o

Racial/Ethnic Differences in the

Prevalence of Atrial Fibrillation Among
Older AdultsA Cross-Sectional Study
Albert Yuh-Jer Shen, MD, MS; Richard Contreras, MS; Serap Sobnosky, MD; Ahmed I. Shah, MD;
Anne M. Ichiuji, MD; Michael B. Jorgensen, MD; Somjot S. Brar, MD; Wansu Chen, MS

Background: Atrial fibrillation affects 4% to 8% of individuals

over 60 years of age based on studies of predominantly white
populations, whether this is true among nonwhite individuals is
not clear. This study was undertaken to define racial/ethnic differences in atrial fibrillation prevalence among a large community cohort.
Methods: This is a cross-sectional study. In 2008, there were
430317 members aged 60 years or older in a large California health maintenance organization. By searching International Classification of Diseases, Ninth Revision codes and
electronic electrocardiographic archives, we identified all
members in this age group with primary, nonvalvular atrial
fibrillation. Race/ethnicity data were assigned using health
plan enrollment, service utilization, Asian/Hispanic surname
and geocoding methods, and was available for 80.5% of
members (79.8% of nonatrial fibrillation and 92% of atrial fibrillation), 99% of which were white, black, Asian, or Hispanic.
We assessed the age- and gender-specific atrial fibrillation
prevalence rates for each racial/ethnic group. The effect of
race/ethnicity on atrial fibrillation was analyzed with logistic
regression methods adjusting for potential confounders.
Results: The overall atrial fibrillation prevalence was 5.3%.
Among members with assigned race/ethnicity data, the
prevalence among whites, blacks, Asians, and Hispanics
was 8.0%, 3.8%, 3.9%, and 3.6%, respectively. The adjusted
odds ratios (95% confidence intervals) of atrial fibrillation
among blacks, Asians, and Hispanics with whites as referent
were 0.49 (0.47-0.52), 0.68 (0.64-0.72), and 0.58 (0.55-0.61),
respectively.
Conclusions: Atrial fibrillation is less prevalent in older nonwhite individuals than whites. White race/ethnicity is associated with significantly greater odds for atrial fibrillation
compared to blacks, Asians, and Hispanics, after adjusting
for comorbidities associated with the development of atrial
fibrillation.
Keywords: race/ethnicity n atrial fibrillation
J Natl Med Assoc. 2010;102:906-913

906 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

Author Affiliations: Department of Cardiology and the Center for Medical

Education and Research, Kaiser Permanente Medical Center, Los Angeles,
California (Drs Shen, Sobnosky, Shah, Ichiuji, Jorgensen, and Brar);
Department of Research and Evaluation, Kaiser Permanente Southern
California, Pasadena, California (Messers Contreras and Chen).
Correspondence: Albert Yuh-Jer Shen, MD, MS; Department of Cardiology,
Kaiser Permanente Medical Center; 1526 Edgemont St, 2nd Floor, Los
Angeles, CA 90027 (albert.y-j.shen@kp.org).

Introduction

trial fibrillation is the most common clinically relevant arrhythmia and affects approximately 1% of the general adult population.1
Extrapolations from population studies have estimated
that approximately 2.3 million Americans have atrial
fibrillation.1,2 Atrial fibrillation is associated with a 4to 5-fold risk of stroke3 and a 2-fold risk of death even
after adjustment for traditional cardiovascular risk factors.4 Atrial fibrillation prevalence is highly dependent
on age: it is present in less than 0.5% of individuals
younger than 50 years, 6% in those older than 65 years,
and approximately 10% in those aged 80 years or older,
with the median age being 75 years. Studies have suggested that the incidence and prevalence of atrial fibrillation has increased over the past few decades.5, 6
Population studies of atrial fibrillation prevalence
have included predominantly white patients, while estimation of national atrial fibrillation burden requires
knowledge of the disease status in diverse racial/ethnic
populations. According to the US Census Bureau,
Hispanics, blacks, and Asians accounted for 15%,
12.8%, and 4.4%, respectively, of the 2006 US population. In California, non-Hispanic whites constitute less
than half (42.7%) of the states populace.7 Prior studies
have suggested that the incidence and prevalence of
atrial fibrillation may be substantially lower among
some nonwhite races/ethnicities than in whites in certain
populations or disease states.2,8-11 The goals of this study
were to assess the prevalence of atrial fibrillation among
the major racial/ethnic groups in a large, multiethnic
community cohort, and to determine whether race/ethnicity is associated with atrial fibrillation.
VOL. 102, NO. 10, OCTOBER 2010

Race/ethnicity and atrial fibrillation

Methods

This is a cross-sectional study of atrial fibrillation

prevalence within the 2008 Kaiser Permanente Southern
California (KPSC) Health Plan membership, a large
integrated health maintenance organization that served
approximately 3.2 million members in that year. KPSC
members comprise approximately 15% of the regions
population. Membership demographics, socioeconomic
status, and racial/ethnic composition are representative
of California (administrative data). Due to the very low
prevalence of atrial fibrillation among younger individuals, we limited our study to members aged 60 years or
older. To be included for analysis, a patient must have
been an active member on July 1, 2008, and continuously enrolled for the prior 12 months except for a gap
of 93 or fewer days. The study protocol complies with
the declaration of Helsinki and was approved by the
local institutional review Board. Waiver of informed
consent was granted.

