- Email: [email protected]
Racial Differences in the Prevalence of Atrial Fibrillation among Males
o
r
i
g
i
n
a
l
c
o
m
m
u
n
i
c
a
t
i
o
n
Racial Differences in the Prevalence of Atrial Fibrillation among Males Ann M. Borzecki, MD, MPH; D. Keith Bridgers, MD, MSc; Jane M. Liebschutz, MD, MPH; Boris Kader, PhD; Lewis E. Kazis, ScD; and Dan R. Berlowitz, MD, MPH
Financial support: We acknowledge the VA Health Services Research and Development Service and Office of Quality and Performance, who originally supported the 1999 Large Health Survey. Background: Despite being the most common cardiac arrhythmia, little is known about racial differences in atrial fibrillation (AF) prevalence and whether differences persist after accounting for known risk factors. Methods: We identified male respondents to the 1999 Large Health Survey of Veteran Enrollees who had an AF diagnosis in the VA administrative database during the preceding two years. Results: Of 664,754 male respondents, 5.3% had AF. By race, age-adjusted prevalence was 5.7% in whites, 3.4% in blacks, 3.0% in Hispanics, 5.4% in native Americans/Alaskans, 3.6% in Asians and 5.2% in Pacific Islanders (p<0.001). Of predisposing conditions, whites were more likely to have valvular heart disease, coronary artery disease and congestive heart failure, blacks had the highest hypertension prevalence; Hispanics had the highest diabetes prevalence. Racial differences remained after adjustment for age, body mass index and these comorbidities. White males were significantly more likely to have AF compared to all races but Pacific Islanders [versus blacks, OR=1.84 (95% CI: 1.71–1.98); versus Hispanics, OR=1.77 (1.60–1.97); vs Native Americans, OR 1.15 [1.04-1.27]; versus Asians, OR=1.41 (1.12–1.77) versus Pacific Islanders, OR=1.16 (0.88–1.53)]. Conclusions: AF prevalence varies by race. White males have the highest AF burden even after adjustment for known risk factors. Recognition of the high AF prevalence, especially among whites, as well as native Americans and Pacific Islanders, should help guide provider practices for screening among older male patients. Further research is necessary to verify and establish reasons for these racial differences. Key words: atrial fibrillation n epidemiology n race/ ethnicity © 2008. From the Center for Health Quality, Outcomes and Economic Research, Bedford VAMC, Bedford, MA (Borzecki, Kader, Kazis, Berlowitz); Department of Health Policy and Management, Boston University School
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
of Public Health, Boston, MA (Borzecki, Kader, Kazis, Berlowitz); Section of General Internal Medicine, Boston University School of Medicine, Boston, MA (Borzecki, Liebschutz, Berlowitz); and Three Rivers Medical Associates, Irmo, SC (Bridgers). Send correspondence and reprint requests for J Natl Med Assoc. 2008;100:237–245 to: Dr. Ann Borzecki, 200 Springs Road (152), Bedford, MA 01730; phone: (781) 687-2870; fax: (781) 687-3106; e-mail: [email protected]
INTRODUCTION
A
trial fibrillation (AF) has been called an emerging cardiovascular epidemic.1 It is the most common arrhythmia, affecting almost 2.3 million Americans.2 Prevalence increases markedly with age, nearly doubling with each decade of adult life.3 Rates are also increasing over time independent of age.2,4 By 2050, >5.6 million Americans will be affected.2 AF may cause significant morbidity and mortality. It results in a three-fold excess stroke and heart failure risk5,6 and a two-fold increased all-cause mortality risk.3,6 Population-based cohort studies have identified age, male sex, diabetes, hypertension, coronary artery disease, heart failure, valvular heart disease and—more recently—family history and obesity as AF risk factors.5,7-10 These studies have been limited by relatively small sample sizes in predominantly white populations.5,8,11 Surprisingly little is known about racial differences in AF prevalence and whether race may be a determinant of AF. Given AF’s chronicity and the immutability of race, cross-sectional studies may be useful to address this issue. However, even the recent Anticoagulation and Risk Factors in Atrial Fibrillation Study (ATRIA), a cross-sectional study of almost 2 million managed care patients, including 18,000 with AF, offered limited information on racial differences.2 ATRIA only compared whites and African Americans, and only with respect to crude rates that lacked adjustment for known risk factors.2 Risk factors such as hypertension and diabetes are more common in blacks than whites,12 yet ATRIA found higher unadjusted rates in whites.2 Thus, there may be racial differences in AF predilection independent of other AF risk factors. Given this condition’s importance from a public VOL. 100, NO. 2, FEBRUARY 2008 237
Race Differences in Atrial Fibrillation
health and cost perspective, and the growing population of U.S. minorities,13 we examined racial differences in AF prevalence in a large integrated healthcare system. We also examined whether differences remained after accounting for potential factors that might explain varying AF risk.
METHODS Study Participants and Design We merged retrospective data from two national Veterans Health Administration (VA) administrative databases—the Outpatient Clinic File (OPC) and the Patient Treatment File (PTF)—with cross-sectional data from the 1999 Large Health Survey of Veteran Enrollees. The OPC contains information on all VA outpatient visits; the PTF contains inpatient information. Both list up to 10 diagnoses per visit or hospital discharge.14 Large survey details are described elsewhere.15,16 Briefly, 1.4 of 3.4 million VA enrollees were randomly sampled between July 1999 and January 2000, and mailed a questionnaire, including demographic and health status questions. Eligible subjects were survey respondents with complete age, race and health status information who had OPC and PTF data between October 1, 1997 and September 30, 1999. Females were excluded from the sample, given their relatively small proportion (4.1%). We further identified subjects with ≥1 OPC- or PTF-listed AF diagnosis, ICD-9-CM code 427.3x, during this period. We previously found that a single OPC-/PTFderived AF code in a two-year period had substantial kappa (0.74), sensitivity (86%) and specificity (97%), compared to a standard of AF diagnosis documented in the provider’s clinic note.17 We also extracted ICD-9-CM codes for predisposing comorbidities (hypertension, heart failure, coronary artery disease, diabetes, valvular heart disease, chronic lung disease, hyperthyroidism) and potential complications (cerebrovascular disease) from the OPC/PTF
files.7,18 (Table 1). Body mass index (BMI) was based on self-reported height and weight. We identified race based on response to a survey question previously used in federal surveys.19 Subjects were asked to mark all applicable options: American Indian/ Alaskan native, Asian, black/African American, Spanish/ Hispanic/Latino, native Hawaiian/other Pacific Islander and white. Subjects indicating >1 racial group were assigned to whichever group represented a smaller portion of the U.S. population; i.e., subjects marking black and Hispanic were considered black.
Analyses We examined bivariate relationships using Chi-squared tests and odds ratios (ORs) for categorical variables, and Student’s t tests or analysis of variance (ANOVA) for continuous variables. We first calculated overall AF frequency and examined bivariate relationships between subjects with and without AF by age (categorical and continuous), race, BMI (continuous and categorical, <25, 25–<30, ≥30 kg/m2)10 and AF-associated comorbidities. (We initially examined mitral stenosis, mitral insufficiency, tricuspid insufficiency and aortic valve disease separately, then as a combined variable after finding a significant association between all valvular heart disease categories and AF.) We computed race-specific age-adjusted rates by the direct method, using the full sample as the standard population with a breakdown by five-year increments. We also compared racial groups by age (continuous variable), BMI and comorbidities for the full sample, and among only those with AF. Next, we used logistic regression to examine the association between AF and race, controlling for age (continuous), BMI and AF-predisposing comorbidities, and performed pairwise comparisons between racial groups using Bonferroni’s method.20 We also examined this full model minus hyperthyroidism, since this may represent a relatively acute condition. Because 7% of our sample were missing BMI data, we additionally examined models without BMI, and with imputed values using
Table 1. Atrial fibrillation-associated comorbidities with ICD-9-CM codes* Diagnosis ICD-9-CM Code Atrial fibrillation 427.3x Cerebrovascular disease 430.x-438.x Chronic lung disease 490-496.x, 500-505.x, 506.4 Congestive heart failure 398.91, 402.01, 402.11, 402.91, 404.01, 404.03, 404.11, 404.13, 404.91, 404.93, 414.8, 428.x Coronary artery disease 410-414.x, 429.2 Diabetes 250.x Hypertension 401, 402, 405 Hyperthyroidism 242.x, 245.x Valvular heart disease 394.0, 394.1, 394.2, 395.0, 395.1, 395.2, 396.0, 396.1, 396.2, 396.3, 396.8, 397.0, 424.0, 424.1, 746.2, 746.3, 746.4, 746.5, 746.6, 747.22 * Based in part on Borzecki et al. and Deyo et al.17,18
238 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 100, NO. 2, FEBRUARY 2008
Race Differences in Atrial Fibrillation
the mean sample BMI. We considered a p value <0.05 as significant for all analyses, except for pairwise comparisons, where we used a significance level of 0.003 [derived by adjusting alpha for the number of comparisons (i.e., p=0.05/15 or 0.003)].20 To test the robustness of findings, we performed several additional analyses: 1. We examined the number of subjects reporting multiple race responses who thus might have been misclassified and examined the impact of excluding these individuals from our model. 2. Since AF was diagnosed using administrative data, not electrocardiogram (EKG) results, we addressed possible misclassification by repeating logistic models, excluding patients with only one AF diagnosis first from the PTF and then from any source. (Subjects with transient AF following cardiac surgery would be eliminated in this model.) 3. We examined possible nonrespondent bias by repeating analyses, including nonrespondents with available OPC/PTF diagnostic data, and race data derived either from the large survey (for those
with partial survey information who did not meet sample inclusion criteria) or OPC/PTF databases as available. 4. Since we identified AF using the ICD-9-CM code 427.3x, comprising both AF (427.31) and atrial flutter (427.32), we repeated analyses, excluding subjects with only 427.32 codes. While these two conditions are related, frequently coexist and are treated similarly, they nevertheless do represent distinct electrical abnormalities.21 5. We addressed possible ascertainment bias by race in several ways. We examined the: a) likelihood of having an EKG by race via a logistic regression model adjusting for age, BMI, AF-related conditions, having an inpatient stay and number of outpatient visits; b) number of admissions and outpatient visits by race, since additional medical encounters should increase the likelihood of having an EKG; c) source of AF diagnosis; d) impact of adjusting for VA enrollment priority category, which is associated with the likelihood of having non-VA insurance,16 on our original model.
