Abstract: Objective. The objective of our study is to compare transcatheter aortic valve replacement (TAVR) complication rates among teaching vs non-teaching centers in the United States. Methods. Using National Inpatient Sample (NIS) data, the largest all-payer database of hospital inpatient stay available in the United States, we identified patients (age ≥18 years) who underwent TAVR from January-December 2012. We constructed multivariable models to determine independent predictors (age, sex, race, Charlson’s comorbidity index, hospital size, hospital location, and TAVR approach) of TAVR-associated complications. Results. We identified 7405 TAVR procedures performed in the United States in 2012. In all, 88% of TAVRs were performed in teaching centers. There was no difference in mortality following TAVR between teaching and non-teaching centers. In-hospital complication rate was lower in teaching centers vs non-teaching centers (42% vs 50%, respectively; P<.001). In adjusted analysis, hemorrhage requiring transfusion (13.2% vs 20.8%; P<.001), renal complications requiring dialysis (1.2% vs 2.3%; P<.01), respiratory complications (7.5% vs 11%; P<.001), and complications requiring open-heart surgery (2% vs 4.6%; P<.001) were lower in teaching centers vs non-teaching centers. Vascular access-site, pacemaker insertion, pericardial, and neurological complications were similar between teaching and non-teaching centers. Conclusion. Institutional design impacts TAVR complications, albeit with no difference in mortality. In general, complication rates are lower in teaching centers compared with non-teaching centers.
J INVASIVE CARDIOL 2016;28(2):67-70
Key words: renal complications, transcatheter aortic valve replacement
In the recent era, transcatheter aortic valve replacement (TAVR) has become a cornerstone in the management of severe aortic stenosis. With advancement in techniques and expanding indications, the utilization of TAVR is rapidly rising among various sites in the United States (including both teaching and non-teaching centers). However, complications may occur in almost one-third of patients1 and include vascular access-site related, cardiac, neurological, respiratory, renal, and death. While appropriate patient selection and better techniques are the keys to better outcomes, the impact of institutional design on TAVR outcomes has not yet been studied. In this study, we analyzed whether the complication rates differed among patients undergoing TAVR at teaching vs non-teaching centers.
We used the Healthcare Cost and Utilization Project – National Inpatient Sample (NIS) database, sponsored by the Agency for Healthcare Research and Quality (AHRQ),2 and performed a retrospective observational cohort study. We identified patients (age ≥18 years) who underwent TAVR from January-December 2012 among teaching and non-teaching centers. The hospital’s teaching status was obtained from the American Heart Association Annual Survey of Hospitals. The facility is considered to be a teaching hospital if it has an American Medical Association-approved residency program, is a member of the Council of Teaching Hospitals (COTH), or has a ratio of full-time equivalent interns and residents to beds of .25 or higher.
The primary outcome of the study was all-cause in-hospital mortality. The secondary outcomes were occurrence of any in-hospital complication (hemorrhage requiring transfusion, vascular access-site complication, pacemaker insertion, pericardial, neurological, respiratory, renal complication requiring dialysis, complications requiring open-heart surgery, or extracorporeal membrane oxygenation), which was identified using respective ICD-9-CM codes (Supplementary Tables 1 and 2). We defined severity of comorbid conditions using Deyo’s modification of Charlson’s comorbidity index. This index contains 17 comorbid conditions with differential weights. The score ranges from 0 to 33, with higher scores corresponding to greater burden of comorbid diseases.3 The bed-size cut-off points into small, medium, and large have been done so that approximately one-third of the hospitals in a given region, location, and teaching status combination would fall within each bed-size category.4
The NIS database is the largest all-payer database (Medicare, Medicaid, private insurance, and uninsured) of hospital inpatient stay available in the United States (excluding federal, institutional, and short-term rehabilitation hospitals). Data from the NIS have been previously used to identify, track, and analyze national trends in health-care utilization, patterns of major procedures, trends in hospitalizations, charges, quality, and outcomes.5-7 We constructed multivariable models to determine the independent association of patient and hospital characteristics (age, sex, race, Charlson’s comorbidity index, hospital size, hospital location, and TAVR approach) with TAVR complications. Stata IC 11.0 (Stata-Corp) and SAS 9.3 (SAS Institute, Inc) were utilized for analyses. Differences between categorical variables were tested using the Chi2 test and differences between continuous variables were tested using student’s t-test, continuous variables with Gaussian distributions, and Kruskal-Wallis rank-sum tests for continuous variables with non-Gaussian distributions. Multivariate logistic regression was generated in order to identify the independent multivariate predictors of in-hospital mortality. All interactions were thoroughly tested. Level of significance was set at P<.05.
Between January and December 2012, we identified 7405 TAVR procedures performed in the United States. The majority of these procedures were performed in teaching centers (88%) (Table 1). The mean age of the patients at teaching and non-teaching centers was 81 years and 82 years, respectively. Gender distribution was similar between the teaching and non-teaching centers. Medicare was the primary payer in 90% of patients. The majority of patients had Charlson’s comorbidity index ≥2 (71% in teaching centers vs 68% in non-teaching centers). The femoral artery was the predominant approach site (86% in teaching center vs 81% in non-teaching centers). More patients had length of stay ≥6 days (58% in teaching centers vs 65% in non-teaching centers).
