JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY

VOL. 68, NO. 9, 2016

2016 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION

ISSN 0735-1097/$36.00

PUBLISHED BY ELSEVIER

http://dx.doi.org/10.1016/j.jacc.2016.05.085

Outcomes of Patients With

Atrial Fibrillation Undergoing
Percutaneous Coronary Intervention
Nadia R. Sutton, MD, MPH,a Milan Seth, MS,a Cyril Ruwende, MD, PHD,b Hitinder S. Gurm, MBBSa,c

ABSTRACT
BACKGROUND Atrial brillation (AF) is increasing in prevalence, and patients with a history of AF commonly undergo
percutaneous coronary intervention (PCI). There is a paucity of contemporary data on the association between AF and
clinical outcomes after PCI.
OBJECTIVES The study sought to evaluate the association between AF and in-hospital adverse outcomes using a large,
prospective multicenter registry.
METHODS Data for consecutive PCI cases from 47 hospitals performed between April 2011 and December 2014 were
utilized for the analysis. Propensity-matched multivariate analysis was used to adjust for differences in baseline
characteristics between patients with and without a history of AF.
RESULTS Of 113,283 PCI cases during the study period, a history of AF was present in 13,912 patients (12%), which
varied by institution (range 2.5% to 18.4%). At baseline, patients with a history of AF were older and were more likely to
have comorbid congestive heart failure, cardiomyopathy, cerebrovascular disease, and chronic lung disease. Patients with
a history of AF were more likely to have in-hospital complications, including in-hospital mortality (3% vs. 1%). In
propensity-matched analysis, patients with a history of AF were more likely to be treated with a bare-metal stent
(27% vs. 18%). In the propensity-matched model, AF remained independently associated with an increased risk of
developing post-procedural bleeding (odds ratio [OR]: 1.32; 95% condence interval [CI]: 1.15 to 1.52), heart failure
(OR: 1.33; 95% CI: 1.17 to 1.52), cardiogenic shock (OR: 1.26; 95% CI: 1.08 to 1.48), and in-hospital mortality (OR: 1.41;
95% CI: 1.18 to 1.68).
CONCLUSIONS AF is common among patients undergoing PCI. AF is associated with older age, the presence of other
comorbidities, and independently associated with in-hospital post-procedural heart failure, cardiogenic shock, and
mortality. (J Am Coll Cardiol 2016;68:895904) 2016 by the American College of Cardiology Foundation.

trial brillation (AF) is the most common

and coronary artery disease (CAD), patients with AF

sustained arrhythmia treated in the United

often have coexistent CAD and are treated with

States

percutaneous coronary intervention (PCI). Patients

(1,2).

The

prevalence

of

AF

is

increasing, reective of increasing numbers of older

with

patients and the pervasiveness of comorbid illness

clinical dilemma. In addition to the need for dual

AF

who

require

PCI

pose

(2,3). Because there is overlap in risk factors for AF

antiplatelet therapy after PCI, patients often warrant

From the aDepartment of Internal Medicine, Division of Cardiovascular Medicine, University of Michigan, Ann Arbor, Michigan;
Listen to this manuscripts

audio summary by

Veterans Affairs Medical Center, Ann Arbor, Michigan. Executive and clinical support for the Blue Cross Blue Shield of Michigan

JACC Editor-in-Chief

Cardiovascular Consortium is provided by Blue Cross and Blue Shield of Michigan (BCBSM), and by Blue Care Network under the

Dr. Valentin Fuster.

aegis of BCBSMs Value Partnerships program. Although Blue Cross and Blue Shield of Michigan (BCBSM) and BMC2 work

Michigan Heart, IHA, St. Joseph Mercy Hospital, Ypsilanti, Michigan; and the cDepartment of Medicine, Section of Cardiology,

collaboratively, the opinions, beliefs and viewpoints expressed by the authors do not necessarily reect the opinions, beliefs, and
viewpoints of BCBSM or any of its employees. Dr. Gurm has received research funding from Blue Cross Blue Shield of Michigan
and the National Institutes of Health; and has served as a consultant to Osprey Medical. All other authors have reported that they
have no relationships relevant to the contents of this paper to disclose.
Manuscript received August 10, 2015; revised manuscript received April 26, 2016, accepted May 18, 2016.

challenging

896

Sutton et al.

JACC VOL. 68, NO. 9, 2016

Atrial Fibrillation and Post-PCI Clinical Outcomes

AUGUST 30, 2016:895904

ABBREVIATIONS

anticoagulation to lower the risk of stroke.

propensity score model are listed in Online Table 1.

AND ACRONYMS

There is a heightened level of interest in

Patients missing covariate data were excluded from

this topic, given the need to weigh the

the matching exercise and thus from the matched

bleeding risk of these therapies against the

cohort. The association between AF and outcomes

risk of future thrombotic events.

was assessed univariately using Fisher exact tests and

AF = atrial brillation
BMC2 = Blue Cross Blue Shield
of Michigan Cardiovascular
Consortium

through multivariable logistic regression models

SEE PAGE 905

tted to the propensity-matched cohort, adjusting for

BMS = bare-metal stent

CAD = coronary artery disease

As AF is associated with higher left atrial

variables included in the propensity score.

