- Email: [email protected]
Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors
IJCA-18160; No of Pages 6 International Journal of Cardiology xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors Angela Sciacqua a,1, Maria Perticone c,1, Giovanni Tripepi b, Sofia Miceli a, Eliezer J. Tassone a, Nadia Grillo a, Giuseppe Carullo a, Giorgio Sesti a, Francesco Perticone a,⁎ a b c
Department of Medical and Surgical Sciences, University Magna Græcia of Catanzaro, Italy CNR, Istituto di Biomedicina, Clinical Epidemiology and Physiopathology of Renal Diseases and Hypertension, Reggio Calabria, Italy Experimental and Clinical Medicine Experimental and Clinical Medicine
a r t i c l e
i n f o
Article history: Received 18 February 2014 Received in revised form 24 April 2014 Accepted 26 April 2014 Available online xxxx Keywords: Atrial fibrillation Chronic kidney disease Atrial volume Left ventricular hypertrophy
a b s t r a c t Objectives: In this prospective population-based study, we tested the possible interaction between chronic kidney disease (CKD) and left atrium volume index (LAVI) in predicting incident atrial fibrillation (AF). Methods: We enrolled 3549 Caucasian subjects, 1829 men and 1720 women, aged 60.7 ± 10.6 years, without baseline AF and thyroid disorders. Echocardiographic left ventricular mass and LAVI were measured. Renal function was calculated by estimated glomerular filtration rate (e-GFR). To test the effect of some clinical confounders on incident AF, we constructed different models including clinical and laboratory parameters. AF diagnosis was made by standard electrocardiogram or 24-h ECG-Holter, hospital discharge diagnoses, and by the all-clinical documentation. Results: During the follow-up (53.3 ± 18.1 months), 546 subjects developed AF (4.5 events/100 patient-years). Progressors to AF were older, had a higher body mass index, blood pressure, LDL-cholesterol, glucose, cardiac mass, and LAVI, and had lower e-GFR. Hypertension, metabolic syndrome, diabetes, cardiac hypertrophy and CKD were more common among AF cases than controls. In the final Cox regression model, variables that remained significantly associated with AF were: cardiac hypertrophy (HR = 1.495, 95% CI = 1.215–1.841), renal disease (HR = 1.528, 95% CI = 1.261–1.851), age (HR = 1.586, 95% CI = 1.461–1.725) and LAVI (HR = 2.920, 95% CI = 2.426–3.515). The interaction analysis demonstrated a synergic effect between CKD and cardiac hypertrophy (HR = 4.040, 95% CI = 2.661–6.133), as well as between CKD and LAVI (HR = 4.875, 95% CI = 2.699–8.805). The coexistence of all three subclinical organ damages significantly increases the arrhythmic risk (HR = 7.185, 95% CI = 5.041–10.240). Conclusions: Our data demonstrate that LAVI and CKD significantly interact in a synergic manner in increasing AF risk. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Chronic kidney disease (CKD) is a relevant health problem in the general population and in other conditions [1] because it is strongly associated with high cardiovascular morbidity and mortality [2,3]. Notably, some evidences demonstrated that mild renal dysfunction, defined as a small increase in serum creatinine (b 2 mg/dl), abnormal urinary albumin excretion and/or a reduction of glomerular filtration rate (GFR) [4,5], is a powerful predictor of cardiovascular events in hypertensives. Probably, hemodynamic and atherogenic factors, such as insulin resistance, endothelial dysfunction and subclinical inflammation
⁎ Corresponding author at: Department of Medical and Surgical Sciences, Campus Universitario di Germaneto, V.le Europa, 88100 Catanzaro, Italy. Tel.: +39 0961 3647149; fax: +39 0961 3647634. E-mail address: [email protected] (F. Perticone). 1 These authors have equally contributed to the work.
may contribute to explain these evidences [3,6–9]. In addition, CKD is characterized by volume expansion due to sodium retention, activation of the renin–angiotensin–aldosterone system (RAAS) and sympathetic overactivity that promote ventricular and atrial remodeling [10,11], favoring the development of atrial fibrillation (AF). In adult and old subjects, AF is the most frequent cardiac rhythm disorder, impacting on the economic costs of caring because it is associated with an increased risk for thromboembolic events [12]. In addition to the different clinical conditions—such as valvular disease, congestive heart failure and myocardial infarction—recent and limited data showed that also CKD is associated with incident AF [13,14]. Usually, cardiac structural remodeling represents an adaptive response of the myocardium to increased cardiac workload in high blood pressure (BP), renal disease [9,15] and other clinical conditions [16,17], even if the increased left ventricular mass (LVM) is recognized as an independent predictor of cardiovascular events in general population and other clinical settings [18]. In addition, it is associated
http://dx.doi.org/10.1016/j.ijcard.2014.04.259 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259
2
A. Sciacqua et al. / International Journal of Cardiology xxx (2014) xxx–xxx
with diastolic dysfunction and left atrium (LA) enlargement that represents an important pathophysiological mechanism involved in incident AF [19] that, in turn, induces anatomical and electrical atrial remodeling and, thus, the maintenance of itself [20]. Taken together, at this moment there are no data about a possible interaction between CKD and LA volume index (LAVI) in predicting new AF. So far, we tested this hypothesis in a large cohort of Caucasian subjects. 2. Materials and methods 2.1. Study population Our study is a population-based prospective study aimed at assessing the possible interaction of LAVI and renal dysfunction on incident AF. Between January 1998 and October 2011 we enrolled 3549 Caucasian outpatients (1829 men, 1720 women; aged 60.7 ± 10.6 years). The study population consists of outpatients with metabolic and cardiovascular risk factors referred to our tertiary care setting for screening program. All patients were prospectively followed to evaluate the development of incident diabetes, AF and fatal and non-fatal cardiovascular events. All patients underwent physical examination, review of their medical history and anthropometrical evaluation: weight, height, and body mass index (BMI). We excluded patients with thyroid disorders or those taking drugs affecting thyroid function, rheumatic and non-rheumatic valvular heart disease, prosthetic valves, congestive heart failure, cardiomyopathies, dialyzed patients and malignant disease. At the time of enrolment none of the patients have experienced myocardial infarction or stroke. Metabolic syndrome (MS) was defined according to NCEP-ATP III criteria [21]. The ethical Committee of University Hospital of Catanzaro approved the protocol and informed written consent was obtained from all participants. All the investigations were performed according with the principles of Helsinki Declaration. 2.2. Blood pressure measurements Patients taking antihypertensive drugs or with a systolic BP (SBP) ≥ 140 mm Hg and/or diastolic BP (DPB) ≥ 90 mm Hg were defined as hypertensive. Secondary forms of hypertension were excluded by a standard clinical protocol. 2.3. Diabetes evaluation Criteria for type 2 diabetes were: (1) presence of more than one classic symptom of hyperglycemia plus either a fasting plasma glucose ≥ 7.0 mmol/l or random plasma glucose ≥11.1 mmol/l, (2) two or more elevated plasma glucose concentrations (fasting plasma glucose ≥7.0 mmol/l, random plasma glucose ≥11.1 mmol/l, or 2-h plasma glucose ≥11.1 mmol/l during oral glucose tolerance testing), and (3) use of an oral hypoglycemic drug or insulin. 2.4. Renal function evaluation Serum creatinine was measured by Jaffé chromogen method and using the URICASE/POD (Boehringer Mannheim, Mannheim, Germany) method implemented in an auto-analyzer. Renal function was evaluated by estimated-GFR (e-GFR) (ml/min−1/1.73 m2), calculated with the equation proposed by investigators in the Chronic Kidney Disease Epidemiology (CKD-EPI) Collaboration [22]. We preferred this equation because it is more accurate in subjects with GFR N60 ml/min−1/1.73 m2. Thus, we considered this value as the cut-off of normalcy. 2.5. Echocardiograms Tracings were taken with the patient in partial left decubitus position, using a VIVID 7 Pro ultrasound machine (GE Technologies, Milwaukee, Wisconsin, USA) with an annular phased-array 2.5-MHz transducer. Having the same experienced sonographer perform all studies in a dimly lit and quiet room optimized the reproducibility of measurements. The measurement of LA volume was performed using the area–length (L) method. We measured single-plane area of the LA from the four-chamber view, at end-ventricular systole, guaranteeing that there was no foreshortening of the LA [23]. The LA area was planimetered with the inferior border defined as the plane of the mitral annulus, excluding the confluence of the pulmonary veins and the LA appendage. Length was measured from back wall to line across hinge points of mitral valve. According to this method, LA volume was calculated by the following formula, assuming that A 1 = A 2 : LA volume = 8(A 1 ) 2 /3π(L). LAVI was obtained indexing LA volume by surface area. Partition values for LAVI were taken with the cut-off value of 28 ml/m2 [24]. Similarly, trans-mitral flow velocities were recorded from the apical four-chamber view using pulsed-Doppler. Peak early (E) and late (A) mitral inflow velocities and E/A ratio were measured at end-expiration. Measurements of interventricular septum thickness, posterior wall thickness and left ventricular internal diameter were made at end-diastole and end-systole, as recommended by the American Society of Echocardiography [25]. LVM was
