Author + information
- Received March 11, 2017
- Revision received June 27, 2017
- Accepted July 13, 2017
- Published online September 27, 2017.
- Patrizio Pascale, MD∗ (, )
- Laurent Roten, MD,
- Ashok J. Shah, MD,
- Daniel Scherr, MD,
- Yuki Komatsu, MD,
- Khaled Ramoul, MD,
- Matthew Daly, MD,
- Arnaud Denis, MD,
- Nicolas Derval, MD,
- Frédéric Sacher, MD,
- Mélèze Hocini, MD,
- Michel Haïssaguerre, MD and
- Pierre Jaïs, MD
- Hôpital Cardiologique du Haut-Lévêque and Université de Bordeaux, IHU LIRYC ANR-10-IAHU-04, Bordeaux-Pessac, France
- ↵∗Address for correspondence:
Dr. Patrizio Pascale, Service de Cardiologie, Centre Hospitalier Universitaire Vaudois–BH 09-792, Rue du Bugnon 46, 1011 Lausanne, Switzerland.
Objectives The purpose of this study was to describe and identify useful electrocardiographic characteristics to help identify the mechanism of atrial tachycardia (AT) occurring after persistent atrial fibrillation (PsAF) ablation.
Background Electrocardiographic analysis to help identify the mechanism of AT after PsAF ablation is much limited by the fact that remodeling and ablation alter the normal activation pattern.
Methods All consecutive patients who underwent mapping and ablation of AT after PsAF ablation were included. Surface P waves were analyzed during higher (>2:1) grades of atrioventricular block.
Results One hundred ninety-six ATs with visible P waves were identified in 127 patients (macro–re-entry in 57%, centrifugal AT in 43%). One-third displayed low-voltage P waves (≤0.1 mV). An isoelectric line >80 ms was more common in centrifugal compared with macro–re-entrant AT (47% vs. 24%, p < 0.001), but its positive predictive value was limited (60%). A minority of peritricuspid ATs displayed the classic saw-tooth pattern (27% [n = 22]). However, the “precordial transition” (a gradual transition from an upright component in lead V1 to a negative component with progression across the precordium) remained often observed and specifically identified peritricuspid AT (specificity, 98%; sensitivity, 59%). Only 2 unique features could help identify perimitral AT (n = 60). First, the presence of a negative or negative-positive P-wave in any of leads V2 to V6 identified perimitral AT with 97% specificity and 30% sensitivity. Second, a “notched” negative component at the beginning of a positive P-wave in the inferior leads specifically identified clockwise perimitral AT (specificity, 98%; sensitivity, 25%).
Conclusions Only few unique electrocardiographic characteristics help identify the mechanism of AT after PsAF ablation. Knowledge of these characteristics may aid in planning and performing ablation.
Catheter ablation of persistent atrial fibrillation (PsAF) often requires adjunctive substrate modification strategies beyond pulmonary vein (PV) isolation to achieve a higher success rate. Strategies to modify the fibrillatory substrate have included electrogram (EGM)–based ablation (1) and left atrial (LA) linear lesions (2). The expansion of such ablation approaches that include a greater amount of LA ablation is associated with a high incidence of atrial tachycardia (AT) (3–6).
The characterization of electrocardiographic (ECG) signature P-wave configurations for these ATs is much limited by spontaneous remodeling and by the fact that substrate ablation and/or the creation of lines of block alter the normal activation pattern of the left atrium. Moreover, LA activation may be either silent or overshadowed by right atrial activation, as ablation often leads to a marked reduction in LA voltage. Better knowledge of the ECG features that apply to these challenging post-ablation ATs may provide the operator with a probabilistic clue of the AT mechanism and help avoid more exhaustive activation and entrainment mapping. It may also allow more appropriate procedural planning and discussion of the success rate and risks with patients.
The purposes of this study were therefore to: 1) describe the ECG characteristics and the features that apply to AT occurring after PsAF ablation; and 2) identify new clues that may potentially assist in the identification of their mechanism.
All consecutive patients who underwent mapping and ablation of sustained AT arising during or after ablation of symptomatic drug-refractory PsAF at our institution from January 2009 to December 2011 were enrolled in the study. The study was approved by the institutional ethics committee, and all patients provided written informed consent.
