Accepted Manuscript

Percutaneous Aortic Balloon Valvotomy in the US: A 13 years perspective

Apurva O. Badheka, M.D Nileshkumar J. Patel, M.D Vikas Singh, M.D Neeraj Shah,M.D Ankit Chothani, M.D Kathan Mehta, M.D Abhishek Deshmukh, M.D AbhijitGhatak, M.D Ankit Rathod, M.D Harit Desai, M.D Ghanshyambhai T. Savani, M.DPeeyush Grover, M.D Nilay Patel, M.D Shilpkumar Arora, M.D Cindy L. Grines,M.D Theodore Schreiber, M.D Raj Makkar, M.D Charanjit S. Rihal, M.D Mauricio G.Cohen, M.D Eduardo De Marchena, M.D William W. O'Neill, M.D

PII: S0002-9343(14)00190-9

DOI: 10.1016/j.amjmed.2014.02.025

Reference: AJM 12418

To appear in: The American Journal of Medicine

Received Date: 2 January 2014

Revised Date: 13 February 2014

Accepted Date: 13 February 2014

Please cite this article as: Badheka AO, Patel NJ, Singh V, Shah N, Chothani A, Mehta K, DeshmukhA, Ghatak A, Rathod A, Desai H, Savani GT, Grover P, Patel N, Arora S, Grines CL, SchreiberT, Makkar R, Rihal CS, Cohen MG, De Marchena E, O'Neill WW, Percutaneous Aortic BalloonValvotomy in the US: A 13 years perspective, The American Journal of Medicine (2014), doi: 10.1016/j.amjmed.2014.02.025.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1

Percutaneous Aortic Balloon Valvotomy in the US: A 13 years perspective

Authors: Apurva O. Badheka M.Da*, Nileshkumar J Patel M.Db*, Vikas Singh M.Dc*, Neeraj

Shah M.Db*, Ankit Chothani M.Dd*, Kathan Mehta M.De, Abhishek Deshmukh M.Df, Abhijit

Ghatak M.Dc, Ankit Rathod M.Dg, Harit Desai M.Dc, Ghanshyambhai T. Savani M.Dc,

Peeyush Grover M.Dc, Nilay Patel M.Da , Shilpkumar Arora M.Dh, Cindy L. Grines M.Da,

Theodore Schreiber M.Da, Raj Makkar M.Dg, Charanjit S. Rihal M.D h, Mauricio G. Cohen

M.Dc, Eduardo De Marchena M.Dc, William W. O'Neill M.Di.

a: Detroit Medical Center, Detroit, MI; b: Staten Island University Hospital, Staten Island,

NY; c: University of Miami Miller School of Medicine, Miami, FL; d: MedStar Washington

Hospital Center, Washington, D.C; e: UPMC Shadyside Hospital, Pittsburgh, PA; f:

University of Arkansas, Little Rock, AR; g: Cedars-Sinai Medical Centre, LA, CA; h: Mayo

Clinic, Rochester, MN; i: Henry Ford Hospital, Detroit, MI.

*: Authors share equal contribution to this manuscript.

Running title: Percutaneous Aortic Balloon Valvulotomy.

Funding resources: None. All authors had access to the data and a role in writing the

manuscript.

Disclosures: William O’Neill received honorarium from Medtronic; Mauricio Cohen

received honoraria from Accumed and Edwards Lifesciences, is a speaker for Terumo

Medical and Abiomed and received research grant from The Medicines Company; Eduardo

de Marchena received honorarium from Aegis and has partnership with Tendyne Medical

Inc; all other authors have no conflict of interest.

Key Words: Percutaneous aortic balloon valvulotomy, complications, mortality, hospital

volume, operator volume, length of stay.

Address for correspondence:

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

2

William O’Neill MD, FACC, FSCAI

Director, Center for Structural Heart Disease Henry Ford Hospital 2799 West Grand Boulevard Detroit, MI 48202 e-mail: raj_ohm2000@yahoo.co.in

Word Count: 2984

ABSTRACT

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

3

Background: We determined the contemporary trends of Percutaneous Aortic Balloon

Valvotomy (PABV) and its outcomes using the nation’s largest hospitalization database.

There has been resurgence in the use of PABV in patients at high surgical risk due to

development of less invasive endovascular therapies.

Methods: This is a cross-sectional study with time trends using the Nationwide Inpatient

Sample (NIS) database between the years 1998-2010. We identified patients using the

International Classification of Diseases, 9th Revision, Clinical Modification procedure code

for valvotomy. Only patients >60 years of age with aortic stenosis were included. Primary

outcome included in-hospital mortality and secondary outcomes included procedural

complications and length of hospital stay.

Results: A total of 2,127 PABVs (weighted n=10,640) were analyzed. The utilization rate of

PABV increased by 158% from 12 PABVs/million elderly in 1998-1999 to 31

PABVs/million elderly in 2009-2010 in the US (p<0.001). The hospital mortality declined by

23% from 11.5% in 1998-99 to 8.8% in 2009-2010 (p <0.001). Significant predictors of in-

hospital mortality were presence of increasing comorbidities (p=0.03), unstable patient

(p<0.001), any complication (p<0.001), weekend admission (p=0.008) whereas increasing

operator volume was associated with significantly reduced mortality (p=0.03). Patients

admitted to hospitals with highest procedure volume & with the highest volume operators had

51% reduced likelihood (p=0.05) of in-hospital mortality when compared to those in

hospitals with lowest procedure volume and lowest volume operators.

Conclusion: This study comprehensively evaluates trends for PABV in the US and

demonstrates the significance of operator and hospital volume on outcomes.

INTRODUCTION

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

4

Percutaneous Aortic Balloon Valvotomy (PABV) was first performed in 1986 as a

palliative treatment for elderly patients with symptomatic AS at high-risk for surgery.1,2

Despite the initial enthusiasm, the use of PABV was largely abandoned and the procedure

was limited to palliative situations due to non-durability and disappointing long-term

survival.3 With the recognition that symptomatic AS remains untreated in 30–60% of patients

coupled with the development of less invasive endovascular therapies including Transcatheter

Aortic Valve Replacement (TAVR), there has been resurgence in the use of PABV in patients

considered inoperable or at high surgical risk.4,5 The procedure is now frequently used as a

bridge to more definitive surgical replacement or TAVR. The aim of our study was to

determine the a) utilization trend of PABV, b) mortality rate and its predictors, c)

complications rate and their predictors, d) length of stay, and e) the effect of operator volume

on outcomes of PABV in US using the nation’s largest hospitalization database.

