topeesp.gif (5672 bytes)

[ Scientific Activities - Actividades Científicas ]

Epidemiology of Atrial Fibrillation: Risk Factors and Hazards

William B. Kannel, MD

Department of Preventive Medicine and Epidemiology
Evans Department of Clinical Researeh
Boston University School of Medicine
Boston , Massachusetts
and the Framingham Heart Study. Framingham

Atrial fibrillation is the most common serious chronic cardiac rhythm disturbance its prevalence doubling with each decade of age from 0.55 at age 50-59 to 9% at age 80-89 years of age. Its prevalence in the population is also increasing in men. The 3-fold increase in men is not attributable to population aging or changes in the prevalence of valve disease or coronary disease. The new onset of AF also doubles per age decade and men are 1. 5 times as likely as women to develop it.

In the Framingham Study only hypertension and diabetes were significant independent predictors of AF increasing the risk 1.5 fold. Because of its high population prevalence, hypertension is responsable for more AF (14%) than any other risk factor. Overt cardiac conditions impose a substantially greater risk than the cardiovascular risk factors with heart failure imposing a risk factor adjusted hazard of 4.5 in men and 5.9 in women. Valvular heart disease also imposed a major hazard that was greater for women (HR= 3,4) than men (HR= 1. 8). Myocardial infarction increased the risk factor adjusted AF hazard by only 40% and only in men.

Independent echocardiographic predictors of AF include left atrial enlargement, left ventricular fractíonal shortening and left ventricular wall thickness. EKG evidence of left ventricular hypertrophy was also a powerful age adjusted predictor, but the risk ratio decreased from 3-4 fold to 1.4 after adjustment for associated conditions.

The chief hazard of AF is a cardiogenic thrornboembolism, increasíng the risk of a stroke 4-5 fold. About 15 % of all ischemic strokes are attributable to AF and at ages 80-89 years about 24% of strokes are AF induced. Survival is also seriously reduced, with mortality rates doubled across a wide age range.

Nonvalvular AF should be recognized as a major public health problem and a major risk factor for presentable strokes. Also, decreasing the risk of cardíovascular conditions that promote AF by controlling predisposing cardiovascular risk factors will curb the íncidence of AF.

Atrial fibrillation is the most common of the serious chroníc cardiac rhythm disturbances (1). lt is responsable for a substancial amount of morbidity, disability and rnortality in the general population. With the aging of the populatíon and improved cardiovascular surveillance, more AF is now being encountered. There is also increasing awareness of the hazards of the condition.

Prevalence, Incidence and Trends:

The prevalence of AF in the adult population doubles with each advancing decade of age from 0.5% at age 50-59 years to 9% at age 80-89 years (Fig. l). As aging proceeds there is gradual loss of nodal fibers and increase of fibrous and adipose tissue in the síno-atrial node (2). Decreased ventricular compliance occurs due to myocardial fibrosis resulting in atrial dilatation that has been shown to predispose to AF (3). Extensive senile amyloid infiltration of the sinoatrial node may occur (4). However, aging also ínvolves longer exposure to predisposing conditions, and even in advanced age some are clearly more vulnerable to the development of AF than others.

There are 2.2 million US citizens who have AF. lt is a rnale dominant affliction with a sex ratio of 1.7. The prevalence is significantly higher in persons with heart murmurs, myocardial infarction and heart failure.

Analysis of secular trends in the prevalence of AF in the in the Framingham Study, where routine ECG's are obtained on each participant biennially, indicates that this rhythm disturbance is becoming more prevalent than forrnerly, particularly in men (Fig. 2). The prevalence of AF in elderly men in the Framingham Study increased 3-fold over 2 decades. The increase in women was less pronounced over this period of time. This increase in the prevalence of AF in the cohort could not be attributed to changes in the population prevalence of valvular heart disease, myocardial infarction or aging of the population. Consístent with the Framingham Study secular trend data, the number of hospitalizations reported nationally for the US between 1982 and 1995 increased sharply (Fig. 3).


fig1a.gif (23974 bytes)

Figure 1: Wolf PA et Al. Stroke. 1991; 22:983-988


fig2a.gif (25705 bytes)

Figure 2: Wolf et al. Am Heart J. 1996; 131:790-795


fig3a.gif (47289 bytes)

Figure 3


The incidence of new onset of atrial fibrillation also doubles with advancing decades of age, independent of any changes in the prevalence of the predisposing conditions. The annual incidence in men increases frorn about 5 per 1000 at age 50-54 to 45 per 1000 at ages 85-94 years
(Fig. 4). in women the incidence increases from 2.5 to 30 per 1 000 at these ages. The size of the gap in AF incídence between the sexes closes with advancing age.

