- Email: [email protected]
Relation of Atrial Fibrillation to Glomerular Filtration Rate
Relation of Atrial Fibrillation to Glomerular Filtration Rate Yasuyuki Iguchi, MD*, Kazumi Kimura, MD, Kazuto Kobayashi, MD, Junya Aoki, MD, Yuka Terasawa, MD, Kenichiro Sakai, MD, Junichi Uemura, MD, and Kensaku Shibazaki, MD Although both atrial fibrillation (AF) and decreasing glomerular filtration rate (GFR) are strongly related to advanced age and share common associated vascular risk factors, few studies have explored the relation between AF and GFR. From residents (age >40 years) in Kurashiki City, a total of 41,417 subjects (median age 72 years; 13,956 men) were enrolled in the Kurashiki City Annual Medical Survey from May to December 2006. The estimated overall prevalence of AF was 1.6% (2.8% in the low-GFR tertile, 1.2% in the middle tertile, and 0.9% in the high tertile, p <0.001). After all subjects were categorized into age tertiles (age thresholds 68 and 76 years), AF was identified in 0.9% in the low-GFR tertile, 0.6% in the middle tertile, and 0.5% in the high tertile in the low-age tertile (p ⴝ 0.018); 2.6% in the low-GFR tertile, 1.2% in the middle tertile, and 1.1% in the high tertile in the middle-age tertile (p <0.001); and 3.9% in the low-GFR tertile, 2.4% in the middle tertile, and 1.7% in the high tertile in the high-age tertile (p <0.001). The odds ratio for AF adjusted for age, gender, vascular risk factors, cardiac disease, and hemoglobin was 1.91 (95% confidence interval 1.54 to 2.38, p <0.001) for the low-GFR tertile versus the high tertile and 1.12 (95% confidence interval 0.88 to 1.42, p ⴝ 0.364) for the middle-GFR tertile versus the high tertile. The prevalence of AF gradually increased with decreasing GFR. In conclusion, AF appears to be associated with decreasing GFR. © 2008 Elsevier Inc. All rights reserved. (Am J Cardiol 2008;102:1056 –1059) The prevalence of chronic kidney disease representing decreased glomerular filtration rate (GFR), particularly in the subclinical stages, is very high in elderly subjects.1 Chronic kidney disease is also associated with a markedly increased risk of cardiovascular event and mortality.2,3 However, epidemiologic uncertainties remain regarding the clinical relation between atrial fibrillation (AF) and kidney dysfunction. The present study investigated whether decreasing GFR was associated with the prevalence of AF in a large community-based epidemiologic study based on the Kurashiki City Annual Medical Survey (KAMS). Methods The KAMS was a prospective population-based investigation of risk factors for the presence of AF in men and women ⱖ40 years old. We enrolled 246,246 adult residents in Kurashiki City who received the official mail request to participate in a screening health test from May to December 2006. The Kurashiki City Public Health Center did not send this notification to employees of private companies, offices, government, and factories because the Labor Standards Law has regulated that employers should survey the health condition of employees. We excluded 19 of the 41,436 residents because of incomplete laboratory data in the KAMS.
Department of Stroke Medicine, Kawasaki Medical School, Kurashiki City, Okayama, Japan. Manuscript received May 3, 2008; revised manuscript received and accepted June 9, 2008. This work was supported by Project 104274 from Kawasaki Medical School, Kurashiki City, Okayama, Japan. *Corresponding author: Tel: ⫹81-86-462-1111; Fax: ⫹81-86-4641128. E-mail address: [email protected] (Y. Iguchi). 0002-9149/08/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2008.06.018
Finally, 41,417 residents (median age 72 years; 13,956 men) participated in the present investigation. All subjects answered questions about medical history (hypertension, diabetes mellitus, hypercholesterolemia, cardiac disease, and liver disease) and smoking status and underwent physical examinations, including blood pressure, electrocardiography, and blood testing (serum hemoglobin, total cholesterol, creatinine, fasting glucose, and hemoglobin A1c). Body mass index was calculated as weight in kilograms divided by height in meters squared. Arterial blood pressure was carefully measured in the arm with the patient in a sitting position after resting for a few minutes, following the guidelines for elderly welfare and health in Kurashiki City.4 Cardiovascular