- Email: [email protected]
Failure to Rescue, Hospital Volume, and In-Hospital Mortality After Transcatheter Aortic Valve Implantation
Accepted Manuscript
Failure to Rescue, Hospital Volume, and In-Hospital Mortality After Transcatheter Aortic Valve Implantation Tomo Ando MD , Oluwole Adegbala MD, MPH , Pedro A Villablanca MD, MS , Mohamad Shokr MD , Emmanuel Akintoye MD, MPH , Alexandros Briasoulis MD, PhD , Hisato Takagi MD. PhD , Theodore Schreiber MD , Cindy L. Grines MD , Luis Afonso MD PII: DOI: Reference:
S0002-9149(18)31192-5 10.1016/j.amjcard.2018.05.020 AJC 23324
To appear in:
The American Journal of Cardiology
Received date: Revised date: Accepted date:
29 March 2018 8 May 2018 8 May 2018
Please cite this article as: Tomo Ando MD , Oluwole Adegbala MD, MPH , Pedro A Villablanca MD, MS , Mohamad Shokr MD , Emmanuel Akintoye MD, MPH , Alexandros Briasoulis MD, PhD , Hisato Takagi MD. PhD , Theodore Schreiber MD , Cindy L. Grines MD , Luis Afonso MD , Failure to Rescue, Hospital Volume, and In-Hospital Mortality After Transcatheter Aortic Valve Implantation, The American Journal of Cardiology (2018), doi: 10.1016/j.amjcard.2018.05.020
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Failure to Rescue, Hospital Volume, and In-Hospital Mortality After Transcatheter Aortic Valve Implantation 1
Tomo Ando MD, 2Oluwole Adegbala MD, MPH, 3Pedro A.Villablanca MD, MSc, Mohamad
Shokr MD 1Emmanuel Akintoye MD, MPH, 4Alexandros Briasoulis MD, PhD, 5Hisato Takagi
CR IP T
MD. PhD, 1Theodore Schreiber MD, 6Cindy L. Grines MD, 1Luis Afonso MD 1
M
AN US
Department of Medicine Division of Cardiology, Wayne State University/Detroit Medical Center, Detroit, Michigan, United States 2 Department of Medicine Department of Internal Medicine, Englewood Hospital and Medical Center, Seton Hall University-Hackensack Meridian School of Medicine, Englewood, New Jersey, United States 3 Department of Medicine Division of Cardiology, New York University Langone Medical Center, New York, New York, United States 4 Department of Medicine Divison of Cardiovascular Medicine, University of Iowa Hospitals and Clinics, Iowa, Iowa, United States 5 Department of Medicine Division of Cardiovascular Surgery, Shizuoka Medical Center, Shizuoka, Japan 6 Department of Medicine Division of Cardiology, North Shore University Hospital, Hofstra Northwell School of Medicine, Manhasset, New York, United States
ED
Running title: Failure to rescue after TAVI Disclosures: Authors have no disclosures
Tomo Ando MD
PT
Corresponding author:
CE
3990, John R, Detroit, Michigan, 48201, United States Detroit Medical Center, Division of Cardiology
AC
Phone: 313-745-2620 Fax: 313-745-8643
Email: [email protected]
1
ACCEPTED MANUSCRIPT
Abstract Failure to rescue (FTR), death after major complications, has been well described in the surgical literature as a source of different outcomes among different hospitals. However, FTR has not been investigated in transcatheter aortic valve implantation (TAVI). Our aim was to assess the
CR IP T
difference of in-patient mortality and FTR among different TAVI volume hospitals. We queried the Nationwide Inpatient Sample database from 2011 to 2015 to identify patients who had transarterial TAVI. FTR was calculated as those who had in-patient mortality with at least one with major perioperative complications. Hospitals were divided into three groups according to annual
AN US
TAVI volume, the lowest quintile (≤30/year), second to fourth quintile (31-130/year), and highest quintile (≥130/year). Multivariate analysis was used to calculate risk adjusted in-patient mortality rate and FTR and was compared between these different volume hospitals. A total of 48,886
M
TAVI procedures were identified (10,407, 28,811, and 9,668 in low, intermediate, and high volume centers, respectively). Mean age, percentage of female, and Elixhauser comorbidity
ED
index was similar across different TAVI volume hospital. Incidence of major perioperative complications did not differ among different volume hospitals. Adjusted rate of in-patient
PT
mortality (2.3%, 1.87%, and 1.57% for low, intermediate, and high volume center, respectively,
CE
p<0.001) were significantly lower with higher hospital volume but FTR (8.24%, 8.20%, and 6.12% for low, intermediate, and high volume center, respectively, p=0.29) were the same among
AC
the three groups. Our results suggest that FTR does not explain the variation of in-hospital mortality among different hospital volumes.
Key words: aortic stenosis, failure to rescue, transcatheter aortic valve implantation Failure to rescue (FTR), characterized as in-patient mortality with at least one with major perioperative complication, has been well described in the surgical literature as a source of
2
ACCEPTED MANUSCRIPT
variation in hospital outcomes (1-5). Transcatheter aortic valve implantation (TAVI) is associated with certain peri-procedural complications that negatively impact the outcomes and therefore assessing the rate of FTR would have incremental value in addition to evaluating periprocedural complication and mortality rate post-TAVI. With increase in-hospital procedural
CR IP T
volume, better outcomes have been reported in the past for various surgical and percutaneous procedures (6-13). Previous studies have reported similar trends, with decreasing adverse
outcomes with increased hospital volume following TAVI (14,15). However, these analyses reflected early United States commercial TAVI experience. With its rapid expansion and center
AN US
experience, TAVI has become a safer procedure and the outcomes of volume-outcome
relationship may have dramatically changed. Our aim was to assess the FTR among different TAVI volume hospital including the most recently available Nationwide Inpatient Sample (NIS)
M
database.
ED
Methods
PT
Data were obtained from the Agency for Healthcare Research and Quality Healthcare
CE
Cost and Utilization Project–NIS files between 2011 and 2015. The data was queried to identify patient demographics and TAVI procedure recipients in the United States using the International
AC
Classification of Diseases-Ninth Revision-Clinical Modification (ICD-9-CM). The NIS was created by the Healthcare Cost and Utilization Project as a discharge database and is maintained by the Agency for Healthcare Research and Quality. NIS data are extracted from a random sample of approximately 20% of all nonfederal, general, and specialty specific hospital inpatient admissions. Approximately 97% of hospitals in the United States are represented by the NIS,
3
ACCEPTED MANUSCRIPT
making it the largest all-payer inpatient discharge database in the United States. Criteria used for stratified sampling of hospitals into the NIS include hospital ownership, patient volume, teaching status, urban or rural location, and geographic region (16). Weighting the patient-level observations in the NIS data sets to account for the complex sampling scheme provides estimates
CR IP T
for the entire US population of hospitalized patients. The data set is publicly available and deidentified. Accordingly, it is a designated exempt from an Institutional Review Board approval.
Patients aged 50 years and older with aortic stenosis who underwent TAVI during the
AN US
study period were identified using International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes (supplemental table 1). We excluded patients with a diagnosis of aortic insufficiency without a diagnosis of aortic stenosis. We did not include the last 3
M
months of 2015 given the non-availability of Elixhauser comorbidity indicators in those months (17). Annual hospital volume was determined by obtaining the total number of TAVI cases
ED
performed by a given institution per annum using the unique Healthcare Cost and Utilization Project (HCUP) hospital number. The total institution was divided into quintiles before analysis.
PT
Low volume was defined by the lowest quintile, intermediate volume was defined by the second
CE
to fourth quintiles, and high volume was defined by the highest quintile. This formula resulted in the following breakdown of volume cutoff points. Low-volume institutions performed ≤ 30
AC
TAVIs/year, intermediate-volume institutions performed 31 – 130 TAVIs/year, and high-volume institutions performed ≥ 130 TAVIs/year.
