- Email: [email protected]
Risk of Mortality Following Catheter Ablation of Atrial Fibrillation
JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY
VOL. 74, NO. 18, 2019
ª 2019 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION PUBLISHED BY ELSEVIER
Risk of Mortality Following Catheter Ablation of Atrial Fibrillation Edward P. Cheng, MD, PHD,a Christopher F. Liu, MD,a Ilhwan Yeo, MD,b Steven M. Markowitz, MD,a George Thomas, MD,a James E. Ip, MD,a Luke K. Kim, MD,a Bruce B. Lerman, MD,a Jim W. Cheung, MDa
ABSTRACT BACKGROUND Although procedure-related deaths during index admission following catheter ablation of AF have been reported to be low, adverse outcomes can occur after discharge. There are limited data on mortality early after AF ablation. OBJECTIVES This study aimed to identify rates, trends, and predictors of early mortality post-atrial fibrillation (AF) ablation. METHODS Using the all-payer, nationally representative Nationwide Readmissions Database, we evaluated 60,203 admissions of patients 18 years of age or older for AF ablation between 2010 and 2015. Early mortality was defined as death during initial admission or 30-day readmission. Based on International Classification of Diseases–9th Revision, Clinical Modification codes, we identified comorbidities, procedural complications, and causes of readmission following AF ablation. Multivariable logistic regression was performed to assess predictors of early mortality. RESULTS Early mortality following AF ablation occurred in 0.46% cases, with 54.3% of deaths occurring during readmission. From 2010 to 2015, quarterly rates of early mortality post-ablation increased from 0.25% to 1.35% (p < 0.001). Median time from ablation to death was 11.6 (interquartile range [IQR]: 4.2 to 22.7) days. After adjustment for age and comorbidities, procedural complications (adjusted odds ratio [aOR]: 4.06; p < 0.001), congestive heart failure (CHF) (aOR: 2.20; p ¼ 0.011) and low AF ablation hospital volume (aOR: 2.35; p ¼ 0.003) were associated with early mortality. Complications due to cardiac perforation (aOR: 2.98; p ¼ 0.007), other cardiac (aOR: 12.8; p < 0.001), and neurologic etiologies (aOR: 8.72; p < 0.001) were also associated with early mortality. CONCLUSIONS In a nationally representative cohort, early mortality following AF ablation affected nearly 1 in 200 patients, with the majority of deaths occurring during 30-day readmission. Procedural complications, congestive heart failure, and low hospital AF ablation volume were predictors of early mortality. Prompt management of post-procedure complications and CHF may be critical for reducing mortality rates following AF ablation. (J Am Coll Cardiol 2019;74:2254–64) © 2019 by the American College of Cardiology Foundation.
C
atheter ablation has been established as an
Furthermore, in selected patients with systolic heart
important treatment for symptomatic atrial
failure, AF ablation has been shown to be superior
fibrillation
improvement
in
(AF),
leading
to
significant
to medical therapy for reduction in all-cause mortal-
patient
quality
of
life
ity and hospitalization for congestive heart failure
(1–3).
From the aWeill Cornell Cardiovascular Outcomes Research Group (CORG), Department of Medicine, Division of Cardiology, Weill Cornell Medicine–New York Presbyterian Hospital, New York, New York; and the bDepartment of Medicine, Icahn School of Listen to this manuscript’s
Medicine at Mount Sinai, New York, New York. This work was supported by grants from the Michael Wolk Heart Foundation, the
audio summary by
New York Cardiac Center, Inc., and the New York Weill Cornell Medical Center Alumni Council. The Michael Wolk Heart Foun-
Editor-in-Chief
dation, the New York Cardiac Center, Inc., and the New York Weill Cornell Medical Center Alumni Council had no role in the
Dr. Valentin Fuster on
design and conduct of the study, the collection, analysis, and interpretation of the data, or the preparation, review, or approval of
JACC.org.
the manuscript. Dr. Cheung has received consulting fees from Abbott and Biotronik; and has received fellowship grant support from Abbott, Biosense Webster, Biotronik, Boston Scientific, and Medtronic. Dr. Markowitz has received consulting fees from Preventice Medical; and has received fees for serving on the Data Safety Monitoring Board from Boston Scientific. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. P. K. Shah, MD, served as Guest Editor-in-Chief for this paper. Manuscript received June 4, 2019; revised manuscript received August 2, 2019, accepted August 9, 2019.
ISSN 0735-1097/$36.00
https://doi.org/10.1016/j.jacc.2019.08.1036
JACC VOL. 74, NO. 18, 2019
Cheng et al.
NOVEMBER 5, 2019:2254–64
Early Mortality After AF Ablation
2255
(CHF) (4–6). As the overall volume of AF ablation pro-
we excluded patients who were discharged in
ABBREVIATIONS
cedures performed worldwide continues to grow, an
December from their index admission for AF
AND ACRONYMS
understanding of the real-world rates of serious com-
ablation to ensure 30-day follow-up after
plications after AF ablation is needed. Recent studies
discharge. In addition, because ICD-10-CM
have suggested an increasing trend in AF ablation-
codes were introduced for the NRD starting
related complication rates (7,8) despite advances in
with October 2015 discharges, all patients
catheter technology and operator experience.
whose index admission discharges occurred
At present, AF ablation-related in-hospital deaths
between September 2015 and December 2015
during index admission have been reported to be in
were excluded. Finally, patients younger
the range of 0% to 0.8% (7,9–15). However, these
than 18 years of age, or those who were
mortality rates are derived from studies arising from
missing mortality or length of stay data, were
single academic centers, regional databases, single-
excluded from the study.
payer databases, and national databases that are confined to outcomes during index admissions. Given that complications after AF ablation—such as esophageal injury, sepsis, and CHF—can occur after discharge, we hypothesized that a significant proportion of early mortality events would occur during readmission, rather than during the index admission for AF ablation. Using a nationally representative allpayer administrative database, we sought to provide real-world evidence on the rate, trends, and predictors of early mortality after AF ablation as defined by combined in-hospital mortality during either index admission or 30-day readmission following catheter ablation of AF. SEE PAGE 2265
METHODS
AF = atrial fibrillation AHRQ = Agency for Healthcare Research and Quality
CAD = coronary artery disease CCS = Clinical Classification Software
CHF = congestive heart failure HCUP = Healthcare Cost and Utilization Project
ICD-9-CM = International
CLINICAL VARIABLES. Patient and hospital
level variables were included in the baseline characteristics for analysis. Age, sex, median
Classification of Diseases–9th Revision-Clinical Modification
NRD = Nationwide Readmissions Database
household income quartiles, primary payer, location, and hospital size were extracted from NRD variables. Patient-level variables and cardiac diagnoses were defined by ICD-9-CM codes, Clinical Classification Software (CCS) codes, and AHRQ comorbidity
measures
based
on
the
Elixhauser
methods as defined in Online Table 2. Annual AF ablation hospital volume was determined on a yearto-year basis, using unique hospital identification numbers to calculate the total number of procedures performed by a particular institution for a given year. Hospitals were grouped into procedural volume tertiles, using annual procedural volume cutoffs based on 33rd and 67th percentiles of the total
DATA SOURCE. All data were obtained from the
number of patients in the dataset between 2010 and
United States Agency for Healthcare Research and
2014 (low-volume tertile: <21 AF ablations per year;
Quality (AHRQ), which administers the Healthcare
middle-volume tertile: 21 to 52 per year; high-
Cost and Utilization Project (HCUP) and Nationwide
volume tertile: >52 per year). For patients in the
Readmissions Database (NRD) from 2010 to 2015 for
2015 dataset, hospitals were grouped into volume
analysis. The NRD allows nationally representative
tertiles with cutoffs specific to 2015, given that only
readmission analyses. The NRD database is binned
ablations
per discharge data during 1 calendar year, and it uses
included in the dataset (low-volume tertile: <10 AF
a verified patient linkage number to track patient
ablations over 8 months; middle-volume tertile: 10
between
January
and
August
were
admissions to any hospital within the same state
to 23 over 8 months; high-volume tertile: >23
during the calendar year. Each admission record in
over 8 months).
the NRD contains diagnoses and procedures performed during hospitalization that are based on the International Classification of Diseases–9th RevisionClinical Modification (ICD-9-CM) codes.
STUDY ENDPOINTS. The primary endpoint of this
study is all-cause early mortality following AF ablation, which was defined as mortality occurring either at index AF ablation admission or at 30-day read-
STUDY POPULATION. All hospitalizations for cath-
mission following ablation (16). Other endpoints
eter ablation of AF were selected by identifying pa-
measured included the following procedural compli-
tients with primary ICD-9-CM diagnosis codes for AF
cations:
(427.31) and a primary ICD-9-CM procedure code for
cardiac complications, central nervous system com-
catheter ablation (37.34). All patients with secondary
plications, vascular complications, and pneumo-
ICD-9-CM codes for other arrhythmias or procedure
thorax (Online Table 2). Only the first readmission
codes for device implantation during the index
within 30 days after discharge from index admission
admission were excluded (Online Table 1). Because
for catheter ablation of AF was included in the study
the NRD is an annual database that is reset annually,
analysis. The primary causes of 30-day readmission
perforation/tamponade,
other
iatrogenic
2256
Cheng et al.
JACC VOL. 74, NO. 18, 2019
Early Mortality After AF Ablation
NOVEMBER 5, 2019:2254–64
F I G U R E 1 Distribution of Time From Ablation to Death Among Patients Who Died
Early After AF Ablation
identify predictors of early mortality, we created multivariable logistic regression models for the outcome of interest by including covariates that had univariate significance for the outcome (p < 0.10).
25
All tests were 2-sided with p values < 0.05 indicating statistical significance.