risk factors were identified as5 (1) hypertension; use of antihypertensive agents, systolic blood pressure ⱖ140 mm Hg, or diastolic blood pressure ⱖ90 mm Hg on admission; (2) diabetes mellitus; use of oral hypoglycemic agents or insulin, fasting blood glucose ⱖ126 mg/dl, random blood glucose ⱖ200 mg/dl, or glycosylated hemoglobin ⬎6.1%; (3) hypercholesterolemia; use of antihyperlipidemic agents or serum total cholesterol ⬎220 mg/ dl; and (4) smoking; any lifetime experience of cigarette use. Past and present illnesses were based on medical interviews of subjects by physicians.6 Cardiac disease included coronary heart disease and heat failure.6 GFR was estimated using the 4-component Modification of Diet in Renal Disease Study equation, incorporating age, race, gender, and serum creatinine as GFR ⫽ 186 ⫻ (serum creatinine [mg/dl])⫺1.154 ⫻ (age [years])⫺0.203 ⫻ 1.233.7,8 For women, the product of this equation was multiplied by a correction factor of 0.742. Liver disease was determined as a patient with chronic hepatitis or liver cirrhosis with disorder on blood testing. A 12-lead electrocardiogram at rest was recorded in each subject to assess the presence of AF, then www.AJConline.org
Arrhythmias and Conduction Disturbances/AF and GFR
1057
Table 1 Clinical backgrounds of subjects Variables Men Age (y) Age group ⬎70 yrs Hypertension Diabetes mellitus Hyperlipidemia Smoker Cardiac disease Liver disease Height (m) Weight (kg) Body mass index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Atrial fibrillation Laboratory data Hemoglobin (g/dl) Total cholesterol (mg/dl) Creatinine (mg/dl) GFR (ml/min/1.73 m2) Glucose (mg/dl) Hemoglobin A1c (%)
Total Cohort (n ⫽ 41,417)
Low Tertile (n ⫽ 13,810)
13,956 (33.7%) 72 (65–78)
5,145 (37.3%) 76 (71–82)
GFR* Middle Tertile (n ⫽ 13,809)
High Tertile (n ⫽ 13,798)
p Value
4,677 (33.9%) 72 (65–82)
4134 (30.0%) 69 (61–75)
⬍0.001 ⬍0.001
24,167 (58.4%) 14,673 (35.4%) 5,058 (12.2%) 18,255 (44.1%) 3,367 (8.1%) 3,598 (8.7%) 1,211 (2.9%) 1.53 (1.47–1.60) 53.9 (47.4–61.0) 22.7 (20.6–24.9) 132 (120–144) 75 (81–68) 676 (1.6%)
10,510 (76.1%) 6,263 (45.4%) 1,804 (13.1%) 5,868 (42.5%) 966 (7.0%) 1,945 (14.1%) 443 (3.2%) 1.52 (1.47–1.60) 54.0 (47.0–61.5) 23.0 (20.9–25.2) 134 (122–146) 75 (81–68) 383 (2.8%)
7,607 (55.1%) 4,464 (32.3%) 1,502 (10.9%) 6,344 (45.9%) 1,047 (7.6%) 942 (6.8%) 384 (2.8%) 1.54 (1.42–1.66) 54.2 (48.0–61.2) 22.8 (20.7–24.9) 132 (120–144) 75 (82–68) 165 (1.2%)
6,050 (43.8%) 3,946 (28.6%) 1,752 (12.7%) 6,043 (43.8%) 1,354 (9.8%) 711 (5.2%) 384 (2.8%) 1.54 (1.48–1.66) 53.2 (48.0–67.0) 22.4 (20.3–24.6) 131 (120–142) 75 (81–68) 128 (0.9%)
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 0.053 ⬍0.001 0.008 ⬍0.001 ⬍0.001 0.021 ⬍0.001
13.2 (12.3–14.1) 207 (184–230) 0.7 (0.6–0.8) 69.2 (59.1–79.3) 96 (88–108) 5.2 (4.9–5.5)
13.0 (12.0–14.0) 204 (181–227) 0.9 (0.7–1.0) 54.3 (47.2–59.1) 97 (89–111) 5.2 (4.9–5.5)
13.3 (12.5–14.2) 209 (187–232) 0.7 (0.6–0.8) 69.2 (65.9–72.3) 95 (88–106) 5.2 (4.9–5.5)
13.3 (12.5–14.1) 208 (185–231) 0.6 (0.5–0.6) 84.2 (79.3–91.7) 95 (88–107) 5.2 (4.9–5.5)
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 0.049
Values expressed as number (percent) or median (25th to 75th percentile). * Thresholds for increasing GFR tertile were 62.6 and 75.5 ml/min/1.73 m2. Table 2 Adjusted odds ratio for the prevalence of atrial fibrillation (AF) according to glomerular filtration rate (GFR) tertile GFR* High tertile Middle tertile Low tertile
AF (n)
Odds Ratio
128 165 383
1.12 1.91
95% Confidence Interval
1 (reference) 0.88–1.42 1.54–2.38
p Value — 0.364 ⬍0.001
Analysis was adjusted for age, gender, vascular risk factors, and cardiac disease. * Thresholds for increasing GFR tertile were 62.6 and 75.5 ml/min/1.73 m2. Figure 1. Bar graph shows the prevalence of AF according to GFR tertile after all subjects were classified into age tertile. In the low-age tertile, AF was observed in 0.9% of the low-GFR tertile, 0.6% of the middle tertile, and 0.5% of the high tertile (p ⫽ 0.018). In the middle-age tertile, AF was observed in 2.6% of the low-GFR tertile, 1.2% of the middle tertile, and 1.1% in the high tertile (p ⬍0.001). Moreover, in the high-age tertile, subjects with AF were 3.9% of the low-GFR tertile, 2.4% of the middle tertile, and 1.7% of the high tertile (p ⬍0.001). The high prevalence of AF in the low-GFR tertile was similar in each age tertile.
