Original Article

Association Between Atrial Fibrillation Symptoms,

Quality of Life, and Patient Outcomes
Results From the Outcomes Registry for Better Informed Treatment of
Atrial Fibrillation (ORBIT-AF)
James V. Freeman, MD, MPH, MS; DaJuanicia N. Simon, MS; Alan S. Go, MD;
John Spertus, MD, MPH; Gregg C. Fonarow, MD; Bernard J. Gersh, MB, ChB, DPhil;
Elaine M. Hylek, MD, MPH; Peter R. Kowey, MD; Kenneth W. Mahaffey, MD;
Laine E. Thomas, PhD; Paul Chang, MD; Eric D. Peterson, MD, MPH;
Jonathan P. Piccini, MD, MHS;
on behalf of the Outcomes Registry for Better Informed Treatment of
Atrial Fibrillation (ORBIT-AF) Investigators and Patients

Background—Instruments to assess symptom burden and quality of life among patients with atrial fibrillation (AF) have not
been well evaluated in community practice or associated with patient outcomes.
Methods and Results—Using data from 10 087 AF patients in the Outcomes Registry for Better Informed Treatment of AF
(ORBIT-AF), symptom severity was evaluated using the European Heart Rhythm Association (EHRA) classification
system, and quality of life was assessed using the Atrial Fibrillation Effect on Quality-of-Life (AFEQT) questionnaire.
The association between AF-related symptoms, quality of life, and outcomes was assessed using Cox regression. The
majority of AF patients (61.8%) were symptomatic (EHRA >2) and 16.5% had severe or disabling symptoms (EHRA
3–4). EHRA symptom class was well correlated with the AFEQT score (Spearman correlation coefficient −0.39). Over
1.8 years of follow-up, AF symptoms were associated with a higher risk of hospitalization (adjusted hazard ratio for
EHRA ≥2 versus EHRA 1 1.23, 95% confidence interval, 1.15–1.31) and a borderline higher risk of major bleeding.
Lower quality of life was associated with a higher risk of hospitalization (adjusted hazard ratio for lowest quartile of
Downloaded from http://ahajournals.org by on May 9, 2020

AFEQT versus highest 1.49, 95% confidence interval, 1.2–1.84), but not other major adverse events, including death.
Conclusions—In a community-based study, most patients with AF were symptomatic and had impaired quality of life.
Quality of life measured by the AFEQT correlated closely with symptom severity measured by the EHRA class.
AF symptoms and lower quality of life were associated with higher risk of hospitalization but not mortality during
follow-up.  (Circ Cardiovasc Qual Outcomes. 2015;8:393-402. DOI: 10.1161/CIRCOUTCOMES.114.001303.)
Key Words: atrial fibrillation ◼ morbidity ◼ mortality ◼ quality of life ◼ symptoms

A trial fibrillation (AF) substantially increases the risk of major

adverse clinical outcomes, such as stroke1,2 and death.3 Yet,
AF can also cause frequent symptoms, affect patient’s functional
status, and impair their quality of life.4 Although prior studies
have reported the range of AF-related symptoms in patient popu-
lations, these studies were generally from highly selected patients
and referral-based practices and may not reflect results in com-
munity practice or results with contemporary AF management..5–8
Additionally, although disease-specific symptom and
quality of life assessment tools have been developed for AF,
these tools have not been well assessed in large community-
based populations. For example, the European Heart Rhythm
Association (EHRA) has proposed a scoring system for
AF-related symptoms (1=asymptomatic, 2=mild, 3=severe,
4=disabling),9,10 but it remains unstudied in a large clinical
cohort. Like the New York Heart Association classification

Received July 24, 2014; accepted May 4, 2015.

From the Yale University School of Medicine, New Haven, CT (J.V.F.); Duke Clinical Research Institute, Durham, NC (D.N.S., L.E.T., E.D.P., J.P.P.);
Division of Research, Kaiser Permanente of Northern California, Oakland, CA (A.S.G.); Saint Luke’s Mid America Heart Institute and University of
Missouri–Kansas City (J.S.); Ronald Reagan-UCLA Medical Center, Los Angeles, CA (G.C.F.); Mayo Clinic Medical Center, Rochester, Minnesota
(B.J.G.); Boston University Medical Center, Boston, MA (E.M.H.); Lankenau Institute for Medical Research and Jefferson Medical College, Philadelphia,
PA (P.R.K.); and Stanford University School of Medicine, Stanford, CA (K.W.M.); Janssen Pharmaceuticals Inc., Bridgewater, NJ (P.C.).
This article was handled independently by Mathew Reeves, DVM, PhD, as a Guest Editor. The editors had no role in the evaluation of this manuscript
or in the decision about its acceptance.
The Data Supplement is available at http://circoutcomes.ahajournals.org/lookup/suppl/doi:10.1161/CIRCOUTCOMES.114.001303/-/DC1.
Correspondence to James V. Freeman MD, MPH, MS, Yale University School of Medicine, PO Box 208017, New Haven, CT 06520. E-mail
james.freeman@yale.edu
© 2015 American Heart Association, Inc.
Circ Cardiovasc Qual Outcomes is available at http://circoutcomes.ahajournals.org DOI: 10.1161/CIRCOUTCOMES.114.001303

