1141

Complex Fractionated Atrial Electrograms Related to Left Atrial Wall Thickness JIN WI, M.D.,∗ ,‡ HYE-JEONG LEE, M.D., Ph.D.,†,‡ JAE-SUN UHM, M.D.,∗ JONG-YOUN KIM, M.D.,∗ HUI-NAM PAK, M.D., Ph.D.,∗ MOONHYOUNG LEE, M.D., Ph.D.,∗ YOUNG JIN KIM, M.D., Ph.D.,†,§ and BOYOUNG JOUNG, M.D., Ph.D.∗ ,§ From the ∗ Division of Cardiology, Department of Internal Medicine; and †Department of Radiology, Research Institute of Radiological Science, Yonsei University College of Medicine, Seoul, Republic of Korea

CFAE and LA Wall Thickness. Introduction: The mechanism of complex fractionated atrial electrogram (CFAE) in patients with atrial fibrillation (AF) remains controversial. This study investigated the relationship between CFAE and left atrial (LA) wall thickness. Methods and Results: LA muscular wall thickness (excluding fat) was measured by cardiac computed tomography in 31 patients with AF (12 paroxysmal, 19 persistent) prior to catheter ablation procedures. Measurements were performed at 31 distinct LA locations: 3 at roof, 3 at floor, 9 at anterior wall, 9 at posterior wall, 3 at lateral wall, 3 at septum, and 1 at base of the anterior appendage. The range of LA wall thickness (LAWT) varied widely (average 2.4 ± 0.4 mm, range 1.5–3.1 mm) between patients. In addition, there were significant regional differences in LAWT. Each patient had an average of 7.3 ± 3.5 CFAE sites. The LA wall was thicker at CFAE sites (227 sites, 3.0 ± 1.0 mm) than at non-CFAE sites (734 sites, 2.2 ± 0.9 mm, P < 0.001). In 23 of 31 (74%) patients, the LA wall was thicker at CFAE area than at non-CFAE area. There was no difference in LAWT between sites where CFAE vanished and those where CFAE persisted after pulmonary vein isolation (PVI) among sites with CFAE before PVI. The LAWT > 2.5 mm predicted CFAE with a sensitivity of 70% and a specificity of 70%. Conclusion: The LAWT correlates well with CFAE areas, suggesting that one of the mechanisms of CFAE might be related to LAWT. (J Cardiovasc Electrophysiol, Vol. 25, pp. 1141-1149, November 2014) atrial fibrillation, atrium, catheter ablation, complex fractionated atrial electrogram, pulmonary vein isolation Introduction It is well recognized that pulmonary vein isolation (PVI) alone is insufficient for the treatment of most persistent atrial fibrillation (AF) cases.1-3 Nademanee et al. were the first to demonstrate that complex fractionated atrial electrogram (CFAE) could be used to target sites for successful ablation of AF.4 Since this seminal report, CFAE ablation has been incorporated into a range of ablation procedures for the treatment of both paroxysmal AF (PAF) and persistent AF (PeAF) with varying success.1-3 In clinical settings, CFAE was originally defined as complex electrograms with 2 or more deflections, continuous ‡J. Wi and H.-J. Lee contributed equally to this work. §Co-correspondence. This study was supported in part by research grants from the Korean Heart Rhythm Society (2011–3), the Basic Science Research Program through the National Research Foundation of Korea, funded by the Ministry of Education, Science and Technology (NRF-2010–0021993, NRF2012R1A2A2A02045367), and the Korean Healthcare Technology R&D Project, Ministry of Health & Welfare (HI12C1552). No disclosures. Address for correspondence: Boyoung Joung, M.D., Ph.D., 50 Yonsei-ro, Seodaemun-gu, Seoul 120-752, Republic of Korea. Fax: +82-2-393-2041; E-mail: [email protected]; Young Jin Kim, M.D., Ph.D., 50 Yonsei-ro, Seodaemun-gu, Seoul 120-752, Republic of Korea. Fax: +82-2-393-3035; E-mail: [email protected] Manuscript received 3 March 2014; Revised manuscript received 5 June 2014; Accepted for publication 9 June 2014. doi: 10.1111/jce.12473