METHODS

Data Source:

The Nationwide Inpatient Sample (NIS) is the largest available all-payer database of

hospital inpatient stays in United States. The 2010 NIS contains all discharge data from 1,051

hospitals located in 45 States, approximating a 20-percent stratified sample of U.S.

community hospitals. Data from the NIS have been used to identify, track, and analyze

national trends in healthcare usage, patterns of major procedures, access, disparity of care,

trends in hospitalizations, charges, quality, and outcomes.6,7 Each individual hospitalization is

de-identified and maintained in the NIS as a unique entry with 1 primary discharge diagnosis

and <24 secondary diagnoses during that hospitalization. Each entry also carries information

on demographic details, insurance status, comorbidities, primary and secondary procedures,

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

5

hospitalization outcome, and length of stay with safeguards to protect the privacy of

individual patients, physicians, and hospitals. Annual data quality assessments are performed

to assure the internal validity of the database. To establish the external validity, database is

compared to the following data sources: the American Hospital Association Annual Survey

Database, the National Hospital Discharge Survey from the National Center for Health

Statistics and the MedPAR inpatient data from the Centers for Medicare and Medicaid

Services.8, 9

Study design and patients:

We queried the Healthcare Cost and Utilization Project’s NIS between 1998 and 2010

using the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-

9-CM) procedure code of 35.96 for percutaneous valvuloplasty. Only patients >60 years of

age with aortic stenosis (AS) (424.1, 395.0, 395.2, 396.2, 746.3) were included. Patients with

concomitant mitral, tricuspid or pulmonic stenosis were excluded (394.0, 394.2, 396.0, 396.1,

396.8, 397.0, 397.1, 746.0, 746.1, 746.5, 424.2, and 424.3).

Utilization rates:

Since NIS represents a 20% stratified random sample of US hospitals, the population

at risk forming the denominator was 20% of the US census population of adults >60 years of

age for any given year.9, 10 Therefore, utilization rates were calculated by dividing the number

of PABV procedures performed, available in the NIS dataset, in a given year divided by 20%

of the US census population >60 years age for that year.

Comorbidities:

We defined severity of comorbid conditions using Deyo modification of Charlson

comorbidity index (CCI).11 This index contains 17 comorbid conditions with differential

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

6

weights. The score ranges from 0 to 33, with higher scores corresponding to greater burden of

comorbid diseases (Supplementary table 1).

Complications:

Procedural complications were identified by Patient Safety Indicators (PSIs) which

have been established by the Agency for Healthcare Research and Quality to monitor

preventable adverse events during hospitalization. These indicators are based on ICD-9-CM

codes and Medicare severity Diagnosis-Related Groups and each PSI has specific inclusion

and exclusion criteria.12,13 PSI individual measure technical specifications, Version 4.4,

March 2012 was used to identify and define preventable complications viz. post-procedure

respiratory failure, post-procedure physiologic and metabolic derangement with acute renal

failure requiring dialysis studied separately, post-procedure pulmonary embolism or deep

venous thrombosis, post- procedure infectious complications which included post-procedure

sepsis & central venous catheter related bloodstream infection, iatrogenic pneumothorax,

complications of anesthesia, pressure ulcers, and accidental puncture or laceration.12 This

methodology has been utilized in earlier studies.14

Other procedure related complications which included post-procedure hemorrhage or

hemorrhage requiring blood transfusion, iatrogenic cardiac complications, complete heart

block, pericardial complications, conversion to open heart surgery, other iatrogenic

respiratory complications (which included ventilator associated pneumonia, post-procedure

aspiration pneumonia and other respiratory complications not elsewhere classified), post-

procedural stroke or transient ischemic attack and vascular complications were identified

using ICD-9-CM codes (Supplementary table 2) in any secondary diagnosis field. In order

to prevent classification of a pre-existing condition (e.g. stroke or heart block) as a

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

7

complication, cases with the ICD-9-CM code for a complication listed as the principal

diagnosis were excluded.

Length of stay and disposition:

The total duration of hospital stay in days was estimated for all patients, after

excluding those who died in the hospital, using the information on length of stay provided in

the NIS dataset. Disposition was classified into 3 categories: (I) those who were discharged

home or with home care services were classified as home-based discharge, (II) those who

were discharged to short or long term nursing home or transferred to another facility were

classified as discharge to another facility and (III) those who died in-hospital were classified

as in-hospital mortality.

Operator procedure volume:

The unique physician-identifying number is specific to this dataset and allowed us to

track the number of PABVs an operator performed in a given year. The operator

identification numbers in NIS do not correlate across years, and hence the same operator

performing the procedure in different years may be recorded under a different identifier,

however the identifiers do not change within the same year. Therefore, annual operator

volume was calculated on a yearly basis by matching the operator identification number

related to a particular procedure to the total number of procedures recorded under that

operator identification number in the given year.12 The data on operator volume were not

recorded for the year 2010 and not all hospitals allow the release of operator specific data.

Hence, operator volume data were available for only 48% (1030/2127) of the population.

High-volume operators were defined as those in the third tertile of annual operator volume

(>3 procedures/year) and low-volume operators were defined as those in the first tertile of

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

8

operator volume (1 procedure/year). Operator volume was incorporated as a continuous

variable in the multivariate model in 5 units per year increments.

Hospital PABV procedure volume:

Annual hospital procedure volume was determined on a year to year basis using the

unique hospital identification number to calculate the total number of PABV procedures

performed by a particular institution in a given year and was stratified by tertile. High volume

centers were defined as those in the third tertile of hospital volume (>19 PABVs/year) & low

volume centers were defined as those in the first tertile of hospital volume (<5 PABVs/year).

Hospital volume was incorporated as a continuous variable in the multivariate model in 10

units per year increments.

Statistical analysis:

Stata IC 11.0 (Stata-Corp, College Station, TX) and SAS 9.2 (SAS Institute Inc, Cary,

North Carolina) was utilized for analyses, which accounted for the complex survey design

and clustering. Weighted values of patient level observations were generated to produce a

nationally representative estimate of the entire US population of hospitalized patients.

Differences between categorical variables were tested using the chi-square test and

differences between continuous variables were tested using student’s t test. P-value of less

than 0.05 was considered significant.

Temporal trend:

For categorical variables like utilization rates, post-procedural complication rates and

in-hospital mortality rates, chi-square test of trend for proportions was used using the

Cochrane Armitage test via the ‘ptrend’ command in Stata.15 For continuous variables like

length of stay, non-parametric test for trend by Cuzick (which is similar to Wilcoxon rank

sum test) using the ‘nptrend’ command in STATA was used.16

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

9

Hierarchical modeling:

Hierarchical mixed effects models were generated in order to identify the independent

multivariate predictors of in-hospital mortality, post-procedural complications and length of

stay. Two level hierarchical models (with patient level factors nested within hospital level

factors) were created with the unique hospital identification number incorporated as random

effects within the model. Hierarchical mixed effects logistic regression models were used for

categorical dependent variables like in-hospital mortality and post-procedural complications

and hierarchical mixed effects linear regression models were used for continuous dependent

variables like length of stay.

Variables with >10% missing data were not included in the multivariate model. We

excluded race (25% missing observations) and type of admission (12% missing

observations). Since 98% of PABV procedures were done in urban hospitals, we did not

include rural/urban location of hospital in the model. In all multivariate models, we included

hospital level variables like hospital region (Northeast, South, Midwest with West as

referent), teaching vs. non-teaching hospital, admission over the weekend and patient level

variables like age, sex, CCI, unstable patient [defined as having a ICD-9 code for shock

(785.5) or ventilator dependence (V461)] & primary payer (with Medicare/Medicaid

considered as referent) in addition to hospital procedure volume or operator procedure

volume or both (see below). All interactions were thoroughly tested. Collinearity was

assessed using variance inflation factor (VIF).

