PERICARDIAL FAT IS STRONGLY ASSOCIATED WITH ATRIAL FIBRILLATION AFTER CORONARY ARTERY BYPASS GRAFT SURGERY

Transcription

1 d PERICARDIAL FAT IS STRONGLY ASSOCIATED WITH ATRIAL FIBRILLATION AFTER CORONARY ARTERY BYPASS GRAFT SURGERY George Drossos 1, Charilaos-Panagiotis Koutsogiannidis 1, Olga Ananiadou 1, George Kapsas 2, Fotini Ampatzidou 1, Athanasios Madesis 1, Kalliopi Bismpa 2, Panagiotis Palladas 2, Labros Karagounis 3 1. Cardiothoracic Surgery Department, General Hospital G. Papanikolaou, Thessaloniki, Greece 2. Department of Radiology, General Hospital G. Papanikolaou, Thessaloniki, Greece 3. Department of Interventional Cardiology, Interbalkan Medical Center, Thessaloniki, Greece Accepted for oral presentation in 27 th EACTS Annual Meeting, Vienna Austria, 5-9 October 2013 Abstract Word Count: 317 Text Word Count: 3840 Corresponding Author George Drossos Cardiothoracic Surgery Department General Hospital G.Papanikolaou Exohi, Thessaloniki, Greece Tel: M: Fax: drogevan@otenet.gr

2 d ABSTRACT Objective: Recent evidence suggests that pericardial fat may represent an important risk factor for cardiovascular disease because of its unique properties and its proximity to cardiac structures. It has been reported that pericardial fat volume is associated with atrial fibrillation. The purpose of this study was to investigate the association between pericardial fat volume and new-onset atrial fibrillation (AF) following coronary artery bypass graft surgery (CABG). Methods: Pericardial fat volume was measured using computed tomography in 83 patients with coronary artery disease scheduled to undergo elective isolated on pump CABG. Patient characteristics, medical history, and perioperative variables were prospectively collected. Any documented episode of new onset postoperative atrial fibrillation until discharge was defined as the study end point. Results: Twenty eight patients (33.7%) developed postoperatively AF during hospital stay. There was no significant difference in demographics and comorbidities among patients that maintained sinus rhythm (SR) and their AF counterparts. In univariate analysis, patients with postoperative AF had significantly more pericardial fat compared with SR patients (195 ± 80 ml vs. 126 ± 47 ml, p=0.0001). Larger left atrial diameter was also associated with postoperative AF (42.4±6.9 mm vs. 39.3±4.8 mm, p=0.017). Additionally, the prebypass use of calcium channel-blocking agents was independently associated with a lower incidence of postoperative AF, confirmed also by multivariate analysis (p = 0.035). In multivariate logistic regression analysis pericardial fat volume was the strongest independent variable associated with the development of postoperative AF (odds ratio (OR) 1.018, 95% confidence interval (CI) , p = ). The best discriminant value assessed by Receiver Operating Characteristic analysis was ml (sensitivity 86% and specificity 56%).

3 d Conclusions: Pericardial fat volume is strongly associated with atrial fibrillation following CABG, independently of many traditional risk factors. Our findings suggest that pericardial fat volume may represent a novel risk factor for postoperative AF. However, the role of pericardial fat in AF mechanism needs to be further delineated. KEYWORDS: pericardial fat volume, postoperative atrial fibrillation, predictors, coronary artery bypass surgery