393
 394   Circ Cardiovasc Qual Outcomes   July 2015
WHAT IS KNOWN
• Atrial fibrillation causes symptoms, such as palpita-
tions and dyspnea with exertion and may decrease
patient quality of life.
WHAT THE STUDY ADDS
• The majority of atrial fibrillation patients (61.8%)
were symptomatic, and 16.5% had severe or dis-
abling symptoms.
• The physician-reported European Heart Rhythm
Association (EHRA) symptom class was well cor-
related with the patient-reported Atrial Fibrillation
Effect on Quality-of-Life (AFEQT) questionnaire
score (Spearman correlation coefficient, −0.39).
• Atrial fibrillation symptoms were associated with
a higher risk of hospitalization (adjusted HR for
EHRA ≥2 versus EHRA 1 1.23, 95% CI 1.15–1.31)
and a borderline higher risk of major bleeding, but
not death.
• Lower quality of life was associated with a higher
risk of hospitalization (adjusted HR for low-
est quartile of AFEQT versus highest 1.49, 95%
CI 1.2–1.84), but not other major adverse events,
including death.
for heart failure, the EHRA reflects physicians’ perspectives
Downloaded from http://ahajournals.org by on May 9, 2020
of patients’ health status (symptoms, function, and quality
of life) and may offer a simple and valuable tool for guiding
treatment decisions and conducting research. Similarly, the
Atrial Fibrillation Effect on Quality-of-Life (AFEQT) ques-
tionnaire11 was recently developed as a disease-specific tool
for assessing quality of life and health status from patients’
perspectives, but has yet to be evaluated in a large commu-
nity-based patient cohort. Finally, the association between
AF-related symptoms or quality of life from these assessment
tools with clinical outcomes, including death, hospitalization,
stroke, myocardial infarction, or major bleeding events, has
not been previously evaluated.
Using the Outcomes Registry for Better Informed
Treatment of Atrial Fibrillation (ORBIT-AF), a large, contem-
porary, prospective, community-based outpatient cohort, we
evaluated the type and frequency of symptoms in patients with
AF. In addition, we measured the degree to which symptom
severity (using the EHRA classification system) was corre-
lated with quality of life (assessed by the AFEQT question-
naire) and the association between symptoms or quality of
life with clinical outcomes, including death, hospitalization,
stroke, and major bleeding.
Methods
ORBIT-AF Registry
The ORBIT-AF is an ongoing national, observational, community-
based, registry of outpatients with AF. The ORBIT-AF study has been
described previously.12 The primary data set for this analysis included
baseline data for 10 132 patients >18 years of age with electrocardio-
graphically documented AF collected between June 2010 and August
2011 from 176 sites (82.4% community-based practice groups and
17.6% academic teaching facilities) throughout the United States.
Trained personnel at participating outpatient practices, including in-
ternal medicine, cardiology, and electrophysiology clinics, abstracted
data on consecutive eligible AF patients and submitted them to the
ORBIT-AF registry via Web-enabled case report forms.
Using standard definitions, data include demographic and clini-
cal characteristics, insurance status, education level, medical history
and prior treatments, type of AF, pharmacological treatment strategy,
antithrombotic therapy and monitoring, EHRA symptom score, and
the patient-reported outcome questionnaire.
Study Population
We enrolled 10 132 patients >18 years of age with electrocardiograph-
ically documented AF. For the current analysis, 45 patients without
documented EHRA status were excluded, leaving 10 087 patients
in the study cohort. Patient quality of life data were derived from a
subsample of 2006 (19.8%) subjects administered a patient-reported
outcome questionnaire at the baseline visit. Of the sites in the main
cohort, 58% participated in the quality of life survey. All patients en-
rolled in the main cohort at these sites were approached to complete
the questionnaire on a voluntary basis until the quality of life subsam-
ple enrollment goal was met. Of the patients with documented EHRA
status, 2006 patients were administered the patient-reported out-
come questionnaire. All subjects provided written, informed consent.
Institutional review boards of the Duke Clinical Research Institute and
the participating enrollment sites approved the study.
Physician-Assessed Symptom Burden, Patient-
Reported Symptoms and Quality of Life, and
Outcomes
At baseline, we assessed physician-assessed AF symptom burden
(EHRA score) and patient-reported AF symptoms and quality of
life (AFEQT score). Individual patient-reported symptoms were as-
sessed, including palpitations, syncope or fainting, dyspnea on exer-
tion, exercise intolerance, lightheadedness or dizziness, dyspnea at
rest, fatigue, and chest tightness or discomfort. In addition, physician-
assessed AF symptom severity and burden was assessed by treating
physicians using the EHRA symptom classification defined as as-
ymptomatic (EHRA 1), mild symptoms (EHRA 2), severe symptoms
(EHRA 3), and disabling symptoms (EHRA 4).9,10 Treating physi-
cians were blinded to the patient AFEQT responses when providing
EHRA scores.
Quality of life data were derived from the patient-reported out-
come questionnaires administered to a subsample of ≈20% of pa-
tients from the overall ORBIT-AF registry population. All patients
at sites agreeing to participate in the questionnaire substudy were
approached until the enrollment goals were met. The baseline char-
acteristics were similar between the overall cohort and the subsample
cohort (Table I in the Data Supplement). AF-related quality of life
was measured in this group using the previously validated AFEQT
questionnaire (St Jude Medical, St Paul, MN). The AFEQT is a 20-
item questionnaire assessing 3 domains of AF-related quality of life,
including activity, symptoms, and treatment concerns.11,13 An overall
summary score can be calculated from these domains and was used
for the primary analyses in this study.
Follow-up data collection occurred at 6-month intervals for a
minimum of 2 years. The primary outcomes for this analysis were
all-cause death, first hospitalization, first stroke or transient ischemic
attack, first myocardial infarction, and first major bleeding event us-
ing standard definitions.12 Primary outcome events were verified by
single-source document submission (eg, hospital discharge report)
and central review at the data coordinating center. Outcomes within
the first 2 years of baseline enrollment were analyzed.
Statistical Analysis
All analyses were performed using SAS statistical software (ver-
sion 9.3, SAS Institute, Cary, NC). We compared the baseline
 Freeman et al   Atrial Fibrillation Symptoms and Outcomes   395
characteristics for patients in the ORBIT-AF study population overall
and according to the EHRA symptom classification as asymptom-
atic (EHRA 1) or symptomatic (EHRA ≥2) using the χ2 test for cat-
egorical variables and the Wilcoxon rank-sum test for continuous/
