intervention, particularly in isolated cases. In the present national analysis, isolated and concomitant cases yielded mortality rates of 3.23% and 3.38%, respectively. High-volume centers were associated with decrements in this risk. Despite lower rates of adverse events, operations at high-volume centers were costlier but resulted in 11% lower rates of 30-day nonelective readmissions. Several of these findings warrant further discussion (Central Illustration).

Central Illustration.

Association of Hospital Volume and Tricuspid Valve Outcomes

HVH = high-volume hospital.

Isolated tricuspid valve procedures are notably reported to be associated with an in-hospital mortality rate of 9% to 10% and a major complication rate of up to 42%.16, 17, 18, 19 This may in part be due to patients with severe tricuspid regurgitation (TR) receiving referrals for operations late in their disease course after exhaustion of medical management, exacerbating adverse outcomes.^5,19 Concerns for ventricular dysfunction and risks of liver or kidney disease perpetuate the hesitation to operate among this patient cohort.²⁰ In addition, concomitant valve procedures also carry substantial risk associated with multivalvular pathologies²¹ in which prior studies have established a mortality rate of up to 15%.²² Although there is a potential benefit to the simultaneous consideration of multivalve repair, these procedures have increased case complexity, operative times, and an ultimate risk. Although transcatheter treatment options have emerged to mitigate adverse events for patients not suitable for traditional surgical correction, standardized care pathways should be implemented to assess patient risk and improve shared decision-making.^17,21, 22, 23, 24, 25

Indeed, the present study demonstrates a volume-outcome relationship exists in tricuspid valve procedures with higher-volume centers, defined as those performing 12 or more operations annually, to demonstrate lower odds of mortality, MAEs, and postoperative complications. Furthermore, adjusted rates of MAE declined with increasing TVR and all cardiac valve volumes, supporting national consensus statements recommending procedural volume to assess heart valve center performance.²⁶ In a statewide analysis, Shang et al²⁷ reported centers with an annual volume of 12 cases, comprising the highest quartile of institutions in New York to maintain lower risk-adjusted rates of mortality. In addition, the greater procedural experience as a result of increased annual valve operations can directly translate into improved patient outcomes through enhanced technical proficiency and effective management of complex postoperative challenges. Persistent disparities in outcomes support the need for regionalization of tricuspid operations despite potential difficulties including limiting access and practice patterns. Efforts to develop appropriate triage and referral networks are needed to reduce the risk of TVR at the population level.

In the present work, we noted a paradoxical increase in hospitalization costs with increasing TVR volume. The provision of advanced perioperative and postoperative care may increase expenditures due to high-volume centers investing in specialized physicians, advanced imaging, and intensive monitoring following surgical intervention.²⁸ Furthermore, higher rates of postoperative survival at high-volume institutions can contribute to greater costs due to longer lengths of stay, requiring increased resources for recovery.²⁹ Moreover, high-volume centers are more commonly metropolitan teaching institutions, which are shown to have 44% greater costs than other hospitals.²⁸ Despite greater index hospitalization costs, high-volume centers demonstrate significantly reduced 30-day nonelective readmissions and nonhome discharge which are key measures of value-based care. Such improved outcomes may offset initial cost differences by minimizing downstream health care utilization and readmission.²⁷ However, access to these benefits is not equitable, given patients with lower socioeconomic status, limited insurance, and those in medically underserved communities are disproportionately treated at low-volume centers, as demonstrated by the present analysis. Since geographical barriers and referral networks exacerbate these disparities, the expansion of high-performing programs is needed to provide accessible and value-based care.

With the rise of less invasive modalities, recent guidelines for the management of TR increasingly incorporates transcatheter repair, supported by the PASTE registry.³⁰ In a recent study, Wild et al³⁰ demonstrated high-risk patients treated with the PASCAL edge-to-edge repair system and achieved durable TR reduction and significant functional improvement at 1 year, with greater procedural success noted at experienced centers. Furthermore, this analysis highlights the importance of early referral and structured evaluation, including of anatomic factors such as coaptation gap and right ventricular dilation predicting residual TR.³⁰ As a result, standardized care pathways for tricuspid valve patients at high-volume centers may benefit from such care options. In parallel, the Tricuspid Valve Academic Research Consortium has established standardized definitions for TR mechanism, grading, and trial endpoints to unify outcome assessment across surgical and transcatheter therapies.³¹ Adoption of the Tricuspid Valve Academic Research Consortium (TVARC) 5-grade TR severity scheme and distinctions between secondary causes of TR, may further refine patient selection and risk stratification.³¹ Together, these advances provide important context for the present analysis in support of ongoing efforts to improve patient stratification as well as structured growth in providing these treatment options at experienced facilities.

Study limitations

This study has several important limitations inherent to its retrospective design.

The National Inpatient Sample (NIS) is an administrative database lacking granularity in clinical data. As such, we are unable to adjust for potentially important prognostic determinants such as the severity of tricuspid valve disease, left ventricular function, or laboratory values.

In addition, ICD-10 coding may be subject to bias by physician and institutional billing preferences. Although risk-adjusted for within the analysis, differences in patient and procedural characteristics between volume groups may confound outcomes. Low-volume centers may commonly treat patients with lower income and greater comorbidities. However, our analysis demonstrated greater individual comorbidities at low-volume centers, but a similar distribution of overall Elixhauser Comorbidity Index across high- and low-volume institutions. Moreover, we excluded endocarditis patients due to a lack of granular information and complexity, a factor that may introduce bias and should be further evaluated in institutional analyses. Furthermore, we are limited by the inclusion of both isolated and concomitant tricuspid valve procedures. Thus, we used doubly robust statistical risk adjustment and the largest all-payer inpatient database to reduce bias and enhance the generalizability of our findings.

In conclusion, this national analysis demonstrates hospital tricuspid valve volume to be associated with reduced mortality and morbidity but incurred greater hospitalization costs. Our findings underscore a complex relationship between improved clinical outcomes and cost expenditures, warranting further investigation to enhance cardiac surgical quality and optimize resource utilization in emerging tricuspid valve programs.

Perspectives.

COMPETENCY IN MEDICAL KNOWLEDGE: Greater institutional procedure volume is associated with improved outcomes among those receiving TVR.

COMPETENCY IN SYSTEMS-BASED PRACTICE: These findings demonstrate the significance of institutional experience and regionalization of care.

TRANSLATIONAL OUTLOOK: As a result, these findings play a role in the development of standardized care pathways that optimize postoperative complications and readmission risk.

Funding support and author disclosures

Dr Benharash has received proctoring fees from AtriCure. This manuscript does not discuss any related products or services. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.