4 d INTRODUCTION Atrial fibrillation (AF) occurs in 20% 40% of patients after coronary artery bypass grafting (CABG) requiring cardiopulmonary bypass (CPB) and it may be a marker of postoperative complications, including stroke, congestive heart failure, and increased mortality. [1-2] Postoperative AF is likely related to a combination of factors. These include pre-existing degenerative myocardial changes and pre- and perioperative conditions that result in abnormalities of electrophysiologic parameters that promote the development of AF. Clinical variables, such as advanced age, hypertension, male gender, right coronary artery (RCA) stenosis, depressed left ventricular function, and a remote history of previous AF may predispose to postoperative AF; intraoperative surgical variables, including combined valve replacement/cabg procedures and prolonged aortic cross-clamp and bypass times, are similarly associated with a more frequent AF incidence. [3-5] Therefore, identifying high-risk patients and facilitating preventive measures is of utmost importance. Recently, attention has been focused on heart adiposity because of its inflammatory and endocrine properties. The theory of a local toxic effect of excess adipose tissue is supported by evidence from basic science, translational science, and epidemiology. [6-8] A significant contribution was made by the rapid development in the field of noninvasive imaging, which has made it possible to quantify fat masses with accuracy. [9-10] Adipose tissue surrounding the heart, within the pericardial sac, is called epicardial or pericardial fat and given its direct apposition to myocardium and coronary arteries, it may play a central role in the pathogenesis of cardiovascular disease, mediated by its inflammatory properties. [11] Several studies have shown a relationship of increased pericardial fat volume (PFV) with coronary artery disease, atherosclerosis and the progression of coronary plaque burden [12-14], major adverse cardiovascular events [15], and atrial fibrillation. [16-19] Adipose tissue secretes both proinflammatory cytokines such as interleukin-6 and anti-inflammatory

5 d mediators such as adiponectin. [6-8, 11] In a recent study Kourliouros et al have reported that epicardial adiponectin is associated with maintenance of sinus rhythm following cardiac surgery. This paper reinforces the inflammatory hypothesis in the pathogenesis of postoperative AF, which may be result of a balance of pro- and anti-inflammatory markers. [20] In the context of cardiac surgery, inflammation due to CPB, ischemia and oxidative stress has been implicated in the development of postoperative AF. [1, 4-5] To the best of our knowledge, the relationship of pericardial fat volume with post CABG AF has not been explored. In view of evidence suggesting the modulatory role of epicardial adipose tissue in arrhythmogenesis and its association with a nearly 40% higher odds of AF, we conducted a study to evaluate the impact of pericardial fat on the development of postoperative AF. The aim of the present study was thus to characterize the relationship between pericardial fat volume, as measured by multidetector computed tomography (MDCT), and the occurrence of new-onset postoperative AF, in a CABG cohort. MATERIAL AND METHODS Patient Population and data collection A total of 83 patients undergoing elective first time coronary revascularization for single-, double-, triple- or more vessel disease, were recruited in our study over a 12-month period (January 2012 to December 2012), following approval from the institutional review board. Patients were selected at the Clinic of Cardiac Surgery by surgeon decision. All subjects were acceptable candidates for elective operation with CPB. They all provided

6 e written informed consent and underwent preoperative MDCT for pericardial fat measurement and echocardiography. Patient characteristics, medical history, intra-operative variables and postoperative outcomes were prospectively collected. Potential predictors of atrial fibrillation were chosen based on a review of the literature. Data prospectively collected from patient medical records included age, sex, body mass index (BMI) (kg/m 2 ), preoperative medication with b-blockers, calcium channel blockers (CCBs) or angiotensin converting enzyme inhibitors (ACEIs) and comorbidities (hypertension [HTN], diabetes mellitus [DM], congestive heart [CHF]-New York Heart Association [NYHA] III-IV, chronic obstructive pulmonary disease [COPD], peripheral vascular disease [PVD], recent myocardial infarction [MI] and history of cerebrovascular accident [CVA]). Preoperative variables also recorded were: presence of critical RCA stenosis (defined as narrowing of the proximal or mid-segment greater than or equal to 90% of lumen diameter and assessed on coronary angiogram), left ventricular hypertrophy and left ventricular ejection fraction (assessed on preoperative echocardiography). Procedural related variables included aortic cross clamp time, pump time, use of retrograde cardioplegia, number of bypass grafts and the perioperative use of intraaortic balloon pump (IABP). Postoperatively levels of potassium and magnesium were daily recorded. Exclusions Exclusion criteria were presence of AF (persistent or paroxysmal) or past history suggestive of AF or any other arrhythmia atrial and concomitant valve surgery along with CABG. We also excluded patients in dialysis, use of anti-arrhythmic medications (other than b blockers), presence of permanent pacemakers, systemic inflammatory or terminal illness,

