independant risk factors for atrial fibrillation in a population-bases cohort

8/10/2019 Independant Risk Factors for Atrial Fibrillation in a Population-Bases Cohort

http://slidepdf.com/reader/full/independant-risk-factors-for-atrial-fibrillation-in-a-population-bases-cohort 1/5

Independent Risk Factors for AtrialFibrillation in a Population-Based CohortThe Framingham Heart StudyEmelia J. Benjamin, MD, ScM; Daniel Levy, MD; Sonya M. Vaziri, MD, MPH; Ralph B. D'Agostino, PhD;Albert J. Belanger, MA; Philip A. Wolf, MD

Objective.\p=m-\Todetermine the independent risk factors for atrial fibrillation.Design.\p=m-\Cohortstudy.Setting.\p=m-\TheFramingham Heart Study.Subjects.\p=m-\Atotal of 2090 men and 2641 women members of the original co-

hort, free of a history of atrial fibrillation, between the ages of 55 and 94 years.Main Outcome Measures.\p=m-\Sex-specificmultiple logistic regression models to

identify independent risk factors for atrial fibrillation, including age, smoking, diabe-tes, electrocardiographic left ventricular hypertrophy, hypertension, myocardial in-farction, congestive heart failure, and valve disease.

Results.\p=m-\Duringup to 38 years of follow-up, 264 men and 298 women devel-oped atrial fibrillation. After adjusting for age and other risk factors for atrial fibrilla-tion, men had a 1.5 times greater risk of developing atrial fibrillation than women.

In the full multivariable model,

the odds ratio(OR)

of atrial fibrillation for each de-cade of advancing age was 2.1 for men and 2.2 for women (P<.0001). In addition,after multivariable adjustment, diabetes (OR, 1.4 for men and 1.6 for women), hy-pertension (OR, 1.5 for men and 1.4 for women), congestive heart failure (OR, 4.5for men and 5.9 forwomen), and valve disease (OR, 1.8 for men and 3.4 forwomen)were significantly associated with risk for atrial fibrillation in both sexes. Myocardialinfarction (OR, 1.4) was significantly associated with the development of atrialfibrillation in men. Women were significantly more likely than men to have valvularheart disease as a risk factor for atrial fibrillation. The multivariable models were

largely unchanged after eliminating subjects with valvular heart disease.Conclusion.\p=m-\Inaddition to intrinsic cardiac causes such as valve disease and

congestive heart failure, risk factors for cardiovascular disease also predispose toatrial fibrillation. Modification of risk factors for cardiovascular disease may have theadded

benefit of diminishing the

incidence of atrial

fibrillation.(JAMA. 1994;271:840-844)

From The Framingham (Mass) Heart Study (DrsBenjamin, Levy, Vaziri, D'Agostino, and Wolf and MrBelanger); the Cardiology Section, Boston (Mass) CityHospital, Boston University School of Medicine (DrBenjamin); National Heart, Lung, and Blood Institute,Bethesda, Md (Dr Levy); Department of Mathematics,Boston University (Dr D'Agostino and Mr Belanger);Departments of Preventive Medicine (Drs Benjamin,Levy, and Wolf) and Neurology (Dr Wolf), Boston Uni-versity School of Medicine; and Divisions of Cardiologyand Clinical Epidemiology, Beth Israel Hospital, Boston(Dr Levy).

Presented in part at the 33rd Annual Conference on

Cardiovascular Disease Epidemiology, AmericanHeart Association, Santa Fe, NM, March 19, 1993.

Reprint requests to Framingham Heart Study, 5Thurber St, Framingham, MA 01701 (Dr Benjamin).

ATRI AL fibrillationoccurs with increas¬ing frequency as people age and is re¬

sponsible for substantial morbidity andmortality.1 Present in only 0.5% of50- to59-year-old subjects, the lifetime preva¬lence of atrial fibrillation is nearly 9%among 80- to 89-year-old subjects.2 Non-valvular atrial fibrillation is associatedwith approximately a threefold to five¬fold increased risk for stroke after ad¬ justing for other stroke risk factors; bythe ninth decade of life, atrial fibrilla¬tion is the most importantrisk factor for

stroke.2 Furthermore, atrial fibrillationhas a deleterious impact on longevity,3with an approximate doubling of all-cause mortality.4

To successfully prevent atrial fibril¬lation, physicians must understand thefactors that predispose to its develop¬ment. Earlier reports have emphasizedvalvular heart disease, coronary heartdisease, congestiveheart failure, and hy¬pertensive heart disease as risk factorsfor atrial fibrillation.1'412 However, thesestudies lacked multivariable analysesand thus were limited in their abilityto

identifyindependentcontributors to thedevelopment of atrial fibrillation. Thepurpose of this study was to examinethe risk factors for atrial fibrillation insubjects of the Framingham HeartStudy, for whom nearly 40 years of fol¬low-up are available.

