Advertisement

Electrogram signature of specific activation patterns: Analysis of atrial tachycardias at high-density endocardial mapping

Published:August 07, 2017DOI:https://doi.org/10.1016/j.hrthm.2017.08.001

      Background

      The significance of fractionated electrograms (EGMs) is object of debate, with multiple mechanisms described.

      Objective

      Using Rhythmia, a high-density mapping system, we sought to investigate the relationship between specific electrophysiological phenomena and EGM characteristics at those sites.

      Methods

      Twenty-five consecutive patients underwent high-density atrial mapping during atrial tachycardias. Bipolar EGMs were recorded with a 64-electrode basket catheter. The following atrial phenomena were identified: slow conduction (SC) areas, lines of block (LB), wavefront collisions (WFC), pivot sites (PS), and gaps. EGMs collected at these predefined areas were analyzed in terms of amplitude, duration, and morphology.

      Results

      Twenty-five atrial maps with 195 sites of interest (1755 EGMs) were object of our analysis. Thirty-five percent were sites of SC: fractionation had low amplitude (0.16 ± 0.07 mV) and long duration (87.8 ± 10.7 ms); wavefront collisions were seen in 38% of sites with EGMs shorter in duration (46.5 ± 4.5 ms) and higher in voltage (0.58 ± 0.13 mV); 17% were lines of block, never responsible for fractionation (0.13 ± 0.05 mV; 122.4 ms ± 24.8 ms); 9% were PS with a high degree of fractionation (0.55 ± 0.15 mV; 85.8 ± 7.9 ms). Two gaps were identified (1%) with a low degree of fractionation.

      Conclusion

      Specific EGM characteristics in atrial tachycardia can be reproducibly linked to electrophysiological mechanisms. High-voltage and short-duration EGMs are associated with collision sites and PS that are unlikely to form critical sites for ablation; long-duration, low-voltage EGMs are associated with SC. However, not all SC regions will lie within the critical circuit and identification by only EGM characteristics cannot guide ablation.

      Keywords

      Background

      Since the inception of invasive electrophysiology, electrograms (EGMs) recorded during mapping of arrhythmias have always provided information about the electrical status of the underlying myocardium. Pathological processes as well as dynamic (not substrate-related) mechanisms
      • Jadidi A.S.
      • Duncan E.
      • Miyazaki S.
      • et al.
      Functional nature of electrogram fractionation demonstrated by left atrial high-density mapping.
      may render any EGMs complex. In the early stages there was some doubt whether fractionated EGMs were the result of artifact,
      • Josephson M.E.
      • Wit A.L.
      Fractionated electrical activity and continuous electrical activity: fact or artifact?.
      but it has subsequently become clear that EGM complexity results from many factors: far-field signal from adjacent structures, anisotropy, alterations in conduction velocity, asynchronous conduction, and meandering rotors.
      • de Bakker M.
      • Wittkampf F.H.
      The pathophysiologic basis of fractionated and complex electrograms and the impact of recording techniques on their detection and interpretation.
      While fractionated potentials can identify areas of slow conduction (SC) in regions of scar in the ventricles,
      • Wiener I.
      • Mindich B.
      • Pitchon R.
      Determinants of ventricular tachycardia in patients with ventricular aneurysms: results of intraoperative epicardial and endocardial mapping.
      • Anter E.
      • Tschabrunn C.M.
      • Buxton A.E.
      • Josephson M.E.
      High-resolution mapping of postinfarction reentrant ventricular tachycardia: electrophysiological characterization of the circuit.
      in atrial fibrillation (AF) different mechanisms have been shown to result in EGM fractionation,
      • Nademanee K.
      • McKenzie J.
      • Kosar E.
      • Schwab M.
      • Sunsaneewitayakul B.
      • Vasavakul T.
      • Khunnawat C.
      • Ngarmukos T.
      A new approach for catheter ablation of atrial fibrillation: mapping of the electrophysiologic substrate.
      arising from active and passive mechanisms
      • Hunter R.J.
      • Diab I.
      • Tayebjee M.
      • Richmond L.
      • Sporton S.
      • Earley M.J.
      • Schilling R.J.
      Characterization of fractionated atrial electrograms critical for maintenance of atrial fibrillation: a randomized, controlled trial of ablation strategies (the CFAE AF trial).
      : for AF not all complex atrial fragmented electrogram (CFAE) sites are critical for its maintenance. It has also been showed that during AF, CFAE can be induced and maintained not only by changes in the local electrophysiological properties but also by the activation of local and distant autonomic neural elements.
      • Lin J.
      • Scherlag B.J.
      • Zhou J.
      • Lu Z.
      • Patterson E.
      • Jackman W.M.
      • Lazzara R.
      • Po S.S.
      Autonomic mechanism to explain complex fractionated atrial electrograms (CFAE).
      However, correlating activation fronts to fractionation in AF remains challenging because of the chaotic nature of the arrhythmia. On the contrary, stable activation patterns, as observed in atrial tachycardias (ATs), may offer the chance to study more closely atrial activation patterns at sites of EGM fractionation.
      The definition of fractionation is difficult because of varying mechanisms and recording strategies. Multiple definitions of fractionation have been suggested in AF and in organized rhythms, leading some authors to believe that a universal definition of fractionation may not even be possible.
      • Van der Does L.J.
      • de Groot N.M.
      Inhomogeneity and complexity in defining fractionated electrograms.
      • Lellouche N.
      • Buch E.
      • Celigoj A.
      • Siegerman C.
      • Cesario D.
      • De Diego C.
      • Mahajan A.
      • Boyle N.G.
      • Wiener I.
      • Garfinkel A.
      • Shivkumar K.
      Functional characterization of atrial electrograms in sinus rhythm delineates sites of parasympathetic innervation in patients with paroxysmal atrial fibrillation.
      The use of small electrode spacing for bipolar recordings seems to be important, and a definition of fractionation of >4 deflections in the bipolar EGM compares favorably with other definitions and has been widely used.
      • Lellouche N.
      • Buch E.
      • Celigoj A.
      • Siegerman C.
      • Cesario D.
      • De Diego C.
      • Mahajan A.
      • Boyle N.G.
      • Wiener I.
      • Garfinkel A.
      • Shivkumar K.
      Functional characterization of atrial electrograms in sinus rhythm delineates sites of parasympathetic innervation in patients with paroxysmal atrial fibrillation.
      • Fukumoto K.
      • Habibi M.
      • Ipek E.G.
      • et al.
      Association of left atrial local conduction velocity with late gadolinium enhancement on cardiac magnetic resonance in patients with atrial fibrillation.
      The aim of our study was to relate the EGM characteristics (including amplitude, duration, and fractionation) to specific activation patterns during AT using high-density mapping.

