CLINICAL RESEARCH: CARDIAC RHYTHM DISORDER
The Relationship Between Stature and the Prevalence of Atrial Fibrillation in Patients With Left Ventricular Dysfunction
Ibrahim R. Hanna, MD*,*,
Brian Heeke, BS*,
Heather Bush, MS ,
Lynne Brosius, MS,
Diane King-Hageman, BS,
John F. Beshai, MD* and
Jonathan J. Langberg, MD*
* Division of Cardiology, Section of Electrophysiology, Emory University, Atlanta, Georgia
REGISTRAT Inc., Lexington, Kentucky.
Manuscript received October 13, 2005;
revised manuscript received November 14, 2005,
accepted November 16, 2005.
* Reprint requests and correspondence: Dr. Ibrahim R. Hanna, Emory University Hospital, 1364 Clifton Road, Suite F414, Atlanta, Georgia 30322. (Email: ihanna2{at}aol.com).
 |
Abstract
|
|---|
OBJECTIVES: This study sought to determine the influence of stature on atrial fibrillation (AF) in high-risk patients with reduced left ventricular (LV) systolic function.
BACKGROUND: Left atrial (LA) enlargement is a potent risk factor for AF. Because LA size is strongly associated with stature, we hypothesized that height and body surface area (BSA) are risk factors for AF, independent of other known associations.
METHODS: Data were obtained from ADVANCENT, a multicenter registry of patients with impaired LV function. Height and BSA were divided into quartiles by gender. Statistical analysis was done using the Cochran Mantel-Haenszel statistic, and multivariable logistic regressions were used to adjust for the effects of known confounders on the association between stature and AF.
RESULTS: A total of 25,268 patients were enrolled. The mean age was 66 years, and the cohort consisted mostly of white men (72%) and patients with ischemic cardiomyopathy (72%). The mean left ventricular ejection fraction was 31%. A history of AF was present in 7,027 patients (27.8%). The AF prevalence increased significantly between the lowest and highest quartiles for height (32% relative increase, p < 0.0001). In the multivariable analysis, the effect of height on AF risk persisted after adjusting for age, gender, race, left ventricular ejection fraction, heart failure class and etiology, hypertension, diabetes, and medication use (odds ratio 1.026/cm, 95% confidence interval [CI] 1.022 to 1.030). In the multivariable analysis, BSA was also an independent predictor of AF risk (odds ratio 4.221/m2, 95% CI 3.358 to 5.306).
CONCLUSIONS: In patients with LV dysfunction, increasing stature portends a higher risk of AF independent of other traditional risk factors for the arrhythmia. This association seems to account for the higher prevalence of AF in men and may be useful for identification of a high-risk population.
|
Abbreviations and Acronyms
| | ACE = angiotensin-converting enzyme | | AF = atrial fibrillation | | ARB = angiotensin receptor blockers | | BMI = body mass index | | BSA = body surface area | | LA = left atrial | | LV = left ventricular | | LVEF = left ventricular ejection fraction | | NYHA = New York Heart Association |
|
Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, with an incidence that increases two-fold with every decade after age 55 years (1). With a prevalence of more than 5% in patients over the age of 65 years, and an aging U.S. population (2), the number of hospital admissions for AF more than doubled from 1984 to 1994 (3). Despite sizeable health care costs, relatively few population-based studies have systematically characterized predictors of AF, a necessary step for the identification of a high-risk group that would constitute a target for the evaluation and implementation of primary preventive strategies.
Data derived from long-term follow-up of the Framingham Heart Study participants identified older age, male gender, hypertension, diabetes, heart failure, valvular heart disease, and prior myocardial infarction as independent clinical predictors of the arrhythmia (4). Echocardiographic parameters from this same cohort, including left atrial (LA) diameter and left ventricular (LV) systolic function, were also strongly associated with the development of AF (4). More recently, the Cardiovascular Health Study reported on the predictors of AF in a cohort over the age of 65 years, confirming the role of traditional risk factors and showing that LA size was among the strongest independent predictors of the arrhythmia. This association was linear, with a 60% increase in relative risk of AF for every 1 cm increase in LA dimension above 3 cm (5).