7 e and patients under treatment with corticosteroids or non-steroidal anti-inflammatory drugs, other than aspirin. CT imaging protocol CT studies were performed using a four slice MDCT scanner (Siemens Sensation 4, Siemens Medical Solutions, Erlangen, Germany). Continuous 2 mm slices of the heart with 1 mm overlap were acquired, from the level of the bifurcation of the pulmonary artery to the diaphragm. The images were analyzed at a dedicated workstation (Siemens Leonardo, Siemens Medical Solutions, Erlangen, Germany), using a semi-automated technique. The operator was obliged to manually trace the pericardium in each of the slices, in order to create a 3D volume of interest. Pericardial fat was defined as the adipose tissue contained in this volume, within the pericardial sack. (Figure 1) A threshold between -200 and -30 Hounsfield units was used to identify the adipose tissue within the volume of interest. The pericardial fat was measured by two independent readers with excellent intra-observer and inter-observer reproducibility (0.95 and 0.97 respectively). Echocardiographic measurements Transthoracic echocardiography was used to assess the structural and functional condition of the heart preoperatively. Left atrial diameter (LAD) was measured at end systole in the parasternal long-axis view. Left atrial enlargement was defined as diameter > 4.2 cm. We also evaluated the thickness of the interventricular septum (IVS) and posterior LV wall in diastole and the end-diastolic LV dimension using M-mode and 2D images from parasternal and apical projections. Left ventricular ejection fraction (LVEF) was measured by biplanar Simpson s method. Surgical protocol

8 e All patients underwent standard median sternotomy CABG performed by a single consultant surgeon and were operated on CPB. All efforts were made to leave aortic fat pad intact but anterior fat pat in the aortopulmonary window was dissected before placing the aortic cross-clamp. After initiation of CPB, myocardial protection was achieved with ice slush and cold blood cardioplegia, with a temperature drift to 34 C. Conduits for bypass included the internal mammary artery in all cases and saphenous vein grafts.the distal anastomoses were performed first. The left internal mammary artery was exclusively used to bypass left anterior descending artery stenosis and major saphenous vein grafts were used to bypass the diseased marginal, diagonal and/or right coronary artery. The proximal anastomoses were constructed with the aorta clamped. Postoperative protocol All patients were nursed initially in the cardiac surgery intensive care unit (ICU) and then on high dependency unit (HDU) and the cardiothoracic wards until discharge. Potassium supplementation was given as necessary to maintain electrolyte balance within the normal range. Patients were continuously monitored in cardiac surgery ICU and HDU and monitoring was continued after being transferred to the ward with electrocardiographic (ECG) telemetry with a central monitor at the nursing station. All hemodynamically stable patients received b- blockade on the first postoperative day and had routine daily 12-lead ECGs. Nursing observations were performed routinely every 4 hours by measuring blood pressure, pulses, assessing the heart rhythm and carrying out 12-lead ECGs whenever a cardiac dysrhythmia was suspected. Postoperative AF was defined as new-onset AF, detected by continuous telemetry monitoring during the postoperative hospital stay. The analysis took into account any single episode of AF requiring treatment or lasting longer than 20 minutes, or for a cumulative total of more than 60 minutes within a 24-hour period. Any documented episode of postoperative atrial fibrillation was defined as the study end point. Patients, who developed

9 e AF, were treated with amiodarone (intravenous and oral, following appropriate loading) while correcting electrolyte and acid-base imbalances. Statistical analysis Univariate analysis was performed to examine the association of patient and operative variables on postoperative AF, followed by multivariate logistic regression analysis to identify independent predictors of postoperative AF. Candidate variables with p < 0.2 upon univariate analysis were entered into the logistic regression model. Dichotomous data were evaluated using Pearson s chi-square test or Fisher s exact test, whereas unpaired t-test and Mann -Whitney U test were used for continuous variables with normal or non-normal distribution, respectively. A p value < 0.05 was considered statistically significant. Odds ratios and 95% CIs were calculated for independent associates of AF. Statistical analysis was performed with SPSS software package. RESULTS Eighty three patients undergoing their first elective on pump CABG surgery were enrolled in our study. The mean age of the patients was 64.6 ± 8.9 years (range, 38 to 82 years); 79.5% of patients were male. The mean BMI was 29.6 ± 4.5 kg/m 2. Comorbid conditions included HTN (81.9%), DM (47%), COPD (18%), PVD (24.1%), history of CVA (8.4%), recent MI (41%) and NYHA III-IV status (13.3%). Fifty (60.2%) patients had critical RCA stenosis, 69 (83.1%) were on preoperative medication with b-blockers, only 20 (24.1%) on CCBs and 31 (37.3%) on ACEIs inhibitors. In this study population, 28 (33.7%) patients developed postoperative AF and 55 (66.3%) maintained sinus rhythm (SR) during hospitalization. Baseline characteristics and procedural differences between patients who developed postoperative AF and those who remained in sinus rhythm are shown in Table 1. As Table 1 demonstrates, patients with AF

