Author + information
- Received May 26, 2017
- Revision received October 9, 2017
- Accepted October 17, 2017
- Published online December 11, 2017.
- Chad J. Zack, MDa,
- Erin A. Fender, MDa,
- Pranav Chandrashekar, MBBSa,
- Yogesh N.V. Reddy, MBBSa,
- Courtney E. Bennett, DOa,b,
- John M. Stulak, MDc,
- Virginia M. Miller, PhDc,d and
- Rick A. Nishimura, MDa,∗ ()
- aDepartment of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
- bDepartment of Critical Care Medicine, Mayo Clinic, Rochester, Minnesota
- cDepartment of Surgery, Mayo Clinic, Rochester, Minnesota
- dDepartment of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, Minnesota
- ↵∗Address for correspondence:
Dr. Rick A. Nishimura, Department of Cardiovascular Diseases, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905.
Background Severe isolated disease of the tricuspid valve (TV) is increasing and results in intractable right heart failure. However, isolated TV surgery is rarely performed, and there are little data describing surgical outcomes.
Objectives The purpose of this study was to evaluate contemporary utilization trends and in-hospital outcomes for isolated TV surgery in the United States.
Methods Patients age >18 years who underwent TV repair or replacement from 2004 to 2013 were identified using the National Inpatient Sample. Patients with congenital heart disease, with endocarditis, and undergoing concomitant cardiac operations except for coronary bypass surgery were excluded.
Results Over a 10-year period, a total of 5,005 isolated TV operations were performed nationally. Operations per year increased from 290 in 2004 to 780 in 2013 (p < 0.001 for trend). In-hospital mortality was 8.8% and did not vary across the study period. Adjusted in-hospital mortality for TV replacement was significantly higher than TV repair (odds ratio: 1.91; 95% confidence interval: 1.18 to 3.09; p = 0.009).
Conclusions Isolated TV surgery is rarely performed, although utilization has increased over time. However, despite an increase in surgical volume, operative mortality has not changed. Mortality is greatest in patients undergoing valve replacement. Given the increasing prevalence of isolated TV disease in the population, research into optimal surgical timing and patient selection is critical.
- isolated tricuspid valve disease
- national inpatient sample
- tricuspid regurgitation
- tricuspid valve repair
- tricuspid valve replacement
The prevalence of severe disease of the tricuspid valve (TV) is increasing, but surgery for isolated TV disease is rarely performed despite guideline recommendations to intervene in patients with symptomatic primary tricuspid regurgitation (TR) or tricuspid stenosis (1–5). Patients with severe TV disease may present with progressive and intractable right heart failure. Isolated TR is associated with increased mortality even in the absence of other cardiopulmonary comorbidities; yet, surgical referral has been limited by concerns over in-hospital mortality, which is reported to be as high as 24% (6–9). Existing observational cohort studies of TV surgery have included patients with congenital heart disease or those undergoing concomitant left-sided valve surgery, and therefore, may not accurately reflect the mortality associated with stand-alone TV surgery in the general adult population. Furthermore, single-center experiences with isolated surgery are limited by small numbers, and as a result include patients treated prior to the adoption of modern surgical practices or referral patterns (10). There is a critical gap in our knowledge of contemporary operative outcomes in patients with noncongenital heart disease undergoing isolated TV surgery. We sought to address these knowledge deficits using a large national cohort of patients undergoing isolated TV surgery. Defining modern surgical outcomes is particularly relevant now, as new percutaneous therapies for treatment of TV disease are in development (11).
Data were obtained from the Agency for Healthcare and Research and Quality Healthcare Cost and Utilization Project National Inpatient Sample (NIS) files from January 1, 2004, through December 31, 2013. The NIS is the largest, publicly available, all-payer administrative claims database in the United States (12). It contains deidentified patient and clinical data from approximately 1,000 nonfederal hospitals in 45 states, including approximately 5 to 8 million discharges annually. These data represent roughly 20% of hospital admissions in the United States, and the collection of data is stratified to ensure equal representation of hospitals by geographic region, urban and rural locations, teaching status, and hospital bed size. Procedure and diagnostic codes are recorded using International Classification of Diseases-9th Revision-Clinical Modification (ICD-9-CM). As deidentified and publicly available data were used, this study was exempt from institutional review board evaluation.