calculated using the Devereux formula [26] and normalized by body surface area (LVMI). Partition values for LVH were taken with the cut-off value of 110 g/m 2 for women and 125 g/m2 for men [27]. 2.6. Follow-up and incident AF Patients were evaluated every 6 months, performing physical examination, standard 12-lead electrocardiogram and routine laboratory analyses. Follow-up was improved mailing a questionnaire to family physicians and contacting patients by phone. In addition, all subjects were asked about previous hospitalizations, discharge summaries and diagnoses. AF diagnosis was made by standard electrocardiogram, hospital discharge diagnoses, and the all-clinical documentation provided by the patients or presents in the general practitioner files. In addition, to detect asymptomatic AF we annually performed a 24-h ECG-Holter. All electrocardiograms were read and approved for the confirmation of AF by an expert cardiologists' team (F.P., A.S.). In this study we did not discriminate the AF type. Additionally, patients developing AF during AMI, congestive heart failure or another acute cardiovascular event were not included in this analysis. 2.7. Statistical analysis Frequencies and percentages for categorical data, and mean ± SD for continuous variables summarized baseline characteristics. To test differences between clinical and biological data, we used the unpaired Student's t-test and the chi-square test for continuous and categorical variables, respectively. Event rate is reported as the number of events/100 patients-year based on the ratio of the number of events observed to the total number of patient-years of exposure up to the terminating event or censor. For patients without events, the date of censor was that of the last contact. For the patients who experienced multiple events, survival analysis was restricted to the first event. The effect of prognostic factors on survival was evaluated by using a multivariable Cox regression model. We tested the impact of the following covariates: age, gender, smoking, BMI, diabetes, hypertension, hypercholesterolemia, e-GFR, LVMI, and LAVI as continuous or dichotomic values (absent/present) on incident AF. For categorical variables, proportional hazards were assessed both by visual inspection and by the log–log method. For continuous variables, proportional risk assumption was tested relating the Schoenfeld residuals of the Cox analysis with survival time. By these approaches we found no violation of proportional hazards (P b 0.50). The multiple Cox regression model was constructed by including all variables that resulted to be associated with the incident risk of AF (P = 0.05) at univariate Cox regression analysis. Data are expressed as hazard ratio (HR), 95% confidence interval (CI) and P value. The interaction (synergism) between increased LAVI, CKD and LVH was defined as a deviation from additivity occurring when the observed HR for the study outcome of patients with increased LAVI, CKD and LVH was higher than that expected by summing up the HR of those with increased LAVI or isolated CKD or isolated LVH minus one [28]. The additional prognostic value, beyond and above that provided by standard risk factors, of cardiac and renal damage for predicting AF was investigated by sequentially adding these biomarkers into the same Cox model (basic model) including age, BMI and LDL-cholesterol. We then tested by − 2 log likelihood statistics [29] whether the sequential addition of these three biomarkers provided significant prognostic information to the basic model. The explained variation in the incidence rate of AF provided by Cox models of increasing complexity (Nested Models) was calculated by −2 log likelihood statistics and compared by chi-square test [30].
3. Results During the follow-up (mean 41.3 ± 21.1 months, range 20–150), we documented 546 new cases of AF (4.5 events/100 patients-year). Baseline characteristics of patients who progressed toward AF (progressors) and those remaining free of arrhythmia (non-progressors) are reported in Table 1. There were no statistically significant differences between groups in gender, smoking, diastolic BP, HDLcholesterol and obesity. On the contrary, progressors are older, have a higher BMI, systolic and pulse pressure, total and LDLcholesterol, fasting glucose, and a lower e-GFR; EF showed a very small, even though statistically significant, difference. In addition, hypertension, MS, diabetes, and CKD were more common among AF cases; while, they show a lower prevalence of hypercholesterolemia. LVMI and LVH, LAVI and diastolic dysfunction were significantly increased in progressors than the control group. In Fig. 1 we graphically report the mean values of e-GFR, LVMI, E/A ratio and LAVI in both progressors and non-progressors to AF. After the first eligibility visit, all subjects were treated according to current guidelines to control all CV risk factors with both
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259
A. Sciacqua et al. / International Journal of Cardiology xxx (2014) xxx–xxx
3
Table 1 Baseline characteristics of the study population stratified as progressors and non-progressors to atrial fibrillation. Variables
All (n = 3549)
Non-progressors (n = 3003)
Progressors (n = 546)
P
Gender, M/F Age, years BMI, kg/m2 Smoking, n (%) Systolic BP, mm Hg Diastolic BP, mm Hg Pulse pressure, mm Hg Total cholesterol, mg/dl LDL-cholesterol, mg/dl HDL-cholesterol, mg/dl Fasting glucose, mg/dl e-GFR, ml/min/1.73 m2 LVMI, g/m2 EF, % LAVI, ml/m2 E/A Hypertension, n (%) Hypercholesterolemia, n (%) Metabolic syndrome, n (%) Diabetes, n (%) Obesity, n (%) CKD, n (%) LVH, n (%)
1859/1690 60.7 ± 10.6 29.1 ± 4.9 1399 (39.4) 142.1 ± 18.1 84.4 ± 11.1 57.6 ± 15.3 201.9 ± 36.1 125.8 ± 34.1 48.6 ± 12.9 113.1 ± 44.4 88.1 ± 26.4 126.1 ± 34.9 65.9 ± 10.5 23.5 ± 10.9 0.88 ± 0.2 2409 (67.9) 1697 (47.8) 1942 (54.7) 582 (16.4) 1110 (31.3) 451 (12.7) 1758 (49.5)
1560/1443 59.2 ± 9.6 28.9 ± 4.1 1191 (39.6) 141.1 ± 18.3 84.3 ± 11.1 56.7 ± 15.0 197.1 ± 28.5 125.2 ± 33.1 48.7 ± 13.2 112.1 ± 45.2 90.6 ± 25.4 121.2 ± 30.9 65.8 ± 10.1 21.2 ± 8.2 0.91 ± 0.20 1970 (65.6) 1523 (50.7) 1606 (53.4) 455 (15.1) 945 (31.4) 269 (8.96) 1339 (44.6)
299/247 68.8 ± 11.9 30.1 ± 7.9 208 (38.1) 147.1 +/- 14.8 85.1 +/- 10.7 62.0 ± 16.3 202.7 ± 37.2 128.9 ± 38.9 48.1 ± 10.6 118.5 ± 39.5 73.9 ± 26.9 152.8 ± 42.3 64.1 ± 12.2 36.4 ± 14.6 0.75 +/- 0.21 439 (80.4) 174 (31.8) 336 (61.5) 127 (23.2) 165 (29.7) 182 (33.3) 419 (76.7)
0.244 b0.0001 b0.0001 0.522 b0.0001 0.119 b0.0001 b0.0001 0.020 0.315 0.002 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 0.597 b0.0001 b0.0001
BMI = body mass index; e-GFR = estimated glomerular filtration rate; LVMI = left ventricular mass index; EF = ejection fraction; LAVI = left atrial volume index; CKD = chronic kidney disease; LVH = left ventricular hypertrophy.
pharmacological and lifestyle interventions, without any significant difference between groups.
3.1. Cox regression analyses for AF On univariate analysis, incident AF was directly related to LVMI (10 g/m2) (HR = 1.030, 95% CI = 1.011–1.049), LDL-cholesterol (10 mg/dl) (HR = 1.042, 95% CI = 1.018–1.065), LAVI (1 ml/m 2 ) (HR = 1.048, 95% CI = 1.042–1.053), and age (10 years) (HR = 1.575, 95% CI = 1.449–1.713), and inversely related to e-GFR (10 ml/min/ 1.73 m2) (HR = 0.940, 95% CI = 0.903–0.977) (Table 2). No association was found between occurrence of AF and fasting glucose (P = 0.054),
smoking (P = 0.089), systolic BP (P = 0.812), gender (P = 0.067), and BMI (P = 0.162) and E/A ratio (P = 0.340). In the multiple Cox regression analysis (model 1) (Table 3), traditional cardiovascular risk factors, and e-GFR, LVMI, and LAVI, as continuous variables, were retained as independent predictors of incident AF: LVMI (10 g/m2) (HR = 1.030, 95% CI = 1.011–1.050), LDL-cholesterol (10 mg/dl) (HR = 1.039, 95% CI = 1.016–1.063), LAVI (1 ml/m2) (HR = 1.048, 95% CI = 1.042–1.053), age (10 years) (HR = 1.575, 95% CI = 1.449–1.713), and e-GFR (10 ml/min/1.73 m2) (HR = 0.940, 95% CI = 0.903–0.977). In the model 2, in which we added hypercholesterolemia, hypertension, diabetes, obesity, e-GFR, LVH and LAVI (N28 ml/m2), as dichotomic variables, a significant difference in HRs for incident AF was found
Fig. 1. Subclinical organ damage and atrial fibrillation. In this picture we graphically report the mean values of estimated glomerular filtration rate (e-GFR), left ventricular mass index (LVMI), E/A ratio and left atrial area (LAA) in both progressors and non-progressors to AF. As evident, patients developing AF have a higher LVMI and LAA, and a lower e-GFR and E/A ratio.
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259
4
A. Sciacqua et al. / International Journal of Cardiology xxx (2014) xxx–xxx
additive effect of these three risk factors and that such a synergic effect was fully independent of potential confounders.