      Methods

      Twenty-five consecutive patients with symptomatic drug-refractory AT were considered for the analysis. The majority of patients (94%) had undergone ablation for AF. Antiarrhythmic drugs were discontinued at least 5 half-lives before the procedure. Patients were anticoagulated for at least 1 month before the procedure and underwent transesophageal echocardiography and/or computed tomography scan within 48 hours before the procedure to exclude the presence of thrombus. An electrophysiology study was performed using conscious sedation introducing the following catheters via the right femoral vein: (1) a decapolar catheter (Xtreme, Sorin, Livanova, Clamart, France, or DYNAMIC XT, Boston Scientific, Marlborough, MA, USA) within the coronary sinus, (2) a 3.5-mm open irrigated–tip ablation catheter (ThermoCool SF catheter, Biosense Webster, Diamond Bar, CA), and (3) an Orion mapping catheter (Boston Scientific, Marlborough, MA) used to generate left or right atrium geometry and to collect bipolar EGMs for AT activation maps. A steerable long sheath (Agilis, St. Jude Medical, or Zurpaz, Boston Scientific) were used to aid catheter manipulation in both chambers. Before introducing the mapping catheter, heparin (100 IU/kg) was administered to achieve an activated clotting time of >300 seconds. Every patient had given written informed consent before the study. The study was approved by the local ethics committee on human research at our institution.

      Description of the mapping system and recording catheter

      Rhythmia (Boston Scientific) was used for data collection. The Orion catheter is an 8.5-F, bidirectional, basket-shaped catheter, with 8 electrodes printed on each of 8 splines for a total of 64 electrodes. It is used to create a shell of the chamber of interest and simultaneously record EGMs at very high density and with low noise. Electrodes are printed onto the surface of the splines, having an area of 0.4 mm2 and a center-to-center spacing of 2.5 mm. Bipolar signals were filtered at 30 and 300 Hz, without a 50-Hz notch filter.

      Map acquisition

      Bipolar EGMs were automatically collected during AT with the following beat acceptance criteria: (1) cycle length stability (±10 ms), (2) propagation ΔR (difference in time in activation between 2 electrodes of the coronary sinus catheter), (3) respiration phase allowing data acquisition at a constant respiratory phase, (4) catheter stability, and (5) tracking quality to ensure correct location of the mapping catheter. Maps and EGMs were acquired only through the Orion. Analysis of all EGMs was performed off-line on the Rhythmia system by using electronic calipers at a speed of 100 mm/s. For our study, only the initial map and relative EGMs were considered for the analysis before any radiofrequency (RF) ablation.