Number of Deaths
20
RESULTS STUDY
15
POPULATION
AND
RATES
OF
EARLY
MORTALITY. A total of 60,203 admission records
from the NRD of patients undergoing catheter abla-
10
tion of AF from January 2010 to August 2015 were included in the study analysis. The overall rate of early mortality after AF ablation was 0.46% (95%
5
confidence interval [CI]: 0.37% to 0.52%). Of the 276 patients who died early after AF ablation, 126 (45.7%) died during index admission, and 150 (54.3%) died
0
0
5 10 15 20 25 30 35 40 45 ≥50 Time from AF Ablation to Death (Days)
during 30-day readmission after AF ablation. The median time to death was 11.6 (IQR: 4.2 to 22.7; range: 0 to 77) days (Figure 1). Compared with survivors, patients who died early after AF ablation were older
Early mortality occurrences were grouped and counted according to time from ablation to death in days. There was a peak in deaths occurring on the day of ablation followed by a wide distribution of events over a 30-day window. AF ¼ atrial fibrillation.
and had a higher burden of comorbidities such as CHF, coronary artery disease, previous placement of pacemakers,
pulmonary
hypertension,
chronic
lung disease, chronic kidney disease, anemia, and coagulopathy (Table 1). Overall, compared with surby organ system (Online Table 3) and by cardiac cause
vivors, those who experienced early mortality had a
(Online Table 4) were identified by ICD-9-CM diag-
higher burden of comorbidities (40.1% vs. 14.4% with
nosis codes. The readmission causes were dichoto-
Elixhauser comorbidity scores $4; p < 0.001). Pa-
mized as cardiac and noncardiac causes. Noncardiac
tients who died early after AF ablation were less likely
causes of readmission were adjudicated into the
to have procedures performed at higher-volume
following
gastrointestinal,
centers and teaching hospitals, resided in lower–
neurological, vascular, renal, respiratory, endocrine,
household-income neighborhoods by ZIP code, and
hematologic, and rheumatologic.
had longer lengths of index hospitalizations.
categories:
infectious,
analyses were per-
The unadjusted rates of specific complications
formed using SAS software, version 9.4 (SAS Insti-
associated with catheter ablation of AF for the
tute,
STATISTICAL
ANALYSIS. All
weight
overall study population stratified by early mortal-
provided by the NRD was used for all analyses to
ity status are shown in Table 2. The overall rate of
obtain national estimates (17). All analyses accoun-
any procedural complication during index admis-
ted for hospital-level clustering of patients and
sion was 6.7%. Compared with patients who sur-
complex survey-sampling design. Both patient and
vived after AF ablation, patients who died early
hospital level variables were used for baseline char-
had higher rates of procedural complications (25.6
Cary,
North
Carolina).
Discharge
acteristic analysis. For descriptive analyses, we
vs. 6.6%; p < 0.001). Patients who experienced
compared baseline patient and hospital-level vari-
early mortality had more cardiac perforation, other
ables of catheter ablation of AF patients stratified by
cardiac complications, neurological complications,
the occurrence of early mortality. Categorical vari-
and pneumothorax.
ables are shown as frequencies, and continuous
TRENDS OF EARLY MORTALITY RATES AFTER AF
variables are presented as mean (standard error) or
ABLATION BETWEEN 2010 AND 2015. Between 2010
median (interquartile range [IQR]). Baseline charac-
and 2015, there was a significant increase in quarterly
teristics were compared by Rao-Scott chi-square test
rates of early mortality from 0.25% to 1.35% (p for
for
linear
trend <0.001) (Figure 2). In addition, during this
nonpara-
period, there was a significant increase in quarterly
metric test was used for continuous variables. To
rates of index procedural complications from 4.8% to
categorical
regression
or
variables.
Survey
specific
Mann-Whitney-Wilcoxon
JACC VOL. 74, NO. 18, 2019
Cheng et al.
NOVEMBER 5, 2019:2254–64
Early Mortality After AF Ablation
T A B L E 1 Comparison of Baseline Characteristics Based on Early Mortality After Catheter Ablation of Atrial Fibrillation
Early Mortality Overall
Yes
No
Ablations
60,203
276
59,927
Age, yrs
64.4 (0.13)
72.0 (1.1)
64.4 (0.13)
<65
28,027 (46.6)
70 (25.2)
27,957 (46.7)
65–74
21,389 (35.5)
81 (29.3)
21,308 (35.6)
$75
10,787 (17.9)
126 (45.5)
10,662 (17.7)
Female
22,852 (38.0)
118 (42.7)
22,734 (37.9)
0.354
History of CHF
10,796 (17.9)
148 (53.5)
10,649 (17.8)
<0.001
CAD
<0.001
p Value
<0.001 <0.001
Age group, yrs
16,170 (26.9)
133 (48.0)
16,038 (26.8)
Previous PCI
4,859 (8.1)
12 (4.5)
4,847 (8.1)
0.083
Previous CABG
3,391 (5.6)
37 (13.2)
3,354 (5.6)
0.009
Previous PPM
7,095 (11.8)
79 (28.7)
7,016 (11.7)
<0.001
Previous ICD
3,248 (5.4)
32 (11.5)
3,216 (5.4)
0.023
Hypertension
35,038 (58.2)
132 (47.7)
34,906 (58.2)
0.050
Diabetes mellitus
12,435 (20.7)
67 (24.4)
12,368 (20.6)
0.326
Hyperlipidemia
27,296 (45.3)
113 (40.9)
27,184 (45.4)
0.392
Obesity
9,912 (16.5)
37 (13.3)
9,876 (16.5)
0.330
History of stroke
3,714 (6.2)
15 (5.4)
3,699 (6.2)
0.723
Valvular disease
8,290 (13.8)
52 (18.7)
8,238 (13.8)
0.118
Peripheral vascular disease
2,094 (3.5)
15 (5.5)
2,078 (3.5)
0.200
Pulmonary hypertension
1,908 (3.2)
28 (10.2)
1,880 (3.1)
0.003
Chronic lung disease
9,190 (15.3)
79 (28.7)
9,111 (15.2)
0.001
Renal disease
4,526 (7.5)
66 (23.9)
4,460 (7.4)
<0.001
Anemia
3,596 (6.0)
65 (23.3)
3,531 (5.9)
<0.001
1,157 (1.9)
26 (9.5)
1,131 (1.9)
<0.001
Elixhauser comorbidity score >4
8,750 (14.5)
110 (40.1)
8,640 (14.4)
<0.001
Index procedural complication
4,028 (6.7)
71 (25.6)
3,957 (6.6)
<0.001
Elective procedure
41,027 (68.2)
129 (46.7)
40,898 (68.3)
<0.001
Teaching hospital
45,138 (75.0)
172 (62.3)
44,966 (75.0)
Coagulopathy
Low-volume tertile
21,642 (35.9)
172 (62.3)
21,470 (35.8)
Middle-volume tertile
19,915 (33.1)
71 (25.7)
19,844 (33.1)
High-volume tertile
18,647 (31.0)
33 (12.0)
18,614 (31.1)
11,835 (20.0)
66 (24.1)
11,769 (20.0) 13,897 (23.6)
Median household income First quartile (lowest)
0.015
Second quartile
13,987 (23.6)
91 (33.2)
Third quartile
15,663 (26.5)
70 (25.9)
15,592 (26.5)
Fourth quartile (highest)
17,723 (29.9)
46 (16.9)
17,677 (30.0)
31,951 (53.1)
205 (74.3)
31,746 (53.0)
Primary payer Medicare Medicaid Private including HMO Self-pay/no charge/other Hospital region, urban
0.001 1,856 (3.1)
*
1,848 (3.1)
24,426 (40.6)
48 (17.3)
24,379 (40.7)
1,950 (3.2)
16 (5.8)
1,934 (3.2)
59,531 (98.9)
271 (98.5)
59,259 (98.9)
Hospital bed size Small
2,222 (3.7)
11 (3.9)
2,211 (3.7)
10,939 (18.2)
59 (21.2)
10,880 (18.2)
Large
47,042 (78.1)
207 (74.8)
46,836 (78.2)
2.5 [0.04]
5.98 [0.78]
2.46 [0.04]
1.0 (1.0–2.4)
3.0 (1.0–7.5)
1.0 (1.0–2.4)
18,172 (30.2)
166 (60.1)
18,006 (30.0)
Prolonged index hospital stay, days $3
0.046 0.761
Medium Length of index hospital stay, days
0.004 <0.001
Hospital procedural volume
<0.001 <0.001
Values are n, mean (SE), n (%), or median (interquartile range). *10 or fewer cases (note: 10 or fewer observations cannot be reported in table cells per AHRQ guidelines on the use of HCUP data). CABG ¼ coronary artery bypass grafting; CAD ¼ coronary artery disease; CHF ¼ congestive heart failure; HMO ¼ health maintenance organization; ICD ¼ implantable cardioverter-defibrillator; PCI ¼ percutaneous coronary intervention; PPM ¼ permanent pacemaker.
2257
Cheng et al.
JACC VOL. 74, NO. 18, 2019
Early Mortality After AF Ablation
NOVEMBER 5, 2019:2254–64
12.3%; p for trend <0.001) (Figure 4). Although the
T A B L E 2 Comparison of AF–Ablation-Related Complication
Rates During Index Admission Among Patients Who Died Early
unadjusted odds ratios for early mortality increased
and Those Who Survived
between 2010 and 2015 (p ¼ 0.022), adjusted odds ratios—which accounted for age, comorbidities, and
Early Mortality Complications
Perforation Other cardiac
Overall
Yes
No
1,130 (1.9)
276 (5.3)
1,115 (1.9)
0.002 <0.001
p Value
procedural hospital volume—did not increase significantly during this period (p ¼ 0.241) (Online Figure 1). Similarly, whereas unadjusted odds ratios for index
690 (1.2)
42 (15.2)
648 (1.1)
Bleeding/vascular
2,588 (4.3)
18 (2.1)
2,570 (4.3)
0.187
complications increased between 2010 and 2015 (p for
Any complication
4,028 (6.7)
71 (25.6)
3,957 (6.6)
<0.001
trend <0.001), adjusted odds ratios did not increase (p for trend ¼ 0.332) (Online Figure 2).
Values are n (%).
PREDICTORS OF EARLY MORTALITY AFTER AF
AF ¼ atrial fibrillation.
ABLATION. The univariate and multivariable pre-
dictors for early mortality are shown in Table 3. After 7.4% (p for trend <0.001) (Figure 3). These trends
adjustment for age, comorbidities, and hospital
between 2010 and 2015 paralleled significant in-
characteristics, procedural complications during in-
creases in the mean age of patients at time of ablation
dex admission were independently associated with
(0.35 years of age per annum; p for trend <0.001),
early mortality (adjusted odds ratio [aOR]: 4.06;
increases in the prevalence of comorbidities such as
95% CI: 2.40 to 6.85; p < 0.001). Furthermore, CHF
CHF (12.3% to 27.3%; p for trend <0.001), coronary
(aOR: 2.20; 95% CI: 1.20 to 4.03; p ¼ 0.011]), anemia
artery disease (24.5% to 30.7%; p for trend <0.001),
(aOR: 1.83; 95% CI: 1.13 to 2.96; p ¼ 0.015), coagul-
for
opathy (aOR: 2.14; 95% CI: 1.04 to 4.39; p ¼ 0.046),
trend <0.001), and chronic kidney disease (6.0% to
and age (aOR: 1.04; 95% CI: 1.00 to 1.07; p ¼ 0.046)
chronic
lung
disease
(13.8%
to
19.2%;
p
F I G U R E 2 National Trends in Quarterly Rates of Early Mortality After AF Ablation Between 2010 and 2015
1.6 P-for-trend < 0.001 1.4 1.2 Percent Died (%)
2258
1 0.8 0.6 0.4 0.2 0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3
2010
2011
2012 Year
2013
2014
2015
There was an upward trend in quarterly rates of early mortality from 0.25% in the first quarter of 2010, to 1.35% in the third quarter of 2015. Dotted line indicates trend. AF ¼ atrial fibrillation; Q ¼ quarter.
JACC VOL. 74, NO. 18, 2019
Cheng et al.