METHODS

Study PopulationThe study is based on the original

cohort of the Framingham Heart Study,which was initiated in 1948 with sub¬

jects aged 28 to 62 years to prospec-tively observe the risk factors for car¬diovascular disease. Details of the studydesignhave been publishedpreviously.13Routine biennial examinations were per¬formed on the original sample of 5209men and women. All subjects gave in¬formed consent. The current investiga¬tion was based on 38 years of follow-up.Because atrial fibrillation was rare inyounger subjects and there were fewsubjects over 94 years old, analysis w as

restricted to subjectsaged 55 to 94 yearsat baseline. In addition, the 18 subjectswith atrial fibrillation at examination 1were excluded.

wnloaded From: http://jama.jamanetwork.com/ by a Rijksuniversiteit Groningen User on 09/11/2014



Clinical Variables

At each Framingham Heart Studyclinic examination, subjects'medical his¬tories, physical examinations, and elec¬trocardiograms were obtained to ascer¬

tain symptoms and findings suggestiveof cardiovascular disease. The recordsof all interim hospitalizations for car¬

diovascular disease were soughtand sub¬sequentlyreviewed. Subjects were con¬

sidered to have had atrial fibrillation ifatrial fibrillation or atrial flutter was

verified on review of the electrocardio¬gram by one of two Framingham HeartStudy cardiologists. Electrocardiogramsfrom the following sources were evalu¬ated: routine Framingham Heart Studyclinicexaminations,hospitalrecords, andoutside physician records. A separateanalysis of subjects with atrial fibrilla¬tion present on routine FraminghamHeart Study clinic electrocardiogramswas undertaken to examine the risk fac¬tors for nonacute atrial fibrillation, andto determine if the inclusion of electro¬cardiogramsperformed duringhospital-ization resulted in ascertainment bias.

Potential risk factors were assessedat clinic examinations. Myocardial in¬farction and congestive heart failurewere determined accordingto previouslypublished criteria14 by a panel of threephysicians after review of FraminghamHeart Study and outside hospital andphysician records. Subjects were clas¬sified as having myocardialinfarction or

congestive heart failure if the conditionwas diagnosed at or prior to the dayatrial fibrillationwas documented. Valvedisease was defined as any diastolic mur¬

mur or at least a grade 3/6 systolic mur¬

mur on Framingham Heart Studyphysi¬cal examination. Because echocardio-graphy was unavailable for the first 30years of the study, it was not utilized toestablish the diagnosis of valve disease.Hypertension was defined for each ex¬

amination as a systolicblood pressure ofat least 160 mm Hg or a diastolic bloodpressure of at least 95 mm Hg on eachof two readings by Framingham Heart

Study physicians, o r the use of antihy-pertensive medication. Diabetes melli-tus was diagnosed if the subject had a

fastingblood glucose level of7.77 mmol/Lor higher (>140mg/dL); a random, non-

fasting blood glucose level of 11.11mmol/L or higher (>200 mg/dL); or was

taking insulin or an oral hypoglycémieagent at the current examination o r atany priorFramingham Heart Study ex¬

amination. Forced expiratoryvolume in1 second (FEVj) was used as an indica¬tor of lung function. Alcohol use was

ascertained by self-report and catego¬rized as ounces of ethanol consumed perweek. Body mass index was calculated

as the weight (in kilograms) divided bythe height (in meters), squared. Elec-trocardiographic left ventricular hyper¬trophy was considered present if thesubject fulfilled voltage criteria for hy¬pertrophy and manifested lateral repo-larization changes.15 Because electrocar-diographicleft ventricular hypertrophycannot be interpreted in the setting ofleft bundle-branch block, subjects were

excluded from the multivariable analy¬ses if left bundle-branch block was

present.Statistical Analyses

Logistic regression based on pooledbiennial person-examination data was

used to evaluate the association of po¬tential risk factors with the developmentof atrial fibrillation.16,17 (Each examina¬tion with its 2-year follow-up period was

considered a separate person-examina¬tion.) At each biennial examination, allsubjects aged 55 to 94 years were evalu¬ated for a history of atrial fibrillation.Incident atrial fibrillation w as defined asthe occurrence of atrial fibrillation docu¬mented by electrocardiogram up to thetime of the next examination. Putativerisk factors were measured at each bi¬ennial examination. Missing data were

imputed by substituting the most recentvalues as long as they were obtainedwithin the two preceding examinations;this occurred for 10% of hypertension,10% of left ventricular hypertrophy, and25%ofsmokingvariables. Potential clini¬cal risk factors for atrial fibrillation were

examined first by sex-specific bivariatemodels adjustingfor age. Variables thatwere significant in the bivariate modelswere entered into sex-specific multivari¬able models. Results of the regressionswere expressed as odds ratios (with 95%confidence intervals) and population-at¬tributable risks: [prevalence(odds ratio

- l)]/[prevalence(odds ratio - 1) + l].18To evaluate the interaction of sex with

the risk factors, the data for men andwomen were combined and sex-inter¬action terms were examined for all po¬tential risk factors. Backwards elimina¬tion was employed to remove all non¬

significant interactions (P>.05was usedfor elimination).

