J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 5 , 2 0 1 9

ª 2 0 1 9 B Y T H E A M E R I C A N C O L L E G E O F C A R D I O L O G Y F O U N D A T I O N

P U B L I S H E D B Y E L S E V I E R

Length of Stay After TransfemoralTranscatheter Aortic ValveReplacement

An Analysis of the Society of Thoracic Surgeons/American College of CardiologyTranscatheter Valve Therapy Registry

Siddharth A. Wayangankar, MD,a Islam Y. Elgendy, MD,a Qun Xiang, MS,b Hani Jneid, MD,c

Sreekanth Vemulapalli, MD,b Tigran Khachatryan, MD,d Don Pham, MD,e Anthony A. Hilliard, MD,d

Samir R. Kapadia, MDf

JACC: CARDIOVASCULAR INTERVENTIONS CME/MOC/ECME

This article has been selected as this issue’s CME/MOC/ECME activity,

available online at http://www.acc.org/jacc-journals-cme by selecting the

JACC Journals CME/MOC/ECME tab.

Accreditation and Designation Statement

The American College of Cardiology Foundation (ACCF) is accredited by

the Accreditation Council for Continuing Medical Education to provide

continuing medical education for physicians.

The ACCF designates this Journal-based CME activity for a maximum

of 1 AMA PRA Category 1 Credit(s)TM. Physicians should claim only

the credit commensurate with the extent of their participation in the

activity.

Successful completion of this CME activity, which includes participation

in the evaluation component, enables the participant to earn up to

1 Medical Knowledge MOC point in the American Board of Internal

Medicine’s (ABIM) Maintenance of Certification (MOC) program.

Participants will earn MOC points equivalent to the amount of CME

credits claimed for the activity. It is the CME activity provider’s

responsibility to submit participant completion information to ACCME

for the purpose of granting ABIM MOC credit.

Length of Stay After Transfemoral Transcatheter Aortic Valve Replacement:

An Analysis of the Society of Thoracic Surgeons/American College of

Cardiology Transcatheter Valve Therapy Registry will be accredited by the

European Board for Accreditation in Cardiology (EBAC) for 1 hour of

External CME credits. Each participant should claim only those hours of

credit that have actually been spent in the educational activity. The

Accreditation Council for Continuing Medical Education (ACCME) and

the European Board for Accreditation in Cardiology (EBAC) have recog-

nized each other’s accreditation systems as substantially equivalent.

Apply for credit through the post-course evaluation. While offering the

ISSN 1936-8798/$36.00

From the aDivision of Cardiovascular Medicine, Department of Medicine, U

diovascular Research Institute, Durham, North Carolina; cDivision of Car

Houston, Texas; dDivision of Cardiovascular Medicine, Loma Linda Unive

Hospital, Houston, Texas; and the fDepartment of Cardiology, Cleveland Clin

Medtronic Valve therapies. All other authors have reported that they have n

to disclose.

Manuscript received October 4, 2018; revised manuscript received Novembe

credits noted above, this program is not intended to provide extensive

training or certification in the field.

Method of Participation and Receipt of CME/MOC/ECME Certificate

To obtain credit for this CME/MOC/ECME, you must:

1. Be an ACC member or JACC: Cardiovascular Interventions subscriber.

2. Carefully read the CME/MOC/ECME-designated article available on-

line and in this issue of the journal.

3. Answer the post-test questions. A passing score of at least 70%must be

achieved to obtain credit.

4. Complete a brief evaluation.

5. Claim your CME/MOC/ECME credit and receive your certificate

electronically by following the instructions given at the conclusion of

the activity.

CME/MOC/ECME Objective for This Article: Upon completion of this

activity, the learner should be able to: 1) understand the incidence of longer

length of stay after TAVR; 2) identify risk factors for longer length of stay

after TAVR; and 3) understand the relationship between length of stay and

outcomes after TAVR.

CME/MOC/ECME Editor Disclosure: JACC: Cardiovascular Interventions

CME/MOC/ECME Editor Michael C. McDaniel, MD, has reported that he is

a Penumbra-Investigator on the EXTACT-PE trial.

Author Disclosures: Dr. Wayangankar is a proctor for Medtronic Valve

therapies. All other authors have reported that they have no relation-

ships relevant to the contents of this paper to disclose.

Medium of Participation: Print (article only); online (article and quiz).

CME/MOC/ECME Term of Approval

Issue Date: March 11, 2019

Expiration Date: March 10, 2020

https://doi.org/10.1016/j.jcin.2018.11.015

niversity of Florida, Gainesville, Florida; bDuke Car-

diovascular Medicine, Baylor College of Medicine,

rsity, Loma Linda, California; eMemorial Hermann

ic, Cleveland, Ohio. Dr. Wayangankar is a proctor for

o relationships relevant to the contents of this paper

r 13, 2018, accepted November 14, 2018.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 5 , 2 0 1 9 Wayangankar et al.M A R C H 1 1 , 2 0 1 9 : 4 2 2 – 3 0 LOS After TAVR

423

Length of Stay After Transf

emoralTranscatheter Aortic Valve Replacement

An Analysis of the Society of Thoracic Surgeons/American College of CardiologyTranscatheter Valve Therapy Registry

Siddharth A. Wayangankar, MD,a Islam Y. Elgendy, MD,a Qun Xiang, MS,b Hani Jneid, MD,c

Sreekanth Vemulapalli, MD,b Tigran Khachatryan, MD,d Don Pham, MD,e Anthony A. Hilliard, MD,d

Samir R. Kapadia, MDf

ABSTRACT

OBJECTIVES The goal of this study was to investigate the trends, predictors, and outcomes of delayed discharge

(>72 h) after transcatheter aortic valve replacement.