Electrophysiological study and ablation sequence for PsAF
All antiarrhythmic drugs except amiodarone were stopped ≥5 half-lives before ablation. Surface electrocardiograms and bipolar intracardiac EGMs were monitored continuously and stored on a computer-based digital amplifier and recorder system (Labsystem Pro, Boston Scientific, Natick, Massachusetts). The following catheters were introduced through the right femoral vein: 1) a steerable decapolar catheter was positioned within the coronary sinus (CS); 2) a 10-pole circumferential catheter (Lasso, Biosense Webster, Diamond Bar, California) was used for PV mapping; and 3) a 3.5-mm externally irrigated-tip ablation catheter (Biosense Webster). The Lasso or the ablation catheter was stabilized with a long sheath (SLO, St. Jude Medical, St. Paul, Minnesota) perfused continuously with heparinized solution.
The index procedure for PsAF was performed by using a stepwise ablation approach with a procedural endpoint of atrial fibrillation (AF) termination. If AF converted into AT, ablation was performed until the restoration of sinus rhythm. Details of this particular ablation approach have been published previously (2). In brief, as the first step, circumferential PV isolation was performed. The second step, EGM-guided LA ablation, targeted sites displaying complex fractionated EGMs and locally short AF cycle length. Linear LA ablation, the third step, targeted the LA roof, followed by the mitral isthmus. In the presence of shorter AF cycle length in the right atrium, EGM-guided ablation was performed in that chamber. Cavotricuspid isthmus (CTI) ablation was performed in most patients.
Classification of AT
AT was classified as macro–re-entry or centrifugal AT, the latter including focal AT and localized re-entry (7). Macro–re-entry was defined as re-entry around a large central obstacle. A post-pacing interval exceeding the AT cycle length by no more than 30 ms on 2 opposite segments of the central obstacle was required for diagnosis. Centrifugal AT was defined as atrial activity originating from a single focus and spreading out centrifugally. If activity accounting for >75% of the AT cycle length was present in an area with a diameter ≤3 cm, localized reentry was considered, and if not, focal AT was considered. The site of origin of centrifugal AT had to be confirmed by successful ablation.
Mapping of post-ablation AT
All ATs were conventionally mapped using a deductive strategy detailed previously (7). Briefly, a combination of activation and entrainment mapping was then used to determine the mechanism and location of AT. As a first step, the possibility of a macro–re-entry was investigated. Atrial activation was mapped by systematically comparing the mapping catheter signals with a reference channel from the decapolar CS catheter. In the presence of a consistent activation sequence of the CS, a perimitral circuit was suspected or ruled out on the basis of the activation sequence of the anterior mitral annulus. In the absence of sequential circumferential activation of the mitral annulus, a roof-dependent re-entry was mapped by looking for a cranial or caudal activation of the anterior and posterior LA walls. In the presence of similar or opposite directions of activation on both walls, roof-dependent re-entry was ruled out or ruled in, respectively. Entrainment mapping was performed in 2 opposite segments as guided by the activation mapping. In the presence of a post-pacing interval exceeding the AT cycle length by more than 30 ms in any segment, LA macro–re-entry was ruled out. Centrifugal AT was then mapped, initially in the left atrium and then in the right atrium. Repeated pacing maneuvers with analysis of the post-pacing interval were also used to progressively approach the site of origin.
Surface ECG analysis
The 12-lead electrocardiogram was read at a sweep speed of 25 mm/s and a standard gain of 1 mV/cm with filter settings of 0.05 Hz (high pass) and 100 Hz (low pass). A sweep speed of 100 mm/s was used for interval measurements. The surface P-wave configurations were analyzed during higher (>2:1) grades of atrioventricular block. Recordings of 2:1 block were used when the P waves that preceded the QRS complexes were free of T waves. Very low amplitude P waves, defined as a maximal recorded voltage lower than 0.05 mV in all 12 leads, were excluded from further analysis. Surface P-wave polarity was designated as being upright, negative, biphasic, isoelectric, or multicomponent. Flat polarity was defined as an amplitude ≤0.03 mV. P-wave duration and the presence of isoelectric intervals between P waves in all 12 leads were also evaluated. Isoelectric intervals were defined as the absence of surface ECG activity for ≥80 ms simultaneously on all 12 leads (8). The timing of the P-wave components was also evaluated between each different limb and precordial lead. The ECG analysis was performed in a blinded fashion.