Atrial Fibrillation Ascertainment

By searching International Classification of Disease,
Ninth Revision, Clinical Modification (ICD-9-CM)
codes (427.31) and our electronic electrocardiographic
(ECG) archives, we identified all members aged 60
years or older with any atrial fibrillation diagnosis documented in the period 2006-2008 in the KPSC clinical
systems or out-of-plan claims. Outpatient diagnoses
were physician coded, while hospitalized patients diagnoses were coded by professional coders based on
review and query of physician diagnoses. We have previously reviewed the medical records, including electrocardiograms of 100 randomly sampled hospitalized
patients, and found a 96% positive predictive value of
ICD-9 codes in diagnosing atrial fibrillation within
KPSC.12 Since the 1990s, all ECGs performed within the
KPSC system have been archived in the MUSE
Cardiology Information System (GE Marquette,
Milwaukee, Wisconsin). The vast majority (>90%) of
ECGs have been confirmed by a cardiologist reading
following computer interpretation. We queried the system and identified all patients with an ECG diagnosis of
atrial fibrillation in 2008. Patients with atrial fibrillation
documented by ECG or ICD-9 codes, or both were

defined as having atrial fibrillation. We excluded from

analysis any patient with any diagnosis of pulmonary
embolism, hyperthyroidism, pericardial disease, rheumatic and nonrheumatic aortic and mitral valve diseases, since the cause and effect of these comorbidities
and atrial fibrillation could not be ascertained from our
data. Atrial fibrillation cases that occurred only within
the 30 days after cardiac surgery with no subsequent
record of atrial fibrillation were also excluded.

Covariate Ascertainment
and Exclusion Criteria
By searching for relevant ICD-9 and Current
Procedural Terminology codes in our administrative
datasets, we identified patients major comorbidities
known to be associated with atrial fibrillation,13 including diabetes, hypertension, heart failure, and coronary
artery disease. We searched the 2 years preceding 2008
to allow sufficient time for documentation of comorbidities. The accuracy of the ICD-9 codes for comorbidities
within the KPSC system was validated by manual review
of 200 medical records (50 randomly selected from each
racial/ethnic group) with a k statistic of 0.82-0.94.
Race/ethnicity data was first assigned using health
plan enrollment and health care services utilization
administrative data, and was available for 76% of members aged 60 years or older (90% and 73% of atrial fibrillation and nonatrial fibrillation members, respectively).
We then successfully linked more than 95% member
addresses to census block groups and assigned black race
to individuals who lived in a census block group where
75% or more residents were black in 2008 by using the
proprietary demographic estimates supplied by Nielsen
Claritas Inc (www.claritas.com).14 Subsequently, census
Hispanic and Asian surname lists were applied to assign
as Hispanic or Asian if 75% of individuals in the census
file with a specific surname were Hispanic or Asian,
respectively. By using this sequential method, the percentage of the cohort with an assigned race/ethnicity
increased to 80.5%. All members of the white category
were as assigned in our administrative datasets without
applying the surname or census method to further define
them. Members with unknown race/ethnicity status in
our health plan or health care records and who were not

Table 1. Characteristics of All Members Aged 60 Years or Older

White

Black

Asian

Hispanic

Unknown

All

N (%)
191860 (44.6) 40293 (9.4) 32325 (7.5) 81738 (19.0) 84101 (19.5) 430317
Median age (interquartile range) 70 (64-77) 68 (64-74) 67 (63-73) 67 (63-73)
66 (62-71) 68 (63-75)
Male, %
45.7
41.3
47.1
47.1
45.8
45.7
Heart failure, %
12.8
13.1
6.5
8.1
3.5
9.6
Coronary artery disease, %
23.5
18.6
14.9
15.9
8.1
18.0
Diabetes mellitus, %
22.8
34.2
30.0
34.0
16.1
25.2
Hypertension, %
75.0
85.6
72.0
71.9
57.1
71.7
Atrial fibrillation, %
8.0
3.8
3.9
3.6
2.2
5.3

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

VOL. 102, NO. 10, OCTOBER 2010 907

Race/ethnicity and atrial fibrillation

assigned a race/ethnicity based on the geocoding or surname methods were grouped into the unknown race/
ethnicity category (19.5% overall, 20.2% of nonatrial
fibrillation members, and 8% of atrial fibrillation
patients). To evaluate the validity of the race/ethnicity
data thus estimated, we compared the results to a study
sample of more than 25000+ KPSC adult male members
from whom self-reported race/ethnicity were collected.15
The positive predictive values (PPV) for whites, blacks,
Asians, and Hispanics were 90%, 93%, 94%, and 82%,
respectively (unpublished results). We then repeated the
comparison to maternal race/ethnicity data self-reported
on birth certificates. The PPVs in this validation process
for whites, blacks, Asians, and Hispanics were 84%,
94%, 91%, and 94%, respectively. Among those members with an assigned race/ethnicity, 99% were white,
black, Asian, or Hispanic. We limited this study to members of these groups.