Table 2 Univariate Cox regression analysis for atrial fibrillation occurrence.
LVMI, 10 g/m2 LDL-cholesterol, 10 mg/dl LAVI, ml/m2 Age, 10 years e-GFR, 10 ml/min/1.73 m2
Hazard ratio
95% CI
1.030 1.042 1.048 1.575 0.940
1.011–1.049 1.018–1.065 1.042–1.053 1.449–1.713 0.903–0.977
LVMI = left ventricular mass index; LAVI = Left atrial volume index; e-GFR = estimated glomerular filtration rate.
for LVH (HR = 1.495, 95% CI = 1.215–1.841), CKD (HR = 1.528, 95% CI = 1.261–1.851), age (10 years) (HR = 1.586, 95% CI = 1.461–1.725) and LAVI (HR = 2.920, 95% CI = 2.426–3.515).
3.3. Explained variation in the incidence rate of AF attributable to biomarkers of organ damage A Cox model including age, BMI and LDL-cholesterol (basic model) provided a 17% explained variation in the incidence rate of AF. By adding LVH to the basic model, the explained variation in the study outcome rose from 17% to 20% and such an increase in prognostic value (+3%) was highly significant (P b 0.001). Remarkably, the addition of LAVI into the model increased the prognostic value and the explained variation in AF rose from 20% to 25% (+5%, P b 0.001). The addition of also CKD to the model increased only by 1% its prognostic performance (from 25% to 26%).
3.2. Interaction analysis
4. Discussion
According to crude analysis (see Fig. 2, upper panel), all risk categories, but particularly that of increased LAVI, were strongly and significantly associated to the risk of AF; notably, patients having an abnormal LAVI, in the absence of reduced e-GFR and LVH, had a relative risk of AF that was about 5 times higher than that of patients in the reference category (HR = 4.767, 95% CI = 3.246–7.001). The interaction analysis revealed that the observed HRs of each risk categories combination were higher than those expected in the absence of interaction under the additive model indicating that the simultaneous occurrence of the two or three risk factors under investigation (increased LAVI, LVH and reduced e-GFR) provided a risk excess for AF higher than that of the simple additive effect of the three risk factors. A multivariate Cox regression analysis, adjusting for a series of potential confounders (age, BMI and LDL-cholesterol) (see Fig. 2, lower panel), confirmed an interaction between increased LAVI and reduced e-GFR for explaining the incidence rate of AF. The interaction between abnormal LAVI and LVH, even if statistically significant, after multivariate data adjustment reduces its strength, suggesting that deranged LAVI and altered LVM are in the same causal pathway conducive to AF in the study population. Remarkably, although reduced e-GFR alone was not significantly associated to the study outcome, it interacted with LVH for amplifying the risk of AF, the observed risk for this outcome associated to these two risk factors being almost double than that expected in the absence of interaction under the additive model. Finally, also on multivariate analysis, a strong interaction was found among abnormal LAVI, reduced e-GFR and LVH for explaining the incidence rate of the study outcome indicating that in patients affected by simultaneous alterations in LA size, LVMI and e-GFR, the risk of AF was consistently higher than the simple
Present data clearly demonstrate that LAVI, CKD and LVH predict the incident AF. The most important novelty of this paper is that we observed, for the first time, that renal dysfunction and LAVI show a synergic effect in predicting the cardiac arrhythmia, totally independent of potential confounders such as traditional cardiovascular risk factors. This finding is consistent with statistical analysis because the observed HR (4.875, 95% CI 2.699–8.805) for incident AF was significantly higher than the expected HR (4.510). Another important finding obtained in this study is the fact that the interaction analysis demonstrated that observed risk for incident AF, associated to LVH and CKD, was almost double (HR = 4.040, 95% CI = 2.661–6.133) (Fig. 2) than that expected in the absence of interaction under the additive model (HR = 2.098). This is not surprising because both CKD and LVM represent wellestablished organ damages reflecting and integrating the long-term cumulative level of activity of several risk factors for cardiovascular disease. In keeping with this, it is interesting that also serum creatinine values, within the normal range, may work as a marker of long-term exposure to elevated BP levels, and therefore are able to predict outcome also in nonrenal target organs [4,9]. The higher rate of incident AF in our population is apparently different from that observed in other studies [31,32]. In particular, our subjects were about 10 years older than those enrolled by Verdecchia et al. [31]; this has a prognostic significance because the Framingham Heart Study demonstrated that the risk of AF doubles for each decade of advancing age, both in men and women [33]. On the other hand, our patients have a higher prevalence of LVH compared to the ONTARGET population (49.5% vs 12.7%). Nevertheless, the ONTARGET study was not designed to detect new onset of AF, and criteria used by Verdecchia et al. for AF detection were deeply different by ours, because they evaluated AF appearance only to ECG performed at follow-up visit. Finally, this prevalence may be justified in part by the fact that we have carefully checked the new AF appearance. Another relevant finding emerged from this study is that an abnormal LA size, in the absence of CKD and LVH, has a relative risk of AF that is about 5 times higher than that of patients in the reference category (HR = 4.767, 95% CI = 3.246–7.001) (Fig. 2), confirming the most relevant role of LA enlargement in the pathogenesis of AF. At first, the LA structural changes reflect the chronicity of exposure to abnormal filling pressures as consequence of diastolic dysfunction [34], similarly to that observed in patients with LVH. According with this, our patients developing AF have a higher LVMI and LA dimension. Next, there are several evidences demonstrating that structural remodeling and atrial dilatation are due to interstitial fibrosis that cause conduction delays between cardiomyocytes favoring the electrical ectopic activity and anisotropic conduction generating non-uniform wave fronts and re-entrant arrhythmias [35]. Notably, results of adjusted interaction analysis showed that the interaction between LA dimension and LVH, even if statistically significant, after multivariate data
Table 3 Multivariate Cox regression analysis for atrial fibrillation occurrence.
Model 1 LVMI LDL-cholesterol LAVI Age e-GFR Model 2 LVH CKD Age LAVI
Increase
HR
95% CI
10 g/m2 10 mg/dl 1 ml/m2 10 years 10 ml/min/1.73 m2
1.030 1.039 1.048 1.575 0.940
1.011–1.050 1.016–1.063 1.042–1.053 1.449–1.713 0.903–0.977
Yes/No Yes/No 10 years Yes/No
1.495 1.528 1.586 2.920
1.215–1.841 1.261–1.851 1.461–1.725 2.426–3.515
Model 1 = smoking (yes/no), fasting glucose, LDL-cholesterol, age, gender (male/female), systolic blood pressure, body mass index, left ventricular mass index (LVMI), E/A ratio, left atrial volume index (LAVI), estimated glomerular filtration rate (e-GFR). Model 2 = smoking (yes/no), age, hypercholesterolemia (yes/no), gender (male/female), hypertension (yes/no), diabetes (yes/no), obesity (yes/no), metabolic syndrome (yes/no), left ventricular hypertrophy (LVH) (yes/no), LAVI increased as ≥28 ml/m2 (yes/no), renal dysfunction as e-GFR b 60 ml/min/1.73 m2 (yes/no).
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259
A. Sciacqua et al. / International Journal of Cardiology xxx (2014) xxx–xxx
5
Fig. 2. Synergic effect of subclinical organ damage on incident AF. Interaction analysis results as crude (upper) and adjusted (bottom) hazard ratios (HR) for incident AF are graphically reported. Expected HRs are represented by white bars, while the observed HRs are represented by the sum of with and dashed bars.
adjustment reduces its biological significance, suggesting that LA enlargement and increased LVM are in the same causal pathway conducive to AF appearance. Present data also demonstrate that CKD was retained as a second independent predictor of AF in the final model of multivariate Cox regression analysis, accounting for 52.8% of the increased risk associated with new appearance of AF. Circulating volume expansion due to sodium retention, activation of RAAS and sympathetic overactivity are important causative mechanism of LA enlargement and AF development [3]. Notably, also small decrements in e-GFR impair regulation of extracellular fluid volume, resulting in impaired renal sodium excretion and increased RAAS activity, conditions associated with subclinical inflammation [8] and cardiovascular events [3,36]. Finally, also the sympathetic activation RAAS-related has been proposed to atrial arrhythmogenesis and AF development [10]. Regarding the association between increased LVM and incident AF, there are strengthened data consistent with pathophysiological mechanisms linking LVH with atrial arrhythmia. In fact, LA size represents a marker of both left ventricular stiffness and diastolic dysfunction in hypertensive, diabetic and elderly subjects. In keeping with this, LA size increases with worsening diastolic dysfunction [34], independently of traditional cardiovascular risk factors and independently of LV ejection fraction. Besides, the final model of multivariate analysis revealed a strong interaction among abnormal LAVI, CKD and LVH, indicating that in patients showing all three alterations in LA size, LVM and renal function, the risk of AF was consistently higher than the simple additive effect of these three organ damages. Interestingly, LAVI has a predictive value higher than that of LVH increasing of 5% the risk of new AF explained by basic model, while LVH adds only a 3%. In keeping with this, these results suggest that interventions simultaneously focused on the treatment and/or prevention of all three subclinical organ damages could have beneficial effects for reducing the risk of AF beyond that expected by the simple sum of the single beneficial effects. Thus, it should be convenient to utilize drugs able to modulate the RAAS, not only for reducing BP values, but also to interfere with the appearance and progression of subclinical organ damage; particularly, the effect of these drugs is more evident on left atrium structure in which they have demonstrated to reduce fibrosis and atrial remodeling [19]. According with this, a recent meta-analysis demonstrated that RAAS inhibition may have an important role in the primary and secondary prevention of AF [37].
