Author + information
- Received July 31, 2001
- Revision received January 7, 2002
- Accepted January 18, 2002
- Published online April 17, 2002.
- Teiichi Yamane, MD*,* (, )
- Dipen C Shah, MD*,
- Pierre Jaïs, MD*,
- M.élèze Hocini, MD*,
- Isabel Deisenhofer, MD*,
- Kee-Joon Choi, MD*,
- Laurent Macle, MD*,
- Jacques Clémenty, MD* and
- Michel Haïssaguerre, MD*
- ↵*Reprint requests and correspondence:
Dr. Teiichi Yamane, Hôpital Cardiologique du Haut-Lévêque, Avenue de Magellan, 33604 Bordeaux-Pessac, France.
Objectives We assessed the anatomical distribution and electrogram characteristics of breakthrough from the left atrium (LA) to the pulmonary veins (PVs).
Background Localization of LA-PV breakthrough is an important technique for PV ablation in patients with atrial fibrillation (AF).
Methods A total of 157 patients with paroxysmal AF underwent PV disconnection guided by mapping with a circumferential 10-electrode catheter. Radiofrequency (RF) current was delivered ostially at the site(s) of earliest activation (113 patients) or electrogram polarity reversal defined by opposite polarity across adjacent bipoles (44 patients). Breakthrough sites were proved by changes in pulmonary vein potential activation sequence occurring as a result of localized RF delivery and were classified into four segments around the ostium (top, bottom, anterior, posterior). Results of mapping and ablation were compared between the two groups.
Results A total of 99% of 411 targeted PVs were successfully disconnected in both groups. Breakthroughs were most frequent at the bottom of superior PVs (85% prevalence) and the top of inferior PVs (75% prevalence). A wide activation front (>5 synchronous bipoles) indicating broad breakthrough was observed in 18% of PVs. Polarity reversal occurred with 88% sensitivity and 91% specificity at breakthrough sites. Polarity reversal was restricted to fewer bipoles (2.0 ± 0.4 bipoles vs. 3.4 ± 2.0 bipoles, p < 0.01) compared with earliest activation. Shorter RF application time was required to disconnect PVs with wide synchronous activation using polarity reversal compared with using conventional earliest activity (10.3 ± 3.0 min vs. 12.3 ± 3.4 min, p < 0.05).
Conclusions Bipolar electrogram polarity reversal allows more precise localization of breakthrough compared with the earliest activation, particularly in cases of wide synchronous PV activation.
The pulmonary veins (PVs) have been shown to trigger paroxysms of atrial fibrillation (AF) (1–11). Radiofrequency (RF) energy ablation at the PV ostia, with the end point of distal PV disconnection, is successful in eliminating AF in 70% of patients (3,4). Although the electrical connection between the left atrium (LA) and each PV has been shown to be often limited to specific breakthroughs (3), their anatomical distribution in the different PVs has not been described. Furthermore, there are reports in the literature (12–15)that bipolar electrogram polarity changes are observed at the insertion site of the left-sided accessory pathway, atrial tachycardia or cavotricuspid isthmus. Such a radial propagation of excitation front was hypothesized in the myocardial sleeves of PV from the breakthrough sites and assessed using a circumferential circular mapping catheter.
This study included 157 consecutive (and previously unreported) patients who underwent PV ablation for multidrug-resistant paroxysmal daily AF. Two mapping techniques were used to localize breakthrough sites. Group 1, consisting of the first 113 consecutive patients (72 men, mean age: 55 ± 8 years), was ablated guided only by mapping the earliest pulmonary vein potential (PVP) activation. Group 2, consisting of the subsequent 44 consecutive patients (22 men, mean age: 53 ± 9 years), underwent ablation guided additionally by electrogram polarity reversal mapping. The clinical characteristics of the patients in both groups are shown in Table 1. Twenty-three patients had evidence of cardiovascular disease (17 in group 1 and 6 in group 2): 12 had hypertension; 6 had coronary artery disease; 3 had dilated cardiomyopathy and 5 had mitral valve regurgitation. Informed consent was obtained from all patients before the procedure according to the protocol approved by the hospital’s Human Research Committee.