Table 2. Baseline characteristics of sample Characteristics Subjects with AF Subjects without AF N=35,470 N=629,284 n (%) n (%) Age, Year [Mean (SD)] 72.0 (8.8) 63.3 (13.1) Age, Group 18–49 642 (1.8) 97,523 (15.5) 50–64 5,291 (14.9) 197,066 (31.3) 65–79 23,206 (65.4) 283,798 (45.1) ≥80 6,331 (17.8) 50,897 (8.1) Body Mass Index, kg/m2 [Mean (SD)] 27.61 (5.2) 27.65 (5.0) Body Mass Index Category* <25 kg/m2 9,664 (29.5) 163,925 (28.1) 25–<30 kg/m2 14,261 (43.7) 267,237 (45.7) ≥30 kg/m2 8,736 (26.7) 153,092 (26.2) Race White 30,752 (86.7) 476,015 (75.6) Black 2,139 (6.0) 80,442 (12.8) Hispanic 1,001 (2.8) 37,068 (5.9) Native American or Alaska native 1,208 (3.4) 26,489 (4.2) Asian 211 (0.6) 5,992 (1.0) Hawaiian or other Pacific Islander 159 (0.5) 3,278 (0.5) Selected Comorbidities Hypertension 26,588 (75.0) 321,856 (51.1) Coronary artery disease 23,368 (65.9) 160,267 (25.5) Diabetes mellitus 11,417 (32.2) 130,734 (20.8) Chronic lung disease 12,042 (33.9) 117,575 (18.7) Congestive heart failure 13,497 (38.1) 37,784 (6.0) Cerebrovascular disease 7,039 (19.8) 46,440 (7.4) Valvular heart disease 3,278 (9.2) 8,439 (1.3) Hyperthyroidism 651 (1.8) 5,033 (0.8)
P Value
<0.001 <0.001
0.18 <0.001
<0.001†
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
AF: atrial fibrillation; * BMI numbers do not sum to total sample size because of missing data; † By 2x6 Chi-squared test; in 2x2 analyses whites also had a statistically higher prevalence of AF (p<0.001) compared to all other race groups (unadjusted).
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 100, NO. 2, FEBRUARY 2008 239
Race Differences in Atrial Fibrillation
RESULTS
Blacks had the highest hypertension prevalence, Hispanics the highest diabetes prevalence and native Americans the highest chronic lung disease prevalence. Among those with AF, similar comorbidity trends were seen except that blacks were more likely to have heart failure than whites (44.8% vs. 37.4%; OR=1.36, 1.24–1.49), and native Americans had similar coronary disease rates to whites (67.0% vs. 66.2%; OR=1.04, 0.92–1.17). (Full data available from authors.) The association between BMI and AF was nonlinear. Thus, we used a categorical BMI variable in final models. After adjustment for age, BMI and predisposing comorbidities, whites still had significantly higher odds of having AF compared to all other racial groups except Pacific Islanders. (There was no appreciable difference between adjusted models with and without hyperthyroidism.) Whites were 1.8 times more likely to have AF compared to blacks or Hispanics (Table 4). The odds in blacks versus Hispanics were not significantly different. In adjusted models, both blacks and Hispanics had lower AF odds compared to native Americans, Asians and Pacific Islanders, although the difference between Hispanics and Asians was not significant in pairwise comparisons (p=0.006; prespecified significance level <0.003 based on multiple comparisons). Pacific Islanders, native Americans and Asians were not significantly different. Imputing or excluding missing BMI values yielded similar results.
Overall, survey response rate among males was 61% (N=814,292), of whom 664,754 had complete demographic, health status and administrative information. Of these, 35,470 (5.3%) had an AF diagnosis. Table 2 shows respondents’ baseline characteristics. Respondents with AF were older and more likely white than respondents without AF. The prevalence of selected cardiovascular comorbidities such as heart failure and coronary artery disease was much higher in AF subjects than non-AF subjects. Only 9% of the sample had valvular heart disease; most had lone aortic disease (6.3%). Overall age-adjusted prevalence was 5.2%. By race, prevalence was 6.1% in whites, 2.6% in blacks, 2.6% in Hispanics, 4.4% in native Americans/Alaskans, 3.4% in Asians and 4.6% in Pacific Islanders/Hawaiians. The unadjusted odds of whites having AF compared to blacks was 2.43 [95% confidence interval (CI): 2.32–2.54] and 2.39 compared to Hispanics (95% CI: 2.24–2.55). Table 3 shows baseline characteristics by race. Whites were significantly older, while blacks were significantly younger than all other racial groups (respective mean ages 65.0, SD 12.6 years and 59.0, SD 13.7 years; p<0.001). Prevalence increased with age across all races (data not shown). See Figure 1 for age-adjusted prevalences. Whites had a significantly higher age-adjusted prevalence compared to all other racial groups except Pacific Islanders. Conditions predisposing to AF were all significantly associated with AF risk in bivariate analyses. Further, all varied significantly in prevalence by race, except hyperthyroidism (Table 3). Whites had a significantly higher prevalence of several predisposing comorbidities, including coronary disease, heart failure and valvular disease.
Additional Analyses 1. Multiple races were reported by 29,660 subjects (4.5%); 72% of these reported being native
Figure 1. Age-adjusted prevalence of atrial fibrillation by race* *Adjusted using whole group as standard population. Bars represent 95% confidence intervals.
7.
Prevalence (%)
6.0
5.7%
5.0 4.0
5.2%
5.4%
3.4%
3.6% 3.0%
3.0 2.0 1.0 0.0
White
Black
Hispanic
Native American
Asian
Pacific Islander
Race
240 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 100, NO. 2, FEBRUARY 2008
Race Differences in Atrial Fibrillation
American, 89% of whom reported this in addition to being white (18,789 subjects). Excluding these subjects from analyses did not alter findings, except that the white/native-American difference was no longer significant in pairwise comparisons. 2. Rerunning logistic models using more stringent AF diagnosis criteria—first by eliminating those with only one PTF-based AF diagnosis (N=2,666), then eliminating subjects with only one diagnosis in either source (N=11,263)—did not change results. Retaining these subjects but reclassifying them as lacking AF also made little difference. 3. Nonresponder bias. Out of a potential sample of 1.3 million men, demographic and diagnostic information was available on 1,183,745 subjects, including 411,278 nonrespondents and 107,713 additional respondents lacking full survey-derived demographic or health status data, and thus not in the study sample. Male nonrespondents were younger than sample respondents (mean age 55.0 vs. 63.0). Among those with available race information, blacks were least likely to respond (47% vs. 67% of whites). Response rates among other racial groups varied from 55% for native Americans to 64% for Hispanics. Nonrespondents also had a lower prevalence of AF-related comorbidities. AF prevalence by race (based on OPC/PTF demographic information) among nonrespondents was actually very similar to our study sample. It was 6.1% in whites,
3.1% in blacks, 3.0% in Hispanics, 3.3% in native Americans and 4.6% in Asians (there is no separate category for Pacific Islanders in the OPC/PTF.) Respondents not included in the sample were intermediate between sample respondents and nonrespondents with respect to the noted characteristics. Rerunning logistic models including all subjects did not substantially alter results. All previously significant associations remained and CIs narrowed; the Asian/Hispanic association became significant (p<0.003) (data not shown; available upon request). 4. Lone atrial flutter (ICD-9-CM 427.32) was documented in 13% of cases (4,760). Excluding these subjects did not alter results. 5. Ascertainment bias. Forty-two percent of the study sample had an EKG during the ascertainment period. This ranged from 33% for Asians to 43% for native Americans (Table 3). Among AF subjects, 72% had an EKG (ranging from 71% among whites to 80% among Hispanics). After adjustment for factors (other than AF) that increase the likelihood of having an EKG, Hispanics and native Americans were more likely to have an EKG than whites [respective ORs 1.18 (1.15, 1.21) and 1.05 (1.02, 1.08)]; Asians, blacks and Pacific Islanders were less likely than whites to have an EKG [respective ORs 0.81 (0.76, 0.85), 0.89 (0.88, 0.91) and 0.86 (0.80, 0.93)]. (These same associations held
Table 3. Baseline comparison by race* Native Characteristic White Black Hispanic American Asian N=506,767 N=82,581 N=38,069 N=27,697 N=6,203 Age, Years [mean (SD)] 65.0 (12.6) 59.0 (13.7) 60.8 (13.8) 60.1 (12.5) 61.6 (14.9) BMI, kg/m2 [mean (SD)] 27.6 (4.8) 27.5 (5.1) 27.7 (4.6) 28.1 (5.2) 26.3 (4.3) Comorbidities (%) Hypertension 51.9 59.0 48.4 50.2 47.5 Coronary artery disease 30.1 17.3 19.4 26.8 17.6 Diabetes mellitus 20.6 23.5 27.5 21.4 20.5 Chronic lung disease 20.6 14.9 13.3 23.0 18.7 Congestive heart failure 8.1 6.7 5.3 7.4 5.1 Cerebrovascular disease 8.4 7.2 6.1 7.4 5.6 Valvular heart disease 1.9 1.0 1.4 1.5 1.2 Hyperthyroidism 0.9 0.8 1.0 0.9 1.0 Outpatient Visits FY 98–99 [mean (SD)] 15.4 (18.2) 17.3 (25.5) 15.9 (21.0) 16.9 (20.1) 15.0 (22.1) EKG (%) 42.0 38.7 42.0 43.3 33.0 Enrollment Priority Group 1, 4 or 5 (%) 62.6 67.0 70.2 66.9 53.5
Pacific Islander N=3,437 61.4 (13.8) 28.2 (5.1) 48.6 23.3 23.6 15.4 7.4 7.1 1.6 0.9 16.5 (28.2) 37.3 60.0
* All among race comparisons by characteristic are statistically significant (p<0.001), except for hyperthyroidism (p=0.13); BMI: body mass index; EKG: electrocardiogram; † All VA-enrolled veterans are assigned to one of seven priority groups based on serviceconnected disability, income and special considerations (e.g., POW). Subjects in the highest priority groups are exempt from copayments for medical care. Subjects in priority groups 1 (service-connected disability of ≥50%), 4 (catastrophically disabled) and 5 (income and net worth below established thresholds) are least likely to have non-VA insurance.
