Author + information
- Received August 18, 2006
- Revision received October 18, 2006
- Accepted October 30, 2006
- Published online April 3, 2007.
- Taro Okada, MD⁎,
- Takumi Yamada, MD†,⁎ (, )
- Yoshimasa Murakami, MD⁎,
- Naoki Yoshida, MD⁎,
- Yuuichi Ninomiya, MD⁎,
- Takeshi Shimizu, MD⁎,
- Junji Toyama, MD⁎,
- Yukihiko Yoshida, MD‡,
- Teruo Ito, MD‡,
- Naoya Tsuboi, MD§,
- Takahisa Kondo, MD∥,
- Yasuya Inden, MD∥,
- Makoto Hirai, MD∥ and
- Toyoaki Murohara, MD∥
- ↵⁎Reprint requests and correspondence:
Dr. Takumi Yamada, Division of Cardiovascular Diseases, Cardiac Rhythm Management Laboratory, University of Alabama at Birmingham, VH B147, 1670 University Boulevard, 1530 3rd Avenue South, Birmingham, Alabama 35294-0019.
Objectives The aim of this study was to investigate the prevalence and severity of left atrial (LA) edema after pulmonary vein (PV) ablation and its effect on the cardiac function.
Background Though extensive LA catheter ablation has been demonstrated to be more effective in curing paroxysmal atrial fibrillation (PAF) than segmental ostial pulmonary vein isolation (S-PVI), it might cause life-threatening complications, including congestive heart failure associated with LA edema.
Methods Fifty patients underwent S-PVI (Group S) and 27 underwent circumferential PV antrum ablation (Group C) for drug-refractory PAF. Enhanced electron beam tomography (EBT) was performed before, 1 or 2 days after, and 1 month after the PV ablation, and transthoracic ultrasound cardiography (UCG) was performed 1 month after the PV ablation in all patients.
Results The EBT assessment revealed LA edema immediately after the PV ablation in 47 Group S patients and all Group C patients. The severity of the LA edema, number of radiofrequency applications, and amount of radiofrequency energy delivered during the PV ablation was significantly greater in Group C than in Group S. One month after the PV ablation, in all patients, the EBT assessment revealed that those edematous changes had disappeared, and the UCG assessment showed no reduction in the cardiac function.
Conclusions Left atrial edema was observed in a large portion of the patients immediately after the PV ablation, and the severity of the LA edema depended on the extent and amount of the radiofrequency energy delivered in the PV ablation. The LA edema soon disappeared naturally and did not reduce the cardiac function.
In recent years, ablation targeting arrhythmogenic pulmonary vein (PV) foci for drug-refractory paroxysmal atrial fibrillation (PAF), initiated by Haissaguerre et al. (1), has been established as the pulmonary vein isolation (PVI) technique (2,3). However, in the initial versions of the technique, the high incidence of PAF recurrence and complications such as PV stenosis (4–7) troubled both patients and operators. To improve the clinical results and avoid those complications, the target of the PV ablation was extended toward the PV antrum (8,9) or left atrium (LA) (10). Extensive LA catheter ablation around the PVs, combined with linear ablation at the superior portion of the LA and mitral isthmus, has recently been demonstrated to be a more effective technique for curing PAF than segmental ostial PV ablation (10). However, that technique might cause life-threatening complications such as an atrioesophageal fistula (11) or congestive heart failure (CHF) associated with LA edema (12). The prevalence and severity of LA edema after PV ablation and its effect on the cardiac function remain unknown. The aim of this study was to investigate LA edema after PV ablation by using electron beam tomography (EBT) and transthoracic ultrasound cardiography (UCG).
