International Journal of Cardiology 187 (2015) 208–215
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Correlation of atrial fibrillation cycle length and fractionation is associated with atrial fibrillation free survival Pim Gal a, Andre C. Linnenbank b, Ahmet Adiyaman a, Jaap Jan J. Smit a, Anand R. Ramdat Misier a, Peter Paul H.M. Delnoy a, Jacques M.T. de Bakker b, Arif Elvan a,⁎ a b
Cardiology Department, Isala, Zwolle, The Netherlands Experimental Cardiology Department, Academic Medical Center, Amsterdam, The Netherlands
a r t i c l e
i n f o
Article history: Received 17 November 2014 Received in revised form 6 January 2015 Accepted 3 March 2015 Available online 20 March 2015 Keywords: Atrial fibrillation CFAE Ablation Atrial fibrillation cycle length Fractionation
a b s t r a c t Aims: Fractionation of electrograms in atrial fibrillation (AF) is associated with structural and electrical remodeling. We hypothesized that fractionation can also be associated with the AF cycle length (AFCL). This study was aimed at calculating the mean AFCL to fractionation correlation coefficient (mAFCC) and assessing its association with AF free survival after pulmonary vein isolation (PVI). Methods: In twenty-eight patients, 15-second electrograms during AF were recorded with a twenty-polar catheter at the left and right atrial appendages. The AFCL was determined manually and the number of activations per second was automatically calculated into a fractionation score. The correlation between AFCL and fractionation was assessed with the mAFCC. Results: Mean age was 53 ± 8 years and 86% had paroxysmal AF. 64% of patients were AF free after a median follow-up of 5.5 years. Baseline characteristics, mean AFCL and fractionation score were not associated with AF free survival after PVI. The mAFCC assessed at the left atrial appendage predicted long-term AF free survival (area under the curve: 0.871. P = 0.002), but the mAFCC recorded at the right atrial appendage did not (0.690, P = 0.131). Conclusion: The mean AFCL mAFCC recorded at the left atrial appendage was a significant predictor of long-term AF free survival. Although not a significant predictor of AF free survival, there was a significant association between mAFCC recorded at the right atrial appendage and AF free survival. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Pulmonary vein isolation (PVI) has become an important treatment modality for “lone” symptomatic atrial fibrillation (AF) [1], although not all patients remain in sinus rhythm, even after multiple PVI attempts [2–6]. Current pathophysiological insights into the mechanisms of AF suggest that this is due to a more extended AF substrate [1]. Several reports have suggested that fractionation of electrograms (EGMs) is a sign of local AF substrate [7,8] and may be associated with AF free survival after PVI. However, other studies reported that fractionation of EGMs is unstable and dynamic [9] and a reduction of AF cycle length (AFCL) was observed prior to a fractionated EGM [10–12]. We hypothesized that in tissue with minimal scar, fractionation may be functional in nature, and the local fractionation score would be correlated to the local AFCL. In contrast, a lack of correlation between AFCL and fractionation may exist in areas with structural remodeling and fibrosis. Therefore, the correlation between AFCL and fractionation may be associated ⁎ Corresponding author at: Isala, Dept of Cardiology, Dr. Van Heesweg 2, 8025 AB Zwolle, The Netherlands. E-mail address: [email protected] (A. Elvan).
http://dx.doi.org/10.1016/j.ijcard.2015.03.284 0167-5273/© 2015 Elsevier Ireland Ltd. All rights reserved.
with the atrial scar burden and may predict a lower AF free survival after PVI. The present study was aimed at delineating the AFCL to fractionation correlation and assessing its potential value as a predictor of AF free survival after PVI. 2. Methods 2.1. Patient characteristics Twenty-eight patients with symptomatic “lone” paroxysmal and persistent AF who were accepted for PVI were included in this study. Written informed consent was obtained from all patients. The local ethics approved the study, which was in accordance with the declaration of Helsinki. Patients with a left atrial dimension of N50 mm were excluded. Class I or III antiarrhythmic drugs of the Vaughan-Williams classification were discontinued N 5 half-lives before the start of the procedure, except for amiodarone, which was discontinued at least 3 months prior to the ablation procedure. β-Blockers were allowed. Patients b18 or N 65 years and patients with thyroid dysfunction were also excluded. Transesophageal echocardiography was performed routinely 1 h before ablation to assess the interatrial septum, determine
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
209
the right and left ventricular function and valvular abnormalities, and to rule out left atrial thrombus.
yielded 150 mean AFCL values. In Figs. 2 and 3, panel A, the calculation process of the mean AFCL value is displayed.
2.2. Catheter setup
2.5. Fractionation measurement
Electrophysiological study was performed under general anesthesia supervised by a cardiac anesthesiologist using propofol in a patient weight-dependent dose of 2–4 mg/kg/min. A pentapolar catheter (Josephson, Bard, USA) was positioned in the coronary sinus. After a standard transseptal puncture, a bolus of 10,000 IU of heparin was administered to prevent thromboembolism. Additional heparin was administered guided by activated clotting time, measured every 30 min, with a target activated clotting time of 300–350 s. A twentypolar steerable mapping catheter (PentaRay, Biosense Webster, Diamond Bar, CA, USA) was inserted in the left atrium via an 8.5 F sheath. The setup is shown in Fig. 1. The twentypolar mapping catheter had 5 soft radiating splines, with four 1-mm electrodes per spline and a 4–4–4-mm interelectrode spacing. The electrodes covered an area of approximately 9.6 cm2.
Commercially available algorithms for defining CFAEs differ between manufacturers and publications. It is usually [7,16] defined as (1) low-voltage multiple potential atrial signals and (2) atrial EGMs with a very short cycle length (70 to 120 ms). However, EGMs can be fractionated without meeting these criteria for CFAE. An EGM is fractionated when it shows more than one deflection per activation cycle length [8] and fractionation is considered to reflect differences in activation time within the recording area of the electrode. In this study, we aimed at an automatic quantification of fractionation into a fractionation score by using the previously mentioned custom made software based on Matlab and a commonly used algorithm for identification of CFAEs [11,17–21]. The algorithm measured the number of discrete, sharp deflections exceeding − 0.015 mV/ms per specified length of time (1 s). The peak-to-peak deflection limit was set just above noise level at 0.05 mV to avoid noise detection while allowing detection of small, fractionated complexes [22]. Blanking interval was set at 8 ms. The fractionation was corrected for the number of main deflections in the corresponding time frame. For all one-second EGMs, the mean fractionation score was calculated for all 10 bipolar EGMs separately. 15 consecutive seconds of AF EGMs were studied. Per recording site, this yielded 150 mean fractionation score values. The settings for fractionation assessment were verified by comparison of automatic to manual fractionation assessment in a subset of patients. The chosen settings gave the best match. In Figs. 2 and 3, panel B, the calculation process for fractionation is displayed.
2.3. Electrophysiological study Mapping was performed during spontaneous AF in 6 patients and pacing-induced AF in 22 patients. EGMs during AF were recorded at least 2 min after AF induction. EGMs were recorded at the left atrial appendage and at the right atrial appendage, under fluoroscopic guidance. Ten bipolar EGMs were transferred to an amplifier (LabSystem Pro, Bard, USA), amplified with a gain of 500 to 1000, and filtered with a band-pass filter set at 30 to 500 Hz. The sample frequency was 1 kHz. 2.4. AFCL measurement
2.6. Pulmonary vein isolation Data analysis was performed off-line with custom made software based on Matlab (Mathworks, Inc., Natick, Mass) [13]. This software utilizes user-defined settings to determine local activation times and were verified by the operator manually. The chosen settings were: dV/dT N 0.05 V/s, amplitude N 0.05 mV and a blanking interval of 120 ms. Previous studies showed a high interobserver correlation of up to 0.99 of manual AFCL determination [14,15]. For all one-second EGM, the mean AFCL was calculated for all 10 bipolar EGMs separately. 15 consecutive seconds of AF EGMs were studied. Per recording site, this
All patients underwent conventional point-by-point PVI. A multipolar steerable circular catheter for circumferential pulmonary vein (PV) mapping (Lasso™, Biosense Webster Inc., Diamond Bar, CA, USA) and a radiofrequency (RF) ablation catheter (Thermocool™, Biosense Webster Inc., Diamond Bar, CA, USA) were inserted transseptally into the left atrium. A left atrial electro-anatomical map was created using 3D software (Carto™, Biosense Webster Inc., Diamond Bar, CA, USA). Templates of PV signals were recorded and pacing from the coronary sinus catheter or from the mapping catheter placed in the left atrial appendage was used whenever deemed necessary to distinguish electrical PV potentials from signals generated by activity of other atrial structures. After entering the PV, the circular catheter was positioned as close as possible to the PV ostium. PVI was achieved by delivering RF energy in a point-by-point fashion to the PV antrum creating a circular ablation lesion. RF energy was applied in a temperature-control mode with a temperature setting up to 43 °C. RF energy was applied at 30 W with a flow rate of 15 ml/min or at 40 W with a flow rate of 30 ml/min, depending on the site of ablation. The endpoint of the ablation procedure was PV isolation, as documented by entrance block or dissociation between PV and atrial activation. No adenosine testing was performed. AFCL and fractionation analysis were done after the procedure and thus did not influence the PVI procedure. 2.7. Follow-up
Fig. 1. Fluoroscopic image of electrophysiological setup. A fluoroscopy image with the twentypolar catheter positioned in the left atrial appendage (PA view).