BACKGROUND Length of stay post–transcatheter aortic valve replacement may have significant clinical and admin-

istrative implications.

METHODS Data from the Transcatheter Valve Therapy Registry were used to identify patients undergoing nonaborted

transfemoral transcatheter aortic valve replacement who survived to discharge, and data linked from the Centers for

Medicare & Medicaid Services were used to provide 1-year events. Patients were categorized to early discharge (#72 h)

versus delayed discharge (>72 h). The trends, predictors, and adjusted 1-year outcomes were compared in both groups.

RESULTS From 2011 to 2015, a total of 13,389 patients (55.1%) were discharged within 72 h, whereas 10,896 patients

(44.9%) were discharged beyond 72 h. There was a significant decline in rates of delayed discharge across the study

period (62% vs. 34%; p < 0.01). This remained unchanged when stratified by Transcatheter Valve Therapy risk scores.

Several factors were identified as independent predictors of early and delayed discharge. After adjustment for in-hospital

complications, delayed discharge was an independent predictor of 1-year all-cause mortality (hazard ratio: 1.45; 95%

confidence interval: 1.30 to 1.60; p < 0.01).

CONCLUSIONS Rates of delayed discharge have declined from 2011 to 2015. Delayed discharge is associated with a

significant increase in mortality even after adjusting for in-hospital complications. Further work is necessary to determine

if predictors of early discharge could be used to develop length of stay scores that might be instrumental in adminis-

trative, financial, or clinical policy development. (J Am Coll Cardiol Intv 2019;12:422–30) © 2019 by the American

College of Cardiology Foundation.

T ranscatheter aortic valve replacement(TAVR) has emerged as a suitable alternativefor surgical aortic valve replacement for in-

termediate- to high-risk patients with severe aorticstenosis (1–5). Recent enhancement in techniques,devices, and improved operator experience haslargely contributed to the improved outcomes andlower procedure-related complications with TAVR(6). These improvements have led to a more wide-spread adoption of a “minimalist” approach usingmonitored anesthesia care/local anesthesia andavoidance of post-procedure intensive care stay forselective transfemoral (TF) TAVRs (7). The biggest

push behind this “minimalist approach” has been tolimit direct procedural costs, but more importantlyto abbreviate the length of stay (LOS) thereby mini-mizing post-procedure costs; the latter being a largecontributor to total TAVR costs. Because the proced-ures are becoming less invasive, the LOS after theprocedure is expected to become shorter. Severalsingle-center studies have suggested that an earlydischarge strategy (<72 h) is safe and feasible afterTF-TAVR (8–10). A strategy of early discharge isimportant from administrative and financial view-points. Thus, the aim of this study was to investigatethe trends, predictors, and outcomes of delayed

SEE PAGE 431

TABLE 1 Baseline Characteristi

Delayed Discharge Groups

Age $85 yrs

Female

RaceWhite non-HispanicBlack non-HispanicHispanicOther

Prior infective endocarditis

Prior permanent pacemaker

Previous ICD

Prior PCI

Prior CABG

Prior aortic valve procedure

Prior mitral valve procedure

Prior stroke or TIA

Peripheral arterial disease

Current/recent smoker

Hypertension

Diabetes mellitus

Severe lung disease

Home oxygen

Immunocompromised

Prior MI

Heart failure in past 2 weeks

NYHA functional class IV

Inotropic/mechanical support

Prior atrial fibrillation/flutter

Hemodialysis

Tricuspid valve morphology

Aortic valve insufficiencyNone/mildModerate/severe

ABBR EV I A T I ON S

AND ACRONYMS

LOS = length of stay

MI = myocardial infarction

STS/ACC TVT = Society of

Thoracic Surgery/American

College of Cardiology

Transcatheter Valve Therapy

Registry

TAVR = transcatheter aortic

valve replacement

TF = transfemoral

Wayangankar et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 5 , 2 0 1 9

LOS After TAVR M A R C H 1 1 , 2 0 1 9 : 4 2 2 – 3 0

424

discharge ($72 h) after TAVR using a nation-ally representative dataset.

