Author + information
- Received July 28, 2015
- Revision received September 1, 2015
- Accepted September 3, 2015
- Published online February 1, 2016.
- Samuel H. Baldinger, MD,
- Saurabh Kumar, BSc(Med)/MBBS, PhD,
- Chirag R. Barbhaiya, MD,
- Koichi Nagashima, MD, PhD,
- Laurence M. Epstein, MD,
- Roy John, MD, PhD,
- Usha B. Tedrow, MD, MSc,
- William G. Stevenson, MD and
- Gregory F. Michaud, MD∗ ()
- Cardiac Arrhythmia Center, Cardiovascular Division, Brigham and Women’s Hospital, Boston, Massachusetts
- ↵∗Reprint requests and correspondence:
Dr. Gregory F. Michaud, Cardiovascular Division, The Cardiac Arrhythmia Center, Brigham and Women’s Hospital, 75 Francis Street, Boston, Massachusetts 02115.
Objectives This study sought to assess loss of pulmonary vein (PV) excitability to pacing relative to the development of entrance block and the anatomic completion of the circumferential radiofrequency ablation (RFA) line.
Background During encircling RFA for PV isolation (PVI), entrance block develops before anatomic completion of encirclement (early) in some patients. We hypothesized that early entrance block may be associated with loss of PV excitability to pacing.
Methods In 30 patients undergoing PV isolation (age 61 ± 10 years, 21 men), excitability to pacing was assessed at predefined PV sites when entrance block developed and after completion of the RFA line.
Results Of 60 PV pairs, 37 developed entrance block early, with a gap ≥10 mm in the RFA line. In only 35% of PV pairs in this subgroup, both PV sleeves captured, and all of the capturing PV pairs showed exit block (no conduction from PV to atrium) despite the presence of an excitable gap. In the remaining 23 PV pairs, entrance block did not occur until encircling RFA was anatomically complete. In 83% of these PV pairs, both sleeves captured with exit block (p < 0.001 compared with early block PVs).
Conclusions The majority of PV pairs develops entrance and exit block before complete anatomic encircling by RFA lesions. Early entrance block is frequently associated with loss of PV sleeve excitability, consistent with a spreading wave of injury or edema rather than a permanent conduction barrier. This may help to explain the significant rate of PV conduction recovery associated with the acute endpoints of entrance and exit block.
It is common practice to place radiofrequency (RF) ablation (RFA) lines widely around pulmonary vein (PV) ostia to include antral tissue, which may target potential triggers and focal sources that may contribute to atrial fibrillation (AF) initiation and persistence (1). Entrance block has been observed during wide area circumferential PV isolation (PVI) despite gaps >3 cm in the RFA line (2,3). Additionally, PVs may become unexcitable to pacing remote from RFA lesions during PVI procedures (4). The reason for loss of excitability in PVs is not clear.
We hypothesized that if a spreading wave of injury or edema contributes to entrance block before anatomic completion of the encircling RFA line, it may also spread into the PV sleeve rendering the vein unexcitable to pacing. If so, then PVs demonstrating entrance block before anatomic completion of the circumferential RFA line would be more likely to become unexcitable to pacing than PV pairs that require a completed RFA line before achieving entrance block. PV unexcitability may be an important mechanism of entrance block in the former case and influence the interpretation of exit block.
We analyzed consecutive patients undergoing RFA for AF who presented for the procedure in sinus rhythm. Patients with prior left atrial (LA) ablation or PVs that were unexcitable to pacing at baseline (see the section on pacing maneuvers) were excluded.
Written informed consent for the procedure was provided by all patients, and procedures were performed according to protocols approved by the Brigham and Women's Hospital Human Subject Protection Committee.