A blanking period of 3 months was defined after PVI. Patients visited the outpatient clinic at 3, 6 and 12 months after PVI. At these times a 24-hour Holter electrocardiogram was made. An attempt was made in all patients to cease AADs after the 3 month visit. After 12 months, additional clinical visits were performed as deemed necessary by the physician. All patients were interviewed telephonically at the end of the study period to assess AF symptoms and AAD use. AF recurrences were defined according to European guidelines [1].
210
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
Fig. 2. Calculation process of AFCL and fractionation. Panel A displays the AFCL calculation process. In this example, the mean AFCL in the left part of the figure is 154.4 ms and the mean AFCL in the right part of the figure is 176.3 ms. Panel B displays the fractionation score calculation process. The custom software identifies deflections that exceed the dV/dT of −0.015 V/s with a peak-to-peak detection limit of 0.05 mV, maintaining a blanking interval of 8 ms. The total number of identified deflections per second is reduced by the total number of main deflections to calculate the corrected fractionation score. In this example, the fractionation score in the left part of the figure is 38, the number of main deflections is 6, so the corrected fractionation score is 32. In the right part of the image, the fractionation score is 17, the number of main deflections is 5, so the corrected fractionation score is 12. Notice how, during the first second of the EGM, all main activations are fractionated. After 1 s, the AFCL prolongs, and immediately, the fractionation is significantly reduced. This is a typical example of AFCL-dependent fractionation. This EGM displays increased fractionation with a reduction of AFCL. The mAFCC for this patient was −0.267, and this patient remained AF free during the entire followup of 77 months. AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; EGM: electrogram; mAFCC: mean AFCL to fractionation correlation coefficient.
2.8. Statistical analysis Continuous variables are reported as mean ± standard deviation. Categorical variables are reported as number and percentage. The correlation between automatic fractionation assessment and manual fractionation assessment was performed with an intraclass correlation coefficient (ICC). EGMs of 1 s duration were recorded for 15 s. Data was retrospectively analyzed and an intraclass correlation coefficient (ICC) was used to assess the relationship between the local mean AFCL and fractionation scores. For each bipolar EGM, this yielded 15 mean AFCL and 15 fractionation scores, resulting in 10 ICCs for each of the 10 bipolar EGM recordings site per patient, which were then averaged into a mean AFCL to fractionation correlation coefficient (mAFCC). The
difference in mAFCC between pacing-induced AF and spontaneous AF as well as the difference in AFCL, fractionation and mAFCC between patients who were AF free and those who were not AF free was assessed with Student's t-test. A receiver operating characteristic analysis was used to determine the area under the curve, specificity and sensitivity of the mAFCC as a predictor of AF free survival. The optimal cut-off for the mAFCC was −0.090. A chi square test was used to determine the association between the mAFCC with a cut-off of −0.090 and AF free survival. Uni- and multivariate Cox proportional hazard models were used to determine predictors of AF free survival. Variables were entered into the model. Statistical analysis was performed in IBM SPSS statistics 20.0 (IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp). A P-value of ≤0.05 was considered statistically significant.
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
211
Fig. 3. mAFCC calculation of AFCL-independent fractionation. This figure displays the mAFCC calculation of a patient with AFCL-independent fractionation. Notice that AFCL is prolonged by 14 ms in the right part of the EGM, although the fractionation score remains virtually the same. The mAFCC of this patient was −0.006. After 2 PVIs, this patient developed persistent AF. Notice that each complex is fractionated, independent of the AFCL, which is typical for AFCL-independent fractionation. AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; mAFCC: mean atrial fibrillation cycle length to fractionation correlation coefficient.
2.9. Ethics committee The internal review board (IRB) reviewed and approved this study. 3. Results 3.1. Patient characteristics Mean age was 53 years with paroxysmal AF in 83% of patients. No patients had structural heart disease on preprocedural echocardiography. No intracardiac thrombi were identified prior to the procedure. Baseline characteristics are shown in Table 1. All patients completed the electrophysiological study and PVI. 3.2. AFCL and fractionation characteristics In total, 8400 mean AFCL values and fractionation scores were included in the analysis. EGMs of 5 (8.9%) patients were excluded from
analysis due to suboptimal quality. The mean AFCL, fractionation score and mAFCC for both sites are displayed in Table 2 and Fig. 4. The ICC between automatic and manual fractionation was 0.883 (95% CI: 0.843–0.913). There was no difference in mAFCC among induced and spontaneous AF episodes for both recording sites. However, the mAFCC recorded in the left atrial appendage was significantly stronger in paroxysmal AF patients compared to persistent AF patients (−0.185 vs. −0.009, P = 0.024). Other baseline characteristics were not associated with the mAFCC recorded in both the left and right atrial appendages. A typical example of AFCL-dependent fractionation is displayed in Fig. 2. As can be observed, the AFCL is shorter in the left part of the figure compared to the right part of the figure. The fractionation is much higher in the left part of the recording compared to the right part of the recording. The sudden transition of the AFCL lengthening corresponds to the change in degree of fractionation. In Fig. 3, a typical example of AFCL-independent fractionation is displayed. The AFCL is significantly shorter in the left part of the recording compared to the right part of
212
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
4. Discussion
Table 1 Baseline characteristics. Age (years) Gender men (%) BMI (kg/m2) LA size in PSLAX (mm) CHA2DS2VASc score 0 1 2 3–7 Co-morbidity (%) Hypertension Diabetes mellitus Previous TIA/stroke Structural heart disease Type AF: paroxysmal Failed AADs (range) Class I Class III AF duration (years) Spontaneous AF
53.1 (±8.3) 21 (75.0%) 25.6 (±2.4) 40.0 (±4.6) 18 (64.3%) 6 (21.4%) 4 (14.3%) 0 4 (14.3%) 1 (3.6%) 1 (3.6%) 0 24 (85.7%) 2 (0–3) 22 (78.6%) 21 (75.0%) 5.8 (±4.6) 22 (78.6%)
Data are presented as absolute numbers or median with their ±SD, percentages or ranges where appropriate. AADs: anti-arrhythmic drugs; AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; BMI: body mass index; LA: left atrial; PSLAX: echocardiographic parasternal long axis view; TIA: transient ischemic attack.
the recording. However, there is virtually no change in fractionation between the left and right parts of the recording.
3.3. Follow-up data After a median follow-up of 5.5 (q1–q3: 4.3–6.1) years, and a mean of 1.71 (±0.85) PVI attempts, 18 patients were free of AF (64.3%). During follow-up, 9 patients (32.1%) underwent a second PVI, 4 (14.3%) underwent a third PVI and 1 (3.6%) patient underwent a fourth PVI.
3.4. Association with AF free survival The mAFCC recorded at the left atrial appendage significantly predicted AF free survival (area under the curve (AUC) 0.871, P = 0.002), and ROC analysis revealed an optimal cut-off point of −0.090. The mAFCC recorded at the right atrial appendage was not a significant predictor of AF free survival (AUC 0.690, P = 0.131). The ROC analysis is displayed in Fig. 5. However, with a cut-off point of −0.090, there was an association between the mAFCC at both recording sites and AF free survival, as displayed in Table 3. In univariate analysis, none of the patient characteristics was associated with AF free survival. However, the mAFCC recorded at the left atrial appendage was associated with AF free survival, with a hazard ratio (HR) of 2.577 per 0.1 increase in mAFCC (P = 0.016). The mAFCC recorded at the right atrial appendage was not associated with AF free survival (HR 1.151, P = 0.608). The analysis is displayed in Table 4.
4.1. Main findings The present study reports the relationship between AFCL and fractionation and found that the mAFCC recorded at the left atrial appendage was a significant predictor of long-term AF free survival. Although not a significant predictor of AF free survival, there was a significant association between mAFCC recorded at the right atrial appendage and AF free survival. This study suggests that the mAFCC may be an important measure predictor of long term AF free survival. However, in line with other studies, the left atrium appears to be the main driver for sustaining AF. 4.2. Fractionation assessment In this study, fractionation was assessed in bipolar EGMs with settings adapted from CFAE assessment. The correlation between automatic and manual fractionation assessment was very strong. Furthermore, 3 operators agreed in consensus that bipolar EGMs yielded the most reliable results in fractionation assessment. Of note, manual assessment should not be considered a gold standard and frequently, comparing unipolar and bipolar EGMs is necessary to discern fractionated EGMs from far-field activations. 4.3. Fractionation Atrial myocardial cells usually have an effective refractory period (ERP) of 150 to 260 ms [23] during sinus rhythm. During AF, the ERP is reduced [24] and the AFCL has been shown to reflect the local atrial ERP [25]. Previous studies showed that AF is associated with an increased dispersion in atrial ERP [26–30]. EGM fractionation during AF can occur due to structural remodeling (e.g. fibrosis) [8], which is unrelated to the type of AF [31–33]. Fractionation can also be functionally determined by spatial dispersion in refractory periods [34] and areas of functional conduction block [8], as also observed by the linear association between pacing interval and conduction velocity [35]. We hypothesize that functional fractionation may be differentiated from structural fractionation by the mAFCC. Potentially, a strong correlation between AFCL and fractionation may be considered functional fractionation, limited atrial fibrosis and subsequently a higher chance of AF free survival after PVI. Conversely, a weak correlation between the AFCL and fractionation would imply structural remodeling and fibrosis, and a reduced AF free survival after PVI. The design of the 20-polar catheter used in this study allowed high resolution mapping of the LAA and RAA. Of note, the standard deviation of the ICC values was 0.15 and 0.13, respectively, implying a strong variation in ICC values between electrodes in the same patient. Hypothetically, averaging these ICC values into an mAFCC may give an estimate of the total scar burden. The scar burden has already been shown to be associated with the AF free survival after ablation. However, future studies are needed to assess the correlation between electrophysiologic findings and histopathological findings. Furthermore, the present study showed that the association
Table 2 Mean AFCL and fractionation score.