lling pressure and myocardial remodeling,

The primary outcome of interest was in-hospital

we hypothesized that AF could be a risk fac-

mortality. Secondary outcomes were in-hospital

tor for poor clinical outcomes in patients

post-procedural myocardial infarction, cardiogenic

undergoing PCI (4,5). Despite the frequency

shock, heart failure, stroke, new requirement for

intervention

with which this scenario is encountered in

dialysis, acute kidney injury, vascular complications

STEMI = ST-segment elevation

clinical practice, there are scant contempo-

requiring treatment, need for blood transfusion, and

myocardial infarction

rary published reports describing the inci-

bleeding events within 72 h of PCI (9). In-hospital

UA = unstable angina

dence and characteristics of patients with AF

death was dened as death from both cardiac and

undergoing PCI. Furthermore, there is minimal data

noncardiac causes during the admission in which PCI

describing the procedural characteristics and in-

was performed. A composite of pseudoaneurysm,

hospital outcomes in this population. In this large,

arteriovenous stula, femoral neuropathy, retroperi-

database-derived study using the Blue Cross Blue

toneal hematoma, access site hematoma requiring

Shield Registry, the goal was to describe the popula-

transfusion or associated with prolonged hospital

tion and to evaluate the association between AF and

stay, drop in hemoglobin >3.0 g/dl, or any access site

in-hospital adverse outcomes.

complication requiring surgical repair comprised

DES = drug-eluting stent

NSTEMI = nonST-segment
elevation myocardial infarction

PCI = percutaneous coronary

vascular complications. Bleeding events were dened

METHODS

as a suspected or conrmed bleeding event observed

and documented in the medical record that was

The Blue Cross Blue Shield of Michigan Cardiovas-

associated with any of the following: 1) hemoglobin

cular Consortium (BMC2) is a prospective, multi-

drop $3 g/dl; 2) transfusion of whole blood or packed

center registry that collects demographic, clinical,

red blood cells; 3) procedural intervention or surgery

procedural, and in-hospital outcomes data from

at the bleeding site to stop bleeding (e.g., surgical

consecutive PCI cases at all nonfederal hospitals

closure or exploration of the arteriotomy site, balloon

in the state of Michigan. Details of BMC2 data

angioplasty to seal an arterial tear, endoscopy with

collection and audit processes have been described

cautery for a gastrointestinal bleed).

For the primary outcome of in-hospital mortality, a

previously (6,7).
All PCI cases in the BMC2 registry performed be-

subgroup analysis was performed evaluating the

tween April 2011 and December 2014 were used for

association of mortality and AF within subgroups of

the analysis. Baseline characteristics between pa-

the matched patient cohort dened by a number of

tients with a history of AF and patients without AF

baseline characteristics, including age, sex, presen-

were compared using Pearson chi-square tests for

tation category, history of heart failure, and pre-

categorical variables and Student t tests for contin-

procedural left ventricular ejection fraction. These

uous variables. Continuous variables are summarized

analyses were conducted using multivariate logistic

using mean
SD. Propensity score matching and

regression models, adjusting for baseline covariates

multivariate regression models were used to adjust

tted within each subgroup. In addition, the extent to

for differences in baseline characteristics. A pro-

which subgroup denitions modied the effect of AF

pensity score was estimated using logistic regression,

on the risk of mortality was assessed statistically

and patients with a history of AF were matched on a

through likelihood ratio tests assessing the inclusion

1:1 basis using a greedy algorithm without replace-

of the subgroup by AF interaction terms in the

ment to patients without AF (8). Matched pairs were

regression model.

required to have propensity scores within a caliper of

Data on oral anticoagulant therapy is available in

0.25 SD, and were exact matched on the clinical pre-

the BMC2 registry for medications administered dur-

sentation category: ST-segment elevation myocardial

ing the hospitalization. Except in rare cases, patients

infarction

elevation

were not taking oral anticoagulation at the time of

myocardial infarction (NSTEMI), unstable angina

PCI. In our study population, the absolute numbers of

(UA), and stable angina. Covariates included in the

inpatients receiving oral anticoagulation was small

(STEMI),

nonST-segment

JACC VOL. 68, NO. 9, 2016

Sutton et al.

AUGUST 30, 2016:895904

Atrial Fibrillation and Post-PCI Clinical Outcomes

(data not shown) and could be under-representative

Compared with patients with a history of AF, patients

of the use of oral anticoagulation prior to admission,

without AF were more likely to smoke. Patients with a

as oral anticoagulants are often stopped in anticipa-

history of AF were more likely to present with

tion of PCI. Anticoagulation is typically not recom-

NSTEMI and less likely to present with STEMI,

mended until 48 h after PCI, and many patients are

compared with patients without AF. Compared with

discharged prior to that time point. Therefore, pa-

patients without AF, patients with a history of AF

tients receiving oral anticoagulation in the hospital

were more likely to present in cardiogenic shock or

are likely not representative of the broader popula-

with cardiac arrest, and to be in cardiogenic shock at

tion, and data on oral anticoagulant use is therefore

the start of PCI.

not reported here.

Procedural characteristics for patients with and

without a history of AF are shown in Table 2. At

RESULTS

baseline, patients with a history of AF were more

Between April 2011 and December 2014, there were

plasty alone and less likely to receive a drug-eluting

113,283 PCIs performed at 47 hospitals in Michigan.

stent (DES), compared with patients without AF.

likely to receive a bare-metal stent (BMS) or angio-

Among those, 13,912 (12.3%) patients had a history of

Data for calculation of CHADS2 score were available

AF. Among the 47 hospitals, there was a wide range

for a subset of patients undergoing PCI after January

(2.5% to 18.4%) in the proportion of patients with a

2013 (10). Use of angioplasty alone correlated posi-

history of AF undergoing PCI (Figure 1). Among the

tively with CHADS 2 score in patients with a history of

presentation categories, 8.0% of STEMI, 13.9% of

AF, but there was no trend in the percent use of DES

NSTEMI, 12.4% of UA, and 12.0% of stable angina

or

patients had a history of AF. UA was the most

Table 2). Compared with patients without AF, pa-

frequent indication for PCI in patients with and

tients with a history of AF were more likely to require

without a history of AF. Baseline demographic and

intra-aortic balloon pump placement or other me-

clinical features are shown in Table 1. Compared with

chanical support. PCI was categorized more often as

patients without AF, patients with a history of AF

urgent and less often as emergent or elective in pa-

were more likely to be older and to have prior

tients with a history of AF, compared with patients

congestive heart failure, cardiomyopathy, cerebro-

without AF.