      EGM characterization

      EGM fractionation over areas of interest was defined as ≥4 deflections lasting >40 ms. To define normal EGM, we collected 250 EGMs per patient (total 2500 EGMs) outside the areas of interest to act as a reference (Table 1).
      Table 1Representation of EGM characteristics of normal areas and of the electrophysiological mechanisms identified
      CharacteristicNormal areasSlow conductionWavefront collisionLines of blockPivot sitesGaps
      Pivot pointsFriction
      n (%)25067 (35%)76 (38%)33 (17%)17 (9%)2 (1%)
      Mean number per patient102.731.30.70.1
      Mean amplitude (mV)1.02 ± 0.390.16 ± 0.070.58 ± 0.130.13 ± 0.050.59 ± 0.130.34 ± 0.070.21 ± 0.11
      Mean duration (ms)31.9 ± 3.087.8 ± 10.746.5 ± 4.5122.4 ± 24.869.5 ± 8.689.8 ± 9.898 ± 11.3
      EGM morphologyBiphasic potential of short durationLong duration with low-amplitude fractionation, voltage- dependentShort duration and high amplitude, double/triple deflections. Low degree of fractionationDouble potentials separated by the isoelectric line, with no fractionationSingle-component EGM, good in amplitude with multiple, fast deflectionsDouble component EGM, lower in amplitude, with fractionation on the second componentNarrow-spaced potentials with fractionation in between
      EGM = electrogram.

      Electrophysiological mechanisms

      During the off-line analysis, regions of interest were selected on the basis of abnormal activation patterns and these were (1) SC, (2) lines of block (LB), (3) pivot sites (PS), (4) wavefront collisions (WFC), and (5) gaps (Figure 1). The phenomena were visually identified and analyzed by 2 electrophysiologists blinded to each other’s analysis. There were no discrepancies between the 2 blinded reviewers. We excluded regions where wavefront propagation proceeded uniformly at a normal conduction velocity. Where paced or sinus rhythm maps were available, EGMs were compared between maps in regions of nondynamic phenomena (ie, LB, gap, and SC). WFC and PS were not compared, as these phenomena are dependent on the direction of wavefront activation and do not appear in sinus or paced rhythm.
      Figure thumbnail gr1
      Figure 1Schematic of the electrophysiological mechanisms identified. Timing of conduction is from the patient, representative of each mechanism (in milliseconds).

      Slow conduction zones

      These were identified by a deceleration of the wavefront on the activation map (Figure 1 and example in Figure 2). The conduction velocities were calculated on the Rhythmia suite from the difference in timing and the known distance between recording points of the entrance and exit of the SC area. We recorded a mean conduction velocity over healthy tissue, confirming the normal conduction velocity >1 m/s recorded by Fukumoto et al.
      • Fukumoto K.
      • Habibi M.
      • Ipek E.G.
      • et al.
      Association of left atrial local conduction velocity with late gadolinium enhancement on cardiac magnetic resonance in patients with atrial fibrillation.
      We therefore defined as SC all areas where the wavefront velocity was half less than the normal velocity (0.5 m/s).
      Figure thumbnail gr2
      Figure 2Example of slow conduction occurred along the posterior wall of the left atrium just below the left inferior pulmonary vein. A: The wavefront travels at normal velocity (wide dark-red front) from the anterior wall crossing the ridge between the left appendage and the left upper vein (5, 15, and 27 ms). B: The wavefront is crossing the area of slow conduction evidenced by the narrow dark red front with significant decrease in speed (activation time 75 ms). C: The wavefront regains speed (105 ms).
      LB were defined as areas where activation was completely stopped, with the front making a detour around the obstacle, the downstream activation proceeding toward the LB being in an opposite direction to the upstream one (Figure 1).
      WFC was identified by simultaneous arrival of 2 opposing wavefronts on the propagation map (Figure 1 and example in Figure 3A). EGMs were collected from each advancing wavefront and at the site of collision.
      Figure thumbnail gr3
      Figure 3A: Representation of wavefront collision phenomenon. B: Visual depiction of the friction site where the activation of 2 wavefronts occur in opposing directions. C: Description of activation occurring around a pivot point. Bottom left: Craniocaudal activation of the posterior wall with an area of slow conduction starting from the left inferior pulmonary vein causes increasing curvature of the advancing wavefront, which rotates around a pivot point to collide at a line of functional block caused by refractoriness from the initial craniocaudal activation. There is therefore caudocranial activation along the same area of slow conduction.
      PS were discrete zones where the wavelets makes a U turn around the end of the arcs of functional block or true block such as SC/LB defining the pivot point (Figure 1 and example in Figure 3C). The resulting activation along the LB/SC of the 90°–180° diverging fronts was defined as “friction area” (Figure 3B). For the EGM analysis, we considered the 2 as distinct zones: the pivot point and the “friction” area.
      Conduction gaps were defined as zones of preserved conduction across LB usually from previous RF ablation, the activation wavefront being able to break through the line (Figure 1 and example in Figure 4).
      Figure thumbnail gr4
      Figure 4Left anterior oblique projection of the left atrium. A wavefront coming from the anterior wall get arrested at the previous mitral isthmus line (A). A gap allows the passage of the wavefront (B), which finally invades and activates the posterior wall (C).
      For each area of interest, the morphology, EGM amplitude (in millivolts), and duration (in milliseconds) were collected. Only bipolar recordings were included for analysis. Data were collected over a mean value of at least 10 EGMs per each area of interest that should have a density of >50 points to validate and ensure the accuracy of the data recorded. Conduction velocity was calculated for SC areas. Activation maps were compared with voltage maps by using a cutoff value of <0.05 mV to define scar tissue and >0.5 mV for healthy tissue. Voltage between 0.05 and 0.5 mV was defined as border zone. For the analysis, we divided the atria into several regions in order to classify phenomena by site. To facilitate the description, the left atrium (LA) was divided into roof, left pulmonary vein (PV) area (including the mitral isthmus), anterior wall, posterior wall, and right PV area (including the septum). The right atrium was divided into free wall, septum, and cavotricuspid isthmus region.