Risk Factors:

Of the major cardiovascular risk factors, only hypertension, and diabetes were significantly and independently related to the rate of development of AF in the Framingham Study after adjustment for age and other predísposing conditions (Table l).


fig4a.gif (40255 bytes)

Figure 4: Benjamin EJ et al. JAMA. 1994; 271: 840-844


table1a.gif (32342 bytes)

Table 1: Benjamin EJ et al. JAMA. 1994; 271:840-844


Hypertension and diabetes each independently ícrease the risk about 1.5 fold, Adjusting only for age, diabetes, ECG-LVH and hypertension were signifícant predictors of AF. Cigarette smoking was related to AF in women only and BMI and alcohol were unrelated in both sexes.

Women who srnoked were 40% more likely to develop AF; those with hypertension had a 70% greater risk,; diabetics were twice as likely to develop it; and those with ECG-LVH had almost a 4-fold increased risk. In men diabetes increased AF risk 70%, hypertension 80% and ECG-LVH 3-fold.

Clinical Characteristíes:

Persons who develop AF are usually elderly, male and more likely than persons of the same age to have coronary disease, valvular heart disease, heart faílure, echocardiographic abnormalities, left ventricular hypertrophy, diabetes or hypertension (5), About 20% of men and a third of women have valvular heart discase, about a quarter of both sexes have heart failure and 28% of men and 14% of women haye myocardial infarctions.

Cardiac Harbingers: Cardiac conditions associated with AF include acute and chronic coronary disease, hypertrophic and dilated cardiomyopathy, congenital heart discase, pericarditis, hypertensíve cardiovascular disease, and valvular heart disease. In the Framingham Study the most common cardiac precursors of AF were heart failure, myocardial infarction and valvular heart disease. These cardiac conditions accounted for 20 % of the AF in men and 3 1% of its occurrence in women (6). Taking age and other predisposing conditions into account, heart failure appeared to impose the greatest hazard, increasíng the risk 4.5 fold in men and 5.9 fold in women (Table 2), Valve disease increased the risk 1.8 fold in men and 3.4 fold in wornen, Myocardial infarction in the Framingham Study independently increased AF risk only 40% in men. In women, the age adjusted 2.4 fold increased risk observed for myocardial infarction, was reduced to only a 20% excess risk on adjustment for associated conditions.

Autopsy studies indicate that AF appearing in the course of acute myocardial infarction ís frequently associated with a coronary occlusion proximal to the sinos node artery (7,8). Whereas acute myocardial infaretión is an undisputed cause of AF, the role of chronic coronary disease is controversias (9). The Coronary Artery Surgery Study found only a 0.6% prevalence of sustained Af in over 18,000 patíents undergoing coronary angiography. Those who had associated AF were older, and had mítral regurgitation or heart failure (10). The Framingham Study found that similar risk factors were domínant in coronary patients who developed AF (1 l). Autopsy studies are also ínconsistent on the role of heart failure as a cause of AF (12,13). Autopsy evidence is quite consístent on the role of mitral valve disease as a cause of AF. lt has been estimated to occur in about 40 % of patients with mitral stenosis and in 75% of those with mitral regurgitation (14).


table2a.gif (29752 bytes)

Table 2:  EJ et al. JAMA. 1994; 271: 840-844


Echocardiographic Precursors:

The Framingham Study has examined the echocardiographic predictors of AF (1 5). The echocardiographic characteristics of persons who developed AF included larger left atrial, and left ventricular dimensions and ventricular wall thickncss and more mitral annular calcification (Table 3). Each 5mm increment in left atrial dímension increased AF rísk 39%. A 5% decrement in left ventricular fractional shortening increased the risk 34% and a 4 mm increment in left ventricular wall thickness íncreased AF risk 28%. Mitral annular calcification doubled the risk, Those with two or more of the foregoing echocardiographic abnormalities had a 4-fold greater risk than those free of them all. These echocardiographic features can be used to provide AF predictíve information beyond that deríved from the clinical conditions known to predíspose.


table3a.gif (26729 bytes)