The pathophysiologic basis for the association between AF and LA size can be derived from the multiple wavelet theory of Moe and Allessie (6). The number of simultaneous activation wave fronts necessary for persistence of the arrhythmia requires a critical LA surface area (6). Thus, for the same degree of atrial myocardial pathology, it is likely that patients with larger left atria would be more prone to develop AF. Given the strong association between body size and LA size (7,8), we hypothesized that in a high-risk population with reduced LV function, larger body size is associated with a higher prevalence of atrial fibrillation, independent of traditional risk factors for the arrhythmia.
|
Methods
|
|---|
Study population.
Patients enrolled in the ADVANCENT registry through September 30, 2004, were included in the study; ADVANCENT is a prospective, longitudinal, multicenter, observational registry designed to collect and report data on the histories, diagnostics, therapies, and interventions for patients with LV dysfunction (ejection fraction  40%). At the time of data analysis, 25,268 patients from 106 U.S. centers had been enrolled.
Baseline data collection.
At enrollment, all ADVANCENT participants were interviewed by medical personnel (physicians, nurse practitioners, and physician assistants), with additional data derived from review of medical records. Registry data contained information about age, gender, race, height, weight, left ventricular ejection fraction (LVEF), etiology of cardiomyopathy, New York Heart Association (NYHA) heart failure class, and coronary artery disease risk factors (tobacco smoking, hypertension, hyperlipidemia). Information was also gathered regarding co-morbidities (diabetes, cancer, renal failure, anemia, cerebrovascular events, reactive airway disease, myocardial infarction, and syncope), medication use, and a variety of clinical events and interventions (angioplasty, coronary bypass surgery, pacemaker/defibrillator implantation).
AF.
The presence of AF was identified at the time of enrollment through patient interviews, electrocardiograms, and review of medical records, and was classified as permanent, paroxysmal, or unknown pattern.
Stature.
Measurements of body size consisted of height (in centimeters) and body surface area (BSA). The BSA was derived using the Mosteller formula (BSA (m2) = [height (cm) x weight (kg)/3,600]) (9). To eliminate any confounding effect of gender, the boundaries of the quartiles for these parameters were defined separately for men and women.
For men, the quartiles for height (cm) were <172.7, 172.7 to 177.8, 177.9 to 182.9, and >182.9, and those for BSA (m2) were <1.93, 1.93 to 2.08, 2.09 to 2.24, and >2.24. For women, the quartiles for height (cm) were <157.5, 157.5 to 162.6, 162.7 to 167.6, and >167.6, and those for BSA (m2) were <1.64, 1.64 to 1.79, 1.80 to 1.97, and >1.97.
Statistical analysis.
A univariate analysis was conducted to compare known risk factors for AF and arrhythmia prevalence. Cochran Mantel-Haenszel statistics were used to test for nonzero correlation between AF prevalence and categorical or ordinal variables. Continuous variables were compared using 95% confidence intervals.
Separate logistic regression models were used to estimate odds ratios for AF and height (cm), and for AF and BSA (m2) using SAS version 8.2 (SAS Institute Inc., Cary, North Carolina). Using multivariable logistic regression models allowed for the estimation of the odds of AF in relation to stature while controlling for the confounding effects of known arrhythmia risk factors (continuous variables: age, LVEF, body mass index [BMI]; ordinal variable: NYHA functional class; dichotomous variables: gender, race, hypertension, diabetes, etiology of heart failure, medication use). The two logistic regression models for height and BSA included the same confounding variables, and differed only in the stature variable. A total of 23,478 patients (92.9% of total cohort) had complete data and were included in the multivariable regression analyses.
|
Results
|
|---|
Cohort characteristics.
A total of 25,268 patients were enrolled as of September 30, 2004. The cohort was 72% male and had a mean age of 66 ± 13 years. The mean height of the population was 172.9 ± 10.4 cm, the mean BSA was 2.02 ± 0.28 m2, and the mean BMI was 28.7 ± 6.4. Ischemic cardiomyopathy was present in 72% of the patients, and the mean LVEF was 31 ± 10%. Most subjects had symptomatic heart failure, with 52% having NYHA functional class II symptoms at the time of enrollment. The rest were almost evenly distributed among NYHA functional classes I and III (20% and 25%, respectively). Hypertension and hyperlipidemia were common, affecting 72% and 71% of the study population, respectively. Diabetes was reported in almost one-third of the patients. Medications used included aspirin in 61%, beta-blockers in 79%, angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers (ARB) in 82%, and lipid-lowering therapy in 67% of the patients.