This study population consisted of 77 consecutive patients (65 men, 58 ± 11 years) with symptomatic PAF refractory to 4 ± 1 class I or class III antiarrhythmic drugs (not including amiodarone). The mean PAF history was 5 ± 4 years (range 1 to 15 years). The mean LA dimension was 34.1 ± 5.0 mm (range 22 to 44 mm), and mean left ventricular ejection fraction (EF) was 67.2 ± 10.1% (range 51% to 80%). Fifty patients (Group S) underwent segmental ostial PVI, and 27 (Group C) underwent circumferential PV antrum ablation. One Group S patient had a history of a direct closure operation of an atrial septal defect, and another had a history of moderate aortic valve regurgitation. Two Group C patients had a history of ischemic heart disease. Written informed consent was obtained from all the patients, and all antiarrhythmic drugs were discontinued for at least 5 half-lives before the study.
The trans-septal procedure was performed with intracardiac echocardiography guidance. Thereafter, intravenous heparin was administered to maintain an activated clotting time between 250 and 300 s. Catheterization into the LA was performed with a 1-puncture and 2-sheath technique (1-sheath, 8-F, FAST CATH SL 1, St. Jude Medical, AF Division, Minnetonka, Minnesota) for an ablation catheter and again for a mapping catheter (8.5-F, Soft Tip EP Sheath, EP Technologies, Boston Scientific, San Jose, California). Selective angiograms of all PVs were performed to identify the PV ostia and antrum.
Target PVs for isolation and PV mapping
The left superior PV, left inferior PV, right superior PV, and right inferior PV were all targeted for a 2-PV isolation technique. A 31-mm multielectrode basket catheter (MBC) (Constellation, EP Technologies, Boston Scientific), which consisted of 8 splines (A to H) with eight 1-mm electrodes and 2-mm spacing, was deployed within 3 to 4 PVs via the atrial septum. An MBC was introduced toward the distal PV and then pulled back with fluoroscopic guidance as proximally as possible without dislodgement until its most proximal electrodes were positioned at the PV ostium or antrum, which was identified by a selective angiogram. A total of 56 bipolar electrograms were recorded by the MBC during sinus rhythm (right PVs) or distal coronary sinus pacing (left PVs). When AF persisted during the electrophysiologic study, internal cardioversion was used to restore sinus rhythm, and an MBC recording of at least 1 beat was obtained during the appropriate rhythm. If an MBC could not be deployed in the inferior PVs, a 20-electrode circular catheter (Lasso, Biosense Webster, Diamond Bar, California) was used for mapping those PVs.
In the Group S patients, segmental ostial PVI targeting a preferential electrical connection between the PVs and LA was performed as we previously described (13). Radiofrequency (RF) energy was delivered with a target temperature of 55°C and maximum power output of 30 to 40 W for 30 to 60 s (EPT-1000TC generator, EP Technologies, Boston Scientific) using an 8-mm tip catheter (Blazer II 5770T, EP Technologies, Boston Scientific).
In the Group C patients, circumferential PV antrum ablation targeting the PV antrum potentials was performed as we previously described (9). Radiofrequency energy was delivered in the same manner as in segmental ostial PVI. If a residual conduction gap was detected after the PV antrum ablation, additional RF applications to the PV side next to the previous RF lesions were delivered. The final end point of those 2 ablation techniques was the complete PV electric disconnection and noninducibility of AF during an isoproterenol infusion (2 to 4 μg/min) and burst atrial pacing (to a cycle length as short as 200 ms).
In the 2 PV ablation techniques, the catheter-dragging technique was not used during the RF applications and only a point-to-point RF application was applied. Therefore, we thought that the number of RF applications might closely correspond to the extent of the RF ablation.