BMS

with

increasing CHAD 2

scores

(Online

vascular disease, prior CAD, hypertension, chronic

Data on intraprocedural medications and in-

lung disease, diabetes, and chronic kidney disease.

hospital antiplatelet therapy is shown in Table 3.

F I G U R E 1 Proportion of Patients With a History of AF Undergoing PCI by Institution

20%

Atrial Fibrillation (%)

15%

10%

5%

0%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Site Number

The range of percentages of patients with a history of atrial brillation (AF) undergoing percutaneous coronary intervention (PCI) at 47
hospitals participating in the Blue Cross Blue Shield of Michigan Cardiovascular Consortium registry.

897

898

Sutton et al.

JACC VOL. 68, NO. 9, 2016

Atrial Fibrillation and Post-PCI Clinical Outcomes

AUGUST 30, 2016:895904

T A B L E 1 Baseline Characteristics of the Unmatched and Propensity-Matched Patients With and Without a History of AF

Unmatched
No AF
(n 99,371)

Propensity Matched

AF
(n 13,912)

Absolute
SD

No AF
(n 13,498)

AF
(n 13,498)

Absolute
SD

Demographic characteristics
64.1
11.9

72.1
10.6

70.75

72.3
10.5

72.0
10.6

2.46

66,187 (66.6)

9,322 (67.0)

0.85

9,032 (66.9)

9,018 (66.8)

0.22

White

84,584 (85.1)

12,546 (90.2)

15.43

12,177 (90.2)

12,159 (90.1)

0.45

African American

11,703 (11.8)

1,049 (7.5)

14.38

1,037 (7.7)

1,027 (7.6)

0.28

5,123 (5.2)

239 (1.7)

18.95

267 (2.0)

229 (1.7)

2.10
1.77

Age, yrs
Male
Race

Lack of health insurance

Clinical history
Body mass index, kg/m2

30.6
7.5

30.5
7.4

2.21

30.6
7.24

30.5
7.40

Diabetes

37,663 (37.9)

6,022 (43.3)

10.99

5,993 (44.4)

5,875 (43.5)

1.76

Hypertension

83,929 (84.5)

12,839 (92.3)

24.64

12,562 (93.1)

12,482 (92.5)

2.29

Dyslipidemia

80,633 (81.2)

11,873 (85.4)

11.23

11,617 (86.1)

11,551 (85.6)

1.40

Current/recent smoker

30,844 (31.1)

2,371 (17.0)

33.22

2,237 (16.6)

2,299 (17.0)

1.23

Prior myocardial infarction

33,972 (34.2)

5,811 (41.8)

15.69

5,669 (42.0)

5,644 (41.8)

0.38

Prior PCI

44,516 (44.8)

6,936 (49.9)

10.17

6,804 (50.4)

6,757 (50.1)

0.70

Prior CABG

17,025 (17.1)

3,942 (28.3)

26.98

3,837 (28.4)

3,844 (28.5)

0.11

Prior congestive heart failure

13,351 (13.4)

5,215 (37.5)

57.46

4,861 (36.0)

5,029 (37.3)

2.58

Cardiomyopathy or LV systolic dysfunction

9,239 (9.3)

3,041 (21.9)

35.17

2,903 (21.5)

2,908 (21.5)

0.09

26,386 (27.2)

6,195 (45.2)

38.07

6,138 (45.7)

6,055 (45.0)

1.29

2157 (2.2)

602 (4.3)

12.18

575 (4.3)

593 (4.4)

0.66
0.52

Chronic kidney disease (GFR <60 ml/min/1.73 m2)

Dialysis

13.5
1.9

12.7
2.0

40.32

12.7
2.0

12.7
2.0

Chronic lung disease

17,696 (17.8)

3,904 (28.1)

24.58

3,796 (28.1)

3,774 (28.0)

0.36

Cerebrovascular disease

14,135 (14.2)

3,565 (25.6)

28.84

3,507 (26.0)

3,461 (25.6)

0.78

1,396 (1.4)

692 (5.0)

20.42

577 (4.3)

650 (4.8)

2.60

15,028 (15.1)

3,253 (23.4)

21.07

3,205 (23.7)

3,167 (23.5)

0.66

Baseline hemoglobin, g/dl

Prior valve surgery

Peripheral arterial disease
Presentation characteristics
STEMI

16,978 (17.1)

1,479 (10.6)

18.77

1,365 (10.1)

1,365 (10.1)

0.00

NSTEMI

21,471 (21.6)

3,459 (24.9)

7.72

3,388 (25.1)

3,388 (25.1)

0.00

Unstable angina

43,728 (44.0)

6,190 (44.5)

0.98

6,055 (44.9)

6,055 (44.9)

0.00

11,022 (11.1)

1,506 (10.8)

0.86

1,472 (10.9)

1,472 (10.9)

0.00

Cardiogenic shock

1,656 (1.7)

467 (3.4)

10.82

395 (2.9)

423 (3.1)

1.21

Cardiac arrest

1,827 (1.8)

412 (3.0)

7.34

345 (2.6)

372 (2.8)

1.24

Stable angina

Cardiogenic shock at start of PCI

1,951 (2.0)

483 (3.5)

9.29

414 (3.1)

441 (3.3)

1.14

Pre-PCI LVEF

52.6
12.5

47.9
14.7

34.20

48.4
14.7

48.0
14.6

2.80

Values are mean
SD, %, or n (%).

AF atrial brillation; CABG coronary artery bypass graft; GFR glomerular ltration rate; LV left ventricular; LVEF left ventricular ejection fraction;
NSTEMI nonST-segment elevation myocardial infarction; PCI percutaneous coronary intervention; STEMI ST-segment elevation myocardial infarction.