Age-, Gender- and Race/

Ethnicity-Specific Atrial
Fibrillation Prevalence Rates

To further investigate any possible racial/ethnic differences in atrial fibrillation prevalence in the study population, we calculated the age- and gender-specific prevalence for each race/ethnicity (age groups: 60-69, 70-79,
and 80 years of age). Since atrial fibrillation prevalence
in any population will be strongly related to the prevalence
of certain atrial fibrillationrelated comorbidities, we also
report prevalence rates of diabetes, hypertension, heart
failure, and coronary artery disease for each stratum.

Effect of Race/Ethnicity on
Atrial Fibrillation Prevalence

patients), duration of KPSC membership, outpatient

clinic visits in the prior year, neighborhood education
attainment (<12 years of education), and neighborhood
household income (<$50000).

Statistical Analysis
Continuous variables are presented as mean standard deviation, while categorical variables are reported
as proportions. Comparisons of categorical data, including demographics and comorbidities, were analyzed
with c2 tests. The association between race/ethnicity and
atrial fibrillation was analyzed with logistic regression
models adjusting for potential confounders. ORs with
95% confidence intervals (CIs) were calculated. P values for association were deemed significant if < .05. All
data were analyzed with SAS software version 9.13
(SAS Institute, Cary, North Carolina).

Results

In 2008, there were 430317 KPSC members aged 60

years or older, with the median age being 68 years. The
racial/ethnic composition is listed in Table 1. The median
age of white members was 2 to 3 years older than nonwhites. There was a predominance of females among all
groups, with males comprising the lowest percentage
among blacks. Excluding patients whose races/ethnicities were unknown, heart failure was more prevalent
among white and black patients than among Hispanics
and Asians. Diabetes was least prevalent, but coronary
artery disease most prevalent, among whites.
Hypertension was most prevalent among blacks.

Atrial Fibrillation Prevalence

To assess the effect of race/ethnicity on atrial fibrillation prevalence with white members as referent, we calculated the crude odds ratio (OR) of atrial fibrillation
among each nonwhite group, the OR adjusted for age
and gender, and OR adjusted for a number of covariates
known or suspected to affect atrial fibrillation diagnosis:
age, gender, diabetes mellitus, hypertension, heart failure, chronic kidney disease, coronary artery disease,
chronic obstructive pulmonary disease, morbid obesity,
sleep apnea, prior stroke (present in 16.4% of atrial
fibrillation patients and 7.2% of nonatrial fibrillation

Race/ethnicity data were unavailable for 19.5% of the

overall study population and 8% of patients with atrial
fibrillation. Patients whose race/ethnicity data were not
captured in our database tended to have fewer comorbidities as well as a lower prevalence of atrial fibrillation. It
was not possible, however, to determine the racial/ethnic
composition of the unknown group despite manual chart
review. The prevalences for atrial fibrillation were 8.0%,
3.8%, 3.9%, 3.6%, and 2.2% for whites, blacks, Asians,
Hispanics, and unknowns, respectively.
As shown in Table 2, white atrial fibrillation subjects
were to 4 years older than nonwhites. Consistent with

Table 2. Characteristics of 22842 Patients of Known Race/Ethnicity With Atrial Fibrillation

White

Black

Asian

Hispanic

Unknown

All

N (%)
15318 (67.1) 1509 (6.6) 1255 (5.5) 2926 (12.8) 1834 (8.0) 22842
Median age (interquartile range) 76 (70-82) 73 (68-80) 72 (66-78) 74 (68-79) 72 (66-79) 75 (69-81)
Male, %
56.1
52.0
62.9
57.7
58.2
56.6
Heart failure, %
36.5
44.2
28.8
36.7
24.9
35.7
Coronary artery disease, %
41.7
41.4
39.8
40.7
26.8
40.2
Diabetes mellitus, %
30.0
43.2
44.2
45.1
25.4
33.2
Hypertension, %
89.9
94.5
91.5
91.2
83.4
89.9

908 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

P Value
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001

VOL. 102, NO. 10, OCTOBER 2010

Race/ethnicity and atrial fibrillation

previous studies,2,6,16 there was a male predominance

overall and in each group. Heart failure was most common in blacks and least so in Asians, coronary artery
disease prevalence was similar across groups, diabetes
mellitus was least common among whites, and hypertension most prevalent in blacks.