METHODS

DATA SOURCE. The Society of Thoracic Sur-gery/American College of Cardiology Trans-catheter Valve Therapy (STS/ACC TVT)Registry was founded in 2011. Facilities per-forming TAVR procedures are required tosubmit commercial cases to the registry forMedicare reimbursement. The data are linked

with inpatient Medicare administrative claims by theCenters for Medicare and Medicaid Services usingdirected patient identifiers. The TVT Registry con-tains clinical information including patient baseline

cs of Patients With Early Discharge and

Total(N ¼ 24,285)

EarlyDischarge

(n ¼ 13,389)

DelayedDischarge

(n ¼ 10,896) p Value

9,868 (40.6) 5,275 (39.4) 4,593 (42.1) <0.001

10,332 (42.5) 5,281 (39.4) 5,051 (46.3) <0.001

<0.00121,566 (88.8) 11,991 (89.5) 9,575 (87.9)

895 (3.7) 413 (3.1) 482 (4.4)937 (3.8) 462 (3.4) 475 (4.3)887 (3.6) 523 (3.9) 364 (3.3)

271 (1.1) 144 (1.1) 127 (1.2) 0.51

3,904 (16.1) 2,285 (17.1) 1,619 (14.8) <0.001

1,155 (4.7) 684 (5.1) 471 (4.3) 0.004

8,449 (34.8) 4,730 (35.3) 3,719 (34.1) 0.04

7,119 (29.3) 4,155 (31.0) 2,964 (27.2) <0.001

3,424 (14.1) 1,897 (14.2) 1,527 (14.0) 0.72

603 (2.5) 265 (2.0) 338 (3.1) <0.001

4,163 (17.1) 2,291 (17.1) 1,872 (17.2) 0.88

5,705 (23.5) 3,099 (23.1) 2,606 (23.9) 0.15

1,081 (4.4) 617 (4.6) 464 (4.2) 0.19

21,497 (88.5) 11,840 (88.4) 9,657 (88.6) 0.65

9,069 (37.3) 4,918 (36.7) 4,151 (38.1) 0.03

2,942 (12.1) 1,466 (10.9) 1,476 (13.5) <0.001

2,615 (10.8) 1,229 (9.2) 1,386 (12.7) <0.001

2,467 (10.1) 1,351 (10.1) 1,116 (10.2) 0.69

5,735 (23.6) 3,240 (24.2) 2,495 (22.9) 0.02

18,638 (76.7) 10,147 (75.8) 8,491 (77.9) <0.001

4,233 (17.4) 2,079 (15.5) 2,154 (19.8) <0.001

9,964 (41.0) 5,193 (38.8) 4,771 (43.8) <0.001

9,539 (39.3) 4,997 (37.3) 4,542 (41.7) <0.001

945 (3.9) 441 (3.3) 504 (4.6) <0.001

21,691 (89.3) 12,014 (89.7) 9,677 (88.8) 0.02

0.8619,148 (78.8) 10,561 (78.9) 8,587 (77.8)4,944 (21.2) 2,720 (21.1) 2,224 (21.2)

Continued on the next page

demographics, comorbidities, functional status, andprocedural details, and 1-year outcomes using stan-dardized definitions on all patients undergoing TAVRat a participating facility. This approach of acquiringthe data and ascertainment of events in the STS/ACCTVT Registry has been previously published (6,11).The Duke University School of Medicine institutionalreview board granted a waiver of informed consentfor this study.

STUDY COHORT. For the purpose of this study, weincluded all patients $18 years who underwentTF-TAVR from November 1, 2011 to September 30,2015, from the TVT Registry. Patients were excludedif: 1) they died during the hospitalization; 2) theprocedure was aborted or canceled; or 3) they weretransferred or discharged to any facility other thanhome (because some institutions might choose todischarge patients to another facility as a method toartificially lower the LOS). Based on the hospital LOS,patients were categorized into early discharge group(i.e., discharge within #72 h) or delayed dischargegroup (i.e., discharge beyond 72 h).

OUTCOMES. The 1-year individual outcomesincluded time to event occurrence of all-cause mor-tality, myocardial infarction (MI), and stroke. Thecomposite endpoint included a composite of majoradverse cardiac events, defined as the composite ofall-cause mortality, MI, stroke, or major bleeding.The data for these outcomes were obtained throughMedicare administrative claims records for rehospi-talization for these events using International Clas-sification of Diseases-9th Revision (n ¼ 12,479)-Clinical Modification codes for MI (410.x0 and410.x1) and stroke (433.x1, 434.x1, 997.02, 436, 437.1,437.9, 430, 431, and 432.x).

STATISTICAL ANALYSIS. Baseline characteristicsand in-hospital complications were presented for theearly discharge versus the delayed discharge groups.Categorical variables were reported as percentages.The baseline characteristics and in-hospital compli-cations of the early discharge group and delayeddischarge group were compared using Pearson chi-square test for categorical variables. Multivariablelogistic regression models with generalized esti-mating equation were constructed to estimate oddsratios and corresponding 95% confidence intervals forearly or delayed discharge. A robust covariance ma-trix estimator was used to account for the effect ofcluster between the hospitals. We calculated theincidence of the delayed discharges in each quarteracross the study period, with the total number of

FIGURE 1 Series Plot of DD Over Time

Rate of DD by quarter. Patients discharged to home. DD ¼ delayed discharge.

