Author + information
- Received February 5, 2018
- Revision received March 6, 2018
- Accepted March 29, 2018
- Published online August 20, 2018.
- Thomas Pambrun, MDa,b,∗ (, )
- Rim El Bouazzaoui, MDa,
- Nicolas Combes, MDa,
- Stéphane Combes, MDa,
- Pedro Sousa, MDc,
- Mathieu Le Bloa, MDa,
- Grégoire Massoullié, MDb,
- Ghassen Cheniti, MDb,
- Ruairidh Martin, MDb,
- Xavier Pillois, PhDb,
- Josselin Duchateau, MDb,
- Frédéric Sacher, MD, PhDb,
- Mélèze Hocini, MDb,
- Pierre Jaïs, MDb,
- Nicolas Derval, MDb,
- Agustín Bortone, MDd,
- Serge Boveda, MDa,
- Arnaud Denis, MDb,
- Michel Haïssaguerre, MDb and
- Jean-Paul Albenque, MDa
- aDépartement de Rythmologie, Clinique Pasteur, Toulouse, France
- bHôpital Cardiologique du Haut-Lévêque, CHU Bordeaux, L’Institut de RYthmologie et modélisation Cardiaque (LIRYC), Université Bordeaux, Bordeaux, France
- cCardiology Department, University Hospital of Coimbra, Coimbra, Portugal
- dService de Cardiologie, Hôpital Privé Les Franciscaines, Nîmes, France
- ↵∗Address for correspondence:
Dr. Thomas Pambrun, Hôpital Cardiologique du Haut-Lévêque, CHU Bordeaux, Avenue de Magellan, 33604 Pessac Cedex, France.
Objectives This study evaluated a new algorithm relying on maximal pre-excitation.
Background Prior knowledge of accessory pathway (AP) location facilitates an individual ablation strategy. Delta-wave analysis on a 12-lead electrocardiogram is recognized as crucial for predicting ablation site, but can be ambiguous at basal state.
Methods An algorithm based on maximal pre-excitation, as induced by atrial pacing during an electrophysiological study, was initially developed in 132 patients with a single manifest AP. The maximally pre-excited QRS features included the global polarity in lead V1 (step 1), inferior leads (step 2), and leads V3 or I (step 3), as well as the morphology in lead II (step 4). Three investigators prospectively tested the new algorithm in 207 consecutive patients by comparing its efficacy to a control algorithm relying on basal pre-excitation.
Results The accuracy, defined as the percent of patients with an exact prediction of AP location, was significantly greater with the new algorithm (90% vs. 63%; p < 0.001). The reproducibility, defined as the level of agreement between investigators in determining AP location, was excellent (κ > 0.75; p < 0.05) with the new algorithm and fair (0.40 < κ < 0.75; p < 0.05) with the control algorithm.
Conclusions An algorithm based on maximal pre-excitation allows accurate and reproducible localization of manifest APs. When ablation is indicated, the analysis of maximal pre-excitation is a sensible approach for giving a head start in endocardial mapping.
Catheter ablation of the accessory pathway (AP) can be indicated in patients with pre-excitation (1). Pre-ablation AP localization inferred from the 12-lead surface electrocardiogram (ECG) facilitates a tailored ablation strategy, whose specific risk is properly assessed (2–11). As they are based on pre-excitation in the basal state, available algorithms can lead to complex and ambiguous analysis, mostly due to the variable influence of the AP over QRS morphology (12–15). The electrophysiological study can maximize this influence by maneuvers such as rapid atrial pacing.
To date, studies considering maximally pre-excited ECG have not proposed a stepwise algorithm (6–11). Therefore, the present study aimed to develop an algorithm based on maximal pre-excitation and compare its efficacy to a widely used algorithm relying on basal pre-excitation.
The study cohort consisted of 207 consecutive patients over 15 years of age, referred to our institutions for assessment of a manifest AP. Patients were included when they fulfilled the following criteria: 1) indication for first ablation of a manifest AP (1); 2) absence of structural heart disease; 3) absence of multiple APs, as assessed by electrophysiological study; and 4) absence of Mahaïm fibers.