NOVEMBER 5, 2019:2254–64
Early Mortality After AF Ablation
F I G U R E 3 National Trends in Procedural Complications Occurring During Index Admission for AF Ablation Between 2010 and 2015
12 P-for-trend < 0.001
Index Complications (%)
10
8
6
4
2
0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3
2010
2011
2012
2013
2014
2015
Year There was an upward trend in quarterly rates of procedural complications identified during index admission from 4.8% in the first quarter of 2010 to 7.4% in the third quarter of 2015. Dotted line indicates trend. Abbreviations as in Figure 2.
were independently predictive of early death after AF
mortality were cardiac (30%), infectious (30%), res-
ablation. Finally, patients undergoing AF ablation
piratory (17%), and neurological (12%). The 4 most
performed at low-volume centers (low-volume vs.
common individual primary readmission diagnoses
high-volume tertile), had significantly higher odds of
were septicemia (15%), CHF (15%), pneumonia (7.4%),
early mortality (aOR: 2.35; 95% CI: 1.33 to 4.15;
and stroke (5.9%). Of the cardiac readmissions, the
p ¼ 0.003). The association among specific AF abla-
most common readmission diagnoses were CHF
tion procedural complications with early mortality
(41%), cardiac arrest (18%), AF/flutter (14%), and
was examined with logistic regression analysis and is
pericarditis (9%). The most common procedures
summarized
performed
in
Table
4.
Cardiac
perforation
during
readmission
were
mechanical
(aOR: 2.98; 95% CI: 1.36 to 6.56; p ¼ 0.007), other
ventilation,
cardiac complications (aOR: 12.8; 95% CI: 6.86 to 23.8;
endoscopy, dialysis, chest tube, right-heart catheter-
p < 0.001), and neurological complications such as
ization, bronchoscopy, mechanical circulatory sup-
stroke/transient ischemic attack (TIA) (aOR: 8.72;
port, and thoracentesis (Online Table 5).
blood
transfusion,
echocardiogram,
95% CI: 2.71 to 28.1; p < 0.001) were independent predictors for early mortality. Vascular complications
DISCUSSION
and pneumothorax were not independently associated with early death after AF ablation.
In this analysis of the real-world, all-payer, nationally
CAUSES OF READMISSION OF PATIENTS WHO DIED
representative NRD, which included >60,000 cath-
EARLY AFTER AF ABLATION. The causes of read-
eter ablation procedures for AF between 2010 and
mission by organ system for the 150 patients who died
2015, we identified several key findings. First, be-
during 30-day readmission after AF ablation are
tween 2010 and 2015, the overall rate of early mor-
summarized in Figure 5. The leading causes of read-
tality after AF ablation was 0.46%, with the majority
mission among patients who experienced early
of these deaths occurring after discharge from the
2259
Cheng et al.
JACC VOL. 74, NO. 18, 2019
Early Mortality After AF Ablation
NOVEMBER 5, 2019:2254–64
F I G U R E 4 National Trends in Major Comorbidities in Patients Undergoing AF Ablation Between 2010 and 2015
35 30 25 Comorbidity (%)
2260
20 15 10 5 P-for-trend < 0.001 for all 0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3
2010
2011 CAD
2012 CHF
2013 2014 Year Lung Disease CKD
2015
Significant increases in the prevalence of CAD, CHF, chronic lung disease, and CKD were seen among patients undergoing AF ablation between 2010 and 2015. CAD ¼ coronary artery disease; CHF ¼ congestive heart failure; CKD ¼ chronic kidney disease; other abbreviations as in Figure 2.
ablation procedure (Central Illustration). Second,
found in our study exceeds the rates of procedural
there was a significant increase in the trend of quar-
death reported by other large studies examining AF
terly rates of early mortality and complications after
ablation outcomes. In a study based on an interna-
AF ablation. These trends paralleled a similar rise in
tional survey of 85 centers performing 20,825
comorbidity burden among patients undergoing AF
procedures between 2003 and 2006, the AF ablation-
ablation during that period. Notably, after adjustment
related early mortality rate was 0.15% (12). As these
for age, comorbidities, and hospital factors, the up-
data were based on voluntary responses to surveys,
ward trends in early mortality and complications
under-sampling of AF ablation centers or under-
were no longer significant. Procedural complications,
reporting of adverse events could have led to signif-
CHF, and low hospital AF ablation volume were sig-
icant underestimation of the true mortality rate.
nificant predictors of early mortality.
Comprehensive data from the Nationwide Inpatient
Catheter ablation has been established as an
Sample administrative database between 2000 and
important and effective treatment for patients with
2013 identified a death rate of 0.24% among 190,398
symptomatic AF or AF associated with CHF (1–6).
AF ablation procedures (8). This is consistent with the
However, AF ablation remains a largely elective pro-
index admission mortality rate of 0.25% noted in our
cedure, and an understanding of the risks of major
study. However, these data did not include deaths
complications and mortality associated with the pro-
that occurred after discharge from index admission
cedure is of paramount importance. Although recent
for AF ablation, which omits a significant number of
studies from highly experienced single academic
deaths that occur during early readmission.
centers have identified zero deaths related to AF
In addition to an early mortality rate of 0.46% after
ablation (13,14), our study provides real-world evi-
AF ablation, we found an alarming rise in the rate of
dence that the rates of early mortality after AF abla-
early deaths between 2010 and 2015. This trend may
tion are not insignificant and can occur in nearly 1 of
be explained by 2 major factors. First, we identified
200 procedures. The 0.46% rate of early mortality
significant increases in comorbidities such as CHF,
JACC VOL. 74, NO. 18, 2019
Cheng et al.
NOVEMBER 5, 2019:2254–64
Early Mortality After AF Ablation
T A B L E 3 Predictors of Early Mortality After Atrial Fibrillation Ablation
Univariate
Any index procedural complication
Multivariable
Unadjusted OR (95% CI)
p Value
Adjusted OR (95% CI)
p Value
4.88 (2.98–7.97)
<0.001
4.06 (2.40–6.85)
<0.001
CHF
5.31 (3.50–8.06)
<0.001
2.20 (1.20–4.03)
0.011
Anemia
4.86 (3.09–7.65)
<0.001
1.83 (1.13–2.96)
0.015
Coagulopathy
5.45 (2.83–10.49)
<0.001
2.14 (1.04–4.39)
0.039
Age
1.07 (1.05–1.10)*
<0.001
1.04 (1.00–1.07)*
0.046
Hospital procedural volume Low-volume tertile
4.51 (2.58–7.87)
<0.001
2.35 (1.33–4.15)
0.003
Middle-volume tertile
2.01 (1.12–3.62)
0.020
1.65 (0.91–2.98)
0.099
High-volume tertile
1.00 (reference)
CAD
1.00 (reference) <0.001
2.53 (1.67–3.83)
CKD
3.90 (2.48–6.13)
<0.001
Previous PPM
3.04 (1.90–4.88)
<0.001
Nonelective procedure
2.46 (1.61–3.77)
<0.001
Length of stay $3 days
3.50 (2.23–5.51)
<0.001 0.001
Chronic pulmonary disease
2.24 (1.38–3.65)
Pulmonary hypertension
3.51 (1.46–8.42)
0.005
Previous CABG
2.57 (1.23–5.36)
0.012
Previous ICD
2.29 (1.10–4.80)
0.028
Urban
1.39 (1.00–1.93)
0.048
HTN
0.66 (0.43–1.00)
0.052
Previous PCI
0.53 (0.26–1.10)
0.088
Quartile 1 (lowest)
2.14 (1.28–3.61)
0.186
Quartile 2
2.51 (1.34–4.68)
0.079
Quartile 3
1.74 (1.03–2.95)
0.974
Quartile 4 (highest)
1.00 (reference)
Median household income
Insurance Medicare
1.00 (reference)
Medicaid
0.61 (0.22–1.71)
0.186
Private including HMO
0.30 (0.18–0.50)
0.003
Self-pay/no charge/other
1.28 (0.30–5.46)
0.284
*Odds ratios associated with a 1-year increment in age. CI ¼ confidence interval; CKD ¼ chronic kidney disease; HTN ¼ hypertension; OR ¼ odds ratio; other abbreviations as in Table 1.
coronary artery disease, lung disease, and chronic
institutions performing these procedures has also
kidney disease among patients undergoing AF abla-
likely increased. Consequently, a larger proportion of
tion. Therefore, increasing comorbidity burden likely
patients may be having procedures performed at
explained—at least in part—the rise in quarterly pro-
lower-volume centers, which has been shown to be
cedural complication rates of from 4.8% to 7.8% be-
associated with higher complication rates (8,9).
tween 2010 and 2015 seen in our study. When we
Several randomized clinical trials have shown that
adjusted for age, comorbidities, and hospital factors,
catheter ablation is superior to medical therapy for
the upward trend in complications rates was no
the management of selected patients with CHF and
longer evident. The adverse impact of comorbidities
AF, with respect to left ventricular remodeling, heart
on complication rate trends have been shown with
failure
treatment
(18),
(4,5,20). We identified a highly significant increase in
catheter-ablation procedures (8), and AF ablation
the prevalence of CHF in our study cohort, rising from
(19). It is notable that, despite advances in catheter-
12.3% to 27.3% between 2010 and 2015. CHF was a
ablation technology, significant decreases in compli-
particularly strong independent predictor of early
cations, such as cardiac perforation, have not yet
mortality after AF ablation, as it was associated with a
been seen in analyses performed of more contempo-
>2-fold increased risk and was the leading cause of
rary cohorts (7). Second, as AF ablation volume has
cardiac readmissions among patients who experi-
increased
enced early mortality. To reduce adverse outcomes
of
in
acute
recent
myocardial
years
(8),
infarction
the
number
of
hospitalization,
and
all-cause
mortality
2261
2262
Cheng et al.