Relationship between patient characteristics and AF status.
A diagnosis of AF was reported in 7,027 subjects (27.8%), and the arrhythmia was paroxysmal in 46.4%, permanent in 41.1%, and of unknown pattern in 12.5% of patients (Table 1). When compared with patients without AF, those with the arrhythmia were more likely to be older white men, with nonischemic cardiomyopathy, and with more advanced heart failure symptoms. The AF patients were significantly more likely to have suffered a stroke or transient ischemic attack, and more often had been placed on antiarrhythmic and anticoagulant medications.
Traditional risk factors and AF prevalence in the ADVANCENT cohort.
As shown in Table 2, older age ( 65 years), male gender, and white race were strongly associated with AF. Patients with valvular heart disease had a two-fold increase in the prevalence of the arrhythmia, and those with hypertension had a nonsignificant trend toward a higher AF prevalence. Beta-blocker and ACE inhibitor/ARB use showed protective effects against the arrhythmia.
Interestingly, diabetic patients and those with a history of coronary artery disease were significantly less likely to have AF than those with LV dysfunction who did not have these problems.
Stature and AF in univariate analysis.
Measurements of stature were complete in 6,994 patients with AF, and only these subjects were included in subsequent analyses. The association between height and AF prevalence is shown in Table 3 and Figure 1. A strong, statistically significant correlation (p < 0.0001) was noted between height and AF, with the prevalence of the arrhythmia increasing from 24% in the shortest to 31.7% in the tallest quartile. This pattern remained true for all forms of the arrhythmia (paroxysmal, permanent, and unknown forms). This strong association was comparable in both men and women.
View larger version (28K):
[in this window]
[in a new window]
|
Figure 1 The prevalence of atrial fibrillation (AF) increases from 24% in the lowest height quartile (Q1) to 31.7% in the highest quartile (Q4) (32% relative increase). This statistically significant correlation (p < 0.0001) remains true for all forms of AF: paroxysmal (18% relative increase from Q1 to Q4), permanent (46% relative increase from Q1 to Q4), and unknown (48% relative increase from Q1 to Q4). Height quartiles were defined differently for men and women (see Methods section).
|
|
Stature and AF in multivariable regression analysis.
To further explore the relationship between stature and AF, two separate multivariable regression analyses, using height and BSA as measures of stature, were performed (Tables 4 and 5). Increasing height was associated with higher AF prevalence independent of other risk factors (odds ratio 1.026/cm, 95% CI 1.022 to 1.030), with a 16-cm (6.2 inches) increase in height translating into a 50% increase in the odds of AF. White race, advanced age, and increasing heart failure class were also independently predictive of AF. Despite a higher prevalence of AF in male patients, gender was not independently associated with the arrhythmia after height was included in the model. In this cohort of patients with reduced LV function, valvular heart disease also showed significantly increased odds of AF (odds ratio 1.802, 95% CI 1.512 to 2.148). Conversely, ACE inhibitor or ARB use and beta-blocker therapy were associated with a lower prevalence of the arrhythmia.
View this table:
[in this window]
[in a new window]
|
Table 4. Multivariate Regression Analysis of Factors Associated With AF Prevalence With Height Representing Stature
|
|
When BSA was substituted for height in the model, which adjusted for confounders such as age and race, a similar positive association was obtained. An increase in BSA of 0.28 m2 would result in a 50% increase in the odds of AF. In a 70-kg, 154-cm-tall individual, this would equate to an increase in height of 53 cm if weight were kept constant, or to an increase in weight of 24-kg if height were kept constant.
There was an unexpected weakly positive association between increasing LVEF and a greater prevalence of AF. This observation resulted from inclusion of both NYHA functional class and ejection fraction, which tightly covary, in the same model. Indeed, when NYHA functional class was removed from the model, this paradoxical result was no longer seen (odds ratio for LVEF: 1.002, 95% CI 0.999 to 1.005).
Stature and LA size.