RESULTSCharacterization of Subjects

During up to 38 years of follow-up of2090 men (whocontributed 18245person-examinations of follow-up) and 2641women (who contributed 26 277 person-examinations of follow-up) with no his¬tory of atrial fibrillation, there were 562incident cases ofatrial fibrillation.Amongsubjectswith atrial fibrillation,298 (53%)were women and 264 (47%) were men.

Incidence of atrial fibrillation, Framingham HeartStudy 38-year follow-up. Darker bars indicate men;and lighter bars, women. Vertical axis indicates bi¬ennial rate per 1000 person-examinations (approxi¬mately 2000 person-years).The biennial incidence of atrial fibrilla¬tion increased with age (Figure), rang¬ingfrom 6.2and 3.8 cases per 1000person-examinations in men and women, respec¬tively, aged 55 to 64 years, to 75.9 and62.8 cases per 1000 person-examinationsin men and women aged 85 to 94 years.Table 1 shows the sex-specific, age-ad¬ justed characteristics ofsubjects accord¬ing to incident atrial fibrillation statusand the age-adjusted odds ratios of de¬veloping atrial fibrillation for each vari¬able. Subjects who developed atrial fi¬brillation were older and more likely tohave diabetes mellitus, left ventricularhypertrophy, hypertension, myocardialinfarction, congestive heart failure, andvalvular heart disease than subjects with¬out atrial fibrillation. Body mass index (ameasure ofobesity) and alcohol use were

not significantly related to the develop¬ment of atrial fibrillation in the age-ad¬

justed models; hence, they were not in¬cluded in the multivariable models.

Multivariable AnalysesIn both sexes, age, diabetes, hyper¬

tension, congestiveheart failure, andval¬vular heart disease were significantpre¬dictors of atrial fibrillation (Table 2).For each advancing decade of age theodds ratio of developing atrial fibrilla¬tion was 2.1 in men and 2.2 in women.

Congestive heart failure was associatedwith an increased risk of developingatrial fibrillation with an odds ratio of4.5 in men and 5.9 in women. The pres¬ence of valvular heart disease was as¬sociated with an increased risk of de¬veloping atrial fibrillation (odds ratio of1.8 in men and 3.4 in women). Myocar¬dial infarction was significantly associ¬ated with the development of atrial fi¬brillation in men (odds ratio, 1.4).

The population-attributablerisk analy¬ses are presented in Table 3. Population-attributable risk incorporates both the




Table 1.—Risk Factors for Developing Incident Atrial Fibrillation (AF)*

Variable

Ment Woment

AF No AFOdds Ratio

(95% CI) AF No AFOdds Ratio

(95% CI)Age (mean), y 72 65 75 66

Cigarette smoking, % 33.7 1.0(0.8-1.4) 28.5 23.4 1.4§ (1.0-2.0)Diabetes, % 16.3 10.2 1.7||(1.2-2.3) 15.5 7.5 2.1H(1.5-2.8)ECG LV hypertrophy, % 10.7 4.4 3.01) (1.9-4.8) 13.6 3.8 3.811 (2.6-5.6)Hypertension. % 44.1 30.9 1.811(1.4-2.3) 51.7 40.7 1.7H(1.3-2.2)Myocardial infarction, % 25.5 13.0 2.2I1 (1.6-2.8) 13.0 4.6 2.411(1.7-3.4)Congestive heart failure, % 3.2 6.111(4.5-8.4) 2.9 8.111(6.1-10.7)Valve disease, % 16.7 6.7 2.211(1.6-3.1) 29.5 8.7 3.611 (2.8-4.6)Body mass index (mean) 26.2 26.0 1.03(0.99-1.06) 26.0 25.7 1.02(1.00-1.05)Ethanol, oz/wk 5.4 5.1 1.01 (0.99-1.03) 1.5 1.8 0.95(0.89-1.02)

"Variables are expressed as age-adjusted % (categorical data) and means (continuous data); odds ratios are

age-adjusted from 2-year pooled logistic regression, subjects with compared with those without AF. CI indicatesconfidence interval; ECG, electrocardiographs; and LV, left ventricular.