LAA
RAA
AFCL Fractionation score mAFCC AFCL Fractionation score mAFCC
Total group
AF free
Not AF free
P
170.7 ± 36.3 ms 10.4 ± 4.8 −0.149 ± 0.15 174.4 ± 38.6 ms 9.6 ± 4.2 −0.126 ± 0.14
167.5 ± 30.1 10.3 ± 4.5 −0.211 ± 0.143 176.0 ± 40.6 10.3 ± 4.8 −0.152 ± 0.137
180.0 ± 46.9 9.9 ± 5.6 −0.053 ± 0.091 166.2 ± 34.8 9.0 ± 3.0 −0.098 ± 0.146
0.772 0.664 0.001 0.557 0.734 0.141
Values are displayed their ±SD. AFCL: atrial fibrillation cycle length, mAFCC: mean AFCL to fractionation correlation coefficient. AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; LAA: left atrial appendage; RAA: right atrial appendage.
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
213
Fig. 4. Association between mAFCC and AF free survival. In the left panel, the mAFCC recorded at the left atrial appendage is displayed, and in the right panel, the mAFCC recorded at the right atrial appendage is displayed. The mAFCC recorded at the left atrial appendage is significantly stronger in patients who are AF free compared to those who are not (−0.211 vs. −0.053, P = 0.001). The mAFCC recorded at the right atrial appendage did not significantly differ among AF free and not AF free patients (−0.152 vs. −0.098, P = 0.141). AF: atrial fibrillation, AFCL: atrial fibrillation cycle length; mAFCC: mean AFCL to fractionation correlation coefficient.
Fig. 5. Sensitivity/specificity decision plot of the mean AFCL to fractionation correlation coefficient as a predictor of AF free survival. In the left panel, the mAFCC recorded at the left atrial appendage is displayed, and in the right panel, the mAFCC recorded at the right atrial appendage is displayed. The mAFCC of the left atrial appendage has an area under the curve of 0.871, P = 0.002. The mAFCC of the right atrial appendage has an area under the curve of 0.690, P = 0.131. The optimal cut-off point for both sites is −0.090. AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; mAFCC: mean AFCL to fractionation correlation coefficient.
between the mAFCC recorded in the left atrium was a predictor of AF free survival, whereas the mAFCC recorded in the right atrium was not. This suggests that the left atrium is the main driver for AF, which is in line with previous studies [36,37].
4.4. Functional nature of complex fractionated atrial electrograms Several studies investigated the dynamic nature of CFAEs. Knecht et al. [11] reported that pharmacologically induced autonomic blockade resulted in an increase in AFCL exceeding 6 ms and a decrease in CFAEs. Shan et al. [38] found that administration of the class IC drug cibenzoline led to a prolongation of AFCL and a reduction in the percentage of fractionated EGMs. Rostock et al. [10] showed that a shortening of N10 ms in AFCL precedes the occurrence of CFAEs. Additionally, an inverse correlation between CFAEs and AFCL was reported. Jadidi et al. [21] showed
Table 3 Association between mAFCC and AF free survival.
mAFCC LAA b −0.090 mAFCC LAA N −0.090 mAFCC RAA b −0.090 mAFCC RAA N −0.090
AF free
Not AF free
17 2 11 3
1 8 2 7
that CFAEs arose passively at sites of wavelet collision, and reported that these sites displayed normal bipolar voltages during sinus rhythm, suggesting the absence of local fibrosis/scar. Stiles et al. reported that increased fractionation is associated with a higher dominant frequency [39]. The present study supports the hypothesis that CFAEs can be functionally determined and suggests that CFAEs are not necessarily associated with local AF substrate. This observation may also provide an explanation for the conflicting results of CFAE ablation [40]. 4.5. Potential clinical impact Potentially, software can be designed that facilitates the on-line assessment of the mAFCC. This may impact the interpretation of fractionation and thus current clinical practice in the ablation of fractionated EGMs. Furthermore, the mAFCC may assist in a patient-tailored treatment plan when patients suffer from an AF recurrence after PVI: either a repeat PVI in the case of a strong mAFCC or attempting other treatment modalities in the case of a weak mAFCC. 4.6. Limitations
P b0.001 0.008
mAFCC: mean AF cycle length to fractionation correlation coefficient; LAA: left atrial appendage; RAA: right atrial appendage; AF: atrial fibrillation.
In the interpretation of this study, several limitations should be considered. This is a retrospective analysis in a limited number of patients. Potentially, an association between the mAFCC recorded at the right atrial appendage and AF free survival is overlooked due to the limited number of patients. EGMs of 5 patients were excluded from analysis
214
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
Table 4 Univariate and multivariate analyses of AF recurrence. Univariate analysis
P
Hazard ratio
95% CI
Multivariate analysis
P
Hazard ratio
95% CI
Age Male gender BMI LA dimension in PSLAX CHA2DS2VASc Hypertension Persistent AF Failed AADs AF duration Spontaneous AF LAA mAFCC (per 0.1 increase) RAA mAFCC (per 0.1 increase)
0.823 0.253 0.615 0.167 0.281 0.828 0.827 0.829 0.393 0.672 0.016 0.608
0.990 0.293 0.882 1.117 0.488 0.793 1.263 1.110 1.059 0.706 2.577 1.151
0.908–1.079 0.036–2.399 0.542–1.437 0.955–1.308 0.132–1.800 0.097–6.452 0.155–10.304 0.432–2.847 0.929–1.207 0.141–3.535 1.192–5.568 0.672–1.972
LA dimension in PSLAX LAA mAFCC (per 0.1 increase)
0.681 0.042
1.034 2.371
0.883–1.210 1.033–5.443
Univariate and multivariate analyses of the association between patient characteristics and AF free survival after PVI. AADs: anti-arrhythmic drugs; AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; BMI: body mass index; LA: left atrial; LAA: left atrial appendage; mAFCC: mean AFCL to fractionation correlation coefficient; PSLAX: echocardiographic parasternal long axis view; RAA: right atrial appendage. P-values between AF free and AF recurrence groups.
due to suboptimal quality. Outpatient clinic visits were only performed during the first year after ablation and at the discretion of the physician thereafter, which may have resulted in an overestimation of the AF free survival. 5. Conclusion In this study, the mAFCC recorded in the left atrial appendage was a significant predictor of long-term AF free survival. Although not a significant predictor of AF free survival, there was a significant association between mAFCC recorded at the right atrial appendage and AF free survival. Potentially, the mAFCC can be used as a predictor of AF free survival in patients undergoing PVI. Funding None. Conflict of interest None. Acknowledgments We acknowledge Ms. Petra Koopmans for the advice on the statistical analysis. We thank Ms. Vera Derks for the final preparation of this manuscript. References [1] H. Calkins, K.H. Kuck, R. Cappato, J. Brugada, A.J. Camm, S.A. Chen, et al., 2012 HRS/ EHRA/ECAS Expert Consensus Statement on Catheter and Surgical Ablation of Atrial Fibrillation: recommendations for patient selection, procedural techniques, patient management and follow-up, definitions, endpoints, and research trial design, Europace 14 (2012) 528–606. [2] H. Calkins, M.R. Reynolds, P. Spector, M. Sondhi, Y. Xu, A. Martin, et al., Treatment of atrial fibrillation with antiarrhythmic drugs or radiofrequency ablation: two systematic literature reviews and meta-analyses, Circ. Arrhythm. Electrophysiol. 2 (2009) 349–361. [3] P. Gal, A.E. Aarntzen, J.J. Smit, A. Adiyaman, A.R. Misier, P.P. Delnoy, et al., Conventional radiofrequency catheter ablation compared to multi-electrode ablation for atrial fibrillation, Int. J. Cardiol. 176 (2014) 891–895. [4] Y. Liu, H. Huang, C. Huang, S. Zhang, C. Ma, X. Liu, et al., Catheter ablation of atrial fibrillation in Chinese elderly patients, Int. J. Cardiol. 152 (2011) 266–267. [5] S. Knecht, C. Sticherling, S. von Felten, D. Conen, B. Schaer, P. Ammann, et al., Longterm comparison of cryoballoon and radiofrequency ablation of paroxysmal atrial fibrillation: a propensity score matched analysis, Int. J. Cardiol. 176 (2014) 645–650. [6] T. Pezawas, R. Ristl, M. Bilinski, C. Schukro, H. Schmidinger, Single, remote-magnetic catheter approach for pulmonary vein isolation in patients with paroxysmal and non-paroxysmal atrial fibrillation, Int. J. Cardiol. 174 (2014) 18–24.