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 100, NO. 2, FEBRUARY 2008 241
Race Differences in Atrial Fibrillation
after excluding subjects with AF.) Regarding inpatient versus outpatient identification, blacks had more admissions on average than whites, and blacks with AF were more likely to be identified based only on inpatient records (13% vs. 9%). Hispanics and blacks also had more outpatient visits than whites (Table 3), but whites with AF were more likely to have this diagnosis made based solely on outpatient records (68% of whites vs. 66% of Hispanics and 61% of blacks). Despite Hispanics being more likely to belong to priority groups without insurance and Asians being less likely (Table 3), adjustment for priority did not appreciably alter results (data available upon request).
DISCUSSION
Within a large ethnically diverse male veteran population, 5.3% had an AF diagnosis. We found significant racial differences in AF prevalence, with the highest prevalence among whites, the lowest among blacks and Hispanics. These differences remained even after adjustment for age, BMI and comorbidities predisposing to AF. Prevalence was also relatively high among native American and Pacific Islanders compared to blacks and Hispanics, although whites still were most affected. Our study expands current knowledge of AF prevalence among racial groups. Surprisingly little data are available on racial differences among the general population or undifferentiated outpatient populations; most
existing studies examined prevalence in populations with specific conditions such as heart failure or stroke, and included only blacks or Hispanics.2,22-24 Further, only one previous study, involving heart failure patients, controlled for predisposing comorbidities.22 No other studies have included less common minorities such as native Americans, Pacific Islanders or Asians. The relative white/black and white/Hispanic differences we found are consistent with previous studies.2,8,22-24 The ATRIA study found AF in 2.2% of whites and 1.5% of blacks ≥5;2 our respective sample numbers were 6.8% and 3.4%. A cross-sectional study of subjects hospitalized with heart failure found AF in 38% of whites and 20% of blacks;22 respective prevalences among our sample’s heart failure patients were 28% and 17%. Their adjusted OR for AF in whites versus blacks was similar to ours at 1.96 (1.29–2.63). Notably, they performed sequential models comparing AF odds by race, initially adjusting only for age and gender, then similar to our study, for known AF risk factors, including hypertension, coronary disease, prior heart failure, mitral stenosis, hyperthyroidism and chronic lung disease, with diabetes, aortic valve disease, mitral insufficiency and ejection fraction added to final models. While many risk factors were significant predictors in models, their inclusion did not appreciably change the OR found with age and gender alone. The only study with data on Hispanics found a prevalence of 8% in whites, 5% in blacks and 4% in Hispanics among the complete sample in a case-control study examining stroke risk factors,23 AF prevalence among stroke cases
Table 4. Multivariate logistic regression model of atrial fibrillation by race Variable Odds Ratio Race White vs. black 1.84 White vs. Hispanic 1.77 White vs. native American 1.15 White vs. Asian 1.41 White vs. Pacific Islander 1.16 Black vs. Hispanic 0.96 Black vs. native American 0.62 Black vs. Asian 0.76 Black vs. Pacific Islander 0.63 Hispanic vs. native American 0.65 Hispanic vs. Asian 0.79 Hispanic vs. Pacific Islander 0.65 Native American vs. Asian 1.23 Native American vs. Pacific Islander 1.01 Asian vs. Pacific Islander 0.82
95% Confidence Interval* 1.71–1.98† 1.60–1.97† 1.04–1.27† 1.12–1.77† 0.88–1.53 0.85–1.09 0.55–0.70† 0.60–0.97† 0.47–0.84† 0.56–0.75† 0.62–1.02 0.49–0.88† 0.95–1.57 0.75–1.35 0.58–1.18
Each odds ratio also adjusted for age, BMI (<25, 25–<30, ≥30 kg/m2) and predisposing comorbidities (hypertension, coronary artery disease, diabetes mellitus, chronic lung disease, valvular heart disease and hyperthyroidism). Odds ratios indicate the likelihood of having AF. All variables entered into the model were significant predictors of AF (p<0.003); * 95% confidence intervals using Bonferroni adjustment for multiple comparisons; † Significant result. Confidence interval does not include 1.0.
242 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 100, NO. 2, FEBRUARY 2008
Race Differences in Atrial Fibrillation
by race was 29%, 11% and 11%, respectively.24 Comparable percentages among our cerebrovascular disease patients were 14%, 8% and 10%. Our relative racial differences in cardiovascular comorbidity prevalence are also consistent with other studies.22,24,25 In the previous heart failure and AF study,22 similar to our study, blacks were more likely to have hypertension, and whites were more likely to have coronary disease. The stroke case-control study also found diabetes most prevalent in Hispanics.24 Among AF subjects, the Atrial Fibrillation Follow-Up Investigation of Rhythm Management study also found hypertension most common in blacks and coronary disease more common in whites. Like our AF cohort, heart failure was more common in blacks, although it was also more common in Hispanics than whites.25 We have shown that racial variation in AF prevalence cannot be explained by traditional AF risk factors. For instance, whites had a higher odds of AF than blacks despite greater hypertension prevalence in blacks and despite the fact that, as in other populations, black veterans with hypertension are more likely to have uncontrolled blood pressure than whites.12,26 The higher risk in whites is also likely not explained by unmeasured factors such as lower medication adherence with resulting poor control of underlying conditions. Several studies in veterans have found higher adherence rates among whites compared to blacks and nonwhites with respect to cardiac medications.27,28 Other VA-based research has similarly found differential racial predilection for chronic conditions, i.e., among veterans with diabetes, blacks and native Americans were more likely to have nephropathy, whites were more likely to have coronary disease and heart failure.29 Some authors have conjectured that AF-related racial differences may be due to ascertainment bias, i.e., differential access to care in minorities, including diagnostic testing, may result in lower AF diagnosis rates.22 We did not find evidence to support major differences in access to clinical services, with high healthcare use among all racial groups as evidenced by a mean of ≥15 visits over the two-year period. (Table 3.) Moreover, several minority groups had more outpatient visits than whites, thereby presenting more opportunities for AF diagnosis. Only 42% of the entire sample had an EKG at any time during the ascertainment period, with blacks less likely, and Hispanics actually more likely, to have an EKG versus whites. Presumably, many individuals had an EKG done either before the ascertainment period or in a nonVA healthcare setting. Why racial bias should occur in the first case is unclear. While non-VA healthcare use does vary by race, lack of information on non-VA use should not have altered our findings. Researchers have found that among veterans, blacks, Hispanics and other races are less likely to have non-VA healthcare coverage than whites.16 However, accounting for enrollment JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
priority as a proxy for having non-VA insurance did not change our results. These findings suggest that racial prevalence differences are real and not because whites are more likely to receive healthcare or EKGs and, thus have AF diagnosed. Reasons for racial variation are unclear but presumably include variations in genetic or environmental factors. According to Ruo, “intrinsic racial differences in atrial membrane stability, atrial conduction pathways or genetic polymorphisms [may] lead to different susceptibility” to AF development.22 Racial polymorphism differences have been associated with heart failure risk.30 The Cardiovascular Health Study found racial differences in left atrial size; average left atrial size was greater in elderly white males compared to blacks.31 Further, there may be racial differences in plasma or brain natriuretic peptide levels, both of which have been linked to AF development.32,33 Our study has several strengths. Our sample of 664,754, comprising 20% of the total VA population, represents one of the largest samples studied for AF prevalence. These large numbers allow for a relatively precise estimation of race-specific AF prevalence. We used self-reported race/ethnicity, which is generally accepted as the most reliable and valid source of such information, especially for ethnicities other than white or black. The large survey data is also more complete than VA administrative racial data.34 New information from our study includes data on AF prevalence and risk among native Americans, Pacific Islanders and Asians. Further, we are the first to adjust for factors that might account for ethnic prevalence differences and reached similar conclusions to unadjusted results. Additionally, we performed several sensitivity analyses that support the robustness of our findings. Study limitations include use of VA males who tend to have more comorbidities, so actual prevalence rates may be less generalizable to other settings.15 However, this should not affect our adjusted results. Moreover, the consistency of our relative AF and comorbidity racial prevalence differences with other studies strengthens the external validity of our findings.22,24,25 Given the relatively few females, we excluded them from analyses since we would be unable to draw reliable conclusions. Even among males there were relatively few subjects in the smallest minority groups with AF such that we may have been unable to find statistical differences when true differences existed. Nonetheless, this represents the largest existing sample of such minorities. Given the cross-sectional study design, we cannot definitively prove that race is a risk factor for AF. However, our large sample, the relative chronicity of AF and the unchanging nature of race make this conclusion highly likely. Our diagnosis is based administrative data, not EKGs. However, the ATRIA study similarly identified AF subjects using administrative data, but suppleVOL. 100, NO. 2, FEBRUARY 2008 243
Race Differences in Atrial Fibrillation
mented with EKGs.2 Of almost 18,000 AF subjects, 90% had AF by ICD-9-CM codes, 87% had EKG confirmation. Subjects with only one AF code were less likely to have EKG confirmation. In our study, findings were unchanged by reclassifying those with only one diagnosis as not having AF; additionally, we did not find a predilection for whites to have EKGs. We also cannot determine from ICD-9-CM coding what AF type (persistent or paroxysmal) patients had, and lack left ventricular function data. However, as noted, adding ventricular function to models in another study minimally impacted adjusted ORs.22 A recent meta-analysis also suggested decreased AF risk among heart failure patients treated with angiotensin blockers.35 While we did not examine medication use, we have no reason to suspect a bias against using/prescribing this medication in high-risk groups such as whites. Finally, our data are several years old. Other data have shown that AF prevalence is increasing over time, presumed partly due to the aging of the population and partly due to the fact that subjects who previously might have died from pre-existing or existing comorbidities are living longer,36 such that we may be slightly underestimating the prevalence in current veterans. However, there is no evidence to suggest a differential effect in prevalence based on race over time. Further, given the limited literature regarding the epidemiology of AF in nonwhite populations and the growth of such minorities in the United States, data on these groups are clearly necessary. By providing such information, this study makes an important contribution to the existing literature. This study establishes race as an independent risk factor for AF among males. The highest risk was seen in whites, followed by Pacific Islanders and native Americans; blacks and Hispanics had the lowest risk. Given the general population’s aging, AF will become an increasingly important cause of morbidity and mortality, especially in these high-risk populations. Recognition of the high AF risk in these racial groups should guide provider practices for screening among older patients and help inform policy on healthcare resource distribution. Further research is also warranted to establish reasons for these racial predispositions. Such research, by establishing mechanisms through which AF develops, should help curb this growing cardiovascular epidemic.