tThere were 264 Incident cases of AF among men In 18 245 follow-up person-examinations.iThere were 298 Incident cases of AF among women in 26 277 follow-up person-examinations.§P=s.05.IIPs.01.IJPs.0001.

odds ratio of the risk factor and the riskfactor's prevalence in the study popula¬tion. It reflects the percentage of atrialfibrillation cases that potentially wouldbe reduced if a causal risk factor could beeliminated from the population. The at¬tributable risk of congestive heart fail¬ure for atrial fibrillation was 10% in men

and 12% in women. The attributable riskfor atrial fibrillation resultingfrom myo¬cardial infarction was greater in men (5%)than in women (1%), while the attribut¬able risk from valve disease was approxi¬mately threefold higher in women (18%)than in men (5%). In part because of thehigh prevalence of hypertension in the

elderly (31% of men, 40% of women), itspopulation-attributable risk for atrial fi¬brillation was 14% in both men andwomen. Electrocardiographic left ven¬

tricular hypertrophy and diabetes eachaccounted for less than 5% of the attrib¬utable risk for atrial fibrillation.

Impact of Sex.—After adjusting forage, smoking, diabetes, left ventricularhypertrophy, hypertension, myocardialinfarction, congestive heart failure, andvalvular heart disease, men had a 1.5times greater likelihood of developingatrial fibrillation than women (95% con¬

fidence interval, 1.3 to 1.8; P<.0001). Toexplore potential sex differences in therisk factors for atrial fibrillation, themodels were reanalyzed with sex-inter¬action terms. While the models for men

and women were similar, valvular heartdisease was a significantly (P<.01) more

potent risk factor for the developmentof atrial fibrillation in women than inmen.

Nonacute Atrial Fibrillation.—To as¬sess risk factors for nonacute atrial fi¬brillation, a separate multivariableanalysis identified the risk factors foratrial fibrillation diagnosed on routineFramingham Heart Study clinic elee-

trocardiograms (Table 4). While thenumber of atrial fibrillation cases was

much smaller, and hence some of therisk factors no longer achieved statis¬tical significance, the odds ratios of theclinical risk factors for atrial fibrillationwere similar to those found in the entiresample, except for myocardial infarc¬tion,which had an odds ratio of less thanone.

Additional AnalysesNonvalvular Atrial Fibrillation.—

Since there is little doubt that valvularheart disease predisposes to atrial fi¬brillation, Table 5 presents multivari¬able models restricting the analysis tosubjects without clinically apparentvalve disease (thereby eliminating 136cases of atrial fibrillation). In men andwomen without valvular heart disease,diabetes was no longer significantly as¬

sociated with the development of atrialfibrillation,while age, hypertension, andcongestive heart failure remained sig¬nificant predictors. In women withoutvalvular heart disease, myocardial in¬farction also was significantly associatedwith the development of atrial fibrilla¬tion.

Lung Disease.—Because of the sug¬gestion of a relation between lung dis¬ease and atrial fibrillation,5·6 the asso¬

ciation of FEVi and incident atrial fi¬brillation was examined. When FEVjwas added to the full multi variable mod¬els, it was not significantly associatedwith atrial fibrillation in either sex (oddsratio, 1.00/0.10 L in both sexes).COMMENT

The incidence of atrial fibrillation ap¬proximately doubled for every 10-yearincrement in age in the FraminghamHeart Study cohort. In both sexes, car¬

diovascular disease risk factors (ie, dia-

Table 2.—Multivariable Risk of Developing IncidentAtrial Fibrillation (2-YearPooled Logistic Regression)

Odds Ratio(95% Confidence Interval)

I-1Variable Men* Womenf

Age (/10 y) 2.1 (1.8-2.5) 2.2 (1.9-2.6)Cigarette smoking 1.1(0.8-1.5) 1.4(1.0-2.0)Diabetes 1.4§ (1.0-2.0) 1.6|| (1.1-2.2)Electrocardiographic

left ventricular

hypertrophy 1.4(0.9-2.4)

1.3(0.9-2.1)Hypertension 1.5|| (1.2-2.0) 1.4§ (1.1-1.8)Myocardial infarction 1.4§ (1.0-2.0) 1.2(0.8-1.8)Congestive heart

failure 4.5$ (3.1-6.6) 5.9$ (4.2-8.4)Valve disease 1.811(1.2-2.5) 3.4$ (2.5-4.5)

*Atrial fibrillation was diagnosed in 226 men in 16 529follow-up person-examinations.

tAtrialfibrillationwas diagnosed in244 women in23 763follow-up person-examinations.