10 dd had significantly more pericardial fat compared with patients in SR (195 ± 80 ml vs 126 ± 47ml, p = ). In addition, left atrial enlargement was also implicated, as patients who developed AF had higher atrial dimension than those who maintained SR (42.5 ± 7 vs 29 ± 4.6mm, p = 0.017). Age and gender were not statistically associated with the development of AF upon univariate analysis. The percentage of patients aged over 70 years old was the same and also body mass index was not statistically different between SR and AF groups. Equally, differences in comorbid factors as DM, CVA, PVD, COPD, CHF and recent MI were not statistically significant. However, interesting enough, patients who remained in SR were more hypertensive, as compared with their AF counterparts, which almost reached statistical significance (p=0.076). As for medication prior CABG, univariate analysis demonstrated that subjects that maintained SR had higher intake of CCBs (p = 0.042). There was no difference in b-blockers intake between groups (0.99). Procedure-related factors, as CPB and cross clamp duration or the use of retrograde cardioplegia and number of bypass grafts, did not demonstrate any correlation with postoperative AF either. No episode of hypomagnesaemia (defined as Mg <2mg/dL and measured once a day) was documented postoperatively. Also, there was no difference in episodes of documented hypokalemia (defined as K <3.5meq/L and measured every 2, 4, 6, 8 hours depending on the postoperative day) between patients that developed AF and those that maintained SR (p=0.54). Variables with p < 0.2 upon univariate analysis that entered into the regression model were pericardial fat volume, preoperative LAD, hypertension, PVD, CCBs and ACEIs intake. Multivariate logistic regression analysis identified PFV as a strong independent predictor of postoperative AF (odds ratio (OR) 1.018, 95% confidence interval (CI) , p = ). This association was completely independent of LA size, hypertension and PVD. Receiver Operating Characteristic (ROC) analysis showed that the best discriminatory level

11 dd of PFV to predict new onset post CABG atrial fibrillation was ml (sensitivity 86% and specificity 56%) (Figure 2). The area under the curve (AUC) was with 95% CI and a p value < The second variable which was also associated with the development of postoperative AF was the absence of calcium channel blockers in patient s preoperative treatment (OR 5.4, 95% CI , p = 0.035). When the PFV was dichotomized by the value 129.5, the logistic regression showed OR 9.8, 95% CI and p = All variables examined in multivariate regression analysis are shown in Table 2. DISCUSSION Atrial fibrillation is one of the most common complications following CABG and the incidence has not changed in the last two decades, despite improvements in myocardial preservation and surgical techniques. The incidence of postoperative AF in our study was 34%, corresponding to the data in the literature. The relatively high mean age of our patients, the gender and BMI didn t have an impact on development of postoperative AF. Also, well known risk factors and comorbidities did not reach statistical significance, mainly due to the small sample size. Previous reports suggest that beta-blockers suppress post-cabg AF; however, preoperative b-blockers intake was not associated with lower incidence of AF. In contrast, preoperative use of CCBs afforded a significant protective effect against arrhythmia in univariate analysis (p = 0.042). Further multivariate logistic analysis showed that CCBs intake was also significantly correlated with reduced incidence of postoperative AF (p = 0.035). Structural changes of the heart, such as LA dilation and LV hypertrophy, have been found to be other factors contributing to postoperative AF. [4-5] One early report investigating the relationship between LA parameters and post-cabg AF stated that only LA