Odd Ratios of Atrial Fibrillation

by Race/Ethnicity
Table 3 lists the gender-specific atrial fibrillation
prevalence of each group as well as the crude, age- and
gender- adjusted, and fully adjusted ORs of atrial fibrillation among nonwhites compared to whites. Atrial
fibrillation was more prevalent among men in all groups.
The crude OR for each nonwhite group was less than
half that of whites. With whites as referent, the crude OR
for the nonwhite groups ranged from 0.43 to 0.47.
Adjusting for age and gender only, it ranged from 0.51
to 0.57. In the fully adjusted models, the ORs for blacks,
Asians, and Hispanics were 0.49, 0.68, and 0.58, respectively, with narrow 95% CIs. In the group of unknown
race/ethnicity, atrial fibrillation prevalence was the lowest of all groups because patients in this group had the
least number of comorbidities (Tables 2 and 4) and also
were least likely to seek medical attention. The relative
racial/ethnic composition of this group could not be
determined.

Age-, Gender- and Race/

Ethnicity-Specific Atrial
Fibrillation Prevalence Rates

Figures 1 to 3 and Table 4 list the age-, gender- and

race/ethnicity-specific atrial fibrillation prevalence and
comorbidities rates for 3 age strata: 60 to 69, 70 to 79,
and at least 80 years. Atrial fibrillation was more prevalent among males than females in all 3 age strata and
became more prevalent with each age stratum for both

genders and for all races/ethnicities. Among individuals

aged 60 to 69 years, the prevalences of atrial fibrillation
in males and females were 3.7% and 1.7%, respectively
(2.7% overall). Among individuals aged 70 to 79, the
prevalences of atrial fibrillation in males and females
were 8.6% and 5.3%, respectively (6.8% overall). Among
individuals aged 80 years or older, the prevalences of
atrial fibrillation in males and females were 15.7% and
11.5%, respectively (13.1% overall). Heart failure, coronary artery disease, and hypertension prevalence all
increased with each stratum in all racial/ethnic groups. In
each stratum, hypertension was most prevalent among
blacks, diabetes least prevalent among whites, and coronary artery disease most prevalent among whites. For
each stratum, the prevalence and adjusted OR of atrial
fibrillation was consistently lower for each nonwhite
group compared to whites in that stratum.

Discussion

Two often cited studies have estimated atrial fibrillation disease burden in the United States by applying
age- and gender-specific atrial fibrillation prevalence
rates to US Census data.1,2 Both studies arrived at a similar estimate of approximately 2.3 million prevalent cases
of atrial fibrillation in this country in the 1990s.
Projections of future atrial fibrillation prevalence have
been made by applying current estimates to US Census
Bureau estimates of population growth.2,6 The investigators reported that by 2050, there may be from 5.6 million
to more than 12 million cases of atrial fibrillation in the
United States. However, none of these studies have taken
into account possible racial/ethnic differences in atrial
fibrillation prevalence. With a growing percentage of the
US populace being nonwhite, we believe it is important
to consider these differences in future studies of societal
burden of atrial fibrillation disease.

Table 3. Atrial Fibrillation Prevalence and Adjusted Odds Ratio of Nonwhites Compared to Whites

Atrial fibrillation prevalence

Males, %
Females, %
OR for atrial fibrillation (95% CI)
Unadjusted

White

Black

Asian

Hispanic

Unknown

9.8
6.5

4.7
3.1

5.2
2.7

4.4
2.9

2.8
1.7

Adjusted for age and gender

Fully adjusteda

0.45
0.47
0.43
0.26
(0.43-0.47) (0.44-0.49) (0.41-0.45) (0.25-0.27)
0.51
0.57
0.51
0.33
(0.48-0.54) (0.53-0.60) (0.49-0.54) (0.30-0.35)
0.49
0.68
0.58
0.50
(0.47-0.52) (0.64-0.72) (0.55-0.61) (0.48-0.53)

All P Value
6.6
4.2

<.0001

<.0001
<.0001
<.0001

Abbreviations: CI, confidence interval; OR, odds ratio.

a

 djusted for age, gender, diabetes mellitus, hypertension, heart failure, chronic kidney disease, coronary artery disease, chronic
A
obstructive pulmonary disease, morbid obesity, sleep apnea, prior stroke, duration of KPSC membership, outpatient clinic visits in the
prior year, neighborhood education attainment (<12 years of education), and neighborhood household income (<$50000).