$P==.0001.§Ps.05.IIPS.01.

betes and hypertension) and structuralcardiac disease

(ie, congestive heart fail¬

ure and valvular heart disease) also were

risk factors for atrial fibrillation.In the present study, men were 1.5

times more likely to develop atrial fi¬brillation than women, even after ad¬

justing for other risk factors for atrialfibrillation. The reason for the greatersusceptibility of men to atrial fibrilla¬tion remains unclear. With rare excep¬tion,6 most prior publications also havenoted that men are at greater risk foratrial fibrillation than women.1'6,9·11·19·20In a retrospective hospital-based analy¬sis,5 men were 1.2 times more likely tohave atrial fibrillation than women. Menwere noted to have a 5.4 times greaterrisk of having atrial fibrillation thanwomen in the CoronaryArterySurgeryStudy registry of patients catheterizedfor suspected coronary artery disease.20In a population-based study, men were

2.7 times more likely to have atrial fi¬brillation than women; however, the au¬

thors did not examine whether thehigherprevalence ofatrial fibrillation inmen merely reflected the greater bur¬den of atrial fibrillation risk factors inmen.11

In the current study, women were

significantly more likely than men tohave valvular heart disease as a riskfactor for atrial fibrillation, despite thesimilar prevalence of valvular heart dis¬ease in men and women (7% and 9%,respectively, in the total sample). Withthe exception of that by Stroud et al,9other studies have found a similar fe¬male predominance in the cases ofatrialfibrillation associated with valvularheartdisease, although these studies did notuse multivariable analyses.5·6·11 The bio¬logic mechanism for the greater suscep¬tibility of women with valvular heartdisease to the development of atrial fi-




brillation may be secondary to the fe¬male predisposition to rheumatic mitralstenosis21 and the propensity of mitralstenosis to cause atrial fibrillation more

frequently than other valve lesions.Previous studies have identified car¬

diac and noncardiac précipitants ofatrialfibrillation. Potential cardiac etiologiesof atrial fibrillation have been noted toinclude valvular heart disease (rheu¬matic, myxomatous, senile calcific, andcongenital valve abnormalities)511; myo¬cardial disease (acute myocardialinfarc¬tion and ischemie heart disease,10,12 con¬

gestive heart failure,11 hypertensiveheart disease,10 myocarditis,5"7 and hy¬pertrophie cardiomyopathy22); conduc¬tion system disease (Wolff-Parkinson-White syndrome)6; congenital heart dis¬ease (atrial septal defect, patent ductusarteriosus, and so forth)6; and pericar-dial disease6 (including pericarditis andconstriction). The noncardiac etiologiesof atrial fibrillation have been reportedto include thyrotoxicosis,510 alcoholuse,5·12 severe infections,5and pulmonarypathology (lung cancer, pulmonary in¬fections, pulmonary embolism, andchronic lung disease).58 While clinicalseries provide insights into the poten¬tial risk factors for atrial fibrillation inthe individual patient, they are limitedin their generalizability to the broaderpopulation for several reasons. Most se¬ries are retrospective and hospital-basedand may select for sicker patients withatrial fibrillation.510 Furthermore, be¬cause these studies are generally de¬

scriptive and lackingin antecedent clini¬cal risk factor information, they are

unable to control for confounding riskfactors.510 Two population-based stud¬ies have examined atrial fibrillation;how¬ever, their analyses were limited bysmall case numbers and a failure to con¬

trol for confounding variables.11,12 Thecurrent studypopulationincluded a largenumber of cases ofatrial fibrillation docu¬mented both in the hospital and in an

ambulatory setting during nearly 40years offollow-up. In addition, potentialrisk factors were assessed prospectively,thereby reducing ascertainment bias.Prior Framingham Heart Study re¬

ports have exploredrisk factors for atrialfibrillation, noting that valvular heartdisease, cardiac failure, hypertensiveheart disease, and coronary heart dis¬ease were predictors of atrial fibrilla¬tion.1,4 However, these investigationswere based on far fewer cases andyounger subjects (with a mean age 10years younger),1,4 used case-controlmethodology(matchingon age and sex),4and did not utilize multivariable analy¬ses.1,4The main discrepancybetween the

current and prior Framingham HeartStudy analyseswas in the estimated im-

Table 3.—Population-Attributable Risk (Prevalenceof Variable In %) of the Risk Factors AssociatedWith Atrial Fibrillation*

Variable Men Women

Cigarette smoking 2 (34) 8 (24)Diabetes 4 (10) 4 (8)Electrocardiographic left

ventricular hypertrophy 2 (4) 1 (4)Hypertension 14 (31) 14(40)Myocardial infarction 5(13) 1(4)Congestive heart failure 10(3) 12(3)Valve disease 5(7) 18(9)

*Two-year pooled logistic regression. Follow-upperson-examination and number with atrial fibrillationare the same as in Table 2.