all electrocardiographic records classified as AF were reviewed by a physician. All data from subjects were registered into the database of Kurashiki City Public Health Center. Personal information identifying an individual was blinded. The protocol of the present investigation was approved by the presidential committee of Kawasaki Medical School and was in accordance with the Declaration of Helsinki. At first, all subjects were categorized into 3 groups for GFR tertiles (relative to the lowest tertile). Baseline and demographic data were compared among GFR tertiles using
Fisher’s exact test and Kruskal-Wallis test, as appropriate. We then classified all subjects according to age tertile and recategorized into GFR tertiles in each age tertile. Finally, to estimate the contribution of decreasing GFR to AF using multivariate logistic regression analysis, we constructed a model adjusted for age, gender, hypertension, diabetes mellitus, hyperlipidemia, smoking, and hemoglobin. After referring to the highest GFR tertile, odds ratios with 95% confidence intervals in association with the presence of AF were calculated for each GFR tertile. All statistical analyses were performed using SPSS (SPSS Inc., Chicago, Illinois) for Windows (Microsoft Corp., Redmond, Washington), version 10.0 software. Results were considered significant for p ⬍0.05. Results Baseline data from our series are listed in Table 1. In 41,417 subjects from the KAMS, estimated overall AF prevalence was 1.6% (676 of 41,417). A total of 24,167 subjects
1 (reference) 0.78–1.68 1.48–2.63 1.14 1.97 Analysis was adjusted for age, gender, vascular risk factors, and cardiac disease. Thresholds for increasing age tertiles were 68 and 76 years. * Thresholds for increasing GFR tertiles in the first age tertile were 68.9 and 80.3 ml/min/1.73 m2. † Thresholds for increasing GFR tertiles in the second age tertile were 62.8 and 74.8 ml/min/1.73 m2. ‡ Thresholds for increasing GFR tertiles in the third age tertile were 55.1 and 69.2 ml/min/1.73 m2.
73 106 172 — 0.997 0.001 1 (reference) 0.68–1.45 1.30–2.59 1.00 1.84 51 56 123 — 0.564 0.032 1 (reference) 0.67–2.11 1.05–3.01 1.19 1.80 22 29 44 High Middle Low
95% Confidence Interval Odds Ratio AF (n)
Odds Ratio
95% Confidence Interval
p Value
AF (n)
Odds Ratio
95% Confidence Interval
p Value
AF (n)
High-Age Tertile‡ Middle-Age Tertile† Low-Age Tertile* GFR Tertile
Table 3 Adjusted odds ratio for the prevalence of atrial fibrillation (AF) according to glomerular filtration rate (GFR) tertile after categorizing among age tertiles
— 0.501 ⬍0.001
The American Journal of Cardiology (www.AJConline.org) p Value
1058
(58.4%) were ⬎70 years old. With thresholds for increasing GFR tertile of 62.6 and 75.5 ml/min/1.73 m2, significant differences among GFR tertiles were seen regarding frequencies of age ⬎70 years, men, vascular risk factors, and cardiac disease, but not liver disease. Concerning blood testing, significant differences in hemoglobin (p ⬍0.001) were identified. We classified all 41,417 subjects into age tertiles (thresholds of 68 and 76 years), then subcategorized each age tertile according to GFR tertile (Figure 1). In the low-age tertile, AF was found in 0.9% in the low-GFR tertile, 0.6% in the middle tertile, and 0.5% in the high tertile (p ⫽ 0.018). In the middle-age tertile, AF was seen in 2.6% in the low-GFR tertile, 1.2% in the middle tertile, and 1.1% in the high tertile (p ⬍0.001). Moreover, in the high-age tertile, subjects with AF comprised 3.9% of the low-GFR tertile, 2.4% of the middle tertile, and 1.7% of the high tertile (p ⬍0.001). The prevalence of AF gradually increased in accordance with decreasing GFR in each age tertile. Table 2 lists associations between GFR and prevalence of AF adjusted by age, male gender, vascular risk factors, cardiac disease, and hemoglobin. Similar to univariate analysis, multivariate regression models showed a trend toward increasing prevalence of AF with decreasing GFR. Similar tendencies were seen when we categorized all subjects into age tertiles (Table 3). Discussion In this population-based study supported by a public annual medical survey, we found a 1.6% prevalence of AF in adults ⱖ40 years old and showed a strong and graded inverse association between GFR and prevalence of AF independent of other vascular risk factors. We identified clinical associations between AF and decreasing GFR unrelated to age increases. To the best of our knowledge, no clear data for kidney dysfunction have been clarified in community- or population-based studies. In cross-sectional and retrospective studies investigating the prevalence of AF in patients with kidney dysfunction, AF was seen in 27% of patients with terminal kidney disease treated using hemodialysis. The AF prevalence in these patients was 3 to 15 times higher than in the Framingham population.9,10 Left atrial dilatation in hemodialysis patients was independently associated with the presence of AF.10 In terms of pathophysiologic mechanism, patients with severe kidney dysfunction on hemodialysis therapy have been considered prone to structural and electrical atrial remodeling.11 Although several limitations exist in reaching a definitive conclusion, decreasing GFR seems to have an important role in increasing the prevalence of AF in general populations. When we inferred the mechanisms of AF, the idea that both atherosclerosis and inflammatory processes are involved in the pathogenesis was not new, but renewed focus was gained from clinical and epidemiologic observations. The first observation of lone AF in patients with biopsy specimens showed a high prevalence of inflammatory infiltrates, myocyte necrosis, and fibrosis, whereas biopsy specimens from control patients were normal.12 Furthermore, an association between increased C-reactive protein and prevalence of AF has been confirmed.13 Progression of kidney
Arrhythmias and Conduction Disturbances/AF and GFR