12 dd diameter was associated with postoperative AF [21] In consistence with that finding, LA size was significantly larger in AF patients compared with SR patients (p=0.017). The higher mean of preoperative LA diameter in patients that developed post-cabg AF could have been a significant inherent risk of postoperative AF, but further analysis didn t confirm LA diameter as an independent predictive factor for AF in our study. However, only one dimension of the LA was assessed and volume of the LA was not evaluated. Additionally, the two patient groups showed no differences in echocardiographic parameters of the left ventricle (LVH or LVEF). To overcome some of the limitations of previous studies, including mixed coronary and valve cases, re-operations and off-pump cases, our population was homogeneous with patients undergoing first time, elective isolated CABG with the use of CPB. Although previous studies have shown a relationship between postoperative AF and CPB and cross clamp duration, this study didn t find significant difference between SR and AF patients. [3,5] Electrolyte imbalance, especially hypokalemia, is also considered to be a triggering factor of postoperative AF [1, 2]. In this study, there was no statistically significant relationship between serum potassium and the incidence of postoperative AF. This may be partly explained by normal serum potassium levels in the subgroups due to routine measurements and prompt supplementation of potassium in all patients after surgery. The key measure of this study was PFV with the aid of noninvasive imaging. Based on the renewed interest on heart adiposity and on current evidence that PFV is associated with AF prevalence, we elected to evaluate the association with postoperative AF. This study, for the first time, showed that patients who developed AF following isolated on pump CABG, had significantly higher PFV values than those who maintained SR. The association of pericardial fat with AF was further confirmed in multivariate logistic regression analysis, indicating increased PFV as a strong independent risk factor of postoperative AF. In

13 dd particular, ROC analysis showed that the best discriminatory predictive level of PFV was ml, with sensitivity of 86% and specificity of 56%. Although statistical analysis confirmed such an association, the role of the MDCT scan as a screening test remains uncertain. Further studies are required to evaluate the balance cost/benefit of this radiologic examination as screening tool before cardiac operations. Multiple studies have implicated the role of pericardial fat, quantified by cardiac CT, in the pathophysiology of coronary atherosclerosis. In the Framingham Heart Study and the Multi-Ethnic Study of Atherosclerosis, the volume of pericardial fat was found to be associated with coronary artery calcium and incident coronary artery disease. [12-14] In a community-based study of nearly 1000 participants undergoing MDCT examination, it was found that pericardial fat volume was associated with LV mass, LVEDV, and LA dimension in women and with LV mass and LA dimension in men. [23] In another report, PFV was found to be positively associated with LA size in men. Subsequent work supported these findings by demonstrating a positive association between pericardial adipose tissue and prevalent atrial fibrillation, which is known to be associated with LA size.[16] Al Chekakie and colleagues demonstrated a significant association of PFV with both paroxysmal and persistent AF that is completely independent of all major risk factors age, gender, hypertension, valvular heart disease, left ventricular ejection fraction, diabetes mellitus, and BMI, including also LA enlargement.[19] Wong and colleagues showed that PFV is associated with the presence and severity of atrial fibrillation, left atrial volumes and poorer outcomes after atrial fibrillation ablation, independent of systematic measures of adiposity. [17] Increasing evidence seems to point toward pericardial fat as a metabolically active tissue modulating the adjacent myocardial tissue, shedding further light on the idea of a local paracrine effect of epicardial tissue with findings of increased expression of numerous

14 dd inflammatory markers. [6-7, 18] Although AF pathophysiology is complex, increasing evidence supports mechanistic links between inflammation and AF. Surgical myocardial revascularization using CPB is still the therapeutic gold standard for multivessel coronary involvement. However, this procedure is associated with the occurrence of systemic inflammatory response syndrome of various possible causes, which may lead to postoperative AF. The intraoperative and postoperative periods are potentially stressful for the heart with extremes in response affecting reperfusion, inflammation, hemostasis, and excitotoxicity. Inflammation is one of the risk factors of postoperative AF by affecting atrial conduction. [1] Inflammatory markers including IL-6, IL-8, C-reactive protein, tumor necrosis factor-α and indices of neutrophil and platelet activation are significantly increased in the systemic bloodstream after CABG. [1, 4-5] In our study, patient population is limited to on pump cases, not allowing definite conclusions to be made on association with the CPB effect. The interrelation though, of epicardial adiposity and CPB related inflammatory response should be further delineated. Although epicardial adipose tissue is considered to be proinflammatory and is associated with AF, not all epicardial fat pads are similar. The heart is reported to have mainly 3 epicardial fat pads with parasympathetic ganglia: a fat pad on the anterior surface of the atria located in the aortopulmonary window, a posterior fat pad between the inferior venacava and left atrium, and the one between the superior vena-cava and right atrium. [6,9,18] In an imaging analysis of 169 patients with cardiac CT, a significant, direct, relationship between AF burden and the thickness of the posterior periatrial fat pad between the esophagus and LA was found, independent of BMI, age and LA area. It was hypothesized that local inflammatory mediators produced by the periatrial epicardial fat in the LA posterior wall promote the activation of ectopic foci in the pulmonary vein ostia. [18] Additionally, reports about the impact of surgical preservation of the anterior fat pad between the aorta and right