EBT scan, UCG, and assessment of the LA wall
To evaluate the severity and thickness of the LA edema after the PV ablation, an enhanced EBT scan was performed before, 1 or 2 days, and 1 month after the PV ablation using an Imatron C-150 system (Imatron, South San Francisco, California). If the enhanced EBT scan performed 1 month after the PV ablation revealed LA edema, a follow-up enhanced EBT scan was added every 3 months after the PV ablation until the complete disappearance of the LA edema was confirmed. The Institutional Review Board of Aichi Prefectural Cardiovascular and Respiratory Center approved the study protocol for the serial EBT scans, and all patients provided written informed consent. The patients were scanned in the supine position, and a total volume of 100 to 120 ml of nonionic iodinated contrast medium (Iopamidol, 370 mg iodine/ml) was injected into the antecubital vein at a rate of 2 ml/s. Thirty seconds after the administration of the contrast medium, 40 axial cross-sectional images of the heart were acquired in the single-slice mode 100 ms after an electrocardiographic trigger signal was generated at 40% of the RR interval (LA end-diastolic phase). The matrix size was 512 × 512 and field of view 18 cm (pixel size 0.35 × 0.35 mm). The slice thickness was 3 mm with table increments of 2 mm. Therefore, the EBT imaging range was limited to 8 cm, which was sufficient to evaluate the entire LA and PV ostia. The images were transferred to a workstation (Allegro, ISG Technologies, Toronto, Canada) and analyzed by 2 experienced evaluators who were blinded to the treatment option employed in each patient.
Left atrial edema was defined as LA wall swelling that had not been observed before the PV ablation. The LA edema was graded semi-quantitatively according to the extent of the edema as we defined (Fig. 1A):Grade 0, no LA wall swelling; Grade 1, LA wall swelling observed only around the PV; Grade 2, LA wall swelling observed around the PV and a part of the LA free wall; and Grade 3, LA wall swelling observed almost all around the LA. The difference in the LA wall thickness measured on the right side of the LA posterior wall (Fig. 1B) before and after the PV ablation was assessed as the thickness of the LA edema. For the assessment of the LA size, the maximum width and length were measured at the slice just below the left inferior PV (Fig. 1C). The incidence of significant PV stenosis (a >50% reduction in the PV ostial size relative to that before the ablation) was also evaluated.
The LA diameter, left ventricular EF, and A-wave velocity from the Doppler mitral inflow pattern (in the patients with sinus rhythm alone) were evaluated by transthoracic UCG before and 1 month after the PV ablation. The UCG evaluators were blinded to the patient’s history.
Follow-up was performed at 2 weeks, 1 month, and every month thereafter using 24-h Holter and cardiac recordings until at least 1 year after the PV ablation. All patients who reported symptoms were given an event monitor to document the cause of the symptoms.
Continuous variables are expressed as the group mean ± SD. Comparisons between groups were performed with either an unpaired Student ttest or analysis of variance or, where a normal distribution could not be assumed, the Mann-Whitney Utest. Categorical variables expressed as numbers and percentages were compared with a chi-square test. Simple regression analysis or a Spearman rank correlation was used to estimate the correlation between the 2 parameters. Statistical significance was selected as a value of p < 0.05.
The results of the PV ablation are shown in Table 1.The number of isolated PVs was 3.86 ± 0.35 in Group S and 3.96 ± 0.19 in Group C (p = NS). The number of RF applications and amount of the RF energy delivered during the PV ablation were significantly greater in Group C than in Group S (p < 0.001 and p < 0.0001, respectively).
The results of the EBT assessment are shown in Table 1. The intraobserver variability with repeated measurements was ≤5%, and the interobserver variability was ≤10%. An EBT assessment at 1 or 2 days after the PV ablation revealed edematous changes in the LA wall in 47 Group S patients and in all 27 Group C patients. The grade and thickness of the LA edema were significantly greater in Group C than in Group S (p < 0.05 and p < 0.0001, respectively) (Fig. 2).In 21 (42%) Group S patients and 18 (67%) Group C patients, the edematous changes were observed almost entirely around the LA wall and extended to regions remote from the PVs, where the RF ablation was not applied. The thickness of the LA edema immediately after the PV ablation was weakly correlated (p = 0.0470, r = 0.23) with the LA wall thickness before the PV ablation. The amount of the RF energy delivered was weakly correlated with the grade (p = 0.0390, r = 0.24) and thickness of the LA edema (p = 0.0008, r = 0.38) over the entire group. The number of RF applications, which might almost correspond to the extent of the RF ablation, was weakly correlated with the thickness of the LA edema (p = 0.0073, r = 0.30). The EBT assessment at 1 month after the PV ablation revealed that those edematous changes had almost disappeared in all of those patients (Fig. 3).Although only a subtle LA wall swelling was still observed at 1 month after the PV ablation in 5 patients, it completely disappeared by 3 months after the PV ablation in all patients. In those 5 patients, neither the thickness of the LA edema immediately after the PV ablation nor the amount of the RF energy delivered were significantly greater as compared to the residual 71 patients (4.8 ± 1.7 mm vs. 3.8 ± 2.6 mm, p = NS, 113 ± 64.2 × 103J vs. 98.9 ± 49.9 × 103J, p = NS, respectively).