Mapping and ablation
PVI was performed as described in the Online Methods. Briefly, procedures were performed under general anesthesia. Intracardiac ultrasound was used for visualizing transseptal punctures and improving catheter contact. Deflectable (Agilis, St. Jude Medical, Saint Paul, Minnesota) and fixed-curve (SR0, St Jude Medical) sheaths were inserted into the LA. Using the CARTO 3 (Biosense Webster, Diamond Bar, California), electroanatomic mapping system (EAMS) a detailed anatomic map was generated for LA anatomy by using a multipolar, multispline diagnostic catheter (Pentaray NAV, Biosense-Webster). Bipolar voltage was considered abnormal when it measured <0.5 mV (Figure 1A) (5). Ablation and pacing were performed using a 3.5-mm-tip catheter (ThermoCool SF or SmartTouch, Biosense Webster). Goals for individual lesions included average contact force >10 g and impedance decrease of 10 Ω or greater. Additional endpoints of the procedure were electrical isolation of all PVs by antral ablation without carinal lesions, contiguous anatomic completion of the RFA line, unexcitability of the RFA line (6), and lack of dormant conduction following 12 mg of intravenous adenosine.
Assessment of pace excitability at baseline
After creation of a detailed voltage map of the LA and the PVs, local excitability to pacing from the ablation catheter was assessed at multiple locations within the normal voltage (>0.5 mV) zone in the PVs, including distal sites, as well as on the posterior LA wall. PV sleeve capture was defined as direct capture of near-field PV potential and far-field capture was carefully excluded. All locations where capture could be confirmed were tagged on the EAMS. All bipolar pacing maneuvers were performed with an output of 10 mA and pulse width of 2 ms. Good catheter contact was ensured by mechanical feedback, catheter motion and contact force measurement (>10 g) in 25 of 30 cases.
Design of the RFA line and assessment of entrance block
Point-by-point lesions approximately 1 cm from the PV ostia were applied contiguously to encircle PV pairs. Importantly, special attention was paid to ensure a continuous RFA line with space <5 mm between the centers of lesion markers. We performed RFA on the posterior LA wall as the last part of the ablation ring. The multispline catheter was initially placed in the corresponding superior PV during RF application to demonstrate entrance block (sudden loss of PV electrograms) and then moved to the ipsilateral inferior PV. RF application was continued until entrance block into both ipsilateral PV sleeves was confirmed. At that time, a new voltage map was acquired using the remap feature of the EAMS that allows acquisition of new voltage points on an existing anatomic shell (Figure 1B) and the pacing maneuvers described below were performed before eventually completing the circumferential line anatomically in all patients.
Categorization of PV pairs
A PV pair was categorized as requiring a completed RFA line for isolation if both ipsilateral PVs showed entrance block coincidentally with completion of circumferential RFA line or with a gap <10 mm between the centers of bordering lesion markers. This cutoff value was used based on the assumption that lesions could have a radius of 5 mm such that a 10-mm gap could actually represent a completed line. A PV pair was categorized as isolated before RFA line completion if there was persistent entrance block into both PV sleeves before the circumferential line was completed (≥10 mm discontinuity between bordering lesions) (Figure 1B). Distances were measured using the measuring tool in the EAMS.
Reassessment of pace excitability
Figure 2 outlines the workflow for all pacing maneuvers. For PV pairs that isolated before completion of the RFA line, pace assessments were performed 3 times: at baseline, after entrance block with an apparent gap in the RFA line, and after the RFA line had been completed anatomically. For PV pairs that required a completed RFA line for isolation, the pacing maneuvers were performed twice: at baseline and after completion of the RFA line.
Excitability along the incomplete RFA line
For PV pairs that isolated before completion of the RFA line, pace excitability was assessed at multiple spots within the apparent gap, within 5 mm of the venous aspect of the gap and within 5 mm of the atrial aspect of the gap (Figure 2). Tissue could either be unexcitable to pacing or excitable and electrically connected to the LA only, to the PVs only, or to both. Pacing points were tagged with different colors accordingly on the EAMS.
Excitability of PV sleeves and PV interconnection
Pace excitability was reassessed after entrance block for both ipsilateral PV sleeves individually by pacing from the ablation catheter at sites where capture was confirmed at baseline while recording electrograms from the multispline catheter positioned in the same PV. If capture was not demonstrated, multiple additional locations were sampled to cover the entire PV sleeve including proximal and distal sites and sites just inside the line of ablation, which may represent antral tissue. Local PV capture was defined as presence of 1:1, near-field PV electrograms observed on the multispline catheter while pacing from the ablation catheter.