[7] K. Nademanee, J. McKenzie, E. Kosar, M. Schwab, B. Sunsaneewitayakul, T. Vasavakul, et al., A new approach for catheter ablation of atrial fibrillation: mapping of the electrophysiologic substrate, J. Am. Coll. Cardiol. 43 (2004) 2044–2053. [8] U. Schotten, S. Verheule, P. Kirchhof, A. Goette, Pathophysiological mechanisms of atrial fibrillation: a translational appraisal, Physiol. Rev. 91 (2011) 265–325. [9] D.H. Lau, B. Maesen, S. Zeemering, S. Verheule, H.J. Crijns, U. Schotten, Stability of complex fractionated atrial electrograms: a systematic review, J. Cardiovasc. Electrophysiol. 23 (2012) 980–987. [10] T. Rostock, M. Rotter, P. Sanders, Y. Takahashi, P. Jais, M. Hocini, et al., High-density activation mapping of fractionated electrograms in the atria of patients with paroxysmal atrial fibrillation, Heart Rhythm. 3 (2006) 27–34. [11] S. Knecht, M. Wright, S. Matsuo, I. Nault, N. Lellouche, F. Sacher, et al., Impact of pharmacological autonomic blockade on complex fractionated atrial electrograms, J. Cardiovasc. Electrophysiol. 21 (2010) 766–772. [12] F. Atienza, D. Calvo, J. Almendral, S. Zlochiver, K.R. Grzeda, N. Martinez-Alzamora, et al., Mechanisms of fractionated electrograms formation in the posterior left atrium during paroxysmal atrial fibrillation in humans, J. Am. Coll. Cardiol. 57 (2011) 1081–1092. [13] M. Potse, A.C. Linnenbank, C.A. Grimbergen, Software design for analysis of multichannel intracardial and body surface electrocardiograms, Comput. Methods Programs Biomed. 69 (2002) 225–236. [14] K.C. Ravi, D.E. Krummen, A.J. Tran, J.R. Bullinga, S.M. Narayan, Electrocardiographic measurements of regional atrial fibrillation cycle length, Pacing Clin. Electrophysiol. 32 (Suppl. 1) (2009) S66–S71. [15] R.J. Hunter, I. Diab, M. Tayebjee, L. Richmond, S. Sporton, M.J. Earley, et al., Characterization of fractionated atrial electrograms critical for maintenance of atrial fibrillation: a randomized, controlled trial of ablation strategies (the CFAE AF trial), Circ. Arrhythm. Electrophysiol. 4 (2011) 622–629. [16] D. Scherr, D. Dalal, A. Cheema, A. Cheng, C.A. Henrikson, D. Spragg, et al., Automated detection and characterization of complex fractionated atrial electrograms in human left atrium during atrial fibrillation, Heart Rhythm. 4 (2007) 1013–1020. [17] A. Verma, R. Mantovan, L. Macle, G. De Martino, J. Chen, C.A. Morillo, et al., Substrate and Trigger Ablation for Reduction of Atrial Fibrillation (STAR AF): a randomized, multicentre, international trial, Eur. Heart J. 31 (2010) 1344–1356. [18] A. Verma, P. Novak, L. Macle, B. Whaley, M. Beardsall, Z. Wulffhart, et al., A prospective, multicenter evaluation of ablating complex fractionated electrograms (CFEs) during atrial fibrillation (AF) identified by an automated mapping algorithm: acute effects on AF and efficacy as an adjuvant strategy, Heart Rhythm. 5 (2008) 198–205. [19] J.F. Viles-Gonzalez, J.A. Gomes, M.A. Miller, S.R. Dukkipati, J.S. Koruth, C. Eggert, et al., Areas with complex fractionated atrial electrograms recorded after pulmonary vein isolation represent normal voltage and conduction velocity in sinus rhythm, Europace 15 (2013) 339–346. [20] A.S. Jadidi, H. Cochet, A.J. Shah, S.J. Kim, E. Duncan, S. Miyazaki, et al., Inverse relationship between fractionated electrograms and atrial fibrosis in persistent atrial fibrillation — a combined MRI and high density mapping, J. Am. Coll. Cardiol. 27 (2013) 802–812. [21] A.S. Jadidi, E. Duncan, S. Miyazaki, N. Lellouche, A.J. Shah, A. Forclaz, et al., Functional nature of electrogram fractionation demonstrated by left atrial high-density mapping, Circ. Arrhythm. Electrophysiol. 5 (2012) 32–42. [22] K.R. Grzeda, S.F. Noujaim, O. Berenfeld, J. Jalife, Complex fractionated atrial electrograms: properties of time-domain versus frequency-domain methods, Heart Rhythm. 6 (2009) 1475–1482. [23] P.R. Podrid PK, Cardiac Arrhythmia: Mechanisms, Diagnosis, and Management, Lippincott Williams & Wilkins, 2001. [24] M.C. Wijffels, C.J. Kirchhof, R. Dorland, M.A. Allessie, Atrial fibrillation begets atrial fibrillation. A study in awake chronically instrumented goats, Circulation 92 (1995) 1954–1968. [25] A. Capucci, M. Biffi, G. Boriani, F. Ravelli, G. Nollo, P. Sabbatani, et al., Dynamic electrophysiological behavior of human atria during paroxysmal atrial fibrillation, Circulation 92 (1995) 1193–1202. [26] Z. Li, E. Hertervig, J. Carlson, C. Johansson, S.B. Olsson, S. Yuan, Dispersion of refractoriness in patients with paroxysmal atrial fibrillation. Evaluation with simultaneous endocardial recordings from both atria, J. Electrocardiol. 35 (2002) 227–234.
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215 [27] Z. Li, E. Hertervig, S. Yuan, Y. Yang, Z. Lin, S.B. Olsson, Dispersion of atrial repolarization in patients with paroxysmal atrial fibrillation, Europace 3 (2001) 285–291. [28] E. Diker, M. Ozdemir, S. Aydogdu, U.K. Tezcan, S. Korkmaz, E. Kutuk, et al., Dispersion of repolarization in paroxysmal atrial fibrillation, Int. J. Cardiol. 63 (1998) 281–286. [29] H. Ramanna, R.N. Hauer, F.H. Wittkampf, J.M. de Bakker, E.F. Wever, A. Elvan, et al., Identification of the substrate of atrial vulnerability in patients with idiopathic atrial fibrillation, Circulation 101 (2000) 995–1001. [30] H. Ramanna, A. Elvan, F.H. Wittkampf, J.M. de Bakker, R.N. Hauer, E.O. Robles de Medina, Increased dispersion and shortened refractoriness caused by verapamil in chronic atrial fibrillation, J. Am. Coll. Cardiol. 37 (2001) 1403–1407. [31] N. Akoum, M. Daccarett, C. McGann, N. Segerson, G. Vergara, S. Kuppahally, et al., Atrial fibrosis helps select the appropriate patient and strategy in catheter ablation of atrial fibrillation: a DE-MRI guided approach, J. Cardiovasc. Electrophysiol. 22 (2011) 16–22. [32] A. Boldt, U. Wetzel, J. Lauschke, J. Weigl, J. Gummert, G. Hindricks, et al., Fibrosis in left atrial tissue of patients with atrial fibrillation with and without underlying mitral valve disease, Heart 90 (2004) 400–405. [33] A.W. Teh, P.M. Kistler, G. Lee, C. Medi, P.M. Heck, S.J. Spence, et al., Electroanatomic remodeling of the left atrium in paroxysmal and persistent atrial fibrillation patients without structural heart disease, J. Cardiovasc. Electrophysiol. 23 (2012) 232–238. [34] K.T. Konings, J.L. Smeets, O.C. Penn, H.J. Wellens, M.A. Allessie, Configuration of unipolar atrial electrograms during electrically induced atrial fibrillation in humans, Circulation 95 (1997) 1231–1241.