insufficiency was also accompanied by increased inflammation in subjects with chronic kidney disease.14 As a result, these previous studies suggested the existence of associations between inflammation and AF or kidney dysfunction. Our results showed a linear correlation between decreasing GFR and AF prevalence, although no such evidence was seen from inflammatory markers. We thus considered the increasing prevalence of AF as attributable to inflammation caused by decreasing GFR. Our results, which included clinical associations between decreasing GFR and AF, may lead to tactics to successfully decrease the incidence of AF. In this context, we were able to suggest such agents as hydroxymethylglutaryl coenzyme A reductase inhibitors (statins), angiotensin-converting enzyme inhibitors (ACEIs), and angiotensin II receptor blockers (ARBs). First, 2 studies were conducted to estimate the efficacy of statins for decreasing the prevalence of AF.15,16 Although 2 studies found an effect of statins on AF, 1 randomized study found no decrease in AF for patents treated with statins compared with patients on standard therapy.17 One limitation of these studies was that none measured markers of inflammation that were mainly controlled by statins. Data for statins and AF thus seemed inconclusive. Second, growing evidence from clinical studies suggested involvement of the angiotensin system in the pathophysiologic process of AF. In a meta-analysis, ARB treatments markedly decreased the risk of the development of new-onset AF.18 The anti-inflammatory actions of ACEIs and ARBs seemed to contribute not only to beneficial antiarrhythmic effects for AF, but also protection of kidney function.19 Thus, considering our results, use of ACEIs or ARBs may offer an adequate option for monitoring kidney function. To investigate whether protection of kidney function decreases the incidence of AF, prospective studies are needed. Several limitations have been identified in the present study. First, baseline characteristics did not include history of stroke; therefore, confirming the contribution of stroke to the prevalence of AF is difficult. Second, the relatively small proportion of men may have resulted in insufficient power to discriminate differences in AF prevalence among GFR quartiles. Finally, because paroxysmal AF was difficult to diagnose from a single electrocardiographic examination, not all asymptomatic subjects with paroxysmal AF would have been identified in this study.
2. 3. 4. 5. 6. 7.
8.
9. 10.
11.
12. 13. 14.
15. 16.
17.
18.
Acknowledgments: We thank Takehiko Tokura, MD, in the Department of Internal Medicine, Kawasaki Medical School. 19. 1. Nissenson AR, Pereira BJ, Collins AJ, Steinberg EP. Prevalence and characteristics of individuals with chronic kidney disease in a
1059
large health maintenance organization. Am J Kidney Dis 2001;37: 1177–1183. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med 2004;351:1296 –1305. Bos MJ, Koudstaal PJ, Hofman A, Breteler MM. Decreased glomerular filtration rate is a risk factor for hemorrhagic but not for ischemic stroke: the Rotterdam Study. Stroke 2007;38:3127–3132. Guidelines of Elders’ Welfare and Health in Kurashiki-City. Kurashiki City, Japan: Kurashiki City Public Health Center, 2003:42–44. Jeong JH. Prevalence of and risk factors for atrial fibrillation in Korean adults older than 40 years. J Korean Med Sci 2005;20:26 –30. Guidelines of Elders’ Welfare and Health in Kurashiki-City. Kurashiki City, Japan: Kurashiki City Public Health Center, 2003:71–72. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999;130:461– 470. Imai E, Horio M, Nitta K, Yamagata K, Iseki K, Hara S, Ura N, Kiyohara Y, Hirakata H, Watanabe T, et al. Estimation of glomerular filtration rate by the MDRD study equation modified for Japanese patients with chronic kidney disease. Clin Exp Nephrol 2007;11:41–50. 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. Genovesi S, Pogliani D, Faini A, Valsecchi MG, Riva A, Stefani F, Acquistapace I, Stella A, Bonforte G, DeVecchi A, et al. Prevalence of atrial fibrillation and associated factors in a population of long-term hemodialysis patients. Am J Kidney Dis 2005;46:897–902. Cozma D, Popescu BA, Lighezan D, Lucian P, Mornos C, Ginghina C, Dragulescu SI. Left atrial remodeling: assessment of size and shape to detect vulnerability to atrial fibrillation. Pacing Clin Electrophysiol 2007;30(suppl 1):S147–S150. Frustaci A, Chimenti C, Bellocci F, Morgante E, Russo MA, Maseri A. Histological substrate of atrial biopsies in patients with lone atrial fibrillation. Circulation 1997;96:1180 –1184. Asselbergs FW, van den Berg MP, Diercks GF, van Gilst WH, van Veldhuisen DJ. C-Reactive protein and microalbuminuria are associated with atrial fibrillation. Int J Cardiol 2005;98:73–77. Goicoechea M, de Vinuesa SG, Lahera V, Cachofeiro V, GomezCampdera F, Vega A, Abad S, Luno J. Effects of atorvastatin on inflammatory and fibrinolytic parameters in patients with chronic kidney disease. J Am Soc Nephrol 2006;17(suppl):S231–S235. Siu CW, Lau CP, Tse HF. Prevention of atrial fibrillation recurrence by statin therapy in patients with lone atrial fibrillation after successful cardioversion. Am J Cardiol 2003;92:1343–1345. Hanna IR, Heeke B, Bush H, Brosius L, King-Hageman D, Dudley SC Jr., Beshai JF, Langberg JJ. Lipid-lowering drug use is associated with reduced prevalence of atrial fibrillation in patients with left ventricular systolic dysfunction. Heart Rhythm 2006;3:881– 886. Tveit A, Grundtvig M, Gundersen T, Vanberg P, Semb AG, Holt E, Gullestad L. Analysis of pravastatin to prevent recurrence of atrial fibrillation after electrical cardioversion. Am J Cardiol 2004;93:780 – 782. Madrid AH, Peng J, Zamora J, Marin I, Bernal E, Escobar C, MunosTinoco C, Rebollo JM, Moro C. The role of angiotensin receptor blockers and/or angiotensin converting enzyme inhibitors in the prevention of atrial fibrillation in patients with cardiovascular diseases: meta-analysis of randomized controlled clinical trials. Pacing Clin Electrophysiol 2004;27:1405–1410. Morimoto S, Yano Y, Maki K, Sawada K. Renal and vascular protective effects of telmisartan in patients with essential hypertension. Hypertens Res 2006;29:567–572.