ordinal variables. We also compared the baseline characteristics for
the subset of patients in the ORBIT-AF study population who com-
pleted the AFEQT patient-reported outcome questionnaire overall
and according to quartiles of the AFEQT score using the χ2 test for
categorical variables and the Kruskal–Wallis test for continuous/
ordinal variables. AF symptom prevalence was calculated using fre-
quencies and percentages overall and by each EHRA symptom class,
which were then compared using the χ2 test for categorical variables.
AFEQT overall scores were calculated and compared across EHRA
symptom classes using the nonparametric Kruskal–Wallis test for
continuous variables. The Spearman rank correlation coefficient was
calculated between overall AFEQT score and EHRA class. AFEQT
overall scores were calculated for each AF symptom (present/absent)
and compared using the nonparametric Wilcoxon rank-sum test for
continuous variables.
Using multiply imputed data, Cox frailty regression modeling
(which takes into account clustering of patient outcomes within a
site by adding a random effect for site) was used to examine the as-
sociation between EHRA classes or AFEQT scores at baseline and
clinical outcomes during follow-up. Previously developed models
for each outcome were used for adjustment.14 Briefly, a multiply
imputed data set was used for model construction. Using the first
imputed data set, all continuous variables were evaluated for nonlin-
earity with the outcome while adjusting for 55 candidate variables.
Those which did not meet the linear relationship criteria (P value
<0.05) were accounted for using linear splines. Backward selection
with stay criteria of P value <0.05 was then performed on the first
imputed data set to construct the Cox frailty regression model for
each outcome using the variables from the list of 55 candidate vari-
ables. Rates of missingness were <2% for all candidate variables in
the model, with the following exceptions: level of education (4%),
Downloaded from http://ahajournals.org by on May 9, 2020
eGFR (8%), left ventricular ejection fraction (11%), hematocrit
(11%), and left atrial diameter (14%). All available follow-up data
were used in the model construction process. Imputed values were
obtained by the Markov chain Monte Carlo method or regression
methods.15 The results from each model were then combined to pro-
duce statistically valid inferences when imputed data sets were used.
Adjusted associations for outcomes were displayed as hazard ratios,
95% confidence intervals, and P values. For the evaluation of the
association between AF symptoms and outcomes, the risk of each
outcome for those with an EHRA score of 1 was compared with the
risk of each outcome for those with an EHRA score of ≥2. For the
evaluation of the association between quality of life and outcomes,
the risk of each outcome for those in the highest quartile of AFEQT
was compared with those in each of the other quartiles. A hazard
ratio and P value were generated for each of these comparisons and a
global P value was generated for the highest quartile compared with
all of the other quartiles.
Kaplan–Meier estimates were used to calculate the mortality
curves by EHRA score and AFEQT quartiles. We estimated the cu-
mulative incidence rate by EHRA score and AFEQT quartiles for the
all-cause hospitalization curves. This method accounts for the com-
peting risk of mortality, which makes it impossible for a patient to
experience a subsequent hospitalization.
All statistical analyses for this study were performed using SAS
software (version 9.3, Cary, NC). All P values were 2-sided.
Results
Baseline Characteristics Stratified by Symptom
Status
Among 10 087 adults who had AF between June 2010 and
August 2011, 6235 (61.8%) were symptomatic at baseline as
defined by the physicians-assessed EHRA classification sys-
tem and 3852 (38.2%) were asymptomatic (Table 1).
Patients with AF symptoms were less likely to be male
(53.6% versus 64.3%, P<0.0001). They were also more likely
to have heart failure, chronic obstructive pulmonary disease,
frailty,16 and obstructive sleep apnea. They were more likely to
have a lower (but normal range) left ventricular ejection frac-
tion. Patients with symptoms were more likely to have par-
oxysmal AF (53.8% versus 45.3%, P<0.001). Symptomatic
patients were treated less with angiotensin-converting enzyme
inhibitors, calcium channel blockers, statins, and warfarin, but
more with antiarrhythmic medications. Other clinical differ-
ences were small in absolute terms (<2.5%) yet were statisti-
cally significant because of the large sample size in the study.
Patient-Reported Symptoms and EHRA Symptom
Severity Class
The most common patient-reported symptoms were pal-
pitations (32.7%), dyspnea with exertion (27.6%), fatigue
(26.4%), and lightheadedness or dizziness (20.6%; Table 2).
Less common symptoms included dyspnea at rest (10.3%),
exercise intolerance (10%), chest tightness or discomfort
(9.4%), and syncope (4.5%).
Using the EHRA classification system, 61.8% of patients
had at least mild symptoms and 16.5% had severe or dis-
abling symptoms. Among those identified by physicians
as asymptomatic using the EHRA score, the vast majority
reported no individual symptoms, but 11% reported at least
one symptom, with palpitations being most common (6%).
Among those assessed by physicians as having mild to dis-
abling symptoms (EHRA 2–4), the most common patient-
reported symptoms were palpitations (49.1%), dyspnea with
exertion (43.1%), fatigue (41.1%), and lightheadedness
or dizziness (32.0%). The prevalence of these symptoms
increased ≤2-fold as EHRA symptom severity increased
from mild (EHRA 2) to disabling (EHRA 4). Less common
symptoms among those with EHRA mild to disabling symp-
toms included dyspnea at rest (16.3%), exercise intolerance
(15.6%), chest tightness or discomfort (14.9%), and syn-
cope (6.9%). The prevalence of these symptoms increased
4- to 5-fold as EHRA symptom severity increased from mild
(EHRA 2) to disabling (EHRA 4).
Baseline Characteristics in the Quality of Life
Substudy
In the quality of life substudy, 2007 adults completed the
patient-reported outcome AFEQT questionnaire and 2006
patients had adequate data for subsequent analysis. There were
no substantial differences between the patients in the overall
cohort and the patients in the quality of life substudy popula-
tion (data not shown). Patients in the highest quartile of qual-
ity of life had AFEQT scores of 93.5 to 100, whereas those in
the lowest quartile of quality of life had AFEQT scores of 0 to
65.7 (Table II in the Data Supplement).
Patients in the lowest quartile of quality of life compared
with those in the highest quartile were younger (73 versus 78
years, P<0.001) and more likely to be female (50.5% versus
36.1%, P<0.0001). Those with worse quality of life were
more likely to have a history of peripheral vascular disease,
congestive heart failure, chronic obstructive pulmonary dis-
ease, and obstructive sleep apnea. They were also more likely
 396   Circ Cardiovasc Qual Outcomes   July 2015