JACC VOL. 74, NO. 18, 2019
Early Mortality After AF Ablation
NOVEMBER 5, 2019:2254–64
T A B L E 4 Unadjusted and Adjusted Odds Ratios for Specific Procedural Complications
information provided by the NRD, we were unable to
During Index Admission as a Predictor for Early Mortality (Absence of Early Mortality
clearly define the primary source of infection that led
as Reference)
to sepsis in those cases. Atrioesophageal (AE) fistula
Complications
is a severe complication after AF ablation that can
Unadjusted OR (95% CI)
p Value
Adjusted OR (95% CI)
p Value
16.30 (8.16–32.70)
<0.0001
12.80 (6.86–23.80)
<0.001
9.52 (3.35–27.10)
<0.0001
8.72 (2.71–28.10)
<0.001
2.97 (1.42–6.21)
0.0038
2.98 (1.36–6.56)
0.007
tunately, there is no dedicated ICD-9-CM code for AE
6.93 (1.56–30.66)
0.0108
3.74 (0.71–19.8)
0.121
fistula. We did not identify any patients who died
Bleeding/vascular
1.55 (0.80–3.01)
0.1910
1.18 (0.60–2.31)
0.627
early who had undergone esophageal surgery. It
Any complication
4.88 (2.98–7.97)
<0.0001
4.06 (2.40–6.85)
<0.001
should be noted that it was possible that the diagnosis
*Other cardiac complications were defined as other iatrogenic cardiac complication, cardiogenic shock and cardiac arrest as per Online Table 2.
were adjudicated as having sepsis as the primary
Other cardiac* Neurologic Cardiac perforation Pneumothorax
Abbreviations as in Table 3.
present weeks after the procedure and can be manifest with fever and other signs of sepsis (21). Unfor-
of AE fistula was not recognized in some patients who readmission diagnosis. Regardless, nosocomial infections, such as pneumonia and urinary tract infection, can still be significant causes of mortality in
related to CHF, a team-based approach among elec-
patients who have had AF ablation, especially in
trophysiologists, general cardiologists, and heart
those with significant comorbidities. Although car-
failure specialists should be implemented for patients
diac electrophysiologists are trained to pay close
with CHF undergoing catheter ablation. It should be
attention to prevention of vascular injuries with
emphasized that the benefits of AF ablation for pa-
ultrasound-guided venous access and prompt treat-
tients with CHF shown in recent clinical trials have
ment of pericardial effusion with emergent peri-
largely involved high-volume academic centers,
cardiocentesis, operators also need to recognize that
whose outcomes may not be replicated by lower-
noncardiac complications, such as infection, can lead
volume operators and centers in the real world.
to major adverse events. Close follow-up after patient
Therefore, as more patients with CHF undergo AF
discharge may be essential for lowering current rates
ablation, the rates of early mortality after the pro-
of early mortality after AF ablation.
cedure may continue to rise on a national level.
STUDY LIMITATIONS. First, this is a retrospective
By examining deaths occurring during 30-day
study based on an administrative database from the
readmission after AF ablation, we were able to
NRD. Therefore, we were limited by the accuracy of
examine the causes of readmission that were associ-
ICD-9-CM codes, in which miscoding, overcoding,
ated with early mortality. Notably, the leading pri-
and missing data can occur and can compromise the
mary readmission diagnosis was septicemia, which
quality of estimates. However, HCUP quality-control
involved 15% of the patients who died during read-
measures are routinely performed to confirm data
mission.
validity and reliability (22). Second, any sudden
Given
the
limitations
of
the
coding
deaths occurring outside the hospital before readmission would not be included in this analysis. F I G U R E 5 Primary Readmission Diagnoses Categorized by Organ System
Among Patients Who Died During 30-Day Readmission Following AF Ablation
Therefore, the early mortality rate after AF ablation reported in this study represents an underestimation of the true rate. Third, clinical variables, such as left ventricular ejection fraction, left atrial volume,
3% 3% 12% 30% 5%
Cardiac
medication, and body mass index, were not available.
Infectious
In addition, procedural details—such as use of general
Respiratory
anesthesia, type of catheter ablation energy used
Renal Neurological
17% 30%
(e.g., radiofrequency vs. cryoenergy), procedure duration, ablation targeting nonpulmonary vein triggers, and creation of linear lesions—were not included
Gastrointestinal
in the NRD. Therefore, we were unable to explore the
Rheumatological
impact of patient cardiac substrate factors or catheter ablation
technique
on
endpoints.
Furthermore,
because the NRD cannot track individual patients Cardiac, infectious, and respiratory causes accounted for the majority of
across calendar years, we were unable to compare
primary readmission diagnoses among patients who died early after AF
outcomes of patients undergoing first-time ablation
ablation. AF ¼ atrial fibrillation.
versus repeat ablation. Fourth, although the NRD is designed to approximate the national distribution of
JACC VOL. 74, NO. 18, 2019
Cheng et al.
NOVEMBER 5, 2019:2254–64
Early Mortality After AF Ablation
C ENTR AL I LL U STRA T I O N Early Mortality After Atrial Fibrillation Ablation in the United States 2010 to 2015
Cheng, E.P. et al. J Am Coll Cardiol. 2019;74(18):2254–64.
The current study analyzed 60,203 admissions for atrial fibrillation (AF) ablation between 2010 and 2015 in the Nationwide Readmissions Database. The overall rate of early mortality after AF (atrial fibrillation) ablation was 0.46%. An upward trend in early mortality, complications, and baseline patient comorbidities was seen during this period. Procedural complications, congestive heart failure, and low AF ablation hospital volume were strongly associated with early mortality. aOR ¼ adjusted odds ratios.
2263
2264
Cheng et al.
JACC VOL. 74, NO. 18, 2019
Early Mortality After AF Ablation
NOVEMBER 5, 2019:2254–64
hospital characteristics, it is derived from a 50% sample of all U.S. hospitals, which may introduce
ADDRESS
over- or under-representation of certain hospital
Cheung, Division of Cardiology, Weill Cornell Medical
types. Also, the NRD only includes data from 22 states
College, 520 East 70th Street, Starr 4, New York, New
in the United States, which may limit generalizability
York
to the entire population. Finally, patients who are
Twitter: @DrJCheungEP.
admitted in one state would not be tracked when
FOR
10021.
CORRESPONDENCE:
E-mail:
Dr. Jim W.
[email protected].
PERSPECTIVES
readmitted in another state.
CONCLUSIONS
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: In a national practice-based sample,
In a contemporary, nationally representative real-
early mortality among patients undergoing catheter-
world cohort, the early mortality rate after AF abla-
based ablation for AF between 2010 and 2015 in the
tion was 0.46%, with the majority of deaths occurring
United States was 0.46%. Heart failure, procedural
during 30-day readmission after discharge from
complications, and low hospital ablation volume were
initial admission for AF ablation. Index procedural
associated with a higher risk of early mortality after
complications and CHF were significant independent
AF ablation.
predictors of early mortality. Sepsis and CHF were the leading primary causes of readmission associated with mortality. Implementation of strategies to reduce
procedural
complications,
optimize
CHF
management, and reduce nosocomial infections may
TRANSLATIONAL OUTLOOK: Protocols that optimize management of heart failure and minimize procedural complications warrant further study to reduce the impact of these factors on the outcomes of AF ablation.
help reduce early mortality after AF ablation.
REFERENCES 1. Kirchhof P, Benussi S, Kotecha D, et al. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur Heart J 2016;37:2893–962.
9. Deshmukh A, Patel NJ, Pant S, et al. In-hospital complications associated with catheter ablation of atrial fibrillation in the United States between 2000 and 2010. Circulation 2013;128:2104–12.
17. Healthcare Cost and Utilization Project (HCUP). NRD Overview. http://www.hcup-us.ahrq.gov/ nrdoverview.jsp. Available at: http://www.hcup-us. ahrq.gov/nrdoverview.jsp. Accessed July 18, 2018.
2. January CT, Wann LS, Alpert JS, et al. 2014 AHA/ACC/HRS Guideline for the management of
10. Srivatsa UN, Danielsen B, Amsterdam EA, et al. CAABL-AF (California Study of Ablation for Atrial Fibrillation). Circ Arrhythm Electrophysiol 2018;11: e005739.
18. Schmidt M, Jacobsen JB, Lash TL, Bøtker HE, Sørensen HT. 25 year trends in first time hospital-
patients with atrial fibrillation: executive summary. J Am Coll Cardiol 2014;64:2246–80. 3. Mark DB, Anstrom KJ, Sheng S, et al. Effect of catheter ablation vs medical therapy on quality of life among patients with atrial fibrillation: the CABANA randomized clinical trial. JAMA 2019;321: 1275–85.
11. Cappato R, Calkins H, Chen S-A, et al. Prevalence and causes of fatal outcome in catheter ablation of atrial fibrillation. J Am Coll Cardiol 2009;53:1798–803.
4. Marrouche NF, Brachmann J, Andresen D, et al. Catheter ablation for atrial fibrillation with heart
12. Cappato R, Calkins H, Chen S-A, et al. Updated worldwide survey on the methods, efficacy, and safety of catheter ablation for human atrial fibrillation. Circ Arrhythm Electrophysiol 2010;3:32–8.
failure. N Engl J Med 2018;378:417–27. 5. Prabhu S, Taylor AJ, Costello BT, et al. Catheter ablation versus medical rate control in atrial fibrillation and systolic dysfunction. J Am Coll Cardiol 2017;70:1949–61. 6. January CT, Wann LS, Calkins H, et al. 2019 AHA/ACC/HRS focused update of the 2014 AHA/ ACC/HRS guideline for the management of patients with atrial fibrillation. J Am Coll Cardiol 2019;74:104–32. 7. Tripathi B, Arora S, Kumar V, et al. Temporal trends of in-hospital complications associated with catheter ablation of atrial fibrillation in the United States: an update from Nationwide Inpatient Sample database (2011–2014). J Cardiovasc Electrophysiol 2018;29:715–24. 8. Hosseini SM, Rozen G, Saleh A, et al. Catheter ablation for cardiac arrhythmias: utilization and inhospital complications. JACC Clin Electrophysiol 2017;3:1240–8.
13. Abdur RK, Wazni OM, Barakat AF, et al. Lifethreatening complications of atrial fibrillation ablation. JACC Clin Electrophysiol 2019;5:284–91. 14. Voskoboinik A, Sparks PB, Morton JB, et al. Low rates of major complications for radiofrequency ablation of atrial fibrillation maintained over 14 Years: a single centre experience of 2750 consecutive cases. Heart Lung Circ 2018;27: 976–83. 15. Piccini JP, Sinner MF, Greiner MA, et al. Outcomes of Medicare beneficiaries undergoing catheter ablation for atrial fibrillation. Circulation 2012;126:2200–7. 16. Agency for Healthcare Research and Quality. Introduction to the HCUP Nationwide Readmissions Database (NRD). Available at: https://www.hcupus.ahrq.gov/db/nation/nrd/Introduction_NRD_20102015.pdf. Accessed July 18, 2018.
isation for acute myocardial infarction, subsequent short and long term mortality, and the prognostic impact of sex and comorbidity: a Danish nationwide cohort study. BMJ 2012;344:e356. 19. Kim LK, Yeo I, Cheung J, Feldman DN, Swaminathan RV, Kamel H. Impact of evolving comorbidities on outcomes after catheter ablation of atrial fibrillation in the United States between 2003 and 2013. JACC Clin Electrophysiol 2018;4:704–6. 20. Di Biase L, Mohanty P, Mohanty S, et al. Ablation versus amiodarone for treatment of persistent atrial fibrillation in patients with congestive heart failure and an implanted device. Circulation 2016;133:1637–44. 21. Han H-C, Ha FJ, Sanders P, et al. Atrioesophageal fistula. Circ Arrhythm Electrophysiol 2017;10:e005579. 22. Barrett ML, Ross DN. HCUP Quality Control Procedures Deliverable #1707.05. Available at: https://www.hcup-us.ahrq.gov/db/quality.pdf. Accessed February 20, 2017.
KEY WORDS atrial fibrillation, catheter ablation, early mortality, outcomes
A PPE NDI X For supplemental tables and figures, please see the online version of this paper.