pact ofcongestive heart failure on atrialfibrillation risk in men. In a prioranaly¬sis, cardiac failure was associated withan age-adjustedrisk ratio of 17.5 in men

and 5.7 in women.4 In the current analy¬sis, after full multivariable adjustment,the odds ratio of developing atrial fi¬brillation with congestive heart failurewas 4.5 in men and 5.9 in women. Con¬gestive heart failure frequently occurs

in the settingof valvular heart disease,myocardialinfarction,hypertension,andleft ventricular hypertrophy, so the im¬pact of heart failure in the current mul¬tivariable analyses may have been di¬minished by its correlation with othervariables in the models. The attenuatedrisk of congestive heart failure in thecurrent study, as opposed to the earlierstudies,1·4 also may reflect the aging ofthe cohort. Because the current analy¬ses used full multivariable models with3931 to 4644 additional cases of atrialfibrillation, the current analyses pro¬vided more stable and reliable estimatesof the impact of a variety of risk factorson the incidence of atrial fibrillation.

Limitations of the StudyThe ascertainment of atrial fibrilla¬

tion is problematic, since some cases ofatrial fibrillation may remain undetec¬ted if they were brief or not severe. Forthe principal analyses, subjects withatrial fibrillation on either FraminghamHeart Study or outside electrocardio¬grams were included. Using hospital-ization electrocardiograms for the diag¬nosis of atrial fibrillation selected forsicker subjects with more symptomaticatrial fibrillation. Conversely, analysesrestricted to subjects with atrial fibril¬lation o n routine clinic examination (non-acute atrial fibrillation) selected forchronic and more benign atrial fibrilla¬tion cases (subjects with fatal or com¬

plicated atrial fibrillation would be un¬

available for follow-up). Nevertheless,the results from the multivariable mod¬els of all atrial fibrillation cases and atrialfibrillation cases documented on Fram¬ingham Heart Study clinic electrocar¬

diograms were

comparable, except formyocardial infarction.

Table 4.—Multivariable Risk of Nonacute AtrialFibrillation Detected on Routine Framingham HeartStudy Clinic Electrocardiogram

Odds Ratio*(95% Confidence Interval)

I-1Variable Ment Women$

Age(/10y) 1.9§ (1.5-2.5) 1.7§ (1.4-2.1)Cigarette smoking 1.3(0.8-2.0) 1.1(0.7-1.8)Diabetes 1.7|| (1.1-2.8) 1.4(0.9-2.4)Electrocardiographic

left ventricular

hypertrophy 1.9(1.0-3.8) 1.2(0.6-2.3)Hypertension 1.2 (0.8-1.8) 1.3 (0.9-1.9)Myocardial infarction 0.6(0.3-1.1) 0.7(0.3-1.3)Congestive heart

failure 4.6§ (2.6-8.2) 5.4§ (3.3-8.9)Valve disease 1.7|| (1.0-2.8) 3.9§ (2.6-5.9)

*Odds ratios are based on multiple logistic regressionanalyses of 2-year pooled observations.

tNonacute atrial fibrillation was diagnosed in 109men In 16 694 follow-up person-examinations.

tNonacute atrial fibrillation was diagnosed in 119women in 23929 person-examinations.

§P==.0001.IIPs.05.

Table 5.—Multivariable Risk of Atrial Fibrillation inSubjects Without Clinical Valve Disease*

Odds Ratio(95% Confidence Interval)

Variable Menf WomentAge (/10 y) 2.2§ (1.8-2.7) 2.6§ (2.1-3.2)Cigarette smoking 1.0 (0.7-1.4) 1.5 (1.0-2.2)Diabetes 1.1(0.8-1.7) 1.5(1.0-2.3)Electrocardiographic

left ventricularhypertrophy 1.6(0.9-3.0) 1.2(0.6-2.4)

Hypertension 1.6|| (1.2-2.2) 1.7|| (1.2-2.4)Myocardial infarction 1.4(1.0-2.1) 2.11| (1.3-3.4)Congestive heart

failure 5.0§ (3.2-7.8) 4.3§ (2.7-7.1)

*Two-year pooled logistic regression.t Atrial fibrillation was diagnosed in 180 men in 15 421

follow-up person-examinations.tAtrial fibrillation was diagnosed in 154 women in21 627 person-examinations.§P==.0001.IIP«.01.

Because of the potential for misclas-sification, chronic and paroxysmalatrialfibrillation were combined as end points,despite plausible differences in etiology.Another potential area of misclassifica-tion was in the clinical ascertainment ofvalvular heart disease. This study be¬gan over 40 years ago, prior to the ad¬vent of echocardiography. Echocardio-

graphic predictors of atrial fibrillationhave been reportedin another Framing¬ham Heart Study investigation.23A fur¬ther potential limitation of the study isits generalizability. The study samplewas largely white, and the age was re¬

stricted to subjects 55 to 94 years ofage.The findings may not be generalizableto other racial groups or younger or oldersubjects.