Data on patient- and hospital-level characteristics were provided for each patient in the NIS. Our primary outcome was FTR defined as in-hospital mortality following major peri-
4
ACCEPTED MANUSCRIPT
operative complications. FTR was calculated by the numbers of patients who experienced inhospital mortality following at least one major perioperative complication divided by the total number of patients who had major perioperative complications. These major peri-operative complications include in-hospital mortality, stroke, acute myocardial infarction with
CR IP T
percutaneous coronary intervention, major bleeding requiring transfusion, acute kidney injury requiring dialysis, cardiogenic shock, fatal arrhythmia (ventricular fibrillation, ventricular tachycardia, or cardiac arrest), mechanical circulatory support, TAVI converted to surgical aortic valve replacement, vascular injury requiring surgery, and acute respiratory failure with
AN US
reintubation,. These mentioned outcomes were captured from the dataset with ICD-9-CM codes (Supplementary Table 1).
M
Age was classified as 50-59, 60-69, 70-79, 80-89 and >90 years. Race was catalogued as White, Black, Hispanic, Asian and others. Risk adjustment for comorbid medical conditions
ED
was performed using the Elixhauser comorbidity index, categorized by number of medical comorbidities (0, 1 -3, or ≥4). Individual household income was classified as low, medium, high,
PT
or highest in the NIS data set. Insurance classification was described as Medicaid, Medicare,
CE
private or others. Hospitals were classified by location (rural, urban), region (Northeast, Midwest, West, South), size (small, medium, large), and teaching status (rural, urban non-
AC
teaching, urban-teaching).
Patients were catalogued by institutional volume; low-, medium-, and high-volume
TAVI centers and were compared. Descriptive statistics were presented and compared across hospital volume levels, with chi-square tests for categorical variables, analysis of variance
5
ACCEPTED MANUSCRIPT
(ANOVA) for normally distributed continuous variables, and Kruskal-Wallis test for continuous variables with skewed distribution (Table 1). Mortality rates, complication rates, and rates of FTR were also reported descriptively across hospital volume levels. The adjusted mortality rate and adjusted failure to rescue rate were calculated by fitting a Poisson regression model with a
CR IP T
robust error variance under generalized estimating equations approach. The robust error variance was estimated by using repeated statement and individual subject identifier while adjusting for patient-level covariates, including age, race, gender, median household income, type of insurance, Elixhauser co-morbidity index, as well as hospital-level covariates, including hospital
AN US
location or teaching status, and year of data publication. Statistical significance of the mortality rates and FTR rate differences between the hospital volume categories was determined by the contrast of the regression coefficients from the Poisson regression model. Subsequently, pairwise
M
comparison was conducted with Bonferroni correction for multiple comparisons. All the data extraction and analyses was done with Statistical Analysis System (SAS V.9.4, SAS
ED
Institute Inc, Cary, NC, US). We chose a p-value of <0.05, reported the effect sizes, 95%
AC
Results
CE
comparisons.
PT
confidence intervals (CI), and p-values or the Bonferroni corrected p-values for multiple
There was a total of 48,886 TAVI patients identified (10,407, 28,811, and 9,668 in in
lowest, intermediate, and highest volume centers, respectively). The lowest and intermediate volume hospitals performed 1-30 and 31-130 TAVIs per year respectively, whereas the highest volume hospitals performed more than 130 TAVI per year. Mean age, percentage of female, and
6
ACCEPTED MANUSCRIPT
Elixhauser comorbidity index did not differ among the three different hospital groups. Prior percutaneous coronary intervention, coronary artery bypass graft, peripheral vascular disease, and race differed significantly across the three hospital groups. Large bed size hospitals, and high median household income were observed more frequently in the highest volume hospitals. These
CR IP T
results are summarized in table 1.
The overall in-hospital morality was 3.27%. The unadjusted rates of in-hospital mortality declined significantly with increasing hospital volume (4.28%, 3.21%, and 2.37% for
AN US
lowest, intermediate, and highest volume center, respectively, p=0.006). After adjustments for clinical and hospital variables, the results remained statistically significant (2.3% [1.53-3.52], 1.87% [1.24-2.81], and 1.57% [1.02-2.42] for lowest, intermediate, and highest volume center,
M
respectively, p=0.0002, supplement 2). Yearly trend of un-adjusted mortality in low, intermediate, and high volume hospitals are summarized in supplement 3. In general, the rate of
ED
major complications decreased with increasing hospital volume (except for acute myocardial infarction with percutaneous coronary intervention). Use of mechanical circulatory support was
PT
higher in low volume hospital. However, other complications were similar across different
CE
hospital volumes. The overall FTR was 12.28%. The unadjusted rate of FTR was almost similar between lowest and intermediate volume hospital (13.59% and 12.88%, respectively) and
AC
numerically lower in highest volume hospital (8.87%) but was statistically not significant (p=0.084). After multivariate adjustments, FTR rates were similar among all hospital volume groups (8.24% [3.39-20.03], 8.20% [3.45-19.52], and 6.12% [2.37-15.81] for lowest, intermediate, and highest volume center, respectively, p=0.29). These results are summarized in table 2.
7
ACCEPTED MANUSCRIPT
Discussion
The major findings of our results were 1: Rate of FTR also did not differ among
CR IP T
different volume hospitals. 2: In-patient mortality was lower in higher volume hospitals but major perioperative complications did not significantly decrease with increased hospital volume. Our results suggest that hospital-volume is an independent predictor of in-hospital mortality and
AN US
FTR is not the source of variation of in-hospital mortality among different hospital volume.
FTR and mortality did not differ among different volume hospitals following TAVI. Previous studies have shown that high procedural volume hospital was associated with higher
M
rates of FTR after cardiovascular surgery (9,18). Similar result was also reported in non-cardiac surgery (19). There could be several reasons why FTR was similar among different TAVI volume
ED
hospitals as opposed to previous studies. First, incident rate of the complications with extremely high mortality, requiring advanced medical resources and highly skilled operators to rescue, such
PT
as cardiac tamponade requiring pericardiocentesis, acute myocardial infarction requiring
CE
percutaneous coronary intervention, and conversion to surgical aortic valve replacement, were very low and may partly explain the similar FTR observed in our study. Secondary, even if the
AC
annual TAVI volume is low, TAVI is predominantly performed at teaching hospitals or large sized bed hospitals where access to sub-specialities and intensive care units are often readily available (20). At these hospitals, the more commonly observed perioperative complications may not be difficult to manage, resulting in similar FTR across different TAVI volume hospitals. Third, the standardization of TAVI endpoints by the Valve Academic Research Consortium may
8
ACCEPTED MANUSCRIPT
have helped physicians recognize the expected complications post-TAVI and expeditiously manage them across all TAVI volume hospitals (21).
Regarding the hospital volume-outcome relationship, our results were with previous
CR IP T
studies that demonstrated decreased in-patient mortality with increased hospital volume after TAVI. Badheka et al. reported from NIS database that with increasing hospital volume, inhospital mortality decreased from 6.4% for lowest volume hospitals to 2.8% for the highest TAVI volume hospitals (14). Similar results were reported by Kim et al. also from the NIS
AN US
database (15). These studies, however, only included data from 2012, which is considered a very early commercial TAVI era in the United States. Our result included more updated data from last 5 years of NIS reports, reflecting more mature TAVI programs and operator experience. Increase
M
in institutional experience has been reported with improved outcomes. Carroll et al. reported that with accumulation of procedural experience up to 400th cases, in-hospital mortality, vascular
ED
complications, and bleeding decreased significantly (22). Wassef et al. reported that > 242 cases of experience resulted in decrease of half of all-cause mortality (p=0.002) (23). Furthermore, in-
PT
hospital mortality of trans-femoral TAVI did not inversely relate with higher procedural
CE
experience according to the registry that included procedure in the United States from 2011 through 2015 (22). Considering these results, it is possible that trans-arterial TAVI has matured
AC
during the study period. Our results further strengthen the previous reports that strongly suggested an association between in-hospital mortality and TAVI volume, despite maturation of trans-arterial TAVI techniques, and adds to the existing literature the fact that FTR is not the source of variation in outcomes. In addition, although previous studies have shown decreased perioperative complications as well with increasing hospital volume, our study did not show
9
ACCEPTED MANUSCRIPT
similar trend (14,15). This is likely the result of better patient selection and learning curve.