The local institutional ethics committees on human research approved the study protocol. Written and informed consent was obtained from all patients.
Pre-excitation maximization and ablation procedure
Maximal pre-excitation was systematically provoked during an electrophysiological study, even when previously documented during spontaneous pre-excited AF. All procedures were performed in sinus rhythm after withdrawal of all antiarrhythmic drugs, with 2 diagnostic catheters percutaneously inserted through the right femoral vein. Atrial pacing aimed at determining the AP anterograde refractory period to risk stratify for life-threatening arrhythmic events and to induce AP-related tachycardia. This maneuver provided the maximal pre-excitation upon which the AP location was predicted by the new algorithm.
When ablation was indicated, 1 diagnostic catheter was moved into the coronary sinus (CS), serving as an anatomical landmark of the CS ostium and the mitral annulus, while the other diagnostic catheter was placed at the His bundle. The ablation catheter position was ascertained in the left anterior oblique view, so that the tip of the His-bundle catheter could identify both the annuli frontal plane by pointing toward the operator and the CS ostium located directly below, at the same sagittal plane. During ablation, the ventricular insertion site was systematically targeted, as previously described (16). Vascular access (femoral vein or artery), additional material (sheath for catheter stability), and energy source (standard radiofrequency, irrigated radiofrequency, or cryothermy) varied with AP location (Figure 1).
The terminology of the 9 anatomical locations of the AP was defined as follows: 1) right anterior (RA), from 10 to 1 o’clock; 2) right lateral (RL), from 8 to 10 o’clock; 3) right posterior (RP), below the CS ostium and extending to 8 o’clock; 4) right paraseptal (RPS) (formerly known as posteroseptal), near the CS ostium, including its terminal portion; 5) nodo-Hisian (NH), above the CS ostium and extending to 1 o’clock, therefore including the atrioventricular node (midseptal area) and His bundle (anteroseptal area); 6) deep CS (DCS), more than 1 cm within the CS, including the middle cardiac vein; 7) left paraseptal (LPS) (formerly known as left posteroseptal), from 6 to 8 o’clock; 8) left posterolateral (LPL), from 4 to 6 o’clock; and 9) left lateral (LL), from 0 to 4 o’clock (Figure 1).
Maximal pre-excitation, obtained by rapid atrial pacing for assessment of AP anterograde effective refractory period, was recorded on the surface ECG at a speed of 25 mm/s and band-pass filter of 0.1 to 50 Hz. The QRS polarity was defined as positive or negative, depending on whether the QRS morphology was mainly above or under the baseline. Of importance, in the particular case of isoelectric QRS complex (QR or RS pattern), the polarity of the initial deflection (negative for QR pattern and positive for RS pattern) determined the classification. Hence, 9 morphologies were observed and categorized as: 1) negative for the QS-smooth, QS-notched, QrS, rS, or QR pattern; and 2) positive for the RS, rsr’, Rs, or R pattern (Figure 2A). The QRS ratio between 2 leads compared the amplitude of their positive component, the greatest being defined as the one displaying the highest amplitude from the onset to the peak of the R-wave (Figure 2B).
The new algorithm was empirically developed in 132 initial patients by correlating QRS polarity during maximal pre-excitation and the exact anatomical location of the AP, defined as the site where energy application led to abolition of the pre-excitation.
The new algorithm was then prospectively tested in 207 consecutive patients. The exact ablation site of the AP being determined by the operator, 3 investigators used the new algorithm to decipher the AP location on the maximally pre-excited ECG. All 3 were blinded to the ablation procedures and to each other’s conclusions.
The results of the new algorithm, based on maximal pre-excitation, were compared with those obtained with an algorithm relying on basal pre-excitation (5). The choice of Arruda’s algorithm was supported by a similar classification of the AP anatomical location, hence simplifying the comparison between both algorithms (Table 1).
Categorical data, expressed as percentage, were compared using the chi-square test. The accuracy of the algorithm was defined as the percent of patients with an exact prediction of the successful ablation site. The reproducibility of the algorithm was defined as the level of agreement between investigators in determining AP location. Kappa values below 0.40, from 0.40 to 0.75, and over 0.75 were considered to indicate poor, fair, and excellent agreement, respectively, and p values <0.05 were considered statistically significant.