Table 1.  Baseline Characteristics of 10 087 Adults With Atrial Fibrillation in the ORBIT-AF Registry Enrolled
Between June 2010 and August 2011 Overall and Stratified by the Presence of Symptoms
Asymptomatic (EHRA 1) Symptomatic (EHRA ≥2)
Characteristic Overall (N=10087) (N=3852) (N=6235) P Value
Demographics
 Age, median, year (IQR) 75 (67–82) 76 (68–82) 74 (66–81) <0.001
 Sex male, % 5814 (57.6) 2475 (64.3) 3339 (53.6) <0.001
 Race, %
  
White 8997 (89.2) 3435 (89.2) 5562 (89.2) <0.001
  
Black 506 (5) 165 (4.3) 341 (5.5)
  
Hispanic 425 (4.2) 196 (5.1) 229 (3.7)
  
Other 143 (1.4) 51 (1.3) 92 (1.5)
Cardiovascular history, %
 Peripheral vascular disease 1344 (13.3) 521 (13.5) 823 (13.2) 0.64
 Sinus node dysfunction 1769 (17.5) 634 (16.5) 1135 (18.2) 0.025
 Stroke/transient ischemic attack 1520 (15.1) 561 (14.6) 959 (15.4) 0.27
 Congestive heart failure 3275 (32.5) 1118 (29) 2157 (34.6) <0.001
 Coronary artery disease 3619 (35.9) 1396 (36.2) 2223 (35.7) 0.55
 Prior myocardial infarction 1591 (15.8) 644 (16.7) 947 (15.2) 0.04
 Prior coronary artery bypass surgery 1476 (14.6) 610 (15.9) 866 (13.9) 0.007
 Prior percutaneous coronary 1723 (17.1) 608 (15.8) 1115 (17.9) 0.007
intervention
Cardiovascular risk factors, %
 Diabetes mellitus 2963 (29.4) 1134 (29.4) 1829 (29.3) 0.91
 Hyperlipidemia 7249 (71.9) 2857 (74.2) 4392 (70.4) <0.001
Downloaded from http://ahajournals.org by on May 9, 2020

 Hypertension 8373 (83) 3217 (83.5) 5156 (82.7) 0.29

Other medical history, %
 Alcohol abuse 407 (4.0) 175 (4.5) 232 (3.7) 0.042
 Cancer 2389 (23.7) 963 (25) 1426 (22.9) 0.015
 Chronic obstructive pulmonary disease 1649 (16.4) 569 (14.8) 1080 (17.3) <0.001
 Frailty 583 (5.8) 152 (4) 431 (6.9) <0.001
 Gastrointestinal bleed 908 (9) 350 (9.1) 558 (9) 0.82
 Obstructive sleep apnea 1832 (18.2) 609 (15.8) 1223 (19.6) <0.001
 Thyroid disease 2264 (22.4) 804 (20.9) 1460 (23.4) 0.003
Implanted devices, %
 Pacemaker 1865 (18.5) 665 (17.3) 1200 (19.2) 0.013
 Implantable cardioverter defibrillator 490 (4.9) 184 (4.8) 306 (4.9) 0.77
 Biventricular pacemaker 91 (0.9) 27 (0.7) 64 (1) 0.093
 Biventricular implantable cardioverter 380 (3.8) 135 (3.5) 245 (3.9) 0.28
defibrillator
Type of AF, %
 New onset 477 (4.7) 137 (3.6) 340 (5.5) <0.001
 Paroxysmal 5096 (50.5) 1743 (45.3) 3353 (53.8)
 Persistent 1695 (16.8) 653 (17) 1042 (16.7)
 Long-standing persistent 2819 (28) 1319 (34.2) 1500 (24.1)
Vital statistics and vital signs (IQR)
 Body mass index, median, kg/m2 29.1 (25.4–34) 29 (25.6–33.5) 29.2 (25.2–34.4) 0.42
 Heart rate, median, bpm 70 (63–80) 70 (62–79) 70 (63–80) 0.004
 Blood pressure–systolic, 126 (116–138) 126 (118–138) 124 (115–138) 0.001
median, mm Hg
(Continued)
 Freeman et al   Atrial Fibrillation Symptoms and Outcomes   397

Table 1.  Continued

Asymptomatic (EHRA 1) Symptomatic (EHRA ≥2)
Characteristic Overall (N=10087) (N=3852) (N=6235) P Value

Echocardiography and
laboratory data (IQR)
 Left ventricular ejection fraction, 55 (50–61) 57 (50–62) 55 (50–60) <0.001
median, %
 Estimated glomerular filtration rate, 67 (52.7–82.3) 67.6 (53.6–82.6) 66.6 (52.3–82.1) 0.20
median, mg/dL
 Hemoglobin, median, g/dL 13.5 (12.3–14.6) 13.6 (12.4–14.7) 13.4 (12.2–14.6) <0.001
Cardiac medications, %
 Aldosterone antagonist 555 (5.5) 200 (5.2) 355 (5.7) 0.28
 Angiotensin-converting enzyme 3567 (35.4) 1485 (38.6) 2082 (33.4) <0.001
inhibitor
 Angiotensin receptor blocker 1800 (17.8) 661 (17.2) 1139 (18.2) 0.16
 Antiarrhythmic therapy 2894 (28.7) 862 (22.4) 2032 (32.6) <0.001
 Antiplatelet therapy 4757 (47.2) 1771 (46) 2986 (47.9) 0.061
 Beta-blockers 6448 (63.9) 2437 (63.3) 4011 (64.3) 0.27
 Calcium channel blockers 3038 (30.1) 1234 (32) 1804 (28.9) 0.001
 Digoxin 2365 (23.5) 885 (23) 1480 (23.7) 0.38
 Diuretic 4938 (49) 1862 (48.3) 3076 (49.3) 0.32
 Statin 5553 (55.1) 2223 (57.7) 3330 (53.4) <0.001
 Warfarin 7192 (71.3) 2830 (73.5) 4362 (70) <0.001
AF indicates atrial fibrillation; EHRA, European Heart Rhythm Association; IQR, interquartile range; and ORBIT-AF, Outcomes Registry for
Better Informed Treatment of AF.

to have new onset AF and less likely to have paroxysmal or highest among those in the lowest quartile of quality of life
Downloaded from http://ahajournals.org by on May 9, 2020

long-standing persistent AF. They had higher body weights
and were more likely to be taking aldosterone antagonists and
diuretics. Other clinical differences were small in absolute
terms (<2.5%), yet were statistically significant because of the
large sample size in the study.
Atrial Fibrillation Effect on Quality-of-Life
Questionnaire Score
As with the overall cohort, the most common symptoms in the
subgroup of patients who completed the AFEQT survey were
palpitations (33.0%), dyspnea with exertion (32.6%), fatigue
(32.0%), and lightheadedness or dizziness (27.3%; Table III in
the Data Supplement). The prevalence of every symptom was
and decreased with increasing quality of life.
All of the symptoms assessed were associated with a sta-
tistically significant decrease in quality of life (P<0.001) with
a decrement of AFEQT score ranging from 9 to 17 (Figure 1).
Dyspnea at rest, exercise intolerance, and chest discomfort or
tightness were associated with the largest decreases in quality
of life as measured by the AFEQT.
Correlation of EHRA Classification and Quality of
Life
Patients assessed by physicians to be asymptomatic
(EHRA=1) had the highest quality of life as measured by
the AFEQT, and the AFEQT score decreased with increasing