TABLE 1 Continued

Total(N ¼ 24,285)

EarlyDischarge

(n ¼ 13,389)

DelayedDischarge

(n ¼ 10,896) p Value

Mitral valve insufficiency <0.001None/mild 17,450 (71.8) 9,760 (72.3) 7,690 (70.6)Moderate/severe 6,688 (28.2) 3,541 (27.7) 3,147 (29.4)

Mitral stenosis 1,172 (4.8) 616 (4.6) 556 (5.1) 0.06

Concurrent PCI 325 (100) 169 (52.0) 156 (48.0) 0.04

Procedure location <0.001Hybrid OR suite 14,360 (59.1) 7,633 (57.0) 6,727 (61.7)Hybrid catheterization laboratory

suite7,044 (29.0) 4,031 (30.1) 3,013 (27.6)

Catheterization laboratory 2,793 (11.9) 1,684 (12.9) 1,109 (10.7)

Total site volume <0.0010–99 5,375 (22.1) 3,139 (23.4) 2,236 (20.5)100–199 7,317 (30.1) 4,098 (30.6) 3,219 (29.5)200–399 9,213 (37.9) 4,851 (36.2) 4,362 (40.0)>400 2,380 (9.9) 1,301 (9.8) 1,079 (10.0)

Valve-in-valve 948 (3.9) 597 (4.4) 351 (3.2) <0.001

Anesthesia type <0.001Moderate sedation 2,783 (11.4) 1,913 (14.3) 870 (8.0)General anesthesia 21,312 (87.8) 11,366 (84.9) 9,946 (91.2)Other 127 (0.9) 75 (0.8) 52 (0.8)

Valve sheath access method <0.001Percutaneous 15,776 (65.0) 9,346 (69.8) 6,430 (59.0)Cutdown 8,403 (34.6) 4,017 (30.0) 4,386 (40.2)Other 88 (0.4) 15 (0.2) 73 (0.8)

5-m test <0.001Unable to walk or >10 s 6,134 (25.2) 2,976 (22.2) 3,158 (29.0)#10 s 11,929 (49.1) 7,216 (53.9) 4,713 (43.2)

Values are n (%).

CABG ¼ coronary artery bypass surgery graft; ICD ¼ implantable cardioverter-defibrillator; MI ¼ myocardialinfarction; NYHA¼ New York Heart Association; OR ¼ operating room; PCI ¼ percutaneous coronary intervention;TIA ¼ transient ischemic attack.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 5 , 2 0 1 9 Wayangankar et al.M A R C H 1 1 , 2 0 1 9 : 4 2 2 – 3 0 LOS After TAVR

425

TF-TAVR procedures as the denominator. TheCochran-Armitage test was used to show the proba-bility of delayed discharge rates over time. The trendanalyses were stratified based on the STS PredictedRisk of Operative Mortality score.

Kaplan-Meier analyses were used to compute theincidences of death and nonfatal outcomes at 1-yearpost-TAVR for patients in the early discharge anddelayed discharge groups. For the survival analysesof the nonfatal outcomes, such as stroke and MI, weincorporated the competing effect of death on the riskof these outcomes using Gray test (12). A Cox pro-portional hazards model was applied to examine theassociation of LOS and the outcomes at 1 year. Themodel was adjusted for variables listed in OnlineTable 1. The risk of outcomes of interest for those inthe early discharge versus delayed discharge groupswere reported using hazard ratios with 95% confi-dence interval. All statistical analyses were conduct-ed using SAS version 9.4 (SAS Institute Inc., Cary,North Carolina), and a p < 0.05 was considered sta-tistically significant for all analyses.

RESULTS

STUDY POPULATION. During the study period fromNovember 1, 2011, to September 30, 2015, a total of24,285 patients underwent nonaborted TF-TAVR andsurvived to discharge: 13,389 patients (55.1%) weredischarged within 72 h (i.e., early discharge group),whereas 10,896 patients (44.9%) were dischargedbeyond 72 h (i.e., delayed discharge group). Data forCenters for Medicare & Medicaid Services linkage forthe 1-year outcomes were available for 12,479 patients(51.4% of the entire cohort). The median follow-upestimate was 365 days (Q1, 217; Q3, 365). Table 1summarizes the baseline characteristics of patientsin both groups. Patients in the early discharge groupwere older, men, white, had diabetes, had a history ofMI, prior revascularization (percutaneous coronaryintervention or coronary artery bypass surgery), andwere more likely to have a history of prior pacemakeror defibrillator and a previous history of aortic valveprocedure. Patients in the late discharge group werelikely to have underlying severe lung disease, orhistory of mitral valve procedure.

TEMPORAL TRENDS OF DELAYED DISCHARGE.

During the study period, there was a significantdecline in the rates of delayed discharge (62.2% in the2012 Q1 vs. 34.4% 2015 Q3; p < 0.01) (Figure 1). Thisdecline in the rates of delayed discharge was notedduring the study period across every STS PredictedRisk of Operative Mortality score subgroup (Figure 2).

FIGURE 2 Series Plot of DD Over Time by STS PROM Risk Score

Rates of delayed discharge by quarter by risk group. Patients discharged to home. Blue

line ¼ STS PROM <4%; red line ¼ STS PROM 4% to 8%; green line ¼ STS PROM 8% to

15%; brown line ¼ STS PROM >15%; PROM ¼ Predicted Risk of Operative Mortality score;

STS ¼ Society of Thoracic Surgeons; other abbreviation as in Figure 1.