VOL. 74, NO. 18, 2019
ª 2019 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION PUBLISHED BY ELSEVIER
Risk of Mortality Following Catheter Ablation of Atrial Fibrillation Edward P. Cheng, MD, PHD,a Christopher F. Liu, MD,a Ilhwan Yeo, MD,b Steven M. Markowitz, MD,a George Thomas, MD,a James E. Ip, MD,a Luke K. Kim, MD,a Bruce B. Lerman, MD,a Jim W. Cheung, MDa
ABSTRACT BACKGROUND Although procedure-related deaths during index admission following catheter ablation of AF have been reported to be low, adverse outcomes can occur after discharge. There are limited data on mortality early after AF ablation. OBJECTIVES This study aimed to identify rates, trends, and predictors of early mortality post-atrial fibrillation (AF) ablation. METHODS Using the all-payer, nationally representative Nationwide Readmissions Database, we evaluated 60,203 admissions of patients 18 years of age or older for AF ablation between 2010 and 2015. Early mortality was defined as death during initial admission or 30-day readmission. Based on International Classification of Diseases–9th Revision, Clinical Modification codes, we identified comorbidities, procedural complications, and causes of readmission following AF ablation. Multivariable logistic regression was performed to assess predictors of early mortality. RESULTS Early mortality following AF ablation occurred in 0.46% cases, with 54.3% of deaths occurring during readmission. From 2010 to 2015, quarterly rates of early mortality post-ablation increased from 0.25% to 1.35% (p < 0.001). Median time from ablation to death was 11.6 (interquartile range [IQR]: 4.2 to 22.7) days. After adjustment for age and comorbidities, procedural complications (adjusted odds ratio [aOR]: 4.06; p < 0.001), congestive heart failure (CHF) (aOR: 2.20; p ¼ 0.011) and low AF ablation hospital volume (aOR: 2.35; p ¼ 0.003) were associated with early mortality. Complications due to cardiac perforation (aOR: 2.98; p ¼ 0.007), other cardiac (aOR: 12.8; p < 0.001), and neurologic etiologies (aOR: 8.72; p < 0.001) were also associated with early mortality. CONCLUSIONS In a nationally representative cohort, early mortality following AF ablation affected nearly 1 in 200 patients, with the majority of deaths occurring during 30-day readmission. Procedural complications, congestive heart failure, and low hospital AF ablation volume were predictors of early mortality. Prompt management of post-procedure complications and CHF may be critical for reducing mortality rates following AF ablation. (J Am Coll Cardiol 2019;74:2254–64) © 2019 by the American College of Cardiology Foundation.
C
atheter ablation has been established as an
Furthermore, in selected patients with systolic heart
important treatment for symptomatic atrial
failure, AF ablation has been shown to be superior
fibrillation
improvement
in
(AF),
leading
to
significant
to medical therapy for reduction in all-cause mortal-
patient
quality
of
life
ity and hospitalization for congestive heart failure
(1–3).
From the aWeill Cornell Cardiovascular Outcomes Research Group (CORG), Department of Medicine, Division of Cardiology, Weill Cornell Medicine–New York Presbyterian Hospital, New York, New York; and the bDepartment of Medicine, Icahn School of Listen to this manuscript’s
Medicine at Mount Sinai, New York, New York. This work was supported by grants from the Michael Wolk Heart Foundation, the
audio summary by
New York Cardiac Center, Inc., and the New York Weill Cornell Medical Center Alumni Council. The Michael Wolk Heart Foun-
Editor-in-Chief
dation, the New York Cardiac Center, Inc., and the New York Weill Cornell Medical Center Alumni Council had no role in the
Dr. Valentin Fuster on
design and conduct of the study, the collection, analysis, and interpretation of the data, or the preparation, review, or approval of
JACC.org.
the manuscript. Dr. Cheung has received consulting fees from Abbott and Biotronik; and has received fellowship grant support from Abbott, Biosense Webster, Biotronik, Boston Scientific, and Medtronic. Dr. Markowitz has received consulting fees from Preventice Medical; and has received fees for serving on the Data Safety Monitoring Board from Boston Scientific. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. P. K. Shah, MD, served as Guest Editor-in-Chief for this paper. Manuscript received June 4, 2019; revised manuscript received August 2, 2019, accepted August 9, 2019.
ISSN 0735-1097/$36.00
https://doi.org/10.1016/j.jacc.2019.08.1036
JACC VOL. 74, NO. 18, 2019
Cheng et al.
NOVEMBER 5, 2019:2254–64
Early Mortality After AF Ablation
2255
(CHF) (4–6). As the overall volume of AF ablation pro-
we excluded patients who were discharged in
ABBREVIATIONS
cedures performed worldwide continues to grow, an
December from their index admission for AF
AND ACRONYMS
understanding of the real-world rates of serious com-
ablation to ensure 30-day follow-up after
plications after AF ablation is needed. Recent studies
discharge. In addition, because ICD-10-CM
have suggested an increasing trend in AF ablation-
codes were introduced for the NRD starting
related complication rates (7,8) despite advances in
with October 2015 discharges, all patients
catheter technology and operator experience.
whose index admission discharges occurred
At present, AF ablation-related in-hospital deaths
between September 2015 and December 2015
during index admission have been reported to be in
were excluded. Finally, patients younger
the range of 0% to 0.8% (7,9–15). However, these
than 18 years of age, or those who were
mortality rates are derived from studies arising from
missing mortality or length of stay data, were
single academic centers, regional databases, single-
excluded from the study.
payer databases, and national databases that are confined to outcomes during index admissions. Given that complications after AF ablation—such as esophageal injury, sepsis, and CHF—can occur after discharge, we hypothesized that a significant proportion of early mortality events would occur during readmission, rather than during the index admission for AF ablation. Using a nationally representative allpayer administrative database, we sought to provide real-world evidence on the rate, trends, and predictors of early mortality after AF ablation as defined by combined in-hospital mortality during either index admission or 30-day readmission following catheter ablation of AF. SEE PAGE 2265
METHODS
AF = atrial fibrillation AHRQ = Agency for Healthcare Research and Quality
CAD = coronary artery disease CCS = Clinical Classification Software
CHF = congestive heart failure HCUP = Healthcare Cost and Utilization Project
ICD-9-CM = International
CLINICAL VARIABLES. Patient and hospital
level variables were included in the baseline characteristics for analysis. Age, sex, median
Classification of Diseases–9th Revision-Clinical Modification
NRD = Nationwide Readmissions Database
household income quartiles, primary payer, location, and hospital size were extracted from NRD variables. Patient-level variables and cardiac diagnoses were defined by ICD-9-CM codes, Clinical Classification Software (CCS) codes, and AHRQ comorbidity
measures
based
on
the
Elixhauser
methods as defined in Online Table 2. Annual AF ablation hospital volume was determined on a yearto-year basis, using unique hospital identification numbers to calculate the total number of procedures performed by a particular institution for a given year. Hospitals were grouped into procedural volume tertiles, using annual procedural volume cutoffs based on 33rd and 67th percentiles of the total
DATA SOURCE. All data were obtained from the
number of patients in the dataset between 2010 and
United States Agency for Healthcare Research and
2014 (low-volume tertile: <21 AF ablations per year;
Quality (AHRQ), which administers the Healthcare
middle-volume tertile: 21 to 52 per year; high-
Cost and Utilization Project (HCUP) and Nationwide
volume tertile: >52 per year). For patients in the
Readmissions Database (NRD) from 2010 to 2015 for
2015 dataset, hospitals were grouped into volume
analysis. The NRD allows nationally representative
tertiles with cutoffs specific to 2015, given that only
readmission analyses. The NRD database is binned
ablations
per discharge data during 1 calendar year, and it uses
included in the dataset (low-volume tertile: <10 AF
a verified patient linkage number to track patient
ablations over 8 months; middle-volume tertile: 10
between
January
and
August
were
admissions to any hospital within the same state
to 23 over 8 months; high-volume tertile: >23
during the calendar year. Each admission record in
over 8 months).
the NRD contains diagnoses and procedures performed during hospitalization that are based on the International Classification of Diseases–9th RevisionClinical Modification (ICD-9-CM) codes.
STUDY ENDPOINTS. The primary endpoint of this
study is all-cause early mortality following AF ablation, which was defined as mortality occurring either at index AF ablation admission or at 30-day read-
STUDY POPULATION. All hospitalizations for cath-
mission following ablation (16). Other endpoints
eter ablation of AF were selected by identifying pa-
measured included the following procedural compli-
tients with primary ICD-9-CM diagnosis codes for AF
cations:
(427.31) and a primary ICD-9-CM procedure code for
cardiac complications, central nervous system com-
catheter ablation (37.34). All patients with secondary
plications, vascular complications, and pneumo-
ICD-9-CM codes for other arrhythmias or procedure
thorax (Online Table 2). Only the first readmission
codes for device implantation during the index
within 30 days after discharge from index admission
admission were excluded (Online Table 1). Because
for catheter ablation of AF was included in the study
the NRD is an annual database that is reset annually,
analysis. The primary causes of 30-day readmission
perforation/tamponade,
other
iatrogenic
2256
Cheng et al.
JACC VOL. 74, NO. 18, 2019
Early Mortality After AF Ablation
NOVEMBER 5, 2019:2254–64
F I G U R E 1 Distribution of Time From Ablation to Death Among Patients Who Died
Early After AF Ablation
identify predictors of early mortality, we created multivariable logistic regression models for the outcome of interest by including covariates that had univariate significance for the outcome (p < 0.10).
25
All tests were 2-sided with p values < 0.05 indicating statistical significance.