Finally, the lack of statistical signifi¬cance of several suspected risk factorsfor atrial fibrillation should not be con¬

strued to mean that the factors are not

possible risk factors in the individualpatient. For example, myocardial in-




farction was not shown to be a signifi¬cant risk factor for atrial fibrillation inwomen in the whole sample (Table 2).Subjectswith myocardialinfarction maydevelop congestive heart failure that inturn precipitates atrial fibrillation; be¬cause congestive heart failure'and myo¬cardial infarction are colinear, adjustingfor congestive heart failure may haveattenuated the known association be¬tween myocardial infarction and atrialfibrillation. In another instance, bingealcohol consumption may precipitateatrial fibrillation in susceptibleindividu¬als.24 The lack of an association betweenalcohol consumption and atrial fibrilla¬tion may reflect the relatively few in¬dividuals affected by "holiday heart,"or

the extreme alcohol use required to pre¬cipitate atrial fibrillation.

Clinical Implications

The incidence of atrial fibrillation in¬creases dramatically with age and is a

significant source ofdisabilityand deathin the elderly. The treatment of atrialfibrillation with anticoagulant or anti-arrhythmic medications may diminishthe symptoms and complications asso¬

ciated with it.2528 However, anticoagu¬lant and antiarrhythmic agents are fre¬quently contraindicated,27·28particularlyin the elderly, and their use is associ¬ated with significant morbidity.29 31 Thesafest and most effective means of pre¬venting the complications of atrial fi¬brillation is to prevent its occurrence.

Multivariableanalysesdemonstrate thatage, diabetes, hypertension, congestiveheart failure, and valve disease are in-

dependent risk factors for atrial fibril¬lation in both sexes. If these precursorsof atrial fibrillation contribute to the in¬cidence of atrial fibrillation in a causalmanner, the analyses of population-at¬tributable risk suggest that modifica¬tion of cardiovascular disease risk fac¬tors and the prevention of structuralcardiac disease may have the added ben¬efit of diminishing the incidence ofatrialfibrillation.

This study wa s supported in part through Na¬tional Institutes of Health/National Heart, Lung,and Blood Institute contract NO1-HC-38038and byNational Institute of Neurological Disorders andStroke grant 2-R01-NS-17950-11.

The authors greatly appreciate the contribu¬tions of Rodney M. Falk, MD, and David M. Pollakfor reviewingthe manuscript; William Kannel, MD,for insight into the analyses; and Mary Healy, LoisAbel, Karen Mutalik, and Deborah Dumphy for as¬

sistance in the collection of atrial fibrillation cases.

References1. Kannel WB,Abbott RD,Savage DD, McNamaraPM. Coronary heart disease and atrial fibrillation:The

Framingham Study. Am Heart J.

1983;106:389-396.2. Wolf PA, Abbott RD, Kannel WB. Atrial fibril-lation as an independent risk factor for stroke: TheFramingham Study. Stroke. 1991;22:983-988.3. Gajewski J, Singer RB. Mortality in an insuredpopulationwith atrial fibrillation. JAMA. 1981;245:1540-1544.4. Kannel WB,Abbott RD, Savage DD, McNamaraPM. Epidemiologic features of chronic atrial fibril-lation: the FraminghamStudy. N EnglJ Med. 1982;306:1018-1022.5. AbergH. Atrial fibrillation: a review of463 cases

from Philadelphia General Hospital from 1955 to1965. Acta Med Scand. 1968;184:425-431.6. Davidson E, Weinberger I, Rotenberg Z, FuchsJ, Agmon J. Atrial fibrillation: cause and time ofonset. Arch Intern Med. 1989;149:457-459.7. GodtfredsenJ. Atrial fibrillation:

course

and prog-nosis\p=m-\afollow-up study of 1212 cases. In: Kulber-tus HE, Olsson SB, Schlepper M, eds. Atrial Fi-brillation. M\l=o"\lndal, Sweden: AB Hassle; 1982:134\x=req-\145.8. McEachern D, Baker BM. Auricular fibrillation:its etiology, age incidence and production by digi-talis therapy. Am J Med Sci. 1932;183:35-48.9. Stroud WD, Laplace LB, ReisingerJA. The eti-ology, prognosis and treatment of auricular fibril-lation. Am J Med Sci. 1932;183:48-60.10. Sawyer CG, Bolin LB, Stevens EL, Daniel LB,O'Neil NC, Hayes DM. Atrial fibrillation: its eti-ology, treatment and association with embolization.South Med J. 1958;51:84-93.11. \l=O"\nundarson PT, Thorgeirsson G, JonmundssonE, Sigfusson N, Hardarson T. Chronic atrial fibril-lation\p=m-\epidemiologicfeatures and 14-year follow\x=req-\up: a case control study.Eur Heart J. 1987;8:521-527.12. 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.