      Statistical analysis

      Software R 3.4.0 (R Foundation for Statistical Computing, Vienna, Austria) and packages lme4 and pROC
      R Development Core Team
      R: A language and environment for statistical computing.
      • Bates D.
      • Mächler M.
      • Bolker B.
      • Walker S.
      Fitting linear mixed-effects models using lme4.
      were used. EGM summary statistics are presented as mean ± SD on a per-region, per-patient basis. The effects of the electrophysiological mechanisms on EGM parameters were assessed by a linear mixed model, with the patient identifier included as a random effect, to account for within-patient effects. R2 values for correlation of nonlinearly distributed values were obtained using a linear least-squares fit of log-transformed data. P values <.05 were considered statistically significant. Sensitivity and specificity were determined by receiving operating characteristic analysis.

      Results

      Study population

      Twenty-five patients were enrolled, of which 19 were men (76%) with a mean age of 63.5 ± 8.0 years. The mean number of collected beats per map was 2069 ± 1079, with a mean number of 16,599 ± 9055 EGMs. The mean procedure time was 178 ± 43 minutes. The mean fluoroscopy time was 58.5 ± 13.0 minutes. There were no procedural complications.

      AT characteristics

      The LA was the main site of reentry circuit or focal activity, occurring in 22 patients (88%). Our study was conducted on the first atrial activation map: 18 ATs were macroreentry circuits (72%), 5 localized reentry circuits (20%), and 2 focal ATs (8%).

      Electrophysiological mechanisms identified

      A total number of 195 areas of interest were identified during off-line analysis. A total of 1755 EGMs were analyzed and characterized. Sixty-seven were areas of SC (35%), 33 LB (17%), 76 WFC sites (38%), 17 PS (9%), and 2 conduction gaps (1%) (Table 1). Sinus or paced rhythm maps were available in 5 patients. LB present in ATs were confirmed in all cases, and similarly gaps were also 100% concordant. Eighty-four percent of SC regions also had SC in sinus or paced rhythm.

      Slow conduction zones

      Sixty-seven sites were identified (mean number 2.7 per map). Conduction velocity was measured at 0.13 ± 0.05 m/s, with a mean EGM duration of 87.8 ± 10.7 ms and a mean amplitude of 0.16 ± 0.07 mV.
      Three classes of SC could be identified on the basis of local conduction velocity (≤0.1 m/S, >0.1 to ≤0.2 m/S, >0.2 m/S). The first class (≤0.1 m/S) had an EGM with low amplitude (mean 0.07 ± 0.05 mV) and long duration (mean 121 ± 11 ms) (Figure 5, site 1). EGM morphology was characterized by fine fractionation with smooth onset/offset and gradual termination into the isoelectric line. This pronounced class of SC was characteristic of scar areas (<0.05 mV at bipolar voltage) with a mean conduction velocity over these areas of 0.08 ± 0.02 m/s and accounted for the 49% of SC identified.
      Figure thumbnail gr5
      Figure 5Slow conduction and relative electrograms recorded over the 3 sites of scar, border zone, and boundaries toward healthy tissue (bipolar voltage set from 0.05 to 0.5 mV).
      The second class (>0.1 ≤0.2 m/S) had an EGM with higher amplitude (mean 0.17 ± 0.06 mV) and shorter duration (mean 96 ± 9 ms) (Figure 5, site 2). EGM morphology was characterized by clear, sharper deflections. The mean conduction velocity was 0.13 ± 0.03 m/s. This type of SC was observed in 45% of cases and occurred along boundaries of scar.
      The third class had an EGM characterized by higher amplitude (mean 0.32 ± 0.11 mV) and shorter duration (mean value 82 ± 9 ms). EGM morphology was characterized by clear multiple deflections preceded by fine fractionation (Figure 5, site 3). The mean conduction velocity was 0.29 ± 0.01 m/s. This class accounted for 6% of cases and always occurred along the border with healthy zones.
      An inverse relationship between duration and voltage was documented (R2 = 0.77) (Figure 6). Mostly, the SC areas were recorded at the antrum of PVs, on areas of previous RF ablation (54%).
      Figure thumbnail gr6
      Figure 6A: Inverse relationship between duration and voltage for slow conduction phenomenon: the higher the voltage, the shorter the duration. B: The 3 phenomena (slow conduction, pivot sites, and wavefront collisions) are plotted in different areas of the graph. EGM = electrogram.

