Eur Radiol DOI 10.1007/s00330-015-3643-1

CARDIAC

Increased epicardial fat is independently associated with the presence and chronicity of atrial fibrillation and radiofrequency ablation outcome Jadranka Stojanovska & Ella A. Kazerooni & Mohamad Sinno & Barry H. Gross & Kuanwong Watcharotone & Smita Patel & Jon A. Jacobson & Hakan Oral

Received: 30 June 2014 / Revised: 17 January 2015 / Accepted: 30 January 2015 # European Society of Radiology 2015

Abstract Objective To determine whether intrathoracic fat volumes are associated with presence and chronicity of atrial fibrillation (AF) and radiofrequency ablation (RFA) treatment outcome. Methods IRB approval was obtained and patient consent was waived for this HIPAA-compliant retrospective study. 169 patients with AF (75 non-paroxysmal and 94 paroxysmal) and 62 control patients underwent cardiac CT examination. Extrapericardial (EPFV) and epicardial fat volumes (EFV) were measured on CT, the sum of which is the total intrathoracic fat volume. Associations between these three fat volumes and presence and chronicity of AF, and outcome after RFA, were evaluated using logistic regression analysis. Results EFV was significantly associated with presence [OR 1.01 (95 % CI 1.003–1.03), p=0.01], chronicity of AF [1.008

J. Stojanovska (*) : E. A. Kazerooni : B. H. Gross : S. Patel Department of Radiology, Division of Cardiothoracic Radiology, University of Michigan Health System, UH B1-132 Taubman/Box 0302, 1500 East Medical Center Drive, Ann Arbor, MI 48109-0302, USA e-mail: [email protected] E. A. Kazerooni e-mail: [email protected] B. H. Gross e-mail: [email protected] S. Patel e-mail: [email protected] M. Sinno : H. Oral Department of Internal Medicine, Cardiovascular Medicine, Electrophysiology Laboratory, Cardiovascular Center, University of Michigan Health System, 1500 East Medical Center Drive, Ann Arbor, MI 48109, USA

(1.001–1.020), p=0.03] and AF recurrence after RFA [1.009 (1.001–1.01), p=0.02] after adjustment for age, gender and BMI. Patients with a larger EFV had a shorter time to AF recurrence (p=0.017) and a higher rate of recurrence (54 % vs 46 %) (p=0.002) after RFA. EPFV had no significant associations. Conclusion Increased epicardial fat is associated with the presence and chronicity of AF, a higher probability of AF recurrence after RFA and a shorter AF-free interval. Key Points • Increased epicardial fat is associated with presence and chronicity of atrial fibrillation • Extensive epicardial fat is associated with earlier recurrences of AF after ablation • Extensive epicardial fat may reduce transmurality of ablation by affecting current dynamics

M. Sinno e-mail: [email protected] H. Oral e-mail: [email protected]

K. Watcharotone Michigan Institute for Clinical & Health Research (MICHR), University of Michigan, 2800 Plymouth Road, Building 400, Ann Arbor, MI 48109-2800, USA e-mail: [email protected]

J. A. Jacobson Department of Radiology, Division of Musculoskeletal Radiology, University of Michigan Health System, 1500 East Medical Center Drive, Ann Arbor, MI 48109, USA e-mail: [email protected]

Eur Radiol

Keywords Epicardial fat volume . Intrathoracic fat volume . Atrial fibrillation . Radiofrequency ablation . Computed tomography

Abbreviations AF atrial fibrillation BSA body surface area CAD coronary artery disease CCTA coronary computed tomography angiogram CVD cardiovascular disease ECG electrocardiographic EFV epicardial fat volume EPFV extrapericardial fat volume ITFV intrathoracic fat volume LA left atrium LV left ventricle PV pulmonary vein RFA radiofrequency ablation ROI region of interest

Methods Study subjects Institutional review board approval was obtained with a waiver of informed consent for this HIPAA-compliant retrospective cohort study. A total of 180 consecutive patients with AF who underwent CT to evaluate the left atrium (LA) and pulmonary vein (PV) anatomy for RFA planning from 2009 to 2011 were identified. Eleven patients with AF who did not undergo RFA because of anticoagulation status, left atrial appendage thrombus and/or anaesthesia issues were excluded. The control group consisted of 62 consecutive patients with no history of AF who had a normal coronary CT angiogram (CCTA) for evaluation of chest pain, and no cardiac events for at least 6 months after CCTA. The patient enrolment process is summarized in Fig. 1.