Table 2.  Atrial Fibrillation–Associated Symptoms in 10 087 Adults With Atrial Fibrillation in the ORBIT-AF Registry Enrolled
Between June 2010 and August 2011 Overall and Stratified by European Heart Rhythm Association (EHRA) Symptom Severity
Classification
Asymptomatic Mild (EHRA=2) Severe (EHRA=3) Disabling (EHRA=4)
Atrial Fibrillation Symptom Overall (N=10 087) (EHRA=1) (N=3852) (N=4575) (N=1474) (N=186) P Value
Palpitations 3296 (32.7) 232 (6) 2135 (46.7) 817 (55.4) 112 (60.2) <0.001
Dyspnea with exertion 2779 (27.6) 95 (2.5) 1756 (38.4) 817 (55.4) 111 (59.7) <0.001
Fatigue 2664 (26.4) 102 (2.7) 1604 (35.1) 840 (57.0) 118 (63.4) <0.001
Lightheadedness/dizziness 2081 (20.6) 86 (2.2) 1254 (27.4) 650 (44.1) 91 (48.9) <0.001
Dyspnea at rest 1040 (10.3) 26 (0.7) 499 (10.9) 414 (28.1) 101 (54.3) <0.001
Exercise intolerance 1005 (10) 33 (0.9) 485 (10.6) 411 (27.9) 76 (40.9) <0.001
Chest tightness/discomfort 948 (9.4) 19 (0.5) 509 (11.1) 337 (22.9) 83 (44.6) <0.001
Syncope/fainting 455 (4.5) 23 (0.6) 215 (4.7) 167 (11.4) 50 (26.9) <0.001
EHRA indicates European Heart Rhythm Association; and ORBIT-AF, Outcomes Registry for Better Informed Treatment of AF.
 398   Circ Cardiovasc Qual Outcomes   July 2015
EHRA symptom severity class with a Spearman’s correlation
coefficient of −0.39 (Figure 2). Those assessed by physicians
as asymptomatic (EHRA=1) had a median AFEQT score of
90 (interquartile range [IQR] 79–97), those with mild symp-
toms (EHRA=2) had a median score of 81 (IQR 67–91), those
with severe symptoms (EHRA=3) had a median score of 63
(IQR 48–82), and those with disabling symptoms (EHRA=4)
had a median score of 61 (IQR 41–78).
AF-Related Symptoms and Quality of Life and
Associations With Outcomes
Over a median of 1.8 years (IQR 1.4–2.1 years) of follow-up
in 9600 adults with follow-up data, patients with AF-related
symptoms at baseline (EHRA≥2) had a higher risk of hospi-
talization (adjusted hazard ratio 1.23, 95% confidence inter-
val 1.15–1.31, P=<0.001, Table 3, Figure 3)) and a borderline
higher risk of major bleeding (adjusted hazard ratio 1.21,
95% confidence interval 1.02–1.45, P=0.03). In contrast, AF-
related symptoms at baseline were not associated with sig-
nificant differences in the risk of death (Figure 4), stroke, or
myocardial infarction.
Among 1925 adults with follow-up data, patients in the
lowest quartile of quality of life at baseline (AFEQT ≤65.7)
had a higher risk of hospitalization compared with those in the
highest quartile (AFEQT >93.1; adjusted hazard ratio 1.49,
95% confidence interval 1.2–1.84, P≤0.001, Table 4, Figure I
in the Data Supplement). This finding was consistent for the
comparison of the second and third quartiles compared with
highest quartile of quality of life, and the global P value for all
Downloaded from http://ahajournals.org by on May 9, 2020
of the other quartiles compared with the highest quartile was
0.001 (Table 4). In addition, this finding was consistent when
Figure 1. Boxplot of overall Atrial Fibrillation Effect on Quality-of-Life questionnaire (AFEQT) score stratified by European Heart Rhythm
Association (EHRA) symptom class (correlation coefficient =−0.40, P<0.0001). AF indicates atrial fibrillation.
AFEQT was modeled as a continuous variable (Table IV in
the Data Supplement). In contrast, quality of life at baseline
was not associated with significant differences in the risk of
death (Figure II in the Data Supplement) or major bleeding.
The number of events for stroke/transient ischemic attack and
myocardial infarction did not differ substantially between
quartiles of quality of life, but they were not adequate to per-
form our regression models for risk assessment.
Discussion
In a large, diverse, community-based cohort of adults with
AF, we found that most patients (61.8%) were symptomatic
as measured by the EHRA classification system. The most
common symptoms were palpitations, dyspnea with exertion,
fatigue, and lightheadedness or dizziness. Symptom severity
measured by physicians using the EHRA classification was
correlated with decreasing patient-reported quality of life as
measured by the AFEQT questionnaire. AF symptoms were
associated with a higher risk of hospitalization and a border-
line higher risk of major bleeding, and decreased quality of
life was associated with a higher risk of hospitalization, but
there were no differences in the risk of death or other major
adverse events.
This study is novel in its use of the EHRA AF symptom
classification system in a large, community-based cohort and its
demonstration of a correlation between physician-assessed AF
symptom status and patient-reported quality of life. Although
most patients in our cohort were assessed by physicians as
symptomatic (EHRA 2–4), 38.2% were reported to be asymp-
tomatic (EHRA 1). Interestingly, 11% of those assessed as
asymptomatic by physicians (EHRA 1) reported experiencing
 Freeman et al   Atrial Fibrillation Symptoms and Outcomes   399

Figure 2. Atrial Fibrillation Effect on Quality-of-Life questionnaire (AFEQT) score stratified by the presence or absence of each atrial
fibrillation–related symptom. EHRA indicates European Heart Rhythm Association.

individual symptoms, with palpitations being the most common,
suggesting that physicians may underestimate patient symp-
toms when they are mild. Regardless, this frequency of asymp-
tomatic patients is substantially higher than has previously
been reported, likely because of selection biases in prior stud-
ies.5–8 Prior studies reporting higher symptom burden included
selected patients from referral-based practices, and therefore our
community-based population likely more accurately represents
Downloaded from http://ahajournals.org by on May 9, 2020
or tightness were associated with the largest decreases. This is
consistent with our finding that these symptoms, although rel-
atively rare overall, were common among those who reported
severe or disabling symptoms (EHRA 3–4) in our cohort and
deserve special attention by treating clinicians.
We demonstrated an inverse correlation between the
EHRA AF symptom severity classification system and quality
of life as measured by the AFEQT in our population. This con-

the prevalence of symptoms in outpatients with AF.
The AFEQT questionnaire11 has been validated as a dis-
ease-specific tool for assessing patient quality of life in AF,
but it has not previously been used in a large community-based
cohort. All of the symptoms assessed in our study were associ-
ated with statistically significant decreases in quality of life,
but dyspnea at rest, exercise intolerance, and chest discomfort
sistency across test instruments supports their clinical use for
assessment of AF symptom burden and quality of life. We also
demonstrated that those with mild symptoms (EHRA 2) had an
AFEQT score of <90 and those with severe or disabling symp-
toms (EHRA 3–4) had an AFEQT score of <64. These AFEQT
thresholds may serve as clinically significant quality of life cat-
egories for future clinical trials or observational studies.