Effect of hospital and operator volume:

Initially, the effects of hospital volume (in 10 unit increments) (n=2127) and operator

volume (in 5 unit increments) (n=1030) were studied separately by creating separate models

incorporating one without the other. Subsequently a third model was created incorporating

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

10

both hospital and operator volume with adjustment for interaction effect between hospital and

operator volume. Hospital ID was incorporated as a random effect in the model to account for

the effect of hospital clustering (meaning that patients treated at the same hospital may

experience similar outcomes as a result of other processes of care).17 Since operator ID did

not remain constant across the years, we could not incorporate it as a random effect in the

model. Separate analysis was performed in the sub-groups of low and high volume hospitals,

with operator volume incorporated as tertiles within the model.

RESULTS

Utilization:

A total of 2,127 PABVs (weighted n=10,640) were available for analysis. Table 1A

and Table 1B demonstrate the baseline characteristics of the study population by categories

of secular time. The utilization rate of PABV has significantly increased by 158% from 12

PABVs/million elderly in 1998-1999 to 31 PABVs/million elderly in 2009-2010 in US

(p<0.001) (Supplementary figure 1), 88.5% of admissions occurred on weekdays and 54.1%

were emergent/urgent. Most of the procedures were performed in large (84.8%), teaching

(81.8%), and urban hospitals (97.6%). Medicare/Medicaid constituted the major primary

expected payer (>90%).

Mortality and predictors:

Overall in-hospital mortality rate was 10.5%. There was a significant observable

decline in hospital mortality among this high-risk population by 23% from 11.5% in 1998-99

to 8.8% in 2009-2010 (p value <0.001) (Supplementary figure 1). The mean CCI of patients

undergoing PABV showed a significant upward trend from 1.5 in 1998 to 2.7 in 2010

(Supplementary figure 2).

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

11

We constructed three multivariable models to determine predictors of mortality:

model 1 included hospital volume, model 2 included operator volume and model 3 included

both (Table 2). In model 3, significant predictors of in-hospital mortality were (OR: 95% CI,

p) presence of increasing CCI (1.16 per unit increase in CCI: 1.01-1.34, p=0.037), unstable

patient (5.75: 3.18-10.38, p<0.001), any complications (3.53: 2.31-5.39, p<0.001) and

weekend admission (2.12: 1.22-3.70, p=0.008) whereas high operator volume was associated

with significantly reduced mortality (0.44 per 5 units increase in operator volume: 0.21-0.92,

p=0.03).

Complications:

Complications were reported in 28.9% of admissions (Table 3). The commonest

complications were respiratory (7.7%), cardiac (7.3%) and vascular (6.8%). Though the

overall rate of complications did not change significantly over the study period (p=0.71);

vascular injury declined significantly from 6% in 1998 to 2.9% in 2010, p <0.0001

(Supplementary figure 3). On multivariate analysis, increasing CCI was associated with

higher in-hospital morbidity (any complication) (OR 1.13, 95% CI 1.02-1.25, p=0.024). The

other independent predictors of in-hospital morbidity were female sex (1.38: 1.03-1.86,

P=0.032) and unstable patients (5.94: 3.37-10.48, P<0.001).

Length of stay and disposition:

The mean length of stay was 9 ± 0.2 days which decreased significantly from 9.1 days

in 1998 to 7.7 days (p=0.002) (Supplementary figure 4). Significant predictors of increased

length of stay were increasing CCI (+0.46: 0.03-0.9, p=0.04), unstable patients (+6.61: 3.6-

9.6, p<0.001), occurrence of any complication (+6.9: 5.6-8.3, p<0.001) and weekend

admission (+5.6: 3.6-7.7, p<0.001). Increased operator volume was associated with decreased

length of stay (-2.81: -4.4-1.2, p<0.001). Majority of the patients were discharged to home

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

12

(62.6%) followed by facility (26.2%). The number patients being discharge to home

remained stable however, there was a significant trend towards increasing discharges to

facilities (p=0.0001) (Supplementary figure 5).

Effect of hospital and operator procedure volume tertiles:

Patients admitted to hospitals with highest procedure volume (> 19 PABVs/year) and

with the highest volume operators (>3 procedures/year) had a 51% reduced likelihood

(p=0.05) of in-hospital mortality when compared to those in hospitals with lowest procedure

volume (<5 PABVs/year) and lowest volume operators (1 procedure/year). High operator

volume (>3 procedures/year) was independently associated with decreased in-hospital

mortality even in the highest tertile of hospital volume (p=0.03) (Figure 1).

DISCUSSION

In summary, our data suggest a significant increase in the number of PABVs being

performed in the US associated with a parallel decline in hospital mortality. Overall

procedural complications have remained similar over the last decade with a significant

decline in vascular injuries. In addition, the procedure was noted to be associated with

significantly decreasing hospital stay. Increasing operator procedure volume was found to be

associated with decreasing mortality and length of stay.

Utilization:

Up to 60% patients with severe, symptomatic AS do not undergo surgical valve

replacement (SVR) due to high operative risk owing to their multiple comorbidities.4

Previously the only option available to this group of patients was palliative PABV. However

PABV was sparingly used due to high rates of complications, recurrence of symptoms and no

survival benefit.18-21 With the recent development of TAVR studies have noted PABV as a

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

13

bridge to valve replacement in up to 37% of the patients who were initially considered

inoperable.22, 23

We noticed a noticeable expansion in the utilization of PABV through 2007- 2010.

We hypothesize this spike to be largely driven by increased TAVR procedures as this period

coincides with the publication of the early TAVR experience and initiation of clinical trials in

US.24, 25 With the FDA approval of Edwards SAPIEN prosthesis (Edwards Lifesciences, CA),

we expect this number to ascend further as clinicians begin to offer TAVR to increasing

number of patients.26, 27 In addition, we observed significant difference in the baseline

characteristics of the patient population undergoing PABVs after 2007. These patients in the

end of the study period were noted to have significantly worse mean CCI and increased

percentage of shock depicting “sicker” patients. The performance of percutaneous coronary

interventions during the same admission also increased significantly during this study period.

Mortality:

Historically, the NHLBI PABV registry in 1991 reported an in-hospital mortality of

10%.17 Subsequent contemporary studies have reported in-hospital mortality of up to 8%.23,

28-31 We noted a trend similar to these studies where the in-hospital mortality associated with

PABV has decreased from 11.5% in 1998 to 8.8% in 2009-10 (p<0.001). It must be borne in

mind that the population undergoing this procedure is very high-risk, elderly and with severe

comorbidities. We noticed a significant increase in the mean CCI of patients undergoing

PABV over the last 13 years especially after 2007. This is also consistent with recent reports

of complicated patients undergoing PABV as a bridge to TAVR.23, 28, 32

Predictors of Mortality:

Our present study found increasing co morbidities (increased CCI) (p=0.007),

unstable patient (p<0.001), any complication during the hospital stay (p<0.001) and weekend

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

14

admission (p=0.002) to be significant predictors of in-hospital mortality as anticipated. The

weekend phenomenon is consistent with previous studies that have suggested that weekend

admission increases the risk of poor outcomes in patients admitted with acute or chronic

cardiac conditions.33-36 We also demonstrate that procedures performed by high volume

operators were associated with significantly reduced mortality. Furthermore, in a subgroup

analysis involving only highest volume centers, the mortality further decreased in the hands

of high volume operators (p=0.03). This operator-volume and outcome relationship with

PABV has not been demonstrated earlier and may emphasize the importance of limiting this

procedure only to high volume centers and operators in order to achieve the best outcomes.