Study design

Introduction The prevalence and incidence of atrial fibrillation (AF) increase with age and the presence of cardiovascular disease (CVD) [1], the latter being the leading cause of mortality among men and women in the USA [2]. Atrial fibrillation has been classified as paroxysmal atrial fibrillation, self-terminating that can last up to 7 days, and non-paroxysmal atrial fibrillation that includes persistent atrial fibrillation that sustains beyond 7 days but can be terminated by cardioversion, and/or permanent atrial fibrillation that persists longer than 1 year [3]. Radiofrequency ablation (RFA) is the preferred therapeutic method for the treatment of atrial fibrillation in patients who do not respond to antiarrhythmic drugs [4]. Patients with excessive abdominal visceral fat have an increased risk of CVD through numerous pathologic conditions, including hypertension, dyslipidaemia, left ventricular dysfunction and coronary artery disease (CAD) [5, 6]. Body mass index (BMI) as a measure of general obesity is an important risk factor for AF, mediated mainly by left atrial dilatation [7]. Recently, pericardial fat as an index of visceral obesity has been shown to be independently associated with AF [8–12]. Computed tomography (CT) angiography can be used to accurately quantify epicardial fat volume [8, 13]. We hypothesized that increased epicardial fat measured from CT examinations is associated with the presence and chronicity of AF, and AF recurrences after RFA.

Intrathoracic fat volumes were quantified from CT using semiautomatic segmentation fat analysis software in this retrospective cohort study to determine (1) the clinical correlates of intrathoracic fat volume, (2) the relationship between intrathoracic fat and AF chronicity and (3) the association between intrathoracic fat and RFA outcome.

Clinical characteristics Height and weight were recorded. Normal weight was defined as a BMI between 18.5 and 25 kg/m2, overweight between 25 and 30 kg/m2 and obese as over 30 kg/m2 [14]. Body surface area (BSA) was calculated using the Mosteller formula [15]. Blood pressure and heart rate during the CT were recorded. Hypertension was defined as systolic blood pressure of at least 140 mmHg, diastolic blood pressure of at least 90 mmHg or ongoing antihypertensive treatment [16]. CAD was defined as a history of myocardial infarction, positive stress test, CCTA with at least one regional stenosis greater than 50 %, or any angina [17]. Dyslipidaemia was defined as blood cholesterol level of at least 240 mg/dL, serum triglyceride level of at least 150 mg/dL or ongoing lipid-lowering treatment [18]. Diabetes mellitus was defined as fasting plasma glucose level of at least 126 mg/dL and/or currently taking antidiabetic medications [19]. In the AF cohort, LA diameter (anterior-posterior dimension on parasternal long axis) and left ventricular ejection fraction (modified Simpson method) were recorded from the pre-RFA echocardiogram. Any prior history of cardioversion or antiarrhythmic drug therapy was noted.