TABLE 2 Predictors of Delayed Discharge

Odds Ratio(95% CI) p Value

Age $85 yrs 1.14 (1.07–1.22) <0.001

Female 1.16 (0.76–1.78) 0.48

RaceBlack vs. white 1.30 (1.12–1.52) 0.0007Hispanic vs. white 1.18 (1.04–1.34) 0.0079Other vs. white 1.00 (0.85–1.17) 0.9726

LVEF per 5 units 0.99 (0.97–1.00) 0.04

Prior infective endocarditis 0.97 (0.74–1.27) 0.82

Prior permanent pacemaker 0.74 (0.68–0.80) <0.001

Previous ICD 0.80 (0.70–0.92) 0.001

Prior PCI 1.00 (0.95–1.06) 0.94

Prior CABG 0.89 (0.84–0.95) 0.0003

Prior aortic valve procedure 1.01 (0.91–1.13) 0.82

Prior mitral valve procedure 1.62 (1.35–1.95) <0.001

Prior stroke or TIA 0.99 (0.93–1.06) 0.83

Peripheral arterial disease 1.09 (1.01–1.18) 0.03

Smoker 0.86 (0.75–1.00) 0.05

Hypertension 1.01 (0.93–1.11) 0.76

Diabetes mellitus 1.09 (1.03–1.15) 0.004

Severe lung disease 1.10 (1.01–1.20) 0.03

Home oxygen 1.32 (1.20–1.46) <0.001

Immunocompromised 0.97 (0.87–1.08) 0.59

Heart failure in last 2 weeks 1.05 (0.97–1.15) 0.23

Continued on the next page

Wayangankar et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 5 , 2 0 1 9

LOS After TAVR M A R C H 1 1 , 2 0 1 9 : 4 2 2 – 3 0

426

PREDICTORS OF EARLY AND DELAYED DISCHARGE.

Table 2 depicts the predictors of delayed dischargeafter multivariate analyses. Age $85 years, AfricanAmericans, Hispanics, patients with prior mitralvalve procedures, presence of diabetes, home oxy-gen, New York Heart Association functional class IV,cardiogenic shock, need for ionotropes, need formechanical support, atrial fibrillation, dialysis,glomerular filtration rate <30, right ventricular sys-tolic pressure >60 mm Hg, institutional volume 200to 399 cases/year, cut down access technique, >10 son walk test, and use of self-expanding valve wereindependent predictors for delayed discharge. Incontrast, presence of prior intracardiac devices, priorcoronary artery bypass surgery, smokers, prior MI,valve-in-valve, and percutaneous access were inde-pendent predictors of early discharge.

OUTCOMES AT 1 YEAR. Figure 3A shows the com-parison of outcomes based on LOS. The composite ofdeath, stroke, MI, or bleeding was significantly higherin patients who had delayed discharge compared withearly discharge. This was mainly driven by increasedmortality in the former group. Delayed dischargepost-TAVR was significantly associated with higher 1-year all-cause mortality with separation of curvesnoted immediately post-procedure (Figure 3B). Theresults of survival analysis (adjusted and unadjusteddata) on mortality and cumulative incidence ofnonfatal outcomes at 1-year without, and with in-hospital complication are shown in Table 3. Asobserved from the tables, even after adjusting for in-hospital complications, delayed discharge was an in-dependent predictor of 1-year all-cause mortality(hazard ratio: 1.45; 95% confidence interval: 1.30 to1.60; p < 0.01).

DISCUSSION

This is the first large-scale study looking into LOSpost-TF-TAVR. Our study showed that from 2011 to2015 around 55% of the patients were dischargedwithin 72 h. There was a significant decline in rates ofdelayed discharge across the study period (62% vs.34%; p < 0.01). Regardless of whether patients hadpost-procedural complications or not, delayeddischarge was an independent predictor of increased1-year all-cause mortality (hazards ratio: 1.45; 95%confidence interval: 1.30 to 1.60; p < 0.01).

DECLINE IN DELAYED DISCHARGE. Our studyshowed that on an average, more than one-half of theTF-TAVRs end up having delayed discharge duringthe study period. However, it is encouraging to notethat the prevalence of delayed discharge has shown a

TABLE 2 Continued

Odds Ratio(95% CI) p Value

Prior MI 0.91 (0.85–0.99) 0.02

NYHA functional class IV 1.20 (1.10–1.31) <0.001

Inotropic/mechanical support 1.16 (1.07–1.26) <0.001

Atrial fibrillation/flutter 1.17 (1.10–1.24) <.0001

GFR and dialysisOn dialysis vs. GFR $60 1.42 (1.22–1.65) <0.001GFR <30 vs. GFR $60 1.58 (1.38–1.80) <0.001GFR 30–59.9 vs. GFR $60 1.18 (1.11–1.26) <0.001

RVSP40–60 vs. <40 1.05 (0.97–1.13) 0.21$60 vs. <40 1.37 (1.24–1.51) <0.0001

Tricuspid aortic valve 0.92 (0.80–1.05) 0.2209

Aortic insufficiency (moderate/severe) 0.93 (0.87–1.00) 0.05

Mitral insufficiency (moderate/severe) 1.00 (0.94–1.07) 0.93

Tricuspid insufficiency (moderate/severe) 1.06 (0.99–1.13) 0.12

Mitral stenosis 1.01 (0.89–1.15) 0.86

Valve-in-valve 0.72 (0.59–0.87) <0.001

Procedure locationHybrid catheterization laboratory vs. hybrid OR 0.86 (0.75–0.99) 0.03Catheterization laboratory vs. hybrid OR 0.94 (0.76–1.16) 0.54Other vs. hybrid OR 1.24 (0.59–2.61) 0.58