Department of Stroke Medicine, Kawasaki Medical School, Kurashiki City, Okayama, Japan. Manuscript received May 3, 2008; revised manuscript received and accepted June 9, 2008. This work was supported by Project 104274 from Kawasaki Medical School, Kurashiki City, Okayama, Japan. *Corresponding author: Tel: ⫹81-86-462-1111; Fax: ⫹81-86-4641128. E-mail address: [email protected] (Y. Iguchi). 0002-9149/08/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2008.06.018
Finally, 41,417 residents (median age 72 years; 13,956 men) participated in the present investigation. All subjects answered questions about medical history (hypertension, diabetes mellitus, hypercholesterolemia, cardiac disease, and liver disease) and smoking status and underwent physical examinations, including blood pressure, electrocardiography, and blood testing (serum hemoglobin, total cholesterol, creatinine, fasting glucose, and hemoglobin A1c). Body mass index was calculated as weight in kilograms divided by height in meters squared. Arterial blood pressure was carefully measured in the arm with the patient in a sitting position after resting for a few minutes, following the guidelines for elderly welfare and health in Kurashiki City.4 Cardiovascular risk factors were identified as5 (1) hypertension; use of antihypertensive agents, systolic blood pressure ⱖ140 mm Hg, or diastolic blood pressure ⱖ90 mm Hg on admission; (2) diabetes mellitus; use of oral hypoglycemic agents or insulin, fasting blood glucose ⱖ126 mg/dl, random blood glucose ⱖ200 mg/dl, or glycosylated hemoglobin ⬎6.1%; (3) hypercholesterolemia; use of antihyperlipidemic agents or serum total cholesterol ⬎220 mg/ dl; and (4) smoking; any lifetime experience of cigarette use. Past and present illnesses were based on medical interviews of subjects by physicians.6 Cardiac disease included coronary heart disease and heat failure.6 GFR was estimated using the 4-component Modification of Diet in Renal Disease Study equation, incorporating age, race, gender, and serum creatinine as GFR ⫽ 186 ⫻ (serum creatinine [mg/dl])⫺1.154 ⫻ (age [years])⫺0.203 ⫻ 1.233.7,8 For women, the product of this equation was multiplied by a correction factor of 0.742. Liver disease was determined as a patient with chronic hepatitis or liver cirrhosis with disorder on blood testing. A 12-lead electrocardiogram at rest was recorded in each subject to assess the presence of AF, then www.AJConline.org
Arrhythmias and Conduction Disturbances/AF and GFR
1057
Table 1 Clinical backgrounds of subjects Variables Men Age (y) Age group ⬎70 yrs Hypertension Diabetes mellitus Hyperlipidemia Smoker Cardiac disease Liver disease Height (m) Weight (kg) Body mass index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Atrial fibrillation Laboratory data Hemoglobin (g/dl) Total cholesterol (mg/dl) Creatinine (mg/dl) GFR (ml/min/1.73 m2) Glucose (mg/dl) Hemoglobin A1c (%)
Total Cohort (n ⫽ 41,417)
Low Tertile (n ⫽ 13,810)
13,956 (33.7%) 72 (65–78)
5,145 (37.3%) 76 (71–82)
GFR* Middle Tertile (n ⫽ 13,809)
High Tertile (n ⫽ 13,798)
p Value
4,677 (33.9%) 72 (65–82)
4134 (30.0%) 69 (61–75)
⬍0.001 ⬍0.001
24,167 (58.4%) 14,673 (35.4%) 5,058 (12.2%) 18,255 (44.1%) 3,367 (8.1%) 3,598 (8.7%) 1,211 (2.9%) 1.53 (1.47–1.60) 53.9 (47.4–61.0) 22.7 (20.6–24.9) 132 (120–144) 75 (81–68) 676 (1.6%)
10,510 (76.1%) 6,263 (45.4%) 1,804 (13.1%) 5,868 (42.5%) 966 (7.0%) 1,945 (14.1%) 443 (3.2%) 1.52 (1.47–1.60) 54.0 (47.0–61.5) 23.0 (20.9–25.2) 134 (122–146) 75 (81–68) 383 (2.8%)
7,607 (55.1%) 4,464 (32.3%) 1,502 (10.9%) 6,344 (45.9%) 1,047 (7.6%) 942 (6.8%) 384 (2.8%) 1.54 (1.42–1.66) 54.2 (48.0–61.2) 22.8 (20.7–24.9) 132 (120–144) 75 (82–68) 165 (1.2%)
6,050 (43.8%) 3,946 (28.6%) 1,752 (12.7%) 6,043 (43.8%) 1,354 (9.8%) 711 (5.2%) 384 (2.8%) 1.54 (1.48–1.66) 53.2 (48.0–67.0) 22.4 (20.3–24.6) 131 (120–142) 75 (81–68) 128 (0.9%)