Using ICD-9-CM codes (Online Table 1), adult patients (age >18 years) were identified who underwent TV repair or replacement (mechanical or bioprosthetic) without concomitant aortic, mitral, or pulmonary valve surgery (Figure 1, Table 1). Patients undergoing TV surgery, with or without coronary artery bypass surgery (CABG) were included. Concomitant CABG was included because patients undergoing operation for their tricuspid valve may have had coronary disease identified on pre-operative angiography. Patients receiving TV surgery combined with aortic or pericardial surgery, orthotopic heart transplant, atrial or ventricular septal defect closure, or ventricular assist device implantation were excluded. Patients were also excluded if they had any diagnosis of congenital heart disease or active endocarditis, because in these cases, clinical outcomes are driven by the underlying disease process and not necessarily by the operation itself.
The main outcomes of interest included the number of operations performed annually, the type of surgery performed (repair vs. replacement), in-hospital mortality, post-operative morbidity, total cost of hospitalization, length of hospital stay, and discharge destination.
The NIS is clustered into individual hospitals, which induces dependence within hospitals. Patients receiving care at a particular institution share similar resources amongst each other that may differ from resources available at another hospital. Data is also stratified by hospital characteristics, such as geographic region, bed size, ownership, urban or rural location, and teaching status. Analysis using survey statistics was structured to account for this complex correlated design.
Descriptive statistics are presented as frequencies with percentages for categorical variables. Continuous data are reported as medians with interquartile range (IQR). Baseline characteristics were compared between groups using a Pearson chi-square test or Fisher exact test when expected counts were <5 for categorical variables, or Kruskal-Wallis H test for continuous variables. A Cochran-Armitage test was used to evaluate trends in isolated TV surgery over time. The chi-square test of independence was used to establish differences in proportions for multiple groups.
Data can be weighted such that the results can be extrapolated as representative of the entire U.S. inpatient population. Although weighting was developed primarily to perform trend analysis, it can also be utilized for all analyses per the Agency for Healthcare and Research and Quality Healthcare Cost and Utilization Project. Unweighted counts were used for all statistical analyses except trend assessments, where nationally weighted estimates were used. Nationally weighted characteristics and outcomes are available in Online Tables 2 and 3.
Risk factors for in-hospital mortality were assessed with SUDAAN version 11.0.1 (RTI International, Research Triangle Park, North Carolina) using univariate and multivariate hierarchical logistic regression with an exchangeable matrix via the generalized estimating equation model-fitting method (Taylor linearization) to estimate odds ratios and their associated 95% confidence intervals (CIs). Variance correction was used to derive valid results in the presence of intracluster correlations. Factors found to have a p value < 0.10 in the univariate analysis (Online Table 4) or those felt to be clinically relevant were included in the multivariate model and assessed in a stepwise fashion. A 2-sided threshold of significance was established at p < 0.05. All statistical analyses were performed using SPSS version 22 (IBM, Armonk, New York), SUDAAN version 11.0.1 (RTI International), and SAS version 9.3 (Cary, North Carolina).
Baseline characteristics of the study population
From January 2004 through December 2013, there were a total of 5,005 admissions nationally for isolated TV surgery. The majority of patients were female (57.5%). The median age was 62.0 years (48.0 to 72.0 years), 11.1% had a pacemaker or intracardiac defibrillator, and 49.7% had atrial fibrillation (Table 1). The Charlson Comorbidity Index score was ≥2 in 57.3% of patients. A total of 12.5% of patients had a history of previous cardiac surgery, with 9.4% having undergone previous valve surgery.