The study was performed as described previously (3). The LA and PVs were explored through either a patent foramen ovale (29 patients) or trans-septal catheterization with two catheters: one for circumferential PV mapping and a quadripolar mapping/ablation catheter. Direct visualization of all four PVs was performed using selective venography. A 6F or 7F NIH angio-catheter (Cordis-Europe, Amersfoort, Netherlands) was introduced through a long sheath into each PV under fluoroscopic guidance. Angiography was performed during held midexpiration by hand injection of contrast media (about 10 ml) in the anteroposterior view (and additional oblique views if needed) and displayed during the procedure to show the venous anatomy and the location of LA-PV junction. Heparin was titrated to maintain a partial thromboplastin time of 60 to 90 s (control = 30 s).
Pulmonary vein mapping was performed with a steerable circular catheter 15 or 20 mm in diameter (choice based on PV diameter as determined by angiography) equipped with ten 1-mm electrodes in a loop made of shape-retaining material (Lasso, Biosense Webster, Diamond Bar, California) orthogonal to the shaft (3). It was straightened to allow introduction into the 8F trans-septal sheath and deployed within the body of the LA, allowed to resume its shape and then pushed into the desired PV. Pulmonary vein muscle potentials were defined as described previously (3)and recorded in bipolar mode from 10 bipoles (1 to 2, 2 to 3, etc., up to 10 to 1 with the initial electrode as the anode and the next electrode as the cathode, as shown in Fig. 1A) through bandpass filters of 30 to 500 Hz and an amplification of 1 to 2 cm/mV on a polygraph (Midas PPG or Labsystem Bard). The activation sequences of the PVPs were assessed in the proximal PV (within the first centimeter from the ostium but distal to the level of ablation).
Definition of electrogram polarity reversal
As shown in Figure 1B, based on the hypothesis that the excitation front propagates radially in the PV myocardium from the breakthrough site, its activation vector along the PV perimeter would diverge in opposite directions from the breakthrough site. Consequently, the bipolar electrograms recorded on either side of the breakthrough by a circular mapping catheter would reflect opposite polarity with the breakthrough site as a boundary. Polarity reversal was defined as a sudden change of the main deflection of PVP from positive to negative when analyzing adjacent bipoles in ascending order (e.g., bipole 2 to 3: positive polarity and bipole 3 to 4: negative polarity, shown in Fig. 1A).
Arrhythmogenic PVs (defined on the basis of documented ectopy or AF initiation) were initially targeted by RF ablation, but in the absence of arrhythmia, 3 or 4 PVs were ablated (right inferior pulmonary vein [RIPV] was ablated in the presence of complex and delayed PVPs). Radiofrequency energy was applied as proximally as possible at the LA-PV junction determined by angiography and confirmed by catheter-tip “drop-off” during withdrawal. Segments of the PV perimeter were targeted on the basis of the bipole(s) from the 10 bipoles of the circular catheter showing the earliest activation (timing of the maximal peak of electrograms) in group 1 patients during sinus rhythm for the right PVs or pacing of the distal coronary sinus (or LA appendage) for the left PVs. In group 2 patients, mapping of PVs was performed in the same manner as in group 1, and, when electrogram polarity reversal was observed across adjacent bipoles showing the earliest activation, RF energy was preferentially delivered at sites with polarity reversal and, in its absence, RF energy was delivered at sites of earliest activation. The ablation catheter was positioned proximal to the site of earliest activation. Radiofrequency energy was delivered at contiguous sites showing synchronous PVP resulting in either no effect, abolition of all PVP (single breakthrough) or PV activation change with a shifting of earliest activity to another ostial sector (secondary breakthrough). The end point was elimination of PV muscle conduction distal to the ablation site(s) indicated either by abolition or dissociation of distal PVPs. Radiofrequency energy was delivered at the distal electrode of the thermocouple-equipped ablation catheter (target: 50°C) with a power limit of 20 to 30 W for 30 to 60 s at each site. If this power could not be reached (presumably because of reduced local blood flow), an irrigated-tip catheter (17 ml/min saline flow) was used with the same target temperature and power. After the electrical disconnection of targeted PVs, provocative maneuvers (isoproterenol and burst pacing) were performed to reveal other remaining foci from the ostium proximal to the ablation sites or from nonablated PV or atrial tissue. Additional RF ablation was performed targeting the remaining foci if necessary. Pulmonary vein angiography was repeated after 20 min of surveillance, with PV “stenosis” defined as a diameter reduction of >50%.