Finally, in our study, subclinical organ damages were emerged as independent predictors of incident AF, interacting too, in a synergic manner, in increasing the arrhythmic risk. This emphasizes that our findings are more important from a pathophysiological point of view than from the perspective of risk prediction. On the basis of these results it is possible to suggest that AF may be considered a time-integrated marker of exposure to multiple cardiovascular risk factors, particularly CKD indicating that its occurrence, simultaneously to the other two subclinical disease provides a risk excess for AF higher than that of the simple additive effect. Thus, strategies for the prevention of AF will have to recognize also CKD as a preventable risk factor for the development of AF in addition to other well-established risk factors. In keeping with this, the diagnosis of CKD should alert the practitioner to routinely assess also renal function to correctly define the global AF risk. However, the evaluation of this risk is often not optimized because screening for CKD is frequently limited to a measurement of serum creatinine, which does not accurately reflect GFR, the best indicator of renal function in healthy and diseased subjects. 4.1. Limitations This study has some limitations. At first, the type of AF was not discriminated in this study. Another limitation consists in the fact that e-GFR was not directly measured, but estimated by creatinine value. In conclusion, in spite of the above limitations, data obtained from this study clearly demonstrate that LAVI and CKD significantly interact in a synergic manner in increasing AF risk. This evidence encourages to detect routinely the renal function in all patients that are at increased risk to develop AF, allowing thus a better stratification of AF risk. References [1] National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification and stratification. Am J Kidney Dis 2002;39: S1–S266. [2] Sarnak MJ, Levey AS, Schoolwerth AC, et al. American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention. Kidney disease as a risk factor for development of cardiovascular disease. A statement from the American Heart Association councils on kidney in cardiovascular disease, high blood pressure research, clinical cardiology, and epidemiology and prevention. Hypertension 2011;42:1050–65. [3] 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;351:1296–305.
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259
6
A. Sciacqua et al. / International Journal of Cardiology xxx (2014) xxx–xxx
[4] Schillaci G, Reboldi G, Verdecchia P. High-normal serum creatinine concentration is a predictor of cardiovascular risk in essential hypertension. Arch Intern Med 2001;161:886–91. [5] Jensen JS, Feldt-Rasmussen B, Strandgaard S, Schroll M, Borch-Johensen K. Arterial hypertension, microalbuminuria, and risk of ischemic heart disease. Hypertension 2000;35:898–903. [6] Meguid El NA, Bello AK. Chronic kidney disease. The global challenge. Lancet 2005;365:331–40. [7] Sciacqua A, Scozzafava A, Pujia A, et al. Interaction between vascular dysfunction and cardiac mass increases the risk of cardiovascular outcomes in essential hypertension. Eur Heart J 2005;26:921–7. [8] Zoccali C, Maio R, Tripepi G, Mallamaci F, Perticone F. Inflammation as a mediator of the link between mild to moderate renal insufficiency and endothelial dysfunction in essential hypertension. J Am Soc Nephrol 2006;17:S64–8. [9] Miceli S, Maio R, Perticone M, Mallamaci F, Perticone F. Creatinine and insulin predict cardiac mass in drug-naive hypertensive patients. Int J Cardiol 2013;167:519–24. [10] Schlaich MP, Socratous F, Hennebry S, et al. Sympathetic activation in chronic renal failure. J Am Soc Nephrol 2009;20:933–9. [11] Perticone F, Maio R, Cosco C, et al. Hypertensive left ventricular remodeling and ACE-gene polymorphism. Cardiovasc Res 1999;43:192–9. [12] Naccarelli GV, Johnston SS, Lin J, Patel PP, Schulman KL. Cost burden of cardiovascular hospitalization and mortality in ATHENA-like patients with atrial fibrillation/atrial flutter in the United States. Clin Cardiol 2010;33:270–9. [13] McManus DD, Corteville DCM, Shlipak MG, Wholeey MA, Ix JH. Relation of kidney function and albuminuria with atrial fibrillation (from the Heart and Soul Study). Am J Cardiol 2009;104:1551–5. [14] Baber U, Howard VJ, Halperin JL, Wholeey MA, Ix JH. Association of chronic kidney disease with atrial fibrillation among adults in the United States: REasons for Geographic and Racial Differences in Stroke (REGARDS) Study. Circ Arrhythmia Electrophysiol 2011;4:26–32. [15] Perticone F, Maio R, Ruberto C, et al. Kidney function and risk factors for left ventricular hypertrophy in untreated uncomplicated essential hypertension. Am J Kidney Dis 2008;52:74–84. [16] Sciacqua A, Miceli S, Carullo G, et al. One-hour postload plasma glucose levels and left ventricular mass in hypertensive patients. Diabetes Care 2011;34:1406–11. [17] Sesti G, Sciacqua A, Scozzafava A, et al. Effects of growth hormone and insulin-like growth factor-1 on cardiac hypertrophy of hypertensive patients. J Hypertens 2007;25:471–7. [18] Lorell BH, Carabello BA. Left ventricular hypertrophy: pathogenesis, detection, and prognosis. Circulation 2000;102:470–9. [19] Wachtell K, Gerdts E, Aurigemma GP, et al. In-treatment reduced left atrial diameter during antihypertensive treatment is associated with reduced new-onset atrial fibrillation in hypertensive patients with left ventricular hypertrophy: The LIFE Study. Blood Press 2010;19:169–75. [20] Eckstein J, Verheule S, de Groot NM, Allessie M, Schotten U. Mechanisms of perpetuation of atrial fibrillation in chronically dilated atria. Prog Biophys Mol Biol 2008;97:435–51.
[21] Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of the National Cholesterol Education Program (NCEP) Expert Panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA 2001;285:2486–97. [22] Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150:604–12. [23] Vieira-Filho NG, Mancuso FJ, Oliveira WA, et al. Simplified single plane echocardiography is comparable to conventional biplane two-dimensional echocardiography in the evaluation of left atrial volume: a study validated by three-dimensional echocardiography in 143 individuals. Echocardiography 2014;31:265–72. [24] Lang RM, Bierig M, Devereux RB, et al. Recommendations for chamber quantification. Eur J Echocardiogr 2006;7:79–108. [25] Cheitlin MD, Armstrong WF, Aurigemma GP, et al. ACC/AHA/ASE 2003 guideline update for the clinical application of echocardiography: summary article: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/ASE Committee to Update the 1997 Guidelines for the Clinical Application of Echocardiography). Circulation 2003;108:1146–62. [26] Devereux RB, Alonso DR, Lutas EM, et al. Echocardiographic assessment of left ventricular hypertrophy comparison to necropsy findings. Am J Cardiol 1986;57:450–8. [27] Marcus R, Krause L, Weder AB, et al. Sex-specific determinants of increased left ventricular mass in the Tecumseh Blood Pressure Study. Circulation 1994;90:928–36. [28] de Mutsert R, Jaeger KJ, Zoccali C, Dekker FW. The effect of joint exposures: examining the presence of interaction. Kidney Int 2009;75:677–81. [29] Collet D. Modelling survival data in medical research. London: Chapman and Hall; 1994 72–85. [30] Hosmer DW, Lemeshow S. Assessment of model adequacy. Applied survival analysis. New York: Wiley & Sons; 1999. p. 196–240. [31] Verdecchia P, Reboldi G, Gattobigio R, et al. Atrial fibrillation in hypertension: predictors and outcome. Hypertension 2003;41:218–23. [32] ONTARGET InvestigatorsYusuf S, Teo KK, et al. Telmisartan, ramipril, or both in patients at high risk for vascular events. N Engl J Med 2008;358:1547–59. [33] Benjamin EJ, Levy D, Vaziri SM, et al. Independent risk factors for atrial fibrillation in a population-based cohort. The Framingham Heart Study. JAMA 1994;271:840–4. [34] Tsang TS, Barnes ME, Gersh BJ, Bailey KR, Sewatd JB. Left atrial volume as a morphophysiologic expression of left ventricular diastolic dysfunction and relation to cardiovascular risk burden. Am J Cardiol 2002;90:1284–9. [35] Zhao W, Zhao T, Chen Y, Ahokas RA, Sun Y. Oxidative stress mediates cardiac fibrosis by enhancing transforming growth factor-beta1 in hypertensive rats. Mol Cell Biochem 2008;317:43–50. [36] Manjunath G, Tighiouart H, Ibrahim H, et al. Level of kidney function as a risk factor for atherosclerotic cardiovascular outcomes in the community. J Am Coll Cardiol 2003;41:47–55. [37] Schneider MP, Hua TA, Böhm M, et al. Prevention of atrial fibrillation by reninangiotensin system inhibition a meta-analysis. J Am Coll Cardiol 2010;55:2299–307.