Temporal trends and outcomes for isolated TV surgery
Between January 2004 and December 2013, the number of isolated TV operations significantly increased from 290 procedures annually in 2004 to 780 procedures annually in 2013 (p < 0.001) (Central Illustration). The median length of stay was 11 days (IQR: 7 to 20 days). The majority of operations were performed at teaching hospitals (86%), and no significant difference in mortality between teaching and nonteaching hospitals was observed (8.4% vs. 11.5%; p = 0.22). Overall in-hospital mortality for isolated TV surgery was 8.8%, and this was consistent across all study years (Table 2). Furthermore, among those that did experience in-hospital mortality after isolated TV surgery, the median time between surgery and death was 14.5 days (IQR: 4.0 to 35.8 days).
Hierarchical multivariate logistic regression was performed to identify independent predictors associated with in-hospital mortality (Table 3). After adjustment, increased mortality was associated with age ≥60 years, coagulopathy, type of TV surgery, and end-stage renal disease.
To address the effect of concomitant CABG, a sensitivity analysis was performed excluding those patients who underwent concomitant CABG as well as subjects with a primary billing diagnosis of coronary artery disease. This led to an exclusion of approximately 12% of patients, but did not change the primary or secondary outcomes (Online Table 5).
Effect of type of valve surgery: Repair versus replacement
TV replacement was performed in 59.2% of patients, whereas TV repair was performed in 40.8%. Of the TV replacements, 60.7% were replaced using a bioprosthetic valve, and 39.3% received a mechanical valve. From 2004 to 2013, the proportion of TV replacements significantly decreased from 67.2% to 57.1% (p = 0.003).
Patients treated with a mechanical valve were significantly younger (age 54.0 years [IQR: 41.8 to 70.3 years]) than patients who were treated with repair (age 64.0 years [IQR: 51.5 to 73.0 years]) or bioprosthetic TV replacement (age 61.5 years [IQR 48.0 to 72.3 years]) (p < 0.001) (Table 4). Patients undergoing TV repair were significantly more likely to have coronary artery disease (45.6% vs. 25.3%; p < 0.001) and receive concomitant CABG (36.5% vs. 8.6%; p < 0.0001) than patients undergoing TV replacement. Conversely, patients undergoing TV replacement were more likely to have cirrhosis (1.6% vs. 5.5%; p < 0.001), an intracardiac device (14.3% vs. 6.6%; p < 0.0001), or history of previous valve surgery (5.9% vs. 11.8%; p < 0.001) than those undergoing TV repair.
Mortality occurred in 5.9% of patients treated with TV repair, 9.1% of patients treated with a bioprosthetic valve, and 13.6% treated with a mechanical replacement (p = 0.003). After adjustment, TV replacement was associated with greater mortality compared with TV repair (OR: 1.91 [95% CI: 1.18 to 3.08]; p < 0.009). There was a significantly higher rate of blood transfusion (39.3% vs. 33.2%; p < 0.001) and permanent pacemaker implantation (35.0% vs. 13.4%; p < 0.001) in patients undergoing TV replacement. Patients undergoing a mechanical TV replacement were more likely to have a post-operative cerebrovascular accident compared with a bioprosthetic TV replacement (4.1% vs. 1.1%; p < 0.001). Median hospital charges were $161,500 (IQR: $105,700 to $261,300) and did not vary by the type of surgery performed (p = 0.32). The median length of stay was 11 days, and was similar for both repair and replacement (p = 0.11).
The current study demonstrates that isolated TV surgery in the United States has increased significantly from 2004 to 2013. However, isolated TV surgery remains rare, particularly in comparison to the overall prevalence of significant tricuspid disease, with nearly 1.6 million Americans estimated to have greater than moderate-severe TR (1,3). The rarity of TV surgery may be driven by a high perceived operative mortality, as high as 24% in previous reports; we observed the mortality to still be high: 8.8% in this contemporary national sample (6,13–15). Despite increased utilization of isolated TV surgery and improvement in operative techniques, operative mortality did not decrease over the study period. The costs of surgery were high, which likely reflects the long length of stay (median 11 days), high utilization of pacemakers (26%), and in-hospital mortality (8.8%). Tricuspid valve replacement, age >60 years, ESRD, and coagulopathy predicted greater in-hospital mortality, with a median time from operation to death of almost 2 weeks. The high mortality rate of isolated TV surgery is in stark contrast to what is seen in isolated aortic or mitral valve surgery. Isolated tricuspid procedures have carried the highest risk of all valve surgery for some time (16). Studies from the Society of Thoracic Surgeons National Database report the contemporary in-hospital mortality of isolated aortic valve and isolated mitral valve surgery as approximately 2% to 3% (17,18). Given the multiple comorbid conditions present in this population, it is possible that the high mortality is from patients not being referred to surgery until they are end-stage with severe right ventricular (RV) dysfunction and end-organ damage, as opposed to the paradigm for the left-sided valves, where surgery is performed prior to the onset of LV dysfunction or symptoms.