LA dimension and cardiac function
The EBT assessment revealed that there were no significant changes in the LA size for those assessments performed before (73.1 ± 7.8 mm × 41.5 ± 7.2 mm), 1 or 2 days after (72.3 ± 6.7 mm × 42.9 ± 6.1 mm), and 1 month (71.5 ± 6.3 mm × 42.0 ± 6.8 mm, p = NS) after the PV ablation. A UCG assessment revealed that there were no significant changes in the LA diameter or left ventricular EF between those results obtained before and 1 month after the PV ablation (LA diameter 34.1 ± 5.0 mm vs. 34.3 ± 5.1 mm; left ventricular EF 67.2 ± 10.1% vs. 66.5 ± 7.0%, p = NS, respectively). During the UCG procedures, 65 patients were in sinus rhythm before the PV ablation, and 67 patients were in sinus rhythm 1 month after the PV ablation. In those patients, there were no significant changes in the A-wave velocity observed on the Doppler mitral inflow pattern between those results obtained before and 1 month after the PV ablation (70.1 ± 21.4 cm/s vs. 61.9 ± 21.1 cm/s, p = NS).
During more than 6 months of follow-up, 25 Group S patients (50%) and 24 Group C patients (89%) were free of symptomatic PAF without any antiarrhythmic drugs (p < 0.001). However, no procedure-related LA tachycardias were observed in either group. There were no significant relations between the thickness and grade of the LA edema and early (<1 month after the ablation) and late (>1 month after the ablation) AF recurrence in either group. Although significant PV stenosis was observed in 2 (4%) Group S patients and 1 (3.7%) Group C patient (p = NS), those patients were asymptomatic. No other extracardiac complications occurred.
In recent years, PVI has been developed as a curative therapeutic method for drug refractory PAF (3,8,9). It has been reported that extensive PV ablation might cause unfavorable and sometimes life-threatening complications such as phrenic nerve injury (14), periesophageal nerve injury (15), atrio-esophageal fistulas (11), or CHF (11); however, the extensive PV ablation has achieved better clinical results than the PVI techniques (8,9,10). Though CHF associated with LA edema after PV ablation (12) seems to be rare compared to the other complications, we believe that it is very important to recognize the prevalence and severity of the LA edema after the PV ablation and its effect on the cardiac function in order to achieve an effective and safer PV ablation.
To the best of our knowledge, this is the first report describing the edematous changes in the LA wall immediately after the PV ablation. The edematous changes in the LA wall were observed in a large portion of the patients immediately after the PV ablation. The extent and amount of the RF energy delivered during the PV ablation were weakly correlated with the severity of the LA edema over the entire group. On the other hand, the severity of the LA edema and extent and amount of the RF energy delivered during the PV ablation were significantly greater in the patients who underwent circumferential PV antrum ablation than in those who underwent segmental ostial PVI. Therefore, the severity of the LA edema might depend more greatly on the PV ablation technique itself than on the extent and amount of the RF energy delivered during the PV ablation. The myocardium sleeve covering the PVs, which was continuous longitudinally, became gradually thinner toward the pulmonary segment of the vein and ended distinctly (16). In the circumferential PV antrum ablation, a thicker myocardium might be targeted and a larger amount of myocardium might be damaged than in segmental ostial PVI. Therefore, the severity of the LA edema might actually be associated with the difference in the amount of damaged myocardium resulting from the different PV ablation techniques.