There are several limitations that should be noted. First, this was a retrospective cohort study that used administrative database. Therefore, there is a possibility of coding error that
CR IP T
could have biased our results. However, the use of NIS database has been applied to wide variety of medical researches and is considered a valid resource. Second, some of the important clinical variables, such as the Society of Thoracic Surgeons score and frailty score were not available and hence not adjusted in multivariable analysis. Third, we were also unable to adjust for key
AN US
echocardiographic parameters such as ejection fraction and pulmonary pressure. Lastly, because the NIS database is formed based on hospitals in the United States, the volume cutoff for
AC
CE
PT
ED
M
different quintile hospitals may not apply to country outside the United States.
Acknowledgement: None
10
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
M
AN US
CR IP T
Disclosures: All authors have no disclosures
References
11
ACCEPTED MANUSCRIPT
1: Mehta A, Efron D, Stevens K, Manukyan MC, Joseph B, Sakran JV. Hospital Variation in Mortality after Emergency Bowel Resections: The Role of Failure-to-Rescue. J Trauma Acute Care Surg 2018;84:702-710. 2: Gani F, Johnston FM, Nelson-Williams H, Cerullo M, Dillhoff ME, Schmidt CR, Pawlik TM.
CR IP T
Hospital Volume and the Costs Associated with Surgery for Pancreatic Cancer. J Gastrointest Surg 2017;21:1411-1419.
3: Malone H, Cloney M, Yang J, Hershman DL, Wright JD, Neugut AI, Bruce JN. Failure to Rescue and Mortality Following Resection of Intracranial Neoplasms. Neurosurgery 2017 Aug 1
AN US
4: Amini N, Spolverato G, Kim Y, Pawlik TM. Trends in Hospital Volume and Failure to Rescue for Pancreatic Surgery. J Gastrointest Surg 2015;19:1581-1592.
5: Wright JD, Herzog TJ, Siddiq Z, Arend R, Neugut AI, Burke WM, Lewin SN, Ananth CV,
J Clin Oncol 2012;30:3976-3982.
M
Hershman DL. Failure to rescue as a source of variation in hospital mortality for ovarian cancer.
ED
6: Luft HS, Bunker JP, Enthoven AC. Should operations be regionalized? The empirical relation between surgical volume and mortality. N Engl J Med 1979;301:1364-1369.
PT
7: Birkmeyer JD, Siewers AE, Finlayson EV, Stukel TA, Lucas FL, Batista I, Welch HG,
CE
Wennberg DE. Hospital volume and surgical mortality in the United States. N Engl J Med 2002;346:1128-1137.
AC
8: Holt PJ, Poloniecki JD, Gerrard D, Loftus IM, Thompson MM. Meta-analysis and systematic review of the relationship between volume and outcome in abdominal aortic aneurysm surgery. Br J Surg 2007;94:395-403. 9: Gonzalez AA, Dimick JB, Birkmeyer JD, Ghaferi AA. Understanding the volume-outcome effect in cardiovascular surgery: the role of failure to rescue. JAMA Surg 2014;149:119-123.
12
ACCEPTED MANUSCRIPT
10: Singh V, Badheka AO, Patel NJ, Chothani A, Mehta K, Arora S, Patel N, Deshmukh A, Shah N, Savani GT, Rathod A, Manvar S, Thakkar B, Panchal V, Patel J, Palacios IF, Rihal CS, Cohen MG, O'Neill W, De Marchena E. Influence of hospital volume on outcomes of
Catheter Cardiovasc Interv 2015;85:1073-1081.
CR IP T
percutaneous atrial septal defect and patent foramen ovale closure: a 10-years US perspective.
11: Badheka AO, Chothani A, Mehta K, Patel NJ, Deshmukh A, Hoosien M, Shah N, Singh V, Grover P, Savani GT, Panaich SS, Rathod A, Patel N, Arora S, Bhalara V, Coffey JO, O'Neill W, Makkar R, Grines CL, Schreiber T, Di Biase L, Natale A, Viles-Gonzalez JF. Utilization and
AN US
adverse outcomes of percutaneous left atrial appendage closure for stroke prevention in atrial fibrillation in the United States: influence of hospital volume. Circ Arrhythm Electrophysiol 2015;8:42-48.
M
12: Kim LK, Swaminathan RV, Looser P, Minutello RM, Wong SC, Bergman G, Naidu SS, Gade CL, Charitakis K, Singh HS, Feldman DN. Hospital Volume Outcomes After Septal
ED
Myectomy and Alcohol Septal Ablation for Treatment of Obstructive Hypertrophic Cardiomyopathy: US Nationwide Inpatient Database, 2003-2011. JAMA Cardiol 2016;1:324-332.
PT
13: Opotowsky AR, Landzberg MJ, Kimmel SE, Webb GD. Percutaneous closure of patent
CE
foramen ovale and atrial septal defect in adults: the impact of clinical variables and hospital procedure volume on in-hospital adverse events. Am Heart J 2009;157:867-874.
AC
14: Badheka AO, Patel NJ, Panaich SS, Patel SV, Jhamnani S, Singh V, Pant S, Patel N, Patel N, Arora S, Thakkar B, Manvar S, Dhoble A, Patel A, Savani C, Patel J, Chothani A, Savani GT, Deshmukh A, Grines CL, Curtis J, Mangi AA, Cleman M, Forrest JK. Effect of Hospital Volume on Outcomes of Transcatheter Aortic Valve Implantation. Am J Cardiol 2015;116:587-594.
13
ACCEPTED MANUSCRIPT
15: Kim LK, Minutello RM, Feldman DN, Swaminathan RV, Bergman G, Singh H, Kaple RK, Wong SC. Association Between Transcatheter Aortic Valve Implantation Volume and Outcomes in the United States. Am J Cardiol 2015;116:1910-1915. 16: HCUP Databases. Healthcare Cost and Utilization Project (HCUP). February 2018. Agency
us.ahrq.gov/nisoverview.jsp.
[Available from: www.hcup-
CR IP T
for Healthcare Research and Quality, Rockville, MD
17: 2. (NIS) OotNNIS. HCUP Databases. Healthcare Cost and Utilization Project (HCUP).
www.hcup-us.ahrq.gov/nisoverview.jsp.
AN US
February 2018. Agency for Healthcare Research and Quality, Rockville, MD. [Available from:
18: Shuhaiber J, Isaacs AJ, Sedrakyan A. The Effect of Center Volume on In-Hospital Mortality After Aortic and Mitral Valve Surgical Procedures: A Population-Based Study. Ann Thorac Surg
M
2015;100:1340-1346.
19: Massarweh NN, Kougias P, Wilson MA. Complications and Failure to Rescue After
2016;151:1157-1165.
ED
Inpatient Noncardiac Surgery in the Veterans Affairs Health System. JAMA Surg
PT
20: Telila T, Akintoye E, Ando T, Merid O, Palla M, Mallikethi-Reddy S, Briasoulis A, Grines C,
CE
Afonso L. Hospital teaching status and trascatheter aortic valve replacement outcomes in the United States: Analysis of the national inpatient sample. Catheter Cardiovasc Interv
AC
2017;90:1200-1205.