Step 1 (lead V1 polarity)
Indicates a right-sided AP (including NH AP) when negative and a left-sided AP when positive.
Step 2 (inferior leads polarity)
When none are positive, it indicates RP or RPS locations (posterior pair) for right-sided APs and DCS, LPS, or LPL locations for left-sided APs. When 1 or 2 are positive, it indicates RL or NH locations (intermediate pair) for right-sided APs, and directly identifies LPL location for left-sided APs. When all 3 are positive, it indicates RA or NH locations (anterior pair) for right-sided APs, and directly identifies LL location for left-sided APs.
Step 3 (lead V3 polarity)
Only examined for the final identification of right-sided APs. A negative and a positive polarity distinguish RP from RPS location, respectively, for a posterior pair; RL from NH location for an intermediate pair; and RA from NH location for an anterior pair.
Step 3 (V1/I ratio)
Only examined in case of posterior left-sided APs. A ratio <1 indicates DCS or LPS location. A ratio ≥1 directly identifies LPL location.
Step 4 (lead II morphology)
Only examined to discriminate DCS from LPS APs. A notched QS indicates DCS location.
The distribution of successful AP ablation sites is shown in Figure 1. The averaged sensitivity, specificity, and predictive values of the 2 algorithms for each location are summarized in Table 2. Regarding the accuracy, AP locations were correctly identified by the 3 investigators in 90% (558 of 621) of patients with the new algorithm and 63% (389 of 621) of patients with Arruda’s algorithm (p < 0.001). Accuracy proved consistently higher with the new algorithm, regardless of which investigator it was (for investigator 1: 92.3% vs. 67.1%, p < 0.001; for investigator 2: 88.9% vs. 63.3%, p < 0.001; for investigator 3: 88.4% vs. 57.5%, p < 0.001). Regarding the reproducibility, agreement between investigators was excellent (κ > 0.75) with the new algorithm and fair (0.40 < κ < 0.75) with Arruda’s algorithm (Table 3). A representative maximally pre-excited 12-lead ECG of the 9 predefined AP locations is illustrated in Figure 5. Detailed information about the prediction of each investigator for the 2 algorithms is provided in the supplementary data (Online Tables 4 to 9).
The present study reports a stepwise algorithm exclusively based on induced maximal pre-excitation. This approach demonstrates accurate and reproducible localization of manifest APs, while being easily performed during a standardized electrophysiological study.
Pre-excitation maximization: An accurate and reproducible approach
Studies comparing available algorithms systematically show variable efficacy, which may be explained by ambiguous interpretation of the Delta wave in the basal state (12–15). Disagreement between algorithms relying on basal pre-excitation are mostly found when QRS complex duration drops below 120 ms, due to competing activation from the His-Purkinje system, resulting in a restriction of the AP influence over QRS morphology (15). Therefore, it has been proposed to circumvent this pitfall by focusing on the initial 20 to 40 ms of the Delta wave. Nevertheless, whether pre-excitation is minimal or maximal, the final part of the QRS complex results either from fast ubiquitous activation through the His-Purkinje system or slow focal activation originating from the annulus. As a result, beyond the Delta wave, which reflects ventricular depolarization at the vicinity of the AP, the rest of the maximally pre-excited QRS complex also contains additional information related to AP location, while the minimally pre-excited QRS complex does not (Figure 6A). Furthermore, confining the analysis to the first 20 ms of the Delta wave raises 2 practical concerns. First, the Delta-wave onset at 1 specific lead, measured from the earliest onset simultaneously identified in any other leads, can be equivocal, depending on the lead subjectively selected as the reference. Second, this indirect method requires classification of Delta-wave polarity as positive, negative, or isoelectric, which, given the short window of analysis, can be difficult to differentiate when the Delta-wave slope is not steep (Figure 6B). Hence, the analysis of the entire QRS polarity after pre-excitation maximization by rapid atrial pacing gives more detailed information and less room for interpretation, as shown by the more accurate and reproducible AP’s localization found in the present study.