Echocardiographic data were not collected as part of the ADVANCENT registry. To confirm the relationship between stature and LA size, the echocardiographic data of a subset of registry patients (n = 362) from our institutions (Emory University and Crawford Long Hospitals) were reviewed. The LA diameter was measured using transthoracic echocardiographs in the standard parasternal two-chamber view. There was a strong association between increasing height and larger LA diameter for both men and women in this subset (Fig. 2).
View larger version (44K):
[in this window]
[in a new window]
|
Figure 2 In 362 patients from the Emory University-Crawford Long Hospitals cohort, mean left atrial (LA) diameter was significantly larger in patients whose height exceeded the population median by gender. This association was true for both men and women (p < 0.005 and p = 0.05, respectively).
|
|
|
Discussion
|
|---|
Atrial fibrillation is a common arrhythmia with potentially serious health effects. In addition to impairing quality of life, AF can lead to stroke and a four-fold increase in mortality, particularly in patients with heart failure (10,11). A better understanding of risk factors for AF may help identify a high-risk population that would benefit from the implementation of preventive measures. Available data remain limited by publication volume, sample size, duration of follow-up, and population homogeneity.
In this study we evaluated a large cohort at high risk for AF, given their reduced LVEF at enrollment (4,12), and the high prevalence of factors associated with occurrence of the arrhythmia. The patients were enrolled at 106 centers throughout the country, making this sample more likely to be representative of the demographic composition of the U.S. population. The prevalence of AF was higher in our patients (27.8%) than in the other two largest U.S. cohorts (Framingham Study and Cardiovascular Health Study) (4,5), likely because of lower rates of LV dysfunction in the latter two studies. The prevalence of AF in the current study was similar to that reported by the National Heart Failure Project sample of Medicare beneficiaries hospitalized for heart failure (29.5%) (13).
Our results showed a statistically significant association between stature and AF, independent of other known risk factors for the arrhythmia. In fact, using multivariable analysis to correct for age, gender, race, LVEF, NYHA functional class, hypertension, diabetes, coronary artery disease, valvular heart disease, and medication use, increasing height remained strongly associated with the arrhythmia. Given the mean age of the population enrolled, adult height was reached before AF was diagnosed, and can thus be considered a risk factor for the arrhythmia.
When divided into quartiles, a 32% increase in risk of AF was noted between the lowest and highest height quartiles. The association between AF and BSA was also independent of traditional AF risk factors. Increasing BMI was also shown to be independently associated with increased odds of the arrhythmia when height but not BSA was used as measure of stature. These results confirmed the association between BMI and AF risk described in a recent study that showed a 4% to 5% increase in AF risk with each unit increase in BMI in patients with predominantly preserved LV function (14).
The disparity in the pattern of association between BMI and AF prevalence in the two models can be readily explained when one considers the following. In the first multivariable regression model, when height is kept constant, an increase in BMI can only be observed when body weight is increased, translating into a larger BSA, and presumably a larger LA surface area. On the other hand, in the second model, when BSA is kept constant, higher BMI can only be achieved by reducing height and increasing body weight. As a result, higher BMI in the setting of a constant BSA translates into reduced height, which exerts a protective effect against AF.
Similar to previously published data in large patient populations, age, white race, and heart failure severity were clear risk factors for the arrhythmia. Using odds ratios to infer the clinical significance of these risk factors and their impact on AF burden, valvular heart disease emerged as one of the strongest predictors of AF (4,5). In contrast with other reports, in our study, coronary artery disease and diabetes were associated with a reduced likelihood of AF, whereas hypertension showed a trend toward increased AF risk that did not reach statistical significance (4,5). This disparity can be explained in part by the clinical characteristics of our patient population. All of our subjects had reduced LVEF, and within such a group, processes that result in increased atrial pressure and stretch, particularly valvular disease, are more conducive to AF. Whereas coronary artery disease and diabetes may be risk factors for AF in a general patient population, once patients have developed heart failure, valvular pathology may have a greater impact on LA pressure for any given amount of LV dysfunction. These findings are supported by data from The Cardiovascular Heart Study. Diabetes and coronary artery disease were risk factors for the development of AF in the general population, but not in patients with underlying cardiovascular disease (5). Interestingly, the multivariable analyses showed a weakly positive association between increased LVEF and AF prevalence. This observation is a result of the inclusion of NYHA functional class in the statistical model. In fact, it is well known that LVEF is an imperfect measure of LV function, correlating poorly with NYHA functional class and oxygen consumption during aerobic exercise. Worsening functional class may better reflect the neurohormonal (increased adrenergic tone, activation of the renin-angiotensin axis) and hemodynamic (increased end-diastolic pressures) milieu, thus accounting for the stronger association with AF. The protective effect of ACE inhibitor/ARB (15) and beta-blockers (16) on AF is in agreement with data from previous publications.