Table 3


AF is associated with an adverse quality of life when symptomatic. In the symptomatic person, the awareness of palpitations and tachycardía, the patient's actívity status and the nature of the underlying disorder all impact the quality of life in the patient with AF. More ominous is the cardiovascular hazard of worsening or precipitatíon of heart failure in those with underlying cardiac disease. The most feared hazard of AF is the occurrence of a stroke. AF is the most common cause of cardiogenic thromboernbolism. lt carries a substantial risk of ischemic stroke even in the absence of valvular heart discase. The lifetirne risk of a stroke in patients with untreated AF is about 35%. Epidemiologic and clinical ínvestigation indícates that AF is a major independent risk factor for stroke, inercasing the likelihood 3-5 fold even after adjusting for other stroke risk factors (Table 4). The risk ratios are substancial at all ages, and in contrast to the other risk factors, the influence of AF on strokc incidente does not diminísh with advancing age. The stroke rate in persons with AF is about 5% per year. The proportion of strokes attributable to AF in the Framingham Study increased from 1.5% at age 50-59 years to 23.5% at age 80-89 years. The percent of stroke events occurring  in association wíth AF increased with each decade of age from 6.5% at age 50-59 years to 30.7% at age 80-89 years (Table 5). Cardiac conditions such as coronary disease and heart failure, that also predispose to strokes, often coexist with AF. However in the Framingham Study it was shown that the presence of AF in conjunction with these cardiac condition further escalated the risk of a stroke about 2-fold 
(Fig. 5), Investigation of the time course of embolic events in AF indicate that the risk of strokes in persons with AF ís highest in the early months after the diagnosis of the disturbed rhythm, (1 6,17). The Framingham Study found that stroke recurrences were likely to occur sooner in persons with than wíthout AF (16). Also, ischemic strokes associated with AF were nearly twíce as likely to be fatal as those unassocíated wíth AF, Stroke recurrences were also more common and the functional deficits more severe (18). There appears to be an increased rísk of thromboembolism in persons with AF when there is left atrial enlargement or decreased left ventricular systolic function (19).


table4a.gif (26634 bytes)

Table 4: Wolf PA et al. Stroke. 1991; 22:983-988


table5a.gif (24400 bytes)

Table 5: Wolf PA et al. Stroke. 1996; 22:983-988


fig5a.gif (28968 bytes)

Figure 5: Wolf PA et Al. Stroke. 1991; 22:983-988


lt has bcen unclear as to whether AF contributes to excess mortality independent of the influence of the cardiovascular conditions that predispose to its occurrence. Some have reported excess mortality when it accompanies myocardial infarction whereas others have not (20,2 l). Likewise for subjects  with heart failure or stroke the reported findings have been inconsistent (22,23 ). The Framingham Study recently demonstrated that AF is independently associated with a 50-90% íncrease in the risk of death for men and women respectively. This excess risk persists after adjustment for coexisting cardiovascular conditions (24). The distribution of the causes of death for subjects with AF was similar to that of matched subjects wíthout AF, but for each of the causes of death, a higher percentage of the AF subjects died in the first year after the AF was diagnosed

Preventive Implications:

AF is a major and growing problem in an aging populatíon. Decreasing the risk of cardiovascular events that predispose to AF by controlling hypertension and diabetes should reduce the prevalence of AF that has been increasing alarmingly in the population. The treatment of AF with anticoagulants and antiarrhythmic medications can diminish the symptoms and thromboembolic events with which it is assocíated. Six randomized trials testing the efficacy of oral anticoagulation consistently demonstrated an impressive 67% reduction in rísk of a stroke in patients with AF (25). By contrast the benefit of aspirin is small, with only a 21% reduction in risk that was not statistically significant in most trials (26). Other untested antiplatelet agents such as clopídogrel rnay be more efficacious. However, all these remedies are frequently contraindicated in the elderly and their use is often associated with significant morbidity. lt would be useful to risk stratify AF patients to select those in whom anticoagulation is mandatory and those that can be treated less aggressively. Pooled data from trials indicates that independent risk factors for stroke in patients with AF are a prior stroke, older age, hypertension, diabetes and poor left ventricular function (26). In patients without these risk factors aspirin or other antiplatelet agents may be adequate .