13. Dawber TR, Meadors GF, Moore FE. Epide-miologicalapproachesto heart disease: the Framing-ham

Study. Am J Public Health.

1951;41:279-286.14. Shurtleff D. Some characteristics related to theincidence of cardiovascular disease and death:Framingham Study, 18-year follow-up. In: KannelWB, Gordon T, eds. The Framingham Study: AnEpidemiological Investigation of CardiovascularDisease. Washington, DC: Dept of Health, Educa-tion, and Welfare; 1974: section 30. DHEW publi-cation NIH 74-599.15. Kannel WB, Gordon T, Offutt D. Left ventricu-lar hypertrophyby electrocardiogram: prevalence,incidence, and mortalityin the Framingham Study.Ann Intern Med. 1969;71:89-105.16. Cupples LA, D'Agostino RB,Anderson K, Kan-nel WB. Comparison of baseline and repeated mea-

sure covariate techniquesin the Framingham HeartStudy. Stat Med. 1988;7:205-218.17. D'Agostino RB, Lee ML, BelangerAJ, CupplesLA, Anderson K, Kannel WB. Relation of pooledlogistic regression to time dependent Cox regres-sion analysis: the Framingham Study. Stat Med.1990;9:1501-1515.18. Walker SD. The estimation and interpretationof attributable risk in health research. Biometrics.1976;32:829-849.19. Sherman DG, Goldman L, Whiting RB, Jur-gensen K, Kaste M, Easton D. Thromboembolismin patients with atrial fibrillation. Arch Neurol.1984;41:708-710.20. Cameron A, Schwartz MJ, Kronmal RA, Ko-sinski AS. Prevalence and significance of atrial fi-brillation in coronary artery disease (CASS Reg-istry). Am J Cardiol. 1988;61:714-717.21. Wood P. An appreciation of mitral stenosis,part I: clinical features. BMJ. 1954;1:1051-1063.22. Robinson K, Frenneaux MP, Stockins B,Karatasakis G, Poloniecki JD, McKenna WJ. Atrialfibrillation in hypertrophic cardiomyopathy: a lon-gitudinal study. J Am Coll Cardiol. 1990;15:1279\x=req-\1285.

23. Vaziri SM, Larson MG, Benjamin EJ, Levy D.Echocardiographicpredictorsofnonrheumatic atrialfibrillation:

Framingham Heart

Study. Circulation.

1994;89:724-730.24. EttingerPO, Wu CF, De la Cruz C, Weisse AB,Ahmed SS, Rega TJ. Arrhythmiasand the 'holidayheart': alcohol associated cardiac rhythmdisorders.Am Heart J. 1978;95:555-562.25. Petersen P, Godtfredsen J, Boysen G, Ander-sen ED, Andersen B. Placebo-controlled, random-ized trial of warfarin and aspirin for prevention ofthromboembolic complications in chronic atrial fi-brillation: the Copenhagen AFASAK Study. Lan-cet. 1989;1:175-179.26. The Boston Area AnticoagulationTrial for AtrialFibrillation Investigators. The effect of low-dosewarfarin on the risk of stroke in patients with non-

rheumatic atrial fibrillation. N Engl J Med. 1990;323:1505-1511.27. Stroke Prevention in Atrial Fibrillation Inves-

tigators. Stroke Prevention in Atrial FibrillationStudy: final results. Circulation. 1991;84:527-539.28. Ezekowitz MD, Bridgers SL, James KE, et al,for the Veterans Affairs Stroke Prevention in Non-rheumatic Atrial Fibrillation Investigators. War-farin in the prevention of stroke associated withnonrheumatic atrial fibrillation.N EnglJ Med.1992;327:1406-1412.29. Coplen SE, Antman EM, Berlin JA, Hewitt P,Chalmers TC. Efficacy and safety of quinidinetherapy for maintenance of sinus rhythmafter car-

dioversion: a meta-analysis of randomized controltrials. Circulation. 1990;82:1106-1116.30. Albers GW, Sherman DG, Gress DR, PaulsethJE, Petersen P. Stroke prevention in nonvalvularatrial fibrillation: a review of prospective random-ized trials. Ann Neurol. 1991;30:511-518.31. Fihn SD, McDonell M, Martin D, et al, for theWarfarin Optimized Outpatient Follow-up StudyGroup. Risk factors for complications of chronicaticoagulation:a multicenter study.Ann Intern Ned.1993;118:511-520.

independant risk factors for atrial fibrillation in a population-bases cohort

Documents