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors Angela Sciacqua a,1, Maria Perticone c,1, Giovanni Tripepi b, Sofia Miceli a, Eliezer J. Tassone a, Nadia Grillo a, Giuseppe Carullo a, Giorgio Sesti a, Francesco Perticone a,⁎ a b c
Department of Medical and Surgical Sciences, University Magna Græcia of Catanzaro, Italy CNR, Istituto di Biomedicina, Clinical Epidemiology and Physiopathology of Renal Diseases and Hypertension, Reggio Calabria, Italy Experimental and Clinical Medicine Experimental and Clinical Medicine
a r t i c l e
i n f o
Article history: Received 18 February 2014 Received in revised form 24 April 2014 Accepted 26 April 2014 Available online xxxx Keywords: Atrial fibrillation Chronic kidney disease Atrial volume Left ventricular hypertrophy
a b s t r a c t Objectives: In this prospective population-based study, we tested the possible interaction between chronic kidney disease (CKD) and left atrium volume index (LAVI) in predicting incident atrial fibrillation (AF). Methods: We enrolled 3549 Caucasian subjects, 1829 men and 1720 women, aged 60.7 ± 10.6 years, without baseline AF and thyroid disorders. Echocardiographic left ventricular mass and LAVI were measured. Renal function was calculated by estimated glomerular filtration rate (e-GFR). To test the effect of some clinical confounders on incident AF, we constructed different models including clinical and laboratory parameters. AF diagnosis was made by standard electrocardiogram or 24-h ECG-Holter, hospital discharge diagnoses, and by the all-clinical documentation. Results: During the follow-up (53.3 ± 18.1 months), 546 subjects developed AF (4.5 events/100 patient-years). Progressors to AF were older, had a higher body mass index, blood pressure, LDL-cholesterol, glucose, cardiac mass, and LAVI, and had lower e-GFR. Hypertension, metabolic syndrome, diabetes, cardiac hypertrophy and CKD were more common among AF cases than controls. In the final Cox regression model, variables that remained significantly associated with AF were: cardiac hypertrophy (HR = 1.495, 95% CI = 1.215–1.841), renal disease (HR = 1.528, 95% CI = 1.261–1.851), age (HR = 1.586, 95% CI = 1.461–1.725) and LAVI (HR = 2.920, 95% CI = 2.426–3.515). The interaction analysis demonstrated a synergic effect between CKD and cardiac hypertrophy (HR = 4.040, 95% CI = 2.661–6.133), as well as between CKD and LAVI (HR = 4.875, 95% CI = 2.699–8.805). The coexistence of all three subclinical organ damages significantly increases the arrhythmic risk (HR = 7.185, 95% CI = 5.041–10.240). Conclusions: Our data demonstrate that LAVI and CKD significantly interact in a synergic manner in increasing AF risk. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Chronic kidney disease (CKD) is a relevant health problem in the general population and in other conditions [1] because it is strongly associated with high cardiovascular morbidity and mortality [2,3]. Notably, some evidences demonstrated that mild renal dysfunction, defined as a small increase in serum creatinine (b 2 mg/dl), abnormal urinary albumin excretion and/or a reduction of glomerular filtration rate (GFR) [4,5], is a powerful predictor of cardiovascular events in hypertensives. Probably, hemodynamic and atherogenic factors, such as insulin resistance, endothelial dysfunction and subclinical inflammation
⁎ Corresponding author at: Department of Medical and Surgical Sciences, Campus Universitario di Germaneto, V.le Europa, 88100 Catanzaro, Italy. Tel.: +39 0961 3647149; fax: +39 0961 3647634. E-mail address: [email protected] (F. Perticone). 1 These authors have equally contributed to the work.
may contribute to explain these evidences [3,6–9]. In addition, CKD is characterized by volume expansion due to sodium retention, activation of the renin–angiotensin–aldosterone system (RAAS) and sympathetic overactivity that promote ventricular and atrial remodeling [10,11], favoring the development of atrial fibrillation (AF). In adult and old subjects, AF is the most frequent cardiac rhythm disorder, impacting on the economic costs of caring because it is associated with an increased risk for thromboembolic events [12]. In addition to the different clinical conditions—such as valvular disease, congestive heart failure and myocardial infarction—recent and limited data showed that also CKD is associated with incident AF [13,14]. Usually, cardiac structural remodeling represents an adaptive response of the myocardium to increased cardiac workload in high blood pressure (BP), renal disease [9,15] and other clinical conditions [16,17], even if the increased left ventricular mass (LVM) is recognized as an independent predictor of cardiovascular events in general population and other clinical settings [18]. In addition, it is associated
http://dx.doi.org/10.1016/j.ijcard.2014.04.259 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259
2
A. Sciacqua et al. / International Journal of Cardiology xxx (2014) xxx–xxx
with diastolic dysfunction and left atrium (LA) enlargement that represents an important pathophysiological mechanism involved in incident AF [19] that, in turn, induces anatomical and electrical atrial remodeling and, thus, the maintenance of itself [20]. Taken together, at this moment there are no data about a possible interaction between CKD and LA volume index (LAVI) in predicting new AF. So far, we tested this hypothesis in a large cohort of Caucasian subjects. 2. Materials and methods 2.1. Study population Our study is a population-based prospective study aimed at assessing the possible interaction of LAVI and renal dysfunction on incident AF. Between January 1998 and October 2011 we enrolled 3549 Caucasian outpatients (1829 men, 1720 women; aged 60.7 ± 10.6 years). The study population consists of outpatients with metabolic and cardiovascular risk factors referred to our tertiary care setting for screening program. All patients were prospectively followed to evaluate the development of incident diabetes, AF and fatal and non-fatal cardiovascular events. All patients underwent physical examination, review of their medical history and anthropometrical evaluation: weight, height, and body mass index (BMI). We excluded patients with thyroid disorders or those taking drugs affecting thyroid function, rheumatic and non-rheumatic valvular heart disease, prosthetic valves, congestive heart failure, cardiomyopathies, dialyzed patients and malignant disease. At the time of enrolment none of the patients have experienced myocardial infarction or stroke. Metabolic syndrome (MS) was defined according to NCEP-ATP III criteria [21]. The ethical Committee of University Hospital of Catanzaro approved the protocol and informed written consent was obtained from all participants. All the investigations were performed according with the principles of Helsinki Declaration. 2.2. Blood pressure measurements Patients taking antihypertensive drugs or with a systolic BP (SBP) ≥ 140 mm Hg and/or diastolic BP (DPB) ≥ 90 mm Hg were defined as hypertensive. Secondary forms of hypertension were excluded by a standard clinical protocol. 2.3. Diabetes evaluation Criteria for type 2 diabetes were: (1) presence of more than one classic symptom of hyperglycemia plus either a fasting plasma glucose ≥ 7.0 mmol/l or random plasma glucose ≥11.1 mmol/l, (2) two or more elevated plasma glucose concentrations (fasting plasma glucose ≥7.0 mmol/l, random plasma glucose ≥11.1 mmol/l, or 2-h plasma glucose ≥11.1 mmol/l during oral glucose tolerance testing), and (3) use of an oral hypoglycemic drug or insulin. 2.4. Renal function evaluation Serum creatinine was measured by Jaffé chromogen method and using the URICASE/POD (Boehringer Mannheim, Mannheim, Germany) method implemented in an auto-analyzer. Renal function was evaluated by estimated-GFR (e-GFR) (ml/min−1/1.73 m2), calculated with the equation proposed by investigators in the Chronic Kidney Disease Epidemiology (CKD-EPI) Collaboration [22]. We preferred this equation because it is more accurate in subjects with GFR N60 ml/min−1/1.73 m2. Thus, we considered this value as the cut-off of normalcy. 2.5. Echocardiograms Tracings were taken with the patient in partial left decubitus position, using a VIVID 7 Pro ultrasound machine (GE Technologies, Milwaukee, Wisconsin, USA) with an annular phased-array 2.5-MHz transducer. Having the same experienced sonographer perform all studies in a dimly lit and quiet room optimized the reproducibility of measurements. The measurement of LA volume was performed using the area–length (L) method. We measured single-plane area of the LA from the four-chamber view, at end-ventricular systole, guaranteeing that there was no foreshortening of the LA [23]. The LA area was planimetered with the inferior border defined as the plane of the mitral annulus, excluding the confluence of the pulmonary veins and the LA appendage. Length was measured from back wall to line across hinge points of mitral valve. According to this method, LA volume was calculated by the following formula, assuming that A 1 = A 2 : LA volume = 8(A 1 ) 2 /3π(L). LAVI was obtained indexing LA volume by surface area. Partition values for LAVI were taken with the cut-off value of 28 ml/m2 [24]. Similarly, trans-mitral flow velocities were recorded from the apical four-chamber view using pulsed-Doppler. Peak early (E) and late (A) mitral inflow velocities and E/A ratio were measured at end-expiration. Measurements of interventricular septum thickness, posterior wall thickness and left ventricular internal diameter were made at end-diastole and end-systole, as recommended by the American Society of Echocardiography [25]. LVM was