      Lines of block

      Thirty-five LB were identified with a mean number of 1.3 per patient. The mean amplitude was 0.13 ± 0.05 mV, with a mean duration of 122.4 ± 24.8 ms. In 51% of cases, LB were localized around the PVs. The signals recorded at these sites were made of 2 EGMs of inverse polarity, well separated by an isoelectric line. EGMs at these sites never generated fragmentation, and there was no significant relationship between voltage and duration.

      Wave front collision

      This phenomenon occurred in 38% of sites (mean number 3 per map). The corresponding signals showed limited fractionation with a mean duration of 46.5 ± 4.5 ms and a mean amplitude of 0.58 ± 0.13 mV. The morphology of the EGMs consisted of an Rr′ pattern; RR′ pattern or triple deflection when the collisional ways occurred in healthy tissue. This pattern resulted from the 2 waves arriving from each side and producing a sum of 2 R waves (Figure 3A). Conduction velocity at the sites adjacent to the collision were normal (>1 m/s). Most WFC (84%) were observed in healthy tissue, in the septum and anterior wall (54%).

      Pivot and friction sites

      Seventeen PS were identified (mean 0.7 per map). The mean EGM duration at those sites was 85.8 ± 7.9 ms, with a mean amplitude of 0.55 ± 0.15 mV. For the analysis, as described in the Methods, we distinguished “pivot points” where EGMs had higher mean amplitude (0.59 ± 0.13 mV vs 0.34 ± 0.07 mV; P < .001) and shorter duration (69.5 ± 8.6 ms vs 89.8 ± 9.8 ms; P < .001) when compared with EGMs at the “friction” areas. The morphology of EGMs varied from pivot points to “friction” areas. At pivot points, where the wavefront turned, the EGM had a single-component signal characterized with multiple deflections (Figure 7). Furthest from the pivot point, where the “friction” area started, the EGM was characterized by a distinct double potential with more fractionation on the second component. This phenomenon was recorded only in the LA, with no specific anatomical localization.
      Figure thumbnail gr7
      Figure 7Schematic of electrogram (EGM) characteristics at pivot sites. EGMs at a pivot point (1) are short in duration and high in amplitude and expressed by multiple deflections. On friction area (2 and 3), EGMs are represented by double potentials, longer in duration and with fractionation.

      Conduction gaps

      We identified 2 gaps with a duration of 96 and 101 ms and an amplitude of 0.21 and 0.23 mV. These were localized respectively at the LA roof and mitral isthmus as a consequence of previous attempts to create an LB. The main morphology of this EGM was a closely spaced double potential with a low degree of fragmentation in between.

      Patterns of fractionations

      EGM duration and amplitude differed significantly depending on the underlying electrophysiological mechanism (P < .001 for both parameters). Pathological mechanisms (PS, gaps, LB, and SC) can be distinguished from nonpathological mechanisms (normal regions and WFC) by EGM duration. EGM duration >63 ms has a 99% sensitivity and 99% specificity for the identification of pathological mechanisms. When nonpathological mechanisms are excluded, EGM voltage allows discrimination between mechanisms with delay (SC, LB, and gap) and pivot areas. EGM amplitude <0.3 mV has an 88% specificity and 97% sensitivity for the identification of mechanisms involving delay within pathological areas (Figure 6B). Gaps, LB, and SC could not be distinguished by EGM amplitude and duration. Within SC regions, there is an inverse relationship between EGM duration and voltage (R2 = 0.77; P < .001) (Figure 6A).

      Discussion

      This study demonstrates using high-density mapping for ATs that activation patterns have an EGM signature that can possibly be used to identify them. This study shows that (1) WFC are the most frequent phenomena associated with limited fractionation and (2) SC and PS are associated with the highest degree of fractionation, but PS have significantly higher voltages.