21: Kappetein AP, Head SJ, Généreux P, Piazza N, van Mieghem NM, Blackstone EH, Brott TG, Cohen DJ, Cutlip DE, van Es GA, Hahn RT, Kirtane AJ, Krucoff MW, Kodali S, Mack MJ, Mehran R, Rodés-Cabau J, Vranckx P, Webb JG, Windecker S, Serruys PW, Leon MB. Updated
14
ACCEPTED MANUSCRIPT
standardized endpoint definitions for transcatheter aortic valve implantation: the Valve Academic Research Consortium-2 consensus document. J Am Coll Cardiol 2012;60:1438-1454. 22: Carroll JD, Vemulapalli S, Dai D, Matsouaka R, Blackstone E, Edwards F, Masoudi FA, Mack M, Peterson ED, Holmes D, Rumsfeld JS, Tuzcu EM, Grover F. Procedural Experience for
CR IP T
Transcatheter Aortic Valve Replacement and Relation to Outcomes: The STS/ACC TVT Registry. J Am Coll Cardiol 2017;70:29-41.
23: Wassef AWA, Alnasser S, Rodes-Cabau J, Webb JG, Barbanti M, Liu Y, Muñoz-García AJ,
AN US
Tamburino C, Dager AE, Serra V, Amat-Santos IJ, Al Lawati H, Urena M, Alonso Briales JH, Benitez LM, Del Blanco BG, Roman AS, Bagai A, Buller CE, Peterson MD, Cheema AN. Institutional experience and outcomes of transcatheter aortic valve replacement: Results from an
AC
CE
PT
ED
M
international multicentre registry. Int J Cardiol 2017;245:222-227.
15
ACCEPTED MANUSCRIPT
Table 1. Univariate distribution of clinical and demographic data of patients stratified by hospital-level TAVI volume Variable
Hospital Volume
Intermediate
No.
of 2,084 (21.29%)
5,771 (58.94%)
of 10,407
28,811
9,668
M
High
>130
observation,
observation, weighted TAVI 1-30
volume, weighted
82 ± 8
82 ± 8
0.150
1.59%
1.62%
1.55%
0.376
60-69
6.00%
6.91%
7.64%
70-79
23.14%
23.52%
22.18%
80-89
52.80%
53.53%
52.96%
? 90
16.48%
14.41%
15.66%
PT
Age, mean (SD) 82 ± 8
31 - 130
ED
Annual
1,255(19.32%)
AN US
Unweighted No.
CR IP T
Low
Age
AC
50-59
CE
(Years)
16
ACCEPTED MANUSCRIPT
47.98%
45.71%
47.05%
0.201
White
88.13%
88.10%
85.14%
0.001
Black
3.93%
3.84%
4.78%
Hispanic
4.65%
3.65%
2.92%
Asia
0.81%
1.20%
0.74%
Others
2.47%
3.20%
6.41%
Dyslipidemia
64.37%
65.85%
67.30%
0.326
Prior myocardial 12.04%
13.77%
13.17%
0.194
Prior
19.15%
AN US
infarction
20.23%
23.08%
0.025
22.41%
21.04%
0.030
10.61%
11.64%
0.394
percutaneous
M
coronary intervention
artery
ED
coronary 19.25% bypass
graft
10.40%
CE
Prior pacemaker
PT
Prior
CR IP T
Women
44.68%
42.73%
43.31%
0.347
Chronic
25.65%
25.97%
24.17%
0.745
7.26%
5.72%
0.133
AC
Atrial fibrillation
obstructive pulmonary disease Carotid
artery 6.90%
17
ACCEPTED MANUSCRIPT
disease Cerebrovascular
12.98%
13.70%
14.44%
0.437
Hypertension
80.58%
80.50%
82.93%
0.163
Peripheral
26.63%
29.33%
24.30%
0.008
Diabetes
35.23%
36.42%
34.47%
0.307
Obese
14.76%
15.82%
13.72%
0.115
Anemia
26.02%
25.02%
23.72%
0.523
8.09%
9.83%
0.378
AN US
vascular diseases
Deficiency Congestive heart 8.79%
CR IP T
disease
34.72%
36.76%
36.03%
0.373
Liver disease
2.11%
2.82%
2.82%
0.251
Electrolyte
24.51%
24.71%
20.50%
0.086
28.83%
31.21%
0.344
6.29%
6.52%
5.82%
0.610
2.55%
2.81%
3.47%
0.215
Oxygen
28.89%
CE
Smoking
PT
derangement
ED
Renal failure
M
failure
AC
dependent
Maintenance dialysis
Elixhauser score 0
0.865 1.59%
1.64%
1.45%
18
ACCEPTED MANUSCRIPT
1-3
46.77%
47.52%
48.94%
?4
51.64%
50.84%
49.61%
Hospital bed size
0.002 6.44%
4.53%
1.66%
Medium
25.55%
17.47%
8.38%
Large
67.02%
78.00%
89.97%
Expected
CR IP T
Small
0.126
primary payer 89.86%
90.86%
93.10%
Medicaid
0.96%
0.88%
0.93%
Private
7.25%
6.52%
4.77%
Others
1.92%
1.74%
1.19%
M
Median
income
in
PT
quartile
AC
3nd
24.53%
20.44%
17.97%
26.47%
25.53%
21.04%
26.59%
26.34%
22.49%
22.41%
27.67%
38.50%
CE
1st
4th
<.0001
ED
Household
2nd
AN US
Medicare
Hospital Region
0.351
Northeast
21.27%
24.31%
34.21%
Midwest
23.86%
23.01%
17.56%
19
ACCEPTED MANUSCRIPT
37.99%
34.30%
28.37%
West
16.87%
18.37%
19.85%
AC
CE
PT
ED
M
AN US
CR IP T
South
20
ACCEPTED MANUSCRIPT
Table 2. Perioperative complications in patients undergoing Transcatheter Aortic Valve
Hospital volume Low 3.86%
Intermediate 3.38%
High 2.58%
1.15%
1.35%
0.93%
0.304
11.39%
8.72%
9.24%
0.105
2.98% 0.14%
2.83% 0.07%
2.17% 0.26%
0.185 0.128
0.58%
0.61%
1.03%
0.120
2.31%
1.18%
0.033
7.40% 0.47%
6.41% 0.16%
0.274 0.148
2.17% 0.14%
2.38% 0.05%
2.37% 0.05%
0.866 0.374
24.75% 27.04%
21.69% 23.70%
20.92% 23.13%
0.058 0.063
4.28% 2.31 (1.53, 3.52)
3.21% 1.87 (1.24, 2.81)
2.37% 1.57 (1.02, 2.42)
0.006 0.0002
CE
AC
P value 0.093
CR IP T
AN US
1.89%
ED
7.62% 0.43%
PT
Variable Acute respiratory failure with reintubation Acute kidney injury requiring dialysis Major bleeding requiring transfusion Cardiogenic shock Cardiac tamponade with pericardiocentesis Acute myocardial infarction with percutaneous coronary intervention Mechanical circulatory support Fatal arrhythmia Conversion to surgical aortic valve replacement Stroke Vascular complication requiring surgery Total complication Unadjusted risk Risk-adjusted Mortality Unadjusted risk Risk-adjusted
M
Implantation stratified by hospital-level TAVI volume
21
ACCEPTED MANUSCRIPT
12.88% (3.39, 8.20 19.52)
8.87% (3.45, 6.12 (2.37 15.81)
0.084 0.291
AC
CE
PT
ED
M
AN US
CR IP T
Failure to Rescue (Weighted, n *10, 849+) Unweight (2,175) Unadjusted risk 13.59% Risk-adjusted 8.24 20.03)
22
Failure to Rescue, Hospital Volume, and In-Hospital Mortality After Transcatheter Aortic Valve Implantation Tomo Ando MD , Oluwole Adegbala MD, MPH , Pedro A Villablanca MD, MS , Mohamad Shokr MD , Emmanuel Akintoye MD, MPH , Alexandros Briasoulis MD, PhD , Hisato Takagi MD. PhD , Theodore Schreiber MD , Cindy L. Grines MD , Luis Afonso MD PII: DOI: Reference:
S0002-9149(18)31192-5 10.1016/j.amjcard.2018.05.020 AJC 23324
To appear in:
The American Journal of Cardiology
Received date: Revised date: Accepted date:
29 March 2018 8 May 2018 8 May 2018
Please cite this article as: Tomo Ando MD , Oluwole Adegbala MD, MPH , Pedro A Villablanca MD, MS , Mohamad Shokr MD , Emmanuel Akintoye MD, MPH , Alexandros Briasoulis MD, PhD , Hisato Takagi MD. PhD , Theodore Schreiber MD , Cindy L. Grines MD , Luis Afonso MD , Failure to Rescue, Hospital Volume, and In-Hospital Mortality After Transcatheter Aortic Valve Implantation, The American Journal of Cardiology (2018), doi: 10.1016/j.amjcard.2018.05.020
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Failure to Rescue, Hospital Volume, and In-Hospital Mortality After Transcatheter Aortic Valve Implantation 1
Tomo Ando MD, 2Oluwole Adegbala MD, MPH, 3Pedro A.Villablanca MD, MSc, Mohamad
Shokr MD 1Emmanuel Akintoye MD, MPH, 4Alexandros Briasoulis MD, PhD, 5Hisato Takagi
CR IP T
MD. PhD, 1Theodore Schreiber MD, 6Cindy L. Grines MD, 1Luis Afonso MD 1
M
AN US