activity, and cycle length (CL) < 120 ms.4 Nevertheless, the mechanism of CFAE in patients with AF remains controversial. Anisotropy or summation of electrograms from overlapping layers of myocardial fibers, autonomic nerves, and rotors might be related to CFAE.5,6 Other studies suggest that CFAE might be unrelated to the primary arrhythmia mechanism and simply represent transient pivoting, wave front collision, or sink-to-source mismatch.7 Most importantly, fractionation is significantly reduced after PVI, with prolongation of AF CL underscoring its passive role.8 Therefore, CFAE sites do not necessarily identify regions of atrial substrate abnormality. The present study aims to identify the relationship between sites with CFAE and the left atrial (LA) wall thickness (LAWT). In addition, we evaluate the change of CFAE after PVI according to LAWT.

Methods Study Population The present study includes a total of 31 patients with nonvalvular AF (25 men, mean age 56.8 ± 9.1 years) who have underwent radiofrequency catheter ablation (RFCA) for drug refractory AF between from August 2011 and August 2012. All patients maintained optimal anticoagulation levels (target INR 2.0–3.0) before the procedure and antiarrhythmic drugs were discontinued for at least 5 half-lives of each drug and for at least 4 weeks especially in amiodarone. We used cardiac computed tomography (CT) to visually define the anatomy of each patient’s LA before RFCA.

1142

Journal of Cardiovascular Electrophysiology

Vol. 25, No. 11, November 2014

Figure 1. Schematic representation of the 31 preselected left atrial (LA) locations, in which LA thickness was measured. Left panel: anterior view; right panel: posterior view. Lower panel, 31 locations including 3 at the atrial roof, 1 at the anterior appendage base, 9 at the anterior wall, 9 at the posterior wall, 3 at the atrial floor, 3 at the lateral wall, and 3 at the septal sites. For a high quality, full color version of this figure, please see Journal of Cardiovascular Electrophysiology’s website: www.wileyonlinelibrary.com/journal/jce

This study protocol was approved by the Institutional Review Board of Severance Hospital, Yonsei University Health System. All patients provided written informed consent. Cardiac CT Contrast-enhanced cardiac CT was performed with a dualsource CT scanner (Somatom Definition Flash, Siemens Healthcare, Forchheim, Germany) within 2 days prior to the ablation procedure. Contrast (Iopamiro 370, Bracco, Milan, Italy) was injected into the antecubital vein at a flow rate of 5 mL/s using triple-phase method (60–80 mL of pure contrast, 30 mL of 70%:30% saline-to-contrast mixture, and 20 mL of pure saline) and the scan delay time was determined by the test-bolus technique. Scanning was performed using the following parameters: prospective electrocardiogram (ECG)-gated axial acquisition targeting end-systolic phase using the absolute delay method,9 80–120 kVp with 280– 450 mAs, and a 64 × 0.6-mm slice collimation. The cardiac CT was reconstructed with slice thickness of 0.75 mm and an increment interval of 0.5 mm. Cardiac CT Image Analysis CT images were reviewed using Aquaris Intuition software version 4.4.6 (Terarecon, San Francisco, CA, USA). Two radiologists (5 and 8 years of experience in cardiac CT), blinded to patient clinical and electrophysiological data, evaluated CT images independently. The image quality of the cardiac CT for LAWT was classified as good (no or minor artifact, good diagnostic quality), fair (moderate ar-

tifacts, acceptable for diagnosis), or poor (severe artifact impairing accurate evaluation). The presence of irregular heart rhythm on ECG during acquisition of cardiac CT was evaluated. The LAWT measurements were obtained in the 31 preselected locations agreed upon by both radiologists and cardiologists (Fig. 1). Observers divided the LA surface on the 3-dimensional volume rendered (3D VR) CT by the equal virtual grids and determined a reference point at the center of each region in every patient. For measurement, we determined a point in the LA wall on the 3D VR, axial, coronal, and sagittal images using a cross cursor and marked an arrow and number at the point on the 3D VR images (Fig. 2A). Afterwards, the observers measured the LAWT on the multiplanar reformatted axial, coronal, and sagittal images to obtain correct perpendicular measurements, and used semiautomated measurement through a histogram to determine inner and outer border points of the LA wall accurately (Fig. 2B). At the determined point, 2 short lines (the vertical red lines) across the epicardial fat and LA wall, and across the LA wall and LA cavity were drawn with CT histograms plotting the CT attenuation (the 2 inset graphs) corresponding to each line. Then the border points of the epicardial fat and LA wall and the LA wall and LA cavity (the white dots on the red line) were automatically drawn after being derived by the CT attenuation difference. Through this method, observers could determine the outer margin and inner margin of the LA wall from the determined points. Subsequently, the thickness of the LA wall was measured automatically from the outer margin to the inner margin through a line segment