Total valve volume100–199 vs. 0–99 1.22 (0.99–1.50) 0.06200–399 vs. 0–99 1.36 (1.05–1.76) 0.02>400 vs. 0–99 1.35 (0.56–3.23) 0.50

Cutdown vs. percutaneous 1.72 (1.53–1.93) <0.001

Anesthesia typeModerate sedation vs. general anesthesia 0.51 (0.44–0.60) <0.001Other vs. general anesthesia 0.84 (0.54–1.30) 0.42

Unable to walk or >10 s on walk test 1.28 (1.19–1.37) <0.001

Core valve vs. Sapien valve 1.53 (1.40–1.67) <0.001

CI ¼ confidence interval; GFR ¼ glomerular filtration rate; LVEF ¼ left ventricular ejection fraction; RVSP ¼ rightventricular systolic pressure; other abbreviations as in Table 1.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 5 , 2 0 1 9 Wayangankar et al.M A R C H 1 1 , 2 0 1 9 : 4 2 2 – 3 0 LOS After TAVR

427

consistent decline from 2011 to 2015 (62% to 34%).The biggest drop was noted across 2014 Q2 (about10%). Although the financial feasibility of TAVR wasalways questioned since its inception, it was during2014 to 2016 that the concept of “minimalistic TAVR”gained wider acceptance. Appropriate patient selec-tion, modified monitored anesthesia care, post-procedure care in post-anesthesia care units, andlower threshold to remove temporary pacemakerpost-procedure have all significantly helped in-stitutions curtail operating costs. More importantly,this trend persisted when patients were analyzedstratified based on the STS risk (Figure 2). This speaksto the nationwide system-based changes aiming tooptimize TAVRs that have been universally adoptedby all TAVR programs across the country. Thesechanges that were designed to optimize TAVR (andimprove financial and administrative feasibility ofTAVR programs) have had an indirect effect innegating the effect of high TVT scores on LOS in pa-tients undergoing TF-TAVR.

PREDICTORS OF DELAYED DISCHARGE. Advanced age($85 years) was found to be an independent predictorfor delayed discharge. Advanced age is usually asso-ciated with advanced frailty, lower functional capac-ity, increased burden of comorbidities, and higherrisk for periprocedural vascular and nonvascularcomplications, all factors that can delay discharge tohome (13). A testimony to this fact that is that >10 son walk test (a surrogate marker for frailty) wasindependently associated with delayed discharge inour study. Also, our study excluded patients dis-charged to nursing homes, a feature common in oc-togenarians and nonagenarians post-surgery oradvanced procedures. Hence the impact of age ondischarge to home may actually be underestimated inour study. As observed in many other cardiovascularstudies (14), racial disparity in outcomes wasobserved in our study as well. African American andHispanic race were associated with delayeddischarge. Whether this was caused by increaseddisease burden in certain races or related to differ-ential demographic and clinical variables amongdifferent races remains a matter of speculation. Ourstudy shows interesting phenomena with respect toLOS and institutional volume. Institutions perform-ing 200 to 399 cases per year were associated withdelayed discharge. This phenomenon was notobserved when institutions performed <100 cases peryear or >400 cases per year. A combination of patientselection and proficiency of TAVR programs may beresponsible for this observed effect. The institutionsperforming <100 cases per year are relatively lower

volume centers who prefer to super-select certainstraightforward TAVR cases or might chose moreappropriate cases and therefore have minimal impacton delayed discharge. However, institutions >400cases per year despite potentially receiving high-riskcase referrals have years of experience and havetherefore streamlined their TAVR programs, whicheffectively blunts the effect of high-risk cases ondelayed discharge. Institutions who are in betweenare in a transition phase where they are trying to in-crease their institutional volumes but have not yetperfected the art of careful patient selection; hencethe strong association with delayed discharge. Inter-estingly, the use of self-expanding valves was foundto be a predictor for delayed discharge. We believethis may have been related to the practice of keepingtemporary pacemakers for >48 h post-procedurein every patient with core-valve, a mandated

FIGURE 3 Outcomes Based on Length of Stay

(A) Incidence of the unadjusted outcomes (Kaplan-Meier estimates for death and the composite and cumulative incidence function for the nonfatal outcomes in the

presence of death as a competing risk) in both groups. (B) Kaplan-Meier plot of 1-year all-cause mortality in both groups. With number of subjects at risk and 95%

confidence limits. ED ¼ early discharge; TAVR ¼ transcatheter aortic valve replacement; other abbreviation as in Figure 1.

Wayangankar et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 5 , 2 0 1 9

LOS After TAVR M A R C H 1 1 , 2 0 1 9 : 4 2 2 – 3 0

428

requirement in some of the self-expanding valve tri-als. In addition, some centers still practice surgicalcut down when using self-expanding valves, whichmight be another contributor for the delayeddischarge seen with this type of device.