1. Kannel WB, Abbott RD, Savage DD, McNamara PM, Coronary heart disease and atrial fibrillation: The Framingham Study, AM Heart J 1983- 106:389-396.
2. Falk RH. Ftiology and complications of atrial fibrillatíon: Insights from pathology studies. Am J Cardiol 1998 82: 10N- 176N.
3. Manyari DE, Patterson C, Johnson D, et al. Atrial and ventricular an-ythrnias in asymptomatic elderly subjects. Correlation with left atrial size and left ventricular mass. Am Heart J 1990; 119:1069-1076.
4. Lie JT, Hammond PI. Pathology of the senescent heart: anatomic observations on 237 autopsy studies of patients 90~105 years old. Mayo Clin Proc 1998; 63:552-564.
5. Kannel WB, Wolf PA, Benjamin EJ, Levy D. Prevalenco, incidence, prognosis, and predisposing conditions for atríal fibrillation: Population-based estimates. Am J Cardiol 1998- 82:2N-9N.
6. Feinberg WM, Blackshear JL, Laupasis A, et al. Prevalence, age distribution and gender in patients with atrial fibrillation: analysis and implications. Arch Intern Med 1995; 155:469-473.
7. Soderstrom N. Myocardial infarction and mural thrombus in the atría of the heart. Acta Med Scan 1948; (suppl): 217.
8. Cushing EH, Feil H, Stanton EJ, Wartman WB. Infarction of cardiac auricles (atria): clinical, pathological and experimental studies. Br Heart J 1942; 4:17-34.
9. Haddad AH, Prehkov VK, Dean DC. Chronic atrial fibrillation and coronary artery disease. J Electrocardiol 1978: 11:67-69.
10. Cameron A, Schwartz MJ, Fronmal RA, Kosinski AS. Prevalence and significance of atrial fibrillation in coronary artery disease (CASS registry). Am J Cardiol 1988; 61:714- 717.
11. Kannel WB, Abbott RD, Savage DD. Coronary heart discase and atrial fibrillation: The Framingham Study. Am Heart J 1983; 106:389-396.
12. Davies MJ, Pomerance A. Pathology of atrial fibrillation in man. Br Heart J 1972, 34:520-525.
13. Hinton RC, Kistler JP, Fallon JT, Friedlich AL, Fisher CM. Incidence of etiology of atrial fibrillation on incidence of systemic embolism, Am J Cardiol 1977; 40:509-513.
14. Probst P, Goldschlager N, Seltzer A. Left atrial size and atrial fibrillation in mitral stenosis: factors influencing their relationship. Circulation 1973; 48:1282-1287.
15. Vazirí SM, Larson MG, Benjamin EJ, Levy D. Echocardiographic predictors of nonrheumatic atrial fibríllation. The Framinghain Study, Circulation 1994; 89:724-730.
16. Wolf PA, Kannel Vffl, McGee DL, et al. Duration of atrial fibrillation and imminence of stroke: The Framingham Study.
17. Lin H-J, Wolf PA, Benjamin EJ, et al, Newly diagnosed atrial fibrillation and acute: stroke. The Framingham Study. Stroke 1995; 26:1527-1530.
18. Lin H-J Wolf PA, Kelly-Hayes M, et al. Stroke severity in atrial fibrillation: The Framingham Study. Stroke 1996; 27:1760-1764.
19. Smith VE, White Vffl, Karímeddiní MK. Echocardiographic assessment of left ventrícular diastolic performance in hypertensive subjects: correlation with changes in left ventricular mass, Hypertension 1987; 9 (suppl 11) 11 81- 11 84.
20. Crenshaw BS, Ward SR, Grangcr CB, et al. For the GUSTO-1 Trial Investigators. Atrial fibrillation in the setting of acute myocardial infarction: the GUSTO-1 experience. J Am Coll Cardiol 1997: 30:406-413.
21. Behar s, Tanne D, Zion M, et al. For the SPRINT Study Group. Incidence and prognostic significance of chronic atrial fibrillation among 5839 consecutive patients wíth acute myocardial infarction. Am J Cardiol 1992; 70:816-818.
22. Carson PE, Johnson GR, Dunkman WB, et al. For the V-HEFT VA Cooperative Studies Group. The influence of atrial fibrillation on prognosis in mild to moderate heart failure: the V-HEFT Studies. Circulation 1993; 87 (Suppl VI) VI-102-VI-1 10.
23. Middlekauff HR, Stevenson WG, Stevenson LW. Prognostíc sígnificance of atrial fibrillation in advanced heart failure: a study of 390 patients. Circulation 1991: 84:40-48.
24. Benjamin EJ, Wolf PA, D'Agostíno RB, et al. Imact of atrial fibrillation on the risk of death, The Framingham Heart Study. Circulation 1998; 98:946-952.
25. Laupacis A, Albers G, Dalen J, et al. Antithrombotic therapy in atrial fibrillation Chest 1998; 14:597S-589S.
26. Atrial Fibrillation Investigators. The efficacy of aspirin in patients with atrial fibrillation. Analysis of pooled data from three randomized trials. Arch Intern Med 1997: 157:1237-1240.

Study research is supported by NIH/NHLBI Contract NOI-HC-38038 and the
Visiting Scientist Program which is supported by ASTRA USA and Servier Canada, Inc.