215
[35] S.L. Chang, Y.C. Chen, C.P. Hsu, Y.H. Kao, Y.K. Lin, Y.J. Lin, et al., Electrophysiological characteristics of complex fractionated electrograms and high frequency activity in atrial fibrillation, Int. J. Cardiol. 168 (2013) 2289–2299. [36] M. Haissaguerre, M. Hocini, A. Denis, A.J. Shah, Y. Komatsu, S. Yamashita, et al., Driver domains in persistent atrial fibrillation, Circulation 130 (2014) 530–538. [37] N.F. Marrouche, D. Wilber, G. Hindricks, P. Jais, N. Akoum, F. Marchlinski, et al., Association of atrial tissue fibrosis identified by delayed enhancement MRI and atrial fibrillation catheter ablation: the DECAAF study, JAMA 311 (2014) 498–506. [38] Z. Shan, P.H. Van Der Voort, Y. Blaauw, M. Duytschaever, M.A. Allessie, Fractionation of electrograms and linking of activation during pharmacologic cardioversion of persistent atrial fibrillation in the goat, J. Cardiovasc. Electrophysiol. 15 (2004) 572–580. [39] M.K. Stiles, A.G. Brooks, P. Kuklik, B. John, H. Dimitri, D.H. Lau, et al., High-density mapping of atrial fibrillation in humans: relationship between high-frequency activation and electrogram fractionation, J. Cardiovasc. Electrophysiol. 19 (2008) 1245–1253. [40] W.J. Li, Y.Y. Bai, H.Y. Zhang, R.B. Tang, C.L. Miao, C.H. Sang, et al., Additional ablation of complex fractionated atrial electrograms after pulmonary vein isolation in patients with atrial fibrillation: a meta-analysis, Circ. Arrhythm. Electrophysiol. 4 (2011) 143–148.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Correlation of atrial fibrillation cycle length and fractionation is associated with atrial fibrillation free survival Pim Gal a, Andre C. Linnenbank b, Ahmet Adiyaman a, Jaap Jan J. Smit a, Anand R. Ramdat Misier a, Peter Paul H.M. Delnoy a, Jacques M.T. de Bakker b, Arif Elvan a,⁎ a b
Cardiology Department, Isala, Zwolle, The Netherlands Experimental Cardiology Department, Academic Medical Center, Amsterdam, The Netherlands
a r t i c l e
i n f o
Article history: Received 17 November 2014 Received in revised form 6 January 2015 Accepted 3 March 2015 Available online 20 March 2015 Keywords: Atrial fibrillation CFAE Ablation Atrial fibrillation cycle length Fractionation
a b s t r a c t Aims: Fractionation of electrograms in atrial fibrillation (AF) is associated with structural and electrical remodeling. We hypothesized that fractionation can also be associated with the AF cycle length (AFCL). This study was aimed at calculating the mean AFCL to fractionation correlation coefficient (mAFCC) and assessing its association with AF free survival after pulmonary vein isolation (PVI). Methods: In twenty-eight patients, 15-second electrograms during AF were recorded with a twenty-polar catheter at the left and right atrial appendages. The AFCL was determined manually and the number of activations per second was automatically calculated into a fractionation score. The correlation between AFCL and fractionation was assessed with the mAFCC. Results: Mean age was 53 ± 8 years and 86% had paroxysmal AF. 64% of patients were AF free after a median follow-up of 5.5 years. Baseline characteristics, mean AFCL and fractionation score were not associated with AF free survival after PVI. The mAFCC assessed at the left atrial appendage predicted long-term AF free survival (area under the curve: 0.871. P = 0.002), but the mAFCC recorded at the right atrial appendage did not (0.690, P = 0.131). Conclusion: The mean AFCL mAFCC recorded at the left atrial appendage was a significant predictor of long-term AF free survival. Although not a significant predictor of AF free survival, there was a significant association between mAFCC recorded at the right atrial appendage and AF free survival. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Pulmonary vein isolation (PVI) has become an important treatment modality for “lone” symptomatic atrial fibrillation (AF) [1], although not all patients remain in sinus rhythm, even after multiple PVI attempts [2–6]. Current pathophysiological insights into the mechanisms of AF suggest that this is due to a more extended AF substrate [1]. Several reports have suggested that fractionation of electrograms (EGMs) is a sign of local AF substrate [7,8] and may be associated with AF free survival after PVI. However, other studies reported that fractionation of EGMs is unstable and dynamic [9] and a reduction of AF cycle length (AFCL) was observed prior to a fractionated EGM [10–12]. We hypothesized that in tissue with minimal scar, fractionation may be functional in nature, and the local fractionation score would be correlated to the local AFCL. In contrast, a lack of correlation between AFCL and fractionation may exist in areas with structural remodeling and fibrosis. Therefore, the correlation between AFCL and fractionation may be associated ⁎ Corresponding author at: Isala, Dept of Cardiology, Dr. Van Heesweg 2, 8025 AB Zwolle, The Netherlands. E-mail address: [email protected] (A. Elvan).
http://dx.doi.org/10.1016/j.ijcard.2015.03.284 0167-5273/© 2015 Elsevier Ireland Ltd. All rights reserved.
with the atrial scar burden and may predict a lower AF free survival after PVI. The present study was aimed at delineating the AFCL to fractionation correlation and assessing its potential value as a predictor of AF free survival after PVI. 2. Methods 2.1. Patient characteristics Twenty-eight patients with symptomatic “lone” paroxysmal and persistent AF who were accepted for PVI were included in this study. Written informed consent was obtained from all patients. The local ethics approved the study, which was in accordance with the declaration of Helsinki. Patients with a left atrial dimension of N50 mm were excluded. Class I or III antiarrhythmic drugs of the Vaughan-Williams classification were discontinued N 5 half-lives before the start of the procedure, except for amiodarone, which was discontinued at least 3 months prior to the ablation procedure. β-Blockers were allowed. Patients b18 or N 65 years and patients with thyroid dysfunction were also excluded. Transesophageal echocardiography was performed routinely 1 h before ablation to assess the interatrial septum, determine
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
209
the right and left ventricular function and valvular abnormalities, and to rule out left atrial thrombus.
yielded 150 mean AFCL values. In Figs. 2 and 3, panel A, the calculation process of the mean AFCL value is displayed.
2.2. Catheter setup
2.5. Fractionation measurement
Electrophysiological study was performed under general anesthesia supervised by a cardiac anesthesiologist using propofol in a patient weight-dependent dose of 2–4 mg/kg/min. A pentapolar catheter (Josephson, Bard, USA) was positioned in the coronary sinus. After a standard transseptal puncture, a bolus of 10,000 IU of heparin was administered to prevent thromboembolism. Additional heparin was administered guided by activated clotting time, measured every 30 min, with a target activated clotting time of 300–350 s. A twentypolar steerable mapping catheter (PentaRay, Biosense Webster, Diamond Bar, CA, USA) was inserted in the left atrium via an 8.5 F sheath. The setup is shown in Fig. 1. The twentypolar mapping catheter had 5 soft radiating splines, with four 1-mm electrodes per spline and a 4–4–4-mm interelectrode spacing. The electrodes covered an area of approximately 9.6 cm2.
Commercially available algorithms for defining CFAEs differ between manufacturers and publications. It is usually [7,16] defined as (1) low-voltage multiple potential atrial signals and (2) atrial EGMs with a very short cycle length (70 to 120 ms). However, EGMs can be fractionated without meeting these criteria for CFAE. An EGM is fractionated when it shows more than one deflection per activation cycle length [8] and fractionation is considered to reflect differences in activation time within the recording area of the electrode. In this study, we aimed at an automatic quantification of fractionation into a fractionation score by using the previously mentioned custom made software based on Matlab and a commonly used algorithm for identification of CFAEs [11,17–21]. The algorithm measured the number of discrete, sharp deflections exceeding − 0.015 mV/ms per specified length of time (1 s). The peak-to-peak deflection limit was set just above noise level at 0.05 mV to avoid noise detection while allowing detection of small, fractionated complexes [22]. Blanking interval was set at 8 ms. The fractionation was corrected for the number of main deflections in the corresponding time frame. For all one-second EGMs, the mean fractionation score was calculated for all 10 bipolar EGMs separately. 15 consecutive seconds of AF EGMs were studied. Per recording site, this yielded 150 mean fractionation score values. The settings for fractionation assessment were verified by comparison of automatic to manual fractionation assessment in a subset of patients. The chosen settings gave the best match. In Figs. 2 and 3, panel B, the calculation process for fractionation is displayed.
2.3. Electrophysiological study Mapping was performed during spontaneous AF in 6 patients and pacing-induced AF in 22 patients. EGMs during AF were recorded at least 2 min after AF induction. EGMs were recorded at the left atrial appendage and at the right atrial appendage, under fluoroscopic guidance. Ten bipolar EGMs were transferred to an amplifier (LabSystem Pro, Bard, USA), amplified with a gain of 500 to 1000, and filtered with a band-pass filter set at 30 to 500 Hz. The sample frequency was 1 kHz. 2.4. AFCL measurement
2.6. Pulmonary vein isolation Data analysis was performed off-line with custom made software based on Matlab (Mathworks, Inc., Natick, Mass) [13]. This software utilizes user-defined settings to determine local activation times and were verified by the operator manually. The chosen settings were: dV/dT N 0.05 V/s, amplitude N 0.05 mV and a blanking interval of 120 ms. Previous studies showed a high interobserver correlation of up to 0.99 of manual AFCL determination [14,15]. For all one-second EGM, the mean AFCL was calculated for all 10 bipolar EGMs separately. 15 consecutive seconds of AF EGMs were studied. Per recording site, this
All patients underwent conventional point-by-point PVI. A multipolar steerable circular catheter for circumferential pulmonary vein (PV) mapping (Lasso™, Biosense Webster Inc., Diamond Bar, CA, USA) and a radiofrequency (RF) ablation catheter (Thermocool™, Biosense Webster Inc., Diamond Bar, CA, USA) were inserted transseptally into the left atrium. A left atrial electro-anatomical map was created using 3D software (Carto™, Biosense Webster Inc., Diamond Bar, CA, USA). Templates of PV signals were recorded and pacing from the coronary sinus catheter or from the mapping catheter placed in the left atrial appendage was used whenever deemed necessary to distinguish electrical PV potentials from signals generated by activity of other atrial structures. After entering the PV, the circular catheter was positioned as close as possible to the PV ostium. PVI was achieved by delivering RF energy in a point-by-point fashion to the PV antrum creating a circular ablation lesion. RF energy was applied in a temperature-control mode with a temperature setting up to 43 °C. RF energy was applied at 30 W with a flow rate of 15 ml/min or at 40 W with a flow rate of 30 ml/min, depending on the site of ablation. The endpoint of the ablation procedure was PV isolation, as documented by entrance block or dissociation between PV and atrial activation. No adenosine testing was performed. AFCL and fractionation analysis were done after the procedure and thus did not influence the PVI procedure. 2.7. Follow-up
Fig. 1. Fluoroscopic image of electrophysiological setup. A fluoroscopy image with the twentypolar catheter positioned in the left atrial appendage (PA view).