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 0.053 ⬍0.001 0.008 ⬍0.001 ⬍0.001 0.021 ⬍0.001
13.2 (12.3–14.1) 207 (184–230) 0.7 (0.6–0.8) 69.2 (59.1–79.3) 96 (88–108) 5.2 (4.9–5.5)
13.0 (12.0–14.0) 204 (181–227) 0.9 (0.7–1.0) 54.3 (47.2–59.1) 97 (89–111) 5.2 (4.9–5.5)
13.3 (12.5–14.2) 209 (187–232) 0.7 (0.6–0.8) 69.2 (65.9–72.3) 95 (88–106) 5.2 (4.9–5.5)
13.3 (12.5–14.1) 208 (185–231) 0.6 (0.5–0.6) 84.2 (79.3–91.7) 95 (88–107) 5.2 (4.9–5.5)
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 0.049
Values expressed as number (percent) or median (25th to 75th percentile). * Thresholds for increasing GFR tertile were 62.6 and 75.5 ml/min/1.73 m2. Table 2 Adjusted odds ratio for the prevalence of atrial fibrillation (AF) according to glomerular filtration rate (GFR) tertile GFR* High tertile Middle tertile Low tertile
AF (n)
Odds Ratio
128 165 383
1.12 1.91
95% Confidence Interval
1 (reference) 0.88–1.42 1.54–2.38
p Value — 0.364 ⬍0.001
Analysis was adjusted for age, gender, vascular risk factors, and cardiac disease. * Thresholds for increasing GFR tertile were 62.6 and 75.5 ml/min/1.73 m2. Figure 1. Bar graph shows the prevalence of AF according to GFR tertile after all subjects were classified into age tertile. In the low-age tertile, AF was observed in 0.9% of the low-GFR tertile, 0.6% of the middle tertile, and 0.5% of the high tertile (p ⫽ 0.018). In the middle-age tertile, AF was observed in 2.6% of the low-GFR tertile, 1.2% of the middle tertile, and 1.1% in the high tertile (p ⬍0.001). Moreover, in the high-age tertile, subjects with AF were 3.9% of the low-GFR tertile, 2.4% of the middle tertile, and 1.7% of the high tertile (p ⬍0.001). The high prevalence of AF in the low-GFR tertile was similar in each age tertile.
all electrocardiographic records classified as AF were reviewed by a physician. All data from subjects were registered into the database of Kurashiki City Public Health Center. Personal information identifying an individual was blinded. The protocol of the present investigation was approved by the presidential committee of Kawasaki Medical School and was in accordance with the Declaration of Helsinki. At first, all subjects were categorized into 3 groups for GFR tertiles (relative to the lowest tertile). Baseline and demographic data were compared among GFR tertiles using
Fisher’s exact test and Kruskal-Wallis test, as appropriate. We then classified all subjects according to age tertile and recategorized into GFR tertiles in each age tertile. Finally, to estimate the contribution of decreasing GFR to AF using multivariate logistic regression analysis, we constructed a model adjusted for age, gender, hypertension, diabetes mellitus, hyperlipidemia, smoking, and hemoglobin. After referring to the highest GFR tertile, odds ratios with 95% confidence intervals in association with the presence of AF were calculated for each GFR tertile. All statistical analyses were performed using SPSS (SPSS Inc., Chicago, Illinois) for Windows (Microsoft Corp., Redmond, Washington), version 10.0 software. Results were considered significant for p ⬍0.05. Results Baseline data from our series are listed in Table 1. In 41,417 subjects from the KAMS, estimated overall AF prevalence was 1.6% (676 of 41,417). A total of 24,167 subjects
1 (reference) 0.78–1.68 1.48–2.63 1.14 1.97 Analysis was adjusted for age, gender, vascular risk factors, and cardiac disease. Thresholds for increasing age tertiles were 68 and 76 years. * Thresholds for increasing GFR tertiles in the first age tertile were 68.9 and 80.3 ml/min/1.73 m2. † Thresholds for increasing GFR tertiles in the second age tertile were 62.8 and 74.8 ml/min/1.73 m2. ‡ Thresholds for increasing GFR tertiles in the third age tertile were 55.1 and 69.2 ml/min/1.73 m2.