Table 3.  Association Between Symptoms and Outcomes in 9600 Adults With Atrial Fibrillation in the ORBIT-AF Registry*
Asymptomatic (EHRA=1) Symptomatic (EHRA≥2)
events (events per 100 events (events per 100
Outcome patient-years) (N=3682) patient-years) (N=5918) Unadjusted HR (95% CI) Adjusted HR (95% CI) Adjusted P Value
Death 311 (5.0) 561 (5.7) 1.16 (1.00–1.34) 1.00 (0.86–1.16) 0.98
First stroke/TIA 99 (1.6) 168 (1.7) 1.12 (0.86–1.45) 1.13 (0.87–1.46) 0.37
First myocardial infarction 47 (0.8) 79 (0.8) 1.09 (0.75–1.58) 1.05 (0.72–1.53) 0.80
First hospitalization 1495 (30.8) 2812 (40.1) 1.32 (1.24–1.42) 1.23 (1.15–1.31) <0.001
First major bleeding 219 (3.6) 397 (4.2) 1.29 (1.08–1.53) 1.21 (1.02–1.45) 0.031
AF indicates atrial fibrillation; BMI, body mass index; CAD, coronary artery disease; CHF, congestive heart failure; CI, confidence interval; COPD, chronic obstructive
pulmonary disease; eGFR, estimated glomerular filtration rate; EHRA, European Heart Rhythm Association; HR, hazard ratio; LAD, left anterior descending artery; LVEF,
left ventricular ejection fraction; MI, myocardial infarction; ORBIT-AF, Outcomes Registry for Better Informed Treatment of AF; OSA, obstructive sleep apnea; PCI,
percutaneous intervention; and TIA, transient ischemic attack.
*Variables included in the adjusted models (baseline covariates): Death adjusted for level of education, rhythm control, cognitive impairment/dementia, hyperlipidemia,
linear spline eGFR ≤80, linear spline eGFR >80, LAD bypass graft type, cancer, diastolic blood pressure, intraventricular conduction delay, frailty, height, heart rate,
hematocrit, diabetes mellitus, smoking, linear spline systolic blood pressure ≤120, COPD, BMI, sex, CHF, age, functional status. Stroke adjusted for peripheral vascular
disease, race, rhythm control, AF type, AV node/His bundle ablation, female, hypertension, age, history of stroke/TIA. MI adjusted for diabetes mellitus, eGFR, peripheral
vascular disease, history of CAD, LVEF. Hospitalization for linear spline age ≤70, linear spline age >70, BMI, weight, osteoporosis, height, PCI, cancer, OSA, anemia,
frailty, insurance status, history of CAD, site specialty, prior antiarrhythmic drug use, peripheral vascular disease, functional status, linear spline heart rate >68, diabetes
mellitus, hematocrit, linear spline eGFR ≤80, COPD, diastolic blood pressure truncated above at 70, EHRA score, CHF. Major bleeding adjusted for COPD, OSA, LAD type,
cancer, functional status, level of education, intraventricular conduction delay, rhythm control, smoking status, significant valvular disease, eGFR, insurance status,
history of gastrointestinal bleed, anemia, hematocrit.
 400   Circ Cardiovasc Qual Outcomes   July 2015
Finally, our study demonstrated an association between physi-
cian-assessed AF symptom burden (EHRA) and patient-reported
decrease in quality of life and a higher risk of hospitalization.
We did not show any differences in the risk of death or other
major adverse events, except for a borderline association between
AF symptoms assessed by EHRA class and hospitalization. This
finding is noteworthy because the patients with the highest bur-
den of symptoms and the lowest quality of life were substantially
Downloaded from http://ahajournals.org by on May 9, 2020
younger and healthier than less symptomatic patients and those
with higher self-reported quality of life. These patients did not
have an increased risk of death, stroke, myocardial infarction, or
major bleeding, suggesting that their hospitalizations were likely
related to their symptoms and lower quality of life and not major
adverse events. Interventions targeted at improvement in symp-
toms and quality of life in these patients, including more aggres-
sive outpatient follow-up or rhythm control therapies may be
important for minimizing resource utilization in this population.
Figure 3. Cumulative incidence curves for
hospitalization in asymptomatic patients
(EHRA=1) compared with symptomatic
patients (EHRA≥2) in 9600 adults with
atrial fibrillation in the Outcomes Registry
for Better Informed Treatment of AF
(ORBIT-AF) Registry. EHRA indicates
European Heart Rhythm Association.
Our study has important limitations. Although our large,
community-based cohort is broadly representative of patients
with AF in the United States, we evaluated patients who are
undergoing treatment for AF, and the results may not be gen-
eralizable to a disadvantaged or untreated population. In addi-
tion, the population of patients who chose to participate in the
quality of life substudy may not be fully generalizable to a
general population with AF. Finally, despite controlling for
a large number of covariates, we cannot exclude residual or
unmeasured confounding in our analyses, evaluating the asso-
ciation between AF symptoms or quality of life and outcomes.
Conclusions
In summary, in a large nationally representative community-
based cohort of individuals with AF, most patients were
reported by physicians to be symptomatic (61.8%) and a sub-
stantial minority (16.6%) had severe or disabling symptoms
Figure 4. Kaplan–Meier curves for
mortality in asymptomatic patients
(EHRA=1) compared with symptomatic
patients (EHRA≥2) in 9600 adults with
atrial fibrillation in the Outcomes Registry
for Better Informed Treatment of AF
(ORBIT-AF) Registry. EHRA indicates
European Heart Rhythm Association.
 Freeman et al   Atrial Fibrillation Symptoms and Outcomes   401