Department of Medicine Division of Cardiology, Wayne State University/Detroit Medical Center, Detroit, Michigan, United States 2 Department of Medicine Department of Internal Medicine, Englewood Hospital and Medical Center, Seton Hall University-Hackensack Meridian School of Medicine, Englewood, New Jersey, United States 3 Department of Medicine Division of Cardiology, New York University Langone Medical Center, New York, New York, United States 4 Department of Medicine Divison of Cardiovascular Medicine, University of Iowa Hospitals and Clinics, Iowa, Iowa, United States 5 Department of Medicine Division of Cardiovascular Surgery, Shizuoka Medical Center, Shizuoka, Japan 6 Department of Medicine Division of Cardiology, North Shore University Hospital, Hofstra Northwell School of Medicine, Manhasset, New York, United States
ED
Running title: Failure to rescue after TAVI Disclosures: Authors have no disclosures
Tomo Ando MD
PT
Corresponding author:
CE
3990, John R, Detroit, Michigan, 48201, United States Detroit Medical Center, Division of Cardiology
AC
Phone: 313-745-2620 Fax: 313-745-8643
Email: [email protected]
1
ACCEPTED MANUSCRIPT
Abstract Failure to rescue (FTR), death after major complications, has been well described in the surgical literature as a source of different outcomes among different hospitals. However, FTR has not been investigated in transcatheter aortic valve implantation (TAVI). Our aim was to assess the
CR IP T
difference of in-patient mortality and FTR among different TAVI volume hospitals. We queried the Nationwide Inpatient Sample database from 2011 to 2015 to identify patients who had transarterial TAVI. FTR was calculated as those who had in-patient mortality with at least one with major perioperative complications. Hospitals were divided into three groups according to annual
AN US
TAVI volume, the lowest quintile (≤30/year), second to fourth quintile (31-130/year), and highest quintile (≥130/year). Multivariate analysis was used to calculate risk adjusted in-patient mortality rate and FTR and was compared between these different volume hospitals. A total of 48,886
M
TAVI procedures were identified (10,407, 28,811, and 9,668 in low, intermediate, and high volume centers, respectively). Mean age, percentage of female, and Elixhauser comorbidity
ED
index was similar across different TAVI volume hospital. Incidence of major perioperative complications did not differ among different volume hospitals. Adjusted rate of in-patient
PT
mortality (2.3%, 1.87%, and 1.57% for low, intermediate, and high volume center, respectively,
CE
p<0.001) were significantly lower with higher hospital volume but FTR (8.24%, 8.20%, and 6.12% for low, intermediate, and high volume center, respectively, p=0.29) were the same among
AC
the three groups. Our results suggest that FTR does not explain the variation of in-hospital mortality among different hospital volumes.
Key words: aortic stenosis, failure to rescue, transcatheter aortic valve implantation Failure to rescue (FTR), characterized as in-patient mortality with at least one with major perioperative complication, has been well described in the surgical literature as a source of
2
ACCEPTED MANUSCRIPT
variation in hospital outcomes (1-5). Transcatheter aortic valve implantation (TAVI) is associated with certain peri-procedural complications that negatively impact the outcomes and therefore assessing the rate of FTR would have incremental value in addition to evaluating periprocedural complication and mortality rate post-TAVI. With increase in-hospital procedural
CR IP T
volume, better outcomes have been reported in the past for various surgical and percutaneous procedures (6-13). Previous studies have reported similar trends, with decreasing adverse
outcomes with increased hospital volume following TAVI (14,15). However, these analyses reflected early United States commercial TAVI experience. With its rapid expansion and center
AN US
experience, TAVI has become a safer procedure and the outcomes of volume-outcome
relationship may have dramatically changed. Our aim was to assess the FTR among different TAVI volume hospital including the most recently available Nationwide Inpatient Sample (NIS)
M
database.
ED
Methods
PT
Data were obtained from the Agency for Healthcare Research and Quality Healthcare
CE
Cost and Utilization Project–NIS files between 2011 and 2015. The data was queried to identify patient demographics and TAVI procedure recipients in the United States using the International
AC
Classification of Diseases-Ninth Revision-Clinical Modification (ICD-9-CM). The NIS was created by the Healthcare Cost and Utilization Project as a discharge database and is maintained by the Agency for Healthcare Research and Quality. NIS data are extracted from a random sample of approximately 20% of all nonfederal, general, and specialty specific hospital inpatient admissions. Approximately 97% of hospitals in the United States are represented by the NIS,
3
ACCEPTED MANUSCRIPT
making it the largest all-payer inpatient discharge database in the United States. Criteria used for stratified sampling of hospitals into the NIS include hospital ownership, patient volume, teaching status, urban or rural location, and geographic region (16). Weighting the patient-level observations in the NIS data sets to account for the complex sampling scheme provides estimates
CR IP T
for the entire US population of hospitalized patients. The data set is publicly available and deidentified. Accordingly, it is a designated exempt from an Institutional Review Board approval.
Patients aged 50 years and older with aortic stenosis who underwent TAVI during the
AN US
study period were identified using International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes (supplemental table 1). We excluded patients with a diagnosis of aortic insufficiency without a diagnosis of aortic stenosis. We did not include the last 3
M
months of 2015 given the non-availability of Elixhauser comorbidity indicators in those months (17). Annual hospital volume was determined by obtaining the total number of TAVI cases
ED
performed by a given institution per annum using the unique Healthcare Cost and Utilization Project (HCUP) hospital number. The total institution was divided into quintiles before analysis.
PT
Low volume was defined by the lowest quintile, intermediate volume was defined by the second
CE
to fourth quintiles, and high volume was defined by the highest quintile. This formula resulted in the following breakdown of volume cutoff points. Low-volume institutions performed ≤ 30
AC
TAVIs/year, intermediate-volume institutions performed 31 – 130 TAVIs/year, and high-volume institutions performed ≥ 130 TAVIs/year.