Study limitations.
Our study has a cross-sectional design, which precludes the ability to establish a temporal association between the various parameters studied and the development of AF. However, with a mean age of 66 years, it can be safely assumed that all of our patients had reached their adult height before the occurrence of the arrhythmia. As a result, height can be considered a risk factor for rather, than just associated with, AF. Moreover, the agreement between this study and prior reports on the role of the most traditional risk factors for AF further strengthens this assertion.
In this study, AF history was derived from patient interviews, review of medical records, and electrocardiograms. This method of data collection could introduce a reporting bias. However, the reliability of this method is supported by the significantly higher prescription of anticoagulants in the AF group, with 65% of patients on such agents. This estimate is comparable with the rates of warfarin use reported by Fang et al. (46.5%) (17) and Smith et al. (50%) (18) in patients with electrocardiographically proven AF.
Finally, echocardiographic data were evaluated in a small subset of patients. The results showed a clear association between height and LA size, a relationship that has been consistently noted in other studies. Indeed, it seems likely that the propensity for tall individuals to develop AF is explained by their larger LA size. This speculation is supported by observations in animals over the past 30 years that have shown the important association between body size and AF. Large animals such as horses frequently develop spontaneous AF (19). In contrast, dogs and swine can have AF induced only after structural, electrophysiological, or inflammatory insults. The AF is very difficult to maintain in rabbits, and impossible to induce in mice (20). This is likely the result of an atrial surface area that falls below a threshold necessary for the coexistence of a critical number of wavelets needed for the maintenance of AF.
In conclusion, in this large patient population with reduced LVEF, stature measured as height and BSA is a strong predictor of AF occurrence. This association is independent of traditional clinical risk factors for the arrhythmia, and may help to identify a group at particularly high risk for the arrhythmia. Because adult height is reached many years before AF is likely to develop, it allows close monitoring of patients and early targeting of modifiable risk factors for the arrhythmia.
|
References
|
|---|
1. Benjamin EJ, Levy D, Vaziri SM, D’Agostino RB, Belanger AJ, Wolf PA. Independent risk factors for atrial fibrillation in a population-based cohort. The Framingham Heart Study JAMA 1994;271:840-844.[Abstract/Free Full Text]2. Feinberg WM, Blackshear JL, Laupacis A, Kronmal R, Hart RG. Prevalence, age distribution, and gender of patients with atrial fibrillation. Analysis and implications Arch Intern Med 1995;155:469-473.[Abstract/Free Full Text] 3. Braunwald E. Shattuck lecture—cardiovascular medicine at the turn of the millenniumtriumphs, concerns, and opportunities. N Engl J Med 1997;337:1360-1369.[Free Full Text] 4. Kannel WB, Wolf PA, Benjamin EJ, Levy D. Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillationpopulation-based estimates. Am J Cardiol 1998;82:2N-9N.[CrossRef][Web of Science][Medline] 5. Psaty BM, Manolio TA, Kuller LH, et al. Incidence of and risk factors for atrial fibrillation in older adults Circulation 1997;96:2455-2461.[Abstract/Free Full Text] 6. Jalife J. Experimental and clinical AF mechanismsbridging the divide. J Interv Card Electrophysiol 2003;9:85-92.[CrossRef][Web of Science][Medline] 7. Pritchett AM, Jacobsen SJ, Mahoney DW, Rodeheffer RJ, Bailey KR, Redfield MM. Left atrial volume as an index of left atrial sizea population-based study. J Am Coll Cardiol 2003;41:1036-1043.