Complications:

In this current study the overall rate of any complications was 28.9%. NHLBI registry

reported significant complications rate as 31% however, up to 61% of the patients developed

a complication after PABV.18 Mansfield registry noted a significant complication rate of

20.5% of the patients.37 Even though the overall complications rate in our study appears

similar to previous studies, it must be noted that definitions for complications in different

studies were variable. We included all in-hospital complications as identified by PSI and not

only selected complications as reported by the previous studies.18, 29-31

We noted blood transfusion in 16.7%, post operative respiratory failure in 6.5%,

complete heart block in 3.1%, pericardial complications such as cardiac tamponade in 0.8%

and vascular complications in 6.8%. Post operative stroke/transient ischemic attack occurred

in 2.9% and renal failure requiring hemodialysis was noted in 1.7%. These rates are

consistent with previously published studies except for reduced rates of bleeding and vascular

complications reported in our study. NHLBI reported need for transfusion in 23%,

atrioventricular block requiring pacing in 4%, cardiac tamponade in 1%, vascular surgery in

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

15

7%, cerebrovascular accidents in 3% and acute tubular necrosis in 1%.18 Mansfield registry

reported atrioventricular block requiring pacing in 0.2%, cardiac tamponade in 1.8%,

vascular injury in 11%, cerebrovascular accidents in 2.2% and acute renal failure in 0.4%.37

Ben-Dor et al recently reported a16.2% severe complication rate in a cohort of 262 patients.28

In another study from the same institution, Ben-Dor et al noticed blood transfusion in 23%,

atrioventricular block requiring pacing in 1.3%, cardiac tamponade in 0.7%, vascular injury

requiring interventions in 6.3%, cerebrovascular accidents in 2.0% and acute kidney injury in

9%.23 However this was a single center experience involving highly experienced operators.

Our present study is a cross sectional “real world” US experience involving both high and

low volume centers. In addition, our rates are calculated according to previously validated

administrative metrics.38 We also included complications such as infections, pressure ulcers,

deep venous thrombosis, etc which have been overlooked in the previous studies.

Another important finding of this study was a significant decline in rates of vascular

injuries (p<0.0001) over the last decade. This observation is also consistent with more recent

studies that have also reported decreased rates of vascular complications 6-7%.23, 29 This is

likely secondary to multiple modifications have been made in the PABV technique over the

last decade. Rapid ventricular pacing during PABV enabling precise and stable balloon

positioning, the newer balloon catheters allowing the use of a 10-F sheath instead of a 13-F

sheath and availability of closure devices and use of anticoagulants such as Bivalirudin are

some of the developments that have lead to a decline in complications especially vascular.39-

43 Additionally, majority of the patients now have an assessment of the peripheral vascular

tree by contrast computer tomography as a part of screening for TAVR, thereby allowing

selection of a safer vascular approach. A high rate of respiratory complications (7.6%) could

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

16

be secondary to volume overload leading to pulmonary congestion and edema in this patient

population. It also underlines the importance of aggressive pulmonary toilet post PABV.

Length of stay:

We found overall mean length of stayto be 9 days. Previous studies have shown an

average length of stay up to 12 days with single experienced centers showing reduced stay.18,

23, 28, 44 In congruence with these findings we observed a significant decline in length of stay

from 9.1 days in 1998 to 7.7 days in 2010 (p=0.002). The decline in length of stay was also

temporally associated with increasing discharges to other facilities. On multivariable analysis,

unstable patients, occurrence of any complication and increased CCI were significant

predictors increased hospital stay.

Similar to our observation in improved mortality outcome, we observed a significant

impact of increased operator volume in decreasing the length of stay (Figure 1). A high

volume operator at a specialized high volume center appears to play a crucial role in a

successful PABV procedure. This association between operator and hospital procedure

volume and patient outcomes is similar to that documented for percutaneous coronary

interventions.45, 46 The results of our study therefore suggest that the best outcome is obtained

when PABV is performed by a high volume operator at a high volume center.

This study has limitations which are inherent to any large post hoc analysis. We used

an administrative database and data were not independently verified. Variables that were not

collected as part of the NIS could not be included in the analysis. Compared to previous

studies we lack information such as baseline functional status of patients, cardiac output,

aortic valve area, hemodynamic data, etc. In addition we did not have post-discharge data on

mortality or functional status as the NIS does not track individual patients over time. We

could not determine if a PABV was performed as "destination" palliative therapy or as a

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

17

"bridge" to definitive therapy (SVR or TAVR). The ICD-9-CM procedure codes for TAVR

were not available during the study period to confirm our hypothesis. The database did not

allow us to delineate the cause of a given complication, and we could not generally

differentiate between operator related, patient-related, and device-related events. Similarly,

we selected for admissions likely to be for the primary purpose of PABV however, we cannot

definitively attribute all complications to this procedure. It is also possible that hospitals

might have been more aggressive in coding secondary diagnoses over time for reimbursement

purposes. In addition, we could not determine the time of PABV during the hospitalization

which could have impacted the length of stay. Some patients may have been admitted with

heart failure and had PABV performed for stabilization while some may have been electively

admitted for PABV as a bridge. Whereas unique physician identifiers were used to identify

the number of procedures performed by an operator for a given year, it was impossible to

determine the extent of participation of the operator or trainees in a given procedure, which

may lead to an unmeasured bias. In addition, operator volume data were only available for

48% of the population. Acknowledging these limitations, the present study has important

strengths including the largest sample size, use of standardized definitions of preventable

adverse events that are established by the Agency for Healthcare Research and Quality. This

is also the first study to investigate the effect of operator and hospital volume on outcomes

after PABV.

In conclusion, this study comprehensively evaluates contemporary utilization,

mortality and complications trends for PABV in the US. The technique and devices used in

the procedure have changed considerably over the last decade however the complication rates

indicate a room for further improvement.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

18

REFERENCES

1. McKay RG, Safian RD, Lock JE, Mandell VS, Thurer RL, Schnitt SJ, Grossman W. Balloon dilatation of calcific aortic stenosis in elderly patients: postmortem, intraoperative, and percutaneous valvuloplasty studies. Circulation. 1986;74:119-25.

2. Cribier A, Savin T, Saoudi N, Rocha P, Berland J, Letac B. Percutaneous transluminal valvuloplasty of acquired aortic stenosis in elderly patients: an alternative to valve replacement? Lancet. 1986;1:63-7.

3. Block PC, Palacios IF. Clinical and hemodynamic follow-up after percutaneous aortic valvuloplasty in the elderly. Am J Cardiol. 1988;62:760-3.