At baseline, patients with a history of AF were more

transfusion, bleeding complications, and in-hospital

likely to receive heparin alone and less likely to

mortality.

receive a glycoprotein IIb/IIIa inhibitor in addition to

OUTCOME ANALYSIS. A total of 13,498 patients with

heparin. Compared with patients without AF, pa-

a history of AF were successfully matched to similar

tients with a history of AF were slightly more likely to

patients without AF. In the matched cohort, the ab-

receive bivalirudin. Those with a history of AF more

solute

frequently received clopidogrel and less frequently

was #2.7, indicating that the cohort was well matched

received prasugrel or ticagrelor, compared with those

(Table 1, Online Figure 1). Differences in choice of

without AF.

stent remained after propensity matching; patients

standardized

difference

for

all

variables

In-hospital outcomes for patients undergoing PCI

with a history of AF were more likely to receive a BMS

are shown in Table 4. Patients with a history of AF

or angioplasty alone, and less likely to receive a DES

were more likely to have post-procedural cardiogenic

compared with the matched cohort (Table 2). After

shock, heart failure, stroke, acute kidney injury,

matching, compared with patients without AF, pa-

dialysis initiation, vascular complications, blood

tients with a history of AF were more likely to have

JACC VOL. 68, NO. 9, 2016

Sutton et al.

AUGUST 30, 2016:895904

Atrial Fibrillation and Post-PCI Clinical Outcomes

T A B L E 2 Procedural Characteristics of Unmatched and Propensity-Matched Patients With and Without a History of AF

Unmatched

Propensity Matched

No AF
(n 99,371)

AF
(n 13,912)

Absolute
SD

p Value

No AF
(n 13,498)

AF
(n 13,498)

Absolute
SD

Bare-metal stent

15,670 (15.8)

3,779 (27.2)

p Value

28.0

<0.001

2,454 (18.1)

3,676 (27.2)

21.7

<0.001

Drug-eluting stent

74,949 (75.4)

8,583 (61.7)

29.9

<0.001

9,679 (71.7)

8,332 (61.7)

21.3

<0.001

Angioplasty only

9,634 (9.7)

1,674 (12.0)

7.52

<0.001

1,497 (11.1)

1,612 (11.9)

2.67

0.030

Stents per procedure

1.33
0.85

1.28
0.85

1.33

<0.001

1.33
0.88

1.29
0.85

1.33

<0.001

Intra-aortic balloon pump placement

2,492 (2.5)

432 (3.1)

3.62

<0.001

419 (3.1)

397 (2.9)

0.95

0.436

825 (0.8)

192 (1.4)

5.26

<0.001

246 (1.8)

181 (1.3)

3.86

0.002

Bifurcation lesion

9,357 (9.4)

1257 (9.0)

1.32

0.153

1,350 (10.0)

1,219 (9.0)

3.31

0.007

Chronic total occlusion

3,070 (3.1)

395 (2.8)

1.48

0.115

375 (2.8)

378 (2.8)

0.13

0.941

16,691 (16.8)

1,741 (12.5)

12.1

<0.001

1,690 (12.5)

1,656 (12.3)

0.76

0.542

Other mechanical support

Thrombus present

22.5
12.8

21.7
12.3

6.25

<0.001

22.1
12.6

21.7
12.3

3.17

0.010

Contrast volume, ml

184.8
76.0

180.9
77.3

5.05

<0.001

183.7
77.6

180.9
77.3

2.34

0.055

Arterial access site femoral

77,055 (77.6)

10,917 (78.5)

2.21

0.015

10,885 (80.7)

10,578 (78.4)

5.68

<0.001

Arterial access site radial

21,969 (22.1)

2,928 (21.0)

2.59

0.004

2,550 (18.9)

2,854 (21.1)

5.62

<0.001

Elective PCI

36,419 (36.7)

4,799 (34.5)

4.52

<0.001

4,809 (35.6)

4,676 (34.6)

2.06

0.090

Urgent PCI

44,911 (45.2)

7,426 (53.4)

16.4

<0.001

7,100 (52.6)

7,265 (53.8)

2.45

0.044

Emergency PCI

17,810 (17.9)

1,645 (11.8)

17.2

<0.001

1,553 (11.5)

1,526 (11.3)

0.63

0.605

178 (0.2)

38 (0.3)

1.98

0.017

36 (0.3)

31 (0.2)

0.74

0.541

Lesion length, mm

Salvage PCI
Values are n (%), %, or mean
SD.
Abbreviations as in Table 1.

had radial arterial access and less likely to have had

conditional logistic regression model to account for

femoral arterial access.

the paired nature of the matched cohort yielded

In-hospital

propen-

results that were consistent with the results of

sity matching are shown in Table 4 and in the

the multivariable logistic regression model (data

Central

adverse

Illustration.

outcomes

After

after

propensity

matching,

not shown).

compared with patients without AF, patients with a

Subgroup analysis revealed that the association

history of AF were more likely to have post-

between AF and risk-adjusted mortality was consis-

procedural bleeding, a need for blood transfusion,

tent across subgroups (Figure 2). No subgroup by AF

heart failure, cardiogenic shock, and in-hospital

interactions (including age) was associated with a

mortality. There were no differences in the inci-

statistically signicant improvement in model t. The

dence of post-procedural stroke, acute kidney injury,

standardized differences of baseline characteristics of

dialysis initiation, or vascular complications in

patients with and without a history of AF, by sub-

the matched cohorts. Outcome analysis using a

group, are shown in Online Figure 2.