Exit block was defined as dissociation of sinus rhythm and captured PV sleeves as judged by electrogram activation sequence and timing on the duodecapolar catheter relative to the multispline catheter.
PV interconnection was defined as presence of 1:1 near-field PV electrograms recorded from the multispline catheter positioned in the superior PV while pacing from the ablation catheter in the ipsilateral inferior PV or vice versa.
Continuous variables are presented as mean ± SD. Categorical variables are expressed as counts (percentage). Differences of continuous variables were tested for statistical significance by using the Student t test and differences in categorical variables by the Pearson’s chi-square test.
Correlation between the gap length and electrical unexcitability of PV pairs was calculated using a generalized estimating equation model to account for multiple observations (left and right PV pairs) in individual patients. A 2-tailed p value of <0.05 was considered statistically significant. Data analysis was performed using IBM SPSS Statistics for Mac, version 22.0 (IBM, Armonk, New York).
A total of 32 consecutive patients were assessed. Two patients were excluded because at least 1 PV sleeve was completely unexcitable at baseline. The baseline voltage maps of these 2 patients showed extensive areas of scar in the LA. Patient characteristics are summarized in Table 1. The initial voltage map of the LA included 680 ± 196 electroanatomic points. At baseline, voltage maps of all PV pairs showed normal voltage (>0.5 mV) at the posterior wall, the PV antra, and PV ostia. On average, circumferential normal voltage ranged 17 ± 4 mm into the left superior PV sleeve, 15 ± 4 mm into the left inferior PV sleeve, 18 ± 3 mm into the right superior PV sleeve, and 15 ± 3 mm into the right inferior PV sleeve, measured from the carina.
Entrance block relative to completion of the RFA line
In total, 37 PV pairs (62%) in 24 patients showed entrance block before RFA line completion (17 left and 20 right). Figure 3 shows a flow diagram of the results. The gap length in the RFA line at the time of entrance block was 17 ± 6 mm for left and 20 ± 8 mm for right PV pairs (p = 0.198).
Voltage maps acquired after entrance block clearly demonstrated a line of conduction block connecting the edges of the incomplete RFA line and spanning the gap either directly or in a convex shape (Figures 1 and 4B), indicating in some cases a relatively small part of the PV antrum was still electrically connected to the LA. Of note, these voltage maps only visualize the line of conduction block. The low-voltage area within the isolated PVs does not indicate scar but rather the absence of local electrical activity.
The remaining 23 PV pairs (38%) required a completed RFA line and showed entrance block with a gap <10 mm or after additional ablation along the line. The depth of circumferential normal voltage in the PV sleeves measured at baseline was 16 ± 3 mm in this group versus 16 ± 4 mm in early isolating PV pairs (p = 0.997). Ablation data are provided in Table 2.
Excitability of atrial tissue near the RFA line and PV pairs isolating before RFA line completion (n = 37)
Tissue within the apparent gap was excitable to pacing and connected only to the LA in 32 of 37 PV pairs (87%).
In 5 PV pairs (14%), excitable tissue that was connected to the LA extended ≥5 mm to the venous aspect of the gap (Figure 4B). In 16 PV pairs (43%), the venous aspect of the gap captured the PVs with exit block despite the incomplete RFA line, and in another 16 PV pairs (43%), the area was unexcitable to pacing.
After RFA line completion, the former gap region (now RFA line) was electrically unexcitable in all PV pairs, which was one of our pre-determined endpoints for PVI. The venous aspect of the gap within 5 mm was now unexcitable in 30 PV pairs (81%) and captured the PVs with exit block in 7 PV pairs (19%).
When pacing the atrial aspect of the gap, capture without connection to the PVs was demonstrated in all 37 PV pairs (100%) both before and after the RFA line was completed.