The highest in-hospital mortality for isolated TV disease was in the TV replacement cohort, who experienced 10.9% inpatient mortality (3,19–21). The higher mortality in the TV replacement cohort occurred despite a mean age that was approximately 3 years younger than the TV repair groups. The higher mortality with valve replacement may represent a selection bias, as some surgeons would place a prosthesis in those sicker patients with end-stage annular dilatation and severe noncoaptation of the valve leaflets with poorer RV function to avoid reoperation. However, as with valve repair in the aortic and mitral position, the valve repair itself may result in better outcome than replacement due to better maintenance of RV function with preservation of tricuspid continuity and fewer prosthetic valve–related complications (18,22). Comprehensive repair techniques for the tricuspid valve now exist, and further investigation will be needed to determine if surgeons should strive to perform an increasing percentage of valve repair in this group of patients with severe primary TR.
The 2014 American College of Cardiology/American Heart Association valvular heart disease guidelines provide a Class Ic recommendation for isolated TV surgery in severe symptomatic tricuspid stenosis, a Class IIa recommendation for isolated TV surgery in primary severe TR patients who are unresponsive to medical therapy, and a Class IIb recommendation for isolated TV surgery in asymptomatic patients with severe primary TR who have evidence of moderate or more RV dilation or dysfunction (4). Despite these recommendations, there remains a large gap between the number of affected patients and the number undergoing definitive correction. Previous studies have reported a wide range of inpatient outcomes following TV surgery. Many of these studies were limited to a single center assessing a small number of patients treated over decades, included TV surgery combined with other valve operations, or included patients undergoing TV surgery for congenital heart disease (3,9,10,23,24). A previous publication by Vassileva et al. (24) described operative outcomes using the NIS database for TV surgery from 1999 to 2008 and found 11% in-hospital mortality. However, this study included a more heterogeneous population with both concomitant and isolated TV surgery, as well as those with infective endocarditis. In this study, patients with congenital heart disease and active endocarditis were excluded to best represent the population seen in clinical practice. Despite the exclusion of patients with congenital heart disease, with active endocarditis, and receiving multivalve surgery, operative mortality remained unexpectedly high. There was a high percentage of patients presenting with severe TR who were in atrial fibrillation or had prior intracardiac devices placed, both of which are a growing cause of TR (15,25). As these 2 causes of TR will continue to increase, there may be a growing number of patients who will present with severe primary TR in the future.
Currently, multiple novel percutaneous TV devices are being developed (11) that may ultimately provide a lower-risk alternative treatment for symptomatic patients who require interventions. However, larger trials with long-term outcomes are needed before these devices can be implemented into broad clinical practice.