      Nature of EGM fractionation

      Our study demonstrates that EGMs are affected by activation patterns and tissue characteristics, which can be structural or functional in nature. It is likely that most SC areas are structural in nature in this study (previous RF lesions in this study) while collisions are functional and PS include both mechanisms. In SC areas we have demonstrated that reduced EGM voltage is associated with increased EGM duration and reduced conduction velocity.
      The most diseased zones (low-voltage scar) are associated with the highest degree of fractionation, longer duration EGMs, and slowest conduction velocity. SC was also seen in border zones and healthy tissue with greater EGM amplitude and shorter EGM duration reflecting faster, albeit still slow, wavefront propagation. The direction of wavefront propagation may also explain the imperfect concordance between regions of SC in ATs and sinus or paced rhythm.
      Fibrosis has been confirmed as a major cause of SC and subsequent EGM fractionation
      • Jacquemet V.
      • Henriquez C.S.
      Genesis of complex fractionated atrial electrograms in zones of slow conduction: a computer model of microfibrosis.
      : EGMs were more fractionated with a higher number of deflections when microfibrosis was more severe (higher density or longer collagenous septa). The mean number of deflections was directly proportional to the transverse conduction velocity.
      EGM detection and characterization of small amplitude signals (<0.1 mV) was facilitated in our study because of the inherent characteristics of the Orion catheter and its small, printed 0.4 mm2 interelectrode distance. Electrode parameters are of critical importance in identifying and localizing fractionation.
      • Van der Does L.J.
      • de Groot N.M.
      Inhomogeneity and complexity in defining fractionated electrograms.
      The small, printed, closely spaced electrodes of the Orion catheter provide a unique resolution compared to previously available mapping catheters, which have electrodes of 3.5–4 mm diameter and spacing of up to 6 mm.
      WFC are phenomena that are, at least partly, functional and dynamic, occurring anywhere in the atria. The resulting fractionation was minimal and always associated with EGMs of short duration (mean duration 46.5 ms). Interestingly, EGM morphology was dependent on voltage: if voltage was >0.5 mV, collisions always generated an EGM with double/triple deflections, while collisions in low-voltage area (<0.5 mV) had limited or no fractionation. No collision was observed in scar area. The collision has been studied previously by our group
      • Jadidi A.S.
      • Duncan E.
      • Miyazaki S.
      • et al.
      Functional nature of electrogram fractionation demonstrated by left atrial high-density mapping.
      : WFC (71%) and regional SC (24%) mostly caused EGM fractionation. Our study is in keeping with those results, showing that WFC are the most common mechanisms and that sites usually display normal bipolar voltages, suggesting the absence of myocardial fibrosis or scar. The collision is therefore linked to a functional and dynamic mechanism rather than substrate and characterized by a short-duration, single-component EGM.
      PS were less frequently observed (17%). At those sites we have recognized 2 distinct EGM morphologies: (1) at pivot points we documented higher fractionation, expressed by multiple deflections, and short duration; (2) at sites of “friction,” EGMs were lower in amplitude and higher in duration (P = .04). At the pivot points we may speculate that EGM fractionation is largely functional because of anisotropy
      • Cabo C.
      • Pertsov A.M.
      • Baxter W.T.
      • Davidenko J.M.
      • Gray R.A.
      • Jalife J.
      Wave-front curvature as a cause of slow conduction and block in isolated cardiac muscle.
      and wavefront curvature.
      • Spach M.S.
      • Miller III, W.T.
      • Dolber P.C.
      • Kootsey J.M.
      • Sommer J.R.
      • Mosher Jr., C.E.
      The functional role of structural complexities in the propagation of depolarization in the atrium of the dog: cardiac conduction disturbances due to discontinuities of effective axial resistivity.
      The wavefront curvature is responsible for a 90°–180° turn around a lesion. So the resulting EGMs at the pivot point are likely to be constructed signals, depending on the characteristics of the tissue (anisotropy), the propagation velocity of different wavefronts, curvature of the wavefront, and transverse conduction velocity. EGMs at friction sites depend more on tissue characteristics, since these originate from areas of SC (because of previous RF lesion or functional obstacle) and transverse activation. EGMs at those sites, in fact, are found to be longer in duration. Further studies should investigate this phenomenon.
      Gaps are a phenomenon that may occur in any patient in whom linear ablation has been previously performed, although this was rarely found in our study. During conventional mapping along LB, low-degree fractionation consisting of 2 closely spaced potentials may represent a gap and be a target for ablation. High-density mapping has the potential to describe gap potentials in greater detail that has been possible in previous studies.
      • Matsuo S.
      • Yamane T.
      • Hioki M.
      • et al.
      Identification of a conduction gap of the mitral isthmus by using a novel high-density mapping catheter.
      LB, characterized by distinct EGMs separated by an isoelectric line, are not expected to produce complex EGMs, because those mapping points correspond to areas of previous transmural RF ablation and true iatrogenic scar. Wavefronts do not cross these lines but arrive at different times, producing the classic double atrial potentials separated by an isoelectric line.