Eur Radiol Fig. 1 Flowchart summarizes the patient enrolment process. AF atrial fibrillation, LA left atrium, PV pulmonary vein, RFA radiofrequency ablation

CT technique After a noncontrast localizer scan, all patients underwent an electrocardiographic (ECG)-gated 64-detector CT scan (Lightspeed VCT; GE Healthcare, Milwaukee, WI) in a single breath hold at end-inspiration, with arms elevated above the head, and scan performed in a craniocaudal direction following administration of iodinated intravenous contrast through a cannula in an antecubital vein using a power injector. AF patients underwent a CT for evaluation of the LA and PVs, and control subjects underwent CCTA. For dedicated left atrium–pulmonary vein CT, a timing bolus with 15 mL nonionic intravenous contrast (iodixanol, GE Healthcare, Canada) injected at 4 mL/s followed by a 20-mL saline bolus with a region of interest (ROI) placed in the LA to determine scan delay was used. Then, 100 mL

iodixanol was administered followed by 50 mL saline at 4 mL/s. Imaging was performed with prospective ECG triggering using the following parameters: 0.35 s gantry rotation speed, 1.25 mm slice thickness and reconstruction interval, 100–120 kVp depending on patient weight, ECG-modulated mA (min/max of 100/400–500) and a 25-cm field of view. Craniocaudal coverage extended from 2 cm above the lung apices to 2 cm below the cardiac apex. For an average-sized male adult, radiation exposure was approximately 12 mSv. For dedicated CCTA, patients with a heart rate over 65 beats per minute received 100 mg metoprolol orally 60 min before CT. All patients received one puff of sublingual nitroglycerin 1–5 min before scanning. Craniocaudal coverage extended from 2 cm above the most cephalad coronary artery to 2 cm below the cardiac apex. Scan parameters included 0.35 s gantry rotation speed, 0.625 mm slice

Eur Radiol

thickness, 100–120 kVp, adjusted for patient size, and isoosmolar contrast (Visipaque 370; GE Healthcare, Milwaukee, Wisconsin). A timing bolus of 15 mL contrast material was injected at 5 mL/s with an ROI in the aortic root with the scan delay time calculated as peak enhancement plus 6 s. CCTA was acquired using a triphasic contrast bolus with 50 mL of contrast material, followed by 30 mL contrast material diluted with 30 mL normal saline, followed by 50 mL of normal saline, all at 5 mL/s. Examinations were performed using retrospective gating with tube current modulation (100 % peak tube current during end-diastole and up to 80 % reduction at end-systole). CT image reconstruction and post-processing Reconstructed images for both groups were post-processed on the same off-line workstation (AW Workstation version 4.5, GE Healthcare, Milwaukee, WI) using the semiautomated fat segmentation Reformat software tool (GE Healthcare, Milwaukee, Wisconsin). Ten to fourteen 10-mm-thick contiguous axial sections were obtained with coverage extending from the superior aspect of the left atrial appendage through the cardiac apex. Fat volumes were obtained by tracing the intrathoracic/ lung interface and pericardial areas in a systematic fashion, and data elaborated using a histogram-based statistical program according to the method described by Borkan et al. [20]. Fat compartments were measured using semiautomatic segmentation. The intrathoracic fat measurements were performed by two independent operators. Bland–Altman analysis demonstrated no inter-reader variability in intrathoracic fat measurements for control (n=62) subjects (p=0.33 for epicardial fat volume). The intrathoracic fat volume (ITFV) represented adipose tissue from the superior aspect of the left atrial appendage through the cardiac apex. The ROI was traced to identify intrathoracic fat by excluding the lungs, skeleton and subcutaneous tissue. ITFV was calculated in cubic centimetres using an attenuation range of −250 to −50 HU [19]. Next, a series of ROIs manually tracing the pericardium were performed on all slices. The epicardial fat volume (EFV) was defined as fat inside the tracing. Subtracting EFV from ITFV generated the extrapericardial fat volume (EPFV). The EFV included fat within the myocardium, since anatomically there is no separate fascia between the epicardium and myocardium. This methodology of obtaining fat measurements is in agreement with defined anatomy of the epicardium and pericardium [21] (Fig. 2). RFA and clinical follow-up All antiarrhythmic drugs were discontinued 5 half-lives prior to RFA except for amiodarone, which was continued for 2 months. After RFA, patients resumed their pre-ablation