A blanking period of 3 months was defined after PVI. Patients visited the outpatient clinic at 3, 6 and 12 months after PVI. At these times a 24-hour Holter electrocardiogram was made. An attempt was made in all patients to cease AADs after the 3 month visit. After 12 months, additional clinical visits were performed as deemed necessary by the physician. All patients were interviewed telephonically at the end of the study period to assess AF symptoms and AAD use. AF recurrences were defined according to European guidelines [1].
210
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
Fig. 2. Calculation process of AFCL and fractionation. Panel A displays the AFCL calculation process. In this example, the mean AFCL in the left part of the figure is 154.4 ms and the mean AFCL in the right part of the figure is 176.3 ms. Panel B displays the fractionation score calculation process. The custom software identifies deflections that exceed the dV/dT of −0.015 V/s with a peak-to-peak detection limit of 0.05 mV, maintaining a blanking interval of 8 ms. The total number of identified deflections per second is reduced by the total number of main deflections to calculate the corrected fractionation score. In this example, the fractionation score in the left part of the figure is 38, the number of main deflections is 6, so the corrected fractionation score is 32. In the right part of the image, the fractionation score is 17, the number of main deflections is 5, so the corrected fractionation score is 12. Notice how, during the first second of the EGM, all main activations are fractionated. After 1 s, the AFCL prolongs, and immediately, the fractionation is significantly reduced. This is a typical example of AFCL-dependent fractionation. This EGM displays increased fractionation with a reduction of AFCL. The mAFCC for this patient was −0.267, and this patient remained AF free during the entire followup of 77 months. AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; EGM: electrogram; mAFCC: mean AFCL to fractionation correlation coefficient.
2.8. Statistical analysis Continuous variables are reported as mean ± standard deviation. Categorical variables are reported as number and percentage. The correlation between automatic fractionation assessment and manual fractionation assessment was performed with an intraclass correlation coefficient (ICC). EGMs of 1 s duration were recorded for 15 s. Data was retrospectively analyzed and an intraclass correlation coefficient (ICC) was used to assess the relationship between the local mean AFCL and fractionation scores. For each bipolar EGM, this yielded 15 mean AFCL and 15 fractionation scores, resulting in 10 ICCs for each of the 10 bipolar EGM recordings site per patient, which were then averaged into a mean AFCL to fractionation correlation coefficient (mAFCC). The
difference in mAFCC between pacing-induced AF and spontaneous AF as well as the difference in AFCL, fractionation and mAFCC between patients who were AF free and those who were not AF free was assessed with Student's t-test. A receiver operating characteristic analysis was used to determine the area under the curve, specificity and sensitivity of the mAFCC as a predictor of AF free survival. The optimal cut-off for the mAFCC was −0.090. A chi square test was used to determine the association between the mAFCC with a cut-off of −0.090 and AF free survival. Uni- and multivariate Cox proportional hazard models were used to determine predictors of AF free survival. Variables were entered into the model. Statistical analysis was performed in IBM SPSS statistics 20.0 (IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp). A P-value of ≤0.05 was considered statistically significant.
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
211
Fig. 3. mAFCC calculation of AFCL-independent fractionation. This figure displays the mAFCC calculation of a patient with AFCL-independent fractionation. Notice that AFCL is prolonged by 14 ms in the right part of the EGM, although the fractionation score remains virtually the same. The mAFCC of this patient was −0.006. After 2 PVIs, this patient developed persistent AF. Notice that each complex is fractionated, independent of the AFCL, which is typical for AFCL-independent fractionation. AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; mAFCC: mean atrial fibrillation cycle length to fractionation correlation coefficient.
2.9. Ethics committee The internal review board (IRB) reviewed and approved this study. 3. Results 3.1. Patient characteristics Mean age was 53 years with paroxysmal AF in 83% of patients. No patients had structural heart disease on preprocedural echocardiography. No intracardiac thrombi were identified prior to the procedure. Baseline characteristics are shown in Table 1. All patients completed the electrophysiological study and PVI. 3.2. AFCL and fractionation characteristics In total, 8400 mean AFCL values and fractionation scores were included in the analysis. EGMs of 5 (8.9%) patients were excluded from
analysis due to suboptimal quality. The mean AFCL, fractionation score and mAFCC for both sites are displayed in Table 2 and Fig. 4. The ICC between automatic and manual fractionation was 0.883 (95% CI: 0.843–0.913). There was no difference in mAFCC among induced and spontaneous AF episodes for both recording sites. However, the mAFCC recorded in the left atrial appendage was significantly stronger in paroxysmal AF patients compared to persistent AF patients (−0.185 vs. −0.009, P = 0.024). Other baseline characteristics were not associated with the mAFCC recorded in both the left and right atrial appendages. A typical example of AFCL-dependent fractionation is displayed in Fig. 2. As can be observed, the AFCL is shorter in the left part of the figure compared to the right part of the figure. The fractionation is much higher in the left part of the recording compared to the right part of the recording. The sudden transition of the AFCL lengthening corresponds to the change in degree of fractionation. In Fig. 3, a typical example of AFCL-independent fractionation is displayed. The AFCL is significantly shorter in the left part of the recording compared to the right part of
212
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
4. Discussion
Table 1 Baseline characteristics. Age (years) Gender men (%) BMI (kg/m2) LA size in PSLAX (mm) CHA2DS2VASc score 0 1 2 3–7 Co-morbidity (%) Hypertension Diabetes mellitus Previous TIA/stroke Structural heart disease Type AF: paroxysmal Failed AADs (range) Class I Class III AF duration (years) Spontaneous AF
53.1 (±8.3) 21 (75.0%) 25.6 (±2.4) 40.0 (±4.6) 18 (64.3%) 6 (21.4%) 4 (14.3%) 0 4 (14.3%) 1 (3.6%) 1 (3.6%) 0 24 (85.7%) 2 (0–3) 22 (78.6%) 21 (75.0%) 5.8 (±4.6) 22 (78.6%)
Data are presented as absolute numbers or median with their ±SD, percentages or ranges where appropriate. AADs: anti-arrhythmic drugs; AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; BMI: body mass index; LA: left atrial; PSLAX: echocardiographic parasternal long axis view; TIA: transient ischemic attack.
the recording. However, there is virtually no change in fractionation between the left and right parts of the recording.
3.3. Follow-up data After a median follow-up of 5.5 (q1–q3: 4.3–6.1) years, and a mean of 1.71 (±0.85) PVI attempts, 18 patients were free of AF (64.3%). During follow-up, 9 patients (32.1%) underwent a second PVI, 4 (14.3%) underwent a third PVI and 1 (3.6%) patient underwent a fourth PVI.
3.4. Association with AF free survival The mAFCC recorded at the left atrial appendage significantly predicted AF free survival (area under the curve (AUC) 0.871, P = 0.002), and ROC analysis revealed an optimal cut-off point of −0.090. The mAFCC recorded at the right atrial appendage was not a significant predictor of AF free survival (AUC 0.690, P = 0.131). The ROC analysis is displayed in Fig. 5. However, with a cut-off point of −0.090, there was an association between the mAFCC at both recording sites and AF free survival, as displayed in Table 3. In univariate analysis, none of the patient characteristics was associated with AF free survival. However, the mAFCC recorded at the left atrial appendage was associated with AF free survival, with a hazard ratio (HR) of 2.577 per 0.1 increase in mAFCC (P = 0.016). The mAFCC recorded at the right atrial appendage was not associated with AF free survival (HR 1.151, P = 0.608). The analysis is displayed in Table 4.
4.1. Main findings The present study reports the relationship between AFCL and fractionation and found that the mAFCC recorded at the left atrial appendage was a significant predictor of long-term AF free survival. Although not a significant predictor of AF free survival, there was a significant association between mAFCC recorded at the right atrial appendage and AF free survival. This study suggests that the mAFCC may be an important measure predictor of long term AF free survival. However, in line with other studies, the left atrium appears to be the main driver for sustaining AF. 4.2. Fractionation assessment In this study, fractionation was assessed in bipolar EGMs with settings adapted from CFAE assessment. The correlation between automatic and manual fractionation assessment was very strong. Furthermore, 3 operators agreed in consensus that bipolar EGMs yielded the most reliable results in fractionation assessment. Of note, manual assessment should not be considered a gold standard and frequently, comparing unipolar and bipolar EGMs is necessary to discern fractionated EGMs from far-field activations. 4.3. Fractionation Atrial myocardial cells usually have an effective refractory period (ERP) of 150 to 260 ms [23] during sinus rhythm. During AF, the ERP is reduced [24] and the AFCL has been shown to reflect the local atrial ERP [25]. Previous studies showed that AF is associated with an increased dispersion in atrial ERP [26–30]. EGM fractionation during AF can occur due to structural remodeling (e.g. fibrosis) [8], which is unrelated to the type of AF [31–33]. Fractionation can also be functionally determined by spatial dispersion in refractory periods [34] and areas of functional conduction block [8], as also observed by the linear association between pacing interval and conduction velocity [35]. We hypothesize that functional fractionation may be differentiated from structural fractionation by the mAFCC. Potentially, a strong correlation between AFCL and fractionation may be considered functional fractionation, limited atrial fibrosis and subsequently a higher chance of AF free survival after PVI. Conversely, a weak correlation between the AFCL and fractionation would imply structural remodeling and fibrosis, and a reduced AF free survival after PVI. The design of the 20-polar catheter used in this study allowed high resolution mapping of the LAA and RAA. Of note, the standard deviation of the ICC values was 0.15 and 0.13, respectively, implying a strong variation in ICC values between electrodes in the same patient. Hypothetically, averaging these ICC values into an mAFCC may give an estimate of the total scar burden. The scar burden has already been shown to be associated with the AF free survival after ablation. However, future studies are needed to assess the correlation between electrophysiologic findings and histopathological findings. Furthermore, the present study showed that the association
Table 2 Mean AFCL and fractionation score.