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

VOL. 102, NO. 10, OCTOBER 2010 909

Race/ethnicity and atrial fibrillation

Previously Reported Atrial

Fibrillation Prevalence
Among individuals older than 60 to 65 years in studies of predominantly white US populations, atrial fibrillation prevalence2,3,6,17,18 ranged from 3.8% to 6%. In the
Cardiovascular Health Study (CHS) of 5200 individuals
aged 65 years or older (95% white),19 the prevalence of
atrial fibrillation was 6.2% in men and 4.8% in women.17
In the AnTicoagulation and Risk Factors in Atrial
Fibrillation (ATRIA) cross-sectional study of a large
health maintenance organization in northern California
(85% white), the prevalence of atrial fibrillation was
3.8% among individuals 60 years or older.2 Analyzing
collective data from 4 population-based studies of atrial
fibrillation, Feinberg et al reported that the prevalence
was 5.9% among those older than 65 years.1 Similarly, in
Figure 1. Prevalence and Adjusted Odds Ratio
of Atrial Fibrillation by Race/Ethnicity in Patients
Aged 60-69 Years
A

A, Prevalence of atrial fibrillation by race/ethnicity in patients

aged 60-69 years. B, Adjusted odds ratio of atrial fibrillation
in patients aged 60-69 years. Bars represent 95% confidence
intervals.

910 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

the Rotterdam study of 6432 individuals older than 55

years, atrial fibrillation prevalence20 was 5.5%.
In the present study of a larger and more diverse population than prior studies, the overall nonvalvular atrial
fibrillation prevalence was 5.3% among KPSC members
aged 60 years or older. We found the prevalences to be
2.7% and 6.8% among members 60 to 69 and 70 to 79
years old, respectively, which were comparable to prior
studies.1-3,6,17,20,21 Among KPSC members aged 80 years or
older, the prevalence of 13.1% was somewhat higher than
most other reports2,3,17 but comparable to that of Olmstead
County, Minnesota, and the Rotterdam study cohort.6,18,20
Of note, among the 191860 white individuals in our study,
the atrial fibrillation prevalence was 8.0%. It was substantially lower in the nonwhite groups (range, 3.6%-3.9%).

Effect of Race/Ethnicity
on Atrial Fibrillation
The most significant finding in the present study was
that older white KPSC members had a greater odds of
atrial fibrillation than nonwhites, whether by crude OR
or adjusted for prominent known atrial fibrillation risk
factors. The adjusted OR of atrial fibrillation for the
nonwhite groups was from 32% to 51% lower than
whites. This finding was consistent across age strata and
gender and with a good degree of confidence in the estimates. Among the nonwhite groups, the adjusted OR for
atrial fibrillation among Asians was higher than blacks
and Hispanics. The clinical significance of this finding is
unclear, however, as Asians constituted the smallest
group and were likely the most heterogeneous, since our
database does not differentiate east from south Asians.
Such racial/ethnic heterogeneity of atrial fibrillation
risk has been suggested by previous studies, although in
less-diverse populations.2,9,22 In the ATRIA study, there
was a higher atrial fibrillation prevalence observed
among whites than blacks older than 50 years (2.2% vs
1.5%, p < .001).2 In a study of hospitalized heart failure
patients, blacks had 50% lower odds of atrial fibrillation
compared to whites (38.3% vs 19.7%).8 In a study of
2123 ECGs of consecutive white and black patients,
atrial fibrillation was found in 2.5% of black and 7.8% of
white subjects.23 Among a large population of US male
veterans, the prevalences of atrial fibrillation in whites,
blacks, Asians, and Hispanics were 5.7%, 3.4%, 3.6%,
and 3.0%, respectively.9 The adjusted ORs for atrial
fibrillation among whites compared to blacks, Asians,
and Hispanics were 1.84, 1.41, and 1.77, respectively.
However, this study included only men, and nonwhites
constituted only 13% of the atrial fibrillation cohort.
In all the previous studies, nonwhites constituted no
more than 16% of the atrial fibrillation cohort. In the
present study, approximately 33% of atrial fibrillation
patients were nonwhite. The findings are nonetheless
remarkably consistent over time and study design: all
have identified a greater atrial fibrillation prevalence
VOL. 102, NO. 10, OCTOBER 2010

Race/ethnicity and atrial fibrillation

among whites. What accounts for this apparent predilection for atrial fibrillation in white individuals is unanswered. Since disease prevalence is a function of incidence and disease duration, some insights may be gained
by examining these 2 components.

Racial/Ethnic Differences in
Atrial Fibrillation Incidence and
Associated Comorbidities

In the Atherosclerosis Risk in Communities study,

among individuals aged 45 to 65 years at baseline,
blacks had a 41% lower incidence of atrial fibrillation
than whites.24 In CHS, black race was associated with a
46% relative risk of new-onset atrial fibrillation compared to whites in their stepwise Cox models.25 In an
Figure 2. Prevalence and Adjusted Odds Ratio
of Atrial Fibrillation by Race/Ethnicity in Patients
Aged 70-79 Years
A

A, Prevalence of atrial fibrillation by race/ethnicity in patients

aged 70-79 years. B, Adjusted odds ratio of atrial fibrillation
in patients aged 70-79 years. Bars represent 95% confidence
intervals.