Table 4.  Association Between Quality of Life and Outcomes in 1925 Adults With Atrial Fibrillation in the ORBIT-AF Registry*
First Quartile QoL Second Quartile Third Quartile QoL Fourth Quartile
(AFEQT ≤65.7) QoL (AFEQT 66.7– (AFEQT 81.9– QoL (AFEQT≥93.5)
(Events per 100 81.5) (Events per 93.1) (Events per (Events per 100 Adjusted HR First Adjusted HR Adjusted HR Third
Patient-Years) 100 Patient- 100 Patient- Patient-Years) vs Fourth Quartile Second vs Fourth vs Fourth Quartile Adjusted
Outcome (N=468) Years) (N=488) Years) (N=480) (N=489) (95% CI) Quartile (95% CI) (95% CI) Global P Value
Death 45 (5.2) 42 (4.5) 36 (3.8) 43 (4.4) 1.21 (0.77–1.90) 1.04 (0.67–1.62) 0.98 (0.62–1.55) 0.81
First hospitalization 247 (44.0) 249 (39.3) 233 (34.4) 199 (26.0) 1.49 (1.20–1.84) 1.42 (1.17–1.74) 1.33 (1.09–1.62) 0.001
First stroke/TIA† 10 (1.2) 14 (1.5) 8 (0.9) 13 (1.3) - - -
First myocardial 8 (0.9) 11 (1.2) 5 (0.53) 3 (0.3) - - -
infarction†
First major bleeding 30 (3.6) 36 (4.0) 28 (3.0) 23 (2.4) 1.47 (0.83–2.58) 1.72 (1.0–2.94) 1.39 (0.79–2.44) 0.27
AF indicates atrial fibrillation; BMI, body mass index; CAD, coronary artery disease; CHF, congestive heart failure; CI, confidence interval; COPD, chronic obstructive
pulmonary disease; eGFR, estimated glomerular filtration rate; EHRA, European Heart Rhythm Association; HR, hazard ratio; LAD, left anterior descending artery; LVEF,
left ventricular ejection fraction; MI, myocardial infarction; ORBIT-AF, Outcomes Registry for Better Informed Treatment of AF; OSA, obstructive sleep apnea; PCI,
percutaneous intervention; and TIA, transient ischemic attack.
*Variables included in the adjusted models (baseline covariates): Death adjusted for level of education, rhythm control, cognitive impairment/dementia, hyperlipidemia,
linear spline eGFR ≤80, linear spline eGFR >80, LAD bypass graft type, cancer, diastolic blood pressure, intraventricular conduction delay, frailty, height, heart rate,
hematocrit, diabetes mellitus, smoking, linear spline systolic blood pressure ≤120, COPD, BMI, sex, CHF, age, functional status. Stroke adjusted for peripheral vascular
disease, race, rhythm control, AF type, AV node/His bundle ablation, female, hypertension, age, history of stroke/TIA. MI adjusted for diabetes mellitus, eGFR, peripheral
vascular disease, history of CAD, LVEF. Hospitalization for linear spline age ≤70, linear spline age >70, BMI, weight, osteoporosis, height, PCI, cancer, OSA, anemia,
frailty, insurance status, history of CAD, site specialty, prior antiarrhythmic drug use, peripheral vascular disease, functional status, linear spline heart rate >68, diabetes
mellitus, hematocrit, linear spline eGFR ≤80, COPD, diastolic blood pressure truncated above at 70, EHRA score, CHF. Major bleeding adjusted for COPD, OSA, LAD type,
cancer, functional status, level of education, intraventricular conduction delay, rhythm control, smoking status, significant valvular disease, eGFR, insurance status,
history of gastrointestinal bleed, anemia, hematocrit.
†There were not enough events for the unadjusted and adjusted models for stroke/TIA and MI.

(EHRA 3–4). The most common patient-reported symptoms
reported in our cohort were palpitations, dyspnea with exer-
tion, fatigue, and lightheadedness or dizziness. We demon-
Downloaded from http://ahajournals.org by on May 9, 2020
Affairs, TEVA Pharmaceuticals, Boston Scientific, outside the sub-
mitted work. E.M. Hylek is advisory capacity for Bayer, Boehringer
Ingelheim, Bristol Myers Squibb, Daiichi Sankyo, Janssen, Pfizer, and