Electrical excitability of PV pairs
At baseline, local pace capture in both PV sleeves could be confirmed consistently in all PV pairs. Of 37 PV pairs that showed entrance block despite a gap ≥10 mm, 19 (51%) were found to have lost excitability to pacing in at least 1 PV sleeve at the time of entrance block (both sleeves in 10 and 1 sleeve in another 9 PV pairs) (Figure 4B). The remaining 18 PV pairs showed local capture with exit block despite the incomplete RFA line.
An additional 5 PV pairs of this group lost excitability after completion of the RFA line, such that 24 of 37 early isolating PV pairs (65%) lost excitability to pacing during the procedure. Of the 13 PV pairs that remained completely excitable in this group, 5 lost interconnection between ipsilateral PV sleeves.
There was a strong correlation between longer gap length at the time of entrance block and unexcitability of a PV pair: With every additional 10 mm in gap length, the odds for PV unexcitability increased by a factor of 2.7 (1.4 to 5.2), p = 0.003.
As shown in Figure 5, significantly fewer PV pairs with no or minimal gaps in the RFA line at the time of entrance block lost excitability to pacing (4 of 23, 17%, p < 0.001) (both sleeves in 1 and 1 sleeve in 3 PV pairs) (Figure 4A). All of the PV pairs that remained completely excitable showed exit block, and 1 lost interconnection between ipsilateral PVs.
Dormant conduction with adenosine
Adenosine was administered in every patient after circumferential RFA lines were completed. Dormant conduction in 1 PV pair was observed in a single patient (4%).
When placing sequential, contiguous RF lesions for circumferential wide antral PVI:
1. The majority of PV pairs develop entrance block before anatomic completion of the encircling RFA line, at a time when there is a large electrically viable gap, suggesting that an anatomically continuous line of RFA lesions is not the primary mechanism for entrance block in such cases.
2. The majority of PV pairs that show entrance block in the absence of a complete circumferential RFA line become unexcitable to pacing, even at sites remote from the RFA line. Those that remain excitable to pacing show exit block despite the anatomically incomplete RFA line.
3. PV pairs that require an anatomically complete RFA line to achieve entrance block generally remain excitable to pacing indicating that the PV muscle sleeve is still functional, but conduction is blocked proximally. All excitable PV pairs that had entrance block also showed exit block in this series.
4. Loss of excitability to pacing in the PVs is more likely if the anatomic gap in the RFA line is larger.
Our data confirm that acute PV entrance and exit block occurs frequently despite an anatomically incomplete circumferential RFA line. The finding that tissue within the ablation gap is excitable and connected to the LA in the vast majority of such PV pairs indicates that the gap area is viable and unlikely to create a permanent conduction barrier if the PV sleeves recover conduction. These observations suggest that initial development of PV entrance and exit block is not a sufficient endpoint for ablation in many cases. This is consistent with randomized studies using this acute endpoint for PVI and found that recurrences are commonly associated with PV conduction recovery (6,7). At histological examination, Kowalski et al. (8) showed that PVs exhibiting conduction recovery after antral ablation frequently had anatomic gaps at the sites of catheter ablation. There is evidence from multiple animal studies that conduction can persist through viable tissue in gaps as small as 0.1 mm, and chronic conduction block rarely occurred in gaps larger than 5 mm (9–11). Miller et al. (3), however, showed that acute PVI can often be achieved without complete circumferential ablation and that more than 25% of these PVs exhibit dormant conduction, suggesting that acute but probably reversible tissue injury caused by tissue stunning, edema, ischemia or other factors participate in acute PVI. Squara et al. (4) showed that loss of local PV capture is common after antral PVI, but the relationship of unexcitability to gaps was not explored. A common assumption is that a continuous circumferential conduction barrier is created by RFA lesions that result in PV sleeves becoming electrically disconnected from the LA. This may be accomplished segmentally within the ostium of PVs, because there may be gaps in electrically viable tissue at this level. During wide antral circumferential ablation, a noncontiguous line is unlikely to produce a conduction barrier because electrically viable tissue exists circumferentially.
The mechanisms of PV entrance block before anatomic completion of the RFA line are not entirely clear. Edema in PVs following RF ablation has been described (12), and Arujuna et al. (13) showed with magnetic resonance imaging that reversible and irreversible injury occurs and that the amount of reversible injury is a predictor of AF recurrence.