Data on patient- and hospital-level characteristics were provided for each patient in the NIS. Our primary outcome was FTR defined as in-hospital mortality following major peri-
4
ACCEPTED MANUSCRIPT
operative complications. FTR was calculated by the numbers of patients who experienced inhospital mortality following at least one major perioperative complication divided by the total number of patients who had major perioperative complications. These major peri-operative complications include in-hospital mortality, stroke, acute myocardial infarction with
CR IP T
percutaneous coronary intervention, major bleeding requiring transfusion, acute kidney injury requiring dialysis, cardiogenic shock, fatal arrhythmia (ventricular fibrillation, ventricular tachycardia, or cardiac arrest), mechanical circulatory support, TAVI converted to surgical aortic valve replacement, vascular injury requiring surgery, and acute respiratory failure with
AN US
reintubation,. These mentioned outcomes were captured from the dataset with ICD-9-CM codes (Supplementary Table 1).
M
Age was classified as 50-59, 60-69, 70-79, 80-89 and >90 years. Race was catalogued as White, Black, Hispanic, Asian and others. Risk adjustment for comorbid medical conditions
ED
was performed using the Elixhauser comorbidity index, categorized by number of medical comorbidities (0, 1 -3, or ≥4). Individual household income was classified as low, medium, high,
PT
or highest in the NIS data set. Insurance classification was described as Medicaid, Medicare,
CE
private or others. Hospitals were classified by location (rural, urban), region (Northeast, Midwest, West, South), size (small, medium, large), and teaching status (rural, urban non-
AC
teaching, urban-teaching).
Patients were catalogued by institutional volume; low-, medium-, and high-volume
TAVI centers and were compared. Descriptive statistics were presented and compared across hospital volume levels, with chi-square tests for categorical variables, analysis of variance
5
ACCEPTED MANUSCRIPT
(ANOVA) for normally distributed continuous variables, and Kruskal-Wallis test for continuous variables with skewed distribution (Table 1). Mortality rates, complication rates, and rates of FTR were also reported descriptively across hospital volume levels. The adjusted mortality rate and adjusted failure to rescue rate were calculated by fitting a Poisson regression model with a
CR IP T
robust error variance under generalized estimating equations approach. The robust error variance was estimated by using repeated statement and individual subject identifier while adjusting for patient-level covariates, including age, race, gender, median household income, type of insurance, Elixhauser co-morbidity index, as well as hospital-level covariates, including hospital
AN US
location or teaching status, and year of data publication. Statistical significance of the mortality rates and FTR rate differences between the hospital volume categories was determined by the contrast of the regression coefficients from the Poisson regression model. Subsequently, pairwise
M
comparison was conducted with Bonferroni correction for multiple comparisons. All the data extraction and analyses was done with Statistical Analysis System (SAS V.9.4, SAS
ED
Institute Inc, Cary, NC, US). We chose a p-value of <0.05, reported the effect sizes, 95%
AC
Results
CE
comparisons.
PT
confidence intervals (CI), and p-values or the Bonferroni corrected p-values for multiple
There was a total of 48,886 TAVI patients identified (10,407, 28,811, and 9,668 in in
lowest, intermediate, and highest volume centers, respectively). The lowest and intermediate volume hospitals performed 1-30 and 31-130 TAVIs per year respectively, whereas the highest volume hospitals performed more than 130 TAVI per year. Mean age, percentage of female, and
6
ACCEPTED MANUSCRIPT
Elixhauser comorbidity index did not differ among the three different hospital groups. Prior percutaneous coronary intervention, coronary artery bypass graft, peripheral vascular disease, and race differed significantly across the three hospital groups. Large bed size hospitals, and high median household income were observed more frequently in the highest volume hospitals. These
CR IP T
results are summarized in table 1.
The overall in-hospital morality was 3.27%. The unadjusted rates of in-hospital mortality declined significantly with increasing hospital volume (4.28%, 3.21%, and 2.37% for
AN US
lowest, intermediate, and highest volume center, respectively, p=0.006). After adjustments for clinical and hospital variables, the results remained statistically significant (2.3% [1.53-3.52], 1.87% [1.24-2.81], and 1.57% [1.02-2.42] for lowest, intermediate, and highest volume center,
M
respectively, p=0.0002, supplement 2). Yearly trend of un-adjusted mortality in low, intermediate, and high volume hospitals are summarized in supplement 3. In general, the rate of
ED
major complications decreased with increasing hospital volume (except for acute myocardial infarction with percutaneous coronary intervention). Use of mechanical circulatory support was
PT
higher in low volume hospital. However, other complications were similar across different
CE
hospital volumes. The overall FTR was 12.28%. The unadjusted rate of FTR was almost similar between lowest and intermediate volume hospital (13.59% and 12.88%, respectively) and
AC
numerically lower in highest volume hospital (8.87%) but was statistically not significant (p=0.084). After multivariate adjustments, FTR rates were similar among all hospital volume groups (8.24% [3.39-20.03], 8.20% [3.45-19.52], and 6.12% [2.37-15.81] for lowest, intermediate, and highest volume center, respectively, p=0.29). These results are summarized in table 2.
7
ACCEPTED MANUSCRIPT
Discussion
The major findings of our results were 1: Rate of FTR also did not differ among
CR IP T
different volume hospitals. 2: In-patient mortality was lower in higher volume hospitals but major perioperative complications did not significantly decrease with increased hospital volume. Our results suggest that hospital-volume is an independent predictor of in-hospital mortality and
AN US
FTR is not the source of variation of in-hospital mortality among different hospital volume.
FTR and mortality did not differ among different volume hospitals following TAVI. Previous studies have shown that high procedural volume hospital was associated with higher
M
rates of FTR after cardiovascular surgery (9,18). Similar result was also reported in non-cardiac surgery (19). There could be several reasons why FTR was similar among different TAVI volume
ED
hospitals as opposed to previous studies. First, incident rate of the complications with extremely high mortality, requiring advanced medical resources and highly skilled operators to rescue, such
PT
as cardiac tamponade requiring pericardiocentesis, acute myocardial infarction requiring
CE
percutaneous coronary intervention, and conversion to surgical aortic valve replacement, were very low and may partly explain the similar FTR observed in our study. Secondary, even if the
AC
annual TAVI volume is low, TAVI is predominantly performed at teaching hospitals or large sized bed hospitals where access to sub-specialities and intensive care units are often readily available (20). At these hospitals, the more commonly observed perioperative complications may not be difficult to manage, resulting in similar FTR across different TAVI volume hospitals. Third, the standardization of TAVI endpoints by the Valve Academic Research Consortium may
8
ACCEPTED MANUSCRIPT
have helped physicians recognize the expected complications post-TAVI and expeditiously manage them across all TAVI volume hospitals (21).
Regarding the hospital volume-outcome relationship, our results were with previous
CR IP T
studies that demonstrated decreased in-patient mortality with increased hospital volume after TAVI. Badheka et al. reported from NIS database that with increasing hospital volume, inhospital mortality decreased from 6.4% for lowest volume hospitals to 2.8% for the highest TAVI volume hospitals (14). Similar results were reported by Kim et al. also from the NIS
AN US
database (15). These studies, however, only included data from 2012, which is considered a very early commercial TAVI era in the United States. Our result included more updated data from last 5 years of NIS reports, reflecting more mature TAVI programs and operator experience. Increase
M
in institutional experience has been reported with improved outcomes. Carroll et al. reported that with accumulation of procedural experience up to 400th cases, in-hospital mortality, vascular
ED
complications, and bleeding decreased significantly (22). Wassef et al. reported that > 242 cases of experience resulted in decrease of half of all-cause mortality (p=0.002) (23). Furthermore, in-
PT
hospital mortality of trans-femoral TAVI did not inversely relate with higher procedural
CE
experience according to the registry that included procedure in the United States from 2011 through 2015 (22). Considering these results, it is possible that trans-arterial TAVI has matured
AC
during the study period. Our results further strengthen the previous reports that strongly suggested an association between in-hospital mortality and TAVI volume, despite maturation of trans-arterial TAVI techniques, and adds to the existing literature the fact that FTR is not the source of variation in outcomes. In addition, although previous studies have shown decreased perioperative complications as well with increasing hospital volume, our study did not show
9
ACCEPTED MANUSCRIPT
similar trend (14,15). This is likely the result of better patient selection and learning curve.