Maximal pre-excitation: Neglected data
Recommendations for catheter ablation of a manifest AP usually refer to situations where maximal pre-excitation is collected de facto, either spontaneously (documentation of a pre-excited AF) or during an electrophysiological study (1). In the latter case, maximal pre-excitation is obtained during assessment of AP anterograde refractory period, either for induction of AP-related tachycardia in symptomatic patients or for risk stratification in asymptomatic patients. Interestingly, the collection of this previously neglected data does not preclude any ablation strategies, as no decision has been taken at this stage regarding the choice of vascular access, additional material, or energy source. It could be argued that, once catheters are in place, a careful analysis of intracardiac electrograms is mandatory for ablation. But in daily practice, rather than randomly starting endocardial mapping at an undetermined area of the annulus, a previous analysis of pre-excitation provides an efficient head start by giving a clear idea of the area to primarily target. From this respect, although basal pre-excitation remains useful during outpatient consultation to inform about the specific risks related to AP location, maximal pre-excitation proves more effective in the EP laboratory to guide endocardial mapping. Of note, maximal pre-excitation can be noninvasively obtained with adenosine infusion. But in contrast to an electrophysiological study, adenosine is contraindicated in patients with severe or poorly controlled airway disease and does not provide any indication for AP ablation.
Relevance of distinct maximally pre-excited QRS features for AP localization
Previous studies have analyzed the relationship between maximal pre-excitation pattern and AP location. However, the vast majority of studies were restricted to septal APs (6–10). To our knowledge, only 1 study considered the entire annulus perimeter, but did not propose an algorithm (11). By providing a global stepwise algorithm, the current study may help to rule out potential similarities of the QRS pattern between adjacent locations. Subsequently, we consider the distinct QRS features identified, with reference to previously published studies, where possible.
Lead V1 polarity, either negative or positive, appears essential to distinguish right from left APs, respectively. Accordingly, Fananapazir et al. (11) reported that >75% of APs display a negative lead V1 polarity when right sided and a positive lead V1 polarity when left sided. Although this seminal work considered the paraseptal region as a single entity, further studies described a significant shift from negative polarity, when APs are located in the CS ostium vicinity (RPS), to positive polarity, when APs are located in the middle cardiac vein (DCS) or at the left endocardial surface (LPS) (6,7). Of note, these findings are consistent with our current results. However, due to the muscular thickness and anatomical complexity of the inferior pyramidal space, paraseptal APs, either with right or left ventricular exit site, can sometimes require an ablation from both sides and therefore decrease the algorithm’s performance.
Inferior leads polarity contributes to APs localization, most notably proving highly predictive for RA and LL APs when all 3 are positive. Fitzpatrick et al. (3) first proposed inferior leads analysis, based on a sum of Delta-wave polarities at basal state (ascertained during the first 40 ms and defined as negative, isoelectric, or positive), but confined it to septal APs. Fananapazir et al. (11) extended this approach to all AP locations during maximal pre-excitation and similarly found that all 3 inferior leads were positive in most RA and LL APs.
Lead V3 polarity, either negative or positive, seems useful for identification of right ventricular free wall and septal APs, respectively. As the right ventricular free wall is located forward of the septum, its subsequent rearward depolarization may account for negativity in lead V3. Consistent with this principle, studies describing the morphology of maximally pre-excited QRS complex in septal APs report that a vast majority of midseptal and parahisian APs display positivity in lead V3 (8,9). Of note, as no data on lead V3 polarity are available for RPS APs during maximal pre-excitation (6,7), the current study is the first to point out its relevance in this location.
The V1/I ratio ensures a reliable classification in case of posterior location of left-sided APs. Although defining 2 different planes, the right-facing lead V1 is positioned opposite to the left-facing lead I. Therefore, the 2 leads behave inversely when the origin of ventricular depolarization shifts from the septal to the lateral part of the mitral annulus, leading to an amplitude increase for lead V1 and a decrease for lead I. In our experience, equalization of the 2 leads is found at 6 o’clock.