Acknowledgements
Special thanks to Mark Glickman, PhD, for statistical assistance, and to Marshall Goff for editorial assistance.
REFERENCES
1. Braunwald E. Shattuck lecture—cardiovascular medicine at the turn of the millennium: triumphs, concerns, and opportunities. N Engl J Med. 1997;337(19):1360-1369. 2. Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke
244 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA. 2001;285(18):2370-2375. 3. Benjamin EJ, Wolf PA, D’Agostino RB, et al. Impact of atrial fibrillation on the risk of death: the Framingham Heart Study. Circulation. 1998;98(10):946952. 4. Wolf PA, Benjamin EJ, Belanger AJ, et al. Secular trends in the prevalence of atrial fibrillation: the Framingham Study. Am Heart J. 1996;131(4):790-795. 5. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke. 1991;22(8):983-988. 6. Stewart S, Hart CL, Hole DJ, et al. A population-based study of the longterm risks associated with atrial fibrillation: 20-year follow-up of the Renfrew/Paisley study. Am J Med. 2002;113(5):359-364. 7. Benjamin EJ, Levy D, Vaziri SM, et al. Independent risk factors for atrial fibrillation in a population-based cohort. The Framingham Heart Study. JAMA. 1994;271(11):840-844. 8. Psaty BM, Manolio TA, Kuller LH, et al. Incidence of and risk factors for atrial fibrillation in older adults. Circulation. 1997;96(7):2455-2461. 9. Fox CS, Parise H, D’Agostino RB Sr, et al. Parental atrial fibrillation as a risk factor for atrial fibrillation in offspring. JAMA. 2004;291(23):2851-2855. 10. Wang TJ, Parise H, Levy D, et al. Obesity and the risk of new-onset atrial fibrillation. JAMA. 2004;292(20):2471-2477. 11. Phillips SJ, Whisnant JP, O’Fallon WM, et al. Prevalence of cardiovascular disease and diabetes mellitus in residents of Rochester, Minnesota. Mayo Clin Proc. 1990;65(3):344-359. 12. Hajjar I, Kotchen TA. Trends in prevalence, awareness, treatment, and control of hypertension in the United States, 1988–2000. JAMA. 2003;290(2):199-206. 13. U.S. Census Bureau. State Population Estimates by Race: July 1, 2002. http://eire.census.gov/popest/data/states/ST-EST2002-ASRO-03.php. Accessed 12/20/03. 14. VA Information Resource Center: Toolkit for new users of VA data. www. virec.research.med.va.gov/Support/Training-NewUsersToolkit/Toolkit.htm. Accessed 05/15/04. 15. Perlin J, Kazis L, Skinner K, et al. Health status and outcomes of veterans. 1999 large health survey of veteran enrollees. Executive report. Washington, DC: Department of Veterans Affairs and Veterans Health Administration, Office of Quality and Performance; 2000. 16. Shen Y, Hendricks A, Zhang S, et al. VHA enrollees’ health care coverage and use of care. Med Care Res Rev. 2003;60(2):253-267. 17. Borzecki AM, Wong AT, Hickey EC, et al. Identifying hypertension-related comorbidities from administrative data: what’s the optimal approach? Am J Med Qual. 2004;19(5):201-206. 18. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613-619. 19. Office of Management and Budget. Standards for maintaining, collecting and presenting federal data on race and ethnicity. www.whitehouse. gov/omb/inforeg/r&e_app-a-update.pdf. 20. Fisher LD, van Belle G. Multiple Comparisons. In: Biostatistics. A Methodology for the Health Sciences. New York, NY: John Wiley & Sons, Inc.; 1993. 21. Waldo AL. The interrelationship between atrial fibrillation and atrial flutter. Prog Cardiovasc Dis. 2005;48(1):41-56. 22. 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(3):429-435. 23. Sacco RL, Boden-Albala B, Abel G, et al. Race-ethnic disparities in the impact of stroke risk factors: the northern Manhattan stroke study. Stroke 2001;32(8):1725-1731. 24. Sacco RL, Kargman DE, Zamanillo MC. Race-ethnic differences in stroke risk factors among hospitalized patients with cerebral infarction: the Northern Manhattan Stroke Study. Neurology. 1995;45(4):659-463. 25. Bush D, Martin LW, Leman R, et al. Atrial fibrillation among African Americans, Hispanics and Caucasians: clinical features and outcomes from the AFFIRM trial. J Natl Med Assoc. 2006;98(3):330-339. 26. Borzecki AM, Glickman ME, Kader B, et al. The effect of age on hypertension control and management. Am J Hypertens. 2006;19(5):520-527.
VOL. 100, NO. 2, FEBRUARY 2008
fKPVSOBMQSFTT
27. Charles H, Good CB, Hanusa BH, et al. Racial differences in adherence to cardiac medications. J Natl Med Assoc. 2003;95(1):17-27. 28. Kopjar B, Sales AE, Pineros SL, et al. Adherence with statin therapy in secondary prevention of coronary heart disease in veterans administration male population. Am J Cardiol. 2003;92(9):1106-1108. 29. Young BA, Maynard C, Boyko EJ. Racial differences in diabetic nephropathy, cardiovascular disease, and mortality in a national population of veterans. Diabetes Care. 2003;26(8):2392-2399. 30. Small KM, Wagoner LE, Levin AM, et al. Synergistic polymorphisms of beta1- and alpha2C-adrenergic receptors and the risk of congestive heart failure. N Engl J Med. 2002;347(15):1135-1142. 31. Manolio TA, Gottdiener JS, Tsang TS, et al. Left atrial dimensions determined by M-mode echocardiography in black and white older (≥65 years) adults (The Cardiovascular Health Study). Am J Cardiol. 2002;90(9):983-987.
Race Differences in Atrial Fibrillation
COMING SOON!
32. Wang TJ, Larson MG, Levy D, et al. Plasma natriuretic peptide levels and the risk of cardiovascular events and death. N Engl J Med. 2004;350(7):655663. 33. Rossi A, Enriquez-Sarano M, Burnett JC Jr, et al. Natriuretic peptide levels in atrial fibrillation: a prospective hormonal and Doppler-echocardiographic study. J Am Coll Cardiol. 2000;35(5):1256-1262. 34. Kressin NR, Chang BH, Hendricks A, et al. Agreement between administrative data and patients’ self-reports of race/ethnicity. Am J Public Health. 2003;93(10):1734-1739. 35. Healey JS, Baranchuk A, Crystal E, et al. Prevention of atrial fibrillation with angiotensin-converting enzyme inhibitors and angiotensin receptor blockers: a meta-analysis. J Am Coll Cardiol. 2005;45(11):1832-1839. 36. Tsang TS, Gersh BJ. Atrial fibrillation: an old disease, a new epidemic. Am J Med. 2002;113(5):432-435. n
b!Xfc!cbtfe!tpguxbsf!uppm!up!tvqqpsu!nbovtdsjqu! tvcnjttjpo-!usbdljoh-!boe!qffs!sfwjfx/
We Welcome Your Comments The Journal of the National Medical Association welcomes your Letters to the Editor about articles that appear in the JNMA or issues To photocopy, e-mail, post on Internet or relevant to minority healthcare. Address distribute this or any part of207-1555 JNMA, please correspondence to [email protected]. Contact Camille Walker, JNMA Publications Manager, at (202)
visitwww.nmanet.org www.copyright.com.for a or [email protected] for more information. Visit ® downloadable PowerPoint presentation of the new manuscript process.
fKPVSOBMQSFTT
COMING SOON!
b!Xfc!cbtfe!tpguxbsf!uppm!up!tvqqpsu!nbovtdsjqu! tvcnjttjpo-!usbdljoh-!boe!qffs!sfwjfx/
Contact Camille Walker, JNMA Publications Manager, at (202) 207-1555 or [email protected] for more information. Visit www.nmanet.org for a downloadable PowerPoint® presentation of the new manuscript process.