73 106 172 — 0.997 0.001 1 (reference) 0.68–1.45 1.30–2.59 1.00 1.84 51 56 123 — 0.564 0.032 1 (reference) 0.67–2.11 1.05–3.01 1.19 1.80 22 29 44 High Middle Low
95% Confidence Interval Odds Ratio AF (n)
Odds Ratio
95% Confidence Interval
p Value
AF (n)
Odds Ratio
95% Confidence Interval
p Value
AF (n)
High-Age Tertile‡ Middle-Age Tertile† Low-Age Tertile* GFR Tertile
Table 3 Adjusted odds ratio for the prevalence of atrial fibrillation (AF) according to glomerular filtration rate (GFR) tertile after categorizing among age tertiles
— 0.501 ⬍0.001
The American Journal of Cardiology (www.AJConline.org) p Value
1058
(58.4%) were ⬎70 years old. With thresholds for increasing GFR tertile of 62.6 and 75.5 ml/min/1.73 m2, significant differences among GFR tertiles were seen regarding frequencies of age ⬎70 years, men, vascular risk factors, and cardiac disease, but not liver disease. Concerning blood testing, significant differences in hemoglobin (p ⬍0.001) were identified. We classified all 41,417 subjects into age tertiles (thresholds of 68 and 76 years), then subcategorized each age tertile according to GFR tertile (Figure 1). In the low-age tertile, AF was found in 0.9% in the low-GFR tertile, 0.6% in the middle tertile, and 0.5% in the high tertile (p ⫽ 0.018). In the middle-age tertile, AF was seen in 2.6% in the low-GFR tertile, 1.2% in the middle tertile, and 1.1% in the high tertile (p ⬍0.001). Moreover, in the high-age tertile, subjects with AF comprised 3.9% of the low-GFR tertile, 2.4% of the middle tertile, and 1.7% of the high tertile (p ⬍0.001). The prevalence of AF gradually increased in accordance with decreasing GFR in each age tertile. Table 2 lists associations between GFR and prevalence of AF adjusted by age, male gender, vascular risk factors, cardiac disease, and hemoglobin. Similar to univariate analysis, multivariate regression models showed a trend toward increasing prevalence of AF with decreasing GFR. Similar tendencies were seen when we categorized all subjects into age tertiles (Table 3). Discussion In this population-based study supported by a public annual medical survey, we found a 1.6% prevalence of AF in adults ⱖ40 years old and showed a strong and graded inverse association between GFR and prevalence of AF independent of other vascular risk factors. We identified clinical associations between AF and decreasing GFR unrelated to age increases. To the best of our knowledge, no clear data for kidney dysfunction have been clarified in community- or population-based studies. In cross-sectional and retrospective studies investigating the prevalence of AF in patients with kidney dysfunction, AF was seen in 27% of patients with terminal kidney disease treated using hemodialysis. The AF prevalence in these patients was 3 to 15 times higher than in the Framingham population.9,10 Left atrial dilatation in hemodialysis patients was independently associated with the presence of AF.10 In terms of pathophysiologic mechanism, patients with severe kidney dysfunction on hemodialysis therapy have been considered prone to structural and electrical atrial remodeling.11 Although several limitations exist in reaching a definitive conclusion, decreasing GFR seems to have an important role in increasing the prevalence of AF in general populations. When we inferred the mechanisms of AF, the idea that both atherosclerosis and inflammatory processes are involved in the pathogenesis was not new, but renewed focus was gained from clinical and epidemiologic observations. The first observation of lone AF in patients with biopsy specimens showed a high prevalence of inflammatory infiltrates, myocyte necrosis, and fibrosis, whereas biopsy specimens from control patients were normal.12 Furthermore, an association between increased C-reactive protein and prevalence of AF has been confirmed.13 Progression of kidney
Arrhythmias and Conduction Disturbances/AF and GFR
insufficiency was also accompanied by increased inflammation in subjects with chronic kidney disease.14 As a result, these previous studies suggested the existence of associations between inflammation and AF or kidney dysfunction. Our results showed a linear correlation between decreasing GFR and AF prevalence, although no such evidence was seen from inflammatory markers. We thus considered the increasing prevalence of AF as attributable to inflammation caused by decreasing GFR. Our results, which included clinical associations between decreasing GFR and AF, may lead to tactics to successfully decrease the incidence of AF. In this context, we were able to suggest such agents as hydroxymethylglutaryl coenzyme A reductase inhibitors (statins), angiotensin-converting enzyme inhibitors (ACEIs), and angiotensin II receptor blockers (ARBs). First, 2 studies were conducted to estimate the efficacy of statins for decreasing the prevalence of AF.15,16 Although 2 studies found an effect of statins on AF, 1 randomized study found no decrease in AF for patents treated with statins compared with patients on standard therapy.17 One limitation of these studies was that none measured markers of inflammation that were mainly controlled by statins. Data for statins and AF thus seemed inconclusive. Second, growing evidence from clinical studies suggested involvement of the angiotensin system in the pathophysiologic process of AF. In a meta-analysis, ARB treatments markedly decreased the risk of the development of new-onset AF.18 The anti-inflammatory actions of ACEIs and ARBs seemed to contribute not only to beneficial antiarrhythmic effects for AF, but also protection of kidney function.19 Thus, considering our results, use of ACEIs or ARBs may offer an adequate option for monitoring kidney function. To investigate whether protection of kidney function decreases the incidence of AF, prospective studies are needed. Several limitations have been identified in the present study. First, baseline characteristics did not include history of stroke; therefore, confirming the contribution of stroke to the prevalence of AF is difficult. Second, the relatively small proportion of men may have resulted in insufficient power to discriminate differences in AF prevalence among GFR quartiles. Finally, because paroxysmal AF was difficult to diagnose from a single electrocardiographic examination, not all asymptomatic subjects with paroxysmal AF would have been identified in this study.