Continuous variables are presented as arithmetic mean ± SD or as median values and ranges where indicated. Categorical variables are expressed as absolute numbers and percentages. Categorical variables were compared using the chi-square test or the Fisher exact test and continuous variables using the unpaired Student t test or the Mann-Whitney U test, as appropriate. For ECG features that may help identify the mechanism of AT, we used bootstrapping to test the stability of significance by calculating the 95% confidence interval of each variable’s p value. All tests were 2-tailed, and statistical significance was assumed for p values ≤ 0.05. Statistical analysis was performed using SPSS version 17.0 (SPSS, Chicago, Illinois).
A total of 227 successfully diagnosed sustained ATs were analyzed in 142 consecutive patients who underwent 162 procedures. Twenty-one ATs (9%) were excluded because surface P waves were constantly obscured by the T-wave or QRS complex despite exhaustive scanning of the ECG recordings. Ten additional ATs (5%) were excluded because of very low amplitude P waves (i.e., maximal voltage <0.05 mV). A total of 196 ATs that occurred in 127 patients were therefore included in the present analysis.
Patients’ mean age was 59 ± 10 years, and 80% were men. Patients had their first episodes of AF a median of 84 months before ablation (range: 5 to 276 months) and had PsAF for a median of 10 months (range: 0.5 to 132 months). Treatment with 2.2 ± 0.9 antiarrhythmic drugs, including amiodarone in 47%, had failed. The mean LA diameter was 46 ± 7 mm in the parasternal window. Forty-two patients (30%) had underlying structural heart disease. The mean left ventricular ejection fraction was 58 ± 10%. AT occurred either during the index ablation procedure (27%) or late after AF ablation (73%).
During the index AF ablation procedure, EGM-guided ablation was performed in 88% of cases. Linear ablation had been performed previously to the occurrence of AT in 78% of cases, with a mean of 1.8 ± 1.2 lines performed.
Mechanism of AT following PsAF ablation
The mechanism of AT was macro–re-entry in 57% of cases (n = 111). Centrifugal AT confirmed by successful ablation was diagnosed in the remaining 43% (n = 85).
Of the 111 macro–re-entrant ATs, 60 were perimitral reentry (54%, 3 of which were figure-of-eight with a roof-dependent loop), 27 were roof-dependent re-entry (24%), and 23 were CTI-dependent re-entry (21%). One additional atypical non-roof-dependent macro–re-entry around the LA appendage was diagnosed.
Of the 85 centrifugal ATs, 68 were localized re-entry (80%) and 17 were focal AT (20%). The origin of centrifugal AT was located in the left atrium and CS for the vast majority of cases (n = 83 [98%]), with only 2 centrifugal ATs originating from the right atrium (from the lateral wall and the superior vena cava).
General ECG findings
Low-amplitude P waves, defined as a maximal voltage ≤0.1 mV, in any of the 12 leads, were found in 57 ATs (29%). As a whole, including patients with very low amplitude P waves (i.e., maximal voltage <0.05 mV), 33% of the total study ATs displayed low-amplitude P waves, as illustrated in Figure 1. ATs displaying low-amplitude P waves had longer cycle lengths (295 ± 78 ms vs. 264 ± 59 ms, p = 0.015).
The absence of any surface ECG activity for >80 ms in all 12 leads was observed in 34% of cases. The proportion of patients displaying this feature was inversely related to the voltage of the surface P-wave: an isoelectric line >80 ms was observed in about half of the patients with maximal P-wave voltage ≤0.1 mV (49%), as opposed to 30% of the patients with P waves exceeding 0.1 mV (p = 0.02). This finding was more often observed in centrifugal ATs, though as many as one-quarter of macro–re-entrant ATs also displayed an isoelectric line >80 ms (47% vs. 24%, p < 0.001) (Figure 2). However, stability testing of significance using bootstrapping found that this feature failed to maintain significance to discriminate centrifugal AT from macro–reentry (97.5th quantile of p values = 0.16). The prevalence of isoelectric intervals was similar in both localized re-entry and focal AT (48% vs. 47%, respectively). In contrast, the presence of a continuously undulating P-wave, with an isoelectric baseline either absent or lower than 30 ms, was more specific of a macro–re-entrant AT, as this diagnosis was confirmed in 78% of cases. Centrifugal AT was diagnosed in the remaining 22% of cases (p < 0.001, 97.5th quantile of p values = 0.02). This feature provided 78% positive predictive value to distinguish macro–re-entry as opposed to centrifugal AT (sensitivity, 43%; specificity, 84%; negative predictive value, 52%).