There are several limitations that should be noted. First, this was a retrospective cohort study that used administrative database. Therefore, there is a possibility of coding error that
CR IP T
could have biased our results. However, the use of NIS database has been applied to wide variety of medical researches and is considered a valid resource. Second, some of the important clinical variables, such as the Society of Thoracic Surgeons score and frailty score were not available and hence not adjusted in multivariable analysis. Third, we were also unable to adjust for key
AN US
echocardiographic parameters such as ejection fraction and pulmonary pressure. Lastly, because the NIS database is formed based on hospitals in the United States, the volume cutoff for
AC
CE
PT
ED
M
different quintile hospitals may not apply to country outside the United States.
Acknowledgement: None
10
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
M
AN US
CR IP T
Disclosures: All authors have no disclosures
References
11
ACCEPTED MANUSCRIPT
1: Mehta A, Efron D, Stevens K, Manukyan MC, Joseph B, Sakran JV. Hospital Variation in Mortality after Emergency Bowel Resections: The Role of Failure-to-Rescue. J Trauma Acute Care Surg 2018;84:702-710. 2: Gani F, Johnston FM, Nelson-Williams H, Cerullo M, Dillhoff ME, Schmidt CR, Pawlik TM.
CR IP T
Hospital Volume and the Costs Associated with Surgery for Pancreatic Cancer. J Gastrointest Surg 2017;21:1411-1419.
3: Malone H, Cloney M, Yang J, Hershman DL, Wright JD, Neugut AI, Bruce JN. Failure to Rescue and Mortality Following Resection of Intracranial Neoplasms. Neurosurgery 2017 Aug 1
AN US
4: Amini N, Spolverato G, Kim Y, Pawlik TM. Trends in Hospital Volume and Failure to Rescue for Pancreatic Surgery. J Gastrointest Surg 2015;19:1581-1592.
5: Wright JD, Herzog TJ, Siddiq Z, Arend R, Neugut AI, Burke WM, Lewin SN, Ananth CV,
J Clin Oncol 2012;30:3976-3982.
M
Hershman DL. Failure to rescue as a source of variation in hospital mortality for ovarian cancer.
ED
6: Luft HS, Bunker JP, Enthoven AC. Should operations be regionalized? The empirical relation between surgical volume and mortality. N Engl J Med 1979;301:1364-1369.
PT
7: Birkmeyer JD, Siewers AE, Finlayson EV, Stukel TA, Lucas FL, Batista I, Welch HG,
CE
Wennberg DE. Hospital volume and surgical mortality in the United States. N Engl J Med 2002;346:1128-1137.
AC
8: Holt PJ, Poloniecki JD, Gerrard D, Loftus IM, Thompson MM. Meta-analysis and systematic review of the relationship between volume and outcome in abdominal aortic aneurysm surgery. Br J Surg 2007;94:395-403. 9: Gonzalez AA, Dimick JB, Birkmeyer JD, Ghaferi AA. Understanding the volume-outcome effect in cardiovascular surgery: the role of failure to rescue. JAMA Surg 2014;149:119-123.
12
ACCEPTED MANUSCRIPT
10: Singh V, Badheka AO, Patel NJ, Chothani A, Mehta K, Arora S, Patel N, Deshmukh A, Shah N, Savani GT, Rathod A, Manvar S, Thakkar B, Panchal V, Patel J, Palacios IF, Rihal CS, Cohen MG, O'Neill W, De Marchena E. Influence of hospital volume on outcomes of
Catheter Cardiovasc Interv 2015;85:1073-1081.
CR IP T
percutaneous atrial septal defect and patent foramen ovale closure: a 10-years US perspective.
11: Badheka AO, Chothani A, Mehta K, Patel NJ, Deshmukh A, Hoosien M, Shah N, Singh V, Grover P, Savani GT, Panaich SS, Rathod A, Patel N, Arora S, Bhalara V, Coffey JO, O'Neill W, Makkar R, Grines CL, Schreiber T, Di Biase L, Natale A, Viles-Gonzalez JF. Utilization and
AN US
adverse outcomes of percutaneous left atrial appendage closure for stroke prevention in atrial fibrillation in the United States: influence of hospital volume. Circ Arrhythm Electrophysiol 2015;8:42-48.
M
12: Kim LK, Swaminathan RV, Looser P, Minutello RM, Wong SC, Bergman G, Naidu SS, Gade CL, Charitakis K, Singh HS, Feldman DN. Hospital Volume Outcomes After Septal
ED
Myectomy and Alcohol Septal Ablation for Treatment of Obstructive Hypertrophic Cardiomyopathy: US Nationwide Inpatient Database, 2003-2011. JAMA Cardiol 2016;1:324-332.
PT
13: Opotowsky AR, Landzberg MJ, Kimmel SE, Webb GD. Percutaneous closure of patent
CE
foramen ovale and atrial septal defect in adults: the impact of clinical variables and hospital procedure volume on in-hospital adverse events. Am Heart J 2009;157:867-874.
AC
14: Badheka AO, Patel NJ, Panaich SS, Patel SV, Jhamnani S, Singh V, Pant S, Patel N, Patel N, Arora S, Thakkar B, Manvar S, Dhoble A, Patel A, Savani C, Patel J, Chothani A, Savani GT, Deshmukh A, Grines CL, Curtis J, Mangi AA, Cleman M, Forrest JK. Effect of Hospital Volume on Outcomes of Transcatheter Aortic Valve Implantation. Am J Cardiol 2015;116:587-594.
13
ACCEPTED MANUSCRIPT
15: Kim LK, Minutello RM, Feldman DN, Swaminathan RV, Bergman G, Singh H, Kaple RK, Wong SC. Association Between Transcatheter Aortic Valve Implantation Volume and Outcomes in the United States. Am J Cardiol 2015;116:1910-1915. 16: HCUP Databases. Healthcare Cost and Utilization Project (HCUP). February 2018. Agency
us.ahrq.gov/nisoverview.jsp.
[Available from: www.hcup-
CR IP T
for Healthcare Research and Quality, Rockville, MD
17: 2. (NIS) OotNNIS. HCUP Databases. Healthcare Cost and Utilization Project (HCUP).
www.hcup-us.ahrq.gov/nisoverview.jsp.
AN US
February 2018. Agency for Healthcare Research and Quality, Rockville, MD. [Available from:
18: Shuhaiber J, Isaacs AJ, Sedrakyan A. The Effect of Center Volume on In-Hospital Mortality After Aortic and Mitral Valve Surgical Procedures: A Population-Based Study. Ann Thorac Surg
M
2015;100:1340-1346.
19: Massarweh NN, Kougias P, Wilson MA. Complications and Failure to Rescue After
2016;151:1157-1165.
ED
Inpatient Noncardiac Surgery in the Veterans Affairs Health System. JAMA Surg
PT
20: Telila T, Akintoye E, Ando T, Merid O, Palla M, Mallikethi-Reddy S, Briasoulis A, Grines C,
CE
Afonso L. Hospital teaching status and trascatheter aortic valve replacement outcomes in the United States: Analysis of the national inpatient sample. Catheter Cardiovasc Interv
AC
2017;90:1200-1205.