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
eJournal Press.indd 814
VOL. 100, NO. 2, FEBRUARY 2008 245
12/21/07 10:40:37 AM
c
a
r
E
E
r
o
P
P
o
r
246 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
t
u
n
i
t
Y
VOL. 100, NO. 2, FEBRUARY 2008
r
i
g
i
n
a
l
c
o
m
m
u
n
i
c
a
t
i
o
n
Racial Differences in the Prevalence of Atrial Fibrillation among Males Ann M. Borzecki, MD, MPH; D. Keith Bridgers, MD, MSc; Jane M. Liebschutz, MD, MPH; Boris Kader, PhD; Lewis E. Kazis, ScD; and Dan R. Berlowitz, MD, MPH
Financial support: We acknowledge the VA Health Services Research and Development Service and Office of Quality and Performance, who originally supported the 1999 Large Health Survey. Background: Despite being the most common cardiac arrhythmia, little is known about racial differences in atrial fibrillation (AF) prevalence and whether differences persist after accounting for known risk factors. Methods: We identified male respondents to the 1999 Large Health Survey of Veteran Enrollees who had an AF diagnosis in the VA administrative database during the preceding two years. Results: Of 664,754 male respondents, 5.3% had AF. By race, age-adjusted prevalence was 5.7% in whites, 3.4% in blacks, 3.0% in Hispanics, 5.4% in native Americans/Alaskans, 3.6% in Asians and 5.2% in Pacific Islanders (p<0.001). Of predisposing conditions, whites were more likely to have valvular heart disease, coronary artery disease and congestive heart failure, blacks had the highest hypertension prevalence; Hispanics had the highest diabetes prevalence. Racial differences remained after adjustment for age, body mass index and these comorbidities. White males were significantly more likely to have AF compared to all races but Pacific Islanders [versus blacks, OR=1.84 (95% CI: 1.71–1.98); versus Hispanics, OR=1.77 (1.60–1.97); vs Native Americans, OR 1.15 [1.04-1.27]; versus Asians, OR=1.41 (1.12–1.77) versus Pacific Islanders, OR=1.16 (0.88–1.53)]. Conclusions: AF prevalence varies by race. White males have the highest AF burden even after adjustment for known risk factors. Recognition of the high AF prevalence, especially among whites, as well as native Americans and Pacific Islanders, should help guide provider practices for screening among older male patients. Further research is necessary to verify and establish reasons for these racial differences. Key words: atrial fibrillation n epidemiology n race/ ethnicity © 2008. From the Center for Health Quality, Outcomes and Economic Research, Bedford VAMC, Bedford, MA (Borzecki, Kader, Kazis, Berlowitz); Department of Health Policy and Management, Boston University School
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
of Public Health, Boston, MA (Borzecki, Kader, Kazis, Berlowitz); Section of General Internal Medicine, Boston University School of Medicine, Boston, MA (Borzecki, Liebschutz, Berlowitz); and Three Rivers Medical Associates, Irmo, SC (Bridgers). Send correspondence and reprint requests for J Natl Med Assoc. 2008;100:237–245 to: Dr. Ann Borzecki, 200 Springs Road (152), Bedford, MA 01730; phone: (781) 687-2870; fax: (781) 687-3106; e-mail: [email protected]
INTRODUCTION
A
trial fibrillation (AF) has been called an emerging cardiovascular epidemic.1 It is the most common arrhythmia, affecting almost 2.3 million Americans.2 Prevalence increases markedly with age, nearly doubling with each decade of adult life.3 Rates are also increasing over time independent of age.2,4 By 2050, >5.6 million Americans will be affected.2 AF may cause significant morbidity and mortality. It results in a three-fold excess stroke and heart failure risk5,6 and a two-fold increased all-cause mortality risk.3,6 Population-based cohort studies have identified age, male sex, diabetes, hypertension, coronary artery disease, heart failure, valvular heart disease and—more recently—family history and obesity as AF risk factors.5,7-10 These studies have been limited by relatively small sample sizes in predominantly white populations.5,8,11 Surprisingly little is known about racial differences in AF prevalence and whether race may be a determinant of AF. Given AF’s chronicity and the immutability of race, cross-sectional studies may be useful to address this issue. However, even the recent Anticoagulation and Risk Factors in Atrial Fibrillation Study (ATRIA), a cross-sectional study of almost 2 million managed care patients, including 18,000 with AF, offered limited information on racial differences.2 ATRIA only compared whites and African Americans, and only with respect to crude rates that lacked adjustment for known risk factors.2 Risk factors such as hypertension and diabetes are more common in blacks than whites,12 yet ATRIA found higher unadjusted rates in whites.2 Thus, there may be racial differences in AF predilection independent of other AF risk factors. Given this condition’s importance from a public VOL. 100, NO. 2, FEBRUARY 2008 237
Race Differences in Atrial Fibrillation
health and cost perspective, and the growing population of U.S. minorities,13 we examined racial differences in AF prevalence in a large integrated healthcare system. We also examined whether differences remained after accounting for potential factors that might explain varying AF risk.
METHODS Study Participants and Design We merged retrospective data from two national Veterans Health Administration (VA) administrative databases—the Outpatient Clinic File (OPC) and the Patient Treatment File (PTF)—with cross-sectional data from the 1999 Large Health Survey of Veteran Enrollees. The OPC contains information on all VA outpatient visits; the PTF contains inpatient information. Both list up to 10 diagnoses per visit or hospital discharge.14 Large survey details are described elsewhere.15,16 Briefly, 1.4 of 3.4 million VA enrollees were randomly sampled between July 1999 and January 2000, and mailed a questionnaire, including demographic and health status questions. Eligible subjects were survey respondents with complete age, race and health status information who had OPC and PTF data between October 1, 1997 and September 30, 1999. Females were excluded from the sample, given their relatively small proportion (4.1%). We further identified subjects with ≥1 OPC- or PTF-listed AF diagnosis, ICD-9-CM code 427.3x, during this period. We previously found that a single OPC-/PTFderived AF code in a two-year period had substantial kappa (0.74), sensitivity (86%) and specificity (97%), compared to a standard of AF diagnosis documented in the provider’s clinic note.17 We also extracted ICD-9-CM codes for predisposing comorbidities (hypertension, heart failure, coronary artery disease, diabetes, valvular heart disease, chronic lung disease, hyperthyroidism) and potential complications (cerebrovascular disease) from the OPC/PTF
files.7,18 (Table 1). Body mass index (BMI) was based on self-reported height and weight. We identified race based on response to a survey question previously used in federal surveys.19 Subjects were asked to mark all applicable options: American Indian/ Alaskan native, Asian, black/African American, Spanish/ Hispanic/Latino, native Hawaiian/other Pacific Islander and white. Subjects indicating >1 racial group were assigned to whichever group represented a smaller portion of the U.S. population; i.e., subjects marking black and Hispanic were considered black.
Analyses We examined bivariate relationships using Chi-squared tests and odds ratios (ORs) for categorical variables, and Student’s t tests or analysis of variance (ANOVA) for continuous variables. We first calculated overall AF frequency and examined bivariate relationships between subjects with and without AF by age (categorical and continuous), race, BMI (continuous and categorical, <25, 25–<30, ≥30 kg/m2)10 and AF-associated comorbidities. (We initially examined mitral stenosis, mitral insufficiency, tricuspid insufficiency and aortic valve disease separately, then as a combined variable after finding a significant association between all valvular heart disease categories and AF.) We computed race-specific age-adjusted rates by the direct method, using the full sample as the standard population with a breakdown by five-year increments. We also compared racial groups by age (continuous variable), BMI and comorbidities for the full sample, and among only those with AF. Next, we used logistic regression to examine the association between AF and race, controlling for age (continuous), BMI and AF-predisposing comorbidities, and performed pairwise comparisons between racial groups using Bonferroni’s method.20 We also examined this full model minus hyperthyroidism, since this may represent a relatively acute condition. Because 7% of our sample were missing BMI data, we additionally examined models without BMI, and with imputed values using
Table 1. Atrial fibrillation-associated comorbidities with ICD-9-CM codes* Diagnosis ICD-9-CM Code Atrial fibrillation 427.3x Cerebrovascular disease 430.x-438.x Chronic lung disease 490-496.x, 500-505.x, 506.4 Congestive heart failure 398.91, 402.01, 402.11, 402.91, 404.01, 404.03, 404.11, 404.13, 404.91, 404.93, 414.8, 428.x Coronary artery disease 410-414.x, 429.2 Diabetes 250.x Hypertension 401, 402, 405 Hyperthyroidism 242.x, 245.x Valvular heart disease 394.0, 394.1, 394.2, 395.0, 395.1, 395.2, 396.0, 396.1, 396.2, 396.3, 396.8, 397.0, 424.0, 424.1, 746.2, 746.3, 746.4, 746.5, 746.6, 747.22 * Based in part on Borzecki et al. and Deyo et al.17,18
238 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 100, NO. 2, FEBRUARY 2008
Race Differences in Atrial Fibrillation
the mean sample BMI. We considered a p value <0.05 as significant for all analyses, except for pairwise comparisons, where we used a significance level of 0.003 [derived by adjusting alpha for the number of comparisons (i.e., p=0.05/15 or 0.003)].20 To test the robustness of findings, we performed several additional analyses: 1. We examined the number of subjects reporting multiple race responses who thus might have been misclassified and examined the impact of excluding these individuals from our model. 2. Since AF was diagnosed using administrative data, not electrocardiogram (EKG) results, we addressed possible misclassification by repeating logistic models, excluding patients with only one AF diagnosis first from the PTF and then from any source. (Subjects with transient AF following cardiac surgery would be eliminated in this model.) 3. We examined possible nonrespondent bias by repeating analyses, including nonrespondents with available OPC/PTF diagnostic data, and race data derived either from the large survey (for those
with partial survey information who did not meet sample inclusion criteria) or OPC/PTF databases as available. 4. Since we identified AF using the ICD-9-CM code 427.3x, comprising both AF (427.31) and atrial flutter (427.32), we repeated analyses, excluding subjects with only 427.32 codes. While these two conditions are related, frequently coexist and are treated similarly, they nevertheless do represent distinct electrical abnormalities.21 5. We addressed possible ascertainment bias by race in several ways. We examined the: a) likelihood of having an EKG by race via a logistic regression model adjusting for age, BMI, AF-related conditions, having an inpatient stay and number of outpatient visits; b) number of admissions and outpatient visits by race, since additional medical encounters should increase the likelihood of having an EKG; c) source of AF diagnosis; d) impact of adjusting for VA enrollment priority category, which is associated with the likelihood of having non-VA insurance,16 on our original model.