T A B L E 3 Pharmacotherapy for PCI in Patients With and Without a History of AF

Unmatched

Propensity Matched

No AF
(n 99,371)

AF
(n 13,912)

Absolute
SD

p Value

Heparin

38,557 (38.8)

6,106 (43.9)

10.4

<0.001

Heparin GP IIb/IIIa inhibitor

22,642 (22.8)

2,428 (17.5)

13.3

<0.001

Bivalirudin

42,753 (43.0)

6,139 (44.1)

Clopidogrel

71,415 (71.9)

11,393 (81.9)

Prasugrel

21,675 (21.8)

1,599 (11.5)

Ticagrelor

11,122 (11.2)

1,209 (8.7)

No AF
(n 13,498)

AF
(n 13,498)

Absolute
SD

5,763 (42.7)

5,943 (44.0)

2.69

0.027

2,525 (18.7)

2,341 (17.3)

3.55

0.004

0.014

5,873 (43.5)

5,971 (44.2)

1.46

0.229

24.0

<0.001

10,870 (80.5)

11,073 (82.0)

3.86

0.002

28.0

<0.001

1,832 (13.6)

1,547 (11.5)

6.38

<0.001

<0.001

1,285 (9.5)

1,165 (8.6)

3.10

0.011

p Value

Intraprocedure

2.23

In-hospital

Values are n (%) or %.

GP glycoprotein; other abbreviations as in Table 1.

8.37

899

900

Sutton et al.

JACC VOL. 68, NO. 9, 2016

Atrial Fibrillation and Post-PCI Clinical Outcomes

AUGUST 30, 2016:895904

T A B L E 4 Unmatched and Propensity-Matched In-Hospital Outcomes of Patients With

and Without a History of AF Undergoing PCI

presentations for PCI (12,16). To our knowledge,

among recent studies in which risk adjustment for
other factors known to inuence poor outcomes was

Unmatched

Propensity Matched

performed, there is only 1 other report of the associ-

No AF
AF
No AF
AF
(n 99,371) (n 13,912) p Value (n 13,498) (n 13,498) p Value

ation of AF with short-term mortality in the setting of

Myocardial infarction

1,806 (1.8)

276 (2.0)

0.173

260 (1.9)

264 (2.0)

0.863

STEMI (14). Here we report a 1% risk of in-hospital

Cardiogenic shock

1,923 (1.9)

405 (2.9)

<0.001

312 (2.3)

383 (2.8)

0.006

mortality in patients without AF, compared with a

Heart failure

2,317 (2.3)

609 (4.4) <0.001

446 (3.3)

575 (4.3)

<0.001

3% risk of mortality in patients with a history of AF,

Stroke

278 (0.3)

67 (0.5) <0.001

51 (0.4)

61 (0.5)

0.345

and 2% versus 3% mortality, respectively, after pro-

New requirement
for dialysis

333 (0.3)

85 (0.6) <0.001

83 (0.6)

81 (0.6)

0.873

pensity matching. It is noteworthy that the exagger-

2,476 (2.9)

701 (5.9)

<0.001

638 (5.5)

673 (5.8)

0.338

505 (0.5)

98 (0.7)

0.003

72 (0.5)

96 (0.7)

0.064

Blood transfusion

2,519 (2.5)

739 (5.3)

<0.001

626 (4.6)

706 (5.2)

0.025

Bleeding event
within 72 h

2,484 (2.5)

518 (3.7)

<0.001

379 (2.8)

494 (3.7)

<0.001

In-hospital mortality

1,309 (1.3)

439 (3.2)

<0.001

321 (2.4)

407 (3.0)

0.001

Acute kidney injury*

Vascular complications
requiring treatment

ated risk of mortality among patients with a history of

AF was noted across the spectrum of clinical presentation, and should be taken into account when
considering risk estimation and benchmarking.
We found that patients with a history of AF were
more likely to present in cardiogenic shock or with
cardiac arrest. In line with prior reports, patients with

Values are n (%). *Creatinine increase of $0.5 mg/dl after PCI, excludes patients on dialysis or where creatinine
value not available.
Abbreviations as in Table 1.

a history of AF were more likely to experience postprocedural

heart failure

and

cardiogenic

shock

(14,17,18,22,23). As with mortality, only 1 other recent

report concluded that AF was associated with post-

DISCUSSION

procedural cardiogenic shock or cardiac arrest after

The key nding in this study is that AF is strongly

though patients had undergone PCI, those with AF

associated with in-hospital adverse outcomes, even

may have had a greater burden of nonrevascularized

after extensive adjustment for other relevant factors.

coronary artery disease, which could have contrib-

A history of AF was present in 12% of patients

uted to this nding. It has been theorized that AF is a

undergoing PCI in this study, highlighting how

clinical manifestation of a brotic atrial cardiomyop-

commonly this scenario occurs in contemporary

athy that can occur independently of age and

risk adjustment in the setting of STEMI (14). Even

clinical practice, although the prevalence of AF in

comorbidities (28,29). Whether AF drives this struc-

patients undergoing PCI varied considerably by

tural process or is a consequence is unclear. The data

institution. A history of AF was associated with the

presented herein are consistent with the theory that

presence of other medical conditions, and with the

AF represents underlying cardiomyopathy, as a his-

presence of cardiac arrest and cardiogenic shock on

tory of AF was found to be an independent risk factor

presentation. Those with a history of AF were found

for poor clinical outcomes related to hemodynamic

to have a higher rate of post-procedural bleeding,

alterations and pump function.

need for transfusion, heart failure, cardiogenic shock,

At baseline, there was a higher risk of bleeding and

and mortality. These ndings indicate that a history

stroke in patients with a history of AF, similar to prior

of AF is a marker of patients who are vulnerable to

reports (1113,16,20,22). Following propensity match-

serious in-hospital complications after PCI.

ing, we did not observe an increased risk of stroke,

Consistent with prior reports, we found that pa-

but an increased risk of bleeding remained. Except for

tients with a history of AF were older and had more

rare circumstances, oral anticoagulation is held pre-

comorbid illness (1116). After propensity matching,

and post-procedurally. Most patients with AF resume

AF remained an independent risk factor for poor

anticoagulation 48 h post-procedure, often after

clinical outcomes. Patients with a history of AF had a

discharge.