Number of Deaths
20
RESULTS STUDY
15
POPULATION
AND
RATES
OF
EARLY
MORTALITY. A total of 60,203 admission records
from the NRD of patients undergoing catheter abla-
10
tion of AF from January 2010 to August 2015 were included in the study analysis. The overall rate of early mortality after AF ablation was 0.46% (95%
5
confidence interval [CI]: 0.37% to 0.52%). Of the 276 patients who died early after AF ablation, 126 (45.7%) died during index admission, and 150 (54.3%) died
0
0
5 10 15 20 25 30 35 40 45 ≥50 Time from AF Ablation to Death (Days)
during 30-day readmission after AF ablation. The median time to death was 11.6 (IQR: 4.2 to 22.7; range: 0 to 77) days (Figure 1). Compared with survivors, patients who died early after AF ablation were older
Early mortality occurrences were grouped and counted according to time from ablation to death in days. There was a peak in deaths occurring on the day of ablation followed by a wide distribution of events over a 30-day window. AF ¼ atrial fibrillation.
and had a higher burden of comorbidities such as CHF, coronary artery disease, previous placement of pacemakers,
pulmonary
hypertension,
chronic
lung disease, chronic kidney disease, anemia, and coagulopathy (Table 1). Overall, compared with surby organ system (Online Table 3) and by cardiac cause
vivors, those who experienced early mortality had a
(Online Table 4) were identified by ICD-9-CM diag-
higher burden of comorbidities (40.1% vs. 14.4% with
nosis codes. The readmission causes were dichoto-
Elixhauser comorbidity scores $4; p < 0.001). Pa-
mized as cardiac and noncardiac causes. Noncardiac
tients who died early after AF ablation were less likely
causes of readmission were adjudicated into the
to have procedures performed at higher-volume
following
gastrointestinal,
centers and teaching hospitals, resided in lower–
neurological, vascular, renal, respiratory, endocrine,
household-income neighborhoods by ZIP code, and
hematologic, and rheumatologic.
had longer lengths of index hospitalizations.
categories:
infectious,
analyses were per-
The unadjusted rates of specific complications
formed using SAS software, version 9.4 (SAS Insti-
associated with catheter ablation of AF for the
tute,
STATISTICAL
ANALYSIS. All
weight
overall study population stratified by early mortal-
provided by the NRD was used for all analyses to
ity status are shown in Table 2. The overall rate of
obtain national estimates (17). All analyses accoun-
any procedural complication during index admis-
ted for hospital-level clustering of patients and
sion was 6.7%. Compared with patients who sur-
complex survey-sampling design. Both patient and
vived after AF ablation, patients who died early
hospital level variables were used for baseline char-
had higher rates of procedural complications (25.6
Cary,
North
Carolina).
Discharge
acteristic analysis. For descriptive analyses, we
vs. 6.6%; p < 0.001). Patients who experienced
compared baseline patient and hospital-level vari-
early mortality had more cardiac perforation, other
ables of catheter ablation of AF patients stratified by
cardiac complications, neurological complications,
the occurrence of early mortality. Categorical vari-
and pneumothorax.
ables are shown as frequencies, and continuous
TRENDS OF EARLY MORTALITY RATES AFTER AF
variables are presented as mean (standard error) or
ABLATION BETWEEN 2010 AND 2015. Between 2010
median (interquartile range [IQR]). Baseline charac-
and 2015, there was a significant increase in quarterly
teristics were compared by Rao-Scott chi-square test
rates of early mortality from 0.25% to 1.35% (p for
for
linear
trend <0.001) (Figure 2). In addition, during this
nonpara-
period, there was a significant increase in quarterly
metric test was used for continuous variables. To
rates of index procedural complications from 4.8% to
categorical
regression
or
variables.
Survey
specific
Mann-Whitney-Wilcoxon
JACC VOL. 74, NO. 18, 2019
Cheng et al.
NOVEMBER 5, 2019:2254–64
Early Mortality After AF Ablation
T A B L E 1 Comparison of Baseline Characteristics Based on Early Mortality After Catheter Ablation of Atrial Fibrillation
Early Mortality Overall
Yes
No
Ablations
60,203
276
59,927
Age, yrs
64.4 (0.13)
72.0 (1.1)
64.4 (0.13)
<65
28,027 (46.6)
70 (25.2)
27,957 (46.7)
65–74
21,389 (35.5)
81 (29.3)
21,308 (35.6)
$75
10,787 (17.9)
126 (45.5)
10,662 (17.7)
Female
22,852 (38.0)
118 (42.7)
22,734 (37.9)
0.354
History of CHF
10,796 (17.9)
148 (53.5)
10,649 (17.8)
<0.001
CAD
<0.001
p Value
<0.001 <0.001
Age group, yrs
16,170 (26.9)
133 (48.0)
16,038 (26.8)
Previous PCI
4,859 (8.1)
12 (4.5)
4,847 (8.1)
0.083
Previous CABG
3,391 (5.6)
37 (13.2)
3,354 (5.6)
0.009
Previous PPM
7,095 (11.8)
79 (28.7)
7,016 (11.7)
<0.001
Previous ICD
3,248 (5.4)
32 (11.5)
3,216 (5.4)
0.023
Hypertension
35,038 (58.2)
132 (47.7)
34,906 (58.2)
0.050
Diabetes mellitus
12,435 (20.7)
67 (24.4)
12,368 (20.6)
0.326
Hyperlipidemia
27,296 (45.3)
113 (40.9)
27,184 (45.4)
0.392
Obesity
9,912 (16.5)
37 (13.3)
9,876 (16.5)
0.330
History of stroke
3,714 (6.2)
15 (5.4)
3,699 (6.2)
0.723
Valvular disease
8,290 (13.8)
52 (18.7)
8,238 (13.8)
0.118
Peripheral vascular disease
2,094 (3.5)
15 (5.5)
2,078 (3.5)
0.200
Pulmonary hypertension
1,908 (3.2)
28 (10.2)
1,880 (3.1)
0.003
Chronic lung disease
9,190 (15.3)
79 (28.7)
9,111 (15.2)
0.001
Renal disease
4,526 (7.5)
66 (23.9)
4,460 (7.4)
<0.001
Anemia
3,596 (6.0)
65 (23.3)
3,531 (5.9)
<0.001
1,157 (1.9)
26 (9.5)
1,131 (1.9)
<0.001
Elixhauser comorbidity score >4
8,750 (14.5)
110 (40.1)
8,640 (14.4)
<0.001
Index procedural complication
4,028 (6.7)
71 (25.6)
3,957 (6.6)
<0.001
Elective procedure
41,027 (68.2)
129 (46.7)
40,898 (68.3)
<0.001
Teaching hospital
45,138 (75.0)
172 (62.3)
44,966 (75.0)
Coagulopathy
Low-volume tertile
21,642 (35.9)
172 (62.3)
21,470 (35.8)
Middle-volume tertile
19,915 (33.1)
71 (25.7)
19,844 (33.1)
High-volume tertile
18,647 (31.0)
33 (12.0)
18,614 (31.1)
11,835 (20.0)
66 (24.1)
11,769 (20.0) 13,897 (23.6)
Median household income First quartile (lowest)
0.015
Second quartile
13,987 (23.6)
91 (33.2)
Third quartile
15,663 (26.5)
70 (25.9)
15,592 (26.5)
Fourth quartile (highest)
17,723 (29.9)
46 (16.9)
17,677 (30.0)
31,951 (53.1)
205 (74.3)
31,746 (53.0)
Primary payer Medicare Medicaid Private including HMO Self-pay/no charge/other Hospital region, urban
0.001 1,856 (3.1)
*
1,848 (3.1)
24,426 (40.6)
48 (17.3)
24,379 (40.7)
1,950 (3.2)
16 (5.8)
1,934 (3.2)
59,531 (98.9)
271 (98.5)
59,259 (98.9)
Hospital bed size Small
2,222 (3.7)
11 (3.9)
2,211 (3.7)
10,939 (18.2)
59 (21.2)
10,880 (18.2)
Large
47,042 (78.1)
207 (74.8)
46,836 (78.2)
2.5 [0.04]
5.98 [0.78]
2.46 [0.04]
1.0 (1.0–2.4)
3.0 (1.0–7.5)
1.0 (1.0–2.4)
18,172 (30.2)
166 (60.1)
18,006 (30.0)
Prolonged index hospital stay, days $3
0.046 0.761
Medium Length of index hospital stay, days
0.004 <0.001
Hospital procedural volume
<0.001 <0.001
Values are n, mean (SE), n (%), or median (interquartile range). *10 or fewer cases (note: 10 or fewer observations cannot be reported in table cells per AHRQ guidelines on the use of HCUP data). CABG ¼ coronary artery bypass grafting; CAD ¼ coronary artery disease; CHF ¼ congestive heart failure; HMO ¼ health maintenance organization; ICD ¼ implantable cardioverter-defibrillator; PCI ¼ percutaneous coronary intervention; PPM ¼ permanent pacemaker.
2257
Cheng et al.
JACC VOL. 74, NO. 18, 2019
Early Mortality After AF Ablation
NOVEMBER 5, 2019:2254–64
12.3%; p for trend <0.001) (Figure 4). Although the
T A B L E 2 Comparison of AF–Ablation-Related Complication
Rates During Index Admission Among Patients Who Died Early
unadjusted odds ratios for early mortality increased
and Those Who Survived
between 2010 and 2015 (p ¼ 0.022), adjusted odds ratios—which accounted for age, comorbidities, and
Early Mortality Complications
Perforation Other cardiac
Overall
Yes
No
1,130 (1.9)
276 (5.3)
1,115 (1.9)
0.002 <0.001
p Value
procedural hospital volume—did not increase significantly during this period (p ¼ 0.241) (Online Figure 1). Similarly, whereas unadjusted odds ratios for index
690 (1.2)
42 (15.2)
648 (1.1)
Bleeding/vascular
2,588 (4.3)
18 (2.1)
2,570 (4.3)
0.187
complications increased between 2010 and 2015 (p for
Any complication
4,028 (6.7)
71 (25.6)
3,957 (6.6)
<0.001
trend <0.001), adjusted odds ratios did not increase (p for trend ¼ 0.332) (Online Figure 2).
Values are n (%).
PREDICTORS OF EARLY MORTALITY AFTER AF
AF ¼ atrial fibrillation.
ABLATION. The univariate and multivariable pre-
dictors for early mortality are shown in Table 3. After 7.4% (p for trend <0.001) (Figure 3). These trends
adjustment for age, comorbidities, and hospital
between 2010 and 2015 paralleled significant in-
characteristics, procedural complications during in-
creases in the mean age of patients at time of ablation
dex admission were independently associated with
(0.35 years of age per annum; p for trend <0.001),
early mortality (adjusted odds ratio [aOR]: 4.06;
increases in the prevalence of comorbidities such as
95% CI: 2.40 to 6.85; p < 0.001). Furthermore, CHF
CHF (12.3% to 27.3%; p for trend <0.001), coronary
(aOR: 2.20; 95% CI: 1.20 to 4.03; p ¼ 0.011]), anemia
artery disease (24.5% to 30.7%; p for trend <0.001),
(aOR: 1.83; 95% CI: 1.13 to 2.96; p ¼ 0.015), coagul-
for
opathy (aOR: 2.14; 95% CI: 1.04 to 4.39; p ¼ 0.046),
trend <0.001), and chronic kidney disease (6.0% to
and age (aOR: 1.04; 95% CI: 1.00 to 1.07; p ¼ 0.046)
chronic
lung
disease
(13.8%
to
19.2%;
p
F I G U R E 2 National Trends in Quarterly Rates of Early Mortality After AF Ablation Between 2010 and 2015
1.6 P-for-trend < 0.001 1.4 1.2 Percent Died (%)
2258
1 0.8 0.6 0.4 0.2 0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3
2010
2011
2012 Year
2013
2014
2015
There was an upward trend in quarterly rates of early mortality from 0.25% in the first quarter of 2010, to 1.35% in the third quarter of 2015. Dotted line indicates trend. AF ¼ atrial fibrillation; Q ¼ quarter.
JACC VOL. 74, NO. 18, 2019
Cheng et al.