The primary outcome of the study (all-cause in-hospital mortality) was not significantly different between teaching centers (5.1%) and non-teaching centers (4.6%; P=.62) (Figure 1). The secondary outcome of the study (the occurrence of any complication) was significantly lower in teaching centers (42%) vs non-teaching centers (50%; P<.001). Hemorrhage requiring transfusion (13.2% vs 20.8%; P<.001), renal complication requiring dialysis (1.2% vs 2.3%; P<.01), respiratory complications (7.5% vs 11%; P<.001), and complications requiring open-heart surgery (2% vs 4.6%; P<.001) were significantly lower in teaching centers vs non-teaching centers, respectively. On the other hand, vascular access-site complications (9.7% vs 9.2%; P=.66), pacemaker insertion (8.6% vs 7.5%; P=.30), pericardial (1.5% vs 1.7%; P=.55), and neurological complications (1.5% vs 1.2%; P=.30) were similar between teaching and non-teaching centers.
After multivariate adjustment for patient and hospital characteristics (age, sex, race, Charlson’s comorbidity index, hospital size, hospital location, and TAVR approach), the occurrence of any complication was lower in the teaching centers compared with non-teaching centers (Figure 2).
Our study has several important findings. First, the majority of TAVR procedures were performed at teaching facilities. Second, baseline patient demographics and characteristics did not differ among patients undergoing TAVR at teaching vs non-teaching hospitals. Third, with the majority of TAVRs being performed in teaching hospitals, there was no difference in in-hospital mortality between teaching and non-teaching centers. Finally, overall complications rates were higher in non-teaching centers, including hemorrhage requiring transfusion, renal complications requiring dialysis, respiratory complications, and complications requiring open-heart surgery.
The relationships presented here should be considered hypothesis generating and therefore require further corroboration. Although we cannot ascertain the reasons behind our findings, we speculate that hospital and operator volume play a major role. Hospital volume plays an important role in TAVR-associated complication rates, as demonstrated in other catheter-based procedures like percutaneous coronary intervention.8 In fact, a recent study by Badheka et al reported that higher annual hospital volumes are clearly predictive of lower complications following TAVR.9 The Center for Medicare and Medicaid Services (CMS) requires TAVR centers to have ≥50 total aortic valve replacements in the year prior to TAVR, including 10 high-risk patients performed by ≥2 physicians with cardiac surgery privileges. Institutions must also meet the CMS requirement of ≥1000 catheterizations per year (including 400 PCIs per year) to initiate a TAVR program in hospitals without current TAVR experience.10 Hence, it appears that important keys to improving institutional outcomes within TAVR centers depend upon the selection of sites with a comprehensive teaching medical center where experienced operators perform large volumes of both surgical and transcatheter procedures.
Study limitations. Our study has major limitations inherent to post hoc analysis of an administrative database, including lack of long-term follow up data, clinical information (medication use, core lab data, etc), and errors related to coding. Furthermore, the absence of such variables limits detailed exploration of causal factors resulting in the differences observed. We used the Charlson comorbidity index as a marker of patient severity instead of using direct comparison of various comorbidities. Finally, our study was based on data from 2012, soon after TAVR was introduced in the United States. As TAVR technology, such as equipment and delivery systems, advances, one can expect the complication rates to decline over the years.
1. Fassa AA, Himbert D, Vahanian A. Mechanisms and management of TAVR-related complications. Nat Rev Cardiol. 2013;10:685-695.
2. HCUP Nationwide Inpatient Sample (NIS): Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality, Rockville, MD (2007). Available at: http://www.hcup-us.ahrq.gov/db/nation/nis/NIS_2007_INTRODUCTION.pdf. Accessed on Nov 12, 2014.
3. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373-383.
4. State and county QuickFacts — U.S. census bureau. Available at: http:// quickfacts.census.gov. Accessed on Jan 5, 2014.
5. Deshmukh A, Kumar G, Pant S, Rihal C, Murugiah K, Mehta JL. Prevalence of Takotsubo cardiomyopathy in the United States. Am Heart J. 2012;164:66-71.e1.
6. Deshmukh A, Pant S, Kumar G, Badheka AO, Paydak H. Impact of obesity on atrial fibrillation hospitalization. Int J Cardiol. 2012;159:241-242.
7. Deshmukh A, Pant S, Kumar G, Bursac Z, Paydak H, Mehta JL. Comparison of outcomes of weekend versus weekday admissions for atrial fibrillation. Am J Cardiol. 2012;110:208-211.
8. Badheka AO, Patel NJ, Grover P, et al. Impact of annual operator and institutional volume on percutaneous coronary intervention outcomes: a 5-year United States experience (2005-2009). Circulation. 2014;130:1392-1406.
9. Badheka AO, Patel NJ, Panaich SS, et al. Effect of hospital volume on outcomes of transcatheter aortic valve implantation. Am J Cardiol. 2015;116:587-594. Epub 2015 May 21.
10. Center for Medicare and Medicaid Services. 2015. National coverage determination (NCD) for transcatheter aortic valve replacement. Available at http://www.cms.gov/medicare-coverage-database/details/ncd. Accessed on May 30, 2015.
11. Gooley RP, Talman AH, Cameron JD, Lockwood SM, Meredith IT. Comparison of self-expanding and mechanically expanded transcatheter aortic valve prostheses. JACC Cardiovasc Interv. 2015;8:962-971.
From the 1University of Louisville, Louisville, Kentucky; 2Western Reserve Health Education, Youngstown, Ohio; 3Staten Island University, Staten Island, New York; 4UT Southwestern, Houston, Texas; 5Yale New Haven Medical Center, New Haven, Connecticut; 6Detroit Medical Center, Detroit, Michigan; 7Saint Peter’s University Hospital, Jersey City, New Jersey; and 8Mayo Clinic, Rochester, Minnesota.
Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.
Manuscript submitted August 14, 2015, provisional acceptance given September 24, 2015, final version accepted December 19, 2015.
Address for correspondence: Dr Sadip Pant, Department of Cardiovascular Medicine, University of Louisville School of Medicine, 550 S Jackson St, ACB 3rd Floor, Louisville, KY 40202. Email: email@example.com