higher rate of in-hospital mortality. This was consis-

bleeding complications may be related to periproce-

tent across subgroups. The association between AF

dural anticoagulation and antiplatelet agents. We

and mortality has been described, with the bulk of the

found that clopidogrel was used more frequently and

literature relating to AF in the setting of acute

prasugrel and ticagrelor less frequently in patients

myocardial infarction. Most prior studies have found

with AF relative to the comparator group. Patients

an association between AF and mortality, with vari-

with a history of AF were less likely to receive epti-

ations in populations studied, including all patients

batide intraprocedurally, were more likely to receive

with acute coronary syndrome (ACS) (11,17), STEMI

heparin alone, and had similar rates bivalirudin uti-

(14,15,1825), NSTEMI (19,2326), and UA (27), and all

lization. This suggests that clinicians are attempting

These

data

suggest

that

in-hospital

JACC VOL. 68, NO. 9, 2016

Sutton et al.

AUGUST 30, 2016:895904

Atrial Fibrillation and Post-PCI Clinical Outcomes

C ENTR AL I LL U STRA T I O N AF and Post-PCI Clinical Outcomes

Sutton, N.R. et al. J Am Coll Cardiol. 2016;68(9):895904.

Propensity-matched in-hospital outcomes of patients with and without a history of atrial brillation (AF) undergoing percutaneous coronary
intervention (PCI).

to mitigate a presumed elevated risk of bleeding in

current guidelines suggest that the benet of DES

this population. Despite this, patients with a history

with regard to restenosis be weighed against the

of AF were still more likely to have a bleeding

bleeding risk associated with triple therapy, and that

complication or require a blood transfusion. The

DES be avoided in patients who have a higher

absence of a detectable difference in the incidence of

bleeding risk and, when possible, in patients who will

stroke and post-procedural myocardial infarction be-

require oral anticoagulation long term (3033).

tween cohorts differs from prior reports describing an

A prior study spanning 2004 to 2006 also found

increased risk of these poor outcomes in patients

increased use of BMS in patients with AF, although

with AF (14,15). This inconsistency could be due to

the absolute percent of patients receiving BMS has

improvements in treatment practices over time or

declined substantially in all patients in the interval

differences in time points examined (1416).

(14). We are unable to discern from these data the

We report here that patients with a history of AF

exact reasons for increased BMS use in patients with a

were far more likely to receive a BMS or angioplasty

history of AF in this study. Additional explanations

alone, and were less likely to receive a DES. This

for this observation are concern for bleeding on triple

nding likely stems from the need for oral anti-

therapy,

coagulation in many patients with AF. Selecting a

bleeding, or other unmeasured confounders.

medication

compliance,

history

of

BMS affords the possibility of shortening the duration

These data are important for several reasons. First,

of dual antiplatelet therapy plus oral anticoagulation

this report denes the prevalence of AF in patients

(triple therapy) if a bleeding event occurs. In general,

undergoing PCI in contemporary clinical practice, in

901

902

Sutton et al.

JACC VOL. 68, NO. 9, 2016

Atrial Fibrillation and Post-PCI Clinical Outcomes

AUGUST 30, 2016:895904

F I G U R E 2 Risk-Adjusted In-Hospital Mortality After PCI in Subgroups

OR (95% CI)

Primary Outcome Subgroup Analysis

Male

1.32 (1.04-1.68)

Female

1.45 (1.10-1.90)

Age < 65

1.51 (0.96-2.37)

Age 65 or older

1.36 (1.12-1.65)

STEMI

1.31 (1.00-1.71)

NSTEMI

1.31 (0.97-1.75)

Unstable Angina

1.71 (1.04-2.82)

Stable Angina

7.33 (0.19-282.27)

Heart Failure

1.50 (1.14-1.97)

No Heart Failure

1.26 (0.99-1.59)

EF < 40

1.27 (0.93-1.72)

EF >= 40

1.35 (0.95-1.90)

EF not available

1.57 (1.16-2.11)
0

2
3
Adjusted Odds Ratio: AF vs. No AF

Adjusted odds ratio (OR) for post-procedural in-hospital mortality in propensity-matched patients with and without a history of AF in
subgroups. CI condence interval; EF ejection fraction; NSTEMI nonST-segment elevation myocardial infarction; STEMI ST-segment
elevation myocardial infarction; other abbreviations as in Figure 1.

itself a useful observation. We found that the preva-

The BMC2 database was sufcient to report on

lence of a history of AF in patients undergoing PCI

>25,000 propensity-matched cases. Patients were not

was 12.3%, which is higher than in recent reports (5%

excluded on the basis of eligibility for clinical trials;

to 11%) (11,12,1416,19). It is possible that this reects

therefore, this is truly reective of patients treated in

the increased prevalence of AF in the general popu-

clinical practice. In this study, AF could have

lation, as well as the increased prevalence of obesity,

occurred at any time prior to PCI, not only peri-

a known risk factor for AF in the PCI population (34).

procedurally. As a result, these data capture a broader

Furthermore, we demonstrate a wide range in the

population of patients with AF. This is of relevance

proportion of patients with a history of AF undergo-

because there are implications for stent choice and

ing PCI by institution. These data strongly suggest

antiplatelet agents in patients with a history of AF, as

that AF should be taken into account for individual

well as in patients with new-onset AF at the time of

patients, as well as when grading institutions on

PCI. Other recent studies have been limited by size,

quality and the outcomes of patients undergoing PCI.