      Significance of complex fractionated EGMs

      Patterns of activation and morphology during AF were previously described by Konings et al
      • Konings K.T.
      • Smeets J.L.
      • Penn O.C.
      • Wellens H.J.
      • Allessie M.A.
      Configuration of unipolar atrial electrograms during electrically induced atrial fibrillation in humans.
      using epicardial recordings in patients without a history of AF; these authors showed that fractionation of unipolar EGM during induced AF occurred at sites of (1) SC, (2) PS, (3) asynchronous activation, and (4) collision. Our results are in agreement with these findings, confirming the correlation between fractionation and the underlying mechanism despite being recorded in a different rhythm and using bipolar recordings.
      High-density helps mapping the most complex EGMs that were suboptimally described so far. These fractionated EGM scans either be actively involved in the arrhythmia circuit or be bystander phenomena. This raises the question whether EGM fractionation should always be targeted during catheter ablation. In AF, Nademanee et al
      • Nademanee K.
      • McKenzie J.
      • Kosar E.
      • Schwab M.
      • Sunsaneewitayakul B.
      • Vasavakul T.
      • Khunnawat C.
      • Ngarmukos T.
      A new approach for catheter ablation of atrial fibrillation: mapping of the electrophysiologic substrate.
      was the first to propose these areas with high-frequency, disorganized activity as ablation target sites with a respectable success rate. However, there are several limitations to CFAE ablation, either in AF or in AT. The interpretation is largely variable, and even important is the fact that fragmented EGMs are the common phenotypes for different electrophysiological mechanisms, most of them being possibly passive. Our study improves the understanding of complex signals by describing the fractionation that commonly arises from collision and PS and SC areas.

      Clinical significance

      The increased spatial density of this novel mapping system allows to annotate automatically most of the components of a complex signal in a small region, thus better describing the local complex activation. Accordingly, 3 main types of fractionated atrial EGMs were clinically relevant: (1) collisions typically have EGMs of shorter duration and higher voltage; (2) PS typically have EGMs of longer duration and higher amplitude; and (3) SC have EGMs of longer duration and low amplitude that are always fragmented. Further investigation of SC and its role in AT circuits is critical to understand the substrate in these patients and helps to identify the best targets for ablation. Our research shows that during stable rhythms, areas of SC could be identified by a certain EGM signature; these may represent optimal ablation sites in an active part of the tachycardia circuit, but it is not possible to discern which area of SC represents the critical isthmus from bystander areas by EGM characteristics alone. In addition, these findings may provide insights into EGM fractionation in AF. The ablation of CFAE as a therapeutic strategy of AF ablation remains controversial, and a greater understanding of the underlying mechanisms and substrates may allow more refined ablation strategies.

      Study limitations

      Our study examines only bipolar EGMs that affect the amplitude of the signal, depending on the direction of the wavefront. The Orion catheter can be difficult to manipulate, particularly in patients with transseptal access via a parent foramen ovale. Further the catheter electrodes can fail, resulting in artifact. However, these failures are promptly recognized by the system and are excluded from mapping and data collection. Conduction velocity as assessed in this study is an approximation, as it is impossible to get access to cell orientation in a clinical environment. The retrospective, unblinded nature of this study may have influenced the collection of data. Furthermore, the results of this study have been observed in post-AF ablation ATs using the Orion catheter. Whether these results can be extended to other arrhythmia or mapping catheters remains to be demonstrated.

      Future directions

      Understanding the nature of complex EGMs is the key of the treatment of arrhythmias. Further investigations in the mapping of stable rhythm could improve our understanding of complex arrhythmias such as AF. Three-dimensional structural analysis may allow further insights into the complex interaction of these phenomena and lead to greater understanding of their role in fractionation. Separation of pathological functional mechanisms from normal ones may be better attempted in sinus rhythm as well, with pacing maneuvers from different sites. These results may help improve AF ablation targeting fractionation.

      Conclusion

      During AT, EGM fractionation occurs with characteristics linked to the underlying activation pattern. High-voltage, short-duration EGMs are associated with WFC and pivot points that are unlikely to form critical sites for ablation, whereas long-duration, low-voltage EGMs are associated with SC. Not all SC regions will lie within the critical circuit, so EGM characteristics cannot guide ablation in isolation.