More than 1 AT was observed during 41 procedures for a total of 50 changes in the mechanism or site of origin of the AT. During these AT transitions, no discernible changes in the surface P-wave configuration could be observed in 15 cases (30%), as illustrated in Figure 3.
The surface P-wave polarity in the inferior leads, II, III, and aVF, was predominantly positive in the majority of cases (81% [n = 158]). This positive polarity could not help localize the origins of ATs, as the majority of the centrifugal ATs originating from the CS or the bottom of the left atrium displayed positive P waves in the inferior leads (83% [n = 18]). Predominantly negative P waves in the inferior leads were relatively seldom observed (15% [n = 29]). In such cases, there was a higher probability of being confronted with counterclockwise (CCW) CTI-dependent AT (34% vs. 7%, p < 0.001). CTI-dependent AT and perimitral AT each represented one-third (34%) of the AT diagnoses with predominantly negative P waves. As opposed to positive P waves, negative polarity provided some clue about the origin of centrifugal ATs, as most of them originated from the inferior left atrium or CS (80% vs. 19% for non-negative P waves, p = 0.001). A flat P wave in the inferior leads was observed in 4% (n = 8).
The surface P-wave configurations recorded in leads V1, I, and aVL, with respect to the AT mechanisms, are summarized in Table 1. In lead I, the P-wave polarity was most often flat (63%), whereas a predominantly negative P wave was most often observed in lead aVL (63%). The P-wave polarity in both leads I and aVL could not help differentiate between the different AT mechanism or site of origin.
In lead V1, the P-wave configuration was most often upright (“all R” pattern, 59%). A biphasic P wave, with the second negative component being equal or smaller than the first positive component (“RS” or “Rs” pattern), was observed in 32% of cases. A biphasic P wave with a predominantly negative component (“rS” pattern) was more rarely observed (7%) of patients. The P-wave pattern in lead V1 was of limited use in orienting the AT diagnosis, except for the fact that an “all R” pattern was less often observed in perimitral AT (24% vs. 40%, p = 0.018). Moreover, though an inverted P-wave in lead V1 (“all S” pattern) was observed in only 2 cases (1%), both were related to right atrial centrifugal AT (100% vs. 0%, p < 0.001).
A minority of CCW peritricuspid ATs displayed the classic saw-tooth pattern in the inferior leads (27% [n = 22]), and an upright flutter wave was observed in 55% of cases. In contrast, the configuration of the P-wave in the precordial leads was more reproducible and specific to peritricuspid ATs. With progression across the precordium, we often observed a gradual transition from an upright component in lead V1 to a negative component that became progressively apparent, usually by lead V3. The positive component recorded in lead V1 usually occurred after the negative component recorded in the left precordial leads. Rarely, both components occurred simultaneously, but the left precordial leads’ negative component never followed the positive component in lead V1. This “precordial transition” was present in 59% of CCW peritricuspid ATs compared with 2% in other ATs (p < 0.001, 97.5th quantile of p values < 0.001) (Figure 4). This pattern identified CCW peritricuspid ATs with 81% positive predictive value (sensitivity, 59%; specificity, 98%; and negative predictive value, 95%).
In the peripheral leads, no P-wave pattern could help distinguish perimitral ATs (n = 60) from other AT mechanisms. Even among perimitral AT cases, no significant differences in the polarity of the P-wave were found whether the activation wave front proceeded clockwise (CW) or CCW. In particular, a predominantly negative P wave in the inferior leads was found in a similar proportion in both CW and CCW perimitral AT (18% vs. 14%, respectively, p = 0.717) (see also Table 1).
We identified only 2 features that could help distinguish perimitral AT from other AT mechanisms. First, in the absence of a “precordial transition,” the presence of either a negative or a negative-positive P-wave in any of leads V2 to V6 was observed in 30% of perimitral ATs as opposed to 3% in the remaining cases (p < 0.001, 97.5th quantile of p values < 0.001) (Figures 5 and 6A⇓⇓). This feature identified perimitral ATs with 82% positive predictive value (sensitivity, 30%; specificity, 97%; negative predictive value, 76%). The second useful feature was the presence of a negative “notched” component at the beginning of a positive P-wave in the inferior leads. This unique feature specifically identified perimitral ATs with CW activation. It was observed in 26% of cases as opposed to only 2% in the remaining cases (p < 0.001, 97.5th quantile of p values = 0.004) (Figure 6). A negative notch in the inferior leads provided 77% positive predictive value to distinguish CW perimitral reentries from other ATs (sensitivity, 25%; specificity, 98%; negative predictive value, 84%).