[Abstract/Free Full Text] 8. Vasan RS, Larson MG, Levy D, Evans JC, Benjamin EJ. Distribution and categorization of echocardiographic measurements in relation to reference limits: the Framingham Heart Study: formulation of a height- and sex-specific classification and its prospective validation Circulation 1997;96:1863-1873.[Abstract/Free Full Text] 9. Mosteller RD. Simplified calculation of body-surface area N Engl J Med 1987;317:1098.[Web of Science][Medline] 10. Benjamin EJ, Wolf PA, D’Agostino RB, Silbershatz H, Kannel WB, Levy D. Impact of atrial fibrillation on the risk of deaththe Framingham Heart Study. Circulation 1998;98:946-952.[Abstract/Free Full Text] 11. Wolf PA, Mitchell JB, Baker CS, Kannel WB, D’Agostino RB. Impact of atrial fibrillation on mortality, stroke, and medical costs Arch Intern Med 1998;158:229-234.[Abstract/Free Full Text] 12. Lloyd-Jones DM, Wang TJ, Leip EP, et al. Lifetime risk for development of atrial fibrillationthe Framingham Heart Study. Circulation 2004;110:1042-1046.[Abstract/Free Full Text] 13. Havranek EP, Masoudi FA, Westfall KA, Wolfe P, Ordin DL, Krumholz HM. Spectrum of heart failure in older patientsresults from the National Heart Failure project. Am Heart J 2002;143:412-417.[CrossRef][Web of Science][Medline] 14. Wang TJ, Parise H, Levy D, et al. Obesity and the risk of new-onset atrial fibrillation JAMA 2004;292:2471-2477.[Abstract/Free Full Text] 15. Bourassa MG. Angiotensin II inhibition and prevention of atrial fibrillation and stroke J Am Coll Cardiol 2005;45:720-721.[Free Full Text] 16. McMurray J, Kober L, Robertson M, et al. Antiarrhythmic effect of carvedilol after acute myocardial infarctionresults of the Carvedilol Post-Infarct Survival Control in Left Ventricular Dysfunction (CAPRICORN) trial. J Am Coll Cardiol 2005;45:525-530.[Abstract/Free Full Text] 17. Fang MC, Stafford RS, Ruskin JN, Singer DE. National trends in antiarrhythmic and antithrombotic medication use in atrial fibrillation Arch Intern Med 2004;164:55-60.[Abstract/Free Full Text] 18. Smith NL, Psaty BM, Furberg CD, et al. Temporal trends in the use of anticoagulants among older adults with atrial fibrillation Arch Intern Med 1999;159:1574-1578.[Abstract/Free Full Text] 19. Van Loon G, Blissitt KJ, Keen JA, Young LE. Use of intravenous flecainide in horses with naturally occurring atrial fibrillation Equine Vet J 2004;36:609-614.[CrossRef][Web of Science][Medline] 20. Friedrichs GS. Experimental models of atrial fibrillation/flutter J Pharmacol Toxicol Methods 2000;43:117-123.[CrossRef][Web of Science][Medline]
This article has been cited by other articles:
|
|
|
|
 
A. Rosengren, P. J. Hauptman, G. Lappas, L. Olsson, L. Wilhelmsen, and K. Swedberg
Big men and atrial fibrillation: effects of body size and weight gain on risk of atrial fibrillation in men
Eur. Heart J.,
May 1, 2009;
30(9):
1113 - 1120.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. Mont, R. Elosua, and J. Brugada
Endurance sport practice as a risk factor for atrial fibrillation and atrial flutter
Europace,
January 1, 2009;
11(1):
11 - 17.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. Mont, D. Tamborero, R. Elosua, I. Molina, B. Coll-Vinent, M. Sitges, B. Vidal, A. Scalise, A. Tejeira, A. Berruezo, et al.
Physical activity, height, and left atrial size are independent risk factors for lone atrial fibrillation in middle-aged healthy individuals
Europace,
January 4, 2008;
(2008)
eum263v1.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. Y Tan, J. Myers, and V. F Froelicher
Abstract 2115: Relationship Between Stature and the Prevalence of Atrial Fibrillation and Cardiovascular Mortality
Circulation,
October 31, 2006;
114(18_MeetingAbstracts):
II_429 - II_429.
|
|
|
|
Top
Abstract
Methods