Definition of electrical breakthrough
Electrical breakthrough sites were defined by changes in PVP activation sequence as a result of localized RF delivery in a segment of PV perimeter and classified fluoroscopically into four segments around the ostium of each PV (top, bottom, anterior and posterior). Any encroachment into an adjacent segment was classified as a breakthrough involving both segments. The distribution of all breakthroughs in each vein was analyzed, and the prevalence in each segment of each vein was estimated.
The presence or absence of polarity reversal at segments with and without breakthrough was checked in all PVs that were successfully disconnected from LA in group 1 patients, and the predictive accuracies (sensitivity, specificity, positive and negative predictive values) for localizing the breakthrough correctly were calculated. The width of each breakthrough showing polarity reversal was estimated by the number of successive bipoles involved in polarity reversal, and both the mean width and the number of breakthroughs were compared among four PVs.
Comparison of the ablation results between two groups
The accuracy of breakthrough localization was assessed by comparing the total RF delivery time necessary for all PV disconnection between two groups and for disconnection of PVs with wide synchronous activation (defined as PVs with the earliest 10 ms of activation encompassing more than five bipoles), suggesting a broad breakthrough.
Retrospective analysis of endocardial PVP recordings (in group 1 patients) was performed by two authors independently. The interobserver agreement in the detection of electrogram polarity reversal was 86%. The difference between the observers was not statistically significant according to the McNemer’s analysis (p = NS). Differences were resolved by consensus.
All values are expressed as mean ± SD. Statistical analysis was done using the Student ttest (paired or unpaired) or chi-square analysis. One way analysis of variance followed by the Scheffe’s post-hoc test was also used when we needed to compare more than three groups. Differences with p < 0.05 were considered statistically significant.
A total of 411 PVs (305 PVs in group 1 and 106 PVs in group 2) were ablated, including 131 left superior pulmonary veins (LSPVs), 125 right superior pulmonary veins (RSPVs), 109 left inferior pulmonary veins (LIPVs) and 46 RIPVs. A total of 99% (301/305) of them were successfully electrically disconnected from the LA. In group 1, the earliest PV activation spanned 3.4 ± 2.0 bipoles; it was wider than 5 bipoles in 54 (18%) PVs (29 LSPVs, 25 RSPVs). The total RF application time required to disconnect the superior PVs was significantly longer than that in the inferior PVs (9.6 ± 4.4 min, 9.3 ± 4.1 min, 6.1 ± 3.8 min and 6.3 ± 3.5 min in LSPV, RSPV, LIPV and RIPV, p < 0.001). A significant difference was also observed in the total RF application time between PVs with and without wide synchronous activation (12.3 ± 3.4 min vs. 8.5 ± 3.8 min, respectively, p < 0.0001). Radiofrequency applications at the initial earliest segment eliminated all PVPs in 93 PVs (31%), indicating a single breakthrough (Fig. 2), whereas ablations at ≥2 segments were required in the other 208 PVs (69%) (Fig. 3).
The prevalence of breakthrough sites in the four anatomic segments for each PV in group 1 patients is shown in Figure 4. Their distribution varied significantly from one anatomic segment to another (p < 0.01) and also between superior and inferior PV (p < 0.01); the highest prevalence of breakthrough was located at the bottom of both superior PVs (85% and 84% in the LSPV and RSPV, respectively) and the top of both inferior PVs (70% and 77% in the LIPV and RIPV, respectively).
A bipolar electrogram polarity reversal was documented in 88% (470/534) of breakthrough segments and in 9% (51/565) of other segments in group 1 patients. Accordingly, the predictive accuracies of the polarity reversal for the presence of electrical breakthrough were as follows: sensitivity, 88%; specificity, 91%; positive and negative predictive values, 91% and 88%, respectively. The polarity reversal was frequently observed in adjacent two bipoles (91%) and less frequently spanned three successive bipoles (9%) (Figs. 2and 3). The number of successive bipoles showing polarity reversal was significantly less than those exhibiting the earliest activation (2.0 ± 0.4 bipoles vs. 3.4 ± 2.0 bipoles, p < 0.01) and was similar in all four PVs (2.0 ± 0.3 bipoles, 2.1 ± 0.4 bipoles, 2.1 ± 0.4 bipoles and 2.0 ± 0.4 bipoles in LSPV, RSPV, LIPV and RIPV, respectively, p = NS). Polarity reversal indicated that there were more breakthroughs in superior PVs compared with inferior PVs (2.1 ± 0.6 breakthroughs, 2.0 ± 0.6 breakthroughs, 1.6 ± 0.6 breakthroughs and 1.7 ± 0.7 breakthroughs in LSPV, RSPV, LIPV and RIPV, respectively, p < 0.01). Furthermore, a significant difference was also observed in the number of breakthroughs between PVs with and without wide synchronous activation (2.5 ± 0.5 breakthroughs vs. 1.9 ± 0.6 breakthroughs in PVs, respectively, p < 0.001).