2. 3. 4. 5. 6. 7.
8.
9. 10.
11.
12. 13. 14.
15. 16.
17.
18.
Acknowledgments: We thank Takehiko Tokura, MD, in the Department of Internal Medicine, Kawasaki Medical School. 19. 1. Nissenson AR, Pereira BJ, Collins AJ, Steinberg EP. Prevalence and characteristics of individuals with chronic kidney disease in a
1059
large health maintenance organization. Am J Kidney Dis 2001;37: 1177–1183. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med 2004;351:1296 –1305. Bos MJ, Koudstaal PJ, Hofman A, Breteler MM. Decreased glomerular filtration rate is a risk factor for hemorrhagic but not for ischemic stroke: the Rotterdam Study. Stroke 2007;38:3127–3132. Guidelines of Elders’ Welfare and Health in Kurashiki-City. Kurashiki City, Japan: Kurashiki City Public Health Center, 2003:42–44. Jeong JH. Prevalence of and risk factors for atrial fibrillation in Korean adults older than 40 years. J Korean Med Sci 2005;20:26 –30. Guidelines of Elders’ Welfare and Health in Kurashiki-City. Kurashiki City, Japan: Kurashiki City Public Health Center, 2003:71–72. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999;130:461– 470. Imai E, Horio M, Nitta K, Yamagata K, Iseki K, Hara S, Ura N, Kiyohara Y, Hirakata H, Watanabe T, et al. Estimation of glomerular filtration rate by the MDRD study equation modified for Japanese patients with chronic kidney disease. Clin Exp Nephrol 2007;11:41–50. 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. Genovesi S, Pogliani D, Faini A, Valsecchi MG, Riva A, Stefani F, Acquistapace I, Stella A, Bonforte G, DeVecchi A, et al. Prevalence of atrial fibrillation and associated factors in a population of long-term hemodialysis patients. Am J Kidney Dis 2005;46:897–902. Cozma D, Popescu BA, Lighezan D, Lucian P, Mornos C, Ginghina C, Dragulescu SI. Left atrial remodeling: assessment of size and shape to detect vulnerability to atrial fibrillation. Pacing Clin Electrophysiol 2007;30(suppl 1):S147–S150. Frustaci A, Chimenti C, Bellocci F, Morgante E, Russo MA, Maseri A. Histological substrate of atrial biopsies in patients with lone atrial fibrillation. Circulation 1997;96:1180 –1184. Asselbergs FW, van den Berg MP, Diercks GF, van Gilst WH, van Veldhuisen DJ. C-Reactive protein and microalbuminuria are associated with atrial fibrillation. Int J Cardiol 2005;98:73–77. Goicoechea M, de Vinuesa SG, Lahera V, Cachofeiro V, GomezCampdera F, Vega A, Abad S, Luno J. Effects of atorvastatin on inflammatory and fibrinolytic parameters in patients with chronic kidney disease. J Am Soc Nephrol 2006;17(suppl):S231–S235. Siu CW, Lau CP, Tse HF. Prevention of atrial fibrillation recurrence by statin therapy in patients with lone atrial fibrillation after successful cardioversion. Am J Cardiol 2003;92:1343–1345. Hanna IR, Heeke B, Bush H, Brosius L, King-Hageman D, Dudley SC Jr., Beshai JF, Langberg JJ. Lipid-lowering drug use is associated with reduced prevalence of atrial fibrillation in patients with left ventricular systolic dysfunction. Heart Rhythm 2006;3:881– 886. Tveit A, Grundtvig M, Gundersen T, Vanberg P, Semb AG, Holt E, Gullestad L. Analysis of pravastatin to prevent recurrence of atrial fibrillation after electrical cardioversion. Am J Cardiol 2004;93:780 – 782. Madrid AH, Peng J, Zamora J, Marin I, Bernal E, Escobar C, MunosTinoco C, Rebollo JM, Moro C. The role of angiotensin receptor blockers and/or angiotensin converting enzyme inhibitors in the prevention of atrial fibrillation in patients with cardiovascular diseases: meta-analysis of randomized controlled clinical trials. Pacing Clin Electrophysiol 2004;27:1405–1410. Morimoto S, Yano Y, Maki K, Sawada K. Renal and vascular protective effects of telmisartan in patients with essential hypertension. Hypertens Res 2006;29:567–572.