Fig. 2 Contrast-enhanced axial cardiac CT outlining the thoracic fat compartments (arrows). CT computed tomography, LA left atrium, LV left ventricle, RA right atrium, RV right ventricle

antiarrhythmic regimen for up to 3 months. All patients received anticoagulation for at least 3 months after RFA. After the trans-septal puncture, antral PV isolation was performed using a 3-dimensional electroanatomical mapping system as described previously [22]. All PVs were mapped with a circular multipolar catheter. Ablation of complex fractionated atrial electrograms or linear ablation was performed at the operators’ discretion. ECG monitoring was performed during an overnight hospital stay after RFA. All patients were seen in clinic 3 months after RFA or sooner. Long-term follow-up was performed at 6- to 12month intervals. Six months after ablation, rhythm status was determined using an auto-triggered event monitor for 3 weeks. Freedom from AF was defined as absence of any atrial arrhythmias for more than 30 s off antiarrhythmic drug therapy. Repeat RFA was offered to patients with recurrent AF or atrial flutter. The mean duration of follow-up was 33± 9 months after the last RFA. Statistical analysis Continuous variables are expressed as mean±SD or median and interquartile range (IQR) where appropriate. Categorical variables are presented as percentages. Comparison between continuous variables was performed using the Student t test or Wilcoxon rank sum test where appropriate. Dichotomous variables were compared using the chi-square test. Multiple general linear model (GLM) analysis was used to study the associations of ITFV and baseline characteristics in control and AF groups. Logistic regression analysis was used to study the association between ITFVs and the presence and severity of AF, and also the likelihood of recurrence after RFA, adjusted for age, gender and BMI. In addition, backward stepwise elimination of variables was used to develop the

Eur Radiol

logistic regression model. The variables included all baseline characteristics (age, gender, BMI, LA diameter, CAD, hyperlipidaemia and diabetes in addition to ITFVs) with a p value of 0.15 or less. All variables were tested for interaction. The Cox proportional hazards regression method was used to study the relationship of EFV and AF recurrence in months after first RFA. A p value less than 0.05 was considered statistically significant. Computations were performed with SAS/STAT (version 9.3, SAS Institute Inc., Cary, NC).

Results Baseline characteristics A total of 231 patients formed the study population: 73 % (n=169) with AF and 27 % (n=62) a control group with no history of AF and free of cardiovascular disease. Of the 169 AF patients, 56 % (n=94) had paroxysmal AF and 44 % (n=75) had non-paroxysmal AF. The mean age was 59±11 years and 25 % (n=40) were women. CAD and hypertension were present in 20 % (n=33) and 53 % (n=89) respectively. None had significant mitral valve disease. The mean age of the control group was 48±11 years, and 53 % (n=33) were women. Clinical characteristics of AF and control patients are shown in Tables 1 and 2.

Table 2 Study sample characteristics of the patients with paroxysmal and non-paroxysmal AF Paroxysmal AF Non-paroxysmal P value* (n=94) AF (n=75) Age, years Women, % (n) Intrathoracic fat, cm3 Epicardial fat, cm3 Extrapericardial fat, cm3 Weight, lbs Height, in BMI, kg/m2 LA diameter, mm

60, 53–66 32 (30) 178, 116–277 70, 52–102 109, 64–168 201, 170–230 70, 66–72 29, 26–33 42, 38–45†

61, 54–69 13 (10) 261, 164–371 91, 63–148 168, 97–250 216, 190–259† 71, 68–74 31, 28–36 46, 43–50†

LVEF, % Obesity, % CAD, % (n) Hypertension, % (n) Diabetes, % (n) Hypercholesterolemia, % (n)

60, 55–65‡ 41 (39) 16 (15) 48 (45) 12 (11) 54 (51)

60, 55–65‡ 56 (42) 24 (18) 59 (44) 15 (11) 60 (45)

0.24 0.006S 0.0009S 0.0009S 0.002S 0.003S 0.01S 0.05S