Table 2. Baseline characteristics of sample Characteristics Subjects with AF Subjects without AF N=35,470 N=629,284 n (%) n (%) Age, Year [Mean (SD)] 72.0 (8.8) 63.3 (13.1) Age, Group 18–49 642 (1.8) 97,523 (15.5) 50–64 5,291 (14.9) 197,066 (31.3) 65–79 23,206 (65.4) 283,798 (45.1) ≥80 6,331 (17.8) 50,897 (8.1) Body Mass Index, kg/m2 [Mean (SD)] 27.61 (5.2) 27.65 (5.0) Body Mass Index Category* <25 kg/m2 9,664 (29.5) 163,925 (28.1) 25–<30 kg/m2 14,261 (43.7) 267,237 (45.7) ≥30 kg/m2 8,736 (26.7) 153,092 (26.2) Race White 30,752 (86.7) 476,015 (75.6) Black 2,139 (6.0) 80,442 (12.8) Hispanic 1,001 (2.8) 37,068 (5.9) Native American or Alaska native 1,208 (3.4) 26,489 (4.2) Asian 211 (0.6) 5,992 (1.0) Hawaiian or other Pacific Islander 159 (0.5) 3,278 (0.5) Selected Comorbidities Hypertension 26,588 (75.0) 321,856 (51.1) Coronary artery disease 23,368 (65.9) 160,267 (25.5) Diabetes mellitus 11,417 (32.2) 130,734 (20.8) Chronic lung disease 12,042 (33.9) 117,575 (18.7) Congestive heart failure 13,497 (38.1) 37,784 (6.0) Cerebrovascular disease 7,039 (19.8) 46,440 (7.4) Valvular heart disease 3,278 (9.2) 8,439 (1.3) Hyperthyroidism 651 (1.8) 5,033 (0.8)
P Value
<0.001 <0.001
0.18 <0.001
<0.001†
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
AF: atrial fibrillation; * BMI numbers do not sum to total sample size because of missing data; † By 2x6 Chi-squared test; in 2x2 analyses whites also had a statistically higher prevalence of AF (p<0.001) compared to all other race groups (unadjusted).
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 100, NO. 2, FEBRUARY 2008 239
Race Differences in Atrial Fibrillation
RESULTS
Blacks had the highest hypertension prevalence, Hispanics the highest diabetes prevalence and native Americans the highest chronic lung disease prevalence. Among those with AF, similar comorbidity trends were seen except that blacks were more likely to have heart failure than whites (44.8% vs. 37.4%; OR=1.36, 1.24–1.49), and native Americans had similar coronary disease rates to whites (67.0% vs. 66.2%; OR=1.04, 0.92–1.17). (Full data available from authors.) The association between BMI and AF was nonlinear. Thus, we used a categorical BMI variable in final models. After adjustment for age, BMI and predisposing comorbidities, whites still had significantly higher odds of having AF compared to all other racial groups except Pacific Islanders. (There was no appreciable difference between adjusted models with and without hyperthyroidism.) Whites were 1.8 times more likely to have AF compared to blacks or Hispanics (Table 4). The odds in blacks versus Hispanics were not significantly different. In adjusted models, both blacks and Hispanics had lower AF odds compared to native Americans, Asians and Pacific Islanders, although the difference between Hispanics and Asians was not significant in pairwise comparisons (p=0.006; prespecified significance level <0.003 based on multiple comparisons). Pacific Islanders, native Americans and Asians were not significantly different. Imputing or excluding missing BMI values yielded similar results.
Overall, survey response rate among males was 61% (N=814,292), of whom 664,754 had complete demographic, health status and administrative information. Of these, 35,470 (5.3%) had an AF diagnosis. Table 2 shows respondents’ baseline characteristics. Respondents with AF were older and more likely white than respondents without AF. The prevalence of selected cardiovascular comorbidities such as heart failure and coronary artery disease was much higher in AF subjects than non-AF subjects. Only 9% of the sample had valvular heart disease; most had lone aortic disease (6.3%). Overall age-adjusted prevalence was 5.2%. By race, prevalence was 6.1% in whites, 2.6% in blacks, 2.6% in Hispanics, 4.4% in native Americans/Alaskans, 3.4% in Asians and 4.6% in Pacific Islanders/Hawaiians. The unadjusted odds of whites having AF compared to blacks was 2.43 [95% confidence interval (CI): 2.32–2.54] and 2.39 compared to Hispanics (95% CI: 2.24–2.55). Table 3 shows baseline characteristics by race. Whites were significantly older, while blacks were significantly younger than all other racial groups (respective mean ages 65.0, SD 12.6 years and 59.0, SD 13.7 years; p<0.001). Prevalence increased with age across all races (data not shown). See Figure 1 for age-adjusted prevalences. Whites had a significantly higher age-adjusted prevalence compared to all other racial groups except Pacific Islanders. Conditions predisposing to AF were all significantly associated with AF risk in bivariate analyses. Further, all varied significantly in prevalence by race, except hyperthyroidism (Table 3). Whites had a significantly higher prevalence of several predisposing comorbidities, including coronary disease, heart failure and valvular disease.
Additional Analyses 1. Multiple races were reported by 29,660 subjects (4.5%); 72% of these reported being native
Figure 1. Age-adjusted prevalence of atrial fibrillation by race* *Adjusted using whole group as standard population. Bars represent 95% confidence intervals.
7.
Prevalence (%)
6.0
5.7%
5.0 4.0
5.2%
5.4%
3.4%
3.6% 3.0%
3.0 2.0 1.0 0.0
White
Black
Hispanic
Native American
Asian
Pacific Islander
Race
240 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 100, NO. 2, FEBRUARY 2008
Race Differences in Atrial Fibrillation
American, 89% of whom reported this in addition to being white (18,789 subjects). Excluding these subjects from analyses did not alter findings, except that the white/native-American difference was no longer significant in pairwise comparisons. 2. Rerunning logistic models using more stringent AF diagnosis criteria—first by eliminating those with only one PTF-based AF diagnosis (N=2,666), then eliminating subjects with only one diagnosis in either source (N=11,263)—did not change results. Retaining these subjects but reclassifying them as lacking AF also made little difference. 3. Nonresponder bias. Out of a potential sample of 1.3 million men, demographic and diagnostic information was available on 1,183,745 subjects, including 411,278 nonrespondents and 107,713 additional respondents lacking full survey-derived demographic or health status data, and thus not in the study sample. Male nonrespondents were younger than sample respondents (mean age 55.0 vs. 63.0). Among those with available race information, blacks were least likely to respond (47% vs. 67% of whites). Response rates among other racial groups varied from 55% for native Americans to 64% for Hispanics. Nonrespondents also had a lower prevalence of AF-related comorbidities. AF prevalence by race (based on OPC/PTF demographic information) among nonrespondents was actually very similar to our study sample. It was 6.1% in whites,
3.1% in blacks, 3.0% in Hispanics, 3.3% in native Americans and 4.6% in Asians (there is no separate category for Pacific Islanders in the OPC/PTF.) Respondents not included in the sample were intermediate between sample respondents and nonrespondents with respect to the noted characteristics. Rerunning logistic models including all subjects did not substantially alter results. All previously significant associations remained and CIs narrowed; the Asian/Hispanic association became significant (p<0.003) (data not shown; available upon request). 4. Lone atrial flutter (ICD-9-CM 427.32) was documented in 13% of cases (4,760). Excluding these subjects did not alter results. 5. Ascertainment bias. Forty-two percent of the study sample had an EKG during the ascertainment period. This ranged from 33% for Asians to 43% for native Americans (Table 3). Among AF subjects, 72% had an EKG (ranging from 71% among whites to 80% among Hispanics). After adjustment for factors (other than AF) that increase the likelihood of having an EKG, Hispanics and native Americans were more likely to have an EKG than whites [respective ORs 1.18 (1.15, 1.21) and 1.05 (1.02, 1.08)]; Asians, blacks and Pacific Islanders were less likely than whites to have an EKG [respective ORs 0.81 (0.76, 0.85), 0.89 (0.88, 0.91) and 0.86 (0.80, 0.93)]. (These same associations held
Table 3. Baseline comparison by race* Native Characteristic White Black Hispanic American Asian N=506,767 N=82,581 N=38,069 N=27,697 N=6,203 Age, Years [mean (SD)] 65.0 (12.6) 59.0 (13.7) 60.8 (13.8) 60.1 (12.5) 61.6 (14.9) BMI, kg/m2 [mean (SD)] 27.6 (4.8) 27.5 (5.1) 27.7 (4.6) 28.1 (5.2) 26.3 (4.3) Comorbidities (%) Hypertension 51.9 59.0 48.4 50.2 47.5 Coronary artery disease 30.1 17.3 19.4 26.8 17.6 Diabetes mellitus 20.6 23.5 27.5 21.4 20.5 Chronic lung disease 20.6 14.9 13.3 23.0 18.7 Congestive heart failure 8.1 6.7 5.3 7.4 5.1 Cerebrovascular disease 8.4 7.2 6.1 7.4 5.6 Valvular heart disease 1.9 1.0 1.4 1.5 1.2 Hyperthyroidism 0.9 0.8 1.0 0.9 1.0 Outpatient Visits FY 98–99 [mean (SD)] 15.4 (18.2) 17.3 (25.5) 15.9 (21.0) 16.9 (20.1) 15.0 (22.1) EKG (%) 42.0 38.7 42.0 43.3 33.0 Enrollment Priority Group 1, 4 or 5 (%) 62.6 67.0 70.2 66.9 53.5
Pacific Islander N=3,437 61.4 (13.8) 28.2 (5.1) 48.6 23.3 23.6 15.4 7.4 7.1 1.6 0.9 16.5 (28.2) 37.3 60.0
* All among race comparisons by characteristic are statistically significant (p<0.001), except for hyperthyroidism (p=0.13); BMI: body mass index; EKG: electrocardiogram; † All VA-enrolled veterans are assigned to one of seven priority groups based on serviceconnected disability, income and special considerations (e.g., POW). Subjects in the highest priority groups are exempt from copayments for medical care. Subjects in priority groups 1 (service-connected disability of ≥50%), 4 (catastrophically disabled) and 5 (income and net worth below established thresholds) are least likely to have non-VA insurance.