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

analysis of 7 trials involving patients with acute coronary syndrome, Asians had a 35% lower risk of incident
atrial fibrillation.10 In most studies that reported the ages
of whites and nonwhites separately,8-10,26 white atrial
fibrillation patients were 5 to 7 years older on average
than other races/ethnicities. Given that atrial fibrillation
incidence is strongly correlated with age,6 the older ages
of white individuals may account for some of the excess
incidence that were not fully adjusted for in the models.
Conversely, in our cohort, blacks had the highest prevalence of hypertension and heart failure and had more
diabetes than whites, yet had one of the lowest rates of
atrial fibrillation. Among the traditional atrial fibrillation
risk factors, white subjects had slightly greater or equal
prevalence of coronary artery disease but less of all
other risk factors. Furthermore, in a large randomized
Figure 3. Prevalence and adjusted Odds Ratio
of Atrial Fibrillation by Race/Ethnicity in Patients
Aged 80 years
A

A, Prevalence of atrial fibrillation by race/ethnicity in patients

aged 80 years. B, Adjusted odds ratio of atrial fibrillation
in patients aged 80 years. Bars represent 95% confidence
intervals.

VOL. 102, NO. 10, OCTOBER 2010 911

Race/ethnicity and atrial fibrillation

trial of rate vs rhythm control strategies,26 both blacks

and Hispanics were substantially less likely than whites
to have lone atrial fibrillation (2.6 vs 13.3%, p < .0001
and 6.8% vs 13.3%, p = .03, respectively). Multiple
studies have consistently shown that white individuals
are at greater risk for atrial fibrillation despite having
less comorbidity associated with the development of
atrial fibrillation. While this finding may suggest that
white subjects have a greater propensity to develop atrial
fibrillation in the absence of traditional risk factors, it
remains unclear if this indicates an independent contribution of white race to atrial fibrillation development.

Duration of Atrial Fibrillation

The younger median age among nonwhite subjects
with atrial fibrillation may indicate a shorter duration of
disease, assuming onset of atrial fibrillation was at
approximately the same age as whites. This is unproven
and cannot be addressed by our data. A healthy migrant
effect cannot be ruled out, in which immigrants have a
generally lower adjusted total and cause-specific mortality.27-29 One possible explanation is that immigrants with
chronic illnesses return to their home country, thus
decreasing the apparent disease prevalence among the
remaining migrant group. However, the vast majority of
KPSC members are locally employed or are family
members of such individuals. The year-over-year

membership attrition rate in this age group is roughly

10%. These characteristics lessen the potential for the
healthy migrant effect to account for the observed
differences.
Atrial fibrillation is associated with a 2-fold risk of
death after adjusting for conventional risk factors.4 If certain groups have a greater risk of other competing causes of
death, it may potentially affect apparent atrial fibrillation
prevalence. For instance, blacks have a 2- to 3-year lower
life expectancy at age 60 than whites.30 However, while the
life expectancy of whites is higher than blacks, it is similar
to Hispanics and actually lower than Asians.31 This will not
explain the consistently observed greater prevalence and
adjusted OR of atrial fibrillation among whites.
Another possible factor is underascertainment of disease. Nonwhites are more likely to be uninsured and
underinsured in the general population.32 However, all our
members belong to a prepaid health maintenance organization and have similar access to care, which should minimize, but possibly not eliminate, ascertainment biases.33

Limitations
There are several limitations to this study. Due to the
nature of our administrative databases, race/ethnicity
data were available in a greater proportion of atrial fibrillation patients than the overall membership. Race/ethnicity data were assigned rather than obtained by

Table 4. Age-, Gender-, and Race/Ethnicity-Specific Comorbidity Rates

White

Black

Asian

Hispanic

Unknown

All

P Value

Age 60-69 y, N (%)

91458 (38.0) 22445 (9.3) 20188 (8.4) 50075 (20.8) 56797 (23.6) 240963
Heart failure, %
7.0
9.2
4.0
5.1
1.9
5.4
<.0001
CAD, %
16.8
14.6
11.6
11.7
5.9
12.5 <.0001
Diabetes mellitus, %
21.5
33.1
28.4
32.6
15.5
24.1 <.0001
Hypertension, %
66.5
82.3
66.5
66.1
52.4
64.6 <.0001
1
0.47
0.71
0.52
0.48
<.0001
Adjusted ORa (95% CI)
(0.43-0.52) (0.64-0.78) (0.48-0.56)
(0.44-0.52)
Age 70-79 y, N (%)
65752 (49) 13105 (9.8) 9388 (7.0) 24733 (18.4) 21197 (15.8) 134175
Heart failure, %
14.4
16.2
8.9
10.7
5.3
12.0 <.0001
CAD, %
27.4
22.5
19.1
20.8
11.4
22.6 <.0001
Diabetes mellitus, %
24.9
37.0
33.3
37.2
18.0
27.9 <.0001
Hypertension, %
80.6
89.0
80.1
80.1
66.0
79.0 <.0001
1
0.49
0.66
0.61
0.52
<.0001
Adjusted ORa (95% CI)
(0.45-0.53) (0.60-0.73) (0.57-0.65)
(0.48-0.57)
Age 80 y, N (%)
34650 (62.8) 4743 (8.6) 2749 (5.0)
Heart failure, %
25.1
22.6
16.8
CAD, %
33.9
26.8
24.3
Diabetes mellitus, %
21.9
31.5
31.0
Hypertension, %
86.9
91.7
84.3
1
0.52
0.67
Adjusted ORa (95% CI)
(0.46-0.48) (0.58-0.76)