NOVEMBER 5, 2019:2254–64
Early Mortality After AF Ablation
F I G U R E 3 National Trends in Procedural Complications Occurring During Index Admission for AF Ablation Between 2010 and 2015
12 P-for-trend < 0.001
Index Complications (%)
10
8
6
4
2
0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3
2010
2011
2012
2013
2014
2015
Year There was an upward trend in quarterly rates of procedural complications identified during index admission from 4.8% in the first quarter of 2010 to 7.4% in the third quarter of 2015. Dotted line indicates trend. Abbreviations as in Figure 2.
were independently predictive of early death after AF
mortality were cardiac (30%), infectious (30%), res-
ablation. Finally, patients undergoing AF ablation
piratory (17%), and neurological (12%). The 4 most
performed at low-volume centers (low-volume vs.
common individual primary readmission diagnoses
high-volume tertile), had significantly higher odds of
were septicemia (15%), CHF (15%), pneumonia (7.4%),
early mortality (aOR: 2.35; 95% CI: 1.33 to 4.15;
and stroke (5.9%). Of the cardiac readmissions, the
p ¼ 0.003). The association among specific AF abla-
most common readmission diagnoses were CHF
tion procedural complications with early mortality
(41%), cardiac arrest (18%), AF/flutter (14%), and
was examined with logistic regression analysis and is
pericarditis (9%). The most common procedures
summarized
performed
in
Table
4.
Cardiac
perforation
during
readmission
were
mechanical
(aOR: 2.98; 95% CI: 1.36 to 6.56; p ¼ 0.007), other
ventilation,
cardiac complications (aOR: 12.8; 95% CI: 6.86 to 23.8;
endoscopy, dialysis, chest tube, right-heart catheter-
p < 0.001), and neurological complications such as
ization, bronchoscopy, mechanical circulatory sup-
stroke/transient ischemic attack (TIA) (aOR: 8.72;
port, and thoracentesis (Online Table 5).
blood
transfusion,
echocardiogram,
95% CI: 2.71 to 28.1; p < 0.001) were independent predictors for early mortality. Vascular complications
DISCUSSION
and pneumothorax were not independently associated with early death after AF ablation.
In this analysis of the real-world, all-payer, nationally
CAUSES OF READMISSION OF PATIENTS WHO DIED
representative NRD, which included >60,000 cath-
EARLY AFTER AF ABLATION. The causes of read-
eter ablation procedures for AF between 2010 and
mission by organ system for the 150 patients who died
2015, we identified several key findings. First, be-
during 30-day readmission after AF ablation are
tween 2010 and 2015, the overall rate of early mor-
summarized in Figure 5. The leading causes of read-
tality after AF ablation was 0.46%, with the majority
mission among patients who experienced early
of these deaths occurring after discharge from the
2259
Cheng et al.
JACC VOL. 74, NO. 18, 2019
Early Mortality After AF Ablation
NOVEMBER 5, 2019:2254–64
F I G U R E 4 National Trends in Major Comorbidities in Patients Undergoing AF Ablation Between 2010 and 2015
35 30 25 Comorbidity (%)
2260
20 15 10 5 P-for-trend < 0.001 for all 0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3
2010
2011 CAD
2012 CHF
2013 2014 Year Lung Disease CKD
2015
Significant increases in the prevalence of CAD, CHF, chronic lung disease, and CKD were seen among patients undergoing AF ablation between 2010 and 2015. CAD ¼ coronary artery disease; CHF ¼ congestive heart failure; CKD ¼ chronic kidney disease; other abbreviations as in Figure 2.
ablation procedure (Central Illustration). Second,
found in our study exceeds the rates of procedural
there was a significant increase in the trend of quar-
death reported by other large studies examining AF
terly rates of early mortality and complications after
ablation outcomes. In a study based on an interna-
AF ablation. These trends paralleled a similar rise in
tional survey of 85 centers performing 20,825
comorbidity burden among patients undergoing AF
procedures between 2003 and 2006, the AF ablation-
ablation during that period. Notably, after adjustment
related early mortality rate was 0.15% (12). As these
for age, comorbidities, and hospital factors, the up-
data were based on voluntary responses to surveys,
ward trends in early mortality and complications
under-sampling of AF ablation centers or under-
were no longer significant. Procedural complications,
reporting of adverse events could have led to signif-
CHF, and low hospital AF ablation volume were sig-
icant underestimation of the true mortality rate.
nificant predictors of early mortality.
Comprehensive data from the Nationwide Inpatient
Catheter ablation has been established as an
Sample administrative database between 2000 and
important and effective treatment for patients with
2013 identified a death rate of 0.24% among 190,398
symptomatic AF or AF associated with CHF (1–6).
AF ablation procedures (8). This is consistent with the
However, AF ablation remains a largely elective pro-
index admission mortality rate of 0.25% noted in our
cedure, and an understanding of the risks of major
study. However, these data did not include deaths
complications and mortality associated with the pro-
that occurred after discharge from index admission
cedure is of paramount importance. Although recent
for AF ablation, which omits a significant number of
studies from highly experienced single academic
deaths that occur during early readmission.
centers have identified zero deaths related to AF
In addition to an early mortality rate of 0.46% after
ablation (13,14), our study provides real-world evi-
AF ablation, we found an alarming rise in the rate of
dence that the rates of early mortality after AF abla-
early deaths between 2010 and 2015. This trend may
tion are not insignificant and can occur in nearly 1 of
be explained by 2 major factors. First, we identified
200 procedures. The 0.46% rate of early mortality
significant increases in comorbidities such as CHF,
JACC VOL. 74, NO. 18, 2019
Cheng et al.
NOVEMBER 5, 2019:2254–64
Early Mortality After AF Ablation
T A B L E 3 Predictors of Early Mortality After Atrial Fibrillation Ablation
Univariate
Any index procedural complication
Multivariable
Unadjusted OR (95% CI)
p Value
Adjusted OR (95% CI)
p Value
4.88 (2.98–7.97)
<0.001
4.06 (2.40–6.85)
<0.001
CHF
5.31 (3.50–8.06)
<0.001
2.20 (1.20–4.03)
0.011
Anemia
4.86 (3.09–7.65)
<0.001
1.83 (1.13–2.96)
0.015
Coagulopathy
5.45 (2.83–10.49)
<0.001
2.14 (1.04–4.39)
0.039
Age
1.07 (1.05–1.10)*
<0.001
1.04 (1.00–1.07)*
0.046
Hospital procedural volume Low-volume tertile
4.51 (2.58–7.87)
<0.001
2.35 (1.33–4.15)
0.003
Middle-volume tertile
2.01 (1.12–3.62)
0.020
1.65 (0.91–2.98)
0.099
High-volume tertile
1.00 (reference)
CAD
1.00 (reference) <0.001
2.53 (1.67–3.83)
CKD
3.90 (2.48–6.13)
<0.001
Previous PPM
3.04 (1.90–4.88)
<0.001
Nonelective procedure
2.46 (1.61–3.77)
<0.001
Length of stay $3 days
3.50 (2.23–5.51)
<0.001 0.001
Chronic pulmonary disease
2.24 (1.38–3.65)
Pulmonary hypertension
3.51 (1.46–8.42)
0.005
Previous CABG
2.57 (1.23–5.36)
0.012
Previous ICD
2.29 (1.10–4.80)
0.028
Urban
1.39 (1.00–1.93)
0.048
HTN
0.66 (0.43–1.00)
0.052
Previous PCI
0.53 (0.26–1.10)
0.088
Quartile 1 (lowest)
2.14 (1.28–3.61)
0.186
Quartile 2
2.51 (1.34–4.68)
0.079
Quartile 3
1.74 (1.03–2.95)
0.974
Quartile 4 (highest)
1.00 (reference)
Median household income
Insurance Medicare
1.00 (reference)
Medicaid
0.61 (0.22–1.71)
0.186
Private including HMO
0.30 (0.18–0.50)
0.003
Self-pay/no charge/other
1.28 (0.30–5.46)
0.284
*Odds ratios associated with a 1-year increment in age. CI ¼ confidence interval; CKD ¼ chronic kidney disease; HTN ¼ hypertension; OR ¼ odds ratio; other abbreviations as in Table 1.
coronary artery disease, lung disease, and chronic
institutions performing these procedures has also
kidney disease among patients undergoing AF abla-
likely increased. Consequently, a larger proportion of
tion. Therefore, increasing comorbidity burden likely
patients may be having procedures performed at
explained—at least in part—the rise in quarterly pro-
lower-volume centers, which has been shown to be
cedural complication rates of from 4.8% to 7.8% be-
associated with higher complication rates (8,9).
tween 2010 and 2015 seen in our study. When we
Several randomized clinical trials have shown that
adjusted for age, comorbidities, and hospital factors,
catheter ablation is superior to medical therapy for
the upward trend in complications rates was no
the management of selected patients with CHF and
longer evident. The adverse impact of comorbidities
AF, with respect to left ventricular remodeling, heart
on complication rate trends have been shown with
failure
treatment
(18),
(4,5,20). We identified a highly significant increase in
catheter-ablation procedures (8), and AF ablation
the prevalence of CHF in our study cohort, rising from
(19). It is notable that, despite advances in catheter-
12.3% to 27.3% between 2010 and 2015. CHF was a
ablation technology, significant decreases in compli-
particularly strong independent predictor of early
cations, such as cardiac perforation, have not yet
mortality after AF ablation, as it was associated with a
been seen in analyses performed of more contempo-
>2-fold increased risk and was the leading cause of
rary cohorts (7). Second, as AF ablation volume has
cardiac readmissions among patients who experi-
increased
enced early mortality. To reduce adverse outcomes
of
in
acute
recent
myocardial
years
(8),
infarction
the
number
of
hospitalization,
and
all-cause
mortality
2261
2262
Cheng et al.