Wi et al. CFAE and LA Wall Thickness

1143

Figure 2. Measurement of left atrial (LA) wall thickness. (A) Determination of a point in the LA wall on the 3-dimensional volume rendered, axial, coronal, and sagittal images. (B) The thickness of the LA wall measured semiautomatically from the outer to inner margins through a histogram and a line segment tool using the software. In inset graphs, the white numbers at the y-axis signify the extreme CT numbers of the 2 short lines across the epicardial fat and LA wall and the LA wall and LA cavity. The red numbers signify the median CT numbers from the full width at half maximum method, which corresponded with the white dots on the line across the LA wall. For a high quality, full color version of this figure, please see Journal of Cardiovascular Electrophysiology’s website: www.wileyonlinelibrary.com/journal/jce

tool available from the used software. We assessed the wall thickness in the manner that 5 points within 5 mm of each reference point were measured and the average was calculated. Finally, fused images were made with semitransparent mode and the CFAE map was overlaid on the cardiac CT 3D VR image marked with the 31 preselected locations (Fig. 3). Two cardiologists determined the presence of CFAE at the 31 preselected locations on the fused images.

Electrophysiological Mapping and Assessment of CFAE Intracardiac electrograms were recorded using Prucka CardioLabTM Electrophysiology system (General Electric Health Care System Inc., Milwaukee, WI, USA). After double transseptal punctures, we generated 3D VR cardiac CT merged 3D electroanatomical mapping (Ensite NavX system, St. Jude Medical Inc., Minneapolis, MN, USA). The

1144

Journal of Cardiovascular Electrophysiology

Vol. 25, No. 11, November 2014

Figure 3. The summation image of complex fractionated atrial electrogram (CFAE) map and 3-dimensional computed tomography. (A) Anteroposterior view of CFAE map. Sites with mean AF cycle length (CL) ࣘ120 ms are color coded red/white, and sites with mean AF CL >120 ms are color coded purple. Representative electrograms are shown below CFAE map. (B) The summation image. For a high quality, full color version of this figure, please see Journal of Cardiovascular Electrophysiology’s website: www.wileyonlinelibrary.com/journal/jce

bipolar electrograms were filtered from 32 to 300 Hz. AF was spontaneously induced prior to CFAE mapping without isoproterenol. The assessment of CFAE during AF was performed using the CL-based automated CFAE algorithm (Ensite NavX system, St. Jude Medical Inc.), chosen because it is now widely used in clinical settings, including an ongoing large outcome study.3 CFAEs were defined as those with CL < 120 ms.4 In brief, the algorithm utilizes a peak–peak sensitivity threshold manually set to exclude baseline noise from the analysis. We set the threshold to 0.05 mV, just above the baseline noise in our laboratory. A user-defined refractory period of 50 ms was used to avoid double counting of signals during atrial refractoriness. A maximum electrogram width of 10 ms from peak-peak was also used to exclude broad far-field signals from analysis.8 The algorithm measures the time between discrete deflections in local electrogram over 5 seconds (based on selectable width and peak-to-peak [>0.03 mV] criteria). The mean CL of the local electrogram is projected onto the LA shell as a colorcoded display. Ablation technique is described in the online supplementary. Statistical Analysis Continuous variables are shown as mean ± SD unless otherwise specified. Categorical data are expressed as absolute values and percentages. Continuous variables were