      References

        • Jadidi A.S.
        • Duncan E.
        • Miyazaki S.
        • et al.
        Functional nature of electrogram fractionation demonstrated by left atrial high-density mapping.
        Circ Arrhythm Electrophysiol. 2012; 5: 32-42
        • Josephson M.E.
        • Wit A.L.
        Fractionated electrical activity and continuous electrical activity: fact or artifact?.
        Circulation. 1984; 70: 529-532
        • de Bakker M.
        • Wittkampf F.H.
        The pathophysiologic basis of fractionated and complex electrograms and the impact of recording techniques on their detection and interpretation.
        Circ Arrhythm Electrophysiol. 2010; 3: 204-213
        • Wiener I.
        • Mindich B.
        • Pitchon R.
        Determinants of ventricular tachycardia in patients with ventricular aneurysms: results of intraoperative epicardial and endocardial mapping.
        Circulation. 1982; 65: 856-861
        • Anter E.
        • Tschabrunn C.M.
        • Buxton A.E.
        • Josephson M.E.
        High-resolution mapping of postinfarction reentrant ventricular tachycardia: electrophysiological characterization of the circuit.
        Circulation. 2016; 134: 314-327
        • Nademanee K.
        • McKenzie J.
        • Kosar E.
        • Schwab M.
        • Sunsaneewitayakul B.
        • Vasavakul T.
        • Khunnawat C.
        • Ngarmukos T.
        A new approach for catheter ablation of atrial fibrillation: mapping of the electrophysiologic substrate.
        J Am Coll Cardiol. 2004; 43: 2044-2053
        • Hunter R.J.
        • Diab I.
        • Tayebjee M.
        • Richmond L.
        • Sporton S.
        • Earley M.J.
        • Schilling R.J.
        Characterization of fractionated atrial electrograms critical for maintenance of atrial fibrillation: a randomized, controlled trial of ablation strategies (the CFAE AF trial).
        Circ Arrhythm Electrophysiol. 2011; 4: 622-629
        • Lin J.
        • Scherlag B.J.
        • Zhou J.
        • Lu Z.
        • Patterson E.
        • Jackman W.M.
        • Lazzara R.
        • Po S.S.
        Autonomic mechanism to explain complex fractionated atrial electrograms (CFAE).
        J Cardiovasc Electrophysiol. 2007; 18: 1197-1205
        • Van der Does L.J.
        • de Groot N.M.
        Inhomogeneity and complexity in defining fractionated electrograms.
        Heart Rhythm. 2017; 14: 616-624
        • Lellouche N.
        • Buch E.
        • Celigoj A.
        • Siegerman C.
        • Cesario D.
        • De Diego C.
        • Mahajan A.
        • Boyle N.G.
        • Wiener I.
        • Garfinkel A.
        • Shivkumar K.
        Functional characterization of atrial electrograms in sinus rhythm delineates sites of parasympathetic innervation in patients with paroxysmal atrial fibrillation.
        J Am Coll Cardiol. 2007; 50: 1324-1331
        • Fukumoto K.
        • Habibi M.
        • Ipek E.G.
        • et al.
        Association of left atrial local conduction velocity with late gadolinium enhancement on cardiac magnetic resonance in patients with atrial fibrillation.
        Circ Arrhythm Electrophysiol. 2016; 9: e002897
        • R Development Core Team
        R: A language and environment for statistical computing.
        (Vienna: R Foundation for Statistical Computing; Available at:) (Accessed September 14, 2017)
        • Bates D.
        • Mächler M.
        • Bolker B.
        • Walker S.
        Fitting linear mixed-effects models using lme4.
        J Stat Softw. 2015; 67: 1-48
        • Jacquemet V.
        • Henriquez C.S.
        Genesis of complex fractionated atrial electrograms in zones of slow conduction: a computer model of microfibrosis.
        Heart Rhythm. 2009; 6: 803-810
        • Cabo C.
        • Pertsov A.M.
        • Baxter W.T.
        • Davidenko J.M.
        • Gray R.A.
        • Jalife J.
        Wave-front curvature as a cause of slow conduction and block in isolated cardiac muscle.
        Circ Res. 1994; 75: 1014-1028
        • Spach M.S.
        • Miller III, W.T.
        • Dolber P.C.
        • Kootsey J.M.
        • Sommer J.R.
        • Mosher Jr., C.E.
        The functional role of structural complexities in the propagation of depolarization in the atrium of the dog: cardiac conduction disturbances due to discontinuities of effective axial resistivity.
        Circ Res. 1982; 50: 175-191
        • Matsuo S.
        • Yamane T.
        • Hioki M.
        • et al.
        Identification of a conduction gap of the mitral isthmus by using a novel high-density mapping catheter.
        Pacing Clin Electrophysiol. 2012; 35: 65-68
        • Konings K.T.
        • Smeets J.L.
        • Penn O.C.
        • Wellens H.J.
        • Allessie M.A.
        Configuration of unipolar atrial electrograms during electrically induced atrial fibrillation in humans.
        Circulation. 1997; 95: 1231-1241