It is noteworthy that different ECG overlaps were observed between CW perimitral and CCW peritricuspid AT. Most important, the typical inferior lead saw-tooth pattern was also observed in 10% of CW perimitral ATs (n = 4) (Figures 5A and 7). This pattern was highly specific to 1 of the 2 diagnoses, as it was never observed with other AT mechanisms (specificity, 100%; sensitivity, 16%; positive predictive value, 100%; negative predictive value, 73%).
We found no P-wave pattern in both peripheral and precordial leads that could help distinguish roof-dependent re-entries from other AT mechanisms.
The few specific ECG features that may help identify the mechanism and localization of ATs occurring after PsAF ablation are summarized in Figure 8.
The present study evaluated the surface ECG characteristics of nearly 200 ATs occurring after PsAF ablation. Our study shows that the interpretation of the surface electrocardiogram post–LA ablation bears major limitations that hinder its use. We could identify a few unique ECG characteristics that could still be used to help identify the mechanism of 2 of the most frequent macro–re-entries after PsAF ablation, namely, perimitral and peritricuspid AT. Knowledge of these features may allow more appropriate procedural planning and guide the initial mapping strategy.
The correlation between the surface P-wave configuration and the underlying AT mechanism is much limited by the fact that the normal activation pattern of the atria is altered by the effect of ablation and that some areas of the atria may become electrically silent. The impact of the latter is reflected by the fact that one-third of ATs displayed low-amplitude P waves, defined as a maximal voltage ≤0.1 mV in any of the 12 leads. There certainly is a direct relationship between the extent of atrial scarring and the amount of ablation performed. Takahashi et al. (9) have shown that after a stepwise ablation strategy similar to the one applied in our study, electroanatomical mapping showed areas of scar and low voltage (defined as <0.5 mV) accounting for about 60% of the total LA surface area. Nevertheless, it should not be assumed that a more conservative ablation strategy would necessarily prevent bias in the interpretation of electrocardiograms. The presence of spontaneous scar in patients with PsAF is well known, and several mechanisms have been implicated (10). Chronic AF itself may result in structural remodeling, and a direct link has been demonstrated between the degree of fibrosis and the disease severity. The proportion of patients with PsAF who have pre-existing atrial scarring thus exceeds one-third, with involvement that can sometimes exceed 50% of the LA surface (11–13).
Atrial scarring, whether spontaneous or iatrogenic, may also have influenced the prevalence of long isoelectric intervals on 12-lead electrocardiography in our study population. The presence of discrete P waves separated by an isoelectric interval in all 12 leads has traditionally been regarded as a clue indicating focal AT (“truly focal”), that is, non–re-entrant AT with centrifugal activation from a discrete point source. In such cases, the isoelectric line usually reflects the absence of electric activity because of the diastolic pause at the site of origin. Alternatively, the isoelectric interval may reflect a time interval when very low amplitude electric activity occurs. Moreover, the surface ECG recognition of electric activity depends on the cellular mass activated at each instant. An “isoelectric” line may therefore result from slow conduction and/or the presence of scar tissue. In the case of localized reentry, the “isthmus” of the circuit is usually composed of strands of slow-conducting viable myocardium within scarred areas. They are recognized by low-amplitude, long-duration EGMs, usually covering about 50% or more of the tachycardia cycle length (7,8,14–16). Because of both the very slow conduction and the narrow size of these myocardial strands, the isthmus activation results in a drop of voltage below detection on the surface electrocardiogram. That part of the cycle therefore coincides with an isoelectric interval on the surface in the vast majority of localized re-entry occurring after AF ablation (8,15,16). Consequently, this feature has been proposed as a relatively specific clue to guide the initial mapping, as discrete P waves separated by an isoelectric interval are rarely observed in macro–reentrant ATs (8,16). The predictive value of this pattern to distinguish the mechanism of AT seems far more limited after PsAF ablation, as only about half of our study population with localized re-entry displayed an isoelectric interval >80 ms. Moreover, as much as about one-quarter of macro–re-entrant ATs also displayed this feature. In contrast, the presence of a continuously undulating P-wave pattern was more useful to identify a macro–re-entrant mechanism of AT and guide the initial mapping.