In 44 group 2 patients, a total of 106 PVs were ablated guided by polarity reversal mapping, and all of them were successfully disconnected from the LA. The distribution of breakthrough in the four segments around the PV ostium was similar to that of group 1 (p = NS), with the highest prevalence located at the bottom of superior PVs (82% and 84% in the LSPV and RSPV, respectively) and at the top of inferior PVs (74% and 67% prevalence in the LIPV and RIPV, respectively). A total of 17 PVs (16%: 10 LSPVs and 7 RSPVs) showed the synchronous activation (similar proportion with group 1, p = NS). A total of 70% (168/238) of RF applications at sites with an electrogram polarity reversal resulted in PVP activation change or abolition (Fig. 5), which was significantly higher than that targeting at sites without a polarity reversal (17%: 12/69 sites, p < 0.001). The mean total RF application time required to disconnect PVs in groups 1 and 2 was not significantly different (all PVs: 8.2 ± 4.4 min vs. 7.6 ± 3.7 min, superior PVs: 9.5 ± 4.3 min vs. 8.9 ± 4.1 min, inferior PVs: 6.2 ± 3.7 min vs. 6.0 ± 2.5 min in group 1 and 2, respectively, p = NS). However, the total RF time required for disconnection of PVs with wide synchronous activation (54 PVs in group 1 and 19 PVs in group 2, as demonstrated in Fig. 5) was significantly shorter when guided by electrogram polarity reversal (12.3 ± 3.4 min vs. 10.3 ± 3.0 min in groups 1 and 2, respectively, p < 0.05). Mean procedure duration was similar in both groups (168 ± 35 min and 158 ± 43 min in groups 1 and 2, respectively, p = NS).
Additional RF ablations were performed afterwards in 57 patients (41 in group 1 and 16 in group 2), targeting 86 other foci revealed by provocative maneuvers, including untargeted PVs in 20 (13 RIPV), the PV ostia proximal to ablation in 38 and the atrial foci in 28. Side effects were observed in three patients: two acute PV stenosis (55% narrowing with pressure gradient of 2 mm Hg) and two pericardial effusions (not requiring drainage).
Sixty-four patients (41%) had recurrence of AF, and a reablation session was performed in 60 patients. No significant difference was observed in the recurrence rate between the two groups (42% and 39% in groups 1 and 2, p = NS) and also between patients with and without cardiovascular disease (39% and 52% in patients with and without cardiovascular disease, p = NS). The ectopy was related to a previously ablated PV in 39 patients (46 PVs) with recovery of all distal PVPs. Other ectopic beats or AF initiations were mapped to multiple sources, including previously untargeted PVs in 15 (11 RIPV), the PV ostia proximal to previous ablation in 21 and the atrial tissue in 18, requiring additional RF applications. With a mean follow-up of 9 ± 5 months after discharge, AF was completely eliminated in 116 patients (74%) without antiarrhythmic drug. No PV stenosis was noted during follow-up procedures.
This study shows that the anatomical distribution of electrical breakthroughs from LA to each PV is dominantly located at the bottom of both superior PVs and the top of both inferior PVs. A bipolar electrogram polarity reversal is observed at breakthrough sites with high specificity and sensitivity and provided useful additional information to identify their precise location.
Recent clinical evidence suggests that the electrical connections between the LA and the PVs are limited to specific breakthroughs rather than fully circumferential around the PV ostium (3). There have been a few reports (16,17)describing the complex architecture of the prolonged myocardial sleeves into PVs with bundles arranged in varying orientations and crossing each other; however, the exact anatomic arrangement providing these electrical breakthroughs from LA to PV has not been elucidated because of the extreme complexity of three-dimensional histologic reconstruction. The findings in the present study that breakthroughs defined electrophysiologically were distributed preferentially at the segments intervening between the superior and inferior PVs, may suggest a possible relationship with the embryologic development. The PV trunks are thought to be derived from a common vessel (common PV), which becomes progressively resorbed within the LA. This incorporation transforms the branches of this common vessel into separately inserting superior and inferior PV trunks (18–20). According to this sequence of embryologic development, the bottom segment of the superior vein and the top segment of the inferior vein being derived from a previously contiguous segment may, therefore, have common characteristics including the presence of LA-PV breakthroughs.