Interestingly, the edematous changes in the LA wall often extended to regions remote from the PVs, where the RF ablation was not applied. Duytschaever et al. (17) reported an electrogram voltage reduction on the posterior wall of the LA not only within but also outside the encircled PVs after circumferential PVI. The mechanism underlying their findings may be explained by the edematous changes in the LA wall remote from the PVs, which the present study revealed immediately after the PV ablation. The mechanism underlying the LA edema remote from the ablated PVs is unclear. Wood et al. (18) reported a case complicated by after-pericardiotomy syndrome early after a left atrial linear ablation. In their case, the cardiac magnetic resonance imaging revealed diffuse intense pericardial edema and thickening. They suggested that in this case, pericarditis was caused by the extensive transmural atrial or pericardial injury from the linear LA lesions. Although, fortunately, our patients did not show such severe symptoms, despite significant edematous changes in the LA wall observed in EBT images, there might be some inflammation in the LA wall extending away from the PVs.
Steel et al. (12) reported that severe LA edema could remain even 2 months after the PV ablation. In their case, atrial mechanical stunning attributable to atrial edema also caused severe CHF. Fortunately, none of our cases showed symptoms or signs associated with heart failure. That is probably because the LA edema had almost completely disappeared in all of our patients within 1 month after the PV ablation. Schwartzman et al. (19) demonstrated that tissue edema of the right atrial wall due to inflammation evolved rapidly and resolved within 4 weeks after a linear ablation of the right atrium in a pig model. The LA edema in our cases might have also been caused by temporal inflammation and thus resolved naturally.
Duytschaever et al. (17) suggested that an electrogram voltage reduction at the LA posterior wall remote from the ablated PVs, probably owing to edematous changes, may have contributed to the modification of the arrhythmia substrate. However, it is unknown whether an electrogram voltage reduction at the LA posterior wall persists or not. Because in the present study, the LA edema disappeared in a large portion of the cases within 1 month after the PV ablation, and no significant correlations were observed between the severity of the LA edema and AF recurrence, we do not think the LA edema favored chronic clinical results. Further investigations are needed to clarify whether these changes in the LA wall after the PV ablation are contributors to the long-term clinical results or mere “collateral damage.”
The complications in the current findings, as well as in most previous cases with major complications (esophageal fistula, and so on), might result from ablation using an 8-mm-tip catheter. The use of irrigation-tip catheters with lower energy (30 to 35 W) levels has the advantage of delivering a constant energy, and perhaps our findings might not be applicable to those catheters.
Left atrial wall swelling itself due to tissue edema may be thrombogenic. Therefore, anticoagulation therapy should be continued for at least 1 month after the PV ablation even in patients without AF recurrence, especially when the ablation is extensive.
In all patients without any symptoms or signs of CHF in the present study, almost complete disappearance of the LA edema was demonstrated by an EBT assessment at 1 month after the PV ablation. On the other hand, severe LA edema remaining even 2 months after the PV ablation could cause severe CHF (12). Therefore, an EBT assessment at 1 month after the PV ablation may be useful to predict any complications of CHF, even if it rarely occurs.
Because a histopathologic analysis had not been performed, we could not conclude the precise mechanisms of the edematous changes observed in the LA wall. Schwartzman et al. (19) reported that tissue edema was pathologically demonstrated to be the cause of early right atrial wall thickening after the linear ablation of the right atrium in a pig model. Mitchell et al. (20) reported that atrial subepicardial coagulation and hemorrhaging were pathologically observed after atrial linear ablation in a canine model. Therefore, the same mechanism as described earlier may be suggested in human cases.