15 dd pulmonary artery on postoperative AF have had conflicting results. It is considered usual practice to dissect it during CABG to clear the field of view and secure the aorta cross-clamp. Canine studies demonstrated that this tissue contains vagal neurons and a small study in humans clearly demonstrated a paradoxical increased incidence of postoperative AF in patients with anterior FP dissection as compared with patients with preservation during CABG, probably due to imbalance of autonomic tone. [24] However, these data are not yet confirmed by similar studies and we proceeded accordingly with our routine operative practice of dissecting anterior fat pad. Furthermore, current knowledge of cardiac lymphatics demonstrates their possible role in triggering arrhythmias in the postoperative period. Coronary bypass surgery requires removal from the ascending aorta of adventitial tissue and the aortic fat pad, which contains the sinoatrial node lymphatic collector. In addition, a cross-clamp applied to aorta obstructs the entire lymphatic drainage from the heart and thus affects the conduction system of the heart. [25] During surgery, we tried to minimize aortic manipulations and right atrial damage and preserve the aortic fat pad to avoid negative effect on node function. Both distal and proximal anastomoses were performed under a single aortic cross-clamping period and duration was not different in the two subgroups. In light of available data, our study further bolsters the association of increased PVF with AF. Our findings are consistent with the hypothesis of a local pathogenic effect of pericardial fat promoting an arrhythmogenic substrate. Study Limitations Our study has a number of important limitations. In particular, the study design limits inferences of causality and the single-institution data limit generalizability. The current study was primarily designed to assess the association of pericardial fat volume with postoperative

16 de AF. Patient population was narrowed to include only patients receiving standard median sternotomy on-pump bypass surgery. The role of statins in postoperative AF was not analyzed. Other unmeasured factors also, may partially account for our findings. We also did not measure intra-thoracic and visceral abdominal fat; these measures may have added incremental information on the effects of local versus systemic adiposity. Finally, because computed tomography scanning was performed without heart rate control, it is possible that measurements of pericardial fat may have been affected by motion artifacts. Our imaging protocol analyzed the pericardial fat surrounding the entire heart and didn t allow for focal fat quantification. Given the small sample size, we were unable to conduct many important analyses including the relation between AF and anatomic distribution of pericardial fat. Future software development may allow studies to include more sophisticated and detailed volumetric assessments of pericardial fat. Therefore, we believe that using imaging data gathered prospectively from an international multicenter population will confirm our findings and examine important subgroup analyses. Conclusions To our knowledge, the present study provides the first report of an association between cardiac adiposity, as measured by pericardial fat, and postoperative AF. We have described the incidence of AF in a population undergoing elective isolated CABG, and identified PFV as a strong independent predictor of postoperative AF, thereby to allow preoperative identification of vulnerable patients. However, the nature of the role that pericardial fat plays in cardiac surgery and postoperative AF pathogenesis requires further investigation. Conflicts of Interest Conflict of interest: none declared.