Electrophysiologically, the breakthrough was indicated by the earliest activation timing among circumferential PVPs. However, this data was sometimes ambiguous, notably with wide areas of synchronous timing. It has been previously reported that the reversal of the deflection of bipolar electrograms at adjacent recording sites indicates the propagation of activation front in opposite directions in patients with accessory pathway, atrial tachycardia or cavotricuspid isthmus (12–15). This condition would be particularly fulfilled by the circular mapping catheter located inside the PV, provided that the electrodes are placed orthogonally to the activation front. Although both mapping methods (the earliest activation and the polarity reversal mappings) are based on the same electrogram recordings, polarity reversal mapping provides complimentary, but different, information by indicating the direction of the activation vector. Electrogram polarity reversal was observed at segments with LA-PV breakthrough with high sensitivity (88%) and specificity (91%) and localized breakthroughs to a more limited (smaller) part of the PV perimeter compared with activation mapping. Radiofrequency applications at such sites changed PVP activation sequence (or abolished them) with relatively high probability (70%) so that a lesser total RF time was required to disconnect PVs showing wide synchronous activation compared with those guided by the earliest activation mapping alone. Furthermore, the criterion of polarity reversal is evident at first sight on the recordings, allowing more rapid mapping.
On the other hand, this criterion was less useful in PVs with narrower breakthroughs in which the region with earliest activation could be recognized easily by the PVP activation sequence. Electrogram polarity reversal was also observed at 9% of sites without electrical breakthrough, probably caused by secondary divergence of the wave front (e.g., branching of distal muscular fascicles); unnecessary RF deliveries at these bystander sites can be avoided by using the polarity reversal criterion in combination with the earliest activation mapping.
The findings of this study have practical clinical implications. Information about the dominant location of breakthroughs is important to reduce the RF energy to disconnect PVs and minimize the risk of occlusive PV stenosis (21). When the wide breakthroughs are noted in activation mapping, the polarity reversal mapping provides important superior information without requiring any other additional catheters or recordings. Its simplicity and effectiveness suggest that polarity reversal mapping can be used routinely as an additional indicator for mapping and ablation of PV, like the unipolar recording morphology or vector mapping (12,13)in the case of the Wolff-Parkinson-White syndrome.
The accuracy in detecting the earliest activation or polarity reversal may be affected by the orientation and position of the circular mapping catheter within the PVs. Lasso catheter can at times be inadequate to explore the electrical activity at the PV ostium, especially in the presence of large ostia with very early branching of the vein. Additional limitations are that the interobserver agreement was only 86% and that a difference in outcome could only be defined to those PVs with broad breakthrough.
Electrical breakthroughs connecting the LA to PVs are preferentially located in defined anatomic segments, particularly at the bottom of both superior PVs and the top of both inferior PVs. A bipolar electrogram polarity reversal, observed at breakthrough sites with 88% sensitivity and 91% specificity, provides a useful indicator to identify their precise location in addition to the earliest activation mapping, and minimize RF delivery.
- atrial fibrillation
- left atrium
- left inferior pulmonary vein
- left superior pulmonary vein
- pulmonary vein
- pulmonary vein potential
- right inferior pulmonary vein
- right superior pulmonary vein
- Received July 31, 2001.
- Revision received January 7, 2002.
- Accepted January 18, 2002.
- American College of Cardiology Foundation
- Chen S.A,
- Hsieh M.H,
- Tai C.T,
- et al.
- Haïssaguerre M,
- Shah D.C,
- Jaïs P,
- et al.
- Scheinman M.M,
- Morady F
- Lau C.P,
- Tse H.F,
- Ayers G.M
- Ernst S,
- Schlutter M,
- Ouyang F,
- et al.
- Pappone C,
- Rosanio S,
- Oreto G,
- et al.
- Natale A,
- Pisano E,
- Shewchik J,
- et al.
- Damle R.S,
- Choe W,
- Kanaan N.M,
- Ehlert F.A,
- Goldberger J.J,
- Kadish A
- Yamabe H,
- Okumura K,
- Misumi I,
- et al.
- Neil C.A
- Robbins I.M,
- Colvin E.V,
- Doyle T.P,
- et al.