6930 (12.6)
20.0
28.9
33.3
85.3
0.64
(0.59-0.70)

6107 (11.1)
11.7
17.6
15.0
70.0
0.54
(0.49-0.61)

55179
22.4
30.4
23.9
85.1

<.0001
<.0001
<.0001
<.0001
<.0001

Abbreviations: CAD: coronary artery disease; CI, confidence interval; OR, odds ratio.
a

 djusted for age, gender, diabetes mellitus, hypertension, heart failure, chronic kidney disease, coronary artery disease, chronic
A
obstructive pulmonary disease, morbid obesity, sleep apnea, prior stroke, duration of Kaiser Permanente Southern California
membership, outpatient clinic visits in the prior year, neighborhood education attainment (<12 years of education), and
neighborhood household income (<$50000).

912 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

VOL. 102, NO. 10, OCTOBER 2010

Race/ethnicity and atrial fibrillation

self-report. However, we validated our assignment

method by comparing the results to a sample of members with self-reported race/ethnicity and found the 2 to
be in excellent agreement. Underascertainment of atrial
fibrillation with our methods is possible. A previous
study suggested that the sensitivity of discharge diagnosis for new-onset atrial fibrillation25 may be as low as
71%. We maximized ascertainment by using both computerized ECG records in addition to ICD-9 codes, as
well as diagnoses documented by physicians during
ambulatory visits. However, we cannot exclude systematic disparities in ascertainment based on race/ethnicity.
One study of racial disparity in Medicare health plans
found that among black and white patients, the majority
of the disparity was found in the same health plan.34
We could not adjust for all possible atrial fibrillation
risk factors due to lack of data, such as excessive alcohol
intake, body mass and body mass index, and left atrial
volume.35-37 Residual confounding may account for some
of the observed differences in atrial fibrillation OR
despite adjustment.

Acknowledgment

The authors thank the investigators of the California

Mens Health Study for the use of their race/ethnicity data
among Kaiser Permanente Southern California members.

References

1. Feinberg WM, Blackshear JL, Laupacis A, Kronmal R, Hart RG. Prevalence, age distribution, and gender of patients with atrial fibrillation. Analysis and implications. Arch Intern Med. 1995;155:469-473.
2. Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke
prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA)
Study. JAMA. 2001;285:2370-2375.
3. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk
factor for stroke: the Framingham Study. Stroke. 1991;22:983-988.
4. Benjamin EJ, Wolf PA, DAgostino RB, et al. Impact of atrial fibrillation on
the risk of death: the Framingham Heart Study. Circulation. 1998;98:946-952.
5. Wolf PA, Benjamin EJ, Belanger AJ, et al. Secular trends in the prevalence
of atrial fibrillation: The Framingham Study. Am Heart J. 1996;131:790-795.
6. Miyasaka Y, Barnes ME, Gersh BJ, et al. Secular trends in incidence of atrial fibrillation in Olmsted County, Minnesota, 1980 to 2000, and implications
on the projections for future prevalence. Circulation. 2006;114:119-125.
7. US Census Bureau. State and County QuickFacts. California. http://
quickfacts.census.gov/qfd/states/06000.html. Accessed March 12, 2010.
8. Ruo B, Capra AM, Jensvold NG, Go AS. Racial variation in the prevalence of atrial fibrillation among patients with heart failure: the Epidemiology, Practice, Outcomes, and Costs of Heart Failure (EPOCH) study. J Am
Coll Cardiol. 2004;43:429-435.
9. Borzecki AM, Bridgers DK, Liebschutz JM, et al. Racial differences in
the prevalence of atrial fibrillation among males. J Natl Med Assoc.
2008;100:237-245.
10. Novaro GM, Asher CR, Bhatt DL, et al. Meta-analysis comparing reported frequency of atrial fibrillation after acute coronary syndromes in Asians
versus whites. Am J Cardiol. 2008;101:506-509.
11. Conway DS, Lip GY. Ethnicity in relation to atrial fibrillation and stroke
(the West Birmingham Stroke Project). Am J Cardiol. 2003;92:1476-1479.
12. Shen AY-J, Yao JF, Brar SS, et al. Racial/Ethnic Differences in Ischemic
Stroke Rates and the Efficacy of Warfarin Among Patients With Atrial Fibrillation. Stroke. 2008;39:2736-2743.
13. Benjamin EJ, Levy D, Vaziri SM, et al. Independent risk factors for atrial

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

fibrillation in a population-based cohort. The Framingham Heart Study.