17 de TABLES Table 1. Patient characteristics and procedural variables according to postoperative rhythm Variable a Sinus Rhythm Atrial Fibrillation P value N Age (years) 64 ± 9 66 ± Age >70yrs old 16 (29) 8 (29) 0.96 Male 44 (80) 22 (79) 0.88 BMI (kg/m 2 ) 30 ± ± Pericardial fat volume (ml) 126 ± ± * Medical History Hypertension 48 (87) 20 (71) 0.076* Diabetes 27 (49) 12 (43) 0.59 Chronic Obstructive 8 (14.5) 7 (25) 0.24 Pulmonary Disease Cerebrovascular 4 (7.3) 3 (11) 0.59 Accident Peripheral Vascular Disease 16 (29) 4 (14) 0.13* Recent myocardial 25 (45) 9 (32) 0.24 infarction NYHA III-IV 6 (11) 5 (18) 0.38 Beta blockers 44 (80) 25 (89) 0.99 Angiotensin-convertingenzyme 24 (44) 7 (25) 0.097* inhibitor intake Calcium channel blockers 17 (31) 3 (11) 0.042* intake Coronary Angiography and Echocardiography findings Right coronary artery 31 (56) 19 (68) 0.31 disease Preoperative LVEF (%) 57 ± ± Preoperative LVH 7 (13) 5 (18) 0.53 Preoperative LAD (mm) 29 ± ± * Procedural variables No of bypass grafts 2.4 ± ± CPB time (min) 83.6 ± ± Cross clamp time (min) 63.7 ± ± Use of retrograde 22 (40) 10 (36) 0.70 cardioplegia Perioperative use of IABP 3 (6) 2 (7) 0.76 Postoperative variables K < 3.5 meq/l 16 (29) 10 (36) 0.54 BMI: Body Mass Index; NYHA: New York Heart Association; LVEF: left ventricular Ejection Fraction; LVH: Left ventricular Hypertrophy; LAD: Left Atrial Diameter; No: number; CPB: Cardiopulmonary Bypass; IABP: Intra-aortic Balloon Pump. a Categorical data are numbers (%); continuous data as means ± standard deviation. * Variables examined in the multivariate regression analysis

18 de Table 2. Variables examined in the multivariate regression analysis Variable Odds Ratio 95% Confidence Interval P value Pericardial fat volume (ml) Absence of calcium channel blockers intake Preoperative LAD (mm) Hypertension Peripheral Vascular Disease Angiotensin-convertingenzyme inhibitor intake

19 de FIGURE LEGENDS Figure 1.Images generated after the semi-automatic segmentation of the pericardial fat and calculation of its volume. Axial (A) and sagittal (B) CT images, in which the pericardial fat is highlighted in orange color. Three dimensional reconstruction (C) of the pericardial adipose tissue of the same subject, displayed with the volume rendering technique. For this subject the pericardial fat volume was 266 cm 3. Figure 2. ROC curve illustrating sensitivity and false positive rate (1-specificity) at all possible cut-off levels to assess the performance of pericardial fat volume to predict post CABG AF.

20 dd REFERENCES [1] Echahidi N, Pibarot P, O Hara G, Mathieu P. Mechanisms, prevention, and treatment of atrial fibrillation after cardiac surgery. J Am Coll Cardiol 2008;51: [2] Bharucha DB, Kowey PR. Management and prevention of atrial fibrillation after cardiovascular surgery. Am J Cardiol 2000;85:20D 24D. [3] Aranki FS, Shaw DP, Adams DH, Rizzo RJ, Couper GS, VanderVliet M et al. Predictors of atrial fibrillation after coronary artery surgery. Current trends and impact of hospital resources. Circulation 1996;94: [4] Hogue CW, Hyder ML. Atrial fibrillation after cardiac operation: risks, mechanisms, and treatment. Ann Thorac Surg 2000;69: [5] Jakubováa M, Mitrob P, Stanˇcákb B, Sabolc F, Kolesárc A, Cisarikd P et al. The occurrence of postoperative atrial fibrillation according to different surgical settings in cardiac surgery patients. Interact CardioVasc Thorac Surg 2012; 15: [6] Iacobellis G, Corradi D, Sharma AM. Epicardial adipose tissue: anatomic, biomolecular and clinical relationships with the heart. Nat Clin Pract Cardiovasc Med 2005;2: [7] Mazurek T, Zhang L, Zalewski A, Mannion JD, Diehl JT, Arafat H et al. Human epicardial adipose tissue is a source of inflammatory mediators. Circulation 2003;108: [8] Britton KA, Fox CS. Ectopic Fat Depots and Cardiovascular Disease. Circulation 2011;124:e837-e841. [9] Dey D, Nakazato R, Li D, Berman DS. Epicardial and thoracic fat - Noninvasive measurement and clinical implications. Cardiovasc Diagn Ther 2012;2(2):85-93