4. Iung B, Cachier A, Baron G, Messika-Zeitoun D, Delahaye F, Tornos P, Gohlke-Bärwolf C, Boersma E, Ravaud P, Vahanian A. Decision-making in elderly patients with severe aortic stenosis: why are so many denied surgery? Eur Heart J. 2005;26:2714-20.

5. Charlson E, Legedza AT, Hamel MB. Decision-making and outcomes in severe symptomatic aortic stenosis. J Heart Valve Dis. 2006;15:312-21.

6. Kumar G, Kumar N, Taneja A, Kaleekal T, Tarima S, McGinley E, Jimenez E, Mohan A, Khan RA, Whittle J, Jacobs E, Nanchal R; Milwaukee Initiative in Critical Care Outcomes Research Group of Investigators. Nationwide trends of severe sepsis in the 21st century (2000-2007). Chest. 2011;140:1223-31.

7. Deshmukh A, Kumar G, Kumar N, Nanchal R, Gobal F, Sakhuja A, Mehta JL. Effect of Joint National Committee VII report on hospitalizations for hypertensive emergencies in the United States. Am J Cardiol. 2011;108:1277-82.

8. Nationwide Inpatient Sample (NIS) related reports. Available at: http://www.hcupus.ahrq.gov/db/nation/nis/nisrelatedreports.jsp Accessed January 31, 2013.

9. Epstein AJ, Polsky D, Yang F, Yang L, Groeneveld PW. Coronary revascularization trends in the united states, 2001-2008. JAMA. 2011;305:1769-1776.

10. US Census Bureau. Annual estimates of the resident population by sex and selected age groups for the United States: April 1, 2000 to July 1, 2009 (NC-EST2009-02). http://www.census.gov/popest/national/asrh/NC-EST2009-sa.html. Accessed January 31, 2013.

11. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613-9.

12. McDonald KM, Romano PS, Geppert J, Davies SM, Duncan BW, Shojania KG, Hansen A. Measures of Patient Safety Based on Hospital Administrative Data—The Patient Safety Indicators. Rockville, MD: Agency for Healthcare Reseach and Quality 2002;Vol 02-0038.

13. Romano PS, Geppert JJ, Davies S, Miller MR, Elixhauser A, McDonald KM. A national profile of patient safety in U.S. hospitals. Health Aff (Millwood). 2003;22:154-66.

14. Deshmukh A, Patel NJ, Pant S, Shah N, Chothani A, Mehta K, Grover P, Singh V, Vallurupalli S, Savani GT, Badheka A, Tuliani T, Dabhadkar K, Dibu G, Reddy YM, Sewani A, Kowalski M, Mitrani R, Paydak H, Viles-Gonzalez JF. In-hospital complications associated with catheter ablation of atrial fibrillation in the United States between 2000 and 2010: analysis of 93 801 procedures. Circulation. 2013;128:2104-12.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

19

15. Armitage P. Tests for Linear Trends in Proportions and Frequencies. Biometrics. 1955;Vol 11, No 3, pp 375-386.

16. Cuzick J. A Wilcoxon-type test for trend. Stat Med. 1985;4:87-90. 17. Panageas KS, Schrag D, Riedel E, Bach PB, Begg CB. The effect of clustering of

outcomes on the association of procedure volume and surgical outcomes. Ann Intern Med. 2003;139:658-65.

18. Percutaneous balloon aortic valvuloplasty. Acute and 30-day follow-up results in 674 patients from the NHLBI Balloon Valvuloplasty Registry. Circulation. 1991;84:2383-97.

19. Safian RD, Berman AD, Diver DJ, McKay LL, Come PC, Riley MF, Warren SE, Cunningham MJ, Wyman RM, Weinstein JS, Grossman W, McKay RG. Balloon aortic valvuloplasty in 170 consecutive patients. N Engl J Med. 1988;319:125-30.

20. Otto CM, Mickel MC, Kennedy JW, Alderman EL, Bashore TM, Block PC, Brinker JA, Diver D, Ferguson J, Holmes DR Jr. Three-year outcome after balloon aortic valvuloplasty. Insights into prognosis of valvular aortic stenosis. Circulation. 1994;89:642-50.

21. Lieberman EB, Bashore TM, Hermiller JB, Wilson JS, Pieper KS, Keeler GP, Pierce CH, Kisslo KB, Harrison JK, Davidson CJ. Balloon aortic valvuloplasty in adults: failure of procedure to improve long-term survival. J Am Coll Cardiol. 1995;26:1522-8.

22. Saia F, Marrozzini C, Moretti C, Ciuca C, Taglieri N, Bordoni B, Dall'ara G, Alessi L, Lanzillotti V, Bacchi-Reggiani ML, Branzi A, Marzocchi A. The role of percutaneous balloon aortic valvuloplasty as a bridge for transcatheter aortic valve implantation. EuroIntervention. 2011;7:723-9.

23. Ben-Dor I, Maluenda G, Dvir D, Barbash IM, Okubagzi P, Torguson R, Lindsay J, Satler LF, Pichard AD, Waksman R. Balloon aortic valvuloplasty for severe aortic stenosis as a bridge to transcatheter/surgical aortic valve replacement. Catheter Cardiovasc Interv. 2013; 82:632-7.

24. Cribier A, Eltchaninoff H, Tron C, Bauer F, Agatiello C, Nercolini D, Tapiero S, Litzler PY, Bessou JP, Babaliaros V. Treatment of calcific aortic stenosis with the percutaneous heart valve: mid-term follow-up from the initial feasibility studies: the French experience. J Am Coll Cardiol. 2006;47:1214-23.

25. Webb JG, Pasupati S, Humphries K, Thompson C, Altwegg L, Moss R, Sinhal A, Carere RG, Munt B, Ricci D, Ye J, Cheung A, Lichtenstein SV. Percutaneous transarterial aortic valve replacement in selected high-risk patients with aortic stenosis. Circulation. 2007;116:755-63.

26. Leon MB, Smith CR, Mack M, Miller DC, Moses JW, Svensson LG, Tuzcu EM, Webb JG, Fontana GP, Makkar RR, Brown DL, Block PC, Guyton RA, Pichard AD, Bavaria JE, Herrmann HC, Douglas PS, Petersen JL, Akin JJ, Anderson WN, Wang D, Pocock S; PARTNER Trial Investigators. Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med. 2010;363:1597-607.

27. Smith CR, Leon MB, Mack MJ, Miller DC, Moses JW, Svensson LG, Tuzcu EM, Webb JG, Fontana GP, Makkar RR, Williams M, Dewey T, Kapadia S, Babaliaros V, Thourani VH, Corso P, Pichard AD, Bavaria JE, Herrmann HC, Akin JJ, Anderson WN, Wang D, Pocock SJ; PARTNER Trial Investigators. Transcatheter versus surgical aortic-valve replacement in high-risk patients. N Engl J Med. 2011;364:2187-98.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

20

28. Ben-Dor I, Pichard AD, Satler LF, Goldstein SA, Syed AI, Gaglia MA Jr, Weissman G, Maluenda G, Gonzalez MA, Wakabayashi K, Collins SD, Torguson R, Okubagzi P, Xue Z, Kent KM, Lindsay J, Waksman R. Complications and outcome of balloon aortic valvuloplasty in high-risk or inoperable patients. JACC Cardiovasc Interv. 2010;3:1150-6.