compared using a Student’s t-test. Categorical variables were compared by χ 2 or Fisher’s exact test. We used the intraclass correlation coefficient (ICC) (0–0.20, poor; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, good; and 0.81–1.00, excellent agreement) to evaluate consistency between 2 observers for measurement of LAWT. Receiver operating characteristic (ROC) curves were used to determine the accuracy of a variable in predicting CFAE. Statistical significance was established at P < 0.05 and all statistical analyses were performed using SPSS version 18.0 (SPSS Inc., Chicago, IL, USA). Results Baseline Characteristics of Patients Table 1 shows the baseline clinical characteristics. Of the 31 enrolled patients, 12 (39%) and 19 (61%) had PAF and PeAF, respectively. Among 12 patients with PAF, initial rhythm was AF in 7 patients (58%) and AF was spontaneously induced in 5 patients (42%). The mean LA size was 44.9 ± 4.8 mm and the mean left ventricular ejection fraction was 63.9 ± 7.1%. Average 369 ± 45 CFAE mapping points were taken in each patient. We took at least 10 CFAE mapping points within 10 mm around each of the 31 preselected locations. These CFAE mapping points were well-distributed all around the LA. Among 31 preselected LA sites, the average number of

Wi et al. CFAE and LA Wall Thickness

1145

TABLE 1 Comparison of Baseline Characteristics According to AF Type Variables Male Age (years) Hypertension Diabetes Heart failure Stroke CHADS2 score LA size (mm) LVEF (%) LA wall thickness (mm) CFAE sites in each patient

Total (n = 31)

PAF (n = 12)

PeAF (n = 19)

P value

25 (81) 56.8 ± 9.1 15 (54) 6 (21) 0 2 (7) 0.94 ± 0.93 45 ± 5 64 ± 7 2.4 ± 1.0 7.2 (23)

9 (75) 57.0 ± 9.7 7 (58) 2 (17) 0 1 (8) 0.92 ± 0.90 43 ± 3 63 ± 7 2.3 ± 1.0 6.3 (20)

16 (84) 56.7 ± 9.0 11 (58) 5 (26) 0 1 (5) 0.95 ± 0.97 46 ± 5 65 ± 8 2.4 ± 1.0 7.8 (25)

0.65 0.94 0.98 0.68 – 1.00 0.94 0.01 0.47 0.06 0.10

Numbers in parentheses are percentages. CFAE = complex fractionated atrial electrogram; LA = left atrium; LVEF = left ventricular ejection fraction; PAF = paroxysmal atrial fibrillation; PeAF = persistent atrial fibrillation.

Figure 4. The example of the relationship between complex fractionated atrial electrogram (CFAE) and left atrial (LA) wall thickness. (A) CFAE map of LA at baseline. (B) Computed tomography image. CFAEs are mainly observed at the appendage base (#1) and the left superior anterior wall (#5), which are thicker than the middle (#6) or right superior anterior wall (#7) without CFAE. Ao = aorta; LA = left atrium; LAA = left atrial appendage; S = spine. *Means the corresponding point, where we measured the wall thickness on the CT. For a high quality, full color version of this figure, please see Journal of Cardiovascular Electrophysiology’s website: www.wileyonlinelibrary.com/journal/jce

sites with CFAE was 7.3 ± 3.5 (24%) in each patient. There was no difference in the number of CFAE sites between patients with PAF and those with PeAF. Mean radiation dose for cardiac CT was 2.9 ± 1.4 mSv. Mean heart rate during cardiac CT was 64.1 ± 11.8 bpm and irregular heart rhythm during acquisition of cardiac CT was present in 65% (20/31) of patients. The image quality of cardiac CT was good in 94% (29/31) and fair in 6% (2/31) of scans, with no images of poor quality. CFAE and LAWT The measurements by 2 independent observers showed excellent agreement (ICC = 0.984, P < 0.001). There was a large variation in LAWT between patients with an average thickness of 2.4 ± 0.4 mm (range 1.5–3.1 mm). In addition, there were significant regional differences in LAWT. The anterior appendage base had the thick-

est wall (3.4 ± 1.2 mm) and the lateral wall was thinnest (1.5 ± 0.4 mm). In a total of 961 sites among 31 patients, the mean LAWT at CFAE sites (n = 226) was greater than that of the nonCFAE sites (n = 735) (3.0 ± 1.0 vs. 2.2 ± 0.9 mm, P < 0.001). In both patients with PAF (3.2 ± 1.0 vs. 2.1 ± 0.9 mm, P < 0.001) and PeAF (3.0 ± 0.9 vs. 2.3 ± 1.0 mm, P