21: Kappetein AP, Head SJ, Généreux P, Piazza N, van Mieghem NM, Blackstone EH, Brott TG, Cohen DJ, Cutlip DE, van Es GA, Hahn RT, Kirtane AJ, Krucoff MW, Kodali S, Mack MJ, Mehran R, Rodés-Cabau J, Vranckx P, Webb JG, Windecker S, Serruys PW, Leon MB. Updated
14
ACCEPTED MANUSCRIPT
standardized endpoint definitions for transcatheter aortic valve implantation: the Valve Academic Research Consortium-2 consensus document. J Am Coll Cardiol 2012;60:1438-1454. 22: Carroll JD, Vemulapalli S, Dai D, Matsouaka R, Blackstone E, Edwards F, Masoudi FA, Mack M, Peterson ED, Holmes D, Rumsfeld JS, Tuzcu EM, Grover F. Procedural Experience for
CR IP T
Transcatheter Aortic Valve Replacement and Relation to Outcomes: The STS/ACC TVT Registry. J Am Coll Cardiol 2017;70:29-41.
23: Wassef AWA, Alnasser S, Rodes-Cabau J, Webb JG, Barbanti M, Liu Y, Muñoz-García AJ,
AN US
Tamburino C, Dager AE, Serra V, Amat-Santos IJ, Al Lawati H, Urena M, Alonso Briales JH, Benitez LM, Del Blanco BG, Roman AS, Bagai A, Buller CE, Peterson MD, Cheema AN. Institutional experience and outcomes of transcatheter aortic valve replacement: Results from an
AC
CE
PT
ED
M
international multicentre registry. Int J Cardiol 2017;245:222-227.
15
ACCEPTED MANUSCRIPT
Table 1. Univariate distribution of clinical and demographic data of patients stratified by hospital-level TAVI volume Variable
Hospital Volume
Intermediate
No.
of 2,084 (21.29%)
5,771 (58.94%)
of 10,407
28,811
9,668
M
High
>130
observation,
observation, weighted TAVI 1-30
volume, weighted
82 ± 8
82 ± 8
0.150
1.59%
1.62%
1.55%
0.376
60-69
6.00%
6.91%
7.64%
70-79
23.14%
23.52%
22.18%
80-89
52.80%
53.53%
52.96%
? 90
16.48%
14.41%
15.66%
PT
Age, mean (SD) 82 ± 8
31 - 130
ED
Annual
1,255(19.32%)
AN US
Unweighted No.
CR IP T
Low
Age
AC
50-59
CE
(Years)
16
ACCEPTED MANUSCRIPT
47.98%
45.71%
47.05%
0.201
White
88.13%
88.10%
85.14%
0.001
Black
3.93%
3.84%
4.78%
Hispanic
4.65%
3.65%
2.92%
Asia
0.81%
1.20%
0.74%
Others
2.47%
3.20%
6.41%
Dyslipidemia
64.37%
65.85%
67.30%
0.326
Prior myocardial 12.04%
13.77%
13.17%
0.194
Prior
19.15%
AN US
infarction
20.23%
23.08%
0.025
22.41%
21.04%
0.030
10.61%
11.64%
0.394
percutaneous
M
coronary intervention
artery
ED
coronary 19.25% bypass
graft
10.40%
CE
Prior pacemaker
PT
Prior
CR IP T
Women
44.68%
42.73%
43.31%
0.347
Chronic
25.65%
25.97%
24.17%
0.745
7.26%
5.72%
0.133
AC
Atrial fibrillation
obstructive pulmonary disease Carotid
artery 6.90%
17
ACCEPTED MANUSCRIPT
disease Cerebrovascular
12.98%
13.70%
14.44%
0.437
Hypertension
80.58%
80.50%
82.93%
0.163
Peripheral
26.63%
29.33%
24.30%
0.008
Diabetes
35.23%
36.42%
34.47%
0.307
Obese
14.76%
15.82%
13.72%
0.115
Anemia
26.02%
25.02%
23.72%
0.523
8.09%
9.83%
0.378
AN US
vascular diseases
Deficiency Congestive heart 8.79%
CR IP T
disease
34.72%
36.76%
36.03%
0.373
Liver disease
2.11%
2.82%
2.82%
0.251
Electrolyte
24.51%
24.71%
20.50%
0.086
28.83%
31.21%
0.344
6.29%
6.52%
5.82%
0.610
2.55%
2.81%
3.47%
0.215
Oxygen
28.89%
CE
Smoking
PT
derangement
ED
Renal failure
M
failure
AC
dependent
Maintenance dialysis
Elixhauser score 0
0.865 1.59%
1.64%
1.45%
18
ACCEPTED MANUSCRIPT
1-3
46.77%
47.52%
48.94%
?4
51.64%
50.84%
49.61%
Hospital bed size
0.002 6.44%
4.53%
1.66%
Medium
25.55%
17.47%
8.38%
Large
67.02%
78.00%
89.97%
Expected
CR IP T
Small
0.126
primary payer 89.86%
90.86%
93.10%
Medicaid
0.96%
0.88%
0.93%
Private
7.25%
6.52%
4.77%
Others
1.92%
1.74%
1.19%
M
Median
income
in
PT
quartile
AC
3nd
24.53%
20.44%
17.97%
26.47%
25.53%
21.04%
26.59%
26.34%
22.49%
22.41%
27.67%
38.50%
CE
1st
4th
<.0001
ED
Household
2nd
AN US
Medicare
Hospital Region
0.351
Northeast
21.27%
24.31%
34.21%
Midwest
23.86%
23.01%
17.56%
19
ACCEPTED MANUSCRIPT
37.99%
34.30%
28.37%
West
16.87%
18.37%
19.85%
AC
CE
PT
ED
M
AN US
CR IP T
South
20
ACCEPTED MANUSCRIPT
Table 2. Perioperative complications in patients undergoing Transcatheter Aortic Valve
Hospital volume Low 3.86%
Intermediate 3.38%
High 2.58%
1.15%
1.35%
0.93%
0.304
11.39%
8.72%
9.24%
0.105
2.98% 0.14%
2.83% 0.07%
2.17% 0.26%
0.185 0.128
0.58%
0.61%
1.03%
0.120
2.31%
1.18%
0.033
7.40% 0.47%
6.41% 0.16%
0.274 0.148
2.17% 0.14%
2.38% 0.05%
2.37% 0.05%
0.866 0.374
24.75% 27.04%
21.69% 23.70%
20.92% 23.13%
0.058 0.063
4.28% 2.31 (1.53, 3.52)
3.21% 1.87 (1.24, 2.81)
2.37% 1.57 (1.02, 2.42)
0.006 0.0002
CE
AC
P value 0.093
CR IP T
AN US
1.89%
ED
7.62% 0.43%
PT
Variable Acute respiratory failure with reintubation Acute kidney injury requiring dialysis Major bleeding requiring transfusion Cardiogenic shock Cardiac tamponade with pericardiocentesis Acute myocardial infarction with percutaneous coronary intervention Mechanical circulatory support Fatal arrhythmia Conversion to surgical aortic valve replacement Stroke Vascular complication requiring surgery Total complication Unadjusted risk Risk-adjusted Mortality Unadjusted risk Risk-adjusted
M
Implantation stratified by hospital-level TAVI volume
21
ACCEPTED MANUSCRIPT
12.88% (3.39, 8.20 19.52)
8.87% (3.45, 6.12 (2.37 15.81)
0.084 0.291
AC
CE
PT
ED
M
AN US
CR IP T
Failure to Rescue (Weighted, n *10, 849+) Unweight (2,175) Unadjusted risk 13.59% Risk-adjusted 8.24 20.03)
22