29. Agarwal A, Kini AS, Attanti S, Lee PC, Ashtiani R, Steinheimer AM, Moreno PR, Sharma SK. Results of repeat balloon valvuloplasty for treatment of aortic stenosis in patients aged 59 to 104 years. Am J Cardiol. 2005;95:43-7.

30. Kvidal PD, Stahle E, Nygren A, Landelius J, Thuren J, Enghoff E. Long-term follow up study on 64 elderly patients after balloon aortic valvuloplasty. J Heart Valve Dis. 1997;6:480-6.

31. Robicsek F. The burst of the balloon. The Mansfield Scientific Balloon Aortic Valvuloplasty Registry. European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery. 1993;7:57-62;discussion 62-4.

32. Martinez CA, Singh V, Londoño JC, Cohen MG, Alfonso CE, O'Neill WW, Heldman AW. Percutaneous retrograde left ventricular assist support for interventions in patients with aortic stenosis and left ventricular dysfunction. Catheter Cardiovasc Interv. 2012;80:1201-9.

33. Kostis WJ, Demissie K, Marcella SW, Shao YH, Wilson AC, Moreyra AE. Weekend versus weekday admission and mortality from myocardial infarction. N Engl J Med. 2007;356:1099-109.

34. Horwich TB, Hernandez AF, Liang L, Albert NM, Labresh KA, Yancy CW, Fonarow GC; Get With Guidelines Steering Committee and Hospitals. Weekend hospital admission and discharge for heart failure: association with quality of care and clinical outcomes. Am Heart J. 2009;158:451-8.

35. Ottesen MM, Kober L, Jorgensen S, Torp-Pedersen C. Determinants of delay between symptoms and hospital admission in 5978 patients with acute myocardial infarction. The TRACE Study Group. Trandolapril Cardiac Evaluation. Eur Heart J. 1996;17:429-37.

36. Deshmukh A, Pant S, Kumar G, Bursac Z, Paydak H, Mehta JL. Comparison of outcomes of weekend versus weekday admissions for atrial fibrillation. Am J Cardiol. 2012;110:208-11.

37. McKay RG. The Mansfield Scientific Aortic Valvuloplasty Registry: overview of acute hemodynamic results and procedural complications. J Am Coll Cardiol. 1991;17:485-91.

38. Eppsteiner RW, Csikesz NG, McPhee JT, Tseng JF, Shah SA. Surgeon volume impacts hospital mortality for pancreatic resection. Ann Surg. 2009;249:635-40.

39. Bui QT, Kolansky DM, Bannan A, Herrmann HC. "Double wire" angio-seal closure technique after balloon aortic valvuloplasty. Catheter Cardiovasc Interv. 2010;75:488-92.

40. Ben-Dor I, Looser P, Bernardo N, Maluenda G, Torguson R, Xue Z, Lindsay J, Pichard AD, Satler LF, Waksman R. Comparison of closure strategies after balloon aortic valvuloplasty: suture mediated versus collagen based versus manual. Catheter Cardiovasc Interv. 2010;78:119-24.

41. David F, Sánchez A, Yánez L, Velásquez E, Jiménez S, Martínez A, Alva C. Cardiac pacing in balloon aortic valvuloplasty. Int J Cardiol. 2007;116:327-30.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

21

42. Bauer J, Groneberg DA. Perception of stress-related working conditions in hospitals (iCept-study): a comparison between physicians and medical students. J Occup Med Toxicol. 2013;8:3.

43. O'Neill B, Singh V, Kini A, Mehran R, Jacobs E, Knopf D, Alfonso CE, Martinez CA, Martinezclark P, O'Neill W, Heldman AW, Yu J, Baber U, Kovacic JC, Dangas G, Sharma S, Sartori S, Cohen MG. The use of vascular closure devices and impact on major bleeding and net adverse clinical events (NACEs) in balloon aortic valvuloplasty: A sub-analysis of the BRAVO study. Catheter Cardiovasc Interv. 2014;83:148-53.

44. Witzke C, Don CW, Cubeddu RJ, Herrero-Garibi J, Pomerantsev E, Caldera A, McCarty D, Inglessis I, Palacios IF. Impact of rapid ventricular pacing during percutaneous balloon aortic valvuloplasty in patients with critical aortic stenosis: should we be using it? Catheter Cardiovasc Interv. 2010;75:444-52.

45. McGrath PD, Wennberg DE, Dickens JD Jr, Siewers AE, Lucas FL, Malenka DJ, Kellett MA Jr, Ryan TJ Jr. Relation between operator and hospital volume and outcomes following percutaneous coronary interventions in the era of the coronary stent. JAMA. 2000;284:3139-44.

46. Moscucci M, Share D, Smith D, O'Donnell MJ, Riba A, McNamara R, Lalonde T, Defranco AC, Patel K, Kline Rogers E, D'Haem C, Karve M, Eagle KA. Relationship between operator volume and adverse outcome in contemporary percutaneous coronary intervention practice: an analysis of a quality-controlled multicenter percutaneous coronary intervention clinical database. J Am Coll Cardiol. 2005;46:625-32.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

22

FIGURE TITLES AND LEGENDS

Figure 1: A) Graph showing % in-hospital mortality by operator volume tertiles in the two

subgroups of hospital volume. LVH=Low volume hospital < 5 PABV/yr, HVH=High volume

hospital > 19 PABV/yr. Adjusted OR for in-hospital mortality for 3rd vs. 1st tertile of

operator volume in LVH= 0.86, 0.25-2.94, p=0.81, Adjusted OR for in-hospital mortality for

3rd vs. 1st tertile of operator volume in HVH= 0.38, 0.17-0.89, p=0.026.

B) Graph showing mean length of stay by operator volume tertiles in the two

subgroups of hospital volume. Adjusted estimate for length of stay for 3rd vs. 1st tertile of

operator volume in LVH= -5.73, -9.94 to -1.52, p=0.008. Adjusted estimate for length of stay

for 3rd vs. 1st tertile of operator volume in HVH= -4.33, -6.77 to -1.89, p=0.001.

Supplementary figure 1: Utilization and mortality of BAV: Significant increase in BAV by

1.25 per million population per year, p<0.001 and significant decrease in mortality by 0.41%

per year, p<0.0001.

Supplementary figure 2: CCI trend: Increasing mean CCI from 1.5 in 1998 to 2.7 in 2010.

P<0.001

Supplementary figure 3: Vascular injury trend: Decrease in vascular injury rate, p<0.0001.

Supplementary figure 4: Length of stay: Significant decrease in length of stay (p=0.002).

Supplementary figure 5: Disposition trend: Increasing trend towards discharge to other facility, slope=0.43% per year, p=0.001.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPTTable 1A: Baseline characteristics of the study population undergoing BAV procedure in the United States over the study period of 13

years from 1998-2010.