strated that physician-assessed AF symptom burden using the
European Heart Rhythm Association (EHRA) classification
system correlated with decreased patient-reported quality of
life as measured by the AFEQT questionnaire, supporting
the clinical use of both instruments. Finally, our study dem-
onstrated an association between AF-related symptoms or
decreased quality of life and a higher risk of hospitalization,
although not mortality, thus identifying an important popula-
tion of patients who may require aggressive interventions to
minimize symptoms and lower resource utilization.
Sources of Funding
This work was supported by a grant from Janssen Scientific Affairs,
LLC. The funder was not involved in the design and conduct of the
study; collection, management, analysis, and interpretation of the
data; or preparation, review, or approval of the article.
Disclosures
J.V Freeman is consultant of Janssen Scientific (modest). D.N. Simon
receives personal fees from Janssen Scientific Affairs. J. Spertus
received grants and contracts from National Institute of Health,
American College of Cardiology Foundation, Abbott Vascular,
Genentech, Amorcyte; Was consultant for Janssen, United Healthcare,
Novartis, and Amgen; and received equity interest in Health Outcomes
Sciences, and this company owns the copyright to the Seattle Angina
Questionnaire, the Kansas City Cardiomyopathy Questionnaire, the
Peripheral Artery Questionnaire, and the Atrial Fibrillation Effect on
QualiTy-of-life (AFEQT) questionnaire (licensed by St. Jude Medical).
G.C. Fonarow is consultant for Janssen Scientific Affairs (mod-
est). B.J. Gersh received personal fees from Medtronic, Inc, Baxter
Healthcare Corporation, Cardiovascular Research Foundation, Merck
& Co, Inc, St Jude Medical, Inc, Ortho-McNeil Janssen Scientific
Roche. P.R. Kowey received grants and personal fees from Johnson
& Johnson not related to the submitted work. K.W. Mahaffey re-
ceived grants and personal fees from Johnson & Johnson, grants from
Regeneron, grants and personal fees from Cubist Pharmaceuticals,
grants and personal fees from Sanofi, grants from Baxter, grants from
Roche Diagnostics, grants from Ikaria, grants from Amgen, grants
from Regado, grants and personal fees from Merck, grants and per-
sonal fees from Glaxo Smith Kline, grants from Amylin, grants from
Novartis, grants and personal fees from AstraZeneca, grants from
Portola, grants and personal fees from Eli Lilly, grants from Edwards
Lifesciences, grants and personal fees from Boehringer Ingelhein,
grants from National Institute of Health, grants from National Heart,
Lung & Blood Institute, grants from National Institute of Allergy &
Infectious Diseases, personal fees from Bayer, personal fees from
Biotronik, personal fees from Daiichi Sankyo, personal fees from
Gilead Sciences, personal fees from Medtronic, personal fees from
Ortho/McNeill, personal fees from Pfizer, personal fees from St Jude,
personal fees from ACC, personal fees from John Hopkins University,
personal fees from South East Area Health Education Center, personal
fees from Sun Pharma, grants and personal fees from Bristol Myers-
Squibb, personal fees from Duke Center for Educational Excellence,
personal fees from University of British Columbia, personal fees
from WebMD, personal fees from Perdue Pharma, personal fees from
Dialouges, personal fees from Springer Publishing, personal fees from
Haemonetics, personal fees from Forest, personal fees from Amgen,
personal fees from Elsevier during the conduct of the study; and Other
Relationships: http://www.dcri.duke.edu/research/coi.jsp and www.
med.stanford.edu/profiles/Kenneth_Mahaffey. P. Chang received sub-
stantial personal fees from Janssen Pharmaceuticals, Inc. during the
conduct of the study; other from Janssen Pharmaceuticals, Inc., out-
side the submitted work. E.D. Peterson received grants from Eli Lilly,
grants and personal fees from Janssen Scientific Affairs, personal fees
from Boehringer Ingelheim. J.P. Piccini received grants for clinical
research from ARCA biopharma, Boston Scientific, GE Healthcare,
Janssen Pharmaceuticals, and ResMed. J.P. Piccini is a consultant to
Johnson & Johnson and Spectranetics.
 402   Circ Cardiovasc Qual Outcomes   July 2015
References
1. Lakshminarayan K, Solid CA, Collins AJ, Anderson DC, Herzog CA.
Atrial fibrillation and stroke in the general medicare population: a 10-year
perspective (1992 to 2002). Stroke. 2006;37:1969–1974. doi: 10.1161/01.
STR.0000230607.07928.17.
2. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk
factor for stroke: the Framingham Study. Stroke. 1991;22:983–988.
3. Benjamin EJ, Wolf PA, D’Agostino RB, Silbershatz H, Kannel WB, Levy
D. Impact of atrial fibrillation on the risk of death: the Framingham Heart
Study. Circulation. 1998;98:946–952.
4. Thrall G, Lane D, Carroll D, Lip GY. Quality of life in patients with atrial
fibrillation: a systematic review. Am J Med. 2006;119:448.e1–448.19. doi:
10.1016/j.amjmed.2005.10.057.
5. Lévy S, Maarek M, Coumel P, Guize L, Lekieffre J, Medvedowsky JL,
Sebaoun A. Characterization of different subsets of atrial fibrillation
in general practice in France: the ALFA study. The College of French
Cardiologists. Circulation. 1999;99:3028–3035.
6. Flaker GC, Belew K, Beckman K, Vidaillet H, Kron J, Safford R, Mickel
M, Barrell P; AFFIRM Investigators. Asymptomatic atrial fibrillation: de-
mographic features and prognostic information from the Atrial Fibrillation
Follow-up Investigation of Rhythm Management (AFFIRM) study. Am
Heart J. 2005;149:657–663. doi: 10.1016/j.ahj.2004.06.032.
7. Kerr C, Boone J, Connolly S, Greene M, Klein G, Sheldon R, Talajic M.
Follow-up of atrial fibrillation: The initial experience of the Canadian
Registry of Atrial Fibrillation. Eur Heart J. 1996;17(suppl C):48–51.
8. Rienstra M, Lubitz SA, Mahida S, Magnani JW, Fontes JD, Sinner MF,
Van Gelder IC, Ellinor PT, Benjamin EJ. Symptoms and functional
status of patients with atrial fibrillation: state of the art and future re-
search opportunities. Circulation. 2012;125:2933–2943. doi: 10.1161/
CIRCULATIONAHA.111.069450.
9. Kirchhof P, Auricchio A, Bax J, Crijns H, Camm J, Diener HC, Goette
A, Hindricks G, Hohnloser S, Kappenberger L, Kuck KH, Lip GY,
Olsson B, Meinertz T, Priori S, Ravens U, Steinbeck G, Svernhage E,
Tijssen J, Vincent A, Breithardt G. Outcome parameters for trials in atrial
Downloaded from http://ahajournals.org by on May 9, 2020
fibrillation: recommendations from a consensus conference organized by
the German Atrial Fibrillation Competence NETwork and the European
Heart Rhythm Association. Europace. 2007;9:1006–1023. doi: 10.1093/
europace/eum191.
10. European Heart Rhythm A, European Association for Cardio-Thoracic
S, Camm AJ, Kirchhof P, Lip GY, Schotten U, Savelieva I, Ernst S, Van
Gelder IC, Al-Attar N, Hindricks G, Prendergast B, Heidbuchel H, Alfieri
O, Angelini A, Atar D, Colonna P, De Caterina R, De Sutter J, Goette A,
Gorenek B, Heldal M, Hohloser SH, Kolh P, Le Heuzey JY, Ponikowski P,
Rutten FH. Guidelines for the management of atrial fibrillation: The task
force for the management of atrial fibrillation of the european society of
cardiology (esc). Eur Heart J. 2010;31:2369–2429.
11. Spertus J, Dorian P, Bubien R, Lewis S, Godejohn D, Reynolds MR,
Lakkireddy DR, Wimmer AP, Bhandari A, Burk C. Development and
validation of the Atrial Fibrillation Effect on QualiTy-of-Life (AFEQT)
Questionnaire in patients with atrial fibrillation. Circ Arrhythm
Electrophysiol. 2011;4:15–25. doi: 10.1161/CIRCEP.110.958033.
12. Piccini JP, Fraulo ES, Ansell JE, Fonarow GC, Gersh BJ, Go AS, Hylek
EM, Kowey PR, Mahaffey KW, Thomas LE, Kong MH, Lopes RD,
Mills RM, Peterson ED. Outcomes registry for better informed treat-
ment of atrial fibrillation: rationale and design of ORBIT-AF. Am Heart J.
2011;162:606–612.e1. doi: 10.1016/j.ahj.2011.07.001.
13. Dorian P, Burk C, Mullin CM, Bubien R, Godejohn D, Reynolds MR,
Lakkireddy DR, Wimmer AP, Bhandari A, Spertus J. Interpreting changes
in quality of life in atrial fibrillation: how much change is meaningful? Am
Heart J. 2013;166:381–387.e8. doi: 10.1016/j.ahj.2013.04.015.
14. Ripatti S, Palmgren J. Estimation of multivariate frailty models using pe-
nalized partial likelihood. Biometrics. 2000;56:1016–1022.
15. Enders CK. Applied Missing Data Analysis. Guilford Press; 2010.
16. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J,
Seeman T, Tracy R, Kop WJ, Burke G, McBurnie MA; Cardiovascular
Health Study Collaborative Research Group. Frailty in older
adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci.
2001;56:M146–M156.