limitation to DES, a single presentation type, or lack

We identied that patients with a history of AF

of a comparator group (12,1416,35). The BMC2 was an

undergoing PCI are more susceptible to developing

ideal database to use to overcome these limitations,

post-procedural complications of bleeding, heart

with detailed information on baseline clinical vari-

failure,

ables, presentation, procedural variables, and clinical

cardiogenic

shock,

and

mortality,

and

acknowledgment of this potential may be valuable

outcomes.

clinically. The noted association is an important

reminder that the presence of AF marks underlying

STUDY

pathology. The association of AF with mortality after

include that this is an analysis of observational,

LIMITATIONS. Limitations

PCI was consistent across a spectrum of presentation

nonrandomized data. This report includes data from 1

types. Resources could be targeted to further evaluate

state, and it is unknown to what degree geographic

the factors responsible for this association and to

variations in treatment patterns may inuence the

lower morbidity and mortality in this high-risk

results. The data is drawn from hospitals partici-

subgroup.

pating in a quality-improvement initiative, and it is

of

this

study

This represents the largest contemporary report on

unclear if the ndings would apply to hospitals not

the association of AF with clinical outcomes after PCI.

participating in such an initiative. Although there

JACC VOL. 68, NO. 9, 2016

Sutton et al.

AUGUST 30, 2016:895904

Atrial Fibrillation and Post-PCI Clinical Outcomes

was extensive adjustment for risk factors, residual

after PCI. AF can be viewed as a harbinger of

measured or unmeasured confounding factors may

in-hospital

exist. The specic reasons for stent choice and choice

after PCI.

complications,

including

mortality,

of antiplatelet agent were not available. We do not

report on the association of oral anticoagulant use

REPRINT REQUESTS AND CORRESPONDENCE: Dr.

with clinical outcomes because patients were rarely

Hitinder S. Gurm, Division of Cardiovascular Medi-

administered these medications periprocedurally. We

cine, University of Michigan Cardiovascular Center,

are unable to correlate with long-term clinical

2A381, SPC 5869, 1500 East Medical Center Drive, Ann

adverse outcomes.

Arbor, Michigan 48109-5869. E-mail: hgurm@med.

umich.edu.

CONCLUSIONS
A history of AF is common among patients presenting for PCI, and is associated with comorbid
illnesses. The prevalence of a history of AF in patients undergoing PCI varies substantially by institution. Patients with a history of AF were more
likely to receive a BMS or angioplasty alone, and
less likely to receive a DES. We found that a history
of AF is associated with a higher risk of in-hospital
complications after PCI, including bleeding, heart
failure, cardiogenic shock, and mortality. These
ndings highlight the importance of AF as an in-

PERSPECTIVES
COMPETENCY IN PATIENT CARE AND PROCEDURAL
SKILLS: A history of AF in patients undergoing PCI is associated with
an increased risk of in-hospital adverse outcomes, including bleeding,
post-procedural heart failure, cardiogenic shock, and mortality.
TRANSLATIONAL OUTLOOK: Prospective clinical studies are
needed to evaluate the impact of prophylactic therapies on postprocedural morbidity and mortality in patients with a history of
AF undergoing PCI.

dependent risk factor for poor clinical outcomes

REFERENCES
1. Kannel WB, Wolf PA, Benjamin EJ, et al. Prevalence, incidence, prognosis, and predisposing
conditions for atrial brillation: population-based
estimates. Am J Cardiol 1998;82:2N9N.
2. Go AS, Hylek EM, Phillips KA, et al. Prevalence of
diagnosed atrial brillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial
Fibrillation (ATRIA) Study. JAMA 2001;285:23705.
3. Miyasaka Y, Barnes ME, Gersh BJ, et al. Secular
trends in incidence of atrial brillation in Olmsted
County, Minnesota, 1980 to 2000, and implications on the projections for future prevalence.
Circulation 2006;114:11925.
4. Sugiura T, Iwasaka T, Ogawa A, et al. Atrial
brillation in acute myocardial infarction. Am J
Cardiol 1985;56:279.
5. Hirsh BJ, Copeland-Halperin RS, Halperin JL.
Fibrotic atrial cardiomyopathy, atrial brillation,
and thromboembolism: mechanistic links and
clinical inferences. J Am Coll Cardiol 2015;65:
223951.
6. Kline-Rogers E, Share D, Bondie D, et al.
Development of a multicenter interventional cardiology database: the Blue Cross Blue Shield of
Michigan Cardiovascular Consortium (BMC2)
experience. J Interv Cardiol 2002;15:38792.
7. Moscucci M, Rogers EK, Montoye C, et al.
Association of a continuous quality improvement
initiative with practice and outcome variations of
contemporary percutaneous coronary interventions. Circulation 2006;113:81422.

8. Sekhon JS, Grieve RD. A matching method for

improving covariate balance in cost-effectiveness
analyses. Health Econ 2012;21:695714.
9. NCDR CathPCI Registry v4. 4 Coders Data
Dictionary - Full Specications. Washington, DC:
American College of Cardiology Foundation,
2008.
10. Gage BF, Waterman AD, Shannon W, et al.
Validation of clinical classication schemes for
predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA 2001;285:
286470.
11. Lopes RD, Pieper KS, Horton JR, et al. Shortand long-term outcomes following atrial brillation in patients with acute coronary syndromes
with or without ST-segment elevation. Heart
2008;94:86773.
12. Chan W, Ajani AE, Clark DJ, et al. Impact of
periprocedural atrial brillation on short-term
clinical outcomes following percutaneous coronary intervention. Am J Cardiol 2012;109:4717.
13. El-Omar MM, Dangas G, Mehran R, et al.
Usefulness of atrial brillation as a marker of
outcome after percutaneous coronary intervention. Am J Cardiol 2003;91:2324.
14. Lopes RD, Elliott LE, White HD, et al. Antithrombotic therapy and outcomes of patients with
atrial brillation following primary percutaneous
coronary intervention: results from the APEX-AMI
trial. Eur Heart J 2009;30:201928.
15. Rene AG, Gnreux P, Ezekowitz M, et al.
Impact of atrial brillation in patients with