calculated using the Devereux formula [26] and normalized by body surface area (LVMI). Partition values for LVH were taken with the cut-off value of 110 g/m 2 for women and 125 g/m2 for men [27]. 2.6. Follow-up and incident AF Patients were evaluated every 6 months, performing physical examination, standard 12-lead electrocardiogram and routine laboratory analyses. Follow-up was improved mailing a questionnaire to family physicians and contacting patients by phone. In addition, all subjects were asked about previous hospitalizations, discharge summaries and diagnoses. AF diagnosis was made by standard electrocardiogram, hospital discharge diagnoses, and the all-clinical documentation provided by the patients or presents in the general practitioner files. In addition, to detect asymptomatic AF we annually performed a 24-h ECG-Holter. All electrocardiograms were read and approved for the confirmation of AF by an expert cardiologists' team (F.P., A.S.). In this study we did not discriminate the AF type. Additionally, patients developing AF during AMI, congestive heart failure or another acute cardiovascular event were not included in this analysis. 2.7. Statistical analysis Frequencies and percentages for categorical data, and mean ± SD for continuous variables summarized baseline characteristics. To test differences between clinical and biological data, we used the unpaired Student's t-test and the chi-square test for continuous and categorical variables, respectively. Event rate is reported as the number of events/100 patients-year based on the ratio of the number of events observed to the total number of patient-years of exposure up to the terminating event or censor. For patients without events, the date of censor was that of the last contact. For the patients who experienced multiple events, survival analysis was restricted to the first event. The effect of prognostic factors on survival was evaluated by using a multivariable Cox regression model. We tested the impact of the following covariates: age, gender, smoking, BMI, diabetes, hypertension, hypercholesterolemia, e-GFR, LVMI, and LAVI as continuous or dichotomic values (absent/present) on incident AF. For categorical variables, proportional hazards were assessed both by visual inspection and by the log–log method. For continuous variables, proportional risk assumption was tested relating the Schoenfeld residuals of the Cox analysis with survival time. By these approaches we found no violation of proportional hazards (P b 0.50). The multiple Cox regression model was constructed by including all variables that resulted to be associated with the incident risk of AF (P = 0.05) at univariate Cox regression analysis. Data are expressed as hazard ratio (HR), 95% confidence interval (CI) and P value. The interaction (synergism) between increased LAVI, CKD and LVH was defined as a deviation from additivity occurring when the observed HR for the study outcome of patients with increased LAVI, CKD and LVH was higher than that expected by summing up the HR of those with increased LAVI or isolated CKD or isolated LVH minus one [28]. The additional prognostic value, beyond and above that provided by standard risk factors, of cardiac and renal damage for predicting AF was investigated by sequentially adding these biomarkers into the same Cox model (basic model) including age, BMI and LDL-cholesterol. We then tested by − 2 log likelihood statistics [29] whether the sequential addition of these three biomarkers provided significant prognostic information to the basic model. The explained variation in the incidence rate of AF provided by Cox models of increasing complexity (Nested Models) was calculated by −2 log likelihood statistics and compared by chi-square test [30].
3. Results During the follow-up (mean 41.3 ± 21.1 months, range 20–150), we documented 546 new cases of AF (4.5 events/100 patients-year). Baseline characteristics of patients who progressed toward AF (progressors) and those remaining free of arrhythmia (non-progressors) are reported in Table 1. There were no statistically significant differences between groups in gender, smoking, diastolic BP, HDLcholesterol and obesity. On the contrary, progressors are older, have a higher BMI, systolic and pulse pressure, total and LDLcholesterol, fasting glucose, and a lower e-GFR; EF showed a very small, even though statistically significant, difference. In addition, hypertension, MS, diabetes, and CKD were more common among AF cases; while, they show a lower prevalence of hypercholesterolemia. LVMI and LVH, LAVI and diastolic dysfunction were significantly increased in progressors than the control group. In Fig. 1 we graphically report the mean values of e-GFR, LVMI, E/A ratio and LAVI in both progressors and non-progressors to AF. After the first eligibility visit, all subjects were treated according to current guidelines to control all CV risk factors with both
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259
A. Sciacqua et al. / International Journal of Cardiology xxx (2014) xxx–xxx
3
Table 1 Baseline characteristics of the study population stratified as progressors and non-progressors to atrial fibrillation. Variables
All (n = 3549)
Non-progressors (n = 3003)
Progressors (n = 546)
P
Gender, M/F Age, years BMI, kg/m2 Smoking, n (%) Systolic BP, mm Hg Diastolic BP, mm Hg Pulse pressure, mm Hg Total cholesterol, mg/dl LDL-cholesterol, mg/dl HDL-cholesterol, mg/dl Fasting glucose, mg/dl e-GFR, ml/min/1.73 m2 LVMI, g/m2 EF, % LAVI, ml/m2 E/A Hypertension, n (%) Hypercholesterolemia, n (%) Metabolic syndrome, n (%) Diabetes, n (%) Obesity, n (%) CKD, n (%) LVH, n (%)
1859/1690 60.7 ± 10.6 29.1 ± 4.9 1399 (39.4) 142.1 ± 18.1 84.4 ± 11.1 57.6 ± 15.3 201.9 ± 36.1 125.8 ± 34.1 48.6 ± 12.9 113.1 ± 44.4 88.1 ± 26.4 126.1 ± 34.9 65.9 ± 10.5 23.5 ± 10.9 0.88 ± 0.2 2409 (67.9) 1697 (47.8) 1942 (54.7) 582 (16.4) 1110 (31.3) 451 (12.7) 1758 (49.5)
1560/1443 59.2 ± 9.6 28.9 ± 4.1 1191 (39.6) 141.1 ± 18.3 84.3 ± 11.1 56.7 ± 15.0 197.1 ± 28.5 125.2 ± 33.1 48.7 ± 13.2 112.1 ± 45.2 90.6 ± 25.4 121.2 ± 30.9 65.8 ± 10.1 21.2 ± 8.2 0.91 ± 0.20 1970 (65.6) 1523 (50.7) 1606 (53.4) 455 (15.1) 945 (31.4) 269 (8.96) 1339 (44.6)
299/247 68.8 ± 11.9 30.1 ± 7.9 208 (38.1) 147.1 +/- 14.8 85.1 +/- 10.7 62.0 ± 16.3 202.7 ± 37.2 128.9 ± 38.9 48.1 ± 10.6 118.5 ± 39.5 73.9 ± 26.9 152.8 ± 42.3 64.1 ± 12.2 36.4 ± 14.6 0.75 +/- 0.21 439 (80.4) 174 (31.8) 336 (61.5) 127 (23.2) 165 (29.7) 182 (33.3) 419 (76.7)
0.244 b0.0001 b0.0001 0.522 b0.0001 0.119 b0.0001 b0.0001 0.020 0.315 0.002 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 0.597 b0.0001 b0.0001
BMI = body mass index; e-GFR = estimated glomerular filtration rate; LVMI = left ventricular mass index; EF = ejection fraction; LAVI = left atrial volume index; CKD = chronic kidney disease; LVH = left ventricular hypertrophy.
pharmacological and lifestyle interventions, without any significant difference between groups.
3.1. Cox regression analyses for AF On univariate analysis, incident AF was directly related to LVMI (10 g/m2) (HR = 1.030, 95% CI = 1.011–1.049), LDL-cholesterol (10 mg/dl) (HR = 1.042, 95% CI = 1.018–1.065), LAVI (1 ml/m 2 ) (HR = 1.048, 95% CI = 1.042–1.053), and age (10 years) (HR = 1.575, 95% CI = 1.449–1.713), and inversely related to e-GFR (10 ml/min/ 1.73 m2) (HR = 0.940, 95% CI = 0.903–0.977) (Table 2). No association was found between occurrence of AF and fasting glucose (P = 0.054),
smoking (P = 0.089), systolic BP (P = 0.812), gender (P = 0.067), and BMI (P = 0.162) and E/A ratio (P = 0.340). In the multiple Cox regression analysis (model 1) (Table 3), traditional cardiovascular risk factors, and e-GFR, LVMI, and LAVI, as continuous variables, were retained as independent predictors of incident AF: LVMI (10 g/m2) (HR = 1.030, 95% CI = 1.011–1.050), LDL-cholesterol (10 mg/dl) (HR = 1.039, 95% CI = 1.016–1.063), LAVI (1 ml/m2) (HR = 1.048, 95% CI = 1.042–1.053), age (10 years) (HR = 1.575, 95% CI = 1.449–1.713), and e-GFR (10 ml/min/1.73 m2) (HR = 0.940, 95% CI = 0.903–0.977). In the model 2, in which we added hypercholesterolemia, hypertension, diabetes, obesity, e-GFR, LVH and LAVI (N28 ml/m2), as dichotomic variables, a significant difference in HRs for incident AF was found
Fig. 1. Subclinical organ damage and atrial fibrillation. In this picture we graphically report the mean values of estimated glomerular filtration rate (e-GFR), left ventricular mass index (LVMI), E/A ratio and left atrial area (LAA) in both progressors and non-progressors to AF. As evident, patients developing AF have a higher LVMI and LAA, and a lower e-GFR and E/A ratio.
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259
4
A. Sciacqua et al. / International Journal of Cardiology xxx (2014) xxx–xxx
additive effect of these three risk factors and that such a synergic effect was fully independent of potential confounders.
Table 2 Univariate Cox regression analysis for atrial fibrillation occurrence.