We describe another possible means of PV entrance block during RFA possibly caused by tissue stunning, edema. or injury wave fronts propagating from the RFA line to PV sleeves to alter electrical excitability of PV tissue. In most cases, this leads to unexcitability of entire PV sleeves such that wave fronts propagating through the gap before the antrum is completely encircled with RFA lesions encounter electrically unexcitable PV sleeves. In some cases, segmental PV–LA or PV–PV electrical connections seem to be affected, leaving the bulk of the PV sleeve electrically excitable, but importantly, disconnected from the atrium (exit block despite a gap).
We hypothesize that if edema is minimal, completion of an anatomic barrier is necessary for PV entrance block and the vast majority of these PVs remain excitable to pacing, that is, a captured PV electrogram is seen during PV pacing.
We believe that when PV entrance and exit block occurs before a complete encircling RFA line, leaving 1 or more gaps is likely to be followed by recovery of PV excitability that would result in electrical reconnection to the LA through the electrically viable gap(s), presumably due to resolution of edema and transient injury. However, if the RFA line is anatomically completed by contiguous, high-quality RFA lesions, recovery of excitability of the PV sleeve alone would not result in electrical reconnection between PVs and the LA unless a lesion on the RFA line recovers excitability. Reported evidence of reconnection at gaps (3) and significant PV conduction recovery rates (14–19) supports the concept that transient injury both on the RFA line and within PV sleeves contribute to acute PVI when entrance and exit block are used as primary endpoints for PVI procedures.
Alternative explanations for the observed phenomenon include PV ischemia or infarct, which may cause permanent unexcitability. Another possible explanation is that a critical mass of tissue is necessary for pace capture and that muscular sleeves of PVs that lose excitability may have a smaller mass of tissue (20). Our findings suggest this is not a major factor because it would not explain the correlation between pace excitability and a larger size gap when entrance block occurs. In addition, PV sleeve length was not different for the gap and the no-gap groups at baseline, and all PV sleeves captured along the length of the normal voltage zone. Also, small areas of PV tissue near the gap, but within the encirclement, may remain excitable and connected to the LA when the bulk of the PV sleeve was isolated and unexcitable (Figure 4B) or excitable but disconnected.
Interestingly, PV tissue within the circumferential lesion set, but remote from lesions, seems more vulnerable to the spreading wave of injury or edema than the atrial tissue outside the lesion ring, which maintained excitability in all patients. A possible explanation may be susceptibility to injury that is intrinsic to PV tissue or its blood supply, although this study does not allow us to determine the exact mechanism for unexcitability.
Edema or transient injury may account for the higher rates of dormant PV conduction exposed by adenosine administration reported after PVI (21). In our study, only 4% of patients exhibited dormant conduction after anatomic completion of the ablation ring with careful attention to lesion contiguity and quality.
Our study has important clinical implications for AF ablation procedures, exposing a potential pitfall of using entrance and exit block assessed from PV pacing as an endpoint for an AF ablation procedure. Entrance and exit block is often seen before an encircling RFA line is completed, and should not be relied on as a marker of a complete conduction barrier caused by RFA lesions. Initial entrance and exit block appears to frequently occur when the PV sleeve remote from the RFA line loses excitability and can involve the interconnection between PVs as well as the LA (20). Interconnection should be confirmed before testing, or each PV sleeve should be tested separately for exit block and for dormant conduction with adenosine. If this PV unexcitability reflects transient injury, as seems likely, a high rate of PV conduction recovery would be expected in this situation.
We completed all the RFA lines at the posterior wall. We chose this method since tissue is difficult to assess on the anterior wall or septum because of thin ridges and adjacent, sometimes overlapping structures. Therefore, all gaps were located at the posterior wall of the LA, and we cannot validate whether our findings related to gaps would be similar in other areas of the LA.
Based on the study by Miller et al. (3), which demonstrated a high rate of conduction recovery through gaps following adenosine injection, we did not administer adenosine until the RFA line was complete. Also, given the historically high rates of PV conduction recovery using PV entrance and exit block as the primary endpoint, the RFA line was anatomically completed in every patient, and the transient nature of the PV entrance and exit block and unexcitability are merely presumed, which the literature strongly supports.