ST-elevation myocardial infarction treated with

percutaneous coronary intervention (from the
HORIZONS-AMI [Harmonizing Outcomes With
Revascularization and Stents in Acute Myocardial
Infarction] trial). Am J Cardiol 2014;113:23642.
16. Bramlage P, Cuneo A, Zeymer U, et al. Prognosis of patients with atrial brillation undergoing
percutaneous coronary intervention receiving drug
eluting stents. Clin Res Cardiol 2013;102:28997.
17. Mehta RH, Dabbous OH, Granger CB, et al.,
GRACE Investigators. Comparison of outcomes of
patients with acute coronary syndromes with and
without atrial brillation. Am J Cardiol 2003;92:
10316.
18. Kinjo K, Sato H, Ohnishi Y, et al., Osaka Acute
Coronary Insufciency Study (OACIS) Group.
Prognostic signicance of atrial brillation/atrial
utter in patients with acute myocardial infarction
treated with percutaneous coronary intervention.
Am J Cardiol 2003;92:11504.
19. Podolecki T, Lenarczyk R, Kowalczyk J, et al.
Effect of type of atrial brillation on prognosis in
acute myocardial infarction treated invasively. Am
J Cardiol 2012;109:168993.
20. Pizzetti F, Turazza FM, Franzosi MG, et al.,
GISSI-3 Investigators. Incidence and prognostic
signicance of atrial brillation in acute myocardial infarction: the GISSI-3 data. Heart 2001;86:
52732.
21. Pedersen OD, Bagger H, Kber L, et al. The
occurrence and prognostic signicance of atrial

903

904

Sutton et al.

JACC VOL. 68, NO. 9, 2016

Atrial Fibrillation and Post-PCI Clinical Outcomes

AUGUST 30, 2016:895904

brillation/-utter following acute myocardial

infarction. Eur Heart J 1999;20:74854.
22. Crenshaw BS, Ward SR, Granger CB, et al.,
GUSTO-I Trial Investigators. Atrial brillation in the
setting of acute myocardial infarction: the GUSTO-I
experience. J Am Coll Cardiol 1997;30:40613.
23. Rathore SS, Berger AK, Weinfurt KP, et al.
Acute myocardial infarction complicated by atrial
brillation in the elderly: prevalence and outcomes. Circulation 2000;101:96974.
24. Saczynski JS, McManus D, Zhou Z, et al. Trends
in atrial brillation complicating acute myocardial
infarction. Am J Cardiol 2009;104:16974.
25. Goldberg RJ, Seeley D, Becker RC, et al. Impact
of atrial brillation on the in-hospital and longterm survival of patients with acute myocardial
infarction: a community-wide perspective. Am
Heart J 1990;119:9961001.
26. Kowey PR, Breithardt G, Camm J, et al.
Physician stated atrial brillation management in
light of treatment guidelines: data from an international, observational prospective survey. Clin
Cardiol 2010;33:1728.
27. Al-Khatib SM, Pieper KS, Lee KL, et al. Atrial
brillation and mortality among patients with
acute coronary syndromes without ST-segment

catheter and surgical ablation of atrial brillation:

recommendations for patient selection, procedural
techniques, patient management and follow-up,
denitions, endpoints, and research trial design.

33. Lip GYH, Huber K, Andreotti F, et al.

Antithrombotic management of atrial brillation patients presenting with acute coronary
syndrome and/or undergoing coronary stenting:

Europace 2012;14:528606.

executive summarya Consensus Document of

the European Society of Cardiology Working
Group on Thrombosis, endorsed by the European
Heart Rhythm Association (EHRA) and the
European Association of Percutaneous Cardiovascular Interventions (EAPCI). Eur Heart J
2010;31:13118.

29. Mahnkopf C, Badger TJ, Burgon NS, et al.

Evaluation of the left atrial substrate in patients
with lone atrial brillation using delayedenhanced MRI: implications for disease progression and response to catheter ablation. Heart
Rhythm 2010;7:147581.
30. OGara PT, Kushner FG, Ascheim DD, et al.
2013 ACCF/AHA guideline for the management of
ST-elevation myocardial infarction: executive
summary: a report of the American College of
Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll
Cardiol 2013;61:485510.
31. Authors/Task Force Members, Steg PG,
James SK, et al. ESC Guidelines for the management of acute myocardial infarction in patients
presenting with ST-segment elevation. Eur Heart J
2012;33:2569619.

34. Buschur ME, Smith D, Share D, et al. The

burgeoning epidemic of morbid obesity in patients
undergoing percutaneous coronary intervention:
insight from the Blue Cross Blue Shield of Michigan Cardiovascular Consortium. J Am Coll Cardiol
2013;62:68591.
35. Ruiz-Nodar JM, Marin F, Hurtado JA, et al.
Anticoagulant and antiplatelet therapy use in 426
patients with atrial brillation undergoing percutaneous
coronary
intervention
and
stent implantation implications for bleeding risk
and prognosis. J Am Coll Cardiol 2008;51:81825.

32. European Heart Rhythm Association, European

KEY WORDS acute coronary syndrome,

28. Calkins H, Kuck KH, Cappato R, et al. 2012

Association for Cardio-Thoracic Surgery, Camm AJ,

et al., ESC Committee for Practice Guidelines.
Guidelines for the management of atrial brillation: the Task Force for the Management of Atrial
Fibrillation of the European Society of Cardiology
(ESC) [correction in Europace 2011;13:1058].

A PPE NDI X For supplemental gures and

tables, please see the online version of this

HRS/EHRA/ECAS expert consensus statement on

Europace 2010;12:1360420.

article.

elevation: results from the PURSUIT trial. Am J

Cardiol 2001;88:A7, 769.

clinical outcomes, stable angina