Actual (projected) number of BAV procedure in

the United States

Total

2127 (10640)

1998-2001

486 (2480)

2002-2006

657 (3279)

2007-2010

984 (4881)

P value

Patient characteristics

Age 82.4 ± 7.9 82.6 ± 7.7 82.7 ± 7.7 82.1 ± 8.2 0.300

Female (%) 53.9% 58.2% 50.9% 53.8% 0.070

Race* (%) <0.001

White 66.8 74.5 62 66.1

Non-white 8.3 5.8 7.7 9.9

Unstable Patient† (%) 7.9 7.0 6.5 9.2 0.100

Charlson/deyo comorbidity index‡ 2.2 ± 1.6 1.6 ± 1.3 1.9 ± 1.3 2.7 ± 1.8 <0.001

Comorbidities§ (%)

Obesity 5.2 § 2.5 7.0 <0.001

History of hypertension 59.4 § 51.7 64.5 <0.001

History of diabetes 28.9 § 23.6 32.4 <0.001

History of congestive heart failure 8.1 § 13.7 4.3 <0.001

History of chronic pulmonary disease 27.5 § 26.7 28 0.610

Pulmonary circulatory disorder 1.5 § 1.1 1.8 0.300

Peripheral vascular disease 15.9 § 14.2 17.1 0.100

Fluid-electrolyte abnormalities and or Renal

failure

39.3 § 28 46.8 <0.001

Neurological disorder or paralysis 4.5 § 4.2 4.7 0.560

Anemia or coagulopathy 29.4 § 27 31 0.070

Hematological or oncological malignancy 8.0 § 9.2 7.3 0.200

Weight loss/cachexia 3.9 § 3.8 4.0 0.900

Rheumatoid arthritis or other collagen vascular

Disease

2.9 § 1.8 3.6 0.030

Depression, psychosis or substance abuse 7.5 § 6.1 8.3 0.100

Median household income category for patient's zip code|| <0.001

1. 0-25th percentile 13.4 2.7 11.9 19.8

2. 26-50th percentile 21.1 12.8 18.4 27.2

3. 51-75th percentile 25.9 26.3 29.3 23.4

4. 76-100th percentile 39.5 58.2 40.4 29.6

Primary Payer 0.02

Medicare/Medicaid 1931 (90.8%) 88.8 92.2 91

Private including HMO 159 (7.6%) 10.4 6.5 6.9

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT Self pay/no charge/other 34 (1.6%) 0.8 1.3 2.1

*Race was missing in 25% of the study population and hence excluded in the multivariable analysis.

†Unstable was defined as (DX1-DX25) having a listed code for shock (ICD-9-CM code 785.5) or ventilator dependence (ICD-9-CM code

V461).

‡Charlson/deyo comorbidity index was calculated as per Deyo classification.

§Variables are AHRQ comorbidity measures, which were only available from 2002 through 2010 (n = 1632)

||This represents a quartile classification of the estimated median household income of residents in the patient's ZIP Code. These values are

derived from ZIP Code-demographic data obtained from Claritas. The quartiles are identified by values of 1 to 4, indicating the poorest to

wealthiest populations. Because these estimates are updated annually, the value ranges vary by year. http://www.hcup-

us.ahrq.gov/db/vars/zipinc_qrtl/nisnote.jsp.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPTTable 1B: Baseline characteristics of the study population undergoing BAV procedure in the United States over the study period of 13

years from 1998-2010.

Actual (projected) number of BAV procedure in

the United States

Total

2127 (10640)

1998-2001

486 (2480)

2002-2006

657 (3279)

2007-2010

984 (4881)

P value

Hospital characteristics

Hospital bed size <0.001

Small 7.7% 10 7.2 5.5

Medium 7.5% 11.9 4.1 7.8

Large 84.8% 78.1 88.7 86.7

Rural hospital 2.4% 1.3 2.5 2.9 0.04

Hospital region <0.001

Northeast 40.1% 47.8 44.7 33

Midwest or North Central 24.4% 15.9 21.3 30.8

South 16.9% 15.7 12.4 20.7

West 18.6% 20.6 21.6 15.5

Teaching hospitals 81.8% 70.7 82.6 86.8 <0.001

Non-elective admission (%)¶ 54.1% 51.3 54.2 55.6 0.330

Weekend admission (%) 11.5% 13.1 11.5 10.8 0.340

Other major cardiac procedures performed during hospitalization (ICD 9 procedure code) (%)

Coronary angiography (88.55, 88.56) 58.5% 55.5% 61.4% 58% 0.170

Percutaneous coronary intervention (36.06, 36.07) 5.2% 4.6 8.7 11.1 <0.001

Intraaortic balloon pump (37.61) 4% 3.1 4.1 4.4 0.450

Left ventricular assist device (37.68, 37.62) 0.5% 0 0.8% 0.6% 0.180

Morbidity mortality and outcome

Length of stay (Means ± SE) 9.0 ± 0.2 9.5 ± 0.7 8 ± 0.4 8.2 ± 0.29 0.06

Peri-procedural complications# (%) 28.9 28.2 28.4 29.7 0.77

Disposition 0.01

Home 62.9% 63 62.3 63.3

Facility 26.6% 25.2 24.7 28.6

Death 10.5 11.8 13 8.1 0.003

¶ Admission type was missing in 12% of the population.

# See supplementary table 2 for details of procedural complications coding.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPTTable 2: Multivariable predictors of in-hospital mortality in the study population undergoing BAV over the study period of 13 years,

from 1998-2010.

Variables Model 1 Model 2 Model 3

OR (95% CI) P value OR (95% CI) P value OR (95% CI) P value

Deyo modification of Charlson

comorbidity index (CCI) 1.13 (1.03-1.24) 0.007 1.17 (1.01-1.35) 0.034 1.16 (1.01-1.34) 0.037

Unstable patient* 4.44 (3.00-6.56) <0.001 5.98 (3.3-10.8) <0.001 5.75 (3.18-10.38) <0.001

Weekend admission 1.86 (1.25-2.77) 0.002 2.20 (1.26-3.82) 0.005 2.12 (1.22-3.70) 0.008

Any complications† 3.57 (2.60-4.88) <0.001 3.53 (2.31-5.38) <0.001 3.53 (2.31-5.39) <0.001

Hospital procedure volume (per 10

procedures increase per year) 0.89 (0.82-0.97) 0.006 - - 0.90 (0.78-1.04) 0.15

Operator volume (per every 5 units

increase per year) - - 0.75 (0.58-0.97) 0.031 0.44 (0.21-0.92) 0.03

Two levels hierarchical mixed effects models were generated (patient level factors nested within hospital level factors) with the unique hospital

identification number incorporated as random effects. Model 1 (n=2123) was adjusted for age, sex, Deyo’s modification of Charlson

Comorbidity Index, unstable patients, occurrence of any post-procedural complication, primary payer, weekend admission, hospital region,

hospital teaching status and hospital volume. In model 2 (n=1028) we included all variables in model 1 but operator procedure volume was

included instead of hospital volume. In model 3 (n=1028) we incorporated hospital and operator volume together plus all other variables used in

model 1 and 2.

Operator and hospital volume were calculated based on the unique operator and hospital identification number on year to year basis.

Model 2 and 3 were not assessed in the full study population due to limited availability of operator ID for all hospitals.