LVMI, 10 g/m2 LDL-cholesterol, 10 mg/dl LAVI, ml/m2 Age, 10 years e-GFR, 10 ml/min/1.73 m2
Hazard ratio
95% CI
1.030 1.042 1.048 1.575 0.940
1.011–1.049 1.018–1.065 1.042–1.053 1.449–1.713 0.903–0.977
LVMI = left ventricular mass index; LAVI = Left atrial volume index; e-GFR = estimated glomerular filtration rate.
for LVH (HR = 1.495, 95% CI = 1.215–1.841), CKD (HR = 1.528, 95% CI = 1.261–1.851), age (10 years) (HR = 1.586, 95% CI = 1.461–1.725) and LAVI (HR = 2.920, 95% CI = 2.426–3.515).
3.3. Explained variation in the incidence rate of AF attributable to biomarkers of organ damage A Cox model including age, BMI and LDL-cholesterol (basic model) provided a 17% explained variation in the incidence rate of AF. By adding LVH to the basic model, the explained variation in the study outcome rose from 17% to 20% and such an increase in prognostic value (+3%) was highly significant (P b 0.001). Remarkably, the addition of LAVI into the model increased the prognostic value and the explained variation in AF rose from 20% to 25% (+5%, P b 0.001). The addition of also CKD to the model increased only by 1% its prognostic performance (from 25% to 26%).
3.2. Interaction analysis
4. Discussion
According to crude analysis (see Fig. 2, upper panel), all risk categories, but particularly that of increased LAVI, were strongly and significantly associated to the risk of AF; notably, patients having an abnormal LAVI, in the absence of reduced e-GFR and LVH, had a relative risk of AF that was about 5 times higher than that of patients in the reference category (HR = 4.767, 95% CI = 3.246–7.001). The interaction analysis revealed that the observed HRs of each risk categories combination were higher than those expected in the absence of interaction under the additive model indicating that the simultaneous occurrence of the two or three risk factors under investigation (increased LAVI, LVH and reduced e-GFR) provided a risk excess for AF higher than that of the simple additive effect of the three risk factors. A multivariate Cox regression analysis, adjusting for a series of potential confounders (age, BMI and LDL-cholesterol) (see Fig. 2, lower panel), confirmed an interaction between increased LAVI and reduced e-GFR for explaining the incidence rate of AF. The interaction between abnormal LAVI and LVH, even if statistically significant, after multivariate data adjustment reduces its strength, suggesting that deranged LAVI and altered LVM are in the same causal pathway conducive to AF in the study population. Remarkably, although reduced e-GFR alone was not significantly associated to the study outcome, it interacted with LVH for amplifying the risk of AF, the observed risk for this outcome associated to these two risk factors being almost double than that expected in the absence of interaction under the additive model. Finally, also on multivariate analysis, a strong interaction was found among abnormal LAVI, reduced e-GFR and LVH for explaining the incidence rate of the study outcome indicating that in patients affected by simultaneous alterations in LA size, LVMI and e-GFR, the risk of AF was consistently higher than the simple
Present data clearly demonstrate that LAVI, CKD and LVH predict the incident AF. The most important novelty of this paper is that we observed, for the first time, that renal dysfunction and LAVI show a synergic effect in predicting the cardiac arrhythmia, totally independent of potential confounders such as traditional cardiovascular risk factors. This finding is consistent with statistical analysis because the observed HR (4.875, 95% CI 2.699–8.805) for incident AF was significantly higher than the expected HR (4.510). Another important finding obtained in this study is the fact that the interaction analysis demonstrated that observed risk for incident AF, associated to LVH and CKD, was almost double (HR = 4.040, 95% CI = 2.661–6.133) (Fig. 2) than that expected in the absence of interaction under the additive model (HR = 2.098). This is not surprising because both CKD and LVM represent wellestablished organ damages reflecting and integrating the long-term cumulative level of activity of several risk factors for cardiovascular disease. In keeping with this, it is interesting that also serum creatinine values, within the normal range, may work as a marker of long-term exposure to elevated BP levels, and therefore are able to predict outcome also in nonrenal target organs [4,9]. The higher rate of incident AF in our population is apparently different from that observed in other studies [31,32]. In particular, our subjects were about 10 years older than those enrolled by Verdecchia et al. [31]; this has a prognostic significance because the Framingham Heart Study demonstrated that the risk of AF doubles for each decade of advancing age, both in men and women [33]. On the other hand, our patients have a higher prevalence of LVH compared to the ONTARGET population (49.5% vs 12.7%). Nevertheless, the ONTARGET study was not designed to detect new onset of AF, and criteria used by Verdecchia et al. for AF detection were deeply different by ours, because they evaluated AF appearance only to ECG performed at follow-up visit. Finally, this prevalence may be justified in part by the fact that we have carefully checked the new AF appearance. Another relevant finding emerged from this study is that an abnormal LA size, in the absence of CKD and LVH, has a relative risk of AF that is about 5 times higher than that of patients in the reference category (HR = 4.767, 95% CI = 3.246–7.001) (Fig. 2), confirming the most relevant role of LA enlargement in the pathogenesis of AF. At first, the LA structural changes reflect the chronicity of exposure to abnormal filling pressures as consequence of diastolic dysfunction [34], similarly to that observed in patients with LVH. According with this, our patients developing AF have a higher LVMI and LA dimension. Next, there are several evidences demonstrating that structural remodeling and atrial dilatation are due to interstitial fibrosis that cause conduction delays between cardiomyocytes favoring the electrical ectopic activity and anisotropic conduction generating non-uniform wave fronts and re-entrant arrhythmias [35]. Notably, results of adjusted interaction analysis showed that the interaction between LA dimension and LVH, even if statistically significant, after multivariate data
Table 3 Multivariate Cox regression analysis for atrial fibrillation occurrence.
Model 1 LVMI LDL-cholesterol LAVI Age e-GFR Model 2 LVH CKD Age LAVI
Increase
HR
95% CI
10 g/m2 10 mg/dl 1 ml/m2 10 years 10 ml/min/1.73 m2
1.030 1.039 1.048 1.575 0.940
1.011–1.050 1.016–1.063 1.042–1.053 1.449–1.713 0.903–0.977
Yes/No Yes/No 10 years Yes/No
1.495 1.528 1.586 2.920
1.215–1.841 1.261–1.851 1.461–1.725 2.426–3.515
Model 1 = smoking (yes/no), fasting glucose, LDL-cholesterol, age, gender (male/female), systolic blood pressure, body mass index, left ventricular mass index (LVMI), E/A ratio, left atrial volume index (LAVI), estimated glomerular filtration rate (e-GFR). Model 2 = smoking (yes/no), age, hypercholesterolemia (yes/no), gender (male/female), hypertension (yes/no), diabetes (yes/no), obesity (yes/no), metabolic syndrome (yes/no), left ventricular hypertrophy (LVH) (yes/no), LAVI increased as ≥28 ml/m2 (yes/no), renal dysfunction as e-GFR b 60 ml/min/1.73 m2 (yes/no).
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259
A. Sciacqua et al. / International Journal of Cardiology xxx (2014) xxx–xxx
5
Fig. 2. Synergic effect of subclinical organ damage on incident AF. Interaction analysis results as crude (upper) and adjusted (bottom) hazard ratios (HR) for incident AF are graphically reported. Expected HRs are represented by white bars, while the observed HRs are represented by the sum of with and dashed bars.
adjustment reduces its biological significance, suggesting that LA enlargement and increased LVM are in the same causal pathway conducive to AF appearance. Present data also demonstrate that CKD was retained as a second independent predictor of AF in the final model of multivariate Cox regression analysis, accounting for 52.8% of the increased risk associated with new appearance of AF. Circulating volume expansion due to sodium retention, activation of RAAS and sympathetic overactivity are important causative mechanism of LA enlargement and AF development [3]. Notably, also small decrements in e-GFR impair regulation of extracellular fluid volume, resulting in impaired renal sodium excretion and increased RAAS activity, conditions associated with subclinical inflammation [8] and cardiovascular events [3,36]. Finally, also the sympathetic activation RAAS-related has been proposed to atrial arrhythmogenesis and AF development [10]. Regarding the association between increased LVM and incident AF, there are strengthened data consistent with pathophysiological mechanisms linking LVH with atrial arrhythmia. In fact, LA size represents a marker of both left ventricular stiffness and diastolic dysfunction in hypertensive, diabetic and elderly subjects. In keeping with this, LA size increases with worsening diastolic dysfunction [34], independently of traditional cardiovascular risk factors and independently of LV ejection fraction. Besides, the final model of multivariate analysis revealed a strong interaction among abnormal LAVI, CKD and LVH, indicating that in patients showing all three alterations in LA size, LVM and renal function, the risk of AF was consistently higher than the simple additive effect of these three organ damages. Interestingly, LAVI has a predictive value higher than that of LVH increasing of 5% the risk of new AF explained by basic model, while LVH adds only a 3%. In keeping with this, these results suggest that interventions simultaneously focused on the treatment and/or prevention of all three subclinical organ damages could have beneficial effects for reducing the risk of AF beyond that expected by the simple sum of the single beneficial effects. Thus, it should be convenient to utilize drugs able to modulate the RAAS, not only for reducing BP values, but also to interfere with the appearance and progression of subclinical organ damage; particularly, the effect of these drugs is more evident on left atrium structure in which they have demonstrated to reduce fibrosis and atrial remodeling [19]. According with this, a recent meta-analysis demonstrated that RAAS inhibition may have an important role in the primary and secondary prevention of AF [37].