21 dd [10] Dey D, Wong ND, Tamarappoo B, Nakazato R, Gransar H, Cheng VY et al. Computeraided non-contrast CT-based quantification of pericardial and thoracic fat and their associations with coronary calcium and metabolic syndrome. Atherosclerosis 2010;209: [11] Iozzo P. Myocardial, Perivascular, and Epicardial Fat. Diabetes Care May; 34 Suppl 2: S [12] Rosito GA, Massaro JM, Hoffmann U, Ruberg FL, Mahabadi AA, Vasan RS et al. Pericardial fat, visceral abdominal fat, cardiovascular disease risk factors, and vascular calcification in a community-based sample: the Framingham Heart Study. Circulation. 2008;117: [13] Ding J, Hsu F-C, Harris TB, Liu Y, Kritchevsky SB, Szklo M et al. The association of pericardial fat with incident coronary heart disease: the Multi-Ethnic Study of Atherosclerosis (MESA). Am J Clin Nutr. 2009;90: [14] Mahabadi AA, Massaro JM, Rosito GA, Levy D, Murabito JM, Wolf PA, et al. Association of pericardial fat, intrathoracic fat, and visceral abdominal fat with cardiovascular disease burden: the Framingham Heart Study. Eur Heart J 2009;30: [15] Cheng VY, Dey D, Tamarappoo B, Nakazato R, Gransar H, Miranda-Peats R et al. Pericardial fat burden on ECG-gated noncontrast CT in asymptomatic patients who subsequently experience adverse cardiovascular events. JACC Cardiovasc Imaging 2010 Apr;3(4): [16] Thanassoulis G, Massaro JM, O Donnell CJ, Hoffmann U, Levy D, Ellinor PT et al. Pericardial fat is associated with prevalent atrial fibrillation: the Framingham Heart Study. Circ Arrhythm Electrophysiol. 2010;3:

22 dd [17] Wong CX, Abed HS, Molaee P, Nelson A, Brooks A, Sharma G et al. Pericardial fat is associated with atrial fibrillation severity and ablation outcome. J Am Coll Cardiol 2011;57: [18] Batal O, Schoenhagen P, Shao M, Ayyad AE, Van Wagoner DR, Halliburton SS et al. Left atrial epicardial adiposity and atrial fibrillation. Circ Arrhythm Electrophysiol 2010;3: [19] Al Chekakie MO, Welles CC, Metoyer R, Ibrahim A, Shapira AR, Cytron J et al. Pericardial Fat Is Independently Associated With Human Atrial Fibrillation. J Am Coll Cardiol 2010;56: [20] Kourliouros A, Karastergiou K, Nowell J, Gukop P, Hosseini MT, Valencia O et al. Protective effect of epicardial adiponectin on atrial fibrillation following cardiac surgery. Eur J Cardiothorac Surg 2011;39(2): [21] Nakai T, Lee RJ, Schiller NB, Bellows WH, Dzankic S, Reeves J 3rd et al. The relative importance of left atrial function versus dimension in predicting atrial fibrillation after coronary artery bypass surgery. Am Heart J 2002 Jan;143(1): [22] Aviles RJ, Martin DO, Apperson-Hansen C, Houghtaling PL, Rautaharju P, Kronmal RA et al. Inflammation as a risk factor for atrial fibrillation. Circulation. 2003; 108: [23] Fox CS, Gona P, Hoffmann U, Porter SA, Salton CJ, Massaro JM et al. Pericardial fat, intrathoracic fat, and measures of left ventricular structure and function: the Framingham Heart Study. Circulation. 2009;119: [24] Cummings JE, Gill I, Akhrass R, Dery MA, Biblo LA, Quan KJ. Preservation of the anterior fat pad paradoxically decreases the incidence of postoperative atrial fibrillation in humans. J Am Coll Cardiol 2004;43:

23 dd [25] Lupinski RW. Aortic fat pad and atrial fibrillation: cardiac lymphatics revisited. ANZ J Surg 2009;79:70-74.

24

25