The lower than expected prevalence of isoelectric intervals in localized re-entries and focal ATs may be explained by the fact that scarring and/or the creation of lines of block may delay the activation of some bystander areas with relatively preserved voltage. This late activation will lead to the shortening, or even concealing, of the isoelectric line. In contrast, as previously discussed, the undetectable slow activation of large bystander scarred areas probably explains the unexpected high prevalence of isoelectric intervals observed in macro–re-entrant AT. The slow activation of isthmuses related to small conduction gaps through previous lines of block may also have contributed (8).
Another important limitation in the interpretation of the surface electrocardiogram post–LA ablation pertains to the fact that the areas of the atria that are of low amplitude, or electrically silent, become overshadowed by areas of relatively preserved voltage. This is illustrated by the absence of any discernible changes in the surface P-wave configuration in nearly one-third of the transitions that may occur from one AT to the other, for instance after successful ablation. This likely reflects the limited extent of areas able to generate manifest surface ECG electric activity (supposedly mainly the appendages and the right atrium after LA ablation). Accordingly, the downstream passive activation of those areas, and hence the P-wave configuration, is more likely to be marginally affected by changes of activation sequences occurring upstream in low-voltage or electrically silent LA areas.
Our study demonstrates that the polarity of the surface P-wave in the peripheral leads provides almost no help to differentiate between different AT mechanism or site of origin. As discussed, this likely reflects the fact that the surface P wave is influenced mostly by the activation of bystander areas with more preserved voltage that are “distant” from the AT circuit or focus. For instance, negative P waves in lead I are typically associated with anterolateral LA centrifugal ATs (LA appendage, mitral annulus) or CCW perimitral ATs (4,17,18). In our study population, the finding of a negative P wave in lead I was unspecific and was observed in ATs originating from different segments of the left atrium, in right atrial macro–re-entries and in perimitral ATs, all of which displayed CW activation instead. Similarly, a positive P wave in the inferior leads did not provide any clue to identify the site of origin or the AT mechanism. The only P-wave polarity pattern in the peripheral leads that provided some diagnostic orientation was the finding of a predominantly negative P-wave in the inferior leads. In such cases, the probability of CCW peritricuspid or perimitral AT was one-third for both. Moreover, in case of centrifugal AT, the negative polarity could be used to orient the initial mapping as most ATs originated from the inferior part of the left atrium.
A negative P-wave in lead V1 has been shown to be highly specific for a right atrial origin of de novo centrifugal ATs (17). Moreover, CW peritricuspid AT is also typically characterized by a negative P-wave in lead V1. Although seldom observed in our population, this feature seems to provide the same specificity after PsAF ablation and should therefore prompt right atrial mapping whenever observed. Moreover, de novo CW peritricuspid ATs are also typically characterized by a negative P-wave in lead V1 (19). This feature may also apply to post-ablation ATs (20) but could not be confirmed in our study, as only 1 CW peritricuspid AT was observed.
Our study shows that peritricuspid ATs most often display atypical ECG findings after PsAF ablation. Specifically, the classic “saw-tooth” pattern in the inferior leads was observed in fewer than one-third of CCW peritricuspid ATs, with an upright flutter wave instead observed in more than half of the cases. This is in line with previous reports on the characteristics of CTI-dependent flutter after LA ablation (20,21). Chugh et al. (20) showed that the mechanism underlying these findings is the reduction of the LA voltage by ablation, which attenuates the negative component related to the septal and LA activation. This activation becomes overshadowed by the right atrial activation, and the unopposed craniocaudal activation of the right atrial wall results in positive inferior waves.
Interestingly, different ECG overlaps were observed between CW perimitral and CCW peritricuspid AT. Notably, though the classic “saw-tooth” pattern was rarely observed, we found that CW perimitral AT could mimic that ECG pattern in 10% of cases, probably because of similar septal activation. Cases of double-loop reentries may theoretically also have contributed to that finding.
Although atypical patterns are almost the rule in the peripheral leads, our study shows that the P wave in precordial leads often displayed a gradual transition from an upright component in lead V1 to a negative component with progression across the precordium (usually by lead V3). This “precordial transition” was observed in about 60% of cases and can be used to specifically identify CCW peritricuspid AT despite previous LA ablation. To the best of our knowledge, this “precordial transition” has not been specifically evaluated after LA ablation. Chugh et al. (20) described a negative component of the P-wave in the left precordial lead in more than half of CCW peritricuspid ATs, and Chang et al. (22) found a higher incidence of negative flutter waves in at least 1 precordial lead in right atrial macro–re-entrant ATs (mainly peritricuspid).