What was actually measured in this study might not be the actual edema thickness of the LA myocardium but the change in the LA wall thickness. Actually, because of the limited spatial resolution of the EBT, the LA myocardium could not be distinguished from the interstitial space and epicardium either before or after the PV ablation in all cases in this study. According to the report by Schwartzman et al. (19), measurement using intracardiac echocardiography demonstrated that the interstitial space was 0 mm before the ablation and the interstitial edema thickness was about 50% to 60% of the right atrial wall edema thickness after the linear ablation of the right atrium in a pig model. Therefore, we believe that the LA wall thickness before the PV ablation was almost identical to the thickness of the LA myocardium, and the change in the LA wall thickness was not entirely but mainly due to the LA myocardial edema.
Actually, multi-slice computed tomography may be more efficient for the qualitative analysis of the edematous changes in the LA wall. However, this study was designed in combination with routine monitoring to detect any PV stenosis after the PV ablation. Therefore, we thought that an EBT with a lower radiation exposure (21) might be appropriate for these investigations.
The learning curve of the ablator might have affected the results of this study. However, we experienced more than 50 cases of segmental ostial PV ablation before this study and used the same mapping and navigation systems in the 2 PV ablation techniques in this study. Therefore, we believe that the impact of the learning curve was minimal enough to be neglected.
Recent reports using transtelephonic and long-term Holter monitoring have revealed that asymptomatic AF recurrences after an AF ablation occur more often than we would expect, and thus freedom from AF after an AF ablation may be overestimated (22,23). In this study, asymptomatic AF recurrences or new occurrences of LA flutter might have been missed, and the cure rate might have been a little overestimated because intermittent Holter recordings alone were performed for the clinical follow-up.
This study revealed that the edematous changes in the LA wall were observed in a large portion of the patients immediately after the PV ablation, and the severity of the LA edema depended on the extent and amount of the RF energy delivered during the PV ablation. The LA edema might have been caused by temporal inflammation and thus resolved naturally. The LA edema disappearing soon after the PV ablation did not reduce the cardiac function, and thus, CHF associated with LA edema may be rare.
The authors are extremely grateful to Shoji Yamakami, MD, Kazuhiko Nakahata, ME, Mitsuaki Sumi, ME, and Hiroyuki Ogura, ME, for their technical support.
- Abbreviations and Acronyms
- congestive heart failure
- electron beam tomography
- ejection fraction
- left atrium
- multielectrode basket catheter
- paroxysmal atrial fibrillation
- pulmonary vein
- pulmonary vein isolation
- ultrasound cardiography
- Received August 18, 2006.
- Revision received October 18, 2006.
- Accepted October 30, 2006.
- American College of Cardiology Foundation
- Haissaguerre M.,
- Shah D.C.,
- Jais P.,
- et al.
- Oral H.,
- Knight B.P.,
- Tada H.,
- et al.
- Chen S.A.,
- Hsieh M.H.,
- Tai C.T.,
- et al.
- Haissaguerre M.,
- Jais P.,
- Shah D.C.,
- et al.
- Robbins I.M.,
- Colvin E.V.,
- Doyle T.P.,
- et al.
- Yamada T.,
- Murakami Y.,
- Okada T.,
- et al.
- Oral H.,
- Scharf C.,
- Chugh A.,
- et al.
- Pappone C.,
- Oral H.,
- Santinelli V.,
- et al.
- Steel K.E.,
- Roman-Gonzalez J.,
- O’Bryan C.L.
- Shah D.,
- Dumonceau J.M.,
- Burri H.,
- et al.
- Duytschaever M.,
- Beeumen K.V.,
- Eynde G.V.,
- Tavernier R.
- Mitchell M.A.,
- McRury I.D.,
- Haines D.E.
- Morin R.L.,
- Gerber T.C.,
- McCollough C.H.
- Hindricks G.,
- Piorkowski C.,
- Tanner H.,
- et al.