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 100, NO. 2, FEBRUARY 2008 241
Race Differences in Atrial Fibrillation
after excluding subjects with AF.) Regarding inpatient versus outpatient identification, blacks had more admissions on average than whites, and blacks with AF were more likely to be identified based only on inpatient records (13% vs. 9%). Hispanics and blacks also had more outpatient visits than whites (Table 3), but whites with AF were more likely to have this diagnosis made based solely on outpatient records (68% of whites vs. 66% of Hispanics and 61% of blacks). Despite Hispanics being more likely to belong to priority groups without insurance and Asians being less likely (Table 3), adjustment for priority did not appreciably alter results (data available upon request).
DISCUSSION
Within a large ethnically diverse male veteran population, 5.3% had an AF diagnosis. We found significant racial differences in AF prevalence, with the highest prevalence among whites, the lowest among blacks and Hispanics. These differences remained even after adjustment for age, BMI and comorbidities predisposing to AF. Prevalence was also relatively high among native American and Pacific Islanders compared to blacks and Hispanics, although whites still were most affected. Our study expands current knowledge of AF prevalence among racial groups. Surprisingly little data are available on racial differences among the general population or undifferentiated outpatient populations; most
existing studies examined prevalence in populations with specific conditions such as heart failure or stroke, and included only blacks or Hispanics.2,22-24 Further, only one previous study, involving heart failure patients, controlled for predisposing comorbidities.22 No other studies have included less common minorities such as native Americans, Pacific Islanders or Asians. The relative white/black and white/Hispanic differences we found are consistent with previous studies.2,8,22-24 The ATRIA study found AF in 2.2% of whites and 1.5% of blacks ≥5;2 our respective sample numbers were 6.8% and 3.4%. A cross-sectional study of subjects hospitalized with heart failure found AF in 38% of whites and 20% of blacks;22 respective prevalences among our sample’s heart failure patients were 28% and 17%. Their adjusted OR for AF in whites versus blacks was similar to ours at 1.96 (1.29–2.63). Notably, they performed sequential models comparing AF odds by race, initially adjusting only for age and gender, then similar to our study, for known AF risk factors, including hypertension, coronary disease, prior heart failure, mitral stenosis, hyperthyroidism and chronic lung disease, with diabetes, aortic valve disease, mitral insufficiency and ejection fraction added to final models. While many risk factors were significant predictors in models, their inclusion did not appreciably change the OR found with age and gender alone. The only study with data on Hispanics found a prevalence of 8% in whites, 5% in blacks and 4% in Hispanics among the complete sample in a case-control study examining stroke risk factors,23 AF prevalence among stroke cases
Table 4. Multivariate logistic regression model of atrial fibrillation by race Variable Odds Ratio Race White vs. black 1.84 White vs. Hispanic 1.77 White vs. native American 1.15 White vs. Asian 1.41 White vs. Pacific Islander 1.16 Black vs. Hispanic 0.96 Black vs. native American 0.62 Black vs. Asian 0.76 Black vs. Pacific Islander 0.63 Hispanic vs. native American 0.65 Hispanic vs. Asian 0.79 Hispanic vs. Pacific Islander 0.65 Native American vs. Asian 1.23 Native American vs. Pacific Islander 1.01 Asian vs. Pacific Islander 0.82
95% Confidence Interval* 1.71–1.98† 1.60–1.97† 1.04–1.27† 1.12–1.77† 0.88–1.53 0.85–1.09 0.55–0.70† 0.60–0.97† 0.47–0.84† 0.56–0.75† 0.62–1.02 0.49–0.88† 0.95–1.57 0.75–1.35 0.58–1.18
Each odds ratio also adjusted for age, BMI (<25, 25–<30, ≥30 kg/m2) and predisposing comorbidities (hypertension, coronary artery disease, diabetes mellitus, chronic lung disease, valvular heart disease and hyperthyroidism). Odds ratios indicate the likelihood of having AF. All variables entered into the model were significant predictors of AF (p<0.003); * 95% confidence intervals using Bonferroni adjustment for multiple comparisons; † Significant result. Confidence interval does not include 1.0.
242 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
VOL. 100, NO. 2, FEBRUARY 2008
Race Differences in Atrial Fibrillation
by race was 29%, 11% and 11%, respectively.24 Comparable percentages among our cerebrovascular disease patients were 14%, 8% and 10%. Our relative racial differences in cardiovascular comorbidity prevalence are also consistent with other studies.22,24,25 In the previous heart failure and AF study,22 similar to our study, blacks were more likely to have hypertension, and whites were more likely to have coronary disease. The stroke case-control study also found diabetes most prevalent in Hispanics.24 Among AF subjects, the Atrial Fibrillation Follow-Up Investigation of Rhythm Management study also found hypertension most common in blacks and coronary disease more common in whites. Like our AF cohort, heart failure was more common in blacks, although it was also more common in Hispanics than whites.25 We have shown that racial variation in AF prevalence cannot be explained by traditional AF risk factors. For instance, whites had a higher odds of AF than blacks despite greater hypertension prevalence in blacks and despite the fact that, as in other populations, black veterans with hypertension are more likely to have uncontrolled blood pressure than whites.12,26 The higher risk in whites is also likely not explained by unmeasured factors such as lower medication adherence with resulting poor control of underlying conditions. Several studies in veterans have found higher adherence rates among whites compared to blacks and nonwhites with respect to cardiac medications.27,28 Other VA-based research has similarly found differential racial predilection for chronic conditions, i.e., among veterans with diabetes, blacks and native Americans were more likely to have nephropathy, whites were more likely to have coronary disease and heart failure.29 Some authors have conjectured that AF-related racial differences may be due to ascertainment bias, i.e., differential access to care in minorities, including diagnostic testing, may result in lower AF diagnosis rates.22 We did not find evidence to support major differences in access to clinical services, with high healthcare use among all racial groups as evidenced by a mean of ≥15 visits over the two-year period. (Table 3.) Moreover, several minority groups had more outpatient visits than whites, thereby presenting more opportunities for AF diagnosis. Only 42% of the entire sample had an EKG at any time during the ascertainment period, with blacks less likely, and Hispanics actually more likely, to have an EKG versus whites. Presumably, many individuals had an EKG done either before the ascertainment period or in a nonVA healthcare setting. Why racial bias should occur in the first case is unclear. While non-VA healthcare use does vary by race, lack of information on non-VA use should not have altered our findings. Researchers have found that among veterans, blacks, Hispanics and other races are less likely to have non-VA healthcare coverage than whites.16 However, accounting for enrollment JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
priority as a proxy for having non-VA insurance did not change our results. These findings suggest that racial prevalence differences are real and not because whites are more likely to receive healthcare or EKGs and, thus have AF diagnosed. Reasons for racial variation are unclear but presumably include variations in genetic or environmental factors. According to Ruo, “intrinsic racial differences in atrial membrane stability, atrial conduction pathways or genetic polymorphisms [may] lead to different susceptibility” to AF development.22 Racial polymorphism differences have been associated with heart failure risk.30 The Cardiovascular Health Study found racial differences in left atrial size; average left atrial size was greater in elderly white males compared to blacks.31 Further, there may be racial differences in plasma or brain natriuretic peptide levels, both of which have been linked to AF development.32,33 Our study has several strengths. Our sample of 664,754, comprising 20% of the total VA population, represents one of the largest samples studied for AF prevalence. These large numbers allow for a relatively precise estimation of race-specific AF prevalence. We used self-reported race/ethnicity, which is generally accepted as the most reliable and valid source of such information, especially for ethnicities other than white or black. The large survey data is also more complete than VA administrative racial data.34 New information from our study includes data on AF prevalence and risk among native Americans, Pacific Islanders and Asians. Further, we are the first to adjust for factors that might account for ethnic prevalence differences and reached similar conclusions to unadjusted results. Additionally, we performed several sensitivity analyses that support the robustness of our findings. Study limitations include use of VA males who tend to have more comorbidities, so actual prevalence rates may be less generalizable to other settings.15 However, this should not affect our adjusted results. Moreover, the consistency of our relative AF and comorbidity racial prevalence differences with other studies strengthens the external validity of our findings.22,24,25 Given the relatively few females, we excluded them from analyses since we would be unable to draw reliable conclusions. Even among males there were relatively few subjects in the smallest minority groups with AF such that we may have been unable to find statistical differences when true differences existed. Nonetheless, this represents the largest existing sample of such minorities. Given the cross-sectional study design, we cannot definitively prove that race is a risk factor for AF. However, our large sample, the relative chronicity of AF and the unchanging nature of race make this conclusion highly likely. Our diagnosis is based administrative data, not EKGs. However, the ATRIA study similarly identified AF subjects using administrative data, but suppleVOL. 100, NO. 2, FEBRUARY 2008 243
Race Differences in Atrial Fibrillation
mented with EKGs.2 Of almost 18,000 AF subjects, 90% had AF by ICD-9-CM codes, 87% had EKG confirmation. Subjects with only one AF code were less likely to have EKG confirmation. In our study, findings were unchanged by reclassifying those with only one diagnosis as not having AF; additionally, we did not find a predilection for whites to have EKGs. We also cannot determine from ICD-9-CM coding what AF type (persistent or paroxysmal) patients had, and lack left ventricular function data. However, as noted, adding ventricular function to models in another study minimally impacted adjusted ORs.22 A recent meta-analysis also suggested decreased AF risk among heart failure patients treated with angiotensin blockers.35 While we did not examine medication use, we have no reason to suspect a bias against using/prescribing this medication in high-risk groups such as whites. Finally, our data are several years old. Other data have shown that AF prevalence is increasing over time, presumed partly due to the aging of the population and partly due to the fact that subjects who previously might have died from pre-existing or existing comorbidities are living longer,36 such that we may be slightly underestimating the prevalence in current veterans. However, there is no evidence to suggest a differential effect in prevalence based on race over time. Further, given the limited literature regarding the epidemiology of AF in nonwhite populations and the growth of such minorities in the United States, data on these groups are clearly necessary. By providing such information, this study makes an important contribution to the existing literature. This study establishes race as an independent risk factor for AF among males. The highest risk was seen in whites, followed by Pacific Islanders and native Americans; blacks and Hispanics had the lowest risk. Given the general population’s aging, AF will become an increasingly important cause of morbidity and mortality, especially in these high-risk populations. Recognition of the high AF risk in these racial groups should guide provider practices for screening among older patients and help inform policy on healthcare resource distribution. Further research is also warranted to establish reasons for these racial predispositions. Such research, by establishing mechanisms through which AF develops, should help curb this growing cardiovascular epidemic.