JACC VOL. 74, NO. 18, 2019
Early Mortality After AF Ablation
NOVEMBER 5, 2019:2254–64
T A B L E 4 Unadjusted and Adjusted Odds Ratios for Specific Procedural Complications
information provided by the NRD, we were unable to
During Index Admission as a Predictor for Early Mortality (Absence of Early Mortality
clearly define the primary source of infection that led
as Reference)
to sepsis in those cases. Atrioesophageal (AE) fistula
Complications
is a severe complication after AF ablation that can
Unadjusted OR (95% CI)
p Value
Adjusted OR (95% CI)
p Value
16.30 (8.16–32.70)
<0.0001
12.80 (6.86–23.80)
<0.001
9.52 (3.35–27.10)
<0.0001
8.72 (2.71–28.10)
<0.001
2.97 (1.42–6.21)
0.0038
2.98 (1.36–6.56)
0.007
tunately, there is no dedicated ICD-9-CM code for AE
6.93 (1.56–30.66)
0.0108
3.74 (0.71–19.8)
0.121
fistula. We did not identify any patients who died
Bleeding/vascular
1.55 (0.80–3.01)
0.1910
1.18 (0.60–2.31)
0.627
early who had undergone esophageal surgery. It
Any complication
4.88 (2.98–7.97)
<0.0001
4.06 (2.40–6.85)
<0.001
should be noted that it was possible that the diagnosis
*Other cardiac complications were defined as other iatrogenic cardiac complication, cardiogenic shock and cardiac arrest as per Online Table 2.
were adjudicated as having sepsis as the primary
Other cardiac* Neurologic Cardiac perforation Pneumothorax
Abbreviations as in Table 3.
present weeks after the procedure and can be manifest with fever and other signs of sepsis (21). Unfor-
of AE fistula was not recognized in some patients who readmission diagnosis. Regardless, nosocomial infections, such as pneumonia and urinary tract infection, can still be significant causes of mortality in
related to CHF, a team-based approach among elec-
patients who have had AF ablation, especially in
trophysiologists, general cardiologists, and heart
those with significant comorbidities. Although car-
failure specialists should be implemented for patients
diac electrophysiologists are trained to pay close
with CHF undergoing catheter ablation. It should be
attention to prevention of vascular injuries with
emphasized that the benefits of AF ablation for pa-
ultrasound-guided venous access and prompt treat-
tients with CHF shown in recent clinical trials have
ment of pericardial effusion with emergent peri-
largely involved high-volume academic centers,
cardiocentesis, operators also need to recognize that
whose outcomes may not be replicated by lower-
noncardiac complications, such as infection, can lead
volume operators and centers in the real world.
to major adverse events. Close follow-up after patient
Therefore, as more patients with CHF undergo AF
discharge may be essential for lowering current rates
ablation, the rates of early mortality after the pro-
of early mortality after AF ablation.
cedure may continue to rise on a national level.
STUDY LIMITATIONS. First, this is a retrospective
By examining deaths occurring during 30-day
study based on an administrative database from the
readmission after AF ablation, we were able to
NRD. Therefore, we were limited by the accuracy of
examine the causes of readmission that were associ-
ICD-9-CM codes, in which miscoding, overcoding,
ated with early mortality. Notably, the leading pri-
and missing data can occur and can compromise the
mary readmission diagnosis was septicemia, which
quality of estimates. However, HCUP quality-control
involved 15% of the patients who died during read-
measures are routinely performed to confirm data
mission.
validity and reliability (22). Second, any sudden
Given
the
limitations
of
the
coding
deaths occurring outside the hospital before readmission would not be included in this analysis. F I G U R E 5 Primary Readmission Diagnoses Categorized by Organ System
Among Patients Who Died During 30-Day Readmission Following AF Ablation
Therefore, the early mortality rate after AF ablation reported in this study represents an underestimation of the true rate. Third, clinical variables, such as left ventricular ejection fraction, left atrial volume,
3% 3% 12% 30% 5%
Cardiac
medication, and body mass index, were not available.
Infectious
In addition, procedural details—such as use of general
Respiratory
anesthesia, type of catheter ablation energy used
Renal Neurological
17% 30%
(e.g., radiofrequency vs. cryoenergy), procedure duration, ablation targeting nonpulmonary vein triggers, and creation of linear lesions—were not included
Gastrointestinal
in the NRD. Therefore, we were unable to explore the
Rheumatological
impact of patient cardiac substrate factors or catheter ablation
technique
on
endpoints.
Furthermore,
because the NRD cannot track individual patients Cardiac, infectious, and respiratory causes accounted for the majority of
across calendar years, we were unable to compare
primary readmission diagnoses among patients who died early after AF
outcomes of patients undergoing first-time ablation
ablation. AF ¼ atrial fibrillation.
versus repeat ablation. Fourth, although the NRD is designed to approximate the national distribution of
JACC VOL. 74, NO. 18, 2019
Cheng et al.
NOVEMBER 5, 2019:2254–64
Early Mortality After AF Ablation
C ENTR AL I LL U STRA T I O N Early Mortality After Atrial Fibrillation Ablation in the United States 2010 to 2015
Cheng, E.P. et al. J Am Coll Cardiol. 2019;74(18):2254–64.
The current study analyzed 60,203 admissions for atrial fibrillation (AF) ablation between 2010 and 2015 in the Nationwide Readmissions Database. The overall rate of early mortality after AF (atrial fibrillation) ablation was 0.46%. An upward trend in early mortality, complications, and baseline patient comorbidities was seen during this period. Procedural complications, congestive heart failure, and low AF ablation hospital volume were strongly associated with early mortality. aOR ¼ adjusted odds ratios.
2263
2264
Cheng et al.
JACC VOL. 74, NO. 18, 2019
Early Mortality After AF Ablation
NOVEMBER 5, 2019:2254–64
hospital characteristics, it is derived from a 50% sample of all U.S. hospitals, which may introduce
ADDRESS
over- or under-representation of certain hospital
Cheung, Division of Cardiology, Weill Cornell Medical
types. Also, the NRD only includes data from 22 states
College, 520 East 70th Street, Starr 4, New York, New
in the United States, which may limit generalizability
York
to the entire population. Finally, patients who are
Twitter: @DrJCheungEP.
admitted in one state would not be tracked when
FOR
10021.
CORRESPONDENCE:
E-mail:
Dr. Jim W.
[email protected].
PERSPECTIVES
readmitted in another state.
CONCLUSIONS
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: In a national practice-based sample,
In a contemporary, nationally representative real-
early mortality among patients undergoing catheter-
world cohort, the early mortality rate after AF abla-
based ablation for AF between 2010 and 2015 in the
tion was 0.46%, with the majority of deaths occurring
United States was 0.46%. Heart failure, procedural
during 30-day readmission after discharge from
complications, and low hospital ablation volume were
initial admission for AF ablation. Index procedural
associated with a higher risk of early mortality after
complications and CHF were significant independent
AF ablation.
predictors of early mortality. Sepsis and CHF were the leading primary causes of readmission associated with mortality. Implementation of strategies to reduce
procedural
complications,
optimize
CHF
management, and reduce nosocomial infections may
TRANSLATIONAL OUTLOOK: Protocols that optimize management of heart failure and minimize procedural complications warrant further study to reduce the impact of these factors on the outcomes of AF ablation.
help reduce early mortality after AF ablation.
REFERENCES 1. Kirchhof P, Benussi S, Kotecha D, et al. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur Heart J 2016;37:2893–962.
9. Deshmukh A, Patel NJ, Pant S, et al. In-hospital complications associated with catheter ablation of atrial fibrillation in the United States between 2000 and 2010. Circulation 2013;128:2104–12.
17. Healthcare Cost and Utilization Project (HCUP). NRD Overview. http://www.hcup-us.ahrq.gov/ nrdoverview.jsp. Available at: http://www.hcup-us. ahrq.gov/nrdoverview.jsp. Accessed July 18, 2018.
2. January CT, Wann LS, Alpert JS, et al. 2014 AHA/ACC/HRS Guideline for the management of
10. Srivatsa UN, Danielsen B, Amsterdam EA, et al. CAABL-AF (California Study of Ablation for Atrial Fibrillation). Circ Arrhythm Electrophysiol 2018;11: e005739.
18. Schmidt M, Jacobsen JB, Lash TL, Bøtker HE, Sørensen HT. 25 year trends in first time hospital-
patients with atrial fibrillation: executive summary. J Am Coll Cardiol 2014;64:2246–80. 3. Mark DB, Anstrom KJ, Sheng S, et al. Effect of catheter ablation vs medical therapy on quality of life among patients with atrial fibrillation: the CABANA randomized clinical trial. JAMA 2019;321: 1275–85.
11. Cappato R, Calkins H, Chen S-A, et al. Prevalence and causes of fatal outcome in catheter ablation of atrial fibrillation. J Am Coll Cardiol 2009;53:1798–803.
4. Marrouche NF, Brachmann J, Andresen D, et al. Catheter ablation for atrial fibrillation with heart
12. Cappato R, Calkins H, Chen S-A, et al. Updated worldwide survey on the methods, efficacy, and safety of catheter ablation for human atrial fibrillation. Circ Arrhythm Electrophysiol 2010;3:32–8.
failure. N Engl J Med 2018;378:417–27. 5. Prabhu S, Taylor AJ, Costello BT, et al. Catheter ablation versus medical rate control in atrial fibrillation and systolic dysfunction. J Am Coll Cardiol 2017;70:1949–61. 6. January CT, Wann LS, Calkins H, et al. 2019 AHA/ACC/HRS focused update of the 2014 AHA/ ACC/HRS guideline for the management of patients with atrial fibrillation. J Am Coll Cardiol 2019;74:104–32. 7. Tripathi B, Arora S, Kumar V, et al. Temporal trends of in-hospital complications associated with catheter ablation of atrial fibrillation in the United States: an update from Nationwide Inpatient Sample database (2011–2014). J Cardiovasc Electrophysiol 2018;29:715–24. 8. Hosseini SM, Rozen G, Saleh A, et al. Catheter ablation for cardiac arrhythmias: utilization and inhospital complications. JACC Clin Electrophysiol 2017;3:1240–8.
13. Abdur RK, Wazni OM, Barakat AF, et al. Lifethreatening complications of atrial fibrillation ablation. JACC Clin Electrophysiol 2019;5:284–91. 14. Voskoboinik A, Sparks PB, Morton JB, et al. Low rates of major complications for radiofrequency ablation of atrial fibrillation maintained over 14 Years: a single centre experience of 2750 consecutive cases. Heart Lung Circ 2018;27: 976–83. 15. Piccini JP, Sinner MF, Greiner MA, et al. Outcomes of Medicare beneficiaries undergoing catheter ablation for atrial fibrillation. Circulation 2012;126:2200–7. 16. Agency for Healthcare Research and Quality. Introduction to the HCUP Nationwide Readmissions Database (NRD). Available at: https://www.hcupus.ahrq.gov/db/nation/nrd/Introduction_NRD_20102015.pdf. Accessed July 18, 2018.
isation for acute myocardial infarction, subsequent short and long term mortality, and the prognostic impact of sex and comorbidity: a Danish nationwide cohort study. BMJ 2012;344:e356. 19. Kim LK, Yeo I, Cheung J, Feldman DN, Swaminathan RV, Kamel H. Impact of evolving comorbidities on outcomes after catheter ablation of atrial fibrillation in the United States between 2003 and 2013. JACC Clin Electrophysiol 2018;4:704–6. 20. Di Biase L, Mohanty P, Mohanty S, et al. Ablation versus amiodarone for treatment of persistent atrial fibrillation in patients with congestive heart failure and an implanted device. Circulation 2016;133:1637–44. 21. Han H-C, Ha FJ, Sanders P, et al. Atrioesophageal fistula. Circ Arrhythm Electrophysiol 2017;10:e005579. 22. Barrett ML, Ross DN. HCUP Quality Control Procedures Deliverable #1707.05. Available at: https://www.hcup-us.ahrq.gov/db/quality.pdf. Accessed February 20, 2017.
KEY WORDS atrial fibrillation, catheter ablation, early mortality, outcomes
A PPE NDI X For supplemental tables and figures, please see the online version of this paper.