JAMA. 1994;271:840-844.
14. Chen W, Petitti DB, Enger S. Limitations and potential uses of census-based data on ethnicity in a diverse community. Ann Epidemiol.
2004;14:339-345.
15. Enger SM, Van den Eeden SK, Sternfeld B, et al. California Mens Health
Study (CMHS): a multiethnic cohort in a managed care setting. BMC Public
Health. 2006;6:172.
16. Kannel WB, Wolf PA, Benjamin EJ, Levy D. Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based
estimates. Am J Cardiol. 1998;82:2N-9N.
17. Furberg CD, Psaty BM, Manolio TA, et al. Prevalence of atrial fibrillation in elderly subjects (the Cardiovascular Health Study). Am J Cardiol.
1994;74:236-241.
18. Phillips SJ, Whisnant JP, OFallon WM, Frye RL. Prevalence of cardiovascular disease and diabetes mellitus in residents of Rochester, Minnesota.
Mayo Clin Proc. 1990;65:344-359.
19. Tell GS, Fried LP, Hermanson B, et al. Recruitment of adults 65 years and
older as participants in the Cardiovascular Health Study. Ann Epidemiol.
1993;3:358-366.
20. Heeringa J, van der Kuip DAM, Hofman A, et al. Prevalence, incidence and lifetime risk of atrial fibrillation: the Rotterdam study. Eur Heart
J. 2006;27:949-953.
21. Lake FR, Cullen KJ, de Klerk NH, McCall MG, Rosman DL. Atrial fibrillation and mortality in an elderly population. Aust N Z J Med. 1989;19:321-326.
22. Lip GY, Bawden L, Hodson R, et al. Atrial fibrillation amongst the IndoAsian general practice population. The West Birmingham Atrial Fibrillation
Project. Int J Cardiol. 1998;65:187-192.
23. Upshaw CB, Jr. Reduced prevalence of atrial fibrillation in black
patients compared with white patients attending an urban hospital: an
electrocardiographic study. J Natl Med Assoc. 2002;94:204-208.
24. Alonso A, Agarwal SK, Soliman EZ, et al. Incidence of atrial fibrillation
in whites and African-Americans: the Atherosclerosis Risk in Communities
(ARIC) study. Am Heart J. 2009;158:111-117.
25. Psaty BM, Manolio TA, Kuller LH, et al. Incidence of and Risk Factors for
Atrial Fibrillation in Older Adults. Circulation. 1997;96:2455-2461.
26. Bush D, Martin LW, Leman R, Chandler M, Haywood LJ. Atrial fibrillation
among African Americans, Hispanics and Caucasians: clinical features
and outcomes from the AFFIRM trial. J Natl Med Assoc. 2006;98:330-339.
27. Stern MP, Wei M. Do Mexican Americans really have low rates of cardiovascular disease? Prev Med. 1999;29:S90-95.
28. Singh GK, Siahpush M. All-cause and cause-specific mortality of
immigrants and native born in the United States. Am J Public Health.
2001;91:392-399.
29. Thom T, Haase N, Rosamond W, et al. Heart disease and stroke statistics2006 update: a report from the American Heart Association Statistics
Committee and Stroke Statistics Subcommittee. Circulation. 2006;113:e85-151.
30. Arias E. United States life tables, 2004. Natl Vital Stat Rep. 2007;56:1-39.
31. Lee H, McConville S. Death in the Golden StateWhy Do Some Californians Live Longer? California CountsPopulation Trends and Profiles.
2007;9:1-27.
32. Thamer M, Richard C, Casebeer AW, Ray NF. Health insurance coverage among foreign-born US residents: the impact of race, ethnicity, and
length of residence. Am J Public Health. 1997;87:96-102.
33. Fiscella K, Franks P, Doescher MP, Saver BG. Disparities in health care by
race, ethnicity, and language among the insured: findings from a national
sample. Med Care. 2002;40:52-59.
34. Trivedi AN, Zaslavsky AM, Schneider EC, Ayanian JZ. Relationship
Between Quality of Care and Racial Disparities in Medicare Health Plans.
JAMA. 2006;296:1998-2004.
35. Wang TJ, Parise H, Levy D, et al. Obesity and the risk of new-onset atrial
fibrillation. JAMA. 2004;292:2471-2477.
36. Dublin S, French B, Glazer NL, et al. Risk of New-Onset Atrial Fibrillation in
Relation to Body Mass Index. Arch Intern Med. 2006;166:2322-2328.
37. Mukamal KJ, Tolstrup JS, Friberg J, Jensen G, Gronbaek M. Alcohol consumption and risk of atrial fibrillation in men and women: the Copenhagen
City Heart Study. Circulation. 2005;112:1736-1742. n

VOL. 102, NO. 10, OCTOBER 2010 913