Lead II notch is more frequent in DCS than in LPS APs. The former are associated with muscular strands arising from the CS myocardial coat and epicardially connecting the left atrium with the left ventricle, further away from the mitral annulus (17). These sleeve-like extensions, interspersed with adipose and fibrous tissue, can display various orientations (18). By favoring local conduction abnormalities, these conditions may lead to inhomogeneous ventricular activation, which is recognized as the underlying mechanism of QRS notching (19). Of note, despite maximal pre-excitation analysis, localization of epicardial APs still displays the lowest performance in terms of sensitivity and positive predictive value.
First, as previously observed, DCS APs were predominantly sited in the middle cardiac vein rather than the diverticulum (1 case in the current study) (17). Therefore, no firm conclusion can be drawn regarding the maximally pre-excited QRS pattern of the latter location. Second, because it is not designed or tested for multiple APs, its applicability to this rare condition remains unknown. Finally, as the oblique course of some AP can lead to substantial discrepancies between atrial and ventricular insertion site (20), the latter only, on which maximal pre-excitation pattern exclusively depends, is targeted by the new algorithm.
The present study demonstrates that the analysis of maximal pre-excitation allows accurate and reproducible localization of manifest APs for a vast majority of patients. Since de facto collected when ablation is indicated, maximal pre-excitation can effectively guide endocardial mapping in daily practice.
COMPETENCY IN MEDICAL KNOWLEDGE: Maximal pre-excitation allows more accurate and reproducible localization of the AP than basal pre-excitation does. There are 2 main reasons for this. Since exclusively reflecting ventricular depolarization originating from the AP, maximally pre-excited QRS obtained during rapid atrial pacing contains more information relative to its location. ECG analysis gives less room for interpretation when applied to the entire QRS rather than confined to the delta-wave’s onset. Since AP anterograde refractory period is obtained for induction of AP-related tachycardia and risk stratification for life-threatening arrhythmias, maximal pre-excitation is systematically available when ablation is indicated. Therefore, the use of a localization algorithm based on maximal pre-excitation appears as a rational approach to start endocardial mapping at the adequate area.
TRANSLATIONAL OUTLOOK: High-density mapping during maximal pre-excitation may offer further insight into the variation of ventricular activation underlying the modification of QRS morphology observed from one AP location to another.
This study received financial support from the French Government as part of the “Investments of the Future” program managed by the National Research Agency (ANR), Grant reference ANR-10-IAHU-04. Dr. Albenque has served as a consultant for Abbott, Biosense Webster, Inc., and ACT. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- accessory pathway
- coronary sinus
- deep coronary sinus
- left lateral
- left posterolateral
- left paraseptal
- right anterior
- right lateral
- right posterior
- right paraseptal
- Received February 5, 2018.
- Revision received March 6, 2018.
- Accepted March 29, 2018.
- 2018 American College of Cardiology Foundation
- Page R.L.,
- Joglar J.A.,
- Caldwell M.A.,
- et al.
- Fitzpatrick A.P.,
- Gonzales R.P.,
- Lesh M.D.,
- Modin G.W.,
- Lee R.J.,
- Scheinman M.M.
- Epstein A.E.,
- Kirklin J.K.,
- Holman W.L.,
- Plumb V.J.,
- Kay G.N.
- Haïssaguerre M.,
- Marcus F.,
- Poquet F.,
- Gencel L.,
- Le Métayer P.,
- Clémenty J.
- Fananapazir L.,
- German L.D.,
- Gallagher J.J.,
- Lowe J.E.,
- Prystowsky E.N.
- Maden O.,
- Balci K.G.,
- Selcuk M.T.,
- et al.
- Haïssaguerre M.,
- Dartigues J.F.,
- Warin J.F.,
- Le Metayer P.,
- Salamon R.
- Sun Y.,
- Arruda M.,
- Otomo K.,
- et al.
- Flowers N.C.,
- Horan L.G.,
- Thomas J.R.,
- Tolleson W.J.
- Otomo K.,
- Gonzalez M.D.,
- Beckman K.J.,
- et al.