Occult PV sleeve capture after entrance block may have led to overestimation of unexcitable PVs. Nevertheless, all efforts were taken to unmask occult capture by repositioning the catheters multiple times to cover the whole sleeve when no PV sleeve capture was evident. Gain settings and display speed were adjusted if necessary.
The reason why certain PVs appear to be more vulnerable to potentially transient injury remains unclear.
The majority of PV pairs develops entrance and exit block before complete anatomic encircling by RFA lesions. Despite the presence of a gap, these PV sleeves frequently lose excitability to pacing. The strong correlation between longer gap length in the RFA line at the time of initial entrance block and the likelihood of PV unexcitability suggests edema or tissue stunning as a potentially important factor leading to PVI. This may help to explain the significant rate of PV conduction recovery after the acute endpoints of entrance and exit block are achieved. In the presence of gaps in the wide area encircling line, entrance and exit block would be expected to be temporary, and efforts to complete the anatomic line with high-quality circumferential RFA lesions to create the conduction barrier should be pursued.
COMPETENCY IN MEDICAL KNOWLEDGE: The common concept of PVI is that a continuous conduction barrier around PVs caused by RFA lesions results in isolation. Our data show that a continuous conduction barrier is often not present despite the presence of entrance and exit block.
TRANSLATIONAL OUTLOOK: Pathophysiological mechanisms responsible for alterations in electrical characteristics of PVs and clinical implications of incomplete RFA lines on PV reconnection and AF recurrence need to be studied. These data suggest that entrance and exit block are a poor indicator of circumferential RFA line completion, thus further study is warranted to verify this speculation.
The authors thank Timothy Campbell, Kathryn Finnerty, and Garren Su of Biosense Webster.
For an expanded Methods section and a supplemental figure, please see the online version of this article.
Dr. Baldinger has received educational grants from the University Hospital of Bern, Switzerland, and the Swiss Foundation for Pacemakers and Electrophysiology. Dr. Epstein is a consultant for BSC, Medtronic, St. Jude Medical, and Spectranetics; and is on the Speakers Bureaus of Medtronic, St. Jude Medical, and Spectranetics. Dr. John receives consulting fees from St. Jude Medical and Biosense Webster; and lecture fees from St. Jude Medical and Boston Scientific. Dr. Tedrow receives consulting fees from St. Jude Medical; and research funding from Boston Scientific and Biosense Webster. Dr. Michaud receives consulting fees from St. Jude Medical; honoraria for speaking/lectures from Boston Scientific, Medtronic, St. Jude Medical, Atricure, and Biosense Webster; and research funding from Boston Scientific and Biosense Webster. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- electroanatomic mapping system
- left atrium
- pulmonary vein
- pulmonary vein isolation
- radiofrequency ablation
- Received July 28, 2015.
- Revision received September 1, 2015.
- Accepted September 3, 2015.
- American College of Cardiology Foundation
- Proietti R.,
- Santangeli P.,
- Di Biase L.,
- et al.
- Miller M.A.,
- d'Avila A.,
- Dukkipati S.R.,
- et al.
- Stiles M.K.,
- John B.,
- Wong C.X.,
- et al.
- Steven D.,
- Sultan A.,
- Reddy V.,
- et al.
- Calkins H.,
- Kuck K.H.,
- Cappato R.,
- et al.
- Kowalski M.,
- Grimes M.M.,
- Perez F.J.,
- et al.
- Arujuna A.,
- Karim R.,
- Caulfield D.,
- et al.
- Nanthakumar K.,
- Plumb V.J.,
- Epstein A.E.,
- Veenhuyzen G.D.,
- Link D.,
- Kay G.N.
- Nilsson B.,
- Chen X.,
- Pehrson S.,
- Kober L.,
- Hilden J.,
- Svendsen J.H.
- Cappato R.,
- Negroni S.,
- Pecora D.,
- et al.
- Ouyang F.,
- Antz M.,
- Ernst S.,
- et al.