Finally, in our study, subclinical organ damages were emerged as independent predictors of incident AF, interacting too, in a synergic manner, in increasing the arrhythmic risk. This emphasizes that our findings are more important from a pathophysiological point of view than from the perspective of risk prediction. On the basis of these results it is possible to suggest that AF may be considered a time-integrated marker of exposure to multiple cardiovascular risk factors, particularly CKD indicating that its occurrence, simultaneously to the other two subclinical disease provides a risk excess for AF higher than that of the simple additive effect. Thus, strategies for the prevention of AF will have to recognize also CKD as a preventable risk factor for the development of AF in addition to other well-established risk factors. In keeping with this, the diagnosis of CKD should alert the practitioner to routinely assess also renal function to correctly define the global AF risk. However, the evaluation of this risk is often not optimized because screening for CKD is frequently limited to a measurement of serum creatinine, which does not accurately reflect GFR, the best indicator of renal function in healthy and diseased subjects. 4.1. Limitations This study has some limitations. At first, the type of AF was not discriminated in this study. Another limitation consists in the fact that e-GFR was not directly measured, but estimated by creatinine value. In conclusion, in spite of the above limitations, data obtained from this study clearly demonstrate that LAVI and CKD significantly interact in a synergic manner in increasing AF risk. This evidence encourages to detect routinely the renal function in all patients that are at increased risk to develop AF, allowing thus a better stratification of AF risk. References [1] National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification and stratification. Am J Kidney Dis 2002;39: S1–S266. [2] Sarnak MJ, Levey AS, Schoolwerth AC, et al. American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention. Kidney disease as a risk factor for development of cardiovascular disease. A statement from the American Heart Association councils on kidney in cardiovascular disease, high blood pressure research, clinical cardiology, and epidemiology and prevention. Hypertension 2011;42:1050–65. [3] 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;351:1296–305.
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259
6
A. Sciacqua et al. / International Journal of Cardiology xxx (2014) xxx–xxx
[4] Schillaci G, Reboldi G, Verdecchia P. High-normal serum creatinine concentration is a predictor of cardiovascular risk in essential hypertension. Arch Intern Med 2001;161:886–91. [5] Jensen JS, Feldt-Rasmussen B, Strandgaard S, Schroll M, Borch-Johensen K. Arterial hypertension, microalbuminuria, and risk of ischemic heart disease. Hypertension 2000;35:898–903. [6] Meguid El NA, Bello AK. Chronic kidney disease. The global challenge. Lancet 2005;365:331–40. [7] Sciacqua A, Scozzafava A, Pujia A, et al. Interaction between vascular dysfunction and cardiac mass increases the risk of cardiovascular outcomes in essential hypertension. Eur Heart J 2005;26:921–7. [8] Zoccali C, Maio R, Tripepi G, Mallamaci F, Perticone F. Inflammation as a mediator of the link between mild to moderate renal insufficiency and endothelial dysfunction in essential hypertension. J Am Soc Nephrol 2006;17:S64–8. [9] Miceli S, Maio R, Perticone M, Mallamaci F, Perticone F. Creatinine and insulin predict cardiac mass in drug-naive hypertensive patients. Int J Cardiol 2013;167:519–24. [10] Schlaich MP, Socratous F, Hennebry S, et al. Sympathetic activation in chronic renal failure. J Am Soc Nephrol 2009;20:933–9. [11] Perticone F, Maio R, Cosco C, et al. Hypertensive left ventricular remodeling and ACE-gene polymorphism. Cardiovasc Res 1999;43:192–9. [12] Naccarelli GV, Johnston SS, Lin J, Patel PP, Schulman KL. Cost burden of cardiovascular hospitalization and mortality in ATHENA-like patients with atrial fibrillation/atrial flutter in the United States. Clin Cardiol 2010;33:270–9. [13] McManus DD, Corteville DCM, Shlipak MG, Wholeey MA, Ix JH. Relation of kidney function and albuminuria with atrial fibrillation (from the Heart and Soul Study). Am J Cardiol 2009;104:1551–5. [14] Baber U, Howard VJ, Halperin JL, Wholeey MA, Ix JH. Association of chronic kidney disease with atrial fibrillation among adults in the United States: REasons for Geographic and Racial Differences in Stroke (REGARDS) Study. Circ Arrhythmia Electrophysiol 2011;4:26–32. [15] Perticone F, Maio R, Ruberto C, et al. Kidney function and risk factors for left ventricular hypertrophy in untreated uncomplicated essential hypertension. Am J Kidney Dis 2008;52:74–84. [16] Sciacqua A, Miceli S, Carullo G, et al. One-hour postload plasma glucose levels and left ventricular mass in hypertensive patients. Diabetes Care 2011;34:1406–11. [17] Sesti G, Sciacqua A, Scozzafava A, et al. Effects of growth hormone and insulin-like growth factor-1 on cardiac hypertrophy of hypertensive patients. J Hypertens 2007;25:471–7. [18] Lorell BH, Carabello BA. Left ventricular hypertrophy: pathogenesis, detection, and prognosis. Circulation 2000;102:470–9. [19] Wachtell K, Gerdts E, Aurigemma GP, et al. In-treatment reduced left atrial diameter during antihypertensive treatment is associated with reduced new-onset atrial fibrillation in hypertensive patients with left ventricular hypertrophy: The LIFE Study. Blood Press 2010;19:169–75. [20] Eckstein J, Verheule S, de Groot NM, Allessie M, Schotten U. Mechanisms of perpetuation of atrial fibrillation in chronically dilated atria. Prog Biophys Mol Biol 2008;97:435–51.
[21] Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of the National Cholesterol Education Program (NCEP) Expert Panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA 2001;285:2486–97. [22] Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150:604–12. [23] Vieira-Filho NG, Mancuso FJ, Oliveira WA, et al. Simplified single plane echocardiography is comparable to conventional biplane two-dimensional echocardiography in the evaluation of left atrial volume: a study validated by three-dimensional echocardiography in 143 individuals. Echocardiography 2014;31:265–72. [24] Lang RM, Bierig M, Devereux RB, et al. Recommendations for chamber quantification. Eur J Echocardiogr 2006;7:79–108. [25] Cheitlin MD, Armstrong WF, Aurigemma GP, et al. ACC/AHA/ASE 2003 guideline update for the clinical application of echocardiography: summary article: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/ASE Committee to Update the 1997 Guidelines for the Clinical Application of Echocardiography). Circulation 2003;108:1146–62. [26] Devereux RB, Alonso DR, Lutas EM, et al. Echocardiographic assessment of left ventricular hypertrophy comparison to necropsy findings. Am J Cardiol 1986;57:450–8. [27] Marcus R, Krause L, Weder AB, et al. Sex-specific determinants of increased left ventricular mass in the Tecumseh Blood Pressure Study. Circulation 1994;90:928–36. [28] de Mutsert R, Jaeger KJ, Zoccali C, Dekker FW. The effect of joint exposures: examining the presence of interaction. Kidney Int 2009;75:677–81. [29] Collet D. Modelling survival data in medical research. London: Chapman and Hall; 1994 72–85. [30] Hosmer DW, Lemeshow S. Assessment of model adequacy. Applied survival analysis. New York: Wiley & Sons; 1999. p. 196–240. [31] Verdecchia P, Reboldi G, Gattobigio R, et al. Atrial fibrillation in hypertension: predictors and outcome. Hypertension 2003;41:218–23. [32] ONTARGET InvestigatorsYusuf S, Teo KK, et al. Telmisartan, ramipril, or both in patients at high risk for vascular events. N Engl J Med 2008;358:1547–59. [33] Benjamin EJ, Levy D, Vaziri SM, et al. Independent risk factors for atrial fibrillation in a population-based cohort. The Framingham Heart Study. JAMA 1994;271:840–4. [34] Tsang TS, Barnes ME, Gersh BJ, Bailey KR, Sewatd JB. Left atrial volume as a morphophysiologic expression of left ventricular diastolic dysfunction and relation to cardiovascular risk burden. Am J Cardiol 2002;90:1284–9. [35] Zhao W, Zhao T, Chen Y, Ahokas RA, Sun Y. Oxidative stress mediates cardiac fibrosis by enhancing transforming growth factor-beta1 in hypertensive rats. Mol Cell Biochem 2008;317:43–50. [36] Manjunath G, Tighiouart H, Ibrahim H, et al. Level of kidney function as a risk factor for atherosclerotic cardiovascular outcomes in the community. J Am Coll Cardiol 2003;41:47–55. [37] Schneider MP, Hua TA, Böhm M, et al. Prevention of atrial fibrillation by reninangiotensin system inhibition a meta-analysis. J Am Coll Cardiol 2010;55:2299–307.
Please cite this article as: Sciacqua A, et al, Renal disease and left atrial remodeling predict atrial fibrillation in patients with cardiovascular risk factors, Int J Cardiol (2014), http://dx.doi.org/10.1016/j.ijcard.2014.04.259