Our study shows that there is no distinctive flutter wave polarity in the peripheral ECG leads to help identify perimitral AT. It is noteworthy that we did not observe significant differences in this respect between CW and CCW perimitral ATs. These findings are in sharp contrast with those of a study by Gerstenfeld and Marchlinski (4), in which a distinctive pattern was found out of 15 perimitral ATs occurring after PV isolation. CCW flutter waves were positive in the inferior and had a significant negative component in leads I and aVL. The converse limb configuration was observed in CW flutter waves. Other investigators have reported data on a limited number of de novo ATs (23,24). Similar to our results, the surface ECG manifestations of perimitral ATs were found to be highly variable. These findings reflect not only the altered atrial activation patterns related to scarring but also the fact that perimitral circuits frequently involve other simultaneous loops.
Despite variable ECG manifestations, we could identify 2 specific features that could help distinguish perimitral AT from other AT mechanisms. First, a negative “notched” component at the beginning of a positive P wave in the inferior leads was specifically associated with CW perimitral AT. Considering the overlap observed with CCW peritricuspid AT (see earlier discussion), it may be possible that this negative component results from a similar pattern of septal activation.
Second, we found that in the absence of a “precordial transition,” the presence of a negative, or negative-positive, P-wave in any of leads V2 to V6 helped identify perimitral ATs. Similar to our study, Gerstenfeld and Marchlinski (4) also reported the occurrence of biphasic P waves with an initial negative component in the precordial leads. In their study, the negative component P-wave was observed in lead V2 for CCW and in the lateral precordial leads for CW mitral flutters. A specific distribution of leads displaying that negative component was not reproduced in our study.
The stepwise ablation approach performed in our study population included EGM-guided ablation and LA linear lesions. Both the prevalence of the different AT mechanisms and their ECG features are closely related to the ablation strategy adopted. Accordingly, our findings may not apply in a patient population with PsAF in which a more conservative strategy is undertaken.
The interpretation of the surface electrocardiogram post–LA ablation bears several major limitations. Only a few unique ECG characteristics may help identify the mechanism of the most frequent macro–reentries. Knowledge of these features may allow more appropriate procedural planning and guide the initial mapping strategy.
COMPETENCY IN MEDICAL KNOWLEDGE: Catheter ablation of PsAF often requires adjunctive substrate modification strategies beyond PV isolation. AT occurring in this setting often poses a diagnostic challenge, as both the spontaneous remodeling and the effect of LA ablation alter the normal activation pattern of the atria. The interpretation of the surface electrocardiogram is therefore much limited, and known ECG features most often do not apply. Only a few unique ECG characteristics may be used to help identify the mechanism of 2 of the most frequent macro–reentries, namely, perimitral and peritricuspid AT. Knowledge of these features may allow more appropriate procedural planning and guide the initial mapping strategy.
TRANSLATIONAL OUTLOOK: Analysis of the activation pattern of both the left and right atria may help understand the potential yield and limitations of surface electrocardiography in AT after PsAF ablation.
The authors are grateful to Adrian Luca for statistical assistance.
Dr. Pascale receives financial support from the Swiss National Science Foundation and the SICPA Foundation. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- atrial tachycardia
- coronary sinus
- cavotricuspid isthmus
- left atrial
- persistent atrial fibrillation
- pulmonary vein
- Received March 11, 2017.
- Revision received June 27, 2017.
- Accepted July 13, 2017.
- 2017 American College of Cardiology Foundation
- Nademanee K.,
- McKenzie J.,
- Kosar E.,
- et al.
- Rostock T.,
- Drewitz I.,
- Steven D.,
- et al.
- Takahashi Y.,
- O’Neill M.D.,
- Hocini M.,
- et al.
- Rolf S.,
- Kircher S.,
- Arya A.,
- et al.
- Oakes R.S.,
- Badger T.J.,
- Kholmovski E.G.,
- et al.
- Takahashi Y.,
- Takahashi A.,
- Miyazaki S.,
- et al.
- Kistler P.M.,
- Roberts-Thomson K.C.,
- Haqqani H.M.,
- et al.
- Chugh A.,
- Latchamsetty R.,
- Oral H.,
- et al.
- Bochoeyer A.,
- Yang Y.,
- Cheng J.,
- et al.
- Jais P.,
- Shah D.C.,
- Haissaguerre M.,
- et al.