* Unstable patients were defined as those having a listed code for shock (ICD-9-CM code 785.5) or ventilator dependence (ICD-9-CM code

V461).

† Please see supplementary table 2 for details of procedural complications coding.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPTTable 3: Post procedural complications related to BAV by ICD 9 codes, 1998-2010.

Post-procedural complications Percentage

Any complication 28.9%

Vascular complications 6.8%

1) Hemorrhage requiring transfusion* 2.7%

2) Vascular injury 4.4%

Cardiac complications 7.3%

1) Iatrogenic cardiac complications 4.6%

2) Complete heart block 3.1%

3) Pericardial complications 0.8%

Conversion to open heart surgery 2.5%

Respiratory complications 7.6%

1) Iatrogenic pneumothorax <0.5%

2) Respiratory failure 6.51%

3) Other iatrogenic respiratory complications 0.6%

Neurological Complications

Stroke/transient ischemic attack 2.9%

Renal and metabolic complications 1.8%

1) Acute renal failure requiring dialysis 1.7%

2) Acute severe metabolic derangement <0.5%

Deep vein thrombosis or pulmonary embolism 2.2%

Infectious complications† 4.4%

Pressure ulcer rate 3.4%

Complications of anesthesia <0.5%

*Hemorrhage requiring transfusion was identified as having any patient having post-operative hemorrhage and also got transfusion.

†Infectious complications were identified as composite of post-operative sepsis, septic shock or catheter related infection.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Figure 1: A) Graph showing % in-hospital mortality by operator volume tertiles in the two

subgroups of hospital volume. LVH=Low volume hospital < 5 PABV/yr, HVH=High volume

hospital > 19 PABV/yr. Adjusted OR for in-hospital mortality for 3rd vs. 1st tertile of operator

volume in LVH= 0.86, 0.25-2.94, p=0.81, Adjusted OR for in-hospital mortality for 3rd vs. 1st

tertile of operator volume in HVH= 0.38, 0.17-0.89, p=0.026.

B) Graph showing mean length of stay by operator volume tertiles in the two subgroups

of hospital volume. Adjusted estimate for length of stay for 3rd vs. 1st tertile of operator volume

in LVH= -5.73, -9.94 to -1.52, p=0.008. Adjusted estimate for length of stay for 3rd vs. 1st tertile

of operator volume in HVH= -4.33, -6.77 to -1.89, p=0.001.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Clinical Significance:

• The utilization rate of Percutaneous Aortic Balloon Valvotomy increased by 158% and

the mortality declined by 23%.

• Presence of multiple comorbidities, unstable patient, any complication and weekend

admission increase in-hospital mortality.

• Patients admitted to hospitals with highest procedure volume and with the highest volume

operators had 51% reduced likelihood of in-hospital mortality.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Supplementary table 1: Deyo modification of Charlson’s co-morbidity index (CCI). Reported ICD-9 CM Codes Condition Charlson Score

410 – 410.9 Myocardial Infarction 1

428 – 428.9 Congestive Heart Failure 1

433.9, 441 – 441.9, 785.4,

V43.4

Peripheral Vascular Disease 1

430 – 438 Cerebrovascular Disease 1

290 – 290.9 Dementia 1

490 – 496, 500 – 505, 506.4 Chronic Pulmonary Disease 1

710.0, 710.1, 710.4, 714.0 –

714.2, 714.81, 725

Rheumatologic Disease 1

531 – 534.9 Peptic Ulcer Disease 1

571.2, 571.5, 571.6, 571.4 –

571.49

Mild Liver Disease 1

250 – 250.3, 250.7 Diabetes 1

250.4 – 250.6 Diabetes with Chronic Complications 2

344.1, 342 – 342.9 Hemiplegia or Paraplegia 2

582 – 582.9, 583 – 583.7, 585,

586, 588 – 588.9

Renal Disease 2

140-172.9, 174-195.8, 200-

208.9

Any malignancy including leukemia

and lymphoma

2

572.2 – 572.8 Moderate or Severe Liver Disease 3

196-199.1 Metastatic solid tumor 6

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

042 – 044.9 AIDS 6

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Supplementary table 2: Post procedural complications by ICD 9 code.

Post-procedural complications ICD 9 CODE

Vascular complications

1) Hemorrhage requiring transfusion 998.11Or 998.12 + 99.0

2) Vascular injury -Injury to blood vessels-900-904

-Accidental puncture-998.2, e870 (PSI*)

-Arteriovenous fistula-447

-Injury to retro-peritoneum 868.04

-Vascular complications requiring surgery-

39.31, 39.41, 39.49, 39.52, 39.53, 39.56,

39.57, 39.58, 39.59, 39.79

-Other vascular complications-999.2, 997.7

Cardiac complications

1) Iatrogenic cardiac complications 997.1

2) Complete heart block 37.71, 37.73, 37.81, 37.82, 37.83,

0051,v450.1, 426.0

3) Pericardial complications 423.0-Hemopericardium

423.3-Cardiac temponade

37.0-Pericardiocentesis

Conversion to open heart surgery 35.1, 35.2, 35.32, 35.33, 35.34, 35.35,

35.42, 35.50, 35.51, 35.52, 35.53, 35.54,

35.6, 35.7, 35.8, 35.91, 35.9, 36.31, 36.32,

37.32, 37.33, 37.34, 37.35, 37.51, 37.52,

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

37.53, 37.54

Respiratory complications

1) Iatrogenic pneumothorax PSI *

2) Respiratory failure PSI *

3) Other iatrogenic respiratory

complications

997.3

Neurological Complications

Stroke/transient ischemic attack 997.0, 997.00, 997.01, 997.02, 435.9, 438.0,

438.10, 438.11, 438.12, 438.19, 438.2,

438.3, 438.4, 438.5, 438.81, 438.82, 438.89,

438.9

Renal and metabolic complications

1) Acute renal failure requiring dialysis PSI *

2) Acute severe metabolic derangement PSI *

Deep vein thrombosis or pulmonary

embolism

PSI *

Infectious complications € PSI *

Pressure ulcer rate PSI *

Complications of anesthesia PSI *

* Post-procedural complications were identified by Patient Safety Indicators (PSIs) which have

been established by the Agency for Healthcare Research and Quality to monitor preventable

adverse events during hospitalization. These indicators are based on ICD-9-CM codes and

Medicare severity Diagnosis-Related Groups and each PSI has specific inclusion and exclusion

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

criteria. PSI individual measure technical specifications, Version 4.4, March 2012 was used to

identify & define preventable complications.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Supplementary figure 1: Utilization and mortality of BAV: Significant increase in BAV by

1.25 per million population per year, p<0.001 and significant decrease in mortality by

0.41% per year, p<0.0001.

Supplementary figure 2: CCI trend: Increasing mean CCI from 1.5 in 1998 to 2.7 in 2010.

P<0.001.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Supplementary figure 3: Vascular injury trend: Decrease in vascular injury rate, p<0.0001.

Supplementary figure 4: Length of stay: Significant decrease in length of stay(p=0.002).

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Supplementary figure 5: Disposition trend: Increasing trend towards discharge to other

facility, slope=0.43% per year, p=0.001.