LAA
RAA
AFCL Fractionation score mAFCC AFCL Fractionation score mAFCC
Total group
AF free
Not AF free
P
170.7 ± 36.3 ms 10.4 ± 4.8 −0.149 ± 0.15 174.4 ± 38.6 ms 9.6 ± 4.2 −0.126 ± 0.14
167.5 ± 30.1 10.3 ± 4.5 −0.211 ± 0.143 176.0 ± 40.6 10.3 ± 4.8 −0.152 ± 0.137
180.0 ± 46.9 9.9 ± 5.6 −0.053 ± 0.091 166.2 ± 34.8 9.0 ± 3.0 −0.098 ± 0.146
0.772 0.664 0.001 0.557 0.734 0.141
Values are displayed their ±SD. AFCL: atrial fibrillation cycle length, mAFCC: mean AFCL to fractionation correlation coefficient. AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; LAA: left atrial appendage; RAA: right atrial appendage.
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
213
Fig. 4. Association between mAFCC and AF free survival. In the left panel, the mAFCC recorded at the left atrial appendage is displayed, and in the right panel, the mAFCC recorded at the right atrial appendage is displayed. The mAFCC recorded at the left atrial appendage is significantly stronger in patients who are AF free compared to those who are not (−0.211 vs. −0.053, P = 0.001). The mAFCC recorded at the right atrial appendage did not significantly differ among AF free and not AF free patients (−0.152 vs. −0.098, P = 0.141). AF: atrial fibrillation, AFCL: atrial fibrillation cycle length; mAFCC: mean AFCL to fractionation correlation coefficient.
Fig. 5. Sensitivity/specificity decision plot of the mean AFCL to fractionation correlation coefficient as a predictor of AF free survival. In the left panel, the mAFCC recorded at the left atrial appendage is displayed, and in the right panel, the mAFCC recorded at the right atrial appendage is displayed. The mAFCC of the left atrial appendage has an area under the curve of 0.871, P = 0.002. The mAFCC of the right atrial appendage has an area under the curve of 0.690, P = 0.131. The optimal cut-off point for both sites is −0.090. AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; mAFCC: mean AFCL to fractionation correlation coefficient.
between the mAFCC recorded in the left atrium was a predictor of AF free survival, whereas the mAFCC recorded in the right atrium was not. This suggests that the left atrium is the main driver for AF, which is in line with previous studies [36,37].
4.4. Functional nature of complex fractionated atrial electrograms Several studies investigated the dynamic nature of CFAEs. Knecht et al. [11] reported that pharmacologically induced autonomic blockade resulted in an increase in AFCL exceeding 6 ms and a decrease in CFAEs. Shan et al. [38] found that administration of the class IC drug cibenzoline led to a prolongation of AFCL and a reduction in the percentage of fractionated EGMs. Rostock et al. [10] showed that a shortening of N10 ms in AFCL precedes the occurrence of CFAEs. Additionally, an inverse correlation between CFAEs and AFCL was reported. Jadidi et al. [21] showed
Table 3 Association between mAFCC and AF free survival.
mAFCC LAA b −0.090 mAFCC LAA N −0.090 mAFCC RAA b −0.090 mAFCC RAA N −0.090
AF free
Not AF free
17 2 11 3
1 8 2 7
that CFAEs arose passively at sites of wavelet collision, and reported that these sites displayed normal bipolar voltages during sinus rhythm, suggesting the absence of local fibrosis/scar. Stiles et al. reported that increased fractionation is associated with a higher dominant frequency [39]. The present study supports the hypothesis that CFAEs can be functionally determined and suggests that CFAEs are not necessarily associated with local AF substrate. This observation may also provide an explanation for the conflicting results of CFAE ablation [40]. 4.5. Potential clinical impact Potentially, software can be designed that facilitates the on-line assessment of the mAFCC. This may impact the interpretation of fractionation and thus current clinical practice in the ablation of fractionated EGMs. Furthermore, the mAFCC may assist in a patient-tailored treatment plan when patients suffer from an AF recurrence after PVI: either a repeat PVI in the case of a strong mAFCC or attempting other treatment modalities in the case of a weak mAFCC. 4.6. Limitations
P b0.001 0.008
mAFCC: mean AF cycle length to fractionation correlation coefficient; LAA: left atrial appendage; RAA: right atrial appendage; AF: atrial fibrillation.
In the interpretation of this study, several limitations should be considered. This is a retrospective analysis in a limited number of patients. Potentially, an association between the mAFCC recorded at the right atrial appendage and AF free survival is overlooked due to the limited number of patients. EGMs of 5 patients were excluded from analysis
214
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215
Table 4 Univariate and multivariate analyses of AF recurrence. Univariate analysis
P
Hazard ratio
95% CI
Multivariate analysis
P
Hazard ratio
95% CI
Age Male gender BMI LA dimension in PSLAX CHA2DS2VASc Hypertension Persistent AF Failed AADs AF duration Spontaneous AF LAA mAFCC (per 0.1 increase) RAA mAFCC (per 0.1 increase)
0.823 0.253 0.615 0.167 0.281 0.828 0.827 0.829 0.393 0.672 0.016 0.608
0.990 0.293 0.882 1.117 0.488 0.793 1.263 1.110 1.059 0.706 2.577 1.151
0.908–1.079 0.036–2.399 0.542–1.437 0.955–1.308 0.132–1.800 0.097–6.452 0.155–10.304 0.432–2.847 0.929–1.207 0.141–3.535 1.192–5.568 0.672–1.972
LA dimension in PSLAX LAA mAFCC (per 0.1 increase)
0.681 0.042
1.034 2.371
0.883–1.210 1.033–5.443
Univariate and multivariate analyses of the association between patient characteristics and AF free survival after PVI. AADs: anti-arrhythmic drugs; AF: atrial fibrillation; AFCL: atrial fibrillation cycle length; BMI: body mass index; LA: left atrial; LAA: left atrial appendage; mAFCC: mean AFCL to fractionation correlation coefficient; PSLAX: echocardiographic parasternal long axis view; RAA: right atrial appendage. P-values between AF free and AF recurrence groups.
due to suboptimal quality. Outpatient clinic visits were only performed during the first year after ablation and at the discretion of the physician thereafter, which may have resulted in an overestimation of the AF free survival. 5. Conclusion In this study, the mAFCC recorded in the left atrial appendage was a significant predictor of long-term AF free survival. Although not a significant predictor of AF free survival, there was a significant association between mAFCC recorded at the right atrial appendage and AF free survival. Potentially, the mAFCC can be used as a predictor of AF free survival in patients undergoing PVI. Funding None. Conflict of interest None. Acknowledgments We acknowledge Ms. Petra Koopmans for the advice on the statistical analysis. We thank Ms. Vera Derks for the final preparation of this manuscript. References [1] H. Calkins, K.H. Kuck, R. Cappato, J. Brugada, A.J. Camm, S.A. Chen, et al., 2012 HRS/ EHRA/ECAS Expert Consensus Statement on Catheter and Surgical Ablation of Atrial Fibrillation: recommendations for patient selection, procedural techniques, patient management and follow-up, definitions, endpoints, and research trial design, Europace 14 (2012) 528–606. [2] H. Calkins, M.R. Reynolds, P. Spector, M. Sondhi, Y. Xu, A. Martin, et al., Treatment of atrial fibrillation with antiarrhythmic drugs or radiofrequency ablation: two systematic literature reviews and meta-analyses, Circ. Arrhythm. Electrophysiol. 2 (2009) 349–361. [3] P. Gal, A.E. Aarntzen, J.J. Smit, A. Adiyaman, A.R. Misier, P.P. Delnoy, et al., Conventional radiofrequency catheter ablation compared to multi-electrode ablation for atrial fibrillation, Int. J. Cardiol. 176 (2014) 891–895. [4] Y. Liu, H. Huang, C. Huang, S. Zhang, C. Ma, X. Liu, et al., Catheter ablation of atrial fibrillation in Chinese elderly patients, Int. J. Cardiol. 152 (2011) 266–267. [5] S. Knecht, C. Sticherling, S. von Felten, D. Conen, B. Schaer, P. Ammann, et al., Longterm comparison of cryoballoon and radiofrequency ablation of paroxysmal atrial fibrillation: a propensity score matched analysis, Int. J. Cardiol. 176 (2014) 645–650. [6] T. Pezawas, R. Ristl, M. Bilinski, C. Schukro, H. Schmidinger, Single, remote-magnetic catheter approach for pulmonary vein isolation in patients with paroxysmal and non-paroxysmal atrial fibrillation, Int. J. Cardiol. 174 (2014) 18–24.