PREDICTORS OF EARLY DISCHARGE. Interestingly, pa-tients with prior MI or prior coronary artery bypasssurgery were found to be independent predictors forearly discharge. The reason for this paradox is un-known but favorable outcomes within the same sub-group have been noted in prior studies withcardiogenic shock patients undergoing percutaneouscoronary intervention (15). In the latter study, itseemed that patients with prior coronary artery dis-ease and revascularization did better than naive pa-tients when they underwent percutaneous coronary

TABLE 3 Survival Analysis on Mortality and Nonfatal Outcomes at 1

Unadjusted Estimates

HR (95% CI) p Value H

All-cause mortality 1.55 (1.39–1.72) <0.0001 1.45

Stroke 1.10 (0.84–1.46) 0.49 1.09

Myocardial infarction 0.98 (0.73–1.32) 0.89 0.97

In-hospital bleeding 1.42 (1.29–1.56) <0.0001 1.32

Composite 1.42 (1.32–1.53) <0.0001 1.34

CI ¼ confidence interval; HR ¼ hazard ratio.

intervention for cardiogenic shock and theauthors in that study attributed to the benefit toischemic pre-conditioning and better integration ofthese patients into local medical systems because oftheir prior coronary artery disease. It is hencepossible that a similar benefit may have accrued inour study patients. Another group of patients whoseemed to have had an impact on the early dischargefavorably were the patients with prior intracardiacdevices. This may have been driven by lesser need forprolonged in-hospital electrical monitoring favoringexpedited discharge. An early discharge in patientswith moderate-severe aortic insufficiency may beexplained by dual hemodynamic benefit of treating 2coexisting valvular lesions. A smoker’s paradox wasalso noted with regards to LOS. This has been noted ina few cardiovascular outcome studies; however, the

Year

Adjusted EstimatesAdjusted

(With In-Hospital Complications)

R (95% CI) p Value HR (95% CI) p Value

(1.30–1.62) <0.0001 1.44 (1.29–1.61) <0.0001

(0.82–1.44) 0.56 1.09 (0.82–1.44) 0.56

(0.71–1.31) 0.83 0.96 (0.71–1.30) 0.77

(1.20–1.46) <0.0001 1.32 (1.19–1.45) <0.0001

(1.24–1.45) <0.0001 1.34 (1.24–1.44) <0.0001

PERSPECTIVES

WHAT IS KNOWN? LOS post-TAVR may have significant clin-

ical and administrative implications.

WHAT IS NEW? LOS for patients undergoing transfemoral

TAVR has significantly decreased from 2011 to 2015. Delayed

discharge is associated with a significant increase in mortality

even after adjusting for in-hospital complications.

WHAT IS NEXT? Predictors of early/delayed discharge could be

used to develop LOS scores, which might be instrumental in

administrative, financial, or clinical policy development.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 5 , 2 0 1 9 Wayangankar et al.M A R C H 1 1 , 2 0 1 9 : 4 2 2 – 3 0 LOS After TAVR

429

exact mechanism for this effect remains speculativeat best. Valve-in-valve therapies seem to do betterbecause the native heart (electrical, circulatory, andannular structure) is relatively protected by the priorsurgical valve and most common reasons for delay indischarge are therefore mitigated (16). Finally,percutaneous access to TF-TAVR has allowed efficientoptimization of TF-TAVR procedures and helped byobviating general anesthesia, expedited early recov-ery, mobilization, and hence early discharge post-procedure. These predictors (of delayed and earlydischarge) could be used to develop risk scores thatcould be used by clinicians for appropriate patientselection and use of resources by administrators, in-surance companies, and policy makers to formulatestratified bundle payments for TAVR procedures infuture.

OUTCOMES. One of the most clinically significantfinding in our study is that the LOS does have sig-nificant association with short- and long-term out-comes. The Kaplan-Meier curves demonstrated thatpatients with delayed discharge had significantlyworse 1-year outcomes compared with earlydischarge cohort. Although this could be as a result ofresidual uncaptured difference in terms of patients’risk profile, the adjusted analyses also demonstratedthat early discharge is associated with a lower 1-yearmortality. Delayed discharge portends a significantlyhigher risk of composite of death, stroke, MI, andbleeding compared with early discharge. This effectwas mainly driven by increased mortality andbleeding (requiring hospitalization) in the formergroup. Because procedural complications have hugeimplications on LOS and morbidity and mortality, weanalyzed our data for outcomes with/withoutadjusting for occurrence of complications and foundsimilar results. This is significant and indicates thatregardless of the occurrence of procedural complica-tions, delayed discharges are a group of TF-TAVRpatients who are at higher risk for mortalitycompared with the early discharge group. Develop-ment of pre-procedural and post-procedural riskscores would help in proper patient selection andmore importantly post-procedurally help in planningcloser follow-up and more rigorous risk factor modi-fication in the delayed discharge to mitigate the riskof higher mortality.