First, the NIS is an administrative database, which classifies patients and procedures according to ICD-9-CM codes and may therefore be prone to coding errors. The NIS attempts to mitigate potential errors by using internal and external quality control measures. Second, the use of ICD-9-CM codes limits the granularity of data that can be collected. The ICD-9-CM codes for TV disease do not distinguish between tricuspid stenosis and regurgitation. Therefore, it is impossible to know how many study patients underwent isolated tricuspid surgery for TV stenosis versus TV regurgitation. We further cannot discern what the exact mechanism for TV disease was. Codes for rheumatic TV disease are available in ICD-9-CM; however, medical coders are taught to presume all TV disease is rheumatic unless explicitly stated otherwise by the clinician. As a result, the code for rheumatic disease is highly unreliable and was not used for this study. Furthermore, important data on RV size and function and pulmonary arterial pressures are not recorded in NIS. These variables could represent the true cause of increased mortality observed in patients undergoing valve replacement. Additionally, the definition of isolated TV surgery in our study included a concomitant CABG. Due to limitation in the ability to distinguish the primary indication for the surgical procedure, it is hard to know if the primary reason for the surgery was coronary artery disease with an adjunct TV repair performed, or if it was severe TV disease. We attempted to address this limitation by performing a sensitivity analysis and found our results did not change. Last, the NIS contains only inpatient data without the ability to identify patients with subsequent hospital readmissions. Long-term changes in functional status or mortality cannot be assessed using NIS. Despite these limitations, this study fills an important gap in the published data by addressing the national prevalence of TV surgery, defining operative mortality, and assessing mortality by valve repair or replacement.
Isolated TV surgery remains rare despite the high prevalence of TV disease in the population; however, the number of procedures performed annually is increasing. Mortality rates are higher than for intervention on the left-sided valves and have not improved despite increasing surgical utilization. TV replacement carries the greatest risk for surgical mortality. Long-term follow-up data after isolated TV surgery is lacking, and data on long-term clinical outcomes are required to determine the ultimate role as well as the type and timing of intervention.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: The prevalence of severe TV disease is increasing. Surgery for isolated tricuspid disease is rarely performed and is associated with in-hospital mortality of 8.8%, much higher than the mortality for isolated aortic or mitral valve surgery. Mortality is highest in patients undergoing TV replacement.
TRANSLATIONAL OUTLOOK: More research is needed to define patient selection and optimum timing for TV surgery. Randomized trials should be designed to determine whether surgery can improve survival and when percutaneous devices might be useful strategies for patients with various forms of TV disease.
This paper was written in partial fulfillment of requirements for the Mayo Clinic Graduate School of Medicine Biomedical Engineering (BME) 6855 Tutorial in Cardiovascular Physiology course.
This publication was made possible by funding from the National Institutes of Health CTSA grant number UL1 TR000135 from the National Center for Advancing Translational Sciences and grant number P50 AG044170 (to Dr. Miller). The contents of this paper are solely the responsibility of the authors and do not necessarily represent the official view of the National Institutes of Health. The authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Zack and Fender contributed equally to this work and are joint first authors.
- Abbreviations and Acronyms
- coronary artery bypass graft surgery
- confidence interval
- International Classification of Diseases-9th Revision-Clinical Modification
- interquartile range
- National Inpatient Sample
- tricuspid regurgitation
- tricuspid valve
- Received May 26, 2017.
- Revision received October 9, 2017.
- Accepted October 17, 2017.
- 2017 American College of Cardiology Foundation
- Nath J.,
- Foster E.,
- Heidenreich P.A.
- Nishimura R.A.,
- Otto C.M.,
- Bonow R.O.,
- et al.
- Vahanian A.,
- Alfieri O.,
- Andreotti F.,
- et al.
- Rodes-Cabau J.,
- Hahn R.T.,
- Latib A.,
- et al.
- ↵Healthcare Cost and Utilization Project. Overview of the Nationwide Inpatient Sample (NIS): Healthcare Cost and Utilization Project (HCUP). Available at: http://www.hcup-us.ahrq.gov/nisoverview.jsp. Accessed February 28, 2016.
- Topilsky Y.,
- Nkomo V.T.,
- Vatury O.,
- et al.
- Brown J.M.,
- O'Brien S.M.,
- Wu C.,
- Sikora J.A.,
- Griffith B.P.,
- Gammie J.S.
- Vassileva C.M.,
- Mishkel G.,
- McNeely C.,
- et al.
- Thourani V.H.,
- Weintraub W.S.,
- Guyton R.A.,
- et al.
- Chang J.D.,
- Manning W.J.,
- Ebrille E.,
- Zimetbaum P.J.