[7] K. Nademanee, J. McKenzie, E. Kosar, M. Schwab, B. Sunsaneewitayakul, T. Vasavakul, et al., A new approach for catheter ablation of atrial fibrillation: mapping of the electrophysiologic substrate, J. Am. Coll. Cardiol. 43 (2004) 2044–2053. [8] U. Schotten, S. Verheule, P. Kirchhof, A. Goette, Pathophysiological mechanisms of atrial fibrillation: a translational appraisal, Physiol. Rev. 91 (2011) 265–325. [9] D.H. Lau, B. Maesen, S. Zeemering, S. Verheule, H.J. Crijns, U. Schotten, Stability of complex fractionated atrial electrograms: a systematic review, J. Cardiovasc. Electrophysiol. 23 (2012) 980–987. [10] T. Rostock, M. Rotter, P. Sanders, Y. Takahashi, P. Jais, M. Hocini, et al., High-density activation mapping of fractionated electrograms in the atria of patients with paroxysmal atrial fibrillation, Heart Rhythm. 3 (2006) 27–34. [11] S. Knecht, M. Wright, S. Matsuo, I. Nault, N. Lellouche, F. Sacher, et al., Impact of pharmacological autonomic blockade on complex fractionated atrial electrograms, J. Cardiovasc. Electrophysiol. 21 (2010) 766–772. [12] F. Atienza, D. Calvo, J. Almendral, S. Zlochiver, K.R. Grzeda, N. Martinez-Alzamora, et al., Mechanisms of fractionated electrograms formation in the posterior left atrium during paroxysmal atrial fibrillation in humans, J. Am. Coll. Cardiol. 57 (2011) 1081–1092. [13] M. Potse, A.C. Linnenbank, C.A. Grimbergen, Software design for analysis of multichannel intracardial and body surface electrocardiograms, Comput. Methods Programs Biomed. 69 (2002) 225–236. [14] K.C. Ravi, D.E. Krummen, A.J. Tran, J.R. Bullinga, S.M. Narayan, Electrocardiographic measurements of regional atrial fibrillation cycle length, Pacing Clin. Electrophysiol. 32 (Suppl. 1) (2009) S66–S71. [15] R.J. Hunter, I. Diab, M. Tayebjee, L. Richmond, S. Sporton, M.J. Earley, et al., Characterization of fractionated atrial electrograms critical for maintenance of atrial fibrillation: a randomized, controlled trial of ablation strategies (the CFAE AF trial), Circ. Arrhythm. Electrophysiol. 4 (2011) 622–629. [16] D. Scherr, D. Dalal, A. Cheema, A. Cheng, C.A. Henrikson, D. Spragg, et al., Automated detection and characterization of complex fractionated atrial electrograms in human left atrium during atrial fibrillation, Heart Rhythm. 4 (2007) 1013–1020. [17] A. Verma, R. Mantovan, L. Macle, G. De Martino, J. Chen, C.A. Morillo, et al., Substrate and Trigger Ablation for Reduction of Atrial Fibrillation (STAR AF): a randomized, multicentre, international trial, Eur. Heart J. 31 (2010) 1344–1356. [18] A. Verma, P. Novak, L. Macle, B. Whaley, M. Beardsall, Z. Wulffhart, et al., A prospective, multicenter evaluation of ablating complex fractionated electrograms (CFEs) during atrial fibrillation (AF) identified by an automated mapping algorithm: acute effects on AF and efficacy as an adjuvant strategy, Heart Rhythm. 5 (2008) 198–205. [19] J.F. Viles-Gonzalez, J.A. Gomes, M.A. Miller, S.R. Dukkipati, J.S. Koruth, C. Eggert, et al., Areas with complex fractionated atrial electrograms recorded after pulmonary vein isolation represent normal voltage and conduction velocity in sinus rhythm, Europace 15 (2013) 339–346. [20] A.S. Jadidi, H. Cochet, A.J. Shah, S.J. Kim, E. Duncan, S. Miyazaki, et al., Inverse relationship between fractionated electrograms and atrial fibrosis in persistent atrial fibrillation — a combined MRI and high density mapping, J. Am. Coll. Cardiol. 27 (2013) 802–812. [21] A.S. Jadidi, E. Duncan, S. Miyazaki, N. Lellouche, A.J. Shah, A. Forclaz, et al., Functional nature of electrogram fractionation demonstrated by left atrial high-density mapping, Circ. Arrhythm. Electrophysiol. 5 (2012) 32–42. [22] K.R. Grzeda, S.F. Noujaim, O. Berenfeld, J. Jalife, Complex fractionated atrial electrograms: properties of time-domain versus frequency-domain methods, Heart Rhythm. 6 (2009) 1475–1482. [23] P.R. Podrid PK, Cardiac Arrhythmia: Mechanisms, Diagnosis, and Management, Lippincott Williams & Wilkins, 2001. [24] M.C. Wijffels, C.J. Kirchhof, R. Dorland, M.A. Allessie, Atrial fibrillation begets atrial fibrillation. A study in awake chronically instrumented goats, Circulation 92 (1995) 1954–1968. [25] A. Capucci, M. Biffi, G. Boriani, F. Ravelli, G. Nollo, P. Sabbatani, et al., Dynamic electrophysiological behavior of human atria during paroxysmal atrial fibrillation, Circulation 92 (1995) 1193–1202. [26] Z. Li, E. Hertervig, J. Carlson, C. Johansson, S.B. Olsson, S. Yuan, Dispersion of refractoriness in patients with paroxysmal atrial fibrillation. Evaluation with simultaneous endocardial recordings from both atria, J. Electrocardiol. 35 (2002) 227–234.
P. Gal et al. / International Journal of Cardiology 187 (2015) 208–215 [27] Z. Li, E. Hertervig, S. Yuan, Y. Yang, Z. Lin, S.B. Olsson, Dispersion of atrial repolarization in patients with paroxysmal atrial fibrillation, Europace 3 (2001) 285–291. [28] E. Diker, M. Ozdemir, S. Aydogdu, U.K. Tezcan, S. Korkmaz, E. Kutuk, et al., Dispersion of repolarization in paroxysmal atrial fibrillation, Int. J. Cardiol. 63 (1998) 281–286. [29] H. Ramanna, R.N. Hauer, F.H. Wittkampf, J.M. de Bakker, E.F. Wever, A. Elvan, et al., Identification of the substrate of atrial vulnerability in patients with idiopathic atrial fibrillation, Circulation 101 (2000) 995–1001. [30] H. Ramanna, A. Elvan, F.H. Wittkampf, J.M. de Bakker, R.N. Hauer, E.O. Robles de Medina, Increased dispersion and shortened refractoriness caused by verapamil in chronic atrial fibrillation, J. Am. Coll. Cardiol. 37 (2001) 1403–1407. [31] N. Akoum, M. Daccarett, C. McGann, N. Segerson, G. Vergara, S. Kuppahally, et al., Atrial fibrosis helps select the appropriate patient and strategy in catheter ablation of atrial fibrillation: a DE-MRI guided approach, J. Cardiovasc. Electrophysiol. 22 (2011) 16–22. [32] A. Boldt, U. Wetzel, J. Lauschke, J. Weigl, J. Gummert, G. Hindricks, et al., Fibrosis in left atrial tissue of patients with atrial fibrillation with and without underlying mitral valve disease, Heart 90 (2004) 400–405. [33] A.W. Teh, P.M. Kistler, G. Lee, C. Medi, P.M. Heck, S.J. Spence, et al., Electroanatomic remodeling of the left atrium in paroxysmal and persistent atrial fibrillation patients without structural heart disease, J. Cardiovasc. Electrophysiol. 23 (2012) 232–238. [34] K.T. Konings, J.L. Smeets, O.C. Penn, H.J. Wellens, M.A. Allessie, Configuration of unipolar atrial electrograms during electrically induced atrial fibrillation in humans, Circulation 95 (1997) 1231–1241.
215
[35] S.L. Chang, Y.C. Chen, C.P. Hsu, Y.H. Kao, Y.K. Lin, Y.J. Lin, et al., Electrophysiological characteristics of complex fractionated electrograms and high frequency activity in atrial fibrillation, Int. J. Cardiol. 168 (2013) 2289–2299. [36] M. Haissaguerre, M. Hocini, A. Denis, A.J. Shah, Y. Komatsu, S. Yamashita, et al., Driver domains in persistent atrial fibrillation, Circulation 130 (2014) 530–538. [37] N.F. Marrouche, D. Wilber, G. Hindricks, P. Jais, N. Akoum, F. Marchlinski, et al., Association of atrial tissue fibrosis identified by delayed enhancement MRI and atrial fibrillation catheter ablation: the DECAAF study, JAMA 311 (2014) 498–506. [38] Z. Shan, P.H. Van Der Voort, Y. Blaauw, M. Duytschaever, M.A. Allessie, Fractionation of electrograms and linking of activation during pharmacologic cardioversion of persistent atrial fibrillation in the goat, J. Cardiovasc. Electrophysiol. 15 (2004) 572–580. [39] M.K. Stiles, A.G. Brooks, P. Kuklik, B. John, H. Dimitri, D.H. Lau, et al., High-density mapping of atrial fibrillation in humans: relationship between high-frequency activation and electrogram fractionation, J. Cardiovasc. Electrophysiol. 19 (2008) 1245–1253. [40] W.J. Li, Y.Y. Bai, H.Y. Zhang, R.B. Tang, C.L. Miao, C.H. Sang, et al., Additional ablation of complex fractionated atrial electrograms after pulmonary vein isolation in patients with atrial fibrillation: a meta-analysis, Circ. Arrhythm. Electrophysiol. 4 (2011) 143–148.