STUDY LIMITATIONS. First, this study is observa-tional in nature. Despite applying statistical analysesto adjust for the differences in the baseline charac-teristics and in-hospital complications, the risk of

residual confounding could not be entirelyexcluded. Second, although there is standardizationand uniformity in the TVT Registry, the data areonly internally validated at sites and not centrallyadjudicated. Third, the field of TF-TAVR had rapidchanges in patient selection, techniques, and post-procedural care; hence, our study results may notentirely be representative of the current practicestandards. Finally, the 1-year outcome data weredriven from an administrative database, thus theoutcomes might have been overestimated. However,this approach has been adopted in previous studiesfrom the TVT Registry (6,11) and other NationalCardiovascular Data Registry databases.

CONCLUSIONS

This is the first study that describes the patterns ofLOS post-TAVR from a large nationwide database.LOS post-TF-TAVR has significantly decreased from2011 to 2015; and with continued refinement intechniques and patient selection, we anticipate thistrend would continue. Predictors of delayeddischarge and early discharge could be used by cli-nicians, administrators, and third-party payers tobetter serve this ever-expanding field of percuta-neous valve therapies. Finally, regardless of thepresence of procedure-related complications,delayed discharge is associated with a significant riskof mortality at 1 year.

ADDRESS FOR CORRESPONDENCE: Dr. Siddharth A.Wayangankar, Division of Cardiovascular Medicine,Department of Medicine, University of Florida, 7919SW 79th Drive, Gainesville, Florida 32610. E-mail:siddharth.wayangankar@medicine.ufl.edu.

Wayangankar et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 5 , 2 0 1 9

LOS After TAVR M A R C H 1 1 , 2 0 1 9 : 4 2 2 – 3 0

430

RE F E RENCE S

1. Smith CR, Leon MB, Mack MJ, et al. Trans-catheter versus surgical aortic-valve replacementin high-risk patients. N Engl J Med 2011;364:2187–98.

2. Adams DH, Popma JJ, Reardon MJ, et al.Transcatheter aortic-valve replacement with aself-expanding prosthesis. N Engl J Med 2014;370:1790–8.

3. Leon MB, Smith CR, Mack MJ, et al. Trans-catheter or surgical aortic-valve replacement inintermediate-risk patients. N Engl J Med 2016;374:1609–20.

4. Reardon MJ, Van Mieghem NM, Popma JJ, et al.Surgical or transcatheter aortic-valve replacementin intermediate-risk patients. N Engl J Med 2017;376:1321–31.

5. Pagnesi M, Chiarito M, Stefanini GG, et al. Istranscatheter aortic valve replacement superior tosurgical aortic valve replacement?: a meta-analysis of randomized controlled trials. J AmColl Cardiol Intv 2017;10:1899–901.

6. Holmes DR Jr., Brennan JM, Rumsfeld JS, et al.Clinical outcomes at 1 year following transcatheteraortic valve replacement. JAMA 2015;313:1019–28.

7. Durand E, Borz B, Godin M, et al. Transfemoralaortic valve replacement with the Edwards SAPIENand Edwards SAPIEN XT prosthesis using exclu-sively local anesthesia and fluoroscopic guidance:

feasibility and 30-day outcomes. J Am Coll CardiolIntv 2012;5:461–7.

8. Durand E, Eltchaninoff H, Canville A, et al.Feasibility and safety of early discharge aftertransfemoral transcatheter aortic valve implanta-tion with the Edwards SAPIEN-XT prosthesis. Am JCardiol 2015;115:1116–22.

9. Serletis-Bizios A, Durand E, Cellier G, et al.A prospective analysis of early discharge aftertransfemoral transcatheter aortic valve implanta-tion. Am J Cardiol 2016;118:866–72.

10. Barbanti M, Capranzano P, Ohno Y, et al. Earlydischarge after transfemoral transcatheter aorticvalve implantation. Heart 2015;101:1485–90.

11. Mack MJ, Brennan JM, Brindis R, et al. Out-comes following transcatheter aortic valvereplacement in the United States. JAMA 2013;310:2069–77.

12. Gray RJ. A class of K-sample tests forcomparing the cumulative incidence of acompeting risk. Ann Stat 1988;16:1141–54.

13. Arsalan M, Szerlip M, Vemulapalli S, et al.Should transcatheter aortic valve replacement beperformed in nonagenarians?: insights from theSTS/ACC TVT Registry. J Am Coll Cardiol 2016;67:1387–95.

14. Wayangankar SA, Kennedy KF, Aronow HD,et al. Racial/ethnic variation in carotid arteryrevascularization utilization and outcomes:

analysis from the National Cardiovascular DataRegistry. Stroke 2015;46:1525–32.

15. Wayangankar SA, Bangalore S, McCoy LA,et al. Temporal trends and outcomes of pa-tients undergoing percutaneous coronary in-terventions for cardiogenic shock in the settingof acute myocardial infarction: a report fromthe CathPCI Registry. J Am Coll Cardiol Intv2016;9:341–51.

16. Tuzcu EM, Kapadia SR, Vemulapalli S,et al. Transcatheter aortic valve replacementof failed surgically implanted bioprostheses:the STS/ACC Registry. J Am Coll Cardiol 2018;72:370–82.

KEY WORDS length of stay, major adversecardiac event(s), mortality, transcatheteraortic valve replacement

APPENDIX For a supplemental table, pleasesee the online version of this paper.

Go to http://www.acc.org/jacc-journals-cme to takethe CME/MOC/ECME quizfor this article.