Acknowledgements
Special thanks to Mark Glickman, PhD, for statistical assistance, and to Marshall Goff for editorial assistance.
REFERENCES
1. Braunwald E. Shattuck lecture—cardiovascular medicine at the turn of the millennium: triumphs, concerns, and opportunities. N Engl J Med. 1997;337(19):1360-1369. 2. Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke
244 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA. 2001;285(18):2370-2375. 3. Benjamin EJ, Wolf PA, D’Agostino RB, et al. Impact of atrial fibrillation on the risk of death: the Framingham Heart Study. Circulation. 1998;98(10):946952. 4. Wolf PA, Benjamin EJ, Belanger AJ, et al. Secular trends in the prevalence of atrial fibrillation: the Framingham Study. Am Heart J. 1996;131(4):790-795. 5. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke. 1991;22(8):983-988. 6. Stewart S, Hart CL, Hole DJ, et al. A population-based study of the longterm risks associated with atrial fibrillation: 20-year follow-up of the Renfrew/Paisley study. Am J Med. 2002;113(5):359-364. 7. Benjamin EJ, Levy D, Vaziri SM, et al. Independent risk factors for atrial fibrillation in a population-based cohort. The Framingham Heart Study. JAMA. 1994;271(11):840-844. 8. Psaty BM, Manolio TA, Kuller LH, et al. Incidence of and risk factors for atrial fibrillation in older adults. Circulation. 1997;96(7):2455-2461. 9. Fox CS, Parise H, D’Agostino RB Sr, et al. Parental atrial fibrillation as a risk factor for atrial fibrillation in offspring. JAMA. 2004;291(23):2851-2855. 10. Wang TJ, Parise H, Levy D, et al. Obesity and the risk of new-onset atrial fibrillation. JAMA. 2004;292(20):2471-2477. 11. Phillips SJ, Whisnant JP, O’Fallon WM, et al. Prevalence of cardiovascular disease and diabetes mellitus in residents of Rochester, Minnesota. Mayo Clin Proc. 1990;65(3):344-359. 12. Hajjar I, Kotchen TA. Trends in prevalence, awareness, treatment, and control of hypertension in the United States, 1988–2000. JAMA. 2003;290(2):199-206. 13. U.S. Census Bureau. State Population Estimates by Race: July 1, 2002. http://eire.census.gov/popest/data/states/ST-EST2002-ASRO-03.php. Accessed 12/20/03. 14. VA Information Resource Center: Toolkit for new users of VA data. www. virec.research.med.va.gov/Support/Training-NewUsersToolkit/Toolkit.htm. Accessed 05/15/04. 15. Perlin J, Kazis L, Skinner K, et al. Health status and outcomes of veterans. 1999 large health survey of veteran enrollees. Executive report. Washington, DC: Department of Veterans Affairs and Veterans Health Administration, Office of Quality and Performance; 2000. 16. Shen Y, Hendricks A, Zhang S, et al. VHA enrollees’ health care coverage and use of care. Med Care Res Rev. 2003;60(2):253-267. 17. Borzecki AM, Wong AT, Hickey EC, et al. Identifying hypertension-related comorbidities from administrative data: what’s the optimal approach? Am J Med Qual. 2004;19(5):201-206. 18. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613-619. 19. Office of Management and Budget. Standards for maintaining, collecting and presenting federal data on race and ethnicity. www.whitehouse. gov/omb/inforeg/r&e_app-a-update.pdf. 20. Fisher LD, van Belle G. Multiple Comparisons. In: Biostatistics. A Methodology for the Health Sciences. New York, NY: John Wiley & Sons, Inc.; 1993. 21. Waldo AL. The interrelationship between atrial fibrillation and atrial flutter. Prog Cardiovasc Dis. 2005;48(1):41-56. 22. 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(3):429-435. 23. Sacco RL, Boden-Albala B, Abel G, et al. Race-ethnic disparities in the impact of stroke risk factors: the northern Manhattan stroke study. Stroke 2001;32(8):1725-1731. 24. Sacco RL, Kargman DE, Zamanillo MC. Race-ethnic differences in stroke risk factors among hospitalized patients with cerebral infarction: the Northern Manhattan Stroke Study. Neurology. 1995;45(4):659-463. 25. Bush D, Martin LW, Leman R, et al. Atrial fibrillation among African Americans, Hispanics and Caucasians: clinical features and outcomes from the AFFIRM trial. J Natl Med Assoc. 2006;98(3):330-339. 26. Borzecki AM, Glickman ME, Kader B, et al. The effect of age on hypertension control and management. Am J Hypertens. 2006;19(5):520-527.
VOL. 100, NO. 2, FEBRUARY 2008
fKPVSOBMQSFTT
27. Charles H, Good CB, Hanusa BH, et al. Racial differences in adherence to cardiac medications. J Natl Med Assoc. 2003;95(1):17-27. 28. Kopjar B, Sales AE, Pineros SL, et al. Adherence with statin therapy in secondary prevention of coronary heart disease in veterans administration male population. Am J Cardiol. 2003;92(9):1106-1108. 29. Young BA, Maynard C, Boyko EJ. Racial differences in diabetic nephropathy, cardiovascular disease, and mortality in a national population of veterans. Diabetes Care. 2003;26(8):2392-2399. 30. Small KM, Wagoner LE, Levin AM, et al. Synergistic polymorphisms of beta1- and alpha2C-adrenergic receptors and the risk of congestive heart failure. N Engl J Med. 2002;347(15):1135-1142. 31. Manolio TA, Gottdiener JS, Tsang TS, et al. Left atrial dimensions determined by M-mode echocardiography in black and white older (≥65 years) adults (The Cardiovascular Health Study). Am J Cardiol. 2002;90(9):983-987.
Race Differences in Atrial Fibrillation
COMING SOON!
32. Wang TJ, Larson MG, Levy D, et al. Plasma natriuretic peptide levels and the risk of cardiovascular events and death. N Engl J Med. 2004;350(7):655663. 33. Rossi A, Enriquez-Sarano M, Burnett JC Jr, et al. Natriuretic peptide levels in atrial fibrillation: a prospective hormonal and Doppler-echocardiographic study. J Am Coll Cardiol. 2000;35(5):1256-1262. 34. Kressin NR, Chang BH, Hendricks A, et al. Agreement between administrative data and patients’ self-reports of race/ethnicity. Am J Public Health. 2003;93(10):1734-1739. 35. Healey JS, Baranchuk A, Crystal E, et al. Prevention of atrial fibrillation with angiotensin-converting enzyme inhibitors and angiotensin receptor blockers: a meta-analysis. J Am Coll Cardiol. 2005;45(11):1832-1839. 36. Tsang TS, Gersh BJ. Atrial fibrillation: an old disease, a new epidemic. Am J Med. 2002;113(5):432-435. n
b!Xfc!cbtfe!tpguxbsf!uppm!up!tvqqpsu!nbovtdsjqu! tvcnjttjpo-!usbdljoh-!boe!qffs!sfwjfx/
We Welcome Your Comments The Journal of the National Medical Association welcomes your Letters to the Editor about articles that appear in the JNMA or issues To photocopy, e-mail, post on Internet or relevant to minority healthcare. Address distribute this or any part of207-1555 JNMA, please correspondence to [email protected]. Contact Camille Walker, JNMA Publications Manager, at (202)
visitwww.nmanet.org www.copyright.com.for a or [email protected] for more information. Visit ® downloadable PowerPoint presentation of the new manuscript process.
fKPVSOBMQSFTT
COMING SOON!
b!Xfc!cbtfe!tpguxbsf!uppm!up!tvqqpsu!nbovtdsjqu! tvcnjttjpo-!usbdljoh-!boe!qffs!sfwjfx/
Contact Camille Walker, JNMA Publications Manager, at (202) 207-1555 or [email protected] for more information. Visit www.nmanet.org for a downloadable PowerPoint® presentation of the new manuscript process.
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
eJournal Press.indd 814
VOL. 100, NO. 2, FEBRUARY 2008 245
12/21/07 10:40:37 AM
c
a
r
E
E
r
o
P
P
o
r
246 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
t
u
n
i
t
Y
VOL. 100, NO. 2, FEBRUARY 2008